Rapid distribution of a liquid column into a matrix of nanoliter wells for parallel real-time quantitative PCR

Sensors and Actuators B 135 (2009) 671–677

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb

Rapid distribution of a liquid column into a matrix of nanoliter wells for parallel real-time quantitative PCR Hao-Bing Liu a,b , Naveen Ramalingam a,b , Yu Jiang a,b , Chang-Chun Dai a,b , Kam Man Hui b , Hai-Qing Gong a,∗ a b

BioMEMS Laboratory, School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Division of Cellular and Molecular Research, National Cancer Centre, Singapore 169610, Singapore

a r t i c l e

i n f o

Article history: Received 5 May 2008 Received in revised form 29 August 2008 Accepted 11 October 2008 Available online 6 November 2008 Keywords: Bio-chip Parallel PCR Nanoliter wells Microfluidics Vacuum

a b s t r a c t A simple and reliable microfluidic method to load and seal a two-dimensional matrix of nanoliter wells, preloaded with primer pairs, for performing PCR without well-to-well cross-contamination was developed. With a vacuum aided microfluidics the distribution of sample mixture into 100 wells was achieved within 0.2 s, and all the wells were fluidically isolated within 0.5 s. The wells were subsequently sealed with poly-(dimethylsiloxane) (PDMS) prepolymer to prevent evaporation of PCR mixture and primer cross-contamination. The performance of the PDMS matrix chip was successfully tested by simultaneous (parallel) real-time quantitative PCR amplification of 10 different gene targets against cDNA template from human hepatocellular carcinoma, in 120 nl volume using EvaGreen fluorescent dye. Negligible crosscontamination between primers in different wells was confirmed with the PCR result, and by investigating the diffusion of the primers. The microfluidic channel design, chip structure, vacuum level, and operation schedule were optimized to achieve successful vacuum aided microfluidics for PCR mixture loading without air trapping. The microfluidics and the well matrix chip platform are suitable for both low- and high-density well applications including PCR related applications as well as immunoassays and cell-based assays. © 2008 Elsevier B.V. All rights reserved.

1. Introduction The completion of Human Genome Project triggered rapid development of high-throughput platform for parallel genomic assays. Currently, two types of DNA microarrays are widely used as a platform for highly parallel genomic assays: microarrays for genome-wide expression profiling and microarrays for single nucleotide polymorphism (SNP) detection and genotyping. The validation of microarray results remains a desirable step due to non-standard methods of data analysis and interpretation. Quantitative reverse transcription PCR (qRT-PCR) is often used as an alternative technology to validate gene expression profiling studies. Moreover, the microarray technology has poor sensitivity and the multiple steps involved in this technology introduce variability in the results [1] when compared with real-time PCR. For SNP detection and genotyping, generally PCR process is used to amplify the samples before it is analyzed by melting curve technique or by hybridization to oligonucleotide probes. For gene expression profiling, it is desirable to integrate the quantitative, sensitive capabilities

∗ Corresponding author. Tel.: +65 6790 4279; fax: +65 6790 4756. E-mail address: [email protected] (H.-Q. Gong). 0925-4005/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2008.10.028

of real-time PCR with high-throughput capability of a microarray as a single platform for highly parallel genomic assays. In addition to the microarray technology for parallel analyses of multiple gene targets, a number of platforms have been developed to integrate PCR with microarray techniques [2–4]. Most of these platforms combine solid-phase PCR with microarray platform. However, studies have shown that solid-phase PCR is less efficient than solution-phase PCR [5]. Immense efforts have been made to develop PCR-based biochips as a high-throughput platform, for analyses of thousands of genes in short time. A few researchers have developed a microchamber array for PCR on silicon substrate [6,7], on silicon–glass hybrid chip [8], and on poly-(dimethylsiloxane) (PDMS) substrate [9]. Matsubara et al. developed a microchamber array in which the DNA sample was loaded by using a nanoliter dispensing robotic system and the PCR data was analyzed using a DNA microarray scanner [7]. Nagai et al. developed a microchamber array on silicon substrate for picoliter PCR using manual sample loading step, and the amplified product was characterized by comparing the fluorescence intensity at the beginning and at the end of the PCR process [8]. Quake and coworkers demonstrated the microfluidics for distribution of 2 ␮l PCR mixture among 400 independent reactors in a PDMS chip, using 2860 integrated hydraulic valves and pneumatic pumps [9].

