M. Shimizu, N. Koibuchi, A. Mizugaki et al.
Drug Metabolism and Pharmacokinetics 38 (2021) 100387
concentrations of trimethylamine N-oxide in blood and athero-
sclerotic cardiovascular disease [14].
Nexis GC-2030 equipped with a headspace sampler HS-20 and a
flame ionization detector (Shimadzu, Kyoto, Japan) using a capitally
column (Inertcap for amines, 0.32 mm i.d. x 60 m, GL Sciences,
Tokyo, Japan) under a carrier helium gas (at a flow rate of 20 mL/
min). Genotyping of 71 subjects with impaired phenotypes (around
7% of the new recruits) was done by sequencing analysis of both
strands of the polymerase chain reaction products from eight
coding FMO3 exons (2e9) and some intronic, flanking, and 30-un-
translated regions of FMO3 (a total of 2.9 kb) in DNA prepared from
buccal cells, as described previously [15e19]. The human FMO3
reference used in the current study was the complete gene
sequence described in GenBank (accession number NC_000001.10
and AL021026); for example, 171086469, and g. 30386 for FMO3
p.(Pro496Ser), respectively. In the current study, seven subjects
with FMO3 metabolic activity toward trimethylamine N-oxygena-
tion <90% of the wild type [15e19] were identified among the 979
newly recruited subjects.
More than 30 different mutations and around 40 poly-
morphisms of FMO3 have been reported to date [5,13]. Many of the
impaired polymorphic FMO3 variants associated with the meta-
bolic disorder trimethylaminuria were determined [13]. Genetic
testing in combination with traditional urinary phenotyping assays
is a useful approach to understanding the molecular basis of the
condition and to detect heterozygous carriers. We previously re-
ported phenotypeegene relationship analyses in 428 [15], 787 [16],
171 [17], 640 [18], and 164 [19] Japanese subjects using traditional
assays. Additionally, the Tohoku Medical Megabank Project brings
together population genomics, medical genetics, and prospective
cohort studies to support the establishment of personalized
healthcare in Japan [20]. We recently identified rare novel single-
nucleotide substitutions in human FMO3, e.g., p.(Gly191Cys) and
p.(Arg492Gln) variants with extremely low frequencies (<~0.1%),
among the whole-genome sequences of the approximately 3500
members of the Japanese population reference panel (3.5K JPN)
curated by the Tohoku Medical Megabank Organization [15].
However, the extent of overlap of these newly detected rare FMO3
alleles between our urinary phenotyped population and the Japa-
nese 3.5K JPN population reference panel is currently unknown.
Against this background, the purpose of the present study was
to further investigate both previously reported and novel FMO3
single-nucleotide variants in Japanese volunteers with self-
reported trimethylaminuria and those identified among the
updated 4.7K JPN whole-genome sequences of the enlarged Japa-
nese population reference panel (with 1200 new panel members)
[20]. Using these two methodologies in different Japanese cohorts,
i.e., in self-reported trimethylaminuria sufferers analyzed using
urinary phenotyping assays and in the panel of an extensive whole-
genome sequence database, the present follow-up study adds to
our knowledge of the common functional FMO3 polymorphisms in
populations. In this study, novel variants found in the seven pro-
bands with impaired trimethylamine metabolism, novel variants
identified in the 4.7K JPN database, and previously unanalyzed
variants were evaluated. We report herein three common FMO3
single-nucleotide or allele variants, namely FMO3 p.(Gly191Cys),
p.(Cys197Ter), and p.(Glu158Lys;Glu308Gly;Pro496Ser) variants,
having impaired trimethylamine N-oxygenation capacities in
combination with phenotyping and whole-genome sequence data
in different Japanese cohorts.
Previously reported and novel FMO3 single-nucleotide variants
were also identified in the whole-genome sequences of the Japa-
nese population reference panel (4.7K JPN) of the Tohoku Medical
Megabank Organization, now updated to include around 4700
subjects, as described previously [15,20].
