Efficient nitrogen removal by simultaneous heterotrophic nitrifying-aerobic denitrifying bacterium in a purification tank bioreactor amended with two-stage dissolved oxygen control

Accepted Manuscript Efficient nitrogen removal by simultaneous heterotrophic nitrifying-aerobic denitrifying bacterium in a purification tank bioreactor amended with two-stage dissolved oxygen control Peng Jin, Yinyan Chen, Tao Xu, Zhiwen Cui, Zhanwang Zheng PII: DOI: Reference:

S0960-8524(19)30327-X https://doi.org/10.1016/j.biortech.2019.02.119 BITE 21147

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

21 January 2019 25 February 2019 26 February 2019

Please cite this article as: Jin, P., Chen, Y., Xu, T., Cui, Z., Zheng, Z., Efficient nitrogen removal by simultaneous heterotrophic nitrifying-aerobic denitrifying bacterium in a purification tank bioreactor amended with two-stage dissolved oxygen control, Bioresource Technology (2019), doi: https://doi.org/10.1016/j.biortech.2019.02.119

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Efficient nitrogen removal by simultaneous heterotrophic nitrifying-aerobic denitrifying bacterium in a purification tank bioreactor amended with two-stage dissolved oxygen control Peng Jin1,*, Yinyan Chen3, Tao Xu3, Zhiwen Cui3, Zhanwang Zheng2,3

1

The College of Agricultural and Food Sciences, Zhejiang A & F University, Hangzhou, 311300,

China 2

Zhejiang Shuangliang Sunda Environment co.,LTD, Hangzhou, 310000, China

3

School of Environmental & Resource, Zhejiang A & F University, Hangzhou, 311300, China

*Corresponding authors: The College of Agricultural and Food Sciences, Zhejiang A & F University, Lin’an 311300, China. Tel.: +86-571-18368187965; Fax: +86-571-88218710; P. Jin. ([email protected])

1

Abstract Nitrogen removal performance of a simultaneous heterotrophic nitrifying-aerobic denitrifying (SND) bacterium (KSND) in a purification tank bioreactor (PTBR) amended with two-stage dissolved oxygen (DO) control was investigated. NH4+–N and total nitrogen (TN) removal efficiencies under aerobic conditions for domestic wastewater treatment were 97.12% and 52.64%, respectively. Under serial aerobic (DO > 4.0 mg/L) and anaerobic (DO < 0.5 mg/L) phases, average TN removal efficiency from effluent was 95.45%, without nitrate and nitrite accumulation. DO control assay demonstrated that anaerobic condition adversely affected nitrification (46.13%), but was conducive to denitrification (93.52%). Transcriptional analysis revealed 2.72-fold increase in hydroxylamine reductase expression under aerobic condition as compared to anaerobic condition. Nitrate reductase and nitric oxide reductase homologs had the additional activity of supporting anaerobic or aerobic denitrification in SND bacteria. Under two-stage DO control, KSND maintained high abundance in oligotrophic PTBR, removing 87.88% TN from low-carbon to nitrogen domestic sewage in 180-days.

Keywords: Heterotrophic nitrification-aerobic denitrification; Dissolved oxygen; Transcription analysis; Abundance; Domestic sewage

2

1. Introduction Owing to rapid increase in human activities, including industrial, agricultural, and domestic activities, a series of environmental aftermaths with massive increases in undesirable inorganic nitrogen within the aquatic environment have ensued (Conley et al., 2009). Considering its adverse effects on the environment and human life, elimination of active nitrogen in wastewater, including ammonia, nitrate, and nitrite, is essential and urgent. Biological nitrogen removal, the most common, efficient, and cost-effective method among modern wastewater treatments, is a critical biochemical process for converting nitrogenous compounds to nitrogen gas. Conventional biological processes are essentially based on nitrification and denitrification by nitrifiers and denitrifiers, respectively. However, these microorganisms show distinct differences in their physiology and biochemistry, such as strict limits for dissolved oxygen (DO) content, growth rate, carbon/nitrogen ratio (C/N), and nitrate and nitrite accumulation and tolerance (Shi et al., 2013), making it difficult to coordinate and balance the removal processes and resulting in inefficient and time-consuming procedures (Wu et al., 2015).

Currently, metabolically versatile bacteria have gained interest owing to their ability to efficiently remove different forms of nitrogen in wastewater under aerobic condition. A large number of simultaneous nitrification and denitrification (SND) bacteria had been recently isolated and investigated (Pal et al., 2015; Wan et al., 2017). These unique heterotrophic bacteria display higher growth rates than autotrophs, and can use organic substrates as carbon sources, converting different inorganic nitrogen 3

sources, such as ammonium, nitrate and nitrite, into nitrogenous gas simultaneously (He et al., 2016). In addition, the alkalinity and acidity compensation during nitrification and denitrification could reduce the extra cost of pH-adjusting (Zhang et al., 2012). Especially, this nitrogen removal process could be carried out in in the same reaction pool, and is no longer limited to the combined action of multiclass microorganisms. Moreover, some SND bacterial isolates have been detected with distinct characteristics (Jin et al., 2018; Lei et al., 2015) such as tolerance to low temperature, low C/N, high ammonium level, and high salinity, making them the most economical and promising candidates for nitrogen removal (Bakar et al., 2018). However, most of these studies had mainly focused on characterization of the physiological capability and nitrogen removal performance of SND bacteria in shake flask culture with synthetic medium (Yang et al., 2019). Hence, it is difficult to categorically reflect the comprehensive performance of SND bacteria under harsh natural environmental conditions. Moreover, the potential applications of these bacteria for wastewater treatment in large-scale and open bioreactors have been rarely reported. Besides, it is difficult to generalize large-scale practical application owing to the limitation of rarely successful tested cases. Several possible reasons for this predicament of SND bacteria may be speculated. For instance, in wastewater treatment, the effects of bioaugmentation using specific bacteria are controversial mainly owing to the severe loss and deterioration of the inoculated bacteria (Joo et al., 2006). Nevertheless, many SND bacteria have been found to exhibit excellent nitrogen removal efficiency in the presence of sufficient amount of easily degradable 4

carbon sources (i.e. glucose, sugar, and sodium acetate). The differences in the utilization efficiency of chemical oxygen demand (COD) (refractory organics carbon sources) have been reported to significantly influence cell growth, colonization, and nitrogen removal efficiency. Furthermore, the DO concentration has also been demonstrated to be an important factor affecting the microbial physiology and SND processes (Deng et al., 2017; Ma et al., 2017). Oxygen-control conditions could improve wastewater biodegradability and reduce the toxicity of refractory compounds. It has been indicated that SND bacteria may adapt their metabolic processes based on the DO levels (Sun et al., 2015), achieving the most advantageous viability and maximum treatment efficiency under nutrient-deficient wastewater. Therefore, further investigation of the influence of DO concentration on SND bacteria in open purification tank bioreactors (PTBR) is necessary.

