Proc. Natl. Acad. Sci. USAVol. 83, pp. 8356-8360, November 1986Medical Sciences

Formation of muconaldehyde, an open-ring metabolite of benzene,in mouse liver microsomes: An additional pathway fortoxic metabolites

(ring opening/benzene metabolism/a,3-unsaturated aldehydes/2,4-hexadienedial)

LouISE LATRIANO*, BERNARD D. GOLDSTEIN, AND GISELA WITZtJoint Graduate Program in Toxicology, Department of Environmental and Community Medicine, University of Medicine & Dentistry of New Jersey-RobertWood Johnson Medical School/Rutgers University, Piscataway, NJ 08854

Communicated by Allan H. Conney, July 18, 1986

ABSTRACT It has been proposed that a ring-opened formmay be responsible for the toxicity of benzene. The presentstudies demonstrate that incubation of ("C]benzene with livermicrosomes (obtained from male CD-1 mice treated withbenzene) in the presence ofNADPH results in the formation ofa ring-opened product. Evidence for the identity of this productwas obtained by derivatizing with 2-thiobarbituric acid (TBA),which resulted in the formation of an adduct with a 490-nmabsorbance maximum. This''maximum is identical to thatobserved after authentic trans,trans-muconaldehyde has react-ed with TBA. Separation of muconaldehyde, both with andwithout trapping with TBA, from other benzene metabolites inthe incubation mixture was accomplished by HPLC. Theradioactivity profile of fractions collected during HPLC anal-ysis contained peaks that eluted with muconaldehyde and themuconaldehyde-TBA adduct. The structure of the ring-openedproduct was confirmed by mass spectrometry, studies in whic~hthe HPLC peak from the microsomal incubation mixture thateluted at the retention time of authentic muconaldehyde wascollected and derivatized with 2,4-dinitrophenylhydrazine.The high-resolution mass spectrum of this sample contained anIon with an m/z of 291.0729, corresponding to muconaldehydemono-dinitrophenylhydrazone. These results indicate that ben-zene is metabolized in vitro to a ring-opened product identifiedas muconaldehyde.

The search for potentially toxic metabolic intermediates ofaromatic compounds has classically focused on oxidationproducts of the ring(s) or of side chains. This productiveapproach has led to evidence that intermediates, such asaromatic oxides, epoxides, oxepins, peroxides, and polyhy-droxylated compounds, play a role in toxicity. Benzene is apotent hematotoxin and a known human leukemogen whosehematotoxic effects depend upon its metabolic transforma-tion (1, 2). However, despite a focus on the many metaboliteswith oxygen added to the benzene ring, the search for theagent(s) causing this toxicity has thus far been unsuccessful(2). While there has been some suggestion that a reactiveearly intermediate, such as benzene oxide-oxepin, might beof importance, and there is good evidence indicating thatoxidative metabolism to catechol, hydroquinone, andbenzoquinone has some role in benzene hematotoxicity, asyet the range of hematopoietic toxicity observed after ben-zene exposure has not been explained by intensive study ofits known metabolites.

This focus on products containing an intact aromatic ringhas ignored the possibility that toxic intermediates might beproduced through the opening of the ring. Our laboratory hasproposed that muconaldehyde, an a,4-unsaturated diene

dialdehyde, resulting from ring opening, is a hematotoxicintermediate in benzene metabolism (3). Benzene ring open-ing and formation of trans,trans-muconaldehyde had previ-ously been demonstrated to occur in irradiated solutions ofbenzene, presumably via the action of hydroxyl radicals (4),and more recently in our hands in a system in which aFenton-type reagent was used to generate free radicals (5).There is indirect evidence that benzene ring opening occursin vivo in that trans,trans-muconic acid, the correspondingdiacid of muconaldehyde, has been found in the urine ofanimals administered benzene, accounting for 1-2% of thedose (6, 7). However, this finding has been thought to be dueto the action of pyrocatechase, an enzyme found in intestinalbacteria, which converts catechol directly into cis,cis-muconic acid (8).

