An arsRB resistance operon confers tolerance to arsenite in the environmental isolate Terribacillus sp. AE2B 122.

Terribacillus sp. AE2B 122 is an environmental strain isolated from olive-oil agroindustry wastes. This strain displays resistance to arsenic, one of the most ubiquitous carcinogens found in nature. Terribacillus sp. AE2B 122 possesses an unusual ars operon, consisting of the transcriptional regulator (arsR) and arsenite efflux pump (arsB) but no adjacent arsenate reductase (arsC) locus. Expression of arsR and arsB was induced when Terribacillus was exposed to sub-lethal concentrations of arsenate. Heterologous expression of the arsB homologue in Escherichia coli∆arsRBC demonstrated that it conferred resistance to arsenite and reduced the accumulation of arsenic inside the cells. Two members of the arsC-like family (Te3384 and Te2854) found in the Terribacillus genome were not induced by arsenic, but their heterologous expression in E. coli ∆arsC and ∆arsRBC increased the accumulation of arsenic in both strains. We found that both Te3384 and Te2854 slightly increased resistance to arsenate in E. coli ∆arsC and ∆arsRBC, possibly by chelation of arsenic or by increasing the resistance to oxidative stress. Finally, arsenic speciation assays suggest that Terribacillus is incapable of arsenate reduction, in agreement with the lack of an arsC homologue in the genome.

INTRODUCTION

Arsenic is a naturally occurring element classified as a class I carcinogen by the International Agency of Research on Cancer. Arsenic contamination is derived from both anthropogenic sources, such as industrial waste or agricultural practices, and natural sources, through processes such as dissolution of arsenic-containing minerals (Hettick et al. 2015). Inorganic arsenic can appear in different oxidation states: -3 (arsine, AsH3), 0 (metallic, As0), +3 (arsenite, AsO33−) and +5 (arsenate, AsO43−), depending on the environmental conditions. Arsenate is the predominant form in soil and surface water, while arsenite prevails under reducing conditions in anaerobic groundwater. Both forms exist in terrestrial and aquatic environments independently of the pH and the redox potential, since the redox chemical reactions between arsenate and arsenite are relatively slow (Zhu, Xue and Kappler 2017). Although arsenite is more toxic than arsenate, the reduction of the oxidized form into arsenite is part of one of the most widespread resistance mechanisms of bacteria to this metalloid (Rosen 2002; Paez-Espino et al. 2009). This reduction has also been described as a common process for solubilizing arsenic from minerals, which results in the contamination of water supplies (Zhu, Xue and Kappler 2017). Additionally, some prokaryotes, such as Chrysiogenes arsenatis, Shewanella sp. strain ANA-3 and Bacillus selenitireducens, can utilize arsenate as a terminal electron acceptor for anaerobic respiration, expressing a respiratory reductase (arrAB), which is also considered to be involved in arsenic release from anaerobic sediments worldwide (Ahmann et al. 1994; Dowdle, Laverman and Oremland 1996; Saltikov and Newman 2003; Tsuchiya et al. 2019; Mazumder et al. 2020).

Bacteria have developed a great variety of mechanisms to avoid the toxicity of arsenic, e.g. (i) peroxidation reactions by membrane lipids, (ii) minimizing the absorption of arsenate by the phosphate uptake system, and (iii) the arsenic microbial detoxification pathway, which involves the acr and ars operons that are widespread in bacteria (Paez-Espino et al. 2009; Ben Fekih et al. 2018; Dunivin, Yeh and Shade 2019; Garbinski, Rosen and Chen 2019). The minimal constituents of the ars operon are three genes organized into a single transcriptional unit: arsR, which encodes the arsenic transcriptional repressor; arsB, which encodes a membrane-bound arsenite permease that exports arsenite; and arsC, which encodes a reductase that converts arsenate into arsenite (Rosen 1999, 2002; Paez-Espino et al. 2009). arsB is widespread in the chromosomes of Actinobacteria, Firmicutes and Proteobacteria (Dunivin, Yeh and Shade 2019). More complex ars operons have also been described in bacteria, with additional genes encoding an ATPase (ArsA), an arsenite chaperon (ArsD), an organoarsenical oxidase (ArsH) and the arsenite methyltransferase (ArsM) (Qin et al. 2006; Paez-Espino et al. 2009; Chen, Bhattacharjee and Rosen 2015; Ben Fekih et al. 2018; Paez-Espino et al. 2020).

The addition of composted olive oil waste has been shown to stimulate soil microbial communities, reducing their stress and increasing their growth, activity and functional diversity. Addition of composted olive oil waste tends to reduce the solubility of trace elements in the amended soils, in general. However, in the particular case of arsenic, different studies have reported the mobilization of arsenic after the addition of olive oil waste compost to trace elements-contaminated soil (Clemente et al. 2015). Additionally, over the past few decades there has been a growing interest in utilizing biomass from trace element-contaminated lands to produce bio-energy (Dastyar et al. 2019). Terribacillus sp. AE2B 122 is an environmental strain isolated from olive-oil agroindustry wastes able to perform transesterification reactions of lipids (Escobar-Nino et al. 2014). This capacity can be used for the production of biofuels, and indeed, this strain was one of the top five biofuel producers found in that study. This bacterium is included in group 1 of the genus Bacillus, together with many other halotolerant and halophilic genera (Escobar-Nino et al. 2014). This genus was reported for the first time in 2007 (An et al. 2007). Until now, only the genome sequences of Terribacilus aidingensis MP602 (Lu et al. 2015) and Terribacillus saccharophilus (Kalinowski et al. 2018) have been available. Here, we show that Terribacillus sp. AE2B 122 displays resistance to arsenite and arsenate, but shows an unusual ars gene organization cluster, lacking an arsenate reductase. A homologue of the arsenite pump arsB is functional in Terribacillus sp. AE2B 122, and can promote resistance to arsenite in E. coli by reducing the intracellular arsenic accumulation.

