Characterization and significance of extracellular polymeric substances, reactive oxygen species, and extracellular electron transfer in methanogenic biocathode

The microbial electrolysis cell assisted anaerobic digestion holds great promises over conventional anaerobic digestion. This article reports an experimental investigation of extracellular polymeric substances (EPS), reactive oxygen species (ROS), and the expression of genes associated with extracellular electron transfer (EET) in methanogenic biocathodes. The MEC-AD systems were examined using two cathode materials: carbon fibers and stainless-steel mesh. A higher abundance of hydrogenotrophic Methanobacterium sp. and homoacetogenic Acetobacterium sp. appeared to play a major role in superior methanogenesis from stainless steel biocathode than carbon fibers. Moreover, the higher secretion of EPS accompanied by the lower ROS level in stainless steel biocathode indicated that higher EPS perhaps protected cells from harsh metabolic conditions (possibly unfavorable local pH) induced by faster catalysis of hydrogen evolution reaction. In contrast, EET-associated gene expression patterns were comparable in both biocathodes. Thus, these results indicated hydrogenotrophic methanogenesis is the key mechanism, while cathodic EET has a trivial role in distinguishing performances between two cathode electrodes. These results provide new insights into the efficient methanogenic biocathode development.


Materials and methods
Experiment. Two single chamber MEC-AD systems (working volume of 360 mL), constructed with plexiglass, were used in this experiment. Carbon fibers (2293-A, 24A, Fibre Glast Development Corp., Ohio, USA) fixed onto a stainless-steel frame (which was not exposed to the liquid medium) was used as an anode electrode in both reactors. A similar carbon fiber module was used as a cathode electrode in one reactor (referred to as 'CF-CF'), and a stainless-steel mesh (304 stainless steel, McMASTER-CARR, USA) was used as a cathode in the other reactor (referred to as 'CF-SS'). The specific surface area provided by the stainless-steel electrode was 4.2 m 2 /m 3 . Considering every single filament in the carbon fiber bundle (Fig. S1), the specific surface area of the carbon fibers module was estimated at 3609 m 2 /m 3 . Moreover, considering all filaments in a bundle as a single fiber, the surface area was estimated at 41 m 2 /m 3 . The detailed calculation of the specific surface areas is provided in the Supporting Information. The Ag/AgCl reference electrode (MF-2052, Bioanalytical System Inc., USA) was positioned close (< 1 cm) to the anode electrode.
Both reactors were inoculated with mesophilic anaerobic digester sludge and effluent from a dual-chamber mother MEC operated with sodium acetate as an electron donor for > 12 months. Initially, sodium acetate (1600 ± 55 mg COD/L) supplemented phosphate buffer (50 mM) and trace minerals were used as a substrate. The details of the trace minerals can be found elsewhere 2 . Both reactors were operated with acetate in fed-batch mode until peak current densities reached ~ 77 A/m 3 . Then, the substrate was switched to glucose (2150 ± 31 mg COD/L), while buffer and trace minerals composition remained the same. With glucose, reactors were operated for about 6 months under batch mode in repetitive cycles. The biogas produced from the reactors were collected in gasbags. A decrease in daily methane production to < 3 mL was considered an indicator for replacing the substrate medium. During experiments, the anode potential was fixed at − 0.4 versus Ag/AgCl with a potentiostat (Squidstat Prime, Admiral Instruments, Arizona, USA). This anode potential was selected to enrich and maintain kinetically efficient EAB as suggested in the literature [40][41][42] . The reactors were operated at room temperature (21 ± 1 °C) with continuous mixing (130 ± 5 rpm) of the liquid medium with magnetic stirrers.
EPS and ROS analyses. For EPS analysis, biomass samples were washed three times with 0.1 M PBS (pH 7.4), and then the supernatant was removed after centrifugation at 3000×g for 15 min at 21 °C. EPS extraction was performed using two methods: cation-exchange resin (CER) and heating method. Applying cation-exchange resins and heating methods was highly efficient in several previous studies for the EPS extraction from biofilms, particularly carbohydrates, proteins, and eDNA 20,23,[43][44][45] . In addition, the pellets were collected to examine the cell lysis interference using a Glucose-6-Phosphate Dehydrogenase kit (Sigma-Aldrich, USA). The details of EPS extraction and analytical methods for various EPS components (proteins, carbohydrates, eDNA, heme-binding proteins, and uronic acid) are provided in the Supporting Information. These five major EPS components were Microbial communities and gene expression analyses. For  For evaluating microbial diversity, the nucleotide sequence reads were sorted out using a data analysis pipeline. Short sequences, noisy reads and chimeric sequences were removed through a denoising step and chimera detection, respectively. Then, each sample was run through the analysis pipeline to establish the taxonomic information for each constituent read. Microbial taxonomy was assigned using the Quantitative Insights Into Microbial Ecology (QIIME) pipeline (QIIME2, Version 2021.2, http:// qiime. org 57 ). Principal component analysis (PCA) of cathodic microbial communities was conducted using weighted Unifrac metrics to show the relation between genera and PCs. The expressions of EET genes (i.e., pilA, omcB, omcC, omcE, omcZ, and omcS) were also quantified (for details and method, see Supporting Information). The primers and design methods are listed in Table S2.
