Review |
From Abteilung für Pharmakologie, Medizinische Fakultät, Ruhr-Universität Bochum, Bochum, Germany.
Correspondence to Doris Koesling, Abteilung für Pharmakologie, Medizinische Fakultät, Ruhr-Universität Bochum, Universitätsstr. 150, 44780 Bochum, Germany. E-mail doris.koesling{at}ruhr-uni-bochum.de
Rudi Busse Editor This Review is part of a thematic series on Cyclic GMP-Generating Enzymes and Cyclic GMP-Dependent Signaling, which includes the following articles:
Regulation of Nitric Oxide-Sensitive Guanylyl Cyclase
Cyclic GMP Phosphodiesterases and Regulation of Smooth Muscle Function
Structure, Regulation, and Function of Membrane Guanylyl Cyclase Receptors, With a Focus on GC-A
Cyclic GMP-Dependent Protein Kinases and the Cardiovascular System: Insights From Genetically Modified Mice
Regulation of Gene Expression by Cyclic GMP
Explaining the Phenomenon of Nitrate Tolerance
| Abstract |
|---|
|
|
|---|
Key Words: cyclic GMP guanylyl cyclase nitric oxide phosphodiesterase sensitization/desensitization
| Introduction |
|---|
|
|
|---|
NO-sensitive GC is a heterodimer consisting of two different subunits termed
and ß. The enzyme was shown to contain heme as a prosthetic group.1 Stimulation of the enzyme by its physiological activator NO leads to a tremendous, up to 200-fold increase in catalytic rate, ie, the conversion of GTP to cyclic guanosine monophosphate (cGMP). NO-sensitive GC, similar to other nucleotide-converting enzymes, requires Mg2+ as cofactor for catalysis. The NO-induced cGMP signal is conveyed intracellularly by the activation of several effector molecules: cGMP-dependent protein kinases, cGMP-regulated phosphodiesterases, and cGMP-gated ion channels. Most of the cGMP effects have been shown to be mediated by the cGMP-dependent kinase. Although the occurrence of cGMP-gated channels has been demonstrated outside sensory cells, their functional role remains to be established. The NO/cGMP signaling cascade (Figure 1) is of importance in the cardiovascular and nervous systems, where it controls smooth muscle relaxation,2,3 modulation of synaptic transmission,46 and inhibition of platelet aggregation.7,8
|
Although cGMP-forming activity was first described in 1969,913 it took until the mid-1970s to find out that there are two different types of GC,14 which were subsequently found to differ not only in their cellular localization (cytosolic versus membrane-bound) but also in their structure and regulation (see review15). The membrane-bound, peptide-activated GCs are structurally related to the cytosolic enzymes but are not stimulated by NO. These GCs belong to the group of receptor-linked enzymes containing one membrane-spanning region. For detailed review of this GC family, see the article "Guanylyl Cyclase Receptors" in this series.
| Isoforms and Structure of NO-Sensitive GC |
|---|
|
|
|---|
1 and ß1 subunits with molecular masses of 73 and 70 kDa, respectively, were cloned and sequenced.1619 The so-called
3 and ß3 subunits of NO-sensitive GC20 represent human variants of the
1 and ß1 subunits rather than different isoforms, and changes in the reading frame account for the differences in amino acid sequences.21
Homology screening yielded another subunit of NO-sensitive GC,
2.22 The occurrence of the
2 subunit was demonstrated on the protein level in human placenta, and the ß1 subunit was identified as the dimerizing partner.23 Despite little sequence homology between the two
subunits in the N-terminal domains, no differences in catalytic activity, NO stimulation, or substrate affinity could be detected between the
1ß1 and the
2ß1 heterodimers. Recent data suggest that the C-terminal peptide of the
2 subunit defines the particular property of this subunit: the ability to interact with PDZ domains. Through the interaction with the PDZ domain-containing protein PSD-95, the
2ß1 heterodimer was shown to be localized to synaptic membranes.24 In accordance with a putative role in synaptic transmission, the highest expression of the
2 subunit was found in brain. Furthermore, the ß1 subunit was detected at the presynaptic membrane of CA1 pyramidal cells, with nNOS being expressed in the postsynaptic side underlining the potential role of NO/cGMP signaling in the regulation of synaptic plasticity.25 Nevertheless,
2 is not the only
subunit in brain;
1 occurred in similar amounts. In all other tissues tested, the
1 subunit was the major occurring
subunit.26 It appears that the expression of
1 subunit correlates with the degree of vascularization.
