Effect of different carbon sources on sulfate reduction and microbial community structure in bioelectrochemical systems

Microbial electrolysis cells (MECs) have rapidly developed into a promising technology to treat sulfate-rich wastewater that lacks electron donors. Hence, a better understanding of the effect on the microbial community structure caused by different sources in bioelectrochemical systems is required. This study sought to investigate the effect of different carbon sources (NaHCO3, ethanol, and acetate were employed as sole carbon source respectively) on the performance of sulfate-reducing biocathodes. The sulfate reduction efficiency enhanced by the bioelectrochemical systems was 8.09 − 11.57% higher than that of open-circuit reference experiments. Furthermore, the optimum carbon source was ethanol with a maximum sulfate reduction rate of 170 mg L−1 d−1 in the bioelectrochemical systems. The different carbon sources induced significant differences in sulfate reduction efficiency as demonstrated by the application of a micro-electrical field. Microbial community structure and network analysis revealed that all three kinds of carbon source systems enriched large proportions of sulfate-reducing bacteria and electroactive bacteria but were significantly distinct in composition. The dominant sulfate-reducing bacteria that use NaHCO3 and acetate as carbon sources were Desulfobacter and Desulfobulbus, whereas those that use ethanol as carbon source were Desulfomicrobium and Desulfovibrio. Our results suggest that ethanol is a more suitable carbon source for sulfate reduction in bioelectrochemical systems.


Introduction
Sulfate is widely present in natural aquatic systems and is associated with several human activities (Chen et al. 2020a, b). However, excessive discharge of untreated sulfate-contaminated wastewater can severely affect public water supplies and human health (Xu et al. 2017). Many approaches have been developed to eliminate or reduce the adverse effects of sulfate wastewater, including ion exchange, electrochemical, metal reduction, and biological methods (Liamleam and Annachhatre 2007;Sarti and Zaiat 2011). Among them, biological methods for sulfate removal have garnered widespread interest due to their high efficiency and low operating costs. The biological method using sulfatereducing bacteria (SRB) converts sulfate to sulfide under anaerobic conditions (Chen et al. 2020a, b;Hao et al. 2014). The theoretical chemical oxygen demand (COD) to sulfate ratio for the sulfate reduction process is 0.67, but the COD to sulfate ratio is often higher than this value to achieve complete sulfate reduction (Liamleam and Annachhatre 2007).
Researchers have discovered that different carbon sources can have varying effects on the migration and transformation of sulfur in constructed wetlands (Guo et al. 2020). A previous study reported that the stability of sulfate reduction and its reduction products are inhibited when carbon sources are restricted (Wiessner et al. 2010). Studies have shown that the sulfate-reducing activity of bioreactors can decrease depending on the carbon source used, including organic acids, volatile fatty acid salts, alcohols, and carbohydrates (Cao et al. 2012). Xia et al. reported that Desulfosporosinus was the dominant genus in a sulfate reduction system in which sulfate was the sole electron acceptor, and citrate and lactate were used as carbon sources (Xia et al. 2019). Some researchers have reported that Geobacter, a representative electron-donating bacterium, was domesticated in ethanol-fed anaerobic sludge and participated in the sulfur cycle (Lin et al. 2017). Moreover, to align with the principles of sustainability and reducing operational costs, many researchers have investigated the feasibility of using organic wastes as alternative substances. For example, some researchers have used crab shells, potatoes, and filter paper as carbon sources, achieving sulfate reduction efficiencies of up to 32%, 82%, and 98%, respectively (Reyes-Alvarado et al. 2017). SRB can degrade some organic wastes such as sugarcane vinasse (Nogueira et al. 2019), plant hydrolysates (Lakaniemi et al. 2010), horse manure, vegetable compost (Castillo et al. 2012) into a mixture of ketones, volatile fatty acids (VFA), alcohols, and other low-molecular-weight compounds in reducing sulfate. However, there is often a lack of sufficient electron donors in actual sulfate-rich wastewater, and therefore, external carbon sources are commonly added, which significantly increases the operation costs and limits the application of biological methods (Hurtado et al. 2018;Sampaio et al. 2019).