In a previous study, a low-C/N- and temperature-tolerant heterotrophic nitrifying-aerobic denitrifying bacterium Klebsiella sp. (KSND), was isolated and characterized (Jin et al., 2018). This strain exhibited excellent capacity for simultaneous removal of high concentrations of ammonium, nitrite, and nitrate under aerobic conditions. However, further investigations are needed to explore the nitrogen removal capacity of KSND under harsh sewage conditions and to elucidate the nitrogen removal mechanisms and potential application of KSND in wastewater treatment. Thus, the aim of this study was to determine the nitrogen removal efficiency of KSND under different DO levels in domestic sewage treatment. The nitrification and denitrification capabilities of the bacterium were respectively 5

investigated to examine the influences of DO concentration on nitrogen removal efficiency. Furthermore, transcriptional analysis of the genes involved in nitrogen metabolism in KSND under anaerobic and aerobic conditions was employed to determine the nitrification and denitrification capabilities of the bacterium, respectively. Besides, to assess the survival, activity, and proliferation of the inoculated bacterium, colony statistical analysis and quantitative real-time PCR (Q-PCR) were performed to dynamically analyze the population and abundance of KSND in open PTBR. Subsequently, the practical application of KSND in large-scale, open, integrated PTBR (daily treatment capacity, 10 t of domestic sewage) was investigated for 6 months. The findings of this study provide a new perspective on application of SND bacteria in domestic sewage treatment.

2. Methods and materials 2.1 Materials

The bacterial heterotrophic nitrifying-aerobic denitrifying strain Klebsiella sp. KSND used in this study is reported in previous work (Jin et al., 2018). The activated sludge was sampled from the wastewater treatment plants (WWTPs) of Lin’an, Hangzhou, China. Domestic wastewater from the living quarters of Zhejiang A & F University was as the influent raw sewage for all experiment. The primers used in this study are listed in Table 1.

2.2 Sewage treatment systems and operation strategy

This study was conducted in Purification Tank Bioreactor (PTBR) with four 6

operating processes (S1, S2, S3 and S4). Each PTBR had a working volume of 800 L with two serial chambers (each 400 L) and daily treating capacity up to 0.8t of domestic sewage. Throughout the operational period of 30 days, the influent containing the initial COD, NH4+–N and TN concentrations were approximately 130–160 mg/L, 50–58 mg/L and 55–65 mg/L respectively, with undetectable amounts of nitrite and nitrate. The pH was approximately 7.1–7.4, and temperature was 25 ± 1.7 °C. A constant influent and effluent rate of 0.56 L/min was maintained and up to 800 L of treated volume for daily. The four operating processes systems were designed as S1-Control, S2-AS, S3-SD3O and S4-SD3OA. In S1-Control, S2-AS (1% v/v inoculation of activated sludge), and S4-SD3OA (0.1% v/v inoculation of KSND seed culture), air flow rate in the reactor chamber 1 was regulated consistently as aerobic phase with the dissolved oxygen (DO) levels at 4.5 ± 0.2 mg/L, and the chamber 2 as anaerobic phase with the DO levels at 0.5 ± 0.1 mg/L. Two chambers (0.1% v/v inoculation of KSND seed culture) of S3-SD3O system was maintained at 4.5 ± 0.2 mg/L of DO as aerobic phase. The seed culture of strain KSND (about 2 × 107 CFU/mL) was inoculated in S3-SD3O and S4-SD3OA at an initial cell density of 2×104 CFU/mL. These effluent parameters were analyzed including NH4+–N, NO2––N, NO3––N and TN in all of above system. All the assays were performed in triplicate.

2.3 Microbial amounts and abundances analysis in S4-SD3OA To better understand the nitrogen removal efficiency of KSND in S4-SD3OA, the dynamic amounts and relative abundances of KSND were investigated. For microbiology counting, 1 mL of water sample from chambers was diluted into 7

sterilized H2O by multiples of 10, and cultured on Luria-Bertani agarose medium (LB, g/L) 10 peptone, 5 Yeast and 10 sodium chloride) at 37 °C. The amounts of KSND in S4-SD3OA were calculated by positive colony identification with gene-specific PCR. The randomly picked 100 clones from the cultured library was identified with specific primer pairs HAR-TF/HAR-TR of KSND hydroxylamine reductase HAR gene. The PCR was carried out as follows: 5 min at 94 °C, 30 cycles of 1 min at 94 °C, 1 min at 53 °C, 1.5 min at 72 °C and a final step of 10 min at 72 °C. The abundances of total microbial communities and KSND strain were determined by quantifying the 16S rRNA gene and HAR gene, respectively. The quantitative PCR (Q-PCR) was performed in a Corbett Real-Time PCR Machine with the Rotor-Gene 6000 series software 1.7 (Qiagen, The Netherlands), using the SYBR Green. Standard curves were constructed for the 16S rDNA and HAR genes using the recombinant plasmids 16S-pMD19 and Har-pMD19 as standards, respectively. The samples and standards were analyzed with qPCR in triplicate, and the specificity of the qPCRs was confirmed with a melting curve analysis, agarose gel electrophoresis, and DNA sequencing. The efficiencies of the PCRs were 92.3%–98.9%, with R2 values > 0.994 for all calibration curves. The quantitative PCR primer pairs for 16S rRNA and HAR genes were 27F/1492R and HAR-F/HAR-R, respectively. 2.4. Analysis of DO experiments KSND cells were per-cultured in LB medium, and then incubated at 30 °C with shaking at 200 rpm for 16h as the seed cultures. To evaluate the effects of DO concentrations on nitrogen removal, 250 mL flasks containing 100 mL of the 8

