Abstract
LY451395 (2-propanesulfonamide, N-[(2R)-2-[4′-[2-[methylsulfonyl)amino]ethyl][1,1′-biphenyl]-4-yl]propyl]-) is a potent and highly selective potentiator of the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptors. It is a biaryl-bis-sulfonamide and is known to be highly metabolized in preclinical species. In those metabolism studies, the metabolite structures were proposed exclusively by the analysis of mass spectrometric data. Although mass spectrometry is clearly a technique of choice for rapid identification of drug metabolites, occasionally, nuclear magnetic resonance spectroscopy is required to unambiguously assign and characterize, particularly, the regio- and stereochemistry of metabolic changes. Nuclear magnetic resonance spectroscopy, in general, is less sensitive than other detection methods and demands several micrograms of material for the analysis. To support full structure characterization of metabolites by NMR, in this study we demonstrated the application of a microbial-based surrogate biocatalytic system to produce sufficient amounts of the mammalian metabolites of LY451395. The results revealed that incubation of LY451395 with Actinoplanes missouriensis NRRL B3342 generated several metabolites that were previously detected in the in vivo metabolism studies of the preclinical species. Subsequent large-scale bioconversion resulted in the isolation of seven mammalian metabolites in milligram quantities for structural characterization by nuclear magnetic resonance spectroscopy. Furthermore, a selected group of metabolites generated from the microbial conversion served as analytical standards to monitor and quantify drug metabolites during clinical investigations.
Glutamate is the major excitatory transmitter in the brain. Several recent breakthroughs in the molecular biology and pharmacology of the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-subtype of glutamate receptors have led to the discovery of selective and potent AMPA receptor potentiators. These molecules enhance synaptic transmission and play important roles in plasticity and cognitive processes (Parsons et al., 2002). At Eli Lilly and Company, we have developed several novel series of biarylsulfonamides, which are potent potentiators of AMPA receptors in vitro (Gates et al., 2001; Miu et al., 2001). They have displayed functional central nervous system activity after systemic administration (Baumbarger et al., 2001a,b; Vandergriff et al., 2001; O'Neill et al., 2005) and are shown to be active in rodent models of cognition (Quirk and Nisenbaum, 2002; O'Neill et al., 2004) and depression (Li et al., 2001, 2003; Skolnick et al., 2001).
LY451395 (2-propanesulfonamide, N-[(2R)-2-[4′-[2-[methylsulfonyl)amino]ethyl][1,1′-biphenyl]-4-yl]propyl]-), a biaryl-bis-sulfonamide, is a selective and potent potentiator of AMPA receptors. In vivo metabolism of LY451395 in mouse, rat, dog, and human has generated several phase I oxidative metabolites (D. Coutant, W. Annes, T. Gillespie, T. Rash, W. Knebel, B. Peterson, and M. Goldberg; and W. Annes, T. Gillespie, B. Peterson, and D. Coutant, manuscripts in preparation), which required unambiguous structural characterization by spectroscopic means, including NMR. In addition, metabolite reference standards were required to monitor and quantify the metabolites during clinical investigations. Even though mammalian liver microsomes and/or hepatocytes can be used to generate drug metabolites, in our case, multiple incubations from different mammalian sources are required to cover the entire range of metabolites. Furthermore, the limited capacity of these systems requires a larger incubation volume to generate several milligrams of the metabolites. The use of microbial biocatalysis to produce mammalian metabolites of drugs has long been well recognized (Azerad, 1999; Lacroix et al., 1999; Preisig et al., 2003). Microbes are attractive biocatalytic systems because of their chemo-, regio-, and stereoselectivity, impressive catalytic efficiency, and capability of accepting a wide range of complex molecules as substrates. In addition, they are amenable to easy scale-up to produce sufficient amounts of metabolites for unambiguous structural determination, and to provide such metabolites as authentic standards.
In this study, we report using the microorganism Actinoplanes missouriensis for the preparation and structural determination of seven metabolites of LY451395, which were detected in the previous in vivo metabolism studies following oral administration of the compound in mice, rats, dogs, and humans (D. Coutant, W. Annes, T. Gillespie, T. Rash, W. Knebel, B. Peterson, and M. Goldberg; and W. Annes, T. Gillespie, B. Peterson, and D. Coutant, manuscripts in preparation). Overall, in those studies, M1 and M2 were detected as the major metabolites and M3 to M6, M8, and M9 were detected as the minor metabolites in preclinical species; and M1, M2, and M8 were detected in humans. In addition, in the present study, two metabolites unique to A. missouriensis are also reported for LY451395.
