Electrochemiluminescence microscopy for real-time, single-cell imaging of surface charge in electroactive bacteria

Abstract

Understanding how applied voltage regulates metabolism in microbial electrochemical systems is challenging due to the lack of adequate electrophysiological tools for microorganisms. Here, we present an imaging platform based on electrochemiluminescence (ECL) that allows real-time, femtocoulomb-scale quantification of surface charge in single bacterial cells. The method exploits the electrostatic enrichment of cationic luminophores at bacterial membranes to amplify ECL signals, thereby linking surface charge dynamics to intracellular electron metabolism and interfacial electron flow. Using Shewanella oneidensis MR-1 as a model, we show voltage-activated metabolic enhancement mediated by outer-membrane cytochromes. By modulating direct and indirect electron-transfer pathways, we reveal that cytochrome-dependent direct electron transfer initiates activation, while mediator-enabled indirect electron transfer prevents charge saturation and sustains elevated metabolic turnover. Furthermore, dual-parameter screening enables identification of bacterial subpopulations with high metabolic activity and efficient electron transfer capability. Thus, our work provides a framework for microbial electrophysiology at the single-cell level and opens potential avenues for engineering high-performance electroactive strains for bio-electrochemical applications.

Our understanding of how applied voltage regulates metabolism in microbial electrochemical systems is limited. Here, Xing et al. present an imaging platform based on electrochemiluminescence that allows real-time, femtocoulomb-scale quantification of surface charge in single bacterial cells.

Introduction

Microbial electrochemical systems (MESs) have emerged as a transformative platform bridging microbiology, electrochemistry, and energy science. By exploiting the ability of electroactive microorganisms to exchange electrons with external electrodes, MESs enables sustainable applications such as bioelectricity generation, bioremediation, and biosensing^1–4. A central determinant of MES performance is extracellular electron transfer (EET), the process by which microbes couple intracellular metabolism to external redox reactions. Applied external voltage serves as a critical tool for regulating EET, thereby influencing overall system efficiency. To gain deeper insights into voltage-driven metabolic behavior, single-cell electrophysiological tools are required, because conventional macroscopic techniques provide only ensemble-level information that obscures cellular heterogeneity and dynamic and spatiotemporal behaviors^5,6.

Current single-cell electrophysiological approaches rely primarily on micro/nanoelectrodes^7–9 and optical microscopy^10–14. While these techniques provide insights into microbe-electrode interactions, they do not capture upstream intracellular metabolic pathways responsible for electron generation and consumption, which are equally essential for a comprehensive understanding of microbial EET. Thus, a single technique capable of monitoring both intracellular electron metabolism and interfacial electron transfer would represent a major advance. Because electron generation, transport, and release necessarily manifest as transient changes in surface charge, real-time tracking of bacterial surface charge offers a direct electrophysiological proxy for both processes. However, the total surface charge of a single bacterium is on the femtocoulomb (fC) scale¹⁵, far below the detection limit of conventional methods such as electrochemical impedance spectroscopy. Scanning probe-based approaches, including Kelvin probe force microscopy¹⁶, can reach higher sensitivities but are limited by low throughput. Thus, new modalities capable of resolving subtle charge fluctuations in living bacteria with high spatiotemporal precision are urgently needed.

Electrochemiluminescence (ECL) imaging presents an attractive solution. As a light-emitting process initiated by electrochemical reactions¹⁷, ECL offers exceptional spatiotemporal sensitivity¹⁸. Recent advances have extended ECL microscopy to visualize biological and chemical processes at the single-cell^19,20, single-particle^21,22, and even single-molecule levels²³, with nanometer spatial resolution and millisecond temporal precision. Recently, Feng and co-workers extended ECL imaging to single bacteria, demonstrating that the combination of single-molecule ECL imaging with machine learning can achieve morphology-based species identification²⁴. However, ECL imaging has not been explored as a tool for microbial electrophysiology. Importantly, because ECL directly transduces electrochemical events into optical signals without external excitation, it minimizes photodamage and enables long-term observation of live cells²⁵. These advantages position ECL imaging for capturing the real-time surface charge dynamics of single electroactive bacteria.

Previous studies in electro-fermentation have shown that electrodes and applied potentials can alter microbial metabolic fluxes and product distributions. However, the mechanisms linking extracellular electron exchange to intracellular metabolic regulation remain poorly understood^26–28. Here, we report a noninvasive and high-throughput single-bacterium surface charge quantification platform based on ECL imaging that enables dynamic tracking of external voltage-driven metabolic behaviors in living microbes. The method exploits electrostatic enrichment of cationic luminophores at negatively charged bacterial membranes and correlates the resulting optical signals with femtocoulomb-scale surface charges. Using S. oneidensis MR-1 as a model, we uncovered a voltage-activated metabolic enhancement mechanism mediated by outer-membrane cytochromes, whereby external electrical stimulation accelerates metabolism toward a saturation state. Based on this discovery, we established a strategy to regulate metabolic rate through the synergistic use of direct and indirect electron transfer (IET) pathways. Moreover, we revealed that the introduction of electron shuttles not only promotes long-range EET processes but also plays a critical role in enhancing the efficiency of short-range EET. Finally, by applying a dual-parameter screening approach, we simultaneously assessed metabolic activity and interfacial electron transfer efficiency to identify high-performance bacterial subpopulations. This work establishes surface-charge electrophysiology as a new paradigm for probing voltage-metabolism coupling in single living microbes and offers new opportunities for engineering electroactive strains for high-performance MESs.

