Accepted Manuscript

Effect of reflux ratio on nitrogen removal in a novel upflow microaerobic sludgereactor treating piggery wastewater with high ammonium and low COD/TNratio: efficiency and quantitative molecular mechanism

Jia Meng, Jiuling Li, Jianzheng Li, Kaiwen Deng, Jun Nan, Pianpian Xu

PII: S0960-8524(17)31148-3DOI: http://dx.doi.org/10.1016/j.biortech.2017.07.052Reference: BITE 18472

To appear in: Bioresource Technology

Received Date: 25 May 2017Revised Date: 6 July 2017Accepted Date: 9 July 2017

Please cite this article as: Meng, J., Li, J., Li, J., Deng, K., Nan, J., Xu, P., Effect of reflux ratio on nitrogen removalin a novel upflow microaerobic sludge reactor treating piggery wastewater with high ammonium and low COD/TNratio: efficiency and quantitative molecular mechanism, Bioresource Technology (2017), doi: http://dx.doi.org/10.1016/j.biortech.2017.07.052

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

1

Effect of reflux ratio on nitrogen removal in a novel upflow

microaerobic sludge reactor treating piggery wastewater with high

ammonium and low COD/TN ratio: efficiency and quantitative

molecular mechanism

Jia Meng1, Jiuling Li

2, Jianzheng Li

*,1, Kaiwen Deng

1, Jun Nan

1, Pianpian Xu

1

1 School of Municipal and Environmental Engineering, State Key Laboratory of Urban

Water Resource and Environment, Harbin Institute of Technology, 73 Huanghe Road,

Harbin 150090, P.R. China

2 Advanced Water Management Centre, The University of Queensland, St Lucia,

Brisbane, QLD, 4072, Australia

*Corresponding Authors. Phone: +86 (451)-86283761; Fax: +86 (451)-86283761;

E-mail: ljz6677@163.com (J. Li)

2

Abstract: A novel upflow microaerobic sludge reactor (UMSR) was constructed to treat

manure-free piggery wastewater with high NH4+-N and low COD/TN ratio. In the light

of the potential effect of effluent reflux ratio (RR) on nitrogen removal, performance of

the UMSR was evaluated at 35°C and hydraulic retention time 8 h with RR decreased

from 45 to 25 by stages. A COD, NH4+-N and TN removal of above 77.1%, 80.0% and

86.6%, respectively, was kept with a RR over 35. To get an effluent of TN not more

than 80 mg/L with a TN load removal above 0.88 kg/(m3·d), the RR should be at least

34. Real-time quantitative polymerase chain reaction of functional bacteria revealed that

the RR of less than 34 stimulated ammonium oxidation but badly inhibited anammox,

the dominant nitrogen removal pathway, resulting in the remarkable decrease of

nitrogen removal in the reactor.

Key words: piggery wastewater; microaerobic process; nitrogen removal; reflux ratio;

molecular mechanism

3

1 Introduction

The increasing demand for pork has stimulated a continual expansion of

pig-breeding, resulting in increasing discharge of piggery wastewater from large-scale

pig farms, especially in developing countries such as China. If not disposed properly, it

would result in significant health and environmental consequences (Zhao et al., 2014).

The strength of piggery wastewater varies greatly due to the manure collection methods

(Wang and Guo, 2009). As a traditional method, urine-free manure collection is widely

used and the flushed wastewater of piggery is defined as manure-free piggery

wastewater (MFPW) (Zhao et al., 2014). This MFPE is characterized by high

ammonium (NH4+-N) and low ratio of chemical oxygen demand (COD) to total nitrogen

(TN), posing a challenge to traditional nitrification-denitrification processes for nitrogen

removal (Zhang et al., 2016). Anaerobic-aerobic combined process has been developed

for nitrogen removal from sewage and proved to be robust (Bernet et al., 2000; Canals

et al., 2013). The combined process could also be used for piggery wastewater treatment,

despite of the costly requisite for physical chemistry pretreatment to get a feasible

COD/TN ratio for biological nitrogen removal (Bernet et al., 2000). It is essential to

develop a novel treatment processes that is more efficient and economical for treating

MFPW.

Microaerobic condition is regarded as a transitional state between aerobic and

anaerobic with a dissolved oxygen (DO) ranging from 0.3 to 1.0 mg/L (Zheng and Cui

2012; Zitomer 1998). Previously, a novel upflow microaerobic sludge reactor (UMSR)

was developed to treat MFPW (Meng et al., 2015). Ammonia-oxidizing bacteria (AOB),

nitrite oxidizing bacteria (NOB), heterotrophic denitrifiers, autotrophic denitrifiers and

phosphate accumulating organisms were found to coexist in the UMSR, and their

4

synergistic action improved microaerobic process in the synchronous removal of COD,

NH4+-N, TN and total phosphorus (TP) (Meng et al., 2016). The microaerobic condition

in the UMSR was achieved by recirculating part of the effluent that was aerated to a DO

of about 3 mg/L (Meng et al., 2015). Obviously, reflux ratio (RR) plays a crucial role in

controlling the microaerobic condition in the reactor. In fact, the microbial community

structure can be remarkably affected by the change of RR, which would result in a

variation in pollutant removal efficiency (Jin et al., 2012). An excessive RR would not

only restrict anaerobic denitrifiers due to the increased DO, but also increase the

treatment cost (Yuan et al., 2002). On the other hand, a RR too low would lead to an

insufficient DO for NH4+-N oxidation, also resulting in a bad TN removal in biological

nitrogen removal processes (He et al., 2006). Therefore, a feasible RR is essential for

the microaerobic treatment system to enhance TN removal and cut down the treatment

cost. However, few researches can be found to investigate the effect of RR on

microaerobic processes.

In the present research, the UMSR was continually performed for another 176 days,

by stages, to evaluate the effect of RR on the synchronous removal of COD, NH4+-N,

TN and TP. To understand the mechanism for biological nitrogen removal in the reactor,

functional microbial populations under different RR were investigated using Illumina

Miseq platform, while the nitrogen removal pathways were also analyzed based on the

absolute abundances of functional bacteria identified by real-time quantitative

polymerase chain reaction (qPCR).

2 Materials and methods

2.1 Microaerobic treatment system

5

The lab-scale UMSR process, a microaerobic treatment system, was illustrated in

Fig. 1. The UMSR was a 0.5-metre-high plexiglass column with a working volume of

4.9 L, while with a 0.5 L circular cone attached to the bottom. A 3 L solid-liquid-gas

separator was designed on the top of the column with the off-gas discharge after a water

lock. Four sampling ports at 10 cm interval from each other were allocated along the

vertical height of the column. The temperature in the reactor was kept at 35±1°C by a

temperature controller (Jetten et al., 2001). The effluent was collected by a 10 L tank, in

which part of the effluent was aerated to a DO of about 3 mg/L and then recirculated

into the reactor from the bottom.

2.2 Wastewater and inoculum

The raw MFPW was collected from a local pig breeding farm in Harbin, China.

