I-Shaped Microchannel Array Chip for Parallel Electrophoretic Analyses

Feb 2, 2007 - microchannel array (IMA) chip by integrating 12 inde- pendent microchannels and 2 electrodes onto a 3 cm ×. 2 cm area in a PDMS-glass ...
0 downloads 0 Views 310KB Size
Download PDF
Anal. Chem. 2007, 79, 2168-2173

I-Shaped Microchannel Array Chip for Parallel Electrophoretic Analyses Akira Inoue,†,‡ Toshiyuki Ito,† Kimiko Makino,‡ Kazuo Hosokawa,*,† and Mizuo Maeda†

Bioengineering Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, and Faculty of Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan

We demonstrate electrophoresis in I-shaped microchannels with a new design and operation principle. Unlike the conventional T- or cross-shaped microchannels, the simple I-shaped design makes it straightforward to integrate parallel microchannels with electrodes onto a microchip. The operation of the I-shaped microchannels has been enabled by the autonomous solution filling technique, which exploits the high gas solubility in poly(dimethylsiloxane) (PDMS). We fabricated an I-shaped microchannel array (IMA) chip by integrating 12 independent microchannels and 2 electrodes onto a 3 cm × 2 cm area in a PDMS-glass hybrid microchip. For autonomous regulation of stable sample plugs in all the microchannels, we discovered that O2 plasma treatment of the PDMS-made reservoirs is effective. On the IMA chip, size-dependent separation of double-stranded (ds) DNA and sequence-specific separation of single-stranded DNA were achieved. Specifically, 10 fragments in a 1001000-bp dsDNA ladder were separated using hydroxyethylcellulose as a sieving matrix within a separation length of 2 mm, and polymerase chain reaction products of the wild-type K-ras gene and its point mutant were separated using a probe DNA-poly(dimethylacrylamide) conjugate on the basis of affinity capillary electrophoresis. The IMA chip presented here opens up a new possibility of large-scale integration of microchannels for highthroughput electrophoretic analyses. Electrophoresis is a fundamental technology in modern life science for separation and analysis of nucleic acids, proteins, and other biomolecules. Over the past decade, microfabricated electrophoretic analysis systems1,2 have been attracting considerable interest because they have many advantages including a small sample volume, high-speed separation, and low-cost mass production. An especially appealing feature is the capability of integration of multiple processes onto a single microchip (lab on a chip). The * To whom correspondence should be addressed. E-mail: [email protected]. Fax: +81-48-462-4658. † RIKEN. ‡ Tokyo University of Science. (1) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z. H.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (2) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 1114-1118.

