INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Microorganisms exposed to xenobiotics often show adaptive activities. As the molecular mechanism for adaptation, gene recruitment is the key to establishing several simultaneous degradation activities and the entire catabolic capacity for the xenobiotics to which microorganisms are exposed. In fact, catabolic gene clusters which have similarities in the levels of gene organization and nucleotide sequences have been reported for phylogenetically unrelated bacterial strains isolated from geographically distinct areas (15, 19, 46, 58, 61). Mobile genetic elements, such as plasmids and transposons, have been considered to play an important role in distributing catabolic capacities by gene recruitment. Previously, many xenobiotic-degradative gene clusters have been reported to be located on mobile genetic elements (55, 57, 62). The recruited gene clusters can also undergo various types of rearrangements including deletion, duplication, inversion, and insertion into other replicons, especially when the bacteria adapt in order to acquire novel catabolic activities (18, 62).

The well-clustered genetic organization of entire pathways or independently functioning pathway segments has been reported for bacteria which degrade easily degradable compounds, such as naphthalene, toluene, and biphenyl (15, 23, 63). It can be assumed that these genetic structures developed through long exposure to corresponding compounds. On the other hand, the degradative genes involved in the degradation of highly recalcitrant compounds seem to be disordered or separated. It can be speculated that the different levels of maturation in genetic structures resulted from the difference in the length of exposure to xenobiotics or rarity of the gene(s) indispensable for degradation. That is, the degradative genes for a highly recalcitrant molecule are useful in revealing the maturation (evolution) stage of xenobiotic-degrading gene clusters.

Carbazole (CAR) is a heterocyclic aromatic compound containing a dibenzopyrrole system derived from coal tar and shale oil (41). CAR is known to possess mutagenic and toxic activities (3) and also to be a recalcitrant molecule (12). Pseudomonas sp. strain CA10 is a microorganism having the ability to utilize CAR as a sole source of carbon, nitrogen, and energy (45). As shown in Fig. 1, strain CA10 degrades CAR to anthranilate and 2-hydroxypenta-2,4-dienoate through angular dioxygenation, meta-cleavage, and hydrolysis (45). The anthranilate produced is attacked by dioxygenase at the 1,2 positions to yield catechol, which is degraded by the -ketoadipate pathway. On the other hand, 2-hydroxypenta-2,4-dienoate is converted to a tricarboxylic acid (TCA) cycle intermediate by the three additional conversions called the meta-cleavage pathway.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Degradation pathway of CAR in Pseudomonas sp. strain CA10. Enzyme designations: CarAaAcAd, carbazole 1,9a-dioxygenase; CarBaBb, 2'-aminobiphenyl-2,3-diol 1,2-dioxygenase; CarC, 2-hydroxy-6-oxo-6-(2'-aminophenyl)-hexa-2,4-dienoic acid (meta-cleavage compound) hydrolase; CarD, 2-hydroxypenta-2,4-dienoate hydratase; CarE, 4-hydroxy-2-oxovalerate aldolase; CarF, acetaldehyde dehydrogenase (acylating); AntABC, anthranilate 1,2-dioxygenase; CatA, catechol 1,2-dioxygenase; CatB, cis,cis-muconate lactonizing enzyme; CatC, muconolactone -isomerase. Compounds: I, CAR; II, 2'-aminobiphenyl-2,3-diol; III, 2-hydroxy-6-oxo-6-(2'-aminophenyl)-hexa-2,4-dienoic acid (meta-cleavage compound); IV, anthranilic acid; V, catechol; VI, cis,cis-muconic acid; VII, muconolactone; VIII, -ketoadipic acid enol-lactone; IX, 2-hydroxypenta-2,4-dienoic acid; X, 4-hydroxy-2-oxovaleric acid; XI, pyruvic acid; XII, acetaldehyde; XIII, acetyl coenzyme A.

Using Tn5 mutagenesis, we have previously shown the presence of car and cat gene clusters encoding the enzymes involved in the initial degradation of CAR and the -ketoadipate pathway, respectively (30). By shotgun cloning, we have isolated the car gene cluster encoding carbazole 1,9a-dioxygenase (CarAaAcAd), the meta-cleavage enzyme (CarBaBb), and the meta-cleavage compound hydrolase (CarC), which catalyze the first three steps in CAR degradation (49, 50). Genetic analysis of the car gene cluster revealed that the 1,263-bp DNA region containing the carAa gene is tandemly duplicated except for one base (49). Also, two copies of carAa genes are separated from the carAc and carAd genes by the 2-kb DNA fragment containing carBaBb and carC. So far, such a gene structure has not been observed in other degradative gene clusters, except for the different alleles of the car gene cluster. The genetic organization of the car gene cluster of strain CA10 implies the possibility that gene rearrangements occurred through evolutionary divergence in the natural environment. Because of the fact that the G+C content of the identified car gene cluster is lower than that of total DNA of strain CA10 (45, 50), the car gene cluster was also hypothesized to have been recruited from other microorganisms.

