Characterization of Novel Carbazole Catabolism Genes from Gram-Positive Carbazole Degrader Nocardioides aromaticivorans IC177

Abstract

Nocardioides aromaticivorans IC177 is a gram-positive carbazole degrader. The genes encoding carbazole degradation (car genes) were cloned into a cosmid clone and sequenced partially to reveal 19 open reading frames. The car genes were clustered into the carAaCBaBbAcAd and carDFE gene clusters, encoding the enzymes responsible for the degradation of carbazole to anthranilate and 2-hydroxypenta-2,4-dienoate and of 2-hydroxypenta-2,4-dienoate to pyruvic acid and acetyl coenzyme A, respectively. The conserved amino acid motifs proposed to bind the Rieske-type [2Fe-2S] cluster and mononuclear iron, the Rieske-type [2Fe-2S] cluster, and flavin adenine dinucleotide were found in the deduced amino acid sequences of carAa, carAc, and carAd, respectively, which showed similarities with CarAa from Sphingomonas sp. strain KA1 (49% identity), CarAc from Pseudomonas resinovorans CA10 (31% identity), and AhdA4 from Sphingomonas sp. strain P2 (37% identity), respectively. Escherichia coli cells expressing CarAaAcAd exhibited major carbazole 1,9a-dioxygenase (CARDO) activity. These data showed that the IC177 CARDO is classified into class IIB, while gram-negative CARDOs are classified into class III or IIA, indicating that the respective CARDOs have diverse types of electron transfer components and high similarities of the terminal oxygenase. Reverse transcription-PCR (RT-PCR) experiments showed that the carAaCBaBbAcAd and carDFE gene clusters are operonic. The results of quantitative RT-PCR experiments indicated that transcription of both operons is induced by carbazole or its metabolite, whereas anthranilate is not an inducer. Biotransformation analysis showed that the IC177 CARDO exhibits significant activities for naphthalene, carbazole, and dibenzo-p-dioxin but less activity for dibenzofuran and biphenyl.

Carbazole is an N-heterocyclic aromatic compound derived from creosote, crude oil, and shale oil (32, 33). Carbazole is known to be both mutagenic and toxic (1) but has been used as an industrial raw material for the production of dyes, medicines, and plastics (3). Various carbazole-degrading bacteria have been reported in studies examining the mechanisms of biodegradation of this recalcitrant compound and the application of bacteria in bioremediation (39). The carbazole biodegradation pathway of Pseudomonas resinovorans CA10 has been elucidated (41), and carbazole catabolic genes (car genes) from this organism have been studied extensively. In CA10, the carbazole 1,9a-dioxygenase (CARDO; encoded by carAa, carAc, and carAd), the meta-cleavage enzyme (encoded by carBa and carBb), and the meta-cleavage compound hydrolase (encoded by carC) together are involved in the conversion of carbazole to anthranilate and 2-hydroxypenta-2,4-dienoate (Fig. 1) (47, 48). 2-Hydroxypenta-2,4-dienoate is converted to pyruvic acid and acetyl coenzyme A by 2-hydroxypenta-2,4-dienoate hydratase (encoded by carD), 4-hydroxy-2-oxovalerate aldolase (encoded by carE), and acetaldehyde dehydrogenase (encoded by carF) (Fig. 1) (40). Dihydroxylation by the initial oxygenase is often a critical step for aerobic bacterial degradation of aromatic compounds. The Rieske nonheme iron oxygenase systems (ROSs) catalyzing dihydroxylation typically consist of two or three components that comprise an electron transfer chain mobilizing electrons from NADH or NADPH, via flavin and [2Fe-2S] redox centers, to the site of dioxygen activation (5, 11, 29). The ROSs have historically been classified into five groups (IA, IB, IIA, IIB, and III) based on their numbers of constituents and the nature of their redox centers (2). The three-component systems consisting of terminal oxygenase, ferredoxin, and ferredoxin reductase have been classified into groups IIA, IIB, and III. In groups IIA and IIB, ROSs have putidaredoxin homologs and the Rieske-type ferredoxin, respectively, as the ferredoxin component but commonly have a ferredoxin reductase component that contains flavin adenine dinucleotide (FAD). The class III ROSs have the Rieske-type ferredoxin and a reductase which contains FAD and a chloroplast-type [2Fe-2S] cluster. The CARDO from CA10 is classified as a class III ROS according to the above classification (35, 47). Sphingomonas sp. strain KA1 possesses the carAaIBaIBbICIAcI genes, which are located on the 254-kb circular plasmid pCAR3 (16, 19, 56). As a result of the determination of the complete nucleotide sequence of pCAR3, a second copy of the car gene cluster (carAaIIBaIIBbIICIIAcII), a ferredoxin gene (fdxI), and two ferredoxin reductase genes (fdrII and fdrII) were identified (56). In the CARDO of KA1, putidaredoxin-type ferredoxins (CarAcI, CarAcII, and FdxI) mediate electron transport from ferredoxin reductases (FdrI and FdrII) to the terminal oxygenases (CarAaI and CarAaII), suggesting that the CARDO from KA1 is affiliated with a class IIA oxygenase.

FIG. 1.

Pathway for degradation of carbazole by gram-negative P. resinovorans CA10 (40, 41) and gram-positive N. aromaticivorans IC177.

