Metal-reducing bacteria gain energy by extracellular electron transfer to external solids, such as naturally abundant minerals, which substitute for oxygen or the other common soluble electron acceptors of respiration. This process is one of the earliest forms of respiration on earth and has significant environmental and technological implications. By performing electron transfer to electrodes instead of minerals, these microbes can be used as biocatalysts for conversion of diverse chemical fuels to electricity. Understanding such a complex biotic-abiotic interaction necessitates the development of tools capable of probing extracellular electron transfer down to the level of single cells. Here, we describe an experimental platform for single cell respiration measurements. The design integrates an infrared optical trap, perfusion chamber, and lithographically fabricated electrochemical chips containing potentiostatically controlled transparent indium tin oxide microelectrodes. Individual bacteria are manipulated using the optical trap and placed on the microelectrodes, which are biased at a suitable oxidizing potential in the absence of any chemical electron acceptor. The potentiostat is used to detect the respiration current correlated with cell-electrode contact. We demonstrate the system with single cell measurements of the dissimilatory-metal reducing bacterium Shewanella oneidensis MR-1, which resulted in respiration currents ranging from 15 fA to 100 fA per cell under our measurement conditions. Mutants lacking the outer-membrane cytochromes necessary for extracellular respiration did not result in any measurable current output upon contact. In addition to the application for extracellular electron transfer studies, the ability to electronically measure cell-specific respiration rates may provide answers for a variety of fundamental microbial physiology questions.

Electron transfer is a fundamental process that governs the reduction-oxidation (redox) reactions critical for all cellular energy conversion pathways, including photosynthesis and respiration.1,2 Respiratory organisms extract free energy from their environment by coupling the oxidation of electron donors (food) to the reduction of electron acceptors (oxidants). While oxygen is the familiar electron acceptor of aerobic respiration, many anaerobic microbes are capable of performing electron transfer to alternative soluble acceptors, such as nitrates and sulfates, which can also diffuse to the electron transport chain inside living cells. In contrast to these intracellular electron transfer processes, we now know that a class of anaerobic microbes called dissimilatory metal-reducing bacteria (DMRB), including Shewanella, are capable of extracellular electron transfer (EET) to external surfaces.3–5 This respiratory strategy is attracting considerable fundamental and applied interest, considering that the external electron acceptors can range from natural oxidized metals in the environment to anodes.5 

From an environmental perspective, microbes performing EET are major players in elemental cycles, including the carbon cycle, occurring at a global earth scale.4,5 From a technological perspective, microbial EET is heavily pursued for interfacing redox reactions to electrodes in multiple renewable energy technologies. These technologies include microbial fuel cells (MFCs),6 where microbial biofilms oxidize diverse fuels and route the resulting electrons to energy-harvesting anodes, and the reverse process of microbial electrosynthesis,7 where renewable electricity drives reductive microbial metabolism for synthesis of high value fuels. The success of these technologies hinges on efficient electron exchange mechanisms between microbes and electrode surfaces. Two decades of research on the most studied DMRB model organism, Shewanella oneidensis MR-1, have revealed multiple mechanisms that can be categorized as either direct or indirect.8 Indirect mechanisms rely on biogenic or naturally occurring molecules, including flavins,9,10 that diffusively shuttle electrons from cells to electrodes. Direct mechanisms route electrons through multiheme cytochromes that are either located on the cell surface11–13 or along micrometer-long membrane extensions known as bacterial nanowires.14–18 

Optimization of microbial fuel cells requires detailed knowledge of the contribution from each EET mechanism within biofilms.19 This includes an understanding of the biological, mass transport, and electrochemical factors that can limit performance,20 down to the fundamental respiration rate by individual cells. Much of our mechanistic knowledge of microbe-to-electrode EET is derived from a combination of electrochemical techniques and genetic approaches. Microbial fuel cells and anodic half cells have been used to quantify and compare the electricity production from wild-type S. oneidensis MR-1 and mutant strains lacking specific proteins that are hypothesized to affect EET.21 The impact of specific mutations on EET, especially when combined with additional biochemical, structural, and spectrometric techniques, can be used to reconstruct the cellular EET pathways.22 One difficulty, however, stems from the bulk nature of most electrochemical techniques, which measure the total current from a large microbial population (e.g., ≫107 cells). While it is tempting to think of an average cell-specific EET rate, obtained by normalizing with (often qualitative) cell density measurements performed at the end of an experiment, such an approach ignores the microbial heterogeneity present in any population.23 For instance, the same total anodic current can result from all cells respiring uniformly at a specific rate, or from only 10% of cells respiring at ten times that specific rate; these two scenarios clearly require completely different strategies for optimization.

The effect of heterogeneity is particularly important in microbial fuel cells that contain a mixture of biofilm and planktonic cells, keeping in mind that a wide statistical distribution of cellular respiration current is possible even within each of these two sub-groups. The causes of microbial heterogeneity include basic genotypic variability resulting from mutations, as well as phenotypic variability resulting from progression through the cell cycle or as a physiological response modulated by the local environment and its history.24 As we seek a more complete understanding of the maximum power densities possible from microbial technologies such as microbial fuel and electrosynthesis cells, there is a clear need for techniques that quantify both the expected per-cell and statistical distribution of EET in microbial cultures.

