Invited Review - ACS Publications - American Chemical Society

toxicology will be presented for the mammalian flavin- containing monooxygenases. Pig Liver Flavin-Containing Monooxygenase. In the late 19709, a hepa...
0 downloads 0 Views 3MB Size
MARCH 1995 VOLUME 8, NUMBER 2 0Copyright 1995 by the American Chemical Society

Invited Review Structural and Catalytic Properties of the Mammalian Flavin-Containing Monooxygenase John R. Cashman” Seattle Biomedical Research Institute, 4 Nickerson Street, Seattle, Washington 98109 Received July 28, 1994

Introduction Prior to the 1960s, it was assumed that most, if not all, NADPH-dependent xenobioticheteroatom-containing compound oxidation was mediated by the microsomal cytochrome P-450 family of monooxygenases. Now, it is recognized that the flavin-containing monooxygenase catalyzes oxygenation of nucleophilic nitrogen-, sulfur-, phosphorus-, and other heteroatom-containing chemicals, drugs, and agricultural agents (1-4). The observation that arylamines were oxidized to more reactive metabolites that possessed increased toxic or carcinogenic potential helped to propagate the view that N-oxygenation of all amines leads to metabolites with increased toxicity (5). Today, it is recognized that while this may be the case for a number of arylamines, the large number of medicinal or therapeutic agents that possess tertiary amine functionalities are often efficiently N-oxygenated to polar metabolites (2, 6). Frequently, formation of polar, oxygenated amine metabolites constitutes an efficient means to terminate the pharmacological action of the parent chemical and provide a route for rapid excretion. Compounds containing sulfur, phosphorus, selenium, or other heteroatoms may also likewise be more efficiently excreted after oxidative metabolism. Many exceptions are known, but generally, oxidative metabolism of heteroatom-containing compounds by cytochrome

* Tel:

(206) 284-8846; Fax: (206) 284-0313.

P-450-dependent processes leads to products with increased potential for toxic or carcinogenic properties. On the other hand, the flavin-containing monooxygenase [(EC 1.14.13.8)dimethylaniline monooxygenase (N-oxide forming)] generally converts lipophilic heteroatomcontaining compounds to polar, readily excreted oxygenated metabolites that possess decreased toxic potential. The amount of oxygenated metabolite will, of course, be dependent upon further metabolic processes, both oxidative and reductive, and notable exceptions to the general statements made above have been observed. In addition, large species differences in the rate of microsomal heteroatom oxygenation (and other oxidative and reductive processes) have been observed which undoubtedly also contribute to determining the way that chemicals and drugs are metabolized to detoxicated metabolites. A detailed description of the structure, catalytic mechanism and function, and role in chemical metabolism and toxicology will be presented for the mammalian flavincontaining monooxygenases.

Pig Liver Flavin-ContainingMonooxygenase In the late 19709, a hepatic flavoprotein was first isolated from pig liver that was called an N-oxidase, or the mixed-function amine oxidase, or simply “Ziegler’s enzyme” after Professor Daniel Ziegler of the University of Texas a t Austin who has done much t o place this oxidase in the proper place with other more exhaustively

0893-228x/95/2708-0165$09.00/0 0 1995 American Chemical Society

166 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

studied monooxygenases. It was soon recognized that the name “amine oxidase” was too narrow for the enzyme because the flavin-containing monooxygenase (FMOY oxygenated substrates as diverse as hydrazines (71, phosphines (81, iodide (91,boron-containing compounds (91, sulfides (1, IO),and selenides (111,as well as many tertiary and secondary amines (1,4 , 6 ) . Evidence for a heteroatom N- or S-oxygenase was first unambiguously shown when the FMO from pig liver was isolated and purified to homogeneity (12). A microsomal FMO (now known as FMO1) devoid of cytochromes or metals was shown to catalyze the NADPH-dependent oxygenation of prototypic substrates such as dimethylaniline or methimazole (1). A property that distinguished FMO from other monooxygenases was the unusual thermal lability of the enzyme isolated from pig liver. Thus, rapid chilling of liver is essential to avoid complete inactivation of FMO and loss of enzyme activity. Because hepatic tissue temperature can rise quickly after death, it is essential to immediately cool the tissue upon exsanguination. In the absence of NADPH, approximately 85% of the activity of pig liver FMO is lost upon standing at 40-45 “C for 3-5 min, conditions readily achieved under post-mortem situations (11. Pig liver FMO has been isolated by extraction of microsomes with detergents, selective fractionation, and purification with column chromatography on ion-exchange resins (12).This procedure can be utilized to provide homogeneous pig FMOl free from other monooxygenases or other FMO forms. As described before (11, purified pig FMOl contains 15.1-15.3 nmol of flavin adenosine dinucleotide (FAD)/mg of protein and is free from heme and other metals although highly purified FMOl contains variable amounts of lipid (13). Lipid alkyl hydroperoxides probably have a negative effect on enzyme action and stability, and addition of antioxidants (i.e., butylated hydroxytoluene) tends to help with FMO enzyme purifications. The visible spectrum of fully oxidized FMOl is similar to that of other flavoproteins (i.e., Amax of 445 and 380 nm and shoulder at 480 nm). In the absence of molecular oxygen, NADPH reduces the FAD prosthetic group to reduced flavin adenosine dinucleotide (FADHz),and the spectrum obtained is similar to the fully oxidized flavoprotein but is shifted to shorter wavelength (i.e., ,A of 440 and 370 nm, with no apparent shoulder) (14-16). Addition of molecular oxygen to the fully reduced flavoprotein generates a spectrum similar to that of the flavin hydroperoxide observed with other flavoproteins (i.e., bacterial luciferase) or synthetic isoalloxazine hydroperoxides (17 ) . The hydroperoxyflavin spectrum of FMOl is stable for 1-2 h a t 2 “C, but within minutes the spectrum decomposes to fully oxidized flavoprotein above 25 “C (1). As described below, the spectral data are consistent with an enzyme mechanism whereby the key flavoprotein species at steady state is a preformed peroxyflavin intermediate. The prediction from spectral observations of FMOl is that substrates that are readily oxygenated by alkyl hydroperoxides, peracids, or other similar agents should be readily oxygenated by FMOl (18-22). Barring steric constraints, this view is largely supported by the experimental data. The pig FMOl hydroperoxyflavin species is remarkably stable (23,241. ‘Abbreviations: FMO, flavin-containing monooxygenase; FAD, flavin adenosine dinucleotide; FADH2, reduced flavin adenosine dinucleotide; RFMO, rat flavin-containing monooxygenase; MPTP, 1-methyl4-phenyl-1,2,3,6-tetrahydropyridine; MPP+, N-methyl-4-phenylpyridinium ion; MPDP-, l-methyl-4-phenyl-2,3-dihydropyridinium ion.

Cashman

Thus, nucleophilic tertiary amines, secondary amines, secondary hydroxylamines, sulfides, thiones, and phosphines are converted to their tertiary amine N-oxides, hydroxylamines, nitrones, and S- and P-oxides, respectively. This is one of the fundamental differences between FMOs and other monooxygenases: on the basis of the chemical product produced by treating a substrate with a peracid or a peroxide, the structure of an FMO metabolite can be predicted with great certainty in most cases. However, exceptions to the general rule cited above have been observed primarily due to rearrangement or elimination reactions of unstable N- or S-oxides (i.e., see ref 25 for examples). Another fundamental difference between FMO and cytochromes P-450, for example, stems from the apparent two versus one electron mechanism of oxygenation for each monooxygenase, respectively (26). As discussed below, the two electron oxidation products produced by FMO do not inactivate FMO but in the presence of cytochromes P-450 sometimes can and do covalently modify cytochromes P-450 and other proteins without affecting FMO activity. For example, sulfenic acids (27-32) and sulfinic acids (30, 33-41) are known reactive metabolites generated by FMO that inactivate cytochromes P-450 or covalently label other proteins. Thus, care must be used in interpreting data concerning cytochrome P-450 inhibition especially by chemicals that are excellent substrates for FMO (i.e., thiols, thioureas, or thioamides, to name a few substrates).

Other Flavin-ContainingMonooxygenases In 1984, two research groups reported evidence for a form of FMO present in lung tissue that possessed many properties that distinguished this “pulmonary” form of FMO from the “hepatic”form of FMO as they were known a t the time (42-45). Today, we recognize that these two forms of FMO actually constitute two members of a family of FMOs, and the previously recognized “hepatic” and “pulmonary” forms of FMO (i.e., FMOl and FM02, respectively) actually are present in numerous tissues aside from their namesake. While both forms of FMO utilize molecular oxygen and the same cofactors (i.e., NADPH and FAD) and are essentially free of metals, the immunological behavior of each FMO form is distinct: antibodies raised to FMOl failed to cross-react with FM02 (42,431. Evidence for immunological differences of FMO forms from the same tissue (i.e., liver) have also been obtained among different species (i.e., anti-mouse liver FMO antibodies do not recognize pig liver FMO1) (431,and antibodies raised to pig liver FMOl do not crossreact with purified mouse liver FMO (46). Western blot analysis of microsomal fractions from mice, rats, rabbits, dogs, and humans shows that clear species differences in hepatic FMOl content are apparent (47,481. It should be pointed out, however, that the anti-pig liver FMOl serum used by Dannan and Guengerich (48)was apparently a unique serum possessing unusual antigenrecognition properties because immunoblot studies with subsequently procured anti-pig liver FMOl sera have not been able to provide such pronounced results. The substrate specificity (42, 43) and substrate stereoselectivity (49-51) for pig FMOl and rabbit FM02 are quite distinct although some classes of substrates are efficiently oxygenated by both monooxygenases. For example, FMOl and FM02 both efficiently N-oxygenate dimethylaniline and S-oxygenate thiobenzamide and

Invited Review thiourea (4, 52). Rabbit FM02 N-oxygenates long, aliphatic primary amines (52). FMOl does not Noxygenate primary amines, but in some cases utilizes aliphatic primary amines as positive effectors of FMO action. On the -other hand, short side chain tertiary amines (i.e., chlorpromazine, imipramine and other 10[(NJV-dimethy1amino)alkyllphenothiazinederivatives)that were readily oxygenated by FMOl were not readily oxygenated by FM02. The conclusion was that the substrate binding channel for FMOl and FM02 was quite distinct, and in this series of substituted phenothiazine substrates, the pulmonary FM02 required a nucleophilic nitrogen center possessing a t least a C6 or C7 alkyl side chain (52). It was proposed that pig FMOl admits larger substrates to within 3 A of the enzyme hydroperoxyflavin, with the binding site channel spanning as much as 12 A in diameter. On the other hand, it was hypothesized that the binding site channel of FM02 from rabbit lung rested 6-8 A below the surface of the substrate binding channel with no more than an 8 A diameter along the longest channel axis (52). The tentative structure for pulmonary FM02 requires a narrow, long substrate binding channel, but pig FMOl probably possesses a much broader and shallower substrate binding channel. FMOl and FM02 stereoselectivity studies of various S- and N-containing substrates have largely confirmed the above suggestions for FMO binding channel dimensions. Thus, sulfur-containing substrates such as 4-bromophenyl-1,3-oxathiolane(53-551, aryl-1,3-dithiolanes (49, 561, and 2-methyl-1,3-benzodithiole(50)showed an almost exclusive preference for rabbit FM02-mediated S-oxygenation of one sulfur atom lone pair. This result is in contrast to pig FMOl that showed considerable stereoselectivity (571, but in general, not the absolute stereoselectivity observed for rabbit FM02. These observations concerning rabbit FM02 were the first examples of total stereoselectivity observed during an enzyme-catalyzed sulfoxidation of a geminal pair of sulfur atoms on a prochiral carbon atom (50). Stereoselective N-1'-oxygenation of (&nicotine by FMOl and FM02 also illustrates the point that binding channels of both monooxygenases are distinct: FM02 forms exclusively trans-(&nicotine N-1'-oxide (581, and FMOl forms approximately a 1:l mixture of cis-:trans-(SI-nicotine N-1'oxide (59-61). FMOl and FM02 are also quite distinct with regard to thermal lability (1,451. In the absence of NADPH, rabbit FM02 is stable at a temperature of 45 "C for 10 min, conditions that abrogate pig FMOl activity after 5 min of treatment. Unlike FMO1, rabbit FM02 activity is quite resistant to inactivation by anionic detergents. This has direct consequences for enzyme isolation, and in fact, FMOl and FM02 are isolated by quite distinct chromatographic methods ( 1 2 , 4 3 , 4 5 ) . In addition, low concentrations of bile salts such as cholate have been shown to stimulate FMO activity in mouse and rat liver preparations (62-65), but decrease FMO activity in hepatic preparations from rabbit (66)and pig liver (45). Mercury is another agent that stimulates rabbit FM02 (67, 68) but decreases rabbit FMOl activity (65, 68). Rabbit and mouse FM02 has optimal enzyme activity at approximately pH 10 (43, 45, 691, while pig FMOl shows a pH optimum around pH 8.3-8.5 (1). Other hepatic FMOs may have slightly higher pH optima (i.e., mouse liver FMO activity has a maximum at pH 8.89.2) (46, 69). However, it is not clear whether the pH optima measured for FMO in various preparations take

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 167 into account the pH dependence for ionization of many substrates utilized for the monooxygenase. For example, the pK, of many commonly used tertiary amine FMO substrates is quite high (i.e., pH 8.5-101, and it is possible that apparent FMO activity could increase due to an increase in deprotonation of the amine as the pH is increased. The pH optima data could also indicate that the structures of different forms of FMO have varying sensitivity to alkalinity. However, this is apparently not the case for pig FMO1, because the temperaturefluorescence curves were the same at pH 7.0 or 8.3 (70). The limited amount of evidence available suggests that irreversible structural change of pig FMOl and loss of activity as a consequence of heating are not limited to the pH optima for enzyme activity (70).