MATERIALS AND METHODS

Strains, media and culture conditions

The arsenic-resistant bacterium, Terribacillussp. AE2B 122 was isolated from ponds containing liquid wastes located in an olive oil mill in Ecija (Sevilla, Spain) (Escobar-Nino et al. 2014). The Escherichia coli strains used and constructed in this study are listed in Table S1. Terribacillussp. AE2B 122 was grown in PYA (peptone 1%, yeast extract 0.5%, K2HPO4 0.1%, MgSO4.7H2O 0.02%) + 1.5% NaCl media at 30°C. Escherichia coli strains were grown in tryptone 10 g/L, yeast extract 5 g/L, NaCl 10 g/L (LB) media at 37°C. Tetracyclin, gentamicin and/or isopropyl β-d-1-thiogalactopyranoside (IPTG) were added when required.

Whole-genome sequencing of Terribacillus sp. AE2B 122

The sequencing was performed by BGI-HongKong Co., Ltd., using Illumina technology and a strategy of 2 × 90 bps. Sequences were first filtered to eliminate low quality sequences, and then the oligonucleotide sequences were eliminated. The SOAP de novo algorithm version 1.05 was used (Li et al. 2008). The results were assembled locally and optimized according to the paired ends and the overlapping relationships, mapping the readings to build contigs and supercontigs. In-house automatic annotation was carried out in our laboratory using the RAST server, which performs rapid annotations using subsystem technology (Aziz et al. 2008).

Phylogenetic analysis

A search using the Batch web CD-search Tool (Marchler-Bauer et al. 2010, 2016) was conducted to select the homologous Ars proteins to be included in the study. Multiple alignments were made with Clustal Omega (Sievers et al. 2011). Maximum-likelihood trees were generated using the IQ-Tree server (Trifinopoulos et al. 2016). Branch support analysis was evaluated by 1000 ultrafast bootstrap replicates and by the SH-aLRT test. ModelFinder, implemented at the IQ-Tree server, was used to find the free rate heterogeneity substitution model that best fit the alignments (Kalyaanamoorthy et al. 2017). The online tool iTOL was used to edit the constructed phylogenetic trees (Letunic and Bork 2016).

Heterologous expression of ars homologues in E. coli

To express the putative arsB and arsC homologues of Terribacillus sp. AE2B 122 in E. coli ars mutants under the control of the lac promoter, the plasmids pTe3675, pTe2733, pTe3384 and pTe2854 were constructed using specific primers (Table S2, Supplementary data). To construct pTe3675 and pTe2733, the corresponding genes were amplified by PCR from Terribacillus sp. AE2B 122 DNA using the primers MAO393 and MAO394 (for Te3675) and MAO398 and MAO399 (for Te2733). EcoRI-KnpI fragments were ligated into the vector pVLT31 (Cases, de Lorenzo and Pérez-Martín 1996) linearized with the same enzymes. Escherichia coli DH5α strain was transformed and selected on LB plates supplemented with tetracycline. To check the construct, we used the primers MAO160 and MAO173 (Te3675), and MAO160 and MAO173 (Te2733) for sequencing.

A similar strategy was followed for the construction of pTe3384 and pTe2854, using the primers MAO403 and MAO404, and MAO406 and MAO407, respectively. The PCR fragments were cloned into pIZ1016 (Martínez-Pérez et al. 2004) using EcoRI-XbaI sites. Positives clones were confirmed by sequencing using the primers MAO160 and MAO404 (Te3384), and MAO160 and MAO407 (Te2854) primers.

Arsenic resistance assays in E. coli strains and Terribacillus sp. AE2B 122

Overnight cultures of Terribacillus sp. AE2B 122 and E. coli were adjusted to a final OD600 of 0.5. Ten µl of the final dilution were streaked onto PYA + NaCl solid medium supplemented with 10 mM arsenate or 5 mM arsenite, containing 10 µM IPTG, 1 mM IPTG or 1% glucose. Plates were incubated at 30°C for 1–3 days. PYA + NaCl plates without arsenic were used as the negative control.

Overnight cultures of Terribacillus sp. AE2B 122 and E. coli strains were diluted to a final OD600 of 0.05. Ten µl of each strain dilution were mixed with 140 µl of PYA + NaCl medium (Terribacillus sp. AE2B) or LB (E. coli strains) containing increasing concentrations of arsenate (2.5, 5 or 10 mM) or arsenite (1, 2.5 or 5 mM). Those cultures were grown in triplicate in 96-well microtiter plates at 37°C with shaking. Growth was monitored by measuring the absorbance at 600 nm every hour for 24 h. Strains growing in PYA + NaCl medium (Terribacillus sp. AE2B 122) or LB (E. coli) without arsenic were used as the negative control.

