Chem. Res. Toxicol. 2005, 18, 317-323
317
Identification of Aldehyde Oxidase as the Neonicotinoid Nitroreductase Ryan A. Dick, David B. Kanne, and John E. Casida* Environmental Chemistry and Toxicology Laboratory, Department of Environmental Science, Policy and Management, University of California, Berkeley, California 94720-3112 Received September 21, 2004
Imidacloprid (IMI), the prototypical neonicotinoid insecticide, is used worldwide for crop protection and flea control on pets. It is both oxidatively metabolized by cytochrome P450 enzymes and reduced at the nitroguanidine moiety by a previously unidentified cytosolic “neonicotinoid nitroreductase”, the subject of this investigation. Two major metabolites are detected on incubation of IMI with rabbit liver cytosol: the nitrosoguanidine (IMI-NO) and the aminoguanidine (IMI-NH2). Three lines of evidence identify the molybdo-flavoenzyme aldehyde oxidase (AOX, EC 1.2.3.1) as the neonicotinoid nitroreductase. First, classical AOX electron donor substrates (benzaldehyde, 2-hydroxypyrimidine, and N-methylnicotinamide) dramatically increase the rate of formation of IMI metabolites. Allopurinol and diquat are also effective electron donors, while NADPH and xanthine are not. Second, AOX inhibitors (potassium cyanide, menadione, and promethazine) inhibit metabolite formation when Nmethylnicotinamide is utilized as an electron donor. Without the addition of an electron donor, rabbit liver cytosol reduces IMI only to IMI-NO at a slow rate. This reduction is also inhibited by potassium cyanide, menadione, and promethazine, as well as by additional AOX inhibitors, cimetidine and chlorpromazine. Finally, IMI nitroreduction by AOX is sensitive to an aerobic atmosphere, but to a much lesser extent than cytochrome P450 2D6. Large species differences are observed in the IMI nitroreductive activity of liver cytosol. While rabbit and monkey (Cynomolgus) give the highest levels of total metabolite formation, human, mouse, cow, and rat also metabolize IMI rapidly. In contrast, dog, cat, and chicken liver cytosols do not reduce IMI at appreciable rates. AOX, as a neonicotinoid nitroreductase, may limit the persistence of IMI, and possibly other neonicotinoids, in mammals.
Introduction Neonicotinoids are the most important new class of insecticides of the last three decades, with worldwide use accounting for 11-15% of the total insecticide market value (1). Imidacloprid (IMI1, Figure 1) was the first neonicotinoid developed, is the most commonly used, and is the prototype to which others are compared (2). It is highly effective for crop protection (1) and flea control on cats and dogs (3). IMI is a potent and selective insect nicotinic acetylcholine receptor agonist (4) with a generally favorable mammalian toxicological profile (1, 5, 6). It is systemic in plants (2) and enters soil and groundwater in the course of its normal use. These factors are taken into account in evaluating dietary exposure from food and drinking water and nondietary exposure in the aggregate risk assessment and determination of safety (5, 6). The metabolic pathways of IMI in laboratory animals have been postulated largely from excretion products (5, 7). IMI metabolism follows one of three pathways: oxidative * To whom correspondence should be addressed. Tel: 510-642-5424. Fax: 510-642-6497. E-mail:
[email protected]. 1 Abbreviations: ACN, acetonitrile; AOX, aldehyde oxidase; DAD, diode array detection; DMAC, p-dimethylaminocinnamaldehyde; IMI, imidacloprid; IMI-NH2, aminoguanidine metabolite of IMI; IMI-Nd CHCH3, acetaldehyde imine adduct of IMI-NH2; IMI-NO, nitrosoguanidine metabolite of IMI; KCN, potassium cyanide; LC/MS/MS, HPLC with tandem mass spectrometry; RLC, rabbit liver cytosol; XO, xanthine oxidase.
