Abstract

Urban rivers are considered as a hot spot of microbial nitrogen cycling due to extensive N loading. However, microbial nitrogen transformation dynamics in urban rivers with different dissolved oxygen (DO) conditions are still unclear. This study investigated the effects of DO concentration changes (anaerobic to aerobic) in overlying water on nitrogen-cycling gene abundance in incubation conditions using sediment from a typical urban river in the Yangtze River Delta. Quantitative polymerase chain reaction (qPCR) results revealed that the abundances of the nitrification gene amoA, denitrification gene nirS/K, norB, nosZ, and anammox gene hzo increased by one to two orders of magnitude from anaerobic to aerobic conditions. Ammonia-oxidizing archaea (AOA) predominated the ammonium oxidation microbial populations, about tenfold more than the ammonia-oxidizing bacteria (AOB) populations. Significant correlations were found among the abundances of AOA-amoA, AOB-amoA, nirS, nirK, and hzo genes, implying a close coupling of aerobic ammonium oxidation (AAO), denitrification, and anammox processes at the molecular level. Moreover, the nitrogen transformation rates were calculated using a box model linking the measured dissolved inorganic nitrogen species. The contribution of anammox to N2 production was 85% under saturated treatment, and the AAO rate was significantly positive correlated to the anammox rate. Our results suggested that coupled AAO and anammox might be the dominant pathway for reactive nitrogen removal in urban rivers with elevated DO levels.

INTRODUCTION

Urban rivers are usually subjected to massive N pollution from point and nonpoint sources, such as sewage and storm water, and are therefore hypothesized as hot spots of nitrogen cycling (Zhang et al. 2015). Dissolved oxygen (DO) is one of the most important factors controlling river water quality (EPA), and could be intensively affected by natural processes such as photosynthesis and respiration of algae (Mulholland et al. 2005), or by human activities including gate projects which might form stagnant anaerobic conditions and ecological remediation projects which might supply oxygen by the root systems of submerged plants (Nykänen et al. 2012). Thus, a comprehensive study of DO concentration impact on nitrogen behavior in urban rivers will enrich our understanding of nitrogen biogeochemical cycles in aquatic ecosystems.

The effect of DO on the nitrogen cycle in river ecosystems is well studied but most of it is limited to ammonia oxidation and denitrification. It is commonly realized that hypoxic conditions (DO < 2.0 mg L−1) can trigger the release of ammonium due to anaerobic fermentation from nitrogen-rich sediments to the overlying water, while higher DO levels could activate nitrifying bacteria, leading to high rates of biological ammonia oxidation and coupled nitrification–denitrification (Palmer et al. 2009). These studies have established a basic understanding of the complicated behaviors of nitrogen under various DO levels in rivers. In recent years, by using molecular biology techniques, some new nitrogen transformation pathways such as ammonia-oxidizing by archaea (Francis et al. 2005), anammox (Hu et al. 2012) and dissimilatory nitrate reduction to ammonia (DNRA) (Cheng et al. 2016) have been discovered in river systems.

Nitrogen transformations were driven by a variety of microbes possessing specific enzymes involving the nitrogen cycle (Canfield et al. 2010) including amoA for ammonia oxidation bacteria or archaea, nirS and nirK for nitrite reduction, norB for NO reduction to N2O, nosZ for N2O reduction to N2, and hzo specific to anammox bacteria which dehydrogenate the unique anammox intermediate, hydrazine, to form N2 (Schmid et al. 2008). These functional genes (Figure S1, available with the online version of this paper) encoding enzymes could indirectly represent the nitrogen transformation pathway (Petersen et al. 2012). For the aerobic ammonium oxidation (AAO) process, it has been reported that DO could affect the dynamics of AOA and AOB lineages and abundances (Abell et al. 2011), with the AOA dominating in anoxic areas and anaerobic estuarine sediments. However, there was no consensus on the effects of DO on anammox activity and its contribution to nitrogen removal. Dale et al. (2009) found that elevated DO concentrations enhance the activity of anammox and contributed more to nitrogen removal in Cape Fear River estuary, while Neubacher et al. (2013) reported that anammox rates remained constant in hypoxia and increased disproportionately under sustained anoxic conditions in the southern North Sea. Most of these studies are focused on estuary systems, however, and the urban river is very different from the estuary system by salinity and hydrologic conditions, and the responses of the coupled nitrogen transformation to various DO levels in the urban river are still unclear.

