Introduction

Ammonia can be oxidized to nitrite under aerobic condition by aerobic ammonia-oxidizing bacteria (AOB) (Winogradsky 1890). Until recently, the discovery and isolations of ammonia-oxidizing archaea (AOA) capable to carry out chemoautotrophic ammonia oxidation, challenge the traditional view that ammonia oxidation is restricted to AOB alone (Könneke et al. 2005; Venter et al. 2004). The aerobic ammonia-oxidizing prokaryotes (including AOA and AOB) play an important role in nitrogen removal as they carry a unique amo gene encoding for the enzyme ammonia monooxygenase, including three small subunits, amoA, amoB and amoC (Norton et al. 2002).

The amoA genes in both AOA and AOB have been used as molecular biomarkers to investigate the corresponding community structures and distribution in diverse ecosystems, including soils (Pett-Ridge et al. 2013; Taylor et al. 2012; Tourna et al. 2011; Wang et al. 2013a, 2013b, 2014b; Zhang et al. 2012; Zhou et al. 2014), oceans (Beman et al. 2008; Hollibaugh et al. 2002; Pitcher et al. 2011; Santoro and Casciotti 2011; Walker et al. 2010; Wuchter et al. 2006), coastal wetlands (Cao et al. 2012; Wang et al. 2013c; Wang and Gu 2013, 2014; Wang et al. 2014a), estuaries (Beman and Francis 2006; Dang et al. 2008), lakes (Jiang et al. 2009), freshwaters (Herrmann et al. 2008, 2009), salts marsh (Dorador et al. 2008; Moin et al. 2009). However, the role of AOA in nitrogen cycle of the freshwater habitat such as freshwater marsh remains unclear (Laanbroek and Speksnijder 2008). Most of the studies showed that AOA amoA gene is predominant over that of AOB in terms of both the richness and abundance in various ecosystems (Adair and Schwartz 2008; Dang et al. 2008; Herrmann et al. 2009; Prosser and Nicol 2008), but some reports showed contradictory results in environments such as activated wastewater treatment bioreactor (Wells et al. 2009), lake sediment (Jiang et al. 2009) and estuary sediment (Mosier and Francis 2008). AOB isolates are divided into two monophyletic groups: beta- and gamma-proteobacteria (Kowalchuk and Stephen 2001). The beta-AOB communities are simply further divided into two genera: Nitrosomonas and Nitrosospira (Purkhold et al. 2000; 2003).

Wetlands are a complex ecosystem consisting of marshes, swamps and bogs and these different environmental types offer a favorable condition for the development of common wetland plants such as Phragmites australis and P. communis and create different niches due to vegetation types, oxygen bioavailability, water coverage, and nutrient variation (Ruiz-Rueda et al. 2009; You et al. 2009). It is suggested that both nitrification and denitrification could be achieved in wetlands but at low rates (Clément et al. 2002; Verhoeven and Meuleman 1999). However, to our knowledge, there is no information about the niche specificity of AOA communities in freshwater marsh although the potential niche differentiation may drive AOB communities (Laanbroek and Speksnijder 2008). It is hypothesized that the wetland vegetation of common reed and the environment physicochemical parameters contribute to the species or ecotypes differentiation of AOA and AOB communities in a freshwater wetland.

Longfeng Freshwater Wetland Nature Reserve (E125°07′–125°15′; N46°28′–46°32′) is one of the main local provincial level marshes in Daqing, Heilongjiang Province, China (Xu and Zhang 2009). This wetland includes 314 species of protected wild plants, 207 species of wild land animals, 45 species of wild provincial capital level protected fish, 24 species of wild state level protected birds, 98 species of wild provincial capital level birds and 19 Heilongjiang province protected animal species in an area of 6,211 hm2 (Hou 2009). However, human disturbance to the wetlands such as Daqing oil field development, urban construction, agriculture activities and other factory discharge may threaten the wetland. Because of that, this wetland could serve as a prime study site for examination of nitrogen removal by ammonia-oxidizing microorganisms. The objectives of this study were to investigate the potential driving factors on the community characteristics of AOA and AOB in the freshwater wetland close to the Daqing city.

