Autonomous Polymer Loading and Sample Injection for Microchip

Then the mixture was placed in an incubator (CTU-N,. Taitec; Koshigaya, Japan) at 50 °C for 1.5 h. The probe DNA-. PDMA conjugate was used for affini...
0 downloads 0 Views 447KB Size
Download PDF
Anal. Chem. 2005, 77, 4759-4764

Autonomous Polymer Loading and Sample Injection for Microchip Electrophoresis Toshiyuki Ito, Akira Inoue, Kae Sato, Kazuo Hosokawa,* and Mizuo Maeda

Bioengineering Laboratory, The Institute of Physical and Chemical Research (RIKEN), 2-1 Hirosawa, Wako, Saitama 351-0198, Japan

Over the past decade, microfluidic chips (microchips) for chemical processing, the so-called “micro total analysis system” or “lab on a chip”, have been attracting considerable interests.1-3 Microchips have been applied to biochemical analysis,4 chemical synthesis,5 clinical analysis,6 and tissue engineering7 and have reduced sample and reagent consumption, processing time, space, energy, wastes, and costs. Microchips have been traditionally fabricated by wet etching of glass plates.8 Alternatively, soft

lithography of poly(dimethylsiloxane) (PDMS) provides a rapid and inexpensive fabrication technology without access to clean room facilities.9 Because of its easy fabrication, PDMS has become one of the standard materials for microchips.10 Electrophoresis is the most classical and representative application of microchips.8,9,11 Microchip electrophoresis is usually carried out through the following three steps: (1) a microchannel is filled with a sieving polymer solution such as polyacrylamide, (2) a sample is injected into the microchannel, and (3) components in the sample are separated by applying an electric field along the microchannel. For step 1, the polymer solution is usually introduced by external pressure. This is not an easy task when the polymer solution has a high viscosity.12 Especially for PDMS microchips, air bubbles are often trapped because of the hydrophobic surfaces of PDMS (water contact angle ∼105°13). For step 2, the most common technique is electrokinetic injection. To obtain a well-defined and reproducible shape of the sample region (sample plug), various microchannel designs (cross, single-T, and double-T injectors) as well as voltage control schemes have been investigated.11,14 On the other hand, pressure-driven12,15,16 and hydrodynamic17 techniques for sample injection have also been reported. With these techniques, samples were not biased by different migration rates of the analyte molecules, and a sample compaction effect at the sample/polymer interface was observed.12 So far, a fully autonomous microchip enabling the above steps 1 and 2 without switching of voltage or pressure has not been reported. Affinity electrophoresis is a separation method in which a specific noncovalent binding interaction between ligand and analyte compounds is utilized to enhance separation selectivity and sensitivity.18 With this method, sequence-specific separation of single-stranded DNA (ssDNA) is possible using an oligonucleotide as an affinity ligand immobilized to cross-linked or

* To whom correspondence should be addressed. E-mail: [email protected]. Tel: +81-48-467-9312. Fax: +81-48-462-4658. (1) Reyes, D. R.; Iossifidis, D.; Auroux, P. A.; Manz, A. Anal. Chem. 2002, 74, 2623-2636. (2) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem. 2002, 74, 2637-2652. (3) Vilkner, T.; Janasek, D.; Manz, A. Anal. Chem. 2004, 76, 3373-3386. (4) Bilitewski, U.; Genrich, M.; Kadow, S.; Mersal, G. Anal. Bioanal. Chem. 2003, 377, 556-569. (5) Watts, P.; Haswell, S. J. Curr. Opin. Chem. Biol. 2003, 7, 380-387. (6) Liu, Y.; Garcia, C. D.; Henry, C. S. Analyst 2003, 128, 1002-1008. (7) Andersson, H.; van den Berg, A. Lab Chip 2004, 4, 98-103. (8) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897.

