Environ. Sci. Technol. 2008, 42, 6730–6735
Downloaded via UNIV OF LOUISIANA AT LAFAYETTE on January 22, 2019 at 14:29:42 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Relation between the Redox State of Iron-Based Nanoparticles and Their Cytotoxicity toward Escherichia coli ´ L A N I E A U F F A N , * ,†,§ ME ´ R Ô M E R O S E , †,§ W A F A A C H O U A K , ‡,§ J E M A R I E - A N N E R O N C A T O , ‡,§ ´ AC,| DAVID T. WAITE,⊥ CORINNE CHANE A R M A N D M A S I O N , †,§ JOSEPH C. WOICIK,# MARK R. WIESNER,∇ A N D J E A N - Y V E S B O T T E R O †,§ ´ CEREGE UMR 6635 CNRS/Aix-Marseille Universite, Europoˆle de l’Arbois, 13545 Aix-en-Provence, France, LEMIRe ´ ´ UMR 6191 CNRS/CEA-Universite´ de la Mediterran ee, 13108 Saint-Paul-lez-Durance, France, ECOREV, FR n°3098, Europoˆle de l’Arbois, 13545 Aix-en-Provence, France, LCMC de Paris UMR 7574 CNRS/UPMC, 75252 Paris, France, School of Civil & Environmental Engineering, UNSW Sydney, NSW-2052, Australia, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, and Civil and Environmental Engineering, Duke University, Durham, North Carolina 27708
Received January 9, 2008. Revised manuscript received April 17, 2008. Accepted April 23, 2008.
Iron-based nanoparticles have been proposed for an increasing number of biomedical or environmental applications although in vitro toxicity has been observed. The aim of this study was to understand the relationship between the redox state of ironbased nanoparticles and their cytotoxicity toward a Gram-negative bacterium, Escherichia coli. While chemically stable nanoparticles(γFe2O3)havenoapparentcytotoxicity,nanoparticles containing ferrous and, particularly, zerovalent iron are cytotoxic. The cytotoxic effects appear to be associated principally with an oxidative stress as demonstrated using a mutant strain of E. coli completely devoid of superoxide dismutase activity. This stress can result from the generation of reactive oxygen species with the interplay of oxygen with reduced iron species (FeII and/or Fe0) or from the disturbance of the electronic and/or ionic transport chains due to the strong affinity of the nanoparticles for the cell membrane.
Introduction Accompanying the rapid emergence of nanotechnology and the increasing use of nanoparticles, a new discipline has emerged, that of the nanotoxicology (1). Indeed, as seen in public concerns around use of genetically modified organisms, our society now demands detailed information on * Corresponding author fax: 0033 442 971 559; e-mail: auffan@ cerege.fr. † CEREGE UMR 6635 CNRS/Aix-Marseille Universite´. § ECOREV. ‡ LEMIRe UMR 6191 CNRS/CEA-Universite´ de la Me´diterrane´e. | LCMC de Paris UMR 7574 CNRS/UPMC. ⊥ UNSW Sydney. # National Institute of Standards and Technology. ∇ Duke University. 6730
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
potential health and environmental effects of most new technologies. Only a few studies have been initiated on the potential toxicity of nanoparticles and it is not clear whether the size of the nanoparticles or their particular surface chemistry is the key determinant of toxicity. Nevertheless, some useful insights into aspects of nanotoxicology have begun to emerge. First, some studies have shown that, once dispersed and surface-passivated, nanoparticles can be internalized by cells without inducing any cytotoxic effect (2, 3). Consequently, the fact that nanoparticles are of similar size to biological compounds does not appear to be a key element of the toxicity of these particles. Rather, it appears that the chemical nature and the physicochemical reactions occurring at the nanoparticle/cell interface are the key to the biological effects (4, 5). In particular, nanoparticles can affect biological targets through generation of reactive oxygen species (ROS) thereby inducing an oxidative stress (6). To date, the causal mechanisms linking the physicochemical properties of nanoparticles with their biological effects are not well defined. Further investigations of these mechanisms must be the primary focus of future nanotoxicology studies. It is known that mineral surfaces may promote the formation of ROS via different pathways: dissolution and release of metals, presence of structural defects, activation of the immune systems generating secondary cellular ROS, and catalysis of the reduction of O2 via electron transfer (redox) reactions (7). In this latter case, the mineral surface (particularly of minerals containing transition metals such as iron) plays an active role. For instance, the