1. Introduction
In recent years, excessive nitrogen discharge has caused global nitrogen pollution, with effects ranging from eutrophication to climate change
[1]. To ensure water safety and restore aquatic ecology, the effluent standards of wastewater treatment plants (WWTPs) in China have become increasingly stringent
[2]. Although very strict regulation can improve environmental quality in a short period, it leads to prohibitive treatment costs and energy consumption, thus contradicting the initial purpose of environmental protection
[3],
[4]. Therefore, cost-effective and environmentally friendly technologies for the purification of tailwater are urgently needed to alleviate the burden in WWTPs.
When used as treatment facilities, constructed wetlands (CWs) offer advantages such as inexpensive construction costs, easy operation, and low energy consumption. For these reasons, CWs have been widely applied for secondary and tertiary wastewater treatment
[5]. Generally, the nitrogen metabolism in CWs relies on nitrification and denitrification processes. Nitrification is an oxidation process driven by ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB), which convert ammonia–nitrogen into nitrate
[6]. Denitrification involves a series of reduction reactions that sequentially convert nitrate into nitrite, nitric oxide, nitrous oxide, and finally to harmless nitrogen gas
[7]. The effluent from WWTPs in China typically has the characteristic of low chemical oxygen demand (COD; ≤ 50 mg·L
−1) and ammonia–nitrogen (≤ 5 mg·L
−1) content and high residual nitrate (10–15 mg·L
−1) content, based on the Grade IA in GB18918–2002
[8],
[9]. However, the desired total nitrogen concentration for surface water receiving and assimilating effluents is 1.5 mg·L
−1 (Standard IV in GB3838–2002). Therefore, the tailwater from WWTPs needs to be treated thoroughly to reduce its impact on the surface water environment. Achieving efficient nitrogen removal under low C/N ratio conditions is of great significance for the application of CWs in treating tailwater from WWTPs.
Generally, the main approach for nitrate removal in CWs is heterotrophic denitrification, which requires the oxidation of organic substrates to produce the electrons and energy to support the catalytic reactions involved in NO
x reduction
[10]. Under low C/N ratio conditions, insufficient substrate will lead to inadequate energy and electron supply for denitrification, posing significant challenges to achieve satisfactory denitrification performance. A common solution is to supply an external carbon source to enhance denitrification performance. However, the external addition of organic compounds can increase carbon emission intensity and operational costs, which does not align with the principles of sustainable development
[11]. More recently, researchers have proposed new regulatory directions, and many studies have reported that functional materials with redox properties can enhance denitrification by facilitating intracellular electron transfer
[12],
[13]. However, the specific mechanisms underlying the enhancement of denitrification under low C/N ratio conditions remain poorly understood. In particular, the interrelationships among the generation, transfer, and ultimate utilization of electrons in the denitrification process are still unclear.
The effectiveness of CWs in removing water pollutants greatly depends on the type of substrate used, as it provides a habitat for microorganisms to colonize and interact with pollutants. Among the many candidate substrates for CWs, biochar has received widespread attention because of its high specific surface area, abundant pores, and redox activity characteristics
[14],
[15]. Most previous studies
[16],
[17] have focused on the influence of the physicochemical properties of biochar on the denitrification efficiency of CWs, with less attention paid to the strategies used to enhance denitrification under low C/N ratios. β-Cyclodextrin (β-CD) is an environmentally friendly cyclic oligosaccharide that is easy to synthesize and widely used. In our previous study, we reported that β-CD promoted the nitrogen removal capacity of the denitrification model strain
Paracoccus denitrificans (
P. denitrificans) by improving carbon metabolism, facilitating electron transfer, and regulating iron acquisition
[16]. Furthermore, after grafting onto biochar to synthesize water-insoluble β-cyclodextrin-functionalized biochar (BC@β-CD), we revealed that BC@β-CD could enhance the substrate utilization for supporting denitrification and promote the denitrification performance of
P. denitrificans under a low C/N ratio
[17]. Therefore, BC@β-CD has potential applications as a substrate for enhancing denitrification in CWs under low C/N ratios, yet its specific effects and mechanisms in complex systems remain unclear.
