Toxic Effects and Molecular Mechanism of Different Types of Silver

Oct 2, 2017 - Silver nanoparticles (AgNPs) have been assessed to have a high exposure risk for humans and aquatic organisms. Toxicity varies considera...
0 downloads 15 Views 2MB Size
Subscriber access provided by UNIV OF ESSEX

Article

Toxic effects and molecular mechanism of different types of silver nanoparticles to the aquatic crustacean Daphnia magna Jing Hou, yue zhou, chunjie wang, Shiguo Li, and Xiangke Wang Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b03918 • Publication Date (Web): 02 Oct 2017 Downloaded from http://pubs.acs.org on October 3, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32

Environmental Science & Technology

1

Toxic effects and molecular mechanism of different types of silver

2

nanoparticles to the aquatic crustacean Daphnia magna

3

Jing Hou a, Yue Zhou a, Chunjie Wang a, Shiguo Li b*, Xiangke Wang a*

4

a

5

Beijing 102206, China

6

b

7

China

8

* Address correspondence to [email protected]

College of Environmental Science and Engineering, North China Electric Power University,

Research Center for Eco-Environmental Sciences, Chinese Academy of Science, Beijing 100085,

9

ACS Paragon Plus Environment

Environmental Science & Technology

10

ABSTRACT

11

Silver nanoparticles (AgNPs) have been assessed to have a high exposure risk for humans and

12

aquatic organisms. Toxicity varies considerably between different types of AgNPs. This study

13

aimed to investigate the toxic effects of AgNPs with different particle sizes (40 nm and 110 nm)

14

and different surface coatings (sodium citrate and polyvinylpyrrolidone) on Daphnia magna and

15

their mechanisms of action. The results revealed that the citrate-coated AgNPs were more toxic

16

than PVP-coated AgNPs and that the 40-nm AgNPs were more toxic than the 110-nm AgNPs.

17

Transcriptome analysis further revealed that the toxic effects of AgNPs on D. magna were related

18

to the mechanisms of ion binding and several metabolic pathways such as “RNA polymerase”

19

pathway and “protein digestion and absorption” pathway. Moreover, the Principal component

20

analysis (PAC) results found that surface coating was the major factor that determines the

21

toxicities compared with particle size. These results could help us to better understand the possible

22

mechanism of AgNP toxicity in aquatic invertebrates at the transcriptome level and establish an

23

important foundation for revealing the broad impacts of nanoparticles on aquatic environments.

24

KEYWORDS: silver nanoparticle, Daphnia magna, RNA-Seq, aquatic toxicology, mechanism

ACS Paragon Plus Environment

Page 2 of 32

Page 3 of 32

Environmental Science & Technology

25

INTRODUCTION

26

With the development and commercialization of nanotechnology, silver nanoparticles (AgNPs)

27

have broad application prospects in chemical, optical, medical and other industries 1-3 due to their

28

advantages of small size, large specific surface area, high catalytic activity, low melting point,

29

good sintering properties, good electrical conductivity and good antibacterial properties.

30

catalyst, AgNPs can change the absorption and conversion of laser energy to the polymer system,

31

resulting in the change of laser dissociation. AgNPs doped in semiconductors or insulators can

32

obtain a large nonlinear polarizability, and can be used to make the color filters of optoelectronic

33

devices such as optical switches and advanced optical devices. In addition, AgNPs can be made

34

into medical dressings for most of the trauma treatment, to protect against microbial infection.7-9

35

Given these applications, they will eventually enter the freshwater environment such as rivers or

36

lakes and affect aquatic organisms. AgNPs have induced developmental, hepatic, endocrine and

37

reproductive toxicities in various aquatic organisms such as algae, daphnia, and fish.10-12 Once

38

accumulated in aquatic animals, they can enter the human body through the food chain. Therefore,

39

it is necessary to understand the potential impacts of AgNPs on aquatic invertebrates and facilitate

40

a better assessment of their impacts on human health.

