|Home | About | Journals | Submit | Contact Us | Français|
N,N-Dimethylformamide (DMF), a toxic solvent used in the chemical industry, is frequently present in industrial wastes. Plasmid pAMI2 (18.6 kb) of Paracoccus aminophilus JCM 7686 carries genetic information which is crucial for methylotrophic growth of this bacterium, using DMF as the sole source of carbon and energy. Besides a conserved backbone related to pAgK84 of Agrobacterium radiobacter K84, pAMI2 carries a three-gene cluster coding for the protein DmfR, which has sequence similarities to members of the LuxR family of transcription regulators, and two subunits (DmfA1 and DmfA2) of N,N-dimethylformamidase, an enzyme of high substrate specificity that catalyzes the first step in the degradation of DMF. Genetic analysis revealed that these genes, which are all placed in the same orientation, constitute an inducible operon whose expression is activated in the presence of DMF by the positive transcription regulator DmfR. This operon was used to construct a strain able to degrade DMF at high concentrations that might be used in the biotreatment of DMF-containing industrial wastewaters. To our knowledge, this is the first study to provide insights into the genetic organization and regulation as well as the dissemination in bacteria of genes involved in the enzymatic breakdown of DMF.
C1 compounds containing one or more carbon atoms but no C—C bonds are the products of decomposition of natural organic matter, or they may be introduced into the environment by human activity. N,N-Dimethylformamide (DMF) is one such compound that does not occur naturally. It is usually synthesized by a one-stage reaction of carbon monoxide with N,N-dimethylamine or by a two-stage reaction with methylformate and N,N-dimethylamine (9). DMF is produced in large quantities throughout the world for use in the chemical industry (mainly textile and pharmaceutical) as (i) a versatile solvent (because of its solubility in water, organic nature, and high dielectric constant), (ii) a solubilizing agent, or (iii) a reaction medium.
The worldwide consumption of DMF in 2001 was approximately 285,000 metric tons (19), and there is an increasing demand for this compound. DMF can be partially or totally recycled by distillation of mother liquors, and DMF-containing wastes can be disposed of by incineration. Nevertheless, because of its widespread use, DMF is commonly found in industrial effluents, leading to environmental pollution (5, 15).
DMF is readily absorbed following oral, dermal, or inhalation exposure. It is toxic to human beings and other organisms (with hepatotoxicity, embryotoxicity, teratogenicity, and possible carcinogenicity). Long-term exposure to DMF might also cause irreversible alterations of mitochondrial DNA (37). As the overall rate of chemical degradation of DMF is extremely slow, biodegradation might represent a viable alternative for the removal of this compound from industrial effluents and DMF-polluted sites. It has already been demonstrated that total biodegradation of DMF with the accumulation of no organic intermediates can be achieved using specialized methylotrophic bacteria (13).
So far, several methylotrophic bacteria that are able to grow using DMF as the sole source of carbon, energy, and nitrogen have been described. All but one (Mycobacterium strain TH-35 ) are members of the Proteobacteria. Among them are strains belonging to the genera Pseudomonas of the Gammaproteobacteria (13, 35), Alcaligenes of the Betaproteobacteria (15), and Ochrobactrum (45), Methylobacterium (43), and Paracoccus (43), three representatives of the Alphaproteobacteria. There are three Paracoccus species known to be capable of DMF degradation: (i) P. aminophilus and P. aminovorans, isolated from soils taken from a factory which had produced DMF for about 20 years (43), and (ii) Paracoccus sp. strain DMF, isolated from activated sludge of domestic wastewater treatment unit in India (41).
It is thought that biodegradation of DMF can proceed via two different pathways (13, 15). The first (used by some Pseudomonas strains) involves repeated oxidative demethylation of this compound to produce methylformamide and then formamide, which is further cleaved to ammonia and formate by formamidase. The second pathway, shown in Fig. Fig.1,1, depends on N,N-dimethylformamidase (DMFase), which splits DMF into formate and N,N-dimethylamine (DMA). DMA is further degraded by dimethylamine dehydrogenase to formaldehyde and methylamine, which is finally converted into ammonia and formaldehyde. This pathway is used by Alcaligenes sp. strain KUFA-1 (15), Ochrobactrum sp. strain DGVK1 (45), some Pseudomonas spp. (13, 35) and, as we will show in this report, by P. aminophilus JCM 7686.
