XENOBIOTICA,
1990, VOL. 20, NO. 7, 657670
Development of a lqF-n.m.r.method for studies on the in vivo and in vitro metabolism of 2-fluoroaniline J- VERVOORTT, P. A. DE JAGERS, J. STEENBERGENT and I. M. C. M. RIETJENStO
t Department of Biochemistry, Agricultural University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands
# Department of Molecular Physics, Agricultural University, Dreijenlaan 3,
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
6703 HA Wageningen, The Netherlands Received 8 September 1989; accepted 1 March 1990 1. A IgF-n.m.r. method has been developed for study of the metabolism of 2fluoroaniline both after in wiwo exposure of rats and in in oitro model systems. 2. From the IgF-n.m.r. spectrum of the 24 h urine it was caIcuIatedthat over 90% of the dose was excreted within 24 h. The metabolic pattern showed that 85% of the metabolites were para-hydroxylated, 72% sulphated, 13% glucuronidated and 29% N-acetylated, 4amino-3-fluorophenyl sulphate being the main urinary metabolite (53%). 3. In witro studies of phase I metabolism of 2-fluoroaniline with rat liver microsomes was representative for the in vim metabolism as hydroxylation in both systems was observed only at the para-position. 4. Phase I I1 metabolism was studied in witro in either isolated rat hepatocytes in suspension or in a 1 h recirculating liver perfusion system. In both these in witro systems para-hydroxylation, N-acetylation, sulphation and glucuronidation of 2-fluoroaniline were observed. The ratio between glucuronidation and sulphation was dependent on sulphate availability 5. Of the in vitro systems tested, hepatocytes in Krebs Ringer (sulphate limited) medium was the best model for in wiwo metabolism. 6. The detection limit for fluoro-containingmetabolites in this IgF-n.m.r. method was 1 p~ for an overnight run using a Bruker CXP 300 spectrometer. From this it can be concluded that lgF-n.m.r. urine analysis is a useful tool in biomonitoring studies. For 2fluoroanilinethe method appears to be more sensitive than currently available h.p.l.c./t.l.c. methods. In addition, concentration of urine samples can result in either lower detection limits, or in shorter times needed for n.m.r. data acquisition. 7. N-acetylation is known to show genetic polymorphism. Therefore, the lgF-n.m.r. method, detecting all 2-fluoroaniline metabolites, has the additional advantage of eliminating the risk of obtaining false negatives for fast acetylators.
+
Introduction 2-Fluoroaniline is used as an intermediate in the manufacturing of herbicides and plant growth regulators. Plant workers may be exposed to this chemical (Eadsforth et aZ. 1986). Although much is known about the metabolism of anilines in general (Boyland et al. 1953,Baldwin and Hutson 1980,Grossman and Jollow 1986), specific knowledge concerning the biotransformation of this fluorinated aniline is restricted to only one in vivo study (Eadsforth et al. 1986). We therefore decided to study the metabolism of 2-fluoroaniline in more detail using microsomal, hepatocyte and perfused liver systems in addition to in viwo exposures. Eadsforth et aZ. (1986) detected eight metabolites in the urine of exposed animals; six of these metabolites were identified. 4-Acetamido-3 -fluorophenyl sulphate was reported to be the main
8 To whom correspondence should be addressed. 0049-8254/90 83.00 0 1990 Taylor & Francis Ltd.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
658
J . Vervoort et al.
urine metabolite. In this study of Eadsforth et al., 14C-labelled 2-fluoroaniline was used and urinary 14C-labelledmetabolites were separated on t.1.c. and/or h.p.1.c. We decided to study the metabolism of 2-fluoroaniline using 'gF-n.m.r. This technique enabled us to perform direct measurements afurine samples, as well as of samples from liver perfusion and from hepatocyte or microsomal incubations. Additional advantages of the "F-n.m.r. technique, making it suitable for xenobiotic biotransformation studies, are the high sensitivity of the chemical shift of Auorocontaining compounds to chemical modifications, the absence of natural background signals and the high sensitivity which almost equals that of 'H-n.m.r. The present paper describes the identification of the various fluoro-containing metabolites of 2fluoroaniline on the basis of their "F-n.m.r. resonances. Furthermore the metabolism of 2-fluoroaniline was studied in several model systems from the in vitro microsomal system to the hepatocyte and the perfused liver system, to urine analysis after in vivo biotransformation. The results clearly demonstrate the usefulness of "F-n.m.r. for biomonitoring and/or for studying the biotransformation of fluorocontaining xenobiotics.
Experimental Reference compounds 2-Fluoroaniline was purchased from Janssen Chimica (Beerse, Belgium). Potassium fluoridewas from Merck (Darmstadt, FRG). 2-Fluoroacetanilide was synthesized as described by Vogel(1978). 2-Amino3-fluorophenol was synthesized from 3-fluorophenol (Janssen, Beerse, Belgium) following the procedure described by Aymes and Paris (1980) for the synthesis of 4-amino-2-fluorophenol. Treatment of the 3fluorophenol with liquid nitrogen dioxide resulted in a reaction mixture containing 2-nitro-3fluorophenol and 2-nitro-5-fluorophenol which were separated on a silicagel 60 column (70-230 mesh) (Merck, Darmstadt, FRG) using toluene as eluant. 2-Nitro-3-fluorophenol was reduced to 2-amino-3fluorophenolunder 4 atm. H, (24 h) with Pd/C (0.1% w/v) as catalyst. Likewise, 4-amino-3-fluorophenol was synthesized from 4-nitro-3-fluorophenol. 4-Acetamido-3-fluorophenolwas synthesized from 4amino 3-fluorophenol by acetylation carried out as described by Vogel (1978). Chemical detection of 4-amino-3-jtuorophenol 4-Amino-3-fluorophenol was determined by the method described by Brodie and Axelrod (1948) for the detection of para-hydroxylated aniline (4-aminophenol), modified as follows: to 0 8 ml precipitated supernatant (1 ml incubation mixture +0.3 ml20% trichloroacetic acid (TCA)), 8Opl5% phenol reagent (5%w/v phenol in 2 . 5 NaOH) ~ and 160p12.5 M Na,CO, were added. 'Ihe absorption at 63Unm was measured after 45 min at room temp. (~630=260rnM-' cm-') (Rietjens and Vervoort 1989). Preparation of microsomes Microsomes were prepared from the perfused livers of male Wistar rats (250-300g), which were untreated (control) or treated with 3-methylcholanthrene (30mg/kg body wt, injected i.p. in olive oil for 3 days). Following homogenizationof the livers in Tris-sucrose buffer (50 mM Tris, 025 M sucrose, pH 7.4 with HCI) and centrifugation at lOO00g (20min) the supernantants were centrifuged for 75min at 105OOOg. The microsomal pellet was washed once with Tris-sucrose buffer and finally suspended in 0 1 M potassium phosphate (pH 7.6) containing 20% glycerol and 1 mM EDTA, immediately frozen in liquid nitrogen and stored at -90°C. Microsomal protein and cytochrome P-450 content were determined as described by Rutten et al. (1987). Microsomal incubations Microsomal incubations were carried out at 37°C in 0 1 M potassium phosphate (pH 7.6) containing 3 mM substrate added as 1% (v/v) of a 0 3 M stock solution in dimethyl sulphoxide (DMSO). The reaction was started by the addition of NADPH (1'0mM final concentration) and terminated after 20min by freezing the reaction mixture in liquid nitrogen, for igF-n.m.r. measurements, or by adding 1.0ml of the reaction mixture to 0 3 ml 20% TCA, for chemical detection of 4-amino-3-fluorophenol. Isolation of hepatocytes Hepatocytes were isolated from the livers of male Wistar rats essentially as described by Paine et al. (1979). In short, the isolation was based on perfusion of the liver with a 0.05%collagenase solution (from Clostridiwn histolyticum) (Boehringer, Mannheim, FRG) in 2.5 mM CaCI,, 7 5 % NaHCO, in Hank's balanced salt solution.
Metabolism of 2-fluoroaniline by 'gF-n.m.r.