Initial toxicological studies showed that 10 AM muconal-dehyde produced marked toxicity to plasma clot cultures ofhuman erythropoiet'c precursors (3), and at similar or lowerconcentrations was toxic to other dividing cells in culture (9).Additionally, administration of trans,trans-muconaldehydeat 2 mg/kg daily for 16 days to CD-1 mice resulted instatistically significant decreases in erythrocyte count, hem-atocrit, and bone marrow cellularity as well as an increase inleukocyte count and spleen weight (10). This is similar to thetoxicity of about a 50- to 100-fold higher concentration ofbenzene in this species.The major missing link in the hypothesis that a ring-opened

product of benzene is important in its toxicity is the demon-stration that the known hematotoxic compound, muconalde-hyde, is indeed formed during the metabolism of benzene. Inthe present study we have evaluated this possibility in amouse liver microsomal system incubated with benzene. Theidentification of trans,trans-muconaldehyde (2, 4-hexadiene-dial) as a product of benzene ring fission was accomplishedby several methods of analysis, including HPLC and massspectrometry.

MATERIALS AND METHODSChemicals. trans,trans-Muconaldehyde was synthesized

according to the procedure of Kossmehl and Bohn (3, 11).The melting point of 119'C is in good agreement with thereported values of 117'C (12) and 121'C (13). The absorbancespectrum of trans,trans-muconaldehyde in phosphate bufferhas a 278-nm maximum; its molar extinction coefficient atthis wavelength is 3.15 x 1041 M-1cm-1 (14). A 10 mMsolution in absolute ethanol served as the stock.

Abbreviation: TBA, 2-thiobarbituric acid.*Present address: Columbia University Health Sciences, Institute ofCancer Research, New York, NY 10032.tTo whom reprints should be addressed at: Department of Environ-mental and Community Medicine, University of Medicine & Den-tistry ofNew Jersey, Robert Wood Johnson Medical School, BuschCampus, Piscataway, NJ 08854.

8356

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

Proc. Natl. Acad. Sci. USA 83 (1986) 8357

Carbonyl-free benzene, phenol, corn oil, HPLC grademonobasic potassium phosphate, and sodium dithionatewere obtained from Fisher. [14C]Benzene was obtained fromICN. Hydroquinone and 2,4-dinitrophenylhydrazine werepurchased from Aldrich. trans,trans-Muconic acid, NADPHas the tetrasodium salt (type IV, 98% pure), and 2-thiobar-bituric acid (TBA) were obtained from Sigma.Microsome Preparation and Incubation Procedure. Male

CD-1 mice weighing 25-35 g, purchased from Charles RiverBreeding Laboratories, were administered benzene at 1100mg/kg in corn oil (1:1) subcutaneously 24 and 18 hr prior tosacrifice, a treatment shown to increase benzene metabolism2- to 3-fold (15). Control animals received corn oil. Theanimals were fasted for 12 hr and the microsomes wereprepared (15). Cytochrome P-450 was determined by themethod of Omura and Sato (16) and protein was determinedby using the Lowry procedure (17).Mouse liver microsomes (1 mg of protein per ml) were

incubated with 4 mM [14C]benzene (specific activity 565dpm/nmol) and 3 mM NADPH in a total volume of 1.5 ml of0.1 M potassium phosphate buffer, pH 7.4. The incubationswere carried out in glass-stoppered test tubes that hold a totalvolume of 5 ml. Benzene was added without dilution in anyorganic solvent. The tubes were capped immediately after theaddition of benzene and incubated at 370C in a shaking waterbath for 5, 10, 15, or 20 min. Trichloroacetic acid (0.2 ml of10%, wt/vol) was added to 1-ml samples from the microsomesuspension. After centrifugation for 30 sec at 50 x g, thesupernatant solution was filtered with nylon HPLC filters andimmediately injected into the HPLC column.

Derivatization with TBA. Previous studies in our laboratoryhave demonstrated that trans, trans-muconaldehyde reactswith TBA to form an adduct with a 490-nm absorbancemaximum (18, 19). One milliliter of the microsomal suspen-sion was added to 1 ml of 0.67% TBA in 10% trichloroaceticacid and centrifuged at 50 x g for 5 min, and the supernatantsolution was incubated in a boiling water bath for 30 min. Themuconaldehyde-TBA standard was made by mixing 10 /iMmuconaldehyde in 0.1 M phosphate buffer, pH 7.4, with theTBA reagent (1:1) and incubating for 30 min in a boiling waterbath. Absorption spectra were recorded on a Perkin-Elmer552 spectrophotometer.

