Materials and Methods

Chemicals.

MAMC, HAMC, and 7-hydroxy-4-(chloromethyl)-coumarin were synthesized as described previously (Onderwater et al., 1999). 7-Methoxy-4-(bromomethyl)-coumarin was obtained from Acros (Geel, Belgium). 7-Hydroxy-4-(methylpyridinium)-coumarin was a generous gift from Yamanouchi (Leiderdorp, The Netherlands). Glucose-6-phosphate dehydrogenase and NADPH were from Roche Molecular Biochemicals (Mannheim, Germany) and Applichem (Darmstadt, Germany), respectively. Furafylline was obtained from RBI (Natick, MA). Quinidine was from Aldrich (Zwijndrecht, The Netherlands). All other chemicals were obtained from Sigma Chemical Co. (St. Louis, MO).

Synthesis of 4-Substituted 7-Methoxycoumarins.

A 5 times molar excess of the corresponding primary or secondary amine or pyridine was added to a suspension of 250 mg of 7-methoxy-4-(bromomethyl)-coumarin in 20 ml of acetonitrile. The reaction mixture was left stirring at room temperature for approximately 1 h, resulting in a colorless solution in the case of the N-alkyl amines. Progress of each reaction was followed by thin-layer chromatography using acetone as mobile phase. After completion of the reaction, the mixture was acidified with 6 drops of 47% hydrobromic acid, and the solvent evaporated. The resulting yellow solid was then recrystallized up to two times in 80% isopropanol, after which the beige product could be obtained. In the case of the 4-methylpyridinium analog, the product was obtained by filtration of the suspension, resulting in a white powder. The identity of each product was established by ¹H NMR (dimethyl sulfoxide-d₆) obtained with a Bruker AC 200 (200.1 MHz) using tetramethylsilane as internal standard. Purities of the products were additionally tested by means of HPLC with UV-absorption detection at 220 nm and fluorescence detection at excitation and emission wavelengths of 330 and 400 nm, respectively. Product yield varied between 25 and 60%, and purities were higher than 98%.

N-Methyl 7-methoxy-4-(aminomethyl)-coumarin(MMAMC): δ 7.75 (1H, d, J = 10 Hz), 7.10–6.90 (2H, m), 6.35 (1H, s), 4.45 (2H, s), 3.85 (3H, s), 2.65 (3H, s).

N,N-Dimethyl 7-methoxy-4-(aminomethyl)-coumarin (diMMAMC): δ 7.90 (1H, d, J = 10 Hz), 7.15–7.00 (2H, m), 6.60 (1H, s), 4.55 (2H, s), 3.95 (3H, s), 2.90 (6H, s).

N-Ethyl 7-methoxy-4-(aminomethyl)-coumarin (EMAMC): δ 7.80 (1H, d, J = 10 Hz), 7.15–7.00 (2H, m), 6.50 (1H, s), 4.50 (2H, s), 3.90 (3H, s), 3.15 (2H, q, J = 9 Hz), 1.30 (3H, t, J = 14 Hz).

N-Propyl 7-methoxy-4-(aminomethyl)-coumarin (PMAMC): δ 7.75 (1H, d, J = 10 Hz), 7.15–7.00 (2H, m), 6.50 (1H, s), 4.45 (2H, s), 3.85 (3H, s), 3.05 (2H, t, J = 10 Hz), 1.8–1.4 (2H, m), 0.95 (3H, t, J = 8 Hz).

N-Butyl 7-methoxy-4-(aminomethyl)-coumarin (BMAMC): δ 7.80 (1H, d, J = 10 Hz), 7.15–7.00 (2H, m), 6.50 (1H, s), 4.45 (2H, s), 3.86 (3H, s), 3.08 (2H, t, J = 7 Hz), 1.75–1.55 (2H, m), 1.50–1.25 (2H, m), 0.95 (3H, t, J= 7 Hz).

7-Methoxy-4-(methylpyridinium)-coumarin (MMPyrC): δ 9.18 (2H, d, J = 8), 8.82–8.65 (1H, m), 8.35–8.20 (2H, m), 7.75 (1H, d, J = 10 Hz), 7.15–7.00 (2H, m), 6.25 (2H, s), 5.75 (1H, s), 3.90 (3H, s).

Synthesis of 4-Substituted 7-Hydroxycoumarins.

