ww.sciencedirect.com

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/watres

Microbial community response of nitrifyingsequencing batch reactors to silver, zero-valentiron, titanium dioxide and cerium dioxidenanomaterials

Yanjun Ma 1, Jacob W. Metch, Eric P. Vejerano, Ian J. Miller 2,Elena C. Leon 3, Linsey C. Marr, Peter J. Vikesland, Amy Pruden*

Via Department of Civil and Environmental Engineering, 418 Durham Hall, Virginia Tech, Blacksburg, VA 24061,

USA

a r t i c l e i n f o

Article history:

Received 23 May 2014

Received in revised form

18 August 2014

Accepted 6 September 2014

Available online 28 September 2014

Keywords:

Environmental impacts of nano-

materials

Waste water treatment

Nitrification

Pyrosequencing

q-PCR

* Corresponding author. Tel.: þ1 540 231 398E-mail addresses: yanjunm6@vt.edu (Y.

Miller), ecleon@stanford.edu (E.C. Leon), lma1 Current affiliation: School of Environmen

Jiangsu 221116, China.2 Current affiliation: University of Wiscon3 Current affiliation: Stanford University, S

http://dx.doi.org/10.1016/j.watres.2014.09.0080043-1354/© 2014 Elsevier Ltd. All rights rese

a b s t r a c t

As nanomaterials in consumer products increasingly enter wastewater treatment plants,

there is concern that they may have adverse effects on biological wastewater treatment.

Effects of silver (nanoAg), zero-valent iron (NZVI), titanium dioxide (nanoTiO2) and

cerium dioxide (nanoCeO2) nanomaterials on nitrification and microbial community

structure were examined in duplicate lab-scale nitrifying sequencing batch reactors

(SBRs) relative to control SBRs that received no nanomaterials or ionic/bulk analogs.

Nitrification function was not measurably inhibited in the SBRs by any of the materials

as dosing was initiated at 0.1 mg/L and sequentially increased every 14 days to 1, 10, and

20 mg/L. However, SBRs rapidly lost nitrification function when the Agþ experiment was

repeated at a continuous high load of 20 mg/L. Shifts in microbial community structure

and decreased microbial diversity were associated with both sequential and high loading

of nanoAg and Agþ, with more pronounced effects for Agþ. Bacteroidetes became more

dominant in SBRs dosed with Agþ, while Proteobacteria became more dominant in SBRs

dosed with nanoAg. The two forms of silver also had distinct effects on specific bacterial

genera. A decrease in nitrification gene markers (amoA) was observed in SBRs dosed with

nanoAg and Agþ. In contrast, impacts of NZVI, nanoTiO2, nanoCeO2 and their analogs on

microbial community structure and nitrification gene markers were limited. TEM-EDS

analysis indicated that a large portion of nanoAg remained dispersed in the activated

sludge and formed AgeS complexes, while NZVI, nanoTiO2 and nanoCeO2 were mostly

aggregated and chemically unmodified. Overall, this study suggests a high threshold of

0; fax: þ1 540 231 7916.Ma), jmetch@vt.edu (J.W. Metch), evejer1@vt.edu (E.P. Vejerano), ijmiller2@wisc.edu (I.J.rr@vt.edu (L.C. Marr), pvikes@vt.edu (P.J. Vikesland), apruden@vt.edu (A. Pruden).tal Science and Spatial Informatics, China University of Mining and Technology, Xuzhou,

sin, School of Pharmacy, Madison, WI 53705-2222, USA.tanford, CA 94305, USA.

rved.

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 788

the four nanomaterials in terms of exerting adverse effects on nitrification function.

However, distinct microbial community responses to nanoAg indicate potential long-

term effects.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

As nanotechnology-based consumer products enter into

widespread use, large quantities of engineered nano-

materials (ENMs) will enter wastewater treatment plants

(Brar et al., 2010; Mueller and Nowack, 2008). Wastewater

treatment processes have been observed to remove the ma-

jority of ENMs in aqueous effluents, primarily concentrating

them into the activated sludge phase (Kiser et al., 2009; Li

et al., 2013). Compared with traditional bulk materials, the

nano-size range imparts unique properties, including greater

reactivity and thus more potential toxicity to biological sys-

tems (Nel et al., 2006). Toxicity of ENMs to microbes has been

well-documented in pure culture studies (Diao and Yao,

2009; Li et al., 2011; Pelletier et al., 2010). Therefore, there is

concern that the accumulation of ENMs in activated sludge

could have adverse effects on microbes responsible for bio-

logical wastewater treatment, especially nitrifying bacteria,

which conduct the critical function of nitrification and are

known for slow growth rate and high sensitivity to toxic

chemicals.

To date, silver nanoparticles have been the most widely

studied with respect to environmental implications and are

thought to be among the ENMs most likely to have an effect.

Toxic effects of silver nanoparticles (nanoAg) have been

observed on Nitrosomonas europaea in pure culture (Arnaout

and Gunsch, 2012; Radniecki et al., 2011) and enriched nitri-

fying microorganisms (Choi et al., 2008) at 0.2e20 mg/L.

However, the extent of reported nitrification inhibition by

nanoAg has varied in lab-scale reactors simulating different

biological wastewater treatment processes. While little nitri-

fication inhibition was observed in a membrane bioreactor

receiving 0.1mg/L nanoAg for over 60 days (Zhang et al., 2014),

nitrification inhibition was observed in response to a 12-h

period of nanoAg shock loading to reach a final peak silver

concentration of 0.75 mg/L in a Modified Ludzacke-Ettinger

process (Liang et al., 2010). In a recent study, a brief decrease

of nitrification efficiency was observed in sequencing batch

reactors (SBR) receiving pulse or continuous loading of 0.2 mg/

L nanoAg, but the function quickly recovered (Alito and

Gunsch, 2014). Interaction of nanoAg with wastewater con-

stituents could alter toxicity (Choi et al., 2009; Xiu et al., 2011).

