Flexible bacterial strains that oxidize arsenite in anoxic or aerobic conditions and utilize hydrogen or acetate as alternative electron donors

Abstract

Arsenic is a carcinogenic compound widely distributed in the groundwater around the world. The fate of arsenic in groundwater depends on the activity of microorganisms either by oxidizing arsenite (As^III), or by reducing arsenate (As^V). Because of the higher toxicity and mobility of As^III compared to As^V, microbial-catalyzed oxidation of As^III to As^V can lower the environmental impact of arsenic. Although aerobic As^III-oxidizing bacteria are well known, anoxic oxidation of As^III with nitrate as electron acceptor has also been shown to occur. In this study, three As^III-oxidizing bacterial strains, Azoarcus sp. strain EC1, Azoarcus sp. strain EC3and Diaphorobacter sp. strain MC, have been characterized. Each strain was tested for its ability to oxidize As^III with four different electron acceptors, nitrate, nitrite, chlorate and oxygen. Complete As^III oxidation was achieved with both nitrate and oxygen, demonstrating the novel ability of these bacterial strains to oxidize As^III in either anoxic or aerobic conditions. Nitrate was only reduced to nitrite. Different electron donors were used to study their suitability in supporting nitrate reduction. Hydrogen and acetate were readily utilized by all the cultures. The flexibility of these As^III-oxidizing bacteria to use oxygen and nitrate to oxidize As^III as well as organic and inorganic substrates as alternative electron donors explains their presence in non-arsenic-contaminated environments. The findings suggest that at least some As^III-oxidizing bacteria are flexible with respect to electron-acceptors and electron-donors and that they are potentially widespread in low arsenic concentration environments.

Introduction

Arsenic (As) is a toxic metalloid that is found in groundwater by the natural weathering of rocks (Smedley and Kinniburgh 2002). Long term exposure to As contaminated drinking water increases the risk of cancer in the skin, liver, bladder and lungs (ATSDR 2007). In response to an ever increasing awareness of the health risks associated with As, the maximum concentration level (MCL) of drinking water standard in the United States was made stricter by lowering it from 50 to 10 μg/L in 2006 (EPA 2006).

Arsenate (As^V, H₂AsO₄^- and HAsO₄^2-) and arsenite (As^III, H₃AsO₃) are the predominant As species in circumneutral environments. As speciation is controlled mainly by redox conditions (Ascar et al. 2008; Beauchemin and Kwong 2006) with As^V being the dominant species in oxidizing environments while As^III predominates in reducing environments. As^III is adsorbed less than As^V in soil common metal oxides such as aluminum oxides (Hering 2005) and clay minerals (Lin and Puls 2000; Goldberg 2002), likewise As^III desorbs faster from iron (hydr)oxides than As^V (Tufano et al. 2008), therefore As^III is more mobile compared to As^V Thus processes governing the transformations between the different As species will have an important impact on the fate of As in the environment. Microbial activity plays a major role in the transformation between As^V and As^III (Oremland and Stolz 2003; Paez-Espino et al. 2009; Rhine et al. 2005). As^V can be reduced to As^III by dissimilatory As^V reducing bacteria, when an electron donor, such as organic matter or hydrogen (H₂), is present in the environment. The dissimilatory As^V reductase (arrA) are the responsible genes found in the As^V reducing bacteria (Saltikov and Newman 2003; Malasarn et al. 2004). On the other hand, As^III also can be oxidized to As^V by As^III oxidizing bacteria when an electron acceptor becomes available. A large variety of bacteria have been reported to contain the As^III oxidase (aroA) responsible for oxidation of As^III under aerobic (Silver and Phung 2005; Inskeep et al. 2007; Santini and vanden Hoven 2004) or anaerobic conditions (Hoeft et al. 2007; Rhine et al. 2007; Sun et al. 2010). The oxidation readily occurs under aerobic conditions. Since 1918, when bacteria capable of oxidizing As^III in aerobic environments was first recognized (Green 1918), numerous aerobic As^III oxidizing bacteria have been identified (Turner 1949; Wang and Suttigarn 2007; Krumova et al. 2008). A biodiversity of As metabolizing bacteria isolated from a variety of soil water systems have detectable As^III-oxidizing genes (Inskeep et al. 2007), indicating As^III oxidation plays an important role in the biogeochemical cycle of As.

