cGMP is generated in endothelial cells after stimulation of soluble guanylyl cyclase (sGC) by nitric oxide (NO) or of particulate guanylyl cyclase (pGC) by natriuretic peptides (NP). We examined whether localized increases in cytosolic cGMP have distinct regulatory roles on the contraction induced by H2O2 treatment in human umbilical vein endothelial cells. cGMP concentrations and temporal dynamics were different upon NO stimulation of sGC or C-type NP (CNP) activation of pGC and did not correlate with their relaxing effects measured as planar cell surface area after H2O2 challenge. cGMP production due to sGC stimulation was always smaller and more brief than that induced by pGC stimulation with CNP, which was greater and remained elevated longer. The NO effects on cell relaxation were cGMP dependent because they were blocked by sGC inhibition with 1H-(1,2,4)Oxadiazolo(4,3-a)quinoxaline-1-one and mimicked by 8-Br-cGMP. An antagonist of the cGMP-dependent protein kinase type-I (PKG-I) also inhibited the NO-induced effects. The cell contraction induced by H2O2 produces myosin light chain (MLC) phosphorylation and NO prevented it completely, whereas CNP only produced a partial inhibition. Transfection with a dominant negative form of PKG type-Iα completely reversed the NO-induced effects on MLC phosphorylation, whereas it only partially inhibited the effects due to CNP. Taken together, these results demonstrate that the NO/sGC/cGMP pathway induces endothelial cell relaxation in a more efficient manner than does CNP/pGC/cGMP pathway, an effect that might be related to a selective stimulation of PKG-1α by NO-derived cGMP. Consequently, stimulated PKG-Iα may phosphorylate important protein targets that are necessary to inhibit the endothelial contractile machinery activated by oxidative stress.
- nitric oxide
- C-type natriuretic peptide
- myosin light chain
- cGMP-dependent protein kinase type Iα
- endothelial cell barrier dysfunction
vascular homeostasis depends on the ability of the endothelium to maintain its integrity, serving as a non-adherent nonthrombogenic surface and as a barrier that regulates the exchange of fluid and macromolecules between the blood and the extracellular tissue (9, 15). During vascular disease, endothelial cells are exposed to excess reactive oxygen species that can alter the endothelial cell phenotype by inducing several signaling pathways to generate second messengers that modulate the structure and organization of cytoskeletal proteins (20). These cytoskeletal alterations lead to changes in cell shape and the formation of paracellular gaps that impair the endothelial cell barrier function (2, 10, 36, 39).
Several laboratories, including ours, have demonstrated that reactive oxygen species can induce endothelial cell contraction, which could be responsible for the increased endothelial permeability that often accompanies ischemia/reperfusion injuries (5, 17, 18, 20). Agents that activate guanylyl cyclases (GC) have been shown to relax smooth muscle cells; similarly, cGMP-elevating agents seem to attenuate oxidant-induced endothelial cell barrier dysfunction in some vascular beds (6, 10, 21, 24).
cGMP is a second messenger involved in many physiological processes such as smooth muscle tone, neural excitability, epithelial electrolyte transport, phototransduction in the retina, and cell proliferation (26). Despite the enormous importance of cGMP in cell physiology, little attention is given to the fact that its formation is not uniformly distributed within the cell. cGMP can be formed from GTP by the action of two distinct GCs: a soluble form (sGC) and a particulate membrane-bound form (pGC) (19), each of which is activated by different agonists. Nitric oxide (NO) and NO donors activate sGC, whereas pGC is a plasma membrane receptor for natriuretic peptides and related hormones. The pathways that control cGMP levels are complex due to the existence of several ubiquitously expressed phosphodiesterases (PDE), which hydrolyze cGMP (4). Some PDE are soluble, whereas others are plasma membrane bound. The intracellular actions of cGMP are primarily mediated by cGMP-dependent protein kinases (PKG), but several types of cyclic nucleotide-activated ion channels also appear to be involved (7, 16).
Given the separate sources of cGMP within the cell, it is possible to conceive a functional compartmentalization of cGMP due to localized elevation of these second messengers within the cell, i.e., membrane and cytosol. The purpose of this study was to examine the endothelial cell-relaxing activities of the cytosolic and particulated pools of cGMP in human endothelial cells exposed to hydrogen peroxide (H2O2) and to determine the mechanism by which cGMP inhibits contractility. We previously demonstrated that NO and natriuretic peptides are able to reverse the contraction induced by H2O2 in bovine endothelial cells, an effect that was mimicked by a cGMP analogue and mediated in part by PKG (17). The present study extends those results and addresses the hypothesis that cGMP generated via activation of soluble and particulate GCs may have differential activities in promoting endothelial cell relaxation. Our study will help to clarify how stimulation of different receptors that act via the same second messenger can elicit the appropriate functional response.