A 5 times molar excess of the corresponding N-alkyl amine was added under a nitrogen atmosphere to 1 g of 7-hydroxy-4-(chloromethyl)-coumarin dissolved in 60 ml of acetonitrile. The reaction mixture was left stirring for 48 h at room temperature, during which the progress of the reaction was followed by thin-layer chromatography using acetone as mobile phase. After completion of the reaction, the mixture was acidified with 6 N hydrochloric acid and the solvent evaporated. The resulting brown-yellow solid was taken up in 50 ml of H₂O and extracted three times with 50 ml of ethyl acetate to remove unreacted 7-hydroxy-4-(chloromethyl)-coumarin. The combined water layers were subsequently evaporated to dryness. The resulting yellow solid was then recrystallized twice in isopropanol, resulting in the beige product. The identity of the products was established by¹H NMR (dimethyl sulfoxide-d₆) obtained with a Bruker AC 200 (200.1 MHz) using tetramethylsilane as internal standard. Purities of the products were additionally tested by means of HPLC with UV-absorption detection at 220 nm and fluorescence detection at excitation and emission wavelengths of 370 and 470 nm, respectively. Product yield varied between 20 and 40%, and purities were higher than 98%.

N-Methyl 7-hydroxy-4-(aminomethyl)-coumarin(MHAMC): δ 10.85 (1H, s), 7.70 (1H, d, J = 10 Hz), 7.00–6.80 (2H, m), 6.40 (1H, s), 4.45 (2H, s), 2.65 (3H, s).

N,N-Dimethyl 7-hydroxy-4-(aminomethyl)-coumarin (diMHAMC): δ 10.90 (1H, s), 7.88 (1H, d, J = 10 Hz), 7.05–6.75 (2H, m), 6.65 (1H, s), 4.55 (2H, s), 2.85 (6H, s).

N-Ethyl 7-hydroxy-4-(aminomethyl)-coumarin (EHAMC): δ 10.85 (1H, s), 7.70 (1H, d, J = 10 Hz), 6.95–6.70 (2H, m), 6.45 (1H, s), 4.45 (2H, s), 3.15 (2H, q, J = 12 Hz), 1.30 (3H, t, J = 10 Hz).

N-Propyl 7-hydroxy-4-(aminomethyl)-coumarin (PHAMC): δ 10.85 (1H, s), 7.70 (1H, d, J = 10 Hz), 7.00–6.75 (2H, m), 6.48 (1H, s), 4.45 (2H, s), 3.05 (2H, t, J = 12 Hz), 1.9–1.55 (2H, m), 0.98 (3H, t, J = 9 Hz).

N-Butyl 7-hydroxy-4-(aminomethyl)-coumarin (BHAMC): δ 10.86 (1H, s), 7.70 (1H, d, J = 10 Hz), 6.95–6.75 (2H, m), 6.48 (1H, s), 4.42 (2H, s), 3.06 (2H, t, J = 13 Hz), 1.80–1.60 (2H, m), 1.50–1.25 (2H, m), 0.94 (3H, t,J = 7 Hz).

Calculation of Lipophilicities.

The lipophilicities of the synthesized 7-methoxy analogs were calculated according to the hydrophobic fragmental constant approach used by Rekker (Rekker and Mannhold, 1992). The log P value of the coumarin structure was obtained from Hansch et al. (1995).

Microsomal Protein.

Heterologously expressed human CYP1A2, -2B6, -2C9, -2C19, -2D6, -2E1, and -3A4 (catalog nos. P203, M110a, P258, P259, P171, M106k, and P202, respectively) were all obtained from Gentest Corp. (Woburn, MA). Human liver microsomes from three subjects (designated A9, A10, and A11) were a generous gift from Ph. Beaune (INSERM U75, Paris, France).

IC₅₀ Measurements.

To determine relative binding affinities of the 4-substituted 7-methoxycoumarins for CYP2D6, IC₅₀ measurements were performed using dextromethorphan as a substrate. In a total volume of 200 μl, 10 μM dextromethorphan, 2 pmol of CYP2D6, and 1 mM NADPH were incubated at 37°C in the presence and absence of various concentrations of the corresponding 4-substituted 7-methoxycoumarin. After 40 min, the reaction was stopped by the addition of 10 μl of 70% perchloric acid. The incubation mixture was then centrifuged for 20 min at 4000g. Subsequently, 50 μl of supernatant was injected on a HPLC column. Chromatographic separation of dextromethorphan and its metabolite, dextrorphan, was achieved with a HPLC system consisting of a model 300 and model 480 pump (Gynkotek, Germering, Germany), a Suplex pKb-100 column (4.6-mm × 15-cm) and guardcolumn (Supelcosil; Supelco Inc., Bellefonte, PA), a Triathlon autosampler (Spark Holland, Emmen, The Netherlands), and a model 821-FP fluorescence detector (Separations B.V., H. I. Ambacht, The Netherlands). As the eluent, 13% acetonitrile/1% triethylamine/86% H₂O, adjusted to pH 3 using 70% perchloric acid, was used at a flow rate of 0.9 ml/min. Detection was at an excitation wavelength of 280 nm (bandwidth 18 nm) and an emission wavelength of 311 nm (bandwidth 18 nm) with an analysis time of 20 min. Dextrorphan formation was linear with time using the described incubation conditions.

