Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Interindividual variation resulting in populationwide differences in the metabolism of foreign compounds (i.e., xenobiotics) may contribute to the susceptibility of humans to adverse chemicals or drug reactions and disease states. Most genetic variation can be accounted for by more prevalent alleles (i.e., alleles with a frequency of greater than 1% in the general population). It is probable that common genetic variants may contribute significantly to disease susceptibility (Risch and Merikangas, 1996).

Pharmacogenomics allows for the identification of genetic variation in drug-metabolizing enzymes and the identification of individuals who will benefit most or least from a given medication (Kalow and Grant, 1995; Brown and Hartwell, 1998; Housman and Ledley, 1998; Kleyn and Vesell, 1998; Persidis, 1998).

Flavin-containing monooxygenases (FMOs; EC 1.14.13.8)¹ are microsomal NADPH-dependent flavoprotein enzymes that catalyze the oxygenation of nucleophilic nitrogen-, sulfur-, phosphorus-, and other heteroatom-containing chemicals, drugs, and pesticides. They represent a family of monooxygenases in mammals designated FMO, forms 1 to 5. FMO3 constitutes the major adult human hepatic isoform (Burnett et al., 1994; Cashman, 1995; Falls et al., 1997). Human FMOs are 532 to 558 amino acids in length, with specific amino acids highly conserved in all species, particularly residues 4 to 32 and 186 to 213, which contain the FAD- and NADPH-binding domains, respectively (Cashman, 1995). FMO enzymes have a broad substrate specificity, efficiently oxygenating various nucleophilic heteroatom-containing compounds that are not precluded from the substrate-binding domain on the basis of steric limitations (Ziegler, 1990). Primary and tertiary amines are good substrates for human FMO3 (Cashman, 1995). The human FMO3 substrates include the tertiary amine trimethylamine (TMA) (Cashman et al., 1997; Lang et al., 1998), clozepine (Tugnait et al., 1997), and (S)-nicotine (Cashman et al., 1992). Commonly used drugs such as methimazole and cimetidine (Cashman, 1995; Ziegler, 1990) and the primary amines tyramine (Lin and Cashman, 1997a) and phenylethylamine (Lin and Cashman, 1997b) are also oxygenated by human FMO3. Although substrates are generally detoxicated by FMO-mediated metabolism, the enzyme has been implicated in the bioactivation of several xenobiotics (Cashman, 1995). Inactivating variants or polymorphisms of human FMO3 may contribute to the pathophysiology of disease and adverse reactions or exaggerated clinical response to specific medications.

Trimethylaminuria is an inborn error of metabolism that causes deficiency in N-oxygenation of the odorous tertiary amine TMA to its nonodorous metabolite TMA N-oxide (Zhang et al., 1995; Treacy et al., 1998). Results from our laboratory and others have recently described mutations of the human FMO3 gene that cause trimethylaminuria (Zhang et al., 1995; Cashman et al., 1997; and Dolphin et al., 1997; Treacy et al., 1998). Human FMO3 mutations include the nonsense mutation E305X and the missense mutations P153L, M66I, and R492W.

During our previous studies directed at human FMO3 mutation detection and haplotype analysis of probands with trimethylaminuria, we identified several prevalent nucleotide polymorphisms of the human FMO3 gene (Treacy et al., 1996). Polymorphisms that are prevalent in French Canadian and Australian populations were studied. In this article, we describe the cDNA-expression analysis for two of the prevalent human FMO3 polymorphisms showing altered N-oxygenase activities, indicating that these are significant pharmacogenetic polymorphisms.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Study Population. The human FMO3 polymorphisms E158K (c.488 G-A) and V257M (c.769 G-A) were previously identified in a cohort of individuals with trimethylaminuria and healthy control subjects from Melbourne, Australia, with single-stranded conformational polymorphism (SSCP) screening and sequencing of the FMO3 gene (Treacy et al., 1998). After institutional ethics approval, the frequency of these substitutions was determined in 170 normal control individuals from Quebec (i.e., 110 Francophones and 60 Anglophones) and 50 normal controls from Victoria, Australia. Genomic DNA was prepared from lymphocyte preparations with standard procedures. Amplified DNA fragments obtained by the polymerase chain reaction (PCR) were subjected to restriction enzymatic digestion and visualized by ethidium bromide staining after agarose or polyacrylamide gel electrophoresis. Table 1 illustrates the conditions used for each diagnostic assay.

