Introduction

Microbial nitrification plays an important role in the biogeochemical nitrogen cycle, which is a key process for ammonia oxidization to nitrate via nitrite. Ammonia-oxidizing bacteria (AOB) was once considered as the only contributor responsible for ammonia oxidation in the environment [1] and there were two groups of AOB, β-proteobacteria such as the genera Nitrosomonas and Nitrosospira and γ-proteobacteria like genus Nitrosococcus [2]. A number of studies showed that β-AOB were predominant in both the wastewater treatment plants and natural environments among the AOB communities [3]. Recently, the discoveries of both aerobic ammonia-oxidizing archaea (AOA) and anaerobic ammonium oxidizing (anammox) bacteria have updated our knowledge about the microbial nitrogen cycle [4, 5].

Some archaea in the Crenarchaeota were discovered to be able to carry out chemoautotrophic ammonia oxidation [6] and the first autotrophic ammonia-oxidizing marine archaeon, Nitrosopumilus maritimus SCM1 clone was isolated [4]. Soon after, AOA were found to be widely distributed and detected in various habitats, including estuary, coast wetlands, marine, soils, lake sediments, hot springs, wastewater treatment plants, and forest soils [7–23]. According to the genome studies of both AOA and AOB, ammonia-oxidizing genes (amo) were found to encode the ammonia monooxygenase responsible for ammonia-oxidation in all strains of both AOA and AOB [4, 24]. As a result, the corresponding α subunit of amo gene (amoA) in AOA and AOB was used as a molecular biomarker for the detection of AOA and AOB communities in environmental samples.

The relative copy numbers and dominance of bacterial and archaeal amoA genes in specific environments are still debatable. Many reports found that the copy number of archaeal amoA genes were higher than that of bacterial amoA genes in different environments like fertilized soils, sandy loam, rhizosphere paddy soil, low-oxygen fresh and brackish estuary sediment and semiarid soils [7, 25–30]. However, others showed that the archaeal amoA gene copy numbers were lower than that of bacterial amoA in various environments including estuary sediment, soils under stands of red alder, lake sediments and activated sludge bioreactor [8, 14, 31–33]. As a result, more investigations are required to elucidate if AOA is really the dominant ammonia-oxidizing archaea and bacteria in soils.

Anammox bacteria were first discovered in 1995 from the wastewater treatment plants where ammonium and nitrite disappeared and nitrogen gas appeared which could not be accounted for through denitrification [34, 35]. The anammox reaction is a chemolithotrophic process in which ammonium is coupled with nitrite to yield nitrogen gas in the absence of molecular oxygen [36]. Subsequently, five genera of the anammox bacteria were reported including Candidatus Kuenenia [37], Ca. Scalindua, Ca. Brocadia [38], Ca. Anammoxoglobus [39], and Ca. Jettenia [40]. Anammox bacteria have been detected in marine, coastal and estuarine sediments [41, 42], mangroves [43, 44], permafrost soil [45], oxygen minimum zones and anoxic basins [46], sea-ice [47], freshwater lakes and rivers [48, 49], and agricultural soils [50].

Pearl River Delta Estuary of China is one of the most complex large-scale estuarine systems, which consists of three sub-estuary bays, the Lingdingyang, the Modaomen and the Huangmaohai, and a reticulated network system [51]. The delta is surrounded by three major tributaries: Xijiang, Beijiang and Dongjiang, and started to form since late Pleistocene with annual average river freshwater discharge of 3.26 × 1011 m3 yr−1 [52]. But the current delta network system was the result of middle Holocene and with the effect of world-wide transgression [53]. The average sedimentation rate of constructing the estuary land was at around 1.8 mm annually and hence the formation of the estuary required a long historic evolutional process via different geographic effects [53, 54].