41

4-6

As a

The physicochemical properties such as size,13 shape,14 surface coating,15 surface charge,16 and 17,18

42

solution chemistry

are important factors to determine AgNPs toxicity. Particle size and

43

surface coating have been considered to be the most determinant factors. For example, 20-nm

44

AgNPs induced more inflammation or cardiac ischemia/reperfusion injury than 110-nm AgNPs in

45

Danio Rerio.19 Locomotor activity experiment demonstrated that Gammarus roeseli exposed to

46

60-nm AgNPs exerted a significantly lower activity compared to those exposed to 30-nm

47

AgNPs.20 The reason why particle size is so determinant of biological toxicity may be dependent

48

on the higher dissolution rate of smaller particles and the pathway by which particles enter the cell

49

membrane. Kim et al. reported that NPs < 50 nm can more easily enter cells through endocytosis

50

than NPs > 50 nm.21 It was also demonstrated that 20-nm AgNPs had significantly greater effects

51

on tail moment, tail DNA intensity and burrowing activity of Nereis diversicolor than 80-nm

52

AgNPs.22 Some contrasting findings demonstrated that AgNPs with larger particle sizes were

53

more toxic.23 It is likely that particle size is not a dominant factor in toxicity. For example,

54

Jimenez-lamana et al. revealed that surface coating rather than size determines the stability of

ACS Paragon Plus Environment

Environmental Science & Technology

Page 4 of 32

55

AgNPs in aquatic ecosystems, with a stability order of PVP-coated AgNPs > citrate-coated

56

AgNPs > lipoic acid-coated AgNPs.24 Similarly, PVP-coated AgNPs were more stable than

57

citrate-coated AgNPs as a result of steric repulsion by the absorbed polymer layer.25 Gum

58

arabic-coated AgNPs has a greater impact on aquatic ecosystems than PVP-coated AgNPs.26 The

59

7-nm PVA-coated AgNPs exhibited higher toxicity to Nitrosomonas europaea than 7-nm

60

Na2ATP-coated AgNPs.27 Thus it can be seen that the influence of AgNPs with different

61

properties on aquatic organisms is complex. To fully understand the toxic effects of AgNPs on

62

aquatic organisms, the physiological and molecular mechanisms behind these complex effects

63

need to be revealed at different levels and perspectives. Although considerable progress has been

64

made towards elucidating the toxicity levels and biological effects, there is less information

65

regarding the molecular mechanisms of toxicity of AgNPs with both different particle sizes and

66

different surface coatings to aquatic organisms, especially aquatic invertebrates.

67

With the rapid development of high-throughput sequencing technologies, the possible

68

physiological functions in specific tissues and organs have explored fully at the transcriptomic

69

level.28-31 RNA sequencing (RNA-Seq) has proven to be a rapid and powerful technique with low

70

noise, high efficiency and massive data output enabling high-throughput sequencing of the whole

71

transcriptome.32 RNA-seq can aid in identifying biomarkers of exposure, differentiating toxicity

72

between AgNPs, detecting casual contaminants in environmental samples, and assessing the

73

mechanisms of response to stressors.33-34 Although RNA-seq has been performed in aquatic

74

species such as Pandalus latirostris,35 Salmo salar,36 Oryzias melastigma,37 Labeo rohita,38

75

Lateolabrax japonicas,39 Litopenaeus vannamei,40 Ostrea edulis,41 Pinctada fucata42 and Patella

76

vulgate,43 the molecular mechanisms underlying the effects of different types of AgNPs on aquatic

77

organisms are still rare.

78

The objectives of the present study were to investigate toxic effects of different types of AgNPs

79

on an aquatic invertebrate D. magna and further explore their mechanisms of action. D. magna is

80

widely used as a model organism to evaluate the invertebrate response to toxicants due to its rapid

81

reproduction, short lifespan, and high sensitivity to environmental pollutants.44-46 To achieve the

82

above objectives, a RNA-seq technology was employed to investigate the toxic effects and

83

molecular

84

110-PVP-AgNP) on D. magna, which will establish an important foundation for revealing the

mechanism

of

AgNPs

(40-cit-AgNP,

110-cit-AgNP,

ACS Paragon Plus Environment

40-PVP-AgNP

and

Page 5 of 32

Environmental Science & Technology

85

toxicity and toxic mechanism of AgNPs in aquatic invertebrates.

86

MATERIALS AND METHODS

87

Daphnia magna Cultivation. D. magna obtained from the Chinese Research Academy of

88

Environmental Sciences were cultured in an illumination incubator with a photoperiod of 16 h/8 h

89

(light/dark) and a temperature of 21±1 °C. D. magna were maintained in filtered and sterilized tap

90

water47 and fed Scenedesmus obliquus at approximately 0.2 mg of carbon per individual per day.