Single cell techniques will also allow more direct comparisons between experiments utilizing different microbes, which may result in different biofilms properties or cell densities, even under similar growth conditions. Perhaps more importantly, given the wealth of knowledge that can be gained from genetic studies, single cell techniques are also critical for studies that compare wild-type to mutant strains lacking putative EET proteins. Gene products hypothesized to perform EET (e.g., surface cytochromes or type IV pili) may also impact the cellular surface charge, attachment ability, and biofilm forming properties. Bulk techniques therefore cannot distinguish whether a decrease in anodic current is due to diminished EET directly through a specific protein, or indirectly by disrupting cellular attachment to electrodes and the subsequent biofilm development.

Only a handful of recent studies address these issues surrounding bulk electrochemical techniques. McLean et al. quantified the average per-cell EET to graphite electrodes by combining live noninvasive imaging with an optically accessible MFC, allowing an accurate cellular count as S. oneidensis MR-1 population of the anode developed from separate cells to mature biofilms.25 This approach revealed a current per cell ranges from 50 to 200 fA, depending on the MFC resistance and growth phase. Another study obtained single cell measurements of EET from a different DMRB, Geobacter sulfurreducens DL-1, using microscale Ti/Au electrode arrays.26 When active cells transiently came into contact with these electrodes, a short circuit current of 92(±33) fA per cell was observed in about 30% of the contact events. Finally, Liu et al. reported an innovative approach taking advantage of optical tweezers to attach single S. oneidensis MR-1 cells to electrodes, revealing a current output of 200 fA per cell at a working electrode potential of +200 mV vs. Ag/AgCl.27 To date, however, there is no technical description of a complete instrument or standardized measurement procedure for detecting single cell extracellular respiration.

Here, we present the first complete description of a system inspired by, and combining elements from, the above-mentioned studies of specific respiration rates. The integrated system combines live imaging, an infrared optical trap, and a new electrochemical chip concept containing an array of indium tin oxide (ITO) microelectrodes that serve as ‘landing pads’ for individually trapped bacteria. The microelectrode potentials are controlled by a potentiostat, allowing the detection of the respiration current concomitant with landing a cell on a specific microelectrode. The entire system can be built from standard commercially available optical and electrochemical components, in combination with standard lithographic microfabrication techniques. The microfabrication procedure was optimized to address several challenges stemming from the biocompatibility requirement and high current sensitivity needed to detect single cells. Furthermore, the biological growth conditions and measurement protocol are described in detail. In addition to allowing single cell measurements of EET from a variety of organisms and mutants, the reported system may find wider applicability for general microbiology studies, including the analysis of how respiration rates in heterogeneous microbial cultures are impacted by specific environmental factors.

The single cell extracellular respiration platform integrates three main components: (1) an infrared optical tweezers system for positioning individual cells, (2) a custom transparent electrochemical chip containing an array of ITO microelectrodes on standard coverslips, and (3) a vacuum-sealed perfusion chamber for loading and handling of bacterial cultures in physiological media. In what follows, we detail the design criteria and construction of each of these components.

Optical trapping stemmed from a series of pioneering experiments by Ashkin,28 demonstrating the effect of laser-induced optical forces on controlling micrometer-sized particles in both air and liquid. The technique has been extensively used to characterize biological samples, ranging from DNA and single protein molecules to whole cells.29 

Our optical tweezers system is constructed from commercial components (combined as #OTKB, ThorLabs) and is based on previously published designs.30 The system (Figure 1) is assembled in an inverted configuration, ideal for in vivo biological work, and equipped with beam directing/steering components, a manual translation stage in combination with three piezo-actuators (20 μm of travel at 20 nm resolution driven by three #TPZ001 piezo drivers, ThorLabs), a 100X Nikon oil immersion objective (1.25 numerical aperture, 0.23 mm working distance), and a 1280 × 1024 pixel CCD camera interfaced via universal serial bus (USB) to a computer for video imaging. The sample is illuminated for imaging by a single emitter white light-emitting diode (LED) through a 10X air condenser lens (0.25 numerical aperture, 7 mm working distance). The trapping laser source is a 980 nm single mode fiber-pigtailed laser diode (1.1 μm spot size, 330 mW maximum power, #PL980P330J, ThorLabs) mounted in a laser diode mount (#LM14S2, ThorLabs) and controlled using a laser diode current driver and temperature controller (#LDC210C and #TED200C, ThorLabs). The laser output is collimated using a FiberPort (#PAF-X-7-B, ThorLabs). Dichroic mirrors are used to transmit visible light (e.g., to the camera for imaging), while reflecting the infrared laser towards the sample for trapping.

FIG. 1.

Optical layout of the trapping setup for single cell extracellular respiration measurements. The infrared laser (980 nm) traps individual bacteria within a perfusion chamber(red) sealed against an electrochemical chip (cyan) containing indium tin oxide (ITO) microscale working electrodes. The microelectrodes are biased at a suitable electron-accepting potential, relative to an integrated Ag/AgCl reference electrode, using a potentiostat. The chip also contains an ITO counter electrode to complete the standard 3-electrode electrochemical cell. The trap is used to land individual bacteria on the microelectrodes, and the single cell respiration current is measured using the potentiostat. For detailed description of the microfabrication and color coding, see Fig. 3.

FIG. 1.

Optical layout of the trapping setup for single cell extracellular respiration measurements. The infrared laser (980 nm) traps individual bacteria within a perfusion chamber(red) sealed against an electrochemical chip (cyan) containing indium tin oxide (ITO) microscale working electrodes. The microelectrodes are biased at a suitable electron-accepting potential, relative to an integrated Ag/AgCl reference electrode, using a potentiostat. The chip also contains an ITO counter electrode to complete the standard 3-electrode electrochemical cell. The trap is used to land individual bacteria on the microelectrodes, and the single cell respiration current is measured using the potentiostat. For detailed description of the microfabrication and color coding, see Fig. 3.

