SUN Pengfei ,LI Kuiran ,YI Shaokui ,LI Hui ,and CHEN Xi,

1) Fourth Institute of Oceanography,Ministry of Natural Resources,Beihai 536000,China

2) College of Environmental Science and Engineering,Ocean University of China, Qingdao 266100,China

3) College of Marine Life Science,Ocean University of China, Qingdao 266003,China

4) College of Life Sciences, Huzhou University, Huzhou 313000,China

Abstract Silver nanoparticles (AgNPs) have been widely used in medicine and consumer products.And it enters the river in different ways,then finally converges to the ocean through the estuary.AgNPs polution can affect NO2- and N2O production by denitrifiers in aquatic system.The effects of AgNPs on denitrification activity,nitrogen transformation and nitrous oxide (N2O) emission were investigated in Dagu River Estuary (DRE) and Jiaozhou Bay (JZB).The results showed that the potential denitrification activity(PDA),NO3- and NO2- reduction rates decreased with an increase of AgNPs concentration in DRE and JZB.However,the N2O accumulation was significantly increased at AgNPs concentrations above 5 mg kg-1 in both areas,and the accumulation rate was greater in estuary than in bay (P <0.05).Moreover,the total bacterial count showed a slightly increasing trend with an increase of AgNPs concentration (P >0.05) in DRE and JZB.Importantly,the relative abundance of narG,nirS and nosZ gene in two areas decreased with the increase of AgNPs concentration,and the negative effect of AgNPs varied in order:nosZ＞nirS＞narG,inferring that the expression of denitrifying related genes could be significantly and differently inhibited by AgNPs addition.Thus,this study demonstrated that the inhibitory effect of AgNPs on different denitrification process,which may lead to the increase of inorganic nitrogen accumulation and N2O realease.This study provides a scientific basis for the further studies of AgNPs on the ecological impact mechanism and environmental effects of offshore sedimentary environment.

Key words nitrous oxide;silver nanoparticles;denitrification;estuarine and marine sediments

1 Introduction

Nitrogen (N) is a fundamental and vital element for the organisms and ecosystems (Hsiaoet al.,2014).The unreasonable application of nitrogen could not only lead to excess discharge of nitrogen,but also lead to eutrophication,deterioration of water quality,decrease of biodiversity,and destruction of aquatic habitats (Rabalais,2002).Denitrification occurs by the bacterial transformation of nitrate into nitrite,and then into nitrogen gas and other gaseous forms of nitrogen.Denitrification is a crucial microbial process in the nitrogen cycle that can permanently remove nitrogen from ecosystems by reducing nitrate (NO3-) to dinitrogen (N2) in hypoxic conditions;it plays a significant role in controlling the fate of nitrogen in estuarine and coastal systems (Jiaet al.,2016).As N2O is an intermediate during denitrification,it may affect atmosphere by acting as a potent ozone-depleting substance as well as greenhouse gas (Crutzen,1970;Dickinson and Cicerone,1986).Besides,global warming potential of N2O is 310-fold higher than that of carbon dioxide (Jiaet al.,2016).

Nanoparticles (NP) are defined as natural,incidental or manufactured material containing particles,one or more external dimensions is in the size range 1 nm–100 nm on a three-dimensional scale.Silver nanoparticles (AgNPs) are the most widely used engineered nanomaterials as biocides in products including medical,textile,home appliances,laundry additives and food packaging (Ahamedet al.,2010;Zhanget al.,2016).With the extensive use of AgNPs,AgNPs enter the river through different ways,and finally converge to the bay through the estuary area(Griffittet al.,2009).The well-documented antimicrobial properties and unique physicochemical properties of AgNPs,including a high electrical and thermal conductivity,are increasingly being applied in the areas of microelectronics and medical imaging (Boneet al.,2012).AgNPs and their released ions can interact with the cell membrane due to electrostatic attraction,and damage the integrity and fluidity of the cell membrane.Silver nanoparticles can also enter cells,damaging their structure and DNA (Samarajeewaet al.,2017).Some studies have shown that AgNPs can significantly affect the growth,activity and community diversity of bacteria in soil and sediments (Bradfordet al.,2009).Vandevoortet al.(2012) studied the NO3-reduction process of soil,and found that AgNPs could significantly inhibite the rate of the NO3-reduction,and the degree of inhibition was related to the concentration of added AgNPs.

