DNA-induced changes in traveling wave dielectrophoresis velocity of microparticles

Dielectrophoresis (DEP) of dielectric particles occurs based on their polarization in non-uniform electric fields.^1,2 It has been applied to the separation, collection, and alignment of dielectric materials, especially for nano-/micro-sized materials.^3–6 Because DEP depends on the polarizability of a target in a specific surrounding medium, researchers have attempted to use this phenomenon for their characterization. Polarization of biological materials, especially cells, which are modeled as core–shell structures, depends on their electrophysiological characteristics, and DEP can be used to identify such characteristics. DEP can characterize various bacterial^7,8 and cancer cells when the frequency of the applied voltage is varied,^9,10 as parameters, such as a cell’s diameter, its ionic strength, the thickness and dielectric constant of the cell membrane and cell wall, and the existence of nuclei, affect its electrophysiology.

The polarization of homogeneous spherical particles in a dielectric medium can be expressed using the Clausius–Mossotti (CM) factor as follows:

where $ε p *$ and $ε m *$ denote the complex permittivity of the dielectric particles and surrounding medium, respectively. Here, complex permittivity is expressed as ε^* = ε − jσ/ω, where ε and σ are the permittivity and conductivity of the relevant material, respectively, and ω is the angular frequency (ω = 2πf, f: frequency) of the applied voltage. The DEP force can be expressed as

where ε_m, r, and E are the permittivity of the surrounding medium, the radius of the particle, and the magnitude of the electric field, respectively. In the case of a homogeneous spherical particle, Re[CM] ranges from −0.5 to 1. If Re[CM] > 0, the particle is attracted to strong electric field regions, referred to as positive DEP (p-DEP). Conversely, if Re[CM] < 0, negative DEP (n-DEP) occurs, i.e., the particle is attracted to weak electric field regions such that the particle is repelled from the microelectrode edge.

For small particles, σ_p can be written as

where K_S is the surface conductance of the particle, comprising the conductance of the Stern layer (proportional to the surface charge density of the particle) and the diffuse layer. σ_b is the bulk conductivity of the particle material. For polymer-based microparticles, such as polystyrene, σ_b is almost zero. Hence, the CM factor varies with particle size because of σ_p, indicating that the magnitude and direction of the DEP force depend on its K_S for small particles.^11,12 Changes in the DEP force of microparticles by attaching biomolecules such as antibody and DNA have been investigated. In the literature, a change in electric charge density on the surface changed DEP.^11,13

Recently, we proposed a new DNA detection method based on altering a microparticle’s DEP by changing its K_S.^14–16 This method employed microparticles (∼3 µm in diameter) that exhibited n-DEP natively. We subsequently labeled these microparticles with DNA specifically amplified from a target detection species and demonstrated that these microparticles exhibited p-DEP. Hence, only DNA-labeled microparticles were collected on the microelectrode, while bare microparticles were repelled from the microelectrode. As the collection of DNA-labeled microparticles can be detected through changes in the microelectrode impedance, the target DNA concentration can be estimated through changes in the temporal response of impedance.¹⁶ 

While Re[CM] gives the in-phase component of the dipole moment of a dielectric particle in an electric field, Im[CM] gives the out-of-phase component of that.¹⁷ When polyphase AC voltages are applied to a planar, horizontal electrode array such as interdigitated microelectrodes, a traveling electrostatic potential wave is generated. This electrostatic potential wave moves the dielectric particle perpendicular to the electrode array through traveling-wave DEP (twDEP). The twDEP force, F_twDEP, is given as

where λ is the wavelength of the traveling wave, whose 90° phase shift in voltage corresponds to the distance between every fourth electrode. Figure S1 shows a schematic illustrating the twDEP on a microparticle when voltages with 90° phase differences are applied to a microelectrode array. Similar to F_DEP, F_twDEP also depends on the CM factor. Hence, twDEP is used to separate different cells.^18,19

This study aimed to demonstrate DNA-induced changes to microparticle twDEP. Our previous studies focused on Re[CM] changes caused by attaching DNA to dielectric microparticles. However, changing some of the parameters related to Re[CM] will also affect Im[CM]. This is illustrated in Fig. 1(a), which depicts plots of the variation in Re[CM] and Im[CM] of a latex microparticle with respect to changes to its K_S. Attaching DNA to a microparticle surface increases its K_S because of the electric charges on the DNA. For our previous method, it was necessary that sufficient DNA is attached to a microparticle such that Re[CM] increases from a negative to a positive value. Therefore, this technique was combined with the DNA amplification reaction, such as polymerase chain reaction (PCR), to increase the DNA amount to requisite levels. In contrast, while twDEP is proportional to Im[CM], the maximum variation in its magnitude occurs when the value of K_S is similar to that of a microparticle (∼1 nS). Hence, twDEP peaks at around the zero-crossing point of Re[CM] suggest that if a small DNA amount can induce a change in twDEP in the n-DEP region, a sensitive DNA detection method can be realized [Fig. 1(b)]. Here, we investigated how the amount of DNA labeling changes the twDEP of a microparticle.

