Competition for tetrahydrobiopterin between phenylalanine hydroxylase and nitric oxide synthase in rat liver.

Tetrahydrobiopterin (BH4) is an important cofactor for two hepatic enzymes, inducible nitric oxide synthase (iNOS) and phenylalanine hydroxylase (PAH), and competition for BH4 between the two enzymes might limit hepatic iNOS or PAH activity. To test this hypothesis, we determined whether conversion of phenylalanine to tyrosine was modified by changes in NO synthase activity, and conversely whether NO synthesis was limited by the rate of phenylalanine conversion to tyrosine in rat hepatocytes and perfused livers. NO production was decreased only slightly, when flux through PAH was maximized in isolated perfused livers, and in isolated hepatocytes only when BH4 synthesis was inhibited. Increases in NO synthesis did not reduce tyrosine formation from phenylalanine. Phenylalanine markedly increased biopterin synthesis, whereas arginine had no effect. Thus, basal BH4 synthesis appears to be adequate to support iNOS activity, whereas BH4 synthesis is increased to support PAH activity.

Nitric oxide (NO) 1 has a diverse array of physiologic effects which may be either beneficial or detrimental, depending on the anatomic site and rate of NO synthesis. Accordingly, regulation of NO synthesis rates, particularly by the inducible form of NO synthase (iNOS), is critical in determining whether the net effect of NO will be good or ill. Rates of NO production by iNOS in intact cells can reflect regulatory events at many steps, including transcription, translational efficiency, and stability of mRNA and protein (1,2). Cellular NO synthesis rates also may be regulated by processes which affect availability of the various substrates and cofactors (e.g. arginine and tetrahydrobiopterin (BH4)) required for NO synthesis (2).
BH4 is an important cofactor for full activity of all NOS isoforms (3)(4)(5). Although initial reports indicated that BH4 directly participated in the enzymatic synthesis of NO, more recent evidence contradicts the earlier reports and indicates that BH4 is an allosteric effector of NOS which stabilizes the active dimeric form of the enzyme (4,6,7). The importance of BH4 in NO synthesis has been demonstrated by experiments which showed that inhibition of BH4 synthesis also inhibits NO synthesis (3,8,9) which is consistent with observations that iNOS expression and BH4 synthesis are coinduced in a number of cell types (10).
One potential mechanism for modulating NOS activity might be via competition for BH4 with another enzyme requiring the same cofactor. Such a possibility exists in several tissues including the liver. Phenylalanine hydroxylase (PAH) is a hydroxylase which catalyzes a coupled reaction in which phenylalanine is oxidized to tyrosine and BH4 is oxidized to the corresponding 4␣-hydroxytetrahydropterin (11,12). Thus, BH4 functions catalytically in this hydroxylation system. Conversion of 4␣-hydroxytetrahydropterin to the quinonoid dihydropterin derivative occurs via a dehydratase, and regeneration of BH4 is completed by action of the dihydropterin reductase which reduces the quinonoid dihydropterin to BH4 (13).
Under normal conditions, hepatocytes express high levels of PAH but no iNOS activity. The expression of iNOS can be induced in the liver by various stimuli such as endotoxin, cytokines or Corynebacterium parvum (14,15), which may result in competition between PAH and iNOS for BH4. However, PAH and iNOS differ in their interactions with BH4 in ways which may affect the competition. First, BH4 functions catalytically in the PAH system but not in the iNOS reaction. Second, the apparent K m for BH4 is much lower for the iNOS (0.02-0.3 M) (4) than for PAH (2 M) (13). Third, BH4 is a negative effector of PAH, binding to a high affinity pterin regulatory site on the enzyme to form an inactive enzyme complex at low phenylalanine concentrations and thereby reducing intracellular free BH4 concentrations (16,17). Furthermore, phenylalanine regulates the BH4 biosynthetic pathway by controlling its first and rate-limiting enzyme, GTP cyclohydrolase I, through interacting with the p35 GTP cyclohydrolase I regulator protein (18). The activity of GTP cyclohydrolase I is stimulated at high phenylalanine concentrations in the presence of p35 protein (18). Thus, free BH4 concentrations are regulated by phenylalanine levels. Consequently, both free BH4 levels and demand for BH4 by PAH will be greatest when phenylalanine levels are high, and free BH4 levels and demand will be low when phenylalanine levels are low. Nonetheless, the low apparent K m of iNOS suggests that it will successfully compete with PAH for BH4 at low and high phenylalanine concentrations.
