Introduction

Top Abstract Introduction Materials and methods Results Discussion References

Tertiary amines frequently have been used as medicines. They are generally oxidized at the nitrogen or the -carbon atom of the tertiary amine moiety and, consequently, are transformed to N-oxides and secondary amines. The N-oxides thus formed are frequently reduced back to the parent tertiary amines. This interconversion between tertiary amines and N-oxides is a well known metabolic pathway (Bickel, 1969). Such interconversion in the body might be effective to maintain the pharmacological effect of amines. The enzyme system involved in the N-oxide reduction has been investigated in the livers of animals, and a role of the cytochrome P-450 system in the microsomal reduction has been demonstrated (Sugiura et al., 1974, 1976; Iwasaki et al., 1977).

Recently, we investigated the reduction of various types of N-oxides to the corresponding amines by mammalian liver cytosols using nicotinamide N-oxide, imipramine N-oxide, cyclobenzaprine N-oxide, and S-()-nicotine-1'-N-oxide as substrates in rats and rabbits. The mammalian liver cytosols exhibited a significant N-oxide reductase activity toward these N-oxides when supplemented with an electron donor of aldehyde oxidase, such as 2-hydroxypyrimidine or N¹-methylnicotinamide, and the flavoenzyme functioned as an N-oxide reductase in liver cytosol (Kitamura and Tatsumi, 1984a,b; Sugihara et al., 1996).

In the course of studies on the mechanism for tertiary amine N-oxide reduction in mammalian liver, it was found that liver cytosol exhibited a significant reductase activity toward tertiary amine N-oxides when a quinone such as menadione, an inhibitor of aldehyde oxidase, was added together with NADH or NADPH. This unexpected result led us to investigate in detail the mechanism of the quinone-mediated tertiary amine N-oxide reduction in liver cytosol. We reported in a brief communication that imipramine N-oxide was reduced to imipramine by rat liver cytosol in the presence of menadione and NADH or NADPH (Kitamura and Tatsumi, 1997). In the present investigation, this novel quinone-dependent tertiary amine N-oxide reduction in the mitochondria, microsomes, and cytosol fractions of rat liver was further examined mainly using imipramine N-oxide as a substrate.

	Materials and Methods

Top Abstract Introduction Materials and methods Results Discussion References

Materials. Menadione, p-benzoquinone, 1,4-dihydroxybenzene, 1,4-naphthoquinone, 9,10-anthraquinone, 9,10-phenanthrenequinone, 1,4-dihydroxynaphthalene, sodium anthraquinone -sulfonate, and protoporphyrin IX disodium salt were purchased from Nacalai Tesque Inc. (Kyoto, Japan). 1,5-Dihydroxynaphthalene, 1,4,9,10-tetrahydroxyanthracene, 1,4-dihydroxy-9,10-anthraquinone, 2,3-dihydroxynaphthalene, 1,3-dihydroxynaphthalene, tetrahydroxy-p-benzoquinone, nicotinamide, nicotinamide N-oxide, and tetramethyl-p-benzoquinone were from Tokyo Chemical Co. Ltd. (Tokyo, Japan). Bovine blood hemoglobin, doxorubicin (Adriamycin; Pharmacia & Upjohn, Kalamazoo, MI), horse skeletal muscle myoglobin, bovine liver catalase, horse heart cytochrome c, horseradish peroxidase, bovine blood hematin, dicumarol, NADH, and NADPH were from Sigma Chemical Co. (St. Louis, MO). Imipramine hydrochloride, 1,2-dihydroxynaphthalene, brucine N-oxide, and brucine were from Aldrich Chemical Co. (Milwaukee, WI). Cyclobenzaprine hydrochloride and cyclobenzaprine N-oxide were donated by Merck Sharp & Dohme Research Laboratories (Rahway, NJ). Imipramine N-oxide was prepared by the H₂O₂-oxidation method (Dietrich, 1967). Reduced menadione was prepared by the method of Fieser (1940).

