Improving Arsenic Tolerance of Pyrococcus furiosus by Heterologous Expression of a Respiratory Arsenate Reductase

Arsenate respiratory reductases (Arr) are much less characterized than the detoxifying arsenate reductase system. The heterologous expression and characterization of an Arr from Pyrobaculum aerophilum in Pyrococcus furiosus provides new insights into the function of this enzyme. From in vivo studies, production of Arr not only enabled P. furiosus to use arsenate [As(V)] as a terminal electron acceptor, it also provided the organism with a higher resistance to arsenate and also, surprisingly, to arsenite [As(III)]. In contrast to the tungsten-containing oxidoreductase enzymes natively produced by P. furiosus, recombinant P. aerophilum Arr was much more active with molybdenum than with tungsten. It is also, to our knowledge, the only characterized Arr to be active with both molybdenum and tungsten in the active site.

ABSTRACT

Arsenate is a notorious toxicant that is known to disrupt multiple biochemical pathways. Many microorganisms have developed mechanisms to detoxify arsenate using the ArsC-type arsenate reductase, and some even use arsenate as a terminal electron acceptor for respiration involving arsenate respiratory reductase (Arr). ArsC-type reductases have been studied extensively, but the phylogenetically unrelated Arr system is less investigated and has not been characterized from Archaea. Here, we heterologously expressed the genes encoding Arr from the crenarchaeon Pyrobaculum aerophilum in the euryarchaeon Pyrococcus furiosus, both of which grow optimally near 100°C. Recombinant P. furiosus was grown on molybdenum (Mo)- or tungsten (W)-containing medium, and two types of recombinant Arr enzymes were purified, one containing Mo (Arr-Mo) and one containing W (Arr-W). Purified Arr-Mo had a 140-fold higher specific activity in arsenate [As(V)] reduction than Arr-W, and Arr-Mo also reduced arsenite [As(III)]. The P. furiosus strain expressing Arr-Mo (the Arr strain) was able to use arsenate as a terminal electron acceptor during growth on peptides. In addition, the Arr strain had increased tolerance compared to that of the parent strain to arsenate and also, surprisingly, to arsenite. Compared to the parent, the Arr strain accumulated intracellularly almost an order of magnitude more arsenic when cells were grown in the presence of arsenite. X-ray absorption spectroscopy (XAS) results suggest that the Arr strain of P. furiosus improves its tolerance to arsenite by increasing production of less-toxic arsenate and nontoxic methylated arsenicals compared to that by the parent.

IMPORTANCE Arsenate respiratory reductases (Arr) are much less characterized than the detoxifying arsenate reductase system. The heterologous expression and characterization of an Arr from Pyrobaculum aerophilum in Pyrococcus furiosus provides new insights into the function of this enzyme. From in vivo studies, production of Arr not only enabled P. furiosus to use arsenate [As(V)] as a terminal electron acceptor, it also provided the organism with a higher resistance to arsenate and also, surprisingly, to arsenite [As(III)]. In contrast to the tungsten-containing oxidoreductase enzymes natively produced by P. furiosus, recombinant P. aerophilum Arr was much more active with molybdenum than with tungsten. It is also, to our knowledge, the only characterized Arr to be active with both molybdenum and tungsten in the active site.

INTRODUCTION

Arsenic is a well-known toxicant. The biologically relevant oxidized pentavalent form, arsenate [As(V)], replaces phosphate in many biochemical pathways due to the similarity in their structures and properties, while the reduced form, trivalent arsenite [As(III)], readily interacts with thiol-containing molecules such as cysteine (1). Microorganisms have developed systems to protect themselves against arsenic toxicity and to survive in natural environments where arsenic compounds are prevalent (2). Two distinct types of arsenate reductases are known, the cytoplasmic ArsC type and the membrane-bound Arr type, both of which reduce arsenate to arsenite (3).

The ArsC type has been studied extensively in bacteria in which it functions in detoxification. The enzyme is encoded by an arsenic resistance (ars) operon that contains at minimum arsR, arsB, and arsC. Additionally, some genomes contain the accessory genes arsD and arsA (4). ArsR functions as a repressor of the operon (5, 6). ArsC is a small cytoplasmic protein that uses a redox-active cysteine residue in the active site to reduce arsenate to arsenite using glutathione as the electron donor (7). Arsenite is then transported out of the cell by the arsenite transporter ArsB, which is energized by either a proton motive force or by the ArsA ATPase (4). ArsD is an arsenite chaperone that lowers the affinity of ArsA for arsenite (8). In contrast to ArsC, the Arr system enables arsenate to be used as a terminal electron acceptor of the respiratory chain. The enzyme consists of ArrA, the catalytic subunit that contains a pyranopterin cofactor and an iron sulfur cluster, and ArrB, which also contains iron-sulfur clusters. Compared to ArsC, the Arr-type arsenate reductase is much less studied. Three ArrABs have been purified from bacteria, but none have been characterized from archaea (9–11), which, as described below, appear to contain a different type of Arr.

