Results and Discussion

Analysis of common marmoset’s genome sequence showed that FMO1–FMO4 and FMO6 genes were localized in the gene cluster; the FMO5 gene was localized outside this cluster, in marmoset chromosome 18 (Supplemental Fig. 1). Marmoset FMO genes had one-to-one orthologous relationships with human FMO genes, even though FMO gene cluster organization is different between marmosets and humans. We successfully isolated FMO1–FMO5 cDNA in the marmoset liver by RT-PCR. Marmoset FMO1–FMO5 contained open reading frames of 532–556 amino acid residues, respectively (Supplemental Fig. 2), and had functionally important regions (FAD- and NADPH-pyrophosphate binding sites), and the two characteristic FMO pentapeptides (EGLEP and FATGY). Marmoset FMO1–FMO5 shared high amino acid sequence identities (93%–95%) with human FMO counterparts (Supplemental Table 1) and were phylogenetically more closely clustered with the corresponding primate orthologs than other species (Supplemental Fig. 3). Interestingly, in cynomolgus monkeys, FMO6 is a functional enzyme that is widely expressed in the kidneys, heart, testes, uterus, and liver (Uno et al., 2013); although we have tried, we failed to clone FMO6 cDNA from marmoset tissues.

To investigate the tissue distribution of FMO1–FMO5 mRNAs and proteins in marmosets, real-time RT-PCR was performed to measure expression levels of FMO1–FMO5 in pooled brains, lungs, livers, kidneys, and small intestine. FMO3 mRNA was the most abundant in the liver and kidneys, followed by the lungs (Fig. 1A), whereas FMO1 mRNA was also expressed abundantly in the liver and kidneys but not as much as FMO3 mRNA, the same as results previously reported by de novo transcriptome analysis (Shimizu et al., 2014). Indeed, FMO3 protein (approximately 50 kDa) was detected immunologically with anti-human FMO3 antibodies in pooled marmoset livers and kidneys (Fig. 1B), with a small nonspecific unknown band. No immunoreactive bands in these tissue microsomes were seen with commercial anti-human FMO1 antibodies (results not shown). Similar to marmosets, FMO3 is postnatally expressed in the kidneys of rabbits and rats (Ripp et al., 1999). FMO1 is reportedly expressed postnatally in the liver of dogs (Lattard et al., 2002), rabbits (Shehin-Johnson et al., 1995), rats (Novick et al., 2009), and mice (Itoh et al., 1997). Human FMO1, FMO2, and FMO3 are predominantly expressed in the kidneys, lungs, and liver, respectively, whereas FMO4 and FMO5 are widely expressed in various tissues (Zhang and Cashman, 2006; Uno et al., 2013). In humans, FMO1 is expressed in the fetal liver, but its expression is rapidly extinguished after birth (Koukouritaki et al., 2002). Apparent sex differences in FMO3 protein expression levels in the liver (Ripp et al., 1999) were found in mice, but not in marmosets (Fig. 1A). Marmoset FMO2 and FMO5 mRNA were dominantly expressed in the lungs and liver, whereas FMO4 mRNA was expressed in the liver, kidneys, and small intestine at very low levels, similar to human and cynomolgus monkey FMO forms (Zhang and Cashman, 2006; Uno et al., 2013). These results suggested that tissue distribution of FMO3 and FMO1 was partially different between marmosets and humans.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Determination of FMO mRNA (A) and FMO protein (B and C) levels in marmoset tissues. (A) Expression levels of marmoset FMO1–FMO5 mRNAs in five marmoset tissues (each pool of six male and six female marmosets) were measured by real-time RT-PCR. Raw values of target gene expression were normalized with the 18S rRNA level. Each data point represents the average and S.D. of triplicate determinations from three representative experiments. (B and C) Pooled tissue microsomes from marmosets (20 μg/lane) (B) and recombinant marmoset FMO3 protein (0.1 pmol FMO3/lane), individual liver microsomes from marmosets (male, lanes 1–2; female, lanes 3–5) (20 μg/lane) (C) were analyzed by immunoblotting using anti-human FMO3 antibodies. Protein disulfide isomerase (PDI) expression was assessed as a loading control.

To assess the importance of FMO forms in drug oxidation in the marmoset liver, drug oxygenation activities by liver microsomes preheated for FMO inhibition were measured. Preheat-sensitive benzydamine N-oxygenation and sulindac sulfide S-oxygenation activities in marmoset liver and kidney microsomes was found (Supplemental Table 2). Recombinant FMO3-mediated benzydamine N-oxygenation activity and its V_max/K_m value were higher than those for recombinant FMO1, suggesting that FMO3 abundantly expressed in the liver plays a role in the N-oxygenation of benzydamine in the marmoset liver (Table 1), similar to the human liver (Taniguchi-Takizawa et al., 2015). Kinetic analyses also indicated that marmoset liver microsomes effectively catalyzed sulindac sulfide S-oxygenation (V_max/K_m, 0.12 ml/min per milligram protein) (Table 1), compared with those of humans (V_max/K_m, 0.02 ml/min per milligram protein), as previously reported (Yamazaki et al., 2014). Marmoset FMO3 was catalytically efficient (V_max/K_m, 0.80 ml/min per nanomole) for sulindac sulfide S-oxygenation and showed a low K_m value (38 μM), comparable to marmoset liver microsomes (23 μM), similar to human FMO3 (Yamazaki et al., 2014). Marmoset FMO1 showed low K_m (43 μM) and high V_max/K_m (0.53 ml/min per nanomole), compared with human FMO1 (K_m, 280 μM; V_max/K_m, 0.01 ml/min per nanomole). Considering this together with tissue distribution, FMO1 might account for the potential species differences in sulindac sulfide S-oxygenation rates between marmoset and human livers. These results indicated that benzydamine N-oxygenation and sulindac sulfide S-oxygenation in the marmoset liver were mainly catalyzed by FMO3.

In summary, marmoset FMO1–FMO5 had amino acid sequences that were highly identical (>93%) to human FMO1–FMO5, as well as a phylogenetically close relationship with human FMO1–FMO5. In contrast with humans, FMO3 and FMO1 mRNA was abundant in the marmoset liver and kidneys among five FMO forms. Recombinant marmoset FMO3 enzymes heterologously expressed in E. coli effectively metabolized typical human FMO substrates, benzydamine and sulindac sulfide. These results indicated that marmoset FMO1–FMO5 has sequences that are similar to those of humans but are partially different from humans in terms of tissue expression. Importantly, marmoset and human FMO3, a major hepatic FMO in both species, had similar enzymatic properties, suggesting the similarity of FMO-dependent drug metabolism for marmosets and humans.