We herein describe in detail the genetic analysis of the car gene cluster and its flanking region. We also discuss the evolutionary events that occurred during the development of a novel genetic structure in this DNA region. In addition, it is reported that the car gene cluster is located on the megaplasmid, although the cat gene cluster is located on the chromosome of strain CA10.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, media, and culture conditions. Bacterial strains, plasmids, and cosmids used in this study are listed in Table 1. A physical map of the plasmids and cosmids constructed in this study is also shown in Fig. 2. For extractions of total DNA and the megaplasmid, strain CA10 was grown on 2xYT medium (48) and nutrient broth (Eiken Chemical Co. Ltd., Tokyo, Japan), respectively, at 30°C with reciprocal shaking (300 strokes/min) for 16 h. To prepare the total RNA of strain CA10, carbon- and nitrogen-free mineral medium (CNFMM) supplemented with CAR as the sole source of carbon, nitrogen, and energy was used. CNFMM had the following composition (grams per liter): Na₂HPO₄, 2.2; KH₂PO₄, 0.8; MgSO₄ · 7H₂O, 0.2; FeSO₄ · 7H₂O, 0.01; CaCl₂ · 2H₂O, 0.01; and yeast extract, 0.05. CAR was dissolved in dimethyl sulfoxide (10 mg/ml) and sterilized by filtration with a 0.2-µm-pore-size membrane filter. Then, 1% (vol/vol) of this dimethyl sulfoxide solution was added to CNFMM.

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Physical map of the 44.3-kb DNA region containing the CAR-degrading car gene cluster. The pentagons and triangles in the physical map indicate the sizes, locations, and directions of transcription of the ORFs derived from the DNA sequence analysis. The shaded pentagons and triangles represent the car and ant gene clusters. The shaded boxes containing hatched pentagons represent the ISs, and the hatched pentagons indicate the sizes, locations, and directions of transcription of the transposase genes. The location of the subcloned DNA fragments and orientation of the lac promoter of the cloning vector are indicated by the lengths and the directions of the arrows, respectively. As for the cosmid clones, the DNA region which was revealed to be contained in the respective cosmids by Southern hybridization analysis is shown by the solid line. The double slash attached to the end of the solid line indicates that the DNA region outside the EcoRI site was contained in the respective cosmid insert. Only the restriction enzyme sites used for the construction of the plasmids are shown. Restriction enzyme site designations: C, ClaI; EI, EcoRI; EV, EcoRV; Hc, HincII; Hd, HindIII; P, PstI; Sl, SalI; Sp, SphI; Xb, XbaI; and Xh, XhoI.

DNA manipulation. Plasmid or cosmid DNA was prepared from the E. coli host strain by the alkaline lysis method. Total DNA of strain CA10 was prepared as described previously (50). Restriction endonuclease and DNA Ligation Kit version 2 (Takara Shuzo Co., Ltd., Kyoto, Japan) were used according to the manufacturer's instructions. DNA fragments were extracted from the agarose gel by using Concert Rapid Gel Extraction Systems (Life Technologies, Rockville, Md.) according to the manufacturer's instructions. Other DNA manipulation procedures were performed according to the standard methods (48).

Gene walking to identify the flanking region of the car gene cluster. A cosmid vector, SuperCos1, was first digested with XbaI, and the digested end of the DNA fragment was dephosphorylated with calf intestinal alkaline phosphatase (CIAP; Roche Diagnostics GmbH, Mannheim, Germany). Next, the resultant XbaI-CIAP-treated DNA fragment was digested with EcoRI. Total DNA of strain CA10 was partially digested with EcoRI and directly ligated to XbaI-CIAP/EcoRI-treated SuperCos1 vector. After ligation, the linear DNA formed was packaged in vitro into phage by using Gigapack III Gold Packaging Extract (Stratagene), which was used to infect host E. coli cells prepared according to the manufacturer's instructions. After the infection, host cells were spread onto the 2xYT agar plate supplemented with ampicillin to screen the clones carrying the flanking region of the car gene cluster. Colony hybridization was carried out according to the methods reported by Sambrook et al. (48) using a Biodyne B nylon membrane (Poll BioSupport Co., East Hills, N.Y.). The probe for hybridization was prepared from the 6.9-kb EcoRI DNA fragment of pUCA1 using the Megaprime DNA labeling system (Amersham Pharmacia Biotech UK Ltd., Little Chalfont, Buckinghamshire, England) and [-³²P]dCTP (110 TBq/mmol; Amersham Pharmacia Biotech UK Ltd.).

Southern hybridization. Southern blotting was performed by using a Biodyne B nylon membrane according to the recommendations of the manufacturer. A nonradioactive digoxigenin DNA labeling and detection kit (Roche Diagnostics GmbH) was used according to the manufacturer's instructions. Prehybridization and hybridization were carried out at 68°C. After hybridization, the membranes were washed three times with 2× SSC-0.1% SDS (20 min, room temperature) and three times with 0.1× SSC-0.1% SDS (20 min, 68°C).

