Introduction

Over the last few years, the usage of artificial N fertilizer has doubled to achieve high crop yields to cater for the rapidly growing human population (Galloway et al. 2008; Lunau et al. 2013; Tilman et al. 2002), leading to increased nitrogen-saturation of the terrestrial ecosystem. Accordingly, significant amounts of ammonium_N (NH4+_N) and nitrate_N (NO3−_N) have entered into water bodies such as lakes, reservoirs, and seas (Zhou et al. 2007) via tributary rivers, resulting in deterioration of aquatic ecosystems, adverse public health outcomes, and local economy worldwide (Hu et al. 2018; Qian et al. 2018a; Wang et al. 2017; Liang et al. 2005). Furthermore, the excess N can cause eutrophication and hypoxia in aquatic ecosystems, destruction of habitats for resident organisms, and reduction of species diversity (Lanoux et al. 2013; Lenihan and Peterson 1998; Roberts et al. 2014). Thus, the removal and reduction of N in aquatic ecosystems are critical.

A biological approach involving autotrophic nitrification and denitrification is extensively applied for N removal from wastewater and nitrogen-polluted water bodies (Kuypers et al. 2018; Qian et al. 2018b). Nitrification is the aerobic oxidation of ammonium to nitrate (NH4+ → NO2− → NO3−) by autotrophs including ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB), whereas denitrification is the anaerobic stepwise reduction of nitrate to N2 (NO3− → NO2− → NO → N2O → N2) by heterotrophs. Because of the different requirements for dissolved oxygen (DO) and organic carbon, the nitrification and denitrification are operated in separate aerobic and anoxic reactors, respectively, making the conventional biological N removal (BNR) system complex and expensive. Thus, the alternative heterotrophic nitrification-aerobic nitrification (HN-AD) system is extensively studied to overcome the drawbacks of BNR.

Thiosphaera pantotropha capable of performing both heterotrophic nitrification and aerobic denitrification was firstly found in a wastewater treatment plant in 1984 (Robertson and Kuenen 1984). Thereafter, several HN_AD strains including Alcaligenes faecalis No.4 (Joo et al. 2005), Providencia rettgeri YL (Taylor et al. 2009), Rhodococcus sp. CPZ24 (Chen et al. 2012), Chryseobacterium sp. R31 (Kundu et al. 2014), and Pseudomonas stutzeri strain T1 (Guo et al. 2013) were isolated and characterized in wastewater treatment plants and natural water ecosystems, suggesting that there is a great potential to apply the HD-AD for the remediation of nitrogen-laden water body. HN_AD have several advantages compared with the autotrophic nitrification or heterotrophic denitrification, including the occurrence of nitrification and denitrification within the same systems, faster growth, and ability to perform denitrification under high DO levels and at low temperatures (Choi et al. 2016).

Consequently, more researchers have focused on N removal using HN-AD strains in nitrogen-laden water body. For example, Duan et al. (2015) used HN-AD strain SF-16 to treat saline wastewater and found that the average NH4+_N and TN removal efficiencies reached 97.14% and 73.92%, respectively. Tang et al. (2018a, b) applied aerobic denitrifiers coupled with a denitrification agent in the sediment of an urban river for the bioremediation of N pollution and explored the mechanism of aerobic denitrifiers on urban river sediment remediation. Guo et al. (2013) demonstrated that HN-AD strain Pseudomonas stutzeri strain T1, which was isolated from a shallow lake, showed both rapid removal rate and efficiency for NH4+_N and NO3−_N at 60% and 75%, respectively. They also confirmed that the water quality of the sample improved from Grade V to Grade II due to the addition of this strain. Consequently, the HN-AD bacteria are a promising alternative for N removal in nitrogen-polluted water body.

However, the HN-AD has never been investigated with mixed cultures such as in the sediments, and the feasibility of HN-AD in the aquatic ecosystem requires a comprehensive investigation. In this study, a mixed cultures-driven HN-AD system was established in a simulated aquatic ecosystem with the objectives to investigate (1) the effects of organic carbon source, C/N ratio, and NH4+_N concentration on HN_AD performance and (2) the mechanism of N loss in HN-AD system.

Materials and methods

Study area

Qixiang Lake (34° 1′ 30″–34° 1′ 40″ N, 108° 45′ 30″–108° 45′ 40″ E) is located in the campus of Northwest Polytechnical University, Shaanxi, China (Fig. 1), covers a surface area of 0.03 km2, and has 1.4-m mean water depth and 3-m maximum water depth. The lake area has a typical continental monsoon climate characterized by hot and rainy summer, and cold and dry winter, with a mean annual temperature of 15.6 °C (Quan et al. 2013). The mean annual precipitation is approximately 554 mm and the rainfall is concentrated in the summer and autumn (Song et al. 2010).

