Characterization of o-demethylations and aromatic hydroxylations mediated by cytochromes P450 in the metabolism of flavonoid aglycons

One of the most important groups of metabolic enzymes is cytochrome P450 superfamily. These enzymes are important in terms of the catalytic diversity and the large number of xenobiotics that are detoxified or activated by converting to reactive metabolites. Flavonoids are xenobiotics to which humans are exposed through diet. Data on their oxidative metabolism mediated by cytochromes P450 are limited. The aim of this study was to determine the enzymatic kinetics of O-demethylation and aromatic hydroxylation of flavonoid aglycons on recombinant cytochrome P450 enzymes and human liver microsomes systems. The study was performed on ten flavonoids, namely 3,7dihydroxyflavone, 7-hydroxyflavone, acacetin, apigenin, flavone, galangin, kaempferol, naringenin, sakuranetin, and tangeretin using liquid chromatography coupled with mass spectrometry and UV detector. Most relevant enzyme involved in metabolism of flavonoid aglycons is CYP1A2, and its catalytic effectiveness ranges from 0.5 to 2.9 × 106 M–1 min–1. Having in mind high expression and involvement of CYP1A2 in metabolism of xenobiotics including drugs, and its intraindividual differences in expression and activity, potential of drug-flavonoid competitive interactions/inhibitions should be considered when consuming dietary supplement and foods rich in flavonoids.


INTRODUCTION
ETABOLISM is one of the ways by which human organisms protect themselves from foreign substances i.e. xenobiotics. Xenobiotics are usually biotransformed to more soluble metabolites susceptible to elimination. [1] One of the most important groups of metabolic enzymes is cytochrome P450 superfamily. These enzymes are important in terms of the catalytic diversity and the large number of xenobiotics that are detoxified or activated by converting to reactive metabolites. [1,2] Although present in all tissues, the highest concentration of cytochrome P450 involved in biotransformation of xenobiotics is found in liver endoplasmic reticulum (microsomes).
The liver microsomal cytochrome P450 have important role in determining the intensity and duration of the drug effect and also play a key role in the detoxification of xenobiotics. [2] The total number of cytochrome P450 substrates reaches several thousand. Such a wide spectrum of substrates is partly due to high number of cytochromes (57 in humans), and promiscuity in substrate selection. On one side, the cytochrome P450 enzyme system has a role in the metabolism of specific endogenous substrates such as fatty acids, steroid hormones and eicosanoids. While this is mainly attributed to mitochondrial cytochrome P450, microsomal cytochromes P450 have a main purpose to clean the body from ingested foreign substances such as terpenes, alkaloids, pyrolysis products, and other xenobiotics. [1][2][3] Cytochromes P450 are major enzymes involved in drug metabolism. Metabolism of 75 % of all known small molecule drugs is mediated by cytochromes P450. Out of that number 90 % of oxidoreductions is mediated by CYP3A4/5, CYP2C9, CYP2C19, CYP2D6, CYP2C8, and CYP1A2. [4] Flavonoids are xenobiotics to which we are exposed through diet. They represent low molecular weight polyphenol compounds characterized by the benzo-γpyrone (chromane) structure. Based on the modifications of benzo-γ-pyrone core, flavonoids are divided into six major subgroups (Table 1). [5][6][7][8][9][10] Most of the beneficial effects of flavonoids on human health are attributed to their antioxidant and healing properties. Flavonoids exhibit unique cardio-protective effects due to their ability to inhibit lipidmediated low density lipoproteins (LDL) oxidation. The protective role of diet rich in flavonoids has been demonstrated in several major prognostic studies. For example, it has been shown that a greater intake of flavonoids in the body lowers mortality due to coronary heart disease and lower incidence of myocardial infarction in older males. In postmenopausal women, the risk of coronary heart disease was reduced by 38 %. [9] Flavonoids can affect platelet aggregation at concentrations that can be achieved in the blood after consumption of foods rich in polyphenols. [11] Flavonoids can also interfere the results of  (platelet aggregation induced  by ristocetine and arachidonic acid), and show antitumoral, antioxidative and vasodilatory effects. [12][13][14][15][16] There are numerous other pharmacological effects of flavonoids described in the literature such as positive effects on capillary permeability and hypolipemic, antihypertensive, antimicrobial, antiviral, antiallergic, anti-inflammatory, antineoplastic, and hepatoprotective effects. [7,17] Although the number of known flavonoids is extremely high, over 4000, most flavonoids have not been characterized in terms of biological effects, and knowledge about metabolism is known only for the most common dietary flavonoids. [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33] In the previous study, screening of metabolic reactions mediated by cytochrome P450 was conducted on thirty flavonoid aglycons, most commonly found in medicinal plants and propolis. [34] Flavonoid aglycons are subject to O-demethylation and aromatic hydroxylation reactions, and their metabolism includes cytochrome P450, predominantly CYP1A2. [34] The aim of this study was to determine the enzymatic kinetics of Odemethylation and aromatic hydroxylation of flavonoid aglycons that have been shown to be susceptible to the metabolism mediated by human liver cytochromes P450 in the previously reported screening analysis. [34] The study was performed on ten flavonoids, namely 3,7-dihydroxyflavone, 7-hydroxyflavone, acacetin, apigenin, flavone, galangin, kaempferol, naringenin, sakuranetin, and tangeretin using the recombinant cytochrome P450 enzymes and human liver microsomes. Reactions were monitored using the liquid chromatography coupled with mass spectrometer and UV detector.

