Vol. 57, No. 4APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 1991, p. 962-9680099-2240/91/040962-07$02.00/0Copyright ©D 1991, American Society for Microbiology

Reduction of Nitroaromatic Compounds by Anaerobic BacteriaIsolated from the Human Gastrointestinal Tract

FATEMEH RAFII,1 WIRT FRANKLIN,' ROBERT H. HEFLICH,2 AND CARL E. CERNIGLIAl*

Divisions of Microbiology' and Genetic Toxicology,2 National Center for Toxicological Research,Food and Drug Administration, Jefferson, Arkansas 72079

Received 8 November 1990/Accepted 24 January 1991

Human intestinal microbial flora were screened for their abilities to reduce nitroaromatic compounds bygrowing them on brain heart infusion agar plates containing 1-nitropyrene. Bacteria metabolizing 1-nitropy-rene, detected by the appearance of clear zones around the colonies, were identified as Clostridium leptum,Clostridium paraputrificum, Clostridium clostridiiforme, another Clostridium sp., and a Eubacterium sp. Thesebacteria produced aromatic amines from nitroaromatic compounds, as shown by thin-layer chromatography,high-pressure liquid chromatography, and biochemical tests. Incubation of three of these bacteria with 1-nitropyrene, 1,3-dinitropyrene, and 1,6-dinitropyrene inactivated the direct-acting mutagenicity associatedwith these compounds. Menadione and o-iodosobenzoic acid inhibited nitroreductase activity in all of theisolates, indicating the involvement of sulfhydryl groups in the active site of the enzyme. The optimum pH fornitroreductase activity was 8.0. Only the Clostridium sp. required added flavin adenine dinucleotide fornitroreductase activity. The nitroreductases were constitutive and extracellular. An activity stain for thedetection of nitroreductase on anaerobic native polyacrylamide gels was developed. This activity stain revealedonly one isozyme in each bacterium but showed that the nitroreductases from different bacteria had distinctelectrophoretic mobilities.

Nitrated polycyclic aromatic hydrocarbons (nitro-PAHs)are ubiquitous environmental contaminants that are formedby the nitration of PAHs from various combustion sources(34). These compounds have been detected in carbon blackphotocopier toners, urban-air particulates, diesel fuel emis-sions, used motor oils, barbecued food, and tea leaves (28,34, 35, 39). Nitro-PAHs, especially nitropyrenes, are potentmutagens in the Salmonella typhimurium test system and inmammalian cells. They have carcinogenic activity in labora-tory animals and may pose a human health risk (6, 7, 9-12,14-16, 20, 27, 34, 38, 39).Nitro-PAHs are metabolized by both reductive and oxida-

tive pathways (nitroreduction, ring oxidation, or a combina-tion of the two) (8, 9, 18, 29, 32, 33, 37). The reduction ofnitro groups is catalyzed by several mammalian enzymesand also by intestinal bacteria (4, 5, 14, 15, 17, 23, 24, 32).Exposure of humans to nitro-PAHs usually occurs via

inhalation of diesel engine exhaust particulates. These com-pounds can also reach the intestinal tract through ingestionof food. Ohnishi et al. (28) have measured the amounts of1-nitropyrene and 1,6-dinitropyrene in several grilled foodsand have estimated the exposure of individuals to nitroaro-matic compounds through inhalation and ingestion of foodduring a 20-year period.

Intestinal microflora, which are dominated by obligateanaerobes, can modify a wide range of environmental chem-icals either directly or through the enterohepatic circulation(17). Various anaerobic human, monkey, and rat intestinalmicroflora are capable of rapidly metabolizing 1-nitropyrene,dinitropyrenes, and 6-nitrochrysene to aromatic amines (4,5, 14, 17, 23, 24). In addition, anaerobic bacteria commonlyassociated with the gastrointestinal tract reduce nitro-PAHsto aromatic amines, indicating nitroreductase activity inthese bacteria (4, 15, 17, 18, 24). Kinouchi and Ohnishi (19)

* Corresponding author.

have studied the enzymes that convert 1-nitropyrene to1-aminopyrene in Bacteroidesfragilis and have purified oneof these enzymes.

In this study, we have developed a screening method todetect anaerobic bacteria capable of producing nitroreduc-tases and have shown that the isolated bacteria decrease themutagenicity of nitro-PAHs. We also have partially charac-terized the nitroreductase activities associated with thesebacteria.

