Distribution patterns of denitrification functional genes and microbial floras in multimedia constructed wetlands

Distribution patterns of denitrification functional genes and microbial floras in multimedia constructed wetlands

Ecological Engineering 44 (2012) 179–188 Contents lists available at SciVerse ScienceDirect Ecological Engineering journal homepage: www.elsevier.co...

1MB Sizes 0 Downloads 16 Views

Ecological Engineering 44 (2012) 179–188

Contents lists available at SciVerse ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Distribution patterns of denitrification functional genes and microbial floras in multimedia constructed wetlands Guodong Ji a,∗ , Rongjing Wang a , Wei Zhi a , Xuexin Liu b , Yaping Kong b , Yufei Tan c a

Key Laboratory of Water and Sediment Sciences, Ministry of Education, Department of Environmental Engineering, Peking University, Beijing 100871, China China Academy of Transportation Sciences, Huixinli 240, Beijing 100029, China c Chinese Research Academy of Environmental Sciences, Dayangfang 8, Anwai Beiyuan, Beijing 100012, China b

a r t i c l e

i n f o

Article history: Received 27 November 2011 Received in revised form 18 February 2012 Accepted 26 March 2012 Available online 1 May 2012 Keywords: Constructed wetlands Functional genes Microbial floras Denitrification Distribution patterns

a b s t r a c t The present study, a quantitative investigation of distribution patterns of functional gene communities, including Anammox bacteria 16S rRNA, amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nas, and nifH in two multimedia constructed wetland systems (CWs) was conducted. DGGE results showed similar distribution patterns in two constructed wetland groups. Lactococcus sp. and Moraxella sp. were the dominant organic nitrogen and denitrification flora; Acinetobacter sp. were the dominant NO3 − to NH4 + transformation flora; and Bacillus sp. and ˇ-Proteobacteria sp. were the dominant NH4 + removal flora in the two constructed wetland groups. Quantitative real-time PCR results showed the qnorB and nas functional genes were predominantly enriched in the 15 cm and 60 cm layers of the CW, which was prepared by multi-media packing of 1% zero-valent iron. The nirK was predominantly enriched in the 60 cm layer of the CW, which was prepared by multi-media packing of 2% zero-valent iron. Other functional nitrogen transformation genes were predominantly enriched in the 15 cm layer of CWs. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Constructed wetlands were developed as a sewage treatment method in Germany during the 1970s (US EPA, 1993). The technique was composed of a hierarchical system of fillers, aquatic plants growing on the fillers, and microorganisms deposited on the fillers and aquatic plants. Consequently, a unique man made matrix-plant – microbe ecosystem was established to effectively treat and dispose of wastewater (US EPA, 1993). Wastewater purification mechanisms by constructed wetlands are complex, including matrix precipitation adsorption, ion exchange, plant uptake and microbial degradation, among other factors (Hsu et al., 2011; Vymazal, 2011). Studies have indicated the practice of constructed wetlands had effective removal performance for biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended solids (TSS), and oils (Ji et al., 2002; Konnerup and Brix, 2010). However, nitrogen removal efficiency showed substantial fluctuations when the wetland was influenced by variable nitrogen and hydraulic loadings, and constructed wetland type (Lin et al., 2002). The North America database documented an average nitrogen removal rate of 44% in wastewater-constructed wetlands (IWA, 2000). Verhoeven and Meuleman (1999) reported the

∗ Corresponding author. Tel.: +86 1062755914 87; fax: +86 1062756526. E-mail address: [email protected] (G. Ji). 0925-8574/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecoleng.2012.03.015

nitrogen removal rate of a classic constructed wetland in Europe was only 35%, and could not exceed 50%, even with optimal design. How to improve nitrogen removal efficiency in constructed wetlands has become a research hotspot in the water science field. Research has demonstrated that nitrogen removal in constructed wetlands resulted from the combined action of physical, chemical, and biological processes. Denitrification pathways included matrix deposition adsorption, ion exchange, ammonia volatilization, plant uptake, animal nourishment, and microbial nitrogen removal, among other components. Konnerup and Brix (2010) reported the key factors to achieve high-efficient denitrification which included matrix deposition adsorption and microbial transformation. A constructed wetland surface substrate exhibited some degree of adsorption activity, and could absorb dissolved nitrogen in wastewater, particularly reduced form steady-state ammonia, which served a role in blocking and filtering (Davidsson and Stahl, 2000). However, the ion exchange of the substrate active site towards ammonia could not act long-term as the confluence for ammonia removal, because the ammonia adsorbed by the substrate will continue to be exploited (Blankenberg et al., 2008). Different substrates had different adsorption capacities; zeolite, clay, and peat soil (field peaty soil) possessed an increased cation-exchange capacity, which could improve nitrification effects, however gravel cation exchange capacity was poor (Liu and Zhou, 2009; Jenssen et al., 2010; Li et al., 2011). The limitations of natural substrate adsorption capacity resulted in the development of bio-ceramic,

180

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

and other low-cost multi-media packing using waste from industry and agriculture as raw materials. These materials have been verified with suitable nitrogen and phosphorus adsorption properties. Matrix deposition adsorption was considered a key denitrification factor in constructed wetlands, however a substantial number of studies have reported microbes were responsible for the removal of most nitrogen (89–96%) (Lin et al., 2002; Konnerup and Brix, 2010). Microbial nitrogen removal included, ammonia oxidation, nitrification, denitrification, and anaerobic ammonium oxidation, among other pathways (McDevitt et al., 2000). Nitrification included the following two steps: (i) aerobic ammonium oxidation, where ammonia was oxidized to nitrite; and (ii) oxidation of nitrite, where nitrite was further oxidized to nitrate (Julie et al., 1991). In addition to aerobic ammonia oxidation, anaerobic ammonium oxidation occurred in the natural environment, and wastewater treatment systems. Anaerobic ammonium oxidation oxidized NH4 + , NO3 − , or (NO2 − ) to N2 with NH4 + as the electron donor, and NO3 − or NO2- as the electron acceptor (Mulder et al., 1995). In addition to the presence of aerobic denitrifiers heterotrophic nitrification, microbial nitrogen fixation, and nitrate respiration in aerobic conditions have been widely confirmed (McDevitt et al., 2000). Denaturing gradient gel electrophoresis (DGGE) has been successfully applied to study ammonia oxidation and denitrifying bacterial diversity, and results have indicated that DGGE was a viable tool to study microbial diversity in environmental samples (Throback et al., 2004). Real-time PCR accurately quantified nitrification and dentrification genes, including amoA, nirK, and nosZ (Henry et al., 2006). However, to date most quantitative studies have only focused on individual functional genes, and the joint-use of PCR-DGGE and real-time PCR has not been reported in denitrification microbes from constructed wetlands, and the distribution patterns of nitrogen transformation functional genes has not been evaluated. In this study, microbial diversity in two multimedia constructed wetland systems with different nitrogen removal efficiencies were researched by using the qualitative PCR-DGGE and real-time PCR method. Secondly, the distribution patterns of amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nas, nifH were inferred, and the anaerobic ammonium-oxidizing bacteria were identified in the two-constructed wetland ecosystems.