By varying phenylalanine concentrations, we tested whether the level of PAH activity could affect BH4 levels and consequently limit hepatic NO production in livers induced to express iNOS. Competition between the two enzymes for BH4 might constitute a way to selectively inhibit hepatic iNOS. Conversely, we also tested whether synthetic rates of hepatic NO synthesis might inhibit hepatic conversion of phenylalanine to tyrosine. Previously, the BH4 requirement for iNOS and PAH activities has been studied independently and mostly in cultured cells. The present study is unique in simultaneously studying NO and tyrosine synthesis in both intact liver and isolated hepatocytes.

EXPERIMENTAL PROCEDURES
Animals-Male Sprague-Dawley rats (200 -240 g) were fasted for 24 h prior to the experiment, with free access to water, and anesthetized with sodium pentobarbital (Nembutal, 50 mg/kg intrapertoneally).
Liver Perfusion-Livers were perfused in situ with a Krebs-Henseleit-bicarbonate (KHB) buffer (118 mM NaCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 4.7 mM KCl, 26 mM NaHCO 3 , 2.5 mM CaCl 2 ) as described in detail previously (19). All components (from Sigma) were chosen to avoid any contamination with nitrate. The perfusate was oxygenated with a mixture of 95% O 2 , 5% CO 2 , and the pH was maintained at 7.40. Arginine and phenylalanine were added to the perfusate as described under "Results." After the surgical procedure, the livers were allowed to recover over 30 min using a nonrecirculation KHB perfusion. Because metabolite measurements are limited by low concentrations in a single pass model, the liver were perfused in a recirculation system during the experimental periods to allow accumulation of the end products of NO (NO 2 Ϫ ϩ NO 3 Ϫ ), tyrosine and total biopterin in the perfusate. The total volume of recirculation was 105 ml (40 ml for reservoir and 65 ml for system and tubing). Three recirculation periods of 20 min each, designated throughout as E 1 , E 2 , and E 3 , were performed. Each recirculation period was preceded by a 5-min rinse period during which the livers were perfused with the same buffer and chemicals as during the following experimental period. The rinse periods allowed the elimination of any toxic metabolites that may have accumulated during the previous period.
Livers (n ϭ 5) were perfused with KHB during the first recirculation period (E 1 ), KHB ϩ 0.5 mM phenylalanine during the second period (E 2 ), and KHB ϩ 5 mM phenylalanine in the third period (E 3 ). In a second group (n ϭ 5), to optimize NO 2 Ϫ ϩ NO 3 Ϫ release, 0.1 mM arginine was added to the previous perfusion conditions. Finally, to determine the effect of NO inhibition on tyrosine release, four additional livers were perfused with 5 mM N G -monomethyl-L-arginine (L-NMMA) in similar conditions (KHB, KHB ϩ 0.5 mM phenylalanine, KHB ϩ 5 mM phenylalanine).
The viability of perfused livers was assessed by oxygen consumption, perfusate potassium, and lactate dehydrogenase release at the end of each 20-min recirculation period as described previously (19). At the end of each experiment, the livers were weighed and then freeze-dried over 48 h. The wet/dry weight ratios were calculated to determine any swelling induced by the perfusion or by the C. parvum administration. All metabolic results were normalized to the dry weight of the liver.
Hepatocyte Isolation and Cell Cultures-Rat hepatocytes were isolated using a modification of the in situ collagenase B (Boehringer Mannheim) perfusion technique of Seglen (20). Hepatocytes were separated from nonparenchymal cells by differential centrifugation at 50 ϫ g, four times. Hepatocyte purity assessed by microscopy was greater than 98%, and viability consistently exceeded 95% by trypan blue exclusion. Hepatocytes were isolated from three separate rats, 5 days after a C. parvum injection (28 mg/kg intravenously).
Assays-To determine the NO release in the cell supernatant or in the perfusate, NO 2 Ϫ ϩ NO 3 Ϫ were measured by a procedure based on the Griess reaction, as described elsewhere (21,22). Tyrosine release and total biopterin were measured with a high performance liquid chromatography. Oxidized and reduced forms of biopterin were analyzed by the differential oxidation method of Fukushima and Nixon (23). The concentration of protein was determined by the method of Smith et al. (24) using the bicinchoninic acid protein assay reagent (Pierce).