In this study, experiments using sediment from a typical nitrogen polluted urban river in the Yangtze River Delta were conducted. Nitrogen transformation rates and the abundance of related functional genes under anaerobic, aerobic, and saturated DO levels were determined to reveal the impact of DO on (1) the rates of nitrification and AOA and AOB contributions to ammonia oxidation and (2) the rates of denitrification and anammox and their contributions to nitrogen removal.

MATERIALS AND METHODS

Experimental design

The Inner Qinhuai River is a typical urban river in Nanjing City, with a total length of 25 km and basin area of 24.2 km2. It is generally disconnected from the regional river (Qinhuai River) by water gates. Data from nine sites (Figure S2, available with the online version of this paper) of the middle part of this river showed that this river received a total amount of sewage discharge of about 20,000 tons per day, which resulted in heavily nitrogen polluted (total nitrogen, TN > 20 mg L−1; NH4+ > 15 mg L−1) and anoxic conditions (DO < 2.0 mg L−1) of the overlying water. The content of organic matter, total nitrogen, NH4+ and NOx in the sediment was 37.39 g kg−1, 1.64 g kg−1, 0.14 g kg−1 and 5.93 mg kg−1, respectively. In November 2013, surface sediment (∼10 cm) samples were collected from the Inner Qinhuai River. The sediment samples were homogenized and placed in a covered PVC column, and distilled water was then injected gently into the column (2 cm sediment, 15 cm overlying water). The oxygen conditions of the overlying water were controlled by regulating the inflow of sterile air or 99.99% N2 to create three different DO conditions: anaerobic (AN), DO < 0.5 mg L−1; aerobic (A), 2 mg L−1 < DO < 4 mg L−1; and saturated (S), DO > 8.0 mg L−1. The overlying water DO concentration was monitored automatically using a DO electrode (Unisense, Denmark).

The sediment core reactors were statically incubated in duplicate at room temperature (20 ± 1 °C) in the dark. Using sterile syringes, about 5 mL water samples were collected from three depths of 2 cm, 7 cm, and 12 cm of the overlying water, respectively, and then homogenized to represent the average overlying water to determine the N species concentration. The same bulk of distilled water was supplemented to the overlying water to keep a constant water volume. On days 1, 7, 20, and 37, the vertical DO profile was measured, and approximately 2.0 g of surface sediment (5 mm) was collected carefully from each treatment for DNA extraction and subsequent analysis.

Chemical analysis

Concentration of NH4+ was measured by sodium salicylate–sodium hypochlorite spectrophotometry. Concentrations of NO2 and NO3 were determined by ion chromatography using a Dionex ICS-2000 system. Each sample was measured in triplicate, and the detection limits of NH4+, NO2 and NO3 were 0.02, 0.005, and 0.05 mg L−1, respectively, with uncertainties less than 5%.

DNA extractions and quantitative PCR

Sediment DNA was extracted in triplicate using the MP FastDNA® Spin Kit for Soil, following the manufacturer's instructions. The real-time fluorescence quantitative polymerase chain reaction (qPCR) assays targeting the nitrogen-cycling genes were carried out in triplicate with an ABI StepOne Plus sequence detection system (Applied Biosystems, USA) using the SYBR green qPCR method (Dang et al. 2010). The primers used for target genes and annealing temperature are listed in Table S1 (available online). Standard curves were obtained with serial dilution of purified target gene fragments. In all treatments, negative controls containing no template DNA were subjected to the same qPCR procedure to exclude or detect any possible contamination or carryover. The amplification efficiency of the standard curves was 0.93 to 1.05, and the correlation coefficient was larger than 0.995.