Analytical methods and statistical analysis. Current and applied voltage/potential were recorded every 4.8 min using a computer connected with the potentiostat. The chemical oxygen demand (COD) was measured with HACH method using UV-spectrophotometer (Model DR 3900, HACH, Germany). The volatile fatty acids, VFAs (acetate, propionate, and butyrate), concentrations were measured with an ion chromatography (Dionex ICS-2100, Thermo Scientific, USA) 42 . Electrochemical impedance spectroscopy (EIS) was performed for both reactors using a multi-channel VSP potentiostat (VSP, Bio-Logic Science Instruments, France). The detailed methodology is provided in the Supporting Information. The biogas produced from reactors was collected with 500 mL gas bags. The composition of biogas (i.e., methane content) was analyzed with a gas chromatograph (7890B, Agilent Technologies, Santa Clara, USA) equipped with a thermal conductivity detector and two columns (Molsieve 5A and Hayesep). To reveal the statistical difference between the results collected from two reactors, the student's paired t-test (JMP Software, Version 11.0.0, SAS Institute Inc., Cary, NC, US, https:// www. jmp. com/) was used.

Results and discussion
MEC-AD performance. The performance of the two configurations was compared based on volumetric current density and methane productivity. As shown in Fig. S3, the maximum current density from the CF-SS reactor reached 34.1 ± 0.3 A/m 3 , which was significantly higher (p = 0.01) than CF-CF (27.6 ± 0.2 A/m 3 ). Although the methane generation patterns were comparable in both reactors (Fig. 1), CF-SS showed higher (p = 0.03) daily methane production than CF-CF throughout the batch cycle. The total cumulative methane production was substantially higher in CF-SS (179.5 ± 6.7 vs. 100.3 ± 7.9 mL CH 4 ; p = 0.01). Both reactors used carbon fiber as the anode electrode and were operated under identical operating conditions (e.g., mixing speed, substrate, inoculation, etc.). Hence, the differences in system performance could be closely tied to the difference in the cathode electrode. As discussed later, stainless steel mesh cathode in CF-SS facilitated denser biofilms formation with more methanogenic biomass.
Anode electrodes providing high specific surface areas have been efficient for enhancing the performance of various bio-electrochemical systems 3,28,[30][31][32] . Therefore, carbon-based electrodes, such as carbon brush, activated carbon, have also been widely used for various biocathode applications, including electro-methanogenesis 3,30,32 . Notably, the rough surface of carbon fiber was found to be efficient for developing EAB biofilms 58 59,60 , the specific surface area provided by electrodes is often considered a critical factor. This study shows that stainless steel cathode having a relatively lower specific surface area than carbon fibers (4.23 vs. 3609 m 2 /m 3 ) resulted in a superior methanogenic activity. It has been previously suggested that the agglomeration of fibers in the liquid phase could reduce the available specific surface area for biofilms formation 61 . Nonetheless, considering all carbon fiber filaments in a bundle as a single fiber, the specific surface area provided by the carbon fiber was still higher than stainless-steel (4.23 vs. 41 m 2 /m 3 ). In general, carbon-based electrodes are considered inferior catalysts for HER than metal and carbon-metal composite electrodes 11,31,37-39 .
Previous MEC-AD studies substantiated the role of hydrogenotrophic methanogenesis. It is also reasonable that acetoclastic methanogens would likely be washed out at low residence time (< 7 days) used in this study 7,62 . EIS analysis also indicated that stainless-steel biocathode could reduce various intrinsic internal resistances in CF-SS compared to CF-CF (see Supporting Information). As shown in the Nyquist plot (see Fig. S4), the overall internal impedance of CF-SS (52.62 Ω) was lower than that of CF-CF (77.28 Ω). Thus, stainless-steel cathode largely influenced the internal resistances, which influenced the HER kinetics and, ultimately, growth and activities of hydrogenotrophic methanogens. Previous studies also suggested that lower ohmic resistance in MECs could provide faster HER kinetics 63,64 . Thus, the inferior methane recovery from the CF-CF reactor than the CF-SS reactor was likely due to the inferior HER on carbon fibers and subsequent hydrogenotrophic methanogenesis.
