Electrotransfection of Mammalian Cells Using Microchannel-Type

Seoul, 151-742, Korea, Digital Bio Technology Co., Institute of Advanced Machinery 1304, San 56-1, Shinlim-dong,. Kwanak-gu, Seoul, 151-742, Korea, Sc...
1 downloads 7 Views 834KB Size
Download PDF
Anal. Chem. 2004, 76, 7045-7052

Electrotransfection of Mammalian Cells Using Microchannel-Type Electroporation Chip Young Shik Shin,† Keunchang Cho,‡ Jung Kyung Kim,‡ Sun Hee Lim,‡ Chan Hee Park,‡ Kyu Baek Lee,§ Yongdoo Park,§ Chanil Chung,‡ Dong-Chul Han,† and Jun Keun Chang*,‡,⊥

School of Mechanical and Aerospace Engineering, Seoul National University, 312-203, San 56-1, Shinlim-dong, Kwanak-gu, Seoul, 151-742, Korea, Digital Bio Technology Co., Institute of Advanced Machinery 1304, San 56-1, Shinlim-dong, Kwanak-gu, Seoul, 151-742, Korea, School of Electrical Engineering and Computer Science, Seoul National University, Seoul, 151-742, Korea, and Department of Biomedical Engineering, Korea University, Anam 5-ga, Sungbuk-gu, Seoul, 136-701, Korea

Transfection of DNA molecules into mammalian cells with electric pulsations, which is so-called electroporation, is a powerful and widely used method that can be directly applied to gene therapy. However, very little is known about the basic mechanisms of DNA transfer and cell response to the electric pulse. We developed a microelectroporation chip with poly(dimethylsiloxane) (PDMS) to investigate the mechanism of electroporation as a first step of DNA transfer and to introduce the benefits of miniaturization into the genetic manipulation. The microelectroporation chip has a microchannel with a height of 20 µm and a length of 2 cm. Owing to the transparency of PDMS, we could in situ observe the uptake process of propidium iodide (PI) into SK-OV-3 cells, which shows promise in visualization of gene delivery in living cells. We also noticed the geometric effect on the degree of electroporation in microchannels with diverse channel width. This experimental result shows that the geometry can be another parameter to be considered for the electroporation when it is performed in microchannels with an exponential decaying pulse generator. Cell culturing is possible within the microelectroporation chip, and we also successfully transfected SK-OV-3 cells with enhanced green fluorescent protein genes, which demonstrates the feasibility of the microelectroporation chip in genetic manipulation. Electroporation is a widely used method to introduce xenomolecules into living cells by applying electric pulses. Under the high electric field, cell membranes are transiently rendered porous, and they become permeable to otherwise impermeable foreign materials. Electropermeabilization depends on several factors: pulse amplitude, pulse duration, the number of pulses, and other experimental conditions. Many research groups have * To whom correspondence should be addressed. Phone: +82-2-875-2205. Fax: +82-2-885-2267. E-mail: [email protected], [email protected]. † School of Mechanical and Aerospace Engineering, Seoul National University. ‡ Digital Bio Technology Co. ⊥ School of Electrical Engineering and Computer Science, Seoul National University. § Department of Biomedical Engineering, Korea University. 10.1021/ac0496291 CCC: $27.50 Published on Web 11/10/2004