The offshore (including estuaries and bays) areas have high concentration and complex types of pollutants.Moreover,a large number of denitrifying bacteria of different types are widely distributed in each layer of the sediment,and the metabolic efficiency of surface sediment is the highest.Besides,the surface sediments are often associated with lower oxygen and higher content of carbon and nitrogen.Therefore,the surface sediment is an important source area for the occurrence of denitrification(Yanget al.,2016).To the best of our knowledge,it remains unclear at present whether the release of AgNPs into the environment causes negative effects on denitrification in estuarine and coastal marine sediments,especially the dose effect curve of denitrification activity,the procession of NO3-removal and N2O emissions.Moreover,the estuary and bay areas are two typical habitats with different physical-chemical characteristics,and whether the responses of denitrification processes to AgNPs exposure were different or not is still unknown.We hypothesized that the AgNPs pollution could cause a inhibiting effect to different denitrification process,which may have a great impact on nitrogen cycle in the offshore areas.In this study,to provide a scientific basis for the in-depth study of AgNPs on the ecological impact mechanism and environmental effects of offshore sedimentary environment,the impact of AgNPs contamination on denitrification activity of sediments in Dagu River Estuary(DRE) and in Jiaozhou Bay (JZB) was assessed,taking the expression levels of transcripts of NO3-(narG),NO2-(nirS) and nitrous oxide reductase (nosZ) into account as well.

2 Material and Methods

2.1 Site Description,Sample Collection and Analyses

The station E (36˚10´07´´N,120˚08´11´´E) and S(36˚07´37´´N,120˚11´02´´E) in DRE and JZB were selected as the sampling sites (Fig.1).The JZB is located in southern coast of Jiaodong Peninsula,China (35˚43´–36˚18´N,120˚04´–120˚23´E),and is the home of Qingdao Port,which is one of the top 20 trading ports in the world.More than 10 rivers flow into JZB,of which the Dagu River’s length and watershed area are the longest and largest,and it is also an important sediment provenance of JZB.As an urban estuary and bay,this ecosystem is also impacted by agricultural run-off and urban and industrial sewage processing units,which discharge nutrients and other waste from domestic,industrial and agricultural areas into the estuary.The nanotechnology and the antimicrobial of AgNPs are developed rapidly in the world,and more and more AgNPs would be released into this area,which result in the potential ecological risk to the local aquatic environment (Liuet al.,2018).

Fig.1 The sampling stations in Dagu River Estuary and Jiaozhou Bay.

The surface sediment (0–5 cm) and bottom seawater samples were collected from two stations by using box corers and water sampler (Comprising a transparent polyvinyl chloride cylinder with a capacity of 2 L) in July 2017,respectively.Immediately after samples collecting,sediments were stored in sterile plastic bags in a dark covered bucket at 4℃,and the seawater samples were stored in polyethylene acid (10% HCl) cleaned (milli-Q water rinsed) bottles.In the laboratory,the sediments were wet-sieved (1 mm) to remove large particles,phytodetritus and zoobenthos,and the homogenized in sterilized plastic bags to avoid heterogeneity before sediment characterization and following experiment,while the seawater samples were immediately filtered (0.2 μm membrane filters) and stored at 4℃ until analyses and following experiment.The culture experiments were commenced within 48 h after sampling.

Dissolved oxygen (DO),pH and salinity of water samples were measured in situ using portable water analyzers(YSI6600).Redox potential (Eh) was determined using portable acidity meter (JENCOYC6010,USA).Total nitrogen (TN) of the sediments was analyzed in a CHN elemental analyzer (CE-440,USA).The content of dissolved inorganic N (DIN:NO3-,NO2-,NH4+) concentration in the seawater was quantified using a nutrient analyzer QuAAtro (Seal Analytical).Water (H2O) percentage(%) of sediment was measured by calculating the weight loss after dry sediment samples to 105℃ for 6 h.Total organic matter (TOM) content was obtained as percentage of weight loss by ignition (500 ℃,4 h),and grain size distribution was measured by using a Mastersizer 2000 Laser Size Analyzer (Malvern Co.,UK).Besides,sediments used in the present study had negligible background levels of silver (0.028 ± 0.004 ng kg-1in DRE and 0.019 ± 0.002 ng kg-1in JZB bas ed on ICP-MS).Seawater and sediment characteristics of DRE and JZB are listed in Table 1.