2168 Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

integration concept can be divided into two categories: serial and parallel. Serial integration refers to incorporation of various treatment steps for a single analyte. So far, the electrophoretic separation has successfully been combined with a variety of treatment steps: amplification,3,4 preconcentration,5-8 and postseparation reaction.9 On the other hand, several research groups have reported parallel integration of electrophoretic separation channels for multiple analytes to realize on-chip capillary array electrophoresis (CAE) mainly for high-throughput genetic analysis.10-17 Most of the previously reported microchannels for electrophoresis have T- or cross-shaped branches on their intermediate positions for sample injection. Such a microchannel requires three or four reservoirs as a single separation unit. On CAE microchips, these millimeter-sized reservoirs occupy considerable areas, thus limiting the maximum number of separation units, although some of the reservoirs can be united into a common (usually waste) reservoir.13 Moreover, the existence of many sample injectors (3) Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086. (4) Burns, M. A.; Johnson, B. N.; Brahmasandra, S. N.; Handique, K.; Webster, J. R.; Krishnan, M.; Sammarco, T. S.; Man, P. M.; Jones, D.; Heldsinger, D.; Mastrangelo, C. H.; Burke, D. T. Science 1998, 282, 484-487. (5) Broyles, B. S.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2003, 75, 27612767. (6) Wang, Y. C.; Stevens, A. L.; Han, J. Y. Anal. Chem. 2005, 77, 4293-4299. (7) Kim, S. M.; Burns, M. A.; Hasselbrink, E. F. Anal. Chem. 2006, 78, 47794785. (8) Shaikh, F. A.; Ugaz, V. M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 48254830. (9) Liu, B.-F.; Ozaki, M.; Hisamoto, H.; Luo, Q.; Utsumi, Y.; Hattori, T.; Terabe, S. Anal. Chem. 2005, 77, 573-578. (10) Simpson, P. C.; Roach, D.; Woolley, A. T.; Thorsen, T.; Johnston, R.; Sensabaugh, G. F.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 2256-2261. (11) Liu, S.; Ren, H.; Gao, Q.; Roach, D. J.; Loder, R. T., Jr.; Armstrong, T. M.; Mao, Q.; Blaga, I.; Barker, D. L.; Jovanovich, S. B. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 5369-5374. (12) Cheng, S.; Skinner, C. D.; Taylor, J.; Attiya, S.; Lee, W.; Picelli, G. Harrison, D. J. Anal. Chem. 2001, 73, 1472-1479. (13) Zhang, C. X.; Manz, A. Anal. Chem. 2001, 73, 2656-2662. (14) Emrich, C. A.; Tian, H.; Medintz, I. L.; Mathies, R. A. Anal. Chem. 2002, 74, 5076-5083. (15) Paegel, B. M.; Emrich, C. A.; Wedemayer, G. J.; Scherer, J. R.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 574-579. (16) Dang, F.; Tabata, O.; Kurokawa, M.; Ewis, A. A.; Zhang, L.; Yamaoka, Y.; Shinohara, S.; Shinohara, Y.; Ishikawa, M.; Baba, Y. Anal. Chem. 2005, 77, 2140-2146. (17) Aborn, J.; El-Difrawy, S.; Novotny, M.; Gismondi, E.; Lam, R.; Matsudaira, P.; Mckenna, B.; O’Neil, T.; Streechon, P.; Ehrlich, D. Lab Chip 2005, 5, 669-674. 10.1021/ac0616097 CCC: $37.00

© 2007 American Chemical Society Published on Web 02/02/2007

requires a complicated electric wiring system, which is difficult to integrate onto the CAE microchip. An I-shaped microchannel is ideal for reducing the complexity of CAE microchips because it requires only two reservoirs as a single separation unit. I-shaped microchannels have been used for the focusing separation modes such as isoelectric focusing18 and temperature gradient focusing,19 but only a limited species of samples have successfully been demonstrated so far. Hong et al. used an I-shaped microchannel for zone electrophoresis of double-stranded (ds) DNA with a solidified agarose gel as a sieving matrix.20 However, their method includes iterative manual operations to exchange the solution in the sample reservoir for formation of the sample plug. Therefore, this approach seems impractical for CAE microchips requiring many sample plugs. Here, we demonstrate zone electrophoresis in I-shaped microchannels with a new design and operation principle eliminating the need for cumbersome solution exchange. This design allows straightforward integration of parallel microchannels and electrodes onto a single microchip with a minimum number of reservoirs. We fabricated an I-shaped microchannel array (IMA) chip by integrating 12 independent microchannels, 2 electrodes, and 24 reservoirs onto a 3 cm × 2 cm area in a poly(dimethylsiloxane) (PDMS)-glass hybrid microchip. The simple operation of the IMA chip has been realized by an improved version of the autonomous solution filling technique. We previously applied the original version of this technique to a single T-shaped microchannel and succeeded in autonomous regulation of the sample plug.21 For the IMA chip, we have introduced an additional procedure to improve channel-to-channel uniformity of the sample plugs: a brief exposure of the IMA chip to O2 plasma. Usefulness of the IMA chip was validated by size-dependent separation of dsDNA and sequence-specific separation of single-stranded (ss) DNA on the basis of affinity capillary electrophoresis (ACE).21-24 EXPERIMENTAL SECTION Design and Fabrication of the Microchip. The design of the IMA chip is illustrated in Figure 1. It consists of a molded PDMS part and a glass plate equipped with thin-film electrodes. The electrodes were fabricated as follows. A microscope slide glass (S9112, 76 × 52 × 1 mm, Matsunami Glass, Osaka, Japan) was cleaned by solvent rinsing (acetone, 2-propanol, and deionized (DI) water) using an ultrasonic rinsing cleaner for each 5 min. A vacuum evaporator (KS-201UT-RK03, K-science, Tsukuba, Japan) was used to evaporate a 20 nm adhesion layer of Cr onto the glass plate, followed by a 200 nm layer of Au. A spin coater was operated at 3000 rpm for 30 s to cover the glass plate with a layer of positive photoresist (OFPR-800, Tokyo Ohka Kogyo, Yokohama, Japan). After prebaking of the photoresist layer in an oven at 90 °C for 15 (18) Cui, H.; Horiuchi, K.; Dutta, P.; Ivory, C. F. Anal. Chem. 2005, 77, 13031309. (19) Ross, D.; Locascio, L. E. Anal. Chem. 2002, 74, 2556-2564. (20) Hong, J. W.; Hosokawa, K.; Fujii, T.; Seki, M.; Endo, I. Biotechnol. Prog. 2001, 17, 958-962. (21) Ito, T.; Inoue, A.; Sato, K.; Hosokawa, K.; Maeda, M. Anal. Chem. 2005, 77, 4759-4764. (22) Anada, T.; Arisawa, T.; Ozaki, Y.; Takarada, T.; Katayama, Y.; Maeda, M. Electrophoresis 2002, 23, 2267-2273. (23) Ito, T.; Inoue, A.; Sato, K.; Hosokawa, K.; Maeda, M. Chem. Lett. 2003, 32, 688-689. (24) Inoue, A.; Ito, T.; Sato, K.; Hosokawa, K.; Maeda, M. Chem. Lett. 2006, 35, 658-659.

