Introduction

In recent years, cerium oxide nanoparticles (CeO2 NPs) have been intensively studied owing to their wide applications and unique properties in UV absorbents and filters, gas sensors, catalytic wet oxidation, catalysts in the fuel cell technology, engine exhaust catalysts, photocatalytic oxidation of water, NO removal, electrolyzers, solid electrolytes and so on (Goharshadi et al. 2011). It has been reported that the ability of CeO2 NPs to express so many unique properties was mostly due to their ability to reversibly oxygenate and deoxygenate without disrupting the fluorite lattice-structure (Bumajdad et al. 2009). As rare metal, CeO2 NP chemical configuration with 4f orbitals buried inside the atom and shielded from the atom’s environment by the 4d and 5p electrons has also been seen as the main core of the expression of unique properties which is impossible with transition and main group metals (Hu et al. 2006; Bouzigues et al. 2011). According to Xu and Qu (2014), cerium atom can easily and drastically adjust its electronic configuration to best fit its immediate environment. Cerium oxide nanoparticles have also shown biological applications as they have been used as enzymes (e.g. superoxide oxidase, catalase, oxidase, etc.) and also as antioxidant or radical scavengers (Asati et al. 2009; Mandoli et al. 2010; Karakoti et al. 2010; Xu et al. 2013; Li et al. 2013). This unique nanoparticle has also been used to treat diseases such as oxidative stress-related diseases (e.g. neurodegenerative disorders), diabetes, retinal diseases, chronic inflammation and cancer (Maritim et al. 2003; Chen et al. 2006; Mariani et al. 2005; Lin et al. 2006; Federico et al. 2007; Hussain et al. 2012). Being a mature engineered nanoparticle with such unique properties and industrial applications, CeO2 NPs health and environmental concerns have mostly been overlooked. Recently, CeO2 NPs have been incriminated of being toxic to rats (Srinivas et al. 2011), freshwater alga (Taylor et al. 2015), human lung cells (Mittal and Pandey 2014), zebrafish (Arnold et al. 2013), and bacteria (Pelletier et al. 2010). Considering the above and through their lifecycles, CeO2 NPs represent a major concern as they are likely to enter natural water bodies channeled via wastewater treatment plants (WWTPs) (Aruoja et al. 2009). It is well-known that biological WWTPs consist of a series of biochemical processes, such as nitrification, denitrification, and phosphorus anaerobic release and aerobic or anoxic uptake engineered for the removal of nitrogen and phosphorus as well as other pollutants. The performance of these processes is directly related to the activities of some key microorganisms present in the activated sludge. However, it is unknown how nCeO2 NPs affect essential activities of these key microorganisms in activated sludge. Since activated sludge is the most widely used wastewater treatment option, the present study aimed at assessing the effect of nCeO2 NPs to activated sludge microorganisms.

Materials and methods

Bioreactors

Fresh activated sludge (1 L each) was collected from the Northern Wastewater Works, Johannesburg, chipped to the laboratory in a cooler box (4C) and used within 24 h. The collected activated sludge (100 mL) was then inoculated in a reactor containing 300 mL of culture media [d-glucose anhydrate (2.5 g/L), MgSO4·7H2O (0.5 g/L) and KNO3 (0.18 g/L) in distilled water] and treated with different concentration of CeO2 NPs (10, 20, 30 and 40 mg/L). In order to assess the impact of cerium oxide nanoparticles on the microbial community of wastewater treatment plants, the non-treated mixed liquor which contained the mixed liquor medium without nCeO2 NP was used as control. Experiments were run at 28 ± 2 °C on a checking incubator at 120 rpm for 5 days under aerobic condition. Aliquots were then taken at the final incubation day and analysis for microbial community. The aliquot samples were also used to determine the chemical oxygen demand (COD), nitrate and phosphate, pH, dissolved oxygen (DO) and electrical conductivity (EC). To test for NO −13 , the sodium salicylate method was used as reported by Monteiro et al. (2003). Briefly, 50 mL of samples was pipetted into milliliter beaker, and mixed with 1 mL of the salicylate solution. The mixture was dried out in an oven at 105 °C to allow the formation of NO +12 from NO −13 . Then, 1 mL of sulfuric acid (17.4 M) was added and allowed to cool for 10 min and 7 mL of the solution containing sodium hydroxide (5 M) and sodium potassium tartrate (149 g/L) were later added. The solution was later made up with water and analysed in a spectrophotometer (Monteiro et al. 2003). For PO −34 , the method 424f standard method as reported by APHA (2001) was used. The method 424f uses ammonium molybdate and potassium antimonyl tartrate in order to react in an acidic medium with orthophosphate to form a heteropoly acid (phosphomolybdic acid) that is reduced to intensely coloured molybdenum blue by ascorbic acid. The closed reflux method was also used to measure COD concentration (APHA 2001), whereas pH, DO, electrical conductivity (EC) and temperature were measured using specific probes (HACH, Germany). All experiment was done in triplicates.