Nucleotide sequence determination, homology search, and alignment. To generate a population of DNA sequencing templates with randomly interspersed primer-binding sites, the GPS-1 Genome Priming System (New England Biolabs, Inc., Beverly, Mass.) was used. Nucleotide sequence determination was carried out by the chain termination method using a Li-Cor Model 4200L-2 auto-DNA sequencer (Li-Cor Inc., Lincoln, Nebr.) according to the manufacturer's instructions, and the nucleotide sequences obtained were analyzed with DNASIS-Mac software (version 3.7; Hitachi Software Engineering Co. Ltd., Yokohama, Japan). Homology searching was performed using the SWISS-PROT amino acid sequence data bank or the DDBJ/EMBL/GenBank DNA databases with the BLAST program (version 2.0.10) (1). Alignment of the deduced amino acid sequences of observed open reading frames (ORFs) was performed by using the CLUSTAL W package (version 1.6) (56).

Northern hybridization. After the cultivation of strain CA10 in 5 ml of nutrient broth, cells were gathered from 4 ml of culture by centrifugation at 7,200 × g and then washed twice using CNF buffer, which contained 2.2 g of Na₂HPO₄ and 0.8 g of KH₂PO₄ per liter. The washed cells were suspended in 4 ml of CNF buffer. The cells were starved by incubation of the resultant cell suspension at 30°C with reciprocal shaking at 300 strokes/min for 6 h. The starved cells were washed twice as described above and finally suspended in 4 ml of CNF buffer. One hundred milliliters of CNFMM supplemented with CAR was inoculated with 10 µl of the resultant cell suspension of strain CA10. After 12 h of incubation with reciprocal shaking (300 strokes/min) at 30°C, the growing cells were harvested and used for extraction of total RNA. For the comparative study, 100 ml of nutrient broth was inoculated with 10 µl of the starved-cell suspension, and the nutrient broth-grown cells were similarly harvested as described above. The strain CA10 cells just after the starvation and after the cultivation on nutrient broth were used for extraction of total RNA. Total RNA from harvested cells was extracted using an RNeasy Mini or Midi kit combined with a QIAshredder and RNase-Free DNase set from Qiagen (Santa Clarita, Calif.) according to the manufacturer's instructions.

Preparation of cell extract. E. coli strain JM109 harboring plasmid pBCA616, pUCA617, or pUCA618 was grown to an optical density at 550 nm of 0.8. After further incubation in the presence of 1 mM isopropyl--D()-thiogalactopyranoside for 5 h at 37°C, the cells were harvested, washed twice with 50 mM sodium phosphate buffer (pH 7.5), and resuspended in 10 ml of the same buffer. The cell suspensions were sonicated and centrifuged at 21,600 × g at 4°C for 60 min, and the supernatants were used as cell extracts. The protein concentration was determined by using the Protein Assay Kit II (Bio-Rad Laboratories, Richmond, Calif.) with bovine serum albumin as a standard.

SDS-PAGE. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was performed with a 13% polyacrylamide gel. Protein staining of the gels was performed with Coomassie brilliant blue R-250.

Determination of enzymatic activities. The hydratase activity for 2-hydroxypenta-2,4-dienoate of the cell extract was assayed by monitoring the decrease in A₂₆₅ with a Beckman DU-7400 spectrophotometer equipped with a thermojacketed cuvette holder and a circulating water bath, as described by Harayama et al. (22). The substrate was prepared enzymatically from DL-allylglycine by the method used for synthesis of 2-oxopent-4-enoate (13). The assay was performed at 25°C using 1 ml of 10 mM Tris-HCl buffer (pH 7.0) containing 3.3 mM MgSO₄ and 50 µg of crude cell extract/ml. The reaction was initiated by the addition of 50 µM substrate. One unit of activity was defined as the amount of the enzyme required to degrade 1 µmol of 2-hydroxypenta-2,4-dienoate per min at 25°C. The molar extinction coefficient of 2-hydroxypenta-2,4-dienoate was taken to be 19,200 cm¹ M¹ (47).

1

1

Isolation of 16S rRNA gene from strain CA10 and determination of the sequence. The 16S rRNA gene was amplified from the total DNA of strain CA10 by PCR using two primers that attached to positions 8 to 27 in the E. coli numbering system (5'-AGAGTTTGATC[A/C]TGGCTCAG-3') and positions 1513 to 1492 (TACGG[A/T/C]TACCTTGTTACGACTT). The amplified DNA fragment was ligated to pT7Blue(R) (Novagen, Madison, Wis.), and then the resultant pTCA16S was used as the template to determine the nucleotide sequence of the partial 16S rRNA gene of strain CA10 as described above.

Extraction of the megaplasmid from strain CA10 cells. After the cultivation using nutrient broth, strain CA10 cells were harvested from 5 ml of culture by centrifugation (7,200 × g for 2 min), washed with 1 ml of sterilized distilled water, and repelleted. According to a previously reported procedure (27), the resultant cells suspended in 50 µl of distilled water were lysed by adding 200 µl of lysing solution (50 mM Tris, 3% SDS, pH 12.6). The solution was incubated at room temperature for 15 min followed by heating at 80°C for 1 min and extracted with 250 µl of a phenol-chloroform solution (1:1, vol/vol) saturated with Tris-EDTA (TE) buffer (48). The resultant lysate was incubated for 3 h. After centrifugation (7,200 × g 15 min), the aqueous phase was subjected to electrophoresis in 0.7% (wt/vol) agarose H (Nippon Gene Co., Tokyo, Japan) in Tris-acetate-EDTA buffer (48).

