Photo-electrofermentation screening was performed in 250 ml single-chamber reactors fed with wastewater from the Mahou San Miguel brewery (Alovera, Spain). In order to evaluate the effect of reduction status on PPB culture, we used two different brewery wastewater: high-reduced wastewater (5.9 e⁻/C) and low-reduced wastewater (3.5 e⁻/C).

In order to explore the effect of polarization on PPB cultivation, different treatments were compared: Open Circuit Potential (OCP), −0.4 V (vs. Ag/AgCl) and –0.8 V (vs. Ag/AgCl; Figure 1).

The polarized reactors (−0.4 V and –0.8 V) were made up of three electrodes: a working electrode (WE) whose electrical potential is controlled by a potentiostat with respect to a reference electrode (RE, 3 M KCl Ag/AgCl; Hanna Instruments S.L.) and a counter electrode (CE) in which the counterreaction takes place. In these systems, the working electrode was made up of a stirred bed of vitreous carbon (Sigradur G, HTW, Germany), using a carbon rod (Mersen, Courbevoie, France) as current collector (Liu et al., 2014; Caizán-Juanarena et al., 2020). Moreover, a platinized titanium mesh (Inagasa S.A., Barcelona, Spain) was used as counter electrode. The bed was stirred with a magnetic bar. Electrochemical measurements were performed with a multipotentiostat NEV6 (Nanoelectra S.L., Alcalá de Henares, Spain) and electric current was measured with a Keithley Series 6,400 multimeter (Tektronix/Keithley, Cleveland, United States). Additionally, a non-polarized reactor without electrically conductive material (Electrode-free) was operated as a control.

Photo-electrofermentation in presence of a fluid-like cathode was performed in microbial electrochemical fluid-like cathode reactors (Tejedor-Sanz et al., 2020) fed with synthetic wastewater containing NaHCO₃ 2.5 g L⁻¹, NH₄Cl 0.5 g L⁻¹, NaH₂PO₄.2·H₂O 0.41 g L⁻¹, KCl 0.1 g L⁻¹, a mixed of vitamins 10 ml/l, a mixed of minerals 10 ml/l. Medium was sparged with N₂:CO₂ (80:20) to remove dissolved oxygen. We tested two organic acids as carbon source to a final concentration of 1 g·L⁻¹ of total organic carbon (TOC): malate (3 e⁻/C) and acetate (4 e⁻/C).

Separation of anodic and cathodic is key in electrofermentation so we designed and built a bicameral ME-FBR (Figure 2). Thus, we operated or ME-FBR, with (a) non-polarized electroconductive fluid-like bed (OCP); (b) electroconductive fluid-like bed polarized at −0.6 V (vs. Ag/AgCl); and (c) an electrode free system as control. The polarized reactor potential was –0.6 V to ensure cathodic polarization throughout the bed.

The reactors, made of borosilicate, had an inlet at the bottom and an outlet at the top to recirculate the liquid upward. At the bottom of the reactor, a fritted plate was used as a diffuser. In the polarized reactor, in addition to the main chamber, a second chamber separated by an ion exchange membrane, was designed, to house the counter electrode.

We used a multichannel peristaltic pump (Dinko, Barcelona, Spain) to recirculate the liquid in all the reactors. In the reactors with electroconductive bed both OCP and polarized at –0.6 V) vitreous carbon particles (Sigradur G, HTW, Germany) were used as bed. In the polarized reactor, following the same three-electrode scheme explained in the previous section, a current collector (Mersen, Courbevoie, France) was used to polarize the working electrode (fluidized bed). A reference electrode (Hanna Instruments S.L.) was placed immersed in the bed. A platinized titanium mesh (Inagasa S.A., Barcelona, Spain) was placed in the anodic chamber as a counter electrode. The ion exchange membrane between the chambers was provided by Nafion (The Chemours Company FC, LLC).

Stirred reactors were operated under IR-illuminated conditions in fed-batch with a hydraulic retention time (HRT) of 5 days during a period enough to reach steady state. After 1 week of operation under steady state, samples were taken to assess biomass production and analyze microbial populations.

The OCP acted as non-artificially polarized control. We used two cathodic potentials: −0.4 V and –0.8 V (vs. Ag/AgCl).

Fluid-like bed reactors were operated under IR-illuminated (13 W/m²) conditions in batch. The systems were operated until the stationary phase of growth was reached. At the end of the experiment, samples were taken to assess biomass production and analyze microbial populations.

Water quality analysis showed similar COD removal in all treatments with both types of tested wastewater (Figure 3). Therefore, cathodic polarization did not exert a dramatic effect in organic pollutants removal, except for the apparent lower removal (−20%) in the high-reduced wastewater (5.9 e⁻/C) under −0.8 V (vs. Ag/AgCl) polarization.