Analysis by HPLC. The HPLC method for separation anddetection of muconaldehyde utilized a 0.46 x 25 cm analyt-ical column (Whatman Chemical Separations) containing5-,um C18 reverse-phase packing. The mobile phase of 10mMpotassium phosphate buffer, pH 6.5, and 30-40% methanolwas delivered isocratically at a flow rate of 1 ml/min. A100-,ul Rheodyne loop was used throughout. A Kratos 773continuously variable absorbance detector was set to 280 nmfor detection of muconaldehyde without derivatization. Forthe detection of the muconaldehyde as the TBA adduct, thedetector was set to 490 nm, the absorbance maximum of thisadduct (19).

Determination of Radioactivity Profile. The HPLC eluatewas collected in 0.5-ml fractions into scintillation vialscontaining 5.0 ml of Liquiscint fluor (National Diagnostics,Somerville, NJ). Disintegrations per minute (dpm) weredetermined on a Tracer Analysis Mark III scintillationcounter. Purity was assessed by injecting 4 mM labeledbenzene in 0.1 M phosphate buffer, pH 7.4, into the columnand using the HPLC conditions previously described.

Derivatization with 2,4-Dinitrophenylhydrazine. The rea-gent was made by dissolving 0.4 g of 2,4-dinitrophenylhy-drazine in 2 ml of concentrated H2SO4, and adding 3 ml ofH20 and 30 ml of 95% ethanol. A 1-ml sample from themicrosome incubation mixture was treated with trichloro-acetic acid and centrifuged for 30 sec as described above. A100-/zl aliquot of the supernatant was immediately injectedinto the HPLC column. At the retention time (tR) previously

established for muconaldehyde, 1 ml of the eluate wascollected and allowed to react with 1 ml of 2,4-dinitrophen-ylhydrazine reagent. The control sample from a microsomalincubation mixture without benzene was prepared in thesame manner. The muconaldehyde dinitrophenylhydrazonestandard was prepared by adding 20 p.l of 10 mMmuconaldehyde in ethanol to 1 ml of 2,4-dinitrophenylhydra-zine reagent. After storage in the dark at room temperaturefor at least 24 hr, the samples were prepared for massspectrometry by extracting twice with 1.0 ml of chloroform.The extracts were pooled and evaporated to dryness undernitrogen at room temperature. The dark orange residue wasdissolved in one drop ofglycerol. Analysis was performed ona VG analytical high-resolution mass spectrometer model7070EQ.Recovery of Muconaldehyde from Microsomes. In these

studies, trans,trans-muconaldehyde was added to micro-somes (1 mg ofprotein per ml) to give final concentrations of50 and 5 A.M. The samples were incubated at 370C for 0-30min in a shaking water bath. One-milliliter aliquots wereprepared as previously described for HPLC analysis, withdetection by absorbance at 280 nm.

RESULTSThe incubation of benzene with liver microsomes obtainedfrom CD-1 male mice, treated with benzene, results in theformation of an open-ring metabolite, identified astrans,trans-muconaldehyde. This finding is supported byseveral lines of evidence employing three different methodsfor the detection of muconaldehyde in microsomal mixtures.

Studies Employing TBA as a Trapping Agent. Previousstudies (18, 19) had determined that the TBA assay was usefulfor the detection of muconaldehyde, which forms a TBAadduct with a 490-nm absorbance maximum. A sample frommicrosomes incubated with [14C]benzene in the presence ofNADPH and allowed to react with TBA contained a productthat had a 490-nm absorbance maximum, identical to thatformed by allowing authentic muconaldehyde to react withTBA. This maximum, however, was also observed afterreaction of a microsomal mixture containing NADPH withTBA. Fig. 1B shows the difference spectrum ofa sample fromthe microsomes incubated with benzene vs. a microsomalsample without benzene; both samples contained NADPH.

A B

1.0

420 460 500 540 42046050050.5C.0 -06

0.3

0.2-0.1

420 460 500 540 420 460 500 540

Wavelength, nm

FIG. 1. Absorbance spectra of solutions obtained from thereaction of TBA with trans,trans-muconaldehyde and microsomalincubation mixtures. (A) trans,trans-Muconaldehyde (0.01 mM)allowed to react with TBA. (B), Difference spectrum of sample frommicrosomes incubated for 20 min with benzene vs. a correspondingsample without benzene, each allowed to react with TBA.

