Mercury Reduction and Methyl Mercury Degradation by the Soil Bacterium Xanthobacter autotrophicus Py2

Abstract

Two previously uncharacterized potential broad-spectrum mercury (Hg) resistance operons (mer) are present on the chromosome of the soil Alphaproteobacteria Xanthobacter autotrophicus Py2. These operons, mer1 and mer2, contain two features which are commonly found in mer operons in the genomes of soil and marine Alphaproteobacteria, but are not present in previously characterized mer operons: a gene for the mercuric reductase (MerA) that encodes an alkylmercury lyase domain typical of those found on the MerB protein, and the presence of an additional gene, which we are calling merK, with homology to glutathione reductase. Here, we demonstrate that Py2 is resistant to 0.2 μM inorganic mercury [Hg(II)] and 0.05 μM methylmercury (MeHg). Py2 is capable of converting MeHg and Hg(II) to elemental mercury [Hg(0)], and reduction of Hg(II) is induced by incubation in sub toxic concentrations of Hg(II). Transcription of the merA genes increased with Hg(II) treatment, and in both operons merK resides on the same polycistronic mRNA as merA. We propose the use of Py2 as a model system for studying the contribution of mer to Hg mobility in soil and marine ecosystems.

INTRODUCTION

Bacterial mercury (Hg) resistance genes (mer) are important drivers in the biogeochemical cycle of Hg. These operons catalyze the conversion of inorganic mercury [Hg(II)] and sometimes methylmercury (MeHg) to elemental mercury [Hg(0)] (1). MeHg, which is more toxic than Hg(II) or Hg(0), is the form that biomagnifies in aquatic food webs (2, 3) and can cause toxicity to humans and animals that consume contaminated fish (4–6). Hg(II) is water soluble but sorbs strongly onto iron oxides and interacts with dissolved organic matter (7). Hg(0), the least toxic form, is both a liquid and a gas at room temperature and can evaporate from surface waters and soils (8, 9).

All mer operons contain a gene encoding mercuric reductase, MerA, which converts Hg(II) to Hg(0), thereby conferring resistance (10). Some mer operons, called broad-spectrum mer operons, contain an additional gene encoding MerB, or alkylmercury lyase (AML), which degrades MeHg to Hg(II) and methane (11). Many mer operons contain genes for the transcriptional regulators MerR and MerD (12), as well the Hg transporters encoded by merT, merC, and merF, as well as additional Hg transfer proteins encoded by merP and merE (11, 13). Regulation of mer and the enzymatic activity of MerA is known in great detail for some operons and enzymes (10).

It has been demonstrated that mer can influence the Hg cycle in lakes (14), but at present we know much less about the contribution of mer to the Hg cycle in soil and the terrestrial subsurface. It was long thought that Hg was relatively immobile in the subsurface due to sorption on sediment components such as iron oxides (1), but this assumption has come under scrutiny, since Hg has been unexpectedly found in groundwater at numerous sites, such as the Kirkwood-Cohansey aquifer in New Jersey (15, 16), the Waquoit Bay near Cape Cod, MA (17), and the coasts of California, northern France, and South Korea (18–21). One attractive hypothesis to explain the mobilization of Hg in the subsurface is that bacteria reduce sediment-bound Hg(II) to Hg(0), possibly catalyzed by MerA.

Several lines of evidence have indicated that Alphaproteobacteria are active in Hg- and uranium-contaminated sediments (22–25). We have developed Xanthobacter autotrophicus Py2 (26) as a model system to study mer function in soil Alphaproteobacteria. The Py2 genome has two potential mer operons, each with two unusual features. The merA genes encode proteins with N-terminal AML domains similar to the one typically found in MerB. In addition, the operons contain a previously unknown potential mer gene with homology to glutathione reductase. We demonstrate here that Py2 can convert both MeHg and Hg(II) to Hg(0) in a manner consistent with detoxification by mer and present evidence for a new mer gene, which we refer to as merK.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

Xanthobacter autotrophicus Py2 was obtained from the the American Type Culture Collection, and cultures were grown aerobically at 28°C using Xanthobacter minimal medium (XMM), which contained the following components per liter: 5.37 g of Na₂HPO₄·12H₂O, 1.36 g of KH₂PO₄, 0.5 g of (NH₄)₂SO₄, 0.2 g of MgSO₄·7H₂O, 0.37 g of NaNO₃, 5 ml of Wolfe's minerals, and 5 mM glucose. Growth was measured by optical density using a spectrophotometer at λ = 660 nm. Cells were killed by heating them to 85°C for 30 min.

