Journal of Hazardous Materials 290 (2015) 127–133

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journa l homepage: www.e lsev ier .com/ locate / jhazmat

Effect of silver nanoparticles on Pseudomonas putida biofilms atdifferent stages of maturity

Pumis Thuptimdanga,b, Tawan Limpiyakornb,c,d, John McEvoye, Birgit M. Prüße,Eakalak Khanf,∗

a International Program in Hazardous Substance and Environmental Management, Graduate School, Chulalongkorn University, Bangkok 10330, Thailandb Center of Excellence on Hazardous Substance Management, Bangkok 10330, Thailandc Department of Environmental Engineering, Chulalongkorn University, Bangkok 10330, Thailandd Research Unit Control of Emerging Micropollutants in Environment, Chulalongkorn University, Bangkok 10330, Thailande Department of Veterinary and Microbiological Sciences, North Dakota State University, Fargo, ND 58108, USAf Department of Civil and Environmental Engineering, North Dakota State University, Fargo, ND 58108, USA

h i g h l i g h t s

• Biofilm stages in static batch con-ditions were similar to dynamicconditions.

• Expression of csgA gene increasedearlier than alg8 gene in biofilm mat-uration.

• AgNPs had higher effect on lessmature biofilms.

• Removal of extracellular polymericsubstance made biofilms susceptibleto AgNPs.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 December 2014Received in revised form 25 February 2015Accepted 26 February 2015Available online 27 February 2015

Keywords:Silver nanoparticlesBiofilmsBiofilm maturityExtracellular polymeric substance

a b s t r a c t

This study determined the effect of silver nanoparticles (AgNPs) on Pseudomonas putida KT2440 biofilmsat different stages of maturity. Three biofilm stages (1–3, representing early to late stages of development)were identified from bacterial adenosine triphosphate (ATP) activity under static (96-well plate) anddynamic conditions (Center for Disease Control and Prevention biofilm reactor). Extracellular polymericsubstance (EPS) levels, measured using crystal violet and total carbohydrate assays, and expression ofthe EPS-associated genes, csgA and alg8, supported the conclusion that biofilms at later stages were olderthan those at earlier stages. More mature biofilms (stages 2 and 3) showed little to no reduction in ATPactivity following exposure to AgNPs. In contrast, the same treatment reduced ATP activity by more than90% in the less mature stage 1 biofilms. Regardless of maturity, biofilms with EPS stripped off were more

∗ Corresponding author at: Civil and Environmental Engineering Department (# 2470), P.O. Box 6050, Fargo, ND 58108-6050, USA. Tel.: +1 701 2317717;fax: +1 701 2316185.

E-mail addresses: pumis.th@gmail.com (P. Thuptimdang), tawan.l@chula.ac.th (T. Limpiyakorn), john.mcevoy@ndsu.edu (J. McEvoy), birgit.pruess@ndsu.edu (B.M. Prüß),eakalak.khan@ndsu.edu (E. Khan).

http://dx.doi.org/10.1016/j.jhazmat.2015.02.0730304-3894/© 2015 Elsevier B.V. All rights reserved.

128 P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133

susceptible to AgNPs than controls with intact EPS, demonstrating that EPS is critical for biofilm toleranceof AgNPs. The findings from this study show that stage of maturity is an important factor to consider whenstudying effect of AgNPs on biofilms.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Silver nanoparticles (AgNPs) are incorporated as an antibacte-rial agent into a wide range of materials, including those used inwound dressings, clothes, medical devices, and water disinfection[1–4]. These extensive uses of AgNPs raise a concern that they maybecome widespread in the environment and have a negative impacton beneficial bacteria in the environment [5]. The toxicity of AgNPsto bacteria involves damage to the cell wall or cell membrane, pene-tration into the cell, and inactivation of DNA via released silver ions[6–8]. Reactive oxygen species generated by AgNPs can damagecells through oxidative stress [9].

To determine the antibacterial effectiveness of AgNPs and con-sequences of their release into environment, studying the effect ofAgNPs on biofilms is important because bacteria are often presentin biofilm communities. To survive in harsh environment, bacteriafacilitate cell interactions to attach to surface, and produce extracel-lular polymeric substance (EPS) to create more complex structures,called biofilms [10]. Compared to planktonic cells, biofilm cells havedifferent phenotypes and genotypes leading to specific biologicalactivities, metabolic pathways, and stress responses [11]. The genesexpressed in biofilms include functions related to surface attach-ment, transition to stationary phase-like cells, and EPS production[12].

The EPS of biofilms, which comprises polysaccharides, proteins,nucleic acids, and other macromolecules, can act as a supportingstructure for bacterial adherence to surfaces and access to nutri-ents. It also protects against antimicrobial agents [13,14]. However,AgNPs show ability to eradicate bacterial biofilms. It was found thatAgNPs are more toxic to phototrophic biofilms than Ag ions andare able to diminish biomass of the biofilms [15]. Smaller AgNPscan reduce more biomass and viability of biofilms, due to betterpenetration into the EPS matrix [16]. The ability of AgNPs to inacti-vate biofilms also increases in dynamic conditions due to increasedbiosorption [17]. Still, mature biofilms have mechanisms to tolerateAgNPs by using EPS-mediated trapping, aggregation, and reduceddiffusion of AgNPs [18–21].

