Article pubs.acs.org/est
Iron(III) Modification of Bacillus subtilis Membranes Provides Record Sorption Capacity for Arsenic and Endows Unusual Selectivity for As(V) Ting Yang,† Ming-Li Chen,† Lan-Hua Liu,† Jian-Hua Wang,†,* and Purnendu K. Dasgupta*,‡ †
Research Center for Analytical Sciences, Box 332, Northeastern University, Shenyang 110819, China Department of Chemistry and Biochemistry, University of Texas at Arlington, Arlington, Texas 76019-0065, United States
‡
S Supporting Information *
ABSTRACT: Bacillus subtilis is a spore forming bacterium that takes up both inorganic As(III) and As(V). Incubating the bacteria with Fe(III) causes iron uptake (up to ∼0.5% w/w), and some of the iron attaches to the cell membrane as hydrous ferric oxide (HFO) with additional HFO as a separate phase. Remarkably, 30% of the Bacillus subtilis cells remain viable after treatment by 8 mM Fe(III). At pH 3, upon metalation, As(III) binding capacity becomes ∼0, while that for As(V) increases more than three times, offering an unusual high selectivity for As(V) against As(III). At pH 10 both arsenic forms are sorbed, the As(V) sorption capacity of the ferrated Bacillus subtilis is at least of 11 times higher than that of the native bacteria. At pH 8 (close to pH of most natural water), the arsenic binding capacity per mole iron for the ferrated bacteria is greater than those reported for any iron containing sorbent. A sensitive arsenic speciation approach is thus developed based on the binding of inorganic arsenic species by the ferrated bacteria and its unusual high selectivity toward As(V) at low pH.
■
INTRODUCTION Arsenic pollution problems are common worldwide. South Asia (Bangladesh and the Gangetic delta in particular) has been so afflicted with natural groundwater arsenic poisoning (>600,000 with diagnosed arsenicosis, >20 million at risk) that the World Health Organization has labeled it the greatest environmental calamity in recorded history.1 The US National Academy of Engineering targeted a sustainable solution to remove arsenic from drinking water for the first Grainger Challenge Award.2 The $1 M winning entry utilized a composite iron matrix as the active element to remove the arsenic.3 The exact form in which arsenic occurs in the Gangetic delta in the aquifers is debated. However, arsenic is putatively present as/with arsenopyrite,4 biotite,5 clay minerals, iron hydroxide-coated sand grains,6 etc.; all contain iron: arsenic is always highly correlated with iron in these waters.7 Iron has a high affinity for arsenic. Iron-based sorbents, e.g., hydrous iron oxide (HFO),8−12 or HFO on matrix sorbents, e.g., activated carbon,13 fibrous ion exchangers,14 cellulose,15 etc., are commonly used to remove waterborne arsenic. When sorbents are treated with aqueous Fe(III) and extensively washed,21−23 Fe likely precipitates as HFO on the matrix.16,17 As a green and sustainable alternative to conventional metal remediation techniques, biosorption has gained an increasing role in the removal of metals, especially arsenic.18 Chitosan,19 macrofungi,20 ferns,21 algae,22 waste biomass,23, etc. have all been tried with differing degrees of success. The authors’ research groups have had independent long-standing interest in trace arsenic determination;24−32 more © 2012 American Chemical Society
