Ultrasensitive PCR and Real-Time Detection from Human

Nov 3, 2007 - ... Gaithersburg, Maryland 20899, and The Manchester Interdisciplinary Biocentre, University of Manchester, 131, Princess Street, Manche...
0 downloads 0 Views 177KB Size
Download PDF
Anal. Chem. 2007, 79, 9185-9190

Ultrasensitive PCR and Real-Time Detection from Human Genomic Samples Using a Bidirectional Flow Microreactor Lin Chen,† Jonathan West,† Pierre-Alain Auroux,‡ Andreas Manz,† and Philip J. R. Day*,†,§

Institute for Analytical Sciences, Bunsen-Kirchhoff Strasse 11, D-44139 Dortmund, Germany, National Institute of Standards and Technology, Electronics and Engineering Laboratory, 100 Bureau Drive, MS 8120, Building 225, Gaithersburg, Maryland 20899, and The Manchester Interdisciplinary Biocentre, University of Manchester, 131, Princess Street, Manchester. M1 7ND, U.K.

In this paper we present a reliable bidirectional flow DNA amplification microreactor for processing real-world genomic samples. This system shares the low-power thermal responsiveness of a continuous flow reactor with the low surface area to volume ratio character of stationary reactors for reducing surface inhibitory effects. Silanization with dimethyldichlorosilane in combination with dynamic surface passivation was used to enhance PCR compatibility and enable efficient amplification. For realtime fragment amplification monitoring we have implemented an epimodal fluorescent detection capability. The passivated bidirectional flow system was ultrasensitive, achieving an RNase P gene detection limit of 24 human genome copies with a reaction efficiency of 77%. This starts to rival the performance of a conventional real-time PCR instrument with a reaction efficiency of 93% and revitalizes flow-through PCR as a viable component of lab on a chip DNA analysis formats. The miniaturization of analytical instrumentation has been progressing rapidly since the pioneering demonstration of gas chromatography in a spiral silicon microchannel by Terry et al.1 Today numerous conventional analytical methods have been scaled down to fit a microchip format and avail of the benefits of miniaturization, including fundamental performance gains and economy. In this presently termed postgenomic era, there is a pressing need for high-throughput and quantitative analytical methods to address a vast number of biological applications. In response there has been enormous activity in the miniaturization of bioanalytical methods. Among the array of bioanalytical methods the polymerase chain reaction (PCR) stands out as the true champion. PCR provides a method for the selective and exponential amplification of DNA fragments from genetic samples. Furthermore, the use of small template quantities combined with the sensitivity of PCR * To whom correspondence should be addressed. E-mail: [email protected]. Fax: +49-231-1392-120. Phone: +49-231-1392-109. † Institute for Analytical Sciences. ‡ National Institute of Standards and Technology. § University of Manchester. (1) Terry, S. C.; Jerman, J. H.; Angell, J. B. IEEE Trans. Electron Devices 1979, 26, 1880-1886. 10.1021/ac701668k CCC: $37.00 Published on Web 11/03/2007