Moreover, considering that wastewater treatment is a

microbial-ecologically driven process, insight into effects of

nanoAg on the broader microbial community is needed. A

recent study used high-throughput pyrosequencing to quan-

tify shifts of microbial community compositions in activated

sludge microcosms exposed to nanoAg or Agþ (Yang et al.,

2014). However, pyrosequencing has not been applied to

quantitatively characterize response of microbial

communities to nanoAg, or to benchmark with respect to

other ENMs, under more realistic wastewater treatment

conditions.

Previous studies have generally focused on effects of

nanoAg at sub-ppm levels. As ENMs entering wastewater

streams are expected to continue to expand in type and con-

centration (Gottschalk et al., 2009), it is necessary to identify

the threshold of ENMs that will exert adverse effects on bio-

logical wastewater treatment processes and also to extend

studies to other ENMs commonly occurring in wastewater.

This study investigated the effects of silver (nanoAg), zero-

valent iron (NZVI), titanium dioxide (nanoTiO2) and cerium

dioxide (nanoCeO2) nanomaterials in duplicate lab-scale ni-

trifying SBRs in two phases of experiments: 1.) sequentially

increased loading of 0.1, 1, 10 and 20 mg/L nanomaterials

every 14 days; and 2.) continuous high loading at 20 mg/L for

nanoAg and nanoCeO2, which indicated the greatest potential

effects based on the sequential loading experiments. Effects of

nanomaterials were compared to control SBRs that were

undosed or received bulk or ionic analogs (Agþ, Fe2þ, bulkTiO2,

or bulkCeO2). Nitrification performance was monitored along

with the abundance of nitrification gene markers by quanti-

tative polymerase chain reaction (q-PCR). Response of micro-

bial communities was characterized by pyrosequencing of

bacterial 16S rRNA genes. Transmission Electron Microscopy

(TEM) equipped with Energy Dispersive X-ray Spectroscopy

(EDS) was used to examine aggregation state and interaction

of nanomaterials with activated sludge to understand poten-

tial relationships between nanomaterial transformation and

toxicity.

2. Materials and methods

2.1. Preparation of nanomaterial suspensions

NanoAg was synthesized by the sodium citrate reduction

method, yielding a concentration of 100 mg/L (Lee and Meisel,

1982). NZVI dispersion with a combination of a biodegradable

organic and inorganic stabilizers was purchased from NANO

IRON s.o.r. (NANOFER 25S, Rajhrad, Czech Republic). Anatase

TiO2 and CeO2 nanopowder (SigmaeAldrich, Saint Louis, MO)

were dispersed into nanopure water by sonication for 20 min

(90 W, 20 KHz, 20 �C) at a concentration of 100 mg/L. Freshly

prepared nanomaterial dispersions were used to dose SBRs.

The properties of the prepared nanomaterial dispersions were

confirmed and characterized by TEM. The average sizes of the

nanoparticles were 52 ± 12 nm for nanoAg, 46 ± 10 nm for

NZVI, 21 ± 12 nm for nanoTiO2, and 33 ± 12 nm for nanoCeO2

based on counting 70 particles per TEM image.

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7 89

2.2. Set-up of SBRs

Lab-scale SBRs were set up in 3 L glass beakers (Fisher,

Suwanee, GA) with an active volume of 2 L in a 20 �C tem-

perature controlled room. Each SBR was operated with two

12 h cycles per day with solids retention time (SRT) of 23 days

and hydraulic retention time (HRT) of 2 days. Each cycle

consisted of 0.1 h feeding, 10.82 h aeration, 0.05 hmixed liquor

wasting (solids wasting), 0.75 h settling, and 0.33 h decant.

Waste sludge was returned at a rate of 0.10 to maintain the

biomass. The SBRs were seeded with return nitrifying acti-

vated sludge from a local municipal wastewater treatment

plant. A synthetic wastewater (Table S1 and S2) served as feed

with an initial chemical oxygen demand (COD) of 450 mg/L

and total nitrogen (TN) of 55mg/L. Stable operating conditions

were defined as: pH in the range of 7.4e7.7, mixed liquor

suspended solids (MLSS) of 3000e4000mg/L, andmixed liquor

volatile suspended solids (MLVSS) of 2200e3400 mg/L. SBRs

were acclimated until ammonium nitrogen removal of >98%was achieved (typically 3e5 weeks) before dosing nano-

materials. Each nanomaterial was tested individually, with

the SBRs re-started prior to the next experiment, in order to

attain equivalent performance among the SBRs and avoid ef-

fects of prior nanomaterials before testing the subsequent

nanomaterial.

2.3. Sequential load of nanomaterials

Three conditions were examined in duplicate SBRs: nano-

material, corresponding bulk or ionic material [Agþ as AgNO3

(Fisher, Suwanee, GA); Fe2þ as FeSO4 (Fisher); bulkTiO2

(~44 mm); bulkCeO2 (~5 mm) (SigmaeAldrich, Saint Louis, MO)],

and undosed controls. The dosing of nanomaterials and bulk/

ionic materials was initiated at an influent concentration of

0.1 mg/L and sequentially increased every 14 days to 1 mg/L,

10 mg/L and finally 20 mg/L. Nanomaterial and bulk/ionic

material dispersions or solutions were directly dosed to the

reactors during the feed cycle.

2.4. High load of nanomaterials

Based on results of sequential load experiments, nanoAg,

nanoCeO2 and the bulk/ionic analogs were selected for high

load experiments. Upon reaching stable operating conditions,

20 mg/L nanoAg, Agþ, nanoCeO2 and bulkCeO2 dispersions or

solutions were dosed to each reactor during the feed cycle for

a total of 42 days.

