Chem. Res. Toxicol. 1990,3, 524-535
524
The Contribution of N-Oxidation to the Metabolism of the Food-Borne Carcinogen 2-Amino-3,8-dimethylimidato[ 4,5-f]quinoxaline in Rat Hepatocytes Robert J. Turesky,*J Ingrid Bracco-Hammer,l Jovanka Markovic,l Urs Richli,2 Anne-Marie Kappeler,2and Dieter H. Welti2 Divisions of Toxicology and Fundamental Science, Nestec L t d . , Nest16 Research Centre, Vers-chez-les Blanc, CH-1000 Lausanne 26, Switzerland Received March 15, 1990
The metabolism of 2-amino-3,8-dimethylimidazo[4,5-flquinoxaline, a potent bacterial mutagen and rodent carcinogen formed in low quantities in cooked meat and fish, was studied in freshly isolated rat hepatocytes. Ten metabolites were characterized by various spectroscopic methods. Sulfamate formation was the major route of metabolism in hepatocytes of untreated rats whereas ring-hydroxylated sulfuric and glucuronic acid conjugates were major metabolites in animals pretreated with the enzyme inducers Aroclor-1254,P-naphthoflavone,or isosafrole. The formation of a mutagenic metabolite through N-oxidation, 2-(hydroxyamino)-3,8-dimethylimidazo[4,5flquinoxaline (HNOH-MeIQx), was a n important route of metabolism in hepatocytes of pretreated animals. Its metastable derivative, the N-hydroxy-N-glucuronide, also was detected. T h e nitro derivative of MeIQx, a direct-acting bacterial mutagen, was readily detoxified by glutathione transferase, forming a conjugate where the thiol group of glutathione displaced the nitro moiety. Low but detectable levels of N-acetyltransferase activity were observed for MeIQx and sulfamethazine in hepatocytes. HNOH-MeIQx and 4- (hydroxyamino)biphenyl (HNOHABP), a recognized human carcinogen, displayed acetyl coenzyme A dependent DNA binding in hepatic cytosol assays. Sulfamethazine decreased the DNA binding of HNOH-MeIQx in hepatocytes, suggesting a competition for acetyltransferase. However, the binding of HNOHMeIQx to DNA in hepatocytes was independent of sulfotransferase since inhibitors of this enzyme, 2,6-dichloro-4-nitrophenol (DCNP) and pentachlorophenol (PCP), did not diminish DNA binding. In contrast, binding of HNOH-ABP to DNA was not decreased by sulfamethazine, but binding was diminished by both sulfotransferase inhibitors. From these inhibition experiments it appears that a major route of binding of HNOH-MeIQx to DNA in hepatocytes is mediated through 0-acetyltransferase while a significant portion of HNOH-ABP bound to DNA is catalyzed by sulfotransferase.
Introduction Many carcinogens must be metabolized in order to exert genotoxicity. A number of different routes of metabolism may occur; generally, one pathway leads to activation while the other metabolic pathways are routes of detoxification. Thus, the carcinogenic potency of a chemical is in part dependent upon the proportion of the dose which is transformed into the biologically active species. Heterocyclic aromatic amines are a class of carcinogens that are formed in meat and fish products prepared under typical household cooking practices (1-3). When fed as part of the daily diet in rodent carcinogenicity assays, these chemicals have been found to induce tumors at multiple sites (4). Studies conducted in vitro using microsomes from rodents have demonstrated that the initial activation step is N-oxidation by cytochrome P-450 ( 5 ) . Esterification with sulfate or acetate is believed to produce the ultimate carcinogenic species (5). Several of these chemicals have been found to be N-oxidized by human hepatic microsomes at rates comparable to that of 4-aminobiphenyl (6), a known human carcinogen, and therefore may be important factors in the etiology of human cancers.
* Correspondence should be addressed to this author.
2-Amino-3,8-dimethylimidazo[4,5-f]quinoxaline ( M ~ I Q x is) ~structurally representative of this class of genotoxins (1-3). MeIQx is rapidly absorbed from the gastrointestinal tract of rodents and transformed into a number of detoxified products by the liver including ring-hydroxylated glucuronide and sulfate conjugates (7-9). Detoxification reactions by direct conjugation to the exocyclic amine group also occur. Reactions include Nacetylation, N-glucuronidation, and the uncommon pathway of sulfamate formation (7-11). Although in vitro assays readily demonstrate formation of the mutagenic N-hydroxy metabolite (12),neither HNOH-MeIQx or its Division of Toxicology. Division of Fundamental Science. Abbreviations: MeIQx, 2-amino-3,8-dimethylimidazo[4,5-flquinoxaline; HNOH-MeIQx, 2-(hydroxyamino)-3,8-dimethylimidazo[4,5-flquinoxaline; N02-MeIQx, 2-nitro-3,8-dimethylimidazo[4,5-flquinoxaline; IQ, 2-amino-3-methylimidazo[4,5-flquinoline; MeIQ, 2amino-3,4-dimethylimidazo[4,5-flquinoline; 4-ABP, 4-aminobiphenyl; HNOH-ABP, 4-(hydroxyamino)biphenyl;HNOH-2-NA, N-hydroxy-2naphthylamine;2-AF, 2-aminofluorene; PCP, pentachlorophenol; DCNP, 2,6-dichloro-4-nitrophenol; SMZ, sulfamethazine; DTT, dithiothreitol; EDTA, ethylenediaminetatraaceticacid; PCB, polychlorinated biphenyl; pNF, p-naphthoflavone; TMS, tetramethylsilane; TSP, sodium 3-(trimethylsily1)tetradeuteriopropionate;EI-MS, electron impact mass spectrometry; FAB-MS, fast atom bombardment mass spectrometry; NOE, nuclear Overhauser effect; COSY, homonuclear correlation NMR spectroscopy.
