Environ. Sci. Technol. 2005, 39, 9569-9573
Environmental Microbes Can Speciate and Cycle Arsenic E. DANIELLE RHINE,† ELIZABETH GARCIA-DOMINGUEZ,† C R A I G D . P H E L P S , †,‡ A N D L . Y . Y O U N G * ,†,‡ Biotechnology Center for Agriculture and the Environment, and Department of Environmental Sciences, Cook College, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08901
Naturally occurring arsenic is found predominantly as arsenate [As(V)] or arsenite [As(III)], and can be readily oxidized or reduced by microorganisms. Given the health risks associated with arsenic in groundwater and the interest in arsenic-active microorganisms, we hypothesized that environmental microorganisms could mediate a redox cycling of arsenic that is linked to their metabolism. This hypothesis was tested using an As(V) respiring reducer (strain Y5) and an aerobic chemoautotrophic As(III) oxidizer (strain OL1) both isolated from a Superfund site, Onondaga Lake, in Syracuse, NY. Strains were grown separately and together in sealed serum bottles, and the oxic/anoxic condition was the only parameter changed. Initially, under anoxic conditions when both isolates were grown together, 2 mM As(V) was stoichiometrically reduced to As(III) within 14 days. Following complete reduction, sterile ambient air was added and within 24 h As(III) was completely oxidized to As(V). The anoxic-oxic cycle was repeated, and sterile controls showed no abiotic transformation within the 28day incubation period. These results demonstrate that microorganisms can cycle arsenic in response to dynamic environmental conditions, thereby affecting the speciation, and hence mobility and toxicity of arsenic in the environment.
Introduction Tasteless, odorless, and colorless, arsenic has been an effective poison for mold and mayhem. Naturally occurring arsenic, derived from the weathering of As minerals and volcanic activity, can be detected in terrestrial and aquatic environments under both oxic and anoxic conditions. Of the four oxidation states (-3, 0, +3, and +5), arsenite [As(III); H3AsO42- or H2AsO3-] and arsenate [As(V); H2AsO4- or HAsO42-] are the forms most commonly detected in the environment. Thermodynamically, As(V) would be the predominant state found under oxic conditions, whereas As(III) would be expected in anoxic environments. Both oxidation states, however, have been detected under both oxic and anoxic conditions (1-3). It is generally thought that As(V) can bind to a variety of minerals over a broad pH range, and As(III) can bind over a tighter pH range mostly to Fe(III)-oxyhydroxides and metal sulfides (4, 5); however, the * Corresponding author phone: (732) 932-8165 ext. 312; fax: (732) 932-0312; e-mail:
[email protected]. † Biotechnology Center for Agriculture and the Environment. ‡ Department of Environmental Sciences. 10.1021/es051047t CCC: $30.25 Published on Web 11/10/2005
2005 American Chemical Society
relative affinity is dependent on the solution composition and iron oxide characteristics, i.e., when phosphate is present As(III) will bind over a wider pH range than As(V) (6). Microorganisms have been isolated from the environment which are able to utilize arsenic as either an electron donor or an electron acceptor. Oxidation of As(III) coupled to O2 is highly energetically favorable for heterotrophic bacteria (7-10) and autotrophs, which can fix CO2 into cell matter under aerobic (1, 2, 11-14) and nitrate reducing conditions (15, 16). To some extent this may be a way for the microbes to detoxify their immediate surroundings, e.g., microbial oxidation would transform As(III) which is more stable and mobile to As(V) which is more strongly sorbed and less toxic (1, 2). On the other hand, As(V) reduction to As(III) is an energetically favorable reaction which various anaerobic bacteria take advantage of by using As(V) as the electron acceptor in respiration (1, 2, 17-19). We have isolated and characterized nine novel prokaryotes which either oxidize As(III) or reduce As(V). The four reducers are anaerobes and use As(V) as a respiratory electron acceptor (strains Y5, AKAR3, ARCL1, GS4). The remaining five are facultative chemoautotrophic oxidizers that are able to fix CO2 and rapidly oxidize As(III) to As(V). The phylogenetic relationship