91

The medium was renewed three times a week and aerated overnight before use. S. obliquus

92

obtained from the Institute of Hydrobiology, Chinese Academy of Sciences were grown in Bold's

93

Basal Medium (BBM) with and initial pH of 8.0. Table S1 summarizes the chemical composition

94

of BBM. The algae were cultured at 25±1 °C with a light intensity of 2500 lx and a photoperiod of

95

14 h/10 h (light/dark). The algal cell density was determined using a hemocytometer under a

96

trinocular microscope (B204TR, Chongqing, China) at a magnification of 400×.

97

Silver Nanoparticle Characterization. Citrate-coated AgNPs (sodium citrate as a stabilizer)

98

with average particle sizes of 40 nm and 110 nm (40-cit-AgNP and 110-cit-AgNP) and

99

PVP-coated AgNPs (0.3% polyvinylpyrrolidone) with average particle sizes of 40 nm and 110 nm

100

(40-PVP-AgNP and 110-PVP-AgNP) were obtained from nanoComposix (San Diego, USA).

101

Transmission electron microscopy (TEM) analysis was used to characterize the size and shape of

102

the AgNPs after 0 h and 24 h of dispersion in the test medium by placing a droplet of the AgNPs

103

on a TEM grid followed by obtaining images using a TEM (Hitachi H-7650B, Japan) at an

104

accelerating voltage of 120 kV. The hydrodynamic size and zeta potential of the AgNPs dispersed

105

in the test medium were measured at 0 h and 24 h using a Zetasizer (Nano ZS, Malvern

106

Instruments, UK). Each time point consisted of three replicates, and five consecutive

107

measurements were made per replicate. Dissolution of the AgNPs was determined by separating

108

dissolved silver from the AgNPs using centrifugal ultrafiltration with a 30-kDa filter (Millipore

109

Amicon Ultra, USA).48 The dispersions were centrifuged at 3000 g for 60 min (Centrifuge 5424,

110

Eppendorf, Germany), and the dissolved silver present in the filtrate was acidified and analyzed in

111

triplicate at 6, 12, 18 and 24 h by ICP-AES (SPECTRO Analytical Instruments GmbH ,

112

Germany).

113

Acute Toxicity Test. D. magna toxicity tests were slightly modified from the U.S. EPA Whole

114

Effluent Toxicity guidelines.49 The 14-day-old D. magna were exposed to six different

ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 32

115

concentrations (0, 2, 4, 8, 16, 32, and 64 µg/L) of AgNPs. Each concentration consisted of thirty D.

116

magna with in each of six replicates. D. magna were incubated in incubator without feeding

117

during the 24 h exposure period and examined for lethality after 24 h. The probit analysis method

118

was employed to estimate the median lethal concentration (LC50) values. All subsequent assays

119

were performed using 2 µg/L (close to 1/10 LC50) of 40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP

120

and 110-PVP-AgNP, because this level was below that which causes nonspecific gene expression

121

responses.50 At predetermined intervals (6, 12, 18, and 24 h), Thirty D. magna was sampled from

122

each treatment and sacrificed to measure silver concentration in the body. Three biological

123

replicates were conducted for each treatment.

124

RNA Sequencing. Exposures were performed on D. magna using 2 µg/L (close to 1/10 LC50)

125

of 40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP and 110-PVP-AgNP for 24 h. Thirty 14-d-old D.

126

magna were pooled together to generate one sample. Three biological replicates were conducted

127

for each treatment. The detailed RNA-Seq method, including total RNA extraction, mRNA

128

enrichment, mRNA fragment, cDNA synthesis, library construction and Illumina sequencing, is

129

shown in the supporting information. RNA-seq library sequencing was carried out on an Illumina

130

HiSeq 4000 instrument with 50 bp based on the cDNA library of D. magna. If the result passed

131

quality control, the NOISeq method was used to screen for differentially expressed genes (DEGs)

132

according to the following criteria: fold change > 1.2 and p-value > 0.5. The website

133

(http://bioinfogp.cnb.csic.es/tools/venny/index.html) was used to generate the Venn diagram.

134

Gene Ontology Enrichment Analysis. GO enrichment analysis provided all GO terms that

135

were significantly enriched in a list of DEGs. All DEGs were mapped to GO terms and gene

136

numbers were calculated for every term in the database (http://www.geneontology.org/).