Modulation of FMO Modulation of FMO by endogenous, hormonal, or dietary influences is an understudied area of investigation. Gender-related differences for mouse liver FMO appear to be due to testosterone repression of the hepatic enzyme (71, 72). Rat liver FMO levels are apparently positively regulated by testosterone and repressed by estradiol (73). Estradiol appears to play a role in determining FMO activity in rat lung and kidney preparations that was distinct from pulmonary FMO preparations (74, 75). FM02 mRNA and FM02 protein expression peaks in the lung of pregnant rabbits at days 15 and 28-31, which correlates with progesterone and corticosterone plasma concentrations (27). Interestingly, sheep FMO may be induced in the liver and repressed in the lung of pregnant animals. Apparently, sheep have distinct forms of hepatic and pulmonary FMO (76). Dietary xenobiotics influence rat liver FMO. In the absence of dietary xenobiotics (i.e., rats maintained on total parenteral nutrition for 7 days), a 75-80% decrease in FMO activity was observed (77). Similar dietary treatment of rats has provided evidence that the pharmacokinetic parameters for ethyl methyl sulfide Soxygenation (78)and trimethylamine N-oxygenation (791, two putative selective functional markers for FMO activity, were significantly decreased after 1 week on a synthetic diet. In summary, evidence for induction of FM02 in rabbit lung as a consequence of plasma hormone levels has been obtained although the physiological role for the increased FM02 has not been ascertained. On the other hand, hepatic FMO appears to be constitutively at a high level in the animal administered a normal diet, but when the animal is switched onto a synthetic diet, de-induction of FMO is apparently observed. Undoubtedly, other biochemical and physiological events are also a t work in the animals administered synthetic diets. Ziegler and co-workers have proposed that FMO is induced by one or more organic nitrogenor sulfur-containing xenobiotic(s)present in food derived from plants, and he has suggested that FMO activity is already maximally induced in animals on commercial rat chow ( 2 ) . It is possible that mammalian FMO evolved to detoxicate organic nitrogen- and sulfur-containing nucleophilic xenobiotics from foodstuffs. It should be pointed out, however, that FMO apparently precludes biologically important nucleophiles including many endogenous thiols and other heteroatom-containing compounds from the active site (10). This point will be discussed further, below.

168 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

The Family of Flavin-Containing Monooxygenases Currently, evidence for five forms of mammalian FMO exist that have deduced amino acid sequences ranging between 52% and 57% identical to that found in rabbit and between 50%and 58% identical across species lines ( 2 7 , 8 0 ) . Thus far, eleven full length sequences of FMO have been reported in the literature based on the cDNAs and three purified proteins that have been sequenced by gas phase peptide sequencing (80). In addition to the five distinct FMOs (Le,, FMO1, FM02, FM03, FM04, and FM05) several other published sequences represent orthologs from other species as well as allelic variants (27). The amino acid composition of various FMOs have been reported (2, 44). Comparative studies are useful to correlate flavoproteins as a class, but it is dangerous to make any structural conclusions based on amino acid composition. For example, the bacterial equivalent of FMO, cyclohexane monooxygenase from Acinetobacter, bears a reasonably close relationship in amino acid composition to rabbit FMO1, but the amino acid sequence is only approximately 25% identical (811. While bacterial cyclohexanone monooxygenase is apparently mechanistically similar to mammalian monooxygenases (82) and may even give rise to the same enzymatic products, it is nevertheless not classified as an FMO because it is only distantly related in amino acid sequence. Based directly on the amino acid sequence or the amino acid sequence deduced from the cDNA data, FMO forms possessing greater than approximately 40% identity from data currently available can be assigned to one of the five gene subfamilies. Mammalian orthologs of FMO are assigned with amino acid identities greater than about 80%. Following the guidelines established for cytochromes P-450 (83),a systematic nomenclature for mammalian FMO based on divergent evolutionary relationships was proposed (80).. Thus, FMOl corresponds to the pig liver FMO enzyme and its orthologs in rabbit (i.e., FMO 1Al o r form l), human (i.e., FMOl), and rat (RFMO1). Rabbit FM02 is the same as rabbit lung FMO or rabbit FMO 1B1. FM03 was previously named FMO form 2 (or FMO 1D1) from rabbit liver, and the human ortholog was named HLFMO 11. FM04 was previously given the name human FM02 as well as FMO 1 E l which is the ortholog in rabbit. Finally, FM05 replaces the name FMO 1C1 (or also designated FMO form 3) from rabbit (80). In addition to cDNA oligonucleotide sequencing, automated Edman degradation sequence and mass spectral sequence analysis has provided substantial sequence information for FMOs. Ozols has made significant contribution t o the isolation, purification, and Edman sequencing of various rabbit liver FMOs (84-86). In addition to the extremely lipophilic and intractable nature of the highly purified FMO proteins, the presence of an N-terminal blockade (Le., an N-terminal acetyl group) of FMO1, FM02, and possibly FM03 has significantly confounded progress in the sequencing of FMOs by traditional Edman sequencing methods. Chemical and enzymatic methods are available for deacetylating N-terminal acetyl groups of FMO, but the most straightforward way to identify N-terminal amino acid sequences even in the presence of modified amino acid sequences is by direct determination of an N-terminal peptide by mass spectrometry. Highly purified pig FMOl was

Cashman reduced, carboxymethylated, and digested with either endoproteinase Glu-C or trypsin (87). A total of over 80 HPLC fractions were collected, and the molecular weights were determined by liquid secondary ion mass spectrometry. Approximately half of the HPLC fractions required derivatization with hexanol to give useful peptide molecular ions (i.e., lipophilic hexyl esters of polar peptides). Because the primary sequence of pig FMOl was known from the cDNA data (88), we focused on searching for important peptide digest fragments (i.e., N-terminal posttranslational modification, N-glycosylation tripeptide consensus sequences, NADP+- and FAD-binding domains, etc.). Tandem mass spectrometry confirmed the amino acid sequence of pig FMO1. The use of mass spectrometry/mass spectrometry techniques for FMO sequence determination demonstrates many advantages of tandem mass spectrometry. The mass spectrometry/ mass spectrometry technique does not require complete separation of peptides and is capable of obtaining amino acid sequence information from modified peptides. While our earlier work was quite labor intensive, recent advances in liquid chromatography-mass spectrometry have markedly accelerated the procedure and have allowed rapid attainment of amino acid sequence data. To date, over 90% of pig FMOl and rabbit FM02 has been sequenced by liquid secondary ion mass spectrometry (87, 89). Both enzymes are N-terminal blocked by an N-acetyl moiety. The data provide no evidence that highly purified FMOl from pig liver contains any other flavoprotein aside from FMO1. No evidence for multiple forms or allelic variants is apparent in the highly purified preparations examined from pig liver although it is possible that a minor amount of a variant escaped detection by sequencing by mass spectrometry. For rabbit FM02 isolated from lung, tandem mass spectrometry and gas phase sequencing studies provided direct evidence for the existence of a 1:l complex of a calcium-binding protein (i.e., calreticulin) to rabbit FM02. The complexation of calreticulin to FM02 from rabbit lung could account for some of the unusual physical properties of purified rabbit FM02. For example, in contrast to the highly lipophilic FMOl from pig liver, FM02 from rabbit lung that is complexed to calreticulin is quite water soluble. Separation of FM02 from calreticulin converts the FMO-calreticulin complex into a lipophilic protein with physical-chemical properties similar to those of FMOl from pig liver. I t is possible that binding of calreticulin to FM02 stabilizes the enzyme to thermal denaturation and obviates the requirements for high concentrations of detergents normally required for hepatic FMOl activity. However, experiments performed combining calreticulin with pig FMOl did not markedly change the physiochemical properties (i.e., thermal stability) of FMO1. This observation again points to the distinct nature of these two FMO forms. It is unlikely that the complexation of FM02 with calreticulin alters the kinetic or substrate properties (i.e., types of products or stereoselectivity). However, it is quite possible that different laboratories investigating FM02 from rabbit lung are using preparations containing substantial quantities of calreticulin complexed to FMO. The presence of equimolar amounts of calreticulin present in various FMO preparations (89, 90) may account for the discrepancy between reported amino acid composition of rabbit FM02 and the amino acid composition deduced from the cDNA data. It is not known what the physiological role of binding of calreti-

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 169

Invited Review

culin to rabbit FM02 is (if any), but the C-terminal sequence of calreticulin is present in several resident luminal endoplasmic reticulum proteins and may be required in the cellular mechanism to recognize resident luminal proteins and distinguish them from other proteins that are not targeted for retention in the endoplasmic reticulum (91). Rabbit lung calreticulin may act as a chaperon protein to help target or translocate rabbit FM02 to the endoplasmic reticulum and modulate enzyme activity. From a structural point of view, calreticulin is markedly zonal in character (i.e., the enzyme resembles the structure of a dumbbell) with an approximately neutral net charge on the N-terminus forming a globular domain attached t o a tail-like proline-rich internal zone and an acidic C-terminus. It is possible that rabbit FM02 binds to calreticulin in a complementary fashion (92).

120k.n-X-SerlThr

FAD I

NADP 2 2 5 ~ s ~

513Leu

Gene Regulation Analyses of genomic DNA by Southern blot hybridization show that each of the five FMO gene subfamilies contains a single gene (27). The existence of additional FMO genes encoding proteins closely related to the known FMOs is likely because orthologs and variants have been previously observed (93). A number of studies have shown that FMO expression is both tissue- and species-dependent (2). Thus, in a given tissue it is probably the particular profile of FMOs present that determines enzyme activity, with the dominant form responsible for substrate specificty and stereoselectivity. For example, in the human, FMOl appears to be present in the fetal liver (94) and is not expressed in the adult human liver (95). Evidence for FMOl activity in the human kidney has been reported (96,971, and it is likely that FMOl comprises at least one of the major FMO forms in human kidney. FMOl is also expressed to a significant extent in rabbit intestinal and nasal mucosa (27). The form of FMO present in rabbit lung (i.e., FM02) is also expressed in the esophagus and nasal mucosa. Rabbit “lung“FM02 is present in rabbit kidney and bladder but absent from rabbit liver. Rat brain also contains a form of FMO that is immunochemically similar to FM02 from rabbit lung (98). Like other monooxygenases, it is likely that the gene encoding FMOs will be of considerable length. However, to date, little information along these lines is available. Important FMO gene regulation questions still remain to be answered. Because FMO does not respond to the types of inducing agents associated with other monooxygenases, it is likely that FMO is not coordinately regulated with these other enzyme systems. Additional factors are probably involved. As described above, hormones and dietary factors appear to regulate FMO expression, but this is species- and tissue-specific and likely to occur independently of other factors. Because of the pronounced differences in tissue-specific levels of mRNA, protein levels, and FMO activity in rabbit,2 it is possible that FMO gene expression is regulated at the transcriptional stage. Regulatory regions on the 5’flanking region must be identified and transcription factors characterized. Several regulatory elements that control FMO expression may be at work. Post-translational modification of FMO (i.e., N-glycosylation) does not appear to play a role in modulating FMO1, FM02, and 2D. E. Williams and R. N. Hines, personal communication.

Primary Structure of FMO The primary structure of FMOl from pig liver is schematically illustrated in Figure 1. Figure 1is a useful representation to discuss the enzyme isoforms from a variety of tissues and sources. The screening of cDNA libraries with other cDNAs or synthetic oligonucleotides encoding FMO has provided cDNA inserts appromiately 2.2-2.6 kb in length or smaller that encoded various FMOs. Currently, the complete nucleotide sequences for 11 full length cDNAs have been reported in the literature or through GenBank (27,80). Generally, the cDNAs encode for an enzyme of approximately 533-535 amino acids, but examples of FMOs with 19 (99) or 25 (100) additional C-terminal amino acids have been observed. It is possible that a point mutation of an ancestral gene that altered the “normal” stop codon provided the extra amino acid residues at the C-terminus. Comparison of the amino acids of the 5 forms of FMOs by hydropathy profiles shows a striking similarity even in regions of the FMO isoform sequence that are only modestly identical (i.e., 25-30% amino acid identity). Several regions of all five FMOs have a relatively high percentage of residues that are identical. For example, residues 1-200 and between 450 and the C-terminus of FMO contain relatively highly conserved regions indicating important amino acid structural and/or functional domains. In addition, the highly conserved FAD- and NADP’-binding domains (i.e., GXGXXG) near deduced amino acid positions 9-14 and between 186 and 196, respectively, apparently are critical for FMO function. Site-directed mutagenesis of the region 9-14 to GXGXXV resulted in expression of an inactive protein to which no FAD was bound (1011. It is notable that rather conservative amino acid changes on the N-terminal side of the FAD-binding domain of FMO result in an FMO that retains activity. Also, addition of a 35 amino acid portion of a P-galactose fusion protein to the N-terminus significantly enhanced the level of expression in Escherichia coli. The N-terminus of FMOs does not have any clearly discernible signal peptide sequence and, coupled with the presence of an FADbinding region, does not suggest this region as a functional membrane insertion domain. Most of the FMO cDNAs sequenced to date contain the highly favorable translation initiation site (i.e,, XXATGG, where X is generally C) (202). The cDNA sequence does not provide

170 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

information about possible N-terminal modifications, and in each case examined carefully (i.e., FMO1, FM02, and FM03) the initiation amino acid methionine is not present and the following amino acid is N-acetylated. Residues such as alanine or glycine near the N-terminus apparently promote the removal of methionine during FMO protein maturation (103). With the exception of FM04, all FMOs sequenced to date have at least a single putative consensus N-glycosylation site (i.e., Asn-XaaSermhr). Korsmeyer et al. (104) have shown that pig FMOl is glycosylated, and subsequent HPLC-mass spectrometry studies showed that a saccharide was attached to Asn 120.3 As discussed below, N-glycosylation does not detectably alter the functional characteristics of FMO because cDNA-expressed FMOl and FM03 in E. coli give essentially identical substrate stereoselectivity as that observed with the microsomal enzyme (54, 105, 106). The highly conserved nature of the putative consensus N-glycosylation sites suggests that the region may still play some important structural and/ or functional role in enzyme action (95). All mammalian FMOs possess very strong membrane association properties and in the highly purified state are extremely intractable proteins with poor solubility characteristics (89). While all FMOs possess a number of hydrophobic regions, only the C-terminus (i.e., residues 318-533) is sufficiently hydrophobic to be a membrane-insertion sequence. However, truncation of the C-terminal 26 amino acids from rabbit FM02, for example, did not lead to loss of membrane association properties when the remaining FMO cDNA was expressed in E. coli (101). Even with as many as 200 C-terminal amino acids deleted from rabbit FM02, membrane association was still observed (101).Clearly, FMO membrane association is not a passive event dictated exclusivelyby hydrophobic C-terminal amino acid residues. Rather, the information for active FMO membrane association is probably encoded in an internal sequence proximal to the N-terminus (i.e., residues 15-230) that signals membrane association. Coding region changes resulting in single amino acid substitutions cause dramatic differences in the physical-chemical properties of FMO (i.e., migration on SDS-PAGE). This is observed for variant microsomal enzymes as well as variant proteins expressed in yeast (93). It is possible that a few or even a single amino acid mutant of FMO may possess markedly different physical-chemical or even catalytic properties. A partial list of variant sites in rabbit liver FMO isoforms has been previously published (86). Other regions of FMO possess notable homology to other well-characterized enzymes (i.e., esterases) (86). That is, all microsomal FMOs from rabbit liver contain a consensus sequence common to serine proteases and esterases. The amino acid sequence of one form has the consensus sequence containing the histidine residue of the catalytic triad of the diisopropyl fluorophosphatesensitive carboxylesterases and thioesterases. It is possible that FMOs may have evolved from esterases or, for some unknown reason, evolved from, or utilized structural or catalytic domains from other proteases. For example, Ozols has noted that the highly conserved Ser 194 of FMOs is among a peptide sequence in FMOs containing the active serine residue found in a number of esterases (i.e., V8 protease, acetylcholine esterase, etc.) 3K. Korsmeyer, A. Falick, D. Ziegler, and J. Cashman, unpublished results.