RT-qPCR analysis

Terribacillus sp. AE2B 122 was grown in PYA + NaCl and induced for 6 h with 0, 0.375 or 1.5 mM arsenate. Total RNA was isolated using Trizol method according to the protocol of the manufacturer and quantified with a NanoDrop (Thermo Fisher Scientific). To eliminate genomic DNA, RNA was incubated with RNase-free DNase (Promega) according to the instructions of the manufacturer. After treatment, the samples were adjusted to 12.5 ng μl−1 and employed in RT-qPCR reactions with One Step SYBR PrimeScript RT-PCR kit (Takara Bio Inc.). Primers used in the RT-qPCR analysis were designed using the software Primer Express (Table S3, Supplementary data). Three independent biological replicates were analyzed for each sample using technical triplicates.

Bacterial growth under oxidative stress

To check the resistance to oxidative stress, the E. coli cultures were adjusted to 107 cell ml−1, and cells were diluted to concentrations of 106, 105 and 104 cells ml−1 with distilled water, and spotted (2 μl) onto LB solid medium added with 1 mM IPTG and H2O2. The inoculated plates were incubated at 37°C for 2 days. LB plates without H2O2 were used as the negative control.

Accumulation of arsenic in E. coli

The E. coli strains W3110, WC3110, AW3110, AW3110/pTe2854, AW3110/pTe3384, AW3110/pTe3675, WC3110/pTe2854 and WC3110/pTe3384 were pre-cultured overnight. Then, 0.6 ml of each overnight culture was used to inoculate Erlenmeyer flasks (250 ml) containing 60 ml of LB medium supplemented with 1 mM arsenate or 3 mM arsenite. The inoculated flasks were incubated with continuous stirring (200 rpm) for 5 h at 37°C. After incubation, a volume of 15 ml of culture was filtered through nitrocellulose filters (0.45-μm pore diameter; Whatman) and the cells retained on the filters were washed twice with 0.05 M potassium phosphate buffer (pH 7.6). Filtering and washing were performed using a vacuum pump (Millipore 230V). The filters, which contained the retained cells, were treated with 10 ml of 65% nitric acid at 70°C for 120 min. Nitric acid (Sigma-Aldrich) was doubly distilled from 65% HNO3 reagent grade with a DST-1000 sub-boiling acid purification system from Savillex (Savillex, Eden Prairie, MN, USA) and diluted with Milli-Q water. Samples were analyzed in an Agilent 8800 ICP-MS/MS (Agilent Technologies) at the General Facilities for Research of the University of Sevilla (http://citius.us.es).

Arsenic speciation

Terribacillus sp. AE2B 122 was pre-cultured overnight. Then, 2 ml of each overnight culture was used to inoculate 10 ml of PYA + NaCl medium and further incubated with continuous stirring (200 rpm) for 90 min at 30°C. Arsenate or arsenite were added to a final concentration of 0.9 mM and incubation continued with continuous stirring (200 rpm) at 37°C for 0, 15, 30 and 60 min. Next, 1.5 ml of culture was withdrawn at each time point and centrifuged for 1 min at 20 238 g, separating cells (intracellular) and supernatant (extracellular). The cells were first washed with 1 ml PYA + NaCl and finally resuspended in 70% MeOH. Arsenic speciation in E. coli strains was essentially performed following the same strategy with the only difference that 1 mM IPTG was added to the cultures to induce the expression of Te2854 and Te3384. Intracellular and extracellular fractions were stored at -20°C. Supernatant samples were directly diluted in mQ water, filtered through a 0.22 µm membrane filter and analyzed for major arsenic species (arsenite [AsIII], dimethylarsinic acid [DMA], monomethylarsonic acid [MMA] and arsenate [AsV]) by high-performance liquid chromatography ([HPLC]; Agilent 1260 Infinity with a PSAC1-PRPX 100SAX 250 mm x 4.6 mm column) coupled to hydride generation atomic fluorescence spectrometry (HG-AFS; PS Analytical 10.055 Millennium Excalibur). MeOH extracts were sonicated for 20 min and homogenized by vortexing. Aliquots of 0.05 ml were taken for total arsenic determination after digestion with concentrated HNO3 (4 ml) at 70°C (2 h), taken to a final volume of 10 ml and finally filtered through 0.45 membrane filters before analysis by HG-AFS. The remaining MeOH extracts were centrifuged (7 min, 15 000 rpm) and aliquots of 0.05 ml were taken to a final volume of 1 ml with mQ water, and filtered through 0.22 membrane filters before arsenic speciation analysis by HPLC-HG-AFS. The species were separated using an isocratic mobile phase with 20 mM phosphate buffer at pH 6.25. The recovery of arsenate and arsenite from the cells was estimated to be 136% on average.