DNA extraction, amplification and sequencing of bacterial 16S rRNA genes

In order to extract the genetic material (DNA) representing the microbial communities of each bioreactor, an aliquot (100 mL) of nCeO2-free and treated mixed liquor from day 5 samples was centrifuged at 10,000×g for 5 min at 4 °C and the collected cells cleaned twice using sterile phosphate buffer solution (1×). The collected cell pellets were re-suspended in 1× TE buffer (pH 8.0), homogenously mixed and DNA was extracted using the ZR Fungal/Bacterial DNA Kit™ (Zymo Research, Pretoria, South Africa) according to the procedures provided by the manufacturer. The integrity and purity of extracted DNA was further assessed on the 1.0% agarose gel and measured using a Nanodrop spectrophotometer (Nanodrop 2000, Thermo Scientific, Japan).

Amplification and sequencing of bacterial 16S rRNA genes

Prior of sequencing, the extracted DNA was amplified in triplicate and the V3 and V4 regions of the 16S rRNA gene were targeted by using the universal primers pairs (341F and 785R) and pooled together in order to better sample rare organisms, and avoid PCR biases (Klindworth et al. 2013; Sekar et al. 2014). Each 50 μL PCR reaction system contained 25 µL of 2X Dream Taq green Master Mix (DNA polymerase, dNTPs and 4 mM MgCl2), 22 µL of sterile Nuclease-free water, 1 µL of forward primer (0.2 µM) and 1 µL of reverse primer (0.2 µM), and 1 µL of DNA (50–100 ng/µL). In order to control nuclease contamination, negative control was included at every reaction. The following PCR reaction was performed: an initial denaturation step at 94 °C for 5 min, followed by 30 cycles of denaturation at 94 °C for 1 min, annealing at 55 °C for 30 s and extension at 72 °C for 1 min 30 s, and a final extension at 72 °C for 10 min, followed by cooling to 4 °C. The PCR products were loaded in 1% (m/v) agarose gel (Merck, SA) stained with 5% of 10 mg/mL ethidium bromide (Merck, SA) and visualized under ultra violet Trans illuminatior (InGenius Bio Imaging System, Syngene, Cambridge, UK). The correct PCR amplicons of bacteria were pooled together for the respective samples at approximately equimolar concentrations and submitted to Inqaba Biotechnology Industries, South Africa for sequencing on an Illumina MiSeq.

