Tolerance to arsenite, arsenate and antimonite were tested in O. tritici SCII24. This strain was able to grow in presence of As(III) up to 50 mM or up to 10 mM Sb(III), and also demonstrated high arsenate resistance, growing at concentrations higher than 200 mM (solubility level of arsenate in this medium).

A fragment carrying the arsB gene and another fragment containing the arsC gene were cloned from a gene library, and the sequences of these two gene clusters were determined using specific oligonucleotide primers. Five potential open reading frames (ORFs) were identified in operon ars1 and six ORFs were found in operon ars2 (Fig. 1). The characteristics of the predicted products of these ORFs are shown in Table 1. Based on the BLAST analysis, the first cluster is composed by four ORFs, which encode proteins that belong to families of well-known proteins involved in arsenic resistance, ArsR, ArsD, ArsA, ArsB and an additional gene coding for a CBS domain (Cystathionine-β-Synthase), usually not related with arsenic resistance. ORF1 (arsR1) encodes an arsenite regulatory protein (ArsR) with 109 aminoacids and shows clear homologies to ArsR proteins from different organisms. The ArsR1 from O. tritici has two cysteines (Cys42 and Cys45) instead of the typical highly conserved putative helix-turn-helix DNA binding motif (ELCVCDLC) well identified in plasmid R773 of E. coli (Fig. 2A). Shi et al. 33 indicated that arsenite only interact with two cysteines of this motif, therefore the presence of 2 cysteines should be enough for arsenic interaction and conformational change of the protein. The second gene, arsD encodes a protein of 124 amino acids, which is homologous of the ArsD proteins recently reported as metallochaperone 1112. Like others ArsD, this protein also has two pairs of vicinal cysteines, Cys12–Cys13 and Cys112–Cys113, normally involved in inducer binding. The next gene of the group, arsA comprises 582 aminoacids, and was described as encoding an ars anion-translocating ATPase (ArsA). ArsA consists of two homologous halves, N-and C-terminal domains connected by a short linker sequence. Each domain contains a consensus sequence (GKGGVGKT) for ATP binding and a metalloid binding domain (Cys113, Cys172, Cys422, His148, His451 and Ser420), in agreement with what it was observed for ArsA proteins of other organisms. The next ORF encodes a 175 aa protein with identity to CBS-binding domains of several families of proteins. The last gene of this cluster (arsB) encodes a putative arsenite efflux transporter ArsB, with 430 aminoacids.

The nucleotide sequence of the second cluster of the ars genes revealed six ORFs that corresponded to arsR2, arsC1, Acr3, arsC2, arsH and arsR3. The first gene of this group encoded a transcriptional regulator belonging to ArsR family without the typical CXCXXC arsenite-binding motif of pR773 ArsR (Fig. 2A). The next gene, arsC1, encoded a predicted protein of 137 aa, which is homologous of the ArsC protein family represented by E. coli pR773 (Fig. 2B), that uses glutathione and glutaredoxin as electron sources. The third gene was a putative arsenite carrier homologous to ACR3 protein and comprises 352 aa. The ORF4 of this cluster was a second arsC gene, encoding a protein (175 aa) with sequence similarity to the thioredoxin-dependent arsenate reductases related to the low molecular weight phosphatase family. In fact, this type of arsC comes from evolution of ancient tyrosine phosphatase protein. Downstream of the arsC2 gene was found a homologous of a gene previously identified as arsH. This gene was recently identified in a few ars operons, although its function remains unclear. Recently, published work demonstrates that ArsH protein is an atypical flavodoxin with a non-canonical FMN binding site that catalyzes oxidation of NADPH, generating H₂O₂ and with a low azoreductase activity 2425. The last ORF also showed similarity with a family of arsenic regulatory proteins and therefore it was named arsR3. Unlike all the other genes this protein is divergently transcribed. Alignments of the three different ArsR proteins (ArsR1, ArsR2 and ArsR3) revealed that ArsR2 was more similar to ArsR3 (36% identity and 55% similarity) than to ArsR1 (33% identity and 48% similarity). However, no significant similarity was found between ArsR1 and ArsR3 or between ArsC1 and ArsC2 proteins. "Neural Network Promoter Prediction" 34 was used to predict the location of the putative ars promoters. In ars1 operon one motif, predicted as a promoter, was identified upstream of arsR1 gene. As shown in Fig. 1, the putative -10 region (TATTAT) was separated from the -35 region (TTGCCA) by 17 nucleotides. In ars2 operon, two motives estimated as putative promoters were found upstream and downstream of arsR2 and arsR3 genes, respectively. In this case, we also identified the putatives -10 (TATATA and TATCCT) and -35 (TTATCA and TTGTCG) promoter boxes separated by 17 nucleotides.