Cashman (86). Further, the highly conserved Asp 225 of FMOs is

contained in an amino acid sequence quite similar to one present in thioesterases and rabbit liver microsomal esterases (86). Finally, Kaderlik has noted that the C-terminus of pig FMOl (present in human FM03 as well) contains an amino acid sequence that is highly homologous to a region designated as the “leucine box” for the T cell antigen receptor (107). Currently, it is unknown what the physiological or functional role of the “transplanted domains” of other proteins is or how FMO utilizes such peptide regions for catalytic or structural purposes.

Heterologous Expression of FMOs Structural and functional studies of the FMOs have been greatly aided by heterologous expression of the FMO cDNAs. Because multiple forms of FMO can exist in a specific tissue that have distinct physical and substrate specificity properties, it is quite useful to be able to investigate one FMO without the possible complications of multiple forms of FMO or other monooxygenases being present. Other confounding properties of FMOs including the possibility of post-translational modification and the requirement of detergents for enzyme activity have prompted numerous approaches to cDNA expression of FMO. Pig FMOl cDNA was the first FMO cDNA to be expressed in E. coli (54). The highly purified pig FMOl from microsomes and the highly purified enzyme expressed in E. coli possessed similar kinetic constants for various typical FMO substrates including the tertiary amine N-oxygenation of chlorpromazine and trifluoperazine (54). However, the stereoselectivity of sulfide Soxygenation for highly purified and cDNA-expressed pig FMOl was not identical in all cases examined and led to the suggestion that N-glycosylation of FMO in the presence of pig liver microsomes led to a functionally distinct enzyme compared with the cDNA-expressed enzyme in E. coli. Further examination of this point by comparing adult human liver microsomes and cDNAexpressed human FM03 substrate specificity (105,106) and N- and S-oxygenation stereoselectivity (59,106,108, 109) did not suggest such a marked effect. The N- and S-oxygenation stereoselectivity observed for human liver microsomes or cDNA-expressed human FM03 was virtually identical. Either human FM03 does not require N-glycosylation for full functional activity and pig FMOl does or (what is more likely) detergents present can contribute in a substrate-dependent fashion to the activity of pig FMOl(13). In conclusion, N-glycosylationmay play a stuctural or functional role for pig FMO1, but the majority of studies suggest that post-translational modification is not essential for a complete complement of FMO activity. To date, the evidence suggests that FMO cDNA expression in E. coli is a perfectly adequate system to examine the properties of this class of monooxygenase. The cDNAs of rabbit FM02, FM03, and FM05 likewise have been expressed in active form in E. coli, and the activity toward a number of traditional substrates was examined (101). cDNA-expressed FM02 was similar to highly purified FM02 for S-oxygenation of methimazole. Surprisingly, a number of typical FMO substrates, including methimazole, chlorpromazine, prochlorperazine, imipramine, Nfl-dimethylaniline, cysteamine, and trimethylamine, were not substrates for cDNA-expressed rabbit FM05, although n-nonylamine and n-octylamine were N-oxygenated (99). To date, cDNA expression of

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 171

Invited Review Table 1. Putative Tissue Levels of FMO Forms Present in Animals and Humansa liver mouse rat rabbit human kidney mouse rat rabbit human lung mouse rat rabbit human

FMOl

FM02

FM03

FM04

FM05

low high high v.low

NPb NP low

high low low high

? ? ? v. low

low low low low

high high low high

? ? low low

high high v. low ?

? high high ?

low low low ?

? ? ? ?

high ? v. high low

v. low ? ? ?

NP NP NP NP

low low NP ?

NADPH

f4 *p2

H20 + NADP’ ?b

a Estimates from refs 2,27,47, 48,94,95,97,99,100,101, and 128. NP,apparently not present. A question mark indicates that no data are available or the presence of a n FMO form is in doubt.

FM04 in active form has not been reported. More recently, rabbit FM02 cDNA was expressed in Saccharomyces cerevisiae and compared with the enzyme expressed in E. coli (101). In addition, rat liver FMO cDNA has been expressed in yeast (110). Finally, rabbit FMOl and FM02 cDNAs were expressed in COS-1 cells, and the physical and catalytic properties were compared with the properties of the native microsomal proteins (111). In agreement with other studies, the data obtained supported the observation that more than one form of FMO is present in rabbit liver and lung and that the presence of more than one enzyme could contribute to the biphasic substrate oxygenation kinetics sometimes oberved with various tissue preparations (69,112,113). Table 1shows a compilation of the “best guess” as to the tissue distribution of FMO forms in experimental male animals and humans. It should be recognized that this guesstimate is largely based on FMO mRNA data, and it should be pointed out that mRNA levels may not always correspond to enzyme activity levels. In addition, as more information becomes available on gender- and tissue-specific FMO expression, Table 1will need to be corrected.

Catalytic Mechanisms of FMO Each step of the catalytic cycle of FMOl action is known in some detail and has been reported previously by the laboratories of Ziegler (4, 14)and Ballou (9, 15, 16). It is likely that the enzymes of the other FMO subfamilies also follow a similar mechanism although this point has not been studied in detail. However, as discussed above, the substrate binding channel of different FMO forms are clearly distinct (52,105,106). Thus, selective functional probes of isoforms of FMO may have to be developed on the basis of steric or stereochemical features because it is unlikely that novel classes of substrates will be completely selective for each form. The reason for this is because of the fundamental similarity of FMO enzyme mechanism among members of the FMO family. Significant stereoselectivity differences have been observed for FMO forms, and stereoselective substrate oxygenation may provide the basis for developing form-selective FMO substrates (106). The major steps in the FMOl catalytic cycle are shown in Figure 2. In the first part of the enzyme reaction (step 11, for a substrate such as dimethylaniline, the fully

Enz-FIH2 + NADP*

Enz-FIHOH

?”; S-O

S

-FImH

Figure 2. Schematic representation of the catalytic cycle of pig FMO1. S a n d S-0 a r e t h e substrate a n d oxygenated substrate, respectively.

oxidized flavoprotein (i.e., Enz(Fl,,)) reacts with NADPH in a fast step (Le., 53 M-l) to give the enzyme in the reduced form (i.e., Enz(F1Hz)). The NADP+ produced apparently remains at the active site of the enzyme. The reaction of the reduced enzyme with molecular oxygen (step 2) is also quite rapid (i.e., 45 M-l) as anticipated from model studies and generates the oxidant used in the enzyme reaction (i.e., the 4a-hydroperoxyflavin of FAD, Enz(F100H)). The formation of the relatively stable hydroperoxyflavin intermediate is notable for at least two reasons: (a) it is unusually resistant to decomposition and is remarkably long-lived which suggests that non-nucleophilic FMO active site amino acids are present to provide an appropriate lipophilic environment to preserve this highly reactive species, and (b) in contrast to cytochromes P-450 that only form oxidizing agents after substrate binds, the preloaded FMO active site oxidant waits in a ready position to oxygenate substrate. Oxygenation of substrate proceeds rapidly (step 3) in a bimolecular reaction (i.e., 4700 M-’ s-l) with attack on the terminal flavin peroxide oxygen to produce the oxygenated product (i.e., S-0) and the 4a-hydroxyflavin form of FAD (i.e., the pseudobase Enz(F1HOH)). The next step is apparently largely rate-limiting (i ,1.9 M-l) and must involve either dehydration of the pseudobase or release of NADP+ (step 4). Because NADP+ is a competitive inhibitor of the pig FMOl cofactor NADPH, data of kinetic studies suggest that NADP+ comes away from the flavoprotein last. It is clear that the slow step in the overall catalytic cycle (i.e., for a good substrate) is not release of oxygenated product. This has important consequences for kinetics of good substrates: since either rate-determining step occurs after product release the mechanism of Figure 2 predicts that all good substrates will possess similar and large V,,, values. Solvent deuterium isotope effects on the kinetics of dimethylaniline N-oxygenation with pig FMOl tend to suggest participation of general acid catalysis (1141, which would tend to support dehydration as the rate-limiting step, but this point is controversial. NADP+ appears to play a “gatekeeper” role in that Enz(F1H2) that reacts with molecular oxygen in the absence of NADP+ produces significant amounts of HzOz that is otherwise not normally formed (15,161.FMO is generally tightly coupled, and only minute amounts of HzOz “ l e a k away from the monooxygenase under normal conditions. If this was not the case, FMO would serve as an NADPH oxidase that would produce copious amounts of HzOz in the absence of substrate and expose the cell to the untoward effects of oxidative stress.

Cashman

172 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

Table 2. Kinetic Constants and Free Energy Values for S-Oxygenation of S-AlkylcysteineDerivatives by Pig Liver Flavin-ContainingMonooxygenase ~

Km

substrate FCM FCM (B)a cis-CPCb trans-CPCb SBC“

(JAM)

22.7 33.9 1111 344 769

Vm,, [nmol min-l (mg of protein)-l]

1319 1001 555 263 384

~~

V”/Km 55.1 29.5 0.5 0.76 0.49

AG 1.92 2.17 4.32 3.60 4.09

Two diastereomeric S-oxides were formed with slightly different kinetic constants from farnesylcysteine methyl ester (FCM). CPC is (3-chloro-2-propenyl)-~-cysteine. SBC is S-benzyl-Lcysteine.

Figure 3. Proposed structure of t h e binding channel and active site of pig FMO1. Compartments A, B, and C a r e hypothetical sites adjacent to t h e flavin cofactor below t h e surface of t h e enzyme. Site A is a site that is not normally accessible to substrates but can be a secondary interaction site for cetain substrates. The binding channel opening is proposed to access the surface of t h e enzyme as depicted in region D.

Ziegler has suggested that only a single point of attachment to the terminal hydroperoxyflavin oxygen is required for substrate oxygenation and that more complex interactions involving precise fit normally required to lower the energy of activation for other monooxygenases are not required (2). The kinetics and proposed mechanism of FMO action are in accord with this suggestion, and such a paradigm does not violate principles of enzyme saturation (i.e., Michaelis-Menten) kinetics. However, a number of studies have shown that some FMOs can be highly stereoselective and other FMO orthologs can metabolize the same substrate with modest stereoselectivity. For example, in the presence of (S)nicotine, FM02 and FM03 exclusively form trans-(S)nicotine N-1’-oxide (58,59) whereas FMOl forms approximately equal amounts of cis- and trans-(S)-nicotine N-1’-oxide (58,60). In addition to binding channel constraints described above, additional interactions must be at work to facilitate the stereoselectivity observed. Further, addition of n-octylamine, an aliphatic primary amine that is a substrate for FM02 and a stimulant of FMOl activity, dramatically changes the stereoselectivity of FMO1-mediated S-oxygenation (56).In addition, it is known that changing the aliphatic substituent of alkyl p-tolyl sulfide to larger and larger aliphatic groups reverses the S-oxygenation stereoselectivity for FMO from adult human liver microsomes (i.e., FM03) (97).The observations cannot be explained by a simple lipophilic effect on FMO action although it is well recognized that increasing substrate lipophilicity is associated with a significant decrease inKm(115).This can clearly be seen in the case of S-oxygenation of S-alkylcysteine derivatives (116). We propose a new explanation for all of these obervations that requires additional points of contact for substrates at the active site of FMO. As shown in Figure 3, we propose a new model for the active site structure for FMOs that incorporates what is currently known about the structure of the binding channel and the active site of FMO.

For pig FMO1, a number of studies published in the literature suggest that the pro-R sulfur atom of simple dialkyl and alkyl aryl sulfides is preferentially S-oxygenated (23,50-52, 54-57, 106, 117, 118). However, a marked difference in the stereochemical outcome for FM03 and FMOl S-oxygenations points to markedly different substrate binding channels and possibly different active site topology for the different classes for FMO (106). For example, FMO1, FM02, and FM03 can differientiate between pro-R and pro-S sulfur atoms in 2-methyl-173-benzodithiole with % pro-R of 84- loo%, loo%, and 51%, respectively. The energy difference of binding and stereoselectivelyS-oxygenating benzodithiole is only approximately 1-1.2 kcal mol-l. The relatively small energy difference for formation of 2-methyl-1,3benzodithiole S-oxide diastereomer products suggests that substrate binding modes are not that different for the FMO isoforms examined (106). Other substrate binding phenomena possibly involving secondary site interactions may be utilized by different FMO isoforms to account for the stereoselectivity observed. As shown in Figure 3, we propose that a sulfide, for example, passes through the FMO substrate binding channel and approaches the hydroperoxyflavin of FMO (i.e., Enz(F100H))in a committed process, orienting the larger alkyl group (i.e., R’ moiety) in compartment C. As such, S-oxygenationby FMOl normally yields the S-oxide with (R)-S-oxide stereochemistry. In the presence of n-octylamine, the preferred site for large lipophilic substituents is filled (i.e., compartment C) and makes the effective size of R’ much larger and compels R and R’ groups to switch positions, thus inverting the “normal” stereoselectivity. Because N-oxygenation of long chain aliphatic primary amines is FMO-form specific (e.g., n-octylamine is a substrate for FM02 (119,120) and possibly FM03 (121)but not for FMOl (I)),it is possible that a preferential binding pocket for such lipophilic compounds exists. The binding domain of long aliphatic lipophilic compounds such as n-octylamine must be close to the hydroperoqdlavin moiety of FAD to afford the rateenhancing and substrate properties shown by FMOl and FM02, respectively. The presence of secondary site interactions that come into play for a chemical series as the substrate becomes longer also may explain the rateenhancing effect of more lipophilic S-alkylcysteine compounds. For example, the K , and corresponding AG values for S-oxygenation of a series of S-alkylcysteine compounds varies greatly depending on the chemical structure of the substrate (116)(Table 2). Thus, the S-farnesyl compound, farnesyl-L-cysteine is stereoselectively S-oxygenated with a 15-16-fold lower Kmthan the cis or trans isomers of (3-chloro-2-propenyl)-~-cysteine, respectively. The corresponding methyl ester of fmesyl,

Invited Review L-cysteine exacerbates the lipophilic effect of the farnesyl moiety, presumably by converting the zwitterionic nature of the amino acid substrate to a simple amine. Other lipophilic endogenous sulfur-containing compounds may also be S-oxygenated by FMO. It is clear, however, that zwitterionic nucleophilic endogenous thiaphiles such as cysteine, glutathione, or methionine are not good substrates (Le., low K, substrates) for FMO (20, 125,122, but see Addendum). Recently, an S-alkylcysteine Soxygenase has been described (223)that appears to be catalytically distinct from the S-alkylcysteine S-oxygenase activity of pig FMOl(116,224).As discussed above, if FMO catalyzed efficient S-oxygenation of even one of these polar, endogenous materials, FMO would serve as an NADPH oxidase, rapidly depleting the cell of its ability to survive. The proposed model of Figure 3 also helps to explain how the stereoselectivity of S-oxygenation of a series of p-tolyl alkyl sulfides can reverse for the same FMO monooxygenase as the alkyl substituent increases in size (51).Thus, change of the alkyl groups of p-tolyl alkyl sulfide from ethyl to tert-butyl may force the aliphatic substituent to adopt residence in the secondary site and reverse the preferred S-oxygenation stereoselectivity. For example, compartment A (Figure 3) may not be generally available to substrates (e.g., for ionic interaction reasons), but larger, lipophilic substrates may cause a "breakthrough'' to this secondary site.