672

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677

The fabrication and thermal characterization of multi-chamber for PCR applications were reported by a few researchers [10,11]. Iordanov et al. reported an array of nanoliter PCR chambers implemented in silicon with a glass cover, comprising integrated thermal and integrated optical detectors [10]. The integrated heater and sensor increases the cost of the chip, affects the temperature distribution on the chip, and hinders the development of a disposable PCR chip. Chaudhari et al. developed a microfabricated 18-vessel array for PCR by sealing the microchambers on the silicon to a glass cover using a PCR-compatible acrylic-based adhesive tape [11]. The temperature fields in this microfabricated PCR vessel array were measured by thermochromic liquid crystals (TLCs). Leamon et al. reported a novel PicoTiterPlateTM platform in which three hundred thousand solid-phase asymmetric PCR with a reaction volume of 39.5 pL was simultaneously performed [12]. Recently, Brenan and co-workers reported a rectilinear array of three thousand and seventy two through-holes (each with a volume of 33 nl) in a stainless steel platen, in which the primer pairs were immobilized in a matrix and the array of through-holes were sealed using UV curable epoxy [13]. In the past, few researchers have used microfluidics for performing PCR on-chip [9,14], among them Mathies and co-workers [14] demonstrated multiple PCR on-chip with a microfluidic distribution of PCR mixture, using a mechanical valve array for sample loading and well sealing. There are two types of high-throughput PCR chips based on their configuration. The first type consists of an array or matrix of fluidic micro chambers [13,15], with typically an inlet channel for sample loading and an outlet channel for air-venting. Such a configuration employs microfluidic channel networks for loading liquid sample into the array of chambers, which significantly simplifies the sample loading process for high-throughput PCR chip. However, it is complicated to isolate and seal all the chambers, because the inlet and outlet channel of all the microchambers has to be sealed to prevent PCR mixture from evaporating during thermocycling and to prevent cross-contamination of primer pairs. In addition, the channels in such configuration, occupies space on the chip and limit the density of chambers on the chip. The second type of highthroughput PCR chip is basically a miniaturized well-plate [16]. In this type of configuration, it is possible to reach high density of microwells. The microwells can be easily isolated and sealed using a thin cover or sealant on top of the wells. However, most of these devices either requires manual loading of PCR sample into individual reaction wells in a PCR chip, similar to that used in a well-plate operation, or requires an expensive liquid dispensing robot. A simple and economic sample loading and reaction well sealing method is desired to reduce the operating cost. In this paper, we present a novel vacuum aided microfluidics for liquid distribution into a matrix of nanoliter wells and isolation of the wells. The PDMS matrix chip comprising of 100 microwells is pre-loaded with primer pairs for simultaneous (parallel) real-time quantitative PCR analyses of multiple genes (Fig. 1) against cDNA template. Concerns about cross-contamination of PCR primer pairs among the wells during the PCR mixture loading step have been addressed. 2. Experimental 2.1. Chip design and fabrication The prototype chip contains one hundred wells for realtime quantitative PCR, and each well has a dimension of 0.5 mm × 0.5 mm × 0.5 mm that hold sample liquid of about 120 nl. The chip also contains one inlet channel, three outlet/air-venting channels and a headspace on top of the wells for microfluidic operation. The chip has a total dimension of 22 mm × 22 mm × 1.2 mm

Fig. 1. Optical micrographs of a PDMS matrix chip comprising 100 wells of 120 nl volume. The channels and wells are fabricated on a PDMS layer, and bonded to a glass substrate and a glass cover to form an enclosed headspace above the wells for microfluidic handling of PCR mixture and sealant.