Fig. 1
figure 1

Sampling sites in the Qixiang Lake of China

Full size image

Eutrophication is a significant problem in the lake for many years, especially in summer (Fig. S1), while the N content was relatively lower in spring without any significant eutrophication. It was hypothesized that microorganisms could have removed N in the lake. Thus, water and sediment samples were collected from the lake in spring to investigate the N removal performance of the microorganisms.

Water and sediment sampling

Water and sediments samples were collected from 20 fixed sites in Qixiang Lake on March 15, 2018, using a TC_Y sampler (TECH Instrument in Shenyang, China) and a home-made grab sampler, respectively, and then mixed and homogenized to form a composite sample. The sediment samples were immediately air-dried and sieved through a 2-mm sieve before further use. The physicochemical properties of water and sediments samples are shown in Table 1.

Table 1 Physicochemical properties of the water and sediment samples [N = 3, mean (std. err.)]
Full size table

Mixed cultures in HN_AD

A total of 300 g sediments were inoculated into 1000_mL brown glass bottles, then 800_mLof lake water was added into each bottle with NH4+_N (or) and NO3-_N as N source. According to Choi et al. (2016), optimum DO in aeration tank for the HN_AD process should be above 2 mg/L. Thus, approximately 0.2 vvm (air volume/liquid volume/min) of mild aeration was supplied using a lab-scale air pump in this study to maintain DO between 2 and 3 mg/L (Fig. 2).

Fig. 2
figure 2

Simulated aquatic ecosystem for mixed cultures

Full size image

Establishment of HN_AD in mixed cultures

As indicated in Table 2, two sets of microbial mixed cultures (each set contained three mixed cultures) were applied to confirm the establishment of HN-AD: (1) the establishment of HN_AD by adding NH4+_N or NO3−_N as the single N source and (2) the establishment of HN_AD by adding NH4+_N and NO3−_N as the mixed N source. For the single N source of mixed cultures, four different media were inoculated into one set of mixed cultures at different periods. The performance of HN-AD was further examined by adding both NH4+_N (205 mg/L) and NO3−_N (230 mg/L) with C/N ratio of 10 in the other set of mixed cultures.

Table 2 Experimental conditions for the establishment of HN-AD
Full size table

Effect of operational factor on the performance of HN_AD

The HN_AD system was operated under different conditions, including different carbon source, TOC/TN mass ratios, and NH4+_N concentrations. Three organic compounds, namely glucose, ethanol and citrate, were employed as the sole carbon source to maintain a TOC/TN mass ratio of 10. The effect of TOC/TN mass ratio (citrate as a carbon source) on HN_AD was examined by varying the ratio between 0 and 20 at a fixed NH4+_N amount of 200 mg/L. Similarly, the NH4+_N concentration in lake water was adjusted to 75.99, 201.91, 267.02, and 482.52 mg/L by adding the citrate as sole carbon source (TOC/TN mass ratio of 10) to study the effect of NH4+_N concentration on HN_AD.

N transformation pathway

Ammonium and nitrate were separately used as the sole N source at TOC/TN mass ratio of 10 to study the possible N loss pathway of the HN-AD in the mixed cultures. Furthermore, Na2WO4 (1.7 mmol L−1) was added to inhibit the nitrate reductase (Nar) to further investigate the aerobic metabolic pathway during the nitrification and denitrification (Chen et al. 2016).

Analytical methods

The pH and oxidation reduction potential (ORP) were determined using a water quality meter (SX700, Shanghai, China) (Wu et al. 2019), while TN and TOC were analyzed using a TOC/TN analyzer (N/C3000ChD, Jena, Germany). NH4+_N, NO2−_N, and NO3−_N concentrations were analyzed after filtering the samples through the disposal 0.45-μm Millipore filters (pore size 0.45 μm, H9059, Heng Odd, Beijing, China). NH4+_N was determined using Nessler’s reagent spectrophotometry (iodine mercury method), while NO2−_N was analyzed by N-(1-naphthalene)-diaminoethane photometry method. NO3−_N was calculated using the absorbance at 220 nm after subtracting the two times background absorbance value at 275 nm. Metal contents (Ca, Na, K, and Mg) were determined by the ion emission spectrometer (ICP-2070, Baird, USA), while Cl− was analyzed by ion chromatography (IC) (792Basic IC, Metrohm, Switzerland).