Before investigating the role of O. tritici ars genes in arsenic resistance, we examined which polypeptides were expressed in E. coli by using a polyacrylamide gel. When the transformant carrying the complete ars1 operon was cultivated in the presence of arsenite or antimonite, only one additional polypeptide was produced in large amount (Fig. 3). This protein identified by comparison as ArsA, had approximately 63 KDa that was consistent with the size predicted from the aminoacid sequence (62.6 KDa).

However, this protein was not found when the same clone was grown in the absence of arsenite or antimonite. Furthermore, all different clones without arsR1 gene cultivated in the presence or absence of these chemicals were always able to produce this ArsA protein, indicating that ArsR worked as a repressor protein in E. coli. In contrast, all of the others ars products (ArsR1, ArsD, CBS domain, ArsB, ArsR2, ArsC1, ACR3, ArsC2, ArsH and ArsR3) were not found in the total proteins of E. coli.

Different constructs of the ars genes were used to evaluate the role of these genes in arsenic resistance in E. coli AW3110. All experiments were done in the presence of 0.5 mM IPTG as inducer and the cultures were analyzed for their ability to grow on increasing concentrations of arsenite, arsenate and antimonite (Fig. 4). Cells carrying the constructs from ars1 operon parsRDAcbsB, parsDAcbsB, parsAcbsB and parscbsB showed equivalent levels of resistance to As(III) (Fig. 4A). On the other hand, resistance to arsenite was not found in cells carrying the empty vector, or containing the constructs with each of the following genes: arsA (parsA), arsB (parsB), cbs (pcbs), arsA and cbs (parsAcbs).

The experiments related to arsenite resistance by ars2 operon, showed that this second operon also conferred As(III) resistance to E. coli cells, but at a lower level than the resistance conferred by operon ars1. Comparing the different constructs of ars2 operon (Fig. 4B), the clone with all genes (parsR2C1Acr3C2HR3) showed the highest As(III) resistance. Cells carrying the constructs parsR2C1Acr3C2H, parsAcr3C2HR3 and parsAcr3C2H showed equivalent levels of resistance to As(III). The constructs parsAcr3C2 and parsAcr3 conferred similar resistance, although at a lower level than the previous constructs and this was more evident for high As(III) concentrations. When the experiments were conducted in the presence of arsenate, we observed that only the cloned ars2 genes increased the level of resistance of the host strain. Therefore, abilities of parsR2C1Acr3C2HR3 and subclones to confer resistance to arsenate were tested (Fig. 4C). All the combinations of the multiple ars genes tested conferred similar levels of resistance to As(V). Constructs containing the individual genes did not confer resistance to the host strain except for cells that expressed the arsC gene, which showed a slight resistance to this oxyanion. These experiments are in agreement with the notion that arsCs encode for arsenate reductases, involved in As(V) detoxification. The ars1 operon was also responsible for antimony resistance, but the ars2 operon was not able to confer resistance to this oxyanion. The figure 4D shows that clones, parsDAcbsB, parsAcbsB and parscbsB conferred Sb(III) resistance to E. coli cells, as well as the construct parsRDAcbsB although at an intermediary level. Clone carrying only arsB grew less than the wild type in the absence of metal and was not resistant to arsenite. One possibility is the toxicity of the ArsB protein in E. coli 9 since in the absence of IPTG, clone carrying only parsB grew as well as the will type (data not shown). On the other hand, parscbsB did not induce toxicity in E. coli and clone was arsenite resistant. It is possible that the co-transcription of CBS is required for a correct function of ArsB. The cbs gene encodes a protein with a CBS domain that probably dimerizes to form a stable globular structure with ArsB 35.

