Abstract

Operational conditions are often manipulated to improve the nitrogen removal performance of wastewater treatment, yet the impacts of operational conditions on microbial communities were still not well understood. There is a pressing need to understand the microbial mechanisms that link operation manipulation and nitrogen removal performance. In this study, high-throughput analysis of 16S rDNA and quantitative polymerase chain reaction of functional genes were used to identify the microbial response to operational manipulations. The results showed that alteration of operational parameters could change the bacterial communities at the genera level and denitrification guild gradually dominated in the activated sludge bacterial communities. Heterotrophic Hyphomicrobium and Chromatiaceae drove the kinetic of dominant genera and denitrification guild. Carbon source supplement was the most efficient strategy for improving nitrogen removal, and greatly increased the abundance of denitrifiers and denitrification genes. However, carbon source supplement inhibited expression activities of denitrification genes, as well as the proliferation of autotrophic denitrifiers, and it was supposed to be unfavorable in terms of cost over the long term. The result should bring new inspiration for improving the effect of WWTP performance through the manipulation of operational parameters.

WATER IMPACT

A2/O is a widely used sewage treatment process in wastewater treatment plants and appropriate operational manipulation could improve the performance. However, the lack of understanding of microbial community activities hampered the application of operation manipulation. This study presented the bacterial communities' responses to the three most used operation manipulations at the community level, functional guild level, and gene activity level, respectively. Carbons resource supplement was demonstrated to be the most efficient manipulation, but it would inhibit the expression activities of denitrification genes and proliferation of autotrophic denitrifiers. To fulfill the increasing demand of energy saving and low-cost sewage treatment process, this finding has practical implication for it can give us cues to manipulate WTTPs based on the health and activity of activated sludge (AS).

INTRODUCTION

Regulations regarding the effluents of wastewater treatment plants (WWTPs) are becoming increasingly stringent in order to protect receiving water bodies. To achieve high-efficiency nutrient removal, operational parameters, such as hydraulic and solids retention time and the internal and external recycling of liquid, are often manipulated to improve the nutrient-removal performance of WWTPs (Brown et al. 2011; Lu et al. 2014; Saunders et al. 2016; Zhang et al. 2016). The microbial communities in the AS may be influenced by the manipulation process, resulting in the variable performance of WWTP (Brown et al. 2011; Ju & Zhang 2015; Zhang et al. 2016). Therefore, knowledge of the reactor's physicochemical parameters alone is not sufficient to provide information that will enable an improved operational manipulation. Studying the responses of the microbial communities to operation manipulations will contribute to WWTPs' operation optimization and avoid problems related to pollutant removal. The process control method selected in this study is based on the actual process operation of the sewage treatment plant. By optimizing the process control operation, the nutrient in the water can be fully utilized under the condition that the design parameters are determined. The actual denitrification effect of each unit of sewage treatment plant and the change of microbial community structure under different operation strategies were studied, which provided support for the actual operation of A2/O wastewater treatment plant.

Recently, high-throughput sequencing has revealed many of the ecological characteristics of microbial communities in AS, including the existence of similar, so-called core communities that are responsible for maintaining the satisfactory and steady performance of WWTPs regardless of plant location, influent type, and sampling time (Saunders et al. 2016). Despite a certain understanding of the microbial communities and biological processes in wastewater treatment, most of the data concerning the response of sludge bacterial communities to operation manipulation were collected from strictly controlled, laboratory-scale bioreactors using synthetic wastewater as the influent (Brown et al. 2011; Ju & Zhang 2015; Zhang et al. 2016). Consequently, much remains to be learned about full-scale AS-based WWTPs, in which the influent and operational conditions are more complicated (Guo et al. 2017). In addition, while several studies have demonstrated the primary roles played by various functional groups of microbiota in nutrient removal, the active members which contributed primarily to nutrient removal are poorly understood. A better understanding of mechanism linking operational manipulation variation and response of WWTP performance is necessary to improve WWTPs' performance.

Bacterial functional groups of nitrifiers, denitrifiers, and anammox cooperatively drive nitrogen (N) removal in WWTPs. Nitrification and denitrification rates had been demonstrated to correlate quantitatively with the abundances of the nitrification genes of amoA and nxrA and denitrification genes of narG, napA, nirS, nirK, norB, and nosZ (Wang et al. 2014a). However, gene-transcript (mRNA) is a more accurate indicator of gene function and response to environmental variables than gene abundance. For example, amoA transcript abundance varied from 0.02% to 40% across different environmental conditions (Tolar et al. 2016). Therefore, analysis of the expression activities of functional genes might provide a method to assess the key mechanisms underlying the response of microbial communities to operation manipulation.