Data Analysis. The statistical significance of the frequency of the human FMO3 E158K and V257M polymorphisms between Canadians (i.e., Anglophones and Francophones) and Australians was calculated. A ² analysis was used for human FMO3 E158K to test the hypothesis that the proportion of chromosomes was not significantly different among the groups of English Canadians, French Canadians, and Australians (², 2 df = 2.52, P = 0.28). The Mantel-Haenszel test for small sample size was used for human FMO3 V257M. There was no significant difference observed between the presence of the variant alleles in the three cohorts (², 2 df = 4.71, P = 0.095).

Chemicals. Aromatic ring 3,5-[³H]tyramine hydrochloride (40 Ci/mmol, 98% radiopurity) was obtained from American Radioactivity Company (St. Louis, MO) and diluted with tyramine hydrochloride to afford working stocks of 1 mCi/mmol. [¹⁴C]TMA hydrochloride (1 mCi/mmol, 99.4% radiopurity) was obtained from Sigma Chemical Co. (St. Louis, MO) and diluted with TMA hydrochloride to afford working stocks of 0.4 mCi/mmol. All chemicals and reagents were purchased from Aldrich Chemical Co. (Milwaukee, WI) in the highest purity commercially available. All of the agents of the NADPH-generating system were obtained from Sigma Chemical Co.

Synthesis. 10-(5-Aminopentyl)-2-(trifluoromethyl)phenothiazine (5-APT), its hydroxylamine, and cis and trans oximes were synthesized by a modification of the procedures previously described (Lin et al., 1996). The tertiary amine 10-(N,N-dimethylaminopentyl)-2-(trifluoromethyl)phenothiazine (5-DPT) and its N-oxide were also synthesized by a procedure (Brunelle et al., 1997) similar to the one previously described (Nagata et al., 1990). The hydroxylamine and cis and trans oximes of tyramine were synthesized as previously described (Lin and Cashman, 1997a).

cDNA Expression and Substrate Analysis. The human FMO3 cDNAs were expressed as maltose-binding fusion proteins. Site-directed mutagenesis for the human FMO3 substitutions E158K and V257M were performed as previously described (Brunelle et al., 1997; Cashman et al., 1997; Treacy et al., 1998). For comparison, a cDNA construct for the truncation variants E305X (previously reported) and 510X were also prepared as the maltose-binding protein (MBP) fusions.

Subcloning Human FMO1 and FMO3 cDNA into the MBP Fusion Expression System. Human FMO3 or FMO1 cDNA was inserted into the expression vector pMAL-c2 and PCR amplification was done in a fashion that allowed for fusion of human FMO3 or FMO1 cDNA at the 3'-end of sequences encoding the MBP as previously described (Brunelle et al., 1997; Treacy et al., 1998). Each cDNA was individually cloned and confirmed by oligonucleotide sequencing of both strands. The creation and cDNA expression of wild-type Glu-158 and the common polymorphic form Lys-158 FMO3 MBPs have been described (Brunelle et al., 1997; Cashman et al., 1997; Treacy et al., 1998). Another polymorphic form of human FMO3 (i.e., Met-257) was created by oligonucleotide-directed mutagenesis and PCR in a similar fashion to that described before (Brunelle et al., 1997). The expression levels of Glu-158, Lys-158, and Met-257 FMO3 MBPs were similar to and agreed with the levels described previously (Brunelle et al., 1997). Human FMO1 MBP was also expressed at about the same level as the human FMO3 MBPs. The cDNA-expressed proteins were stable and retained FAD under the storage and incubation conditions. The amount of FAD per milligram of protein present in cDNA-expressed human FMO1 MBP and Glu-158, Lys-158, and Met-257 FMO3 MBPs after purification was similar.

Electrophoresis and Immunoblotting. Overproduction of the affinity-purified human FMO3 MBPs and truncation variants was shown by fractionation on 12% SDS-polyacrylamide gel electrophoresis. Immunoblots were done according to a previously described procedure with an affinity-purified rabbit polyclonal antibody that was directed against the wild-type human FMO3 MBP fusion protein (Guan et al., 1991). An affinity-purified goat polyclonal antibody that was directed against the human FMO1 fusion protein was used for human FMO1 MBP.

Enzyme Assays. Assay and analysis of human FMO3 MBP fusion protein and analysis of 5-DPT N-oxygenation activity for the variant enzymes were done by a procedure described previously (Brunelle et al., 1997). The analysis of N-oxygenation of 5-APT was done with an HPLC method essentially identical with the one described previously for a closely related compound (Lin et al., 1996). For K_m and V_max determinations with human FMO1 or FMO3 MBPs, six 5-APT or 5-DPT substrate concentrations were used ranging from 5 to 500 µM. TMA and tyramine N-oxygenation was determined with the radiometric assays described below.