2. Materials and methods 2.1. Multimedia constructed wetland Two multi-media wetlands systems (each with four parral, total 8 CWs) were established with 5.00 m length × 1.00 m width × 1.70 m depth. The following composition comprised the surface to bottom layer: the rhizosphere layer (0–30 cm), filled with native sandy loam; water distribution layer (30–50 cm), filled with 10 mm particle sized ash; multimedia filter layer (50–130 cm), comprised of permeable filler, and four multi-block layers filled with nested brick, each multimedia block measuring100 cm length × 30 cm width × 13 cm depth. The horizontal and vertical multimedia block spacing was respectively 10 cm and 7 cm apart; and spaces between blocks were occluded with permeable filler. Permeable filler was composed of 10 mm diameter gravel, and 2–4 mm diameter natural clinoptilolite, with a 7:1 volume ratio; multimedia block is mixed by bio-ceramic, 2–4 mm diameter natural clinoptilolite, and 1–2 mm diameter natural clinoptilolite, with a 1:1:1 volume ratio; and the water collection layer level (130–170 cm), was filled with 32 mm ash particles, and 10 mm

diameter gravel, with a 2:1 volume ratio. Yellow iris (Iris pseudacorus (Iridaceae)) was planted at the constructed wetland surface layer, with a planting density of 25 clusters m−2 (the space between clusters was 0.15 cm). A different treatment between the two-wetland groups was established. Bio-ceramics filled the multimedia block, and the bioceramic material had similar composition with the following main ingredients: fly ash, sawdust, natural clinoptilolite and calcium carbonate. However, bio-ceramic in constructed wetland No. 1 included the addition of 1% metallic iron, and bio-ceramic in constructed wetland No. 2 included the addition of 2% metallic iron. An 8 mm diameter bio-ceramic was used in two constructed wetland systems, and the production process and performance of the product was reported in Ji et al. (2010a). The two constructed wetland systems were established in the water-saving irrigation demonstration base of the Chinese Ministry of Water Resources, Shunyi District, Beijing, China, and used to treat the daily domestic sewage of the base buildings. During the tests, the constructed wetland systems were not under temperature control, and the effluent temperature ranged from 18 to 25 ◦ C. The commissioning and operation of the two-wetland systems was initiated on May 11, 2009. NH4 + , NO3 − , NO2 − , total nitrogen (TN), and COD transformation efficiencies were analyzed over a 16 week period. The operation was divided into four different phases. Baseline data was measured to determine nitrogen removal efficiencies, including four operation phases. The first was the domestication phase from May 11 to May 31, with hydraulic loading of 10.0 cm d−1 . The second was the commissioning phase with 20.0 cm d−1 of hydraulic loading from July 1 to July 18. The third was the hydraulic shock loading phase from July 19 to August 31 with hydraulic loading of 40.0 cm d−1 . TN shock loading was the fourth phase from September 1 to September 28. During this period, TN loading increased 3.5 times, and hydraulic loading was maintained. 2.2. Sample collection and determination Throughout the operation, water samples were collected once a week, and measured on-site for NH4 + , NO2 − , NO3 − , TN, COD, pH, dissolved oxygen (DO), and redox potential. A HACH DR2800 multi-function water quality tester was used to measure these parameters applying standard protocols (State Environmental Protection Administration, 2002). The microbial flora in two construced wetland systems was sampled on September 28, 2009. Seven samples from the rhizosphere, water distribution, and water collection layers, and each layer of the four layer multimedia blocks in the multimedia filter layer were collected in the two constructed wetlands. The samples were labeled 15 cm layer, 40 cm layer, 60 cm layer, 80 cm layer, 100 cm layer, 120 cm layer, and 140 cm layer. The sampling column extracted samples from each corresponding layer. The samples were subsequently stored in an ice incubator, and sent to the Laboratory of Environmental Engineering, Peking University. D5625-01 Soil DNA Kits (Omega, USA) were used to extract and purify total genomic DNA from the samples. Extracted genomic DNAs were maintained in a −20 ◦ C freezer until further use, and detected by 1% agarose gel electrophoresis. 2.3. PCR-DGGE (denaturing gradient gel electrophoresis) 2.3.1. PCR (polymerase chain reaction) The general PCR primers for the 16S rDNA V3 variable region of most bacteria are 338F (5 -TACGGGAGGCAGCAG-3 ) and 518R (5 -CCATACGGGAGGCAGGCAG-3 ); and the addition of the GC clip CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCC to the 5 end of the reverse primer to prepare for denaturing

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

181

Table 1 Primers for quantitative real-time PCR assays. Functional genes

Primer name

Primer sequence (5 –3 )

Amplification size (bp)

Source

ANO 16S rRNA

AMX809F AMX1066R amo598f amo718r F1nxrA R1nxrA 1960m2f 2050m2r napA V17F napA 4R nirK583 F nirK909R qnorB2F qnorB5R NosZ 1527F NosZ 1773R nas F nas R nifH F

GCCGTAAACGATGGGCACT AACGTCTCACGACACGAGCTG GAATATGTTCGCCTGATTG CAAAGTACCACCATACGCAG CAGACCGACGTGTGCGAAAG TCYACAAGGAACGGAAGGTC TA(CT)GT(GC)GGGCAGGA(AG)AAACTG CGTAGAAGAAGCTGGTGCTGTT TGGACVATGGGYTTYAAYC ACYTCRCGHGCVGTRCCRCA TCATGGTGCTGCCGCGKGACGG GAACTTGCCGGTKGCCCAGAC GGNCAYCARGGNTAYGA ACCCANAGRTGNACNACCCACCA CGCTGTTCHTCGACAGYCA ATRTCGATCARCTGBTCGTT CARCCNAAYGCNATGGG ATNGTRTGCCAYTGRTC TGXGAXCCYAAZGCYGA

257

Tsushima et al., 2007

120

Dionisi et al., 2002

322

Poly et al., 2008

100

Lopez-Gutierrez et al., 2004

152

Bru et al., 2007

326

Yan et al., 2003

262

Braker and Tiedje, 2003

250

Scala and Kerkhof, 1998

769

Allen et al., 2001

359

Julie et al., 1991

amoA nxrA narG napA nirK qnorB nosZ nas nifH

gradient gel electrophoresis (Ding et al., 2011). The PCR fragments were approximately 200 bp in length. V3-16S rDNA amplification used 50 ␮L of the PCR reaction. The PCR reaction mixture contained 5.0 ␮L 10× PCR buffer (with Mg2 + ), 4 ␮L 2.5 mmol L−1 dNTP, 1.0 ␮L 10 pmol ␮L−1 primer pairs, 2.0 ␮L amplified DNA, and 0.5 ␮L Taq polymerase (2.5 U ␮L−1 ). PCR product sizes were determined by 1% agarose gel electrophoresis at constant voltage.

periplasmic nitrate reductase gene napA (napA), copper-containing nitrite reductase gene nirK (nirK), nitric oxide reductase gene qnorB (qnorB), nitrous oxide reductase gene nosZ (nosZ), biological nitrogenase gene nifH (nifH), and assimilation nitrate reductase gene nas (nas) (Table 1). These primers were synthesized by Shanghai Invitrogen Biotechnology Co. Ltd., China. The primer concentration was 10 pmol ␮L−1 .

