Arthrobacter aurescens TC1 Atrazine Catabolism Genes trzN, atzB, and atzC Are Linked on a 160-Kilobase Region and Are Functional in Escherichia coli

Abstract

Arthrobacter aurescens strain TC1 metabolizes atrazine to cyanuric acid via TrzN, AtzB, and AtzC. The complete sequence of a 160-kb bacterial artificial chromosome clone indicated that trzN, atzB, and atzC are linked on the A. aurescens genome. TrzN, AtzB, and AtzC were shown to be functional in Escherichia coli. Hybridization studies localized trzN, atzB, and atzC to a 380-kb plasmid in A. aurescens strain TC1.

The complete nucleotide sequences of over 100 bacterial plasmids are now available, but most of these plasmids are relatively small and of interest primarily in microbial pathogenesis (for example, in the context of antibiotic resistance or toxin production). By contrast, only a limited number of large catabolic plasmids or catabolic islands on chromosomes have been sequenced. Examples include plasmids pNL1 from Sphingomonas aromaticivorans strain F199 (23), pAO1 from Arthrobacter nicotinovorans (11), pACR1 from Pseudomonas resinovorans strain CA10 (13), the TOL plasmid pWW0 from Pseudomonas putida (9), pBD2 from Rhodococcus erythropolis BD2 (31), pADP-1 from Pseudomonas sp. strain ADP (16), and the 105-kb catabolic island from Pseudomonas sp. strain B13 (28). Large catabolic plasmids can be difficult to isolate for sequencing. Thus, to facilitate widespread sequencing of large catabolic gene regions, more innovative strategies are needed. For example, plasmids have been identified via whole-genome shotgun sequencing of Deinococcus radiodurans (41) and Enterococcus faecalis (19).

A wide variety of bacteria degrading the herbicide atrazine have been reported (2, 14, 21, 22, 32, 36, 37, 43), and in all cases examined except one (3) the genes involved are located on plasmids (5, 25, 36, 39). In Pseudomonas sp. strain ADP (14), the six atrazine-catabolic genes (atzA, atzB, atzC, and atzDEF) have been localized to plasmid pADP-1, from which the complete nucleotide sequence is now available (16). The atzA, atzB, and atzC genes have also been localized to different-sized plasmids in the gram-negative bacteria Chelatobacter heintzii, Stenotrophomonas maltophilia, and Pseudaminobacter spp., and hybridization analyses indicate that the atzB and atzC genes reside on a 117-kb plasmid in the gram-positive bacterium Arthrobacter crystallopoietes (25, 36). The presence of nearly identical atrazine genes in Agrobacterium, Clavibacter, Rhizobium, Pseudomonas, Alcaligenes, and Ralstonia strains (2, 5, 22, 24, 33, 39) further suggests that horizontal gene transfer is involved in the dissemination of atrazine-catabolic genes.

Some differences have been observed in the gene content of the gram-positive atrazine-degrading bacteria Nocardioides sp. strains C190, SP12, and C157, Arthrobacter crystallopoietes, and Arthrobacter sp. strain AD1 (3, 21, 24, 25, 32, 37). Nocardioides sp. strain C190 (37) does not contain atzA, atzB, or atzC, and atrazine degradation is initiated by the product of the trzN gene, which has been cloned and sequenced previously (17). A. crystallopoietes was shown to contain trzN, atzB, and atzC (25).

The isolation and characterization of an atrazine-degrading gram-positive bacterium, Arthrobacter aurescens strain TC1, which contains atzB and atzC but not atzA, was previously reported (32). Strain TC1 grows on atrazine as the sole source of carbon and nitrogen, liberating cyanuric acid into the growth medium. In the present study, we report that Arthrobacter sp. strain TC1 catabolizes atrazine to cyanuric acid via three enzymatic steps encoded by the hybrid pathway involving the trzN, atzB, and atzC genes. The genes were functionally expressed in Escherichia coli; this is the first report of heterologous expression of the trzN gene. All three atrazine-catabolic genes were localized to a single 161,264-nucleotide region of A. aurescens TC1 DNA, and the complete sequencing and annotation of that region was carried out. Sequence, gel electrophoresis, and hybridization data are all consistent with the idea that trzN, atzB, and atzC reside on a 380-kb plasmid in A. aurescens TC1.

Presence of trzN in A. aurescens TC1.