Role of FMO in Human Chemical and Drug Metabolism The majority of studies have implicated cytochromes P-450 in the oxidation of chemicals and drugs in humans (26).Recently, evidence from in vitrohn vivo correlations of selective functional probes of FMO activity have shown that FMO also may be a key component of the monooxygenase system that converts nucleophilic amines, sulfides, and other heteroatom-containing drugs and chemicals into relatively polar, readily excreted metabolites in humans. A number of studies have shown that adult human liver microsomes are capable of tertiary amine N- and sulfide S-oxygenation (58,59,96,97,105,106,108,109, 225-1271. Adult human liver FMO activity is thermally labile (59,225-127), and activity is maximal above pH 8.4, although considerable intersample variation has been observed. Post-mortem loss of FMO activity is related to the length of time that the tissue remains at elevated temperature andor under anaerobic conditions (108,109, 225). For adult human liver tissue that was rapidly snap-frozen after procurement, a good relationship between FMO activity and FMO immunoreactivity was observed (208,109). Utilizing these same snap-frozen tissue preparations, we determined that adult human liver FMO activity was not dependent on the gender, age, smoking history, or previous drug administration history of the subject that the tissue was obtained from (58,108). Currently, there is evidence for three FMOs in adult human liver. FM03 appears to be the dominant form of FMO present in adult human liver microsomes, and FM02 and FM04, if present, are expressed at very low levels (128).N,N-Dimethylaniline N-oxygenase activity has been observed in the presence of fetal human liver (129)and adult human renal microsomes (97),and it is likely FMOl is largely responsible for the fetal liver and adult renal FMO activity observed. However, other

Chem. Res. Toxicol., Vol. 8, No. 2, 2995 173

C H f h ,

CHf

kCH,

A Figure 4. Chemical structure of lO-[(N,N-dimethylamino)alkyll-2-(trifluoromethyl)phenothiazine(compound A) and chlorpromazine (compound B).

forms of FMO are undoubtedly also present and active in adult human kidney. One study showed that imipramine N-oxygenation, considered a selective functional probe of FMOl activity, was quite elevated in adult kidney microsomes, but largely absent in adult human liver microsomes (96).Another study showed that stereoselective sulfide S-oxygenations in the presence of adult human kidney and fetal human liver microsomes were comparable while stereoselective sulfide S-oxygenation in the presence of adult human liver microsomes was quite distinct (97).The conclusion from the studies was that prominent, functionally distinct FMO forms were expressed in adult human liver and kidney. Currently, no form of human FMO has been isolated and purified to homogeneity from liver microsome preparations. However, an FMO from macaque liver microsomes has been purified to homogeneity (130).Based on immunoreactivity and limited N-terminus amino acid sequencing studies, the conclusion is that the major macaque liver FMO is an ortholog of FM03 from human liver. A polyclonal antibody to the macaque liver FMO selectivelyrecognizes a form present in adult human liver microsomes as well as cDNA-expressed human FM03 but does not recognize rabbit FMOl or rabbit FM02. To get around the difficulties associated with isolating and purifying FM03 from adult human liver microsomes, we cloned, sequenced, and expressed adult human liver FM03 cDNA in E. coli (105,106). As described above, the approximate dimensions of the substrate binding channel were determined for cDNA-expressed and adult human liver microsomal FM03 by examining the regioselective N-oxygenation of a series of lO-[(N,N-dimethylamino)alkyl]-2-(trifluoromethyl)phenothiazines. N-10Substituted phenothiazines with long aliphatic side chains (i.e., C5- and C6-alkyl) were better substrates for cDNA-expressed or microsomal adult human liver FM03 than tertiary amines with shorter side chains (i.e., chlorpromazine or imipramine) (Figure 4). The conclusion was that human FM03 possesses a distinct substrate binding channel and a distinct substrate specificity compared with pig FMO1. Like FM02 from rabbit lung (52), human FM03 utilizes a much narrower and longer substrate binding channel compared with pig FMO1. Thus, structural differences between the FMO isoforms are manifested in functional differences as well. As discussed above, (SI-nicotine N-1'-oxygenation can be used as a selective functional probe of FMO activity. In the presence of adult human liver microsomes supplemented with cytosolic aldehyde oxidase, (5')-nicotine is converted to (5')-cotinine and @+nicotine N-1'-oxide (59). Of the two N-1'-oxide diastereomers that could form, only trans-(9)-nicotine N-1'-oxide was observed. Other monooxygenases, including cytochromes P-450 from rat liver, mouse liver, and rabbit lung, catalyzed the formation of @)-nicotine N-l'-oxides albeit with lower diastereoselectivities (i.e., cis-:trans-nicotine N-1'-oxide ratios

174 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

with an average value of 82:18 (58)). In the presence of cDNA-expressed adult human liver FM03, as in the case of adult human liver microsomes, only trans-(S)-nicotine N-1‘-oxide was observed (58). Highly purified rabbit FM02 and FMOl from pig liver formed (S)-nicotineN-1’oxide with trans-:cis-($)-nicotineN-1’-oxide ratios of 1OO:O and 57:43, respectively (59, 60). The conclusion is that pig FMOl and the cytochrome P-450 enzymes examined thus far were considerably less stereoselective toward (5’)nicotine N-1’-oxygenation than rabbit FM02 or FM03 from adult human liver. In agreement with studies described above, rabbit FM02 appears to be functionally much more similar to adult human FM03 than to pig FMO1, based on substrate specificity and stereoselectivity studies thus far examined. To verify that human FM03 was solely responsible for forming trans-@)-nicotine N-1’-oxide in the presence of adult human liver microsomes, we administered (SInicotine to healthy male smokers by three routes of administration (Le., free smoking, intravenous infusion, and dermal patch administration) and determined the stereoselectivity of the (SI-nicotine N-1’-oxide metabolite in the urine (58). In all cases examined, only trans-(S)nicotine N-1’-oxide was observed in accord with our in vitro findings. That @)-nicotine N-1’-oxide was quite stable to further metabolic oxidation and reduction reactions stems from the observation that infusion of (5’)nicotine N-l‘-oxide-dz with a known cis:trans stereochemical ratio was recovered essentially unchanged in the urine. The conclusion is that, in adult male humans, (5’)-nicotine is N-1’-oxygenated with absolute stereoselectivity to produce soley trans-(S)-nicotineN-1’-oxide and that formation of this metabolite is a selective functional marker of adult human liver FM03. That no cis-(SInicotine N-1’-oxide was observed to be formed suggests that nonhepatic (SI-nicotineN-1’-oxygenation metabolism in humans (i.e., kidney, intestine, or elsewhere) is not occurring and that FMOl makes an insignificant contribution to (S)-nicotine N-1’-oxygenation in humans. Cimetidine is a safe, clinically useful H2-receptor antagonist that is widely used for the treatment of peptic ulcer disease and gastric hypersecretory syndromes in humans (131). Because cimetidine is a known inhibitor of cytochromes P-450 (132) and because the principal route of metabolism of cimetidine is S-oxygenation in humans (133), we reasoned that FMO was the likely monooxygenase that afforded the observed S-oxide metabolite. Cimetidine S-oxide possesses a center of chirality and as such could be utilized as a stereoselective functional probe of monooxgenase activity in vitro and in vivo. It should be pointed out that, in our original publication (134), we assigned absolute stereochemistry on the basis of circular dichroism studies, and more recent HPLC-optical rotatory dispersion studies showed that our original assignment was opposite the correct one. Below, we use the correct assignment. In the presence of highly purified FMOl and microsomes from pig liver, cimetidine S-oxygenation (Le., (--):(+I, 57:43) was essentially identical. The conclusion was that FMOl was solely responsible for cimetidine S-oxygenation in the presence of pig liver microsomes. In contrast, adult human liver microsomes formed the opposite cimetidine S-oxide enantiomer with significant stereoselectivity (i.e., (+):(-I, 84:16) (134). The major cimetidine S-oxygenase in adult human liver microsomes (Le., FM03) gives opposite product S-oxide enantioselectivity as that observed for pig FMOl (134). Thus, cimetidine S-oxygen-

Cashman

ation represents another FMO substrate case where enzyme structural differences are manifested in functional differences in enzyme stereoselectivity (134,135). To verify that cimetidine S-oxygenation was a functional marker for adult human FM03 activity, we examined the stereochemistry of cimetidine S-oxide isolated and purified from the urine of human subjects. The average cimetidine S-oxygenation stereopreference for 8 adult males was (+I:(-), 7525, which was in relatively good agreement with the enantiomeric composition of cimetidine S-oxide produced in the presence of adult human liver microsomes (134). The conclusion is that adult human liver FM03 is largely responsible for both in vitro and in vivo transformations. It is possible that the slight difference in stereoselectivity for formation of cimetidine S-oxide observed in comparing the in vitro and in vivo result stems from enantioselective oxidation or reduction of one of the S-oxide enantiomers. Another possibility is that enantioselective renal elimination obscures the stereoselectivity result. Regardless, stereoselective formation of cimetidine S-oxide (or formation of trans-(S)nicotine N-1’-oxide)may be a useful bioindicator of the functional contribution of FMOl or FM03 in a particular species. For example, after oral administration of cimetidine, species with a preponderance of FMOl would be predicted to have mainly urinary (-)-cimetidine S-oxide present. In addition, after administration of (S)-nicotine by smoking or intravenous or transdermal routes of administration, the prediction is that species with a predominance of urinary trans-(S)-nicotine N-1‘-oxide would have mainly FM02 or FM03 activity present. In summary, knowledge of the stereoselective oxygenation of cimetidine and/or (SI-nicotine has been used as a diagnostic indicator of functional FMO activity in humans and animals. It is likely that a convenient in vitro-in vivo probe of FMO activity will be useful in the evaluation of animal models of human drug and chemical metabolism for the investigation and development of new drugs.

Role of FMO in the Toxicology of Amine-Containing Chemicals and Drugs In humans, substantial evidence has been reported that FMO serves an important role in the detoxication of tertiary amines (Le., trimethylamine (136) and (5’)nicotine (58, 137)) to polar, readily excreted tertiary amine N-oxides. As discussed above, however, it is clear that the major form of FMO present in adult humans is structurally and functionally quite distinct from the major form of FMO present in many animals. For example, as shown in Table 1, the prediction is that mouse and not rat or rabbit would be a more appropriate animal model for human FMO action. Nevertheless, the suggestion that FMO evolved to oxidize and detoxicate nucleophilic heteroatom-containing chemicals or their metabolites present in many plants probably also applies to the adult human FMOs as well (2). As Ziegler has stated before, based on work done mainly with animal liver FMO1, the tissue distribution of FMO and the recalcitrance of FMO to inactivation by many chemicals present in plants that otherwise readily inactivate cytochromes P-450 and other hemoproteins suggest that the broad multisubstrate FMO is intended to prevent the untoward effects of bioactivation of chemicals by cytochromes P-450 ( 1 , 2 , 4 , IO). In large part, FMO oxygenates naturally occurring nitrogen- and sulfur-containing

Invited Review

compounds to less biologically toxic materials (138,139). For example, the pyrrolizidine alkaloid plant toxins senecionine, retrorsine, and monocrotaline are efficiently detoxicated by tertiary amine N-oxide formation (140)in a species such as the guinea pig with a relatively high level of FMO activity and a low level of pyrrole-forming cytochrome P-450 activity (141,142). In a species such as rat that is highly susceptible to the toxic properties of the plant alkaloids, just the opposite is true: pyrroleforming cytochrome P-450 activity is very great and detoxicating N-oxide-forming enzyme activity is very low. N-Oxygenation of another plant alkaloid, (5’)-nicotine, as discussed above, probably constitutes a detoxication route in animals (and humans) as well, shunting alkaloid substrate from the metabolic pathway mediated by cytochromes P-450 that generates the electrophilic (SInicotine A1’j5’-iminiumion. The neurotoxin 1-methyl-4phenyl-1,2,3,6-tetrahydropyridine(MPTP) is an excellent substrate for pig FMOl (143, 1441, and tertiary amine N-oxygenation of MPTP affords a polar metabolite that represents a major route for detoxication (145, 146). Presumably, in species with a low level of FMO activity, the majority of MPTP is metabolized via monoamine oxidase to the neurotoxic metabolites N-methyl-4-phenylpyridinium ion (MPP’) and l-methyl-4-phenyl-2,3dihydropyridinium ion (MPDP+). The conclusion from in vitro and in vivo studies is that FMO represents a detoxication pathway for MPTP in mice (145-147). The multisubstrate FMO from animal liver (i.e., FMO1) catalyzes the N-oxygenation of a wide array of tertiary and secondary N-containing compounds (148-153). In only a few cases does FMO catalyze the bioactivation to more reactive N-oxygenated metabolites. For example, N-alkyarylamines can be N-oxygenated by FMO to Nhydroxyarylamines required for subsequent metabolic activation to reactive esters implicated in the carcinogenic properties of arylamines in animals (154). FMO also N-oxygenates 1,l-dialkylhydrazines in a process that could contribute to the toxic properties of these materials (7). One of the determinants of the overall toxicity of a primary arylamine chemical may be dependent on the tendency of the amine to undergo N-methylation versus direct N-oxidation. N-Methylation of a secondary amine to provide a tertiary amine would then provide a substrate for FMO-mediated tertiary amine N-oxygenation, giving a tertiary amine N-oxide that would represent a detoxication pathway. There are a few examples where FMO may promote the formation of electrophilic metabolites due to nonenzymatic rearrangement (i.e., Cope-type elimination reactions) of enzymatically-generated tertiary amine N-oxides (25, 149, 151). For example, verapamil N-oxide is efficiently formed by FMO from the tertiary amine verapamil but is not indefinitely stable and undergoes decomposition to a hydroxylamine and 3,4-dimethoxystyrene (148). It is possible that formation of these unanticipated metabolites of verapamil may contribute to the cardiotoxicity observed with the parent drug. Another example of an unexpected elimination product arising from FMO-catalyzed formation of a tertiary amine N-oxide comes from homoallylic zimeldine (149). Copetype elimination of homoallylic zimeldine tertiary amine N-oxide produces a diene metabolite that could be further biotransformed into electrophilic materials. The closely related allylic tertiary amine antidepressant, zimeldine, was withdrawn from the market because of an unusual incidence of Guillain Barr6 syndrome (155). However,

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 175

it is not known what the relationship between zimeldine metabolism and the neurotoxicity is although the immunotoxicity data argue against an apparent causal relationship (156). Finally, secondary hydroxylamines can be metabolized by FMO to nitrones that can hydrolyze to electrophilic aldehydes (1,152). Hydroxylamines can also be oxidized to oximes by FMOl (157). The importance of these latter metabolic routes have not been exhaustively examined.