Figure 1. Design of the IMA chip. (a) Top view. The device consists of a 2 mm thick PDMS part with microchannels and a 1 mm thick glass plate with Au thin-film electrodes. (b) Magnified view of the channel layout. The 12 I-shaped microchannels are arranged symmetrically with respect to both the horizontal and vertical axes. Each channel is 80 µm wide, 25 µm deep, and 28 mm long. The reservoirs with a diameter of 2 mm were punched in the PDMS part. The cathode is 80 µm wide and 200 nm thick. The anode is 400 µm wide and 200 nm thick. The distance between the cathode and the anode along the microchannels is 14.4 mm. (c) Magnified view around the cathode. Each channel has a passive stop valve (8 µm wide and 80 µm long). The distance between adjacent channels (center to center) is 160 µm. The distance from the cathode and the passive stop valves is 0.4 mm. (d) Magnified view around the corners of the microchannels.

min, the designed pattern was transferred onto the photoresist layer by exposure to UV for 30 s. The exposed portion of the photoresist layer was removed with the developer. After postbaking in the oven at 120 °C for 15 min, the exposed Au and Cr films were removed by soaking in a Au etchant (8% (w/v) ammonium iodide and 1.2% (w/v) iodine in a 6:4 (v/v) mixture of methanol and DI water) for 1 min and then a Cr etchant (18% (w/v) diammonium cerium(IV) nitrate and 2.8% (w/v) ammonium perchlorate in DI water) for 20 s. The remaining photoresist layer was stripped with O2 plasma in a plasma etcher (RIE-10NR, Samco International, Kyoto, Japan). Molding of PDMS was described elsewhere.25 Briefly, a negative master was fabricated on a silicon wafer (Osaka Special Alloy, Osaka, Japan) with negative photoresist (SU-8 25, MicroChem, Newton, MA). After passivation of the master using the plasma etcher, a prepolymer of PDMS (Sylgard 184, Dow Corning, Midland, MI) was cast onto the master with a frame for holding the prepolymer solution. The cured PDMS (30 × 30 × 2 mm) was peeled from the master, and through holes for reservoirs were punched in the PDMS part using a metal pipe. The PDMS part was reversibly bonded to the electrode-patterned side of the glass plate. Sample Preparation. For size-dependent separation, a 1001000-bp dsDNA ladder containing 10 fragments (Bio-Rad Laboratories, Hercules, CA) was stained with SYBR Gold (Molecular Probes, Eugene, OR). For sequence-specific separation of oligonucleotides, 5′-fluorescein isothiocyanate (FITC)-labeled oligonucleotides were purchased from Espec Oligo Service (Tsukuba, Japan). Sequences are as follows: “W(12)” (FITC-5′-GGA GCT (25) Hosokawa, K.; Fujii, T.; Endo, I. Anal. Chem. 1999, 71, 4781-4785.