Organics removal and VFAs profiles. The effluent COD concentration from CF-SS (215 ± 2.8 mg/L) was considerably lower (p = 0.001) than that of CF-CF (382 ± 3.0 mg/L) (Fig. S5a). Correspondingly, COD removal efficiency in CF-SS (89.9 ± 0.5%) was significantly higher (p = 0.012) than that of CF-CF (83 ± 1.7%). Throughout the batch operation, propionate concentrations in CF-SS remained relatively lower than those observed in CF-CF, indicating faster conversion of propionate in the CF-SS. The fermentation of propionate to acetate is a vital process towards anodic respiration (by EAB) and acetoclastic methanogenesis. However, propionate fermentation to acetate is energetically unfavorable in terms of Gibbs free energy 65 . Thus, maintaining lower hydrogen partial pressure would be critical for propionate fermentation to acetate. Even though stainless-steel cathode would be expected to provide superior HER than carbon fibers 11,38 , no hydrogen was detected in biogas from both reactors. This might be due to the rapid consumption of hydrogen by hydrogenotrophic methanogens, as suggested in previous studies 11,66 . Moreover, enhanced homoacetogenic activity (H 2 + CO 2 → acetate) could assist in maintaining lower hydrogen partial pressure in biocathode 11,66 . Microbial community analysis also coincided with these notions (discussed later). Thus, the VFA profiles suggest that the microbiome in CF-SS more rapidly utilized hydrogen produced via fermentation and cathodic HER. Fig. 2a, the EPS composition of anode biofilms in both reactors was quite similar and was not affected by the different cathode materials used. Protein was found as the major EPS   23,46 . Geobacter species were also abundant in anode biofilms in both reactors in this study (discussed later). The concentrations of major EPS components (carbohydrates, proteins, and hemes) in the cathode biofilms in CF-SS were higher than those of CF-CF. Notably, carbohydrates and proteins in cathodic EPS were markedly higher in CF-SS than CF-CF (carbohydrates: 52.2 ± 0.2 vs. 25.8 ± 0.5 mg/cm 2 ; proteins: 212.8 ± 3.4 vs. 170 ± 1.3 mg/cm 2 ). The heme-binding proteins, uronic acid, and eDNA also showed the same patterns. Overall, cathodic biofilms developed on the stainless-steel electrode exhibited markedly higher EPS levels (p = 0.03). Moreover, electrode surfaces were visualized with CLSM (Fig. 3). The CLSM images showed that EPS was more uniformly distributed on the stainless-steel biocathode in CF-SS than the carbon fiber electrodes in both reactors. The biovolume of cathode biofilms in CF-SS was estimated at 30.2 ± 4.2 µm 3 /µm 2 , which was two times higher than that estimated for cathode biofilms in CF-CF (13.5 ± 2.8 µm 3 /µm 2 ) (Fig. 2b). The biovolumes estimated for anode biofilms in both reactors were comparable (p = 0.007). The intensities of EPS and eDNA were also quantified (Fig. 2c). Like estimated biovolume, EPS and eDNA intensities in stainless steel biocathode were higher than those estimated for carbon fiber biocathode (p = 0.008). Simultaneously, EPS and eDNA intensities were comparable for anode biofilms in both reactors (p = 0.20). Thus, CLSM imaging and COMSTAT (COMSTAT2, Version 2.1, Dk, http:// www. comst at. dk/) 51-53 analysis further confirmed that the stainless-steel biocathode resulted in the highest EPS production. The SEM imaging of biofilms also corroborated these results (Fig. S6). The biofilms did not fully cover the surfaces of carbon fibers, while biofilms grown on stainless steel cathode in CF-SS were evenly denser than anode/cathode biofilms grown on carbon fiber electrodes. The anode/ cathode biofilms grown on carbon fibers exhibited substantial heterogeneity. In contrast, a large secretion of EPS could accelerate the surface attachment of cells on the stainless-steel.

