Urinary metabolites of [1,2,3-13C]acrylonitrile in rats and mice

Susan C. J. Sumner, Timothy R. Fennell, Timothy A. Moore, Brian Chanas, ... Sara K. Nauhaus, Timothy R. Fennell, Bahman Asgharian, James A. Bond, and ...
0 downloads 0 Views 1MB Size
Download PDF
Chem. Res. Toxicol. 1991, 4 , 678-687

678

Urinary Metabolites of [1,2,3-13C]Acrylonitrile in Rats and Mice Detected by 13C Nuclear Magnetic Resonance Spectroscopy Timothy

R. Fennell, Gregory L. Kedderis, and Susan C. J. Sumner*

Chemical Industry Institute of Toxicology, P.O. Box 12137, Research Triangle Park, North Carolina 27709 Received June 6, 1991 Acrylonitrile, a carcinogen in rats, undergoes extensive metabolism via two routes: direct glutathione conjugation or epoxidation. Metabolism to cyanoethylene oxide may mediate the carcinogenic and toxic activity of acrylonitrile. T o characterize comprehensively the metabolism in vivo of acrylonitrile, the detection and identification of metabolites in urine of rodents dosed with acrylonitrile have been carried out using NMR spectroscopy. Following administration of [1,2,3-13C]acrylonitrile to male Fischer 344 rats (10 or 30 mg/kg, PO) or B6C3F1 mice (10 mg/kg, PO), urine samples were collected for 24 h. Carbon-13 NMR spectra were acquired directly on the urine samples after centrifugation and addition of 1 0 4 5 % DzO. Resonances were assigned to carbons of acrylonitrile metabolites on the basis of chemical shift, proton multiplicity, carbon-carbon coupling, and calculated values of shift, and by comparison with standards. The proton multiplicity of each carbon was determined by heteronuclear 2D J-resolved spectroscopy (HETBDJ),and the carbon-carbon connectivities of resonances were determined using incredible natural abundance double quantum transfer spectroscopy (INADEQUATE). The metabolites identified in rat urine were thiocyanate, N-acetyl-S-(2-~yanoethyl)cysteine, N-acetyl-S-(2hydroxye thyl) cysteine, N-acetyl-S-( 1-cyano-2-hydroxyethyl) cysteine, thiodiglycolic acid, thionyldiacetic acid, and S-(carboxymethy1)cysteine or its N-acetyl derivative. These metabolites were also identified in mouse urine. Metabolites were quantitated by integrating metabolite carbon resonances with respect to that of dioxane added at a known concentration. Thiodiglycolic acid and (carboxymethy1)cysteine (or its N-acetyl derivative) were the major metabolites in the mouse, while N-acetyl-S-(2-~yanoethyl)cysteine and N-acetyl-S-(2-hydroxyethyl)cysteinewere the major metabolites in the rat. Metabolites derived from cyanoethylene oxide (CEO) accounted for approximately 60% of the products excreted in rat urine, compared with 80% in the urine from mice. Differences between rat and mouse in the further metabolism of CEO were also observed. The proportion of the dose metabolized via CEO may be an important determinant of the toxicity and carcinogenicity of acrylonitrile.

Introduction The elucidation of the routes of metabolism of small organic molecules can present considerable methodological challenges because the polar nature of the metabolites complicates their separation, identification, and quantitation. The purpose of this study was to analyze the metabolism of acrylonitrile (AN)' in rats and mice. In order to accomplish this, a strategy was developed for characterizing the metabolites directly in urine by NMR spectroscopy, following administration of 13C-labeledAN to rats and mice. AN is widely used in the manufacture of plastics and fibers and has been found to be carcinogenic on administration orally or by inhalation to rats ( 1 ). No bioassays have been carried out in other species. The mouse is more susceptible than the rat to the acute toxicity of AN ( 2 ) . The majority of studies on the metabolism of AN have been conducted in the rat. The metabolism of AN occurs by two major pathways: direct conjugation with glutathione to yield eventually N-acetyl-S-(2-~yanoethyl)cysteine ( 3 , 4 )and oxidation by cytochrome P-450 to give cyanoethylene oxide (CEO), which undergoes a variety of further reactions. CEO has been reported to undergo hydrolysis to form cyanide and glycolaldehyde (5, 61, and to undergo conjugation with glutathione ( 4 , 7,8). Comprehensive studies have not yet Abbreviations: AN, acrylonitrile; CEO, cyanoethylene oxide; Cys-S, cysteinyl residue;GS, glutathionylresidue; N-AcCys-S, N-acetylcysteine residue.

been carried out on the relative contributions of the various pathways. From observations on the toxicity of AN, it has been suggested that there are significant differences in metabolism to cyanide (and the epoxide metabolite from which the cyanide is derived) with different routes of exposure and between rats and mice (9). The formation of reactive metabolites, and their reaction with DNA to form adducts and eventually cause mutations, is thought to be the major way in which chemical carcinogens exert their effects (10). CEO is mutagenic ( 1 1 ) and thus is thought to be involved in the carcinogenic activity of AN. NMR spectroscopy has been used in the study of metabolic processes, to follow the conversion of compounds with natural abundance isotopes such as 'H, 19F,and 31P, or isotopically labeled compounds containing *Hor 13C (12). lH NMR has been used to study the metabolism of foreign compounds, as well as changes in the metabolism of endogenous compounds (13,14). However, the observation of metabolites of foreign compounds in urine by 'H NMR is complicated by the signals arising from endogenous compounds, dispersion of signals over a small frequency range, and a dynamic range problem arising from the large water signal. Investigating other nuclei can avoid these problems, but such studies are often limited by sensitivity and low natural abundance of the detected isotope, as is the case for 13C. The use of I3C-labeled compounds can eliminate the problems associated with the low natural abundance (1.170)of this isotope. A number of studies have used 13C-labeledcompounds for tracking the metabolism of exogenous compounds, e.g., form0 1991 American Chemical Society

Chem. Res. Toxicol., Vol. 4, No.6, 1991 679

Urinary Metabolites of Acrylonitrile

aldehyde (15), 1,2-dibromo-3-chloropropane(16), fluoroacetate (1 7), and S-(carboxymethy1)cysteine (18). The 13C NMR spectra of metabolites in the urine of rats and mice that received [1,2,3J3C]AN by gavage have been examined to determine the extent of metabolism of AN by the two major pathways in rats and mice, to verify the formation of previously described metabolites, to identify new metabolites, and to provide quantitative information on their formation. The structures of the urinary metabolites were established by determining the bonding between carbons and their proton multiplicity, using twodimensional NMR experiments, and by comparison of chemical shift data with calculated values. Quantitative analysis of the metabolites was carried out by comparison of peak areas of carbon signals for metabolites and an internal standard.

resonances for the metabolites and dioxane were used to calculate metabolite concentrations. Dioxane was also used as a spectral reference (66.5 ppm). Two-dimensional heteronuclear &resolved spectroscopy (HET2DJ) (22) was used to determine the proton multiplicity of each carbon resonance with the pulse sequence:

Materials and Methods

where D1 is the relaxation delay (5-15 s) and T a 1/45. Broadband decoupling was employed throughout the experiment. The data were acquired in the phase-sensitive mode with 1024 complex points in t z and 32 complex points in t,, with T values corresponding t o coupling constants of 40 and 60 Hz. Calculated shift values were obtained using tables of incremental shift values for alkanes (24,25), and using the 13C NMR Database (STN International, Columbus, OH). Determination of Glucose Concentrations. Urinary glucose concentrations were determined using both the NMR method described above for metabolite quantitation, and by a method involving hexokinase and glucose-6-phosphate dehydrogenase coupled reduction of NAD' (Roche Diagnostics, kit 44557, Montclair, NJ) in a Roche Cobas Fara I1 analyzer.