Nitrogen transformation rate modelling and statistical analysis

The rates of ammonification (R1), ammonia oxidation (R2), nitrite oxidation (R3), denitrification (R4), and anammox (R5) were modelled using the measured nutrient concentrations by the method described previously (Babbin & Ward 2013). Briefly, a box model linking the three measured dissolved inorganic nitrogen (DIN) species via these five biological processes was implemented. The DIN measurements were smoothed using a Savitzky–Golay filter to minimize sampling noise. Time derivatives were numerically calculated for each of the three DIN species, and a least squares non-negative fit of the rates was determined using the algorithm in MATLAB software. As the model used only three inorganic nitrogen concentrations, which are not enough to estimate five rates, we use the Michaelis–Menten rate equation to overcome this limitation.

Spearman correlation coefficients between nitrogen compounds and nitrogen-cycling gene abundances were analyzed using SPSS Statistics 20 (IBM, USA).

RESULTS

Concentrations of nitrogen compounds

Each of the treatments had a rapid flux of NH4+ over the first 7 days (from 0 to 35 mg L−1), while the NO2 and NO3 concentrations were relatively lower during the incubation. However, the inorganic nitrogen (NH4+, NO2, NO3) in the overlying water revealed different time-series trends in three treatments (Figure 1).

Figure 1

Concentrations (mg·L−1) of NH4+, NO2, and NO3 in the water column during the incubation (43 days) under (a) anaerobic (DO < 0.5 mg·L−1), (b) aerobic (2.0 mg·L−1 < DO < 4.0 mg·L−1), and (c) saturated (DO >8.0 mg·L−1) conditions. Calculated rates of ammonia oxidation, nitrite oxidation, denitrification, and anammox in (d) anaerobic, (e) aerobic, (f) saturated treatment. The rates were calibrated well using the Michaelis–Menten rate equation in the aerobic and saturated treatments, excepting the anaerobic treatment.

Figure 1

Concentrations (mg·L−1) of NH4+, NO2, and NO3 in the water column during the incubation (43 days) under (a) anaerobic (DO < 0.5 mg·L−1), (b) aerobic (2.0 mg·L−1 < DO < 4.0 mg·L−1), and (c) saturated (DO >8.0 mg·L−1) conditions. Calculated rates of ammonia oxidation, nitrite oxidation, denitrification, and anammox in (d) anaerobic, (e) aerobic, (f) saturated treatment. The rates were calibrated well using the Michaelis–Menten rate equation in the aerobic and saturated treatments, excepting the anaerobic treatment.

In the saturated treatment, the NH4+ decreased rapidly to a low level (<0.3 mg L−1) from day 7. The NO2 concentration reached the peak value of 2.48 mg L−1 at day 30 and then decreased to the detection limit (0.02 mg L−1) on day 43. The NO3 increased steadily to 2.08 mg L−1 at day 37 and decreased to 1.01 mg L−1 at the end of the incubation.

In the aerobic treatment, the NH4+ showed a slightly decreasing trend from day 7. The NO3 increased gradually to approximately 0.8 mg L−1, while the nitrite NO2 began to increase sharply (day 37) from the detection limit (0.02 mg L−1) to approximately 0.6 mg L−1.

In the anaerobic treatment, the NH4+ showed a slightly increasing trend from day 7. NO2 and NO3 concentrations were very low during the incubation.

Changes of gene abundances

The DO concentrations of the 5 mm surficial sediment increased from approximately 0 mg L−1 to 1.25 mg L−1 and 3.73 mg L−1 in the anaerobic, aerobic and saturated treatments at day 37 (Figure 2(a)). The abundance of key genes involved in nitrogen-cycling pathways changed differently according to DO concentration.

