Subscriber access provided by READING UNIV
Article
CuO nanoparticle dissolution and toxicity to wheat (Triticum aestivum) in rhizosphere soil Xiaoyu Gao, Astrid Avellan, Stephanie N Laughton, Rucha Vaidya, Sónia Morais Rodrigues, Elizabeth A. Casman, and Gregory V. Lowry Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b05816 • Publication Date (Web): 31 Jan 2018 Downloaded from http://pubs.acs.org on February 2, 2018
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 24
Environmental Science & Technology
1
CuO nanoparticle dissolution and toxicity to wheat (Triticum aestivum) in rhizosphere soil
2 3
Xiaoyu Gao†, §, Astrid Avellan †, §, Stephanie Laughton †, §, Rucha Vaidya †, §, Sónia M. Rodrigues‡, Elizabeth A. Casman§, #, and Gregory V. Lowry†, §, *.
4 5
†
6
§
7
States
8
‡
9
#
Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States Center for Environmental Implications of NanoTechnology (CEINT), Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United
Centre for Environmental and Marine Studies (CESAM), Department of Chemistry, Universidade de Aveiro, 3810-193 Aveiro, Portugal Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States
10
*Address correspondence to
[email protected] 11
Abstract:
12
It has been suggested, but not previously measured, that dissolution kinetics of soluble nanoparticles such
13
as CuO NPs in soil affect their phytotoxicity. An added complexity is that such dissolution is also
14
affected by the presence of plant roots. Here, we measured the rate of dissolution of CuO NPs in bulk soil,
15
and in soil in which wheat plants (Triticum aestivum) were grown under two soil NP dosing conditions:
16
(a) freshly added CuO NPs (500 mg Cu/kg soil), and (b) CuO NPs aged for 28d before planting. At the
17
end of the plant growth period (14 days), available Cu was measured in three different soil compartments:
18
bulk (not associated with roots), loosely attached to roots, and rhizosphere (soil firmly attached to roots).
19
The labile Cu fraction increased from 17mg/kg to 223mg/kg in fresh treatments and from 283 mg/kg to
20
305mg/kg in aged treatments over the growth period due to dissolution. Aging CuO NPs increased the
21
toxicity to Triticum aestivum (reduction in root maximal length). The presence of roots in the soil had
22
opposite and somewhat compensatory effects on NP dissolution, as measured in rhizosphere soil. pH
23
increased 0.4 pH units for fresh NP treatments and 0.6 pH units for aged NPs. This lowered CuO NP
24
dissolution in rhizosphere soil. Exudates from T. aestivum roots also increased soluble Cu in porewater.
25
CaCl2 extractable Cu concentrations in bulk vs. rhizosphere soil increased from 1.8mg/kg to 6.2mg/kg
26
(fresh treatment), and from 3.4mg/kg to 5.4mg/kg (aged treatments). Our study correlated CuO NP
27
dissolution and the resulting Cu ion exposure profile to phytotoxicity, and showed that plant-induced
28
changes in rhizosphere conditions should be considered when measuring the dissolution of CuO NP near
29
roots.
30
TOC art:
1 ACS Paragon Plus Environment
Environmental Science & Technology
31 32 33
Introduction
34
The anticipated benefits of nano-enabled agrochemicals include slow and controlled release of
35
micronutrients, plant tissue-specific targeted release of micronutrients or pesticides, reduced amounts of
36
agrochemicals being required, and generally lower toxicity compared to more soluble products1,2. Copper-
37
based nanoparticles (NPs) are already on the agrochemical market3,4. Copper is an essential crop
38
micronutrient. Deficiency may lead to reduced disease resistance5 and decreased crop yields6. However,
39
at high concentrations, Cu can also be toxic to plants,7 the surrounding microbial communities,8 and soil
40
invertebrates9. Due to its relatively slow dissolution, CuO NPs have been studied as a potential candidate
41
for agrochemical use. It behaves differently from dissolved Cu2+ in soil, potentially affecting copper
42
bioavailability, the release of Cu ions over time, and potential associated risks10–12. However, the
43
connection between NP dissolution, the resulting dose of Cu ions and its toxicity to terrestrial plants, and
44
the role of root exudates on this process have not been well elucidated due to a lack of appropriate
45
characterization of the dissolution of the NPs in soil. Ideally, application rates of these novel materials
46
should be based on their fate and effects in the terrestrial environment, their bioavailability and potential
47
toxicity to plants. The toxic effect of Cu species is reflected in physiological changes in plant roots and
48
shoots, such as decreased root length, increased root compactness, change in root color, shorter leaf
49
length and decreased shoot biomass13–15. Hyperspectral imaging has been used to visualize NPs in plants
50
and to confirm macroscopic evidence of NP toxicity16,17.
