- Email: [email protected]
Cadmium (II) alters the microbial community structure and molecular ecological network in activated sludge system
Environmental Pollution 255 (2019) 113225
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Cadmium (II) alters the microbial community structure and molecular ecological network in activated sludge system* Xiaohui Wang a, Tao Ya a, Minglu Zhang b, Lin Liu a, Pengfei Hou a, Shaoyong Lu c, * a Beijing Engineering Research Center of Environmental Material for Water Purification, College of Chemical Engineering, Beijing University of Chemical Technology, Beijing, 100029, China b Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry, Beijing Technology and Business University, Beijing, 100048, China c State Environmental Protection Scientific Observation and Research Station for Lake Dongtinghu (SEPSORSLD), National Engineering Laboratory for Lake Pollution Control and Ecological Restoration, State Key Laboratory of Environmental Criteria an Risk Assessment, Research Centre of Lake Environment, Chinese Research Academy of Environmental Sciences, Beijing, 100012, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 April 2019 Received in revised form 8 September 2019 Accepted 8 September 2019 Available online 18 September 2019
Cadmium (II) can potentially alter the microbial community structure and molecular ecological network in activated sludge systems. In this study, we used Illumina sequencing combined with an RMT-based network approach to show the response of the microbial community and its network structure to Cd (II) in activated sludge systems. The results demonstrated that 1 mg/L Cd (II) did not have chronic negative effects on chemical oxygen demand (COD) reduction and denitrification processes, but negatively affected the nitrification process and phosphorus removal. In contrast, 10 mg/L Cd (II) adversely affected both COD and nutrient removal, and reduced the microbial diversity and changed the overall microbial community structure. The relative abundances of Nitrosomonadaceae, Nitrospira, Accumulibacter and Acinetobacter, which are involved in nitrogen removal, significantly decreased with increases in the Cd (II) concentration. In addition, molecular ecological network analysis showed that the networks sizes in the presence of higher levels of Cd (II) were smaller than in the control, but the nodes were more closely connected with neighbors. These shifts in bacterial abundance and the bacterial network structure may be responsible for the deterioration of COD and nutrient removal. Overall, this study provides new insights into the effects of Cd (II) on the bacterial community and its interactions in activated sludge systems. © 2019 Published by Elsevier Ltd.
Keywords: Cadmium (II) Activated sludge system Microbial community structure Molecular ecological network
1. Introduction Cd (II) is a toxic heavy metal that ranks seventh in the priority list of hazardous substances (Deycard et al., 2014). It can be transported into cells and damage both the structure and functionality of proteins, leading to teratogenicity. In addition, Cd (II) can replace some essential metals at metabolic sites and inhibit the function of some physiological cations (Amor et al., 2001). However, Cd (II) is of great commercial significance, and is widely used in various industrial processes (e.g. as a chemical stabilizer as well as in batteries, pigments and solar panels) (Chen et al., 2014). This usage of Cd
* This Paper has been recommended for acceptance by Christian Sonne. * Corresponding author. Chinese Research Academy of Environmental Sciences, Beijing, 100012, China. E-mail address: [email protected] (S. Lu).
https://doi.org/10.1016/j.envpol.2019.113225 0269-7491/© 2019 Published by Elsevier Ltd.
(II) inevitably leads to its discharge into the wastewater collection system from where it eventually reaches biological wastewater treatment plants (WWTPs). It has been reported that Cd (II) influent concentrations at eight WWTPs in central China ranged from 0.54 to 1.39 mg/L (Xu et al., 2017). Because of its toxic properties, Cd (II) can adversely affect the microbial communities within a biological WWTP and thus affect its performance. Some researchers have examined the influences of Cd (II) on the performance of activated sludge systems and microbial community structures. Mertoglu et al. reported that 10 mg/L Cd (II) led to a drop in ammonia removal efficiency from 99% to 10% (Mertoglu et al., 2008). Correspondingly, a study by Chen et al. (2014) showed that 10 mg/L Cd (II) significantly worsened nitrification and phosphorus removal, and decreased the abundance of nitrite oxidizing bacteria (NOB) and polyphosphate accumulating organisms (PAOs) in a lab-scale sequencing biological reactor (SBR). Although these studies provide valuable insights into the potential influence of Cd
2
X. Wang et al. / Environmental Pollution 255 (2019) 113225
(II) on the performance and microbial communities of SBRs, little is known about the interactions among various populations and their response to Cd (II) perturbation in activated sludge systems. Activated sludge is a highly complex ecosystem of bacteria, eukaryotes, archaea and viruses, in which bacteria are the dominant microbes. Microorganisms in activated sludge systems do not exist in isolation but coexist within complicated networks with a multitude of interactions (Faust et al., 2012). These interactions can have a positive impact, a negative impact, or no impact on other species involved in the interaction web (Faust & Raes, 2012; Liang et al., 2016). Most positive impacts are attributable to commensalism or mutualism, while negative impacts can be due to predation, amensalism, and competition (Faust et al., 2012). Through these microbial interactions, activated sludge systems can remove organics, nutrients, and toxins (Xiaohui et al., 2012). A high concentration of Cd (II) entering activated sludge systems can affect the growth and metabolism of the microbes within them. It has been reported that Cd (II) exposure can promote reactive oxygen species (ROS) production and accumulation, which might inhibit microbial enzymatic activities. This might promote microbial cooperation or competition and further influence microbial interactions (Ghoul & Mitri, 2016; Li et al., 2018). However, deciphering the potential network interaction among microbial communities remains challenging, because it is difficult to examine microbial interactions directly. Recently, researchers have developed a novel random matrix theory (RMT)-based network approach to elucidate molecular ecological networks (MENs) (Deng et al., 2012; Deng et al., 2016). This approach has been used to describe the potential interactions of complex microbial communities in various environments, including soil (Deng et al., 2016; Ding et al., 2015), anaerobic bioreactors (Wu et al., 2016), and activated sludge (Li et al., 2017; Wang et al., 2017b). In this study, three lab-scale SBRs fed with 0 (SBR 0), 1 (SBR 1), and 10 (SBR 10) mg/L Cd (II) were operated for 80 days. High throughput MiSeq sequencing and an RMT-based MEN analysis were used to answer the following questions about activated sludge systems: (i) how does Cd (II) influence the reactors’ performance, and (ii) how does Cd (II) change the microbial community and its network structure? 2. Materials and methods
wastewater was pumped into the SBRs in the first 10 min of a cycle. In the aerobic stage, air was provided steadily using an electromagnetic air pump and an air flow meter to obtain a typical range of dissolved oxygen (DO) concentration between 2 and 2.5 mg/L, and the synthetic wastewater pH was controlled at around 7.5 ± 0.2 by adding of 1 M HCl or 1 M NaOH throughout the experiment. The hydraulic retention time was set at about 16 h and the solids retention time was controlled at approximately 20 d in the three SBRs. 2.3. Exposure experiments The three reactors were operated for 40 d until the wastewater treatment performance reached a stable state. Subsequently, 80 d exposure experiments were conducted to evaluate the performance, microbial components, and their interactions under different Cd (II) concentrations. The SBRs were exposed to no Cd (II) (SBR 0), 1 (SBR 1) and 10 (SBR 10) mg/L Cd (II). During the experiment period, DO and pH were measured every day while COD, TP, NHþ 4 -N, and NO3 -N were measured every two days, and MLSS and MLVSS were measured every five days. The measurement of the pH, þ MLSS, MLVSS, COD, NO 3 -N, NH4 -N and TP were performed according to the methods described by Gilcreas (2012). 2.4. DNA extraction and PCR amplification Biomass samples were taken daily from SBR 0, SBR 1 and SBR 10 during the final eight days of the study (i.e. days 73e80). A total of 24 samples were collected. The pretreatment of each activated sludge sample was based on our previous publication (Hai et al., 2014). DNA extraction was then operated using a PowerSoil DNA Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA) following the manufacturer's instructions. The forward primer and reverse primer were 50 -ACTCCTACGGGAGGCAGCAG-30 and 50 -GGACTACHVGGGTWTCTAAT-30 , which targeted the V4 region of the bacterial 16S rRNA and were used for polymerase chain reaction (PCR) amplification. The detailed PCR program was described previously (Hai et al., 2014). The reaction system was performed in 20 mL mixtures, and the amplification of each sample was repeated three times.
2.1. Synthetic wastewater
2.5. MiSeq sequencing
The influent wastewater used in this experiment was synthetic wastewater. COD, NHþ 4 -N, and phosphorus were supplemented to the medium in the form of glucose, NH4Cl, and KH2PO4 at approximately 500, 25, 10 mg/L, respectively. The medium was made up of 10 mg/L of MgSO4$7H2O and 5 mg/L of CaCl2 as well as 0.5 mL/L of trace element solution, according to Chen et al. (2014). The trace elements solution contained 1500 g/L FeCl3$6H2O, 30 g/L CuSO4$5H2O, 180 g/L KI, 150 g/L H3BO3, 120 g/L MnCl2$4H2O, 120 g/L ZnSO4$7H2O, 60 g/L Na2MoO4$2H2O, 150 g/L CoCl2$6H2O, and 3000 g/L EDTA.
An AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, CA, USA) was used for the purification of PCR products and 16S rRNA sequencing was conducted using an IlluminaMiSeq by the commercial company Majorbio (Shanghai, China). Low-quality sequences were removed following Wang et al. (2017a). High-quality sequences were assigned to operational taxonomic units (OTUs) using QIIME software with a 97% identity threshold. Annotation taxonomic information of the OTU representative sequence was performed using RDP Classifier (Version 2.4). 2.6. Statistical analysis
2.2. Reactor set-up and operation conditions Three identical lab-scale SBRs were operated in this study, each with a working volume of approximately 4 L. All three reactors were inoculated with 2 L activated sludge collected from Gaobeidian WWTP (Beijing, China). All SBRs were maintained at ambient temperature (20 ± 2 C) and operated with a cycle time of 8 h, including an anaerobic period of 100 min, an aerobic period of 140 min, an anoxic period of 60 min, a settling period of 90 min, a discharge period of 10 min and an idle period of 80 min. Synthetic
Alpha-diversity indices were used to evaluate bacterial richness and diversity. Principal coordinate analysis (PCoA) was performed to examine the overall variation among bacterial communities of the 24 samples. All statistical analyses were carried out using various packages within the statistical program R (http://www.rproject.org). To illuminate the effect of Cd (II) on microbial interactions, three MEN construction and network topology characterizations were processed using the online MENA pipeline that is available at http://
X. Wang et al. / Environmental Pollution 255 (2019) 113225
ieg2.ou.edu/MENA (Deng et al., 2016; Tu et al., 2015). Briefly, a Pearson correlation matrix was constructed based on the standardized relative abundance (SRA) of individual OTUs and then the correlation matrix was transformed into a similar matrix (Steve & Dong, 2008). Next, RMT was used to automatically identify an appropriate similarity threshold (St) (Feng et al., 2006; Luo et al., 2007). An adjacency matrix encoded with the connection strength can be subsequently obtained by retaining all of the OTUs whose abundance similarity values were greater than the determined threshold. For statistical comparisons of the networks, the Maslov-Sneppen procedure was explored to construct the random networks corresponding to all three experimental networks (Maslov & Sneppen, 2002). Based on the adjacency matrix, a network graph was constructed to visualize the microbial interactions using Cytoscape 3.2.1 software.
3. Results and discussion 3.1. Effects of Cd (II) on the SBR performance Fig. 1 summarizes the effluent COD concentrations during the 80 d operation. It was found that 1 mg/L Cd (II) had no detectable impact on COD removal, as the average effluent concentration of COD in SBR 1 was similar to the control (SBR 0). However, the effluent COD concentrations in SBR 10 obviously increased during the last 20 days from 28.1 to 35.64 mg/L, which indicates that 10 mg/L Cd (II) can eventually worsen organic matter removal efficiency. As Fig. 1 shows, the effluent NHþ 4 , NO3 , and TP concentrations after adding 10 mg/L Cd (II) gradually increased with operation time, reaching 6.68, 7.61, and 1.73 mg/L, respectively. This suggests that 10 mg/L Cd (II) induces serious deterioration of nutrient
3
removal in the activated sludge treatment process. Similar trends for NHþ 4 -N and TP were found in SBR 1. 3.2. Effects of Cd (II) on the overall bacterial communities The high-quality sequences for the 24 sludge samples ranged from 27,590 to 41,113. The library size of each sample was then normalized to 27,590 to conduct the data analysis. The average OTU number of SBR 10 was 382, significantly lower than that of the control (700), while the average OTU number of SBR 1 (673) was similar to the control. Similarly, the Shannon-Wiener index in SBR 10 (2.522) was significantly lower than that of the control (4.365) and SBR 1 (4.204) was similar to the control. These results indicate that 10 mg/L Cd (II) reduced bacterial diversity and richness. The OTU-based PCoA plot (Supplementary Fig. S1) showed that samples from different concentrations of Cd (II) were generally distributed separately, especially for samples from 10 mg/L Cd (II). Nonparametric multivariate statistical methods including analysis of similarities (ANOSIM) and Adonis (Supplementary Table S2) also showed that the bacterial communities of SBR 10 and SBR 1 were both significantly different from the control. Overall, 10 mg/L Cd (II) not only significantly reduced bacterial diversity but also changed the bacterial community structure. Although 1 mg/L Cd (II) did not reduce bacterial diversity, it significantly altered the bacterial community structure. 3.3. Significantly changed populations in the presence of Cd (II) At the phylum level, the bacterial communities of the three reactors were primarily composed of Saccharibacteria (28.06e80.31%), Proteobacteria (6.90e33.74%), Chloroflexi (4.08e17.13%), Acinobacteria (2.74e4.74%) and Bacteroidetes
Fig. 1. The effluent concentrations of COD, TP, NHþ 4 and NO3 in three SBRs during 80 d operation.
4
X. Wang et al. / Environmental Pollution 255 (2019) 113225
(2.79e8.51%) (Fig. 2). The relative abundance of Saccharibacteria drastically increased, while Proteobacteria and Chloroflexi prominently decreased in the presence of Cd (II). The average relative abundances of Saccharibacteria in SBR 10 and SBR 1 were 80.31% and 42.07%, respectively, significantly higher than that of the control (28.06%). This finding is consistent with the previous study by Wang et al. (2016), who found that Saccharibacteria were the predominant bacteria in biofilm systems under 10e50 mg/L Cd (II). Previous studies have shown that Saccharibacteria are responsible for the degradation of various organic compounds and were also widely detected in WWTPs (Kindaichi et al., 2016). In contrast, the relative abundances of Proteobacteria in SBR 10 and SBR 1 were 6.90% and 24.1%, significantly lower than in the control (33.74%). It has been reported that Proteobacteria is also a dominant phylum in the presence of other heavy metals such as Ag, Cu, and Fe (Reis et al., 2016; Sun et al., 2016; Y et al., 2014). Furthermore, Proteobacteria is closely associated with the removal of organics and nutrients; thus the negative impact of some heavy metals on their abundance might adversely affect the performance of WWTPs (Jeon et al., 2003; Y et al., 2014). Cd (II) also decreased the richness of Chloroflexi which are involved in carbon oxidation and nitrification in wastewater treatment process (Kragelund et al., 2007). A heat map of the top 50 abundant genera was used to intuitively show the differences in their relative abundances (Fig. 3). Most genera accounted for a lower proportion in SBR 10 than in SBR 1 and the control. Specifically, one genus from Saccharibacteria showed high tolerance to Cd (II), and its relative abundances under 10 mg/L Cd (II) were 74%, higher than in SBR 0 (32%) and SBR 1 (48%). Thiothrix is widely detected in WWTPs and can cause filamentous sludge bulking and foaming problems in activated sludge system (Gillan & Dubilier, 2004). The abundance of Thiothrix in SBR 1 (1%) was lower than in SBR 0 (2%), and it was almost absent from SBR 10. Some genera, such as Nannocystis, Haliangium, Aeromonas, Cytophagaceae, Terrimonas and members of Saprospiraceae, were enriched in SBR 1 and decreased in SBR 10. Saprospiraceae, typical sludge bulking microorganisms that are also denitrifiers, can produce extracellular polymeric substances (EPSs), and the higher relative abundance in SBR 10 compared to SBR 1 and SBR 0 indicates poorer sludge sedimentation performance. 10 mg/L Cd (II) significantly reduced the richness of Nitrosomonadaceae and Nitrospira, which are ammonia-oxidizing bacteria (AOB) and NOB in activated sludge, which might partly explain the deterioration of
nitrification in SBR 10. Furthermore, compared with the SBR 1, some denitrifiers such as Rhodobacter, Haliangium and Hyphomicrobium, were significantly decreased in SBR 10, which could be responsible for the relatively poor denitrification performance. In addition, after adding 1 mg/L Cd (II), the relative abundances of Accumulibacter and Acinetobacter, known as PAO, were 0.05% and 0.08%, respectively, which were significantly lower than those in the control (0.1% and 0.3%). When exposed to 10 mg/L Cd (II), Accumulibacter and Acinetobacter could not be detected. The abundance reduction of these PAOs may lead to a deterioration of phosphorus removal (Bond et al., 1999; García et al., 2006; He et al., 2007). 3.4. Effects of Cd (II) on the microbial interactions Three networks for SBR 0, SBR 1 and SBR 10 were constructed to determine the effects of Cd (II) on activated sludge microbial interactions. As shown in Table 1, the clustering coefficient, geodesic distance, and modularity of the three networks were significantly higher than random networks, indicating that the three networks were obviously different from the corresponding random networks. Cd (II) greatly decreased the network size regarding the nodes and higher concentrations of Cd (II) led to smaller network sizes with the SBR 1 and SBR 10 networks containing 457 nodes and 296 nodes, respectively, which was smaller than that of the control (544). Also, the network nodes under Cd (II) were more closely connected with each other, because SBR 10 and SBR 1 had a higher average connectivity (6.58 and 6.27) and clustering coefficient (0.24 and 0.21), and a lower geodesic distance (5.59 and 5.78) than those of the control network (3.94, 0.18 and 7.13, respectively). We explored the top seven nodes with higher connectivity in each SBR (Fig. 4). In the control network (Fig. 4A), the top seven nodes with the highest connectivity were linked with 169 first neighbors and formed a relatively complex network. In contrast, the seven nodes with higher connectivity in SBR 1 (Fig. 4B) and SBR 10 (Fig. 4C) were linked with 231 and 238 first neighbors, respectively, indicating that exposure to Cd (II) led to a more complex microbial network. All of these results indicate a potential structural change in the activated sludge microbial community in the presence of Cd (II). We speculate that Cd (II) probably made the nodes more closely connected with first neighbors to resist the negative effect of Cd (II). In addition, the seven most highly linked nodes in each network were
Fig. 2. Relative abundances of major phyla in the activated sludge samples. CK. (control sludge sample); Cd 1 (sludge under 1 mg/L Cd (II)); Cd 10 (sludge under 10 mg/L Cd (II)). 1e8 after “-” represent the eight replicate sludge samples taken from the same reactor.
X. Wang et al. / Environmental Pollution 255 (2019) 113225
5
Fig. 3. Heatmap analysis of bacterial community based on top 50 genera. CK (control sludge sample); Cd 1 (sludge under 1 mg/L Cd (II)); Cd 10 (sludge under 10 mg/L Cd (II)). 1e8 after “-” represent the eight replicate sludge samples taken from the same reactor.