Chemicals. [ 1,2,3J3C]AN was purchased from Cambridge Isotopes Ltd. (Cambridge, MA). A sample of [1,2,3-13C]ANin CDC13 gave the expected proton-decoupled 13C NMR spectrum = 77.8 Hz), doublet consisting of a doublet (116.90 ppm, C-1, JCC of doublets (107.60 ppm, C-2, JC.c = 71.2,77.8 Hz),and doublet (137.19 ppm, (2-3, JC.c= 71.3 Hz) (not shown). DzO, DMSO-d,, CDC13,cyanoacetic acid, pyruvonitrile, thiodiglycolic acid, cyanoacetaldehyde, cyanoethanol (3-hydroxypropionitrile),ethyl carbonate, glycolaldehyde dimer, and tetramethylsilane were purchased from the Aldrich Chemical Co. (Milwaukee, WI). and glyCysteine, N-acetylcysteine, S-(carboxymethyl)cysteine, colic acid were obtained from Sigma Chemical Co. (St.Louis, MO). N-Acetyl-S-(2-~yanoethyl)cysteine was synthesized by reaction of N-acetylcysteine with AN (19). S-(2-Hydroxyethyl)cysteine and N-acetyl-S-(2-hydroxyethyl)cysteine were synthesized by reaction of ethyl carbonate with cysteine or N-acetylcysteine (20). Thionyldiacetic acid was prepared by the oxidation of thiodiglycolic acid by hydrogen peroxide (21). N-Acetyl-S-(1-cyano2-hydroxyethy1)cysteine methyl ester was prepared by Dr. V. Amarnath, Department of Pathology, Duke University, Durham, NC, by reaction of 2-bromo-3-hydroxypropionitrilewith Nacetylcysteine methyl ester, using the method of Linhart et al. (7).

Animals. Male Fischer 344 rats (220-270 g) and male B6C3Fl mice (20-35 g) were purchased from Charles River Laboratories (Raleigh, NC). They were supplied food (NIH 07 diet) and deionized water ad libitum and maintained in a 12-h light-dark cycle, a t a temperature of 22 h 2 "C and relative humidity of 55

* 5%.

Three rats and 11 mice were administered [1,2,3-13C]AN(10 mg/kg) dissolved in distilled water (1.0 or 10 mL/kg body weight for rata and mice, respectively) by gavage. Four rats received an oral dose of 30 mg/kg [1,2,3-13C]ANdissolved in distilled water (1.0mL/kg body weight). They were placed in all-glass metabolism cages, and urine was collected for 24 h. Control urine samples were collected from 3 rats and 4 mice during the 24-h period prior to dosing. The urine was centrifuged a t 2000g for 20 min and either analyzed immediately or stored at -20 "C until analyzed. NMR Spectroscopy. NMR spectra were acquired with a dual proton-multinuclear probe on a Varian VXR 300 spectrometer (Varian Instruments, Palo Alto, CA). 13C NMR spectra were obtained with a spectral width of 16 500 Hz and 30 000-50 000 data points. A 45O or 90" pulse width was used, with a relaxation delay of 5 s and Waltz-16 decoupling of protons. DzO (10-25% final concentration) was added to an aliquot of each urine sample which was then placed in a 5-mm i.d. NMR tube. Standardization was carried out using a coaxial capillary (Wilmad, Buena, NJ) containing TMS in CDC13. The most intense peak in the urine spectra, due to urea, was referenced a t 162.5 ppm. Quantitation of metabolites was carried out by adding dioxane as standard to the urine sample. The relaxation times (T,) of the metabolite resonances and dioxane were determined. Spectra were then acquired with a delay of 4 times the longest TIof a metabolite to be measured (delay time = 24 s) and proton decoupling only during acquisition. The integrals of the carbon

"C:

Dl-90°-(t1/2)-1800-(t,/2)-A~q(t,)

and gated decoupling was applied to protons. Two-dimensional NMR data were acquired with 2048 complex points in t , and 64 complex points in tl and were Fourier transformed with zero filling in both dimensions. The two-dimensional incredible natural abundance double quantum transfer experiment (INADEQUATE) was used to determine the connectivity between carbon resonances, with the pulse sequence (23):

l3c:

~1-90°-~-1800-T-900-t,-900-Acq(tp)

Results Metabolites in Rat Urine. The 'H-decoupled 13C spectrum of control rat urine (Figure 1A) shows an intense peak at 162.5 ppm which has been assigned to urea. Other peaks present in the spectrum, due to endogenous compounds such as creatinine and hippurate, have also been previously identified (13). Resonances not present in spectra of control samples were detected in urine from rats administered both 10 and 30 mg/kg AN (Figure 1, panels B and C). Resonances from AN-derived carbon atoms in metabolites are easily recognized by their multiplet patterns. Carbon atoms in normal urinary compounds (1.1% 13C) give single resonances. Adjacent enriched carbon atoms give multiplet patterns as a result of carbon-carbon coupling. The chemical shift values (for the center of multiplet patterns) and carbon-carbon coupling constants for the metabolites of AN are listed in Table I. Metabolites which contain an isolated enriched carbon atom give rise to single resonances, such as the singlet at 133.2 ppm which is assigned to thiocyanate, a metabolite of AN. Natural abundance signals at 61, 70-78,92, and 96 ppm due to a- and P-Dglucose were also detected, with intensities that increased with increasing dose of AN (Figure 1B,C). Treatment of rats with AN causes a dose-dependent increase in glucose excretion as a result of nephrotoxicity (26, 27). The direct measurement of coupling constants in the 1-D spectrum may be used to determine the connection between carbon resonances. In mixtures, however, resonance patterns for carbon atoms from different compounds can give rise to similar coupling constants. The INADEQUATE experiment provides an unambiguous determination of carbon connectivity. For several metabolites of AN, connectivities are established in the INADEQUATE spectrum in Figure 2. The 7 value used in the acquisition of the spectrum shown resulted in the largest contours arising from carbons with coupling constants of approximately 40 Hz. Additional spectra were acquired with

Fennel1 et al.

680 Chem. Res. Toxicol., Vol. 4, No. 6, 1991

urea

A

rc 3b

JL

3a 2b

C

lb

Y I , ) I I , I I I 1 1 / 1 1 I I I / I l /I I I I j I I 1 1 1 1 I I I j I I I I I I I I / / I I I I 1 1 1 1 1 1 I I I I

I80

160

140

120

100

80

I

I I I / / I I I I I I l l / I I I 1 , I I I 1 1 1 I,I,lT

60

40

20 PPY

Figure 1. ‘H-decou led 13C NMR spectra of control rat urine (A) and urine collected for 24 h following administration of 10 (B)or 30 (C)mg/kg (1,2,3-PSC]AN.Signals from metabolites are labeled according to metabolite number (see Table I and Scheme I) and the letter of the carbon derived from AN (,CH,=&H--,CN).

