2024 Volume 92 Issue 2 Pages 022015
To perform on-site bacterial testing at food and pharmaceutical manufacturing sites, it is desired to develop a new method that can quickly measure the number of viable bacterial cells. We have succeeded in measuring the number of viable bacteria using small and inexpensive disposable electrode chips focusing on electrochemical methods that realize quick detection and device miniaturization. The oxidized form of the tetrazolium salt, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), is soluble and highly permeable through cell membranes. MTT is a useful indicator for evaluating cell activity that not only turns color as a result of structural changes related to intracellular metabolism but also causes a clear current response. The carbon-screen-printed electrode chip provides a distinct current response related to the MTT redox reaction in a small volume of bacterial suspension (50 µL). Based on the fact that the current reaction of MTT was strongly dependent on intracellular metabolism, the number of viable cells in a bacterial suspension could be measured electrochemically. Current changes for live cells occurred within 10 min and increased with the incubation time. After only 60 min of incubation, we successfully estimated the number of viable cells in a bacterial suspension of 103 CFU mL−1. This technology eliminates the need for complicated testing, expensive equipment, and lengthy culture testing times, thereby enabling the confirmation of food safety before shipping to prevent food poisoning.
Pathogenic Escherichia coli (E. coli) is a well-known causative organism that causes food poisoning.1,2 The ingestion of food and drinking water contaminated with these pathogens causes various symptoms, such as diarrhea, abdominal pain, vomiting, and, in severe cases, hemolytic uremic disease.3 According to a World Health Organization (WHO) and World Bank study, 600 million cases of foodborne illness occur worldwide annually, costing low- and middle-income economies approximately US $110 billion in lost productivity and medical costs. Therefore, rapid detection of pathogenic bacteria is necessary to prevent health hazards such as food poisoning and infectious diseases, as well as reduce economic losses such as medical expenses. Current microbiological testing methods, such as polymerase chain reaction (PCR), enzyme-linked immunosorbent assay (ELISA), and colony counting, require skilled experimentalists as well as complex and time-intensive culture processes.4,5 In recent years, many studies have reported on the development of biosensors using novel materials and technologies. For highly sensitive and rapid detection of pathogenic bacteria, biosensors based on physical or chemical signals, such as light scattering,6–9 fluorescence,10,11 electrochemistry,12–15 and piezoelectricity,16 have been developed. In particular, electrochemical biosensors are particularly useful in fields where on-site testing is required because of their high sensitivity, rapid measurement, and ease of device miniaturization.17–19 In addition to identifying bacterial species such as E. coli, Salmonella enterica, and Staphylococcus aureus at food manufacturing sites, microbial contamination must be determined based on the total number of viable bacteria, regardless of the bacterial species. Therefore, we developed a method to evaluate the total number of viable bacteria based on intracellular electron mediators.20 In living cells, the reduced coenzymes nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FADH2) function to transfer electrons through redox reactions.21–24 Furthermore, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT), which is the oxidized form of the tetrazolium salt, is a soluble dye molecule and has excellent cell membrane permeability; MTT permeates the cell membrane and forms a reduced form in living cells.25–30 MTT receives the electrons produced by the enzyme reaction and is converted into formazan, the reduced form of MTT, which is deposited inside the cell based on its insolubility. The MTT colorimetric assay is useful for assessing the activity of large cells (∼20 µm), such as fungi26,27 and mammalian cells,28–30 as well as bacteria (a few µm).31–34
In a previous study, we measured viable bacteria using a glassy carbon (GC) electrode based on the reduction current of MTT remaining in the suspension and established an electrochemical method for evaluating the viable bacterial activity of bacterial species with different bacterial metabolic processes and cell wall structures.35 However, when measuring live bacteria using a commercially-available GC electrode (total length: 65 mm, outer diameter: 6 mm, electrode diameter: 3 mm) in a field where microorganisms are handled, a large sample solution volume (>5.0 mL) is required. To avoid this, if the electrode is made smaller, the current response decreases; therefore, a highly sensitive potentiostat is required. In this study, we designed a small disposable tip electrode to enable measurements using a small standard potentiostat and a small volume of sample solution (50 µL).