Statistical Analysis-The results are expressed as mean Ϯ S.E. In each group, data were analyzed using one-way analysis of variance with repeated measurements. When the results were significant, the mean values were compared by the Fischer test. Comparisons between groups were analyzed using one-way analysis of variance or Student's t test.
Significance was established at a p value Ͻ0.05.

RESULTS
C. parvum injection was used to induce hepatic iNOS expression since we have previously characterized the in vivo induction of iNOS in rat liver during chronic inflammation triggered by C. parvum. A high level of iNOS induction is achieved in the intact liver by 5 days postinjection (19), and hepatocytes have been shown to be the major site of NO production in this model (14,15).

Tyrosine, and Biopterin Release in Perfused Livers
In the absence of phenylalanine, tyrosine release ( Fig. 1, upper) was low and unaffected by addition of arginine or L-NMMA to the perfusate. Phenylalanine (0.5 or 5 mM) increased tyrosine release, in KHB, KHB ϩ 0.1 mM L-arginine, and KHB ϩ 5 mM L-NMMA perfusions. Tyrosine release was not affected by the addition of arginine, whereas L-NMMA perfusion, which completely blocked NO 2 Ϫ ϩ NO 3 Ϫ release, increased tyrosine release at 5 mM phenylalanine concentration compared to KHB perfusion.
Consistent with our previous observations (19), in the absence of added arginine, perfused livers released high levels of NO 2 Ϫ ϩ NO 3 Ϫ (Fig. 1, middle). This release was increased by the addition of 0.1 mM arginine, whereas L-NMMA, as already mentioned, completely blocked this release. Addition of phenylalanine (0.5 or 5 mM) slightly decreased NO 2 Ϫ ϩ NO 3 Ϫ release in the three conditions (KHB, KHB ϩ 0.1 mM arginine, and KHB ϩ L-NMMA). Both concentrations of phenylalanine induced a marked rise in biopterin release (Fig. 1, lower), whereas neither arginine or L-NMMA had an effect on the release of biopterin from the perfused livers.
When phenylalanine (0.5 and 5 mM) was added to the KHB perfusate, hepatic oxygen consumption significantly increased from 5.8 Ϯ 0.5 to 7.5 Ϯ 0.5 mol/min/dry weight g, probably as a result of the catabolism of the carbon skeleton of phenylalanine. Addition of arginine or L-NMMA had no effect on the hepatic oxygen consumption. Liver injury increased slightly during the KHB perfusion since lactate dehydrogenase release increased from 17.8 Ϯ 2.7 to 87.2 Ϯ 9.0 milliunits/min/dry weight g from E 1 to E 3 . However, this injury was not associated with a decline in the biosynthetic capacity of the livers, since tyrosine and biopterin release were similar during the early (E 2 ) and late (E 3 ) perfusion periods. L-NMMA perfusion significantly worsened the hepatic injury compared to the KHB perfusion (257.2 Ϯ 51.7 versus 87.2 Ϯ 9.0 milliunits/min/dry weight g, during E 3 -perfusion).

NO 2 Ϫ ϩ NO 3 Ϫ , Tyrosine, and Biopterin Release in Hepatocytes
In the Absence of Phenylalanine-In order to allow more specific manipulation of the two pathways, the relationships between iNOS and PAH activities and BH4 availability were also studied in hepatocytes isolated from rats previously in-jected with C. parvum. Cultured hepatocytes, in amino acidfree incubation medium, released small amount of tyrosine (42.0 Ϯ 7.6 nmol/1.5 ϫ 10 6 cells/18 h) and moderate levels of NO 2 Ϫ ϩ NO 3 Ϫ (164.0 Ϯ 4.0 nmol/1.5 ϫ 10 6 cells/18 h) in the supernatant. Addition of 0.5 mM arginine caused a 2.7-fold increase in NO 2 Ϫ ϩ NO 3 Ϫ release but had no effect on tyrosine release. NO synthesis in the hepatocytes exhibited some dependence on the de novo biosynthesis of BH4, since NO 2 Ϫ ϩ NO 3 Ϫ release decreased by 16.5% when DAHP was added in the medium. DAHP did not influence the low basal levels of tyrosine synthesis seen in the absence of phenylalanine. Thus, iNOS in the hepatocytes is slightly dependent on de novo BH4 biosynthesis but any increase in BH4 requirement for iNOS activity with arginine addition had no consequence on PAH activity.