659
Before use, cells were washed twice with William's E medium (Flow, Irvine, Scotland) or with a modified Krebs Ringer solution (see below). The viability of the cells (determined by trypan blue exclusion) was 85-95%. In vitro metabolism of 2-jwoamline by hcpatocytes After washing, hepatocytes were resuspended in William's E medium or Krebs Ringer solution up to 1.5 x lo6 viable cells per ml. At t=Omin, 1% (v/v) 03M 2-fluoroaniline in DMSO was added to the hepatocytes in suspension to give a h a l concentration of 3 mM 2-fluoroaniline. Incubations were carried out at 37°C and the pH and oxygen content of the medium was stabilized by flushing with O,/CO, (20% IS%). At 15min time intervals, samples taken from the cell suspension were frozen in liquid nitrogen.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Liver perfusion For liver perfusion the procedure used for the isolation of hepatocytes (Paine et al. 1979)was modified by replacing the collagenasesolution by 100ml perfusion buffer containing 3 mM 2-fluoroanilineadded as 1 ml 0-3 M stock solution in DMSO. This buffer was either:
1. a modified Krebs Ringer solution containing 131mM NaCI, 6 m KCl. ~ 1mM MgCl,, 1 mM CaCl, and 2 0 m glucose, ~ buffered with 25mM HEPES/NaOH pH7-3 (Evelo et al. 1984) or 2. a buffer containing sulphate based on the one described for liver perfusion by Eyer et al. (1980) containing 11OmM NaCl, 5 - 9 m KCl, ~ 1.2mM MgCl,, 2 3 m CaCI,, ~ 1.2m~ NaH,PO,, 6 m ~ Na,SO,, 75mM glucose and 25mM NaHCO,, equilibrated with 0 2 / C 0 , (2Vk/5%). The flow used for liver perfusion was 30ml/min. At different time intervals samplesfrom the recirculating perfusion buffer were frozen in liquid nitrogen. In vivo exposure to 2-jwoamline Male Wistar rats (250-300g) were exposed to 5Omg 2-fluoroaniline per kg body weight by oral administration of 2-fluoroaniline in olive oil. Before the administration a 24 h control urine sample was collected. After oral dosing of 2-fluoroaniline, 24 h urine samples were collected. En2yme hydrolysis of urine samples Enzyme hydrolysis of urine samples was carried out essentially as described by Evelo et al. (1984). In short, for /?-glucuronidasetreatment, 1200pl02 M KH,PO,/Na,HPO, pH 6 2 containing 8 units of /?glucuronidase (from E. coli Klz) (Boehringer, Mannheim, FRG) were added to 1 2 0 0 ~of 1 urine sample and the mixture was incubated for 1 h at 37°C. For arylsulphatase//?-glucuronidasetreatment 4Opl of the enzyme mixture from Helix pomutxa (Boehringer, Mannheim, FRG) were added to 1200pl of urine sample diluted with 1 2 0 0 ~ 10 - 1 ~ potassium acetate pH 52. Samples were incubated for 16h at 37°C.
"F-n.m.r. measurements "F-n.m.r. measurements were performed on a Bruker CXP 300 spectrometer operating at 282.3 MHz with a lOmm Bruker 19Fobserve (free of fluor background), 'H decouple probe and a 2 20A 19F preamplifier tuned at 282MHz. Wilmad (Buena, USA) lOmm n.m.r. tubes were used. The sample volume was 1-7ml, containing 100pl 'H,O for locking the magnetic field. Proton decoupling was achieved with the waltz-16 pulse sequence (Shaka et at. 1983)at - 16 dB from 20 W. Nuclear Overhauser effects on the signals were eliminated using the inverse gated decoupling technique. Spectra were obtained with 30" pulses (6 p),a 20 kHz spectral width, repetition time of 1 s, quadrature phase detection and quadrature phase cycling (CYCLOPS). Between loo00 and 4OOOO scans were recorded. The free induction decays were multiplied by an exponential decay resulting in an increase of the linewidth by 5 Hz. 4-Fluoroaniline (3 mM in lOOmM potassium phosphate, pH 76,7"C) served as a reference for the 19F measurements. Chemical shifts are reported relative to CFCl, using a conversion of - 1300ppm (64fluoroaniline-6CFC1, = - 130.0ppm) (Wray 1983). Intensities of the various metabolites observed in the "F-n.m.r. spectra were determined from the integrals of their ''F-n.m.r. resonances. Parafluorobenzoate was used as an internal standard for the determination of absolute concentrations. During the acquisition of the 19Fsignal the extremely strong 'H decouple signal is present in the probe. Although the leakage of the 'H decouple signal in the I9Foutput of the probe is small, the output of the 'H transmitter was passed through a Bruker 300 MHz bandpass and a 282 MHz bandstop filter to be sure that a clean 300 MHz signal reaches the probe. A Bruker 300 MHz bandstop and a 282 MHz bandpass filter is placed at the input of the I9F preamplifier to prevent the 300MHz signal from overloading the preamplifier. These two filters were carefully tuned to maximum sensitivity of the 19F signal. Further attenuation of the interfering signal (of the 'H decoupler) was obtained by placing a home-built 90MHz bandstop filter in the 72MHz intermediate frequency amplifier of the ''F preamplifier. Furthermore, the gain of the preamplifierwas set to the low position and the gain of the low-frequency amplifier of the main receiver was set to the high position to prevent the phase-sensitive detectors of the main receiver from overloading.
J. Vervoort et al.
660
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Results Identification of the lgF-n.rn.r. resonances The 'gF-n.m.r. resonances of the fluoride ion (F-), 2-fluoroaniline, 2-fluoroacetanilide, 4-amino-3-fluorophenol and 2-amino-3-fluorophenol were identified on the basis of their "F-n.m.r. spectra acquired under our standard experimental conditions (0.1 M potassium phosphate pH 7-6, 7°C) (table 1). The "F-n.m.r. resonance of 4-acetamido-3-fluorophenolwas identified based on results obtained from a microsomal incubation with 2-fluoroacetanilide (figure 1). From the "Fn.m.r. spectrum presented in figure 1 it is seen that the microsomal metabolism of 2fluoroacetanilide results in deacetylation, yielding 2-fluoroaniline, in the formation of 4-amino-3-fluorophenol and in one additional metabolite at - 125.7 ppm, representing 4-acetamido-3-fluorophenol.This is concluded from the fact that chemical acetylation of 4-amino-3-fluorophenol resulted in one main product, having its lgF-n.m.r. resonance at - 125.7ppm (data not shown). Finally, the resonances of 4-acetamido-3-fluorophenylsulphate, 4-amino-3fluorophenyl sulphate, 4-acetamido-3-fluorophenyl glucuronide and 4-amino-3Ruorophenyl glucuronide were identified from the lgF-n.m.r. spectra of urine samples in which these glucuronidated or sulphated urine metabolites were PY
C-0
&' I
OH
I I I I
Y YO
&'
i I
I
I
I
I
f
I
F
c
F-
OH
1
I
!
I
-115
I
-120
I
I
-125
-130
I
I
-135
-140
PPM (6) "F-n.m.r. spectrum of a microsomal suspension incubated with 2-fluoroacetanilide.
Figure 1. Microsomes were from 3-methylcholanthrene-treated rats. The resonances marked with an asterisk were present in control experiments without NADPH.
66 1
Metabolism of 2-jhoroaniline by "F-n.m.r.
Table 1. Chemical shifts of "F-n.m.r. resonances of identified 2-fluoroaniline metabolites in 0.1 M potassium phosphate pH 76; 7°C. ~~
~
Chemical shift (pprn)'
Compound
- 1384 - 128.5 - 123.0 - 1349 - 1347
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
2-Fluoroaniline 2-Fluoroacetanilide Fluoride ion 4-Amino-3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Acetamido-3-fluorophenol 4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate
-135.1 - 1257 -1250 125.1
-
'The chemical shifts are relative to CFCl,.
Table 2.
''F-n.m.r.
characteristics of urine samples of a 2-fluoroaniline-exposed rat. Urine sample (concentration in mM)
Compound Fluoride ion (F-) 4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate 4-Acetamido-'3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Amino-3-fluorophenol Unknown (- 1354ppm) Unknown (- 1382 ppm) 2-Fluoroanaline
Untreated 0.36 084 1.25 0.15
0.44 3.74 0.36 0.53 067
fl-Glucuronidase treated
Arylsulphatasel fl-glucuronidase treated
0.31 1.14 1.00 3.71 0.41 0.40 0.40 0.80
0.32 210 401 0.24 0-06 0.93
Urine samples were either untreated, treated with B-glucuronidase or with arylsulphataselflglucuronidase as described in the Experimental section. para-Fluorobenzoate was added as internal standard for determination of concentrations. -=Not observed.
hydrolysed to their corresponding phenols using fl-glucuronidase or arylsulphatase, respectively. fl-Glucuronidase treatment of urine resulted in the loss of the resonance at - 1347ppm with a concomitant proportional increase of the resonance at - 134.9 ppm representing 4-amino-3-fluorophenol (table 2). In the same experiment the resonance at - 125.0 ppm disappeared with a new resonance of similar intensity appearing at - 125-7ppm representing 4-acetamido-3-fluorophenol.Incubation of an urine sample containing fluorinated metabolites with an arylsulphatase/fl-glucuronidase preparation resulted not only in the loss of the signals of the glucuronidated compounds but also in the disappearance of the resonances at - 135.1 ppm and at - 125.0ppm, giving rise to proportional increases of the resonances at respectively - 1349ppm (4-amino-3-fluorophenol)and at - 125.7ppm (4-acetamido-3-fluorophenol)(table 2). Table 1 summarizes the "F-n.m.r. chemical shifts of various 2-fluoroaniline derived metabolites.
J . Vervoort et al.
662
2-FA
&' Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
OH
-115
-120
-125
-130
-135
-140
PPM (6) Figure 2. ''F-n.m.r. spectrum of a microsomal suspension incubated with 2-fluoroaniline. The resonances marked with an asterisk were present in control experiments without NADPH.
