Integrated Affinity Capture, Purification, and Capillary Electrophoresis

Anal. Chem. 2007, 79, 8549-8556

Integrated Affinity Capture, Purification, and Capillary Electrophoresis Microdevice for Quantitative Double-Stranded DNA Analysis Nicholas M. Toriello,† Chung N. Liu,‡ Robert G. Blazej,† Numrin Thaitrong,§ and Richard A. Mathies*,†,§

UCSF/UC Berkeley Joint Graduate Group in Bioengineering, Department of Chemical Engineering, and Department of Chemistry, University of California, Berkeley, California 94720

A novel injection method is developed that utilizes a thermally switchable oligonucleotide affinity capture gel to mediate the concentration, purification, and injection of dsDNA for quantitative microchip capillary electrophoresis analysis. The affinity capture matrix consists of a 20 base acrydite modified oligonucleotide copolymerized into a 6% linear polyacrylamide gel that captures ssDNA or dsDNA analyte including PCR amplicons and synthetic oligonucleotides. Double stranded PCR amplicons with complementarity to the capture probe up to 81 bases from their 5′ terminus are reproducibly captured via helix invasion. By integrating the oligo capture matrix directly with the CE separation channel, the electrophoretically mobilized target fragments are quantitatively captured and injected after thermal release for unbiased, efficient, and quantitative analysis. The capture process exhibits optimal efficiency at 44 °C and 100 V/cm with a 20 µM affinity capture probe (TM ) 57.7 °C). A dsDNA titration assay with 20 bp fragments validated that dsDNA is captured at the same efficiency as ssDNA. Dilution studies with a duplex 20mer show that targets can be successfully captured and analyzed with a limit of detection of 1 pM from 250 nL of solution (∼150 000 fluorescent molecules). Simultaneous capture and injection of amplicons from E. coli K12 and M13mp18 using a mixture of two different capture probes demonstrates the feasibility of multiplex target capture. Unlike the traditional cross-injector, this method enables efficient capture and injection of dsDNA amplicons which will facilitate the quantitative analysis of products from integrated nanoliterscale PCR reactors. As microfluidic systems evolve from simple one-step separation devices1,2 to fully integrated platforms3-12 capable of performing sample amplification, processing, and detection, the focus in * Corresponding author. Department of Chemistry, MS 1460, University of California, Berkeley, CA 94720. Phone: (510) 642-4192. Fax: (510) 642-3599. E-mail: [email protected]. † UCSF/UC Berkeley Joint Graduate Group in Bioengineering. ‡ Department of Chemical Engineering. § Department of Chemistry. (1) Woolley, A. T.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 1134811352. (2) 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. 10.1021/ac0712547 CCC: $37.00 Published on Web 10/12/2007

