Ecotoxicology and Environmental Safety 206 (2020) 111405 Contents lists available at ScienceDirect Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv Impact of Ag2S NPs on soil bacterial community – A terrestrial mesocosm approach S. Peixoto a, Z. Khodaparast a, G. Cornelis b, E. Lahive c, A. Green Etxabe c, M. Baccaro d, A. G. Papadiamantis e, f, S.F. Gonçalves a, I. Lynch c, M. Busquets-Fite g, V. Puntes h, i, j, S. Loureiro a, *, I. Henriques k a CESAM-Centre for Environmental and Marine Studies & Department of Biology, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal b Department of Soil and Environment, Swedish University of Agricultural Sciences, 75651 Uppsala, Sweden c UK Centre for Ecology and Hydrology, Maclean Building, Benson Lane, Crowmarsh Gifford, Wallingford OX10 8BB, UK d Division of Toxicology, Wageningen University, 6708 WE Wageningen, The Netherlands e School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston B15 2TT, Birmingham, UK f NovaMechanics Ltd., 1065 Nicosia, Cyprus g Applied Nanoparticles SL, C Àlaba 88, 08018 Barcelona, Spain h Institut Català de Nanociència i Nanotecnologia (ICN2), CSIC, The Barcelona Institute of Science and Technology (BIST), Campus UAB, Bellaterra, Barcelona, Spain i Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain j Vall d’Hebron Institut de Recerca (VHIR), 08035 Barcelona, Spain k Department of Life Sciences, Faculty of Sciences and Technology, University of Coimbra, Portugal A R T I C L E I N F O A B S T R A C T Edited by Richard Handy Soils might be a final sink for Ag2S nanoparticles (NPs). Still, there are limited data on their effects on soil bacterial communities (SBC). To bridge this gap, we investigated the effects of Ag2S NPs (10 mg kg− 1 soil) on the Keywords: structure and function of SBC in a terrestrial indoor mesocosm, using a multi-species design. During 28 days of Silver sulfide nanoparticle exposure, the SBC function-related parameters were analysed in terms of enzymatic activity, community level Soil bacterial community physiological profile, culture of functional bacterial groups [phosphorous-solubilizing bacteria (P-SB) and het­ Denaturing gradient gel electrophoresis erotrophic bacteria (HB)], and SBC structure was analysed by 16S rRNA gene-targeted denaturing gradient gel Soil enzymatic activity Community level physiological profile electrophoresis. Indoor Mesocosm The SBC exposed to Ag2S NPs showed a significative decrease of functional parameters, such as β-glucosidase activity and L-arginine consumption, and increase of the acid phosphatase activity. At the structural level, significantly lower richness and diversity were detected, but at later exposure times compared to the AgNO3 treatment, likely because of a low dissolution rate of Ag2S NPs. In fact, stronger effects were observed in soils spiked with AgNO3, in both functional and structural parameters. Changes in SBC structure seem to negatively correlate with parameters related to phosphorous (acid phosphatase activity) and carbon cycling (abundance of HB, P-SB, and β-glucosidase activity). Our results indicate a significant effect of Ag2S NPs on SBC, specifically on parameters related to carbon and phosphorous cycling, at doses as low as 10 mg kg-1 soil. These effects were only observed after 28 days, highlighting the importance of long-term exposure experiments for slowly dissolving NPs. 1. Introduction antimicrobials has been predicted to grow from 0.79 billion dollars in 2014–2.54 billion dollars in 2022 (Pachapur et al., 2016). The Silver nanoparticles (AgNPs) are broadly used in many applications increasing application and consumption of AgNPs-containing products and products, due to their anti-microbial, anti-inflammatory, and anti- results in AgNPs release into wastewater treatment plants (WWTPs) cancer properties (McGillicuddy et al., 2017). The global production (European Commission Science for Environment Policy, 2017). In these of AgNP-based products has been estimated at 450 tons per year plants, it has been shown that most of the AgNPs are captured and (McGillicuddy et al., 2017), whereas the market of AgNP-based settled in the sludge (Kaegi et al., 2011), where the estimated AgNPs * Corresponding author. E-mail address: sloureiro@ua.pt (S. Loureiro). https://doi.org/10.1016/j.ecoenv.2020.111405 Received 12 May 2020; Received in revised form 18 September 2020; Accepted 19 September 2020 Available online 30 September 2020 0147-6513/© 2020 Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 concentration ranges between 1 and 6 mg kg− 1 (Gottschalk et al., 2009). spots, as well as their toxicity to soil microorganisms. In fact, changes in During wastewater treatment, the majority of AgNPs (79% as estimated oxygenation levels and soil pH might occur due to the organism’s by Wang et al., 2016) are transformed to Ag sulfide NPs (Ag2S NPs), behaviour (burrowing, ingestion and egestion, among other activities) which are then incorporated into biosolids. Additionally, this sulfidation (El-Wakeil, 2015), promoting changes in microbial communities’ process also was suggested to occur for silver in ionic form (Ag+) as a abundance and structure. Thus, microbial responses to the contaminants means to reduce silver toxicity (Kent et al., 2014). can be altered and pose a risk for the terrestrial ecosystem functions. In many countries, these biosolids are amended to agricultural soils Our study aims to investigate the effects of Ag2S NPs on SBC at to improve soil fertility (Doolette et al., 2016). In the United Kingdom structural (Denaturing Gradient Gel Electrophoresis - DGGE) and func­ for example, of the 1.05 million tons of dry biosolids produced per year, tional (culture of functional bacterial groups, soil enzymatic activity, 85% on average are applied on land (European Commission Science for and Community Level Physiological Profiling - CLPP) levels, using a Environment Policy, 2017). Agricultural soils are often considered a sink more realistic exposure scenario than has been reported previously, in for Ag2S NPs, characterized by a low dissolution rate of these Ag2S in order to obtain relevant information at the ecosystem level. Thus, indoor biosolids and most likely also in soils. For instance, Wang et al. (2016) terrestrial mesocosms experiments were performed in which a 1000x demonstrated that the ≥ 87% of Ag2S NPs from biosolids remained in predicted environmental concentration estimated for 2050 (10 µg kg− 1 - this form at least up to 400 days after soil spiking. The dissolution rate of sludge treated soils) (Giese et al., 2018) was tested in the presence of Ag2S NPs depends on the degree of sulfidation (Reinsch et al., 2012) and biota, simulating a worst-case scenario. Besides SBC, wheat plants a higher degree of sulfidation of silver results in lower dissolution and a (Triticum aestivum L.) and terrestrial organisms, such as isopods (Por­ lower toxic effect on soil organisms (Starnes et al., 2015; Wang et al., cellio scaber), mealworms (Tenebrio molitor), and earthworms (Lumbricus 2017) and microorganisms (Reinsch et al., 2012; Doolette et al., 2015). rubellus) were also included in the mesocosms to simulate a real edaphic Indeed, the toxicity of AgNPs strongly depends on its speciation (Doo­ scenario. The impact of Ag2S NPs on SBC was evaluated as a function of lette et al., 2015), and insoluble Ag2S NPs (Ksp = 8 × 10− 51) are less toxic time (0, 14 and 28 days) and the impact was compared with that of the and less bioavailable compared to non-sulfidated AgNPs (Doolette et al., ionic Ag form (AgNO3) and an unexposed control. 2015). Only a limited number of studies have thus far explored the toxic 2. Material and methods effects of Ag2S NPs on terrestrial organisms. Wang et al. (2017) showed that Ag2S NPs were bioavailable to some extent and caused toxic effects 2.1. Silver exposure in planta. A substantial increase of Ag accumulation on leaves (3.8–5.8 μg Ag g− 1 dry mass) was demonstrated when wheat (Cucumis sativus) The study was conducted with well defined sulfidized silver nano­ and cucumber (Triticum aestivum L.) plants were exposed to 10 mg (Ag) particles colloids (Ag2S NPs) and AgNO3 (Sigma-Aldrich, 99% purity, L− 1 as Ag2S NPs for one week (Wang et al., 2017). However, Ag2S NPs CAS 7761–88–8, Germany). AgNO3 was used as a Ag control for fate and toxicity to plants are still a matter of debate since different comparing the toxicity of the NPs with the toxicity of the ionic Ag form. studies have used different plant species and different silver concen­ The polyvinylpyrrolidone (PVP) coated Ag2S NPs colloids were syn­ trations (Wang et al., 2017), obtaining conflicting results. A study by thesized and characterized by Applied Nanoparticles (Barcelona, Spain). Starnes and their collaborators suggests that the exposure to Ag2S NPs These particles were used in this work as a model of sulphidized Ag2S (at 10 mg Ag L− 1) resulted in a 20% increase in mortality rate of Cae­ NPs, in which no degree of sulfidation was demonstrated. These parti­ norhabditis elegans (Starnes et al., 2015). However, this observed toxic cles were supplied in suspension with a stock concentration of 1320 ± effect is mainly due to the solubilization of Ag+ (released from Ag2S 48 mg L− 1. These particles presented an average diameter of 20.4 ± particles) in soils. 