2.5. Analytical methods

Nitrification was monitored by quantifying concentrations of

nitrite nitrogenðNO2� �NÞ, nitrate nitrogen ðNO3

� �NÞ and

ammonia nitrogen (NH3eN) in the aqueous effluent by Ion

Chromatography (Thermo Scientific, Tewksbury, MA) and

using an Ammonia Nitrogen Test Kit (Hach, Loveland, CO). pH,

MLSS and MLVSS of SBRs were monitored regularly as in-

dicators of reactor performance using standard methods

(APHA, 1998). Concentrations of nanomaterials and bulk/ionic

materials in aqueous effluent were analyzed using ICP-MS

(Thermo Scientific, Tewksbury, MA) to quantify total Ag, Fe,

Ti and Ce. Aqueous effluents were filtered through 0.45 mm

mixed-cellulose ester (MCE) membrane filters (EMD Millipore,

Billerica, MA) prior to analysis.

2.6. q-PCR

Activated sludge was sampled from each SBR just prior to

commencing dosing and every 7 days during both sequential

load and high load experiments. Total nucleic acid was

extracted in duplicate from 175 mL activated sludge using a

MagMAX™ Total Nucleic Acid Isolation Kit (Life Technologies,

Grand Island, NY) according to manufacturer protocol. Q-PCR

was used to quantify amoA genes encoding the active-site

polypeptide of ammonia monooxygenase, 16S rRNA genes of

Nitrobacter and Nitrospira (two common genera of NOB), and

universal bacterial 16S rRNA genes. Specific primers and re-

action conditions are described in Supporting Information

(Tables S3eS4). All q-PCR standard curves were constructed

from serial dilutions of cloned genes ranging from 108 to 102

gene copies per mL. The presence of PCR inhibitors was

checked by serially diluting a representative sub-set of sam-

ples and comparing PCR efficiencies with standards. It was

found that a 1:50 dilution worked well to minimize inhibitory

effects and was applied across all samples. Samples were

analyzed in triplicate with a standard curve and negative

control included in each run. Averages and standard de-

viations of all data were determined using Microsoft Excel®

2007. t tests were conducted using R 3.0.3 software (http://

cran.r-project.org/bin/windows/base/) to determine statisti-

cal differences between samples (p < 0.05).

2.7. Pyrosequencing

Nucleic acid extracts from just prior to dosing and at the end

of both sequential load and high load experiments were

diluted to 20 ng/mL and subjected to pyrosequencing of 16S

rRNA genes using primers 341F/907R on a Roche 454 FLX Ti-

tanium platform (Roche, Nutley, NJ) by the Research and

Testing Laboratory (Lubbock, TX). Duplicate extracts were

combined for pyrosequencing analysis. The retrieved se-

quences were processed using MOTHUR 1.32.1 software ac-

cording to publishedmethods (Schloss et al., 2011) and on-line

454 SOP (http://www.mothur.org/wiki/454_SOP). In brief,

primers and barcodes were removed, and sequences with

ambiguous base pairs, more than two mismatches in the

primer sequence, more than one mismatch in the barcode

sequence, more than nine homopolymers, or less than 150

base pairs were filtered out. Alignment of sequences was

carried out using the SILVA reference database (http://www.

mothur.org/wiki/Silva_reference_alignment). Chimeras were

removed using UCHIME. Sequences corresponding to mito-

chondria, chloroplast, Archaea, Eukarya or unknown were

removed. This processing resulted in an average of 3541 se-

quences per sample (n ¼ 64, sd ¼ 2108). The sequences were

clustered into operational taxonomy units (OTUs) at 3% cutoff

and further assigned to phylotypes from phylum to genus

level. The outputs were subsampled to the smallest library

size (n ¼ 906). A total of 315 genera belonging to 23 bacterial

phyla were identified. At genus level, inverse Simpson di-

versity index (Peet, 1974) was calculated using MOTHUR, and

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 790

multidimensional scaling (MDS) analysis was performed

using Primer 6 software (Primer-E, Plymouth, UK) to compare

phylogenetic distance of microbial communities based on the

calculation of BrayeCurtis dissimilarity (Bray and Curtis,

1957). Heat maps displaying relative abundance of specific

bacteria taxonomies were generated in Explicet 2.8.1 (http://

www.explicet.org/).

2.8. TEM-EDS mapping

Activated sludge from SBRs dosed with nanoAg, NZVI, nano-

TiO2, and nanoCeO2 was freeze-dried for TEM grid preparation

as previously described (Kim et al., 2010). A JEOL 2100 TEM

operated at 200 kV was used to detect and characterize ag-

gregation state of nanoparticles in the activated sludge. The

microscope was equipped with EDS, which was used to

analyze the chemistry of the particles of interest under the

scanning TEM (STEM) mode.

3. Results

3.1. Effects of nanomaterials on nitrification function

Nitrification functionality was not impacted by dosing nano-

materials or bulk/ionic analogs when the load was sequen-

tially increased from 0.1 to 1, 1 to 10 or 10 to 20 mg/L, with

NH3eN and NO2�eN effluent concentrations remaining near

the detection limit in all SBRs (data not shown). There was

similarly no observable effect of nanomaterials on ammonia

removal in the high load experiments for nanoAg, nano CeO2

or bulk CeO2 (data not shown). However, in the high load Agþ

condition, effluent NH3eN rose above 10 mg/L and NO3-eN

decreased to ~15 mg/L within seven days (Fig. 1). NO2-eN

concentrations remained near the detection limit throughout

dosing, indicating that NO2- oxidizationwas not impacted even

by Agþ (Fig. 1). The SBRs receiving the high load of Agþ failed at

about 14 days and were no longer operational, with high COD

Fig. 1 e Effluent concentrations of (a) NH3eN, NO3¡-N, and

(b) NO2¡eN in duplicate SBRs dosed with continuous high

load of 20 mg/L nanoAg or Agþ relative to undosed

controls.

