Regioselectivity of cytochrome P-450 catalyzed hydroxylation of fluorobenzenes predicted by calculated frontier orbital substrate characteristics

Metalloporphyrin-Catalyzed Oxidation of 2-Methylnaphthalene to Vitamin K(3) and 6-Methyl-1,4-naphthoquinone by Potassium Monopersulfate in Aqueous Solution

The metalloporphyrin-catalyzed oxidation of 2-methynaphthalene (1) by potassium monopersulfate produced mainly two naphthoquinones: 2-methyl-1,4-naphthoquinone (2) (menadione or vitamin K(3)) and 6-methyl-1,4-naphthoquinone (3). In aqueous solution and at room temperature in the presence of 5 mol % of the water-soluble metalloporphyrins MnTPPS or FeTMPS, 2-methylnaphthalene was quantitatively oxidized to quinones 2 and 3. Based on experiments performed in(18)O-labeled water and according to the "redox tautomerism" mechanism previously described for such catalysts, the oxidation to quinones is proposed to be mainly due to a cytochrome P-450-type oxygenation reaction (oxygen atom transfer), rather than a peroxidase-type oxidation (electron transfer).

Microperoxidase/H2O2-catalyzed aromatic hydroxylation proceeds by a cytochrome-P-450-type oxygen-transfer reaction mechanism

The mechanism of aromatic hydroxylation of aniline and phenol derivatives in a H2O2-driven microperoxidase-8(MP8)-catalyzed reaction was investigated. It was shown that the reaction was not inhibited by the addition of scavengers of superoxide anion or hydroxyl radicals, which demonstrates that the reaction mechanism differs from that of the aromatic hydroxylation catalyzed by a horseradish peroxidase/ dihydroxyfumarate system. Additional experiments with 18O-labelled H2 18O2 demonstrated that the oxygen incorporated into aniline to give 4-aminophenol originates from H2O2. Furthermore, it was found that the addition of ascorbic acid efficiently blocks all peroxidase-type reactions that can be catalyzed by the MP8/H2O2 system, but does not inhibit the aromatic hydroxylation of aniline and phenol derivatives. Together, these observations exclude reaction mechanisms for the aromatic hydroxylation that proceed through peroxidase-type mechanisms in which the oxygen incorporated into the substrate originates from O2 or H2O. The mechanism instead seems to proceed by an initial attack of the high-valent iron-oxo intermediate of MP8 on the pi-electrons of the aromatic ring of the substrate leading to product formation by a cytochrome-P-450-type of sigma-O-addition or oxygen-rebound mechanism. This implies that MP8, which has a histidyl and not a cysteinate fifth axial ligand, is able to react by a cytochrome-P-450-like oxygen-transfer reaction mechanism.

Comparative MO-QSAR studies in various species including man

In the present study it is demonstrated that MO-QSARs (quantitative structure activity relationships based on calculated molecular orbital substrate characteristics) of cytochrome P450-catalysed biotransformation of benzene derivatives obtained in previous studies for Wistar rats, can be extrapolated to other species, including man. First, it was demonstrated that the regioselectivity of the in vivo aromatic hydroxylation of two fluorobenzene derivatives can be quantitatively predicted, on the basis of the calculated density distribution of the reactive pi-electrons in the aromatic ring of the fluorobenzene derivative, for all experimental animal species tested. Second, it was investigated whether the preferential site for in vitro aromatic hydroxylation of 3-fluoroaniline could be predicted on the basis of the same calculated parameter. This was done because extrapolation to human systems requires in vitro instead of in vivo experiments. The results obtained indicated that the variation in the regioselectivity of the aromatic hydroxylation of 3-fluoroaniline by liver microsomes from different species, including man, was only a few percent, and was mainly directed by calculated chemical reactivity characteristics of the 3-fluoraniline substrate. Finally, possibilities for the extrapolation from rat to other species, of the MO-QSAR for the rate of in vitro C4 hydroxylation of a series of aniline derivatives converted in an iodosobenzene-supported microsomal cytochrome P450 system, were investigated. Experiments with liver microsomes from rats, mice, rabbit and man resulted in clear MO-QSARs with correlation coefficients for the relationship between the 1n k(cat) and the E(HOMO) of the aniline substrates that were > or = to 0.97 in all cases. Thus, the results of the present study clearly demonstrate that MO-QSARs previously described for Wistar rats can be extrapolated to mice, rabbit, guinea pig and even to man. Regioselectivities obtained and QSAR lines for the rate of conversion plotted against calculated E(HOMO) values of the aniline derivatives are similar for the various species investigated. Altogether, these results strongly support the conclusion that the conversion of the relatively small benzene derivatives in the relatively large and aspecific active sites of the mammalian cytochromes P450, even when derived from various species, are mainly dependent on chemical reactivity parameters of the substrates. Therefore, the results of the present study support the hypothesis that MO-based QSARs obtained in rat for the cytochrome P450 catalysed aromatic hydroxylation of benzene derivatives can provide a basis for prediction of biotransformation pathways in different species, including man.