Enzyme essay

In order to investigate the impact of nCeO2 on functional microbial population in the bioreactor, enzymes catalysing the degradation of polyphosphate such as adenylate kinase (ADK) and polyphosphate kinase (PPK) as well as those involved in the denitrification process namely nitrate reductase (NaR) and nitrite reductase (NiR)were assessed. Prior to assess enzymatic activities, activated sludge aliquots was taken and cleaned three times with 1.5 M NaCl buffer 5 M NaCl buffer consisted with 0.01 M EDTA and 1 mM NaF (pH 7.4). Cell structure of activated sludge were later broken down by resuspending pellets and sonicating for 5 min at 20 kHz and 4 °C, and centrifuged for 10 min at 12,000 rpm as reported by Chen et al. (2012). ADK was determined by mixing 0.16 mL of cell extract per mL with 7 mM MgCl2, 90 mM Tris hydrochloride (Tris–HCl, pH 7.0), 200 mM d-glucose, 0.6 mM NADP (Sigma), 3.4 U of hexokinase (HK, Wako Chemical, Osaka, Japan), and 1.7 U of glucose 6-phosphate dehydrogenase (G6P-DH, Wako Chemical, Osaka, Japan) per mL. Adenosyne diphosphate (1 mM ADP) was later added to the mixture in order to start the enzymatic reaction and the production of NADPH2 was measured at 340 nm by microplate reader (BioTek, USA). For PPK activity, the polyphosphate utilization approach was used and the reaction was carried out at 30 °C after mixing 150 µL crude extracts with 100 mM Tris–HCl buffer (pH 7.4), 8 mM MgCl2, 200 mM d-glucose, 0.5 mM NADP, 150 µg of sigma Type 45 poly-P, 1 unit of HK and 1 U of G6P-DH (Chen et al. 2012). The enzymatic activity of PPK and ADK were defined as the production of _µmol NADPH/(min mg protein). As for denitrification process enzymes such as NaR and NiR enzyme, their enzymatic activities were assayed according to Kenji et al. (1981). One unit of enzyme activity for NaR and NiR was defined as the production of 1 μmol/(min mg protein).

Assessment of nCeO2-NPs impact on microbial morphology

To further determine the impact of nCeO2-NPs on the microbial population, a scanning electron microscopy (SEM) was used. After 5 days of incubation, nCeO2-NPs treated and not treated samples were centrifuged (10 mL) at 7000×g at 4 °C for 10 min. Microbial pellets were later washed five times using 0.1 M phosphate buffered saline (1× PBS) and fixed for 24 h in 2% glutaraldehyde (prepared in 1× PBS). Pellets were further dehydrated through a series of ethanol starting from 60% to absolute, and for each series samples were held for 30 min. Samples were placed on a brass stub, sputter-coated with gold and examined by SEM.

Statistical data processing

Prior to be used, artificial replicate reads and low quality reads were removed from the dataset using Mothur pipeline (Schloss et al. 2009). Good quality reads were further pre-screened for ribosomal identity (at approximately 70% identity) using Qiime-uclust and chimeras removed through UCHIME according to de novo method (Edgar et al. 2011). All rRNA non-chimeric reads were later been analyzed at a confidence threshold of 97% for taxonomic classification using RDP pyrosequencing pipeline. In addition, reads with similarity more than 97% were clustered within the same operational taxonomic unit (OTU) and rarefaction curves were also determined (Wang et al. 2007; Cole et al. 2014). The diversity index Shanon and richness estimator Chao1 were also performed to estimate the microbial diversity and richness from each water samples. The relative abundance (%) of individual taxa within each community was calculated by comparing the number of sequences assigned to a specific taxon against the number of total sequences obtained for that sample. The similarity and dissimilarity in bacterial community structure within both wastewater treatment plants were analyzed using Jaccard index (Cole et al. 2014). Generated data was later made publicly available at the DDBJ Sequence Read Archive (DRA) under the accession number PSUB005615.

Results

Community species richness and diversity indices

The present study generated approximately 28,201 reads from the control samples but when stressed with an increase nCeO2 concentration, samples showed an approximately 28.6% decrease (20,135 reads) to a 57.1% decrease (12,082 reads) in the samples treated with 10 mg/L-CeO2 and 40 mg/L-CeO2, respectively. Similar observation was noted with the operational taxonomic units (OTUs) as a total of 27,967 OTUs was generated from the control samples while the sample with highest nCeO2 NP revealed a total of 6433 OTUs. The impact of nCeO2 NPs on the microbial complexity and abundance in the samples was also revealed by using the Shannon–Weaver index and Chao1 richness estimator at 3% cutoff (Table 1). The diversity index (Shannon) revealed a fluctuation in diversity as Shannon values for each samples were not inversely proportional to the increase of nCeO2 NP in the reactors as sample containing 40 mg/L-nCeO2 had high diversity index (8.178) while those with 30 mg/L-nCeO2 NPs was the lowest (7.689). Besides the fact that control samples had the highest diversity index (10.267), no significant difference (p > 0.05) between treated samples in terms of diversity index was observed and this revealed that nCeO2 NPs impacted more on the microbial abundance than on the diversity. The evenness highlighting the complexity of individual microbial population within samples also revealed that no statistical difference between samples in terms of microbial complexity as the values ranged from 0.885 to 0.999. A species richness test conducted using Chao1 richness estimator showed a drastic decrease of species richness of approximately 97.23–98.48% when comparing the control samples to nCeO2 NP treated samples.