and consists of three layers: a glass substrate layer, a glass cover layer, and a PDMS structure layer sandwiched in between. The PDMS layer is replicated from a stainless steel die, which was fabricated by an electrical discharge machining (EDM) process. To replicate the PDMS layer, 2 ml of liquid containing 10:1 PDMS Sylgard Silicone Elastomer 184 and Sylgard Curing Agent 184 (Dow Corning Corporation Midland, MI, USA) was mixed homogenously in a beaker using magnetic stirrer at 150 rpm for 1 h. The PDMS prepolymer was poured onto the surface of the metal die and subjected to vacuum for 20 min to degas. Following this, another metal block with a smooth surface was placed on top of the PDMS prepolymer and the entire assembly was placed at 80 ◦ C for 2 h for PDMS curing. The PDMS layer was peeled off and cut to a desired size. It was permanently bonded to a 0.1-mm thick acid-washed borosilicate glass substrate (Herenz Medizinalbedarf, Hamburg, Germany) using a 4-␮m thick spin-coated layer of PDMS prepolymer as adhesive [17]. Following this, liquors containing different primer pairs were manually loaded into wells using micro-tubing and dried by incubating the chip at 80 ◦ C for 10 min. Finally, a 0.1-mm thick acid-washed borosilicate cover glass was adhesive bonded to the top of the chip to form enclosed microchannels and a headspace for microfluidic handling of PCR mixture and PDMS prepolymer. 2.2. Vacuum system set-up The vacuum system for PCR mixture loading and isolation of the nanoliter wells consists of a vacuum enclosure made of acrylic plastic plates, a pipe connected to the vacuum pump, a valve to control the vacuum, and a pinch valve to control the sample liquid flow (Fig. 2). For the loading of the PCR mixture into the wells, a 2cm long Teflon tubing (PTFE 30 TW, Cole Parmer, USA) was inserted into the inlet channel of the chip. A pipette tip, which acts as a reservoir for holding PCR mixture was connected with the Teflon tubing through a 6.5-cm long silicone tubing (I.D. 0.020 and O.D. 0.083 , Cole Parmer, USA). Following this, the chip was mounted inside the vacuum enclosure; and the silicone tubing extended out of the enclosure and passed through a pinch valve, for control of the liquid flow by constricting or by releasing the elastic silicone tubing. The PCR mixture reservoir (pipette tip), Teflon and silicon tubing are low-cost disposable parts of our system. Prior to PCR mixture loading into the wells, the chamber is cleaned with DNA AWAY

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677

673

the feedback control. The optics of the instrument was designed to measure the fluorescence of SYBR Green I/EvaGreen, DNA intercalating dye. The EvaGreen fluorophore was excited using an array of blue LED (Marl International Ltd., Cumbria, UK) centred at 480 nm and fixed at an angle of 45◦ to the plane of the PDMSglass chip surface to prevent interference of the excitation light on the light path of the detection unit. The excitation light from the blue LED array was filtered using a bandpass filter (465–495 nm, Chroma Technologies Corp., Brattleboro, USA), and the emission light from the microreactor was filtered using another bandpass filter (515–555 nm, Chroma Technologies Corp., Brattleboro, USA) before being detected by a CCD camera (DTA, Pisa, Italy). The system is fully automated and integrated with multiple functions including thermal cycling control, real-time fluorescence imaging, on-line image processing, and data analysis. Fig. 2. The vacuum system set-up for PCR mixture loading into the PDMS matrix chip. The parts shown in the inset are disposable.

(Molecular Bioproducts) to avoid any aerosol cross-contamination of nucleic-acid template from the previous run.