Table 1 Network properties of the three empirical MENs and their corresponding random networks. System Empirical networks No. of original St OTUs SBR 0 1569 SBR 1 1285 SBR 10 1012
Random networks Network size (n)
0.94 544 0.95 457 0.92 296
Avg Avg geodesic connectivity distance
Avg clustering coefficient
Modularity (No. of modules)
Avg geodesic distance ± SD
Avg clustering coefficient ± SD
Avg modularity ± SD
3.94 6.27 6.58
0.18 0.21 0.24
0.81 (74) 0.60 (59) 0.58 (23)
4.16 ± 0.05 3.38 ± 0.04 3.16 ± 0.03
0.016 ± 0.003 0.06 ± 0.007 0.06 ± 0.007
0.51 ± 0.01 0.34 ± 0.01 0.33 ± 0.01
7.13 5.78 5.59
almost never shared among the three networks (with the exception of OTU 38, which existed both in SBR 0 and SBR 1), indicating that Cd (II) significantly affected the overall architecture of the network. Cd (II) had different influences on the microbial interactions of different phylogenetic groups. Within Proteobacteria, the top five nodes with higher connectivity under 1 mg/L Cd (II) had more complex interactions (77 nodes with 162 links) than those in the control (64 nodes with 100 links). However, in the presence of 10 mg/L Cd (II), the top five nodes with higher connectivity formed a less complicated network (61 nodes with 74 links) (Fig. 5). It appears that low concentrations of Cd (II) can promote microbial interactions, while higher concentrations of Cd (II) reduce the microbial interactions among Proteobacteria. A decreasing complexity of network interactions among microbial populations could lead to
decreased functional stability in an ecological system with different interaction types (Mougi & Kondoh, 2012). In an activated sludge system, Proteobacteria play a crucial role in the removal of organic matter and nutrients, and the decreased network complexity of Proteobacteria was probably a contributor to the deterioration of COD, nitrogen, and phosphorus removal. For the Saccharibacteria, however, the top five OTUs in the presence of Cd (II) had much more complex network structures than in the control (38 links), and the network interaction of SBR 10 (124 links) was more complicated than in SBR 1 (101 links) (Supplementary Figure S2). Meanwhile, the relative abundance of Saccharibacteria also increased with increasing Cd (II) concentration, which could have contributed to maintaining system stability in the presence of Cd (II).
6
X. Wang et al. / Environmental Pollution 255 (2019) 113225
Fig. 4. Variation of the network interactions of bacterial community under different Cd (II) concentration. The interactions of the top seven nodes (in the center) with the highest links and their neighbors in control (A), SBR 1 (B), SBR 10 (C). The different colours of nodes represent different phylum. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Chloroflexi. OTU 754 belonged to the family Saprospiraceae, which has been shown to occur in activated sludge and is responsible for the degradation of complex organic compounds (especially the aromatic compounds) as a carbon source (Mcilroy & Nielsen, 2014). OTU 885 was from genus Haliangium. Members of Haliangium were identified as active denitrifiers by 13C stable isotope probing (Mcilroy et al., 2016). These keystone species in the control network (SBR 0) were speculated to play important roles in organic compounds degradation and denitrification, and maintaining the reactor stability. Fewer module hubs in SBR 1 were detected (OTUs 31, 382, 1162 and 1261). OTU 1162 was closely related to family Anaerolineaceae, which was previously found to be abundant in an anaerobic reactor where it had the function of degradation of carbohydrates and other cell tissues (Lei et al., 2016). OTUs 382 and 1261 belong to phylum Proteobacteria. OTU 382 was derived from Defluviicoccus, which was identified as a glycogen accumulating organism (GAO). GAOs compete with PAOs for volatile fatty acid (VFA) uptake under anaerobic conditions in activated sludge systems, but they do not contribute to phosphorus removal (Mcilroy & Seviour, 2010; Oyserman et al., 2016), The increase of Defluviicoccus may be one explanation for the deterioration in phosphorus removal in SBR 1. OTU 1261 was related to unclassified NB1-j. The role of NB1-j in activated sludge is not yet well described. OTU 31 was from the genus Saccharibacteria and was enriched in SBR 10. Only three OTUs (19, 48 and 339) were detected as module hubs in SBR 10. OTU 19 belonged to the Caldilineaceae, which have been identified as filamentous bacteria responsible for the bulking and foaming problems in activated sludge systems (Guo & Zhang, 2012; Meng et al., 2006). OTU 339 was similar to the genus SB-5, and OTU 48 was related to Family-XIII. The roles of these two OTUs in biological wastewater treatment systems are not well described. In addition, none of the connectors in the SBR 0 network was detected. Three connectors (OTUs 14, 61 and 236) with higher links to other modules were detected in the network of SBR 1. OTU 14 belonged to Verrucomicrobia, which has been associated with the degradation of carbohydrates (Herlemann et al., 2013; Laura et al., 2012). OTU 236 was closely related to the Saprospiraceae, typical sludge-bulking microorganisms and denitrifiers. OTU 61 was closely related to Levilinea, which is an obligately anaerobic, filamentous microbe capable of growth on a range of carbohydrates (Yamada et al., 2006). Only one connector (OTU 2721) in the SBR 10 network was detected. It was derived from the Saprospiraceae. Interestingly, no network hubs were found in the three networks and no module hubs or connectors were shared among the three networks. These results suggested that Cd (II) greatly altered the key microbial populations and network structure.
4. Conclusion Fig. 5. Variation of the network interactions of Proteobacteria under different Cd (II) concentration. The interactions of the top five nodes (in the center) with the highest links and their neighbors in control (A), SBR 1 (B), SBR 10 (C). The different colours of nodes represent different phylum. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
The Z-P plot (Fig. 6) shows that the different nodes play distinct topological roles in the networks. Here, the majority of the nodes (about 99.1%, 98.5% and 98.6% of SBR 0, SBR 1 and SBR 10, respectively) were peripheral. Only 12 and 4 nodes were categorized as module hubs and network hubs in all three networks. Hubs have been proposed to be keystone taxa due to their important roles in network topology (Zhang et al., 2018). The five module hubs detected in the control network (SBR 0) were OTUs 56, 495, 754, 885, 1188. OTUs 56, 1188 and 495 were related to uncultured
In this study, 1 mg/L Cd (II) did not have chronic negative effects on COD removal and denitrification but led to a negative performance of both the nitrification process and phosphorus removal in the activated sludge system. Long-term operation in the presence of 10 mg/L Cd (II) adversely affected COD and nutrient uptake. Furthermore, 10 mg/L Cd (II) reduced the microbial diversity and changed the microbial community structure. The abundance of some genera associated with nutrient removal, including Nitrosomonadaceae, Nitrospira, Rhodobacter, Accumulibacter and Acinetobacter, decreased after the addition of 10 mg/L Cd (II). In addition, Cd (II) significantly decreased the network size but induced higher complexity of microbial interactions. This study increases our knowledge of the responses of bacterial community structures and their molecular ecological networks to Cd (II) in activated sludge systems.
X. Wang et al. / Environmental Pollution 255 (2019) 113225
7
Fig. 6. Classification of nodes to identify putative keystone species within the activated sludge system networks. (All nodes were divided into four categories according to the among-module connectivity (Pi) and within-module connectivity (Zi)).
Acknowledgements This study was supported by the National key research and development program of China (2016YFC0401105), Beijing Municipal Science and Technology Project (Z181100002418017) and the Open Research Fund Program of Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry (CP-2018-YB4). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.envpol.2019.113225. References Amor, L., Kennes, C., Veiga, M.C., 2001. Kinetics of inhibition in the biodegradation of monoaromatic hydrocarbons in presence of heavy metals. Bioresour. Technol. 78 (2), 181e185. Bond, P.L., Erhart, R., Wagner, M., Keller, J., Blackall, L.L., 1999. Identification of some of the major groups of bacteria in efficient and nonefficient biological phosphorus removal activated sludge systems. Appl. Environ. Microbiol. 65 (9), 4077e4084. Chen, H.B., Wang, D.B., Li, X.M., Yang, Q., Luo, K., Zeng, G.M., Tang, M.L., 2014. Effects of Cd(II) on wastewater biological nitrogen and phosphorus removal. Chemosphere 117 (1), 27e32. Deng, Y., Jiang, Y.H., Yang, Y., He, Z., Luo, F., Zhou, J., 2012. Molecular ecological network analyses. BMC Bioinf. 13 (1), 113. Deng, Y., Zhang, P., Qin, Y., Tu, Q., Yang, Y., He, Z., Schadt, C.W., Zhou, J., 2016. Network succession reveals the importance of competition in response to emulsified vegetable oil amendment for uranium bioremediation. Environ. Microbiol. 18 (1), 205e218. Deycard, V.N., Sch€ afer, J., Blanc, G., Coynel, A., Petit, J.C.J., Lanceleur, L., Dutruch, L., Bossy, C., Ventura, A., 2014. Contributions and potential impacts of seven priority substances (As, Cd, Cu, Cr, Ni, Pb, and Zn) to a major European Estuary (Gironde Estuary, France) from urban wastewater. Mar. Chem. 167, 123e134. Ding, J., Zhang, Y., Deng, Y., Cong, J., Lu, H., Sun, X., Yang, C., Yuan, T., Nostrand, J.D.V., Li, D., 2015. Integrated metagenomics and network analysis of soil microbial community of the forest timberline. Sci. Rep. 5 (7994), 7994. Faust, K., Raes, J., 2012. Microbial interactions: from networks to models. Nat. Rev. Microbiol. 10 (8), 538.