Materials and Methods
Materials. LY451395 (see Fig. 1 for the structure) was prepared at Eli Lilly and Company (Indianapolis, IN). All the microorganisms used for the screening were from the Lilly Culture Collection (Eli Lilly and Company), The American Type Culture Collection (Manassas, VA), the National Center for Agricultural Utilization Research (Peoria, IL), and Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). Actinoplanes missouriensis NRRL B3342 used for the scale-up incubation was obtained from the National Center for Agricultural Utilization Research. All other reagents and solvents were either analytical or HPLC grade.
Screening of Microorganisms. Each culture was started using 0.2 ml of the frozen stock inoculated into 13.5 ml of the appropriate growth medium based on the organism type (Goodhue, 1982). After 2 days of growth at 25°C on a rotary shaker at 165 rpm, approximately 10 mg of LY451395 in 0.2 ml of dimethyl sulfoxide was added to each culture. Five days later, the cultures were harvested by adding an equal volume of ethanol. For the bioconversion with A. missouriensis, ActinoBio production medium was used (see below for medium components).
Scale-up Incubation of LY451395 with A. missouriensis. Frozen stock culture (0.2 ml) of A. missouriensis was inoculated into seed medium consisting of 30 g/l trypticase soy broth, 3 g/l yeast extract, 2 g/l MgSO4·7H2O, 5 g/l glucose, and 4 g/l maltose (pH adjusted to 7.5 with NaOH) for 3 days at 30°C on a rotary shaker at 160 rpm. The seed culture (7 ml each) was then inoculated into four flasks each containing 140 ml of ActinoBio production medium consisting of 20 g/l dextrose, 5 g/l soybean meal, 5 g/l yeast extract, and 5 g/l K2HPO4 (pH adjusted to 7 with HCl). After a day of growth at 30°C on a rotary shaker at 250 rpm, 50 mg of LY451395 in 1 ml of sterile dimethyl sulfoxide was added to each flask. Five days later, the fermentation was harvested.
Isolation of Metabolites. The whole fermentation broth was centrifuged, mycelia were washed with 100 ml of water and centrifuged again, and the filtrates were combined (400 ml) and evaporated to yield fraction A. Water (100 ml) was added to fraction A and the suspension was mixed with 10 g of Amberchrom CG-161 resin (Rohm and Haas, Philadelphia, PA). The mixture was stirred for 1 h and filtered. This process was repeated with an additional 175 ml of water followed by 150 ml each of 3:1 water/methanol, 1:1 water/methanol, 1:3 water/methanol, and methanol. LC/MS analysis revealed significant amounts of metabolites in the latter three washes. These washes were combined and evaporated to yield fraction B.