2.5. Tissues specimens, RNA isolation and cDNA amplification Cancerous human hepatocellular carcinoma (HCC) surgical biopsies were obtained and were immediately snap frozen in liquid

2.3. Microfluidic operation of the PCR matrix chip In order to visualize primer re-suspension and PCR mixture loading into the wells, a blue dye (0.1% of xylene cyanol) was selectively loaded into some wells (to form “NTU” characters) and dried. The microfluidic operation procedure is shown in Fig. 3, and the results for fluidic sample loading and sealing are shown in Fig. 4. The microfluidic operation process is described below. Step 1: the liquid control pinch valve was closed. The vacuum pump was switched ON and subsequently the vacuum control valve was opened to achieve a vacuum level of 26 in.Hg. Following this PCR mixture (70 ␮l) was loaded into the reservoir using pipette (Figs. 3(A) and 4(1)). Step 2: the pinch valve was opened. The sample liquid was jetted into the headspace of the chip due to vacuum. The presence of walls surrounding the well matrix helped the liquid column to redistribute to cover the entire area of the well matrix (Figs. 3(B, C) and 4(2, 3)). Air followed the liquid column into the headspace, and purged the remaining sample (dead-volume) in the headspace. After this step, all the wells that have been loaded with liquid were isolated (Figs. 3(D) and 4(4, 5)). The entire sample loading and well isolation process can be completed in a fraction of a second, e.g. in 0.44 s as shown in Fig. 4. Step 3: the vacuum valve was closed. The dried blue dye resuspended into the PCR mixture. Only the wells preloaded with dye were rendered blue colour (Fig. 4(6)). This shows that the rapid sample loading and well isolation method is free of primer crosscontamination. Step 4: following this, liquid sealant (a PDMS prepolymer) was injected through the reservoir into the headspace of the chip, in order to seal all the wells (Figs. 3(E) and 4(7–9)). Step 5: the vacuum enclosure was opened and the chip was retrieved for PCR thermocycling. 2.4. Real-time quantitative PCR instrument A prototype real-time PCR instrument for the PDMS-glass chips was constructed (Fig. 5). The temperature of the microchip was cycled using a thermoelectric cooler (TEC) (Melcor Corp., Trenton, NJ, USA). A RTD (resistive temperature detector) was mounted on the TEC surface to measure the temperature and it was used as

Fig. 3. The fluidic operation process for PCR mixture loading, isolation and sealing. (A) A column of PCR mixture is loaded into the reservoir, and vacuum is established in the headspace and wells. (B and C) The pinch valve is opened; the sample column is filled in to the headspace. (D) Under vacuum, air followed the liquid into the headspace purging out the liquid in the headspace and isolated the wells. (E) Vacuum closed, and the sealant is injected into the headspace.

674

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677 Table 1 Primer sequence for amplification of 10 different gene targets of liver cDNA. Gene symbol

Gene name

Primer sequence

Ade-f

Adenylyl cyclase-associated protein 2

TCAATGGTGTCATTGCAGGT

Ade-r Aldo-f Aldo-r Anx-f Anx-r HSP90-f HSP90-r PK-f PK-r PD2-f PD2-r APC-f APC-r FBJ-f

Fig. 4. Visualization of the sample loading, well isolation and sealing process on the chip recorded at 25 fps. (1) The chip was placed in the vacuum enclosure before PCR mixture loading. Some of the wells were preloaded with blue dye dried on the well surface. (2 and 3) Under vacuum, the liquid is injected into the headspace. (4 and 5) Air followed the sample liquid immediately and purged the excess liquid out of the headspace through the venting channels, leaving all the wells isolated from each other. Subsequently, the vacuum is shut OFF. (6) Resuspension of the dried dye rendered the wells blue colour. (7–9) The sealant (PDMS prepolymer) was injected into the headspace from the tubing to seal the wells.