Q893-228~/90/ 27Q3-Q524$Q2.5Q/Q0 1990 American Chemical Society
H e p a t o c y t e Metabolism of
MeZQx
metastable derivatives have been identified in vivo. This suggests either that very low levels of the N-hydroxy derivatives are formed in vivo or that they are not stable and rapidly decompose. Such findings are in contrast to those for other aromatic amines such as 4-ABP and 2-AF, where metastable derivatives of the N-hydroxylated derivatives have been detected (13-15). In the present study we investigated the contribution of metabolic activation through N-oxidation to the overall metabolism of MeIQx using an in vitro system that closely simulates the metabolism that occurs in vivo. The use of freshly isolated hepatocytes is ideal for such a study as a number of different metabolic pathways are operational (16). We also examined the effect of cytochrome P-450 inducers, the metabolic fate of HNOH-MeIQx in intact hepatocytes, and the role of sulfotransferase and 0-acetyltransferase, two enzymes that may be involved in binding of HNOH-MeIQx to DNA (5).
Experimental Procedures MeIQx and 2-14C-labeled MeIQx (50 mCi/mmol) were purchased from Toronto Research Chemicals. T h e radiochemical purity was 96% as determined by HPLC and liquid scintillation counting. 2-Nitro-3,8-dimethylimidazo[4,5-flquinoxaline was synthesized by the method of Grivas (17)(EI-MS accurate mass of M+ ion: measured m/z 243.0869, calculated 243.0756). 2(Hydroxyamino)-3,8-dimethylimidazo[4,5-~quinoxaline was prepared by the controlled reduction of the nitro compound with hydrazine using P d / C as a catalyst (18). T h e reaction was monitored by HPLC using the conditions described below. T h e reaction was terminated by removing the P d JC by centrifugation. An E1 mass spectrum confirmed the reduction to the hydroxylamine, showing a M+ ion at m/z 229.0956 (calculated 229.0964), a [M - O]+ ion at m/z 213, a characteristic [M - OH]+ ion a t m/z 212, a [M - HNO]+ ion a t m / z 198, and a [M - H,NO]+ ion a t m / z 197. One milliliter of tetrahydrofuran solution was diluted with 15 mL of 100 pM EDTA saturated with argon and applied to a Sep-Pak cartridge which had been prewashed with methanol followed by the EDTA solution. The cartridge was washed twice with 10 mL of EDTA solution to remove hydrazine, and the hydroxylamine was eluted with a minimum volume of DMSO. T h e hydroxylamine was used immediately or stored under an atmosphere of argon in liquid nitrogen. T h e purity of the hydroxylamine exceeded 90% as determined by HPLC with unreacted nitro derivative as the principal contaminant. Identity of HNOH-MeIQx also was confirmed by formation of the azoxy conjugate, 2-(phenylazoxy)-3,&dimethylimidazo[4,5-flquinoxaline (19) (EI-MS M+ ion: measured m / z 318.1164, calculated 318.1229). W-Acetyl-MeIQx was prepared by reacting 1 mg of MeIQx in 0.8 mL of pyridine with 0.2 mL of acetic anhydride a t 60 "C for 1 h. T h e mass spectrum displayed a M+ ion a t m / z 255.1117 (39%) (calculated 255.1120), a [M - CH,]+ ion a t m / z 240 (100%), and a [M - COCH3]+ ion a t m / z 212 (18%). T h e W-sulfamic acid of MeIQx was synthesized and characterized as previously described (1I). N-Acetylsulfamethazine was prepared by incubating 50 mg of sulfamethazine in 1mL of acetic anhydride for 1h at room temperature. The product was crystallized in 95% ethanol, and its identity was confirmed by mass spectroscopy. [2,2'-3H]-4-(hydroxyamino)biphenyl (119 Ci/mmol) was a generous gift from Dr. Fred Kadlubar, National Center for Toxicological Research, Jefferson, AR. The purity exceeded 95% as determined by HPLC. Animals. Male Sprague-Dawley rats (250-350 g) were obtained from Iffa Credo, L'Arbresle (France). T h e rats were pretreated with a single ip injection of Aroclor-1254 (500 mg/kg) in corn oil 5 days prior t o isolation of cells. Pretreatment with BNF (80 mg/kg) and isosafrole (150 mg/kg) was done by three sucessive i.p. injections in corn oil at 24-h intervals. Control animals also received three ip injections of corn oil. Animals were sacrificed 24 h following the final injection. Diethyl ether was used as the anaesthetic. Hepatocyte Metabolism Studies. Hepatocytes were isolated by collagenase perfusion (20) with final purification by Percoll treatment (21). The cell incubation medium was Krebs-Hensleit buffer containing serum albumin (2% w/v) with 0.2 m M me-