Fraction B was divided into two portions, and one portion was purified over a Symmetry C18 column (19 × 300 mm, 5-μm particle size; Waters Corporation, Milford, MA) with a flow rate of 20 ml/min. The mobile phase consisted of 0.1% aqueous trifluoroacetic acid (mobile phase A) and 0.1% trifluoroacetic acid in acetonitrile (mobile phase B). Compounds were eluted using a gradient method: 0 min/30% B, 22 min/60% B, and 24 min/100% B. The second portion was purified with a slightly modified gradient method: 0 min/20% B, 22 min/50% B, and 24 min/100% B. The preparatory HPLC system consisted of a Waters 600 E pump and 2487 UV-VIS detector (Waters Corporation), a Gilson 215 autosampler and fraction collector (Gilson Inc., Middleton, WI), and a Micromass ZMD mass spectrometer (Waters Corporation). Fractions were collected by triggering on the relevant masses of the metabolites, and appropriate fractions were combined to yield from the first run M2 (5.1 mg; 93% purity), M4 (12 mg; 72% purity), M5 (21 mg; 95% purity), M8 (27.7 mg; 95% purity), and compounds 1 and 2 (9.8 mg; 94% purity). The second run yielded M3 (10.1 mg; 87%), M4 (11.5 mg, 95% purity), M5 (10.9 mg; 55% purity), M6 (6.7 mg; 95% purity), M8 (9.1 mg; 77% purity), and M9 (7.0 mg; 92% purity). The purity of metabolites was determined by analytical chromatography over a Discovery C18 column (2.1 × 150 mm, 5-μm particle size; Supelco, Bellefonte, PA) at a flow rate of 0.25 ml/min using a Shimadzu VP series HPLC system (Shimadzu Scientific Instruments Inc., Columbia, MD) connected to a Finnigan LCQ mass spectrometer (Thermo Electron Corporation, San Jose, CA). The mobile phase used was 10 mM aqueous ammonium formate (mobile phase A) and methanol (mobile phase B), and the analytes were eluted using the gradient method: 0 min/10% B and 45 min/90% B. The purity was determined by integrating peak areas using maximum UV absorbance. A portion of M3 was repurified to increase its purity over an Inertsil ODS3 column (4.6 × 250 mm, 5-μm particle size; MetaChem Technologies Inc., Torrance, CA) with a flow rate of 1 ml/min using 0.2% formic acid (mobile phase A) and methanol (mobile phase B). Analytes were eluted using the gradient method: 0 min/40% B, 25 min/57% B, 26 min/90%, and 28 min/90% B. The HPLC system consisted of a Waters 2690 Separation Module and a 966 Photodiode Array Detector (Waters Corporation). In all instances, the purity of metabolites submitted for NMR analysis was >90%.
NMR Analysis. Flow NMR experiments were carried out on an Inova NMR spectrometer equipped with a triple resonance flow probe (Varian Inc., Palo Alto, CA), operating at 600 MHz for 1H. Samples for NMR experiments typically contained approximately 0.5 mg of compound dissolved in 1 ml of 3:7 CH3CN/H2O. The technique of column trapping was used to introduce the compound into the flow probe. The compound in 1 ml of the above solvent mixture was transferred to a 2-ml HPLC injector loop, which was connected to the LC/NMR apparatus. The flow rate was set to 2 ml/min with D2O as the solvent. Using a flow splitter, 25% of this flow was directed through the loop and was then combined with the remainder of the flow in a 50-μl static mixer. From the mixer, the flow was directed to a trapping column (Aquasil, 30 × 3 mm, C18, 3μ particle size; Thermo Electron Corporation), and the compound elution was monitored with the UV detection at 260 nm. When the transfer was complete, the flow was stopped and the trapping column was reoriented to reverse the flow. The solvent was changed to 1:1 D2O/CD3CN to elute the compound from the column into the flow probe. Solvent suppression was achieved using a Varian standard WET solvent suppression sequence. All chemical shifts were referenced to the residual solvent signal (CD3CN) at 2.00 ppm. Two-dimensional WET-DQFCOSY, TOCSY, and ROESY experiments were performed using Varian standard pulse sequences.
Tube NMR experiments were carried out on an Inova NMR spectrometer equipped with a pulsed-field gradient and a Nalorac Z-SPEC microdual 3-mm probe (Varian Inc.), operating at 500 MHz for 1H and 125.7 MHz for 13C. Samples for NMR typically contained 0.2 mg to 3 mg of compound dissolved in 170 to 200 μl of dimethyl sulfoxide-d6 or CD3OD. Chemical shifts were referenced to the residual solvent signals: in the case of DMSO-d6, 2.49 ppm for 1H and 39.5 ppm for 13C, and in the case of CD3OD, 3.3 ppm for 1H and 49.0 ppm for 13C. Two-dimensional DQFCOSY, TOCSY, HSQC, HMBC, and CIGAR-HMBC experiments were performed using Varian standard pulse sequences.
Results
Screening of Microbes. Initial biotransformation of LY451395 was carried out using 32 microbes consisting of bacterial, actinomycetous, and fungal strains. Among them, a single organism, Actinoplanes missouriensis, produced the most number of products that corresponded to metabolites observed in the preclinical species, including mice, rats, and dogs. Figure 2 shows the LC/UV/MS chromatograms of a small-scale bioconversion of LY451395 with A. missouriensis. This organism was chosen for subsequent large-scale bioconversion studies to generate sufficient amounts of the metabolites for unambiguous structure determination by NMR, and to provide a selected group of metabolites as authentic standards during clinical studies.