In this study, we used BC@β-CD as a functional material in CWs to assess its impact on denitrification and explore how it enhances this process, in the treatment of wastewater with a low C/N ratio condition. The nitrogen removal performance of CWs with traditional gravel, BC-amended gravel, and BC@β-CD-amended gravel substrates was compared under low C/N ratios (4 and 2) to reveal details of the performance of BC@β-CD optimized CWs. First, variations in the microbial community among different groups were investigated based on metagenomic data, to investigate whether the impact of BC@β-CD on nitrogen removal was because of the enrichment of functional microorganisms. Subsequently, the ability of BC@β-CD to promote electron generation, transportation, and utilization in denitrification was elucidated from a metabolic perspective based on functional gene abundance and enzymology analyses. Lastly, by analyzing the contributions of essential factors to denitrification performance using a structural equation model, we revealed how BC@β-CD regulates the reallocation of carbon metabolism to nitrogen metabolism under low C/N ratios. The results of this study not only represent a novel application of BC@β-CD in CWs but also, more importantly, provide new insights into strategies to strengthen the denitrification of low C/N ratio wastewater.
2. Materials and methods
2.1. Preparation of biochar substrate and synthetic wastewater
In this study, we prepared biochar using reed straw waste from Baiyang Lake (China). According to our previous study
[17], the straw was cut into short-columnar particles (1–2 cm) and pyrolyzed in a nitrogen atmosphere at 800 ℃ for 2 h to obtain the original biochar (BC). Subsequently, β-CD was loaded onto the biochar surface using epichlorohydrin as a cross-linking reagent to prepare functional biochar (BC@β-CD). Details of the process of biochar activation, the grafting reaction, and biochar characterization are provided in Texts S1 and S2 in Appendix A. The BC and BC@β-CD used in this study were both in the form of short-columnar particles and possessed good mechanical strength.
To investigate the effect of BC and BC@β-CD as substrates on the processing of WWTPs tailwater, we prepared synthetic wastewater containing nitrate, ammonia–nitrogen, and total phosphorus at approximately 10, 2, and 1.5 mg·L−1 to simulate effluent that meet the specifications of the Grade IA in GB18918–2002 (Text S3 in Appendix A). To explore the effect of the C/N ratio on the performance of CWs in pollutant removal, glucose was used as a carbon source to adjust the C/N ratio during different operational stages.
2.2. Construction and operating conditions of constructed wetland microcosms
Three types of laboratory-scale CWs were established in cylindrical polyvinyl chloride tanks (height 50 cm, inner diameter 20 cm), namely the control, BC, and BC@β-CD. The bottom supporting layer of CWs was filled with gravel (diameter 1–2 cm) to a depth of 10 cm. For intermediate functional layers, gravel with a diameter of 0.3–0.5 cm mixed with BC and BC@β-CD (volume ratio of 10%) was filled to a height of 30 cm in the BC and BC@β-CD groups, respectively. In the control group, gravel of the same size without biochar addition was added to the same height as in other groups. The tops of the CWs were covered with 10 cm gravel (diameter 1–2 cm) for immobilizing plant roots (Acorus calamus). In the center of the system, a perforated polyvinyl chloride pipe (height 50 cm, inner diameter 5 cm) was installed to collect biological samples from the substrates (Fig. S1(a) in Appendix A).
Before the experiment, all CWs were inoculated with activated sludge collected from the Nanshan Municipal Wastewater Treatment Plant (Shenzhen, China). During the inoculation (Stage I) and start stage (Stage II), the CWs were continuously supplied with sequential batches of synthetic wastewater at a hydraulic retention time (HRT) of 2 d until stable operation was achieved. Subsequently, a formal experiment was carried out to investigate the effect of different substrates on pollutant removal by the CWs. The detailed experimental flowchart and operating conditions are shown in Fig. S1(b) and
Table 1, respectively.
2.3. Collection and analysis of samples during operation
Effluent samples from each finished batch were collected, mixed evenly, and filtered through 0.45 μm filters. The pollutant concentration in each water sample (50 mL) was determined using standard methods established by the American Public Health Association (APHA)
[18].
At the end of Stages III and IV, samples of the substrates in the functional layer in each system were collected and sonicated to extract the biofilm from surface. The activities of functional enzymes related to nitrogen removal and oxidative phosphate processes were determined using microorganism enzyme-linked immunosorbent assay kits (Shanghai Enzyme Co., Ltd., China).
To collect the gas samples, cylindrical polyvinyl chloride hoods (height 50 cm, inner diameter 24 cm) were designed using the water sealing method. The components of gas samples were analyzed using a gas chromatograph (Thermo Fisher Scientific, USA) equipped with a thermal conductivity detector and an electron capture detector
[19].
2.4. Analyses of electron transfer system activity and carbon metabolism supporting denitrification
Electron transfer system (ETS) activity was assessed by determining the efficiency of microorganisms in reducing 2-(
p-iodophenyl)-3-(
p-nitrophenyl)-5-phenyl tetrazolium chloride (INT) to formazan
[20].