Animals. Male Wistar (Slc:Wistar/ST) rats (180-230 g), ddY mice (25-32 g), Syrian golden hamsters (110-140 g), Hartley guinea pigs (215-260 g), and Japanese albino rabbits (2.6-3.2 kg) were used. The former three and the latter two species were fed standard pellet diets MM-3 and RM-4 (Funabashi Farm, Chiba, Japan), respectively.

Liver Preparations. Animals were stunned by a blow on the head and exsanguinated. Livers were immediately perfused extensively with saline and homogenized in 4 volumes of the KCl solution, using a Potter-Elvehjem homogenizer. The homogenate was centrifuged at 900g for 10 min. Mitochondrial fraction was obtained from the supernatant by centrifugation at 9000g for 20 min. Microsome and cytosol fractions were obtained from the supernatant by centrifugation at 105,000g for 60 min. The mitochondria and microsome fractions were washed by resuspension in the KCl solution and resedimentation. The pellets of these fractions were suspended in the KCl solution to make 1 ml equivalent to 1 g of liver.

Purification of DT-Diaphorase from Rat Liver. DT-diaphorase was prepared from livers of 3-methylcholanthrene-treated rats according to the method of Yoshimura et al. (1987) with a slight modification. The cytosol obtained from 22 g of liver was applied to diethylaminoethanol cellulose (DE-52), hydroxyapatite, and TSK-gel Toyopearl 650 ML (AF-Red) columns successively for purification. The purified enzyme showed a single protein band upon SDS-PAGE.

Assay of Tertiary Amine N-Oxide Reduction. The incubation mixture consisted of 0.2 µmol of imipramine N-oxide, 0.2 µmol of quinone, 0.5 µmol of NADH or NADPH, and mitochondria, microsomes, or cytosol equivalent to 2 to 4 mg of protein, 0.1 to 0.5 mg of hemoproteins, or 3 µg of hematin in a final volume of 1 ml of 0.1 M potassium/sodium-phosphate buffer (pH 7.4). In some experiments, 1 µmol of reduced quinone such as menadiol or 1,4-dihydroxynaphthalene was used instead of both a quinone and a reduced pyridine nucleotide. The incubation was performed using a Thunberg tube under anaerobic conditions. The side arm contained the N-oxide, and the body contained all other components. The tube was gassed for 5 min with deoxygenated nitrogen, evacuated with an aspirator for 5 min, and again gassed with the nitrogen. The reaction was started by mixing the components of the side arm and the body, continued for 30 min at 37°C under anaerobic conditions, and stopped by adding 0.1 ml of 1 N NaOH. The mixture, after addition of 10 µg of phenacetin as an internal standard, was extracted once with 5 ml of ethyl acetate, and the extract was evaporated to dryness in vacuo. The residue was dissolved in 0.1 ml of methanol, and then a 5-µl aliquot was subjected to high-performance liquid chromatography (HPLC)¹ on a Hitachi L-6000 chromatograph fitted with a 125- × 4-mm Inertsil ODS-2 (GL Sciences Inc., Tokyo, Japan) column. The mobile phase was acetonitrile, 0.1 M KH₂PO₄ (3:7). The chromatograph was operated at a flow rate of 0.8 ml/min at ambient temperature, with detection at 254 nm. The elution times of imipramine N-oxide and its reduction product, imipramine, were 19.5 and 17.7 min, respectively. The reduction product was determined from its peak area. Some properties of the reaction were examined by using rat liver cytosol in the presence of NADH and menadione. The reaction proceeded linearly for at least the initial 60 min and with increasing amounts of protein of the cytosol up to 8 mg.

Assay of Quinone Reduction. Quinone reductase activity was assayed by measuring the absorbance at 550 nm due to the formation of reduced cytochrome c (Yoshimura et al., 1987). The NADH- and NADPH-linked quinone reductase activities in rat liver cytosol were 206 and 186 nmol/min/mg protein, respectively.

Measurement of Heme Contents. Heme content was determined by the pyridine hemochrome method with hematin as a standard heme (De Duve, 1948). The heme content in rat liver cytosol was 0.93 µg/ml.