Table 1 Diversity indices of samples treated with nCeO2 NPs during 5 days of incubations
Full size table

An additional confirmatory test on species richness conducted using rarefaction analysis also revealed a difference in the number of reads and OTUs between samples and control highlighting a high dissimilarity in bacterial diversity with control having more OTUs and reads than the treated samples. When comparing treated samples among them, no significant difference was noted (Fig. 1). However, the absence of plateau on the bacterial samples indicated that sequencing depth was still not enough to cover the entire bacterial diversity and a large fraction of the different species remains to be discovered. A pairwise community similarity between samples was assessed based on the absence and presence of each OTU using a Jaccard index (Additional file 1: Table S1). The Jaccard index exhibited a moderate or no similarity between all bacterial samples ranging with values from 0.479 to 0.999 with the S_A (10 mg/L) and S_C (30 mg/L) bacterial community showing the most similarity (0.479) as compared to others.

Fig. 1
figure 1

Rarefaction curves showing the dissimilarity levels among reactors treated and not treated with nCeO2 NPs

Full size image

Change of bacterial community in activated sludge over nCeO2 NP effects

In the present study, Proteobacteria has been noted as the most predominant phylum in our samples with an average number of reads of 18,330 out of 28,201 assigned to it in the control samples. Moreover, Proteobacteria dominated by Gammaproteobacteria (80.57% of the all population), Alphaproteobacteria (5.19%) and Betaproteobacteria (3.19%) was followed by unclassified bacteria (19.6%), Firmicutes (11.567%), Actinobacteria (2.55%) and other additional 11 phyla occupying only 1.5% of the all populations (Figs. 2, 3; Additional file 1: Table S2). The control showed an overall 15 phyla, 36 classes, 54 orders, 107 families and 240 genera. Furthermore, number of reads assigned for Proteobacteria appeared to decrease in the nCeO2 NP-treated samples as the concentration of test NPs increases. However, Proteobacteria was still noted to be the predominant phylum in the presence of 10 mg-nCeO2/L (53%) and 20 mg-nCeO2/L (48%). Unlike in control samples, in the nCeO2 NPs-treated samples, Firmicutes was the second most predominant phylum compared to unclassified bacteria in the control. This situation revealed that in our reactors nCeO2 NPs could promote the growth of some type of microorganisms while slowing the growth of others. Moreover, Firmicutes phylum was dominated by classes of Bacilli (29.49–41.86%) followed by Clostridia or unclassified Firmicutes (Fig. 3).

Fig. 2
figure 2

Taxonomic distribution of the most abundant bacterial phyla in both nCeO2 NPs-treated and not treated (control) samples. Analysis of 16S rRNA gene sequences was done in comparison with the RDP II database

Full size image
Fig. 3
figure 3

Relative abundance (%) of bacterial classes in nCeO2 NPs-treated and not treatment samples