Recently, As^III oxidation under anoxic conditions utilizing nitrate (NO₃^-) as an efficient alternative electron acceptor has been studied. Results from an urban lake have shown that nitrate (NO₃^-) levels in the anoxic zone are positively correlated with the formation of both As^V and As^V-adsorbing hydrous ferric oxides (Senn and Hemond 2002). NO₃^- injected into As-contaminated groundwater of Bangladesh was also shown to effectively lower the aqueous As concentration (Harvey et al. 2002). These results demonstrated the importance of NO₃^- as a controlling factor of As mobility in anoxic environments. In 2001, Hoeft et al. reported microbial As^III oxidation coupled with the reduction of NO₃^- to nitrite (NO₂^-) in an arsenic-contaminated soda lake in California (Hoeft et al. 2002). Oremland et al. identified the strain Alkalilimnicola ehrlichi sp. MLHE-1 as a bacterial species responsible for oxidizing As^III when linked to NO₃^- reduction to NO₂^- (Oremland et al. 2002). More recently, in 2006, Rhine et al., identified two anaerobic As^III oxidizing strains from an As-contaminated soil, Azoarcus sp. DAO1 and Sinorhizobium sp. DAO10 (Rhine et al. 2006; Rhine et al. 2007). Based on the stoichiometry of As^V-formed to NO₃^--consumed as well as the presence of nitrous oxides reductase (nosZ) genes, these strains, DAO1 and DAO10, appear to link the As^III oxidation to the complete denitrification of NO₃^- to dinitrogen gas (N₂).

The three strains characterized in this paper are Azoarcus sp. EC1-pb1, Azoarcus sp. EC3-pb1 and Diaphorobacter sp. MC-pb1 (to be referred as EC1, EC3 and MC in the rest of the manuscript). These strains were isolated from enrichment cultures (ECs) which were originally derived from sediments or sludge from pristine environments as inoculum (Sun et al. 2009). The ECs, from which strains EC1 and EC3 were isolated, linked As^III oxidation with complete denitrification as evidenced from measurements of N₂ production (Sun et al. 2009). The fact that these strains originate from environments with no known As contamination suggests they must have metabolic flexibility. The scope of this research was to better understand the flexibility of EC1, EC3 and MC to utilize different electron acceptors in the oxidation of As^III as well as different electron donors for the reduction of nitrate.

Materials and Methods

As^III-oxidizing pure cultures

Azoarcus sp. EC1-pb1, Azoarcus sp. EC3-pb1 and Diaphorobacter sp. MC-pb1were isolated from enrichment cultures EC1, EC3 and MC, respectively (Sun et al. 2009). The sequences of three isolated have been deposited in the GenBank with accession numbers: HM177479, FJ514096 and FJ514095, respectively.

Medium composition

The standard basal medium (pH 7.0-7.2) was prepared using ultra pure water (Milli-Q system; Millipore). The final composition of the basal medium in the batch experiments was (mg L¹): K₂HPO₄ (8.33); NH₄Cl (617.5); MgCl₂·6H₂O (173.3); MgSO₄·7H₂O (23.3); CaCl₂ (23.3), and trace elements in concentration (mg L^-1): FeC1₃·4H₂O (0.4); CoCl₂·6 H₂O (0.4); MnCl₂·4 H₂O (0.1); AlCl₃·6 H₂O (0.018); CuCl₂·2H₂O (0.006); ZnCl₂ (0.01); H₃BO₃ (0.01); (NH₄)₆Mo₇O₂₄·4 H₂O (0.01); Na₂SeO₃·5 H₂O (0.032); NiCl₂·6 H₂0 (0.01); EDTA (0.2); resazurin (0.04); HCl 36% (0.2 μL). 10 mM HCO₃^- (NaHCO₃) was used to buffer the pH of the medium.

