A three-dimensional hybrid electrode with electroactive microbes for efficient electrogenesis and chemical synthesis

Significance Addressing the global challenge of sustainability calls for cost-effective and eco-friendly pathways to go beyond the existing energy-intense synthetic routes. Biohybrid electrochemical systems integrate electroactive bacteria with synthetic electrodes to leverage the power of biocatalysis for energy conversion and chemical synthesis. This work presents a three-dimensional electrode scaffold to couple the intracellular metabolism with extracellular redox transformations by means of electrochemistry. The large population of bacteria actively metabolizing within the electrode scaffold produces a benchmark current density. The biohybrid electrode can also carry out synthetic reactions within or beyond biochemical pathways driven by solar light. This hierarchical electrode provides a robust and versatile platform to wire bacteria’s intrinsic physiological functionalities with artificial electronics for sustainable energy conversion and chemical production.

S3 solution (DSMZ 826, see the following for details) using sodium acetate (20 mM) as the electron donor and sodium fumarate (50 mM) as the electron acceptor. The bacterial strain was inoculated in 20 mL of the medium solution in a sterilized vial and was purged with N 2 :CO 2 (80:20 v:v%) for 1 h. The inoculated media were kept in a shaking incubator (30 °C, 180 rpm) for 5 days. S. loihica was cultured aerobically in Luria-Bertani medium solution (see the following for details) by keeping in a shaking incubator (30 °C, 180 rpm) overnight. The concentration of bacterial suspension was determined by measuring optical density (OD) at 600 nm using a UV-vis spectrometer (Varian Cary 50, Agilent Technologies). To suppress the outer-membrane c-type cytochromes without altering their genes, G. sulfurreducens was cultivated in the standard medium solution containing 2,2'-bipyridine (30 µM) as the iron chelator to limit the iron availability.(S4) G. sulfurreducens was first cultured in an iron-lacking medium at 30 °C for 5 days. Then 30 µM 2,2'-bipyridine was added into the medium and the bacteria were cultured anaerobically at 30 °C for another 5 days.

Components of the bacterium culturing mediums
(1) Medium for Geobacter sulfurreducens (DSMZ 826) NH 4 Cl 1.50 g Na 2 HPO 4 0.60 g KCl 0.10 g Trace element solution (see below) 10.00 mL NaHCO 3 2.50 g Vitamin solution (see below) 10.00 mL Distilled water 980.00 mL (2)  ). An ITO glass slide with a pre-defined area of 0.25 cm 2 was used for a comparison and an IO-ITO electrode without bacteria was used for control. Cyclic voltammetry was carried out after the anodic current reaching a plateau at a scan rate of Tryptone 10.00 g Yeast extract 5.00 g NaCl 10.00 g S5 5 mV s −1 . To co-culture with S. loihica, sodium lactate (40 mM) was added in the medium solution (13 mL) as the electron donor to S. loihica, and purged with N 2 :CO 2 (80:20 v:v%) for 40 min. G. sulfurreducens suspension (1 mL, final OD: 0.6) and S. loihica (1 mL, final OD: 0.6) were inoculated into the medium solution and purged for 20 min. The IO-ITO electrodes were used as the working electrode and poised at a potential of 0.4 V vs. SHE.

Protein quantification
The proteins in the electrodes were quantified with a colorimetric assay following a previously-reported protocol.(S5) The bacteria-colonized electrode was immersed in a sodium dodecyl solution (5 mL, 10 wt%) solution at 99 °C for 15 min to extract proteins from the electrode. Then the solution was centrifuged (14,000 rpm, 10 min) to remove the impurities. The supernatant was used to quantify the protein using the Bio-Rad protein assay. Typically, 100 μL of the protein solution was added into a clean test tube, followed by adding 5 mL of the diluted dye reagent and incubating at 25 °C for 10 min. Then the light absorbance at 600 nm was measured using a UV-vis spectrometer. Bovine serum albumin was used as the standard protein to make the correlation curve. The protein concentration was calculated from the standard curve. Each protein solution was assayed in triplicate.
Differential gene expression analysis IO-ITO|G. sulfurreducens electrodes were prepared at different potentials following the method reported above. The plateau current density attained at 0.1 V and 0.4 V vs. SHE was 1.12 ± 0.05 mA cm -2 and 2.82 ± 0.25 mA cm -2 , respectively. The biohybrid electrode was scratched off the ITO glass slide and cells in electrodes were lysed with TissueLyser II immediately after the chronoamperometry was stopped. The mixture was sonicated for 1 min and centrifuged at 8000 ×g for 15 min. The resulting supernatant was used for RNA extraction. G. sulfurreducens anaerobically cultured in acetate (20 mM) and fumarate (50 mM) at 30 °C was used as control samples. RNA of G. sulfurreducens were extracted using an RNeasy protect bacteria mini kit (Qiagen, USA) at room temperature. The extracted RNA were eluted in nuclease free water. The RNA sequencing and data analysis was conducted by Cambridge Genomic Services (CGS). The quality control showed the q-score across all samples are above 30, signifying high sample quality. Reads were mapped with G. sulfurreducens reference genome using STAR v2.5.2a. (S6) Using the genes annotation defined in the G. sulferreducens gtf file from Ensembl Bacteria. The reads mapping into genomics features were counted using HTSeq v0.6.0. A feature is considered as the union of all gene's exons whose genomic coordinates are determined from the G. sulfurreducens gtf. Reads with a mapping quality less than 10, or those that map to multiple loci or to overlapping gene regions are discarded to avoid ambiguity and false S6 positives. Differential gene expression analysis was performed using the counted reads and the R package edgeR version 3.16.5 for the 3 pairwise comparisons.

