Electro-catalysis by immobilised human flavin-containing monooxygenase isoform 3 (hFMO3)

Bioelectrochemical platform with human monooxygenases: FMO1 and CYP3A4 tandem reactions with phorate

It is highly advantageous to devise an in vitro platform that can predict the complexity of an in vivo system. The first step of this process is the identification of a xenobiotic whose monooxygenation is carried out by two sequential enzymatic reactions. Pesticides are a good model for this type of tandem reactions since in specific cases they are initially metabolised by human flavin-containing monooxygenase 1 (hFMO1), followed by cytochrome P450 (CYP). To assess the feasibility of such an in vitro platform, hFMO1 is immobilised on glassy carbon electrodes modified with graphene oxide (GO) and cationic surfactant didecyldimethylammonium bromide (DDAB). UV-vis, contact angle and AFM measurements support the effective decoration of the GO sheets by DDAB which appear as 3 nm thick structures. hFMO1 activity on the bioelectrode versus three pesticides; fenthion, methiocarb and phorate, lead to the expected sulfoxide products with K_M values of 29.5 ± 5.1, 38.4 ± 7.5, 29.6 ± 4.1 µM, respectively. Moreover, phorate is subsequently tested in a tandem system with hFMO1 and CYP3A4 resulting in both phorate sulfoxide as well as phoratoxon sulfoxide. The data demonstrate the feasibility of using bioelectrochemical platforms to mimic the complex metabolic reactions of xenobiotics within the human body.

N- and S-oxygenation activity of truncated human flavin-containing monooxygenase 3 and its common polymorphic variants

Human flavin-containing monooxygenase 3 (FMO3) is a membrane-bound, phase I drug metabolizing enzyme. It is highly polymorphic with some of its variants demonstrating differences in rates of turnover of its substrates: xenobiotics including drugs as well as dietary compounds. In order to measure its in vitro activity and compare any differences between the wild type enzyme and its polymorphic variants, we undertook a systematic study using different engineered proteins, heterologously expressed in bacteria, purified and catalytically characterized with 3 different substrates. These included the full-length as well as the more soluble C-terminal truncated versions of the common polymorphic variants (E158K, V257M and E308G) of FMO3 in addition to the full-length and truncated wild-type proteins. In vitro activity assays were performed with benzydamine, tamoxifen and sulindac sulfide, whose products were measured by HPLC. Differences in catalytic properties between the wild-type FMO3 and its common polymorphic variants were similar to those observed with the truncated, more soluble versions of the enzymes. Interestingly, the truncated enzymes were better catalysts than the full-length proteins. The data obtained point to the feasibility of using the more soluble forms of this enzyme for in vitro drug assays as well as future biotechnological applications possibly in high throughput systems such as bioelectrochemical platforms and biosensors.

Production of drug metabolites by human FMO3 in Escherichia coli

Background

In the course of drug discovery and development process, sufficient reference standards of drug metabolites are required, especially for preclinical/clinical or new therapeutic drugs. Whole-cell synthesis of drug metabolites is of great interest due to its low cost, low environmental impact and specificity of the enzymatic reaction compared to chemical synthesis. Here, Escherichia coli (E. coli) JM109 cells over-expressing the recombinant human FMO3 (flavin-containing monooxygenase isoform 3) were used for the conversions of clomiphene, dasatinib, GSK5182 and tozasertib to their corresponding N-oxide metabolites.

Results

The effects of NADPH regeneration, organic solvents as well as C-terminal truncations of human FMO3 were investigated. Under the optimized conditions, in excess of 200 mg/L of N-oxide metabolite of each of the four drugs could be produced by whole-cell catalysis within 24 h. Of these, more than 90% yield conversions were obtained for the N-oxidation of clomiphene and dasatinib. In addition, FMO3 shows high regio-selectivity in metabolizing GSK5182 where only the (Z) isomer is monooxygenated.