reaction between FeII and O2 is much faster when FeII ions are adsorbed at the surface of a mineral than when FeII ions are dissolved (8). Because of the large specific surface area (SSA) of nanoparticles, the rate of surface-mediated reactions such as this will be enhanced. Consequently, it is necessary to determine the underlying mechanisms and key factors responsible for the oxidative stress induced by nanoparticles in order to properly ameliorate risk studies (9). The present work is focused on three iron-based nanoparticles characterized by different redox states: maghemite (γFe2IIIO3), magnetite (Fe3II/IIIO4), and zerovalent iron nanoparticles. These nanoparticles are used for a variety of applications, particularly in the biomedical and environmental fields. For example, Fe° nanoparticles are used in soil and groundwater decontamination (10) while the superparamagnetic and adsorbing properties of γFe2O3 and Fe3O4 nanoparticles permit their use in biomedicine and water treatment (11, 12). Following their development, preliminary investigations of their potential toxicity have been undertaken (3, 13, 14). While the results are rather inconsistent, in vitro experiments suggest that iron-based nanoparticles can be toxic. Intercomparison between the various studies that have been undertaken is problematic because of differences in the experimental protocols used including purity of particles, cellular strain, and conditions under which the nanoparticles are presented to cells. In this paper, we compare the toxic effects associated with Fe3O4, γFe2O3, and Fe° nanoparticles toward the Gram-negative bacterium Escherichia coli and its double mutant sodA sodB which is completely devoid of superoxide dismutase activity. The objectives were (i) to determine the concentrations at which the nanoparticles suspensions would be toxic as a function of the Fe redox state, and (ii) to characterize (from the macroscopic to the atomic scale) the physicochemical mechanisms responsible for this toxicity. 10.1021/es800086f CCC: $40.75
2008 American Chemical Society
Published on Web 07/23/2008
Experimental Section Nanoparticles (NPs). Magnetite nanoparticles (nMagnetite) were synthesized by coprecipitation of ferric and ferrous ion according to ref 15. Maghemite nanoparticles (nMaghemite) were obtained via oxidation of magnetite NPs under acidic conditions. Iron atoms are distributed into the octahedral (Oh) and tetrahedral (Td) sites according to [Fe3+]Td[Fe3+Fe2+]OhO4 and [Fe3+]Td[Fe5/33+V1/3]OhO4 in nMagnetite and nMaghemite, respectively (15). These particles are well crystallized and roughly spherical with a mean coherent diameter of 6 ( 1 nm and an SSA of 172 m2/g (16) (see Supporting Information). The zerovalent iron NPs (nZVI) were prepared by adding NaBH4 to FeCl3 · 6H2O solution as described by ref 17. nZVI were washed three times with 10-4 M HCl. They have a SSA of 32 m2/g (18) and an average primary particle size of 50 nm. Size Measurement. The average hydrodynamic sizes as a function of pH were determined by dynamic light scattering using a Nanotrac (Microtrac, North Largo, FL). Particle size was determined by suspending nanoparticles in 0.01 M NaCl solution at 25 °C and the pH adjusted using solutions of 0.1 M HCl and 0.1 M NaOH. Toxicity Assessment. Escherichia coli strain Qc1301 and its double mutant sodA sodB Qc2472 were used in this study (19). Preculture of the two strains were performed in Luria-Bertani (LB) medium containing 5 g/L of yeast extract, 10 g/L of bactotryptone and 5 g/L of NaCl and appropriate antibiotics. Cultures were inoculated from precultures and grown at 37 °C for 16 h to an OD630 of 1. For the survival assays, 1 mL of bacterial culture (109 bacteria) was washed, and different concentrations of bacteria (from 1.25 × 103 to 5 × 107 bacteria/mL) were incubated with from 7 to 700 mg/L of NPs solutions diluted in ultrapure water at pH 5-5.5. Two controls of NPs-free bacteria were prepared by incubating both strains for 1 h in ultrapure water at physiological pH 7.4 and at pH 5. After 1 h incubation at 30 °C, serial dilutions of the bacterial suspensions were plated on solid LB-agar plates and incubated overnight at 37 °C, and colonies were counted. All treatments were prepared in triplicate and repeated two times. The Student’s t-test was used for significance testing. Transmission Electron Microscopy (TEM). The two strains of E. coli were incubated with 100 mg/L of NPs for 1 h, and bacteria were then washed three times in ultrapure water, fixed with 4% glutaraldehyde