Arsenic Methylation and Volatilization by Arsenite S-Adenosylmethionine Methyltransferase in Pseudomonas alcaligenes NBRC14159

Abstract

Inorganic arsenic (As) is highly toxic and ubiquitous in the environment. Inorganic As can be transformed by microbial methylation, which constitutes an important part of the As biogeochemical cycle. In this study, we investigated As biotransformation by Pseudomonas alcaligenes NBRC14159. P. alcaligenes was able to methylate arsenite [As(III)] rapidly to dimethylarsenate and small amounts of trimethylarsenic oxide. An arsenite S-adenosylmethionine methyltransferase, PaArsM, was identified and functionally characterized. PaArsM shares low similarities with other reported ArsM enzymes (<55%). When P. alcaligenes arsM gene (PaarsM) was disrupted, the mutant lost As methylation ability and became more sensitive to As(III). PaarsM was expressed in the absence of As(III) and the expression was further enhanced by As(III) exposure. Heterologous expression of PaarsM in an As-hypersensitive strain of Escherichia coli conferred As(III) resistance. Purified PaArsM protein methylated As(III) to dimethylarsenate as the main product in the medium and also produced dimethylarsine and trimethylarsine gases. We propose that PaArsM plays a role in As methylation and detoxification of As(III) and could be exploited in bioremediation of As-contaminated environments.

INTRODUCTION

Arsenic (As) is a toxic metalloid widely distributed in the environment (1). Arsenic is introduced into the environment from both geogenic sources and anthropogenic activities, such as mining, combustion of fossil fuels, and applications of arsenic-based pesticides or herbicides (2, 3). Arsenic poisoning affects more than 140 million of people worldwide (4), leading to health problems such as cancers, diabetes, and cardiovascular and kidney diseases (5, 6). Major sources of human As exposure are drinking water and food. In south Asia and parts of northern China, many people are exposed to high levels of As in drinking water (4, 7). Mining and irrigation of As-contaminated groundwater have also led to widespread soil contamination, resulting in elevated As accumulation in food crops (8, 9). In the United States, the U.S. Environmental Protection Agency ranks As first on its Superfund List of Hazardous Substances (http://www.atsdr.cdc.gov/SPL/index.html).

Arsenic is a redox sensitive metalloid that can also be methylated by different organisms. The As biogeochemical cycle involves various redox and methylation reactions (3). Arsenic methylation has been proposed to play an important role in As cycling among terrestrial, aquatic, and atmospheric environments (10, 11). It is thought that As methylation involves sequential transformation of inorganic As to mono-, di-, and tri-methylated species (12, 13). Methylated arsine species such as dimethylarsine [DMAs(III)H: (CH₃)₂AsH] and trimethylarsine [TMAs(III): (CH₃)₃As] are volatile and have been detected in the air above paddy soils (14). Therefore, As methylation can lead to volatilization of As (15). Many bacteria, archaea, fungi, and animals are able to methylate As. The biotransformation of arsenite [As(III)] to TMAs(III) by fungi was first documented in the 1890s (16). Of the arsenite-methylating bacteria, Rhodopseudomonas palustris (17), Streptomyces sp. strain GSRB54 (18) and Methanosarcina acetivorans C2A (19) have been studied in detail. The enzymes that catalyze As methylation, As(III) S-adenosylmethionine methyltransferases (ArsMs), have been purified and characterized from these bacteria. ArsM has also been studied in a number of eukaryotic organisms, including the human and several algal species (20–22). In the amino acid sequences of ArsM, four cysteine residues are highly conserved in different organisms and are likely to be involved in the catalytic function (23). Genes encoding ArsM have also been identified and sequenced in several microorganisms, showing certain genetic organization. For example, the arsM genes from Rhodopseudomonas palustris CGA009 and Methanosarcina mazei Go1 appear to be downstream of and regulated by the arsenical resistance operon repressor (ArsR) (24).