RT-PCR was used to test if each ars gene cluster forms a unique transcriptional unit. RT-PCR products related to all ars1 genes were obtained showing that genes were transcribed from one independent mRNA (Fig. 5). The RT-PCR experiments for ars2 also showed products for all genes demonstrating that all genes were transcribed from the same operon (data not shown).

Specific probes for structural ars genes (arsA, cbs domain, arsB, arsC1, Acr3, arsC2 and arsH) were designed to determine the presence of the ars genes in other strains belonging to the genus Ochrobactrum. Total DNA from these strains was analyzed by Southern blot (Fig. 6). The genes arsA, arsB and the gene coding for the CBS domain were only detected in strain O. tritici SCII24 and the arsC1 gene was detected in both strains of O. tritici (SCII24 and 5bvl1). Using the Acr3 probe a hybridization signal was obtained from preparations of O. tritici (SCII24 and 5bvl1) and type strain O. anthropi. Southern blot experiments performed with arsC2 and arsH probes yielded a signal for all of the Ochrobactrum strains tested except for type strain O. intermedium.

The strains and plasmids used in this study are listed in Table 2. E. coli XL1-Blue and E. coli AW3110 were used as hosts for the cloning vectors or expression vectors. E. coli was grown in LB medium containing, per litre, 10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl. Ampicillin (100 μg/ml) and isopropyl-thio-ß-D-galactoside (IPTG) (0.5 mM) was added to the medium, as required. For some experiments, cells were grown in a low-phosphate medium.

This medium contains in g/l, Tris 6.06 (pH 7.0), NaCl 4.68, KCl 1.49, NH₄Cl 1.07, Na₂SO₄ 0.43, MgCl₂.6H₂O 0.2, CaCl₂.2H₂O 0.03, Na₂HPO₄.12H₂O 0.23 and glucose 5.0. Analytical-grade salts of NaAsO₂, KH₂AsO₄ and C₄H₄KO₇Sb.0.5H2O (Sigma) were used to prepare 0.5 M stock solutions. Metalloid-containing medium was solidified with 20 g of Bacto-Agar (Difco Laboratories) per litre.

The following degenerate pairs of 5 oligonucleotides were designed based on conserved regions of several prokaryotes arsenic resistance genes, arsB and arsC, available from public databases: forward primers 5'-GTG/CATC/TTGGCAA/GCCG/CAAA/GGG-3' and 5'ACG/CATC/TTAC/TCAC/TAAC/TCCG-3'; reverse primers 5'GTG/CGGCATA/GTTA/GTTCATA/GAT-3' and 5'TCGCCA/GTCC/TTCC/TTTG/CGTA/GAA-3', respectively. PCR fragments of 0.97 Kb for arsB gene and 0.35 Kb for arsC gene were amplified from O. tritici with these primers (purchased from Sigma Genosys). These fragments were subsequently used as probes for the isolation of the complete gene from the library. Total DNA was isolated from O. tritici as previously described 42. Two partial genomic libraries from this strain, containing DNA fragments selected by Southern blot analysis of genomic DNA with the 0.97 Kb and 0.35 Kb fragment probes, were obtained by complete digestion of total DNA with restriction enzyme HindIII (Roche Molecular Biochemicals), followed by ligation of purified fragments of around 5–9 Kb into the HindIII site of pUC18 (Invitrogen) and subsequent transformation of E. coli XL1-Blue with the ligation mixture. Clones were selected for their resistance and positive clones were confirmed by colony hybridization with a digoxigenin-labeled probe (Roche Molecular Biochemicals). Inserts of positive pUC18 clones and pGEM-T Easy (Promega) clones were sequenced by using vector-and insert-specific oligodeoxynucleotide primers. DNA sequencing was performed by using a BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems) using a model ABI310 automatic DNA sequencer. Nucleic acid and protein sequence analyses were performed with BioEdit editor 43. Comparison searches were performed by using the BLAST program of the National Center for Biotechnology Information (NCBI).