(109 and 92 mg/L) and turbid effluent, while the remaining

SBRs continued to operate at high load with no apparent

inhibitory effects on nitrification.

3.2. Changes in relative abundance of nitrifying bacteria

To assess the abundance of nitrifying bacteria proportional to

the overall bacterial population, copy numbers of amoA and

Nitrobacter and Nitrospira 16S rRNA genes were normalized to

universal bacterial 16S rRNA genes. In the sequential load

experiments, relative abundance of amoA genes decreased in

SBRs dosed with nanoAg or Agþ at the 10 and 20 mg/L level

relative to the undosed control (Fig. 2, p < 0.05) and also in

NZVI, Fe2þ, nanoCeO2 and bulkCeO2 SBRs at the final 20 mg/L

dosing point (Fig. 2, p < 0.05). amoA genes were not signifi-

cantly impacted by dosing of nanoTiO2 or bulkTiO2 (Fig. S2a,

p > 0.05). Relative abundance of Nitrobacter 16S rRNA genes

decreased in SBRs dosed with nanoAg from 10 to 20 mg/L

(Fig. S3a, p < 0.05), but were not significantly impacted by Agþ

or the other three nanomaterials/bulk/ionic analogs (Fig. S3,

p > 0.05). Nitrospira 16S rRNA gene levels were not impacted by

any of the four nanomaterials or bulk/ionic materials (Fig. S4,

p > 0.05).

In the high load experiments, relative abundance of amoA

genes decreased significantly in SBRs dosed with Agþ relative

to SBRs dosed with nanoAg or undosed controls (Fig. 2,

p < 0.05). However, Nitrobacter and Nitrospira 16S rRNA genes

were not significantly impacted by Agþ (Fig. S5a-b, p> 0.05). No

effects of nanoAg or nano/bulkCeO2 on any of the three

nitrifier marker genes were observed (Fig. 2, Fig. S5, p > 0.05).

3.3. Shifts in microbial community composition

Upon commencing dosing, microbial communities were

clustered closely, with similarities greater than 70% based on

MDS analysis at the genus level, confirming that initial con-

ditions were relatively consistent (Fig. 3). Proteobacteria was

the most dominant phylum in the initial microbial commu-

nities, with a relative abundance of 44.3e63.0% (Fig. 4, Fig. S7).

Other abundant phyla in the initial communities included

Bacteroidetes (14.0e23.8%), Actinobacteria (6.1e24.4%), Acid-

obacteria (0.4e3.4%), Nitrospira (0.1e7.8%), Firmicutes (0.3%e

2.3%), and Chloroflexi (0.3%e1.8%).

By the end of dosing, microbial communities measurably

shifted in all SBRs, with similarities less than 70% relative to

the initial conditions (Fig. 3). In both sequential and high load

experiments, dosing with nanoAg and Agþ resulted in distinct

community structures relative to the other SBRs, with a more

dramatic effect of Agþ (Fig. 3). Based on the inverse Simpson

diversity index, microbial diversities decreased by 52%e77%

in SBRs dosed with Agþ and decreased by 22%e40% in SBRs

dosed with nanoAg (except one of the nanoAg SBRs that

maintained initial diversity) (Fig. S6a-b). Bacteroidetes became

more dominant in SBRs dosed with Agþ, with increases of

19.4% and 35.0% under sequential load, and 43.4% and 42.1%

under high load (Fig. 4). All other phyla that were dominant at

the initial conditions generally decreased, although there

were some variations between duplicate SBRs (Fig. 4). In

particular, Proteobacteria decreased by 9.8% and 22.8% under

high load of Agþ (Fig. 4). At the class level, an increase of

Fig. 2 e Quantities of amoA genes normalized to 16S rRNA genes in conditions that indicated an effect of the dosed

materials: (a) sequential load of nanoAg and Agþ; (b) high load of 20 mg/L nanoAg and Agþ; (c) sequential load of NZVI and

Fe2þ; and (d) sequential load of nanoCeO2 and bulkCeO2. Error bars represent standard deviation of duplicate DNA extracts

with triplicate q-PCR runs.

Fig. 3 e Multidimensional Scaling (MDS) analysis of microbial community phylogenetic distance at genus level derived from

pyrosequencing of bacterial 16S rRNA genes at the beginning (Begin) and at the end (End) of dosing: (a) sequential load and

high load of nanoAg and Agþ; (b) sequential load and high load of nanoCeO2 and bulkCeO2; (c) sequential load of NZVI and

Fe2þ; (d) sequential load of nanoTiO2 and bulkTiO2.

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7 91

Fig. 4 e Heat-maps of relative abundances of phyla derived from pyrosequencing of bacterial 16S rRNA genes at the

beginning (Begin) and at the end (End) of: (a) sequential load and (b) high load of nanoAg and Agþ. Unit of numbers:

percentage (%).

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 792

Flavobacteria and unclassified classes mainly accounted for

the increase of Bacteroidetes (Fig. S8a-b).

In contrast to the Agþ conditions, the Proteobacteria phylum

became more dominant in SBRs dosed with nanoAg, with in-

creases of 16.8% in one SBR under sequential load (the dupli-

cate SBR had no observable effect), and 16.6% and 21.4% under

high load (Fig. 4). All other phyla generally decreased with

some variations between duplicate SBRs (Fig. 4). At the class

level, while d-Proteobacteria mainly accounted for the increase

of Proteobacteria under sequential load of nanoAg, a- and b-

Proteobacteria mainly accounted for the increase under high

load (Fig. S8a-b).

As an indicator of resistance or sensitivity, the five genera

that exhibited the greatest increases or decreases in relative

abundances in SBRs dosed with nanoAg or Agþ are compared

in Fig. 5. Notably, the identified genera were generally

distinct for nanoAg versus Agþ and for sequential and high

load. Acidovorax, Ohtaekwangia, Chryseobacterium and Fla-

vobacteriumwere markedly resistant to Agþ under sequential

and high load, but they were not resistant to nanoAg.