Table 1 The seawater and sediment indexes of Dagu River Estuary and Jiaozhou Bay

2.2 Experimental Preparation and Setup

Sediments used in the present study had negligible background levels of Ag (0.028 ± 0.004 mg Ag/Kg sediment in DRE and 0.019 ± 0.002 mg Ag/Kg sediment in JZB based on ICP-MS) compared with the spiked AgNPs concentration.AgNPs were purchased from Sigma-Aldrich Company,the average diameter of the particles size was 100 nm.To avoid C and N limitation for denitrifying communities in subsequent exposure experiments,the filtered water was amended with 20 mmol L-1KNO3and 22 mmol L-1glucose as incubation water.AgNPs were added to 1 L incubation water to produce sub-samples with AgNPs levels of 0 (non-spiked control),5,10,30,70,150,500 and 1500 mg L-1.Then the samples were sonicated (50 kHz and 250 W) for 0.5 h to break down aggregates at room temperature.The nanoparticle suspension was freshly prepared based on above-mentioned procedures just prior to the addition.

The effects of different AgNPs concentrations on denitrification in sediments were probed using the acetylene(C2H2) inhibition technique for slurries according to Sørensen (1978) and Magalhaeset al.(2005) with some modifications.In detail,15 g wet marine sediments and 15 mL sub-samples with different concentration of AgNPs were transferred to 50 mL serum bottles.Then shaken thoroughly,generating final AgNPs levels of 0 (nonspiked control),5,10,30,70,150,500 and 1500 mg kg-1wet sediment.Each treatment was set in sextuplicates.All serum bottles were hermetically sealed with a butyl stopper and aluminum crimp,and purged for 15 min with nitrogen (N2) to remove oxygen (O2).The sextuplicates of different concentrations of AgNPs were divided into two sets of subsamples to be incubated with and without acetylene (20% v/v).As the C and N would not limit denitrification,the measured potential denitrification activity (PDA) reflects the potential rates of the processes(Sørensen,1978).All exposure treatments were incubated in the dark for 24 h upon shaking (rotary shaker,80 r min-1,25 ℃).At the end of the experiment,gas samples (12mL)from all serum bottles were collected after headspace equilibration via vigorous shaking.The gas samples were then injected into evacuated serum vials (18 mL) for N2O analysis.The PDA was calculated by rate of N2O concentration change in the treatments with C2H2.N2O reduction(N2produced)viadenitrification was calculated as the difference between the N2O produced with and without acetylene.(Joyeet al.,1996).5 mL sediment slurries were centrifuged (8000 r min-1) and frozen for later DNA extraction.5 mL sediment slurries were collected by adding 3 to 5 drops HgCl (1 g L-1),and then centrifuged,0.22 μm filtered and frozen (-20 ℃) for later qua ntification the concentration of NO3-and NO2-.NO3-reduction rate was calculated by the difference between concentrations of NO3--N compounds measured atT0andT24h.NO2-reduction rate was calculated by the difference between NO3-reduction and NO2-residues after the exposure experiment.Moreover,2 g wet sediment was transferred to 10 mL sterilized centrifuge tube containing 5 mL of saline solution (0.22 μm filtered,9 g L-1NaCl,200 μL of 0.2 μm filtered,12.5% v/v Tween 80) and fixed with 100 μL of formaldehyde (0.2 μm filtered,4% v/v) for total cell count by flow cytometry.

