Abstract
This article is a report on a symposium sponsored by the American Society for Pharmacology and Experimental Therapeutics and held at the Experimental Biology 01 meeting in Orlando, FL. The presentations addressed the mechanisms of inhibition and regulation of cytochrome P450 and flavin monooxygenase enzymes by nitric oxide. They also highlighted the consequences of these effects on metabolism of drugs and volatile amines as well as on important physiological parameters, such as control of blood pressure, renal ion transport, and steroidogenesis. This is achieved via regulation of P450-dependent prostacyclin, hydroxyeicosatetraenoic acid, and epoxyeicosatrienoic acid formation. Conversely, the mechanisms and relative importance of nitric oxide synthases and P450 enzymes in NO production from endogenous and synthetic substrates were also addressed.
Nitric oxide is a short-lived radical gas with important roles as a cellular messenger in the cardiovascular and nervous systems (Liaudet et al., 2000). NO is synthesized from arginine through the actions of three different forms of nitric oxide synthase (NOS1) in response to various physiological stimuli (Liaudet et al., 2000). NO forms stable nitrosyl complexes with metal ions, most notably with ferrous iron in hemoproteins. Indeed, high affinity binding of NO to the heme iron of soluble guanylyl cyclase results in activation of this enzyme and the formation of cyclic GMP, which is one mechanism by which NO produces vasodilatation (Liaudet et al., 2000). It is not surprising, then, that NO and NO donors are capable of inhibiting the catalytic activities of hepatic microsomal and purified cytochrome P450 (P450) enzymes (Wink et al., 1993; Minamiyama et al., 1997; Morgan, 1997).
Together with the reversible inhibition of P450s caused by nitrosyl complex formation, an apparently irreversible inhibition is also observed, and this has been suggested to be due to nitration of tyrosine residues or to oxidation of P450 protein thiols (Roberts et al., 1998; Takemura et al., 1999). Furthermore, in addition to the well documented effects of NO on P450 catalytic activities, NO synthesized by the inducible NOS (iNOS) during inflammatory episodes has also been proposed to be responsible for the down-regulation of hepatic P450 proteins and mRNAs that occurs during an inflammatory response (Morgan, 2001). The participation of NO in hepatic P450 down-regulation is the subject of the section below by Morgan.
Inhibition of hepatic cytochrome P450 enzymes by NO released from NO donor drugs or by NO generated physiologically during an inflammatory reaction has obvious potential consequences for clinical drug metabolism. However, hepatic and especially extrahepatic P450 enzymes are involved in the synthesis of biologically active molecules with important physiological functions. These include the generation of arachidonic acid metabolites that regulate blood pressure and the synthesis of prostacyclins and steroid hormones. The inhibition of these P450 enzymes by NO and the physiological consequences of such inhibition are the subjects of the ensuing sections by Drs. Ullrich, McGiff, Hanke, and their respective coworkers.
Little is known about the effects of NO on drug-metabolizing enzymes other than the P450s. In the section by Dr. Cha and colleagues, this subject will be addressed by their studies on the effects of NO on human flavin monooxygenase enzymes.
NOS enzymes catalyze the formation of NO and citrulline from arginine via the intermediate N-hydroxyarginine, and the chemistry of these reactions is typical of cytochrome P450 reactions (Marletta, 1994). Although NOS enzymes are not homologous to P450s, they exhibit a reduced-CO difference spectrum identical to that of P450s, indicating the presence of a heme thiolate ligand (Marletta, 1994). Microsomal P450s can catalyze the formation of NO and citrulline fromN-hydroxyarginine but not from arginine (Renaud et al., 1993). Other studies have suggested that P450 enzymes participate in the physiological formation of NO during an inflammatory response (Fantuzzi et al., 1995; Kuo et al., 1995). Thus, P450 enzymes may be a source of NO as well as a target. In the final section of this report, Dr. Mansuy and colleagues present their findings on the mechanism and efficiency of NO formation by P450 enzymes compared with that by NOS enzymes.
Isoform-Selective Role of NO in Hepatic Cytochrome P450 Down-Regulation (E.T.M.)
Infectious or inflammatory stimuli cause the down-regulation of multiple P450 mRNAs and proteins in rat and human liver or hepatocytes (Morgan, 2001). In most instances, this is accompanied by the induction of iNOS activity and the subsequent production of nitric oxide in both hepatocytes and Kupffer cells. Thus, the role of NO in the suppression of P450 activity and expression has been a topic of study in many laboratories. As reviewed previously (Morgan, 1997; Morgan, 2001), different groups have reported apparently contradictory results of the effects of NOS inhibition on hepatic or hepatocyte P450 down-regulation. Various studies in our laboratory using NOS inhibitors in primary hepatocytes or in vivo or comparing the responses of wild-type and iNOS-null mice to bacterial lipopolysaccharide (LPS) have failed to find evidence for a role of NO in the down-regulation of constitutively expressed CYP2E1 or enzymes of the CYP2C or CYP3A subfamilies (Morgan, 2001, 1997).
In contrast to our negative findings on the role of nitric oxide in the down-regulation of constitutive P450s by inflammatory stimuli, we recently obtained compelling evidence for a NO-dependent down-regulation of the phenobarbital (PB)-induced expression of CYP2B1 protein in hepatocytes treated with LPS (Ferrari et al., 2001). Previously, Khatsenko et al. (1997) reported that administration of a NOS inhibitor to rats treated with PB and LPS blocked the down-regulation of CYP2B1/2 activity, mRNA, and protein. Carlson and Billings (1996) showed that inhibition of NOS blocked the suppression of CYP2B1/2 proteins by a cytokine cocktail in short-term cultures of rat hepatocytes.