Pseudomonas species are widespread in both aquatic environments and soils and may play an important role in the As biogeochemical cycle (25–28). Some bacterial strains have been proposed to methylate As to different chemical species. Cheng et al. reported that a soil bacterial strain Pseudomonas sp. produced only arsine (AsH₃, i.e., no methylation) when incubated anaerobically with arsenate [As(V)] or As(III) (25). An arsenic-resistant Pseudomonas putida, isolated from a contaminated Chlorella culture, produced a nonvolatile trimethylarsenic species and other mono- and dimethylated species (27). It appears that a wide genetic diversity exists among organisms capable of arsenic methylation. Attempts have been made by researchers to design specific PCR primer sets for the detection of the arsM genes. More recently, degenerate primers based on the known diversity of arsM sequences from both prokaryotic and eukaryotic microbes have been designed for the purpose of identifying additional arsM genes (29). However, Pseudomonas arsM sequences were not included in the design of degenerate primers due to the lack of reference sequences from isolates in the genus (29, 30).

Given the importance of Pseudomonas sp. in the environment, the present study was conducted to investigate the As methylation ability of P. alcaligenes. A gene encoding the As(III) S-adenosylmethionine methyltransferase was identified in the draft genome of P. alcaligenes and functionally characterized. The role of ArsM in As(III) detoxification was also investigated.

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions.

The strains, vectors, and recombinant plasmids used in the present study are listed in Table 1. The lysogeny broth (LB) medium used here contains the following substances (in g liter⁻¹) at pH 7.0: yeast extract, 5.0; tryptone, 10.0; and NaCl, 5.0. The Pseudomonas strain and its derivatives were aerobically grown at 30°C in LB medium. The Escherichia coli strains were routinely cultured in LB medium at 37°C. Antibiotics were used at the following concentrations: for P. alcaligenes, chloramphenicol (Cm) at 34 mg liter⁻¹, and for E. coli, kanamycin (Km) at 50 mg liter⁻¹ and gentamicin (Gm) at 50 mg liter⁻¹. Kanamycin and gentamicin were prepared in sterile deionized water, whereas chloramphenicol was dissolved in ethanol.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Characteristicsa	Source or reference
Strains
P. alcaligenes WT	Cmr; wild type, arsenite-methylating strain; G−	Lab stock
E. coli
BL21(DE3)	F− ompT hsdS(rB− mB−) gal dcm lacY1(DE3)	Novagen
AW3110(DE3)	Arsenic-hypersensitive strain of E. coli lacking the chromosomal arsRBC operon	35
SM10λpir	thi thr leu tonA lacY supE recA::RP-4-Tc::Mu (λpir)	Lab stock
Plasmids
pET29a(+)	Kmr; expression vector	Novagen
pEX18Gm	Gmr; allelic exchange suicide vector sacB oriT(RP4) lacZ	47
pEXΔarsM	pEX18 carrying a partial PaarsM fragment	This study
pET-PaarsM	Kmr; NdeI-EcoRI fragment containing arsM inserted into pET29a(+)	This study

^aCm^r, chloramphenicol resistance; Gm^r, gentamicin resistance; Km^r, kanamycin resistance.

Time course experiment for As transformation by P. alcaligenes.

Arsenic was supplied as sodium arsenite (NaAsO₂) to investigate As transformation by P. alcaligenes. An aliquot of the seed culture (1% [vol/vol]) was inoculated into 50 ml of LB medium supplemented with 20 μM As(III) in a 250-ml Erlenmeyer flask. The cultures were incubated at 30°C and 180 rpm on a rotary shaker. The samples were collected at different time intervals for arsenic speciation analysis. Culture samples were centrifuged (5 min, 10,000 × g) at 4°C, and supernatant was passed through a 0.22-mm nylon membrane filter. Volatile methylarsines released from the culture were trapped in AgNO₃ impregnated filters according to the method of Yin et al. (21). Arsenic species in the solution phase were determined using high-performance liquid chromatography coupled to inductively coupled plasma mass spectrometry (HPLC-ICP-MS). Each treatment was performed in three replicates, and the control experiments without inoculation and substrate were carried out under the same conditions.

Construction of an arsM gene deletion mutant in P. alcaligenes.

To investigate the function of ArsM in P. alcaligenes, the arsM gene was disrupted through a single-crossover event. A 600-bp DNA fragment in the middle of arsM was generated by PCR using the genomic DNA of P. alcaligenes as the template and the primers KO-F and KO-R. The resulting product was cloned between SalI and EcoRI sites of the suicide plasmid pEX18 (31) to give pEXDA. pEXDA was delivered into strain P. alcaligenes from E. coli SM10_λpir via conjugal transfer, and the transconjugants were selected on LB plates supplemented with Cm and Gm. The PaarsM-disrupted mutant, designated PaΔarsM, was confirmed by PCR. The abilities of the mutant to tolerate and methylate arsenic were tested using a cell suspension assay as described below.