© 2004 American Chemical Society

performed various theoretical as well as experimental studies for those parameters to understand the mechanism of electroporation and to enhance the efficiency of the transfection. The electric field intensity was reported as the decisive parameter inducing membrane permeabilization and controlling the extent of the cell surface where the transfer can take place.1 Zimmermann et al. considered the electrode formation forces for high transfection yields and demonstrated the parameters involved in the generation of electrode formation forces.2 Canatella et al. collected a comprehensive set of data for more than 200 different electroporation conditions by using flow cytometry.3 Luo and Saltzman reported that increased DNA concentration at the cell surface enhanced transfection efficiency.4 Numerical models that provide relationships between theory and experiments were also provided by several groups.5-9 The visualization of electroporation is an important issue, considering the lack of theoretical explanation for the mechanism of electroporation. Golzio et al. investigated the DNA transfer process at the single-cell level by using fluorescence microscopy digitized imaging.10 Deng et al. assessed the effect of intense submicrosecond electric pulses on cells by means of temporally resolved fluorescence and light microscopy.11 Gabriel and Teissie´ directly observed exchanges of calcium ions between the cytosol of a single cell and the extracellular medium by using an ultrarapid intensifying video system.12 A cuvette equipped with two parallel electrode plates is commonly used to contain the mixture of cell suspension and (1) Rols, M.-P.; Teissie´, J. Biophys. J. 1998, 75, 1415-1423. (2) Zimmermann, U.; Friedrich, U.; Mussauer, H.; Gessner, P.; Ha¨mel, K.; Sukhorukov, V. IEEE T. Plasma Sci. 2000, 28 (1), 72-82. (3) Canatella, P. J.; Karr, J. F.; Petros, J. A.; Prausnitz, M. R. Biophys. J. 2001, 80, 755-764. (4) Luo, D.; Saltzman, W. M. Nat. Biotechnol. 2000, 18, 893-895. (5) DeBruin, K. A.; Krassowska, W. Biophys. J. 1999, 77, 1213-1224. (6) DeBruin, K. A.; Krassowska, W. Biophys. J. 1999, 77, 1225-1233. (7) Kotnik, T.; Miklavcˇicˇ, D. Biophys. J. 2000, 79, 670-679. (8) Wachner, D.; Simeonova, M.; Gimsa, J. Bioelectrochemistry 2002, 56, 211213. (9) Gimsa, J. Bioelectrochemistry 2001, 54, 23-31. (10) Golzio, M.; Teissie´, J.; Rols, M.-P. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (3), 1292-1297. (11) Deng, J.; Schoenbach, K. H.; Buescher, E. S.; Hair, P. S.; Fox, P. M.; Beebe, S. J. Biophys. J. 2003, 84, 2709-2714. (12) Gabriel, B.; Teissie´, J. Biophys. J. 1999, 76, 2158-2165.