In a recent study, in order to understand the spatial distribution of the ammonium in the aquifer and the mechanism of abnormally high ammonium groundwater in the Pearl River Delta, a total of 40 boreholes were drilled and core samples of the aquitard and groundwater samples in the basal aquifer were collected [55]. The results showed that the aquifer-aquitard system contains an exceptionally high ammonium concentration. This ammonium was naturally originated in the overlying organic-rich Holocene-Pleistocene aquitard [55]. Hence, it is interesting to explore the status of ammonium-related microorganisms in this special habitat. In the present study, aerobic/anaerobic ammonia/ammonium-oxidizing microorganisms (AOM, including AOA, AOB and anammox bacteria) in the high nitrogen environment in the aquitard were for the first time investigated to understand these microorganisms in such environment.

Materials and Methods

Sample Collection

The sediment samples were collected from the cores of the borehole SDZK14 (N 22° 44′ 41″, E 113° 21′ 59″) drilled in the aquifer-aquitard to the north of Zhongshan, Guandong, China (Fig. 1). The borehole was ca. 50 m deep, drilled by the Department of Earth Sciences, The University of Hong Kong in January of 2008. The samples were immediately put onto ice bags in a heat-insulated cooler after collection and transported to the laboratory. Three sediment depths at 6.9, 23.5 and 37.4 m from the ground level were sampled for this study. Each sample was homogenized and split into three equal parts for chemical analysis, DNA extraction and stored at −20 °C.

Fig. 1
figure 1

The distribution of the ammonium-rich groundwater with NH4 + >10 mg/L (in brown) and the location of SDZK14 (N 22° 44′ 41″, E 113° 21′ 59″) in the Pearl River Delta. The white area is a fluvial and alluvial platform about 5–10 m in elevation and the grey areas are mountain areas with higher elevation (simplified from Jiao et al., 2010)

Full size image

Environmental Physicochemical Analyses

Depth and pH of the sediments were measured in situ at various sediment depths with microelectrodes using the modified protocol at boreholes [56]. Five grams of each sediment sample was extracted with 2.0 M KCl by incubation at 150 rpm at 20 °C on an incubation shaker (Innova 4340, New Brunswick Scientific) overnight. Parameters including ammonium-N, nitrate-N, and nitrite-N were measured with an autoanalyzer (QuikChem, Milwaukee, WI) using standard flow injection analysis (FIA) technique according to the standard extraction procedures.

DNA Extraction and PCR Amplification

Total genomic DNA was extracted from 100 to 200 mg of sediment samples using the SoilMasterTM DNA Extraction Kit, according to the manufacturer’s protocol (Epicentre Biotechnologies, Madison, WI). Soil DNA was amplified with PCR primers targeting different genes (primers listed in Table 1).

Table 1 PCR primers and amplificaiton protocols
Full size table

In order to amplify the anammox 16S rRNA gene using the first pair of PCR primer set (Brod541F/Amx820R), the initial PCR amplification was performed with Pla46F/1037R which corresponded to one of the Planctomycetes 16S rRNA genes and 23S rRNA genes, respectively (Table 1) with GoTaq® DNA Polymerase (Promega, Madison, WI) in a 25-μl volume containing 10 μl of 5× colorless GoTaq® Flexi Buffer, 25 mM of MgCl2, 10 mM of each deoxyribonucleoside triphosphate, 2 μM of each primer, 1.25 U GoTaq® DNA Polymerase, 40 ng ml−1 BSA and 3 μl of DNA as templates (10–100 ng). When amplifying the anammox 16S rRNA gene using the second pair of PCR primer set (Amx368F/Amx820R), the initial PCR amplification was performed with Pla46F/1037R primers. The nested PCR was performed with the anammox-specific primers, Amx368F/Amx820R. The nested PCR condition was identical to that for Pla46F/1037R primers, except 1 min of extension at 72 °C and the cycle number increased to 36. The compositions of archaeal amoA and bacterial amoA PCR in 25 μl reaction mixture were identical to that of anammox 16S rDNA, except using different PCR primer pairs. The amplifying protocols are listed in Table 1.