To investigate the effects of BC and BC@β-CD on carbon metabolism supporting denitrification, the theoretical COD for denitrification was calculated based on our previous study
[17], with some modifications. The specific procedure and calculation are provided in Text S4 in Appendix A.
2.5. Metagenomic analysis
The biofilm samples extracted from the functional layers of systems at the end of Stages III and IV were filtered, immediately frozen in liquid nitrogen, and cryogenically transported to Majorbio Biopharm Technology Co., Ltd. (China) for metagenomic sequencing.
MEGAHIT (v1.1.2) and Prodigal (v2.6.3) were used to assemble clean reads and to predict open reading frames (ORFs), respectively. The gene sequences predicted from the samples were clustered by CD-HIT (v4.7) to obtain non-redundant gene sequences. Subsequently, Diamond (v2.0.13) was used to obtain species and abundance information at each taxonomic level by comparison with the Non-Redundant (NR) Protein Sequence Database. Functional annotations and statistical information for the sequences were obtained by searching the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.
2.6. Statistical analysis
All the experiments were conducted in triplicate, and the results are expressed as mean ± standard deviation. Significant difference analysis and data mapping were performed using SPSS Statistics (v23, IBM, USA) and Origin 2021 software (OriginLab, USA). A structural equation model was used to analyze the contribution of critical factors to nitrogen removal (AMOS 24, IBM).
3. Results and discussion
3.1. Characterization of BC and BC@β-CD
Substrates provide an area for microorganisms in CWs, and their physicochemical properties affect microbial activity. As shown in
Figs. 1(a) and
(b), the BC surface had many pores and a flat topography at the edges, while the surface of BC@β-CD had many polymer particles. We conducted energy dispersive spectrometry (EDS; ZEISS Sigma 300, Germany) analyses to explore the distribution of carbon, oxygen, phosphorus, and iron on the different substrates. The oxygen content on the surface of BC@β-CD was significantly higher than that on the surface of BC (Figs. S2 and S3 in Appendix A), indicating that more oxygen-containing functional groups were introduced by the β-CD grafting process.
To further analyze the differences in chemical composition between BC and BC@β-CD, Fourier transform infrared spectroscopy (FTIR; Nicolet iS5, Thermo Fisher Scientific) and X-ray photoelectron spectroscopy (XPS; Nexsa, Thermo Fisher Scientific) were employed to analyze variations in the functional group constituents and specific chemical bond compositions. As shown in
Fig. 1(c), the major functional groups on the BC and BC@β-CD surfaces appeared at around 3430–3433, 1590–1617, and 1086–1094 cm
−1, arising from –OH stretching vibration, –COO stretching vibration, and C–O stretching vibration, respectively
[21]. The variations of chemical bond compositions are shown in
Figs. 1(d)–(f). In the C 1s level of the XPS spectrum, the peaks with binding energies of 284.8, 285.6, 286.5, and 288.7 eV are associated with C–C, C–OH, C–O–C, and O–C–O, respectively
[22],
[23]. During the grafting process, epichlorohydrin (EPI) was used to link β-CD to the hydroxyl groups on the BC surface, which explains why C–OH was replaced by C–O–C on the BC@β-CD surface
[21]. This was also demonstrated by the O 1s level of the XPS spectrum, where β-CD grafting dramatically increased the relative amount of C–O–C on the BC@β-CD surface. Biochar prepared from plant biomass may be enriched with phosphorus, so it may release phosphorus into CWs causing pollution. According to the EDS and XPS analyses, after the grafting process, the relative weight ratio and atomic ratio of phosphorus decreased from 1.23% to 1.09% and from 0.41% to 0.18%, respectively. These results indicate that the β-CD grafting process reduced the phosphorus content of BC@β-CD, which is potentially beneficial in decreasing the amount of phosphorus released in CWs. Moreover, the water contact angle reflects the hydrophilicity/hydrophobicity of the substrate, which is related to the number and types of oxygen-containing functional groups on the substrate surface
[24]. As shown in
Figs. 1(g) and
(h), β-CD grafting increased the hydrophilicity of the material surface, indicating that BC@β-CD has better biocompatibility and mass transfer properties
[25].
3.2. Enhancement of pollutant removal performance in CWs
The operation of the CWs was divided into four stages based on the operational parameters. Stage I lasted from day 1 to 10, during which the systems were inoculated with activated sludge. Stage II lasted from day 10 to 40, during which the systems were operated with influent with a C/N of 4 and HRT of 2 d to ensure sufficient nutrient supply for microbial colonization. Formal experiments were conducted once the effluent of each system stabilized after the inoculation and start-up period (0–40 d).