Conclusions

The study shows the successful use of human FMO3-based whole-cell as a biocatalyst for the efficient synthesis of drug metabolites including regio-selective reactions involving GSK5182, a new candidate against type 2 diabetes mellitus.

Capillary electrophoresis integrated immobilized enzyme reactor for kinetic and inhibition assays of β-secretase as the Alzheimer's disease drug target

Capillary electrophoresis integrated immobilized enzyme reactors are becoming an increasingly popular alternative for enzyme kinetic and inhibition assays thanks to their unique set of features including cost effectiveness, repeated use of the enzyme, minuscule sample consumption, rapid analysis time and easy automation. In this work we present the development and application of a capillary electrophoresis integrated immobilized enzyme reactor based on magnetic particles for kinetic and inhibition studies of β-secretase, a key enzyme in the development of Alzheimer's disease and a promising drug target. We document the optimization of the immobilization procedure, characterization of immobilized β-secretase, optimization of a mutually compatible incubation protocol and separation method as well as the production of the capillary electrophoresis integrated immobilized enzyme reactor. The applicability of the capillary electrophoresis integrated immobilized enzyme reactor was demonstrated by kinetic assay with an unlabelled substrate and by inhibition assays using three structurally different reference inhibitors. The resulting kinetic and inhibition parameters clearly support the applicability of the herein presented method as well as document the fundamental phenomena which need to be taken in account when comparing the results to other methods.

An improved design to capture magnetic microparticles for capillary electrophoresis based immobilized microenzyme reactors

In this paper, we demonstrate the effectiveness of a new 3D printed magnet holder that enables capture of magnetic microparticles in commercially available capillary electrophoresis equipment with a liquid or air based coolant system. The design as well as the method to capture magnetic microparticles inside the capillary are discussed. This setup was tested at temperature and pH values suitable for performing enzymatic reactions. To demonstrate its applicability in CE- immobilized microenzyme reactors (IMER) development, human flavin-containing monooxygenase 3 and bovine serum albumin were immobilized on amino functionalized magnetic microparticles using glutaraldehyde. These microparticles were subsequently used to perform in-line capillary electrophoresis with clozapine as a model substrate. This setup could be used further to establish CE-IMERs of other drug metabolic enzymes in a commercially available liquid based capillary coolant system. The CE-IMER setup was successful, although a subsequent decrease in enzyme activity was observed on repeated runs.

Inactivation mechanism of N61S mutant of human FMO3 towards trimethylamine

Human flavin-containing monooxygenase 3 (hFMO3) catalyses the oxygenation of a wide variety of compounds including drugs as well as dietary compounds. It is the major hepatic enzyme involved in the production of the N-oxide of trimethylamine (TMAO) and clinical studies have uncovered a striking correlation between plasma TMAO concentration and cardiovascular disease. Certain mutations within the hFMO3 gene cause defective trimethylamine (TMA) N-oxygenation leading to trimethylaminuria (TMAU) also known as fish-odour syndrome. In this paper, the inactivation mechanism of a TMAU-causing polymorphic variant, N61S, is investigated. Transient kinetic experiments show that this variant has a > 170-fold lower NADPH binding affinity than the wild type. Thermodynamic and spectroscopic experiments reveal that the poor NADP⁺ binding affinity accelerates the C4a-hydroperoxyFAD intermediate decay, responsible for an unfavourable oxygen transfer to the substrate. Steady-state kinetic experiments show significantly decreased N61S catalytic activity towards other substrates; methimazole, benzydamine and tamoxifen. The in vitro data are corroborated by in silico data where compared to the wild type enzyme, a hydrogen bond required for the stabilisation of the flavin intermediate is lacking. Taken together, the data presented reveal the molecular basis for the loss of function observed in N61S mutant.

