Translating Extracellular Electron Transfer Activities with Organic Electrochemical Transistors

Summary

Here, we present a protocol for using organic electrochemical transistors (OECTs) to translate extracellular electron transfer (EET) activity in Shewanella oneidensis into electrical signals. The hybrid OECT system provides enhanced robustness, sensitivity, and potential for rapid, high-throughput testing, making it an effective tool for EET measurements.

Abstract

Extracellular electron transfer (EET) is a process through which certain microorganisms can transfer electrons across their cell membranes to external electron acceptors, linking cellular metabolism to their environment. While Geobacter and Shewanella have been the primary models for EET research, emerging studies reveal that EET-active species are also associated with fermentation and the human gut microbiome. Leveraging the ability of EET to bridge biological and electronic systems, we present a protocol for using organic electrochemical transistors (OECTs) to translate microbial EET activity into easily detectable electrical signals. This system enables the use of cellular responses to external stimuli for biosensing and biocomputing applications. Specifically, we demonstrated the de-doping of the p-type poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT: PSS) channel in the OECT is driven by cellular EET from Shewanella oneidensis. By transcriptionally controlling EET flux by genetic circuits, we establish the biosensing capability of this hybrid OECT system to detect chemical stimuli, such as inducer molecules. Furthermore, we introduce plasmid-based Boolean logic gates within the cells, allowing them to process environmental signals and drive current changes in the OECTs, further demonstrating the biocomputing potential of these devices. This method provides a novel interface between biological systems and electronics, enabling future high-throughput screening, biosensing, and biocomputing applications.

Introduction

Devices that can transduce and amplify biological and chemical activities into electrical signals are crucial across various fields, such as sensing^1,2, neuromorphic computing^3,4, and wearable electronics⁵. Among these, organic electrochemical transistors (OECTs) have emerged as exceptional interfaces between biological systems and electronic readouts due to their compatibility with aqueous environments and low operating voltages^6,7. OECTs differ from conventional electronics by utilizing ions in an electrolyte to modulate the conductivity of an organic channel, which couples ionic and electronic transport to achieve exceptional transconductance⁸. These characteristics make OECTs ideal for interfacing biological systems with electronics, as they can amplify weak biological signals and translate them into electrical readouts.

OECTs operate by altering the doping state of a mixed ionic-electronic conducting channel through ionic diffusion, typically controlled by applying a voltage at the gate electrode. However, biological or redox reactions can also change the channel's conductivity, enabling OECTs to respond to a variety of chemical and biological stimuli. Functionalizing OECTs with lipid bilayers, ion channels, or biomolecules allows them to detect specific analytes, making them useful for sensing applications^9,10. For instance, OECTs have been integrated with redox-active enzymes like glucose oxidase to directly transfer electrons to the channel, tuning its conductivity in response to glucose concentration¹¹. While such configurations are effective for biosensing, they are limited in their computational capacity due to the relatively simple behavior of individual enzymes or proteins.

In contrast, living cells, particularly bacteria, offer a versatile platform capable of performing complex and robust computations^{4,12,13,14,15}. Electroactive bacteria such as Shewanella oneidensis and Geobacter sulfurreducens have the unique ability to transfer electrons across their cell membranes in a process known as extracellular electron transfer (EET). Under anaerobic conditions, these bacteria couple their metabolic processes to the reduction or oxidation of external electron acceptors, including metals, metal oxides, and synthetic materials like conducting polymers¹⁶ and nanoparticles¹⁷ (Figure 1A). This capability has been exploited in microbial fuel cells for power generation and offers the potential for more advanced bioelectronic applications^18,19. Additionally, advancements in synthetic biology have enabled precise genetic manipulation of electroactive bacteria to control EET pathways. By engineering genetic circuits that regulate the expression of EET-related genes, researchers can modulate electron flux in response to specific environmental cues or computational logic operations^20,21. This genetic control over EET opens avenues for creating bio-hybrid systems where bacterial computations are directly interfaced with electronic devices like OECTs. For example, bacterial genetic circuits could be designed to respond to combinations of chemical inputs, turning on or off EET pathways and thereby modulating the conductivity of an OECT channel. This would allow for direct electrical readouts of bacterial computations, bypassing the need for fluorescent or other traditional biological reporters.

Recent advances have demonstrated the potential of coupling OECTs with electroactive bacteria. For example, Méhes et al. used S. oneidensis to monitor real-time EET activity with a p-type OECT, illustrating how bacterial metabolism could be tracked electrically²². While this work highlights the possibility of using OECTs to detect bacterial activity, the potential for the hybrid system for biosensing and biocomputing remains underexplored. To address this, we recently developed hybrid transistors incorporating genetically engineered S. oneidensis into p-type OECTs²³. The results demonstrated that the OECT channel could be de-doped through bacterial EET activities (Figure 1B). To further enhance biosensing capabilities and gain mechanistic insights into the de-doping process, we engineered S. oneidensis strains with genetic circuits that regulate EET flux. This results in predictable OECT output changes in response to environmental cues such as inducer molecules. Moreover, by integrating Boolean logic into these genetic circuits, we enabled direct electrical readouts of complex bacterial computations.