Measurement of Protein Contents. Protein contents were determined by the method of Lowry et al. (1951) with bovine serum albumin as a standard protein.

	Results

Top Abstract Introduction Materials and methods Results Discussion References

Quinone-Dependent Tertiary Amine N-Oxide Reduction. The reduction of imipramine N-oxide to imipramine by rat liver subcellular preparations was examined by addition of several cofactors. The liver mitochondria, microsomes, or cytosol exhibited only weak N-oxide reductase activity even with NADH or NADPH under anaerobic conditions. The NAD(P)H-linked N-oxide reductase activities in mitochondria, microsomes, and cytosol of rat livers were enhanced significantly by the addition of menadione, except for the mitochondrial NADPH-linked activity. In particular, the activity in liver cytosol was enhanced markedly by menadione, and, in this case, the NADH- and NADPH-linked activities were equally enhanced (Table 1). Boiling the liver mitochondria, microsomes, or cytosol destroyed the activity for reduction in the presence of both menadione and the reduced pyridine nucleotide. The quinone-dependent reduction in the presence of NAD(P)H was also observed under aerobic conditions, but the activity was less than one-tenth of that under anaerobic conditions (data not shown). The liver cytosols of mice, hamsters, rabbits, and guinea pigs, like those of rats, had the ability to reduce the tertiary amine N-oxide in the presence of both NAD(P)H and menadione under anaerobic conditions. Among the mammalian species examined, the highest activity was observed with mice (Table 2).

Ability of Various Quinones to Replace Menadione in the Tertiary Amine N-Oxide Reduction. The cytosolic N-oxide reductase activity of rat liver was also observed when 1,4-naphthoquinone, 9,10-anthraquinone, sodium anthraquinone -sulfonate, or 9,10-phenanthrenequinone was added instead of menadione in the presence of NADH under anaerobic conditions. However, p-benzoquinone and tetramethyl-p-benzoquinone were not effective in the reduction system (Table 3).

Components Responsible for the Tertiary Amine N-Oxide Reduction. The menadione-dependent tertiary amine N-oxide reduction in the presence of NADH or NADPH was inhibited markedly by dicumarol, cupric sulfate, and carbon monoxide, but not by potassium cyanide or sodium arsenite (Fig. 1). Of these inhibitors, dicumarol is an effective inhibitor of DT-diaphorase. These observations suggest the involvement of DT-diaphorase and hemoproteins in the N-oxide reduction. The diaphorase, a liver cytosolic enzyme, with NADH or NADPH plays a well recognized role in the conversion of quinones, including menadione, to their reduced forms by two-electron reduction (Ernster, 1967, 1987; Iyanagi and Yamazaki, 1970). In fact, catalase or hemoglobin as a model hemoprotein exhibited a significant N-oxide reductase activity when DT-diaphorase purified from rat liver was added in the presence of both NADH and menadione under anaerobic conditions. The N-oxide-reducing activities of hemoproteins were inhibited by the addition of dicumarol or carbon monoxide (Fig. 2). These activities were not observed upon addition of boiled DT-diaphorase (data not shown).

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Influence of some chemicals on menadione-dependent N-oxide reduction by rat liver cytosol. Each value represents the mean ± S.D. of four experiments. Reactions were conducted with 0.2 ml of liver cytosol equivalent to 0.04 g of liver in the presence of both NAD(P)H and menadione at 37°C for 30 min under anaerobic conditions. Dicumarol, cupric sulfate, potassium cyanide, and sodium arsenite were added at the concentrations of 1 × 104, 2 × 104, 1 × 103, and 1 × 103 M, respectively. Other details are described in Materials and Methods.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Reduction of imipramine N-oxide by the combination of hemoglobin or catalase and DT-diaphorase. Each value represents the mean ± S.D. of four experiments. Reactions were conducted with 0.1 mg of hemoglobin or 0.5 mg of catalase with DT-diaphorase (0.06 mg protein) in the presence of both NADH and menadione at 37°C for 30 min under anaerobic conditions. Dicumarol was added at the concentration of 1 × 104 M. Other details are described in Materials and Methods. ND, not detected.