During biofilm formation, biofilms develop themselves to dif-ferent stages. There are at least four stages of biofilm formation:planktonic, attachment (reversible and irreversible), maturation(microcolonies and macrocolonies), and dispersion [22]. Thesestages occur dynamically during biofilm formation. The formationof various phenotypes related to each stage is regulated by differentgene expressions. First, bacteria use organelles like flagella to moveonto the surface. To attach irreversibly to the surface, the flagellargenes are repressed, followed by the expression of adhesion pro-teins such as curli, pilli, and type I fimbriae [23,24]. After irreversiblyattached to the surface, exopolysaccharide biosynthesis genes areexpressed such as the ones encoding capsule and alginate to con-struct the mature biofilms [24,25]. Since biofilms show differentcharacteristics during maturation, biofilms in different stages mayhave different susceptibility to AgNPs.

Various studies have proven different levels of susceptibilityof biofilms in different stages to other antimicrobial agents. Tré-Hardy et al. [26], studied the co-administration of antibiotics onbiofilms at different stages of maturation. They found that moremature biofilms were less susceptible to antibiotics. Other studies

have shown that older biofilms are less susceptible to chlorhexi-dine and various sanitizers [27,28]. However, the effect of biofilmmaturity on their susceptibility to AgNPs has not been elucidatedand should be studied in order to understand the adverse effect ofAgNPs on environmental biofilms.

The objective of this study was to determine the AgNPs suscep-tibility of Pseudomonas putida KT2440 biofilms at different stagesof maturity. P. putida KT2440 was selected because it is an effec-tive biofilm-producer found in soil and aquatic environments, andcomprehensive physiological and genetic data are available [29,30].The study was divided into two parts. Firstly, biofilm matura-tion was observed in biofilms grown under static (96-well plate)and dynamic (Center for Disease Control and Prevention (CDC)biofilm reactor) conditions. Secondly, biofilms at different matura-tion stages were exposed to AgNPs and the effect on biofilm viabilitywas determined.

2. Materials and methods

2.1. Preparation of AgNPs

AgNPs were synthesized according to the method by Choi et al.[31], using sodium borohydride to reduce silver nitrate with 0.06%of polyvinyl alcohol (PVA) as a capping agent. The concentrationof total Ag from the calculation was 26.3 mg/l. The formation ofAgNPs was verified by scanning the absorbance of the solutionbetween 250 and 700 nm with a UV–vis spectrophotometer [32].The particles were characterized for size and zeta potential usinga zetasizer (Malvern Instruments, Worcestershire, UK). To mea-sure the amount of Ag ion, the AgNPs solution was centrifuged at165,000 × g, 4 ◦C, for 1 h [33]. The supernatant was collected anddissolved with HNO3 before measurement by inductively coupledplasma mass spectrometry (ICP-MS).

To observe the release of Ag ion after the exposure, an experi-ment was performed using a polystyrene, flat-bottom, 6-well plate(Thermo Scientific). Each well contained 5 ml of 0.5X Luria–Bertani(LB; 1% tryptone, 0.5% NaCl, 0.5% yeast extract) and 50 �l of the P.putida KT2440 inoculum prepared according to the next subsec-tion. The plate was incubated at room temperature (20 ◦C) withoutshaking for 24 h to allow biofilm formation. After that, 4 ml of mediawas removed before adding 4 ml of AgNPs. The plate was incubatedfurther at room temperature. After 48 h, the media was taken formeasurement of total Ag and Ag ion by ICP-MS.

2.2. Bacterial strain and culture preparation

Before each experiment, P. putida KT2440 (ATCC 47054) wascultivated at 37 ◦C overnight in LB medium. The suspension wascentrifuged and the pellet was re-suspended in phosphate buffersaline (PBS). The optical density of the culture, measured at 600 nm,was adjusted to 0.4 with PBS (approximately 107 CFU/ml) beforeuse as an inoculum in experiments.

2.3. Biofilm formation

A polystyrene, flat-bottom, 96-well microtiter plate (GreinerBio-One Frickenhausen, Germany) was used to support biofilm for-

P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133 129

mation under static conditions. Each well contained 100 �l of 1X LBmedium (final concentration = 0.5X), 95 �l of deionized (DI) water,and 5 �l of the prepared P. putida KT2440 inoculum. The plate wasincubated at room temperature without shaking to allow biofilmformation.

A CDC biofilm reactor (Model 90-1, Biosurface Technologies,Bozeman, MT) was used to examine biofilm formation underdynamic conditions. The reactor is a one-liter glass vessel with alid that can hold 8 polyethylene rods. Each rod holds three remov-able polycarbonate coupons serving as biofilm growth surfaces.One milliliter of the P. putida KT2440 inoculum was pipetted intothe reactor containing 500 ml of 0.5X LB medium. The reactor wasoperated in a batch mode (100 rpm stirring) and was kept at roomtemperature to allow biofilm formation.