© 2007 American Chemical Society

uniquely enables the quantitative assessment of nucleic acids.2 Since its invention in 1985 by Saiki et al.,3 PCR has already revolutionized the life sciences and has become a ubiquitous genetic sample preparation and analysis tool. The rapid thermocycling requirement has made polymerase-mediated biochemistry a prime candidate for miniaturization.4 For a more in-depth appreciation of the development of different PCR microreactor systems some excellent reviews are available.5,6 Primarily the onus has been on the reduction of parasitic heat capacities to provide a small thermal mass for near instantaneous heating and cooling. For example, silicon cantilevers were used to suspend and nearly perfectly isolate 100 nL thermocycling microreactors for ultrafast heating (175 °C s-1) and cooling (-125 °C s-1).7 Alternatively noncontact infrared (IR)-mediated PCR can be used to directly heat the reagents for rapid thermal cycling.8,9 Taking a radically different approach, Kopp et al. reduced the thermal mass to that of the sample alone by using a miniature continuous flow format. Here a microchannel passes repetitively through spatially defined temperature zones such that the microchannel geometry and flow rate determines the temperature profile that the reactants experience. At miniature dimensions, heat transfer times become minimal and complete run times within minutes can be attained.10,11 Miniature and rapid thermocycler systems ideally lend themselves to point-of-care applications that are beyond the reach of the traditional diagnostic or research laboratory. In these scenarios the integration of sample preparation and downstream detection (2) Day, P. J. R. Expert Rev. Mol. Diagn. 2006, 6, 23-28. (3) Saiki, R. K.; Scharf, S.; Faloona, F.; Mullis, K. B.; Horn, G. T.; Erlich, H. A.; Arnheim, N. Science 1985, 230, 1350-1354. (4) de Mello, A. J. Lab Chip 2001, 1, 24N-29N. (5) Auroux, P.-A.; Koc, Y.; de Mello, A.; Manz, A.; Day, P. J. R. Lab Chip 2004, 4, 534-546. (6) Roper, M. G.; Easley, C. J.; Landers, J. P. Anal. Chem. 2005, 77, 38873894. (7) Neuzil, P.; Zhang, C.; Pipper, J.; Oh, S.; Zhuo, L. Nucleic Acids Res. 2006, 34, e77. (8) Oda, R. P.; Strausbauch, M. A.; Huhmer, A. F. R.; Borson, N.; Jurrens, S. R.; Craighead, J.; Wettstein, P. J.; Eckloff, B.; Kline, B.; Landers, J. P. Anal. Chem. 1998, 70, 4361-4368. (9) Giordano, B. C.; Ferrance, J.; Swedberg, S.; Hu ¨ hmer, A. F. R.; Landers, J. P. Anal. Biochem. 2001, 291, 124-132. (10) Kopp, M. U.; de Mello, A. J.; Manz, A. Science 1998, 280, 1046-1048. (11) Hashimoto, M.; Chen, P.-C.; Mitchell, M. W.; Nikitopoulos, D. E.; Soper, S. A.; Murphy, M. C. Lab Chip 2004, 4, 638-645.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007 9185