Nucleotide sequence accession numbers

The nucleotide sequence of the genome is deposited under accession number PRJEB33524 in the European Nucleotide Archive and NCBI (GenBank: CABVGR000000000.1). The nucleotide sequences of Te3675 (arsB) and Te3676 (arsR) were deposited under accession numbers MN068029 (BankIt2234525 ArsB) and MN068030 (BankIt2234525 ArsR), respectively, in the GenBank database (https://www.ncbi.nlm.nih.gov/genbank/).

RESULTS

Arsenic resistance in Terribacillus sp. AE2B 122

In order to test tolerance to arsenate and arsenite, Terribacillussp. AE2B 122 was grown in PYA + NaCl medium in the presence of different concentrations of these compounds and absorbance was monitored every hour for 24 h (Fig. 1). Terribacillus sp. AE2B 122 was more sensitive to arsenite than arsenate. This strain tolerated a concentration of 1 mM arsenite (Fig. 1A) and showed very slight growth at 2.5 mM arsenite. The tolerance to arsenate was higher, as is usually the case in other bacteria, with growth still observed at 10 mM (Fig. 1B).

Figure 1.

Arsenic resistance assays of Terribacillus sp. AE2B 122. Terribacillus sp. AE2B 122 was inoculated in PYA + 1.5% NaCl with the indicated concentrations of (A) arsenite, As(III) or (B) arsenate, As(V), at 37°C. Data shown are the mean of at least three independent experiments performed in technical duplicates ± standard errors of the mean (SEM).

Terribacillus sp. AE2B 122 genome sequencing and bioinformatic analysis

The genome sequence of Terribacillus sp. AE2B 122 was assembled into 91 contigs and 22 supercontigs using the SOAP de novo program, to give a total genome size of 3718 326 bps (accession number PRJEB33524). The genome was annotated automatically using the RAST server and found to contain a total of 3769 open reading frames (ORFs). Among these annotated ORFs we found eight homologues of genes related to arsenic resistance. Two of these ORFs are homologues of arsB (Te2733 and Te3675), two are homologues of arsC (Te3384 and Te2854) and four are homologues of arsR (Te2310, Te3140, Te3676 and Te3724). BLAST searches using several ars genes and proteins (arsH, arsM, arsR, arsC, arsB, arsD, arsI and arsP) from reference strains did not reveal additional ars homologues in the genome of Terribacillus sp. AE2B 122. In addition, we also found homologues of genes involved in resistance to several metals: (i) eight homologues of chromate resistance genes, such as chrA (Viti et al. 2014) and nitroreductases, which may have secondary chromate reductase activity (Viti et al. 2014); (ii) seven homologues of Co-Zn-Cd resistance genes, including zntR (Nies & Silver 1995) and czcD (Moore et al. 2005); (iii) five homologues of cobalt resistance genes, such as corA and corC (Smith et al. 1998; Sponder et al. 2010); and (iv) five homologues of copper resistance genes, including cutC (Gupta et al. 1995), csoR (Marcus et al. 2016), copA and copZ (Reyes et al. 2006).

Characterization of the arsenic resistance genes in the genome of Terribacillus sp. AE2B 122

Out of the eight ORFs annotated as ars genes in the genome of Terribacillus sp. AE2B 122, only one homologue of the repressor arsR (Te3676) and one homologue of the arsenite pump arsB (Te3675) appeared contiguous in the genome. No arsC homologue was found in their proximity, and consequently this species did not show the typical minimum operon arrangement (arsRBC) found in many organisms with resistance to this metalloid (Kaur, Kamli and Ali 2011). Te3675 and Te3676 are located in a 12 kb contig. RAST annotation predicted that the genes located upstream of Te3676 are a hypothetical protein (Te3677), a Na+/H+ antiporter (Te3678), another hypothetical protein (Te3679), a sodium/solute symporter family protein (Te3680) and a glycerate kinase (Te3681). The genes located downstream of Te3675 are predicted to be a pyridine nucleotide-disulfide oxidoreductase NADH dehydrogenase (Te3674), a transcriptional regulator of the AraC family (Te3673), a hypothetical protein (Te3672) and a short-chain dehydrogenase/reductase (SDR) (Te3671).

The arsR-like genes (Te2310, Te3140, Te3676 and Te3724) encoded predicted proteins showing the highest level of identity to ArsR family transcriptional regulators (Table S4, Supplementary data). All of them showed the typical motif HTH_ARSR identified in this family of repressor proteins. Te3676p showed the best hit, with 100% identity to a member of the ARSR subfamily of winged helix-turn-helix transcriptional regulators from T. saccharophilus (WP_095220794.1) (Table S4, Supplementary data). It also displayed a significant level of identity to the metal sensing ArsR-SmtB repressor from Bacillus subtilis (NP_388414.1; 52.8% identity) and a DNA-binding transcriptional repressor ArsR from E. coli K12 (NP_417958.1; 41.8% identity).