2.4. Adenosine triphosphate (ATP) assay

An ATP based BacTiter-GloTM microbial cell viability assay(Promega, Madison, WI) was used to monitor changes in bacte-rial activity during biofilm formation [34]. In 96-well plates, ATPconcentration was measured every 3 h for the first 24 h and every12 h between 24 and 72 h. Media was removed and the biofilmwas rinsed twice with 200 �l of PBS. One hundred microliters ofBacTiter-GloTM reagent was added to the well and mixed brieflywith the biofilm by pipetting. After incubation at room temperaturefor 5 min, the bioluminescence was measured as relative light units(RLU) using a TD-20/20 luminometer (Turner Designs, Sunnyvale,CA).

Under CDC reactor conditions, the ATP concentration was mea-sured every 12 h for 72 h. A rod was removed from the reactor andcarefully dipped in two consecutive tubes containing 25 ml of PBSto remove the planktonic cells. The three coupons on each rod rep-resented three replicates for the same time point. Each coupon wasremoved and put in a tube containing 2.5 ml of PBS. The biofilm wasdetached from the coupon by vortex mixing for 30 s. One hundredmicroliters of BacTiter-GloTM reagent was mixed with 100 �l of thecell suspension before measuring the bioluminescence as describedabove.

2.5. Biofilm amount

Two different methods were used for determination of biofilmamount in the 96-well plate and the CDC reactor. The biofilmamount in a 96-well plate was quantified by crystal violet (CV)staining according to the method by Sule et al. [34]. In the CDCreactor, the biofilm amount was determined from total carbohy-drate by a phenol-sulfuric acid method modified from Masuko et al.[35]. The samples were prepared by the method described in theATP assay subsection. A 1.5 ml aliquot of concentrated H2SO4 wasadded to 500 �l of the sample and incubated for 30 min. A 300 �laliquot of 5% (w/v) phenol in water was added, and the sample washeated at 90 ◦C in a water bath for 10 min. The sample was cooledat room temperature for 15 min before measuring the absorbanceat 492 nm.

2.6. RNA extraction and qPCR

To extract RNA from the 96-well plate, media was removed andthe biofilms were rinsed twice with PBS. One hundred microlitersof PBS were added to each well and the biofilms were scrapedwith an inoculating needle. Disrupted biofilms were removed fromthe wells by a pipette. RNA was extracted from 500 �l of sus-pended biofilm using an RNeasy mini kit (Qiagen, Valencia, CA) inaccordance with manufacturer’s instructions. Genomic DNA con-tamination was removed by treatment with DNase I (Qiagen).

Biofilms were collected from the CDC reactor by the samplingmethod described above for the ATP assay. To prepare a samplewith an adequate number of cells, 20 ml of a cell suspension, pre-pared from 8 rods (24 coupons), was centrifuged and the pellet wasre-suspended in 2.5 ml of PBS. RNA was extracted from 500 �l ofthe sample with an RNeasy Plus Micro kit according to the protocolsprovided by the manufacturer (Qiagen).

cDNA was synthesized using random primers (Promega) andMoloney murine leukemia virus reverse transcriptase (MMLV-RT,Promega). The reverse transcription process was carried out at 37 ◦Cfor 60 min followed by heating at 70 ◦C for 10 min for enzyme inac-tivation. Samples without the reverse transcriptase were used as anegative control. Fragments of csgA, alg8, and 16S ribosomal RNA(rRNA; used to normalize expression) transcripts were amplifiedusing a SYBR green qPCR approach according to the method byHorne and Prüß [36]. The fluorescence signal was monitored in aniQ5 thermocycler Real-Time PCR detection system (Biorad). For-ward and reverse primers for csgA were 5′-ATA AAT CCA CCG TGTGGC AGG ACA-3′ and 5′-AGG TCT GTT CGA TGA AAG CCT CGT-3′,respectively. Forward and reverse primers for alg8 were 5′-GTGACC TCG CCA GCT TTC AAC AAT-3′ and 5′-TGA ACA GCA CAG CAACGA AGA TGC-3′, respectively. Forward and reverse primers for 16SrRNA were 5′-CCA GGG CTA CAC ACG TGT TA-3′ and 5′-TCT CGCGAG GTC GCT TCT-3′, respectively. Expression data were analyzedby the comparative Ct method (��Ct), where Ct is the thresholdcycle [37].

2.7. Exposure of biofilms to AgNPs

For the 96-well plate, 150 �l of the media was removed beforeadding 150 �l of the AgNPs solution. Biofilms were exposed toAgNPs for 48 h at room temperature. At 0, 3, 6, 9, 12, 24, and 48 h,the solution was removed; biofilms were rinsed twice in PBS beforemeasuring the ATP concentration. Control experiments were car-ried out in a similar manner, with the exception that 150 �l of0.06% PVA solution was used instead of the AgNPs solution. Theeffect of AgNPs on biofilms was determined by comparing the ATPconcentrations of treatment and control samples.

