Environmental Science and Pollution Research

, Volume 20, Issue 5, pp 3456–3463

Toxicity of two types of silver nanoparticles to aquatic crustaceans Daphnia magna and Thamnocephalus platyurus

Authors

    • Laboratory of Molecular GeneticsNational Institute of Chemical Physics and Biophysics
  • Jukka Niskanen
    • Laboratory of Polymer Chemistry, Department of ChemistryUniversity of Helsinki
  • Paula Kajankari
    • Department of Environmental SciencesUniversity of Helsinki
  • Liina Kanarbik
    • Laboratory of Molecular GeneticsNational Institute of Chemical Physics and Biophysics
    • Department of Chemical and Materials TechnologyTallinn University of Technology
  • Aleksandr Käkinen
    • Laboratory of Molecular GeneticsNational Institute of Chemical Physics and Biophysics
    • Department of Chemical and Materials TechnologyTallinn University of Technology
  • Heikki Tenhu
    • Laboratory of Polymer Chemistry, Department of ChemistryUniversity of Helsinki
  • Olli-Pekka Penttinen
    • Department of Environmental SciencesUniversity of Helsinki
  • Anne Kahru
    • Laboratory of Molecular GeneticsNational Institute of Chemical Physics and Biophysics
Research Article

DOI: 10.1007/s11356-012-1290-5

Cite this article as:
Blinova, I., Niskanen, J., Kajankari, P. et al. Environ Sci Pollut Res (2013) 20: 3456. doi:10.1007/s11356-012-1290-5

Abstract

Although silver nanoparticles (NPs) are increasingly used in various consumer products and produced in industrial scale, information on harmful effects of nanosilver to environmentally relevant organisms is still scarce. This paper studies the adverse effects of silver NPs to two aquatic crustaceans, Daphnia magna and Thamnocephalus platyurus. For that, silver NPs were synthesized where Ag is covalently attached to poly(vinylpyrrolidone) (PVP). In parallel, the toxicity of collargol (protein-coated nanosilver) and AgNO3 was analyzed. Both types of silver NPs were highly toxic to both crustaceans: the EC50 values in artificial freshwater were 15–17 ppb for D. magna and 20–27 ppb for T. platyurus. The natural water (five different waters with dissolved organic carbon from 5 to 35 mg C/L were studied) mitigated the toxic effect of studied silver compounds up to 8-fold compared with artificial freshwater. The toxicity of silver NPs in all test media was up to 10-fold lower than that of soluble silver salt, AgNO3. The pattern of the toxic response of both crustacean species to the silver compounds was almost similar in artificial freshwater and in natural waters. The chronic 21-day toxicity of silver NPs to D. magna in natural water was at the part-per-billion level, and adult mortality was more sensitive toxicity test endpoint than the reproduction (the number of offspring per adult).

Keywords

EcotoxicologyCrustaceansSilver nanoparticlesCollargolBioavailabilityNatural water

Introduction

Silver nanoparticles are increasingly used in consumer applications. The Woodrow Wilson Database (2011), although not comprehensive, has listed 1,317 nanoparticle (NP)-based consumer products currently on the market, 311 of which contain silver nanoparticles (AgNPs). Colloidal silver has been used for more than 100 years (Nowack et al. 2011). Already since the late nineteenth century, colloidal silver such as collargol (protein-stabilized nanosilver) was used for numerous medical purposes (Fung and Bowen 1996). According to Mueller and Nowack (2008), approximately 500 tonnes of nanosilver is produced per year. Thus, there is a high risk of environmental pollution by nanosilver due to its leaching from consumer products, as well as through industrial waste streams. Even though an increasing number of reports on the toxicity of different AgNPs to aquatic species have been published during the last decade, the information is still scarce (Fabrega et al. 2011; Kahru and Dubourguier 2010; Wijnhoven et al. 2009). Among the papers on silver NPs, the number of articles published on the effect of nanosilver to bacteria and environmentally relevant aquatic organisms such as crustaceans, algae, and fish is still low. Majority of the reports concern the effects on bacteria, mostly Escherichia coli, showing the importance of research on the antibacterial applications of nanosilver (Fig. S1 in Supplementary material). Often, AgNPs used in various commercial products are coated with organic compounds to promote their dispersability. The coatings/capping agents may in turn influence the bioavailability and fate of the particles in the environment.