Two different sets of experiments were carried out. In the first set of experiments, the electron acceptor was NO₃^- as KNO₃ (1.5 mM) and different electron donors were tested to be able to reduce NO₃^-, including NaAsO₂ (0.5 mM), CH₃COONa (0.11 mM), H₂ (0.4 mmol L_liq^-1), FeCl₂.4H₂O (1mM), Na₂S.9H₂O (0.125 mM) or MnSO₄.H₂O (0.5 mM). In the second set of experiments, the electron donor was As^III added as NaAsO₂ (0.5 mM) and different electron acceptor were tested as potential As^III oxidants, at the concentration of KNO₃ (1.5 mM), NaNO₂ (0.5 mM), O₂ (0.4 mmol L_liq^-1) or NaClO₃ (0.25 mM), respectively. The basal medium with the electron acceptor was sterilized by autoclaving, 20 min at 121°C, while NaHCO₃ and the electron donor solution were sterilized using membrane filtration (0.22 μm).

Pure cultures maintenance

The three pure cultures were maintained in the absence of oxygen (O₂) in a basal medium amended with 0.5 mM As^III as electron donor and 5 mM NO₃^- as electron acceptor. The pure cultures were incubated in 160 mL serum bottles, with a total liquid volume of 110 mL. After the medium was autoclaved, bottles were closed using butyl rubber septa in order to ensure an anaerobic atmosphere. A flush gas mixture of N₂/CO₂ (80%:20%) was used to purge the headspace and the medium for exclusion of O₂. The gas was introduced through a 0.22 μm filter to sterilize it (needle in-needle out). As^III solution was added by using a 0.22 μm filter in order to sterilize the solution and then the medium was inoculated with the pure culture (6.4% vol:vol). The culture bottles were incubated in the dark on an orbital shaker (115 rpm) at 30°C. Pure cultures were transferred to fresh medium every 10-14 d once the complete oxidation of As^III to As^V was proved with 85% As^III converted to As^V.

Experimental incubations

Batch experiments were performed in 160 mL serum bottles. All assays were conducted in duplicate. In order to avoid the contamination of carry-over NO₃^- from old cultures to fresh medium during the inoculation, the cultures (10% volume of previous culture) were centrifuged in 15 mL sterilized centrifuge tubes at 1,400 g for 20 min. The pellets were resuspended into same volume of sterilized MiliQ water, and after two cleaning cycles they were transferred to the experiments. Several controls were run in parallel. Abiotic controls were prepared without adding pure culture to prove biological nature of the reaction. Controls lacking the electron donor or the electron acceptor but inoculated were included to demonstrate that the reaction just occurs when both reactants are added to the medium. The culture conditions were the same as those described for culture maintenance. He/CO₂ (80%:20%) was utilized as an alternative flushing gas in the experiment in which aerobic oxidation of As^III was studied to avoid N₂ interference when measuring O₂. In the experiments in which H₂ was added as electron donor, a H₂ stock bottle was prepared. An empty 160 mL serum bottle closed with a butyl rubber septa was flushed through a 0.22 μm filter with pure N₂ gas during 20 min and with H₂/CO₂ (80%:20%) for an extra 20 min. In a similar way, a 100% O₂ stock bottle was prepared to use in the aerobic experiments, but the serum bottle was flushed with pure O₂ for 30 min. From these bottles, the adequate volume of gas for each experiment (1.6 mL H₂/CO₂ and 1 mL O₂) was taken with a syringe and injected in the corresponding assays.

Analytical Methods

Aliquot samples were taken from sealed anaerobic serum flasks by piercing the stoppers using sterile syringes with 16-gauge needles. All samples were centrifuged (10 min, 14,000 g) immediately after sampling and stored in polypropylene vials. As^V, sulfate (SO₄^2-), chlorate (ClO₃^-), NO₃^- and NO₂^- were analyzed by suppressed conductivity ion chromatography using a Dionex IC-3000 system (Sunnyvale, CA, USA) fitted with a Dionex IonPac AS11 analytical column (4 × 250 mm) and AG16 guard column (4 mm × 40 mm). During each injection the eluent (KOH) used was 30 mM for 10 min. Mn²⁺ was measured by using an inductively coupled plasma-optical emission spectrometry (ICP-OES) system model Optima 2100 DV from Perkin–Elmer TM (Shelton, CT, USA). Fe²⁺ was quantified by the 5-ortho-phenanthroline colorimetric method, using an UV-visible spectrophotometer (Agilent 8453, Palo Alto, CA, USA). Total Fe was obtained by reducing Fe³⁺ with a hydroquinone solution in pH 4.5 (0.05 M acetate buffer) (Fortune and Mellon 1938).