The DMFases of Pseudomonas sp. strain DMF 3/3 (35) and Alcaligenes KUFA-1 (15) have been isolated and characterized. Both are α2β2-type enzymes composed of two light and two heavy chain subunits, but they differ in their sizes and enzymatic properties, such as specific activities and thermostability. These enzymes exhibit a narrow substrate spectrum, but besides DMF, the DMFase of Alcaligenes KUFA-1 also hydrolyzes substituted amides, such as N,N-dimethylacetamide and N,N-diethylacetamide (closely related to DMF), which are not viable substrates for the DMFase of Pseudomonas spp. The dmfA1 and dmfA2 genes that encode the light (α; 132 amino acids [aa]) and heavy (β; 762 aa) chains, respectively, of the Alcaligenes KUFA-1 DMFase have been cloned and sequenced.
In this study we analyzed Paracoccus aminophilus JCM 7686, which carries seven indigenous plasmids (pAMI1 to pAMI7) ranging in size from 5.6 kb to approximately 500 kb. Plasmid pAMI2 (18.6 kb) was the focus of a previous study in which its minireplicon pAMI209 was constructed, sequenced, and analyzed (8). We demonstrate here that pAMI2 contains genetic information which is crucial for the first step in the degradation of DMF. The identified gene cluster consists of three genes, which code for the two subunits of DMFase and a transcription regulator of the LuxR family. Altogether, these genes constitute an inducible operon. Our findings provide the first insights into the genetic organization, regulation of expression, and dissemination of genes involved in the enzymatic breakdown of DMF.
P. aminophilus JCM 7686 (43) was the host strain of the analyzed plasmid pAMI2. Other paracoccal strains used in this study were P. aminophilus JCM 7686R (a rifampin-resistant derivative of the wild-type JCM 7686), P. aminophilus ABW 7686 (JCM 7686R deprived of pAMI2), P. versutus UW225 (3), P. alcaliphilus JCM 7364R, and P. aminovorans JCM 7685R (8). Escherichia coli TG1 was used for plasmid construction. All strains were grown in Luria-Bertani (LB) medium (33) at 30°C (Paracoccus spp.) or 37°C (E. coli). Where necessary, the medium was supplemented with antibiotics at the following concentrations: kanamycin, 50 μg/ml; rifampin, 50 μg/ml; tetracycline, 1 μg/ml for Paracoccus spp. and 20 μg/ml for E. coli strains. The minimal medium used for propagating Paracoccus spp. was that described by Wood and Kelly (49), with N,N-dimethylformamide (1,500 to 50,000 mg/liter) or arabinose (0.2%) added as the carbon source. The Paracoccus spp. formed colonies on solid medium after 48 h of incubation. Plasmids used in this study are described in Table Table11.
Overnight cultures of Paracoccus spp. in synthetic media were harvested by centrifugation, and the cells were washed twice and resuspended in appropriate medium to obtain an A600 of ~0.05. Incubation was continued for a further 72 or 96 h, and the growth of the cultures was monitored by A600 measurement. In addition, viable cell counts were determined every 24 h by plating appropriate culture dilutions on LB agar plates.
DMF concentration in culture supernatants was determined by high-performance liquid chromatography (HPLC) using a Waters Alliance Integrity System 996 PDA with a C8 Hypersil MOS-1 column and UV detection at 196.9 nm. The flow rate was 0.4 ml/min. The mobile phase was 50 mM sodium dihydrogen phosphate (in double-distilled water) containing 0.5% acetonitrile (the pH of the eluent was 4.5).
The ammonia concentration in culture supernatants was determined colorimetrically by using the ammonium test (Spectroquant; Merck) with a Beckman DU-65 spectrophotometer.
Transformation of E. coli TG1 was performed according to the method of Kushner (23). For triparental mating, the E. coli TG1 donor strain carrying the mobilizable vector, E. coli DH5α carrying the helper plasmid pRK2013, and a suitable Paracoccus sp. recipient strain were mixed at a ratio of 1:1:2. A volume of 100 μl of this mixture was spread onto a plate of solidified LB medium and incubated overnight at 30°C. The bacteria were washed off this plate, and suitable dilutions were plated on selective medium containing rifampin (selectable marker for the recipient strains) and kanamycin and/or tetracycline to select for transconjugants. The plasmids carried by transconjugants were identified by screening colonies using a rapid alkaline extraction procedure (4) and agarose gel electrophoresis.