2.3.2. DGGE analysis Bio-RAD’s mutation detection system was used for DGGE analysis. The polyacrylamide (PA) concentration gradient was 6% to 12% (acrylamide: dual acrylamide = 15:4), and the denaturant range was 25–60% (100% denaturant = 40% formamide + 7 M urea). PCR products (20 ␮L) were electrophoresed in 1× TAE buffer for 5 h under 200 V at 62 ◦ C. Samples were stained with SYBR Green1 nucleic acid staining solution for 30 min, and photographed with a UV imager. Bio-Quantity One 4.4.0 (Bio-Rad) software was used to process the images.

2.4.2. Polymerase chain reaction (PCR) PCR products of genes were obtained via PCR amplification using the primers listed in Table 1. The polymerase was the specific cloning enzyme TransTaq-T DNA Polymerase (Beijing TransGen Biotechnology Co. Ltd.). The PCR mixture contained 2.5 ␮L of 10× buffer (with Mg2+ ), 2.0 ␮L 2.5 mmol L−1 dNTP, 0.25–1.0 ␮L 10 pmol ␮L−1 primer pairs, 1.0 ␮L amplified DNA, and 0.25 ␮L Taq polymerase (2.5 U ␮L−1 ). The reaction reached a total volume of 25 ␮L. The PCR program for anammox bacteria 16S rRNA amplification was pre-denaturation at 94 ◦ C for 5 min, denaturation at 95 ◦ C for 30 s, annealing at 57 ◦ C for 45 s, and extension at 72 ◦ C for 1 min, for a total of 40 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of amoA gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 94 ◦ C for 15 s, annealing at 57 ◦ C for 15 s, an extension at 72 ◦ C for 15 s, a total of 40 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of nxrA gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 94 ◦ C for 30 s, annealing at 53 ◦ C for 45 s, an extension at 72 ◦ C for 45 s, a total of 40 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of narG gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 95 ◦ C for 15 s, annealing from 63 ◦ C to 58 ◦ C for 30 s (the annealing temperature dropped from 63 ◦ C to 58 ◦ C and decreased 1 ◦ C each cycle, 40 cycles at the last 58 ◦ C), an extension at 72 ◦ C for 30 s, a total of 45 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of napA gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 95 ◦ C 15 s, annealing from 63 ◦ C to 58 ◦ C for 30 s (the annealing temperature dropped from 63 ◦ C to 58 ◦ C and decreased 1 ◦ C each cycle, 40 cycles at the last 58 ◦ C), an extension at 72 ◦ C for 30 s, a total of 45 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of nirK gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 95 ◦ C for 15 s, annealing from 64 ◦ C to 61 ◦ C for 30 s (the annealing temperature dropped from 64 ◦ C to 61 ◦ C and decreased 1 ◦ C every two cycles per, 35 cycles at the last 61 ◦ C), an extension at 72 ◦ C for 30 s, a total of 41 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of qnorB gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 95 ◦ C

2.3.3. Sequencing and phylogenetic analysis DGGE gel bands were rinsed with sterile by TE buffer, dissolved in 20 ␮L TE buffer, and stored at 4 ◦ C overnight. Recovered DNA was used as a template for PCR amplification. The amplification primers were F338 and R518 without a GC clip, which are the general 16S rDNA V3 variable region primers for most bacteria, and the reaction system, and conditions were the same as above. The PCR products were sent to Saogon Biotech Co., Ltd. (Shanghai) for sequencing. Based on National Center for Biotechnology Information (NCBI) standards, sample sequences were compared with sequences in the Genbank database using BLAST analysis. We subsequently selected strain sequences with the highest similarities to our sample sequences, and conducted analyses using CLUSTAL X, Bioedit, and MEGA 3.1. Neighbor joining analysis was used to construct a phylogenetic tree, and 1000 bootstrap replications were applied to provide support for the tree topology structure. 2.4. Real-time PCR 2.4.1. Primer design Quantitative analysis on samples from multimedia constructed wetlands was performed on 10 target functional gene fragments, including anammox bacteria (ANO16S rRNA), ammonia monooxygenase gene amoA (amoA), nitrite oxidoreductase gene nxrA (nxrA), membrane-bound nitrate reductase gene narG (narG),

182

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

80

80

(a)

influent effluent(1#) effluent(2#)

60

TN (mg L-1)

NH4 + (mg L-1 )

60

40

40

20

20

0 2-Jul-09

2-Aug-09

2-Sep- 09

15

(c)

2-Aug-09

2-Sep-09

1.5

3-Oct-09

(d)

1.2

NO2- (mg L-1)

NO3- (mg L-1)

0 2-Jul-09

3-Oct-09

12

9

0.9

6

0.6

3

0 2-Jul-09

(b)

0.3

2-Aug-09

2-Sep-09

3-Oct-09

0.0 2-Jul-09

2-Aug-09

2-Sep-09

3-Oct-09

Fig. 1. Nitrogen in the constructed wetlands ((a) NH4 + ; (b) TN; (c) NO3 − ; (d) NO2 − ).

for 15 s, annealing from 62 ◦ C to 58 ◦ C for 40 s (the annealing temperature dropped from 62 ◦ C to 58 ◦ C and decreased 1 ◦ C every two cycles, 35 cycles at the last 58 ◦ C), an extension at 72 ◦ C for 30 s, a total of 43 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of nosZ gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 95 ◦ C for 15 s, annealing from 64 ◦ C to 61 ◦ C for 30 s (the annealing temperature dropped from 64 ◦ C to 61 ◦ C decreased 1 ◦ C every two cycles, 35 cycles at the last 61 ◦ C), an extension at 72 ◦ C for 30 s, a total of 41 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of nifH gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 94 ◦ C for 1 min, annealing at 54 ◦ C for 1 min, an extension at 72 ◦ C for 1 min, a total of 35 cycles, and a final extension at 72 ◦ C for 8 min. The procedure of nas gene PCR was a post-denaturation at 94 ◦ C for 5 min, a denaturation at 94 ◦ C for 18 s, annealing at 55 ◦ C for 25 s, an extension at 72 ◦ C for 1 min, a total of 40 cycles, and a final extension at 72 ◦ C for 7 min. The PCR products were verified by electrophoresis on a 1.0% agarose gel at constant pressure. PCR product sizes were determined by 1% agarose gel electrophoresis at constant voltage. 2.4.3. Real-time PCR standard curves The plasmids containing specific ANO 16S rRNA genes, and functional genes (i.e. amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nifH, and nas) were used as quantitative standards. The specific gene fragments amplified from environmental samples were connected to pEASY-T3 (Beijing TransGen Biotechnology Co. Ltd.), and cloned

