Label-Free Detection of Single Living Bacteria via Electrochemical Collision Event

We detected single living bacterial cells on ultramicroelectrode (UME) using a single-particle collision method and optical microscopic methods. The number of collision events involving the bacterial cells indicated in current-time (i-t) curves corresponds to the number of bacterial cells (i.e., Escherichia coli) on the UME surface, as observed visually. Simulations were performed to determine the theoretical current response (75 pA) and frequency (0.47 pM−1 s−1) of single Escherichia coli collisions. The experimental current response (83 pA) and frequency (0.26 pM−1 s−1) were on the same order of magnitude as the theoretical values. This single-particle collision approach facilitates detecting living bacteria and determining their concentration in solution and could be widely applied to studying other bacteria and biomolecules.

Single-particle collisions have attracted considerable attention for over a decade because this technique provides important information on single-particle properties, such as size, concentration, surface charge, and lifetime 1 . Initially, micro-/nanoscale beads composed of non-conducting rigid materials were employed to demonstrate this method's working principle, which is based on blocking the flux of redox species toward an ultramicroelectrode (UME) through the collision of single particles at the UME, thereby decreasing the electrode reaction 2,3 . Interesting properties of various rigid particles have been observed using this unique analytical platform [4][5][6] .
Recent publications have broadened the range of materials that can be studied via this single-particlecollision-based technique, including soft particles (e.g., emulsions) 7,8 . Since the introduction of the emulsion particle collision technique, biomolecule-detection studies investigating viruses, proteins, and DNA have been reported 9,10 . Indeed, this technique enabled us to characterize individual particles consisting of such soft materials 11 . Interesting features of various single biomolecules have been observed 12,13 , but label-free detection of single living microbes has not been reported to date. Because a single unit of living matter (e.g., a bacterium) can proliferate over time and ultimately generate significant phenomena (e.g., disease), the rapid, real-time detection of such units via single-particle-collision-based analysis can provide valuable information about and/or control over living microbes.
The detection and quantification of bacteria are in high demand in the fields of clinical pathology, food science, and biology. However, it is difficult to monitor or capture living bacteria, which are small and move by swimming, swarming, gliding, and twitching in solution 14 . Furthermore, the rapid self-proliferation of living cells increases the difficulty of estimating their exact concentrations 15 . Currently, a variety of approaches are available to measure bacteria in samples 16,17 : optical density, microscopy techniques, and cell cultivation [18][19][20][21] . Although each technique has unique advantages, they are associated with some limitations that hinder their wide application. For example, using cultivation methods, the bacterial concentration can be obtained in ~5 days; thus, these methods are time consuming. Spectrophotometrically measuring the optical density and correlating it with the bacterial concentration is rapid and simple, but the specific response factors for the investigated bacteria and media must first be calculated; additionally, the sensitivity and accuracy of this method are not sufficiently high. Microscopic counting is a straightforward method to quantify bacteria. However, this method cannot be used for small (less than 2 μ m) bacteria. Therefore, a fast and reliable technique able to detect specific bacteria and determine their concentrations is urgently needed.
Here, we report the detection of a single living bacteria on a UME. The concentration of living bacteria can be measured based on the number of captured cells at the UME surface, and the electrochemical signal can be used to estimate the size of the bacteria. Escherichia coli (E. coli) which is a rod-shaped, gram-negative bacteria with a length of ca. 2 μ m was chosen as a model bacteria in this work. The mobile living bacteria were detected at the UME according to two sequential strategies: 1) electrophoretic migration and 2) blockage of the electroactive area. To observe an E. coli collision event, a redox species is continuously oxidized at the UME surface ( Fig. 1). When the ferrocyanide (Fe(CN) 6 4─ ) is oxidized at the electrode surface, a positive electric field is developed by the steady-state current flow. Therefore, the negatively charged E. coli is attracted to the UME surface through electrophoretic migration. When an E. coli collides with and then attaches to the UME surface, the level of the steady-state current decreases immediately because the flux of the redox species is blocked by the E. coli. Therefore, a staircase current response is observed (Fig. 1).