Microbial electrosynthesis
The electrosynthesis were performed with a potentiostat (MultiEmStat3+) in three-electrode configuration in a three-neck round bottom flask in a water bath (30 °C) under continuous stirring (200 rpm). A platinum mesh was used as the counter electrode and Ag/AgCl (in 3 M NaCl solution, +0.2 V vs. SHE) as the reference electrode. The IO-ITO|G. sulfurreducens electrode was replaced with a medium solution containing sodium fumarate (10 mM) or GO (0.1 mg mL −1 ) and purged with N 2 :CO 2 (80:20 v:v%) for 40 min. IO-ITO electrodes without bacteria were used for control experiments.
The Faraday efficiency was determined by the ratio of electrons derived from succinate (two electrons per succinate) to the electrons consumed based on chronoamperometry. We note that such calculation might include charge contributions from reducing equivalents being stored in G. sulfurreducens during the anodic growth.(S7) A potential of −0.45 V vs. SHE was applied on the electrode for fumarate reduction and −0.3 V vs. SHE for GO reduction.
Microbial photoelectrosynthesis IO-TiO 2 (pore size: 10 µm; thickness: 40 µm; geometrical area: 0.25 cm 2 ) were cleaned with a UV/ozone cleaner for 15 min before use. The IO-TiO 2 |RuP photoanode was prepared by immersing the IO-TiO 2 electrodes in [Ru II bis(2,2'-bipyridine)(2,2'-bipyridine-4,4'-diylbis(phosphonic acid))]Br 2 (RuP,(S8) 0.25 mM in H 2 O) for 16 h in the dark, followed by rinsing with water to remove weakly adsorbed dyes. Stepped chronoamperometry of the IO-TiO 2 |RuP electrode was conducted in TEOA (25 mM, pH 7.2, in 0.1 M NaCl solution) in a N 2 :CO 2 atmosphere (80:20, v:v%) with Pt and Ag/AgCl as the counter and reference electrode, respectively. The electrode was under periodic irradiation (10 s in light, 30 s in dark, I = 100 mW cm −2 , AM 1.5G) at stepped potentials from −0.7 V to 0.4 V vs. SHE every 40 s. The BiVO 4 -CoO x electrode (1.0 cm 2 ) was directly used as the photoanode without any pre-treatment. Linear sweep voltammetry of the BiVO 4 -CoO x electrode was carried out in a phosphate buffered saline solution (20 mM Na 2 HPO 4 , 3.6 mM KH 2 PO 4 , 5.4 mM KCl, 0.274 M NaCl, pH 7.3) in N 2 :CO 2 (80:20 v:v%) under periodic irradiation (5 s in light, 5 s in dark, I = 100 mW cm −2 , AM 1.5G) at a scan rate of 5 mV s −1 with Pt and Ag/AgCl as the counter and reference electrode, respectively. The photoanode and the IO-ITO|G. sulfurreducens electrode were connected in a two-compartment, two-electrode photoelectrochemical cell separated by a Nafion membrane. A bare IO-ITO electrode without bacteria and an IO-ITO hybrid electrode with G. sulfurreducens killed by 0.1% glutaraldehyde were used for control experiments. TEOA (25 mM, pH 7.2, in 0.1 M NaCl solution) was used as the S7 electrolyte and electron donor for the IO-TiO 2 |RuP electrode, whereas a phosphate buffered saline solution (20 mM Na 2 HPO 4 , 3.6 mM KH 2 PO 4 , 5.4 mM KCl, 0.274 M NaCl, pH 7.3) was used as the electrolyte for the BiVO 4 -CoO x electrode. Sodium fumarate (20 mM, in medium solution) was used as the substrate and electrolyte for the cathode. The photoelectrochemical cell was purged with N 2 :CO 2 (80:20 v:v%) for 40 min. The photoelectrosynthesis was performed using simulated solar light (I = 100 mW cm −2 , AM 1.5G) (LOT Quantum Design) from a 150 W Xe lamp (Newport). Light intensity was calibrated by a thermal sensor (S302C (Thorlabs)) and power meter console (PM100D (Thorlabs)). Zero bias (U = 0) was applied between the photoanode and the cathode during the two-electrode photoelectrochemical experiment. Dark current was recorded for 30 min before and after irradiation for 24 h. Solution in the cathode chamber before and after the light experiment was extracted for product quantification.
Physical characterization SEM and STEM images were acquired on a scanning electrode microscope (TESCAN MIRA3) at an accelerating voltage of 5 kV and 30 kV respectively. X-ray microscopy image of the IO-ITO electrode was acquired on a 3D Xray microscope (Zeiss Xradia 510 Versa). Cross-sectional SEM images were acquired on a focused ion beam-scanning electron microscope (FIB-SEM, ZEISS Crossbeam 540) at an acceleration voltage 1.6 kV. Serial sectioning was carried out using FIB milling at a 3 nA ion current and a slice thickness of 500 nm. Before SEM imaging, biohybrid electrodes were treated with 2.5 wt% glutaraldehyde and 2 wt% osmium tetraoxide, and then dehydrated with a series of ethanol solutions with increasing concentrations (30, 50, 70, 90 and 100%) and dried in air. The iron contents of bacteria and hybrid electrodes were determined by inductively coupled plasma-optical emission spectrometry (ICP-OES, Thermo scientific). Fluorescence images of IO-ITO|G. sulfurreducens electrodes were acquired on a confocal laser scanning microscope (Leica TCS SP8) using a 488 nm laser and a hybrid detector (600−650 nm). 100 μL of 5-cyano-2,3-ditolyl tetrazolium chloride (10 mM in medium solution) was dropcast on an IO-ITO|G. sulfurreducens electrode that was then incubated in the dark for 30 min at 25 °C. The AFM image was acquired on an atomic force microscope (NanoIR2, ANASYS Instrument) using a gold-coated silicon tip. 1 H NMR spectroscopy analysis was conducted on a Bruker 400 MHz NMR spectrometer in D 2 O. Sodium 3-(trimethylsilyl)propionate-2,2,3,3-d 4 (TMSP-d 4 , 1 mM) was used as the reference and internal standard for quantification. NMR spectra were processed with MestReNova v12.0. Raman spectra were recorded on a confocal Raman microscope (HORIBA LabRAM HR Evolution) with an excitation laser of 633 nm.