Headspace samples were taken with a pressure lock gas tight syringe (1710RN, 100 μL (22s/2″/2), Hamilton Company). N₂, O₂ and nitrous oxide (N₂O) were analyzed using a Hewlett Packard 5890 Series II gas chromatograph fitted with a Carboxen™ 1010 Plot column (30 m × 0.32 mm) and a thermal conductivity detector. Helium (He) was used as the carrier gas. H₂ and acetic acid were detected in an Agilent Technologies 7890A gas chromatography system. A Carboxen™ 1010 Plot Fused Capillary Column (30 m × 0.53 mm) with a thermal conductivity detector was used to analyze H₂ with N₂ used as the carrier gas. A Restek Stabilwax®-DA Column (30 m × 0.35 mm, ID 0.25 um) with flame ionization detector, and He used as a carried gas, was used to detect acetic acid.

DNA concentration was measured with a TBS-360 Mini-Fluorometer (Turner Biosystems, Sunnyvale, CA, USA) for DNA quantification using PicoGreen dsDNA Quantitation Reagent (Molecular Probes, Inc., Eugene, OR, USA) (Ahn et al. 1996).

Metabolizing genes including nosZ, RuBisCO genes, aroA and arrA genes were PCR amplified with specifically designed primer sets. Successfully purified PCR products were cloned to identify the sequences. The methodology and conditions for these techniques are described in the Electronic Supplementary Material (ESM).

Results

Anoxic As^III oxidation linked to nitrate reduction

Fig. 1A shows that As^III was oxidized to As^V in the full treatments amended with 0.5 mM As^III as electron donor and 1.5 mM NO₃^- as electron acceptor in the presence of inoculum. As^V formation began after a lag phase of 1.5 d, and complete oxidation was achieved after 3 d of incubation by observing the formation of 0.49 ± 0.01 mM of As^V as an average for the three pure cultures. No conversion of As^III to As^V was observed in control treatments run in parallel.

Fig. 1.

As^V formation by pure cultures EC1, EC3 and MC for oxidation of of As^III (0.5 mM) linked to NO₃^- (1.5 mM) reduction. Legends: panel A, As^V formation, complete treatments containing As^III and NO₃^- and inoculated with EC1 (■), EC3 (●) and MC (▲); biological controls (dashed lines) containing As^III but lacking NO₃^-: EC1 (□), EC3 (○) and MC (△); non inoculated control (x); panel B, NO₃^- consumption; panel C, NO₂^- formation. Biological treatments supplied with NO₃^- and As^III , EC1 (■), EC3 (●) and MC (▲); biological controls (dashed lines) supplied with As^III but no NO₃, EC1 (□), EC3 (○) and MC (△); non inoculated control (x).

The end product of the NO₃^- reduction was evaluated. NO₃^- and NO₂^- concentrations were measured throughout the experiment (shown in Fig. 1B and 1C, respectively). NO₃^- concentration decreased from 1.7 ± 0.2 mM to 1.1 ± 0.1 mM, which occurred concomitantly with NO₂^- production of 0.49 ± 0.01 mM in the full treatments, whereas no appreciable NO₂^- formation was observed in control assays. These results imply that for each mole of NO₃^- reduced, approximately one mole of As^III is oxidized and one mole of NO₂^- is formed (Table 1A) in accordance with the reaction shown in equation 1:

$${\mathrm{H}}_3{\mathrm{AsO}}_3+{{\mathrm{NO}}_3}^{\text{-}}\to {{\mathrm{HAsO}}_4}^{2\text{-}}+{{\mathrm{NO}}_2}^{\text{-}}+2{\mathrm{H}}^{+}$$ [eq. 1]

Table 1.

Stoichiometric calculations of the biological catalyzed reactions mediated by three pure culture strains EC1, MC and EC3^*

Part A	Δ(NO3-)	Δ(NO2-)	Δ(AsV)	$\frac{\Delta \left({\mathrm{As}}^{\mathrm{V}}\right)}{\Delta \left({\mathrm{NO}}_{3^{-}}\right)}$
	mM	mM	mM
EC1	0.480	0.487	0.481	1.00
MC	0.531	0.494	0.483	0.91
EC3	0.583	0.490	0.468	0.80