for 24 h at 4 °C, and postfixed with osmium at 25 °C. Bacteria were embedded in Epon resin. Ultrathin sections of 50 nm were examined with a JEOL/JEM-1220 instrument. X-ray Diffraction (XRD). XRD patterns of NPs before and after contact with bacteria were recorded with a PANanalytical X’Pert PRO diffractometer with Co KR radiation (1.79Å) at 40 kV and 40 mA. A counting time of 25s per 0.05° step was used for the 2θ range 15-90°. X-ray Absorption Spectroscopy (XAS). XAS spectra were recorded at room temperature at beamline X23A2 at the National Synchrotron Light Source (NSLS, New York, NY). Spectra were acquired in transmission mode using a Si(311) monochromator from 100 to 800 eV above the Fe K-edge (7.112 KeV). The spectra are the sum of three to five scans, and the energy was calibrated using an iron foil. XANES (Xray absorption near edge structure) spectra were obtained after performing standard procedures for pre-edge subtraction and normalization using the IFEFFIT software package (20). Spectra were fitted using least-squares optimization of linear combination fits (LCF) of model compounds (Fe°, nMagnetite, nMaghemite, and lepidocrocite). A fit was retained when the residual (r) decreased by at least 20% relative to the previous simulation. The precision of the LCF method was estimated at (10%.
FIGURE 1. Influence of the pH on the colloidal stability of nMaghemite, nMagnetite, and nZVI (0.01 mol/L NaCl, 25 °C, incubation time: 1 h). Inset pictures: transmission electron micrograph of nMaghemite, nMagnetite, and nZVI before contact with bacteria. Black dash: 90 nm.
Results Toxicity Assessment. In order to maximize the interactions between the NPs and bacteria, it was important to welldisperse iron-based NPs before used. Since the point of zero net charge (PZC) of iron oxide and zerovalent iron particles are close to physiological pH (∼7.4), particles typically strongly aggregate in biological medium (3). An indication of the colloidal stability of the three NPs as a function of suspension pH is given in Figure 1. For pH lower than 5.4-5.6, the nMaghemite and nZVI suspensions are stable with a mean hydrodynamic diameter (Dh) of 50 ( 5 nm and 320 ( 30 nm, respectively. The suspension of nMagnetite is stable until pH 4.2 with slight aggregation occurring for pHs between 4.2 and 5.7. In this range the mean Dh of nMagnetite is 380 ( 40 nm. For pHs above these values, the NPs are strongly destabilized and the mean Dh of the aggregated assemblages reach 3 to 5 µm. Hence, to ensure that the NPs are fully dispersed during the cytotoxic assays, it was necessary to undertake studies with pH between 5 and 5.5. This pH was considered close enough to physiological pH such that minimal disturbance to bacterial metabolism was expected. The biological effects of well-dispersed iron-based NPs toward the two strains of E. coli were investigated. First, the potential toxicity due to the NP-free bacterial incubation at pH 5-5.5 during 1 h was checked (see Supporting Information). The cytotoxicity assay yielded, respectively, 85% ( 11% and 100% ( 20% of survival with respect to Qc1301 and Qc2472 incubated under physiological conditions (pH 7.4). The percentage of survival given in the rest of the study is calculated from the number of bacterial colonies for each NP concentration normalized to the number of bacterial colonies counted for the NP-free control incubated at pH 5.5. The results of the survival tests for the two E. coli genotypes in the presence of the various NPs are shown in Figure 2. In the case of the wild type, the toxicity increased with dose for nMagnetite and nZVI whereas nMaghemite gave no evidence of toxicity under our experimental conditions. Cytotoxic effects appear for nMagnetite and nZVI for doses upper than 700 mg/L and 70 mg/L, respectively. At the same concentrations, nMagnetite and nMaghemite showed a higher toxicity toward the double mutant sodA sodB that is completely devoid of the antioxidant enzymes iron and manganese superoxide dismuatse (Figure 2). Whereas the wild type strain can withstand up to 350 mg/L of nMagnetite and 700 mg/L of nMaghemite, the double mutant sodA sodB VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6731
FIGURE 2. Number of colonies of the strain QC1301 and the double mutant sodA sodB Qc2472 formed on the agar plate as a function of the concentration of NPs (mg/L). Statistical significance: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
FIGURE 4. XRD patterns of nZVI before and after 1 h of incubation in water at pH 5-5.5.