Medical Sciences: Latriano et al.

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

8358 Medical Sciences: Latriano et al.

Though the difference spectrum was consistent with theformation of muconaldehyde from benzene (Fig. 1), theconfounding factor ofan apparent NADPH-TBA adduct withsimilar spectral characteristics necessitated further studies.The HPLC chromatogram of the supernatant from

microsomes incubated with 4 mM ['4C]benzene in the pres-ence of 3 mM NADPH for 10 min and then allowed to reactwith TBA is shown in Fig. 2A. When detection is byabsorbance at 490 nm, a peak with tR = 7.9 min is observed.A small peak at this retention time is also present in thecorresponding sample without benzene, which was alsodetermined to be due to NADPH. Since NADPH reacts withTBA to form an adduct with the same absorption maximumand HPLC retention time as the muconaldehyde-TBAadduct, it was necessary to identify muconaldehyde derivedfrom benzene by radioactive material eluting at the retentiontime previously established for muconaldehyde.The HPLC profile of muconaldehyde and the other known

metabolites of benzene, in the absence of TBA and using280-nm detection, is shown in Fig. 2B (solid line). Achromatogram of the muconaldehyde-TBA adduct, using490-nm detection (broken line), is also shown in Fig. 2B. Thechromatograms show that the muconaldehyde-TBA adduct,in contrast to muconaldehyde, is well separated from hydro-quinone, benzoquinone, catechol, and phenol, resulting ineasier identification of this peak. Fig. 2C shows the radio-activity profile of the HPLC eluate collected during analysisof the sample shown in Fig. 2A. Peak assignment based onstandards is as follows: tR = 5.0 min, muconaldehyde(without TBA); tR = 8.0 minm muconaldehyde-TBA; tR = 9.5min, phenol. The peaks with tR = 4.0 and 6.5 min cannot beidentified with certainty. Under these HPLC conditions,trans,trans-muconic acid elutes immediately after the solventfront (tR = 3.5 min). Assessment of the purity of theradiolabeled benzene reveals a contaminant that elutes at tR= 4.0 min. Therefore, the early peak may be a contaminant,muconic acid, or another highly polar metabolite.

Detection of Muconaldehyde Without Chemical Derivatiza-tion. The studies employing radioactively labeled benzeneand TBA demonstrated that trans,trans-muconaldehyde

v

CZ-D0

could also be detected without the use of a trapping agent. Aslight adjustment of the HPLC mobile phase allowed forbaseline separation from the other benzene metabolites, asshown by the HPLC profile of the standards shown in Fig.3B. Fig. 3A shows a chromatogram of the supernatant frommicrosomes incubated with benzene for 5 min in the presenceofNADPH. Two very small peaks absorbing at 280 nm withretention times of 6.7 and 7.0 min are present. The materialwith tR = 7.0 min elutes at the same time as the muconalde-hyde standard (Fig. 3B) and is not present in the correspond-ing sample without benzene (not shown). Fig. 3C shows theradioactivity profile of the HPLC eluate that was collectedduring analysis of the sample shown in Fig. 3A. Peakassignment in the radioactivity profile is based on comparisonwith the standards: the peak at tR = 5.5 min corresponds tohydroquinone, the peak at tR = 7.0 min corresponds tomuconaldehyde, and the peak at tR = 14.0 min correspondsto phenol. Assignment of the peak at tR = 4.5 min iscomplicated by a radioactive contaminant present in the[14C]benzene. The radioactivity profiles of chromatogramsfrom microsomal samples incubated with benzene for 10 and15 min were similar except that all peaks contained moreradioactivity, indicating increased product formation overtime. For muconaldehyde, the difference in dpm over timewas only a few counts and could not be considered signifi-cantly different from the value observed in the 5-minmicrosomal incubation of benzene.

In the above studies employing radioactive benzene, ex-periments were carried out to ascertain that the radioactivityeluting at the retention time of muconaldehyde was associ-ated only with muconaldehyde and not with any otherbenzene metabolite. These studies included determining theretention times of phenol, catechol, hydroquinone, benzo-quinone, 1,2,4-benzenetriol and muconic acid in both thepresence and absence ofTBA, with detection at both 280 and490 nm. None of these metabolites eluted with muconalde-hyde or the muconaldehyde-TBA adduct.