Chemicals. CAR and anthranilic acid were purchased from Tokyo Kasei Kogyo Co., Ltd. (Tokyo, Japan). All other chemicals were of the highest purity commercially available.

Nucleotide sequence accession number. The nucleotide sequence data of the car gene cluster flanking region and the 16S rRNA gene of strain CA10 reported in this study were registered in the DDBJ, EMBL, and GenBank nucleotide sequence databases with the accession no. AB047548 and AB047273, respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning of the genes containing the flanking region of the car gene cluster. A cosmid library of the strain CA10 genome was constructed, and approximately 1,800 of the resultant colonies were screened for the car gene cluster by probing with the 6.9-kb EcoRI insert DNA fragment of pUCA1. Thus, nine positive clones (pSCos701 to pSCos709) were isolated, and restriction mapping combined with Southern hybridization analyses was performed. The resultant restriction maps of these clones are shown in Fig. 2.

Determination of ORFs by nucleotide sequence analysis. To determine the nucleotide sequence of the upstream region of the carAa gene, we constructed the deletion plasmids pUCA601, pUCA610, pBCA620, pBCA711, pBCA722, pBCA731, pUCA741, and pBCA751 from the cosmid inserts. The series of Tn7 mutant derivatives of each plasmid were constructed by using the GPS system, and the flanking region of the Tn7 insertion was sequenced. Consequently, the nucleotide sequences of 27,937-bp-long upstream and 9,448-bp-long downstream regions of the already-known car gene cluster were determined. There existed 23 (ORF9 to ORF31) and 9 (ORF32 to ORF40) additional ORFs in the upstream and downstream regions of the car gene cluster as shown in Fig. 2 and Table 2. The putative Shine-Dalgarno sequences (51) were found in the upstream regions of the putative initiation codons of ORFs except for ORF9, ORF10, ORF14, ORF15, ORF16, ORF22, ORF23, ORF25, ORF29, ORF31, ORF38, and ORF39 (data not shown).

Homology search analyses. To determine the functions of identified ORFs, we performed homology search analyses, the results of which are summarized in Table 2.

(i) ORF9 to ORF15. The deduced amino acid sequence of ORF9 showed 40% identity with the large subunit (AntA) of the oxygenase component in anthranilate dioxygenase from Acinetobacter sp. strain ADP1 (11). However, the similarity with AntA was observed only in the N-terminal region of a putative protein encoded by ORF9, and the C terminus showed no homology to previously reported proteins. In addition, while homologous proteins (AntA [11], XylX [21], and BenA [40]) have 450 to 470 amino acid residues, ORF9 encodes a polypeptide with 567 amino acids.

View larger version (70K): [in this window] [in a new window]   FIG. 3.   Alignments of TnpA1 and TnpA3 identified from strain CA10 with their homologous transposases. The aligned transposases are contained in ISs classified in the IS5 family (32). The identical and similar amino acid residues in all proteins are indicated by asterisks and dots, respectively. The gaps introduced into the alignment are indicated by hyphens. The critical DDE consensus motif and the well-conserved R (or K) residue 7 amino acids downstream from the E residue in N2, N3, and C1 domains (32) are indicated by the white type. The DDBJ, EMBL, and GenBank nucleotide sequence accession numbers are indicated after the names of transposases and the names of ISs or bacterial strains (in parentheses).

(ii) ORF16 to ORF21. As summarized in Table 2, it is thought that the putative proteins encoded by three ORFs, ORF17 to ORF19, showed homology to the reductase component and the large () and small () subunits of the terminal oxygenase component of the multicomponent oxygenase system, respectively. Alignment analyses revealed that a putative plant-type iron-sulfur protein consensus sequence (CX₄CX₂CX₂₈C) for the binding of a [2Fe-2S] cluster (44) and a possible NAD-ribose binding region were contained in the deduced amino acid sequence of ORF17 (data not shown). In the N terminus of the deduced amino acid sequence of ORF18, the consensus sequence of Rieske-type iron-sulfur proteins for the binding of a [2Fe-2S] cluster (CXHX_16-17CX₂H) (28) was found (Fig. 4). The consensus sequence for the binding of mononuclear iron (28) was also observed in ORF18 (Fig. 4). Based on these results, the proteins encoded by these ORFs constitute the multicomponent dioxygenase system classified in class IA by Batie et al. (6).

View larger version (52K): [in this window] [in a new window]   FIG. 4.   Alignments of the large () subunits of class I terminal oxygenase components with the postulated ORF18 and antA gene products. The identical and similar amino acid residues in all proteins are indicated by asterisks and dots, respectively. The gaps introduced into the alignment are indicated by hyphens. The amino acids conserved for the Rieske-type [2Fe-2S] cluster binding site and the mononuclear iron coordination site (28) are indicated by white type and shaded boxes, respectively. The DDBJ, EMBL, and GenBank nucleotide sequence accession numbers are indicated after the names of enzymes and the names of plasmids or bacterial strains (in parentheses).