51
Previous studies of the toxicity of CuO NPs to terrestrial plants assumed, but did not measure,
52
dissolution behavior of CuO NP in soil. This has led to conflicting conclusions on the toxicity of CuO
53
NPs. While some studies attributed the toxic effect of CuO NP to released ionic Cu15,18,19, others
54
concluded the opposite20. For example, Servin et al. chose a Cu ion control concentration based on the
55
assumption that only 10% of the CuO NP would dissolve in soil, the same fraction that dissolved in pure
56
sand, rather than measuring CuO dissolution in soil. They concluded that dissolution of CuO NPs could
57
not fully explain the plant toxicity because the plant responses differed from their Cu ion control .20 Much
58
more than 10% CuO NP could have dissolved in soil because soil organic matter (SOM) acts as a Cu 2 ACS Paragon Plus Environment
Page 2 of 24
Page 3 of 24
Environmental Science & Technology
59
sink, increasing the amount of CuO NP that can be dissolved11. This weakens their conclusions about a
60
NP-specific effect. Similar problems occurred in other studies 15,18,21–23. Breaking with this trend, Dimkpa
61
et al. (2013) evaluated the total CuO NP dissolved in soil using a water-extraction method. 24
62
Unfortunately, the water-extraction does not extract Cu bound to the soil solid matrix which accounts for
63
most of the dissolved Cu in soils25–27. Thus, their assertion of a CuO NP-specific toxicity in soil is
64
confounded by the potential that more Cu had dissolved than was assumed or measured. Recently Qiu et
65
al. found that the toxicity of CuO NP, CuO bulk particles and soluble Cu (Cu(AC)2) depends on their
66
solubility in soil, and that the distinction in solubility diminished after a 90-day aging period. However,
67
the actual dissolution during the incubation periods (one day vs. 90 day) was not quantified. They
68
successfully correlated the toxicity of NP to roots of Hordeum vulgare L. (5-day root elongation
69
experiment) with ‘free Cu ions’ in soil pore water measured at a single time point before seeding;19
70
though convincing, it should be noted that the dissolution during the 5-day toxicity test was not
71
considered. While the relatively slow dissolution of CuO NP may result in unobservable impacts on
72
toxicity during a relatively short 5-day toxicity test, dissolution at this rate would probably affect toxicity
73
of NPs in longer tests.
74
The dissolution of CuO NPs is a dissolution rate-limited process. Experimental approaches, such
75
as extraction with CaCl2 or with diethylenetriaminepentaacetic acid (DTPA), have been used to predict
76
the bioavailability or toxicity of metals in soil.28–31 CaCl2 extracts the Cu ions in soil pore water that are
77
considered ‘readily available,’ while DTPA extracts the “labile” fraction including dissolved Cu in soil
78
pore water (free Cu2+and Cu2+ complexed with soluble ligands such as dissolved organic matter (DOM)),
79
but also the Cu2+ associated with soil solid phases, such as soil organic matter (SOM), clay particles, and
80
iron oxide minerals.29–31 Whereas CaCl2 extracts metals that are ‘readily available’ to plants29, DTPA
81
extracts this pool as well as the pool that may eventually become bioavailable in soil, the so called
82
‘potentially available’ fraction32. One problem with using these extraction methods to predict the
83
bioavailability of Cu based nanomaterials is that a single time point extraction does not capture the
84
temporal dynamics of the CuO NP dissolution process. Our recent study used extraction methods at
85
different times to monitor the kinetics of release of Cu ions from CuO NP in soil. In that study, the
86
increase in DTPA extractable Cu over 30 days in soils was used to estimate the dissolved pool of Cu in
87
soil.11 The availability of Cu ions increased with time over a 30d period, which may explain why previous
88
efforts to correlate the extractable metals in metal-based NP-amended soils with their bioavailability or
89
toxicity have generally failed33–35.