Preparation and Isolation of Metabolites. A larger-scale bioconversion using A. missouriensis with 200 mg of LY451395 was conducted using a slightly modified fermentation medium as detailed under Materials and Methods. After bioconversion, the fermentation broth was separated into filtrate and mycelia. LC/MS analysis of the filtrate and the mycelial extract revealed that the mycelia contained predominantly LY451395 and the filtrate contained most of the metabolites. Thus, the filtrate was used for the subsequent isolation of metabolites. To remove most of the fermentation media-related components and to concentrate the metabolites, the filtrate was subjected to a capture step using CG-161 resin. After the capture step, the fractions enriched in the metabolites were combined and subjected to preparative reversed phase HPLC to afford metabolites M2, M4, M5, M6, M8, and M9 in >90% purity for NMR analysis. The sample containing M3 required additional purification to obtain >90% purity for NMR analysis. In addition, compounds 1 and 2 (detected only in the scale-up bioconversion), unique to A. missouriensis, were also isolated. Table 1 lists the amounts of the metabolites isolated from the bioconversion of LY451395 (200 mg) using A. missouriensis and their estimated purity.
Structure Determination. The structures of the nine isolated compounds were deduced by using LC/MS, MSn, and NMR (one- and two-dimensional) methods. To facilitate rapid identification of metabolites, MSn and NMR measurements were also carried out on the parent compound. Detailed analysis of one- and two-dimensional NMR data led to the accurate assignment of protons and carbons of LY451395 and its metabolites as listed in Tables 2 and 3, respectively. In general, deviations in the chemical shifts of key protons and carbons in each metabolite relative to the parent compound in combination with the MS data (Table 4) were used to determine the structure of all the metabolites.
Metabolite M2. M2 showed an ammoniated molecular ion at m/z 393. Overall, this ion represents the loss of the methyl sulfonamide group from the parent compound (ammoniated molecular ion at m/z 456) plus addition of molecular oxygen. MS2 of m/z 393 produced an intense ion at m/z 253 indicative of the loss of the isopropyl sulfonamide group and ammonia. In addition, MS3 of m/z 393 → m/z 253 showed a product ion at m/z 207 due to the loss of formic acid. These data suggested metabolic changes in the methyl sulfonamide region of the molecule and led to the proposal of the structure as shown in Fig. 1. The structure of M2 was further confirmed by NMR. Comparison of the 1H NMR data of M2 with those of LY451395 (Table 3) revealed that the isopropyl sulfonamide region of the molecule remains intact. After taking into account the aromatic protons, the only signal not accounted for in M2 integrated for two protons and appeared as a singlet at 3.61 ppm (H2-4). In the HSQC spectrum, this signal showed a direct correlation to a carbon at 40.0 ppm (C-4) and in the HMBC spectrum, in addition to two aromatic carbon signals at 133.9 (C-3) and 129.7 (C-4 and C-8) ppm, to a carbonyl carbon at 172.5 ppm (C-3). These data and the absence of the methyl resonance of the methyl sulfonamide moiety clearly revealed oxidative cleavage of the carbon-nitrogen bond leading to the carboxylic acid as shown in the formula (Fig. 1).
Metabolite M3. M3 showed an ammoniated molecular ion at m/z 381. MS2 of m/z 381 produced significant ions at m/z 346 and 267, the former representing the loss of water and ammonia and the latter, loss of CH3SO2 as a radical from m/z 346. These data suggested metabolic changes in the isopropyl sulfonamide region of the molecule. Accordingly, in the 1H NMR spectrum, the doublet signal integrating for 6 protons due to the isopropyl group was absent. Furthermore, the significant downfield shift experienced by the H3-18 (Δδ + 0.24 ppm) coupled with the appearance of H-17 resonance as a quartet (J = 7 Hz) as opposed to a sextet (J = 7 Hz) in LY451395 suggested that the carbon-nitrogen bond (nitrogen bearing the isopropyl sulfonamide group) was oxidatively cleaved and C-19 was oxidized to a carboxylic acid. The appearance of a CH proton signal at δ 4.79 (δC 73.8) spin-coupled to CH2 protons at δ 3.29 (δC 51.2) indicated additional oxidation either at the benzylic carbon (C-4) or carbon α to the amide nitrogen (C-3) of the methyl sulfonamide segment. The fact that the proton at δ 4.79 and the methylene protons at δ 3.29 attached to carbons at 73.8 and 51.2 ppm, respectively, strongly supported oxidation at the benzylic position (C-4) in M3.