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 7045

genes. By applying a high electric field across those two electrodes, it is possible to deliver genes into cells. Aluminum is the commonly used material for disposable cuvettes. In this case, however, the release of Al3+ ions from the aluminum electrode, which has adverse effects on cells, has been observed by several research groups. Field pulses solubilize metals from the electrode plates, and the produced metal ions can thus be introduced.13 Friedrich et al. explained that the enhanced solubilization of Al3+ ions from the electrodes was presumably due to electrolysis and correspondingly to changes in the local pH at the electrodes.14 In the case of using aluminum electrodes, a different field strength is expected because of the significant voltage drop across the oxide layer on the electrode surface. Therefore, although they are more expensive, platinum or gold electrodes are more advantageous.13 Several research groups induced microelectromechanical systems (MEMS) fabrication techniques to miniaturize the electroporation system. Through the miniaturization, it is possible to transfect DNA into cells with voltage as low as 10 V to achieve the necessary electric field. A simple power supply is thus required. Furthermore, the heat generated from voltage pulses can be dissipated more quickly due to a large surface-to-volume ratio.15-19 Because cells can be thermally damaged and the electroporation, applying electric pulses, itself probably stresses cells, the additional stress owing to heating may be significant.20-22 Lin and Huang demonstrated in vitro electroporation using a microchip that integrated cell culture cavity and thin-film electrodes.15 Lin et al. developed a flow-type electroporation microchip to overcome the limitation of the cell numbers and the heating effect.16 Huang and Rubinsky reported a novel electroporation chip that incorporates a live biological cell in the electrical circuit.17-19 By passing electrical currents through the cell, the electroporation status of cells can be measured electrically. Recently, single-cell electroporation using microelectrodes, such as solid carbon fiber microelectrodes and electrolyte-filled capillaries, has been investigated with keen interest.23-25 It fosters opportunities to manipulate the biochemical content of single cells as well as to transfer selective membrane-impermeable solutes into single cells. In this work, we designed a microelectroporation chip that has a microchannel and investigated the characteristics of electroporation induced by that specific design. The microchip is fabricated with poly(dimethylsiloxane) (PDMS) by using MEMS techniques. (13) Zimmermann, U.; Davey, G. A. Electroporation of Cells; CRC Press: Boca Raton, FL, 1996. (14) Friedrich, U.; Stachowicz, N.; Simm, A.; Fuhr, G.; Lucas, K.; Zimmermann, U. Bioelectrochem. Bioenerg. 1998, 47, 103-111. (15) Lin, Y.-C.; Huang, M.-Y. J. Micromech. Microeng. 2001, 11, 542-547. (16) Lin, Y.-C.; Jen, C.-M.; Huang, M.-Y.; Wu, C.-Y.; Lin, X.-Z. Sens. Actuators, B 2001, 79, 137-143. (17) Huang, Y.; Rubinsky, B. Biomed. Microdev. 1999, 2 (2), 145-150. (18) Huang, Y.; Rubinsky, B. Sens. Actuators, A 2001, 89, 242-249. (19) Huang, Y.; Rubinsky, B. Sens. Actuators, A 2003, 104, 205-212. (20) Pliquett, U. F.; Martin, G. T.; Weaver, J. C. Bioelectrochemistry 2002, 57, 65-72. (21) Pliquett, U.; Gift, E. A.; Weaver, J. C. Bioelectrochem. Bioenerg. 1996, 39, 39-53. (22) Gervais, P.; de Maranˇon, I. M. Biochim. Biophys. Acta 1995, 1235, 52-56. (23) Olofsson, J.; Nolkrantz, K.; Ryttse´n, F.; Lambie, B. A.; Weber, S. G.; Orwar, O. Curr. Opin. Biotechnol. 2003, 14 (1), 29-34. (24) Lundqvist, J. A.; Sahlin, F.; A° berg, M. A. I.; Stro ¨mberg, A.; Eriksson, P. S.; Orwar, O. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 10356-10360. (25) Karlsson, M.; Nolkrantz, K.; Davidson, M. J.; Stro ¨mberg, A.; Ryttse´n, F.; A° kerman, B.; Orwar, O. Anal.Chem. 2000, 72, 5857-5862.