90 91
Plants also may affect the dissolution behavior and availability of CuO NP in soil, especially in the rhizosphere. Previous studies using extraction methods to predict the bioavailability or toxicity of 3 ACS Paragon Plus Environment
Environmental Science & Technology
92
metal-based ENMs or the dissolution of ENMs in soil did not typically consider the impact of roots on Cu
93
availability.11,12,33–35 Plant roots exude organic acids 36–38 that may affect the pH in rhizosphere. 39,40
94
Although soil pH and organic carbon are known to be important factors influencing the dissolution
95
behavior of CuO NPs in soil11,12, and previous studies have proposed that exudates from plant roots may
96
affect the dissolution of CuO NP in the rhizosphere41, no studies have quantified this. Given that the
97
rhizosphere is where plants interact with soil for nutrient uptake,42,43 a better understanding of how the
98
roots impact NP dissolution and metal availability in the rhizosphere is needed to design nano-enabled
99
agrichemicals with optimal properties for delivering nutrients.
100
The objectives of this study are to quantify the influence of time and near-root chemical
101
conditions on dissolution and lability of CuO NPs in rhizosphere soil, and to determine the influence of
102
this dissolution on the toxicity of CuO NPs to Triticum aestivum during a 14-day plant growth period in
103
soil. Wheat (Triticum aestivum) was used in this study because it is the 2nd most cultivated plant in the
104
world, and it is sensitive to Cu deficiency44 or excess45. To evaluate the toxicity of CuO NP to plants, we
105
measured the dissolution behavior of CuO NPs in soil in the presence of plants with emphasis on the soil-
106
plant interface (rhizosphere) where roots interact with soil. The toxicity of Cu was evaluated by
107
physiological changes in plant roots and shoots.
108
Methods and materials
109
Chemicals. Calcium chloride (≥99.0%, ACS grade) and hydrogen peroxide (30%, certified ACS) were
110
purchased from Fisher Scientific. DTPA (>99%) and triethanolamine (TEA, ≥99.0% (GC)) were
111
purchased from Sigma-Aldrich. Trace metal grade nitric acid (65%-70%) was purchased from VWR.
112
Triticum aestivum seeds (Pembroke 2014) were bred by Dr. David Van Sanford (Department of Plant and
113
Soil Sciences, University of Kentucky).
114
Nanoparticles and Characterization of Nanoparticle properties. CuO NPs (~40 nm primary particle
115
size), were purchased from Sigma-Aldrich. The primary size of particles, zeta potential, isoelectric point
116
and hydrodynamic diameter have been characterized and reported in our previous study11. The details of
117
characterization methods can be found in SI.
118
Soils and Characterization of Soil Properties. Standard Lufa 2.2 soil (loamy sand) was purchased from
119
Lufa Speyer, Germany. Lufa 2.2 soil contains 1.6 wt. % soil organic matter, and little total and available
120
Cu (see supporting information SI, Table S1 and Table S2, control treatment). Using a well-characterized
121
standard soil allows comparisons between studies. The high carbon organic content (about 1.6%) of Lufa
122
2.2 makes this soil good for agricultural studies. Soil was air dried and sieved < 2mm before shipping.
123
The soil was further air-dried for at least 24 hours before all experiments. Soil pH in different treatments 4 ACS Paragon Plus Environment
Page 4 of 24
Page 5 of 24
Environmental Science & Technology
124
was determined by the CaCl2 extraction method (see ‘Extraction methods’ section). Soil moisture content
125
(1% for the air dried soil) was determined gravimetrically after oven-drying the soil at 105 ºC for 24 h46.
126
Soil field moisture capacity (21%) was determined using a Haines apparatus with 0.1 bar pressure
127
difference between the wet soil and the atmosphere.
128
Soil amendment. The CuO NP suspension (containing Na2SO4), CuSO4 solution, or Na2SO4 solution
129
(control treatment) were mixed with soil and brought to a moisture content of 21.7% (corresponding to
130
~50% of the water holding capacity). The soil was mixed with wooden sticks in a beaker for 20min. The
131
homogeneity was confirmed with the low standard deviation for the total Cu content measured by soil
132
digestion data (SI, table S1). To test if CuO NP and CuSO4 treatments resulted in different Cu
133
bioavailability and toxicity, the Cu ion concentration had to be high enough to ensure some CuO NPs
134
remained in the soil during the study period. We chose 500mg/kg (as Cu) for the CuO NP treatment, and
135
300 mg/kg (as Cu) for the CuSO4 treatment based on a preliminary study to assess the solubility of the
136
CuO NPs in the Lufa 2.2 soil (SI, Figure S1). The results showed that the solubility of CuO NPs in Lufa
137
2.2 soil was ~300mg/kg. Therefore, the selected concentrations provided a similar concentration of added
138
Cu ion in both treatments after one month.
