Bacterial detection based on polymerase chain reaction and microbead dielectrophoresis characteristics

Abstract

In this study, an electrical DNA detection method was applied to bacterial detection. DNA was extracted from bacteria and amplified by polymerase chain reaction. The microbeads were labelled with amplicons, altering their surface conductance and therefore their dielectrophoresis characteristics. Amplicon‐labelled microbeads could thus be trapped within a high‐strength electric field, where they formed a pearl chain between the electrodes, resulting in an increased conductance between the electrodes. This method reduces the amplicon detection time from 1–2 h to 15 min, compared with the conventional method. The presented method realised quantitative detection of specific bacteria at concentrations above 1 × 10⁵ and 2.4 × 10⁴ CFU/ml for bacterial solutions with and without other bacterial presence, respectively.

1 Introduction

The method of dielectrophoresis (DEP), which refers to the motion of electrically polarised particles in non‐uniform electric fields [1], has been widely used to manipulate biological microparticles, such as bacteria [2], biological cells [3], and DNA [4]. The authors have developed and demonstrated a novel electrical detection method for DNA by DEP of microbeads [5]. In this method, DNA are chemically bound to the surface of dielectric microbeads so that DNA functionalisation alters the DEP characteristics of the microbeads. DNA‐labelled microbeads are trapped between microelectrodes under the action of positive DEP, whereas native beads are not trapped. Combining this dramatic alteration in DEP characteristics with impedance measurement allows rapid and quantitative detection of the amplicons. An electrical detection technique called dielectrophoretic impedance measurement (DEPIM), which was originally developed by our group for bacterial and viral inspection, can be used for the impedance measurement [6].

In this study, we aim to apply this novel electrical detection method to bacterial detection. Contamination of food or water with bacteria, such as Escherichia coli O157:H7, is a serious problem worldwide [7]. To determine the source of a contamination, rapid and sensitive detection of bacteria is required. Immunological methods [8, 9] and genetic methods such as the polymerase chain reaction (PCR) [10] and real‐time PCR [11] have been used for bacterial detection. The high sensitivity, specificity, and speed of PCR‐based methods have led to the development of DNA‐based bacterial detection methods [12]. PCR is used to amplify specific regions of DNA or RNA via enzymatic reaction. Theoretically, PCR exponentially amplifies template DNA or RNA because each thermal cycle doubles the template number [13]. DNA amplified by PCR, or ‘amplicons’, are generally separated by size and detected by agarose gel electrophoresis. Although this method is well established and reliable, it requires rather complicated and time‐consuming manual operations, carried out by experts. Therefore, detection of amplicons using this novel electrical detection method can lead to simple, rapid, and specific detection of bacteria.

The authors have demonstrated the possibility of using this novel electrical detection method for amplicons that amplified from synthesised RNA [5]. However, it has not been applied until now to the detection of DNA extracted from bacteria. In this study, DNA extracted from bacteria was amplified by PCR and the amplicons were used to label microbeads for detection by this novel electrical method. The relationship between the bacterial concentration and the electrodes conductance variation was discussed in this study. Furthermore, the sensitivity and selectivity of bacterial detection by this method were also investigated.

2 Theory

2.1 DEP of microbeads

The DEP force acting on a spherical particle suspended in a medium is given by [14]

$$F_{\mathrm{DEP}}=2\pi r^3{\epsilon }_{\mathrm{m}}\text{Re}\left[K\left(\omega \right)\right]\nabla E^2,$$ (1)

where r is the radius of a spherical particle, ε _m is the permittivity of the medium, E is the magnitude of the applied field, and Re[K (ω)] is the real part of the Clausius–Mossotti (CM) factor. The CM factor is defined as

$$K\left(\omega \right)=\frac{{\epsilon }_{\mathrm{p}}^{\ast }-{\epsilon }_{\mathrm{m}}^{\ast }}{{\epsilon }_{\mathrm{p}}^{\ast }+2{\epsilon }_{\mathrm{m}}^{\ast }},$$ (2)

where ${\epsilon }_{\mathrm{p}}^{\ast }$ is the complex permittivity of the particle and ${\epsilon }_{\mathrm{m}}^{\ast }$ is the complex permittivity of the surrounding medium. For a real dielectric, the complex permittivity is defined as ${\epsilon }^{\ast }=\epsilon -j(\sigma /\omega )$, where ε is the permittivity, σ is the conductivity of the dielectric, and ω is the angular frequency of the applied electric field.