Incubations for Characterization of Metabolic Reactions Mediated by Human Liver Microsomes
Evaluation of enzyme kinetics was conducted using human liver microsomes (HLM) as source of enzyme. A range of concentrations from 2 to 200 μM flavonoid aglycons (2-800 μM in experiment with 3,7-dihydroxyflavone and 7hydroxyflavone) was prepared; appropriate aliquots of 20 mM flavonoid methanolic solution were transferred to glass tubes and evaporated to dryness on a water bath equipped with mechanical shaking. After evaporation of the solvent, a 100 μL incubation mixture was prepared in each of the tubes by adding human liver microsomes (HLM, amount corresponding to 100 pmol cytochrome P450 enzyme), 50 mM potassium phosphate buffer pH 7.4 and ultra-pure water. Generating system containing glucose-6phosphate, NADP + and glucose-6-phosphate dehydrogenase in a ratio of 100:50:2 (V/V/V) was used as a source of coenzyme, and its addition marked the beginning of the reaction (15 % volume in final incubation, V/V). In the negative control, 15 μL of ultra-pure water was added instead of generating system. After 15 minutes at 37 °C of incubation in water bath with mechanical shaking, the reaction was stopped by adding 60 μL of a mixture of acetonitrile and 25 % hydrochloric acid solution in a ratio of 5:1 (V/V). After stopping the reaction, each incubation mixture was centrifuged at 10,000 rpm for 10 minutes in a MiniSpin centrifuge (Eppendorf AG, Hamburg, Germany). The pure supernatant was transferred to the HPLC vial and analyzed on LC/MS/UVD.

Incubations for Characterization of Metabolic Reactions Mediated by Recombinant Cytochromes P450
Incubations with recombinant cytochromes P450 were conducted in the same manner as with the human liver microsomes. Instead of human liver microsomes, cytochromes P450 were used in incubations (10 pmol of CYP3A4 and CYP2D6 in experiment with 7-hydroxyflavone and CYP2C19 in the experiment with galangin; otherwise 3 pmol). Other incubation components were used in the same concentrations as stated above i.e. 50 mM potassium phosphate buffer pH 7.4, and 15 % volume of generating system. Volume of ulta-pure water was adjusted to final volume of 100 μL. Reaction was stopped and samples prepared for the LC/MS/UVD analysis, as described above.