(iii) ORF22 to ORF31. Homology search and alignment analyses revealed that ORF25 encodes a protein identical to TnpA1 and that the nucleotide sequence of ORF25 was identical to that of ORF10 except for one base. Because ORF25 was considered to encode transposase, we designated ORF25 as tnpA2. On the other hand, the deduced amino acid sequence of ORF29 showed 99% identity with those of both TnpA1 from P. stutzeri strain AN10 (7) and TnpA of IS1384 (AF052751) (Table 2). Ninety-six percent identity was observed between the deduced amino acid sequence of ORF29 and that of TnpA1 (TnpA2) of strain CA10. The critical DDE consensus motif and the R (or K) residue 7 amino acids downstream from the E residue (32) were also conserved in the deduced amino acid sequence of ORF29, as shown in Fig. 3. Based on these results, ORF29 was designated tnpA3.

(iv) ORF32 to ORF36. The deduced amino acid sequences of ORF32, ORF35, and ORF36 showed significant homology to those of 2-hydroxypenta-2,4-dienoate hydratase, acetaldehyde dehydrogenase (acylating), and 4-hydroxy-2-oxovalerate aldorase, which were identified from various aromatic-compound-degrading bacteria (Table 2). Similar to other reported 2-hydroxypenta-2,4-dienoate hydratases, the pentapeptide (ADNAS), which has been suggested to have either a catalytic or a structural role in the respective enzymes (31), was observed near the central portion of the deduced amino acid sequence of ORF32 (data not shown). Based on the above results, ORF32, ORF35, and ORF36 were designated carD, carF, and carE, respectively.

(v) ORF37 to ORF40. Pairwise comparison among the amino acid sequences of the transposases identified from a neighboring region of the car gene cluster revealed that the deduced amino acid sequence of ORF38 is identical to those of TnpA1 and TnpA2. Therefore, ORF38 was designated tnpA4. The deduced amino acid sequence of ORF39 showed homology (36 to 38% identity) with C-terminal regions (about 300 amino acids in length) of proteins of unknown function, YkoW in Bacillus cereus strain ATCC 14579 (43) and hypothetical protein slr1305 in Synechocystis sp. strain PCC6803 (S76238), although these two proteins have 840 and 892 amino acid residues, respectively. The deduced amino acid sequences of ORF37 and ORF40 did not show significant homology with those of any other reported proteins.

Analysis of ISs. As shown above, four copies of homologous transposase genes were found in the flanking region of the car gene cluster of strain CA10. As a result of the alignment analysis of the neighboring region of transposase genes, inverted repeat sequences (IRs; 16 bp) were found as shown in Fig. 5. In addition, the transposases identified from strain CA10, TnpA1, TnpA2, TnpA3, and TnpA4, had significant homology with the IS5 family of transposases as shown in Table 2 and Fig. 3. Based on the above results, IS-like DNA regions containing tnpA1, tnpA2, tnpA3, and tnpA4 were designated IS5car1, IS5car2, IS5car3, and IS5car4, respectively, and considered to be classifiable in the IS5 subgroup of the IS5 family (32). The nucleotide sequence of IS5car1 is identical to that of IS5car4 and nearly identical to that of IS5car2 with one base pair mismatch. On the other hand, the nucleotide sequence of IS5car3 showed only 81.6% identity to that of IS5car1, although it showed 98% identity to the IS-like sequence containing tnpA1 from P. stutzeri strain AN10 (7). In the flanking region of three ISs, IS5car2, IS5car3, and IS5car4, a target site duplication consisting of 4-bp-long direct repeat sequences (DRs; CTAG, CTAA, and TTAG for IS5car2, IS5car3, and IS5car4, respectively) was observed, although the 4-bp-long DNA sequence located on the left border (TTAA) of IS5car1 was not identical with that of the right border (CTAG) (Fig. 5).

View larger version (16K): [in this window] [in a new window]   FIG. 5.   Terminal regions of ISs and their flanking sequences identified from strain CA10. The IRs and DRs are shown by white type and underlining, respectively. The complementary nucleotides between the left and right IRs are shown by asterisks. The direction of transposase genes is shown by the arrow.

Transposition of IS5car2 with its flanking sequence. As shown in Fig. 6, the 2,279-bp-long DNA regions (8,878 to 6,600 in AB047548) containing IS5car2 and the 5' portion of the antA gene were quite similar to the 2,115-bp-long DNA regions (25,306 to 27,420 in AB047548) containing IS5car1 and the 5' portion of ORF9. The 888-bp-long 5' portion of antA corresponds to the 698-bp-long 5' portion of ORF9 and the 26-bp-long 5' upstream sequence of ORF9. Deletion of a 163-bp-long DNA fragment and replacement of AT bases with a single G base were shown to occur in ORF9. The 3' half of ORF9 following the transposed antA sequence showed no homology to that of the antA gene. These results strongly suggest that the transposition of IS5car2 and the 5' portion of the antA gene resulted in the formation of IS5car1 and the fusion gene ORF9. However, immediately 3' of the 698-bp-long ORF9 portion, we could not find the target site duplication which was considered to be formed by the transposition (data not shown).