For CDC reactor experiments, the lid of the reactor containingthe polyethylene rods was transferred to another reactor contain-ing 400 ml of AgNPs solution and 100 ml of 1X LB medium. Controlexperiments were carried out in a similar manner, with the excep-tion that 400 ml of 0.06% PVA solution was used instead of theAgNPs solution. The reactor was operated in a batch mode (100 rpmstirring) for 6 h at room temperature, and it was sampled after 0,1, 3, and 6 h. During sampling, one rod was taken from the reactorand was replaced by a new rod to balance the fluid shear stress inthe reactor. The ATP concentration in the biofilm was determined asdescribed earlier. Biofilms also were examined using a conventionalplate count method [38].

2.8. Effect of EPS on biofilm susceptibility to AgNPs

Experiments to examine the effect of EPS on biofilm suscepti-bility to AgNPs were conducted in a 96-well plate with biofilmsgrown for 6, 12, and 48 h. At each time point, media was removed,the biofilm was rinsed twice with PBS, and part of EPS was removedusing 200 �l of 2% (w/v) ethylenediaminetetraacetic acid (EDTA)[39]. The reduction of biofilm amount was observed by the CV assay.Control experiments were carried out using DI water instead ofEDTA. Treatment and control plates were incubated at 4 ◦C for 3 hbefore rinsing the biofilms with PBS and treating with 200 �l of theAgNPs solution for 2 h. The effect was determined by comparing theATP concentrations of treatment and control samples as describedearlier.

130 P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133

Fig. 1. Time course of ATP amount of P. putida KT2440 biofilm.

2.9. Statistical analysis

Experimental data were statistically analyzed using GraphPadPrism® software version 6.01 (GraphPad Software, La Jolla, CA). Inevery experiment, the standard deviation of the triplicate data wascalculated and presented as error bars. The multiple t-test was usedto analyze the statistical differences. To correct the errors from mul-tiple comparisons of t-test, the Holm–Sidak method was used overthe t-test at 5% significance level.

3. Results and discussion

3.1. Characterization of AgNPs

AgNPs showed the characteristic absorbance at 395 nm similarto a previous report [40]. The particle size range was 40–60 nm,and a zeta potential range of –2 to –6 mV indicated a nearly neu-tral charge. The concentration of synthesized AgNPs was 25.86 mgof total Ag/l from the measurement by ICP-MS (26.3 mg/l from thecalculation). The synthesized AgNPs released 1.24 mg/l of Ag ion,which was 4.8% of total Ag. The 6-well plate experiment showed thereduction in total Ag from 20.61 to 17.94 mg/l after 48 h of exposure,suggesting the transport of AgNPs into biofilms. At 48 h, the con-centration of Ag ion showed the higher release at 1.76 mg/l, whichwas 9.8% of the total Ag. Therefore, the synthesized AgNPs shouldhave the mechanisms of toxicity through both nanoparticles andAg ion.

3.2. Stages of P. putida KT2440 biofilm maturation

Fig. 1 presents ATP levels for P. putida over a 72 h period understatic (96 well plate) and dynamic (CDC reactor) conditions. Asimilar temporal pattern of ATP activity was observed in biofilmsgrown under both conditions, with the exception that ATP activityfor dynamic conditions was not detectable before 12 h and peakactivity was not observed until 30 h. Three stages of biofilm devel-opment were identified from these ATP activity data. The first stage(stage 1) represents early development, when metabolic activity isincreasing (6 and 12 h under static and dynamic conditions, respec-tively). The second stage (stage 2) represents a biofilm at peakmetabolic activity (12 and 30 h under static and dynamic condi-tions, respectively). The third stage (stage 3) represents the stable,lower metabolic activity of a mature biofilm (48 h under both staticand dynamic conditions).

The biofilm amount should increase with maturity of thebiofilms as EPS is produced for cell adhesion to surfaces and pro-tection from environmental stresses [41]. Therefore, the amountof biofilms at selected stages was determined for maturity understatic and dynamic growth conditions using the CV and total car-bohydrate assays, respectively (Fig. 2). Under static conditions, theamounts of biofilms at stage 2 (12 h) and stage 3 (48 h) were 6 and 5

Fig. 2. Biofilm amount of P. putida KT2440 biofilms at different stages. A600 is theabsorbance at 600 nm for CV measured in 96-well plate experiments; total carbo-hydrate was measured in CDC reactor.

times higher than at stage 1 (6 h) (p = 0.016 for 6 h vs. 12 h; p = 0.006for 6 h vs. 48 h). Similarly, under dynamic conditions, the amountsof biofilms at stage 2 (30 h) and stage 3 (48 h) were 2 and 3 timeshigher, respectively, than at stage 1 (12 h) (p = 0.019 for 12 vs. 30 h;p = 0.008 for 12 vs. 48 h).

Biofilms at stage 2 had higher amounts of biomass than thoseat stage 3 under static conditions, while it was the opposite underdynamic conditions. This might be due to different biofilm quan-tification methods used (the CV assay for static conditions and thetotal carbohydrate assay for dynamic conditions). The total carbo-hydrate assay measured only carbohydrate from the EPS of biofilms,whereas the CV assay measured total biomass from live cells, deadcells, and EPS. According to the activity from Fig. 1, biofilms in stage2 should have much higher cell numbers than in stage 3, which waslikely to give more CV staining. However, there was no statisticaldifference between the amounts of biofilms in stages 2 and 3 underboth conditions (p = 0.401 and p = 0.093 under static and dynamicconditions, respectively).