We treated rat hepatocytes cultured on Matrigel with 1 mM PB to obtain maximal induction of CYP2B enzymes, then stimulated the cultures with LPS in the continued presence of the inducer. LPS treatment for 24 h caused a >80% suppression of CYP2B protein levels in the hepatocytes, and the concentration dependence of this suppression coincided with that for induction of NO production in the cells. The observed EC50 values for both parameters were approximately 3 ng/ml (Ferrari et al., 2001). Down-regulation of CYP2B protein occurred rapidly following LPS treatment, attaining 50% of control levels within only 6 h of stimulation and was prevented by the inclusion of the NOS inhibitorsNω-monomethylarginine or aminoguanidine in the media (Ferrari et al., 2001), suggesting that the rapid decline in CYP2B is caused by NO generation. This was supported by experiments showing that the effect of the competitive NOS inhibitors was reversed by the NOS substrate arginine and that the NO donors S-nitroso-N-acetylpenicillamine andS-nitrosoglutathione mimicked the effect of LPS (Ferrari et al., 2001). Furthermore, LY83583, a drug that inhibits NO production by a different mechanism (inhibition of iNOS induction), also blocked the down-regulation of CYP2B1 protein measured 24 h after LPS or interleukin-1 stimulation of the cells (Fig.1) (Ferrari et al., 2001).
Unlike the rapid, NO-dependent suppression of CYP2B1 proteins by LPS, the down-regulation of CYP2B1 mRNA by LPS was slower and unaffected by NOS inhibition. Furthermore, LPS was about 1000-fold more potent in the suppression of CYP2B1 mRNA than in suppression of CYP2B proteins or induction of NO formation (Ferrari et al., 2001). Therefore, we concluded that LPS causes suppression of inducible CYP2B expression by two different mechanisms: 1) a rapid and NO-dependent suppression of CYP2B proteins that occurs at higher concentrations of LPS, and 2) a slower and NO-independent suppression of CYP2B1 mRNA that occurs at lower LPS concentrations.
The NO-dependent and NO-independent suppression of CYP2B1 mRNA and proteins are conceptualized in the model shown in Fig.2. At low concentrations of LPS, CYP2B1 protein suppression is dependent on the potent, but slow, suppression of CYP2B1 mRNA. Consistent with our results, no significant effect of LPS on CYP2B1 protein is predicted after 24 h because the half-life of CYP2B1 protein in hepatocytes is >24 h (Roberts, 1997). At high-LPS concentrations, we hypothesize that the degradation of CYP2B1 protein is stimulated in an NO-dependent manner such that it declines faster than CYP2B1 mRNA levels. When this NO stimulated degradation is inhibited by inhibition of NO synthesis, the kinetics of protein suppression revert to those that occur at low-LPS concentrations (Fig. 2). Note that this model predicts that at longer time points (e.g., 48 h) after stimulation with high concentrations of LPS, CYP2B protein expression will be suppressed even in the presence of NO inhibitors due to the reduced levels of the mRNA and consequent reduction of protein synthesis. We have confirmed this prediction experimentally (Ferrari et al., 2001).
Incubation of CYP2B1 with peroxynitrite in vitro resulted in the formation of 2 mol of nitrotyrosine/mol of CYP2B1, and the formation of nitrotyrosine correlated with loss of enzyme activity (Roberts et al., 1998). A single major nitrotyrosine-containing peptide was identified, containing Tyr residues 190 and 203 (Roberts et al., 1998). Thus, it is feasible that nitration of these tyrosines by peroxynitrite formed by the reaction of NO and superoxide ions could be responsible for the stimulated degradation of CYP2B proteins caused by stimulation of hepatocytes with LPS. Interestingly, the tyrosines 190 and 203 are conserved in the mouse CYP2B10 and human CYP2B6 enzymes. Alternatively, modification of CYP2B1 by nitrosation or thiol oxidation could be the NO-dependent events that trigger the postulated degradation.
The conflicting reports from different laboratories on the role of NO in P450 regulation in cultured cells are likely to be explained by isoform-specific effects of NO and/or the use of different hepatocyte culture conditions. However, the disparate reports on the role of NO in P450 regulation in vivo are more difficult to explain. Our work on CYP2B1 in hepatocytes suggests that one possible explanation is that NO-dependent mechanisms of P450 suppression in vivo may occur only at high doses of LPS. In support of this idea, our in vivo experiments that failed to find NO-dependent effects on P450 regulation in vivo used a moderate dose of LPS (1 mg/kg), which is sufficient to down-regulate several P450 mRNAs, to induce acute phase protein expression and NF-κB activation in hepatocytes and to produce fever and inhibit food intake. On the other hand, almost all of the studies that found NO-dependent effects have used LPS doses of 2 mg/kg or more. However, it remains to be determined whether the dose-dependent dual mechanisms of CYP2B1 down-regulation by LPS observed in cultured hepatocytes also occur in the whole animal, and studies are under way to address this question.
Reactions of P450 Proteins with Peroxynitrite (V.U., A.D., P.S., T.N., and H.S.)