values of T sufficient to establish connectivity between carbons with other coupling constants. The presence of two contours with the same double quantum frequency (i.e., aligned horizontally along the F1 axis) but with different chemical shifts (E2axis) indicates coupling between the carbons giving rise to these signals. For example, the two contours at the chemical shift of 2b are aligned horizontally with contour peaks at the shift positions of 2a and 2c, while 2a and 2c are each only aligned with 2b. This indicates that metabolite 2 has a C,-Cb-C, structure for the portion of the metabolite derived from AN. The single contours for resonances 3a and 3b indicate that this metabolite contains only two adjacent enriched carbon atoms. The connectivities of the other carbon resonances were determined in a similar manner and are listed in Table I. The heteronuclear 2-D J-resolved spectrum permits identification of the number of hydrogen atoms attached to each carbon atom. An expanded region of the HETBDJ spectrum of a sample of rat urine is shown in Figure 3. The 13C spectrum is shown on the F2 axis, and carbon-

proton coupling is shown by multiple contour peaks in the F1 dimension. The number of hydrogens attached to a carbon at a chemical shift position equals the number of contour peaks minus one. The two signals for carbon 3a each have three contours located a t their shift positions along the F1 axis and are assigned to methylene carbons. The proton multiplicities of the metabolite carbon signals are listed in Table I. The signals from 4b and 4’b are from methine (CH) carbons, and the resonances for 5a are from methylene (CH,) carbons. The multiplets are displaced from 0 Hz on the F1 axis, with the displacement proportional to the carbon-carbon coupling constant. This enables further distinction of the metabolite signals from those of natural abundance carbons, since this displacement occurs only for adjacent carbons derived from the 13C-enriched AN. Structural Assignments. The assignment of metabolite structures was carried out using the data obtained on proton multiplicity, carbon connectivity, and calculated values of shift. Since the NMR data are obtained for the

Chem. Res. Toxicol., Vol. 4, No. 6, 1991 681

Urinary Metabolites of Acrylonitrile

Table I. *%NMR Chemical Shifts (ppm), C a r b o n C a r b o n Coupling Constants (Hz)and Multiplicities, C a r b o n C a r b o n Correlations, and Proton Multiplicities for Resonances of Metabolites in Rat a n d Mouse Urine Produced after Administration DO of *%-Enriched Acrylonitrile (CH,=CHCN)a carbon coupling constant, Hz, proton and compd chemical connectivity multiplicity calcd Database! no. shift, ppm multiplicityb (INADEQUATE)’ (HET2DJ)d shiftse calcd shifts 1 2 2 3 4 4‘ 5 5’ 1 6 3 4 4‘ 1 4 4’ 2 7 6 5,5’

10.84 18.11 26.89 33.84 35.33 35.64 36.97 37.20 48.54 57.85 60.09 61.20 61.28 nd 118.91 119.06 120.17 133.13 171.18 177.35

58, 35 56, 34 35 38 59, 36 59, 36 53 46 35 51 38 36 36

d,d d,d d d d,d d,d d d d d d d d

Ib 2b 2a 3a 4b 4’b 5a 5’a la 6a 3b 4a 4’a

nda CH2 CH2 CH2 CH CH CH2 CH2 nd CH2 CH2 CH2 CH2

60 60 57

d d d

4c 4’c 2c

C C C C C C

S

51 54

d d

6b 5b

17 16 28 35 36 36 33 33 45 58 60 63 63 ncJ nc nc nc

16 (18) 19 28 35 36 36 34 35 44 (53) 50-53 (61) 61 67 67 119 (120) 118 118

118-122 nc 180 (166) 172-177

nc nc

Metabolite structures: 1, XCH2CH2CN;2, RSCH2CH2CN3, RSCH2CH20H;4,4‘, RSCH(CH,OH)CN; 5, S(CH2C02H)2;5’, RSCH2C02H; 6, XCH2C02H; 7, SCN- [R = N-acetylcysteine, X = NHR’ or R(SO)]. Molecular structures 1-6 are proposed from the connectivity and chemical shift data. bCarbon multiplicities were obtaied by inspection of the I-D spectrum; d = doublet, d,d = doublet of doublets. ‘Carbon connectivities were obtained through INADEQUATE spectroscopy and interpretation of 1-D carbon coupling constants. The carbon atom is assigned a number corresponding to the metabolite and a letter designating its deviation from acrylonitrile (a, b, or c corresponding to the 3-, 2-, or 1-carbons of acrylonitrile, respectively). Proton multiplicities were obtained through HET2DJ spectroscopy. e Calculated values obtained through addition of incremental effects for alkanes. Metabolites 1 and 6 are calculated only for X = NHR’. !Values for shift obtained through the 13C NMR Database. For metabolites 1 and 6, the first value is calculated for X = Gly (NHR), and the second value is calculated for the sulfoxide of N-aceylcysteine. and, not determined; nc, no calculated.

Fl (P

*

I

59 1 160

,

-

180 4

200 -

3a 200

180

160

140

120

100 BO F2 IPPM)

60

40

20

0

Figure 2. INADEQUATE spectrum of rat urine collected for 24 h following administration of 10 mg/kg [1,2,3-lSC]AN. T h e urine was concentrated as described under Materials and Methods. The spectrum was acquired with 2048 points in tP,32 increments in t l , a relaxation delay of 15 s, and T corresponding to coupling of 40 Hz. Correlations of carbon signals (on the F2 axis) are shown by the contours aligned horizontally along the F1 axis.

portion of each metabolite derived from the carbon-13 labels of AN, the remainder of each metabolite structure was assigned by consideration of substituent effects on chemical shift of the observed carbon signals and comparison with chemical shifts for previously reported metabolites or calculated values of shift for mechanistically feasible metabolites. These assignments were compared with standards.

Using the assignment of metabolite 2 as an example, the INADEQUATE spectrum indicates connectivity between three labeled carbons derived from AN at 18,27, and 120 ppm. The HET2DJ experiment indicates that the peaks at 18 and 27 ppm arise from methylene groups. Since the chemical shift value for 2c is indicative of a cyano carbon, the resulting structure is XCH2CH2CN.Using incremental shift values for substituent groups on ethanes, the nature

Fennel1 et al.

682 Chem. Res. Toxicol., Vol. 4 , No. 6, 1991

35.035.5-

4,db 36.036.5-

37.5 38.0 0 Fl

(HZ)

Figure 3. A section of the HET2DJ spectrum of rat urine obtained after administration of [1,2,3-13C]AN (30mg/kg). The urine was concentrated as described under Materials and Methods. The spectrum was acquired with 2048 points in t 2 and 64 increments in tl and with gated decoupling. The number of hydrogens attached to metabolite carbons is determined from the number of contours positioned along the F1 axis a t the chemical shift of the carbon. Table 11. *FNMR Spectra of Standard Compounds chemical shift, ppm

N-acetyl-S-(2-~yanoethyl)cysteine N-acetyl-S-(2-hydroxyethyl)cysteinea N-acetyl-S-(l-cyano-2-hydroxyethyl)cysteine methyl ester

S-(carboxymethyl)cysteineb thiocyanate thiodiglycolic acidb thionyldiacetic acidb glycolic acidb ethylene glycol acetic acid cyanoacetic acid cyanoethanol glycoaldehyde pyruvonitrile

176.48 (CONH) 33.70 (CH,) 60.30 (CHzOH)d 21.86 (CH3) 173.9 (CONH) 53.07 (OCHB) 21.5 (CHS) 180.7 (C02H) 37.1 (SCH.Jd 133.20 (SCN-) 177.56 (C02H) 171.18 (COZH) 179.64 (COzH) 62.67 (CHzOH) 176.21 (COzH) 169.70 (C02H) 20.83 (CHJ 92.19 [CH(OH)Z] 19.88 (CHJ

173.56 (COzH) 27.03 (SCH.Jd 53.70 (CH,)

120.20 (CN)d 54.36 (CH,) 21.99 (CHJ 18.30 (CHzCN)d 33.95 (SCHz)d 32.19 (CHV)

171.66 (C02H) 52.3 (CH,)

118.62 (CN)d 35.6 (SCH)d

177.9 (C02H)d

55.1 (CH,)

61.26 (CHzOH)d 32.4 (CH,) 37.8 (CH,)