modified medium (g/L: 2.50 glucose, 0.50 KH2PO4, 0.50 K2HPO4, 0.20 MgSO4·7H2O and 2.00 mL trace elements) with NH4Cl and KNO3 of initial 100 mg/L as the sole nitrogen source respectively, were inoculated with 1 mL of seed culture and then cultured at 30 °C with shaking at 50 rpm for aerobic condition and static cultivation for anaerobic condition. To evaluate the effects of different carbon sources on nitrogen removal efficiency, glucose, sodium acetate, sodium bicarbonate, sodium citrate and soluble COD of domestic sewage were used as the sole carbon source with added initial ammonium of 50 mg/L as C/N of 4:1, respectively. Nitrogen removal efficiency (ammonium and total nitrogen), accumulation of nitrate and cell growth rate of the bacterium were measured at regular interval for a period of 24 h. All experiments were carried out in triplicate. The effects of two-stage DO of PTBR on simultaneous nitrification and denitrification were examined in 10 PTBRs (daily sewage treatment capacity of 5t) for two weeks. Various DO concentrations (0.25, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0 and 6.0 mg/L) were set at chamber 1 and chamber 2 separately. A constant influent and effluent rate of 3.5 L/min was maintained for daily, and the influent COD, NH4+–N and TN concentrations were approximately 130–160 mg/L, 50–58 mg/L and 55–65mg/L respectively, with undetectable amounts of nitrite and nitrate. The operation was at temperature of 25 ± 2 °C and pH of 7.4 ± 0.2. 2.5. Real-time PCR assay To quantify the gene transcriptional levels during nitrogen removal process under aerobic/anaerobic conditions, cell was collected from the modified medium with 100 mg/L NH4+–N or NO3––N used as sole nitrogen source at mid-logarithmic growth 9

phase (cultivated 16 h after inoculation) under aerobic and anaerobic conditions, respectively. The equivalent cells of 3 OD600 were harvested and frozen immediately in liquid nitrogen. Total mRNA of strains was extracted using a Bacterial Total RNA extraction kit (Hangzhou Biosci Co., Ltd, China.), and cDNA was synthesised with the M-MlV first-strand RT kit (Hangzhou Biosci Co., Ltd, China) according to manufacturer’s specifications. For all RT-PCR reactions, according to its consistent expression and little variability, 16S rDNA was used as a control for normalization between samples. qRT-PCR data were treated using the 2¯∆∆Ct method described previously (Kang et al., 2011) with the software of Opticon Monitor 3. The experiment was carried out with three replicates under identical experimental conditions and appropriate negative controls were maintained.

2.6 Application of KSND strain in Amplifying PTBR

To further investigate the application potential of KSND in the enlarged, open and frequently fluid reactors, we performed this study at an integrated purification tank bioreactor (daily treatment capacity of 10 t domestic sewage). 20 L seed culture was mixed with 20 L domestic wastewater with aeration for 8 h. The mixtures were added into PTBR. Subsequently, domestic wastewater incessantly flowed into the chamber at a constant flow rate of 6.9 L/min, and up to 104 L of total volume, and automatically overflowed out with the same rate. The experiment was carried out 6 months from July 1 to Dec. 1, 2018. The DO concentrations of chamber 1 and chamber 2 were stably maintained at 5.0 ± 0.3 mg/L and 0.4 ± 0.2 mg/L, respectively. 10

Nitrogen removal efficiency was measured (ammonium, total nitrogen and COD) at regular interval of 3 days for a period of 180 days. 2.7 Analytical methods and calculation The concentrations of NH4+–N, NO2––N, NO3––N and TN were measured by Nessler's reagent spectrophotometry, N-(1-naphthalene)-diaminoethane photometry, ultraviolet spectrophotometric method and potassium persulfate digestion UV spectrophotometric method, respectively, and chemical oxygen demand (COD) was analyzed using dichromate method. Dissolved oxygen (DO) was determined by a dissolved oxygen meter (Bante820, Bante instruments, China). The NH4+–N, NO2––N, NO3––N, TN and COD removal ratio formula was (ρ0–ρn)/ ρ0×100%, where ρ0 is the initial nitrogen concentration and ρn is the final concentration of NH4+–N, NO2––N, NO3––N, TN and COD at n hour. Statistical analysis and graphical work were carried out by using Excel and Origin 9.0. The draft genome sequence has been deposited at NCBI/GenBank under the accession NWVG00000000.

3 Results and discussion 3.1 Performances of the four PTBR systems

The four sewage treatment systems (daily sewage treatment capacity, 0.8 t) were established to evaluate the reactor performance with respect to COD and nitrogen removal for 30 days after inoculation (Fig. 1). For the entire duration of the experiment, these modified reactor systems (S1-Control, S2-AS, S3-SD3O, and 11

S4-SD3OA) were operated in two stages (Chamber 1 and Chamber 2) at controlled DO concentrations. The variations in the COD concentration and removal efficiencies of all the reactors are shown in Fig. 2. In S1-Control, the average effluent COD, NH4+–N, and TN concentrations were 84.1–120.8, 41.80–48.70, and 45.71–53.80 mg/L,

which

corresponded

to

removal

efficiencies

of

16.69%–42.00%,

9.81%–22.59%, and 10.3%–23.82%, respectively. The low removal efficiency of S1-Control can be attributed to the poor microbial population colonized in the two chambers. Notably, the other three bioreactors with microbial inoculants had different COD and nitrogen removal. When compared with the control, the three systems displayed equally excellent COD removal performances, with average COD removal efficiencies of >79.29%, indicating that the changes in the DO concentration had no adverse influence on COD removal. This observation is similar to the removal efficiency demonstrated in previous reports (He et al., 2009; Yan et al., 2019).

However, the residual nitrogen in the effluent of the three systems exhibited significant differences. In S2-AS bioreactor with activated sludge inoculation, the residual TN and NH4+–N concentrations were 16.00–19.93 and 9.09–13.50 mg/L, corresponding to average removal efficiencies of 63.25 ± 3.53% and 79.09 ± 4.09 %, respectively. Meanwhile, the redundant NO3−–N from NH4+–N oxidation in the reactors did not decrease at all, resulting in simultaneous nitrogen accumulation (3.24–6.56 mg/L) in the effluent. This accumulation could be attributed to the low COD/N ratio causing incomplete nitrification and denitrification. It must be noted that activated sludge contains abundant active microorganisms involved in biological 12