pH in the sediment was determined using a soil:water (1:1) extract (Pei et al. 2010), whereas total organic matter (TOC) in the sediment phase was analyzed by using a TOC analyzer (N/C3000ChD, Jena, Germany). Total N (TN) was measured using the Kjeldahl method after digesting samples in a digestor using sulfuric acid/mercuric oxide catalyst. To extract mineral N in the sediments, 5 g aliquots of moist sediments were taken and shaken in 25 mL of 1 M KCl for 1 h before filtration through a membrane filter (Pei et al. 2010). NH4+_N, NO2−_N, and NO3−_N concentrations in the filtered extracts were then determined using the relevant methods described above. Specific surface area and microporosity of the sediments were analyzed using the N2-BET method (Tristar II 3020, Micromeritics, USA). Metal contents (Ca, Na, K, and Mg) of sediment were determined by an ion emission spectrometer (ICP-2070, Baird, USA).

Results and discussion

Establishment of HN-AD in mixed cultures

Utilization of NH4+_N or NO3−_N

Four different media were added into one set of mixed cultures to examine whether HN-AD could occur in the stimulated aquatic ecosystem (Fig. 3). The ORP value fluctuated, but was above 100 mV in the mixed cultures during the test (Fig. S2), suggesting an aerobic condition in the simulated ecosystem.

Fig. 3
figure 3

Changes in the concentration of N species at different periods when the N source was either ammonium (Period I and II) or nitrate (Period III and IV) in the mixed cultures

Full size image

In the mixed cultures with NH4+_N as the sole N source in Period I, the NH4+_N concentration decreased gradually at a removal rate of approximately 4.32 mg/L/day in the 42-day period by the heterotrophic nitrification due to the simultaneous consumption of total organic carbon (TOC) (Fig. S3). Although NO3−_N was present due to nitrification with NH4+_N as the initial N source (Eq. 1), the increased concentration of generated NO3−_N was significantly lower compared with the decreased NH4+_N concentration. In other words, N was unexpectedly lost via pathways such as aerobic denitrification. A temporary low level accumulation of NO2−_N was observed, confirming that nitrification took place during Period I.

$$ {{\mathrm{NH}}_4}^{+}+{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\to {{\mathrm{NO}}_3}^{-}+{4\mathrm{e}}^{-}+{6\mathrm{H}}^{+} $$
(1)

Heterotrophic nitrification was confirmed by adding both NH4+_N and external organic carbon from the 43rd to 85th days in Period II. The maximum NH4+_N removal rate was 12.01 mg/L/day when citrate was used as the external carbon source, which was 2.78 times higher than that in the absence of an external carbon source. During this process (Period II), NO3−_N, and NO2−_N accumulations were low, indicating that citrate could significantly promote aerobic denitrification. Guo et al. (2013) suggested that citrate could contribute to either NH4+_N or NO3−_N removal by HN_AD strain with a TOC/TN mass ratio of 10.

A total of 130.00 mg/L NO3−_N source was added as the sole N in the mixed cultures during Period III since the 86th day to further confirm the existence of aerobic denitrification. The NO3−_N concentration decreased gradually at a rate of 1.15 mg/L/day from the 86th to 123rd days, suggesting that aerobic denitrification occurred in the mixed cultures. Furthermore, approximately 85% of TN was removed from the mixed cultures in Period III. Unlike the autotrophic nitrification, NO2−_N concentration was lower than the detection limit of 0.08 mg/L in the aerobic denitrification process.

When citrate was added to the mixed cultures after the 123rd day in Period IV, the NO3−_N concentration decreased dramatically with a removal rate of 14.10 mg/L/day. A significant accumulation of NO2−_N was accompanied by the decrease of NO3−_N. The maximum NO2− concentration was 40.08 mg/L within 4 days from the 123rd to the 127th day and was completely removed since the 129th day, confirming the nitrite generation and consumption during this process. The trend of NO2−_N variation is similar to the NO3−_N concentration variation pattern, further confirming the aerobic denitrification in the mixed cultures.

The findings confirmed that HN_AD occurred in the simulated natural aquatic ecosystem and that the system could remove N using either NH4+_N or NO3−_N as N source under aerobic conditions. A total of 201.91 mg/L NH4+_N and 130.00 mg/L NO3−_N were completely transformed to other N species and finally to N2O or N2 with the TN removal efficiencies reaching approximately 99.61% and 100%, respectively, which are higher than that by the pure HN_AD strain (Table 3). This may be attributed to the abundance of functional microorganism in our mixed cultures and is in agreement with the study by Choi et al. (2016), which reported that the mixed strains were more effective than the pure strain to remove N with TOC/TN of 12.5.