Peroxynitrite (PN) originates from a combination of nitric oxide (⋅NO) with superoxide (⋅O
Using model investigations, we were able to prove that heme proteins react with PN as shown in Fig. 3. The primary ferryl complex can be reduced to its ferric form by an endogenous tyrosine or by exogenously added phenols, which then, as phenoxy radicals, add the ⋅NO2 radical and form nitrated phenols as products. This reaction has been shown for P450CAM to nitrate tyrosine residues at the active site (Daiber et al., 2000a), and also the heme portion of P450BM−3 was nitrated in the vicinity of the heme (Daiber et al., 2000b). The nitration was even higher when the active site located at F87 in P450BM−3 was replaced by Tyr, indicating that the spatial factor is important for an effective nitration. This allows the conclusion that a Tyr residue occurs in the direct neighborhood of the heme-iron of PGI2 synthase, which is in agreement with a block of nitration by a substrate analog (Zou et al., 1997). In view of the important regulatory role of PN for this P450 enzyme, we have further extended our studies on PN interaction to other heme thiolate proteins. Of special interest was the recently crystallized P450NOR (Nakahara et al., 1993; Park et al., 1997) for which NO is a natural substrate, being reduced to N2O by NADH. The organism Fusarium oxysporum may also become exposed to superoxide, and hence peroxynitrite may be a second substrate for the enzyme. We therefore investigated its reaction with PN and compared the data with those of other P450 enzymes and heme proteins.
Nitration of tyrosine residues in a protein can be conveniently followed by Western blots with staining by anti-nitrotyrosine antibodies (Ye et al., 1996). Such monoclonal or polyclonal antibodies are available, but they differ in their specificities and sensitivities depending on the probed protein and the antigen used for their generation. By this technique PN-treated P450 proteins were found to be tyrosine-nitrated (Roberts et al., 1998; Mehl et al., 1999). Surprisingly, we found that P450NOR only reacted positively at very high-PN concentrations (Fig.4, left side). If NO and⋅O
To determine the exact quantity of 3-NT formed, complete pronase digests of PN-treated P450CAM and P450NOR were separated by high-pressure liquid chromatography and detected at 360 nm (Fig.5). From such chromatograms, 3-NT could be detected in both P450 proteins but at comparable bolus PN concentrations P450NOR was 3- to 4-fold less nitrated. Using SIN-1 at a concentration of 500 μM in the presence of 5 μM P450 protein, the quantitation for P450CAMwas 1.51 ± 0.03 μM compared with 0.26 ± 0.11 μM for P450NOR (2-h incubations in 0.1 M NaPi; pH 7.4). In the mutant of P450BM−3 in which F87 was exchanged for Y, the extent of 3-NT reached 6.5 ± 0.5 μM.
The primary conclusion from the data is that P450NOR is less sensitive to PN with regard to an autocatalytic nitration of endogenous tyrosines. Reasons for this could be a lack of accessible Tyr residues or a decreased reactivity with PN. The latter possibility could be ruled out since previous measurements had shown the highest rate of PN decomposition for P450NOR compared with P450BM−3, P450CAM, or CPO (Zou et al., 2000).
We then checked the possibility that the enhanced turnover with PN involved a different route of decomposition or a preferred isomerization of PN to nitrate (Mehl et al., 1999). Rapid mixing experiments, however, yielded the same ferryl intermediate (compound II) as had been observed with P450BM−3 or CPO (Mehl et al., 1999; Daiber et al., 2000a).
A final experiment was designed to confirm that P450NOR reacted faster with a second molecule of PN than other P450 proteins as judged from its rapid degradation of PN. When the decomposition of PN was carried out in the presence of phenol, the yields of 2- and 4-nitrophenols were highest with CPO and lower but comparable with P450NOR and P450BM−3 (Zou et al., 2000). This again confirmed that P450NOR behaved qualitatively similar to other P450 proteins but reacted with faster kinetics. It can be assumed that the easy access of PN to the active site is the main reason for its enhanced turnover of PN compared with other P450 and heme proteins. In view of the high stability of peroxidase compound II of peroxidase in the presence of PN, it is likely that the reactivity of the ferryl complexes of P450 enzymes is high and possibly due to the special properties of the S-Fe-O entity.
In summary, nitric oxide and superoxide anion form PN, which can react with heme proteins and features especially high turnovers with heme-thiolate (P450) proteins. We studied the reactions of PN with various P450 enzymes as models for the Tyr nitration of prostacyclin synthase, which was found to be Tyr-nitrated at very low levels of PN. We report that of all P450 proteins tested the P450-dependent NO-reductase (P450NOR) shows the highest rate of PN decomposition with a very low rate of auto-Tyr nitration. The catalytic cycle involves a ferryl species that can either react with a second molecule of PN or with exogenous phenol. The open active site favors rapid kinetics, and the obvious absence of active-site-located Tyr residues keeps auto-nitration low. This reaction of P450NOR may be of physiological significance. Thus, in addition to its catalysis of NO reduction by NADH, P450NOR may also be able to lower PN concentrations in F. oxysporum when superoxide causes the extremely fast combination with the natural substrate NO to yield PN.
NO Exerts a Tonic Inhibitory Effect on 20-HETE Formation: Renal Functional Implications (J.C.McG. and A.O.)
The ability of NO to inhibit cytochrome-P450 monooxygenase activity was initially recognized by Duthu and Schertzer (1979), who related the inhibitory action of nitrates on hepatic drug metabolism to formation of NO from nitrates. NO reacts with superoxide (O
Rationale.