EPS characteristics. As shown in
Studies on the EPS in electroactive biofilms received less attention and primarily focused on understanding their role in anodic EET. A few reports revealed redox-active features of anodic EPS in model EAB biofilms (e.g., Geobacter sulfurreducens, Shewanella oneidensis, and Pseudomonas putida) 20,23,46 . Notably, higher levels of proteins in anode biofilms were correlated with higher EET efficiency. In this study, despite differences in volumetric current densities, both EPS composition and concentrations were quite similar in anode biofilms in both reactors. Instead, the difference in cathodic EPS levels was likely linked to current densities and methane productivity. As mentioned earlier, EPS can serve as immobilized redox cofactors (i.e., electron carriers) for facilitating EET in anodic EAB biofilms 20,23 . EAB can also regulate EPS generation to balance EET and protect cells 48 . The existing literature provides limited information on the roles of EPS in methanogenic biocathode. However, a few reports suggested that EPS could play similar roles (EET and cell protection) in archaeal biofilms in conventional   www.nature.com/scientificreports/ of iron-based conductive materials in conventional anaerobic bioreactors could enhance redox-active EPS contents in methanogenic biomass 24 , which was positively correlated with methanogenesis rates. Conductive materials promote the syntrophic DIET from bacteria to archaea and thereby enhance methanogenesis 68 . Therefore, the CV of biocathode EPS from two reactors was performed to identify their redox activity (Fig. S7) qualitatively. As shown in Fig. S7, the voltammograms of cathodic EPS extracted from both reactors showed distinct redox peaks, indicating their redox capability. However, redox peaks were observed at different potentials, suggesting that redox properties would be different for EPS extracted from two biocathodes. The peak current from stainless steel biocathode EPS was considerably higher than the EPS extracted from carbon fiber biocathode. This difference could be associated with higher levels of redox-active EPS in stainless steel biocathode, as previously suggested in the literature for anodic EAB biofilms 20,46 . Despite higher EPS levels in stainless-steel biocathode and differences observed in CV patterns, the expressions of genes associated with EET were comparable in both biocathodes (discussed later). Thus, it can be inferred that redox activities of EPS did not play a decisive role in differentiating between the performances observed from the two systems. Instead, EPS variations might be more associated with the protection of cells from harsh metabolic environments. Nonetheless, future investigation is warranted to reach a more thorough understanding and quantitative characterization of redox properties of EPS.
A recent study reported that the current from anode biofilms was positively associated with EPS protein content and negatively correlated to carbohydrates in EPS 48 . In this study, both carbohydrates and proteins in EPS were considerably higher in stainless steel biocathode than that of carbon fiber biocathode (see Fig. 2). The secretion of carbohydrates could be associated with harsh environmental conditions 19,25 to provide a protective layer and maintain the redox activity of proteins involved in EET 22,25 . It is possible that enhanced HER in stainless steel cathode could create highly alkaline conditions near the cathode 7,69,70 , which might induce more EPS secretion. Based on a recent report, hydrogenotrophic methanogenesis could be the dominant pathway under alkaline pH 71 . As discussed later, stainless steel biocathode also showed a higher abundance of hydrogenotrophic methanogens in this study. Thus, it appeared that higher enrichment of hydrogenotrophic methanogens promoted by faster HER kinetics on stainless steel cathode was possibly associated with higher EPS excretion. While further investigation is needed to get more insights into the function of EPS on electro-methanogenesis, these results suggested that different cathode materials could influence EPS secretion and methanogenic activity due to differences in HER kinetics.

ROS levels.
The quantitative measurement of ROS demonstrated a significant difference between biofilms grown on stainless steel and carbon fibers (Fig. 2d). The lowest ROS level was observed for cathode biofilms formed on stainless steel, while ROS levels were very similar in anode/cathode biofilms developed on carbon fibers. Recent studies reported ROS accumulation in anaerobic digesters 21,72 , while ROS is usually thought to be produced during aerobic metabolism. It has been suggested that unfavorable metabolic conditions (e.g., inhibition by toxicants, pH changes) could lead to ROS accumulation in digesters 21,72 . ROS accumulation may suppress metabolic activities, leading to the deterioration of digester performance. As we used synthetic glucose medium as a substrate, the potential unfavorable metabolic conditions induced by any toxic compounds can be ruled out. Thus, potential local pH changes by HER can be considered as an unfavorable metabolic condition. The HER in both biocathode can lead to alkaline pH due to protons reduction (2H + + 2e − → H 2 ), while effects will likely be more intense on stainless steel cathode 11,31,[37][38][39] . Thus, the lowest ROS level in stainless steel biocathode suggests that higher EPS levels provided some degree of protection to the cathodic microbiome from potential environmental stress (e.g., local alkaline pH due to superior HER). However, potential mechanisms relating to EPS and ROS levels should be further explored.