Results

Principle of surface charge imaging for individual bacteria

Bacterial surfaces inherently carry substantial negative surface charges, which generates local electrostatic fields that facilitate the adsorption of cationic ECL luminophores. This local electrostatic enrichment remarkably enhances ECL emission at bacterial surfaces relative to electrode regions^29,30. By recording the spatiotemporal evolution of ECL signals from single-bacterium regions, we achieve in-situ, dynamic tracking of bacterial surface charge states. S. oneidensis MR-1 was selected as a representative electroactive strain to demonstrate the working principle (Fig. 1a) (the measurement apparatus is shown in Supplementary Fig. 1). Bacteria were immobilized on the electrode by natural sedimentation without chemical fixation or adhesive coating, and ECL imaging was performed under applied potentials ranging from 0.85 to 1.40 V (vs. Ag/AgCl) using low concentrations of Ru(bpy)₃²⁺ luminophores and DBAE co-reactants, thereby maintaining a balance between signal strength and bacterial viability (Supplementary Fig. 2). This result also confirms that the applied short ECL-triggering pulses does not cause detectable damage to the bacteria, although it is much higher than is needed for EET studies.

Fig. 1. Principle of surface charge imaging for individual bacteria.

a Illustration of individual bacterial surface charge mapping technique. b ΔI/I_B diagram of a single S. oneidensis MR-1 bacterium at different potentials. Scale bar: 1 μm. The calibration bar is ΔI/I_B value. Electrolyte is 2 mM PB (pH 7.0) containing 0.1 mM Ru(bpy)₃²⁺ and 1 mM DBAE. The exposure time is 1 s. c Statistical analysis of difference on ΔI/I_B among bacteria incubation in solutions with different PB concentrations (n = 22). Data are presented as mean ± s.d. d Surface charge distribution of individual bacteria (n = 100).

Representative normalized ECL snapshots acquired during potential scans are displayed in Fig. 1b. Below 1.05 V, no detectable emission was observed, suggesting that the low oxidation potential (LOP) pathway of the Ru(bpy)₃²⁺-DBAE system did not contribute under our experimental conditions (Supplementary Fig. 3). At 1.10 V, brighter ECL signal emerged at bacterial surfaces compared with the surrounding electrode. However, at potentials above 1.10 V, the normalized bacterial ECL signal gradually decreased due to competition with the ECL reaction at the electrode (Supplementary Fig. 4). Furthermore, time-resolved ΔI/I_B imaging under constant potential confirmed that bacterial-localized enhanced ECL signals reach a rapid steady state and remain stable over 0–5 min across 0.90–1.40 V, excluding slow adsorption/desorption or release of luminophores on the minute timescale (Supplementary Fig. 5). These observations suggest that bacterial membranes preferentially adsorb Ru (III) intermediates, which subsequently participate in ECL generation via an oxidative-reduction pathway (Supplementary Fig. 3b)²⁴. This conclusion was further corroborated by comparing the fluorescence and ECL imaging results of the bacteria with those of Ru(bpy)₃²⁺-doped silica nanoparticles (Supplementary Figs. 6 and 7).

To verify that the observed signal enhancement arises from bacterial surface charge, we varied the ionic strength of the medium to modulate the Debye length. As shown in Fig. 1c, increasing ionic strength shortened the Debye length and progressively attenuated the bacterial ECL enhancement (see also in Supplementary Fig. 8). The positive relationship between Debye length and signal enhancement confirms that the observed ECL amplification originates from electrostatic interactions between the bacterial surface charge and the cationic ECL luminophores. Furthermore, we replaced the cationic Ru(bpy)₃²⁺-DBAE system with the electrically neutral luminophore L012 and H₂O₂ as its co-reactant. The results showed no bacterial-localized ECL enhancement in the L012-H₂O₂ ECL system, supporting the bacterial ECL enhancement reported here arises from electrostatic enrichment of cationic luminophores at negatively charged bacterial envelopes (Supplementary Fig. 9). On this basis, we established a quantitative relationship between ECL signal and bacterial surface charge (see Supplementary Fig. 10 for details). Applying this calibration across a large population of single cells (Supplementary Fig. 11), we obtained a statistical distribution of surface charge values (Fig. 1d). The mean surface charge measured by our technique (−76 ± 23 fC) closely matched the value estimated from zeta potential-based analysis (~−82 fC, Supplementary Fig. 12). This agreement validates the quantitative accuracy of our framework and demonstrates the establishment of an ultrasensitive, non-invasive platform for femtocoulomb-scale single-bacterium charge quantification.

