Impact of light on anoxic/oxic reactors: performance, quorum sensing, and metagenomic characteristics

Eutrophication has been one of the most serious environmental concerns in recent years (Huang et al. 2017; Chen et al. 2020b), especially in China's Dianchi Lake. Consequently, the wastewater discharge standards are becoming increasingly stringent. In China, the latest ammonia and total nitrogen (TN) discharge standards have been decreased from 5 to 1 mg/L and 15 to 10 mg/L, respectively (Liu & Wang 2017). The newly built wastewater treatment plants (WWTPs) are required to achieve this standard, and the existing WWTPs have to be upgraded to ensure that this standard is achieved. Thus, nitrogen removal has become a major priority of the WWTPs. The anoxic/oxic (A/O) process is one of most renowned and commonly applied techniques in engineering because of its relatively simple operation and stable pollutant removal effect, such as total nitrogen (TN) removal rate up to 50%–70% (Jin et al. 2021).

Considerable amount of work has already been conducted to investigate the factors, such as the hydraulic retention time (HRT), dissolved oxygen (DO), carbon source, step feed process, nitrate recirculation, and sludge retention time (SRT), affecting the A/O performance (Hu et al. 2013; Rasool et al. 2014; Sun et al. 2017). As we have known, the majority of the WWTPs that adopt activated sludge processes are built underground or semi-underground (Hou 2017), so they are not affected by light. If the WWTPs are illuminated for a long time, whether the system is influenced by long time exposure to light is a topic worth investigation. Recently, photo bioreactors have been proposed and have gained increasing attention. Many studies have reported that photosynthetic organisms proliferate in such reactors and that these reactors present an important advantage of simultaneously enhancing carbon, nitrogen, and phosphorus removal (Zhang et al. 2019; Chen et al. 2020a). This indicates that light significantly affects such systems; however, little research has concentrated on the influences of light on activated sludge (Fan et al. 2019).

As is well known, activated sludge is an ecosystem with abundant microorganisms (Daims et al. 2006), which typically results in cell–cell interactions. Numerous studies have shown that cell–cell communication is a phenomenon of quorum sensing (QS), which is mediated by producing signaling molecules (Ding et al. 2015; Maddela et al. 2019). QS is widely observed in various microorganisms and mediates a variety of physiological behaviors, such as nutrient removal and extracellular polymeric substances (EPS) production (Ma et al. 2018; Li et al. 2019). Therefore, it has gained even more attention with respect to wastewater treatment (Tan et al. 2014; Zhao et al. 2018; Maddela et al. 2019; Sun et al. 2019; Wang et al. 2019). Valle et al. (2004) and Chong et al. (2012) demonstrated the presence of QS in activated sludge systems. Yong & Zhong (2013) investigated the mechanism for QS regulation on aromatics degradation, and Li et al. (2015) investigated how QS influenced the cell adhesion, nitrification efficiency of the sludge. However, the QS regulation on activated sludge is still ill-informed.

The 16S rRNA sequencing has been widely applied for exploring community composition; however, 16S rRNA analyses provide limited available information, and do not provide insights into the metabolic function (Gu et al. 2019). Recently, Tax4Fun, a software package for functional prediction profiling based on 16S rRNA gene identification, was proposed by Aßhauer et al. (2015). This approach could use the operational taxonomic units (OTUs) mapped to the SILVA database to predict the functions of the microbial community by assigning the OTUs to the Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology (KO). In many recent researches, Tax4Fun has been successfully applied to predict the functions of a microbial community (Wang et al. 2020a, 2020b).

In this study, the effects of light on the performance, QS, and metagenomic characteristics of two A/O reactors were investigated. Two reactors operating under light and dark conditions were used and the differences in nitrogen removal performance, sludge properties, extracellular polymeric substances (EPS), and N-acyl-homoserine lactones (AHLs) were evaluated. Furthermore, the potential interactions among them are assessed by conducting Pearson correlation analyses and the function prediction profiles were identified based on 16S rRNA genes sequencing to reveal the biological mechanism by which light regulates the reactor performance.

In this study, two identical A/O reactors with 10 L working volumes (a 4 L anoxic tank and a 6 L oxic tank) are operated in parallel, as illustrated in Fig. S1 in Supplementary Information. One reactor (AO1) was operated in the dark, whereas the other reactor (AO2) was illuminated by a 30 W daylight lamp (TL-D, Philips, Netherlands). Further, seed sludge was collected from a WWTP in Tianjin (China), then added to both reactors, yielding initial mixed liquor suspended solid (MLSS) and volatile suspended solid (VSS) concentrations of approximately 3.1 and 1.9 g/L, respectively, in both the reactors.