recently, we jointly embarked on the characterization of biosorbents for the measurement and removal of arsenic. We found a live HeLa cell, a kind of human cervical cancer cell, could take up arsenic; both surface and intracellular accumulation are involved.32 Interestingly, while both As(III) and As(V) were taken up at high pH, at low pH, As(V) was taken up with a 40:1 selectivity over As(III). The sorbed arsenic was readily eluted by strong acid; HeLa cell-packed beds were used for arsenic analysis. As a safer alternative with respect to the HeLa cell, we explored Bacillus subtilis (B. subtilis), a spore forming grampositive rod-shaped bacterium that naturally occurs in soil and many vegetations. B. subtilis is easily grown, and its culturing is well controlled under laboratory conditions. More importantly, its pathogenic potential is generally regarded as low or absent.33 Indeed, it is presently marketed as a probiotic.34 In sorbentimmobilized form it has been studied for copper and cadmium uptake.35 We observed that native B. subtilis can remove even very low concentrations of inorganic arsenic of either oxidation states. Incubating the bacteria with iron(III) enhanced significantly the selectivity for As(V) at acid condition and in the meantime achieved record sorption capacity for arsenic at a higher pH. Received: Revised: Accepted: Published: 2251
August 12, 2011 December 31, 2011 January 19, 2012 January 19, 2012 dx.doi.org/10.1021/es204034z | Environ. Sci. Technol. 2012, 46, 2251−2256
Environmental Science & Technology
Article
nongrowth conditions. The ferrated B. subtilis (hereinafter called Fe-bac) was then harvested and washed thoroughly with DI water until neutral supernatant washing effluent was obtained, i.e., ∼pH 6. Arsenic Adsorption and Concentration by the Iron(III) Loaded Bacillus subtilis. A 100 μL Fe-bac suspension (containing ca. 2.2 mg of Fe-bac) was added into a 1.5 mL centrifuge tube. The Fe-bac was separated by centrifuging at 4000 rpm for 5 min, followed by removal of the supernatant. A 1000 μL standard arsenic solution or real water sample solution was then added, and the mixture was shaken vigorously for 3 min. Afterward, the Fe-bac were separated again from the mixture by centrifuging at 4000 rpm for 5 min, followed by adding 100 μL of HNO3 (0.8 mol L−1) for the recovery of the sorbed arsenic. After shaking for 2 min, the supernatant is separated, and the concentration of the recovered arsenic is determined by GFAAS. Samples and Sample Pretreatment. Certified reference material (CRM): 0.0100 g of CRM (human hair, GBW 09101, http://www.ncrm.org.cn/Home/Index.aspx) was taken into a PTFE digestion vessel with 5 mL of HNO3 (65%, w/v). After soaking for 2 h, the mixture was digested on a microwave oven with the following program: 150 °C/20 atm for 5 min; 180 °C/ 22 atm for 5 min. Thereafter, the mixture was heated to near dryness on a sand bath at 150 °C. The residue was dissolved with DI water and diluted to 10 mL. The As(III) was converted to As(V) by 50 μmol L−1 KMnO4.36 A sample blank was processed by following the same procedure. Spring waters collected from GuanMen Moutain (Benxi, China) and QiPan Mountain (Shenyang, China) were used for performing arsenic speciation. The water samples were filtered through a 0.22 μm membrane filter and adjusted to pH 3 before subject to the above preconcentration process. Other Assays. The viability of B. subtilis cells was tested by using MTT (3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay) viability test kit under instructions of the manufacturer.37 The iron bound on the surface of Fe-bac was stripped with 3 mol L−1 HCl, and the iron content was then determined by o-phenanthroline spectrophotometry.38
These observations entail the development of an inorganic arsenic separation/preconcentration and speciation approach. Iron(III) related sorption materials have been frequently developed for the removal of arsenic from water bodies.13−17 However, to the best of our knowledge, they have never been used for high selective preconcentration of arsenic and its speciation.