Figure 2

Nitrogen-cycling gene abundances and average DO concentration in the 5 mm surficial sediment in the anaerobic, aerobic and saturated treatments: (a) average DO concentrations in the 5 mm surficial sediment, (b) ammonia-oxidizing archaeal amoA (AOA-amoA) gene, (c) ammonia-oxidizing bacterial amoA (AOB-amoA) gene, (d) nitrite-oxidizing-related nxrA gene, denitrification (e) nirS and (f) nirK gene, N2-production-related (g) norB, (h) nosZ gene, and anammox-specific (i) hzo gene.

Figure 2

Nitrogen-cycling gene abundances and average DO concentration in the 5 mm surficial sediment in the anaerobic, aerobic and saturated treatments: (a) average DO concentrations in the 5 mm surficial sediment, (b) ammonia-oxidizing archaeal amoA (AOA-amoA) gene, (c) ammonia-oxidizing bacterial amoA (AOB-amoA) gene, (d) nitrite-oxidizing-related nxrA gene, denitrification (e) nirS and (f) nirK gene, N2-production-related (g) norB, (h) nosZ gene, and anammox-specific (i) hzo gene.

In all treatments, the abundance of AOA-amoA was about one order of magnitude higher than the abundance of AOB-amoA. Both of them increased quickly by two orders of magnitude at the beginning (7 days) in the aerobic and saturated treatments (Figure 2(a) and 2(b)). From day 21, the saturated treatment revealed a decreasing trend for both the AOA-amoA and AOB-amoA genes. The nxrA gene increased after 7 days' incubation in the aerobic and saturated treatments. During incubation, DO concentrations showed different impacts on the abundances of two genes encoding nitrite reductases (nirS and nirK genes). In aerobic and saturated treatments, the nirS gene abundance initially increased and then decreased (Figure 2(d)), but the nirK gene abundance showed a lag in growth (Figure 2(e)). In the anaerobic treatment, the nirS gene showed a moderate growth, while the nirK gene revealed no significant difference during incubation. The abundances of genes associated with N2 production (norB and nosZ) changed within one order of magnitude during the incubation. The abundance of the nosZ gene had a slightly decreasing trend in anaerobic treatment but increased moderately in aerobic and saturated treatments. The anammox-specific hzo gene also showed an increasing trend in all the treatments, ranging from 1.30 × 107 copies g−1 to 3.62 × 107 copies g−1.

Nitrogen transformation rates and their integrated amounts

The R1, R2, R3, R4, and R5 rates were calculated and validated in each treatment (Figure 1(d)–1(f); Figure S3, available with the online version of this paper). Generally, the ammonification process was high at the beginning, and decreased rapidly. The R2, R4, and R5 rates changed synchronously, and R4 was lower than R2 and R5. The peak values of each rate from high to low were the saturated, aerobic, and anaerobic treatments, respectively.

However, in the saturated experiment, from day 20 to day 30, R2 was significantly higher than R3 and R4, and R5, which resulted in the accumulation of NO2; from day 30 to day 35, R2 and R5 were lower than R3 and R4, while from day 35 to day 43, R4 and R5 were higher than R2 and R4.

The total amount of ammonium oxidation, nitrite oxidation, denitrification, and anammox increased with the DO levels, but the amount of ammonium release decreased inversely. The contributions of anammox to nitrogen losses varied in a small range from 83.36% to 89.19% (Table S2, available online).

Relationships between nutrients, rates, and gene abundances

The average DO concentrations in surface sediment were significantly correlated with NO2 and NO3 concentrations and the gene abundances of almost all the nitrogen genes except for nxrA (Table S3, available online).

The nitrogen-cycling genes' abundances were closely related to each other. The AOA-amoA and AOB-amoA genes' abundances were positively related to each other (r2 = 0.98), and both of them had positive correlations with nirS, norB, and hzo gene abundance. Moreover, the hzo gene abundance showed positive relationships to all the other genes.

The NO3 and NO2 concentrations were related to most of the genes, while the NH4+ showed no significance with DO nor with the nitrogen genes, which implied that NO3 and NO2 might play an important role in the microbial nitrogen cycle under change of DO level.