The crenarchaeon Pyrobaculum aerophilum grows optimally at 100°C using arsenate as well as several other compounds, such as oxygen, nitrate, and thiosulfate, as a terminal electron acceptor (12, 13). Based on transcriptional analyses in response to these respiratory substrates, the genes PAE1263 to PAE1265 were predicted to encode an arsenate respiratory reductase (Arr). Their expression increased up to 8-fold when arsenate was supplied in the growth medium (13, 14). Although the P. aerophilum Arr operon was not detected using degenerate PCR primers to bacterial Arr genes, the catalytic domain of bacterial Arr is conserved in arrA (PAE1265) (15). A model of P. aerophilum Arr based on the sequence analysis is shown in Fig. 1. ArrA is the catalytic subunit containing a pyranopterin-binding site and a [4Fe-4S] cluster, while ArrB harbors three [4Fe-4S] clusters. In further contrast to the characterized bacterial Arr, which forms heterodimers, the enzyme from P. aerophilum contains a third subunit, ArrC, which is predicted to be a membrane-bound anchor for ArrAB. ArrC is thought to oxidize quinol and transport the electrons to ArrA through the iron-sulfur clusters in ArrB (16). In contrast to this archaeal system, bacteria that respire arsenate, such as Shewanella species, do not have a gene encoding ArrC but contain the membrane-bound menaquinol-oxidizing protein CymA, which does not show any sequence similarity to ArrC. CymA has been shown by mutational studies to be crucial for arsenate reduction by Shewanella, although it has not been biochemically characterized (17).

FIG 1.

SDS-PAGE of purified Arr and the proposed cofactor content of ArrABC based on sequence analysis. ArrA and ArrB were identified by MALDI-TOF from bands cut from the gel. ArrC was not identified in the purified protein and is shown with a dashed outline in the model.

Since a genetic system is not available for P. aerophilum, we used the euryarchaeon Pyrococcus furiosus to heterologously produce its Arr with an affinity tag to facilitate purification. P. furiosus is a strict anaerobe that also grows optimally near 100°C (18), but unlike P. aerophilum, it grows only by fermenting carbohydrates or peptides. Elemental sulfur (S⁰) is essential as a terminal electron acceptor when cells are grown solely on peptides, while H₂ rather than H₂S is produced during growth on sugars in the absence of S⁰ (19). The production of both H₂ and H₂S involves homologous respiratory complexes that use reduced ferredoxin (Fd) as the electron donor (20). P. furiosus has a well-developed genetic system, and genes from other euryarchaeota have been heterologously expressed that enable the organism to oxidize formate and carbon monoxide and produce H₂ (21, 22).

The use of P. furiosus as an expression system also raises the issue of which metal it would insert into the recombinant Arr. The three bacterial ArrA enzymes contain molybdenum (Mo) (9–11); indeed, almost all life forms have one or more pyranopterin enzymes that all contain molybdenum (23). However, P. furiosus is very unusual in that it expresses five pyranopterin-containing oxidoreductases that instead contain tungsten (W) (24), a chemically analogous metal (23). P. furiosus does not produce active native molybdoenzymes (25), and it has been proposed that, in general, hyperthermophilic microbes growing near the boiling point are predisposed to use tungsten rather than molybdenum as the metal in their pyranopterin enzymes (23). Previous work from our lab investigated the metal requirements of the closely related thiosulfate reductase (Tsr) complex from P. aerophilum heterologously purified from P. furiosus and found that despite the fact that P. aerophilum requires tungsten for growth, the Tsr was more active with molybdenum than with tungsten (26, 27). Hence, key questions were (i) does P. furiosus insert tungsten or molybdenum into P. aerophilum Arr, and (ii) if both can be purified, which is the most active form?

To gain a deeper understanding of the biochemical properties of an archaeal Arr, we therefore heterologously expressed the arr operon (PAE1263-1265) from P. aerophilum in P. furiosus. The recombinant Arr enzyme was purified from P. furiosus cells grown on medium with added molybdenum or tungsten, and the recombinant enzymes were affinity purified and characterized. The response of P. furiosus with and without expression of the P. aerophilum Arr genes to growth in the presence of arsenate [As(V)] and arsenite [As(III)] was also investigated.