Chem. Res. Tonicol., Vol. 3, No. 6, 1990 525 thionine to ensure high levels of intracellular glutathione (22). Cell viability was 95% or greater as determined by trypan blue exclusion. Cells were maintained under an atmosphere of 95% O2 and 5% COz during the entire course of the experiment. Substrates were dissolved in DMSO, and final volumes did not exceed 0.3% of the cell incubation medium. Studies measuring formation rates of detoxified metabolites were conducted with 10 mL of cell suspension at 2.5 X 106 cells/mL with 35 pM MeIQx. The reaction was terminated by lysing 1 mL of suspension with 2 volumes of chilled ethanol followed by centrifugation. The pellet was resuspended in distilled-deionized water, precipitated a second time with ethanol, and removed by centrifugation. T h e pooled supematants were rotary evaporated to dryness and stored a t -80 "C. Appropriate control experiments demonstrated that the metabolites were stable and could be stored under these conditions prior to analysis. However, prolonged storage did result in a slow hydrolysis of the N-hydroxy-N-glucuronide, metabolite 6 (Chart I). Experiments designed for measuring HNOH-MeIQx were performed differently due to the instability of the metabolite. At the appropriate times 0.6 mL of cell suspension was added t o 0.4 mL of chilled methanol containing 1 m M DTT. T h e mixture was vortexed and immediately centrifuged for 20 s at 15000g. T h e supernatant was immediately analyzed by HPLC for HNOH-MeIQx. Recovery of radioactivity from the cell incubations by both cell lysis methods averaged 90% or better, irrespective of animal pretreatment with cytochrome P-450 inducers. Large quantities of HNOH-MeIQx were obtained by using 10-mL batches of cells and lysing with methanol (40% v/v) as above. Following centrifugation, the methanol supernatant was diluted with 50 mL of 100 pM EDTA, immediately applied to a C-18 Sep-Pak, and washed with 100 pM EDTA. The N-hydroxy metabolite was eluted with 3 mL of methanol, then concentrated by rotary evaporation at room temperature, and subjected to HPLC. HNOH-MeIQx was collected and derivatized to form the azoxy conjugate as above. Large-scale quantites of detoxified metabolites were obtained for spectroscopic purposes by incubating 40 mL of 2.5 x lo6 cells/mL with MeIQx for 3 h. Metabolites were recovered as above, except that lipids were removed by extraction with hexane. Hepatocytes could be maintained for at least 3 h without noticeable changes in rate of MeIQx metabolite formation or 0-deethylase activity using ethoxycoumarin as a standard (23). Estimates of metabolite formation were based upon the amount of metabolites recovered in the cell lysate supematant. Studies designed to examine the metabolic fate of HNOHMeIQx and mechanism of DNA binding were conducted as above. PCP, DCNP, or SMZ were added to the cell suspensions 5 min prior to the addition of 10 pM HNOH-MeIQx or HNOH-ABP. After a 15-min incubation, the cells were centrifuged a t 500g for 1 min at 4 "C. T h e pellet was washed with chilled phosphatebuffered saline and centrifuged a second time. T h e pooled supernatants containing extracellular metabolites were analyzed by HPLC while the DNA was isolated as described below. Covalent B i n d i n g t o D N A a n d Protein. DNA was isolated from the incubation medium by the method described by Corbett et al. (24) except that RNA was removed by treatment with RNase A and RNase T1. Then the DNA was treated with proteinase K followed by organic extractions with phenol reagent (phenol, chloroform, isoamyl alcohol, and 8-hydroxyquinoline, 25/24/1/ 0.05) and chloroform and then precipitated with 2 volumes of ethanol following addition of NaCl (0.25 M final salt concentration). DNA concentration was measured by a modified diphenylamine method (25). Protein was isolated from the cell incubation medium as described by King et al. (26). Nonspecific binding was determined by the addition of radiolabeled substrates t o untreated hepatocytes after the cell lysis step. Statistical Analysis. Statistical significance of DNA binding was determined with the Student t test for the analysis of matched pairs using the raw data. Instrumentation. T h e HPLC system was either a Hewlett Packard 1090M system containing a diode array detector or a Varian 5500 system. T h e UV absorbance was monitored at 275 nm. A Supelco (2-18reverse-phase column (4.6 mm i.d. X 25 cm, 5 pm particle size) was used for kinetic measurements of detoxified metabolites. The solvent conditions were 9.5% methanol in 50 mM ammonium acetate, p H 5.8, for 10 min followed by a linear gradient to 16% methanol at 45 min followed by a linear gradient
526 Chem. Res. Toxicol., Vol. 3, No. 6, 1990
Turesky et al.