MATERIALS AND METHODS
Materials. E-199 medium, fetal calf serum, l-glutamine, penicillin, streptomycin sulphate, and Hanks balanced salt solution were purchased from Biomedia (Boussens, France). Collagenase type IA from Clostridium histolyticum, H2O2, natriuretic peptide type C (CNP), sodium nitroprusside (SNP, NO donor), 8-Br-cGMP, 1H-(1,2,4)Oxadiazolo(4,3-a)quinoxalin-1-one (ODQ, sGC inhibitor), Rp-8-[(4-Clorophenil)tyo]-cGMPs triethylamine (Rp-cGMPs, PKG type I inhibitor), 3-isobutyl-1-methylxanthine (IBMX, PDE inhibitor), Zaprinast (PDE type-5 inhibitor), and monoclonal antimyosin (light chains) were obtained from Sigma Chemical (St. Louis, MO). Spermidine-NONOate (sp-NO, NO donor) was purchased from Alexis Biochemicals (San Diego, CA). Eagle's modified essential medium (EMEM) was from BioWhittaker (Walkersville, MD). Pansorbin was obtained from Calbiochem (La Jolla, CA). cGMP RIA kit and [32P]orthophosphate were purchased from Amersham Pharmacia (Chalfont St. Giles, UK). Vasodilator-stimulated phosphoprotein (VASP) and phosphorylated VASP (P-VASP) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Plasmids. Flag-tagged PKG type Iα-regulatory region (fcGK-IαR), which acts as a dominant negative mutant when overexpressed in cells, was a kind gift from Dr. D. Browning (3). pcDNA 3.1 plasmid, obtained from Invitrogen (Carlsbad, CA), was used as a control plasmid in the transfection experiments.
Cell culture. Human endothelial cells from umbilical vein (HUVEC) were obtained and cultured as described previously (8). Cells were seeded on dishes coated with 0.2% gelatin at 37°C in a humidified atmosphere of 95% O2-5% CO2. Individual clones were established and subcloned to obtain pure cell populations. Clones were characterized by their typical cobblestone morphology, by the presence of factor VIII-related antigen, and by the uniform uptake of fluorescent acetylated low-density lipoprotein, as described (13). Cells were fed every 2 days with E-199 medium supplemented with 20% fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 20 mM HEPES, and 300 μg/ml endothelial cell growth factor. Cells were passaged when confluence was reached with trypsin-EDTA. Toxicity was evaluated in every experimental condition by the trypan blue dye exclusion method.
Measurement of planar cell surface area. Cells were grown at low density in 20-mm plates and studied before confluence. In every experiment, cells were washed twice, discarding the culture medium, and placed in buffer A (20 mM Tris, 130 mM NaCl, 5 mM KCl, 10 mM sodium acetate, 5 mM glucose, pH 7.45) containing 2.5 mM Ca2+ and maintained at room temperature. After 15 min of temperature equilibration, the experiments were started.
In the first group of experiments, the cells were preincubated with buffer, 1 μM sp-NONOate, or 0.1 μM CNP for 5 min. H2O2 (100 μM) was subsequently added to each treatment. Microphotographs were taken before H2O2 addition (time 0) and 30 min after this addition (time 30).
In the second group, cells were preincubated with buffer, 10 μM 8-Br-cGMP (cGMP analogue), sp-NONOate, or CNP. Five minutes later, H2O2 was added and the experiment was carried out as described previously.
In the third group of experiments, cells were preincubated with buffer, 1 μM ODQ (sGC inhibitor), or 1 μM Rp-cGMPs (PKG type I inhibitor) for 5 min. Then, sp-NONOate was added and cells were incubated for an additional 5 min, after which H2O2 was added. Microphotographs were taken as described previously, just before H2O2 addition and 30 min afterwards.
During each experiment, cells were observed under phase contrast with an inverted photomicroscope (Olympus IMT 2, Tokyo, Japan) with a ×150 magnification. Photographs of the same cells were taken under the experimental conditions cited above. Every cell with a sharp margin suitable for the planimetric analysis was considered, and 6-12 cells were analyzed per photograph. Planar cell surface area (PCSA) was determined by computer aid planimetric techniques (17, 35). Measurements were performed by two different researchers in a blind fashion. The intraobserver and interobserver variations were 2 and 5%, respectively.