into Trans1-T1 competent cells (Beijing TransGen Biotechnology Co. Ltd.). The positive suspicious strains were identified as follows: the strains were screened on ampicillin (50 mg L−1 ) plates and incubated at 37 ◦ C overnight, blue-white selection was used for colony/stain selection; PCR and restriction endonuclease analyses were applied for plasmid identification/verification, and products were subsequently sent to Shanghai Biological Engineering Co. Ltd. for sequencing. Sequencing results were used for BLAST homology analysis. Recombinant plasmids were extracted and purified using TIAN pure Mini plasmid kits (Tiangen Biotechnology Co., Ltd.). Mass concentrations of purified recombinant plasmids were measured using nucleic acid and protein quantitative systems. The standard gradients for ANO 16S rRNA, and amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nifH, and nas functional genes are listed in Table 1. The PCR products were stored at −20 ◦ C. Sterile water was used as the negative control. The real-time PCR standard curves for the ANO specific gene, and amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nifH, and nas functional genes were repeated three times. Each R2 -value of each standard curve for each replicate exceeded 0.99, and the variation coefficient (CV) was relatively small. The standard curve melting curves were drawn according to the recombinant plasmids. 2.4.4. Quantification by real-time PCR Real-time PCR was performed in 20 ␮L reaction mixtures containing 10 ␮L SYBR Green I PCR master mix (Applied Biosystems), 0.25 ␮L primers (5 ␮mol L−1 ), and 0.5 ␮L template DNA (DNA

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

183

sample or cloned DNA for standard curves). Each PCR was 40 cycles in total. After each reaction was completed, the melting curve procedure began, i.e. 95 ◦ C for 15 s, 60 ◦ C for 1 min, 95 ◦ C for 15 s, and 60 ◦ C for 15 s. 3. Results and discussion 3.1. Nitrogen removal efficiency The average NH4 + and TN removal over the commissioning and operation period (1 June to 18 July 2009) at 20.0 cm d−1 hydraulic loading was achieved at the following average rates for constructed wetlands Nos. 1 and 2, respectively: 98.3% and 95.5%; and 79.8% and 73.9%. NH4 + and TN removal was respectively decreased to rates of 77.7% (No. 1) and 82.8% (No. 2); and 55.0% (No. 1) and 62.3% (No. 2) during the hydraulic shock loading phase, when hydraulic loading was elevated to 40.0 cm d−1 (19 July to 31 August 2009). NH4 + and TN removal rates over the steady running period (1 September to 28 September 2009) achieved 55.7% (No. 1) and 70.0% (No. 2); and 58.9% (No. 1) and 66.2% (No. 2) (Fig. 1), respectively, and the NH4 + and TN volumetric loading increased by a respective 7.2 and 4.5 times the former levels. Results showed that at lower hydraulic loading, constructed wetland No. 1 achieved more efficient NH4 + and TN removal than No. 2; however hydraulic loading and nitrogen volumetric loading anti-shock capabilities in constructed wetland No. 2 were significantly (P < 0.01) better than in No. 1. In September 2009, the average influent for COD, TN, NH4 + , NO3 − , and NO2 − was 210.1, 68.9, 58.6, 10.1, and 0.23 mg L−1 , respectively. The average COD, TN, NH4 + , NO3 − and NO2 − effluent values were respectively 91.2, 28.4, 25.9, 2.3, and 0.23 mg L−1 in constructed wetland No. 1 samples. The average COD, TN, NH4 + , NO3 − , and NO2 − effluent values in constructed wetland No. 2 samples were respectively 46.9, 23.3, 17.8, 5.0, and 0.49 mg L−1 . Results indicated COD, TN, and NH4 + removal efficiencies were greater in constructed wetland No. 1 relative to No. 2. 3.2. Microbial community 3.2.1. Microbial diversity 16S rDNA DGGE fingerprints derived from the microbial community amplified by PCR showed a variable number of electrophoretic bands isolated from the two multimedia groups in the constructed wetlands. In addition, signal intensity and migration positions differed (Fig. 2). Tan and Ji (2010) suggested that in principle, each isolated fragment represents a microbial species; and biodiversity increases with the number of bands in a sample. In addition, Ji et al. (2011) reported that band signal intensity strengthened with a rise in the number of species in a sample. In the present study, the number of bands detected from the 15 cm layer (1–15, the rhizosphere zone), 60 cm layer (1–60), and 120 cm layer (1–120) in constructed wetland No. 1 was 17, 24, and 23, respectively. The band number generated from the 15 cm layer (2–15), 60 cm layer (2–60) and the 120 cm layer (2–120) in constructed wetland No. 2 was respectively 21, 26, and 25. In No. 2, the band number in the 60 cm layer was greater than in the 120 cm layer, and the 120 cm layer exhibited more bands than the 15 cm layer. Bio-RAD QUANTITY ONE software was used to process the DGGE images. Days coefficient (Cs) of each band was calculated to quantitatively show the degree of similarity between each band. The greater the value of Cs, the higher the similarity and the smaller the difference with layer depth; a 68% similarity was found between the 15 cm and 60 cm layers, and a 57% similarity between the 15 cm and 120 cm layers. Wetland microbial community No. 2 also decreased with layer depth; a 67% similarity was detected between

Fig. 2. DGGE profile comparisons of 16S rDNA amplified fragments ((a) No. 1–15 cm layer; (b) No. 1–60 cm layer; (c) No. 1–120 cm layer; (d) No. 2–15 cm layer; (e) No. 2–60 cm layer; (f) No. 2–120 cm layer).

the 15 cm and 60 cm layers, and 59% similarity between the 15 cm and 120 cm layers. The similarity among the 15 cm, 60 cm, and 120 cm layers of the microbial communities within the two constructed wetland groups was 88%, 79%, and 69%, respectively. 3.2.2. Microbial homology Homology analyses are qualitative assessments, however the comparisons and analyses between multiple nucleic acid sequences are vital to identify and interpret the conserved regions of nucleic acid sequences, and infer evolutionary relationships. Furthermore, the generation of these data is required to draw molecular evolutionary trees (Tan and Ji, 2010). The representative gel bands were selected from the DGGE profiles, and excised from the gel. The cleaned DNA was amplified to fragments of approximately 150 bp, and sequenced. Nine sequences were obtained (Fig. 2), and based on NCBI standards, were compared with data in the Genbank database using BLAST analysis (Table 2). Band A was primarily distributed in the 60 cm and 120 cm layers of constructed wetland No. 1, and the 120 cm layer of constructed wetland No. 2, and the maximum abundance of Band A was in the 60 cm layer of wetland No. 2. The sequence composition of Band A exhibited a 98% similarity with Lactococcus sp. (Fig. 2, Table 2). Lactococcus sp. deoxidize nitrate, and transform nitro compounds into amino compounds (Shin et al., 2005). Band A comprised the dominant microflora for organic nitrogen degradation, and nitrogen denitrification in the multimedia constructed wetland No. 2. Band B distribution was primarily in the 15 cm and 60 cm layers of constructed wetland No. 2, and its abundance decreased with depth (Fig. 2, Table 2). Band B sequence composition showed 99% similarity with Moraxella sp. (Table 2). Moraxella sp. shows high frequency in animal feces, and expresses organophosphorus hydrolase at the cytomembrane surface. In addition, Moraxella sp. are effective in pesticide and nitrobenzene acetic acid degradation (Shimazu et al., 2001), and were the dominant microflora conducting organic nitrogen degradation in both multimedia constructed wetlands. Bands C and D were widely distributed in the 15 cm, 60 cm and the 120 cm layers of both constructed wetland groups (Fig. 2, Table 2). The similarity of Bands C-Acinetobacter sp., and Band D-Acinetobacter sp. was 99% and 96%, respectively. Acinetobacter sp. are the dominant microflora of organic contaminated soils and the species was reported as the primary decomposer of non-degradation organic compounds, and grease substances in