Dihydroquinone-Dependent Tertiary Amine N-Oxide Reduction. When menadione was replaced with reduced menadione, menadiol, the reduction of imipramine N-oxide by rat liver cytosol occurred significantly in the absence of NADH or NADPH under anaerobic conditions. The tertiary amine N-oxide reduction in the presence of menadiol was inhibited markedly by carbon monoxide and oxygen, but not dicumarol (Fig. 3). The N-oxide reductase activity was also observed when 1,4-dihydroxynaphthalene, 1,4,9,10-tetrahydroxyanthracene, 1,4-dihydroxy-9,10-anthraquinone, or doxorubicin was added instead of menadiol. However, 1,4-dihydroxybenzene, tetrahydroxy-p-benzoquinone, 1,2-dihydroxynaphthalene, 1,3-dihydroxynaphthalene, 1,5-dihydroxynaphthalene, or 2,3-dihydroxynaphthalene was not so effective in the cytosolic tertiary amine N-oxide reduction (Table 4).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Reduction of imipramine N-oxide by rat liver cytosol in the presence of menadiol. Each value represents the mean ± S.D. of four experiments. Reactions were conducted with rat liver cytosol in the presence of 1 µmol of menadiol at 37°C for 30 min under anaerobic conditions. Dicumarol was added at the concentration of 1 × 104 M. Other details are described in Materials and Methods.

Activity of Various Hemoproteins and Hematin in the Dihydroquinone-Dependent Tertiary Amine N-Oxide reduction. Hemoproteins such as hemoglobin, myoglobin, peroxidase, cytochrome c, and catalase catalyzed the reduction of imipramine N-oxide in the presence of menadiol or 1,4-dihydroxynaphthalene to varying degrees under anaerobic conditions (Table 5). The N-oxide-reducing activity of hemoglobin was inhibited completely by carbon monoxide (data not shown). The activity was not abolished with boiled hemoproteins; on the contrary, the activity was increased when boiled catalase, peroxidase, or cytochrome c was used. Hematin was also effective in the presence of the dihydroquinones, whereas protoporphyrin and iron chlorides were ineffective in the N-oxide reduction (Table 5). The activity of hematin was not abolished by boiling or in an atmosphere of carbon monoxide (data not shown). The results led us to speculate that the heme group of hemoproteins catalyzes the reduction of imipramine N-oxide with dihydroquinones as an electron donor.

Tertiary Amine N-Oxide Reduction by Hemoglobin and SH Reagents. The N-oxide reductase activity of hemoglobin was also examined using SH reagents instead of dihydroquinones. The imipramine N-oxide reductase activity was exhibited in the presence of hemoglobin and dithionite, dithiothreitol, or 2-mercaptoethanol to varying degrees (Fig. 4). The N-oxide reductase activity of hemoglobin with dithionite was inhibited by carbon monoxide. These activities were not observed in the absence of hemoglobin (data not shown). The observations suggested that a mild reducing agent such as diol or SH reagent is involved in the cytosolic N-oxide reduction by heme.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Reduction of imipramine N-oxide by hemoglobin in the presence of SH reagents. Each value represents the mean ± S.D. of four experiments. Reactions were conducted with 0.1 mg of hemoglobin in the presence of SH reagents at 37°C for 30 min under anaerobic conditions. Other details are described in Materials and Methods.

Substrate Specificity of the Quinone-Dependent N-Oxide Reduction. Tertiary amine N-oxides such as cyclobenzaprine N-oxide and brucine N-oxide were also reduced, like imipramine N-oxide, to the corresponding amines by rat liver cytosol when NADH and menadione were added. However, the cytosolic reducing activity toward nicotinamide N-oxide, an aromatic N-oxide, was not enhanced by the addition of menadione (Table 6).