HPLC-MS-UV Analysis
High performance liquid chromatography (HPLC) coupled with ultraviolet detector (UV) and mass spectrometer (MS) was performed on the Agilent 1200 RR (Agilent Technologies, Waldbronn, Germany) LC system coupled to Agilent 6530 Accurate Mass Q-TOF LC/MS Agilent Technologies (Waldbronn, Germany) using an electrospray ionization (ESI) interface. Separation was achieved on the Poroshell EC-C18 column (100 × 3.0 mm i.d., 2.7 μm particle size, Agilent). The temperature of column was set to 40 °C. The flow rate was 0.4 mL/min; mobile phase A consisted of water, methanol and formic acid at a ratio of 93:5:

Determination of Enzyme Kinetic Parameters
All incubations were conducted in triplicate. The results were expressed as the amount of generated metabolite based on LC-MS-UVD analysis (vide supra

RESULTS
In this study enzyme kinetics was determined for 3,7dihydroxyflavone, 7-hydroxyflavone, acacetin, apigenin, flavone, galangin, kaempferol, naringenin, sakuranetin, and tangeretin. These were flavonoids that showed metabolite formation in the amount of at least 10 % (m/m) in comparison with the amount of substrate in the human liver microsomes screening experiments. [34] An example of methodology used to determine kinetic parameters is provided in Figure 1.
Enzyme kinetics was determined for 10 flavonoids on human liver microsomes and for 8 flavonoids it was also characterized on recombinant cytochromes P450. Summarized results are presented in Table 2.
As shown on Figure 1, apigenin undergoes aromatic hydroxylation to luteolin and as expected from experiments with acacetin this conversion is mediated by CYP1A2. The catalytic effectiveness determined on human liver microsomes was (0.017 ± 0.009) × 10 6 M -1 min -1 , and in case of recombinant CYP1A2 this value was (0.5 ± 0.1) × 10 6 M -1 min -1 .
In case of flavone enzyme kinetic parameters could not be determined, as the formation of flavone metabolite was linearly dependent to the concentration used in this experiment (higher concentration of flavone could not be used due to solubility issue in physiological water media). While Vmax and Km values could not be determined, catalytic effectiveness could be extrapolated as a slope value of linear regression (under low concentrations of substrate rate of metabolite production is linearly dependent to substrate concentration). [35] Thus, catalytic effectiveness of 4'-hydroxylation of flavone catalyzed by human liver microsomes was (0.0015 ± 0.0001) × 10 6 M -1 min -1 . Minor products of aromatic hydroxylation are also generated at the positions 6 and 7, and overall catalytic effectiveness was (0.0028 ± 0.0002) × 10 6 M -1 min -1 .
Naringenin is metabolized to eriodictyol in the reaction of aromatic hydroxylation catalyzed by cytochrome P450 HLM in appropriate incubations. The catalytic effectiveness constant was (0.025 ± 0.009) × 10 6 M -1 min -1 . CYP1A2 is major cytochrome P450 involved in this reaction with catalytic effectiveness of (0.003 ± 0.002) × 10 6 M -1 min -1 . Since this is a slow reaction characterized by low efficiency, a very small amount of metabolite is formed in the incubations and the error of measurement is more pronounced in the results of the kinetic parameters.   Therefore, the obtained results can be taken as a rough estimate of the kinetics for this reaction. Sakuranetin undergoes sequential biotransformation, first demethylation to naringenin, then aromatic hydroxylation to eriodictyol. At the same time, minor product of direct aromatic hydroxylation of sakuranetin is formed i.e. 5,3',4'-trihydroxy-7-methoxyflavanone. Both metabolic pathways are observed in incubations with human liver microsomes and sequential biotransformation is 10 times more catalytic effective ((0.020 ± 0.005) × 10 6 M -1 min -1 ) when compared to hydroxylation at the positon 3' ((0.0022 ± 0.0002) × 10 6 M -1 min -1 ). Interestingly, only minor reaction was directly linked with cytochromes P450 used in this study i.e. CYP3A4 and CYP1A2 for which catalytic effectiveness constants were (0.06 ± 0.03) × 10 6 M -1 min -1 and (0.7 ± 0.4) × 10 6 M -1 min -1 , respectively.