139
Germination and plant growth. The seeds of Triticum aestivum were surface sterilized by submerging
140
them in 10% sodium hypochlorite solution for 10 minutes, and then washed with DI water three times.
141
The seeds were then kept immersed in DI water overnight on an end-to-end rotator. The following day,
142
the seeds were transferred to a petri-dish containing moist tissue paper. The petri-dishes were covered
143
with aluminum foil and incubated in the growth chamber for 7 days, until 90% germination was achieved.
144
Germinated seeds were transplanted into syringes containing 120g of amended soils either immediately
145
after adding the Cu (fresh treatment) or 28 days after the Cu was added (aged treatment).The plants were
146
incubated in a growth chamber with constant moisture content and 16h-light/8h-dark cycle (25 °C for
147
daytime and 21 °C for night time). A diluted Cu-free Hoagland solution (quarter strength) was added
148
(1ml/day) to each syringe to maintain the moisture content of the soil as well as provide nutrients to
149
plants. The concentration of Cu in soil and plant tissue was determined using a standard digestion method
150
(EPA Method 3050b47) and ICP-MS analysis of Cu in the digestate. See SI. Adding moisture content did
151
not induce any vertical transport of Cu, as suggested by Figure S4.
152
Sampling of soil and plant tissue. Prior to transplanting the germinated seeds in soils, subsamples of
153
each soil were collected from all treatments for DTPA extraction (2g of soil per extraction) to measure the
154
labile metal fraction. After 14d of growth, rhizosphere soil, "loosely attached soil," and bulk soil (Figure
155
S2) were collected for DTPA and CaCl2 extraction to determine the total dissolved metal and readily
156
available metal, respectively, as described below. After the plants and roots were removed from the 5 ACS Paragon Plus Environment
Environmental Science & Technology
157
syringe, the soil remaining inside the syringe was defined as bulk soil, presumably minimally affected
158
by the plant roots. The bottom 5mm of bulk soil was also collected to determine if there was significant
159
vertical transport of Cu. No vertical transport of Cu was observed (SI Figure S4). The roots were
160
separated from shoots. Both roots and shoots were photographed with a scale bar for determination of
161
length. For each treatment, one plant root replicate was washed with 1mM KCl three times for Cytoviva
162
analysis (described below). The remaining roots were shaken by hand in a 50 ml centrifuge tube, and the
163
soil that detached during shaking was defined as loosely attached soil48 (Figure S2). After shaking, the
164
roots were placed on aluminum foil and air dried in a fume hood for 24 hours. The roots were then shaken
165
again in a 50 ml centrifuge tube, and the soil that detached during the air-drying process and the second
166
shaking process was defined as rhizosphere soil49. Due to the small amount of rhizosphere soil collected
167
per treatment, not all replicates were suitable for DTPA and CaCl2 extraction. The details of which
168
samples were analyzed can be found in SI (Table S3 and Table S4). For CuSO4 treatments, the roots were
169
highly compacted, precluding the collection of rhizosphere soil. The shoots and roots were oven-dried at
170
105 ºC for 24 h. The mass of the dried roots and shoots was recorded before digestion for total Cu
171
analysis (details in SI).
172 173
Soil extraction. Two standard extraction fluids were used in this study. DTPA extractant was composed
174
of 0.01M CaCl2, 0.005M DTPA and 0.1M triethanolamine (TEA) (pH=7.6). CaCl2 extractant was 0.01M
175
CaCl2 without pH adjustment. All extractions were done using a reciprocal shaker at 180 rpm for 2 hours.
176
It is important to note that the centrifuge tubes were laid horizontally in the shaker rather than vertically to
177
provide the best extraction efficiency. For soil samples collected before the plant growth experiments, 2g
178
of soil were extracted with 4ml DTPA extractant. For bulk soil samples, loosely attached soil samples and
179
rhizosphere soil samples, 0.4g of soil was extracted with 0.8ml DTPA extractant, while 0.35g of soil was
180
extracted with 3.5ml CaCl2 extractant. After extraction, all samples were centrifuged at 3000 rpm for 10
181
min, and the supernatants were filtered using a 0.45 µm PTFE filter. The pH of the CaCl2 extracts for
182
each soil fraction was measured (Figure 5). All samples were acidified with 20% HNO3 (final HNO3
183
concentration, 2%) and Milli-Q-water before being analyzed by ICP-MS. The method for ICP-MS is
184
provided in detail in the SI.
