Article pubs.acs.org/est
Bioaccumulation Dynamics and Modeling in an Estuarine Invertebrate Following Aqueous Exposure to Nanosized and Dissolved Silver Farhan R. Khan,*,† Superb K. Misra,‡,§ Javier García-Alonso,†,¶ Brian D. Smith,† Stanislav Strekopytov,‡ Philip S. Rainbow,† Samuel N. Luoma,†,⊥ and Eugenia Valsami-Jones*,‡,§ †
Department of Zoology, Natural History Museum, Cromwell Road, London SW7 5BD, England Department of Mineralogy, Natural History Museum, Cromwell Road, London SW7 5BD, England § School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, England ⊥ John Muir Institute of the Environment, University of California at Davis, Davis, California, United States ¶ Biodiversity Group, Centro Universitario Regional Este, Universidad de la República, Maldonado 20000, Uruguay ‡
S Supporting Information *
ABSTRACT: Predicting the environmental impact of engineered nanomaterials (ENMs) is increasingly important owing to the prevalence of emerging nanotechnologies. We derived waterborne uptake and efflux rate constants for the estuarine snail, Peringia ulvae, exposed to dissolved Ag (AgNO3) and silver nanoparticles (Ag NPs), using biodynamic modeling. Uptake rates demonstrated that dissolved Ag is twice as bioavailable as Ag in nanoparticle form. Biphasic loss dynamics revealed the faster elimination of Ag from Ag NPs at the start of depuration, but similar slow efflux rate constants. The integration of biodynamic parameters into our model accurately predicted Ag tissue burdens during chronic exposure with 85% of predicted values within a factor of 2 of observed values. Zeta potentials for the Ag NPs were lower in estuarine waters than in waters of less salinity; and uptake rates in P. ulvae were slower than reported for the freshwater snail Lymnaea stagnalis in similar experiments. This suggests aggregation of Ag NPs occurs in estuarine waters and reduces, but does not eliminate, bioavailability of Ag from the Ag NPs. Biodynamic modeling provides an effective methodology to determine bioavailable metal concentrations (originating from metal and metal-oxide nanoparticles) in the environment and may aid future ENM risk assessment.
■
INTRODUCTION Silver nanoparticles (Ag NPs) are commonly used in a wide variety of consumer products (clothing, cosmetics, and household appliances) and medical devices as a result of their broad-spectrum biocidal properties. The release of nanosilver from these products has been demonstrated in the case of socks1 and commercially available washing machines, with the effluent of the washing machines severely reducing the abundance of a natural bacterial community.2 A few studies have addressed the bioavailability3−5 and toxic responses6,7 of Ag NPs, but in marine and estuarine waters not enough information is available to guide the regulatory process. Engineered nanomaterials (ENMs) may, in principle, fall under regulations such as the EU’s REACH (Regulation, Evaluation, Authorisation, and Restriction of Chemical substances), but the unique properties of nanosized materials that differentiate them from their bulk or ionic equivalents are not accounted for.8 The U.S. Environmental Protection Agency is currently assessing the risk of nanopartcles, including Ag NPs, which may result from specific nanoscale properties.9 One potential option for environmental regulators would be to base © 2012 American Chemical Society
guidelines on models that predict the impact of nanoparticles in much the same way that the U.S. EPA has adopted the use of biotic ligand models (BLMs) to derive acute water quality criteria based on the bioavailability of dissolved metals (e.g., Cu10). By deconstructing accumulation into singular processes and experimentally addressing processes unique to the nanoparticles (like aggregation) the biodynamic model may offer a viable predictive model for risk assessment. Toxic effects at the cellular, tissue, or whole organism level may be caused by either by the Ag NP itself or by the nanoparticle acting as a delivery vehicle for bioavailable ionic Ag (the “Trojan horse” effect11); but both are tied to bioaccumulation processes determined by uptake and loss dynamics. Biodynamic modeling has become an established method for studying trace metal accumulation in aquatic organisms by characterizing the individual processes of uptake and Received: Revised: Accepted: Published: 7621
March 30, 2012 June 11, 2012 June 14, 2012 June 14, 2012 dx.doi.org/10.1021/es301253s | Environ. Sci. Technol. 2012, 46, 7621−7628
Environmental Science & Technology
Article