Microbial electrolysis cells (MECs) are a promising technology that integrates electrochemistry with microbiological catalysis and has a wide range of promising applications in wastewater treatment (Blázquez et al. 2017. Microorganisms in MEC can accept electrons to reduce chemical materials. SRB have been proposed as a potential source for biocathode enrichment, and previous studies have confirmed that SRB can directly accept electrons on the cathode surface for sulfate reduction . For example, the cytochrome c of Desulfovibrio desulfuricans and G. sulfurreducens has been reported to efficiently attach to the electrode and participates in direct electron transfer (Eaktasang et al. 2016). Specifically, bioelectrochemistry facilitates sulfate reduction by providing continuous electrons and a favorable redox environment (Wang et al. 2017a). More recently, many studies have demonstrated the feasibility of this approach. With the dominance of Desulfovibrionaceae in the community of the biocathode of a dual-chamber bioelectrochemical system (BES), 94.2% sulfate was removed within 12 days with a cathodic potential of − 0.8 V (vs. standard hydrogen electrode (SHE)) (Xiang et al. 2017). The synergistic effect of neutral red and graphite felt was reported to increase the bioelectrocatalytic activity of microbial electrolysis, thus maximizing sulfate reduction efficiency (Wang et al. 2017a). Moreover, a recent study reported an SRB-biocathode that also achieved effective sulfate removal when coupled with elemental sulfur in bioelectrochemical reactors (BERs) (Blázquez et al. 2016). Above all, these studies on sulfate reduction in MECs mainly focused on electrochemical parameters including potential regulation, electrode materials, and applied current (Gao et al. 2022;Hu et al. 2019;Wang et al. 2017b).
More importantly, electrodes are essential in bioelectrochemistry because they serve as key electron donors and can partially offset the consumption of carbon sources (Wang et al. 2017b). The performance of biocathode is largely determined by the metabolic activity of biofilms, and the types of carbon sources can have different effects on the biocathode by directly influencing the microbial activity and regulating the energy available for bacterial growth (Hu et al. 2019). Additionally, variations in the molecular weight and chemical structure of carbon sources may significantly influence their utilization mechanism by key microorganism (Xu et al. 2018a, b;Zhou et al. 2020). The addition of easily degradable organic matter can enhance the capacity of electrochemically active bacteria to degrade a wide range of pollutants in bioelectrochemistry systems. A previous study revealed that three systems enriched with large amounts of dechlorinators and cathode-respiring bacteria exhibited distinct bacterial compositions due to difference in carbon sources . Some studies have confirmed that the reduction rate of nitrobenzene and the formation rate of aniline in a bicarbonate biological cathode were significantly lower than that of a glucose biological cathode (Liang et al. 2014). Therefore, different carbon sources can affect the microbial community structure in electrical stimulation, resulting in various in pollutant removal efficiency. Previous studies have investigated the effect of various carbon sources on the performance and microbial community structure of sulfate reduction systems. However, very few studies have characterized the changes in the bacterial structure of bioelectrochemical systems in response to different external carbon sources, including sulfate reduction efficiencies, characteristics, and the mechanisms underlying the variations in microbial community structure.
This study aimed to assess which types of carbon sources could better synergize with MEC to obtain higher sulfate reduction rate. The objectives of the present study were to determine the impacts on sulfate-reducing laws in bioelectrochemical systems as well as microbial community structure and function corresponding to different carbon sources through comprehensive microbial analysis. This study put forward a new perspective for efficient decontamination of sulfate-rich wastewater deficient in electron donors.

SRB enrichment
Sludge collected from the East Sewerage Plant in Yancheng was used as inoculums to SRB enrichment. The sludge was acclimatized in 100-mL anaerobic bottles filled with 50 mL sulfate medium. The basal medium contained 0.74 g/ L Na 2 SO 4 (approximately 500 mg/L of SO 4 2− ), 0.74 g/L acetate, 0.3 g/L NH 4 Cl, 0.04 g/L MgCl 2 , 6.84 g/L NaH 2 PO 4 ·2H 2 O, 0.1 g/L KCl, 2.2 g/L Na 2 HPO 4 ·12H 2 O, and 1 ml/L trace element solution (Table S1). The basal medium was aerated with nitrogen gas for 20 min before adding to the anaerobic bottles to maintain anaerobic conditions. Then batch-mode acclimation was carried out on 1 week cycle and half of the supernatant was replaced with fresh basal medium every week. Samples were taken every 24 h to analyze sulfate concentrations. When the sulfate reductive rate in the anaerobic bottles became stable, the acclimatized sludge was employed for inoculation of the biocathode in the MEC for further experiments.