For our experiments, we prepared short (391 bp) double-stranded DNA segments using PCR. The DNA was amplified from pUC 19 DNA (Takara bio) using forward (5′-biotin-TEG-TTG CCG GGA AGC TAG AGT AA) and reverse (5′-GCT ATG TGG CGC GGT ATT AT) primers. The 5′-terminus of the forward primer was modified with biotin. After the PCR, the amplicon was purified by using a Microspin S-300 HR Column (GE Healthcare) to remove excess primers. The amplicon was confirmed by an agarose gel electrophoresis. The concentration of the amplicon was measured by using a Qubit 3.0 fluorometer (Life Technologies). The DNA segments were labeled with biotin at one of the 5′ ends to enable their attachment onto 2.8-μm-wide magnetic microbeads (Dynabeads M-280 Streptavidin, Life Technologies) used as microparticles. The microparticle surfaces were coated with streptavidin for specific binding to biotin. For labeling, the microparticles (1.2 × 10⁴/μl) and DNA were suspended in 50 µl binding solution [5 mM Tris–HCl (pH 7.5), 0.5 mM EDTA, 1M NaCl]. Following incubation at room temperature for 30 min, the DNA-labeled microparticles were washed and suspended in 100 µl deionized water. The electric conductivity of the final microparticle suspension was 4 × 10⁻⁴ S/m.

A microelectrode array consisting of 30 interdigitated electrodes with a 5 µm gap length was fabricated from indium tin oxide (ITO) on a glass substrate using standard photolithography processes (Figure S2 a). A liquid reservoir, in which the microparticle solution was placed, was constructed by placing a round silicone rubber sheet (0.5 mm thick) with a 7-mm-wide hole on the microelectrode. The microelectrode was connected to a function generator (WF1968, NF) controlled by using a LabView program. Macroscopic view of the microelectrode with the connector is shown in Figure S2 b. The microparticles were observed using an inverted microscope equipped with a charge-coupled device camera. Microparticle movement was recorded at 100 fps for velocity analysis.

First, 20 µl of the DNA-labeled microparticle mixture was poured in the reservoir. Once the microparticles had completely sunk to the bottom, polyphase AC voltage (6 V_PP, 100 kHz) was applied to the microelectrode array. To generate the twDEP force on the microparticles, the microelectrodes were operated with a 90° phase shift successive neighbor.

The velocity of microparticles was analyzed using Image-Pro (Media Cybernetics, Inc.), which identifies individual particles in an image and tracks their movement frame by frame. In this way, the trajectories of hundreds of microparticles could be analyzed in a set of captured images.

While this study focuses on twDEP, it should be noted that an n-DEP force is also generated when voltage is applied, which determines the microparticles’ buoyancy. Because, as shown in Eq. (4), twDEP depends on E² and the magnitude of the twDEP force on a microparticle depends on its position. A previous study has revealed that the twDEP force is not steady near the microelectrode.²⁰ Hence, in this study, the microscope was focused to 10 µm above the microelectrode surface. During particle tracking, unfocused microparticles were removed manually such that only particles on the same vertical plane (i.e., particles experiencing identical twDEP forces) were selected for image analysis. A typical example of trajectories determined from particle tracking analysis is shown in Figure S2 c.

On applying voltage, microparticles in the suspension moved perpendicular to the microelectrode. Velocity profiles for microparticles functionalized with varying amounts of DNA (0, 10²,10³, and 10⁴ DNA molecules per microparticle (DNA/mp)) are shown in Figs. 2(a)–2(d). Approximately 100 particles were analyzed for each DNA labeling condition. Our experiments clearly show that the velocity of the microparticles increased with an increasing amount of DNA, suggesting that |Im[CM]| increased. The direction of twDEP motion was opposite to the propagation of the traveling wave of electric potential, suggesting that Im[CM] < 0. Figure 2(e) shows a plot of the average twDEP velocity of the microparticles (from measurements) with respect to the number of DNA molecules on its surface, demonstrating that this velocity is proportional to the DNA amount.