Hg analysis.

Hg(II) was provided as HgCl₂. MeHg was provided as methylmercury chloride, dissolved in ethanol. All glassware was cleaned overnight with nitric acid, 8 M for glassware, and 4 M for caps and tubing. Water (18 MΩ) was provided with a Hydro PicoPure2 and trace-metal-grade (TMG) acids were used for analysis. Digestion of samples containing Hg was performed using TMG sulfuric acid, TMG nitric acid, and potassium permanganate as previously described (27). Hg was analyzed with a PSA Analytical Millennium Systems Merlin mercury analyzer with PSA 20.400 Autosampler, using the parts per billion (ppb) protocol. The machine was calibrated using 18 MΩ water as a blank. Standard curves were generated using a NIST Hg standard solution of 1,000 mg/liter (Hach Company, Loveland, CO), and curves were accepted with an R² value of >0.99.

Hg(II) loss assays.

For Hg reduction experiments performed in XMM, overnight cultures were used to inoculate triplicate 6-ml cultures at an equivalent optical density (OD) of <0.05, with the indicated amount of Hg(II) or MeHg. For reduction experiments, cells were grown to an optical density of 0.1. The cultures were split: half were exposed to 0.02 μM Hg in growth medium for 1 h, and the other half were incubated in growth medium without Hg. Cells were harvested by centrifugation and washed three times with 50 mM HEPES buffer (pH 7.4). The cell pellets were resuspended to an OD of 0.1 in HEPES containing 0.75 μM Hg(II).

Hg trapping.

To trap Hg(0), cultures of Py2 were placed in a 10-ml amber serum bottle containing 5 ml of XMM and either Hg(II) or MeHg. The serum bottle was connected to a midget bubbler (Ace Glass, Vineland, NJ; catalog no. 75320-06) containing acidified potassium permanganate as previously described (27). The apparatus was flushed continuously with in-house compressed air filtered with a 0.2-μm-pore-size nylon syringe filter.

Bioinformatic analyses.

Candidate mer operons were recovered from the NCBI or IMG databases by BLAST searches using the MerA protein from the transposon Tn5041 (NCBI accession number CAA67451) or the Bacillus RC607 MerA sequence (NCBI accession number AAA83977.1) as a query. All recovered subjects were aligned with known functional MerAs using MUSCLE (28) and included in analyses only if they had both a conserved active site and C-terminal cysteine pairs known to be required for catalytic activity (29). Active sites were further elucidated by aligning Py2 proteins with known MerAs using TCoffee (30). AML domains were localized within MerA₁ and MerA₂ by querying the Pfam database (31).

Transcript analysis.

RNA was extracted from both Hg(II)-treated and -unexposed Py2 cultures using Mo-Bio bacterial RNA isolation kit (MO BIO, catalog no. 15800-50). For each sample, 500 ng of RNA was reverse transcribed to cDNA using a QuantiTect reverse transcription kit (Qiagen, catalog no. 205310). Aliquots of RNA (500 ng) were removed for mock treatment without the addition of reverse transcriptase.

Primers for the amplification of cDNA of merA₁, merA₂, merK₁, merK₂, gorA, and recA (32) were designed using Primer3 (33, 34). Primer pairs were selected that amplified a 200-bp fragment in each gene (see Table S1 in the supplemental material). The specificity of the primers was checked using NCBI Primer-BLAST against the Py2 genome (NCBI accession number NC_009720.1) and by sequencing PCR products. Sequencing was performed by Macrogen (Cambridge, MA).

PCRs were performed in a 10-μl PCR, with 0.01 nmol of primers, 0.4 μl of cDNA (or no-RT control samples), 2 nM deoxynucleoside triphosphate, and Taq-Pro DNA polymerase (1 U; Denville Scientific, catalog no. CB4050-7). Thermocycling conditions for intragenic regions were as follows: a 5-min incubation at 95°C, followed by 30 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 30 s, and culminating in a 5-min incubation at 72°C. Primers 0797qF and 0796qR (see Table S1 in the supplemental material) were used to amplify the intergenic region between merA₁ and merK₁. Primers 2903qF and 2904qR (see Table S1 in the supplemental material) were used to amplify the intergenic region between merA₂ and merK₂. Thermocycling conditions for intergenic regions were as follows: a 5-min incubation at 95°C, followed by 30 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 2.5 min, and culminating in a 20-min incubation at 72°C. PCR products were analyzed using a 1.2% agarose gel for intragenic regions and 0.8% for the intergenic regions.