To characterize and define the basis of these effects ofl-NAME on renal function, we addressed the responses to inhibition of NO production in terms of 1) renal metabolism of AA, 2) the expression of ω-hydroxylase, and 3) the efflux of 20-HETE from the kidney (Oyekan et al., 1999). In microsomal suspensions from the renal cortex and medulla obtained from rats treated withl-NAME for 10 days, conversion of [14C]AA to HETEs was increased, as was the expression of ω-hydroxylase (CYP4A) protein (Oyekan et al., 1999). CYP4A protein is constitutively expressed in renal microsomes of control rats and is inducible in clofibrate-treated rats. In microsomes from rats treated with l-NAME for 10 days, expression of CYP4A protein increased by ∼40%. The magnitude of the effect of inhibition of NOS was strikingly evident in the renal efflux of 20-HETE in rats treated with l-NAME, viz., a 4-fold increase in 20-HETE release from the rat kidney occurred. On the other hand, conversion of AA to epoxyeicosatrienoic acids was unaffected byl-NAME treatment. However, inclusion of 2% NaCl in the drinking water for 7 to 10 days greatly elevated the conversion of AA to epoxyeicosatrienoic acids (by 20-fold or more), whereas it reduced the capacity of renal cortical microsomes to convert AA to HETEs (Oyekan et al., 1999).
Proximal Tubular Transport: 20-HETE and NO.
The natriuresis-diuresis that accompanied inhibition of NO formation was rapid in onset and was related to increased 20-HETE synthesis in two key segments of the nephron, proximal tubules (PT) and thick ascending limb, that are responsible for an estimated ca. 80% of sodium chloride absorption. These segments are heavily invested with ω-hydroxylase (Carroll et al., 1991; Omata et al., 1992). In the PT, 20-HETE serves as a second messenger for several vasoactive hormones, parathyroid hormone (Ribeiro et al., 1994), endothelin-1 (ET-1) (B. A. Escalante, J. C. McGiff, and A. O. Oyekan, unpublished observations), and dopamine (Satoh et al., 1993), that inhibit transport in this segment. Evidence for 20-HETE acting as a second messenger for ET-1 to mediate inhibition of ion and fluid absorption in the PT, effects that are potentiated by inhibition of NOS in response to the peptide, will be examined.
In freshly isolated rat PT, 86Rb uptake, an index of ion transport, increased in a time-dependent manner, reaching saturation after 10 min. Ouabain and amiloride inhibited86Rb uptake by ∼70%, indicating that an active Na+-K+-ATPase and Na+/H+ exchanger regulate ion transport in rat PT. To examine the participation of a P450-dependent transport mechanism mediated by an arachidonate metabolite, the effect of AA on 86Rb uptake by PT was studied in the presence and absence of DBDD, the selective inhibitor of ω-hydroxylase (B. A. Escalante, J. C. McGiff, and A. O. Oyekan, unpublished observations). AA decreased86Rb uptake (ng of Rb/10 μg of protein/2 min), an effect abolished by DBDD, suggesting that 20-HETE mediated the AA effect. This suggestion was supported by demonstrating that 20-HETE mimicked the action of AA on 86Rb uptake. DBDD did not affect the response to 20-HETE. Ouabain was used as a negative control; l-NAME did not affect ouabain-induced inhibition of 86Rb uptake, which ranged between 60 to 70% with or without l-NAME treatment.
To examine the effects of disinhibition of the tonic inhibitory action of NO on ω-hydroxylase activity, rats were treated withl-NAME for 4 days before isolating the rat PT (B. A. Escalante, J. C. McGiff, and A. O. Oyekan, unpublished observations). l-NAME increased 20-HETE production by PT associated with potentiation of the inhibitory effect of AA on86Rb uptake. Thus, 0.1 μM AA did not inhibit86Rb uptake by rat PT unless the rats were treated with l-NAME, resulting in a ca. 40% inhibition of86Rb uptake. Furthermore, this effect of AA was mediated by 20-HETE because inhibition of 20-HETE synthesis by DBDD attenuated the inhibitory action of AA on86Rb uptake.
The principal target of 20-HETE in the PT is Na+-K+-ATPase. 20-HETE inhibits Na+-K+-ATPase via activation of protein kinase C, which in turn phosphorylates the α subunit of the Na pump, preventing extrusion of Na at the basolateral surface (Nowicki et al., 1997). This mechanism in which 20-HETE acts as a second messenger is also the basis for the inhibitory action of parathyroid hormone (Ribeiro et al., 1994), ET-1, and dopamine (B. A. Escalante, J. C. McGiff, and A. O. Oyekan, unpublished observations) and is subject to modulation by NO. Additional support for the regulation of proximal tubular transport by NO is based on the similarity of the effects on transport in this segment of neuronal NOS (nNOS) knock-out mice and inhibition of NOS withl-NAME (Wang et al., 2000). Inhibition of NO increased urine volume and Na+ excretion and decreased the rate of fluid and HCO3 absorption, effects which were identical to those observed in nNOS knock-out mice (Wang et al., 2000). Furthermore, NO donors abolished the stimulation of fluid uptake by luminal angiotensin II (Eitle et al., 1998).
NO Modulates Vascular ω-Hydroxylase.