Previous studies showed that A. aurescens TC1 contained homologs to the atzB and atzC genes from Pseudomonas sp. strain ADP, but it lacked the atzA gene (32). In the present study, PCR done with primers directed against the trzN gene (17) resulted in the amplification of a 0.4-kb DNA fragment. The sequence of this PCR product was 100% identical to a region of trzN from Nocardioides sp. strain C190 (data not shown). These data are consistent with the atrazine degradation pathway shown in Fig. 1A being operative in A. aurescens TC1.

FIG. 1.

(A) Complete catabolic pathway for atrazine degradation in A. aurescens strain TC1. Enzymes involved in each catabolic step are indicated. (B) Activity of TrzN, AtzB, and AtzC in E. coli(BAC6O10) measured by adding purified AtzD from Pseudomonas sp. strain ADP (30), which releases 1 mol each of [¹⁴C]CO₂ and biuret from 1 mol of [¹⁴C]atrazine.

Linkage of trzN, atzB, and atzC on a single BAC clone.

With the knowledge that trzN, atzB, and atzC gene homologs are found in A. aurescens TC1, we hypothesized that these genes might be closely linked and could be captured on a single bacterial artificial chromosome (BAC) clone. A BAC library was constructed using pCUGIBAC1 (http://www.genome.clemson.edu) (35, 42) and high-molecular-weight insert DNA from A. aurescens strain TC1, isolated from agarose plugs and partially digested with HindIII (18, 35). High-density BAC library filters, containing 3,072 colonies, were robotically gridded in duplicate spots on membranes and were screened by hybridization to radiolabeled single-stranded probes prepared from PCR products of the atzB (5), atzC (5′-AGTCAGCGAAGGGCGTAGGTATCA-3′ and 5′-GACAAATCCGGGAGACACAAGGTT-3′), and trzN genes (17) (http://www.genome.clemson.edu). This screening identified 89, 72, and 91 clones containing the trzN, atzB, and atzC genes, respectively. Sixty-four BAC clones hybridized to atzB and atzC gene probes, and one clone hybridized to atzC and trzN but not to the atzB gene probe. Only one clone, BAC6O10, containing a 160-kb insert, hybridized to all three gene probes.

Functional expression of trzN, atzB, and atzC in E. coli.

The functionality of the three atrazine-catabolic genes identified on BAC6O10 was determined using [¹⁴C]atrazine. E. coli (BAC6O10) was grown overnight in Luria-Bertani medium (27) containing 12.5 μg of chloramphenicol per ml, harvested, and disrupted by French press to yield a crude protein extract which was incubated at 25°C for 18 h with approximately 13,000 dpm of [¹⁴C-UL ring]atrazine (15.1 mCi per mmol) in 100 mM phosphate buffer, pH 8.0. Different experimental treatments contained purified atrazine-degrading enzymes TrzN (20), AtzB (J. Osborne, personal communication), and AtzC (30) as complete mixtures or with one component missing. All reaction mixtures contained purified AtzD (8), which catalyzes ring opening and release of one equivalent of CO₂ (Fig. 1B) which was trapped in NaOH. Trapped ¹⁴CO₂ was quantified via liquid scintillation counting. In control reactions containing purified TrzN, AtzB, AtzC, and AtzD, 0.7 mol of carbon dioxide per mol of atrazine was detected in the NaOH trap (Table 1). When crude protein extract from E. coli (BAC6O10) was substituted for AtzB or AtzC, 0.8 mol of carbon dioxide per mol of atrazine was liberated (Table 1). When purified TrzN was complemented with the E. coli(BAC6O10) extract, levels of CO₂ obtained were lower but were significantly greater than those of the negative control. These data are consistent with a previous report by Mulbry et al. (17), who reported that TrzN is difficult to express in E. coli. This is the first report of TrzN activity expressed in recombinant form. In total, these data indicate that TrzN, AtzB, and AtzC are functional on BAC6O10, because there are no other genes on the cloned region encoding amidohydrolase superfamily enzymes (accession no. AY456696), and all enzymes metabolizing atrazine to cyanuric acid observed to date fall within this enzyme class.

TABLE 1.

Functional analysis of BAC6O10-encoded TrzN, AtzB, and AtzC activity in E. coli

Enzyme activity examined	BAC6O10 crude extract	Enzyme addition				Radioactivity in NaOH trap (cpm)a	% Atrazine degradationb
		TrzN	AtzB	AtzC	AtzD
TrzN	+	−	+	+	+	620 ± 80	14.9 ± 1.3
AtzB	+	+	−	+	+	3400 ± 900	77.3 ± 12.9
AtzC	+	+	+	−	+	3300 ± 300	80.1 ± 9.6
Complete pathway	+	−	−	−	+	670 ± 120	17 ± 3.2
Positive control (purified enzymes)	−	+	+	+	+	2600 ± 200	68.3 ± 5.5
Negative control (no enzymes)	−	−	−	−	−	0 ± 14	0 ± 14

^aValues are the means of three replicates ± standard deviations and are corrected for background (112 cpm).