Figure 1. Nitroreduction of IMI by rabbit liver cytosol to the nitrosoguanidine (IMI-NO) and aminoguanidine (IMI-NH2).
cleavage at the methylene bridge connecting the heterocyclic substituents; hydroxylation and dehydration of the ethylene moiety; and reduction and cleavage of the nitroguanidine. The first steps have been taken in identifying the IMImetabolizing enzymes and possible reactive intermediates. Of the human P450s, 3A4 is largely responsible for 5-hydroxylation and subsequent dehydration, while several carry out anaerobic IMI nitroreduction (8). A cytosolic “neonicotinoid nitroreductase” from rabbit liver with superior aerobic activity has been characterized, but not identified (9). This enzyme is inhibited by several aldoketo reductase inhibitors and does not require nicotinamide adenine dinucleotide cofactor electron donors. Additionally, recombinant human aldehyde and aldose
10.1021/tx049737i CCC: $30.25 © 2005 American Chemical Society Published on Web 01/28/2005
318
Chem. Res. Toxicol., Vol. 18, No. 2, 2005
reductases and bovine xanthine oxidase (XO) do not exhibit IMI nitroreductive activities (9). The goal of this study was to identify the “neonicotinoid nitroreductase” of rabbit liver cytosol (RLC) using IMI as the prototype. Specific electron donor substrates and cofactors provide evidence that it is the molybdoflavoenzyme aldehyde oxidase (AOX, EC 1.2.3.1), while inhibitors confirm this proposal (10-12). As determined by HPLC with tandem mass spectrometry (LC/MS/MS), IMI is reduced to both a nitrosoguanidine (IMI-NO) and an aminoguanidine (IMI-NH2) (Figure 1). The metabolic profile is also related to oxygen sensitivity, cytosol concentration, and time. Finally, differences in the IMIreductive activities of AOX from nine different species, including human, dog, and cat, are reported.
Materials and Methods Reagents. Sources of IMI, IMI-NO and IMI-NH2 were described previously (9). All other substrates and inhibitors were obtained from Sigma-Aldrich (St. Louis, MO). Acetonitrile (ACN) was purchased from Fisher Scientific, dichloromethane from Sigma-Aldrich, and ethyl acetate from EMD chemicals (Gibbstown, NJ); all were HPLC grade. Rabbit livers were either purchased from Pel-Freez Biologicals (Rogers, AR) or isolated from freshly euthanized New Zealand whites. Mouse, cow, and chicken livers were also obtained from Pel-Freez. Pooled human and rat (Sprague-Dawley) liver cytosols, and monkey (Cynomolgus) and dog (beagle) liver S9 fractions were obtained from BD Biosciences (Bedford, MA). Cat liver was provided by Jeffery Winer (University of California, Berkeley). Recombinant 2D6 expressed in SUPERSOMES was also obtained from BD Biosciences. Enzyme Preparations and Incubations. Cytosol was prepared from liver homogenates or S9 fractions in 100 mM potassium phosphate buffer, pH 7.4, as described previously (9). Partial purification of AOX from RLC involved heat treatment (60 °C for 10 min), ammonium sulfate (50% saturation) precipitation, and extensive dialysis against 100 mM phosphate buffer, pH 7.4 (13). This procedure results in purification of AOX approximately 20-fold (13, 14). Protein concentrations were determined using the bicinchoninic acid reagent from Pierce (Rockford, IL). Incubation mixtures and final concentrations were as follow unless otherwise noted: cytosol containing 1 mg of protein; 400 µM IMI (added in Me2SO); and 2 mM electron donor substrate in 200 µL 100 mM potassium phosphate buffer, pH 7.4. Concentrations of 10 mM N-methylnicotinamide and 1.25 mM IMI were found to be saturating for IMI reduction by RLC. Diagnostic inhibitors were dissolved in Me2SO or phosphate buffer. The Me2SO concentration in incubations never exceeded 1.5% and did not affect IMI nitroreduction. Unless otherwise noted, incubations were aerobic at 37 °C for 2 h. Anaerobic conditions were created using the following protocol: phosphate buffer was degassed under vacuum and sparged with argon for 30 min; the reaction minus electron donor was assembled in a sample vial sealed with a septum; argon was blown into the vial for 1 min; and electron donor was added using a syringe to start the reaction. For incubations of IMI with P450 the following conditions were used: 10 pmol 2D6, 1 mM NADPH and 1.25 mM IMI in 200 µL 100 mM phosphate buffer, pH 7.4. Metabolite Identification. Metabolites were identified by LC/MS/MS comparisons with synthetic standards. Incubations were stopped by addition of ice-cold ACN (400 µL). They were then deproteinized by incubation at 4 °C for 10 min, followed by centrifugation at 20,000g at 4 °C for 10 min. Supernatants were isolated and evaporated using a SpeedVac (Thermo Savant, Milford, MA). The residues were resuspended in water and desalted with Oasis HLB columns (Waters, Milford, MA). Metabolites were separated by a Luna C8(2) 3µ 100 × 2.0 mm HPLC column (Phenomenex, Torrance, CA), at a flow rate of
Dick et al. Table 1. LC/MS/MS and HPLC-DAD tR Values and Molecular Ions of IMI and Its Metabolites LC/MS/MS
HPLC-DAD
compound
tR (min)a
[M + H]+/ [M + H + ACN]+ b
tR (min)a
IMI-NH2c IMI-NdCHCH3c IMI-NO IMI
11.2 19.1 22.0 24.0
226/267 252/293 240/281 256/297
10.6 13.1 14.2 18.4
a Different chromatographic conditions and instrumentation used (see Materials and Methods). b m/z with predominant ion in italics. c IMI-NdCHCH3 is an artifact from acetaldehyde introduced during analysis. It is included with IMI-NH2 for metabolite quantification (see Materials and Methods).