In an attempt to define the relative contributions of reduced 20-HETE synthesis (elimination of the vasoconstrictor effect of 20-HETE) versus cGMP generation (vasodilator) to the renal vascular response to NO, Alonso-Garcia et al. (1997) studied the renal vasodilator response to an NO donor, sodium nitroprusside (SNP). In this study, SNP (10−7 to 10−3 M) increased the diameter of isolated perfused interlobular arterioles preconstricted with phenylephrine in a concentration-dependent manner to 82% of control. As SNP increased vascular diameter by only 17% after inhibition of 20 HETE synthesis with DBDD, the renal vasodilator action of the NO donor was considered to be dependent primarily on prevention of 20-HETE formation. This study endorses unambiguously the importance of a cGMP-independent effect of NO on renal vascular tone and reactivity. Confirmation of this view was obtained in an in vivo study of the renal circulatory response to an NO donor in the face of inhibition of ω-hydroxylase. After inhibition of ω-hydroxylase, the fall in renal vascular resistance in response to the NO donor was reduced by more than 70%. The authors concluded that the minor component of the renal vasodilator response to NO represents a cGMP effect (Alonso-Galicia et al., 1997).
In summary, NO exerts a braking action on ω-hydroxylase. Inhibition of NO production by l-NAME removes this braking effect, allowing enhancement of 20-HETE synthesis with attendant renal functional effects.
Nitric Oxide Inhibits Steroid Hormone Synthesis in Adrenal Zona Glomerulosa Cells (C.J.H. and W.B.C.)
The regulation of aldosterone synthesis by the adrenal zona glomerulosa (ZG) cell involves a complex interaction between a wide variety of endogenous stimulatory and inhibitory factors. Angiotensin II (AII), adrenocorticotropic hormone, and potassium ion are the primary secretagogues stimulating aldosterone synthesis (Quinn and Williams, 1988). Atrial natriuretic peptide and decreasing oxygen concentration have been identified as inhibitory factors (Campbell et al., 1985; Raff et al., 1989). Recent investigations in a number of laboratories have indicated the inhibitory effects of NO on the synthesis of various steroid hormones (Adams et al., 1992; Natarajan et al., 1997; Cymeryng et al., 1998). The mechanism of NO inhibition of aldosterone synthesis involves a direct interaction with the cytochrome P450 enzymes required for the multistep conversion of cholesterol into aldosterone (Hanke et al., 1998). The inhibitory effects of NO and the ability of nitric oxide to bind to the cytochrome P450 heme site have been previously described (Wink et al., 1993). This article will present data supporting the inhibitory effects of NO on aldosterone synthesis and the possible physiologic interactions of NO in the adrenal gland.
The inhibition of aldosterone synthesis by NO was examined in bovine adrenal ZG cells maintained in primary culture. Both, type-A natriuretic peptide (10−10 to 10−6 M) and the NO donor (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl) amino]diazen-1-ium-1,2-diolate (DETA nonoate; 10−6 to 10−3 M) stimulated concentration-related increases in ZG cell cGMP. Type-A natriuretic peptide and DETA nonoate also attenuated AII-stimulated aldosterone production over the same range of concentrations. The selective inhibitor of soluble guanylyl cyclase 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one completely prevented DETA nonoate-stimulated cGMP production without altering the inhibitory effect of DETA nonoate on AII-stimulated steroidogenesis. 1H-[1,2,4]Oxadiazolo[4,3-a]quinoxalin-1-one did not block type-A natriuretic peptide-stimulated cGMP synthesis or type-A natriuretic peptide inhibition of steroidogenesis.
The regulation of aldosterone synthesis has been divided into the study of the early pathway, characterized by the conversion of cholesterol into pregnenolone, and the late pathway conversion of progesterone into aldosterone. The direct inhibitory effects of DETA nonoate on steroidogenic cytochrome P450 enzymes were determined by treating ZG cells with exogenous 25-hydroxy cholesterol and progesterone. These treatments bypass the signal transduction cascades and directly stimulate steroid hormone synthesis within the cell. DETA nonoate (10−3 M) completely blocked 25-hydroxy cholesterol and progesterone-stimulated aldosterone synthesis in ZG cells and inhibited the conversion of 25-hydroxy cholesterol to pregnenolone in the mitochondrial fraction of bovine adrenal cortex. DETA nonoate-derived NO binding to cytochrome P450 enzymes of isolated ZG mitochondria produced an absorbance maximum of 453 nm and blocked the formation of the carbon monoxide-cytochrome P450 complex with a characteristic absorbance maximum at 450 nm. These data suggest that DETA nonoate reduces steroidogenesis independent of guanylyl cyclase activation and that NO has a direct effect to inhibit the activity of cytochrome P450 enzymes by binding to the heme groups of the enzymes.
Localization of NOS within the outermost zones of the adrenal cortex was examined in cultures of adrenal cortical cells. Cell lysates of adrenal fibroblasts and ZG cells did not demonstrate immunoreactive bands to inducible NOS, neuronal NOS, or endothelial NOS antibodies by Western blotting. Adrenal endothelial cell lysates contained a 133-kDa band that was immunoreactive to endothelial NOS but not inducible NOS or neuronal NOS antibodies. Only endothelial cells demonstrated significant NOS activity measured as the conversion of [3H]l-arginine to [3H]l-citrulline.
The inhibitory effect of NO on ZG cell aldosterone synthesis was enhanced 10-fold by decreasing oxygen concentrations from 21 to 8%. The IC50 for DETA nonoate inhibition of AII-stimulated aldosterone was approximately 3 × 10−4 M in 21% oxygen and 2.5 × 10−5 M in 8% oxygen. Coincubation of endothelial and ZG cells resulted in NO-mediated inhibition of basal and AII-stimulated aldosterone synthesis in 8% oxygen but not 21% oxygen. These data indicate that the NO inhibition of aldosterone synthesis is oxygen-sensitive and that, in decreased oxygen environments, adrenal endothelial cell NO production inhibits ZG cell aldosterone synthesis.