7046

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

PDMS has great potential as a fabrication material. It is easy to fabricate, inexpensive, transparent, and biocompatible. The transparency of PDMS enables us to observe the process of transfection in real time and in situ. Furthermore, PDMS chips can be easily integrated with other systems, according to many published works using PDMS, such as PCR,26 CE,27 mixer,28 and filter.29 To explore the functional advantages of the novel microelectroporation chip, we introduced propidium iodide (PI) into SK-OV-3 cells. We also changed the width of microchannels and investigated the process of PI uptake under the same electric field conditions. The feasibility of the microelectroporation chip in genetic manipulation was evaluated by experiments using on-chip cell culturing and transfection with enhanced green fluorescent protein (EGFP) genes. MATERIALS AND METHODS Fabrication of the Microelectroporation Chip. The microelectroporation chip has a simple design that contains a microchannel and two reservoirs for the inlet and outlet. The channel height is 20 µm, and the length is 2 cm. We varied the channel width from 100 to 500 µm. Figure 1a shows the fabricated microelectroporation chip. The microelectroporation chip was fabricated using the replica molding method.30 The microchannel patterns were fabricated by photolithography using chrome photo mask. The fabrication process of the microelectroporation chip is described in our previous work.26 Briefly, negative photoresist (SU-8, MicroChem, MA) was spin-coated onto a silicon wafer to create a mold master of 20-µm-thick structure. After a soft bake, the pattern of the mask was transferred to the SU-8-coated silicon wafer by mask aligner (MA-6, Karl Suss GmbH, Germany). Postexposure bake, developing, and hard baking of the exposed SU-8 patterns were followed by pouring the mixture of PDMS and curing agent onto the pattern (Sylgard 184, Dow Corning Co., U.S.A.). The curing condition was 90 °C for 30 min. The PDMS replica was bonded with a glass substrate to form a microchannel by 25 W oxygen plasma treatments. Cell Preparation and Culture. SK-OV-3 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS, Sigma), penicillin (100 units/mL), streptomycin (100 µg/mL), and Lglutamine (4 mM) at 37 °C in a humidified 5% CO2 incubator. Cells were dissociated from the 25 cm2 tissue culture flask by using trypsin-EDTA. The final cell concentration was adjusted to 1 × 107 cells/mL. The ability of cells to grow after the pulsation was used as direct evidence of their viability. Before applying a voltage pulse, PI (1.0 mg/mL) was added to the cell media at a ratio of 1:20 (v/v). We used a Plasmid Isolation Kit (Promega, U.S.A.) to extract and purify the pEGFPN1 plasmids carrying the green fluorescent protein (GFP) from (26) Shin, Y. S.; Cho, K.; Lim, S. H.; Chung, S.; Park, S.-J.; Chung, C.; Han, D.C.; Chang, J. K. J. Micromech. Microeng. 2003, 13, 768-774. (27) Heo, Y. S.; Chung, S.; Cho, K.; Chung, C.; Han, D.-C.; Chang, J. K. J. Chromatogr., A 2003, 1013, 111-122. (28) Park, S.-J.; Kim, J. K.; Park, J.; Chung, S.; Chung, C.; Chang, J. K. J. Micromech. Microeng. 2004, 14, 6-14. (29) Park, J.; Chung, S.; Chung, C.; Han, D.-C.; Chang, J. K. Proc. µTAS 2002 2002, 204-206. (30) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974-4984.

Figure 1. (a) Microelectroporation chip fabricated with MEMS techniques; (b) picture of system setup for the electroporation: the electrode holder and a microchip; (c) schematic drawing of components used in the experimental setup.

competent Escherichia coli (DH5-R). The extracted plasmid DNA was confirmed by electrophoresis on agarose gel. The concentration was determined by measuring the absorbance at 260 nm in a spectrophotometer. Before pulsing, the plasmid pEGFP-N1 was added to the pulse media at a concentration of 0.1 µg/µL. The expression of the reporter genes was used to evaluate the successful transfection. To check the expression of EGFP, the cells exposed to the voltage pulse were cultured. After the pulsation, the microchip was immersed in DMEM and placed in an incubator for 24 h. Cells in microchips were observed using fluorescent microscopy to evaluate the expression level of EGFP. Experimental Setup and Procedure. The experimental setup for the electroporation was made up of a homemade pulse generator, Pt electrodes, and an electrode holder. Because the electrode holder was mounted on a microscope, we could observe the process of electropermeabilization while applying the voltage pulse. The pulse generator delivers pulses of exponentially decaying voltages and consists of a high-voltage power supply (model 3125, Canberra, U.S.A.), a charging capacitor, and fast high-voltage relays. The pulse generator was connected to a computer via an analogue output board (COMI-CP301, COMIZOA, Korea) and controlled with LabVIEW version 6.1 (National Instruments, U.S.A.). The operation of the pulse generator was

confirmed by oscilloscope (54825A Infiniium, Agilent, U.S.A.). To verify the performance of the microelectroporation chip, a commercial electroporation system (ECM 830, BTX) and a 2-mm gap cuvette with parallel-plate aluminum electrodes (BTX) were used as a reference. To analyze the performance of two systems in the same electric field, we applied 200 V for the cuvettes, and 2 kV for the chips, respectively. The resulting electric field was 1 kV/cm, and the pulse duration was 10 ms for both systems. We performed the experiments with PI for five cases by changing the channel width from 100 to 500 µm by 100-µm increases. For the GFP transfection and expression, experiments were performed under various pulse conditions: from 0.75 to 0.25 kV/cm for 10 ms. Figure 1b shows the electrode holder and a microchip on the microscope, and Figure 1c shows the schematic features of components used in the experimental setup. To observe the PI uptake, we used an inverted fluorescence microscope (IX70, Olympus, U.S.A.) equipped with a 100-W mercury lamp and 20×/0.4 NA objective. The excitation light was optically filtered by a 530 ( 20-nm band-pass filter, and the induced fluorescence from the electroporated cells was filtered with a 590nm long-pass filter. 640 × 480 pixel images were acquired by a 12-bit cooled CCD camera (Cooke, MI) at 15 frames/s. The exposure time was 10 ms for all cases. Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