1H and 19F-nmr spectroscopic studies on the metabolism and urinary excretion of mono- and disubstituted phenols in the rat

1. 1H and 19F-nmr spectroscopy was used to investigate quantitatively the urinary excretion of the metabolites of 15 substituted phenols in the rat. The compounds studied were: 2-, 3-, and 4-fluorophenols; 2-, 3-, and 4-trifluoromethylphenol; 2,4-, 2,6- and 3,4-difluorophenol; 2-fluoro-5-trifluoromethylphenol, 3-fluoro-5-trifluoromethylphenol, 2-trifluoromethyl-4-fluorophenol; 3-chloro-4-fluorophenol, 3-fluoro-4-chlorophenol, and 3-methyl-4-fluorophenol. All compounds were dosed to the Sprague-Dawley rat (10 mg/kg i.p.) and urine was collected over the periods 0-8, 8-24 and 24-48 h post-dosing and analyzed using nmr spectroscopy. 2. The compounds were excreted in the urine mainly as glucuronide or sulphate conjugates or as the unchanged parent compound. There was considerable variation in the urinary excretion of the compounds over 48 h ranging from 22.1 to 93.6% of the dose. There was no apparent relationship between the molecular weight of compounds or their metabolites and the percentage molar recovery of each in the urine. 3. Ortho-substituted phenols in general showed a greater propensity for glucuronidation than did either meta- or para-substituted compounds, irrespective of the substituent group. The molar glucuronide-to-sulphate ratio for ortho-substituted compounds was found to be 2.2 +/- 0.9 whereas the ratio for both meta- and para-substituted compounds was 0.8 +/- 0.2 (p < 0.0001). 4. There were characteristic substituent effects of phenolic glucuronidation or sulphation on the 19F-nmr chemical shifts for both F- and CF3-substituted phemols and these substituent effects were a useful aid to metabolite signal assignment. 5. These studies show that nmr spectroscopy provides a rapid and convenient approach to the construction of metabolic databases of simple xenobiotics for the investigation of structure-metabolism relationships.

Quantitative structure-metabolism relationships for substituted benzoic acids in the rabbit: prediction of urinary excretion of glycine and glucuronide conjugates

1. Quantitative relationships between molecular physicochemical properties of 24 substituted benzoic acids and their metabolic fate in the rabbit have been investigated using computational chemistry and multivariate statistical methods. 2. A total of 34 molecular properties were calculated for each compound using computational chemistry and were related statistically to the % molar recovery of glycine and glucuronide conjugates excreted in the urine of rabbits for the 24 compounds. 3. Compounds were successfully classified according to their dominant metabolic fate based on urinary excretion data, where stepwise linear regression analysis of the theoretical property data achieved good predictive fits for observed versus predicted % molar urinary recovery as glucuronide conjugates (r2 = 0.79) and % molar urinary recovery as glycine conjugates (r2 = 0.66). 4. Quantitative prediction of the urinary excretion of glucuronide and glycine conjugates of the parent compounds was achieved based on a statistical model using calculated molecular physicochemical parameters. Interpretation of the molecular properties, which are important for predicting the metabolic class, should give new insights into basic mechanisms of drug metabolism and underlying molecular recognition events that determine disposition and metabolism.