Surface charge evolution under external voltage stimulation

The bacterial membrane serves both as a carrier of charges and as a mediator for electron transport between the intracellular and extracellular environments, playing a pivotal role in EET. Monitoring its surface charge dynamics under applied voltage offers mechanistic insights into voltage-driven metabolism. Therefore, we examined the real-time surface charge responses of S. oneidensis MR-1 to external voltage stimulation. In the absence of applied voltage, single bacteria remained stable surface charges, excluding potential interference from factors other than the external voltage in measuring the bacterial surface charge, and demonstrating that the short voltage pulses used to trigger ECL are insufficient to perturb the surface charge of the bacteria (Supplementary Fig. 13). When a +0.2 V voltage was applied, a rapid accumulation of negative surface charge was observed, which eventually reached a plateau, suggesting a saturation limit (Fig. 2a). Conversely, under −0.2 V stimulation, the surface charge decreased before stabilization, implying a limit to electron uptake from the electrode (Fig. 2b). To assess reproducibility, we performed back-and-forth stepwise potential switching in a bacterial population (Fig. 2c). However, the population-averaged change appears moderate, this mainly reflects the substantial cell-to-cell heterogeneity in both the initial surface charge and the magnitude of voltage-induced responses. To better demonstrate this response effect, as shown in Supplementary Fig. 14, at the single-cell level, individual bacteria exhibited clear and reversible surface-charge changes upon voltages switching. In addition, bright-field tracking confirmed that the applied potential protocol does not cause bacterial detachment or movement on the electrode surface over the measurement period (Supplementary Fig. 15), excluding bacterial motion-induced errors in the extracted surface charge density. A consistent trend emerged: positive potentials promoted negative charge accumulation, whereas negative potentials induced charge depletion (Fig. 2d). In addition, it is worth noting that although S. oneidensis can form biofilms under growth conditions, the present experiments monitored freshly deposited cells over minutes in a non-growth measurement electrolyte, mature biofilm formation is not expected to contribute appreciably to the observed results. To exclude non-metabolic factors that may also contribute to the surface charges, mesoporous silica nanoparticles (MSNs) subjected to the same potential protocols showed no measurable change in surface charge (Supplementary Fig. 16), excluding voltage-triggered charging/discharging effect. External voltage may induce charge redistribution on bacterial membranes. Using tunable ECL emission layer thickness (TEL) (the detailed protocol of TEL regulation is discussed in Supplementary Fig. 17), we found bacterial ΔI/I_B increased as TEL matched cell height but declined below it (Fig. 2e), indicating uniform rather than localized charge accumulation. This confirms voltage-driven surface charge changes arise from metabolic electron transfer, not electrostatic polarization. Furthermore, glutaraldehyde-fixed bacteria that arrest metabolic activity while preserving the structural integrity of the cell envelope exhibited no charge variations under identical voltage conditions (Supplementary Fig. 18), which provides definitive evidence that the observed surface charge dynamics are metabolism-dependent and originate from intracellular electron generation rather than passive membrane charging.

Fig. 2. Surface charge evolution under external voltage stimulation.

The variations of surface charge density of three S. oneidensis MR-1 bacteria as time elapsed when continuously stimulated by (a) +0.2 V and (b) −0.2 V (vs. Ag/AgCl), respectively. Each cycle consisted of 1.1 V (vs. Ag/AgCl) applied for 2 s, followed immediately by 0.2 V or –0.2 V applied for 8 s. This cycle was repeated 30 times. c Statistical analysis of difference on surface charge density of individual S. oneidensis MR-1 bacteria upon distinct voltages stimulation. First, an open-circuit potential (0 V) was applied for 300 s and then a 1.1 V pulse was applied for 2 s to record the initial surface charge (Original state). After that, the potential was switched to either 0.2 V or –0.2 V and held constant for 300 s (n = 12). Data are presented as box-and-whisker plots overlaid with individual data points. Boxes represent the 25^th–75^th percentiles, central lines represent medians, and whiskers represent the minimum and maximum values. Open squares denote mean values. d Illustration of bacterial surface charge variations upon positive and negative voltage stimulation. e ΔI/I_B diagram of a single S. oneidensis MR-1 bacterium under different thickness of ECL emission layer. Scale bar: 1 μm. Electrolyte is 2 mM PB (pH 7.0) containing 0.1 mM Ru(bpy)₃²⁺ and 1 mM DBAE. The exposure time is 1 s. The calibration bar denotes values of ΔI/I_B.

In principle, this platform is not restricted to S. oneidensis MR-1 and can be extended to other bacterial types. To examine the generality of this technique, we further tested a representative Gram-positive bacterium, B. subtilis. Using the same ECL imaging conditions, we also observed pronounced signal enhancement at the surface of individual B. subtilis cells (Supplementary Fig. 19a), confirming that the method is broadly applicable across bacterial types, regardless of Gram-negative and Gram-positive bacterial. To further evaluate the versatility of the platform, we applied the same protocol to B. subtilis. Previous research has reported that B. subtilis also possess EET capabilities, although often weaker than those of well-known electroactive bacteria like Shewanella or Geobacter³¹. Under identical +0.2 V stimulation for 300 s, B. subtilis also showed an increase in surface charge, but the magnitude (~15%) was markedly lower than that of S. oneidensis MR-1 (~32%) (Supplementary Fig. 19b, c). These results indicate that the present method can not only visualize surface charges of distinct bacteria, but also quantitatively compare their voltage-responsive electrophysiological behaviors.