Materials and methods

Sampling site description and collecting method

Surface and bottom soil/sediment samples were collected from different sites in the Daqing Longfeng Freshwater Wetland Natural Reserve, Heilongjiang, China in summer of 2008 (Fig. 1). The sampling sites were classified as four types of niches: (1) Samples 23 and 24 sediments were collected from a wetland covered with reeds and the surface sediment was characterized of rich nitrate. This site was labeled as R; (2) Sample 25 and 26 sediments were collected from a wetland near the niche R, covered with standing water but without any plantations. This site served as the reed control and was labeled as R-Con; (3) Sample 2S and 2B sediments were collected from a marsh characterized of relatively high concentration of ammonia, located near a main highway and further apart from niches R and R-Con. This site was labeled RW; (4) Sample 3S and 3B soils were collected from a dry land nearby the niche RW with mixed vegetation of mainly common poplar trees and labeled as D. The soils/sediments between 0.5 cm and 10 cm were classified as surface while that below 10 cm were classified as bottom. All samples were transported to the laboratory on the dry-ice in a cooling box.

Fig. 1
figure 1

Map of the sampling sites in the Longfeng Freshwater Natural Reserve, Northeast China

Full size image

Analysis of chemical characteristics

Concentrations of ammonium-N, nitrate–N, and nitrite-N were measured by extracting sediments in 2.0 M KCl (Zuber et al. 2005) and the extracts were analyzed by an auto-analyzer (QuikChem, Milwaukee, WI) using standard flow injection analysis (FIA) techniques according to the standard extraction procedures. Replicate chemical tests were performed for each sample.

DNA extraction and PCR amplification

DNA was extracted from 100 to 200 mg of soil/sediment using the SoilMaster™ DNA Extraction Kit according to the manufacturer’s protocol (Epicentre Biotechnologies, Madison, WI). DNA extracts were stored at −20 °C before using as a template for PCR reaction.

The archaeal amoA genes were amplified using the PCR primers Arch-amoAF (5′–STAATGGTCTGGCTTAGACG–3′) and Arch-amoAR (5′–GCGGCCATCCATCTGTATGT–3′). Based on the standard procedures described in the manufacturer’s instructions and results of previous studies (Francis et al. 2005), the optimized PCR reaction mixture in a 25 μl volume contained 5 μl of 5× colorless GoTaq®Flexi Buffer, 12.5 mM MgCl2, 5 mM of each deoxyribonucleotide triphosphate, 0.5 μM of each primer, 1.25U GoTaq® DNA Polymerase and 3 μl of purified DNA as template (10–100 ng). PCR conditions were set as follows: 94 °C for 3 min; 40 cycles of 94 °C for 30 s, 53 °C for 1 min, and 72 °C for 1 min; and finally 72 °C for 15 min.

The bacterial amoA genes were amplified using the PCR primers amoA-1F (5′–GGGGGTTTCTACTGGTGGT–3′) and amoA-2R (5′–CCCCTCKGSAAAGCCTTCTTC–3′). The optimized PCR reaction mixture in a 25 μl volume contained 5 μl of 5× colorless GoTaq®Flexi Buffer, 12.5 mM MgCl2, 5 mM of each deoxyribonucleotide triphosphate, 0.5 μM of each primer, 1.25U GoTaq® DNA Polymerase and 3 μl of purified DNA as template (10–100 ng). PCR conditions were set as follows: 94 °C for 3 min; 37 cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s; and finally 72 °C for 10 min.