RESULTS

Recombinant P. aerophilum Arr produced in P. furiosus.

The P. furiosus strain expressing the P. aerophilum Arr genes will be referred to as the Arr strain. It was constructed using the parent strain COM1 by homologous recombination of the knock-in cassette containing PAE1263 to PAE1265 into the intergenic space between PF0265 and PF0266, as shown in Fig. S1 in the supplemental material. Initially, a strain was constructed containing a sequence encoding a 9×His affinity tag at the N terminus of ArrA, but we were unable to purify the Arr enzyme using a Ni-nitrilotriacetic acid (NTA) column. A second strain was therefore constructed (MW0548) with the His tag at the C terminus of ArrA, and this enabled affinity purification of the recombinant Arr protein as described below. The expression of arrA (PAE1265), driven by the constitutive promoter P_slp, in the Arr strain was confirmed by quantitative PCR (qPCR) analysis, where the level of expression was similar to that of pyruvate oxidoreductase (γ subunit) (data not shown), an abundant enzyme involved in fermentation.

The P. furiosus Arr strain was initially grown in a medium containing both 10 μM W and 10 μM Mo, and the arsenate [As(V)] reduction activity, using reduced methyl viologen as the artificial electron donor, was measured in the cell extract. As shown in Fig. S2, no significant activity was observed in the parent strain (Par). To investigate if all three subunits were present as an intact complex in the Arr strain, the arsenate reduction assay was carried out using the cytoplasmic and membrane fractions, as shown in Fig. S3. Unexpectedly, 80% of the activity was found in the cytoplasmic extract, and ArrC was not identified by liquid chromatography-tandem mass spectrometry (LC-MS/MS) in the active Arr-containing fractions purified from either the cytoplasmic or the membrane fraction (Fig. 1). qPCR analysis indicated that its gene was transcribed at the expected level, suggesting that a stable ArrC subunit is not produced in the Arr strain.

Characterization of recombinant Arr containing Mo or W.

The catalytic subunit ArrA is predicted to contain a pyranopterin cofactor. To investigate whether the active enzyme contained Mo or W bound to the pyranopterin site, the Arr strain was grown under two different conditions: (i) a so-called Mo condition, in which 10 μM Mo but no W was added to the growth medium, and (ii) a W condition, in which 10 μM W but no Mo was added. Cells (25 g wet weight.) grown under the Mo or W condition were used for the affinity purification of Arr, and the results are summarized in Tables S1 and S2, respectively. With both cell types, after two chromatography steps (Ni-NTA and hydrophobic interaction), each enzyme was purified more than 1,400-fold with at least a 60% yield of activity. Surprisingly, we were able to purify active forms of the enzyme from both cell types, but the Mo-containing enzyme (Arr-Mo) had >140-fold higher arsenate reduction activity than the W-containing enzyme (Arr-W). Metal analysis revealed that the Fe-to-active site metal ratio was 17:1 for both Arr-Mo and Arr-W, consistent with the presence of 4 [4Fe-4S] clusters. The Mo-to-W ratios in these enzymes were 9:1 and 1:86 for Arr-Mo and Arr-W, respectively, and kinetic analysis showed that Arr-W had a 20-fold higher affinity for arsenate than Arr-Mo (Table 1). The arsenite [As(III)] reduction activity and ferredoxin-linked arsenate [As(V)] reduction activities (using the ferredoxin from P. furiosus in place of methyl viologen) were also measured for Arr-Mo, and the activities were more than 100-fold lower than the arsenate [As(V)] reduction activity using the artificial mediator (Table 1). Arr proteins from Alkalilimnicola ehrlichii and Shewanella sp. strain ANA-3 were reported to exhibit arsenite [As(III)] oxidation activity (28). However, we were unable to measure arsenite oxidation activity with Arr-Mo using the high-potential dye dichlorophenolindophenol as the electron acceptor. Because the arsenate reduction activity of Arr-Mo was much higher than that of Arr-W, the Mo growth condition was used to carry out the following in vivo experiments.

TABLE 1.

Properties of Arr

Property	Value
	Arr-Mo	Arr-W
Km for arsenate (μM)	340 ± 50	17.9 ± 2
Vmax (U/mg)	1,800 ± 160	13.0 ± 2.0
Metal content
Mo:W	9:1	1:86
Fe:Mo or W	17:1	17:1
Reduction (U/mg)
Ferredoxin-linked arsenate [As(V)]	20.9 ± 1.7	NDa
Arsenite [5 mM As(III)]	6.79 ± 0.36	ND

^aND, not determined.