Measurement of cGMP synthesis by HUVEC. Cells were washed twice with buffer A. Cells were then preincubated in the same buffer containing 2.5 mM Ca2+ at room temperature. Reactions were started after addition of the reagents as indicated previously. At different intervals from 30 s to 30 min, the medium was aspirated and 1 ml of ice-cold ethanol was added to the plates, which were maintained at 4°C for 30 min. Cell extracts were centrifuged for 20 min at 2,000 g, the supernatant fraction was evaporated to dryness, and cGMP levels were determined with the use of a commercial [125I]cGMP radioimmunoassay kit as described (28). Protein concentration in the pellets was determined according to the Bradford method.
Measurement of myosin light chain phosphorylation. Phosphorylation of the myosin light chain (MLC) was determined after immunoprecipitation and protein separation by SDS-polyacrylamide gel electrophoresis, as reported previously (34). Briefly, after the cells were labeled with 50 μCi/ml of neutralized, carrier-free sodium [32P]orthophosphate (3 h, 37°C), incubations were performed under the conditions detailed elsewhere (see figure legends). Thereafter, the incubation medium was removed and cells were precipitated with ice-cold ethanol. After the proteins were solubilized with a pyrophosphate buffer (100 mM NaF, 8 mM sodium pyrophosphate, 250 mM NaCl, 5 mM EDTA, 10 mM EGTA, 1 mM phenylmethylsulphonyl fluoride, 50 μg/ml leupeptin, and 1% Nonidet P-40), the samples were centrifuged. The supernatants were collected and incubated with human antiplatelet myosin antibody at 4°C for 90 min, and Pansorbin was used to precipitate the immunolinked MLC. This fraction was separated by 12% SDS-polyacrylamide gel electrophoresis, and the gel was frozen and exposed to X-OMAT films. The phosphorylated MLC was identified on the autoradiographs, and the absorbance of the 20-kDa band was measured by densitometry. Results were calculated in arbitrary density units and corrected for the protein concentration in the sample.
Transient transfection experiments. HUVEC were plated at 65% confluency on either 100-mm dishes or six-well plates. The cells were transfected to express exogenous DNA using LipofectAMINE (Invitrogen) according to the manufacturer's instructions. Subconfluent cell cultures were transfected with 2 μg of the fcGK-IαR plasmid, which expresses a dominant negative form of PKG type Iα. The pcDNA3.1 plasmid DNA was used as a control. Transfection was performed for 6 h, after which regular medium was added. Cells were treated with the corresponding reagents 24 h after transfection.
Western blotting. HUVEC were washed briefly in PBS and solubilized in lysis buffer (10 mM Tris · HCl, pH 7.4, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 500 nM sodium orthovanadate, 50 nM sodium fluoride, 1 μg/ml pepstatin/leupeptin/aprotinin, 1 mM phenylmethylsulfonyl fluoride) for 40 min at 4°C. Lysates were spun down for 5 min and the supernatants were collected. Protein concentration was determined by the Bradford method. Proteins (100 μg) were separated in a 15% SDS-polyacrylamide gel overnight and transferred to a PDVF membrane (Polyscreen; Dupont). For protein detection, the membranes were incubated with a phosphomyosin light chain-specific antibody, generously supplied by Dr. J. Staddon (25), at a dilution of 1:500 at 4°C overnight. After washing, the blots were incubated with secondary antibody and ECL detection was performed using the manufacturer's instructions. The membranes were reprobed with anti-VASP and anti-P-VASP antibodies and developed as described above.
Statistical analysis. Every experimental condition was duplicated within each experiment, and each experiment was repeated at least three times. The data are expressed as means ± SE. Comparisons were made with analysis of variance followed by Dunnett's modification of the t-test whenever comparisons were made with a common control, whereas the unpaired two-tailed Student's test was used for other comparisons. The level of statistically significant difference was defined as P < 0.05.