Clone Library Construction and Phylogenetic Analysis

Purified 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/M13R. 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 by Tech Dragon Ltd (Hong Kong). DNA sequences were examined and edited using BioEdit (Tom Hall, North Carolina State University, NC) and MEGA, version 5.2 [60]. NCBI BLAST (http://www.ncbi.nih.gov) was used to find the closest related 16S rRNA gene and amoA gene sequences in the GenBank. The multiple alignments of partial 16S rRNA gene and amoA gene sequences were done and neighbor-joining phylogenetic trees with bootstrap values based on 1000 replications [61] were produced using MEGA program, version 5.2.

Quantitative PCR Assay

Quantitative PCR (q-PCR) was performed on ABI 7000 Real Time PCR System (Applied Biosystems, USA) to estimate the abundance of ammonia oxidizing archaea and bacteria in all three samples. The quantification was based on the fluorescent dye SYBR-Green I, which binds to double-stranded DNA during PCR amplification. In the process, bacterial and archaeal amoA gene were amplified by the same PCR primer as above. Archaeal and bacterial amoA gene fragments were cloned into pMD18-T plasmids (TaKaRa) and were extracted with QIAprep® Spin Miniprep Plasmid Kit (Qiagen, USA) according to the manufacturer’s protocols. The plasmid concentration was detected by UV-Visible Biophotometer (Eppendorf, Germany).

Tenfold serial dilution was carried out with the known copy number of plasmid of the archaeal and bacterial amoA genes and generated the standard curve over six orders of magnitude. The 25 μl archaeal amoA q-PCR reaction mixture contained 1–10 ng of DNA, 40 ng ml−1 BSA, 0.2 μM of each primer and 12.5 μl of FastStart Universal SYBR Green Master (ROX) (Roche). The modified q-PCR protocol was as follows: 3 min at 94 °C, followed by 50 cycles of 30 s at 94 °C, 1 min at 53 °C with the extension step of 1 min at 72 °C with fluorescence detection at the end of each cycle [13]. Bacterial amoA q-PCR was carried out in 25 μl reaction mixture with 1–10 ng of DNA, 40 ng ml−1 BSA, 0.2 μM of each primer and 12.5 μl of FastStart Universal SYBR Green Master (ROX) (Roche). The modified protocol was as follows: 3 min at 94 °C, followed by 50 cycles of 30 s at 94 °C, 30 s at 55 °C with 45 s extensions at 72 °C with fluorescence detection at the end of each cycle [59]. The intensity of fluorescence was measured at 83 °C and the melting curve analysis was carried out to confirm the specificity of the q-PCR products. The standard curves for q-PCR assays and data analysis were generated with ABI Prism 7000 SDS software (version 1.1).

Statistical Analysis

The DOTUR (Distance-Based OTU and Richness) program was employed to compare diversity for each anammox 16S rRNA and amoA sequence from each sample [62]. Operational taxonomic units (OTU) for community analysis were defined by a 3 % sequence variation in detecting anammox bacteria, AOA and AOB. DOTUR was also used to generate diversity analyses such as coverage, Chao1, Simpson index (D) and Shannon index (H) for each site. The H and D indices were calculated as the diversity indices by the program [62, 63]. The estimated coverage of the constructed anammox related 16S rRNA, 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 stands 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 [64, 65]. Correlation analysis between the abundance of archaeal and bacterial amoA genes and environmental variables were conducted using Microsoft Excel program.

Nucleotide Sequence Accession Numbers

All the amoA genes sequences of ammonia oxidizing bacteria and archaea, and 16S rRNA gene sequences of anammox bacteria determined in this study were deposited in GenBank under accession numbers HM537238 to HM537467.