Table 3 N loss percentage through HN-AD
Full size table

Simultaneous utilization of NH4+_N and NO3−_N

Both NH4+_N and NO3−_N were used as mixed N sources at TOC/TN ratio of 10 to further confirm HN_AD in the other set of mixed cultures. As per Fig. 4, the NH4+_N and NO3−_N concentrations declined immediately with the removal rate of 24.48 and 28.56 mg/L/day, respectively, while the TN removal efficiency increased with the decrease of NH4+_N and NO3−_N concentrations. It is worthy to note that a lag phase of 2-day period existed for the NO3− removal and decreased rapidly in the following days similar to Alcaligenes sp. TB (Chen et al. 2016). Yao et al. (2013) and Zhao et al. (2010) also demonstrated that the pure HN-AD strains growth had a lag phase when nitrate was supplied as the N source. The simultaneous NH4+_N and NO3−_N degradation demonstrated that the mixed cultures possessed HN_AD capability.

Fig. 4
figure 4

The NH4+-N, NO3—N, and TN concentration changes when both NH4+-N and NO3−-N were N sources in the mixed cultures

Full size image

Effect of experimental parameter on HN_AD

Types of organic carbon

Organic carbon serves as the energy and electron source for HN_AD bacteria (Joo et al. 2005). As shown in Fig. 5, three organic carbon types such as glucose, citrate, and ethanol could enhance the activities of HN-AD. Similar NH4+_N and TN removal efficiencies suggested that HN_AD bacteria were insensitive to these organic carbons in the mixed cultures. Nevertheless, glucose was not beneficial for N removal by P. stutzeri strain T1 (Guo et al. 2013) at the TOC/TN mass ratio of 10 and inhibit the HN_AD bacterial activity. The growth of the strain L7 could be enhanced by glucose when citrate is with the TOC/TN mass ratio of 6 (Zhang et al. 2012), indicating that the mixed cultures had good tolerance to different organic carbon types. This might be explained by the interaction of many kinds of microorganisms in the studied aquatic ecosystem (Zhang et al. 2012).

Fig. 5
figure 5

Effect of organic carbon on the HN-AD efficiency at the 18th day with NH4+-N as the N source

Full size image

TOC/TN mass ratio

Figure 6 shows the effect of TOC/TN mass ratio on HN-AD performance within 18 days. Significant differences were observed with TOC/TN ratios ranging from 0 to 20. Within the suitable ratio of TOC/TN (0–10), more NH4+_N and TN were lost at higher TOC/TN. When external organic carbon was absent (TOC/TN mass ratio of 0) in the mixed cultures, N loss was observed with NH4+_N and TN removal percentages of 74.10% and 65.86%, respectively. The maximum NH4+_N and TN consumption occurred at TOC/TN mass ratio of 10, although further increase in TOC/TN mass ratio yielded a decrease in N removal efficiency indicating that the N loss occurs at an appropriate TOC/TN mass ratio, whereas the mixed cultures enter a nutritional state at high organic carbon (Becquevort et al. 2007). Most investigations on N loss by heterotrophic nitrifying bacteria were also conducted at TOC/TN mass ratio of 10 (Guo et al. 2013; Joo et al. 2005; Kundu et al. 2014; Taylor et al. 2009). The optimized TOC/TN mass ratio could lead to efficient denitrification without any nitrate and nitrite accumulation (Chen et al. 2016). Therefore, TOC/TN mass ratio of 10 was used in the subsequent tests.

Fig. 6
figure 6

Effect of TOC/TN mass ratio on the HN-AD efficiency at the 18th day with NH4+-N as the N source

Full size image

NH4+-N concentration

Rapid N loss occurred when the initial NH4+_N concentration was higher in the mixed cultures with the TN removal rates of 4.69, 11.01, and 21.03 mg/L/day, at the initial NH4+_N concentrations of 85.53, 201.91, and 493.12 mg/L, respectively (Table 4). The NH4+_N and TN removal rates increased with the increase in initial NH4+_N concentration, suggesting that high NH4+_N concentration did not inhibit the performance of HN_AD in the mixed cultures. The tolerance of HN-AD mixed cultures to NH4+_N was higher than that of heterotrophic nitrifier Providencia rettger YL, which had the maximum tolerance threshold of 300 mg/L NH4+_N (Taylor et al. 2009). This finding confirmed that the mixed cultures were superior to the pure strains in NH4+_N removal at high concentration, indicating the practical applications in actual aquatic ecosystems where ammonium concentration tends to be high.