DISCUSSION

Gene abundances under different DO conditions

The dynamics of AOA and AOB lineages and abundances were correlated strongly with DO concentrations (Lian et al. 2014). In this study, both of the amoA gene (AOA-amoA and AOB-amoA) abundances increased by two orders in the aerobic and saturated experiments compared with the anaerobic experiments, which is similar to reports from mudflat (Luo et al. 2014) and estuarine sediment (Abell et al. 2011).

Denitrification bacteria favor anoxic or anaerobic habitats (Jensen et al. 2009), however, in this study, significant increases of denitrification gene abundance (nirS, nirK, norB, and nosZ) were found with DO elevation, which might be due to the increase of aerobic denitrification bacteria (Gao et al. 2009).

The abundances of the nirS and nirK genes were increased about tenfold in aerobic and saturated treatments (Figure 2(a) and 2(b)), while the norB and nosZ genes increased about two fold, compared with the anaerobic treatment. The difference might lead to partial denitrification and an increase in N2O emission (Wang et al. 2015). Further study on the existence and effects of possible aerobic denitrification and N2O emission is needed.

The anammox bacteria live in anaerobic habitats, and they can also survive in facultative environments (Jensen et al. 2009), but no significant correlation was found between the abundance and DO concentration (Zhang et al. 2014). In this study, the anammox-specific hzo gene increased with DO elevation (from 0 to 3.73 mg/L) under anoxic and facultative anaerobic conditions.

AOA and AOB for ammonium oxidation

The AOA populations predominate the ammonium oxidation microbial populations, as reported for many freshwater sediments (Zhang et al. 2014). The predomination of the AOA populations might imply that the AOA contribute more to the ammonium oxidation process than the AOB. A previous study (Luo et al. 2014) has found that the nitrification rate was positively correlated to the abundance of AOA rather than AOB, and the AOA were primarily involved in ammonium oxidation in estuarine mudflat sediments. To determine the contribution of AOA or AOB to ammonium oxidation, further studies on the activities of AOA and AOB should be performed.

Denitrification and anammox

The DO concentration is thought to be an important factor for the contribution of anammox to nitrogen removal (Dale et al. 2011). However, in this study, the calculated contributions of anammox to nitrogen losses (%Ramx) varied in a small range (83.36% ∼ 89.19%). The DO levels showed no significant impact on the contribution of anammox to nitrogen losses, which might be due to the uniformity of sediments with the same C/N ratio used in the incubation (Burgin & Hamilton 2007; Babbin & Ward 2013).

The calculated contributions of anammox to N2 production in DO saturated treatment (85.35%) is much higher than previously reported for marine ecosystems (29%), but is close to the ratio for a bioreactor of simultaneous anammox and heterotrophic denitrification (SAD) (85%) (Bi et al. 2015). It is reported that the anammox preferred lower C/N (Algar & Vallino 2014), and the river received untreated sewage as low C/N (∼2.5) as the SAD (∼2.0), higher than marine ecosystems (∼6.6). In addition, the gene (nosZ, ∼105 copies g−1) of nitrogen production in denitrification bacteria was lower than the hzo gene related to N2 production in anammox bacteria (hzo, ∼106 copies per gram), which might reveal a higher N2-production potential for anammox than for denitrification. Moreover, considering the DNRA process, which was not included in the calculation model, the accurate contributions of anammox to N2 production should be confirmed by isotope-pairing technology in future work.

Coupled nitrogen loss pathway

Significant correlations were found between the abundances of AOA-amoA, AOB-amoA, nirS, nirK, and hzo (Table S3), implying a close coupling of AAO, denitrification and anammox at the molecular level. These couplings of genes have also been found in high-altitude lakes (Vila-Costa et al. 2014), hyper-nutrified estuarine tidal flats (Zhang et al. 2014) and a tidal flow constructed wetland (Zhi & Ji 2014). It is reported that the anammox is denitrification-dependent in the sediment of inland rivers (Zhou et al. 2014). However, in this study, the calculated AAO rates coupled tightly with the anammox rates reveal that the coupling of AAO and anammox was the dominant pathway of nitrogen.