As shown in
Fig. 2(a), compared with the control, both BC and BC@β-CD significantly improved the removal efficiency of total nitrogen in CWs. When the C/N ratio of the influent was 4, the total nitrogen removal rate was lower in the control (32.53%) compared to the BC group (54.16%) and the BC@β-CD group (78.42%) (Fig. S4 in Appendix A). As the C/N ratio decreased from 4 to 2, the average total nitrogen removal rate of the BC@β-CD group decreased to 58.17%, suggesting that the low C/N ratio of influent limited the nitrogen removal capacity of the system. However, the total nitrogen removal rate was still significantly higher in the BC@β-CD group than in the control (15.37%) and the BC group (39.51%). The variation in nitrate removal among the systems was consistent with total nitrogen (
Fig. 2(b)). During Stages III and IV, the nitrate removal rate in the BC@β-CD group was 79.54% and 54.18%, respectively, which was significantly higher than that of the control and BC group at the same stages (Fig. S5 in Appendix A). Moreover, nitrite and ammonia were present at low levels in the effluents of all systems (
Figs. 2(c) and
(d)). These results show that denitrification, especially the nitrate reduction process, dominates the effectiveness of systems for nitrogen removal. Notably, as the C/N ratio decreased from 4 to 2, the nitrate removal efficiency of the control group declined significantly from 27.38% to 7.29%, indicating that the denitrification function of CWs is severely inhibited, especially under very low C/N ratios. Another noteworthy aspect is that nitrous oxide, the important nitrogen removal intermediate, has a potent greenhouse effect and needs to be controlled carefully. As shown in Fig. S6 in Appendix A, nitrous oxide produced by nitrogen removal was well controlled in the BC@β-CD group. Specifically, during Stages III and IV, the nitrous oxide emissions in the BC@β-CD group were 70.57% and 85.45% lower, respectively, than those in the control. These results suggest that BC@β-CD not only enhances the nitrogen removal capacity of CW but also mitigates excessive nitrous oxide emission caused by low C/N ratios. These findings are consistent with our previous study that BC@β-CD promotes the denitrification effect under low C/N conditions, further demonstrating its effectiveness in CW systems
[17]. In addition, Fig. S7 in Appendix A shows that the chemical functional groups on BC@β-CD did not change over the experiment. This indicates that BC@β-CD exhibits good stability in CW systems.
Heterotrophic denitrification is the primary pathway for nitrate reduction, and the carbon source is an essential substance in this process because it directly affects the denitrification rate
[26]. Although the COD removal rates of all systems exceeded 90% in Stages III and IV (
Fig. 2(e) and Fig. S8 in Appendix A), the nitrogen removal performance of each system decreased as the influent C/N ratio dropped, once again demonstrating that the organic substrate content affects the denitrification efficiency in CWs. Interestingly, even with the C/N ratio dropping to two and similar substrate consumption rates among the treatment groups, the denitrification efficiency remained significantly higher in the BC@β-CD group than in the control and BC group. This suggests that BC@β-CD enabled microbes to overcome some of the potential limitations between carbon metabolism and the nitrogen cycle.
Other than nitrogen, excessive phosphorus release also leads to eutrophication and algal blooms
[27]. Hence, we also investigated the effectiveness of each system for total phosphorus removal. As shown in
Fig. 2(f), the removal efficiency of phosphorus was significantly higher in the control and BC@β-CD group than in the BC group. However, in the BC group, the total phosphorus content was higher in the effluent water than in the influent water, indicating that BC slowly released phosphorus into the environment. The release of phosphorus from biochar is mainly affected by the phosphorus content of the feedstock and the preparation process
[28]. The direct use of biochar prepared from plant biomass as the substrate in CWs would have a potential risk of increasing phosphorus release. In contrast, BC@β-CD, after the activation and grafting process, can effectively reduce phosphorus leaching when used as the substrate in CWs.
3.3. Microbial community and nitrogen metabolism pathway analysis
The removal of pollutants in CWs relies mainly on microorganisms that colonize the substrates. Therefore, the composition of the microbial community is one of the potential determinants of the function and efficiency of pollutant removal by CWs. As shown in Table S1 in Appendix A, there was no significant difference in the Chao 1 index among the three systems, whereas the Shannon’s and Simpson’s indexes of the BC@β-CD group at low C/N ratios significantly increased and decreased compared with other systems, respectively. These results demonstrate that BC@β-CD did not affect species richness in the system, but contributed to the conservation of microbial diversity in the systems with low C/N ratios.