Part B	Δ(O2)	Δ(AsV)	$\frac{\Delta \left({\mathrm{As}}^{\mathrm{V}}\right)}{\Delta \left({\mathrm{NO}}_2\right)}$
	mmol Lliq-1	mM
EC1	0.217	0.488	2.25
MC	0.2436	0.556	2.28
EC3	0.233	0.561	2.41

Part C	Δ(H2)	Δ(NO2-)	Δ(NO3-)	$\frac{\Delta \left({\mathrm{H}}_2\right)}{\Delta \left({\mathrm{NO}}_{3^{-}}\right)}$
	mmol Lliq-1	mM	mM
EC1	0.355	0.353	0.387	0.92
MC	0.352	0.364	0.328	1.07
EC3	0.352	0.359	0.336	1.05

^*The stoichiometry has been calculated as the moles of electron donor consumed (moles of As^V produced are equivalent to the moles of As^III oxidized) per mole of electron acceptor utilized. Molar relationship of the As^III oxidation under denitrifying conditions (Part A), and aerobic conditions (Part B). Molar relationship of the H₂ oxidation under denitrifying conditions (Part C)

Aerobic As^III oxidation

The three pure cultures were tested for their capacity to oxidize As^III in aerobic conditions. As^V formation was observed when 0.35 ± 0.06 mmol O₂ L_liq^-1 was added as electron acceptor as shown in Fig. 2. In Fig. 2A, the first evidence of As^V formation was observed after only 1 d of incubation and the complete oxidation was achieved after 2.5 d in all the cultures. No As^III conversion was observed in the control assays (lacking either inocula or O₂). The O₂ concentration was also monitored (Fig. 2B). As shown in Table 1B, O₂ concentration was 0.32 ± 0.03 mmol L_liq^-1 in the controls after 3.5 d while the concentration in the full treatments was 0.112 ± 0.01 mmol L_liq^-1 , confirming the fact that biological As^III oxidation is linked to the O₂ consumption. The measured molar ratios (As^III:O₂) of the reaction were 2.25 to 2.4, depending on the strain, approximately corresponding to the reaction shown in equation 2:

$$2{\mathrm{H}}_3{\mathrm{AsO}}_3+{\mathrm{O}}_2\to 2{{\mathrm{HAsO}}_4}^{2\text{-}}+4{\mathrm{H}}^{+}$$ [eq. 2]

Fig. 2.

As^V production by pure cultures EC1, EC3 and MC with O₂ as electron acceptor. Oxidation of 0.5 mM of As^III with 0.35 mmol O₂ L_liq^-1. Legends: panel A, As^V formation. Biological treatments supplied with O₂ and As^III, EC1 (■), EC3 (●) and MC (▲); controls (dashed lines) supplied with As^III but no O₂, EC1 (□), EC3 (○) and MC (△); non inoculated control (x); panel B, O₂ consumption. Full treatments supplied with O₂ and As^III, EC1 (■), EC3 (●) and MC (▲); controls (dashed lines) supplied with O₂ but no As^III, EC1 (□), EC3 (○) and MC (△); non inoculated control (x).

Oxidation of As^III by alternative electron acceptors

Aside from NO₃^- and O₂, NO₂^- and ClO₃^- were also tested as potential electron acceptors. The summary of the results obtained from testing electron acceptors is shown in Table 2. Strains MC and EC3 were able to link a partial oxidation of As^III when NO₂^- was present in the active treatments as electron acceptor. An As^V concentration of 0.24 ± 0.04 mM was reached in less than 18 d, and no further increase was observed after an additional 7 d of incubation. As^V formation was not detected when strain EC1 was used as the inoculum. No changes in As speciation occurred in the control bottles (lacking either NO₂^- or inocula) with an As^V concentration of 0.068 ± 0.01 mM at the end of the experiment. ClO ^-₃ was not utilized as an electron acceptor by any of the pure cultures even after 24 d of incubation for strains MC and EC3. A small As^III oxidation was observed by strain EC1, but it accounted for less than a 10% in 54 days of incubation; the concentration of As^V in the full treatment was 0.121 ± 0.051 mM, while in the control bottles was constant at 0.070 ± 0.02 mM throughout the entire experiment.

Table 2.

