Abundance and community structure of ammonia-oxidizing bacteria in activated sludge from different geographic regions in China

As the first and rate-limiting step of the nitrification process, ammonia oxidation plays an important role in the treatment of activated sludge containing organic nitrogen wastewater process (Zhao et al. 2014a). This process is mainly completed by ammonia-oxidizing bacteria (AOB) and ammonia-oxidizing archaea (AOA) (Beman & Francis 2006). But many previous studies have indicated that the abundance of AOA is much lower than those of AOB in activated sludge samples of wastewater treatment plants (WWTPs) (Siripong & Rittmann 2007; Li et al. 2013). Li et al. (2013) suggested that AOA may be a major contributor to ammonia oxidation in natural habitats but play minor roles in highly aerated activated sludge. Limpiyakorn et al. (2011) also indicated that the archaeal amoA gene is below the limit of detection in WWTPs with higher influent ammonium levels. Therefore, AOB received more attention in studies of nitrification in activated sludge samples. The abundance and community structure of AOB in WWTPs have been previously studied (Otawa et al. 2006; Limpiyakorn et al. 2011; Lu et al. 2014; Fitzgerald et al. 2015). Nevertheless, little information on the relationships between AOB communities and water quality parameters have been reported, and the characterization of AOB communities in WWTPs from different geographic locations has never been considered.

Unlike other environmental microorganisms, the abundance and structure of AOB communities in WWTPs are widely considered to be associated with various plant-specific parameters. For instance, Wells et al. (2011) found that the operation and wastewater conditions were related to the abundance of AOB. Therefore, most previous studies paid attention to the distribution of AOB in plant-specific wastewater conditions or water quality parameters. Limpiyakorn et al. (2011) investigated the abundance of archaeal and bacterial amoA genes in activated sludge of full-scale WWTPs, and found a fourfold difference among the two groups in favor of AOB. Fitzgerald et al. (2015) investigated the microbial communities in reactors with efficient nitrification at low concentrations of dissolved oxygen (DO) and concluded that Pseudomonas, Xanthomonadaceae, Rhodococcus, and Sphingomonas are involved in nitrification under low DO conditions. However, these studies did not consider the effects of other water quality parameters, especially geographical localization of WWTPs. Sloan et al. (2006) and Lozupone (2007) indicated that the local microbial communities in influent wastewater might also affect microbial communities in WWTPs, because they were selected by local environmental conditions according to neutral processes. In addition, different operational parts in a similar WWTP may also affect microbial communities, where the activated sludge post-processing and mechanical purifying sections had the highest amount of bacteria (Kowalski et al. 2017). In recent years, few studies have investigated the difference of bacterial communities in geographic distribution WWTPs (Zhang et al. 2012; Zhao et al. 2014a; Niu et al. 2016). We hypothesized that the geographic distribution has relevant effects on the abundance and community structure of AOB in activated sludge of distinct WWTPs. To confirm this hypothesis, activated sludge samples were collected from different regions to reveal the relationships between geographic distribution and other water quality parameters with the abundance and community structure of AOB in activated sludge of WWTPs.

In this study, activated sludge samples were collected from seven WWTPs in seven cities of China. These sampling sites are located in different regions of eastern, central and western China. Real-time quantitative polymerase chain reaction (qPCR), as well as construction of clone libraries combined with correlation-based data analysis, were employed. We are trying to investigate: (1) if the abundance and community structure of AOB in activated sludge samples differ in geographically distinct WWTPs in China; (2) the relationships between water quality parameters and ecological distribution of AOB.

Samples of activated sludge were collected from the aeration tanks of seven municipal WWTPs in seven cities in China, including Wuxi (WX), Ma'anshan (MAS), Hefei (HF), Nanjing (NJ), Taiyuan (TY), Xi'an (XA) and Urumqi (ULMQ). These sampling sites are located in different parts of China, of which WX, MAS, HF, NJ are located in the eastern part; XA, TY are located in the central part; and ULMQ is located in the western part. All the samples were collected in July and August of 2012. Most WWTPs of this study mainly treat domestic wastewater, but HF and XA also process of industrial wastewater (ca. 30–40% of total). Details of the seven WWTPs, such as wastewater treatment processes, wastewater influent parameters and effluent parameters, are shown in Table S1 (available with the online version of this paper).