11.9 nm [mean (n = 3) ± standard deviation] (Fig. S1), a ζ-potential of Although a few studies have investigated the stability (Reinsch et al., − 23.8 ± 4.5 mV in miliQ water (16.12 mg Ag2S NPs mL− 1, conductivity 2012), toxicity, fate (Schlich et al., 2018), and (bio)availability (Wang 0.174 mS cm− 1, pH 8.72). Additionally, Ag2S NPs colloids, measured by et al., 2016) of Ag2S NPs in soils, the impact on soil bacterial commu­ ICP-MS (fully described in Supplemental materials S1), showed high nities (SBC) and microbiological processes is still unknown. SBC play a stability in ultrapure water over-time, and a very low proportion of vital role in soil ecological functions and have been considered a sen­ dissolved Ag during the 48 h (<0.1%, Table S1). sitive target for assessing the impact of manufactured nanomaterials on The Lufa 2.2 soil (LUFA-Speyer 2.2, LUFA Speyer, Speyer, Germany) terrestrial compartments (Doolette et al., 2016). Some microbial end­ was used as a natural and standard soil, because this soil is suggested as a points have been used to infer on these effects such as soil enzyme ac­ reference soil for testing NPs toxicity (Geitner et al., 2020). This soil was tivities (Samarajeewa et al., 2017), biological nitrification, microbial spiked with Ag2S NPs or AgNO3 to a final concentration of 10 mg Ag respiration rates and organic matter decomposition (Samarajeewa et al., kg− 1 soil. For this, an Ag2S NPs stock suspension and an AgNO3 stock 2017), community level physiological profiling (Samarajeewa et al., solution were prepared, using deionized water, with a final concentra­ 2017) and microbiome structure and diversity (Doolette et al., 2016). tion of 1150 mg Ag L− 1 and 1148 mg Ag L− 1, respectively. These stock Currently, the effects of Ag2S NPs on SBC are difficult to assess, since the solutions were added to the soil, to maximize homogeneity by mixing few reported studies used different NP concentrations, exposure times, thoroughly, and the soil was brought to 55% of the water holding ca­ soil types, and were conducted at different scales (e.g., microcosms, pacity (WHC) using de-ionised water. Spiked and watered soils were mesocosms). Even so, most studies showed a stronger impact of the ionic mixed manually which has been shown to result in a homogeneous silver when compared to AgNPs and Ag2S NPs (Doolette et al., 2016). exposure throughout replicates. Once prepared, soils were transferred to Although Ag2S NPs were the least toxic form, impacts on the structure each replicate (column) and incubated for two days at 20 ◦ C± 2, to (Judy et al., 2015) and composition (Doolette et al., 2016) of SBC were achieve a chemical equilibrium stage of the soil. This is a common confirmed. In terms of function-related parameters, a decrease in the procedure for equilibrium in soil for metal exposure in ecotoxicological abundance of the genes nirK, encoding nitrite reductase, and amoA, tests and it has also been tested for nanomaterials (see Neves et al., encoding ammonia oxidizing bacteria, was demonstrated as well as a 2019). Indeed, silver transportation, speciation, and availability might reduction of the nitrification rate (Doolette et al., 2016). Studies with change in soils due to silver exchange between the solid (soil) and liquid single microbial species also reported toxic effects, e.g. on Escherichia phase (pore water). Each mesocosm column was watered daily with coli (E. coli) (Reinsch et al., 2012) and Arthrobacter globiformis (Schultz artificial rainwater (16 mL per day; NaCl (0.01 mM), (NH4)4 SO4∙H2O et al., 2018). (0.0053 mM), NaNO3 (0.0059 mM), and CaCl2∙H2O (0.0039 mM); pH= In a terrestrial ecosystem, the presence of different organisms might 5.1) dripping/distributing the water volume over the whole corer sur­ change the bioavailability of contaminants in soils, in some localized face by a syringe avoiding the edges. The soil pH was measured at 2 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 sampling times (day 0, 14 and 28); five grams of soil were shaken with reagent blank, n = 22) and the instrument method had a precision of 25 mL of a 0.01 M CaCl2 solution, and pH measured after a 2 h settling 0.28% (n = 22). period. Soil pH did not change with Ag2S NPs or AgNO3 exposure in time To analyse Ag concentration (total and soluble) in soil pore water, (Table S2). the procedure was adapted from Lahive et al. (2017). After sampling, the soil was moistened (to 100% of its water holding capacity) to achieve an 2.2. Mesocosm setup extractable aqueous sample volume, and incubated overnight at 20 ◦ C. The soils (35 g) were then transferred to inserts with 0.02 g glass wool The terrestrial mesocosm was established in indoor conditions and 0.45 µm PVDF filters, previously soaked with a Cu solution (0.1 M [(temperature 20 ± 2 ◦ C and photoperiod 16 h (light):8 h (dark)] and CuSO4) to minimize Ag+ adsorption (Cornelis et al., 2010), and the pore three soil treatments were used: non-exposed soil, as the negative con­ water was removed by 1 h centrifugation at 2000 g. Part of this pore trol (CT), soil spiked with Ag2S NPs and soil spiked with AgNO3 (as the water extract was sampled for total metal analysis and the remaining ionic control, Ag+). For each treatment, 10 replicates were made. Each water was centrifuged in a Cu soaked 10 kDa ultra-filtration tube at mesocosm replicate (column = 20 cm long and 11 cm diameter) was 4000 g for 30 min to measure the soluble Ag concentration. Ag analyses filled with 2.6 kg of Lufa 2.2 soils: 1.3 kg (wet soil) of control or spiked were completed as described above by ICP-MS. soil (with Ag2S NPs or AgNO3) in the upper layer (1–8 cm), and 1.3 kg of non-treated soil in the bottom layer (8–16 cm). Total soil height for the 2.4. Culture-dependent analysis of SBC two layers was 16 cm. The bottom of the corer was closed with a PVP nylon mesh (1 mm) To assess Ag2S NPs effects on the culturable fraction of the SBC, and a funnel was placed underneath with a 50 mL tube to collect Colony Forming Units (CFUs) were counted in Nutrient Agar (NA) leachate. To the upper layer of each replicate ten isopods (Porcellio (Merck, Darmstadt, Germany), which is a general purpose medium for scaber), ten mealworms (Tenebrio molitor), six earthworms (Lumbricus heterotrophic bacteria, and in the National Botanical Research In­ rubellus) and four plants (Triticum aestivum L.) were included. A stitute’s phosphate (NBRIP) medium (Nautiyal, 1999), which was destructive sampling approach was used at each sampling time. At day designed to count bacteria able to solubilize phosphate (P-SB). Media 0 (after 2 days of spiking equilibrium), 14 and 28 (after organisms were were sterilized at 121 ◦ C for 20 min and supplemented with 0.2 mg L− 1 introduced), several replicates were removed from the design (three, cycloheximide (95% purity, Acros organics, New Jersey) to prevent the three and four, respectively at each sampling time and from each growth of fungi. treatment) and the 0–4 cm, 4–6 cm, 10–14 cm, and 14–16 cm mesocosm From samples collected at 0 and 28 days, three grams of soil were layers were destructively sampled for all invertebrates, plants, soil suspended in 25 mL of phosphate saline buffer (PBS, pH = 7.4), in chemical analysis and SBC (Fig. 1). triplicate. Soil samples were shaken at 304 g with twenty sterile glass During the experiment, organisms’ mortality (isopods, mealworms, beads (4 mm) for 10 min at 20 ◦ C, followed by fold serial dilutions (10− 2 and earthworms) was recorded and was reported to be similar among dilution) in PBS. From each suspension, 100 µL was plated on NA or treatments (data not shown). NBRIP (three replicates, with four technical replicates per medium), and the plates were incubated at 20 ◦ C, for three days. Colony counts were 2.3. Soil chemical analysis reported as % of the variation of Log10 [colony forming units (CFU) per gram of soil]. The total soil Ag concentration was determined using an aqua regia To assess the usefulness of the medium in the context of this study, digestion (3:1 HCl:HNO3) using a microwave (Milestone) (US-EPA, the identity of a randomly selected colony from NBRIP was confirmed by 1996). Digests were diluted to an acid concentration of ca. 3% and 16S rRNA gene sequencing (data not shown). measured using Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) (Perkin Elmer Nexion 350 D). Calibration procedures used 2.5. Enzymatic activity known Ag concentrations matrix-matched in 3% HCl solutions, and only the correlation coefficient greater than 0.999 was considered. 