MEC setup and operation
The two-chamber reactor was made of plexiglass with a sufficient volume of 150 mL for each chamber. The two compartments were separated by proton exchange membranes (Nafion®N-115 membrane, 0.127 mm thick). The silicone ring and silicone gasket were placed at the interface for leakage prevention treatment. The anode (counter electrode) electrode was made out of a 4.0 cm × 5.0 cm × 0.1 cm piece of carbon felt (Haote New Material Co. Ltd., Jingzhou, China), which was placed vertically in the middle of the anode chamber (Fig. S1). The cathode electrode (working electrode) was made of a 3 cm diameter graphite brushes (3 cm × 12 cm) (PANEX33 160 K, ZOLTEK) coupled with a 0.2 mm diameter titanium wire. Before the experiments, the graphite brushes were thermally treated at 450 °C for 30 min to enhance the biomass adhesion. The carbon felt was first immersed in 1 mol/L HCl for 24 h, and then put into deionized water for 24 h, and calcined in muffle furnace at 450 °C for 30 min prior to application (Chen et al. 2022). The top of the reactor was sealed with butyl stoppers and aluminum seals and connected to a 100 mL gas bag with nitrogen gas to maintain anaerobic conditions. An Ag/AgCl reference electrode was used as a reference electrode (RE-1B, BAS Inc., 197 mV vs. SHE). Potential control and data acquisition were conducted with a multichannel potentiostat (CHI 1000C, Chenhua Instrument, China). Unless otherwise stated, all the potentials were in reference to those of the Ag/AgCl electrode.
The anode chamber was filled up with 100 mmol/L potassium ferrocyanide. The main function of potassium ferricyanide in the anolyte is to provide electrons to the cathode through the external circuit. The composition of the basal cathodic medium was as follows: 0.74 g/L Na 2 SO 4 , 0.3 g/L NH 4 Cl, 6.84 g/L NaH 2 PO 4 ·2H 2 O, 0.04 g/L MgCl 2 , 2.20 g/L Na 2 HPO 4 ·12H 2 O, 0.1 g/L KCl, 1 ml/L trace element solution, and 1 mol/L NaOH was used to adjust the initial pH to 7.0. The cathode medium was purged using nitrogen gas for 20 min to eliminate any traces of dissolved oxygen, after which it was added to the reactor. Three BERs were operated in parallel and maintained at − 0.5 V cathodic potential, including a closedcircuit BER (CCB) with added NaHCO 3 (CCB-NaHCO 3 ), a CCB with the addition of ethanol (CCB-Ethanol), and a CCB with the addition of acetate (CCB-NaAc). Three control experiments were also performed: an open-circuit BER (OCB) with the addition of NaHCO 3 (OCB-NaHCO 3 ), an OCB with the addition of ethanol (OCB-Ethanol), and an OCB with added acetate (OCB-NaAc) ( Table S2). The amounts of NaHCO 3 , ethanol, and acetate added were 2 g/L, 0.74 ml/L, and 1.724 g/L. The ethanol and acetate concentrations were selected to achieve a COD to sulfate in the reactor was 2 to ensure sufficient electron donors for sulfate reduction. A magnetic stirrer was used to obtain homogeneous and stable sample solutions in each cathode chamber. All experiments were conducted at laboratory temperature (25 ± 3 °C). Ten batches of experiments were conducted under each carbon source condition. Representative steady-state data for plotting when the data and trends were similar in the last three or four batches.