0.2 mL/min, using as the mobile phases, 0.1% formic acid in water (solvent X) and ACN (solvent Y) in a linear gradient: 0 min 0% Y; 21 min to 60% Y; 25 min 60% Y; and 27 min to 0% Y. A Finnegan TSQ 700 mass spectrometer (ThermoFinnegan, San Jose, CA) was used with electrospray ionization to characterize IMI and its metabolites. All samples were monitored full scan (200-400 amu, 2 s scan time). A collision voltage of -25 eV and argon collision gas at a pressure of 2.3 Torr were utilized for MS/MS. Metabolite Analysis. Incubations for analysis by HPLC with diode array detection (DAD) were stopped by the addition of ethyl acetate (400 µL) containing 1 nmol of clothianidin [another neonicotinoid (1) used here as an internal standard]. Samples were vortexed vigorously and then centrifuged at 10,000 g for 10 min. After the organic layer was removed, 5 M NaOH (10 µL) and dichloromethane:ACN (5:2, 400 µL) were added to the aqueous layer. Samples were again vortexed and centrifuged, and the organic layer combined with that of the previous step. Solvent was evaporated using a SpeedVac, and the metabolic residues resuspended in trifluoroacetic acid:water: ACN (0.1:75:25, 100 µL). They were then injected into a HewlettPackard 1050 HPLC-DAD (Palo Alto, CA) and chromatographed using a Luna C18 5µ 250 × 4.6 mm column. A flow rate of 1 mL/min and mobile phases of 0.1% trifluoroacetic acid in water (solvent A) and ACN (solvent B) were utilized with the following linear gradient: 0 min 5% B; 20 min to 40% B; 25 min to 80% B; 30 min 80% B; and 33 min to 5% B. Authentic standards of IMI, IMI-NO, and IMI-NH2 were detected at 254 nm and used to quantify metabolites. Additionally, a significant percentage (ranging from 35 to 88%) of the IMI-NH2 metabolite was detected as an artifactual acetaldehyde imine adduct (IMI-Nd CHCH3). Using both IMI-NH2 and 2,4-dinitrophenylhydrazine, acetaldehyde was determined to be a minor impurity of the HPLC-grade ethyl acetate used for extractions. IMI-NdCHCH3 was synthesized from IMI-NH2 using a general procedure (15), characterized by LC/MS/MS (16) and used as a standard to quantify IMI-NH2. Oxidative Activity of AOX. The oxidative activity of AOX was assayed spectrophotometrically using p-dimethylaminocinnamaldehyde (DMAC) as reported previously (17). Briefly, 0.3 mg of liver cytosolic protein was added to 100 mM potassium phosphate buffer, pH 7.4 in a well of a 96-well plate. DMAC (100 µM) was added to a final volume of 200 µL to start the reaction. Incubations at 25 °C were monitored using a VersaMax microplate reader (Molecular Devices, Sunnyvale, CA) at 398 nm for 400 s.