Chart I. Proposed Structures, Names, and Abbreviations of Isolated Metabolitesa
n,cl::;~iN'i
MelQx
I
-
-1
-5
2
oso;
oso;
Metabolites: 1, N-~8-(hydroxymethyl)-3-methylimidazo[4,5-flquinoxalin-2-yl)sulfamic acid, 8-CH2OH+NSO;; 2, 2-amino-8-(hydroxymethyl)-3-methylimidazo[4,5-flquinoxalin-5-yl sulfate, 8-CH20H+5-OS0,-; 3, 2-amino-5-(~-l-glucosiduronyloxy)-3,8-dimethylimidazo[4,5flquinoxaline, 0-G1; 4, N2-(~-1-glucosiduronyl)-2-amino-3,8-dimethylimidazo[4,5-flquinoxaline, N-GI; 5 , 2-amino-3,8-dimethylimidazo[4,5flquinoxalin-5-yl sulfate; 5-OSO;; 6, ~-(~-1-glucosiduronyl)-N-hydroxy-2-amino-3,8-dimethylimidazo[4,5-flquinoxaline, HO-N-GI; 7, N ~3,8-dimethylimidazo[4,5-flquinoxalin-2-yl)sulfamic acid, NSO,; 8, N-acetyl-2-amino-3,8-dimethylimidazo[4,5-flquinoxaline, N-acetyl; 9, 2-(hydroxyamino)-3,8-dimethylimidazo[4,5-flquinoxaline, HNOH-MeIQx; 10, 2-(glutathion-S-yl)-3,8-dimethylimidazo[4,5-flquinoxaline, GS. to 90% methanol a t 65 min at a flow rate of 1 mL/min. The analysis of NHOH-MeIQx was performed with the same column, but the solvent was held a t 50 mM ammonium acetate, p H 6.8, containing 100 pM EDTA and 37.5% methanol. At 15 min, the percentage of methanol increased linearly to 90% a t 25 min where it was held for 5 min. A semipreparative Supelco C-18 column was used (10mm i.d. x 25 cm, 5 pm particle size) for large-scale purification of detoxified metabolites. The solvent conditions were as above except that the flow rate was 3 mL/min. Metabolites which were isolated for spectroscopy were further purified by rechromatography using the same conditions as above except for an increase in the p H of the ammonium acetate buffer to 6.8. Synthetic products were purified by using the analytical C-18 column with a linear gradient of 5% methanol in 50 mM ammonium acetate, p H 6.8, to 100% methanol over 40 min a t 1 mL/min. Radioactivity measurements were made by either liquid scintillation counting using a LKB 1219 Rackbeta counter or a Berthold LB 506 C-1 radioactivity monitor. Enzyme Assays. Metabolites were characterized for enzyme stability against P-glucuronidase, arylsulfatase, or y-glutamyltranspeptidase as previously described except that the amount of sulfatase was increased to 150 units and the incubation was for 8 h (7). Acid hydrolysis of metabolites was performed a t 60 "C in 1 N HC1 for 1 h. Assay for Biological Activity. The induction of umu gene expression as expressed by @galactosidase activity in Salmonella typhimurium TA1535/pSK1002 was used to examine biological activity (27). Spectroscopy. Proton NMR spectra were recorded on a 360-MHz Bruker spectrometer with a 5-mm 'H probehead. The
sample temperature was 21.0 f 0.5 "C. Typical parameters for basic one-dimensional spectroscopy were spectral width 7576 Hz, data table size 64K, pulse angle ca. 65", and pulse interval 14.3 s. A 0.05-Hz exponential line broadening was applied. Several hundred to several thousand transients were acquired, depending upon sample quantity (typically 20-500 pg of MeIQx equivalent). The samples were prepared in DMSO-& CD30D (sulfamates), or D20 (glutathione) with T M S or TSP as an internal standard. The nuclear Overhauser effect (NOE) difference spectra (28) were acquired with the standard Bruker software which cycled through a list of irradiation frequencies several times in order to reduce long-term drifts. Typically, a relaxation delay of 12 s and ca. 2.5 s of homonuclear gated irradiation with a power level just sufficient to suppress the irradiated signals preceded each transient. Acquisition time for one free induction decay was 4.19 s; the spectral width was 3906 Hz and the data table size 32K. Several hundred transients were usually acquired per irradiation frequency. The homonuclear correlation experiments (COSY) (29)using two 90" pulses (5.6 ps) were typically acquired with a spectral width of 2660 Hz and 2K data points for each of the 512 experiments, using a relaxation delay of 7.4 s and a full 16-phase cycle. Zero filling was applied in the F1 direction only, and a n unshifted sine filter was used in both dimensions before Fourier transformation. The spectra were plotted in the magnitude mode. Mass spectra were acquired with a Finnigan/MAT 8430/SS300 mass spectrometer. E1 spectra were obtained by direct sample introduction a t 70 eV with an EI/CI source at 200 "C. Positive and negative ion FAB-MS were obtained with the same source a t 60 "C using glycerol or diethanolamine as a matrix. Accurate
Chem. Res. Toxicol., Vol. 3, No. 6,1990 527
Hepatocyte Metabolism of MeZQx
I 1
7,
0
5
10
I5
20
25
30
35
10
45
50
MelQx
55
60
Time [min]
B
l
7
I
54-
"
Wavelength lnml
20 E 40
-i o ^r 0
5
10
15
M
25
30
35
40
45
50
55
ffi
50
100
150
WZ
200
250
300
350
Time [min]
Figure 1. HPLC analysis of metabolites isolated (A) from hepatocytes of control rats and (B) from hepatocytes of rats pretreated with PCB. Identified metabolites are numbered in order of increasing retention time. (Metabolite 6, the N-hydroxy-Nglucuronide of MeIQx, has partially hydrolyzed due to prolonged sample storage prior to analysis.)
mass measurements were performed with a resolution of about 3000, and perfluorokerosene was the reference compound in E1 mode and the used matrix in FAB mode. In E1 mode the accurate masses were measured by the data system whereas in FAB mode the mass of the ions was manually checked by using a peak matching unit. Infrared spectra were recorded with a Perkin Elmer 1600 series Fourier transform infrared spectrometer, and samples were prepared as KBr pellets using approximately 30 pg of metabolite and 15 mg of KBr.