Close modal

The laser wavelength (980 nm) was chosen to minimize damage to the cells during manipulation before placement on the microelectrodes. Previous studies have characterized possible photodamage to the trapped cells.31,32 The damage can be largely mitigated by using the near infrared region (790-1064 nm), with a photodamage minimum at 970 nm. In addition, since oxygen is implicated in the photodamage pathway, anaerobic conditions reduce cell damage down to background levels.31 This is ideal for our purposes, since our goal is to perform EET measurements under anaerobic conditions to exclude oxygen as an alternate electron acceptor. By taking appropriate precautions including a near-IR laser, anaerobic conditions, and <100 mW of trapping power at the sample plane, previous studies have manipulated individual E. coli cells for durations up to an hour32 (in excess of the time needed to perform EET measurements) and even confirmed the long-term viability by observing the division of trapped cells.33 These conditions allowed us to trap individual S. oneidensis MR-1 cells, which remained viable judging by their continued motility after turning off the trap.

Our chip concept facilitates respiration measurements from individual bacteria in a standard three-electrode electrochemical configuration. An array of nine ITO working electrodes (WEs) are lithographically patterned on standard glass coverslips, in addition to a central counter electrode (CE) (Figure 2). An Ag/AgCl reference electrode with porus Teflon tip (#CHI111P, CH Instruments) is connected just upstream of the chamber.

FIG. 2.

(a) The microfabricated electrochemical chip containing an array of ITO electrodes. Inset shows a tapping mode atomic force microscopy image of one of the windows fabricated at the tips of the 9 working electrodes (WEs). These windows provide microscale openings to the ITO underlying an SU-8 passivation layer. (b) Schematic of the device, showing the 9 WEs (WE 1–WE 9) in yellow, the central counter electrode (CE) in light purple, as well as the overlying passivation layer in dark purple. Inset shows a larger view of one of the windows, which allows the attachment of optically trapped bacteria on ITO.

FIG. 2.

(a) The microfabricated electrochemical chip containing an array of ITO electrodes. Inset shows a tapping mode atomic force microscopy image of one of the windows fabricated at the tips of the 9 working electrodes (WEs). These windows provide microscale openings to the ITO underlying an SU-8 passivation layer. (b) Schematic of the device, showing the 9 WEs (WE 1–WE 9) in yellow, the central counter electrode (CE) in light purple, as well as the overlying passivation layer in dark purple. Inset shows a larger view of one of the windows, which allows the attachment of optically trapped bacteria on ITO.

Close modal

Since our goal is to position individual bacteria on defined electrodes, the majority of the chip’s surface is passivated with a layer of SU-8 epoxy based photoresist (#2000.5, MicroChem), except for exposed ITO regions that allow contact to the biological growth medium by the central 0.5 mm radius CE and defined 15 × 15 μm windows at the WE tips. In addition to discouraging accidental contact with more than one (or few) individual bacteria, the small exposed WE window translates to a small background electrochemical current before cellular contact. The small baseline current is critical, allowing us to select a small total current measurement range (e.g., <60 pA) by the potentiostat (Reference 600, Gamry), which in turn facilitates detection of the sub-pA currents expected from individual bacteria (e.g., ∼10 fA steps can be detected in the 60 pA total range using the potentiostat’s 16-bit analog/digital converter). Figure 2(a) is a picture of the fabricated chip, with an inset image (atomic force microscopy (AFM), see below) showing the ITO window at a WE’s tip, which serves as the cellular landing pad. Figure 2(b) is a schematic of the same chip. Note the centralized design, allowing minimal translational motion while switching between WEs on the optical stage, while keeping all the WEs equidistant from the central CE. Next, we detail the microfabrication workflow leading to the finished chip.

The microfabrication process is divided into three stages, each requiring a lithography mask, as shown in Figure 3. The first stage (steps 1-5 in Figure 3) results in the desired ITO pattern on the coverslip. The second stage (steps 6 and 7) produces a preliminary SU-8 passivation layer, which keeps only desired ITO features exposed. Finally, the third stage (steps 8-9) produces the second SU-8 passivation layer, which define the small 15 × 15 μm WE windows. The chip is then sealed against the perfusion chamber (step 10), ahead of optical trapping and electrochemical measurements (step 11).

FIG. 3.

Microfabrication workflow for the electrochemical chip. Steps 1-5 show ITO deposition and patterning. Steps 6-9 detail the passivation of the surface with two SU-8 layers, while leaving microscale windows to the ITO for cellular attachment. Steps 10 and 11 show perfusion chamber assembly and trapping.

FIG. 3.

Microfabrication workflow for the electrochemical chip. Steps 1-5 show ITO deposition and patterning. Steps 6-9 detail the passivation of the surface with two SU-8 layers, while leaving microscale windows to the ITO for cellular attachment. Steps 10 and 11 show perfusion chamber assembly and trapping.

Close modal

Before microfabrication, the glass coverslips are cleaned with a series of sonication steps in acetone, isopropyl alcohol, and distilled de-ionized (DDI) water, for 5 min each, and then finally rinsed in DDI water and dried using N2 gas. Any residual water is removed by heating the chips at 150 °C for 10 min on a hot plate. To remove all organic matter, the coverslips are then exposed to an O2 plasma (Tegal Plasma Asher 421) for 2 min at 200 W with a reflective power of ≤5 W. The coverslips are then treated with hexamethyldisilazane (HMDS) vapor in a HMDS vapor deposition tank to promote the adhesion of photoresist in the subsequent steps.