In this study, three kinds of operation manipulation, i.e. influent distribution, multipoint nitrification liquor recycling, and carbon source supplementation, were examined for their abilities to optimize the nitrogen removal performance of a modified A2/O. The specific goals of this study were to discern the responses of active nitrifiers and denitrifiers, as well as functional genes associated with nitrogen removal to particular operational manipulation. The results will contribute to the evaluation and optimization of WWTPs' manipulation in order to comply with stricter effluent criteria and sustainable operation.

MATERIALS AND METHODS

Pilot modified A2/O bioreactor

A pilot modified A2/O bioreactor (6 × 2.1 × 2.4 m (L × W × H)) equipped with an automatic control system was built within the Yixing municipal WWTP (YXM). The total hydraulic retention time (HRT) of the bioreactor was 17.3 h and its sludge retention time (SRT) was 22 d (Table 1). A detailed configuration of the reactor, which was designed in parallel with the modified A2/O of YXM, was described elsewhere (Figure S1, available with the online version of this paper) (Chen et al. 2016). To simulate the full-scale WWTP, the raw wastewater was the same as that of YXM, and the inoculum was derived from the excess sludge of YXM.

Table 1

The size and hydraulic retention time (HRT) of each reactor unit in a pilot A2/O bioreactor

Reactor unit HRT (h) Size (L × W × H, in meters) 
Pre-anoxic (PRAN) 0.82 × 0.7 × 2.1 
Anaerobic (ANA) 0.82 × 0.7 × 2.1 
Anoxic (AN) 5.5 4.48 × 0.7 × 2.1 
Aerobic (AE) 5.70 × 0.7 × 2.1 
Post-anoxic (POAN) 2.5 2.00 × 0.7 × 2.1 
Post-aerobic (POA) 0.3 0.6 × 0.7 × 2.1 
Reactor unit HRT (h) Size (L × W × H, in meters) 
Pre-anoxic (PRAN) 0.82 × 0.7 × 2.1 
Anaerobic (ANA) 0.82 × 0.7 × 2.1 
Anoxic (AN) 5.5 4.48 × 0.7 × 2.1 
Aerobic (AE) 5.70 × 0.7 × 2.1 
Post-anoxic (POAN) 2.5 2.00 × 0.7 × 2.1 
Post-aerobic (POA) 0.3 0.6 × 0.7 × 2.1 

Operation manipulation

After inoculation, the reactor was started to achieve a stable stage that met the effluent criteria of first class A (COD: 50 mg/L, total nitrogen (TN): 15 mg/L, ammonia nitrogen: 5 (8) mg/L, and TP: 0.5 mg/L) (stage I). Thereafter, the reactor was manipulated as follows: (1) influent distribution to the pre-anoxic (PRAN) and anaerobic (ANA) units at ratio of 2:8 (stage II); (2) influent distribution to the PRAN, ANA, and post-anoxic (POAN) units at ratio of 2:5:3 (stage III); (3) two-point nitrification liquor internal return to the starting and mid points of the anoxic (AN) unit (stage IV); (4) POAN unit supplementation with 20 mg acetate/L (stage V); (5) POAN and AN unit supplementation with 20 mg acetate/L (stage VI). The detailed operational manipulation of each stage are summarized in Table S1 (available with the online version of this paper). Each manipulation lasted 10 days.

Physicochemical parameters analysis

Samples were collected from the AN unit at the end of each stage. Dissolved oxygen (DO), pH, and oxidation-reduction potential (ORP) were measured directly on-site. The amount of mixed liquor suspended solids (MLSS) was also analyzed daily. Liquid samples were filtered through Millipore filter units (0.45-μm pore size, Millipore, USA) and analyzed in the laboratory for TN, NH4+-N, NO3-N, chemical oxygen demand (COD), and total phosphate according to the standard methods (National Environmental Bureau 2002).