7047

Figure 2. PI inflow process within a 100-µm-wide microchannel before the pulsation (a) and after the pulsations of (b) 1 ms, (c) 10 ms, (d) 5 s, (e) 10 s, (f) 15 s, (g) 25 s, and (h) 35 s.

To observe the fluorescence for the cell viability and transfection of GFP, the excitation light was filtered by a 475 ( 15-nm band-pass filter, and the induced fluorescence was filtered with a 520-nm long-pass filter. 640 × 480 pixel images were acquired by a color 3IT CCD camera (AW-E300, Panasonic, U.S.A.). For all electroporation experiments, DMEM was used as a pulsing buffer. To exclude the movement of the electroporation media, we waited 15 min after sample injection and applied voltage pulses. RESULTS AND DISCUSSION Electrolysis Effect While Applying Electric Pulse. During the conventional electroporation within a cuvette, a two-phase layer of liquid and gas was created because bubbles were generated electrochemically at the surface of the electrodes. The bubbles were generated very quickly and caused complex liquid motion. This bubble movement, combined with electrophoresis during pulse, induced an inhomogeneous condition of the bulk media as well as the cells. Our hypothesis for the mechanism of the bubble formation is that the oxide layer on the aluminum electrodes acts as a high-resistance layer. Aluminum is the material that forms the oxide layer (Al2O3) very easily by intervening oxidizable electrolyte. Pliquett et al. showed a hypothesis similar to ours and explained it in great detail.21 7048

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

In the microelectroporation chip, we could find neither bubble generation nor complex movement of the media (data not shown). We may be able to explain this phenomenon in terms of the chemical stability of the electrode metals. Pt that we use for the electrodes in microelectroporation experiments is a very stable material; therefore, it does not form an oxide layer easily, and even if that oxide layer was formed, it could be removed with ease. Because the motion of the bulk media within a cuvette is very intense, the stable condition of the media within the microelectroporation chip can be advantageous, especially for the case of investigating the mechanisms of electroporation by visualization. Intercalation Rate of PI and Detection of theElectropermeabilization Process. A localized inflow of PI into cells in milliseconds was observed within the microchannels after the pulsation. PI is a commonly used fluorescent marker that intercalates into nucleic acid molecules as an indicator for surface membrane integrity in living cells. When a cell membrane becomes permeable, PI enters the cell and emits red fluorescence by binding to nucleic acids. Quantitative study is possible because the intensity of the red fluorescence varies with the amount of PI bound to the nucleic acids.11 The permeabilization of PI was

Figure 3. Fluorescence microscopic images of cells for two different channels: (a) 100 and (b) 500 µm in width. The images were taken 30 s after the pulsation.