functionalities for robust automated operation is critical.12 This is the central dogma of the micro total analysis system (µTAS) concept,13 and accordingly, PCR microreactors have been combined with capillary electrophoresis for fragment sizing14-16 and more recently with real-time fluorescent monitoring17 or microarrays for multiplexed sequence determination.18 Sample preparation functions such as cell isolation using microfilters19 or immunomagnetic beads,20 thermal cell lysis,21 and solid-phase DNA extraction22 have also been coupled to thermocycler microreactors. Already fully integrated systems have been developed that can analyze challenging samples.23 This is exemplified by Easley et al. who have developed a system for the detection of Bacillus anthracis from whole blood and Bordetella pertussis from nasal aspirates.24 Sensitivity is arguably the most important performance criteria. During the development of most PCR microreactors artificial DNA templates have been used to optimize thermocycling and PCR compatibility. These templates are commonly amplification fragments produced by conventional thermocycling or fragments contained within plasmids for convenience of replication. This approach has been used to show stochastic single-molecule detection in 280 nL reaction volumes.25 However, real-world genomic samples are significantly more complex, with bacterial genomes of millions of base pairs and the human genome some 3 orders of magnitude larger (3.3 × 109 bp). Primer target interactions amidst large genomes are initially infrequent, and reaction sensitivity becomes challenged.26 Nevertheless, Belgrader et al. have demonstrated B. anthracis detection from a single genome copy,27 and Matsubara et al. achieved stochastic amplification of 0.4 human genome copies per reaction microchamber.28 To date, flow-through PCR microsystems have not achieved high levels of selectivity and sensitivity. In this paper we present a linear bidirectional shuttle PCR system, with a favorable surface (12) Chen, L.; Manz, A.; Day, P. J. R. Total nucleic acids analysis integrated on microfluidic devices. Lab Chip 2007, 7, 1413-1423. (13) Manz, A.; Graber, H.; Widmer, H. M. Sens. Actuators, B 1990, 1, 244-248. (14) Burns, M. A.; Johnson, B. N.; Brahmasandra, S. N.; Hanique, 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. (15) Lagally, E. T.; Emrich, C. A.; Mathies, R. A. Lab Chip 2001, 1, 102-107. (16) Woolley, A. T.; Hadley, D.; Landre, P.; de Mello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086. (17) Northrup, M. A.; Benett, B.; Hadley, D.; Landre, P.; Lehew, S.; Richards, J.; Stratton, P. Anal. Chem. 1998, 70, 918-922. (18) Liu, Y,; Rauch, C. B.; Stevens, R. L.; Lenigk, R.; Yang, J.; Rhine, D. B.; Grodzinski, P. Anal. Chem. 2002, 74, 3063-3070. (19) Wilding, P.; Kricka, L. J.; Cheng, J.; Hvichia, G.; Shoffner, M. A.; Fortina, P. Anal. Biochem. 1998, 257, 95-100. (20) Lien, K.-Y.; Lee, W.-C.; Lei, H.-Y.; Lee, G.-B. Biosens. Bioelectron. 2007, 22, 1739-1748. (21) Waters, L. C.; Jacobson, S. C.; Kroutchinina, N.; Khandurina, J.; Foote, R. S.; Ramsey, J. M. Anal. Chem. 1998, 70, 158-162. (22) Cady, N. C.; Stelick, S.; Kunnavakkam, M. V.; Batt, C. A. Sens. Actuators, B 2005, 107, 332-341. (23) Liu, R. H.; Yang, J.; Lenigk, R.; Bonanno, J.; Grodzinski, P. Anal. Chem. 2004, 76, 1824-1831. (24) 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. (25) Lagally, E. T.; Medintz, I.; Mathies, R. A. Anal. Chem. 2001, 73, 565-570. (26) Garner, T. W. Genome 2002, 45, 212-215. (27) Belgrader, P.; Elkin, C. J.; Brown, S. B.; Nasarabadi, S. N.; Langlois, R. G.; Milanovich, F. P.; Colston, B. W., Jr. Anal. Chem. 2003, 75, 3446-3450. (28) Matsubara, Y.; Kerman, K.; Kobayashi, M.; Yamamura, S.; Morita, Y.; Takamura, Y.; Tamiya, E. Anal. Chem. 2004, 76, 6434-6439.