denitrification (Cokgor et al., 2009) and phosphorus removal (Yang et al., 2019). Therefore, the DO levels of Chambers 1 and 2 were strictly controlled to achieve aerobic (DO > 4 mg/L) and anaerobic (DO < 0.5 mg/L) conditions under continuous wastewater inflow, which were beneficial for rapid accumulation of dominant nitrification and denitrification species, respectively. However, the TN and NH4+–N removal efficiencies achieved with conventional biological nitrification and denitrification are generally very low in domestic sewage treatment owing to the low C/N in sewage (Jin et al., 2018), which is not favorable for complete nitrification process during the reaction stage. For instance, the average concentration of effluent COD was only 32.60 ± 3.86 mg/L, with a C/N of about 1.5:1. Thus, conventional biological nitrification and denitrification may not be suitable for achieving almost no residual nitrogen in low-C/N domestic sewage (Wett and Rauch, 2003; Zhou et al., 2017). In S3-SD3O and S4-SD3OA, the NH4+–N removal rates were reliable and stable at 96.61% under aerobic condition. However, unexpectedly, when compared with the control and S2-AS bioreactors, obvious accumulation of NO3−–N and NO2−–N was noted in the S3-SD3O effluent, with average concentrations of 20.45 ± 2.13 and 4.78 ± 1.74 mg/L throughout the test period, respectively, resulting in average TN removal efficiency of 52.64 ± 4.19%. In contrast, when the DO level in Chamber 2 of S4-SD3OA was maintained at 0.3 ± 0.2 mg/L (anaerobic condition), the average TN removal efficiency was at a high level of 95.45 ± 1.37%, especially without NO3−–N and NO2−–N accumulation. These results indicated that high DO concentration 13

favored nitrification by KSND, but inhibited the denitrification process. Therefore, alternating the DO concentration in the bioreactor could effectively improve the effluent quality with almost no residual nitrogen (Yan et al., 2019).

3.2 Nitrogen removal performance of the modified S4-SD3OA Figure 3 shows the variations in the NO3−–N and NO2−–N accumulation and TN and NH4+–N removal efficiencies in the aerobic (Chamber 1) and anaerobic (Chamber 2) phases of the modified S4-SD3OA. The concentration of NH4+–N in Chamber 1 effluent was 2.60–6.74 mg/L with average NH4+–N removal efficiency of 91.35 ± 3.83% (Fig. 3A). However, effluent NH4+–N concentration in Chamber 2 was slightly decreased further to 1.06–3.35 mg/L, suggesting that the nitrification capacity of KSND (the dominant microorganism in Chamber 2) may have been suppressed by carbon-limited and anaerobic conditions (Chen et al., 2018). In particular, large amounts of NO2−–N and NO3−–N were also observed in Chamber 1 effluent, with average concentrations of 3.60 ± 0.92 and 18.64 ± 3.02 mg/L, respectively (Fig. 3B and 3C). The concentration of TN in Chamber 1 effluent was only 34.11 ± 2.48 mg/L, which corresponded to the average TN removal efficiencies of approximately 43.15 ± 4.13% (Fig. 3D). This removal efficiency under aeration for 12 h was similar to that of S3-SD3O under aeration for 24 h (Fig. 2C), suggesting that extending the reaction duration was ineffective in improving TN removal, which may be owing to the inhibition of simultaneous nitrification and denitrification process by excess oxygen content (Zhu et al., 2012). 14

Therefore, Chamber 1 effluent (predominantly comprising NO2−–N and NO3−–N) was subjected to anaerobic condition (Chamber 2) to achieve efficient removal of NO2−–N and NO3−–N. The results revealed approximately 100% NO2−–N removal (Fig. 3B) and gradual decrease in the effluent NO3−–N concentration to 0.56 ± 0.43 mg/L, with a removal efficiency of > 93.59% in Chamber 2. However, the finding of increased NO2−–N accumulation at high DO concentration is inconsistent with that reported in a previous study in which low DO concentrations were noted to be favorable for NO2−–N accumulation (Liu et al., 2017). This inconsistent observation could probably be owing to the fact that the nitrogen removal pathways in SND bacteria are different from conventional nitrification and denitrification (i.e. activated sludge) (Jin et al., 2018). Moreover, the TN concentration in Chamber 2 of the modified S4-SD3OA decreased from 34.11 ± 2.48 to 2.68 ± 0.87 mg/L with a removal efficiency of > 88.78%, suggesting that low DO concentration is beneficial for providing an anoxic environment for denitrification by SND bacteria. In a previous study (Jin et al., 2018), KSND was noted to exhibit a high NH4+–N removal efficiency of 86.56% in flask culture containing glucose and 160 mg/L NH4+–N, which could be owing to the relatively easily degradable carbon source and suitable DO concentration allowing effective nitrification and denitrification by KSND. Thus, although numerous SND bacteria with excellent performances have been reported, their low nitrogen removal efficiency in practical domestic sewage treatment could be attributed to the significant inhibitory effects of the carbon source (underutilized and low C/N) and aeration environment. 15

3.3 Dynamic variation in the abundance of KSND in PTBR

To better understand the nitrogen removal efficiency of KSND in PTBR (S4-SD3OA) (Fig. 2C), the habitat population and relative abundance of KSND were investigated. As shown in Fig. 4, DO concentration significantly affected the bacterial community richness and diversity. Figure 4A illustrates the relative richness of KSND inoculated into Chamber 1 at the startup phase (about 2.00 × 104 CFU/mL); however, the abundance sharply decreased to 3.40 × 102 CFU/mL within the first 5 days, and subsequently remained at a low level of about 0.42 × 102 CFU/mL in the later stage of the experiment. Furthermore, the results of gene-specific Q-PCR analysis (Fig. 4B) indicated that KSND was predominant (about 81% of the sequences belonged to KSND) during the first 5 days, but its abundance decreased to about 15% after 20 days. Although only less than 0.5% of KSND completely colonized the aerobic Chamber 1, its relative abundance was high, accomplishing the entire nitrifying process during the reaction stage. This predominance of KSND may be owing to the washout of the majority of the other microbial cells as well as competition for space and nutritional conditions (especially carbon and nitrogen sources) between KSND and other bacteria, resulting in colonization of appropriate amount of KSND in the aerobic Chamber 1 for balanced nitrogen metabolism.