Summary of As^III oxidation^* with various electron acceptors tested

	Electron acceptors+
	KNO3		NaNO2		O2		NaClO3
Pure Cultures	AsV formation (mM)	Time to completion (days)	AsV formation (mM)	Time to completion (days)	AsV formation (mM)	Time to completion (days)	AsV formation (mM)	Time to completion (days)
EC1	0.496 ± 0.011	4	0.053 ± 0.052	54	0.473± 0.002	2.5	0.121 ± 0.051	54
MC	0.496 ± 0.004	3	0.240 ± 0.008	24	0.500 ± 0.010	2.5	0.072 ± 0.002	24
EC3	0.485 ± 0.028	3	0.284 ± 0.116	24	0.472 ± 0.018	2.5	0.063 ± 0.021	24

^*As^III concentration was 0.5 mM.

⁺Concentrations of the different electron acceptor tested were 1.5 mM KNO₃, 0.5 mM NaNO₂, 0.35 mmol O₂ L_liq^-1 and 0.25 mM NaClO₃. Electron acceptors that have been positively utilized by the cultures to oxidize As^III are shown in bold text.

Use of alternative electron donors under nitrate-reducing and aerobic conditions

In order to test the ability of the three pure cultures to reduce NO₃^- with different substrates under anaerobic conditions, four inorganic electron donors and an organic electron donor were evaluated. The three strains were able to reduce NO₃^- when H₂ or acetate was amended to the treatment, but the other electron donors tested, ferrous iron (Fe²⁺), sulfide (S^2-) and manganese (Mn²⁺), were not utilized by the cultures to reduce NO₃^-.

Fig. 3A shows the consumption of H₂ linked to the NO₃^- consumption represented in Fig. 3B. H₂ and NO₃^- removal were first noticeable after 1 d, and the reaction was completed by day 1.5. H₂ concentration decreased from 0.356 ± 0.014 mmol H₂ L_liq^-1 to 0.0036 ± 0.0001 mmol H₂ L_liq^-1 in the complete treatments, while it was not oxidized in the controls treatments (lacking either NO₃^- or inocula). In parallel, NO₃^- concentration decreased an average of 0.350 ± 0.024 mM in the course of the experiment in the complete treatments and no NO₃^- removal was observed in the control treatments (lacking either H₂ or inocula). NO₃^- was reduced reduction to NO₂^- which accumulated to a concentration of 0.359 ± 0.004 mM at the end of the experiment (Fig. 3C). The experimental results indicated a molar ratio of H₂ consumed and NO₃^- reduced of 1.12 (Table 1C). This was close to the stoichiometric molar ratio of 1 shown in equation 3:

$${\mathrm{H}}_2+{{\mathrm{NO}}_3}^{\text{-}}\to {{\mathrm{NO}}_2}^{\text{-}}+{\mathrm{H}}_2\mathrm{O}$$ [eq. 3]

Fig. 3.

H₂ oxidation linked to NO₃^- reduction. Treatments with 0.35 mmol H₂ L_liq^-1 and 1.5 mM of NO₃^-. Legends: panel A, H₂ oxidation biological treatments supplied with NO₃^- and H₂, EC1 (■), EC3 (●) and MC (▲); controls (dashed lines) supplied with H₂ but no NO₃^-, EC1 (□), EC3 (○) and MC (△); non inoculated control (x); panel B, decrease of NO₃^- concentration; panel C, NO₂^- formation. Full treatments supplied with NO₃^- and H₂ and inoculated with EC1 (■), EC3 (●) and MC (▲); controls (dashed lines) with only H₂ and no NO₃^-: EC1 (□), EC3 (○) and MC (△); non inoculated control (x).

H₂ was also tested as a possible electron donor under aerobic conditions. To pursue this experiment, the three strains were incubated with an atmosphere consisting of 0.503 mmol H₂ L_liq^-1 and 0.251 mmol O₂ L_liq^-1 and the H₂ consumption was monitored. The initial H₂ concentration was completely consumed in less than 20 hours in bottles containing O₂ and inoculated with pure cultures. No H₂ consumption was observed in the controls (lacking either inoculum or O₂), confirming a biological mediated oxidation of H₂ by O₂.