9186

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

area to volume ratio for the ultrasensitive detection of DNA targets from typical human genomic samples. DESIGN RATIONALE The continuous flow PCR format offers excellent heat exchange time constants and ease of fabrication (a single photolithographic step). However, the continuous flow approach can be criticized. Notably, there is no dwell time and no cycle number versatility of these preprogrammed systems. In addition, an enormous channel length is required (1-2 m) and results in a sizable device footprint (e.g., 50 mm × 40 mm) even when folded into a serpentine9 or spiral10 arrangement. This impacts fabrication economy and is not suited to either parallel sample processing or a simple real-time detection scheme. To address these drawbacks several reconfigurations of the continuous flow format have been developed. For example, cycle number flexibility, ease of real-time detection implementation, and significant footprint reductions can be achieved using an annular microchannel or microflow with literal reagent cycling.29-32 However, like the continuous flow format this approach does not overcome Taylor dispersion produced by the flows parabolic velocity profile, and this results in marked temperature dwell time variability. To combat this problem a plug flow can be employed and has been used within bidirectional flow PCR systems. This format like continuous flow reactors offers the benefit of rapid heat exchange, with the small footprint and moderate surface area to volume ratio of static microreactors. These bidirectional flow systems have been used for complete thermocycling reactions in minutes.33-35 These reaction times compare favorably with those of the static microcavity reactors. To date, however, these systems have only been used to demonstrate limits of detection of 100 000’s of copies of artificial template molecules. In this paper we report the use of a bidirectional flow system with a low surface area to volume ratio to alleviate surface-induced reaction inhibition. EXPERIMENTAL SECTION Device Assembly and Operation. To define the linear fluidic channel a standard 1.0 mm i.d. glass capillary with a wall thickness of ∼100 µm was used and cast within a sheet of poly(dimethylsiloxane) (PDMS, Sylgard 184, Dow Corning). Critically, the capillary was positioned at one face, with the PDMS sheet used to enable conformal contact with the heater elements. Two 12 mm × 18 mm footprint Peltier heating elements (QC-17-1.0-3.9M, Quick-Ohm-Ku¨pper & Co. GmbH, Germany) were used and contacted the device via 2.5 mm thick aluminum heat sinks. A 4 mm length PDMS section between the two temperate zones was removed using a scalpel to reduce thermal cross-talk. For feedback temperature control, sensors (AD 590, Analog Devices, MA) were mounted on each aluminum block and used (29) West, J.; Karamata, B.; Lillis, B.; Gleeson, J. P.; Alderman, J.; Collins, J. K.; Lane, W.; Mathewson, A.; Berney, H. Lab Chip 2002, 2, 224-230. (30) Liu, J.; Enzelberger, M.; Quake, S. Electrophoresis 2002, 23, 1531-1536. (31) Krishnan, M.; Ugaz, V. M.; Burns, M. A. Science 2002, 298, 793. (32) Chen, J.; Wabuyele, M.; Chen, H.; Patterson, D.; Hupert, M.; Shadpour, H.; Nikitopoulos, D.; Soper, S. A. Anal. Chem. 2005, 77, 658-666. (33) Chiou, J.; Matsudaira, P.; Sonin, A.; Ehrlich, D. Anal. Chem. 2001, 73, 20182021. (34) Auroux, P.-A.; Day, P. J. R.; Manz, A. NSTIsNanotech 2004, 1, 67-69. (35) Frey, O.; Bonneick, S.; Hierlemann, A.; Lichtenberg, J. Biomed. Microdevices 2007, 9, 711-718.

Figure 1. Schematic illustration of the bidirectional flow thermocycling system.

with a PID. The capillary device and heaters were assembled within a poly(methyl methacrylate) (PMMA) housing to ensure good positional contact. The system and housing was mounted on a home-made x-y stage and adapted to the Caliper 42 microfluidics workstation (Caliper Europe GmbH, Germany) that is equipped with a mercury arc lamp and filter cubes for epifluorescence. For detection of the FAM reporter dye, a 490 nm filter was used for excitation and a 520 nm filter for collection of emitted light. A schematic of the bidirectional flow PCR system is shown in Figure 1. For temperature calibration within the capillary a K-type thermocouple (Pt30, TC Ltd., U.K.) was inserted and temperatures were recorded (TC-08, Pico Technology Ltd., U.K.). A 2 µL PCR volume plug was flanked by mineral oil to avoid droplet break up and evaporation during thermocycling. The reaction plug was transported back and forth using a Kloehn syringe pump (V6, Kloehn Ltd., Las Vegas, NV) connected via Teflon tubing (390 µm i.d.) and controlled using a custom Labview program. Surface Modification. To reduce reagent adsorption and reaction inhibition the glass capillary was silanized with dimethyldichlorosilane (DMDCS) according to the protocol described by Obeid et al.36 Briefly, the glass capillary was rinsed with deionized H2O, washed with acetone, and dried with a N2 gas stream. The glass capillary was filled with 5% (v/v) DMDCS (Fluka) in chloroform and placed in a desiccator for brief vacuum boiling, with the silanization mixture removed by evaporation. Finally, the glass capillary was cleaned by employing a series of chloroform, acetone, and deionized H2O washes with drying again with a N2 gas stream. For resilanization, additional treatment with 10% (w/v) NaOH for 30 min was required to completely remove the previous silane monolayer coating.37 PCR Chemistry. To evaluate biocompatibility a PCR chemistry was designed using Primer Express (Applied Biosysems, Foster City, CA) with forward (5′-TGGTTGTGAAGACGTCGTTGA-3′) and reverse primers (5′-TGCATATCCTCGCTCTCCAGA3′) for the amplification of a 66 bp RNase P mimic template oligo possessing uracyl moieties for endonuclease digestion if required (5′-TGGUTGUGAAGACGTCGUUGAACAACCCATACAUCATCCGCUGGAGCGCTGGAGAGCGGGAATGCA-3′). All (36) Obeid, P. J.; Christopoulos, T. K.; Crabtree, H. J.; Backhouse, C. J. Anal. Chem. 2003, 75, 288-295. (37) Prakash, R.; Kaler, K. V. I. S. Microfluid. Nanofluid. 2007, 3, 177-187.