Results Reductive Metabolites of IMI. IMI metabolism by RLC produces only IMI-NO and IMI-NH2 metabolites in significant amounts (Figure 1). The retention times, mass spectra (Table 1), and daughter ion spectra with MS/MS (data not shown) of these metabolites are identical to those of authentic standards. These standards are subsequently used with HPLC-DAD to quantify IMI me-
Identification of the Neonicotinoid Nitroreductase
Figure 2. Effect of electron donor substrates (2 mM) on IMI nitroreduction by rabbit liver cytosol (1 mg protein) in order of increasing overall effectiveness. Data are mean ( SD, n ) 4. ND ) Not detected.
tabolite formation. Because of its hydrazine functional group, IMI-NH2 is a potent scavenger of aldehydes and ketones. A significant percentage of IMI-NH2 was routinely detected as an imine adduct of acetaldehyde, which is present in trace amounts in some organic solvents, including ethyl acetate (Materials and Methods), and on thermal degradation of plastics (18). Electron Donor Substrates. Specific electron donor substrates and cofactors were used at 2 mM to determine their effect on IMI nitroreduction by RLC (1 mg protein, Figure 2). NADH and NADPH did not previously enhance IMI reduction in this system (9); however, under the conditions used in this study, NADPH slightly increases the rate of IMI-NO formation. In marked contrast, classical electron donor substrates of AOX, benzaldehyde, N-methylnicotinamide, and 2-hydroxypyrimidine, dramatically increase the rate of IMI-NO and IMI-NH2 formation. IMI-NH2 levels are diminished when benzaldehyde is used, presumably because the hydrazine moiety of the IMI metabolite reacts with the aldehydic electron donor. IMI and RLC were also incubated with xanthine and allopurinol, a natural substrate and potent inhibitor of XO, respectively, to verify that nitroreduction is not due to cross reactivity of AOX-specific substrates with this closely related enzyme. While coincubation with xanthine does produce very low levels of metabolites, allopurinol is a superior electron donor substrate for IMI reduction. These data indicate that XO does not play a significant role in IMI reduction in rabbit liver. Two major bipyridylium herbicides, diquat and paraquat, are tested here for the first time as possible electron donor substrates of AOX. Diquat is the most active of the compounds assayed, while coincubation with paraquat appears to only slightly increase the rate of IMI nitroreduction. Partial Purification of AOX. Partial purification of AOX from RLC (see Materials and Methods) leads to nearly 20-fold enrichment of IMI-nitroreductive activity when N-methylnicotinamide is used as the electron donor (data not shown). Of the IMI consumed, most (>90%) is metabolized to IMI-NH2; minor amounts of IMI-NO are detected. Because of the purification process, which involves ammonium sulfate precipitation and extensive dialysis, these samples are depleted of endogenous co-
Chem. Res. Toxicol., Vol. 18, No. 2, 2005 319
Figure 3. Effect of diagnostic inhibitors (KCN at 500 µM and others at 125 µM) on IMI nitroreduction by rabbit liver cytosol (1 mg protein) with N-methylnicotinamide (2 mM) electron donor. Data are mean ( SD, n ) 4. ND ) Not detected. Table 2. Effect of Diagnostic Inhibitors on IMI Nitroreduction by Rabbit Liver Cytosol (4 mg of Protein) without Added Electron Donor
a
inhibitor (µM)
IMI-NO (% of control ( SD)
none menadione (200) menadione (500) cimetidine (500) chlorpromazine (500) promethazine (500) KCN (1000)
100a 78 ( 7 48 ( 6 50 ( 7 48 ( 3 28 ( 2 18 ( 3
1.5 ( 0.1 pmol/(min‚mg of protein).