Transduction of ZG cells with adenovirus encoding endothelial NOS (AdeNOS) resulted in the synthesis of NOS protein and enzymatic activity. Cells transduced with β-galactosidase-encoding adenovirus and untransduced cells did not demonstrate detectable NOS protein or activity. AII-stimulated aldosterone synthesis was decreased in AdeNOS-transduced cells compared with β-galactosidase-encoding adenovirus control cells. Treatment with the NOS inhibitor thiocitrulline (3 × 10−5 M) restored AII-stimulated aldosterone synthesis following AdeNOS transduction. These data demonstrate that adenovirus-mediated gene transfer of endothelial NOS in ZG cells results in the expression of active NOS enzyme and that this endogenous NO production within ZG cells decreases aldosterone synthesis.
The results of this study, summarized in Fig.7, indicate that NO inhibits ZG cell aldosterone synthesis by a direct interaction with multiple steroidogenic cytochrome P450 enzymes. The inhibition of aldosterone synthesis does not require activation of the soluble guanylyl cyclase enzyme of the ZG cell. Within the physiologic setting, ZG cell aldosterone synthesis may be inhibited by NO released from the adjacent endothelial cell, but this effect appears to require decreased oxygen concentrations.
Overproduction of Nitric Oxide Causes Reduction of Hepatic Flavin-Containing Monooxygenase (FMO) Activity and Trimethylaminuria in Patients with Chronic Viral Hepatitis (C.S.P, J.S.K., H.G.Y., and Y.N.C.)
FMOs play a major role in hepatic N- andS-oxidation of various endogenous and exogenous compounds (Ziegler, 1988). FMOs are responsible in large part for the oxidation of the volatile odorous sulfur and nitrogen metabolites, produced by metabolism of dietary methionine, and choline, produced by intestinal microflora, to nonvolatile hydrophilic metabolites, which are excreted predominantly in urine without any aroma.
When the liver fails to oxidize the absorbed volatile substances and allows them to escape via breath, sweat, and urine, “foetor hepaticus”, “fish-odor syndrome”, or “trimethylaminuria (TMAU)” may occur (Mitchell et al., 1999). FMO3 is the predominant FMO enzyme present in adult human liver and catalyzes the oxidation of volatile trimethylamine (TMA) produced by intestinal microflora from dietary choline to the nonvolatile TMA N-oxide (TMAO). Reduced activity of FMO3 may be caused either by inheritable defects in the FMO3 gene (primary form) or by reduction of FMO3 activity in chronic liver diseases (secondary form), and the resulting failure to oxidize TMA produces fish-odor syndrome or TMAU. In this respect, the breath, urine, and body odors of patients with chronic viral hepatitis (CVH) have a fecal or rotten-fish odor, associated with severely depressed P450 and FMO activities. Mechanisms underlying the loss of these drug-metabolizing enzyme activities are not well understood.
CVH is characterized by parenchymal infiltration of activated cytotoxic T-lymphocytes and also by the abundant presence of proinflammatory cytokines like tumor necrosis factor-α and interferon-γ and adhesion molecules in the liver. Inflammatory cytokines enhance iNOS expression in hepatocytes and overproduce NO (Geller et al., 1993), which has been shown to decrease hepatic contents of P450 in vivo either by suppressing the expression of P450 mRNA or by direct inhibition of liver microsomal P450 activity (Wink et al., 1993;Carlson and Billings, 1996). The overproduced NO has also been demonstrated to decrease hepatic content of FMO in vivo by suppressing the expression of FMO mRNA, without directly interacting with the flavin adenine dinucleotide of FMO (Park et al., 1999).
This study was conducted, therefore, to understand the mechanisms involved in the development of secondary TMAU in CVH patients thought to have stimulated immune systems and induction of hepatic iNOS. We measured the in vivo activity of FMO in 12 healthy volunteers and 22 CVH patients diagnosed with chronic hepatitis B virus infection (n = 8), liver cirrhosis (n = 7), or cirrhotic hepatocarcinoma (n = 7; HCC) and correlated FMO activity with plasma levels of nitrite + nitrate (NOx), the stable end products of NO, in the same patients. In vivo FMO activities were determined by urinary ratios of TMAO to TMA (Zhang et al., 1992) and of ranitidine (RA) N-oxide to RA (Kang et al., 2000) 8 h after ingestion of 168 mg of RA hydrochloride. These ratios are expressed in Table 1 as the percentages of TMAO in urinary TMAO + TMA and of RANO in urinary RA + RANO, respectively. The mean plasma NOx concentrations of CVH patients were elevated 2.2-fold compared with those of the healthy volunteers (Table1). Concomitantly, the fraction of nonvolatile TMAO in excreted TMA + TMAO was reduced in the urine of CVH patients compared with healthy patients (74% compared with 96%; Table 1). The urinary ratios of RANO to RA provided further support for a reduced in vivo FMO activity in CVH patients (Table 1). FMO activity observed in CVH patients without cirrhosis or hepatoma was not significantly different from those with cirrhosis or hepatoma (data not shown).