(9) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974-4984. (10) Sia, S. K.; Whitesides, G. M. Electrophoresis 2003, 24, 3563-3576. (11) Ugaz, V. M.; Elms, R. D.; Lo, R. C. Shaikh, F. A.; Burns, M. A. Philos. Trans. R. Soc., London A 2004, 362, 1105-1129. (12) Tabuchi, M.; Ueda, M.; Kaji, N.; Yamasaki, Y.; Nagasaki, Y.; Yoshikawa, K.; Kataoka, K.; Baba, Y. Nat. Biotechnol. 2004, 22, 337-340. (13) Ocvirk, G.; Munroe, M.; Tang, T.; Oleschuk, R.; Westra, K.; Harrison, D. J. Electrophoresis 2000, 21, 107-115. (14) Roddy, E. S.; Xu, H.; Ewing, A. G. Electrophoresis 2004, 25, 229-242. (15) Solignac, D.; Gijs, M. A. M. Anal. Chem. 2003, 75, 1652-1657. (16) Lee, N. Y.; Yamada, M.; Seki, M. Anal. Sci. 2004, 20, 483-487. (17) Backofen, U.; Matysik, F. M.; Lunte, C. E. Anal. Chem. 2002, 74, 40544059. (18) Heegaard, N. H. H. Electrophoresis 2003, 24, 3879-3891.

We have developed an extremely simple method for microchip electrophoresis. Loading of a sieving polymer solution and injection of a sample solution are autonomously executed by a microchip fabricated in poly(dimethylsiloxane) (PDMS). In advance, the energy for the fluid pumping is stored in bulk PDMS by evacuating air dissolved in PDMS, and the information for the sample plug regulation is coded into the microchannel design. Besides the simplicity, our method brings about an advantageous effect: sample compaction due to the discontinuous electrophoretic mobility at the sample/ polymer interface. The sample compaction effect was moderate in ordinary size-dependent separation for doublestranded DNA and was extreme in affinity electrophoresis for single-stranded DNA (ssDNA). In the latter separation mode, ssDNA components were sequence-specifically separated by difference in affinity to a probe oligonucleotide immobilized to the sieving polymer matrix. We separated up to 60-mer ssDNA mixtures based on singlebase substitutions. The separation processes included typically 100-fold sample compaction and were completed within 15-30 s. This technology provides easy, simple, and sensitive method for detection of gene point mutations and typing of single-nucleotide polymorphisms.