103Rh was Dehydrogenase (DHA), β-glucosidase (βG), acid phosphatase (AP), applied in all samples as internal standard quality control. The ICP-MS and arylsulfatase (AS) activities were determined according to the instrument detection limit for Ag was 0.028 μg l− 1 (mean blank + 3σ methods described by Dick et al. (1996), with some modifications. The Fig. 1. Scheme of the temporal procedures during the 28 days of the mesocosm experiment. 3 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 DHA activity was evaluated by suspending soil (2.5 g) in a 3% triphenyl Portugal), each primer (0.75 µL of a 10 mM solution), and a DNA sample tetrazolium chloride (TTC; Sigma-Aldrich, 98%) solution (1 mL), fol­ (1 µL). The amplification conditions consisted of an initial denaturation lowed by 24 h incubation at 37 ◦ C, in darkness (Dick et al., 1996). The step (94 ◦ C for 3 min), followed by 30 cycles comprising the following triphenyl formazan (TPF) produced was extracted with methanol. Then, steps: denaturation (94 ◦ C for 1 min), annealing (52 ◦ C for 1 min) and soil samples were centrifuged (3000 rpm for 15 min) and the superna­ extension (72 ◦ C for 2 min), and a final extension step (72 ◦ C for tant was measured photometrically at a wavelength of 546 nm. The 10 min). The second PCR was conducted as described above, using 1 µL results were expressed as μg of TPF g− 1 24 h− 1. The βG, AS and AP ac­ of the first PCR product as template, and the universal primers [338f-GC tivities were measured using specific substrate solutions: 4-nitrophenyl (5’-GACTCCTACGGGAGGCAGCAG-3’) and 518r (5’- β-D-glucopyranoside substrate (0.05 M) (Acros organics, 99%), 4-nitro­ ATTACCGCGGCTGCTGG-3’)], (Muyzer et al., 1993), and the amplifi­ phenyl sulfate (0.05 M) (Acros organics, 99%), and p-nitrophenyl cation conditions were: denaturation (94 ◦ C for 5 min), annealing (52 ◦ C phosphate (0.05 M) (Sigma-Aldrich), respectively (Dick et al., 1996; for 30 s) and extension (72 ◦ C for 30 s), and a final extension step (72 ◦ C Tabatabai, 1994). The soil samples (0.5 g) were incubated at 37 ◦ C for for 30 min). Positive and negative controls were included, consisting of 1 h with 0.5 mL of substrate solution and 2 mL of the modified universal identical reactions but using purified bacterial DNA from Escherichia coli buffer (MUB) with pH 6.5 (βG) and pH 6.0 (AP); or acetate buffer (0.5 M; ATCC 25922 and nuclease-free water as templates, respectively. pH = 5.8) for AS activity (Tabatabai, 1994). The reaction was stopped Amplification was confirmed by electrophoresis (1.5% agarose). with 0.5 M CaCl2 (0.5 mL) and 0.5 M NaOH (2 mL) and the soil samples PCR products were loaded into 8% (w/v) polyacrylamide (37.5:1, were centrifuged at 4562 g for 15 min. The supernatant was subse­ acrylamide:bisacrylamide) gels with denaturing gradient ranging from quently measured spectrophotometrically at wavelengths of 400 nm 35% to 62.5% [100% denaturant corresponded to 7 M urea and 40% (v/ (βG) and 410 nm (AP, AS). v) formamide]. A DGGE marker composed of 8 bands (Henriques et al., All enzymatic activities were measured in 96 wells microplates, in 2004) was included in the extremities of each gel. The electrophoresis three replicates (with four technical replicates) using a microplate was performed on a D-Code™ System (Bio-Rad, Hercules, CA, USA) with spectrophotometer (Thermo Scientific Multiskan Spectrum, USA). 1X TAE buffer (Sigma-Aldrich, Germany) at 60 ◦ C in two steps, (1) for 15 min at 20 V and (2) for 16 h at 70 V. Gels were stained in a solution of 2.6. Community level physiological profiling (CLPP) ethidium bromide (0.5 μg mL− 1) for 5 min and rinsed in distilled water (20 min). Images were captured by the Molecular Imager® Gel Doc™ Community-level physiological profiling (CLPP) was performed XR+ System (BioRad Laboratories, Hercules, California, USA). using 96-well Biolog® Ecoplates that contained 31 different carbon DGGE patterns were analysed using Bionumerics Software (Applied sources (Biolog, Hayward, CA, USA). The rate of utilization of these Maths, Belgium). carbon sources was verified by the reduction of tetrazolium violet redox dye, which changed from colourless to purple as a function of microbial 2.8. Statistical analysis activity (Garland, 1997). Following the method by Samarajeewa et al. (2017), three grams of soil were collected and added to 27 mL of sterile All microbial endpoints [CFU counts, enzymatic activity, and CLPP] water with 20 sterile glass beads. This soil suspension was shaken at were analysed using Sigma plot V.12.5 (SysStat software Inc., CA, USA). 304 g for 10 min, at 20 ◦ C and then left for 45 min. The suspension was Assumptions of homogeneity of variance and test for normality of dis­ diluted 100-fold to attain a cellular density of approximately 107 cell tributions were verified using Levene’s test and Shapiro-Wilk’s test, mL− 1. One hundred microliters of the suspension were then inoculated respectively. When the normality test failed, a Kruskal-Wallis test fol­ into each well, and plates incubated at 20 ◦ C. The colour development lowed by the appropriate post-hoc (Holm-Sidak) test was used. A level of was measured by reading optical density at a wavelength of 590 nm at p = 0.05 was considered to assume statistical significance. A two-way 24 h intervals over 192 h of plates incubation, using a microplate ANOVA analysis was conducted to assess the effect of the two factors spectrophotometer (Biolog, MicroStation TM, USA). (soil treatments vs. time of exposure). Additionally, the Tukey’s HSD Measurements were calibrated against blank wells (water) (Garland, (honestly significant difference) and Dunnett’s method test was applied 1997), and the optical density values below 0.06 were set as 0. Optical for the post-hoc analysis, to obtain multiple comparisons between density curves of CLPP were constructed for each technical replicate treatments or toward the respective control, respectively. (n = 3). The total area under the curve (AUC) was calculated using the DGGE band matrix was analysed in PRIMER v6 software (Primer-E trapezoidal integration function, included in the trapz {pracma} pack­ Ltd., Plymouth, UK). Band position and intensity was used to calculate age (version 2.2.2) in R program (version 3.5.1). the richness, the Shannon-Wiener diversity and Pielou’s indexes. A one- Additionally, the 31 carbon substrates from Biolog®Ecoplate were way ANOVA followed by the Dunnett’s method was used to discriminate divided into six groups following the method of Sala et al. (2006): (1) differences in diversity indices. Further, two-dimensional Principal Co­ carbohydrates, (2) carboxylic acids, (3) amines and amides, (4) amino ordinate Analysis (PCoA) was performed based on the Bray-Curtis sim­ acids, (5) polymers, and (6) phenolic acids. Thus, the substrate well ilarity after the matrix was square root transformed. Additionally, a averaged colour development (SAWCD) index was calculated based on PCoA analysis was performed as described above for DGGE patterns of optical density values obtained after 192 h of Biolog®Ecoplate incuba­ samples obtained after 28 days of exposure. Functional variables that tion for each soil treatment. presented a good correlation (R≥0.6 based on Pearson correlation) with SBC structure were represented as vectors in the PCoA plot. 2.7. PCR-DGGE Differences in SBC structure between treatments were evaluated through PERMANOVA based on 999 permutations. Soil (0.25 g) was collected from each corer, and the total DNA was extracted using the UltraClean® Power Soil DNA Isolation Kit (MoBio 3. Results Laboratories, Inc., Carlsbad, CA) following the instructions of the manufacturer. The effects of Ag2S NPs on SBC were assessed in different layers The V3 region of the 16S rRNA gene was amplified using a nested (0–4 cm, 4–6 cm, 10–14 cm, and 14–16 cm) of the mesocosm corer. As PCR strategy. First, the 16S rRNA gene was amplified using the universal expected, effects were clearer in the top-soil layer (0–4 cm), for which primers 27 F (5′ -AGAGTTTGATCCTGGCTCAG-3′ ) and 1492R (5′ - significant differences were observed between SBC subjected to different GGTTACCTTGTTACGACTT-3′ ). The PCR mixture (25 µL) contained: treatments [non-exposed soil - CT and soil exposed to Ag2S NPs or nuclease-free water (16.25 µL), NZYTaq 2x Green Master Mix (6.25 µL; AgNO3] (Fig. S2A–D). Moreover, a significant decrease in Shannon- 2.5 mM MgCl2; 200 mM dNTPs; 0.2 U/µL DNA polymerase) (NZYTech, Wiener and richness indexes was observed for soils exposed to Ag2S 4 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 NPs, in comparison to the CT, exclusively for top-soil samples (Table S1). 