Full size image

Even though the bacterial community appeared to be more diverse as the sequences were classified into lower taxonomic levels, their relative abundances were affected (Additional file 1: Tables S2–S5). Up to the order level, control samples (approximately 21,521 reads) revealed high abundance than the treated samples (19,303, 14,023, 13,840 and 11,501 reads from S_A, S_B, S_C and S_D, respectively). However, the control samples showed more unclassified sequences as compared to the treated samples leading to lower abundance at the family and genus level. When considering each sample individually, and the bacterial proportion in the lower taxonomic level “genera”, control samples were populated by 239 genera with unclassified_Comamonadaceae (26.61%), unclassified_Moraxellaceae (8.93%), unclassified_Pseudomonadaceae (7.08%), Novispirillum (5.88%), Fusibacter (4.88%), unclassified_Enterobacteriaceae (4.48), unclassified_Xanthomonadaceae (3.86%), Shewanella (3.05%), Proteocatella (2.93%), unclassified_Carnobacteriaceae (2.9%), Acinetobacter (2.84%), Proteiniclasticum (2.76%), Trichococcus (2.28%) and the remaining occupying a total of 21.51% (with < 2% each). Contrary in the nCeO2 NP-treated samples, unclassified genera appeared to be affect as their relative abundance drastically reduced to a range of 10 and 0%, while other classified genera showed their abundances increase (Additional file 1: Table S5). At the presence of 10 mg-nCeO2 NP/L, 123 genera were observed with Trichococcus (38.25%) was the most dominant genus followed by Acinetobacter (32.29%), unclassified_Pseudomonadaceae (8.9%), Pseudoxanthomonas (3.09%) and unclassified_Enterobacteriaceae (2.44%). In the samples treated with 20 mg-nCeO2 NPs/L, 115 genera were found with Acinetobacter (35.87%), Trichococcus (28.28%), unclassified_Moraxellaceae (6.9%), unclassified_Comamonadaceae (4.04%), Aerococcus (3.49%), Shewanella (2.72%), Comamonas (2.66%), unclassified_Carnobacteriaceae (2.1%) and the remaining with abundance <2% each. In the sample treated with 30 mg-nCeO2 NPs/L, the fluctuation was also noted as 108 genera was generated with Trichococcus (36.35%), Acinetobacter (33.85%), Comamonas (5.26%), unclassified_Comamonadaceae (5%), Pseudoxanthomonas (4.5%) and unclassified_Moraxellaceae (4.08%) as the most abundant genera in the samples. Similar to other treated samples, a total of 99 genera were observed in the 40 mg-nCeO2 NP/L treated samples with Acinetobacter (29.68%), Trichococcus (28.73%), unclassified_Comamonadaceae (10.1%), Pseudoxanthomonas (5.19%), Comamonas (4.74%), unclassified_Moraxellaceae (4.25%), Aerococcus (2.61%), Cloacibacterium (2.35%), and unclassified_Pseudomonadaceae (2.15%) as the most predominant genera. Despites the observed alternance, the treated samples have the same most two dominant genera as compared to the control samples. The study revealed that nCeO2 NPs affected the diversity of the microbial population while enhancing the growth of particular microbial species.

Change of enzymatic activities in activated sludge stressed with nCeO2 NP

Figure 4 illustrates the impact of nCeO2-NP on the enzymatic activities in the activated sludge. It was observed that enzymatic activities of enzymes catalyzing the denitrification of nitrate (NaR and NiR) were less affected by nCeO2-NP than those catalyzing the degradation of polyphosphate (ADK and PPK). The statistical significant difference (p > 0.05) between activities of NaR-NiR and ADK-PPKwas noted. Despite the significant effect of nCeO2-NPs to ADK-PPK than to NaR–NiR, all enzymatic activities appeared to decrease over the increase of nCeO2-NPs in the media. Activities of ADK was the most affected with a decrease of 91.41–99.54% when compared to the control, while NaR showed the lowest decrease on activities ranging from 11.29 to 92.26%.

Fig. 4
figure 4

Effects of nCeO-NPs on activities of key enzymes associated with denitrification and polyphosphate degradation in EBPR. The unit is µmol NADPH/(min mg protein)

Full size image

Effects of nCeO2-NPs on microbial structure

The representative SEM images of the microbial biomass stressed for 5 days to nCeO2-NPs compared to the control (nCeO2-NPs free-sample) are shown in Fig. 5. The integrity of bacterial cell structure appeared to be disrupted leading to the agglomeration and lyse of microbial cells when compared to the microbial cells from the control sample. Control sample showed a high microbial biomass compared to nCeO2-treated samples with a decrease on microbial biomass as the concentration of nCeO2-NPs increases. SEM images further showed a heterogeneous microbial morphology in both nCeO2-NPs-treated and nCeO2-NPs-free samples. Rod-shaped microorganisms were noted to be predominant to the microbial community followed by cocci-shaped microorganisms.