View larger version (21K): [in this window] [in a new window]   FIG. 6.   Comparison of 2,279-bp-long DNA regions (8,878 to 6,600 in AB047548), containing IS5car2 and the 5' portion of the antA gene, with 2,115-bp-long DNA regions (25,306 to 27,420 in AB047548), containing IS5car1 and the 5' portion of ORF9. ISs and ORFs are shown by the boxes. Shaded pentagons within the boxes represent the locations, lengths, and directions of translation of transposase genes. The hatched box represents the deleted region of the antA gene in the nucleotide sequence of ORF9. Dashed lines represent the neighboring region of transposed DNA.

Genomic Southern hybridization of the identified gene. The above-mentioned results indicate the possibility that the IS plays an important role in the development of the car and ant gene clusters. To reveal the copy numbers of IS5car2 and IS5car3 in the strain CA10 genome, Southern hybridization analyses were conducted. Using the probe made from the 332-bp PvuI-HindIII fragment in IS5car2, we detected four positive bands (Fig. 7A). Based on the determination of the hybridized band, three of the four bands were revealed to correspond to DNA fragments containing IS5car1, IS5car2, and IS5car4. Because the hybridization and washing were carried out under high-stringency conditions, IS5car3 could not be detected. Therefore, the additional bands indicate the presence of another copy of a nearly identical IS.

View larger version (69K): [in this window] [in a new window]   FIG. 7.   Southern blot analyses using the 332-bp PvuI-HindIII fragment (IS5car2) of pBCA711 (A), the 880-bp PstI-DraI fragment (IS5car3) of pBCA731 (B), and the 1,119-bp SmaI fragment (antA) of pBCA711 (C) as probes. Restriction endonucleases used for digestion of total DNA of strain CA10 are shown in the panels.

Detection of the products of carD, carE, and carF. To confirm if ORF32, ORF35, and ORF36 could be translated to proteins with the predicted sizes in E. coli, the cellular proteins from the recombinant E. coli cells harboring plasmids encoding the respective genes were loaded onto an SDS-13% polyacrylamide gel (Fig. 8). Expression of ORF32, ORF35, and ORF36 yielded peptides having molecular masses of 27 (lane 2), 35 (lane 4) and 36 (lane 6) kDa, respectively. The estimated molecular masses were in accordance with the predicted molecular masses of CarD (27,458 Da), CarF (33,265 Da), and CarE (36,651 Da).

View larger version (61K): [in this window] [in a new window]   FIG. 8.   Detection of CarD, CarF, and CarE produced by E. coli cells. Total cellular proteins of E. coli cells were analyzed by SDS-PAGE. Lanes: 1 and 8, molecular mass standards of 94, 67, 43, 30, 20.1, and 14.4 kDa (top to bottom, respectively); 2, JM109(pBCA616); 3, JM109[pBluescript II SK()]; 4, JM109(pUCA617); 5, JM109(pUC18); 6, JM109(pUCA618); 7, JM109(pUC19).

Enzymatic activities of CarD and CarF. CarD showed significant hydratase activity for 2-hydroxypenta-2,4-dienoate, and its enzymatic activity for 2-hydroxypenta-2,4-dienoate was 3.05 U/mg of protein in crude extracts from E. coli cells harboring pBCA616. The hydratase activity of negative-control E. coli cells harboring pBluescript II SK() was almost negligible (0.076 U/mg of protein).

Transcriptional analysis of the ant genes. In order to determine whether the transcripts of the ant genes exist in the CAR-grown strain CA10 cells, Northern hybridization analyses were performed. As shown in Fig. 9, lane 3, a strong and clear hybridization, with the 200-bp SmaI fragment being specific for antA and not for ORF9, was observed, although the band detected was very blurry. A similar hybridization was observed when the antC gene was used as a probe (data not shown). In sharp contrast, no hybridization band was detected when equal amounts of total RNA extracted from the strain CA10 cells grown on nutrient medium were loaded in the lanes (as confirmed by ethidium bromide visualization of 16S and 23S rRNA bands before transfer [data not shown]) (Fig. 9, lane 2). These results indicate that the ant genes are strongly induced in the strain CA10 cells grown on CAR, although it is impossible to tell whether there is more than one band or whether the mRNA was degraded.

View larger version (36K): [in this window] [in a new window]   FIG. 9.   Northern hybridization analysis using the antA gene as a probe. Total RNA was extracted from the CA10 cells grown on nutrient medium (lane 2) and CAR (lane 3) to late log phase. As the control, total RNA was similarly extracted from strain CA10 cells just after starvation culture (lane 1).

Analysis of the 16S rRNA gene of strain CA10. For use as the authentic gene encoded on the chromosome of strain CA10 in Southern hybridization analysis, we amplified the 16S rRNA gene by using PCR. After cloning the amplified DNA fragment into the plasmid vector, we determined the nucleotide sequence of the insert.