Plasmid DNA was isolated according to the procedure of Birnboim and Doly (4) and, when required, further purified by CsCl-ethidium bromide gradient centrifugation (33). Megaplasmid visualization was achieved by in-gel lysis and DNA electrophoresis according to the procedure of Hynes and McGregor (17). Common DNA manipulation methods were performed as described by Sambrook and Russell (33). DNA probes for Southern hybridization were labeled with digoxigenin (Roche). Hybridization and visualization of bound probes were carried out as recommended by the supplier.
Pairs of forward and reverse primers were used to amplify DNA regions carrying the PdmfR, PdmfA1, and PdmfA2 promoters, respectively: (i) LP3DMF and RP3DMF, (ii) LPDMF and RP2DMF, (iii) LP4DMF and RP4DMF (Table (Table2).2). For amplification of the dmfR gene, primers RP3DMF and LPDMF were used. Amplification was performed in standard reaction mixtures containing pairs of synthetic oligonucleotide primers, appropriate template DNAs, deoxynucleoside triphosphates, and Pfu polymerase (Qiagen) with the supplied buffer. A Mastercycler (Eppendorf) was used to perform the PCR thermocycle. Amplified DNA fragments were analyzed by electrophoresis on 0.8 or 2% agarose gels and, where necessary, purified using a Gel Out kit (A&A Biotechnology), digested with BamHI and EcoRI, and cloned in appropriate vectors.
Total RNA was extracted from a cell pellet of a P. aminophilus JCM 7686 culture (grown on minimal medium supplemented with arabinose or DMF as the source of carbon) using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Contaminating DNA was removed by a 30-min digestion at 37°C with 20 units of RNase-free DNase I (MBI Fermentas). First-strand cDNA was synthesized from 1 μg of total RNA using 1 pmol of primer CDNADMF (Table (Table2)2) and 200 units of SuperScript II reverse transcriptase (Invitrogen) in a reaction that was incubated at 42°C for 50 min. The cDNA products were then amplified in 25-μl PCR mixtures with 2.5 μl of the reverse transcription (RT) reaction mixture as the template using different combinations of primers LPDMF, RP2DMF, RP4DMF, LP4DMF, RP1DMF, LDMF7, and LDMF6 (Table (Table22).
In vitro-coupled transcription-translation was carried out using the E. coli S30 extract system for circular DNA as described by the supplier (Promega). Proteins synthesized in the reactions were radiolabeled by incorporation of [35S]methionine, separated on a 10% Tricine-SDS-PAGE gel as described by Schagger and von Jagow (34), and visualized by autoradiography.
β-Galactosidase activity in Paracoccus spp. was measured by the conversion of o-nitrophenyl-β-d-galactopyranoside into nitrophenol as described by Miller (27). Individual assays were repeated three times.
The nucleotide sequence of plasmid pAMI2 was determined in the DNA Sequencing and Oligonucleotide Synthesis Laboratory at the Institute of Biochemistry and Biophysics, Polish Academy of Sciences, using a dye terminator sequencing kit and an automated sequencer (ABI 377; Perkin-Elmer). Plasmids for DNA sequencing were prepared by (i) cloning restriction fragments of pAMI2 into compatible sites of vector pBGS18, (ii) digestion of the resulting plasmids with appropriate restriction endonucleases, and (iii) generation of unidirectional nested deletions within the cloned inserts using exonuclease III and S1 nuclease (ExoIII/S1 deletion kit; MBI Fermentas). Primer walking was used to verify the entire nucleotide sequence of pAMI2.
