Comparing abundance, composition and environmental influences on prokaryotic ammonia oxidizers in two subtropical constructed wetlands

Ecological Engineering 90 (2016) 336–346

Contents lists available at ScienceDirect

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

Comparing abundance, composition and environmental influences on prokaryotic ammonia oxidizers in two subtropical constructed wetlands Lan-Feng Fan a , Hsing-Ju Chen a,b , Hwey-Lian Hsieh a , Hsing-Juh Lin b,1 , Sen-Lin Tang a,∗ a b

Biodiversity Research Center, Academia Sinica, No. 128 Academia Rd., Sec. 2, Nankang, Taipei 115, Taiwan Department of Life Sciences, National Chung Hsing University, Taichung 402, Taiwan

a r t i c l e

i n f o

Article history: Received 4 March 2015 Received in revised form 17 November 2015 Accepted 26 January 2016 Available online 19 February 2016 Keywords: Ammonia-oxidizing archaea (AOA) Ammonia-oxidizing bacteria (AOB) Constructed wetland (CW) Denaturing gradient gel electrophoresis (DGGE) Quantitative PCR (qPCR)

a b s t r a c t Microbial communities in wastewater treatment can be critical indicators for proper design and functional management of constructed wetlands (CWs). The aim of this study was to examine changes in abundance and composition of ammonia-oxidizing prokaryotes (ammonium-oxidizing archaea and bacteria, AOA and AOB, respectively) and how these relate functionally to nutrient depletion, by comparing a series of treatment cells in two CWs in subtropical Taiwan. Abundance and composition of AOA and AOB were measured by amoA-gene targeted denaturing gradient gel electrophoresis (DGGE) and quantitative PCR, respectively. Prokaryotic composition (including archaea and bacteria) was also determined by 16S rRNA-gene targeted DGGE, while concurrently measuring environmental characteristics. Dominant AOA and AOB were related to uncultured Nitrososphaera and Nitrosomonas groups, respectively. Treatment cells of the two CWs had special clusters with regards to composition of bacteria and AOA, whereas composition of archaea and AOB were evenly distributed in the wetland enclosing a small treatment area. Composition of AOA reflected distinct environmental requirements compared to archaeal composition. Based on abundance analysis, there were more AOB in middle treatment cells, with abundance highly correlated to concentrations of ammonium and phosphate. However, AOA was distributed throughout most of the treatment cells, and was poorly correlated to measured environmental characteristics. With regards to distribution, community complexity, environmental affiliation and environmental demands, we inferred that AOA may be more critical than AOB for depleting nutrients in the CWs. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Constructed wetlands (CWs) are often nutrient-enriched, with higher nutrient concentrations than natural wetlands (Kadlec and Wallace, 2009). A prominent function of CWs is nitrogen removal (Faulwetter et al., 2009; Villalobos et al., 2013). Inside CWs, nitrogen removal mainly occurs through microbial transformations, i.e. nitrification and denitrification (Ruiz-Rueda et al., 2009; Wang et al., 2011). Ammonia-oxidizing archaea and bacteria (AOA and AOB) are commonly regarded as major drivers of nitrification in sediment.

∗ Corresponding author. Tel.: +886 227893863. E-mail addresses: [email protected] (H.-J. Lin), [email protected] (S.-L. Tang). 1 Equal contribution. http://dx.doi.org/10.1016/j.ecoleng.2016.01.034 0925-8574/© 2016 Elsevier B.V. All rights reserved.

Ammonia oxidation is the first step in nitrification; this reaction is a rate-limiting step for transforming ammonia to nitrite, linking mineralization of organic matter-derived nitrogen in estuarine and continental shelf sediments to its ultimate removal as gaseous products by denitrification or anammox (Fan et al., 2006). The key enzyme in the reaction is ammonia monooxygenase, which was discovered in AOBs in the late 19th century and used since 1997 as a molecular marker of AOB (Rotthauwe et al., 1997). Using metagenomics approaches, ammonia monooxygenase has also been identified in AOAs in recent years and a culturable marine AOA (categorized into an archaeal phylum, Thaumarchaeota; Treusch et al., 2005) known to have specific activity of ammonia oxidation. Quantitative PCR (qPCR) of the amoA gene, which encodes subunit A of ammonia monooxygenase, was done to determine AOA and AOB abundance. The QPCR data reflected relative abundance of AOA and AOB in various environments, e.g. aquatic and terrestrial

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

environments including forest soil, estuarine sediment, mangrove, salt marsh, and other natural wetlands (Bernhard et al., 2007; Moin et al., 2009; Li et al., 2011; Sims et al., 2012). Environmental characteristics are critical in shaping the ammonia-oxidizer community. Although AOA usually grow at a wide ammonium range, growth of AOB was prominent only at a higher concentration of 14.3 mM (Verhamme et al., 2011). Increasing oxygen concentrations enhanced growth of AOB, whereas growth of AOA was almost oxygen insensitive. In addition, AOA are detected over a wide pH range, whereas AOB are neutrophilic and their highest growth rate occurs at pH 7 to 7.5 (Prosser and Nicol, 2012). Biotic factors are also important for prokaryotes. For example, based on analysis of potential activity and diversity of ammonia-oxidizer communities in CWs, variation of plant species influences structure and function of ammonia oxidizers (RuizRueda et al., 2009; Wang and Gu, 2013). To improve efficiency with which CWs remove redundant nutrients, a series of treatment cells were designed, namely depositional, heavily vegetated, open-water and ecological ponds (Faulwetter et al., 2009; Hsu et al., 2011). Varying the organization of treatment cells facilitates comparisons of their distinctive functionality and unique chemistry. Wetland hydrological conditions and dimensions are crucially important, both from engineering and ecological perspectives; to maintain treatment efficiency, surface flow design requirements are normally based either on hydraulic retention time or total wetland area (Economopoulou and Tshrintzis, 2004; Kadlec and Wallace, 2009). The Hsin-Hai II Constructed Wetland (HS2) and Daniaopi Constructed Wetland (DN) are free-water-surface integrated wetlands in the riparian zone of the Danshuei River system in northern Taiwan. Although both are effective for nutrient removal,

337

performance of these two CWs varies, as does that of various sub-cells within each CW (Hsu et al., 2011; Hsueh et al., 2014). Therefore, we hypothesized that the abundance and composition of prokaryotic ammonium-oxidizers differs between the two constructed wetlands and among treatment cells of each constructed wetland. The 16S rRNA- and amoA-genes were targeted in denaturing gradient gel electrophoresis (DGGE) as a proxy to describe composition of prokaryotes and ammonium oxidizers. Relative abundance of archaeal- and bacterial-ammonium oxidizers was measured by amoA-gene targeted qPCR. In addition to categorizing compositional variation of ammonium oxidizers in the two CWs, environmental characteristics were also measured to correlate these with possible factors influencing the presence of ammonium-oxidizing communities and their relative abundance. 2. Materials and methods 2.1. Study sites The Hsin-Hai II Constructed Wetland (HS2) and Daniaopi Constructed Wetland (DN) are located in the riparian zone on the eastern bank of the Da-Han River, one tributary of the Danshuei River in northern Taiwan (Fig. 1a). With <6 km between site, these two CWs have less difference than channel versus riparian features. Both CWs are five-cell, free-water-surface integrated wetlands and were constructed in September and November 2006, respectively. The five sequential cells in HS2 consist of a depositional pond, heavily vegetated pond I, an open-water pond, heavily vegetated pond II, and an ecological pond (cells HS2-1–HS2-5, Fig. 1b). In the DN, the depositional pond (DN1) was followed by an open-water pond (DN2). The heavily vegetated pond I (DN3) was divided into

Fig. 1. Locations (a) and configurations of the Hsin-Hai II Constructed Wetland (HS2) (b) and Daniaopi Constructed Wetland (DN) (c), and two natural wetlands (NW): Gu-Ting (GT) and Xiu-Lang (XL).