Results and Discussion
Determination of experimental conditions for electrochemical collision events. To distinguish a single living microbial collision event based on UME surface blockage, the experimental conditions must be carefully chosen. First, the solution composition should be adjusted. The single-particle blocking experiment was typically performed in the presence of a high concentration of redox species (e.g., 400 mM Fe(CN) 6

4─
) to obtain a high current intensity 7 . However, such high concentrations of redox species may cause unexpected damage to the living E. coli. Therefore, relatively low Fe(CN) 6 4─ concentrations able to produce observable collision responses were investigated. Four different concentrations of redox species were tested for single E. coli collisions. As shown in Fig. 2, staircase current responses were observed upon single E. coli collision for all studied concentrations, and the current step height was proportional to the redox species concentration ( Supplementary Fig. S1). In the presence of 20 mM Fe(CN) 6 4─ , E. coli collision resulted in a response of tens of pA, which exceeds the instrumental lower limit (a few pA) by one order of magnitude. This concentration is low enough to prevent osmotic cell death but high enough to result in observable signal intensity. The stability of bacterial cells under this condition (20 mM K 4 Fe(CN) 6 ) was confirmed by a cell growth experiment ( Supplementary Fig. S2). Next, the particle sizeto-UME diameter ratio must be optimized to produce an obvious staircase current 3,22 . When the UME diameter was 10 μ m, spherical beads with diameters of 500 nm were marginally detectable by the blocking method. As a practical guideline, this ratio (UME diameter to analyte dimension) should be less than approximately 20. Thereby, a 10 μ m UME was chosen to detect E. coli.

Luria-Bertani medium test for control experiments. A control experiment was performed using an
E. coli-free Luria-Bertani (LB) medium which was used for E. coli cultivation ( Supplementary Fig. S3). No meaningful staircase signal was observed over 500 s in the presence of LB medium which contains several small molecules that are smaller than 340 nm ( Supplementary Fig. S4). These small particles did not cause any detectable staircase current decrease under our experimental conditions. Therefore, this system can be used to detect E. coli without removing LB, which would be highly experimentally useful.
Visual confirmation of sigle E. coli collision on C-UME. To visually confirm the attachment of E. coli on the UME surface, enhanced green fluorescent protein (EGFP)-expressing E. coli was used. After completing the electrochemical measurements, the electrode was gently washed with distilled water to remove the electrolyte salts. The UME surface was carefully observed using various microscopy techniques to identify the existence of E. coli remaining on the UME surface. As shown in Fig. 3A, one collision signal was detected for 90 s. After collecting the electrochemical collision signal, the electrode surface was observed by fluorescent microscopy (Fig. 3B). The green-colored E. coli on the UME surface were clearly observabled. Scanning electron microscopy (SEM) also supports the adsorption of E. coli at the same position ( Supplementary Fig. S5). These microscopic images confirmed that the adsorbed particle on the electrode surface is an E. coli and that the cell cannot be easily detached. Therefore, this method can be used to investigate attached single bacteria.
Detection of multiple E. coli collisions on C-UME. The stable adsorption of E. coli on the electrode surface eliminated the need to develop an alternative method to stabilize the cells and allowed us to conveniently    6 . The C-UME (10 μ m diameter) was biased at + 0.6 V vs Ag/AgCl. Black circles in the fluorescent images indicate the surface of the C-UME. elucidate the relationship between the number of electrochemical staircase current responses and the actual number of EGFP-expressing E. coli that collided with the electrode. An electrode that exhibited multiple collision responses was also investigated. In this case, 7 collision responses were observed over 500 s (Fig. 3C). As shown in Fig. 3D, 7 fluorescent E. coli were present on the UME surface (black circles). The number of adsorbed E. coli was in good agreement with the number of electrochemical collision responses. Thus, this approach can be used to estimate the number of adsorbed E. coli on the UME based on the current transient.
Simulation for analysing current step decrease and collision frequency. The magnitude of the staircase current decrease resulting from a single E. coli collision was estimated via 3-D Comsol Multiphysics simulations (Fig. 4). Depending on the alignment of the E. coli, different current responses would be expected. Thereby, the alignment of the E. coli must also be considered (Fig. 4A). The simulation results imply that the relative size of the current step (∆ i/i lim ), where i lim is the limiting current and ∆ r is the distance from the center of the UME to the E. coli, tends to increase as the landing position nears the edge of the UME, regardless of the E. coli alignment (Fig. 4C). Based on Fig. 4C, the maximum and minimum staircase current heights can be estimated when E. coli of a specific size collide with the UME (in our simulation, an E. coli was assumed to have a cylindrical shape with a radius of 0.4 μ m and a length of 2 μ m). The height of the experimentally observed staircase current The experimental and theoretical collision frequencies of E. coli can be related to the concentration of cells. A 2-D simulation considering the flux of the symmetric UME was performed to obtain the theoretical collision frequency under migrational transport with a steady-state current at the UME (simulation details are shown in the Supplementary Dataset). The model developed here is fairly similar to that described in our previous report 3, 23,24 . Based on this simulation, the expected frequency was ca. 0.47 pM −1 s −1 , which was calculated based on the total flux of E. coli observed on the UME. The experimental (0.26 pM −1 s −1 ) and expected (0.47 pM −1 s −1 ) frequencies are on the same order of magnitude. This simulation considered only the diffusional and attractive migrational movement of E. coli (between positively developed electric field near UME and negatively charged E. coli). This result implies that the complex movement of E. coli hardly affects the overall collision frequency. 25 Opposite electric field (negatively developed electric field near UME) generated by reduction of ruthenium(III) hexamine was tested to confirm the effect of migrational movement of E. coli (Supplementary Fig. S6). As shown in Fig. S6, there was no collision signal detected during the measurement time. From these results, we can confirmed that the collision signals under ferrocyanide condition originated from negatively charged E. coli. Various E. coli concentrations were measured in the same manner ( Supplementary Fig. S7). Supplementary Fig. S7 (collision frequency vs. E. coli concentration) shows good linearity in the concentration range from 10 to 100 fM. The collision frequency was saturated at 212 fM of E. coli because the fully E. coli-covered UME, which had reduced electroactive area, exhibited reduced migrational E. coli flux to its surface. The concentration of cells in the sample could thus be predicted using the theoretical single-particle collision frequency.