Nucleic acid extraction and cDNA synthesis

Genomic DNA was extracted from the AS using the FastDNA® spin kit for soil (MP Biomedicals, CA, USA) according to the manufacturer's instructions. Total RNA was extracted from the AS using the FastRNA Pro Soil-Direct kit (MP Biomedicals, CA, USA) and the FastPrep instrument according to the manufacturer's instructions. Nucleic acid concentration and purity were determined micro-spectrophotometrically (NanoDrop® ND-1000, NanoDrop Technologies, Wilmington, DE, USA).

The extracted RNA was immediately reverse transcribed into cDNA using the PrimeScriptTM RT reagent kit (TaKaRa, Dalian, China) according to the manufacturer's protocol. DNA and cDNA were stored at −80 °C until further analysis.

Sequencing of 16S rRNA genes and phylogenetic classification

The DNA samples were sent out to Majorbio (Shanghai, China) for Illumina high-throughput sequencing of the 16S rRNA genes using the MiSeq platform (Illumina, USA) and standard protocols. The generated raw sequences of all AS samples were assigned using the Ribosomal Database Project (RDP) (http://rdp.cme.msu.edu/) to roughly trim off the adapters and barcodes.

The sequences were clustered into operational taxonomic units (OTUs) by setting a distance limit of 0.03 (equivalent to 97% similarity) using the MOTHUR program. BLAST and MEGAN were used to obtain species-level information based on the assignments.

Net growth rate of nitrifiers and denitrifiers

To evaluate the contribution of active nitrifiers and denitrifiers to the function of the AS communities, the net growth rate of each group was calculated using amplicon data. Growth and decay can be described as first-order processes, assuming that AS is at steady state, with no net change in the number of cells in the AS biomass (Nx, AS). Then the net growth rate could be calculated using Equation (1) (Saunders et al. 2016):  
formula
(1)
where Nx, AS is the number of cells of organism x in the AS; K is the rate constant (d−1); and nx, SP is the number of cells of organism x removed (excess sludge) per unit time (d−1). In this study, the average SRT was 22 days, thus, the growth rates of the bacteria, 1/22 per day, were used in the calculation.

Gene expression quantitation using RT-qPCR

The abundance and expression of bacterial 16S rDNA and the genes involved in N cycling (amoA, nxrA, narG, napA, nirS, nirK, norB, and nosZ) were determined using SYBR Green I qPCR. The primer pairs and qPCR programs are summarized in Table S2 (available with the online version of this paper). Plasmids containing these genes were prepared as described in a previous study (Zhi & Ji 2014). Each reaction was performed in triplicate.

Statistical analysis

Pearson correlation analyses were performed using SPSS 23. A p value <0.05 was considered to indicate statistical significance.

RESULTS AND DISCUSSION

Performance of the pilot modified A2/O under different operation conditions

After its start-up, the reactor was operated for 10 more days (stage I), during which time the average removal efficiency of COD, TN, and NH4+-N reached 73.50 ± 13.02%, 64.40 ± 14.56%, and 98.04 ± 1.06%, respectively. The maximum concentration of COD, TN, NH4+-N and NO3-N in the effluent was 49.6 mg L−1, 9.62 mg L−1, 0.42 mg L−1, and 8.59 mg L−1, respectively (Figure 1 and Figure S2, available with the online version of this paper). The results confirmed the successful start-up of the reactor and that the quality of the effluent met the WWTP effluent criteria of first class A. Operation manipulations were then conducted to assess their effects on the performance of WWTP. Among the different operation stages, the removal efficiency and effluent concentration of COD did not differ significantly, indicating the stability of COD removal (Figure 1(a)). However, the removal of TN, NH4+-N, and NO3-N differed under the different operating conditions.

Figure 1

The profiles of chemical oxygen demand (COD) (a), total nitrogen (TN) (b), NH4+-N (c), and NO3-N (d) in the influent and effluent during reactor stages I–VI.

Figure 1

The profiles of chemical oxygen demand (COD) (a), total nitrogen (TN) (b), NH4+-N (c), and NO3-N (d) in the influent and effluent during reactor stages I–VI.