Comparison of As(III) resistance between PaΔarsM and WT strains.

Arsenic was supplied as sodium arsenite (NaAsO₂) at the indicated concentrations (0, 100, 300, 500, and 1,000 μM) to investigate As tolerance of PaΔarsM and wild-type (WT) cells. Both PaΔarsM and WT cells were grown to the early log phase and exposed to different concentrations of As(III) in triplicate. After 24 h, the cell density was determined at 600 nm using a UV-vis spectrophotometer (Shimadzu UV 2450). Log-logistic equations relating the relative growth rates to the logarithms of As(III) concentrations were established using the OriginPro8 program, and the As(III) concentration corresponding to 50% inhibition of the growth rate (EC₅₀) was calculated for both strains.

RNA isolation and RT-PCR.

An aliquot of the cells of P. alcaligenes was inoculated in 5 ml of LB medium with or without 20 or 40 μM As(III), each with three replicates. The cultures were incubated at 30°C and 180 rpm on a rotary shaker. After 6 or 12 h, the cultures were harvested by centrifugation (5, 000 × g, 10 min at 4°C). Total RNA was extracted using an RNA isolation kit (Omega, China) and treated with gDNA Eraser (TaKaRa, China) according to the manufacturer's instructions. A reverse transcription (RT) reaction was performed using a PrimeScript RT reagent kit (TaKaRa). Quantitative real-time PCR was performed in a Realplex2 system (Eppendorf, Germany) in a reaction mixture of 20 μl of SYBR green master mix (SYBR Premix Ex Tag TMII; TaKaRa Biotechnology) according to the manufacturer's instructions (TaKaRa Biotechnology). The conditions used were 95°C for 5 min, followed by 35 cycles of 95°C for 8 s, 58°C for 30 s, and 72°C for 30 s. Each quantitative real-time PCR assay was tested in a dissociation protocol to ensure that each amplicon was a single product. The 2^−ΔΔCT threshold cycle (C_T) method was used to calculate relative changes in gene expression (32). The gyrB gene (GenBank accession no. AB039388) was used as the internal control because its expression was unaffected by As(III). The primers used for quantitative RT-PCR are listed in Table 2. All PCR products were checked by electrophoresis and sequenced by Invitrogen (Shanghai, China) to confirm their identities.

TABLE 2.

PCR primers used in this study

Primer	DNA sequence (5′–3′)a	Purpose
GyrF	GGCGCTCTGGATATTGCA	Amplification of 171-bp gyrB gene by qRT-PCR
GyrR	GTTGAGGATCTTGCCCTTGAG	Amplification of 171-bp gyrB gene by qRT-PCR
ArsF	AGGACGAGGTGCTGTATGG	Amplification of 180-bp PaarsM gene by qRT-PCR
ArsR	GGTAGGTGGCCGAGTAGAA	Amplification of 180-bp PaarsM gene by qRT-PCR
KOF	ATACAGTCGACCTGCCGCGATAGATCACCGCCTGA	Amplification of 600-bp fragment of PaarsM for gene knockout
KOR	ATTGAATTCCCGGTCGTGACTGCTACGTGCTTGCG	Amplification of 600-bp fragment of PaarsM for gene knockout
AMF	CGGGGTACCGGCGCAGCTCAGCAGCGTATT	Amplification of PaarsM gene for expression
AMR	CCCGAGCTC CTCCAAACGCTGCGTTTGAAT	Amplification of PaarsM gene for expression

^aRestriction sites are underlined.

PaArsM purification.

For construction of plasmids for PaArsM expression and purification of PaArsM proteins, a 1.1-kb fragment containing the ATG start codon but excluding the stop codon was PCR amplified from the genomic DNA of P. alcaligenes using PrimeSTAR HS DNA polymerase (TaKaRa) with the primers AMF and AMR shown in Table 2. The PCR product was digested with NdeI and XhoI and ligated into pET-29a(+) to generate the recombinant plasmid pET-PaarsM, which was then transformed into E. coli BL21(DE3). Recombinant His-tagged PaArsM was purified using Ni-nitrilotriacetic acid affinity chromatography according to the methods described by Sambrook and Russell (33). The protein concentration was quantified with the Bradford method using bovine serum albumin (Sigma) as the standard (34).