In DEP, the real part of the CM factor is usually used to qualitatively define the direction of DEP. When Re[K (ω)] > 0, the particle is propelled towards the high electric field region and such a motion is termed as positive DEP (p‐DEP). When Re[K (ω)] < 0, the particle is repelled from the high field region, and this motion is called negative DEP (n‐DEP).

The conductivity of a solid dielectric particle, σ _p, can be expressed by the following equation [15, 16]:

$${\sigma }_{\mathrm{p}}={\sigma }_{\mathrm{b}}+\frac{2K_{\mathrm{s}}}{r},$$ (3)

where σ _b and K _s are the bulk conductivity and the surface conductance of the particle, respectively. Equation (3) implies that the conductivity of a solid dielectric particle should be more dependent on the surface conductance when the radius of the particle is small. Therefore, the surface conductance of a small particle would affect the dielectric properties of the particle and the resultant DEP force acting on the particle. Nakano et al. [5] showed that the DEP force changes from negative to positive with increasing surface conductance by theoretical calculation of the dependence of Re[K (ω)] on the surface conductance of dielectric particles of various diameters.

2.2 Concept of bacterial detection using DEP

In our previous study, we demonstrated a new electrical detection method for DNA amplified by PCR [5]. This method is based on the alteration of particle surface conductance when the particle is labelled with DNA. DNA is negatively charged because of the negatively charged phosphate ions in its structure (with one unit of negative charge for each phosphate group). Therefore, DNA labelling on small particles will alter the surface conductance of a particle and the DEP force acting upon the particle.

We applied this new DNA detection method for bacterial detection as shown in Fig. 1. The DNA extracted from bacteria, which should be proportional to bacterial concentration, was amplified by PCR. After amplification, the amplicons were conjugated to the surface of the microbeads. The amplicon‐labelled microbeads were detected by an electrical detection technique called DEPIM, originally developed by Suehiro et al. [6] for bacterial and viral inspection [17, 18], and which acts under positive DEP. Therefore, the particles would be attracted to the edges of an interdigitated electrode array and bridge the electrode gap via a pearl chain formation, resulting in an increased conductance between the electrodes. To ensure positive DEP, the bead must be made more polarisable than the surrounding medium. This was achieved in our experiments by increasing the surface conductance of microbeads by DNA labelling.

Fig. 1.

New DNA detection method for bacterial detection

3 Experimental methods

3.1 Bacterial preparation and DNA extraction

We used a mixed suspension of E. coli strain K12 and yeast cells for experiments. E. coli, which was used as a target bacteria to be selectively detected against background cells, was grown in liquid Luria broth (LB) at 37°C overnight. The yeast, which served as background cells, was grown in liquid yeast peptone dextrose (YPD) at 30°C for 1 day. Before each experiment, 1 ml of E. coli and yeast solution were harvested from the LB and YPD, respectively. The E. coli solution was serially diluted and 50 μl of each diluted solution was inoculated on an LB agar plate and grown at 37°C overnight for colony counting. Similarly, yeast was serial decimal diluted and 50 μl of each diluted solution was inoculated on a YPD agar plate and grown at 30°C for 1 day. Colonies were counted and concentrations of bacterial solutions were calculated. Genomic DNA was extracted from each diluted bacterial solution using the Cell Ease 2 Bacteria kit (Biocosm, Kobe, Japan) following the manufacturer's instructions. Five microliters of bacterial solution were mixed with 4 μl solution comprising CellEase A solution 2 μl and CellEase B solution 2 μl, and incubated at 90°C for 3 min, followed by 70°C for 6 min.

3.2 PCR

Primer sequences were pA (5′‐TAG AGT TTG ATC CTG GCT CAG) and pH (5′‐AAG GAG GTG ATC CAG CCG CA) [19], specifically targeting the 1.5‐kbp segment of the 16S rRNA of E. coli. The 5′ end of pA was tagged with biotin. Five microliters of extracted DNA solution were mixed with PCR solution containing 5 μl of 10 × Ex Taq buffer, 5 μl of dNTPs (2 mM each), 1 μl of pA (10 μM), 1 μl of pH (10 μM), 0.5 μl Takara Ex Taq polymerase, and 32.5 μl double deionised water. Cycling conditions consisted of heating to 94°C for 1 min, followed by 35 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min, followed by a final elongation step at 72°C for 4 min. These amplicons were confirmed by standard agarose gel electrophoresis and the concentration of the amplicons was confirmed using a Qubit^® 3.0 Fluorometer (Life Technologies Japan Ltd, Tokyo, Japan).

