Chem. Res. Toxicol. 1990, 3, 8-16
8
Metabolism of Methoxychlor by Hepatic P-450 Monooxygenases in Rat and Human. 1. Characterization of a Novel Catechol Metabolitet David Kupfer,*J William H. Bulger,t and A n t h o n y D. Theoharideso T h e Worcester Foundation for Experimental Biology, Shrewsbury, Massachusetts 01545, and Walter Reed A r m y Institute of Research, Washington, DC 20307-5100 Received M a y 16, 1989 Previous investigations demonstrated that the incubation of the chlorinated hydrocarbon with rat liver microsomes pesticide methoxychlor [l,l,l-trichloro-2,2-bis(4-methoxyphenyl)ethane] generates phenolic estrogenic metabolites. The current study shows that the incubation of liver microsomes from untreated and phenobarbital-treated rats and human donors, in the presence of NADPH, yields three phenolic metabolites. Identification of the metabolites was achieved by TLC, HPLC, GC/MS, and LC/MS and by hydrodynamic voltammetric analysis. These metabolites were identified as the mono- and didemethylated phenolic derivatives (mono-OH-M and bis-OH-M, respectively) and as a novel trihydroxy derivative (tris-OH-M). The tris-OH-M was demonstrated to be a catechol [ l,l,l-trichloro-2-(4-hydroxyphenyl)-2-(3,4-dihydroxyphenyl)ethane]. Furthermore, the tris-OH-M becomes radiolabeled by [ m e t h y L 3 H 3 ] - S adenosylmethionine (SAM) in a reaction catalyzed by catechol 0-methyltransferase (COMT), indicating that tris-OH-M behaves like a catechol. Incubation of the monohydroxy metabolite with liver microsomes from phenobarbital-treated rats (PB microsomes) yields the dihydroxy and the trihydroxy metabolites. Furthermore, the time course of methoxychlor metabolism by PB microsomes demonstrated a rapid appearance and disappearance of the monohydroxy metabolite with the subsequent formation of the dihydroxy and trihydroxy metabolites. On the basis of these findings, it is proposed that the metabolic route of methoxychlor by monooxygenases involves sequential demethylations t o the dihydroxy derivative and a subsequent ring hydroxylation.
Introduction Methoxychlor [l,l,l-trichloro-2,2-bis(4-methoxyphenyl)ethane] is a broad-spectrum pesticide that is being used as a substitute for certain pesticidal functions of DDT, which has been largely banned in the industrially developed countries. The favorable features of methoxychlor are its low overt toxicity and its relatively short half-life in the biosphere in general and in mammals in particular, classifying methoxychlor as a biodegradable pesticide (1-6). However, there are certain features of methoxychlor that appear undesirable. The commercial pesticidal preparations of methoxychlor contain numerous contaminants, some of which are estrogenic (7, 8 ) , and though pure methoxychlor is not estrogenic per se, its metabolites are estrogenic (7-9, 12, 13). Most recently, methoxychlor was shown to induce estrogen-like changes in behavior and reproductive tract and to inhibit fertility in the female rat (26, 27). Additionally, methoxychlor undergoes oxidative metabolic activation by the hepatic cytochrome P-450 monooxygenases, resulting in covalent binding to microsomal proteins (10, 11). Previous studies in mice and rats indicated that methoxychlor is metabolized by mono- and didemethylation (2, 12, 13). However, the metabolites in the rat were not rigorously characterized, and identification of the highly polar metbolite(s) (12) was not attempted. The current study identifies the three phenolic metabolites of methoxychlor, which are formed by liver microsomes from unA preliminary account of this work was presented at the FASEB Meeting in New Orleans, LA, March 19-23, 1989 (30). The Worcester Foundation for Experimental Biology. 8 Walter Reed Army Institute of Research.
*
0893-228x/90/2703-0008$02.50/0
treated and phenobarbital-treated rats and humans. The sequence of the metabolic transformation of methoxychlor by the hepatic monooxygenases is also examined.
Materials and Methods EDTA disodium salt, NADPH, glucose 6-phosphate,and glucose-6-phosphate dehydrogenase were from Sigma Chemical Co. (St. Louis, MO). Phenobarbital sodium and reagent-grade solvents were purchased from Mallinckrodt (St. Louis, MO). Scintillation cocktails were Insta-Gel from Packard (Downers Grove, IL) and Liquifluor from New England Nuclear (Boston, MA). All other chemicals were of reagent-grade quality. [ring-UL-'*C]Methoxychlor (5.85 mCi/mmol) was obtained from Sigma Chemical Co. (St. Louis, MO) and was used at the original specific activity or was diluted (as indicated in text) with radioinert methoxychlor obtained from Chem Service (West Chester, PA). Routinely, the [14C]methoxychlorwas purified to a radiochemical purity of 99+% as previously described (12). [metho~y-~H]Methoxychlor (4.4 Ci/mmol) was obtained by custom synthesis from New England Nuclear (Boston, MA) and was used as described in the text. Catechol 0-methyltransferasefrom porcine liver was obtained from Sigma Chemical Co., and [meth~l-~H]-S-adenosyl-~methionine (11.2 Ci/mmol) was from New England Nuclear. 14C-Labeled l,l,l-trichloro-2-(4-methoxyphenyl)-2-(4hydroxypheny1)ethane (['4C]mono-OH-M)' and 14C-labeled Abbreviations: TLC, thin-layer chromatography; HPLC, high-performance liquid chromatography; GC/MS, gas chromatography mass spectrometry; CI-GC/MS,chemical ionization mode GC/MS; L /MS, liquid chromatography/massspectrometry; LC/EC, liquid chromatography/electrochemical detection; mono-OH-M, l,l,l-trichloro-2-(4hydroxyphenyl)-2-(4-methoxyphenyl)ethane;bis-OH-M, IJJ-trichloro2,2-bis(4-hydroxyphenyl)ethane; tris-OH-M, l,l,l-trichloro-2-(4hydroxyphenyl)-2-(3,4-dihydroxyphenyl)ethane;COMT, catechol 0methyltransferase; SAM, S-adenosyl-L-methionine; PB microsomes, liver microsomes from phenobarbital-treated rats; control microsomes, liver microsomes from untreated rats or receiving the same regimen of the vehicle as the PB-treated rats.