The nucleotide sequence of pAMI2 was analyzed using the ORF Finder program available at the National Center for Biotechnology Information (NCBI) website (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). Similarity searches were performed using the BLAST programs (1) provided by the NCBI (http://www.ncbi.nlm.nih.gov/BLAST). Conserved domains in DmfR-related proteins were identified with the Simple Modular Architectural Research Tool (SMART) program (http://smart.embl-heidelberg.de/) (24). A G+C plot was created using the program Artemis (32) with a window setting of 500 nucleotides. Helix-turn-helix (HTH) prediction was performed using the Helix-Turn-Helix Motif Prediction program (7) at the PBIL-IBCP Lyon-Gerland website (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_hth.html). Insertion sequences were annotated using the ISfinder database (http://www-is.biotoul.fr/is.html). Phylogenetic analysis was performed using the maximum parsimony method (programs SEQBOOT, PROTPARS, and CONSENSE in the software package PHYLIP, version 3.65) (11). The constructed trees were drawn using Tree View (29).
The nucleotide sequence of pAMI2 determined in this study has been annotated and deposited in the GenBank database with the accession number GQ410978.
P. aminophilus JCM 7686 is a methylotrophic bacterium that was isolated by Urakami et al. (43) from DMF-contaminated soil. The DMF degradation activity of this strain was examined by monitoring the decrease in DMF concentration in the culture supernatant during growth in minimal medium supplemented with this compound as the sole source of carbon and energy by using HPLC (Fig. (Fig.2).2). The observed degradation of DMF correlated with the increase in levels of ammonia (a product of DMF degradation [Fig. [Fig.1]1] in the medium) (reference 44 and data not shown). Interestingly, we found that loss of plasmid pAMI2 from the JCM 7686 cells completely abolished the ability to grow for the pAMI2-less strain (LDZ1) on minimal medium supplemented with DMF as the source of carbon and energy (pAMI2 was removed by incompatibility with its miniderivative, pABW1-REP+PAR+TA). This finding indicated a crucial role for pAMI2 in the degradation of this compound.
To identify the genetic module responsible for the degradation of DMF, the complete nucleotide sequence of pAMI2 (18,563 bp) was determined. A summary of the predicted open reading frames (ORFs) of pAMI2, including their position, the size of the encoded proteins, and their closest homologues, is presented in Table S1 in the supplemental material.
The core region of the plasmid comprises a replication system (which contains ORF2, encoding a putative protein with similarities to a number of plasmid replication initiation proteins) and three stabilization systems responsible for (i) active partitioning of plasmid molecules into daughter cells upon cell division (PAR, ORF3, and ORF4), (ii) recombinational resolution of plasmid multimers (MRS and ORF1), and (iii) postsegregational elimination of plasmidless cells from the bacterial population (ORF5 and ORF6) (Fig. (Fig.3).3). The in silico predictions concerning the role of the REP, PAR, and TA modules were confirmed by functional analysis in our previous study (8).
The transfer region of pAMI2 consists of three genes. The predicted proteins show similarity to TraA, TraG, and MobC proteins, encoded by many mobilizable plasmids. The predicted origin of conjugal transfer (oriT) of pAMI2 is between the traA and mobC genes, and its location and nucleotide sequence are conserved among a group of related traA-mobC-traG transfer systems of, for example, plasmids pSmeSM11a of Sinorhizobium meliloti Rm1021 (39), pAgK84 of Agrobacterium radiobacter K84 (accession no. NC_006277), and the linear chromosome of Agrobacterium tumefaciens C58 (14).
Database comparisons revealed that the backbone of pAMI2 described above is conserved in pAgK84 of Agrobacterium radiobacter K84, a plasmid responsible for the production of the antiagrobacterial compound agrocin 84 (21). Plasmids pAMI2 and pAgK84 carry related replication, partitioning, and transfer systems in synteny, which strongly suggests that they originate from a common ancestor replicon (data not shown).
An additional “load” of pAMI2, which is not present in pAgK84, consists of three ORFs (ORF15 to ORF17) placed in the same transcriptional orientation (Fig. (Fig.3).3). The predicted proteins of ORF15 (762 aa) and ORF16 (141 aa) show significant similarity to the large DmfA2 and small DmfA1 subunits of DMFase encoded by Alcaligenes sp. KUFA-1 (15), respectively, while that encoded by ORF17 (344 aa) shows somewhat less similarity to a number of LuxR-related transcription regulators. These three predicted genes are bordered by two putative insertion sequences (ISs). A search of the ISfinder database revealed that one of the elements, designated ISPam3 (contains ORF13 and ORF14), belongs to the IS407 group of the IS3 family, while the other, ISPam4 (contains ORF18 and ORF19), is a member of the IS427 group of the IS5 family (data not shown).