FIGURE 3. TEM pictures of E. coli ultrathin sections. Control bacteria (a) incubated in water at pH 5 during 1 h. E. coli observed after 1 h of incubation with 100 mg/L of nMaghemite (b), nMagnetite (c), and nZVI (d). showed a significant decrease of bacterial survival in the presence of 175 mg/L of nMagnetite and 700 mg/L of nMaghemite. To be sure that this toxicity is related to the NPs themselves and not to the presence of residue from the synthesis, the toxicity of the supernatant was evaluated (see Supporting Information). NPs were centrifuged and survival tests were done on the liquid phase. No toxicity of the supernatant of the NPs suspension was observed, confirming that the toxic effect was due to iron-based NPs. Nanoparticle Localization. TEM analysis allows direct visualization of morphological changes of E. coli due to the presence NPs and/or changes in the shape of NPs after incubation with bacteria. E. coli is seen to be a moderately sized (1.5 × 0.5 µm) Gram-negative bacillus with a tubular form (Figure 3). The iron-based NPs were clearly visible and distinct from the cellular matrix because of the high electron density of iron. The bacterial cell wall clearly displays an outer shell of high electronic density corresponding to a layer of NPs adsorbed on the cell surface. No evidence of cellular 6732
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
internalization of NPs has been found. Another interesting result is the modification of the shape of nZVI after incubation with bacteria. While the primary nZVI particles are pseudospherical before contact with E. coli, they are transformed into lamella structure after 1 h of incubation. No modification of the shape of nMaghemite or nMagnetite is revealed by TEM analysis. Structural Modifications of Iron-Based NPs. An X-raybased study was performed to follow the structural evolution of the iron-based nanoparticles during incubation with E. coli. Whereas the nZVI are only composed by Fe0 in their mother solution as demonstrated by XRD (Figure 4), they very quickly transform to oxy-hydroxides when they are placed in solution. This high sensitivity toward oxidation of nZVI in aerobic conditions is a well-documented spontaneous electrochemical process (21, 22). Analysis by XRD of nZVI after incubation for 1 h in an abiotic solution at pH 5–5.5 indicates the appearance of two oxy-hydroxide mineral phases. The strong XRD peaks at 16.5°, 31.6°, 41.6°, and 74.7° 2θ are in agreement with the crystallographic d-spacing of reference standard lepidocrocite (γFeOOH) and a spinel phase of magnetite and/or maghemite (Figure 4). Since the two spinel phases are isostructural, they are indistinguishable from XRD patterns. XANES at the Fe-K-edge is highly sensitive to the redox and coordination states of iron atoms.