Carbazole degradation by gram-negative bacteria has been studied genetically in detail. However, there are no reports about carbazole catabolic genes from gram-positive bacteria. Some gram-positive bacteria, such as actinomycetes, can mineralize various recalcitrant molecules (30). Nocardioides species, which belong to the Actinomycetales, have been shown to degrade a variety of aromatic compounds, including 2,4,6-trinitrophenol, 2,4,5-trichlorophenoxyacetic acid, p-nitrophenol, phenanthrene, and dibenzofuran (9, 13, 15, 20, 25, 45). Recently, we isolated a carbazole degrader, Nocardioides aromaticivorans (formally Nocardioides sp.) IC177 (18). IC177 accumulates anthranilate in culture during biodegradation of carbazole as well as the gram-negative carbazole degrader CA10 (18, 41). This strain can also grow on protocatechuate, dibenzothiophene sulfone, and anthranilate as the sole source of carbon (18). Degenerate PCR experiments using primers for amplifying the internal fragment of the carAa gene revealed that IC177 possesses carAa and carC homologs (18).

For this study, we characterized the genes for carbazole degradation and determined the catabolic pathway of carbazole to anthranilate in a gram-positive bacterium. We also demonstrated the operonic structure of the car gene clusters and their inducibility by reverse transcription-PCR (RT-PCR) and quantitative RT-PCR (qRT-PCR). The substrate preference of the initial carbazole oxygenase from IC177 was also investigated.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The bacterial strains, plasmids, and cosmids used in this study are listed in Table 1. The medium used for bacterial growth was Luria-Bertani (LB) medium (46). Escherichia coli JM109 (46) and DH5α (46) were used as host strains for the plasmids pBluescript II SK(−), SuperCos1, and their derivatives. Ampicillin (50 μg/ml), kanamycin (50 μg/ml), and chloramphenicol (30 μg/ml) were added to selective media. For plate cultures, the above-described media solidified with 1.6% (wt/vol) agar were used.

TABLE 1.

Bacterial strains, plasmids, and cosmids used in this study

Strain, plasmid, or cosmid	Relevant characteristics	Source or reference
Bacterial strains
Escherichia coli JM109	recA1 supE44 endA1 hsdR17 Δ(lac-proAB) gyrA96 relA1 thi [F′ traD36 proAB+lacIqZΔM15]	46
Escherichia coli DH5α	supE44 ΔlacU169 φ80dlacZΔM15 hsdR17 recA1 endA1 gyrA96 thi-1 relA1	46
Nocardioides aromaticivorans IC177	Car+a	18
Plasmids
pBluescript II SK(−)	AprlacZ, pMB9 replicon	Stratagene
pB177101	Apr, pBluescript II SK(−) with 12.2-kb NotI insert of IC177 DNA	This study
pB177102	Apr, pBluescript II SK(−) with 6.5-kb NotI-EcoRV insert of IC177 DNA	This study
pB177103	Apr, pBluescript II SK(−) with 3.0-kb XhoI insert of IC177 DNA	This study
pB177104	Apr, pBluescript II SK(−) with 5.5-kb EcoRV-NotI insert of IC177 DNA	This study
pB177105	Apr, pBluescript II SK(−) with 2.5-kb SacI insert of IC177 DNA	This study
pB177106	Apr, pBluescript II SK(−) with 3.7-kb BamHI insert of IC177 DNA	This study
pB177107	Apr, pBluescript II SK(−) with 1.2-kb BamHI insert of IC177 DNA	This study
pB177301	Apr, pBluescript II SK(−) with 1.2-kb insert containing carAa gene of IC177	This study
pB177302	Apr, pBluescript II SK(−) with 0.3-kb insert containing carAc gene of IC177	This study
pB177304	Apr, pBluescript II SK(−) with 1.6-kb insert containing carAaAc genes of IC177	This study
pB177305	Apr, pBluescript II SK(−) with 2.4-kb insert containing carAaAd genes of IC177	This study
pB177306	Apr, pBluescript II SK(−) with 1.6-kb insert containing carAcAd genes of IC177	This study
pB177307	Apr, pBluescript II SK(−) with 2.8-kb insert containing carAaAcAd genes of IC177	This study
pB177308	Apr, pBluescript II SK(−) with 3.1-kb insert containing carAaAcAdBa genes of IC177	This study
pB177309	Apr, pBluescript II SK(−) with 3.9-kb insert containing carAaAcAdBaBb genes of IC177	This study
pB177310	Apr, pBluescript II SK(−) with 4.8-kb insert containing carAaAcAdBaBbC genes of IC177	This study
pT177001	Apr, pT7Blue(R) with 991-bp degenerate PCR-amplified DNA fragment containing partial carAaC genes of IC177	18
pT177201	Apr, pT7Blue(R) with 1.2-kb PCR-amplified DNA fragment containing carAa gene of IC177	This study
pT177202	Apr, pT7Blue(R) with 0.4-kb PCR-amplified DNA fragment containing carAc gene of IC177	This study
pT177203	Apr, pT7Blue(R) with 1.2-kb PCR-amplified DNA fragment containing carAd gene of IC177	This study
pT177204	Apr, pT7Blue(R) with 0.3-kb PCR-amplified DNA fragment containing carBa gene of IC177	This study
pT177205	Apr, pT7Blue(R) with 0.8-kb PCR-amplified DNA fragment containing carBb gene of IC177	This study
pT177206	Apr, pT7Blue(R) with 0.9-kb PCR-amplified DNA fragment containing carC gene of IC177	This study
pT7Blue(R)	AprlacZ	Novagen
pUCARA	Apr, pUC119 with 5.6-kb EcoRI insert of CA10 DNA	47
SuperCos1	Apr Kmr cos	Stratagene
Cosmid
pSC1770033	Apr, SuperCos1 with Sau3AI insert of IC177 DNA containing car gene cluster	This study

^aCar⁺ indicates an ability to grow on carbazole as the sole source of carbon, nitrogen, and energy.