Faust, K., Sathirapongsasuti, J.F., Izard, J., Segata, N., Gevers, D., Raes, J., Huttenhower, C., 2012. Microbial co-occurrence relationships in the human microbiome. PLoS Comput. Biol. 8 (7), e1002606. Feng, L., Zhong, J., Yang, Y., Scheuermann, R.H., Zhou, J., 2006. Application of random matrix theory to biological networks. Phys. Lett. A 357 (6), 420e423. García, M.H., Ivanova, N., Kunin, V., Warnecke, F., Barry, K.W., Mchardy, A.C., Yeates, C., He, S., Salamov, A.A., Szeto, E., 2006. Metagenomic analysis of two enhanced biological phosphorus removal (EBPR) sludge communities. Nat. Biotechnol. 24 (10), 1263. Ghoul, M., Mitri, S., 2016. The ecology and evolution of microbial competition. Trends Microbiol. 24 (10), 833e845. Gilcreas, F.W., 2012. Standard methods for the examination of water and waste water. Am. J. Public Health Nation's Health 56 (3), 387e388. Gillan, D.C., Dubilier, N., 2004. Novel epibiotic thiothrix bacterium on a marine amphipod. Appl. Environ. Microbiol. 70 (6), 3772e3775. Guo, F., Zhang, T., 2012. Profiling bulking and foaming bacteria in activated sludge by high throughput sequencing. Water Res. 46 (8), 2772e2782. Hai, R., Wang, Y., Wang, X., Du, Z., Li, Y.J.P.O., 2014. Impacts of multiwalled carbon nanotubes on nutrient removal from wastewater and bacterial community structure in activated sludge, 9 (9), e107345. He, S., Gall, D.L., Mcmahon, K.D., 2007. “Candidatus Accumulibacter” population structure in enhanced biological phosphorus removal sludges as revealed by polyphosphate kinase genes. Appl. Environ. Microbiol. 73 (18), 5865. Herlemann, D.P.R., Daniel, L., Matthias, L., Klaus, J., Zongli, Z., Henrik, A., Andersson, A.F., Mbio, %J., 2013. Metagenomic de novo assembly of an aquatic representative of the verrucomicrobial class Spartobacteria, 4 (3), e00569. Jeon, C.O., Lee, D.S., Park, J.M., 2003. Microbial communities in activated sludge performing enhanced biological phosphorus removal in a sequencing batch reactor. Water Res. 37 (9), 2195e2205. Kindaichi, T., Yamaoka, S., Uehara, R., Ozaki, N., Ohashi, A., Albertsen, M., Nielsen, P.H., Nielsen, J.L., 2016. Phylogenetic diversity and ecophysiology of Candidate phylum Saccharibacteria in activated sludge. FEMS Microbiol. Ecol. 92 (6), fiw078. Kragelund, C., Levantesi, C., Borger, A., Thelen, K., Eikelboom, D., Tandoi, V., Kong, Y., Van, d.W.J., Krooneman, J., Rossetti, S., 2007. Identity, abundance and ecophysiology of filamentous Chloroflexi species present in activated sludge treatment plants. FEMS Microbiol. Ecol. 59 (3), 671e682. Laura, A.S., Waller, A.S., Mende, D.R., Kevin, B., Hanna, F., Yager, P.L., Connie, L., Jean Eric, T., Marianne, P., Friederike, H., 2012. Role for urea in nitrification by polar marine Archaea. Proc. Natl. Acad. Sci. U. S. A. 109 (44), 17989e17994. Lei, M., Wang, S., Li, B., Cao, T., Zhang, F., Zhong, W., Peng, Y., 2016. Effect of carbon source type on intracellular stored polymers during endogenous denitritation (ED) treating landfill leachate. Water Res. 100, 405e412. Li, Y., Lin, C.Q., Wang, Y.B., Gao, X.P., Xie, T., Hai, R.T., Wang, X.H., Zhang, X.W., 2017. Multi-criteria evaluation method for site selection of industrial wastewater discharge in coastal regions. J. Clean. Prod. 161, 1143e1152.
8
X. Wang et al. / Environmental Pollution 255 (2019) 113225
Li, S.S., Xu, Q.Y., Ma, B.R., Guo, L., She, Z.L., Zhao, Y.G., Gao, M.C., Jin, C.J., Dong, J.W., Wan, Y.P., 2018. Performance evaluation and microbial community of a sequencing batch reactor under divalent cadmium (Cd(II)) stress. Chem. Eng. J. 336, 325e333. Liang, Y., Zhao, H., Deng, Y., Zhou, J., Li, G., Sun, B., 2016. Long-term oil contamination alters the molecular ecological networks of soil microbial functional genes. Front. Microbiol. 7 (66111), 60. Luo, F., Yang, Y., Zhong, J., Gao, H., Khan, L., Thompson, D.K., Zhou, J., 2007. Constructing gene co-expression networks and predicting functions of unknown genes by random matrix theory. BMC Bioinf. 8 (1), 299-299. Maslov, S., Sneppen, K., 2002. Specificity and stability in topology of protein networks. Science 296 (5569), 910e913. Mcilroy, S.J., Nielsen, P.H., 2014. The Family Saprospiraceae. Springer Berlin Heidelberg. Mcilroy, S., Seviour, R.J., 2010. Elucidating further phylogenetic diversity among the Defluviicoccus-related glycogen-accumulating organisms in activated sludge. Environ. Microbiol. Rep. 1 (6), 563e568. Mcilroy, S.J., Starnawska, A., Starnawski, P., Saunders, A.M., Nierychlo, M., Nielsen, P.H., Nielsen, J.L., 2016. Identification of active denitrifiers in full-scale nutrient removal wastewater treatment systems. Environ. Microbiol. 18 (1), 50. Meng, F., Zhang, H., Yang, F., Li, Y., Xiao, J., Zhang, X., 2006. Effect of filamentous bacteria on membrane fouling in submerged membrane bioreactor. J. Membr. Sci. 272 (1), 161e168. Mertoglu, B., Semerci, N., Guler, N., Calli, B., Cecen, F., Saatci, A.M., 2008. Monitoring of population shifts in an enriched nitrifying system under gradually increased cadmium loading. J. Hazard Mater. 160 (2), 495e501. Mougi, A., Kondoh, M., 2012. Diversity of interaction types and ecological community stability. Science 337 (6092), 349e351. Oyserman, B.O., Noguera, D.R., Rio, T.G.D., Tringe, S.G., Mcmahon, K.D., 2016. Metatranscriptomic insights on gene expression and regulatory controls inCandidatusAccumulibacter phosphatis. ISME J. 10 (4), 810e822. Reis, M.P., Dias, M.F., Costa, P.S., Avila, M.P., Leite, L.R., Araújo, F.M.G.D., Salim, A.C.M., Bucciarelli-Rodriguez, M., Oliveira, G., Chartone-Souza, E., 2016. Metagenomic signatures of a tropical mining-impacted stream reveal complex microbial and metabolic networks. Chemosphere 161, 266e273. Steve, H., Dong, J., 2008. Geometric interpretation of gene coexpression network analysis. PLoS Comput. Biol. 4 (8), e1000117.
Sun, F.L., Fan, L.L., Xie, G.J., 2016. Effect of copper on the performance and bacterial communities of activated sludge using Illumina MiSeq platforms. Chemosphere 156, 212e219. Tu, Q., Yuan, M., He, Z., Deng, Y., Xue, K., Wu, L., Hobbie, S.E., Reich, P.B., Zhou, J., 2015. Fungal communities respond to long-term CO2 elevation by community reassembly. Appl. Environ. Microbiol. 81 (7), 2445e2454. Wang, Z., Gao, M., Wei, J., Ma, K., Zhang, J., Yang, Y., Yu, S., 2016. Extracellular polymeric substances, microbial activity and microbial community of biofilm and suspended sludge at different divalent cadmium concentrations. Bioresour. Technol. 205, 213e221. Wang, X., Li, J., Liu, R., Hai, R., Zou, D., Zhu, X., Luo, N., 2017a. Responses of bacterial communities to CuO nanoparticles in activated sludge system. Environ. Sci. Technol. 51 (10). Wang, X., Zheng, Q., Yuan, Y., Hai, R., Zou, D., 2017b. Bacterial community and molecular ecological network in response to Cr 2 O 3 nanoparticles in activated sludge system. Chemosphere 188, 10. Wu, L., Yang, Y., Chen, S., Zhao, M., Zhu, Z., Yang, S., Qu, Y., Ma, Q., He, Z., Zhou, J., 2016. Long-term successional dynamics of microbial association networks in anaerobic digestion processes. Water Res. 104, 1e10. Xiaohui, W., Man, H., Yu, X., Xianghua, W., Kun, D.J.A., Microbiology, E., 2012. Pyrosequencing analysis of bacterial diversity in 14 wastewater treatment systems in China, 78 (19), 7042e7047. Xu, Q., Li, X., Ding, R., Wang, D., Liu, Y., Wang, Q., Zhao, J., Chen, F., Zeng, G., Yang, Q., Li, H., 2017. Understanding and mitigating the toxicity of cadmium to the anaerobic fermentation of waste activated sludge. Water Res. 124, 269e279. Y, Y., J, Q., J, M., Q.,W., J, W., M.,L., JM, T., PJ, A., 2014. Pyrosequencing reveals higher impact of silver nanoparticles than Agþ on the microbial community structure of activated sludge. Water Res. 48 (1), 317e325. Yamada, T., Sekiguchi, Y., Hanada, S., Imachi, H., Ohashi, A., Harada, H., Kamagata, Y., 2006. Anaerolinea Thermolimosa Sp. nov., Levilinea Saccharolytica Gen. nov., Sp. Nov. And Leptolinea Tardivitalis Gen. nov., Sp. nov., Novel Filamentous Anaerobes, and Description of the New Classes Anaerolineae Classis Nov. and Caldilineae Classis Nov. in the B. Electrical Transmission in A New Age Conference. Zhang, S., Zhou, Z., Li, Y., Meng, F., 2018. Deciphering the core fouling-causing microbiota in a membrane bioreactor: low abundance but important roles. Chemosphere 195, 108e118.