Arsenate serves as a terminal electron acceptor for the Arr strain.

P. furiosus shows significant growth on peptides only in the presence of S⁰, which is reduced to hydrogen sulfide. S⁰ serves as a terminal electron acceptor with a reduction potential (E₀′ ∼ −280 mV) higher than that of hydrogen (E₀′ = −420 mV) (19). The P. furiosus Arr strain (grown with Mo) was examined for the ability to use arsenate (E₀′ = 60 mV) as a terminal acceptor to support peptide metabolism. The experiment was performed using a growth medium containing only peptides as the carbon source, and the Arr strain grew as well as the parent in the presence of S⁰ (Fig. 2A). However, the Arr strain showed increased growth on peptides (by more than 50%, measured by cell protein) compared to that of the parent strain when 10 mM arsenate was present, although a long lag phase (30 h) was observed before there was an effect. Stimulation of growth by arsenate was highly reproducible with the Arr strain, although under such conditions, the strain did not reach the growth attained in the presence of S⁰ (Fig. 2A). These data suggested that the Arr strain was able to reduce arsenate [As(V)] to arsenite [As(III)] to support growth. However, the parent strain also reduced arsenate to arsenite, as shown by the increase in the concentration of arsenite in the growth medium, although the Arr strain produced slightly more arsenite than the parent (Fig. 2B). Accordingly, as shown in Fig. 2C, in the presence of arsenate, both the Par and Arr strains produced much less hydrogen gas during growth if arsenate was present. These results, therefore, suggest that P. furiosus can nonspecifically reduce arsenate and that the presence of Arr allows arsenate to stimulate growth, an effect not seen with the parent strain.

FIG 2.

Growth of P. furiosus on peptides with and without arsenate (10 mM). (A) Red, the Arr strain with the addition of arsenate; black, Par with the addition of arsenate; blue, Par without arsenate; green, the Arr strain without arsenate. (B) Arsenite [As(III)] concentration in the spent media after the addition of 10 mM arsenate. Black line, Par; red line; the Arr strain. (C) Hydrogen production normalized by protein concentration during growth. Blue, Par; green, the Arr strain; black, Par with the addition of arsenate; red, the Arr strain with the addition of arsenate. The error bars represent standard deviations obtained using biological triplicate samples.

Arsenic toxicity.

As shown in Fig. 3A, the growth of P. furiosus by sugar (maltose) fermentation was strongly inhibited by the presence of 10 mM arsenate, but the Arr strain was more resistant than the parent strain. Note that such inhibition of the parent strain was not observed during growth on peptides, presumably because there was minimal growth and the effect was not significant enough to be observed. As shown in Fig. 3A, the maximum growth of the parent was approximately one-half that of the Arr strain in the presence of arsenate, although both strains exhibited a long lag phase compared to that of the control in the absence of arsenate [As(V)]. In addition, arsenite [As(III)] was detected in the growth medium of arsenate-inhibited cells. However, although the Arr strain generated slightly more (approximately 0.6 mM) than the parent strain (approximately 0.5 mM), arsenite production was largely abiotic, since a similar concentration was measured using the P. furiosus growth medium in the absence of cells (Fig. 3B).

FIG 3.

Inhibition of P. furiosus growing on maltose by arsenate. (A) Blue, Par; green, the Arr strain; black, Par with the addition of 10 mM arsenate; red, the Arr strain with the addition of 10 mM arsenate. (B) Arsenite [As(III)] concentration in the spent media with the addition of 10 mM arsenate. Black line, Par; red line, the Arr strain; purple line, abiotic control without addition of inoculum. The error bars represent standard deviations obtained using biological triplicate samples.

These data raised the issue of which species, arsenate or arsenite, was responsible for the growth inhibition. As shown in Fig. 4, arsenite is a much more powerful inhibitor of growth than arsenate, and surprisingly, the Arr enzyme system once more conferred high resistance. For example, in the presence of 1.0 mM arsenite, the growth of the parent was completely inhibited, while that of the Arr strain decreased by approximately 30% (in terms of total cellular protein), although there was a significant lag phase. With 0.5 mM arsenite, the parent strain had a growth phenotype similar to that observed in the presence of 10 mM arsenate, and maximum growth was approximately one-half that of the Arr strain (Fig. 4). These data therefore suggest that P. furiosus is inhibited not by arsenate (using 10 mM) (Fig. 3A) but by the arsenite (using 0.5 to 1.0 mM) (Fig. 4) that is generated abiotically (∼0.5 mM) (Fig. 3B), and Arr confers resistance to arsenite rather than to arsenate.