Here, we present the comprehensive protocol for the hybrid transistor operation, covering the device fabrication, cell culture preparation, measurement procedures, and data analysis. Additionally, we address key considerations such as device cleaning, reusability, and the potential for automation in high-throughput testing.

Protocol

NOTE: All chemicals were used as received without further purification. If not specified, analytical-grade chemicals were used.

1. OECT device fabrication

NOTE: The OECTs are fabricated on quartz microscope slides using standard microfabrication techniques adapted from prior work²⁴. As shown in Figure 2A, eight OECT devices are arranged on a single standard slide. Pre-cut polydimethylsiloxane (PDMS) sheets are placed on the slide to form OECT chambers and access ports. The p-type conducting polymer poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT: PSS) is used for the OECT channel and the gate electrode, with dimensions of 150 µm x 10 µm for the channel and 500 µm x 500 µm for the gate tip ( Figure 2B). A summary of the OECT channel fabrication and device assembly process is provided in Figure 2C.

1. Gold electrode fabrication
   1. Clean the quartz slides by rinsing them with acetone, isopropyl alcohol (IPA), and deionized (DI) water. Blow dry with N₂ gas.
   2. Perform O₂ plasma cleaning with 50 sccm at 150 W for 120 s.
   3. Spin-coat AZ5209e photoresist (PR) with 2 stages: first at 500 rpm, 100 rpm/s acceleration for 5 s, and the second at 2000 rpm, 1000 rpm/s acceleration for 45 s.
   4. Hard bake at 90 °C for 150 s.
   5. Turn on the mask aligner (Suss MA6) and select i-line (365 nm) as the UV source, set the parameters: exposure time as 3.5 s (13 W), alignment gap at 15 µm, WEC type to Contact. and exposure type to Hard.
   6. Load the lithography mask and the slide into the aligner, align the slide to the mask, then perform the initial exposure.
   7. Reverse bake at 110 °C for 180s.
   8. Remove the lithography mask and perform flood exposure with the following parameters: exposure time as 25 s (13 W), exposure type as Flood E.
   9. Develop the PR with the AZ 400K 1:4 developer pre-diluted by the manufacturer. Pour the appropriate amount of the developer into a crystallizing dish to submerge the slide, and constantly shake the dish with circular motions for 45 s. Then, immediately remove the slide and soak it in DI water for 45 s, rinse with DI water, and blow dry with N₂.
      NOTE: The process can be paused here.
   10. Move slides into the e-beam evaporator. Pump to 2.7 x 10^-4 Pa (2 x 10^-6 T) or lower.
   11. Deposit the Ti adhesive layer (10 nm) and Au (100 nm) sequentially. The deposition speed may vary depending on the specific evaporator used. A deposition speed of 0.1 nm/s is recommended to achieve a good balance between deposition quality and processing time.
   12. Remove the excessive metal by soaking the slides in acetone overnight. After soaking, use a pipette to jet acetone streams on the slide surface to dislodge any remaining excessive metal. The quart slides obtained after the above steps are referred to as OECT electrode slides.
       NOTE: Alternatively, use sonication to assist in dislodging the excessive metal.
2. PEDOT: PSS channel fabrication
   1. Clean the OECT electrode slides by rinsing them with acetone, IPA, and DI water. Blow dry with N₂ gas.
   2. Bake the slides for at least 5 min at 160 °C. Allow the slides to cool down before spin-coating.
   3. Spin-coat AZ5209e PR with 2 stages: first at 500 rpm, 100 rpm/s acceleration for 5 s followed by 2000 rpm, 1000 rpm/s acceleration at 45 s.
   4. Hard bake at 90 °C for 150 s.
   5. Start initial exposure with i-line in hard contact mode for 3.5 s at 13 W.
   6. Reverse bake at 110 °C for 180 s.
   7. Perform flood exposure for 25 s at 13 W.
   8. Develop with AZ 400K 1:4 developer for 45 s with constant circular shaking of the dish. Then, soak the slides in DI water for 45 s, rinse with DI water, and blow dry with N₂.
      NOTE: The process can be paused here.
   9. Sonicate the PEDOT: PSS dispersion for 5 min by floating the bottle on DI water bath filled to the fill-line of the sonicator. Filter the PEDOT: PSS dispersion with a 0.22 µm filter to remove clumps.
   10. Spin-coat PEDOT: PSS at 3500 rpm with acceleration at 3500 rpm/s for 60 s. Dry at 90 °C for 15 min.
   11. Remove excessive PEDOT:PSS film by soaking the slides in acetone for 12 min, followed by 1 min sonication.
   12. Rinse the slides with acetone and then IPA and blow dry with N₂.
   13. Perform ethylene glycol (EG) treatment by placing the slides over the hot plate at 90 °C, then drop sufficient EG to cover all the PEDOT: PSS for 5 min. Rinse with DI water and blow dry with N₂. The quart slides obtained after the above steps are referred to as OECT slides.
3. PDMS device chamber fabrication
   1. Make PDMS sheets by mixing the base and curing agent at a 10:1 ratio. Cast the solution into the mold to reach a thickness of 3 mm and degas for 1 h in a 2-quarter degas chamber connected to a vacuum pump.
   2. Cure the mixture by placing the mold on a leveled bench for 48 h at room temperature to ensure smooth PDMS surfaces. Cover the mold to avoid contamination from air-borne debris.
   3. Cut out the PDMS with razor blades and create the device chambers and fluid exchange ports with circular hole punches with diameters of 5 mm and 1.5 mm, respectively. Ensure the height of the PDMS is 3 mm as controlled during mold casting (step 1.3.1)

