Article pubs.acs.org/est
Light-Independent Reactive Oxygen Species (ROS) Formation through Electron Transfer from Carboxylated Single-Walled Carbon Nanotubes in Water Hsin-Se Hsieh,† Renren Wu,‡ and Chad T. Jafvert*,† †
Purdue University, Lyles School of Civil Engineering and Division of Environmental and Ecological Engineering, West Lafayette, Indiana 47907, United States ‡ Ministry of Environmental Protection, South China Institute of Environmental Sciences, Guangzhou 510655, P. R. China S Supporting Information *
ABSTRACT: Promising developments in application of carbon nanotubes (CNTs) have raised concern regarding potential biological and environmental effects upon their inevitable release to the environment. Although some CNTs have been reported to generate reactive oxygen species (ROS) under light, limited information exists on ROS generation by these materials in the dark. In this study, generation of ROS was examined, initiated by electron transfer from biological electron donors through carboxylated single-walled carbon nanotubes (C-SWCNT) to molecular oxygen in water in the dark. In the presence of C-SWCNT, the oxidation of NADH (β-nicotinamide adenine dinucleotide, reduced form) and DTTre (DL-dithiothreitol, reduced form) was confirmed by light absorbance shifts (340 nm to 260 nm during oxidation of NADH to NAD+, and increased light absorbance at 280 nm during oxidation of DTTre). Production of superoxide anion (O2•−) was detected by its selective reaction with a tetrazolium salt (NBT2+), forming a formazan product that is visible at 530 nm. A modified acid-quenched N,N-diethyl-p-phenylenediamine (DPD) assay was used to measure the accumulation of H2O2 in C-SWCNT suspensions containing O2 and NADH. In the same suspensions (i.e., containing C-SWCNT, NADH, and O2), pBR322 DNA plasmid was cleaved, although •OH was not detected when using •OH scavenging molecular probes. These results indicate that the oxidation of electron donors by C-SWCNT can be a light-independent source of ROS in water, and that electron shuttling through CNTs to molecular oxygen may be a potential mechanism for DNA damage by this specific CNT and potentially other carbon-based nanomaterials.
■
INTRODUCTION Manufacturing methods and commercial applications of carbon nanotubes (CNTs) are in rapid development due to their remarkable electrical and chemical properties. Applications in some electronic and medical fields appear to be especially promising.1,2 At the same time, the potential for extensive commercial mass production and use have raised concern regarding their possible adverse health and environmental effects because of inevitable environmental release and exposure.3−5 Indeed, several studies have indicated potential cytotoxicity and DNA damage by CNTs, under both in vivo and in vitro conditions.6,7 In these studies, induction of reactive oxygen species (ROS) causing oxidative stress was suspected as the major mechanism responsible for cytotoxicity, although the mechanism for ROS generation and subsequent DNA damage caused by CNTs has not been elucidated.8 Some CNT formulations are known to generate ROS (e.g., 1O2, O2•−, •OH) in sunlight;9 however, ROS generation involving CNTs during dark reactions is not well characterized. Moreover, it has been reported that CNTs significantly can scavenge free radicals, © 2014 American Chemical Society
making the effect of CNTs on ROS generation more complicated.10,11 Due to the large array of conjugated π bonds, previous research has demonstrated that single-walled carbon nanotubes (SWCNTs) can enhance the electrochemical reactivity of important biomolecules, including promoting electron-transfer reactions in proteins.12−14 These studies indicate that SWCNTs can act as redox mediators by shuttling electrons from biological electron donors (i.e., reducing agents) through SWCNTs to electron acceptors (i.e., oxidizing agents). From an electronic standpoint, SWCNTs can be thought of as solid matrices that can have different Fermi levels based on how many electrons have been transferred to each SWCNT. An additional complication to measuring their electrochemical properties is that commercial preparations are usually a mixture of different chiralities (i.e., rollReceived: Revised: Accepted: Published: 11330
June 30, 2014 August 28, 2014 August 29, 2014 August 29, 2014 dx.doi.org/10.1021/es503163w | Environ. Sci. Technol. 2014, 48, 11330−11336
Environmental Science & Technology
Article