Table 4 HN-AD ability with different NH4+_N concentrations
Full size table

N transformation pathway via HN_AD

Ammonium and nitrate were separately used as the N sources in the mixed cultures to determine the possible N loss pathways (Fig. 3). When NH4+_N was the N source, N loss in the mixed cultures showed simultaneous HN-AD. The accumulation of N species suggested that the rate of heterotrophic nitrification was higher than that of aerobic denitrification; thus, the N transformation pathways can be described as NH4+ → NH2OH → NO2− → NO3− → NO2− → N2O (N2). When NO3−_N was the N source, N was lost with the accumulation of NO2−_N, further confirming the aerobic denitrification pathway of NO3− → NO2− → N2O (N2).

To further investigate the aerobic metabolic pathway of HN_AD, tests without inhibitor (Fig. 7a) and with Na2WO4 as the Nar reductase inhibitor (Fig. 7b) were conducted with NH4+_N as the sole N source under aerobic conditions. As shown in Fig. 7a, NH4+_N was completely removed within 30 days in the mixed cultures without Na2WO4, indicating that aerobic nitrification had occurred. In contrast, NO3−_N was a final product with its concentration increasing to the maximum of 80 mg/L on 32nd day and then gradually decreasing below its detection limit, indicating subsequent aerobic denitrification. In the mixed cultures with Na2WO4 as nitrate reductase inhibitor, the denitrification was completely inhibited with an accumulation of NO3−_N concentration of 235.13 mg/L (Fig. 7b). As per the N mass balance, 100% NH4+_N were converted to NO3−_N, confirming that N was not lost during the nitrification process and that NO3−_N was an intermediate during the N transformation process.

Fig. 7
figure 7

Utilization of NH4+_N a without inhibition and b with Na2WO4 as the Nar reductase inhibitor. ([NH4+_N = 230 mg/L, 8.4 g K2HPO4, 3.6 g KH2PO4, [Na2WO4] = 500 mg/L)

Full size image

Production of N2 has been widely accepted as a result of HN_AD in wastewater and nitrogen-polluted water bodies (Huang et al. 2013; Joo et al. 2005; Zhang et al. 2012). However, N2O is generally the main end product instead of N2 due to the inhibition of nitrous oxide reductase by O2 (Aboobakar et al. 2013). The two main pathways for aerobic N removal and gaseous N production by heterotrophic bacteria (Richardson et al. 1998) are the heterotrophic nitrification and coupled aerobic denitrification in wastewater (Chen et al. 2016; Chen and Ni 2012; Taylor et al. 2009) or nitrogen-polluted water bodies (Huang et al. 2013). In addition, conversion of ammonium to N2O and/or N2 via NH2OH (Frear and Burrell 1955) by pure HN-AD strains isolated from wastewater treatments is also considered as another route (Joo et al. 2005; Zhao et al. 2010). In such case, the conversion of NH4+_N to NO3−_N should not be detected by 100% with Na2WO4 as nitrate reductase inhibitor (Fig. 7b). Nevertheless, the N balance showed that almost all of NH4+_N was converted to NO3−_N in this simulated aquatic ecosystem when Na2WO4 was added, suggesting that N was not lost via partial nitrification/denitrification pathway (Chen et al. 2016), i.e., NH4+ → NH2OH → NO2− → N2O (N2), instead via the full nitrification/denitrification pathway, i.e. NH4+ → NH2OH → NO2− → NO3− → NO2− → N2O (N2). Concentrations of NH4+_N, NO3-_N, and NO2−_N varied in the HN_AD processes (Fig. 3) further confirming the N transformation pathway. This pathway is generally consistent with that of Acinetobacter sp. Y16 isolated from source water (Huang et al. 2013).

Conclusions

This study demonstrates that HN_AD could occur in the mixed culture of a stimulated aquatic ecosystem. With citrate as organic carbon source at 10 C/N mass ratio, the N loss percentages of NH4+_N (201.91 mg/L) and NO3−_N (130.00 mg/L) were 99.61% and 100.00%, respectively, especially, the high efficiencies were kept at low C/N mass ratio and high NH4+_N concentration (493.12 mg/L). Using Na2WO4 as the nitrate reductase inhibitor, the study found that N loss was not via NH4+ → NH2OH → NO2− → N2O (N2), instead via NH4+ → NH2OH → NO2− → NO3− → NO2− → N2O (N2). NH4+_N, NO3−_N, and NO2−_N concentration variations in the HN_AD processes further confirmed the N transformation pathway, confirming that HN_AD may spontaneously occur and cause N loss in natural aquatic ecosystems. Further, future works will focus on the gene analysis of microorganisms through high-throughput sequencing, which will permit us to get more information on the heterotrophic nitrification-aerobic denitrification in aquatic ecosystems.