To begin the first stage (steps 1-5 in Figure 3), positive photoresist (AZ 5214 IR, Clariant) is spin coated on the coverslips for 5 s at 500 rpm, using a ramp of 100 rpm/s, to remove the bulk of the photoresist, and then at 3000 rpm for 30 s. The resulting photoresist layer is 1.6 μm thick. The chips are then baked for 2 min at 100 °C in order to remove the solvent and set the resist. The first photomask, which defines the ITO WEs and CE, is aligned with each chip using a Karl Suss Aligner MA6 set for 15 s exposure time at 8 W lamp power, with an alignment gap of 30 μm. The exposed chips are then submerged for 1 min in a solution of 5 parts DDI water to 1 part positive photoresist developer (AZ 400K, Clariant). The chips are rinsed in DDI water and plasma cleaned again as described above. The result is a coverslip with photoresist present everywhere except where conductive electrode contacts are desired. A 300 nm thick ITO film is then deposited (2400 s at 180 W in Denton Discovery Sputterer 550) on the entire surface. The thickness was chosen based on prior work showing favorable electrical properties at this thickness while still maintaining high optical transmittance for both visible and infrared light.34 After ITO deposition, the chip is submerged in acetone and sonicated for 5 min to lift off the resist. It is then cleaned again using isopropyl alcohol and DDI water. At this point, the chip consists of the desired ITO array pattern on the coverslip. The chips are then baked at 400 °C in N2. Annealing helps create oxygen vacancies, which donate two free electrons, as well as increases the density of the crystal structure.

In the second stage of microfabrication (steps 6 and 7 in Figure 3), SU-8 2000.5 negative photoresist (MicroChem) is spin coated on the chips for 5 s at 500 rpm, using a 100 rpm/s ramp, and then for 30 s at 1800 rpm with 1000 rpm/s ramp. The chips are then baked at 90 °C for 1 min in order to remove the solvent. The second photomask, which defines the openings for ITO features, is aligned with each chip using the Karl Suss Aligner set for 23 s exposure with a 23 μm gap on soft contact. The chips are post baked for 10 min at 90 °C in order to cross link the SU-8 polymer and harden the exposed resist. After post baking, the chips are submerged in SU-8 developer for 1 min, rinsed with isopropyl alcohol for 10 s, rinsed with DDI water, and developed again for 10 s in fresh developer. The second development step ensures the removal of any traces which might have reattached from the solution during the first development step.

Large area development of SU-8 can often introduce microscale flaws (e.g., cracks or holes) that may leave ITO exposed in undesirable areas. The third stage of microfabrication (steps 6 and 7 in Figure 3) addresses this issue by applying a second SU-8 passivation layer on top of the first. Using a different photomask, the second layer features smaller 15 × 15 μm windows designed to fit within the larger windows of the first layer. The coating and exposure details are similar to the second stage, described above, except that the exposure time is reduced to 21 s from 23 s. The reduced exposure time ensures the unexposed SU-8 (now relatively small 15 × 15 μm) features are developed away more easily, avoiding any surface residue that may passivate the desired window. Furthermore, the chip is now sonicated during development for 10 s, unlike the first layer of SU-8. After development, the chips are treated with O2 plasma in the Tegal Plasma Asher 421 once more, but at a lower power (100 W forward power with reflective power ≤5W for 1 min) so as not to damage the thin SU-8 layer. The result is a ready-to-use mostly passivated surface with small exposed windows at the tips of the ITO WEs and a large exposed central ITO CE.

A perfusion chamber, for loading and handling of growth medium and bacterial cultures, was designed and constructed to interface to the electrochemical chip. The key design criterion is a good seal, in order to minimize oxygen permeability into the system. While S. oneidensis MR-1 is a facultative anaerobe, capable of growing with both oxygen and other soluble or insoluble electron acceptors, the availability of any alternate oxidant in the system will minimize the EET current to the microelectrodes. The round polycarbonate chamber body (Figure 4(a)) seals against the bottom chip and a top coverslip by pulling a vacuum between two concentric o-rings at the bottom and top interfaces. Figure 4(b) shows the chip (with attached 32 gauge stranded wires that connect to the potentiostat) and chamber (with inlet, outlet, reference electrode and vacuum line) assembly mounted on top of the optical trap’s microscope stage. Finally, Figure 4(c) shows the whole integrated system, which is enclosed within a home-made metallic enclosure serving as a Faraday cage to minimize any external interference detrimental for sub-pA measurements.

FIG. 4.

(a) Schematic of the perfusion chamber sealed against the bottom electrochemical chip and a coverslip allowing illumination from the top. A seal is achieved against both surfaces by pulling a vacuum between dual o-rings. (b) The entire chamber-chip assembly mounted on a microscope stage. The reference electrode/inlet, vacuum line, and stage are labeled as i, ii, and iii, respectively. (c) The integrated platform combining the perfusion chamber and electrochemical chip, labeled i, with the optical trapping system. The translation stage and infrared laser fiber are indicated by ii and iii.

FIG. 4.

(a) Schematic of the perfusion chamber sealed against the bottom electrochemical chip and a coverslip allowing illumination from the top. A seal is achieved against both surfaces by pulling a vacuum between dual o-rings. (b) The entire chamber-chip assembly mounted on a microscope stage. The reference electrode/inlet, vacuum line, and stage are labeled as i, ii, and iii, respectively. (c) The integrated platform combining the perfusion chamber and electrochemical chip, labeled i, with the optical trapping system. The translation stage and infrared laser fiber are indicated by ii and iii.