© 2007 American Chemical Society

microfluidics is shifting to the challenging tasks of minimizing sample/reactor volumes and the efficient integration of multiple process steps. The polymerase chain reaction (PCR) is the most common sample preparation step integrated onto microdevices. PCR coupled with microchip-based capillary electrophoretic (CE) separation systems provides rapid, high-resolution amplicon sizing of double-stranded DNA (dsDNA).3-13 However, the reproducibility, efficiency, and quantitative capabilities are compromised because of intrinsic limitations in the cross injection process. It is thus desirable to improve microchip-based CE and especially the sample injection process such that it is capable of quantitatively purifying, concentrating, and separating dsDNA derived from integrated low-volume microchip PCR. The ability to manipulate well-defined subnanoliter injection volumes using the cross-injector introduced in 1992 by Harrison, Manz and co-workers was one of the central elements contributing to the early success of microchip electrophoretic systems.14-16 A cross-injector, formed by crossing two orthogonal channels, enables sample loading in the first dimension and sample separa(3) Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086. (4) Lagally, E. T.; Emrich, C. A.; Mathies, R. A. Lab Chip 2001, 1, 102-107. (5) Koh, C. G.; Tan, W.; Zhao, M. Q.; Ricco, A. J.; Fan, Z. H. Anal. Chem. 2003, 75, 4591-4598. (6) Pal, R.; Yang, M.; Lin, R.; Johnson, B. N.; Srivastava, N.; Razzacki, S. Z.; Chomistek, K. J.; Heldsinger, D. C.; Haque, R. M.; Ugaz, V. M.; Thwar, P. K.; Chen, Z.; Alfano, K.; Yim, M. B.; Krishnan, M.; Fuller, A. O.; Larson, R. G.; Burke, D. T.; Burns, M. A. Lab Chip 2005, 5, 1024-1032. (7) Lagally, E. T.; Scherer, J. R.; Blazej, R. G.; Toriello, N. M.; Diep, B. A.; Ramchandani, M.; Sensabaugh, G. F.; Riley, L. W.; Mathies, R. A. Anal. Chem. 2004, 76, 3162-3170. (8) Blazej, R. G.; Kumaresan, P.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 7240-7245. (9) Liu, C. N.; Toriello, N. M.; Mathies, R. A. Anal. Chem. 2006, 78, 54745479. (10) Toriello, N. M.; Liu, C. N.; Mathies, R. A. Anal. Chem. 2006, 78, 79978003. (11) Liu, P.; Seo, T. S.; Beyor, N.; Shin, K.-J.; Scherer, J. R.; Mathies, R. A. Anal. Chem. 2007, 79, 1881-1889. (12) Easley, C. J.; Karlinsey, J. M.; Bienvenue, J. M.; Legendre, L. A.; Roper, M. G.; Feldman, S. H.; Hughes, M. A.; Hewlett, E. L.; Merkel, T. J.; Ferrance, J. P.; Landers, J. P. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 19272-19277. (13) Roper, M. G.; Easley, C. J.; Landers, J. P. Anal. Chem. 2005, 77, 38873893. (14) Manz, A.; Harrison, D. J.; Verpoorte, E. M. J.; Fettinger, J. C.; Paulus, A.; Ludi, H.; Widmer, H. M. J. Chromatogr. 1992, 593, 253-258. (15) Harrison, D. J.; Manz, A.; Fan, Z.; Luedi, H.; Widmer, H. M. Anal. Chem. 1992, 64, 1926-1932. (16) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z. H.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007 8549