Arsenic tolerance and transformation assays using recombinant E. coli cells.

Arsenic resistance assays were performed as previously described (17). Single colonies of the E. coli strain AW3110(DE3) (35) with deletion in the As tolerance operon (ΔarsRBC) and AW3110(DE3)pET-PaarsM were inoculated into 4 ml of LB medium containing 50 μg of kanamycin ml⁻¹ and incubated with shaking at 180 rpm at 37°C overnight. Cells at the late exponential phase were diluted by 100-fold into 50 ml of LB medium containing 50 μg of kanamycin ml⁻¹, 0.3 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG), and different concentrations of As(III) (0 to 1,000 μM, each with three replicates). Growth was determined as the optical density at 600 nm using a UV-vis spectrophotometer (Shimadzu UV 2450).

Arsenic methylation was measured in E. coli strains AW3110 (ΔarsRBC) bearing plasmid PaarsM or vector plasmid pET-29a(+). Cultures were grown overnight in LB medium containing 50 μg of kanamycin ml⁻¹ and then diluted by 100-fold into 50 ml of fresh LB medium containing 50 μg of kanamycin ml⁻¹. When the cells reached the exponential phase, 0.3 mM IPTG and 10 μM As(III) were added into the medium (three replicates). At a 12-h interval after As(III) addition, the culture samples were centrifuged at 10,000 × g for 5 min. Arsenic species were analyzed by HPLC-ICP-MS. Volatile methylarsines released from E. coli strain AW3110 expressing PaarsM were trapped in AgNO₃ impregnated filters according to the method of Yin et al. (21).

As(III) methylation assay using purified PaArsM protein.

Arsenic methylation in vitro was performed in triplicate in a phosphate-buffered saline (PBS; 50 mM phosphate [pH 7.5]) containing 3 μM PaArsM protein, 5 mM reduced glutathione (GSH) as the reductant, 0.9 mM S-adenosylmethionine chloride (SAM) as the methyl donor, and 10 μM As(III). Arsenic species during the reaction process was monitored during 0.5 to 12 h. To determine the optimal pH for As methylation by purified PaArsM in vitro, three different buffering systems were used: 50 mM citric acid-sodium phosphate (pH 3 to 5.3), 50 mM KH₂PO₄-NaOH buffer (pH 5 to 9), and 50 mM glycine-NaOH (pH 8.8 to 11). The reaction was maintained at 37°C for 12 h. For determination of the thermostability, As(III) methylation with 3 μM purified PaArsM was performed in triplicate in PBS (pH 7.5), containing 5 mM GSH, 0.9 mM SAM, and 10 μM As(III) at 20 to 60°C for 12 h (17). Arsenic species were analyzed with HPLC-ICP-MS.

Chemotrapping of volatile As.

Arsenic species volatilized by P. alcaligenes or by the purified PaArsM enzyme were trapped as described previously (21). Silica gel (0.5 mm diameter) was soaked in 5% HNO₃ overnight, washed with Millipore deionized water, impregnated with 10% AgNO₃ solution (wt/vol) overnight, and then dried at 70°C. The AgNO₃ impregnated silica gel was loaded into a 3-ml burette and held in by a small quantity of quartz wool at each end. The trap tube was covered with aluminum foil to avoid photodecomposition of AgNO₃ and connected to a gas outlet attached to Erlenmeyer flasks (250 ml) used for P. alcaligenes culture. The gas inlet was connected to a super silent adjustable air pump (ACO-008; 2W power) to ensure aerobiosis and gas exchange with the atmosphere. After 96 h, As species trapped by the silica gel were eluted in 1% HNO₃ using a microwave digestion system (CEM Microwave Technology, Ltd., Matthews, NC). The working program was as follows: 55°C for 10 min, 75°C for 10 min, and 95°C for 30 min, with 5 min of ramp time between each stage (8). The supernatant was filtered through a 0.22-μm-pore-size filter prior to As species analysis. In the assay with purified PaArsM, the assay solution was kept in a 20-ml vessel connected to a glass joint with an inlet of air from an air pump and an outlet connected to a silica gel trap tube.