Statistical information
All experiments were performed in three individual replicates unless otherwise mentioned. Data are presented in bar diagrams as mean ± standard error of the mean (s.e.m.). The mean values and standard errors of the mean were calculated from the number of repeats of independent experiments. Statistical analyses were conducted using GraphPad Prism v.6.0 g. Statistical significance was determined by one-sided analysis of variance (ANOVA) for multiple groups of samples and Student's t tests for unpaired two samples. P < 0.05 was deemed statistically significant. Significance values: n.s., not significant (P > 0.05), ****P < 0.0001.

Supplementary text
(1) Significance summary of differential gene expression analysis (3) Extended discussion of differential gene expression analysis Microorganisms survive and thrive under different environments thanks to their acute and flexible regulatory mechanisms, which enable them to maintain their homeostasis in response to environmental variations. Through various sensing and transduction pathways, extracellular changes can be translated into intracellular signals that induce transcriptional responses.(S9) RNA sequencing is widely employed for gene expression profiling,(S10) which can provide a global view on microbial strategies to cope with environmental perturbations.

G. sulfurreducens can use various forms of soluble and insoluble electron
acceptors.(S11) Soluble electron acceptors such as fumarate can diffuse into a cell and participate in intracellular metabolism, whereas insoluble electron acceptors such as metal oxides cannot permeate through the cellular membrane and can only be utilized for anaerobic respiration by extracellular electron transfer. In a biohybrid system, an electrode with a poised electrochemical potential substitute for natural minerals as an inexhaustible electron sink, which forms the basis of microbial electrogenesis. When G. sulfurreducens is transferred to a new environment deprived of soluble electron acceptors (fumarate), bacteria will regulate gene expression to align their metabolism with the physiological needs of using an electrode as the electron acceptor. Although their cellular regulation has not been completely deciphered due to many genes with unknown functionality,(S12) several genes that are found responsible for extracellular electron transport could shed light on part of their adaptation strategies. Genome-wide study has S10 identified 111 gene encoding putative Cyt c (109 of them found in this study) in G. sulfurreducens, (S6) but only 16 multi-haem Cyt c were found essential for extracellular electron transfer.(S13) Here, we conducted differential gene expression analysis for G. sulfurreducens grown with fumarate (the control) and cultured on electrodes with different potentials (0.1 V and 0.4 V vs. SHE). Amongst more than 3500 genes sequenced, ~25% of genes were induced with significant changes (FDR < 0.05) in expression on electrodes, with ~75% of them down-regulated (see Significance summary). The down regulation of genes indicates that bacteria saved more energy when respiring on electrodes.