2. Media preparation

1. Autoclave 500 mL of ultrapure water, four stainless steel dispensing needles, two rubber flask caps (for the 250 mL round-bottom flasks), two 250 mL round-bottom flasks, and one 500 mL glass media bottle. Store the sterile ultrapure water at 4 °C for future use.
2. Prepare the 2x Shewanella basal medium (SBM) amended with 2x Wolfe's Trace Mineral Mix 0.1% casamino acids according to Table 1. Filter sterilize the medium into the sterile glass media bottom (obtained in Step 2.1) with a bottle top filter. This medium is referred to as 2x SBM++, with each + indicating the addition of Trace Mineral Mix and casamino acids. Store at 4 °C for future use.
3. Filter sterile (0.22 µm filter) each of the following chemicals separately: 50 mL of 1 M sodium fumarate into a sterile 50 mL centrifuge tube; 1 mL of 2.5 mg/mL kanamycin (KAN), referred to as 100x KAN into a sterile microcentrifuge tube. Allocate 1 mL of 60% w/w sterile sodium lactate (used as received) into a microcentrifuge tube. Store all prepared solutions at 4 °C.
4. Separately prepare 1 mL of the following chemicals in sterile microcentrifuge tubes: 100 mM isopropyl ß-D-1-thiogalactopyranoside (IPTG), referred to as 100x IPTG; 10 mM 3-oxohexanoyl-homoserine lactone (OC6), referred to as 100x OC6; 1 mM anhydrotetracycline hydrochloride (aTc), referred to as 100x aTc. Store at 4 °C for future use.
5. Using the sterilized flasks, rubber caps, and needles prepared in Step 2.1, purge the medium stock. Allocate 200 mL each of sterile 2X SBM++ and ultrapure water into separate 250 mL round-bottom flasks. Seal the flasks with rubber caps and insert sterile dispensing needles to create a gas washing setup (air inlet needle submerged in the liquid; gas outlet needle above the liquid surface). Purge with N₂ for 15 min.
6. Transfer the purged 2x SBM++ and ultrapure water, lactate, 1 M sodium fumarate, 100x KAN, 100x IPTG, 100x OC6, and 100x aTc into the glovebox for further use with the OECT.

3. Monitoring cellular EET with OECT devices

NOTE: To ensure anaerobicity, the OECT experiment is conducted in a glovebox without humidity control at room temperature.