oligodeoxyribonucleotides were purchased from Sigma Proligo (Paris, France). The 4 µL PCR mixture consisted of 1× TaqMan master mix (Applied Biosystems, Foster City, CA) containing buffer, dNTPs, modified Taq DNA polymerase, 3.5 mM MgCl2, 330 nM of each primer, and 0.1 ng/µL DNA template. In addition to silanization and to further protect the reaction from surface inhibitory effects a dynamic passivation method was used for the sacrificial surface adsorption of chemical additives comprising 0.1% (v/v) Tween 20, 0.01% (w/v) poly(vinylpyrrolidone) (PVP, MW 10 000), and 0.2 µg/µL bovine serum albumin (BSA). The thermocycling program consisted of 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 60 s at 60 °C. Amplicons generated using bidirectional flow and the conventional PCR instrument were both separated and detected using the Experion automated electrophoresis system (Bio-Rad Laboratories GmbH, Munich, Germany). Real-Time Bidirectional Flow PCR. Following the biocompatibility study, some real-world human genomic DNA samples were used as the template. Here, the complexity and size (3.3 × 109 bp) of the human genome presents a challenge for reaction selectivity and thus amplification efficiency. The single-copy RNase P PCR chemistry consisted of 1× TaqMan Master Mix with 1× RNase P primer-probe (FAM) mix (PN: 4316831, Applied Biosystems, Foster City, CA) for the amplification of a larger, 87 bp, fragment. Additionally, and as before, 0.1% (v/v) Tween 20, 0.01% (w/v) PVP, and 0.2 µg/µL BSA were added to further protect the reaction from surface inhibitory effects. Ten-fold serial dilutions of human genomic DNA ranging from 24 000 to 2.4 copies (80 ng to 8 pg) were used as the samples, to evaluate the real-time detection capability and determine the limit of detection. The bidirectional flow PCR experiments used a 2 µL volume of the reaction cocktail. The thermocycling program for bidirectional PCR consisted of a 10 min hot start at 95 °C, followed by 40 cycles of 15 s at 95 °C for denaturation and 30 s at 60 °C for primer annealing and extension. For real-time detection, excitation light (490 nm) was focused on the capillary between the two temperature zones. During each cycle a shutter system was used to record the fluorescence signal of the PCR plug only as it was transported from the 60 to the 95 °C zone (i.e., the coolest stage for improved signal-to-noise). The performance of the bidirectional flow PCR device was evaluated by gel electrophoresis (as described above) and by direct data comparison with a conventional real-time PCR instrument (ABI 7900HT). With the conventional instrument all thermocycling conditions were as above with the exception of a 60 s dwell time at 60 °C instead of 30 s. Reactions within the conventional instrument involved a larger volume, and here a template quantification curve with decade intervals ranging from 60 000 to 6 copies was used. RESULTS AND DISCUSSION Bidirectional Thermocycling. A thermocouple was inserted into the capillary to monitor the temperature offset, relative to the Peltier element, and was used to determine the Peltier element temperatures required to attain the desired 60 and 95 °C reaction temperature zones. The Peltier elements contact the device via aluminum heat sinks that aid temperature stability. However, these do not represent an infinite thermal mass, and the arrival of the liquid caused transient temperature perturbations. The extent of Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9187