Role of FMO in the Toxicology of Sulfur-ContainingChemicals and Drugs The sulfur atom of sulfur-containing xenobiotics and drugs represents the preferred site for FMO oxygenation presumably because of the enhanced nucleophilicity of the heteroatom (1,2,4). As such, this class of chemical provides more examples of reactive metabolites produced by FMO. For example, depending on the structure, thiols, thioamides, 2-mercaptoimidazoles, thiocarbamates, and thiocarbamides all are efficiently S-oxygenated by FMO to electrophilic reactive intermediates. Without exception, however, the reactive metabolites do not inactivate FMO, but in some cases are sufficiently stable and electrophilic to covalently modify other nearby proteins including cytochromes P-450. For example, spironolactone thiol is S-oxygenated by FMO to a sulfenic acid (28). In the presence of microsomes depleted of glutathione, FMO-generated spironolactone sulfenic acid inactivates cytochrome P-450 (28). In the presence of glutathione, FMO-catalyzed glutathionyl-spironolactone disulfide formation was observed and fully characterized (28). This represents a good example whereby a thiol participates in a futile cycle, generating oxidized glutathione (in vivo) or a mixed disulfide if trapped by a thiaphile in vitro. Thioamides are among the best substrates for FMO and sequentially form mono- and diS-oxides (33,158,159). Remarkably, even thioamide S,Sdioxides (i.e., the chemical equivalent of a sulfur dioxide adduct to a carbene) does not inactivate FMO, but does efficiently covalently modify microsomal proteins (1601, presumably by acylation of the amide carbon atom (37, 158, 1611. 2-Mercaptoimidazoles are efficiently S-oxygenated to sulfenic acids by FMO as well as chemical oxidants that are subsequently S-oxygenated again to sulfinic acids ( I , 30,381. The intermediate sulfenic acid readily forms thiol adducts, resulting in disulfides that serve as subsequent sites for disulfide exchange and net thiol oxidation and substrate regeneration (32). Work of Ziegler and co-workers has shown that thiols that establish such a futile cycle catalyzing the oxidation of cellular thiols (Le., glutathione) and NADPH may render the cell susceptible to the toxic properties of other chemicals (2). The prediction that increased excretion of oxidized glutathione should be observed has been experimentally verified. For example, administration of chemicals including thioureas, mercaptoimidazoles, and spironolactone to animals has resulted in elevated levels of oxidized glutathione in vivo. The thiocarbamate functionality is widely used in agricultural products and is also sequentially S-oxygenated to S-oxides and sulfones (162-167). S-Oxygenation adjacent to a ketone functionality significantly increases the potential for generating powerful electrophilic acylating agents (168). It is not known whether thiocarbamate S-oxides or thiocarbamate sulfones are responsible for the toxic properties of many herbicides and fungicides in commercial use

176 Chem. Res. Toxicol., Vol. 8, No. 2, 1995

today, but FMO activity has been observed in a number of aquatic organisms (166, 167, 169, 170) possibly contaminated by thiocarbamates (171). Exposure of certain fish to thiocarbamates has been associated with toxic consequences ( I 721, but it remains t o be seen whether fish die-offs are associated with exposure to herbicides or pesticides that undergo oxidative bioactivation in vivo. Thioureas are another class of nucleophilic compound that are extremely efficiently S-oxygenated by FMO (30, 39, 138, 173). Depending on the substituents on the nitrogen atoms or whether the thiourea moiety is part of an aromatic ring system, sequential S-oxygenation by FMO may afford electrophilic sulfine metabolites that are either rapidly hydrolyzed (and nontoxic) or sufficiently stable to acylate biological macromolecules and exert the maximal toxicity observed for this class of chemical (27, 34,35,38). In summary, the relative rate of sulfenic acid oxidation (i,e., to reactive electrophilic sulfines) compared with the propensity for attack by a thiol or hydrolysis of the corresponding sulfines probably determines in large part the toxic potential to mammalian systems of thioureido-containing chemicals and drugs. While the physiological role of FMO is unknown, a few endogenous sulfur-containing compounds are S-oxygenated by FMO. Cysteamine is efficiently converted to the disulfide, cystamine, and this has been proposed to serve as an endogenous source of disulfide equivalents in the face of the strongly reducing environment of the cell (2, 174, 175). It has been proposed that an endogenous disulfide such as cystamine formed by FMO serves as a low molecular weight disulfide exchange agent, possibly required for the naturation of proteins that require a disulfide for their structure or function (274). Other endogenous sulfides are also substrates for FMO. As a model for endogenously prenylated peptides and proteins, farnesylcysteine methyl ester and derivatives were examined as substrates for FMO (116). The study was prompted by the observations that S-alkenylated cysteines and mercapturates arising from glutathione Sconjugate formation from haloalkenes were S-oxygenated by FMO (123, 124). S-Alkenylated cysteine sulfoxides and mercapturate sulfoxides are not indefinitely stable and may undergo spontaneous rearrangements and/or elimination reactions. We reasoned that if the terminal cysteine of the farnesylated ras protein was a substrate for FMO, S-oxygenation may result in cleavage of the farnesyl group to provide an inactivated protein that would not be associated with the membrane. Farnesylcysteine and farnesylcysteine methyl ester are efficiently S-oxygenated in a stereoselective fashion by FMO (Table 2), but further studies are required to establish a possible role of FMO-mediated S-oxygenation of farnesylated proteins and peptides in the membrane association and homeostasis of proteins such as the ras protein.

FMO Activity and Endogenous Substrates Based on several studies of FMO activity in animals and limited observations regarding human FMO, it is nevertheless possible that human FMO may be a factor in several human disease states. For example, the literature describes a few examples of individual’s who, instead of N-oxygenating trimethylamine to the polar, readily excreted trimethylamine (TMA)N-oxide, excrete large amounts of unmetabolized TMA in the urine and excrete the volatile and malodorous TMA in their breath, sweat, and skin (136, 137, 276-1791. Exogenous or

Cashman dietary supply of TMA or TMA obtained from other dietary sources (e.g., choline by metabolism) is normally almost completely N-oxygenated and excreted (180). TMA smells like the essence of rotting fish, and people who suffer from this apparent metabolic disorder have what is referred to as the “fish odor” syndrome. In humans, the only metabolic pathway observed for TMA is that of N-oxygenation to TMA N-oxide (181). The enzyme responsible for TMA N-oxygenation was initially attributed to human FMOl(94), but we now know that this cannot be correct because in adult human liver, very little FMOl is expressed (95, 97, 108, 109, 128). It is likely that adult human liver FM03 is the major TMA N-oxygenase, but this needs to be directly examined. A recent study reported that, for 11 subjects (6 females) examined with the syndrome, none had any obvious physical or mental abnormality although the subjects showed various psychosocial reactions including low selfesteem, frustration, anxiety, clinical depression, paranoia, suicidal personality, and addiction to cigarettes, alcohol, and other drugs (179). In addition to the psychosocial consequences associated with “fish odor” syndrome, the possibly more important consequences of trimethylaminuria on drug and endogenous amine metabolism as a result of decreased or absence of adult human FM03 activity has not been examined. It is possible that adult human FM03 deficiency may serve to foretell about other important human metabolic problems. The ability of humans to N-oxygenate TMA is distributed polymorphically (i.e., at least in some Caucasian populations studied thus far), and people with “fish odor” syndrome apparently are homozygous carriers for an allele that determines an individual’s ability to carry out the TMA N-oxygenation reaction (178). Unlike the X-linked deficiency linking monoamine oxidase and biogenic amine metabolic disorders (1821, trimethylaminuria is linked to an autosomal recessive gene. In this regard, it is notable that the localization of genes encoding human FMO1, FM02, and FM03 have been mapped to the long arm of chromosome l q (183). Thus, the genes that encode human FMO may form part of a gene cluster. It is possible that the trimethylaminuria metabolic defect represents a mutually common mutation because a large percentage of the cases examined thus far were Caucasian and apparently of Northern European descent. As discussed above, all FMOs including human FMOs are likely to be quite sensitive to post-mortem thermal inactivation and anaerobiosis. Kaderlik and Ziegler have shown that removal of molecular oxygen from FMOl preparations especially in the absence of NADPH results in an irreversible and precipitous loss of enzyme activity (77)even at ambient temperature. This observation may have important consequences for human disease states such as ischemic heart disease or ischemic bowel disease where a loss of molecular oxygen and blood flow to an organ may result in significant loss of FMO activity. Even temporary loss of FMO activity in a human or animal tissue may have significant effects on the ability of a tissue to detoxicate chemicals and drugs and otherwise participate in normal cellular homeostasis (184). On the other hand, during the ischemic condition it is possible that suppression of FMO activity decreases the cells’ normal propensity to form sulfenic acids from dietary foodstuffs and thus preserve cellular stores of NADPH and abolishes formation of reactive sulfine metabolites.

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 177

Invited Review

It is not known what effect immunosuppression has on FMO activity. It is clear, however, that in human conditions such as AIDS, the structure and function of intestinal villi disappear and place the subject at an enormous disadvantage with regard to normal absorption and metabolism of chemicals, drugs, and nutrients. If intestinal or other organ FMO activity turns out to be a key component in human xenobiotic metabolism and detoxication, the AIDS patient may be at a significant disadvantage with respect to the normal, protective processes of human intestinal FMO (27).

Future Directions The rapid progress that has been reported in the literature in the last 5 years concerning advances in molecular and structural biology of FMO is impressive. The role of FMO in the biotransformation of chemicals and endogenous substances including dietary agents will likely receive increased attention. Studies showing the contribution of FMO to human and animal drug and chemical metabolism will likely continue to increase. With the emergence of selective functional probes for different FMO forms and the future possible development of FMO isoform-selective inhibitors, the role of indivual FMO forms will undoubtedly be made more clear. As such, in vitro-in vivo correlations will undoubtedly provide needed information for the correct choice of experimental animals in the drug design and drug development regime. Ultimately, this fundamental enzyme metabolism information will be useful in the development of new human therapeutics. Preparations of substantially pure FMO either as cDNA-expressed enzymes in lower organisms or as stably transfected mammalian cell lines will continue to facilitate the evaluation of a role of FMO in the biotransformation processes of chemicals. Stable expression systems used to study single FMO forms or in combination with other forms or other monooxygenases will find widespread use in the evaluation of possible metabolic, toxic, or mutagenic properties of chemicals and drugs. Hepatocyte studies where FMO activity has been obtained in a stable and reproducible form will also be important in the evaluation of metabolism of chemicals (185). Studies with hepatocytes and stably transformed cells may be especially important in examining agricultural or environmental agents including many heteroatom-containing chemicals that may be good substrates for FMO but have not been exhaustively studied in the past. In addition, heterologous and stably transformed cDNA-expression systems may allow the evaluation of forms of FMO including brain and intestine or placental enzymes that may have new or interesting properties. It is possible that with more sensitive molecular probes of FMOs from a variey of sources, the fundamental physiological role of the enzyme will become apparent. Significant progress has been made toward the elucidation of the primary structure and binding channel topology (186) of FMO. As a membrane-associated protein, however, determination of the X-ray crystallographic structure may be a significant challenge. Future studies directed to the solution of structures of “model”flavoproteins such as bacterial cyclohexane monooxygenase may be an attractive alternative. Knowledge of a more detailed structural framework of FMO will enable a more rational approach to structurefunction studies including site-directed mutagenesis and

elaboration of chimeric proteins. Development of structural analogs or chimeric FMO proteins may prove to be very useful in producing biocatalysts for industrial chemical or biotechnology applications. In the future, important questions about gene regulation of FMO will undoubtedly lead to a more complete understanding about a role of FMO in human disease states (187).

Acknowledgment. I am grateful to the many coworkers and collaborators of my laboratory throughout the years who have made valuable contributions to the work discussed. The names of many of these co-workers are listed in the references. The author acknowledges the financial support of the National Institutes of Health (GM36426) and the cigarette and tobacco surtax fund of the State of California through the Tobacco-Related Disease Research Program of the University of California (21T0071). Addendum During the preparation of this review a report describing the S-oxygenation of L-methionine by rabbit FMO appeared in the literature suggesting that rabbit FM03 was responsible (188). The K,,,values reported for L-methionine S-oxygenation were 48.0,30.0, and 6.5 mM for cDNA-expressed rabbit FMO1, FM02, and FM03, respectively.

References (1) Ziegler, D. M. (1980) Microsomal flavin-containing monooxyge-

nase: oxygenation of nucleophilic nitrogen and sulfur compounds. In Enzymatic Basis ofDetoxication (Jakoby, W. B., Ed.) Vol. 1, pp 201-277, Academic Press, New York, NY. (2) Ziegler D. M. (1993)Recent studies on the structure and function of multisubstrate flavin-containing monooxygenases. Annu. Rev. Pharmacol. Toxicol. 33,179-199. (3) Hodgson, E., and Levi, P. E. (1992) The role of the flavincontaining monooxygenase in the metabolism and mode of action of agricultural chemicals. Xenobiotica 22, 1175-1183. (4) Ziegler, D. M. (1988) Flavin-containingmonooxygenases: catalytic mechanism and substrate specificities. Drug Metab. Rev. 19, 1-32. (5) Weisburger, J . H., and Weisburger, E. K. (1973) Biochemical formation and pharmacological, toxicological, aand pathological properties of hydroxylamines and hydroxamic acids. Pharmacol. Rev. 26, 1-66. (6) Damani, L. A. (1988) The flavin-containing monooxygenase as an amine oxidase. In Metabolism of Xenobiotics (Gorrod, G. W., Oelschlaeger, H., and Caldwell, J., Eds.) pp 59-70, Taylor and Francis, London. (7) Prough, R. A., Freeman, P. C., and Hines R. N. (1981). The oxidation of hydrazine derivatives catalyzed by the purified microsomal FAD-containing monooxygenase. J . Biol. Chem. 266, 4178-4184. (8) Smyser, B. P., and Hodgson, E. (1985) Metabolism of phosphorouscontaining compounds by pig liver microsomal FAD-containing monooxygenase. Biochem. Pharmacol. 34, 1145-1150. (9) Jones, K. C., and Ballou, D. P. (1986) Reactions of the 4ahydroperoxide of liver microsomal flavin-containing monooxygenase with nucleophilic and electrophilic substrates. J. Biol. Chem. 261,2553-2559. (10)Ziegler, D. M. (1990) Flavin-containing monooxygenases: enzymes adapted for multisubstrate specificity. Trends Pharmacol. Sci. 2, 321-324. (11) Ziegler, D. M., Graf, P., Poulsen, L., Sies, H., and Stal, W. (1992) NADPH-dependent oxidation of reduced ebselen, 2-selenylbenzanilide, and of 2-(methylseleno)benzanilidecatalyzed by pig liver flavin-containing monooxygenase. Chem. Res. Toxicol. 5, 163166. (12)Ziegler, D. M., and Poulsen, L. L. (1978) Hepatic microsomal mixed function amine oxidase. In Methods in Enzymology (Fleischer, S., and Packer, L., Eds.) Vol. 52, Part C, pp 142-151, Academic Press, New York, NY.