Arsenic speciation analysis.

Arsenic speciation was determined by HPLC-ICP-MS (NexION 300X; Perkin-Elmer, USA) as described previously (8). Arsenic species were separated using an anion exchange column (Hamilton PRP X-100, 250 mm in diameter). The mobile phase consisted of 8.5 mM ammonium di-hydrogen phosphate (NH₄H₂PO₄) and 8.5 mM ammonium nitrate (NH₄NO₃) prepared in Millipore deionized water, and the pH was adjusted to 6.0 with ammonium hydroxide. The mobile phase was pumped through the column isocratically at a flow rate of 1 ml min⁻¹. Indium, prepared from a high-purity stock solution supplied by Perkin-Elmer, was added to the post-column solution and measured by ICP-MS as the internal standard. ICP-MS was set up in the He gas collision mode to minimize polyatomic interferences on m/z 75 (As). Arsenic species in the samples were identified by comparing their retention times with those of the standards, including arsenite [As(III)], arsenate [As(V)], dimethylarsenate [DMAs(V)], monomethylarsonate [MAs(V)], trimethylarsenic oxide [TMAs(V)O], and quantified by external calibration curves with peak areas.

RESULTS

Arsenite methylation by P. alcaligenes.

Biotransformation of arsenic by P. alcaligenes was measured in liquid medium containing 20 μM As(III) under oxic conditions. During the initial 48 h, the concentration of As(III) in the medium decreased and that of DMAs(V) increased rapidly (Fig. 1A and B). After 48 h, 79.4% of As(III) was methylated to DMAs(V) (Fig. 1A). There was no further increase in DMAs(V) concentration from 48 to 96 h in the medium. DMAs(V) was the major species, accounting for 79.4 to 83.8% of the total As in the medium. Monomethylarsonate [MAs(V)] was not detected in the medium during the experiment. In the As speciation method used, TMAs(V)O and As(III) coeluted. To separate the two As species, As(III) was oxidized to As(V) by H₂O₂ to allow the detection of TMAs(V)O (see Fig. S1 in the supplemental material). During the entire incubation time, small amounts of TMAs(V)O were detected in the growth medium (see Fig. S1 in the supplemental material), accounting for 0.44 to 2.04% of the total As in the medium (Fig. 1A). In a time-dependent fashion, the cells produced dimethylarsine [DMAs(III)H, Me₂AsH] and trimethylarsine [TMAs(III), Me₃As] gases in the headspace above the culture medium; these were oxidized to DMAs(V) and TMAs(V)O, respectively, by AgNO₃-impregnated silica gel in the gas collection filter. The amounts of TMAs(III) and DMAs(III)H volatilized were 58.9 and 880.4 ng of As, respectively (Fig. 1C), accounting for 1.3% of the total As(III) added to the assay solution.

FIG 1.

Biotransformation of arsenite by Pseudomonas alcaligenes NBRC 14159. (A) Arsenic species in LB medium with P. alcaligenes NBRC14159 exposed to 20 μM arsenite for the indicated times. Symbols: ●, As(III); ▲, DMAs(V); ▼, TMAs(V)O; ■, As(V). (B) HPLC-ICP-MS chromatograms of As species in the assay solution. Curve 1, standards; curve 2, 0 h; curve 3, 24 h; curve 4, 96 h. (C) Volatile As collected by AgNO₃ impregnated silica gel and quantified by HPLC-ICP-MS. P. alcaligenes was grown with 20 μM As(III) in LB medium for 96 h.

Cloning and functional analysis of the PaarsM gene.

Genome sequence data for P. alcaligenes NBRC14159 (GenBank accession number BATI00000000) was obtained from the National Institute of Technology and Evaluation. Using the arsM sequence from the eukaryotic photosynthetic alga Chlamydomonas reinhardtii and the photosynthetic prokaryote Rhodopseudomonas palustris, we identified an orthologous sequence in P. alcaligenes. The gene was designated PaarsM (GenBank accession number WP_021703038). PaarsM encoded a protein of 347 amino acids, with a predicted molecular mass of 38 kDa. The deduced protein was searched against the Protein Data Bank (PDB; http://blast.ncbi.nlm.nih.gov) using the BLASTP program. Sequence homology analysis showed that PaArsM has a similarity with the ArsM of Chlamydomonas reinhardtii (amino acid identity of 55%) and Rhodopseudomonas palustris (amino acid identity of 36%). We examined multiple alignments of deduced amino acid sequences of microbial ArsM from different species (Fig. 2). PaArsM contains four fully conserved cysteines at the residues 24, 52, 145, and 195 (indicated by triangles in Fig. 2), which are characteristic of all ArsM proteins investigated to date and are probably involved in As(III) binding (23, 36, 37). Three sequence motifs possibly involved in the interactions with S-adenosylmethionine (SAM) are underlined Fig. 2, showing both the conserved amino acid residues and differences between the different ArsM proteins.

