The Xingchang Wastewater Treatment Plant in Zhejiang, China, provided the sludge that was used to enrich and cultivate SAD-associated microorganisms. The sludge was cultivated with an enrichment culture solution comprising 0.62 g/L Na₂S₂O₃·5H₂O, 0.36 g/L KNO₃, 0.21 g/L NaHCO₃, 0.34 g/L KH₂PO₄, 0.086 g/L NH₄Cl, 0.076 g/L MgCl₂, 0.003 g/L CaCl₂, and 0.25 mL/L trace elements (Pang and Wang, 2021). The trace element compositions included 0.1 mg/L FeSO₄·7H₂O, 0.03 mg/L H₃BO₃, 0.12 mg/L MnCl₂·4H₂O, 0.12 mg/L CoCl₂·6H₂O, 0.024 mg/L NiCl₂·6H₂O, 0.07 mg/L ZnCl₂, 0.015 mg/L CuSO₄·5H₂O, and 0.036 mg/L Na₂MoO₄·2H₂O. Serum bottles (500 mL), each containing 250 mL of the enriched culture fluid and 100 mL of sludge, were used to domesticate the sludge. Following N₂ flushing for 15 min, the bottles were promptly sealed to establish an anaerobic environment. According to previous studies, S⁰ oxidation under anaerobic conditions can proceed through the following pathways (Pang and Wang, 2021): (1) Biological conversion of S⁰ to SO₃²⁻ (Equation 2); (2) Reaction of S⁰ with SO₃²⁻ to form S₂O₃²⁻(Equation 3); and (3) Biological oxidation of S₂O₃²⁻ to SO₄²⁻ (Equation 4).

Artificial water distribution was employed in the experiments to simulate natural groundwater, comprising 0.45 g/L KNO₃, 0.14 g/L KH₂PO₄, 0.03 g/L MgCl₂, 0.001 g/L CaCl₂, 0.42 g/L NaHCO₃, and 0.1 mL/L trace elements (Pang and Wang, 2021). Each of the 100 mL serum vials was filled with 50 mL of artificially made water. Moreover, 2.5 mM Fe²⁺, 2.5 mM Mn²⁺, and 0.05 g S⁰ were added to the blank control group (CK) to assess the interactions of Mn²⁺, Fe²⁺, and S⁰ in the absence of microorganisms. Each experimental group received 0.05 g of S⁰ and 2 g of cultivated anaerobic sludge following centrifugation of the sludge for 20 min at 5,000 rpm. Five experimental groups and a control group which did not contain Fe²⁺ or Mn²⁺ additions were established (Zeng et al., 2024). The initial Fe²⁺: Mn²⁺ ratios in the five experimental groups were as follows: 1:9 (0.5 mM:4.5 mM), 3:7 (1.5 mM:3.5 mM), 5:5 (2.5 mM:2.5), 7:3 (3.5 mM:1.5 mM), 9:1 (4.5 mM:0.5 mM), 7:3 (3.5 mM:1.5 mM), and 9:1 (4.5 mM:0.5 mM). All experimental and control groups were conducted in triplicate (Zeng et al., 2024). To create an anaerobic environment, each serum vial was sealed with a cap after 15 min of N₂ washing. The mixture was subsequently incubated in a thermostatic shaking chamber for 12 d at 30 °C and 150 rpm. The NO₃⁻-N, NO₂⁻-N, NH₄⁺-N, N₂O, Fe²⁺, Mn²⁺, and SO₄²⁻ concentrations were assessed by sampling the supernatants on Days 1, 2, 4, 6, 8, 10, and 12. On Day 12, samples underwent 16S rRNA high-throughput sequencing and functional gene abundance assays, while the pH of the supernatant was assessed on Days 1 and 12.