RESULTS

Two regions in the genome of Xanthobacter autotrophicus Py2 (NCBI accession number NC_009720.1) are candidates for mer operons, mer1 and mer2 (Fig. 1), forming two distinct loci rather than a coverage or annotation error. Although both copies of MerA are 767 amino acids long, they have only 83% homology at both the nucleotide and protein level (see Fig. S1A and B in the supplemental material). They exist in distinct regions of the chromosome; mer1 is flanked by an ardC homolog (Xaut_0795) and a hypothetical protein (Xaut_0801), but mer2 is flanked by a homolog of an FAD containing monooxygenase (Xaut_2899) and an integrase (Xaut_2905).

FIG 1.

Arrangement of potential mer operons in Xanthobacter autotrophicus Py2. R, merR; T, merT; P, merP; A, merA; K, potential glutathione reductase. Locus tags are indicated above the genes.

Both mer1 and mer2 have genes with homology to merR, merT, merP, and merA (Fig. 1). In both operons, a gene with homology to glutathione reductase resides promoter distal to the merA homolog, which will be referred to as merK. Additional mer operons containing merK are present in the genomes of at least 23 other Alphaproteobacteria, including 4 isolates from uranium-contaminated soil and 16 marine bacteria (Table 1; see also Table S2 in the supplemental material). We were only able to find this arrangement in the genome of one bacterium outside the phylum Alphaproteobacteria, in the gammaproteobacterium Halomonas sp. strain KO116, between the loci Ga0054578_113120 and Ga0054578_113128.

TABLE 1.

Presence of similar mer features in other Alphaproteobacteria

^a A, merA; B, merB; C, merC; D, merD; Dh, dehalogenase; F, merF; P, merP; T, merT; K, merK; Th, thiol-related protein; Tr, transporter; x, hypothetical protein.

Both potential MerAs have all of the residues known to contribute to catalytic activity. The pyridine nucleotide disulfide oxidoreductase domain spans amino acids 307 to 614 in both proteins, followed by a pyridine nucleotide disulfide dimerization domain, from amino acids 643 to 752 (see Fig. S1B in the supplemental material). The active site of MerA in Bacillus RC607, located between amino acids 201 to 218, has key cysteine residues at 207 and 212 (see Fig. S1B in the supplemental material) (10). This region spans amino acids 337 to 354 in MerA₁ and MerA₂ from Py2, with key cysteine residues present at amino acids 343 and 348. These regions are identical to those in RC607 except for isoleucines at amino acids 339 and 346 in the Py2 proteins, replacing valines at amino acids 203 and 210 of the RC607 protein. The same substitution occurs in MerA in other species, such as Acidithiobacillus ferrooxidans (NCBI accession number BAA14139.3) and Pseudomonas fluorescens (NCBI accession number CAA51542.1). Both MerA₁ and MerA₂ possess the two C-terminal cysteines (amino acids 764 and 765, corresponding to amino acids 628 and 629 in RC607) known to be required for catalytic activity (35).

Although residues required for reduction of Hg(II) are present in the MerAs from Py2, both proteins lack the GMTCXXC-containing metal-binding domain NmerA (36, 37) and have, in its place, an AML domain from amino acids 128 to 258, similar to MerB. To date, no MerA proteins with an AML domain have been investigated. An AML domain is common among MerAs in soil and rhizosphere isolates in the phylum Alphaproteobacteria (Table 1; see also Table S3 in the supplemental material) but is not apparently present in MerA proteins from organisms outside Alphaproteobacteria.

We characterized the resistance of Py2 to Hg(II) and MeHg. Py2 has MICs of 0.50 μM for Hg(II) (Fig. 2A) and 0.1 μM for MeHg (Fig. 2B). To determine whether growth of Py2 occurred concurrently with loss of Hg(II), we grew Py2 with or without 0.2 μM Hg(II) (Fig. 3A). In both exposed and unexposed cultures, cells were in logarithmic phase between 5 and 25 h. In cultures containing Py2 and Hg(II), 90.2% ± 1.4% of Hg(II) was lost from the culture medium during the first 7.5 h of the experiment compared to only 5.0% ± 0.8% from uninoculated media.

FIG 2.

MICs of Hg(II) and MeHg in Py2. (A) Hg(II) MICs. The 0 and 0.1 μM Hg(II) growth curves are overlapping. (B) MeHg MICs. The 0, 0.02, and 0.05 μM MeHg growth curves overlap. Error bars represent the standard deviations of the means from three independent cultures.

FIG 3.