The nitrogen turnover processes are limited by ammonium oxidation because the concentrations of NO2 and NO3 are very low compared with the NH4+ at the beginning of the incubation. Higher DO concentrations increased AOA and AOB populations by two orders of magnitude (Figure 2), significantly raising the ammonia oxidation rates (R2) and ultimately leading to higher N2 production. However, when the average DO concentration in the 5 mm surficial sediment is greater than 2.13 mg L−1 (day 21 in the saturated treatment), the NO2 is accumulated in the incubation system, which might be due to the inhibition of anammox and denitrification rates at higher DO concentrations. The river restoration to improve DO conditions would also lead to greater N removal and retention which would improve downstream water quality.

CONCLUSIONS

In the hypoxic nitrogen-rich urban river, elevated DO concentrations of the overlying water could stimulate nitrogen-cycling bacteria using functional genes as proxy. All the gene abundances showed positive correlation with average DO concentrations in the surficial sediment, especially for the AOA-amoA gene and AOB-amoA gene. Significant correlation between amoA, hzo, nirK, and nirS showed that nitrogen-cycling pathways were tightly coupled at the molecular level. In the DO saturated treatment, the calculated anammox rate is tightly coupled to the nitrification rate and contributed 85.35% to N2 production. Thus, the coupling of nitration and anammox was the main nitrogen-removal pathway. When the overlying water changed from anaerobic to aerobic, nitration rates and subsequent anammox or denitrification rates increased because the growth of aerobic ammonium-oxidation bacteria and archaea was stimulated by increased DO concentration. These results suggest that river restoration to improve DO conditions would also lead to greater N removal and retention which would improved downstream water quality.

ACKNOWLEDGEMENTS

Funding: This work was supported by the National Natural Science Foundation of China (grant number 41230640).