As^III oxidation after pre-incubation of cultures with alternative electron donors and acceptors

The ability of the pure cultures to oxidize As^III with NO₃^- was evaluated after being previously incubated with H₂/NO₃^- or H₂/O₂ (Fig. 4A and 4B, respectively). In both cases, As^III oxidation linked to NO₃^- reduction was confirmed following the pre-incubation. Complete As^III oxidation took place in less than 1 day in both experiments. NO₃^- consumption was also confirmed in the full treatments with inoculum and As^III. These results indicate that the use of alternative electron acceptors and donors did not suppress the capacity of the bacteria to oxidize As^III. The pre-incubation with H₂/NO₃^- or H₂/O₂ likely increased the cell concentration in the assays, which may have been the reason for a faster oxidation of As^III compared to earlier experiments in which As^III grown cells were used as inoculum.

Fig. 4.

Ability of the pure cultures to oxidize As^III under denitrifying conditions after being incubated with: panel A, 0.3 mM NO₃^- and 0.4 mmol L_liq^-1 H₂ or panel B, 0.45 mmol H₂ L_liq^-1 and 0.3 mmol O₂ L_liq^-1 . Treatment with 0.5 mM of As^III and 0.3 mM of NO₃^-: EC1 (■), EC3 (●) and MC (▲). Controls (dashed lines) with only 0.5 mM of As^III and no NO₃^-: EC1 (□), EC3 (○) and MC (△). Non inoculated controls (x).

The difference in cell concentration when the cultures are incubated either with NO₃^-/As^III or H₂/O₂ was estimated by quantifying the DNA at the beginning and at the end of the experiments. These two experiments were the two extreme scenarios, the lowest cell yield (NO₃^-/As^III) and the highest cell yield (H₂/O₂). In both cases, DNA production was shown to be clearly linked to the use of As(III) or H₂ as a growth substrate (Fig S1 and Fig S2 in ESM). There was no increase in DNA concentration in non-inoculated controls, nor in controls lacking either the electron acceptor or electron donor. The concentration of DNA was approximately five times greater at the end of the experiment when the pure cultures were incubated with H₂/O₂ compared with NO₃^-/As^III. This outcome also confirms the hypothesis that the cell yield was significantly increased with pre-incubation with the better growth substrate, H₂.

Functional gene PCR and cloning

Functional gene PCR targeting aroA and arrA genes, nosZ and RuBisCO genes were performed on all three As^III-oxidizing, NO₃^--reducing pure cultures. Various primer sets were used to detect the existence of aroA or arrA genes. No PCR products were observed for any of the primer sets for aroA, nosZ which are genes responsible for As^III oxidation (Inskeep et al. 2007) and nitrous oxide reductase (Scala and Kerkhof 1998), respectively. PCR products were found for primer set #1 of arrA, a gene encoding for dissmilatory As^V reductase (Saltikov and Newman 2003) but is related in structure to As^III oxidizing enzymes believed to catalyze the reverse reaction of the dissimilatory reductase (Richey et al. 2009). However, the sequence of the amplicons did not correspond to arrA. PCR products were obtained from a primer set targeting the large unit, cbbL, of ribulose-1,5-biphosphate carboxylase oxygenase genes (RuBisCO genes, an indicator of autotrophy) in all the three pure cultures (Fig. S3 in ESM). The amplicons were 800 bp in agreement with the expected size (Elsaied and Naganuma 2001). The amplicons of two of the strains, EC1 and EC3, were also checked by cloning and sequencing. The sequences were deposited in GenBank with accession numbers JN008173 and JN008174, respectively. The EC1 and EC3 derived sequences corresponded 95% and 80% similarity to the RuBisCO large unit cbbL genes of Nitrosomonas sp. ENI-11 [GenBank accession number AB061373] (Hirota et al. 2002) and an uncultured organism [GenBank accession number AB505078] (Kojima et al. 2009), respectively.

Discussion

The three As^III-oxidizing bacteria presented in this study, EC1, EC3 and MC, were found to be flexible in their ability to use different electron acceptors and electron donors. The ability to oxidize As^III in anoxic nitrate-reducing as well as aerobic conditions is unique and has not been observed before by a single strain. The strains therefore have the ability to adapt to changing redox conditions in the environment. Functional gene PCR was used to determine which genes might be involved in arsenic metabolism. As^III oxidase (Aro) and As^V respiratory reductase (Arr) are two As metabolizing enzymes within the dimethyl sulfoxide (DMSO) reductase family (Silver and Phung 2005) that are involved in growth linked metabolism of As^III and As^V, respectively. However this approach provided no definitive evidence for the presence of aroA nor arrA genes in any of the three pure cultures.