338

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

four parallel sub-cells (DN3-A, -B, -C and -D); each of the sub-cells received wastewater from the DN2 and discharged outflow into the heavily vegetated pond II (DN4) and then into the ecological pond (DN5; Fig. 1c). After treatment, wastewater effluent from these two CWs flowed into a drainage channel and then into the Da-Han River. Although both CWs had similar configurations, the DN covered a larger area with a higher flow rate and shorter hydraulic retention time than HS2 (11000 vs 2700 m3 day−1 for flow rate and 5.0 vs 5.5 day for hydraulic retention, Hsu et al., 2011). In addition, for control sites, we studied two natural wetlands, namely Gu-Ting (GT) and Xiu-Lang (XL) located in the riparian zone on the eastern bank of the Xin-Dian River, one of the tributaries of the Danshuei River. 2.2. Environmental characteristics On June 16–18, 2009, samples were collected from the series of treatment cells in HS2 and DN on the Da-Han River, as well as the two natural wetlands chosen as control sites (Fig. 1a) to analyze microbial communities and record environmental measurements. The latter were measured in the field using portable instruments, including dissolved oxygen (DO; WTW oxi 3302, Weilheim, Germany), conductivity (WTW Cond 330i), pH (WTW pH340), and redox potential (Clean ORP30, Los Angeles, CA, USA). Water samples were collected from the outflow of each cell, immediately placed on ice in a cooler, and brought back to the laboratory for analysis of ammonium (NH4 + ), oxidized nitrogen (NOx − = NO3 − + NO2 − ), and phosphate (PO4 −3 ) content, following standard methods of the APHA (Clesceri et al., 1998). 2.3. Extraction and quantification of DNA Approximately 2 g of sediment were used for environmentalDNA extraction and purification using an SDS-chloroform method, as described with some modifications. Concentrations of DNA were quantified using a NanoDrop ND-1000 UV–vis Spectrophotometer (Thermo Scientific, Rockwood, TN, USA). 2.4. Polymerase chain reaction (PCR) of amoA and 16S rRNA genes of archaea and bacteria The universal primers of 16S ribosomal RNA gene, 21F/915R and 341F/907R, were used for archaeal and bacterial prokaryotes, respectively, and the universal primers of amoA-gene, ArchamoAF/Arch-amoAR and amoA-1F/amoA-2R, for AOA and AOB communities, respectively (DeLong, 1992; Muyzer et al., 1993; Muyzer et al., 1995; Rotthauwe et al., 1997; Chan et al., 2002; Francis et al., 2005). The PCR was conducted in a PxE Thermal Cycler (Thermo Electron Corp., Milford, MA, USA). The major cycling program for each primer set was conducted according to previous protocols (Francis et al., 2005). Furthermore, a 36–41-bp GC-clamp was added at the 5 -end of the same forward primers, used in the PCR as mentioned above (Muyzer et al., 1993). 2.5. Denaturing gradient gel electrophoresis (DGGE) of 16S rRNA and amoA genes for archaea and bacteria The DGGE revealed microbial compositions based on amoA and 16S rRNA genes. Gels were made with denaturing gradients ranging from 15 to 55% and from 25 to 60% for archaeal and bacterial 16S RNA genes, respectively, and were run at 65 V for 13 h. Other gels with 30–50% and 40–60% denaturing gradients for archaeal and bacterial amoA genes were run at 50 V for 20 h. All gels were stained with SYBR Gold solution and high-resolution images were captured and analyzed using an ImageQuant 300 Imager system with ImageQuant TL software (Amersham Bioscience, Buckinghamshire, UK). After normalizing gels, only bands with a peak height intensity

exceeding 5.0% of the strongest band in each lane were adopted for further analyses. The number of bands present was recorded per 5 mm of gel scale. 2.6. QPCR of amoA gene for archaea and bacteria A qPCR for amoA genes of archaea (AOA) and another for bacteria (AOB) were done using SYBR Permix Ex Taq (TaKaRa). Two duplicated reactions were performed in each sample. To avoid pipetting errors prior to qPCR, total DNA of the sediment sample was diluted 10-fold (5 ␮l DNA was mixed with 45 ␮l ddH2 O). The primers used are described above. The qPCR reaction solution contained 2 ␮l of diluted sediment DNA, 0.2 ␮M of each primer and 1× SYBR Permix Ex Taq in a total volume of 20 ␮l. The QPCR was performed in an ABI 7300 Real-Time PCR System (Applied Biosystems) under the following conditions: 50 ◦ C for 2 min, 94 or 95 ◦ C for 5 min; the major cycling program for each primer set was as above, followed by 95 ◦ C for 15 s, 60 ◦ C for 1 min, 95 ◦ C for 15 s, and 60 ◦ C for 15 s for dissociation. A standard curve was constructed with a series dilution of template DNA from PCR products of known concentration. Data were analyzed using ABI 7300 System SDS Software (Applied Biosystems). 2.7. Statistical analysis Differences in DGGE profile patterns among sampling settings were examined using one-way analyses of similarities (ANOSIM) based on Bray–Curtis similarity coefficients generated from the abundance matrix after presence-absence transformation (Clarke and Ainsworth, 1993). Three settings were examined: (a) three sites, two constructed wetlands, HS2 and DN, and a natural wetland group; and (b) treatment cells in each constructed wetland. Because pairwise tests for each two treatment cells did not detect significance, cells with similar prokaryotic composition were combined for another ANOSIM test: (c) cell groups in each constructed wetland. To examine spatial variation among prokaryote compositions, non-metric multidimensional scaling (NMDS) were also applied to similarity matrixes to determine similarity of sites with respect to DGGE patterns. Relationships between seven environmental characteristics (conductivity, DO, pH, redox potential, ammonium, oxidized nitrogen and phosphate) and each of the four DGGE profile patterns were determined using the BIO-ENV routine, which uses a generalized Mantel test to examine associations between biotic and abiotic data expressed as Spearman rank correlation coefficients (Clarke and Ainsworth, 1993). Matrices of DGGE profile patterns came directly from the aforementioned analysis, and environmentalfactor matrices were based on Euclidean distance (Clarke and Gorley, 2001). All of the ANOSIM, NMDS and BIO-ENV analyses were performed using PRIMER version 6 software package (Clarke and Gorley, 2001). Copy numbers of AOA and AOB of various wetland sites were also compared using a nonparametric ANOVA (Kruskal–Wallis test; Hollander and Wolfe, 1999). When results were significant, a multiple-means comparison was performed using Dunn’s test (Hollander and Wolfe, 1999). To evaluate relationships between amoA abundance and environmental characteristics, Pearson correlation analyses were done using log10 transformed amoA copy numbers. All statistical analyses were conducted using the software package SAS 9.1 (SAS Institute, 2003, Cary, NC, USA). 2.8. Sequencing and phylogenetic analysis The most prominent major band of DGGE profile patterns, widely distributed in each site, was sequenced to confirm the amplicon of four target microbial communities. The DGGE bands