FIGURE 5. (A) Fe-K-edge XANES spectra of nMagnetite, nMaghemite, and nZVI after 1 h incubation in water and with E. coli. The dotted line is the best fit line to the sample. The results and the quality (χ2) of the fits are summarized in Table 1. (B) Fe-K-edge XANES spectra of reference compounds. The inset corresponds to the pre-edge area. (C) Evolution of the position and the area of the pre-edge for nMagnetite and nMaghemite before and after incubation in water and with E. coli. The four poles Fe6(III), Fe6(II), Fe4(III), Fe4(II) are given by ref 24. Hence, the presence of FeII in Fe3O4 and only FeIII in γFe2O3 can be discriminated. The differences between the XANES spectra of γFe2O3, Fe3O4, and γFeOOH reference compounds with respect to the intensity and the position of the pre-edge and the main edge and the position of the absorption ramp are well understood (23, 24). Linear combination fits of reference spectra (Figure 5, Table 1) are in close agreement with experimental data. A contribution of 60% of γFeOOH and 38% of Fe3O4 was necessary to fit the spectra of nZVI in the abiotic system. The ratio γFeOOH/Fe3O4 decreased in the presence of E. coli Qc1301 and Qc2472 for which 21-25% of γFeOOH and 78-82% of Fe3O4 were necessary to satisfactorily reproduce the XANES experimental data (Figure 5, Table 1). No significant structural changes were observed as a function of the bacterial strain present. These strong structural modifications of nZVI that occur under abiotic and biotic aerobic conditions were confirmed by the rapid change in shape of the nZVI particles as observed by TEM with the globular nZVI particles adopting a lath-like appearance typical of lepidocrocite, γFeOOH. In the case of nMagnetite, XANES data also present differences before and after incubation with bacteria. To adequately fit the XANES spectra of nMagnetite after exposure to water for 1 h at pH 5, a contribution of 77% of Fe3O4 and 22% of γFe2O3 was necessary (Figure 5A, Table 1). When nMagnetite particles were placed in contact with both genotypes of E. coli, the ratio Fe3O4/γFe2O3 increased with a contribution of 50-53% of Fe3O4 and 46-48% of γFe2O3 (Figure 5A, Table 1). Given that 30-40% of the iron atoms present are localized at the surface of the NPs, it is likely that the entire surface and subsurface region of the nMagnetite is fully oxidized. These results are also confirmed by the evolution of the pre-edge feature (Figure 5C). The position and the intensity of the prepeak strongly depend on the geometry and redox states of Fe (inset in Figure 5B). When
TABLE 1. Mineralogical Composition of the Iron-Based Nanoparticles before and after 1 h of Contact with Escherichia coli Determined by XANES and XRD samples stock solution in water with Qc1301 with Qc2472 stock solution in water with Qc1301 with Qc2472 stock solution in water with Qc1301 with Qc2472
mineralogical composition nMaghemite 100% γFe2O3 98% γFe2O3 95% γFe2O3 96% γFe2O3 nMagnetite 100% Fe3O4 77% Fe3O4 + 22% γFe2O3 53% Fe3O4 + 46% γFe2O3 50% Fe3O4 + 48% γFe2O3 nZVI 100% Fe0 38% Fe3O4 + 60% γFeOOH 82% Fe3O4 + 25% γFeOOH 78% Fe3O4 + 21% γFeOOH
% of error
residue χ2
(10%
0.016 0.021 0.020 0.020
(10%
0.017 0.025
0.025 (10%
0.032 0.020
nMagnetite particles are placed in water and in contact with E. coli, the pre-edge features move to a fully Fe-oxidized pole close to that of γFe2O3. In comparison, nMaghemite appeared highly stable after direct contact with bacteria (Figure 5, Table 1). No contribution of mineral phases other than maghemite was necessary to reproduce the XANES experimental data, and no structural modifications were observed. VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6733
Discussion XAS results combined with XRD revealed that iron-based NPs react in differing ways on incubation with E. coli. The extent of structural modifications increase in the order nMaghemite < nMagnetite < nZVI. The most stable NPs examined here are the nMaghemite. They are composed of fully oxidized crystals and consequently are highly stable under aerobic conditions even in the presence of bacteria. In comparison, the surface of nMagnetite is unstable against oxidation because of the high mobility of electrons within the structure and the diffusion of FeII from the solid into the solution. This leads finally to a core/shell structure with a core of magnetite and a stable structure at the surface close to that of maghemite. Molecular modeling suggests that electron transfers are facilitated in magnetite by the large overlap of the FeIII-FeII octahedral d-orbitals found in the edge-sharing environment (25). The presence of E. coli enhances this structural transformation and, after 1 h, 50% of nMagnetite particles (corresponding to all the surface and subsurface atoms) are oxidized compared to only 20% in water. nZVI particles are the most sensitive to oxidation under both abiotic and biotic aerobic conditions. Lepidocrocite (γFeIIIOOH) and magnetite particles are the end products of this oxidation process. Similar corrosion products of ZVI powder have been noted in previous studies (18). These oxyhydroxides can only be obtained at room temperature and neutral pH through an intermediate oxidation of Fe0 via a precursor phase named green rust (FeII4FeIII2(OH) · 12X, nH2O) (15, 26–28). This hydroxide is formed via a dissolution and recrystallization process. The