In
Fig. 3(a), at the phylum level, Actinomycetota, Pseudomonadota, Candidatus_Saccharibacteria, Chloroflexota, and Bacteroidota were the predominant microorganisms in all systems, accounting for 80.80%–93.23% of the total bacterial community. These phyla are among the dominant bacteria in WWTPs that participate in carbon and nitrogen cycles
[29],
[30],
[31]. Specifically, Pseudomonadota and Bacteroidota are common denitrifying bacteria and Actinomycetota are widespread in nutrient-poor environments
[32],
[33]. In this study, the low nutrient levels of in the influent are likely unfavorable for the growth of other microorganisms, resulting in a relative increase in the abundance of Actinomycetes. The phyla Chloroflexota and Bacteroidota contain many heterotrophic species that can degrade organic compounds
[34].
Fig. 3(b) demonstrates the microbial composition at the genus level in the CW systems. The major potential denitrifiers in the systems included
Nakamurella (Actinomycetota phylum)
[35],
unclassified_p_Candidatus_Saccharibacteria (Candidatus_Saccharibacteria phylum)
[36],
unclassified_o_Burkholderiales (Pseudomonadota phylum), and
unclassified_c_Betaproteobacteria (Pseudomonadota phylum)
[37].
Microlunatus (Actinomycetota phylum) has been reported to be associated with polyphosphate-accumulating organisms
[38]. Notably, the abundance of
unclassified_c_Actinomycetes (Actinomycetota phylum) was 9.62% (Stage III) and 4.90% (Stage IV) in the BC@β-CD group, significantly higher than its abundance in the control (1.43% and 1.87%) and the BC group (1.72% and 3.34%). Because
unclassified_c_Actinomycetota are able to mineralize organic substrates under nutrient-poor conditions, the increased abundance of
unclassified_c_Actinomycetota in the BC@β-CD group probably promoted the utilization of organic substrates.
One potential factor that can impact the effectiveness of pollutant removal in CWs is the abundance of functional genes
[39].
Fig. 3(c) shows the major nitrogen metabolic pathways in CWs, including nitrification and denitrification, according to the KEGG Orthology database. Among the numerous metabolic pathways, there was a high abundance of genes related to the nitrate (EC:1.7.2.1) and nitrite (EC:1.7.5.1 reduction metabolic pathways. The relative abundance of nitrate reduction (EC:1.7.2.1) genes during Stages III and IV was 1.19- and 1.32-fold higher, respectively, in the BC@β-CD group compared to the control. Additionally, the relative abundance of nitrite reduction (EC:1.7.5.1) genes was 1.14- (Stage III) and 1.06-fold (Stage IV) higher in the BC@β-CD group than in the control. These results suggest that BC@β-CD probably enhanced the performance of nitrogen removal by promoting nitrate and nitrite reduction. Considering the significant enhancement of BC@β-CD on nitrogen removal in CWs, while its impact on the abundance of biological communities and functional genes in CWs is limited. As the denitrification efficiency of CWs also depends on other factors such as denitrifying enzyme activity, substrate metabolism, or electron transfer, perhaps other mechanisms regulated by BC@β-CD affect denitrification efficiency in CWs.
3.4. Response of nitrogen metabolic enzyme activity to BC@β-CD
The activities of various enzymes can reflect the functional performance of microbial communities. To better understand the impact of BC@β-CD on microbial activity in the nitrogen cycle, we determined the activities of nitrogen metabolic enzymes. The nitrogen removal process is attributed to two typical bio-reactions (
Fig. 4(a)). One is the nitrification process, which is driven by ammonia monooxygenase (AMO), hydroxylamine oxidase (HAO), and nitrite oxidoreductase (NXR) to convert ammonia–nitrogen to nitrate
[40]. Ammonia oxidation, driven by AMO and HAO, is the first and rate-limiting step in this process and directly affects nitrification efficiency
[41]. As shown in
Figs. 4(b) and
(c), although the AMO and HAO activities in the BC@β-CD group were slightly greater than those of the control, the concentration of ammonia nitrogen in effluent had no difference. This indicates that the removal of ammonia nitrogen in this study did not impact the nitrogen removal in CWs, which may be due to the low ammonia nitrogen content in the influent.