We then studied the expression and activity of iNOS and FMO3 in cirrhotic and cancerous areas of liver tissues from five cirrhotic hepatocarcinoma patients compared with normal human livers. iNOS mRNA and iNOS protein were not detectable in the normal human liver tissue but were elevated in the surgically removed cirrhotic and cancerous liver tissues, as observed by Majano et al. (1998). Conversely, the contents of FMO3 mRNA and protein were higher in the normal human liver tissue than in cirrhotic or cancerous liver tissues, with contents in the cancerous liver tissues being much lower than those in surrounding cirrhotic tissues. FMO3 (RA oxidation assay) activities in both cirrhotic and cancerous liver tissues were depressed severely compared with normal liver. These results suggest that the overproduced NO might have suppressed the hepatic expression of FMO3 mRNA and protein in CVH patients, causing a reduction of in vivo FMO3 activity.
To determine whether overproduced NO in CVH patients may have decreased the hepatic FMO3 activity directly, we pretreated normal human liver microsomes with the NO generatorS-nitroso-N-acetylpenicillamine (SNAP; 1 mM) and measured the remaining microsomal RA N-oxidation activity in vitro after centrifugation and resuspension. Pretreatment with SNAP inhibited FMO3 activity by 30 to 40%, and this was completely prevented by copretreatment with 100 μM hemoglobin (NO scavenger) but not with 500 μM butylated hydroxyanisole (BHA) (⋅O
Pretreatment of microsomes with 1.0 mM SIN-1, which generates ONOO− by releasing both NO and⋅O
In conclusion, our results indicate that the reduction of FMO3 activity (TMAU and reduced RA N-oxidation) observed in vivo in CVH patients may be caused by the overproduced NO. The overproduced NO, resulting from induction of iNOS in hepatocytes of CVH patients, may decrease the expression of FMO3 mRNA indirectly (suppression). It may also inhibit FMO3 activity directly via reversible nitrosylation of cysteine residues when superoxide anion is absent (modulation), or it may also cause the destruction of FMO3 protein via irreversible nitration of tyrosine residues when superoxide anion is present. In the liver of CVH patients, all these mechanisms (destruction, modulation, and suppression) may operate in concert to compromise the FMO3 activity, and this may be responsible, at least in part, for the secondary form of trimethylaminuria observed in CVH patients.
Formation of Nitric Oxide by Cytochromes P450: Comparison with NO Synthases (D.M. and J.-L. B.)
The only route of biosynthesis of NO discovered so far in mammals, as in most living organisms, is the NOS-catalyzed oxidation ofl-Arg to citrulline and NO (Pfeiffer et al., 1999). This reaction occurs in two steps. The first is a monooxygenation of Arg toNω-hydroxy-Arg (NOHA) with consumption of 1 mole of O2 and 1 mole of NADPH. The second step is a three-electron oxidation of NOHA that leads to an oxidative cleavage of the CNOH bond of NOHA. It consumes 1 mol of O2 and only 0.5 mol of NADPH.
As far as the second question is concerned, cytochrome P450-dependent monooxygenases appeared to be good candidates to catalyze oxidations similar to those performed by NOS because of the great analogy between these two classes of heme-thiolate proteins that use identical prosthetic groups, if one excepts tetrahydrobiopterin (BH4), which is only present in NOS. In fact, microsomal cytochromes P450 were found to catalyze the oxidation of NOHA with formation of citrulline and NO (Boucher et al., 1992; Renaud et al., 1993). In a more general manner, microsomal cytochromes P450 catalyze the oxidative cleavage of the CNOH bond ofN-hydroxyguanidines, amidoximes, ketoximes, and aldoximes, with formation of the corresponding products with a CO bond, and nitrogen oxides, including NO (Jousserandot et al., 1995).
Similar experiments have been done recently on recombinant inducible NOS, which does not contain BH4(BH4 free-iNOS), and variousN-hydroxyguanidines (Moali et al., 2001). Some of theseN-hydroxyguanidines were related to NOHA, such asNω-hydroxyhomo-l-arginine (homo-NOHA),Nω-hydroxynor-l-arginine (nor-NOHA), andNω-hydroxydinor-l-arginine (dinor-NOHA), whereas others did not contain an α-amino acid function, such as N-aryl-N′-hydroxyguanidines andN-hydroxyguanidine itself. BH4free-iNOS catalyzes the oxidation of all theseN-hydroxyguanidines with formation of NO2− and NO3− at rates between 20 and 80 nmol · min−1 · mg of protein−1 (Moali et al., 2001).
In those reactions, BH4 free-iNOS and microsomal cytochromes P450 exhibit a strikingly similar behavior: 1) they are not substrate selective, as all N-hydroxyguanidines are oxidized; 2) they are not selective in terms of products because the oxidation of N-aryl-N′-hydroxyguanidines leads not only to the corresponding N-arylureas but also to the corresponding cyanamides; and 3) the corresponding reactions are strongly inhibited by SOD, indicating that they are mainly due to the oxidase function of these hemeproteins (Moali et al., 2001). It is noteworthy that the oxidation of NOHA by BH4-free iNOS, NADPH, and O2 leads to the formation ofNδ-cyanoornithine in addition to citrulline (Rusche et al., 1998).