Results

Physicochemical Characteristics of Sediment Samples

The stratigraphy of the sampling site consisted of over 30 m of Quaternary aquitard of silty sediments. Detailed nutrient concentrations, C/N ratios and other chemical parameters of the three samples from SDZK14 are listed in Table 2. Samples from the site were enriched in ammonium (NH4 +), ranged from 144.8 to 568.6 mg kg−1 dry weight. The pH values of the three sediment samples were between neutral to slightly alkaline (from 6.98 to 7.79). The salinity of the groundwater in the aquifer below 40 m depth was 15.9‰ which was defined as intermediate salinity within the range of the subterranean estuary brackish waters [26]. The nitrite concentration remained low among the three samples with a range from 6.69 to less than 0.1 mg kg−1 dry weight which is even under the detection limit and did not show any characteristic trend. However, the concentrations of nitrate were relatively much higher from 124.3 to 142.1 mg kg−1 dry weight, showing a decreasing trend with the depths.

Table 2 Chemical properties of the high N sediment samples
Full size table

Effect of Depth on Phylogenetic Distribution and Diversity of AOA, AOB and Anammox

The operational taxonomic unit (OTU) was defined based on a 3 % sequence difference cut-off to represent the phylotype diversity of anammox bacteria, AOA and AOB (Table 3). The estimated coverage of the clone libraries was high enough to represent the majority of AOM in the samples. The diversity of AOA in the high ammonium aquitard sediment samples was generally low (Fig. 2a), all archaeal amoA gene sequences fell into the sub-cluster AI which is affiliated with the Water column/sediment sequences [13, 66]. A few archaeal amoA sequences were grouped together and formed a sub-cluster AII and three representatives of culturable archaeal isolates formed the soil/sediment clade (cluster B), defined by two previous studies [66, 67]. All cluster classifications were well supported by high bootstrap value (>96 %). Majority of the clone sequences (84 %) in sub-cluster AI were related to the environmental clone (GQ414591) found from natural drinking water column of Dongjiang River of the Pearl River Delta and the rest 16 % clone sequences were all from 6.9 m depth where the closest representative relative was Ca. Nitrosopumilus maritimus (EU239959) from tropical marine aquarium tank belonging to the ubiquitous marine group 1 Crenarchaeota [4]. Another closely related clone was an environmental archaeal amoA clone (EU860282) recovered from wastewater treatment plants [68]. Surprisingly, none of the clones was closely related with terrestrial clones but all were the same or similar to AOA detected in water related environment.

Table 3 Diversity and richness of (a) archaeal and bacterial amoA; (b) anammox 16S rRNA genes libraries
Full size table
Fig. 2
figure 2

Phylogenetic relationship of the major phylogenetic groups of AOA (a), AOB (b), and anammox bacteria (c). The numbers at the bracket are the clone number of the related phylotype. The numbers at the nodes are percentages that indicate the levels of bootstrap support based on 1000 resample data sets (only values greater than 50 % are shown). Branch lengths correspond to sequence differences as indicated by the scale bar. For Fig. 2a, the clustering reference was defined elsewhere [13, 66, 67]. For Fig. 2c, sequences amplified by Brod541F/Amx820R primers are labeled with filled triangle and those by Amx368F/Amx820R primers are labeled with filled squares correspondingly

Full size image

Evenly distributed and low diversity of AOB in the high ammonium aquitard sediment was detected (Fig. 2b). The distribution of AOB communities was uniform in general among the three depths based on the results in the three AOB clone libraries constructed. All sequences belonged to Nitrosomonas marina/Nitrosomonas oligotropha lineage [69, 70] instead of Nitrosospira group which was only found in acidic soils with between pH values from 4 to 6 [71]. Previous studies reported that Nitrosomonas marina/Nitrosomonas oligotropha lineage was dominant in wastewater treatment sludge and bioreactors [31, 72–74], in wastewater contaminated soil and freshwater sediments [70, 75]. Some species in this lineage adapted to the environment with high ammonium concentrations [76–79] as they have high ammonium affinity with Km values between 0.075 and 0.03 mM [79]. It was divided into two sub-clusters among the defined lineage, with 29 % of all sequenced clones (30 out of the 104 clones) in sub-cluster B closely associated with environmental clones from coastal eutrophication bay (EU244543) (unpublished data) and eutrophic subtropical estuary [63]. But 71 % of those (74 out of the 104 clones) were closely related to Nitrosomonas marina (AAG37813) [70] which was firstly retrieved from sea water, South Pacific [77] in cluster A. Both sub-clusters A and B provided evidence that AOB communities in these samples were related to the eutrophic environment and the libraries were dominated by sequences related to Nitrosomonas, which supports high ammonia concentration has a greater benefit for the Nitrosomonas species growth than that of Nitrosospira [80, 81].