The other bioreaction involved in nitrogen removal is denitrification, which reduces nitrate to harmless nitrogen via the activities of nitrate reductase (NAR), periplasmic nitrate reductase (NAP), nitrite reductase (NIR), nitric oxide reductase (NOR), and nitrous oxide reductase (NOS)
[42]. NAR and NAP catalyze the first step of the denitrification process: the conversion of nitrate to nitrite
[43]. As shown in
Fig. 4(d), the activity of NAR in the BC@β-CD group was 427.89 (Stage III) and 462.02 (Stage IV) U·mg protein
−1, significantly higher than the corresponding values in the control (274.62 and 327.26 U·mg·protein
−1) and the BC group (372.45 and 399.7 U·mg·protein
−1). Moreover, NAP activity was the same in each period in all treatment groups and its activity level was much lower than that of NAR (Fig. S9 in Appendix A). These results suggest that BC@β-CD promotes the nitrate reduction process by enhancing NAR activity. The activities of downstream denitrification functional enzymes, NIR and NOR, were similar among the three groups (
Figs. 4(e) and
(f)). Notably, although the relative abundance of nitrite reduction (EC: 1.7.5.1) functional genes was slightly higher in the BC@β-CD system than in the control and BC group, there was no significant difference in NIR activity among the groups, suggests that the process of nitrite reduction is not the rate-limiting step of denitrification in CWs. Furthermore, NOS activity in the BC@β-CD group was 58.86 (Stage III) and 57.99 U·mg·protein
-1 (Stage IV), which was 43.64% and 34.92% higher than its corresponding values in the control at the same stages (
Fig. 4(g)). Interestingly, NOS activity did not match well with the relative abundance of genes involved in nitric oxide reduction (EC: 1.7.2.4). The activity of denitrifying enzymes not only depends on the abundance of functional gene but also related to other impact factors such as functional genes transcription, substrate metabolism, or electron consumption. So, these findings indicate that BC@β-CD enhances denitrification mainly by regulating metabolic processes rather than solely relying on the abundance of denitrifying functional genes. Furthermore, the decrease of the C/N ratio in influents caused a significant inhibition of the denitrification performance, denitrification enzyme activity did not decrease in each group. This further illustrates that the generation and transfer of electrons are the vital factors influencing the efficiency of CWs nitrogen removal. Therefore, subsequently, we will focus on studying the regulatory role of BC@β-CD in these aspects.
3.5. Impact of BC@β-CD on carbon metabolism to generate electron donors
The denitrification process relies on the electrons and energy derived from carbon metabolism and oxidative phosphorylation
[44]. Microorganisms metabolize glucose as the organic substrate mainly through the Embden−Meyerhof−Parnas (EMP) pathway and the tricarboxylic acid (TCA) cycle to generate electron donors (NADH) and energy carriers (adenosine triphosphate (ATP)) (
Fig. 5(a)). As shown in
Figs. 5(b) and
(c), the results of pathway enrichment analysis showed that the genes associated with EMP (glycolysis) and TCA cycling were significantly more abundant in the BC and BC@β-CD groups compared to the control. These results suggest that the enhancement of nitrogen removal performance in the BC and BC@β-CD groups might be attributed to better substrate metabolism.
To obtain insights into the mechanisms through which BC and BC@β-CD enhance carbon metabolism in CWs, we analyzed the changes in the abundance of genes related to the EMP pathway and the TCA cycle.
Fig. 5(d) shows that the abundance of most genes participating in the EMP pathway was higher in the BC and BC@β-CD groups than in the control, suggesting that BC and BC@β-CD promoted the primary metabolism of glucose in CWs. During Stage III, compared with the BC group, the BC@β-CD group had a significantly higher abundance of genes encoding hexokinase (EC:2.7.1.1), glucokinase (EC:2.7.1.2), polyphosphate glucokinase (EC:2.7.1.63), glucose-6-phosphate isomerase (EC:5.3.1.9), fructose-1,6-bisphosphatase II (EC:3.1.3.11), and glyceraldehyde-3-phosphate dehydrogenase (EC:1.2.1.59) (Table S2). Glyceraldehyde-3-phosphate dehydrogenase (EC:1.2.1.59) is an important participant in the EMP pathway for the synthesis of NADH. Therefore, BC@β-CD can promote the biosynthesis of NADH by enhancing the EMP pathway. Notably, more highly abundant EMP pathway-related functional genes were found in Stage IV in the BC@β-CD group, with the abundance of 21 and 18 genes increased among the 24 detected genes compared with the control and the BC group, respectively (Table S2). These results suggest that BC@β-CD promoted glucose metabolism by enhancing the EMP pathway under low C/N ratio conditions.