137

Significantly enriched GO terms for the input list of DEGs were calculated using the

138

hypergeometric

139

(http://search.cpan.org/~sherlock/GO-TermFinder/).

test

with

p-value

correction

based

on

'GO::TermFinder'

140

Kyoto Encyclopedia of Genes and Genomes Pathway Analysis. The KEGG was used to

141

identify the biological pathways of DEGs that were affected by AgNPs via comparison to the

142

background genome, which helped to further predict the functions of these DEGs. A KEGG

143

enrichment analysis (hypergeometric test, q-value < 0.05) was performed on all DEGs using the

144

KEGG database (http://www.genome.jp/kegg/).

ACS Paragon Plus Environment

Page 7 of 32

Environmental Science & Technology

145

Hierarchical Clustering Analysis (HCA) and Principal Component Analysis (PCA). HCA

146

is a chemometric tool used to calculate the distance between samples via the Sum of Squares of

147

Deviations algorithm. Genes with similar expression values are clustered at both the row and

148

column level. Only genes that were expressed in all samples of the cluster plan were used to build

149

the heatmap. The heatmap was produced with cluster and javaTreeview software. Expression

150

values for each transcript were log2 transformed before cluster formation. The heatmap was

151

clustered using the complete linkage hierarchical analysis based on Euclidean distance. For the

152

PCA, all data were first unit-variance scaled and mean centered. A two-component model was

153

selected for the distance measurement, which scattered in oval shapes when the datasets were

154

strongly related.

155

Real-time Quantitative PCR. RT-qPCR was performed to validate the reliability of the

156

RNA-seq data. A total of eight DEGs with various expression patterns, including two DEGs

157

involved in 40-cit-AgNP, two DEGs involved in 110-cit-AgNP, two DEGs involved in

158

40-PVP-AgNP, and two DEGs involved in 110-PVP-AgNP, were selected for validation. Total

159

RNA was extracted and purified as described above. The first-strand cDNA was synthesized using

160

the Superscript III First-Strand Synthesis kit (Invitrogen, Carlsbad, CA) following the

161

manufacturer’s instructions. RT-qPCR was performed using SYBR Premix Ex TaqII (Takara,

162

Japan) on a 7500 Fast Real-Time PCR system (Applied Biosystems, Foster, CA). The cycling

163

conditions were as follows: one cycle of 50 °C for 5 min and 95 °C for 2 min for initial

164

denaturation, and 45 cycles of 95 °C for 15 s and 60 °C for 30 s. Each sample was run in triplicate.

165

DmACTIN was used as a reference gene, and primers were designed with Primer-BLAST

166

(www.ncbi.nlm.nih.gov/tools/primer-blast/). The relative quantitative method (2-∆∆CT) was used to

167

calculate the fold change of the selected genes. Data are expressed as the mean ± standard error

168

from samples performed in triplicate.

169

RESULTS

170

Physicochemical Characterization of AgNPs. Figure 1 displays the size, shape, and

171

agglomeration state of the four types of AgNPs after 0 h and 24 h of dispersion in the D. magna

172

test medium. The representative TEM images demonstrated that the primary AgNPs were

173

spherical in shape and uniform in size, confirming the manufacturer’s description of 40 nm or 110

174

nm for all four particles (Figure 1a-d). After 24 h of dispersion, the 40-cit-AgNP aggregated

ACS Paragon Plus Environment

Environmental Science & Technology

175

significantly into irregular shapes (Figure 1e). The 110-cit-AgNPs were discrete without

176

aggregation, although some small spherical particles were obviously dispersed around each

177

nanoparticle (Figure 1f). Both 40-PVP-AgNP and 110-PVP-AgNP showed a monodisperse

178

nanoparticle size distribution and spherical form without clusters or aggregates, indicating greater

179

stabilization of the PVP-coated AgNPs (Figure 1g-h).

180

The TEM results were also confirmed by the hydrodynamic sizes and zeta potentials (Table 1).