FIG 2.

Multiple alignment of ArsM homologues. ArsM from Pseudomonas alcaligenes NBRC14159 (WP_021703038) was aligned with homologues from Rhodopseudomonas palustris (NP_948900), Chlamydomonas reinhardtii (AY286122), Cyanidioschyzon sp. strain 5508 (CmarsM7; FJ476310), Nostoc sp. strain PCC7120 (NsArsM; HQ891147), human (hA53MT; AAI19639), and rat (rA53MT; EDL94361). Underlining is used to indicate three conserved motifs that are involved in interactions with S-adenosylmethionine (SAM); triangles are used to indicate cysteine residues that are probably involved in As binding.

Transcription of the PaarsM gene.

Quantitative real-time PCR showed that PaarsM of P. alcaligenes was expressed in the absence of As(III). The expression was further enhanced by 40 μM As(III) at 6 h and by both 20 and 40 μM As(III) at 12 h (see Fig. S2 in the supplemental material).

PaArsM enzyme plays a role in As(III) detoxification.

To investigate the role of PaarsM in As(III) detoxification, a deletion mutation was created in PaarsM. No methylated As was detected in PaΔarsM samples in the presence of 20 μM As(III), indicating that PaArsM is specifically responsible for As methylation in the wild-type (WT) P. alcaligenes. The mutant and WT grew similarly in the absence of As(III). At 100 to 1,000 μM As(III), mutant cells grew significantly (t test, P < 0.01) less well than the WT (Fig. 3). Based on the 24 h growth data, the EC₅₀ was calculated as 0.55 and 0.20 mM for WT and PaΔarsM strains, respectively, suggesting that As(III) tolerance of the mutant was decreased by ca. 60% compared to the WT.

FIG 3.

Organization and functional analysis of the arsM gene of Pseudomonas alcaligenes NBRC14159. (A) Open arrows show gene orientations. The ars genes are shown in black; non-ars genes are shown in gray. (B) Growth profiles of the wild-type strain and the ΔarsM mutant in the presence of As(III). The black bar represents the wild-type strain; the gray bar represents the ΔarsM mutant. The results are averages of three independent assays. Asterisks denote significant difference between the WT and the mutant (∗∗, P < 0.01; ∗∗∗, P < 0.001).

To examine whether the ArsM confers As(III) resistance in E. coli, the PaarsM gene was expressed in E. coli strain AW3110, which is an As(III)-hypersensitive strain of E. coli lacking the chromosomal arsRBC and does not have an arsM gene (35). When grown on As(III)-free LB medium, no difference was observed between cells bearing vector plasmid and pET-PaarsM. At 50 or 100 μM As(III), strain AW3110 expressing PaarsM grew significantly (Student t test, P < 0.05) better than that bearing vector plasmid alone at the 8- to 20-h time points (Fig. 4), demonstrating that the gene product confers the tolerance to As(III).

FIG 4.

ArsM protein confers resistance to As(III) in E. coli strain AW3110 (ΔarsRBC). Absorbance was monitored at 600 nm. Expression of PaarsM confers tolerance to As(III), bearing vector plasmid pET29a(+) or pET29a-PaarsM with the indicated As(III) concentrations. Filled symbols, pET29a-PaarsM; open symbols, vector plasmid pET29a(+); circles, 0 μM As(III); squares, 50 μM As(III); triangles, 100 μM As(III).

As(III) methylation and volatilization in E. coli expressing PaarsM.