MATERIALS AND METHODS

Chemicals. 4-Nitrobenzoic acid and 4-aminobenzoic acidwere obtained from Chem Service, Media, Pa. 1-Nitropy-rene, 1,3- and 1,6-dinitropyrene, 1-aminopyrene, and pyrenewere obtained from Chemsyn Science Laboratories, Le-nexa, Kans., and Aldrich Chemical Co., Milwaukee, Wis.1-Nitro[4,5,9,10-14C]pyrene (specific activity, 57.4 mCi/mmol; radiochemical purity, >99%) was obtained fromChemsyn. Tetracycline hydrochloride, menadione, o-io-dosobenzoic acid, flavin adenine dinucleotide (FAD), andNADH were purchased from Sigma Chemical Co., St.Louis, Mo. N-(1-Naphthyl)ethylenediamine dihydrochloride(NEDD) and high-pressure liquid chromatography(HPLC)-grade solvents were purchased from Fisher Chemical Co.,Pittsburgh, Pa. All other chemicals were of reagent gradeand of the highest purity available.Media and cultural conditions. Brain heart infusion (BHI)

agar from Difco Laboratories, Detroit, Mich., was preparedas previously reported (31) and used for the isolation anddetection of nitroreductase-producing bacteria from intesti-nal microbial flora. BHI broth, prereduced and anaerobicallysterilized (PRAS), from Carr-Scarborough Microbiologicals,Stone Mountain, Ga., was used for the growth and mainte-nance of nitroreductase-producing bacteria.

Isolation of nitroreductase-producing bacteria. Bacteriafrom human intestinal microflora were isolated on BHI agar

962

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC BACTERIA REDUCE NITROAROMATIC COMPOUNDS

plates containing 16 ,ug of 1-nitropyrene per ml, whichproduced a yellow background in the plates. Fecal materialfrom one individual was collected in a degassed containerand used as the source of the isolates. One gram of feces wasadded to 9 ml of BHI (PRAS) broth. Serial dilutions ofhuman feces in BHI (PRAS) broth were made, and 0.1 ml ofeach dilution was spread on each of the plates. The plateswere incubated at 37°C under both aerobic and anaerobicconditions. They were observed after 2 days for clearance of1-nitropyrene around bacterial colonies. Clostridium perfrin-gens ATCC 3626, from the American Type Culture Collec-tion, Rockville, Md., was used as a positive control fornitroreductase activity (4). Colonies surrounded by clearzones were transferred to BHI plates containing 1-nitropy-rene for single-colony isolation. The isolated colonies weretransferred to BHI (PRAS) broth for growth, maintenance,and identification.

Identification of nitroreductase-producing anaerobic bacte-ria. The bacteria capable of reducing 1-nitropyrene wereidentified by colony morphology, Gram stain, biochemicaltests, volatile and nonvolatile fatty acids, and protein anal-ysis according to the methods described elsewhere (13, 21,26, 36).

Assay of enzyme activity. Nitroreductase activity wasassayed by adding 4-nitrobenzoic acid to an enzyme-contain-ing solution and measuring the amount of 4-aminobenzoicacid produced (40). 4-Nitrobenzoic acid was added to tubescontaining combined supernatants and cell extracts fromovernight cultures of each bacterial isolate. NoninoculatedBHI (PRAS) broth was used as a control. The tubes wereincubated at 37°C for 1 h and then acidified with 0.21% (finalconcentration) trichloroacetic acid. Sodium nitrite (final con-centration, 0.007%) was added to the mixtures to form adiazonium salt with 4-aminobenzoic acid. The mixtures wereincubated at room temperature for 20 min. Next, ammoniumsulfamate (final concentration, 0.04%) was added to neutral-ize the sodium nitrite. Following incubation for 3 min atroom temperature, NEDD (final concentration, 0.35%) wasadded. The diazonium salt formed from 4-aminobenzoic acidcouples with NEDD in acidic solutions and produces ared-purple azo dye. The A540 was measured with a BeckmanDU-7 spectrophotometer. The amount of 4-aminobenzoicacid produced was calculated from a standard curve.One unit of enzyme was defined as the amount of enzyme

necessary to produce 1 ,ug of 4-aminobenzoic acid at 37°Cunder anaerobic conditions in 1 h. The specific activity of theenzyme was measured by calculating the units of enzymaticactivity per milligram of total soluble protein in supernatantsand cell extracts.

Detection of arylamines in bacterial colonies followinggrowth with nitroaromatic compounds. 1-Nitropyrene or 6-ni-trochrysene, dissolved in dimethyl sulfoxide, was added tothe BHI agar at a concentration of 16 to 32 ,ug/ml of themedium before incubation with bacteria. The cultures wereincubated in the dark at 37°C for 48 h and were tested for thepresence of arylamine. Thirteen milliliters of a solutioncontaining 0.21% trichloroacetic acid and 0.007% sodiumnitrite was added to each 9-cm-diameter culture plate. Theplates were incubated at room temperature for 20 min. Onemilliliter of 0.5% ammonium sulfamate was added to theplates, which were incubated for another 3 min, and then 5ml of 0.1% NEDD in distilled water was added to each plate.The plates were incubated at room temperature and ob-served for the appearance of a red-purple color in thecolonies.

Analyses of 1-nitropyrene metabolites by TLC and HPLC.