Every day, synthetic wastewater was prepared and pumped into the A/O systems. Glucose was used as the carbon source, providing a COD concentration of 300 mg/L. Otherwise, the synthetic wastewater composition was as follows: glucose (0.2813 g), KH₂PO₄ (0.0219 g), (NH₄)₂SO₄ (0.1414 g), NaHCO₃ (0.5000 g), CaCl₂ (0.0300 g), MgSO₄·7H₂O (0.200 g), which was the same as that mentioned by Li et al. (2019). A total of 20 L of synthetic wastewater was persistently pumped into each A/O reactor each day, resulting in an HRT of approximately 12 h. The SRT was maintained at approximately 20 days by ensuring daily sludge discharge. Agitation was used to keep the sludge in suspension, and air diffusers were used to supply oxygen in the aeration tank, where the DO was maintained at 2–3 mg/L. Both the reactors were operated at ambient temperature.

Sludge was obtained from the ends of the aerobic tanks of both the reactors over a period of approximately 15–20 days to analyze the EPS. EPS were extracted according to the previous literature (Han et al. 2018). Briefly, a 15 mL sample was centrifuged to remove the supernatant, then ultrapure water was added to recover the original volume and it was shaken to re-suspend the sample. Subsequently, it was heated in a water bath at 80 °C for 60 min. After cooling, the sample was centrifuged again and the supernatant was filtered by a 0.45 μm membrane (Jinteng, China). The polysaccharide (PS) and protein (PN) content of EPS were determined by the anthrone method and Modified Lowry Protein Assay Kit (Sangon, China), respectively. The detailed procedure of excitation-emission matrix (EEM) and parallel factor (PARAFAC) analysis were as similarly reported by Li et al. (2019).

On the 15th, 35th, 55th and 70th days, the effluent was collected from both reactors for AHLs extraction. Extraction of AHLs was conducted according to the procedure proposed by Tang et al. (2015) with minor modifications. Briefly, collect 200 mL effluent from the reactor, and filter it using 0.45 μm membrane (Millipore, USA), then extract 3 times with equal volume ethyl acetate acidified by adding 0.1% (v/v) formic acid. Subsequently, collect the organic layer and evaporate it using a rotary evaporator (Yarong, China). Finally, dissolve the residues in 2 mL methanol for subsequent AHLs quantification. The high performance liquid chromatography mass spectrometer (HPLC-MS) method was used to quantify AHLs and the details can be found elsewhere (Li et al. 2020).

On the 35th and 70th days, sludge samples were collected from both reactors and stored at −80 °C for performing further high-throughput sequencing. DNA was extracted using the Hipure Soil DNA Kit (Magen, China) and the V3–V4 region of the bacterial 16S rRNA genes was amplified by polymerase chain reaction (PCR) according to the mentioned procedure of Li et al. (2018). Subsequently, the MiSeq sequencing platform (Illumina, USA) was used to sequence the purified PCR products. Detailed bioinformatics analysis can be found elsewhere (Li et al. 2018). Finally, Tax4Fun was used to identify the metagenomics characteristics of the sequenced data. The raw data of sequencing have been submitted to Sequence Read Archive of National Center for Biotechnology Information with a BioProject ID PRJNA555142.

The influent and effluent samples were collected from both the reactors every one to two days and immediately analyzed. The water quality (including NH⁺₄-N and TN) and sludge characteristics (including MLSS, VSS, and sludge volume index (SVI)) were analyzed using the standard methods (APHA 2005). The HQ30d multiparameter analyzer (HACH, USA) was used to monitor the DO and temperature regularly. Finally, the sludge particle size was determined by a Mastersizer 2000 laser particle analyzer (Malvern, UK).

Paired t-tests were conducted using SPSS 21.0 as well as Pearson correlation analyses using the online OmicShare tools (www.omicshare.com/tools) to statistically analyze the relations between the two reactors. Here, the relation between the two datasets is considered to be significant if p < 0.05.