The influent distribution ratio influences the HRT, pollutant distribution, and the carbon source in each unit (Ge et al. 2012). The TN removal efficiency in stage II and III did not differ significantly from that of the control (stage I) (P > 0.05). In stage II, although TN in the effluent was significantly higher than in stage I (P < 0.01), the higher TN content of the raw wastewater could be the reason (Figure 1(b) and Figure S2c). In stage III, the average effluent concentration of COD, TN, and NH4+-N was significantly higher than in stage I (P < 0.05) (Figure S2), partly because the distribution of influent to POAN unit was unfavorable to WWTP performance. Although the distribution of influent could provide carbon compounds to the POAN unit, an HRT of raw wastewater that was too short in the POAN unit resulted in the leakage of NH4+-N into the effluent. In addition, a reduced HRT may lead to a much lower anoxic contact time for denitrification and increase the nitrate load rate in the AN unit (Zhang et al. 2015). The significantly lower NO3-N content of the effluent in stage III (P < 0.01) could be attributed to the deficient removal of NH4+-N by nitrification (Figure S2f).

The nitrification liquor in an aerobic tank contains NO3-N, and nitrification liquor return from the aerobic to the anoxic tank is the key step in the denitrification occurring in an A2/O (Ge et al. 2012). The concentration and removal efficiency of COD, NH4+-N, and NO3N in stage IV had no significant difference with stage I (P > 0.05, Figure S2). The TN removal efficiency decreased from 64.4 ± 14.6% to 52.98 ± 19.43% (Figure 1(b) and Figure S2d); however, no significant difference was detected between stage I and stage IV. The HRT might partly explain the unsatisfied performance of TN removal in stage IV, for the calculated HRT in the AN unit was only 5.25 h compared to the 7-h HRT of single-point return.

The POAN unit is the last step in the denitrification occurring in a modified A2/O, and the denitrification in POAN was limited by carbon source scarcity. The strategy of distributing influent into the POAN unit in stage III was unacceptable due to the high concentration of NH4+-N in the effluent (Figure 1). In stage V, in which 20 mg acetate/L was added to the POAN unit, NO3-N decreased significantly in the POAN outlet, from 10.38 mg/L (POAN inlet) to 7.08 mg/L (POAN outlet). Since 1 mg NO3-N results in the consumption of 7.6 mg COD (Mohseni-Bandpi et al. 2013), the removal 3.3 mg NO3N L−1 from the POAN unit corresponds to a stoichiometric demand of 25.08 mg COD L−1. Supplementation with 20 mg acetate L−1 only amounted to a COD of 21.34 mg L−1, implying that the usage efficiency of acetate during the denitrification process was as high as 100% and had no influence on the effluent COD level. However, the TN removal efficiency in stage V had no significant difference from the control (Figure S2d). When both the AN and POAN units were supplemented with 20 mg acetate/L (stage VI), the TN removal efficiency could be as high as 88.08% ± 6.87%, together with a significant decrease in TN in the effluent, from 7.68 ± 1.47 mg/L to 3.34 ± 1.76 mg/L (P < 0.01). Meanwhile, NO3N in stage VI was significantly lower than that in stage I (P < 0.05, Figure S2g).

Among the examined operational conditions, carbon source supplementation (stages V and VI) yielded the most effective improvement in the nitrogen removal performance of the WWTP (Figure 1 and Figure S2), indicating that a lack of carbon source was the important constraint in denitrification and nitrogen removal now. For heterotrophic denitrifiers, carbon compounds serve as both an energy source for bacterial growth and an electron donor for sequential reduction of NO3 to NO2, nitric oxide (NO), N2O and/or N2 (Sun et al. 2010). Although carbon addition incurs additional costs, it can promptly increase the denitrification capacity and thus lower the amount of TN in the effluent.

Effect of operation manipulation on microbial communities

To reveal response of bacterial communities and functional groups in AS to operation manipulation, high-throughput sequencing was employed to analyze turnover of bacterial communities. Of the 45 phyla identified in the AS, the abundances of Proteobacteria (34.96–50.63%), Chloroflexi (8.95–22.97%), Acidobacteria (3.26–12.35%), Bacteriodetes (6.61–11.09%), Ignavibacteriae (1.98–4.59%), Chlorobi (3.60–9.37%), Saccharibacteria (2.57–5.05%), Gemmatimonadetes (1.91–3.80%), and Actinobacteria (1.57–3.52%) collectively accounted for 89.9–92.5% of the total effective reads in each stage (Figure 2). The results indicated that variation of different operation conditions did not significantly alter the bacterial core composition at the phylum level, which was consistent with previous studies (Xia et al. 2010; Saunders et al. 2016). Although rare ones could play important roles in helping a bacteria community to adapt to the environment changing (Hua et al. 2015), the core community of abundance bacteria in AS guaranteed the performance of WWTPs. Furthermore, previous studies had shown that a large abundant core community was also present in AS systems in Denmark, the USA and China (Saunders et al. 2016).