Conclusions
In this paper, we have demonstrated the electrochemical label-free detection of single living bacteria using the blocking methodology, as was visually confirmed using EGFP-expressing E. coli. This blocking method is faster and simpler than the conventional methods used to detect living bacteria in solution. The number of collision responses observed in i-t curves correlates with the number of E. coli observed via fluorescence microscopy and predicted by theoretical calculations. The theoretical current responses of single E. coli collisions were obtained by Comsol Multiphysics simulations. These simulation results are also in good agreement with the experimental results. Therefore, this methodology can be applied to measure the concentrations of specific-sized microbes in solution. By observing the cells adhered to the UME, this method can provide a platform to study the properties of single cell after separation and/or attachment onto the UME surface. This approach can also be applied to study other biomolecules. Instrumentation. The electrochemical experiments were performed using CHI900B and 750E potentiostats (CH Instruments, Austin, TX) with a three-electrode cell placed in a Faraday cage. A 1 mm diameter Pt wire was used as a counter electrode, and the reference electrode was Ag/AgCl (3 M KCl). Scanning electron microscopy (SEM) was performed using a Sirion FEI XL FEG/SFEG microscope (FEI Co., The Netherlands) with an accelerating voltage of 2 kV. Fluorescence microscopy images were taken using an Olympus IX71 inverted microscope with a mercury lamp and a CCD camera system (DP70 camera unit and PCI board). Dynamic laser scattering (DLS) data was obtained by Zetasizer Nano ZS (Malvern, Westborough, MA). Nanoparticle tracking analysis (NTA) data was obtained by NanoSight NS300 (Malvern, Westborough, MA). 2D and 3D simulation were performed using Comsol Multiphysics.

Reagents and
Preparation of Carbon Fiber Ultramicroelectrodes (C-UMEs). C-UME was prepared following the general procedure developed in our laboratory. Briefly, a 10 μ m (diameter) carbon fiber was sealed in a borosilicate glass tubing (1.5 mm o.d. × 0.75 mm i.d.) after rinsing with hexane, toluene, IPA, ethanol, and water. The electrode was then polished with an alumina powder water suspension to a mirror finish. The surface area was checked with standard redox electrochemistry in ferrocenemethanol solution. Before every experiment, all electrodes were polished with alumina (0.05 μ m) paste on microcloth pads (Buehler, Lake Bluff, IL) prior to use.