Interaction between the substrate and the high-valent-iron-oxo porphyrin cofactor as a possible factor influencing the regioselectivity of cytochrome P450 catalysed aromatic ring hydroxylation of 3-fluoro(methyl)anilines

In the present study the in vitro and in vivo aromatic ring hydroxylation of a series of amino and/or methyl containing fluorobenzenes, i.e. 3-fluoro(methyl)anilines, was investigated and compared to the calculated density distribution of the reactive frontier pi-electrons of the aromatic substrate. This was done (1) to study to what extent the regioselectivity of the aromatic ring hydroxylation of the 3-fluoro(methyl)anilines could be predicted on the basis of the calculated chemical reactivity, as was previously observed for a series of fluorinated benzenes and monofluoroanilines, and (2) to investigate which factors contribute to possible deviations from the predictions on the basis of the calculated chemical reactivity. Results obtained show that the in vitro and in vivo aromatic ring hydroxylation of the series of 3-fluoro(methyl)anilines correlates qualitatively with the calculated frontier orbital density distribution for electrophilic attack by the cytochrome P450(FeO)3+ species. These results indicate that the HOMO/HOMO-1 frontier orbital densities, i.e. the chemical reactivity of the carbon centres for an electrophilic attack, predict the preferential as well as the non-reactive sites for cytochrome P450 catalysed aromatic ring hydroxylation of the tested model compounds. The absolute values, however, deviated in a systematic way; C4 para hydroxylation being observed to a higher extent than expected on the basis of chemical reactivity and C2/C6 ortho hydroxylation being observed to a lower extent than expected. Additional experiments were performed using different microsomal preparations and microperoxidase-8. The latter is a mini-heme protein of eight amino acids without a substrate binding site. In incubations of the model compounds with different types of microsomal preparations, as well as with MP-8 and purified reconstructed cytochrome P4502B1, similar systematic deviations between the predicted and observed regioselectivity of aromatic hydroxylation were observed. These results show that the regioselectivity of aromatic ring hydroxylation of the 3-fluoro(methyl)anilines cannot be predominantly ascribed to an interaction between the substrate and the substrate binding site of the cytochromes P450 dictating a specific stereoselective positioning of the substrate in the active site. More likely, the systematic deviations between the observed and predicted regioselectivity of hydroxylation of the tested model substrates should be ascribed to an (orienting) interaction between the substrate and the activated cytochrome P450(FeO)3+ cofactor.

Prediction of urinary sulphate and glucuronide conjugate excretion for substituted phenols in the rat using quantitative structure-metabolism relationships

1. The quantitative urinary excretion of the sulphate and glucuronide metabolites of 15 substituted phenols dosed to rat has been determined using high resolution 19F-nmr spectroscopy. 2. The urinary metabolic fate of each of the compounds was related to a series of calculated physicochemical properties for each compound to produce quantitative structure-metabolism relationships (QSMRs). Using these calculated molecular properties it was possible to predict the urinary recovery of xenobiotic material as a percentage of the administered dose, to classify the compounds according to their 'dominant' metabolite pattern and to predict quantitatively the proportions of glucuronide and sulphate conjugates in the urine by the use of multiple linear regression. 3. The quantitative predictions were tested by cross-validation and good prediction of total xenobiotic urinary recovery as a percentage of the administered dose was achieved based on an equation involving the electrophilic superdelocalizability at C4 (para to the hydroxyl function), the smallest principal ellipsoid axis dimension and the heat of formation. The largest moment of inertia and the electrophilic superdelocalizability at C3 were found to be the most significant factors for the prediction of the percentage glucuronide in the urine, and the urinary excretion of sulphate conjugates as a percentage of total urinary recovery was negatively correlated with the glucuronide excretion as little parent compound was excreted.

1H NMR T1 relaxation rate study on substrate orientation of fluoromethylanilines in the active sites of microsomal and purified cytochromes P450 1A1 and 2B1

The present study describes 1H NMR T1 relaxation rate studies on fluoromethylanilines bound to the active sites of microsomal and purified cytochromes P450 1A1 and 2B1. From the data obtained, insights into the average orientation of the substrates with respect to the paramagnetic Fe3+ centre in the cytochromes P450 could be derived. Particular attention was paid to a possible extra relaxation pathway for methyl protons compared to the aromatic protons, due to the rotational motion of the CH3 around the sigma-C-CH3 bond. However, this effect appeared to be minimal and to result in at most a few percent underestimation of the actual distance of the methyl protons to the Fe3+ centre. Furthermore, the data obtained demonstrate that all aromatic protons are at about the same average distance from the paramagnetic centre. The results also demonstrate that the fluromethylanilines are bound to the active sites of cytochromes P450 1A1 and 2B1 in a similar way. A time-averaged orientation of the substrate with the Fe3+ above the aromatic ring, with the pi-orbitals of the aromatic ring and those of the porphyrin rings in a parallel position, providing possibilities for energetically favourable pi-pi interaction defines the orientation which best fits the results of the present study. Possibilities for a flip-flop rotation around an axis in the plane of the aromatic ring can be included in this picture, as such rotations would still result in a similar average distance of all aromatic protons to the Fe3+ paramagnetic centre. The results obtained also indicate that possible differences in metabolite patterns resulting from conversion of the fluoromethylanilines by different cytochromes P450, especially P450 1A1 and 2B1, are unlikely to be caused by a specific orientation of the substrate imposed by the substrate binding site of the enzyme.