Figure 2

The profile of bacterial abundance on phyla level in the different operational manipulation stages.

Figure 2

The profile of bacterial abundance on phyla level in the different operational manipulation stages.

The abundances of 19 dominated genera accounted for >50% of the total effective reads in all samples (Figure 3). Bacteria belonging to Hyphomicrobium, Sarccharibacteria, Chromatiaceae, and Comamonadaceae in stages IV, V and VI outnumbered their counterparts in stages I, II, and III, whilst the abundance of Nitrosomonadaceae decreased continuously. While the above results suggested that the composition of bacterial communities was rather stable, evidence of intense variation of bacterial communities among different operation stages was observed at genera level (Figure 3). These data also indicated that AS microbiomes are plastic in response to operation manipulation when low taxonomic level was used.

Figure 3

Sanky diagram of the most abundant bacterial genera in the different operational manipulation stages.

Figure 3

Sanky diagram of the most abundant bacterial genera in the different operational manipulation stages.

Effect of operation manipulation on dynamics and growth of nitrifiers

The role of nitrifiers in nitrogen removal was very important for most of the nitrogen in WWTP influent was in the form of NH4+-N, which must be converted to NO3 by nitrification for denitrification. In this study, five OTUs related to nitrification were identified (Figure 4). Nitrosomonadaceae uncultured dominated the ammonia oxidizers guild, whereas nitrite oxidizers were dominated by Nitrospira, indicating that these two groups essentially controlled the performance of N removal. Some of Nitrospira was recently found to be complete nitrifier, which were species able to conduct the whole process by which ammonia is converted first to nitrite and then to nitrate (Kits et al. 2017). The dominance of Nitrospira in the nitrifier guild can therefore facilitate NH4+-N removal.

Figure 4

Kinetics of abundance (a) and growth rates (b) of nitrifiers in the different operational manipulation stages.

Figure 4

Kinetics of abundance (a) and growth rates (b) of nitrifiers in the different operational manipulation stages.

Our results demonstrated that different nitrifiers responded differently to operation manipulation, and stages V and VI witnessed high growth rate of nitrifiers (Figure 4(b)). However, no significant relationship between the abundance and growth rate of these nitrifiers and the NH4+-N transformation rate (P > 0.05) was observed in our study (Table 2). Previous studies attributed this kind of observation to gene activity or the cell-specific activity of the ecotype (Santoro et al. 2010; Tolar et al. 2016). Furthermore, the NH4+-N transformation rate determined in this study was an integrated value that included ammonia oxidization, anammox, and cellular intake, and the integrated process might decouple the significantly direct correlation of ammonia oxidizers with NH4+-N transformation rate.

Table 2

Pearson correlation of the growth rate of denitrifiers and nitrogen transformation

  Bradyrhizobiaceae Defluviicoccus Hyphomicrobiaceae Thiothrix Rhodocyclaceae Sulfuritalea 
NH4+-N −0.428 −0.416 −0.018 0.962** −0.630 0.928* 
NO3-N −0.991** −0.902* 0.935* 0.371 −0.926* 0.516 
TN −0.471 −0.346 0.155 0.595 −0.538 0.598 
  Bradyrhizobiaceae Defluviicoccus Hyphomicrobiaceae Thiothrix Rhodocyclaceae Sulfuritalea 
NH4+-N −0.428 −0.416 −0.018 0.962** −0.630 0.928* 
NO3-N −0.991** −0.902* 0.935* 0.371 −0.926* 0.516 
TN −0.471 −0.346 0.155 0.595 −0.538 0.598 

TN means total nitrogen. *Indicates a significant difference at P < 0.05 and **indicates a significant difference at P < 0.01.