nitrogen. All tissue samples employed in this study were approved and provided by the Tissue Repository of the National Cancer Centre, Singapore. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, USA) by following the manufacturer’s protocol. cDNA was synthesized with random hexamers (Invitrogen, Carlsbad, USA) and oligo(dT) 12–18 primers using SuperScript II reverse transcriptase. All reverse transcription reaction was performed with a starting amount of 1–5 ␮g of DNase I-treated total RNA in a final volume of 20 ␮l. To validate the primer-cDNA template system, compare and evaluate the real-time PCR on the PDMS matrix chip, ten different gene fragment was PCR amplified both in polypropylene (PP) microfuge tube on a commercial real-time PCR instrument (RotorGene 3000, Corbett Research, Sydney, Australia) and in our PCR matrix chip on the prototype real-time quantitative PCR instrument with the following thermal cycling profile: initial denaturation at 95 ◦ C for 900 s; followed by 40 cycles of denaturation at 94 ◦ C for 15 s, annealing at 55 ◦ C for 30 s and extension at 72 ◦ C for 30 s. The fluorescence of EvaGreen dye was measured at the extension step of every PCR cycle. The PCR mixture contained 10 mM Tris–HCl (pH 9.0), 50 mM KCl, 0.1% Triton X-100, 2.5 mM MgCl2 , 0.2 mM each of dATP, dTTP, dGTP, and dCTP, 1.0 ␮g/␮l of BSA, 2X EvaGreen fluorescent dye (Biotium, USA), 1.0 ␮M each of gene

FBJ-r Osn-f Osn-r Ras-f Ras-r

Aldo-keto reductase family Annexin A2 Heat shock protein 90 kDa alpha (cytosolic) Protein kinase DNA activated PD2 domain Adapter related protein complex FBJ murine osteosarcoma viral oncogene homolog B Cystein rich osteonectin Ras GTPase activating protein

CTCGTGGGGTTGTTGGTACT ATCACCGTTACGGCCTACAG CGTGCTGGTGTCACAGACTT ATGTTCCCAAGTGGATCAGC CATACAGCCGATCAGCAAAA GGCCGACAAGAATGATAAGG GGGGGATCTCATCAGGAACT TGCAAGCAGATGTTCCTCAC CAACACACCAGCTGCATTTT TGGGCTAGAGGATGAGGATG AGGGGTGTCAAGTGGATCAG AATCTGAGGCTGGCTTAGCA GATGACTCTTCCCCCTGGAT AAACTCTGGCTCAGCCTGTC AAAATCTCATGTCCCCAACG GAGAAGGTGTGCAGCAATGA AGGACGTTCTTGAGCCAGTC GCAGATGCAGTCTACGGACA CCTCAGGAGCAAGGACAAAC

specific primer, 0.1 U/␮l of Platinum hot-start Taq DNA polymerase (Invitrogen, Carlsbad, USA), and cancerous liver cDNA template. Primer pairs were designed with similar Tm (melting temperature) of ∼55 ◦ C (see Table 1). The desired PCR products from the RotorGene 3000 PCR instrument was confirmed on a capillary electrophoresis chip (DNA Labchip 500) using Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) 3. Results and discussion 3.1. Real-time PCR on PDMS matrix chip To test the performance of the PDMS matrix chip and to investigate cross-contamination between wells, we preloaded primer pairs specific for 10 different target regions of cancerous liver cDNA. The samples were run in duplicates. In addition to the PDMS matrix chip with wells containing positive template control reactions (PC), a different PDMS matrix chip with wells containing no-template control reactions (NTC) was also thermal cycled. Using the vacuum aided microfluidic scheme, the primer pre-loaded PDMS matrix chip containing positive template control reactions was loaded with PCR mixture containing a fixed concentration of cDNA template and was sealed using PDMS prepolymer. The chips were then thermal cycled on the in-house real-time quantitative PCR instrument. The threshold cycle (Ct) values for amplification of 10 different regions of cDNA from cancerous human hepatocellular

Fig. 5. Diagram and photos of the real time PCR instrument.