From each individual clone library constructed, there were 1 and 2–3 anammox bacterial 16S rDNA OTUs using PCR primers Brod541F/Amx820R and Amx368F/Amx820R, respectively; 1–2 archaeal amoA OTUs and 2 bacterial amoA OTUs. Among the anammox bacteria clone libraries, the highest diversity of anammox bacteria was found at 23.5 and 37.4 m depths using both anammox PCR primer sets. At 6.9 m depth, anammox bacteria were below detection using the PCR primer set Brod541F/Amx820R; however, two anammox OTUs were found using the PCR primer set Amx368F/Amx820R. On the other hand, the highest diversity of archaeal amoA was detected at depth 6.9 m and evenly distributed bacterial amoA phylotypes among different depths, while the lowest diversity of anammox 16S rRNA genes was at 6.9 m depth (Table 3).

The phylogenetic study of the anammox bacterial community was based on PCR amplification of anammox related 16S rDNA in the samples after extraction. Initial PCR amplification did not result in any sequence with the primers Pla46F/1037R. However, in the subsequently nested PCR amplification reactions, PCR products with a size of 452 bp were obtained from all three samples using primers Amx368F/Amx820R. By amplifying anammox 16S rRNA gene using primers Brod541F/Amx820R, for sample at 6.9 m depth none of the sequences retrieved was related to any of the known anammox bacteria but all the amplicons fell into the Actinobacteria phylum; for samples at depths of 23.5 and 37.4 m, all sequences were affiliated with one anammox cluster, Ca. Scalindua group associated with the clones found in the marsh and river sediment.

From the phylogenetic tree constructed with 16S rDNA sequences recovered using primers Amx368F/Amx820R and Brod541F/Amx820R (Fig. 2c), two clusters amplified by Amx368F/Amx820R were detected in the three clone libraries. All clone sequences in the three samples fell into the two clades, Ca. Scalindua brodae/sorokinii clade and Ca. Scalindua sea sediment-like sequences clade. The closely related cultured relative to the sequenced clones was the annmmox bacterium isolate Ca. Scalindua brodae Strain EN 8 [38] with a high similarity from 93 to 99 %.

Effect of Depth on Abundance of AOA, AOB

The abundances of AOA and AOB were estimated by quantifying the amoA gene copy numbers. The highest archaeal amoA gene copy number was detected in depth of 23.5 m at 1.30 × 105 copies g−1 of dry soil and the lowest archaeal amoA gene copy number was found in depth of 37.4 m, which was eight times lower than that of 23.5 m depth (Fig. 3). The highest bacterial amoA gene copy numbers was found in depth of 23.5 m at 2.89 × 105 copies per gram of dry sediment and the depth of 6.9 m had the lowest bacterial amoA gene copy numbers with two times lower than that of depth of 23.5 m (Fig. 3). Surprisingly, the copy numbers of bacterial amoA gene in all samples ranged from 1.33 × 105 to 2.89 × 105 copies g−1 of dry sediment were higher than those of archaeal amoA copy numbers ranged from 1.62 × 104 to 1.30 × 105 copies g−1 of dry sediment. The AOA/AOB amoA gene ratios in 6.9, 23.5 and 37.4 m were 0.34, 0.45 and 0.07, respectively.