HPLC Analysis of Products Formed by CYP2D6 and CYP1A2.

To investigate the metabolism of the 4-substituted 7-methoxycoumarins by CYP2D6 and CYP1A2, 100 μM of each compound was incubated in a 100 mM phosphate buffer (pH 7.4) with 20 nM microsomal protein at 37°C. After 10 min of preincubation, the reaction was started by the addition of a preincubated NADPH-regenerating system, resulting in final concentrations of 0.1 mM NADPH, 0.3 mM glucose 6-phosphate, 0.4 mM MgCl_2, and 0.4 U/ml glucose-6-phosphate dehydrogenase. After 0 and 45 min, samples were drawn, and the reaction was stopped by the addition of 10 μl of 70% perchloric acid. After centrifugation for 20 min at 4000g, the supernatant was analyzed by means of HPLC using the system described above. The eluent consisted of 2% methanol/1% triethylamine/97% H₂O, adjusted to pH 3 using 70% perchloric acid, for MMAMC, diMMAMC, and EMAMC delivered at a flow rate of 0.6 ml/min. For PMAMC and BMAMC, a gradient was used. After eluting for 10 min with 2% methanol/1% triethylamine/97% H₂O, adjusted to pH 3 using 70% perchloric acid, the linear gradient was started, increasing the methanol content to 15% at 25 min using a flow rate of 0.6 ml/min. Fluorescence detection was performed at excitation and emission wavelengths of 370 and 470 nm (bandwidths 18 nm), respectively, for detecting O-dealkylated products, as well as 330 and 400 nm (bandwidths 18 nm) for detectingN-dealkylated products of the N-alkyl 7-methoxy-4-(aminomethyl)-coumarins. MMPyrC and its metabolite were detected with an Applied Biosystems model 759A UV-absorbance detector (Separations B.V.) using a wavelength of 325 nm because this appeared to be more sensitive than measuring fluorescence. The eluent applied for this compound consisted of 10% methanol/1% triethylamine/89% H₂O, adjusted to pH 3 using 70% perchloric acid, with a flow rate of 0.5 ml/min.

Enzyme Kinetics of O-Dealkylation by CYP2D6 and CYP1A2.

The O-dealkylation reactions of the N-alkyl 7-methoxy-4-(aminomethyl)-coumarins and unsubstituted MAMC by CYP2D6 were performed in Costar 3595 96-well plates using a 100 mM phosphate buffer (pH 7.4) and a final volume of 200 μl. Each compound was preincubated in a final concentration range of 1.25 to 160 μM (MAMC and diMMAMC) or 0.31 to 40 μM (MMAMC, EMAMC, PMAMC, and BMAMC) with 20 nM CYP2D6 at 37°C. After 10 min, the reaction was started by adding a preincubated NADPH-regenerating system, resulting in final concentrations of 0.1 mM NADPH, 0.3 mM glucose 6-phosphate, 0.4 mM MgCl₂, and 0.4 U/ml glucose-6-phosphate dehydrogenase. Product formation was subsequently followed in real-time using a Victor² 1420 multilabel counter (Wallac Oy., Turku, Finland) equipped with a plate heater set at 37°C. The fluorescence at excitation and emission wavelengths of 405 (bandwidth 8 nm) and 460 nm (bandwidth 30 nm), respectively, was measured at 5-min intervals for 45 min.

The enzyme kinetic parameters for the O-dealkylation of the MAMC analogs by CYP1A2 were essentially determined analogous to those for CYP2D6. However, the compounds were incubated in a final concentration range of 0.31 to 20 μM, and a CYP1A2 concentration of 10 nM was used. An exception was MMAMC, for which a CYP1A2 concentration of 30 nM was used. In the case of MAMC, the CYP1A2 samples, also containing 30 nM enzyme, were analyzed by HPLC with fluorescence detection because the rate of product formation was too low to be detected by means of a microplate reader.