Microbial quantity and diversity. Figure 4 shows the quantitative assessment of microbial communities performed with qPCR. The total microbial cell counts (16S) in anode biofilms in CF-SS were slightly higher than that of CF-CF (9 × 10 8 vs. 8 × 10 8 cells/cm 2 ) (Fig. 4a). An almost similar pattern was observed for cathode biofilms; however, the difference was more prominent (1 × 10 11 vs. 6 × 10 8 cells/cm 2 ). The archaeal cell numbers also showed similar patterns, with the highest archaeal cell numbers for the stainless steel biocathode.
Furthermore, mcrA gene copies were quantified (Fig. 4b), considered a biomarker for hydrogenotrophic methanogenesis 73 . A few recent reports also confirmed the positive link between mcrA gene copies and methanogenesis rates in MEC-AD reactors 56,73,74 . The highest number of mcrA gene copies was observed for the stainless steel biocathode (4 × 10 6 cells/cm 2 ; 100 times higher than carbon fiber biocathode). The mcrA gene copies in anodic biofilms for both systems were comparable. Thus, the higher abundance of mcrA gene copies within the stainless steel biocathode corroborated with higher methane productivity in the CF-SS reactor.
The alpha diversity of microbial communities was also estimated (Table S3). The higher values of Chao 1, phylogenetic distance, OTUs, Pielou's evenness, and Shannon index clearly showed that the richness and diversity indices were relatively higher in CF-SS than CF-CF. Notably, cathode biofilms in CF-SS showed more diversity with the Shannon index of 5.10, as compared to CF-CF (3.95). These results indicated that the stainless-steel electrode persuaded the richness and diversity of the microbial communities.
Principal component analysis. The PCA analysis of biocathode bacterial and archaeal communities was performed to evaluate the relation between genera and PCs (Fig. S9). Based on 16S rRNA bacterial sequencing of biocathode, the superior performance of CF-SS was related to the enrichment of homoacetogenic Acetobacterium (Fig. S9a). However, the other genera might have an indirect relation to the superior performance of CF-SS. Based on archaeal sequencing of biocathode, hydrogen-consuming Methanobacterium and Acetobacterium primarily contributed to the superior performance of stainless steel biocathode (Fig. S9b).
Expression of EET genes. The gene expression for pilA and c-type cytochromes (Fig. 6) shows trivial differences in their expression levels in anode/cathode biofilms between both reactors. Moreover, compared to anode biofilms, the EET-associated genes were less expressed in cathode biofilms in both reactors. Based on the authors' knowledge, this study first reports the expression of EET genes for methanogenic biocathode. The EET from EAB to the anode has been demonstrated to be facilitated via c-type cytochromes and conductive nanowire or pili 77 , while the significance of EET in methanogenic biocathode is still ambiguous. However, previous reports postulated that conductive pili and c-type cytochromes could play an important role in DIET from EAB to methanogens 78,79 . Notably, some bacteria (e.g., Enterobacteriaceae, Desulfovibrio, etc.) found in biocathode in this study could express different cytochromes and/or conductive pili 77,80 . Furthermore, a recent study suggested that Methanobacterium species could produce methane via DIET 81 . Nonetheless, the expressions of these EET genes were quite comparable in both systems, indicating higher current density and methane productivity from the CF-SS reactor was not attributed to the overexpression of EET genes.
Implications. This study provides new insights into the characteristics and significance of EPS and expressions of EET genes in methanogenic biocathode. As compared to the carbon fiber, significantly higher EPS levels were observed in the stainless steel biocathode. Protons reduction to H 2 during HER can create local alkaline pH on the cathode. Thus, it could be posited that the highest EPS secretion in stainless steel biocathode could be linked with faster HER. One important finding of this current study is that EET may not play a decisive role in differentiating performances in MEC-AD systems using different electrode materials. Instead, the effective catalysis of HER, lower internal resistance, and higher abundances of H 2 -utilizing methanogens and homoacetogens on stainless steel cathode appeared to be the primary reason behind the higher methanogenic activity. Nonetheless, based on EET gene expression patterns and redox activity of biocathode-derived EPS, EET would still be involved in cathodic electro-methanogenesis.
Regarding the engineering significance of the results, carbon-based cathode electrodes have been mostly used in MEC-AD systems due to their excellent biocompatibility and higher surface area over metal-based electrodes 3,30-32 . While carbon fibers provided a higher specific surface area, stainless steel mesh outperformed www.nature.com/scientificreports/ carbon fibers under similar operating conditions (e.g., anode electrode, inoculum, mixing, etc.). Given that most of the single-chamber MEC-AD studies used carbon-based biocathode 5,7 , the results of this study are significant for selecting efficient cathode materials to realize improved performance. However, it should be noted that the results presented here are from specific operating conditions with two selected electrode materials. Hence, further research is warranted with more carbon and metal electrodes with similar textures and surface areas.