A more thorough substrate metabolism process uses the product of the EMP pathway, acetyl-CoA, to continue producing NADH and ATP in the TCA cycle
[45]. Similar to the pattern observed for the EMP pathway, in the TCA cycle, most of the functional genes were more abundant in the BC and BC@β-CD groups compared to the control, suggesting that glucose was more thoroughly metabolized in the BC and BC@β-CD systems (
Fig. 5(e)). Specifically, there was a higher abundance of genes encoding ATP citrate (pro-S)-lyase (EC:2.3.3.8) and succinyl-CoA synthetase beta subunit (EC:6.2.1.4 and 6.2.1.5) in the BC@β-CD group. However, the number of functional genes related to the TCA cycle that increased and decreased was roughly equal in the BC and BC@β-CD groups, indicating that both BC and BC@β-CD have a similar effect on the TCA cycle. Therefore, BC and BC@β-CD led to an increased synthesis of NADH and ATP for the denitrification process by promoting carbon metabolism. Nevertheless, compared to BC, BC@β-CD must have other advantages that result in the better denitrification performance.
3.6. Contribution of efficient carbon utilization to nitrogen removal
The above results in Section 3.5 demonstrate that BC and BC@β-CD improved the generation of electron donors (NADH) and energy carriers (ATP) by enhancing glucose metabolism. However, the similar promotion of glucose metabolism in the BC and BC@β-CD groups led us to suspect that the better denitrification effect in the BC@β-CD group might be attributed to better electron utilization and energy efficiency. Given that the electrons for denitrification are generated from the oxidation of the electron donor (NADH) by NADH dehydrogenase in the process of oxidative phosphorylation, we further investigated the abundance of functional genes and the actual activity of NADH dehydrogenase in CWs. As shown in
Fig. 6(a) and Table S4, the abundance of all functional genes related to NADH dehydrogenase was higher in the BC and BC@β-CD groups than in the control, at both Stages III and IV. Notably, compared to the BC group, the BC@β-CD group exhibited a significantly higher number of genes associated with NADH dehydrogenase with increased abundance, as indicated in Table S4. These findings were consistent with the variations in the activity of NADH dehydrogenase in CWs. As depicted in
Fig. 6(b), the activity of NADH dehydrogenase was higher in the BC@β-CD group than in both the control and BC groups during the same period. Specifically, there was 39.87% (Stage III) and 61.60% (Stage IV) increase compared with the control, as well as 13.76% (Stage III) and 28.62% (Stage IV) increase compared with the BC group. These results demonstrate that BC@β-CD promoted the generation of electrons used for denitrification by enhancing the activity of NADH dehydrogenase. Another important function of the oxidative phosphorylation process is ATP synthesis, which is catalyzed by F-type ATP synthase. It has been reported that more than 90% of intracellular ATP is synthesized by ATP synthase
[46].
Fig. 6(c) and Table S5 show that the expression of nearly all ATP synthase-related genes was higher in the BC and BC@β-CD groups than in the control at Stage IV. Meanwhile, among the 13 detected functional genes related to ATP synthase, the abundance of 10 was higher in the BC@β-CD group than in the BC group, while only three exhibited lower abundance. Correspondingly, adenosine triphosphate (ATPase) activity in the BC@β-CD group was 17.77% higher than that in the control, and 15.84% higher than that in the BC group at Stage IV (
Fig. 6(d)). Hence, BC@β-CD results in enhanced ATPase activity under low C/N ratio conditions, which provides energy for microbial growth and metabolism.
The denitrification process is directly linked to ETS activity (ETSA), as the electrons necessary for denitrification are generated through NADH oxidation and subsequently transferred to denitrifying functional enzymes via ETS
[20]. As shown in
Fig. 6(e), the BC@β-CD group exhibited the highest level of ETSA. Compared with the control and BC groups, the BC@β-CD group showed 25.35% and 12.89% higher ETSA at Stage III, and 22.06% and 14.27% higher ETSA at Stage IV, respectively. These results suggest that BC@β-CD not only promotes carbon metabolism to produce electron donors (NADH) but also promotes the oxidation of NADH to produce electrons and delivers sufficient electrons to denitrify functional enzymes ultimately enhancing denitrification.
Considering the presence of various other biochemical processes in microorganisms that require energy and electron consumption, BC@β-CD may enhance the competitiveness of the denitrification processes for electrons by promoting the efficiency of electron generation and transfer. To validate this hypothesis, the ratio of theoretical COD consumed by the denitrification process to actual COD consumed was calculated to evaluate the efficacy of carbon metabolism in supporting the denitrification process. As shown in
Fig. 6(f), the ratios of carbon metabolism supporting the denitrification process were 64.13% (Stage III) and 55.57% (Stage IV) in the BC@β-CD group, significantly higher than those in the control (21.13% for Stage III and 4.94% for Stage IV) and BC group (39.57% for Stage III and 33.30% for Stage IV).