1. Day 1
   1. Device sterilization and preparation
      1. Autoclave the OECT slides, PDMS sheets, and Ag/AgCl reference electrode (RE) in separate containers. Dry by baking at 80 °C. Bring the autoclaved OECT slides and Ag/AgCl RE into the glovebox.
      2. Desorb trapped oxygen in the PDMS sheets by subjecting it to the vacuum in the antechamber of the glovebox for 4 h. Assemble the OECT device by placing the PDMS sheets over the OECT slides in the glovebox.
      3. Inject 45 µL of appropriate SBM medium into each OECT chamber. For most experiments, use 1x SBM++ supplemented with 20 mM lactate. For example, prepare 1 mL of 1x SBM++ by mixing 500 µL of 2x SBM++, 497 µL of sterile ultrapure water, and 2.85 µL of 60% w/w sodium lactate.
      4. Cover the OECT chamber ports with PDMS sheets to avoid medium evaporation and contamination.
      5. Monitor the OECT channel current I_DS under a constant channel voltage V_DS of -0.05 V and gate voltage V_GS of 0.2 V until the current stabilizes. The voltage selection can vary depending on the OECT device or biological samples. Use the same continuous bias voltages for the gate and channel throughout the measurement. Stabilization may take hours to a few days, depending on the vacuum level and desorption duration in the antechamber.
   2. Cell cultures
      1. Streak cell stocks (stored at -80 °C) onto agar plates containing LB (for non-plasmid-harboring strains) or LB with 25 µg/mL kanamycin (for plasmid-harboring strains) and grow aerobically overnight at 30 °C^20,23.
2. Day 2
   1. Cell preparation (aerobic cultures)
      1. Randomly pick single colonies from the LB agar plates (obtained in step 3.1.2.1) in 2 mL of 1x SBM++ supplemented with 20 mM lactate in 15 mL culture tubes. Keep cells overnight at 30 °C and 250 rpm shaking. Prepare 6 mL growth medium without cell as abiotic control which will be used later in cell washing.
   2. Cell preparation (anaerobic culture)
      1. Transfer LB agar plates containing the strains to be tested into the glovebox.
      2. Randomly pick single colonies from LB agar plates in 1 mL of appropriate SBM supplemented with 20 mM lactate and 40 mM fumarate in 15 mL culture tubes. For example, prepare 1 mL of medium for IPTG inducible strain requiring 1 mM IPTG and 25 µg/mL of KAN by mixing 500 µL of 2x SBM++, 437 µL of sterile ultrapure water, 2.85 µL of 60% w/w sodium lactate, 40 mM of 1 M sodium fumarate, 10 µL of 100 mM IPTG, and 10 µL of 2.5 mg/mL KAN.
      3. Keep cells at 30 °C for 18 - 24 h without shaking. To ensure comparability across experiment batches, maintain consistent growth times.
3. Day 3
   1. Initial OECT electrochemical measurement
      NOTE: The type of OECT characterizations varies from static types, such as constant voltage bias, to dynamic types, such as gate pulsing. Here, we use the gate voltage step test to obtain both dynamic and static characteristics. Reconfiguring the electrical connection between OECTs and the instrument (a multichannel potentiostat in this case) is assisted by customized manual multiplex (MUX) circuit boards. Automated MUX boards can be developed for high-throughput measurement.
      1. Measure the OECT channel current I_DS corresponding to the gate voltage step from 0 V to 0.2 V with the settings described below (exact names may vary depending on the instrument).
      2. Instrument channel 1 measures the OECT channel current I_DS. For this, set Technique name: Fast Amperometry; Equilibration time: 1 s; Equilibration voltage: - 0.05 V; Bias voltage: -0.05 V; Run time: 14 s; Sample interval: 0.5 ms.
      3. Instrument channel 2 controls the OECT gate voltage V_GS. For this, set Technique name: Mixed Mode; Condition time: 1 s; Condition voltage: 0 V; Stage 1 mode: Constant E; Stage 1 bias voltage: 0 V; Stage 1 run time: 4 s; Stage 2 mode: Constant E; Stage 2 bias voltage: 0.2 V; Stage 2 run time: 10 s; Sample interval: 0.5 ms.
      4. Repeat the gate voltage step test (steps 3.3.1.1 to 3.3.1.3) for all the OECTs. This step can be automated for high-throughput testing.
      5. Apply constant channel voltage V_DS of -0.05 V and gate voltage V_GS of 0.2 V to all the OECTs until inoculation. Ensure that the bias voltages are the same as the ones used during device stabilization (step 3.1.1.5).
   2. Inoculum preparation (aerobic cultures)
      1. Triple-wash the cells by centrifuging at 1503 x g for 4 min and resuspending in 1 mL of fresh medium used for the cell growth (step 3.2.1.1). After the last (3^rd) spin, resuspend the cells in 0.5 mL (or half the original culture volume) to obtain concentrated cell suspension at OD₆₀₀ of 1 - 3.5. Adjust the final suspension volume accordingly.
      2. Transfer the cell suspension into the glovebox.
      3. To obtain the inoculum, dilute the cells to an intended OD₆₀₀ of 0.1 (or 10x the inoculation OD₆₀₀). The delusion medium has the same constitution as the growth medium but prepare it fresh in the glovebox with purged media stocks. For example, if cells were aerobically grown in SBM++ supplemented with lactate, prepare fresh SBM++ supplemented with lactate in the glovebox for the inoculum dilution.
      4. Stop the potentiostat. Inject 5 µL of the inoculum into the OECT chamber (containing 45 µL of appropriate SBM) to achieve a final OD₆₀₀ of 0.01 in the OECT. Cover the OECT chamber ports with PDMS sheets to avoid medium evaporation and contamination.
   3. Inoculum preparation (anaerobic cultures)
      1. To obtain the inoculum, dilute the cell cultures 10-fold. Use a delusion medium that has the same constitution as the growth medium but lacks the fumarate. For example, if cells were anaerobically grown in SBM++ supplemented with lactate, fumarate, KAN, and IPTG, prepare fresh SBM++ supplemented with lactate, KAN, and IPTG in the glovebox for the inoculum dilution.
      2. Stop the potentiostat. Inject 5 µL of the inoculum into the OECT chamber (containing 45 µL of appropriate SBM). Cover the OECT chamber ports with PDMS sheets to avoid medium evaporation and contamination.
   4. Continuous measurement and timepoints
      1. Apply constant bias voltages to the OECT channels (V_DS = -0.05 V) and gate (V_GS = 0.2 V) throughout the experiment, except during timepoint measurements when the potentiostat is connected to individual OECTs for characterization.
      2. Run the experiment for 24 h after inoculation. Consider the measurement before inoculation (see step 3.3.1) as the 0 h timepoint.
      3. Significant changes in the OECT channel doping state typically occur during the initial 8 h. Select 5-7 time points for measurements during this period, followed by one measurement between 12 h and 16 h, and the final time point at the 24 h mark. Example time points could include immediately after inoculation and at 1 h, 2 h, 3 h, 4.5 h, 6 h, 8 h, 18 h, and 24 h time points post-inoculation.
      4. To avoid late-night measurements, start the inoculation (see step 3.3.2 or 3.3.3) early in the day to accommodate the frequent timepoint measurements during the first 8 h.
      5. Repeat the step outlined in 3.3.1 for each timepoint measurement.
4. Day 4
   1. Transfer curve measurement at the last timepoint (24 h)
      1. Insert the Ag/AgCl RE into the OECT chamber by gently twisting and pushing the RE through the fluid exchange port.
      2. Measure the transfer curves of the OECT by sweeping the gate voltage from -0.1 V to 0.6 V while monitoring the channel current I_DS with constant bias V_DS at -0.05 V. Measure the accurate gate and source electrode potentials against the Ag/AgCl RE.
      3. Configure the instrument as described below (exact names may vary depending on the instrument).
      4. Instrument channel 1 measures OECT channel current I_DS, use the following setting: Technique name: Chronoamperometry (CA), Equilibration time: 5 s, Equilibration voltage: - 0.05 V, Bias voltage: -0.05 V, Run time: 35 s, Sample interval: 0.09915 s.
      5. Instrument channel 2 controls the OECT gate voltage V_GS, use the following setting: Technique name: Linear Sweep Voltammetry (LSV), Equilibration time: 5 s, Equilibration voltage: - 0.1 V, Begin voltage: -0.1 V, End voltage: 0.6 V, Voltage step: 0.002 V, Scan rate: 0.02 V/s.
      6. Instrument channel 3 measures OECT source potential V_S against Ag/AgCl RE, use the following setting: Technique name: Open Circuit Potentiometry (OCP), Run time: 40 s, Sample interval: 0.09915 s.
      7. Instrument channel 4 measures OECT gate potential V_G against Ag/AgCl RE, use the following setting: Technique name: Open Circuit Potentiometry (OCP), Run time: 40 s, Sample interval: 0.09915 s.
      8. Repeat the transfer curve measurement for all OECTs. After each measurement, rinse the Ag/AgCl reference electrode (RE) with 70% ethanol and wipe it with a new low-lint towel. The Ag/AgCl RE can be reused within the same sample group for the triplicates but should not be reused between different sample groups.