0 1990 American Chemical Society
(i
Methoxychlor Metabolism by Mammalian P-450
~,l,~-trichloro-2,2-bis(4-hydroxyphenyl)ethane ([14C]bis-OH-M) were prepared by incubating [14C]methoxychlorwith control rat liver microsomes for a short incubation time for enriching the mono-OH-M, as described below under Incubation Procedure. These compounds were puified by reverse-phase HPLC, using a Whatman 10-pm C8 (RAGI) column and a mobile phase of 60% CH3CN/H,0 for the bis-OH-M and 45% CH3CN/H20 for the mono-OH-M, a t 2 mL/min. The fractions of interest, detected by UV a t 228 and 280 nm and by radioactivity, were collected, and after evaporation under a stream of nitrogen the residue was subjected to an additional HPLC purification, using a 10-pm C8 analytical column (Whatman) with the above mobile phases at 2.0 mL/min. The eluted fractions were analyzed by TLC, and their radiochemical purity was determined to be approximately 95%. After evaporation of the solvents as above, the compounds were stored a t -20 "C under argon. The radioinert mono-OH-M was kindly provided by Dr. Vernon J. Feil (USDA, Fargo, ND). Male Sprague-Dawley CD rats (9C-100 g) were purchased from Charles River Breeding Laboratories (Wilmington, MA) and were kept in a room with controlled temperature (22 "C) and light (12-h light/dark cycle; lights off at 7:OO p.m. EDT). Phenobarbital (PB) treatment (37.5 mg/kg ip in 0.2 mL of H 2 0 twice daily) was for 4 days. Livers were removed 12 h after the last injection. Human liver samples from eight kidney transplant donors were kindly provided by Prof. Urs Meyer (Basel, Switzerland). Microsomes were prepared from a homogenate of 1g of liver per 5 mL of 0.25 M aqueous sucrose, as previously described (19),and represented a pool of four to eight rat or human livers. The resulting microsomal pellets were washed by resuspension in 1.15% aqueous KCl followed by centrifugation a t 105000g for 1 h. The supernatant was discarded, and the microsomal pellet was covered with 2 mL of 1.15% KCl and stored a t -70 "C, until use. Protein concentrations were determined by a modified Lowry procedure (20, 21). Incubation Procedure. Incubations were carried out in 20-mL glass scintillation vials containing the following constituents: 0.6 mL of sodium phosphate buffer (pH 7.4,60 pmol); microsomal suspension (0.81 mg of protein in 0.1 mL of 1.15% aqueous KCl); EDTA (1 pmol); [14C]methoxychloror [3H]methoxychlor [usually 25 nmol (about 100000 dpm for 14Cand 250000 dpm for [3H]methoxychlor)] added in 10 pL of ethanol; NADPH-regenerating system (glucose 6-phosphate, 10 pmol; NADPH, 0.5 pmol; glucose-6-phosphate dehydrogenase, 2 IU) in 0.1 mL of phosphate buffer (pH 7.4, 10 pmol); and 0.1 mL of HzO, for a final volume of 1.0 mL. After a 2-min preincubation a t 37 OC, the reaction was initiated by adding the NADPH-regenerating system, and the vials were incubated a t 37 "C in a water bath shaker; see text for duration of incubation. The reaction was terminated by adding 10 mL of ethanol and processed by the following modification of a previously described method (10). T o remove the proteins, the resulting suspension was filtered through a 2.4-cm Whatman GF/C glass microfiber filter (Whatman Ltd., Maidstone, England) in a filter holder (Schleicher & Schuell, Inc., Keene, NH) attached to a vacuum filter flask (18). The filtrate was evaporated at room temperature to dryness under a stream of nitrogen, and the residue was stored under argon, at 4 "C, until use for chromatography. The residue was subjected to thin-layer chromatography (TLC), using chloroform/acetone (91), essentially as previously described (13),except that Analtech silica gel G plates were used. Radioinert authentic compounds were chromatographed in adjoining lanes and were detected by exposure to iodine, and the radioactive substrate and metabolites were detected with a Vangard 930 autoscanner. Usually, about 3000-9000 dpm were loaded on each lane of the TLC plate. The gel in the radioactive zones was scraped with a razor blade into individual vials, to which was added 1.0 mL of methanol and 5 mL of Liquifluor. The radioactivity was determined by liquid scintillation counting in a Packard 460C spectrometer. Quantitation of products was obtained from their respective ratio on TLC, assuming comparable loss of the various compounds. Alternatively, the residue was used for GC/MS, LC/MS, and LC/EC analysis as described below. At times, the residue was subjected to liquid-liquid partition between hexane and aqueous base (pH -11). The aqueous phase was acidified with HC1 to pH 3.0 and evaporated to dryness under a stream of nitrogen, or the acidified phase was extracted with hexane and the organic phase was evaporated under nitrogen. The