The experimentally measured toxicity of nanosilver varies depending on the test species. For example, a recent review by Kahru and Dubourguier (2010) showed that AgNPs were toxic to algae and crustaceans even at very low concentration (EC50 < 1 mg/L), but the toxicity to protozoa was relatively low, EC50~40 mg Ag/L. Therefore, any additional information on the toxicity of different formulations of silver NPs to environmentally relevant test species in varying test conditions will be helpful for realistic risk assessment of AgNPs. As the bioavailability of silver NPs owes to numerous factors determined already in the synthesis stage (Tolaymat et al. 2010; Xia et al. 2011), the need to report in detail the preparation methods is evident.

In the current study, the toxicity of two types of silver NPs, i.e., (1) poly(vinylpyrrolidone) (PVP)-stabilized AgNPs and (2) collargol (protein-coated silver NPs), are reported for two aquatic crustaceans, Daphnia magna and Thamnocephalus platyurus. The toxicity of silver NPs was compared with the toxic effect of a soluble silver salt, AgNO3, to evaluate the contribution of the dissolved silver to the overall toxic effect of silver NPs.

PVP is a nontoxic polymer widely used as an additive in pharmaceutical applications (Haaf et al. 1985), in cosmetics (Goddard and Gruber 1999), and even as a blood plasma expander (Kuo et al. 1997). It is important to note that most often PVP is used as a stabilizing agent for various types of NPs, as silver, without binding it covalently to the particles (Greulich et al. 2011; Ledwith et al. 2007). In the current work, PVP was first synthesized by reversible addition–fragmentation chain transfer radical polymerization (RAFT) (Chiefari et al. 1998), and next, the obtained polymer was used in the preparation of water-dispersible silver NPs. In this way, the polymer was covalently attached to the particles via sulfur linkages. This is a technique that has been formerly used in the preparation of gold nanoparticles (Lowe et al. 2002; Shan et al. 2003). Essentially, the same chemistry to graft silver nanoparticles with rubbery polymeric (meth)acrylates was recently used by Niskanen et al. (2010).

D. magna was chosen as a test species because it is among the most sensitive freshwater test species to silver (Ratte 1999). In addition, for D. magna, an OECD test method has been developed both for short-term and long-term endpoints. Both of these endpoints are listed as necessary environmental ones for hazard evaluation of synthetic nanomaterials in Phase-One of the OECD Sponsorship Programme for the Testing of Manufactured Nanomaterials (List of manufactured nanomaterials and list of endpoints for phase one of the sponsorship programme for the testing of manufactured nanomaterials: revision 2010). The majority of the information published so far on the toxicity of silver NPs to environmentally relevant aquatic species concerns D. magna; this information allows the comparison of results. Another crustacean, T. platyurus, was used as an additional test species to compare species sensitivity. To increase the environmental relevance of our studies, the toxicity of silver NPs to crustaceans was evaluated also in natural water.

Materials and methods

Chemicals

For the synthesis of PVP-Ag, the following chemicals were used. The initiator azobisisobutyronitrile (Fluka, >98.0 %) was recrystallized from methanol. The monomer N-vinyl-2-pyrrolidone (Polysciences, 99 %) and solvent 1,4-dioxane (Aldrich, >99.5 %) were distilled prior to use. The chain transfer agent cyanopentanoic acid dithiobenzoate was synthesized according to Shan et al. (2003). Diethyl ether (J.T. Baker, anhydrous), silver nitrate (VWR International, Ph. Eur.), sodium borohydride (Sigma-Aldrich, >98.5 %), tetrahydrofuran (THF; Lab-Scan, 99.8 %), silver nitrate (Fluka), and collargol (protein-coated colloidal silver, purchased from an Estonian drugstore) were used for the toxicity testing as received.