Results The HPLC analysis of metabolites produced from hepatocytes of control and PCB-pretreated rats is shown in Figure 1. Animal pretreatment with inducers greatly increased the number of metabolites formed. The metabolites which have been identified are numbered in order of increasing retention time. They have been characterized spectroscopically, through enzyme hydrolysis assays and acidlbase stability, and by cochromatography with synthesized standards or with metabolites which were previously isolated from bile of rats (7). Their chemical names, abbreviations and structures are shown in Chart I. The chemial characterizations for several of these metabolites are in excellent agreement with our previous findings (7, 111, except for metabolite 5, 2-amino-3,8-dimethylimidazo[4,5-flquinoxalin-5-yl sulfate, which is discussed below. The short-lived HNOH-MeIQx metabolite also was detected. The metabolite has an identical retention time and possesses an UV spectrum indistinguishable from that of the synthetic product. The mass spectrum of the azoxy conjugate formed from reaction of
302 316 I
50
100
150
miz
200
250
300
350
Figure 2. HPLC and UV spectrum of HNOH-MeIQx produced
from hepatocytes and electron impact mass spectrum of 2(phenylazoxy)-3,8-dimethylimidazo[4,5-~quinoxaline.
the metabolite or synthetic HNOH-MeIQx and nitrosobenzene also is in excellent agreement (Figure 2). Several other metabolites were detected, particularly from pretreated animals. However, these metabolites were not formed in amounts sufficient for spectroscopic analyses. The rate of disappearance of MeIQx and the time course of formation of metabolites are shown in Figure 3. Disappearance of MeIQx was most rapid with hepatocytes from animals pretreated by PCB, followed by animal pretreatment with PNF and isosafrole, and lastly from the control cells. A concentration of 35 pM MeIQx was sufficient to saturate all cells except those from animals pretreated with PCB as judged by the linear decrease in MeIQx and the constant rate of formation of detoxified metabolites for at least 30 min. The sulfamate was found to serve as a substrate for cytochrome P-450 in control hepatocytes. There was a noticeable increase in the formation of 8-(hydroxymethy1)-MeIQx sulfamate during hours 2-3, while the rate of formation of the sulfamate decreased. The lifetime of HNOH-MeIQx was short in all hepatocytes except in the control cells where there was an accumulation. The accumulation would have to be attributed to a rate of N-oxidation of MeIQx which exceeded the rate of decomposition of HNOH-MeIQx. The initial rates of formation for the principal metabolites are presented in Table I. Hepatocytes from control cells had a low capacity for cytochrome P-450 metabolism. The major routes of metabolism were sulfamate formation and N-acetylation, while ring-hydroxylated sulfate and
528 Chem. Res. Toxicol., Vol. 3, No. 6, 1990
Turesky et al.
Rate of Disappearance of MelQx in Hepatocytes 1
I"
0
30
120
60
1
180
minY t e I
Corn Oil Pretreatment
BNF Pretreatment
looo0lf
/
5.000l
5 000
0
30
0
60
120
180
0
30
60
minutes
120
180
minutes
PCB Pretreatment
lsosafrol Pretreatment
10 000
30
0
180
120
60
0
30
120
60
ino
1260
1000
760
5oQ
250
0 0
15
30 mlnuleb
45
60
minutes
F i g u r e 3. Time course for the disappearance of MeIQx and formation of major metabolites in hepatocytes. .Average of 4 experiments for detoxified metabolites and 2 experiments for HNOH-MeIQx.
Chem. Res. Toxicol., Vol. 3, No. 6, 1990 529
Hepatocyte Metabolism of MeIQx
Table I. Initial Rates of Metabolite Formation of MeIQxa,b
pretreatment corn oil isosafrole BNF PCB
3 O-GI 70 f 30 1270 f 260 2340 f 670 4080 f 395
4 N-GI 50 f 10 115 f 45 220 f 95 595 f 180
pmo1/[(2.5 X lo6 ce1ls)alO min] 5 6 7 NSO15-0S0,HO-N-GI 100 f 45 nd 645 f 320 95 f 25 1675 f 170 2105 f 280 255 f 95 1125 f 160 2930 f 495 5650 f 1375 2175 f 300 1625 f 510
8 N-acetyl 175 f 50 nd nd nd
macromolecular binding 9 pmol/mg of pmol/mg HNOH-MeIQx Drotein of DNA 90-130 43f6 33f3 1500-2000 123 f 14 42 f 3 3775-5400 145 f 22 102 f 18 4675-5750 110 f 14 53 f 6
"2.5 x IO6 cells/mL were incubated with 35 pM MeIQx. Average f SD of 4 animals except for HNOH-MeIQx analysis, which is the average of 2. The amount of HNOH-MeIQx which would be produced a t 10 min was estimated by extrapolating the amount formed following a l-min incubation from induced animals and by linear regression of the slope for control animals. Detoxified metabolite formation was found to be linear for 30 min for all animal pretreatments except for PCB. Protein and DNA modification were analyzed after a 3-h incubation. (nd = not detected.) *Rate of N-acetylsulfamethazine formation was 480 f 370 pmo1/[(2.5 X lo6 cells).lO min] under the same incubation conditions.