FAD Determination. A 0.5-ml aliquot of protein was combined with an equivalent amount of cold acetonitrile, mixed, and centrifuged at 12,000g. The pellet was washed three times with cold acetonitrile. Fifty microliters of trifluoroacetic acid was added to the protein pellet, which was then mixed thoroughly and heated in a sealed tube at 55°C for at least 40 min. At the end of the reaction, 100 µl of cold acetonitrile was added, as was a sufficient amount of NH₄OH to neutralize the reaction. A portion of the mixture was injected directly onto a Hitachi HPLC system with a Rainin Microsorb MV C-18 reversed-phase column (Rainin, Emeryville, CA). An eluant of water (containing 0.05% HClO₄) for the first 20 min followed by CH₃CN (containing 0.05% HClO₄) thereafter was used to separate FAD from other minor impurities by HPLC. FAD eluted during the first phase of the gradient and was detected at 450 nm with a retention time of 9.2 min. The amount of FAD was determined from a comparison of peak heights from a standard curve of FAD. The correlation coefficient for these analyses was generally r² = 0.99.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Genotype Frequencies. The genotype frequencies for the two prevalent human FMO3 polymorphisms (i.e., alleles occurring at a frequency greater than 1%), E158K and V257M, in healthy populations from Quebec, Canada, and Victoria, Australia, are listed in Table 2. On the basis of statistical analysis, the distribution of the two codon 158 polymorphic variants were found to be similar in Canadian Francophones and Anglophones and in the Australian population. Although the numbers were small, the methionine variant involving codon 257 appeared at higher frequency in the Australian population than in the Canadian population.

Substrate N-Oxygenation. Previous studies showed that nontransformed host bacteria or bacteria transformed with pMAL alone did not contain any detectable human FMO activity when grown in the presence or absence of isopropyl -D-thiogalactoside. Previously, we showed that the relative activity of human FMO3 MBP and some variants exhibited varying degrees of TMA N-oxygenation activity (Cashman et al., 1997; Treacy et al., 1998). To examine this point more carefully and to quantify differences among the cDNA-expressed alleles, K_m and V_max values were obtained from double reciprocal plots of velocity versus substrate concentration. For the substrates examined, the formation of tertiary amine N-oxide, hydroxylamine, or oxime metabolite was a linear function of protein concentration and with incubation time for at least 10 min. As shown by the kinetic constants listed in Table 3, wild-type human FMO3 MBP efficiently N-oxygenated 5-DPT, TMA, and tyramine. Activities for human FMO1 MBP (i.e., a fetal hepatic FMO isoform that is expressed in adult kidney and intestine but not in adult liver) were evaluated in a similar kinetic fashion. The kinetic constants are listed in Table 3 for comparison. Human FMO1 MBP N-oxygenated 5-DPT with a very low K_m . In contrast, TMA N-oxygenation was significant, but in agreement with a previous study (Lang et al., 1998), the concentration of substrate required for half-maximal activity of FMO1 MBP was 15-fold greater than that of human FMO3 MBP. The K_m for human FMO1 MBP N-oxygenation of TMA compared favorably with that of TMA N-oxygenation for pig FMO1 (Cashman, 1995) or human FMO1 (Lang et al., 1998).

1

1

1

1

1

1

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Human FMO3 N-oxygenates nitrogen-containing chemicals and drugs to form more polar materials that are readily excreted. Recently, we reported mutations in the human FMO3 gene that segregate with and are specific for diminished capacity to N-oxygenate TMA (Cashman et al., 1997; Treacy et al., 1998). In humans, decreased TMA N-oxygenation results in a condition called trimethylaminuria. For individuals with deficient human FMO3 N-oxygenase activity, decreased clearance of nucleophilic hetereoatom-containing endogenous, dietary, or medicinal chemicals may increase the possibility of adverse reactions.

We previously reported that the enzyme encoded by the human Glu-158 FMO3 polymorphic allele is more active for the tertiary amine substrates TMA and 5-DPT than the enzyme encoded by the less prevalent human Lys-158 FMO3 allele (Brunelle et al., 1997; Treacy et al., 1998). Others have confirmed this observation by using TMA as a substrate (Dolphin et al., 1997). We now report another polymorphism. The presence of these prevalent FMO3 polymorphisms implies that some individuals may be more susceptible to the effects of environmental chemicals and to idiosyncratic drug reactions than others. Thus, if a drug or chemical is dependent on human FMO3 for detoxication and an individual possesses an impaired polymorphic form of FMO3, then the drug or chemical may produce an exaggerated clinical response and this could lead to adverse reactions.