Linearity and Enzyme Kinetics of MMPyrCO-Dealkylation by CYP2D6.

Before determining the enzyme kinetic parameters of HMPyrC formation by CYP2D6, the linearity of the reaction was investigated. MMPyrC was incubated in a final concentration of 100 μM with 20 nM CYP2D6 at 37°C. After 10 min of preincubation, the reaction was started by the addition of the NADPH-regenerating system described above. Samples of 200 μl were drawn after 0, 5, 10, 15, 30, and 45 min and added to 10 μl of a 70% perchloric acid solution. After centrifugation at 4000g for 20 min, the supernatant was analyzed by HPLC as described above.

To determine the enzymatic parameters of metabolism by P450 2D6, MMPyrC was incubated in a final concentration range of 5 to 160 μM with 20 nM CYP2D6 at 37°C, using a final volume of 200 μl. After 10 min of preincubation, the reaction was started by adding the NADPH-regenerating system described above. After 45 min, the reaction was stopped by adding 10 μl of a 70% perchloric acid solution. After centrifugation at 4000g for 20 min, the supernatant was analyzed by HPLC as described above.

P450 Isozyme Selectivity of O-Dealkylation of 4-Substituted 7-Methoxycoumarins.

The P450 isozyme selectivities of the N-alkyl 7-methoxy-4-(aminomethyl)-coumarins and MAMC were determined by incubating the compounds with CYP1A2, -2B6, -2C9, -2C19, -2D6, -2E1, and -3A4. The incubations were carried out at 37°C in a Shimadzu RF-5000 spectrofluorometer with excitation and emission wavelengths of 405 (5-nm slit) and 480 nm (10-nm slit), respectively. Twenty-five microliters of a NADPH-regenerating system (see above) and 10 μl of microsomal protein (final P450 concentration of 10 nM) were added to 955 μl of a 100 mM phosphate buffer (pH 7.4) preheated at 37°C. After equilibration at 37°C, 10 μl of an N-alkyl 7-methoxy-4-(aminomethyl)-coumarin solution was added, resulting in a final concentration of 5 times the K_m value for CYP2D6, i.e., 60, 20, 50, 6, 1, and 0.5 μM for MAMC, MMAMC, diMMAMC, EMAMC, PMAMC, and BMAMC, respectively. After following the real-time increase in fluorescence for several minutes, 10 μl of the corresponding 7-hydroxy compound was added in a concentration of 2.5 μM. The resulting increase in fluorescence was used as a scaling factor in the calculation of reaction rates.

MMAMC and BMAMC O-Dealkylation in Human Liver Microsomes.

In addition to the heterologously expressed P450s, P450 enzyme selectivity of MMAMC and BMAMC O-dealkylation was also tested with human liver microsomes in essentially the same manner as described above. MMAMC and BMAMC were incubated in a final concentration of 5 and 0.1 μM, respectively, in the absence and presence of the CYP1A2 inhibitor furafylline (30 μM) and the CYP2D6 inhibitor quinidine (1 μM). The incubations, with a total volume of 1.0 ml, were performed with three different batches of human liver microsomes, using 25 μl of microsomal protein. Furafylline was added 5 min before the addition of the corresponding coumarin. After following the reaction, samples were drawn for the purpose of metabolite identification.

Kinetic Analyses.

Apparent K_m andV_max parameters were determined by using the nonlinear least-squares fitting method for one-site binding, as implemented in Prism 2.0 (GraphPad Software Inc., San Diego, CA). In the case of the microplate reader assay, the time-dependent increase in fluorescence was plotted against the substrate concentration. V_max values were quantified by using calibration curves of the corresponding metabolite.

Modeling.

Structures of the analogs of MAMC and their low energy conformers were generated as described previously (Onderwater et al., 1999). An active site model of CYP2D6, based on a multi-alignment of CYP2D6 (to be published; available upon request) with the crystallographically resolved bacterial P450s terp (Hasemann et al., 1994), cam (Poulos et al., 1985), BM3 (Ravichandran et al., 1993), eryF (Cupp-Vickery and Poulos, 1995), and nor (Park et al., 1997), was built using the Homology Modeling Module of InsightII (Biosym/MSI, San Diego, CA). The protein model was relaxed by first minimizing the side chains, followed by a minimization with the backbone tethered using the default settings of the Discover program (Biosym/MSI).