Synthesis of PVP-stabilized silver nanoparticles (PVP-Ag)

For the preparation of polyvinyl pyrrolidone (PVP) and for the reduction of the polymer dithiobenzoate end group into a thiol (PVP-SH), N-vinyl-2-pyrrolidone (4 mL, 37.5 mmol), 1,4-dioxane (5 mL), cyanopentanoic acid dithiobenzoate (50.48 mg, 0.18 mmol), and azobisisobutyronitrile (5.25 mg, 0.03 mmol) were placed in a Schlenk flask. Oxygen was removed by four freeze–thaw cycles using a Schlenk line. The reaction flask was placed in an oil bath at 60 °C for 20 h. The polymer was purified by precipitation in diethyl ether twice and freeze-dried. Altogether 1.98 g of polymer was obtained. The molar mass (Mn) of the polymer was determined by size exclusion chromatography to be 13,600 g/mol with a polydispersity index of 1.2. For the reduction of the polymer end group, 1.0 g polymer was dissolved in 100 mL THF/water (1:1) mixture. NaBH4 (59.3 mg, 1.6 mmol) was then added to the solution, and the reaction was left overnight in an open vessel. The polymer obtained was purified by dialysis against water and freeze-dried.

Altogether, four batches of PVP-Ag were produced using PVP with a dithiobenzoate end group or PVP with a thiol end group (PVP-SH). Three of them were used for the current study, designated as PVP-Ag1, PVP-Ag3, and PVP-Ag4 (Table 1). Briefly, PVP (234 mg) was dissolved in 40 mL water followed by the addition of AgNO3 (141 mg, 0.83 mmol). A solution of NaBH4 (312 mg, 8.3 mmol) in 10 ml water was slowly added to the polymer solution. The formation of NPs could be observed as the color changed from colorless to brown. After stirring overnight in an open vessel a clear, dark brown solution of nanoparticles was obtained. The particles were purified by dialysis against water and freeze-dried.
Table 1

Characterization of AgNPs

AgNP

Coatinga

Size of the NPsb, nm

%

Mean

St. dev.

PVP-Ag1

73

6.3

2.3

PVP-Ag3

56

6.0

3.0

PVP-Ag4

71

8.4

2.8

Collargol

30

12.5

4.0

aMass fraction, analyzed by thermogravimetry (see also Table S1)

bMeasured from the TEM images (see Figures S25)

Characterization of the Ag nanoparticles used for the toxicity tests

The PVP-AgNPs were characterized by thermogravimetric analysis (TGA) and transmission electron microscopy (TEM). The hydrodynamic diameter and zeta (ζ) potential of collargol and PVP-Ag4 (concentration 10 mg Ag/L) in media used for the toxicity testing, i.e., in distilled water, OECD Daphnia medium and natural water, were determined at 25 °C using Zetasizer Nano ZS equipped with a 633-nm He–Ne laser (Malvern Instruments, UK). Before the addition of NPs, natural waters were filtered through PTFE filters (pore size, 0.45 μm). The intensity weighted average hydrodynamic radii were obtained using the software by the manufacturer. ζ potential measurements were performed using Doppler electrophoresis as the basic principle of operation. The instruments for TGA as well as the methods used to determine the molar masses of the polymers are presented in Supplementary material, Annex 2.

Toxicity tests

In the acute immobilization tests with D. magna (water flea), the neonates less than 24 h old obtained by the hatching of ephippia were exposed to different concentrations of Ag compounds at 20 °C in the dark for 48 h (OECD 202). Differently from the standard test procedure (OECD 202), neonates used for the toxicity tests were not from a laboratory culture but were hatched from dormant eggs. Both, the ephippia of D. magna and cysts of T. platyurus were purchased from MicroBio-Tests, Inc. (Mariakerke-Gent, Belgium). In the acute mortality test with the crustacean T. platyurus (fairy shrimp), the larvae hatched out from the cysts were exposed at 25 °C in the dark for 24 h. The tests were performed three to four times each in four (D. magna) and three (T. platyurus) replicates.

In the D. magna reproduction test (OECD 211), the neonates less than 24 h old were exposed at 21 ± 1 °C with 16:8 h light/dark photoperiod during 21 days. Endpoints were the mortality of the parents during the test and the total number of juveniles produced per parent alive at the end of the test. D. magna was fed daily with the algae Pseudokirchneriella subcapitata (0.1–0.2 mg C/Daphnia/day). The test medium was renewed every 3 days; the parent animals were transferred to vessels with fresh medium with a glass pipette.