Cloning, sequencing and phylogenetic analysis

Purified amoA DNA fragments were ligated into a pMD18-T vector (TaKaRa, Japan) and transformed into competent Escherichia coli DH5α cells. The library clones were screened directly by PCR for the presence of inserts using two M13 universal primers: M13F (5′-GTTTCCCAGTCACGAC-3′) and M13R (5′-TCACA CAGGAAACAGCTATGAC-3′). The positive clones from each library were randomly selected and the PCR products were purified using a PCR Purification Kit (Qiagen, USA). The PCR products were sequenced (Tech Dragon Ltd, Hong Kong). DNA sequences were examined and edited using BioEdit (Tom Hall, North Carolina State University, NC) and MEGA, version 5.0 (Tamura et al. 2011). NCBI BLAST (http://www.ncbi.nih.gov) was used to find the closest related amoA gene sequences in the GenBank. The multiple alignments of amoA gene sequences were considered and neighbor-joining phylogenetic trees with bootstrap values based on 1,000 replications (Felsenstein 1985) were produced using MEGA program, version 5.0.

Data statistical analysis

The DOTUR (Distance-Based OTU and Richness) program was employed to compare diversity for amoA sequences from each sample (Schloss and Handelsman 2005). Operational taxonomic units (OTU) for community analysis were defined by a 5 and 1 % sequence variation in detecting AOA and AOB respectively, as determined using the furthest neighbor algorithm in DOTUR. DOTUR was also used to generate diversity analyses such as coverage, Chao1, Simpson index (D) and Shannon index (H) numbers for each site. The H index and the D index were calculated as the diversity indices by the program (Beman and Francis 2006; Schloss and Handelsman 2005). Another recently developed and more advanced community analysis program MOTHUR was employed to conduct the UniFrac Environment Clustering Analysis (Schloss et al. 2009). The estimated coverage of the constructed archaeal amoA gene and bacterial amoA gene libraries were calculated using the formula of C = [1−(n1/N)] × 100, where n1 represents the number of OTUs detected in one clone library and N strands for the total number of clones in that particular library. This coverage estimates the probability that all the unique sequences present in a given sample were represented at least once in the library (Dang et al. 2008; Mullins et al. 1995). Both AOA and AOB community differences between the niches were determined using non-weighted principal co-ordinates analysis (PCoA) and Jackknife Environment Clustering Analysis (Lozupone et al. 2007). Both the PCoA and Jackknife Environment Clustering Analysis were generated using online UniFrac tool program (http://bmf2.colorado.edu/unifrac/index.psp) according to the sequences molecular evolutionary distances and microbial community similarity based on the environmental occurrences. Correlation analysis between the diversity of archaeal and bacterial amoA genes and environmental variables were conducted using Microsoft Excel program.

Nucleotide sequence accession numbers

All the partial archaeal and bacterial amoA gene sequences of ammonia-oxidizing bacteria and archaea determined in this study were deposited in GenBank with accession numbers HM537746 to HM537915 and HM537916 to HM538185, respectively.

Results and discussion

Physicochemical characteristics of sediments/soils

Sediment/soil cores were obtained from four selected locations and the concentrations of N-associated parameters were measured (Table 1). Samples 2S and 2B, classified as ammonia rich niches (RW), had highest concentration of ammonia. In contrast, samples 25 and 26, defined as non-vegetated ones (R-Con), had the lowest except the sample 24. In general, surface samples had higher concentrations of ammonium and nitrate than bottom ones except in the RW niche. Nitrite concentrations in the soils were low in non-vegetated niche, but slightly higher in vegetated niches. Nitrate concentrations in all samples were generally low and even below detection limit (0.1 µg g−1 DW) in the RW niches and the bottom soil of the dry-land samples.

Table 1 Sampling sites and physicochemical parameters of the sediments/soils
Full size table