178 Chem. Res. Toxicol., Vol. 8, No. 2, 1995 (13) Venkatesh, K., Levi, P. E., and Hodgson, E. (1991) The effects of detergents on the purified flavin-containing monooxygenase of mouseliver, kidney and lungs. Gen. Pharmacol. 22, 549-552. (14) Poulsen, L. L., and Ziegler, D. M. (1979) The liver microsomal FAD-containing monooxygenases. Spectral characterization and kinetic studies. J. B i d . Chem. 254, 6449-6455. (15) Beaty, N. B., and Ballou, D. P. (1981)The reductive half-reaction of liver microsomal FAD-containing monooxygenase. J. Biol. Chem. 256,4611-4618. (16) Beaty, N. B., and Ballou, D. P. (1981) The oxidative half-reaction of liver microsomal FAD-containing monoxygenase. J. Biol. Chem. 256,4619-4625. (17) Kemal, C., and Bruice, T. C. (1976) Simple synthesis of a la-hydroperoxy adduct of a 1,5-dihydroflavine: Preliminary studies of a model for bacterial luciferase. Proc. Natl. Acad. Sci. U.S.A. 73,995-999. (18) Miller, A. (1982) A model for FAD-containing monooxygenase: the oxidation of thioanisole derivatives by an isoalloxazine hydroperoxide. Tetrahedron Lett. 23, 753-756. (19) Ball, S., and Bruice, T. C. (1980) Oxidation of amines by a 4a102,6498-6503. hydroxyperoxyflavin. J . Am. Chem. SOC. (20) Doerge, D. R., and Corbett, M. D. (1984) Primary arylamine oxidation by a flavin hydroperoxide. Biochem. Pharmacol. 33, 3615-3619. (21) Miller, A. E., Bischoff, J. J . Bizub, C., Luminoso, P., and Smiley, S. (1986) Electronic and steric effects in oxidations by isoallox108, 7773-7778. azine 4a-hvdro~eroxides.J . Am. Chem. SOC. Doerge, D."R., and Corbett, M. D. (1984) Hydroperoxy-mediated oxidations of organosulfur compounds. Mol. Pharmacol. 26,348352. Merenyi, G., and Lind, J. (1991) Chemistry of peroxidic tetrahedral intermediates of flavin. J. Am. Chem. SOC. 113,3146-3153. Bach, R. D., Owensby, A. L., Gonzalez, C., Schlegel, H. B., and McDouall, J . J.W. (1991) Nature of the transition structure for oxygen atom transfer from a hydroperoxide. Theoretical comparison between water oxide and ammonia oxide. J . Am. Chem. SOC. 113, 6001-6011. Cashman, J. R. (1991) Metabolism of tertiary amines by rat and hog liver microsomes: role of enzymatic Cope-elimination to a N-dealkylated product. Prog. Pharmacol. Clin. Pharmacol. 8, 117-126. Guengerich, F. P. (1987) Cytochrome P-450 enzymes and drug metabolism. Prog. Drug Metab. 10, 1-54. Hines, R. N., Cashman, J. R., Philpot, R. M., Williams, D. E., and Ziegler, D. M. (1994) The mammalian flavin-containing monooxygenases: molecular characterization and regulation of expression. Toxicol. Appl. Pharmacol. 125, 1-6. Decker, C. J., Cashman, J . R., Sugiyama, K., Maltby, D., and Correia, M. A. (1991) Formation of glutathionyl-spironolactone disulfide by rat liver cytochrome P-450 or hog liver monooxygenase: A functional probe of two-electron oxidations of the thiosteroid? Chem. Res. Toxicol. 4, 669-677. Decker, C. J., and Doerge, D. R. (1991) Rat hepatic microsomal metabolism of ethylene thiourea. Contributions of the flavincontaining monooxygenase and cytochrome P-450 isozymes. Chem. Res. Toxicol. 4, 482-489. Decker, C. J.,Doerge, D. R., and Cashman, J. R. (1992) Metabolism of benzimidazoline-2-thionesby rat hepatic microsomes and hog liver flavin-containing monooxygenase. Chem. Res. Toxicol. 5, 726-733. Poulsen, L. L., Hyslop, R. M., and Ziegler, D. M. (1979) SOxygenation of N-substituted thioureas catalyzed by the pig liver microsomal FAD-containing monooxygenase. Arch. Biochem. Biophys. 198, 78-88. Kreiter, P. A., Ziegler, D. M., Hill, K. E., and Burk, R. P. (1984) Increased biliary GSSG efflux from rat livers perfused with thiocarbamide substrates for the flavin-containing monooxygenases. Mol. Pharmacol. 26, 122-127. Hanzlik, R. P., and Cashman, J . R. (1982) Oxidation and other reactions of thiobenzamide derivatives of relevance to their hepatotoxicity. J . Org. Chem. 47, 4645-4650. Hui, Q. Y., Armstrong, C., Laver, G., and Iverson, F. (1988) Monooxygenase-mediated metabolism and binding of ethylene thiourea to mouse liver microsomal protein. Toxicol. Lett. 42,231237. Decker, C. J., and Doerge, D. R. (1992) Covalent binding of 14C and 35S-labeledthiocarbamides in rat hepatic microsomes. Biochem. Pharmacol. 43, 881-888. (36) Ziegler, D. M. (1982) Functional groups bearing nitrogen. In Metabollic Basis of Detoxication (Jakoby, W. B., and Caldwell, J., Eds.) pp 171-184, Academic Press, New York, NY. (37) Dryoff, M. C., and Neal, R. A. (1981) Identification of the major protein adduct formed in rat liver after thioacetamide administration. Cancer Res. 41, 3430-3435.

Cashman (38) Miller, A. E., Bischoff, J . J., and Pae, K. (1988) Chemistry of aminoiminomethanesulfinic and sulfonic acids related to the toxicity of thioureas. Chem. Res. Toxicol. 1, 169-174. (39) Kedderis, G. L., and Rickert, D. E. (1985) Loss of rat liver microsomal cytochrome P-450 during methimazole metabolism. Role of flavin-containing monooxygenase. Drug Metab. Dispos. 13, 58-61. (40) Neal, R. A. (1980) Microsomal metabolism of thiono-sulfur compounds: Mechanisms and toxicological significance. Reu. Biochem. Toxicol. 2, 131-171. (41) Chiele, E., and Malvaldi, G. (1984) Role of the microsomal FADcontaining monooxygenase in the liver toxicity of thioacetamide S-oxide. Toxicology 31, 41-52. (42) Williams, D. E., Ziegler, D. M., Nordin, D. J., Hale, S. E., and Masters, B. S. S. (1984) Rabbit lung flavin-containing monooxygenase is immunochemically and catalytically distinct from the liver enzyme. Biochem. Biophys. Res. Commun. 125, 116-122. (43) Tynes, R. E., Sabourin, P. J., and Hodgson, E. (1985) Identification of distinct hepatic and pulmonary forms of microsomal flavincontaining monooxygenase in the mouse and rabbit. Biochem. Biophys. Res. Commun. 126, 1069-1075. (44) Hlavica, P., and Golly, I. (1991) On the genetic polymorphism of the flavin-containing monooxygenase. In N-Oxidation of Drugs (Hlavica, P., and Damani, L. A., Eds.) pp 71-90, Chapman and Hall, London. (45) Williams, D. E., Hale, S. E., Muerhoff, A. S., and Masters, B. S. S. (1985) Rabbit lung flavin-containing monooxygenase. Purification, characterization and induction during pregnancy. Mol. Pharmucol. 28, 381-390. (46) Sabourin, P. J., and Hodgson, E. (1984) Characterization of the purified FAD-containing monooxygenase from mouse and pig liver. Chem.-Biol. Interact. 51, 125-139. (47) Tynes, R. E., and Philpot, R. M. (1987) Tisue- and speciesdependent expression of multiple forms of mammalian microsomal flavin-containing monooxygenase. Mol. Pharmacol. 31,569574. (48) Dannan, G. A., and Guengerich, F. P. (1982) Immunochemical comparison and quantitation of microsomal flavin-containing monooxygenase in various hog, mouse, rat, rabbit, dog and human tissues. Mol. Pharmacol. 22, 787-794. (49) Cashman, J. R., and Williams, D. E. (1990) Enantioselective S-oxygenation of 2-aryl-1,3-dithiolanes by rabbit lung enzyme preparations. Mol. Pharmacol. 37,333-339. (50) Cashman, J. R., Olsen, L. D., Boyd, D. R., McMordie, R. A. S., Dunlop, R., and Dalton, H. (1992) Stereoselectivity of enzymatic and chemical oxygenation of sulfur atoms in 2-methy-1,3-benzothiol. J. Am. Chem. SOC. 114, 8772-8777. (51) Rettie, A. E., Bogucki, B. D., Lim, I., and Meier, G. P. (1990) Stereoselective sulfoxidation of a series of alkyl p-tolysulfides by microsomal and purified flavin-containing monooxygenases. Mol. Pharmacol. 37, 643-651. (52) Nagata, T., Williams, D. E., and Ziegler, D. M. (1990) Substrate specificities of rabbit lung and porcine liver flavin-containing monooxveenases: Differences due to substrate suecificitv. Chem. Res. T o k k . 3, 372-376. (53) Cashman. J. R.. Proudfoot. J., Ho, Y. K., Chin. M. S., and Olsen, L. D. (1989) Chemical and enzymatic oxidation of 2-aryl-1,3oxathiolanes: mechanism of the hepatic flavin-containing monooxygenase. J . Am. Chem. SOC. 111, 4844-4852. (54) Lomri, N., Thomas, J., and Cashman, J. R. (1993) Expression in Escherichia coli of the cloned flavin-containing monooxygenase from pig liver. J. Biol. Chem. 268, 5048-5059. ( 5 5 ) Cashman, J. R., Olsen, L. D., and Bornheim, L. M. (1990) Enantioselective S-oxygenation by flavin-containing and cytochrome P-450 monooxygenases. Chem. Res. Toxicol. 3,344-349. (56) Cashman, J. R., and Olsen, L. D. (1990) Stereoselective Soxygenation of 2-aryl-1,3-dithiolanesby the flavin-containing and cytochrome P-450 monooxygenases. Mol. Pharmacol. 38, 573578. (57) Light, D. R., Waxman, D. J., and Walsh, C. (1982) Studies on the chirality of sulfoxidation catalyzed by bacterial cyclohexanone monooxygenase and hog liver flavin-adenine dinucleotide containing monooxygenase. Biochemistry 21, 2490-2498. (58) Park, S. B., Jacob, P., 111, Benowitz, N. L., and Cashman, 3. R. (1993) Stereoselective metabolism of (S)-(-)-nicotine in humans: Formation of trans-(S)-(-)-nicotineN-1'-oxide. Chem. Res. Toxicol. 6, 880-888. (59) Cashman, J . R., Park, S. B., Yang, Z.-C., Wrighton, S. A,, Jacob, P., 111, and Benowitz, N. L. (1992) Metabolism of nicotine by human liver microsomes: Stereoselective formation of trans nicotine &"-oxide. Chem. Res. Toxicol. 5, 639-646. (60) Damani, L. A,, Pool, W. F., Crooks, P. A,, Kaderlik, R. K., and Ziegler, D. M. (1988) Stereoselectivity in the "-oxidation of nicotine isomers by flavin-containingmonooxygenase. Mol. Pharmacol. 33, 702-705.