ensure the separation of dissolved silver from the particulate silver species, the synthesized particles were dialyzed in deionized water (MW cutoff = 12 kDa). Ag NPs were thereafter stored in the dark until further use. Ag NPs were imaged by TEM (Hitachi 700, 100 KV) on carbon-coated copper grids and size distribution was determined using Gatan software (Gatan Inc., CA, USA). The hydrodynamic diameter and zeta potential (suspension stability) of the particles were assessed (Zetasizer Nano ZS, Malvern Instruments, UK). The UV−vis absorption spectra of the Ag NPs were collected in the range 300−500 nm (Shimadzu spectrophotometer UV1800, Shimadzu Scientific Instruments, MD, USA). These characterizations were performed after synthesis and then before use (7 days later). Characterization was also performed in the experimental media. The hydrodynamic size of the particles and the suspension stability were determined in deionized water, synthetic estuarine water (17 ‰ salinity, pH = 7.4 added as Tropic Marin Sea Salt, Dr Biener GmbH, Germany), and synthetic marine water (33 ‰ salinity, pH = 7.6) at 2 μg L−1 and 22 °C by dynamic light scattering and zeta potential, respectively. Changes in hydrodynamic size and zeta potential were measured at 0, 1, and 7 days to include all experimental renewal time points. Prior to each use the Ag NP stock solution was mixed by inverting several times and an aliquot was removed as a working solution. Aliquots were sonicated for 2 × 30 s at 40 kHz using a benchtop water bath sonicator (Branson B-200, Branson Ultrasonics, CT, USA). Biodynamic Modeling and Membrane Transport Characteristics. Biodynamic models deconstruct bioaccumulation into singular quantifiable processes of uptake and loss from diet and food, and growth where appropriate (see Luoma and Rainbow12 for the full model equation). In this study, we were concerned with predicting Ag concentration in P. ulvae following waterborne exposures to dissolved Ag and Ag NPs. Accumulation is defined by ku (the rate constant for metal uptake from solution in L g−1 d−1), ke (the rate constant of loss in d−1), Cw (concentration in the water in μg L−1) and time (t in d, eq 1).
elimination that collectively result in steady state tissue concentrations.12 The application of biodynamic principles has allowed researchers to discriminate the relative importance of dietary and waterborne exposure routes13,14 and to investigate the differing rates of uptake and efflux between metal tolerant and naive populations.15 Modeled predictions of bioaccumulation have agreed well with measured tissue burdens from field collected samples for a wide range of metals and species.12,16 More recently, Croteau and colleagues have used biodynamics to study the bioavailability and toxicity of ZnO nanoparticles following dietary exposure17 and the influence of capping agents on the bioavailability of Ag NPs in comparison to dissolved Ag following exposure via water and food4 to the freshwater snail Lymnaea stagnalis. The use of hydrobiid snails in ecotoxicology has primarily focused on their ingestion of sediments with varying quality,18,19 but as dominant members of coastal environments, particularly along Northern Atlantic coasts, they are likely to be among a group of organisms that are at risk from ENM exposure. The importance of the estuarine environment is increasingly recognized as a sink for a variety of contaminants.20 Like other trace metal forms that associate with and settle within sediments, Ag NPs may also eventually be trapped within the estuarine environment due to their propensity to aggregate at higher salinities.21 Yet there are few studies of Ag NPs in marine environments that consider the processes that affect Ag bioavailability from those particles.7,22 Physico−chemical properties are known to greatly modify the bioavailability and subsequent toxicity of trace metals and there is evidence to suggest that nanoparticle characteristics, such as dissolution and aggregation, are also affected by salinity, pH, and temperature.23 Salinity reduced the bioavailability and uptake of silver to salmonids24 as the free silver ion (Ag+) is readily complexed by chloride to form AgCl(aq) and AgCl2− complexes.25 In estuarine water the AgCl2− complex is bioavailable, at least to invertebrates, but uptake rates are slower than in circumstances where free ion Ag predominates bioavailability.26 Therefore “dissolved” is a more accurate description than “ionic” for aqueous Ag in estuarine water25 and that term is used here. Characterization of nanoparticles within environmentally relevant media (i.e., estuarine water) is also vital to determine whether bioavailability is affected by differences in environmental chemistry.26 In the present study, we employed a biodynamic approach in exposing the estuarine snail Peringia ulvae (formerly Hydrobia ulvae) to Ag NPs and dissolved Ag via water. In doing so we ask five questions fundamental to understanding bioavailability and effects of Ag NP in estuarine environments: (1) Are Ag NPs bioavailable to an organism in brackish waters where aggregation is likely to be enhanced? (2) How do uptake rates of nanosilver compare to uptake rates of aqueous Ag in estuarine waters? (3) Does the form of Ag (nanoparticle vs dissolved) affect rates of Ag loss from a marine organism? (4) Is it possible to model longer-term bioaccumulation of the Ag when an organism is exposed to Ag NP? and (5) How well do predictions from such models compare to actual uptake?