■
MATERIALS AND METHODS Instrumentation. A SP-3530GA graphite furnace atomic absorption spectrometer (GFAAS) with deuterium background correction was used for arsenic detection (Shanghai Spectrum Instruments, China). An arsenic hollow cathode lamp (Rayleigh Analytical Instrument, Beijing, China) was used as a light source at a wavelength of 193.7 nm and operated at 6.0 mA, with a 0.7 nm slit width. Pyrolytically coated graphite tubes with L’vov platforms were used for quantification. The GFAAS temperature program for the detection of arsenic was as follows: drying at 120 °C (10/10 s ramp/holding time), pyrolysis at 700 °C (10/15 s), atomization at 2500 °C (0/3 s), and cleaning at 2600 °C (0/3 s). 57 Fe Mössbauer spectra were measured using a FH-1918 Mössbauer spectrometer (Beijing Nuclear Instrument Factory, China), with a room temperature palladium matrix cobalt-57 source and was calibrated with a natural α-iron foil. The spectra were fitted with symmetric quadrupole doublets using a standard least-squares fit procedure. ATR-FT-IR spectra were obtained using a Nicolet 6700 Fourier transform infrared spectrometer (Thermo Fisher, USA) within the range of 4000−400 cm−1. Scanning electron microscopy/energy dispersive X-ray analysis was conducted with a Superscan SSX550-SEMEDX instrument (Shimadzu Scientific, Kyoto, Japan). The accelerating voltage was 15.0 kV throughout, and the spatial resolution of is 3.5 nm. UV−vis reflectance spectra were obtained with a 90 degree angle between the incident and the reflected beam. A high power Xenon light source (HPX-2000) and a CCD-array spectrometer optics (Spectra USB4000, Ocean Optics Inc., Dunedin, FL, USA) were used. Reagents. Unless specified otherwise, all chemicals used in the present experiments were analytical reagent grade. Stock solutions of As(III) and As(V) (1000 mg L−1) were prepared by dissolving appropriate amounts of anhydrous Na2AsO3 and Na2AsO4 in 1.0 mol L−1 HCl. Working standards of lower concentrations were prepared by serial dilution of the stock solutions. The pH of the sample solution was adjusted by using 0.1 mol L−1 HNO3 and/or 0.1 mol L−1 NaOH, and measured with an Orion 818 pH meter (Thermo Fisher Scientific, USA). Deionized (DI) water (18.2 MΩ·cm) was used throughout. Bacillus subtilis Cell Surface Modification by Iron(III) Binding. Gram-positive bacteria Bacillus subtilis (Wild type, obtained from Biological Department, Northeastern Univ., China, hereinafter called bac) is cultured in 100 mL of LB (Luria−Bertani) broth at 37 °C until OD600 reaches 1.2; this corresponded to 0.9 g dry biomass L−1. The B. subtilis cells were harvested and washed twice with DI water followed by centrifugation at 4000 rpm for 5 min. A Fe(NO3)3·9H2O solution was then added into the mixture to make a final Fe(III) concentration of 8 mM. The mixture was shaken by an orbital shaker for 1 h to promote Fe(III) sorption onto the B. subtilis cell surface. Since no nutrient source was added into the Fe(III) solution, the sorption of Fe(III) took place under
■
RESULTS AND DISCUSSION The binding of metal ions on a biomembrane can affect its ability to sorb desired solutes.39 Fe(III)-microbe systems have been much studied previously.40,41 In one study, HFO, two bacterial species, and each bacteria, surface modified with HFO were studied separately for their abilities to sorb Sr2+. The bacteria-HFO composites exhibited less sorption capacity for the target than the bacteria alone, for both microbes.42 As HFO itself displays high affinity for arsenic, we hoped for better results than that above. Many types of iron oxide (HFO/ferrihydrite/oxyhydroxide) have been studied for arsenic removal. The relative affinities for As(III) and As(V) of such sorbents are pH dependent. For As(III), ferrihydrite has a greater adsorption capacity than As(V), especially at pH > 7;8 highest sorption maxima of approximately 0.60 (0.58) and 0.25 (0.16) mol As/mol Fe were reported for As(III) and As(V), respectively, at pH 4.6 (9.2 in parentheses).9 For HFO and goethite (FeOOH), As(V) is favored at pH 7−8.10 Unusual Selectivity of Fe-bac for As(V) at Low pH. The iron loading initially increased rapidly with increasing [Fe(III)] in the incubating solution. More than 0.5% w/w iron loading is achieved at 8 mM Fe(III) (Figure 1A). It is surprising that up 2252