Figure 1A demonstrates how the coexistence of various Fe²⁺ and Mn²⁺ ratios affects the NO₃⁻-N concentration in the SAD system. The NO₃⁻-N concentration in systems with varying 5 mM Fe²⁺ and Mn²⁺ ratios were higher than that in the system without Fe²⁺ and Mn²⁺, indicating that the coexistence of 5 mM Fe²⁺ and Mn²⁺ inhibited the removal of NO₃⁻-N in the SAD system. As the Fe²⁺ concentration increased, the NO₃⁻-N removal of the SAD system steadily declined. On the sixth day of incubation, the NO₃⁻-N removal rate from the system decreased significantly from 92.73% without adding Fe²⁺ and Mn²⁺ to 60.96% when an Fe²⁺: Mn²⁺ ratio of 9:1 was applied. Zeng et al. (2024) found that an SAD system with an Fe²⁺: Mn²⁺ ratio of 0:20 achieved a 90.91% NO₃⁻-N removal rate. Similarly, the removal rate of NO₃⁻-N decreased substantially to 86.15% when the Fe²⁺: Mn²⁺ ratio reached 5:5, indicating that NO₃⁻-N removal declined with increasing Fe²⁺ concentration. High Fe²⁺ concentrations may result in this behavior by lowing the pH of the system and adversely affecting microorganisms (Wang et al., 2020). This reduces the effectiveness of the SAD system in removing NO₃⁻-N by limiting the growth of associated denitrifying microbes. Nonetheless, by the 12th day of incubation, the NO₃⁻-N removal rate in all SAD systems with varying Fe²⁺ and Mn²⁺ ratios exceeded 94%. This phenomenon can be attributed to the declines in both Fe²⁺ and Mn²⁺ concentrations toward the conclusion of the reaction, thereby reducing toxicity to denitrifying microorganisms. Additionally, microbial adaptation to the environment with coexisting Fe²⁺ and Mn²⁺ facilitated the effective removal of NO₃⁻-N from the system.

Both NO₂⁻-N and N₂O serve as typical intermediate products of denitrification. N₂O functions as a greenhouse gas with potentially detrimental effects on the ozone layer, whereas low levels of NO₂⁻-N accumulation can also limit microbial activity (Albina et al., 2019). In the pre-experimental phase, the NO₂⁻-N concentrations in systems with varying Fe²⁺ and Mn²⁺ ratios were lower than those in the systems without any additions. Additionally, these concentrations decreased as Fe²⁺ concentration increased (Figure 1B). Fe²⁺ is more prone to electron loss during redox processes, facilitating the NO₂⁻-N reduction (Tekerlekopoulou et al., 2013). At the conclusion of the experiment, the system without Fe²⁺ and Mn²⁺ addition exhibited no NO₂⁻-N accumulation, whereas the systems with 5 mM Fe²⁺ and Mn²⁺ addition at various ratios displayed various levels of NO₂⁻-N accumulation. The system with an Fe²⁺: Mn²⁺ ratio of 5:5 exhibited the highest NO₂⁻-N accumulation at 2.16 mM. The NO₂⁻-N and N₂O concentrations of all experimental systems showed an initially increasing and subsequently decreasing trend (Figures 1B,C). This can be explained by the high NO₃⁻-N concentration in the initial phases of the experiment, which prevents denitrification and causes temporary NO₂⁻-N and N₂O accumulation (Li et al., 2023). Luo et al. (2018) indicated that the NO₂⁻-N reduction rate increased only when Pseudomonas sp. H117 cells were almost completely exhausted with 10 mg/L NO₃⁻-N. This suggests that high NO₃⁻-N concentrations inhibit the NO₂⁻-N reduction process. At the conclusion of the experiment, no N₂O accumulation occurred in the system without Fe²⁺ and Mn²⁺ addition, whereas the systems with Fe²⁺ and Mn²⁺ addition had different degrees of N₂O accumulation. Among these, the system with an Fe²⁺: Mn²⁺ ratio of 1:9 had the highest N₂O accumulation at 1.59 μM. In conclusion, insufficient denitrification in the SAD system may result from introducing 5 mM varying Fe²⁺ and Mn²⁺ ratios, potentially leading to NO₂⁻-N and N₂O accumulation (Wang et al., 2022).