The artificial fresh water (AFW) and natural water from rivers and lakes were used as test media. The moderately hard AFW (US EPA 2005) was used in tests with T. platyurus and ISO medium (OECD 202) for D. magna. Samples of natural water were collected in Estonia and Finland, filtered through 0.45-μm-pore-size filters and stored at 4 °C.

The stock suspensions of AgNPs (500 mg/L) were prepared in MilliQ water, sonicated for 30 min, and stored in the dark. The silver concentrations in the stock solutions were determined with a Varian SpectrAA 220Z Atomic Absorption Spectrometer.

Results and discussion

Physical–chemical characterization of AgNPs

Poly(vinyl pyrrolidones) with narrow molar mass distributions and dithiobenzoate end groups were prepared by the RAFT polymerization technique. In the syntheses of AgNPs, the polymers were used as such or after reducing the end groups into thiols, both methods being as effective. Silver ions were reduced into metallic silver in the presence of the end-functionalized polymers, and in this way, stable NPs stabilized with PVP were obtained. Several batches of water-dispersible silver NPs with PVP bound to the particles with sulfur bonds were produced. It is worth noticing that usually PVP has been used in stabilizing metallic NPs by simply adsorbing the polymer to the particle surfaces, and thus, the concept of using covalently bound polymers as stabilizing agents is fairly novel. Some variation in the amount of the grafted polymer in different batches of the silver nanoparticles was observed. PVP-SH was used in the first batch (PVP-Ag1). This produced smaller particles with a higher amount of polymer than what was obtained in the last two batches with PVP carrying dithiobenzoate end groups (PVP-Ag3 and 4). Both methods yielded similar particles where the polymer is bound to the metal surface with a sulfur bond (Table S1 in the Supplementary material).

For AgNPs, TEM micrographs and light scattering data are presented in the Supplementary material (Fig. S25). As shown in Table 1, the core size of synthesized PVP-AgNPs (6.3–8.4 nm) was slightly smaller than the size of protein-coated AgNPs (collargol), whereas the mass fraction of the organic coating in the case of PVP-Ag was noticeably larger (56–73 %) than that in collargol (30 %).

Acute toxicity of AgNPs to crustaceans

The toxicity of three batches of PVP-Ag (Table 1) was first screened with the crustacean T. platyurus acute assay as this test needs a smaller sample volume and can be conducted more rapidly than the D. magna acute assay. The results revealed that small differences in particle core size, percentage of synthetic polymer on the surface, and thus hydrodynamic size (Tables 1 and S1) did not influence the acute toxic properties of PVP-AgNPs to T. platyurus (data not shown); therefore, only one batch of three (PVP-Ag 4) was used for further investigation.

Acute toxicity of PVP-AgNPs, collargol, and silver nitrate was tested with two crustaceans. Five natural water samples with different hydrochemical characteristics were used in addition to AFW to study the bioavailability/toxicity of the silver compounds to aquatic crustaceans in environmentally relevant test media (Table 2). The wide concentration range of inorganic salts (observed as conductivity varying 4-fold) and different amount of dissolved organic carbon (DOC) in natural waters used as test media reflect the hydrochemical diversity of surface waters in Estonia and Finland (Table 2).
Table 2

Characterization of the test media

Test medium

Conductivity

DOCa

pH

20 °C (μS/cm)

mg C/L

River 1

560

15.3

8.3

River 2

421

35

7.45

Lake 1

439

11.8

8.3

Lake 2

143

22

7.6

Lake 3

315

5.3

7.1

AFWDaphnia

640

0

7.8–8.0

AFWThamnocephalus

665

0

7.8–8.0

aDissolved organic carbon

The analysis of the hydrodynamic diameter (Dh) and zeta potential of the silver NPs (10 ppm Ag) showed that the Dh of collargol in all tested media was about half of that of PVP-Ag4 (50.7 versus 105.7 nm in AFW). Apparently, the protein coating of collargol behaves differently from electroneutral PVP in stabilization of nanosilver suspensions. Also, the higher absolute zeta potential of collargol (−16… −42 mV) than that of PVP-Ag4 (about −5 mV) shows that the behavior of the NPs in aqueous media is different. However, the chemical composition of different test media did not affect the Dh of AgNPs (collargol), or the effect was not remarkable (PVP-Ag4) (Table 3; Fig. 1).
Table 3