tion in the second dimension. This injection method has been used for the rapid and efficient separation of DNA, amino acids,16,17 and proteins in two-dimensions.18,19 Despite its widespread adoption, the cross-injection scheme has two fundamental limitations. First, the sample loading is extremely inefficient. A cross injector requires establishment of an electrophoretic quasi-steady state in the first dimension channel during sample loading to achieve an unbiased injection.20 However, unpurified samples generated from PCR contain high salt and an excess of short unreacted primers leading to long loading times and inefficient sample use. Second, injection timing becomes critical and highly variable when one is injecting limiting amounts of analyte from small nanoliter on-chip PCR reactors through a cross injector.7-11 In fully integrated microsystems, where the transfer efficiency between each nanoliter process must be optimized, more efficient injection methods are needed. Modified injection techniques that improve the efficiency of the sample loading process and enhance the sensitivity in CE can generally be categorized as stacking or extraction based.21 A few of the sample stacking approaches include gated injection,22 field amplified sample injection,23 basemediated stacking,24 the staggered T configuration,25 and pressuredriven injection.12,26 The simplicity of these approaches has led to their broad adoption in many microchip CE devices. However, no injection stacking approach has yet demonstrated efficient sample injection which is critical for highly sensitive and quantitative analysis. Extraction based techniques such as membrane filtration,27-29 solid-phase extraction,30,31 and liquid-liquid extraction32 offer a more scalable platform for complete sample injection. However, the implementation of these techniques in microfluidic systems for DNA analysis has yet to be fully realized because they are physically difficult to integrate and the chemistries required for extraction are typically incompatible with those needed for separation. One promising approach based on solidphase extraction is the use of affinity capture and purification based on DNA hybridization. Paegel et al. utilized an affinity capture matrix containing an acrylamide-copolymerized oligonucleotide to purify and concentrate single stranded (ssDNA) (17) Skelley, A. M.; Scherer, J. R.; Aubrey, A. D.; Grover, W. H.; Ivester, R. H. C.; Ehrenfreund, P.; Grunthaner, F. J.; Bada, J. L.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 1041-1046. (18) Gottschlich, N.; Culbertson, C. T.; McKnight, T. E.; Jacobson, S. C.; Ramsey, J. M. J. Chromatogr., B 2000, 745, 243-249. (19) Emrich, C. A.; Tian, H.; Medintz, I. L.; Mathies, R. M. Anal. Chem. 2002, 74, 5076-5083. (20) Ermakov, S. V.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2000, 72, 35123517. (21) Breadmore, M. C. Electrophoresis 2007, 28, 254-281. (22) Jacobson, S. C.; Koutny, L. B.; Hergenroeder, R.; Moore, A. W. J.; Ramsey, J. M. Anal. Chem. 1994, 66, 3472-3476. (23) Chien, R. L.; Burgi, D. S. J. Chromatogr. 1991, 559, 141-152. (24) Kim, D. K.; Kang, S. H. J. Chromatogr., A 2005, 1064, 121-127. (25) Bharadwaj, R.; Santiago, J. G.; Mohammadi, B. Electrophoresis 2002, 23, 2729-2744. (26) Easley, C. J.; Karlinsey, J. M.; Landers, J. P. Lab Chip 2006, 6, 601-610. (27) Khandurina, J.; Jacobson, S. C.; Waters, L. C.; Foote, R. S.; Ramsey, J. M. Anal. Chem. 1999, 71, 1815-1819. (28) Kim, S. M.; Burns, M. A.; Hasselbrink, E. F. Anal. Chem. 2006, 78, 47794785. (29) Kelly, R. T.; Li, Y.; Woolley, A. T. Anal. Chem. 2006, 78, 2565-2570. (30) Strausbauch, M. A.; Landers, J. P.; Wettstein, P. J. Anal. Chem. 1996, 68, 306-314. (31) Tian, H. J.; Huhmer, A. F. R.; Landers, J. P. Anal. Biochem. 2000, 283, 175-191. (32) Pedersen-Bjergaard, S.; Rasmussen, K. E.; Halvorsen, T. G. J. Chromatogr., A 2000, 902, 91-105.

8550

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

fragments for DNA sequencing.33 Blazej et al. adapted this ssDNA affinity capture technique and developed a microfabricated bioprocessor for integrated nanoliter-scale Sanger DNA sequencing requiring only 1 fmol of template.8 Despite this progress, improved methods for the concentration and injection of double-stranded amplicons derived from PCR based genetic assays are needed. Off-chip affinity capture of duplex DNA has been investigated through Hoogsteen triplex formation, frayed end capture, and helix invasion. Triplex DNA formation offers sequence specific capture and purification but is limited to fragments containing extended homopurine sequences.34 RecA mediated affinity capture of dsDNA is not limited by sequence structure but exhibits poor fidelity.35 Helix invasion shows the greatest promise for dsDNA affinity capture and has been demonstrated through the use of peptide nucleic acids (PNAs) in high molar excess.36 Here we present a quantitative integrated affinity capture, purification, concentration, and injection CE analysis microdevice and method that addresses the aforementioned issues for dsDNA analysis. The ssDNA affinity capture technique developed by Paegel et al. is extended to the integrated capture, direct injection, and CE analysis of dsDNA.33 We exploit helix invasion into duplex DNA for sequence specific capture by utilizing a high molar excess of ssDNA capture probe. This capture process demonstrates high selectivity for amplicons containing the complement to the capture sequence and excellent mismatch discrimination. The captured products are thermally released and efficiently injected into a CE channel for sized-based separation. The optimum capture temperature and electric field are determined, and the effect of capture probe design is explored. Capture and injection of two different PCR amplicons from E. coli K12 cells and M13mp18 illustrates the multiplexing capabilities of our approach. This affinity capture, injection, and separation technique lays the foundation for the quantitative multiplex analysis of PCR amplicons from nanoliter integrated PCR reactors. EXPERIMENTAL SECTION Microdevice Design. A schematic of the four-channel capture, purification, and direct injection CE separation microdevice is shown in Figure 1. The fabrication protocol is similar to previously described microdevices and the details are in the Supporting Information.9,10 The device is a four-layer glass-polydimethylsiloxane (PDMS) hybrid structure comprised of a glass manifold layer, PDMS valving membrane, glass fluidic/channel layer, and blank glass layer. The etched glass manifold layer actuates the PDMS microvalves for fluidic control.37 The glass fluidic/channel layer contains etched fluidic channels on the top side while chambers and separation channels are etched on the back side. Thermal compression bonding is used to fuse the fluidic/channel layer to a blank glass wafer to form enclosed glass channels. The three-layer micropump is formed by sandwiching the PDMS membrane between the diced glass manifold and the bonded fluidic/channel structure. (33) Paegel, B. M.; Yeung, S. H. I.; Mathies, R. A. Anal. Chem. 2002, 74, 50925098. (34) Povsic, T. J.; Dervan, P. B. J. Am. Chem. Soc. 1989, 111, 3059-3061. (35) Malkov, V. A.; Sastry, L.; Camerini-Otero, R. D. J. Mol. Biol. 1997, 271, 168-177. (36) Demidov, V. V.; Yavnilovich, M. V.; Belotserkovskii, B. P.; Frankkamenetskii, M. D.; Nielsen, P. E. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 2637-2641. (37) Grover, W. H.; Skelley, A. M.; Liu, C. N.; Lagally, E. T.; Mathies, R. A. Sens. Actuators, B 2003, 89, 315-323.