Accumulation at time t = (k uCwt )/(1 − ket )
(1)
In many such studies a one-component loss model is employed assuming the slower component dominates over long periods of time. In calculating bioacumulation over shorter periods, as is done here, it is important to include both fast (higher rate of efflux over a short period of time) and slow (lower efflux rate over a greater time period) elimination phases if they are present (e.g., ref 28). To predict the steady state concentration of P. ulvae exposed to dissolved Ag and Ag NPs using a two-component loss model (eq 2), the single efflux component in eq 1 was separated into “fast” and “slow” phases, where ke1 is the fast efflux rate constant for the duration of the fast elimination phase (t1), ke2 is the slow efflux rate constant for the duration of the slow elimination phase (t2), and t3 is the total length of depuration.
■
Css = Cwku(1 − (ke1t1) + (ke2t 2))t3
MATERIALS AND METHODS Synthesis and Characterization of Silver Nanoparticles (Ag NPs). Citrate-capped Ag NPs were synthesized based on the method of Doty et al.27 Briefly, 10 mM NaBH4 was added to a mixture of sodium citrate (1.25 mM) and AgNO3 (10 mM) at 25 °C and stirred continuously for 3 h. To
(2)
Metal influx (influxorganism in nmol g−1 d−1) at the site of uptake (i.e., gill for waterborne exposure) can also be interpreted in terms of membrane transporter characteristics (eq 3) where Bmax is the binding site density (in nmol g−1), Kmetal is the transporter affinity of each biding site (in nmol g−1) 7622
dx.doi.org/10.1021/es301253s | Environ. Sci. Technol. 2012, 46, 7621−7628
Environmental Science & Technology
Article
Table 1. Ag NP Behavior Measured as Hydrodynamic Diameter (nm) and Suspension Stability (Zeta Potential in mV) in Deionized, Estuarine (17 ‰), and Marine (33 ‰) Waters over 96 h (n = 3 Replicates Per Measure)a hydrodynamic diameter (nm)
Zeta potential (mV)
water
0h
1 day
7 days
0h
1 day
7 days
d-H20 estuarine marine
32 ± 1 79 ± 13 162 ± 21
32 ± 2 145 ± 6 339 ± 6
31 ± 1 164 ± 6 266 ± 3
−41 ± 1 −12 ± 2 −13 ± 1
−32 ± 2 −16 ± 1 −15 ± 2
−28 ± 2 −16 ± 1 −14 ± 2
a
Results demonstrate an increase in aggregation at higher salinities. The consistency of the zeta potentials over time would indicate that colloidal instability continues over time and hydrodynamic diameters are subject to fluctuations.
from which the log K is derived as an affinity constant and [M]exposure is the exposure concentration (nmol L−1).4,29 Influx organism = Bmax [M]exposure /K metal + [M]exposure
This was necessary in order to reveal subtle differences in efflux rates over the depuration phase. The depuration phase was also increased to 24 days. In the loading phase, eight groups of 48 snails were exposed to 556 nmol L−1 (60 μg L−1) as dissolved Ag or Ag NP for 5 days. Exposure solutions were renewed daily and snails were not fed during this period. Post exposure, as previously described, snails were transferred to clean water to depurate and were subsampled at t = 1, 2, 4, 8, 12, 16, and 24 days. Feeding and water changes occurred as previously described. Chronic Exposures and Model Validation. Longer term net uptake experiments were used to validate model predictions derived from eqs 1 and 2 using calculated uptake and efflux rate constants. Static renewal 21 day exposures were conducted at 186 nmol L−1 (20 μg L−1) and 464 nmol L−1 (50 μg L−1) with dissolved Ag and Ag NPs. Exposures were performed in acid washed presoaked 50 mL glass beakers as previously described. Exposure solutions were renewed every 4 days and snails were subsampled on days 0, 2, 6, 10, 16, and 21 for exposures at 186 nmol L−1 and days 0, 1, 2, 6, 12, and 21 for exposures at 464 nmol L−1. P. ulvae were allowed to feed on C. mulleri for 2 h prior to water changes. As with uptake and elimination experiments control and background samples were analyzed, and postexposure snails were sacrificed by freezing. Sample Preparation and Analysis. Ag in the snail’s soft tissue was determined by ICP-MS (Varian 810) following HNO3 digestion. Details of the tissue dissection, digestion protocol, and sample analysis can be found in the Supporting Information.