To confirm the role of the pAMI2 ORF15 to ORF17 cluster in degradation of DMF, it was cloned into broad-host-range vector pBBR1MCS-2 to produce plasmid pDMF. After the insertion of a tetracycline resistance cassette, the construct pDMF1 was introduced by conjugation into the pAMI2-less strain P. aminophilus ABW 7686 (resulting in the LDZ1 strain). Growth experiments were performed in minimal medium (50 ml) supplemented with DMF (1,500 mg/liter), and the kinetics of DMF biodegradation were monitored by HPLC. P. aminophilus LDZ1(pDMF1) was found to degrade DMF more rapidly than the parental strain JCM 7686 (Fig. (Fig.22).
The influence of high concentrations of DMF, as the sole source of carbon and energy, on the growth of P. aminophilus strains was also tested. The wild-type JCM 7686 was unable to grow in the presence of DMF at 3,000 mg/liter, whereas for LDZ1(pDMF1) an analogous “toxic” effect was observed only in the presence of a much higher concentration of 10,000 mg/liter (data not shown).
Mutational analysis was performed to investigate whether ORF17 (coding for a putative transcription regulator) is an integral part of the analyzed DMFase module. A short out-of-frame insertion (4 bp) was made at a PscI site within the coding region of ORF17 (in plasmid pDMF1) (Fig. (Fig.4A)4A) (mutation generated by PscI digestion and filling the protruding ends by the Klenow fragment of DNA polymerase I followed by DNA ligation), and the resulting construct, pDMF1ΔdmfR, was introduced into strain ABW 7686. HPLC analysis revealed that this mutation reduced the ability of the strain to degrade DMF by approximately 50% compared to the LDZ1(pDMF1). In contrast, disruption of ORF15 (coding for the large subunit of DMFase) in pDMF1 (pDMF1ΔdmfA2; mutation generated by digestion of pDMF1 with the AatII, blunt ending with T4 DNA polymerase, and DNA ligation) (Fig. (Fig.4A)4A) completely abolished the ability to degrade DMF (data not shown). These results strongly suggest a regulatory role for the putative ORF17 product.
To confirm that ORF15, ORF16, and ORF17 are expressed, the coupled in vitro transcription-translation system was used and three polypeptides, with molecular masses that are in good agreement with those predicted for the proteins encoded by the corresponding ORFs (38.3, 17.2, and 84.4 kDa, respectively), were produced (data not shown). Taking into account the sizes of these polypeptides as well as the aforementioned sequence similarities, plus the results of the mutational analysis, the three ORFs were designated dmfR (ORF17), dmfA1 (ORF16), and dmfA2 (ORF15).
As mentioned previously, DmfA1 and DmfA2 of pAMI2 are highly similar to subunits of the DMFase of Alcaligenes sp. KUFA-1. According to the sequence annotation (accession no. AB028874), the DMFase genes of the KUFA-1 strain are not accompanied by a dmfR homologue. However, detailed sequence analyses revealed the presence of an ORF placed in synteny with dmfR (upstream of the dmfA1 gene), which was not distinguished at the time of sequence deposition. This ORF (position 3098 to 4171 of accession no. AB028874) encodes a putative protein of 357 aa (predicted molecular mass, 40 kDa) which shares substantial similarity with DmfR (54% identity) and other LuxR-related proteins (data not shown).
To investigate whether the dmfR-dmfA1-dmfA2 module is conserved in other bacterial genomes, sequence database searches were carried out using BLASTP. DmfA1 and DmfA2 homologues are present in 19 bacterial strains representing the Proteobacteria (alpha and beta subgroups) and Actinobacteria (data not shown). Interestingly, three dmfA1-dmfA2 loci were identified at different locations in the chromosome of Verminephrobacter eiseniae EF01-2 (Betaproteobacteria). On the other hand, the genome of Sphingomonas wittchii RW1 (Alphaproteobacteria) contains four copies of the dmfA2 relatives, but only one is accompanied by the dmfA1 partner gene. This analysis also revealed that, besides P. aminophilus and Alcaligenes sp. KUFA-1, dmfR-like genes could be identified in only two other strains: Sphingomonas wittchii RW1 and Methylobacillus flagellatus KT (Alphaproteobacteria). In the latter case, the gene (luxR) is placed downstream of the dmfA2 gene, which is different from the arrangement of the other three loci.