3.2. Treatment effects on enzymatic activity Based on these results and given that all the main biota groups (that were part of the mesocosms) were also in this layer, subsequent analyses The effects of Ag2S NPs treatment on soil enzymatic activities (DHA, were focused on the top-soil layer (0–4 cm). βG, AP, and AS) were assessed over time and are represented in Fig. 4 Soil analysis of the top 0–4 cm of the soil corer confirmed the ac­ (A–D). curate soil spiking procedures (Fig. 2A), with the control soil presenting The DHA activity (Fig. 4A) showed the greatest sensitivity to AgNO3 a basal value of 0.051 mg Ag kg− 1, while the AgNO3 and Ag2S NPs treatment resulting in significant activity decrease in the three sampling spiked soils presenting 9.97 and 12.20 mg Ag kg− 1, respectively. At day times. For this enzyme, an activity reduction of 72.8% was observed 28, the spiked soils still presented similar concentrations, with average after 28 days of exposure (Tukey HSD: CT vs. AgNO3: p < 0.001). No values of 10.7 ± 0.5 (AgNO3 addition) and 13 ± 1 mg Ag kg− 1 (Ag2S significant effect from Ag2S NPs exposure was observed (Tukey HSD: CT addition). The control soil maintained a similar basal concentration vs. Ag2S NPs: p = 0.852). Accordingly, significant differences between (0.045 mg Ag kg− 1). AgNO3 and Ag2S NPs treatments were observed (Tukey HSD: AgNO3 vs. Porewater analysis from top-soil (Fig. 2B) showed similar average Ag2S NPs: p < 0.001). values for non-exposed soil [0.03 ± 0.008 µg Ag L− 1] and soil exposed to A time-dependent effect was verified for βG activity (Fig. 4B). Thus, Ag2S NPs [0.10 ± 0.08 µg Ag L− 1], at day 0. Soil spiked with AgNO3 at an earlier time of exposure, the AgNO3 treatment resulted in a sig­ showed an initial Ag dissolution and presented an average value of nificant (Tukey HSD: CT vs. AgNO3: p = 0.020) activity reduction 2.47 ± 1.64 µg Ag L− 1. (AgNO3 0 day vs. CT 0 day: 75.3%). Despite an apparent recovery after After 28 days of exposure, a similar Ag content in porewater samples 14 days for the AgNO3 exposure, a significant reduction on βG activity was verified for CT and Ag2S NPs, although a slight increase in Ag was again observed after 28 days, but now for both treatments [Tukey dissolution was detected when compared with day 0, presenting an HSD: (CT > AgNO3: p < 0.001); (CT > Ag2S NPs: p = 0.001)]. average concentration of 0.33 ± 0.08 µg Ag L− 1 and 0.67 ± 0.02 µg Ag Moreover, the level of inhibition on βG activity observed for soils L− 1, respectively. Soils spiked with AgNO3 revealed a significant (Tukey exposed to Ag2S NPs was lower than for soils exposed to AgNO3 (Tukey test: F = 35.474; p < 0.001) free Ag+ concentration, when compared HSD: AgNO3 < Ag2S NPs: p = 0.019). with other treatments, presenting an average value of 1.74 ± 0.22 µg Ag After day 28, AP activity (Fig. 4C) significantly increased in soils L− 1 . exposed to Ag2S NPs (Tukey HSD: CT < Ag2S NPs: p = 0.007), while no significant effect was observed for soils exposed to AgNO3. 3.1. Effects on culturable bacteria The activity of AS significantly decreased after 14 days in soils spiked with AgNO3 (Tukey HSD: CT > AgNO3: p = 0.047), recovering to levels The abundance of heterotrophic bacteria and of bacteria able to similar to the control at day 28 (Fig. 4D). No significant effects were solubilize phosphate was determined using a culture-based method observed on the AS activity from Ag2S NPs exposure. (Fig. 3A and B). The Ag2S NPs treatment did not significantly affect the CFUs counts, in either culture medium, although, a slight reduction was observed 3.3. Effects on carbon source consumption after 28 days of exposure [less 3.4% and 5.9% of Log10(CFU.g− 1 soil) for NA and NBRIP culture media, respectively]. In contrast, the AgNO3 The Ag2S NPs effects on carbon substrate utilization were assessed treatment resulted in a significant (Dunnett’s Method: df = 0.1; using the Biolog® Ecoplate and are shown in Fig. 5(A–C). q = 6.363; p < 0.001) reduction of 10% of the CFUs counted in NA The analysis of AUC (Fig. 5A) revealed an over-time significant in­ medium after 28 days (Fig. 3A). For P-SB an even stronger impact of crease on substrate consumption for non-exposed soils (AUC: CT AgNO3 was observed (Fig. 3B), corresponding to a significant reduction 0 days < CT 14 days; Tukey HSD: F = 554.985, p < 0.001) until 14 of 71.3% (Dunnett’s Method: df = 0.66; q = 10.7; p < 0.001) and 91.8% days, and a slight decrease between 14 and 28 days of incubation (AUC: of Log10 (CFU.g− 1 soil) (Dunnett’s Method: df = 0.918; q = 14.691; CT 14 days > CT 28 days; Tukey HSD: F = 554.985, p = 0.915). p < 0.001), at sampling day 0 and 28 after the beginning of the meso­ Time-dependent effects on AUC calculated for soils exposed to silver cosm trial, respectively. were observed. A significant inhibitory effect of Ag2S NPs on L-arginine consumption (Fig. 5B) was observed after 28 days (p < 0.05, Table S2). On the other hand, after 14 days the AgNO3 treatment caused a signif­ icative reduction (p < 0.05, Table S2) of the consumption of the following substrates: tween 80, L-erythritol, D-glucosaminic acid and L- Fig. 2. Silver content in soils (A) and silver dissolution into soil porewater (B), after 0 (2 days after soil spiking) and 28 days of exposure. The data was expressed as mg (Ag) kg− 1 of soil (A) and as µg (Ag) L− 1 of porewater [average (n = 3) ± standard deviation]. The treatments included: the non-exposed soil (CT), soil exposed to Ag2S NPs and soil exposed to AgNO3. Different letters (a,b) indicate statistical significance (p < 0.05) between soil treatments and the time of exposure, using the two- way ANOVA (Tukey post-hoc test, p < 0.05). 5 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 Fig. 3. Soil colony forming units (CFUs) measured in Lufa 2.2 soil, sampled at day 0 and 28 from the mesocosm experiment, previously spiked with AgNO3 and Ag2S NPs. CFUs are presented as % of variation on CFUs (Log10 g− 1 soil) (± standard deviation) from the respective control (non-exposed soil), n = 3. Heterotrophic bacteria in nutrient agar (NA) medium (A), and P-solubilizing bacteria in National Botanical Research Institute Phosphate (NBRIP) medium (B). The grey bars represent the soil samples collected at the beginning of the mesocosm soil experiment (day 0), and the white bars represent samples after 28 days. Asterisks (*) indicate significant differences using two-away ANOVA (Dunnett’s post-hoc test), p < 0.05) relative to the respective control. Fig. 4. Soil enzymatic activity measured in Lufa 2.2 soil, sampled at days 0, 14 and 28 from the mesocosm experiment, previously spiked with AgNO3 and Ag2S NPs. Also, the non-exposed soil was included in this experiment (CT). The data represent the mean (n = 3) of enzymatic activity (± standard deviation) of dehydrogenase (A), β-glucosidase (B), acid phosphatase (C) and arylsulfatase (D) activities. Different letters (a, b) indicate a significant difference using two-away ANOVA (Tukey post-hoc test, p < 0.05), among treatments. serine. At day 28 AgNO3 treatment resulted in a significant decrease of between sampling times but not between treatments versus control. the consumption of pyruvic acid methyl ester, β-methyl-D-glucoside, Considering exposure time, SAWCD results were in accordance with and L-arginine (p < 0.05, Table S4). AUC, with the pattern of substrate utilization being 0 days > 14 days In terms of SAWCD analysis, a significant effect was observed > 28 days of exposure, for phenolic acid and amides/amines classes. For 6 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 Fig. 5. Community level and physiological profile (CLPP) measured in Lufa 2.2 soil, sampled at days 0, 14 and 28 from the mesocosm experiment, previously spiked with AgNO3 and Ag2S NPs. For further comparison the non-exposed soil (CT) was included. The area under the curve (AUC) was calculated based on a trapezoidal integration function (A) over time; (B) after day 28, the Pyruvic acid methyl ester, β-methyl-D-glucoside, and L-arginine consumption; and (C) substrate average well colour development (SAWCD) after 14 and 28 days of exposure. The data represent the mean (n = 3) of AUC ( ± standard deviation). Letters (a, b) represent sig­ nificant differences using two-away ANOVA (Tukey post-hoc test, p < 