37.28 (SCH2) 57.98 (CHZ) 61.27 (CH,OH) 20.39 (CH3) 117.74 (CN) 26.61 (CHZ) 56.52 (CHzOH) 120.01 (CN) 67.01 (CHZOH) 111.19 (CN)' 175.75 (CO)

OSignals from the carbonyl carbons were not detected because of contaminating acetic acid in the preparation. bAs the sodium salt, in

DzO/HzO with NaOH. 'Triplet, J = 42 Hz,coupling to l'N. dSignals which would be derived from AN carbons in conjugated metabolites.

of the functional group X can be determined. A cyano group has (Y and 0effects of 4 and 3 ppm, respectively. In order to fit the experimental values of 18 and 27 ppm, the functional group X must have CY and 0effects of approximately 18 and 8 ppm, respectively. An RS group is the only feasible substituent which satisfies these conditions, resulting in an assignment of RSCH2CH2CN. Given the extensive conversion of glutathione conjugates to Nacetylcysteine derivatives prior to excretion (28), and the previous description of N-acetyl-S-(2-cyanoethyl)cysteine as a urinary metabolite of AN ( 3 , 4 ) ,it is probable that RS represents an N-acetylcysteinyl substituent. The signals observed for this metabolite have the same chemical shifts as the carbon atoms of the cyanoethyl group in the standard N-acetyl-S-(2-~yanoethyl)cysteine (Table 11). From the INADEQUATE, HETBDJ, and 1-D spectral data, metabolite 3 contains two adjacent 13Catoms, both of which were methylene carbons. An alkylthio group on carbon 3a and a hydroxy group on 3b give calculated values consistent with the observed chemical shifts (Table I).

Similar chemical shifts were obtained for the hydroxyethyl group in authentic N-acetyl-S-(2-hydroxyethyl)cysteine (Table 11), which has been previously described as a metabolite of AN ( 4 ) . The signals for metabolite 4 (designated 4 and 4') at 35, 61, and 119 ppm are two sets of similar complex multiplets with nearly identical coupling constants. This is apparent from the HET2DJ spectrum in which two sets of methine signals for 4,4% are observed with equal JcH and Jcc values (Figure 3, eight pairs of doublets between 35 and 37 ppm). The two sets of doublet signals at 119 ppm for 4,4'c indicate the presence of a cyano group. The resonances at 61 ppm arise from a methylene carbon. The combined data suggest a structure of XCH(CH2Y)CN,and calculated values indicate that metabolites 4 and 4' are diastereomers of HOCH,CH(SR)CN, where R contains an additional asymmetric center. The chemical shift values agree with those of N-acetyl-S-(l-cyano-2-hydroxyethyl)cysteine methyl ester, described previously as a metabolite of AN (7), and with those obtained for an authentic standard

Urinary Metabolites of Acrylonitrile Table 111. Quantitative Determination of [ 1,2,3-'F]AN Metabolites Excreted in Rat Urine' chemical concentra% of total metabolite shift, ppm tion, mM metabolitesb 3.99 f 0.76 42.8 f 4.8 2 26.9 2.33 f 0.13 26.7 i 1.8 3 60.1 1.50 f 0.10 17.4 f 2.2 4,4' 61.3 5,5' 37.2 0.64 f 0.13 7.4 f 0.9 0.50 i 0.21 5.7 i 2.1 6 57.9 ORats were administered a dose of 123.2 rmol of [1,2,3-l3C]AN (30 mg/kg), PO, and metabolites measured in 0-24-h urine using dioxane as standard. The total amount excreted in 24 h was 68.9 f 3.0 rmol (the sum of amounts of the metabolites listed above). Values represent mean f SD (n = 3). bThe amount of each metabolite expressed as a percentage of the total excreted in the urine in 24 h.

(Table 11). The presence of the methyl ester group would not be expected to significantly alter the resonances of the cyanohydroxyethyl carbons. The isomeric compound, N-acetyl-S-(2-cyano-2-hydroxyethyl)cysteine, formed on reaction of CEO with N-acetylcysteine in vitro, also gives signals with chemical shifts similar to those observed for the AN-derived carbons at 119 (CN), 61 (CHOH), and 36 (SCH2) ppm (29). However, the carbon coupling and proton multiplicities of the signals in the rat urine at 61 and 35 ppm are consistent with the l-cyano-2-hydroxyethyl isomer. Resonances for metabolites 5 and 5' are overlapping in the NMR spectra of rat urine, but are clearly distinguished in the NMR spectra of mouse urine (see below). The data for metabolites 5 and 5' are consistent with structures of RSCH2COOH. One metabolite has chemical shifts similar to those obtained for an authentic standard of thiodiglycolic acid (Table 11). The other metabolite is assigned to S-(carboxymethy1)cysteine (or ita N-acetyl derivative). Both of these products have been reported as metabolites of AN (8). Metabolite 6 has a structure of XCH2COOH, where X may represent an amine group or a sulfoxide. Oxidation of metabolites 5 or 5' could result in the formation of such a sulfoxide. A similar metabolite at 58.5 ppm (J = 58 Hz) has been found in the urine of rats on administration of fluoroacetate or S-(carboxymethyl)glutathione, or on peracid oxidation of S-(carboxymethyl)glutathione,suggesting that it may be a sulfoxide derivative of RSCH2COOH(17). Thionyldiacetic acid gave a spectrum with resonances at similar chemical shift to those of metabolite 6, suggesting that it is a sulfoxide. This metabolite does not appear to be a degradation product, since its concentration does not increase with time. Two carbon signals were detected for metabolite 1. This compound must contain a third enriched carbon to produce the observed coupling patterns at 48 (d, la) and 11 (d,d, lb) ppm. The structure proposed for this metabolite is XCH2CH2CN. From the chemical shift data, X is consistent with either an RNH group or a sulfoxide (Table I). In addition to the signals listed in Table I, several other signals with coupling associated with 13C have been detected but not fully characterized at this time. The signal at 54 ppm, for example, has characteristics associated with a W-enriched compound, but insufficient data are available for a structural assignment. Quantitative Analysis of Rat Urinary Metabolites. For quantitation of metabolites, the relaxation times (TJ of the standard dioxane and the metabolites of interest were determined in urine. The Tl for dioxane in urine was approximately 6 s, and the T1 values for most of the metabolites were less than 4 s. Quantitation experiments were

Chem. Res. Toxicol., Vol. 4, No. 6, 1991 683

A

s Figure 4. 'H-decoupled 'SC NMFt spectra of control mouse urine (A) and urine collected for 24 h after administration of [1,2,313C]AN (10 mg/kg, PO) (B). Signals from metabolites are labeled according to metabolite number (see Table I and Scheme I) and the letter of the carbon derived from AN (,CH,=&H--,CN). Table IV. Excretion of Glucose in Rat Urine Measured by NMR Spectroscopy or by an Enzymatic Method glucose concn, mM, determined by NMR enzyme control 1 ND 0.1 2 ND 1.25 30 mg/ kg AN 1 33.1 32.2 2 28.5 27.4 3 12.2 10.9

run with a delay time of 24 s, a 90' pulse width, and decoupling only during acquisition to ensure complete relaxation and to avoid NOE accumulation. Quantitation of thiocyanate was not carried out, due to ita long relaxation time (about 60 s). The concentrations of metabolites detected in rat urine following administration of 30 mg/kg AN are shown in Table 111. Metabolites 2, 3,4,4', 5 , 5', and 6 amounted to 55% of the total dose excreted in 24 h. The signals for metabolite 1 were not quantitated because of low signal:noise, but its concentration is estimated to be approximately 0.1 mM. For several of the metabolites (2, 3,4, and 49, quantitation was carried out for two carbon resonances within the same molecule, yielding similar resulta (data not shown). The major metabolite detected was metabolite 2, formed by direct conjugation of AN with GSH. The other metabolites, derived from cyanoethylene oxide, accounted for approximately 60% of the metabolites. Similar observations were made at 10 mg/kg, with 46 f 7% of the urinary metabolites derived from direct conjugation (metabolite 2) in three individual animals. Comparison of the concentrations of glucose measured from the NMR spectra and measurements by a standard enzymatic assay was conducted as a validation of the quantitative determinations. The data obtained from the two methods were nearly identical (Table IV), indicating the validity of the NMR quantitation. Metabolites in Mouse Urine. The spectrum of control mouse urine was similar to that observed for the rat (Figure 4A). Following administration of 10 mg/kg [1,2,3-'SC]AN, 13C resonances were detected in mouse urine (Figure 4B) with chemical shifts similar to those found in rat urine,

Fennell et al.