Chem. Res. Toxicol., Vol. 3, No. I, 1990 9 dry extracts were analyzed by TLC. High-performance liquid chromatography utilizing oxidative electrochemical, ultraviolet, and/or thermospray mass spectrometry detection was performed with a Waters liquid chromatograph equipped with two Model M-6000 solvent delivery systems, a Waters Model 490 multichannel variable-wavelength UV detector (254, 280 nm), a Waters Model U6K injector, and a Waters 840 data system (Waters Associates, Milford, MA). Electrochemical detection was performed with a BAS LC-17A dual-electrode amperometric detector (Bioanalytical Systems, West Lafayette, IN), which was serially connected to the liquid chromatograph described above. A glassy carbon electrode and Ag/AgCl reference electrode were used for all electrochemical studies. Waters pBondapak CI8columns (4.6 mm X 30 cm) were used in all reverse-phase separations. Linear gradient chromatographic runs beginning a t acetonitrile/0.05 M ammonium acetate buffer (4060 v/v, pH 4.5, flow 1.0 mL/min) for 12 min, and ending with acetonitrile/ammonium acetate (8020 v/v, 30-min ramp), were performed on all samples for thermospray liquid chromatography/mass spectrometry (LC/MS). Isocratic separation (50% acetonitrile/0.05 M ammonium acetate) was used to separate and detect polar metabolites of methoxychlor by electrochemical detection for determination of the hydrodynamic voltammograms, whereas 0.05 M ammonium acetate was used for electrochemical characterization of model dihydroxybenzene compounds. Gas chromatography/mass spectrometry (GC/MS) and LC/MS were performed with a Nermag R10-1OC quadrupole mass spectrometer (Delsi, Inc., Fairfield, NJ) equipped with a Nermag Thermospray source and Vestec Thermospray probe and gradient controller (Vestec Inc., Houston, TX) for LC/MS and a Varian Vista gas chromatograph (Varian, Palo Alto, CA) for capillary GC/MS. The Finnigan Super INCOS data system was used for data acquisition (Finnigan Corp., Sunnyvale, CA). With the chromatographic conditions described above, LC/MS characterization of metabolites was accomplished with negative ion detection with the thermospray source operated a t 200 "C, probe, T1 = 105 "C, and T2 = 190 "C. A repeller, voltage 200 V, was applied for all runs. Application of the repeller voltage had no effect on fragmentation but increased sensitivity 5-10-fold with this system. No discharge or filament current was used. GC/MS was performed with a dry solids injector (Allen Scientific, Boulder, CO) and a 15-m SPB5 fused silica capillary column (Supelco, Inc., Bellafonte, PA), which was plumbed directly to the ion source. Helium was used as carrier gas a t a flow rate of 40 cm/s. Positive ion chemical ionization (CI) mass spectrometry was performed for metabolite identification at 90 eV with methane as reagent gas at a source pressure of 0.25 Torr. Microsomal extracts were derivatized with bis(trimethylsily1)trifluoroacetamide (Supelco, Inc., Bellafonte, PA) by adding 20 pL of reagent directly to the thoroughly dried extract and heating in an oven for 1h at 70 O C . An aliquot of the derivatized extract was then injected onto the column. The GC was operated as follows: injector, 275 "C, oven, 100 "C for 2 min, followed by a linear gradient to 275 "C a t 10 "C/min. Catechol 0-Methyltransferase (COMT). The method to determine whether the tris-OH-M behaves like a typical catechol and is methylated by S-adenosylmethionine (SAM) through catalysis by COMT involved modifications of a previously described procedure for quantitation of catechol estrogens (24,25). The incubation was in 1mL containing 3.3 pM [ m e t h ~ l - ~ H ~ ] S and AM 25 units of COMT in sodium phosphate buffer (50 mM, pH 7.4) and ascorbate (50 pM), MgC1, (10 mM), dithiothreitol(70 pM), and the respective substrate 2-hydroxyestradiol [2-OH-E2] or bis-OH-M or the mixture of isolated polar metabolites of methoxychlor (bis-OH-M and tris-OH-M), a t 5-30 pM. After preincubation for 2 min a t 37 "C, 25 units of COMT was added and the incubation continued for 20 min. The reaction mixture was placed on ice and extracted with 6 mL of cold hexane. A 2-mL aliquot of the hexane phase was used for scintillation counting. The residual hexane extract, in samples containing a significant amount of radioactivity, was evaporated to dryness a t room temperature under a stream of nitrogen gas, and the residue was subjected to TLC and analyzed for radioactive zones (13). The mixture of unlabeled metabolites (bis-OH-M and tris-OH-M) used in the above experiment with COMT was obtained from an in-
10 Chem. Res. Toxicol., Vol. 3, No. 1, 1990 3 n
M
n
2
M
+
Kupfer et al. M
I
n
2
I
c
+
Figure 2. Radioscans of TLC of an extract from an incubation
Figure 1. Radioscan of TLC of aliquots of extracts from incubations of [14C]methoxychlor(20-25 pM, 80 000-100 000 dpm) with rat or human liver microsomes. Upper panel: PB rat liver microsomal protein (0.8 mg/incubation). Middle panel: Control rat liver microsomal protein (0.5 mg). Bottom panel: Human liver microsomal protein (0.8 mg). Incubation time 60 min. Symbols: 0 = origin, S = solvent front; arrows in other panels correspond to origin (right) and solvent front (left); M, methoxychlor.
cubation of unlabeled methoxychlor with PB liver microsomes and NADPH, as described above under Incubation Procedure. The metabolites were isolated by differential extraction. The quantitation of the unlabeled bis-OH-M and tris-OH-M was deduced from analysis of metabolites formed by parallel incubations of [14C]methoxychlor.The radioactive extract was analyzed by TLC and scintillation spectrometry of the scraped radioactive zones corresponding to bis-OH-M and tris-OH-M as previously described (12),and the ratio of the metabolites was quantitated. In incubations with SAM and COMT, the labeling of unlabeled 2-OH-Ezserved as positive control and incubations lacking substrates or containing 2-OH-Ezbut lacking COMT served as negative controls.