The low energy conformers of each analog were manually docked into the active site model of CYP2D6 in such a way that the basic nitrogen atom of the substrate could form a hydrogen bond with either Asp³⁰¹ or Glu²¹⁶ and with the site of oxidation directed toward the heme iron atom. The complexes were subsequently energy-minimized and evaluated by means of the interaction energy between the substrate and active site residues and the distance between the site of oxidation and heme iron atom. The criterion for the latter was set at a maximal distance of 4.5 Å.

All minimizations were carried out with the Discover program (Biosym/MSI) using the conjugate gradient method and the consistent valence force field implemented with the heme parameters obtained fromPaulsen and Ornstein (1991, 1992).

Results

Fluorescent Properties of the 4-Substituted Coumarins.

The 4-substituted 7-methoxycoumarins and their putative 7-hydroxy metabolites, as shown in Fig. 1, were synthesized and their fluorescent characteristics determined. As shown in Table1, the introduction of N-alkyl substituents consistently shifted the excitation maximum of the 7-methoxy analogs from 326 to 342 nm. The nature of the incorporated substituent did not affect the excitation maxima. Effects ofN-substitution on emission maxima were less pronounced. In the case of the putative 7-hydroxy metabolites, alkylation of the nitrogen atom resulted in a smaller shift of the excitation maxima to longer wavelengths. In contrast, incorporation of a positively charged pyridine moiety at the 4-methyl position in the case of HMPyrC resulted in a shift of the excitation maximum to lower wavelengths. As a result, this product cannot be accurately quantified by fluorescence using a microplate reader due to interference with NADPH fluorescence. The relative fluorescence of the various 7-hydroxy compounds at a concentration of 1 μM, and excitation and emission wavelengths of 405 (bandwidth 8 nm) and 460 nm (bandwidth 30 nm), respectively, was also determined (Fig. 2). Interestingly, the intrinsic fluorescence of the 7-hydroxy N-alkyl analogs was 3.8 to 5.5 times higher than HAMC, the O-dealkylated metabolite of MAMC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Fluorescent yield of 1 μM 7-hydroxycoumarins measured in a microplate reader at excitation and emission wavelengths of 405 (bandwidth 8 nm) and 460 nm (bandwidth 30 nm), respectively, in a 100 mM phosphate buffer (pH 7.4).

The excitation wavelength of 405 nm was chosen to eliminate interference from background fluorescence of NADPH. HMPyrC could not be detected under these conditions.

Competition Experiments.

The relative binding affinities of MAMC and its synthesized analogs for CYP2D6 were determined by competition experiments in the presence of 10 μM of the CYP2D6 model substrate dextromethorphan. The observed IC₅₀ values, as listed in Table2, indicate that addition of a singleN-alkyl substituent to MAMC, and subsequent enlargement of its chain length, markedly increases the affinity of the coumarins for CYP2D6. However, the introduction of a second N-methyl group in diMMAMC resulted in a reduction in affinity for CYP2D6 when compared with MAMC and the mono-N-methyl analog. Interestingly, MMPyrC had an affinity toward CYP2D6 similar to that of MAMC. When examining the calculated lipophilicities of the compounds (Table 2), an excellent correlation (R² = 0.978) was obtained between Σf and the log IC₅₀values (Fig. 3). On statistical grounds, diMMAMC was excluded from this correlation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Correlation between log IC₅₀ and the calculated lipophilicities (Σf) of the mono-N-alkyl analogs of MAMC and MAMC itself (R² = 0.978).

DiMMAMC (○) was not included in this correlation.

Metabolite Identification of CYP2D6-Mediated Metabolism.

Because all of the MAMC analogs displayed affinity toward CYP2D6, our study was continued by investigating the metabolism of the structural analogs by this enzyme. Metabolite identification by HPLC analysis of the incubation mixtures, containing 100 μM of each analog and 20 nM CYP2D6, showed that in all cases O-dealkylation to the corresponding 7-hydroxy compounds occurred. Formation of MAMC and HAMC, which were anticipated as metabolites resulting fromN-dealkylation (Fig. 4), in the case of the mono-N-alkyl analogs, was not detected. DiMMAMC was not N-dealkylated by CYP2D6, either.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Metabolic routes of theN-alkyl analogs of MAMC mediated by CYP2D6 and CYP1A2.

O-dealkylation of the N-alkyl 7-methoxycoumarins resulted in the corresponding 7-hydroxy compounds, whereas N-dealkylation resulted in MAMC formation. A combination of O- and N-dealkylation resulted in the formation of HAMC.