185 186
Cytoviva analysis. The interaction between roots and NPs were visualized in fresh roots after a rinsing step
187
in 10-3 M KCl, using a darkfield-based hyperspectral imaging (DF-HSI) system (CytoViva Inc., USA). See
188
SI for additional details.
189 6 ACS Paragon Plus Environment
Page 6 of 24
Page 7 of 24
Environmental Science & Technology
190
Results
191
Nanoparticle characterization. The properties of the CuO NPs have been previously described11.
192
Briefly, the primary particle size was determined by TEM to be 38nm ± 1.7nm (278 particles counted,
193
95% CI). The hydrodynamic diameter of an 80mg/kg CuO NP in an aqueous 5mM NaHCO3 suspension
194
(pH=7) was measured to be 560nm±103nm (3 replicates, 95% CI, intensity averaged), and the zeta
195
potential was -16.1mV±1.7mV (3 replicates, 95%CI) in the same suspension. The pH of the isoelectric
196
point (pHiep) of the CuO NPs in a 5mM NaNO3 solution was 8.8, while the pHiep was 5.8 in the 5mM
197
NaHCO3 solution11. The hydrodynamic diameter and zeta potential likely change after they are added to
198
the soils due to interactions with soil components such as natural organic matter and calcium50,51.
199 200
Change in extractability of Cu in bulk soil during the plant growth experiment. The DTPA
201
extractable Cu in the bulk soil for CuO NP and CuSO4 amended soils are shown in Fig.1. The DTPA
202
extractable Cu represents the Cu that was released from the CuO NPs during the treatment. The
203
extractable Cu vs. time is shown for the 14d growth period for both the freshly added CuO NPs, and for
204
the aged CuO NPs, where plants were added after the CuO NPs had aged for 28d prior to planting the
205
germinated seeds. The total Cu concentration in the two treatments and in the control (unamended) soil is
206
provided in the SI (Table S1).
207
For the CuO NP treatments, the DTPA extractable Cu from bulk soil increased over time (Figure
208
1a) (ANOVA test, P0.05). The DTPA extractable Cu in the
217
aged CuO NP treatment was statistically significantly higher than both fresh and aged CuSO4 treatment
218
(ANOVA test followed by Fisher’s LSD test for multiple comparison, P≤0.05) (Figure 1c). The CaCl2
219
extractable Cu revealed a different order, with fresh CuSO4 treatment having the highest CaCl2
220
extractable Cu, followed by the aged CuO NP treatment and the aged CuSO4 treatment, with the fresh
221
CuO NP treatment having the lowest amount of CaCl2 extractable Cu. The CaCl2 extractable copper
222
represents the “readily available” Cu in the porewater. 7 ACS Paragon Plus Environment
Environmental Science & Technology
223 224 225
Figure 1. Change in DTPA extractable Cu over time for each treatment: a) CuO NP treatment, b) CuSO4
226
treatment, and comparison of mean of extractable Cu for each Cu treatments at the end of the plant
227
growth period: c) DTPA extraction, d) CaCl2 extraction. Error bars show ± 1 standard deviation. In a and
228
b, capital letters indicate significant differences between DTPA extractable Cu at four time points for
229
CuO NP treatments (a) and CuSO4 treatments (b). In c and d, capital letters indicate significant
230
differences in DTPA extractable Cu (c) and CaCl2 extractable Cu (d) among soils collected after plant
231
harvesting (ANOVA test followed by Fisher’s LSD test for multiple comparisons, P≤0.05).
232
Toxicity of CuSO4 and CuO NP. Root maximal length, root compactness (root mass/root maximal
233
length), leaf lengths, shoot mass (Figure 2) and root morphology (SI Figure S5 and Figure S6) were used
234
to evaluate the toxic effect of CuSO4 and CuO NPs.