As biofilms mature, they produce not only more EPS but alsodifferent components. Among various components, curli is a pro-tein component used for bacterial adhesion to surfaces [42]. Sixproteins encoded by the csgBA and csgDEFG operons contribute tothe formation of curli fiber [43]. For P. putida KT2440, the csgA geneencodes the major subunit of curli. During irreversible attachment,the csgA gene should be highly expressed [24]. A polysaccharidecomponent of EPS, alginate, also contributes to the development,structure, and resistance of biofilms [44]. The alginate biosynthe-sis protein is encoded by the alg8 gene for P. putida KT2440. Asbiofilms produce polysaccharides to form the structure of biofilms,the expression of alg8 should be higher in mature biofilm.

Fig. 3 shows that, analogous to EPS levels, expressions of csgAand alg8 were higher in biofilms at later stages. csgA expres-sion increased significantly between stages 1 and 2 (p = 0.001 andp = 0.013 under static and dynamic conditions, respectively) andagain between stages 2 and 3 (p = 0.024 and p = 0.030 under staticand dynamic conditions, respectively). alg8 expression did not dif-fer between stages 1 and 2, but increased significantly between

Fig. 3. Expressions of csgA and alg8 genes of biofilms at different stages.

P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133 131

Fig. 4. Effect of AgNPs on biofilms at different stages.

stages 2 and 3 (p = 0.002 and p = 0.002 under static and dynamicconditions, respectively).

The earlier increase in csgA expression relative to alg8 may beexplained by the specific roles these genes play in biofilm mat-uration. Bacteria had to adhere to the surface (expression of csgAgene) before they could form the structure of biofilms by producingpolysaccharide components such as alginate (expression of alg8).This resulted in different levels of gene expressions at differentstages as seen in Fig. 3. Collectively, the EPS and gene expres-sion data support the conclusion that biofilm development stagesidentified from ATP activity data represent stages of increasing mat-uration. The second part of this study examined the effect of AgNPson biofilms at different stages of maturity.

3.3. Effect of biofilm maturity on the susceptibility to AgNPs

The effect of AgNPs on biofilms was measured as a reductionin ATP activity relative to that in a non-treated control. A platecount was used in addition to ATP activity for biofilms grown under

dynamic conditions. Fig. 4 shows that the least mature biofilms(stage 1) were most susceptible to AgNPs, with greater than 90%reduction in ATP activity and plate count. ATP was not reducedin the more mature stages 2 and 3 biofilms under static conditions,and small reductions in ATP and plate count were observed in stages2 and 3 biofilms grown under dynamic conditions.

Several factors may explain the increased resistance of maturebiofilms to AgNPs. Firstly, bacterial cells in mature biofilms arelikely to be in the stationary growth phase and, therefore, less sus-ceptible to antimicrobial agents [45]. To prove this, the exposureexperiment was conducted on planktonic cells at different stages(Fig. S1, Supplementary data). With the same starting cell num-ber, after 3 h of exposure to 20 mg/l of AgNPs, the log-phase cells(6 h) of P. putida KT2440 were not observed by the plate countmethod while the stationary-phase cells (16 h) were still at 105

CFU, showing more tolerance to AgNPs. Secondly, cells that die inthe outer layers of mature biofilms could provide nutrients thatenhance the growth of cells in deeper layers [46]. A previous studyon the effects of single-walled carbon nanotubes on Escherichia coli

Fig. 5. ATP activity and biofilm amount (represented by A600 which is the absorbance at 600 nm for CV) before and after EDTA treatment.

132 P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133

Fig. 6. Effect of EPS on biofilm susceptibility to AgNPs.

biofilm showed that dead bacterial cells could cause aggregation ofthe nanotubes and at the same time release intracellular substancesto serve as nutrients for other cells [47]. Also, the high thickness orhigh amount of EPS in mature biofilms may have a role in transportlimitations of AgNPs through biofilms.

3.4. Role of EPS in biofilm susceptibility to AgNPs

To determine how EPS affects biofilm susceptibility to AgNPs,the EPS of biofilms were partly removed by EDTA [48]. Fig. 5 showsthe reduction of ATP activity and biofilm amount after EDTA treat-ment. There was a statistically significant reduction of ATP in allstages of biofilms (6 h: p = 0.008, 12 h: p = 0.0005, 48 h: p = 0.009).However, the amount of 6 h biofilms (based on the CV assay results)did not get reduced by the EDTA treatment, while the older biofilmsshowed high reduction of biomass (48 h: p = 0.003). After EPS strip-ping by EDTA treatment, the biofilms were exposed to AgNPs andthe effect was measured by reduction in ATP activity (Fig. 6). Theresults showed the critical role of EPS in the protection of biofilmcommunities from AgNPs. The EPS-stripped biofilms in all threestages showed significantly higher reduction in ATP than biofilmswith intact EPS (control) at every time point of exposure (p < 0.05).