Each activated sludge sample (about 1 L) was collected in sterile plastic bottles, and transported on dry ice to the laboratory within 24 h. The samples were dried using a freeze dryer (ALPHA1-2, CHRIST, Germany) and stored at -80 °C for further analysis.

For each sample, about 0.3 g dry activated sludge was collected in triplicate according to a method previously described (Zhou et al. 1996) for DNA extraction. DNA was extracted from each sample (triplicate) using the Power Soil DNA Isolation kit (MoBio Laboratories, Solana Beach, CA, USA) according to the manufacturer's instructions. The extracted DNA from the triplicate samples was combined. Then, they were verified by electrophoresis in 0.8% (w/v) agarose gel. Finally, the quality of the extracted DNA was determined using a Bio Photometer (Eppendorf, Hamburg, Germany).

In order to investigate the abundance of AOB in activated sludge samples, real-time qPCR was employed to quantify the bacterial amoA gene copies in triplicate in different activated samples.

The fluorescent dyes of SYBR Green I were used to react on the IQ5 Thermocycler (RG65HD, Corbett, Australia). The primers amoA-1F/amoA-2R were used for the bacterial amoA PCR amplification (Rotthauwe et al. 1997). A series of 10 times dilutions of plasmid DNA were used to construct the standard curves, covering a known concentration range (1.59 × 10²–1.59 × 10⁸ of bacterial amoA gene copies per microliter).

The 20 μL reaction mixtures for bacterial amoA PCR amplification, included 5 ng DNA template, 1 × SYBR Premix Ex Taq^TM buffer (Takara, Japan), 0.2 μM forward and reverse primers. Thermalcycler protocol was set at: 95 °C for 3 min; 45 cycles of 95 °C for 30 s, 50 °C for 1 min, 72 °C for 20 s, and a final extension at 72 °C for 7 min.

The specificity of the PCR products was checked with melting curve analysis and 2% (w/v) agarose gel electrophoresis. Data analysis was carried out using the Rotor-Gene 6,000 software package. The amplification efficiencies of bacterial amoA gene were 0.95–1.04 (R² = 0.994–0.996). The obtained data were normalized with the mass of the samples to determine the number of bacterial amoA gene copies/g dry activated sludge.

DNA extracted from activated sludge samples of each WWTP were used to construct clone libraries. The primers described above were used to perform the PCR amplification reaction of the bacterial amoA gene. The 25-μL PCR products from triplicate reactions were pooled, the gel bands were purified using the Axygen PCR cleanup purification kit, and cloned using the pGEM-T vector (Promega, Madison, WI, USA). The ligation products were transformed into competent Escherichia coli cells (DH5a, Takara, Japan) and grown overnight on LB agar plates containing 100 μg/mL ampicillin, 40 μg/mL X-Gal and 24 μg/mL IPTG. Positive clones were selected and detected by PCR amplification using vector primers (T7 and SP6). Then, they were sequenced at the Shanghai Majorbio Bio-technology Co., Ltd.

The DNAStar software package was used to remove the vector sequences by bioinformatics analysis. Then we blasted all the obtained sequences with the online GenBank BLAST program according to Altschul et al. (1997), and the highest match sequences were retrieved. Operational taxonomic units (OTUs) were defined as sequences with <3% nucleotide differences. Multiple sequence alignments of the most related environmental sequences and clone sequences were carried out by ClustalX (Thompson et al. 1997). The diversity indices (Shannon-Weiner, H) and nonparametric richness estimator (bias-corrected Chao1 (S_Chao1)) were calculated for each library by the software program DOTUR (Schloss & Handelsman 2005). The library coverage was calculated according to Mullins et al. (1995). A representative sequence with highest quality for each OTU was exported for phylogenetic analysis. Neighbor-joining phylogenetic trees based on Jukes–Cantor distances were built by MEGA 4.0 (Tamura et al. 2007).