Loss of Hg(II) in the presence of Py2. (A) Loss of Hg(II) by growing cultures of Py2. (B) Loss of Hg(II) by Py2 suspended in buffer. Pretreated cells were exposed to 0.02 μM Hg(II) in growth media for 1 h prior to the assay. Mock treated cells were incubated for 1 h in media with no Hg(II).

One of the hallmarks of reduction of Hg(II) via merA is that previous exposure to low levels of Hg(II) initiates transcription of the operon, which increases the initial rate of Hg(II) reduction (38). This occurs because Hg(II) forms a complex with MerR, changing it from a transcriptional repressor to an activator. When resuspended in HEPES buffer with no carbon source, cells that were previously exposed to 0.020 μM Hg(II) for 1 h lost 75.6% ± 5.5% of 0.7 μM Hg(II) in 1 h (Fig. 3B). In contrast, no significant loss of Hg(II) was observed in live cells that were not exposed to Hg(II). No loss of Hg(II) was observed in buffer only or in the presence of heat-killed cells (Fig. 3B).

To demonstrate that the loss of Hg(II) in the presence of Py2 is the result of reduction of Hg(II) to Hg(0), we captured Hg(0) in a strongly oxidizing solution of sulfuric acid and potassium permanganate. Py2 cells were grown to mid-log phase, incubated for 1 h with 0.02 μM Hg(II), and placed in a serum bottle containing XMM and 132.8 ± 16.7 nM Hg(II). Air was then passed through the serum bottle and into a bubbler containing the acidified permanganate. The total amount of Hg(II) added to the system was 0.66 ± 0.11 nmol. After 2 h, 107% of the Hg was lost from the Py2 cultures, and 111% (0.74 ± 0.09 nmol) of the Hg(II) was captured in the trapping solution (Fig. 4A). When Py2 cells were heat killed prior to treatment with Hg(II), 100% of the 0.72 ± 0.04 nmol of Hg(II) remained in the culture medium, indicating that the observed reduction of Hg(II) is not a chemical process.

FIG 4.

Capture of Hg(0) in Py2 cultures exposed to Hg(II) or MeHg. (A) Data from experiments conducted with Hg(II). (B) Data from experiments conducted with MeHg. Gray bars represent Hg present in the reaction vessel, and white bars represent Hg present in the trapping solution. Error bars represent the standard deviations of the means for three independent reactors and traps.

Py2 also converts MeHg to Hg(0). Py2 cells were grown to mid-log phase and incubated for 1 h with 5 nM Hg(II), harvested via centrifugation, and resuspended in 5 ml of medium containing 49.7 ± 3.8 nM MeHg (188.7 pmol total). After 2 h, 17.2% ± 1.4% of the MeHg remained in the culture medium, and 81.7% ± 6.8% of the Hg was captured in the trapping solution, yielding 98.9% recovery of mercury (Fig. 4B). With heat-killed cells, using 49.7 ± 3.4 nM MeHg, 99.9% of the MeHg remained in the culture medium.

We observed increases in the transcription of merA₁ and merA₂ upon exposure to Hg(II). We exposed mid-log-phase cultures of Py2 to 0.1 μM Hg(II) for 1 h, collected cells by centrifugation, and then prepared cDNA from those cultures and untreated controls. We amplified ∼200-bp fragments from merA₁ and merA₂ in cDNA prepared from cultures that were exposed to Hg(II), but these amplicons were absent in cultures that were not exposed to Hg(II) (Fig. 5). Sequencing confirmed that the bands detected were from merA₁ and merA₂. The transcription of merK₁ and merK₂ followed a similar pattern (Fig. 5). The gene for the chromosomal copy of glutathione reductase, gorA, and the housekeeping gene recA were transcribed in the presence or absence of Hg, confirming the presence of cDNA in all preparations. No bands were observed in aliquots processed with no reverse transcriptase, confirming that the amplification products were not a result of genomic DNA contamination.

FIG 5.

Gel electrophoresis of reverse transcription-PCR (RT-PCR) products from intragenic regions of the Py2 mer operon. Three independent cultures (1, 2, and 3) were split into two aliquots, one of which (Hg) was exposed to 0.1 μM Hg(II) and one of which was not exposed Hg(II) (No Hg). After preparing mRNA, half of each sample was treated with reverse transcriptase (RT), and half was not (No-RT). gDNA indicates the use of Py2 genomic DNA as a template.

The observation that merK₁ and merK₂ are not constitutive, like gorA, and are regulated similarly to merA₁ and merA₂ suggests that these gene pairs are transcribed on a polycistronic mRNA. To confirm this, we selected primers spanning the intergenic region between these two gene pairs and were able to amplify sequence spanning this region from the cDNA described above (Fig. 6). No amplification products were obtained from samples in which the mRNA was not treated with reverse transcriptase (Fig. 6).