Selection criteria analytical methods
The sample was filtered through a 0.22 mm syringe filter (Millipore, USA) for subsequent index analysis. The sulfate concentrations of the filtered samples were analyzed using a Dionex ICS-1100 ion chromatograph with a Dionex AS-DV Autosampler, an IonPac AS19 column and an IonPac AG19 pre-column (Thermo-Scientific, USA). The suppressor operating current was 57 mA. The pH and oxidation-reduction potential (ORP) were determined with a pH meter (PHS-3C, Shanghai Precise. Sci. Instru. Co., Ltd., China) equipped with an ORP and pH composite electrode (model-501, Shanghai Precise. Sci. Instru. Co., Ltd., China). The concentration of sulfide ions (S 2− ) was analyzed via the methylene blue method at 664 nm and the sample was measured immediately to minimize oxidation and sulfide stripping.

Microbial community analysis
The biofilm samples in the six reactors were scraped with sterile scissors for high-throughput sequencing to investigate the effects of different carbon sources on microbial communities. The steps for DNA extraction, PCR amplification, and 16S rRNA gene sequencing are outlined in the Supplemental Materials. The correlation between microbial communities and carbon sources was assessed by principal component analysis (PCA) using the Canoco (version 4.5) software (Chen et al. 2019a). The constructed networks were evaluated with the Cytoscape (version 3.7.1) software (Chen et al. 2019b). Network-based visualization was carried out to investigate the interrelationships between the generality and mutability of all operational units (OTUs) and visualize the relationships between groups and populations.

Sulfate reduction performance under different carbon sources in MECs
Cathodic current and sulfate reduction efficiency under different carbon sources Figure 1 illustrates the cathodic current variations in the CCB-NaHCO 3 , CCB-ethanol, and CCB-NaAc biocathode reactors after 45 days of acclimation. The current of the three biocathode reactors gradually decrease and then tended to stabilize from 12 to 48 h. This decrease in current may be due to the consumption of sulfate and carbon sources at the cathode chamber, after which the stabilization of the current indicated that the sulfate-reducing biocathode systems had been successfully established (Teng et al. 2016). Ethanol was reported as the most favorable carbon source for SRB in an anaerobic reactor due to its direct interspecies electron transfer capacity. Particularly, the cathodic current of the CCB-ethanol was obviously higher than that of the CCB-NaHCO 3 and CCB-NaAc during the reaction cycle. Previous researchers demonstrated that some typical electron-donating bacterium, was acclimated in ethanol-fed anaerobic sludge and capable of direct interspecies electron transfer (Rotaru et al. 2014;Lin et al. 2017). These findings indicated that CCB-ethanol could improve the efficiency of the electron transfer between bacteria and electrodes, which in turn may enhance sulfate reduction efficiency. Figure 2a illustrates the changes in sulfate reduction capacity under different carbon sources in the potentiostatic experiments (− 0.5 V) and the open-circuit reference experiments were compared within 48 h of operation. The accumulated sulfate reduction changing in CCB-NaHCO 3 , CCB-ethanol, CCB-NaAc, OCB-NaHCO 3 , OCB-ethanol, and OCB-NaAc were 98 ± 10.5, 340 ± 16.3, 179.6 ± 14.5, 47.3 ± 12.9, 246 ± 32.6, and 77.135 ± 22.7 mg/L, respectively. Our findings indicated that sulfate reduction mainly occurred in the first 24 h, after which it remained largely constant in the later period of the cycle. This may have been caused by the depletion of potassium ferrocyanide in the anode chamber, which presumably prevented the biocathode from accepting enough electrons for the sulfate reduction. Previous study reported that sulfate reduction mainly occurred in the first 30 h; it was speculated that the cathode solution could not accept enough electrons for sulfate reduction due to the consumption of potassium ferrocyanide in the anode solution ). These results indicated that the electron supplied by anolyte affects the sulfate reduction efficiency in the cathode chamber. Additionally, the accumulation of hydrogen sulfide during the reduction process may be toxic to SRB, which inhibited sulfate reduction from 24 to 48 h (Luo et al. 2014;Teng et al. 2016). The sulfate reduction efficiency and the average sulfate removal rate in the six reactors are shown in Fig. 2b. The sulfate reduction efficiency in CCB-NaHCO 