AOA community composition and phylogenetic affiliation

Seven archaeal amoA gene libraries among four locations were constructed except the bottom sample in the eutrophic reed site with water covering (RW), which failed to amplify amoA gene successfully. Possible reasons might be that AOA were either absent or under the PCR amplification detection limit. Among these seven clone libraries, 270 clones were obtained and 26 OTUs were identified. A total of 6–14 OTUs were observed in each niche type using a 5 % cutoff at the nucleotide level (Table 2). The high OTUs value indicated that the four niches harbored diverse AOA. The estimated coverage of all libraries was quite high (Table 2), indicating the clone libraries represented majority of archaeal amoA gene sequences present in the samples, and this information is in agreement with rarefaction analyses (Fig. S1a). The dry-land site (D) had the highest Shannon index values while non-vegetated reed site (R-Con) had the lowest. These results may raise a possibility that the terrestrial soil supports a higher AOA diversity than the freshwater covered niches, where terrestrial soil has a higher oxygen concentration and hence facilitate the ammonia oxidation (Christensen et al. 1990). The presence of plants may release oxygen in the rhizosphere through the roots (Herrmann et al. 2008; Laanbroek and Speksnijder 2008) and provide aerobic zones sustaining the growth of AOA communities. Hence common weed plant species in vegetated niches except the R-Con might favor the development of various AOA species.

Table 2 Diversity of archaeal and bacterial amoA sequences
Full size table

The phylogenetic tree of AOA (Fig. 2) showed that all the archaeal amoA gene sequences were grouped into two main clusters: Water column/sediment clade (Cluster A1) and Soil/sediment clade (Cluster B) defined earlier (Francis et al. 2005). In the phylogenetic tree, Cluster A2 consisted of several amoA gene sequences from estuarine samples and soils. Only a small proportion (about 2 %) of sequences fell into Cluster A1 while the rest majority fell into Cluster B. It is reasonable that the majority of archaeal amoA gene sequences recovered from the freshwater wetland soils are associated with the Soil/sediment AOA sequences. In Cluster A1, all the sequences were retrieved from the niche R-Con except one clone from the niche common reed (R). The closest representative relative in Cluster A1 was Nitrosopumilus Maritimus, which was firstly found in substratum of a tropical marine tank and inhabited at oligotrophic environment (Könneke et al. 2005).

Fig. 2
figure 2

Phylogenetic tree of AOA constructed with amoA gene fragments (594 bp). The numbers in the brackets are the clone number of the related phylotype which S represents surface sediment and B represents bottom sediment. The numbers at the nodes are percentages that indicate the levels of bootstrap support based on 1,000 resample data sets (only values greater than 50 % are shown). Branch lengths correspond to sequence differences as indicated by the scale bar. The clustering reference was defined by Francis et al. (2005), Herrmann et al. (2008), and (2009), and modified with several sub-clusters in Cluster B in this study

Full size image

AOA community distribution in response to niches and physicochemical parameters

The Jackknife Environment Clusters Analysis (Fig. S2a) showed that the vegetated niches RW, D and R were grouping together, and the non-vegetated niche R-Con was distinct from vegetated ones. This supported the previous studies in that plantation could stimulate the rhizosphere-association nitrification, in turn cause the community shift of AOA (Herrmann et al. 2008; Laanbroek and Speksnijder 2008; Ottosen et al. 1999; Riis and Sand-Jensen 1998). PCoA analysis of AOA communities showed that individual niches did not group together (Fig. 3a), indicating that unique patterns of AOA community structures occurred in each niche independently. The results are consistent with the idea proposed by Herrmann et al. (2009). The phylogeny of AOA in freshwater marsh environment might be associated with different niche characteristics such as presence of water and plantation. Environmental factors might also contribute to the community diversity of AOA in the wetland. The results of Pearson moment correlation test further supported this suggestion (Table 3). Shannon index of AOA had significant positive correlation to the concentration of ammonia in soils (Pearson moment correlation: r = 0.74, P < 0.05, n = 6), suggesting that ammonia input could be a pivotal factor enhancing AOA diversity in freshwater wetland. Previous studies showed that a higher abundance and diversity of archaeal amoA genes were detected under higher total ammonium concentrations (Park et al. 2006; Treusch et al. 2005). In contrast, some studies found that low ammonium concentrations might restrict the growth of AOA in some low N-containing environments (Erguder et al. 2009; Könneke et al. 2005; Reigstad et al. 2008).