All the bacterial amoA gene sequences obtained in this study have been deposited in the GenBank database under the accession numbers JX999156–JX999365.

Correlation analysis (two-tailed Pearson correlation coefficients) between the bacterial amoA gene copies and water quality parameters was performed with the SPSS 13.0 software (SPSS, Chicago, IL). One-way analysis of variance (ANOVA) and post-hoc comparisons were used to test statistical differences of the wastewater parameters and bacterial amoA gene abundance among samples with software SPSS 13.0. Statistical significance was defined at P < 0.05. In order to investigate the relationships between the differences in AOB community structures and the geographic distribution among sampling WWTPs, Mantel test was carried out with the vegan package in R 2.15.0 environment with permutation number of 999. Canonical correspondence analysis (CCA) was performed using the vegan package in R (version: 2.15.0) to investigate the effects of wastewater parameters on AOB community structure in activated sludge samples of WWTPs (Oksanen et al. 2010).

Duncan's multiple range test method was used to analyze the results of qPCR. As the results show in Table 1, the highest bacterial amoA gene abundance (1.85 × 10⁷ copies/g of dry activated sludge) was detected in the HF sample, whereas sample collected from TY showed the lowest abundance of bacterial amoA gene (6.58 × 10⁵ copies/g of dry activated sludge). Among the different samples, we observed significant differences in the bacterial amoA gene abundance (P < 0.05). To be exact, the abundance of bacterial amoA gene in the HF sample was significantly higher than other samples and the samples collected from TY, XA and ULMQ were significantly lower than the other four samples (P < 0.05).

A total of 210 bacterial amoA sequences of seven clone libraries were identified in the present study, which could be divided into 36 OTUs. The library of each sample included 26–32 sequences (Table 2). OTUs were identified based on <3% nucleotide differences by using the DOTUR software package (Beman & Francis 2006). The number of OTUs in each clone library varied from four to 15. Shannon–Weiner (H) and S_Chao1 richness estimators were calculated with the OTU data using DOTUR to characterize the diversity of AOB and shown in Table 2. The coverage of the seven clone libraries ranged from 61.5% to 93.8%, demonstrating adequate sequencing depth. According to the number of OTUs and Shannon-Weiner (H) index, the TY sample had the highest bacterial amoA gene diversity, followed by NJ sample, whereas the MAS sample had the lowest bacterial amoA gene diversity.

In Table 3, Pearson's correlation coefficients were employed to estimate that how the abundance of bacterial amoA gene related to water quality parameters. Significant negative correlations were detected between the bacterial amoA gene abundance and the influent concentrations of total nitrogen (TN) and ammonia nitrogen (NH₄⁺-N) in the activated sludge samples (P < 0.05). However, the number of OTUs was not significantly correlated to any water quality parameters. Furthermore, the NH₄⁺-N removal rate also showed insignificant correlations with the abundance and diversity of bacterial amoA gene.

The phylogenetic tree of bacterial amoA gene sequences is shown in Figure S1 (available with the online version of this paper). All the sequences collected from the seven samples could be assigned into three lineages: Nitrosospira, N. oligotropha and N. europaeal/Nc. mobilis lineages. The relative abundance of different bacterial amoA gene groups in the seven activated sludge samples are shown in Figure 1. In all the samples, N. oligotropha lineage was the superior group, which covered 43.33–96.88% of bacterial amoA gene sequences in each clone library. All sequences in WX and HF samples were assigned to the N. oligotropha lineage, except for one sequence affiliated with N. europaea and Nc. mobilis lineages. In the MAS sample and the XA sample, only a few sequences belonging to Nitrosospira lineage were detected except for the N. oligotropha lineage, with the relative percentages of 3.12% and 10.34%, respectively. The sequences recovered from NJ, TY and ULMQ sample fell into all three lineages, and the N. oligotropha lineage had the same relative percentages (43.33%) as the Nitrosospira lineage in the sequences isolated from NJ.