Table 11. 360-MHz 'H NMR Spectral Parameters for MeIQx and Metabolites",' metabolites proton assignment aromatic
MeIQx 8.78, s, 1, H-7 7.88, d, 1, H-4; J = 8.8 7.76, d, 1, H-5; J = 8.8 7.48, s, 2
NH2 NH NCHS 3.71, s, 3 8-CH3 2.74, s, 3 8-CHzOH CIH (glucuronide)
2 3 O-G1 8-CHzOH+5-OS0c 8.63, S, 1, H-7 8.83, s, 1, H-7 7.50, S, 1, H-4 7.80, S, 1, H-4
4 5 6 N-G1 5-OSOc HO-N-G1 8.68, S, 1, H-7 8.81, S, 1, H-7 8.71, S, 1, H-7 7.78, d, 1, H-4; J = 8.7 7.90, S, 1, H-4 7.89, d, 1, H-4; J = 8.9
7.61, d, 1, H-5; J = 8.7 6.40, s, 2 3.60, s, 3
6.40, s, 2
7.74, d, 1, H-5; J = 8.9 7.88, s, 2d
3.62, s, 3 2.69, s, 3
7.55, d, 1; J = 9.2 3.72, s, 3 2.72, s, 3
5.18, d, 1; J = 7.7
5.22, "t", 1; "s" = 8.7c
3.67, s, 3 2.75, s, 3
3.91, s, 3 2.73, s, 3
4.79, d, 2; J = 6.0b 5.07, d, 1; J = 8.4
"Chemical shifts are reported in ppm downfield from TMS, and absolute values of coupling constants J are given in hertz (f0.2 Hz because of digitization). The solvent is DMSO-d6. The assignment of the 4 and 5 protons of metabolites 4 and 6 is based on shift analogy with the parent compound only, since NOE experiments were not performed. The spectra on the two sulfamate derivatives, metabolites 1 and 7, are consistent with previous findings (7). For metabolite 3, no NOE could be observed upon irradiation of the 3-N-methyl; therefore, the assignment of the signal at 7.50 ppm is not unequivocal. bSeen as a doublet, coupling to (CH,)OHgroup. OH signal present a t 5.66 ppm (t, J = 6.0). cFollowing a D,O exchange, this signal collapsed to a doublet. The "s" is a mean value between 9.2 and 8.2 Hz. The "t" is actually a dd, hut the lines in the center are not resolved. dBroad signal; both shift and line width are temperature dependent. eMetabolite 8, N-acetyl-MeIQx, was not characterized by 'H NMR. The high-resolution EI-MS and UV spectra were indistinguishable from those of the synthetic derivative.
glucuronide conjugates were major routes of metabolism in hepatocytes from pretreated animals. The initial rate of formation of the mutagenic N-hydroxy metabolite was as great as the rates of formation of the detoxified ringhydroxylated glucuronide and sulfate conjugates and even greater than the rate of sulfamate formation in hepatocytes from pretreated animals. The metastable metabolite of the hydroxylamine, (HO-N-G1)-MeIQx,was most predominant in hepatocytes from animals pretreated with PCB but also could be detected in hepatocytes obtained following other pretreatments. Additionally, the rates of formation of the sulfamate and the W-glucuronide were increased by animal pretreatment. Higher levels of protein and DNA modification were observed in hepatocytes of pretreated animals but were less than might be expected for a 20-50-fold increase in the initial rate of formation of the reactive HNOH-MeIQx metabolite (Table I). However, the rate of HNOH-MeIQx formation is not constant in hepatocytes of pretreated animals (Figure 3), and the amount formed in control hepatocytes during the 3-h incubation may well approach that formed in hepatocytes from pretreated animals. The 'H NMR spectral parameters of MeIQx and identified metabolites are presented in Table 11. The two ring-hydroxylated sulfates and the two N-glucuronic acid conjugates are described in detail. The parent compound contains the isolated 7 proton a t 8.78 ppm, a set of doublets, assignable to the 4,5 protons at 7.88 and 7.76 ppm, two isolated singlets upfield at 3.71 and 2.74 ppm
representing the NCH3 and 8-CH3, and the D20-exchangeable NH, at 7.48 ppm. An NOE difference experiment resulted in a 10.4% increase in the signal height of the H-7 when the 8-CH3 group was irradiated, while irradiation of the NCH, group yielded an NOE of 8.2% for the proton at 7.88 ppm and 0% for the proton a t 7.76 ppm. This finding permits unambiguous assignment of the 4 proton at 7.88 ppm and the 5 proton at 7.76 ppm. In the spectrum of MeIQx-hulfate, metabolite 5, the two methyl groups, the amine, and the H-7 are evident. However, the AB system of the 4 and 5 proton is missing and only a single resonance is present, indicating oxidation. The protons on the NH, group integrated to 2 relative to the aromatic protons, indicating that the site of conjugation is to the oxygen atom and not to the exocyclic nitrogen group. When the NCH3 group was irradiated, an NOE of 6.9% was observed for the aromatic proton singlet at 7.90 ppm. This assigned the 5-position and not the 4-position as the site of oxidation. These results agree with our previous findings (7). However, on the original analysis of this metabolite obtained from bile, there was thought to be an acetyl function based upon HPLC-MS and NMR spectroscopy (7). Upon further purification of the original biliary product we found parameters that were chromatographically and spectroscopically identical with this metabolite, and the acetyl function, regretably, was an incorrect assignment attributable to an artifact. The identity of a sulfate moiety was confirmed by enzyme hydrolysis assays, infrared spectroscopy, and FAB-
530 Chem. Res. Toxicol., Vol. 3, No. 6, 1990
Turesky et al.