0.05) between treatments. the other classes (Fig. 5C) a significant difference over time was not 3.5. Relationship between structure and functions of SBC observed. Due to late-term effects of Ag2S NPs on SBC, at structural level, the 3.4. Effects on SBC structure and diversity PCoA based on relative band intensity from DGGE profiles (Fig. 6) was performed only considering the 28 days of exposure (Fig. 7). Moreover, The impact of Ag2S NPs and AgNO3 on the SBC structure was different functional parameters were imposed as a vector, to obtain the assessed over time and results are presented in Fig. 6 (A–B). correlation between both levels (structural and functional). Thus, the The cluster analysis (Fig. 6A) showed a temporal effect among soil association between SBC structure and functional variables was assessed treatments, in which the Ag2S NPs treatment was completely separated for samples collected at 28 days of exposure and results are shown in from the other treatments only at day 28. The soil treated with AgNO3 Fig. 7. conversely revealed a strong impact on SBC from the experiment start As explained above, a significant difference among treatments was onwards. revealed. The two main PCoA axes (Fig. 7) explained 65.1% of the total The first two axis of the principal coordinate analysis (PCoA) variation [PCO1 axis (42.5%) and PCO2 axis (22.6%)]. The PCO1 axis explained 49.6% of the total variation in the data: PCO1 axis (32.1%) separated soil samples according to treatment. The highest negative and PCO2 axis (17.5%) (Fig. 6B). Along the PCO1 axis, samples tend to correlation with axis 1 was determined for AP activity (Pearson corre­ cluster as a function of exposure time while along the PCO2 axis sepa­ lation: R = − 0.79, p = 0.023), consistent with a significantly increased ration seems to be influenced by treatment. At 28 days of exposure, a AP activity in soils exposed to Ag2S NPs in comparison to other treat­ clear separation between soil exposed to Ag2S NPs from the other ments (Fig. 4C). On the other hand, the axis 2 separated the soils exposed treatments was observed. to AgNO3 from the other treatments, and the highest negative correla­ The Permanova analysis confirmed a significant spatial separation tions with this axis were observed for βG, heterotrophic bacteria and P- between soil treatments (CT, AgNO3 and Ag2S NPs) according to the SB (Pearson correlation: R = − 0.67, p = 0.144; R = − 0.74, p = 0.796; time of exposure (day 0, 14, 28) (PERMANOVA: F = 4.76; P = 0.001), R = − 0.87, p = 0.864, respectively). The other functional variable except for non-exposed soils that did not show a significant difference, in assessed correlated with both axes, namely L-arginine (PCO1: R = 0.35 spatial separation, between day 14 and 28 (PERMANOVA: CT14 vs. and PCO2: R = − 0.50; p = 0.949). CT28: t = 1.93; P = 0.052). In terms of diversity indices, a significant decrease in richness and 4. Discussion Shannon-Wiener diversity was induced by exposure to Ag2S NPs after 28 days (p < 0.05; Table 1). On the other hand, significant effects of AgNO3 Currently, there is a lack of research data indicating the potential were only observed in richness after 48 h of soil spiking (Table 1), which effects of Ag2S NPs on SBC, under realistic scenarios. There is also corresponds to the day 0 of mesocosm experiment (Fig. 1), suggesting insufficient knowledge on the effects of Ag2S NPs compared to those of binding of the Ag+ to soil and thus lowered bioavailability. Ag ionic forms. To address these knowledge gaps, we conducted an 7 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 Fig. 6. Denaturing gradient gel electrophoresis (DGGE) analysis of the V3 region of the 16S rRNA gene amplified from SBC from non-exposed soil (CT), soil exposed to Ag2S NPs and AgNO3, after 0, 14 and 28 days of exposure. Cluster analysis (A) and Principal component analysis (PCoA) (B) were constructed based on Bray-Curtis similarity after square root transformation of DGGE relative abundance data (n = 3). Table 1 Diversity indexes from SBC structure analysis were analysed over time and considering the following soil treatments: CT (non-exposed soil), and spiked soil with AgNO3 and Ag2S NPs. Diversity indexes Soil Treatment Sampling time (day) S J’ H’ a a CT 0 44 ± 1 0.922 ± 0.003 3.483 ± 0.022a AgNO3 0 41 ± 2a 0.897 ± 0.006a 3.321 ± 0.051b Ag2S NPs 0 44 ± 1a 0.919 ± 0.004a 3.469 ± 0.030a CT 14 48 ± 1a 0.911 ± 0.006a 3.527 ± 0.038a AgNO3 14 45 ± 1ab 0.903 ± 0.005a 3.431 ± 0.020a Ag2S NPs 14 44 ± 2b 0.898 ± 0.013a 3.403 ± 0.076a CT 28 45 ± 1a 0.888 ± 0.028a 3.380 ± 0.091a AgNO3 28 43 ± 1a 0.911 ± 0.014a 3.428 ± 0.070a Ag2S NPs 28 37 ± 2b 0.885 ± 0.018a 3.193 ± 0.042b The diversity indexes included: the richness (S), Pielou (J’) and Shannon-Wiener index (H’). The data represent the mean of each diversity index (n = 3 ± standard deviation). Different letters (a, b) indicate statistical difference using two-away ANOVA (Tukey post-hoc test, p < 0.05) among treatments and within exposure time. 8 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 Fig. 7. Principal coordinates analysis (PCoA) ordinations based on Bray-Curtis similarity showing the spatial distribution of SBC struc­ tures among the different soil treatments: non- exposed soil (CT) and exposed soil to Ag2S NPs or AgNO3 (n = 3), after 28 days of expo­ sure. The vectors represent a functional variable (Black: enzymatic activity and CLPP) or a chemical variable (Red: silver ion presence (Fig. 2) in soil pore-water) that presented a correlation ≥ 0.6 with SBC structure: AP - Acid Phosphatase; βG - β-Glucosidase; HB - Hetero­ trophic bacteria; P-SB - Phosphatase- Solubilizing Bacteria. indoor mesocosm experiment to evaluate the impact of Ag2S NPs on SBC DHA in soils exposed to citrate-coated AgNPs at different concentrations during 28 days, at a nominal exposure concentration of 10 mg Ag kg− 1 (0, 1, 10, 100, and 1000 µg g− 1). Also, McGee et al. (2017), using a lower soil. Structure and function of AgNO3− and Ag2S NPs-exposed SBC were Ag concentration (1 mg kg− 1soil) and a non-sulfidized form of silver compared to those of communities in the control soil and related to the [AgNPs (20–30 nm and 0.5–1 µm)], observed a similar sensitivity of content/dissolution of Ag+ in the soil pore water over time. DHA. Although these studies demonstrated a significant impact of AgNPs on DHA activity, our study used the AgNPs in sulfidized form (Ag2S NPs), which presents distinct physical and chemical properties. 4.1. Effects of silver on SBC function Consequently, the Ag+ dissolution is also different; for instance, Ag2S NPs were shown to be more stable than AgNPs (Wang et al., 2016). Colony bacterial counts are considered a suitable indicator for Additionally, Ag2S NPs particles can be aggregated in soils, resulting in measuring Ag2S NPs effects on SBC, as they usually present a fast increased particle sizes (micro-particles) (Liu et al., 2019), with less response to metal contamination (Vasileiadis et al., 2015). Our study surface area for oxidative dissolution and consequently decreased demonstrated that Ag2S NPs did not significantly affect counts of het­ toxicity (Liu et al., 2019). The study by Liu et al. (2019) reported that erotrophic bacteria and P-SB, albeit after 28 days of exposure a slight hetero-aggregation, tested by combining the exposure of TiO2 and reduction was observed. Prior studies demonstrated that Ag2S NPs are AgNPs in soils, exhibit a higher toxic effect than the homo-aggregation sparingly soluble in soils over a long period (> 1 year) (Wang et al., of each of the single TiO2 or Ag NPs. Although this was discussed 2016), which may justify our results. However, the toxic effect of Ag2S based on the size of aggregates, the joint toxicity of a binary combina­ NPs on specific bacterial groups was detected in previous studies and tion of metals cannot be disregarded. related to their sulfidation degree (Schultz et al., 2018). For instance, a Besides, DHA is only expressed in intact cells and does not accu­ lower degree of sulfidation of Ag2S NPs was reported to result in faster mulate extracellularly in the soils (Dick, 1996), so its activity has been dissolution and release of Ag+, thus enhancing the toxic effects of Ag2S related with microbial biomass, respiratory activity, microbial oxidative NPs on SBC (Wang et al., 2017; Schlich et al., 2018). A similar trend was activity, and microbial viability (Dick, 1996; McGee et al., 2017; also documented for Escherichia coli (Reinsch et al., 2012); Arthrobacter Samarajeewa et al., 2017). Accordingly, in our study the lower abun­ globiformis and Pseudomonas putida (Schultz et al., 2018). The study by dance of heterotrophic bacteria (NA medium) in