Salinity stress results in ammonium and nitrite accumulation during the elemental sulfur-driven autotrophic denitrification process

In this study, we investigated the effects of salinity on elemental sulfur-driven autotrophic denitrification (SAD) efficiency, and microbial communities. The results revealed that when the salinity was ≤6 g/L, the nitrate removal efficiency in SAD increased with the increasing salinity reaching 95.53% at 6 g/L salinity. Above this salt concentration, the performance of SAD gradually decreased, and the nitrate removal efficiency decreased to 33.63% at 25 g/L salinity. Approximately 5 mg/L of the hazardous nitrite was detectable at 15 g/L salinity, but decreased at 25 g/L salinity, accompanied by the generation of ammonium. When the salinity was ≥15 g/L, the abundance of the salt-tolerant microorganisms, Thiobacillus and Sulfurimonas, increased, while that of other microbial species decreased. This study provides support for the practical application of elemental sulfur-driven autotrophic denitrification in saline nitrate wastewater.


Introduction
As anthropological activity increases, an increasing amount of nitrogen-containing wastewater is discharged into the environment, causing eutrophication, which threatens ecosystems, physical health, and ultimately human survival (Jiang et al., 2020;Xu et al., 2022).Nitrate-containing wastewater can cause the quality of water bodies to deteriorate thereby increasing the risk of methemoglobinemia, non-Hodgkin lymphoma, and heart disease in humans (Nuhogl et al., 2002).Consequently, treating nitratecontaining wastewater to meet the emission standards of different industries is essential.In recent years, biological denitrification technology has been identified as an efficient and cost effective method for removing nitrate-nitrogen (NO − 3 -N) from water bodies (Li S. et al., 2020).Biological denitrification technology includes heterotrophic and autotrophic denitrification in accordance with differences in denitrifying microbial carbon sources (Ren et al., 2021).Heterotrophic denitrification (HD) technology has the advantage of a rapid reaction rate and high denitrification efficiency.However, additional carbon is required during the HD process, and a lot of sludge and CO 2 are produced, which limits the achievement of carbon reduction and neutrality goals (Pang and Wang, 2021).Conversely, autotrophic denitrification, particularly sulfur autotrophic denitrification (SAD), does not require additional carbon sources.In addition, low amounts of sludge and CO 2 are produced during this process.Therefore, SAD gradually became a promising nitrate-reducing wastewater treatment method (Tian and Yu, 2020).
Salinity has various effects on SAD, which are driven by different reductants.Usually, most nitrate-containing wastewaters such as contaminated groundwater (Zhang et al., 2023), aquaculture wastewater (Ching and Redzwan, 2017), and agricultural production wastewater (Lefebvre and Moletta, 2006) contain high salt concentrations (0-50 g/L).Salinity affects microbial activity, thus influencing the NO − 3 -N removal efficiency (Navada et al., 2020;Li et al., 2023).Zhu et al. (2023) found that a salinity of 80 g/L significantly inhibited the growth of Sulfurimonas and Thiobacillus, thereby decreasing the effectiveness of using bacteria for NO − 3 -N removal in the SAD system, with thiosulfate as the reductant.Li et al. (2023) found that anammox and shortcut sulfur autotrophic denitrification (SSADN, with sulfide as the reductant) activity increases with increasing salinity when the salinity was <2%.When the salinity reached 5%, the nitrogen removal efficiency of this system decreased significantly, and the microbial structure and abundance changed noticeably.Shen et al. (2023) explored the effects of salinity on double short-cut sulfur autotrophic denitrification (DSSADN, using sulfide as the reductant).A salinity of 1.5% strongly stimulated NO − 3 -N removal by microorganisms, but once the salinity reached 3% it had an inhibiting effect.
In addition, different types of influent have significant differences in the treatment of pollutants in saline wastewater.Current reactor inlet water types are divided into the sequential batch reactor (SBR) and continuous flow reactor (CFR).SBR has good biomass sedimentation and effective resistance to water and pollutant impacts (Yan et al., 2022), and has been widely used to treat NO − 3 -N in saline wastewater (Wang et al., 2020;Zhang et al., 2022).However, compared to the CFR, the SBR has the disadvantage of a complex design, unstable effluent, and the need for regular maintenance (Corsino et al., 2016).For largescale treatment plants requiring a large amount of continuous wastewater, the CFR operation has higher treatment efficiency and better treatment consistency than the SBR.Zhu et al. (2022) compared the ability of the SBR and the CFR to treat high-salinity wastewater, and the results showed that the CFR exhibited better pollutant removal performance than the SBR.Compared to other reductants in the SAD process, the elemental sulfur-driven SAD process has exhibited low toxicity, low cost, stability, and it is widely used in practical wastewater treatment plants that rely on a continuous inflow (Li Y. et al., 2020;Zhou et al., 2021).Nonetheless, little attention has been paid to the influence of salinity on the SAD process driven by S 0 continuous flow bioreactors.
In this study, elemental sulfur was used as the reductant in a SAD reactor.NaCl was added to study the effect of salinity on nitrogen transformation.The relationship between salinity and microbial community structure was also analyzed.The results provide a reference for the application of SAD in high-nitrate high-salinity wastewater.