181

The hydrodynamic sizes of 40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP, and 110-PVP-AgNP in

182

the D. magna test medium were found to change from 43.2 ± 0.3 nm, 94.3 ± 0.6 nm, 45.2 ± 0.2

183

nm, and 89.6 ± 0.9 nm after 0 h of exposure to 209.7 ± 26.5 nm, 89.4 ± 5.6 nm, 38.7 ± 3.1 nm, and

184

86.1 ± 0.7 nm after 24 h of exposure, respectively. Zeta potentials were used to characterize the

185

stability of the dispersed system. The absolute value of the zeta potentials for 40-cit-AgNP and

186

110-cit-AgNP decreased from -31.3 ± 3.9 and -36.7 ± 2.8 after 0 h of exposure to -11.3 ± 3.2 and

187

-18.6 ± 2.2 after 24 h of exposure, respectively, suggesting a significant decrease in the AgNP

188

stability. 40-PVP-AgNP and 110-PVP-AgNP had zeta potentials of -28.3 ± 2.4 and -31.6 ± 2.6

189

after 24 h of exposure, indicating good stability of the PVP-coated AgNPs.

190

Acute Toxicity. The 24-h LC50 values for 40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP, and

191

110-PVP-AgNP were 8.90 µg/L, 17.43 µg/L, 24.97 µg/L, and 38.35 µg/L, respectively (Figure S1).

192

The rank order of toxicity to D. magna based on the LC50 values was 40-cit-AgNP >

193

110-cit-AgNP > 40-PVP-AgNP > 110-PVP-AgNP. Two trends related to toxicity were

194

summarized: citrate-coated AgNPs were more toxic than PVP-coated AgNPs, and 40-nm AgNPs

195

were more toxic than 110-nm AgNPs.

196

Dissolution of Silver Nanoparticles. The particle dissolution rates were determined because

197

particle dissolution plays a key role in the toxicity of AgNPs to aquatic organisms. All four

198

particles demonstrated a significant increase in Ag shedding after 6, 12, 18, and 24 h of incubation

199

in the D. magna test medium (Figure 2). The dissolution rates within 24 h followed the sequence

200

40-cit-AgNP > 110-cit-AgNP > 40-PVP-AgNP > 110-PVP-AgNP. The citrate-coated AgNPs

201

exhibited a higher level of dissolution than the PVP-coated AgNPs, and the 40-nm AgNPs were

202

more rapid than the 110-nm AgNPs for the same type of surface coating with the exception of

203

40-PVP-AgNP at 6 h and 40-cit-AgNP at 24 h.

204

Uptake of Silver Nanoparticles. As shown in Figure 3, the uptake of the four types of AgNPs

ACS Paragon Plus Environment

Page 8 of 32

Page 9 of 32

Environmental Science & Technology

205

showed a general increase during the 24 h. The uptake of 40-nm AgNPs by D. magna is greater

206

than that of 110-nm AgNPs at the same time point, while there was no significant difference in the

207

uptake between citrate-coated AgNPs and PVP-coated AgNPs. The difference in body burden

208

during this period could be explained by particle size. The body burden values of 40-cit-AgNP at

209

6, 12, 18, and 24 h increased by 116%, 138%, 108% and 122% compared with 110-cit-AgNP. The

210

body burden values of 40-PVP-AgNP at 6, 12, 18, and 24 h increased by 172%, 224%, 66% and

211

141% compared with 110-PVP-AgNP.

212

Analysis of Differentially Expressed Genes. To discriminate the toxic effects induced by the

213

different types of AgNPs, a total of 1262, 974, 664, and 594 DEGs for 40-cit-AgNP,

214

110-cit-AgNP, 40-PVP-AgNP, and 110-PVP-AgNP were identified using a cutoff of p-value > 0.5

215

and fold change > 1.2 (Table 2). The lists of DEGs for each exposure are presented in File S1. The

216

gene expression data were correlated with the acute toxicity data. The 40-cit-AgNP caused the

217

most abundant candidate genes, followed by 110-cit-AgNP and then 40-PVP-AgNP;

218

110-PVP-AgNP resulted in the weakest response. The number of down-regulated genes exceeded

219

the number of up-regulated genes for all four exposures. The distributions and possible relations

220

of DEGs between all treatments are demonstrated in a Venn diagram (Figure 4). Venn diagram

221

analysis revealed that 120 (5.2%) DEGs were commonly altered by all four exposures. There were

222

729 (31.6%), 444 (19.2%), 147 (6.4%), and 214 (9.3%) DEGs for 40-cit-AgNP, 110-cit-AgNP,

223

40-PVP-AgNP, and 110-PVP-AgNP, respectively, that were altered by only one of the treatments

224

and not by any other treatments. The number of DEGs altered by citrate-coated AgNPs (319 genes,

225

13.8%) was greater than the number of altered by PVP-coated AgNPs (223 genes, 9.6%).