The wastewater quality fluctuated following the changes in breeding seasonality with a

COD, NH4+-N, nitrite (NO2

--N), nitrate (NO3--N), TN, TP and pH ranging from 241-459,

237.4-343.3, 0.0-2.2, 0.0-5.8, 285.4-458.2, 6.4-21.4 mg/L and 7.0-8.4, respectively.

Obviously, the high concentration of NH4+-N contributed mainly to the TN in the

wastewater.

The seed sludge inoculated in the UMSR was anaerobic sludge collected from an

upflow anaerobic sludge blanket (UASB) treating swine wastewater in the State Key

Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology,

China. The initial biomass in the UMSR was 5.68 g/L in terms of mixed liquor

suspended solids (MLSS), i.e. 1.95 g/L in terms of mixed liquor volatile suspended

solids (MLVSS). The UMSR previously started up and performed for 200 days (Meng

et al., 2015). At the start of the present research, MLVSS in the UMSR was 3.5 g/L with

6

a MLVSS/MLSS ratio of 0.66. The sludge was dominated by flocs, among which some

small-sized granular sludge (<2 mm) could also be observed.

2.3 Operation of the UMSR

In the present research, the UMSR was continued to be operated for another 176

days in a 4-stage procedure with a consistent hydraulic retention time (HRT) of 8 h.

Quality of the raw MFPW and the other controlling parameters in the 4 stages are

illustrated in Table 1. The influent and effluent were sampled daily for analysis.

Activated sludge in the reactor was sampled at the end of each stage for measurement of

biomass (MLVSS and MLSS) and analysis of high-throughput pyrosequencing and

qPCR.

2.4 Analytical methods

COD, NH4+-N, NO2

--N, NO3--N, TP and biomass (MLSS and MLVSS) were

detected by closed reflux titrimetric method, Nessler's reagent colorimetric method,

colorimetric method, ultraviolet spectrophotometric screening method,

vanadomolybdophosphoric acid colorimetric method and gravimetric method,

respectively, according to the Standard Methods (APHA, 2002). DO and pH were

measured with a dissolved oxygen meter (Taiwan Hengxin, AZ 8403) and a pH meter

(Switzerland Mettler Toledo, DELTA320), respectively. TN was measured on a total

nitrogen analyzer (Germany Analytikjena Multi, N/C 2100S).

2.5 High-throughput pyrosequencing

Sludge sampled from the UMSR at each end of the 4 stages (Table 1) was named

7

S1, S2, S3, and S4 in sequence, and the total DNA of the samples were extracted,

respectively, using the Bacteria DNA Isolation Kit Components (MOBIO) according to

the manufacturer’s instruction. Bacterial 16S rRNA genes for the V3-V4 regions were

amplified with primer pairs 341f (5′-CCTACGGGAGGCAGCAG-3′) and 805r

(5′-GACTACHVGGGTATC TAATCC-3′) (Herlemann et al., 2011; Hugerth et al., 2014).

Composition of the PCR products of V4 region were determined by pyrosequencing on

the Illumina Miseq sequencer platform (Shanghai Sangon, China).

2.6 qPCR analysis

To determine the amount of nitrifiers and denitrifiers in the sludge samples, seven

independent qPCR assays were conducted in a real-time PCR system (USA Foster, ABI

7500) by quantifying the 16S rRNA of total bacteria, anammox bacteria, Nitrobacter

spp., Nitrospira spp., and gene of amoA, nirS and narG, respectively. The primers used

are illustrated in Table 2. Each of the PCR reactions was performed in a 20 µL system

loaded with 2 µL template DNA (at 22 - 88 ng/µL), 0.4 µL forward and 0.4 µL reverse

primers (with the same 10 µM), 10 µL SYBR Premix (Applied Biosystems), and 7.2 µL

PCR-grade sterile water. The PCR program was as follows: 3 min at 95°C followed by

40 cycles each including 15 s at 95°C, 20 s at 55°C and 30s at 72°C in sequence. The

PCR was controlled by a reaction without template and each of the qPCR was

performed in triplicate.

All of the plasmid standards of the 16S rRNAs and the functional genes were

constructed by Sangon Biotechnology Incorporated Company (Shanghai, China). The

standard curves were prepared by ten-fold serial dilutions of the plasmid DNA with

known copy numbers, and the correlation coefficients were all above 0.997. The PCR

8

amplification efficiency for the genes was between 90% and 110%.

2.7 Data analysis

The absolute abundance of the seven functional genes (i.e. total bacterial,

anammox, Nitrobacter spp., Nitrospira spp., amoA, nirS and narG) obtained from each

of the triplicate tests were averaged, respectively, and then used as the basic candidate

variables in nonlinear regression analyses (SPSS 22, USA) for the correlations with

nitrogen transformation rates. Standard deviations were calculated using the three

replicate measurements and plotted as error bars for analyzing data variations and

measurement errors.

3 Results and discussion

3.1 Pollutant removal in the UMSR

After the previous 200-day performance (Meng et al., 2015), the UMSR was

continued to be operated for another 176 days following a progressive decrease in RR

(Table 1) with a consistent HRT of 8 h at 35±1°C. The performance in pollutant

removal within the 176 days was illustrated in Fig. 2. The results showed that, in stage 1

with the RR 45, NH4+-N removal in the reactor exhibited a rapid decrease phase within

the first 6 days, and then increased gradually in fluctuation thereafter and reached a

steady phase within the last 13 days (55th

- 67th

day) (Fig. 2A). The average NH4+-N

removal in the steady phase reached 86.2% with a residual of about 40.2 mg/L in

effluent. The NH4+-N removal was remarkably affected and decreased sharply to 37.7%

when the RR decreased to 35 in the first day of stage 2, but rebounded and reached a

new steady state since the 96th day with the removal and effluent concentration

9

averaging 77.1% and 67.4 mg/L, respectively. When RR was further decreased to 30 in

stage 3 and 25 in stage 4, the same variation pattern of NH4+-N removal with time were

also observed. The NH4+-N removal in the steady phase of stage 3 (130th - 141th days)

and stage 4 (164th - 176th days) was about 73.1% and 53.3%, with a NH4+-N of about

81.3 and 137.8 mg/L in effluent, respectively. Obviously, the decrease of RR had badly

affected the ammonium oxidation in the reactor.

Variation of TN removal (Fig. 2B) in the UMSR found the same trend as that of

NH4+-N removal, with the steady states obtained synchronously. In the steady phase of

stage 1, 2, 3 and 4, the TN removal averaged 87.2%, 80.0%, 75.4% and 60.9%, with an

effluent concentration of about 45.8, 73.2, 89.2 and 140.6 mg/L, respectively. The

average TN load removal (TNLR) in the four steady phases was calculated to 0.94, 0.88,

0.82 and 0.66 kg/(m3·d), respectively. The inhibited NH4+-N oxidization by the

decreased RR (Fig. 2A) was considered as the main reason for the decrease of TN

removal because no obvious NOx--N (including NO2

--N and NO3--N) acclimation was

found in the four steady phases. However, NO2--N accumulation could be found in the

first days all of the stages. In stage 1, for example, NO2--N accumulated to 50.4 mg/L on

the 31th

day, but decreased rapidly with time and no observable accumulation was

detected in the steady phase. The same phenomenon was also observed in the next 3

stages, though the accumulation of NO2--N decreased slightly with the stages. This

performance accorded with the cultural characteristics of anammox bacteria (van

Dongen et al., 2001), indicating anammox was an important approach to nitrogen

removal in the UMSR.