In a system containing 5 mM of coexisting Fe²⁺ and Mn²⁺ in varying ratios, the NH₄⁺-N concentration in the system initially increased and subsequently decreased, exhibiting the highest rate of increase on the first day and then leveling off (Figure 1D). Protein hydrolysis in sludge can yield NH₄⁺-N, with the sharp increase in NH₄⁺-N during the 1st phase attributed to the increase in total protein and elevated hydrolysis levels observed during the early period of increased sludge concentration (Jiang et al., 2024). The decrease in NH₄⁺-N levels in the later stages may be attributed to the sludge entering the methanogenic stage, characterized by a reduction in protein content drops and partial consumption of NH₄⁺-N by microbial growth. The system without Fe²⁺ and Mn²⁺ addition generated the most NH₄⁺-N, potentially because the addition of 5 mM varying Fe²⁺ and Mn²⁺ ratios inhibited sludge hydrolysis. Fe²⁺ is essential for the heme c synthesis, which aids anammox metabolism by facilitating the production of associated enzymes, including hydroxylamine oxidoreductase (HAO), nitrite reductase (NIR), hydrazine synthase (HZS), and hydrazine dehydrogenase (Kartal and Keltjens, 2016). In the system with varying Fe²⁺ and Mn²⁺ ratios, NH₄⁺-N production initially decreased and subsequently increased as the Fe²⁺ concentration increased. This is likely because a low Fe²⁺ concentration increases the related enzyme activity, thereby promoting anammox activity and increasing NH₄⁺-N utilization. Conversely, a high Fe²⁺ concentration inhibits anammox activity. Jiang et al. (2023) found that Fe²⁺ toxicity and the inhibition of anammox activity occur at Fe²⁺ concentrations of 70–80 mg/L in anammox bacteria. Sindhu et al. (2021) also demonstrated that 5 mM Fe²⁺ was detrimental to anammox bacteria, whereas 1 mM Fe²⁺ enhanced anammox activity and increased anammox bacteria quantities.

The SO₄²⁻ concentration increased as the NO₃⁻-N concentration decreased, suggesting that sulfur oxidation and NO₃⁻-N reduction was coupled (Pang and Wang, 2021). Theoretically, 7.54 mg of SO₄²⁻ is produced for every 1 mg of NO₃⁻-N eliminated (Zhao et al., 2024). In this research, the maximum theoretical SO₄²⁻ production was observed in the system without Fe²⁺ and Mn²⁺ addition (Figure 2A). The yield of SO₄²⁻ and removal rate of NO₃⁻-N were positively correlated; the higher the removal rate of NO₃⁻-N, the more SO₄²⁻ was generated (Wang et al., 2019). When 5 mM Fe²⁺ and Mn²⁺ were added to the system, SO₄²⁻ production decreased as the Fe²⁺ concentration increased, with the levels far below those predicted. In addition, Figure 1A shows that NO₃⁻-N removal declined as the Fe²⁺ concentration increased in all treatments. Accompanying this decline, the production of SO₄²⁻ also decreased. This suggests that the SAD system for all treatments exhibits NO₃⁻-N reduction coupled with sulfur oxidation, which aligns with the results of earlier research (Wang et al., 2019). Although SO₄²⁻ is not inherently toxic, excessive consumption can lead to organ damage in humans, and high SO₄²⁻ emissions may considerably disrupt ecosystem stability.

pH can influence the denitrification rate by altering the charged state of substrates and bacterial enzyme proteins within the culture system, thereby influencing nutrient absorption by cells and the corresponding enzyme activity (Song T. et al., 2021). The ideal pH range for denitrification is 6–8. Beyond this range, the N removal rate may be limited, leading to potential NO₂⁻-N and N₂O accumulation (Baeseman et al., 2006). In the absence of Fe²⁺ and Mn²⁺ addition, the pH of the SAD system significantly decreased at the conclusion of the incubation process. When various Fe²⁺ and Mn²⁺ ratios were added, the pH of the system decreased more considerably as the Fe²⁺ concentration increased (Figure 2B), with a minimum value of 5.67 observed at Fe²⁺: Mn²⁺ = 9:1. The above phenomenon could be because Fe²⁺ is a propensity to lose electrons during a redox process, resulting in the formation of H⁺. Additionally, introducing a small quantity of oxygen during sampling can lead to the Fe²⁺ oxidating to Fe³⁺ and subsequent hydrolysis, resulting in a lower pH.