dismutase (SOD), catalase, agarose, and ethidium bromide (EtBr) were obtained from Sigma-Aldrich (St. Louis, MO). pBR322 DNA plasmid (4361 bp, 0.5 μg/μL) was purchased from Thermo Fisher Scientific Inc. (Waltham, MA). To detect several of the molecular probes, HPLC mobile phases were used that contained HPLC-grade methanol and acetonitrile (ACN), and trifluoroacetic acid (TFA) (Acros Organics). All water was purified with a Barnstead Nanopure ultrapure water system after R/O pretreatment. Sample Preparation. Stock dispersions of 50 mg/L were prepared by dispersing C-SWCNT in water with a low energy (80 W) bath sonicator (1210R-MT, Branson Ultrasonics, Danbury, CT) for 30 min. The suspensions were stable with no settling after storage at 4 °C in the dark for several weeks. Stock dispersions were used in experiments without further treatment. In most experiments, 7 mL O2-saturated aqueous volumes containing aliquots of NADH or DTTre, and/or C-SWCNT were mixed in 18 mL borosilicate amber glass vials wrapped in aluminum foil and sealed with PTFE-lined caps. To further reduce light exposure, capped vials were placed in the dark until sacrificed for analysis. Most samples were buffered to pH 7.0 by adding phosphate salts (i.e., KH2PO4 and K2HPO4) to a total phosphate concentration of 5 mM, and the ionic strength was adjusted to 20 mM by adding NaCl. All experiments were performed in duplicate or triplicate. Oxidation of Electron Donors. The rate of oxidation of the electron donors was used as an indication of the catalytic function of C-SWCNT in these systems. To this end, the oxidation of NADH to NAD+ was monitored over time by measuring the light absorbance of samples in quartz cuvettes of 1 cm optical path length with a Varian Cary-300 UV−vis spectrophotometer at 340 and 260 nm, where NADH and NAD+ have absorbance maxima, respectively.23 The oxidation of DTTre to its oxidized form (DTTox) was analyzed by quantifying DTTox by HPLC using a PDA detector, monitoring at 280 nm. In both cases, samples were analyzed after removing the C-SWCNT by filtering each sample through a 0.22 μm PTFE filter. Additional information on the HPLC analysis of DTTox is provided in the SI. Reactive Oxygen Species. The production of different ROS was measured using specific molecular probes and assays. The production of O2•− was detected by adding NBT2+ to some samples before incubation as it selectively reacts with O2•− to form a formazan product. The light absorbance caused by the formation of NBT formazan was measured spectrophotometrically in quartz cuvettes at 530 nm, where it has a molar absorption coefficient of 12,800 M−1 cm−1.24 To detect •OH, pCBA (2 μM) or TPA (1 μM) was added to samples before incubation, and the remaining pCBA or TPA was quantified over time by HPLC analysis, providing necessary information for calculating the steady-state concentration of •OH.25,26 The accumulation of H2O2 in the suspensions was determined by an assay in which N,N-diethyl-p-phenylenediamine (DPD, 714 μM) is oxidized by H2O2 in a peroxidase (HRP, 714 U/L) catalyzed reaction. The light absorbance of the stable product DPD•+ was measured at 551 nm where the molar absorption coefficient is reported to be 21,000 M−1 cm−1.27 All samples measured by this HRP/DPD assay were quenched with 0.2 M HCl to pH < 2.5 for 15 min in order to remove any remaining NADH before analysis.28 Gel Electrophoresis. To assay whether ROS production is significant enough to cause DNA damage, gel electrophoresis was performed to monitor DNA cleavage in C-SWCNT
up angles and diameters) and lengths, with some being semiconducting and others being conducting (i.e., metallic). As a result, the measured redox potential of SWCNTs have been reported within the range −0.14 to −0.60 V vs SHE (standard hydrogen electrode) by applying cyclic voltammetry, redox titrations, or interfacial corrosion (see summary in Supporting Information (SI), Table S1).15−17 However, for biological electron donors with redox potentials less than −0.2 V, such as NADH (β-nicotinamide adenine dinucleotide at −0.32 V)17 and DTTre (dithiothreitol at −0.33 V),18 transfer of some electrons to SWCNTs appears to be thermodynamically favorable, resulting in negatively charged SWCNTs.12,17 Compared to nonfunctionalized SWCNTs, the preparation of carboxylated single-walled carbon nanotubes (C-SWCNT) with a strong oxidant (e.g., HNO3 or H2SO4) leads to hole-doped nanotubes and shifts the Fermi level down (i.e., the redox potential up).19 Therefore, C-SWCNT appear to have greater tendency to accept electrons from electron donors, catalyzing the oxidation of the electron donors. Addition of electrons to each nanotube will result in an incremental increase in the Fermi level for each electron transferred. These electrons in the SWCNTs can be discharged by introducing other reactive electron acceptors with redox potentials lower than the Fermi level of the charged SWCNTs.20,21 In air-saturated aqueous solutions, molecular oxygen is one of the most important electron acceptors, which we hypothesize can receive electrons from the negatively charged CSWCNT once the Fermi level of the C-SWCNT is higher than the redox potential necessary for water-dissolved molecular oxygen (O2,aq) to accept electrons. A one electron transfer would produce superoxide anion (O2•−), whereas transfer of a second electron would produce hydroperoxide (HO2−).22 In this study, we have