3 , CCB-ethanol, CCB-NaAc, OCB-NaHCO 3 , OCB-ethanol, and OCB-NaAc reached 19.6 ± 2.1%, 74.71 ± 4.8%, 38.57 ± 3.6%, 9.04 ± 1.5%, 66.62 ± 6.1%, and 26.97 ± 2.4%. The sulfate reduction efficiencies in CCB-NaHCO 3 , CCB-ethanol, and CCB-NaAc were significantly higher than those in OCB-NaHCO 3 , OCB-ethanol, and OCB-NaAc by 10.05%, 8.09%, and 11.6%. Recent many studies have suggested that biological sulfate reduction can also be driven by electricity as the sole electron source by using bioelectrochemical systems, but the process may not be thermodynamically spontaneous; an additional voltage was required (Blázquez et al. 2016). This result suggested that stimulating the microbial In the biological cathode systems with three different carbon sources, the reactors with NaHCO 3 as a carbon source exhibited an appreciable sulfate reduction ability, albeit lower than the other two organic carbon sources. These results indicated that an electrode and inorganic NaHCO 3 could be respectively used as the electron donor and carbon source for sulfate reduction, which is consistent with the result of previous studies (Blázquez et al. 2019, Hu et al. 2018. The reduction efficiency of the NaHCO 3 system was relatively lower than that of the hydrogen-initiated sulfate reduction system. The sulfate reduction rate in CCB-ethanol was significantly higher than those in the CCB-NaHCO 3 , and CCB-NaAc by 71.23% and 47.26% (Fig. 2b), respectively, demonstrating that ethanol might be a more favorable carbon source for sulfate reduction. Similarly, previous studies reported that the sulfate reduction efficiency of an ethanol-fed reactor was higher than that of its acetate-fed counterpart (Xing et al. 2020;Yildiz et al. 2019). Additionally, the incomplete oxidation of ethanol (ΔG 0 = − 66.4 kJ/ mol) provides more energy than the complete oxidation of acetate (ΔG 0 = − 47.6 kJ/mol) according to the Gibbs freeenergy calculation, demonstrating that using ethanol as a carbon source could enhance sulfate reduction efficiency (Thauer et al. 1977). Moreover, the products of ethanol metabolism, such as acetic acid and propionic acid, can be used as electron donors to further promote sulfate reduction. Figure 2c illustrates the concentration of sulfide ions (S 2− ) in the six reactors, and the maximum concentration of 54.75 ± 3.6 mg L −1 was observed in CCB-ethanol. The S 2− concentration trend was consistent with that of sulfate in the cathode chamber and exhibited the following order: CCB-ethanol > CCB-NaAc > CCB-NaHCO 3 . Some reports have confirmed that some of the S 2− may react with metal ions in the catholyte to form precipitates at elevated pH conditions, suggesting that the biocathode MEC can be utilized to treat wastewater containing both sulfate and heavy metals (Pozo et al. 2017;Zhang et al. 2018). Collectively, among the three carbon sources, ethanol was found to be the best as carbon source to promote sulfate reduction. Figure 3a illustrates the ORP of different carbon sources in both closed-circuit and open-circuit BERs. Some studies have pointed out that ORP of − 100 mV is the key factor for SRB growth and we found that ORP of all reactors could be kept below − 200 mV (Hwang and Jho 2018). Compared with several carbon sources, the reactors with organic carbon sources showed more negative ORP value, which benefited to the growth of anaerobic microorganisms to promote sulfate reduction. Furthermore, ORP exhibited a slight increase trend after 24 h, and a previous study reported that the competition of microorganisms for carbon sources becomes fierce with the increase of ORP (Wang et al. 2017b). This may lead to lower sulfate reduction efficiency during the later stage of the reaction. More importantly, these results also showed that the ORP of the closed-circuit BERs was lower than that of the open-circuit BERs, which could maintain a more favorable redox environment for SRB to promote sulfate reduction under weak electrical stimulation. Biocathode systems can manipulate the redox potential to create a favorable environment for reduction. The applied voltage can regulate the redox potential and impact the overall environment (Srivastava et al. 2021). It has been demonstrated that electrical stimulation can maintain the ORP environment of SRB systems, thereby enhancing the competitiveness of SRB in microbial communities (Wang et al. 2017a, b). Therefore, we suggested that the applied potential (− 0.5 V vs. Ag/AgCl) might affect the redox environment around the electrode and render the ORP more negative.