Phylogenetic analysis separated the DMFases into two main subgroups, well supported by bootstrap values. All of the predicted DMFase modules encoding putative DmfR-like proteins are clustered in a single branch of the phylogenetic tree. Multiple alignment of the amino acid sequences of these proteins revealed that the area of highest similarity is within the C-terminal part of these proteins. Using the SMART database this region was found to contain a conserved HTH DNA-binding domain, characteristic of the LuxR family of transcription regulators (data not shown).
As shown in Fig. Fig.4,4, dmfR, dmfA1, and dmfA2 are transcribed in the same direction. The genes dmfR and dmfA1 are separated by a 31-bp-long DNA spacer, while those coding for the two subunits of the putative DMFase (dmfA1 and dmfA2) overlap by 4 bp. The positioning of these genes strongly suggests that they may be transcribed together as a single transcriptional unit (operon).
In order to analyze the promoter(s) located within the gene cluster, DNA sequences upstream of the dmfR, dmfA1, and dmfA2 genes were separately amplified by PCR and inserted into the mobilizable broad-host-range promoter probe vector pCM132 to generate transcriptional fusions with a promoterless lacZ reporter gene. The resulting constructs were introduced into P. aminovorans JCM 7685R (routinely used in our laboratory as a host strain for paracoccal plasmids), and β-galactosidase activity assays were used to examine promoter strength. The results indicated the presence of three weak promoters: (i) PdmfR, for the gene encoding the putative transcription regulator DmfR (and possibly for the entire operon) (β-galactosidase activity, 1.4 ± 0.3 Miller units [mean ± standard deviation), (ii) PdmfA1, for the dmfA1 gene (9.1 ± 0.8 Miller units), and (iii) PdmfA2, for the dmfA2 gene (4.7 ± 0.6 Miller units). In comparison, the β-galactosidase activity produced by P. aminovorans JCM 7685R carrying “empty” vector pCM132 was 0.3 ± 0.2 Miller units (negative control).
The aforementioned pCM132 derivatives were also tested in two other related hosts: P. alcaliphilus JCM 7364R and P. versutus UW225. In both cases, β-galactosidase activities similar to those of the corresponding P. aminovorans strains were obtained (data not shown).
To examine whether expression of the dmfR-dmfA1-dmfA2 genes can be induced by DMF, specific primers were designed for RT-PCR analyses using cDNA of P. aminophilus JCM 7686 as the template (Fig. (Fig.4A)4A) (see Materials and Methods for details). Total RNA of the strain JCM 7686 was isolated from cultures grown either methylotrophically on minimal medium containing DMF as the sole source of carbon and energy or on minimal medium containing arabinose instead of DMF. RT-PCR of the former template with (i) primers A and D, which are specific for a region that spans the dmfR and dmfA1 genes and (ii) primers E and G, spanning the dmfA1 and dmfA2 genes, as well as (iii) primer pairs specific for internal parts of the dmfR (B and C) or dmfA1 (E and F) genes resulted in amplified fragments of the expected sizes (Fig. (Fig.4).4). In contrast, with the latter template, PCR products were not detected with any of the primer pairs, which suggests that expression of these genes is completely repressed in the absence of DMF. Alternatively, the operon could be expressed, but at an extremely low level that is below the limit of detection by RT-PCR. Each of the primer pairs used in the RT-PCR was also used in standard PCRs with pAMI2 DNA (positive control) and RNA isolated from P. aminophilus JCM 7686 grown in the presence of DMF and treated with DNase (negative control). As expected, PCR products with the predicted sizes were obtained exclusively in the former case (Fig. (Fig.4B),4B), which confirms the specificity of the amplification reactions. Additional control RT-PCRs are illustrated in Fig. Fig.4B4B (E + H reactions).
These results also demonstrated that the dmfR, dmfA1, and dmfA2 genes are cotranscribed from the promoter PdmfR, which suggests that despite the presence of internal promoters this gene cluster constitutes an operon. Furthermore, the regulation of the operon seems to occur mainly at the level of transcription, which appears to be inducible by DMF.