Like
2, the ß2 subunit was identified by homology screening.27 However, existence of the ß2 subunit on the protein level appears unlikely; in the human gene, a base deletion has been described that results in a shift of the reading frame excluding functional expression of the protein.28 Quantitative PCR analysis of mouse tissue revealed virtually no mRNA for the ß2 subunit.26 In addition, no convincing evidence of a functionally active ß2-containing dimer has been presented so far.
NO-sensitive GC is catalytically active only as a heterodimer.29,30 Homodimeric enzymes (
1
1and ß1ß1) have been purified from Sf9 cells; however, no enzymatic activity of these homodimers has been detected.31
A comparison of the primary structure shows that the subunits can be divided into three domains: a C-terminal catalytic domain, a central part, and an N-terminal region. The catalytic C-terminal domains of each subunit of NO-sensitive GC display the highest degree of homology.32 These domains are also very similar to the respective regions in the peptide-activated, membrane-bound guanylyl cyclases and in the adenylyl cyclases (AC). The catalytic domains of AC and NO-sensitive GC are structurally closely related as shown by the functional expression of AC/GC heterodimers33 and the conversion of nucleotide specificity by the mutation of only three amino acids.34 In the crystal structure obtained for these AC domains,35,36 the binding site for the activator forskolin has been identified. The analogous region in NO-sensitive GC may be of special interest with regard to the possible binding site of the activator YC-1 (see later section).
The central regions preceding the catalytic domains show considerable homology to the membrane-bound enzymes.37 In analogy, these regions are probably involved in the dimerization of the subunits, although detailed investigation is still missing.
The N-terminal regions of the subunits of NO-sensitive GC comprise the heme-binding domain. The prosthetic heme group mediates the NO-induced stimulation subsequent to binding NO. A histidine residue (His-105) within this region of the ß1 subunit was shown to bind to the heme iron, thereby acting as the proximal ligand. Deletion of His-105 has been shown to abrogate NO activation of the enzyme.38
| Prosthetic Heme Group |
|---|
|
|
|---|
Both, the
and ß subunits are required for proper binding and orientation of the heme group. Sequential truncation showed that the subunits contribute unequally to heme binding with the ß1 subunit playing a more important role45; in addition, homodimers of the N-terminal part of the ß1 subunit (the first 385 amino acid) expressed in bacteria were shown to bind heme in a manner similar to the wild-type enzyme.46
The His-105 of the ß1 subunit was identified as the heme coordinating residue.38,47 Mutation of this histidine led to the generation of an NO-insensitive, heme-depleted enzyme with intact basal activity. In addition, two conserved cysteines adjacent to the His-105 on the ß1 subunit appear to play a role in the formation of the proper heme pocket. Mutation of these residues led to loss of enzyme-bound heme and NO responsiveness, which could be regained after heme reconstitution.48
| Activation of NO-Sensitive GC by NO |
|---|
|
|
|---|
Despite the stimulatory effect of these NO-containing compounds, the physiological significance of NO-induced activation of the enzyme did not become apparent until the identification of endothelium-derived relaxing factor (EDRF) as NO.51,52 Formation of EDRF had been shown to occur in endothelial cells in response to vasodilatory agonists such as acetylcholine, histamine, or bradykinin, leading to vasodilation via the activation of NO-sensitive GC in smooth muscle cells.53 After the discovery of NO in the vascular system, NO formation was reported to occur throughout the body.54 The enzymes responsible for the synthesis of NO were identified and termed NO synthases of which three isoforms are known to date. The inducible isoform produces relatively high NO concentrations and thereby exhibits direct toxic effects. The enzyme is mainly expressed in macrophages and plays a role in the nonspecific immune response. The neuronal and endothelial NO synthases are constitutively expressed enzymes that are regulated by the intracellular calcium concentration. The endothelial NO synthase is also activated by Akt/PKB-dependent phosphorylation triggered by shear forces exerted on the endothelial cells.55 Both isoforms produce relatively low NO concentrations (1 to 100 nmol/L). At these low concentrations, NO functions as a signaling molecule, and most of its effects are mediated via the activation of NO-sensitive GC.