All through the 176-day operation, COD removal in the UMSR was above 65%

with an effluent COD less than 125 mg/L (Fig. 2C). It was found that COD removal in

10

the steady phases of stage 1 and stage 3 were almost the same (77.9% and 74.5%,

respectively), though the RR (45 and 30, respectively) was the remarkable difference.

The COD removal in the steady phases of stage 2 (RR 35) and stage 4 (RR 25) were

also similar (86.6% and 83.8%, respectively). It was noticed that COD/TN ratio in stage

1 (0.84) and 3 (0.82) were almost the same, as well as those in stage 2 (0.95) and 4

(0.94). The results suggested that COD/TN ratio had a stronger effect than RR on the

COD removal in the UMSR. Within the 176-day operation of the UMSR, no excess

sludge was discharged and the biomass in terms of MLVSS at each end of the four

stages was 5.20, 7.15, 4.11, and 6.26 g/L, respectively. Evidently, the higher COD/TN

ratio in stage 2 and stage 4 had stimulated the growth of chemoheterotrophic bacteria,

resulting in the better COD removal in the reactor (Li et al., 2016).

The stimulated growth of biomass was favorable for TP removal in the UMSR. As

shown in Fig. 2D, TP removal in the second steady phase reached 83.5% with the

highest biomass (7.15 g/L). With a lower biomass in stage 1 (5.20 g/L), stage 3 (4.11

g/L) and stage 4 (6.26 g/L), the TP removal decreased to 46.8%, 52.5% and 66.6% in

their steady phases, respectively. The effluent TP in the four steady phases averaged 5.1,

2.1, 4.4 and 5.0 mg/L, respectively. Observably, TP removal in the UMSR was

positively correlated with the biomass, indicating sludge growth was the most important

approach for TP removal in the microaerobic treatment system.

Discharge standards for various wastewater have been promulgated in many

countries. According to the Discharge Standard of Pollutants for Livestock and Poultry

Breeding issued by the Ministry Environmental Protection of China (MEPPRC, 2001),

the discharge standard for NH4+-N should not be more than 80 mg/L. All through the

176-day operation of the UMSR, the effluent COD and TP could meet the standard

11

required very well, while the effluent NH4+-N and TN could meet the standard only with

a RR not less than 35 (Fig. 2). To get an economical RR for treating manure-free

piggery wastewater, the influence of RR on nitrogen removal (NH4+-N and TN) and

TNLR in the four steady phases were analyzed as illustrated in Fig. 3, in which an

inhibitory threshold for nitrogen removal was defined when TN removal dropped to

around 78.0%, with an effluent TN more than 80 mg/L. Based on the quadratic

polynomial regression equations (Fig. 3), the RR threshold should be 34, at which a

TNLR of 0.88 kg/(m3·d) could be obtained in the UMSR.

3.2 Diversity of microbial community and synchronous pollutant removal

To understand the biological mechanism of synchronous removal of COD, NH4+-N,

TN and TP in the UMSR, DNA extraction and PCR amplification of the sludge sampled

at the end of each stage (Table 1) were carried out. As shown in Table 3, 41342, 36383,

37906 and 19515 reads were obtained from the sludge sampled from stage 1 (S1), stage

2 (S2), stage 3 (S3) and stage 4 (S4), respectively. The identified operational taxonomic

units (OTUs) in the four sludge samples reached 5632, 5211, 6542 and 1459,

respectively. The results indicated that better species diversity and richness had been

kept in the UMSR, though the Chao1 and ACE index, as well as the Shannon index,

varied with the decrease of RR.

The microbial community structure and composition under different RR were

characterized by phylogenetic classification. As shown in Fig. 4A, Proteobacteria was

the most predominant phylum, followed by Firmicutes and Bacteroidetes, all the time in

the UMSR. Fig. 4B indicated that the main subgroups of phyla Proteobacteria in S1, S2

and S4 were the same β-Proteobacteria, while γ-Proteobacteria dominated the S3. It is

12

reported that β- and γ-Proteobacteria play a critical part in the biological removal of

nitrogen and phosphorus (Yang et al., 2014).

Phylogenetic classification of the four sludge samples at genus level (Fig. 5)

showed that the microbial community structure in the UMSR was distinguished by the

difference of RR. Even so, heterotrophic bacteria (including aerobic and anaerobic

heterotrophic bacteria), nitrifiers (including AOB and NOB), and denitrifiers (including

heterotrophic and autotrophic denitrifiers) were found to coexist and flourish in the

microaerobic system, resulting in the synchronous removal of COD, NH4+-N and TN.

However, no genera related to anammox had been identified from the activated sludge

samples by high-throughput pyrosequencing (Fig. 5). Thus, the nitrogen removal

mechanism had to be further investigated.

3.3 Nitrogen removal pathways

Nitrification-denitrification, shortcut nitrification-denitrification and anammox are

well known as the approaches to the TN removal in biological denitrification processes

(Xie et al., 2016). The reduction of nitrate to nitrogen gas (N2 or N2O) needs more

carbon source to supply electron than that of nitrite reduction, and nitrous oxide (N2O)

production would need less carbon source than the production of N2. Theoretically,

without assimilation, 2.86 g or 1.71 g COD is needed to reduce 1 g NO3--N or 1 g

NO2--N to N2, respectively (Li et al., 2012; Saggar et al., 2013). For the reduction of 1 g

NO3--N or 1 g NO2

--N to N2O, 2.29 g or 1.14 g COD would be consumed respectively.

Based on the cooperation of the theoretical COD/TN ratio with the measured

CODremoved/TNremoved ratio in the UMSR, the theoretical TN removal efficiencies via

various pathways in the microaerobic system were calculated and illustrated in Table 4.

13

The results showed that the TN removal in the UMSR was remarkably lower than that

of the actual efficiency even all of the removed COD were used for the denitrification

via nitrite reduction or nitrate reduction to N2. Even for nitrite reduction to N2O, the

least carbon-source-consumption denitrification, the theoretical TN removal was also

observably lower (stage 1, stage 2 and stage 3) or similar (stage 4) to that of the actual

efficiency in the UMSR. In the steady phase of stage 1, for example, the theoretical TN

removal efficiency via N2O production by nitrite reduction, N2 production by nitrite and

nitrate reduction were 59.0%, 39.1% and 23.5%, respectively, considerably lower than

the actual 87.2%. Hence, the autotrophic anammox must be one of the most important

pathways to nitrogen removal in the UMSR.