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

of archaeal and bacterial 16S rRNA were excised, cloned and sequenced from the profile HS2-1.1, HS2-5.1, respectively, whereas bands of archaeal and bacterial amoA were from profiles DN5.2 and SL.2, respectively. Excised bands were dissolved in 20 ␮l of deionized H2 O and incubated overnight at 4 ◦ C to elute the DNA from the gel, and 5 ␮l of the solution was used in a PCR reaction with corresponding GC-clamped primers. These amplicons were purified with a Clean Up-M kit (Viogene, New Taipei City, Taiwan) and cloned with a TOPO TA Cloning Kit (Invitrogen Carlsbad, CA, USA), as specified by manufacturers. Using dye terminator chemistry, DNA sequences were determined by direct sequencing with a DNA sequencer model 3730 (Applied Biosystems, Foster City, CA, USA). We used BLAST to compare sequences obtained with public database of sequences to determine potential taxonomic affiliations (Altschul et al., 1990). These sequences were deposited in the NCBI nucleotide sequence databases under accession numbers KM491494–KM491507. The 16S rRNA gene sequences were aligned as DNA sequences before they were aligned against their closest relatives. Fourteen sequences for archaea were chosen from the database and 16 sequences for bacteria. For the amoA gene, sequences were aligned as amino acid sequences, with seven amoA gene sequences chosen from the database for AOA, and 12 for AOB. Phylogenetic analyses were performed using MEGA version 5.0 (Tamura et al., 2011) using only unambiguously aligned positions. These matrices were analyzed by distance-matrix of neighbor joining (NJ; Saitou and Nei, 1987), maximum parsimony (MP; Fitch, 1971), and maximum likelihood (ML; Felsenstein, 1981). At this point, 1000 bootstraps were performed for NJ, MP, and ML algorithms to analyze relatedness of the 16S rRNA and amoA genes. Multifurcations were created at the appropriate basal node only when branching patterns were supported in >70% of bootstrap re-samplings.

3. Results 3.1. Environmental characteristics Environmental characteristics differed by site, especially at treatment cells of constructed wetlands in the Danshuei River system (Table 1). The DO detected little variation among treatment cells in the HS2 wetland, and was lower than that in the natural wetlands. The DO in the DN wetlands generally increased with the cell series and the level in the latter cells was higher than that in the natural wetlands. Water conductivity was higher in the two CWs than in natural sites. With the exception of extreme values at the third cell, water conductivity generally decreased with the series of the treatment cells in the two CWs. Water pH had limited variation in cell series in HS2, which was lower than that in the other wetlands, whereas there was a comparatively high pH value at the end cell of the DN wetland. Water redox potential decreased from the first cell, peaked at the third cell in the HS2 wetland, and was lower than that in the natural wetland. Water redox potential in the HS2 wetland followed a similar trend as that in the DN wetland. However, the level was comparatively higher and near that of natural sites. Concentration of ammonium was higher in both CWs than in the natural sites. Nearly 24 and 87% of ammonium concentrations were removed after treatment in the cells of the HS2 and DN wetlands, respectively (Table 1). Ammonium concentration was comparatively high in the HS2 wetland, where the concentration peaked in the third cell and then decreased with the cell series after the third cell. In the DN wetland, ammonium concentration peaked in two parallel cells of the third treatment series and decreased to the level near that in the natural wetlands. With the exception of a peak in the last cell, oxidized-nitrogen concentration was generally

339

low in the HS2 wetland. In the DN wetland, oxidized-nitrogen concentration decreased to the lowest in the third-cell group, but peaked in the cell series that followed. Phosphate concentration increased in the second cell and remained at a constant level in the subsequent cells of the HS2 wetland. In the DN wetland, phosphate concentration was generally low, except for a peak in the last cell. 3.2. Phylogenetic affiliation and spatial composition of archaeal and bacterial 16S rRNA genes Based on band patterns of DGGE profiles, composition of archaeal 16S rRNA genes varied between constructed and natural wetlands groups. Sequences from the most major band of these profile patterns corresponded to the 16S rRNA gene of uncultured Methanosarcinales sp., Methanosaeta concilii and Methanothrix soehngenii (96–98% similarity of DNA sequence; Eggen et al., 1990; Barber et al., 2011; Zhang et al., 2011b). Band sequences were also similar to samples from estuarine, lake and peatland sediments (95–99%), and digested sludges (95–99%; Fig. 2; Schwarz et al., 2007; Li et al., 2011; Yavitt et al., 2012). The DGGE profile patterns of archaeal 16S rRNA gene were significantly different for the HS2 wetland compared to the DN wetland and natural wetlands groups (Fig. 4a, ANOSIM Global R = 0.57, p < 0.001). However, there was no significant difference in archaeal 16S rRNA gene profiles among treatment cells in the HS2 wetland (Fig. 5a left, Global R = 0.005, p = 0.5). In the DN wetland, the pairwise test of these profiles had significant differences among cell groups of the former (DN1, DN2), middle (DN3-A, DN3-B, DN3-C) and middle-latter (DN3-D, DN4, DN5) ones (Fig. 5a right, Global R = 0.55, p < 0.05). Based on DGGE profile patterns, bacterial 16S rRNA gene composition varied between the constructed and the natural wetlands groups. From the most dominant band of these profiles, one sequence was close to the 16S rRNA gene of cyanobacteria, such as sSalicola marasensis and other Salicola sp. (98% sequence identity, Fig. 2; Maturrano et al., 2006; Sabet et al., 2009; Atanasova et al., 2012). Other sequences were similar to the chloroplast gene of pinnate diatom (Phaeodactylum tricornutum, 98%; Rampen et al., 2009), and sequences from the bacterial-algal mat from oilpolluted sediments and lakes (99–100%; Fig. 2; Chaudhary et al., 2009; Acosta-González et al., 2013). The DGGE profile patterns of bacterial 16S rRNA gene in the HS2 wetland differed significantly from those in the DN wetland and natural wetlands (Fig. 4b, Global R = 0.61, p < 0.001). These profile patterns also differed among treatment cells in both HS2 and DN. Based on pairwise tests, bacterial 16S rRNA gene profiles were significantly different between groups of the former (HS2-1, HS2-2) and latter (HS2-4, HS2-5) cells in the HS2 wetland (Fig. 5b left, Global R = 0.91, p < 0.05). Furthermore, based on pairwise tests, these profiles significantly differed among the former (DN1, DN2), middle (DN3-A, DN3-B, DN3-C) and middle-latter (DN3-D, DN4, DN5) cells in the DN wetland (Fig. 5b right, Global R = 0.80, p < 0.01). 3.3. Phylogenetic affiliation and spatial composition of the amoA genes in archaea and bacteria The amoA genes in archaeal (AOA) composition produced varied DGGE profile patterns for constructed versus natural wetlands. Sequences from the most major band of these profile patterns were close to the amoA gene of Nitrososphaera viennensis EN76 and Nitrososphaera sp. JG1 (94% similarity of protein sequence; Kim et al., 2012; Stieglmeier et al., 2014). Band sequences were also similar to samples from nitrogen-rich wetland, paddy soils, and river sediment (91–99%, Fig. 3 right; Liu et al., 2011; Wang et al., 2011; Baolan et al., 2012).