AgNO3-exposed soils Schultz et al. (2018) confirmed that the AgNO3 treatment caused a may explain at least in part the decreased DHA activity. higher toxic effect on bacterial counts (Arthrobacter globiformis and βG activity has been also considered as an indicator of soil quality Pseudomonas putida) when compared to exposures to several NPs: Ag2S (Turner et al., 2002), related to carbon cycling in soils. Thus, the hy­ (36 nm), Ag-PVP (49 nm) or Ag citrate (58 nm). These distinct toxic drolysis products resulting from βG activity are an essential energy effects on SBC might be related with different free Ag+ concentrations, source for microorganisms in soils (Dick, 1996). In accordance, studies in which a higher free Ag+ concentrations resulting from AgNO3 was from Kim et al. (2018) showed that exposure to silver-graphene oxide documented that arising from AgNPs dissolution in soil. In accordance, (0.1–1 mg g− 1 soil) decreased the βG activity up to 80% in soils. Simi­ our study demonstrated this distinct toxic effect, considering the silver larly, Samarajeewa et al. (2017) showed a significant decrease of this form present, on bacterial counts. The effect of AgNO3 on P-SB may enzyme’s activity upon exposure to 49 mg of AgNPs per kg of soil. In our indicate an impairment of phosphorous solubilization in soils, with study, a decrease on βG activity was observed for both treatments after consequences for crop production and soil fertility (Richardson and 28 days of exposure, with approximately 56% and 33% decrease for the Simpson, 2011). AgNO3 and Ag2S NPs treatments, respectively. Ag+ dissolution did not To assess the impact of nanomaterials on soil functions it is essential explain the effect of Ag2S NPs, and some specific-particle effects might to evaluate the enzymatic activities related to nutrient cycling (Samar­ occur, as suggested Starnes et al. (2015). In fact, the ratio between the ajeewa et al., 2017). DHA is one of the most sensitive enzymes to metal silver total content in soil (Fig. 2A) and the dissolved/dissociated Ag+ in contamination (Samarajeewa et al., 2017). In our study, only the AgNO3 the pore water (Fig. 2B) showed the highest value for Ag2S NPs (19.8) treatment caused a strong inhibition of DHA activity over time, a result than for AgNO3 (6.15), suggesting a specific-particle effect for the Ag2S once again probably related with distinct free Ag+ concentrations. The NPs treatment. Although previous studies reported a decrease of AP investigation conducted by Shin et al. (2012), found a high sensitivity of 9 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 activity in soil exposed to AgNPs, regardless of the concentrations tested soils and may release exudates that can directly change the soil condi­ (Colman et al., 2013; Samarajeewa et al., 2017), our study showed a tions (pH, for instance) (Liu et al., 2019), affecting the toxicity of AgNO3 significant increase in this enzymatic activity, after 28 days of exposure and Ag2S NPs on SBC. Also, invertebrates may alter soil conditions, to 10 mg kg− 1 of Ag2S NPs. Although our study did not evaluate the root essentially through their behaviour in soils (Tourinho et al., 2015), biomass, we hypothesize that root growth and proliferation in soils may contributing to the microbial composition/abundance (decreasing, for contribute to the increased AP activity in our experiment. On the other instance) changes, as a result of the inhibition of oxygenation level. hand, the presence of different invertebrates possibly contributes to the complexation of the Ag2S NPs in soils (e.g. ligation with organic matter 4.2. Effects of silver on SBC structure and diversity or with clay), and therefore, protected/stimulate the activity of extra­ cellular enzymes (Shin et al., 2012). The effects of Ag2S NPs on the structure, composition, and diversity Although study from Vasileiadis et al. (2018) demonstrated that the of SBC were previously assessed in only a limited number of studies AS enzyme is more sensitive to Ag+ exposure (from AgNO3 or acetate, (Judy et al., 2015; Doolette et al., 2016). The present study is a respectively), our results corroborate this AS sensitivity to the AgNO3 considerable contribution given its higher tier approach, using an indoor treatment but only after 14 days. After this period this enzymatic ac­ mesocosm including several species of invertebrates and a plant species tivity recovered, becoming similar to that in the non-exposed soil. This making them more representative of field conditions. Furthermore, our time-dependent effect might have been directly related to a rapid flux of study revealed a distinct impact of Ag2S NPs and AgNO3 on the SBC free Ag+ from AgNO3 in the first days (short period) or by directly structure, in a time-dependent manner. During the exposure, a faster affecting the sulfatase producer microorganisms (Vasileiadis et al., increase of Ag+ concentration in the soil pore water from the AgNO3 2018) followed by a gradually stronger adsorption of Ag+ to organic S treatment was observed and might be responsible for an early structural groups or even a reduction to metallic Ag0 making Ag+ less available change, as previously explained. Although Ag+ dissolution from Ag2S (Settimio et al., 2014). NPs was detected in very low amounts at 28 days of exposure, this Ag+ Some fluctuation in soil enzymatic activity was observed, even in dissolution wasn’t enough to explain the latter’s effects on SBC non-exposed soil (CT), over time. The SBC is dynamic, and the soil structure. conditions (e.g. daily watering) and the invertebrates and plants (root The impact of Ag2S NPs on SBC structure has been already evaluated growth) presence in our experiment might influence this fluctuation. on natural soils after 28 days using metagenomics (Doolette et al., 2016) Soil water content influences the SBC composition, resulting in an and phospholipid fatty acids (PLFAs) analysis (Judy et al., 2015). In this impact on oxygen concentrations and nutrient availability. The watering regard, Doolette et al. (2016) observed that the impact of Ag2S NPs might reduce soil oxygen levels and promote anaerobic microorganisms (152 nm) on the number of operational taxonomic units (OTUs) was (facultative and obligate) growth, whilst decreasing the total microbial lower compared to non-sulfidized AgNPs (44 nm) and AgNO3, when activity (Sylvia et al., 2005). On the contrary, the presence of in­ using the toxicity threshold hazardous concentrations (HC20) approach vertebrates can promote soil aeration (El-Wakeil, 2015), which indi­ [5.9 (Ag2S NPs); 1.4 (AgNPs) and 1.4 (AgNO3) mg Ag kg− 1]. On the rectly influences the proliferation of different bacteria/fungi, ultimately other hand, the study by Judy et al. (2015) demonstrated that microbial resulting in a structural and composition change of the SBC. Moreover, biomass (inferred from PFLA analysis) decreased when Ag2S NPs these organisms have a crucial role in the decomposition of the organic (30.1 nm) was applied in sandy loam soils (pH=6.8), at 1 and matter. Soil and organic matter are also ingested, digested and excreted 100 mg kg− 1soil, however, the silver in sulfidized form was less toxic by invertebrates, producing a substrate that is more accessible to SBC compared to the other silver NP (PVP-Ag - 20.6 mm) and ionic silver but that may also contain Ag (as free Ag+, and/or Ag2S depending on the (AgNO3). Although Ag toxicity is strongly dependent on soil character­ treatment plus some or all of the range of speciation forms noted above), istics (e.g. pH, soil type), nanomaterial characteristics (e.g. size, inducing different microbial mobilizations. Indirectly, this can induce coating), and the specific microbial methodologies (Samarajeewa et al., bacterial community adjustments, in terms of abundance and compo­ 2017; Vasileiadis et al., 2018), both studies showed a similar toxicity sition, according to the bacteria nutrient/metabolic requirements. Be­ pattern (Ag2S NPs < AgNPs < AgNO3) on microbial structure. sides, plants and associated microorganisms alter the physicochemical Moreover, a reduction of the alpha diversity of the bacterial com­ properties of the root-soil interface, such as pH and redox conditions munity exposed to AgNPs has already been reported (Samarajeewa (Liu et al., 2019). et al., 2017). Regardless of silver form, our study revealed a significant The physiological profiles were used to assess the impact of both Ag reduction in diversity and richness, after 14 days of exposure. In fact, the forms on bacteria consortia. The significant reduction in L-arginine free Ag+ concentration might explain the AgNO3 impact on these in­ consumption (amino acid class) for both silver