1-Nitropyrene (final concentration, 16 ,ug/ml) was added toBHI (PRAS) broth and inoculated with 0.2 ml of an over-night bacterial culture. After incubation, the cells werebroken by sonic oscillation and the cultures were extractedthree times with equal volumes of ethyl acetate. The com-bined extracts were dried by rotary evaporation underreduced pressure and dissolved in methanol for thin-layerchromatography (TLC) and HPLC.The ethyl acetate extracts from the culture media were

spotted onto a TLC plate (silica gel, 0.2-mm thickness; E.Merck AG, Darmstadt, Germany). 1-Aminopyrene wasspotted on the plate as a standard, and the plates weredeveloped in ethyl acetate-benzene (1:9 [vol/vol]). Theplates were then observed for 1-aminopyrene spots with UVlight.For study of metabolites by HPLC, 3 VjCi of4C-labeled

1-nitropyrene and 8 ,ug of nonlabeled 1-nitropyrene per mlwere added to BHI (PRAS) broth and the medium wasinoculated with the bacterial isolates. The cultures wereincubated at 37°C for 3 days, sonicated for 12 min (foursonications of 3 min each), and extracted with ethyl acetate.The extracts were dried under anhydrous sodium sulfate,and the residues were dissolved in methanol. The samplesthen were analyzed by HPLC, using a Beckman model 100Aliquid chromatograph with a diode array detector as previ-ously described (14). One-minute fractions were collected,and the radioactivity present in each fraction was measured(14, 24, 32). Authentic 1-aminopyrene and 1-nitropyrenewere used as chromatographic markers.

Mutagenicity assay. 1-Nitropyrene (final concentration, 8,ug/ml) and 1,3-dinitropyrene and 1,6-dinitropyrene (finalconcentrations, 1 jig/ml) were added to separate tubescontaining BHI (PRAS) medium. The media were inoculatedand incubated for 0 to 5 days at 37°C. The metabolites wereextracted with ethyl acetate and concentrated as describedabove. Tubes of uninoculated BHI (PRAS) medium contain-ing nitro-PAHs were used as controls. Nitro-PAHs were alsoadded to ethyl acetate extracts of bacterial cultures and usedin the mutagenicity assay as controls. Ethyl acetate-ex-tracted metabolites were dried, dissolved in dimethyl sulfox-ide, and tested for mutagenicity by using S. typhimuriumTA98 (5).

Effect of pH on nitroreductase activity. Tubes containing 10ml of BHI broth were adjusted to pHs from 4.5 to 9.5 withHCl and NaOH and were inoculated with 0.2 ml of anovernight culture of each isolate. 4-Nitrobenzoic acid (0.29mg/ml) was added to all of the cultures. Nitroreductaseactivity was assayed by testing for the formation of 4-ami-nobenzoic acid with NEDD as described elsewhere (40).

Protein determination and cell count. After sonication andcentrifugation, the soluble protein in the combined superna-tant and cell extract from each isolate was measured (22).The amount of protein detected in noninoculated BHI(PRAS) broth was subtracted from the total amount ofprotein. Bacterial cells were enumerated with a Petroff-Hausser bacteria counter.

Cofactor requirement. Two tubes of M10 medium pre-pared by the method of Caldwell and Bryant (3) wereinoculated with each isolate, and FAD (final concentration,50 ,ug/ml) was added to one tube. 4-Nitrobenzoic acid (finalconcentration, 0.29 mg/ml) was added to both tubes. Thecultures were incubated for 2 days and assayed for nitrore-ductase activity by testing for the production of 4-aminoben-zoic acid from 4-nitrobenzoic acid.

Effects of inhibitors on nitroreductase activity. Overnightcultures of bacteria in BHI (PRAS) broth were sonicated

VOL. 57, 1991 963

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

with an ultrasonic cleaner for a total of 12 min and centri-fuged at 5,000 x g for 10 min. Tetracycline hydrochloride(final concentration, 15 ,ug/ml) and menadione or o-io-dosobenzoic acid (final concentration, 0.32 mg/ml) were

added to the combined supernatant and cell extract. Cultureswithout inhibitors were used as controls. Replicate culturetubes were incubated for 15 min at 37°C, and then 4-ni-trobenzoic acid was added to the tubes with or withoutinhibitors (final concentration, 0.29 mg/ml). The cultureswere incubated at 37°C for 1 h and assayed for the formationof 4-aminobenzoic acid.

Nitroreductase activity assay on native polyacrylamide gels.A nondenaturing anaerobic gel was prepared and run as

previously described (30). The gel was then incubated, withoccasional shaking, anaerobically for 1.5 h at 37°C in a

Ziploc bag containing 30 ml of degassed M10 mediumsupplemented with 0.24 mg of 4-nitrobenzoic acid per ml andwith FAD and NADH (final concentration of each, 50,ug/ml). The solution was decanted, and the gel was incu-bated for another hour. The gel then was incubated for 20min at room temperature in a solution containing 0.21%trichloroacetic acid and 0.007% sodium nitrite. Ammoniumsulfamate (final concentration, 0.004%) was then added tothe gel. After 3 min, NEDD (final concentration, 0.027%)was added to the bag to develop the nitroreductase bands.The gel was gently shaken at room temperature in the baguntil the bands of red-purple color, indicating 4-aminoben-zoic acid, had developed on the gel.