10.1021/ac050122f CCC: $30.25 Published on Web 06/14/2005

© 2005 American Chemical Society

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 4759

linear polymer matrix. This idea has been implemented in capillaries19-22 as well as in microchannels.23-25 Separation based on a single-base substitution (SBS) under isothermal conditions has been realized in either format.20,21,23 Other groups achieved concentration enrichment of target sequences by 2 orders of magnitude in microchannels.24,25 However, separation based on an SBS accompanied by a large factor of concentration enrichment has not yet been reported. In this paper, we describe a PDMS microchip that executes the above steps 1 and 2 in a fully autonomous fashion. The polymer and the sample solutions are simultaneously introduced by energy stored in the microchip, and the shape of the sample plug is regulated by the geometry of the microchannel. Two separation modes for DNA, size-dependent and sequence-specific, are demonstrated. A key technology here is the power-free pumping technique,26 which exploits the high solubility and the rapid diffusion of air in PDMS.27 Air dissolved in PDMS can be evacuated in a vacuum. When the microchip is brought back to the atmosphere, air is dissolved into PDMS again. The redissolution through the microchannel walls drives the solution movement and prevents air bubble trapping. The degassed microchip can be preserved in an airtight packaging. Unlike the one-way pumping in the previous report,26 formation of stationary interfaces of the solutions is demonstrated here. Combination of the new fluid-handling technique and affinity electrophoresis has produced an easy method enabling separation of ssDNA based on an SBS accompanied by concentration enrichment of 2 orders of magnitude. EXPERIMENTAL SECTION Materials. Silicon wafers were obtained from Osaka Special Alloy (Osaka, Japan). Ultrathick photoresist (SU-8 25) and its developer were obtained from MicroChem (Newton, MA). Prepolymer of PDMS (Sylgard 184) was obtained from Dow Corning (Midland, MI). A 100-1000-bp double-stranded DNA (dsDNA) ladder containing 10 fragments was obtained from Bio-Rad Laboratories (Hercules, CA). Hydroxyethylcellulose (HEC) was obtained from Polyscience (Warrington, PA). SYBR Gold was obtained from Molecular Probes (Eugene, OR). For samples, four types of 5′-fluorescein isothiocyanate (FITC)-labeled ssDNA were obtained from Espec Oligo Service (Tsukuba, Japan). Sequences are as follows: N(12) (FITC-5′-GGA GCT GGT GGC-3′), M(12) (FITC-5′-GGA GCT AGT GGC-3′), N(60) (FITC-5′-G ACT GAA TAT AAA CTT GTG GTA GTT GGA GCT GGT GGC GTA GGC AAG AGT GCC TTG ACG AT-3′), and M(60) (FITC-5′-G ACT GAA TAT AAA CTT GTG GTA GTT GGA GCT AGT GGC GTA GGC (19) Muscate, A.; Natt, F.; Paulus, A.; Ehrat, M. Anal. Chem. 1998, 70, 14191424. (20) Katayama, Y.; Arisawa, T.; Ozaki, Y.; Maeda, M. Chem. Lett. 2000, 29, 106107. (21) Anada, T.; Arisawa, T.; Ozaki, Y.; Takarada, T.; Katayama, Y.; Maeda, M. Electrophoresis 2002, 23, 2267-2273. (22) Murakami, Y.; Maeda, M. J. Biosci. Bioeng. 2003, 96, 95-97. (23) Ito, T.; Inoue, A.; Sato, K.; Hosokawa, K.; Maeda, M. Chem. Lett. 2003, 32, 688-689. (24) Olsen, K. G.; Ross, D. J.; Tarlov, M. J. Anal. Chem. 2002, 74, 1436-1441. (25) Paegel, B. M.; Yeung, S. H. I.; Mathies, R. A. Anal. Chem. 2002, 74, 50925098. (26) Hosokawa, K.; Sato, K.; Ichikawa, N.; Maeda, M. Lab Chip 2004, 4, 181185. (27) Merkel, T. C.; Bondar, V. I.; Nagai, K.; Freeman, B. D.; Pinnau, I. J. Polym. Sci. Pol. Phys. 2000, 38, 415-434.

4760 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Figure 1. Microchannel design. The microchannel has a single-T injector, three reservoirs (2-mm diameter), and a cross section of 100 µm width × 25 µm depth except for the two pinched regions (20-µm width). Two types of microchannels were fabricated: L1 ) 10 mm and L2 ) 1 mm for ordinary electrophoresis, and L1 ) 5 mm and L2 ) 10 mm for affinity electrophoresis.