The two homologues of arsB (Te2733 and Te3675) contain conserved domains typical of ArsB proteins, which belong to the ArsB_NhaD superfamily of permeases (cI21473). This family includes plasma membrane arsenite-efflux transporter proteins that pump arsenite from the cytosol to the extracellular medium in different microorganisms (Garbinski, Rosen and Chen 2019). Te3675 encodes a predicted protein of 432 amino acids that showed high levels of identity to the predicted ArsB permease of T. saccharophilus (WP_095220793.1; 99.8% identity), and to functional ArsB permeases of different bacteria, such as Staphylococcus aureus pI258 (WP 000153629; 68.2% identity) and E. coli plasmid R773 (WP_063116263.1; 55.2% identity) (Table S4, Supplementary data). Te2733 encodes a predicted protein of 452 amino acids that showed a high level of identity to a predicted ArsB protein from T. saccharophilus (WP_095218693.1; 99% identity), and functional ArsB proteins of S. aureus pI258 (WP 000153629; 30.4% identity) and E. coli plasmid R773 (WP_063116263.1; 28.8% identity) (Table S4, Supplementary data), but also with a putative Na+/H+ antiporter from B. subtilis (NP_391484.1, 56.4% identity).

Members of the ArsB family of permeases typically present 8–13 transmembrane helices. Te3675p with 12 predicted transmembrane helices and Te2733p with 13 predicted transmembrane helices fall within this range. Fig. 2 shows a phylogenetic tree depicting the relationship of both ArsB proteins (Te3675p and Te2733p) with other bacterial permeases classified into the four subgroups identified in the ArsB_NhaD_permease superfamily (cI21473). Although both proteins (Te3675p and Te2733p) seemed to be more closely related to members of the subgroup cd01118 (ArsB_permease), only Te3675p clustered together with ArsB proteins from Gram positive and Gram negative bacteria of that subgroup. Te2733p formed an outgroup from this cluster and was more distantly related to the members of this subgroup of ArsB proteins (yet it did not cluster with the other proteins of other groups in the ArsB_NhaD_permeases superfamily). The closest relative to Te3675p found in this analysis was ArsB of S. aureus plasmid pI258 (AAA25637).

Figure 2.

Phylogenetic reconstruction of the ArsB arsenite transporter homologues. The tree shows the phylogenetic relationships of eight representative transporter proteins of the ArsB_NhaD_permease superfamily (cI21473) and the two putative ArsB homologues (Te3675 and Te2733) of Terribacillus sp. AE2B. Maximum-likelihood trees were generated using the IQ-Tree server. Branch support analysis was evaluated by 1000 ultrafast bootstrap replicates. ModelFinder was used to find the free rate heterogeneity substitution model that best fit the alignment and online iTOl was used for tree editing.

The automatic annotation of the genome sequence by RAST predicted the presence of two arsC homologues (Te3384 and Te2854) scattered in the chromosome of the bacterium, which contained sequences similar to protein domains of the cd02977 ArsC_family (subfamilies: cd03033 ArsC_15kD; cd03034 ArsC_ArsC; cd03035 ArsC_Yfb; cd03032 ArsC_Spx; and cd3036 ArsC-like) belonging to the thioredoxin (TRX)-like superfamily (cI00388) (Table S4, Supplementary data). The genomic context of both genes did not provide any useful information about a putative function for these genes. Phylogenetic reconstruction revealed that Te2854p clustered together with proteins classified in the cd03032 ArsC_Spx subfamily and Te3384p was closely related to a protein from Lactococcus lactis (NP 266944) included in the cd03036 ArsC_like subfamily. None of the hypothetical ArsC-homologues of Terribacillus clustered together with the bona fide E. coli ArsC proteins (chromosomal and plamid R773) (Fig. S1 and Table S4, Supplementary data).

Expression of arsenic resistance genes in Terribacillus sp. AE2B 122

RT-qPCR analysis was employed to quantify the expression levels of all the eight homologues of the ars genes in Terribacillus sp. AE2B 122 in the presence and absence of arsenate. Only Te3675 and Te3676 were inducible, showing an increase in expression when the concentration of arsenate in the culture media increased (Fig. 3). No induction was observed with the other homologue of arsB (Te2733) and the other homologues ofarsR(Te2310, Te3140 and Te3724). None of the arsC homologues (Te3384 and Te2854) were induced by arsenate (Fig. 3B). Analysis of absolute gene expression without normalization to the control samples (i.e. in the absence of arsenate) revealed that Te2854 was expressed 10–20-fold higher than Te3384 (Fig. S2, Supplementary data).

Figure 3.

Gene expression of the Terribacillus sp. AE2B 122 ars homologues in response to arsenate induction. Terribacillus sp. AE2B 122 was grown in PYA and induced for 6 h with 0, 0.375 or 1.5 mM arsenate. The expression of the (A) arsB homologues (Te2733 and Te3675), (B) arsC homologues (Te3384 and Te2854) and (C) arsR homologues (Te2310, Te3140, Te3676 and Te3724) was quantified by RT-qPCR and normalized to the expression of the housekeeping gene gyrA. Data shown are relative to gene expression in media without arsenate (n = 3). Differences were considered to be significant in a Student's t-test: P ≤ 0.01 (***), P ≤ 0.02 (**) or P ≤ 0.05 (*).