Quantification of bacterial colonies revealed that the overall bacterial population in the anaerobic Chamber 2 was significantly higher than that in the aerobic Chamber 1 (Fig. 4A), with predominantly high relative abundances (>90%) (Fig. 4B), similar to 16

that noted previously (data not published). This increase in microbial abundance might be owing to the following reasons. First, the anaerobic environment in Chamber 2 would have provided favorable conditions for cell attachment and growth, thus significantly reducing washout of microbial cells. Second, oligotrophic domestic sewage (low C/N) may not contain adequate nutrients for second-phase consumption via first-phase consumption, which would have led to severe shortage of carbon sources and inhibition of the growth of most of the other microorganisms. Third, the high concentrations of nitrate and nitrite in Chamber 1 effluent (Fig. 3B) would have produced toxic effects on environmental bacteria (Alvaro and Camargo, 2009), inhibiting the growth of most of the other microorganisms as well. However, KSND has been reported to effectively remove high concentrations of nitrate and nitrite, without being inhibited by their toxic effects (Jin et al., 2018). Furthermore, it has been shown that chemolithotrophs such as Klebsiella spp. could use nitrate or nitrite as electron acceptor to oxidize inorganic or organic sulfur compounds (Jing et al., 2017; Pal et al., 2015). Taken together, it can be concluded that KSND, as a metabolically versatile bacterium, could maintain relatively high abundance of more than 90% in the oligotrophic and anaerobic Chamber 2, creating favorable conditions for efficient TN removal in PTBR.

3.4 Effects of DO concentration on nitrification and denitrification performances of KSND

While the nitrogen removal performance of SND bacteria had been extensively 17

studied in shake-flask experiments, the effect of DO concentration on the nitrification and denitrification performances of these bacteria had been rarely investigated. In the present study, the influence of DO concentration on the TN removal efficiencies of the low-C/N-tolerant KSND was further investigated. The nitrification and denitrification efficiencies of KSND under aerobic (50 rpm, DO of about 3.8 mg/L) and anaerobic (static culture, DO of about 0.5 mg/L) shake-flask culture were investigated with ammonium chloride and sodium nitrate, respectively (Fig. 5A). KSND exhibited similar growth phases and nearly identical optical density (OD) values (OD600 of about 1.23–1.38) at 48 h under aerobic and anaerobic conditions, but presented significantly lower cell growth (OD600 of about 2.5–2.7) at 24 h under constant shaking at 200 rpm (Jin et al., 2018). However, the NH4+–N removal efficiency (90.01%) and rate of nitrification (1.87 mg/L/h) for the bacterium were higher than that of denitrification (78.50% and 1.64 mg/L/h, respectively) under aerobic condition (Fig. 5A). Surprisingly, anaerobic condition had an important effect on the nitrification capability of KSND, causing significant decrease in the NH4+–N removal efficiency (46.13%) and rate (0.96 mg/L/h). However, the maximal denitrification efficiency of KSND significantly increased to 93.52%, suggesting that anaerobic condition was more conducive to denitrifying metabolism.

Interestingly, when compared with the relatively high TN removal capability of KSND (>90%) under aerobic shake-flask culture (DO > 4 mg/L), that in PTBR under aerobic condition was only 45.80% (DO > 4). This might be owing to the differences in the availability of carbon sources (i.e. glucose in shake-flask culture and soluble 18

COD in PTBR), which could have exerted noticeable effects on the simultaneous nitrification and denitrification capacity of the bacterium. Hence, the effects of different carbon sources on the nitrogen removal efficiency of KSND were assayed. As shown in Fig. 5B, the TN removal efficiency of the bacterium obviously decreased when bicarbonate, sodium citrate, or soluble COD of domestic sewage was the sole carbon source, with TN removal rates of 65.4%, 47.4%, and 67.8%, respectively, when compared with that achieved using glucose as the sole carbon source (87.8%). These results indicated that the carbon conversion efficiency of KSND had severe effects on cell metabolism. Thus, moderate DO concentrations could provide a better nitrifying and denitrifying environment for microbial cellular energy metabolism (Sun et al., 2015). Furthermore, accumulation of nitrite was undetectable during the aerobic and anaerobic NH4+–N removal process, which may be attributed to the rapid removal of trace nitrite (produced by denitrification) by nitrite reductase. The accumulation of nitrate during aerobic nitrification (1.32–1.98 mg/L) was higher than that (0.12–0.25 mg/L) during anaerobic nitrification (data not shown). To further estimate the effect of DO concentration on SND, the study was performed using 10 PTBR with two stages (daily sewage treatment capacity of 5 t) for 2 weeks. Figure 5C shows the variations in the DO concentration and efficiencies of nitrification and denitrification in the two stages of modified PTBR. When the DO concentration in Chamber 1 was maintained at 4.0 mg/L, the effluent NH4+–N removal efficiency was 89.87 ± 2.03%, which further increased to 94.65 ± 1.43% with the increase in DO level. However, when the concentration of DO gradually 19

decreased to 1.0 mg/L, a significant decrease in NH4+–N removal efficiency from 90% to 39.62 ± 0.89% was noted. Under anaerobic condition, KSND exhibited relatively high nitrification activity. At 0.25 mg/L DO, KSND maintained almost 35% of NH4+–N removal efficiency, indicating its strong nitrification capacity and tolerance to DO to accomplish metabolism of limited nitrogen source. Subsequently, the denitrification capacity of KSND was further verified by batch experiments. The denitrification efficiency (NO4––N and NO3––N removal in Chamber 2) of KSND was 93.85 ± 1.78% and 91.38 ± 0.72% at 0.25 and 0.50 mg/L DO, respectively. With further reduction in the DO concentration from 0.5 to 3.0 mg/L, the denitrification ability of KSND significantly decreased from 91.38% to 14.37%; in particular, high DO content (> 4.0 mg/L) resulted in low denitrification efficiency of about 3%. It must be noted that the nitrification activity was enhanced, whereas denitrification activities were inhibited at high DO concentration, causing a negative impact on KSND, which is consistent with the results reported in a previous study (He et al., 2009). Therefore, the DO concentration gradient in modified PTBR is favorable for simultaneous nitrogen removal from domestic wastewater by SND bacteria.