Ohtaekwangiawas one of themost sensitive genera under the

high load nanoAg condition (decreases of 7.2% and 12.0% in

duplicate SBRs). Nannocystis was the most resistant to

nanoAg under sequential load conditions (increases of 24.5%

and 6.5% in duplicate SBRs), but was not measurably affected

by Agþ. Nannocystis was not detected in any of high load ex-

periments. Zoogloea showed the greatest sensitivity to

sequential load of nanoAg and Agþ, with decreases of 3.1%e

6.2%. Zoogloea was also sensitive to Agþ in the high load

condition (decreases of 1.7% and 1.0%), but was the most

resistant genera to high load nanoAg (increases of 12.6% and

13.4%). Nitrospira was one of the most sensitive genera to

high loads of nanoAg and Agþ, but was not measurably

impacted by sequential loading.

Fig. 5 e Heat-maps of relative abundance of (a) five most resistant and (b) five most sensitive classified genera derived from

pyrosequencing of bacterial 16S rRNA genes at sequential and high load of nanoAg and Agþ. Unit of numbers: percentage

(%).

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7 93

Dissimilarities in the microbial community structure

following sequential load of nano and bulk CeO2 were also

apparent in the MDS analysis (Fig. 3), but the microbial di-

versities did not shift significantly (Fig. S6c). a-Proteobacteria

increased markedly in dosed SBRs relative to undosed SBRs

(Fig. S8c). Genera that showed the greatest increase and

decrease in dosed SBRs were also identified, but the changes

were all within 5.0%, with no obvious trends among the

different conditions (Fig. S9). For SBRs dosed with nanoCeO2

and bulkCeO2 under high load and SBRs dosed with NZVI,

Fe2þ, nanoTiO2 and bulkTiO2 under sequential load, although

therewas shift inmicrobial community structure at the end of

dosing relative to the beginning, there was no significant dif-

ference among the final microbial community structures

(similarities greater than 70%, Fig. 4) or microbial diversities

(Fig. S6d-f) relative to undosed controls. Changes of phyla

compositions in these experiments were also consistent be-

tween SBRs receiving nanomaterials, analogs, and undosed

controls (Fig. S7b-d).

Finally, because SBRs were re-started with each experi-

ment to avoid bias of prior dosed materials, it was of interest

to compare the similarities of SBR microbial communities

across experiments. Notably, variation in the initial microbial

community structures across experiments were greater than

the variation associated with dosing of nanomaterials or bulk/

ionic analogs, except SBRs dosed with Agþ (Fig. 6).

3.4. Partitioning and transformation of nanomaterialsin SBRs

Results of ICP-MS indicated that the concentrations of Ag, Fe,

TiO2 and CeO2 in the aqueous effluents of SBRs receiving

nanomaterials or analogs were all less than 1% relative to the

influent (data not shown), except that Ag concentration

increased to 0.74 and 1.49 mg/L in the aqueous effluent of

duplicate SBRs receiving high loads of Agþ (Fig. S1). In acti-

vated sludge, a large portion of nanoAg remained dispersed

(Fig. 7). EDSmapping revealed spatial correlation of silver with

sulfur, but not with chlorine, phosphorus, oxygen or carbon,

which could also form complexes with silver (Fig. S10). In

contrast, the other three nanomaterials: NZVI, nanoTiO2, and

nanoCeO2, were mostly aggregated (Fig. 7). EDS mapping did

not indicate correlation of iron with oxygen or other elements

(Fig. S11). Titanium and cerium were observed to be spatially

correlated only with oxygen and not with other elements

(Fig. S12-13).

Fig. 6 e Multidimensional Scaling (MDS) analysis of

bacterial community similarities derived from

pyrosequencing in the beginning (Begin) and at the end

(End) of dosing in all sequential load and high load

experiments. Green: sequential load of nanoAg and Agþ;Red: sequential load of nanoTiO2 and bulkTiO2; Blue:

sequential load of NZVI and Fe2þ; Yellow: sequential load

of nanoCeO2 and bulk CeO2; Gray: high load of nanoAg and

Agþ; Pink: high load of nanoCeO2 and bulk CeO2. (For

interpretation of the references to colour in this figure

legend, the reader is referred to the web version of this

article.)

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 794

4. Discussion

4.1. Fate of nanoAg in the SBRs

Although>99%ofnanoAgwas retained in theactivated sludge,

its impact on nitrification appeared to be negligible under the

conditions of this study. The spatial correlation of silver with

sulfur in the EDS maps indicated that nanoAg was largely

transformed to AgeS species, which is consistent with other

studies based on pilot-scale wastewater treatment processes

(Kaegi et al., 2011;Ma et al., 2014). Sulfidation of nanoAg results

in highly insoluble products and has been observed to reduce

toxicity of nanoAg to enriched nitrifying bacteria (Choi et al.,

2009) and Escherichia coli (Reinsch et al., 2012). In this study,

sulfide was not added to the synthetic wastewater feed, but it

did contain sulfate as a nutrient and pyrosequencing

confirmed the presence of sulfate-reducing bacteria (Desulfo-

bulbus and Desulfovibrio). Thus, sulfidation of nanoAg could

occur under the conditions of this study and thus acted to

stabilize the nanoAg. TEM-EDS analysis revealed that a large

portion of silver-based nanoparticles remained dispersed in

activated sludge. This outcome was possibly the result of the

high colloidal stability of citrate coated nanoAg (Arnaout and

Gunsch, 2012), or to the sulfidation of the nanoAg. Dispersion

of nanoAg has been found to enhance sulfidation, relative to

highly aggregated nanoparticles (Reinsch et al., 2012).