RESULTS

Isolation and identification of nitroreductase-producing bac-teria. Following incubation of bacteria from human intestinalmicroflora on the medium containing 1-nitropyrene, manydifferent bacterial colonies developed on the plates underboth anaerobic and aerobic conditions. There were no clearzones on the plates that were incubated aerobically, but on

plates incubated under anaerobic conditions, some colonieswere surrounded by clear zones on the yellow background ofthe plates. These colonies were transferred anaerobically toplates containing 1-nitropyrene for single-colony isolation.The 1-nitropyrene was utilized, and clear zones surroundingthe bacteria developed on the plates (Fig. 1).The bacterial isolates were identified as a Clostridium sp.,

Clostridium leptum, Clostridium paraputrificum, Clostrid-ium clostridiiforme, and a Eubacterium sp. One of theisolates, designated NP4, was not identified. Table 1 lists theanaerobic bacteria and the amount of 4-nitrobenzoic acidutilized per milligram of protein, as calculated from dupli-cates of the combined supernatants and cell extracts fromeach bacterium.The bacteria listed in Table 1 also metabolized 6-nitro-

chrysene, 1-amino-7-nitrofluorene, 4-nitrobenzoic acid, and1,3- and 1,6-dinitropyrene, as was shown by the develop-ment of clear zones, amine production, and mutagenicitytests (data not shown).

Identification of metabolites. HPLC, TLC, and biochemi-cal tests were used for identification of the metabolitesproduced by the isolates in BHI (PRAS) broth containing1-nitropyrene. HPLC analysis of the ethyl acetate extractsfrom C. perfringens and five of the isolates revealed severalUV-absorbing peaks. The compounds eluting from 2 to 16min were derived from the culture medium and not from1-nitropyrene (Fig. 2). This was determined by experimentswith 14C-labeled 1-nitropyrene that revealed only two peaksof radioactivity (data not shown). The retention time and UV

FIG. 1. Clearance of 1-nitropyrene around colonies of bacteriathat had been isolated from human intestinal microflora and showednitroreductase activity.

spectrum of the peak eluting at 17 min were identical to thoseof authentic 1-aminopyrene (Fig. 2). The peak eluting at 31min corresponded to 1-nitropyrene (Fig. 2).On TLC plates, the major metabolite from each of the

isolates migrated with authentic 1-aminopyrene and wasfluorescent, as was the authentic 1-aminopyrene standard.We concluded that 1-nitropyrene was reduced by nitroreduc-tases from all of the different bacteria to 1-aminopyrene.The ability of the isolates to convert other nitroaromatic

compounds to amines was also tested. After the addition of4-nitrobenzoic acid to broth cultures or after growth ofbacterial isolates on agar plates containing 1-nitropyrene or6-nitrochrysene, the arylamine in the bacterial cultures wasdetected by the addition of NEDD. The purple azo dye wasproduced after the reduction of 4-nitrobenzoic acid in brothcultures or after the reduction of 1-nitropyrene or 6-nitro-chrysene in bacterial colonies. There was no color develop-ment in bacterial isolates grown without nitroaromatic com-pounds or in bacteria that did not reduce these nitroaromaticcompounds.

Mutagenicity tests. The S. typhimurium reversion assayusing strain TA89 was used to determine the direct-actingmutagenic activity of the ethyl acetate-extracted metabolitesof bacteria incubated with 1-nitropyrene, 1,3-dinitropyrene,

TABLE 1. Nitroreductase activity of anaerobic bacteriaisolated from human feces

Reduction of 4-nitrobenzoicacid (,ug/mg of protein/h)'

C. leptum ................................ 20.3 ± 8.0Eubacterium sp ................................ 17.8 ± 2.4C. clostridiiforme ................................ NDC. perfringens ................................ 14.0 ± 1.1C. paraputrificum ................................ 14.6 ± 4.0Clostridium sp ................................ 30.4 ± 8.8

" Mean ± standard deviation. ND, Not done.

964 RAFII ET AL.

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC BACTERIA REDUCE NITROAROMATIC COMPOUNDS

NH2

- N2

ilb 15 20 5 3g0Time (min.)

TABLE 3. Nitroreductase activity of proteins from anaerobicbacteria isolated from human gastrointestinal tract

U' of enzymeBacterial strain

Per mg of protein Per 1010 bacteria

C. leptum 16.7 ± 6.6 86.4 ± 12.4Eiubacteriium sp. 14.6 ± 1.9 188.7 ± 17.6C. clostridiiforme 16.8 ± 2.7 11.0 ± 4.7C. perfringens 11.5 + 1.0 147.3 ± 79.0C. paraputrificum 12.0 ± 3.2 85.6 + 11.6Clostridium sp. 25.0 + 7.1 80.0 + 2.0

" One unit of enzyme activity was defined as the amount of enzymenecessary to produce 1 ,ug of 4-aminobenzoic acid at 37°C in 1 h underanaerobic conditions.

35 40 45

Time (min.)