Fig. 5
figure 5

Scanning electron micrograph of microbial biomass in activated sludge stressed with cerium oxide nanoparticles (nCeO2-NPs)

Full size image

Variation of physicochemical parameters

Table 2 illustrates changes physicochemical parameters in the media under nCeO2 NP effects. It was noted that microbial community’s abilities in removing nitrate, phosphate and COD decreased as the nCeO2 NP increased in the media. Carbon oxygen demand was the parameters with the highest removal at 89.74% (initial concentration: 593.5 mg/L) followed by nitrate (75%) and phosphate (13.81%). Microbial community also showed uptake of dissolved oxygen at up to 65.09% from the culture media containing 6.2 mg-DO/L. However, the uptake was hampered as nCeO2 NP concentration was increasing. The present study revealed that there was not significant variation of electric conductivity and pH with the percentage variation ranging from 643 to 717 µS/cm and 7.21–7.18 pH unit, respectively during the experimental period.

Table 2 physicochemical changes in the culture media at the end of incubation (5 days)
Full size table

Discussion

For many decades now, activated sludge under the Enhanced biological phosphorus removal (EBPR) has widely been used as means for the treatment of wastewater (Seviour et al. 2003; Oehmen et al. 2007). Even though many EBPR wastewater treatment plant (WWTP) is considered as a cost-effective and eco-friendly process for the removal of phosphorus and other pollutants, this process is also prone to instability and unreliability which is often attributed to competition between detrimental and beneficial microbes in the plants (Sidat et al. 1997; Kamika et al. 2014). Regardless of the fact that other factors such as environmental as well as anthropogenic have been investigated for their role in the deterioration of EBPR, very few studies have been carried out on the impact of nanoparticles on the activated sludge microbial community (Khan et al. 2002; Thomsen et al. 2007: Zhang et al. 2012). To our knowledge, the impact of nCeO2 NPs on activated sludge microbial community as well as on the bacteria is still unknown. In the study, the impact of nCeO2 NP on the bacterial community structure and species richness using Illimina sequencing of 16S rRNA gene in order to understand the influences of NPs on the useful bacterial community in an activated sludge system. The present study revealed that out of the 28,201 reads generated from the control samples, 18,330 reads (64.77%) were assigned to Proteobacteria phylum while 5527 reads (19.6%), 3260 reads (11.56%), and 719 reads (2.55%) were assigned to unclassified_Bacteria, Firmicutes and Actinobacteria, respectively (Fig. 1). In general, a decrease on microbial abundance was noted in samples treated with nCeO2 NPs with 10,856 reads (38.49%), 9256 reads (32.82%), and 7671 reads (27.2%) assigned to Proteobacteria phylum in samples treated with 10, 20, 30 and 40 mg/L, respectively. Similarly to the present study, common phyla Proteobacteria and Actinobacteria have been reported in the activated sludge (EBPR) as they have involved in several mechanism such as phosphorus and nitrate removal from the waste (Liu et al. 2005; Sanz and Kochling 2007; Kamika et al. 2014). According to Kamika et al. (2014), classes belong to the Proteobacteria phylum such as Gammaproteobacteria (80.57% of the all population), Alphaproteobacteria (5.19%) and Betaproteobacteria (3.19%) have been reported as functional bacteria for EBPR. The present study also agreed with Chen et al. (2014) who reported that the addition of NPs such as ZnO-NP and Ag-NP have a remarkable impact to the functional bacterial community in activated sludge. To further investigate the impacts of nCeO2 NPs on the bacterial community/diversity, it was revealed that 18 phyla were generated from the control samples whereas in the nCeO2 NPs-treatment samples over 11 phyla, 13 phyla, 10 phyla and 10 phyla, in S_A (10 mg/L), S_B (20 mg/L), S_C (30 mg/L) and S_D (40 mg/L) samples, respectively. This was also confirmed as the diversity index (Shannon) and Chao1 richness estimator revealed a significant different (p < 0.05) between treated samples and the control samples. Unlike the control samples, no significant difference (p > 0.05) was noted within treated samples. A further confirmation was noted as the species richness test indicated a drastic decrease of approximately 97.23–98.48% when comparing the control samples to nCeO2 NP treated samples.