Complementation of an ars mutant of E. coli with the Terribacillus sp. AE2B 122 ars homologues

Heterologous expression experiments were performed to analyze the functional contribution of the putative arsB and arsC homologous genes from Terribacillus sp. AE2B 122 to arsenate and arsenite resistance. First, we performed growth tests of E. coli ars mutants expressing the putative transporters, reductases or the four different combinations of the transporters and reductases. Te3675, but not Te2733, complemented the ∆arsB mutation in E. coli AW3110 (∆arsRBC) and provided resistance to 1 mM arsenite and partial resistance to 2.5 mM arsenite (Fig. 4A-D). Similar results were obtained in the experiments performed on solid medium (Fig. S3, Supplementary data). This result is consistent with Te3675, but not Te2733, being inducible in Terribacillus sp. AE2B 122 (Fig. 3).

Figure 4.

Arsenic resistance assays of E. coli W3110 (WT), E. coli ∆arsRBC and E. coli ∆arsC strains expressing the arsB and/or arsC homologues of Terribacillus sp. AE2B 122. (A) E. coliW3110 (wild type; E. coli WT); (B) E. coli AW3110 (E. coli ΔarsRBC); (C-F) E. coli AW3110 (E. coli ΔarsRBC) expressing the arsB homologue Te2733 (C) or Te3675 (D) or a combination of Te3675 with the arsC homologues Te2854 (E) or Te3384 (F); (G-I) E. coli WC3110 (E. coli ΔarsC) expressing the arsC homologue Te2854 (H) or Te3384 (I). The strains were inoculated in LB medium with 1 or 2.5 mM arsenite, As(III) or 2.5 or 5 mM arsenate, As(V), at 37°C. Data shown are the mean of at least three independent experiments performed in duplicates ± standard errors of the mean (SEM).

Since no other homologues of arsC were found in the genome of Terribacillus sp. AE2B 122, we also decided to test whether Te3384 and Te2854 could confer arsenate resistance to an E. coli ∆arsC mutant. Both Te3384 and Te2854 seemed to confer a slight resistance to 2.5 mM arsenate to E. coli ∆arsC in liquid media (Fig. 4A, G-I). This observation was independently confirmed on solid media in the case of Te3384 (but not of Te2854), which allowed the E. coli ∆arsRBC mutant to grow in 10 mM arsenate (Fig. S3, Supplementary data). When both the arsB-homologue Te3675, and either Te3384 or Te2854, were co-expressed in E. coli ∆arsRBC (Fig. 4E-F), the results obtained were similar to the expression of either gene individually, i.e. recovery of wild-type growth in arsenite (1 and 2.5 mM), and a slight increase in resistance to 2.5 and 5 mM arsenate, accounting for 12–20% of the wild-type growth under the same conditions. Co-expression of Te2733 and Te3384 or Te2854 did not improve the resistance of E. coli ∆arsRBC to arsenite or arsenate compared with expression of the single genes (Fig. S4, Supplementary data). Together, these data suggest that Te3675 provides the E. coli ars mutant with resistance to arsenite, and that Te3384 and Te2854 provide partial resistance to arsenate that is independent of the arsenite pump.

Arsenic accumulation in E. coli ars mutants expressing the Terribacillus sp. AE2B 122 ars homologues

To fully characterize the arsB homologue Te3675, and the members of the ArsC-like family Te3384 and Te2854 of Terribacillus sp. AE2B 122, the capacity of different E. coli strains to accumulate arsenic was evaluated by exposing them to arsenite (expression of arsB homologue) or arsenate (expression of the ArsC-like family members). Expression of the arsB homologue Te3675 in the E. coli ∆arsRBC mutant reduced the accumulation of arsenic to wild-type levels when cells were grown in the presence of 3 mM arsenite (Fig. 5A). Together with the arsenic resistance assays in E. coli and the expression data in Terribacillus sp. AE2B 122, these data strongly suggest that Te3675p functions as an arsenite efflux pump.

Figure 5.

Total arsenic accumulation assays of E. coli mutant strains expressing the ars homologues. The wild-type strain (WT), and the ∆arsC and ∆arsRBC mutant strains with the vector plasmid (Vector) or plasmids expressing the indicated arsC (Te2854 and Te3384) or arsB (Te3675) homologues were grown in LB overnight, and then 3 mM arsenite (A) or 1 mM arsenate (B) were added. The strains were further incubated for 5 h. Accumulation of total intracellular arsenic was quantified by ICP-MS/MS. Data shown are the average of three independent experiments ± SEM. Differences were considered to be significant in a Student's t-test: P ≤ 0.01 (**), P ≤ 0.05 (*).

In order to test the effect of expressing the members of the ArsC-like family of Terribacillus sp. AE2B 122 Te2854 and Te3384 on arsenic accumulation in E. coli, they were first expressed in the ∆arsC mutant, with 1 mM arsenate added to the cultures. The E. coli strain expressing either Te2854 or Te3384 presented slightly higher levels of arsenic accumulation than E. coli ∆arsC harboring only the plasmid vector (Fig. 5B), and all accumulated more arsenic than wild type (E. coli W3110). Unexpectedly, when Te2854 and Te3384 were expressed in the E. coli ∆arsRBC mutant strain, arsenic accumulation increased 2–2.3-fold compared with levels in the control strain ∆arsRBC harboring the empty plasmid, and over 3-fold compared with the wild-type strain (Fig. 5B). These experiments, together with the lack of arsenic induction in Terribacillus, suggest that neither Te2854 nor Te3384 encode arsenate reductases, but rather that they help minimize the toxic effects of arsenic by other means, one of which could be the chelation of the metalloid (Fig. 5B).