oxidation of green rust leads to γFeOOH and Fe3O4 crystalline phases which are formed via two pathways (15, 27, 28): (i) γFeOOH particles are obtained if the oxidation of FeII in the green rust is fast enough and the ratio O2/FeII ratio is high, and (ii) Fe3O4 particles are obtained when the ratio O2/FeII decreases and the oxidation of FeII is slower. This reconstructive transformation can explain the change of morphology observed in this study for the nZVI before (pseudospherical) and after (pseudolamellar) 1 h-contact with E. coli (Figures 1 and 3). Moreover, we have shown that, in the abiotic suspension, γFeOOH particles are the main product of nZVI oxidation whereas Fe3O4 particles predominate in the presence of bacteria. Two hypotheses can explain the change in the Fe3O4/γFeOOH ratio. First, bacterial respiration may well consume a portion of the oxygen present thereby decreasing the O2/FeII ratio, favoring the crystallization of Fe3O4 particles. Second, the range of pH suited for γFeOOH crystallization is known to be very narrow (between 5 and 7) (27). It is possible that an increase of the pH occurs in the presence of E. coli either because of close contact between nanoparticles and the bacterial membrane (pH ∼7) or because of the initial corrosion process (18). Under our experimental conditions, the zeta potential is on the order of 10-20 mV for the three iron-based NPs (26), the affinity of the three NPs for the bacterial wall is elevated (Figure 3), and the aggregation state is similar (Figure 1). Hence, the difference in the toxicity response cannot be related to these parameters. However, it is known that Fe2+ ions exhibit a stronger toxicity than Fe3+ ions toward E. coli. The concentration of Fe2+ necessary to inhibit the reduction of cytochrome c is 10 times larger than the concentration of Fe3+ (29). Nishimo et al. have found that Gram-negative bacteria resist to 3.2 mmol/L of Fe3+. Hence, the extent of structural modifications observed for the nanoparticulate materials examined and the redox changes between Fe0/ FeII/FeIII can be directly related to the increase of cytotoxicity observed with nMaghemite < nMagnetite < nZVI. nZVI were entirely oxidized and exhibited the highest toxicity. It is worth noting that iron oxidation and phase 6734
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
transformation mechanisms of nZVI and nMagnetite involve ionic and/or electronic transfers at the NPs surface. Since the contact between NPs and the cell membrane is strong (Figure 3), such contact and chemical transformations can disturb the bacterial metabolism. Several hypotheses can be developed concerning the mode of toxicity such as the disruption of the membrane integrity, the disturbance of the electron and ionic transport chains between the intra- and extracellular media, or the induction of an oxidative stress toward E. coli by generating reactive oxygen species (ROS). The hypothesis of an oxidative stress is directly supported by the greater sensitivity of the SOD-deficient strain of E. coli. SOD is the first line of defense against oxidative stress. In the absence of SOD, the reaction between oxygen free radicals and biological structures are not inhibited (30) with a resultant stronger cytotoxic effect. Reduced-iron oxides are known to be efficient ROS producer. ROS can be directly generated on exposure of nMagnetite to oxic solutions as a result of oxidation of either structural FeII at the Fe3O4 surface or FeII released to solution (7) and also during the corrosion of nZVI (eqs 1 and 3) via the Fenton reaction (18). Moreover, nZVI are able to catalyze this reaction via eq 2) and the regeneration of FeII. Fe0 + O2 + 2H+ f Fe2+ + H2O2
(1)
Fe0 + H2O2 f Fe2+ + 2OHFe
2+
+ H2O2 f Fe
3+
•
+ OH + OH
(2) -
(3)
Although the nMaghemite are the more physicochemically stable and not toxic toward the wild strain of E. coli, they were toxic toward the mutant devoid of SOD, indicating a slight oxidative stress effect but only for very large amount of NPs. Hence, we have shown that exposure to increasing concentrations of iron-based NPs containing either FeII or Fe0 results in a dose-dependent decrease in survival of E. coli. The main source of toxicity is an oxidative stress induced by oxygenation of reduced Fe species (FeII and/or Fe0) either within the NPs lattice or on release of FeII to solution. These results highlight the importance of further research into the mechanisms accounting for the generation of ROS and into factors influencing the extent of ROS production. Other toxic mechanisms may be responsible for the effects observed including the disruption of the membrane integrity or the disturbance of the chains of transport of electron and ions after the strong adsorption of iron-based NPs onto the bacteria wall and warrant further study before definitive assignment of the major mechanism of toxicity can be made. Irrespective of the precise mode of attack, this high toxicity observed with nZVI highlights the particular care that must be taken to fully understand potential ecotoxicological impacts of these materials given their increasing use in environmental and manufacturing applications. Indeed, these results raise many questions concerning the toxicity of nanometric crystallites with a low oxidation degree and consequently a high reactivity at the contact with living organisms.