Detection of Muconaldehyde by Derivatization with 2,4-Dinitrophenylhydrazine. A third method for the detection ofmuconaldehyde employed post-column derivatization ofmuconaldehyde with 2,4-dinitrophenylhydrazine followed bymass spectral analysis. Fig. 4A shows the fragmentationpattern of the muconaldehyde dinitrophenylhydrazone stan-

Wu

ctCZ.00.0

2 4 6 8 10 2 6 10Time, min

FIG. 2. HPLC chromatogram and radioactivity profile of asample from the microsomal incubation mixture containing['4C]benzene, and HPLC chromatogram of a mixture of standards ofbenzene metabolites. The mobile phase was 40:60 (vol/vol)methanol/phosphate buffer, pH 6.5. (A) HPLC chromatogram ofmicrosomal incubation mixture containing [14C]benzene (565dpm/nmol). Samples were allowed to react with TBA. Detection wasby absorbance at 490 nm. (B) -, HPLC chromatogram of a mixtureof 1 mM muconic acid (tR = 3.5 min), 1 mM hydroquinone (tR = 4.6min), 1 mM benzoquinone (tR = 5.6 min), 1 mM catechol (tR = 5.6min), 1 mM phenol (tR = 9.6 min), and 0.01 mM muconaldehyde (tR= 5.1 min). Samples were not allowed to react with TBA. Detectionwas by absorbance at 280 nm. ----, HPLC chromatogram of 0.01 mMtrans,trans-muconaldehyde allowed to react with TBA (tR = 7.9min). Detection was by absorbance at 490 nm. (C) Radioactivityprofile of HPLC eluate collected during chromatographic analysis ofsample in A.

4 8 12 16

0.5 B

4, 812 164 8 1216

0.

C

-180

-80

340 12

4 8 12

Time, min

FIG. 3. HPLC chromatogram and radioactivity profile of asample from the microsomal incubation mixture containing[14C]benzene, and HPLC chromatogram of a mixture of standards ofbenzene metabolites. The mobile phase was 30:70 (vol/vol)methanol/phosphate buffer, pH 6.5. (A) HPLC chromatogram ofmicrosomal incubation mixture containing ['4C]benzene (565dpm/nmol). Detection was by absorbance at 280 nm. (B) HPLCchromatogram of 1 mM muconic acid (tR = 3.5 min), 1 mMhydroquinone (tR = 5.5 min), 1 mM catechol (tR = 8.9 min), 1 mMphenol (tR = 15.9 min), and 0.01 mM muconaldehyde (tR = 6.9 min)in 0.01 M phosphate buffer, pH 7.4. Detection was by absorbance at280 nm. (C) Radioactivity profile of HPLC eluate collected duringchromatographic analysis of sample in A.

Proc. Natl. Acad. Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

Proc. NatL. Acad. Sci. USA 83 (1986) 8359

dard obtained by fast atom bombardment (FAB) mass spec-trometry in which glycerol was the matrix. The material withan m/z of 291 corresponds to the protonated muconaldehydemono-dinitrophenylhydrazone. The mass spectrum of thestandard contains two other peaks-i.e., the peak with anm/z of 185, which corresponds to a dimer ofglycerol plus oneproton, and a peak with an m/z of 199, which corresponds toprotonated dinitrophenylhydrazine. Fig. 4B shows the massspectrum of a sample from the microsomal incubation mix-ture with benzene, collected from the HPLC column at theretention time characteristic for muconaldehyde and subse-quently derivatized with dinitrophenylhydrazine. This spec-trum contains a peak with an m/z of 291, which correspondsto the protonated muconaldehyde mono-dinitrophenylhydra-zone. The high-precision mass value (291.0729) obtained onthis sample was consistent with that of the muconaldehydedinitrophenylhydrazone standard. Fig. 4B also contains apeak with an m/z of 281. The identity of this peak is unknown,but it is also found in the analysis ofa sample from a similarlytreated microsomal incubation mixture without benzene (Fig.4C).