2.3 Total Bacterial Cells

For the total count of bacterial cells,the subsamples of sediment slurries were stirred at 150 r min-1for 15 min,and then sonicated for 20–30 s at a low intensity (0.5 cycle,20% amplitude).The clear supernatant was then stained with SYBR Green I in anhydrous dimethylsulfoxide (DMSO) at a final concentration of 0.1 μmol L-1.After the addition of the dyes the samples were gently vortexed and incubated in a dark room for 15 min at room temperature (Foladoriet al.,2010;Colladoet al.,2017).Before analyzed by the flow cytometer,the subsamples were diluted 1:10 or 1:100 with saline solution (0.22 μm filtered,9 g L-1NaCl) to ensure the concentration of bacterial cells less than 106cells mL-1.Then the diluted samples were vortexed by vortex mixer,and analyzed using a BD Accuri C6 flow cytometer (Becton,Dickinson and Company) equipped a 488 nm solid state laser.Green fluorescence was collected at 530 nm in the FL1 channel and red fluorescence at 630 nm in the FL3 channel with the trigger set on the green fluorescence.Data were processed using the BD Accuri C6 Plus software (Becton,Dickinson and Company).Moreover,electronic gating was used to separate the desired events.Total bacterial cells were selected according to their FL1/FL3 signal to reduce signal background and instrument noise (Hammes and Egli,2005).

2.4 Nucleic Acid Extraction and Real-Time Quanti-tative PCR (RT-qPCR)

Total genomic DNA (gDNA) was extracted from 1 g wet weight of control and AgNPs spiked sediments (5,10,30,70,150,500 and 1500 mg kg-1) in triplicates using a Fast DNA SPIN kit for soil (SK8233,Sangon Biotech Co.,Ltd.) following the manufacturer’s instructions.Then the DNA solution was stored at -80 ℃ for the following RT-qPCR.ThenirG,nirSandnosZgenes were selected for denitrifiers.narG,nirSandnosZgene fragments were amplified with the primer pairs 1960m2F and 2050m2RGC (López-Gutiérrezet al.,2004),Cd3aF and R3cdR-GC(Throbäcket al.,2004),F2 and R2 (Henryet al.,2006),respectively.The product length ofnarG,nirSandnosZgenes was 110,406 and 267 bp,respectively.A Roche LightCycler 480 system was used to detect the gene abundance.In details,the 25 μL reaction mixtures contained 10 μL of SYBR Green Fast qPCR Master Mix(Roche,Switzerland),0.4 μL of each 20 mmol L-1primer,7.2μL ddH20,and 2μL of template DNA.The thermal cycling conditions were as follows:pre-incubation at 95℃ for 3 min,45 cycles consisting of denaturation at 94℃ for 7 s,annealing at 57 ℃ for 10 s,extension at 72 ℃for 30 s,followed by melting curve analysis at 60–95 ℃with a heating rate of 0.11℃ s-1and a continuous fluorescence measurement) and finally a cooling step to 40℃ .Three replicate amplifications were performed for RTqPCR each sample,and the mean of replicate values was used for following analysis.The relative gene abundance was analyzed by the 2-△△CTmethod (Livak and Schmittgen,2001).

2.5 Statistical Analysis

A gas chromatograph (Shimadzu GC-14B,Japan)equipped with an electron-capture detector was used to detected N2O.The N2O concentration was quantified using daily standard curves generated from certified gas standards,and the detection limit of the method was approximately 20 nmol L-1N2O.Inhibition rate (%) of all the parameters indexes was calculated using the following Eq.(1):

whereRcmeans the value of control treatment,Rimeans the value of Phe treatment.The N2O reduction ratio was calculated using the Eq.(2):

where total N2O means PDA,and accumulation N2O was calculated by N2O produced without C2H2.

In this study,all tests were analyzed in triplicate,and the mean and standard deviation values of the results were calculated.The statistical significance of changes was analyzed by one-way analysis of variance (ANOVA)using SPSS software (version 22.0).The SPSS software was also used to calculate the Spearman Indices (R) and Significant Correlation (P) of the correlations among correlation parameters.Data were considered statistically significant whenP<0.05.