Oxidative and non-oxidative metabolism of 4-iodoanisole by rat liver microsomes

1. The oxidative metabolism of 4-iodoanisole (1) by liver microsomes from beta-naphthoflavone-treated rats yields 4-iodophenol (2) 2-iodo-5-methoxyphenol (3), 2-methoxy-5-iodophenol (4), 4-methoxyphenol (5), and 3-methoxyphenol (6) in relative yields of 5:2:4:1:1 respectively. 2. [3 5-2H2]-1 was converted to the same five metabolites in the same proportions; formation of 2, 4 and 5 involved no loss of deuterium, but formation of 3 and 6 involved respectively 55 and 28% loss of one deuterium. 3. When metabolism of 1 was carried out in buffers containing D2O or H2(18)O, no incorporation of these isotopes into 2-6 could be detected. Nor was it possible to detect formation of iodinating intermediates derived from 1 by trapping with 2,6-dimethylphenol. 4. The P450-catalysed hydroxylative de-iodination of 1-5 and 6 is suggested to involve C-O bond formation via attack of the ferry moiety on the aromatic ring followed by reductive cleavage of the C-iodine bond, with electrons coming from P450 reductase.

Conversion of phenol derivatives to hydroxylated products by phenol hydroxylase from Trichosporon cutaneum. A comparison of regioselectivity and rate of conversion with calculated molecular orbital substrate characteristics

This study describes the regioselective hydroxylation and the rates of conversion of a series of fluorinated phenol derivatives by phenol hydroxylase from the yeast Trichosporon cutaneum. The natural logarithm of the kcat value for the conversion of the phenolic substrates correlates with the calculated energy of the reactive electrons in the highest occupied molecular orbital of the substrate (r = 0.85). This observation supports the hypothesis that at physiological pH (7.6) and 25 degrees C, in the absence of monovalent anions, the nucleophilic attack of the electrons in the highest occupied molecular orbital of the substrate on the C(4a)-hydroperoxyflavin enzyme intermediate is of major importance in determining the overall rate of catalysis. Results from 19F-NMR analysis of the incubation mixtures demonstrate for phenols with two identical ortho substituents, that the ortho position which becomes preferentially hydroxylated is the one with the highest density of the reactive electrons in the highest occupied molecular orbital. A halogen substituent at a meta position decreases the chances for hydroxylation at the adjacent ortho position further than expected on the basis of the calculated reactivity. This result indicates a contribution of a protein/substrate dipolar interaction, influencing the time-averaged orientation of the substrate with respect to the reactive C(4a)-hydroperoxyflavin intermediate.

Different metabolic pathways of 2,5-difluoronitrobenzene and 2,5-difluoroaminobenzene compared to molecular orbital substrate characteristics