FIG. 2. Elution profile of 1-aminopyrene and 1-nitropyrene on

reversed-phase HPLC. (A) Extract of Clostridium sp. metabolitesafter incubation with 1-nitropyrene. (B) Standard mixture of 1-ami-nopyrene and 1-nitropyrene. (Inset) UV spectrum of 1-aminopy-rene.

and 1,6-dinitropyrene. Mutagenic activity decreased duringgrowth of the bacteria with all three compounds (Table 2).The decrease in mutagenicity was consistent with the rela-tive mutagenic activities of nitro- and amino-PAHs.

Effect on pH on growth and nitroreductase activity. C.perfringens grew and produced an active enzyme in BHIbroth media between pHs 5.5 and 9.0. The other isolates

grew at neutral to slightly alkaline pHs. The enzyme wasactive at all the pHs at which the bacteria grew, but theoptimum pH for activity was 8.0.

Nitroreductase characteristics in bacterial isolates. Thesupernatants of overnight cultures of bacteria incubated inthe absence of 4-nitrobenzoic acid were tested for nitrore-ductase activity. The conversion of 4-nitrobenzoic acid to4-aminobenzoic acid under anaerobic conditions occurred inall of the cultures tested. The exclusion of oxygen wasrequired in this procedure. These results indicate that theenzyme was constitutive and extracellular and requiredanaerobiosis for activity. Only the enzyme from one of theisolates (Clostridium sp.) required added FAD for activity.Comparison of enzyme activity in different isolates. The

amounts of enzyme produced by different bacterial isolateswere compared by measuring the units of enzyme producedper milligram of soluble protein and the numbers of cells thatproduced 1 U of enzyme in two replicate cultures (Table 3).There were differences in the amount of enzyme producedby each bacterium and in the enzymatic activity per milli-gram of protein produced by each bacterium. The Elubacte-rium sp. produced the highest amount of enzyme and C.clostridiiforine produced the least.

Effect of inhibitors on enzymatic activity of nitroreductase.The effects of sulfhydryl-modifying chemicals on nitroreduc-tase activity are shown in Table 4. Menadione and o-io-dosobenzoic acid inhibited the enzymatic activity of nitrore-ductases from all of the isolates. The percentage of inhibitionwith menadione varied from 38.4% for the Eubacterium sp.to 77.4% for isolate NP4. With o-iodosobenzoic acid, inhi-bition varied from 58.4% for the Eubacterium sp. to 77.0%for Clostridium sp.Comparison of nitroreductases from different isolates on

native gels. Ammonium sulfate-fractionated crude extracts

TABLE 2. Effect of nitroreductase-producing bacterial isolates on the mutagenicity of different nitro-PAHsafter incubation with these compounds

b~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Mutagenic activity'

No. ofChemx cal bacteria in C. clostridiiforme Isolate NP4 C. leptirm(amt of extract equivalent) control'

Day 0 Day 5 Day 0 Day 5 Day 0 Day S

1-Nitropyrene' (2 ,ug) 6,673 ± 64 4,528 + 179 34 ± 2 4,299 ± 680 35 ± 6 4,684 ± 801 395 ± 431,3-Dinitropyrene (2.5 ng) 2,457 + 104 1,949 ± 422 16 ± 4 1,934 ± 15 8 ± 6 2,218 + 127 15 ± 31,6-Dinitropyrene (2.5 ng) 1,368 ± 159 1,368 ± 159 10 ± 10 1,408 ± 259 19 ± 2 1,362 ± 212 9 ± 4

"Uninoculated BHI broth and nitro-PAH.b Number of S. typhimurium TA98 revertants per piate after treatment with ethyl acetate-extractable metabolites of bacteria incubated for 0 or 5 days with

1-nitropyrene or with 1,3- or 1,6-dinitropyrene.' Incubation time was 3 days.

A

-gD

,IntD4:)

EcstN

0

n

VOL. 57, 1991 965

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 4. Effects of menadione and o-iodosobenzoic acid onnitroreductase activity

% Inhibition by:Bacterial strain

Menadione o-lodosobenzoic acid

Clostridium sp. 69.8 77.0C. leptum 71.8 74.0C. perfringens 92.3 77.6Eubacterium sp. 38.4 58.3Isolate NP4 77.2 72.5

from cultures grown in BHI (PRAS) broth were used for thedetection of enzymes from three strains on the gels. Theenzymes from different isolates were detected by usingnative gels under anaerobic conditions. Enzyme bands werered-purple on the gels where 4-nitrobenzoic acid had beenreduced to 4-aminobenzoic acid and converted to azo dye byNEDD. Five units of enzyme was sufficient to produce ared-purple band on the gels. There was only one enzymeband in each lane, indicating that there was only onenitroreductase isozyme in each bacterium. The electropho-retic mobilities of the enzyme bands were slightly differentfor each strain, indicating that different forms of nitroreduc-tase were produced by these strains (Fig. 3).