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677

675

Fig. 6. (A) Amplification plot of ten different gene targets of cDNA from human hepatocellular carcinoma. (B) Melt curve analysis to test the purity of PCR product from the PDMS matrix chip. Information about abbreviations is given in Table 1.

carcinoma PDMS matrix chip (Fig. 6(A)), were comparable to corresponding reactions performed in a commercial RotorGene 3000 real-time PCR instrument. Melting curve analysis was performed to evaluate the purity of PCR product from our real-time PCR instrument (Fig. 6(B)). The Tm (melting temperature) for the desired PCR products and primer–dimer in NTC reaction were comparable to the values determined by RotorGene 3000 commercial real-time PCR machine. 3.2. Primer cross-contamination During the PCR mixture loading step, the headspace above all the wells is filled with liquid, Hence, all the PCR wells are fluidically connected with each other. There is a possibility of primer crosscontamination among the wells as the dried pre-loaded primer pairs inside the wells re-suspends into the PCR mixture. Hence, the timing of liquid removal in the headspace is critical factors in performing primer cross-contamination free PCR in our PDMS matrix chip. Based on our PCR experiments, we found that primer cross-contamination occurred if the PCR wells were fluidically con-

nected for a few minutes. When we reduced the incubation time of fluidically connected wells to 25 s, we observed negligible primer cross-contamination from the PCR result. In the sample loading process of our chip, the liquid in the wells connects for only 0.5 s, which is adequate to avoid primer cross-contamination. To further study the primer diffusion in our chip, we have monitored the fluorescence from HPLC purified 20 mer long oligonucleotides (primers) tagged with FAM fluorophore (5 -(6 FAM)-TCG TGC GTG GAT TGG CTT TG). The oligonucleotide tagged with FAM was preloaded and dried in selected number of PDMS wells (with reference to Fig. 7) while the remaining wells were kept empty. The PCR mixture without EvaGreen dye was loaded and the wells were isolated using the vacuum aided microfluidics. The fluorescence from the primers tagged with FAM was monitored for 5 min on a fluorescence microscope setup (Olympus optical, Germany, Olympus BX 60). The images of FAM tagged primer movement in two wells at different time points are shown in Fig. 7, indicating that there is no significant primer movement to the adjacent wells. Fig. 7 also shows that even long after the microfluidic operation is completed, the primers still stay near the bottom of the wells. 3.3. Optimization of parameters for successful vacuum aided sample loading

Fig. 7. Visualization of primer (oligonucleotide) resuspension process during sample loading. The fluorescence images show no observable cross-contamination of primers between wells after sample loading. (A) The top view of the shape of the wells. (B) Fluorescence image of the wells before sample loading. The wells at the upper-left and the lower-left corners have been preloaded with HPLC purified 20 mer long FAM tagged oligonucleotides and they have been dried in the wells. (B–C) the fluorescence images shots at selected time intervals after sample loading show the resuspension process of the oligonucleotides. (D) enlarged view of the fluorescence image before sample loading. The dried FAM tagged oligonucleotide appears concentrated along the corners of the bottom of the well. (E) 9 s after sample loading. (F) 99 s after loading, further diffusion makes the fluorescence intensity even in the well. The boundary of the fluoresced region shows that the oligonucleotides still remain near the bottom of the well. (G) 279 s after loading, a larger fluorescence area compare to F indicates the oligonucleotides have diffused upward from bottom.

3.3.1. Chip structure and fluidic design Currently, the fluidic design of the chip consists of three outlet channels (0.6 mm width and 0.1 mm depth), one inlet channel (1 mm width and 0.5 mm depth) and a headspace above the wells. The dimensions of the outlet channels determine the residencetime of the sample liquid inside the headspace. The residence-time is an important parameter, which ensures complete filling of the wells, before the liquid in the headspace is purged out through the outlet channels. If the outlet channels are larger, the liquid in the headspace may be purged out by the following air, before the liquid completes the filling of the wells, resulting in partially filled wells or even unfilled wells. For the same purpose of reducing loss of the liquid in the headspace while maintaining the high speed of the sample loading, the orientation of the outlet channels are designed perpendicular to the circulating flow directions. To avoid air trapping at the corners during sealant injection, we arranged the three outlet channels at the three corners of the square shaped chip. 3.3.2. Vacuum level and bubble formation Insufficient vacuum level results in trapping of air at the bottom of the wells which may lead to PCR failure due to expansion of