4. Data analysis

1.  Data extraction
   1. Use custom MATLAB code for data processing to automate the extraction and logging of data from the instrument's native software, significantly reducing processing time. MATLAB scripts are available in the Texas Data Repository (https://doi.org/10.18738/T8/MNKO8D).
2. Data processing
   1. Fit the OECT channel current I_DS data to obtain the rate constants k for further analysis and comparison (Figure 3). For EET-capable cell samples, we find the single exponential model provided the best fit. Therefore, we used the one-phase exponential decay model in analysis software to fit the I_DS time series data and obtain the fitted rate constant k. This I_DS time series data can be the raw value or normalized to the 0 h timepoint, as the rate constant k reflects for the rate of change rather than the absolute values.
   2. Obtain the normalized I_DS by dividing the I_DS values by the 0 h timepoint value (I_DS0), referred to as I_DS/I_DS0.
3. Determining rate constant k
   1. Open the analysis software (Prism 10, GraphPad) and create a new project by clicking File > New > New Project File. Under the Create tab, choose XY. In the Options section, set X to Numbers and Y to Enter and plot a single Y value for each point.
   2. Import the OECT channel current I_DS data to the table. Fit the I_DS data with the one-phase exponential decay model by clicking Analyze in the Analysis tab. Find the Nonlinear regression (curve fit) method in XY analyses. Make sure to select All the Data Sets that Need to be Fitted on the Right, then click OK.
   3. In the new Parameters window, select One phase decay model in the Exponential group, then click OK. The model equation is as follows:
      Y=(Y0 - Plateau)*e^-k*X + Plateau
      ​Where: Y is the OECT channel's current I_DS, T is time, k is the rate constant.
   4. Once the k values are obtained, review the fit quality by checking the R squared value for each fit. A good fitting should have an R² higher than 0.95.
4. Unpaired two-tailed Student's t-test
   1. To assess the statistical significance between sample groups, perform unpaired two-tailed student's t-tests.
   2. Open the analysis software and create a new project by clicking File > New > New Project File. Under the Create tab, choose XY. In the Options section, set X to Numbers and Y to Enter 3 replicate values in side-by-side subcolumns.
   3. Import the fitted rate constant k values into the table. Perform the t-test by clicking Analyze in the Analysis tab. Find the t-test (and nonparametric tests) method in Column analyses. Make sure to select Two Data Sets Each Time, then click OK.
   4. In the Experimental Design manual, select Unpaired, Yes. Use parametric test and Unpaired t-test. Assume both populations have the same SD options. The p-values will be generated for each comparison. In figures, p-values are represented as follows: n.s.: p > 0.05 (not significant), *: p ≤ 0.05, **: p ≤ 0.01, ***: p ≤ 0.001, ****: p ≤ 0.0001.
   5. Repeat steps 4.4.3 -4.4.4 for all relevant data sets.

5. OECT device cleaning and reusability

1. After the experiment, disassemble the device and boil the OECT slides in soapy water (1:10, soap: DI water) for 15 min.
2. Quickly move the slides into a beaker containing room-temperature DI water and wait 5 min for the slides to cool down. Rinse the slides with DI and blow dry with clean air.
3. Measure the resistance of the OECT channel with a multimeter. If the channel resistance is within 400% of the original value, the device can be reused for the experiment after sterilization (Step 3.1.1). The OECT slides can also be cleaned using the following steps and are suitable for channel fabrication using Step 1.2.
   CAUTION: Piranha solution is highly corrosive and can generate significant heat. Always use appropriate personal protective equipment (PPE), including gloves, goggles, and a lab coat. Handle with extreme care and obey institution's safety guidance.
4. Prepare the piranha solution in a well-ventilated fume hood. Carefully and slowly add 30% (w/w) hydrogen peroxide to 98% sulfuric acid in a 1:4 ratio while stirring by gently tilting the crystallizing dish back and forth.
5. Place the crystallizing dish on a hot plate heated at 70 °C. Use tweezers to bring and submerge the slides in piranha solution for 1 min.
6. Carefully transfer the slides to a beaker containing DI water and allow them to sit for at least 1 min. Thoroughly rinse the slides with DI water and blow dry with air.
7. Dispose of the piranha solution according to the institutional chemical waste disposal guidelines. Neutralize the piranha solution by slowly adding it to a beaker containing an ice-water mixture about 4x the volume of the solution. Gradually add a neutralizing agent (e.g., 10 M NaOH) while monitoring the pH with test strips until neutral.