The highest growth rate was that of Nitrosomonas: 0.11 d−1 on average and 0.33 d−1 during stage VI (Figure 4(b)). These rates were comparable to the previously reported growth rates of ammonia oxidizers (Tolar et al. 2016). Nitrospira was a commonly reported member of the core community of AS (Xia et al. 2010; Saunders et al. 2016), and its high growth rate (0.06–0.15 d−1) indicated its high level of activity and its adaptation to all of the operational conditions tested. It was surprising to find high growth rate of Nitrosomonas and Nitrospira in stage V and stage VI for the higher growth rate of nitrifiers presumably reflected the primary role of them for more efficient oxidation of ammonia (Figure 4). Recently, direct nitrogen removal by nitrifiers alone also attracted more and more attention (Caranto & Lancaster 2017). Ammonia oxidizing bacteria and archaea contributed to nitrogen removal by producing substantial amounts of nitric oxide (NO) and nitrous oxide (N2O) (Caranto & Lancaster 2017). Although a community analysis failed to reveal a significant activity of ammonia oxidizers that correlated with TN removal (Table 2), the N-removal processes driven by ammonia oxidizers could reduce the time needed for nitrogen removal in a bioreactor, compared to the time required by a sequential chain of nitrification–denitrification, and thus accelerate the TN-removal rate.

Effect of operation manipulation on dynamics and growth of denitrifiers

Twenty-eight OTUs of potential denitrifiers were detected and their abundances increased steadily, from 9.5% at stage I to 19.6% at stage VI (Figure 5(a)). The increase of denitrifier abundance was mainly explained by the dominant Hyphomicrobium and Chromatiaceae (Figure 5), which also dominated in core genera under all operation conditions (Figure 3), supporting the improving denitrification performance in stages V and VI. Hyphomicrobium harbour the denitrifying functional genes nirS and nosZ (Wang et al. 2014b), which could accumulate in AS with increasing SRTs and NO3-N concentrations (Ju & Zhang 2015).

Figure 5

Kinetics of abundance (a) and growth rates (b) of denitrifiers in the different operational manipulation stages.

Figure 5

Kinetics of abundance (a) and growth rates (b) of denitrifiers in the different operational manipulation stages.

It is well known that environmental conditions such as concentration of DO (Tiedje 1988), organic carbon source and nitrate availability could influence growth of denitrifiers (Bernhardt & Likens 2002). The growth rates of the 28 OTUs of denitrifiers responded differently to each operational manipulation (Figure 5(b)). In stages VI, Bacillus, Rhizomicrobium, Rhodobacter, and Rhodobacteraceae_Unclassified were found in high growth (Figure 5(b)). Generally, the growth of Bradyrhizobiaceae_Unclassified, Defluviicoccus, and Rhodobacteraceae_Unclassified correlated significantly with the removal of NO3-N (Table 2), indicating the importance of these bacterial groups in denitrification.

We also found that autotrophic denitrifiers of Hydrogenophaga and Thioalkalispira were inhibited in stage VI. This result indicated that the niche altered by carbon addition was more favorable for heterotrophic denitrifiers (Nielsen et al. 2010; Calvano et al. 2014). Despite the greater energy savings of autotrophic denitrifiers, autotrophic denitrifiers could be removed gradually because of their low growth rate, especially in the case of sufficient organic carbon supplementation as shown in our study (Hao et al. 2013; He et al. 2015). Thus, the results suggested that carbon supplementation was a disadvantage strategy over the long run. The increasing abundance of heterotrophic denitrifiers in AS may trigger a positive feedback loop for the demand of a carbon source and thereby elevate the cost of wastewater treatment.

Effect of operation manipulation on abundance and transcriptional activity of genes associated with nitrogen transformation

The amoA and nxrA genes represented two steps in nitrification, i.e. ammonia and nitrite oxidization. The abundance of amoA and nxrA increased under all manipulation; however, the activity of amoA and nxrA responded differently to manipulation (Figure 6). The activity of nxrA was significantly inhibited under all manipulation conditions (P < 0.01), whilst 40 mg L−1 acetate addition increased expression of amoA significantly (P < 0.01), confirming the selective effect of manipulation on nitrifier guild. The positive influence of carbon supplementation on amoA gene abundance and activity might contribute to the coupled nitrification–denitrification process as evidenced by elevated nitrogen removal (Figure 1). Some ecotype of ammonia oxidizers could adapt to relatively low DO below 0.5 mg L−1, which could contribute to the transformation of NH4+ to NO2 at low DO (Fitzgerald et al. 2015).

Figure 6

Gene abundance (a) and expression (b) activity of amoA and nxrA in the different operational manipulation stages.

Figure 6

Gene abundance (a) and expression (b) activity of amoA and nxrA in the different operational manipulation stages.