In the Presence of Phenylalanine-In the absence of arginine (Fig. 2, upper), increasing phenylalanine concentrations in the medium increased tyrosine release. In these conditions, the high requirement of BH4 for PAH has only a slight impact on NO 2 Ϫ ϩ NO 3 Ϫ release at 0.5 mM phenylalanine concentration.
When 0.5 mM arginine was added in the medium (Fig. 2, middle), the addition of phenylalanine had no effect on NO 2 Ϫ ϩ NO 3 Ϫ . When de novo BH4 biosynthesis was inhibited by DAHP (Fig. 2, lower), the release of tyrosine associated with phenylalanine addition was dramatically reduced. In the presence of DAHP, the addition of phenylalanine, which only slightly increased tyrosine release, resulted in a concentration-dependent decrease in NO 2 Ϫ ϩ NO 3 Ϫ . This indicates that iNOS activity is likely to be limited only when BH4 levels are dramatically lowered by the inhibition of BH4 synthesis plus increased flux through PAH.
We then tested the capacity of exogenous BH4 to overcome the effects of DAHP and phenylalanine. As shown in Table I, the addition of 100 M BH4 increased basal NO 2 Ϫ ϩ NO 3 Ϫ release and also overcame the inhibitory effects of DAHP and phenylalanine on NO 2 Ϫ ϩ NO 3 Ϫ release. BH4 also overcame the inhibitory effect of DAHP on tyrosine release in the presence of phenylalanine but did not increase basal levels of tyrosine release in the presence of phenylalanine. As shown in Table II, the addition of phenylalanine increased total biopterin release, although intracellular levels were not affected. Also shown is the suppression of intracellular biopterin and BH4 levels in DAHP-treated cells and even further reductions when phenyl- alanine is combined with DAHP, which is consistent with the effects of DAHP on NO 2 Ϫ ϩ NO 3 Ϫ and tyrosine release.

DISCUSSION
BH4 is an essential cofactor for several enzymes, including aromatic acid hydroxylases and all three NO synthases (13). When enzymes requiring BH4 coexist in a cell, it is not known how BH4 is regulated or whether competition for BH4 will take place, limiting enzyme activity. It is unlikely that BH4 would be limiting in sites, such as the brain or adrenal glands, where a low output or constitutive NO synthase coexists with aromatic acid hydroxylases, due to the relatively low activity of the constitutive NO synthase enzymes. In this regard, hepatocytes are unique, containing high PAH activity in the presence of activating concentrations of phenylalanine (25) and expressing high levels of the high output or inducible NOS during inflammatory conditions (14). Using both isolated perfused livers and cultured hepatocytes from rats prestimulated to express high levels of iNOS, we studied whether the availability of BH4 could limit NO or tyrosine synthesis. We found little evidence that competition for BH4 limited the activity of either PAH or iNOS. This finding was true even if flux through PAH or iNOS were maximized by the addition of excess substrate. Under conditions in which de novo BH4 production was limited by DAHP, PAH was more susceptible to inhibition than iNOS. Furthermore, only when de novo BH4 synthesis was inhibited did increased flux through PAH limit NO synthesis. Our results also indicate that, whereas basal BH4 production was adequate to maintain near-maximal NO synthesis, PAH activity was maintained by increasing biopterin synthesis.
Several reasons may explain that increased flux through PAH has little effect on NO synthesis. First, even though the utilization of BH4 by PAH is relatively high (1 mol of BH4 per mol of tyrosine produced) (11), phenylalanine stimulates an increase in biopterin synthesis to meet the BH4 requirement of PAH (17,18). This production is likely to take place through the interaction of phenylalanine with the recently described feedback regulator protein for GTP cyclohydrolase I, resulting in increased activity of this rate-limiting enzyme for biopterin biosynthesis (18). Second, the K m for BH4 for the NOS is 0.02 to 0.3 M (4) and minor changes in BH4 availability would not be expected to have a major effect on NOS activity. Furthermore, there is no evidence for stoichiometric BH4 turnover by NOS. In fact, Giovanelli et al. (4) demonstrated that the NOS produced 18 mol of NO per mol of BH4. Studies on murine (6) and human (7) iNOS indicate that BH4 functions to maintain the enzyme in an active configuration.