Microsomal conversion of 2-fluoroaniline The "F-n.m.r. spectrum presented in figure 2 clearly indicates 4-amino-3fluorophenol as the main metabolite formed during the incubation of 2-fluoroaniline with rat liver microsomes. In addition to the major para-hydroxylated 2-fluoroaniline metabolite, fluoride (F-) was a minor metabolite in these incubations. No 2-amin0-3-fluorophenol(- 129.9 ppm) that would have been formed upon orthohydroxylation was observed. It is therefore concluded that microsomal metabolism of 2-fluoroaniline results in formation of para-hydroxylated 2-fluoroaniline (4amino-3-fluorophenol) as the main metabolite. Biotransformation of 2-fluoroaniline by isolated rat hepatocytes Results obtained upon incubation of rat hepatocytes with 2-fluoroaniline are presented in figure 3. Metabolism by hepatocytes in Krebs Ringer buffer produces a substantial number of reaction products (figure 3 A). N-Acetylation is the metabolic pathway occurring at the highest rate, although para-hydroxylation of 2-fluoroaniline also occurred to a significant extent. Small amounts of phase I1 reaction products of 4-amino-3-fluorophenol were also observed. Using rat hepatocytes in Krebs Ringer buffer, glucuronidation of 4-amino-3-fluorophenol resulting in 4amino-3-fluorophenyl glucuronide was the major phase I1 reaction (figure 3). In these incubations 4-amino-3-fluorophenyl sulphate and 4-acetamido-3fluorophenyl sulphate were formed at a relatively low level. Furthermore, only a limited percentage of all 4-amino-3-fluorophenol appeared to be conjugated in these incubations. In a sulphate-rich medium, however (figure 3 B), all phenolic metabolites appeared to be conjugated with sulphate as no 4-amino-3-fluoropheno1 or 4amino-3-fluorophenyl glucuronide could be observed. Conversion of 2-fluoroaniline by a perfused rat liver system Figure 4 shows the results of the metabolism of 2-fluoroaniline in perfused rat liver. The similarity of the spectra presented in figure 4 and in figure 3 (conversion by
663
Metabolism of 2-fluoroaniline by IgF-n.m.r.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
A
n B
-115
-120
-125
-130
-135
-140
PPM ( 6 ) ''F-n.m.r. spectra obtained upon incubation of hepatocyteswith 2-fluoroaniline (A) in Krebs Ringer buffer and (B) in William's E medium. Both contained 3 mM 2-fluoroaniline. Spectra are from the samples taken after Wmin incubation. The resonances marked with an asterisk were present in the incubations at t =0 min.
Figure 3.
664
. I . Vmwoort et al.
&' I
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
A
I
I
i B
I -115
I
-120
I
-125
I -130
I -135
I -140
PPM ( 6 ) Figure 4. lgF-n.m.r. spectrum of the metabolism of 2-fluoroaniline in a perfused rat liver system. The perfusion buffer (100 ml) was recirculated and contained 3 mM 2-fluoroaniline in a modified Krebs Ringer solution without sulphate (A) or in an O,/CO, buffered perfusion buffer containing 6 . 0 m sulphate ~ (B). Spectra were taken after 60min incubation. The resonances marked with an asterisk were already observed at t=Omin.
Metabolism of 2-fluoroaniline by lgF-n.m.r.
665
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
rat hepatocytes) is evident. The main difference is that under sulphate-limited conditions liver perfusion, in contrast to hepatocyte incubations, did not give rise to formation of glucuronidated species.
In vivo metabolism of 2-fluoroaniline by Wistar rats Figure 5 shows the lgF-n.m.r. spectrum of a urine sample of a rat exposed to 2-fluoroaniline. After addition of 2-fluoroaniline the parent compound was rapidly metabolized and over 90% of the administered dose was detected in 24 h urine. The lgF-n.m.r. spectrum (figure 5 ) shows formation of about nine major urinary metabolites. Seven of these metabolic products have been unambiguously assigned (table 1). The relative amount of the various fluoro-containing metabolites is presented in table 3.4-Amino-3-fluorophenylsulphate appears to be the major urine metabolite. In addition to these major metabolites at least 16 minor products (each > 0 5 % of the total "F intensity) can be observed. No attempt was made to assign these resonances. Metabolites carrying an amino group can be easily discriminated from N acetylated products as the latter do not shift in the lgF-n.m.r. spectrum on acidifying the urine. The apparent overlap of the resonances of 4-acetamido-3-fluorophenyl-sulphate and -glucuronide is dependent on conditions (pH, temm). The difference between
I
-115
I
-120
I
-125
I -130
I -135
I
-140
PPM ( 6 ) Figure 5. lgF-n.m.r.spectrum of the 24 h urine of a male Wistar rat orally exposed to 2-fluoroaniline. 2-Fluoroaniline (50mg/kg) was administered dissolved in olive oil.
J. Vervoort et al.
666 Table 3.
Comparison of rat urine metabolite patterns of 2-fluoroaniline as determined by I9F-n.m.r. and by using radioactive t.1.c. analysis as described by Eadsforth et ul. (1986). Present study: percentage of total "F-intensity mean & SEM (n=4 rats)
Identified compound
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
F-
-120
?
1.6k0.5 9.9 f 2.3 18.1 f1.3 05f01 3.3 It05 53.5 f27 0.0 3.8 f 0 2 2.6 k1.2 4.9 & 1.4
4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate 4-Acetamido-3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Amino-3-fluorophenol Unknown ( - 135.4ppm) Unknown (- 138.2 ppm) 2-Fluoroaniline
-115
Eadsforth et ul. (1986): percentage of dose recovered in urine
-125
-130
48 44.1 08 1.o 27.8 00 63
1.8 0.0
-135
-140
PPM ( 6 ) Figure 6. "F-n.m.r. spectrum of the urine of a 2-fluoroaniline exposed rat showing the detection limits for a sample measured in an overnight run on a Bruker CXP-300 spectrometer using a dedicated 19F-n.m.r. lOmm probehead. ) added as an internal reference for quantitative Para-fluorobenzoate (final concentration 10p ~ was measurements. The peaks labelled with an asterisk are present in a concentration of 1 PM. This spectrum is identical to the one shown in figure 5, with the vertical scale magnified about 150times.
Metabolism of 2-fluoromiline by IgF-n.m.r.
667
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
the chemical shift of the two resonances varies between 20 and 40Hz (0070.13ppm). Because the intrinsic linewidth of both resonances is 5 Hz, the two peaks are well enough resolved to determine their integrals accurately. In addition, this quantification of 4-acetamido-3-fluorophenyl-glucuronide and -sulphate was checked by the addition of 20% (v/v) acetonitrile to the urine sample. Under these conditions the two resonances were separated by 0-3 ppm, making quantification easier. Identical data were obtained. Sensitivity of the lgF-n.m.r. method The results presented in figure 6 clearly demonstrate that the detection limit of the lgF-n.m.r. method for fluoro-containingcompounds is 1 /AMfor an Overnight run using a dedicated lOmm "F probehead and a Bruker CXP 300 spectrometer. The main metabolites observed in the urine spectra are present in concentrations ranging between 0.1 and 4 mM, whereas the concentration of the minor products is between 1 and 100PM. It was checked that, during an overnight incubation of a urine sample at 7"C, no decomposition and/or bacterial degradation of labile urinary fluorocontaining compounds occurs. Finally, figure 7 presents a calibrationcurve of 2-fluoroacetanilidein urine from a non-exposed animal. It can be seen that urine components do not disturb the measurements. Linearity is observed for a broad concentration range.
[2-fluoroacetanilide]
in mM
Figure 7. Calibration curve for 2-fluoroacetanilide in urine from an untreated rat. 2-Fluoroacetanilide was added as 05% (v/v) of a 200 times concentrated stock in dimethyl sulphoxide. Relative "F-n.m.r. intensities were quantified from spectra obtained in 30 min.
668
f . Vwvoort et a).
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Discussion After administration of 2-fluoroaniline to rats, the compound is rapidly metabolized and over 90% of the dose is excreted in the urine within 24 h. In the 19Fn.m.r. spectra of these urine samples nine major products were observed of which seven have been assigned (figure 5 , table 3). From these assignments it can be seen that the major biotransformation route is via para-hydroxylation. No orthohydroxylated metabolites, which are known to be formed to a substantial amount upon biotransformation of aniline derivatives (Eyer et al. 1980, Boyland et al. 1953), were observed as major excretion products. This observation of para-hydroxylation being by far the dominating pathway is in agreement with results obtained with microsomal incubations (figure 2), hepatocyte incubations (figure 3) and liver perfusions (figure 4). Microsomal incubation studies give similar results to those presented recently for monofluorinated anilines which showed that parahydroxylation is by far the dominant reaction (Rietjens and Vervoort 1989). In addition, the present paper has shown that upon in vivo exposure to 2fluoroaniline, para-hydroxylation is mainly followed by sulphation and to a minor extent by glucuronidation. Sulphation being the dominating phase I1 reaction in rats is in accordance with other published data on metabolism of aniline derivatives in rats (Baldwin and Hutson 1980, Eadsforth et al. 1984, 1986). In the present study these sulphated species were also clearly observed in hepatocyte incubations and in perfused liver (figures 3 and 4). No glucuronidated metabolites were observed in perfused liver, not even when using perfusion buffers with low sulphate concentration. The absence of glucuronidated species in perfused liver may be related to a higher intracellular concentration of sulphate (in the cells of the perfused liver) than occurs in isolated hepatocytes. In addition, Evelo et al. (1984) has shown, when studying phase I I dependent reactions of 4-aminophenol in isolated hepatocytes, that the K, for sulphation of the hydroxy function was lower than for glucuronidation. This implies that, even at low sulphate concentration, sulphation of 4aminophenol is favoured over glucuronidation. Excretion of glucuronidated metabolites into bile during liver perfusion cannot explain the fact that these metabolites are not observed in the liver perfusion. This follows from the fact that perfusion was carried out in such a way that excreted bile would end up in the perfusion medium. Furthermore, the observation that more than 90% of the dose of 2-fluoroaniline was detected in the urine excludes the faeces (and thus the bile) as a major route of excretion for glucuronidated 2-fluoroaniline derivatives. In lgF-n.m.r. measurements of faeces no fluorinated metabolites were observed. In incubation experiments with hepatocytes or in perfused liver systems (figures 3 and 4) N-acetylation is the major phase 11type reaction. However, inurine samples only 29% of the assigned metabolites were N-acetylated (table 3). So the ratio between N-acetylation and para-hydroxylation of 2-fluoroaniline observed in vitro in favour of N-acetylation, is not reflected in a high percentage of N-acetylated urine metabolites. Deacetylation of N-acetylated products of 2-fluoroaniline in the liver and/or the kidney could explain this discrepancy. In the present study deacetylation was indeed observed in microsomal incubations with 2-fluoroacetanilide (figure 1). Furthermore, during the 60 min of hepatocyte incubations or 90 min of liver perfusion, hydroxylation of 2-fluoroacetanilide was not observed to any significant extent because, compared to 2-fluoroaniline, the concentration of 2-fluoroacetanilide is low. However, during the in vivo experiment, metabolism continues for longer than
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Metabolism of 2-fluoroaniline by 'gF-n.m.r.
