Abstract
A polymorphism for the phase I drug-metabolizing enzyme, flavin-containing monooxygenase isoform 2 (FMO2), encoding either truncated inactive protein, FMO2X472 (FMO2.2A), or full-length active enzyme, FMO2Q472 (FMO2.1), is known and exhibits significant interethnic differences in allelic frequency. FMO2 is the major or sole FMO isoform expressed in the lung of most mammals, including nonhuman primates. To date, FMO2.1 has been found only in African-American and Hispanic populations, rendering individuals with this allele subject to drug metabolism that is potentially different from that of the general population. Approximately 26% of African-Americans (n = 180) possess the FMO2*1 allele. In preliminary studies, we initially estimated that 5% of Hispanics (n = 40) have the FMO2*1 allele, but access to large cohorts of individuals of defined national origin has allowed us to determine the occurrence among Mexican-American and Puerto Rican-American groups. We used allele-specific genotyping to detect FMO2*1 from 632 Hispanic individuals, including 280 individuals of Mexican origin and 327 individuals of Puerto Rican origin. Statistical analysis indicated that results from Mexican (five sample sources) and Puerto Rican (three sample sources) samples were consistent with the hypothesis of homogeneity within each group from different sources. Data were subsequently pooled across sources to test for evidence of a difference in occurrence of FMO2*1 between ethnic groups. There was strong evidence (p = 0.0066) that FMO2*1 is more common among Puerto Ricans (7%) than among individuals of Mexican descent (2%). The overall occurrence of FMO2*1 among Hispanics of all origins is estimated to be between 2 and 7%.
Flavin-containing monooxygenases (FMOs) (EC 1.14.13.8) oxygenate numerous drugs and xenobiotics containing soft nucleophiles, predominantly nitrogen and sulfur (Cashman, 1995, 2003; Ziegler, 2003). Eleven distinct FMO genes exist, encoding five active enzymes (FMO1-5), the remaining six representing functional pseudogenes (Lawton et al., 1994; Hines et al., 2002; Hernandez et al., 2004). FMO2 is expressed at high levels in the lung of most mammals, including nonhuman primates (Williams et al., 1985; Yueh et al., 1997; Karoly and Rose, 2001; Krueger et al., 2001). In humans, a genetic polymorphism has been described (Dolphin et al., 1998; Whestine et al., 2000; Krueger et al., 2002b; Furnes et al., 2003) in which all Caucasians and Asians genotyped to date have two FMO2*2A alleles. The FMO2*2A allele has a C>T transition at codon 472 (g.23,238C>T) resulting in a conversion of a glutamine to a stop codon (p.Q472X), yielding inactive protein (Dolphin et al., 1998). Twenty-six percent of African-Americans possess at least one FMO2*1 allele (p.Q472) encoding a full-length, functional enzyme (FMO2.1). Preliminary studies documented that the FMO2*1 allele was also present in individuals of Hispanic descent at a frequency of about 5% (Krueger et al., 2002b). In this study, we present genotyping data from 632 Hispanic individuals and conclude that a significant difference exists in FMO2*1 allelic frequency between individuals of Puerto Rican and Mexican descent. We have recently demonstrated that FMO2.1 functions in the bioactivation of thioureas such as ethylenethiourea and α-naphthylthiourea (Henderson et al., 2004b) and in the potential detoxication of organophosphate insecticides such as phorate and disulfoton (Henderson et al., 2004a). Thus, Hispanic and African-American individuals with the FMO2*1 allele may exhibit distinct outcomes compared with the general population following environmental or occupational exposures to such chemicals or in therapeutic response to drugs that are substrates for FMO2.1.
Materials and Methods
Genomic DNA used for genotyping experiments was obtained, without identifiers, from seven sources (Table 1). Hispanic samples were obtained as donor organs, or from living donors as blood or DNA. Ethnicity was self-reported by living donors. Criteria used included two generations of immediate ancestors from the same country (Zheng et al., 2003), three generations of immediate ancestors from the same country (Genomics Collaborative, Inc.), and unascertained extended ancestry (balance of sample sources). The appropriate institutional review boards approved studies before initiation of experiments. All donor samples were obtained with informed consent. DNA was isolated from donor samples as described elsewhere (Krueger et al., 2002b; Zheng et al., 2003).
Sequence variations in human FMO2 were described according to the recommendations of den Dunnen and Antonarakis (2001). GenBank accession number AL021026 was used as the genomic reference sequence for FMO2 numbering (g.1 to g.23,432 begins with nucleotide A from M1 and ends with nucleotide G from X536). Reference to the cDNA or the inferred protein sequence is indicated with a lowercase c or p prefix, respectively. Allelic and protein designations used (e.g., FMO2*1 and FMO2.1, respectively) are consistent with recommendations of the Human Nomenclature Committee (Shows et al., 1987) and were first applied to FMO2 variants by Whetstine et al., (2000).