DISCUSSION

The abilities of pure cultures of bacteria and of intestinalmicroflora from humans and several animal species to re-duce nitro-PAHs to amino-PAHs has been shown previously(4, 5, 14, 17, 23, 24). However, the specific bacteria in thehuman gastrointestinal tract that reduce these environmentalpollutants have not been previously identified. In this study,a simple method for rapid detection and isolation of bacteriacapable of reduction of nitro-PAHs has been described.Although we used feces from only one individual for

isolation of these microorganisms, other bacteria probably

1

FIG. 3. Activity staining of nitroreductase on native polyacryl-amide gel. Crude extracts were used for the detection of enzymes.Lane 1, C. paraputrificum; lane 2, a Eubacterium sp.; lane 3, C.perfringens. Arrows show the nitroreductase bands.

would be detected by using this method with other sourcesof isolates. Four of the bacterial strains isolated belonged tothe genus Clostridium. Some of these bacteria were notpreviously known to produce nitroreductases. HPLC, TLC,and biochemical tests revealed that aromatic amines werethe major metabolites produced from nitro compounds bythese bacteria. The presence of 1-aminopyrene, 6-ami-nochrysene, and 4-aminobenzoic acid as major metabolitesfrom the corresponding nitroaromatic compounds indicatesthat the first reaction in the bioconversion of nitroaromaticcompounds by intestinal bacteria is accomplished by anitroreductase .The Ames (S. typhimurium) assay was used to monitor the

effect of biotransformation by three of the isolates on themutagenicity of 1-nitropyrene and of 1,3- and 1,6-dinitropy-rene. The three isolates were capable of metabolizing allthree nitro-PAHs to products that were considerably lessmutagenic. Kinouchi and Ohnishi (19) attributed the de-crease in the mutagenicity of 1-nitropyrene after incubationwith B. fragilis to the rapid metabolism of 1-nitropyrene to1-aminopyrene, which has little direct-acting mutagenicity.The isolates used in the present study virtually eliminated

the mutagenicity of 1-nitropyrene and of 1,3- and 1,6-dinitropyrene. No traces of active N-hydroxylamines weredetected in the metabolites of these bacteria after exposureto nitro-PAHs (HPLC results showed only two peaks ofradioactivity). Thus these organisms could be useful foranaerobic detoxification of nitro-PAHs under appropriateconditions. Previously it was shown that intestinal micro-flora from different animal species produce compounds from1,8-dinitropyrene that are less mutagenic than the parentcompound (5). Moller et al. (25) administered 2-nitrofluoreneto germfree and conventional rats and found that a smalleramount of mutagenic metabolites of this environmentalcontaminant was excreted in the urine and feces of conven-tional rats than of germfree rats. Their study clearly demon-strates the importance of intestinal microflora in the metab-olism of nitro-PAHs in vivo. In our study, the specificorganisms contributing to detoxification have been identi-fied.

Nitroreductases from these organisms seemed to have awide spectrum of substrates, as shown by the reduction ofseveral nitroaromatic compounds, but the efficiency of re-duction varied with different substrates (data not shown).The bacterial isolates grew in a narrow pH range under

anaerobic conditions and produced active enzymes with apH optimum of 8.0. Arora et al. (1) noted similarly that theoptimum pH for nitroreductase activity of Vibrio choleraewas 8.0 in phosphate buffer.Our isolates grew and produced the enzyme in a low-

nutrient broth medium (M1O medium). The activity waslower in M10 medium than in the rich BHI (PRAS) medium(data not shown), probably because of differences in growthrates. FAD was necessary for nitroreductase activity only inthe Clostridium sp. Similarly, nitroreductase activity of B.fragilis is markedly enhanced by the addition of flavin (19). Asmall amount of riboflavin (0.01 ,ug/ml), present in the yeastextract in M10 medium, conceivably may have satisfied aflavin requirement for the nitroreductases from the otherbacterial isolates.The enzymes from these isolates were inhibited by both

menadione and o-iodosobenzoic acid, which bind to sulfhy-dryl groups. This indicates that sulfhydryl groups are essen-tial for the catabolic function of the enzyme and suggeststhat there may be cysteine residues in the active site.Kinouchi and Ohnishi (19) also noted the inhibition of

966 RAFII ET AL.

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC BACTERIA REDUCE NITROAROMATIC COMPOUNDS

nitroreductases from B. fragilis by menadione and o-io-dosobenzoic acid.A simplified activity assay for direct detection of nitrore-

ductase activity on native polyacrylamide gels after electro-phoresis was developed by using 4-nitrobenzoic acid. Thiscompound is a model substrate for nitroreductase (29).Previously, the detection of nitroreductase in Escherichiacoli on gels was accomplished by slicing the gel and assayingfor reductase activity (2). The method described here detectsstructural differences among nitroreductases from differentbacteria by using crude extracts and can be used for purifi-cation of enzymes in preparative gels. The need for purifi-cation of highly oxygen-sensitive nitroreductase enzymes byconventional methods for this purpose thus has been elimi-nated.The presence of different nitroreductases in different iso-

lates was indicated by the distinct electrophoretic mobilitiesof the enzyme in an activity assay on native polyacrylamidegels. We detected only one nitroreductase isozyme in eachbacterium. Kinouchi and Ohnishi (19) detected four differentnitroreductase isozymes (nitroreductases I to IV) from B.fragilis. If more than one isozyme of nitroreductase was

present in our isolates, they either migrated together on thenative gels or could not be detected by this method.