Close modal

The microelectrodes’ fabrication and electrochemical activity were characterized in advance of the biological measurements. Tapping mode-AFM (TM-AFM) was used to confirm that the lithographically fabricated WE windows developed properly, using a Bruker Innova AFM and Olympus silicon probes (AC240TS) with 2 N/m spring constant and 9 ± 2 nm tip radius of curvature. Figure 2(a) contains the AFM topography profile of a typical window, showing a central smooth (1 nm RMS roughness) ITO bottom surface with a sharp drop off at the edge of the window and the relatively rough (5 nm RMS roughness) surface of the SU-8 passivation layer. The AFM images also revealed that the square design developed with rounded edges, rather than sharp corners, making the window geometry closer to a circle with 8.4 μm diameter.

After ensuring the successful microfabrication of the ITO WEs, we turned our attention to confirming the electrochemical activity of these microelectrodes using the standard ferri/ferro cyanide redox couple (Fe3CN63− + e↔Fe3CN64−) in our system, rather than the bacteria. Figure 5 shows the cylic voltammetry (CV) from two separate WE windows exposed to a solution of 10 mM K3Fe(CN)6 in 1M KNO3. The CVs exhibited the classic expected ultramicroelectrode35 (UME, i.e., with a critical dimension smaller than the length scale of the surrounding depletion later) behavior with no peaks but a sigmoidal quasi-reversible transition between the oxidized and reduced states. Under these conditions, starting in the fully oxidized state and applying a reductive potential results in a maximum current described by i = n A F m O C O , where n = 1 for this one-electron process, F is Faraday’s constant, A is our nominal electrode area, and C O is the bulk concentration of ferricyanide in our CV measurement. mO is a geometric mass-transfer coefficient given by 4 D O π r O for a UME disk shape with radius rO, and DO is the ferricyanide diffusion coefficient, which is on the order of 10−9 m2/s from prior studies.35–37 This calculation, which is intended for heuristic comparison since the disk shape only approximates our rounded-corner WEs and because of the inherent window-to-window size variability (up to a few micrometers) in our SU-8 photolithography process, reveals an expected current on the order of tens of nA. This expected range is in good agreement with the results in Figure 5 for two separate WEs, confirming that most, if not all, surface area of the ITO windows is electrochemically active. Shorting the two WEs gave the expected additive current result, with no extraneous contributions. With the electrochemical validation in hand, we now describe the cellular EET measurements.

FIG. 5.

Cyclic voltammetry demonstrating the electrochemical activity of the lithographically patterned ITO working electrodes exposed to 10 mM K3Fe(CN)6 in 1M KNO3 and using the chip’s central ITO counter electrode. Potential is shown against Ag/AgCl (1M KCl). Scan rate 10 mV/s.

FIG. 5.

Cyclic voltammetry demonstrating the electrochemical activity of the lithographically patterned ITO working electrodes exposed to 10 mM K3Fe(CN)6 in 1M KNO3 and using the chip’s central ITO counter electrode. Potential is shown against Ag/AgCl (1M KCl). Scan rate 10 mV/s.

Close modal

Wild-type S. oneidensis MR-1 and its markerless multi-deletion mutant ΔOMC, which lacks the outer membrane cytochromes necessary for EET,38 were first grown aerobically in 50 ml LB medium from a frozen (−80 °C) stock up to an optical density at 600 nm (OD600) of 1.5. These aerobic pre-cultures were inoculated at 1% (v/v) into anaerobic serum bottles containing 100 ml PIPES-buffered (pH 7.0) M1 medium with 18 mM lactate and 30 mM fumarate as the electron donor and acceptor, respectively.21 Prior to inoculation, anaerobic conditions were reached by purging with N2 for 1 h and the serum bottles, sealed with butyl stoppers and aluminum seals, were sterilized by autoclaving at 120 °C for 15 min. Each anaerobic culture was grown to mid-exponential phase (OD600 0.07), after which cells were injected into a defined non-growth trapping medium (at 106 fold dilution for low cell density) containing PIPES buffer, 30 mmol/l; sodium hydroxide, 7.5 mmol/l; ammonium chloride, 28.04 mmol/l; potassium chloride, 1.34 mmol/l; sodium phosphate monobasic, 4.35 mmol/l; sodium chloride, 30 mmol/l; sodium lactate, 18 mmol/l. This composition is identical to the anaerobic M1 media except that it that lacks fumarate as well as the vitamins, minerals, and amino acid trace solutions of M1 medium. Fumarate was removed to ensure that no soluble electron acceptor diminishes EET, while the trace solutions were removed to avoid their possible redox activity at the microelectrode surfaces. The low cell density in the trapping medium reduces the likelihood of cells contacting or accidentally landing on the microelectrodes before and during single cell measurements. Despite the low cell density, it should be noted that we did not encounter difficulty finding cells in the vicinity of WEs, likely due to a previously reported phenomenon where cells congregate near insoluble electron acceptors.39 Finally, this trapping culture is introduced into the chamber-chip assembly that is integrated into the optical trap.