detected from almost every cell in the microchannel when an electric field in the same range as the conventional system was applied. Figure 2 shows the PI inflow process within a 100-µmwide microchannel. Right after the pulsation, PI inflow was present only on the side facing the anode. This phenomenon shows that only the cell membrane on the left side was altered. The same observation was reported by Golzio et al.10 As time goes on, the fluorescence spreads into the whole cell interior (c-d), and after 10 s, the nucleus begins to emit fluorescence (e-h). This observation directly reflects the PI characteristics, binding to nucleic acids. The observation of the real-time uptake process of a single living cell provides important information, including fundamental cellular processes. Tsuji et al. observed the hybridization of the pair of oligoDNAs to c-fos mRNA in the cytoplasm by detecting the fluorescence resonance energy transfer (FRET).31 Tsukamoto et al. investigated the dynamic morphological changes in chromatin structure through direct visualization of gene expression in living cells.32 For these kinds of studies, the electroporation and real-time monitoring of cellular uptake is a basic experimental step; therefore, our electroporation chip is very useful due to its in situ real-time monitoring capability. The Effect of Channel Width on the PI Uptake Efficiency in the Microelectroporation Chip. For microchips, the intensity of fluorescence by dye uptake was variable according to the channel width. As the channel width increases, the intensity of the gray scale unit for cell area decreases, even though the same voltage pulses were applied. Figure 3 shows the representative microscopic images of cells for two different channels: 100 and 500 µm in width, respectively. The electric field was 1 kV/cm, and pulse duration was 10 ms. Pictures were taken at 30 s after the pulsation. It can be clearly noticed that PI uptake in a narrow microchannel (100 µm) is higher than that in a wide microchannel (500 µm). To compare the PI uptake for the five microchannels with different widths, images were acquired at 15 frames/s and stored while performing the experiments. The image processing was (31) Tsuji, A.; Koshimoto, H.; Sato, Y.; Hirano, M.; Sei-lida, Y.; Kondo, S.; Ishibashi, K. Biophys. J. 2000, 78, 3260-3274. (32) Tsukamoto, T.; Hashiguchi, N.; Janicki, S. M.; Tumbar, T.; Belmont, A. S.; Spector, D. L. Nat. Cell Biol. 2000, 2, 871-878.

Figure 4. Average intensity of gray scale unit for cell areas relative to the background for the five different cases of microchannels. The relative intensity was calculated by image processing every 50 frames. Images were acquired at 15 frames/s, and exposure time was 10 ms.

performed for images every 50 frames. The average intensity of gray scale unit for the background was subtracted from that for the cell area by using graphic software (Paint Shop Pro 7.0, Jasc Software, U.S.A.) and MATLAB program (MathWorks, Inc., U.S.A.). The comparative data for the PI intensity in cells are shown in Figure 4. The PI uptake rate depending on the channel width is obvious. In a conventional cuvette-based system, the geometric parameters except for the electrode gap are not considered serious. Thus, this unique phenomenon in a microchannel is noteworthy. Sugar and Schmukler developed the lowvoltage electroporator, which contains filter pores to increase local current density.33 Owing to the large increase in current density (33) Sugar, I. P.; Lindesay, J.; Schmukler, R. E. J. Phys. Chem. B 2003, 107, 3862-3870.

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

7049

Figure 5. Cell images before (left) and after (right) the pulsation for two different channel widths: (a) 150 and (b) 500 µm.

in the filter pores, cells could be transfected by applying only 25 V. Their study was, however, based on local geometric changes for the reduction of current shunt pathways. Therefore, it is different from our observations in this microchip. The morphological changes of cells after the electric pulsation were also analyzed with bright field images. The bright field analysis was performed under two conditions: 150 and 500 µm in width. The pulse condition was the same as the other experiments (1 kV/cm electric field for 10 ms). Images after being exposed to a voltage pulse were acquired at 25 s after the pulsation. When exposed to a voltage pulse, cells were swelled instantaneously. Figure 5 shows the result of the image analysis for two cases: (a) 150 and (b) 500 µm. We measured the increase of the cell diameter with AutoCAD 2002 (Autodesk, Inc., U.S.A.) and calculated the percentage of increase in diameter relative to that of before the pulsation. The diameter of cells in a 150-µmwide channel was increased about 23%; that in a 500-µm-wide channel was increased about 10%. This different extent of swelling with the change of channel width can be considered as evidence for the difference in the degree of electroporation. This different degree of electroporation within microchannels with various widths can be explained by the shape of the voltage pulses. For this study, we used the exponentially decaying pulse generator, which is commonly used in electroporation. The voltage pulses exponentially decay, and the discharge time constant is dependent upon the charging capacitance and the media resistance. Even though the charging capacitance is constant, the resistance of the media between the electrodes is changed by the channel width in our experiments. The effects of electrophoretic drive as well as membrane permeabilization are reported by Bureau et al.34 Because the electrophoretic drive depends on the electric field intensity, the different exponentially decaying electric field would induce differences in the transport of charged molecules and the degree of electroporation, as shown in Figure 3. We also performed the same experiments with a homemade square-wave pulse generator, and noticeable differences were not found in the intensity of fluorescence according to the channel width. From the results, we could elucidate that the channel width is closely related to the cellular uptake rates in a microchannel when using an exponential decaying pulse generator. Cell Culture in a PDMS Microchip. PDMS is known as a promising material for the on-chip cell culture system due to its biocompatibility and permeability.35-38 Several studies have been (34) Bureau, M. F.; Gehl, J.; Deleuze, V.; Mir, L. M.; Scherman, D. Biochim. Biophys. Acta 2000, 1474, 353-359.