ACKNOWLEDGMENTS

Special thanks are due to Tangelyn E. Black and Allison L. Selbyfor technical assistance and to John B. Sutherland, K. BarryDelclos, Ming-Hsiung Lu, and Frederick A. Beland for helpfulcomments.

REFERENCES1. Arora, K. L., C. R. Krishna Murti, and D. L. Shrivastava. 1959.

Studies in the enzyme make-up of Vibrio cholerae. Part XIV.Organic nitroreductase. J. Sci. Ind. Res. 18C:105-109.

2. Bryant, D. W., D. R. McCalla, M. Leeksma, and P. Laneuville.1980. Type I nitroreductases of Escherichia coli. Can. J. Micro-biol. 27:81-86.

3. Caldwell, D. R., and M. P. Bryant. 1966. Medium without rumenfluid for nonselective enumeration and isolation of rumen bac-teria. Appl. Microbiol. 14:794-801.

4. Cerniglia, C. E. 1985. Metabolism of 1-nitropyrene and 6-ni-trobenzo[a]pyrene by intestinal microflora. Prog. Clin. Biol.Res. 181:133-137.

5. Cerniglia, C. E., K. J. Lambert, G. L. White, R. H. Heflich, W.Franklin, E. K. Fifer, and F. A. Beland. 1988. Metabolism of1,8-dinitropyrene by human, rhesus monkey, and rat intestinalmicroflora. Toxicity Assessment 3:147-159.

6. Delclos, K. B., R. P. Walker, K. L. Dooley, P. P. Fu, and F. F.Kadlubar. 1987. Carcinogen-DNA adduct formation in the lungsand livers of preweanling CD-1 male mice following administra-tion of [3H]-6-nitrochrysene, [3H]-6-aminochrysene, and [3H]-1,6-dinitropyrene. Cancer Res. 47:6272-6277.

7. Eddy, E. P., P. C. Howard, G. D. McCoy, and H. S. Rosenkranz.1987. Mutagenicity, unscheduled DNA synthesis, and metabo-lism of 1-nitropyrene in the human hepatoma cell line HepG2.Cancer Res. 47:3163-3168.

8. Foster, G. D., and D. G. Crosby. 1987. Comparative metabolismof nitroaromatic compounds in freshwater, brackish water andmarine decapod crustaceans. Xenobiotica 17:1393-1404.

9. Fu, P. P. 1990. Metabolism of nitro-polycyclic aromatic hydro-carbons. Drug Metab. Rev. 22:209-268.

10. Fu, P. P., C. E. Cerniglia, K. E. Richardson, and R. H. Heflich.1988. Nitroreduction of 6-nitrobenzo[a]pyrene: a potential acti-vation pathway in humans. Mutat. Res. 209:123-129.

11. Heflich, R. H., Z. Djuric, E. K. Fifer, C. E. Cerniglia, and F. A.Beland. 1986. Metabolism of dinitropyrenes to DNA-bindingderivatives in vitro and in vivo. Dev. Toxicol. Environ. Sci.13:185-197.

12. Heflich, R. H., P. C. Howard, and F. A. Beland. 1985. 1-Ni-trosopyrene: an intermediate in the metabolic activation of1-nitropyrene to a mutagen in Salmonella typhimurium TA1538. Mutat. Res. 149:25-32.

13. Holdeman, L. V., E. P. Cato, and W. E. C. Moore (ed.). 1977.Anaerobe laboratory manual, 4th ed. Virginia Polytechnic In-stitute and State University, Blacksburg.

14. Howard, P. C., F. A. Beland, and C. E. Cerniglia. 1983.Reduction of the carcinogen 1-nitropyrene to 1-aminopyrene byrat intestinal bacteria. Carcinogenesis 4:985-990.

15. Howard, P. C., R. H. Heflich, F. E. Evans, and F. A. Beland.1983. Formation of DNA adducts in vitro and in Salmonellatyphimurium upon metabolic reduction of the environmentalmutagen 1-nitropyrene. Cancer Res. 43:2052-2058.

16. King, C. M., L. K. Tay, M. S. Lee, K. Imaida, and C. Y. Wang.1986. Mechanisms of tumor induction by dinitropyrenes in thefemale CD rat. Dev. Toxicol. Environ. Sci. 13:279-290.

17. Kinouchi, T., K. Nishifuji, and Y. Ohnishi. 1987. In vitrointestinal microflora-mediated metabolism of biliary metabolitesfrom 1-nitropyrene-treated rats. Microbiol. Immunol. 31:1145-1159.

18. Kinouchi, T., K. Nishifuji, and Y. Ohnishi. 1990. Biliary excre-tion of glutathione conjugates of 4,5-epoxy-4,5-dihydro-1-nitro-pyrene and 9,10-epoxy-9,10-dihydro-1-nitropyrene in rats ad-ministered 1-nitropyrene orally and their further metabolism inthe intestinal tract. Carcinogenesis 11:1381-1387.

19. Kinouchi, T., and Y. Ohnishi. 1983. Purification and character-ization of 1-nitropyrene nitroreductases from Bacteroides fragi-lis. Appl. Environ. Microbiol. 46:596-694.