Invited Review (61) Thompson, J. A,, Norris, K. J., and Peterson, D. R. (1985)Isolation and analysis of N-oxide metabolites of tertiary amines: quantitation of nicotine-1’-N-oxide formation in mice. J . Chromatogr. 341,349-359. (62) Dixit, A., and Roche, T. E. (1984) Spectrophotometric assay of the flavin-containing monooxygenase and changes in its activity in female mouse liver with nutritional and diurnal conditions. Arch. Biochem. Biophys. 233, 50-63. (63) Ziegler, D. M., and Petit, F. H. (1964) Formation of an intermediate N-oxide in the oxidative demethylation of N,N-dimethylaniline catalyzed by liver microsomes. Biochem. Biophys. Res. Commun. 15, 188-193. (64) Arrhenius, E. (1969) Effects of various in vitro conditions on hepatic microsomal N- and C-oxygenation of aromatic amines. Chem.-Biol. Interact. 18, 361-380. (65) Willi, P., and Bickel, M. H. (1973) Liver metabolic reactions: tertiary amine N-dealkylation, tertiary amine N-oxidation, Noxide reduction, and N-oxide and N-dealkylation. 11. N,N-dimethylaniline. Arch. Biochem. Biophys. 156, 772-779. (66) Gorrod, J. W., Temple, D. J., and Beckett, A. H. (1975) The differentiation of N-oxidation and N-dealkylation of N-ethyl-Nmethylaniline by rabbit liver microsomes as distinct metabolic routes. Xenobiotica 5, 465-474. (67) Deveroux, T. R., and Fouts, J . R. (1974) N-Oxidation and demethylation of N,N-dimethylaniline by rabbit liver and lung microsomes. Effects of age and metals. Chem.-Biol. Interact. 8, 91-105. (68) Deveroux, T. R., Philpot, R. M., and Fouts, J. R. (1977) The effect of Hg+z on rabbit hepatic and pulmonary solubilized, partially purified N,N-dimethylaniline N-oxidases. Chem.-BWl. Interact. 18, 277-287. (69) Tynes, R. E., and Hodgson, E. (1985) Catalytic activity and substrate specificity of the flavin-containing monooxygenase in microsomal systems: characterization of the hepatic, pulmonary and renal enzymes of the mouse, rabbit and rat. Arch. Biochem. Biophys. 240, 77-93. (70) Kitchell, B. B., Rauckman, E. J., and Rosen, G. M. (1978) The effect of temperature on mixed function amine oxidase intrinsic fluoresence and oxidative activity. Mol. Pharmacol. 14, 10921098. (71) Wirth, P. J., and Thorgeirsson, S. S. (1978) Amine oxidase in mice-sex differences and developmental aspects. Biochem. Pharmacol. 27, 601-603. (72) Duffel, M. W., Graham, J. M., and Ziegler, D. M. (1981) Changes in dimethylaniline N-oxidase activity of mouse liver and kidney induced by steroid sex hormones. Mol. Pharmacol. 19, 134-139. (73) Dannan, G. A., Guengerich, F. P., and Waxman P. J. (1986) Hormonal regulation of rat liver microsomal enzymes. Role of gonadal steroids in programming maintenance and suppression of A4-steroid5-a-reductase,flavin-containing monwxygenase, and sex-specificcytochromes P-450.J. Biol. Chem. 261,10728-10735. (74) Cashman, J. R., Olsen, L. D., Lambright, C. E., and Presas, M. J. (1990)Enantioselective S-oxygenation ofpara-methoxy phenyl1,3-dithiolane by various tissue preparations: Effect of estradiol. Mol. Pharmacol. 37,319-327. (75) Lemoine, A,, Williams, D. E., Cresteil, T., and Leroux, 3. P. (1991) Hormonal regulation of microsomal flavin-containing monooxygenase: tissue dependent expression and substrate specificity. Mol. Pharmacol. 40, 211-217. (76) Williams, D. E., Meyer, H. H., and Dutchuk, M. S. (1989) Distinct pulmonary and hepatic forms of flavin-containingmonooxygenase in sheep. Comp. Biochem. Physiol. B 93, 465-470. (77) Kaderlik, R. F., Weser, E., and Ziegler, D. M. (1991) Selective loss of liver flavin-containing monooxygenases in rats on chemically defined diets. Prog. Pharmacol. Clin. Pharmucol. 3,95-103. (78) Nnane, I. P., and Damani, L. A. (1992) Ethylmethyl sulfide: a potential pharmacokinetic probe for monitoring the activity of the flavin-containingmonooxygenase in vivo in experimental animals. Br. J. Clin. Pharmacol. 34, 16OP-161P. (79) Nnane, I. P., and Damani L. A. (1992) Pharmacokinetics of trimethylamine: An alternative approach for monitoring the activity of the flavin-containing monooxygenase in vivo. J. Pharm. Pharmacol. 44, 1060. (80) Lawton, M. P., Cashman, J. R., Cresteil, T., Dolphin, C., Elfarra, A,, Hines, R. N., Hodgson, E., Kimura, T., Ozols, J., Phillips, I., Philpot, R. M., Poulsen, L. L., Rettie, A. E., Williams, D. E., and Ziegler, D. M. (1994)A nomenclature for the mammalian flavincontaining monooxygenase gene family based on amino acid sequence identities. Arch. Biochem. Biophys. 308, 254-257. (81) Donoghue, N. A., Norris, D. B., and Trudgill, P. W. (1976) The purification and properties of cyclohexane oxygenase from Nocardia globerula CL1 and Acinetobacter NCIB 9871. Eur. J . Biochem. 63, 175-192. (82) Ryerson, C. C., Ballou, D. P., and Walsh, C. (1982) Mechanistic studies on cyclohexanone oxygenase. Biochemistry 21,2644-2655.

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 179 (83) Nelson, D. R., Kamataki, T., Waxman, D. J., Guengerich, F. P., Estabrook, R. W., Feyereisen, R., Gonzalez, F. J., Coon, M. J., Gunsalus, I. C., Gotoh, O., Okuda, K., and Nebert, D. W. (1993) The P-450 superfamily: Update on new sequences, gene mapping, accession numbers, early trivial names of enzymes and nomenclature. DNA Cell Biol. 12, 1-51. (84) Ozols, J. (1990) Covalent structure of liver microsomal flavincontaining monooxygenase form 1. J . Biol. Chem. 265, 1028910299. (85) Ozols, J. (1991) Multiple forms of liver microsomal flavincontaining monooxygenases: Complete covalent structure of form 2. Arch. Biochem. Biophys. 290, 103-115. (86) Ozols, J. (1994) Isolation and structure of a third form of liver microsomal flavin-monooxygenase. Biochemistry 33,3751-3757. (87) Guan, S., Falick, A. M., and Cashman, J . R. (1990) N-Terminus determination: FAD and NADP+ binding domain mapping of hog liver flavin-containing monooxygenase by tandem mass spectrometry. Biochem. Biophys. Res. Commun. 170, 937-943. (88) Gasser, R., Tynes, R. E., Lawton, M. P., Korsmeyer, K. IC,Ziegler, D. M., and Philpot, R. M. (1990) The flavin-containing monooxygenase expressed in pig liver: Primary sequence, distribution, and evidence for a single gene. Biochemistry 29, 119-124. (89) Guan, S., Falick, A. M., Williams, D. E., and Cashman, J. R. (1991) Evidence for complex formation between rabbit lung flavincontaining monooxygenase and calreticulin. Biochemistry 30, 9892-9900. (90) Hlavica, P., Kellerman, J., Henschen, A,, Mann, K.-H., and Knuzel-Mulas, U. (1990). Evidence for the existence of structurally distinct hepatic and pulmonary forms of microsomal flavincontaining monooxygenase in the rabbit. Biol. Chem. Hoppe-Seyler 371, 521-526. (91) Fliegel, L., Burns, K., MacLennan, D. H., Reithmeier, R. A., and Michalak, M. (1989) Molecular cloning of the high affinity calcium-binding protein (calreticulin) of skeletal muscle sarcoplasmic reticulum. J. Biol. C h m . 264, 21522-21528. (92) Smith, M. J., and Koch, G. L. (1989) Multiple zones in the sequence of calreticulin (CRP55, calreulin, HACBP), a major calcium binding EWSR protein. EMBO J . 8, 3581-3586. (93) Nikbakht, K. N., Lawton, M. P., and Philpot, R. M. (1992) Guinea pig or rabbit lung flavin-containingmonooxygenases with distinct mobilities in SDS-PAGE are allelic variants that differ at only two positions. Pharmacogenetics 2, 207-216. (94) Dolphin, C., Shepard, E. A., Povey, S., Palmer, C. N., Ziegler, D. M., Ayesh, R., Smith, R. L., and Phillips, I. R. (1991) Cloning, primary sequence, and chromsomal mapping of a human flavincontaining monooxygenase (FMO1).J. Biol. Chem. 266, 1237912385. (95) Lomri, N., Gu, Q., and Cashman, J. R. (1992) Molecular cloning of the flavin-containing monwxygenase (form 11) cDNA from adult human liver. Proc. Natl. Acad. Sci. U.S.A. 89, 1685-1689. (96) Lemoine, A., Johann, M., and Cresteil, T. G . (1990) Evidence for the presence of distinct flavin-containing monooxygenases in human tissue. Arch. Biochem. Biophys. 276, 336-342. (97) Sadeque, A. J. M., Eddy, A. C., Meier, G. P., and Rettie, A. E. (1992) Stereoselective sulfoxidation by human flavin-containing monooxygenase. Drug Metab. Dispos. 20, 832-839. (98) Bhamre, S., and Ravindranath, V. (1991) Presence of flavincontaining monooxygenase in rat brain. Biochem. Pharmacol. 42, 442 -444. (99) Atta-Asafo-Adjei, E., Lawton, M. P., and Philpot, R. M. (1993) Cloning, sequencing, distribution and expression in E. coli of flavin-containing monooxygenase 1C1. Evidence for a third gene subfamily in rabbits. J. Biol. Chem. 268, 9681-9689. (100) Dolphin, C. T., Shephard, E. A., Povey, S., Smith, R. L., and Phillips, I. R. (1992) Cloning, primary sequence and chromosomal localization of human FM02, a new member of the flavincontaining monooxygenase family. Biochem. J. 287, 261-267. (101) Lawton, M. P., and Philpot, R. M. (1993) Functional characterization of flavin-containing monooxygenase 1B1 expressed in Saccharomyces cerevisiae and Escherichia coli and analysis of proposed FAD- and membrane binding domains. J. Biol. Chem. 268, 5728-5734. (102) Kozak, M. (1987) An analysis of 5’-noncoding sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res. 15, 81258133. (103) Flinta, C., Persson, B., Jornvall, H., and von Heijne, G. (1986) Sequence determinants of cytosolic N-terminal protein processing. Eur. J . Biochem. 154, 193-196. (104) Korsmeyer, K. K., Poulsen, L. L., and Ziegler, D. M. (1991) In Flavins and Flavoproteins (Curti, B., Ronchi, S., and Zanetti, G., Eds.) pp 243-246, de Gruyter, Berlin. (105)Lomri, N., Yang, Z.-C., and Cashman, J . R. (1993) Expression in Escherichia coli of the flavin-containing monooxygenase from adult human liver. Determination of a distinct tertiary amine substrate specificity. Chem. Res. Toxicol. 6, 425-429.

180 Chem. Res. Toxicol., Vol. 8, No. 2, 1995 (106) Lomri, N., Yang, Z.-C., and Cashman, J . R. (1993) Regio- and stereoselective oxygenations by adult human liver flavincontaining monooxygenase 3. Comparison with forms 1 and 2. Chem Res. Toxicol. 6, 800-807. (107) Kaderlik, K. (1990) Potential mechanisms for regulation of hepatic flavin-containing monooxygenases, Ph.D. Thesis, University of Texas, Austin. (108) Cashman, J. R., Yang, Z.-C., Yang, L., and Wrighton, S. (1993) Stereo- and regioselective N- and S-oxidation of tertiary amines and sulfides in the presence of adult human liver microsomes. Drug Metab. Dispos. 21, 492-501. (109) Wrighton, S. A,, Vandenbranden, M., Stevens, J . C., Shipley, L. A,, Ring, B. J., Rettie, A. E., and Cashman, J. R. (1993) In vitro methods for assessing human hepatic drug metabolism: their use in drug development. Drug Metab. Reu. 25, 453-484. (110) Itoh, K., Kimura, T., Yokoi, T., Itoh, S., and Kamataki, T. (1993) Rat liver flavin-containing monooxygenase (FMO): cDNA cloning and expression in yeast. Biochim. Biophys. Acta 1173,165171. (111) Lawton, M. P., Kronbach, T., Johnson, E. F., and Philpot, R. M. (1991) Properties of expressed and native flavin-containing monooxygenases: evidence of multiple forms in rabbit liver and lung. Mol. Pharmacol. 40, 692-698. (112) Cashman, J . R., and Hanzlik, R. P. (1981) Microsomal oxidation of thiobenzamide: a photometric assay for the flavin-containing monooxygenase. Biochem. Biophys. Res. Commun. 98,147-153. (113) Ziegler, D. M., Poulsen, L. L., and McGee, E. M. (1971) Interaction of primary amines with a mixed-function amine oxidase isolated from pig liver microsomes. Xenobiotica 1,523531. 14) Fujimori, K., Yaguchi, M., Mikami, A., Matsuura, T., Furukawa, N., Oae, S., and Iyanagi, T. (1986) Kinetic solvent deuterium isotope effect on the oxygenation of N,N-dimethylaniline with the pig liver microsomal FAD-containing monooxygenase. Tetrahedron Lett. 27, 1179-1182. 15) Taylor, K. H., and Ziegler, D. M. (1987) Studies on substrate specificity of the hog liver flavin-containing monooxygenase: anionic organic sulfur compounds. Biochem. Pharmacol. 36, 141-146. 16) Park, S. B., Howald, W. N., and Cashman, J. R. (1994) SOxidative cleavage of farnesylcysteine and farnesylcysteine methyl ester by the flavin-containing monooxygenase. Chem. Res. Toxicol. 7, 191-198. (117) Holland, H. L. (1988) Chiral sulfoxidation by biotransformation of organic sulfides. Chem. Reu. 88, 473-485. (118) Levi, P. E., and Hodgson, E. (1988) Stereospecificity in the oxidation of phorate and phorate sulfoxide by purified FADcontaining monooxygenase and cytochrome P-450 isozymes. Xenobiotica 18, 29-39. (119) Poulsen, L. L., Taylor, K., Williams, D. E., Masters, B. S. S., and Ziegler, D. M. (1986) Substrate specificity of the rabbit lung flavin-containing monooxygenase for amines: Oxidation products of primary alkylamines. Mol. Pharmacol. 30, 680-685. (120) Tynes, R. E., Sabourin, P. J., Hodgson, E., and Philpot, R. M. (1986) Formation of hydrogen peroxide and N-hydroxylated amines catalyzed by pulmonary flavin-containing monooxygenase in the presence of primary alkylamines. Arch. Biochem. Biophys. 251, 654-664. (121) Yamada, H., Yuno, K., Oguri, K., and Yoshimura, H. (1990) Multiplicity of hepatic microsomal flavin-containing monooxygenase in the guinea pig: its purification and characterization. Arch. Biochem. Biophys. 280, 305-312. 22) Ziegler, D. M., Poulsen, L. L., and York, B. M. (1983) Role of the flavin-containing monoxygenase in maintaining cellular thiol: disulfide balance. In Function of Glutathione: Biochemical, Physiological, Toxicological and Clinical Aspects (Larson, A,, et al.,Eds.) pp 297-305, Raven Press, New York. 23) Sausen, P. J., and Elfarra, A. A. (1990) Cysteine conjugate S-oxidase: characterization of a novel enzymatic activity in rat hepatic and renal microsomes. J . Biol. Chem. 265, 6139-6145. 24) Park, S. B., Osterloh, J . D., Vamvakas, S., Hashmi, M., Anders, M. W., and Cashman, J. R. (1992) Flavin-containing monooxygenase-dependent stereoselective S-oxygenation and cytotoxicity of cysteine S-conjugates and mercapturates. Chem. Res. Toxicol. 5, 193-201. (125) Gold, M. S., and Ziegler, D. M. (1973) Dimethylaniline N-oxidase and aminopyrine N-demethylase activities of human liver tissues. Xenobiotica 3, 179-189. (126) McManus, M. E., Stupans, I, Burgess, W., Koenig, J. A,, Hall, P. dela M., and Birkett, D. J . (1987) Flavin-containing monooxygenase activity in human liver microsomes. Drug Metab. Dispos. 15, 256-261. (127) Ziegler, D. M., and Gold, M. S. (1971) Oxidative metabolism of tertiary amines by human liver tissue. Xenobiotica 1,325-326.