of our isolates and other arsenic-active prokaryotic organisms is shown in Figure 1. Based on 16S rRNA gene sequences, most of the chemoautotrophic oxidizers tend to be members of the R-Proteobacteria subdivision (including strain OL1), whereas the heterotrophic strains belong to the subdivision β-Proteobacteria. Heterotrophic As(III) oxidation is typically thought of as a detoxification process in which As(III) oxidation occurs on the outer surface of the inner membrane, thereby making transport into the cell less likely (20, 21). Preliminary biochemical studies of an autotrophic As(III)-oxidizer, strain NT-26, showed that the arsenite oxidase (Aro) is periplasmic and As(III) induces its synthesis (12, 22). These results suggest that As(III) oxidation may have evolved separately for the two types of metabolic systems. Unlike the oxidizers, the As(V)-respiratory reducers are more phylogenetically diverse (Figure 1), including Grampositive strains (including strain Y5), γ-Proteobacteria, and -Proteobacteria, suggesting that this may be a widespread process and not a phylogenetic descriptor. All of the known As(V)-respiratory reducing isolates are able to use other electron acceptors for growth (i.e., sulfate, sulfur, thiosulfate, nitrate, nitrite, Fe(III), and selenate), and to our knowledge no obligate As(V)-respiratory reducer has been reported (1, 2). Preliminary studies of As(V)-respiratory reducing strains Shewanella ANA-3 (23), Chrysiogenes arsenatis (24), and Desulfitobacterium hafniense (25) show that the As(V)respiratory reductase, arrAB, gene sequences are similar, suggesting that these genes define a new class of reductases specific for As(V)-respiratory reduction. The biochemical process for As(V)-respiratory reduction may have evolved a long time ago, or the arrAB genes might have been widely distributed through horizontal gene transfer (23). Interestingly, at some environmental sites both arsenic oxidizers and reducers have been found (e.g., Onondaga Lake), suggesting that the two processes, As(III) oxidation and As(V)-respiratory reduction can both occur within the same habitat (Figure 2). Thus, it is conceivable that if conditions are oxic, then arsenic oxidizing strains could grow and rapidly oxidize As(III) to As(V). A shift to anoxia would then allow those microorganisms that are able to reduce As(V) to As(III) through anaerobic respiration to grow. As long as the arsenic remains within the habitat of the microbes, VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9569
FIGURE 1. An unrooted phylogenetic tree of the known arsenic-utilizing bacteria. A neighbor-joining tree derived using ClustalX comparing 1200 base pair segments of the 16S rRNA gene of the known arsenic-utilizing bacteria, the bar equals 10% difference. As(V) respiratory reducers (red lines), aerobic As(III) oxidizers (blue lines), anaerobic As(III) oxidizers (green lines). All color-coded and bold strains have been isolated in our laboratory.
FIGURE 2. Conceptual model of how an As(III)-oxidizer, strain OL1, and an As(V) respiratory reducer, strain Y5, may be able to cycle arsenic under changing redox potentials. in theory these processes can continue to cycle as a function of oxic-anoxic changes, thus affecting arsenic speciation, mobility, and/or toxicity. In an unsaturated soil, microbially mediated aerobic arsenic cycling has been demonstrated; however, the arsenic was transformed through detoxification processes and was not used for cell growth (26). Here we test the hypothesis that microorganisms that can utilize arsenic for cell growth, via autotrophic As(III)-oxidation or As(V)-respiratory reduction, could cycle arsenic within the environment (Figure 2). Both isolated from Onondaga Lake, a Superfund site in Syracuse, NY, strain Y5 (Desulfosporosinus sp.), an arsenaterespiratory reducing obligate anaerobe, and OL1 (Ancylobacter sp.), an aerobic facultative chemoautotrophic arseniteoxidizer, were used to test this hypothesis. We grew them together in sealed serum bottles, provided all essential nutrients, and changed only the oxic/anoxic conditions. Each strain was also grown separately and subjected to exactly the same cycle of anoxic and oxic conditions to confirm their activity in parallel when in pure culture.