Enzyme Kinetics of CYP2D6-Mediated O-Dealkylation.

For determining the enzyme kinetics of O-dealkylation of theN-alkyl MAMC analogs by heterologously expressed CYP2D6, the incubations were carried out in real-time at 37°C, with the exception of MMPyrC. For this compound, enzyme kinetics was determined by means of HPLC analysis.

Product formation was linear up to at least 45 min and displayed apparent Michaelis-Menten kinetics for all analogs. As shown in Table3, all MAMC analogs are good substrates of CYP2D6, with no K_m values above 16.31 ± 0.73 μM. It can also be seen that addition and elongation of a single N-alkyl chain led to a smallerK_m value, resulting in extremely low values of 0.20 ± 0.03 and 0.080 ± 0.020 μM for theN-propyl and N-butyl analog, respectively. Addition of a second N-methyl group in diMMAMC led to a somewhat higher K_m value when compared with the mono-N-substituted analog MMAMC, i.e., 10.56 ± 0.36 versus 5.09 ± 0.94 μM. Another observation made was thatN-alkylation and subsequent elongation of the alkyl substituent resulted in a reduction in the maximal rate of metabolism by CYP2D6. The intrinsic clearance of the compounds, however, showed a more than 30-fold increase going from unsubstituted MAMC to theN-butyl-substituted analog. MMPyrC displayed the lowest intrinsic clearance.

P450 Selectivity in the Metabolism of 4-Substituted 7-Methoxycoumarins.

The P450 selectivity of the novel substrates was first investigated with heterologously expressed P450s. Each of the MAMC analogs was incubated in a concentration of approximately 5 times itsK_m value for CYP2D6 in the presence of 10 nM of the most important drug-metabolizing enzymes, i.e., CYP1A2, -2B6, -2C9, -2C19, -2D6, -2E1, and -3A4. No metabolism could be detected in the case of CYP2B6, -2E1, and -3A4. The observed rates of metabolism by the other P450s is shown in Fig. 5. It can be seen that, at saturating conditions of CYP2D6, both MAMC and the analogs were also metabolized by recombinant CYP1A2 and CYP2C19, albeit to varying degrees. In the case of CYP1A2, only MAMC and MMAMC were metabolized at a lower rate than CYP2D6. The activity of CYP2C19 was, however, lower than that of CYP2D6 for all analogs except diMMAMC and BMAMC. Interestingly, CYP2C9 only metabolized the N-butyl analog of MAMC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

O-dealkylation of the 4-substituted 7-methoxycoumarins by microsomes of heterologously expressed human P450s at saturating concentrations of CYP2D6 (i.e., 60, 20, 50, 6, 1, and 0.5 μM MAMC, MMAMC, diMMAMC, EMAMC, PMAMC, and BMAMC, respectively).

All P450s were obtained from Gentest Corp. ▧, CYP2D6; ░, CYP1A2; ▩, CYP2C9; ▪, CYP2C19.

Metabolism of 4-Substituted 7-Methoxycoumarins in Human Liver Microsomes.

In addition to heterologously expressed P450s, the metabolism of MMAMC and BMAMC was also investigated in human liver microsomes to determine relative contributions of the various P450s toO-dealkylation in this system. Liver microsomes of three subjects (designated A9, A10, and A11) were used. The incubations were performed with 25 μl of microsomal protein and MMAMC and BMAMC concentrations of 5 and 0.1 μM (i.e., their approximateK_m values for heterologously expressed CYP2D6), respectively. MMAMC and BMAMC metabolism was reduced to 73.32 ± 22.12 and 56.42 ± 10.62%, respectively, after addition of 1.0 μM quinidine, a selective CYP2D6 inhibitor. After 5 min of preincubation with 30 μM furafylline, a selective CYP1A2 inhibitor, MMAMC and BMAMC metabolism was reduced to 22.10 ± 14.88 and 46.61 ± 15.50%, respectively. Residual activity after preincubation with furafylline could be completely inhibited by the addition of 1.0 μM quinidine. HPLC analysis of incubation samples only showed the formation of the corresponding O-dealkylated product.

Enzyme Kinetics of CYP1A2-Mediated O-Dealkylation.