A Mantel test was employed to investigate the relationships between AOB community structure and geographic distribution of the WWTPs by Euclidean distance. According to the results, geographic distribution showed no significantly correlation with the variations of bacterial community structure (Spearman's rank correlation coefficients, R = 0.2137, P = 0.108). Thus, we speculated that the community structure of AOB in activated sludge of WWTPs might be affected by other water quality parameters.

In order to investigate which water quality indicators had the greatest impact on community structure, CCA was carried out (Figure 2). The results indicate that the influent concentration of TN and NH₄⁺-N were the dominated effects of factors on the AOB communities in the activated sludge samples of this study, followed by the influent concentration of total phosphorus (TP) and the removal rate of NH₄⁺-N. The influent concentration of chemical oxygen demand (COD_Cr) had the least effect on the AOB community structure.

In the present study, the abundance of bacterial amoA gene showed significant differences in distinct geographical areas, and the differences mainly exhibited that the copy numbers of bacterial amoA gene of activated sludge samples in the eastern of China (HF, MAS, WX, NJ) were significantly higher than those in the central and western regions (TY, XA, ULMQ) (P < 0.05). Although there were no direct conclusions in previous studies that could be compared to the present study, Zhao et al. (2014a) indicated that both the source microorganisms and the resident microorganisms in the system would be affected by the local environment and climate, which explains why the geographical distance had an impact on the abundance of AOB. Niu et al. (2016) also found that nitrogen removal associated bacteria in high altitude WWTPs showed a great sensitivity to water quality parameters. In this study, the bacterial amoA gene abundance in the samples collected from high-altitude plants (TY, XA, ULMQ) was significantly lower than those collected from low-altitude plants (HF, MAS, WX, NJ) (P < 0.05), which is consistent with the result of Niu et al. (2016).

A significantly negative correlation was found between the influent concentrations of TN and NH₄⁺-N and the bacterial amoA gene abundance (P < 0.05). Although little information on the relationship between water quality parameters and the abundance of AOB has been previously reported in activated sludge, the significant correlation of AOB abundance with environmental variables, as dissolved oxygen (DO) (Fitzgerald et al. 2015), temperature (T) (Zeng et al. 2014) and nitrate nitrogen (NO₃⁻-N) (Zhao et al. 2014a), have been discussed for other habitats. For instance, previous studies showed that ammonia concentrations positively influence the bacterial amoA gene abundance in soil ecosystem (Zhao et al. 2015; Wang et al. 2009). However, a negative correlation was found between the concentrations of TN and NH₄⁺-N and the bacterial amoA gene abundance in a eutrophic lake (Zhao et al. 2014b) and the sediments of Taihu (Zeng et al. 2012). Ammonia maintain probably positive effects on the abundance and diversity of AOB as for the substrate for ammonia oxidizers (Adair & Schwartz 2008), but further water quality parameters not considered in this study could also play different roles, such as temperature (T), which could cause both inhibitory and stimulative effects on the abundance of bacterial amoA gene (Zeng et al. 2014).

According to the results of Mantel test, geographic distribution had no significant effects on AOB communities. However, in this study, we observed that the samples collected from different WWTPs had different community structure of AOB (Figure 1). There might be a number of factors that can cause this result, such as treatment process and treatment scale. Dytczak et al. (2008) found that the different reactors are dominated by different AOB communities. In this study, MAS, HF and XA were using oxidation ditch (OD) process, and they had similar AOB community structure, which supported the results mentioned. Rowan et al. (2003) analyzed the AOB community structure in different scale WWTPs and found that there were similarities in AOB structure in the two wastewater treatment systems, but the AOB community structure of the large-scale WWTP was more stable and more diverse. In the present study, TY had the largest scale and the highest diversity of AOB community, indicating that the scale of the wastewater treatment plant might have a certain impact on AOB community structure.