Allele discrimination methods were designed that would yield allele-specific polymerase chain reaction (PCR) products detectable by electrophoresis. Allele-specific oligonucleotides hF2-15 and hF2-20, differing in length and discriminating FMO2*1 and FMO2*2A/*B, respectively, on the basis of the g.23,238C>T single nucleotide polymorphism (SNP), were combined in a single tube reaction with the common reverse primer hF2-18r, as already described (Krueger et al., 2002b), for 290 samples (Table 1, first six sample sources). Allele-specific oligonucleotide primer hF2-42 (5′-CGGACCCTGCAACTCCTtTC-3′) was substituted for hF2-15 for the remaining samples. The two alternative FMO2*1-specific forward primers are identical except for the incorporation of a single nucleotide mismatch (lower case t) designed into hF2-42 to enhance product specificity in late PCR cycles. PCR amplification conditions were identical to those already reported; however, the cycle number was dropped from 40 to 30. The 230-bp FMO2*1 PCR product was separated from the 245-bp FMO2*2A (or the 246-bp FMO2*2B) product by electrophoresis using 6% polyacrylamide gels (Invitrogen, Carlsbad, CA) and 1× Tris/borate/EDTA running buffer (89 mM Tris base, 89 mM boric acid, 2 mM EDTA).
A 198-bp region from AA471 to AA536 was sequenced (Krueger et al., 2002b) for all individuals identified with the FMO2*1 allele to confirm the allele and to identify individuals possibly exhibiting the c.1588_1589insT (g.23,412_23,413insT) variant, given that previous results suggested a high degree of linkage between the g.23,238C>T and g.23,412_23,413insT SNPs (Whestine et al., 2000). Additional random samples were also sequenced as part of methods verification. PCR products with the g.23,412_23,413insT base change were cloned into a TA vector (Invitrogen) and resequenced to determine whether the insertion occurred on the same allele as the g.23,238C>T SNP.
Statistical analyses were performed using StatXact version 5 (Statistical Solutions, Saugus, MA). The Fisher-Freeman-Halton test of homogeneity was utilized to determine whether or not samples of a single ethnic group from multiple sources could be pooled before comparing ethnic groups. Data that were consistent with equal proportions across sources were pooled within an ethnic group before comparing ethnic groups. Exact permutation p values generated by StatXact were used throughout. Binomial-based Blyth-Still-Casella confidence intervals for pooled proportions were calculated, as were Clopper-Pearson intervals. Genotyping data for the g.23,238C>T SNP were tested for agreement with expected values assuming the Hardy-Weinberg equilibrium and using a chi square test with 1 degree of freedom.
Results and Discussion
In total, we obtained and analyzed genomic DNA from 632 Hispanic individuals (Table 1). We were able to assemble large sample numbers (Tables 2 and 3) from the two most populous Hispanic groups, Mexicans and Puerto Ricans, in the United States. The remaining samples included 10 individuals representing six additional countries and 15 individuals (organ donors) of unknown Hispanic origin. Cubans, the third most populous Hispanic group in the United States, were not represented in our study.
The Mexican samples (n = 274) were from five sources (Table 2), and the Puerto Rican (n = 304) samples were from three sources. The test of homogeneity among samples from different sources indicated that the data were consistent with the null hypothesis that there were no differences in the proportion of individuals with FMO2*1 among Mexican (p = 0.3) and Puerto Rican (p = 0.57) samples, each from their different sources.
Since the data were consistent with equal proportions across sources, data within an ethnic group were pooled before comparing the two ethnic groups. Results of this comparison are shown in Table 3. There was strong evidence (p = 0.0066) of a difference in the proportion of individuals having FMO2*1 among Hispanics of either Mexican or Puerto Rican descent. Puerto Ricans had a higher occurrence of FMO2*1 than did Mexican individuals. Blyth-Still-Casella 95% confidence intervals did not overlap (Table 3), whereas Clopper-Pearson intervals (not shown) show the same pattern, but intervals just overlap. Since there were significant differences between ethnic groups and since the ethnic proportions in our test group do not reflect the overall distribution of Hispanics in the United States, no attempt was made to estimate an overall confidence interval for the occurrence of FMO2*1 in the complete study group. The Mexican, Puerto Rican, and combined samples were all consistent with Hardy-Weinberg equilibrium expectations (p > 0.37 for all three, chi square, 1 degree of freedom).