The oxidation of the same N-hydroxyguanidines by complete iNOS, which has been reconstituted after incubation with BH4, exhibits very different characteristics (Moali et al., 2001). Oxidation by iNOS is substrate selective because NOHA and homo-NOHA are efficiently transformed with formation of NO, whereas nor- and dinor-NOHA are not. Interestingly, someN-aryl-N′-hydroxyguanidines are also efficiently oxidized with formation of NO. The iNOS-dependent reactions are much more efficient than the corresponding BH4-free iNOS- and cytochrome P450-dependent reactions (rates from 100–400 instead of 20–80 nmol · min−1 · mg of protein−1). They are also much more selective because they only lead to the urea corresponding to the starting N-hydroxyguanidine and NO in stoichiometric amounts. In fact, oxidations catalyzed by iNOS lead to a clear formation of NO in an NO-urea ratio close to 1, whereas oxidations by BH4-free iNOS only lead to very low levels of NO. Finally, contrary to the BH4 free-iNOS and cytochrome P450 reactions, they are not inhibited by SOD. Thus, the oxidations of N-hydroxyguanidines by microsomal cytochromes P450 and BH4 free-iNOS appear to be mainly performed by O
The different behavior exhibited by BH4-free iNOS or microsomal cytochromes P450, and by BH4-sufficient iNOS could be interpreted by considering the different possible fates of the Fe(II)-O2 intermediate of these hemeproteins (Mansuy et al., 1995; Moali et al., 2001). This weak oxidizing species either dissociates its iron-dioxygen bond leading to Fe(III) and O
In most cytochrome P450- or BH4-free iNOS-dependent oxidations, dissociation to Fe(III) and O
As far as the first question mentioned in the introduction (i.e., are there substrates other than Arg and NOHA that may be oxidized by NOS with formation of NO?) is concerned, the aforementioned results show that some N-aryl-N′-hydroxyguanidines do act as NOS substrates. Thus, the iNOS-catalyzed oxidation ofN-(4-chlorophenyl)-N′-hydroxyguanidine by NADPH and O2 selectively leads to the corresponding urea and NO in a 1:1 ratio, with a Vm value only 4 times lower than that found for NOHA (Renodon-Corniere et al., 1999). We have found that iNOS also catalyzes the oxidation of several other N-aryl-N′-hydroxyguanidines, with efficient formation of NO.
The best substrate for iNOS in this series so far isN-(4-fluorophenyl)-N′-hydroxyguanidine. TheVm of its oxidation is 50% of theVm of NOHA oxidation, even though itsKm remains 20-fold higher that the one measured for NOHA. Interestingly enough, such completely exogenous substrates have been also found for recombinant neuronal and endothelial NOS (nNOS and eNOS). These results suggest that it should be possible to find efficient and selective exogenousN-hydroxyguanidine substrates for each class of NOS isoforms. Such compounds could be used as selective NO donors after in situ oxidation by a given NOS.
In conclusion, our results show that microsomal cytochromes P450 and recombinant BH4-free NOS are able to catalyze the oxidative cleavage of the CNOH bond of a great number ofN-hydroxyguanidines, with formation of nitrogen oxides. Most of these reactions appear to derive from nonselective oxidations of these compounds by O
Acknowledgments
The authors (J.C.McG. and A.O.) thank Melody Steinberg for preparation of the manuscript and editorial assistance.
Footnotes
-
This work was supported in part by United States Public Health Service Grants GM53093 (E.T.M.), HL34300 and HL25394 (J.C.M.), HL59884 (A.O.), and HL52159 and DK58145 (W.B.C.); by UH1 03674 and an Established Investigator Award (0040119N) from the American Heart Association (A.O.); and by the Deutsche Forschungsgemeinschaft Schwerpunktprogramm “Radikale in der enzymatischen Katalyse” and the Fonds der Chemischen Industrie (V.U.).
- Abbreviations used are::
- NOS
- nitric oxide synthase
- P450
- cytochrome P450
- iNOS
- inducible NOS
- LPS
- bacterial lipopolysaccharide
- PB
- phenobarbital
- PN
- peroxynitrite
- SOD
- superoxide dismutase
- PGI2
- prostacyclin
- SIN-1
- 3-morpholinosydnonimine N-ethylcarbamide
- 3-NT
- 3-nitrotyrosine
- CPO
- chloroperoxidase
- AA
- arachidonic acid
- HETE
- hydroxyeicosatetraenoic acid
- l-NAME
- l-nitroarginine methyl ester
- GFR
- glomerular filtration rate
- UV
- urine volume
- UNaV
- urinary sodium excretion
- DBDD
- 12,12-dibromododec-11-enoic acid
- PT
- proximal tubules
- ET-1
- endothelin-1
- nNOS
- neuronal NOS
- SNP
- sodium nitroprusside
- ZG
- zona glomerulosa
- AII
- angiotensin II
- DETA nonoate
- (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl) amino]diazen-1-ium-1,2-diolate
- AdeNOS
- adenovirus encoding endothelial NOS
- FMO
- flavin-containing monooxygenase
- TMAU
- trimethylaminuria
- TMA
- trimethylamine
- TMAO
- trimethylamineN-oxide
- CVH
- chronic viral hepatitis
- RA
- ranitidine
- RANO
- ranitidine N-oxide
- SNAP
- S-nitroso-N-acetylpenicillamine
- BHA
- butylated hydroxyanisole
- DTT
- dithiothreitol
- NOHA
- Nω-hydroxyarginine
- BH4
- tetrahydrobiopterin
- homo-NOHA
- Nω-hydroxyhomo-l-arginine
- nor-NOHA
- Nω-hydroxynor-l-arginine
- dinor-NOHA
- Nω-hydroxydinor-l-arginine
- Received June 21, 2001.
- Accepted August 6, 2001.
- The American Society for Pharmacology and Experimental Therapeutics