In the absence of microbes, Mn²⁺, Fe²⁺, and S⁰ showed no observable reaction (Figure 2A). Upon the addition of Fe²⁺, several biochemical reactions may occur (Pang and Wang, 2021; Chen et al., 2024): (1) Fe²⁺ can react with NO₃⁻-N if iron-autotrophic denitrifying bacteria become enriched (Equation 5); (2) Fe²⁺ is oxidized to Fe³⁺ by oxygen (Equation 6); (3) The resulting Fe³⁺ may react with S⁰ (Equation 7); and (4) Fe³⁺ can hydrolyze to form Fe(OH)₃ (Equation 8). Similarly, the addition of Mn²⁺ may lead to the following reactions (Bai et al., 2022; Yu and Leadbetter, 2020): (1) Mn²⁺ can react with NO₃⁻-N upon later enrichment of manganese-autotrophic denitrifying bacteria (Equation 9); (2) The generated MnO₂ may react with S₂O₃²⁻ (Equation 10); and (4) In an anoxic manganese-rich environment, MnO₂ can also react with NH₄⁺ (Equations 11, 12).

Both Fe²⁺and Mn²⁺ can act as electron donors for denitrification because they release electrons during oxidative processes (Wang Y. N. et al., 2023). Upon completion of the experiment, the concentrations of Fe²⁺ and Mn²⁺ decreased in all systems, regardless of the initial Fe²⁺/Mn²⁺ ration. As the proportion of Fe²⁺ increased and that of Mn²⁺ decreased, the removal rate of Fe²⁺ progressively declined, while that of Mn²⁺ gradually rose (Figures 3A,B). Bai et al. (2020a) discovered that increasing the initial concentration of Mn²⁺ from 5 to 60 mg/L substantially decreased Mn²⁺ removal from 95.79 to 75.53% via the denitrifying strain Cupriavidus sp. HY129. The strain may exhibit increased activity, utilizing most of the Mn²⁺ ions for denitrification at low Mn²⁺ concentrations. However, high Mn²⁺ concentrations restrict bacterial activity, lower the rate of Mn²⁺ oxidation, and decrease electron release. Zhang et al. (2019) observed that excessive Mn²⁺ disrupted the equilibrium of the system that regulates the metabolism of reactive oxygen species, thereby decreasing biomass. Similarly, as the Fe²⁺ concentration increased, the microorganism activity was suppressed, reducing the removal of Fe²⁺ in the SAD system. According to Dixon et al. (2012), introducing large amounts of Fe²⁺ generates reactive oxygen species that damage cell and organelle membranes and inhibit microbial activity. Chang et al. (2021) reported that a low pH affects specific microbial functions and reduces Fe-oxidizing microorganism activity, thereby decreasing Fe²⁺ removal from the system. This corresponds to the significant decrease in pH along with the increase in the Fe²⁺ concentration.

In systems where S⁰, Fe²⁺ and Mn²⁺ coexist, all three can act as electron donors to drive denitrification. Denitrifying bacteria preferentially utilize electron donors with stronger reducing capacities. Among them, Fe²⁺ possesses the highest reducing power, with an electron release rate significantly exceeding that of S⁰, whereas Mn²⁺ exhibits the weakest reducing capacity, even lower than that of S⁰ (Lin et al., 2022). When S⁰, Fe²⁺ and Mn²⁺ are present together, denitrifying bacteria primarily activate the Fe²⁺ oxidation pathway to transfer electrons to NO₃⁻. This suppresses S⁰ oxidation and limits the electron supply for sulfur autotrophic denitrification, ultimately lowering the nitrate reduction rate. Additionally, electrons released from Fe²⁺ oxidation can elevate the system’s oxidation–reduction potential (ORP), creating conditions unfavorable for the growth of sulfur autotrophic denitrifying bacteria (Yu et al., 2022). Such ORP shifts may also indirectly inhibit the activity of key denitrifying enzymes, such as nitrate reductase (nar) and nitrite reductase (nir), thereby impeding critical steps in NO₃⁻-N reduction. Consequently, even in the presence of S⁰, the electron transport chain cannot function efficiently. Moreover, Fe²⁺ can rapidly transfer electrons to NO₃⁻ by via direct binding to cytochrome c in the bacterial electron transport chain (Cui et al., 2021). In contrast, S⁰-mediated electron transfer requires sulfur oxidases (e.g., sulfide quinone oxidoreductase), which exhibit lower affinity for electron carriers compared to Fe²⁺ (Bobadilla Fazzini et al., 2013). Furthermore, Fe³⁺ produced from Fe²⁺ oxidation may react with S⁰ to form FeS precipitates that coat S⁰ particles, physically blocking contact between S⁰ and sulfur oxidases and further inhibiting electron release from S⁰. When these precipitates deposits on microbial surfaces, they can obstruct NO₃⁻-N entry into cells and potentially hinder nutrient uptake and metabolism (Coby and Picardal, 2005). Therefore, in a coexisting system of Fe²⁺ and Mn²⁺, the NO₃⁻-N removal rate of the SAD system progressively decreased with increasing Fe²⁺ concentration. In comparison, Mn²⁺ has a relatively minor influence on NO₃⁻-N reduction, largely due to its lower reducing capacity.