235
Root maximal length and root compactness indicated no visual toxic effect from the fresh CuO
236
NP treatment. For other treatments, significant decreases in root maximal length (a decrease of 6.6cm,
237
8.2cm, and 6.8cm compared to the control treatment for aged CuO NPs, fresh CuSO4, and aged CuSO4,
238
respectively) were observed. Increased root compactness was observed for the aged CuO NP treatment
239
(an increase of 4.0 mg/cm compared to the control) and for the fresh CuSO4 treatment (an increase of 5.1
240
mg/cm compared to the control). Examples of shortened roots and compactness of roots are shown in the
241
SI (Figure S5). Evidence of Cu toxicity was also observed in Cytoviva images. In comparison to the roots 8 ACS Paragon Plus Environment
Page 8 of 24
Page 9 of 24
Environmental Science & Technology
242
exposed to CuSO4 (SI, Figure S6), the roots exposed to CuO NP (fresh or aged) did not present the same
243
damaged physiology. Roots exposed to CuSO4 (both fresh and aged) showed a brown damaged (necrotic)
244
zone that was not found on any of the CuO NP exposed roots. No effects of Cu on the shoots (leaf length,
245
biomass) were observed for the CuO NP treatments. Both the freshly amended and aged CuSO4
246
treatments resulted in shorter third leaves (shortened by 5.4cm and 4.0 cm compared to the control for
247
fresh and aged CuSO4 treatments, respectively). The freshly amended CuSO4 treatment also had less total
248
shoot biomass compared to the control treatment.
249
Some indication of toxicity was evident in all treatments except for the fresh CuO NP treatment.
250
Aging of CuSO4 decreased its toxicity to Triticum aestivum, while aging of CuO NP increased its toxicity.
251
Overall, the two CuSO4 treatments showed more toxic effects to Triticum aestivum compared with the
252
two CuO NP treatments, even though the CuSO4 was added at a significantly lower Cu concentration
253
(300mg/kg for CuSO4 treatments vs. 500mg/kg for CuO NP treatment).
254
255 256 257
Figure 2. a) Root compactness and b) leaf length (leaf growth stage is noted with number, from 1 being
258
the youngest to 3 the oldest) of wheat seedlings grown in freshly amended and aged CuO NP, CuSO4-
259
amended soil, and control treatments. Error bars show ±1SD, * indicates P≤0.05; ** indicates P≤0.01.
260
(ANOVA test followed by Fisher’s LSD test for multipal comparisions) compared to the control
261
treatment.
262
Cu root association and Cu uptake. The presence of CuO NPs associated with the roots after 2 weeks of
263
plant growth in both fresh and aged CuO NP amended soils was investigated using enhanced dark-field
264
hyperspectral imaging (DF-HSI) as shown in Figure 3. The pixels containing CuO NP have been
265
highlighted in red. In both fresh and aged CuO NP amended soils (Figure 3), CuO NPs were found 9 ACS Paragon Plus Environment
Environmental Science & Technology
266
associated to specific locations on the roots, either to the root tip mucilage (Figure3 a, b, f, g), or to soil
267
aggregates attached to the root hairs or root tips (Figure 3 a, c-i). For the concentration of Cu in roots, all
268
Cu treatments were significantly higher than the control treatment. The Cu concentration in roots
269
(577mg/kg, s.d.=46mg/kg, 6 replicates) was statistically significantly higher in the freshly amended CuO
270
NP treatment than in the aged CuO NP treatment (400mg/kg, s.d.=60mg/kg, 6 replicates) or either ionic
271
treatment (278mg/kg, s.d.=51mg/kg, 6 replicates for fresh CuSO4 and 442mg/kg, s.d.=67mg/kg, 6
272
replicates for aged CuSO4) (Figure S7, a). For the shoot concentrations, no statistically significant
273
differences were found for all Cu treatments (53mg/kg-88mg/kg) (Figure S7, b).
274 10 ACS Paragon Plus Environment
Page 10 of 24
Page 11 of 24
Environmental Science & Technology
275 276
Figure 3. Hyperspectral imaging of plant roots grown in soil with freshly amended CuO NPs (a-e) or
277
after aging (f-i). The b, c and g views are magnified views from a and f. Pixels containing the reflectance
278
spectra specific to CuO NPs are highlighted in red. CuO NPs and their aggregates were found associated
279
to mucilage, root tissues and root hairs (red arrow), and to soil aggregates attached to those locations
280
(yellow arrows). Scale bars: 25µm
281 282
Effect of near-root environment on Cu availability from CuO NP treatment. Figure 4 shows the
283
differences in extractable Cu in rhizosphere soil, loosely attached soil and bulk soil for fresh and aged
284
CuO NP treatments. For the CaCl2 extraction in both fresh and aged CuO NP treatments (Figure 4a, b),
285
the extractability of Cu in the rhizosphere soil was significantly higher than the extractability of Cu in the
286
loosely attached soil or bulk soil (ANOVA test, P≤0.05).