Results

Fitted rate constant k
The fitted rate constant k of the OECT channel current I_DS serves as a reliable metric for assessing the EET activity of the sample. While the constant gate bias voltages would affect the rate constants, we choose 0.2 V to ensure positive bias to encourage bacterial electron transfer while avoiding fast de-doping at higher gate voltages and providing and minimizing electrochemical stress on the bacteria cells²³. The drain voltage -0.05 V is selected to ensure reliable, current measurement with minimal electrochemical stress on the bacteria cells. Under these conditions, the fitted rate constants highlight significant differences in EET activity between strains. For example, the fitted rate constants for the EET-deficient strains ΔmtrC and ΔMtr were significantly lower than those of the wild-type S. oneidensis MR-1 (Figure 4A,B). To restore EET flux, the ΔmtrC and ΔMtr mutants were complemented by incorporating mtrC and mtrCAB Buffer gates, which transcribe the corresponding EET genes in response to the inducers IPTG and OC6, respectively. The OECT channel current data shown in Figure 4C showed the distinctive profile between the induced and uninduced samples. Further analysis of the fitted rate constant k (Figure 4D) revealed that the induced strains achieved rates comparable to S. oneidensis MR-1, with a statistically significant difference compared to their uninduced counterparts. However, because data points are collected manually, the accuracy of the fitted rate depends on the frequency of measurements. Due to the rapid changes in I_DS during the first 8 h of the experiment, we recommend more frequent measurements during this period (Figure 4C,D). As shown in Figure 4A,B, insufficient data points can result in poor fitting quality, which may hinder comparability between independent experiments. Nevertheless, within a single experiment, the fitted rate constants remain useful for comparing EET activity across different samples.

Scalability and consistency of hybrid transistor systems in biocomputing demonstrations
Comparing results from independent experiments is a critical feature of the hybrid transistor system, enabling scalability and high-throughput testing. In this biocomputing demonstration, we employed previously developed NAND and NOR transcriptional logic gates to regulate EET flux in response to combinations of inducer stimuli²⁰. The two-input NAND and NOR gates control mtrC gene expression using customized sensing modules responsive to common small molecule inducers, IPTG and OC6 for the NAND gate, and OC6 and aTc for the NOR gate. These genetic gates serve as foundational components for constructing more complex genetic circuits, allowing precise control of gene expression in response to diverse stimuli. Due to the limitation on the number of OECTs that can be run per experiment, the data presented in Figure 5A,B were obtained from three and two independent experiments, respectively. As PEDOT OECTs are highly sensitive to both ionic and redox changes, maintaining tight control over the inoculum is essential to ensure the quality and comparability of results. To achieve this, precise control of cell density (OD₆₀₀) in the inoculum and a two-step dilution process is employed. For aerobically grown cultures (OD₆₀₀ ranging from 1 to 3.5), an initial dilution is performed to reduce the cell density (e.g., 0.1) to 10 times the final desired OD600 (e.g., 0.01). This is followed by a further 10-fold dilution upon inoculation, resulting in a total dilution of at least 100-fold. This step is critical to minimize the carry-over of dissolved oxygen from the inoculum into the OECT. For anaerobically grown cells, where determining precise cell density is more challenging, inoculum control is achieved by strictly regulating the growth time. A 100-fold dilution is used to prevent the carry-over of electron acceptors, such as fumarate, which may interfere with the OECT measurements.

The consistency and comparability of these results also attributable to the stringent control of experimental conditions, such as inoculum cell density and the number and timing of measurements, ensuring that the inducer concentrations were the primary factor influencing the rate constant k in these experiments.

Fast response from OECT channel current
Another advantage of the hybrid OECT system is the fast response of the OECT channel current to EET activity, originating from the high sensitivity of the PEDOT: PSS to both ionic and redox changes. As shown in Figure 6A,B, the channel currents of the ΔmtrC+mtrC strains were extracted from Figure 4C at the 10 min and 1.5 h timepoints, respectively. Statistically significant differences between the strains are detectable 1.5 h post-inoculation. The relatively small error bars are a result of normalizing the channel current I_DS to its initial value prior to inoculation I_DS0, which effectively reduces variability caused by differences in device fabrication. For broader adoption, standardization of experimental conditions such as inoculum density, incubation time, and environmental parameters (e.g., temperature and humidity) will be critical to ensure reproducibility and consistency across devices. Additionally, incorporating abiotic reducing agents into the OECT during experiments could be a reliable standardized benchmark for comparative analysis. Although we have not optimized or explicitly utilized the current change (I_DS/I_DS0) data for rapid testing, we believe the inherent high sensitivity of PEDOT: PSS combined with the internal referencing of the channel current make the hybrid OECT system well-suited for rapid and high-throughput testing applications.