7050

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

reported about a PDMS-based cell culture system: endothelial cells,35 liver cells,36 bacterial cells,37 and mast cells.38 In our microelectroporation chip, cell culture function is required for EGFP transfection experiments because it usually takes more than 24 h for EGFP to be expressed in a cell after the electroporation. To check the feasibility of the microchannel as a cell culture device, we injected cells into microchannels and immersed the whole PDMS chip into a cell medium, DMEM. They were kept in an incubator for 7 days, and Figure 6 shows the result of culturing. The surface of the PDMS microchip was treated only with O2 plasma for bonding of PDMS and glass. After 7 days, cells at the end of the microchannel and in the wells proliferated and spread well on the bottom. However, the condition of cells at the center of a 50-µm-wide microchannel was not that good (Figure 6a). It seems that this phenomenon is highly related to the channel width. The small physical space in a narrow channel contains a low level of oxygen and CO2, and the perfusion in this microchannel depends solely on diffusion. The diffusion process is very slow, and we could not expect fast circulation. As a result, the concentrations of metabolites would be higher within narrow microchannels, and that might be the adverse influence upon culturing cells. In wider microchannels, cells were observed to attach, proliferate, and migrate successfully (Figure 6: (b) 150, (c) 200, and (d) 250 µm). From these results, we could notice the possibility of cell culturing in our microdevices, and this functionality can be expected to offer advantages for the simultaneous study of multiple cells over long periods of time combined with the real-time visualization function. Furthermore, quantum dots (QDs)ssemiconductor nanocrystalssare watched with the keen interest of researchers due to their excellent properties as fluorescent probes for cellular imaging.39 Jaiswal et al. demonstrated the use of QDs in long-term multicolor imaging of live cells,40 and our group has studied the conjugation of streptavidincoated QDs for the real-time imaging of gene transfer into live cells.41 Considering this trend, our microdevice for electroporation (35) Borenstein, J. T.; Terai, H.; King, K. R.; Weinberg, E. J.; Kaazempur-Mofrad, M. R.; Vacanti, J. P. Biomed. Microdev. 2002, 4 (3), 167-175. (36) Leclerc, E.; Sakai, Y.; Fujii, T. Biomed. Microdev. 2003, 5 (2), 109-114. (37) Chang, W.-J.; Akin, D.; Sedlak, M.; Ladisch, M. R.; Bashir, R. Biomed. Microdev. 2003, 5 (4), 281-290. (38) Matsubara, Y.; Murakami, Y.; Kobayashi, M.; Morita, Y.; Tamiya, E. Biosens. Bioelectron. 2004, 19 (7), 741-747. (39) Parak, W. J.; Gerion, D.; Pellegrino, T.; Zanchet, D.; Micheel, C.; Williams, S. C.; Boudreau, R.; Gros, M. A. L.; Larabell, C. A.; Alivisatos, A. P. Nanotechnology 2003, 14, R15-R27. (40) Jaiswal, J. K.; Mattoussi, H.; Mauro, J. M.; Simon, S. M. Nat. Biotechnol. 2003, 21, 47-51.