Results Incubation of [14C]methoxychlor with liver microsomes from untreated (control) and phenobarbital (PB) treated rats and with human livers, in the presence of NADPH, revealed radiolabeled zones on TLC representing the formation of at least three metabolites. Figure 1depicts representative TLC radioscans of extracts from incubations of methoxychlor with liver microsomes from control rats (control microsomes), from phenobarbital treated rats (PB microsomes), and from human liver microsomes. Solvent-solvent partition of the radiolabeled metabolites from basic and acidic solutions indicated that metabolites 1and 2 are acidic (Figure 2). Apparently the high polarity of metabolite 3 prevented its extraction by hexane. On the basis of R values of authentic synthetic derivatives [ 1,1,l-tricklor o- 2-(4-hydroxyphenyl)- 2-(4-methoxyphenyl)ethane, referred to as mono-OH-M, and l,l,l-trichloro-2,2-bis(4-hydroxyphenyl)ethane, referred to as bis-OH-MI, it was assumed that the least polar metabolite 1 is the monodemethylated product (mono-OH-M) and
of [14C]methoxychlorwith control rat liver microsomes (metabolites partitioned between the aqueous and organic phases). Upper panel: Hexane extract from the incubation mixture extract suspended in 0.1 N NaOH. Middle panel: second hexane extract, after acidificationwith HC1 of the aqueous phase from the above mixture. Bottom panel: Aqueous residue after the second hexane extraction. M, methoxychlor; peak 1,putative mono-OH-M;peak 2, putative bis-OH-M; and peak 3, polar metabolite(s1. Symbols as in Figure 1. metabolite 2 is the didemethylated product (bis-OH-M). These metabolites were found to exhibit identical chromatographic behavior as the respective authent,ic synthetic compounds on TLC and HPLC, providing support for the proposed structure assignment. The observation that metabolite 3 is highly polar (low Rfon TLC and insolubility in hexane) and exhibits a considerably shorter retention time than bis-OH-M on reverse-phase HPLC tempted the speculation that this compound is a tri- or tetrahydroxy derivative. To identify the metabolites, multiple incubations were performed with methoxychlor and liver microsomes from PB-treated (PB microsomes) and untreated (control microsomes) rats and pooled separately. These pooled extracts were subjected to GC/MS, LC/MS, and LC/EC analysis. Figure 3 shows the CI mass spectra of the trimethylsilyl (TMS) derivatives of authentic bis-OH-M and of metabolite 2 in the P B microsome extract, whereas Figure 4 shows the negative ion LC/MS spectra of the authentic bis-OH-M and of metabolite 2 in both control and PB microsomes. These results show that the spectra of the authentic bis-OH-M were essentially identical with those obtained with the isolated metabolite 2, using the TMS derivative in GC/MS or the underivatized metabolite in LC/MS. Both the authentic bis-OH-M and the metabolite are dechlorinated during gas chromatography; however, an additional minor peak was also detected in both samples in which all three chlorine atoms were retained. It is also interesting that these compounds form dimer adducts as well as acetate adducts under thermospray LC/MS conditions. These findings established that metabolite 2 is bis-OH-M. The incubation of the HPLCpurified metabolite 1with liver microsomes and NADPH yielded the bis-OH-M and a polar metabolite 3 (not shown), supporting our initial supposition that metabolite 1is the mono-OH-M. Additionally, the LC/MS analysis of the extract from incubations with control microsomes
Chem. Res. Toxicol., Vol. 3, No. I , 1990 11
Methoxychlor Metabolism by Mammalian P-450 A
[M+CH,COO]-
A
375
1000,
,
l:::/,),
MI2
;,
, ,
,,,
,
I
150
100
, , , , 12;g/
200
250
[MI+ 42s
100.07
r
50.0 [MI-
E
375 MI2
300
350
400
316
100 0,
453
,I,, ,i, j
450
B
500],
,
MI2
,
,
, , ,
,'p",
200
150
,
:_, ,
z;!, ,
250
F
[M+CHICOO]375
I
300
t
,
,
350
400
20Oq
F
50.0
loo.{, MI2
'0°'0
, ,
I
,
,
, ,
100
I
,
,
, ,
150
,
200
,
;\g/ 250
0
Mil-
[ M+CH,COOl-
C
1
315
100.0,
[M-15]+
50.0 -
352
389
I
453
b 20 0
id
Figure 3. Chemical ionization (CI) GC MS of trimethylsilyl (TMS)derivatives of authentic bis-OH- (A) and of metabolite 2 (B) in extract from PB rat microsome incubations. Table I. Mass Spectra of the Polar Methoxychlor Metabolite with GC/MS and LC/MS mass % base peak proposed fragment GC/MSn 541 7.0 [M + 291' 514 65.1 [MI+ (C1 cluster) 497 29.1 [M minus 151' 475 43.8 [M minus Cl]' 459 85.8 [M minus (CH3 and Cl)]+ 385 59.4 unassigned 371 100.0 unassigned 431 100 [M minus CCl3It 513 98.3 [M minus C1]+ (C1 cluster) 535 70.4 [M minus 15]+ (C1 cluster) 549 51.6 [MIt (C1 cluster)
b
c
332 225 295 (296) 391 202 151 665
LC/MSd 100 [MI- (C1 cluster) 81.4 [M minus 3C1]22.5 [M minus C1]- (C1 cluster) 20.9 [M plus CH3CO]- (C1 cluster) 18.9 unassigned 15.5 unassigned 6.7 [2M]- (C1 cluster)
GC/MS was carried out on TMS-derivatized metabolite. *Fragmentation of the major GC peak in which Mt represents a loss of a single chlorine. cFragmentation of a minor GC peak in which Mt retained the chlorines. dLC/MS was carried out on an underivatized metabolite. (I
showed the presence of significant amounts of mono-OH-M (Figure 5). The spectrum showed a molecular ion cluster a t m/z 330, an acetate adduct a t m / z 389, and a dimer adduct a t m / z 661, analogous to the spectrum observed
j
1
o [M+~CH~COO~-
10
MIZ
1
419495
450
1
[2Ml633
437
Y I
650
600
550
500
700
Figure 4. Negative ion LC/MS: (A) authentic bis-OH-M; (B) metabolite 2 in PB microsomes; and ( C ) metabolite 2 in control microsomes.
11;
100.0, 50.0
I-
[ M+CH&OO 389
j;l-J I F
~,
,9&,
I_,
I ,
209 221 MI2
100.0
150
200
250
300
350
400
1 13 i [2Ml-
50'0
MI2
661
450
500
550
600
650
700
Figure 5. Negative ion LC/MS of metabolite 1 from control rat liver microsome incubations.
with bis-OH-M. These results demonstrated that metabolite 1 is mono-OH-M. Since there were no authentic compounds available, which could facilitate the identification of metabolite 3, the characterization of this metabolite required an exten-
12 Chem. Res. Toxicol., Vol. 3, No. 1, 1990
Kupfer et al.
A 100.0
50.0 103 129 147 L j t 1
MI2
MI2
I -1
I
100
I
I
1
,
'
I
400
1'l
I
200
150
350
219 205 j 239
.
450
I
:
I
i
i
I
250
500
9
299 313 I ' 300
' '
9
550 ..