4.2. Effects of nanoAg and Agþ on nitrification functionand microbial community structure

A recent study reported that both nanoAg and Agþ dosed to a

pilot-scale activated sludge plant were converted to Ag2S (Ma

et al., 2014). Despite the similar partitioning of nanoAg and

Agþ in the sludge versus aqueous phase, Agþ exhibited more

toxicity than nanoAg in the present study. Agþ disrupted

nitrification function within 7 days at high load, with no cor-

responding effect for nanoAg. While both nanoAg and Agþ

resulted in microbial community shifts, their precise effects

were distinct. The phylum of Bacteroidetes became increas-

ingly dominant in SBRs dosed with Agþ, while Proteobacteria

became more dominant in SBRs dosed with nanoAg. Corre-

spondingly, different genera were identified to be the most

resistant to nanoAg or Agþ. The genera associated with

resistance to Agþ, but not to nanoAg, such as Flavobacterium

and Chryseobacterium, have been observed by others to be

dominant in metal polluted wastewaters (Bestawy et al., 2013;

Dilek and Gokcay, 1996). This suggests that bioavailability of

any potential Agþ released by nanoAg is minimal compared to

direct dosing of Agþ. Increase of Bacteroidetes and decrease of

Proteobacteria in high load Agþ SBRs was associated with fail-

ure of ammonia and COD removal, as well as increased

effluent turbidity. Excessive growth of some filamentous

bacteria in the phylum of Bacteroidetes (Manz et al., 1996) could

cause sludge bulking and disrupt sludge settling, while Pro-

teobacteria includes a variety of functionally important bacte-

ria for organic and nitrogen removal (Schmidt et al., 2003;

Nogueira et al., 2002). For the genera classified as resistant to

nanoAg, Nannocystis are known to produce flocculating sub-

stances with potential wastewater treatment applications

(Zhang et al., 2002). Metal resistance of Nannocystis has not

previously been reported, suggesting other potential selection

mechanisms of nanoAg besides metal toxicity. Overall, the

results indicate that nanoAg and Agþ interact distinctly with

microbes in a complex activated sludge environment. Liu et al.

(2011) suggested direct sulfidation of nanoAg at high sulfide

concentrations without dissolution to Agþ as intermediate.

This mechanism may contribute to the greater tolerance of

the SBRs to nanoAg than Agþ, and their different impacts on

specific bacterial groups. However, Agþ did not inhibit nitrifi-

cation at 20 mg/L in the sequential loading experiment. It is

possible that as the concentration was sequentially increased

from 0.1 mg/L to 20 mg/L, microbes acclimated to the Agþ

toxicity.

MDS analysis revealed that the variation of microbial

community structure associated with dosing of nanoAg was

not greater than the variation among the initial communities

when the SBRs were re-started. This observation suggests

functional redundancy, which is also likely at least a partial

explanation for why microbial shifts induced by the dosed

materials were not generally associated with adverse effects

on SBR function. However, both nanoAg and Agþ dosing were

generally associated with a decrease in microbial diversity,

especially in SBRs receiving Agþ. Shifts in diversity can serve

as an indicator of the overall health of a microbial community

(Philippot et al., 2013). Moreover, the relative abundance of

nitrification gene markers amoA and Nitrobacter 16S rRNA

genes decreased significantly at 10 and 20 mg/L nanoAg in the

sequential load experiment, despite the absence of detectable

nitrification inhibition. Together the results suggest the value

of examining the effects of nanoAg over longer exposure pe-

riods. With continuous loading over time, nanoAg could build

up in the sludge and further decrease the diversity and

Fig. 7 e EDS maps of (a) silver, (b) iron, (c) titanium, and (d) cerium in activated sludge after dosing with nanoAg, NZVI,

nanoTiO2 and nanoCeO2. Color intensity is indicative of the number of X-ray counts. Corresponding STEM images appear in

the gray insets. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version

of this article.)

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7 95

nitrifying bacteria and ultimately impact functionality of the

wastewater treatment process.

The findings of this study are inconsistent with a very

recent report by Yang et al. (2014), which indicated that

nanoAg had a stronger impact than Agþ on the microbial

community structure of activated sludge. Inconsistencies be-

tween the two studies could be associated with different

behavior of nanoAg under the batch microcosm conditions

applied in the Yang study, compared to the more realistic

wastewater treatment conditions of the present study. While

Yang et al. (2014) suggested that their effectsmay be the result

of a higher propensity of Agþ to be scavenged by inorganic

ligands and organic matter than AgNPs, nanoAg appeared to

be more effectively stabilized by sulfidation in the present

study. Also, the two studies identified different genera that

showed the most resistance or sensitivity to nanoAg or Agþ.This observation may be a result of the distinct initial micro-

bial communities used in the two studies. Specific microbial

community compositions often vary between wastewater

treatment plants (Wang et al., 2012). Even in this study, MDS

analysis revealed distinct initial microbial community struc-

tures each time the SBRs were re-started. Differences in the

initial microbial community structure could also have played

a role in the observation that Nannocystis increased in the

sequential nanoAg load condition, but was not detected in any

of the high load samples. Overall, this study suggests that

initial microbial community structure will play in a role in the

ultimate genera that display resistance versus sensitivity to

the dosed materials.

4.3. Fate and effects of NZVI, nanoTiO2, and nanoCeO2

in the SBRs

It is important to note that negligible inhibition of nitrification

was observed in SBRs dosedwithNZVI, nanoTiO2, nanoCeO2 or

their ionic/bulk analogs. Their impacts on the relative abun-

dances of nitrification gene markers were also limited.