184

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

Table 2 Nucleotide sequence identity of DGGE fragments. Sequenced bands

The strains which have the highest identity from NCBI

A B C D E F G H I

Lactococcus piscium partial 16S rRNA gene, strain R-31597 Moraxella osloensis strain AKAV 10 16S ribosomal RNA gene, partial sequence Acinetobacter junii strain M-1 16S ribosomal RNA gene, partial sequence Acinetobacter soli strain LCR52 16S ribosomal RNA gene, partial sequence Bacillus sp. SM1 16S ribosomal RNA gene, partial sequence Beta proteobacterium OR-214 16S ribosomal RNA gene, partial sequence Brevundimonas sp. X08 16S ribosomal RNA gene, partial sequence Exiguobacterium aurantiacum strain M-4 16S ribosomal RNA gene, partial sequence Exiguobacterium sp. MR-R5 16S ribosomal RNA gene, partial sequence

contaminated soils (Al-Saleh et al., 2009). Furthermore, Bands C and D exhibited the highest sequence similarity to NO3 − transformation microfloras in both constructed wetland samples. Band E was widely distributed in the 15 cm, 60 cm, and 120 cm layers of constructed wetland No. 2. The 60 cm layer abundance was highest, followed by the 120 cm layer, and the 15 cm layer showed the lowest levels (Fig. 2, Table 2). The similarity between Band E and Bacillus sp. was only 91%. Generally speaking, species identification by gene sequences classifies a species as one taxon when consistency is greater than 93%. However, species must exhibit values equal to or exceeding 93% (Tan and Ji, 2010). Thus, Band E could be a new species cultured by the multimedia constructed wetland ecosystem. Bacillus sp. was widely distributed in the contaminated soils, which were polluted by organic matter (Aditi and Aloke, 2010), and showed a positive correlation with NH4 + removal performance in the constructed wetlands (Dong and Reddy, 2010). Band F was widely distributed in the 15 cm and 60 cm layers of the constructed wetland No. 2. The similarity between Band F and ˇ-Proteobacteria sp. was 96%. ˇ-Proteobacteria sp. was the most important microflora in ammoxidation for the water treatment system (Yin and Xu, 2009). Thus, Band F was the dominant microflora of NH4 + removal by ammoxidation matrix in the two groups of constructed wetlands. Band G was also widely distributed in the 15 cm and 60 cm layers of constructed wetland No. 2. Band G and Brevundimonas sp. similarity was 98%. Bands I and H were primarily distributed in the 15 cm and 60 cm layers of constructed wetland No. 2. Bands H and I abundance in the 15 cm layer was not significantly different from the 60 cm layer in constructed wetland No. 1, however Bands I and H abundance in the 60 cm layer was significantly (P < 0.01) greater than Bands I and H in the 15 cm layer in No. 2 (Fig. 2, Table 2). The similarities between Bands H and I to Exiguobacterium sp. were 98% and 99%, respectively. The microflora provided primary soil organic matter degradation (Lopez-Gutierrez et al., 2004; Ji et al., 2010b). Sequencher 5.0 (Gene Codes, Ann Arbor, MI), Clustal X, and the Neighbor-Joining method in the Mega4 software package were adopted to map the phylogenetic tree (Fig. 3). The phylogenetic relationships of Bands A, and H were very close, but band differences were distinct. Phylogeny reconstruction showed Bands A and B in separate lineages, but band similarity was high. Bands B, Band C, and Band D exhibited high band similarity and phylogenetic affinity. 3.3. Distribution patterns of denitrification genes 3.3.1. Distribution patterns In constructed wetland No. 1, absolute qnorB abundance in the rhizosphere zone (15 cm layer) was equivalent to the 60 cm layer. Absolute abundance of qnorB and nas was ∼1.00E + 06 in the 15 cm and 60 cm layers, and ∼1.00E + 05 in than the 120 cm layer (Fig. 4a). In constructed wetland No. 2, nirK was dominantly enriched in the 120 cm layer, and the absolute abundance of this layer was

Similarity 98% 99% 99% 96% 91% 96% 98% 99% 98%

Accession AM943029.1 HQ130446.1 HM030745.1 FJ976560.1 EF424399.1 HM163254.1 GQ426314.1 HM030747.1 GU201835.1

one order of magnitude greater than the rhizosphere zone (15 cm layer) and the 60 cm layer (Fig. 4b). In addition, in both constructed wetlands, ANO, amoA, nxrA, narG, napA, nirK, qnorB, nosZ, nas, and nif functional genes were dominantly enriched in the rhizosphere zone (15 cm layer), and the absolute abundance of the rhizosphere zone (15 cm layer) showed increased average absolute abundance relative to the non-rhizosphere zones (60 cm and 120 cm layers) (Fig. 4). Therefore, the rhizosphere zones (15 cm layer) of both constructed wetlands were more conducive to the absolute enrichment of nitrogen transformation functional genes. Two primary explanations can be provided for increased nitrogen transformation functional genes in the rhizophere zone. First dead roots, dead root shedding, and inorganic and organic matter are secreted by roots in this zone, and provide an important source of nutrients and energy for microbes (Dunbar et al., 2000). Second, due to root interpenetration and continued water and nutrient supply, rhizosphere aeration and moisture conditions are more suitable than other soil profiles (Dunbar et al., 2000; Zhang et al., 2011). Transformation of NH4 + and NO2 − into N2 can occur by anammox bacteria under anoxic conditions, and its 16S rRNA has potential as a marker for anammox processes (Stramma et al., 2008). In the 15 cm layer, the absolute abundance of anammox bacteria 16S rRNA in constructed wetland No. 1 was significantly (P < 0.01) higher than constructed wetland No. 2 (Fig. 4a and b). In the 60 cm layer, absolute abundance of anammox bacteria 16S rRNA in constructed wetland No. 1 was double that of constructed wetland No. 2. However, at 120 cm the absolute abundance of anammox bacteria in constructed wetland No. 2 were twice the value of constructed wetland No. 1. Consequently, in the 60 cm layer, the anammox NH4 + and NO2 − into N2 transformation activity in constructed wetland No. 1 was higher than in constructed wetland No. 2, and in the 15 cm and 120 cm layers, constructed wetland No. 2 showed increased activity. NH4 + to NO2 − oxidation catalyzed by ammonia monooxygenase encoded by the amoA gene is often regarded as the aerobic ammonia oxidation marker (Dionisi et al., 2002). In the 15 cm and 120 cm layers, the absolute abundance of amoA in constructed wetland No. 2 was approximately one and one half times that of constructed wetland No. 1; in the 60 cm layer, the abundance of amoA in constructed wetland No. 1 was more than two times that of constructed wetland No. 2 (Fig. 4a and b). In comparison, in the 15 cm and 120 cm layers, the aerobic ammonia oxidation activity in constructed wetland No. 2 was higher than in wetland No. 1, however increased activity was observed in the 60 cm layer of wetland No. 1. The nitrite oxidase gene (nxrA) is the key gene for nitriteoxidizing bacteria (NOB) in the oxidation of NO2 − to NO3 , and can be a viable marker for nitrite oxidation processes (Poly et al., 2008). In the 15 cm, 60 cm, and 120 cm layers, nxrA absolute abundance in constructed wetland No. 1 showed a respective increase of 8.5, 2.0, and 2.0 relative to wetland No.2 levels (Fig. 4a and b). The