3.3 DNA and microbeads combination

Magnetic microbeads (2.8 μm in diameter, Dynabeads^® M‐280 Streptavidin, Life Technologies Ltd, Tokyo, Japan) were used in this experiment to simplify the procedures of beads washing. The surface of the microbeads was coated with streptavidin, which binds specifically to biotin. One microliter of microbeads (7 × 10⁵ beads/μl) were resuspended in 10 μl 2× BW buffer (10 mM Tris‐HCl (pH 7.5), 1 mM EDTA, 2 M NaCl), and then mixed with 10 μl of amplicons. The mixtures of the amplicons and microbeads were incubated at room temperature for 15 min, resulting in the conjugation of amplicons via the biotin–streptavidin interaction. The amplicon‐labelled microbeads were then suspended in 50 μl of deionised water.

3.4 Optical observation of DEP behaviours of DNA‐labelled microbeads

A castellated microelectrode with a narrowest gap of 5 μm was used. DEP behaviours of amplicon‐labelled microbeads were observed under an inverted microscope equipped with a charge‐coupled device camera. Five microliters of the solution containing the amplicon‐labelled microbeads was placed on the microelectrode and covered with a glass cover slip. An AC voltage of 5 V_P–P (peak‐to‐peak) amplitude at 100 kHz was applied to the microelectrode to generate the DEP force.

3.5 DEPIM of DNA‐labelled microbeads

DEPIM for electrical measurement of DEP‐trapped amplicon‐labelled microbeads was performed with an electrical circuit similar to that used in previous studies [17, 18]. The impedance change of the microelectrode was measured on a real‐time basis. Fifty microliters of a solution containing the amplicon‐labelled microbeads was placed on the microelectrode for 1 min prior to DEP and impedance measurements to allow the DNA‐labelled beads to settle onto the microelectrode surface by gravity. This provided experimental consistency by ensuring that most beads were within the influence of the DEP field. An AC voltage of 2 V_P–P and 100 kHz was applied to the microelectrode to generate DEP forces and perform DEPIM.

3.6 Selective detection of target bacteria

Detection of specific bacteria from environmental or clinical samples can require detection of a low number of target bacteria in the presence of a large number of other bacteria. Therefore, the specificity of our detection method was tested with yeast solutions mixed with serially diluted E. coli solutions. The yeast and E. coli were cultured as described before and the concentration of bacterial solutions were confirmed by colony count. The cultured E. coli solution was serially diluted in liquid LB. Then, 2.5 μl of each diluted E. coli solution was mixed with 2.5 μl of cultured yeast solution and used immediately. The DNA extraction and PCR procedures were similarly performed for pure E. coli suspension, as described in Section 3.1. The amplicons were confirmed by standard agarose gel electrophoresis and the concentration of amplicons was confirmed using the Qubit^® 3.0 Fluorometer. The amplicons were labelled on microbeads and used for electrical detection, as described above.

4 Results

Fig. 2 shows the concentration of amplicons from serially diluted E. coli. When the bacterial concentration was above 2.4 × 10⁴ CFU/ml, the concentration of amplicons increased in a semi‐logarithmic way with bacterial concentration. However, DNA extracted from bacteria was not amplified efficiently when bacterial concentration was below 2.4 × 10⁴ CFU/ml.

Fig. 2.

Concentration of amplicons from serially diluted E. coli

Fig. 3 shows optical micrographs of the DEP behaviour of amplicon‐labelled microbeads. Microbeads labelled with amplicons from E. coli at a concentration of 2.4 × 10³ CFU/ml, were repelled from the high electric field regions under the action of n‐DEP. In contrast, microbeads labelled with more amplicons from E. coli at a concentration of 2.4 × 10⁴ CFU/ml, were trapped within the high electric field regions under the action of p‐DEP.

Fig. 3.