To demonstrate that EDTA did not make the planktonic cellsmore susceptible to AgNPs after 3 h of treatment, a test was con-ducted and the results are presented in Fig. S2 (Supplementarydata). Between cells with EDTA treatment and without EDTA treat-ment, there was no significant difference of ATP percentage after1 h of AgNP exposure (p = 0.971). However, after 2 h of AgNP expo-sure, cells treated with EDTA showed lower susceptibility to AgNPsthan cells without EDTA treatment (p = 0.031). From these results,it can be concluded that EDTA did not increase the susceptibility ofcells to AgNPs. Therefore, the reduction in ATP of biofilms after EPSstripping should be from the EPS removal. Similarly, in a study oneffects of AgNPs on wastewater biofilms, greater bacterial reduc-tions were achieved after loosely-bound EPS was removed [20].This is consistent with the findings by Peulen and Wilkinson thatEPS density reduces the diffusion of AgNPs into biofilms [19].

4. Conclusions

In this study, we characterize three stages of biofilm maturitybased on cell number, expression of biofilm-associated genes, andEPS amount, and we show that more mature biofilms have greatlyreduced susceptibility to AgNPs compared to immature biofilms.These findings have important implications for environmentalsystems where biofilm maturity varies, including wastewater treat-ment plants at different phases of operation. AgNPs will be less toxicin steady-state systems with mature biofilms, but systems duringstart-up, when biofilms are becoming established, will be vulnera-ble to AgNPs. It should be noted that this study only focused on theeffect of AgNPs on single-species biofilms growing in batch con-ditions. In environment or wastewater treatment system, various

species of bacteria are present together under the continuous con-ditions. It is possible that stage of maturity will be different from theresults in this study, leading to different effect of AgNPs on biofilms.Therefore, these two points should be considered for future stud-ies in order to better understand the effect of AgNPs on biofilms atdifferent stages of maturity.

Acknowledgments

This work was supported by the 90th Anniversary of Chula-longkorn University Fund (Ratchadaphiseksomphot EndowmentFund) and was conducted under the research cluster “Control ofEmerging Micropollutants in Aquacultural and Feedstock Industry”granted by Center of Excellence for Hazardous Substance Man-agement (HSM) and Special Task Force for Activating Research(STAR) program of Chulalongkorn University. The iCycler iQ qPCRdetection system for RT-qPCR was purchased with grant 2009-35201-05010 from the USDA/NIFA. The authors would like to thankShane Stafslien and Justin Daniels from the Center for NanoscaleScience and Engineering, North Dakota State University, for sug-gestions and opinions on CDC reactor operation.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at http://dx.doi.org/10.1016/j.jhazmat.2015.02.073.

References

[1] J. Tian, K.K.Y. Wong, C.-M. Ho, C.N. Lok, W.Y. Yu, C.M. Che, J.F. Chiu, P.K.H. Tam,Topical delivery of silver nanoparticles promotes wound healing,ChemMedChem 2 (1) (2007) 129–136.

[2] T.M. Benn, P. Werterhoff, Nanoparticle silver released into water fromcommercially available sock fabrics, Environ. Sci. Technol. 42 (2008)4133–4139.

[3] D. Roe, B. Karandikar, N. Bonn-Savage, B. Gibbins, J.B. Roullet, Antimicrobialsurface functionalization of plastic catheters by silver nanoparticles, J.Antimicrob. Chemother. 61 (4) (2008) 869–876.

[4] D. Gangadharan, K. Harshvardan, G. Gnanasekar, D. Dixit, K.M. Popat, P.S.Anand, Polymeric microspheres containing silver nanoparticles as abactericidal agent for water disinfection, Water Res. 44 (18) (2010)5481–5487.

[5] J. Dobias, R. Bernier-Latmani, Silver release from silver nanoparticles innatural waters, Environ. Sci. Technol. 47 (9) (2013) 4140–4146.

[6] I. Sondi, B. Salopek-Sondi, Silver nanoparticles as antimicrobial agent: a casestudy on E. coli as a model for gram-negative bacteria, J. Colloid Interface Sci.275 (1) (2004) 177–182.

[7] J.R. Morones, J.L. Elechiguerra, A. Camacho, K. Holt, J.B. Kouri, J.T. Ramírez, M.J.Yacaman, The bactericidal effect of silver nanoparticles, Nanotechnology 16(10) (2005) 2346–2353.

[8] Q.L. Feng, J. Wu, G.Q. Chen, F.Z. Cui, T.N. Kim, J.O. Kim, A mechanistic study ofthe antibacterial effect of silver ions on Escherichia coli and Staphylococcusaureus, J. Biomed. Mater. Res. 52 (4) (2000) 662–668.

[9] A. Kora, J. Arunachalam, Assessment of antibacterial activity of silvernanoparticles on Pseudomonas aeruginosa and its mechanism of action, WorldJ. Microbiol. Biotechnol. 27 (5) (2011) 1209–1216.

[10] H.C. Flemming, J. Wingender, The biofilm matrix, Nat. Rev. Microbiol. 8 (2010)623–633.