Materials and methods . Test equipment and operating process
The experimental setup consisted of a cylindrical Plexiglas denitrification bioreactor (outer diameter: 8.6 cm; inner diameter: 7.6 cm; total height: 100 cm; and volume: 4 L) (Figure 1).The reactor was filled with elemental sulfur and limestone particles in a 1:1 volume ratio, and the fillers were 2-3 mm in size.Elemental sulfur served as the reductant (electron donor) in the denitrification process, while limestone regulated the reactor's pH balance.The reactor was continuously fed with water, and wastewater was pumped into the device through a peristaltic pump from the bottom of the reactor.After the reaction was complete, the effluent flowed out near the top of the reactor.The up-flow water intake created anaerobic conditions for microorganismal growth, which helped gases to the escape of gases and ensured interaction between the microorganisms and wastewater.The reactor was wrapped in tinfoil to avoid photosynthetic bacterial propagation.The inoculated sludge (MLSS) was obtained from the anaerobic section-activated sludge of a sewage treatment plant in Nanning, Guangxi, China, with a ∼3,000 mg MLSS/L concentration.The sludge inoculation ratio was 20%.

. Artificial configuration of synthetic wastewater
The wastewater contained synthetic NO 3 -N derived from stepwise addition of 30 ± 2 mg/L of NaNO 3 to the influent.Additionally, 210 ± 2 mg/L of HCO − 3 was added to the influent to provide an inorganic source of carbon.When the SAD reactor operated stably, the volumetric load of nitrogen was controlled so that it remained unchanged, and the salinity of the wastewater was gradually increased from 0 to 25% (Tan et al., 2019;Zhang et al., 2021).The SAD reaction was maintained continuously throughout the experimental process.The hydraulic retention time (HRT) of the reactor was 4 h.Micronutrients were added to maintain microbial growth (1 mL/L), and the nutrient fluid (1 L) contained: 0.12 g ZnSO 4 -7H 2 O, 0.12 g MnCl 2 -4H 2 O, 0.15 g H 3 BO 3 , 0.03 g CuSO 4 -5H 2 O, 0.15 g CoCl 2 -6H 2 O, 0.06 g Na 2 MoO 4 -2H 2 O, and 10 g EDTA (Chauhan et al., 2022).The pH was adjusted to 7.2-7.3using sodium hydroxide and hydrochloric acid.All dosages were configured to the required concentration.To prevent temperature changes from affecting the system's denitrification, the reactor operated stably at room temperature (25 ± 1 • C).

. Water quality determination
The reactor effluent was collected every 24 h, and water samples were stored at −20 • C.After centrifugation, the supernatant of the water samples were taken for water quality analysis based on the guidelines provided by Eaton (2002) to determine the following water quality indicators.TN concentrations were determined using the alkaline potassium persulfate digestion-UV spectro-photometric method; NO − 3 -N using the hydrochloric acid photometry method; nitrite nitrogen (NO − 2 -N) using the N-(1naphthalene)-dia-minoethane spectrophotometry method; ammonium nitrogen (NH + 4 -N) using the indophenol blue method; and sulfate (SO 2− 4 ) using the barium chromate photometric method; the inlet and outlet water pH was measured with a benchtop pH meter (S220 Micro, Mettler Toledo, Columbus, OH, USA); and salinity was measured using a salinometer (SSM-500, China).
The NO − 3 -N and TN removal efficiencies were calculated using the Equation 1: where R V is the removal efficiency (%), and S 1 and S 2 are the initial and final concentrations, respectively.Microsoft Excel 2020 and Origin 2021 were used for statistical and graphical analyses. .

EPS extraction and analysis
Biofilm-attached packing samples were collected from the reactors on days 15, 30, 45, 60, and 75 to determine the extracellular polymer (EPS).Biofilms were collected using a previously described sampling method (Hao et al., 2022).EPS was extracted from the biofilm using a heating method (Li and Yang, 2007).The amount was calculated as follows: EPS = protein (PN) content + polysaccharide (PS) content.The PS and PN contents were determined using the phenol-sulfuric acid and Lowry methods, respectively (Shi et al., 2020).