Optical micrographs of the DEP behaviour of un‐labelled and amplicon‐labelled microbeads

Fig. 4 compares the DEPIM results obtained with microbeads labelled by amplicons from serially diluted E. coli. For E. coli at concentrations above 2.4 × 10⁴ CFU/ml, the conductance between electrodes increased rapidly after AC voltage was applied and the rate of conductance increase became larger with bacterial concentration. In contrast, for E. coli concentrations below 2.4 × 10⁴ CFU/ml, conductance between electrodes decreased after AC voltage was applied.

Fig. 4.

DEPIM results obtained with microbeads labelled by amplicons from serially diluted E. coli

Fig. 5 shows the amplification of extracted DNA from mixtures of yeast solution and serially diluted E. coli. The gel electrophoresis results indicate that the concentration of amplicons increased with the mixed E. coli concentration and DNA extracted from yeast was not amplified by PCR with primer pA and pH.

Fig. 5.

Amplification of extracted DNA from mixtures of yeast solution and serially diluted E. coli

Fig. 6 shows DEPIM results obtained with microbeads labelled by amplicons of mixtures from yeast and serially diluted E. coli. For solutions containing only yeast, at a concentration of 6.5 × 10⁷ CFU/ml, the conductance between the electrodes decreased after AC voltage was applied. When the mixture contained E. coli at concentrations above 10⁵ CFU/ml, the conductance between the electrodes increased rapidly after the AC voltage was applied and the rate of conductance increase became larger with increasing E. coli concentrations.

Fig. 6.

DEPIM results obtained with microbeads labelled by amplicons of mixtures from yeast and serially diluted E. coli

5 Discussion

As shown in Fig. 2, the concentration of amplicons depended on bacterial concentration. It is reasonable to expect that a higher bacterial concentration would result in a higher concentration of extracted DNA, providing higher levels of template DNA for PCR amplification. The amount of template DNA in the reaction determines the amount of amplified product because PCR is theoretically an exponential amplification of template DNA [13]. Therefore, a higher bacterial concentration results in a higher amplicon concentration, providing greater numbers of amplicons for microbead labelling. The surface conductance of the microbeads would increase as more DNA is attached to their surfaces because of the negative charges of DNA, resulting in an increase in Re[K (ω)], as shown in our previous study [5]. Therefore, the p‐DEP force will become stronger as more DNA is labelled on the microbeads, based on (1)–(3). The microbeads would be attracted to the gap between the electrodes more rapidly with stronger p‐DEP forces, leading to a higher rate of increase in conductance between the electrodes. Therefore, the rate of increase in conductance between the electrodes depends on the initial bacterial concentration, as shown in Fig. 4. For E. coli at concentrations below 2.4 × 10⁴ CFU/ml, the conductance between electrodes decreased after AC voltage was applied. This is because the concentration of amplicons is not detectable when the bacterial concentration is below 2.4 × 10⁴ CFU/ml, as shown in Fig. 2. The quantity of amplicon labelling would thus not be sufficient to alter the DEP behaviour of microbeads from n‐DEP to p‐DEP and microbeads would consequently be repelled from the high electric field region under n‐DEP, as shown in Fig. 3 a, while the conductance between microelectrodes would decrease. Therefore, as shown in Figs. 4 and 6, E. coli can be detected by this method at concentrations above 1 × 10⁵ and 2.4 × 10⁴ CFU/ml for bacterial solutions with and without yeast presence, respectively.