For denitrifier guild, the response of gene abundance was similar among all of the denitrification genes, including narG, napA, nirS, nirK, norB, and nosZ. The abundance of these genes increased significantly under the tested operation manipulation especially for the carbon resource addition in stage V and stage VI (Figure 7). This result corroborated the results from the taxa analysis by 16S rDNA, which showed that carbon resource addition could increase the relative abundance of heterotrophic denitrifiers (Figure 5), and compete with autotrophic denitrifiers for nitrate as electron acceptor, resulting in the inhibition to the growth of autotrophic denitrifiers. At the same time, our results demonstrated that the expression of denitrification genes also showed an inhibitory effect and the inhibition differs with different species. It may be that the addition of carbon source affects the growth of different denitrifier species and results in the change of gene expression activity. However, expression activities of these denitrifying genes showed more various responses to operation manipulation conditions. First of all, both of the activities of narG and napA were significantly inhibited, indicating the specific activities of narG and napA decreased although their absolute abundance increased. For nirS and nirK, the genes associated with nitrite reduction to NO, external carbon addition significantly inhibited their activities (P < 0.01), whilst nitrification liquor recycling increase the expression of nirS significantly (P < 0.01). The activity of norB determined the ability of transformation from NO to N2O, and the result showed that activity of norB increased significantly in stage IV and stage V. In contrast to the increase of gene abundance, the activity of nosZ was significantly inhibited in stage V and stage VI (P < 0.01).

Figure 7

Gene abundance and expression activity of denitrification genes in the different operational manipulation stages.

Figure 7

Gene abundance and expression activity of denitrification genes in the different operational manipulation stages.

The increase in gene abundance confirmed that carbon addition could increase the abundance of denitrifiers; however, at the level of gene expression activity, the unfavorable external carbon addition to denitrification in the long run was once again demonstrated by decreasing expression activity of narG, napA, nirS, nirK, and nosZ. The variation of environmental factors under different manipulation conditions, such as DO, carbon source, nutrient level and stoichiometric, etc., could result in difference of gene expression and affect the growth of different denitrifiers and shift of microbial composition (Figures 3 and 5) (Yu & Zhang 2011). Denitrifiers could have different combinations of genes involved in the denitrification pathway (Jones et al. 2008). Some denitrifiers possess complete denitrification genes and can potentially perform the complete denitrification pathway, while others lack some of the denitrification genes, such as nitrous oxide reductase gene and only produce N2O (Philippot et al. 2011). Denitrification is an energy production process using NOx as the electron acceptor instead of oxygen. However, the expression pattern of denitrification genes implied a reduced cell-specific denitrification ability of the AS communities as well as higher potential of N2O production whilst the biomass of denitrifiers increased.

CONCLUSION

The results showed that alteration of operational parameters could rapidly change the composition of bacterial communities at low taxonomic level and the gene expression activities. Carbon source supplementation was the most effective manipulation to improve nitrogen removal in the short-term. However, supplementing the carbon source increased the abundance and growth of heterotrophic denitrifiers and inhibited the growth of autotrophic denitrifiers, as well as denitrification genes' expression. The correlation of WWTP performance with operational manipulations at communities and molecular levels opens up new possibilities to optimize reactor operation and design based on engineered communities. For example, combined with different types of A2/O process in the actual operating conditions, the optimal allocation of carbon source can be realized by properly adjusting the water inlet of the system, the removal of nutrient salt can be optimized by adjusting the backflow inside and outside the system, and the carbon source can be used more efficiently by adjusting the position of carbon source input and addition, so as to reduce the demand for external carbon source as far as possible and achieve greater nitrogen removal effect. Thus, the operation manipulation should be based on the health of microbial communities in AS which the operator should bear in mind.

CONFLICTS OF INTEREST

There are no conflicts to declare.

ACKNOWLEDGEMENT

This work was supported by grants from the National Key Research and Development of China (No. 2016YFC0502801).