Sludge settleability and particle size are two important sludge indices for assessing the operation of a system. Figure 1 depicts the SVI and particle size of both the reactors during the experimental period. The SVIs of two reactors were both in the range of 140–250 mL/g during days 1–35 but rapidly increased from 200 mL/g to more than 600 mL/g during days 36–70, indicating the occurrence of serious sludge bulking. The increasingly cold weather might have been the main reason for the occurrence of sludge bulking after day 35. It was also interesting to note that the particle sizes in both the reactors tended to increase during the experimental period. In this case, the particle sizes might be related to the settleability. To verify this hypothesis, the relationship between the sludge particle size and SVI was studied. During the experimental period, linear relations could be observed between the particle size and SVI in both the reactors (Figure 1(b)), denoting a stronger relation (y = 5.254x − 305.34, R² = 0.9608) in the illuminated reactor (AO2) and a weaker relation (y = 7.5666x − 516.14, R² = 0.7787) in the dark reactor (AO1). In addition, as depicted in Figure 1(b), the linear relations were significantly different between AO1 and AO2, demonstrating that light clearly affected the properties of the sludge. It should not be ignored that the particle size was much larger in AO1 when compared with that in AO2 for the first 40 days; subsequently, the trend reversed. The possible reason for the trend reverse might be the occurrence of sludge bulking. After 40 days, sludge bulking occurred in both reactors, but the sludge bulking in AO2 was more severe than that in AO1. Filamentous bacteria would attach small flocs to its surface to form larger particles. More filamentous bacteria in AO2 than AO1 would make the particle size of AO2 larger than that of AO1.

It has been reported that under specific environmental conditions bacteria will secrete EPS, which has a considerable impact on the properties of activated sludge, such as the settleability and spatial structure of sludge (Sheng et al. 2010). The EPS variations in both reactors are shown in Figure 1(c). The results showed the EPS contents of AO2 (50.88–89.80 mg/g VSS) were much higher when compared with those of AO1 (49.41–67.68 mg/g VSS), and the PN/PS ratios were also higher in AO2 (3.90–8.64) than those in AO1 (3.52–6.50). Thus, evident differences could be observed between the EPS contents and PN/PS ratios of the two reactors, suggesting that light substantially affected EPS production. The PARAFAC analysis combined with EEM was used to identify main EPS components. This analysis indicated that there were two components, as depicted in Fig. S2. Component 1 includes a peak located at Ex/Em of 285/345 nm, which coincided with protein-like substances (Jacquin et al. 2017). Whereas component 2 includes two peaks at approximate Ex/Em values of 230/345 and 270/340 nm, which were identified as tryptophan-like substances (Yang et al. 2015; Maqbool et al. 2016).

The PARAFAC-derived components intensity profiles of EPS in both the reactors during the experimental period were illustrated in Figure 1(d). Here, the two components’ profiles exhibit different trends in both the A/O reactors. The F_max values of both the components were higher in AO2 when compared with those in AO1 during the whole process, consistent with the EPS results. This indicates that the system produced more EPS under illumination because of the promotion of the synthesis of protein-like and tryptophan-like substances by light.

Figure 2 depicts the effluent concentrations and removal efficiency of NH⁺₄-N and TN in both A/O reactors during the experimental period. As indicated in Figure 2(a), the effluent NH⁺₄-N concentrations in both the reactors were less than 5 mg/L and the NH⁺₄-N removal efficiency was more than 90% (averages of 98.68 ± 0.29% and 98.79 ± 0.29%, respectively). This suggests that both the reactors performed well at NH⁺₄-N removal. The significance analysis on NH⁺₄-N removal efficiencies of AO1 and AO2 (p > 0.05) (Table 1) indicated that light did not significantly affect the NH⁺₄-N removal.

Although there was no significant difference in NH⁺₄-N removal between the two reactors, clear differences could be observed with respect to TN removal. As depicted in Figure 2(b), the effluent concentration of TN was much higher in AO2 when compared with that in AO1 during days 1–48. Subsequently, the TN concentrations gradually deteriorated in both the reactors. Correspondingly, the TN removal efficiency of AO1 was significantly higher than that of AO2 over the whole operational period (p = 0.003), especially at 1–35 days (p = 0.000) (Table 1). This suggested that light significantly affected TN removal. However, no significant differences could be observed during days 36–70 (p > 0.05). Interestingly, sludge bulking (SVI > 250 mg/L) was observed in both the reactors after day 40, which became increasingly serious over time (Figure 1(a)). The previous report showed that sludge bulking would result in a significant shift in bacterial compositions (Wang et al. 2016). Therefore, the overgrowth of filamentous bacteria could result in a decrease in the denitrifying bacteria abundance with the occurrence of sludge bulking (Jiang et al. 2016), so the TN removal in both reactors decreased, reducing the difference in TN removal between the two reactors.