ORP and pH under different carbon sources
With different carbon sources, the pH values of the reactors ranged from 6.8 to 7.8 (Fig. 3b). The pH value of the closed-circuit BERs was generally higher than that of the open-circuit BERs due to proton consumption and production of OH − , which contributed to the enhancement of SRB biofilm activity to enhance sulfate reduction (Liang et al. 2013). The results showed that the use of bioelectrochemical treatment of wastewater containing sulfate can maintain the neutral pH of cathode liquid, which was more conducive to the long-term operation of the system and had greater significance for engineering application (Shi et al. 2022). The lowest pH value of OCB-ethanol may be due to the accumulation of acetic acid and propionic acid in ethanol metabolism. Moreover, the pH condition had a significant effect on the composition of sulfate reduction products. Weak alkaline conditions can reduce the toxic reduction product hydrogen sulfide, and a recent study investigated the feasibility of mitigating sulfide inhibition and increasing CH 4 production by utilizing an anaerobic reactor with built-in MECs (Yuan et al. 2020). These results indicated that electrical stimulation can create weak alkaline conditions to alleviate acidification in the reactor with ethanol as carbon source, which is beneficial to improve the sulfate reduction efficiency.

Microbial community succession under different carbon sources
The biological samples from the six reactors were collected at the end of the experiment. The V3-V4 region of the 16S rRNA was examined via Illumina high-throughput sequencing technology to qualify the dominant strains and the changes in the structure of the microbial community. The α diversity index was used to examine the biodiversity of the microbial community, and the results were summarized in Table S3. We found that the addition of different carbon sources and the diversity of the microbial community varied. The NaHCO 3 -fed biocathode with the highest overalldiversity had the highest Shannon index (9.02) and lowest Simpson index (0.005) among the closed-circuit BERs. The reactor with NaAc addition showed the lowest Shannon index (6.00) and highest Simpson index (0.059) among the open-circuit BERs, as well as the lowest overall-diversity. The sequencing coverage was ≥ 0.99 indicating that the collected gene sequences were a good reflection of the bacterial OTUs in the six biofilm samples.
The microbial community abundances among the six biofilm samples were identified at the phylum level (Fig. 4a). The relative abundance of dominant bacteria varied significantly in response to the different carbon sources. Proteobacteria, Desulfobacterota, Bacteroidota, Firmicutes, and Actinobacteria were the most dominant bacteria. Similarly, the bacteria of Proteobacteria, Desulfobacterota, Bacteroidota, Firmicutes, and Actinobacteria have been reported to be the most commonly found in sulfate reduction bioreactors  (Tang et al. 2020;Xiang et al. 2017). Proteobacteria were the most dominant phylum of bacteria in CCB-NaHCO 3 and CCB-NaAc, including many electrochemically active bacteria that play a key role in the carbon and sulfur cycle of wastewater treatment . It is also worth noting that a significant variation (38.9%) in the relative abundance of Desulfobacterota was observed in the CCBethanol, which was higher than that in other reactors. Desulfobacterota can reportedly utilize sulfate, sulfite, thiosulfate, and elemental sulfur as electron acceptors to reduce sulfide under anoxic conditions (Murphy et al. 2021). This result suggests that the increase in the sulfate reduction rate may be related to the high abundance of aforementioned bacteria. Firmicutes accounted for 4.2%, 8.0%, and 6.2% of the bacterial community in the CCB-NaHCO 3 , CCBethanol, and CCB-NaAc, but only 3.8%, 2.1%, and 2.5% in OCB-NaHCO 3 , OCB-ethanol, and OCB-NaAc, suggesting that electronic stimulation may provide favorable anaerobic conditions for its growth, which is consistent with previous reports (Ailijiang et al. 2016). Additionally, the relative abundance of this bacteria in the reactor with organic carbon sources was significantly higher than that in the reactors with inorganic carbon source, and previous studies reported that Firmicutes can degrade organic acids and provide shortchain fatty acids or H 2 to other bacteria (Nascimento et al. 2018). Compared with open-circuit BERs, the relative abundance of Campylobacterota presented a significant decrease in closed-circuit BERs. These bacteria have previously been linked with the process of sulfur oxidation (Sun et al. 2021), suggesting that electrical stimulation can create a lower ORP for microorganisms, which is consistent with the hypothesis of ORP decline in closed-circuit BERs. These observations further confirmed the strong influence of carbon sources on microbial community structure, which results in various in sulfate reduction efficiency.