Localization of the car gene cluster, the cat gene cluster, and transposase genes. Electrophoresis of the cell lysate of strain CA10 revealed two bands (Fig. 10A). As shown in Fig. 10B, the probe made from the 16S rRNA gene hybridized to the lower band, suggesting that this lower band is a chromosome. This result indicated that strain CA10 cells have at least one megaplasmid. Clear and specific hybridization to the upper band (megaplasmid) was observed with the car gene (pUCA1 insert) used as the probe (Fig. 10C). These results indicate that the car and ant gene clusters which are responsible for CAR degradation by strain CA10 are encoded on this megaplasmid. Therefore, we designated this megaplasmid pCAR1.

View larger version (36K): [in this window] [in a new window]   FIG. 10.   Southern hybridization analysis of the chromosome and plasmid detected in strain CA10 cells using the 16S rRNA gene (EcoRI-PstI fragment of pTCA16S) (B), the car gene (EcoRI insert of pUCA1) (C), the cat gene (SalI fragment of pTE11) (D), and the tnpA2 gene (PvuI-HindIII fragment in IS5car2) (E) as probes. A photograph of the ethidium bromide-stained agarose gel is shown in panel A.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Whole structure of the car gene cluster. Homology searching, SDS-PAGE analyses, and the determination of enzymatic activity revealed that the genes encoding the meta-cleavage pathway enzymes involved in the conversion of 2-hydroxypenta-2,4-dienoate to a TCA cycle intermediate are located in the downstream region of the carAd gene (Fig. 2). Consequently, the car gene cluster consists of 10 genes (carAaAaBaBbCAcAdDFE) which are involved in the conversion of CAR to anthranilate and 2-hydroxypenta-2,4-dienoate and the degradation of 2-hydroxypenta-2,4-dienoate to a TCA cycle intermediate. In the case of bacterial gene clusters for the degradation of biphenyl, toluene, isopropylbenzene, phenol, and naphthalene (8, 17, 20, 25, 29, 31, 52), the genes for hydratase, acetaldehyde dehydrogenase (acylating), and aldolase which constitute the meta-cleavage pathway are assembled in this order and tandemly linked. Interestingly, the functionally unknown ORF33 and ORF34 are inserted between carD and carFE in the case of the car gene cluster of strain CA10. Such unusual organization of corresponding genes has been reported elsewhere for Rhodococcus sp. strain RHA1 (33) and Comamonas testosteroni strain TA441 (2). In the biphenyl-degrading bph operon of Rhodococcus sp. strain RHA1, the genes encoding the meta-cleavage enzyme (etbC), the meta-cleavage compound hydrolase (bphD), hydratase (bphE), and aldolase (bphF) are clustered, while the putative acetaldehyde dehydrogenase (acylating) (bphG) is not encoded in the neighboring region of the etbCbphDEF operon. In the 3-(3-hydroxyphenyl)propionic acid-degrading mhp operon of C. testosteroni strain TA441, two ORFs (ORF4 and ORF5) encoding proteins of unknown function are inserted between mhpD and mhpFE. Because genes similar to the two ORFs have been found in the gene clusters involved in the chlorocatechol degradation pathway and biosynthesis of poly(3-hydroxybutyrate-co-4-hydroxybutylate) (2), the products of ORF4 and ORF5 might have these functions, although disruption of the genes did not affect the 3-(3-hydroxyphenyl)propionic acid-degrading activity of strain TA441. As shown in Table 2, CarF and CarE showed significantly high homology with MhpF and MhpE, respectively, of strains TA441 and E. coli K-12. In contrast, CarD showed the highest identity with the isofunctional enzymes involved in isopropylbenzene degradation (Table 2), while the identities of CarD with MhpD proteins from C. testosteroni strain TA441 (2) and E. coli strain K-12 (AE000142) were only 45 and 40%, respectively. In addition, the nucleotide and deduced amino acid sequences of ORF33 and ORF34 showed no homology with those of mhp genes including ORF4 and ORF5 identified from strain TA441. On the basis of this observation and the fact that carAaBaBbCAcORF7Ad also showed no phylogenetic affiliation with mhp genes at both DNA and amino acid sequence levels (data not shown), we can speculate that carFE genes have been recruited from other loci, resulting in the present structure of the car gene cluster (Fig. 2). This hypothesis is in accordance with the fact that carEF is transcribed as an mRNA different from carAaAaBaBbCAcORF7AdD (M. Urata, H. Nojiri, T. Yoshida, H. Habe, and T. Omori, unpublished data).