287
There were no statistically significant differences among DTPA extractable Cu measurements
288
from rhizosphere soil, loosely attached soil and bulk soil in the freshly amended CuO NP treatment
289
(Figure 4c). However, the DTPA extractable Cu in the rhizosphere soil in the aged CuO treatment was
290
significantly lower than the DTPA extractable Cu in bulk soil, but similar to that measured in loosely
291
attached soil (Figure 4d). In control experiments (Na2SO4), the CaCl2 extractable Cu was below the
292
detection limit (0.08mg/kg in soil, 4ug/L for the diluted samples) in all soil samples.
293
Aging increased the concentration of CaCl2 extractable Cu and DTPA extractable Cu in loosely
294
attached soil and bulk soil, and increased the concentration of DTPA extractable Cu in rhizosphere soil (t
295
test, P0.05).
11 ACS Paragon Plus Environment
Environmental Science & Technology
296 297 298
Figure 4. CaCl2 and DTPA extractable Cu in fresh (left side) and aged (right side) CuO NP amended
299
rhizosphere soil, loosely attached soil and bulk soils. Error bars show ± 1 SD. Capital letters indicate
300
significant differences between groups (One way ANOVA test followed by Fisher’s LSD test for multiple
301
comparison, P≤0.05).
302 303
Soil pH in bulk soil, rhizosphere soil and loosely attached soil. For all CuO NP treatments and the
304
control treatment (no addition), the pH of the rhizosphere soils was significantly higher than the pH of the
305
loosely attached and bulk soils (Figure 5a, b and c). In freshly amended CuO NP treatments and control
306
treatments, the pH of the loosely attached soils were not statistically significantly different than the pH in
307
the bulk soils. However, in aged CuO NP treatments, the pH of the loosely attached soil was statistically
308
significantly higher than the pH in bulk soil. In bulk soil, the pH was the highest in freshly amended CuO
12 ACS Paragon Plus Environment
Page 12 of 24
Page 13 of 24
Environmental Science & Technology
309
NP treatments, followed by aged CuO NP treatment, followed by the control treatment, followed by the
310
aged CuSO4 treatment, with freshly amended CuSO4 treatment having the lowest soil pH (Figure 5d).
311
312
313 314 315
Figure 5. Mean ± SD of soil pH (measured using CaCl2 extraction) in rhizosphere soil, loosely attached
316
soil and bulk soil in a) soil freshly amended with CuO NP, b) aged CuO NP treatment c) control soil, and;
317
d) Comparison of pH of bulk soil among all treatments. Capital letters indicate significant differences
318
(ANOVA test followed by Fisher’s LSD test for multiple comparison, P≤0.05).
319 320
Discussion
321
CuO NP dissolution is linked to toxicity. Compared to Cu ions, the dynamic dissolution process of CuO
322
NP in soil led to a very different Cu exposure profile for plants. At the end of the two growth periods, the
323
DTPA-extracted Cu concentrations in CuO NP treatments were similar or even higher than in the CuSO4 13 ACS Paragon Plus Environment
Environmental Science & Technology
324
treatment. However, a decreasing trend in DTPA extractable Cu on CuSO4 treatments during the two
325
plant growth periods was observed. This decrease can be attributed to the soil-organic matter interactions,
326
solid-state diffusion of Cu ions into iron minerals or metal (co)precipitates52–54. Conversely, an increase in
327
DTPA-extractable Cu over time was shown in CuO NP treatments (fresh treatment and aged treatment)
328
during the two plant growth periods and the aging period. This can be attributed to the dissolution of CuO
329
NP11,12. Thus, the plants in the freshly amended CuO NP soil were exposed to lower amounts of labile Cu
330
compared to either the two CuSO4 treatments or the aged CuO NP treatment. These findings suggest that
331
when evaluating the chemical availability or toxic effect of metal-based NPs in soil, single-time point
332
chemical extractions at the end of the plant growth period cannot capture the dissolution process of NPs
333
in soil, and thus may fail to predict the toxicity or bioavailability of NPs11,12,55. Considering that it is not
334
feasible to uniformly dose Cu ions into soil over time to precisely mimic the dosing rate from NP
335
dissolution, toxicity studies with soluble NPs should measure the dissolution rate in soil and monitor the
336
behavior of soluble ions in soil, and interpret their results in light of the different dosing conditions that
337
manifest.