In the SAD system without Fe²⁺ and Mn²⁺ addition, Proteobacteria (40.33%), Bacteroidetes (34.22%), Desulfobacterota (7.3%), Chloroflexi (6.03%), and Acidobacterota (4.03%) were the most predominant phyla (Figure 4A). Figure 4A shows that the system with varying Fe²⁺ and Mn²⁺ proportions was similar to the dominating phylum of the system absent Fe²⁺ and Mn²⁺. The most prevalent functional bacterial phylum of the SAD system was identified to be Proteobacteria, which are involved in the denitrification process (Lin et al., 2022; Liu et al., 2021). Proteobacteria remained the dominating phylum in the SAD system, exhibiting abundances of 37.83% (Fe²⁺: Mn²⁺ = 1:9), 39.96% (3:7), 35.58% (5:5), 42.85% (7:3), and 37.62% (9:1), despite a decline in their abundance with the addition of various Fe²⁺: Mn²⁺ ratios. Fe²⁺ and Mn²⁺ addition decreased NO₃⁻-N removal from the SAD system, suggesting a positive correlation between the relative abundance of Proteobacteria and the decrease in the N removal rate. Desulfobacterota contributes approximately 12% of the N pathway genes, indicating its importance in the N cycle (Nie et al., 2021). Compared to the system without Fe²⁺ and Mn²⁺ addition, the addition of 5 mM of various ratios of both Fe²⁺ and Mn²⁺ increased the relative abundance of Desulfobacterota. These findings indicate that the coexisting Fe²⁺ and Mn²⁺ may encourage some aspects of N cycling and enhance the denitrification capacity of the system at the conclusion of the experiment. Chloroflexi can decompose soluble organic materials and chemicals that contribute to cellular degradation. They can also collaborate to form network architectures that promote the production of anaerobic ammonia oxidation particles (Kindaichi et al., 2012). Jiang et al. (2023) demonstrated that Fe²⁺ addition enhanced the settling of anammox sludge, promoted its granulation, and increased the quantity of Chloroflexi. However, in this study, the Chloroflexi abundance was lower in systems with different Fe²⁺ and Mn²⁺ ratios than in those without Fe²⁺ and Mn²⁺ addition. This could be explained by the addition of varying 5 mM Fe²⁺ and Mn²⁺ ratios, inhibiting the activity of anammox bacteria. Acidobacterota contributes to the Fe cycle by promoting the transformation of Fe²⁺ into Fe³⁺ (Lu et al., 2010). The relative abundance of Acidobacterota in systems with varying Fe²⁺ and Mn²⁺ ratios was lower than that in systems without Fe²⁺ and Mn²⁺ addition. This can be attributed to the coexistence of Fe²⁺ and Mn²⁺ inhibiting the oxidation of Fe²⁺ in the system.