3 Results and Discussion

3.1 Effect of AgNPs on Denitrification Activity,Nitrogen Transformation and N2O Emission

3.1.1 Potential denitrification activity

The change of PDA in the sediments of DRE and JZB is shown in Fig.2.After one day AgNPs incubation,the PDA of the control treatments in DRE and JZB was the highest among all treatments,0.453 mg N kg-1h-1and 0.592 mg N kg-1h-1,and lowest for sediment of DRE and JZB treated with 1500 mg kg-1AgNPs,0.124 mg N kg-1h-1and 0.186 mg N kg-1h-1,respectively.Moreover,the inhibition rate of PDA for DRE and JZB ranged from 4.22% to 72.82% and 3.89% to 68.96%,respectively.The PDA of each treatment group in the two study areas decreased with an increase of AgNPs concentration,and was negatively correlated with AgNPs concentration(DRE:r=-0.969,P<0.001;JZB:r=-0.975,P<0.001),which revealed that the PDA could be inhibited by experimental range of AgNPs concentrations for both study areas.The PDA of 5 mg kg-1AgNPs treatments in the two areas had no significant difference compared with control treatment (ANOVA,P>0.05),while the PDA of AgNPs treatments above 10 mg kg-1showed significant difference compared with control treatment (ANOVA,P<0.05).According to the logistic curve,the mean EC50value of AgNPs for PDA of DRE and JZB was 114.45 mg kg-1(r2=0.99,P<0.0001) and 146.09 mg kg-1(r2=0.99,P<0.0001),respectively.Previous studies have showed that physical and biogeochemical factors in different sediment environment can affect the function,community structures and diversity of denitrifiers,which can explain the different responses of denitrifiers in the two study areas(Wanget al.,2014;Neubaueret al.,2019).These differentces between two study areas coupled with the other indicators (e.g.,NO3-reduction,NO2-reduction,the abundance of denitrifying genes and so on) were further discussed in Section 3.2.

Fig.2 Potential denitrification activity (PDA,presented as mg N kg-1 sediment h-1) under different AgNPs concentrations in Dagu River Estuary and Jiaozhou Bay.Notes:Bars represent standard deviation.The fitting curve and the mean EC50 value were obtained by applying ‘Four Parameter Logistic Curve’ (y=min+(max -min)/ (1+(x/EC50)Hillslope) to the relation of AgNPs concentration and PDA.

3.1.2 Nitrate and nitrite reduction

The NO3-and NO2-reduction rates of experimental treatments in the two study areas decreased with the increase of AgNPs concentration (Fig.3).Both of the NO3-and NO2-reduction rates in DRE and JZW were significantly negatively correlated with AgNPs concentration(DRE:r=-0.988,P<0.01,r=-0.990,P<0.001;JZB:r=-0.985,P<0.001,r=-0.980,P<0.001).Moreover,the NO3-and NO2-reduction rates of 5 mg kg-1AgNPs concentration in both areas had no significant difference compared with control treatment (ANOVA,P>0.05),while it showed a significant different at AgNPs concentrations above 10 mg kg-1compared with control treatment (ANOVA,P<0.05).Furthermore,for both study areas,NO3-reduction rate of control treatment was highest,and lowest in the 1500 mg AgNPs kg-1treatments(DRE:7.439 and 4.486 mg N kg-1h-1;JZB:8.470 and 5.527 mg N kg-1h-1).The inhibition rate of NO3-reduction rate in 1500 mg AgNPs kg-1treatment for DRE and JZW was 36.88% and 32.00%,respectively.Furthermore,the NO2-reduction rate was also highest and lowest in the control and 1500 mg AgNPs kg-1treatments (DRE:3.222 and 1.563 N kg-1h-1;JZB:3.793 and 2.200 mg N kg-1h-1),and the 1500 mg AgNPs kg-1treatments in DRE and JZW inhibited NO2-reduction rate by 36.88% and 32.00%.In addition,the NO3-and NO2-reduction rates were generally lower in DRE than in JZB (ANOVA,P<0.05),which implied that the NO2-reductase may be more sensitive to AgNPs than NO3-reductase.

Fig.3 Effect of AgNPs on NO3- (a),NO2- reduction (b) and inhibition rates in Dagu River Estuary (DRE) and Jiaozhou Bay(JZB).Note:* Represent the significant difference between experimental and the control group (ANOVA,P <0.05;the same below).