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

2169

GGT GGC-3′) and “M(12)” (FITC-5′-GGA GCT AGT GGC-3′). For sequence-specific separation of polymerase chain reaction (PCR) products, two DNA templates (ras mutant set, c-Ki-ras codon 12, Takara Bio, Otsu, Japan) containing the same sequences as W(12) and M(12), respectively, were independently amplified. Each reaction mixture contained 25 µL of 2× Ready Mix (Sigma-Aldrich Japan, Tokyo, Japan), 2.5 µL of 20 µM FITC-labeled forward primer (FITC-5′-GAC TGA ATA TAA ACT TGT GG-3′, Espec Oligo Service, Tsukuba, Japan), 2.5 µL of 20 µM biotinylated reverse primer (biotin-5′-ATC GTC AAG GCA CTC TTG CC-3′, Operon Biotechnologies, Huntsville, AL), 1 µL of 1 ng/µL template, and 19 µL of DI water. After preheating at 95 °C for 4 min, 35 cycles were performed using steps of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s in a thermocycler (TGradient, Biometra, Go¨ttingen, Germany). The solution was kept at 72 °C for 7 min and then cooled. After purification with a spin column (AutoSeq G-50, GE Healthcare Bio-Sciences, Piscataway, NJ), the FITC-labeled strand was isolated from the rest of the PCR product using streptavidincoated magnetic beads (MBs, diameter 1 µm, New England Biolabs, Ipswich, MA) as follows:26,27 A 400 µL sample of 4 mg/ mL MB solution was mixed with 400 µL of binding/washing buffer (10 mM Tris-HCl, 1 mM EDTA, and 2.0 M NaCl, pH 7.5). After magnetic separation, the MBs were resuspended in 200 µL of the binding/washing buffer. Then the PCR product was added, and the mixture was subsequently incubated at room temperature (∼25 °C) for 1 h with occasional mixing. After another magnetic separation, the supernatant was replaced by 50 µL of 1× TrisEDTA (TE) buffer (pH 8.0). To denature the MB-bound duplex, the TE buffer was replaced by 30 µL of 0.25 M NaOH, and the solution was incubated at room temperature for 10 min. After another magnetic separation, the supernatant including the detached FITC-labeled strand was collected and purified with a spin column (Microcon YM-30, Millipore, Bedford, MA). Sieving Polymer Preparation. For size-dependent separation, hydroxyethylcellulose (HEC; Wako Pure Chemical Industries, Osaka, Japan) was dissolved into 1× Tris-borate-EDTA (TBE) buffer (pH 8.0) to obtain a 1.2% (w/v) solution. For sequencespecific separation, probe DNA-poly(N,N-dimethylacrylamide) (PDMA) conjugates21 were synthesized as described below. Oligonucleotides modified with a methacryloyl group at their 5′ ends were purchased from Integrated DNA Technologies (Coralville, IA). Sequences are as follows: “P(6)” (5′-ACC AGC-3′), “P(8)” (5′-C ACC AGC T-3′), “P(10)” (5′-CC ACC AGC TC-3′), and “P(12)” (5′-GCC ACC AGC TCC-3′). A mixture consisting of 20 µL of N,N-dimethylacrylamide (Wako Pure Chemical Industries, Osaka, Japan), 10 µL of 1 mM 5′-methacryloyl-modified oligonucleotide, 2 µL of 100 mM MgCl2, and 148 µL of 25 mM Trisborate buffer (pH 7.4) was prepared, and copolymerization was initiated by adding 10 µL of 10% (w/v) ammonium persulfate and 10 µL of N,N,N′,N′-tetramethylethylenediamine. Then the mixture was placed in an incubator (CTU-N, Taitec, Koshigaya, Japan) at 50 °C for 1.5 h. The resulting probe DNA-PDMA conjugate was used without further purification or dilution. Solution Filling and Electrophoresis. For surface modification of the reservoirs, the assembled IMA chip was treated with (26) Murphy, M. B.; Fuller, S. T.; Richardson, P. M.; Doyle, S. A. Nucleic Acids Res. 2003, 31, e110. (27) Hultin, E.; Kaller, M.;, Ahmadian, A.; Lundeberg, J. Nucleic Acids Res. 2005, 33, e48.