Cashman (128) Burnett, V. L., Lawton, M. P., and Philpot, R. M. (1994) Cloning and sequencing of flavin-containingmonooxygenases FM03 and FM04 from rabbit and characterization of FM03. J . Biol. Chem. 269, 14314-14322. (129) Rane, A. (1973) N-Oxidation of a tertiary amine (N,N-dimethylaniline) by human fetal liver microsomes. Clin. Pharmacol. Ther. 15,32-38. (130) Sadeque, A. J., Thummel, K. E., and Rettie, A. E. (1993) Purification of macaque liver flavin-containing monooxygenase: a form of the enzyme related immunologically t o an isozyme expressed selectively in adult human liver. Biochim. Biohys. Acta 1162, 127-134. (131) Lipsy, R. J., Fennerty, B., and Fagan, T. C. (1990) Clinical review of histamine 2 receptor antagonists. Arch. Intern. Med. 150, 745-751. (132) Winzor, D. J., Ioannoni, B., and Reilly, P. E. B. (1986) The nature of microsomal monooxygenase inhibition by cimetidine. Biochem. Pharmacol. 35, 2157-2161. (133) Taylor, D. C., and Cresswell, P. R. (1975) The metabolism of cimetidine in the rat, dog and man. Drug Metab. Dispos. 6, 2130. (134) Cashman, J. R., Park, S. B., Yang, Z.-C., Washington, C. B., Gomez, D. Y., Giacomini, K. M., and Brett, C. M. (1993) Chemical, enzymatic and human enantioselective S-oxygenation of cimetidine. Drug Metab. Dispos. 21, 587-597. (135) Stevens, J. C., Shipley, L. A., Cashman, J. R., Vandenbranden M., and Wrighton, S. A. (1993)Comparison of human and rhesus monkey in vitro phase I and phase I1 hepatic drug metabolism activities. Drug Metab. Dispos. 21, 753-760. (136) Cholerton S., and Smith, R. L. (1991) Human pharmacogenetics of nitrogen oxidation. In N-oxidation of Drugs: Biochemistry, Pharmacology, Toxicology (Hlavica, P., and Damani, L. A,, Eds.) pp 107-131, Chapman and Hall, London. (137) Ayesh, R., Al-Waiz, M., and Crother, N. J . (1988) Deficient nicotine N-oxidation in two sisters with trimethylaminuria. Br. J . Clin. Pharmacol. 25, 664P-665P. (138) Poulsen, L. L. (1991) The multisubstrate FAD-containing monooxygenase. In Chemistry und Biochemistry of Flavoenzymes (Miller, F., Ed.) Vol. 2, pp 87-100, CRC Press, Boca Raton, FL. (139) Yuno, K., Yamada, H., Oguri, K., and Yoshimura, H. (1990) Substrate specificity of guinea pig liver flavin-containing monooxygenase for morphine, tropane, and Strychnos alkaloids. Biochem. Pharmacol. 40, 2380-2382. (140) Williams, D. E., Reed, R. L., Kedzierski, B., Ziegler, D. M., and Buhler, D. R. (1989)The role of flavin-containing monooxygenase in the N-oxidation of the pyrrolizidine alkaloid senecionine. Drug Metab. Dispos. 17, 380-386. (141) Mirand, C. L., Chung, W., Reed, R. E., Zhao, X., Henderson, M. C., Wang, J.-L.,Williams, D. E., and Buhler, D. R. (1991)Flavincontaining monooxygenase: A major detoxifying enzyme for the pyrrolizidine alkaloid senecionine in guinea pig tissues. Biochem. Biophys. Res. Commun. 178, 546-552. (142) Williams, D. E., Reed, R. L., Kedzierski, B., Guengerich, F. P., and Buhler, D. R. (1989) Bioactivation and detoxication of the pyrrolizidine alkaloid senecionine by cytochrome P-450 isozymes in rat liver. Drug Metab. Dispos. 17, 387-392. (143) Cashman, J . R., and Ziegler, D. M. (1986) Contribution of N-oxygenation to the metabolism of MPTP (l-methyl-l-phenyl1,2,3,6-tetrahydroppyridine) by the flavin-containing monooxygenase by various liver preparations. Mol. Pharmacol. 29, 163167. (144) Cashman, J. R. (1988) Facile N-oxygenation of l-methyl-4phenyl-1,2,3,6-tetrahydropyridineby the flavin-containing monooxygenase. A convenient synthesis of 3H-MPDP+.J . Med. Chem. 31, 1258-1261. (145) Chiba, K., Kubota, E., Miyakawa, T., Kato, Y., and Ishizaki, T. (1988) Characterization of hepatic microsomal metabolism as an in vivo detoxication pathway of l-methyl-4-phenyl-l,2,3,6tetrahydropyridine in mice. J . Pharmacol. Exp. Ther. 246,11081115. (146) Chiba, R, Horii, H., Kubata, E., Ishizaki, T., and Kato, Y. (1990) Effects of N-methylmercaptoimidazole on the disposition of MPTP and its metabolites in mice. Eur. J . Pharmacol. 180,5967. (147) Di-Monte, D. A,, Wu, E. Y., Irwin, I., Delanney, L. E., and Langston, J. W. (1991) Biotransformation of 1-methyl-4-phenyl1,2,3,6-tetrahydropyridinein primary cultures of mouse astrocytes. J . Pharmacol. Enp. Ther. 258, 594-600. (148) Cashman, J. R. (1989) Enantioselective N-oxygenation of verapamil by the hepatic flavin-containing monooxygenase. Mol. Pharmacol. 36, 497-503. (149) Cashman, J. R., Proudfoot, J., Pate, D. W., and Hogberg, T. (1988) Stereoselective N-oxygenation of zimeldine and homozimeldine by the flavin-containingmonooxygenase. Drug Metab. Dispos. 16, 616-622.

Chem. Res. Toxicol., Vol. 8, No. 2, 1995 181

Invited Review (150) Cashman, J. R., Celestial, J. R., and Leach, A. R. (1992) Enantioselective N-oxygenation of chlorpheniramine by the flavin-containing monooxygenase from pig liver. Xenobiotica 22, 459-469. (151) Mani, C., and Kupfer, D. (1991) Cytochrome P-450-mediated activation and irreversible binding of the antiestrogen tamoxifen to proteins in rat and human liver: possible involvement of flavin-containine monooxvgenases in tamoxifen activation. Can"cer Res. 51, 6Og-6058. (152) Cashman. J. R.. Yane. Z.-C.. and Hogberg, T. (1990) Oxidation of N-hydroxynorzimidine to a stabie n&one by hepatic monooxygenases. Chem. Res. Toxicol. 3, 428-432. (153) Vyas, K. P., Kari, P. H., Ramjit, H. G., Pitzenberger, S. M., and Hichens, M. (1990)Metabolism of antiparkinson agent dopazinol by rat liver microsomes. Drug Metab. Dispos. 18, 1025-1030. (154) Ziegler, D. M., Ansher, S. S., Nagata T., Kadlubar, F. F., and Jakoby, W. B. (1988) N-Methylation: potential mechanism for metabolic activation of carcinogenic primary arylamines. Proc. Natl. Acad. Sci. U.S.A. 85, 2514-2517. (155) Fagius, J., Osterman, P. O., Siden, A., and Wilholm, B.-E. (1985) Guillain-Barre syndrome following zimeldine treatment. J. Neurol., Neurosurg. Psychiatry 65-69. (156) Thomas, C., Groten, J., Kammuller, M. E., DeBakker, J. M., Seinen, W., and Bloksma, N. (1989) Popliteal lymph node reactions in mice induced by the drug Zimeldine. Int. J. ImmunoPharmacol. 11, 693-702. (157) Clement, B., Lustig, K. L., and Ziegler, D. M. (1993) Oxidation of desmethylpromethazine catalyzed by pig liver flavin-containing monooxygenase. Number and nature of metabolites. Drug Metab. Dispos. 21, 24-29. (158) Hanzlik, R. P., and Cashman, J. R. (1983) Microsomal metabolism of thiobenzamide and thiobenzamide S-oxide. Drug Metab. DZSPOS.11, 201-209. (159) Cashman, J. R. (1989) Thioamides. In Sulfur-containing drugs and related organic compounds (Damani, L. A., Ed.) Vol. 1,Part B, pp 35-48, Ellis Honvood, Chichester. (160) Hanzlik, R. P. (1986)Chemistry of covalent binding: studies with bromobenzene and thiobenzamide. Adv. Exp. Med. Biol. 197, 31-40. (161) Cashman, J. R., Parikh, K. K., Traiger, G. J., and Hanzlik, R. P. (1983) Relative hepatotoxicity of ortho and meta monosubstituted thiobenzamides in the rat. Chem.-Biol. Interact. 45, 341-347. (162) Ziegler, D. M. (1989) S-Oxygenases, I: Chemistry and biochemistry. In Sulfur-containing drugs and related organic compounds (Damani, L. A., Ed.) Vol. 2, Part A, pp 53-66, Ellis Horwood, Chichester. (163) Hutson, D. H. (1981) S-Oxygenation in herbicide metabolism in mammals. In Sulfur in Pesticide Action and Metabolism, ACS Svmuosium Series. DD 65-82. American Chemical Societv. Wasiington, DC. (164) Chen. S. Y.. and Casida, J. E. (1978) Thiocarbamate herbicide metabolism: microsomal'oxygenase metabolism of E W C involving mono-and di-oxygenation at the sulfur and hydroxylation a t each alkyl carbon. J . Agric. Food Chem. 26, 263-267. (165) Casida, J.,E., Gray, R. A., and Tilles, H. (1974) Thiocarbamate sulfoxides: uotent. selective and biodezradable herbicides. Science 184,533-574. (166) Cashman. J. R.. Olsen, L. D., Young, G.. and Bern, H. A. (1989) S-Oxygenation of eptam in hepatic-mirosomes from fresh- and saltwater striped bass (Morone saxatilis) and mammalian systems. Chem Res. Toxicol. 2, 392-399. (167) Cashman, J. R., Olsen, L. D., Nishioka, R. S., Gray, E. S., and Bern H. A. (1990) S-Oxygenation of Thiobencarb (Bolero) in hepatic preparations from striped bass (Morone saxatilis) and mammalian systems. Chem. Res. Toxicol. 3, 433-440. (168) Casida, J. E., and Ruzo, L. 0. (1986) Reactive intermediates in pesticide metabolism: peracid oxidations as possible biomimetic models. Xenobiotica 16, 1003-1015. (169) Schlenk, D., and. Buhler, D. R. (1993) Immunological characterization of flavin-containing monooxygenases from the liver of a rainbow trout (Oncorhynchus mykiss): sexual- and age-dependent differences and the effect of trimethylamine on enzyme regulation. Biochim. Biophys. Acta 1156, 103-106. I

..

-

(170) Schlenk, D., and Buhler, D. R. (1991) Role of flavin-containing monooxygenase in the in vitro biotransformation of aldicarb in rainbow trout (Oncorhynchus mykiss). Xenobiotica 21, 15831589. (171) Cashman, J. R., Maltby, D. A., Nishioka, R. S., Bern, H. A., Gee, S. J., and Hammock, B. D. (1992) Chemical contamination and the annual summer die-off of striped bass (Moronne saxatilis) in the Sacramento-San Joquin Delta. Chem. Res. Toxicol. 5, 100-105. (172) Schimmel, S. C., Garnas, R. L., Patrick, J. M., and Moore, J. C. (1983) Acute toxicity, bioconcentration and persistence of AC 222,705, benthiocarb, chlorpyrifos, fenvalerate, methyl parathion, and permethrine in the estuarine environment. J. Agric. Food Chem. 31,104-113. (173) Guo, W. X., and Ziegler, D. M. (1991) Estimation of flavincontaining monooxygenase activities in crude tisue preparations by thiourea-dependent oxidation of thiocholine. Anal. Biochem. 198, 143-148. (174) Poulsen, L. L., and Ziegler, D. M. (1977) Microsomal mixed function oxidase-dependent renaturation of reduced ribonuclease. Arch. Biochem. Biophys. 183, 563-570. (175) Ziegler, D. M. (1985) Role of reversible oxidation-reduction of enzyme thiol-disulfides in metabolic regulation. Annu. Rev. Biochem. 64, 305-329. (176) Chen, H., and Aiello, F. (1993) Trimethylaminuria in a girl with Prader-Willi syndrome and de1(15)(qllq13).Am. J. Med. Genet. 45, 335-339. (177) Humbert, J. R., Hammond, K. B., Hathaway, W. E., Marcoux, J. G., and O'Brien, D. (1970) Trimethyaminuria-the fish odor syndrome Lancet i, 770-771. (178) Al-Waiz, M., Ayesh, R., Mitchell, S. C., Idle, J. R., and Smith, R. L. (1987) Clin. PharmacoE. Ther. 42, 588-594. (179) Ayesh, R., Mitchell, S. C., Zhang, A., and Smith, R. L. (1993) The fish odor syndrome: biochemical, familial and clinical aspects. Br. Med. J. 307, 655-657. (180) Al-Waiz, M., Ayesh, R., Mitchell, S. C., Idle, J. R., and Smith, R. L. (1989) Trimethylaminuria: the detection of carriers using a trimethylamine load test. J . Inherited Metab. Dis. 12, 80-85. (181) Al-Waiz, M., Mitchell, S. C., Idle, J. R., and Smith R. L. (1987) The metabolism of 14C-labeledtrimethylamine and its N-oxide in man. Xenobiotica 17, 551-558. (182) Brunner, H. G., Nelen, M. R., van Zandvoort, P., Abeling, N. G. G. M., van Gennip, A. H., Wolters, E. C., Kuiper, M. A,, Ropers, H. H., and van Oost, B. A. (1993) X-Linked borderline mental retardation with prominent behavioral disturbance: phenotype, genetic localization, and evidence for disturbed monoamine metabolism. Am. J . Hum. Genet. 52, 1032-1039. (183) Shephard, E. A,, Dolphin, C. T., Fox, M. F., Povey, S., Smith, R., and Philips, I. R. (1993)Localization of genes encoding three distinct flavin-containing monooxygenases to human chromsome lq. Genomics 16, 85-89. (184) Brodfuehrer, J. I., and Zannoni, V. G. (1986) Modulation of the flavin-containing monooxygenase in guinea pig by ascorbic acid and food restriction J. Nutr. 117, 286-290. (185) Croeke, S., Mertens, K., Sagaert, A., Callaerts, A,, Vercruysse, A,, and Rogiers, V. (1992) Spectrophotometric measurement of flavin-containingmonooxygenase activity in freshly isolatied rat hepatocytes and their cultures. Anal. Biochem. 205, 285-288. (186) Cashman, J. R., Celestial, J. R., Leach, A. R., Newdoll, J., and Park, S. B. (1993) Preferred conformations for N-oxygenation by the hog liver flavin-containing monooxygenase. Pharm. Res. 10, 1097-1105. (187) Cashman, J. R., Park, S. B., Berkman, C., and Cashman, L. E. (1994)Role of hepatic flavin-containing monooxygenase 3 in drug and chemical metabolism in adult humans. Chem.-Biol. Interact. (in press). (188) Duescher, R. J., Lawton, M. P., Philpot, R. M., and Elfarra, A. A. (1994) Flavin-containing monooxygenase (FM0)-dependent metabolism of methionine and evidence for FM03 being the major FMO involved in methionine sulfoxidation in rabbit liver and kidney microsomes. J. Biol. Chem. 269, 17525-17530.

TX940091B