As^III was not the only substrate that can be used as electron donor by the isolated strains. H₂ and acetate were readily oxidized under NO₃^--reducing conditions. Also H₂ was shown to be utilized as electron-donor with O₂. The flexibility to utilize different substrates in order to support growth seems to be a common characteristic of many As^III oxidizing isolates. The facultative, As^III oxidizing chemoautotroph Alkalimnicola ehrlichii strain MLHE-1, belonging to the γ-Proteobacteria group, was able to grow with As^III, H₂, sulfide, thiosulfate and acetate while reducing NO₃^- to NO₂^-. Aerobic oxidation was also possible with H₂ and acetate (Hoeft et al. 2007). Two autotrophic As^III oxidizing denitrifying bacteria, Azoarcus strain DAO1 and Sinorhizobium strain DAO10 were also capable of using H₂, acetate, glucose, lactate and other organic substrates as electron donors (Rhine et al. 2006). Three newly isolated aerobic As^III oxidizing strains Ancylobacter strain OL-1, Thiobacillus strain S-1 and Hydrogenophaga strain CL-3 can also oxidize different sulfur species (thiosulfate, elemental sulfur and sulfide) to sulfate (Garcia-Dominguez et al. 2008). Ammonium was also found to be an electron donor for strains S-1 and CL-3 and the latter can also oxidize NO₂^- to NO₃^- under aerobic conditions.

None of the three bacterial strains characterized in this study were able to completely denitrify NO₃^- to N₂. Instead, NO₃^- was partially reduced to NO₂^-, which accumulated as a product at the end of the experiments. The enrichment cultures, EC1 and EC3, from which Azoarcus sp. strains EC1 and EC3 were isolated, respectively, were shown to produce N₂ gas as the final product of denitrification linked to As^III oxidation (Sun et al. 2009). Also, strain DAO1 isolated by Rhine et al. (Rhine et al. 2006), closely related to Azoarcus (96% 16S rRNA gene sequence similarity to strains EC1 and EC3) was able to perform complete denitrification. The dilution to extinction technique used to isolate the pure cultures may have favored subpopulations lacking the full set of denitrifying genes, in particular the nosZ gene. The functional PCR assay confirmed the absence of nosZ genes which corresponds to their inability to completely denitrify of NO₃^- to N₂.

The presence of RuBisCO genes, a key enzyme of CO₂ fixation for cell growth (Miziorko and Lorimer 1983) in all of the strains studied here, provides DNA-based evidence of autotrophic nature of pure cultures. Two amplicon sequences were confirmed to be related to RuBisCO large unit cbbL genes. There are several examples of autotrophic anoxic As^III-oxidizing bacteria reported before. Azoarcus sp. DAO1, Sinorhizobium sp. DAO10 and Alkalimnicola ehrlichii sp. MLHE1 were shown to contain RuBisCO genes (Hoeft et al. 2007; Rhine et al. 2006). Likewise, many aerobic As^III-oxidizing bacteria also contain RuBisCO genes such as Ancylobacter strain OL-1, Thiobacillus strain S-1 and Hydrogenophaga strain CL-3 (Garcia-Dominguez et al. 2008).

The most well studied aerobic As^III-oxidizing strains have been isolated from the As bearing sediments of a gold mine in Australia (Santini et al. 2000; Santini et al. 2002). Earlier anoxic As^III-oxidizing isolates have either been isolated from an alkaline lake, with high levels of As, in California (Oremland et al. 2002; Oremland et al. 2004; Oremland et al. 2005) or were isolated from As contaminated soil and sediments (Rhine et al. 2006). The three As^III-oxidizing bacteria presented in this study (EC1, EC3 and MC) were isolated from anoxic As^III-oxidizing EC and a mixed culture originating from pristine, non-contaminated environments (Sun et al. 2009).

The ability of As^III-oxidizing strains to survive in environments in which As is present in negligible concentrations is most likely related to their flexibility in using different electron donors. The isolates are able to survive with environmentally relevant electron donors such as H₂ and acetate, which can be expected in anaerobic conditions where organic matter is decomposing. Therefore, anoxic As^III oxidizing bacteria like those described in this study are probably more widespread in the environment than previously thought due to the high level of flexibility in the use of substrates.