P. Thuptimdang et al. / Journal of Hazardous Materials 290 (2015) 127–133 133

[11] P.S. Stewart, M.J. Franklin, Physiological heterogeneity in biofilms, Nat. Rev.Microbiol. 6 (2008) 199–210.

[12] C. Beloin, J.M. Ghigo, Finding gene-expression patterns in bacterial biofilms,Trends Microbiol. 13 (2005) 16–19.

[13] A. Pal, A.K. Paul, Microbial extracellular polymeric substances: centralelements in heavy metal bioremediation, Indian J. Microbiol. 48 (2008) 49–64.

[14] N. Høiby, T. Bjarnsholt, M. Givskov, S. Molin, O. Ciofu, Antibiotic resistance ofbacterial biofilms, Int. J. Antimicrob. Ag. 35 (4) (2010) 322–332.

[15] A. González, S. Mombo, J. Leflaive, A. Lamy, O. Pokrovsky, J.L. Rols, Silvernanoparticles impact phototrophic biofilm communities to a considerablyhigher degree than ionic silver, Environ. Sci. Pollut. (2014) 1–13,http://dx.doi.org/10.1007/s11356-014-3978-1.

[16] M.B. Habash, A.J. Park, E.C. Vis, R.J. Harris, C.M. Khursigara, Synergy of silvernanoparticles and aztreonam against Pseudomonas aeruginosa PAO1 biofilms,Antimicrob. Agents Chemother. 58 (2014) 5818–5830.

[17] H.J. Park, S. Park, J. Roh, S. Kim, K. Choi, J. Yi, Y. Kim, J. Yoon,Biofilm-inactivating activity of silver nanoparticles: a comparison with silverions, J. Ind. Eng. Chem. 19 (2013) 614–619.

[18] O. Choi, C.P. Yu, G.E. Fernández, Z. Hu, Interactions of nanosilver withEscherichia coli cells in planktonic and biofilm cultures, Water Res. 44 (20)(2010) 6095–6103.

[19] T.O. Peulen, K.J. Wilkinson, Diffusion of nanoparticles in a biofilm, Environ.Sci. Technol. 45 (8) (2011) 3367–3373.

[20] Z. Sheng, Y. Liu, Effects of silver nanoparticles on wastewater biofilms, WaterRes. 45 (18) (2011) 6039–6050.

[21] N. Joshi, B.T. Ngwenya, C.E. French, Enhanced resistance to nanoparticletoxicity is conferred by overproduction of extracellular polymeric substances,J. Hazar. Mater. 241–242 (2012) 363–370.

[22] R.D. Monds, G.A. O’Toole, The developmental model of microbial biofilms: tenyears of a paradigm up for review, Trends Microbiol. 17 (2009) 73–87.

[23] B.A. Lazazzera, Lessons from DNA microarray analysis: the gene expressionprofile of biofilms, Curr. Opin. Microbiol. 8 (2005) 222–227.

[24] B.M. Prüß, C. Besemann, A. Denton, A.J. Wolfe, A complex transcriptionnetwork controls the early stages of biofilm development by Escherichia coli, J.Bacteriol. 188 (11) (2006) 3731–3739.

[25] A. Ghafoor, I.D. Hay, B.H.A. Rehm, Role of exopolysaccharides in Pseudomonasaeruginosa biofilm formation and architecture, Appl. Environ. Microbiol. 77(2011) 5238–5246.

[26] M. Tré-Hardy, C. Macé, N. El Manssouri, F. Vanderbist, H. Traore, M.J.Devleeschouwer, Effect of antibiotic co-administration on young and maturebiofilms of cystic fibrosis clinical isolates: the importance of the biofilmmodel, Int. J. Antimicrob. Agents 33 (1) (2009) 40–45.

[27] Y. Shen, S. Stojicic, M. Haapasalo, Antimicrobial efficacy of chlorhexidineagainst bacteria in biofilms at different stages of development, J. Endod. 37 (5)(2011) 657–661.

[28] C.S. Beauchamp, D. Dourou, I. Geornaras, Y. Yoon, J.A. Scanga, K.E. Belk, G.C.Smith, G.J.E. Nychas, J.N. Sofos, Sanitizer efficacy against Escherichia coliO157:H7 biofilms on inadequately cleaned meat-contact surface materials,Food Prot. Trends 32 (4) (2012) 173–182.

[29] M. Martínez-Gil, J.M. Quesada, M.I. Ramos-González, M.I. Soriano, R.E. deCristóbal, M. Espinosa-Urgel, Interplay between extracellular matrixcomponents of Pseudomonas putida biofilms, Res. Microbiol. 164 (5) (2013)382–389.

[30] K.E. Nelson, C. Weinel, I.T. Paulsen, R.J. Dodson, H. Hilbert, V.A.P. Martins dosSantos, D.E. Fouts, S.R. Gill, M. Pop, M. Holmes, L. Brinkac, M. Beanan, R.T.DeBoy, S. Daugherty, J. Kolonay, R. Madupu, W. Nelson, O. White, J. Peterson,H. Khouri, I. Hance, P.C. Lee, E. Holtzapple, D. Scanlan, K. Tran, A. Moazzez, T.Utterback, M. Rizzo, K. Lee, D. Kosack, D. Moestl, H. Wedler, J. Lauber, D.Stjepandic, J. Hoheisel, M. Straetz, S. Heim, C. Kiewitz, J. Eisen, K.N. Timmis, A.