	Discussion

Top Abstract Introduction Materials and methods Results Discussion References

Quinones occur naturally in animals, plants, and fungi. The importance of quinones in such basic metabolic processes as respiration and photosynthesis is well established (Mitchell, 1976; Nohl et al., 1986; Cadenas and Hochstein, 1992). Anthraquinones (e.g., daunorubicin) are used clinically for the treatment of a wide variety of malignancies (Driscoll et al., 1974; Nohl et al., 1986; Riley and Workman, 1992). In mammalian livers, both endogenous and exogenous quinones are reduced by NADPH- or NADH-oxidizing flavoenzymes such as microsomal NADPH-cytochrome P-450 reductase and NADH-cytochrome b₅ reductase, or mitochondrial NADH dehydrogenase, which are one-electron-transferring quinone reductases (Iyanagi and Yamazaki, 1969; Bachur et al., 1979; Powis and Appel, 1980; Turrens and Boveris, 1980). DT-diaphorase, a two-electron-transferring quinone reductase, is also present in liver cytosol (Iyanagi and Yamazaki, 1970; Hosoda et al., 1974). The present study indicated that quinone reductases, which are widely distributed among liver subcellular fractions, play a role in the tertiary amine N-oxide reduction with NAD(P)H and quinones in rat liver preparations.

The quinone-dependent tertiary amine N-oxide reduction was observed widely in mitochondria, microsomes, and cytosol of rat liver in the presence of both NAD(P)H and menadione. These activities were also widely distributed among species, i.e., rats, mice, guinea pigs, hamsters, and rabbits. The quinone-dependent N-oxide-reducing activity was one order of magnitude greater than the N-oxide reductase activity of the aldehyde oxidase or cytochrome P-450 system described in the literature (Kitamura and Tatsumi, 1984b). Thus, the quinone-dependent reduction appears to function as a major system in the reduction of tertiary amine N-oxides in liver preparations. It is noteworthy that the N-oxidation of tertiary amines was suggested to be a route of detoxification (Bickel, 1969). If this is so, the system demonstrated in this study would result in the pharmacological reactivation of tertiary amines from the N-oxides.

It is well known that tertiary amines are oxidized to the corresponding N-oxides in liver microsomes (Ziegler, 1980). The reduction kinetics of tertiary amine N-oxides to the amines is complex because of the reoxidation of tertiary amine to the parent N-oxide. In our preliminary study, when imipramine was incubated with liver preparations in the presence of NADPH under aerobic conditions, imipramine was oxidized to imipramine N-oxide. The oxidation rates in mitochondria, microsomes, and cytosol of rat liver were 9.6, 10.3, and 0.3 nmol/30 min/mg protein, respectively. These activities are very low compared with those of the reduction of imipramine N-oxide to the amine in liver preparations. High activity in the retroreduction is of major relevance for the maintenance of the concentration of active tertiary amines. When imipramine was incubated similarly under anaerobic conditions, the formation of imipramine N-oxide was not observed in liver mitochondria, microsomes, or cytosol, even in the presence of NADPH. Under aerobic conditions, the velocity of the N-oxide reduction in liver preparations decreased to about one-tenth of that under anaerobic conditions. Perhaps this is because of the competition of the substrate with oxygen at the active site. However, the reoxidation of imipramine to the N-oxide may contribute to the decreased formation of the reduction product.

It has been reported that amidines, amidinohydrazones, and guanidines are oxidized to their N-hydroxy metabolites, and retroreduction to the parent compounds occurs in liver microsomes (Clement et al., 1993, 1994; Clement and Jung, 1994; Clement and Demesmaeker, 1997). The reducing activity was much higher than the oxidative in liver microsomes, which is similar to the case of interconversion between tertiary amines and their N-oxides. This may be the reason why N-hydroxy metabolites were not detected in the in vivo experiments. The reduction system resembles in many respects the microsomal hydroxylamine reductase system demonstrated by Kadlubar et al. (1973) and Kadlubar and Ziegler (1974), which is composed of NADH, NADH-cytochrome b₅ reductase, cytochrome b₅, and a third, unidentified protein component. Recently, Clement et al. (1997) purified the third component from pig liver microsomes as a constituent of microsomal benzamidoxime reductase and demonstrated that the protein has many cytochrome P-450-like characteristics. Benzamidoxime was reduced to benzamidine by the reconstituted system, which required NADH, NADH-cytochrome b₅ reductase, cytochrome b₅, phospholipid, and cytochrome P-450-like benzamidoxime reductase. Our tertiary amine N-oxide reduction system is different from the microsomal benzamidoxime reductase or hydroxylamine reductase system with respect to the oxygen sensitivity, the cofactor requirement, and the components. It would be interesting to see whether the quinone-dependent tertiary amine N-oxide reduction system demonstrated in this study functions as a reductase of N-hydroxylated metabolites of guanidines, amidines, and amidinohydrazones or hydroxylamines.