340 Table 1 Mean (±SEM) of water variables in the two natural wetlands of Xiu-Lang and Gu-Ting used as controls, and in each treatment cell of the Hsin-Hai II (HS2) and Daniaopi (DN) constructed wetlands. DO, dissolved oxygen; WT, water temperature; BDL, below detection limit. WT (◦ C)

Site\variable

Control Natural Wetlands Xiu-Lang Gu-Ting

Daniaopi Constructed Wetland (DN) Depositional pond Open-water pond Heavily vegetated pond I-A Heavily vegetated pond I-B Heavily vegetated pond I-C Heavily vegetated pond I-D Heavily vegetated pond II Ecological pond

DO (mg l−1 )

pH

Redox potential (mV) Water

Sediment

NH4 + N (mg l−1 )

NOx − N (mg l−1 )

PO4 −3 –P (mg l−1 )

AOA (copies g−1 )

AOB (copies g−1 )

AOA/AOB Ratio

0.78 15.1

XL GT

24.2 ± 0.2 25.7 ± 0.5

0.14 ± 0.02 0.16 ± 0.01

7.6 ± 0.5 6.7 ± 0.8

7.6 ± 0.3 7.3 ± 0.4

35.8 ± 0.5 39.1 ± 0.4

−168.5 ± 2.1 153.8 ± 3.1

0.12 ± 0.08 0.51 ± 0.23

4.52 ± 0.04 6.02 ± 0.87

0.73 ± 0.73 0

6.4 × 103 ± 6.8 × 102 1.6 × 105 ± 4.9 × 103

8.2 × 103 ± 8.5 × 102 1.0 × 104 ± 4.4 × 102

HS2-1

30.0 ± 1.8

0.47 ± 0.04

2.0 ± 0.7

7.1 ± 0.2

−15.8 ± 1.0

−180.9 ± 4.9

13.65 ± 0.31

0.32 ± 0.03

1.29 ± 0.06

8.2 × 102 ± 1.7 × 102

BDL

HS2-2

28.6 ± 1.4

0.38 ± 0.03

1.3 ± 0.8

7.1 ± 0.01

−185.2 ± 0.8

−123.3 ± 1.2

13.11 ± 0.05

0.38 ± 0.06

4.54 ± 0.08

6.1 × 102 ± 5.4 × 101

BDL

HS2-3

30.0 ± 0.7

0.47 ± 0.07

1.5 ± 0.7

7.2 ± 0.3

−129.8 ± 1.9

−113.9 ± 1.0

15.02 ± 0.11

0.26 ± 0.09

4.21 ± 0.02

1.4 × 103 ± 6.1 × 102

2.9 × 104 ± 1.6 × 104

HS2-4

29.0 ± 1.7

0.37 ± 0.07

0.6 ± 0.9

7.2 ± 0.1

−168.5 ± 0.6

−172.2 ± 2.7

13.56 ± 0.07

0.31 ± 0.04

4.64 ± 0.01

2

2.3 × 10 ± 8.9 × 10

BDL

HS2-5

29.7 ± 1.0

0.36 ± 0.04

2.6 ± 1.5

7.1 ± 0.2

−8.9 ± 0.3

−172.4 ± 3.1

10.44 ± 2.07

0.47 ± 0.02

4.20 ± 0.06

5.1 × 103 ± 8.0 × 102

BDL

DN1

30.0 ± 0.7

0.43 ± 0.16

1.9 ± 0.4

7.3 ± 0.2

91.5 ± 0.6

152.5 ± 1.9

1.84 ± 0.11

10.42 ± 0.05

0.45 ± 0.02

1.2 × 103 ± 2.7 × 102

BDL

DN2

29.5 ± 0.9

0.41 ± 0.07

1.8 ± 1.0

7.4 ± 0.2

−36.6 ± 0.6

−117.0 ± 2.1

4.38 ± 0.16

2.86 ± 0.58

1.02 ± 0.04

BDL

BDL

DN3-A

30.3 ± 0.1

0.33 ± 0.01

4.1 ± 1.1

7.2 ± 0.3

75.4 ± 0.3

−128.0 ± 2.5

4.42 ± 0.73

2.45 ± 0.00

DN3-B

29.6 ± 0.4

0.39 ± 0.07

8.6 ± 0.5

7.6 ± 0.3

41.3 ± 0.2

−171.2 ± 3.9

4.71 ± 0.07

DN3-C

28.8 ± 2.2

0.32 ± 0.10

1.9 ± 0.6

7.1 ± 0.2

40.6 ± 0.3

126.7 ± 3.4

DN3-D

28.3 ± 0.4

0.27 ± 0.07

3.0 ± 1.1

7.1 ± 0.5

108.9 ± 0.5

DN4

28.4 ± 1.4

0.36 ± 0.07

4.1 ± 1.1

7.2 ± 0.3

DN5

31.2 ± 1.3

0.30 ± 0.01

10.3 ± 2.1

8.8 ± 1.2

3

0.05

1.17 ± 0.00

2

7.5 × 10 ± 6.2 × 10

7.5 × 103 ± 2.9 × 103

0.10

1.29 ± 0.15

0.88 ± 0.03

5.6 × 102 ± 1.3 × 102

1.4 × 102 ± 3.1 × 101

3.84

0.50 ± 0.03

0.92 ± 0.04

0

4.4 × 102 ± 3.2 × 102

2.6 × 101 ± 7.6 × 100

−15.8 ± 4.4

0.27 ± 0.15

0.69 ± 0.08

0.03 ± 0.03

1.0 × 104 ± 8.2 × 102

2.1 × 103 ± 5.0 × 102

26.9 ± 0.4

−132.4 ± 2.5

1.06 ± 0.77

6.94 ± 2.18

0

BDL

BDL

65.0 ± 0.4

−101.2 ± 3.4

0.23 ± 0.05

1.78 ± 0.43

0.13 ± 0.13

1.1 × 103 ± 7.2 × 102

BDL

2

16.8

4.84

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

Hsin-Hai II Constructed Wetland (HS2) Depositional pond Heavily vegetated pond I Open-water pond Heavily vegetated pond II Ecological pond

conductivity (ms cm−1 )

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

341

Fig. 2. Phylogenetic tree of 16S rRNA genes for archaea and bacteria, based on DGGE bands using the neighbor-joining method (NJ). Bootstrap values (>50%) of NJ and maximum parsimony and consensus values (>50%) of maximum likelihood method are shown in the aforementioned order on the left of relevant nodes.

Profile patterns of the archaeal amoA gene differed among the two CWs and the natural wetlands group (Fig. 4c, Global R = 0.38, p < 0.05). These profile patterns also differed among treatment cells in HS2 and DN. Based on pairwise tests, there was a significant difference of these profiles between the former (HS2-1, HS2-2) and latter (HS2-4, HS2-5) cell groups in the HS2 wetland (Fig. 5c left, Global R = 0.59, p < 0.05). In the DN wetland, the archaeal amoA gene profiles in the former-middle cells (DN1, DN2, DN3-A, DN3-B, DN3C, DN3-4) differed from those in the latter cells (i.e., DN4, DN5; Fig. 5c right, Global R = 0.41, p < 0.05). Based on DGGE profile patterns, amoA genes in bacterial (AOB) composition also varied between constructed and natural wetlands. Sequences from the most dominant band of these profiles were close to the amoA gene of uncultured Nitrosomonas sp. clones (87–99%; Zeng et al., 2011). These sequences were also similar to samples from animal manure compost, leachate from landfills, membrane bioreactors, and active sludge (87–100%; Fig. 3 right; Purkhold et al., 2000; Zhu et al., 2007; Figuerola and Erijman, 2010; Ozdemir et al., 2011; Yapsakli et al., 2011; Zhang et al., 2011a; Yamamoto et al., 2012).