Figure 1. Layout of the four-channel capture, purification, and direct injection CE separation microdevice. (a) The four-layer glassPDMS-glass-glass microdevice consists of a three valve pump for metering nanoliter volumes of analyte, etched hold chambers, capture chambers, CE separation channels, and drilled fluidic access points at the sample, waste, cathode, and anode reservoirs. (b) Microphotograph of the three-layer pump used to meter analyte onto the device. Each complete cycle pumps 10 nL of liquid from the sample reservoir toward the hold chamber. (c) Microphotograph of the capture and hold chambers and the injection structure.

The four-channel microdevice comprises three main regions for dsDNA PCR amplicon sizing and quantitation. The first region is a three-valve pump for reproducible and precise metering of analyte onto the device (Figure 1b). The second element is the capture region, where dsDNA is affinity-captured, purified, and concentrated by the acrydite modified oligonucleotide matrix to generate a well-defined capture plug (Figure 1c). The third region is the analysis region, where thermally released DNA is directly injected, separated, and detected for peak area quantitation. Sample Preparation. PCR amplicons from E. coli K12 cells and M13mp18 template are used to characterize dsDNA capture, purification, and injection. For E. coli analysis, K12 cells (MG1655, American Type Culture Collection, Manassas, VA) transfected with a 3.9 kb PCR 2.1-TOPO vector (Invitrogen) for ampicillin and kanamycin resistance are grown overnight on ampicillin plates. Colonies exhibiting resistance are picked and grown overnight in LB medium containing 25 µg of kanamycin. Cultured E. coli K12 cells are washed three times with 1x PBS. A 50 µL reaction cocktail is prepared and is comprised of Platinum Taq Supermix kit (22 U/mL complexed recombinant Taq DNA polymerase with Platinum Taq Antibody, 22 mM Tris-HCl (pH 8.4), 55 mM KCl,