Based on the present results, the following mechanism for the reduction of tertiary amine N-oxides by mammalian liver preparations is proposed. The reductase system consists of NAD(P)H, a quinone, quinone reductase, and heme. In the first step, the quinone is reduced to the dihydroquinone by a quinone reductase such as DT-diaphorase in the presence of NAD(P)H. In the second step, the tertiary amine N-oxide is catalytically reduced by the heme group of a hemoprotein in the presence of the dihydroquinone as the electron donor; in this step, SH reagents also function as reducing agents of heme in place of dihydroquinones. The reduction of the quinone by quinone reductase in the first step is assumed to be the rate-limiting step in tertiary amine N-oxide reduction. We cannot rule out the possibility of a low level of contamination with hemoglobin from blood in the liver preparations, and this hemoprotein may act as a catalyst in the tertiary amine N-oxide reduction. However, the tertiary amine N-oxide reduction mediated by hemoglobin appears to occur in blood with dihydroquinones or SH reagents. In preliminary experiments, we found that when hematin was incubated with 1,4-dihydroxynaphthalene, the UV maximum at 450 nm due to the oxidized form of hematin was changed to 380 nm due to the reduced form and, upon addition of imipramine N-oxide, returned to 450 nm. This suggested that interconversion between oxidized and reduced forms of hematin was involved in the catalytic action (Fig. 5). In the second step, p-dihydroxy aromatic hydrocarbons were effective in the tertiary amine N-oxide reduction, but o- or m-dihydroxy hydrocarbons were not. The p-dihydroxy structure of the aromatic rings might be essential for the catalytic action by heme.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Proposed mechanism for quinone-dependent N-oxide reduction in rat liver cytosol.

There are only a few reports of the involvement of hemoproteins in the reduction of tertiary amine N-oxides. They cover the reduction of trimethylamine N-oxide by hemoglobin in the presence of cysteine, the reduction of indicine N-oxide by cytochrome c in the presence of ascorbic acid and by denatured hemoglobin, and the reduction of 4-bromo-N,N-dimethylaniline N-oxide by ferrihemoglobin or ferricytochrome c (Vaisey, 1956; Terayama, 1963; Powis and DeGraw, 1980). However, the quinone-dependent tertiary amine N-oxide reduction demonstrated in this study has not been reported previously. This quinone-dependent reduction catalyzed by heme appears to function as a major N-oxide-reducing system for various tertiary amine N-oxides, including the three N-oxide compounds listed above.

Tertiary amine N-oxides, such as imipramine N-oxide, cyclobenzaprine N-oxide, and brucine N-oxide, were reduced markedly by the N-oxide-reducing system of liver cytosol described in this study. However, nicotinamide N-oxide, an aromatic-type N-oxide, was not reduced by this system. Tertiary amine N-oxides were also reduced by the cytochrome P-450 system in liver microsomes, although nicotinamide N-oxide was not (Kitamura and Tatsumi, 1984a,b). In contrast, aldehyde oxidase exhibited reductase activities toward both tertiary amine N-oxide and nicotinamide N-oxide, and extraordinarily high reductase activity toward nicotinamide N-oxide was observed in this case (Kitamura and Tatsumi, 1984a,b). There is a distinct difference between tertiary amine N-oxides and aromatic N-oxides from the viewpoint of biological reduction.