3.1.3 N2O accumulation and reduction

The N2O accumulation rate of the control treatment in DRE and JZB was 0.026 mg N kg-1h-1and 0.038 mg N kg-1h-1,which was lowest among all treatments in both areas (Fig.4).The N2O accumulation rate in both areas generally increased and then decreased with the increase of AgNPs concentration.For both of the study areas,the N2O accumulation rate in 5 mg kg-1AgNPs treatment had no significant difference compared with control treatment(ANOVA,P>0.05),while it had significant difference in AgNPs treatments above 10 mg kg-1compared with control treatment (ANOVA,P<0.05).The N2O accumulation rate in AgNPs treatment was 1.78–4.49 and 1.68–4.18 folds of that in the control group for DRE and JZW,respectively (Fig.4),which means N2O realease was inceeased with the AgNPs pollution aggravating.The highest value of N2O accumulation rate appeared at the 70 and 150 mg AgNPs kg-1treatments,which was 0.108 mg N kg-1h-1and 0.147 mg N kg-1h-1for DRE and JZW,respectively.It implied that the N2O realease was more affected by AgNPs in DRE than in JZB (ANOVA,P<0.05).The results revealed that the toxic effect of AgNPs on N2O accumulation rate and N2O realease was greater in estuary than in bay (ANOVA,P<0.05).

Fig.4 N2O accumulation rate and fold of control in different AgNPs concentration treatments in Dagu River Estuary (DRE) and Jiaozhou Bay (JZB).

The N2O reduction rate in both study areas showed a decreasing trend with the increase concentration of AgNPs addition (Fig.5),and the N2O reduction rate was negatively correlated with AgNPs concentration (DRE:r=-0.993,P<0.001;JZB:r=-0.990,P<0.001).The N2O reduction rate of all AgNPs treatments in two study areas was significantly different from that of the correspond control treatment (ANOVA,P<0.05).The N2O reduction rate of the control treatment in DRE and JZW was both the highest,0.430 mg N kg-1h-1and 0.561 mg N kg-1h-1,respectively,and the N2O reduction ratio was 94.39% and 93.67%.The N2O reduction ratio was lowest at the 1500 mg kg-1AgNPs concentration for DRE and JZW,31.11%and 38.92%,respectively.Some studies have showed that lower N2O reduction ratio could lead to the higher N2O accumulation (Guoet al.,2011).Zhenget al.(2017)found that the addition of AgNPs could promote the N2O production in estuarine sediments,which could lead to an increase of N2O accumulation,and in turn endangering the estuarine ecosystem.The N2O accumulation and reduction rate of each treatment group in DRE was higher than that of the treatment group in JZB (ANOVA,P<0.05).

Fig.5 N2O reduction rate and ratio in different AgNPs concentration treatments in Dagu River Estuary (DRE)and Jiaozhou Bay (JZB).

This study showed that AgNPs have different degrees of inhibiting effect to different denitrification process,and the higher AgNPs addition could lead to stronger inhibiting effect.This inhibiting effect results in excessive accumulation of N-containing nutrients (such as NO3-and NO2-) in water environment,which could cause profound toxic effect on marine organism,and further affect the nutrient cycle and energy flow of marine ecosystem(Alinsafiet al.,2008).For the highest AgNPs addition treatments (1500 mg kg-1),the inhibition rate of PDA,NO3-and NO2-reduction in DRE and JZB was 72.82%and 68.96%,36.88% and 32.00%,31.19% and 26.22%,respectively.The results of inhibition rate of PDA,NO3-and NO2-reduction in two study areas showed that the inhibition effect of AgNPs on PDA was the largest,followed by NO2-reduction,and the NO3-reduction was the weakest.It may be related to the cascade effect of nutrient,pollutants or enzymatic reactions as well as the acting position of reductase in different denitrification processes and their sensitivity to AgNPs,which are usually sequentially induced under anaerobic conditions (Philippot and Hallin,2005).

3.2 Effect of AgNPs on Total Bacterial Count and Relative Abundance of narG,nirS and nosZ Genes

The total bacterial count in the sediments of the two study areas both showed a slightly increasing trend with an increase of AgNPs concentration,whereas the increasing trend is not evident (Fig.6).The inhibition rate of the total bacterial count in AgNPs treatments for DRE ranged from -0.27% to -3.95%,while it ranged from -0.28% to-7.53% for JZB.The results indicated that AgNPs can promote bacterial growth,especially in JZB area.After 24 h incubation,the total bacterial count of AgNPs treatments in two study areas both had no significant difference with the control treatments (ANOVA,P>0.05).The total bacterial count in DRE and JZB was highest in the 1500 mg AgNPs kg-1treatments (DRE:5.28 × 108cells g-1;JZB:5.71 × 108cells g-1),and lowest in the control treatments (DRE:5.49 × 108cells g-1;JZB:6.14 × 108cells g-1).Some studies reported that AgNPs addition can significantly affect the growth and activity of bacteria in soil and sediment (Bradfordet al.,2009).Baoet al.(2016)investigated two freshwater sediments and found that the total bacterial count in 35 nm,75 nm and PVP wrapped AgNPs treatments increased in different degrees compared with the control group after 24 h exposure.In this study,the short addition time of AgNPs may lead to less silver ions (Ag+) released,which resulted in less toxicity.Besides,the short-term stimulation effect of AgNPs may also promote the growth of bacteria,and some researchers have found that low concentration of Ag+can stimulate the activity of microorganisms and promote the growth of bacteria (Arnaout and Gunsch,2012).