The relationship between bacterial concentration and the initial rate of increase in microelectrode conductance (tangent slope of conductance increase at time t  = 0) in Fig. 4, is depicted in Fig. 7. For bacteria at concentrations below 2.4 × 10⁴ CFU/ml, the rate of conductance increase was nearly zero. However, for bacteria at concentrations above 2.4 × 10⁴ CFU/ml, the rate of conductance increase grew in a semi‐logarithmic way with bacterial concentration. Therefore, this electrical detection method has potential for quantitative detection of bacterial at concentrations above 2.4 × 10⁴ CFU/ml. Based on (3), it would seem possible that the sensitivity of this method would improve if smaller beads were used. However, reduction of microbeads size has turned out to be an ineffective way to improve the sensitivity, as shown in Fig. 8. Fig. 8 shows the DEPIM results obtained with different size microbeads (2.8 and 1 μm, respectively) labelled by amplicons at different concentration (24.5 and 4.14 ng/μl, respectively). It indicated that the conductance change between electrodes would reduce when smaller beads were used. As smaller beads were used, it requires more beads to form the pearl chain between electrode gaps. Therefore, although smaller beads may lead to higher surface conductivity, it would not increase the sensitivity of this system. As shown in Fig. 2, DNA extracted from bacteria was not amplified efficiently when the bacterial concentration was below 2.4 × 10⁴ CFU/ml. Therefore, the sensitivity of this method depends strongly on the PCR sensitivity. However, PCR sensitivity depends on the primer used and the reaction conditions. For instance, the sensitivity of the conventional PCR primer targeting the fgbE gene specifically for E. coli is 2.95 × 10³ copies/μl [20], which is much lower than the sensitivity of the PCR primer targeting 16S rRNA used in this study. The sensitivity of this method for bacterial detection can be improved by changing the primer and PCR reaction conditions as required. In this research, we used 5 μl of bacterial solution for detection. The present detection limit of E. coli at concentrations of 2.4 × 10⁴ CFU/ml indicates that E. coli could be detected at a level of 120 CFU. As the detection limit of this detection method depends on the amount of bacteria used for detection, the sensitivity of this detection method could be improved by applying pre‐concentration, such as centrifugation or immunomagnetic separation [21].

Fig. 7.

Relationship between bacterial concentration and the initial rate of increase in microelectrode conductance

Fig. 8.

DEPIM results obtained with different size microbeads

The specificity of our assay is a function of PCR. As shown in Fig. 5, DNA was not amplified from yeast by PCR with primers pA and pH, which specifically target 16S rRNA for E. coli. Without amplicon labelling of the microbeads, the microbeads would be repelled from the high electric field region and the conductance between microelectrodes would decrease, as shown in Fig. 6. The relationship between E. coli concentration mixed with yeast cells and the rate of conductance increase at time t  = 0 obtained from Fig. 6 is depicted in Fig. 9. The rate of conductance increase grew in a semi‐logarithmic way with mixed E. coli concentrations above 10⁵ CFU/ml. This implies that this electrical detection method has potential for quantitative detection of specific bacteria at concentrations above 10⁵ CFU/ml, in the presence of other bacteria in the solution. This method can also be applied for more specific detection of bacteria by applying primers targeting specific gene sequences. For example, primers for the eaeA gene, which is specific to E. coli O157:H7 and O55:H7 [22], could lead to specific detection of these species.

Fig. 9.

Relationship between E. coli concentration mixed with yeast cells and the initial rate of increase in microelectrode conductance

Traditionally, methods for detection of bacteria are culture based, which is far more time consuming than PCR‐based methods. For example, Arthur et al. [23] showed that the detection times of culture‐based methods (21–48 h) were at least 9 h longer than those of a PCR‐based method (7.5–12 h) to detect E. coli in ground beef. However, after DNA amplification, amplicons are generally detected by agarose gel electrophoresis, which take from 1 to 2 h. Our new detection method, using DEP of microbeads, could reduce the amplicon detection time to 15 min by simply mixing amplicons with microbeads for electric detection. This feature offers a more rapid and simple bacterial detection. Furthermore, this method has the potential for automatic detection through use of magnetic microbeads, which allow automated handling for separation and re‐suspension. Although we used magnetic microbeads in this study, we believe it is possible to use non‐magnetic beads instead of magnetic beads in this method since the DEP change is mainly based on the surface conductance change of the microbeads. In our method, the amplicons are immobilised on dielectric microbeads through the strong biotin–avidin affinity reaction. Holmberg et al. [24] reported that the biotin–streptavidin interaction can be reversibly broken in water at elevated temperatures. Therefore, amplicons detected by this method can be retrieved and used for further applications, such as DNA sequencing for bacterial identification [25].

6 Conclusions

We have applied a new method for the detection of amplicons to bacterial detection. The method is based on the surface conductance dependence of microbead DEP. The DEP force of amplicon‐labelled microbeads is strengthened with increasing quantities of labelled DNA, which is related to the initial concentration of bacteria. Therefore, this method has strong potential for highly selective quantitative detection of bacteria. The sensitivity and selectivity of this detection method depend strongly on the efficiency of the PCR step, which can be adjusted for various purposes. The proposed microbead‐based assay reduces the amplicon detection time from 1–2 h to 15 min, compared with conventional methods, and offers the possibility of quantitative and automated diagnosis of specific bacteria.