To confirm the existence of anammox and understand the nitrogen removal

mechanism in depth, the absolute abundances of amoA, nirS, narG and Nitrobacter 16S

rRNA, Nitrospira 16S rRNA, anammox 16S rRNA and total bacteria 16S rRNA in the

sampled sludge (S1, S2, S3 and S4) were quantified using qPCR assays and the results

were illustrated in Fig. 6. Among the seven functional genes, amoA was responsible for

NH4+-N conversion and NO2

--N production, while nirS for NO2

--N reduction, narG for

NO3--N reduction, both Nitrobacter 16S rRNA and Nitrospira 16S rRNA for NO2

--N

oxidation, and anammox 16S rRNA for NH4+-N and NO2

--N consumption. Absolute

abundances of the seven functional genes were used as basic candidate variables for

nonlinear regression analyses (SPSS 22, USA) to get the correlation with nitrogen

conversion rates. Then the following equations are obtained.

NH4+-N=-4.706(amoA/(anammox + Nitrobacter + Nitrospira)) + 0.780 (1)

TN=-2.881(amoA/(anammox + Nitrobacter + Nitrospira)) + 23.625(nirS +

narG)/Bacterial + 0.831 (2)

14

The R2 in equation (1) and (2) was 0.977 and 0.989, respectively. It was found that

the NH4+-N removal had a negative correlation with amoA/(anammox + Nitrobacter +

Nitrospira). The NH4+-N regression model suggested that NH4

+-N removal in the

UMSR was not only influenced by amoA gene but also by anammox, Nitrobacter and

Nitrospira 16S rRNA, and the increased amoA would reduce the NH4+-N removal in the

UMSR. TN removal in the UMSR was related to denitrification, too (Li et al., 2016).

Both nirS/(total bacteria) and narG/(total bacteria), involved in NO2--N and NO3

--N

reduction respectively, were positively correlated with TN removal in the UMSR

(equation (2)). Above all, the qPCR assays indicated that shortcut nitrification,

anammox, and denitrification coexisted in the UMSR, and their combined action

resulted in the better TN removal without supply of extra carbon source.

3.4 Quantitative molecular mechanism of pollutant removal

As illustrated in Fig. 5, though the RR decreased from 45 to 25 by stages, the

heterotrophic bacteria were the foremost populations all along the 176-day performance

of the UMSR. The flourishing heterotrophic population resulted in the better COD

removal in the UMSR all through the four operation stages (Fig. 2C). Among the

heterotrophic bacteria, Variovorax (Prasad and Suresh, 2012), Novosphingobium

(Kertesz and Kawasaki, 2010), Stenotrophomonas (Binks et al., 1995) and

Altererythrobacter (Park et al., 2011) are strict aerobic bacteria, and the total proportion

decreased from 4.66% in S1 to 2.27% in S4 with the decreased RR from 45 to 25 stage

by stage. Obviously, it was the decreased RR that had reduced the DO concentration and

resulted in the decline of aerobic heterotrophic bacteria in the UMSR.

It was interesting to find that the decrease of RR enriched AOB and NOB in the

15

reactor. With the RR 45, genera related to AOB in S1 were Sphingomonas (Zheng et al.,

2005), unclassified_Nitrosomonadaceae (Prosser et al., 2014), Nitrosococcus (Stein et

al., 2013), Nitrosomonas (Bock et al., 1995) and Nitrosospira (Hollibaugh et al., 2002),

with a total proportion of 1.07% (Fig. 5A). When the RR was decreased to 35 and then

to 30, the total proportion of AOB in S2 (Fig. 5B) and S3 (Fig. 5C) was enhanced to 1.53%

and 1.43%, respectively. Meanwhile, the total proportion of NOB was also improved

from 0.17% in S1 to 0.25% in S2 and then to 0.43% in S3. The results showed that the

microaerobic condition in the UMSR was feasible for the growth of aerobic autotrophic

bacteria. However, the total proportion of NOB was remarkably decreased to 0.08%

with the reduction of RR to 25 in stage 4, though the proportion of AOB was further

increased to 2.91% at the same time (Fig. 5D).

As aerobic polyphosphate accumulating bacteria (Barker and Dold, 1996),

Acinetobacter was also found in the UMSR. The proportion in total bacteria was

remarkably decreased from 0.84% to 0.05% by stages, indicating Acinetobacter was

more sensitive to DO. However, as shown as Fig. 2D, TP removal in the four steady

phases of the UMSR was 46.8%, 83.5%, 52.5% and 66.6%, respectively. Obviously, the

TP removal in the reactor was not positively correlated with the abundance of

Acinetobacter, but with the biomass as expressed in Section 3.1.

In fact, aerobic bacteria in the UMSR had played an important role in consuming

the limited DO in forming an anoxic environment that allowed anaerobic bacteria

located deep in the activated sludge flocs (Meng et al., 2016). As illustrated in Fig. 5,

heterotrophic denitrification was one of the most important approaches to nitrogen

removal in the UMSR. In stage 1 with the RR 45, the proportion of heterotrophic

denitrifiers in S1 was 10.44% (Fig. 5A). With the RR decreased to 35, 30 and 25 by

16

stages (Table 1), the total oxygen supplement to the microaerobic system decreased and

resulted in an enhancement of heterotrophic denitrifiers to 16.01%, 10.76% and 15.54%,

respectively (Fig. 5B, 5C and 5D). In addition to the heterotrophic denitrifiers,

Thiobacillus as autotrophic denitrifiers (Claus and Kutzner, 1985) was also observed in

the microaerobic system. Its proportion decreased from 0.83% to 0.68% stage by stage,

positively correlated with the decrease of TN removal (Fig. 2B).

qPCR analysis of the four sampled activated sludge (Fig. 6) showed that: 1) the

total bacteria in the UMSR decreased gradually from 7.87×105 copies/ngDNA in S1 to

2.35×105 copies/ngDNA in S4 with the decreased RR from 45 to 25 by stages, 2) the

amoA gene in S4 was 5.6 times as that of 5.28×101 copies/ngDNA in S1, 3) Nitrospira

16S rRNA decreased obviously from 2.10×102 copies/ngDNA in S1 to only 1.10

copies/ngDNA in S4, 4) Nitrobactor 16S rRNA also showed a decreased trend with 10

copies/ngDNA in S4. With the decrease of RR in the UMSR, both absolute abundance

variation of amoA and the total 16S rRNA of Nitrobacter and Nitrospira were

consistent with the change of AOB and NOB revealed by the high throughput (Fig. 5). It

is known that AOB has a better oxygen affinity than NOB (Blackburne et al., 2008).

Thus, it could be understood that the ratio of amoA/(Nitrobacter + Nitrospira) increased

with the decrease of RR. It was found that anammox 16S rRNA was the most abundant

nitrogen functional gene (Fig. 6) with the copies of 6.00×103, 4.19×10

3, 5.81×10

3 and

4.44×103 in S1, S2, S3, S4, respectively. This result indicated that anammox was the most

important approach for NH4+-N and TN removal in the UMSR.

Since nirS gene was more abundant than counterpart nirK gene in wastewater

treatment system (Kim et al., 2011; Zhi et al., 2015), nirS along with narG were

selected as the denitrifying genes in the present study. Because the dominant anammox

17

would consume the most NH4+-N in the MFPW, only a small amount of NOx

--N was

produced (Fig. 2B). Without enough NOx--N as substrate for denitrifiers, both narG and

nirS genes exhibited a downward trend in the UMSR with the decrease of RR by stages

(Fig. 6). Therefore, anammox was identified as the main mechanism for nitrogen

removal in the UMSR, and the decrease of RR had a significant negative effect on

anammox bacteria, resulting in a less removal for both NH4+-N and TN.