Quantitation of Benzene-Derived Muconaldehyde. Theamount of recoverable metabolites formed during a 10-minincubation of [14C]benzene with microsomes is shown inTable 1. The peak numbers correspond to the radioactivepeaks depicted in Fig. 3C. Recoverable muconaldehydeformed in this system can be directly calculated from thespecific activity of the radioactively labeled benzene (565dpm/nmol) and is expressed as benzene equivalents/mg ofprotein. Recoverable muconaldehyde formed from 4 mM[14C]benzene is 0.84 nmol/mg of protein or 0.87 nmol/nmolof cytochrome P-450.

10-

5.

10

5

101

Muconaldehyde2, 4-dinitrophenylhydrazone A

291

With benzene B

IL 1291Without benzene C

260 290

&d.

180 200m/z

FIG. 4. Mass spectrometry of the chloroform extracts fromreaction of 2,4-dinitrophenylhydrazine with samples from themicrosomal incubation mixture and the trans,trans-muconaldehydestandard. (A) Standard trans,trans-muconaldehyde (0.2 mM finalconcentration) that had reacted with 2,4-dinitrophenylhydrazone.(B) Microsomal incubation mixture containing 15 mM benzene. (C)Microsomal incubation mixture without benzene. The backgroundpeaks were higher in this sample compared with the correspondingsample with benzene because a greater amount of sample wasanalyzed in the mass spectrometer.

Table 1. Amounts of recoverable metabolites formed duringa 10-min incubation of [14C]benzene in liver microsomesobtained from CD-1 mice treated with benzene

Peak Benzene equivalents, % ofno. Identity nmol/mg protein total metabolites1 Unknown 0.44 52 Hydroquinone 1.12 133 Muconaldehyde 0.84 94 Phenol 6.24 72

Specific activity of ['4C]benzene was 565 dpm/nmol. The peaknumbers describe the peaks observed in the radioactivity profile afterincubating liver microsomes with [14C]benzene (Fig. 3C). Results aremeans of two experiments, done in duplicate. Microsomes wereprepared from homogenates pooled from 11-12 animals.

The recovery studies indicated that the recoverableamount decreases as the concentration of muconaldehydeadded to the microsomal suspension decreases. For example,68% of 50 ,uM muconaldehyde incubated with microsomes (1mg of protein per ml) could be recovered after 10 min, butonly 20% of 5 uM muconaldehyde could be recovered at thattime.

Incubation of benzene in the presence of NADPH in asystem containing no microsomes or boiled microsomesresulted in no detectable formation of muconaldehyde, asdetermined from the radioactivity profile of the elutingmaterial. In the absence ofNADPH, no muconaldehyde wasdetected, nor was it detected in incubations carried out withmicrosomes obtained from animals that had not been treatedwith benzene. This latter result may very well be due to thelimits of detection by radioactivity in these studies. In thepresent studies, the radioactivity in the peak correspondingto muconaldehyde formed from benzene by microsomes frombenzene pretreated mice was 2-3 times above the back-ground.

DISCUSSIONThe present studies demonstrate that liver microsomes fromCD-1 mice catalyze benzene ring opening, resulting in theformation of a previously unidentified metabolite. This me-tabolite has been identified as muconaldehyde, on the basisof comparison with an authentic standard. Detection ofmuconaldehyde formed from benzene in a microsomal me-tabolizing system was accomplished with and without chem-ical derivatization, employing several methods of analysis.These studies utilized microsomes from CD-1 mice, a

strain sensitive to the direct hematotoxic and leukemogenicactions of benzene (20). The overall metabolism comparedwell with published data from liver microsomal systemsobtained from benzene-treated rats. In the present studies,0.25% of the incubated benzene was metabolized, comparedwith 0.4% in a study using microsomes from benzene-treatedSprague-Dawley rats (21). In the latter study highly polarwater-soluble metabolites, predominantly hydroquinone andcatechol, accounted for 18% of the total metabolites (21). Inthe present studies, hydroquinone and an unknown com-pound in the solvent front also accounted for 18% of the totalmetabolites. In rats, the ratio of phenol to highly polarwater-soluble metabolites was 5.3:1 (21), a ratio very similarto 5.4:1 observed in the present work.The amount of recoverable muconaldehyde formed in a

10-min incubation with benzene is approximately 0.8nmol/mg of protein (Table 1). This is 12.8% of the amount ofphenol formed in this system. In recent studies on benzenemetabolism in CD-1 mice administered [14C]benzene at 220mg/kg, trans,trans-muconic acid was found to be 14% of thetotal amount of phenol formed (7). As no muconic acid wasdetected in the urine of CD-1 mice administered phenol,catechol, or hydroquinone (7), it appears that a major

Medical Sciences: Latriano et al.