Evaluation of immobilized hFMO3 on magnetic nanoparticles by capillary zone electrophoresis

Aim: In this paper, we developed a method to immobilize human flavin-containing monooxygenase-3 (hFMO3) using glutaraldehyde as a cross-linker onto amino-functionalized magnetic nanoparticles. Materials & methods: All the analyses were done using capillary electrophoresis coupled with a diode array detector using clozapine as a substrate. Results: The apparent Km with clozapine as substrate and inhibition of hFMO3 by methimazole were explored for immobilized hFMO3 and were found to be comparable to literature values. The immobilized enzyme could be used three-times continuously at 37°C with no loss in enzyme activity. Conclusion: A method to immobilize hFMO3 on magnetic nanoparticles has been described and evaluated in terms of enzyme activity, inhibition, pH stability and reusability.

Effect of human flavin-containing monooxygenase 3 polymorphism on the metabolism of aurora kinase inhibitors

Aurora kinases were recently identified as a potential target in anticancer therapy and, amongst their available inhibitors, Tozasertib (VX-680) and Danusertib (PHA-739358) have been indicated as possible substrates of human flavin-containing monooxygenase 3 (hFMO3). Here we report the in vitro rate of oxidation of these drugs by wild-type hFMO3 and its polymorphic variant V257M. The conversion of Tozasertib and Danusertib to their corresponding metabolites, identified by LC-MS, by the purified wild-type and V257M hFMO3 show significant differences. In the case of Tozasertib, the V257M variant shows a catalytic efficiency, expressed as k_cat/K_m, similar to the wild-type: 0.39 ± 0.06 min⁻¹μM⁻¹ for V257M compared to 0.33 ± 0.04 min⁻¹μM⁻¹ for the wild type. On the other hand, in the case of Danusertib, V257M shows a 3.4× decrease in catalytic efficiency with k_cat/K_m values of 0.05 ± 0.01 min⁻¹μM⁻¹ for V257M and 0.17 ± 0.03 min⁻¹μM⁻¹ for the wild type. These data reveal how a simple V257M substitution ascribed to a single nucleotide polymorphism affects the N-oxidation of relevant anticancer drugs, with important outcome in their therapeutic effects. These findings demonstrate that codon 257 is important for activity of the hFMO3 gene and the codon change V to M has an effect on the catalytic efficiency of this enzyme.

Engineering Macaca fascicularis cytochrome P450 2C20 to reduce animal testing for new drugs

In order to develop in vitro methods as an alternative to P450 animal testing in the drug discovery process, two main requisites are necessary: 1) gathering of data on animal homologues of the human P450 enzymes, currently very limited, and 2) bypassing the requirement for both the P450 reductase and the expensive cofactor NADPH. In this work, P450 2C20 from Macaca fascicularis, homologue of the human P450 2C8 has been taken as a model system to develop such an alternative in vitro method by two different approaches. In the first approach called "molecular Lego", a soluble self-sufficient chimera was generated by fusing the P450 2C20 domain with the reductase domain of cytochrome P450 BM3 from Bacillus megaterium (P450 2C20/BMR). In the second approach, the need for the redox partner and also NADPH were both obviated by the direct immobilization of the P450 2C20 on glassy carbon and gold electrodes. Both systems were then compared to those obtained from the reconstituted P450 2C20 monooxygenase in presence of the human P450 reductase and NADPH using paclitaxel and amodiaquine, two typical drug substrates of the human P450 2C8. The K(M) values calculated for the 2C20 and 2C20/BMR in solution and for 2C20 immobilized on electrodes modified with gold nanoparticles were 1.9 ± 0.2, 5.9 ± 2.3, 3.0 ± 0.5 μM for paclitaxel and 1.2 ± 0.2, 1.6±0.2 and 1.4 ± 0.2 μM for amodiaquine, respectively. The data obtained not only show that the engineering of M. fascicularis did not affect its catalytic properties but also are consistent with K(M) values measured for the microsomal human P450 2C8 and therefore show the feasibility of developing alternative in vitro animal tests.