332
1
225
t
:001391 -
I
[2Ml-
10.0-
665
sive investigation. The mass spectrum of the TMS derivative of metabolite 3 in CI-GC/MS is shown in Figure 6A and Table I and is consistent with a trihydroxy trisTMS, dichloro metabolite. An additional minor peak (see Table I) was also detected in which the spectrum is consistent with a trihydroxy, tris-TMS, trichloro metabolite. Similar observations were also made with authentic bisOH-M and the metabolite which are also dechlorinated during gas chromatography as described above. The LC/MS spectrum of metabolite 3 is shown in Figure 6B and Table I. This spectrum is consistent with a didemethylated, hydroxylated metabolite. In addition to the chlorine isotope clusters, the spectrum of this metabolite D
2
1
I
. e
00-
-$
60-
A METABOLITE AUETA-DHB
0
K c
I
0ORTH+DHE
2
v
I
100-
-r
.E
has the acetate and dimer adduct ions which were also observed in the spectra of mono- and bis-OH-M. These spectra suggest that the chlorine atoms are retained; however, no information on the exact position of hydroxylation was obtained. Further characterization of metabolite 3 was obtained by HPLC with UV and oxidative electrochemical detection. The metabolite had a higher ratio of UV absorbance at 280:254 nm in HPLC (being 0.89) than that of the dihydroxy derivative (0.69), a finding that indicated that the third hydroxyl in metabolite 3 was most probably a ring hydroxyl; namely, there was one phenolic hydroxyl and two hydroxyls in the second ring. This raised the question of whether the two hydroxyls were ortho or meta to each other. To resolve this question, metabolite 3 was subjected to reverse-phase HPLC and oxidative LC/EC analysis. The hydrodynamic voltammograms of model compounds (0- and m-dihydroxybenzene) were compared to those of bis-OH-M (metabolite 2) and metabolite 3. Figure 7, left panel, demonstrates the high similarity of the voltammograms of metabolite 3 to that of the o-dihydroxybenzene. By contrast, the voltammogram of mdihydroxybenzene differed substantially, requiring a much higher potential for oxidation. Additionally, the voltammogram of m-dihydroxybenzene resembled that of the bis-OH-M (Figure 7 , right panel), indicating that mhydroxyls are not readily oxidized and behave autonomously like single phenolic hydroxyls. Moreover, whereas both the electrochemically oxidized forms of the o-dihydroxybenzene (catechol) and of metabolite 3 underwent reduction at a second in-series electrode at 0.0 V, oxidized m-dihydroxybenzene was not reduced (not shown). These findings demonstrated that metabolite 3 is a trihydroxy catechol [l,l,l-trichloro-2-(4-hydroxyphenyl)-2-(3,4-dihydroxyphenyl)ethane, to be referred to as tris-OH-MI. To determine whether tris-OH-M behaves as a typical catechol, we examined whether this metabolite could be ( [3H]SAM) labeled by [methyl-3H,]-S-adenosylmethionine when incubated in the presence of catechol O-methyltransferase (COMT). To prevent the potential autoxidation of the tris-OH-M in air, we isolated a mixture of bis-OH-M and tris-OH-M, from incubations of unlabeled methoxychlor with P B microsomes by differential extraction without further purification, as described under Materials and Methods. This mixture containing both tris-OH-M and bis-OH-M was incubated in the presence of [3H]SAM and COMT. Results indicate that whereas the mixture containing both the bis-OH and the tris-OH metabolites yielded labeled products in the hexane phase, synthetic bis-OH-M alone did not generate a significant amount of radioactive metabolites (Table 11). The level of the radioactive metabolite in the hexane phase from the
loo': 80
6o
40 --
i
X
0
Y
40: 20
207-
I
Y
/
w
O
O 0.1
A
0.4
0.7
POTENTIAL (VOLTS)
1
.o
0 0.1
I m
0.4
0.7
1 .o
POTENTIAL (VOLTS)
Figure 7. Hydrodynamic voltammograms of metabolite 3 and model compounds (dihydroxybenzenes and bis-OH-M). Left panel: metabolite, o-dihydroxybenzene (ortho-DHB), and m-dihydroxybenzene (meta-DHB). Right panel: Bis-OH-M and meta-DHB.
Chem. Res. Toxicol., Vol. 3, No. 1, 1990 13
Methoxychlor Metabolism by Mammalian P-450 Table 11. Labeling of Catechols by [sH]-S -Adenosylmethionine (SAM) Catalyzed by Catechol O-Methyltransferase (COMT)' pmol substrate, pM COMT (labeled)d + 1.6 none tris-OH-M, 16.7, plus bis-OH-M, 10.5* + 7.5e tris-OH-M, 10.9, plus bis-OH-M, 6.6*sC + 7.3e 2-OH-Ez, 20 0.06 + 802.8 2-OH-Ez, 20 + 410.3 2-OH-E2, 10 + 0.7 bis-OH-M, 5 + 1.3 bis-OH-M, 10 "Each incubation contained [3H]SAM (0.5 pCi, 3.3 pM) and when indicated 25 units of COMT. Conditions where as described under Materials and Methods. bRepresents a mixture of tris-OH-M plus bis-OH-M isolated from the incubation of methoxychlor with PB microsomes. The quantitation of methoxychlor metabolites used in the COMT reaction was deduced from determinations of metabolites formed from [14C]methoxychlor and hence represents only an approximation (see Materials and Methods). Represents a mixture of metabolites isolated from another incubation with PB microsomes. Radioactivity in the hexanephase extract. 'Radioactivity in the hexane phase, from the incubation of the mixture of tris-OH-M plus bis-OH-M; when subjected to TLC, the radioactivity exhibited characteristics of both a highly polar product that remained at the origin and a product less polar than bis-OH-M. By contrast, there was little or no radioactivity when bis-OH-M only was used in the incubation.