REFERENCES

REFERENCES
Abell
G. C. J.
,
Banks
J.
,
Ross
D. J.
,
Keane
J. P.
,
Robert
S. S.
,
Revill
A. T.
&
Volkman
J. K.
2011
Effects of estuarine sediment hypoxia on nitrogen fluxes and ammonia oxidizer gene transcription
.
FEMS Microbiology Ecology
75
(
1
),
111
122
.
Algar
C. K.
&
Vallino
J. J.
2014
Predicting microbial nitrate reduction pathways in coastal sediments
.
Aquatic Microbial Ecology
71
(
3
),
223
238
.
Babbin
A. R.
&
Ward
B. B.
2013
Controls on nitrogen loss processes in Chesapeake Bay sediments
.
Environmental Science & Technology
47
(
9
),
4189
4196
.
Burgin
A. J.
&
Hamilton
S. K.
2007
Have we overemphasized the role of denitrification in aquatic ecosystems? A review of nitrate removal pathways
.
Frontiers in Ecology and the Environment
5
(
2
),
89
96
.
Canfield
D. E.
,
Glazer
A. N.
&
Falkowski
P. G.
2010
The evolution and future of Earth's nitrogen cycle
.
Science
330
(
6001
),
192
196
.
Cheng
L.
,
Li
X. F.
,
Lin
X. B.
,
Hou
L. J.
,
Liu
M.
,
Li
Y.
,
Liu
S.
&
Hu
X. T.
2016
Dissimilatory nitrate reduction processes in sediments of urban river networks: spatiotemporal variations and environmental implications
.
Environmental Pollution
219
,
545
554
.
Dale
A. W.
,
Sommer
S.
,
Bohlen
L.
,
Treude
T.
,
Bertics
V. J.
,
Bange
H. W.
,
Pfannkuche
O.
,
Schorp
T.
,
Mattsdotter
M.
&
Wallmann
K.
2011
Rates and regulation of nitrogen cycling in seasonally hypoxic sediments during winter (Boknis Eck, SW Baltic Sea): sensitivity to environmental variables
.
Estuarine, Coastal and Shelf Science
95
(
1
),
14
28
.
Francis
C. A.
,
Roberts
K. J.
,
Beman
J. M.
,
Santoro
A. E.
&
Oakley
B. B.
2005
Ubiquity and diversity of ammonia-oxidizing archaea in water columns and sediments of the ocean
.
Proceedings of the National Academy of Sciences of the United States of America
102
(
41
),
14683
14688
.
Gao
H.
,
Schreiber
F.
,
Collins
G.
,
Jensen
M. M.
,
Kostka
J. E.
,
Lavik
G.
,
de Beer
D.
,
Zhou
H.-y.
&
Kuypers
M. M. M.
2009
Aerobic denitrification in permeable Wadden Sea sediments
.
ISME J.
4
(
3
),
417
426
.
Hu
B.
,
Shen
L.
,
Zheng
P.
,
Hu
A.
,
Chen
T.
,
Cai
C.
,
Liu
S.
&
Lou
L.
2012
Distribution and diversity of anaerobic ammonium-oxidizing bacteria in the sediments of the Qiantang River
.
Environmental Microbial Reports
4
(
5
),
540
547
.
Jensen
M. M.
,
Petersen
J.
,
Dalsgaard
T.
&
Thamdrup
B.
2009
Pathways, rates, and regulation of N2 production in the chemocline of an anoxic basin, Mariager Fjord, Denmark
.
Marine Chemistry
113
(
1–2
),
102
113
.
Nykänen
A.
,
Kontio
H.
,
Klutas
O.
,
Penttinen
O.-P.
,
Kostia
S.
,
Mikola
J.
&
Romantschuk
M.
2012
Increasing lake water and sediment oxygen levels using slow release peroxide
.
Science of the Total Environment
429
,
317
324
.
Palmer
H.
,
Beutel
M.
&
Gebremariam
S.
2009
High rates of ammonia removal in experimental oxygen-activated nitrification wetland mesocosms
.
Journal of Environmental Engineering
135
(
10
),
972
979
.
Petersen
D. G.
,
Blazewicz
S. J.
,
Firestone
M.
,
Herman
D. J.
,
Turetsky
M.
&
Waldrop
M.
2012
Abundance of microbial genes associated with nitrogen cycling as indices of biogeochemical process rates across a vegetation gradient in Alaska
.
Environmental Microbiology
14
(
4
),
993
1008
.
Schmid
M. C.
,
Hooper
A. B.
,
Klotz
M. G.
,
Woebken
D.
,
Lam
P.
,
Kuypers
M. M. M.
,
Pommerening-Roeser
A.
,
op den Camp
H. J. M.
&
Jetten
M. S. M.
2008
Environmental detection of octahaem cytochrome c hydroxylamine/hydrazine oxidoreductase genes of aerobic and anaerobic ammonium-oxidizing bacteria
.
Environmental Microbiology
10
(
11
),
3140
3149
.
Vila-Costa
M.
,
Bartrons
M.
,
Catalan
J.
&
Casamayor
E. O.
2014
Nitrogen-cycling genes in epilithic biofilms of oligotrophic high-altitude lakes (central Pyrenees, Spain)
.
Microbial Ecology
68
(
1
),
60
69
.
Wang
J.
,
Chen
N.
,
Yan
W.
,
Wang
B.
&
Yang
L.
2015
Effect of dissolved oxygen and nitrogen on emission of N2O from rivers in China
.
Atmospheric Environment
103
,
347
356
.
Zhang
X.
,
Wu
Y.
&
Gu
B.
2015
Urban rivers as hotspots of regional nitrogen pollution
.
Environmental Pollution
205
,
139
144
.
Zhou
S.
,
Borjigin
S.
,
Riya
S.
,
Terada
A.
&
Hosomi
M.
2014
The relationship between anammox and denitrification in the sediment of an inland river
.
Science of the Total Environment
490
,
1029
1036
.

Supplementary data