684 Chem. Res. Toricol., Vol. 4, No. 6,1991 Table V. Quantitative Determination of [ 1,2,3-'C]AN Metabolites Excreted in Mouse Urinea concentra% of total metabolite chemical shift, w m tion, mM metabolitesb 0.29 f 0.04 20.5 f 2.0 2 26.9 3 60.1 0.32 f 0.03 22.3 f 1.2 0.20 f 0.05 13.9 f 3.2 4,4' 61.3 5,5' 37.2 0.61 f 0.03 43.2 f 3.5 DMicewere administered a dose of 4.64 f 0.18 pmol of [1,2,3'3C]AN (10mg/kg), PO, and metabolites measured in 0-24-h urine using dioxane as a standard. The total amount excreted in the urine in 24 h was 2.52 f 0.57 pmol (the sum of amounts of the metabolites listed above). Values represent mean f SD (n = 3). bThe amount of each metabolite expressed as a percentage of the total excreted in the urine in 24 h.

indicating the formation of similar metabolites. However, quantitative differences were found in the relative intensities of resonances. Signals for glucose were observed in the urine of some of the mice treated with 10 mg/kg AN, suggesting that AN is nephrotoxic in mice as well as rats. The spectra of urine samples from mice treated with 10 mg/kg [ 1,2,3-13C]ANcontained resonances assigned to metabolites 2, 3,4,4', 5,5', 6 and thiocyanate (Figure 4B). The most intense metabolite resonances detected were assigned to thiodiglycolic acid and N-acetyl-S- (carboxymethy1)cysteine (or its deacetylated product). Signals corresponding to metabolite 1were not detected in samples of mouse urine. In some of the samples examined, metabolite 6 was detected at low levels. In the mouse, the presence of a second RSCH2CH2CNmetabolite was indicated by an additional set of resonances similar to those of metabolite 2. Quantitative Analysis of Mouse Urinary Metabolites. The greater toxicity of AN in the mouse restricted the highest dose used in these studies to 10 mg/kg. Quantitation of metabolites in mouse urine was complicated by the resulting lower concentration of metabolites excreted. Metabolites contained in one mouse urine sample were quantitated using a long delay time with proton decoupling only during acquisition. The calculated concentrations were similar to those determined from spectra acquired with a 5-s delay, a 45" pulse width, and proton decoupling. Experiments with rat urine comparing quantitation under both conditions produced similar results. Therefore, the mouse quantitation data reported here are derived from spectra acquired with a delay of 5 s and 'H decoupling. In the 24 h following administration of AN to mice, the metabolites excreted in the urine accounted for approximately 55% of the administered material. In general, the concentrations of metabolites in mouse urine were considerably lower than those observed in rat urine, which is to be expected from the lower dose administered and the lower body weight. In the mouse, the excretion of S(carboxymethy1)cysteine (or its N-acetyl derivative) and thiodiglycolic acid (metabolites 5 and 5') was more extensive than that of the other metabolites (Table V) and was approximately 6-fold higher than in the rat. The formation of metabolite 2 was considerably less in the mouse (20%)than in the rat (43%, Table 111). These data indicate that a greater proportion of AN is metabolized via CEO in the mouse than in the rat.

Dlscusslon This study has demonstrated the utility of 13C NMR in the detection, identification, and quantitation of urinary metabolites. Using the approach described in this paper, the presence of previously described metabolites has been

' -

Scheme I. Proposed Metabolism Scheme for AN in the Rat a n d Mousea CHz-CH-CN

1 0 I \

GS-CHz-CH&N R-CHz-CHp-CN Metabollte 17

H~c-CH-CN

I

CNe

'I

c~~.SCH~.COOH

I

----

N-AcC~S-SCHZ-COOH Metabollte 5'

N-AcCys-SCH&H,.CN Metabollte 2 HOOCCH+-CHz-CHz-CN

CHZ-OH

GS-~H-CN

GS-CHz-CHOH-CN

----

FHZ-OH

----

N-AcCyS-S-CH-CN Metabolite 4,4'

SCNe Metabollte 7

HOOCCHZ-S-CHZ-COOH Metabollte 5

'4

1,

HOOC-CHz-S-CH&OOH Metabollte 6

" G S represents a glutathionyl residue, and Cys-S and N-AcCys-S represent cysteinyl and N-acetylcysteinyl residues, respectively. The broken arrows represent processes which involve several transformations.

verified, and the identification of additional new metabolites has been accomplished. Separation of the metabolites was not required, and considerable structural information was obtained from the spectral data without the need for metabolite standards. A parallel study,2 using [2,3-14C]AN,with HPLC and GC/MS analysis of the dose dependence of AN metabolism utilized the information from this study in the preparation of standards and provided confirmation of metabolite identification. Additional support is provided by studies reported by others, discussed below, which have previously assigned some of these metabolites in the rat. The overall metabolism scheme deduced from this study and those discussed below is shown in Scheme I. In both the rat and the mouse, the major metabolite formed by direct reaction of AN with glutathione is excreted as N acetyl-S-(2-~yanoethyl)cysteine, an observation in agreement with previous studies (3-5). In the mouse, signals indicating the presence of an additional RSCH2CH2CN metabolite were detected. This is probably S-(2-cyanoethy1)thioacetic acid, which has been detected by other techniques in urine of mice dosed with ANS2On administration of 30 mg/kg AN in the rat, there was a metabolite formed which may be due to direct reaction with an amine or formation of a sulfoxide (metabolite 1). Similar observations have been made for acrylic acid, where a major metabolite thought to be formed by direct reaction with an amine has been reported (30). Oswald et al. (31)described the urinary excretion of tertiary aminomethylenedioxypropiophenones(Mannich bases) after the administration of safrole to rats or guinea pigs. These metabolites were suggested to arise from the direct addition of the appropriate secondary amine to 1'-oxosafrole, an a,&unsaturated ketone. The second major pathway, via epoxidation of AN to G. L. Kedderis, S. C.J. Sumner, S. D. Held, R. Batra, M. J. Turner, Jr., A. E. Roberts, and T. R. Fennell, unpublished results.