The in vivo metabolite patterns of 2,5-difluoroaminobenzene and of its nitrobenzene analogue, 2,5-difluoronitrobenzene, were determined using 19F NMR analysis of urine samples. Results obtained demonstrate significant differences between the biotransformation patterns of these two analogues. For the aminobenzene, cytochrome P450 catalysed aromatic hydroxylation presents the main metabolic pathway. 2,5-Difluoronitrobenzene was predominantly metabolised through glutathione conjugation leading to excretion of 5-fluoro-2-(N-acetylcysteinyl)-nitrobenzene and fluoride anions, and, to a minor extent, through cytochrome P450 catalysed hydroxylation and nitroreduction. Pretreatment of the rats with various inducers of cytochrome P450 enzymes, known also to influence glutathione S-transferase enzyme patterns, followed by exposure to the 2,5-difluoroamino- or 2,5-difluoronitrobenzene, generally resulted in metabolite patterns that varied only to a small (< or = 12%) extent. Based on these results it was concluded that the biotransformation enzyme pattern is not the predominant factor in determining the metabolic route of these two model compounds. Additional in vitro microsomal and cytosolic incubations with 2,5-difluoroaminobenzene and 2,5-difluoronitrobenzene qualitatively confirmed the in vivo results. NADPH/oxygen supported microsomal cytochrome P450 catalysed hydroxylation was observed only for 2,5-difluoroaminobenzene whereas cytosolic GSH conjugation occurred only in incubations with 2,5-difluoronitrobenzene as the substrate. Outcomes from molecular orbital calculations provided a working hypothesis that can explain the difference in metabolic pathways of the nitro- and aminobenzene derivative on the basis of their chemical characteristics. This hypothesis states that the chances for a nitro- or aminobenzene derivative to enter either a cytochrome P450 or a glutathione conjugation pathway are determined by the relative energy levels of the frontier orbitals of the compounds. The aminobenzene derivative has relatively high energy molecular orbitals leading to an efficient reaction of its highest occupied molecular orbital (HOMO) with the singly occupied molecular orbital of the cytochrome P450 (FeO)3+ intermediate, but a low reactivity of its lowest unoccupied molecular orbital (LUMO) with the HOMO of glutathione. The nitrobenzene, on the other hand, has molecular orbitals of relatively low energy, explaining the efficient interaction, and, thus, reaction between its LUMO and the HOMO electrons of glutathione, but resulting in low reactivity with the SOMO electron of the cytochrome P450 (FeO)3+ reaction intermediate.

Catabolism of 4-hydroxybenzoate in Candida parapsilosis proceeds through initial oxidative decarboxylation by a FAD-dependent 4-hydroxybenzoate 1-hydroxylase

The first two steps in the catabolism of 4-hydroxybenzoate by the ascomycetous yeast Candida parapsilosis CBS604 were investigated. In contrast to the well-known bacterial pathways and to what was previously assumed, metabolism of 4-hydroxybenzoate in C. parapsilosis proceeds through initial oxidative decarboxylation to give 1,4-dihydroxybenzene. This reaction is catalyzed by a NAD(P)H and FAD-dependent 4-hydroxybenzoate 1-hydroxylase. Further metabolism of 1,4-dihydroxybenzene to the ring-fission substrate 1,2,4-trihydroxybenzene is catalyzed by a NADPH-specific FAD-dependent aromatic hydroxylase acting on phenolic compounds. 19F-NMR experiments with cell extracts and 2-fluoro-4-hydroxybenzoate as the model compound confirm this metabolic pathway and exclude the alternative pathway proceeding through initial 3-hydroxylation followed by oxidative decarboxylation in the second step.

Influence of the type of halogen substituent on in vivo and in vitro phase II metabolism of 2-fluoro-4-halophenol metabolites formed from 3-halo-fluorobenzenes

1. The influence of a change in the type of halogen substituent on phase II metabolism of 2-fluoro-4-halophenol metabolites formed from 3-halo-fluorobenzenes was studied in vivo and in vitro using 19F nmr and spectroscopic assays. 2. The ratio of sulphation to glucuronidation of 2-fluoro-4-halophenol metabolites formed from 3-halofluorobenzenes decreased from 48 to 13 to 6 when the halogen substituent varied from fluorine to chlorine to bromine. 3. When the 2-fluoro-4-halophenols themselves were administered to the rats, the ratio of sulphation to glucuronidation was not affected by the type of halogen substituent at C4 and at a constant value of 0.6, i.e. significantly lower. 4. Kinetic data for P450 catalysed hydroxylation of the 3-halo-fluorobenzenes and for sulphation and glucuronidation of their 2-fluoro-4-halophenol metabolites were obtained from in vitro microsomal and cytosolic incubations. These data demonstrate that the effects of varying the halogen substituent on phase II metabolism of the 2-fluoro-4-halophenol metabolites can be mainly ascribed to an apparently decreased Km for the glucuronidation of the 2-fluoro-4-halophenols with a change in the halo substituent from fluorine to chlorine to bromine. 5. Results from calculations on electronic and structural characteristics of the three 4-halo-2-fluorophenols demonstrate that the best explanation for the decrease in the apparent Km of the glucuronidation from 2,4-difluoro- to 4-chloro-2-fluoro- to 4-bromo-2-fluorophenol might be an increase in the hydrophobicity of the phenol. An increase in the hydrophobicity of the phenol would provide an increased possibility for substrate accumulation in the hydrophobic membrane environment of the UDP-glucuronyltransferases, resulting in an apparently decreased Km.