Figure 7. Images of transfected SK-OV-3 cells with EGFP in the 250-µm-wide microchannel: (a) bright-field image of cells electroporated at 0.75 kV/cm, 10 ms; (b) overlay of bright-field and fluorescence image of cells electroporated at 0.4 kV/cm, 10 ms; (c) fluorescence image of cells in (b).

Figure 6. Images of cultured cells in microchannels for 7 days: (a) 50, (b) 150, (c) 200, and (d) 250 µm.

is expected to be a powerful tool for simultaneous tracking of multiple proteins in live cells for long periods, especially by using QDs. Studies relevant to the microelectroporation chip and QDs are in progress. Transfection of EGFP Genes into SK-OV-3 Cells. Even though PI uptake was successfully visualized in the microelectroporation chip, it is not clear that cells with electropulsation can survive and express the gene or protein. Therefore, we transfected pEGFP plasmids to the cells and monitored the expression levels. The GFP from the jellyfish Aequorea victoria is widely used in biochemistry and cell biology owing to its highly visible and efficiently emitting internal fluorophore. The GFP has been used as a marker of gene expression and protein targeting in intact cells and organisms.42 At first, we applied a 1.5-kV voltage pulse, which induced the electric field, as much as 0.75 kV/cm for 10 ms. This condition is appropriate for SK-OV-3 cells to be transfected using a BTX (41) Kim, J. K.; Lim, S. H.; Lee, Y.; Shin, Y. S.; Chung, C.; Yoo, J. Y.; Chang, J. K. Proc. NanoTech 2004 2004, 3, 379-382. (42) Tsien, R. Y. Annu. Rev. Biochem. 1998, 67, 509-544.

commercial electroporator. This electric field condition, however, was too severe for cells in a microchannel. Cells were inspected after 24 h, and the result is shown in Figure 7a. Most of the cells remained round and did not spread on the surface. Moreover, no fluorescence was detected, reflecting that transfection was not successful. When an electric field identical to the conventional machine is applied to the microelectroporation chip, the electric stress is too high for cells to survive and be transfected. We changed the electric field from 0.25 to 0.75 kV/cm. As a result, we observed that cells under 0.4-0.5 kV/cm were successfully transfected with EGFP and emitted green fluorescence. The best condition was 0.4 kV/cm; Figure 7b and c shows the results. From these results, we can recognize that the effect of electropulsation in a microchannel is far stronger than that in a cuvette, which corresponds to the results of PI experiments. High efficiency as well as effectiveness is expected when optimal conditions for the electroporation in a microchannel are found. CONCLUSIONS We developed a microdevice to perform electroporation in a microchannel. Owing to the advantageous material PDMS, we could directly visualize the whole intercalation process in real time. In a microchannel, bubble generation and the complex motion of the cell media as well as the cells were not observed. The Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

7051

homogeneous internal milieu within the microchannel may result in the high intercalation rate of PI into SK-OV-3 cells, as compared to the commercial electroporator under the same electric field condition. The geometric effect on the degree of electroporation was identified through the experiments within microchannels with various channel widths. These results reflect that the channel geometry should be another parameter to be considered for electroporation when it is performed in a microchannel with an exponentially decaying pulse generator. Cell culturing in the microelectroporation chip was examined, and this functionality is expected to be applied for the simultaneous tracking of multiple proteins in live cells for long periods. The successful transfection of SK-OV-3 cells with EGFP demonstrated the feasibility of the microelectroporation chip in genetic manipulation. The electric field applied for EGFP transfection into SK-OV-3 cells (0.4 kV/

7052

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

cm) in a microchannel was far less than that in a cuvette-based system (1 kV/cm). With these experimental results, we may use the microelectroporation chip as a good tool that is more suitable for high-throughput gene delivery and more useful for investigating the mechanism of gene transfection. ACKNOWLEDGMENT The authors thank Kee Chun Shin and Kyung Mi Lee at Digital Bio Technology for their help on microfabrication and PMM service. We also appreciate Dr. Seok Chung and Yongku Lee for their valuable suggestions and technical advice. Received for review March 9, 2004. Accepted August 3, 2004. AC0496291