2.2. Expression of recombinant FMO3 variant proteins and enzyme
assays
Variant and wild-type FMO3 cDNAs were prepared as previ-
ously described [15]. To produce the variant FMO3 proteins, site-
directed mutagenesis using a QuikChange II Site-Directed Muta-
genesis Kit (Stratagene, La Jolla, CA, USA) was carried out with the
designated primers shown in Table 1. The complete coding regions
of the mutagenized and wild-type FMO3 cDNAs were confirmed by
repeat sequencing of both strands with primers [15]. Modified and
wild-type FMO3 cDNAs were inserted into the pTrc99A expression
vector (Pharmacia Biotechnology, Milwaukee, WI, USA) and then
transformed into Escherichia coli strain JM109. Bacterial membrane
fractions expressing FMO3 were prepared from bacterial pellets
after centrifugation, as previously described [19]. The amounts of
recombinant FMO3 (0.010e0.27 nmol FMO3/mg bacterial protein,
Supplemental Table 1) were determined by immunochemical
quantification (Supplemental Fig. 1) using an anti-FMO3 antibody
(ab126790, Abcam, Cambridge, UK) and comparing the results with
those of a recombinant human FMO3 standard (Corning, Woburn,
MA, USA), as reported previously [15].
2. Materials and methods
Trimethylamine and benzydamine N-oxygenation rates were
evaluated as described previously [15e19]. Briefly,
a typical
2.1. Phenotype/genotype analyses of individuals and a survey of
genetic FMO3 variants in an extensive database
Table 1
Sequences of eight sets of primers used for mutagenesis of FMO3.
Between April 2018 and December 2020, 979 unrelated Japa-
nese subjects (367 male and 612 female subjects) responded to our
Internet article seeking volunteers with fish-like body odor and
were recruited. The current study follows up our previous in-
vestigations of the phenotypeegene relationship in a total of 2190
Japanese subjects as described in five previous reports covering 428
[15], 787 [16], 171 [17], 640 [18], and 164 [19] subjects. Of these
2190 subjects, 1040 were genotyped based on their phenotype.
Informed consent was obtained from the subjects and/or from their
parents. The current study was approved by the Ethics Committee
of Showa Pharmaceutical University. Phenotyping of the newly
recruited 979 potential trimethylaminuria sufferers was carried out
by evaluating the ratio of the amount of trimethylamine N-oxide in
urine compared with the total amount of trimethylamine and tri-
methylamine N-oxide as measured by gas chromatography [16,21].
Briefly, the head-space gas was subjected to gas chromatograph
Primer
Sequence
Met66Val-S
50-CAACTCTTCCAAAGAGGTGATGTGTTTCCCAG-30
50-CTGGGAAACACATCACCTCTTTGGAAGAGTTG-30
50-GGGTGATGAGCCAGGTCTGGGACAATG-30
50-CATTGTCCCAGACCTGGCTCATCACCC-30
50-CAAGCATGAAAACCATGGCTTGATGCC -30
50-GGCATCAAGCCATGGTTTTCATGCTTG -30
50-CAAGCATGAAAACTTTGGCTTGATGCC -30
50-GGCATCAAGCCAAAGTTTTCATGCTTG-30
50-GTGGCATTGTGTCCATAAAGCCTAACGTG-30
50-CACGTTAGGCTTTATGGACACAATGCCAC-30
50-GTAATAAAGGGAACTAGTACTTTGCCTTCTA-30
50-TAGAAGGCAAAGTACTAGTTCCCTTTATTAC-30
50-GCAAAAGCGAGACCACACAGACAGATTACAT-30
50-ATGTAATCTGTCTGTGTGGTCTCGCTTTTGC-30
50-CCGGTCGTTGAAATCCATGCAGACACG-30
50-CGTGTCTGCATGGATTTCAACGACCGG-30
Met66Val-AS
Arg223Gln-S
Arg223Gln-AS
Tyr269His-S
Tyr269His-AS
Tyr269Phe-S
Tyr269Phe-AS
Val299Ile-S
Val299Ile-AS
Cys397Ser-S
Cys397Ser-AS
Ile426Thr-S
Ile426Thr-AS
Pro496Ser-S
Pro496Ser-AS
2