1.65 mM MgCl2, 220 µM dNTPs, and stabilizers), 0.4 µM of forward and reverse primers (Integrated DNA Technologies, Coralville, IA) along with 1000 cells. The primer set amplifies a 237 bp FAM-labeled product as shown in the Supporting Information (Table S-1). The thermal cycling protocol employed is comprised of an initial activation of the Taq polymerase at 95 °C for 40 s, followed by 35 cycles of 95 °C for 20 s, 52 °C for 30 s, and 72 °C for 30 s, and a final extension step for 5 min at 72 °C. For M13mp18 analysis, forward and reverse primers along with 1000 copies of template are added to the PCR supermix using the same conditions listed above. This primer set amplifies a 197 bp FAM-labeled product. All PCR amplicon working solutions are created by diluting the amplified product 1:25 in a PCR supermix buffer. Synthetic dsDNA fragments (20 bp and 60 bp, IDT) are used for quantitative dsDNA capture efficiency characterization. Synthetic dsDNA samples are formed by hybridizing two complementary oligonucleotides in a PCR buffer solution containing the same composition as described above at 55 °C for 15 min and subsequently holding at 44 °C. The sequences of each sample are shown in the Supporting Information, and samples are made fresh daily to prevent template degradation. Matrix Synthesis. A DNA affinity-capture gel is synthesized by copolymerizing linear polyacrylamide (LPA) with a 5′ acrydite modified capture oligonucleotide. The affinity-capture matrix is synthesized at 4 °C by sparging a 2 mL solution containing 6% w/v acrylamide, 1x TTE, and 40 nmol of the acrydite-modified oligonucleotide (IDT) for 2 h with argon followed by the addition of 0.015% w/v APS and TEMED giving a final concentration of 20 µM of capture oligo. K12-specific affinity matrix contains only the E. coli K12 capture oligo (5′-Acry-CCA GTA ATC ATC GTC TGG AT-3′, TM ) 57.7 °C; 50 mM monovalent salt, 20 µM) and the M13mp18-specific affinity-capture matrix contains only the M13mp18 capture oligo (5′-Acry-ACT GGC CGT CGT TTT ACA A-3′, TM ) 61.2 °C; 50 mM monovalent salt, 20 µM). A mixed affinity-capture matrix contains both M13mp18 and K12 capture probes at 20 µM to enable capture from both templates. The K12 capture oligo is complementary to a 20 base sequence in the 237 bp amplicon, 81 bases from the 5′ terminus. The M13 capture oligo is complementary to a 20 base sequence in the 197 bp amplicon, 49 bases from the 5′ terminus. Affinity-capture matrixes containing acrydite oligo concentrations ranging from 1 to 30 µM are also synthesized to investigate its effect on capture efficiency. A nonfunctionalized 5% LPA gel is used as the separation matrix for all experiments. Multichannel Direct Injection Microdevice Operation. As depicted in Figure 2, the microdevice operation begins by treating the channels for 1 min with a dynamic coating diluted in methanol (1:1; The Gel Company, San Francisco, CA, DEH-100) to suppress electroosmotic flow. Next, affinity capture matrix (20 µM, 6%, yellow) is loaded from each cathode (C) reservoir up to the separation channel cross at room temperature. Separation matrix (red) is then loaded from the central anode (A) past the capture chamber to the sample load cross. With all the valves opened, the rest of the system is hydrated by adding 3 µL of 1x TTE at the sample ports and applying a vacuum at the waste (W) reservoirs. The microdevice is then placed on a 44 °C temperature controlled stage. An additional 2 µL of 1x TTE is flushed through Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8551

lab/ under the Valves and Control Systems menu. Once the sample is loaded, the analyte is pumped toward the hold chamber (Figure 2c). A constant 100 V/cm field between the waste (W) and cathode (C) reservoirs electrophoretically drives the analyte toward the capture chamber. Analytes complementary to the capture probe are hybridized at the entrance of the capture chamber creating a sample plug (Figure 2d). The electric field between the waste and cathode is maintained until residual PCR reactants (excess primer, salts, and buffer) are washed into the cathode reservoir thus resulting in a purified amplicon sample plug (Figure 2e). A 23 s delay is employed between each pump cycle during the capture process to allow sufficient time for the analyte to migrate into the capture region and to prevent analyte accumulation in the hold chamber. Thirty pump cycles are used resulting in a total capture and wash time of 12.2 min. After the capture process is completed, the temperature of the entire device is raised to 80 °C to thermally release the captured DNA fragments from the affinity capture gel and the sample is separated with a field of 150 V/cm between the cathode and the anode (Figure 2f). The electrophoretically separated FAM-labeled products from all four lanes are detected using laser-induced fluorescence with the Berkeley rotary confocal scanner.38