When considering the lower taxonomic levels “genus”, it was observed that nCeO2 NPs could mostly affect the bacterial diversity and abundance of bacterial community as the control samples showed 239 genera whereas treated samples have genera decreasing from 123 to 99 genera. It was also revealed that nCeO2 NPs was affecting some bacteria especially unclassified ones while enhancing others and this was revealed when the abundance was higher in treated samples than in the control. The present study revealed the control samples were dominated by unclassified_Comamonadaceae, unclassified_Moraxellaceae, unclassified_Pseudomonadaceae, Novispirillum, Fusibacter, unclassified_Enterobacteriaceae, unclassified_Xanthomonadaceae, Shewanella, Proteocatella, unclassified_Carnobacteriaceae, Acinetobacter, Proteiniclasticum and Trichococcus occupying approximately 78.49% of the total community. This was also confirmed while investigating the impact of nCeO2-NPs on microbial cell structure using SEM. SEM images revealed that the microbial biomass were damaged and decreased over the increase of nCeO2-NPs concentration. Furthermore, samples had more rod-shaped microorganism that can be associated to Acinetobacter, Comamonadaceae, Moraxellaceae, Pseudomonadaceae despite of the presence of cocci-shaped microorganism such as Trichococcus (Fig. 5). Although most the dominant genus was unclassified, it was reported that genera and species belonging to Comamonadaceae family are considered as functional bacteria as they classified as denitrifiers (Khan et al. 2002; Sadaie et al. 2007). These authors revealed that the species belonging to these genera can be involved into the removal of phosphate in wastewater. Furthermore, previous studies also reported the predominance of several genera and species belonging to Moraxella, Pseudoxanthomonas, Comamonadas in activated sludge (Naili et al. 2015). Khan et al. (2002) also reported that species belong to comamonadaceae are primary degrading denitrifiers in activated sludge.