In fact, an EC50 value for NO-sensitive GC as low as 2 nmol/L NO has been determined in cerebellar cell suspensions using a technique to deliver constant concentrations.56 Thus, activation of NO-sensitive GC in intact cells can be achieved at concentrations that are below those potentially inhibitory for other cellular targets such as mitochondrial cytochromes.56 NO donors, commonly used in in vitro and in vivo studies, differ in parameters like NO release or thiol requirement. For this reason, EC50 values reported for these substances vary considerably depending on the donor and the assay system used. Whereas NO donors such as GSNO or SNP have to be used at micromolar concentrations for half-maximal activation, the frequently used compound DEA-NO, which releases NO spontaneously, shows an EC50 value of approximately 300 nmol/L.24,57
| Mechanism of Activation |
|---|
|
|
|---|
, NO+), only the uncharged NO radical (NO·) has been shown to significantly activate NO-sensitive GC.58 Activation of the enzyme by NO involves the binding to the enzymes prosthetic heme group, which leads to the up to 200-fold activation of the enzyme.43,59 The heme group of NO-sensitive GC exhibits an absorbance maximum at 431 nm, which is indicative of a five-coordinated ferrous heme with a histidine as the axial ligand at the fifth coordinating position.43 Several models exist to explain the activation of NO-sensitive GC, the simplest is thought to occur in two steps: binding of NO to the sixth coordination position of the heme results in a six-coordinated NO-Fe2+-His-complex. The subsequent breakage of the histidine-to-iron bond leads to the formation of a five-coordinated nitrosyl-heme complex with an absorbance maximum at 398 nm. The opening of the histidine-to-iron bond is considered to initiate a conformational change resulting in the activation of the enzyme. In support of this simple model, protoporphyrin IX activates NO-sensitive GC independently of NO.60 Protoporphyrin IX, the iron-free precursor of heme, structurally resembles the NO-heme complex in which the iron is moved out of the plane of the porphyrin ring. Thus, in both structures, the axial histidine is unbound. However, the release of the histidine-iron bond appears not to be sufficient for the activation of NO-sensitive GC as a mutant lacking the proximal histidine did not show an increased catalytic rate. Therefore, the proximal histidine not only coordinates the heme group but also mediates the stimulatory action.
Recently, a more complex model of NO activation has been introduced. By using stopped-flow spectroscopy, Zhao et al61 investigated the subsecond kinetics of NO activation of NO-sensitive GC and provided data that the transition from the six-coordinate to the five-coordinate complex depended on the NO concentration present. This results in a model where NO not only activates the enzyme by binding to the heme but also regulates the velocity of activation via binding to a second non-heme binding site. However, Bellamy et al,62 reinterpreted the data by Zhao et al in favor of the simple two-state model and claimed the postulated additional NO site as unnecessary. These two models are currently disputed between the two groups.62,63
| Deactivation of NO-Sensitive GC |
|---|
|
|
|---|
NO dissociation measured spectrophotometrically resulted in a half-life of the NO-GC complex of approximately 2 minutes, which was accelerated to a half-life of
5 seconds by the addition of the substrate MgGTP.64,65 This accelerating effect of GTP is corroborated by a study using resonance Raman spectroscopy, which suggested binding of GTP in the proximity of bound NO, thereby possibly regulating NO binding.66 A half-life of approximately 3 minutes but no increase in the NO dissociation rate on the addition of MgGTP was shown in a different spectrophotometric study.67
The half-life of approximately 5 seconds for the NO-GC complex was confirmed monitoring deactivation, ie, the decline in cGMP forming activity of the purified enzyme68 or in cytosolic preparations of retina.69 Measurements in intact cerebellar cells showed an even 10-fold faster deactivation with a half-life of 0.2 seconds.70 In conclusion, deactivation of NO-sensitive GC was shown to be in the range required for rapid signaling in various systems. Additional research is necessary to unravel the special feature of the heme group of NO-sensitive GC permitting such fast NO dissociation.