Profile patterns of the bacterial amoA gene differed significantly between the HS2 and DN wetlands and the natural wetlands group (Fig. 4d, Global R = 66, p < 0.01). However, there was no significant difference for these profile patterns in treatment cells in the HS2 wetland (Fig. 5d left, Global R = 0.29, p = 0.14). In the DN wetland, based on the pairwise test, bacterial amoA gene profiles in the former-middle ((DN1, DN2, DN3-A, DN3-B, DN3-C) cells were significantly different from those in the middle-later (DN3-D, DN4, DN5) ones (Fig. 5d right, Global R = 0.77, p < 0.01). 3.4. Spatial distribution of copy numbers of amoA in archaea and bacteria The qPCR of amoA gene as a proxy for AOA and AOB had spatial variation among wetlands (Table 1). The AOA abundances ranged from below detection level to 1.6 × 105 ± 4.9 × 103 gene copies g−1 soil. They were widely distributed in each wetland, except for cells DN2 and DN4 of the DN wetland. The AOA in both CWs had lower copy numbers than that in the natural wetlands. The copies in the DN were lower than those

342

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

Fig. 3. Phylogenetic tree of amoA genes for AOA and AOB, based on DGGE bands using the neighbor-joining method (NJ). Bootstrap values (>50%) of NJ and maximum parsimony and consensus values (>50%) of maximum likelihood method are shown in the aforementioned order on the left of relevant nodes.

Fig. 4. Non-metric multi-dimensional scaling ordination (NMDS) plot showing similarity in DGGE banding patterns for archaea (a), bacteria (b), AOA (c) and AOB (d) communities among the two CWs (HS2 and DN) and the natural wetlands (NW). Sites enclosed within the same red outline are not significantly different from one another. The result of the ANISIM test is shown in each graph corner. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

343

Fig. 5. Non-metric multi-dimensional scaling ordination (NMDS) plot showing similarity among treatment cells of the HS2 (left) and DN (right) wetlands in DGGE banding patterns of archaea (a), bacteria (b), AOA (c) and AOB (d) communities. Cells enclosed within the same red outline are not significantly from one another (ANISIM test, shown in each graph corner). Purple dash arrows refer to the flow direction of treatment cell series. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

in the natural wetlands (Kruskal–Wallis test, p < 0.05). The AOA copy number also varied among treatment cells in both CWs. The AOA abundances in the HS2 wetland decreased from 8.2 × 102 ± 1.7 × 102 to 6.1 × 102 ± 5.4 × 101 copies g−1 soil in the former two cells (HS2-1 to HS2-2), and then increased to 5.1 × 103 ± 8.0 × 102 copies g−1 soil in the latest cell (HS2-5). In the

DN wetland, AOA abundance had a higher copy number at sub-cell DN3-D (1.0 × 104 ± 8.2 × 102 copies g−1 soil), but was below detection at cells DN2 and DN4. For AOB, copy numbers ranged from below detection level to 2.9 × 104 ± 1.6 × 104 copies g−1 soil. The AOB in both CWs had similar abundances compared to those of natural wetlands. Most of

344

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

Table 2 Relationships between prokaryotic community composition (AOA, AOB, archaea and bacteria), as expressed in denaturing gradient gel electrophoresis (DGGE) pattern profiles, and the best combination of environmental characteristics. Prokaryotic community composition

rs

The best combination of environmental characteristics

Archaea Bacteria AOA AOB

0.664 0.617 0.538 0.671

Conductivity, ORP, NH4 + , PO4 −3 Conductivity, NH4 + Conductivity, pH, NOx − Conductivity, NH4 +

Seven environmental characteristics: conductivity, dissolved oxygen (DO), pH, ammonium (NH4 + ), water redox potential (ORP), oxidized nitrogen (NOx − ), phosphate (PO4 −3 ). rs , Spearman rank correlation.

the AOB abundance was restricted, with high levels in the middle cells (HS2-3, and DN3-A, DN3-B, DN3-C and DN3-D), whereas levels were below detection in the former and last cells of the two CWs. Copy numbers in the middle cell of the HS2 wetland were higher than those in the middle cells of the DN wetland (Kruskal–Wallis test and Dunn’s test, p < 0.05). Moreover, ratios of AOA copy numbers to AOB copy numbers ranged from 0.05 to 16.8 (Table 1). 3.5. Relationship between microbial patterns and environmental characteristics Band patterns of DGGE profiles of the 16S rRNA genes had associations with the surrounding environment (Table 2). The DGGE profile patterns of archaeal 16S rRNA gene were mostly correlated with single ammonium, and a combination of conductivity, redox potential, ammonium and phosphate (matching correlation, rs = 0.66, respectively). Furthermore, DGGE patterns of bacterial 16S rRNA gene were highly correlated, with a combination of conductivity and ammonium (rs = 0.62). The DGGE profile patterns of amoA genes were also strongly related to the surrounding environment (Fig. 4). The DGGE profile patterns of archaeal amoA gene mostly correlated with a combination of conductivity, pH and oxidized nitrogen (rs = 0.54). In addition, profile patterns of bacterial amoA gene were affected by a combination of conductivity and ammonium (rs = 0.67). The AOA abundance had less correlation with environmental characteristics, whereas AOB abundance was positively correlated to ammonium and phosphate concentrations (r = 0.82, p < 0.05 and r = 0.88, p < 0.01), but negatively correlated with water redox potential (r = −0.87, p < 0.05). Moreover, the AOA/AOB ratio was positively correlated to sediment redox potential (r = 0.93, p < 0.01). 4. Discussion In this study, both abundance and composition of ammoniaoxidizer (AO) communities varied among natural wetlands and two CWs. Furthermore, there were significant differences in archaeal and bacterial compositions between HS2 and natural wetlands. Among the treatment-cell series of these CWs, AOA and AOB communities had distinct spatial zonation in composition. Compositions of archaea and bacteria also had different zonation among treatment-cells. These discrepancies were attributed to wastewater inflow, nutrient removal, wetland scale, treatment cell series in constructed wetland, and specific designs for treatment cells. In contrast to sharp distinctions among AOs communities, band patterns of DGGE profiles of the 16S rRNA gene (revealing archaeal and bacterial community composition) differed less between the DN and natural wetland (Figs. 3b and 4a). Microbial communities were typically no more similar than if they had been randomly assembled from a common-source community (Baptista et al., 2008). Similar environmental characteristics were also observed in the latter cells of the DN wetland and in natural wetlands (Table 1).