REFERENCES

REFERENCES
Calvano
C. D.
,
Italiano
F.
,
Catucci
L.
,
Agostiano
A.
,
Cataldi
T. R. I.
,
Palmisano
F.
&
Trotta
M.
2014
The lipidome of the photosynthetic bacterium Rhodobacter sphaeroides R26 is affected by cobalt and chromate ions stress
.
Biometals
27
(
1
),
65
73
.
Caranto
J. D.
&
Lancaster
K. M.
2017
Nitric oxide is an obligate bacterial nitrification intermediate produced by hydroxylamine oxidoreductase
.
Proceedings of the National Academy of Sciences of the United States of America
114
(
31
),
8217
8222
.
Fitzgerald
C. M.
,
Camejo
P.
,
Oshlag
J. Z.
&
Noguera
D. R.
2015
Ammonia-oxidizing microbial communities in reactors with efficient nitrification at low-dissolved oxygen
.
Water Research
70
,
38
51
.
Hua
Z. S.
,
Han
Y. J.
,
Chen
L. X.
,
Liu
J.
,
Hu
M.
,
Li
S. J.
,
Kuang
J. L.
,
Chain
P. S.
,
Huang
L. N.
&
Shu
W. S.
2015
Ecological roles of dominant and rare prokaryotes in acid mine drainage revealed by metagenomics and metatranscriptomics
.
The ISME Journal
9
(
6
),
1280
1294
.
Kits
K. D.
,
Sedlacek
C. J.
,
Lebedeva
E. V.
,
Han
P.
,
Bulaev
A.
,
Pjevac
P.
,
Daebeler
A.
,
Romano
S.
,
Albertsen
M.
,
Stein
L. Y.
,
Daims
H.
&
Wagner
M.
2017
Kinetic analysis of a complete nitrifier reveals an oligotrophic lifestyle
.
Nature
549
(
7671
),
269
272
.
Lu
H.
,
Chandran
K.
&
Stensel
D.
2014
Microbial ecology of denitrification in biological wastewater treatment
.
Water Research
64
(
7
),
237
254
.
Mohseni-Bandpi
A.
,
Elliott
D. J.
&
Zazouli
M. A.
2013
Biological nitrate removal processes from drinking water supply – a review
.
Iran Journal of Environmental Health
11
(
1
),
35
35
.
National Environmental Bureau
2002
Methods for the Analysis of Water and Wastewater
,
4th edn
.
Chinese Environmental Science Press
,
Beijing, China
.
Nielsen
P. H.
,
Mielczarek
A. T.
,
Kragelund
C.
,
Nielsen
J. L.
,
Saunders
A. M.
,
Kong
Y.
,
Hansen
A. A.
&
Vollertsen
J.
2010
A conceptual ecosystem model of microbial communities in enhanced biological phosphorus removal plants
.
Water Research
44
(
17
),
5070
5088
.
Philippot
L.
,
Andert
J.
,
Jones
C. M.
,
Bru
D.
&
Hallin
S.
2011
Importance of denitrifiers lacking the genes encoding the nitrous oxide reductase for N2O emissions from soil
.
Global Change Biology
17
(
3
),
1497
1504
.
Santoro
A. E.
,
Casciotti
K. L.
&
Francis
C. A.
2010
Activity, abundance and diversity of nitrifying archaea and bacteria in the central California current
.
Environmental Microbiology
12
(
7
),
1989
2006
.
Saunders
A. M.
,
Albertsen
M.
,
Vollertsen
J.
&
Nielsen
P. H.
2016
The activated sludge ecosystem contains a core community of abundant organisms
.
The ISME Journal
10
(
1
),
11
20
.
Sun
S. P.
,
Nàcher
C. P.
,
Merkey
B.
,
Zhou
Q.
,
Xia
S. Q.
,
Yang
D. H.
,
Sun
J. H.
&
Smets
B. F.
2010
Effective biological nitrogen removal treatment processes for domestic wastewaters with low C/N ratios: a review
.
Environmental Engineering Science
27
(
2
),
111
126
.
Tiedje
J. M.
1988
Ecology of denitrification and dissimilatory nitrate reduction to ammonium
.
Biology of Anaerobic Microorganisms
717
,
179
244
.
Tolar
B. B.
,
Ross
M. J.
,
Wallsgrove
N. J.
,
Liu
Q.
,
Aluwihare
L. I.
,
Popp
B. N.
&
Hollibaugh
J. T.
2016
Contribution of ammonia oxidation to chemoautotrophy in Antarctic coastal waters
.
The ISME Journal
10
(
11
),
2605
2619
.
Xia
S.
,
Duan
L.
,
Song
Y.
,
Li
J.
,
Piceno
Y. M.
,
Andersen
G. L.
,
Alvarez-Cohen
L.
,
Moreno-Andrade
I.
,
Huang
C. L.
&
Hermanowicz
S. W.
2010
Bacterial community structure in geographically distributed biological wastewater treatment reactors
.
Environmental Science & Technology
44
(
19
),
7391
7396
.

Supplementary data