DNA manipulations.

DNA techniques were performed according to standard methods (46). Total DNA of IC177 grown on LB medium at 30°C and plasmid and cosmid DNAs from E. coli host cells were prepared as described previously (40). DNA fragments were extracted from agarose gels by using a QIAEX II DNA extraction kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions.

Cloning of DNA containing the car genes.

We constructed a genomic cosmid library of IC177 as described previously (40), except that we used the restriction endonuclease Sau3AI instead of EcoRI to prepare the vector arm of SuperCos1 and partially digested genomic DNA. To screen the positive cosmid clones containing the car gene clusters of IC177, we conducted colony hybridization using probes prepared from the 991-bp DNA fragment containing partial carAa and carC from IC177 (insert of pT177001) (Table 1), as previously described (40). Southern hybridization analysis of endonuclease-digested DNA of IC177 was performed under stringent conditions according to previously described methods (40). A positive cosmid clone designated pCS1770033 was used for further analyses.

Nucleotide sequence determination, homology searches, and alignment analysis.

Nucleotide sequences were determined using the GPS-1 genome priming system (New England Biolabs, Ipswich, MA) as described previously (40). The nucleotide sequences obtained were analyzed with DNASIS-Mac software (version 3.7; Hitachi Software Engineering, Yokohama, Japan). Nucleotide and deduced amino acid sequences were searched for similarities by using the BLAST program (http:/www.ncbi.nlm.nih.gov/BLAST/). Amino acid sequence alignment analysis and phylogenetic analysis were performed by using the ClustalW program at the DDBJ website (http://www.ddbj.nig.ac.jp). Phylogenetic trees were produced with the TreeView program (42).

RNA extraction and RT-PCR.

Cells for RNA extraction were harvested from 5 ml NMM2 culture containing carbazole (0.1% [wt/vol]) after 5 days of incubation. NMM2 had the following composition: Na₂HPO₄, 2.2 g/liter; KH₂PO₄, 0.8 g/liter; MgSO₄ · 7H₂O, 0.2 g/liter; NH₄NO₃, 3 g/liter; FeSO₄ · 7H₂O, 0.01 g/liter; and CaCl₂ · 2H₂O, 0.01 g/liter. Harvested cells were resuspended in 100 μl distilled water containing 2 mg/ml lysozyme and incubated at 37°C for 1 h. Total RNA was prepared with Sepasol-RNA I Super (Nacalai Tesque, Kyoto, Japan) according to the manufacturer's instructions. Contaminating DNA in total RNA preparations was removed with RQ1 RNase-free DNase (Promega, Madison, WI). RT-PCR was performed in 25-μl reaction mixtures containing 200 ng total RNA and 10 pmol of each primer with a OneStep RNA PCR kit (Takara, Tokyo, Japan). The thermocycler program used for RT-PCR was as follows: 60°C for 30 min; 94°C for 2 min; 30 cycles of 96°C for 30 s, 55°C for 30 s, and 72°C for 2 min; and 72°C for 7 min. The nucleotide sequences of the eight primer sets used for RT-PCR are shown in Table S1 in the supplemental material.

qRT-PCR.

After cultivation in 5 ml of NMM2 containing succinate (0.1% [wt/vol]) at 30°C for 4 days, succinate (0.1% [wt/vol]), succinate (0.08% [wt/vol]) plus anthranilate (0.02% [wt/vol]), or succinate (0.08% [wt/vol]) plus carbazole (0.02% [wt/vol]) was added to the resultant IC177 cultures. The concentration of all added stock solutions was 10% (wt/vol). After incubation of each culture at 30°C for 1, 4, and 8 h, total RNA was isolated as described above. qRT-PCR was performed using Quantitect SYBR green RT-PCR (QIAGEN) on an ABI PRISM 7700 sequence detection system (Applied Biosystems, Foster City, CA). The total volume of the reaction mixture was 20 μl and contained 80 ng (or 8 pg for rrn gene detection) RNA. The primer sets used for amplification of partial carAa, carD, and rrn genes in qRT-PCR analysis are shown in Table S1 in the supplemental material. The thermal cycling conditions were as follows: 30 min at 55°C, 15 min at 95°C, and 40 cycles of 30 s at 96°C, 30 s at 55°C, and 30 s at 72°C. A negative control (distilled water instead of the RNA solution) and an RNA sample without the reverse transcription step (to determine genomic DNA contamination) were included. For each measurement, a standard curve was generated by using appropriately diluted target DNA fragments. Relative quantification was based on the levels of carD and carAa mRNAs compared to that of rrn as a reference gene. For each RNA extraction, measurements of gene expression were obtained in triplicate.

Construction of expression plasmids.