2170

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

O2 plasma in the plasma etcher for 7 s under the following conditions: O2 gas flow of 20 cm3/min, pressure of 20 Pa, and electric power of 150 W. The IMA chip was placed in a vacuum desiccator (model VS, As One, Osaka, Japan) equipped with a mechanical pump (DAL-5D, Ulvac, Chigasaki, Japan). Air dissolved in PDMS was evacuated at 10 kPa for 10-15 min. The microchip was taken from the vacuum desiccator, and then 1 µL of the sample solution (dsDNA ladder, oligonucleotides, or PCR products) and 1 µL of the sieving polymer solution (HEC or probe-PDMA conjugate) were pipetted into the reservoirs. Autonomous filling of the solutions21 was confirmed, and then electrophoresis was carried out at room temperature (∼25 °C) by applying an electric potential of 9.0 V between the thin-film electrodes. Fluorescence images were monitored using an inverted fluorescence microscope (Eclipse TE2000-U, Nikon, Tokyo, Japan) equipped with a 100 W high-pressure mercury lamp, a dichroic mirror block (excitation 465-495 nm, emission 515555 nm), and a cooled CCD camera (C5958, Hamamatsu Photonics, Hamamatsu, Japan). The images were recorded by a digital video recorder (DCR-TRV50, Sony, Tokyo, Japan) and were analyzed using image analysis software (ImageJ 1.28u, National Institutes of Health). RESULTS AND DISCUSSION Design and Fabrication of the Microchip. Figure 1 shows the design of the IMA chip consisting of a PDMS part and a glass plate. They were reversibly bonded to each other. The PDMS part has 12 independent I-shaped microchannels. At the middle of each channel, a pinched region was constructed as a passive stop valve.21 Each channel is equipped with two reservoirs for the sample and sieving polymers. Although we employed the most versatile layout with 24 reservoirs, the number of reservoirs may further be reduced by introducing common reservoirs,13 depending on specific applications. All the microchannels share two Au thin-film electrodes patterned on the glass surface. The anode was set under the waste reservoirs, whereas the cathode was laid across the microchannels. For electrophoresis, the sample plug was defined by the stop valve and the cathode. The manual assembly of the microchip under the microscope yielded the sample plug length of 0.40 ( 0.15 mm. We observed no solution leakage along the cathode; the 200 nm step at the edges of the cathode seemed to be successfully sealed by the contact with PDMS. Solution Filling. Figure 2 shows the operation principle of the I-shaped microchannel. By ordinary procedures, an I-shaped channel is difficult to fill with different solutions because of the entrapped air in the channel. One can overcome this problem by employing the autonomous solution filling technique,21 which is based on the high gas solubility in PDMS.28 Specifically, air dissolved in PDMS is evacuated in a vacuum chamber. After removal of the microchip from the vacuum chamber, the sample solution and the sieving polymer solution are dispensed into the reservoirs using a pipet (Figure 2a). The redissolution of air into the PDMS channel walls causes spontaneous introduction of the solutions (Figure 2b). The solutions are eventually merged together at the passive stop valve (Figure 2c), which stops the (28) Hosokawa, K.; Sato, K.; Ichikawa, N.; Maeda, M. Lab Chip 2004, 4, 181185.