examined whether carboxylated singlewalled carbon nanotubes (C-SWCNT) suspended in cell-free aqueous physiological buffers can act as facile electron shuttles in dark reactions, receiving electrons from NADH or DTTre, effectively oxidizing them, and then donating these electrons to molecular oxygen to produce various ROS, including O2•−, •OH and/or H2O2. C-SWCNT were chosen because aqueous colloidal dispersions can be prepared easily without added dispersing agents, and this type of CNT has been suggested as a potential agent for drug delivery and diagnostic applications.1 Additionally, these CNTs have relatively low metal catalyst content compared with many nonfunctionalized SWCNTs.9 NADH and DTTre were chosen as reducing agents because of their suitable redox potentials and critical enzymatic roles in biological systems.18,23
■
MATERIALS AND METHODS Materials. C-SWCNT, synthesized by using Nickel/Yttrium catalyst and purified by using nitric acid, were purchased from Carbon Solutions, Inc. (Riverside, CA) and used without further purification. According to the manufacturer, these C-SWCNT contain 1.0−3.0 atomic% carboxylic acid, 5−8 wt % metal content, and >90% carbonaceous purity. The elemental composition of C-SWCNT was previously reported.9 Other chemicals were of the highest purity available and used as received. β-nicotinamide adenine dinucleotide (reduced form, NADH) was purchased from EMD Chemicals, Inc. (Gibbstown, NJ). DL-dithiothreitol (reduced form, DTTre), trans-4,5dihydroxy-1,2-dithiane (oxidized form, DTTox), nitro blue tetrazolium salt (NBT2+), N,N-diethyl-p-phenylenediamine hemioxalate salt (DPD), p-chlorobenzoic acid (pCBA), terephthalic acid (TPA), horseradish peroxidase (HRP), superoxide 11331
dx.doi.org/10.1021/es503163w | Environ. Sci. Technol. 2014, 48, 11330−11336
Environmental Science & Technology
Article
suspensions containing NADH, O2 and pBR322 DNA. First, a 30 μL pBR322 DNA solution (0.5 μg/μL) was diluted by adding 170 μL water to obtain a 0.075 μg/μL DNA solution. Typically, 4 μL of C-SWCNT suspension (50 mg/L), 4 μL of NADH (2 mM), 8 μL of diluted pBR322 DNA (0.075 μg/μL), and 8 μL of phosphate buffer (25 mM, pH 7.0) were mixed with water in microcentrifuge vials to a total volume of 40 μL and incubated in the dark for 5 h. After mixing with 8 μL gel loading dye (6×, purchased from New England Biolabs, Inc.), 10 μL stained samples were loaded onto a 1% agarose gel containing ethidium bromide (0.5 μg/mL). Gel electrophoresis was performed in 1× TBE solutions at 100 V for 60 min in the horizontal gelelectrophoresis system (Thermo EC Class CSSU78) and visualized under a UV trans-illuminator. Images were captured with a UVP BioDoc-It imaging system with an instant camera (Upland, CA).
■
RESULTS AND DISCUSSION Oxidation of NADH and DTTre. Figure 1 shows absorbance spectra as a function of time for O2-saturated water containing C-
Figure 2. HPLC chromatograms (monitored at 280 nm) showing formation of DTTox in suspensions containing 0.1 mM DTTre and 5 mg/L C-SWCNT. The inset shows the temporal trends for a DTTre control, and DTTre suspensions containing 5 mg/L C-SWCNT, and an equivalent-volume of C-SWCNT filtrate, applying the conversion factor reported in the SI, Figure S1.
oxidation of DTTre occurred but at a slower rate than that in samples containing C-SWCNT. Consistent with spectrophotometric results,12,17 the oxidation of each molecule of NADH to NAD+, or DTTre to DTTox, involves transfer of 2 electrons, which presumably are both transferred to C-SWCNT, n (NADH + H+) + C‐SWCNT 2 n → (NAD+ + 2H+) + C‐SWCNTn − (1) 2 n DTTre + C‐SWCNT 2 n → (DTTox + 2H+) + C‐SWCNTn − (2) 2 ROS Production Mediated by C-SWCNT. To similar CSWCNT suspensions containing NADH or DTTre, nitro blue tetrazolium salt (NBT2+) was added prior to sample incubation as a reactive O2•− scavenger. The effects of C-SWCNT on O2•− generation during NADH oxidation is shown in Figure 3. The increase in absorbance at 530 nm in suspension containing CSWCNT indicates formazan formation, whereas in the absence of C-SWCNT (i.e., NADH alone), only a slight increase in absorbance occurred. To test whether formazan formation was caused directly by O2•−, superoxide dismutase (SOD) was added to some samples to facilitate conversion of O2•− to hydrogen peroxide (H2O2), thereby reducing its pseudo steady-state concentration. Figure 3A shows that SOD completely inhibited the production of the formazan product, providing further evidence that O2•− was formed in the NADH/C-SWCNT/O2 systems. By monitoring the change in absorbance over time at 530 nm at different NADH and C-SWCNT concentrations, formation of O2•− is shown to be dose-dependent on both NADH and C-SWCNT (Figure 3B). At 2 mg/L C-SWCNT, significant O2•− generation still occurred in the presence of 0.2 mM NADH. Also the generation of O2•− was shown to be feasible in NADH/C-SWCNT suspensions at pH values between 6 and 9 (see SI Figure S2).