669
these time periods. Taking into account the high rate of N-acetylation, observed in both in vitro systems during the first hours of metabolism, the in vivo concentration of 2-fluoroacetanilide available for hydroxylation can be expected to increase. As a result the relative amount of para-hydroxylated acetylatedmetaboliteswill be higher upon prolonged in vivo metabolism than in in vitro metabolism. Comparison of our results with those of Eadsforth et al. (1986) with 14C-2fluoroanilineshows that the same type of metabolites are formed (table 3). The major difference between these two studies is in the amounts of N-acetylated products observed (table 4). 4-Acetamido-3-fluorophenylsulphate is reported to be the main metabolite present in the urine in the study of Eadsforth et al., whereas we observed 4-amino-3-fluorophenyl sulphate to be the main metabolite. On performing our experiment under the conditions used by Eadsforth et al. (1986) (i.e. oral administrationof 23-5mg 2-fluoroaniline/kgbody-weight, dissolved in olive oil) we observed the same metabolic pattern as shown in table 3, so that the difference between the two studies cannot be a dose-related effect. Another possible explanation might be that the two major metabolites, 4-amino-3-fluorophenyl sulphate and 4-acetamido-3-fluorophenylsulphate, have been confused in the study of Eadsforth et al. (1986) as the authors assumed that the two peaks, although not well resolved on t.l.c., eluted in the same order from the h.p.1.c. column. Alternatively, the rats used in our study and those of Eadsforth et al. (1986) are genotypically different. It is well known that within species fast and slow acetylators occur (Weber and Levy 1988). The differencesobserved, with regard to the amount of N-acetylated products present in the urine spectra, may therefore be explained by either a slow acetylator genotype (our study), or a fast acetylator genotype (Eadsforth et al. 1986) of the rats used. Finally it is concluded that the IgF-n.m.r. method described in the present paper has important advantagesfor both biomonitoringand biotransformationstudies, not only of 2-fluoroaniline but also of other fluoro-containing compounds. Firstly, "Fn.m.r. spectra, and thus metabolic patterns, can be obtained directly from urine samples without any treatment. Secondly, the method is shown to be a sensitive tool as the lower limit of detection is 1 PM, which can be decreased to about 50 nM (or even less) if concentration procedures such as freeze-drying of urine, are incorporated. Alternatively concentration procedures can be used to limit the "Fn.m.r. acquisition time, needed to obtain 1PM detection limit in the original urine. It can be calculated that upon concentration of urine samples by a factor 10, 1PM sensitivity in the original urine (i.e. 1 O p in ~ the concentratedsample) is obtained in a 15-min run. Finally, all fluoro-containing metabolites can be observed in one single Table 4. Relative routes of in viw metabolism of 2-fluoroaniline by male Wistar rats observed using I9F-n.rn.r. and reported by Eadsforth ct al. (1986) using a radioactive t.1.c. method. Percentage of total I9F metabolities (n=4 rats)
Reaction ~
Eadsforth et al. (1986)
~~
para-Hydroxylation N-Acetylation Sulphation Glucuronidation
100 33 84
15
100
68 96 8
Calculations are based on the relative amounts of identified metabolites presented in table 3, taking para-hydroxylation as 100% reference.
670
Metabolism of 2-jtuoroaniline by 'F-n.m.r.
spectrum, making the method independent of the relative amount of products present and of genetic polymorphisms, in contrast to alternative methods where urine samples are screened for the presence of one metabolite only. N-acetylation, for example, is known to show genetic polymorphism. The "F-n.m.r. method, detecting all 2-fluoroaniline metabolites, and not only 4-amino-3-fluorophenylsulphate, has the advantage of reducing the risk of false negatives for fast acetylators.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Acknowledgements Thanks are extended to Ms L. Ausma and Ms S. Fulton for typing the manuscript, to Mr M. M. Bouwmans for preparing the figures and to Mr J. W. M. Haas for help with animal handling.
References AYMES,J. D., and PARIS,M. R., 1980, Etude des complexes du cobalt (11) transporteurs d'oxygene: nouvelle synthese du fluoro-3-hydroxy-2-benurldehyde.Bulletin de la Societe Chimique de France, 3 4 , 11-175-11-178. BALDWIN, M. K., and HUTSON,D. H., 1980, The metabolism of 3-chloro-4-fluoro-aniline in dog and rat. Xenobiotica, 10, 135-144. BOYLAND, E., MANSON, D., and SIMS,P., 1953, The preparation of 0-aminophenyl sulphates. Journal of the Chemical Society, 3623-3628. BRODIE, B. B., and AXELROD, J., 1948, The estimation of acetanilide and its metabolic products, aniline, N-acetylp-aminophenol andp-aminophenol (free and conjugated) in biological fluids and tissues. Journal of Pharmacological and Experimental Therapeutics, 94,22-28. EADSFORTH, C. V., LOGAN, C. J., MORRISON, B. J., and WARBURTON, P. A., 1984,2-4-Diffuoroanilineand 4-fluoroaniline exposure: monitoring by methaemoglobin and urine analysis. International Archives of Occupational and Environmental Health, 54, 223-232. EADSFORTH, C. V., COVENEY, P. C., HUTSON,D. H.,LOGAN,C. J., and SAMUEL, A. J., 1986, The metabolism of o-fluoroaniline by rats, rabbits and marmosets. Xenobiotica, 16, 555-566. EVELO, C. T. A., VERSTEEGH, J. F. M., and BLAAUBOER, B. J., 1984, Kinetics of the formation and secretion of the aniline metabolite 4-aminophenol and its conjugates by isolated rat hepatocytes. Xenobiotica, 14, 409-416. EYER,P., KAMPFFMEYER, H., MAISTER,H., and ROSCH-OEHME, E., 1980, Biotransformation of nitrosobenzene, phenylhydroxylamine, and aniline in the isolated perfused rat liver. Xenobiotica, 10,499-516. GROSSMAN, S . J., and JOLLOW, D. J., 1986, Use of the NIH shift to determine the relative contribution of competing pathways of aniline metabolism in the rat. Drug Metabolism and Disposition, 14, 689691. PAINE, A. J., WILLIAMS, L. J., and LEGC,R. F., 1979, In The Liver: Quadtotiwe Aspects of Structure and Function, edited by R. Preisig and J. Bircher (FRG, Aulendorf), 215-263. RIETJENS, I. M. C. M., and VERVOORT, J., 1989, Microsomal metabolism of fluoroanilines. Xenobiotica, 19, 1297-1305. RUTTEN, A. A. J. J. L., FALKE,H. E., CATSBURG, J. F., TOPP,R., BLAAUBOER, B. J., HOLSTEIJN, L. VAN DOORN,L., and LEEUWEN, F. X. R. VAN,1987, Interlaboratory comparison of total cytochrome P-450and protein determination in rat liver microsomes. Archives of Toxicology, 61, 27-33. SHAKA, A. J., KEELER, J., and FREEMAN, R., 1983, Evaluation of a new broad band decoupling sequence: Waltz 16. Journal of Magnetic Resonance, 53,313-340. VOGEL, A. J., 1978, In Textbook of Practical Inorganic Chemistry, (Harlow: Longman), p. 684. WEBER,W. W., and LEVY,G. N., 1988, In Metabolism of Xenobiotics, edited by Gorrod, J. W., Oelschlager, H., and Caldwell, J. (London: Taylor & Francis), pp. 209-215. WRAY,V., 1983, Annual Reports on N M R Spectroscopy, edited by Webb, G. A. (London: Academic Press), Vol. 14, p. 252.