incubation of the mixture of bis-OH-M and tris-OH-M was small by comparison with that obtained with synthetic 2-OH-E2,suggesting that tris-OH-M is a poor substrate for COMT or that a substantial portion of the metabolite has been oxidized during isolation and could not serve as a substrate for COMT. The radioactivity found in the hexane phase when subjected to TLC exhibited both a highly polar component (at the origin) and a component less polar than bis-OH-M. Though not identified, this finding suggested that the radioactive product contained methylated tris-OH-M. The possibility that the majority of the radiolabeled methylated tris-OH-M has remained in the aqueous phase after hexane extraction is unlikely, since ether extraction did not yield any additional amount of the less polar metabolite. To diminish the likelihood of oxidation of the tris-OH-M during its isolation, an attempt was made to carry out the COMT-SAM-mediated methylation sequentially after the formation of the trisOH-M, without the necessity for its prior isolation. That procedure, which employed a coupled enzymatic reaction involving the incubation of methoxychlor with PB microsomes, NADPH, 13H]SAM,and COMT, was not successful; i.e., it did not yield radiolabeled products. Similarly, the incubation of estradiol with P B microsomes under the same conditions did not generate radiolabeled methylated products, normally detected in the hexane extract (not shown). This surprising observation was contrary to previous findings, which demonstrated the formation of a catechol from estradiol by liver microsomes, using SAM and COMT (24,25). The possibility that under our conditions liver microsomes inhibit the COMT-mediated methylation of catechol estrogens was examined. Indeed, liver microsomes, with or without NADPH, markedly inhibited the methylation (labeling) of added 2-OH-E2by L3H]SAM and COMT (not shown). The reason for this discrepancy among different laboratories is not apparent. The question was raised whether tris-OH-M is formed by demethylation of methoxychlor to bis-OH-M, followed by ring hydroxylation, or vice versa by ring hydroxylation prior to the mono- or didemethylation. On the basis of the observation that the incubation of the mono-OH-M
A-ATRIS-OH
I 0
0
O
..-.-.
I
\\
30
60
90
120
INCUBATION TIME (Minutes)
Figure 8. Time course of methoxychlor metabolism by liver microsomes from phenobarbital-treated rats. Each incubation contained 0.86 mg of microsomal protein.
with PB rat liver microsomes yields the bis-OH-M and the polar metabolite (not shown) and on a time course of methoxychlor metabolism by PB microsomes (Figure a), resulting in a rapid appearance and disappearance of the mono-OH-M with the subsequent formation of the bisOH-M and tris-OH-M, we propose that tris-OH-M is formed by hydroxylation of bis-OH-M. However, the possibility that under certain conditions the sequence is reversed, with ring hydroxylation occurring first, was not ruled out. Additionally it should be noted that a small portion of the metabolite in the highly polar tris-OH-M fraction is another compound(s), possibly resulting from side-chain hydroxylation and/or ring hydroxylation. This is based on the observation that a certain amount of highly polar metabolite is extractable from base by hexane and on the finding that a small amount of metabolite from incubations of meth~xy-~H-labeled methoxychlor yields labeled metabolite migrating on TLC in the tris-OH-M zone. No attempt was made to characterize these minor metabolites. Because of paucity of metabolites from incubations with human liver microsomes and due to limitation in amounts of human tissues, identification of human liver microsomal metabolites was based on chromatographic characteristics on TLC. Furthermore, to determine whether metabolites 1and 2 in human liver microsomes were indeed the monoand didemethylated products, respectively, incubations were conducted with two radiolabeled preparations of methoxychlor: one preparation labeled with m e t h ~ x y - ~ H and the other labeled with ring-14C. A similar experiment was conducted with control and PB rat liver microsomes. On the basis of TLC analysis of the metabolites from the 3H-labeled versus 14C-labeledmethoxychlor, it was evident that the products from incubations with human liver microsomes were the same as from rat liver microsomes. There was a dramatic decrease in the recovery of radioactivity in extracts from incubations of [3H]methoxychlor with PB microsomes due to loss of [3H]methoxyas formaldehyde2 (Figure 9), which most probably evaporated during the workup. The drop in radioactivity paralleled the disappearance of methoxychlor. There was a much lesser decrease in radioactivity in control rat and human microsomes, because of the accumulation of mono-OH-M Using the Nash reagent, we have observed that the incubation of methoxychlor with untreated rat liver microsomes generates formaldehyde (12) and that phenobarbital treatment causes a relatively small increase in formaldehyde accumulation (unpublished). By contrast, the quantitation of methoxychlor metabolites on TLC demonstrated a marked increase in phenolic metabolites. We attribute this discrepancy to the insensitivityof the Nash method, particularly since we used a low methoxychlor concentration (20-25 p M ) due to the low K, of demethylation, being about 9 pM (unpublished).
14 Chem. Res. Toxicol., Vol. 3, No. I , 1990
Kupfer et al. "V
0-0
10
RMlOACmnn
-0
0
30
90
60
120
\
0
5
t
\
40i
\
1 ,
\
o----
--
0
0 120
0
30 60 90 INCUBATION TIME (Minutes)
Figure 9. Time course of loss of radioactivity and disappearance of methoxychlor in extracts from incubations of [ m e t h o ~ y - ~ H ] methoxychlor with control and PB rat and human liver microsomes. Top left panel: Control rat liver microsomes. Top right panel: PB microsomes. Bottom panel: Human microsomes.
METHOXYCHLOR
MONO-OH-~IETHOXYCHLOR
81 S-OH-flETHOxYCHLoR
TR I S-OH-RETHOXYCHLOR
Figure 10. Proposed pathway of metabolism of methoxychlor by liver microsomes.
containing one labeled [3H]methoxy. Furthermore as expected, in the incubation of [3H]methoxychlorwith human liver microsomes the TLC zone corresponding to metabolite 1 was radioactive and the zone corresponding to metabolite 2 was devoid of 3H radioactivity; by contrast, both zones from concomitant incubations with the [14C]methoxychlor were radioactive. These findings together with the R values of the metabolites on TLC demonstrated that metaLolites 1 and 2 from incubations with human liver microsomes were the same as with control and P B rat microsomes. Currently, we have no information about the highly polar metabolite (a minor component) from human liver incubations.