676

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677

Fig. 8. Optical micrograph of two chips and their fluorescence images with and without air bubble formation after PCR thermocycling for 20 cycles. (A) When the vacuum level is not adequate, air may be trapped at the bottom of the wells during sample loading. Trapped air expands to form bubbles during PCR thermocycling, causing the failure of PCR. (B) With sufficient vacuum level, no bubble is formed during PCR thermocycling.

trapped air during thermocycling (Fig. 8). Adequate vacuum level (2.5 Torr in our practice by a vane vacuum pump in our practice) in the chip contributes to the high-speed flow of the sample liquid into the chip immediately after opening the pinch valve, and foaming the liquid to ease the spreading process in the headspace and the wells. This vacuum level is also necessary for a rapid purging of the liquid in the headspace to prevent primer cross-contamination among the wells. 3.3.3. Microfluidic operation scheme The operation scheme described in Section 2.3 was optimized for successful sample loading and sealing in the chip. One example of incorrect operation is closing the vacuum before opening the pinch valve and using vacuum remaining in the vessel to draw the liquid in, which will result in an insufficient vacuum level during sample loading. Another inappropriate operation scheme is to close the pinch valve after sample loading, while the vacuum is still ON, this may be performed in order to degas any air bubble trapped in the wells. In such a situation, we found that rapid increase in vacuum level inside the headspace of the PCR matrix chip generated large amounts of foams. The introduction of foams led to primer cross-contamination and loss of PCR mixture. 4. Conclusion In this paper, we report a low cost PDMS matrix chip comprising an array of a hundred microwells, and successfully demonstrated the potential of the vacuum aided microfluidics in generating a large number of solution-phase wells for real-time quantitative PCR. The vacuum aided microfluidics eliminates arduous manual sample loading or costly liquid dispensing robots, and involves only a few steps of operation over a couple of valves. In addition, the vacuum aids in accomplishing reliable bubble-free PCR in PDMS matrix chip without cross-contamination of the PCR primers preloaded into the wells. Since PDMS is inert to many chemicals and biologically compatible, the matrix chip with microwells can be used for other analytical processes. For applications requiring high well density, we do not foresee any limitations imposed by our vacuum aided microfludics on well density and size. We have also tested the vacuum aided microfluidics in a higher well density matrix chip on a substrate of 45 mm × 24 mm, containing 80,000 wells, each having 64 pL in