1990, VOL. 20, NO. 7, 657670
Development of a lqF-n.m.r.method for studies on the in vivo and in vitro metabolism of 2-fluoroaniline J- VERVOORTT, P. A. DE JAGERS, J. STEENBERGENT and I. M. C. M. RIETJENStO
t Department of Biochemistry, Agricultural University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands
# Department of Molecular Physics, Agricultural University, Dreijenlaan 3,
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
6703 HA Wageningen, The Netherlands Received 8 September 1989; accepted 1 March 1990 1. A IgF-n.m.r. method has been developed for study of the metabolism of 2fluoroaniline both after in wiwo exposure of rats and in in oitro model systems. 2. From the IgF-n.m.r. spectrum of the 24 h urine it was caIcuIatedthat over 90% of the dose was excreted within 24 h. The metabolic pattern showed that 85% of the metabolites were para-hydroxylated, 72% sulphated, 13% glucuronidated and 29% N-acetylated, 4amino-3-fluorophenyl sulphate being the main urinary metabolite (53%). 3. In witro studies of phase I metabolism of 2-fluoroaniline with rat liver microsomes was representative for the in vim metabolism as hydroxylation in both systems was observed only at the para-position. 4. Phase I I1 metabolism was studied in witro in either isolated rat hepatocytes in suspension or in a 1 h recirculating liver perfusion system. In both these in witro systems para-hydroxylation, N-acetylation, sulphation and glucuronidation of 2-fluoroaniline were observed. The ratio between glucuronidation and sulphation was dependent on sulphate availability 5. Of the in vitro systems tested, hepatocytes in Krebs Ringer (sulphate limited) medium was the best model for in wiwo metabolism. 6. The detection limit for fluoro-containingmetabolites in this IgF-n.m.r. method was 1 p~ for an overnight run using a Bruker CXP 300 spectrometer. From this it can be concluded that lgF-n.m.r. urine analysis is a useful tool in biomonitoring studies. For 2fluoroanilinethe method appears to be more sensitive than currently available h.p.l.c./t.l.c. methods. In addition, concentration of urine samples can result in either lower detection limits, or in shorter times needed for n.m.r. data acquisition. 7. N-acetylation is known to show genetic polymorphism. Therefore, the lgF-n.m.r. method, detecting all 2-fluoroaniline metabolites, has the additional advantage of eliminating the risk of obtaining false negatives for fast acetylators.
+
Introduction 2-Fluoroaniline is used as an intermediate in the manufacturing of herbicides and plant growth regulators. Plant workers may be exposed to this chemical (Eadsforth et aZ. 1986). Although much is known about the metabolism of anilines in general (Boyland et al. 1953,Baldwin and Hutson 1980,Grossman and Jollow 1986), specific knowledge concerning the biotransformation of this fluorinated aniline is restricted to only one in vivo study (Eadsforth et al. 1986). We therefore decided to study the metabolism of 2-fluoroaniline in more detail using microsomal, hepatocyte and perfused liver systems in addition to in viwo exposures. Eadsforth et aZ. (1986) detected eight metabolites in the urine of exposed animals; six of these metabolites were identified. 4-Acetamido-3 -fluorophenyl sulphate was reported to be the main
8 To whom correspondence should be addressed. 0049-8254/90 83.00 0 1990 Taylor & Francis Ltd.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
658
J . Vervoort et al.
urine metabolite. In this study of Eadsforth et al., 14C-labelled 2-fluoroaniline was used and urinary 14C-labelledmetabolites were separated on t.1.c. and/or h.p.1.c. We decided to study the metabolism of 2-fluoroaniline using 'gF-n.m.r. This technique enabled us to perform direct measurements afurine samples, as well as of samples from liver perfusion and from hepatocyte or microsomal incubations. Additional advantages of the "F-n.m.r. technique, making it suitable for xenobiotic biotransformation studies, are the high sensitivity of the chemical shift of Auorocontaining compounds to chemical modifications, the absence of natural background signals and the high sensitivity which almost equals that of 'H-n.m.r. The present paper describes the identification of the various fluoro-containing metabolites of 2fluoroaniline on the basis of their "F-n.m.r. resonances. Furthermore the metabolism of 2-fluoroaniline was studied in several model systems from the in vitro microsomal system to the hepatocyte and the perfused liver system, to urine analysis after in vivo biotransformation. The results clearly demonstrate the usefulness of "F-n.m.r. for biomonitoring and/or for studying the biotransformation of fluorocontaining xenobiotics.
Experimental Reference compounds 2-Fluoroaniline was purchased from Janssen Chimica (Beerse, Belgium). Potassium fluoridewas from Merck (Darmstadt, FRG). 2-Fluoroacetanilide was synthesized as described by Vogel(1978). 2-Amino3-fluorophenol was synthesized from 3-fluorophenol (Janssen, Beerse, Belgium) following the procedure described by Aymes and Paris (1980) for the synthesis of 4-amino-2-fluorophenol. Treatment of the 3fluorophenol with liquid nitrogen dioxide resulted in a reaction mixture containing 2-nitro-3fluorophenol and 2-nitro-5-fluorophenol which were separated on a silicagel 60 column (70-230 mesh) (Merck, Darmstadt, FRG) using toluene as eluant. 2-Nitro-3-fluorophenol was reduced to 2-amino-3fluorophenolunder 4 atm. H, (24 h) with Pd/C (0.1% w/v) as catalyst. Likewise, 4-amino-3-fluorophenol was synthesized from 4-nitro-3-fluorophenol. 4-Acetamido-3-fluorophenolwas synthesized from 4amino 3-fluorophenol by acetylation carried out as described by Vogel (1978). Chemical detection of 4-amino-3-jtuorophenol 4-Amino-3-fluorophenol was determined by the method described by Brodie and Axelrod (1948) for the detection of para-hydroxylated aniline (4-aminophenol), modified as follows: to 0 8 ml precipitated supernatant (1 ml incubation mixture +0.3 ml20% trichloroacetic acid (TCA)), 8Opl5% phenol reagent (5%w/v phenol in 2 . 5 NaOH) ~ and 160p12.5 M Na,CO, were added. 'Ihe absorption at 63Unm was measured after 45 min at room temp. (~630=260rnM-' cm-') (Rietjens and Vervoort 1989). Preparation of microsomes Microsomes were prepared from the perfused livers of male Wistar rats (250-300g), which were untreated (control) or treated with 3-methylcholanthrene (30mg/kg body wt, injected i.p. in olive oil for 3 days). Following homogenizationof the livers in Tris-sucrose buffer (50 mM Tris, 025 M sucrose, pH 7.4 with HCI) and centrifugation at lOO00g (20min) the supernantants were centrifuged for 75min at 105OOOg. The microsomal pellet was washed once with Tris-sucrose buffer and finally suspended in 0 1 M potassium phosphate (pH 7.6) containing 20% glycerol and 1 mM EDTA, immediately frozen in liquid nitrogen and stored at -90°C. Microsomal protein and cytochrome P-450 content were determined as described by Rutten et al. (1987). Microsomal incubations Microsomal incubations were carried out at 37°C in 0 1 M potassium phosphate (pH 7.6) containing 3 mM substrate added as 1% (v/v) of a 0 3 M stock solution in dimethyl sulphoxide (DMSO). The reaction was started by the addition of NADPH (1'0mM final concentration) and terminated after 20min by freezing the reaction mixture in liquid nitrogen, for igF-n.m.r. measurements, or by adding 1.0ml of the reaction mixture to 0 3 ml 20% TCA, for chemical detection of 4-amino-3-fluorophenol. Isolation of hepatocytes Hepatocytes were isolated from the livers of male Wistar rats essentially as described by Paine et al. (1979). In short, the isolation was based on perfusion of the liver with a 0.05%collagenase solution (from Clostridiwn histolyticum) (Boehringer, Mannheim, FRG) in 2.5 mM CaCI,, 7 5 % NaHCO, in Hank's balanced salt solution.
Metabolism of 2-fluoroaniline by 'gF-n.m.r.