dx.doi.org/10.1021/es204034z | Environ. Sci. Technol. 2012, 46, 2251−2256
Environmental Science & Technology
Article
basic media might be due to the unfavorable surface charge of B. subtilis cells in addition to cell death. The arsenic sorption behavior by Fe-bac is much different. The high selectivity of Fe-bac for As(V) is maintained at pH < 3, down to pH 1 (at this point the cell membranes lyse, iron begins to leach). At pH >4, more As(III) is sorbed but it does not become comparable to that of As(V) until pH 10 (Figure 1B). The protonation constants for HFO are pK1 5.3 and pK2 8.8.14 At pH < 5.3, HFO is predominated as positively charged FeOH2+ which is unfavorable for the sorption of neutral H3AsO3, as As(III) exists as H3AsO3 until pH reaches its pKa (9.2). On the other hand, when pH increases to 8.8, HFO exists in the form of neutral FeOH, which can partly explain why As(III) uptake increases while that for As(V) declines slightly. Sorption isotherms were measured for both bac and Fe-bac (0.5% w/w Fe) for As(III) and As(V) at pH 3.0, 8.0, and 10.0. In all cases in the present investigation Langmuir adsorption behavior is followed well. The calculated adsorption capacities are shown in Table 1. Clearly, the presence of very small Table 1. Sorption Capacity for As(III) and As(V) as a Function of pH native B. subtilis
(▲)
and As(V) (●) by Fe-bac
Figure 1. (A) Removal of As (III) produced by incubation with different concentrations of Fe(III). Iron content in Fe-bac as a function of [Fe(III)] in the incubation solution was shown as a dashed line. Experiments were conducted at pH 3.0, with 2.2 mg Fe-bac shaking for 3 min. Initial arsenic concentration: 20 μg L−1, sample volume: 200 μL. (B) Equilibrium sorption efficiencies of As(III) (▲) and As(V) (●) by native live B. subtilis (dash line) and Fe-bac (0.5% w/w Fe) (solid line) as a function of pH at low As concentrations. Experiments were conducted at pH 1−10, with 2.2 mg Fe-bac or B. subtilis shaking for 3 min. Initial arsenic concentration: 20 μg L−1, sample volume: 200 μL.
Fe-loaded B. subtilis (0.5% w/w Fe)
pH
As(III) Qmax μmol/g biomass
As(V) Qmax μmol/g biomass
As(III) Qmax μmol/g (mol As/ mol Fe)
As(V) Qmax μmol/g (mol As/ mol Fe)
3.0
36.9 ± 2.1
22.8 ± 0.1
a
8.0
68.6 ± 2.2
31.4 ± 1.2
10.0
15.3 ± 0.7
11.9 ± 0.1
127 ± 1 (1.4 ± 0.01) 110 ± 5.4 (1.2 ± 0.06)
87.6 ± 1.0 (0.98 ± 0.01) 172 ± 4 (1.9 ± 0.04) 137 ± 3.3 (1.5 ± 0.04)
a
No detectable uptake at low levels of arsenic and short contact periods. At high As(III) levels or long contact periods, Langmuir isotherm is not followed, As(III) oxidation is likely.43
amounts of iron on the bacterium changes both the selectivity and the capacity markedly. What is remarkable is that in terms of sorption capacity for arsenic per unit iron content, the Feloaded B. subtilis outperforms any hitherto reported ironcontaining sorbent, including purpose-made ion exchangers, with sorption capacity of 0.88 mol As(III)/mol Fe and 0.78 mol As(V)/mol Fe, and iron loaded sponge, with sorption capacity of 0.24 mol As(III)/mol Fe and 1.83 mol As(V)/mol Fe.16,17 This is particularly true at pH 8 (close to the pH of most natural water), where the maximum sorption capacities for As(III) and As(V), Qmax values, are 4.5-fold and 7.9-fold of those obtained by HFO.10 Table S1 depicts this in comparison with many other iron-containing sorbents, showing the record sorption capacity for arsenic. The sorption process is fairly fast. Taking As(V) at pH 3 as an example (the sorption capacity is the smallest at this pH), even at low arsenic levels of 20 μg L−1, 96% was removed from the solution in 2 min. A further increase of the sorption time gives rise to 98% removal at 7 min (Figure S1). The Nature of As Binding by Fe-bac. Surface complexation is the most acceptable mechanism for arsenate-iron interactions.8,12 However, this is used in neutral medium, which might not be applicable for acid conditions.12 As Table 1 indicates, Fe-bac treated with 150 mg/L arsenic can sorb on a molar basis more arsenic than the iron they contain. Note that the native bacteria are capable of binding significant amounts of arsenic in either oxidation state, more