Similar arguments are likely to explain the minimal effect of increased flux through iNOS on the conversion of phenylalanine to tyrosine by PAH. Here, the low utilization of BH4 by iNOS would be expected to have little impact in the face of increased BH4 production, resulting from the interaction of phenylalanine and GTP cyclohydrolase I (18). Furthermore, as BH4 is utilized, any negative feedback exerted by BH4 on GTP cyclohydrolase I would be removed. This may explain our finding that BH4 and biopterin content within the cells remain constant. We did observe a slight increase in tyrosine production when NOS activity was blocked with L-NMMA. Whether this is due to increased BH4 availability is unclear. Since NO can interfere with a number of enzyme systems, such as mitochondrial respiration (26), nonspecific effects on cellular metabolism must also be considered.
Our in vitro experiments using excess exogenous BH4 also suggest important differences in the utilization and actions of this cofactor by the two enzymes. Added BH4 increased NO 2 Ϫ ϩ NO 3 Ϫ release both in the presence and absence of phenylalanine, whereas the addition of phenylalanine alone, which stimulates BH4 synthesis, did not. The observation that only extracellular release of biopterin and not intracellular levels of BH4 increased with phenylalanine addition suggests that the higher quantities of BH4 produced with phenylalanine addition are rapidly oxidized by PAH and released into medium. Thus, this BH4 may not be available to iNOS. It is also interesting to note that excess BH4 suppressed tyrosine release. The reason for this is not readily apparent, but high levels of BH4 have been a Hepatocytes from C. parvum-treated rats were harvested and incubated (4.5 ϫ 10 6 cells/plate) in the incubation medium (5 ml) containing no amino acids (except 0.5 mM arginine) and substrates as indicated. Results represent mean Ϯ standard error of the mean of two separate cultures of hepatocytes from three separate animals. b p Ͻ 0.05 versus none. c p Ͻ 0.05 versus phenylalanine.
a Hepatocytes from C. parvum-treated rats were harvested and incubated (4.5 ϫ 10 6 cells/plate) in the incubation medium (5 ml) containing no amino acids (except 0.5 mM arginine) and substrates as indicated. Results represent mean Ϯ standard error of the mean of two separate cultures of hepatocytes from three separate animals. b p Ͻ 0.05 versus similar substrates without DAHP. c Not determined.
shown to block phenylalanine-induced activation of PAH in a preparation of purified enzyme (28). Our cell culture experiments also show that up to 40% of the biopterin in the cells is not BH4. If it were BH2, it might be expected that the BH2 could inhibit NOS activity (29); however, because BH2 binds to NOS with an affinity 10-fold lower than BH4, it is unlikely to effectively compete for binding to NOS in the presence of higher levels of BH4. Furthermore, much of the measured total biopterin may be in the form of biopterin which does not bind to NOS. Any BH2 in the cultures should be converted to BH4 by dihydrofolate reductase. The high levels of oxidized biopterin in the cultured cells (40%) compared to that reported for whole liver (10%) may be due to uptake of oxidized biopterin in the medium by the cells in culture. When one considers the major differences in the functions of iNOS and PAH, as well as the differences in the functions of BH4 in each enzyme, it is not surprising that BH4 availability to support the activity of each enzyme is also differentially regulated. PAH is a constitutively expressed enzyme which must respond to rapid changes in ambient phenylalanine concentrations. PAH also has a higher K m for BH4 than iNOS and utilizes BH4 at a much faster rate. To support the high demand for BH4 in this metabolic pathway, increases in phenylalanine lead to a rapid increase in BH4 availability for the PAH enzyme. Whereas PAH is part of a constitutive metabolic pathway, iNOS, which functions to form NO from arginine, is expressed in hepatocytes only under inflammatory conditions and is a biosynthetic pathway. We have found that arginine for this reaction is most likely derived from exogenous sources, although an endogenous source not associated with the urea cycle also exists (19). As the K m of iNOS for BH4 is low, ambient intracellular BH4 levels, which appear to remain constant, are adequate to support iNOS activity. However, our finding that exogenous BH4 increases NO synthesis in hepatocytes raises the possibility that intracellular levels of BH4 may not support maximal NO synthesis, at least in cultured cells. We recently showed that exogenous BH4 increased NOdependent killing of Plasmodium falciparum in human hepatocytes (27). Whether these effects are only an in vitro phenomenon and whether exogenous BH4 will be a useful therapeutic approach to modify NO synthesis remains to be determined.
Thus, we conclude that the rate of flux through either iNOS or PAH is unlikely to limit the activity of the other enzyme.