Figure 2. Schematic of direct injection microdevice operation. (a) Affinity capture matrix (yellow) comprised of 6% linear polyacrylamide copolymerized with a 20 µM oligonucleotide capture probe is loaded from the cathode (C) reservoir up to the separation channel cross. Separation matrix (red) is loaded from the anode (A) past the capture chamber to the sample load cross. (b) Analyte is loaded into the sample reservoirs and the pump is primed. (c) An amount of 10 nL of sample is moved into the hold chamber with each pump cycle. (d) Analyte is electrophoretically driven toward the capture matrix by applying a constant 100 V/cm field between the waste (W) and cathode reservoirs, and complementary fragments are captured at the entrance of the capture chamber creating a sample plug, followed by electrophoretic washing (e). (f) The temperature of the entire device is raised to 80 °C to thermally release the immobilized DNA fragments from the affinity capture gel. The sample is separated with a field of 150 V/cm between the cathode and the anode.

the system to remove thermally expanded gel from the sample load cross. Next, the valves are closed and the analyte is added to the sample reservoir followed by actuating the micropump for three cycles (Figure 2b). A five step pump cycle with a 350 ms actuation time is used as described previously,37 and each pump cycle loads 10 nL of analyte into the system. The PDMS micropump is controlled through a custom Labview program via a NI-USB adapter (USB-6008, NI) to independently address three solenoid valves during the pumping sequence. The pressure and vacuum levels used for the actuation are 7 and 5 psi, respectively. Details on the fabrication of the PDMS valve/pump controllers and software can be found at http://zinc.cchem.berkeley.edu/ 8552

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

RESULTS AND DISCUSSION Electrophoretic Analysis. Electropherograms resulting from the integrated capture, purification and separation of ss and dsDNA are presented in Figure 3. FAM-labeled DNA fragments are metered (30 nL from three pump cycles), captured, and purified at 44 °C and 100 V/cm using a 6% LPA gel copolymerized with a 20 base (20 µM) acrydite-modified oligonucleotide capture probe. Purified fragments are then thermally released from the capture gel and electrophoretically separated at 80 °C and 150 V/cm. Figure 3a presents the electropherogram of a purified 20mer ssDNA (500 pM) with complete complementarity (O′) to the capture probe (O, see Table S-1 for sequences), generating the expected single peak. Figure 3b demonstrates the ability of the affinity capture matrix to also bind a dsDNA input target. A 20 bp dsDNA fragment (500 pM) complementary to the capture probe generates a single peak with a peak area similar to that in the ssDNA experiment. The high affinity capture probe concentration (20 µM) successfully displaces its competitor from the initially double-stranded 20mer and captures the strand of interest. A study was performed with a ssDNA 20mer (500 pM) containing a one base mismatch to the capture probe (G for T substitution three bases from 3′ terminus) to explore the sequence specificity of this capture (Figure 3c). The lack of a captured peak in the electropherogram verifies that a capture temperature of 44 °C ensures highly specific hybridization and purification. To study the effect of DNA length on capture efficiency, a dsDNA 60mer (500 pM) with complementarity at its 5′ terminus was captured and injected (Figure 3d). The electropherogram shows a single peak with a comparable peak area to the ss- and dsDNA 20mer experiments, suggesting that larger DNA fragments are captured with similar efficiencies. Figure 3e presents an electropherogram for the capture and direct injection of a 237 bp PCR amplicon generated from E. coli K12 cells. The amplicon (38) Shi, Y. N.; Simpson, P. C.; Scherer, J. R.; Wexler, D.; Skibola, C.; Smith, M. T.; Mathies, R. A. Anal. Chem. 1999, 71, 5354-5361.

Figure 3. Electropherograms from FAM-labeled DNA fragments captured and purified (44 °C and 100 V/cm) in a 6% linear polyacrylamide gel copolymerized with a 20 base (20 µM) oligonucleotide capture probe. Purified fragments are thermally released at 80 °C and electrophoretically separated at 150 V/cm. A depiction of each captured DNA fragment is also shown with the location of the 5′ FAM label (b) and the complementary sequence (black). The spectra are offset for viewing; the same intensity scale is used for all plots. (a) Capture and separation of a 500 pM ssDNA 20mer complementary to the 20 base capture probe. (b) Capture and separation of a 500 pM dsDNA 20mer complementary to the capture probe. (c) Analysis of a 500 pM ssDNA 20mer with one bp mismatch to the capture probe. (d) Capture and separation of a 500 pM dsDNA 60mer with complementarity at its 5′ terminus to the capture probe. (e) Capture and separation of a 237 bp ds PCR amplicon generated from E. coli K12 cells with complementarity 81 bp from the 5′ terminus. (f) A crossinjection with the same 237 bp ds PCR amplicon results in a very intense unreacted primer peak at 90 s and a product peak that is 1/30 in area.