3.6 Effects of DO levels on genes expression in KSND

The draft genome showed the involvement of hydroxylamine reductase (HAR), nitrate reductase (NarH' and NapA), and nitric oxide reductase (NOR) [flavorubredoxin (FlRd) and FlRd-NAD(+) reductase (FlRd-red)] homologous genes 20

in nitrogen metabolism pathway in KSND. Transcriptional analysis was performed to understand the effects of DO levels on the enzymes involved in nitrogen removal processes. As shown in Fig. 5D, in ammonium medium, the expression level of the HAR gene under aerobic condition significantly increased (2.72 fold), when compared with that under anaerobic conditions, whereas HAR expression levels were slightly downregulated in nitrate medium. This result indicated that HAR may play an important role contributing to the high nitrification capability of KSND (Padhi et al., 2017), causing a significant decrease in nitrification (46.13%) under anaerobic condition (Fig. 5A). With regard to the expression of Nar-type homologs, NapA and NarH', aerobic conditions resulted in a 1.46-fold (ammonium medium) and a 0.97-fold (nitrate medium) increase in NapA expression, whereas a 3.08-fold (ammonium medium) and a 2.87-fold (nitrate medium) decrease in NarH' expression occurred, when compared with those under anaerobic condition. It has been reported that NarH' can act as a substitute for the NAR enzyme, allowing utilization of nitrate as an electron acceptor in Escherichia coli during anaerobic growth (Sven et al., 2009). The results of the present study demonstrated that KSND possesses potential substitute pathways for supporting anaerobic or aerobic reduction of nitrate. Intriguingly, two novel NOR homologs, FlRd and FlRd-red, in KSND were found to metabolize NO to N2O during denitrification. FlRd-red gene expression was significantly upregulated by 2.35 fold and 1.63 fold in ammonium and nitrate media, respectively, under aerobic conditions, when compared with that under anaerobic conditions. In contrast, FlRd gene expression levels were obviously lower by 1.39 21

fold and 1.97 fold in ammonium and nitrate media, respectively, under aerobic conditions, when compared with those under anaerobic conditions. In denitrifying bacteria, inducible O2-sensitive NOR activity detoxifies NO under anaerobic and microaerobic conditions, and failure of this activity may result in poor denitrification function (Gardner et al., 2003). However, KSND was noted to exhibit excellent nitrogen removal performance under anaerobic and aerobic conditions, which may be owing to the homologous enzyme pathways found in SND bacteria, which are clearly different from the conventional anaerobic denitrification process (Gomes et al., 2002; Pal et al., 2015). To the best of the authors’ knowledge, the present study is the first to perform

transcription-level

analysis

of

heterotrophic

nitrification-aerobic

denitrification efficiency of SND bacteria by DO regulation. In practical applications, the low utilization efficiency of soluble COD could severely affect microbial growth and nitrogen removal; hence, regulation of the DO levels, which can provide a microaerobic/anoxic environment for high-efficiency denitrification and energy metabolism, may be a good approach to promote TN removal efficiency (Sun et al., 2015).

3.7 Feasibility and stability of PTBR system for treating low-C/N wastewater

Despite numerous reports on heterotrophic denitrifying bacteria with efficient nitrogen removal capacity, only a few studies have focused on assessing their practical applications in large-scale and open bioreactors. In the present study, the application potential of KSND was further investigated in integrated PTBR (daily 22

treatment capacity, 10 t of domestic sewage) for 180 days. As shown in Fig. 6, the daily influent concentrations of COD, TN, and NH4+–N were approximately 117.20–137.90, 47.71–56.08, and 39.34–47.2 mg/L respectively, with untraceable amounts of NO3––N and NO2––N. Following KSND inoculation, the effluent concentrations of COD and NH4+–N immediately dropped to about 17.00 and 3.50 mg/L within 24 h, respectively. After 3 days of experiment, the effluent concentrations of TN and NO3––N decreased to 6.06 and 0.35 mg/L, respectively. Remarkably, the effluent COD, TN, and NH4+–N concentrations remained steady throughout the rest of the experiment (140 days; from July to November). After 180 days of the experiment, the NO3––N concentration was undetectable.

Temperature is an important factor that can influence the nitrogen removal capacity of SND bacteria. The maximum temperature of the PTBR system was observed to be 23.23 ± 0.2°C and the average temperature was 21.97 ± 1.55°C from July to November. At this temperature, the average COD, TN, and NH4+–N removal efficiencies were 90.90%, 90.56%, and 94.08%, respectively, during the 140-day experiment. These results indicated the entire nitrogen removal efficiency of KSND was higher than some other new processes, such as 72.28% TN removal of oxygen-limited SBR (Yan et al., 2019), 52.70% TN removal of vertical flow constructed wetlands (Zhou et al., 2017). However, the reactor temperature gradually decreased to 12.03 ± 2.61°C at the end of the experiment (from November to December), which had a significant impact on nitrogen removal by KSND. Although the average effluent COD, TN, and NH4+–N concentrations increased to 19.3, 7.68, 23

and 4.75 mg/L, respectively, at the end of 40 days of the experiment, the corresponding average COD, TN, and NH4+–N removal efficiencies remained at 84.87%, 85.20%, and 89.02% respectively. This may be owing to the low temperature tolerance of KSND (Jin et al., 2018). These results showed that KSND inoculated into PTBR could maintain its survival capability and efficiently eliminate nitrogen released from domestic sewage for 180 days, without requiring addition of external carbon source and supplementary inoculation of the bacterium.

Aerobic nitrification and anaerobic denitrification is the most widely adopted biological nitrogen removal proceeds. Unfortunately, these enhanced processes have generally not performed well and do not meet the current discharge standard (10 mg/L TN) of domestic wastewater, which was ascribed to the slow growth of microorganisms, the low stability and activity of bacteria and insufficient/unavailable carbon source (Khin and Annachhatre, 2004). In this study, TN removal capacity was significantly increased in S4-SD3OA compared with S1-Control and S2-AS, which was slightly lower than that under the lab-conditions. As reported before, the threshold theory postulates the denitrification rate is not affected by the DO over a certain range, but that the activity of denitrification enzymes increases significantly when DO concentration decreases (Patureau et al., 2000). Therefore, compared with the simultaneous nitrification and denitrification of aerobic denitrifiers in single aerobic reactors, two-stage DO control can provide a significant advantage for KSND bacterium over conventional systems and would keep efficient nitrogen removal efficiency and save time. In addition, the process of PTBR is easy to operate stably by 24

aerated in Chamber 1 and unaerated in Chamber 2. The low C/N-tolerant KSND strains could maintain faster growth rates in both alternating DO levels than that of those bacteria in conventional aerobic/anaerobic conditions, which achieve the entire nitrogen removal for low C/N domestic wastewater without added COD. This greatly reduces the treatment time and energy requirements. Overall, the present study demonstrated the excellent performances of KSND combined with two-stage DO control of PTBR, which could have significant potential application in the removal of nitrogen from domestic wastewater, especially from food, aquaculture, and agriculture sectors.