226

Similarly, the number of DEGs altered by 40-nm AgNPs (312 genes, 13.5%) was greater than the

227

number altered by 110-nm AgNPs (213 genes, 9.2%). Sequencing data have been submitted to the

228

Sequence Read Archive (SRA, https://submit.ncbi.nlm.nih.gov/subs/sra/) under the accession

229

number PRJNA389206.

230

GO Enrichment Analysis of DEGs. To further study the functions of the DEGs, they are

231

subjected to gene ontology (GO) functional enrichment analysis, which is a gene-function

232

classification system used to describe the properties of genes and their products according to three

233

characteristics: cellular components, molecular function, and biological processes. The GO

234

enrichment analysis results indicated that 29, 15, 8, and 24 GO terms were significantly enriched

ACS Paragon Plus Environment

Environmental Science & Technology

235

in the D. magna exposed to 40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP, and 110-PVP-AgNP,

236

respectively (File S2). The GO terms that existed in at least two exposure conditions are

237

summarized in Figure 5. Most of the GO terms are related to ion binding. Transition metal ion

238

binding (GO:0046914), ion binding (GO:0043167), and iron ion binding (GO:0005506) are the

239

common GO terms for three of the exposure conditions. Cation binding (GO:0043169) is the only

240

common GO term for the 40-nm AgNPs (40-cit-AgNP and 40-PVP-AgNP). Protein complex

241

(GO:0043234), transporter activity (GO:0005215), substrate-specific transporter activity

242

(GO:0022892), gas transport (GO:0015669), single-organism transport (GO:0044765), and

243

single-organism localization (GO:1902578) are common GO terms for the 110-nm AgNPs

244

(110-cit-AgNP and 110-PVP-AgNP). Notably, gas transport (GO:0015669) is the most

245

statistically significant term for both 110-cit-AgNP (p = 7.61×10-9) and 110-PVP-AgNP (p =

246

3.05×10-8).

247

KEGG Pathway Analysis of DEGs. KEGG analysis was further performed to identify the

248

enrichment pathways in D. magna for each treatment. DEGs for 40-cit-AgNP, 110-cit-AgNP,

249

40-PVP-AgNP, and 110-PVP-AgNP were found to be enriched in 252, 225, 184, and, 192 KEGG

250

pathways, respectively (File S3). Taking a q-value < 0.05 as a threshold, the most represented

251

pathways are summarized in Table 3. Hippo signaling (q = 0.0211) is the most statistically

252

significant pathway for 40-cit-AgNP. RNA polymerase (q = 8.55×10-30) is the most statistically

253

significant pathway for 110-cit-AgNP. Pyrimidine metabolism (q = 1.60×10-5) and lysosom (q =

254

8.62×10-5) are the most statistically significant pathways for 40-PVP-AgNP. Protein digestion and

255

absorption (q = 7.64×10-5) is the most statistically significant pathway for 110-PVP-AgNP. RNA

256

polymerase is the only in-common pathway for the citrate-coated AgNPs (40-cit-AgNP and

257

110-cit-AgNP). Protein digestion and absorption is the only in-common pathway for the

258

PVP-coated AgNPs (40-PVP-AgNP and 110-PVP-AgNP).

259

Hierarchical Clustering Analysis. HCA was performed to obtain more information about the

260

similarities or differences between groups and between samples by calculating the distance

261

between them. As shown in Figure S2, the dendrogram presented in the top axis was generated by

262

identifying relatively homogenous clusters of samples based on their similarities. Each row

263

represents a single gene and each column represents an experimental sample. Different colors

264

indicate different gene expression levels. As seen in the dendrogram, the expression profiles

ACS Paragon Plus Environment

Page 10 of 32

Page 11 of 32

Environmental Science & Technology

265

induced by the fifteen samples were clustered into two distinct clusters. The left cluster can be

266

further divided into two subclusters. One is comprised of two 110-PVP-AgNP samples. The other

267

is comprised of one 110-PVP-AgNP sample and three control samples. The right cluster is formed

268

by the 40-cit-AgNP group, 40-PVP-AgNP group, and 110-cit-AgNP group. Three replicate

269

samples are clustered in one group. Each group is independent of the others and has no overlap.

270

The 40-cit-AgNP group is clearly discernible from the rest of groups and forms a subcluster. The

271

110-cit-AgNP and 40-PVP-AgNP groups are mixed to form another subcluster. These results

272

indicate that most of the treatments exhibited a good reproducibility, and the different types of

273

AgNPs can be discriminated from each other at relatively low concentrations.