Fig.6 Effect of AgNPs on total bacterial count and inhibition rate in Dagu River Estuary and Jiaozhou Bay.

The effects of different concentrations of AgNPs on relative abundance ofnarG,nirSandnosZgenes in DRE and JZB are shown in Fig.7.The results showed that the relative abundance of the three functional genes in two areas decreased with the increase of AgNPs concentration,which was different from the trends of total bacterial count.It is possible that the denitrifying genes were more sensitive to AgNPs than total bacterial count,which means AgNPs has specific effects on denitrifiers.The relative abundance ofnarG,nirSandnosZgenes was lowest in the 1500 mg AgNPs kg-1treatments of DRE(0.42,0.35 and 0.29) and JZB (0.44,0.37 and 0.31).Several studies have demonstrated that the abundance of denitrification functional genes (nirKandnosZ) could be reduced by heavy metal pollution (Magalhaeset al.,2012;Liuet al.,2016;Liuet al.,2018).Zhenget al.(2014)demonstrated that the expression of denitrifying genes could be significantly inhibited by AgNPs addition.The inhibition rate ofnosZgene was the highest,followed bynirSgene and thennarGgene (ANOVA,P<0.05).The EC50values of AgNPs for relative abundance ofnarG,nirSandnosZgenes were 84.00,70.68 and 55.41 mg kg-1(r2>0.95,P<0.0001) in DRE,and 96.90,82.55 and 66.76 mg kg-1(r2>0.95,P<0.0001) in JZB,respectively,which was higher than the corresponding EC50values of AgNPs for PDA (ANOVA,P<0.05).The obtained results illustrated that the denitrifiers harboringnarG,nirSandnosZgene were all more sensitive to AgNPs than PDA.Besides,in the two areas,the EC50value ofnarGgenewas the highest,followed bynirSgene,thennosZgene,implying that the tolerance ofnarGgene to AgNPs is the strongest,followed bynirSgene,thennosZgene.In addition,the inhibitory effect of AgNPs onPDAwas strongest,followed by NO2-reduction rate,and then NO3-reduction rate,which is consistent with the sensitivity order of denitrifying genes (narG,nirSandnosZ) participated in different reduction processes to AgNPs.The toxic effects of AgNPs on denitrification processes (PDA,NO3-reduction,NO2-reduction,N2O accumulation and reduction)and related functional genes (narG,nirSandnosZ) in DRE was greater than that in JZB,which may be related to various environmental factors (such as pH,salinity,OM,Eh and sediment particle size) in the two study areas.Previous studies have shown that different environmental factors could change the toxic effect of AgNPs through influencing its complexation,and then indirectly affect the biogeochemical process in sediments (Kimet al.,2010;Pokhrelet al.,2013).Generally,increasing oxygen and absolute high concentration of nitrate,or decreasing pH and carbon/nitrate ratio could lead to incomplete denitrification process by inhibit N2O reducing to N2,and accumulate N2O (Shaoet al.,2011),in this research,the influence of these factors on denitrification activity may not be as strong as salinity,Eh,OM,silt and clay.Beddowet al.(2017) found that the nitrification of estuarine sediments was inhibited by AgNPs,besides,the inhibition effect of AgNPs on nitrification rate in low salinity sediments is stronger than that in high salinity sediments.Previous research showed that only AgNPs with a diameter of 1–10 nm could directly influence the microorganisms by destroying cell membrane functions through adsorption,which implied the toxic action mode AgNPs with a diameter of 100 nm was probably mainly based on Ag+release and oxidative stress,etc.(Recordatiet al.,2016;Liuet al.,2018).Previous research showed that the toxicity of AgNPs could be more reduced by seawater with high ionic strength and salinity than that with low ionic strength and salinity (Oukarroum,2012).Moreover,clay minerals and OM content are the most influential factors on the absorption,sequestration,and bioavailability of contaminants,and denitrification is carried out mostly by anaerobic and facultative microorganisms (Zhuet al.,2016;Zhenget al.,2018).