Hydrodynamic diameter and zeta potential of collargol and PVP-Ag4 suspensions (10 ppm Ag) in three different test media

Analysis medium

Collargol

PVP-Ag 4

Hydrodynamic diameter (Dh), nm

Zeta potential, mV

Hydrodynamic diameter (Dh), nm

Zeta potential, mV

Milli Q water

50.7

−42.7

122.4

−6.97

AFWDaphnia

50.7

−16.5

105.7

−4.86

River 2

50.7

−16.1

141.8

−4.93

https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1290-5/MediaObjects/11356_2012_1290_Fig1_HTML.gif
Fig. 1

Hydrodynamic diameter (Dh) of PVP-Ag4 NPs (right panel) and collargol (left panel) in three different test media. Analysis was performed using dynamic light scattering

Acute toxicity (EC50) of collargol, PVP-AgNPs, and AgNO3 to crustaceans D. magna and T. platyurus in six different test media (see also Table 2) is presented in Table 4. The E(L)C50 values for AgNO3 in AFW were very low: 1.4 ppb for D. magna and 3.6 ppb for T. platyurus (Table 4) and comparable with the earlier reported data on toxicity of soluble silver salts to D. magna (Erickson et al. 1998).
Table 4

Acute toxicity, E(L)C50, of the Ag compounds to crustaceans in six different test media, in micrograms Ag/L

Test compound

Test medium

Daphnia magna

Thamnocephalus platyurus

48 h EC50

24 h LC50

Mean

STD

Mean

STD

Collargol

AFW

49.4 (36.6)

19.7 (14.6)

256 (189)

51.8 (38.3)

River 1

59.4 (44.0)

9.4 (6.9)

178 (132)

41.0 (30.4)

River 2

40.2 (29.8)

16.1 (11.9)

147 (109)

5.9 (8.0)

Lake 1

74.9 (55.5)

18.6 (13.8)

n.d.

n.d.

Lake 2

50.8 (37.6)

1.8 (1.3)

250 (185)

13.8 (10.2)

Lake 3

65.7 (48.7)

7.0 (5.2)

n.d.

n.d.

PVP-Ag4

AFW

54.0 (15.7)*

1.4 (0.4)

68.8 (20.0)*

1.4 (0.4)

River 1

191 (55.5)

80.5 (23.4)

191 (55.5)

13.4 (3.9)

River 2

98.7 (28.7)

31.3 (9.1)

n.d.

n.d.

Lake 1

176 (51.1)

15.1 (4.4)

252 (73.3)

62.3 (18.1)

Lake 2

236.3 (68.7)*

97.7 (28.4)

605 (176.0)*

113 (33.1)

Lake 3

162 (47.2)

59.8 (17.4)

n.d.

n.d.

AgNO3

AFW

2.2 (1.4)*

0.5 (0.3)

5.7 (3.6)*

0.6 (0.4)

River 1

12.4 (7.9)

4.6 (2.9)

10.7 (6.8)

1.7 (1.1)

River 2

15.9 (10.1)

2.4 (1.5)

n.d.

n.d.

Lake 1

8.3 (5.3)

0.5 (0.3)

11.6 (7.4)

0.5 (0.3)

Lake 2

12.9 (8.2)

4.1 (2.6)

24.3 (15.5)

7.2 (4.6)

Lake 3

6.8 (4.3)

0.2 (0.1)

n.d.

n.d.

AFW artificial freshwater, n.d. not determined

*p < 0.05 (statistically significant differences from other test media)

The results of acute tests revealed that toxicity of both studied AgNPs (collargol and PVP-Ag) to crustaceans was similar when calculated based on the silver content. Here, it is important to note that the share of the organic coating of PVP-AgNPs was up to 73 % whereas in the case of collargol, the share of the coating was 30 %. Silver from both types of NPs (Table 4) was up to 10-fold less bioavailable to crustaceans in all test media than Ag from the soluble silver salt (AgNO3). The lower toxicity of different silver NPs to daphnids compared to a soluble silver salt has also been reported by Allen et al. (2010), Griffitt et al. (2008), and Zhao and Wang (2011).