Because CYP1A2, in addition to CYP2D6, was found to be the only physiologically relevant P450 enzyme contributing to theO-dealkylation of MMAMC and BMAMC, the metabolism of theN-alkyl analogs by CYP1A2 was investigated further. Product formation was linear for 30 min and 20 min, respectively, for MAMC and its analogs. Using these incubation times, apparent Michaelis-Menten kinetics was observed. As can be derived from the data in Table4, addition and elongation of theN-alkyl chain had little influence on theK_m values of the compounds, nor did the introduction of a second N-alkyl substituent. However, in contrast to CYP2D6, an increase in the correspondingV_max values was observed with increasing chain length. The effect of elongation of the mono-N-alkyl chain on the intrinsic clearance of the analogs was not as notable as the one observed for CYP2D6. Investigation of the time-dependent metabolism of MMPyrC, by HPLC analysis, showed that this compound is not metabolized by CYP1A2.

Metabolite Identification of CYP1A2-Mediated Metabolism.

Metabolite identification of incubation mixtures containing 100 μM compound and 20 nM CYP1A2 demonstrated that all N-alkyl analogs were mainly metabolized to the correspondingO-dealkylated metabolites. After 45 min of incubation, metabolite peaks corresponding to MAMC, HAMC, and MMAMC (in the case of diMMAMC) could also be detected (Fig. 6). Thus, apart from O-dealkylation reactions, CYP1A2 also mediates the N-dealkylation of the analogs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Overlay of the HPLC traces of the metabolism of 100 μM BMAMC by CYP1A2 (20 nM) at t = 0 and t = 45 min.

Peaks corresponding to HAMC (A), MAMC (B), BHAMC (C), and BMAMC (D) are visible.

Discussion

Recently, we described the development of a novel CYP2D6 substrate, 7-methoxy-4-(aminomethyl)-coumarin (Onderwater et al., 1999). MAMC is selectively O-dealkylated by CYP2D6 to the corresponding 7-hydroxy compound, HAMC. Due to the significantly different fluorescent properties of HAMC and MAMC and the favorable enzyme kinetic properties, the latter is suitable for high-throughput screening of ligands of CYP2D6, as recently demonstrated (Venhorst et al., 2000). In the current study we have used these advantages of the 4-(aminomethyl)-coumarin structure to investigate the effects of structural variations of MAMC on cytochrome P450 metabolism and selectivity.

The investigated set of structurally related compounds, shown in Fig.1, was obtained by introducing alkyl substituents to the nitrogen moiety of the 4-aminomethyl group of MAMC. The increase in affinity toward CYP2D6, observed for all of the secondary amine analogs, upon addition and enlargement of a single N-alkyl group, is most likely due to additional hydrophobic interactions of the substituents with active site residues. This hypothesis is supported by the excellent correlation observed between the log IC₅₀ values and calculated lipophilicities of the compounds (R² = 0.978, Table 2 and Fig. 3), as well as by docking studies of the substrates in an active site model of CYP2D6. The latter indicated that the alkyl substituents of the mono-alkylated analogs mainly interact with hydrophobic residues located in the I helix (Ile²⁹⁷ and Ala³⁰⁰) of the enzyme (Fig.7). The introduction of the second methyl group in diMMAMC resulted in a reduction in affinity. DiMMAMC was found to adopt a different orientation in the active site of CYP2D6 due to steric hindrance of the additional N-methyl group with I helix residues. Modeling studies suggested that, instead of forming a hydrogen bond with Asp³⁰¹, diMMAMC interacts with Glu²¹⁶ (Fig.8). Amino acid Glu²¹⁶ has also been proposed to play a role in the binding of other ligands of CYP2D6 (Lewis et al., 1997; De Groot et al., 1999). The different binding orientation of diMMAMC may explain why this analog does not fit the above-mentioned correlation. MMPyrC was found to have a binding orientation similar to those of the mono-N-alkyl-substituted analogs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Proposed binding orientation of BMAMC in the active site of CYP2D6.

The basic nitrogen atom of BMAMC is hydrogen bonded to Asp³⁰¹ (purple) in the I helix of CYP2D6. TheN-butyl group of BMAMC is involved in hydrophobic interactions with Ile²⁹⁷ (orange) and Ala³⁰⁰(yellow), which are also located in the I helix. The heme moiety of CYP2D6 is shown in gray.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8

Docking results of diMMAMC in the active site of CYP2D6.

DiMMAMC adopts a binding orientation distinctly different from that of BMAMC (Fig. 7). Instead of forming a stabilizing hydrogen bond with Asp³⁰¹, modeling studies suggested that the basic nitrogen atom of this analog interacts with Glu²¹⁶ (purple), which is located in the F helix of the enzyme. The heme moiety of CYP2D6 is shown in gray.