In theory, if the external voltages only activate electron transfer process at the bacteria–electrode interface, a positive potential stimulus would be expected to drive surface charges away from the bacteria, thereby reducing the amount of surface charge, whereas a negative potential would increase the overall surface charge. However, the trend we observed was entirely opposite to this expectation. Together, these results demonstrate that external voltage directly modulates bacterial electron transport and redox activity, activating interfacial electron exchange and concomitantly altering metabolic behavior. The mechanism underlying this voltage-driven metabolic behavior will be investigated in detail in the following sections.

Although electrochemical current is a standard readout in MESs, it is not suitable for our single-cell analytical platform. To optically resolve individual bacteria, cells must be sparsely distributed on the ITO electrode; under such low surface coverage and within our short measurement period, the faradaic contribution from bacteria is expected to be far below the non-faradaic background current. Moreover, any detectable current would represent an ensemble-average signal and would obscure the cell-to-cell heterogeneity that is the central aim of this work. Consistent with this rationale, under our imaging conditions, it is showed no measurable current difference between bare ITO and ITO with adhered MR-1 cells, both in phosphate buffer and in the full ECL reagent solution (Supplementary Fig. 20).

Voltage stimulation-activated bacterial metabolism

Previous studies on voltage effects at bacterial membranes have largely attributed charge changes to non-metabolic factors, such as lipid rearrangements or conformational shifts of membrane proteins^32,33. Optical imaging approaches have been limited to capturing voltage-triggered electron exchange events at the microbe–electrode interface. It has long been recognized that introducing electrodes or external potentials can shift microbial metabolic profiles and enhance metabolite production^26–28. However, these observations have largely been made at the process or community level, and the underlying molecular mechanisms remain unclear. In contrast, our single-cell measurements directly visualize how extracellular electrical stimulation is coupled to intracellular metabolic responses. Establishing this intrinsic connection between electrical stimulation and metabolism represents a new direction in EET research and offers guidance for the rational design of MESs.

The tricarboxylic acid (TCA) cycle is the central metabolic pathway for intracellular electron generation, transfer, and utilization³⁴. We therefore hypothesized that voltage-driven metabolic changes would be reflected in altered TCA cycle activity. To test this, we introduced malonate, a known TCA cycle inhibitor³⁵, and examined its effect on single-bacterium surface charge dynamics under applied voltage. When stimulated at +0.2 V (vs Ag/AgCl), untreated bacteria showed progressive negative charge accumulation (Fig. 3a), consistent with previous observations. In contrast, in the presence of malonate, the surface charge gradually decreased (Fig. 3b), indicating that inhibition of the TCA cycle prevented electron replenishment from intracellular metabolism, while pre-accumulated electrons were depleted at the anodic electrode. Similar opposing trends were observed under –0.2 V stimulation (Fig. 3c, d). Statistical analysis confirmed that TCA inhibition reversed the polarity of surface charge responses under both positive and negative voltage stimulation (Fig. 3e). To quantitatively verify the metabolic inhibition upon malonate addition, we performed microbial fuel cell (MFC) chronoamperometry as a readout of electron output. The presence of 10 μM malonate markedly suppressed electrons donation to ~3.4% of the control (Supplementary Fig. 21), consistent with strong inhibition of TCA-driven electron supply.

Fig. 3. Voltage stimulation-activated bacterial metabolism.

The variations of surface charge density of the same three S. oneidensis MR-1 bacteria as time elapsed (a, c) before and (b, d) after TCA inhibitor treatment when continuously stimulated by (a, b) +0.2 V and (c, d) −0.2 V (vs. Ag/AgCl). e–g The variations of surface charge density of various S. oneidensis MR-1 bacteria under different conditions when continuously stimulated by +0.2 V or −0.2 V (vs. Ag/AgCl) for 300 s (e normal/inhibited TCA; f with/without MtrC protein, g with/without CymA protein, n = 10). Data are presented as mean ± s.d. Variations of distribution of bacterial surface charges upon different external redox molecules treatments (h potassium ferricyanide; i potassium ferrocyanide n = 50). IM and OM denote inner membrane and outer membrane, representatively. O and R denote oxidized and reduced state of molecules.

We next investigated mutant strains of S. oneidensis MR-1 lacking key electron transfer components: the outer-membrane cytochrome MtrC (ΔMtrC) and the cytoplasmic membrane cytochrome CymA (ΔCymA)³⁶. In both positive and negative voltage stimulation, the ΔMtrC strain exhibited negligible surface charge variation (Fig. 3f and Supplementary Fig. 22), indicating that MtrC plays an essential role in transducing external voltage stimulation into enhanced intracellular metabolic activity, not merely serving as a passive electron conduit. In contrast, the ΔCymA strain displayed no response to positive voltage stimulation but exhibited surface charge accumulation under negative voltage stimulation (Fig. 3g and Supplementary Fig. 23). Under positive voltage, intracellular electrons could not be transferred to the outer membrane through the CymA-mediated electron transport pathway, preventing surface charge changes. Conversely, under negative voltage, electrode-supplied electrons accumulated on the bacterial surface due to the lack of inward metabolic coupling. This asymmetry implies that CymA is essential for coupling electrode-derived electrons into metabolism under negative voltage conditions.