is complementary to the affinity capture sequence 81 bp from its 5′ terminus. The observation of a single peak at 250 s shows that the 237 bp PCR amplicon is effectively captured, purified, and injected. An electropherogram from a cross-injection of the same 237 bp PCR amplicon is presented in Figure 3f for comparison. The amplicon peak resulting from the cross-injection is significantly weaker than the affinity captured and injected PCR amplicon but is narrower and exhibits better resolution. In addition, this separation of unpurified products contains the expected strong unreacted FAM-labeled primer peak at 90 s. The oligonucleotide-mediated affinity capture process demonstrated here has a number of fundamental advantages. First, it allows for purification and concentration of PCR amplicons from complex input solutions containing unreacted primers, high salt, and cellular debris. By capturing the input target DNA in a small

concentrated zone, the integrated capture technique enables higher sample loading capacity. The directly injected product generated from 30 nL of metered analyte results in >30× improvement in fragment signal strength compared to the traditional cross-injection procedure. In addition, this technique offers the ability to capture both ssDNA and dsDNA fragments expanding its application beyond DNA sequencing to a wide range of biochemical assays. Importantly, the capture process is not limited to targeting the frayed end of dsDNA fragments; a PCR amplicon from E. coli K12 cells containing sequence complementarity 81 bp from its 5′ terminus was reproducibly captured, purified, and detected. The cause of the reduced resolution for the captured PCR product is addressed below and in the Supporting Information. Double-Strand DNA Capture Optimization. The dependence of dsDNA capture on temperature, field strength, and capture oligo concentration was studied to maximize the capture efficiency. Capture optimization was performed with the FAMlabeled 237 bp E. coli PCR amplicon using a 6% LPA gel copolymerized with a 20 bp, 20 µM oligonucleotide capture probe targeted 81 bp from the 5′ terminus. After capture, purified fragments are thermally released from the capture gel and electrophoretically separated at 80 °C and 100 V/cm. Optimal capture efficiency is accessed by maximizing the fluorescent CE separation peak areas and minimizing the full width half-maximum (fwhm). Figure 4a presents the temperature dependence of the capture processes. The peak area is maximized at 44 °C, and the peak width is minimized at temperatures e44 °C. At temperatures 100× increase in signal strength because of the improved sample loading capacity and concentration provided by affinity capture. This 100-fold increase in detection limit will enable analysis of products from smaller PCR reactors operating at the stochastic template limit. This capture and injection process also ensures an unbiased injection of all products and offers the ability to correlate fluorescent peak area to number of fluorescent molecules. By integration of this efficient injection process with our RT-PCR microsystem10 and our single-cell capture techniques,43 it should be possible to perform quantitative genotyping and gene expression profiling from individual cells. These enhanced capabilities illustrate the power of specific capture and direct injection as an alternative to the traditional cross injector. ACKNOWLEDGMENT The first two authors contributed equally to this work. We thank Eric Chu for assistance with microfabrication and Stephanie Yeung, Erik Douglas, Palani Kumaresean, Nate Beyor, and Tom Cheisl for valuable discussions. N.M.T. and C.N.L. were supported by a NIH Molecular Biophysics Training Grant (Grant T32GM08295) and Chevron-Texaco Graduate Fellowship, respectively. Microfabrication was performed in the UC Berkeley Microfabrication Laboratory. This work was supported by NIH Grants R01HG01399 and HG003329 and by the Chemical Sciences Division of the U.S. Department of Energy under Contract DE-AC03-76SF00098. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review June 13, 2007. Accepted August 27, 2007. AC0712547 (43) Toriello, N. M.; Douglas, E. S.; Mathies, R. A. Anal. Chem. 2005, 77, 69356941.