A potential of +400 mV vs. Ag/AgCl (1M KCl) was applied to the chips’ WEs, and the system was left to stabilize for up to 3 h before trapping, in order to achieve the low electrochemical background (typically <60 pA and a current change rate <10 fA s−1) necessary for detecting the expected respiration currents, as discussed above. Such a background allowed a simple detrending procedure to quantify the effect of single cell respiration, where the current before cell attachment is fit with a linear or low-order polynomial and this trend is subtracted from the time series. Prior to the trapping experiments, we confirmed, using standard plating and colony counting methods, that this stabilization time in the acceptor-free medium did not result in any loss of viability. The WE potential was chosen to be compatible with Shewanella’s direct EET pathways, being more positive than the redox potential window of the outer membrane multiheme cytochromes.40,41

In a typical scenario, a single cell is trapped in the vicinity of a WE after the stabilization time and the stage’s piezo-actuators are used to land the cell on the ITO window. The time of contact was determined by using the visible light image to microscopically observe the cell as it lands, which typically caused the cell to fluctuate in the trap when it first encountered the normal force from the surface. Out of 27 wild-type S. oneidensis MR-1 cells landed in this manner, 4 cells resulted in a clear current increase above background, all in the range of 15 fA to 100 fA and sustained over >10 min (Figure 6). Electrochemical data were recorded for up to 15 min or until cells appeared to detach from the surface. An increase in current after attachment was observed in a few additional cases, but with a much slower rise time (>15 min post attachment) than Figure 6, making it difficult to definitively assign these responses to respiration rather than instrument drift, using our simple detrending procedure. Our observation that only a fraction of the cells is actively respiring at electrodes is consistent with a previous study not based on optical trapping.26 It remains unclear whether such a percentage of actively electrode-respiring cells reflects the typical heterogeneity in these cultures, or whether it stems from the particular growth procedure used in this and previous studies; our finding clearly motivates additional single cell studies of the statistical distribution in EET as a function of various growth conditions. In comparison, no EET current was detected from any of 20 ΔOMC cells tested under identical conditions, further validating our measurement system, as the outer membrane multiheme cytochromes are known to be necessary for mediating EET.

FIG. 6.

Current measurements for four distinct Shewanella oneidensis MR-1 cell contacts with microelectrodes biased at +400 mV vs. Ag/AgCl (1M KCl). An arrows marks the time of contact. Inset shows an image of a single S. oneidensis MR-1 cell (indicated with dashed circle) trapped on the microelectrode window. 5 μm scale bar.

FIG. 6.

Current measurements for four distinct Shewanella oneidensis MR-1 cell contacts with microelectrodes biased at +400 mV vs. Ag/AgCl (1M KCl). An arrows marks the time of contact. Inset shows an image of a single S. oneidensis MR-1 cell (indicated with dashed circle) trapped on the microelectrode window. 5 μm scale bar.

Close modal

Finally, we discuss the magnitude of the observed respiration rates (15-100 fA per cell), in light of previous EET studies and more recent measurements of the Shewanella cytochrome complexes. The single cell currents detected here are consistent with previous respiration measurements in whole planktonic and biofilm cultures (normalized by cell number), using O2 and graphite electrodes as electron acceptors, respectively.15,25 Furthermore, the respiration currents detected here translate to an electron transfer rate up to 6.2 × 105 s−1 per cell. Recent measurements of Shewanella’s cytochrome complexes assembled into proteoliposomes indicated average electron transfer rates up to 104 s−1 per complex to external minerals.13 At this rate, 10 — 100 cytochromes out of the experimentally estimated average density of 103-104 outer membrane cytochromes per cell42–44 are needed to satisfy our measured extracellular respiration rates in whole cells. From these results, and assuming a uniform distribution of cytochromes on the outer membrane, <10% of the cell’s surface area is required to make contact with the electron-accepting surface. The type of measurement demonstrated here therefore provides a figure of merit that can be used to quantitatively assess the validity and impact of specific EET pathways.

We described an integrated experimental system that combines optical trapping, transparent electrochemical chips displaying potentiostatically controlled indium tin oxide microelectrodes, and a biological perfusion chamber for single cell measurements of extracellular electron transfer in bacteria. The system is capable of manipulating individual bacteria and detecting their specific extracellular respiration rates down to currents <100 fA. The trapping, microfabrication, and measurement protocols were optimized to address several challenges surrounding the biocompatibility, feature resolution, and electrochemical sensitivity needed to detect single cell response. We demonstrated the system with single cell respiration measurements of the dissimilatory-metal reducing bacterium Shewanella oneidensis MR-1, which ranged from 15 fA to 100 fA per cell, compared to no current output from mutants lacking the outer-membrane cytochromes necessary for extracellular electron transfer. The system helps define the fundamental limit of respiration in microbial cultures and biofilms, down to the single cell level, which is critical for predicting the maximum power densities of technologies that rely on electron exchange between bacteria and electrode surfaces, including microbial fuel and electrosynthesis cells. Applications of this instrument also include discovering the extracellular electron transfer pathways in living cells, by studying specific mutations, and quantifying the impact of environmental factors on the heterogeneity inherent in microbial populations.

The design and construction of the platform were funded by a 2010-2013 Air Force Office of Scientific Research Young Investigator Research Program Grant No. FA9550-10-1-0144. Optimization of the electrochemical chips, respiration measurements, and support for BJG was funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and Biosciences Grant No. DE-FG02-13ER16415. Chip fabrication was performed at the Nanoelectronics Research Facility at the University of California, Los Angeles. Atomic force microscopy and electron microscopy were performed at the University of Southern California Centers of Excellence in NanoBioPhysics and Electron Microscopy and Microanalysis. The authors acknowledge Chuan Huang and James Lu for their support in assembling microelectrode arrays and assistance in cultivation of MR-1 and ΔOMC.