to 30% of the B subtilis cells remained viable at such a high concentration of Fe(III). This clearly suggests the viability of highly Fe-tolerant strains in successive generations. Even with as little as 0.08% Fe-loading (1 mM Fe(III) incubation medium), Fe-bac displayed an unexpected favorable affinity for As(V) over As(III), increasing with the increase of iron loading (Figure 1A). This is particularly striking in view of the fact that neither the native bac (see the [Fe(III)] = 0 data point in Figure 1A) nor HFO displayed such a pronounced high selectivity for As(V). Under the experimental conditions as detailed in Figure 1A, 1.5 mg of HFO removed either As(III) or As(V) essentially quantitatively (95 ± 2 and 97 ± 1%, respectively). At low arsenic levels as in these experiments (relevant to environmental samples), a similar sorption efficiency of As(V) was observed by Fe-bac at pH 3, while almost no As(III) is removed by highly Fe-loaded Fe-bac (initial iron concentration at 8 mmol L−1) at this low pH, providing potentials for the selective separation of As(V) and As(III). pH Dependent Arsenic Sorption Selectivity and Capacity. Figure 1B shows that live native B. subtilis is capable of removing even very low concentrations of arsenic of either oxidation states. Functional groups on B. subtilis cell surface and active uptake by living B. subtilis are responsible for the adsorption process. The decline of arsenic uptake in acidic or 2253
dx.doi.org/10.1021/es204034z | Environ. Sci. Technol. 2012, 46, 2251−2256
Environmental Science & Technology
Article
band shifts in a minor fashion from 1538 cm−1 (bac) to 1517 cm−1 (Fe-bac), suggesting possible involvement. Obviously, a new absorption band at 818 cm−1 appears to Febac-As(V). Aqueous As(V) absorbs at 878 and 907 cm−1, assigned to the stretching of As−O bonds.11 For arsenate adsorbed on ferrihydrite, the As−O stretching in As−O−Fe linkage appears at ∼833 cm−1.46 The As−O stretching vibration is sensitive to its chemical environment, and it is reasonable to conclude that this is the same As−O stretching in As−O−Fe linkage that may be present in the ferrated B. subtilis Fe-bacAs(V). This would logically result from the replacement of Febound −OH groups with −OAsO(OH)2 moieties. Room temperature 57Fe Mössbauer spectroscopy showed that both Fe-bac and Fe-bac-As(V) exhibit paramagnetic doublet states. An increase in the s-electron density around the Fe atom will cause the isomer shift (IS) to decrease.47,48 With decreasing ligand electronegativity, the d-electrons shield less, and thus the s-electron density around Fe increases. Replacing a Fe-bound −OH group with an −O−As moiety (H atom is more electronegative than As atom) should decrease IS. This was indeed observed: IS decreased from 0.3854 ± 0.001 to 0.3715 ± 0.001 mm/s (relative to α-Fe) upon As exposure (Figure S6). The quadrupole split also slightly decreased from Fe-bac to Fe-bac-As(V) (see details in the Supporting Information) consistent with a greater electron density symmetry around Fe atom after arsenic sorption, in accordance with an Fe−O− As bond. 57 Fe Mössbauer spectroscopy easily identifies the iron oxidation states. Active reduction of Fe(III) to Fe(II) has been observed at B. subtilis cell surface upon 24 h of aging.41 In the present case, the short incubation periods in our study did not create detectable Fe(II). SEM studies (Figure 2B) show a clear change in the nature of the B. subtilis cell surface from bac to Fe-bac with additional amorphous nanomaterial that is likely HFO being separately present. These micrographs echo the vision that sorption of Fe(III) and the precipitation of HFO on the cell membrane is a continuum.40 Upon arsenic exposure, the biomass always appear cemented together, possibly by −Fe−O−As−O−Fe− bridges between cell-bound Fe and amorphous HFO. Visibly, the slight color of the Fe-bac becomes paler on arsenic exposure; this is more readily seen in the reflectance spectra, the same loss of color is seen when