Figure 2. Temperature perturbations caused by the transport of a 5 µL water plug (to simulate the total volume of the sample plug and mineral oil) moving from 95 to 60 °C (lower) and from 60 to 95 °C (upper) for flow velocities ranging from 16.0 to 1.6 mm s-1.

perturbation was influenced by transport velocity and is shown in Figure 2 for samples arriving at the 60 °C zone (lower) and the 95 °C zone (upper) for velocities ranging from 16.0 to 1.6 mm s-1. High velocities do not enable sufficient heat exchange during transport between temperature zones and result in g10 °C temperature corruption requiring ∼10 s before desirable reaction temperatures are attained. Operating at the lower 1.6 mm s-1 velocity does not cause significant temperature perturbation. This produced a zone-to-zone transport time of 10 s, with a modest mean thermal ramping rate of 3.5 °C s-1. In the remainder of the study we opted to undertake the bidirectional flow PCR experiments using a 1.6 mm s-1 mean velocity for known dwell times at the given temperatures. This velocity resulted in a cycle duration of 65 s, or 43 min for the entire 40 cycle chemistry. Enhanced Biocompatibility. Glass is typically chemically inert. However, with the high surface area to volume ratios of both static and flow-through miniaturized PCR systems38 glass is well-known to cause reaction fouling by sequestering Taq polymerase and other reaction components. With static mode reactors this problem can be considerably ameliorated by coating the chamber surfaces with hydrophobic silanes, typically methylated chlorosilanes.39 The silanization of continuous flow microchannels only partially promotes PCR effectiveness, with moderate amplicon yields from high template inputs.36 In these systems the complete surface area the reagents are exposed to throughout the reaction is enormous relative to the volume. The surface area is greatly reduced with the bidirectional shunting method. However, silanizing the capillary with DMDCS to present a hydrophobic surface with a 1 µL sessile water droplet contact angle of ∼105° was insufficient to prevent reaction inhibition. Chemical additives are commonly used to reduce reaction inhibition and improve reaction efficiency. Protein adsorption on silica surfaces can be inhibited using PVP.40,41 The use of high molecular weight PVP at a concentration of 0.4% has been used (38) Kricka, L. (39) Felbel, J.; 333-338. (40) Robinson, (41) Giordano, 334-340.

J.; Wilding, P. Anal. Bioanal. Chem. 2003, 377, 820-825. Bieber, I.; Pipper, J.; Ko ¨hler, J. M. Chem. Eng. J. 2004, 101, S.; Williams, P. A. Langmuir 2002, 18, 8743-8748. B. C.; Copeland, E. R.; Landers, J. P. Electrophoresis 2001, 22,

9188 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

Figure 3. Electropherogram of RNase P amplification products using 8 ng of human genomic DNA as the reaction template: DNA ladder (L), positive control using a conventional instrument (1), no template control (NTC) (2), and the product from the bidirectional flow PCR system (3).