AAG AGT GCC TTG ACG AT-3′). For polymerase chain reaction (PCR), two DNA templates (ras mutant set, c-Ki-ras codon 12) containing the same sequences as N(60) and M(60), respectively, were obtained from Takara Bio (Otsu, Japan). FITC-labeled forward primer (FITC-5′-GAC TGA ATA TAA ACT TGT GG-3′), and unlabeled reverse primer (5′-ATC GTC AAG GCA CTC TTG CC-3′) was obtained from Espec Oligo Service. Pyrobest DNA polymerase, Pyrobest reaction buffer, and a deoxyribonucleoside triphosphate (dNTP) mixture were obtained from Takara Bio. For recognition probes, 5′-methacryloyl-modified oligonucleotides were obtained 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′), P(11) (5′-GCC ACC AGC TC-3′), and P(12) (5′-GCC ACC AGC TCC-3′). Tris(hydroxymethyl)aminomethane (Tris) was obtained from SigmaAldrich Japan (Tokyo, Japan). Boric acid, N,N-dimethylacrylamide (DMA), ammonium persulfate (APS), and N,N,N′,N′-tetramethylethylenediamine (TEMED) were obtained from Wako Pure Chemical Industries (Osaka, Japan). Magnesium chloride hexahydrate was obtained from Kanto Kagaku (Tokyo, Japan). Deionized water was produced with a Direct-Q system (Nihon Millipore; Tokyo, Japan). Microchip Fabrication. Figure 1 expresses two microchannel layouts, which are almost the same but different in the position of the single-T injector (L1) and the interval length between the two pinched regions (L2). Soft lithography of PDMS was described elsewhere.28 Briefly, a negative master was fabricated on a silicon wafer with the ultrathick photoresist using a transparency mask. After passivation of the master using a plasma etcher (RIE-10NR, Samco International, Kyoto, Japan), prepolymer of PDMS was cast onto the master with a frame for holding the prepolymer solution. The cured PDMS (30 × 30 × 2 mm) was peeled off from the master, and through holes for reservoirs were punched in the PDMS chip using a metal pipe. The PDMS chip with surface relief was reversibly bonded to a microscope slide glass (S9112, 76 × 52 × 1 mm, Matsunami Glass; Osaka, Japan). (28) Hosokawa, K.; Fujii, T.; Endo, I. Anal. Chem. 1999, 71, 4781-4785.

Figure 2. Major steps of the experiment. (a) The sample solution and the polymer solutions in the reservoirs are autonomously introduced into the microchannel due to the redissolution of air into PDMS channel walls. (b) The solutions are merged together at the passive stop valves. (c) Electrophoresis is started by applying voltages to the reservoirs.

Synthesis of Probe DNA-Polymer Conjugate. To synthesize probe DNA-poly(N,N-dimethylacrylamide) (PDMA) conjugate, 20 µL of 98% DMA solution, 20 µL of 1 mM 5′-methacryloylmodified oligonucleotide, 20 µL of 100 mM MgCl2, and 120 µL of 25 mM Tris-borate buffer (pH 7.4) were mixed, and copolymerization was initiated by adding 10 µL of 10 wt % APS and 10 µL of TEMED. Then the mixture was placed in an incubator (CTU-N, Taitec; Koshigaya, Japan) at 50 °C for 1.5 h. The probe DNAPDMA conjugate was used for affinity electrophoresis without further purification or dilution. Asymmetric PCR. The two DNA templates were independently amplified. Each reaction mixture contained 1 ng of the template, 20 pmol of the FITC-labeled forward primer, 5 pmol of the reverse primer, 2.5 units of Pyrobest DNA polymerase, and 20 nmol of each dNTP in a 100-µL volume. After preheating at 95 °C for 10 min, 35 cycles were performed using steps of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s in a thermocycler (TGradient, Biometra, Go¨ttingen, Germany). The solutions were kept at 72 °C for 5 min and then cooled. Without further purification, the PCR products were mixed together and diluted twice with 25 mM Tris-borate buffer (pH 7.4). Solution Filling and Electrophoresis. PDMS microchip 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 5 min. After the microchip was brought back to the atmosphere, 3-µL aliquots of DNA sample and sieving polymer solutions were dispensed into the reservoirs within 2 min. Autonomous filling of the solutions were observed with an inverted fluorescence microscope (Eclipse TE2000-U, Nikon, Tokyo, Japan). Electrophoresis was carried out at room temperature (∼25 °C) by

Figure 3. Size-dependent separation of the 100-bp dsDNA ladder. (a) Electropherogram. (b) A representative image. Channel cross section, 100 µm × 25 µm; electric field, 50 V/cm; buffer, 1 × Trisborate-EDTA; polymer matrix, 2% HEC; initial sample concentration, 50 µg/mL; initial sample plug length, 1 mm; label, SYBR Gold; 4× objective lens.