E. coli K12 was purchased from the National Institute of Technology and Evaluation Biological Resource Center (NBRC, Japan). Bacterial cultures and all experiments were developed and managed in a Biosafety Level 2 laboratory in accordance with the appropriate safety regulations (WHO Laboratory Biosafety Manual).6,7 The bacterial strain was cultured in an agar growth medium (E-MC35, Eiken Chemical, Japan) at 310 K for 18 h. A single colony was selected, placed in a liquid growth medium (5.0 mL), and incubated at 310 K for 6 h. The precipitate was dispersed in sterile ultrapure water and the E. coli K12 suspension concentration was adjusted to 3.6 × 1011 CFU mL−1.
2.2 Preparation of MTT-containing electrolyte solutionUltrapure water (>18 MΩ cm) was sterilized using an autoclave (393 K, 20 min). Reagent-grade MTT (Dojindo, Japan) was used as-purchased without further purification. The MTT solution (0.50 µL) was adjusted to a concentration of 20 mM with a 0.10 M (= mol L−1) potassium chloride (KCl) aqueous solution after dissolving MTT salts in a small amount of dimethyl sulfoxide (0.050 µL). To obtain 0.10 mM MTT solution, the MTT solution (0.50 µL) was added to a mixture of 0.20 % glucose-added nutrient broth medium (25 µL) and 0.10 M phosphate-buffered saline (PBS) solution (pH 7.4, 25 µL). To prepare a sample solution, a suspension with a predetermined bacterial density (0.50 µL) was added to a 0.10 mM MTT solution (50 µL).
2.3 Electrochemical measurementsAll the electrochemical measurements were performed using a carbon screen-printed electrode (Fig. 1A). The patterns for the working electrode (surface area: 14 mm2, electrical resistance: 25 Ω/square) and counter electrode (surface area: 20 mm2, electrical resistance: 25 Ω/square) were drawn using carbon ink on the same surface of a resin substrate (30 × 6.0 mm). The Ag/AgCl ink was applied to the reference electrode and allowed to dry for 5 min. After irradiating the electrode with 10 mA plasma for 1 min, a constant potential of −0.5 V was applied for 180 s as a pretreatment. Electrochemical measurements were performed by the dropwise addition of 50 µL of PBS solution (pH 7.4) onto the electrode chip as the electrolyte. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) were performed using a single-channel potentiostat (ECstat-302; EC Frontier Co., Kyoto, Japan). The CV measurements were performed by dropping a 0.10 mM MTT solution (50 µL) onto the electrode chip at a scanning rate of 50 mV s−1 and a starting potential of +0.20 V (vs. Ag|AgCl). As post-processing, DPV measurements were performed in the potential range from 0 to −0.5 V at a scan rate of 10 mV s−1 under the following conditions: pulse amplitude: 50 mV, increasing potential: 5 mV, pulse width: 50 ms, sample width: 17 ms, and pulse period: 500 ms.
(A) Photograph of a carbon screen-printed electrode and (B) CV of 0.10 mM MTT in 0.10 M PBS solution (pH 7.4). The potential sweep ranges from −0.5–0.6 V. Scan rate was 50 mV s−1.
A mixture of MTT and bacterial suspension (5.0 mL, 1.0 × 109 CFU mL−1) was incubated at 310 K for the indicated time (Fig. 2). Based on the absorbance of MTT at 380 nm the yellow-colored suspension gradually turned bluish as the incubation time increased and finally turned deep purple based on the absorbance of formazan at 650 nm.35 The suspension was centrifuged at 10,000 g for 10 min at 278 K, yielding a yellowish supernatant and a purple precipitate. The absorbance at 380 nm based on the residual MTT in the supernatant decreased with incubation time, whereas the absorbance at 650 nm based on deposited MTT formazan in the suspension increased with incubation time. The absorbance of residual MTT decreased moderately during the early stages of incubation (approximately 5 min) and significantly after 10 min. Furthermore, the absorption peak almost disappears after 60 min. This suggests that the minimum total reaction time required for MTT to permeate the cell membrane and generate MTT formazan is less than 10 min, with a maximum time of approximately 60 min. This also indicates that almost all MTT molecules in the suspension were converted to formazan within the cells.