The top 15 dominant genera at the genus level, including Bacteroidetes-vadinHA17, Thiobacillus, SC-I-84, Limnobacter, Longilinea, PHOS-HE36, Ellin6067, and Herminiimonas, were all associated with the N cycle, indicating that most genus-level microbial communities participated in denitrification and hydrolysis (Figure 4B). Bacteroidetes_vadinHA17 is an unclassified anaerobic genus within the phylum Bacteroidetes. It can degrade complex organic matter in wastewater into readily available carbon sources, which can be utilized by denitrifying bacteria, thus indirectly improving nitrogen removal efficiency in wastewater bioremediation systems (Li et al., 2022; Zhang et al., 2022). The genus Thiobacillus, associated with sulfur-oxidizing bacteria (SOB), was found in all treatments. Previous studies have indicated that Thiobacillus ranks among the most prevalent SAD taxa identified in denitrifying systems across diverse settings (Xu et al., 2020). The relative abundance of Thiobacillus in the system without Fe²⁺ and Mn²⁺ reached 32.32% (Figure 4B). Yang et al. (2018) employed anaerobic sludge from a municipal wastewater treatment facility and thiosulfate as a substrate for enrichment culture studies. Their findings indicated that Thiobacillus was the dominant genus in sludge, with an autotrophic denitrification rate reaching 21 mg N₂/(g VSS·d). In the system with coexisting 5 mM Fe²⁺ and Mn²⁺, the relative abundance of Thiobacillus gradually decreased as the Fe²⁺ concentration increased. Additionally, the SAD effect gradually diminished. The system with the lowest relative abundance of Thiobacillus, which had an Fe²⁺: Mn²⁺ ratio of 9:1, showed a decrease of 19.10%. Pang and Wang (2021) discovered that raising the Fe²⁺ concentration from 0 to 5 mM resulted in a dramatic reduction in the relative abundance of Thiobacillus, from 81.6 to 27.4%. The relative abundance of SC-I-84, an anammox bacterium (Song Y. P. et al., 2021), increased across all treatments as the Fe²⁺ concentration rose, indicating that Fe²⁺ addition enhanced anammox activity. Both PHOS-HE36 and Limnobacter are denitrifying bacteria (Liang et al., 2023), and their relative abundances increased with rising Fe²⁺ concentration (Figure 4B), suggesting a gradual increase in denitrification. In conclusion, SAD activity, anammox, and denitrification predominantly facilitated NO₃⁻-N removal, while changes in the Fe²⁺ and Mn²⁺ ratios influenced the microbial community and distinct reactions involved.

Figure 4C illustrates the genus-level correlation analysis performed between the denitrification-related microbial communities and initial concentrations of Fe²⁺, Mn²⁺, NO₃⁻-N removal rate (NRR), and SO₄²⁻. 76.48% of the variation in microbial community structure at the genus level was represented by the horizontal axis, while 16% was denoted by the vertical axis. The horizontal axis was closely associated with the initial Fe²⁺ and Mn²⁺ concentrations, NRR, and SO₄²⁻ production. In the system without Fe²⁺ and Mn²⁺ addition, the microbial community structure differed significantly from that of the systems containing 5 mM of various Fe²⁺ and Mn²⁺ ratios. The system was more sensitive to Fe²⁺ during NO₃⁻-N removal, as evidenced by the substantial negative correlation between the NRR and SO₄²⁻ production and the initial Fe²⁺ concentration and the positive correlation with the initial Mn²⁺ concentration. Introducing large quantities of Fe²⁺ appears to inhibit the growth of Thiobacillus, consequently affecting SAD function. The negative correlation with the initial Fe²⁺ concentration and the positive correlation with the initial Mn²⁺ concentration for the relative abundance of Thiobacillus evidences this relationship. The significant positive correlation between the NRR and relative abundance of Thiobacillus implies that SAD is primarily responsible for removing NO₃⁻-N from the system, which is consistent with prior research findings (Pang and Wang, 2021).