To express the car genes under control of the lac promoter in E. coli, we constructed pB177301 (for the expression of carAa), pB177302 (carAc), pB177304 (carAaAc), pB177305 (carAaAd), pB177306 (carAcAd), pB177307 (carAaAcAd), pB177308 (carAaAcAdBa), pB177309 (carAaAcAdBaBb), and pB177310 (carAaAcAdBaBbC) (Table 1). Each car gene, including an artificial Shine-Dalgarno (SD) sequence (51), was amplified by PCR from pSC1770033 and its derivatives, using the primers shown in the supplemental material. Each PCR product containing carAa, carAc, carAd, carBa, carBb, or carC was ligated into the pT7Blue T-vector (Novagen, Madison, WI), and the resultant plasmids were designated pT177201, pT177202, pT177203, pT177204, pT177205, and pT177206, respectively. The nucleotide sequence of each insert was checked with the original sequence. All of the inserts of the plasmids possessed an XbaI site and an SD sequence upstream of each car gene, with a SpeI site immediately downstream of the termination codon for each car gene. pT177201 was digested with XbaI and SpeI. A 1.2-kb XbaI-SpeI fragment containing the SD sequence and the carAa gene was gel purified. The 1.2-kb XbaI-SpeI fragment was ligated into the SpeI site of pBluescript II SK(−) to give pB177301, taking advantage of the identical nucleotide sequences at the staggered end of the SpeI site and that of XbaI. The resultant plasmid was digested with the appropriate enzymes to check the orientation of the insert. The carAa gene in pB177301 was under the control of the lac promoter and an SD sequence. The 0.4-kb XbaI-SpeI fragment containing the SD sequence and the carAc gene was prepared from pT177202 and ligated into the SpeI site of pB177301 (carAa) to give pB177304 (carAaAc). The pB177302 (carAc), pB177305 (carAaAc), pB177306 (carAcAd), pB177307 (carAaAcAd), pB177308 (carAaAcAdBa), pB177309 (carAaAcAdBaBb), and pB177310 (carAaAcAdBaBbC) plasmids were constructed in a similar manner.

Biotransformation analysis.

Escherichia coli JM109 cells carrying plasmid pBluescript II SK(−) (negative control), pB177301, pB177304, pB177305, pB177306, pB177307, pB177310, or pUCARA (positive control) (47) were cultured for 8 h at 37°C in 5 ml LB medium containing ampicillin (50 μg/ml). The culture was transferred to 250 ml fresh LB medium and further cultured at 37°C. After 2 h of incubation, isopropyl-β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM to induce gene expression, and the cells were harvested after 8 h by centrifugation. The collected cells were washed twice with CNF buffer and resuspended in the buffer at an optical density at 600 nm of 20 to 30. CNF buffer had the following composition: Na₂HPO₄, 2.2 g/liter; and KH₂PO₄, 0.8 g/liter. A reaction mixture consisting of 3 ml cell suspension and 0.01% (wt/vol) (0.46 to 0.78 mM) substrate was placed in a test tube and shaken at 285 strokes per min at 37°C. After 24 h of incubation, the mixture was extracted with an equal volume of ethyl acetate. For detection of anthranilate formed from carbazole, the incubated culture was acidified to pH 3 with 1 N HCl prior to the ethyl acetate extraction. The ethyl acetate layer was dried over anhydrous sodium sulfate. A portion of each extract was analyzed using a JMS-Automass 150 gas chromatography-mass spectrometry (GC-MS) system (JEOL, Tokyo, Japan) as previously described (18). The substrates and products were identified by comparison with previously analyzed retention times and MS spectra (18, 21, 38, 53).

Nucleotide sequence accession numbers.

The nucleotide sequences reported in this study have been deposited in the DDBJ, EMBL, and GenBank nucleotide sequence databases. The nucleotide sequence data for the 17,582-bp NotI-BamHI fragment of IC177 was registered under accession no. AB244528.

RESULTS

Cloning and sequencing of car genes from IC177.

The nucleotide sequence of the 16S rRNA gene from IC177 is most similar to that of N. aromaticivorans H-1^T (99.6% identity), as recently described (25). Due to this similarity, we concluded that IC177 belongs to this species, and we renamed it accordingly.

A 991-bp DNA fragment containing tandemly linked partial carAa and carC homologs was previously obtained by PCR experiments using degenerate primers for carAa (18). A cosmid clone containing the above-described 991-bp region was selected from an IC177 cosmid library by colony hybridization and designated pSC1770033. The nucleotide sequence of a 17,582-bp (NotI-BamHI) region in pSC1770033 containing the carAa and carC gene homologs was determined. The positions and lengths of the inserts of the subclones used for sequencing analysis are shown in Fig. 2. The G+C content of the 17,582-bp region was uniformly high (67.3%). The 17,582-bp region contained 19 complete open reading frames (ORFs) and a partial ORF (ORF9). Similarities between the ORFs and representative homologs are summarized in Table 2.

FIG. 2.

Genetic organization of car gene clusters and their flanking regions in N. aromaticivorans IC177 and electrophoresis of RT-PCR products. The pentagons indicate the sizes, locations, and directions of ORF transcription. The car genes are shown by black pentagons. Gray pentagons indicate the putative transcriptional regulators. Black bars indicate the locations of the subcloned DNA fragments or locations of the expected RT-PCR products for each lane. Restriction sites are indicated as follows: B, BamHI; E, EcoRV; S, SacI; N, NotI; and X, XhoI. Only restriction sites relevant to the clones and constructs described in this figure are shown.

TABLE 2.