659
Before use, cells were washed twice with William's E medium (Flow, Irvine, Scotland) or with a modified Krebs Ringer solution (see below). The viability of the cells (determined by trypan blue exclusion) was 85-95%. In vitro metabolism of 2-jwoamline by hcpatocytes After washing, hepatocytes were resuspended in William's E medium or Krebs Ringer solution up to 1.5 x lo6 viable cells per ml. At t=Omin, 1% (v/v) 03M 2-fluoroaniline in DMSO was added to the hepatocytes in suspension to give a h a l concentration of 3 mM 2-fluoroaniline. Incubations were carried out at 37°C and the pH and oxygen content of the medium was stabilized by flushing with O,/CO, (20% IS%). At 15min time intervals, samples taken from the cell suspension were frozen in liquid nitrogen.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Liver perfusion For liver perfusion the procedure used for the isolation of hepatocytes (Paine et al. 1979)was modified by replacing the collagenasesolution by 100ml perfusion buffer containing 3 mM 2-fluoroanilineadded as 1 ml 0-3 M stock solution in DMSO. This buffer was either:
1. a modified Krebs Ringer solution containing 131mM NaCI, 6 m KCl. ~ 1mM MgCl,, 1 mM CaCl, and 2 0 m glucose, ~ buffered with 25mM HEPES/NaOH pH7-3 (Evelo et al. 1984) or 2. a buffer containing sulphate based on the one described for liver perfusion by Eyer et al. (1980) containing 11OmM NaCl, 5 - 9 m KCl, ~ 1.2mM MgCl,, 2 3 m CaCI,, ~ 1.2m~ NaH,PO,, 6 m ~ Na,SO,, 75mM glucose and 25mM NaHCO,, equilibrated with 0 2 / C 0 , (2Vk/5%). The flow used for liver perfusion was 30ml/min. At different time intervals samplesfrom the recirculating perfusion buffer were frozen in liquid nitrogen. In vivo exposure to 2-jwoamline Male Wistar rats (250-300g) were exposed to 5Omg 2-fluoroaniline per kg body weight by oral administration of 2-fluoroaniline in olive oil. Before the administration a 24 h control urine sample was collected. After oral dosing of 2-fluoroaniline, 24 h urine samples were collected. En2yme hydrolysis of urine samples Enzyme hydrolysis of urine samples was carried out essentially as described by Evelo et al. (1984). In short, for /?-glucuronidasetreatment, 1200pl02 M KH,PO,/Na,HPO, pH 6 2 containing 8 units of /?glucuronidase (from E. coli Klz) (Boehringer, Mannheim, FRG) were added to 1 2 0 0 ~of 1 urine sample and the mixture was incubated for 1 h at 37°C. For arylsulphatase//?-glucuronidasetreatment 4Opl of the enzyme mixture from Helix pomutxa (Boehringer, Mannheim, FRG) were added to 1200pl of urine sample diluted with 1 2 0 0 ~ 10 - 1 ~ potassium acetate pH 52. Samples were incubated for 16h at 37°C.
"F-n.m.r. measurements "F-n.m.r. measurements were performed on a Bruker CXP 300 spectrometer operating at 282.3 MHz with a lOmm Bruker 19Fobserve (free of fluor background), 'H decouple probe and a 2 20A 19F preamplifier tuned at 282MHz. Wilmad (Buena, USA) lOmm n.m.r. tubes were used. The sample volume was 1-7ml, containing 100pl 'H,O for locking the magnetic field. Proton decoupling was achieved with the waltz-16 pulse sequence (Shaka et at. 1983)at - 16 dB from 20 W. Nuclear Overhauser effects on the signals were eliminated using the inverse gated decoupling technique. Spectra were obtained with 30" pulses (6 p),a 20 kHz spectral width, repetition time of 1 s, quadrature phase detection and quadrature phase cycling (CYCLOPS). Between loo00 and 4OOOO scans were recorded. The free induction decays were multiplied by an exponential decay resulting in an increase of the linewidth by 5 Hz. 4-Fluoroaniline (3 mM in lOOmM potassium phosphate, pH 76,7"C) served as a reference for the 19F measurements. Chemical shifts are reported relative to CFCl, using a conversion of - 1300ppm (64fluoroaniline-6CFC1, = - 130.0ppm) (Wray 1983). Intensities of the various metabolites observed in the "F-n.m.r. spectra were determined from the integrals of their ''F-n.m.r. resonances. Parafluorobenzoate was used as an internal standard for the determination of absolute concentrations. During the acquisition of the 19Fsignal the extremely strong 'H decouple signal is present in the probe. Although the leakage of the 'H decouple signal in the I9Foutput of the probe is small, the output of the 'H transmitter was passed through a Bruker 300 MHz bandpass and a 282 MHz bandstop filter to be sure that a clean 300 MHz signal reaches the probe. A Bruker 300 MHz bandstop and a 282 MHz bandpass filter is placed at the input of the I9F preamplifier to prevent the 300MHz signal from overloading the preamplifier. These two filters were carefully tuned to maximum sensitivity of the 19F signal. Further attenuation of the interfering signal (of the 'H decoupler) was obtained by placing a home-built 90MHz bandstop filter in the 72MHz intermediate frequency amplifier of the ''F preamplifier. Furthermore, the gain of the preamplifierwas set to the low position and the gain of the low-frequency amplifier of the main receiver was set to the high position to prevent the phase-sensitive detectors of the main receiver from overloading.
J. Vervoort et al.
660
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Results Identification of the lgF-n.rn.r. resonances The 'gF-n.m.r. resonances of the fluoride ion (F-), 2-fluoroaniline, 2-fluoroacetanilide, 4-amino-3-fluorophenol and 2-amino-3-fluorophenol were identified on the basis of their "F-n.m.r. spectra acquired under our standard experimental conditions (0.1 M potassium phosphate pH 7-6, 7°C) (table 1). The "F-n.m.r. resonance of 4-acetamido-3-fluorophenolwas identified based on results obtained from a microsomal incubation with 2-fluoroacetanilide (figure 1). From the "Fn.m.r. spectrum presented in figure 1 it is seen that the microsomal metabolism of 2fluoroacetanilide results in deacetylation, yielding 2-fluoroaniline, in the formation of 4-amino-3-fluorophenol and in one additional metabolite at - 125.7 ppm, representing 4-acetamido-3-fluorophenol.This is concluded from the fact that chemical acetylation of 4-amino-3-fluorophenol resulted in one main product, having its lgF-n.m.r. resonance at - 125.7ppm (data not shown). Finally, the resonances of 4-acetamido-3-fluorophenylsulphate, 4-amino-3fluorophenyl sulphate, 4-acetamido-3-fluorophenyl glucuronide and 4-amino-3Ruorophenyl glucuronide were identified from the lgF-n.m.r. spectra of urine samples in which these glucuronidated or sulphated urine metabolites were PY
C-0
&' I
OH
I I I I
Y YO
&'
i I
I
I
I
I
f
I
F
c
F-
OH
1
I
!
I
-115
I
-120
I
I
-125
-130
I
I
-135
-140
PPM (6) "F-n.m.r. spectrum of a microsomal suspension incubated with 2-fluoroacetanilide.
Figure 1. Microsomes were from 3-methylcholanthrene-treated rats. The resonances marked with an asterisk were present in control experiments without NADPH.
66 1
Metabolism of 2-jhoroaniline by "F-n.m.r.
Table 1. Chemical shifts of "F-n.m.r. resonances of identified 2-fluoroaniline metabolites in 0.1 M potassium phosphate pH 76; 7°C. ~~
~
Chemical shift (pprn)'
Compound
- 1384 - 128.5 - 123.0 - 1349 - 1347
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
2-Fluoroaniline 2-Fluoroacetanilide Fluoride ion 4-Amino-3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Acetamido-3-fluorophenol 4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate
-135.1 - 1257 -1250 125.1
-
'The chemical shifts are relative to CFCl,.
Table 2.
''F-n.m.r.
characteristics of urine samples of a 2-fluoroaniline-exposed rat. Urine sample (concentration in mM)
Compound Fluoride ion (F-) 4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate 4-Acetamido-'3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Amino-3-fluorophenol Unknown (- 1354ppm) Unknown (- 1382 ppm) 2-Fluoroanaline
Untreated 0.36 084 1.25 0.15
0.44 3.74 0.36 0.53 067
fl-Glucuronidase treated
Arylsulphatasel fl-glucuronidase treated
0.31 1.14 1.00 3.71 0.41 0.40 0.40 0.80
0.32 210 401 0.24 0-06 0.93
Urine samples were either untreated, treated with B-glucuronidase or with arylsulphataselflglucuronidase as described in the Experimental section. para-Fluorobenzoate was added as internal standard for determination of concentrations. -=Not observed.
hydrolysed to their corresponding phenols using fl-glucuronidase or arylsulphatase, respectively. fl-Glucuronidase treatment of urine resulted in the loss of the resonance at - 1347ppm with a concomitant proportional increase of the resonance at - 134.9 ppm representing 4-amino-3-fluorophenol (table 2). In the same experiment the resonance at - 125.0 ppm disappeared with a new resonance of similar intensity appearing at - 125-7ppm representing 4-acetamido-3-fluorophenol.Incubation of an urine sample containing fluorinated metabolites with an arylsulphatase/fl-glucuronidase preparation resulted not only in the loss of the signals of the glucuronidated compounds but also in the disappearance of the resonances at - 135.1 ppm and at - 125.0ppm, giving rise to proportional increases of the resonances at respectively - 1349ppm (4-amino-3-fluorophenol)and at - 125.7ppm (4-acetamido-3-fluorophenol)(table 2). Table 1 summarizes the "F-n.m.r. chemical shifts of various 2-fluoroaniline derived metabolites.
J . Vervoort et al.
662
2-FA
&' Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
OH
-115
-120
-125
-130
-135
-140
PPM (6) Figure 2. ''F-n.m.r. spectrum of a microsomal suspension incubated with 2-fluoroaniline. The resonances marked with an asterisk were present in control experiments without NADPH.