FIG 4.

Inhibition of P. furiosus grown on maltose by arsenite [As(III)]. Par and the Arr strain were grown on 5 g/liter maltose in the presence of 0.5 mM or 1 mM arsenite. The error bars represent standard deviations obtained using biological triplicate samples.

Higher intracellular arsenic concentrations in the Arr strain.

The intracellular arsenic concentration was determined in the parent and Arr strains when grown in the presence of 0.5 mM arsenite. As shown in Fig. 5, the arsenic concentration in the cell extract of the Arr strain was almost an order of magnitude higher than that of the parent strain. A similar result was observed in the presence of 0.1 mM arsenite, where no lag phase was found before reaching stationary phase in the parent (data not shown). This suggests that the higher concentration of intracellular arsenic in the Arr strain might be due to the presence of the Arr enzyme. To investigate this, the cytoplasmic extract of the Arr strain grown with 0.5 mM arsenite was applied to a desalting column, and the distribution of arsenic was determined. As shown in Fig. S4, most of arsenic was recovered in the low-molecular-weight fractions (<5 kDa), suggesting that the majority of the arsenic in the cell is not associated with proteins.

FIG 5.

Intracellular arsenic concentrations in the parent and Arr strains grown on maltose with 0.5 mM arsenite [As(III)]. The error bars represent standard deviations obtained using biological triplicate samples.

To determine the identity of the arsenic species present in both the parent and Arr strains, cells grown in the presence of 0.5 mM arsenite were analyzed by X-ray absorption spectroscopy (XAS). The results are shown in Fig. 6. The major component contributing to both spectra is arsenite, with the rest of the fitting represented by pentavalent methylated arsenicals, including mono- and dimethyl arsenate (MMA and DMA, respectively), trimethylarsine oxide (TMAO), and trimethylarsine sulfide (TMAS) (Tables S3 and S4; Fig. S5). However, the major difference between the samples was the distribution of these species, with methylated species representing only 27% ± 3% of the sample in the parent strain and 39% ± 7% in the Arr strain, indicating that the Arr strain was converting a larger percentage of arsenite to the methylated species. Because the Arr strain accumulates approximately 10× the intracellular arsenic compared to that by the parent (20 versus 2 nmol/mg cell protein), this corresponds to approximately 9 nmol/mg of methylated arsenicals, compared to ∼0.5 nmol/mg in the parent. The implications of the increased production of methylated arsenicals by the Arr strain on arsenite toxicity in P. furiosus are discussed below.

FIG 6.

X-ray absorption spectroscopy results. Normalized As-K near-edge X-ray absorption spectra collected in cell cultures exposed to 0.5 mM arsenite. The Arr strain is denoted by a red line with solid circles and the parent strain by a black line with empty circles.

DISCUSSION

Here, we present results of archaeal arsenate respiratory reductase (Arr) from P. aerophilum heterologously expressed in P. furiosus. The enzyme was affinity purified and is a heterodimer (ArrAB), although it is predicted to contain a third and membrane-bound subunit (ArrC). Even though the arrC gene was transcribed as expected, most of the Arr activity was in the cytoplasmic extract rather than in the membranes (see Fig. S3 in the supplemental material), and ArrC was not detected by mass spectrometry in the purified enzyme. This is probably because P. aerophilum uses a twin-arginine translocation (TAT) system for membrane transport, and arrC has a TAT signal motif for translocation to the periplasmic side of the membrane (13). However, P. furiosus does not contain the TAT export system (29), and in its absence, ArrC appears to be unstable and disassociates from ArrAB.

Although both Arr-Mo and Arr-W are active in reducing arsenate [As(V)], the Mo form is approximately 100-fold more active, suggesting that the native enzyme contains molybdenum rather than tungsten. Nitrate reductase (Nar) and thiosulfate reductase (Tsr), both dimethyl sulfoxide (DMSO) reductase family enzymes, were previously characterized from P. aerophilum and found to be more active with Mo than with W (26, 30). However, a W-containing aldehyde ferredoxin oxidoreductase (AOR) has also been characterized from P. aerophilum, and its growth has been shown to be dependent on tungstate and inhibited by molybdate (27, 31). While Arr-W has a lower catalytic activity than Arr-Mo, its affinity for arsenate is 20-fold higher, which is consistent with different metals in the catalytic site and could allow P. aerophilum to use arsenate as a terminal electron acceptor under tungsten growth conditions. This makes P. aerophilum Arr, to our knowledge, the only example of an active tungsten-containing arsenate reductase.