Figure 5 illustrates the impact of coexisting Fe²⁺ and Mn²⁺ on the number of SAD-related functional genes. When the concentration of Fe²⁺ and Mn²⁺ increased and decreased, respectively, the relative expression of 16S rRNA genes initially increased and subsequently declined (Figure 5A). This suggests that a moderate level of Fe²⁺ benefits microbial enrichment, whereas excessive Fe²⁺ concentrations inhibit microbial growth. The narG gene encoding nitrate reductase catalyzes the NO₃⁻-N reducing to NO₂⁻-N (Zhi and Ji, 2014). The relative expression of the narG gene in the systems with varying ratios of Fe²⁺ and Mn²⁺ was significantly higher than that in the system without Fe²⁺ and Mn²⁺ addition (Figure 5B). When the Fe²⁺: Mn²⁺ ratio was 3:7, the narG gene showed the highest relative expression. This suggests that the addition of varying ratios of Fe²⁺ and Mn²⁺ may enhance the ability of the SAD system to reduce NO₃⁻-N to NO₂⁻-N. Li et al. (2024) constructed a S-Mn carbonate ore denitrification (SMCD) reactor by combining Mn²⁺-rich manganese carbonate ore with SAD and discovered that the 16S rRNA and narG gene abundances in the SMCD reactor were noticeably greater than those in the reactor without manganese carbonate ore addition.

Nitrite reductase catalyzes the NO₂⁻-N reducing to NO, with the corresponding genes being nirK and nirS. Compared with the system without Fe²⁺ and Mn²⁺, the relative expression of the nirS gene substantially decreased with the addition of various Fe²⁺ and Mn²⁺ ratios (Figure 5C). Moreover, the relative expression of the nirS gene exhibited an initially decreasing and subsequently increasing trend in response to rising Fe²⁺ and falling Mn²⁺ concentrations. This suggests that increasing Fe²⁺ concentrations strongly inhibited nirS gene expression, whereas decreasing Mn²⁺ concentrations might have reinstated nirS gene expression. Among them, an Fe²⁺: Mn²⁺ ratio of 5:5 inhibited the relative expression of nirS by up to 45.32%. Compared to the system without Fe²⁺ and Mn²⁺ addition, the low proportion of Fe²⁺ in the system with coexisting Fe²⁺ and Mn²⁺ did not significantly impact the relative expression of the nirK gene. The relative expression of the nirK gene was only significantly suppressed when the Fe²⁺: Mn²⁺ ratio was 9:1, with a suppression rate of 25.84% (Figure 5D). Furthermore, the relative expression of the nirK gene in all systems was lower than that of the nirS gene, indicating that the nirS gene is the dominant gene responsible for reducing NO₂⁻-N in the SAD system (Zhi and Ji, 2014).

In all systems (except the system with an Fe²⁺: Mn²⁺ ratio of 1:9), the relative expression of the norB gene, which encodes nitric oxide reductase and catalyzes the reduction of NO to N₂O, was significantly decreased following the addition of varying Fe²⁺ and Mn²⁺ ratios (Figure 5E). This suggests that the coexisting Fe²⁺ and Mn²⁺ inhibit the reduction of NO to N₂O in the SAD system. The Fe²⁺: Mn²⁺ ratio of 9:1 demonstrated the highest suppression of the relative expression of the norB gene, amounting to a 71.83% reduction. When the Fe²⁺: Mn²⁺ ratio was greater than 1:9 (Figure 5F), the relative expression of the nosZ gene was significantly lower than that in the system without Fe²⁺ and Mn²⁺ addition. This suggests that an increase in the Fe²⁺ concentration in the system with coexisting Fe²⁺ and Mn²⁺ inhibited the systemic reduction of N₂O. The Fe²⁺: Mn²⁺ ratio of 9:1 exhibited the largest inhibition (52.23%) of the relative expression of the nosZ gene. Zheng et al. (2024) investigated the relationship between the autotrophic denitrification rate and Fe/N ratio. They found that increasing the ratio from 2 to 4 decreased the relative abundances of the nosZ and norB genes. In conclusion, adding the appropriate Fe²⁺ and Mn²⁺ addition can boost microbial diversity and abundance, enhance narG gene expression, and promote the NO₃⁻-N reducing to NO₂⁻-N. However, by suppressing the expression of the nirS and nosZ genes, adding 5 mM of various Fe²⁺ and Mn²⁺ ratios prevented the NO₂⁻-N and N₂O conversing, thereby resulting in the NO₂⁻-N and N₂O accumulation.