2.
D. N.
Beratan
,
J. N.
Onuchic
,
J. R.
Winkler
, and
H. B.
Gray
,
Science
258
,
1740
(
1992
).
3.
C. R.
Myers
and
K. H.
Nealson
,
Science
240
,
1319
(
1988
).
4.
K. H.
Nealson
,
A.
Belz
, and
B.
McKee
,
Antonie van Leeuwenhoek
81
,
215
(
2002
).
5.
J. K.
Fredrickson
,
M. F.
Romine
,
A. S.
Beliaev
,
J. M.
Auchtung
,
M. E.
Driscoll
,
T. S.
Gardner
,
K. H.
Nealson
,
A. L.
Osterman
,
G.
Pinchuk
,
J. L.
Reed
,
D. A.
Rodionov
,
J. L. M.
Rodrigues
,
D. A.
Saffarini
,
M. H.
Serres
,
A. M.
Spormann
,
I. B.
Zhulin
, and
J. M.
Tiedje
,
Nat. Rev. Microbiol.
6
,
592
(
2008
).
6.
B. E.
Logan
,
Nat. Rev. Microbiol.
7
,
375
(
2009
).
7.
K.
Rabaey
and
R. A.
Rozendal
,
Nat. Rev. Microbiol.
8
,
706
(
2010
).
8.
M. Y.
El-Naggar
and
S. E.
Finkel
,
The Scientist
27
(
5
), 38 (May 2013), available at http://www.the-scientist.com/?articles.view/articleNo/35299/title/Live-Wires/.
9.
E.
Marsili
,
D. B.
Baron
,
I. D.
Shikhare
,
D.
Coursolle
,
J. A.
Gralnick
, and
D. R.
Bond
,
Proc. Natl. Acad. Sci. U. S. A.
105
,
3968
(
2008
).
10.
H.
von Canstein
,
J.
Ogawa
,
S.
Shimizu
, and
J. R.
Lloyd
,
Appl. Environ. Microbiol.
74
,
615
(
2008
).
11.
C. R.
Myers
and
J. M.
Myers
,
J. Bacteriol.
174
,
3429
(
1992
).
12.
R. S.
Hartshorne
,
C. L.
Reardon
,
D.
Ross
,
J.
Nuester
,
T. A.
Clarke
,
A. J.
Gates
,
P. C.
Mills
,
J. K.
Fredrickson
,
J. M.
Zachara
,
L.
Shi
,
A. S.
Beliaev
,
M. J.
Marshall
,
M.
Tien
,
S.
Brantley
,
J. N.
Butt
, and
D. J.
Richardson
,
Proc. Natl. Acad. Sci. U. S. A.
106
,
22169
(
2009
).
13.
G. F.
White
,
Z.
Shi
,
L.
Shi
,
Z. M.
Wang
,
A. C.
Dohnalkova
,
M. J.
Marshall
,
J. K.
Fredrickson
,
J. M.
Zachara
,
J. N.
Butt
,
D. J.
Richardson
, and
T. A.
Clarke
,
Proc. Natl. Acad. Sci. U. S. A.
110
,
6346
(
2013
).
14.
Y. A.
Gorby
,
S.
Yanina
,
J. S.
McLean
,
K. M.
Rosso
,
D.
Moyles
,
A.
Dohnalkova
,
T. J.
Beveridge
,
I. S.
Chang
,
B. H.
Kim
,
K. S.
Kim
,
D. E.
Culley
,
S. B.
Reed
,
M. F.
Romine
,
D. A.
Saffarini
,
E. A.
Hill
,
L.
Shi
,
D. A.
Elias
,
D. W.
Kennedy
,
G.
Pinchuk
,
K.
Watanabe
,
S.
Ishii
,
B.
Logan
,
K. H.
Nealson
, and
J. K.
Fredrickson
,
Proc. Natl. Acad. Sci. U. S. A.
103
,
11358
(
2006
).
15.
M. Y.
El-Naggar
,
G.
Wanger
,
K. M.
Leung
,
T. D.
Yuzvinsky
,
G.
Southam
,
J.
Yang
,
W. M.
Lau
,
K. H.
Nealson
, and
Y. A.
Gorby
,
Proc. Natl. Acad. Sci. U. S. A.
107
,
18127
(
2010
).
16.
S.
Pirbadian
and
M. Y.
El-Naggar
,
Phys. Chem. Chem. Phys.
14
,
13802
(
2012
).
17.
K. M.
Leung
,
G.
Wanger
,
M. Y.
El-Naggar
,
Y.
Gorby
,
G.
Southam
,
W. M.
Lau
, and
J.
Yang
,
Nano Lett.
13
,
2407
(
2013
).
18.
S.
Pirbadian
,
S. E.
Barchinger
,
K. M.
Leung
,
H. S.
Byun
,
Y.
Jangir
,
R. A.
Bouhenni
,
S. B.
Reed
,
M. F.
Romine
,
D. A.
Saffarini
,
L.
Shi
,
Y. A.
Gorby
,
J. H.
Golbeck
, and
M. Y.
El-Naggar
,
Proc. Natl. Acad. Sci. U. S. A.
111
,
12883
(
2014
).
19.
C. I.
Torres
,
A. K.
Marcus
,
H. S.
Lee
,
P.
Parameswaran
,
R.
Krajmalnik-Brown
, and
B. E.
Rittmann
,
FEMS Microbiol. Rev.
34
,
3
(
2010
).
20.
A. K.
Marcus
,
C. I.
Torres
, and
B. E.
Rittmann
,
Biotechnol. Bioeng.
98
,
1171
(
2007
).
21.
O.
Bretschger
,
A.
Obraztsova
,