arsenate is added to Fe(NO3)3 (see Figures S7 and S8). Fe-bac for Arsenic Preconcentration and Speciation. The unusual selectivity toward As(V) by Fe-bac at low pH is readily feasible for the development of an inorganic arsenic speciation approach. In practice, As(V) is selectively retained by Fe-bac, which can then be recovered and quantified by GFAAS. After converting As(III) to As(V), the total amount of arsenic is derived similarly. The As(III) content is obtained by difference. The adsorption kinetics have shown that As(V) sorption is a fast process, and 3 min is enough for obtaining a reasonable sorption efficiency. The retained As(V) can be recovered by nitric acid, NaOH, and complexing reagent (EDTA). Considering that NaOH and EDTA cause serious background interferences when using GFAAS for detection, nitric acid is most suitable for collecting the retained As(V), and 0.8 mol L−1 HNO3 gives rise to a quantitative recovery of the As(V). The Fe-bac amount used not only affects the adsorption efficiency of As(V) but also closely relates to its ensuing recovery. Insufficient Fe-bac results in low sorption efficiency, while excessive Fe-bac requires a large eluent volume for
than that could be accounted for surface binding alone. Note also the pH dependence of As(III) uptake (Table 1) strongly suggests that the uptake neither works by surface ion exchange nor by transport of neutral As(OH)3 into the B. subtilis cell. Thus, some other mechanisms must be operative. In the present case, arsenic sorption involves two successive processes, i.e., Langmuir adsorption followed by transferring the sorbed arsenic inside the cell via metabolism activity of the live B. subtilis cells.44,45 That is, part of the arsenic was shuttled inside the cell after sorption, by slower transport of the surfacecaptured arsenic into the cell interior through cell membranes or special transport channels, and in the cell nucleus the metal species can potentially interact with proteins/enzymes or DNA. The presence of iron on the cell surface apparently promotes this path. We conducted microscopic (scanning electron microscopy coupled with energy dispersive X-ray, SEM-EDX) analysis of highly arsenic-loaded Fe-bac (see the Table of Content Graphic). It is not sufficiently sensitive to create elemental maps with a resolution that can determine the arsenic distribution inside versus outside individual cells. However, we were able to map the relative iron and arsenic X-ray intensities in randomly chosen 30 × 40 μm viewing frames (each frame containing more than 100 bacteria). The arsenic and iron signals showed no correlation whatsoever (r2 < 0.003, n = 15, Figure S5). We also examined the bac, Fe-bac, and Fe-bac after As(V) sorption at pH 3 by Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) spectroscopy. Figure 2A
Figure 2. (A) ATR-FTIR spectra. See Table S2 for assignments. (B) Scanning electron micrograph (scale bar is 500 nm): (a) native B. subtilis, (b) Fe-bac, (c) Fe-bac-As(V), note general adhesion in (c). For (b), the Fe(III) incubation concentration is 8 mmol L−1. For (c), the Fe-bac is pretreated with 200 mg L−1 As(V) solution (pH 3.0) under vigorous agitation for 60 min to facilitate adsorption.
shows the spectra of the native B. subtilis, Fe-bac, and Fe-bacAs(V), and Table S2 lists the band assignments. There were no shifts in the bands attributable to amide CO stretching, CH2 scissoring, and phosphate PO stretching from bac to Fe-bac. A broad band is present at ∼1050 cm−1 that cannot be unequivocally assigned. The 1398 cm−1 absorption band due to the symmetric stretching of COO−, decreases, however, after Fe-loading, and carboxylate groups are likely involved in binding Fe. In addition, amide II (N−H stretching) absorption 2254
dx.doi.org/10.1021/es204034z | Environ. Sci. Technol. 2012, 46, 2251−2256
Environmental Science & Technology
Article
the reflectance spectra of Fe-bac and Fe-bac-As(V). This material is available free of charge via the Internet at http:// pubs.acs.org.
recovery the sorbed As(V), thus reducing the enrichment factor. A compromised amount of 2.2 mg Fe-bac is appropriate. In such a case, an elution time of