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

8360 Medical Sciences: Latriano et al.

portion, if not all, of urinary muconic acid is derived frommuconaldehyde, rather than through the action of intestinalbacteria on catechol.Although the absolute amounts of radioactivity in meta-

bolically derived muconaldehyde and its TBA adduct col-lected from the HPLC were low, in each case they were atleast four standard deviations above background (P < 0.001).As only a small amount ofbenzene was recovered in the urineas muconic acid, the low amount of muconaldehyde found isnot surprising in view of the extent to which benzene ringopening, as indicated by muconic acid formation, might occurin vivo, and in view of the reactive nature of muconaldehyde.The hematotoxicity of trans,trans-muconaldehyde (3, 10)

is not unexpected in view of the reported toxicity of othera43-unsaturated aldehydes to dividing cells. It has beenshown that this class of compounds reacts readily withnucleophiles, particularly sulfhydryl groups (22). The toxiceffects reported include inhibition of cellular metabolism aswell as inhibition of DNA, RNA, and protein synthesis.The mechanism for the formation of muconaldehyde from

benzene is not known. Fig. 5 shows a number ofpathways formuconaldehyde formation. The classical pathway ofbenzeneoxidation involving the initial formation of benzene oxidecould provide two routes for muconaldehyde formation, oneby way of ring opening of the dihydrodiol and one involvingrearrangement of the oxidized benzene oxide-oxepin (13).Mechanisms that depend on the direct interaction of activestates of oxygen with benzene may also be involved. Thushydroxyl radicals generated by pulse radiolysis (4, 23) or ina Fenton-type system (5) cause oxidative ring opening ofbenzene in aqueous solutions and have been implicated in themetabolism of benzene by rabbit microsomes in vitro (24).Singlet oxygen, an active oxygen species that has beensuggested to account for muconaldehyde formation frombenzene during photooxidation (25), is known to formdioxetanes that decompose to aldehydes by carbon-carbonbond cleavage (26).As in the case of benzene, hydroxyl radical-mediated ring

opening and fragmentations resulting in the formation ofdicarbonyl and aldehydic compounds have also been report-ed for toluene and xylene (27). Unsaturated dicarbonyls havealso been reported to be oxidation products of polycyclicaromatic hydrocarbons, such as naphthalene (28) andphenanthrene (29). Recent reports on the carcinogenicity of

Benzene oxi

eBenzene "

IFO | "OH

0

ide Dioxetane

0)-0-Hr~HlO-HH

Hydroperoxide

'b 0 0

-C=C-C =C-C-H

H H

transtrans-Muconaldehyde

0 H O

H-O-C-C=C -C=C- C-0-H

H H

transirrans-Muconic acid

FIG. 5. Potential pathways for the formation of muconaldehydefrom benzene. MFO, mixed-function oxidase; A02, singlet oxygen.

benzene indicate that it is a multipotent carcinogen, capableof producing leukemia and a variety of other neoplasias. Inthis study, similar tumors were observed in rats administeredxylene and ethyl benzene (30). The formation of muconal-dehyde, a known hematotoxin, in liver microsomes suggeststhat a quantitatively minor metabolic route may be animportant pathway leading to the formation of toxic metab-olites from benzene. The probable formation of hydroxylradicals during microsomal oxidation (31), and the propensityof hydroxyl radicals to open aromatic rings, suggest that thismay be a general mechanism for the production of toxicintermediates from aromatic compounds.

We thank Dr. R. Rosen, Center for Advanced Food Technology,Cook College, Rutgers University, for performing and interpretingthe mass spectral analysis and Ms. N. Lawrie for help in preparationof this manuscript. This work was supported by National Institutesof Health Grant ES02558.

1. Goldstein, B. D. (1977) J. Toxicol. Env. Health Suppl. 2, 69-105.2. Snyder, R., Kocsis, J. J. & Witmer, C. M. (1981) Rev. Biochem.