Proposed functions of proteins encoded by ORFs identified in a 17,582-bp region of the IC177 genome

ORF (gene)	Positiona	Probable function	Homology
			Amino acid identity (%)b	Name of protein	Source	Accession no.
ORF9	1-432	Oxidoreductase	59	Putative oxidoreductase	Streptomyces coelicolor A3(2)	CAC44594
ORF10	429-866	Thioesterase	26	Probable 4-hydroxybenzoyl-CoA thioesterase	Pirellula sp. strain 1	CAD78650
ORF11	872-1,672	Transcriptional regulator	37	Putative regulator	Streptomyces avermitilis MA-4680	BAC70019
ORF12	1,780-2,358	Unknown		Hypothetical protein
ORF13	2,403-2,948	Transcriptional regulator	23	TetR family transcriptional regulatory protein	Gloeobacter violaceus PCC 7421	BAC88612
ORF 14 (carRII)	3,085-3,849	Transcriptional regulator	36	CarRI	N. aromaticivorans IC177	This study
ORF15 (carD)	3,919-4,734	2-Hydroxypenta-2,4-dienoate hydratase	49	Putative 2-keto-4-pentenoate hydratase	Pseudomonas putida strain RR21	CAE92879
ORF16 (carF)	4,731-5741	Acetaldehyde dehydrogenase (acylating)	64	Acetaldehyde dehydrogenase	Azotobacter vinelandii	ZP_00092919
ORF17 (carE)	5,738-6,760	4-Hydroxy-2-oxovalerate aldolase	67	XylK	N. aromaticivorans F199	AAD03993
ORF18	6,881-7,501	Unknown
ORF19	7,568-9,910	Transcriptional regulator	36	Putative regulator	S. coelicolor A3(2)	CAB93384
ORF20 (carRI)	9,917-10,705	Transcriptional regulator	36	CarRII	N. aromaticivorans IC177	This study
ORF1 (carAa)	10,860-12,026	Terminal oxygenase component of carbazole 1,9a-dioxygenase	49	CarAa	Sphingomonas sp. strain KA1	BAC56759
ORF2 (carC)	12,023-12,862	meta-Cleavage compound hydrolase	45	CarC	Sphingomonas sp. strain KA1	BAC56762
ORF3 (carBa)	12,859-13,134	Subunit of meta-cleavage enzyme	29	CarBa	Janthinobacterium sp. strain J3	BAC56743
ORF4 (carBb)	13,134-13,949	Subunit of meta-cleavage enzyme	39	CarBb	Sphingomonas sp. strain KA1	BAC56761
ORF5 (carAc)	13,949-14,296	Ferredoxin component of carbazole 1,9a-dioxygenase	36	Putative dioxygenase ferredoxin subunit	S. avermitilis MA-4680	BAC74038
ORF6 (carAd)	14,296-15,468	Ferredoxin reductase component of carbazole 1,9a-dioxygenase	38	Ferredoxin reductase	Mycobacterium tuberculosis	AAK44942
ORF7	15,528-16,058	Unknown
ORF8	16,253-17,461	Transposase	34	Putative transposase	S. avermitilis MA-4680	BAC68012

^aPositions in 17,582-bp DNA region sequenced in this study (accession no. AB244528).

^bValues were calculated only for the sequenced region.

The amino acid sequence of ORF1 showed similarity to CarAa, the terminal oxygenase component of carbazole 1,9a-dioxygenase, from KA1 and CA10 (49 and 45% identity, respectively). The consensus sequences of Rieske-type iron-sulfur proteins for the binding of a [2Fe-2S] cluster and mononuclear nonheme iron (36, 43) were determined. The amino acid sequences of ORF2, ORF3, and ORF4 from IC177 showed similarities to the meta-cleavage compound hydrolase CarC (45% identity), the 2′-aminobiphenyl-2,3-diol 1,2-dioxygenase small subunit CarBa (19% identity), and the large subunit CarBb (39% identity), respectively, from KA1. The deduced amino acid sequence of ORF5 contained a conserved sequence for Rieske-type [2Fe-2S] clusters and showed similarity to CarAc from CA10 (31% identity) (34) and the ferredoxin component of biphenyl dioxygenase BphF from Acidovorax sp. strain KKS102 (20% identity) (10). The deduced amino acid sequence of ORF6 possessed FAD-binding and NADH-binding domains (29, 49) and showed major similarity to the ferredoxin reductase components of initial oxygenases of aromatic compounds, such as AhdA4 (ferredoxin reductase component of salicylate dioxygenase) from Sphingomonas sp. strain P2 (37% identity) (44) and BphA4 (ferredoxin reductase component of biphenyl dioxygenase) from Acidovorax sp. strain KKS102 (28%) (22). Thus, following the nomenclature of CA10, these six ORFs were designated carAa, carC, carBa, carBb, carAc, and carAd (Fig. 2).

In a region 4.1 to 6.9 kb upstream of the initiation codon of the carAa gene, ORF15, ORF16, and ORF17 were identified (Fig. 2). The deduced amino acid sequences of ORF15, ORF16, and ORF17 showed similarities to 2-hydroxypenta-2,4-dienoate hydratase, acetaldehyde dehydrogenase (acylating), and 4-hydroxy-2-oxovalerate aldolase, respectively, whose genes have been identified in various aromatic-compound-degrading bacteria, such as bphEGF in Acidovorax sp. strain KKS102 (23) and Rhodococcus sp. strain RHA1 (28), todGIH in P. putida F1 (26), dmpEFG in Pseudomonas sp. strain CF600 (52), and carDFE in P. resinovorans CA10 (40). Putative crucial residues for dehydrogenase activity and putative metal iron ligands were identified on the basis of the crystal structures of DmpF and DmpG, which functionally correspond to CarF and CarE, respectively (27). The carDFE gene products were thought to function as meta-pathway enzymes for carbazole catabolism in IC177 (Fig. 1 and 2).