338
A significantly higher toxicity in CuSO4 treatments compared to the fresh CuO NP treatment is
339
explained by the higher exposure of roots to labile Cu species, even though the CuO NP treatment had a
340
higher total Cu concentration. Also, dissolution of CuO NPs over time gradually increased the available
341
Cu in soil, leading to higher toxicity in the aged treatment. The opposite has been observed with CuSO4,
342
where the available Cu in soil decreased over time, leading to lower toxicity to the plants in the aged
343
treatment. The effects of time on toxicity of CuO NP and CuSO4 have already been observed19. The
344
authors attributed this to the dissolution behavior of CuO NP, although without quantification. Here, we
345
clearly showed that in order to correlate the chemical availability of CuO NPs with toxicity, the
346
dissolution kinetics, i.e. predicting the total Cu released to soil during the growth period, should be
347
considered. The dissolution kinetics can be modeled as first-order dissolution, with the rate constant fit to
348
the extractable Cu over time11, and the total amount of Cu ion released from CuO NPs can be estimated
349
by integrating the expression relating the change of extractable Cu over time. This observation is relevant
350
to NP formulations of fungicides and micronutrients, so the release rate of the active ingredients can be
351
better timed to the plant’s needs.
352 353
CaCl2 extractable Cu correlates with toxicity of CuO NP to wheat. Although DTPA extractable Cu
354
gives a better indication of CuO NP dissolution because it extracts most of the Cu species in soil, CaCl2
355
extractable Cu is better for correlating toxicity, since it measures dissolved Cu in pore water that can
356
directly interact with plant roots. DTPA extraction would predict the toxicity of the aged CuO NP to be
357
higher than the CuSO4 treatment (Figure 1 a, b). However, this was not the case. The aged CuO NP had 14 ACS Paragon Plus Environment
Page 14 of 24
Page 15 of 24
Environmental Science & Technology
358
lower toxicity compared to both the aged and the fresh CuSO4 treatment, indicating that the DTPA
359
extraction cannot accurately predict toxicity for the CuO NPs. The CaCl2 extraction ranked them correctly
360
(Figure 1d). Considering that the extractable Cu in CuO NP amended soil increased over time while the
361
extractable Cu in CuSO4 amended soil decreased over time in Lufa 2.2 soil (SI, Figure S1), the wheat
362
plants were exposed to a lower overall ‘readily available’ Cu (i.e. Concentration x time) in CuO NP
363
treatments compared to the CuSO4 treatments. The lower CaCl2 extractable Cu in aged CuO NP treatment
364
is a result of higher soil pH in aged CuO NP treatment compared to the fresh and aged CuSO4 treatment.
365
Higher soil pH has been previously shown to lower Cu concentration in soil pore water 27,56.
366 367
Root-associated CuO NP modulates toxicity. In the freshly amended CuO NP soil, although being
368
exposed to a lower concentration of labile Cu, the roots of Triticum aestivum were actually exposed to
369
higher total Cu concentration (Figure S7) than the other treatments. This is mainly due to CuO NPs'
370
association with plant roots (Figure 3b). This exposure to CuO NPs did not lead to any detected toxic
371
effects, indicating a low or de minimis level of toxic effects from the particle itself over the 14d growth
372
period.
373 374
Root exudates affect CuO NP dissolution and availability. The increase in the pH of rhizosphere soil
375
compared to bulk soil in our study indicates that the rhizosphere region was indeed influenced by the
376
plant roots. Excretion of organic acid (dissociated ions) by plant roots , nitrogen uptake and ionic
377
exchanges by plant roots may explain the higher pH of the rhizosphere soil compared to the pH of bulk
378
soi39,40,57,58. The observed pH change in rhizosphere soil was not likely a result from the presence of CuO
379
NP, as similar pH changes occurred with both the fresh CuO NP treatment and the negative control
380
treatment (0.4pH unit, ANOVA test, P>0.05). However, in the aged CuO NP treatment, the pH increase
381
was higher (0.6pH unit, ANOVA test, P