Arsenate reduction in E. coli ars mutants expressing the Terribacillus Te3384 and Te2854

To fully characterize the members of the ArsC-like family Te3384 and Te2854 of Terribacillus sp. AE2B 122, the capacity of different E. coli strains to reduce arsenate was evaluated by exposing them to arsenate (Fig. 6). After 60 min of incubation of the E. coli wild-type strain with 1 mM arsenate, we did not observe changes in the accumulation of arsenite inside the cells. On the other hand, the ∆arsRBC mutant, which is incapable of pumping arsenite out of the cells, showed an increased accumulation of intracellular arsenite (Fig. 6A). This increased accumulation of arsenite was not observed in the ∆arsC mutant, which expressed the native arsenite pump, arsB. Expression of either Te3384 or Te2854 did not affect the levels of intracellular arsenite in either of the two mutant strains. Furthermore, the amount of arsenite in the supernatant of the cultures after 60 min of incubation with arsenate was 3.3–3.7-fold higher in the ∆arsC mutant and 5.8–6.5-fold higher in the ∆arsRBC mutant strains. Expression of the Terribacillus Te3384 and Te2854 genes in these mutants did not affect the amount of arsenite in the supernatant of the cultures (Fig. 6B). The most plausible explanation is that the arsenite found inside the cells and in the supernatants originates from ArsC-independent mechanisms, and therefore, taking all these results together, it can be concluded that Te3384 and Te2854 from Terribacillus sp. AE2B 122 are not arsenate reductases.

Figure 6.

Arsenic speciation in E. coli mutant strains expressing Te2854 and Te3384. Escherichia coli strains were grown overnight, and then 1 mM IPTG and 1 mM arsenate were added to the cultures. The strains were further incubated for 1 h, and samples were withdrawn. Arsenic speciation was performed from intracellular samples (A) and supernatants (B) to quantify arsenite. Data shown are the average of three independent experiments ± SEM.

Role of Terribacillus sp. AE2B 122 members of the ArsC-like family in resistance to oxidative stress in E. coli strains

Analysis of the domains of Te3384p and Te2854p of Terribacillus sp. AE2B 122 predicted that these proteins are members of the TRX-like superfamily (cl00388) and ArsC-like (arsenate reductase) or Spx regulator family (Table S4 and Fig. S1, Supplementary data). The thioredoxin-like superfamily is a large and diverse group of proteins containing a TRX fold that perform different redox reactions. To check whether Te3384 and Te2853 perform any anti-oxidative role under arsenic-derived oxidative stress, we assayed resistance to hydrogen peroxide. Surprisingly, the ∆arsRBC but not the ∆arsC mutant of E. coli was more sensitive to H2O2 than the wild-type strain. Interestingly, expression of either Te2854 or Te3384 removed this sensitivity (Fig. S5, Supplementary data).

Arsenic speciation in Terribacillus sp. AE2B 122

The data obtained strongly suggested that neither Te2854 nor Te3384 function as arsenate reductases in Terribacillus sp. AE2B 122, and yet this bacterium could tolerate at least 10 mM arsenate. However, since the genome sequence is not complete, we cannot discard the possibility that another member of the ArsC-like family exists in the genome of Terribacillus, and that it could play the alleged arsenate reductase role. Therefore, we investigated whether Terribacillus sp. AE2B 122 can reduce arsenate by performing arsenic speciation assays. Next, 1 mM arsenite or 1 mM arsenate was added to Terribacillus cultures and samples were taken at different time points. Arsenic speciation was performed in the intracellular fractions. When arsenate was added to the media, only arsenate could be detected inside the cells (Fig. 7A). However, when arsenite was added to the media, small amounts of arsenate were also detected inside the cells (Fig. 7B). Since the amount of arsenate did not change over the time course of the experiment and was equivalent to the control sample at time 0, we concluded that the arsenate was probably due to oxidation of arsenite to arsenate in the presence of ambient oxygen, and that it was likely to be present in the commercial powder. Arsenite was not pre-reduced before starting the experiment (Cánovas et al. 2003). To rule out the possibility that arsenate is reduced to arsenite, and immediately pumped out of the cell, we also analyzed the arsenite and arsenate content of the supernatant (Fig. 7C). Similar to what we observed inside the cells when arsenite was added to the media, residual amounts of arsenate were found in the supernatant, which remained constant over time. When arsenate was added to the media, only arsenate was detected in the supernatant. Taken together, these results suggest that Terribacillus sp. AE2B 122 is incapable of reducing arsenate to arsenite.

Figure 7.

Arsenic speciation in Terribacillus sp. AE2B 122. Terribacillus was grown overnight, and then 1 mM arsenate (A, C) or 1 mM arsenite (B, C) were added. The strains were further incubated for up to 1 h, and samples were withdrawn at the indicated times. Arsenic speciation was performed from intracellular samples (A, B) and supernatants (C) to quantify arsenite and arsenate inside the cells (A, B) or pumped out to the culture media (C). Data shown are the average of three independent experiments ± SEM.