The complexity of the compounds in the brewery wastewater, such as sulfur compounds or some metals, may hinder the evaluation of the organic matter removal by COD analysis (Watts and Dean Adams, 1983). The analysis of total organic carbon (TOC), on the other hand, allowed a more accurate quantification of organic matter. Our TOC results revealed that the different COD removal observed in the high-reduced wastewater does not show a negative impact in the biodegradation performance (Figure 3). We hypothesize that the potential –0.8 V vs. Ag/AgCl could have favored electrofermentation processes (Prévoteau et al., 2020), giving rise to the transformation of the organics into other more reduced compounds. Although this finding requires further exploration, it was observed under hydrogen-mediated conditions (−0.8 V) and not through direct extracellular electron uptake (−0.4 V).

Nitrogen removal showed highly variable results throughout the assays with no significant impact of polarization (Supplementary Figure S1).

Extracellular electron uptake from a cathode is strongly connected to carbon dioxide fixation in PPB (Guzman et al., 2019). Therefore, we hypothesized that providing extra electrons to PPB via a cathode in brewery wastewater treatment could maximize carbon fixation and thus increase biomass production. We quantified the amount of biomass produced at the end of each tested condition. The results showed slightly different behaviors depending on the type of wastewater used as growth culture (Figure 4). In the high-reduced wastewater, we observed a 2-fold increase at –0.4 V and a 3-fold increase for the –0.8 V potential. In contrast, in the low-reduced wastewater there was a 4-fold increase at –0.4 V and a 7-fold increase in the reactor polarized at –0.8 V. These results showed that, unlike the water quality parameters (TOC and COD), biomass production was highly affected by cathodic polarization.

The impact in biomass production for both types of wastewater was consistent with PPB metabolism previously reported in presence of a high-reduced carbon source. Indeed, this condition activates the electron sinks so the microorganisms become less sensitive to cathodic polarization (McKinlay and Harwood, 2010; Vasiliadou et al., 2018; Guzman et al., 2019). This may explain why polarization had a greater impact when PPB where grown using low-reduced carbon sources.

Despite the impact of polarization on biomass production, we observed less consumption of electrical current than expected (ca. −50 to −100 μA; Supplementary Figure S2). We hypothesize that the cathode could be acting as a metabolic regulatory element rather than as a true electron donor, as previously reported in PPB electrofermentation processes (Manchon et al., 2023b). Under polarization conditions where hydrogen should be produced (ca. −0.8 V), we monitor higher current consumption (−80 to −100 μA). We are certainly producing less hydrogen than the one expected with non-carbonaceous materials like platinized titanium; unfortunately, we doubt this material could be fluidized and still be biocompatible.

Electrofermentation processes using microbial consortia led to changes in microbial population structure and metabolic activity. Therefore, we explored the microbial community using 16S Illumina sequencing to understand the role of polarization in microbial selection. The analyses revealed differences at all taxonomic levels although we focused on the genus level as it is the most informative regarding functionality.

The mere presence of electrically conductive material has been reported to favor the exchange of electrons between electroactive microorganisms even without external polarization (Rotaru et al., 2021). This phenomenon, which has been described in non-phototrophic bacteria either using fixed bed (Prado et al., 2022) or fluid-like beds (Tejedor-Sanz et al., 2018). In contrast, our stirred reactors showed similar pollutant removal efficiency and biomass production in both free-electrode reactors and stirred-electrode reactors under Open Circuit Potential (OCP; Supplementary Figures S3A,B). However, the presence of electroconductive material had an impact on the microbial population composition (Supplementary Figure S4). The electrode-free reactors showed a majority presence of microorganisms characteristic of wastewater, with Pseudomonas and Azonexus genera and a very low presence (<0.5%) of PPB genera such as Rhodopseudomonas or Ectothiorhodospira and electroactive bacteria such as Geobacter. On the contrary, Rhodopseudomonas sp., the electroactive PPB model genus (Bose et al., 2014), was present when stirred electrons were used under OCP (Figure 5). In addition, other genera such as Pseudomonas and Acidovorax, considered core genera in electroactive communities (Xiao et al., 2015), were also present. In the OCP-based reactor fed with high-reduced wastewater (5.9 e⁻/C), we found Thauera genus, which was described as electroactive due to its extracellular respiration of AQDS (Liu et al., 2006). In both OCP reactors, Propionivibrio was found, a genus also present in other electrofermentation systems, capable of producing propionic acid and of interacting with extracellular redox species (Yang et al., 2021).