Although nano and bulk CeO2 were associated with distinct

microbial community shifts in response to sequential loading

relative to the undosed controls, themicrobial diversity did not

decrease, and the shifts were not greater than the variation

acrossdifferent experiments. To theauthors'knowledge, this is

the first study to report fate and impacts of NZVI during bio-

logical wastewater treatment. Under aerobic conditions, NZVI

could be rapidly oxidized and inactivate microbes by physical

attachment, disruption of the membrane, or generation of

reactive oxygen species (Diao and Yao, 2009). However, in this

study, NZVI appeared to be quickly aggregated upon dosing to

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 796

the SBRs, which could reduce the reactivity and overall toxicity

of NZVI. Toxicity of nanoTiO2 and nanoCeO2 has mainly been

attributed to generation of reactive oxygen species and pene-

tration of cells, especially in small size ranges (Jin et al., 2011;

Zhang et al., 2011). Aggregation could limit the size effect of

TiO2 and CeO2 particles on microbes, and thus reduce the

overall toxic effects.

5. Conclusions

This study observed negligible inhibitory effects of nanoAg,

NZVI, nanoTiO2 and nanoCeO2 in lab-scale SBRs as concen-

trations increased from 0.1 to 20 mg/L, and suggested a high

threshold of the four nanomaterials for adverse effect on

nitrification function. As NZVI, nanoTiO2 and nanoCeO2 had

limited impacts onmicrobial community, cautious optimism is

suggested for disposing of these nanomaterials into waste-

water treatment plants. In contrast, NanoAg, and to a greater

extent Agþ, had more evident effects on nitrification gene

markers, microbial diversity, and microbial community com-

positions. The two forms of silver appeared to have nearly

opposite effects on shifts in specific phyla and genera in

response to the dosed materials, indicating distinct in-

teractions of nanoAg versus Agþ with microbial cells. Further

studies are needed to understand the mechanisms by which

nanoAg andAgþ specifically affectmicrobial communities, and

also to examine the effects of nanoAg inwastewater treatment

over the longer term, to verify if the continuous impacts on

microbial communities will reach a threshold of disrupting

wastewater treatment functions.

Acknowledgments

Funding for this research was provided by the U.S. Environ-

mental Protection Agency Star Grant #834856, the National

Science Foundation (NSF) Center for the Environmental Im-

plications of Nanotechnology (CEINT) (EF-0830093), the

Microbiology in the Post-Genome Era NSF REU site at Virginia

Tech (NSF Award #1156954), and the Virginia Tech Institute for

Critical Technology and Applied Science (ICTAS). The findings

do not reflect the views of the sponsors.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2014.09.008.

r e f e r e n c e s

Alito, C.L., Gunsch, C.K., 2014. Assessing the effects of silvernanoparticles on biological nutrient removal in bench-scaleactivated sludge sequencing batch reactors. Environ. Sci.Technol. 48 (2), 970e976.

APHA, 1998. Standard Methods for the Examination of Water andWastewater, twentieth ed. American Public HealthAssociation, Washington, DC.

Arnaout, C.L., Gunsch, C.K., 2012. Impacts of silver nanoparticlecoating on the nitrification potential of Nitrosomonas europaea.Environ. Sci. Technol. 46 (10), 5387e5395.

Bestawy, E.E.1, Helmy, S., Hussien, H., Fahmy, M., Amer, R., 2013.Bioremediation of heavy metal-contaminated effluent usingoptimized activated sludge bacteria. Appl. Water Sci. 3,181e192.

Brar, S.K., Verma, M., Tyagi, R.D., Surampalli, R.Y., 2010.Engineered nanoparticles in wastewater and wastewatersludge e evidence and impacts. Waste Manag. 30 (3), 504e520.

Bray, J.R., Curtis, J., 1957. An ordination of the upland forestcommunities of Southern Wisconsin. Ecol. Monogr. 27 (4),325e349.

Choi, O., Deng, K.K., Kim, N.J., Ross, L., Surampalli, R.Y., Hu, Z.Q.,2008. The inhibitory effects of silver nanoparticles, silver ions,and silver chloride colloids on microbial growth. Water Res. 42(12), 3066e3074.

Choi, O., Cleuenger, T.E., Deng, B.L., Surampalli, R.Y., Ross, L.,Hu, Z.Q., 2009. Role of sulfide and ligand strength incontrolling nanosilver toxicity. Water Res. 43 (7), 1879e1886.

Diao, M.H., Yao, M.S., 2009. Use of zero-valent iron nanoparticlesin inactivating microbes. Water Res. 43 (20), 5243e5251.

Dilek, F.B., Gokcay, C.F., 1996. Microbiology of activated sludgetreating wastewater containing Ni(II) and Cr(VI). Water Sci.Technol. 34 (5e6), 183e191.

Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009.Modeled environmental concentrations of engineerednanomaterials (TiO2, ZnO, Ag, CNT, Fullerenes) for differentregions. Environ. Sci. Technol. 43 (24), 9216e9222.

Jin, C., Tang, Y., Yang, F.G., Li, X.L., Xu, S., Fan, X.Y., Huang, Y.Y.,Yang, Y.J., 2011. Cellular toxicity of TiO2 nanoparticles inanatase and rutile crystal phase. Biol. Trace Elem. Res. 141(1e3), 3e15.

Kaegi, R., Voegelin, A., Sinnet, B., Zuleeg, S., Hagendorfer, H.,Burkhardt, M., Siegrist, H., 2011. Behavior of metallic silvernanoparticles in a pilot wastewater treatment plant. Environ.Sci. Technol. 45 (9), 3902e3908.

Kim, B., Park, C.S., Murayama, M., Hochella, M.F., 2010. Discoveryand characterization of silver sulfide nanoparticles in finalsewage sludge products. Environ. Sci. Technol. 44 (19),7509e7514.

Kiser, M.A., Westerhoff, P., Benn, T., Wang, Y., Perez-Rivera, J.,Hristovski, K., 2009. Titanium nanomaterial removal andrelease from wastewater treatment plants. Environ. Sci.Technol. 43 (17), 6757e6763.

Lee, P.C., Meisel, D., 1982. Adsorption and surface-enhancedraman of dyes on silver and gold sols. J. Phys. Chem. 86 (17),3391e3395.