C. A.
Sturm
,
I. S.
Chang
,
Y. A.
Gorby
,
S. B.
Reed
,
D. E.
Culley
,
C. L.
Reardon
,
S.
Barua
,
M. F.
Romine
,
J.
Zhou
,
A. S.
Beliaev
,
R.
Bouhenni
,
D.
Saffarini
,
F.
Mansfeld
,
B.-H.
Kim
,
J. K.
Fredrickson
, and
K. H.
Nealson
,
Appl. Environ. Microbiol.
73
,
7003
(
2007
).
22.
D. J.
Richardson
,
J. N.
Butt
,
J. K.
Fredrickson
,
J. M.
Zachara
,
L.
Shi
,
M. J.
Edwards
,
G.
White
,
N.
Baiden
,
A. J.
Gates
,
S. J.
Marritt
, and
T. A.
Clarke
,
Mol. Microbiol.
85
,
201
(
2012
).
23.
M. E.
Lidstrom
and
M. C.
Konopka
,
Nat. Chem. Biol.
6
,
705
(
2010
).
24.
H. M.
Davey
and
D. B.
Kell
,
Microbiol. Rev.
60
,
641
(
1996
).
25.
J. S.
McLean
,
G.
Wanger
,
Y. A.
Gorby
,
M.
Wainstein
,
J.
McQuaid
,
S. I.
Ishii
,
O.
Bretschger
,
H.
Beyenal
, and
K. H.
Nealson
,
Environ. Sci. Technol.
44
,
2721
(
2010
).
26.
X. C.
Jiang
,
J. S.
Hu
,
E. R.
Petersen
,
L. A.
Fitzgerald
,
C. S.
Jackan
,
A. M.
Lieber
,
B. R.
Ringeisen
,
C. M.
Lieber
, and
J. C.
Biffinger
,
Nat. Commun.
4
,
2751
(
2013
).
27.
H. A.
Liu
,
G. J.
Newton
,
R.
Nakamura
,
K.
Hashimoto
, and
S.
Nakanishi
,
Angew. Chem., Int. Ed.
49
,
6596
(
2010
).
28.
A.
Ashkin
,
IEEE J. Sel. Top. Quantum Electron.
6
,
841
(
2000
).
29.
K. C.
Neuman
and
S. M.
Block
,
Rev. Sci. Instrum.
75
,
2787
(
2004
).
30.
D. C.
Appleyard
,
K. Y.
Vandermeulen
,
H.
Lee
, and
M. J.
Lang
,
Am. J. Phys.
75
,
5
(
2007
).
31.
K. C.
Neuman
,
E. H.
Chadd
,
G. F.
Liou
,
K.
Bergman
, and
S. M.
Block
,
Biophys. J.
77
,
2856
(
1999
).
32.
T. L.
Min
,
P. J.
Mears
,
L. M.
Chubiz
,
I.
Golding
,
Y. R.
Chemla
, and
C. V.
Rao
,
Nat. Methods
6
,
831
(
2009
).
33.
A.
Ashkin
,
J. M.
Dziedzic
, and
T.
Yamane
,
Nature
330
,
769
(
1987
).
34.
L.
Hao
,
X. G.
Diao
,
H. Z.
Xu
,
B. X.
Gu
, and
T. M.
Wang
,
Appl. Surf. Sci.
254
,
3504
(
2008
).
35.
A. J.
Bard
and
L. R.
Faulkner
,
Electrochemical Methods: Fundamentals and Applications
, 2nd ed. (
Wiley
,
New York
,
2001
), p.
833
.
36.
R. R.
Unocic
,
R. L.
Sacci
,
G. M.
Brown
,
G. M.
Veith
,
N. J.
Dudney
,
K. L.
More
,
F. S.
Walden
,
D. S.
Gardiner
,
J.
Damiano
, and
D. P.
Nackashi
,
Microsc. Microanal.
20
,
452
(
2014
).
37.
H. S.
Harned
and
R. M.
Hudson
,
J. Am. Chem. Soc.
73
,
5083
(
1951
).
38.
C.
Bucking
,
F.
Popp
,
S.
Kerzenmacher
, and
J.
Gescher
,
FEMS Microbiol. Lett.
306
,
144
(
2010
).
39.
H. W.
Harris
,
M. Y.
El-Naggar
, and
K. H.
Nealson
,
Biochem. Soc. Trans.
40
,
1167
(
2012
).
40.
J. N.
Roy
,
S.
Babanova
,
K. E.
Garcia
,
J.
Cornejo
,
L. K.
Ista
, and
P.
Atanassov
,
Electrochim. Acta
126
,
3
(
2014
).
41.
V. B.
Wang
,
N. D.
Kirchhofer
,
X. F.
Chen
,
M. Y. L.
Tan
,
K.
Sivakumar
,
B.
Cao
,
Q. C.
Zhang
,
S.
Kjelleberg
,
G. C.
Bazan
,
S. C. J.
Loo
, and
E.
Marsili
,
Electrochem. Commun.
41
,
55
(
2014
).
42.
B. H.
Lower
,
L.
Shi
,
R.
Yongsunthon
,
T. C.
Droubay
,
D. E.
McCready
, and
S. K.
Lower
,
J. Bacteriol.
189
,
4944
(
2007
).
43.
J.
Borloo
,
B.
Vergauwen
,
L.
De Smet
,
A.
Brige
,
B.
Motte
,
B.
Devreese
, and
J.
Van Beeumen
,
FEBS J.
274
,
3728
(
2007
).
44.
D. E.
Ross
,
S. L.
Brantley
, and
M.
Tien
,
Appl. Environ. Microbiol.
75
,
5218
(
2009
).