Figure 1: Schematic illustration of extracellular electron transfer (EET) and the de-doping mechanism of organic electrochemical transistor (OECT) mediated by cellular EET. (A) The MtrCAB pathway facilitates electron transfer to extracellular electron acceptors. (B) Channel de-doping mechanisms through direct EET. Metabolic electrons can be transferred either directly to the PEDOT: PSS channel or indirectly to the gate, followed by transfer to the source electrode and the channel through external circuits. Cations in the electrolyte are associated with the negatively charged poly(styrenesulfonate) (PSS-) to neutralize the positively charged poly(3,4-ethylenedioxythiophene) (PEDOT+). Created with BioRender.com. This figure has been modified from²³. Please click here to view a larger version of this figure.

Figure 2: Photo and microscopy images of the organic electrochemical transistor (OECT). (A) The eight OECTs on a microscope slide with PDMS sheets to form the OECT chambers, and (B) the top view of a single OECT. (C) Cartoon illustration of the major OECT channel fabrication and assembly steps. Created with BioRender.com. This figure has been modified from²³. Please click here to view a larger version of this figure.

Figure 3: Organic electrochemical transistor (OECT) output changes induced by S. oneidensis MR-1. (A) The I_DS/I_DS0 curves of S. oneidensis MR-1 inoculated OECTs with varying final cell optical density (OD₆₀₀) values. (B) Corresponding fitted rate constants. Data show the mean ± SD of 3 biological replicates, unpaired two-tailed Student's t-tests were performed without adjustments for multiple comparisons, n.s. represents p > 0.05. This figure has been modified from²³. Please click here to view a larger version of this figure.

Figure 4: Organic electrochemical transistor (OECT) responds to different S. oneidensis mutant strains. (A) The I_DS/I_DS0 curves of ΔmtrC, ΔMtr, and MR-1 strains with final cell OD₆₀₀ of 0.1. (B) Corresponding fitted rate constants. (C) The I_DS/I_DS0 curves of ΔmtrC and ΔMtr strains complemented with mtrC and mtrCAB Buffer gates. (D) Corresponding fitted rate constants. Data show the mean ± SD of 3 biological replicates, unpaired two-tailed Student's t-tests were performed without adjustments for multiple comparisons, n.s. represents p > 0.05. Panels A, B, and D have been modified from²³. Please click here to view a larger version of this figure.

Figure 5: Organic electrochemical transistor (OECT) responds to strains carrying Boolean logic gates. Fitted rate constants as a function of combinatorial inducer concentrations for ΔmtrC mutants carrying (A) NAND and (B) NOR Boolean logic gates. Data show the mean ± SD of 3 biological replicates. Created with BioRender.com. This figure has been modified from²³. Please click here to view a larger version of this figure.

Figure 6: Channel current measurements from organic electrochemical transistor (OECT) inoculated with different S. oneidensis strains. Channel current IDS/IDS0 from ΔmtrC knockout strains with induced and uninduced mtrC Buffer gates extracted at (A) 10 min and (B) 1.5 h timepoints. Data show the mean ± SD of 3 biological replicates, unpaired two-tailed Student's t-tests were performed without adjustments for multiple comparisons, n.s. represents p > 0.05. Please click here to view a larger version of this figure.

Figure 7: Gate current measurements from organic electrochemical transistor (OECT) inoculated with different S. oneidensis strains. (A) Gate current I_GS from ΔmtrC knockout strains with induced and uninduced mtrC Buffer gates, measured 24 h post-inoculation. The shaded region indicates the range of standard deviation. Extracted I_GS at V_GS of (B) 0.3 V and (C) 0.4 V. (D) Gate current I_GS from ΔMtr knockout strains with induced and uninduced mtrCAB Buffer gates, measured 24 h post-inoculation. The shaded region indicates the range of standard deviation. Extracted I_GS at V_GS of (E) 0.3 V and (F) 0.4 V. Data show the mean ± SD of 3 biological replicates, unpaired two-tailed Student's t-tests were performed without adjustments for multiple comparisons, n.s. represents p > 0.05. Panel D has been modified from²³. Please click here to view a larger version of this figure.

Ingredient	Quantity per 1 L
K2HPO4	225 mg
KH2PO4	225 mg
NaCl	460 mg
(NH4)2SO4	225 mg
MgSO4.7H2O	117 mg
HEPES	100 mL of 1 M HEPES
Casamino acids	5 mL of 10% casamino acids in ddH2O, if needed
Wolfe's Mineral Mix	5 mL of Wolfe's Mineral Mix, if needed
ddH2O	Adjust volume to 1 L and pH to 7.2

Table 1: Shewanella Basal Medium (SBM) formulation.