forms, indicates a po­ dexes, as above mentioned. In our experiment, the negative impact of tential effect on carbon and nitrogen cycles that are closely related to Ag2S NPs on diversity indexes may result from the Ag+ released (from plant fertility success. L-arginine plays a key role in nitrogen distribution these particles), or the accumulation of sulfidized particles in the plant and recycling in plants (Slocum, 2005), and consequently, might be tissues (Wang et al., 2017). In fact, this accumulation (at 10 mg kg− 1) related to the plant homeostasis. The study by Yang et al. (2018) reduces plant growth (Wang et al., 2017), changing the interaction be­ demonstrated that AgNPs decreased the arginine and histidine content tween root exudates and microbes, which can result in an indirect in wheat (Triticum aestivum L.), during 4 months in 2000 mg kg− 1 change in the abundance of microorganisms in the soil. After day 28, the amended soil. On the other hand, the decrease of L-arginine consump­ effect of AgNO3 on SBC diversity and richness was gradually attenuated tion might be partially explained by the decreased bacterial counts to values similar to the control, suggesting SBC resilience to this silver observed in our results, at least for AgNO3, because heterotrophic bac­ form or/and a decrease in Ag bioavailability. This pattern has been teria are able to catabolize the soil arginine, involved with ammonifi­ observed before for invertebrate species, where Ag+ from AgNO3 cation processes (nitrogen and carbon cycles) (Alef and Kleiner, 1986). decreased significantly over time in porewater, reflecting a decrease of Other studies reported a significant decrease of general carbon substrate Ag bioavailability. Due to the increased silver adsorption to soils, the utilization (AWCD) by the microbial community from soils exposed to free Ag+ concentration decreases in less than 14 days (Settimio et al., AgNPs (uncoated, 20 nm) (Sillen et al., 2015) or decreased utilization of 2014). specific substrates such as amino acids and carbohydrates (Kumar et al., 2011), for soils exposed to metal NPs. Our results did not indicate a 4.3. Environmental relevance and integrative analysis between structure significant impairment of SBC activity and the effects attenuation may and functional levels of SBC be related to the presence of different soil invertebrates and plants in this mesocosms approach. Plants may change the availability of metals on Shifts of SBC structure may lead to a change in the biological 10 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 function, albeit the functional redundancy reported for SBC (Vasileiadis Environment; grant no. 646002). SL, IH, SP and ZK received additional et al., 2015). The relationship between the SBC structure and soil financial support from Fundação para a Ciência e a Tecnologia/Minis­ function is not easy to predict and is poorly documented. Our results tério da Ciência, Tecnologia e Ensino Superior (FCT/MCTES), through demonstrated that Ag2S NPs change the SBC structure and this structural national funds, to CESAM (UIDP/50017/2020 + UIDB/50017/2020). change may affect, for instance, the phosphorous cycle, resulting in an Also, this work was supported by FCT through a PhD grant to Sara increased AP activity. Also, the P-SB counts showed a good correlation Peixoto (SFRH/BD/117738/2016). (R>0.6) with the SBC structure from soils exposed to Ag2S NPs. In accordance with these findings, the study by Fanin et al. (2015) showed Appendix A. Supporting information a significant correlation (Mantel test) between CLPPs and SBC structure from different fertilization levels (N and P input) in undisturbed tropical Supplementary data associated with this article can be found in the forests. online version at doi:10.1016/j.ecoenv.2020.111405. Few studies are available that relate the N cycle with the SBC structure exposed to metal contamination (Kumar et al., 2014; Doolette References et al., 2016), and other nutrient cycles are even less explored. Never­ theless, the effects of AgNPs (at 10 and 100 mg kg− 1) on SBC structure Alef, K., Kleiner, D., 1986. Arginine ammonification, a simple method to estimate microbial activity potentials in soils. Soil Biol. Biochem. 18 (2), 233–235. might impact the carbon cycle, through a decrease in microbial biomass Colman, B.P., Arnaout, C.L., Anciaux, S., Gunsch, C.K., Hochella Jr, M.F., Kim, B., (carbon) (Kumar et al., 2014), soil enzymatic activity (e.g. β-glucosi­ Unrine, J.M., 2013. Low concentrations of silver nanoparticles in biosolids cause dase) and/or soil microbial respiration. adverse ecosystem responses under realistic field scenario. PloS One 8 (2), e57189. Cornelis, G., Kirby, J.K., Beak, D., Chittleborough, D., McLaughlin, M.J., 2010. A method for determination of retention of silver and cerium oxide manufactured 5. Conclusions nanoparticles in soils. Environ. Chem. 7 (3), 298–308. Dick, R.P., Breakwell, D.P., Turco, R.F., 1996. Soil enzyme activities and biodiversity measurements as integrative microbiological indicators. Methods for Assessing Soil This study provided evidence that Ag2S NPs affect the SBC and that Quality. SSSA 49th, Madison, USA, pp. 247–271. these effects become more pronounced over-time. Ag2S NPs effects on Doolette, C.L., Gupta, V.V., Lu, Y., Payne, J.L., Batstone, D.J., Kirby, J.K., McLaughlin, M. the SBC are expected to potentially imbalance soil functions, namely J., 2016. Quantifying the sensitivity of soil microbial communities to silver sulfide nanoparticles using metagenome sequencing. PLoS One 11 (8), 1–20. those related to the carbon (L-arginine consumption and β-glucosidase Doolette, C.L., McLaughlin, M.J., Kirby, J.K., Navarro, D.A., 2015. Bioavailability of activity) and phosphorus cycles (P-SB and PA). The structure of SBC and silver and silver sulfide nanoparticles to lettuce (Lactuca sativa): effect of agricultural diversity indexes were significantly impacted by Ag2S NPs, after 28 days amendments on plant uptake. J. Hazard. Mater. 300, 788–795. of exposure. El-Wakeil, K.F.A., 2015. Effects of terrestrial isopods (Crustacea: Oniscidea) on leaf litter decomposition processes. J. Basic Appl. Zool. 69, 10–16. Function and structure of SBC were also impacted by AgNO3, pre­ European Commission Science for Environment Policy (2017) In-Depth Report 14 dominantly at earlier times of exposure due to the initial high concen­ Assessing the environmental safety of manufactured nanomaterials Science for tration of free Ag+, with the expected decrease of Ag bioavailability over Environment Policy Environment. Fanin, N., Hättenschwiler, S., Schimann, H., Fromin, N., 2015. Interactive effects of C, N time, as a result of binding to lignin and clay particles, being a likely and P fertilization on soil microbial community structure and function in an cause for the weakening of effects. Amazonian rain forest. Funct. Ecol. 29 (1), 140–150. This study emphasizes the usefulness of the terrestrial mesocosm Garland, J.L., 1997. Analysis and interpretation of community-level physiological profiles in microbial ecology. FEMS Microbiol. Ecol. 24 (4), 289–300. approach to evaluate the effects of sulfidation state of AgNPs on mi­ Geitner, N., Hendren, C., Cornelis, G., Kaegi, R., Lead, J., Lowry, G., Lynch, I., crobial parameters. Moreover, the observed temporal effects (28-days) Nowack, B., Petersen, E., Bernhardt, E., Brown, S., Chen, W., de Garidel-Thoron, C., from both ionic Ag and Ag2S NPs highlight the importance of long-term Hanson, J., Harper, S., Jones, K., von der Kammer, F., Kennedy, A., Kidd, J., Matson, C., Metcalfe, C., Pedersen, J., Peijnenburg, W., Quik, J., Rodrigues, S.M., exposure experiments under realistic environmental conditions. Rose, J., Sayre, P., Simonin, M., Svendsen, C., Tanguay, R., Tufenkji, N., van Teunenbroek, T., Thies, G., Tian, Y., Rice, J., Turner, A., Liu, J., Unrine, J., CRediT authorship contribution statement Vance, M., White, J., Wiesner, M., 2020. Harmonizing across environmental nanomaterial testing media for increased comparability of nanomaterial datasets. Environ. Sci. Nano 7, 13–36. S. Peixoto: Methodology, Formal analysis, Investigation, Writing - Giese, B., Klaessig, F., Park, B., Kaegi, R., Steinfeldt, M., Wigger, H., Gottschalk, F., 2018. original draft, Visualization. Z. Khodaparast: Methodology, Investiga­ Risks, release and concentrations of engineered nanomaterial in the environment. Sci. Rep. 8 (1), 1–18.. tion. G. Cornelis: Methodology, Validation, Writing - review & editing. Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009. Modeled environmental E. Lahive: Methodology, Validation, Writing - review & editing. A. concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, fullerenes) for Green Etxabe: Methodology, Writing - review & editing. M. Baccaro: different regions. Environ. Sci. Technol. 