Results and discussion
.

Startup of the SAD system
At the beginning of the reactor start-up, the effluent concentrations of NO − 3 -N decreased gradually.After 20 days of reactor operation, the NO − 3 -N removal efficiency increased to 90%, and no NO − 2 -N was detected (Figure 2).These results indicated that the reactor had successfully started (Shen et al., 2023;Zhu et al., 2023).Previous research has demonstrated that SO 2− 4 production confirms the presence and degree of SAD (Chen et al., 2018;Li et al., 2022).According to Equation (2), for every 1 mg/L of NO − 3 -N consumed, 7.83 mg/L of SO 2− 4 is produced (Capua et al., 2019).Using the calculations, the theoretical SO 2− 4 yield should be 214 mg/L when 27.3 mg/L of NO − 3 -N is removed in the SAD system.The observed effluent concentration of SO 2− 4 was 219 mg/L (Figure 3A).This result demonstrated that NO − 3 -N removal in the system was mainly via sulfur autotrophic denitrification.Owing to the addition of limestone to neutralize the H + produced in the SAD process (Guo et al., 2022), the pH value of the effluent decreased slightly but stabilized at approximately pH 6.7 (Figure 3B).The reactions were as follows.
After the reactor was stabilized, the NO − 3 -N removal efficiency remained >90%, corresponding to a TN removal efficiency of 90.5%.This left almost undetectable NO − 2 -N and NH + 4 -N in the effluent (Figure 4).This illustrated that the microorganisms in the SAD system could efficiently remove NO − 3 -N through to almost complete denitrification (Wang S. et al., 2021;Guo et al., 2022). .

E ect of salinity on nitrogen and sulfur transformations in SAD system
As the salinity increased from 0 to 6 g/L, the NO − 3 -N conversion capability increased continuously (Figure 4).The NO − 3 -N and TN removal efficiencies increased to 95.5 and 95.3%, respectively.Moreover, the effluent NO − 2 -N and NH + 4 -N concentrations also remained undetectable.This phenomenon indicated that the low salinity (2 and 6 g/L) stimulation could enhance NO − 3 -N removal.Ding et al. (2023) also reported that 7 g/L salinity could increase the NO − 3 -N removal efficiency (from 54.1 ± 12.5% to 74.4 ± 11.53%) in wetlands.This might be because appropriate amounts of salt can promote the enzymatic reactions of the microorganism, maintain their membrane balance, and regulate their cellular osmotic pressure (Bassin et al., 2017).The effluent pH was maintained at ∼6.6-6.8, and the effluent SO 2− 4 concentration increased gradually as the NO − 3 -N removal efficiency increased (Figure 3).When the salinity increased to 15 g/L, the concentration of NO − 3 -N in the effluent of the system began to increase gradually and stabilized at 5.5 mg/L (Figure 4).The NO − 3 -N removal efficiency initially declined to 66.8% and then increased to 81.7% after the reactor had been running for 15 days.The effluent SO 2− 4 concentration dropped to 150 mg/L, then increased to ∼175 mg/L (Figure 3A).The effluent pH maintained at ∼6.6-6.8 (Figure 3B).Therefore, a salinity of 15 g/L would not have an inhibitory effect on functional microorganisms in the SAD system.The reactor would recover once these microorganisms adapted to the high-salt environment (Zhu et al., 2023).The effluent NO − 2 -N concentration increased to 10 mg/L then decreased and remained at 4 mg/L during this process (Figure 4).Similarly, Shen et al. (2023) found that gradually increasing salinity to 1.5% promoted NO − 2 -N accumulation in the SAD system, probably because the nitrite reductase was sensitive to the increase in salinity.When the salinity continued to increase to 25 g/L, the concentration of NO − 3 -N in the effluent continually increased to ∼20 mg/L.The corresponding NO − 3 -N and TN removal efficiencies decreased to 33.6 and 28.3%, respectively (Figure 4).The effluent SO 2− 4 concentration dropped to 53 mg/L (Figure 3A).This indicated that a salinity of 25 g/L would significantly inhibit the function of microorganisms in the SAD system.The NO − 2 -N concentration in the effluent dropped to 0.1 mg/L, and the NH + 4 -N concentration increased to 1 mg/L (Figure 4).This might be because the high salinity stress stimulates dissimilatory nitrate reduction to ammonium (DNRA) (Giblin et al., 2010).The effluent pH therefore increased to 6.8-7.0 (Figure 3B).