Absorption spectra of (a) suspension and (b) supernatant obtained from E. coli K12 suspension during incubation. Supernatant was obtained from bacterial suspension (1.0 × 109 CFU mL−1) comprising liquid growth medium containing 0.10 mM MTT in 0.10 M PBS solution (pH 7.4). The spectra were obtained every 10 min.
One soluble MTT molecule accepts one proton (H+) and two electrons to form reduced MTT formazan (Scheme 1), based on the following reaction (1):25
| \begin{equation} \text{MTT}^{+} + \text{H}^{+} + 2\text{e}^{-} = \text{formazan} \end{equation} | (1) |
A distinct reduction current peak was observed at −0.2 V (vs. Ag|AgCl) when the MTT was reduced to form formazan (Fig. 1B). We performed electrochemical measurements on live and dead bacteria to clarify the relationship between the electrochemical response of MTT and cellular metabolism (Fig. 3). The DPV was performed with the addition of MTT (0.10 mM) to bacterial suspensions (1 × 108 CFU mL−1) with and without autoclaving (394 K, 20 min). The bacteria in the suspension were killed by autoclaving. The current response of MTT in bacterial suspension without autoclaving decreased by 2.4 µA before and after incubation for 2 h, whereas it was constant in sterilized suspension (0.018 µA). In viable cells, MTT molecules were taken up and reduced one after another by the redox cycle of coenzymes, producing more and more formazan. Insoluble formazan was deposited inside the cells without being eluted outside. In other words, MTT in suspension was converted into formazan and concentrated inside the cells. As a result, the current decreased with the decrease of residual MTT in the suspension. On the other hand, since MTT reduction reaction did not occur in dead cells, no change in the current response was observed. It was thought that MTT might react with free coenzymes in a sterile suspension, but it was concluded that the effect was small due to a unidirectional reduction reaction without a redox cycle.
The chemical structures of MTT+ (left) and formazan (right).
DPVs (A) with and (B) without autoclave sterilization (394 K, 20 min), and (C) difference between live and dead bacteria in electrochemical response at −0.2 V. DPVs obtained (a) before and (b) after incubation for 2 h with 0.10 mM MTT in the bacterial suspension (1 × 108 CFU mL−1). The ΔI was obtained as a difference in current responses at −0.2 V before and after the incubation.
It is well known that the DPV provides a lower background by suppressing the charging current and sensitive current response (refer to Fig. 1). Therefore, in this paper, we performed measurements using DPV. Figure 4A shows the time dependence of the electrochemical response of MTT in E. coli K12 suspension (3.6 × 109 CFU mL−1). The DPV of MTT showed a sharp reduction peak without incubation (0 min). During incubation, the current response gradually decreased. This indicates that the MTT taken up into the cell was reduced to formazan inside the cell and the formed formazan was deposited without diffusion into the electrolyte due to its insolubility.20 Therefore, the net MTT concentration in the electrolyte decreased.35 After 10 min, the yellowish MTT electrolyte solution started to change color, eventually becoming a dark purple suspension after 60 min of incubation. The ΔI was obtained as the difference between the current response before (Io) and after a given time of incubation (I). The ΔI increased rapidly until 60 min, then slowly increased until the MTT reduction current disappeared at 120 min. The time course of the response was strongly dependent on the bacterial density of the suspension (refer to Section 3.1). As mentioned above, the reduction of residual MTT in the suspension upon incubation was attributed to the formation of formazan that progressed sequentially within the cells. It was clear that the current response obtained in the suspension as well as the absorbance of the supernatant (refer to Fig. 2) depended on the residual MTT. The disappearance of the current response after incubation for 2 h indicates that all the MTT in the suspension was incorporated into the bacterial cells (3.6 × 109 CFU mL−1).
(A) DPVs of 0.10 mM MTT in 0.10 M PBS solution (pH 7.4) containing E. coli K12 (3.6 × 109 CFU mL−1) at 310 K. The potential sweep ranges from 0 to −0.5 V at a scan rate of 10 mV s−1. The voltammograms were obtained every 10 min. (B) Dependence of the ΔI on the incubation time. The ΔI was obtained as a difference in current responses at −0.2 V before and after the incubation.