Fig.7 The relative abundance of narG,nirS and nosZ genes in treatments with different AgNP concentrations in Dagu River Estuary (a) and Jiaozhou Bay (b).

3.3 Correlations Between Microbial Abundance and Denitrification

Previous studies have shown that the denitrification enzyme activity was correlated with the abundance ofnarG,nirSandnosZgenes (Hallinet al.,2009;Chon and Cho,2015),which means denitrifiers harboring these functional genes played an important role in the denitrification process,and the abundance of the denitrifying genes could predict the corresponding process (Hallinet al.,2009).The correlation between total bacterial count,relative abundance of denitrifying genes and denitrification in the two study areas was shown in Table 2.The total bacterial count had no significant correlation with PDA,N2O,NO3-and NO2-reduction rates (P>0.05) in the two areas;by contrast,the relative abundance ofnarG,nirSandnosZgenes was all strongly positively correlated (P<0.001) with them.In addition,thenarG,nirSandnosZgenes were strongly positively correlated between each other (P<0.001).This may be related to that thenarG,nirSandnosZgenotype denitrifiers were more sensitive to AgNPs than the other bacteria,which means that AgNPs may have specific effects for these denitrifiers.Besides,it may also be related to the short-term incubition and the stress response of denitrifiers (Trautweinet al.,2008).

Table 2 Correlations between total bacterial count,relative abundance of denitrifying genes (narG,nirS and nosZ)and denitrification in Dagu River Estuary and Jiaozhou Bay (Spearman correlation).

3.4 Environmental Implications and Prospective

Owing to the antimicrobial properties,AgNPs has been widely used as engineered nanomaterials over the world(Fabregaet al.,2011).Currently,there is no evidence to suggest that humans are being adversely affected by AgNPs through their use in consumer products,but the extensive use of AgNPs inevitably leads to their release into aquatic environment due to their abundant use in antibacterial products.Nitrifying bacteria are highly susceptible to AgNPs,and the releases of AgNP can result in the stimulation of N2O release and affect the nitrogen cycle (Michelset al.,2015).In this study,we demonstrated that the inhibitory effect of AgNPs on different denitrification process,which may lead to the increase of inorganic nitrogen accumulation and N2O realease,and this may in turn affect the nitrogen cycle.Meanwhile,the Ag+was considered the most toxic form of silver in water,and AgNPs in aquatic system potentially show dose-dependent toxicity to aquatic organisms,such as bacteria,algae,macrophytes,invertebrates and fishes (Jianget al.,2017).For instance,in a natural estuary,photosynthesis of the phytoplankton community was significantly reduced when the AgNPs concentration was higher than 0.5 mg L-1after a 24 h exposure (Baptistaet al.,2015).Moreover,AgNPs strongly accumulated in aquatic biotic components that could be consumed by other higher trophic animals;therefore AgNPs might transfer to food webs which imply risks for human health (Bruneauet al.,2016).

The use of AgNPs and their potential environmental and human health risks are of increasing concern.It’s essential to develop effective techniques for detecting and quantifying AgNPs in aquatic system.Meanwhile,indepth studies of AgNPs toxicity (e.g.,Ag+ion and NPs toxicity) are needed to better understanding the potential risks for aquatic environments.Most importantly,the detailed information on the use of AgNPs in consumer products is indispensable,which could be strictly tracked by the governments.

Acknowledgements

This work was supported by the Special Project of Guangxi Science and Technology Base and Talent (GUI KE AD20297065),the National Natural Science Foundation of China (No.U20A20103),the National Major Project of Water Pollution Control and Management Technology in China (No.2013ZX07202-007),and the Science and Technology Planning Projects of Beihai,Guangxi,China (Nos.201995002,201995076,202082031,2020 82022 and 202082032).