DISCUSSION
In this study we have characterized kinetic parameters of O-demethylations and aromatic hydroxylations to which flavonoid aglycons are susceptible. The range of values for rate constant kcat was 0.021 to 14.2 min -1 . The Michaelis-Menten constants, Km, was obtained in the range of 0.7 to 492 μM, and calculated catalytic effectiveness, kcat/Km, was in the range from 0.0015 to 2.9 × 10 6 M -1 min -1 . The range of results obtained for all kinetic parameters is very wide, and within a three-orders of magnitude. If we have in mind the structural homogeneity of the analyzed compounds, these results indicate a relatively high selectivity of human liver cytochrome P450 to certain substrates from this group of compounds. This conclusion is supported by the fact that only a few recombinant cytochromes P450 (CYP1A2, CYPC19, CYP2D6 and CYP3A4) catalyze the observed metabolic reactions of the analyzed flavonoids.
The enzyme that is most efficient is CYP1A2. With the exception of metabolic conversion of naringenin to eriodictyol, CYP1A2 catalytic effectiveness, kcat/Km, was in the range from 0.5 to 2.9 × 10 6 M -1 min -1 , indicating the high catalytic specificity or efficacy of this cytochrome P450 in flavonoid metabolism. This result is in agreement with published data on the metabolism of flavonoids mediated by cytochromes P450, in which most authors of the studies concluded that CYP1A2 plays a major role in the oxidative metabolism of flavonoids in in vitro conditions. [19,20,22,23,25,26,28,30,32] Given the low concentrations generally achieved by oral intake of flavonoids, a very important kinetic parameter to consider is the Km value. This is especially important in sense that flavonoids are known to have very low solubility in water and, most importantly, poor bioavailability after oral administration. Consequently, the usual concentrations achieved in the blood after the flavonoids consumption from different sources are nanomolar to micromolar, with large interindividual differences, and most often do not exceed 10 μM values. [36] Flavonoids for which the low Km value has been determined have a higher chance of being metabolized in vivo after their oral administration compared to those having higher catalytic conversion rate (kcat) and high Km. In that sense, flavonoids that have higher potential of competitive interactions with other xenobiotics are: apigenin, acacetin, kaempferol, 3,7dihydroxyflavone and galangin; all having Km values bellow 10 μM ( Table 2).
Out of ten analyzed flavonoids, seven of them are susceptible to the metabolism mediated by CYP1A2 (Table  2), which further confirms the hypothesis that this cytochrome P450 is the most important for the metabolism of flavonoids in humans. However, for a better estimate of the possible clinical significance of these metabolic reactions, it is not enough to consider only the Km parameter. This parameter needs to be put in relation to the rate constant (kcat) to provide kcat/Km ratio known as catalytic effectiveness. E.g. kaempferol, galagnin, and naringenin, have low Km values, but given the relatively low rate of catalyst constant, it is expected that the extent of these reactions will not be so significant.
The range of the kcat/Km values of the studied reactions catalyzed by human liver microsomes was (0.0015 -2.9) × 10 6 M -1 min -1 . It has been previously pointed out that the values of kcat/Km ratio catalyzed by CYP1A2 enzymes ranged from 0.003 to 2.9 × 10 6 M -1 min -1 . Therefore, it is not surprising that reactions with the highest catalytic efficiency are catalyzed by CYP1A2. Tangeretin demethylation reactions catalyzed by CYP3A4 and CYP2D6 are only secondary metabolic pathways of this flavonoid, while major metabolic pathways are catalyzed by CYP1A2. Considering that CYP1A2 is one of the most commonly found cytochromes P450 in the liver and one of the most important enzyme involved in drug metabolism, [4] potential