To further confirm that outer-membrane cytochromes (OMCs) serve as the primary sites of surface charge modulation, we replaced external voltage stimulation with exogenous redox reagents. Potassium ferricyanide, an electron acceptor, oxidized OMCs and induced a marked increase in bacterial surface charge (Fig. 3h and Supplementary Fig. 24a), consistent with the response to positive voltage stimulation. In contrast, treatment with potassium ferrocyanide acted as electron donors to reduce OMCs, leading to a pronounced decrease in surface charge (Fig. 3i and Supplementary Fig. 24b). These results demonstrate that direct redox stimulation of OMCs is essential for modulating surface charge states and thereby enhancing bacterial metabolic activity. Collectively, these findings establish that cytochrome-mediated electron transfer pathway couples external electrical or chemical redox inputs to TCA cycle regulation, providing the mechanistic basis for voltage-activated bacterial metabolism.

Logic control of bacterial metabolic levels

These observations prompted us to examine whether bacterial metabolic rates could be regulated by modulating the electron transfer pathway. We first increased the external voltage to facilitate the direct electron transfer (DET) process. As shown in Fig. 4a, applying +0.2 V induced a pronounced increase in surface charge, consistent with previous results. However, further increasing the potential to +0.4 V or +0.6 V produced no additional charge accumulation, indicating that higher voltages rates cannot surpass the intrinsic saturation capacity of bacterial surface charges (see also Supplementary Fig. 25). Statistical analysis across multiple single bacteria (Fig. 4b) confirmed that the final charge levels remained constant at elevated voltages. We attribute this plateau to limitations within the intracellular electron transport chain. Once the surface charge accumulation reaches its upper limit, the capacity for electron flux from the cytoplasmic redox network is saturated, resulting in a steady metabolic rate despite stronger external electron acceptors. Therefore, simply raising the applied voltage cannot accelerate metabolism beyond this threshold, and excessive voltages may even compromise cell viability.

Fig. 4. Logic control of bacterial metabolic levels.

a The variations of surface charge mapping of the same S. oneidensis MR-1 bacterium when continuously stimulated by +0.2 V, +0.4 V, and +0.6 V (vs. Ag/AgCl), respectively. Scale bar: 2 μm. The calibration bar denotes surface charge density (unit of fC/μm²). b Variations of distribution of bacterial surface charge density upon different external voltage stimulation. c The variations of surface charge density of the same S. oneidensis MR-1 bacterium as time elapsed when continuously stimulated by +0.2 V (vs. Ag/AgCl) in the absence and presence of riboflavin (2 μM), respectively. d The variations of surface charge density increase rate of S. oneidensis MR-1 bacteria as time elapsed before/after adding riboflavin. Statistical analysis of surface charge density variations of multiple S. oneidensis MR-1 bacteria when stimulated by +0.2 V (vs. Ag/AgCl) in the (e) absence and (f) presence of riboflavin (2 μM), respectively, and in the presence of 2 μM riboflavin when continuously stimulated by (g) +0.2 V and (h) +0.4 V (vs. Ag/AgCl), respectively (n = 4). Data in (e–h) are presented as box-and-whisker plots overlaid with individual data points. Boxes represent the 25^th–75^th percentiles, central lines represent medians, and whiskers represent the minimum and maximum values. Open squares connected by gray dashed lines denote mean values. i Illustration of strategy for tuning bacterial metabolic rates through combined modulation of EET pathways. RF denotes riboflavin. DET and IET denote direct and indirect electron transfer pathway, respectively.

To overcome this, we introduced an additional electron dissipation route, IET pathway, by adding riboflavin, a redox mediator that shuttles electrons between bacteria and electrodes³⁷. As shown in Fig. 4c, without riboflavin, surface charge gradually increased to saturation under +0.2 V. After addition of 2 μM riboflavin, the surface charge further increased then declined to a plateau near to the initial level, and a higher concentration of riboflavin led to a shorter time for this process (Supplementary Fig. 26). Statistical analysis of the saturated surface charge revealed a pronounced increase upon riboflavin addition (Supplementary Fig. 27). We attributed this elevated saturated level to the formation of riboflavin complexes with outer membrane cytochromes that generates additional electron storage sites³⁸. It is worth noting that the saturated charges induced by riboflavin were similar between different concentrations of riboflavin, suggesting that 2 μM riboflavin already occupies most cytochromes (Supplementary Fig. 27). Notably, we confirmed that riboflavin itself does not measurably perturb the electrochemistry and ECL process of Ru(bpy)₃²⁺-DBAE system (Supplementary Fig. 28), indicating that the signal changes observed upon riboflavin addition in bacterial experiments arise from biological effects rather than a direct chemical enhancement of the ECL system. Based on these results, we conclude that the further metabolic enhancement induced by introducing IET through riboflavin addition is primarily achieved by accelerating the dissipation of surface charge and increasing the surface charge capacity of the bacteria. Further, a step of voltage to +0.4 V led to more surface charges accumulation (Fig. 4d), signifying elevated activity, which is contrast to that without riboflavin (Fig. 4a). These results indicate that DET via cytochromes is essential for initiating active metabolic upregulation, while IET complements by preventing intracellular electron flux saturation. However, IET alone can only induce a modest increase in metabolic activity in a passive manner (Fig. 4d). Statistical analysis of surface charge variations across multiple bacteria under these conditions revealed trends consistent with the aforementioned observations, thereby demonstrating the generality of this DET- and IET-based metabolic regulation (Fig. 4e–h).