| Inhibitors |
|---|
|
|
|---|
|
Other inhibitors such as methylene blue and LY-83583 show less specificity than ODQ and, eg, inhibit the olfactory cyclic nucleotide-gated ion channel80 or interfere with NO formation and NO release from NO synthases81; therefore, ODQ should be preferred as specific GC inhibitor.
| Novel Activation Mechanisms of NO-Sensitive GC |
|---|
|
|
|---|
|
YC-1 (3-(5'-hydroxymethyl-2'-furyl)-1-benzylindazole)90 was the first compound of this series to be identified. Originally, YC-1 had been published as an activator of NO-sensitive GC in rabbit and human platelets leading to an increase in the intracellular cGMP concentration.91,92 YC-1 causes a 10-fold activation of purified NO-sensitive GC independent of NO.82,93 Despite this NO independence, YC-1 failed to stimulate the heme-free enzyme, showing that the effect of YC-1 required the presence of the prosthetic heme group. Most interestingly, YC-1 sensitized NO-sensitive GC toward NO as YC-1 shifted the EC50 for NO to the left by one order of magnitude.82,93 Mechanistically, YC-1 was shown to slow down deactivation of NO-stimulated GC, which is sufficient to explain the sensitizing effect of YC-1.68,83 Rather unexpectedly, YC-1 also turned CO into an effective activator, despite the fact that by itself, CO is hardly able to activate the purified enzyme.82,94
By spectral analysis and enzyme studies, YC-1 was shown to bind to a site different from the heme group, suggesting a so far unknown allosteric site on the enzyme.83 Controversial reports exist on the localization of this allosteric site. In order to identify this site, Stasch et al95 used a novel high-affinity YC-1 analogue, BAY 41-2272, as photoaffinity label. An N-terminal region of the
1 subunit (amino acids 236 to 290) was identified to comprise the YC-1 binding site. However, this region is not conserved in the
2 subunit albeit the
2-containing heterodimer exhibits virtually identical sensitivity toward YC-1.23 In theory, the structural similarity between AC and GC in combination with the sensitizing effect of the respective activators forskolin and YC-1 suggest the C-terminal catalytic region as the YC-1 binding site (see previous section); this notion is supported by different mutations within this region of GC either mimicking or abolishing the effects of YC-1.68,96 Nevertheless, definitive conclusions about the allosteric site involved in NO sensitization of the enzyme await protein crystallography.
The effects of YC-1 on NO/cGMP signaling have been investigated in a variety of cells and tissues. In vascular smooth muscle cells, YC-1 was reported to increase cGMP levels and to induce a concentration-dependent relaxation of endothelium-denuded rat aortic rings.93,97,98 Other tissues in which YC-1 induced cGMP elevations include guinea pig trachea,99 guinea pig colon,100 corpus cavernosum,101 and pig urethra.102 In endothelial cells, the YC-1 induced cGMP increases were most pronounced as in these cells YC-1 did not only stimulate NO-sensitive GC directly but also potentiated the stimulatory effect of the endogenously synthesized NO.103
In platelets, YC-1 was shown to inhibit adhesion and aggregation.92,104,105 In these cells, YC-1 led to a drastic, over 1000-fold increase in cGMP in the presence of NO. Closer analysis revealed that YC-1 inhibited various PDEs.98,105 Due to the high catalytic rate of PDEs, some of the YC-1 effects observed in intact cells may be caused by the inhibition of PDEs rather than solely by the stimulation of NO-sensitive GC.
A second novel mechanism of activation has been shown to occur using the substance BAY 58-2667.89 This compound, an amino dicarboxylic acid, alone led to a moderate activation of NO-sensitive GC; this stimulation was additive to that of NO, which is in contrast to the potentiation seen with YC-1. Surprisingly, BAY 58-2667 stimulated the heme-oxidized or heme-depleted purified enzyme almost 200-fold. This indicates that BAY 58-2667 may be able to substitute for the heme group or stabilize the heme-free conformation; the occurrence of the heme-free form of NO sensitive GC under physiological conditions remains to be elucidated. Photoaffinity labeling and binding studies suggest binding of the compound to both subunits of NO-sensitive GC.
| Sensitization/Desensitization Within NO/cGMP Signaling |
|---|
|
|
|---|
Sensitization/desensitization describe an alteration of the NO/cGMP response toward a given NO challenge. This process may be caused by a change in enzyme reactivity, eg, by posttranslational modification, or an alteration of the amount of the enzyme present, the latter being discussed in the next section. Sensitization/desensitization of the cGMP response has been mostly attributed to a changed responsiveness of NO-sensitive GC. However, the impact of cGMP degradation has not been examined in most cases.