Both CYP2D6 and CYP1A2 were found to catalyze theO-dealkylation reaction of the studied mono- and di-N-alkyl analogs of MAMC. N-Dealkylation was not observed for CYP2D6, although this enzyme is intrinsically capable of performing this metabolic route (Coutts et al., 1994; De Groot et al., 1995). Thus, it can be concluded that the regioselectivity of the metabolism by CYP2D6 is not compromised by substitutions at the 4-aminomethyl group of MAMC. In contrast to CYP2D6, an additional metabolic route was observed for CYP1A2, as this enzyme also mediated the N-dealkylation of the compounds, resulting in several products (Fig. 4).

The present kinetic studies on the O-dealkylation of the analogs by CYP2D6 demonstrated that N-alkyl substitution of MAMC notably affected the K_m andV_max values (Table 3) of the compounds. The increase in intrinsic clearance values (i.e.,V_max/K_m, Table3) with elongation of the mono-N-alkyl chain was most marked. The trend in the K_m values of theN-alkyl analogs corresponds well with the one observed for the IC₅₀ values (Table 2), indicating that theK_m value reflects the affinity of the compounds for CYP2D6. An interesting feature with respect to the high-throughput screening prospective of the present analogs is that the fluorescent signal of the N-substituted 7-hydroxy compounds is higher than that of HAMC (Fig. 2). Thus, despite their somewhat lower V_max values, MMAMC, diMMAMC, and EMAMC O-dealkylation can be measured in the microplate reader with a higher sensitivity than MAMC. In the case of MMAMC, the sensitivity was increased more than 3-fold (Fig.9).

• Download figure
• Open in new tab
• Download powerpoint

Figure 9

Michaelis-Menten plot for the CYP2D6-mediated O-dealkylation of MAMC (●) and MMAMC(▴).

The rate of metabolite formation is given in arbitrary fluorescence units to demonstrate the higher sensitivity of MMAMC. Experiments were performed at 37°C in a 100 mM phosphate buffer (pH 7.4) using CYP2D6 expressed in human lymphoblasts.

Apart from its high turnover, the most important advantage of MAMC over other fluorescent probes is its selectivity for CYP2D6 (Onderwater et al., 1999). Because the use of human liver microsomes as an in vitro system is imperative in the process of drug discovery and development, we also investigated whether the isozyme selectivity of MAMC was influenced by the introduction of N-alkyl substituents. In heterologously expressed systems and at saturating conditions of CYP2D6, MAMC and all of the N-alkyl analogs were also metabolized by CYP1A2 and CYP2C19 (Fig. 5). Only BMAMC was also metabolized by CYP2C9. Thus, the identity of participating P450s has remained unchanged upon addition of an N-alkyl group, with the exception of BMAMC. The ratio of metabolism by the various heterologously expressed P450s did differ between the different analogs (Fig. 5).

Incubations of MMAMC and BMAMC in human liver microsomes showed that these substrates were exclusively metabolized by CYP2D6 and CYP1A2. Regardless of its relative contribution, the latter could be totally inhibited by preincubating with furafylline, a selective CYP1A2 inhibitor. As MAMC was also shown to be exclusively metabolized by CYP1A2 and CYP2D6 in human liver microsomes (Onderwater et al., 1999), it can be concluded that introduction of an N-alkyl chain does not qualitatively affect the P450 selectivity of the investigated compounds in this system.

In conclusion, the addition of mono-N-alkyl substituents to MAMC leads to higher affinities toward CYP2D6. Although the structural variation introduced did not result in altered metabolic profiles, the CYP2D6-mediated O-dealkylation was strongly influenced by the substituents. The P450s contributing to O-dealkylation of the N-alkyl analogs of MAMC were found to be identical to those of MAMC itself (Onderwater et al., 1999) in both human liver microsomes and heterologously expressed P450s, with the exception of BMAMC. Thus, next to CYP2D6 only CYP1A2 contributed to theO-dealkylation of the N-alkyl compounds in human liver microsomes. Further investigation of the CYP1A2-mediatedO-dealkylation showed that, in sharp contrast to CYP2D6, theK_m value of the MAMC analogs was virtually unaffected by the structural variation introduced, whereas the turnover increased with elongation of the N-alkyl chain. An additional metabolic pathway was observed in the case of CYP1A2, asN-dealkylation was also detected.