Above results summarized a logic regulation circuits for bacterial metabolic rates through both DET and IET pathways. As shown in Fig. 4i, the presence of an external stimulus capable of exchanging electrons with the bacteria is a prerequisite for elevating their metabolic activity. However, when either the DET or IET pathway operates alone, the extent of metabolic enhancement remains limited, with IET alone leading to a much weaker increase than DET alone. Only when both pathways coexist can the bacteria achieve the highest metabolic level. This tripartite paradigm offers a powerful approach to rationally modulate bacterial metabolism in real time. Moreover, the findings provide direct single-cell evidence that flavin-mediated IET contributes to electron transfer even in electrode-contacted bacteria. This observation is consistent with previous reports proposing that flavin-cytochrome complexes can accelerate interfacial electron transfer, while further demonstrating that such enhancement is accompanied by sustained metabolic activation³⁹.

High-efficient screening of bacterial subpopulations

Having established that voltage-driven different EET pathways cooperatively modulate bacterial metabolism, we next asked whether these charge dynamics could serve as phenotypic fingerprints to identify high-performance subpopulations. Given that stimulation via DET followed by IET led to the maximal metabolic activation in bacteria, we utilized the surface charge accumulation rate during the sequential stimulation process to probe bacterial-limited metabolic capacities. Our earlier observations showed that upon IET introduction, the bacterial surface charge continued to rise until reaching saturation, after which it declined (Fig. 4c). This decline reflects a saturated level of intracellular metabolism and enhanced electrode-bacteria electron exchange with the help of IET. Therefore, we defined another parameter: the charge dissipation rate post-saturation during IET predominant process, indicative of the interfacial electron transfer efficiency. By measuring the both parameters across individual bacteria, as shown in Fig. 5a, we observed no correlation between charge accumulation and depletion rates. This result suggests that metabolic activity and interfacial electron transfer capacity of bacteria are independent. In our study, we distinguish between two sub-processes contributing to overall EET efficiency: (i) the intracellular generation of electrons from metabolism (reflected in surface charge accumulation), and (ii) the interfacial transfer of electrons to an external electrode (reflected in surface charge dissipation). The above result suggests that a bacterium can be metabolically active but inefficient in electron export, or vice versa. This independence does not imply that surface charge is unrelated to EET. Rather, surface charge accumulation is a direct readout of intracellular electron generation, while EET flux depends on both the supply and export of electrons.

Fig. 5. High-efficient screening of bacterial subpopulations.

a Distribution of charge accumulation and depletion rate of S. oneidensis MR-1 bacteria (n = 40) when stimulation with voltage of +0.2 V (vs. Ag/AgCl) and riboflavin after metabolic activation. The solid lines represent the medians. b Spatial distributions of S. oneidensis MR-1 bacteria with dual-high or dual-low activity of metabolism and interfacial electrons transfer. Scale bar: 10 μm. c Chronoamperometry of two batches of S. oneidensis MR-1 bacteria at +0.2 V (vs. Ag/AgCl) (batch 1 is freshly cultured cells; batch 2 is aged cells). d Distribution of charge accumulation and depletion rate of corresponding two batches of S. oneidensis MR-1 bacteria (n = 40) when stimulation with voltage of +0.2 V (vs. Ag/AgCl) and riboflavin after metabolic activation. The ellipse centers represent mean values and the peripheries represent the standard deviations.

A high-performing electroactive bacterium must exhibit both high metabolic capacity (high accumulation rate) and high interfacial electron transfer efficiency (high dissipation rate), in such a context, screening electroactive bacterial subpopulations that exhibit superior performance in both individual steps is of critical importance. It not only enables the construction of highly efficient MFCs but also provides fundamental insights into the regulatory mechanisms of EET from a biological perspective. Building on our synergistic activation strategy that can achieve the highest microbial metabolic activity and interfacial electron exchange efficiency, and the high-throughput as well as visualization capabilities of our imaging platform. From Fig. 5a, cells in the upper-right quadrant exhibit both high metabolic flux and efficient EET, representing elite subpopulations for bio-electrochemical performance. Then, as shown in the representative imaging results (Fig. 5b), the dual-high cells (those exhibiting both high surface charge accumulation and high charge dissipation rates) can be spatially identified within the imaging field. By mapping the population of bacteria according to both intracellular metabolic capacity and interfacial electron transfer efficiency, we can screen cells with superior EET capability. This dual-parameter approach allows us to identify a subpopulation of bacteria that excel in both electron generation and transfer, providing a more effective method for screening high-performing electroactive bacteria and a foundation for engineering microbial strains with optimized bio-electrochemical performance.