It was noted that biodiversity of wetlands constructed for wastewater treatment can be enhanced by proper design and management (Hsu et al., 2011). Based on the current study, with better design and management of constructed wetlands, ecological benefits should approach that of natural wetlands (Becerra-Jurado et al., 2012). 4.1. Why AO community composition differed in the two CWs Performance of constructed wetlands is variable (Hsu et al., 2011). In the present study, the AO community composition defined by the DGGE profile patterns of the amoA gene in the two CWs could be classified by these differences (Fig. 4c and d). In addition, AOA community composition changing from former cells to the latter ones was noted in HS2 and DN (Fig. 5c). Regardless, both constructed wetlands were effective at nutrient removal and effects of such removal led to an environmental gradient among treatment cells of the two CWs (Table 1; Hsu et al., 2011); this was probably the main cause of the spatial distribution of AOA community composition (Fig. 5c). That AOB community composition differed from the former and latter cells was also apparent in the DN wetland but not in the HS2 wetland (Fig. 5d); this indirectly reflected differences in removal efficiency of the two CWs (Table 1). Such results were attributed to the size of constructed wetlands and vegetation design of treatment cells. 4.2. Effect of CWs size and cells vegetation design on AOs composition Species richness of multicellular organisms might be correlated with wetland area, including the things living in it, such as macroinvertebrates, herptiles, birds, mammals and plants (Hsu et al., 2011). However, the relationship between microbial diversity and wetland area remains unknown. Distinct microbial compositions were first observed in constructed wetlands with different areas (Table 1, Fig. 5c and d). To maintain similar hydraulic time, wetlands enclosing a smaller treatment area always have a slower flow rate, lower treatment performance, and less environmental gradient (Hsu et al., 2011). Outflow ammonium concentration in the HS2 wetland was over 40-fold higher than that in the DN wetland (Table 1). Gradients of other environmental parameters, such as DO, sediment redox potential and oxidizing nitrogen concentration, were also less in the HS2 wetland than in the DN wetland (Table 1). This is one reason why AOB diversity was lower in the HS2 wetland, which contained a smaller treatment area, compared to that in the DN wetland (Table 1, Fig. 5d). A similar pattern was also observed in spatial distribution of archaeal compositions in these two wetlands (Table 1, Fig. 5c). In terms of vegetation design, both uptake of nutrient and supply of oxygen may influence microbial composition (Ruiz-Rueda et al., 2009; Skiba et al., 2011; Wang and Gu, 2013; Villalobos et al., 2013). As plants can transport oxygen into rhizospheres (RuizRueda et al., 2009; Wang and Gu, 2013), redox characteristics of the rhizospheres are usually different from those of the non-vegetated sediments. Since the extent of the process is determined by oxygen availability (Ruiz-Rueda et al., 2009), oxic rhizosphere may accelerate nitrification and benefit AOA and AOB in CWs. It has also been reported that nitrification in the soil was suppressed by the shortage of ammonium that was assimilated by plants (Skiba et al., 2011). Typha latifolia was usually applied for uptake of nutrient in CWs (Villalobos et al., 2013). There were much greater abundances of AOs in the rhizosphere of Typha angustifolia in freshwater CW of Hong Kong (Wang and Gu, 2013). Similar species, Typha orientalis, occupied a large proportion of vegetation coverage in both HS2 and DN (Hsu et al., 2011). Therefore relatively abundant and diverse AOs were expected (Table 1). This potential mechanism provides an explanation as to why AOs were very different in the

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

latter cells with vegetation in the DN wetlands from those in the non-vegetated former cells. 4.3. Abundances AOA and AOB in the two CWs Using a copy number of the amoA gene to infer numbers of AOA and AOB in this study should be interpreted with caution. In that regard, it is inevitable that some strains of ammonia-oxidizer contain two or more nearly identical copies of target genes. This situation would skew the ratio for AOA and AOB. Since it is hard to identify and weight all amoA genes under environmental complexity, this situation was assumed as an average error of AO abundance estimation in every sampling site of this study. Compared to sediment from other natural habitats such as coasts, bays, estuaries, freshwater wetlands, mangroves, and salt marshes, abundance of AOs in the two CWs in this study were 1to 6-orders and 2- to 4-orders lower, respectively (Bernhard et al., 2007; Moin et al., 2009; Li et al., 2011; Sims et al., 2012). Such low abundance of AOs was also reported in other constructed wetlands used for biological treatment of wastewater (Correa-Galeote et al., 2013; Huang et al., 2013). Previous studies suggested that low AO abundance reflected extremely sparse spatial distribution, possibly caused by a lack of ecological niche variables in constructed wetlands (Correa-Galeote et al., 2013). It was also reported that relative composition of AO abundances were affected by the type of main substrates in constructed wetlands (Huang et al., 2013). In the two CWs of this study, highly environmental gradients formed different ecological niches and two AO abundances had distinct spatial distributions. Most AOB was restricted to the middle cells of the constructed wetlands, reflecting a distinct ecological niche in this study (Table 1). The AOB abundance was comparatively higher in the HS2 than in the DN wetland, with ammonium and phosphate concentrations that were much higher (Table 1). Similar observations have been reported, suggesting that the high abundance of AOB was caused by the relatively high ammonium concentration in mesocosm experiments (Verhamme et al., 2011). We also confirmed that the affiliated genus Nitrosomonas of the dominant AOB phylotype can be engineered in high-ammonia environments (Zeng et al., 2011). Comparatively speaking, AOA was widely distributed in each of the study wetlands. Their affiliated strains also had an ability to adapt to a wide variety of terrestrial and aquatic environments (Liu et al., 2011; Wang et al., 2011; Baolan et al., 2012). The AOA seemed to possess more physiological versatility in evolution and adaptation to particular environmental sets within sediment, compared to AOB (Prosser and Nicol, 2012). 4.4. Effects of environmental and biological characteristics on AOs communities AOA and AOB communities were correlated to distinct parameter combinations (conductivity, pH, NOx − , rs = 0.54 for AOA; conductivity, NH4 + , rs = 0.67 for AOB, Table 2). These results confirmed the diverse availability of nitrogen nutrients for the two types of AO communities (Webster et al., 2002). Compared to AOA, that the AOB community were highly correlated to ammonium, which reflected their dominance of ammonia oxidation in neutral environments with high nitrogen input (Verhamme et al., 2011). It is well documented that AOA are less sensitive than AOB to environmental variations, e.g. pH, oxygen and ammonium (Webster et al., 2002; Schmidt et al., 2007; Alam et al., 2013). In contrast, AOA community composition was affected by pH (Table 2), consistent with the suggestion that AOA existence and abundance over AOB occurs in low-pH and/or low-phosphate-containing environments (Prosser and Nicol, 2012).