Urinary Metabolites of Acrylonitrile

Chem. Res. Toxicol., Vol. 4, No.6, 1991 685

idence for the formation of such a product was obtained give CEO, is involved in the generation of the remaining in this study. urinary metabolites (6). A number of metabolites are derived from nucleophilic attack by glutathione at the The excretion of unchanged AN in urine has been re3-position of CEO. Cyanide is released from the resulting ported in several inhalation studies (8, 40), but was not cyanohydrin and converted to thiocyanate (2). The reseen in this study. The previous observations in both maining S-(2-oxoethyl)glutathionecan undergo either reanimals and humans suggest that AN in the urine did not duction to N-acetyl-S-(2-hydroxyethyl)cysteine( 4 ) or oxarise by direct transfer from the exposure atmosphere. idation, eventually yielding a number of S-(carboxymethyl) The presence of AN in the urine may be related to the products, namely, thiodiglycolic acid, N-acetyl-S-(carbroute of exposure, or it may have been generated by degoxymethy1)cysteine (81, and a further metabolite tentaradation of metabolites during sample workup for analysis. tively identified in this study as thionyldiacetic acid. The Quantitative analysis of the metabolites of AN produced method of detection used by Muller et al. (8) did not by direct conjugation with glutathione and by formation distinguish between N-acetyl- and S-(carboxymethyl)of CEO indicates that more CEO is produced in the mouse cysteine. N-Acetyl-S-(carboxymethy1)cysteinehas been than in rat. This is consistent with a number of previous reported as a metabolite of ethylene oxide, bromoethanol, observations. Microsomes from mouse liver have a higher and S-(carboxymethy1)cysteine ( 3 2 , 3 3 ) ,suggesting that V,, than those from rat liver for oxidation of AN to CEO the N-acetyl form is the more probable in this study. in vitro (41, 42). Following administration of AN iv, ip, Thiodiglycolic acid was the major metabolite of (carboxor PO, greater excretion of thiocyanate was observed in ymethy1)cysteine in man, with its sulfoxide present in mice compared with rats (9),suggesting a difference in the smaller amounts ( 18). Chloroethanol undergoes conversion extent of metabolism to CEO, or its further metabolism to thiodiglycolic acid and thionyldiacetic acid in the rat to release cyanide. Lambotte-Vandepaer et al. (39) re(34). Thus, thiodiglycolic acid may undergo oxidation to ported a similar difference in thiocyanate excretion after thionyldiacetic acid in animals receiving AN. The conip administration of AN to rats and mice, but observed version of a-thiocarboxylic acids to the corresponding little difference between species for the excretion of Nsulfoxides has been reported in rat liver microsomes (35). acetyl-S-(2-hydroxyethyl)cysteine or N-acetyl-S-(2Comparison of the ratios of metabolites derived from recyanoethy1)cysteine. The results of our studies and other duction and oxidation of S-(2-oxoethy1)glutathione(meinvestigations2 indicate that the major quantitative diftabolite 3 vs 5,5‘ and 6) for rat (2.05) and mouse (0.52) ferences between rat and mouse become fully apparent suggests marked species differences in the activities of the only on measurement of additional metabolites such as enzymes responsible for metabolism of this aldehyde thiodiglycolic acid, S-(carboxymethy1)cysteineor its Nconjugate. acetyl derivative, thionyldiacetic acid, and N-acetyl-S-(lcyano-2-hydroxyethy1)cysteine.The ratio of metabolites Conjugation at the 2-position of CEO resulted in the formation of N-acetyl-S-(l-cyano-2-hydroxyethyl)cysteine, derived from glutathione conjugation of CEO at the 2- and 3-positions will determine the amount of cyanide released. as described by Linhart et al. (7). This metabolite had previously been identified as 4-acetyl-3-carboxy-5-cyano- The difference in the ratios of metabolites 4,4’vs 3,5,5’, and 6 in rats (0.43) and mice (0.21) suggests that a greater tetrahydro-1,4-2H-thiazine by mass spectrometry as its percentage of the CEO produced in mice is metabolized methyl ester (3). However, further investigation indicated to produce cyanide. This, together with the greater extent that an isomeric compound, N-acetyl-S-(l-cyanoof metabolism via CEO in the mouse, may be responsible etheny1)cysteine methyl ester, was the compound detected for the greater acute toxicity of AN in the mouse compared (7). It was suggested that this compound was not present with the rat (2). Since the carcinogenicity of AN in the in the urine but was in fact formed by facile dehydration mouse and the role of CEO in AN carcinogenesis have not of N-acetyl-S-(l-cyano-2-hydroxyethyl)cysteinein the GC yet been investigated, the significance of the greater extent injector. The NMR analysis carried out in this study has of metabolism of AN to CEO in mice is currently not provided direct evidence for the presence of a metabolite known. with the structure of RSCH(CN)CH20H. Comparison of the ratios of metabolites derived from conjugation at the The route of administration of AN is also known to 2- and 3-positions (metabolites 4,4’vs 3, 5,5’, and 6) suginfluence its metabolism, with greater production of gests a difference in site specificity for conjugation between thiocyanate in rats after inhalation exposure or oral adrats (0.43) and mice (0.21). ministration than on ip or iv injection (43,44). Similarly, higher levels of N-acetyl-S-(2-hydroxyethyl)cysteineand A number of metabolites of AN which have been dewere lower levels of N-acetyl-S-(2-~yanoethyl)cysteine scribed previously either in vivo or in vitro were not deobserved on inhalation exposure of rats to AN, compared tected as urinary metabolites in this study. Acetic acid with ip or iv administration (44,45).The reasons for these has been reported as an in vitro metabolite derived from differences merit further investigation. pyruvonitrile via CEO (36). Neither of these compounds was detected in this study. Any acetic acid formed in vivo Various metabolites have been proposed as biomarkers would presumably enter normal metabolic processes and of exposure for acrylonitrile (45-47). Since a biomarker would be exhaled as C02. Metabolism studies with [2,3should be highly selective for exposure to a given com14C]AN did not result in the exhalation of significant pound, a number of the metabolites detected in this study amounts of 14C02(37). Cyanoacetaldehyde has also been would be unsuitable as biomarkers. A considerable portion suggested as an intermediate in the production of cyanoof administered AN is metabolized to N-acetyl-S-(2acetic acid and cyanoethanol,both in vitro and in vivo (38). cyanoethyl)cysteine, and since this is thought to be specific However, neither of these end products was detected as to AN, this metabolite would represent a suitable marker for exposure to AN (45, 46). Many of the products of metabolites in this study. Metabolites derived from the further metabolism of CEO, such as N-acetyl-S-(2hydrolysis of CEO, such as glycolaldehyde (36) or its hydroxyethyl)cysteine, thiodiglycolic acid, and N-acetylfurther metabolites, glycine and ethanolamine, were not S-(carboxymethyl)cysteine, are excreted as metabolites of observed. The formation of a glucuronide metabolite has other compounds, including vinyl chloride, ethylene oxide, been suggested from the increase in mutagenicity of urine chloroethanol, and dichloroethane (32, 34, 47, 48). The following treatment with @-glucuronidase(39),but no ev-

686 Chem. Res. Toxicol., Vol. 4, No. 6, 1991

only metabolite produced exclusively from CEO was N acetyl-S-(l-cyano-2-hydroxyethyl)cysteine. This compound could thus serve as an indicator of the metabolism of AN to CEO in vivo. A major use of AN is in the manufacture of acrylamide (49) which could result in exposure to both chemicals. Interestingly, none of the metabolites produced from AN were identical to those of acrylamide, although the latter is metabolized via similar routes (50). The use of Wlabeled exogenous compounds together with NMR spectroscopy can provide considerable information about their metabolism. In this study, a compound labeled in all three carbon atoms was administered and metabolites were examined directly in urine by 1-and 2-D NMR techniques to determine metabolite structures. Such an approach permits the simultaneous detection of all the metabolites excreted in the urine, as observed for compounds labeled in a single position (1618). The use of 2-D NMR experiments, such as HET2DJ and INADEQUATE, permits characterization of metabolites directly in the urine. Techniques for isolation of metabolites may be selective and result in the loss of volatile, polar, or chemically labile products. Direct analysis of metabolites without isolation thus avoids these potential problems. Using appropriate experimental conditions, quantitative analysis of the metabolites can be readily accomplished. The resolution afforded by NMR spectroscopy and the large range of intensities of signals which can be examined can result in the detection of metabolites which would be difficult to resolve by other techniques, particularly those that represent a small percentage of the total material present.