Metabolite M4. The MS spectrum for M4 showed an ammoniated molecular ion at m/z 409, 16 Da higher than the corresponding ion of M2. MS2 of m/z 409 produced an intense product ion at m/z 374 due to the loss of water and ammonia. MS3 of m/z 409 → m/z 374 produced ions at m/z 310 (loss of SO2) and 268 (loss of [CH3]2C=SO2 from m/z 374). These data, taken together with the independent loss of formic acid (m/z 222) from m/z 268 and isopropyl sulfonamide (m/z 251) from m/z 374, suggested oxidation of M2 leading to M4 without further commitment as to the exact site of the oxidation. The 1H NMR data of M4 overall resembled quite well those of M2 except that the methyl doublet at 1.24 ppm (H3-18) in M2 was replaced by a methyl singlet at 1.56 ppm (H3-18) in M4 revealing hydroxylation at C-17. Thus, the structure of M4 was represented as shown in the formula (Fig. 1).
Metabolite M5. Mass spectroscopic data of M5 revealed addition of molecular oxygen (ammoniated molecular ion peak at m/z 488) to the parent compound. An intense peak at m/z 435 was observed in the MS2 spectrum of m/z 488 as a result of the loss of two water moieties and ammonia. In addition, MS3 of m/z 488 → m/z 435 produced ions at m/z 371 and 329 due to the losses of SO2 and [CH3]2C=SO2, respectively. In addition, the ions at m/z 312 (loss of isopropyl sulfonamide from m/z 435) and m/z 250 (loss of CH3SO2 radical from m/z 329) suggested hydroxylations at both the benzylic positions (C-4 and C-17) and/or α to the nitrogen atoms (C-5 and C-19).
The appearance of a methyl singlet at δ 1.46 (H3-18) similar to M4 and the presence of a CH signal at δ 4.66 in the 1H NMR spectrum of M5 indicated hydroxylations at C-17 and C-3 or C-4. The fact that in the HSQC spectrum the CH signal at δ 4.66 and CH2 protons at δ 3.12 correlated to carbons at 71.7and 50.2 ppm, respectively, clearly indicated the site of oxidation as C-4, similar to M3. This was further supported by the appearance of both NH resonances as triplet (J = 6 Hz), indicating that the carbons adjacent to the nitrogens remain unsubstituted as in the parent compound.
Metabolite M6. The LC/MS spectrum for M6 showed an ammoniated molecular ion at m/z 409 identical to that of M4. As with M4, the MS2 of m/z 409 also produced an intense ion at m/z 374 due to the loss of water and ammonia. Furthermore, the MS3 of m/z 409 → m/z 374 showed an ion at m/z 251 indicative of the loss of the isopropyl sulfonamide and additional ions at m/z 205 and 328 due to loss of formic acid from m/z 251 and 374, respectively. These data again suggested hydroxylation of M2 leading to M6. With the available data, however, the position of the hydroxyl group in M6 could not be determined. The presence of resonances diagnostic of the isopropyl sulfonamide segment in the 1H NMR spectrum of M6 have demonstrated metabolism of the methyl sulfonamide region of the molecule. The presence of a CH proton at δ 4.96 as a singlet directly attached to a carbon at 76.7 ppm suggested oxidative cleavage of the carbonnitrogen bond followed by subsequent hydroxylation at C-4. This conclusion was further supported by long-range correlations observed from the proton at δ 4.96 (H-4) to aromatic carbons C-5 (140.6 ppm), C-6 and C-10 (126.7 ppm), and a carbonyl carbon at 176.8 ppm (C-3) in the HMBC spectrum, and the structure of M6 was established as shown in Fig. 1.