Figure 2. Schematic representation of the operation principle of the I-shaped microchannel. One straight microchannel is drawn for simplicity. (a) The sample and the sieving polymer solutions are dispensed into the reservoirs of the degassed microchip. (b) The solutions are spontaneously introduced into the microchannel by dissolution of air into the PDMS channel walls. (c) The solutions are eventually merged together at the passive stop valve. The sample plug is defined by the cathode and the passive stop valve. (d) Electrophoresis is started by applying an electric field.

first arriving solutionsby an abrupt increase in the Laplace pressure25,29suntil it is merged together with the other solution. When an electric potential is applied between the electrodes, the sample plug defined by the cathode and the stop valve is electrophoresed toward the anode, in the case of a negatively charged analyte (Figure 2d). The IMA chip is intended to be used only once in practical applications, because both exchange of the solutions and repeated injections of the sample are impossible. In reality, we recycled one electrode-patterned glass plate several times after removing the used PDMS part and washing with DI water. For the multiple I-shaped microchannels, an improved solution filling procedure was necessary to obtain uniform sample plugs. Because of the manual pipetting, the 12 microchannels did not simultaneously start and finish the solution filling. Experimentally, the time lag between the first and the last finishing times was about 1 min. Without improvement, we observed that significant bulk flow disturbed the sample plugs that were formed earlier. (Note: the stop valve loses its primary function after the complete filling, although it raises the viscous drag.) This situation is shown in Figure 3a, in which 10% PDMA solutions spiked with food dyes have been used for visualization. These random motions made it impossible to control sample plugs. The bulk flow was probably caused by an imbalance of pressure between the two reservoirs. A similar phenomenon was reported by Kim et al.30 They attributed this imbalance to the variation in Laplace pressure at the reservoirs. We discovered that the random bulk flow can dramatically be reduced by surface treatment of the reservoirs. In the improved procedure, we exposed the whole IMA chip, with the PDMS side up, to O2 plasma for 7 s immediately prior to the (29) Yamada, M.; Seki, M. Anal. Chem. 2004, 76, 895-899. (30) Kim, S.-J.; Lim, Y. T.; Yang, H.; Shin, Y. B.; Kim, K.; Lee, D.-S.; Park, S. H.; Kim, Y. T. Anal. Chem. 2005, 77, 6494-6499.

Figure 3. Autonomous filling of 10% PDMA solutions visualized with food dyes (red, New Coccine; blue, Brilliant Blue FCF). (a) Without O2 plasma treatment, random bulk flow disturbed some of the sample plugs. Two severe displacements are indicated by the red arrows. (b) With O2 plasma treatment, the bulk flow was suppressed, and stable sample plugs were obtained in all the microchannels. (c) Hypothetical mechanism of the random bulk flow. In general, contact angles in the connected reservoirs may differ from each other because of the large contact angle hysteresis of native PDMS. This difference causes an imbalance of the Laplace pressure. (d) The difference in contact angles can be eliminated by surface treatment of the reservoirs.

degassing. As a result, uniform sample plugs were reproducibly obtained (Figure 3b). In general, the Laplace pressure is determined by the surface tension, contact angle, and geometry. The above results indicate that the variation in contact angle was the major source of the pressure imbalance in our microchip. Native PDMS has a large contact angle hysteresis,31 which means that various contact angles may occur in the reservoirs depending on irreproducible initial conditions such as the position of the pipet tip and the speed of dispensing (Figure 3c). The O2 plasma treatment makes oxidized PDMS surfaces which are completely hydrophilic and have a negligible contact angle hysteresis.32 This resulted in a uniform Laplace pressure in all the reservoirs (Figure 3d) and hence eliminated the random bulk flow. The above explanation can further be supported by another set of experiments (see the Supporting Information (SI)), in which PDMS-poly(methyl methacrylate) (PMMA) hybrid microchips were used. We fabricated two types of microchips: Type 1 has reservoirs in a hydrophilized PMMA plate, whereas type 2 has reservoirs in a native PDMS part. As a result, type 1 caused much smaller random bulk flow than type 2. These data strongly support our hypothesis: the surface property of the reservoirs is the main (31) Uilk, J. M.; Mera, A. E.; Fox, R. B.; Wynne, K. J. Macromolecules 2003, 36, 3689-3694. (32) Kim, J.; Chaudhury, M. K.; Owen, M. J. J. Colloid Interface Sci. 2000, 226, 231-236.