274

Principal Component Analysis. PCA is commonly used to obtain more information about the

275

main variables and classify different samples by reducing redundant information in the data. As

276

illustrated in Figure 6, dots with the same color represent the same treatment. Each treatment

277

consists of three replicates. The PCA result is in accord with the HCA result. Dots with the same

278

color are relatively close to each other and far from the other dots, indicating a closer relationship

279

among the replicate samples, thus further verifying the reproducibility and reliability of the

280

RNA-seq data. The fifteen sample dots are classified into five groups. The principal component 1

281

values for all of the samples are negative. Note that principal component 2 separates 40-cit-AgNP

282

and 110-cit-AgNP (positive scores) from 40-PVP-AgNP and 110-PVP-AgNP (negative scores),

283

the former being citrate-coated AgNPs and the latter being PVP-coated AgNPs. Furthermore,

284

40-PVP-AgNP is split between positive and negative scores (close to zero) for principal

285

component 2, indicating that the difference is mainly between the group formed by 40-cit-AgNP

286

and 110-cit-AgNP and the group formed by 110-PVP-AgNP and the control. Therefore, surface

287

coating may be the major factor that determines the toxicity of AgNPs to D. magna compared

288

with particle size.

289

Validation of RNA-Seq. To validate the reliability of the data obtained from RNA-Seq, a total

290

of eight DEGs, including four up-regulated DEGs and four down-regulated DEGs from the

291

40-cit-AgNP, 110-cit-AgNP, 40-PVP-AgNP and 110-PVP-AgNP treatments, were selected and

292

validated by RT-qPCR. Detailed information about the primers including gene ID, gene symbol,

293

gene description, product size and primer sequences is provided in Table S2. As shown in Figure

294

S3, the RT-qPCR results for the eight DEGs show a high degree of correlation with the trend

ACS Paragon Plus Environment

Environmental Science & Technology

295

revealed by RNA-Seq. A positive correlation exists (R2=0.933) between the expression level of

296

DEGs from RNA-Seq and RT-qPCR, confirming the high reproducibility and reliability of the

297

RNA-Seq data.

298

DISCUSSION

299

Coating-specific and size-dependent toxicities of AgNPs were demonstrated in the present study.

300

The LC50 data showed that citrate-coated AgNPs were more toxic than PVP-coated AgNPs and

301

that 40-nm AgNPs were more toxic than 110-nm AgNPs. Similar results were obtained from

302

previous studies. Poynton et al. found that the LC50 values of AgNPs in 10-day-old D. magna

303

were 1.8 µg/L for citrate-coated AgNPs and 10.6 µg/L for PVP-coated AgNPs.51 The acute

304

toxicity of AgNPs in D. magna demonstrated that citrate-coated AgNPs were less toxic than

305

uncoated AgNPs, and the larger citrate-coated AgNPs (60 nm and 100 nm) exhibited

306

approximately 7.5-fold lower toxicity relative to the smaller citrate-coated AgNPs (20 nm).52 Our

307

results were also in agreement with a study on Caenorhabditis elegans, where it was demonstrated

308

that the PVP coating reduced the toxicity of the AgNPs significantly and that 8 nm PVP-coated

309

AgNPs were more toxic than 38-nm PVP-coated AgNPs.53

310

Moreover, the toxicity of the AgNPs is directly related to the aggregation of their particles.

311

Aggregation can reduce toxicity.54,55 The aggregated AgNPs formed larger clusters that reduced

312

the uptake by cells, whereas the dispersed AgNPs more easily crossed cell membranes and

313

inhibited cell growth.23 However, in our study, 40-cit-AgNPs were more toxic than 40-PVP-

314

AgNPs even though 40-cit-AgNPs tended to aggregate more. Because the toxicity of AgNPs is not

315

determined by a single factor, a number of factors act in combination to mediate the toxicity of

316

AgNPs. This is in accordance with previous investigation which has confirmed that SO2-coated

317

silver nanowires (AgNWs) were more toxic to D. magna than PVP-coated AgNWs, but the

318

SO2-coated AgNWs aggregated and settled more.