Acknowledgment. We thank J. P. MacNeela, M. J. Turner, Jr., and S. D. Held for their assistance and P. W. Ross, C. U. Parkinson, and T. Shepard for their help with animal experiments. This research was supported in part with funding from BP America, American Cyanamid, and Sterling Chemical.

References (1) Maltoni, C., Ciliberti, A,, Cotti, G., and Perino, G. (1987) Experimental research on acrylonitrile carcinogenesis, Princeton Scientific Publishing, Princeton. (2) Ahmed, A. E., and Patel, K. (1981) Acrylonitrile: in vivo metabolism in rata and mice. Drug Metab. Dispos. 9, 219-222. (3) Langvardt, P. W., Putzig, C. L., Braun, W. H., and Young, J. D. (1980) Identification of the major urinary metabolites of acrylonitrile in the rat. J. Toxicol. Enuiron. Health 6, 273-282. (4) van Bladeren, P. J., Delbressine, L. P. C., Hoogerterp, J. J., Beaumont, A. H. G. M., Breimer, D. D., Seutter-Berlage, F., and van der Gen, A. (1981) Formation of mercapturic acids from acrylonitrile, crotononitrile, and cinnamonitrile by direct conjugation and via an intermediate oxidation process. Drug Metab. Dispos. 9, 246-249. (5) Kopeky, J., Gut, I., Nerudova, J., Zachardova, D., and Holecek, V. (1980) Two routes of acrylonitrile metabolism. J. Hyg., Epidemiol., Microbiol., Immunol. 24, 356-362. (6) Guengerich, F. P., Geiger, L. E., Hogy, L. L., and Wright, P. L. (1981) In vitro metabolism of acrylonitrile to cyanoethylene oxide, reaction with glutathione, and irreversible binding to proteins and nucleic acids. Cancer Res. 41, 4925-4933. (7) Linhart, I., Smejkal, J., and Novak, J. (1988) N-Acetyl-S-(1cyano-2-hydroxyethyl)-L-cysteine,a new urinary metabolite of acrylonitrile and oxiranecarbonitrile. Arch. Toxicol. 61, 484-488. (8) Muller, G., Verkoyen, C., Soton, N., and Norpoth, K. (1987) Urinary excretion of acrylonitrile and its metabolites in rats. Arch. Toxicol. 60, 464-466. (9) Gut, I., Nerudova, J., Kopecky, J., and Holecek, V. (1975) Acrylonitrile biotransformation in rats, mice, and Chinese hamsters as influenced by the route of administration and by phenobarbital, SKF 525-A,-cysteine,dimercaprol, or thiosulfa& Arch. Toxicol. 33, 151-161. (10) Miller, E. C., and Miller, J. A. (1981) Searches for ultimate chemical carcinogens and their reactions with cellular macro-

Fennel1 et al. molecules. Cancer 47, 2327-2345. (11) Recio, L., and Skopek, T. (1988) Mutagenicity of acrylonitrile and its metabolite 2-cyanoethylene oxide in human lymphoblasts in vitro. Mutat. Res. 206, 297-305. (12) Lundberg, P., Harmsen, E., Ho, C., and Vogel, H. J. (1990) Nuclear magnetic resonance studies of cellular metabolism. Anal. Biochem. 191, 193-222. (13) Nicholson, J. K., and Wilson, I. D. (1987) High resolution nuclear magnetic resonance spectroscopy of biological samples as an aid to drug development. Prog. Drug Res. 31,427-479. (14) Nicholson, J. K., and Wilson, I. D. (1989) High resolution proton magnetic resonance spectroscopy of biological fluids. Prog. Nucl. Magn. Reson. Spectrosc. 21,449-501. (15) Hunter, B. K., Nicholls, K. M., and Sanders, J. K. M. (1984) Formaldehyde metabolism by Escherichia coli. In vivo carbon, deuterium, and two-dimensional NMR observations of multiple detoxifying pathways. Biochemistry 23, 508-514. (16) Dohn, D. R., Graziano, M. J., and Casida, J. E. (1988) Metabin male rata studied olites of [3-I3C]1,2-dibromo-3-chloropropane by and IH-13C correlated two-dimensional NMR spectroscopy. Biochem. Pharmacol. 37, 3485-3495. (17) Tecle, B., and Casida, J. E. (1989) Enzymatic defluorination and metabolism of fluoroacetate, fluoroacetamide, fluoroethanol, and (-)-fluorocitrate in rata and mice examined by IgF and 13C NMR. Chem. Res. Toxicol. 2,429-435. (18) Meese, C. O., and Fischer, P. (1990) Tracing the human metabolism of stable isotope-labelled drugs by ex vivo NMR spectroscopy. A revision of S-carboxymethyl-L-cysteinebiotransformation. 2.Naturforsch, 45C, 1171-1175. (19) Friedman, M., Cavins, J. F., and Wall, J. S. (1965) Relative nucleophilic reactivities of amino groups and mercaptide ions in addition reactions with a$-unsaturated compounds. J. Am. Chem. SOC. 87, 3672-3682. (20) Zilkha, A., and Rappoport, S. (1963) Syntheses of hydroxyalkyl and mercaptoalkyl derivatives of sulfur-containing amino acids. J. Org. Chem. 28, 1105-1107. (21) Oddy, H. G., and Dodson, V. H. (1949) Thionyldiacetic acid. Ohio J. Sci. 49, 149-153. (22) Muller, L., Kumar, A,, and Ernst, R. R. (1975) Two-dimensional carbon-13 NMR spectroscopy. J. Chem. Phys. 63, 5490-5491. (23) Bax, A., Freeman, R., and Frenkiel, T. A. (1981) An NMR technique for tracing out the carbon skeleton of an organic molecule. J. Am. Chem. SOC. 103, 2102-2104. (24) Wehrli, F. W., and Wirthlin, T. (1976) Interpretation of carbon-13 NMR spectra, Heyden & Son, New York. (25) Breitmaier, E., and Voelter, W. (1987) Carbon-23 NMR spectroscopy: high resolution methods and applications in organic chemistry and biochemistry, VCH, New York. (26) Rouisse, L., Chakrabarti, S., and Tuchweber, B. (1986) Acute nephrotoxic potential of acrylonitrile in Fischer-344 rata. Res. Commun. Chem. Pathol. Pharmacol. 53, 347-360. (27) Vodicka, P., Gut, I., and Frantik, E. (1990) Effects of inhaled acrylic acid derivatives in rata. Toxicology 65, 209-221. (28) Chasseaud, L. F. (1976) Conjugation with glutathione and mercapturic acid excretion. In Glutathione: metabolism and junction (Arias, I. M., and Jakoby, W. B., Eds.) pp 77-114, Raven Press, New York. (29) Kopecky, J., Smejkal, J., Linhart, I., Hanus, V., and Turecek, F. (1984) Reaction of oxiranecarbonitrile with L-cysteine methyl ester. Tetrahedron Lett. 25, 4295-4298. (30) Winter, S. M., MacKenzie, N. E., Gooley, P. R., and Sipes, I. G. (1990) Identification of the urinary metabolites of [1,2,313C3]acrylicacid in the rat by 13C nuclear magnetic resonance spectroscopy. Toxicologist 10, Abstract 1331. (31) Oswald, E. O., Fishbein, L., Corbett, B. J., and Walker, M. P. (1971) Identification of tertiary aminomethylenedioxypropiophenones as urinary metabolites of safrole in the rat and guinea pig. Biochim. Biophys. Acta 230, 237-247. (32) Jones, A. R., and Wells, G. (1981) The comparative metabolism of 2-bromoethanol and ethylene oxide in the rat. Xenobiotica 11, 763-770. (33) Waring, R. H. (1978) The metabolism of S-carboxymethylcysteine in rodents, marmosets and humans. Xenobiotica 8, 265-270. (34) Grunow, W., and Altmann, H.-J. (1982) Toxicokinetics of chloroethanol in the rat after single oral administration. Arch. Toxicol. 49, 275-284. (35) Lee, Y. C., Joiner Hayes, M. G., and McCormick, D. B. (1970) Microsomal oxidation of a-thiocarboxylic acids to sulfoxides. Biochem. Pharmacol. 19, 2825-2832.