Supporting Information Available Survival tests and nanoparticle characterization. This information is available free of charge via the Internet at http:// pubs.acs.org..
Acknowledgments This program has been funded by the French National Program ACI-FNS (ANR) “ECCO” supported by INSU (French National Institute for Earth Sciences and Astronomy). The XAS work was performed at the NSLS synchrotron (USA). We are grateful to the machine and beamline groups who have
made these experiments possible. We also thank Joe¨l Courageot for the TEM experiments and Sam Dukan for the E. coli strains.
Literature Cited (1) Oberdo¨rster, G.; Oberdo¨rster, E.; Oberdo¨rster, J. Nanotoxicology: An Emerging Discipline Evolving from Studies of Ultrafine Particles. Environ. Health Perspect. 2005, 113 (7), 823–839. (2) Auffan, M.; Decome, L.; Rose, J.; Orsiere, T.; DeMeo, M.; Briois, V.; Chaneac, C.; Olivi, L.; Berge-Lefranc, J. L.; Botta, A.; Wiesner, M. R.; Bottero, J. Y. In Vitro Interactions between DMSA-Coated Maghemite Nanoparticles and Human Fibroblasts: A Physicochemical and Cyto-Genotoxical Study. Environ. Sci. Technol. 2006, 40 (14), 4367–4373. (3) Limbach, L. K.; Li, Y.; Grass, R. N.; Brunner, T. J.; Hintermann, M. A.; Muller, M.; Gunther, D.; Stark, W. J. Oxide Nanoparticle Uptake in Human Lung Fibroblasts: Effects of Particle Size, Agglomeration, and Diffusion at Low Concentrations. Environ. Sci. Technol. 2005, 39 (23), 9370. (4) Adams, L. K.; Lyon, D. Y.; Alvarez, P. J. J. Comparative ecotoxicity of nanoscale TiO2, SiO2, and ZnO water suspensions. Water Res. 2006, 40 (19), 3527. (5) Thill, A.; Zeyons, O.; Spalla, O.; Chauvat, F.; Rose, J.; Auffan, M.; Flank, A. M. Cytotoxicity of CeO2 Nanoparticles for Escherichia coli. Physico-Chemical Insight of the Cytotoxicity Mechanism. Environ. Sci. Technol. 2006, 40 (19), 6151–6156. (6) Xia, T.; Kovochich, M.; Brant, J.; Hotze, M.; Sempf, J.; Oberley, T.; Sioutas, C.; Yeh, J. I.; Wiesner, M. R.; Nel, A. E. Comparison of the Abilities of Ambient and Manufactured Nanoparticles To Induce Cellular Toxicity According to an Oxidative Stress Paradigm. Nano Lett. 2006, 6 (8), 1794–1807. (7) Schoonen, M. A. A.; Cohn, C. A.; Roemer, E.; Laffers, R.; Simon, S. R.; O’Riordan, T. Mineral-induced formation of reactive oxygen species. In Medical Mineralogy and Geochemistry, Reviews in mineralogy and geochemistry; Sahai, N., Schoonen, M. A. A., Eds.; Mineralogical Society of America and the Geochemical Society: Washington, DC, 2006; Vol. 64, p 349. (8) Wehrli, B.; Sulzberger, B.; Stumm, W. Redox processes catalyzed by hydrous oxide surfaces. Chem. Geol. 1989, 78 (3), 167–179. (9) Robichaud, C. O.; Tanzil, D.; Weilenmann, U.; Wiesner, M. R. Relative Risk Analysis of Several Manufactured Nanomaterials: An Insurance Industry Context. Environ. Sci. Technol. 2005, 39 (22), 8985–8994. (10) Zhang, W. X. Nanoscale Iron Particles for Environmental Remediation: An Overview. J. Nanopart. Res. 2003, V5 (3), 323. (11) Gupta, A. K.; Gupta, M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 2005, 26 (18), 3995. (12) Wiesner, M. R.; Bottero, J. Y. Environmental nanotechnology: Applications and impacts of nanomaterials; MacGraw Hill: New York, 2007; p 540. (13) Gupta, A. K.; Gupta, M. Cytotoxicity suppression and cellular uptake enhancement of surface modified magnetic nanopar-