applying voltages to platinum wires (0.5-mm diameter) dipped into the microchip reservoirs using a computer-controlled high-voltage power supply developed and fabricated in-house. Fluorescence images were monitored using the fluorescence microscope (100-W high-pressure mercury lamp, excitation 465-495 nm, emission 515-555 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 Figure 2 illustrates autonomous polymer loading and sample injection followed by electrophoresis. At atmospheric pressure, the degassed microchip absorbs air through all the PDMS surfaces including the walls of the microchannel. When all the reservoirs are filled with the solutions, the redissolution of air through the walls starts to reduce the air pressure in the microchannel, because air is not supplied from the outside any more. The pressure reduction causes spontaneous introduction of the solutions26 (Figure 2a), and they are eventually merged together (Figure 2b). Under the experimental conditions described above, the complete filling took 2-5 min. This time should be dependent on channel geometry, extent of degassing, and viscosities of the solutions. An important advantage of this filling technique is that no air bubble is trapped at the channel walls. If an air bubble is accidentally generated, it adheres to the hydrophobic PDMS surface rather than the hydrophilic glass surface, and it is quickly absorbed into PDMS. In the filling procedure above, the most critical issue is regulation of the sample plug defined by the locations where the solutions are merged together. For autonomous and precise regulation, we constructed the two pinched regions (20 µm wide; Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

4761

Figure 4. Schematics around the right stop valve illustrating sequence-specific separation based on a SBS by affinity microchip electrophoresis. (a) The sample solution containing two types of ssDNA and the probe DNA-PDMA conjugate solution are merged at the stop valve by the autonomous filling technique. The black dot indicates the SBS. (b) The sample plug is compacted at the entrance of the conjugate region by electrophoresis. (c) M migrates faster than N because of difference in affinity to the probe DNA. Mobility of N is assumed to be zero in this figure.

see Figure 1) within the microchannel as passive stop valves. When one of the solutions reaches a passive stop valve first, the solution stops there until it is merged with the second solution, because the abrupt change in cross section impedes the passage of the meniscus by capillary force.16,28 (Supporting Information (SI) includes a 15-s movie extracted from an original length of 150 s.) Therefore, the length of the sample plug is basically determined by the interval between two passive stop valves (L2 in Figure 1). We fabricated two microchips with different intervals: 1 mm for ordinary electrophoresis of dsDNA and 10 mm for affinity electrophoresis of ssDNA. Another consideration for the sample plug regulation is suppression of diffusion between the merged solutions. For this purpose, the left reservoir was filled with a polymer solution, which was not necessarily the same as the sieving polymer solution in the right reservoir. Figure 3a is an electropherogram of the 100-bp dsDNA ladder sieved with HEC, which has a surface-coating effect for suppression of electroosmosis.29 The detection point was set at 5 mm downstream from the right stop valve. All peaks were clearly separated within 3 min. Passage of the sample through the stop valve did not seem to disturb the electrophoretic separation seriously. A representative fluorescence image is shown in Figure 3b (a movie is included in SI). The image indicates that the sample plug was significantly compacted from its initial length, 1 mm. This sample compaction, caused by the lower DNA mobility in the HEC solution than in the sample buffer, is advantageous for signal enhancement as well as shortening of separation channel length.30 (29) Tian, H.; Landers, J. P. Anal. Biochem. 2002, 309, 212-223. (30) Brahmasandra, S. N.; Ugaz, V. M.; Burke, D. T.; Mastrangelo, C. S.; Burns, M. A. Electrophoresis 2001, 22, 300-311.

4762 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Figure 5. Sequence-specific separation of N(12) and M(12) using P(10)-PDMA conjugate. (a) Fluorescence images showing the separation process. (b) Concentration profiles. Channel cross section, 100 µm × 25 µm; electric field, 250 V/cm; buffer, 15 mM Tris-borate (pH 7.4) with 10 mM MgCl2; polymer matrix, 10% PDMA with 100 µM probe DNA; initial sample concentration, 0.1 µM for each; initial sample plug length, 10 mm; label, FITC; 10× objective lens.