Involvement in CAR degradation by the proteins encoded by the genes located in the flanking region of the car gene cluster. Southern hybridization analysis revealed that strain CA10 has only one copy of intact antABC genes on the genome (Fig. 7C). In addition, Northern hybridization analysis revealed that antABC genes were strongly induced in CAR-grown strain CA10 cells (Fig. 9). These results strongly suggest that AntABC functions as the critical enzyme catalyzing the conversion of anthranilate to catechol in CAR degradation by strain CA10. However, we have not succeeded in the expression of antABC in E. coli cells and detection of anthranilate 1,2-dioxygenase activity (data not shown). To support the evidence of the function of AntABC in CAR degradation, it is important to detect anthranilate 1,2-dioxygenase activity by using the expression system in other bacterial host strains, such as pseudomonads. In addition, it would also be interesting to determine the effects of gene disruption of antABC on the metabolic capacity of CAR and anthranilate in strain CA10. By identifying the ant gene cluster in addition to the car gene cluster, we have elucidated the genes encoding enzymes associated with the catabolic pathway from CAR to catechol. Several CAR-utilizing bacteria have already been isolated, and the degradation pathway, degradative enzymes, and corresponding genes have been studied (10). However, this is the first report on an entire gene set specific for CAR degradation in bacteria.

Identification of ISs. Sequencing analysis revealed the presence of four copies of homologous ISs in the flanking region of the car and ant gene clusters (Fig. 2 and Table 2). These ISs are considered to belong to the IS5 subgroup of the IS5 family (32) based on homology searching and alignment analyses. Among the four ISs, it is thought that IS5car2 and IS5car3 constitute the composite transposon containing the entire antABC genes (Fig. 2). The DR observed at either border of IS5car2 was not identical to that observed at IS5car3, although only the central TA was conserved between the two DRs. For the IS5 subgroup, a target sequence ([C/T]TA[G/A]) has been observed in which either all four base pairs or the central TA is duplicated on insertion (32). Therefore, this DNA region might be formed by the transposition of this composite transposon. These results also imply the possibility that antABC genes have been acquired in this locus during the evolution of the CAR degradation pathway in strain CA10. It is well known that many bacterial strains have the capability to convert catechol to a TCA cycle intermediate through meta or ortho cleavage (24). Therefore, the composite transposon containing only the anthranilate 1,2-dioxygenase gene is considered to be effective enough to distribute anthranilate metabolic capacity in the environment.

Novel structure of the car gene cluster and its flanking region. A proposed model of formation of the present novel structure of the car gene cluster and its flanking region is shown in Fig. 11. The present study revealed that IS5car1 and the 5' portion of ORF9 were formed by transposition of IS5car2 and the 5' portion of antA. This phenomenon has been termed one-ended transposition and has been reported for Tn3, Tn21, and Tn1721, which lack one terminal IR (4, 5, 36). Also, van der Ploeg et al. reported that IS1247 was able to transpose itself together with the dhlB gene without the requirement for a second insertion element to be present downstream of dhlB (59). In contrast to the case with the transposition of IS1247 and the dhlB gene (59), we could not find the target site duplication or central TA duplication flanking the ends of the transposed DNA region containing IS5car1 and the 5' portion of ORF9 (data not shown). In general, the one-ended transposition occurs at a much lower frequency than does a normal transposition. We must analyze the frequency of transposition (or one-ended transposition) of identified ISs and the putative composite transposon-containing antABC genes. Also, it is quite interesting to reveal the change of genetic structure in the flanking region of these putative transposable elements by their transposition. In addition, the deletion of the 163-bp DNA region and the replacement of AT bases with a G base occurred in the 5' portion of ORF9, as shown in Fig. 6. In the 163-bp DNA regions deleted, we could not find DRs or IRs, which possibly promote the deletion by homologous recombination or conservative transposition. It is also interesting to reveal whether these deletions are linked to this one-ended transposition. On the other hand, IS5car1 and IS5car4 form the composite transposon structure containing the whole car gene cluster (Fig. 2). Although IS5car1 is formed by transposition of IS5car2 as shown in Fig. 6, the resultant composite transposon structure might function as a mobile genetic element. Considering the facts that the car gene cluster encodes the entire genes responsible for growth on CAR as a carbon and energy source and that carbazole 1,9a-dioxygenase can attack the dioxin skeleton (42, 49), it is quite interesting to reveal whether this putative composite transposon can transpose into other loci.

View larger version (22K): [in this window] [in a new window]   FIG. 11.   Proposed model of the developmental process of the novel structure in the car gene cluster and its flanking region. Shaded boxes represent the car genes and several neighboring ORFs. Shaded boxes with triangles indicate the ISs. Striped boxes represent the ant genes or the transposed DNA region from the antA gene within ORF9. The arrows with solid lines indicate the transposition (recombination). The arrows with dashed lines indicate the duplication or homologous recombination.

Identification of CAR-degrading megaplasmid pCAR1. As shown in Fig. 10C, the car gene cluster is located on the newly identified megaplasmid pCAR1. Considering that the catechol-degradative cat genes were located on the chromosome (Fig. 10D), it can be considered that strain CA10 had acquired the metabolic capacity for CAR by recruitment of pCAR1. In newly identified CAR-degrading bacteria, we have also detected the pCAR1-like plasmid containing the car gene cluster homologue, although chromosomally encoded car gene clusters were also found (J. Widada, H. Nojiri, S. Nakai, T. Yoshida, H. Habe, and T. Omori, unpublished data). This phenomenon suggests the possibility that pCAR1 plays an important role in the distribution of CAR-degrading activity in the environment.