4 Conclusion

RR had a remarkable effect on the nitrogen removal in the UMSR. The TN

removal presented a decreasing trend with the decrease of RR from 45 to 25 by stages.

To get an effluent TN of not more than 80 mg/L with a TN load removal above 0.88

kg/(m3·d), the RR should be at least 34. The RR of less than 34 badly inhibited the

growth of anammox bacteria. Though the lower RR enriched AOB, the nitrogen

removal remarkably decreased in the reactor due to the inhibited anammox that was the

dominant nitrogen removal pathway.

Acknowledgements

The authors gratefully acknowledge the China Postdoctoral Science Foundation

funded project (Grant No. 2017M611376), the National Natural Science Foundation of

China (Grant No. 51478141) and the State Key Laboratory of Urban Water Resource

and Environment, Harbin Institute of Technology (Grant No. 2016DX06) for valuable

financial support.

References

18

1. APHA (American Public Health Association), 2005. Standard Methods for the

Examination of Water and Wastewater (19th edition). Washington DC, USA.

2. Barker, P.S., Dold, P.L., 1996. Denitrification behaviour in biological excess

phosphorus removal activated sludge systems. Water Res. 30, 769-780.

3. Bernet, N., Delgenes, N., Akunna, J.C., Delgenes, J., Moletta, R., 2000. Combined

anaerobic-aerobic SBR for the treatment of piggery wastewater. Water Res. 34,

611-619.

4. Binks, P.R., Nicklin, S., Bruce, N.C., 1995. Degradation of hexahydro-1, 3,

5-trinitro-1, 3, 5-triazine (RDX) by Stenotrophomonas maltophilia PB1. Appl.

Environ. Microbiol. 61, 1318-1322.

5. Blackburne, R., Yuan, Z., Keller, J., 2008. Partial nitrification to nitrite using low

dissolved oxygen concentration as the main selection factor. Biodegradation 19,

303-312.

6. Bock, E., Schmidt, I., Stüven, R., Zart, D., 1995. Nitrogen loss caused by

denitrifying Nitrosomonas cells using ammonium or hydrogen as electron donors

and nitrite as electron acceptor. Arch. Microbiol. 163, 16-20.

7. Braker, G., Fesefeldt, A., Witzel, K.P., 1998. Development of PCR primer systems

for amplification of nitrite reductase genes (nirK and nirS) to detect denitrifying

bacteria in environmental samples. Appl. Environ. Microbiol. 64, 3769-3775.

8. Canals, O., Salvadó, H., Auset, M., Hernández, C., Malfeito, J., 2013. Microfauna

communities as performance indicators for an A/O shortcut biological nitrogen

removal moving-bed biofilm reactor. Water Res. 47, 3141-3150.

9. Claus, G., Kutzner, H.J., 1985. Physiology and kinetics of autotrophic denitrification

by Thiobacillus denitrificans. Appl. Microbiol. Biotechnol. 22, 283-288.

19

10. Dionisi, H.M., Layton, A.C., Harms, G., Gregory, I.R., Robinson, K.G., Sayler, G.S.,

2002. Quantification of Nitrosomonas oligotropha-like ammonia-oxidizing bacteria

and Nitrospira spp. from full-scale wastewater treatment plants by competitive PCR.

Appl. Environ. Microbiol. 68, 245-253.

11. Graham, D.W., Knapp, C.W., Van Vleck, E.S., Bloor, K., Lane, T.B., Graham, C.E.,

2007. Experimental demonstration of chaotic instability in biological nitrification.

ISME J. 1, 385-393.

12. He, L. S., Liu, H. L., Xi, B. D., Zhu, Y. B., 2006. Effects of effluent recirculation in

vertical-flow constructed wetland on treatment efficiency of livestock wastewater.

Water Sci. Technol. 54, 137-146.

13. Herlemann, D.P., Labrenz, M., Jürgens, K., Bertilsson, S., Waniek, J.J., Andersson,

A.F., 2011. Transitions in bacterial communities along the 2000 km salinity gradient

of the Baltic Sea. ISME J. 5, 1571-1579.

14. Hollibaugh, J.T., Bano, N., Ducklow, H.W., 2002. Widespread distribution in polar

oceans of a 16S rRNA gene sequence with affinity to Nitrosospira-like

ammonia-oxidizing bacteria. Appl. Environ. Microbiol. 68, 1478-1484.

15. Hugerth, L.W., Wefer, H.A., Lundin, S., Jakobsson, H.E., Lindberg, M., Rodin, S.,

Engstrand, L., Andersson, A.F., 2014. DegePrime, a program for degenerate primer

design for broad-taxonomic-range PCR in microbial ecology studies. Appl. Environ.

Microbiol. 80, 5116-5123.

16. Jetten, M.S., Wagner, M., Fuerst, J., van Loosdrecht, M., Kuenen, G., Strous, M.,

2001. Microbiology and application of the anaerobic ammonium oxidation

(anammox) process. Curr. Opin. Biotechnol. 12, 283-288.

17. Jin, R.C., Yang, G.F., Ma, C., Yu, J.J., Zhang, Q.Q., Xing, B.S., 2012. Influence of

20

effluent recirculation on the performance of Anammox process. Chem. Eng. J. 200,

176-185.

18. Kertesz, M.A., Kawasaki, A., 2010. Hydrocarbon-Degrading Sphingomonads:

Sphingomonas, Sphingobium, Novosphingobium, and Sphingopyxis. in: Handbook

Of Hydrocarbon And Lipid Microbiology, (Ed.) K. Timmis, Springer Berlin

Heidelberg, pp. 1693-1705.

19. Kim, Y.M., Cho, H.U., Lee, D.S., Park, D., Park, J.M., 2011. Influence of

operational parameters on nitrogen removal efficiency and microbial communities

in a full-scale activated sludge process. Water Res. 45, 5785-5795.

20. Li, H., Zhou, S., Ma, W., Huang, G., Xu, B., 2012. Fast start-up of ANAMMOX

reactor: operational strategy and some characteristics as indicators of reactor

performance. Desalination 286, 436-441.

21. Li, J., Meng, J., Li, J., Wang, C., Deng, K., Sun, K., Buelna, G., 2016. The effect

and biological mechanism of COD/TN ratio on nitrogen removal in a novel upflow

microaerobic sludge reactor treating manure-free piggery wastewater. Bioresour.

Technol. 209, 360-368.

22. López-Gutiérrez, J.C., Henry, S., Hallet, S., Martin-Laurent, F., Catroux, G.,

Philippot, L., 2004. Quantification of a novel group of nitrate-reducing bacteria in

the environment by real-time PCR. J. Microbiol. Meth. 57, 399-407.