FIG 6.

Gel electrophoresis of RT-PCR products from intergenic regions the Py2 mer operon. (A) Intergenic region between merA₁ and merK₁. (B) Intergenic region between merA₂ and merK₂. The template cDNA and no-RT samples were identical to the Hg(II)-exposed samples in Fig. 6. gDNA indicates the use of Py2 genomic DNA as a template. The rightmost lane is the 2-log ladder from New England BioLabs.

DISCUSSION

Here we present the first characterization of Hg(II) reduction and MeHg degradation from a member of the phylum Alphaproteobacteria (Fig. 3 and 4). Because reduction of Hg(II) is inducible (Fig. 3B) and occurs with increased transcription merA₁ and merA₂ (Fig. 5), we propose that these loci represent novel broad-spectrum mer operons. Supporting this, when these regions were transformed into an Hg(II)-sensitive strain of Escherichia coli, they conferred the loss of Hg(II) and MeHg from the culture medium (Y. Wang and H. Wiatrowski, unpublished data).

Alphaproteobacteria have not traditionally been found in screens for mercury resistant microorganisms and have only recently been discovered in surveys from soil (22–25). The lack of representation of Alphaproteobacteria among Hg(II)-resistant microorganisms is possibly explained by our observation that Py2 is not able to grow at Hg(II) concentrations higher than 0.2 μM. This is considerably lower than the level of Hg(II) resistance typically afforded by mer on a plasmid. For example, the mer operon present on the plasmid VS1 confers resistance to ∼20 μM Hg(II) to Pseudomonas aeruginosa (39). When investigators attempt to isolate bacteria containing the mer operons, they typically plate serial dilutions of a sample on solid culture medium containing Hg(II) at 10 μg/ml (24, 40). Despite obtaining 102% plating efficiency at a concentration of 1 μg of Hg(II)/ml, no colonies of Py2 grew after plating on media containing Hg(II) at a concentration of 10 μg/ml (data not shown). The lower level of resistance in Py2 could possibly be explained by the observation that most of the highly Hg-resistant organisms possess a plasmid-borne copy of mer, whereas in Py2 they are chromosomal. We should therefore be cautious about drawing conclusions regarding Hg resistance based on the discovery of mer in soil metagenomic surveys.

The present study represents the first demonstration of Hg(II) reduction and mercury demethylation by an organism that has an AML domain on the N terminus of the mercuric reductase (Fig. 2 and 4). Typically, AML function is conferred on a separate protein, MerB. To date, mercuric reductases have come in two classes: those with an N-terminal domain involved in delivering Hg(II) to the catalytic core (37) and those that lack an N-terminal extension, consisting solely of the core catalytic domains (41). The MerAs in Py2 represent a third group, including MerAs from several other isolates from the phylum Alphaproteobacteria. Notably several of these isolates are from uranium and nitrate contaminated soils which have experienced transient Hg plumes, such as Afipia sp. strain 1NLS2 and Hyphomicrobium denitrificans 1NES1 (42). In addition, this class of merA is found in the genomes of the rhizosphere and soil bacteria Ochrobactrum rhizosphaerae SJY1, Rhizobium leguminosarum YR374, Sinorhizobium arboris LMG 14919, and Sinorhizobium fredii HH103 (Table 1; see Tables S2 and S3 in the supplemental material).

Our work identifies the presence of a new mer gene, which we will call merK. These genes encode proteins with homology to members of the pyridine nucleotide disulfide oxidoreductase family, and are most similar to a glutathione reductase (Wang and Wiatrowski, unpublished). They are likely part of the mer operon as they exist on a polycistronic mRNA with merA (Fig. 6). We propose that merK is a novel addition to mer operons in Alphaproteobacteria, and this gene is particularly common in marine isolates. At present, we do not know the function of merK in mercury resistance or transformation, but we have observed that it can transform oxidized glutathione to reduced glutathione (Y. Wang and H. Wiatrowski, unpublished data).

The biogeochemical cycle, as well as the role of mer, is well understood in freshwater lakes (14); however, it is much less clear how mer impacts the biogeochemical cycle of Hg in subsurface sediments and the ocean (23–25, 43). We propose that the mer operons in Py2 are representative of those that may be found in the genomes of bacteria living in the ocean and contaminated subsurface sediments (Table 1; see also Tables S2 and S3 in the supplemental material), and we support the utilization of Py2 as a model system for understanding Hg transformations in these environments.