Stirred-electrode reactors polarized at –0.4 V (vs. Ag/AgCl) had a completely different microbial structure, showing maximum predominance of PPB genus (Figure 5). Under polarization at this potential (−0.4 V) the production of hydrogen is thermodynamically not possible (E = −0.63 V vs. Ag/AgCl), so extracellular electron uptake was not hydrogen-mediated. Therefore, extracellular electron transfer could only occur through direct electron transfer or through self-produced redox mediators (Logan et al., 2019). This is consistent with the high presence of electroactive genera: classical electroactive genus such as Geobacter (Koch and Harnisch, 2016) and other electroactive PPB genera such as Rhodopseudomonas (Bose et al., 2014). In these reactors (−0.4 V) wastewater composition was a key factor in microbial selection (Figure 5). Thus, using high-reduced wastewater (5.9 e⁻/C), we detected a predominance of Ectothiorhodospira a PPB also found among electroactive communities (Manchon et al., 2023b). On the contrary, Rhodopseudomonas, model electroactive PPB genus (Bose et al., 2014; Guzman et al., 2019) predominated in presence of low-reduced wastewater (3.5 e⁻/C).

The microbial community analysis revealed a different scenario in the reactors polarized at –0.8 V (Figure 5). In both reactors, the analyses showed 50% Pseudomonas and only in presence of high-reduced wastewater fed reactor, the PPB Caenispirillum was observed. This population structure was, in fact, very similar to the free-electrode reactors used as control (Supplementary Figure S4). At –0.8 V, hydrogen-mediated electron transfer is feasible (E = −0.63 V vs. Ag/AgCl). Although another extracellular electron pathway (H₂-mediated) could explain the different community observed at this potential, hydrogen does not inhibit the growth of PPB (Vasiliadou et al., 2018; Supplementary Figure S5). These results suggest that –0.8 V polarization inhibits PPB growth and although hydrogen is not involved in such inhibition, further study is needed.

We studied the main parameter in microbial cultivation: biomass production. Biomass analysis revealed higher production under fluid-like cathodic polarization at –0.6 V (vs. Ag/AgCl) in comparison with the electrode-free control (Figure 6). Such enhancement was confirmed for both acetate-feeding (+75%) and malate-feeding (+130%). Therefore, the results of this experiment confirmed our previous observation with stirred-electrode assays. Furthermore, they are consistent with the role of having an extra-source of electrons (cathode) on PPB metabolism, promoting carbon fixation and thus maximizing biomass production (Guzman et al., 2019).

In previous assays with stirred-electrode reactors fed with wastewater, we did not observe an impact in the biomass production (Figure 4) when the fluid-like electrode was non-polarized (OCP) in comparison with free-electrode conditions. In contrast, our fluid-like electrode (OCP) reactor fed with malate revealed higher biomass production (ca., 50%) in comparison with the electrode-free control (Figure 6). This could indicate that, with oxidized carbon sources (3 e⁻/C), the electrically conductive material could specially facilitate syntrophies between microorganisms as reported elsewhere (Rotaru et al., 2014, 2018). These syntrophic relationships seem to favor microbial growth, allowing the consortium to deal more easily with redox stress situations. This hypothesis is consistent with the absence of differences in biomass production when acetate (4 e⁻/C), a more balanced electron-carbon ratio (McKinlay and Harwood, 2010) was feeding our free-electrode or OCP reactors.

Once the effect of fluid cathode on biomass production was demonstrated, we analyzed microbial population structure by Illumina 16S sequencing. We focused on the population structure at the genus level as it is the most informative from the point of view of function and metabolism.

As we described in the assays with stirred-electrode reactors fed with wastewater (Figure 5), the mere presence of conductive material can select for a different microbial community. In this case, the differences were more evident in the malate-fed OCP reactor, with a higher relative abundance of Rhodopseudomonas (24.3%) compared to the free-electrode reactor (Figure 7). The acetate-fed reactor, on the other hand, showed a similar abundance of Rhodopseudomonas regardless the presence of a non-polarized electrode. These results were consistent with those shown in biomass production (Figure 6). Malate is a more unbalanced organic compound, so we hypothesized that the conductive material would promote electroactive bacteria from genus Rhodopseudomonas to deal with redox stress through electrosyntrophies (Liu et al., 2021; Rotaru et al., 2021). In the acetate-fed reactor, more redox-balanced substrate, electrically conductive material does not provide a competitive advantage to electroactive bacteria and therefore its effect was less remarkable.

Additionally, we observed a high presence of Geosporobacter in all the reactors, including the electrode-free control (Figure 7). Some bacteria of this genus have been described as an anaerobic iron-reducing bacteria (Hong et al., 2015) capable of colonizing electrodes (Miceli et al., 2012; Moscoviz et al., 2018). These results could indicate some type of syntrophic relationship between both genera where Geosporobacter, in absence of electron acceptor, may transfer electrons to Rhodopseudomonas which is well known for its great ability to dissipate reducing power via electron sinks (Guzman et al., 2019). The possible electrosyntrophy between the genera Geosporobacter and Rhodopseudomonas requires further study.