The Mo-form of Arr was extremely active in vitro in reducing arsenate with reduced methyl viologen and exhibited significant activity with reduced ferredoxin as the electron donor (Table 1). We assume that ferredoxin was also an electron donor in vivo and enabled P. furiosus to exhibit significant growth on peptides using arsenate as the electron acceptor compared to that of the parent strain (Fig. 2A). This is, to our knowledge, the only known example of P. furiosus utilizing a Mo enzyme to support growth. The decreased H₂ production also supported the diversion of reducing equivalents for arsenate reduction (Fig. 2C). Interestingly, we also observed a similar result for the parent strain in H₂ production, but in contrast to that with the Arr strain, there was no stimulation of growth in the presence of arsenate. This lack of growth may be due to the toxicity of abiotically generated arsenite [As(III)] to the parent strain but not the Arr strain. As discussed below, the physiological effects of heterologously expressed Arr appear to be related to removing arsenite rather than arsenate.

Purified Arr-Mo had significant arsenite reduction activity (methyl viologen linked) (Table 1), and we hypothesize that this is responsible for the Arr strain of P. furiosus being much more resistant to arsenite than the parent strain, both from endogenous [growth on 10 mM As(V) produces ∼0.5 mM As(III)] (Fig. 3B) and exogenous [1.0 mM As(III)] (Fig. 4) sources. For example, in arsenate-resistant bacteria, arsenate [As(V)] enters cells via phosphate transporters and is reduced by ArsC to arsenite [As(III)]. The arsenite is then extruded outside the cell by either an energy-dependent ArsB transporter or the Acr3 efflux permease (32, 33). In arsenite resistance, arsenite enters cells through aquaglyceroporins (34), and an arsenite methyltransferase (ArsM) is involved in the process of detoxification, where arsenite is sequentially reduced and oxidatively methylated to the volatile trimethylarsine (TMA) that leaves cells (35). Homologs of both ArsC and Acr3 are annotated in the P. furiosus genome, but parent cells did not contain detectable arsenate reductase activity. While the genome of P. furiosus does not contain an annotated ArsM, it is possible that it used nonspecific metal methylases or radical S-adenosylmethionine (SAM) enzymes to methylate reduced arsenic species. Best-fit calculations of the XAS spectra of both parent and Arr cells showed a mixture of methylated arsenicals, including MMA, DMA, and TMAS. MMA and DMA are both intermediates in the production of TMA, which readily reacts with sulfur species to produce the nontoxic TMAS (36). The overexpression of P. aerophilum Arr likely increased the pool of reduced arsenic species available for methylation, as the Arr strain showed an ∼22% shift from As(III) to methylated species compared to that by the parent strain, corresponding to 9 nmol of methylated arsenicals per mg of protein in the Arr strain (compared to 0.5 nmol/mg in the parent). Because the Arr strain accumulates approximately 10× the intracellular arsenic compared to that in the parent, this corresponds to approximately 9 nmol of methylated arsenicals per 1 mg of cells compared to ∼0.5 nmol/mg in the parent. Hence, the overexpression of Arr likely explains the resistance of the P. furiosus Arr strain to both arsenate and arsenite.

MATERIALS AND METHODS

Strain construction.

A genetically tractable Pyrococcus furiosus strain, COM1, was used for the heterologous expression of the arsenate respiratory reductase (Arr) from Pyrobaculum aerophilum (37). A knock-in cassette, containing the fragments shown in Fig. S1 in the supplemental material, was assembled using overlapping PCR (38). The upstream flanking region (UFR), the promoter of the gene encoding the S-layer protein (P_slp of PF1399), and the downstream flanking region (DFR) were amplified from P. furiosus genomic DNA, and the selection marker (P_gdh-pyrF) was amplified from pGLW021 (37). PAE1263 to PAE1265 were amplified using P. aerophilum genomic DNA as the template. A 9×His tag was placed at the C terminus of PAE1265, and the ribosomal binding sites (RBS) of P. furiosus pyruvate oxidoreductase γ subunit (PORγ; PF0971) and S-layer protein were placed upstream of arrC (PAE1264) and arrA (PAE1265), respectively. The genomic DNA of P. furiosus and P. aerophilum was isolated by using a ZymoBead Genomic DNA kit (Zymo Research). The P. aerophilum strain was purchased from DSMZ (DSMZ-7523), and medium 611 was used for cultivation. The transformants were grown as previously described (37). Genomic DNA was also used for PCR screening using GXL polymerase (TaKaRa, Clontech). The PCR screening was performed using a pair of primers outside the UFR and DFR. The PCR-confirmed colony was sequence verified using the Genewiz service.