1 H-7
j:: -T
x 8.5
8.0
75
70
6.5
A
5.5 F?M
6.0
5.0
4.5
4h
3:5
3b
4.0
3.5
3.0
7
/
J L 83
00
25
7b
6.5
55 PFm
6.0
5.0
4.5
4.0
3.5
Figure 4. lH NMR spectra of W-(~-1-glucosiduronyl)-2amino-3,8-dimethylimidazo[4,5-flquinoxaline. The inset shows of the triplet signal of the CIH glucuronide to a the collapse . - e nniiniet iinnn I 1-1 I e w n n n u e
. ..
i
3.0
.
MS. The metabolite was slowly hydrolyzed by arylsulfatase, but not by j3-glucuronidase or y-glutamyltranspeptidase. An IR spectrum showed broad absorption bands a t 1020-1100 cm-l, indicative of a sulfoxide group (30). A strong absorption band also was observed in this region of the spectrum for MeIQx sulfamate, but not for MeIQx. The negative ion FAB mass spectrum of the sulfate derivative showed a [M - HI- ion a t m / z 308.0454 (checked by peak matching), the [(M - H) - SO3]-fragment ion at m / z 228, and an ion a t m/z 80 corresponding to [SO3]-which support the proposed structure. The 'H NMR of metabolite 2, 8-(hydroxymethyl)-3methylimidazo[4,5-flquinoxalin-5-y1 sulfate, reveals that the amine group as well as the 4 and 7 protons are present, but the 5 proton is missing (Table 11). An NOE experiment gave about an 8% increase in signal height for the aromatic proton at 7.80 ppm upon irradiation of the NCH3 group. The metabolite could be hydrolyzed with arylsulfatase, indicating a sulfate conjugate. The &methyl has been replaced by a methylene group which is shifted downfield to 4.79 ppm and present as a doublet. The coupling to the D20-exchangeable OH proton seen as a triplet a t 5.66 ppm demonstrates that sulfate conjugation must be to the oxygen a t the C-5 position and not to the oxygen atom of the methylene group. The negative ion FAB mass spectrum supports this structure. Accurate mass measurements of the [M - HI- ion (mlz 324.0403) were in agreement with the expected atomic composition (CllH10N505S). Characteristic fragment ions [M - H SO3]-at m/z 244 and [SO3]- a t m / z 80 were detected and indicate the presence of an SO3 group. The 'H NMR spectra of the two N-glucuronide conjugates are shown in Figures 4 and 5. All of the aromatic protons are present in the spectrum of MeIQx-N2-glucuronide (metabolite 4, Figure 4). The assignment of the 4 and 5 protons in Table I1 is based on shift analogy with the parent compound only, since an NOE experiment was not done for this metabolite, and the assignment is therefore not unambiguous. The NH, signal was diminished in intensity to one proton and found a t 7.55 ppm (doublet, J = 9.2 Hz). A COSY experiment showed that this signal is coupled to the anomeric proton of the glucuronide at 5.22 ppm, seen as a "triplet" ("J" = 8.7 Hz) due to the additional coupling with the glucuronide H-2 proton. This "triplet" collapsed to a doublet (J = 8.2 Hz) upon addition of D20(see spectrum inset of Figure 4) and was therefore an unresolved doublet of doublets, the "Svalue being an average of the two coupling constants involved. These findings prove the 2-N substitution by a p-glucuronic acid (31). From the COSY spectrum we know that
14-7
8.5
ClH
0.0
75
70
6.5
6.0
5.5
5.0
4.5
PPM
Figure 5. lH NMR spectra of W-(j3-l-glucosiduronyl)-Nh y d r o x y - 2 - a m i n ~ 3 , 8 d e ~ y ~ ~ [ 4 , 5 f(A) l qbefore e and (B) following D20exchange. The inset displays the coupling pattern of the glucuronic ring following rechromatography and addition of D20.