Materials and Methods Initial Growth and Medium Conditions. To grow the initial stock culture of the As(V)-respiratory reducing strain Y5, a mineral salts medium, with pH adjusted to 7.2 was used as described by Evans et al. (27), with the exception that sodium 9570
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 24, 2005
nitrate was not included. The culture was prepared under strict anaerobic conditions and amended with 5 mM As(V) (Na2HAsO4‚7H2O) as the electron acceptor and 1mM benzoate as the carbon source. A 100 mL aliquot of the mineral salts medium was anaerobically dispensed into a 160-mL serum bottle with argon in the headspace. The bottle was sealed with a rubber stopper and aluminum crimp seal to prevent a loss of argon or introduction of O2 into the headspace, and incubated statically at 30 °C in the dark. Periodically, the culture was sampled and measured for As(V) reduction to As(III). In a 500-mL Erlenmeyer flask with a foam stopper, the As(III)-oxidizing strain OL1 was grown in 100 mL of the defined mineral salts medium, which contained 10 mM HCO3- (NaHCO3) as the sole carbon source, 5mM As(III) (NaAsO2) as the electron donor, and O2 as the electron acceptor. Incubation was carried out at 30 °C in the dark on a rotary shaker (120 rpm) to ensure complete aeration. The culture was periodically sampled and measured for As(III) and As(V) concentration to determine when complete oxidation had occurred. Following complete utilization of the added arsenic by each strain, cultures were concentrated 5× by centrifugation, and cells were washed with phosphate buffered saline to remove residual arsenic. Arsenic Cycling Batch Culture Experiment. Under strict anaerobic conditions, aliquots of 45 mL of the mineral salts medium were anaerobically dispensed into 160-mL serum bottles with argon in the headspace, and the bottles were sealed with rubber stoppers and aluminum crimp seals to maintain an anaerobic environment. All bottles were inoculated with 2 mL of concentrated (5×) stock culture of either both strains together or each individually. To the bottles in which the As(III) oxidizer, strain OL1, was incubated alone, the medium contained 2 mM As(III) and 10 mM HCO3-. An additional 2 mM As(III) was added on day 27. When strain Y5, the As(V)-respiratory reducing isolate, was incubated
alone the medium was amended with 2 mM As(V) and 1 mM benzoate, and an additional 2 mM As(V) was added on incubation day 15. When both strains were incubated together the medium initially contained 2 mM As(V), 10 mM HCO3-, and 1 mM benzoate. Sterile controls, autoclaved three times on consecutive days, were also included for each condition. All bottles were incubated at 30 °C in the dark on a rotary shaker (120 rpm) for 28 days. To begin the arsenic cycling batch culture experiment, all bottles were initially incubated under anaerobic conditions with As(V) supplied as the only electron acceptor, in the case of OL1 alone As(III) was added. Samples were periodically analyzed for As(V) transformation to As(III) by monitoring the reduction of As(V) concentration and increase in As(III) concentration. Once the As(V) had been completely reduced to As(III), sterile ambient air was added to each bottle to change to an oxic environment. In a biological hood 60 mL of ambient air was passed through a filter and injected into each of the bottles. Each bottle was then swirled and vented to allow for the exchange of gases. The process was repeated twice, with a total of 180 mL of sterile ambient air added, to ensure complete aerobiasis. Samples were again periodically taken to monitor As(III) and As(V) concentrations, and once oxidation of As(III) was complete the bottles were degassed with sterile argon to change the redox state back to anoxic. The anoxic/oxic cycle occurred over a 28-day incubation period. Analytical Methods. To ensure anaerobic conditions, all samples were taken by piercing the stoppers using 1.0-mL sterile plastic syringes with 16-gauge needles flushed with argon. A 0.7-mL sample from each bottle was transferred to a Spin X vial with 0.2-µm pore-size nylon filters (VWR Scientific, West Chester, PA) and centrifuged at 14 000 rpm for 1 min in a benchtop microcentrifuge to remove all suspended materials. The filtered samples were then transferred to glass GC vials, capped, and analyzed for As(III) and As(V) concentrations. A high-performance liquid chromatograph (HPLC) (Shimadzu, Columbia, MD) was used to measure arsenite [As(III)] and arsenate [As(V)] concentrations. The HPLC had a Hamilton PRP-X100 anion exchange column, and a UV absorbance detector set at 195 nm. The mobile phase used was 30 mM sodium monobasic phosphate buffer (pH adjusted to 6), with a flow rate of 1 mL min-1, and the injection volume was 10 µL (19). The peaks were quantified by comparison with external standards using integrated calibration curves.
FIGURE 3. Arsenic cycling by As(III) oxidizing strain OL1 and As(V) respiratory reducing strain Y5 under cycling anoxic/oxic conditions. The addition of sterile ambient air was on days 14 and 27 (f), and the bottles were degassed with sterile argon prior to inoculation and on day 15 (dashed arrow). OL1 and Y5 cells were incubated in triplicate bottles with 10 mM HCO3-, 1 mM benzoate, and 2 mM As(V). Active bottles were sampled for As(III) (9) and As(V) (0), and killed controls were also sampled for As(III) (b) and As(V) (O).