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Cadmium (II) alters the microbial community structure and molecular ecological network in activated sludge system* Xiaohui Wang a, Tao Ya a, Minglu Zhang b, Lin Liu a, Pengfei Hou a, Shaoyong Lu c, * a Beijing Engineering Research Center of Environmental Material for Water Purification, College of Chemical Engineering, Beijing University of Chemical Technology, Beijing, 100029, China b Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry, Beijing Technology and Business University, Beijing, 100048, China c State Environmental Protection Scientific Observation and Research Station for Lake Dongtinghu (SEPSORSLD), National Engineering Laboratory for Lake Pollution Control and Ecological Restoration, State Key Laboratory of Environmental Criteria an Risk Assessment, Research Centre of Lake Environment, Chinese Research Academy of Environmental Sciences, Beijing, 100012, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 April 2019 Received in revised form 8 September 2019 Accepted 8 September 2019 Available online 18 September 2019
Cadmium (II) can potentially alter the microbial community structure and molecular ecological network in activated sludge systems. In this study, we used Illumina sequencing combined with an RMT-based network approach to show the response of the microbial community and its network structure to Cd (II) in activated sludge systems. The results demonstrated that 1 mg/L Cd (II) did not have chronic negative effects on chemical oxygen demand (COD) reduction and denitrification processes, but negatively affected the nitrification process and phosphorus removal. In contrast, 10 mg/L Cd (II) adversely affected both COD and nutrient removal, and reduced the microbial diversity and changed the overall microbial community structure. The relative abundances of Nitrosomonadaceae, Nitrospira, Accumulibacter and Acinetobacter, which are involved in nitrogen removal, significantly decreased with increases in the Cd (II) concentration. In addition, molecular ecological network analysis showed that the networks sizes in the presence of higher levels of Cd (II) were smaller than in the control, but the nodes were more closely connected with neighbors. These shifts in bacterial abundance and the bacterial network structure may be responsible for the deterioration of COD and nutrient removal. Overall, this study provides new insights into the effects of Cd (II) on the bacterial community and its interactions in activated sludge systems. © 2019 Published by Elsevier Ltd.
Keywords: Cadmium (II) Activated sludge system Microbial community structure Molecular ecological network
1. Introduction Cd (II) is a toxic heavy metal that ranks seventh in the priority list of hazardous substances (Deycard et al., 2014). It can be transported into cells and damage both the structure and functionality of proteins, leading to teratogenicity. In addition, Cd (II) can replace some essential metals at metabolic sites and inhibit the function of some physiological cations (Amor et al., 2001). However, Cd (II) is of great commercial significance, and is widely used in various industrial processes (e.g. as a chemical stabilizer as well as in batteries, pigments and solar panels) (Chen et al., 2014). This usage of Cd
* This Paper has been recommended for acceptance by Christian Sonne. * Corresponding author. Chinese Research Academy of Environmental Sciences, Beijing, 100012, China. E-mail address: [email protected] (S. Lu).
https://doi.org/10.1016/j.envpol.2019.113225 0269-7491/© 2019 Published by Elsevier Ltd.
(II) inevitably leads to its discharge into the wastewater collection system from where it eventually reaches biological wastewater treatment plants (WWTPs). It has been reported that Cd (II) influent concentrations at eight WWTPs in central China ranged from 0.54 to 1.39 mg/L (Xu et al., 2017). Because of its toxic properties, Cd (II) can adversely affect the microbial communities within a biological WWTP and thus affect its performance. Some researchers have examined the influences of Cd (II) on the performance of activated sludge systems and microbial community structures. Mertoglu et al. reported that 10 mg/L Cd (II) led to a drop in ammonia removal efficiency from 99% to 10% (Mertoglu et al., 2008). Correspondingly, a study by Chen et al. (2014) showed that 10 mg/L Cd (II) significantly worsened nitrification and phosphorus removal, and decreased the abundance of nitrite oxidizing bacteria (NOB) and polyphosphate accumulating organisms (PAOs) in a lab-scale sequencing biological reactor (SBR). Although these studies provide valuable insights into the potential influence of Cd
2
X. Wang et al. / Environmental Pollution 255 (2019) 113225
(II) on the performance and microbial communities of SBRs, little is known about the interactions among various populations and their response to Cd (II) perturbation in activated sludge systems. Activated sludge is a highly complex ecosystem of bacteria, eukaryotes, archaea and viruses, in which bacteria are the dominant microbes. Microorganisms in activated sludge systems do not exist in isolation but coexist within complicated networks with a multitude of interactions (Faust et al., 2012). These interactions can have a positive impact, a negative impact, or no impact on other species involved in the interaction web (Faust & Raes, 2012; Liang et al., 2016). Most positive impacts are attributable to commensalism or mutualism, while negative impacts can be due to predation, amensalism, and competition (Faust et al., 2012). Through these microbial interactions, activated sludge systems can remove organics, nutrients, and toxins (Xiaohui et al., 2012). A high concentration of Cd (II) entering activated sludge systems can affect the growth and metabolism of the microbes within them. It has been reported that Cd (II) exposure can promote reactive oxygen species (ROS) production and accumulation, which might inhibit microbial enzymatic activities. This might promote microbial cooperation or competition and further influence microbial interactions (Ghoul & Mitri, 2016; Li et al., 2018). However, deciphering the potential network interaction among microbial communities remains challenging, because it is difficult to examine microbial interactions directly. Recently, researchers have developed a novel random matrix theory (RMT)-based network approach to elucidate molecular ecological networks (MENs) (Deng et al., 2012; Deng et al., 2016). This approach has been used to describe the potential interactions of complex microbial communities in various environments, including soil (Deng et al., 2016; Ding et al., 2015), anaerobic bioreactors (Wu et al., 2016), and activated sludge (Li et al., 2017; Wang et al., 2017b). In this study, three lab-scale SBRs fed with 0 (SBR 0), 1 (SBR 1), and 10 (SBR 10) mg/L Cd (II) were operated for 80 days. High throughput MiSeq sequencing and an RMT-based MEN analysis were used to answer the following questions about activated sludge systems: (i) how does Cd (II) influence the reactors’ performance, and (ii) how does Cd (II) change the microbial community and its network structure? 2. Materials and methods
wastewater was pumped into the SBRs in the first 10 min of a cycle. In the aerobic stage, air was provided steadily using an electromagnetic air pump and an air flow meter to obtain a typical range of dissolved oxygen (DO) concentration between 2 and 2.5 mg/L, and the synthetic wastewater pH was controlled at around 7.5 ± 0.2 by adding of 1 M HCl or 1 M NaOH throughout the experiment. The hydraulic retention time was set at about 16 h and the solids retention time was controlled at approximately 20 d in the three SBRs. 2.3. Exposure experiments The three reactors were operated for 40 d until the wastewater treatment performance reached a stable state. Subsequently, 80 d exposure experiments were conducted to evaluate the performance, microbial components, and their interactions under different Cd (II) concentrations. The SBRs were exposed to no Cd (II) (SBR 0), 1 (SBR 1) and 10 (SBR 10) mg/L Cd (II). During the experiment period, DO and pH were measured every day while COD, TP, NHþ 4 -N, and NO3 -N were measured every two days, and MLSS and MLVSS were measured every five days. The measurement of the pH, þ MLSS, MLVSS, COD, NO 3 -N, NH4 -N and TP were performed according to the methods described by Gilcreas (2012). 2.4. DNA extraction and PCR amplification Biomass samples were taken daily from SBR 0, SBR 1 and SBR 10 during the final eight days of the study (i.e. days 73e80). A total of 24 samples were collected. The pretreatment of each activated sludge sample was based on our previous publication (Hai et al., 2014). DNA extraction was then operated using a PowerSoil DNA Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA) following the manufacturer's instructions. The forward primer and reverse primer were 50 -ACTCCTACGGGAGGCAGCAG-30 and 50 -GGACTACHVGGGTWTCTAAT-30 , which targeted the V4 region of the bacterial 16S rRNA and were used for polymerase chain reaction (PCR) amplification. The detailed PCR program was described previously (Hai et al., 2014). The reaction system was performed in 20 mL mixtures, and the amplification of each sample was repeated three times.
2.1. Synthetic wastewater
2.5. MiSeq sequencing
The influent wastewater used in this experiment was synthetic wastewater. COD, NHþ 4 -N, and phosphorus were supplemented to the medium in the form of glucose, NH4Cl, and KH2PO4 at approximately 500, 25, 10 mg/L, respectively. The medium was made up of 10 mg/L of MgSO4$7H2O and 5 mg/L of CaCl2 as well as 0.5 mL/L of trace element solution, according to Chen et al. (2014). The trace elements solution contained 1500 g/L FeCl3$6H2O, 30 g/L CuSO4$5H2O, 180 g/L KI, 150 g/L H3BO3, 120 g/L MnCl2$4H2O, 120 g/L ZnSO4$7H2O, 60 g/L Na2MoO4$2H2O, 150 g/L CoCl2$6H2O, and 3000 g/L EDTA.
An AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, CA, USA) was used for the purification of PCR products and 16S rRNA sequencing was conducted using an IlluminaMiSeq by the commercial company Majorbio (Shanghai, China). Low-quality sequences were removed following Wang et al. (2017a). High-quality sequences were assigned to operational taxonomic units (OTUs) using QIIME software with a 97% identity threshold. Annotation taxonomic information of the OTU representative sequence was performed using RDP Classifier (Version 2.4). 2.6. Statistical analysis
2.2. Reactor set-up and operation conditions Three identical lab-scale SBRs were operated in this study, each with a working volume of approximately 4 L. All three reactors were inoculated with 2 L activated sludge collected from Gaobeidian WWTP (Beijing, China). All SBRs were maintained at ambient temperature (20 ± 2 C) and operated with a cycle time of 8 h, including an anaerobic period of 100 min, an aerobic period of 140 min, an anoxic period of 60 min, a settling period of 90 min, a discharge period of 10 min and an idle period of 80 min. Synthetic
Alpha-diversity indices were used to evaluate bacterial richness and diversity. Principal coordinate analysis (PCoA) was performed to examine the overall variation among bacterial communities of the 24 samples. All statistical analyses were carried out using various packages within the statistical program R (http://www.rproject.org). To illuminate the effect of Cd (II) on microbial interactions, three MEN construction and network topology characterizations were processed using the online MENA pipeline that is available at http://
X. Wang et al. / Environmental Pollution 255 (2019) 113225
ieg2.ou.edu/MENA (Deng et al., 2016; Tu et al., 2015). Briefly, a Pearson correlation matrix was constructed based on the standardized relative abundance (SRA) of individual OTUs and then the correlation matrix was transformed into a similar matrix (Steve & Dong, 2008). Next, RMT was used to automatically identify an appropriate similarity threshold (St) (Feng et al., 2006; Luo et al., 2007). An adjacency matrix encoded with the connection strength can be subsequently obtained by retaining all of the OTUs whose abundance similarity values were greater than the determined threshold. For statistical comparisons of the networks, the Maslov-Sneppen procedure was explored to construct the random networks corresponding to all three experimental networks (Maslov & Sneppen, 2002). Based on the adjacency matrix, a network graph was constructed to visualize the microbial interactions using Cytoscape 3.2.1 software.
3. Results and discussion 3.1. Effects of Cd (II) on the SBR performance Fig. 1 summarizes the effluent COD concentrations during the 80 d operation. It was found that 1 mg/L Cd (II) had no detectable impact on COD removal, as the average effluent concentration of COD in SBR 1 was similar to the control (SBR 0). However, the effluent COD concentrations in SBR 10 obviously increased during the last 20 days from 28.1 to 35.64 mg/L, which indicates that 10 mg/L Cd (II) can eventually worsen organic matter removal efficiency. As Fig. 1 shows, the effluent NHþ 4 , NO3 , and TP concentrations after adding 10 mg/L Cd (II) gradually increased with operation time, reaching 6.68, 7.61, and 1.73 mg/L, respectively. This suggests that 10 mg/L Cd (II) induces serious deterioration of nutrient
3
removal in the activated sludge treatment process. Similar trends for NHþ 4 -N and TP were found in SBR 1. 3.2. Effects of Cd (II) on the overall bacterial communities The high-quality sequences for the 24 sludge samples ranged from 27,590 to 41,113. The library size of each sample was then normalized to 27,590 to conduct the data analysis. The average OTU number of SBR 10 was 382, significantly lower than that of the control (700), while the average OTU number of SBR 1 (673) was similar to the control. Similarly, the Shannon-Wiener index in SBR 10 (2.522) was significantly lower than that of the control (4.365) and SBR 1 (4.204) was similar to the control. These results indicate that 10 mg/L Cd (II) reduced bacterial diversity and richness. The OTU-based PCoA plot (Supplementary Fig. S1) showed that samples from different concentrations of Cd (II) were generally distributed separately, especially for samples from 10 mg/L Cd (II). Nonparametric multivariate statistical methods including analysis of similarities (ANOSIM) and Adonis (Supplementary Table S2) also showed that the bacterial communities of SBR 10 and SBR 1 were both significantly different from the control. Overall, 10 mg/L Cd (II) not only significantly reduced bacterial diversity but also changed the bacterial community structure. Although 1 mg/L Cd (II) did not reduce bacterial diversity, it significantly altered the bacterial community structure. 3.3. Significantly changed populations in the presence of Cd (II) At the phylum level, the bacterial communities of the three reactors were primarily composed of Saccharibacteria (28.06e80.31%), Proteobacteria (6.90e33.74%), Chloroflexi (4.08e17.13%), Acinobacteria (2.74e4.74%) and Bacteroidetes
Fig. 1. The effluent concentrations of COD, TP, NHþ 4 and NO3 in three SBRs during 80 d operation.
4
X. Wang et al. / Environmental Pollution 255 (2019) 113225
(2.79e8.51%) (Fig. 2). The relative abundance of Saccharibacteria drastically increased, while Proteobacteria and Chloroflexi prominently decreased in the presence of Cd (II). The average relative abundances of Saccharibacteria in SBR 10 and SBR 1 were 80.31% and 42.07%, respectively, significantly higher than that of the control (28.06%). This finding is consistent with the previous study by Wang et al. (2016), who found that Saccharibacteria were the predominant bacteria in biofilm systems under 10e50 mg/L Cd (II). Previous studies have shown that Saccharibacteria are responsible for the degradation of various organic compounds and were also widely detected in WWTPs (Kindaichi et al., 2016). In contrast, the relative abundances of Proteobacteria in SBR 10 and SBR 1 were 6.90% and 24.1%, significantly lower than in the control (33.74%). It has been reported that Proteobacteria is also a dominant phylum in the presence of other heavy metals such as Ag, Cu, and Fe (Reis et al., 2016; Sun et al., 2016; Y et al., 2014). Furthermore, Proteobacteria is closely associated with the removal of organics and nutrients; thus the negative impact of some heavy metals on their abundance might adversely affect the performance of WWTPs (Jeon et al., 2003; Y et al., 2014). Cd (II) also decreased the richness of Chloroflexi which are involved in carbon oxidation and nitrification in wastewater treatment process (Kragelund et al., 2007). A heat map of the top 50 abundant genera was used to intuitively show the differences in their relative abundances (Fig. 3). Most genera accounted for a lower proportion in SBR 10 than in SBR 1 and the control. Specifically, one genus from Saccharibacteria showed high tolerance to Cd (II), and its relative abundances under 10 mg/L Cd (II) were 74%, higher than in SBR 0 (32%) and SBR 1 (48%). Thiothrix is widely detected in WWTPs and can cause filamentous sludge bulking and foaming problems in activated sludge system (Gillan & Dubilier, 2004). The abundance of Thiothrix in SBR 1 (1%) was lower than in SBR 0 (2%), and it was almost absent from SBR 10. Some genera, such as Nannocystis, Haliangium, Aeromonas, Cytophagaceae, Terrimonas and members of Saprospiraceae, were enriched in SBR 1 and decreased in SBR 10. Saprospiraceae, typical sludge bulking microorganisms that are also denitrifiers, can produce extracellular polymeric substances (EPSs), and the higher relative abundance in SBR 10 compared to SBR 1 and SBR 0 indicates poorer sludge sedimentation performance. 10 mg/L Cd (II) significantly reduced the richness of Nitrosomonadaceae and Nitrospira, which are ammonia-oxidizing bacteria (AOB) and NOB in activated sludge, which might partly explain the deterioration of
nitrification in SBR 10. Furthermore, compared with the SBR 1, some denitrifiers such as Rhodobacter, Haliangium and Hyphomicrobium, were significantly decreased in SBR 10, which could be responsible for the relatively poor denitrification performance. In addition, after adding 1 mg/L Cd (II), the relative abundances of Accumulibacter and Acinetobacter, known as PAO, were 0.05% and 0.08%, respectively, which were significantly lower than those in the control (0.1% and 0.3%). When exposed to 10 mg/L Cd (II), Accumulibacter and Acinetobacter could not be detected. The abundance reduction of these PAOs may lead to a deterioration of phosphorus removal (Bond et al., 1999; García et al., 2006; He et al., 2007). 3.4. Effects of Cd (II) on the microbial interactions Three networks for SBR 0, SBR 1 and SBR 10 were constructed to determine the effects of Cd (II) on activated sludge microbial interactions. As shown in Table 1, the clustering coefficient, geodesic distance, and modularity of the three networks were significantly higher than random networks, indicating that the three networks were obviously different from the corresponding random networks. Cd (II) greatly decreased the network size regarding the nodes and higher concentrations of Cd (II) led to smaller network sizes with the SBR 1 and SBR 10 networks containing 457 nodes and 296 nodes, respectively, which was smaller than that of the control (544). Also, the network nodes under Cd (II) were more closely connected with each other, because SBR 10 and SBR 1 had a higher average connectivity (6.58 and 6.27) and clustering coefficient (0.24 and 0.21), and a lower geodesic distance (5.59 and 5.78) than those of the control network (3.94, 0.18 and 7.13, respectively). We explored the top seven nodes with higher connectivity in each SBR (Fig. 4). In the control network (Fig. 4A), the top seven nodes with the highest connectivity were linked with 169 first neighbors and formed a relatively complex network. In contrast, the seven nodes with higher connectivity in SBR 1 (Fig. 4B) and SBR 10 (Fig. 4C) were linked with 231 and 238 first neighbors, respectively, indicating that exposure to Cd (II) led to a more complex microbial network. All of these results indicate a potential structural change in the activated sludge microbial community in the presence of Cd (II). We speculate that Cd (II) probably made the nodes more closely connected with first neighbors to resist the negative effect of Cd (II). In addition, the seven most highly linked nodes in each network were
Fig. 2. Relative abundances of major phyla in the activated sludge samples. CK. (control sludge sample); Cd 1 (sludge under 1 mg/L Cd (II)); Cd 10 (sludge under 10 mg/L Cd (II)). 1e8 after “-” represent the eight replicate sludge samples taken from the same reactor.
X. Wang et al. / Environmental Pollution 255 (2019) 113225
5
Fig. 3. Heatmap analysis of bacterial community based on top 50 genera. CK (control sludge sample); Cd 1 (sludge under 1 mg/L Cd (II)); Cd 10 (sludge under 10 mg/L Cd (II)). 1e8 after “-” represent the eight replicate sludge samples taken from the same reactor.