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

185

Fig. 3. Phylogenetic tree of 16S rDNA sequences.

nitrite-oxidizing bacteria in constructed wetland No. 1 were more active and oxidized more NO2 − to NO3 − , which was favorable to active NO2 − removal, but not suitable to NO3 − removal. This may explain why effluent NO2 − in constructed wetland No. 1 was significantly (P < 0.01) lower than constructed wetland No. 2. Aerobic and anaerobic growth of microbes directly affects the activity and enrichment of membrane-bound nitrate reductase (NAR), and periplasmic nitrate reductase (NAP) enzymes (Bell and Ferguson, 1991). Under hypoxic conditions, NAR expression plays a dominant role (Lopez-Gutierrez et al., 2004). NAP plays a dominant role under aerobic conditions (Bell and Ferguson, 1991; Bru et al., 2007). In addition, most of aerobic denitrifying bacteria exhibited heterotrophic nitrification (Robertson et al., 1988). Absolute abundance among the of 15 cm, 60 cm, and 120 cm layers in both groups of constructed wetlands was not significantly different (Fig. 4a

and b). Therefore, the anaerobic denitrification activity of NO2 − to NO3 − denitrification activity in constructed wetland No. 2 was roughly equal to that in constructed wetland No. 1. In the 15 cm and 120 cm layers, the difference in napA absolute abundance difference between thetwo groups was less than double; in the 60 cm layer, the napA absolute abundance and relative richness in wetland No. 2 was 12 times more than constructed wetland No. 1 (Fig. 4c and d). These results indicated that the enrichment of heterotrophic nitrification-aerobic denitrification bacteria resulted in substantially higher NO3 − to NH4 + transformation activity relative to wetland No. 1, via aerobic heterotrophic nitrification and aerobic denitrification pathways. This might be the mechanism increasing NO3 − and NH4 + removal efficiency in constructed wetland No. 2. The Cu-containing nitrite reductase (nirK) gene, which catalyses the reduction reaction of NO2 − to NO, is now widely used as the

186

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

Fig. 4. The richness and relative abundance of nitrogen transformation genes in constructed wetland ((a) No. 1 abundance; (b) No. 2 abundance; (c) No. 1 relative richness; (d) No. 2 relative richness).

marker in NO to NO2 − denitrification reduction (Lam et al., 2007). In the 15 cm and 120 cm layers, nirK absolute abundance in constructed wetland No. 1 was 3.5 and two times that of constructed wetland No. 2; and in wetland No. 2, nirK in the 60 cm layer was double the abundance of wetland No. 1 (Fig. 4a and b). Therefore, the absolute enrichment of the nirK gene flora in the 60 cm layer of constructed wetland No. 2 exhibited high NO2 − reduction activity, catalyzed by the Cu-containing nitrite reductase (nirK) gene. NO reductase, which encodes the qnorB gene, catalyzes NO to N2 O. The qnorB gene has potential to serve as a marker for NO transformation by autotrophic bacteria. Fujiwara and Fukumori, 1996 reported this process is not adequate to control emission of the greenhouse gas N2 O (Fujiwara and Fukumori, 1996). In the 15 cm and 120 cm layers, the absolute abundance of qnorB in constructed wetland No. 2 was 34 and 19 times more than wetland No. 1, and in the 60 cm layer, qnorB showed an increase of 19 times the absolute abundance in group No.1 relative to group No. 2 (Fig. 4a and b). Consequently, the 15 cm and 120 cm layers in constructed wetland No. 2 may release more N2 O through NO transformation, and the 60 cm layer will release decreased N2 O through NO transformation. The transformation reaction of N2 O to N2 catalyzed by N2 O reductase encoded by the nosZ gene is conducive to facilitate the control of greenhouse gas emissions. In the 15 cm layer, nosZ absolute abundance was increased by 2.5 times in wetland No. 1 compared with constructed wetland No. 2; in the 60 cm layer, nosZ

absolute abundance in constructed wetland No. 2 exhibited a 3-fold increase relative to wetland No. 1; in the 120 cm layer, the difference in nosZ absolute abundance was not significant. In constructed wetland No. 1, based on inferred qnorB and nosZ, we determined microfloras in the 15 cm layer rich in the nosZ gene transformed N2 O released by the 60 cm layer into N2 . In constructed wetland No. 2, most N2 O released by the 120 cm layer was transformed into N2 by microfloras exhibiting the nosZ gene in the 60 cm layer. The coupling transformation of the nosZ functional gene community along the layer facilitated emission control of N2 O in wetland No. 2 (Fig. 4a and b). The nas codase dissimilatory nitrate reductase gene catalyzes the NO2 − to NO3 − via NH4 + reaction process (Cabello et al., 2004). This process is conducive to NO3 − transformation in wastewater, but not favorable to NH4 + and TN removal (Cabello et al., 2004). In the 15 cm and 120 cm layers, the nas absolute abundance in constructed wetland No. 2 exhibited a two to 3.5 level increase relative to constructed wetland No. 1; in the 60 cm layer, nas absolute abundance was seven times more than in wetland No. 2. Therefore, the high NO3 − to NH4 + transformation activity areas in group No. 2 were the 15 cm and 120 cm layers, and the 60 cm layer exhibited increased activity in group No. 1 (Fig. 4a and b). In the two constructed wetland groups, nifH was only distributed in the rhizosphere zone. nifH is the bacterial reductase gene of the nitrogen fixation molecule. Bacterial nitrogen fixation is

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

187

Fig. 5. The Simpson dominance index and Margalef index of nitrogen transformation gene ((a) No. 1 constructed wetland; (b) No. 2 constructed wetland).