The deduced amino acid sequence of ORF20, located immediately upstream of the carAaCBaBbAcAd gene cluster, showed high homology with transcriptional regulators classified in the IclR family. The putative regulatory gene (tentatively designated carRI) was transcribed in the opposite direction from that of the carAaCBaBbAcAd genes (Fig. 2), implying that this gene product functions as a transcriptional regulator of the carAaCBaBbAcAd genes. Members of the IclR family of regulatory proteins are generally transcriptional repressors, but those that control catabolic pathways have all been described as activators (54). Moreover, a putative transcriptional regulatory gene (designated carRII) located immediately upstream of the carD gene was identified (Fig. 2). The deduced amino acid sequence of CarRII also showed similarity to the transcriptional regulators in the IclR family, with the greatest similarity to CarRI (36% identity) (Table 2).

Operonic structures and car gene inducibility.

To confirm that the identified car genes are actually expressed as a single transcriptional unit in response to carbazole exposure, RT-PCR experiments were performed with total RNA extracted from IC177 cells grown on carbazole. The primers were designed to generate PCR products encompassing the tandemly linked genes, as shown in Fig. 2. Whereas PCR without RT did not show any amplification products, RT-PCR produced products of the expected sizes (Fig. 2). These data showed that the carAa, -C, -Ba, -Bb, -Ac, and -Ad genes and the carD, -F, and -E genes are transcribed as respective operons.

To analyze the inducibility of the car operons, we performed quantitative RT-PCR, using two pairs of primers specific for carAa and carD. IC177 cells grown on succinate were exposed to succinate alone, anthranilate plus succinate, or carbazole plus succinate. Total RNA was extracted from each sample after 1, 4, or 8 h of incubation. Compared to the negative control samples containing no reverse transcriptase or template, very low levels of carAa and carD mRNAs were detected at all stages in cells incubated with succinate alone or anthranilate plus succinate (data not shown). The expression levels of both carAa and carD did not change upon exposure to anthranilate. These results show that anthranilate, an intermediate of the carbazole degradation pathway (Fig. 1), is not the inducer of the car operons of IC177, although anthranilate is an inducer of the CA10 car operon (55). carAa expression was inducible in cells exposed to carbazole (Fig. 3A), and the carAa mRNA level at 1 h in the presence of carbazole plus succinate was approximately 42-fold higher than that in the presence of anthranilate plus succinate or succinate alone (Fig. 3A). Further incubation with carbazole plus succinate led to a higher level of carAa expression (approximately 26-fold higher than that after 1 h of incubation). carD expression was also induced in cells incubated with carbazole plus succinate (Fig. 3B). However, the incubation time for induction of carD expression was 4 h, which differed from that for carAa (Fig. 3).

FIG. 3.

Quantitative RT-PCR analysis of carAa (A) and carD (B) mRNAs. mRNA levels of carAa and carD in N. aromaticivorans IC177 were normalized to that of rrn. Times for incubation and exposure to the substrates in succinate-grown IC177 cells are shown.

Expression of carAa, carAc, carAd, carBa, carBb, and carC genes in E. coli.

The sequencing analysis and gene expression data indicated that the carAa, -Ac, and -Ad genes encode a carbazole dioxygenase responsible for the growth of IC177 on carbazole. To confirm the role of initial oxygenation of carbazole, the carAa, -Ac, and -Ad genes were cloned and expressed in E. coli. The carAa, carAaAc, carAaAd, carAcAd, and carAaAcAd genes were cloned into pBluescript II SK(−), and the resultant plasmids were designated pB177301, pB177304, pB177305, pB177306, and pB177307, respectively. The ability of E. coli JM109 carrying each plasmid to transform carbazole to 2′-aminobiphenyl-2,3-diol after IPTG induction was analyzed. E. coli JM109(pB177307), expressing carAaAcAd, completely (>99%) transformed carbazole to 2′-aminobiphenyl-2,3-diol in 24 h. These data indicate that the gene products of carAaAcAd function as the angular dioxygenase for carbazole. The angular dioxygenase catalyzes the dihydroxylation of carbazole at an angular positioned carbon (C-9a) binding to the imino nitrogen and the adjacent C-1 carbon (Fig. 1). After 24 h of incubation, E. coli JM109(pB177304), expressing carAaAc, oxidized 2.3% of the carbazole, suggesting that the ferredoxin (CarAc) had the ability to accept electrons from unidentified E. coli reductases with a much lower efficiency than that from its own natural reductase (CarAd). E. coli JM109 carrying pB177301 (carAa), pB177305 (carAaAd), and pB177306 (carAcAd) did not transform carbazole at all. These data demonstrate that carAc and carAd encode a specific ferredoxin and ferredoxin reductase, respectively, which act together with the terminal oxygenase component to transform carbazole to 1,9a-dihydroxy-1-hydrocarbazole, which is considered to be spontaneously converted to 2′-aminobiphenyl-2,3-diol (Fig. 1).

The functionality of the carBaBbC gene products was also tested by measuring their enzymatic activity in E. coli. We constructed plasmids containing the carAaAcAdBa, carAaAcAdBaBb, and carAaAcAdBaBbC genes under control of the lac promoter and designated them pB177308, pB177309, and pB177310, respectively. A colony of E. coli harboring pB177309 grown in LB medium containing 1 mM IPTG showed production of a yellow compound after being sprayed with ethereal 2,3-dihydroxybiphenyl, which is an analog of 2′-aminobiphenyl-2,3-diol, although no production of the yellow compound was observed on E. coli harboring pB177308 prepared as a negative control (data not shown). This result indicates that the carBaBb gene products function as meta-cleavage enzymes in the catabolism of carbazole in IC177, as shown in Fig. 1.