Table 1 Network properties of the three empirical MENs and their corresponding random networks. System Empirical networks No. of original St OTUs SBR 0 1569 SBR 1 1285 SBR 10 1012
Random networks Network size (n)
0.94 544 0.95 457 0.92 296
Avg Avg geodesic connectivity distance
Avg clustering coefficient
Modularity (No. of modules)
Avg geodesic distance ± SD
Avg clustering coefficient ± SD
Avg modularity ± SD
3.94 6.27 6.58
0.18 0.21 0.24
0.81 (74) 0.60 (59) 0.58 (23)
4.16 ± 0.05 3.38 ± 0.04 3.16 ± 0.03
0.016 ± 0.003 0.06 ± 0.007 0.06 ± 0.007
0.51 ± 0.01 0.34 ± 0.01 0.33 ± 0.01
7.13 5.78 5.59
almost never shared among the three networks (with the exception of OTU 38, which existed both in SBR 0 and SBR 1), indicating that Cd (II) significantly affected the overall architecture of the network. Cd (II) had different influences on the microbial interactions of different phylogenetic groups. Within Proteobacteria, the top five nodes with higher connectivity under 1 mg/L Cd (II) had more complex interactions (77 nodes with 162 links) than those in the control (64 nodes with 100 links). However, in the presence of 10 mg/L Cd (II), the top five nodes with higher connectivity formed a less complicated network (61 nodes with 74 links) (Fig. 5). It appears that low concentrations of Cd (II) can promote microbial interactions, while higher concentrations of Cd (II) reduce the microbial interactions among Proteobacteria. A decreasing complexity of network interactions among microbial populations could lead to
decreased functional stability in an ecological system with different interaction types (Mougi & Kondoh, 2012). In an activated sludge system, Proteobacteria play a crucial role in the removal of organic matter and nutrients, and the decreased network complexity of Proteobacteria was probably a contributor to the deterioration of COD, nitrogen, and phosphorus removal. For the Saccharibacteria, however, the top five OTUs in the presence of Cd (II) had much more complex network structures than in the control (38 links), and the network interaction of SBR 10 (124 links) was more complicated than in SBR 1 (101 links) (Supplementary Figure S2). Meanwhile, the relative abundance of Saccharibacteria also increased with increasing Cd (II) concentration, which could have contributed to maintaining system stability in the presence of Cd (II).
6
X. Wang et al. / Environmental Pollution 255 (2019) 113225
Fig. 4. Variation of the network interactions of bacterial community under different Cd (II) concentration. The interactions of the top seven nodes (in the center) with the highest links and their neighbors in control (A), SBR 1 (B), SBR 10 (C). The different colours of nodes represent different phylum. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Chloroflexi. OTU 754 belonged to the family Saprospiraceae, which has been shown to occur in activated sludge and is responsible for the degradation of complex organic compounds (especially the aromatic compounds) as a carbon source (Mcilroy & Nielsen, 2014). OTU 885 was from genus Haliangium. Members of Haliangium were identified as active denitrifiers by 13C stable isotope probing (Mcilroy et al., 2016). These keystone species in the control network (SBR 0) were speculated to play important roles in organic compounds degradation and denitrification, and maintaining the reactor stability. Fewer module hubs in SBR 1 were detected (OTUs 31, 382, 1162 and 1261). OTU 1162 was closely related to family Anaerolineaceae, which was previously found to be abundant in an anaerobic reactor where it had the function of degradation of carbohydrates and other cell tissues (Lei et al., 2016). OTUs 382 and 1261 belong to phylum Proteobacteria. OTU 382 was derived from Defluviicoccus, which was identified as a glycogen accumulating organism (GAO). GAOs compete with PAOs for volatile fatty acid (VFA) uptake under anaerobic conditions in activated sludge systems, but they do not contribute to phosphorus removal (Mcilroy & Seviour, 2010; Oyserman et al., 2016), The increase of Defluviicoccus may be one explanation for the deterioration in phosphorus removal in SBR 1. OTU 1261 was related to unclassified NB1-j. The role of NB1-j in activated sludge is not yet well described. OTU 31 was from the genus Saccharibacteria and was enriched in SBR 10. Only three OTUs (19, 48 and 339) were detected as module hubs in SBR 10. OTU 19 belonged to the Caldilineaceae, which have been identified as filamentous bacteria responsible for the bulking and foaming problems in activated sludge systems (Guo & Zhang, 2012; Meng et al., 2006). OTU 339 was similar to the genus SB-5, and OTU 48 was related to Family-XIII. The roles of these two OTUs in biological wastewater treatment systems are not well described. In addition, none of the connectors in the SBR 0 network was detected. Three connectors (OTUs 14, 61 and 236) with higher links to other modules were detected in the network of SBR 1. OTU 14 belonged to Verrucomicrobia, which has been associated with the degradation of carbohydrates (Herlemann et al., 2013; Laura et al., 2012). OTU 236 was closely related to the Saprospiraceae, typical sludge-bulking microorganisms and denitrifiers. OTU 61 was closely related to Levilinea, which is an obligately anaerobic, filamentous microbe capable of growth on a range of carbohydrates (Yamada et al., 2006). Only one connector (OTU 2721) in the SBR 10 network was detected. It was derived from the Saprospiraceae. Interestingly, no network hubs were found in the three networks and no module hubs or connectors were shared among the three networks. These results suggested that Cd (II) greatly altered the key microbial populations and network structure.
4. Conclusion Fig. 5. Variation of the network interactions of Proteobacteria under different Cd (II) concentration. The interactions of the top five nodes (in the center) with the highest links and their neighbors in control (A), SBR 1 (B), SBR 10 (C). The different colours of nodes represent different phylum. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
The Z-P plot (Fig. 6) shows that the different nodes play distinct topological roles in the networks. Here, the majority of the nodes (about 99.1%, 98.5% and 98.6% of SBR 0, SBR 1 and SBR 10, respectively) were peripheral. Only 12 and 4 nodes were categorized as module hubs and network hubs in all three networks. Hubs have been proposed to be keystone taxa due to their important roles in network topology (Zhang et al., 2018). The five module hubs detected in the control network (SBR 0) were OTUs 56, 495, 754, 885, 1188. OTUs 56, 1188 and 495 were related to uncultured
In this study, 1 mg/L Cd (II) did not have chronic negative effects on COD removal and denitrification but led to a negative performance of both the nitrification process and phosphorus removal in the activated sludge system. Long-term operation in the presence of 10 mg/L Cd (II) adversely affected COD and nutrient uptake. Furthermore, 10 mg/L Cd (II) reduced the microbial diversity and changed the microbial community structure. The abundance of some genera associated with nutrient removal, including Nitrosomonadaceae, Nitrospira, Rhodobacter, Accumulibacter and Acinetobacter, decreased after the addition of 10 mg/L Cd (II). In addition, Cd (II) significantly decreased the network size but induced higher complexity of microbial interactions. This study increases our knowledge of the responses of bacterial community structures and their molecular ecological networks to Cd (II) in activated sludge systems.
X. Wang et al. / Environmental Pollution 255 (2019) 113225
7
Fig. 6. Classification of nodes to identify putative keystone species within the activated sludge system networks. (All nodes were divided into four categories according to the among-module connectivity (Pi) and within-module connectivity (Zi)).
Acknowledgements This study was supported by the National key research and development program of China (2016YFC0401105), Beijing Municipal Science and Technology Project (Z181100002418017) and the Open Research Fund Program of Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry (CP-2018-YB4). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.envpol.2019.113225. References Amor, L., Kennes, C., Veiga, M.C., 2001. Kinetics of inhibition in the biodegradation of monoaromatic hydrocarbons in presence of heavy metals. Bioresour. Technol. 78 (2), 181e185. Bond, P.L., Erhart, R., Wagner, M., Keller, J., Blackall, L.L., 1999. Identification of some of the major groups of bacteria in efficient and nonefficient biological phosphorus removal activated sludge systems. Appl. Environ. Microbiol. 65 (9), 4077e4084. Chen, H.B., Wang, D.B., Li, X.M., Yang, Q., Luo, K., Zeng, G.M., Tang, M.L., 2014. Effects of Cd(II) on wastewater biological nitrogen and phosphorus removal. Chemosphere 117 (1), 27e32. Deng, Y., Jiang, Y.H., Yang, Y., He, Z., Luo, F., Zhou, J., 2012. Molecular ecological network analyses. BMC Bioinf. 13 (1), 113. Deng, Y., Zhang, P., Qin, Y., Tu, Q., Yang, Y., He, Z., Schadt, C.W., Zhou, J., 2016. Network succession reveals the importance of competition in response to emulsified vegetable oil amendment for uranium bioremediation. Environ. Microbiol. 18 (1), 205e218. Deycard, V.N., Sch€ afer, J., Blanc, G., Coynel, A., Petit, J.C.J., Lanceleur, L., Dutruch, L., Bossy, C., Ventura, A., 2014. Contributions and potential impacts of seven priority substances (As, Cd, Cu, Cr, Ni, Pb, and Zn) to a major European Estuary (Gironde Estuary, France) from urban wastewater. Mar. Chem. 167, 123e134. Ding, J., Zhang, Y., Deng, Y., Cong, J., Lu, H., Sun, X., Yang, C., Yuan, T., Nostrand, J.D.V., Li, D., 2015. Integrated metagenomics and network analysis of soil microbial community of the forest timberline. Sci. Rep. 5 (7994), 7994. Faust, K., Raes, J., 2012. Microbial interactions: from networks to models. Nat. Rev. Microbiol. 10 (8), 538.