for the dynamic (i.e., during the reaction) passivation of surfaces to enable efficient PCR in a PDMS microreactor.42 However, higher concentrations resulted in reaction inhibition. In this study, 0.01% PVP was used to obtain modest reaction efficiency in a silanized capillary. Both passive and dynamic passivation with BSA are popular choices to render surfaces PCR compatible.36,43 By the further addition of 0.2 µg µL-1 BSA as well as 0.1% (v/v) Tween 20, the on gel concentration of the RNase P mimic PCR amplicon was 84% in comparison to that produced by the conventional thermocycling instrument. Use of these additives in isolation was not effective, indicating the requirement for using the additives in concert. Successive reactions fail as the silane monolayer is disrupted by thermocycling temperatures in the presence of the reaction buffer, and resilanization is required prior to subsequent runs.37,39 These reports are corroborated by our findings which are documented in Figure S-1 of the Supporting Information. Here amplicons produced using a conventional thermocycler are compared with those produced by bidirectional flow PCR with a fresh silane layer and those without resilanization following a prior PCR run. Amplification from Human Genomic Samples with RealTime Detection. The achievement of enhanced PCR compatibility permitted the evaluation of the bidirectional flow system to amplify the RNase P target from real-world human genomic template inputs. Initially an 8 ng genomic sample was evaluated with a 40 cycle reaction protocol. The amplification products were compared with those from a conventional instrument by gel electrophoresis. As shown in Figure 3 both systems amplified the 87 bp target fragment without cross-reactivity with other potential primer targets. The linear bidirectional flow system is ideally suited for online reaction monitoring with a single excitation source and detector. By interfacing the system with an epimodal fluorescence system (Caliper 42 microfluidics workstation) using a custom assembly real-time product monitoring was undertaken, again using an 8 ng human genomic input. The raw data comprises a fluorescent signal peak each cycle, corresponding to the passage of the reaction plug across the intertemperature zone after leaving the annealing/extension zone. With the use of this method (42) Xia, Y.-M.; Hua, Z.-S.; Srivannavit, O.; Ozel, A. B.; Gulari, E. J. Chem. Technol. Biotechnol. 2007, 82, 33-38. (43) Koh, C. G.; Tan, W.; Zhao, M.-Q.; Ricco, A. J.; Fan, Z. H. Anal. Chem. 2003, 75, 4591-4598.

Figure 4. Real-time PCR for detection of RNase P using human genomic DNA template inputs using the bidirectional flow device. Curves 1-5 are for genome copy number inputs of 24 000, 2400, 240, 24, and the NTC, respectively.

the signal from the FAM-derivatized reporter probes, and hence the amplicon yield, was monitored during each cycle throughout the entire reaction protocol. Following normalization to remove signal noise, a classic PCR amplification curve results. This is documented in Figure S-2 of the Supporting Information. A standard curve comprising 10-fold dilutions of human genomic DNA spanning 80 ng to 8 pg, or 24 000 to 2.4 genome copies, was subsequently used to assess RNase P (chosen because of the absence of pseudogenes) amplification in the bidirectional flow device. Amplification curve for 24 000 to 24 genome copy inputs are documented with the curve from a no template control in Figure 4. The bidirectional flow system demonstrated a 24 genome copy limit of detection, with the 2.4 genome copy reaction failing to produce a detectable signal after 40 cycles. In contrast, the conventional instrument demonstrated a 6 genome copy limit of detection, albeit with an annealing/extension dwell time of 60 s and not 30 s as with the described microreactor. To the best of our knowledge this is the most sensitive flow-through PCR system reported in the literature. We attribute this sensitivity to the combination of static and dynamic passivation in addition to the low surface area to volume ratio (∼25 mm2/µL) of the system. For comparison, the seminal continuous flow microreactor had a surface area to volume ratio of ∼60 mm2/µL (with the use of a large 10 µL reaction volume) and was able to detect 108 template copies.10 More recently, a 57 mm2/µL system was able to amplify, with the aid of static and dynamic passivation chemistries, a target fragment from a sample containing 106 human genomic copies.36 Cycle Threshold Analysis. For each of the amplification curves, a baseline and a line for the exponential stage were drawn using a linear fit function (OriginPro 7.0). The Ct value of each curve was defined as the intercept of these lines. The Ct values obtained by both the bidirectional flow PCR system and the conventional instrument for the different template inputs are plotted in Figure 5. This real-time detection capability can be used, in parallel, for target quantification. For example, target quantification can find application for point-of-care viral diagnostics where the viral titer impacts patient care management.44 The concentration of DNA molecules at the signal-to-noise threshold can be approximated by the following: (44) Kaigala, G. V.; Huskins, R. J.; Preiksaitis, J.; Pang, X.-L.; Pilarski, L. M.; Backhouse, C. J. Electrophoresis 2006, 27, 3753-3763.

Figure 5. Ct value vs the logarithm of the input genome copies for the conventional real-time PCR instrument (ABI 7900HT) and the bidirectional flow PCR system.