Except for the treatment process and treatment scale, the influent water quality of WWTPs could also affect the AOB community structure. Sangchul et al. (2005) showed that the community structure of AOB in municipal WWTPs were more diverse than that in industrial WWTPs. Among all WWTPs investigated in this study, only HF and XA involved industrial wastewater (ca. 30–40% of total). And the samples of HF and XA did show lower diversity than others (except for WX and MAS), which is consistent with Sangchul et al. (2005). Otawa et al. (2006) and Limpiyakorn et al. (2011) found that the concentration of ammonia nitrogen (NH₄⁺-N) in the original wastewater had a certain effect on AOB communities in the activated sludge, in which the Nitrosomonas europaea-eutropha cluster was the main cluster at high NH₄⁺-N concentrations, whereas the Nitrosomonas ureae oligotropha-marina cluster was the main cluster at low NH₄⁺-N concentrations. In this study, there was no significant difference in the concentration of ammonia nitrogen in the considered WWTPs, but a significant difference of the community structure of AOB still existed among WWTPs as determined by Mantel test (P < 0.05). We hypothesized that the water quality parameters influencing the community structure of the AOB in the WWTPs were multifaceted rather than singular, as AOB communities may have different responses to different water quality parameters and the superposition of these water quality parameters may result in the final community structure of the AOB. The results of the CCA suggested that this was the case, and the influent concentrations of TN and NH₄⁺-N were the most significant influencing factors, followed by TP and NH₄⁺-N removal rate.

The results shown in Figure 1 revealed that N. oligotropha lineage dominated the AOB communities of each library. The lineage had previously been reported from many ecosystems such as WWTPs, fresh water, sediments, soil and others (Regan et al. 2002; Hallin et al. 2005; Limpiyakorn et al. 2005; LaPara & Ghosh 2006; Qin et al. 2008; Zhao et al. 2013). Zhao et al. (2013) also found that the three lineages existed in the sediments of a eutrophic lake, and N. oligotropha lineage was the dominated lineage. A previous study showed that N. oligotropha lineage has the advantage of fast growth, and a low saturation constant (K_s) for free ammonia, indicating that it is suitable for survival in the environment with low ammonia nitrogen (Lydmark et al. 2007). In this study, the influent NH₄⁺-N of the seven WWTPs were low, except for TY, which explain why N. oligotropha lineage was the dominant population of AOB. However, in the industrial WWTPs that have a high concentration of NH₄⁺-N, N. europaeal and Nc. mobilis lineages may be the dominant populations (Koops & Andreas 2001).

The data obtained in this study demonstrated that geographic distribution did have an effect on the distribution of AOB in WWTPs. However, this was rarely discussed in previous studies. Although geographic distribution only affects the abundance of AOB, the effect can-not be ignored. At the same time, the results further clarified which water quality parameters of the WWTP have the most significant effects on the abundance and community structure of AOB.

The information obtained in this study would be useful to elucidate the abundance and community structure of AOB in activated sludge. Further investigations are needed to explore the factors affecting the ecological distribution of AOB in activated sludge samples. This will need to include more WWTPs in further areas of China, as well as further water quality parameters such as dissolved oxygen (DO) or nitrate nitrogen (NO₃⁻-N).

In this study, we investigated the abundance and community structure of AOB in the activated sludge of seven WWTPs in seven cities throughout China, and the influence of various water quality parameters on the abundance and community structure of AOB were examined. It was observed that the geographic distribution of WWTPs had a significant effect on the abundance of the bacterial amoA gene. The community structure of AOB was not significantly affected by geographic distribution, but most probably by the treatment process, treatment scale and influent wastewater quality, especially by the concentrations of TN and NH₄⁺-N. N. oligotropha lineage was the superior AOB population in the considered wastewater treatment system in this study.