319

Most of the previous studies showed that AgNP toxicities were caused by the release of silver

320

ions, whether they are different coatings or particle sizes. The dissolution rates were

321

time-dependent and varied considerably between the different types of AgNPs. Citrate-coated

322

AgNPs were more reactive than PVP-coated AgNPs. It might be that the stability of the PVP

323

coating enabled the silver to be fully encapsulated.56 The stability of the PVP-coated AgNPs was

324

also confirmed by the zeta potentials. The absolute values of the zeta potentials for the

ACS Paragon Plus Environment

Page 12 of 32

Page 13 of 32

Environmental Science & Technology

325

citrate-coated AgNPs decreased significantly after 24 h of exposure, suggesting a significant

326

decrease in the AgNP stability and explaining the observed aggregation; in contrast, no obvious

327

changes were observed for the PVP-coated AgNPs, indicating good stability. This result is

328

consistent with previous studies in which citrate-coated AgNPs had a higher dissolution rate than

329

PVP-coated AgNPs.57 The PVP-coated AgNPs were more stable than the citrate-coated AgNPs as

330

a result of steric repulsion by the absorbed polymer layer.25 These findings indicated that the

331

surface coating was not an impervious barrier and that Silver ion can be released from AgNPs and

332

induce toxicity. Meanwhile, there was a significant difference in the toxicity between 40-nm

333

AgNPs and 110-nm AgNPs. The 40-nm AgNPs had higher dissolution rates than 110-nm AgNPs

334

for the same types of surface coating. This may be caused by differences in the release of silver

335

ions.

336

110-nm citrate-coated AgNPs,58 and in which 7-nm AgNPs showed higher dissolution rates than

337

40-nm AgNPs due to the higher specific surface areas and higher enthalpies of formation of the

338

smaller particles.59 Silver ion can be also released from AgNPs in gut or fluids, which is the main

339

mechanism of the toxic effects of AgNPs and AgNPs sterilization. AgNP can be ionized in body

340

fluids or gut, and their interaction with H2O2 is considered to be one of the factors of the release of

341

Ag+ in gut, From the mechanism, the surface of AgNPs can quickly release free Ag+ after

342

oxidation. Therefore,the toxic effects could be also explained by the body burden values. The

343

uptake of 40-nm AgNPs by D. magna is greater than that of 110-nm AgNPs at the same time point.

344

Body burden of the four types of AgNPs showed a general increase during the first 24 h. The

345

similar results were also concluded from the uptake behaviors of carbon nanotubes by Eisenia

346

foetida.60 The uptake of graphene by D. magna across the range of concentrations (25, 50, 100 and

347

250 µg/L) increased during the first 24 h.61

For

example,

20-nm citrate-coated

AgNPs

released

silver

ions

faster

than

348

The toxicity effects of different AgNPs to D. magna were further confirmed by comparative

349

transcriptome analysis. The number of DEGs in D. magna that were altered by citrate-coated

350

AgNPs was significantly greater than that by PVP-coated AgNPs, and the number of DEGs in D.

351

magna that were altered by 40-nm AgNPs was significantly greater than the number altered by

352

110-nm AgNPs, supporting the coating-specific and size-dependent toxicity of AgNPs to D.

353

magna. GO enrichment analysis of the DEGs indicated that the number of GO terms in the

354

citrate-coated AgNPs was greater than the number in the PVP-coated AgNPs, which was

ACS Paragon Plus Environment

Environmental Science & Technology

355

consistent with the results of silver ion release. However, 40-nm was not greater than 110-nm for

356

the PVP-coated AgNPs, which may be caused by the large number of unannotated genes in the D.

357

magna genome database. Note that most of the DEGs in D. magna that were induced by AgNPs

358

were enriched in ion binding-related terms, including “transition metal ion binding”, “ion binding”,

359

“iron ion binding” and “cation binding”, which is consistent with the release trend of silver ion.

360

Ion binding is crucial for numerous biological processes, particularly protein activities involved in

361

these processes. The binding processes between metal ions and active sites of key enzymes in

362

vertebrates are regarded as the basis for enzyme activation. Approximately 1/3 of proteins in

363

vertebrates bind at least one metal ion, including potassium, calcium, sodium, magnesium, copper,

364

iron, zinc and manganese ions.62,

365

proteins upon binding, which stabilize the protein structure and participate in enzyme catalysis.64

366

In this study, the GO results were consistent with earlier studies on aquatic animals. For example,

367

the metal ion binding related genes of zebrafish embryos were significantly influenced by

368

AgNPs.65 D. magna that were exposed to CuO NPs for 96 h (