To validate the discriminatory power of our dual-parameter screening platform, we compared two MR-1 populations with different EET activities: one is cells with better EET activity (batch 1) and the other one is cells with lower EET activity (batch 2). Conventional MFC chronoamperometry confirmed that the cells from batch 2 exhibited delayed current onset and markedly reduced current output (Fig. 5c). When analyzed by our single-cell assay, the cells from batch 2 showed a systematic shift toward lower accumulation and dissipation rates, whereas the cells from batch 1 retained a pronounced fraction of “dual-high” cells (Fig. 5d). These results provide a population-level validation that the dual-parameter approach faithfully report EET-relevant physiological states and can distinguish high-activity versus low-activity subpopulations.

Discussion

In conclusion, we developed a non-invasive and ultrasensitive ECL imaging platform for real-time quantification of single-bacterium surface charge, introducing quantitative surface-charge electrophysiology as a general strategy to visualize and regulate microbial metabolism through electrical stimuli. The technique exploits electrostatic enrichment of cationic luminophores and the quantitative correlation between the optical signals and surface charge, realizing accurate, high-throughput, and dynamic measurement of femtocoulomb-scale bacterial surface charge. Using this approach, we discovered that external voltage stimulation induces metabolism-dependent surface charge modulation in S. oneidensis MR-1. By analyzing the response of bacterial surface charge dynamics to various conditions, we identified a voltage-activated metabolic enhancement mechanism, wherein external voltage stimulation activate outer-membrane cytochrome proteins, triggering intracellular electron flux and accelerating metabolism until a saturated level is reached. Based on the above discoveries, we established a strategy for tuning bacterial metabolic rates through combined modulation of DET and IET pathways. Furthermore, we also confirmed riboflavin-mediated IET can prevent charge saturation and sustain higher metabolic activity by facilitating electron drainage and expanding surface electron storage capacity, which is consistent with earlier findings proposing flavin–cytochrome complex-mediated electron transfer pathways³⁹. Importantly, our single-cell surface-charge imaging platform enables direct real-time visualization of this process in real time and further reveals its coupling to intracellular metabolic activation. These findings provide mechanistic and spatiotemporal evidence supporting the role of flavin-mediated pathways in enhancing both interfacial electron transfer and cellular metabolism. Finally, we applied a dual-parameter optical screening paradigm to realize in-situ distinguishing high-performance bacterial subpopulations by simultaneously assessing maximum metabolic activity and interfacial electron exchange efficiency. In addition, the successful imaging of a Gram-positive strain (B. subtilis) demonstrates that the platform is not restricted to S. oneidensis MR-1, but is applicable to bacteria with distinct cell-envelope architectures. This work introduces surface-charge electrophysiology as an integrative readout linking intracellular redox flux with EET, offering a unified framework for elucidating EET mechanisms. Coupling this platform with single-cell omics in the future may reveal biological determinants of EET efficiency and guide rational engineering of electroactive bacteria for enhanced bio-electrochemical and synthetic-biology applications.

Methods

Cell culture

Shewanella oneidensis MR-1 was cultured in 15 mL LB broth medium at 37 °C for 24 h on a shaker. The bacterial cells were then centrifuged (2650 × g, 4 min) and dispersed in sterile water for further use. The mutant bacteria were cultured with the same method.

ECL imaging of single bacteria

The ITO electrodes were initially sterilized by ultraviolet lamp irradiation and then the bacteria suspension was added onto the ITO surface for deposition by natural sedimentation, without chemical fixation or adhesive coating. After several minutes, the ITO with bacteria was washed with sterile water for further measurement. Next, electrolyte containing 2 mM PBS buffer (pH = 7.0), 0.1 mM Ru(bpy)₃²⁺ and 1 mM DBAE were injected into the electrochemical cell. Finally, sCMOS was triggered and synchronized by the potentiostat when applying a specific voltage on the electrode, then collected ECL images of bacteria. All data analyzing was performed with Origin, Matlab, and ImageJ.

Test of cell viability

After external voltage stimuli and immersion in ECL reagents for a period of time, the bacterial cells were incubated with fluorescent dyes SYTO 9 (for live cells) and propidium iodide (for dead cells) in the dark for 20 min. Then, the cells were imaged with suitable fluorescent filter cubes using Nikon Ti-E microscope.

Simultaneous voltage modulation and bacterial signals collection

Double potential step chronoamperometry was adopted to stimulate bacteria and generate ECL, respectively. In one sweeping cycle, the duration of modulation potential is 8 s and that of 1.1 V is 2 s. To accept electrons from bacteria, the stimulation voltage was set at 0.2 V, 0.4 V, or 0.6 V. Similarly, to donate electrons to bacteria, the stimulation voltage was set at −0.2 V, −0.4 V, or −0.6 V. For better ECL signal of bacteria, the ECL triggering voltage was set at 1.1 V.

Preparation of mesoporous silica NPs (MSNs)

The synthesis of MSNs was according to previously reported protocols⁴⁰. To prepare MSNs, 0.96 g CTAT and 0.17 g TEA were firstly dissolved in 50 mL water and then stirred for 1 h at 80 °C. After that, 2 mL TEOS was added and the solution was stirred for 2 h at the speed of 1000 rpm. Finally, MSNs were centrifugated and washed with ethanol for three times, followed by dispersing in 5 mL ethanol and drying in the oven at 60 °C overnight.