Fig. 3
figure 3

Ordination diagrams of the sedimentary a archaeal amoA and b bacterial amoA assemblages calculated with unweighted UniFrac PCoA analyses using deduced amoA protein sequences. See Table 1 for detailed explanation for niche codes

Full size image
Table 3 Pearson moment correlation statistical analysis of physicochemical parameters and AOA and AOB communities
Full size table

AOB community composition and phylogenetic affiliation

The phylogenetic analysis was based on bacterial amoA gene with representative sequences retrieved from the GenBank database (Fig. 4). Eight bacterial amoA libraries were constructed among four niches, a total of 170 clones were obtained and 13 OTUs were identified. The estimated coverage of all libraries was quite high (Table 3), which was in agreement with the result of rarefaction analyses (Fig. S1b). Similar to the Shannon index of AOA, the dry-land site had the highest Shannon index values of AOB than other niches and the non-vegetated reed site had the lowest.

Fig. 4
figure 4

Phylogenetic tree of AOB constructed with amoA gene fragments (450 bp). The numbers in the brackets are the clone number of the related phylotype which S represents surface sediment and B represents bottom sediment. The numbers at the nodes are percentages that indicate the levels of bootstrap support based on 1,000 resample data sets (only values greater than 50 % are shown). Branch lengths correspond to sequence differences as indicated by the scale bar. The clustering reference was defined by Avrahami et al. (2002), Purkhold et al. (2000), and Purkhold et al. (2003), and modified in this study

Full size image

Compared to AOA, a relatively lower diversity of AOB was found in different niches, for both surface and bottom sediment/soil. All the sequences were classified and grouped into two β-proteobacteria: Nitrosospira-like and Nitrosomonas-like sequences. Most of the sequences (about 62 %) belonged to the Nitrosospira-like cluster while the Nitrosomonas-like cluster contained the rest clone sequences (about 38 %). Among the Nitrosospira-like cluster, 99 % sequences were similar to Nitrosospira-like sequences retrieved from sea water column in Gulf of Mexico. As defined by the previous studies, a sequence identity of less than 80 % (partial amoA gene sequence) and 88 % (whole amoA gene sequence) respectively has been suggested as a novel AOB clusters/lineages (Kim et al. 2008; Purkhold et al. 2000). The 105 amoA sequences found in this study had 75 % DNA identity to the referencing sequences Nitrosospira multiformis (DQ228454) and Nitrosomonas marina Nm22 (AF272405). As a consequence, these sequences together with the amoA sequences from Gulf of Mexico are proposed to be a novel independent cluster. This proposed cluster was found from all tree methods independently and supported by high parsimony bootstrap analysis (100 %). According to the phylogenetic tree, this novel cluster consisted of amoA sequences from all niches, indicating that it might be widely distributed in the freshwater marsh. One sequence from eutrophic reed (RW) niche was related to the Nitrosospira briensis Nsp10 and C-57 (AY123821 and Z97858 respectively) and grouped into cluster 3b. This finding was consistent with the dominance of Nitrosospira cluster 3 found in the relatively high ammonium concentrations in soils (Avrahami et al. 2002; Bruns et al. 1999; Kowalchuk et al. 2000).

Around 70 % sequences in the Nitrosomonas-like clade were affiliated to the Nitrosomonas communis cluster and had 98 % identity similar to an uncultured ammonia-oxidizing bacterium, isolate S38 (AM851051) retrieved from the environmental clone in rice field soil affected by nitrogen fertilization (Wang et al. 2009). Among the published AOB isolated species, these amoA sequences in the N. communis cluster were the most similar to the Nitrosomonas nitrosa. Studies showed that N. nitrosa prefer to inhabit in eutrophic freshwater environment, while N. communis prefer to inhabit in soils (Koops and Pommerening-Röser 2001). Two sequences from the non-vegetated niche R-Con were related to a sequence found in the coastal aquaculture environment (EU156173). In the eutrophic reed niche RW, 14 clone sequences were related to the N. marina Nm22 (AF272405). All these 16 clones were clustered into the Nitrosomonas oligotropha/Nitrosomonas marina lineage. The isolates among this lineage were retrieved from either wastewater or freshwater habitats, and they had low ammonium affinity constants (Koops and Pommerening-Röser 2001; Pommerening-Röser et al. 1996; Purkhold et al. 2003; Stehr et al. 1995). Four sequences in dry-land (D) niche were related to the Nitrosomonas halophila (AF272398) and uncultured AOB clone 17-11 (EU116356), with 96 % identities. These four sequences fell into the Nitrosomonas europaea/Nitrosococcus mobilis lineage whose members preferred to inhabit in sewage disposal plants, eutrophic freshwater and brackish water with relatively high ammonium substrate (Koops and Pommerening-Röser 2001).