DISCUSSION

Terribacillus sp. AE2B 122 is an environmental strain isolated from olive oil agroindustry waste. The resistance assays in liquid medium showed that this microorganism was resistant to up to 2.5 mM arsenite and at least 10 mM arsenate under our assay conditions (Fig. 1). To gain insights into the complex system underpinning the bacterial contribution to the biogeochemical cycling of arsenic, we analyzed the mechanism used by Terribacillus sp. AE2B 122 to tolerate this toxic metalloid. The bioinformatics analysis of Terribacillus sp. AE2B 122 genome revealed that there are eight ORFs that showed sequence homology to typical arsenic resistance genes: two arsB-like genes (Te2733, Te3675); two arsC-like genes (Te3384, Te2854); and four arsR-like genes (Te2310, Te3140, Te3676, Te3724). Interestingly, a global survey of arsenic resistance genes reported that many microorganisms harbor several copies of arsB, and in particular several Bacillus strains harbor two arsB copies of different clades (Dunivin, Yeh and Shade 2019). Indeed, previous reports showed that Pseudomonas putida harbors two ars operons, and that both are functional depending on the environmental conditions (Canovas, Cases and de Lorenzo 2003; Paez-Espino, Durante-Rodriguez and de Lorenzo 2015). Our bioinformatics analysis using the RAST server also revealed two genes in the genome of Terribacillus that display homology to arsB. However, several lines of evidence strongly suggest that only Te3675 is an arsenite efflux pump: (i) it appears in an operon together with a homologue of arsR (Te3676); (ii) it is induced in response to arsenate in Terribacillus sp. AE2B 122 (Fig. 3); (iii) it promoted arsenite resistance in an E. coli ∆arsRBC mutant (Fig. 4); and (iv) it reduced arsenic accumulation in E. coli ∆arsRBC grown in the presence of arsenite (Fig. 5). Therefore, this example suggests that functional analyses are essential to determine whether all arsB copies of a species are bona fide arsenite pumps. It has been suggested that an ancestral arsRB operon could have evolved in anaerobic environments during the early stages of the evolution of life, when arsenite was the predominant arsenic oxyanion (Rosen 1999). The existence of this minimal arsRB operon was first described in Ferroplasma acidarmanus, which presented an additional gene, arsA, located separately from the arsRB operon (Gihring et al. 2003; Baker-Austin et al. 2007). However, the functionality of this arsRB operon was not demonstrated. In F. acidarmanus, homologues of the arsenate reductase arsC were not found, indicating that additional, unknown arsenate-resistance genes exist for this organism, as our results indicate could be the case in Terribacillus sp. AE2B 122.

Neither Te3384 nor Te2854 were able to fully revert the arsenate sensitivity of a ∆arsC mutant, they did not reduce arsenate accumulation, and they did not reduce arsenate when they were expressed in E. coli ∆arsRBC and ∆arsC strains. Furthermore, their expression was not induced by arsenic. We cannot rule out the possibility that other genes displaying homology to arsC are present in the genome of Terribacillus sp. AE2B 122, as the genome sequence is not complete. However, additional arsenic speciation experiments revealed that Terribacillus sp. AE2B 122 does not reduce arsenate to arsenite (Fig. 7). Therefore, we can conclude that they are not bona fide arsenate reductases. However, both conferred some increased resistance to arsenate in E. coli Δars mutants, while accumulation assays showed that E. coli ΔarsRBC expressing Te3384 and Te2854 accumulated higher amounts of arsenic (Fig. 5B). Both Te3384p and Te2854p proteins contain predicted domains of the TRX-like superfamily as well as GSH-transferases (GST) and glutaredoxins (GRX). One hypothesis is that both Te3384 and Te2854 mediate formation of GSH-arsenate/arsenite complexes, resulting in increased arsenic accumulation in the cell, like the GRX or the GST reported in Pteris vittata (Sundaram et al. 2008) or E. coli(Chrysostomouet al.2015). Another hypothesis is that Te3384 and Te2854 contribute to tolerance to oxidative stress (see Fig S5, Supplementary Data). Both proteins contain Spx-related domains, and they appeared closely related to Spx members in the ArsC evolutionary relationship tree. GRX, GST and Spx functions have been associated with protection against oxidative stress (Sharma et al. 2004; Zuber 2004; Reder et al. 2008; Jozefczak et al. 2012; Berndt and Lillig 2017; Sundaram et al. 2008). A recent article also reports the role of arsH (a NADPH-dependent flavin mononucleotide reductase encoded in the ars operons of P. putida) in relieving the oxidative stress generated by arsenite and arsenate (Paez-Espino et al. 2020).

In conclusion, the environmental strain Terribacillus sp. AE2B 122 harbors an unusual and simple ars gene cluster. How does arsenic homeostasis work in Terribacillus sp. AE2B 122? Arsenite is taken up by the cell and subsequently pumped out of the cell by the ArsB efflux system, resembling what is happening in most bacteria. However, the absence of a bona fide arsenate reductase suggests that alternative, yet to be identified mechanisms must operate for resistance to arsenate. Te2854 and especially Te3384 seem to contribute to arsenate resistance, possibly by reducing the oxidative stress generated by arsenic and/or by the formation of arsenic complexes inside the cells.

Click here for additional data file.