ticles. Biomaterials 2005, 26 (13), 1565. (14) Hussain, S. M.; Hess, K. L.; Gearhart, J. M.; Geiss, K. T.; Schlager, J. J. In vitro toxicity of nanoparticles in BRL 3A rat liver cells. Toxicol. in Vitro 2005, 19 (7), 975. (15) Jolivet, J. P.; Chaneac, C.; Tronc, E. Iron oxide chemistry. From molecular clusters to extended solid networks. Chem. Commun. 2004, (5), 481–487. (16) Jolivet, J.-P.; Tronc, E.; Chaneac, C. Synthesis of iron oxidebased magnetic nanomaterials and composites. C. R. Chim. 2002, 5 (10), 659–664. (17) Wang, C.-b.; Zhang, W.-x. Synthesing nanoscale iron particles for rapid and complete dechlorination of TOE and POBs. Environ. Sci. Technol. 1997, 31 (7), (18) Joo, S. H.; Feitz, A. J.; Sedlak, D. L.; Waite, T. D. Quantification of the Oxidizing Capacity of Nanoparticulate Zero-Valent Iron. Environ. Sci. Technol. 2005, 39 (5), 1263–1268. (19) Carlioz, A.; Touati, D. Isolation of superoxide dismutase mutants in Escherichia coli: is superoxide dismutase necessary for aerobic life. EMBO J. 1986, 5 (3), 623–630. (20) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537–541. (21) Kanel, S. R.; Greneche, J. M.; Choi, H. Arsenic(V) Removal from Groundwater Using Nano Scale Zero-Valent Iron as a Colloidal Reactive Barrier Material. Environ. Sci. Technol. 2006, 40 (6), 2045–2050. (22) Zhang, T. C.; Huang, Y. H. Profiling iron corrosion coating on iron grains in a zerovalent iron system under the influence of dissolved oxygen. Water Res. 2006, 40, 2311–2320. (23) Manceau, A.; Gates, W. P. Surface structural model for ferrihydrite. Clays Clay Miner. 1997, 45 (3), 448–460. (24) Wilke, M.; Partzsch, G. M.; Bernhardt, R.; Lattard, D. Determination of the iron oxidation state in basaltic glasses using XANES at the K-edge. Chem. Geol. 2005, 220 (1-2), 143. (25) Williams, A. G. B.; Scherer, M. M. Spectroscopic Evidence for Fe(II)-Fe(III) Electron Transfer at the Iron Oxide-Water Interface. Environ. Sci. Technol. 2004, 38 (18), 4782–4790. (26) Cornell, R. M.; Schwertmann, U. The iron oxide: structure, properties, reactions, occurence and uses; Wiley-VCH: Weinheim, Germany, 1996. (27) Morgan, B.; Lahav, O. The effect of pH on the kinetics of spontaneous Fe(II) oxidation by O2 in aqueous solution - basic principles and a simple heuristic description. Chemosphere 2007, 68, 2080–2084. (28) Refait, P.; Abdelmoula, M.; Genin, J.-M.; Sabot, R. Green rusts in electrochemical and microbially influenced corrosion of steel. C. R. Geosci. 2006, 338, 476–487. (29) Nagano, T.; Hirano, T.; Hirob, M. Superoxide Dismutase Mimics Based on Iron in Vivo. J. Biol. Chem. 1989, 264 (16), 9243–9249. (30) Chaudiere, J.; Ferrari-Iliou, R. Intracellular Antioxidants: from Chemical to Biochemical Mechanisms. Food Chem. Toxicol. 1999, 37 (9-10), 949.
ES800086F
VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6735