Microsomal conversion of 2-fluoroaniline The "F-n.m.r. spectrum presented in figure 2 clearly indicates 4-amino-3fluorophenol as the main metabolite formed during the incubation of 2-fluoroaniline with rat liver microsomes. In addition to the major para-hydroxylated 2-fluoroaniline metabolite, fluoride (F-) was a minor metabolite in these incubations. No 2-amin0-3-fluorophenol(- 129.9 ppm) that would have been formed upon orthohydroxylation was observed. It is therefore concluded that microsomal metabolism of 2-fluoroaniline results in formation of para-hydroxylated 2-fluoroaniline (4amino-3-fluorophenol) as the main metabolite. Biotransformation of 2-fluoroaniline by isolated rat hepatocytes Results obtained upon incubation of rat hepatocytes with 2-fluoroaniline are presented in figure 3. Metabolism by hepatocytes in Krebs Ringer buffer produces a substantial number of reaction products (figure 3 A). N-Acetylation is the metabolic pathway occurring at the highest rate, although para-hydroxylation of 2-fluoroaniline also occurred to a significant extent. Small amounts of phase I1 reaction products of 4-amino-3-fluorophenol were also observed. Using rat hepatocytes in Krebs Ringer buffer, glucuronidation of 4-amino-3-fluorophenol resulting in 4amino-3-fluorophenyl glucuronide was the major phase I1 reaction (figure 3). In these incubations 4-amino-3-fluorophenyl sulphate and 4-acetamido-3fluorophenyl sulphate were formed at a relatively low level. Furthermore, only a limited percentage of all 4-amino-3-fluorophenol appeared to be conjugated in these incubations. In a sulphate-rich medium, however (figure 3 B), all phenolic metabolites appeared to be conjugated with sulphate as no 4-amino-3-fluoropheno1 or 4amino-3-fluorophenyl glucuronide could be observed. Conversion of 2-fluoroaniline by a perfused rat liver system Figure 4 shows the results of the metabolism of 2-fluoroaniline in perfused rat liver. The similarity of the spectra presented in figure 4 and in figure 3 (conversion by
663
Metabolism of 2-fluoroaniline by IgF-n.m.r.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
A
n B
-115
-120
-125
-130
-135
-140
PPM ( 6 ) ''F-n.m.r. spectra obtained upon incubation of hepatocyteswith 2-fluoroaniline (A) in Krebs Ringer buffer and (B) in William's E medium. Both contained 3 mM 2-fluoroaniline. Spectra are from the samples taken after Wmin incubation. The resonances marked with an asterisk were present in the incubations at t =0 min.
Figure 3.
664
. I . Vmwoort et al.
&' I
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
A
I
I
i B
I -115
I
-120
I
-125
I -130
I -135
I -140
PPM ( 6 ) Figure 4. lgF-n.m.r. spectrum of the metabolism of 2-fluoroaniline in a perfused rat liver system. The perfusion buffer (100 ml) was recirculated and contained 3 mM 2-fluoroaniline in a modified Krebs Ringer solution without sulphate (A) or in an O,/CO, buffered perfusion buffer containing 6 . 0 m sulphate ~ (B). Spectra were taken after 60min incubation. The resonances marked with an asterisk were already observed at t=Omin.
Metabolism of 2-fluoroaniline by lgF-n.m.r.
665
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
rat hepatocytes) is evident. The main difference is that under sulphate-limited conditions liver perfusion, in contrast to hepatocyte incubations, did not give rise to formation of glucuronidated species.
In vivo metabolism of 2-fluoroaniline by Wistar rats Figure 5 shows the lgF-n.m.r. spectrum of a urine sample of a rat exposed to 2-fluoroaniline. After addition of 2-fluoroaniline the parent compound was rapidly metabolized and over 90% of the administered dose was detected in 24 h urine. The lgF-n.m.r. spectrum (figure 5 ) shows formation of about nine major urinary metabolites. Seven of these metabolic products have been unambiguously assigned (table 1). The relative amount of the various fluoro-containing metabolites is presented in table 3.4-Amino-3-fluorophenylsulphate appears to be the major urine metabolite. In addition to these major metabolites at least 16 minor products (each > 0 5 % of the total "F intensity) can be observed. No attempt was made to assign these resonances. Metabolites carrying an amino group can be easily discriminated from N acetylated products as the latter do not shift in the lgF-n.m.r. spectrum on acidifying the urine. The apparent overlap of the resonances of 4-acetamido-3-fluorophenyl-sulphate and -glucuronide is dependent on conditions (pH, temm). The difference between
I
-115
I
-120
I
-125
I -130
I -135
I
-140
PPM ( 6 ) Figure 5. lgF-n.m.r.spectrum of the 24 h urine of a male Wistar rat orally exposed to 2-fluoroaniline. 2-Fluoroaniline (50mg/kg) was administered dissolved in olive oil.
J. Vervoort et al.
666 Table 3.
Comparison of rat urine metabolite patterns of 2-fluoroaniline as determined by I9F-n.m.r. and by using radioactive t.1.c. analysis as described by Eadsforth et ul. (1986). Present study: percentage of total "F-intensity mean & SEM (n=4 rats)
Identified compound
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
F-
-120
?
1.6k0.5 9.9 f 2.3 18.1 f1.3 05f01 3.3 It05 53.5 f27 0.0 3.8 f 0 2 2.6 k1.2 4.9 & 1.4
4-Acetamido-3-fluorophenylglucuronide 4-Acetamido-3-fluorophenylsulphate 4-Acetamido-3-fluorophenol 4-Amino-3-fluorophenyl glucuronide 4-Amino-3-fluorophenyl sulphate 4-Amino-3-fluorophenol Unknown ( - 135.4ppm) Unknown (- 138.2 ppm) 2-Fluoroaniline
-115
Eadsforth et ul. (1986): percentage of dose recovered in urine
-125
-130
48 44.1 08 1.o 27.8 00 63
1.8 0.0
-135
-140
PPM ( 6 ) Figure 6. "F-n.m.r. spectrum of the urine of a 2-fluoroaniline exposed rat showing the detection limits for a sample measured in an overnight run on a Bruker CXP-300 spectrometer using a dedicated 19F-n.m.r. lOmm probehead. ) added as an internal reference for quantitative Para-fluorobenzoate (final concentration 10p ~ was measurements. The peaks labelled with an asterisk are present in a concentration of 1 PM. This spectrum is identical to the one shown in figure 5, with the vertical scale magnified about 150times.
Metabolism of 2-fluoromiline by IgF-n.m.r.
667
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
the chemical shift of the two resonances varies between 20 and 40Hz (0070.13ppm). Because the intrinsic linewidth of both resonances is 5 Hz, the two peaks are well enough resolved to determine their integrals accurately. In addition, this quantification of 4-acetamido-3-fluorophenyl-glucuronide and -sulphate was checked by the addition of 20% (v/v) acetonitrile to the urine sample. Under these conditions the two resonances were separated by 0-3 ppm, making quantification easier. Identical data were obtained. Sensitivity of the lgF-n.m.r. method The results presented in figure 6 clearly demonstrate that the detection limit of the lgF-n.m.r. method for fluoro-containingcompounds is 1 /AMfor an Overnight run using a dedicated lOmm "F probehead and a Bruker CXP 300 spectrometer. The main metabolites observed in the urine spectra are present in concentrations ranging between 0.1 and 4 mM, whereas the concentration of the minor products is between 1 and 100PM. It was checked that, during an overnight incubation of a urine sample at 7"C, no decomposition and/or bacterial degradation of labile urinary fluorocontaining compounds occurs. Finally, figure 7 presents a calibrationcurve of 2-fluoroacetanilidein urine from a non-exposed animal. It can be seen that urine components do not disturb the measurements. Linearity is observed for a broad concentration range.
[2-fluoroacetanilide]
in mM
Figure 7. Calibration curve for 2-fluoroacetanilide in urine from an untreated rat. 2-Fluoroacetanilide was added as 05% (v/v) of a 200 times concentrated stock in dimethyl sulphoxide. Relative "F-n.m.r. intensities were quantified from spectra obtained in 30 min.
668
f . Vwvoort et a).
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Discussion After administration of 2-fluoroaniline to rats, the compound is rapidly metabolized and over 90% of the dose is excreted in the urine within 24 h. In the 19Fn.m.r. spectra of these urine samples nine major products were observed of which seven have been assigned (figure 5 , table 3). From these assignments it can be seen that the major biotransformation route is via para-hydroxylation. No orthohydroxylated metabolites, which are known to be formed to a substantial amount upon biotransformation of aniline derivatives (Eyer et al. 1980, Boyland et al. 1953), were observed as major excretion products. This observation of para-hydroxylation being by far the dominating pathway is in agreement with results obtained with microsomal incubations (figure 2), hepatocyte incubations (figure 3) and liver perfusions (figure 4). Microsomal incubation studies give similar results to those presented recently for monofluorinated anilines which showed that parahydroxylation is by far the dominant reaction (Rietjens and Vervoort 1989). In addition, the present paper has shown that upon in vivo exposure to 2fluoroaniline, para-hydroxylation is mainly followed by sulphation and to a minor extent by glucuronidation. Sulphation being the dominating phase I1 reaction in rats is in accordance with other published data on metabolism of aniline derivatives in rats (Baldwin and Hutson 1980, Eadsforth et al. 1984, 1986). In the present study these sulphated species were also clearly observed in hepatocyte incubations and in perfused liver (figures 3 and 4). No glucuronidated metabolites were observed in perfused liver, not even when using perfusion buffers with low sulphate concentration. The absence of glucuronidated species in perfused liver may be related to a higher intracellular concentration of sulphate (in the cells of the perfused liver) than occurs in isolated hepatocytes. In addition, Evelo et al. (1984) has shown, when studying phase I I dependent reactions of 4-aminophenol in isolated hepatocytes, that the K, for sulphation of the hydroxy function was lower than for glucuronidation. This implies that, even at low sulphate concentration, sulphation of 4aminophenol is favoured over glucuronidation. Excretion of glucuronidated metabolites into bile during liver perfusion cannot explain the fact that these metabolites are not observed in the liver perfusion. This follows from the fact that perfusion was carried out in such a way that excreted bile would end up in the perfusion medium. Furthermore, the observation that more than 90% of the dose of 2-fluoroaniline was detected in the urine excludes the faeces (and thus the bile) as a major route of excretion for glucuronidated 2-fluoroaniline derivatives. In lgF-n.m.r. measurements of faeces no fluorinated metabolites were observed. In incubation experiments with hepatocytes or in perfused liver systems (figures 3 and 4) N-acetylation is the major phase 11type reaction. However, inurine samples only 29% of the assigned metabolites were N-acetylated (table 3). So the ratio between N-acetylation and para-hydroxylation of 2-fluoroaniline observed in vitro in favour of N-acetylation, is not reflected in a high percentage of N-acetylated urine metabolites. Deacetylation of N-acetylated products of 2-fluoroaniline in the liver and/or the kidney could explain this discrepancy. In the present study deacetylation was indeed observed in microsomal incubations with 2-fluoroacetanilide (figure 1). Furthermore, during the 60 min of hepatocyte incubations or 90 min of liver perfusion, hydroxylation of 2-fluoroacetanilide was not observed to any significant extent because, compared to 2-fluoroaniline, the concentration of 2-fluoroacetanilide is low. However, during the in vivo experiment, metabolism continues for longer than
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Metabolism of 2-fluoroaniline by 'gF-n.m.r.