As the concentration of nCeO2 NP increased, samples showed a decrease of approximately 28.6% (20,136 reads) to 57.1% (12,084 reads) reads in the samples treated with 10 mg/L-CeO2 and 40 mg/L-CeO2, respectively. This was also noted with the number of OTUS which appeared to be approximately 27,967 OTUs from the control samples while the sample with highest nCeO2 NP revealed a total of 6433 OTUs. However, the relative abundance of two functional bacterial genera (Trichococcus and Acinetobacter) was found to alternatively dominate treated sample populations whereas most of those from the control samples saw their growth slowing down and inhibited. Vande Walle et al. (2012) disagreed with the findings from control samples by reporting that Acinetobacter, Aeromonas and Trichococcus as the predominant functional bacterial genera within urban sewer infrastructure. According to Lv et al. (2014), Trichococcus is among the most abundant genera responsible for denitrifying and aerobic phosphorus removal in the activated sludge. This genus was found to be enhanced in the present study highlighting that nCeO2 NPs are beneficial to their growth in the activated sludge and this similarly to Acinetobacter. The importance of Trichococcus species was further reported by Scheff et al. (1984) who revealed that their presence from bulking sludge. Despite their presence, the inhibition of phosphate removal from the treated samples as compared to nitrate removal could be due to the drastic inhibition of the activities of enzyme catalysing the degradation of polyphosphate such as adenylate kinase (ADK) and polyphosphate kinase (PPK) (Table 2). These enzymes have been reported as responsible in releasing and taking up phosphorus from the activated sludge, respectively (Chen et al. 2012). Furthermore, since unclassified bacteria appeared to be sensitive to nCeO2-NPs and this coupled with the inhibition of phosphate removal, it can be hypothesized that these unclassified bacteria were phosphate accumulating organisms (PAOs). It should be mentioned that the inhibition of phosphate removal is of great concern since this pollutant is considered the main responsible of eutrophication (Kamika et al. 2014). The effect of nCeO2 NPs was mostly observed with less abundant bacterial species such as sludge bulking bacterial species (Dechloromonas and Thauera), ammonia-oxidizing bacterial species (Zoogloea, Methyloversatilis), denitrifying bacterial species (Thauera, Azoarcus, Acidovorax, Comamonas, Pseudomonas, Paracoccus, Ochrobactrum, Hyphomicrobium and Nitrospira), Sulfate-reducing bacterial genera (Desulfomicrobium and Paracoccus), phosphate removing bacteria genera (Dechloromonas, Azospira, unclassified_Burkholderiales_incertae_sedis), and bacteria involved in flocs stabilization (Caldilinea) which showed an significant decrease over the gradual increase of nCeO NPs (Juretschko et al. 2002; Daims et al. 2006). Nevertheless, this did not affect the removal of COD and nitrate from the treated samples. This appeared to be contradictory as the enzymes associated with denitrification were affected by the increase of nCeO2 (Fig. 4). However, these enzymes have differently been affected with respect to nCeO2 NPs concentration. Nitrite reductase was less sensitive toward nCeO2 NPs increase than nitrate reductases. It has been reported that denitrifying bacteria convert nitrate into nitrogen gas via an enzymatic pathway consisting of four successive steps involving nitrate reductase (NaR), nitrite reductase (NiR), nitric oxide reductase, and nitrous oxide reductase in the periplasm and/or cytoplasm (Adav et al. 2010). Although the nCeO2 NPs were noted to promote the growth of some bacterial species while slowing those of others, it was unclear to know the real cause of such behavior as unclassified bacteria were mostly affected by the toxic effects of test NPs. Similar to the present study, Das et al. (2012) reported that bacteria community have four general exposure responses namely (1) intolerant, (2) impacted but recovering, (3) tolerant, and (4) stimulated when exposed to nanoparticles such as nAg-NP. Meli et al. (2016) also revealed that moderate concentrations of nanoparticles such as nZnO could accelerate the growth of some types of denitrifying bacteria and promote the growth of some pathogenic bacteria, and can also destroy the integrity of the cell membrane of Nitrosomonas europaea. Although, very little information is available on how these nCeO2 NPs affect microbial communities in activated sludge, effect of other NPs have been reported. The impact of nCeO2 NP on microbial community has also been reported by Antisari et al. (2013) who revealed that though microbial biomass was not statistically affected by nCeO2 NPs, the microbial stress or changes was noted. Beside of nCeO2, other engineered metal oxides-NPs such as nAg NPs (Das et al. 2012), nZnO NPS (Meli et al. 2016) and TiO2 NPs (Shah et al. 2014) have also been reported to have toxic effects on microbial community from several ecosystem. Jeong et al. (2014) also revealed the impact of nAg-NPs on bacterial community from wastewater treatment systems. These authors revealed that nitrifying bacteria are most susceptible to NPs such as nAg.

In conclusion, the present study provided a comprehensive insight in the effect of nCeO-NPs to bacterial community structure of activated sludge using Illumina sequencing. The present results revealed that Proteobacteria was the most predominant phylum in both treated and not-treated samples with nCeO2 NPs with exception in the 30 mg-nCeO2/L and 40 mg-nCeO2/L treated samples. The number of genus in control samples was found to be the lowest compared to treated samples as a large number of orders could not be classified. Despite of inhibiting some bacterial species especially the less abundant and unclassified ones, nCeO2 NPs appeared to enhance the growth of some bacterial species such as Trichococcus and Acinetobacter. Nevertheless, this enhancement did not increase the removal of phosphate in the treated samples. The results can extend our biological knowledge by revealing that nCeO2 NPs at moderate concentration could be beneficial as they enhanced some bacterial species involved nitrification, denitrification, and phosphorylation cycles in EBPR. More studies are needed to further understand the mechanism involved in the enhancement of bacteria growth by nCeO2 NPs as well as the inhibition of phosphate due to continuous addition of nCeO2-NPs.