Malonate treatment for TCA-cycle inhibition

Malonate was introduced by directly spiking it into the Ru(bpy)₃²⁺/DBAE measurement electrolyte to a final concentration of 10 μM⁴¹. Prior to ECL imaging, bacteria were pre-incubated in the malonate-containing electrolyte for 10 min to allow sufficient uptake and competitive inhibition of succinate dehydrogenase. During ECL imaging and voltage stimulation, the electrolyte was kept unchanged and continuously contained 10 μM malonate.

Microbial fuel cell (MFC) chronoamperometry current measurement

First, the MFC setup is constructed with a carbon cloth as working electrode, a platinum wire as counter electrode, and an Ag/AgCl as reference electrode. The electrolyte is M9 minimal medium supplemented with 20 mM sodium lactate as carbon source, with or without 10 μM malonate. Then, the entire electrochemical cell was purged with nitrogen gas (N₂) for 20 min prior to measurement to remove dissolved oxygen. Finally, chronoamperometry at a constant potential of +0.2 V (vs. Ag/AgCl) was performed for 10,000 s, and the real-time current was recorded⁴².

Stimulating bacteria using external redox molecules

After bacteria fall on the electrode surface, for comparation, electrolyte containing 2 mM PBS buffer (pH = 7.0), 0.1 mM Ru(bpy)₃²⁺ and 1 mM DBAE were injected into the electrochemical cell. sCMOS was triggered and synchronized by the potentiostat when applying a pulse voltage (1.1 V for 2 s) on the electrode, then collected ECL images (exposure time is 1 s). Next, these bacteria were incubated in solutions containing specific redox molecules (10 μM disodium fumarate/sodium lactate/potassium ferricyanide/potassium ferrocyanide) for 5 min, then the solutions were replaced by the aforementioned ECL reagents. sCMOS recorded ECL signal to get the images of bacteria after external redox molecules stimulation with the same conditions and parameters.

Dual-parameter screening strategy

Using diluted bacterial samples, we employed the real-time surface charge imaging technique to assess EET activity at the single-bacterium level. After bacteria deposited onto the electrode surface, a potential of +0.2 V was applied for 5 min to activate bacterial enhanced metabolism. Subsequently, 2 μM riboflavin was introduced into the solution and double potential step chronoamperometry (0.2 V for 8 s, followed by 1.1 V for 2 s per cycle) was applied to monitor the real-time variations in surface charge of individual bacteria.

Firstly, the surface charge accumulation rate of individual bacterium during the initial 10 s following the onset of the measurement was recorded, which indicates the maximum metabolic activity. After reaching the saturation state, the surface charge depletion rate of individual bacterium during the initial 10 s following the onset of the surface charge decreasing process was then recorded, which reflects the interfacial electron transfer capability between bacterium and electrode. Finally, the charge accumulation and depletion rates for each individual bacterium were statistically analyzed. To identify bacteria with high EET performance, we first plotted the surface charge accumulation rate against the charge dissipation rate. To classify bacterial subpopulations, we use the median values of each parameter as natural dividers. “Dual-high subpopulation” represents bacteria with both accumulation rate above the median and dissipation rate above the median (upper-right quadrant). These cells exhibit high metabolic electron supply and high interfacial electron transfer efficiency. “Dual-low subpopulation” represents bacteria with both parameters below their respective medians (lower-left quadrant). These cells have lower EET capability. In such a context, we then selected the bacteria located in the upper-right quadrant, representing the “dual-high” population, facilitating the screening of bacterial subpopulations possessing concurrently enhanced metabolic capacity and interfacial electron transfer efficiency.

Statistical analysis

All the data were denoted as mean ± s.d. The difference between two groups was analyzed with Student’s t-test and the threshold for difference significance was P < 0.1. All the analyses were performed with Origin 2018 and PRISM 8.0 software. To ensure the reliability and reproducibility of our single-cell measurements, all statistical analyses were performed with the following replication strategy: each experimental group was independently repeated on three different ITO electrodes. For each group, we randomly selected no fewer than 10 single bacteria from the entire population imaged on the three electrodes, and the data from all selected cells were pooled for statistical analysis.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Z.X. designed and analyzed data. Z.X., Z.Z., R.L. and J.Z. performed experiments. Z.X., X.G. and C.M. designed and established the imaging setup. J.Z. and Z.C. directed the research. Z.X., L.J., C.M., J.Z. and Z.C. wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks Zhiwen Wang, Matthew Yates, and the other anonymous reviewer for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the National Natural Science Foundation of China (Grant numbers 22574074, 22374072, 22374125 and 223B2406), Key Research and Development Projects of Ningxia Province of China (2024FRD05098), the Jiangsu Young Talent Lifting Project, the Yangzhou University Interdisciplinary Research Foundation for Chemistry Discipline of Targeted Support (yzuxk202009), the Lvyangjinfeng Talent Program of Yangzhou, and Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX25_0184).

Data availability

All data generated and/or analyzed in this study are included in the manuscript. Source Data is available as a Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-75226-y.