Growth conditions.

The P. furiosus strains constructed and used in this study are listed in Table 2. The growth medium was used as previously described with some modifications (37). The inoculum was first transferred to the medium containing no exogenous Mo or W for 10 generations before being used in the experiments. All media contained 1× base salt, 1× trace minerals without molybdenum, 2 g/liter yeast extract, 5 g/liter maltose, 0.17 g/liter cysteine, 1 g/liter sodium bicarbonate, 1 mM potassium phosphate (pH 6.8), and 1 mM morpholinepropanesulfonic acid (MOPS) buffer (pH 6.8). Molybdate or tungstate (to a final concentration of 10 μM) and sodium arsenate (10 mM) or sodium arsenite (0.5 and 1.0 mM) were added as indicated in the text and figures. For the growth study with peptides, the medium contained 1× base salt, 1× trace minerals without molybdenum, 0.1 g/liter yeast extract, 5 g/liter casein, 0.17 g/liter cysteine, 1 g/liter sodium bicarbonate, 1 mM potassium phosphate (pH 6.8), 1 mM MOPS buffer (pH 6.8), 10 μM molybdate, and 10 mM sodium arsenate. The medium for the 20-liter fermentations contained 1× base salt, 1× trace minerals without molybdate, 5 g/liter yeast extract, 5 g/liter maltose, 5 g/liter casein, 1 g/liter cysteine, and 1 mM potassium phosphate (pH 6.8) with 10 μM molybdate or tungstate supplied where indicated. The 20-liter fermentations were carried out as previously described (39), and the harvested cells were frozen in liquid nitrogen and stored at −80°C. For growth studies, 1 ml of cultures was harvested by centrifugation at 14,000 × g for 5 min. The cells were lysed osmotically using 1 ml distilled H₂O, and cell debris was removed by centrifugation at 14,000 × g for 1 min. The supernatant was taken for protein estimation using protein assay reagent (Bio-Rad).

TABLE 2.

Strains used and constructed in this study

Trivial name	Strain name	Description	Reference
Par	COM1	ΔpyrF::Pgdh pyrF	37
Arr	MW0548	ΔpyrF::Pgdh pyrF Pslp (PAE1263-PAE1265-9× His)	This study

Purification of Arr.

All purification steps were performed under anaerobic conditions. Frozen cells were lysed osmotically in 25 mM sodium phosphate, pH 7.5, containing 1 mM dithiothreitol (DTT) and 50 μg/ml DNase I (5 ml per g frozen cells) for 1 h at room temperature in an anaerobic chamber (Coy, Grass Lake, MI). The cell extract was centrifuged at 100,000 × g for 1 h using a Beckman-Coulter Optima L-90K ultracentrifuge. The supernatant was collected and applied to a 5-ml His-Trap FF Ni-NTA column (GE Healthcare) using an NGC chromatography system (Bio-Rad). The column was equilibrated with 25 mM sodium phosphate, pH 7.5, containing 300 mM NaCl and 1 mM DTT (buffer A) before loading the sample, and 2% buffer B (buffer A containing 500 mM imidazole) was loaded with the cytoplasmic extract. The column was washed with buffer A, and the bound protein was eluted with a 20-column volume gradient from 0% to 100% buffer B. Active fractions containing methyl viologen (MV)-linked arsenate reduction were pooled and applied to an equilibrated 1-ml phenyl HP column (GE Healthcare) while diluting 5-fold with 50 mM sodium phosphate, pH 7.0, containing 1.5 M ammonium sulfate. A 30-column volume gradient from 20% to 100% 50 mM sodium phosphate, pH 7.0, was used to elute the bound protein.

Enzyme assays.