CNP and NO display different relaxing activities on H2O2-induced HUVEC contraction. Preliminary experiments were performed to determine the amount of cGMP released by different concentrations of CNP and two NO donors, SNP and sp-NO, in human endothelial cells in the absence of the phosphodiesterase inhibitor IBMX. CNP (10-5 M to 10-9 M) was used to activate pGC, and the NO donors SNP and sp-NO (10-5 M to 10-9 M) were used to activate sGC. After a 15-min incubation, cGMP levels were determined by radioimmunoassay (RIA) and the results are represented in Fig. 1A. Both SNP and sp-NO generated a concentration-dependent elevation of cGMP levels. sp-NO (10-6 M) increased cGMP levels from a control value of 2.5 ± 0.1 to 7 ± 0.06 fmol/μg protein, whereas incubation of the endothelial cells with CNP led to an increase of intracellular cGMP from a control value of 2.5 ± 0.1 to 33 ± 0.045 fmol/μg protein at the highest concentration used (half-maximal effective concentration EC50 = 1.8 × 10-7 M). The cGMP production in response to CNP was much more pronounced than that produced in response to NO even when the highest doses of the NO donors were employed. In view of these results, endothelial cells were incubated with sp-NO 10-6 M and CNP 10-7 M in all subsequent experiments.
To examine the differential effects of cGMP on H2O2-induced endothelial cell contraction, HUVEC were preincubated for 5 min with exogenous NO from sp-NO (NO, 10-6 M), CNP (10-7 M), or vehicle in buffer A without IBMX, after which H2O2 (10-4 M) was added (time 0) and the incubation continued for another 30 min. Microphotographs were taken at times 0 and 30, and PCSA was measured. As shown in Fig. 1B, H2O2 (10-4 M, 30 min) induced a significant contraction of cultured HUVEC expressed as a reduction of the PCSA, which was abolished by preincubation with the NO donor. CNP preincubation also diminished the H2O2 contractile effect, but to a lesser extent than NO. Similar results were obtained when another NO donor (SNP, 10-6 M) was used instead of sp-NO (PCSA after SNP plus H2O2: 100.2 ± 2% of control value).
To determine whether a correlation exists between intracellular cGMP levels and the relaxing potencies of the GC agonists, we measured cGMP production under the same experimental conditions on which microphotograph experiments were carried out. To this aim, HUVEC were preincubated with NO and CNP with or without H2O2 as described above, in the absence of the PDE inhibitor IBMX. cGMP levels after NO or CNP stimulation varied both in magnitude and temporal pattern. Whereas the NO donor induced low cGMP levels that peaked at 2 min after the start of the experiment (Fig. 2A) and faded at 5 min, CNP-mediated cGMP production resulted in levels that were 10 times higher and remained elevated for at least 10 min after the start of the experiment, declining to control levels at 30 min probably due to PDE activity (Fig. 2B). H2O2 addition did not affect cGMP production induced by either NO or CNP, because cGMP produced in CNP/H2O2- and NO/H2O2-treated cells was comparable to those obtained in CNP- or NO-treated cells. cGMP determination was conducted after 30 min of NO or CNP treatment in the presence of IBMX (Fig. 2C). PDE inhibition caused a more pronounced difference between cGMP levels obtained after GC stimulation and the levels obtained via sGC, thus confirming that the decline in cGMP observed in Fig. 2, A and B, is due to PDE activity. Similar results were obtained using Zaprinast (10-6 M), a selective PDE type 5 (PDE-5) inhibitor (data not shown). In both cases, H2O2 did not have any effect on cGMP production when used in combination with CNP or NO.
Taken together, these results show that the cGMP levels produced by CNP stimulation of pGC, which were one order of magnitude higher and remained elevated longer than the levels obtained by NO activation of sGC, did not correlate with the corresponding relaxing response elicited by this second messenger.
Recent findings indicate that H2O2 treatment of endothelial cells increases MLC phosphorylation, suggesting that endothelial contraction plays an important role in the oxidative stress-induced endothelial barrier dysfunction (18, 39). cGMP-dependent relaxation mechanisms involve MLC dephosphorylation via PKG activation (33). PKG-I is expressed in HUVEC in our experimental conditions (data not shown). To analyze whether this is the mechanism involved in our study, we labeled HUVEC with [32P]orthophosphate, as described in materials and methods. H2O2 increased phosphate incorporation into MLC, an effect that was completely prevented by preincubation with NO. By contrast, preincubation with CNP was unable to significantly prevent MLC phosphorylation in response to H2O2 (Fig. 3).
Effects of NO on endothelial cell relaxation depend on cGMP production. Because of the differences observed between NO- and CNP-relaxing potencies and cGMP production levels, we decided to test whether NO could produce its effects by means of an additional mechanism such as direct protein modification. As shown in Fig. 4, the effects of NO on PCSA were mimicked by the addition of a soluble cGMP analogue (8-Br-cGMP, 10-5 M) and were inhibited by treatment with a sGC inhibitor (ODQ, 10-6 M). In addition, a PKG-I inhibitor (Rp-cGMPs, 2.5 10-6 M) blocked the NO inhibitory effect on PCSA reduction obtained after HUVEC treatment with H2O2. This result suggests the involvement of the cGMP/PKG signaling pathway in the observed effects.