(S14) However, there were much less differences in gene expression between electrodes at different potentials, suggesting potential variations cannot stimulate extensive transcriptional responses. S11 Fig. S1. Co-assembly method to prepare the IO-ITO electrodes. PS beads (10 μm) were mixed with ITO nanoparticles (average size < 50 nm) and an aliquot of the colloidal mixture was dropcast on an ITO glass substrate. The IO-ITO electrodes were obtained by annealing at 500 °C to remove the PS template and sinter the mesoporous ITO skeleton (see Methods for details). The same procedure has also been employed for IO-TiO 2 and IO-ZrO 2 electrodes used in this study. hybrid electrode (B). The electrodes were stained with 5-cyano-2,3-ditolyl tetrazolium chloride (CTC, 10 mM) and incubated in the dark for 30 min at 25 °C. As conventional fluorescent dyes (such as SYTO 9 and propidium iodide) adsorbed on the IO-ITO scaffold would be unable to differentiate between live and dead cells, we used an alternative dye to assess the bacterial viability. CTC is a soluble non-fluorescent compound whereas its reduced form, CTC formazan, is insoluble and produces red fluorescence. Living bacteria respiring via the electron transport chain adsorb CTC dyes and reduce them into insoluble CTC formazan that precipitates in the cell. Dead bacteria and IO-ITO scaffolds cannot reduce CTC, thereby producing much less fluorescence.(S15) This method allows us to assess the bacterial viability inside the IO-ITO electrode by means of evaluating their respiratory activity, without being interfered by the scaffold.   pH 7.4;30 °C,purged with N 2 :CO 2 (80:20,v:v%). The non-conductive IO-ZrO 2 scaffold was prepared on an ITO glass substrate following a similar co-assembly method using 10 µm PS beads and ZrO 2 nanoparticles (20-30 nm).

Fig. S14. A.
Stepped chronoamperometry of the IO-TiO 2 |RuP photoanode in TEOA (25 mM, pH 7.2), with Pt and Ag/AgCl as the counter and reference electrode, respectively. The electrode was under periodic irradiation (10 s in light, 30 s in dark, I = 100 mW cm −2 , AM 1.5G) at stepped potentials from −0.7 V to 0.4 V vs. SHE. The experiment was performed in N 2 :CO 2 (80:20, v:v%). B. Energy level diagram of an IO-TiO 2 |RuP photoanode coupled with an IO-ITO|G. sulfurreducens cathode with respect to an electrochemical potential scale at pH 7.0. Energy levels of TiO 2 and RuP are taken from Ref. (S16). As the mechanism of electron intake by bacterium remains elusive, here we tentatively assume electrons are transferred inward via c-type cytochromes with redox potentials centered at −0.2 V vs. SHE. (S17, 18) The reduction potential of fumarate (0.03 V vs. SHE) is sourced from Ref. (S19). C. Linear sweep voltammetry trace of the BiVO 4 -CoO x electrode in phosphate buffered saline (20 mM Na 2 HPO 4 , pH 7.3) under periodic irradiation (5 s in light, 5 s in dark, I = 100 mW cm −2 , AM 1.5G). Scan rate: 5 mV s −1 . Pt and Ag/AgCl were employed as the counter and reference electrode, respectively. The experiment was performed in N 2 :CO 2 (80:20, v:v%). D. Energy level diagram of a BiVO 4 -CoO x photoanode coupled with an IO-ITO|G. sulfurreducens cathode with respect to an electrochemical potential scale at pH 7.0. The conduction band (CB) edge of BiVO 4 was taken from Ref. (S20) S1 Movie S1.

Movie S2.
Serial cross-sectional SEM images of an IO-ITO electrode acquired from FIB-SEM.

Movie S3.
Serial cross-sectional SEM images of an IO-ITO|G. sulfurreducens hybrid electrode prepared at 0.1 V vs. SHE for 80 h.