Consistent with the experimental results of Pang and Wang (2021), the relative expression of Sulfur-oxidizing genes (dsrA and soxB) gradually decreased as the concentration of Fe²⁺ increased (Figures 5G,H). The dsrA gene is engaged in the oxidation of S⁰ to SO₃²⁻, with S⁰ and SO₃²⁻ able to subsequently react to generate S₂O₃²⁻ (Bobadilla Fazzini et al., 2013). Additionally, SOB possessing the soxB gene can oxidize S₂O₃²⁻ to SO₄²⁻ (Chen et al., 2020). Compared with the system without Fe²⁺ and Mn²⁺, the relative expression of the dsrA gene was considerably lower in systems with varying Fe²⁺ and Mn²⁺ ratios. However, the relative expression of the soxB gene was considerably reduced solely when the Fe²⁺ concentration exceeded 1.5 mM (Fe²⁺: Mn²⁺ = 3:7). Furthermore, the relative expression of the soxB gene corresponded to the production of SO₄²⁻ in Section 3.2 (Figure 2A) and the relative abundance of the bacterial community in Section 3.4 (Figures 4A,B). These results indicate that a high Fe²⁺ concentration in the system with coexisting Fe²⁺ and Mn²⁺ inhibits the expression of the soxB gene, which lowers the relative abundance of the SOB genus Thiobacillus, ultimately resulting in decreased SO₄²⁻ production.

To identify the primary driving factors of the NO₃⁻-N removal in the SAD system with 5 mM of varying ratios of coexisting Fe²⁺ and Mn²⁺, using a linear regression analysis to determine the relationship between the removal rate of NO₃⁻-N in the SAD system and relative expression of relevant genes. Figure 6 presents the results, with the linear fitting line indicated by the straight red line. The figure illustrates a significant positive correlation between the removal rate of NO₃⁻-N in the SAD system and relative expression of norB, nosZ, soxB, nosZ/narG, nosZ/nirK, norB/nirK, dsrA/16S rRNA, soxB/nirK, and dsrA/nirK (p = 0.003–0.021). The NO₃⁻-N reducing was mediated by narG, the NO₂⁻-N reducing by nirK, the NO reducing by norB, and the N₂O reducing by nosZ. Consequently, nosZ/narG, nosZ/nirK, and norB/nirK denote the degree of complete denitrification for NO₃⁻-N reduction to N₂ (Pang and Wang, 2021), thereby suggesting that higher levels of complete denitrification are more favorable for the denitrification of the SAD system. Conversely, incomplete denitrification produces intermediates that are toxic to microorganisms, including NO₂⁻-N. This lowers the N removal performance of the system, in which NO reduction and N₂O reduction are the rate-limiting reactions.

Because dsrA is involved in S⁰ oxidation, dsrA/16S rRNA indirectly indicates the capacity of the system to oxidize S⁰ to SO₃²⁻ and the relative abundance of SOB in the system. As previously mentioned, the addition of varying ratios of Fe²⁺ and Mn²⁺ reduced the relative abundance of the SOB Thiobacillus (Figure 4B) and the relative expression of Sulfur-oxidizing genes (dsrA and soxB) (Figures 5G,H). This resulted in the inhibition of the sulfur oxidation process within the system, potentially reducing the denitrification effect of the SAD system. soxB is crucial for S₂O₃²⁻ oxidation, and the rate of denitrogenation in the SAD system was positively correlated with its relative expression (Figure 6C). Therefore, the couple of NO₂⁻-N reduction and sulfur oxidation (S⁰, S₂O₃²⁻) are reflected in soxB/nirK and dsrA/nirK (Pang and Wang, 2021). As described in Sections 3.1 and 3.2, SO₄²⁻ production in the SAD system was positively correlated with the NO₃⁻-N removal rate, implying that the denitrification process was coupled with the sulfur oxidation process in the system. In summary, denitrogenation in the SAD system was primarily driven by nosZ/narG, nosZ/nirK, norB/nirK, dsrA/16S rRNA, soxB/nirK, and dsrA/nirK in the coexisting Fe²⁺ and Mn²⁺, whereas the rate-limiting reactions were NO reduction and N₂O reduction.