Ten kinds of AHLs were detected in both the reactors during the entire experimental period, as depicted in Figure 3(a). In both the reactors, relatively high concentrations of C4–HSL (3.18 ± 0.32 ng/L (AO1) and 3.29 ± 0.21 ng/L (AO2), respectively), C7–HSL (2.72 ± 0.81 ng/L (AO1) and 2.45 ± 0.38 ng/L (AO2), respectively), C8–HSL (2.41 ± 0.30 ng/L (AO1) and 2.31 ± 0.16 ng/L (AO2), respectively), and C10–HSL (3.34 ± 0.17 ng/L (AO1) and 3.07 ± 0.06 ng/L (AO2), respectively) could be observed, indicating that they were dominant AHLs. In previous studies, C4–HSL, C8–HSL, and C10–HSL were the most frequently reported AHLs (Tan et al. 2014; Sun et al. 2019; Zhang et al. 2019). However, only the concentration of C10–HSL in AO1 showed significantly higher than that in AO2 (p < 0.05) (Figure 3(a)), indicating that production of C10–HSL was significantly affected by light. Because light substantially affected nitrogen removal, sludge properties, and EPS production, it should be investigated that whether there were certain relationships among them.

To investigate the potential correlation among the nitrogen removal, sludge properties, EPS production, and AHLs, Pearson correlation analyses were performed on AO1 and AO2, and the results are presented in Figure 3(b) and 3(c), respectively. Here, it can be observed that 3OC6–HSL is significantly positively correlated (p < 0.05) with protein-like substances (component 1) in both the reactors. Significant relations could be observed between C4–HSL and EPS (p < 0.01) and between tryptophan-like substances (component 2) and 3OC12–HSL (p < 0.05) in AO2. These results suggested that large amount of AHLs were involved in mediating EPS production, which might be a reason for more production of EPS in AO2. In case of nitrogen removal, NH⁺₄-N removal had significant negative correlation with C8–HSL (p < 0.05) but positive correlation with C10–HSL (p < 0.05) in AO1. Moreover, a significant negative correlation could be observed between TN removal and SVI in AO1. However, NH⁺₄-N removal was negative related to both 3OC10–HSL and C12–HSL (p < 0.001 and p < 0.05, respectively), and the TN removal was also negatively related to 3OC12-HSL (p < 0.05) in AO2. In addition, the particle size was negatively related to both 3OC10–HSL and 3OC12–HSL. These results demonstrated that AHLs mediated the nitrogen removal, sludge properties, and EPS production; however, the mediated AHLs and behaviors were different depending on whether the reactor was illuminated. Furthermore, it should be noticed that there were significant correlations among the AHLs in both reactors. Thus, it can be assumed that AHLs with different carbon sidechains can be converted to each other but further research is required.

The composition of the microbial community at the genus level in the two reactors is displayed in Figure 4(a). The genus level composition of the microbial communities revealed there were significant differences in the relative abundance of the microbial community, indicating light significantly influenced the microbial community. Chryseobacterium and Thiothrix are filamentous bacteria responsible for sludge bulking (Nielsen et al. 2009), occupied dominant position in both reactors, which was consistent with severe sludge bulking in both reactors. Meanwhile, it was noted that the relative abundance of Chryseobacterium in D35AO2 (i.e., sludge sample collected from the reactor with light and sampled on day 35) and D70AO2 (i.e., sludge sample collected from the reactor with light and sampled on day 70) was lower than that in D35AO1 (i.e., sludge sample collected from reactor with no exposure of light and sampled on day 35) and D70AO1 (i.e., sludge sample collected from reactor with no exposure of light and sampled on day 70). However, the relative abundance of Thiothrix in D35AO2 and D70AO2 was higher than that in D35AO1 and D70AO2. Therefore, the light might promote the growth of Thiothrix, but inhibit the growth of Chryseobacterium. Ferruginibacter, Terrimonas, Acinetobacter, contributed to the degradation of organic substrates and nitrogen removal in denitrification process (Li et al. 2020), were obtained in both reactors. It should be noticed that the relative abundance of Ferruginibacter, Terrimonas and Acinetobacter in D70AO1 (0.27%, 0.24% and 0.1%, respectively) and D70AO2 (0.87%, 0.64% and 0.08%, respectively) were much lower than that in D35AO1 (2.30%, 2.23% and 0.65%, respectively) and D35AO2 (2.12%, 1.28% and 2.81%, respectively). These results also revealed that the overgrowth of filamentous bacteria could result in a decrease in the abundance of denitrifying bacteria, resulting in the lower TN removal.