Figure 1. Absorbance spectra of NADH (0.2 mM) in pH 7 phosphate buffer (5 mM, ionic strength = 20 mM) in the presence of C-SWCNT (5 mg/L). The inset shows the temporal trend in relative NADH concentration (monitored at 340 nm where NADH has a maximum molar absorption coefficient).
SWCNT (5 mg/L) and NADH (0.2 mM). The absorbance of NADH at 340 nm decreased as the absorbance of NAD+ at 260 nm increased with time. During irradiation with light, NADH is known to undergo a one-electron photo-oxidation, forming NADH•+ with electron transfer to O2 forming superoxide anion (O2•−).29 However, as shown in the inset of Figure 1, oxidation of NADH in the dark in the absence of C-SWCNT was negligible, even after 36 h. A similar result was found for the oxidation of DTTre to DTTox in the presence of C-SWCNT and O2 (Figure 2). A much more rapid increase in DTTox concentration occurred, monitoring DTTox at 280 nm. Because CNTs preparations often contain metal impurities and some amorphous carbon, filtrates (0.22 μm) of the C-SWCNT stock suspensions were used in some experiments in place of the C-SWCNT suspensions. It is noteworthy that the addition of an equivalent volume of CSWCNT filtrate had almost no effect on the absorbance spectra of NADH (i.e., similar to NADH controls), whereas some 11332
dx.doi.org/10.1021/es503163w | Environ. Sci. Technol. 2014, 48, 11330−11336
Environmental Science & Technology
Article
Figure 3. (A) Evidence of O2•− production is shown due to the formation of NBT formazan in systems containing 0.2 mM NBT2+ and: NADH (0.2 mM) and C-SWCNT (5 mg/L) (■); NADH (0.2 mM), CSWCNT (5 mg/L), and SOD (40 U/mL) (□); NADH (0.2 mM) and an equivalent volume of C-SWCNT filtrate (○); NADH (0.2 mM) alone (△); C-SWCNT (5 mg/L) alone (◊). (B) Temporal trends in NBT2+ (0.4 mM) reduction in systems containing NADH (0.2 mM), and C-SWCNT at 0 mg/L (△); 2 mg/L (□); 5 mg/L (■), and 10 mg/L (○). The inset shows the effect of NADH concentration in 5 mg/L CSWCNT suspensions. On panel B and the inset, the y-axis represents the increase in absorbance at 530 nm after subtracting the intrinsic light attenuation of the initial C-SWCNT suspension as background.
Figure 4. (A) Evidence of O2•− production is shown due to the formation of NBT formazan in systems containing 0.1 mM NBT2+ and: DTTre (0.1 mM), and C-SWCNT (5 mg/L) (■); DTTre (0.1 mM), CSWCNT (5 mg/L) and SOD (40 U/mL) (□); DTTre (0.1 mM) and an equivalent volume of C-SWCNT filtrate (○); DTTre (0.1 mM) alone (△). (B) Temporal trends in NBT2+ (0.1 mM) reduction, in systems containing DTTre (0.1 mM), and C-SWCNT at 0 mg/L (△), 1 mg/L (□), 5 mg/L (■), and 10 mg/L (○). The inset shows the effect of DTTre concentration in 5 mg/L C-SWCNT suspensions. On panel B and the inset, the y-axis represents the increase in absorbance at 530 nm after subtracting the intrinsic light attenuation of the initial C-SWCNT suspension as background.
In the experiments using DTTre as the electron donor with and without SOD, similar results were obtained (Figure 4), again indicating O2•− also was formed in the DTTre/C-SWCNT/O2 systems. Although the C-SWCNT filtrate resulted in some DTTre decay (as should in Figure 2), the filtrate had no effect on O2•− generation in either system (i.e., NADH or DTTre) as shown in Figures 3A and 4A, indicating that generation of O2•− requires the carbon nanotubes rather than any impurities that dissolve upon adding them to water. Although we cannot totally exclude the possible involvement of insoluble metal nanoparticles (NPs) in O2•− generation, the metal catalyst of CSWCNT used in this study are Ni and Y (