23. Meng, J., Li, J., Li, J., Antwi, P., Deng, K., Wang, C., Buelna, G., 2015. Nitrogen

removal from low COD/TN ratio manure-free piggery wastewater within an upflow

microaerobic sludge reactor. Bioresour. Technol. 198, 884-890.

24. Meng, J., Li, J., Li, J., Sun, K., Antwi, P., Deng, K., Wang, C., Buelna, G., 2016.

Efficiency and bacterial populations related to pollutant removal in an upflow

21

microaerobic sludge reactor treating manure-free piggery wastewater with low

COD/TN ratio. Bioresour. Technol. 201, 166-173.

25. MEPPRC (Ministry Environmental Protection of People’s Republic of China), 2001.

Discharge Standard of Pollutants for Livestock and Poultry Breeding. Beijing,

China.

26. Muyzer, G., De Waal, E.C., Uitterlinden, A.G., 1993. Profiling of complex microbial

populations by denaturing gradient gel electrophoresis analysis of polymerase chain

reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59,

695-700.

27. Park, S.C., Baik, K.S., Choe, H.N., Lim, C.H., Kim, H.J., Ka, J.O., Seong, C.N.,

2011. Altererythrobacter namhicola sp. nov. and Altererythrobacter aestuarii sp.

nov., isolated from seawater. Int. J. Syst. Evol. Micr. 61, 709-715.

28. Prasad, B., Suresh, S., 2012. Biodegradation of phthalate esters by Variovorax sp.

Apcbee Procedia 1, 16-21.

29. Prosser, J.I., Head, I.M., Stein, L.Y., 2014. The Family Nitrosomonadaceae. in: The

Prokaryotes, Springer, pp. 901-918.

30. Rotthauwe, J.H., Witzel, K.P., Liesack, W., 1997. The ammonia monooxygenase

structural gene amoA as a functional marker: molecular fine-scale analysis of

natural ammonia-oxidizing populations. Appl. Environ. Microb. 63, 4704-4712.

31. Saggar, S., Jha, N., Deslippe, J., Bolan, N., Luo, J., Giltrap, D., Kim, D.G., Zaman,

M., Tillman, R., 2013. Denitrification and N2O:N2 production in temperate

grasslands: processes, measurements, modelling and mitigating negative impacts.

Sci. Total Environ. 465, 173-195.

32. Stein, L., Campbell, M., Klotz, M., 2013. Energy-mediated vs. ammonium-regulated

22

gene expression in the obligate ammonia-oxidizing bacterium, Nitrosococcus oceani.

Front. Microbiol. 4, 277.

33. Tsushima, I., Kindaichi, T., Okabe, S., 2007. Quantification of anaerobic

ammonium-oxidizing bacteria in enrichment cultures by real-time PCR. Water Res.

41, 785-794.

34. van Dongen, U., Jetten, M.S., Van Loosdrecht, M., 2001. The

SHARON-ANAMMOX process for treatment of ammonium rich wastewater. Water

Sci. Technol. 44, 153-160.

35. Wang, L., Guo, Z.Y., 2009. The biological fermentation bed raises pigs technical the

application present situation. Livestock and Poultry Industry 3, 006.

36. Xie, G.J., Cai, C., Hu, S., Yuan, Z., 2016. Complete nitrogen removal from synthetic

anaerobic sludge digestion liquor through integrating anammox and denitrifying

anaerobic methane oxidation in a membrane biofilm reactor. Environ. Sci. Technol.

51, 819-827.

37. Yang, S., Guo, W., Chen, Y., Zhou, X., Zheng, H., Feng, X., Yin, R., Ren, N., 2014.

Simultaneous nutrient removal and reduction in sludge from sewage waste using an

alternating anaerobic-anoxic-microaerobic-aerobic system combining

ozone/ultrasound technology. RSC Adv. 4, 52892-52897.

38. Yuan, Z., Oehmen, A., Ingildsen, P., 2002. Control of nitrate recirculation flow in

predenitrification systems. Water Sci. Technol. 45, 29-36.

39. Zhang, Y., Wang, X.C., Cheng, Z., Li, Y., Tang, J., 2016. Effect of fermentation

liquid from food waste as a carbon source for enhancing denitrification in

wastewater treatment. Chemosphere 144, 689-696.

40. Zhao, B., Li, J., Leu, S.Y., 2014. An innovative wood-chip-framework soil infiltrator

23

for treating anaerobic digested swine wastewater and analysis of the microbial

community. Bioresour. Technol. 173, 384-391.

41. Zheng, S., Cui, C., 2012. Efficient COD removal and nitrification in an upflow

microaerobic sludge blanket reactor for domestic wastewater. Biotechnol. Lett. 34,

471-474.

42. Zheng, X.S., Yang, H., Li, D.T., 2005. Change of microbial populations in a

suspended-sludge reactor performing completely autotrophic N-removal. World J.

Microb. Biot. 21, 843-850.

43. Zhi, W., Yuan, L., Ji, G., He, C., 2015. Enhanced long-term nitrogen removal and its

quantitative molecular mechanism in tidal flow constructed wetlands. Environ. Sci.

Technol. 49, 4575-4583.

44. Zitomer, D.H., 1998. Stoichiometry of combined aerobic and methanogenic COD

transformation. Water Res. 32, 669-676.

24

Figure captions

Fig. 1 Schematic representation of the lab-scale UMSR process.

Fig. 2 Removal of NH4+-N (A), TN (B), COD (C) and TP (D) in the UMSR with the

variation of reflux ratio.

Fig. 3 Effect of reflux ratio on nitrogen removal in the UMSR.

Fig. 4 Bacterial community structure of the activated sludge at phylum (A) and the

classes in Proteobacteria (B). S1, S2, S3 and S4 was the sludge sample taken from the

UMSR at the end of stage 1, stage 2, stage 3 and stage 4, respectively.

Fig. 5 Bacterial community structure at genus level of the activated sludge sampled

from the UMSR at the end of the first stage (A), the second stage (B), the third stage (C)

and the fourth stage (D).

Fig. 6 Copies per ngDNA of nitrogen functional genes with the reflux ratios in the

UMSR.

Table captions

Table 1 Stages and operating parameters of the UMSR

Table 2 Primers in the qPCR

Table 3 Diversity of microbial community in the activated sludge sampled from the

UMSR

Table 4 Actual TN removal efficiency in the UMSR comparing with the theoretical

values

25

Figure 1:

26

Figure 2:

0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Rem

ov

al (

%)

TP

(m

g/L

)

Time (d)

Influent

Effluent

Removal

stage 1 stage 2 stage 3

Reflux ratio 45 Reflux ratio 35 Reflux ratio 30

stage 4

Reflux ratio 25

B

C

A

D

0

10

20

30

40

50

60

70

80

90

100

0

50

100

150

200

250

300

350

400

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Rem

ov

al (

%)

NH

4+-N

(m

g/L

)

Time (d)

Influent Effluent Removal

0

10

20

30

40

50

60

70

80

90

100

0

50

100

150

200

250

300

350

400

450

500

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Rem

oval

(%

)

CO

D (

mg

/L)

Time (d)