345

Moreover, in this study, both AOs communities were correlated to electrical conductivity. Previous studies have discussed diverse influences of the environmental characteristics described above on AOs communities (Schmidt et al., 2007; Verhamme et al., 2011; Prosser and Nicol, 2012; Alam et al., 2013). In this study, electrical conductivity was highly correlated with many environmental characteristics, including ammonium, oxidized nitrogen, redox potential and phosphate (r = 0.87, −0.62, −0.65, and 0.73, respectively). Electric conductivity may be regarded as a cumulative effect of several environmental characteristics. Furthermore, electrical conductivity physically influences cell osmosis (Kadlec and Wallace, 2009). Apart from environmental characteristics, interspecies interactions may be another factor affecting microbial abundances (Truu et al., 2005). In this study, archaeal and AOA compositions reflected distinct environmental combinations (Table 2). Dominant phylotypes for the CW archaea were affiliated with anaerobic chemoautotroph methanogens (Fig. 2). In contrast to assimilation of ammonium for AO growth, methanogens usually use methane an energy source (Eggen et al., 1990; Barber et al., 2011; Zhang et al., 2011b). Due to unique source availability for each chemoautotrophic process, both AOA and methanogen communities were likely to co-exist spatially in the CW sediment. Compared to the archaeal and AOA communities, the environmental combinations related to bacterial and AOB composition were similar (Table 2). Affiliated bacterial phylotypes also reflected potential competition for nutrient assimilation (Fig. 2), which restricted development of a more abundant AOB community and its composition (Table 1). In conclusion, differing performance in terms of nutrient removal in the two constructed wetlands determined the existence of AO communities. Both AO communities also reflected divergent adaption to such distinct environmental gradients. Compared to the AOB community, AOA exhibited a complex community composition and was widely distributed, had high environmental affiliation and unique environmental demands. Therefore, we speculated that AOA may be more critical for nutrient depletion in constructed wetlands. References Acosta-González, A., Rosselló-Móra, R., Marqués, S., 2013. Characterization of the anaerobic microbial community in oil-polluted subtidal sediments aromatic biodegradation potential after the Prestige oil spill. Environ. Microbiol. 15, 77–92. Alam, M.S., Ren, G.D., Lu, L., Zheng, Y., Peng, X.H., Jia, Z.J., 2013. Conversion of upland to paddy field specifically alters the community structure of archaeal ammonia oxidizers in an acid soil. Biogeosciences 10, 5739–5753. Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Basic local alignment search tool. J. Mol. Biol. 215, 403–410. Atanasova, N.S., Roine, E., Oren, A., Bamford, D.H., Oksanen, H.M., 2012. Global network of specific virus–host interactions in hypersaline environments. Environ. Microbiol. 14, 426–440. Baolan, H., Shuai, L., Lidong, S., Ping, Z., Xiangyang, X., Liping, L., 2012. Effect of different ammonia concentrations on community succession of ammonia-oxidizing microorganisms in a simulated paddy soil column. PLoS ONE 7, 1–7. Baptista, J.C., Davenport, R.J., Donnelly, T., Curtis, T.P., 2008. The microbial diversity of laboratory-scale wetlands appears to be randomly assembled. Water Res. 42, 3182–3190. Barber, R.D., Zhang, L., Harnack, M., Olson, M.V., Kaul, R., Ingram-Smith, C., Smith, K.S., 2011. Complete genome sequence of Methanosaeta concilii a specialist in aceticlastic methanogenesis. J. Bacteriol. 193, 3668–3669. Becerra-Jurado, G., Harrington, R., Kelly-Quinn, M., 2012. A review of the potential of surface flow constructed wetlands to enhance macroinvertebrate diversity in agricultural landscapes with particular reference to Integrated Constructed Wetlands (ICWs). Hydrobiologia 692, 121–130. Bernhard, A.E., Tucker, J., Giblin, A.E., Stahl, D.A., 2007. Functionally distinct communities of ammonia-oxidizing bacteria along an estuarine salinity gradient. Environ. Microbiol. 9, 1439–1447. Chan, O.C., Wolf, M., Hepperle, D., Casper, P., 2002. Methanogenic archaeal community in the sediment of an artificially partitioned acidic bog lake. FEMS Microbiol. Ecol. 42, 119–129. Chaudhary, A., Haack, S.K., Duris, J.W., Marsh, T.L., 2009. Bacterial and archaeal phylogenetic diversity of a cold sulfur-rich spring on the shoreline of Lake Erie, Michigan. Appl. Environ. Microbiol. 75, 5025–5036.

346

L.-F. Fan et al. / Ecological Engineering 90 (2016) 336–346

Clarke, K.R., Ainsworth, M., 1993. A method of linking multivariate community structure to environmental variables. Mar. Ecol. Prog. Ser. 92, 205–219. Clarke, K.R., Gorley, R.N., 2001. Primer V5: User Manual/Tutorial. Plymouth Marine Laboratory, Plymouth, UK. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association, Washington, DC. Correa-Galeote, D., Marco, D.E., Tortosa, G., Bru, D., Philippot, L., Bedmar, E.J., 2013. Spatial distribution of N-cycling microbial communities showed complex patterns in constructed wetland sediments. FEMS Microbiol. Ecol. 83, 340–351. DeLong, E.F., 1992. Archaea in coastal marine environments. Proc. Natl. Acad. Sci. U.S.A. 89, 5685–5689. Economopoulou, M.A., Tshrintzis, V.A., 2004. Design methodology of free water surface constructed wetlands. Water Resour. Manage. 18, 541–565. Eggen, R., Harmsen, H., de Vos, W.M., 1990. Organization of a ribosomal RNA gene cluster from the Archaebacterium Methanothrix soehngenii. Nucleic Acids Res. 18, 1306. Fan, L.F., Shieh, W.Y., Wu, W.F., Chen, C.-P., 2006. Distribution of nitrogenous nutrients and denitrifier strains in estuarine sediment profiles of the Tanshui River, northern Taiwan. Estuar. Coast. Shelf Sci. 69, 543–553. Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F., Burr, M.D., Brisson, J., Camper, A.K., Stein, O.R., 2009. Microbial processes influencing performance of treatment wetlands: a review. Ecol. Eng. 35, 987–1004. Felsenstein, J., 1981. Evolutionary trees from DNA sequences: a maximum likelihood approach. J. Mol. Evol. 17, 368–376. Figuerola, E.L.M., Erijman, L., 2010. Diversity of nitrifying bacteria in a full-scale petroleum refinery wastewater treatment plant experiencing unstable nitrification. J. Hazard Mater. 181, 281–288. Fitch, W.M., 1971. Toward defining the course of evolution: minimum change for specified tree topology. Syst. Zool. 20, 406–416. 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. Proc. Natl. Acad. Sci. U.S.A. 102, 14683–14688. Hollander, M., Wolfe, D.A., 1999. Nonparametric Statistical Methods, second ed. John Wiley, New York, NY. 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. Hsueh, M.L., Yang, L., Hsieh, L.Y., Lin, H.J., 2014. Nitrogen removal along the treatment cells of a free-water surface constructed wetland in subtropical Taiwan. Ecol. Eng. 73, 579–587. Huang, X., Liu, C., Wang, Z., Gao, C., Zhu, G., Liu, L., 2013. The effects of different substrates on ammonium removal in constructed wetlands: a comparison of their physicochemical characteristics and ammonium-oxidizing prokaryotic communities. Clean Soil Air Water 41, 283–290. Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, second ed. CRC Press, Boca Raton, FL. Kim, J.G., Jung, M.Y., Park, S.J., Rijpstra, W.I., Sinninghe, D.J.S., Madsen, E.L., Min, D., Kim, J.S., Kim, G.J., Rhee, S.K., 2012. Cultivation of a highly enriched ammoniaoxidizing archaeon of thaumarchaeotal group I.1b from an agricultural soil. Environ. Microbiol. 14, 1528–1543. Li, M., Cao, H., Hong, Y., Gu, J.-D., 2011. Spatial distribution and abundances of ammonia-oxidizing archaea (AOA) and ammonia-oxidizing bacteria (AOB) in mangrove sediments. Appl. Microbiol. Biotechnol. 89, 1243–1254. Liu, Z., Huang, S., Sun, G., Xu, Z., Xu, M., 2011. Diversity and abundance of ammoniaoxidizing archaea in the Dongjiang River, China. Microbiol. Res. 166, 337–345. Maturrano, L., Valens-Vadell, M., Rosselló-Mora, R., Antón, J., 2006. Salicola marasensis gen. nov., sp. nov., an extremely halophilic bacterium isolated from the Maras solar salterns in Peru. Int. J. Syst. Evol. Microbiol. 56, 1685–1691. Moin, N.S., Nelson, K.A., Bush, A., Bernhard, A.E., 2009. Distribution and diversity of archaeal and bacterial ammonia oxidizers in salt marsh sediments. Appl. Environ. Microbiol. 75, 7461–7468. 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 encoding for 16S rRNA. Appl. Environ. Microbiol. 59, 695–700. Muyzer, G., Teske, A., Wirsen, C.O., Jannasch, H.W., 1995. Phylogenetic relationship of Thiomicrospira species and their identification in deep-sea hydrothermal vent samples by denaturing gradient gel electrophoresis of 16S rDNA fragments. Arch. Microbiol. 164, 165–172. Ozdemir, B., Mertoglu, B., Yapsakli, K., Aliyazicioglu, C., Saatci, A., Yenigun, O., 2011. Investigation of nitrogen converters in membrane bioreactor. J. Environ. Sci. Health, A: Tox. Hazard Subst. Environ. Eng. 46, 500–508. Prosser, J.I., Nicol, G.W., 2012. Archaeal and bacterial ammonia-oxidisers in soil: the quest for niche specialisation and differentiation. Trends Microbiol. 20, 523–531. Purkhold, U., Pommerening-Röser, A., Juretschko, S., Schmid, M.C., Koops, H.-P., Wagner, M., 2000. Phylogeny of all recognized species of ammonia oxidizers based on comparative 16S rRNA and amoA sequence analysis: implications for molecular diversity surveys. Appl. Environ. Microbiol. 66, 5368–5382.