To confirm the function of the carC gene product, we also analyzed its reaction in resting cells. The ability of E. coli JM109 expressing CarC in conjunction with CarAaAcAdBaBb to transform carbazole to anthranilate was analyzed. Although no production of anthranilate was detected in E. coli cells carrying pBluescriptII SK(−), anthranilate (approximately 2% of the remaining carbazole) was clearly detected in the cell suspension of E. coli expressing carAaAcAdBaBbC incubated with carbazole. Thus, it was confirmed that the carBaBbC gene products possess the expected functions in the conversion of carbazole to anthranilate, as shown in Fig. 1.

Substrate preference of CarAaAcAd.

To analyze the substrate preference of the CARDO of IC177, E. coli carrying pB177307 (containing carAaAcAd) was tested for the ability to convert several aromatic compounds, including biphenyl, naphthalene, 9-fluorenone, fluorene, dibenzothiophene sulfone, dibenzo-p-dioxin, dibenzofuran, and carbazole. The products were identified by comparison to the retention times and MS patterns of GC-MS data from previous experiments (21, 38, 53). Among the seven aromatic compounds, naphthalene was the most efficiently transformed (Table 3). The lateral dioxygenation product, cis-1,2-dihydroxy-1,2-dihydronaphthalene, and its putative dehydrated derivative, 1-naphthol, were detected (Table 3). Significant transformation was also observed for dibenzo-p-dioxin, and only the angular dioxygenation product (2,2′,3-trihydroxydiphenyl ether) was detected. Weak transformation of biphenyl to cis-2,3-dihydroxy-2,3-dihydrobiphenyl (with 2-hydroxybiphenyl, a putative dehydrated derivative from cis-2,3-dihydroxy-2,3-dihydrobiphenyl), fluorene to 9-hydroxyfluorene and dihydrodiol (lateral dioxygenation product) (53), and dibenzofuran to 2,2′,3-trihydroxybiphenyl was also detected (Table 3). Other putative oxygenation products were not detected in this analysis.

TABLE 3.

Biotransformation analysis using E. coli expressing carbazole 1,9a-dioxygenase from N. aromaticivorans IC177

Substrate	Detected compounda	Conversion ratio (%)b
Naphthalene	cis-1,2-Dihydroxy-1,2-dihydronaphthalene	47 (±4.8)
	1-Naphthol	53 (±4.8)
Carbazole	2′-Aminobiphenyl-2,3-diol	96 (±4.6)
Dibenzo-p-dioxin	2,2′,3-Trihydroxydiphenyl ether	69 (±2.5)
Dibenzofuran	2,2′,3-Trihydroxybiphenyl	8.8 (±0.75)
Biphenyl	cis-2,3-Dihydroxy-2,3-dihydroxybiphenyl	3.5 (±0.63)
	2-Hydroxybiphenyl	2.7 (±0.24)
	Biphenyl dihydrodiol	0.50 (±0.15)
Fluorene	9-Hydroxyfluorene	1.6 (±0.25)
	Dihydrodiol	1.9 (±0.51)
9-Fluorenone	NDc
Dibenzothiophene sulfone	NDc

^aProducts were identified by GC-MS analysis and compared to reported GC-MS data (21, 38, 53). These compounds also contained the products derived by nonenzymatic dehydration.

^bConversion ratio (%) = 100 × (peak area for the total ion current [TIC] of the product)/(peak area for substrate TIC + peak area for the TIC of all identified products). Values in parentheses represent standard deviations calculated from at least three assays.

^cND, no product was detected in this analysis.

DISCUSSION

We have studied the diversity of the carbazole catabolic genes and isolated several carbazole degraders and their car genes. To explain the evolutionary relationships of each carbazole catabolic gene, we analyzed the structures of the car genes and their flanking regions from several carbazole-degrading bacteria (18, 19, 40). The gene order of the carC and carBaBb genes in IC177 was inverted relative to that in well-characterized car gene clusters of the gram-negative bacteria P. resinovorans CA10, Janthinobacterium sp. strain J3, and Sphingomonas sp. strain KA1 (Fig. 4). In the carAaCBaBbAcAd and carDFE operons in IC177, all of the genes overlap each other by 1 or 4 bp (data not shown), while such complete overlap was not observed in car gene clusters from gram-negative bacteria. In addition, the carAc gene was separated from the carAd gene by an ORF of unknown function in CA10 and J3 (Fig. 4). As for the two car gene clusters of KA1, there was no gene corresponding to carAd and encoding NAD(P)H:ferredoxin oxidoreductase in the neighboring region of the car gene cluster (Fig. 4). Moreover, in CA10, carD was separated from the carFE genes by two unknown ORFs, and these genes were located downstream of the carAaBaBbCAcAd gene cluster, implying that the carFE genes were recruited into the car gene cluster during the late phase of the evolutionary process (19). However, the carDFE genes were closely linked and located upstream of the carAaCBaBbAcAd gene cluster in IC177. These results suggest a unique evolutionary history of carbazole catabolic genes in IC177. Many genes involved in the degradation of aromatic compounds are organized into operons or functional units (57). The carbazole catabolic operons of IC177 are organized more orderly as functional units than those in gram-negative strains, such as CA10, J3, and KA1.