Fig. 3
figure 3

Abundance of bacterial and archaeal amoA genes in different depths of high N site. Error bars indicate standard deviation

Full size image

Archaeal amoA was much more abundant than bacterial amoA in many different environments, for example, semiarid soil, pristine and agricultural soils in climatic zone and sandy soil, and rhizosphere [25, 28–30, 66]. However, in this study the bacterial amoA, rather than archaeal amoA, was predominant in the high ammonium sediment. This is in the consistency with some other reports that AOB amoA copy numbers were greater than AOA amoA in some habitats [8, 14, 26, 32, 82]. AOB was thought to be obligate aerobes [83] before studies showed that Nitrosomonas species in AOB could survive under anoxic conditions [84, 85]. Though AOA appeared to be capable of survival in suboxic condition [9, 10], the role and effectiveness of AOA in nitrifier denitrification under low aeration environment are still unclear [86].

Effect of Environmental Variables on Diversity and Abundance of AOA, AOB and Anammox

The archaeal amoA abundance had a significant positive relationship with pH (Pearson correlation r = 0.97, p < 0.05) and ammonium (Pearson correlation r = 0.96, p < 0.05) (Table 4). The bacterial amoA abundance had a significant negative relationship with C/N ratio (Pearson correlation r = −0.96, p < 0.05). The ratios of AOA to AOB ranged from 0.07 to 0.45 and showed a significant positive correlation to the pH (Pearson correlation r = 0.97, p < 0.05).

Table 4 Statistical analysis of the relationship between physicochemical parameters and archaeal and bacterial amoA gene abundance
Full size table

Oxygen concentration might affect the sizes of the ammonia-oxidizer communities when low oxygen was available across the depth of sediment [87]. The effect of depth to the copy number of AOA amoA gene did not change significantly comparing to that of AOB amoA gene in either fertilized or unfertilized soils [28]. These previous findings might support higher amoA gene abundance of AOB over AOA detected in anoxic sediments in this study with the evidence of better utilization of ammonia by AOB under anoxic conditions.

The bacterial amoA gene copy numbers had a slightly positive correlation with the concentration of ammonia (Pearson correlation r = 0.82, p < 0.05), but a negative correlation with nitrite concentration (Pearson correlation r = 0.92, p = 0.01). It suggested that ammonia input could be a key factor determining the AOB abundance in natural soil environment, consistent with the results found in another study [27].

The inverse correlation of AOB bacterial amoA gene copy number with nitrite concentration in this study agreed with the study of activated sludge bioreactor published recently [31]. It was previously reported that a higher level of nitrite accumulation than this study would inhibit the ammonia-oxidizing activity of AOB by causing the loss of AMO activity under both aerobic and anaerobic conditions and more activity was lost under alkaline than under acidic conditions [31, 88, 89].

The AOB amoA copy number negatively correlated to the C/N ratios (Pearson correlation r = 0.96, p < 0.05), consistent with the results found in an estuary bay [32], freshwater sediment [67] and semiarid soils [25]. It was suggested that the negative relationship between AOB and soil C/N ratio was due to heterotrophs and AOB have high demand of nitrogen and face competition of AOA for available substrates under high C/N conditions [25].

Conclusion

Lower amoA gene abundance of AOA than that of AOB was detected, and anammox bacteria of low diversity at this site were dominated by Ca. Scalindua-related phylotypes. Among the AOA community, all archaeal amoA gene sequences were affiliated with the water column/sediment sequences. For the AOB community, only Nitrosomonas-like sequences with very low diversity were detected, all grouped into the Nitrosomonas marina/Nitrosomonas oligotropha lineage. The archaeal amoA abundance had a significant positive relationship with pH and ammonium, while the bacterial amoA abundance had a significant negative relationship with C/N ratio. The ammonia-oxidizing microorganism communities might cooperate with each other forming a stable dynamic partnership under the stresses of rich ammonia and low oxygen conditions in the aquitard.