The influence of the putative regulator DmfR on transcription driven by the three identified promoters was analyzed by using the pCM132 derivatives described above (carrying PdmfR, PdmfA1, or PdmfA2) together with a compatible plasmid, pBBR-dmfR, as a source of the DmfR protein (contains the dmfR gene under the transcriptional control of its native promoter). The plasmid pBBR-dmfR was introduced in trans with the individual pCM132 derivatives, and the activities of the promoters were tested in β-galactosidase assays. Control experiments were performed with plasmid pBBR-ΔdmfR, carrying a short out-of-frame insertion in the proximal part of dmfR, which failed to express functional DmfR protein.
The promoter activities were tested in cells grown in medium supplemented with either arabinose (repression of the DMFase operon) or DMF (induction of the system) as the sole source of carbon and energy. The host strain P. aminovorans JCM 7685R is able to use DMF for methylotrophic growth, but it does not contain a DMFase operon homologous to that of pAMI2, as was confirmed by DNA-DNA hybridization analysis (data not shown).
All three of the analyzed promoters exhibited low activities in the arabinose medium (corresponding to those previously determined for the individual pCM132 derivatives), which were not affected by the presence in trans of the wild-type dmfR gene (Fig. (Fig.5).5). This was as anticipated, since in the absence of DMF, transcription of dmfR is blocked (as shown by RT-PCR analysis), which precludes formation of the DmfR protein. In the medium containing DMF the operon promoter (PdmfR) was highly induced, but the induction was strictly correlated with the presence of pBBR-dmfR in the cells. Mutation of the dmfR gene (pBBR-ΔdmfR) completely abolished the observed effect, which confirms the regulatory role of the DmfR protein. Such an effect was not observed for the PdmfA1 and PdmfA2 promoters. Indeed, the transcription driven by these promoters was slightly downregulated in the presence of DMF (Fig. (Fig.55).
It may be concluded that DmfR (in the presence of DMF) acts as a positive regulator of expression of the DMFase operon from the PdmfR promoter. This points to a role for DMF as a specific inducer of DmfR.
DMF is an anthropogenic compound that does not occur in nature. It was synthesized for the first time in 1893 (20), and in a very short period of time it became widely used in the chemical industry as a versatile solvent. The release of this toxic compound into the environment as a result of human activity has caused the appearance of bacterial strains able to use DMF as the sole source of carbon and energy. The DMFase proteins do not show any significant similarities to other known proteins present in sequence databases. Therefore, their occurrence most probably represents an example of rapid evolutionary change which has led to the generation of novel proteins exhibiting unique properties.
The gene cluster responsible for DMF degradation examined in this study is located on the mobilizable plasmid pAMI2 (18.6 kb), which potentially enables its horizontal transfer among different, even distantly phylogenetically related hosts. Homologous genes are present in 18 bacterial strains from the Proteobacteria and Actinobacteria. The closest homologue of the operon was identified in Alcaligenes sp. KUFA-1 (15), but the specific genomic location of these genes (plasmid versus chromosome) has not been determined.
In a previous study, Hasegawa et al. (15) showed that DMFase is a tetrameric protein composed of two light and two heavy chain subunits encoded by dmfA1 and dmfA2 genes, respectively. Here, we found that the DMFase module of pAMI2 (and that of the strain KUFA-1) also contains a third component of the system: the dmfR gene coding for a transcription regulator of the LuxR family.
The majority of LuxR-type proteins are transcription activators, although some can act as repressors or have a dual role, depending on their conformation or site of action (40). Many of these proteins have been shown to bind specifically to N-acyl homoserine lactones (AHL; synthesized by a LuxI protein), which are secreted signal molecules involved in quorum sensing in a variety of Gram-negative bacteria. AHL modulate activity of the LuxR proteins (e.g., through conformational changes of the protein) and enable binding to specific promoter sequences with dyad symmetry (lux boxes), which affects the expression of the target genes. One of the best-studied examples is the activation by LuxR of the Vibrio fischeri luminescence operon, which occurs exclusively in the presence of sufficient concentrations of the signal molecule N-(3-oxohexanoyl)homoserine lactone (3-oxo-C6 HSL) (47). LuxR-type proteins have also been shown to regulate the expression of genes involved in (i) the degradation of acetoin, testosterone, and dibenzofuran, (ii) the synthesis of galactoglucan, antibiotics, and dyes, (iii) virulence, (iv) glycolysis, and (v) conjugal transfer of Ti plasmids (12, 16, 18, 26, 28, 30, 36, 42, 46).