To confirm As(III) methylation activity, the recombinant cells (AW3110 bearing PaarsM) were grown in LB medium containing 10 μM As(III), and changes in the arsenic species in the medium were analyzed over 48 h (Fig. 5). The increases in the DMAs(V) concentration coincided with the decrease in the As(III) concentration in the medium. At 48 h, DMAs(V) accounted for 59.6% of the total As in the medium. Small amounts of MAs(V) were also detected in the medium. No methylated species were observed in the cells bearing vector plasmid pET29a(+).

FIG 5.

Biomethylation of arsenite by E. coli expressing PaarsM. Cells of E. coli strain AW3110(DE3) with plasmid pET29a-PaarsM were grown in LB medium in the presence of 10 μM sodium As(III) for up to 48 h. Arsenic species in the medium were determined by HPLC-ICP-MS. Symbols: ●, As(III); ▲, DMAs(V); ▼, MAs(V); ■, As(V).

In vitro As(III) methylation using purified ArsM protein.

Consistent with the predicted molecular mass, purified PaArsM had a molecular mass of ∼40 kDa (see Fig. S3A in the supplemental material). The purified enzyme showed As(III) methylation activity at pHs between 6 and 10, with highest activity being observed at pH 7.5, whereas no methylated species were formed between pH 3.0 and 5.0 (see Fig. S3B in the supplemental material). Purified PaArsM converted most of As(III) to DMAs(V) at 37°C and pH 7.5 (see Fig. S3C in the supplemental material). At higher temperatures, the rate of As methylation was reduced. Small amounts of MAs(V) were detected at temperatures between 37 and 60°C. At 60°C more MAs(V) than DMAs(V) was produced, but at this temperature the overall As methylation activity was very low.

Figure 6 shows the time course of in vitro As methylation by purified PaArsM protein maintained at 37°C and pH 7.5. Immediately after the addition of the protein to the assay solution, As(III) concentration decreased from the initial 10 μM to 6.5 μM, possibly due to the binding of As(III) to the protein (23). DMAs(V) was detected in the assay solution after 0.5 h (Fig. 6A), and its concentration increased steadily over the 12 h time course concomitant with the decrease of inorganic As. MAs(V) was detected after 6 h and its concentration also increased with time, although MAs(V) concentration was much lower that DMAs(V) concentration. At 12 h, DMAs(V) and MAs(V) accounted for 58 and 9.8% of the total As recovered in the assay solution. TMAsO(V) and DMAs(V) were detected in the AgNO₃ impregnated silica gel trap (Fig. 6B), with the latter being the predominant species. These two As species were the oxidized products of trimethylarsine [TMAs(III), Me₃As] and dimethylarsine [DMAs(III)H, Me₂AsH] gases, respectively. Over12 h, 4.5 and 44.3 ng of As were volatilized as TMAs(III) and DMAs(III)H, respectively (Fig. 6C), accounting for 1.3% of the total As(III) added to the assay solution.

FIG 6.

Methylation of As(III) by purified PaArsM in PBS buffer (pH 7.5) at 37°C. (A) Changes in As species in the assay solution over 12 h. Symbols: ■, total arsenic; △, inorganic arsenic; ○, DMAs(V); ▼, MAs(V). The initial concentration of As(III) was 10 μM; the concentration of the PaArsM protein was 3 μM. (B) HPLC-ICP-MS chromatograms of volatile As trapped on AgNO₃ impregnated silica gel. Curve 1, standards; curve 2: zero time; curve 3, 12 h. (C) Amount of volatile As after 12 h.

DISCUSSION

Although putative arsM orthologs have been found in the genomes of many bacteria, archaea, fungi, and algae (24), only ArsMs from R. palustris CGA009, Cyanidioschyzon sp. strain 5508, Methanosarcina acetivorans C2A, and three cyanobacterial species have been functionally characterized previously (17, 19, 21, 22). In the present study, we showed that P. alcaligenes was able to biotransform As(III) into methylated As species, with DMAs(V) being the predominant product. We further demonstrated that this biotransformation was catalyzed by the As(III) S-adenosylmethionine methyltransferase PaArsM because the PaarsM knockout strain lost the ability to methylate As, and E. coli expressing PaarsM and the purified protein was able to convert As(III) into methylated As species.