of interactions of drug-flavonoid interaction cannot be ruled out, especially in case of acacetin, 3,7-dihydroxyflavone, 7-hydroxyflavone, tangeretin, sakuranetin, apigenin, and galangin, with drugs that are primarily metabolized by CYP1A2 such as clozapine, duloxetine, fluvoxamine, haloperidol, imipramine, naproxen, olanzapine, zolmitriptan, etc. [15] When considering these interactions, it is important to keep in mind the great interindividual differences observed in CYP1A2 expression and activity, which ranges up to 40 folds. [4,37] It has already been pointed out that very little information on the metabolism of flavonoids mediated by cytochrome P450 is available in the literature. Data on kinetics of flavonoid metabolism is even less available. Bursztyka et al. and Hu et al. have studied kinetics of aromatic hydroxylation of genistein mediated by human liver microsomes and CYP1A2. [30,31] Wen and Walle studied the kinetics of demethylation of 5,7-dimethoxyflavone and 3',4'-dimethoxyflavone mediated by HLM. [38] Gradolatto et al. investigated the kinetics of aromatic hydroxylation of apigenin and the formation of luteolin in the incubation with rat liver microsomes. [21] Galangin metabolism and the kinetics of its conversion to kaempferol mediated by HLM were studied by Otake and Walle, and Otake et al. [26,39] The latter group of authors also studied the same reaction using CYP1A1, CYP1A2 and CYP2C9 enzymes. In all of these studies there is a common problem that makes it difficult to interpret the results of the enzyme kinetics. All incubations based on which the kinetic parameters were determined, lack the exact molar concentrations of the cytochrome P450 enzyme used in incubations. Instead of the molar concentration, the authors of the aforementioned studies used the protein mass unit in the incubations to express the results of these kinetic parameters. In these cases, kcat/Km constants included amounts of protein expressed in mg. This parameter is termed intrinsic clearance, which, according to Guengerich, is not good for comparison as it causes confusion. [16] The term "clearance" has a specific meaning in the in vivo pharmacokinetics.
In the literature only two studies have reported the kinetics of the metabolic reactions of flavonoids in which the actual parameters of their enzymatic kinetics kcat and kcat/Km were comparable to the results obtained in this study. Androutsopulos et al. studied kinetics of diosmetin metabolism and its conversion into luteolin by the CYP1A1, CYP1A2 and CYP1B1 enzymes. [22] The kinetic parameters of this reaction were determined and the catalytic effectiveness constant was 8.1 × 10 6 M -1 min -1 . However, our previous results do not indicate extensive metabolism of diosmetin which could be due to different enzyme sources used in these studies. Another extensive study of the ipriflavone metabolism was conducted by Moon et al. using recombinant CYP1A2, CYP3A4, CYP2C9, and CYP2C19. [8] Values of catalytic effectiveness were in the range of (0.02 -0.22) × 10 6 M -1 min -1 , and these are comparable to the results we obtained in our study.

CONCLUSION
Kinetics of O-demethylation and aromatic hydroxylation reactions of ten flavonoids were characterized on human liver microsomes cytochrome P450 system. For eight of them characterization was also conducted on recombinant cytochromes P450. Most relevant enzyme involved in metabolism of flavonoid aglycons is CYP1A2, and it catalyzes biotransformation of acacetin, 3,7-dihydroxyflavone, 7-hydroxyflavone, tangeretin, sakuranetin, apigenin and galangin (aligned according to the catalytic effectiveness). Having in mind high expression and involvement of CYP1A2 in metabolism of xenobiotics including drugs, and its intraindividual differences in expression, potential of drug-flavonoid competitive interactions/inhibitions should be considered when consuming dietary supplement and foods rich in flavonoids.