Chem. Res. Toxicol. 1991,4, 687-691 (36) Duverger-Van Bogaert, M., Lambotte-Vandepaer, M., De Meester, C., Rollmann, B., Poncelet, F., and Mercier, M. (1981) Effect of several factors on the liver extract mediated mutagenicity of acrylonitrile and identification of four new in vitro metabolites. Toxicol. Lett. 7, 311-318. (37) Ahmed, A. E., Farooqui, M. Y. H., Upreti, R. J., and El-Shabrawy, 0. (1983) Comparative toxicokinetics of 2,3-14C- and 114C-acrylonitrilein the rat. J. Appl. Toxicol. 3, 39-47. (38) Lambotte-Vandepaer, M., Duverger-Van Bogaert, M., de Meester, C., Rollman, B., Poncelet, F., and Mercier, M. (1981) Identification of two urinary metabolites of rats treated with acrylonitrile; influence of several inhibitors on the mutagenicity of those urines. Toxicol. Lett. 7, 321-328. (39) Lambotte-Vandepaer, M., Duverger-Van Bogaert, M., de Meester, C., Poncelet, F., and Mercier, M. (1985) Mutagenicity of urine from rats and mice treated with acrylonitrile. Toxicology 16,67-71. (40) Sakurai, H., Onodera, M., Utsunomiya, T., Minakuchi, H., Iwai, H., and Matsumura, H. (1978) Health effects of acrylonitrile in acrylic fibre factories. Br. J. Znd. Med. 35, 219-222. (41) Roberts, A. E., Lacey, S. A., Pilon, D., Turner, M. J., and Rickert, D. E. (1989) Metabolism of acrylonitrile to 2-cyanoethylene oxide in F-344 rat liver microsomes, lung microsomes, and lung cells. Drug Metab. Dispos. 17, 481-486. (42) Roberts, A. E., Kedderis, G. L., Turner, M. J., Rickert, D. E., and Swenberg, J. A. (1991) Species comparison of acrylonitrile epoxidation by microsomes from mice, rats, and humans: relationship to epoxide concentrations in mouse and rat blood. Carcinogenesis 12, 401-404. (43) Silver, E. H., Kuttab, S. H., Hassan, T., and Hassan, M. (1982) Structural considerations in the metabolism of nitriles to cyanide

687

in vivo. Drug. Metab. Dispos. 10, 495-498. (44)Tardif, R., Talbot, D., GBrin, M., and Brodeur, J. (1987) Urinary excretion of mercapturic acids and thiocyanate in rata exposed to acrylonitrile: influence of dose and route of administration. Toxicol. Lett. 39,255-261. (45) GBrin, M., Tardif, R., and Brodeur, J. (1988) Determination of specific urinary thioethen derived from acrylonitrile and ethylene oxide. In Methods for Detecting DNA Damaging Agents in Humans: Applications in Cancer Epidemiology and Prevention (Bartsch, H., Hemminki, K., and O'Neill, I. K., Eds.) IARC Scientific Publications 89, pp 275-278, International Agency for Research on Cancer, Lyon. (46) Jakubowski, M., Linhart, I., Pielas, G., and Kopecky, J. (1987) 2-Cyanoethylmercapturicacid (CEMA) in the urine as a possible indicator of exposure to acrylonitrile. Br. J. Ind. Med. 44, 834-840. (47) Vermeulen, N. P. E., de Jong, J., van Bergen, E. J. G., and van Welie, R. T. H. (1989) N-Acetyl-S-(2-hydroxyethyl)-L-cysteine88 a potential tool in biological monitoring studies? A critical evaluation of possibilities and limitations. Arch. Toxicol. 63, 173-184. (48) Green, T., and Hathway, D. E. (1977) The chemistry and biogenesis of the S-containing metabolites of vinyl chloride in rata. Chem.-Bid. Interact. 17, 137-150. (49) International Agency for Research on Cancer (1986) Acrylamide. In IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans, Vol. 39, pp 41-66, International Agency for Research on Cancer, Lyon. (50) Sumner, S. C. J., MacNeela, J. P., and Fennell, T. R. (1989) Urinary metabolites of [ 1,2,3-13C]acrylamidedetermined by 13C nuclear magnetic resonance spectroscopy. Toxicologist 10, 332.

Quantitation of 5 4 Hydroxymethyl)uracil in DNA by Gas Chromatography with Mass Spectral Detection Zora Djuric,* Domenico A. Luongo, and Dorcas A. H a r p e r Division of Hematology and Oncology, Department of Internal Medicine, Wayne State University, P.O. Box 02188, Detroit, Michigan 48201 Received July 9, 1991

5-(Hydroxymethy1)uracilis a product of oxidative DNA damage. This hydroxylated base was quantified in DNA by GC-MS using either acid or enzymatic hydrolysis of the DNA and isotopically labeled internal standards. Both 5-(hydroxymethy1)uracil and thymine were quantified in each DNA sample and the results expressed as a ratio. This procedure controlled for possible errors in the quantitation of DNA prior to hydrolysis and derivatization. In addition, quantitation of thymine was important due to possible variations in DNA hydrolysis efficiency for each sample. The isotopically labeled internal standards controlled for compound instability through the procedure and for variations in derivatization efficiency. The conditions used for acid hydrolysis of the DNA resulted in considerable degradation of 5-(hydroxymethy1)uracil; however, since isotopically labeled 5-(hydroxymethy1)uracil was added prior to acid treatment, 5-(hydroxymethy1)uracil still could be quantified. The degradation of 5-(hydroxymethy1)uracil was avoided using enzymatic hydrolysis of the DNA. In DNA that had been treated with hydrogen peroxide and iron in the presence of EDTA, the observed level of 5-(hydroxymethy1)uracil using enzymatic hydrolysis was 1.6-fold higher than when using acid hydrolysis of the DNA. With analysis of 2 pg of DNA, the detection limit for 5-(hydroxymethyl)uracilwas 3/105 thymines.

I ntroductlon Oxidative DNA damage results from various endogenous biological processes as well as from chemical exposures. This type of DNA damage has been associated with the processes of aging and cancer (I). When DNA is treated with hydrogen peroxide and iron, the pyrimidines may be relatively more susceptible to oxidation than purines (2, 3),although under certain conditions high levels of mod0893-228x/91/2704-0687$02.50/0

ified purines also may be obtained (4). The stability of these products is an important consideration for their quantitation. Many hydroxylated DNA bases fragment readily ( 5 ) . For example, thymine glycol is decomposed by further oxidation with hydrogen peroxide and iron (6). Numerous free-radical-inducedproducts of thymine have been identified and their biological effects investigated. One product, 5-(hydroxymethyl)uracil,appears to be more 0 1991 American Chemical Society