We also were interested in the occurrence of g.23,412_23,413insT (c.1588_1589insT). To date, this SNP has only been demonstrated in combination with g.23,238C>T (p.Q472X) and is designated as FMO2*2b (Whetstine et al., 2000; Krueger et al., 2002b; Furnes et al., 2003); however, if it were to occur as part of the p.Q472 haplotype, the resulting frameshift would have the potential of altering the last 6 amino acids and adding 23 amino acids to the FMO2 carboxy terminus. We sequenced genomic DNA (p.AA471-AA536) from the 32 individuals with the FMO2*1 allele (33, FMO2*1 alleles; 31, FMO2*2 alleles) and an additional 27 individuals homozygous for FMO2*2 (54 alleles). The g.23,412_23,413insT SNP was present in DNA from 7 of the 59 individuals (118 chromosomes) examined. Three of these individuals were heterozygous for FMO2*1/*2; therefore, this DNA was cloned to determine the phase of the two SNPs. In each case it was determined that the frameshift occurred in combination with the p.X472 haplotype (FMO2*2b). Thus, the frequency of the frameshift was 8.2% of the FMO2*2 allele (7 of 85 FMO2*2 alleles sequenced).
In addition to the frameshift variant, three SNPs encoding silent sequence changes to their respective third codon positions were identified from the unphased sequencing data. Two of the SNPs, c.1497C>G (g.23,321C>G) and c.1587T>C (g.23,411T>C), were rare, occurring in only one and two individuals, respectively. Neither of these alterations, which, respectively encode p.L499 and p.F529, has previously been documented. The third SNP, c.1476A>G (g.23,300A>G), which encodes p.K492, was recently documented in an African-American population (Furnes et al., 2003), but frequency information was not provided. This SNP may be an additional example of a variant with a frequency of occurrence that is dependent on the race or ethnicity of the group (Table 4). Whereas the c.1476A variant was present in 40% of 58 Mexican chromosomes analyzed, it was present in 87% of the 46 Puerto Rican chromosomes sequenced. The apparent ethnic difference remains even if the assumption is made that the c.1476G SNP is secondary to the FMO2*2 allele, in which case the incidence of c.1476G would rise from 60 to 62% for the FMO2*2 allele of Mexicans and from 13% to 27% incidence for the FMO2*2 allele of Puerto Ricans. Overall, the c.1476A variant was present in 61% of the 118 chromosomes sequenced.
Although we have demonstrated that active FMO2.1 is present in the lung from an FMO2*1/*2 individual (Krueger et al., 2002a), our hypothesis, as yet untested, is that individuals of African or Hispanic descent expressing the FMO2*1 allele will exhibit either increased or decreased sensitivity to toxicity upon exposure to environmental chemicals for which the lung is a major route of exposure and/or target organ. For example, exposure to ethylenethiourea or α-naphthylthiourea may result in enhanced pulmonary toxicity since the sulfenic acid metabolites formed from thioureas by FMO are themselves reactive and promote oxidative stress through glutathione oxidation (Krieter et al., 1984; Scott et al., 1990; Decker and Doerge, 1991; Smith and Crespi, 2002; Henderson et al., 2004b). In contrast, exposure to phorate or disulfoton through inhalation may result in less toxicity for these individuals since FMO is the predominant pulmonary phase I enzyme and the sole FMO metabolite is the sulfoxide, a considerably less toxic metabolite compared with the oxon metabolites produced by cytochrome P450 (Levi and Hodgson, 1988; Hodgson et al., 1998; Henderson et al., 2004a; Usmani et al., 2004). However, the extent of toxicity will depend on the regulation of FMO2 and the relative contribution by lung cytochromes P450, since cytochromes P450 can further metabolize the oxon metabolite to the more toxic sulfoxide oxon (Kinsler et al., 1988).
Acknowledgments
We thank Genomics Collaborative, Inc. and the Human Biomolecular Research Institute for sample donations and information. We also thank the New York Cancer Project, which made biological samples and information available in connection with the publication of this study. The New York Cancer Project is administered and funded by Academic Medicine Development Corporation Foundation, Inc.
Footnotes
-
This study was supported by United States Public Health Service Grants HL38650 and ES00210. Part of this study was presented at the 41st Annual Meeting of the Society of Toxicology, Nashville, Tennessee, March 2002, an abstract of which appeared in The Toxicologist 66(1-S),189. Part of this study was also presented at the Second International workshop on Trimethylaminuria, Washington, D.C., USA, March 2002.
-
doi:10.1124/dmd.104.001099.
-
ABBREVIATIONS: FMO, flavin-containing monooxygenase; PCR, polymerase chain reaction; SNP, single nucleotide polymorphism
- Received June 18, 2004.
- Accepted September 7, 2004.
- The American Society for Pharmacology and Experimental Therapeutics