To determine whether the NO effect on MLC phosphorylation was also dependent on the cGMP/PKG pathway, HUVEC were preincubated with the sGC antagonist (ODQ) or the PKG-I antagonist Rp-cGMPs. MLC phosphorylation levels were analyzed by [32P]orthophosphate cell labeling. As shown in Fig. 5, MLC phosphorylation induced by H2O2 was inhibited by NO and ODQ, whereas Rp-cGMPs treatment reversed the NO-induced effects. These results indicate that NO mediates HUVEC relaxation mainly through the activation of sGC followed by cGMP production and the activation of PKG.
Iα isoform of PKG is involved in the differential effects of NO/cGMP and CNP/cGMP on H2O2-induced HUVEC contraction. It has recently been shown that PKG-Iα activates the MLC phosphatase by phosphorylating its myosin-binding subunit, thereby inhibiting MLC phosphorylation and contraction (33). To clarify the mechanism involved in the differential relaxing effect of NO-derived cGMP compared with CNP-derived cGMP, we transfected HUVEC with a dominant negative form of PKG type-Iα (fcGK-IαR). The overexpression of this form of PKG, which lacks the PKG catalytic subunit, is able to block cGMP-stimulated activity of the endogenous kinase, having no basal kinase activity itself (3). Figure 6 shows MLC phosphorylation of transfected cells using either fcGK-IαR or an empty vector (pcDNA 3.1). In the pcDNA 3.1-transfected cells, H2O2-induced MLC phosphorylation was completely prevented by NO, whereas CNP produced only a moderate inhibition. By contrast, fcGK-IαR transfection resulted in a complete reversal of the NO inhibitory effect on MLC phosphorylation. However, the CNP effects were not completely reversed by fcGK-IαR. The endogenous PKG activity of transfected HUVEC was assessed by examining the phosphorylation status of its vascular substrate, VASP, at serine (239) (31). Figure 6 shows an increase of P-VASP levels in NO- or CNP-treated cells transfected with an empty plasmid. By contrast, P-VASP levels are lower in cells transfected with a PKG negative dominant construct, even in the presence of NO or CNP, which confirms the biological activity of the transfected constructs.
The main finding of this study is that different pools of cGMP produced in human endothelial cells via pGC activation by CNP or NO stimulation of sGC have different functional activities. CNP has a moderate relaxing effect compared with NO on endothelial cells of human origin exposed to H2O2. However, cGMP production by the CNP/pGC system is higher (in all the concentration ranges studied) than that due to sGC stimulation by NO. We demonstrate that this discrepancy is due to a more efficient stimulation of PKG-Iα by the NO/sGC-derived cGMP. This result strongly suggests a role for the Iα isoform of PKG in transducing the signals of certain pools of cGMP to produce differential effects.
The concept of functional compartmentalization of second messengers is not new. One well-established example is the spatial control of [Ca2+] and cAMP signals (14, 38). However, parallels between those systems and cGMP have not been properly established. Zolle and coworkers (40) recently demonstrated that the activation of pGC inhibits Ca2+ extrusion via a plasma membrane Ca2+ ATPase, whereas the activation of sGC leads to an increase of Ca2+ uptake into the intracellular stores. The specific source of cGMP seems to be important for this effect. In our study, NO-derived cGMP produced a more intense relaxing effect than CNP-derived cGMP on endothelial cell contraction induced by H2O2. cGMP levels evoked by stimulation of sGC with NO or of pGC with CNP differ both in amplitude and duration. Whereas NO produced a typical pulse of cGMP during the first 5 min of incubation, which faded 15 min after the start of the experiment, CNP produced cGMP levels that were 10 times greater and remained elevated throughout the experiment. Despite the differences in cGMP production, there was no correlation with the observed cellular effect.