primarily associated with root systems (Rubio and Ludden, 2002). In the rhizosphere zone of the two groups, nifH abundance was 7.01 × 107 copies g−1 (No. 1) and 9.83 × 106 copies g−1 (No. 2). Microbial N2 transformation to NH4 + in constructed wetland No. 1 was seven times higher than constructed wetland No. 2, which did not promote NH4 + wastewater removal. 3.3.2. Genes richness and dominance The genetic richness index, Margalef (Ma), estimates gene species richness, and the number of gene species increases with an increase in the richness index. Ma values for the 60 cm layer were higher than other layers in constructed wetland No. 1, whereas, in constructed wetland No. 2, the 120 cm layer exhibited the highest Ma (Fig. 5a and b). These results indicated the 60 cm layer of No.1 and the 120 cm layer of No. 2 were beneficial to rare gene enrichment in each system. The following rare genes exhibited increased relative abundance: amoA, nxrA, narG, qnorB, and nas (Fig 4c and d). The reaction pathway catalyzed by codase encoded by these functional genes was the major rate-limiting process in nitrogen transformation for each individual system. The Simpson dominance index provides a dominance level index of genes and bacterial community of increased frequency in a community (Keylock, 2005). The higher the C value of the Simpson dominance index, the more concentrated the genes and bacterial species. The Simpson dominance index (C) in the rhizosphere zone (15 cm layer) was higher than all other layers of constructed wetland No. 1; in the non-rhizosphere zone, and C in the 120 cm layer was higher than the 60 cm layer. In constructed wetland No. 2, the 60 cm layer C value was higher than the other layers (Fig. 5a and b). Therefore, the 15 cm and 60 cm layers of wetland No. 1 were favorable for relative enrichment of dominant genes, and the 120 cm layer of wetland No. 2 was conducive to the relative enrichment of dominant genes. ANO, napA, nirK, nosZ, and nifH showed increased relative abundance (Fig. 4c and d). The reaction pathway catalyzed by codase encoded by these functional genes was the key process in nitrogen transformation for each individual system. 4. Conclusion The results of this study indicated that constructed wetlands comprised of overall increased total filling with raw material containing zero-valent iron multimedia filler, resulted in better hydraulic and nitrogen volume loading shock resistance. The two groups of constructed wetland shared similar distribution patterns, and Lactococcus sp. and Moraxella sp. were the dominant

organic nitrogen and denitrification flora; Acinetobacter sp. were the dominant NO3 − to NH4 + transformation flora; and Bacillus sp. and ˇ-Proteobacteria sp. were the dominant NH4 + removal flora in the two constructed wetland groups. The qnorB and nas functional genes were predominantly enriched in the 15 cm and 60 cm layers of the constructed wetland, which was prepared by multi-media packing of 1% zero-valent iron. The nirK was predominantly enriched in the 60 cm layer of the constructed wetland, which was prepared by multi-media packing of 2% zero-valent iron. Other functional nitrogen transformation genes were predominantly enriched in the 15 cm layer of constructed wetlands. These results provide useful information for identifying microbal microbial floras, denitrification functional genes distribution, and improve denifrification in constructed wetlands. Acknowledgements This study was supported by the National Natural Science Foundation of China (no. 51179001), the Transportation Construction Project of Science and Technology (no. 20113186701290), and Jilin Province Transprotation Science and Technology Projects (no. 2011-1-7). References Aditi, B., Aloke, K.G., 2010. Isolation and characterization of hyper phenol tolerant Bacillus sp. from oil refinery and exploration sites. J. Hazard. Mater. 176, 85–91. Allen, A.E., Booth, M.G., Frischer, M.E., Verity, P.G., Zehr, J.P., Zani, S., 2001. Diversity and detection of nitrate assimilation genes in marine bacteria. Appl. Environ. Microbiol. 67, 5343–5348. Al-Saleh, E., Drobiova, H., Obuekwe, C., 2009. Predominant culturable crude oildegrading bacteria in the coast of Kuwait. Int. Biodeterior. Biodegrad. 63, 400–406. Bell, L.C., Ferguson, S.J., 1991. Nitric and nitrous oxide reductases are active under aerobic conditions in cells of Thiosphaera pantotropha. Biochem. J. 273, 423–427. Blankenberg, A.G.B., Haarstad, K., Sovik, A.K., 2008. Nitrogen retention in constructed wetland filters treating diffuse agriculture pollution. Desalination 226, 1–3. Braker, G., Tiedje, J.M., 2003. Nitric oxide reductase (norB) genes from pure cultures and environmental samples. Appl. Environ. Microbiol. 69, 3476–3483. Bru, D., Sarr, A., Philippot, L., 2007. Relative abundances of proteobacterial membrane-bound and periplasmic nitrate reductases in selected environments. Appl. Environ. Microbiol. 73, 5971–5974. Cabello, P., Pino, C., Olmo-Mira, M.F., Castillo, F., Roldan, M.D., Moreno-Vivian, C., 2004. Hydroxylamine assimilation by Rhodobacter capsulatus E1F1requirement of the hcp gene (hybrid cluster protein) located in the nitrate assimilation nas gene region for hydroxylamine reduction. J. Biol. Chem. 279, 45485–45494. Dong, X.L., Reddy, G.B., 2010. Soil bacterial communities in constructed wetlands treated with swine wastewater using PCR-DGGE technique. Bioresour. Technol. 101, 1175–1182.