Human FMO1 is not functionally active in adult human liver (Cashman, 1995). Based on the kinetic parameters observed herein, human FMO1 does not appear to make a significant contribution to the metabolism of TMA. Our results suggest that human FMO1 probably does not "rescue" an individual who is deficient in FMO3 from suffering the consequences of trimethylaminuria. Human FMO1 does not significantly N-oxygenate primary amines such as tyramine or 5-APT but can participate in the N-oxygenation of chemicals or drugs containing the tertiary amine functionality.

The substrate-binding domain of the human FMO3 enzyme is not yet known. The structural consequence of the methionine for valine substitution at codon 257 has not yet been elaborated. From our in vitro data, the codon 257 polymorphism appears to show substantial differences in N-oxygenation kinetics for the biogenic amine substrate tyramine and the dietary amine TMA. This may have clinical consequences. For example, consumption of large quantities of cruciferous vegetables that contain substantial quantities of indolemethyl glucosinolates has been shown to inhibit human FMO3 in vivo and cause elevated levels of urinary TMA (Cashman et al., 1999). This effect may be more apparent in individuals that are homozygous or heterozygous for human FMO3 polymorphisms.

The maintenance of the prevalent codon 158 polymorphism in the white populations examined may not result solely from "genetic drift" but perhaps as a consequence of "molecular drive," whereby particular polymorphisms with selective advantages persist, for example, to combat exposure to plant toxins in particular geographic regions (Nebert, 1997). The distribution of the two codon 158 alleles is almost in equilibrium in these populations, suggesting that this may be an older polymorphism that is now balanced (Treacy et al., 1996). Human V257M FMO3 is less prevalent and may represent a founder effect.

Variation in human drug metabolism by genetic polymorphisms may increase the risk for acquiring exposure-related disease, including cancer, with important public health consequences (Wolf, 1990; Kalow, 1991; Motulsky, 1991; Vesell, 1991; Wolf et al., 1992; Daly et al., 1994; Smith et al., 1994). Although human FMOs generally convert lipophilic heteroatom-containing compounds to polar, readily excreted oxygenated metabolites that possess decreased toxic potential, they may also catalyze the N-oxygenation of a wide array of xenobiotics such as plant toxins, organic chemicals, and aromatic amines associated with carcinogenesis (Ziegler et al., 1988).

Consistent with the fact that human drug-metabolizing enzymes have endogenous substrates and are prevalent not as neutral balanced polymorphisms but for their selective advantages, we have previously shown that human FMO3 metabolizes biogenic amines such as tyramine and phenethylamine, resulting in formation of their oxime metabolites. Formation of oxime metabolites generally terminates the pharmacological activity of the parent amine (Lin and Cashman, 1997a,b). Herein, we show that the methionine variant at codon 257 of human FMO3 shows decreased N-oxygenation for the substrate tyramine. Tyramine is an indirectly acting sympathomimetic that exerts its pressor effect through amine uptake into the sympathetic nervous system with release of norepinephrine from synaptic vesicles. It is thus possible that human FMO3 polymorphisms affecting tyramine or other biogenic amine metabolism may predispose humans to variable tolerance to tyramine or other biogenic amine-containing foods and the associated symptoms (Reddy and Hayes, 1989; Stratton et al., 1991).

We have previously shown that human FMO3-null alleles (i.e., Ile-66 and Leu-153) are rare and cause a severe trimethylaminuria phenotype with abrogation of N-oxygenation of human FMO3 substrates such as TMA and tyramine in vitro (Cashman et al., 1997; Treacy et al., 1998). On the basis of the in vitro data described herein, the human FMO3 common polymorphisms have been shown to have milder effects on N-oxygenation activity. The polymorphisms are of higher frequency, however, and thus are more likely to show specific associations with complex diseases in populations.

In conclusion, we present evidence of prevalent pharmacogenetic polymorphisms of the FMO3 gene in two white populations. Individuals with variants at the human FMO3 locus may have predisposition to spontaneous disease associated with common environmental exposures including pro-carcinogens. In addition, individuals with abnormal or defective human FMO3 function may show diminished capacity to metabolize environmental chemicals and may show exaggerated clinical effects and idiosyncratic drug reactions.