Results When both arsenic-utilizing strains were incubated together initially under anaerobic conditions (Figure 3), 2 mM As(V) (open squares) was reduced stoichiometrically to As(III) (closed squares) over the course of 14 days. When reduction was complete, sterile ambient air was added to the vessels (solid arrow) and within 24 h of aerobiosis As(III) was completely oxidized to As(V). At this point in time, day 15, the only action taken was to degas the headspace with sterile argon for a second cycle of anaerobiosis (dotted arrow). This was followed by complete reduction of As(V) to As(III) within 12 days. A second cycle of aerobiosis was initiated on day 27 (solid arrow), resulting in complete oxidation of As(III) to As(V) which again occurred within 24 h. Hence, two full anoxic-oxic cycles were carried out. No reduction of As(V) (open circles) nor generation of As(III) (closed circles) occurred in heat killed controls that had been inoculated with both OL1 and Y5 prior to autoclaving. When the aerobic strain OL1 was incubated alone (Figure 4A), no oxidation of the added 2 mM As(III) took place during the initial anoxic cycle. Similar to Figure 3, following the addition of the sterile air on day 14 (oxic cycle) rapid and
FIGURE 4. Arsenic transformation by (A) strain OL1 cells inoculated alone in triplicate bottles containing 10 mM HCO3- and 2 mM As(III) initially, and (B) strain Y5 alone amended with 1 mM benzoate and 2 mM As(V) initially. Prior to incubation and on day 15 the bottles were degassed with sterile argon (dashed arrow). Sterile ambient air was added on days 14 and 27 (f). In the bottles inoculated with OL1 only (A), As(III) (2) and As(V) (4) concentrations in the active bottles, and As(III) (b) and As(V) (O) in the killed controls were measured. In the Y5 only bottles (B), As(III) ([) and As(V) (]) concentrations were again measured and compared to Y5 only killed controls As(III) (b) and As(V) (O). complete oxidation of the As(III) occurred. After degassing with sterile argon on day 15 (second anoxic cycle), no change in As(V) levels occurred. On day 27, when sterile ambient air was again added, an additional 2 mM As(III) was also provided since none was available in the absence of the reducing organism. Again, stoichiometric oxidation of As(III) to As(V) occurred in 24 h of aerobic incubation. Noteworthy is the VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9571
observation that strain OL1 readily withstands substantial periods of anoxia multiple times, without apparent harm and is able to rapidly oxidize As(III) once oxygen is available. This is apparent in both the single organism pure culture system and when both strains were incubated together. In parallel cultures, the anaerobic strain Y5 when grown alone (Figure 4B) completely reduced the available As(V) to As(III) within the 14-day anoxic cycle. No changes in arsenic speciation occurred during the 24-hour oxic cycle that followed. During the second anoxic cycle, an additional 2 mM of As(V) was added and fully reduced to As(III) by Y5 again. Hence, despite the fact that Y5 is an obligate anaerobe, transient oxic conditions do not inhibit its activity once anoxic conditions are reestablished.
Discussion Changing redox conditions are only one of many fluctuating environmental parameters encountered by natural communities of microorganisms. To rapidly respond to fluctuating redox gradients, one can hypothesize that a diverse spectrum of species is present but they are not all active. As conditions become unfavorable for some species, they enter dormancy, and as conditions become favorable for others, those microorganisms ramp up their metabolism. Selection for environmentally successful organisms then would favor those that can withstand these frequent and repeated fluctuations, such as strains Y5 and OL1. To our knowledge, this is the first demonstration in vitro and proof of principle that microorganisms can mediate redox cycling of arsenic. Oxygen alone cannot oxidize As(III) in our experimental system as shown in the controls in Figures 3, 4A, and 4B. Only when the active organism, strain OL1, was added did rapid and complete arsenic oxidation take place (Figures 3 and 4A). Abiotically, arsenic can be oxidized by soil or sediment mineral components such as manganese oxide (28). As(III) oxidizers such as strain OL1 could potentially play a significant part in transforming arsenic from the more toxic form of As(III) to the more sorptive As(V), thereby making the arsenic less bioavailable. Microbially mediated As(V) reduction can be readily observed in the near-surface hyporheic zones in freshwater streams (29), and 104 to 105 cells of As(V)-reducers g-1 have been detected in arsenic-contaminated lake sediment by most-probable-number estimations (30). Within aerobic waters and surface sediments, it is known that microbial degradation of organic matter can deplete the oxygen concentration, thereby establishing reducing conditions which are favorable for the reduction of As(V) sorbed to hydrous iron and manganese oxides (31-34). The reduction of As(V) bound to a variety of minerals potentially could lead to the mobilization of arsenic, by making the more toxic form As(III) bioavailable. Recently it was shown that the addition of acetate to aquifer material from the Bengal Delta led to a release of As(III) and Fe(II), resulting from the microbial reductive dissolution of solid-phase Fe(III) and As(V) (35). Unlike hazardous organic compounds which can be microbially mineralized to CO2, metals cannot be removed from the system. They can, however, be transformed to more or less hazardous forms as a consequence of fluctuations or changes in the environment. Given that both oxidizers and reducers have been isolated from the same environment, a microbially mediated arsenic cycle as a function of the oxidation-reduction potential in the environment is reasonable and not unlike that for nitrogen and sulfur. A recent study involving algal periphyton (Cladophora sp.) from a freshwater stream containing