Influent

Effluent

Removal

0

10

20

30

40

50

60

70

80

90

100

0

50

100

150

200

250

300

350

400

450

500

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

TN

rem

oval

(%

)

NO

x-N

(m

g/L

)

TN

(m

g/L

)

Time (d)

Inf. TN Eff. TN TN removal Eff. nitrite Eff. nitrate

27

Figure 3:

y = -0.096x2 + 8.282x - 92.449

R² = 0.9628

y = -0.0724x2 + 6.334x - 51.429

R² = 0.981

y = -0.0009x2 + 0.0772x - 0.6953

R² = 0.988

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0

10

20

30

40

50

60

70

80

90

100

20 25 30 35 40 45 50

To

tal

nit

rogen l

oad

rem

oval

(kg

/(m

3·d

))

Nit

rog

en r

emov

al (

%)

reflux ratio

Ammonium removal

TN removal

Total nitrogen load removal

28

Figure 4:

S1 S2

S3 S4

Proteobacteria

unclassified

Firmicutes

Bacteroidetes

Chloroflexi

Cyanobacteria

Synergistetes

Planctomycetes

Verrucomicrobia

Euryarchaeota

Chlorobi

Gemmatimonadetes

Nitrospirae

Deferribacteres

Spirochaetae

others

Actinobacteria

Acidobacteria

Candidate_division_TM7

1.30

0.92

46.15

18.92

16.9410.34

1.85

1.17

0.630.58

0.44

1.210.191.30

47.71

19.97

13.377.73

4.57

0.432.40

42.96

22.05

13.9910.07

1.413.24

1.901.85

49.88

16.18

16.157.80

0.261.06

1.123.25

1.12

0

10

20

30α-Proteobacteria

β-Proteobacteria

γ-Proteobacteria

δ-Proteobacteria

ε-Proteobacteria

ζ-Proteobacteria

S1

S2

S3

S4

A

B

29

Figure 5:

Lacto

baci

llus

Pant

oea

Ignav

ibac

terium

Alis

tipes

Bifi

doba

cter

ium

Solit

alea

Klebs

iella

Turicib

acte

r

Bac

illus

Rho

doba

cter

Var

iovo

rax

Nov

osph

ingo

bium

Sten

otro

phom

onas

Paeni

bacillu

s

Vib

rio

Hym

enob

acte

r

Hel

icob

acter

Bac

tero

ides

Stre

ptoc

occu

s

Prot

eini

philu

m

Lon

gilin

ea

Sphi

ngom

onas

uncl

assifie

d_N

itros

omon

adac

eae

Nitr

osom

onas

Nitr

osoc

occu

s

unclas

sified

_Nitr

ospi

nace

ae

Nitr

ospi

ra

unclas

sified

_Nitr

ospi

race

ae

Nitr

obac

ter

Pseud

omon

as

Lim

noba

cter

Den

itrat

isom

a

Citr

obac

ter

Azo

spirill

um

Azo

spira

Are

nim

onas

Flav

obac

terium

Hyp

hom

icro

bium

Arc

obac

ter

Thi

obac

illus

Aci

neto

bact

er

Des

ulfo

bulb

us

Ent

erob

acter

Azo

arcu

s

Cal

othr

ix

Rhi

zobi

um

Proc

hlor

on

Lautrop

ia

Bla

utia

Esc

heric

hia-

Shig

ella

Gem

mob

acte

r

Clo

acib

acterium

Rue

geria

othe

rs

uncl

assified

0

1

2

3

4

20

25

30

Rela

tiv

e ab

un

dance (

%)

Lacto

bacillu

s

Panto

ea

Solit

alea

Met

hano

saeta

Tur

icib

acte

r

Alis

tipes

Petri

mon

as

Bifi

doba

cter

ium

Kle

bsie

lla

Igna

viba

cter

ium

Smith

ella

Bac

illus

Rho

doba

cter

Erysip

elot

hrix

Var

iovo

rax

Nov

osph

ingo

bium

Sten

otro

phom

onas

Simpl

icispi

ra

Paeni

bacillu

s

Ferrug

inib

acte

r

Protein

iphi

lum

Longi

linea

Sphin

gom

onas

uncl

assifie

d_N

itros

omon

adac

eae

Nitr

osoc

occu

s

Nitr

osom

onas

Nitr

osos

pira

Nitr

ospi

ra

uncl

assifie

d_N

itros

pina

ceae

Nitr

obac

ter

uncl

assifie

d_N

itros

pira

ceae

Pseud

omon

as

Limno

bacter

Def

luvi

imon

as

Azo

spiri

llum

Citr

obac

ter

Azo

spira

Com

amon

as

Den

itrat

isom

a

Para

cocc

us

Vul

cani

baci

llus

Adv

enel

la

Are

nim

onas

Thiob

acill

us

Aci

neto

bact

er

Des

ulfo

bulb

us

Azo

arcu

s

Enter

obac

ter

Cal

othr

ix

Rhi

zobi

um

Lau

tropi

a

Bla

utia

Clo

acib

acteriu

m

Fastid

iosipi

la

Ros

eom

onas

othe

rs

uncl

assifie

d

0

1

2

3

20

25

30

35

Rela

tiv

e ab

un

dance (

%)

Lac

toba

cillu

s

Panto

ea

Igna

viba

cter

ium

Orn

atili

nea

Solit

alea

Alis

tipes

Bifid

obac

teriu

m

Tur

icib

acte

r

Smith

ella

Bac

illus

Rhodo

bact

er

Gelria

Var

iovo

rax

Sten

otro

phom

onas

Paeni

bacillu

s

Hym

enob

acter

Prot

eini

philu

m

Longi

linea

Sphin

gom

onas

Nitr

osom

onas

Nitr

osoc

occu

s

uncl

assif

ied_

Nitr

osom

onad

acea

e

Nitr

osos

pira

unclas

sifie

d_Nitr

ospi

nace

ae

Nitr

ospi

ra

unclas

sifie

d_N

itros

pira

ceae

Nitr

obac

ter

Limno

bact

er

Den

itrat

isom

a

Pseud

omon

as

Tha

uera

Are

nim

onas

Azo

spira

Azo

spiri

llum

Com

amon

as

Citr

obac

ter

Thiob

acill

us

Aci

neto

bact

er

Des

ulfo

bulb

us

Azo

arcu

s

Enter

obac

ter

Cal

othr

ix

Lautro

pia

Gem

mob

acter

Lacto

cocc

us

Bla

utia

Esche

richi

a-Sh

igel

la

Met

allib

acteriu

m

Clo

acib

acte

rium

12up

Ekhid

na

othe

rs

uncl

assifie

d

0

2

4

6

20

22

24

26

28

30

Rela

tive a

bu

nd

ance

(%)