The sample compaction effect became dramatic in affinity electrophoresis of ssDNA. Figure 4 illustrates affinity electrophoresis in our microchip for separation of two types of ssDNA, normal type (N) and mutant type (M) (Figure 4a). The two types are identical except for an SBS indicated by the black dot. The probe DNA, immobilized to the polymer matrix, has a complementary sequence to a part of N and makes a single-base mismatch with M. Migrating through the polymer region, N and M are separated by difference in affinity to the probe DNA (Figure 4c). Depending on the affinity design (length and concentration of the probe DNA, ionic strength, etc.), the mobility in the polymer region can be very low, which means a high degree of sample compaction in our microchip. For full utilization of this advantage, we adopted the second microchannel layout producing an extraordinarily long initial sample plug (L2 ) 10 mm). For N, we chose a 12-mer model sequence identical to human K-ras gene codons 10-13, denoted by N(12). Its mutant, M(12), has an SBS at the middle of the whole sequence. The SBS is indicated by the underline in the sequence in the Experimental Section. An equimolar mixture of N(12) and M(12) was used as a sample solution. For recognition probes, we designed five

Figure 6. Fluorescence images showing the migration behavior of N(12), M(12), or both under the same electrophoresis conditions as in Figure 5. (a) Pure sample of N(12) in P(10)-PDMA. (b) Pure sample of M(12) in P(10)-PDMA. (c) Mixture sample in P(6)-PDMA. (d) Mixture sample in P(8)-PDMA. (e) Mixture sample in P(12)PDMA.

sequences: P(6), P(8), P(10), P(11), and P(12). The numbers denote the chain lengths; P(12) is fully complementary to N(12), and other probes are partial sequences of P(12). All the probes make single-base mismatches with M(12). The probes were separately conjugated with PDMA, which also has the surfacecoating effect.31 Figure 5a shows the separation process of N(12) and M(12) using the P(10)-PDMA conjugate. The fluorescence images were taken at the right stop valve. In Figure 5a (0 s), corresponding to Figure 4a, the sample solution on the left side of the stop valve did not have sufficient concentration to be detected with our setup. By applying an electric field, the sample was concentrated just after entering into the conjugate region (Figure 5a (5 s)). The sample plug was compacted from 10 to 0.1 mm: roughly 100-fold compaction. We also observed a transient increase in sample concentration just before entering into the conjugate region. The mechanism of this unexpected phenomenon is now under investigation. Subsequently, M(12) slowly migrated toward the anode, (31) Madabhushi, R. S. Electrophoresis 1998, 19, 224-230.

Figure 7. Fluorescence images showing the separation processes of N(60) and M(60). (a) Chemically synthesized N(60) and M(60) in P(10)-PDMA under the same electrophoresis conditions as in Figure 5. (b) Asymmetric PCR-amplified N(60) and M(60) in P(11)-PDMA under the same electrophoresis conditions as in Figure 5 except for the sample concentrations: less than 50 nM for each.

while N(12) was strongly immobilized. As a result, the two types of ssDNA were separated by the SBS within analysis time of 15 s and separation length of 1 mm. Intensity profiles in gray scale values were obtained from another set of images. (The maximum signal is saturated in Figure 5a.) The gray scale values were converted to concentration values by calibration with standard solutions diluted from another equimolar mixture of N(12) and M(12). The resulting concentration profiles are shown in Figure 5b. The peak concentrations of N(12) and M(12) were calculated as 13 and 7 µM, corresponding to enrichment factors of 130 and 70, respectively. These values are consistent with the roughly estimated sample compaction factor, 100. Here we used the conventional fluorescence microscope for detection. If the detection setup is optimized, our method would be applicable to very dilute samples, because the sample compaction effect enhances the detection capability by 2 orders of magnitude. The band identification described above was confirmed by two independent runs for pure samples of N(12) and M(12). As expected, N(12) was strongly trapped at the entrance of the conjugate region (Figure 6a), whereas M(12) slowly migrated Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