Several schemes have been proposed to explain the process of As biomethylation (13, 23, 38, 39) via sequential methylation steps producing mono-, di-, and trimethylated species. The main product of the PaArsM protein activity was found to be DMAs(V) in both P. alcaligenes and E. coli expressing PaarsM. Small amounts of MAs(V) were detected in the in vitro enzyme activity assays. It is possible that MAs(V), the first methylation product, may bind to the PaArsM protein more strongly than the subsequent methylation products, thus explaining the low level of MAs(V) measured in the medium (23). In both the assay with P. alcaligenes and the in vitro assay using purified PaArsM enzyme, volatile As species were produced. DMAs(III)H was found to be the main form of the volatile As with TMA(III) being the minor form. These results suggest that PaArsM is efficient at carrying out the methylation steps from As(III) to dimethylarsenic species but not to the trimethylarsenic species. This is different from a previous study on the ArsM of the bacterium Rhodopseudomonas palustris, which produces mainly TMAs(V)O in the medium and TMAs(III) in the headspace (17). The As(III) methylation mechanism in P. alcaligenes appears to be similar to that in the protozoan Tetrahymena pyriformis with both producing mainly DMAs(V) and DMAs(III)H (40). Although PaArsM contains all four conserved cysteine residues required for As binding and the three motifs for methyl donor binding typical of ArsM proteins (Fig. 2), its sequence identity to the R. palustris ArsM is only 36%, which may explain the differences in the methylation products.

There are debates on whether As methylation is a detoxification mechanism (41, 42). In the present study, P. alcaligenes could tolerate up to about 1 mM As(III). As(III) tolerance was decreased by ca. 60% in the PaarsM knockout strain PaΔarsM. Furthermore, overexpression of PaarsM in the As(III)-hypersensitive E. coli strain AW3110 significantly enhanced its As(III) tolerance. Taken together, our results provide strong evidence that As methylation is a detoxification mechanism in P. alcaligenes. This could include the conversion of As(III) to the much less toxic species DMAs(V) and its subsequent extrusion into the growth medium, as well as volatilization of DMAs(III)H and TMAs(III) gases.

In many bacteria and archaea, arsM genes are often present in clusters adjacent to other genes encoding As-resistance proteins (17, 24). For example, the As resistance gene cluster in Rhodopseudomonas palustris contains genes encoding arsenical resistance operon repressor (arsR), arsM, and arsenate reductase (arsC). RparsM expression was induced by As under the regulation of ArsR (43), which has been shown to control the expression of arsenic-resistance operons. In contrast, in the draft genome of P. alcaligenes, only a putative arsenate reductase (arsC) was found 4 kb downstream of the arsM and transcribed in the same direction. Three putative arsR genes exist in the draft genome of P. alcaligenes, but they are present in different scaffolds of the draft genome and are at least two open reading frames upstream of the PaarsM gene (Fig. 3A). The PaarsM gene was found to be expressed in the absence of As(III), and the expression was further enhanced by As(III) treatments. This pattern is different from the observations in R. palustris and three cyanobacteria species, in which arsM was expressed only in the presence of As(III) (17, 21). Whether the As(III)-enhanced PaarsM expression is regulated by the distant arsR genes requires further investigation.

Pseudomonas alcaligenes is abundant in the aquatic and soil environments (44). It may therefore play an important role in the biogeochemistry of As cycling. arsM genes have been detected in soils using degenerate primers (29) or Geochip which contains probes for more than 60 arsM genes but not PaarsM (30). However, the degenerate primers used by Jia et al. (29) could not amplify PaarsM from the genomic DNA of P. alcaligenes, likely due to the phylogenetic diversity and the limited As methylation sequences available. This means that the degenerate primers likely underestimate the true abundance of arsM in soil. Degenerate primers that can detect most arsM genes are needed for studies of functional gene diversity and abundance in the environment.

In conclusion, we have functionally characterized a novel arsenite S-adenosylmethionine methyltransferases gene, PaarsM in P. alcaligenes. PaArsM is able to methylate As(III) rapidly into DMAs(V) and produce volatile DMAs(III)H and TMAs(III). Arsenic methylation is also an important mechanism of As detoxification in P. alcaligenes. Differences between PaArsM and other microbial ArsM were observed, including the regulation of gene expression and the methylation products, which provides a better understanding of the genetic and biochemical diversity of As methylation mechanisms in microorganisms. The function of PaarsM could be exploited in bioremediation of As-contaminated soil or water, as has been demonstrated for other arsM genes (45, 46).