Toxicol. 3, 123-154.3. Goldstein, B. D., Witz, G., Javid, J., Amoruso, M., Rossman, T. &

Wolder, B. (1982) in Biological Reactive Intermediates, eds.Snyder, R., Parke, D. V., Kocsis, J. J., Jallow, D., Gibson, G. G.& Witmer, C. M. (Plenum, New York), Vol. 2, Part A, pp. 331-339.

4. Loeff, I. & Stein, G. (1959) Nature (London) 184, 901.5. Latriano, L., Zaccaria, A., Goldstein, B. D. & Witz, G. (1986) J.

Free Radicals Biol. Med. 1, 363-371.6. Parke, D. V. & Williams, R. T. (1953) Biochem. J. 54, 231-

238.7. Gad-El Karim, G. G., Ramanujum, S. & Legator, M. S. (1985)

Xenobiotica 15, 211-220.8. Itada, N. (1965) Biochem. Biophys. Res. Commun. 20, 149-

153.9. Lee, E. G., Laskin, J. D., Amoruso, M. A., Witz, G. & Goldstein,

B. D. (1983) Toxicologist 3, 93 (abstr.).10. Witz, G., Rao, G. & Goldstein, B. D. (1985) Toxicol. Appl.

Pharmacol. 80, 511-516.11. Kossmehl, G. & Bohn, B. (1974) Chem. Ber. 107, 710-711.12. Karrer, P., Eugster, C. H. & Perl, S. (1949) Helv. Chim. Acta 32,

1013-1015.13. Davies, S. G. & Whitham, G. H. (1977) J. Chem. Soc. Perkin

Trans. 1, 1346-1347.14. Nakajima, M., Tomida, I. & Takei, S. (1959) Chem. Ber. 92,

163-171.15. Snyder, R., Uzuki, F., Gonasun, L., Bromfeld, E. & Wells, A.

(1967) Toxicol. Appl. Pharmacol. 11, 346-360.16. Omura, T. & Sato, R. (1964) J. Biol. Chem. 239, 2370-2378.17. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.

(1951) J. Biol. Chem. 193, 265-275.18. Witz, G., Zaccaria, A., Lawrie, N. J., Ferran, H. J., Jr., &

Goldstein, B. D. (1985) Toxicologist 5, 230 (abstr.).19. Yu, L. W., Latriano, L., Duncan, S., Hartwick, R. A. & Witz, G.

(1986) Anal. Biochem. 156, 326-333.20. Snyder, C. A. & Goldstein, B. D. (1982) in Genotoxic Effects of

Airborne Agents, eds. Tice, R. R., Costa, D. L. & Schaich, K. M.(Plenum, New York), pp. 277-289.

21. Post, G. B. & Synder, R. (1983) J. Toxicol. Environ. Health. 11,619-630.

22. Schauenstein, E., Esterbauer, H. & Zollner, H. (1977) Aldehydes inBiological Systems (Metheun, New York), pp. 25-88.

23. Loeff, I. & Stein, G. (1963) J. Chem. Soc., 2623-2633.24. Johansson, I. & Ingelman-Sundberg, M. (1983) J. Biol. Chem. 258,

7311-7316.25. Wei, K. & Mani, J. C. (1967) J. Am. Chem. Soc. 89, 4225-4227.26. Foote, C. S. (1980) Molecular Basis of Environmental Toxicity

(Ann Arbor Science Publishers, Ann Arbor, MI), pp. 37-50.27. Shepson, P. B., Edney, E. 0. & Corse, E. W. (1984) J. Phys.

Chem. 88, 4122-4126.28. Smith, J. R. L., Shaw, B. A., Jeffrey, A. M. & Jerina, D. (1977) J.

Chem. Soc. Perkin Trans. 2, 1583-1589.29. Dowty, B. J., Brightwell, N. E., Laseter, J. L. & Griffin, G. W.

(1974) Biochem. Biophys. Res. Commun. 57, 452-456.30. Maltoni, C., Conti, B. & Cotti, G. (1983) Am. J. Ind. Med. 4,

589-630.31. Cohen, G. & Cederbaum, A. (1979) Science 204, 66-67.

Proc. NatL Acad. Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1