^bValues are expressed as percentages of theoretical; 1 mol of CO₂ released per mol of atrazine is 100%.

Sequence and topology of the 160-kb region.

In order to examine the physical relationship of the atrazine degradation genes to each other and to other flanking genes, a 2.6- to 4-kb, 8× shotgun sequencing library was created from BAC6O10 by using pBluescriptII KS+ (Stratagene, La Jolla, Calif.). Plasmid DNA (15) from 1,920 clones was sequenced using ABI BigDye 2 chemistry and a model 3700 DNA sequencer (Applied Biosystems, Foster City, Calif.). Gap closure was done by primer walking. Sequences were assembled with the Phred/Phrap/Consed software package (version 12); open reading frames (ORFs) were predicted using GeneMark.hmm (12), ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html), and GLIMMER 2.1 (IMM) (6) programs as previously described (16); and predicted coding regions were searched against nonredundant GenBank CDS, PDB, SwissProt, PIR, and PRF protein databases (http://www.ncbi.nlm.nih.gov/) and were analyzed by BLASTP (http://seqsim.ncgr.org/newBlast.html).

The complete nucleotide sequence of BAC6O10, deposited under GenBank accession number AY456696, was found to be 161,264 bp in length, with an overall mean G+C content of 63.5 mol%. A total of 178 ORFs were identified, and a genetic map of BAC6O10 is shown in Fig. 2. The overall topology of BAC6O10, revealed by sequencing and annotation, consisted of two regions of genes: (i) region I (15,873 to 42,322 bp; ORFS 25 through 48) had significant identity to plasmid pADP-1, surrounding the atzB and atzC genes in Pseudomonas sp. strain ADP (accession no. U66917); and (ii) region II, upstream of trzN, consisted of a region (127,401 to 143,000 bp; ORFs 146 through 165) having high similarity to plasmid pAO1 in A. nicotinovorans (accession no. ANI507836). The remainder of the ORFs were closely related to genes identified in the gram-positive bacteria of the genera Streptomyces, Corynebacterium, and Rhodococcus and to genes unique to A. aurescens TC1.

FIG. 2.

Diagram showing positions of ORFs present on BAC6O10. Arrows within each ORF indicate the direction of transcription. Numbers below the lines refer to ORF numbers, and values on the right and left of each line refer to nucleotide sequence numbers. Genes of the same type (transposases [blocks with horizontal striping], atrazine-catabolic genes [black blocks], regulatory genes [blocks with wavy lines], metal resistance genes [blocks with vertical striping], and transfer genes [blocks with vertical and horizontal striping]) have similar shading. ORFs have been numbered relative to their nucleotide positions, and those with putative functions have designators above the lines.

Gene annotation.

Sequence analysis identified only three proteins in the amidohydrolase superfamily (TrzN, AtzB, and AtzC), and no other genes were implicated in atrazine catabolism (Fig. 2). TrzN (ORF 175) had 99% amino acid identity to TrzN from Nocardioides sp. strain C190 (accession no. AF416746). AtzB and AtzC were encoded by ORFs 33 and 38 and had 100 and 99% amino acid identity, respectively, to the AtzB and AtzC proteins from Pseudomonas sp. strain ADP (accession no. U66917). The trzN, atzB, and atzC genes were flanked by different insertion sequence elements or transposases. While linked on the same BAC, the three genes were not organized in an operon-like structure or flanked upstream by a common regulatory element. The trzN gene was located 129 and 122.5 kb from the atzB and atzC genes, respectively, consistent with the BAC clone hybridization data. Taken together, our results indicate that atrazine degradation genes in A. aurescens strain TC1 are located on different regions of BAC6O10 and most likely are transcribed independently.