It is necessary to analyze the function of the microbial community for further understanding the influence of light on the reactors. Tax4Fun was used to predict the functions of the microbial community by assigning operational taxonomic units (OTUs) mapped to the SILVA database to the Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology (KO) and the corresponding pathways to investigate the biological mechanism by which light regulates the reactor performance. The KEGG pathway (at level 2) of microbial community in both A/O is presented in Figure 4(b). In both the reactors, the dominant functions were observed to be carbohydrate metabolism (12.28% ± 0.08% and 13.22% ± 0.72%), amino acid metabolism (12.05% ± 0.04% and 11.64% ± 0.82%), and membrane transport (11.18% ± 0.56% and 12.17% ± 0.02%). The KO abundance of carbohydrate metabolism and membrane transport was much higher in AO2 when compared with that in AO1, indicating that light enhanced the carbohydrate metabolism and membrane transfer potential of the system. Furthermore, the KO abundances of lipid metabolism, nucleotide metabolism, translation, and replication and repair were also higher in AO2 when compared with those in AO1. However, the comparisons between AO1 and AO2 demonstrate that light decreased many KO abundances such as the energy metabolism and amino acid metabolisms. To reveal the effects of light, further level KOs or associated genes should be analyzed. Considering the main differences caused by light were with respect to TN removal, EPS, and AHLs, only the KOs or genes associated with nitrogen removal, synthesis of EPS and AHLs were analyzed in this study.

As is well known, one of the most important functions of activated sludge is nitrogen removal, including the nitrification and denitrification processes. In the process of nitrification, NH⁺₄-N was converted to -N, whereas in the process of denitrification, -N was successively reduced to -N, NO, and N₂ (Waheed et al. 2013). Six genes involved nitrogen removal were identified (Figure 4(c)), namely ammonia monooxygenase [EC:1.13.12.-] (amo), hydroxylamine oxidase [EC:1.7.3.4] (hao), nitrite oxidoreductase, nitrate reductase [EC:1.7.1.1], nitrate reductase [EC:1.7.99.4], and nitrite reductase (NO-forming) [EC:1.7.2.1] (Yu & Zhang 2012; Püttker et al. 2015).

The amo and hao genes mainly relate to the transformation of NH⁺₄-N into -N, whereas nitrite oxidoreductase is involved in nitrite oxidation (⁠-N to -N), and the three forms of nitrate reductase [EC:1.7.1.1], nitrate reductase [EC:1.7.99.4] and nitrite reductase (NO-forming) [EC:1.7.2.1] are functional genes associated with the denitrification process. There were no significant differences in NH⁺₄-N removal, and amo and hao were considerably less abundant than others, so they are not discussed further.

As Figure 4(c) shows, denitrification genes were abundant in both the reactors, indicating that denitrification played dominant roles in nitrogen removal. In addition, the three forms of nitrate reductase were more abundant in AO1, suggesting that the denitrification potential of AO1was much higher than that of AO2. In contrast, nitrite oxidoreductase was more abundant in AO2 when compared with that in AO1. Based on these results, it can be inferred that light increased the nitrite oxidation ability of the activated sludge but decreased its denitrification ability. Consequently, the most probable explanation for the difference in TN removal was that the higher nitrite oxidation potential and lower denitrification potential resulted in a greater amount of nitrate that could not be denitrified in AO2, leading to a high effluent TN concentration. In addition, the mass transfer of nutrients was significantly affected by EPS, suggesting that the higher EPS contents in AO2 were supposed to contribute to the lower TN removal (Sheng et al. 2010).

EPS is the main component of an activated sludge floc and is considered to play an important role in its formation by acting as glue (Li & Yang 2007). Amino sugar and nucleotide sugar metabolism (ko00520) and lipopolysaccharide biosynthesis (ko00540) are correlated with EPS production (Sun et al. 2019). The availability of the sugar nucleotides is an important factor affecting PS biosynthesis (Boels et al. 2001). In AO2, high amino sugar and nucleotide sugar metabolism abundance (Figure 4(d)) indicated that a large amount of sugar nucleotides was produced, resulting in more PS secretion, consistent with the results presented in Figure 2(c). Lipopolysaccharides are another important component of PS (Boels et al. 2001). Therefore, the amino sugar and nucleotide sugar metabolism and lipopolysaccharide biosynthesis were mainly responsible for the PS biosynthesis. However, lipopolysaccharide biosynthesis was less abundant in AO2 when compared with that in AO1. Consequently, the higher content of PS in AO2 was likely because of light promoting the amino sugar and nucleotide sugar metabolism.