At the class level (Fig. 4b), the relative abundances of Acidimicrobiia, Anerolineae, Clostridia, and Desulfovibrionia were higher in the closed-circuit reactors, suggesting that certain species can be selectively enriched by electrical stimulation. Compared with organic carbon sources, Alphaproteobacteria was relatively enriched in CCB-NaHCO 3 (9.25%) and CCB-NaHCO 3 (10.28%), whereas it only accounted for 3. 36%,5.19%,4.34%,respectively. Previous studies have also demonstrated that Anaerolineae can act synergistically with hydrogenotrophic microorganisms to degrade pollutants (Xu et al. 2018a, b). Desulfovibrionia exhibited a high relative abundance of 25.3% and 10.4% in the CCB-ethanol and OCB-ethanol. It has been reported that this bacterial class is composed of sulfate reducers and can partially oxidize organic compounds to acetic acid. Moreover, Desulfovibrionia is known to utilize ethanol as a carbon source and can achieve a satisfactory sulfate reduction performance (Nagpal et al. 2000).
Further studied analysis of the microbial community at the genus level was exhibited in Fig. 4c. Some similarities of dominant bacteria were observed in the two in which NaHCO 3 was added as a carbon source. Particularly, the dominant bacteria exhibited similar proportions in CCB-NaHCO 3 and OCB-NaHCO 3 : Bacteroidetes vadinHA17norank (4.64%, 3.86%), > Desulfobulbus (3.10%, 2.81%), > Dechloromonas (2.51%, 2.48%), > Desulfobacter (2.37%, 2.42%). The high relative abundance of Desulfomicrobium, Desulfovibrio, Geobacteraceae, and Desulfobulbus accounted for 17.91%, 7.40%, 6.15%, and 3.59% in the CCBethanol. Desulfomicrobium, and Desulfovibrio, Desulfobulbus are the typical SRB that were widely found in sulfate wastewater (Nogueira et al. 2021). Among them, the relative abundance of Desulfovibrio was 8.73 times higher than in the CCB-ethanol (7.40%) than in the OCB-ethanol (0.76%), indicating that weak electrical stimulation significantly promoted the enrichment of Desulfovibrio. These results were consistent with previous studies in which Desulfovibrio was detected as an electroactive bacterium in bioelectrochemical systems that can use an electrode as the sole electron donor for sulfate reduction (Rago et al. 2015). Acinetobacter can utilize a wide range of organic compounds as carbon and energy sources. The high relative abundance of Acinetobacter in the two reactors in which NaAc was added accounted for 8.8%, 27.93% in CCB-NaAc and OCB-NaAc. These results demonstrated that the microbial structure and composition varied with different carbon sources, electrical stimulation can promote sulfate reduction under the same carbon source conditions, and ethanol was the best carbon source for sulfate reduction. Several sulfate-reducing bacteria including Desulfobacter, Desulfobulbus, Desulfomicrobium, and Desulfovibrio were detected in all microbial samples. Some electroactive genera including Desulfovibrio, Geobacteraceae, Acinetobacter, and Pseudomonas were also detected in all microbial samples. The relative abundance of Geobacteraceae accounted for 0.07%, 6.15%, and 0.03% in the CCB-NaHCO 3 , CCBethanol, and CCB-NaAc, which was higher than the proportions in OCB-NaHCO 3 , OCB-ethanol, and OCB-NaAc. Geobacteraceae are anaerobic chemo-organotrophic mesophiles that oxidize small organic acids in MECs fed with wastewater. Our findings indicated that the weak electrical Sulfate-reducing bacteria Electroactive genera Sulfate-reducing bacteria Electroactive genera stimulation promoted the selective enrichment of these genera. Desulfovibrio accounted for 7.40% and 0.24% of the bacterial communities of the CCB-ethanol and CCB-NaAc, which was higher than the values in OCB-ethanol and OCB-NaAc (6.64% and 0.09%, respectively). Desulfovibrio, an electroactive deltaproteobacteria, has been detected in bioelectrochemical systems for sulfate reduction and has one of the highest affinities for hydrogen among SRB. Importantly, many electroactive bacteria have low relative abundance but