Discussion The present study describes the oxidative metabolism of methoxychlor by the hepatic monooxygenases from control and PB-treated rat and human into three phenolic metabolites. On the basis of a time course of formation of [14C]methoxychlormetabolites and the observation that bis-OH-M is converted to the tris-OH-M, it is concluded that the pathway of metabolism of methoxychlor proceeds by a two-step demethylation to mono-OH-M and bis-OHM, followed by hydroxylation of bis-OH-M, resulting in accumulation of the tris-OH-M (Figure 10). This is particularly evident in incubations with rat PB microsomes (Figure 11,where tris-OH-M becomes the major product. Preliminary observations, based on inhibition studies with liver microsomes and liver slices, suggest that ring hydroxylation is catalyzed by a different monooxygenase than
the demethylation reactions. Further studies are needed to establish this with certainty and to determine which cytochrome P-450sare involved in demethylation and ring hydroxylation of methoxychlor. The identification of the metabolites was somewhat complicated because of the ease of loss of chlorines, particularly in the GC/MS, resulting in a molecular ion lacking one chlorine atom. However, in the LC/MS (negative ions), the presence of all the side-chain chlorines was helpful, since a cluster of peaks was usually evident around the major fragments, signifying the presence of the two stable chlorine isotopes (37Cland 35Cl). The hydrodynamic voltammetry studies to characterize the tris-OH-M further demonstrated the ease of use and the capability of the technique to yield important information with excellent sensitivity and with very small sample preparation. Other investigators (22,23) have also used hydrodynamic voltammetry to characterize unambiguously liver microsomal metabolites of 2- and 3hydroxyacetanilide and hydroxylated metabolites of benzene (catechol, hydroquinone, and phenol). Their results were similar to the results obtained with model hydroxylated benzene derivatives used in the current study. Previous studies showed that the mono- and bis-OH metabolites are estrogenic (8, 12, 13). Also, our earlier studies, albeit involving poor chromatographic separation techniques, suggested that the highly polar metabolite (HPM) is weakly estrogenic (12). Tris-OH-M is most probably identical with the HPM or represents the major component of HPM. Studies are needed to determine the
Methoxychlor Metabolism by Mammalian P-450 estrogenic potency of a highly purified preparation of tris-OH-M and whether its estrogenic characteristics are similar to those of other catechols, such as 2- and 4hydroxyestradiol. Catechol estrogens have been implicated in hormonal carcinogenesis (28,29);hence it would be of interest to determine whether tris-OH-M is involved in carcinogenesis. In earlier studies, we observed that methoxychlor is metabolically activated by cytochrome P-450 monooxygenase to bind covalently to microsomal proteins and that PB microsomes catalyze that activity at approximately a 10-fold higher rate than control microsomes (10,11). On the basis of the molecular weight of the radioactive protein adducts on SDS-PAGE, it appears that methoxychlor binds primarily to proteins with M, values of 45-60K (11); this molecular weight region corresponds to a variety of proteins, among these cytochrome P-450s. Though the incubations of methoxychlor with liver microsomes did not diminish the total level of cytochrome P-450, assayed spectrally, preliminary results suggest that methoxychlor inactivates in a time-dependent fashion certain enzymatic activities catalyzed by P-450 (not shown). The formation of tris-OH-M is of particular interest, since this catechol becomes the major metabolite within 30 min in incubations of methoxychlor with P B microsomes. Catechols are able to undergo oxidative activation and covalent binding to macromolecules, possibly via semiquinone radicals (15-1 7); hence it is conceivable that the tris-OH-M could undergo similar activation and covalent binding. Preliminary studies indicate that demethylation is not essential for covalent binding (11). If proven true, such a finding will suggest only a minor role for tris-OH-M in that reaction. Nevertheless, it is possible that a minor portion of the covalent binding is due to the activation of the tris-OH-M catechol. Studies to determine whether tris-OH-M can undergo such activation have been delayed by difficulties in the isolation of adequate amounts of radiolabeled metabolite. In preliminary studies, with a limited amount of highly purified [14C]tris-OH-Mat our disposal, we demonstrated binding to P B microsomes. This irreversible binding, presumably covalent, did not appear to require NADPH. Currently, we do not know whether the binding involves enzymatic or nonenzymatic mechanisms. It is likely that a nonenzymatic mechanism is involved, since in earlier studies (lo), we observed that a crude preparation of 14C metabolites, containing primarily HPM, exhibited irreversible binding to liver microsomal proteins. That reaction appears to be nonenzymatic, since binding occurred even with boiled microsomes. Future studies are planned to isolate or synthesize sufficient amounts of tris-OH-M, to determine whether this compound is estrogenic, and to determine the nature of its binding to proteins.
Acknowledgment. We sincerely thank Drs. Vernon Feil (Fargo, ND), James Sanborn (Urbana, IL),and Toshio Fujita (Kyoto University, Japan) for providing authentic compounds used in these studies. We are grateful to Ms. Debrah Luper, who participated in early phases of this investigation. We also thank Ms. Chitra Mani for the help with some of the experiments. We thank Dr. Claire 0’Connor for generously providing the SAM. This study was supported in part by USPHS Grant ES00834 from the National Institute of Environmental Health Sciences, NIH. Registry No. [14C]Mono-OH-M, 124042-16-6; [14C]bis-OH-M, 124042-17-7; methoxychlor, 72-43-5;mono-OH-M, 28463-03-8; bis-OH-M, 2971-36-0; tris-OH-M, 124042-15-5; monooxygenase, 9038-14-6;cytochrome P-450,9035-51-2.