. Microbial community changes under di erent salinities
To further explore the influence of salinity on the performance of the SAD system, SEM and microbial community analysis were performed on the filler samples collected at the reactor startup (0 g/L salinity), 2, 6, 15, and 25 g/L salinity levels, respectively.
Figures 5a-e shows the structures and microbial morphologies of the filler samples at the different salinities.At 0 g/L salinity, short bacilli and cocci dominated the bacterial morphology (Figure 5a), and their size and morphology were consistent with those of thiobacillus azotizans, reported in an earlier study (Guo et al., 2022).When the salinity increased to 2 and 6 g/L, the numbers of short bacilli and spherical bacteria on the filler increased significantly and were attached to the entire surface of the filler (Figures 5b, c).This result indicated that low salinity could stimulate the growth of microorganisms in the system, and promote the removal efficiency of NO − 3 -N (Figure 4).However, at salinities of 15 and 25 g/L, the number of microorganisms attached to the filler decreased significantly and were dominated by short bacilli (Figures 5d, e).Similarly, the Chao index increased in low-salinity wastewater (2 and 6 g/L), but significantly decreased in high-salinity wastewater (15 and 25 g/L) compared with non-saline water (Table 1).This indicates that appropriate salinity increases the bacterial community richness in sewage.In contrast, high salinity can kill salt-intolerant microorganisms, thereby reducing community richness (Zhu et al., 2023).The Shannon index decreased and the Simpson index increased as the salinity increased (Table 1).This indicated that some microorganisms in the SAD system could not adapt to the high-salinity environment; however, some salt-tolerant microorganisms continued to increase (He et al., 2020).
The dominant phyla in the salt-free SAD system were Proteobacteria (58.2%),Campilobacterota (20.0%),Bacteroidota Frontiers in Microbiology frontiersin.org(10.1%), and Verrucomicrobiota (6.5%) (Figure 6).The relative abundances of Bacteroidota and Verrucomicrobiota decreased significantly with the increase in salinity.However, the abundances of Proteobacteria increased in 0-15 g/L salinity, and then decreased at 25 g/L salinity, indicating initial significant stimulation and subsequent inhibition owing to salt stress (Ding et al., 2023;Shen et al., 2023).In contrast, the abundances of Campilobacterota decreased gradually with the increasing salinity but increased at 25 g/L salinity, indicating that some of the microorganisms in this phyla are salt-tolerant, and they adapt to salinity stress over time (Ohore et al., 2022).At the genus level, the SOB-related genera Thiobacillus (21.1%),Sulfurimonas (16.4%), and Thiomonas (10.5%), which are commonly present in the SAD system (Blázquez et al., 2019;Polizzi et al., 2022;Yuan et al., 2022;Shen et al., 2023), were detected (Figure 7).Thiobacillus was dominant because the SAD system was driven by sulfur (S 0 ) (Polizzi et al., 2022), and its abundance increased significantly from 21.1 to 71.1% when the salinity increased from 0 to 15 g/L.However, high salinity (25 g/L) inhibited the growth of Thiobacillus with a slight reduction in its abundance.Sulfurimonas could barely grow efficiently with sufficient S 2− (Polizzi et al., 2022), and was therefore less abundant than the genus Thiobacillus in the S 0driven SAD system.Moreover, the abundance of Sulfurimonas decreased gradually with increasing salinity but increased at 25 g/L salinity, which was similar to the variation trend of the abundance of Proteobacteria (Figure 6).Sulfurimonas belongs to Proteobacteria, and is widely distributed in global deep-sea hydrothermal environments (Wang S. S. et al., 2021).Therefore, the genus Sulfurimonas in the phylum Proteobacteria might be a salt-tolerance microorganism in the SAD system.Thiomonas was highly abundant in 0 and 15 g/L salinity (10.5 and 10.2%, respectively) but decreased to 0.81% in 25 g/L salinity (Figure 7), indicating that it was intolerant to high-salinity environments.
In addition, the abundance of other autotrophic denitrifying bacteria, such as Ferritropicum and Cloacibacterium, had the ability to remove NO − 3 -N from the system (Li Y. et al., 2020;Chauhan et al., 2022) also changed with the increase of salinity.The abundance of Ferritropicum increased from 0.31 to 4.47% when the salinity increased from 0-6 g/L, but decreased when the salinity was ≥15 g/L.Song et al. (2020) also found that low salinity (20‰) can promote the growth of Ferritropicum.Furthermore, the abundance of Cloacibacterium remained unchanged when the salinity increased from 0 to 6 g/L but decreased significantly when the salinity reached 15 and 25 g/L.Moreover, a few heterotrophic bacteria unclassified_f_Rhodocyclaceae (4.53%), unclassified_f_Xanthomonadaceae (5.01%), Cloacibacterium (3.56%), and Simplicispira (0.31%) were detected in the SAD system (Figure 7), and have been identified as denitrifying bacteria in prior studies (Fitzgerald et al., 2015;Si et al., 2021;Liang et al., 2022;Zhang et al., 2023).The abundance of these three genera in the SAD system when the salinity was ≤6 g/L was higher than that when the salinity was ≥15 g/L.These results indicated that the heterotrophic bacteria in the SAD system were salt-intolerant.Unclassified_f_Rhodocyclaceae and unclassified_f_Xanthomonadaceae are dominant bacterial species that use NO − 2 -N as an electron acceptor and can be used to reduce NO − 2 -N, with a higher affinity for NO − 2 -N (Hou et al., 2022).When the salinity exceeds 15 g/L, the significant decrease in their abundance may be the reason for the inability to reduce NO − 2 -N, which resulted in a large NO − 2 -N accumulation (Figure 4).Low salt conditions stimulated the growth of some autotrophic and heterotrophic bacteria, resulting in the coexistence of these bacteria in the system.The co-participation of autotrophic and heterotrophic bacteria at low salinity (≤6 g/L) promotes the NO − 3 -N removal efficiency.When the salinity was ≥15 g/L, Thiobacillus and Sulfurimonas that closely related to the DNRA process (Robertson et al., 2016;Lai et al., 2020) occupied a dominant position in SAD (Figure 4).Usually, a correlation between the ammonium flux and salinity  strengthens with increasing salinity, but high salinity leads to an inhibition of the denitrification process and enhancement of the DNRA process (Giblin et al., 2010).In this study, high salinity (25 g/L) inhibited denitrification in the SAD system, but Thiobacillus and Sulfurimonas could reduce NO − 3 -N to NH + 4 -N via the DNRA pathway, resulting in the accumulation of NH + 4 -N (Figure 4).Microbial community changes at phylum level in S -driven SAD system at di erent salinity. .