(3)
Uptake and Elimination Experiments. Details on the collection and maintenance of P. ulvae can be found in the Supporting Information. All laboratory glassware was soaked in 20% HCl and thoroughly rinsed in ultrapure water before use. Exposure chambers (50 mL glass beakers) were presoaked at the exposure concentration (with either dissolved Ag or Ag NPs) for 24 h to saturate any surface binding. Uptake and elimination experiments were based on the methods of previous studies that have employed a biodynamic approach to understanding metal accumulation from solution.4,14,16 In all exposures dissolved Ag (added as AgNO3, VWR, UK) and Ag NPs were added at equivalent Ag concentrations to estuarine water (17 ‰ salinity). To characterize silver uptake, and compare the uptake rate constants (ku), P. ulvae were exposed to Ag concentrations ranging from 12 to 1855 nmol L−1 (1.25−200 μg Ag L−1) for 24 h (one exposure of 48 snails per concentration). Snails were not fed during the 24 h exposure. Postexposure snails were rinsed with ultrapure water, blotted dry, and sacrificed by freezing. To determine the unidirectional efflux rate (ke) between the different forms of Ag, snails were first loaded with Ag via waterborne exposure to dissolved Ag or Ag NPs and then depurated in clean water for 16 or 24 days. In elimination experiments providing a single ke (as used in the onecomponent efflux model (eq 1)), snails (6 or 7 groups of 48 individuals) were exposed for 24 h at 556 and 928 nmol L−1 (60 and 100 μg L−1) as dissolved Ag or Ag NPs. These exposures were conducted using the same procedure as the uptake experiments. Following the loading phase one set of snails was immediately sacrificed (t = 0) to determine the initial accumulated concentration and the remaining groups were rinsed in ultrapure water and transferred to clean 50 mL acid washed plastic beakers (closed by 1 mm mesh at the top and bottom) which were partially submerged in a plastic tank containing 3 L of aerated estuarine water. Groups exposed to dissolved Ag and Ag NPs were kept in separate tanks. Snails were depurated for 16 days during which time one group was removed at each sampling time point (t = 1 (928 nmol L−1 exposures only), 2, 4, 8, 12, and 16 days). During depuration snails were allowed to feed on diatoms (Chaetoceros mulleri) for 2 h before water changes which were made by complete renewal of the 3 L tank water. Water changes were made at the sampling time points. In the experiment used to determine rate constants associated with two-component efflux, a longer loading phase was used to increase the initial accumulated Ag body burden.
■
RESULTS Characterization of Ag NPs. The intrinsic properties of the synthesized Ag NPs are shown in the Supporting Information (S2). The mean size, as determined by TEM, was 16.5 ± 4.5 nm (n = 200 Ag NPs). UV−vis spectra demonstrated a maximum absorbance at 391 nm immediately after synthesis and at 7 days (time elapsed before use) indicating that there was no change in size. Similarly, there was a no significant change in the hydrodynamic size of the NPs (Day 1 = 32 nm, Day 7 = 36 nm) and the colloidal suspension stability remained high (Day 1 = −35 ± 1 mV, Day 7 = 35.4 ± 1.2 mV). The nanoparticles’ aggregation potential and suspension stability were measured as hydrodynamic diameter (DLS) and suspension stability (zeta potential) in deionized, estuarine (17 ‰), and marine (33 ‰) waters over 7 days (Table 1). The change in zeta potential indicated the relative aggregation of the Ag NPs which was confirmed by the increase in hydrodynamic diameter (Table 1, Figure S3). Dissolution of the Ag NPs was not measured in estuarine water, but recent studies have concluded that in solutions of near-neutral pH (measured pH 7623
dx.doi.org/10.1021/es301253s | Environ. Sci. Technol. 2012, 46, 7621−7628
Environmental Science & Technology
Article
Figure 1. Ag influx into Peringia ulvae soft tissue following 1 day waterborne exposures to dissolved Ag (A, black circles) and Ag NPs (B, open circles) for (mean values (nmol g−1 (dw) d−1) ± S.D., n = 6 pooled replicates per data-point from 48 individuals). Concentrations are shown following the subtraction of the background Ag concentration. Linear regression (solid line) was used to determine the uptake rate constants (nmol g−1 d−1), and nonlinear regression (Michaelis−Menten) fits were used to derive metal binding characteristics.
Table 2. Biodynamic Parameters (± 95% C.I.) and Metal Binding Characteristics (± S.E.) for the Estuarine Mud Snail P. ulvae Derived from Exposures to Dissolved Ag and Ag NPsa P. ulvae dissolved Ag biodynamic parameters ku (L g−1 d−1) ke (d−1) ke1 (d−1) ke2 (d−1) metal binding characteristics Bmax (nmol g−1) Kmetal (nmol L−1) Log K a
0.15 0.026 0.13 0.016
± ± ± ±
0.037 0.013 0.082 0.004
139 ± 8 672 ± 91 6.17
L. stagnalis cit-Ag NPs
dissolved Ag
cit-Ag NPs
± ± ± ±
1.1 ± 0.1 0.004 ± 0.0013
0.35 ± 0.01 0.058 ± 0.019
0.074 0.027 0.21 0.017
0.031 0.024 0.795 0.02
62 ± 6 489 ± 134 6.31
31 ± 5 20 ± 7 7.70
18 ± 6 29 ± 23 7.54
Comparative figures for the freshwater snail, Lymnaea stagnalis, are also shown.4
of estuarine water was 7.4) dissolution of nanosilver is slow,30 and typically