4 Conclusion DO levels had significant effects on nitrification and denitrification performances of SND bacteria in oligotrophic domestic sewage treatment. The ammonium degradation efficiency of KSND decreased, whereas nitrate and nitrite degrading capacity increased with decreasing DO content. At optimal DO levels, KSND abundance in PTBR was high, indicating steady survival of the bacterium. Transcriptional analysis suggested correlation between DO concentration and nitrifying and denitrifying ability of KSND. Thus, DO concentration, oligotrophic status, open bioreactor, and domestic sewage flow affect the nitrogen removal efficiency of SND bacteria. These findings provide novel insights into nitrogen removal by SND bacteria for domestic sewage treatment.

Acknowledgments 25

This work was supported by the National Natural Science Foundation of China (Grant 31700078, 21276235), the Scientific Research Foundation for Talent program of Zhejiang Agricultural and Forestry University (W20170029).

Notes The authors declare no competing financial interest. References 1. Alvaro, A., Camargo, J.A., 2009. Effects of pulse duration and post-exposure period on the nitrite toxicity to a freshwater amphipod. Ecotox. Environ. Safe. 72, 2005-2008. 2. Bakar, S.N.H.A., Hasan, H.A., Mohammad, A.W., Abdullah, S.R.S., Haan, T.Y., Ngteni, R., Yusof, K.M.M., 2018. A review of moving-bed biofilm reactor technology for palm oil mill effluent treatment. J. Clean. Prod. 171, 1532-1545. 3. Chen, H., Zhao, X., Cheng, Y., Jiang, M., Li, X., Xue, G., 2018. Iron robustly stimulates simultaneous nitrification and denitrification under aerobic conditions. Environ. Sci. Technol. 52, 1404-1412. 4. Cokgor, E., Oktay, S., Tas, D., Zengin, G., Orhon, D., 2009. Influence of pH and temperature on soluble substrate generation with primary sludge fermentation. Bioresour. Technol. 100, 380-386. 5. Conley, D.J., Paerl, H.W., Howarth, R.W., Boesch, D.F., Seitzinger, S.P., Havens, K.E., Christiane, L., Likens, G.E., 2009. Controlling eutrophication: nitrogen and phosphorus. Science 323, 1014-1015. 6. Deng, Y.L., Ruan, Y.J., Zhu, S.M., Guo, X.S., Han, Z.Y., Ye, Z.Y., Liu, G., Shi, M.M., 2017. The impact of DO and salinity on microbial community in poly(butylene succinate) denitrification reactors for recirculating aquaculture system wastewater treatment. AMB Express 7, 113. 7. Gardner, A.M., Gessner, C.R., Gardner, P.R., 2003. Regulation of the nitric oxide reduction operon (norRVW) in Escherichia coli. Role of NorR and sigma54 in the nitric oxide stress response. J. Biol. Chem. 278, 10081-10086. 8. Gomes, C.M., Alessandro, G., Elena, F., Vicente, J.O.B., Saraiva, L.M., Maurizio, B., Miguel, T., 2002. A novel type of nitric-oxide reductase. Escherichia coli flavorubredoxin. J. Biol. Chem. 277, 25273-25276. 9. He, S.B., Xue, G., Wang, B.Z., 2009. Factors affecting simultaneous nitrification and de-nitrification (SND) and its kinetics model in membrane bioreactor. J. Hazard. Mater. 168, 704-710. 10. He, T., Li, Z., Sun, Q., Xu, Y., Ye, Q., 2016. Heterotrophic nitrification and aerobic denitrification by Pseudomonas tolaasii Y-11 without nitrite accumulation during nitrogen conversion. Bioresour. Technol. 200, 493-499. 26

11. Jin, P., Chen, Y., Zheng, Z., Du, Q., 2018. Evaluation of a novel low-carbon to nitrogen- and temperature-tolerant simultaneously nitrifying–denitrifying bacterium and its use in the treatment of river water. RSC Adv. 8, 27417-27422. 12. Jing, Z., Feng, L., Yang, G., Chen, D., Mu, J., 2017. Development of simultaneous nitrification-denitrification (SND) and organics removal in biofilm reactors with partially coupled a novel biodegradable carrier for nitrogen-rich water purification. Bioresour. Technol. 243, 800-809. 13. Joo, H.S., Hirai, M., Shoda, M., 2006. Piggery wastewater treatment using Alcaligenes faecalis strain No. 4 with heterotrophic nitrification and aerobic denitrification. Water Res. 40, 3029-3036. 14. Kang, Z., Wang, Y., Gu, P., Wang, Q., Qi, Q., 2011. Engineering Escherichia coli for efficient production of 5-aminolevulinic acid from glucose. Metab. Eng. 13, 492-498. 15. Khin, T., Annachhatre, A.P., 2004. Novel microbial nitrogen removal processes. Biotechnol. Adv. 22, 519-532. 16. Lei, Y., Ren, Y., Liang, X., Zhao, S., Wang, J., Xia, Z., 2015. Nitrogen removal characteristics of a heterotrophic nitrifier Acinetobacter junii YB and its potential application for the treatment of high-strength nitrogenous wastewater. Bioresour. Technol. 193, 227-233. 17. Liu, J., Yuan, Y., Li, B., Zhang, Q., Wu, L., Li, X., Peng, Y., 2017. Enhanced nitrogen and phosphorus removal from municipal wastewater in an anaerobic-aerobic-anoxic sequencing batch reactor with sludge fermentation products as carbon source. Bioresour. Technol. 244, 1158-1165. 18. Ma, W., Han, Y., Ma, W., Han, H., Zhu, H., Xu, C., Li, K., Wang, D., 2017. Enhanced nitrogen removal from coal gasification wastewater by simultaneous nitrification and denitrification (SND) in an oxygen-limited aeration sequencing batch biofilm reactor. Bioresour. Technol. 244, 84-91. 19. Padhi, S.K., Tripathy, S., Mohanty, S., Maiti, N.K., 2017. Aerobic and heterotrophic nitrogen removal by Enterobacter cloacae CF-S27 with efficient utilization of hydroxylamine. Bioresour. Technol. 232, 285-296. 20. Pal, R.R., Khardenavis, A.A., Purohit, H.J., 2015. Identification and monitoring of nitrification and denitrification genes in Klebsiella pneumoniae EGD-HP19-C for its ability to perform heterotrophic nitrification and aerobic denitrification. Funct. Integr. Genomics 15, 63-76. 21. Patureau, D., Bernet, N., Delgenès, J.P., Moletta, R., 2000. Effect of dissolved oxygen and carbon–nitrogen loads on denitrification by an aerobic consortium. Appl. Microbiol. Biotechnol. 54, 535-542. 22. Shi, Y., Hu, S., Lou, J., Lu, P., Keller, J., Yuan, Z., 2013. Nitrogen removal from wastewater by coupling anammox and methane-dependent denitrification in a membrane biofilm reactor. Environ. Sci. Technol. 47, 11577-11583. 23. Sun, Y., Li, A., Zhang, X., Fang, M., 2015. Regulation of dissolved oxygen from accumulated nitrite during the heterotrophic nitrification and aerobic denitrification of Pseudomonas stutzeri T13. Appl. Microbiol. Biot. 99, 27