the signal at c a 3.64 ppm belongs to one of the glucuronide protons, while the other three are hidden under the dominating water signal at 3.33 ppm. They become visible after addition of D20, but the system is so strongly coupled that it could not be easily analyzed. The spectrum of (HO-N-GI)-MeIQx displays all of the aromatic and methyl protons of MeIQx, and conjugation must be to the exocyclic amine group (Figure 5). The assignment of the 4 and 5 protons in Table I1 is again tentative, based on shift analogy with MeIQx. The glucuronide moiety also is evident with the anomeric proton a t 5.07 ppm and three more signals between 3.15 and 3.65 ppm, while the fourth (3.38 ppm) only can be seen only upon shifting the HDO signal by D 2 0 exchange. After another HPLC separation without acetate buffer and D20 exchange, a clearly resolved spectrum could be obtained (see inset in the upper part of Figure 5). This allowed the assignment of the glucuronide protons by a COSY experiment and confiimed the identity of the &glucuronide, since all the vicinal couplings were very similar (5.11 ppm, d, J = 8.8 Hz, Cl-H; 3.59 ppm, t, J = 9.0 Hz, C,-H; 3.52 ppm, d, J = 9.8 Hz, C5-H; 3.38 ppm, t, J = 8.9 Hz, C3-H; 3.19 ppm, dd, J,, = 9.3 Hz, C,-H; coupling constants f0.2 Hz). Although an OH resonance on the exocyclic nitrogen of MeIQx could not be clearly identified, the glucuronide may be bound as a C1-N(0H)Ar or by a C1-O-NH-Ar bond. The identification of metabolite 6, (HO-N-Gl)MeIQx, as having a Cl-N(0H)Ar linkage and not a C10-NH-Ar linkage is based upon chemical stability, chemical reactivity and FAB-MS. The metabolite was stable in methanol or distilled water but could be hydrolyzed in acid to generate HNOH-MeIQx. Following an 18-h hydrolysis in 0.1 N HCl a t 37 "C approximately 70% of the metabolite was recovered as HNOH-MeIQx but nearly 30% still remained as the unhydrolyzed conjugate. This stability is suggestive of a C1-N[OH]Ar linkage, as C1-ONH-Ar glucuronide conjugates of N-hydroxy aromatic amines have been found to be highly unstable even at -20 "C (14). Metabolite 6 tested positive with pentacyanoamine ferroate, indicating a free NHOH group (141, and treatment with K3Fe(CN), resulted in formation of a purple product which is consistent with nitrone formation (32). The UV spectrum of the oxidized product obtained on line with HPLC also was consistent with nitrone formation. Maxima for (HO-N-G1)-MeIQxwere 275 (100%) and 338 (15%) nm while the oxidized product had maxima at 278 (50%),310 (85%), and 340 (s) nm. Unfortunately,
Hepatocyte Metabolism of MeIQx
Chem. Res. Toxicol., Vol. 3, No. 6, 1990 531
Table 111. Metabolic Fate of HNOH-MeIQx in Hepatocytes and Influence of DCNP, PCP, and SMZ on Metabolism and DNA Binding“ 90 90 distribution in supernatant 70relative binding 1 (pmol/mg of DNA) recovery in 8CH,OH+ 6 7 10 HNOHHNOHNSOc HO-N-GI NS03substrate inhibitor supernatant GS MeIQx MeIQx ABP HNOH0 67.7 f 4.6 1.6 f 0.5 32.1 f 4.2 9.6 f 0.5 42.5 f 1.9 100 (38 f 1.6 f 0.3 100 (326 f MeIQx 20) 221) HNOH10 pM DCNP 72.3 f 5.6 0.6 f 0.1 2.6 f 1.0 11.9 f 0.3 10.0 f 0.6 60.7 f 2.8 122 f 11 80 f 14 MeIQx HNOH100 pM DCNP 70.5 f 5.4 0.2 f 0.1 2.8 f 1.0 1.8 f 0.2 9.7 f 0.3 69.0 f 3.8 106 f 11 58 5b MeIQx HNOH500 pM DCNP 70.4 f 3.5 0.1 f 0.1 3.0 f 0.6 0.6 f 0.4 11.4 f 0.8 69.2 f 2.8 109 f 18 not tested MeIQx 64.5 f 5.2 1.4 f 0.4 1.8 f 0.3 31.2 f 0.1 10.1 f 0.1 42.9 f 8.2 99 f 15 89 21 10 pM PCP HNOH: MeIQx 68.3 f 4.3 0.7 f 0.1 2.3 f 0.7 17.3 f 0.8 9.7 f 0.6 55.6 f 2.5 123 f 11 58 f 14b HNOH100 pM PCP MeIQx HNOH500 pM PCP 71.0 f 6.0 0.1 f 0.1 2.7 f 0.8 3.3 f 0.5 10.6 f 0.3 70.2 f 2.3 123 f 43 not tested MeIQx HNOH50 pM SMZ 68.7 f 9.5 1.4 f 0.3 3.4 f 0.5 33.0 f 1.7 7.6 f 1.2 42.3 f 2.8 95 f 18 109 f 2 MeIQx HNOH5000 pM SMZ 73.1 f 9.1 0.9 f 0.2 0.9 f 0.1 36.9 f 1.4 8.6 f 0.6 38.6 f 2.3 57 f 15b 105 f 8 MeIQx MeIQx 0 72.9 f 10.5 0.5 f 0.3 nd 9.1 f 1.8 nd 75.9 f 3.5 7 (3 f 1) ~
* *
One milliter of 2.5 X lo6 cells was incubated with 10 pM MeIQx, HNOH-MeIQx, or HNOH-ABP for 15 min a t 37 O C . Inhibitors were added 5 min prior to addition of substrates. n = 5 animals f SD. P values