Faust, K., Sathirapongsasuti, J.F., Izard, J., Segata, N., Gevers, D., Raes, J., Huttenhower, C., 2012. Microbial co-occurrence relationships in the human microbiome. PLoS Comput. Biol. 8 (7), e1002606. Feng, L., Zhong, J., Yang, Y., Scheuermann, R.H., Zhou, J., 2006. Application of random matrix theory to biological networks. Phys. Lett. A 357 (6), 420e423. García, M.H., Ivanova, N., Kunin, V., Warnecke, F., Barry, K.W., Mchardy, A.C., Yeates, C., He, S., Salamov, A.A., Szeto, E., 2006. Metagenomic analysis of two enhanced biological phosphorus removal (EBPR) sludge communities. Nat. Biotechnol. 24 (10), 1263. Ghoul, M., Mitri, S., 2016. The ecology and evolution of microbial competition. Trends Microbiol. 24 (10), 833e845. Gilcreas, F.W., 2012. Standard methods for the examination of water and waste water. Am. J. Public Health Nation's Health 56 (3), 387e388. Gillan, D.C., Dubilier, N., 2004. Novel epibiotic thiothrix bacterium on a marine amphipod. Appl. Environ. Microbiol. 70 (6), 3772e3775. Guo, F., Zhang, T., 2012. Profiling bulking and foaming bacteria in activated sludge by high throughput sequencing. Water Res. 46 (8), 2772e2782. Hai, R., Wang, Y., Wang, X., Du, Z., Li, Y.J.P.O., 2014. Impacts of multiwalled carbon nanotubes on nutrient removal from wastewater and bacterial community structure in activated sludge, 9 (9), e107345. He, S., Gall, D.L., Mcmahon, K.D., 2007. “Candidatus Accumulibacter” population structure in enhanced biological phosphorus removal sludges as revealed by polyphosphate kinase genes. Appl. Environ. Microbiol. 73 (18), 5865. Herlemann, D.P.R., Daniel, L., Matthias, L., Klaus, J., Zongli, Z., Henrik, A., Andersson, A.F., Mbio, %J., 2013. Metagenomic de novo assembly of an aquatic representative of the verrucomicrobial class Spartobacteria, 4 (3), e00569. Jeon, C.O., Lee, D.S., Park, J.M., 2003. Microbial communities in activated sludge performing enhanced biological phosphorus removal in a sequencing batch reactor. Water Res. 37 (9), 2195e2205. Kindaichi, T., Yamaoka, S., Uehara, R., Ozaki, N., Ohashi, A., Albertsen, M., Nielsen, P.H., Nielsen, J.L., 2016. Phylogenetic diversity and ecophysiology of Candidate phylum Saccharibacteria in activated sludge. FEMS Microbiol. Ecol. 92 (6), fiw078. Kragelund, C., Levantesi, C., Borger, A., Thelen, K., Eikelboom, D., Tandoi, V., Kong, Y., Van, d.W.J., Krooneman, J., Rossetti, S., 2007. Identity, abundance and ecophysiology of filamentous Chloroflexi species present in activated sludge treatment plants. FEMS Microbiol. Ecol. 59 (3), 671e682. Laura, A.S., Waller, A.S., Mende, D.R., Kevin, B., Hanna, F., Yager, P.L., Connie, L., Jean Eric, T., Marianne, P., Friederike, H., 2012. Role for urea in nitrification by polar marine Archaea. Proc. Natl. Acad. Sci. U. S. A. 109 (44), 17989e17994. Lei, M., Wang, S., Li, B., Cao, T., Zhang, F., Zhong, W., Peng, Y., 2016. Effect of carbon source type on intracellular stored polymers during endogenous denitritation (ED) treating landfill leachate. Water Res. 100, 405e412. Li, Y., Lin, C.Q., Wang, Y.B., Gao, X.P., Xie, T., Hai, R.T., Wang, X.H., Zhang, X.W., 2017. Multi-criteria evaluation method for site selection of industrial wastewater discharge in coastal regions. J. Clean. Prod. 161, 1143e1152.
8
X. Wang et al. / Environmental Pollution 255 (2019) 113225
Li, S.S., Xu, Q.Y., Ma, B.R., Guo, L., She, Z.L., Zhao, Y.G., Gao, M.C., Jin, C.J., Dong, J.W., Wan, Y.P., 2018. Performance evaluation and microbial community of a sequencing batch reactor under divalent cadmium (Cd(II)) stress. Chem. Eng. J. 336, 325e333. Liang, Y., Zhao, H., Deng, Y., Zhou, J., Li, G., Sun, B., 2016. Long-term oil contamination alters the molecular ecological networks of soil microbial functional genes. Front. Microbiol. 7 (66111), 60. Luo, F., Yang, Y., Zhong, J., Gao, H., Khan, L., Thompson, D.K., Zhou, J., 2007. Constructing gene co-expression networks and predicting functions of unknown genes by random matrix theory. BMC Bioinf. 8 (1), 299-299. Maslov, S., Sneppen, K., 2002. Specificity and stability in topology of protein networks. Science 296 (5569), 910e913. Mcilroy, S.J., Nielsen, P.H., 2014. The Family Saprospiraceae. Springer Berlin Heidelberg. Mcilroy, S., Seviour, R.J., 2010. Elucidating further phylogenetic diversity among the Defluviicoccus-related glycogen-accumulating organisms in activated sludge. Environ. Microbiol. Rep. 1 (6), 563e568. Mcilroy, S.J., Starnawska, A., Starnawski, P., Saunders, A.M., Nierychlo, M., Nielsen, P.H., Nielsen, J.L., 2016. Identification of active denitrifiers in full-scale nutrient removal wastewater treatment systems. Environ. Microbiol. 18 (1), 50. Meng, F., Zhang, H., Yang, F., Li, Y., Xiao, J., Zhang, X., 2006. Effect of filamentous bacteria on membrane fouling in submerged membrane bioreactor. J. Membr. Sci. 272 (1), 161e168. Mertoglu, B., Semerci, N., Guler, N., Calli, B., Cecen, F., Saatci, A.M., 2008. Monitoring of population shifts in an enriched nitrifying system under gradually increased cadmium loading. J. Hazard Mater. 160 (2), 495e501. Mougi, A., Kondoh, M., 2012. Diversity of interaction types and ecological community stability. Science 337 (6092), 349e351. Oyserman, B.O., Noguera, D.R., Rio, T.G.D., Tringe, S.G., Mcmahon, K.D., 2016. Metatranscriptomic insights on gene expression and regulatory controls inCandidatusAccumulibacter phosphatis. ISME J. 10 (4), 810e822. Reis, M.P., Dias, M.F., Costa, P.S., Avila, M.P., Leite, L.R., Araújo, F.M.G.D., Salim, A.C.M., Bucciarelli-Rodriguez, M., Oliveira, G., Chartone-Souza, E., 2016. Metagenomic signatures of a tropical mining-impacted stream reveal complex microbial and metabolic networks. Chemosphere 161, 266e273. Steve, H., Dong, J., 2008. Geometric interpretation of gene coexpression network analysis. PLoS Comput. Biol. 4 (8), e1000117.
Sun, F.L., Fan, L.L., Xie, G.J., 2016. Effect of copper on the performance and bacterial communities of activated sludge using Illumina MiSeq platforms. Chemosphere 156, 212e219. Tu, Q., Yuan, M., He, Z., Deng, Y., Xue, K., Wu, L., Hobbie, S.E., Reich, P.B., Zhou, J., 2015. Fungal communities respond to long-term CO2 elevation by community reassembly. Appl. Environ. Microbiol. 81 (7), 2445e2454. Wang, Z., Gao, M., Wei, J., Ma, K., Zhang, J., Yang, Y., Yu, S., 2016. Extracellular polymeric substances, microbial activity and microbial community of biofilm and suspended sludge at different divalent cadmium concentrations. Bioresour. Technol. 205, 213e221. Wang, X., Li, J., Liu, R., Hai, R., Zou, D., Zhu, X., Luo, N., 2017a. Responses of bacterial communities to CuO nanoparticles in activated sludge system. Environ. Sci. Technol. 51 (10). Wang, X., Zheng, Q., Yuan, Y., Hai, R., Zou, D., 2017b. Bacterial community and molecular ecological network in response to Cr 2 O 3 nanoparticles in activated sludge system. Chemosphere 188, 10. Wu, L., Yang, Y., Chen, S., Zhao, M., Zhu, Z., Yang, S., Qu, Y., Ma, Q., He, Z., Zhou, J., 2016. Long-term successional dynamics of microbial association networks in anaerobic digestion processes. Water Res. 104, 1e10. Xiaohui, W., Man, H., Yu, X., Xianghua, W., Kun, D.J.A., Microbiology, E., 2012. Pyrosequencing analysis of bacterial diversity in 14 wastewater treatment systems in China, 78 (19), 7042e7047. Xu, Q., Li, X., Ding, R., Wang, D., Liu, Y., Wang, Q., Zhao, J., Chen, F., Zeng, G., Yang, Q., Li, H., 2017. Understanding and mitigating the toxicity of cadmium to the anaerobic fermentation of waste activated sludge. Water Res. 124, 269e279. Y, Y., J, Q., J, M., Q.,W., J, W., M.,L., JM, T., PJ, A., 2014. Pyrosequencing reveals higher impact of silver nanoparticles than Agþ on the microbial community structure of activated sludge. Water Res. 48 (1), 317e325. Yamada, T., Sekiguchi, Y., Hanada, S., Imachi, H., Ohashi, A., Harada, H., Kamagata, Y., 2006. Anaerolinea Thermolimosa Sp. nov., Levilinea Saccharolytica Gen. nov., Sp. Nov. And Leptolinea Tardivitalis Gen. nov., Sp. nov., Novel Filamentous Anaerobes, and Description of the New Classes Anaerolineae Classis Nov. and Caldilineae Classis Nov. in the B. Electrical Transmission in A New Age Conference. Zhang, S., Zhou, Z., Li, Y., Meng, F., 2018. Deciphering the core fouling-causing microbiota in a membrane bioreactor: low abundance but important roles. Chemosphere 195, 108e118.