The AOB detected in the freshwater wetland soils were mainly Nitrosospira-like amoA sequences, which was consistent with the results of other wetland studies that Nitrosospira species dominated in the wetland system (Ibekwe et al. 2003; Laanbroek and Speksnijder 2008; Moin et al. 2009). The phylogenetic tree showed a high AOB community composition overlapping among different niches from the freshwater marsh (Fig. 4). In the present study, AOB communities in the freshwater wetland sediments consisted of five stable lineages: Nitrosospira cluster 3b, Nitrosomonas marina/Nitrosomonas oligotropha cluster, N. communis cluster, Nitrosomonas europaea/Nitrosococcus mobilius cluster, and a proposed novel Nitrosospira cluster. The discovery of the novel potential AOB cluster suggested that a more wide distribution of AOB communities is present in natural environments.

AOB community distribution in response to depths and niches

A number of studies found that Nitrosospira dominated over Nitrosomonas in terrestrial environments (Avrahami and Conrad 2003; Avrahami et al. 2002; Kowalchuk et al. 2000; Mintie et al. 2003; Wang et al. 2009), which were consistent with this study showing a majority of Nitrosospira-like amoA sequences occurred in this wetland. Nitrosomonas were often found in high nitrogen or contaminated environments and even predominated in wastewater treatment plants (Dionisi et al. 2002; Geets et al. 2006; Wagner et al. 1995; Wang et al. 2009). A portion of Nitrosomonas-like amoA gene sequences were also detected in this study might give evidence that domestic discharge contamination from the nearby city influenced the ammonia-oxidizing oxidizers community ecology.

UniFrac environmental clustering of the bacterial amoA gene clone libraries showed clear grouping of separate AOB communities in surface and bottom sediments/soils except the samples 2S and 3B (Fig. S3). The surface samples 23, 25 and 3S were grouped together while the bottom samples 24, 26 and 2B were grouped together. In this study, more Nitrosospira-like sequences inhabited the bottom soils whereas more Nitrosomonas-like sequences were detected in upper soils (data not shown). This finding suggested a depth separation of the AOB community in the freshwater wetland. The factor of O2 availability might partly govern the AOB community distribution in soil system (Briones et al. 2002; Nicolaisen et al. 2004). Previous reports showed that oxygen concentration across soil depth gradient showed an inverse proportion (Freney et al. 1992; Lüdemann et al. 2000; Wang et al. 2009). Nitrosospira had a relatively lower O2 affinity compared to Nitrosomonas (Schramm et al. 1998) so that the former strains probably preferred partially oxic surface soils and the latter had better adaption of relatively less oxic deeper soil layer. The finding of depth separation of AOB community was also supported by Wang et al. (2009).

For AOB distribution in response to niches, niches R and R-Con were clustered together; niches D and RW formed the second and third groups respectively (Fig. S2b). Relatively high sequence similarity between niches R and R-Con suggested that these two niches had low within-group variation of AOB. That is to say, the presence of plants did not affect the AOB community composition since the niche R was grown with reeds while the niche R-Con had no plants. This finding was in agreement with previous studies that plantation had little influence on the composition of AOB community (Herrmann et al. 2009; Laanbroek and Speksnijder 2008). In contrast, the results from Jackknife Environment Clusters Analysis (Fig. S2b) supported that ammonium might be an important factor on the AOB community structure since niche RW was separated from other three niches which grouped together. Because the niche RW had a significant higher ammonium concentration than the other three niches, it is proposed that the different community composition of AOB in RW niche from other three niches was due to the effect of ammonium.