The first question that arises from this observation is whether the NO-induced effects are dependent on cGMP production. NO can act through cGMP-independent pathways to directly modify amino acid residues in several proteins and, thus, alter their function (32). In support of this hypothesis, Hart and coworkers (11) showed that NO effects on endothelial barrier dysfunction in porcine coronary artery endothelial cells stimulated by H2O2 were cGMP independent. In our study, incubation of HUVEC with inhibitors of either sGC or PKG and incubation with a cGMP analogue demonstrated that the NO-induced effects were mainly due to cGMP production rather than to a protein modification.
cGMP regulates cell responsiveness through PKG stimulation, which comprises a major mechanism for cGMP action (12). There are two reported forms of PKG: a soluble type I PKG, which is expressed and present in endothelial cells, smooth muscle cells, and neurons; and a membrane-bound type II PKG (16). Two isoforms of PKG type I (Iα and Iβ) are produced by alternate splicing of the same gene and differ only in their amino terminus. Both PKG type Iα and Iβ isoforms are involved in the control of smooth muscle cell relaxation. PKG type Iβ-dependent phosphorylation of inositol 1,4,5-triphosphate receptor-associated G kinase (IRAG) decreases Ca2+ release from the sarcoplasmic reticulum (1), whereas PKG-Iα phosphorylates the myosin-binding subunit of MLC phosphatase, activating it and therefore inhibiting MLC phosphorylation and contraction (33). The similarities between the contractile apparatus in smooth muscle and endothelial cells prompted us to investigate the role of the two PKG isoforms in the differential effects of cGMP originated by the activation of the two sets of GCs. Transfections with the dominant negative form of PKG-Iα completely abrogated the responses elicited by NO on MLC phosphorylation induced by H2O2. In this case, the responses to CNP showed only partial inhibition. Therefore, these results suggest that PKG-Iα is involved in the transduction of NO/cGMP signaling rather than the CNP/cGMP system. This could be simply due to a different spatial confinement of GCs and PKG-Iα. PKG-Iα was originally described as a cytosolic enzyme. However, under certain situations, it is partially associated with the cytoskeleton (31, 37). During oxidant injury, there are several changes that promote endothelial cell contraction and rearrangement of the actin cytoskeleton that compromises the endothelial barrier function, producing tissular edema (20). PKG-Iα is able to phosphorylate VASP, a protein member of the ENA/VASP family of proteins involved in the regulation of the actin cytoskeleton, causing its detachment from sites of focal adhesions. This could explain some of the effects of NO on H2O2-induced endothelial cell contraction because focal adhesions provide additional adhesive forces in the endothelial barrier regulation (22).
In addition to MLC kinase, MLC phosphorylation in endothelial cells can be induced by the Rho/Rho kinase pathway. PKG-Iα can phosphorylate Rho in vitro and in vivo, causing its inhibition (29). The role of a selective inhibition of Rho kinase in our results awaits further investigation.
Besides the possible phosphorylation by PKG-Iα of different substrates involved in cell relaxation, PKG-Iα can control the level and subcellular distribution of cGMP directly by regulating PDE activity. PDE-5 is the major PDE that degrades cGMP, and PKG-Iα can phosphorylate PDE-5 both in vivo and in vitro, activating it (23, 27). Although the physiological function for the phosphorylation and activation of PDE-5 has not been analyzed in endothelial cells, a similar mechanism has been described for platelets on NO sensitization (30). Activation of PDE-5 may then provide a negative feedback regulation of cGMP and PKG-Iα, because PDE-5 activation by PKG can also control the ability of PKG to phosphorylate other substrates when the intracellular concentration of cGMP reaches a high level, such as after CNP stimulation.
In summary, our results demonstrate that cGMP originating from separate sources within the endothelial cell plays a different role in the control of cell relaxation. Additional studies will be necessary to address whether these differences are due to a different subcellular location of the cGMP effectors, such as PKG and its substrates and cGMP degrading systems. This has potential implications for understanding the role of natriuretic peptides vs. NO in endothelial-dependent vascular relaxation and endothelial barrier function upon oxidant injury.
This work was supported in part by grants from the Ministerio de Ciencia y Tecnología (PM 97-0067, BFI 2001-1036), Comunidad de Madrid (08.4/0012/2001-2), and Fondo de Investigación Sanitaria (FISS 01-0434). F. J. Rivero-Vilches is a recipient of a postgraduate research scholarship from the Ministerio de Educación, Cultura y Deporte. S. De Frutos is a recipient of a postgraduate research scholarship from the Comunidad de Madrid. M. Saura is a research investigator from the Ramón y Cajal program (MCYT).
We thank Drs. Federico Gago, Carlos Zaragoza, and Lillian Puebla for review of this manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵* F. J. Rivero-Vilches and S. De Frutos contributed equally to this work.
- Copyright © 2003 the American Physiological Society