First reports on an increased sensitivity of the NO-induced cGMP response have been performed on isolated blood vessels that were either endothelium-denuded or in which endothelial NO production was impaired by inhibitor treatment.106108 In all these studies, removal of NO led to an increase in the potency of vasodilators used for relaxation. These in vitro data were corroborated in in vivo experiments by Moncada et al109 who found an enhanced antihypertensive response to glyceryltrinitrate in rats treated with inhibitors of eNOS. This "supersensitive" state did not require long-term treatment but was also detected in vessels treated with L-NAME for only 15 minutes. More recent reports show similar results in mice lacking the endothelial NO synthase.110 Thus, reduction of the NO concentration in vitro or in vivo leads to an increase in sensitivity of the signaling pathway.
Vice versa, decreased sensitivity of the NO/cGMP system has been demonstrated under various conditions after treatment with NO-releasing agents.111113 Transgenic mice overexpressing eNOS showed a decreased response toward endothelium-dependent and -independent vasorelaxation, which was attributed to a decreased activity of NO-sensitive GC.114 In a recent report, Bellamy et al115 showed a rapid NO-induced desensitization of NO-sensitive GC within seconds.
Despite these reports, factors responsible for sensitization/desensitization of NO-sensitive GC have not been identified so far (for phosphorylation, see next section). Although regulation of NO-sensitive GC by thiol-modifying or redox-active substances like superoxide has been postulated for more than 30 years, the functional consequences have not been unequivocally demonstrated on the purified enzyme, and the underlying molecular modifications are unknown. Theoretically, NO sensitivity of GC could be controlled by the heme content of the enzyme. However, the heme/protein ratio of 1 mole per mole obtained in several purification procedures from different tissues argues against a varying heme content. Besides, mice with limited heme synthesis were shown to contain a normally functioning NO-sensitive GC.116 Thus, a molecular mechanism for sensitization/desensitization of NO-sensitive GC is still lacking.
In most of these cases, sensitization/desensitization of the cGMP response was assessed by measuring intracellular cGMP levels not taking into account the downstream cGMP-degrading activity of phosphodiesterases (PDE). Recent results on cGMP-induced stimulation of PDE5 provide another possibility to explain the sensitization/desensitization within NO/cGMP signaling. In platelets, the NO-induced cGMP response was fast (less than a minute) and controlled by PDE5 activity.117 Desensitization did not occur on the level of NO-sensitive GC, which remained fully active but was attributed to an increased activity of PDE5. PDE5 is directly activated by cGMP binding to one of the enzymes regulatory GAF domains (GAF-A).118 Surprisingly, PDE5 activation was very sustained and still observed 1 hour after the removal of NO.119 PDE5 activation was paralleled by phosphorylation by the cGMP-dependent protein kinase I, which increases the cGMP affinity of the GAF-A domain.120 The exact mechanism of cGMP-induced activation of PDE5 is outlined in more detail in another article of this series. In sum, there is evidence for a new, so far unknown, feedback inhibition within NO/cGMP signaling (Figure 3). Within this feedback loop, cGMP bound to the regulatory GAF domain of PDE5 serves as a memory for the amount of NO the cell has been exposed to adjusting the subsequent cGMP response toward further incoming NO signals. Further experimental evidence is required to find out whether PDE5 activation is a general mechanism explaining the above described changes in the sensitivity of NO/cGMP signaling.