Li, M., Zhu, L.Z., Lin, D.H., 2011. Toxicity of ZnO nanoparticles toEscherichia coil: mechanism and the influence of mediumcomponents. Environ. Sci. Technol. 45 (5), 1977e1983.

Li, L.X.Y., Hartmann, G., Doblinger, M., Schuster, M., 2013.Quantification of nanoscale silver particles removal andrelease from municipal wastewater treatment plants inGermany. Environ. Sci. Technol. 47 (13), 7317e7323.

Liang, Z.H., Das, A., Hu, Z.Q., 2010. Bacterial response to a shockload of nanosilver in an activated sludge treatment system.Water Res. 44 (18), 5432e5438.

Liu, J.Y., Pennell, K.G., Hurt, R.H., 2011. Kinetics and mechanismsof nanosilver oxysulfidation. Environ. Sci. Technol. 45 (17),7345e7353.

Ma, R., Levard, C., Judy, J.D., Unrine, J.M., Durenkamp, M.,Martin, B., Jefferson, B., Lowry, G.V., 2014. Fate of zinc oxide

wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 8 7e9 7 97

and silver nanoparticles in a pilot wastewater treatment plantand in processed biosolids. Environ. Sci. Technol. 48 (1),104e112.

Manz, W., Amann, R., Ludwig, W., Vancanneyt, M., Schleifer, K.H.,1996. Application of a suite of 16S rRNA-specificoligonucleotide probes designed to investigate bacteria of thephylum cytophaga-flavobacter-bacteroides in the naturalenvironment. Microbiology-Uk 142, 1097e1106.

Mueller, N.C., Nowack, B., 2008. Exposure modeling of engineerednanoparticles in the environment. Environ. Sci. Technol. 42(12), 4447e4453.

Nel, A., Xia, T., Madler, L., Li, N., 2006. Toxic potential of materialsat the nanolevel. Science 311 (5761), 622e627.

Nogueira, R., Melo, L.F., Purkhold, U., Wuertz, S., Wagner, M.,2002. Nitrifying and heterotrophic population dynamics inbiofilm reactors: effects of hydraulic retention time and thepresence of organic carbon. Water Res. 36 (2), 469e481.

Peet, R.K., 1974. The measurement of species diversity. Annu.Rev. Ecol. Syst. 5, 285e307.

Pelletier, D.A., Suresh, A.K., Holton, G.A., McKeown, C.K.,Wang, W., Gu, B.H., Mortensen, N.P., Allison, D.P., Joy, D.C.,Allison, M.R., Brown, S.D., Phelps, T.J., Doktycz, M.J., 2010.Effects of engineered cerium oxide nanoparticles on bacterialgrowth and viability. Appl. Environ. Microbiol. 76 (24),7981e7989.

Philippot, L., Spor, A., Henault, C., Bru, D., Bizouard, F.,Jones, C.M., Sarr, A., Maron, P.A., 2013. Loss in microbialdiversity affects nitrogen cycling in soil. Isme J. 7 (8),1609e1619.

Radniecki, T.S., Stankus, D.P., Neigh, A., Nason, J.A., Semprini, L.,2011. Influence of liberated silver from silver nanoparticles onnitrification inhibition of Nitrosomonas europaea. Chemosphere85 (1), 43e49.

Reinsch, B.C., Levard, C., Li, Z., Ma, R., Wise, A., Gregory, K.B.,Brown, G.E., Lowry, G.V., 2012. Sulfidation of silver

nanoparticles decreases Escherichia coli growth inhibition.Environ. Sci. Technol. 46 (13), 6992e7000.

Schloss, P.D., Gevers, D., Westcott, S.L., 2011. Reducing the effectsof PCR amplification and sequencing artifacts on 16S rRNA-based studies. Plos One 6 (12).

Schmidt, I., Sliekers, O., Schmid, M., Bock, E., Fuerst, J.,Kuenen, J.G., Jetten, M.S.M., Strous, M., 2003. New concepts ofmicrobial treatment processes for the nitrogen removal inwastewater. Fems Microbiol. Rev. 27 (4), 481e492.

Wang, X.H., Hu, M., Xia, Y., Wen, X.H., Ding, K., 2012.Pyrosequencing analysis of bacterial diversity in 14wastewater treatment systems in China. Appl. Environ.Microbiol. 78 (19), 7042e7047.

Xiu, Z.M., Ma, J., Alvarez, P.J.J., 2011. Differential effect of commonligands and molecular oxygen on Antimicrobial activity ofsilver nanoparticles versus silver ions. Environ. Sci. Technol.45 (20), 9003e9008.

Yang, Y., Quensen, J., Mathieu, J., Wang, Q., Wang, J., Li, M.Y.,Tiedje, J.M., Alvarez, P.J.J., 2014. Pyrosequencing reveals higherimpact of silver nanoparticles than Agþ on the microbialcommunity structure of activated sludge. Water Res. 48,317e325.

Zhang, J., Liu, Z., Wang, S., Jiang, P., 2002. Characterization of abioflocculant produced by the marine myxobacteriumNannocystis sp NU-2. Appl. Microbiol. Biotechnol. 59 (4e5),517e522.

Zhang, H.F., He, X.A., Zhang, Z.Y., Zhang, P., Li, Y.Y., Ma, Y.H.,Kuang, Y.S., Zhao, Y.L., Chai, Z.F., 2011. Nano-CeO2 exhibitsadverse effects at environmental relevant concentrations.Environ. Sci. Technol. 45 (8), 3725e3730.

Zhang, C.Q., Liang, Z.H., Hu, Z.Q., 2014. Bacterial response to acontinuous long-term exposure of silver nanoparticles at sub-ppm silver concentrations in a membrane bioreactor activatedsludge system. Water Res. 50, 350e358.