669
these time periods. Taking into account the high rate of N-acetylation, observed in both in vitro systems during the first hours of metabolism, the in vivo concentration of 2-fluoroacetanilide available for hydroxylation can be expected to increase. As a result the relative amount of para-hydroxylated acetylatedmetaboliteswill be higher upon prolonged in vivo metabolism than in in vitro metabolism. Comparison of our results with those of Eadsforth et al. (1986) with 14C-2fluoroanilineshows that the same type of metabolites are formed (table 3). The major difference between these two studies is in the amounts of N-acetylated products observed (table 4). 4-Acetamido-3-fluorophenylsulphate is reported to be the main metabolite present in the urine in the study of Eadsforth et al., whereas we observed 4-amino-3-fluorophenyl sulphate to be the main metabolite. On performing our experiment under the conditions used by Eadsforth et al. (1986) (i.e. oral administrationof 23-5mg 2-fluoroaniline/kgbody-weight, dissolved in olive oil) we observed the same metabolic pattern as shown in table 3, so that the difference between the two studies cannot be a dose-related effect. Another possible explanation might be that the two major metabolites, 4-amino-3-fluorophenyl sulphate and 4-acetamido-3-fluorophenylsulphate, have been confused in the study of Eadsforth et al. (1986) as the authors assumed that the two peaks, although not well resolved on t.l.c., eluted in the same order from the h.p.1.c. column. Alternatively, the rats used in our study and those of Eadsforth et al. (1986) are genotypically different. It is well known that within species fast and slow acetylators occur (Weber and Levy 1988). The differencesobserved, with regard to the amount of N-acetylated products present in the urine spectra, may therefore be explained by either a slow acetylator genotype (our study), or a fast acetylator genotype (Eadsforth et al. 1986) of the rats used. Finally it is concluded that the IgF-n.m.r. method described in the present paper has important advantagesfor both biomonitoringand biotransformationstudies, not only of 2-fluoroaniline but also of other fluoro-containing compounds. Firstly, "Fn.m.r. spectra, and thus metabolic patterns, can be obtained directly from urine samples without any treatment. Secondly, the method is shown to be a sensitive tool as the lower limit of detection is 1 PM, which can be decreased to about 50 nM (or even less) if concentration procedures such as freeze-drying of urine, are incorporated. Alternatively concentration procedures can be used to limit the "Fn.m.r. acquisition time, needed to obtain 1PM detection limit in the original urine. It can be calculated that upon concentration of urine samples by a factor 10, 1PM sensitivity in the original urine (i.e. 1 O p in ~ the concentratedsample) is obtained in a 15-min run. Finally, all fluoro-containing metabolites can be observed in one single Table 4. Relative routes of in viw metabolism of 2-fluoroaniline by male Wistar rats observed using I9F-n.rn.r. and reported by Eadsforth ct al. (1986) using a radioactive t.1.c. method. Percentage of total I9F metabolities (n=4 rats)
Reaction ~
Eadsforth et al. (1986)
~~
para-Hydroxylation N-Acetylation Sulphation Glucuronidation
100 33 84
15
100
68 96 8
Calculations are based on the relative amounts of identified metabolites presented in table 3, taking para-hydroxylation as 100% reference.
670
Metabolism of 2-jtuoroaniline by 'F-n.m.r.
spectrum, making the method independent of the relative amount of products present and of genetic polymorphisms, in contrast to alternative methods where urine samples are screened for the presence of one metabolite only. N-acetylation, for example, is known to show genetic polymorphism. The "F-n.m.r. method, detecting all 2-fluoroaniline metabolites, and not only 4-amino-3-fluorophenylsulphate, has the advantage of reducing the risk of false negatives for fast acetylators.
Xenobiotica Downloaded from informahealthcare.com by Nyu Medical Center on 05/18/15 For personal use only.
Acknowledgements Thanks are extended to Ms L. Ausma and Ms S. Fulton for typing the manuscript, to Mr M. M. Bouwmans for preparing the figures and to Mr J. W. M. Haas for help with animal handling.
References AYMES,J. D., and PARIS,M. R., 1980, Etude des complexes du cobalt (11) transporteurs d'oxygene: nouvelle synthese du fluoro-3-hydroxy-2-benurldehyde.Bulletin de la Societe Chimique de France, 3 4 , 11-175-11-178. BALDWIN, M. K., and HUTSON,D. H., 1980, The metabolism of 3-chloro-4-fluoro-aniline in dog and rat. Xenobiotica, 10, 135-144. BOYLAND, E., MANSON, D., and SIMS,P., 1953, The preparation of 0-aminophenyl sulphates. Journal of the Chemical Society, 3623-3628. BRODIE, B. B., and AXELROD, J., 1948, The estimation of acetanilide and its metabolic products, aniline, N-acetylp-aminophenol andp-aminophenol (free and conjugated) in biological fluids and tissues. Journal of Pharmacological and Experimental Therapeutics, 94,22-28. EADSFORTH, C. V., LOGAN, C. J., MORRISON, B. J., and WARBURTON, P. A., 1984,2-4-Diffuoroanilineand 4-fluoroaniline exposure: monitoring by methaemoglobin and urine analysis. International Archives of Occupational and Environmental Health, 54, 223-232. EADSFORTH, C. V., COVENEY, P. C., HUTSON,D. H.,LOGAN,C. J., and SAMUEL, A. J., 1986, The metabolism of o-fluoroaniline by rats, rabbits and marmosets. Xenobiotica, 16, 555-566. EVELO, C. T. A., VERSTEEGH, J. F. M., and BLAAUBOER, B. J., 1984, Kinetics of the formation and secretion of the aniline metabolite 4-aminophenol and its conjugates by isolated rat hepatocytes. Xenobiotica, 14, 409-416. EYER,P., KAMPFFMEYER, H., MAISTER,H., and ROSCH-OEHME, E., 1980, Biotransformation of nitrosobenzene, phenylhydroxylamine, and aniline in the isolated perfused rat liver. Xenobiotica, 10,499-516. GROSSMAN, S . J., and JOLLOW, D. J., 1986, Use of the NIH shift to determine the relative contribution of competing pathways of aniline metabolism in the rat. Drug Metabolism and Disposition, 14, 689691. PAINE, A. J., WILLIAMS, L. J., and LEGC,R. F., 1979, In The Liver: Quadtotiwe Aspects of Structure and Function, edited by R. Preisig and J. Bircher (FRG, Aulendorf), 215-263. RIETJENS, I. M. C. M., and VERVOORT, J., 1989, Microsomal metabolism of fluoroanilines. Xenobiotica, 19, 1297-1305. RUTTEN, A. A. J. J. L., FALKE,H. E., CATSBURG, J. F., TOPP,R., BLAAUBOER, B. J., HOLSTEIJN, L. VAN DOORN,L., and LEEUWEN, F. X. R. VAN,1987, Interlaboratory comparison of total cytochrome P-450and protein determination in rat liver microsomes. Archives of Toxicology, 61, 27-33. SHAKA, A. J., KEELER, J., and FREEMAN, R., 1983, Evaluation of a new broad band decoupling sequence: Waltz 16. Journal of Magnetic Resonance, 53,313-340. VOGEL, A. J., 1978, In Textbook of Practical Inorganic Chemistry, (Harlow: Longman), p. 684. WEBER,W. W., and LEVY,G. N., 1988, In Metabolism of Xenobiotics, edited by Gorrod, J. W., Oelschlager, H., and Caldwell, J. (London: Taylor & Francis), pp. 209-215. WRAY,V., 1983, Annual Reports on N M R Spectroscopy, edited by Webb, G. A. (London: Academic Press), Vol. 14, p. 252.