188

G. Ji et al. / Ecological Engineering 44 (2012) 179–188

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 competitive PCR. Appl. Environ. Microbiol. 68, 245–253. Ding, L.L., Zhou, Q.X., Wang, L., Zhang, Q.R., 2011. Dynamics of bacterial community structure in a full-scale wastewater treatment plant with anoxic-oxic configuration using 16S rDNA PCR-DGGE fingerprints. Afr. J. Biotechnol. 10, 589–600. Dunbar, J., Ticknor, L.O., Kuske, C.R., 2000. Assessment of microbial diversity in four southwestern United States soils by 16S rRNA gene terminal restriction fragment analysis. Appl. Environ. Microbiol. 66, 2943–2950. Fujiwara, T., Fukumori, Y., 1996. Cytochrome cb-type nitric oxide reductase with cytochrome c oxidase activity from Paracoccus denitrificans ATCC 35512. J. Bacteriol. 178, 1866–1871. Henry, S., Bru, D., Stres, B., Hallet, S., Philippot, L., 2006. Quantitative detection of the nosZ gene, encoding nitrous oxide reductase, and comparison of the abundances of 16S rRNA, narG, nirK, and nosZ genes in soils. Appl. Environ. Microbiol. 72, 5181–5189. Hsu, C.B., Hsieh, H.L., Yang, L., Wu, S.H., Chang, J.S., Hsiao, S.C., Su, H.C., Yeh, C.H., Ho, Y.S., Lin, H.J., 2011. Biodiversity of constructed wetlands for wastewater treatment. Ecol. Eng. 37, 1533–1545. IWA, 2000. Constructed wetlands for pollution control: processes, performance, design and operation. International Water Association Scientific and Technical Report No. 8. London, UK: IWA Publishing. Jenssen, P.D., Krogstad, T., Paruch, A.M., Maehlum, T., Adam, K., Arias, C.A., Heistad, A., Jonsson, L., Hellstrom, D., Brix, H., Yli-Halla, M., Vrale, L., Valve, M., 2010. Filter bed systems treating domestic wastewater in the Nordic countries – performance and reuse of filter media. Ecol. Eng. 36, 1651–1659. Ji, G.D., Sun, T., Zhou, Q.X., Sui, X., Chang, S.J., Li, P.J., 2002. Constructed subsurface flow wetland for treating heavy oil-produced water of the liaohe oilfield in China. Ecol. Eng. 18, 459–465. Ji, G.D., Zhou, Y., Tong, J.J., 2010a. Nitrogen and phosphorus adsorption behavior of ceramsite material made from coal ash and metallic iron. Environ. Eng. Sci. 27, 871–878. Ji, G.D., Zhai, F.M., Wang, R.J., Ni, J.R., 2010b. Sludge granulation and performance of a low superficial gas velocity sequencing batch reactor (SBR) in the treatment of prepared sanitary wastewater. Bioresour. Technol. 101, 9058– 9064. Ji, G.D., Tong, J.J., Tan, Y.F., 2011. Wastewater treatment efficiency of a multi-media biological aerated filter (MBAF) containing clinoptilolite and bioceramsite in a brick-wall embedded design. Bioresour. Technol. 102, 550–557. Julie, D., Kirshtein, H.W., Paerl, J.Z., 1991. Amplification, cloning, and sequencing of a nifH segment from aquatic microorganisms and natural communities. Appl. Environ. Microbiol. 57, 2645–2650. Keylock, C.J., 2005. Simpson diversity and the Shannon–Wiener index as special cases of a generalized entropy. OIKOS 109, 203–207. Konnerup, D., Brix, H., 2010. Nitrogen nutrition of Canna indica: effects of ammonium versus nitrate on growth, biomass allocation, photosynthesis, nitrate reductase activity and N uptake rates. Aquat. Bot. 92, 142–148. Lam, P., Jensen, M.M., Lavik, G., McGinnis, D.F., Muller, B., Schubert, C.J., Amann, R., Thamdrup, B., Kuypers, M.M.M., 2007. Linking crenarchaeal and bacterial nitrification to anammox in the Black Sea. Proc. Natl. Acad. Sci. U.S.A. 104, 7104–7109. Li, W., Hua, T., Zhou, Q.X., Zhang, S.G., Rong, W.Y., 2011. Toxicity identification and high-efficiency treatment of aging chemical industrial wastewater from the hangu reservoir. Chin. J. Environ. Qual. 40, 1714–1721. Liu, R., Zhou, Q.X., 2009. Fluxes and influencing factors of ammonia emission from monosodium glutamate production in Shenyang. Chin. Bull. Environ. Contam. Toxicol. 85, 279–286.

Lin, Y.F., Jing, S.R., Wang, T.W., Lee, D.Y., 2002. Effects of macrophytes andexternal carbon sources on nitrate removal from groundwater in constructed wetlands. Environ. Pollut. 119, 413–420. Lopez-Gutierrez, 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. Methods 57, 399–407. McDevitt, C., Burrell, P., Blackall, L.L., McEwan, A.G., 2000. Aerobic nitrate respiration in a nitrite-oxidising bioreactor. FEMS Microbiol. Lett. 184, 113–118. Mulder, A., Van de Graaf, A.A., Robertson, L.A., Kuenen, J.G., 1995. Anaerobic ammonium oxidation discovered in a denitrifying fluidized-bed reactor. FEMS Microbiol. Ecol. 16, 177–183. Poly, F., Wertz, S., Brothier, E., Degrange, V., 2008. First exploration of Nitrobacter diversity in soils by a PCR cloning-sequencing approach targeting functional gene nxrA. FEMS Microbiol. Ecol. 63, 132–140. Robertson, L.A., Vanniel, E.W.J., Torremans, R.A.M., Kuenen, J.G., 1988. Simultaneous nitrification and denitrification in aerobic chemostat cultures of thiosphaerapantotropha. Appl. Environ. Microbiol. 54, 2812–2818. Rubio, L.M., Ludden, P.W., 2002. The gene products of the nif regulon. In: Leigh, G.J. (Ed.), Nitrogen Fixation at the Millennium. Elsevier, New York, pp. 421–445. Shimazu, M., Mulchandani, A., Chen, W., 2001. Simultaneous degradation of organophosphorus pesticides and p-nitrophenol by a genetically engineered Moraxella sp. with surface-expressed organophosphorus hydrolase. Biotechnol. Bioeng. 76, 318–324. Shin, K.H., Lim, Y., Ahn, J.H., Khil, J., Cha, C.J., Hur, H.G., 2005. Anaerobic biotransformation of dinitrotoluene isomers by Lactococcus lactis subsp lactis strain 27 isolated from earthworm intestine. Chemosphere 61, 30–39. State Environmental Protection Administration, 2002. Water and Exhausted Water Monitoring Analysis Method, Fourth ed. Environmental Science Press, China. Stramma, L., Johnson, G.C., Sprintall, J., Mohrholz, V., 2008. Expanding oxyenminimum zones in the tropical oceans. Science 320, 655–658. Scala, D.J., Kerkhof, L.J., 1998. Nitrous oxide reductase (nosZ) gene-speci¢c PCR primers for detection of denitri¢ers and three nosZ genes from marine sediments. FEMS Microbiol. Lett. 162, 61–68. Tan, Y.F., Ji, G.D., 2010. Bacterial community structure and dominant bacteria in activated sludge from a 70 ◦ C ultrasound-enhanced anaerobic reactor for treating carbazole-containing wastewater. Bioresour. Technol. 101, 174–180. Throback, I.N., Enwall, K., Jarvis, A., Hallin, S., 2004. Reassessing PCR primers targeting nirS, nirK and nosZ genes for community surveys of denitrifying bacteria with DGGE. FEMS Microbiol. Ecol. 49, 401–417. Tsushima, I., Kindaichi, T., Okabe, S., 2007. Quantification of anaerobic ammoniumoxidizing bacteria in enrichment cultures by real-time PCR. Water Res. 41, 785–794. Davidsson, T.E., Stahl, M., 2000. The influence of organic carbon on nitrogen transformations in five wetland soils. J. Am. Soil Sci. Soc. 64, 1129–1136. US EPA, 1993. Office of Water. Subsurface flow constructed wetlands for wastewater treatment: a technology assessment. Verhoeven, J.T.A., Meuleman, A.F.M., 1999. Wetlands for wastewater treatment: opportunities and limitations. Ecol. Eng. 12, 5–12. Vymazal, J., 2011. Long-term performance of constructed wetlands with horizontal sub-surface flow: ten case studies from the Czech Republic. Ecol. Eng. 37, 54–63. Yan, T.F., Fields, M.W., Wu, L.Y., Zu, Y.G., Tiedje, J.M., Zhou, J.Z., 2003. Molecular diversity and characterization of nitrite reductase gene fragments (nirK and nirS) from nitrate and uranium-contaminated groundwater. Environ. Microbiol. 5, 13–24. Yin, J., Xu, W.F., 2009. Ammonia biofiltration and community analysis of ammoniaoxidizing bacteria in biofilters. Bioresour.Technol. 100, 3869–3876. Zhang, H.G., Cui, B.S., Hong, J.M., Zhang, K.J., 2011. Synergism of natural and constructed wetlands in Beijing. China. Ecol. Eng. 37, 128–138.