Chem. Res. Toxicol., Vol. 3, No. 1, 1990 15
References (1) Weikel, J. H. (1957)The metabolism of methoxychlor (1,lJtrichloro-2,2-bis-(p-methoxyphenyl)ethane).I. The role of the liver and biliary excretion in the rat. Arch. Znt. Pharmacodyn. Ther. 110,423-432. (2) Kapoor, I. P., Metcalf, R. L., Nystrom, R. F., and Sangha, G. K. (1970)Comparative metabolism of methoxychlor, methiochlor, and DDT in mouse, insects, and in a model ecosystem. J. Agric. Food Chem. 18,1145-1152. (3) Hayes, W. J. (1971)Clinical Handbook on Economic Poisons, Environmental Protection Agency Pesticide Program, Chamblee, GA. (4) Metcalf, R. L. (1976)in The Future for Insecticides (Metcalf, R. L., and McKelvey, J. J., Eds.) pp 223-285, John Wiley and Sons, New York. (5) Metcalf, R. L., Kapoor, I. P., and Hirwe, A. S. (1971)Biodegradable analogues. Bull. WHO 44, 363-377. (6) Murphy, S. D. (1980)Pesticides. In Toxicology, the Basic Science of Poisons (Doull, J., Klaassen, C. D., and Amdur, M. O., Eds.) pp 357-408,Macmillan Publishing Co., New York. (7) Bulger, W. H., Muccitelli, R. M., and Kupfer, D. (1978)Interactions of methoxychlor, methoxychlor base-soluble contaminant, and 2,2-bis(p-hydroxyphenyl)-l,l,l-trichloroethane with rat uterine estrogen receptor. J. Toxicol. Enuiron. Health 4, 881-893. (8) Bulger, W. H., Feil, V. J., and Kupfer, D. (1985)Role of hepatic monooxygenases in generating estrogenic metabolites from methoxychlor and from its identified contaminants. Mol. Pharmacol. 27,115-124. (9) Kupfer, D., and Bulger, W. H. (1979)A novel in vitro method for demonstrating proestrogens. Metabolism of methoxychlor and o,p’-DDT by liver microsomes in the presence of uteri and effects of intracellular distribution of estrogen receptors. Life Sci. 25, 975-984. (10) Bulger, W. H., Temple, J. E., and Kupfer, D. (1983)Covalent binding of 14C-methoxychlormetabolite(s) to rat liver microsomal components. Toxicol. Appl. Pharmacol. 68,367-374. (11) Bulger, W. H., and Kupfer, D. (1989)Characteristics of monooxygenase-mediated covalent binding of methoxychlor in human and rat liver microsomes. Drug Metab. Dispos. 17,487-494. (12) Bulger, W. H., Muccitelli, R. M., and Kupfer, D. (1978)Studies on the in uiuo and in vitro estrogenic activities of methoxychlor and its metabolites. Role of hepatic mono-oxygenase in methoxychlor activation. Biochem. Pharmacol. 27,2417-2423. (13) Ousterhout, J. M.,Struck, R. F., and Nelson, J. A. (1981) Estrogenic activities of methoxychlor metabolites. Biochem. Pharmacol. 30,2869-2871. (14) Lamoureux, C. H., and Feil, V. J. (1980)Gas chromatographic and mass spectrometric characterization of impurities in technical methoxychlor. J. Assoc. Off. Anal. Chem. 63,1007-1037. (15) Sasame, H. A., Ames, M. M., and Nelson, S. D. (1977)Cytochrome P-450and NADPH cytochrome C reductase in rat brain: Formation of catechols and catechol metabolites. Biochem. Biophys. Res. Commun. 78,919-926. (16) Nelson, S. D. (1982)Metabolic activation and drug toxicity. J. Med. Chem. 25,753-765. (17) Kalyanaraman, B., Korytowski, W., Pilas, B., Sarna, T., Land, E. J., and Truscott, T. G. (1988)Reaction between ortho-Semiquinone and Oxygen: Pulse radiolysis electron spin resonance and oxygen uptake studies. Arch. Biochem. Biophys. 266,277-284. (18) Juedes, M.J., Bulger, W. H., and Kupfer, D. (1988)Monooxygenase-mediated activation of Chlorotrianisene (TACE) in covalent binding to rat hepatic microsomal proteins. Drug Metab. Dispos. 15,786-793. (19) Burstein, S. H., and Kupfer, D. (1971)Hydroxylation of trans delta 1-tetrahydrocannabinol by hepatic microsomal oxygenase. Ann. N.Y. Acad. Sci. 191,61-67. (20) Lowry, 0.H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951)Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193,265-275. (21) Stauffer, C. E. (1975)A linear curve for the Folin Lowry determination of protein. Anal. Biochem. 69,646-648. (22) Hamilton, M., and Kissinger, P. T. (1986)The metabolism of 2-and 3-hydroxyacetanilide. Drug Metab. Dispos. 14,5-12. (23) Roston, D. A., and Kissinger, P. T. (1982)Isolation and identification of benzene metabolites in vitro with liquid chromatography/electrochemistry. Anal. Chem. 54,1798-1802. (24) Hoffman, A. R., Paul, S. M., and Axelrod, J. (1980)Estrogen2-hydroxylase in the rat. Biochem. Pharmacol. 29,83-87.
16 Chem. Res. Toxicol., Vol. 3, No. 1, 1990 (25) Li, S. A., Purdy, R. H., and Li, J. J. (1989) Variations in catechol 0-methyltransferase in rodent tissues: possible role in estrogen carcinogenicity. Carcinogenicity 10, 63-67. (26) Gray, L. E., Jr., Ostby, J. S., Ferrell, J. M., Sigmon, E. R., and Goldman, J. M. (1988) Methoxychlor induces estrogen-like alterations of behavior and reproductive tract in the female rat and hamster: effects on sex behavior, running wheel activity, and uterine morphology. Toxicol. Appl. Pharmacol. 96, 525-540. (27) Cummings, A. M., and Gray, L. E., Jr. (1989) Antifertility effect
Kupfer et al. of methoxychlor in female rats: Dose- and time-dependent blockade of pregnancy. Toxicol. Appl. Pharmacol. 97,454-462. (28) Li, J. J., and Li, S. A. (1987) Estrogen Carcinogenesis in the hamster tissues: role of metabolism. Fed. Proc., Fed. Am. SOC. Exp. Biol. 46, 1858-1863. (29) Liehr, J. G., Fang, W.-F., Sirbasku, D. A., and Ari-Ulubelen, A. (1986) Carcinogenicity of catechol estrogens in Syrian hamster. J. Steroid Biochem. 24, 353-356. (30) Kupfer, D., et al. (1989) FASEB J . 3 (4), Abstract 5592.