In the present study we have shown, for the first time, that the LuxR-type protein DmfR can activate transcription of the DMFase operon. Detailed in silico analysis of the DNA region upstream of the dmfR gene permitted identification of the putative DmfR-binding site, whose structure and location are similar to the well-defined lux boxes (10). As shown in Fig. Fig.4,4, it is composed of perfect 7-bp-long inverted repeats (5′-TGGGTAAGTTTACCCA3′) placed in close proximity to the predicted PdmfR core promoter responsible for transcription of the entire operon. The −35 and −10 hexamers of this promoter, gcGACA and TAcAca, respectively (residues not matching the canonical E. coli promoter sequences, TTGACA and TATAAT, are shown in lowercase), are separated by 15 bp.
The AHL-binding domain is placed in the less-well-conserved N-terminal part of the LuxR-type proteins, while their C terminus is responsible for binding to DNA. The amino-terminal domain of DmfR does not share any sequence similarities with other members of this protein family, which suggests that it binds a unique inducer molecule. Despite the lack of sequence similarity, this region appears to form an α/β/α sandwich structure (data not shown), which is typical for TraR and other LuxR-type proteins (50).
In this study we found that activation of transcription of the DMFase operon by DmfR takes place exclusively in the presence of DMF, which suggests that this compound may act as an inducer. It is also possible that DmfR is activated by other compounds, such as DMA, which is a first product of the degradation of DMF (Fig. (Fig.1).1). Besides PdmfR the DMFase operon was shown to contain two very weak internal promoters, which may ensure constitutive transcription of the dmf genes at a very low level in the absence of DMF. It may be speculated that a basal level of DMFase in bacterial cells might be necessary for the conversion of the small fraction of DMF into DMA, which could, in turn, activate DmfR and the entire operon. An analogous type of regulation has been shown for the lactose operon, where 1,6-allolactose (a product of lactose breakdown) binds with high affinity to the LacI repressor, which “induces” the repressor protein by lowering its affinity for the operator sequence (48). The identification of the DmfR inducer molecule, which is essential to the understanding of the DMFase operon, is an immediate goal of our future studies.
In our study we focused our interest on the first step of degradation of DMF (Fig. (Fig.1).1). The genomic locations of other genes of the pathway (coding for dimethylamine dehydrogenase and methylamine dehydrogenase) were not determined. We have obtained the nucleotide sequences of three plasmids of the strain JCM 7686: pAMI2 (18.6 kb), pAMI3 (5.6 kb), and pAMI7 (20.5 kb), and we know that none of the plasmids contains these genes. Therefore, the genes might be placed within a chromosome(s) or megaplasmids.
It is important to emphasize that the operon identified in this study has great potential as a functional cassette for the construction of novel plasmids and strains to be used for the bioremediation of DMF-contaminated environments. The P. aminophilus strain LDZ1(pDMF1) constructed in this study was able to degrade DMF much more efficiently than the parental wild-type strain JCM 7686, and it represents a good starting point to examine the application of this operon in bioremediation. Considering the environmental release and application of the modified strain, we performed a plasmid stability assay, which revealed that pDMF1 is very stably maintained in bacterial cells during growth under nonselective conditions (no pDMF1-less cells were detected after approximately 60 generations of growth without antibiotic selection). It seems likely that this catabolic plasmid can be stably maintained even after deletion of the antibiotic resistance cassettes, which would prevent dissemination of the resistance genes in the environment.
This work was supported by (i) the University of Warsaw, Poland (grants BW1720/16, BW1755-17, and BW179-113) and (ii) the State Committee for Scientific Research, Poland (grants no. 2 P04A 073 29 and PBZ-MNiSW-04/I/2007).
We acknowledge A. Sklodowska for enabling the performance of HPLC analyses in her lab.
Published ahead of print on 29 January 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.