FIG. 4.

Comparison of gene organization around the car gene clusters of N. aromaticivorans IC177, P. resinovorans CA10, Janthinobacterium sp. strain J3, and Sphingomonas sp. strain KA1. The pentagons indicate the sizes, locations, and directions of gene or ORF transcription.

Various ROSs have been identified and extensively studied, mainly as the initial oxygenases of aromatic compounds (5, 11, 14, 17, 29, 58). The known CARDOs from P. resinovorans CA10, Sphingomonas sp. strain KA1, and Sphingomonas sp. strain GTIN11 are class III, IIA, and IIA enzymes, respectively (19, 24, 47). In addition, a phylogenetically less-related CARDO was isolated from Sphingomonas sp. strain CB3 (50). Although the car genes of CB3 are transcribed in carbazole-grown CB3 cells, the enzymatic activity of the CARDO was not directly detected (50). Thus, we excluded the CARDO of CB3 from our discussion on the electron transfer functions of CARDOs. The terminal oxygenase of IC177 was closely related to the CARDOs from CA10 and J3 (class III) and from KA1 (class IIA) (>45% identity) (see Fig. S1A in the supplemental material), and the ferredoxin of IC177 was a Rieske-type ferredoxin similar to that of CA10, even though they did not have phylogenetically very close relationships with each other (see Fig. S1B in the supplemental material). The amino acid sequence of the ferredoxin reductase component (CarAd) from IC177 showed high homology with the NAD(P)H:ferredoxin oxidoreductases from class IIB ROSs (see Fig. S1C in the supplemental material). It did not contain the consensus sequence of plant-type iron-sulfur proteins for the binding of a [2Fe-2S] cluster which is found in the ferredoxin reductase component of CA10. On the basis of these findings, the CARDO of IC177 belongs to class IIB. It is noteworthy that the terminal oxygenase components of the CARDOs from CA10/J3, KA1, and IC177 have substantial homology with each other (>45%) but that their electron transfer components are completely different. It would be interesting to learn how such similar terminal oxygenases can accept electrons from such diverse electron transfer components among the three CARDOs. Further biochemical and protein structural studies of these CARDOs will provide important information on the interactions of the electron transfer proteins and the enzyme evolutionary mechanisms.

The substrate preference of IC177 CARDO differed from those of CA10 and KA1. The CARDOs from CA10 and KA1 have high enzyme activities toward biphenyl and dibenzofuran (38, 53, 56), although these compounds were not good substrates for IC177 CARDO. Therefore, the CARDO of IC177 has a stronger preference for carbazole than those of CA10 and KA1. Recently, the crystal structure of a terminal oxygenase of CARDO from Janthinobacterium sp. strain J3 was determined (36). Some hydrophobic residues, such as Ile184, Ile186, Leu270, Val272, Phe275, Phe329, and Ile334, in CarAa of J3 were found to comprise substrate-binding pockets. In the terminal oxygenase of IC177, the residues corresponding to Ile186 and Ile334 of J3 are replaced by Met192 and Val338, respectively, although these residues are conserved in KA1. The difference in substrate specificity may result from the replacement of these amino acid residues. These residues for substrate binding are not conserved in either the naphthalene or biphenyl dioxygenase (6, 12); therefore, it will be important to determine which amino acid residues are involved in determining CARDO's substrate preference. In a previous study, we demonstrated that IC177 can grow on dibenzothiophene sulfone as the sole carbon source (18). There are several reports on dibenzothiophene sulfone- and 9-fluorenone-degrading bacteria suggesting that they convert these two compounds to angular dioxygenation products (4, 37). However, no plausible oxidation product was formed by IC177 CARDO with dibenzothiophene sulfone and 9-fluorenone. Thus, there may be a series of genes for the degradation of dibenzothiophene sulfone in another locus of the IC177 genome.

The (putative) regulatory proteins of the car gene clusters of IC177, CA10, J3, and KA1 are located differently in the neighboring regions of each car gene cluster (Fig. 4). Moreover, CarRI of IC177 showed similarity to transcriptional regulators in the IclR family, although the transcriptional regulators of the car operons of CA10 and J3 are the AraC/Xyl family protein AntR (55) and the GntR family protein CarR (31), respectively. Therefore, the transcriptional regulatory systems of the car genes of IC177, CA10, and J3 are all different, although the carbazole catabolic genes are similar. Many related catabolic pathways do not carry the same regulatory system, which led to the hypothesis that regulatory systems and their target operons do not necessary coevolve but seem to become associated independently (7, 8, 54). The differences in the regulatory systems of carbazole catabolic genes in CA10, J3, and IC177 also support this hypothesis. It would be interesting to determine which conditions cause the regulators from each strain to activate their car operons and how the activities of the translated products differ from each other.

From the results of qRT-PCR, the carAa gene was induced by exposure to carbazole within 1 h, although it took 4 h to detect the transcription of carD, implying a difference between the inducers of the carAaCBaBbAcAd and carDFE operons. In the carbazole degradation pathway, the gene products of the carAaCBaBbAcAd operon are responsible for the upper pathway of carbazole degradation, while those of the carDFE operon are responsible for the lower pathway (meta-pathway) (Fig. 1). Thus, it is reasonable that the carAaCBaBbAcAd operon is more promptly transcribed than the carDFE operon after application of the substrate.