E ects of salinity on the EPS content and composition
EPS is a self-protective layer of microorganisms, helping to maintain bacterial growth and reactor stability when environmental pressure changes (Han et al., 2020).In this study, we found that Thiobacillus and Sulfurimonas were the main salt-tolerant microorganisms in the SAD system.Therefore, the EPS was further analyzed to explore why these two genera could tolerate high salinity.
The results showed that the EPS content gradually increased with increasing salinity (Figure 8), consistent with the phenomenon observed by Song et al. (2016).PN is usually considered a hydrophobic substance crucial for resisting external osmotic pressures and improving sludge performance (Han et al., 2018).Previous studies found that microorganisms in the system secrete more PN under salinity stress (Lin et al., 2022;Sun et al., 2022).This is because the amino and carboxyl groups in PN increase the hydrophobicity of the cell surface, which helps to maintain the structure and stability of the biofilm (Wang et al., 2013;He et al., 2019).In this study, the PN content increased from 35.5 to 79.3 mg/g SS, and the PS content from 18.6 to 30.8 mg/g SS with the contentious increase in salinity (Figure 8).Furthermore, the PN/PS ratio consistency increased from 1.91 in non-saline to 2.57 at 25.0 g/L salinity conditions, suggesting that the PN content exceeded the PS content.Therefore, possibly the dominant strains in the SAD system resisted the high salinity stress by releasing PN.

Conclusions
Salinities of 2 and 6 g/L promoted the NO − 3 -N removal efficiency of the elemental sulfur-driven SAD system, which increased from 91.1 to 94.8% and 95.5%, respectively.However, when the salinity increased to 15 and 25 g/L, the NO − 3 -N removal efficiency decreased significantly to 81.7 and 33.7%, respectively.This was accompanied by accumulation of nitrite and ammonium.Thiobacillus and Sulfurimonas play important roles when the system is in a high salt environment and they release EPS, particularly PN, to resist the salinity.

FIGURE
FIGURESchematic diagram of the S -driven SAD reactor.

FIGURE
FIGUREStartup of S -driven autotrophic denitrification reactor.

FIGURE
FIGUREVariation of sulfate concentration (A), pH and salinity (B) in S -driven SAD system at di erent salinity.

FIGUREE
FIGUREE ect of salinity on nitrogen transformation in S -driven SAD system.

FIGURE
FIGURE Scanning electron microscope (SEM) images of the filter samples.Salinity of (a) g/L, (b) g/L (c) g/L (d) g/L, and (e) g/L.
TABLE Bacterial community diversity and richness of filter samples.