In addition to the atrazine-catabolic genes, several ORFs on BAC6O10 corresponded to those known to function in metal resistance and transport in gram-positive bacteria. ORF 8 showed 59% amino acid identity to mercuric reductase (MerA) from Streptomyces sp. strain CHR28 (accession no. AAF64138), and ORF 9 had 43% identity to MerR from Mycobacterium tuberculosis CDC1551 (accession no. NP_337968). ORF 123 showed 58% identity to Co/Zn/Cd cation transport protein (CZCT) of Corynebacterium glutamicum ATCC 13032 (accession no. NP_601976), and ORF 59 showed 50% identity to the CzcD efflux protein of Streptomyces coelicolor A3 (accession NP_630824). ORFs 60 and 122 showed 55 and 74% identity to transcriptional regulators of the ArsR family from Streptomyces coelicolor A3 (2) (accession no. NP_627722) and C. glutamicum ATCC 13032 (accession no. NP_601975), respectively. ORF 120 showed 73% identity to a metal transporter ATPase of S. coelicolor A3 (2) (CAB96031), and ORF 85 had 58% identity to a H⁺-stimulated manganese transporter, MntH, from Bifidobacterium longum NCC2705 (accession no. NP_695458).

Sequence analysis also indicated that ORFs 134, 167, and 178 encoded homologs to the TraA proteins of plasmid p103 from Rhodococcus equi ATCC 103 (accession no. NP_066783) with 28, 31, and 34% identity, respectively. In addition, ORF 143 had 30% identity with a TraG homolog from Streptomyces avermitilis MA-4680 (accession no. NP_825686), and ORFs 146 and 165 had 26 and 87% identity to TrbL (accession no. CAD47985) and soj (ParA) (accession no. CAD48001), respectively, from A. nicotinovorans. While each of these genes have occasionally been found on the bacterial chromosome, their presence together on the same 160-kb cloned region suggests that the region might be part of a large catabolic plasmid.

Plasmid localization of the 160-kb region.

Based on the above data, we sought physical evidence for the plasmid localization of the sequenced atrazine-catabolic genes, metal resistance genes, and genes possibly involved in plasmid transfer. Because previous attempts using conventional methods failed to reveal a plasmid in this bacterium, we used a more gentle and rigorous lysis procedure that was developed to isolate large plasmids in many bacteria (1). Plugs containing high-molecular-weight Arthrobacter DNA were prepared as described above, incubated at 37°C for 30 or 45 min with 1 U of Aspergillus oryzae S1 nuclease (Invitrogen), and loaded into wells of a Bio-Rad DRII contour-clamped homogeneous electric field (CHEF) gel containing 1.0% agarose. Plasmids were separated by electrophoresis at 200 V for 22 h at 14°C, with switch times ramped from 1 to 40 s. Gels were blotted and hybridized to single-stranded radiolabeled DNA probes (Rediprime II kit; Amersham, Piscataway, N.J.) prepared from PCR products of the trzN (17), atzB (5), and soj (ParA) (primers sojF 5′-ATGATCCTGCGCGACGTGAG-3′ and sojR 5′-CGGCACCTTCAAGTTCGTGG-3′) genes.

This method revealed that A. aurescens strain TC1 contained two large plasmids, pAA1 and pAA2, of approximately 380 and 290 kb, respectively (Fig. 3). Southern hybridization analyses done with trzN, atzB, and soj (ParA) as probes indicated that all three genes are located on pAA1. In addition, sequence analysis revealed that the 160-kb region contains merA, merR, arsR, soj, parB, and traA gene homologs, all of which have been found almost exclusively on plasmids, and regions of pAA1 had significant homology to pADP-1 from Pseudomonas sp. strain ADP and pAO1 from A. nicotinovorans (11). In other studies, catabolic genes in Arthrobacter sp. strains have been found on plasmids (4, 7, 10, 26, 29, 34, 38, 40). In total, sequencing and physical analysis data strongly suggest that the 160-kb region cloned into BAC6O10 is part of a 380-kb plasmid in A. aurescens strain TC1 and that this plasmid contains the atrazine degradation genes trzN, atzB, and atzC.

FIG. 3.

Localization of atrazine degradation genes to plasmid pAA1 in A. aurescens strain TC1. (A) Plasmids were separated by CHEF gel electrophoresis after incubation of plugs with S1 nuclease. Lane 1, λ pulsed-field gel electrophoresis molecular size marker; lane 2, plug without S1 nuclease digestion; and lane 3, plug after digestion with S1 nuclease for 45 min. Plasmids pAA1 and pAA2 are indicated by arrows, and values in margins are in kilobase pairs. (B) Southern hybridization of atzB, trzN, and soj (ParA) gene probes to blotted CHEF gels. Probes used in hybridizations are indicated at the bottom of each autoradiogram.

Nucleotide sequence accession number. The complete nucleotide sequence of BAC6O10 was deposited in GenBank under accession number AY456696.