20. Kinouchi, T., and Y. Ohnishi. 1986. Metabolic activation of1-nitropyrene and 1,6-dinitropyrene by nitroreductases fromBacteroidesfragilis and distribution of nitroreductase activity inrats. Microbiol. Immunol. 30:979-992.

21. Krieg, N. R., and J. G. Holt (ed). 1984. Bergey's manual ofsystematic bacteriology, vol. 1. The Williams & Wilkins Co.,Baltimore.

22. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

23. Manning, B. W., W. L. Campbell, W. Franklin, K. B. Delclos,and C. E. Cerniglia. 1988. Metabolism of 6-nitrochrysene byintestinal microflora. Appl. Environ. Microbiol. 54:197-203.

24. Manning, B. W., C. E. Cerniglia, and T. W. Federle. 1986.Biotransformation of 1-nitropyrene to 1-aminopyrene andN-formyl-1-aminopyrene by the human intestinal microbiota. J.Toxicol. Environ. Health 18:339-346.

25. Moller, L., M. Corrie, T. Midtvedt, J. Rafter, and J.-A. Gus-tafsson. 1988. The role of the intestinal microflora in the forma-tion of mutagenic metabolites from the carcinogenic air pollut-ant 2-nitrofluorene. Carcinogenesis 9:823-830.

26. Moore, W. E. C., D. E. Hash, L. V. Holdeman, and E. P. Cato.1980. Polyacrylamide slab gel electrophoresis of soluble pro-teins for studies of bacterial floras. Appl. Environ. Microbiol.39:900-907.

27. Odagiri, Y., S. Adachi, H. Katayama, H. Matsushita, and K.Takemoto. 1986. Carcinogenic effects of a mixture of nitropy-renes in F344 rats following its repeated oral administrations.Dev. Toxicol. Environ. Sci. 13:291-307.

28. Ohnishi, Y., T. Kinouchi, H. Tsutsu, and M. Uejima. 1986.Mutagenic nitropyrenes in foods, p. 107-118. In Y. Hayashi etal. (ed.), Diet, nutrition and cancer. Japan Science SocietyPress, Tokyo.

29. Orna, M. V., and R. P. Mason. 1989. Correlation of kineticparameters of nitroreductase enzymes with redox properties ofnitroaromatic compounds. J. Biol. Chem. 264:12379-12384.

30. Rafii, F., and C. E. Cerniglia. 1990. An anaerobic nondenaturinggel assay for the detection of azoreductase from anaerobicbacteria. J. Microbiol. Methods 12:139-148.

31. Rafii, F., W. Franklin, and C. E. Cerniglia. 1990. Azoreductaseactivity of anaerobic bacteria isolated from human intestinalmicroflora. Appl. Environ. Microbiol. 56:2146-2151.

32. Richardson, K. E., P. P. Fu, and C. E. Cerniglia. 1988. Metab-olism of 1-, 3-, and 6-nitrobenzo[a]pyrene by intestinal micro-

967VOL. 57, 1991

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

flora. J. Toxicol. Environ. Health. 23:527-537.33. Rickert, D. E. 1987. Metabolism of nitroaromatic compounds.

Drug Metab. Rev. 18:23-53.34. Rosenkranz, H. S., and R. Mermelstein. 1983. Mutagenicity and

genotoxicity of nitropyrenes. All nitro-containing chemicalswere not created equal. Mutat. Res. 114:217-267.

35. Rosenkranz, H. S., and D. R. Sanders. 1980. Nitropyrenes:isolation, identification, and reduction of mutagenic impuritiesin carbon black toners. Science 209:1039-1043.

36. Sneath, P. H. A., N. S. Mair, M. E. Sharpe, and J. G. Holt (ed).1986. Bergey's manual of systematic bacteriology, vol. 2. TheWilliams & Wilkins Co., Baltimore.

37. Tatsumi, K., and H. Amano. 1987. Biotransformation of 1-nitro-pyrene and 2-nitropyrene to novel metabolites, the correspond-

ing formylamino compounds, in animal bodies. Biochem.Biophys. Res. Commun. 142:376-382.

38. Tokiwa, H., R. Nakagawa, K. Horikawa, and A. Ohkubo. 1987.The nature of the mutagenicity and carcinogenicity of nitratedaromatic compounds in the environment. Environ. Health Per-spect. 73:191-199.

39. Tokiwa, H., T. Otofuji, R. Nakagawa, K. Horikawa, T. Maeda,N. Sano, K. Izumi, and H. Otsuka. 1986. Dinitro derivatives ofpyrene and fluoranthene in diesel emission particulates and theirtumorigenicity in mice and rats. Dev. Toxicol. Environ. Sci.13:253-270.

40. Zachariah, P. K., and M. R. Juchau. 1974. The role of gut florain the reduction of aromatic nitro-groups. Drug Metab. Dispos.2:74-78.

968 RAFII ET AL.

on February 6, 2020 by guest

http://aem.asm

.org/D

ownloaded from