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

2171

Figure 4. Size-dependent separation of the 100-1000-bp dsDNA ladder fragments on the IMA chip. (a-d) Consecutive fluorescence images showing the initial stages of the separation. (e) Electropherograms. Conditions: electric field, 6.3 V/cm; effective separation length, 2 mm; buffer, 1× TBE (pH 8.0); polymer matrix, 1.2% HEC; sample concentration, 10 µg/mL; label, SYBR Gold; 4× objective lens; temperature ∼25 °C.

cause for the random bulk flow. At the same time, these data suggest a possibility of further improvement of the IMA chip. Currently, we expose the IMA chip to O2 plasma immediately prior to use because the oxidized PDMS surfaces revert back to their original state over time.32 The need for the oxidation step on the analysis site may be inconvenient in practical applications, although a facile oxidation method for PDMS has recently been reported.33 (Note: the degassing step is unnecessary on the analysis site if a predegassed PDMS microchip has been kept in a portable airtight package.34) The on-site oxidation could be avoided by adopting an intrinsically hydrophilic material for the reservoirs. For example, this could be done by fitting glass cylinders with an appropriate outer diameter into the through holes in the PDMS part. Size-Dependent Separation of dsDNA. We carried out sizedependent separation of dsDNA in the 12 microchannels simultaneously. As the sieving polymer, 1.2% HEC was used. The 1001000-bp dsDNA ladder was stained with SYBR Gold and diluted with an equal volume of the sieving polymer. Without this dilution, significant depletion of fluorescence in the sample plug was observed, because of nonspecific adsorption onto the microchannel walls.35 The nonspecific adsorption was suppressed by the dynamic coating effect of the HEC36 in the sample solution. Figure 4e shows electropherograms obtained from the video image. All peaks were separated within 700 s in all the microchannels. Representative resolutions are listed in Table 1. The IMA chip yielded resolutions comparable to those of microchips with a conventional cross injection method (see the SI) and two other recently published methods. The relative standard devia(33) Haubert, K.; Drier, T.; Beebe, D. Lab Chip 2006, 6, 1548-1549. (34) Hosokawa, K.; Omata, M.; Sato, K.; Maeda, M. Lab Chip 2006, 6, 236241. (35) Fosser, K. A.; Nuzzo, R. G. Anal. Chem. 2003, 75, 5775-5782. (36) Tian, H.; Landers, J. P. Anal. Biochem. 2002, 309, 212-223.

2172 Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

Table 1. Comparison of the Resolutions of On-Chip Electrophoreses for dsDNAa method

R100-200b

R500-600b

R900-1000b

ref

IMA chip cross injection Ito et al. Shaikh et al.

1.68 ( 3.07 ( 0.97d 2.77 1.26f

1.36 ( 1.64 ( 0.20d 1.70 1.12f

0.82 ( 0.95 ( 0.21d 0.83 0.83f

this work SIe 21 8

0.07c

0.10c

0.07c

a A 100-1000-bp dsDNA ladder containing 10 fragments. b R 100-200, R500-600, and R900-1000 denote resolutions between 100- and 200-bp fragments, 500- and 600-bp fragments, and 900- and 1000-bp fragments, respectively. c N ) 12. d N ) 4. e Supporting Information. f Measured from the electropherogram (Figure 2G in ref 8).

tion (RSD) of migration times on the IMA chip ranged from 4.5% (100 bp) to 6.7% (1000 bp). The RSD of peak heights ranged from 17.5% (300 bp) to 33.7% (1000 bp). Sequence-Specific Separation of ssDNA. We utilized the IMA chip for sequence-specific separation of ssDNA on the basis of ACE, in which probe DNA immobilized to the sieving polymer matrix binds to specific sequences in the sample.21-24 Migrating through the probe-polymer conjugate, the sample components are separated by the difference in affinity to the probe. As model sequences, we adopted human K-ras gene and its codon 12 point mutant (GGT/AGT). First, we optimized the length of the probe DNA using 12-mer oligonucleotides as the samples: W(12) has a sequence identical to that of codons 10-13 of the wild type, whereas M(12) has a sequence identical to that of the point mutant. Four probe DNAs with different chain lengths, P(6), P(8), P(10), and P(12), were designed. P(12) is fully complementary to W(12), and other probes are partial sequences of P(12). All the probes make single-base mismatches with M(12). The probe DNAs were covalently bound to PDMA chains at their 5′ ends. An equimolar mixture of W(12) and M(12) was