43 (24), 9216–9222. Methodology, Writing - review & editing. A.G. Papadiamantis: Vali­ Henriques, I.S., Almeida, A., Cunha, Â., Correia, A., 2004. Molecular sequence analysis of prokaryotic diversity in the middle and outer sections of the Portuguese estuary Ria dation, Formal analysis, Investigation. S.F. Gonçalves: Methodology, de Aveiro. FEMS Microbiol. Ecol. 49 (2), 269–279. Formal analysis. I. Lynch: Validation, Investigation, Writing - review & Judy, J.D., Kirby, J.K., Creamer, C., McLaughlin, M.J., Fiebiger, C., Wright, C., Bertsch, P. editing. M. Busquets-Fite: Methodology, Formal analysis, Investiga­ M., 2015. Effects of silver sulfide nanomaterials on mycorrhizal colonization of tomato plants and soil microbial communities in biosolid-amended soil. Environ. tion. V. Puntes: Validation, Investigation. S. Loureiro: Conceptualiza­ Pollut. 206, 256–263. tion, Validation, Investigation, Resources, Writing - review & editing, Kaegi, R., Voegelin, A., Sinnet, B., Zuleeg, S., Hagendorfer, H., Burkhardt, M., Supervision, Funding acquisition. I. Henriques: Conceptualization, Siegrist, H., 2011. Behavior of metallic silver nanoparticles in a pilot wastewater treatment plant. Environ. Sci. Technol. 45 (9), 3902–3908. Validation, Formal analysis, Investigation, Writing - review & editing, Kent, R.D., Oser, J.G., Vikesland, P.J., 2014. Controlled evaluation of silver nanoparticle Visualization, Supervision, Funding acquisition. sulfidation in a full-scale wastewater treatment plant. Environ. Sci. Technol. 48 (15), 8564–8572. Kim, M.J., Ko, D., Ko, K., Kim, D., Lee, J.Y., Woo, S.M., Chung, H., 2018. Effects of silver- Declaration of Competing Interest graphene oxide nanocomposites on soil microbial communities. J. Hazard. Mater. 346, 93–102. The authors declare that they have no known competing financial Kumar, N., Palmer, G.R., Shah, V., Walker, V.K., 2014. The effect of silver nanoparticles on seasonal change in arctic tundra bacterial and fungal assemblages. PLoS One 9 interests or personal relationships that could have appeared to influence (6), e99953. the work reported in this paper. Kumar, N., Shah, V., Walker, V.K., 2011. Influence of a nanoparticle mixture on an arctic soil community. Environ. Toxicol. Chem. 31 (1), 131–135. Acknowledgements Lahive, E., Matzke, M., Durenkamp, M., Lawlor, A.J., Thacker, S.A., Pereira, M.G., Lofts, S., 2017. Sewage sludge treated with metal nanomaterials inhibits earthworm reproduction more strongly than sludge treated with metal metals in bulk/salt forms. All the authors were funded by the European Union Horizon 2020 Environ. Sci. Nano 4 (1), 78–88. project Nano-FASE (Nanomaterial Fate and Speciation in the 11 S. Peixoto et al. Ecotoxicology and Environmental Safety 206 (2020) 111405 Liu, J., Williams, P.C., Goodson, B.M., Geisler-Lee, J., Fakharifar, M., Gemeinhardt, M.E., Settimio, L., McLaughlin, M.J., Kirby, J.K., Langdon, K.A., Lombi, E., Donner, E., 2019. TiO2 nanoparticles in irrigation water mitigate impacts of aged Ag Scheckel, K.G., 2014. Fate and lability of silver in soils: effect of ageing. Environ. nanoparticles on soil microorganisms, Arabidopsis thaliana plants, and Eisenia fetida Pollut. 191, 151–157. earthworms. Environ. Res. 172, 202–215. Shin, Y.J., Kwak, J.I., An, Y.J., 2012. Evidence for the inhibitory effects of silver McGee, C.F., Storey, S., Clipson, N., Doyle, E., 2017. Soil microbial community responses nanoparticles on the activities of soil exoenzymes. Chemosphere 88 (4), 524–529. to contamination with silver, aluminium oxide and silicon dioxide nanoparticles. Sillen, W.M., Thijs, S., Abbamondi, G.R., Janssen, J., Weyens, N., White, J.C., Ecotoxicology 26 (3), 449–458. Vangronsveld, J., 2015. Effects of silver nanoparticles on soil microorganisms and McGillicuddy, E., Murray, I., Kavanagh, S., Morrison, L., Fogarty, A., Cormican, M., maize biomass are linked in the rhizosphere. Soil Biol. Biochem. 91, 14–22. Morris, D., 2017. Silver nanoparticles in the environment: Sources, detection and Slocum, R.D., 2005. Genes, enzymes and regulation of arginine biosynthesis in plants. ecotoxicology. Sci. Total Environ. 575, 231–246. Plant Physiol. Biochem. 43 (8), 729–745. Muyzer, G., Waal, E., C., Uitterlinden, A., G., 1993. Profiling of complex microbial Starnes, D.L., Unrine, J.M., Starnes, C.P., Collin, B.E., Oostveen, E.K., Ma, R., Tsyusko, O. populations by denaturing gradient gel electrophoresis analysis of polymerase chain V., 2015. Impact of sulfidation on the bioavailability and toxicity of silver reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59 (3), nanoparticles to Caenorhabditis elegans. Environ. Pollut. 196, 239–246. 695–700. Sylvia, D.M., Fuhrmann, J.J., Hartel, P.G., Zuberer, D.A., 2005. Principles and Nautiyal, S.C., 1999. An efficient microbiological growth medium for screening Applications of Soil Microbiology (No. QR111 S674 2005). Pearson. phosphate solubilizing microorganisms. FEMS Microbiol. Lett. 170 (436), 265–270. Tabatabai, M.A., 1994. Methods of soil analysis: microbiological and biochemical Neves, J., Cardoso, D.N., Malheiro, C., Kah, M., Soares, A.M., Wrona, F.J., Loureiro, S., properties. Soil Enzymes. SSSA Book series,, pp. 775–883. 2019. Copper toxicity to Folsomia candida in different soils: a comparison between Tourinho, P.S., van Gestel, C.A., Jurkschat, K., Soares, A.M., Loureiro, S., 2015. Effects of nano and conventional formulations. Environ. Chem. 16 (6), 419–429. soil and dietary exposures to Ag nanoparticles and AgNO3 in the terrestrial isopod Pachapur, V.L., Larios, A.D., Cledón, M., Brar, S.K., Verma, M., Surampalli, R.Y., 2016. Porcellionides pruinosus. Environ. Pollut. 205, 170–177. Behaviour and characterization of titanium dioxide and silver nanoparticles in soils. Turner, B.L., Hopkins, D.W., Haygarth, P.M., Ostle, N., 2002. β-Glucosidase activity in Sci. Total Environ. 563–564, 933–943. pasture soils. Appl. Soil Ecol. 20 (2), 157–162. Richardson, A.E., Simpson, R.J., 2011. Soil microorganisms mediating phosphorus US-EPA, 1996. Method 3052: Microwave Assisted Digestion of Soloceous and availability update on microbial phosphorus. Plant Physiol. 156 (3), 989–996. Organically Based Matrices. United States Environmental Protection Agency,, Reinsch, B., Levard, C., Li, Z., Ma, R., Wise, A., Gregory, K.B., Lowry, G.V., 2012. Washington. Sulfidation of silver nanoparticles decreases Escherichia coli growth inhibition. Vasileiadis, S., Brunetti, G., Marzouk, E., Wakelin, S., Kowalchuk, G.A., Lombi, E., Environ. Sci. Technol. 46 (13), 6992–7000. Donner, E., 2018. Silver toxicity thresholds for multiple soil microbial biomarkers. Sala, M.M., Pinhassi, J., Gasol, J.M., 2006. Estimation of bacterial use of dissolved Environ. Sci. Technol. 52 (15), 8745–8755. organic nitrogen compounds in aquatic ecosystems using Biolog plates. Aquat. Vasileiadis, S., Puglisi, E., Trevisan, M., Scheckel, K.G., Langdon, K.A., McLaughlin, M.J., Microb. Ecol. 42 (1), 1–5. Donner, E., 2015. Changes in soil bacterial communities and diversity in response to Samarajeewa, A.D., Velicogna, J.R., Princz, J.I., Subasinghe, R.M., Scroggins, R.P., long-term silver exposure. FEMS Microbiol. Ecol. 91 (10), 1–11. Beaudette, L.A., 2017. Effect of silver nano-particles on soil microbial growth, Wang, P., Lombi, E., Sun, S., Scheckel, K.G., Malysheva, A., McKenna, B.A., Kopittke, P. activity and community diversity in a sandy loam soil. Environ. Pollut. 220, M., 2017. Characterizing the uptake, accumulation and toxicity of silver sulfide 504–513. nanoparticles in plants. Environ. Sci. Nano 4 (2), 448–460.. Schlich, K., Hoppe, M., Kraas, M., Schubert, J., Chanana, M., Hund-Rinke, K., 2018. Wang, P., Menzies, N.W., Dennis, P.G., Guo, J., Forstner, C., Sekine, R., Kopittke, P.M., Long-term effects of three different silver sulfide nanomaterials, silver nitrate and 2016. Silver nanoparticles entering soils via the wastewater-sludge-soil pathway bulk silver sulfide on soil microorganisms and plants. Environ. Pollut. 242, pose low risk to plants but elevated Cl concentrations increase Ag bioavailability. 1850–1859. Environ. Sci. Technol. 50 (15), 8274–8281. Schultz, C.L., Gray, J., Verweij, R.A., Busquets-Fité, M., Puntes, V., Svendsen, C., Yang, J., Jiang, F., Ma, C., Rui, Y., Rui, M., Adeel, M., Xing, B., 2018. Alteration of crop Matzke, M., 2018. Aging reduces the toxicity of pristine but not sulphidised silver yield and quality of wheat upon exposure to silver nanoparticles in a life cycle study. nanoparticles to soil bacteria. Environ. Sci. Nano 5 (11), 2618–2630. J. Agric. Food Chem. 66 (11), 2589–2597. 12