Rampen, S.W., Schouten, S., Panoto, F.E., Brink, M., Andersen, R.A., Muyzer, G., Abbas, B., Damsté, J.S.S., 2009. Phylogenetic position of Attheya longicornis and Attheya septentrionalis (Bacillariophyta). J. Phycol. 45, 444–453. 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. Microbiol. 63, 4704–4712. ˜ Ruiz-Rueda, O., Hallin, S., Baneras, L., 2009. Structure and function of denitrifying and nitrifying bacterial communities in relation to the plant species in a constructed wetland. FEMS Microbiol. Ecol. 67, 308–319. Sabet, S., Diallo, L., Hays, L., Jung, W., Dillon, J.G., 2009. Characterization of halophiles isolated from solar salterns in Baja California, Mexico. Extremophiles 13, 643–656. SAS Institute, 2003. SAS User’s Guide: Statistics, Release 9.1. SAS Institute, Cary, NC. Saitou, N., Nei, M., 1987. The neighbour-joining method: a new method for reconstructing phylogenetic trees. Mo. Biol. Evol. 4, 406–425. Schmidt, C.S., Hultman, K.A., Robinson, D., Killham, K., Prosser, J.I., 2007. PCR profiling of ammonia-oxidizer communities in acidic soils subjected to nitrogen and sulphur deposition. FEMS Microbiol. Ecol. 61, 305–316. Schwarz, J.I.K., Lueders, T., Eckert, W., Conrad, R., 2007. Identification of acetateutilizing bacteria and archaea in methanogenic profundal sediments of Lake Kinneret (Israel) by stable isotope probing of rRNA. Environ. Microbiol. 9, 223–237. Sims, A., Horton, J., Gajaraj, S., McIntosh, S., Miles, R.J., Mueller, R., Reed, R., Hu, Z., 2012. Temporal and spatial distributions of ammonia-oxidizing archaea and bacteria and their ratio as an indicator of oligotrophic conditions in natural wetlands. Water Res. 46, 4121–4129. Skiba, M.W., George, T.S., Baggs, E.M., Daniell, T.J., 2011. Plant influence on nitrification. Biochem. Soc. Trans. 39, 275–278. Stieglmeier, M., Klingl, A., Alves, R.J., Rittmann, S.K., Melcher, M., Leisch, N., Schleper, C., 2014. Nitrososphaera viennensis gen. nov., sp. nov., an aerobic and mesophilic, ammonia-oxidizing archaeon from soil and a member of the archaeal phylum Thaumarchaeota. Int. J. Syst. Evol. Microbiol. 64, 2738–2752. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., Kumar, S., 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol., http://dx.doi. org/10.1093/molbev/msr121. Treusch, A.H., Leininger, S., Kletzin, A., Schuster, S.C., Klenk, H.P., Schleper, C., 2005. Novel genes for nitrite reductase and amo-related proteins indicate a role of uncultivated mesophilic crenarchaeota in nitrogen cycling. Environ. Microbiol. 7, 1985–1995. Truu, J., Nurk, K., Juhanson, J., Mander, Ü., 2005. Variation of microbiological parameters within planted soil filter for domestic wastewater treatment. J. Environ. Sci. Health 40, 1191–1200. Verhamme, D.T., Prosser, J.I., Nicol, G.W., 2011. Ammonia concentration determines differential growth of ammonia-oxidising archaea and bacteria in soil microcosms. ISME J. 5, 1067–1071. ˜ Villalobos, R.M., Zúniga, J., Salgado, E., Schiappacasse, M.C., Maggi, R.C., 2013. Constructed wetlands for domestic wastewater treatment in a Mediterranean climate region in Chile. Electron. J. Biotechnol., doi.2225/vol16-issue4-fulltext-5. Wang, Y.-F., Gu, J.-D., 2013. Higher diversity of ammonia/ammonium-oxidizing prokaryotes in constructed freshwater wetland than natural coastal marine wetland. Appl. Microbiol. Biotechnol. 97, 7015–7033. Wang, S., Wang, Y., Feng, X., Zhai, L., Zhu, G., 2011. Quantitative analyses of ammoniaoxidizing archaea and bacteria in the sediments of four nitrogen-rich wetlands in China. Appl. Microbiol. Biotechnol. 90, 779–787. Webster, G., Embley, T.M., Prosser, J.I., 2002. Grassland management regimes reduce small-scale heterogeneity and species diversity of ␤-proteobacterial ammonia oxidizer populations. Appl. Environ. Microbiol. 68, 20–30. Yamamoto, N., Oishi, R., Suyama, Y., Tada, C., Nakai, Y., 2012. Ammonia-oxidizing bacteria rather than ammonia-oxidizing archaea were widely distributed in animal manure composts from field-scale facilities. Microbes Environ. 27, 519–524. Yapsakli, K., Aliyazicioglu, C., Mertoglu, B., 2011. Identification and quantitative evaluation of nitrogen-converting organisms in a full-scale leachate treatment plant. J. Environ. Manage. 92, 714–723. Yavitt, J.B., Yashiro, E., Cadillo-Quiroz, H., Zinder, S.H., 2012. Methanogen diversity and community composition in peatlands of the central to northern Appalachian Mountain region, North America. Biogeochemical 109, 117–131. Zeng, T., Li, D., Zhang, J., 2011. Characterization of the microbial community in a partial nitrifying sequencing batch biofilm reactor. Curr. Microbiol. 63, 543–550. ˜ Zhang, Y., Canas, E.M.Z., Zhu, Z., Linville, J.L., Chen, S., He, Q., 2011a. Robustness of archaeal populations in anaerobic co-digestion of dairy and poultry wastes. Bioresour. Technol. 102, 779–785. Zhang, T., Ye, L., Tong, A.H.Y., Shao, M.-F., Lok, S., 2011b. Ammonia-oxidizing archaea and ammonia-oxidizing bacteria in six full-scale wastewater treatment bioreactors. Appl. Microbiol. Biotechnol. 91, 1215–1225. Zhu, S., Chan, G.Y.S., Cai, K.-L., Qu, L.-H., Huang, L.-N., 2007. Leachates from municipal solid waste disposal sites harbor similar, novel nitrogen-cycling bacterial communities. FEMS Microbiol. Lett. 267, 236–242.