The above results demonstrate that BC@β-CD boosts denitrification in CWs through various approaches, including optimized utilization of carbon sources, enhanced electron generation and transfer capacity, and heightened denitrifying enzyme activity. We used a structural equation model to further understand the interrelationships among the enhancement approaches. During Stage III, the ratio of carbon metabolism supporting denitrification had the most significant effect on the denitrification rate (
Fig. 6(g)). The denitrification functional enzymes represented by NAR and NOS were supported by NADH dehydrogenase and ETSA. This enhanced their activities and increased the ratio of carbon metabolism supporting denitrification. Notably, the original interrelationships among the enhancement approaches changed significantly as the C/N ratio decreased in Stage IV (
Fig. 6(h)). Microorganisms rely more on enhancing NADH dehydrogenase activity to facilitate substrate metabolism conversion into electrons that support denitrification reactions. Overall, despite variations in the emphasis of regulatory pathways, the BC@β-CD system favors the reallocation of carbon metabolism to more strongly support denitrification, thus improving the effectiveness of nitrogen removal in CWs under low C/N ratio conditions.
3.7. Environmental implications
Herein, we used CWs to treat the low C/N ratio simulated tailwater in depth. However, due to the constraints of the C/N ratio, the denitrification performance of traditional gravel substrate CWs is unsatisfactory, with a significant amount of nitrate remaining difficult to remove. This drawback results in a large amount of nitrate entering the aquatic environment with the effluent, leading to water eutrophication, especially when the influent C/N ratio fluctuates. Moreover, the emission of the notorious greenhouse gas nitrous oxide from CWs, caused by nitrogen removal, increased over 2-fold during the C/N ratio dropped from 4 to 2 in the influents, probably exacerbating global warming.
The BC@β-CD-amended gravel substrates significantly enhance the denitrification efficiency of traditional CWs for wastewater with low C/N ratios and dramatically mitigate the release of nitrous oxide. However, even with the BC@β-CD-amended substrates, the nitrate cannot be removed completely. Additionally, the nitrogen concentration in the effluents was higher than 1.5 mg·L−1 (GB3838–2002) when the influent C/N ratio was 2. For in-depth denitrification of low C/N sewage using CWs, we propose two suggestions from the perspectives of organic substrate sourcing and reducing organic substrate dependence by constructing multi-stage CWs: ① Organic substrate sourcing: utilize the root exudates produced by plants as a supplementary inner carbon source; ② reducing organic substrate dependence: employ the denitrification capabilities of autotrophic microorganisms in multi-stage CW systems to remove the remaining traces of nitrate from the water.
4. Conclusions
A low C/N ratio always constrains the denitrification performance of CWs, hindering their ability to improve the water quality of effluent in WWTPs. Therefore, resolving the bottleneck of poor denitrification efficiency caused by insufficient organic substrate in the influent is crucial. In this study, we applied BC@β-CD as a substrate in CWs and found that it can allocate more carbon metabolic flow to support denitrification, thereby enhancing denitrification efficiency and mitigating N2O emissions. Our results show that BC@β-CD enhances the carbon and nitrogen metabolism activity of individual functional microorganisms in the system, rather than altering the original microbial community composition. A mechanistic approach combining metagenomics and enzymology analyses reveals that, compared with BC, BC@β-CD not only enhances carbon metabolism but also strategically allocates more carbon metabolism flow to denitrification by elevating both the oxidative phosphorylation process and electron transfer efficiency. This enables denitrification to receive maximum support when carbon sources are insufficient, further ensuring nitrogen removal under low C/N ratio conditions. Moreover, structural equation modeling analysis confirmed that the core strategy for enhancing nitrogen removal efficiency under low C/N ratio conditions is to increase the allocation of carbon metabolism towards supporting denitrification. Certainly, functional materials will help the denitrification process overcome substrate limitations and reduce dependence on additional carbon sources from the perspective of rational carbon metabolism allocation and efficient utilization, providing new solutions for low-carbon wastewater treatment.
Acknowledgments
This study was financially supported by the National Natural Science Foundation of China (52321005), the Guangdong Basic and Applied Basic Research Foundation (2023A1515012383 and 2024A1515030138), the State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology; 2021TS30), and the Shenzhen Science and Technology Program (KQTD20190929172630447 and KCXFZ20211020163404007).
Compliance with ethics guidelines
Hong-Tao Shi, Xiao-Chi Feng, Zi-Jie Xiao, Chen-Yi Jiang, Wen-Qian Wang, Qin-Yao Zeng, Bo-Wen Yang, Qi-Shi Si, Qing-Lian Wu, and Nan-Qi Ren declare that they have no conflict of interest or financial conflicts to disclose.
Appendix A. Supplementary material
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2024.07.020.
PDF
(4691KB)