factors (such as NAD(P)H) that could provide reducing equivalents to competing nitroreductases. No metabolites are detected when N-methylnicotinamide is not added to incubations. These data establish that AOX is capable of reducing IMI by two electrons to form IMI-NO and by six electrons to form IMI-NH2 (Figure 1). Diagnostic Inhibitors. IMI was incubated with RLC (1 mg protein), 2 mM N-methylnicotinamide, and either an AOX inhibitor or Me2SO control (Figure 3). Potassium cyanide (KCN) at 500 µM decreases IMI-NO formation by nearly 7-fold, and IMI-NH2 by 56-fold. AOX inhibitors, menadione and promethazine at 125 µM, also inhibit all IMI metabolite formation. Menadione is stronger with IMI-NO formation decreased 71-fold, compared to only 5.3-fold by promethazine. These data further support the proposal that AOX is a neonicotinoid nitroreductase. Because the physiological substrate(s) of AOX have not been identified (12), it is difficult to ascertain in vivo relevance from in vitro reactions with high levels of artificial electron donors. To this end, IMI nitroreduction was investigated without added electron donor (Table 2). IMI-NO is the only metabolite detected by HPLC-DAD from a 2-h incubation of IMI with 4 mg of RLC protein, giving a rate of 1.5 ( 0.1 pmol/min/mg cytosol. Following the trend exhibited in Figure 3, AOX inhibitors also inhibit IMI reduction without added electron donor (Table 2). Incubations containing 1000 µM KCN retain only 18% of control activity, while those with 200 and 500 µM menadione retain 78 and 48%, respectively. It is assumed that the difference in the inhibitory potency of menadione between Figure 3 and Table 2 is due to the much higher protein concentration, and thus nonspecific protein binding, of the latter experiments. Additional
320
Chem. Res. Toxicol., Vol. 18, No. 2, 2005
Figure 4. Effect of anaerobic atmosphere on IMI nitroreduction by rabbit liver cytosol (1 mg protein) with N-methylnicotinamide (2 mM) or by recombinant 2D6 Supersomes (10 pmol P450) with NADPH (1 mM). Data are mean ( SD, n ) 4. ND ) Not detected.
Dick et al.
Figure 6. Effect of protein concentration on IMI nitroreduction by rabbit liver cytosol. Inset relates total metabolites to cytosol concentration. Saturating concentrations of IMI (1.25 mM) and N-methylnicotinamide (10 mM) were incubated with 1-4 mg cytosolic protein. Data are mean ( SD, n ) 4. Table 3. Species Differences in IMI Nitroreduction by Liver Cytosol (3 mg of Protein) with N-Methylnicotinamide (2 mM) metabolites (pmol/(min‚mg of protein) ( SD)
Figure 5. Time course of IMI nitroreduction by rabbit liver cytosol (1 mg protein) with N-methylnicotinamide (2 mM). Representative data from one of three experiments.
AOX inhibitors, chlorpromazine and cimetidine, inhibit IMI reduction to nearly 50% at concentrations of 500 µM. These data suggest that AOX is a metabolizer of IMI under physiological conditions. Oxygen Sensitivity. The effect of an anaerobic atmosphere on IMI nitroreduction was compared for RLC with N-methylnicotinamide and 2D6 with NADPH (Figure 4). IMI-nitroreduction by RLC is moderately oxygen sensitive. Anaerobic incubations with 2 mM N-methylnicotinamide yield 4.9-fold more IMI-NH2, but less IMI-NO. It appears that reduction past IMI-NO is oxygen sensitive while reduction of IMI to IMI-NO is not. Recombinant 2D6 was previously found to catalyze the nitroreduction of IMI and not its 5-hydroxylation (8). Incubation of 2D6 with IMI and 1 mM NADPH under anaerobic conditions produces a 46-fold increase in IMI-NO formation over aerobic conditions. Similarly, IMI-NH2 is detected in anaerobic, but not in aerobic, incubations. Because recombinant 2D6 in Supersomes is used, incubations contain much higher IMI reducing activity per mg of protein (the normalization factor) than those with RLC. Thus, direct comparisons of specific activities between the two data sets may be misleading; only the aerobic/ anaerobic trend is important here. Time Course. Incubations of IMI with 1 mg RLC and 2 mM N-methylnicotinamide were stopped at various time points and analyzed by HPLC-DAD to construct a time course of IMI nitroreduction (Figure 5). At early points (e20 min) IMI-NO formation dominates with little IMI-NH2 detected. However, after this initial lag phase, formation of IMI-NH2 appears to increase at a linear rate. IMI-NO formation peaks at 60 min and begins to decrease at later time points.
no.b
speciesa
IMI-NO
IMI-NH2
total
1 2 3 4 5 6 7 8 9
rabbit monkey (Cynomolgus) human mouse cow rat dog chicken cat
10 ( 1 80 ( 6 26 ( 3 60 ( 10 40 ( 2 23 ( 1