The acute toxicity (E(L)C50 values) of all studied silver compounds to D. magna (water flea) and T. platyurus (fairy shrimp) was of the same order of magnitude, but the latter test species was slightly less sensitive to silver. As shown previously by Blinova et al. (2010), the sensitivity of these two aquatic species was similar also to CuO nanoparticles. Both crustacean species showed similar trends for toxicity of investigated silver compounds in different test media (Table 4). The big variation (STD in Table 4) of E(L)C50 values may be explained firstly by the steep slope of the dose–response curve, i.e., a small increase in silver dose caused a large increase in toxicity, and, secondly, by the instability of NP suspensions (dissolution, aggregation, settling) in the test media (Liu and Hurt 2010). The measurements showed that the absolute value of zeta potential of both nanosilver suspensions in MilliQ water was higher (indicating higher electrostatic stability) than in other test media (Table 3). The observed remarkably high toxicity of nanosilver compounds (in the parts-per-billion range) to crustaceans indicates that these organisms are a vulnerable link in the aquatic food chain concerning contamination by nanosilver.

Bioavailability of all the studied silver compounds (Table 4) decreased in natural water by a factor of 3–5 as compared with AFW. These results are in general agreement with the data of other authors, showing that in natural water, the toxicity of silver salts (Erickson et al. 1998) and silver NPs (Gao et al. 2009) was remarkably lower than in AFW. The interaction with different organic and/or inorganic components (water hardness, DOC, sulfides, etc.) in natural water may significantly modify the silver speciation and bioavailability/toxicity to living organisms (Bianchini and Wood 2008; Ratte 1999). In the current study, the relationship between the silver toxicity and concentration of dissolved organic matter was revealed only for AgNO3 (R2 = 0.88). A higher DOC was accompanied by a lower toxicity (Table 4). This indicates that the behavior of AgNPs in natural water may differ from that of Ag ions, and thus, the fate of silver NPs in an aquatic environment most probably cannot be predicted on the basis of the existing knowledge on behavior of silver ions in natural water. The commercially available nanosilver particles have different coatings and correspondingly, different surface properties. Thus, it may be assumed that their fate in the environment may be different. Bone et al. (2012) have recently shown that the behavior of different AgNPs in the same test media was coating dependent.

Chronic toxicity of silver nanoparticles to D. magna

The dose–effect data on the mortality of the daphnids and the reduction of the offspring per surviving adult upon 21 days of exposure to PVP-Ag and collargol are presented in Table 5. A very steep dose–effect response was typical for D. magna mortality. The data show that in the current test setting, it was not possible to determine the EC50 values for the reproduction endpoint, because the tested organisms died before a 50 % decrease in reproduction was reached. Thus, in the case of AgNPs, the 21-day mortality endpoint could be considered more relevant and more sensitive than the reduction of offspring per adult (reproduction). We agree with Nebeker et al. (1983) that for silver, “a simpler and less expensive test using just long-term survival might be adequate to predict toxicity.” In the case of chronic tests with silver compounds, some additional endpoints such as body length of daphnids may be more informative than mortality.
Table 5

Long-term toxic effect of silver nanoparticles to D. magna exposed in two types of natural water for 21 days

Test compound

Test media

(μg Ag/L)

Mortality of the adults, %

Decrease of the offspring per adult, %

PVP-Ag4

River 2

50.0 (14.5)

0

No effect*

PVP-Ag4

River 2

100 (29)

0

No effect*

PVP-Ag4

River 2

200 (58)

0

No effect*

PVP-Ag4

River 2

300 (87)

90

No effect*

PVP-Ag4

Lake 3

5 (1.4)

0

No effect*

PVP-Ag4

Lake 3

10 (2.9)

0

No effect*

PVP-Ag4

Lake 3

25 (7.2)

0

No effect*

PVP-Ag4

Lake 3

50 (14.5)

0

No effect*

PVP-Ag4

Lake 3

85 (25)

0

No effect*

PVP-Ag4

Lake 3

100 (29)

0

23 %**

PVP-Ag4

Lake 3

175 (50)

100

Collargol

River 2

50.0 (37)

0

No effect*

Collargol

River 2

100 (74)

20

No effect*

Collargol

River 2

200 (148)

100

No effect*

Collargol

Lake 3

50.0 (37)

0

No effect*

Collargol

Lake 3

100 (74)

30

43 %**

Collargol

Lake 3

200 (148)