CT ) C0(1 + e)Ct

(1)

where CT and C0 are the concentration of PCR products after t cycles and the initial concentration, respectively, e is the reaction efficiency, and Ct is the cycle threshold. The Ct value can be directly expressed:45

Ct ) -[log(1 + e)]-1(log CT + log C0)

(2)

The Ct value is therefore proportional to the logarithm of the initial DNA template concentration. Linear regression of the Ct versus log C0 yields a slope m of -3.51 for the conventional instrument and -4.04 for the bidirectional flow system. These values can be used to calculate the PCR efficiency:45

e ) 10-1/m - 1

(3)

From this the reaction efficiency was calculated to be 93% for the conventional instrument and 77% for the bidirectional flow PCR system, or 83% as efficient compared to the conventional instrument. The reduced efficiency could be attributed to a number of factors, including the reduction of the annealing/extension time by 30 s (50%) and imperfect surface passivation measures. Miniaturization Potential. The present system has a modest cycle duration of 65 s, with 20 s for transportation between the temperature zones, producing a 43 min, 40 cycle total run time. The thermal mass of the plug and the cylindrical reaction cavity both retard heat exchange. Miniaturization of the system using standard glass microfabrication techniques could be used to define a 100 nL reaction plug, again 4 mm in length, with a depth of 100 µm and a width of 250 µm. With a reduced thermal mass, the shallow microchannel and the relatively large heater element contacting face (1 mm2) such a system should achieve rapid heat transfer, enabling faster reagent transport and shorter thermocycling times. However, improved or more advanced passivation chemistries will be required to tolerate the surface area to volume ratio gain. (45) Wang, Z.; Sekulovic, A.; Kutter, J. P.; Bang, D. D.; Wolff, A. Electrophoresis 2006, 27, 5051-5058.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9189

To enable complete thermocycling flexibility the linear system can readily be reconfigured with a third heater element for standard three-temperature phase reactions. Moreover, the inherent fluidic transport capability of the described bidirectional flow PCR amplification system is also suited for integration with additional sample preparation functions such as DNA extraction and reverse transcription. Another key aspect of miniaturization is the ability to undertake reactions in parallel. This is also important for the coamplification of a standard curve for comparative target copy quantification. Many laboratory-based applications require high-throughput capabilities. For example, multilayer soft lithography was used for 72 RT-PCRs,46 which was shortly followed by a massively parallel (14 112 simultaneous reactions) format for the multigene analysis of environmental bacteria.47 As demonstrated by Frey et al. with a 10-channel system,35 the linear channel of the bidirectional reactor can also readily be parallelized. For laboratory-based implementation the key driver will be to standardize interfaces for compatibility with existing fluid handling and optical detection formats. Coincidentally, a standard microtiter plate can accommodate 2 rows of 96 of the 1.2 mm o.d. capillary reactors described in this paper.

CONCLUSIONS In this paper we report a significant step in the development of bidirectional flow PCR systems. The presented device combines the advantages of both continuous flow and stationary PCR, offering rapid heat exchange for fast temperature ramping and a sufficiently small surface area to volume ratio to reduce surfaceassociated reaction inhibition. Hydrophobic surface modification in combination with the addition of a cocktail of dynamic passivation agents was required to render the surface environment PCR compatible for efficient target fragment amplification. The system was used for the amplification and real-time detection of a single-copy target gene from 24 human genomes, or 12 cells. The current device is ripe for miniaturization for ultrafast processing times, integration, and parallelization for quantitative nucleic acids analysis.

(46) Marcus, J. S.; Anderson, W. F.; Quake, S. R. Anal. Chem. 2006, 78, 956958. (47) Ottesen, E. A.; Hong, J. W.; Quake, S. R.; Leadbetter, J. R. Science 2006, 314, 1464-1467.

Received for review August 6, 2007. Accepted September 28, 2007.

9190

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

ACKNOWLEDGMENT The authors are grateful for invaluable discussions with Ingo Feldmann and Dr. Norbert Jakubowski and to Norman Ahlmann for construction of the Peltier element assembly. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

AC701668K