Igna

viba

cter

ium

Orn

atili

nea

Solita

lea

Bac

illus

Rho

doba

cter

Nov

osph

ingo

bium

Var

iovo

rax

Alte

rery

thro

bact

er

Ato

poco

ccus

Hym

enob

acte

r

Simpl

icispi

ra

Lacto

cocc

us

Prote

inip

hilu

m

Lon

gilin

ea

Nitr

osom

onas

Sphi

ngom

onas

Nitr

osoc

occu

s

unclas

sified

_Nitr

osom

onad

acea

e

uncl

assifie

d_N

itros

pina

ceae

Nitr

obac

ter

Nitr

ospi

ra

Den

itratisom

a

Limno

bacter

Azo

spira

Are

nim

onas

Com

amon

as

Thaue

ra

Adv

enel

la

Acido

vora

x

Thiob

acill

us

Aci

neto

bact

er

Des

ulfo

bulb

us

Des

ulfo

mic

robi

um

Halot

hiob

acill

us

Des

ulfo

tom

acul

um

Azo

arcu

s

Gra

nulo

sico

ccus

vadi

nBC27

_was

tew

ater

-slu

dge_

grou

p

Fluvi

icol

a

Ow

enwee

ksia

Lau

trop

ia

Gem

mob

acter

Can

dida

tus_

Clo

acam

onas

Fastid

iosipi

la

BD1-

7_cl

ade

Ros

eom

onas

Pire

llula

Can

dida

tus_

Sacc

harim

onas

Rub

ellim

icro

bium

Clo

acib

acte

rium

othe

rs

uncl

assifie

d

0

2

4

15

18

21

24

Rela

tiv

e a

bu

nd

ance (

%)

Heterotrophic bacterium

19.06%

Ammonia oxidizer

1.07%

Nitrite oxidizer

0.17%Denitrification

10.44%

A

B

C

D

Heterotrophic bacterium

16.14%Ammonia oxidizer

1.53%

Nitrite oxidizer

0.25%

Denitrification

16.01%

Heterotrophic bacterium

19.31%

Ammonia oxidizer

1.43%

Nitrite oxidizer

0.43%

Denitrification

10.76%

Heterotrophic bacterium

18.93%Ammonia oxidizer

2.91%

Nitrite oxidizer

0.08%

Denitrification

15.54%

Acinetobacter

0.84%

Acinetobacter

0.44%

Acinetobacter

0.40%

Acinetobacter

0.05%

30

Figure 6:

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.0E+05

1.0E+06

1.0E+07

45 35 30 25

Cop

ies/

ng

DN

A

Reflux ratio

bacteria anammox amoA Nitrobacter

Nitrospira nirS narG

31

Table 1:

Stage 1 Stage 2 Stage 3 Stage 4

Reflux ratio 45 35 30 25

Days 1-67 68-109 110-141 142-176

NH4+-N (mg/L) 299.7±16.4 304.1±19.9 298.7±20.4 299.2±17.5

NO2--N (mg/L) 0.2±0.3 0.2±0.2 0.2±0.1 0.0±0.0

NO3--N (mg/L) 0.9±0.4 2.6±1.3 0.5±0.3 1.5±0.9

TN (mg/L) 366.9±19.9 370.8±33.6 352.2±29.8 365.0±23.4

COD (mg/L) 307±35 350±46 285±29 342±40

COD/TN 0.84±0.10 0.95±0.14 0.82±0.10 0.94±0.11

NLR a (kg/(m

3·d)) 1.10±0.06 1.11±0.10 1.06±0.09 1.09±0.07

OLR b (kg/(m

3·d)) 0.92±0.11 1.05±0.14 0.86±0.09 1.02±0.12

pH 8.0±0.1 7.7±0.2 8.0±0.2 7.8±0.2

a, nitrogen loading rate in terms of TN.

b, organic loading rate in terms of COD.

32

Table 2:

Target Primer Sequence (5’-3’) References

Bacterial

16S rRNA

338F ACTCCTACGGGAGGCAGCAG Muyzer et al.

1993 518R ATTACCGCGGCTGCTGG

Anammox 16S rRNA

AMX809F GCCGTAAACGATGGGCACT

Tsushima et al.

2007

AMX1066

R

AACGTCTCACGACACGAGCTG

amoA

gene

amoA-1F GGGGTTTCTACTGGTGGT Rotthauwe et

al. 1997 amoA-2R CCCCTCKGSAAAGCCTTCTTC

Nitrobacte

r spp. 16S

rRNA

Nitro-1198

F

ACCCCTAGCAAATCTCAAAAAACC

G Graham et al.

2007 Nitro-1423

R

CTTCACCCCAGTCGCTGACC

Nitrospira

spp. 16S

rRNA

NSR 1113F CCTGCTTTCAGTTGCTACCG Dionisi et al.

2002 NSR 1264R GTTTGCAGCGCTTTGTACCG

nirS gene nirS-1F TACCACCCSGARCCGCGCGT Braker et al.

1998 nirS-3R GCCGCCGTCRTGVAGGAA

narG gene narG 1960m2F

TAYGTSGGGCAGGARAAACTG López-Gutiérre

z et al. 2004 narG

2050m2R

CGTAGAAGAAGCTGGTGCTGTT

33

Table 3:

Stage

(reflux ratio)

Steady phase

(days)

Sludge

sample*

Reads 3% distance

Raw

reads

Clean

reads

OTUs Chao1 ACE Shannon

Good's

coverage (%)

1 (45) 55th

- 67th

S1 41342 41328 5632 12481 18929.92 6.83 90.92

2 (35) 96th

- 109th S2 36383 36373 5211 12728 19749.52 6.63 90.62

3 (30) 130th

- 141th

S3 37906 37882 6542 15204 23992.14 6.92 88.48

4 (25) 164th

- 176th

S4 19515 19512 1459 2619 3112.58 5.68 96.43

*, S1, S2, S3 and S4 was the sludge sample taken from the UMSR at the end of stage 1, stage 2, stage 3

and stage 4, respectively.

34

Table 4:

COD

removal a

(kg/(m3·d))

TN

removed a

(kg/(m3·d))

CODrem./TNrem.

ratio a

Actual

TN

removal a

(%)

TN removal via

N2O production

from NO2--N

b

(%)

TN

removal

via nitrite

reduction

to N2 c (%)

TN removal

via nitrate

reduction to

N2 d (%)

Stage 1 0.72 0.94 0.76 87.2 59.0 39.1 23.5

Stage 2 0.89 0.88 1.01 80.0 70.9 47.0 28.3

Stage 3 0.63 0.82 0.77 75.4 50.9 33.8 20.3

Stage 4 0.86 0.66 1.30 60.9 69.4 46.0 27.7

a, the average during steady phase of the stage.

b, denitrification via nitrous oxide according to the theoretical 1.14 g COD/g N without assimilation (Saggar

et al. 2013).

c, denitrification via nitrite according to the theoretical 1.72 g COD/g N without assimilation (Li et al. 2012).

d, denitrification via nitrate according to the theoretical 2.86 g COD/g N without assimilation (Li et al. 2012).

35

Highlights

1. To treat piggery wastewater with high NH4+ density and low C/N ratio.

2. An innovative upflow microaerobic sludge reactor was constructed.

3. Effect of reflux ratio (RR) and mechanism of nitrogen removal were investigated.

4. RR 34 was favorable for NH4+ and TN removal with anammox as the dominant

mechanism.

5. Decrease in RR stimulated ammonia-oxidization but inhibited anammox.