3243-3248. 24. Sven, M., Yvonne, T., Julia, M., Sonja, S., Insa, J., Isabel, W., Sarah, H., Birgit, A., Mascha, S., Wolfgang, W., 2009. The roles of the nitrate reductase NarGHJI, the nitrite reductase NirBD and the response regulator GlnR in nitrate assimilation of Mycobacterium tuberculosis. Microbiology 155, 1332-9. 25. Wan, W., He, D., Xue, Z., 2017. Removal of nitrogen and phosphorus by heterotrophic nitrification-aerobic denitrification of a denitrifying phosphorus-accumulating bacterium Enterobacter cloacae HW-15. Ecol. Eng. 99, 199-208. 26. Wett, B., Rauch, W., 2003. The role of inorganic carbon limitation in biological nitrogen removal of extremely ammonia concentrated wastewater. Water Res. 37, 1100-1110. 27. Wu, Y., Shukal, S., Mukherjee, M., Cao, B., 2015. Involvement in denitrification is beneficial to the biofilm lifestyle of comamonas testosteroni: a mechanistic study and its environmental implications. Environ. Sci. Technol. 49, 11551-11559. 28. Yan, L., Liu, S., Liu, Q., Zhang, M., Liu, Y., Wen, Y., Chen, Z., Zhang, Y., Yang, Q., 2019. Improved performance of simultaneous nitrification and denitrification via nitrite in an oxygen-limited SBR by alternating the DO. Bioresour. Technol. 275, 153–162. 29.Yang, J., Liu, X., Wang, D., Xu, Q., Yang, Q., Zeng, G., Li, X., Liu, Y., Gong, J., Ye, J., 2019. Mechanisms of peroxymonosulfate pretreatment enhancing production of short-chain fatty acids from waste activated sludge. Water research 148, 239-249. 30. Zhang, Q., Liu, Y., Ai, G., Miao, L., Zheng, H., Liu, Z., 2012. The characteristics of a novel heterotrophic nitrification–aerobic denitrification bacterium, Bacillus methylotrophicus strain L7. Bioresour. Technol. 108, 35-44. 31. Zhou, X., Wang, X., Zhang, H., Wu, H., 2017. Enhanced nitrogen removal of low C/N domestic wastewater using a biochar-amended aerated vertical flow constructed wetland. Bioresour. Technol. 241, 269-275. 32. Zhu, L., Feng, L.J., Dai, X., Xu, X.Y., 2012. Characteristics of an aerobic denitrifier that utilizes ammonium and nitrate simultaneously under the oligotrophic niche. Environ. Sci. Pollut. R. 19, 3185-3191.

28

Table 1 The primers used in this work. Primer

Sequence (5'-3')

HAR-TF HAR-TR 27F 1492R HAR-F HAR-R

GACCATGTTTTGTGTGCAATGTGAAC CAGGCGCGGCAGCCCTTCGATATCGC AGAGTTTGATCCTGGCTCAG TACGGCTACCTTGTTACGACTT GCGATTATTCTGGCGGTCA AGCAGCGTCAGCAGGATGA

NapA-F NapA-R NarH'-F NarH'-R

GCTGTCCAACGACATGCCTTGCT GCCGCAGTTGGTCCCGCATTTCA GGCGATGGACTGGAAGCTGG GCAGATTCGCCAGGTATTGC

FlRd-F FlRd-R FlRd-red-F FlRd-red-R

TGTACGAACAGTGCCAGCGCTAC CGAGGTATTTCTCGACGATTTGC GCGACTGTGCGGAAATCAA CACCTTCACCAGCATTGGC

29

Figure 1 Flow chart of the sewage treatment equipment. (A) Schematic diagram of the PTBR. a1, a2 and a3 are influent pipe, overflow pipe and effluent pipe, respectively; b1 and b2 are aeration control tube; c1 and c2 are Chamber covers. (B) Schematic diagram of the four operating processes. S1: control; S2: Activated sludge was inoculated into Chamber 1 and Chamber 2; S3 and S4: KSND strain was inoculated into Chamber 1 and Chamber 2.

30

Figure 2 Profiles of effluent nitrogen concentrations and COD in four modified PTBR system. (A) S1-Control with Chamber 1 aerobic and Chamber 2 anaerobic, (B) S2-AS with Chamber 1 aerobic and Chamber 2 anaerobic, (C) S3-SD3O with Chamber 1 aerobic and Chamber 2 aerobic, (D) S4-SD3OA with Chamber 1 aerobic and Chamber C2 anaerobic.

31

Figure 3 The nitrogen removal characteristics in Chamber 1 and 2 of S4-SD3OA/PBTR. The effluent concentrations of ammonium (A), accumulated nitrite (B) and nitrate (C) and total nitrogen (D) from Chamber 1 and 2,respectively.

32

Figure 4 The habitat population (A) and relative abundances (B) of KSND bacteria in PTBR.

33

Figure 5 Nitrogen removal performance of KSND in DO and carbon source defined conditions. Nitrification and denitrification efficiency under different DO (A) and carbon sources (B) in various carbon sources. (C) The ammonium and TN removal of KSND in PTBR under DO-control conditions. (D) Expression analysis of pathway genes of nitrogen metabolism under aerobic and anaerobic conditions. The relative fold changes are represented as a ratio of the targeted mRNA expression at anaerobic condition.

34

Figure 6 The profile for the low C/N wastewater treatment with KSND in the purification tank reactors (PTBR).

35

Highlights

1 Alternating DO levels in PTBR achieved 95% TN removal of domestic sewage by KSND.

2 Anaerote adversely affected nitrification of KSND, but conducive to denitrification.

3 Transcription analysis revealed correlation between DO level and genes expression.

4 With two-stage DO control, KSND removed 88% TN from low-C/N domestic sewage in 180-d.

36

37