Metabolite M8. MS data of M8 showed an ammoniated molecular ion at m/z 472, 16 Da higher than LY451395, suggesting a single oxidation of the parent molecule. MS3 of m/z 472 → m/z 437 (loss of water and ammonia) produced ions at m/z 373 and 331 indicative of the losses of SO2 and [CH3]2C=SO2 from m/z 437. In addition, the presence of an ion at m/z 236 represented the loss of methyl sulfonamide from m/z 331. These data suggested oxidation of the parent molecule either at C-17/18 or C4. The exact site of the hydroxylation was defined as C-17 as evidenced by the appearance of the H3-18 resonance as a singlet at δ 1.45 (Table 3) and C-17 at 72.8 ppm in the NMR spectra.
Metabolite M9. The MS data of M9 also showed an ammoniated molecular ion at m/z 472, 16 Da higher than the parent drug. The MS2 of m/z 472 produced a product ion at m/z 437 due to loss of ammonia and water and an intense product ion at m/z 314 due to further loss of isopropyl sulfonamide. The MS3 of m/z 472 → m/z 437 produced ions at m/z 373 (loss of SO2) and 331 (loss of isopropyl sulfonamide from m/z 437) similar to those of M8. In contrast to M8, the site of oxidation in M9 was identified as C-4 on the basis of the appearance of the H-4 signal at 4.65 ppm (δC 71.6) in the NMR spectra. Further evidence for this assignment was provided by the amide proton resonances at 7.01 and 7.05 ppm, which in each instance showed as a triplet (J = 6 Hz), suggesting that both amide nitrogens are adjacent to a CH2 group.
Compounds1 and 2. The mass spectra showed ammoniated molecular ion peaks at m/z 411, a deviation of 45 Da from the ammoniated molecular ion of LY451395, suggesting multiple structural changes in the metabolite. The 1H NMR spectrum was complicated because of the presence of two sets of resonances in approximately a 2:1 ratio. However, careful examination of the two-dimensional NMR data led to the assignment of all the protons and carbons (Tables 3 and 4; only the chemical shift assignments of the major isomer are shown) and led to the structural assignment of 1 and 2. The presence of two methyl doublets at 1.14 and 1.13 ppm (H3-22 and H3-23), one methyl singlet at 1.47 ppm (H3-18), and one amide resonance coupled to a CH2 group in the proton NMR spectrum revealed hydroxylation at C-17 and no additional changes in the isopropyl sulfonamide segment of the molecule. In contrast, the absence of the methyl resonance of the methyl sulfonamide group coupled with the presence of a low-field CH signal at δH 5.40 (δC 75.7) spin-coupled to an oxygenated CH2 group (δH 4.48 and 3.84, δC 71.0) suggested significant changes to the other half of the molecule. The chemical shifts of these protons and their HMBC correlation to an aromatic carbon at δ 140.5 (C-5) led to the assignment of a triol structure to compound 1 as shown in Fig. 1. Compound 2 is a diastereomer (epimeric at C-4) of compound 1. For the minor diastereoisomer 2, distinct resonances were observed for H2-3 at 3.47 ppm (δC 67.2) and H-4 at 4.57 ppm (δC 73.4). However, a stereochemical distinction between the two diastereoisomers and the stereochemistry at C-17 for both the diastereoisomers could not be achieved from the available data.
The metabolic changes in M3 to M6, M8, and M9 have either introduced an additional chiral center or modified the existing chiral center, or both. With the available data, however, the stereochemistry of these metabolites could not be determined.
Discussion
Biocatalysis has emerged as an important tool in the preparation of pharmaceutical intermediates and semisynthetic drugs. Both isolated enzymes and whole cell systems have successfully been used for the production of several important industrial products, including enantiomerically pure building blocks for pharmaceuticals and agrochemicals, active pharmaceutical ingredients, antibiotics, and food ingredients (Powell et al., 2001; Schmid et al., 2001; Panke and Wubbolts, 2005). Moreover, the concept of “microbial models of mammalian metabolism” has been well recognized (Smith and Rosazza, 1975; Azerad, 1999), and the use of microorganisms in drug metabolism studies is rapidly gaining importance (Lacroix et al., 1999; Preisig et al., 2003; Barbuch et al., 2006; Zhang et al., 2006). One of the key advantages of using microbial models to generate mammalian metabolites is that it provides a convenient way to produce sizable amounts of metabolites, otherwise very difficult to synthesize by chemical methods, for further investigations. Yet, despite the advantages, the number and diversity of microbial models applied to drug metabolism studies remain rather modest. We describe here an additional example of the application of microbial models in the preparation of several mammalian metabolites of a clinical candidate.