The ars genes were amplified by PCR using specific forward and reverse primers containing additional XbaI and HindIII recognition sites, respectively. The PCR products were purified after digestion with XbaI and HindIII and ligated into the corresponding sites into the pTRC 99A expression vector (Amersham Biosciences). Then, each construction was transformed into competent E. coli AW3110. Cells bearing each plasmid were grown at 37°C overnight in LB medium containing ampicillin. Overnight cultures were diluted 100-fold into fresh LB medium containing ampicillin alone or with arsenite, or with arsenate, or with antimonite. When the absorbance at 600 nm reached 0.8 to 1.0, 0.5 mM IPTG was added to the culture to induce gene expression, and cultivation was continued for 6 hours. Cells were pelleted by centrifugation and dissolved in a loading buffer for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). Total proteins were analyzed on 12% – 15% polyacrylamide gels.

Arsenite and antimonite resistance assays were carried out in LB medium, while for arsenate resistance, cells were grown in low phosphate medium. Overnight cultures were diluted 100-fold into fresh medium containing the ampicillin and different concentrations of arsenite, antimonite and arsenate. The bacterial suspensions were incubated at 37°C with shaking for 5 hours for experiments in LB medium or for 15 hours for assays performed in low phosphate medium and the absorbance at 600 nm was measured.

Total RNA was obtained from mid exponential phase O. tritici cells grown in low-phosphate medium exposed for 2 h to 1 mM of arsenite, antimonite or arsenate. Control cells were not exposed to added compounds. Total RNA isolated from those cells by the Rneasy Mini kit (Qiagen), according to the manufacturer's instructions, was then digested with RQ1 Rnase-free DNase (Promega) to remove the residual DNA. cDNA synthesis was done with Sensiscript Reverse Transcriptase (Qiagen) according to the manufacturer's instructions. Briefly, the reverse primers from arsB and arsH genes were used to create the cDNA for each ars gene clusters. cDNAs were used as template for the second PCR amplification reaction. For monitoring the expression of ars1, were used 5 μl of the RT reaction and specific primers designed to amplify a region spanning the junction of the different genes of operon ars1 and operon ars2. RT-PCR products were examined by 1% agarose gel electrophoresis. DNA contamination of the mRNA was determined by PCR using Taq polymerase without reverse transcriptase.

Southern blot analysis was performed as described by Sambrook et al. 42. Purified DNAs (10 μg) of strains O. tritici SCII24, O. tritici 5bvl1, O. grignonense OgA9a, O. anthropi LMG 3331 and O. intermedium LMG 3301 were subjected to digestion with restriction enzymes and electrophoresed on agarose gel [0.8% (w/v)]. DNA was capillary transferred for approximately 16 h to a nylon membrane in 0.4N NaOH, 1 M NaCl buffer followed by neutralization. DNA probes for structural genes [arsA, cbs domain, arsB, arsC1, Acr3, arsC2 and arsH] were amplified by PCR using gene-specific primers, recovered from agarose gels and purified with Wizard SV Gel and PCR Clean-Up System (Promega). Each probe was labelled with dioxigenin-dUTP nonradioactive (Roche Molecular Biochemicals) and the subsequent pre-hybridization and hybridization with the membrane were performed at 45°C to 58°C using DIG High Prime DNA Labelling and Detection Starter Kit II, following the manufacturer's instructions. The membrane was washed twice with 2× SSC-0.1% SDS at room temperature and twice with 0.2× SSC-0.1% SDS for 15 min at 68°C by constant agitation. Membranes were autoradiographed after incubation at 37°C for 45 min.

The nucleotide sequences of the DNA fragments containing the ars1 operon and ars2 operon have been submitted to GenBank under the accession no. DQ490089 and DQ490090, respectively.