4763

(Figure 6b). The velocity of M(12) was nearly equal to that in the mixture case shown in Figure 5a. Effect of probe length was also investigated by applying P(6)-, P(8)-, and P(12)-PDMA conjugates to the mixture sample. Figure 6c-e shows the results, which are understandable from the fact that a longer probe DNA has stronger affinity to the sample DNA and, hence, yields slower migrations. With P(6)-PDMA (Figure 6c), both components migrated, and the sample compaction was not as effective as with P(10)-PDMA. Separation was barely possible within the range of imaging. With P(8)-PDMA (Figure 6d), the migration behaviors were intermediate between those with P(6)- and P(10)PDMA. This result reflects the intermediate affinity produced by P(8). The clarity of separation is comparable to that in Figure 5a. With P(12)-PDMA (Figure 6e), both N(12) and M(12) were so strongly trapped that no separation was observed at least for 1 min. Longer ssDNA samples were also examined. We purchased a set of chemically synthesized 60-mer: N(60) and M(60). The former has the same sequence as K-ras codons 1-21, while the latter contains the same type of SBS as M(12) at the corresponding position. Separation process of N(60) and M(60) using P(10)PDMA conjugate is shown in Figure 7a, which closely resembles Figure 5a. The sample compaction and the sequence-specific separation were not affected by the 5-fold elongation of sample DNA. In affinity electrophoresis, the migration behavior is considered to be dominated by local interaction between probe DNA and recognition sequence in sample DNA. This interaction is not affected by chain lengths of nonrecognition sequences in principle. However, we do not expect that this is always true for extremely long samples; the recognition sequence might be sterically masked by higher-order structures. This limitation is common to all assay methods based on DNA hybridization. To confirm the practical applicability of our method to genetic analysis, we tried to separate a mixture of unpurified PCR products. The two DNA templates and the set of primers were

chosen to amplify identical sequences to N(60) and M(60). Asymmetric PCR32 was conducted to obtain FITC-labeled ssDNA by adding the excess amount of the FITC-labeled forward primer. The final concentrations of the ssDNA products were obviously limited by the amounts of the forward primer and were calculated as less than 50 nM for each. The separation process of the PCRamplified N(60) and M(60) using P(11)-PDMA conjugate is shown in Figure 7b. After passage of a large band of impurities (probably dsDNA and unreacted primer), two distinct bands corresponding to N(60) and M(60) were observed. Using P(10)PDMA conjugate as in the case of chemically synthesized sample, only two bands corresponding to N(60) and the impurities were observed (data not shown). We suppose that the band of M(60) was masked by the large band of impurities. The existence of impurities made separation difficult. However, by optimizing the probe length, we realized simultaneous sample purification, concentration, and sequence-specific separation. Our method is promising for practical application to genetic analysis.

(32) Rehbein, H.; Mackie, I. M.; Pryde, S.; Gonzales-Sotelo, C.; Perez-Martin, R.; Quinteiro, J.; Rey-Mendez, M. Electrophoresis 1998, 19, 1381-1384. (33) Nollau, P.; Wagener, C. Clin. Chem. 1997, 43, 1114-1128. (34) Syvanen, A. C. Nat. Rev. Genet. 2001, 2, 930-942.

Received for review January 21, 2005. Accepted May 17, 2005.

4764

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

CONCLUSIONS We have achieved a great simplification for microchip electrophoresis by the new solution filling technique. Loading of the sieving polymer and formation of the sample plug are autonomously carried out by the degassed PDMS microchip. The microchip is especially suitable for affinity electrophoresis of ssDNA, because the extreme sample compaction and the sequencespecific separation based on an SBS are easily attainable in a short period. The most promising applications of this technology are detection of gene point mutations33 and typing of single-nucleotide polymorphisms.34 SUPPORTING INFORMATION AVAILABLE A movie demonstrating the solution merging at the stop valves and a movie demonstrating the separation of dsDNA. This material is available free of charge via the Internet at http://pubs.acs.org.

AC050122F