Figure 5. Sequence-specific separation of ssDNA using ACE on the IMA chip. (a) Optimization of the chain length of probe DNA using a mixture of W(12) and M(12) as a sample. Channels 4 and 9 were used for control experiments without probe DNA. Channels 5, 6, 7, and 8 contained P(6)-PDMA, P(8)-PDMA, P(10)-PDMA, and P(12)-PDMA, respectively. Conditions: electric field, 6.3 V/cm; buffer, 40 mM Tris-borate (pH 7.4) with 1.0 mM MgCl2; polymer matrix, 10% PDMA with 50 µM probe DNA; sample concentration, 1.0 µM for each; label, FITC; 4× objective lens; temperature, ∼25 °C. (b) Typing of PCR-amplified samples using P(8)-PDMA. Channels 1, 4, 7, and 10 were used for the WT/MT mixture, channels 2, 5, 8, and 11 were used for WT, and channels 3, 6, 9, and 12 were used for MT. Conditions were the same as those in (a) except that the sample concentration was not determined.

diluted with an equal volume of 10% PDMA (without probe) for dynamic coating of the channel walls. We simultaneously tested the four probe DNAs in the four central microchannels (channels 5-8 in Figure 1b; channels 4 and 9 were used for control experiments without a probe). The results are shown in Figure 5a. At the beginning of the electrophoresis (20 s), we observed significant sample compaction due to the discontinuous mobility at the sample/polymer interface. The positions of the bands at 90 s can be explained by the affinities between the probes and the sample components: the affinity of P(6) in channel 5 was too weak to separate W(12) and M(12), whereas the affinity of P(12) in channel 8 was too strong. Separation was possible with P(8) in channel 6 and P(10) in channel 7; the best separation performance was obtained with P(8). On the basis of the above probe optimization, we separated PCR products with the same type of single-base substitution to confirm the practical usefulness of ACE on the IMA chip. The two DNA templates and the primer set were chosen to amplify

K-ras codons 1-21 (60 bp). The PCR products were checked by electrophoresis in a slab gel, and a single band was observed for each (data not shown). We obtained the FITC-labeled single strands using the streptavidin-coated MBs; the wild-type and mutant-type products are denoted by WT and MT, respectively. To mimic the heterozygote, we mixed WT and MT by a 1:1 volumetric ratio. WT, MT, and the mixed samples were diluted with an equal volume of 10% PDMA (without probe) for dynamic coating of the channel walls. The samples were loaded into the following microchannels (see Figure 1b): the mixed sample into channels 1, 4, 7, and 10, WT into channels 2, 5, 8, and 11, and MT into channels 3, 6, 9, and 12. Simultaneous ACE was carried out using P(8)-PDMA in all the microchannels. The results are shown in Figure 5b. As expected, in the corresponding microchannels, MT migrated toward the anode, whereas WT was trapped around the stop valves. Using the IMA chip, we were able to discriminate the 3 samples in the 12 independent microchannels at the same time. CONCLUSIONS We have demonstrated a novel IMA chip and parallel electrophoreses on it. The unique I-shaped channel design has the following advantages: straightforward integration of the channels onto a minimum area of a microchip, a simple electric wiring system that can be integrated onto the same microchip, and autonomous regulation of the sample plugs. The usefulness of the IMA chip was verified with size-dependent separation of dsDNA and sequence-specific separation of ssDNA. We believe that the utility of the IMA chip can easily be expanded to proteins and other biomolecules. The IMA chip will open up a new possibility of large-scale integration of microchannels for highthroughput electrophoretic analyses. ACKNOWLEDGMENT We thank Dr. Masaki Ihara at RIKEN for his technical assistance for purification of PCR products. SUPPORTING INFORMATION AVAILABLE Measurements of fluid displacements on PDMS-PMMA hybrid microchips and separation of dsDNA with a conventional cross injection method. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review August 29, 2006. Accepted January 3, 2007. AC0616097

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

2173