100

Dose–effect data are presented

*p > 0.05 (no statistically significant differences from the control)

**p < 0.05 (statistically significant decrease of the offspring per adult)

The comparison of the acute and chronic toxicity data for silver NPs toward D. magna in two different natural waters (river 2 and lake 3; Table 2) shows that upon the long-term (21 days) exposure, the toxicity of PVP-AgNPs was reduced compared with the acute tests (Tables 4 and 5). For example, the 48-h EC50 of PVP-Ag4 (river 2) was 28.7 ppb Ag whereas in the chronic assay, all the daphnids were alive even after 21-day exposure to 58-ppb Ag. Analogously, the 48-h EC50 for collargol (river 2) was 51.3, whereas the 21-day exposure to 37-ppb collargol did not cause any mortality. The 48-h EC 50 of collargol (lake 3) was 48.7 whereas the 21-day exposure to 74-ppb collargol (lake 3) caused only 30 % mortality.

One reason for the reduced toxicity in the chronic test could be the addition of algae as food. It should be mentioned that toxicity of silver NPs to P. subcapitata (used as food in the current test setting) is nearly 5-fold lower than that to the crustaceans (Griffitt et al. 2008). It has been shown that the addition of algae may decrease the toxicity of silver ions and AgNPs up to ten times (Allen et al. 2010; Erickson et al. 1998; Nebeker et al. 1983). Koukal et al. (2007) observed that P. subcapitata exudates markedly decreased the toxicity of metals. Unrine et al. (2012) reported that the presence of the aquatic plants Potamogeton diversifolius and Egeria densa in the microcosms reduced the toxicities of both gum arabic-coated and PVP-coated AgNPs. This was suggested to owe to the changes in the surface chemistry of the particles upon the release of organic substances from the plants.

Thus, the presence of algae in the test media may have contradictory effects. On one hand, the algae may reduce the concentration of toxic silver ions/NPs in the test medium. On the other hand, silver adhered on ingested algae may increase the dietary intake of AgNPs by crustaceans. Zhao and Wang (2010) showed that more than 70 % of AgNP accumulated in the daphnids was through ingestion of silver sorbed to algae. The finding highlights the importance of AgNP transport along the food chain as the AgNPs could not be completely depurated from the daphnids.

It has been shown that the observed toxicity of AgNPs to bacteria (Fabrega et al. 2009), algae (Navarro et al. 2008), and Daphnia pulex (Griffitt et al. 2008) was the result of both Ag+ ions and particles of nanosilver. To reveal the role of the silver ions in the overall toxic effect of AgNPs, the release of metal ions from NPs must be quantified. However, currently, the appropriate analytical methods to detect and to quantify NPs in the complex matrix (e.g., natural water) are limited (Weinberg et al. 2011). Moreover, ingestion and excretion of NPs and adhesion of aggregates of NPs to the outer surface of test organisms (such as the exoskeleton of crustaceans) during the test may change the speciation of silver. Thus, complex chemical and biological processes that constantly modify Ag speciation during the test as well as different exposure pathways for tested silver to organisms (via food or from the test medium) make the analysis of the contribution of silver ions to the net toxicity of AgNPs very complicated.

The major finding of our research, based on the data on two crustacean species, is that there is apparently no reason to consider silver NPs more dangerous for aquatic ecosystems than silver ions. Therefore, the environmental risks of manufactured AgNPs assumingly do not exceed the risks related to environmental contamination by soluble Ag salts.

However, these conclusions are based on the results from conventional OECD test procedures and may need to be reconsidered in case of using more relevant endpoints for toxicity testing of nanomaterials containing silver.

Acknowledgments

This study was supported by the Estonian Science Foundation Projects ETF8066, ETF8561, and EU Central Baltic INTERREG IVA programme 2007–2013 project: Risk Management and Remediation of Chemical Accidents (RIMA). Julien Pinaton is acknowledged for participation in the synthesis of PVP-grafted silver nanoparticles and Prof. Damjana Drobne (University of Ljubljana) for the TEM image of collargol.

Supplementary material

11356_2012_1290_MOESM1_ESM.docx (1.4 mb)
ESM 1(DOCX 1,440 kb)

Copyright information

© Springer-Verlag Berlin Heidelberg 2012