PN, with a relatively higher content, is more predominant than PS in EPS. The previous study has reported that amino acids are positively correlated with PN (Feng et al. 2018), especially phenylalanine, leucine, alanine, threonine, glycine, glutamate, and valine, which are considered to be the main synthetic monomers of PN (Zhao et al. 2018). The Tax4Fun function predictions yielded three pathways related to amino acid metabolism, including phenylalanine, tyrosine and tryptophan biosynthesis, valine, leucine and isoleucine biosynthesis, as well as lysine biosynthesis (Figure 4(d)). In AO2, the phenylalanine, tyrosine and tryptophan biosynthesis and lysine biosynthesis pathways were more activated. Furthermore, this high abundance of phenylalanine, tyrosine, and tryptophan biosynthesis corresponded with the result that the intensity of tryptophan-like substances was higher in AO2 than that in AO1 (Figure 2(d)). Therefore, it can be noted that light stimulated the pathways related to the amino acid metabolism in AO2, and these high amino acid synthesis potentials contributed to the high PN content. In addition, another reason for higher EPS production in AO2 shouldn't be ignored. According to the metagenomics function prediction data (Figure 4(b)), light increased the membrane transport function. The results were obtained that QS mediated EPS production in Section “Variations of AHLs”. Therefore, it is possible that the increased membrane transport function has caused more efficient AHLs to mediate EPS production in AO2.

It was reported that the synthase utilized acyl-coenzyme A (acyl-CoA) and acyl-acyl carrier protein (acyl-ACP) as substrates for AHLs synthesis (Huang et al. 2016; Sun et al. 2019). However, the acyl-CoA and acyl-ACP was mediated by the function of fatty acid biosynthesis and metabolism, indicating that the synthesis of AHLs largely depended on the fatty acid biosynthesis and metabolism. Therefore, the level of fatty acid biosynthesis and metabolism can reflect the ability of AHLs synthesis. Based on Tax4fun analysis, the fatty acid biosynthesis pathway was detected at a low level in AO2, indicating that light down-regulated the fatty acid biosynthesis, which resulted in the decrease of AHL synthesis. This result was consistent with the results that the overall AHL concentrations were lower in AO2 when compared with those in AO1. It is suggested that the effect of light on QS is mainly mediated by fatty acid biosynthesis.

Therefore, light not only influenced the reactor performance and sludge properties but also affected the QS and the microbial metabolism pathways. According to existing results, a schematic depicting the likely mechanism was suggested by this study regarding the effect of light on the bioreactor, which is shown in Figure 5. In summary, the light up-regulated the amino sugar and nucleotide sugar metabolism and amino acid synthesis, promoting the production of PS and PN, respectively. Meanwhile, the genes of denitrification and AHLs synthesis were down-regulated by light, resulting in poor nitrogen removal and less AHLs content. Furthermore, the AHLs likely has potential mediation in nitrogen removal and EPS production, and the EPS contents may affect the nitrogen removal performance. This study just provides some preliminary results to understand the influence of light on bioreactor, the mechanism should be future studied and improved in the next work.

This study demonstrated that light significantly influenced nitrogen removal, sludge properties, EPS production, and QS. The TN removal was significantly reduced by light, but the production of protein-like and tryptophan-like substances were increased by light, which resulted in more EPS production. Moreover, the production of AHLs was also reduced by light, especially C10-HSL, resulting in the difference of AHLs-mediated behaviors. The metagenomic function prediction by Tax4Fun provided an in-depth understanding on the manner in which light affects reactors. The amino sugar and nucleotide sugar metabolism and amino acid synthesis were up-regulated by light, increasing the production of PS and PN, respectively. Meanwhile, the pathways related to denitrification and fatty acid biosynthesis were down-regulated by light, resulting in poor TN removal and less AHLs production. These observations will extend insights into the factors affecting the performance of wastewater treatment process and give some inspiration for other scientists. Moreover, future researches should be done to enrich the mechanism of light's influence on bioreactor.

This work was financially supported by the National Natural Science Foundation of China (grant numbers 51778398 and 51578360) and Science and Technology Project of Henan Province (grant number 212102310277).

All relevant data are included in the paper or its Supplementary Information.