Discussion

Electrochemical Cell (EC) Comparison
One major advantage of fitting the rate constant from the OECT channel current is that it minimizes device variation by focusing on the underlying dynamics of I_DS changes rather than the raw output. Combined with the inherent signal amplification capability of OECTs, this approach enhances the robustness of hybrid OECT systems compared to traditional electrochemical cells (EC). For example, at the end of each 24-hour OECT experiment, transfer curves are measured by sweeping the gate electrode voltage V_GS from -0.1 to 0.6 V. This is achieved by connecting the gate to the working electrode (WE) and the source to the counter electrode (CE) and reference electrode (RE), forming a two-electrode EC between the gate and source, with the measured I_GS being the WE current. Thus, I_GS represent the conventional EC results that can be compared to the fitted rate constants of the OECT in the same experiment. As demonstrated in Figure 7A, the I_GS of induced and uninduced mtrC+mtrC strains had more overlap than the Mtr+mtrCAB strains (Figure 7D). To further distinguish the WE current, I_GS was extracted at V_GS of 0.3 V and 0.4 V. At V_GS = 0.3 V, the I_GS data of ΔmtrC+mtrC strains showed no discernible difference between induced and uninduced samples (Figure 7B), and a similar result to that of the OECT (Figure 4D) was obtained at V_GS = 0.4 V (Figure 7C). On the other hand, the I_GS data for ΔMtr+mtrCAB strains (Figure 7E,F) consistently demonstrated distinguishable differences between induced and uninduced strains, though only at V_GS = 0.4 V did I_GS data show comparable statistical significance to the OECT rate constants.

These results suggest that, under the same conditions, caution is needed when interpreting EC data, as different conclusions could be drawn from the same dataset depending on the measurement conditions (WE bias potentials). We attribute this variability to the inherent noise in the EC tests. Contributing factors include device variations such as electrode surface defects, differences in electrode surface area (due to the hand-cut PDMS chambers that may not perfectly center the OECT electrodes), and weak biological signals due to the lack of biofilm formation during the 24 h test period. Conversely, under the same experimental conditions, the OECT channel current-derived rate constants proved more reliable results by avoiding many of the shortcomings seen in the EC tests.

Automatic measurement
Automation is a major potential improvement to this protocol. Currently, measurements must be taken manually at each time point. This process could be automated using an auto MUX circuit, such as relays controlled by a microcontroller unit (MCU)²⁵. By integrating the MCU with the potentiostat, the measurement resolution would be significantly enhanced, as the instrument could cycle through the OECTs continuously at much shorter intervals and more consistency than manual operations. Furthermore, since the raw OECT channel current typically ranges from tens to thousands of µA, cost-effective potentiostat or multichannel data loggers could be used or developed to continuously measure all OECT channels without the need for MUX circuits^26,27. Coupling automated measuring systems with scalable OECT arrays, such as those in 96-well plate formats²⁸, could transform the hybrid OECT system into a high-throughput testing platform compatible with standard biological experimental setups, significantly broadening the applicability of the OECT platform.

Device sealing and test conditions
One limitation of the current setup is oxygen sealing. Although PDMS is a versatile and user-friendly material for microfluidics and device fabrication, it is permeable to oxygen. This is why the OECTs are placed in anaerobic environments during experiments, such as a glovebox. The development of compatible, non-oxygen-permeable seals or the integration of oxygen scavenger systems with this OECT platform would significantly expand its applicability, enabling its use in aerobic conditions.

Temperature is another critical factor influencing the experimental results. Since bacterial metabolism is highly temperature-dependent, maintaining precise temperature control over the device could potentially shorten the experiment duration by optimizing cell activity at their ideal growth temperature. This control can be achieved using simple heater coils or thermoelectric components, such as Peltier modules, which provide both heating and cooling capabilities.

Scope of application and device design
The OCET design used in this protocol allows direct contact between living cells and the electrical components: electrodes and channels. We believe that this direct contact simplifies device fabrication and operation, making it suitable for general sensing applications with other bacteria strains and other living components such as enzymes and eukaryote cells. However, if direct contact with the electrical components needs to be avoided due to issues such as electrochemical stimuli and stress from the electrode or material incompatibility, it may be advantageous to segregate the biological and electrical components. While our current approach does not isolate bacteria to a specific region of the device, future OECT designs could achieve this through the use of light-patternable biofilms²⁹, ion-permeable membranes, or alternative device architectures such as floating gates^30,31. Additionally, exploring new channel materials such as n-type or inorganic semiconductors and introducing electron shuttles to facilitate electron transfer^32,33 could provide valuable insights into the redox behaviors of biological components and expand device design possibilities by enabling more targeted biological-electrical interactions.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
3-oxohexanoyl-homoserine lactone	Sigma-Aldrich
anhydrotetracycline hydrochloride	VWR
casamino acids	VWR
Equipment
Ethylene glycol	Sigma-Aldrich		anhydrous 99.8%,
HEPES buffer solution	VWR		1 M in water, pH = 7.3
isopropyl ß-D-1-thiogalactopyranoside	Teknova
kanamycin sulfate	Growcells
Magnesium(II) sulfate heptahydrate	VWR
PEDOT:PSS aqueous suspension	Heraeus Epurio LLC	Clevios PH1000
Potassium phosphate dibasic	Sigma-Aldrich
Potassium phosphate monobasic	Sigma-Aldrich
Potentiostat	PalmSens BV	MultiPalmSens4
Quartz microscopic slides	AdValue	FQ-S-003
Quartz microscopic slides
Sodium chloride	VWR
Sodium DL-lactate	TCI		60% in water
Sodium fumarate	VWR		98%
Sulfuric acid	Sigma-Aldrich		95.0%-98.0%
Two-part silicone elastomer	Electron Microscopy Sciences	Sylgard184
Wolfe's Trace Mineral Mix	ATCC