All assays were carried out anaerobically at 80°C. The routine arsenate [As(V)] reduction assay was carried out anaerobically using sealed cuvettes that contained 2 ml 100 mM MOPS, pH 7.5, 150 mM NaCl, 5 mM sodium arsenate, and 1 mM methyl viologen (MV). The assay cuvettes were preheated to 80°C, 0.1 mM titanium citrate was added to reduce MV, and the activity was measured by monitoring the decrease in absorbance at 600 nm using a 100 Cary UV-Vis spectrophotometer with a Peltier-based temperature controller from Agilent Technologies (Santa Clara, CA). The reaction was started with the addition of enzyme. An extinction coefficient of 8.25 mM⁻¹cm⁻¹ was used for reduced MV. One unit of activity was defined as 1 μmol of substrate reduced per min. The arsenite [As(III)] reduction assay was performed similarly with 5 mM sodium arsenite, which was purified using high-performance liquid chromatography (HPLC) as described below, as the substrate. For arsenite oxidation activity, 1 mM dichlorophenolindophenol (DCPIP) was used to measure the activity in the presence of 5 mM sodium arsenite. For P. furiosus ferredoxin (Fd)-dependent arsenate reduction activity, Fd was reduced using titanium citrate, and the oxidation of Fd was monitored at 425 nm (ϵ = 13 mM⁻¹cm⁻¹).

Metal concentration determination.

The concentrations of arsenic in spent media and the concentrations of Mo and W in purified proteins were measured using octopole-based inductively coupled plasma (ICP)-MS (7500ce; Agilent Technologies). Arsenite [As(III)] concentration was measured using HPLC for speciation followed by ICP-MS for concentration determination. A Zorbax SAX column (Agilent Technologies) was used in the speciation with a mobile phase of 60 mM KH₂PO₄ and a flow rate of 1 ml/min.

Intracellular arsenic concentration determination.

Cultures grown with 5 g/liter maltose, 2 g/liter yeast extract, and 0.5 mM sodium arsenite were centrifuged at 5,000 × g for 1 min to remove insoluble materials. The cells were washed three times with 1× P. furiosus base salt medium. Washed cells were lysed osmotically with 50 mM Tris-HCl, pH 8.0, containing 50 μg/ml DNase I (5 ml per g of wet cells) for 1 h at room temperature in an anaerobic chamber. Cell debris was removed by centrifugation at 14,000 × g for 10 min. The supernatant was collected, and the arsenic concentration was determined by ICP-MS. To estimate the size of the arsenic species in the cytoplasmic extract, the Arr strain was grown with 0.5 mM sodium arsenite and prepared as described above. After a 1-h lysis, the cell extract was centrifuged at 100,000 × g for 1 h, and the supernatant was applied to a 5-ml HiTrap desalting column (GE Healthcare) that was equilibrated with 50 mM Tris-HCl, pH 8.0. ICP-MS was used to measure the arsenic concentration.

XAS studies.

Cultures grown with 5 g/liter maltose, 2 g/liter yeast extract, and 0.5 mM sodium arsenite were centrifuged at 10,000 × g for 10 min to pellet cells. Cells were washed three times with 2× P. furiosus base salts medium and resuspended in the same medium. The Lucite cuvettes with a 2-mm-wide window were insulated with metal-free Mylar tape, filled with a syringe, frozen in liquid nitrogen, and sent on dry ice to the Canadian Light Source for XAS examination.

The XAS experiments were conducted at the BioXAS main station of the spectroscopy sector of the newly commissioned BioXAS facility (07ID-2) at the Canadian Light Source. Spectra of standard compounds were collected at Stanford Synchrotron Radiation Lightsource using beamline 7-3. In both cases, an Si(220) double crystal monochromator was used, and XAS data were measured as the primary fluorescence excitation spectrum using a 32-element germanium array detector (Canberra Industries, Meriden, CT, USA) positioned at 90° to the incident X-ray beam, with the sample at 45°. Samples were maintained at a temperature of 10 K in a liquid helium-flow cryostat (Oxford Instruments, Concord, MA, USA). The incident energy was calibrated with reference to the lowest energy As K-edge inflection point of elemental arsenic, which was assumed to be 11,867.0 eV. Primary data reduction and the least-square fitting of the samples’ near-edge XAS spectra were conducted using the EXAFSPAK suite of programs. Selected As-K near-edge XAS spectra, used for the least-square fitting, are shown in Fig. S5.

Other methods.

Hydrogen was measured in the headspace for growth studies with the 6850 Network GC system from Agilent Technologies. Purified ArrAB was analyzed by SDS-PAGE using precast Criterion TGX (4% to 15%) gels (Bio-Rad). Matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) and LC-MS/MS were performed by the PAMS Facility at the University of Georgia.

Data availability.

All data are included here and in the supplemental material. All strains used are stored as glycerol stocks internally and are available upon request.