We encountered the problem of large variations in the current response between the electrode chips (Fig. 5). When performing DPV in a 0.10 mM MTT solution using five different electrode chips, we observed a variation of ±0.266 µA (±13.6 %) in the current value. This indicates that electrode surface properties such as electrode area, electrical resistance, and hydrophobicity are nonuniform among the fabricated electrode chips. To accurately evaluate the number of viable bacteria, it was necessary to correct for the variations between electrodes. Therefore, we evaluated the number of viable bacteria using the current values obtained with the 0.10 mM MTT solution as the respective standards (Io) for these five different chips. The ΔI was obtained by subtracting the current response (I) at a given number of viable cells from the (Io) and normalized by dividing it by Io (= ΔI/Io). The variation in ΔI obtained by measuring the same bacterial suspension (103 CFU mL−1) with different electrode chips was ±12.4 %, but normalization reduced the variation (±10.4 %).
DPVs of 0.10 mM MTT in 0.10 M PBS solution (pH 7.4) with five different carbon screen-printed electrodes.
After incubation for 30 min, a significant current response (ΔI/Io) was observed at bacterial densities above 104 CFU mL−1, which increased with the bacterial density (Fig. 6). However, no change in absorbance was observed at 104 CFU mL−1, and a dense bacterial suspension (>107 CFU mL−1) was required to obtain a change. For 60 min of incubation, a significant current response was observed at a bacterial density of 103 CFU mL−1, reaching approximately 1.0 at 109 CFU mL−1. After 90 min of incubation, the current response increased over the entire cell density range. Although there was a trade-off between the incubation time and current sensitivity, it was confirmed that sufficient sensitivity could be obtained at least 60 min. After 60 min of incubation, the current response 0.61 (= ΔI/Io) was obtained corresponding to the number of viable bacteria 1.22 × 108 CFU mL−1 from the response curve. The same suspension was tested using 3M Petrifilm aerobic counting plates, commonly used for colony counting (1.20 × 108 CFU mL−1). We could find a good agreement between the results obtained with our method and the conventional method. Therefore, our method is considered to be useful for realizing faster and more sensitive counting of viable bacteria than conventional methods.
Dependence of the ΔI/Io on the cell density of viable E. coli K12 at the indicated incubation time. The ΔI was obtained as a difference in current responses at −0.2 V before and after the incubation (n = 3). The Io was the current response to 0.10 mM MTT at each electrode.
We developed a new electrochemical method using the MTT dye to assess bacterial viability. For practical applications, we evaluated the bacterial activity using a small and inexpensive disposable carbon screen-printed electrode. The variation in the response between electrodes can be reduced through normalization, making it possible to evaluate the net viable bacterial count. Because the current response obtained using this method is related to intracellular metabolism, it is applicable to all viable bacteria, and even low-density bacteria (approximately 103 CFU mL−1) can be measured after 60 min of incubation. In contrast to existing techniques, this method does not require complex operations, expensive equipment, or lengthy testing times. This technology ensures safety at food and pharmaceutical manufacturing sites and contributes to creating a safe, secure, and comfortable life.
This study was financially supported by a JST START Grant (grant no. JPMJST1916). We also gratefully acknowledge the financial support provided by the Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (A) (KAKENHI grant no. 21H04963) and Grants-in-Aid for Challenging Exploratory Research (grant no. 22K18442).
Hikaru Ikeda: Data curation (Lead), Investigation (Lead), Writing – original draft (Lead)
Akira Tokonami: Data curation (Supporting), Methodology (Supporting)
Akihiro Nakao: Data curation (Supporting), Methodology (Supporting)
Shigeki Nishii: Resources (Lead), Supervision (Supporting)
Masashi Fujita: Supervision (Supporting)
Yojiro Yamamoto: Conceptualization (Supporting), Resources (Supporting), Supervision (Supporting)
Yasuhiro Sadanaga: Supervision (Supporting)
Hiroshi Shiigi: Conceptualization (Lead), Funding acquisition (Lead), Project administration (Lead), Supervision (Lead), Writing – review & editing (Lead)
The authors declare no conflict of interest in the manuscript.
Japan Science and Technology Agency: JPMJST1916
Japan Society for the Promotion of Science: 21H04963
Japan Society for the Promotion of Science: 22K18442
H. Ikeda: ECSJ Student Member
M. Fujita and H. Shiigi: ECSJ Active Members