Düsterhöft, B. Tümmler, C.M. Fraser, Complete genome sequence andcomparative analysis of the metabolically versatile Pseudomonas putidaKT2440, Environ. Microbiol. 4 (12) (2002) 799–808.

[31] O.K. Choi, K.K. Deng, N.J. Kim, L. Ross Jr., R.Y. Surampalli, Z.Q. Hu, Theinhibitory effects of silver nanoparticles, silver ions, and silver chloridecolloids on microbial growth, Water Res. 42 (12) (2008) 3066–3074.

[32] L. Mulfinger, S.D. Solomon, M. Bahadory, A.V. Jeyarajasingam, S.A. Rutkowsky,C. Boritz, Synthesis and study of silver nanoparticles, J. Chem. Educ. 84 (2)(2007) 322–325.

[33] C. Beer, R. Foldbjerg, Y. Hayashi, D.S. Sutherland, H. Autrup, Toxicity of silvernanoparticles – nanoparticle or silver ion? Toxicol. Lett. 208 (2012)286–292.

[34] P. Sule, T. Wadhawan, N.J. Carr, S.M. Horne, A.J. Wolfe, B.M. Prü�, Acombination of assays reveals biomass differences in biofilms formed byEscherichia coli mutants, Lett. Appl. Microbiol. 49 (2009) 299–304.

[35] T. Masuko, A. Minami, N. Iwasaki, T. Majima, S.I. Nishimura, Y.C. Lee,Carbohydrate analysis by a phenol–sulfuric acid method in microplateformat, Anal. Biochem. 339 (1) (2005) 69–72.

[36] S. Horne, B. Prüß, Global gene regulation in Yersinia enterocolitica: effect ofFliA on the expression levels of flagellar and plasmid-encoded virulencegenes, Arch. Microbiol. 185 (2) (2006) 115–126.

[37] T.D. Schmittgen, K.J. Livak, Analyzing real-time PCR data by the comparativeCT method, Nat. Protoc. 3 (2008) 1101–1108.

[38] W.K. Jung, H.C. Koo, K.W. Kim, S. Shin, S.H. Kim, Y.H. Park, Antibacterialactivity and mechanism of action of the silver ion in Staphylococcus aureusand Escherichia coli, Appl. Environ. Microbiol. 74 (7) (2008) 2171–2178.

[39] I. Bourven, G. Costa, G. Guibaud, Qualitative characterization of the proteinfraction of exopolymeric substances (EPS) extracted with EDTA from sludge,Bioresour. Technol. 104 (2012) 486–496.

[40] Z. Yuan, J. Li, L. Cui, B. Xu, H. Zhang, C.-P. Yu, Interaction of silver nanoparticleswith pure nitrifying bacteria, Chemosphere 90 (4) (2013) 1404–1411.

[41] B. Vu, M. Chen, R. Crawford, E. Ivanova, Bacterial extracellular polysaccharidesinvolved in biofilm formation, Molecules 14 (7) (2009) 2535–2554.

[42] Z. Saldana, J. Xicohtencatl-Cortes, F. Avelino, A.D. Phillips, J.B. Kaper, J.L.Puente, J.A. Girón, Synergistic role of curli and cellulose in cell adherence andbiofilm formation of attaching and effacing Escherichia coli and identificationof Fis as a negative regulator of curli, Environ. Microbiol. 11 (4) (2009)992–1006.

[43] M.M. Barnhart, M.R. Chapman, Curli biogenesis and function, Annu. Rev.Microbiol. 60 (1) (2006) 131–147.

[44] E. Diaz, H. Haaf, A. Lai, J. Yadana, Role of alginate in gentamicin antibioticsusceptibility during the early stages of Pseudomonas aeruginosa PAO1 biofilmestablishment, J. Exp. Microbiol. Immunol. 15 (2011) 71–78.

[45] J.N. Anderl, J. Zahller, F. Roe, P.S. Stewart, Role of nutrient limitation andstationary-phase existence in Klebsiella pneumoniae biofilm resistance toampicillin and ciprofloxacin, Antimicrob. Agents Chemother. 47 (4) (2003)1251–1256.

[46] A. Ito, A. Taniuchi, T. May, K. Kawata, S. Okabe, Increased antibiotic resistanceof Escherichia coli in mature biofilms, Appl. Environ. Microbiol. 75 (12) (2009)4093–4100.

[47] D.F. Rodrigues, M. Elimelech, Toxic effects of single-walled carbon nanotubesin the development of E. coli biofilm, Environ. Sci. Technol. 44 (12) (2010)4583–4589.

[48] Z. Liang, W. Li, S. Yang, P. Du, Extraction and structural characteristics ofextracellular polymeric substances (EPS), pellets in autotrophic nitrifyingbiofilm and activated sludge, Chemosphere 81 (5) (2010) 626–632.