The measured microparticle velocities are distributed around a range of values, as is typical of particles in a flow.^21,22 However, in this case, we suggest that the non-uniform microparticle size and labeling efficiency also affect variation particle velocity variation. Because of imperfect labeling, it is difficult to distinguish the average velocity of sparsely coated microparticles (100 DNA/mp) from that of bare microparticles. Figure 2(a) shows the velocity distribution of bare microparticles, representing an average velocity of 56.4 ± 7.25 µm/s. When the labeling reaction was conducted at a ratio of 100 DNA/mp, a new peak was observed at a higher velocity [see Fig. 2(b)], in addition to the original peak. In this condition, the DNA concentration in the labeling mixture was very low. It is, thus, possible that the labeling reaction did not achieve saturation in 30 min. This provides an explanation for bimodal distribution in Fig. 2(b), which may reflect the motion of a mixture of bare and labeled microparticles.

For all other labeling conditions, we observed a negatively skewed non-Gaussian velocity distribution, similar to that of particles in a Newtonian fluid.^23,24

Fitting curves to the measured velocity distributions, based on negatively skewed Gaussian distributions, are shown in Figs. 2(a)–2(d).^25,26 For Fig. 2(b) (100 DNA/mp), the curve was estimated by calculating the sum of two negatively skewed Gaussian distributions. The peak on the right of this image was determined to be 73.2 µm/s, whereas that of bare microparticles was 55.0 µm/s. Hence, by analyzing this distribution, the modal velocity of a microparticle coated with 100 DNA molecules could be distinguished from that of a bare microparticle. The modes of different distributions are plotted in Fig. 2(f).

This study investigated the twDEP for 100–10 000 DNA/mp concentrations. Preliminary experiments demonstrated that the velocity of microparticles in ∼10 DNA/mp could not be distinguished from that of bare microparticles (data not shown). This inability to distinguish velocities could be due to DNA labeling efficiency at lower DNA concentrations. This can be overcome using a microchannel, or through compartmentalization in small droplets, to increase the effective DNA concentration.^27–30 In a separate study, we observed that pp-DEP occurred when the concentration of DNA exceeded 100 000 DNA/mp.³¹ 

There are still some limitations to address. Prior to automatic particle analysis, unfocused particles, suspended at focal planes differing from the one analyzed, were removed manually. In future experiments, this process will be automated in the particle tracking analysis software. For the labeling reaction, amounts of microparticle and DNA used for a concentration of 100 DNA/mp were 6 × 10⁵ and 6 × 10⁷, respectively. Thus, as a DNA detection method, the twDEP measurement cannot be considered to be particularly sensitive at this ratio. Recent progress in analysis of microparticles in the microchannel could improve the handling of smaller numbers of microparticles to increase the technique’s sensitivity.^32,33

The twDEP of DNA labeled microparticles can be used as a DNA detection method with DNA amplification methods such as PCR because this method requires the attachment of DNA to microparticles. During the amplification, the modification for the attachment onto the amplicon can be carried out. Once the amplicon-labeled microparticles are obtained, the amount of the amplicon can be measured easily within a few minutes by the method of this study. An agarose gel electrophoresis, which is a standard method to identify the amplicon, takes several hours. Other fluorescent-based amplicon detection techniques require expensive fluorescent dye and sophisticated optical setting. Electrical equipment and observation system used in this study were simple and easy to prepare.

As we mentioned above, our previous technique using the amplicon-labeled microparticle based on Re[CM]. Although it detects the attached DNA on the microparticle quickly, it requires 10⁵ DNA/mp to convert microparticle DEP from negative to positive. In the case of twDEP based method, 100 DNA/mp can be distinguished because twDEP is based on Im[CM] of the microparticle.

To the best of our knowledge, this is the first study in which the relationship between the twDEP of microparticle and its surface labeling has been investigated. Attachment of DNA to a microparticle surface increased the magnitude of the twDEP force, enabling us to quantitatively distinguish the twDEP velocity of microparticles coated with 100–10 000 DNA molecules. Based on the theoretical explanation of this force, we suggest that this change is due to the increased K_S of the microparticle, caused by the electric charge on the DNA. This increased K_S is reflected as an increase in |Im[CM]|. The method can, thus, be utilized for characterizing other electrically charged biomolecules, such as proteins, phospholipid membranes, and sugar chains. Moreover, when combined with microchannel technologies, such molecules can be detected sensitively with our technique. Hence, the twDEP measurement shows promise as a novel DNA detection method.

See the supplementary material for the illustration of twDEP (S1), images of the experimental setup (S2), and movie of twDEP.

This work was partly supported by JSPS KAKENHI (Grant No. JP17H03277) and a grant from the Hattori Hokokai Foundation.