The in vivo metabolism and disposition of LY451395 in mice, rats, dogs, and humans and preliminary identification of its metabolites M1 to M6, M8, and M9 by LC/MS and tandem mass spectrometry have been described previously (D. Coutant, W. Annes, T. Gillespie, T. Rash, W. Knebel, B. Peterson, and M. Goldberg; and W. Annes, T. Gillespie, B. Peterson, and D. Coutant, manuscripts in preparation). In those studies, M1 and M2 were reported to be the major metabolites, the structures of which were determined by the interpretation of MS data; and M3 to M6, M8, and M9 were reported to be the minor metabolites, which required full characterization (except M6, the structure of which was discernable from the MS data) by NMR. Furthermore, metabolite reference standards were deemed essential to monitor and quantify metabolite concentrations in plasma and excreta during clinical investigations. To this end, we carried out a large-scale (200 mg) bioconversion of LY451395 with A. missouriensis. The bioconversion generated a total of nine compounds, of which seven were found to correlate with the in vivo metabolites detected in mice, rats, and dogs. However, M1, reported as a major metabolite in the preclinical species, was not detected in the microbial biotransformation using A. missouriensis. The isolated yields and estimated purity of these metabolites are listed in Table 1. As indicated in Table 1, an efficient bioconversion of LY451395 by A. missouriensis was observed (>65% turnover). This was accomplished without any systematic optimization of the organism's growth conditions or enhancing the organism's endogenous P450 enzymes.
The structures of the seven purified metabolites, M2 to M6, M8, and M9, and two compounds, 1 and 2, unique to microbial metabolism, were fully deduced (with the exception of stereochemistry) by NMR spectroscopy in conjunction with tandem mass spectrometry. The structures of metabolites M2 and M6 proposed by LC/MS and MSn in the earlier metabolism studies (D. Coutant, W. Annes, T. Gillespie, T. Rash, W. Knebel, B. Peterson, and M. Goldberg; and W. Annes, T. Gillespie, B. Peterson, and D. Coutant, manuscripts in preparation) were confirmed by NMR in the present study. In addition, the present study led to the assignment of the exact site of the metabolic changes (hydroxylations) in metabolites M3 to M5, M8, and M9. Overall, the metabolic pathways found to be common to the preclinical species and the microbe include oxidation of the carbons α to both the sulfonamide nitrogens leading to carboxylic acid metabolites due to the C-N bond cleavage (M2), direct oxidation at the benzylic carbons (M5, M8, and M9), and a combination of benzylic oxidation and C-N bond cleavage leading to carboxylic acid metabolites (M3, M4, and M6). More significantly, the present work enabled us to provide a selected group of metabolites as reference standards during clinical investigations.
In summary, we successfully used a microbial culture, A. missouriensis, to efficiently produce several mammalian metabolites of LY451395 in sizable amounts. This led to the unambiguous characterization of several mammalian metabolites detected previously in the in vivo metabolism studies of the preclinical species by MS and NMR methods, and enabled us to provide a selected group of the metabolites as reference standards.
Acknowledgments
We thank DeLise Douglas and Patrick Baker for contributions during the initial purification of the metabolites. We also thank Robert Barbuch and Dr. William Ehlhardt for the critical review of the manuscript.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.106.009522.
-
ABBREVIATIONS: AMPA, α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid; HPLC, high-performance liquid chromatography; LC/MS, liquid chromatography/mass spectrometry; MSn, n-stage mass spectrometry; WET, water eliminated through transverse gradients; DQFCOSY, double-quantum filtered correlation spectroscopy; ROESY, rotating frame nuclear Overhauser effect spectroscopy; HSQC, heteronuclear single-quantum coherence; HMBC, heteronuclear multiple bond correlation; CIGAR-HMBC, constant time inverse-detected gradient accordion rescaled long-range heteronuclear multiple bond correlation.
- Received January 27, 2006.
- Accepted February 22, 2006.
- The American Society for Pharmacology and Experimental Therapeutics