play a key role in biodegrading pollutants. Different carbon sources resulted in different enrichment levels of sulfatereducing bacteria, including Desulfobacter, Desulfobulbus, Desulfomicrobium, and Desulfovibrio. The dominant sulfatereducing bacteria in CCB-NaHCO 3 and CCB-NaAc were Desulfobacter and Desulfobulbus. Desulfobacter can reduce sulfate and completely oxidize organic acids such as acetate to carbon dioxide. Desulfobulbus can reportedly utilize H 2 as an electron donor and acetate as an organic carbon source (Brandt and Ingvorsen 1997). The dominant sulfate-reducing bacteria in CCB-ethanol were Desulfomicrobium and Desulfovibrio. The members of the genus Desulfomicrobium are known to use sulfate as an electron acceptor to incompletely oxidize ethanol to acetate and can tolerate alkaline environments ). These results suggested that different carbon sources resulted in different enrichment patterns of related functional genera, and CCB-ethanol enriches sulfate-reducing bacteria with higher abundance, enhancing the sulfate-reducing ability.

Effect of different carbon sources on functional genus
PCA results indicated that the microbial communities were well separated by carbon source, further revealing the difference in the compositions of their functional microbial community structures (Fig. 6a). Several functional genera such as Desulfovibrio, Desulfomicrobium, and Geobacteraceae were positively correlated with the CCB-ethanol, whereas Lentimicrobium and Arcobacter showed a significantly positive correlation with the CCB-NaAc. These results demonstrated that several functional genera that are known to promote sulfate reduction -had a strong correlation with the CCB-ethanol.
Analysis of the microbial sharing network indicated that the type of carbon sources played an essential role in the formation of network topology (Fig. 6b). We found that 1604 OTUs were distributed under all enrichment conditions among a total of 3499 OTUs. Upon comparing the OTU species exclusive to different carbon sources, CCB-NaHCO 3 was found to exhibit the highest number of OTU species, which is consistent with previous α diversity results. However, the CCB-ethanol exhibited the largest proportion of functional genera, and therefore, this had the highest sulfate reduction ability. The shared genera, including Arcobacter, Desulfobulbus, Desulfobacter, Desulfococcus, Desulfomicrobium, and Desulfovibrio were gathered in the center. Among these, Desulfobulbus, Desulfobacter, Desulfococcus, and Desulfomicrobium were SRB genera. These results suggest that large proportions of functional genera were enriched, but the composition varied greatly in response to different carbon sources.

Conclusion
Approximately 1.12 − 2.16 times higher sulfate reduction efficiency was achieved in the closed-circuit reactors than those in open-circuit reactors, thus confirming that electrical stimulation can promote sulfate reduction. The highest sulfate reduction efficiency was observed in the closed-circuit reactor fed with ethanol. In biocathodic systems, the addition of different carbon sources can affect the composition of bacterial communities. Moreover, ethanol can be used as a low-cost carbon source for large-scale operations, once a reasonable biomass yield has been achieved. Our findings demonstrated that different carbon sources directly affect the bacterial communities, which consequently impacts the performance of sulfate pollutant treatment. Besides, the real wastewater is much more complicated; we need to consider interactive effects between different pollutants. In the future, the material and size of the electrode, the influence range of the electrode, and the type of inoculum used can be further investigated. Based on energy consumption and cost considerations, a new type of low-voltage photoelectric microbial electrolytic cell (PMEC) can be used for further research. This study provides a foundation for the application of the bioelectrochemical system in sulfate wastewater.