|
| Regulation of the Expression of NO-Sensitive GC |
|---|
|
|
|---|
1 subunit of NO-sensitive GC in rat aorta.124 The authors identified HuR (Human R, embryonic lethal abnormal visual [ELAV]-like RNA-binding protein) as a factor stabilizing GC
1 mRNA. YC-1-induced activation of GC decreased HuR expression, thereby inducing rapid degradation of GC
1 mRNA and lowering
1 subunit expression as a negative feedback response. Expression of GC subunits was also studied under conditions of low NO availability. Although animals in which NO production was impaired either by gene knockout110 or NO synthase inhibitor treatment125 showed an enhanced sensitivity toward additional NO, aortic GC content was not found to be altered.110
Besides the availability of NO, the effects of various other conditions on GC expression have been studied. Hypertension and ageing appeared to be factors negatively influencing the expression of GC.126129 High salt intake by spontaneously hypertensive rats led to a reduced vascular GC content in the aorta, which was paralleled by impairment of the vascular relaxation response to NO donors.130 Various additional factors like cytokines, estradiol, and ß-amyloid peptides appear to reduce GC protein.131134 Yet, also upregulation of aortic NO-sensitive GC in Wistar rats after myocardial infarction was reported.135
| Putative Activators or Inhibitors |
|---|
|
|
|---|
Before the identification NO as of the physiological stimulator of NO-sensitive GC, a multitude of mostly redox-active substances have been proposed as activators (see review15). Several mechanisms of activation have been postulated and subsumed as "redox regulation."139 With respect to the small stimulatory potential of these compounds, their physiological role as activators appears minor. In addition, one has to consider that GC is very sensitive to NO and is even stimulated by the minute amounts NO present in the normal atmosphere. Therefore, many of the putative GC activators may act indirectly by increasing the amount of available NO in the solution, eg, by a reduction of the superoxide concentration thereby preventing the inactivation of NO to peroxynitrite.140
Phosphorylation of NO-sensitive GC has been published by a few laboratories over the last 20 years. Enzymes possibly involved include PKC, PKA and PKG.141145 Further work will be needed to clarify the role of GC phosphorylation.
For many enzymes, product inhibition is a relevant mechanism of regulating catalytic activity. Accordingly, cGMP has been suggested to inhibit NO-sensitive GC.146 However, the IC50 values for cGMP were >5 mmol/L, concentrations that are unlikely to occur in vivo.
Moreover, calcium has been suggested to regulate the activity of NO-sensitive GC.70,147,148 The calcium concentrations used to inhibit NO-sensitive GC in these studies (5 to 250 µmol/L) appear to be too high to assume a relevant role under physiological conditions. A recent report,149 however, shows interesting data on a calcium-dependent membrane association of NO-sensitive GC. In human platelets, agonist-induced increase in free cytosolic calcium concentration was shown to lead to the translocation of NO-sensitive to the membrane accompanied by increased sensitivity toward NO. Whether these data apply as a general mechanism of regulating NO-sensitive GC will be evaluated in the near future.
| Conclusion and Perspective |
|---|
|
|
|---|
2ß1, a role in synaptic transmission is conceivable as this protein was shown to be mainly expressed in brain and to be targeted to synaptic membranes. In contrast, the
1ß1 heterodimer is prominent in vascularized tissues. Thus, the two GC isoforms differ in subcellular and tissue localization, which may reflect the association to the neuronal and endothelial NO synthases, respectively. Furthermore, besides the physiological and pharmacological NO-based activation, two so far unknown routes to sensitize and stimulate the enzyme have been presented. The modulation of the NO responsiveness of NO-sensitive GC has important pharmacological implications for the therapy of cardiovascular diseases. The relevance of the respective substances has been already been substantiated in various animal models. A better understanding of the underlying molecular mechanisms will greatly improve our current knowledge of the enzyme, which in turn may also help to elucidate the physiology and pathophysiology of NO/cGMP signaling.
| Acknowledgments |
|---|
| Footnotes |
|---|
| References |
|---|
|
|
|---|
-subunit of soluble guanylyl cyclase. FEBS Lett. 1991; 292: 217222.[CrossRef][Medline]
[Order article via Infotrieve]
2ß1 guanylyl cyclase to synaptic membranes. J Biol Chem. 2001; 276: 4464744652.
2ß1 isoform of NO-sensitive guanylyl cyclase in brain. Cell Signal. 2003; 15: 189195.[CrossRef][Medline]
[Order article via Infotrieve]