volume with a dimension of 40 ␮m × 40 ␮m × 40 ␮m. The same microfluidics scheme described was demonstrated successfully in loading, isolating and sealing of these wells. The future work will include the development of a picoliter volume droplet deposition system of either contact or non-contact type for loading PCR primers into these wells. The upper limit in well density and well size on our chip depends on the improvement in chip operating parameters, such as, parameters to control primer crosscontamination during PCR mixture loading process, and ability to perform efficient PCR in PDMS well in nanoliter and picoliter volumes. The advantages of the chip include reduced reagent consumption, parallel analyses of multiple genes, and a potential to integrate with a nucleic-acid sample preparation system. Comparing with other PCR matrix well array chips using either micro valves or robotic liquid handling system, this chip and the related microfluidics are simple and economic to fabricate as well as to operate. Acknowledgment The authors acknowledge the financial support of the Biomedical Research Council of Singapore under project 04/1/31/19/365. References [1] J.E. Larkin, B.C. Frank, H. Gavras, R. Sultana, J. Quackenbush, Independence and reproducibility across microarray platforms, Nat. Methods 2 (2005) 337–343. [2] B.N. Strizhkov, A.L. Drobyshev, V.M. Mikhailovich, A.D. Mirzabekov, PCR amplification on a microarray of gel-immobilized oligonucleotides: detection of bacterial toxin- and drug-resistant genes and their mutations, BioTechniques 29 (2000) 844–857. [3] S.V. Tillib, B.N. Strizhkov, A.D. Mirzabekov, Integration of multiple PCR amplifications and DNA mutation analyses by using oligonucleotide microchip, Anal. Biochem. 292 (2001) 155–160. [4] G. Mitterer, M. Huber, E. Leidinger, C. Kirisits, W. Lubitz, M.W. Mueller, W.M. Schmidt, Microarray-based identification of bacteria in clinical samples by solid-phase PCR amplification of 23S ribosomal DNA sequences, J. Clin. Microbiol. 42 (2004) 1048–1057. [5] D.H. Bing, C. Boles, F.N. Rehman, M. Audeh, M. Belmarsh, B. Kelley, C.P. Adams, Bridge amplification: a solid phase PCR system for the amplification and detection of allelic differences in single copy genes, in: Genetic Identity Conference Proceedings, Seventh International Symposium on Human Identification, 1996. [6] J.H. Daniel, S. Iqbal, R.B. Millington, D.F. Moore, C.R. Lowe, D.L. Leslie, M.A. Lee, M.J. Pearce, Silicon microchambers for DNA amplification, Sensors Actuat. A: Phys. 71 (1998) 81–88. [7] Y. Matsubara, K. Kerman, M. Kobayashi, S. Yamamura, Y. Morita, E. Tamiya, Microchamber array based DNA quantification and specific sequence detection from a single copy via PCR in nanoliter volumes, Biosens. Bioelectron. 20 (2005) 1482–1490. [8] H. Nagai, Y. Murakami, Y. Morita, K. Yokoyama, E. Tamiya, Development of a microchamber array for picoliter PCR, Anal. Chem. 73 (2001) 1043–1047. [9] J. Liu, C. Hansen, S.R. Quake, Solving the “world-to-chip” interface problem with a microfluidic matrix, Anal. Chem. 75 (2003) 4718–4723. [10] V.P. Iordanov, M. Malatek, J. Bastemeijer, I.T. Young, A. Bossche, G.W.K. Van Dedem, P.M. Sarro, M.J. Vellekoop, PCR array on chip-thermal characterization, Proc. IEEE Sensors (2003) 1045–1048. [11] A.M. Chaudhari, T.M. Woudenberg, M. Allin, K.E. Goodson, Transient liquid crystal thermometry of microfabricated PCR vessel arrays, J. MEMS 7 (1998) 345–354. [12] J.H. Leamon, W.L. Lee, K.R. Tartaro, J.R. Lanza, G.J. Sarkis, A.D. deWinter, J. Berka, K.L. Lohman, A massively parallel PicoTiterPlate based platform for discrete picoliter-scale polymerase chain reactions, Electrophoresis 24 (2003) 3769–3777. [13] T. Morrison, J. Hurley, J. Garcia, K. Yoder, A. Katz, D. Roberts, J. Cho, T. Kanigan, S.E. Ilyin, D. Horowitz, J.M. Dixon, C.J.H. Brenan, Nanoliter high throughput quantitative PCR, Nucleic Acids Res. 34 (2006). [14] E.T. Lagally, P.C. Simpson, R.A. Mathies, Monolithic integrated microfluidic DNA amplification and capillary electrophoresis analysis system, Sensors Actuat. B 63 (2000) 138–146. [15] M. Allen Northrup, B. Benett, D. Hadley, P. Landre, S. Lehew, J. Richards, P. Stratton, A miniature analytical instrument for nucleic acids based on micromachined silicon reaction chambers, Anal. Chem. 70 (1998) 918–922. [16] A. Dahl, M. Sultan, A. Jung, R. Schwartz, M. Lange, M. Steinwand, K.J. Livak, H. Lehrach, L. Nyarsik, Quantitative PCR based expression analysis on a nanoliter scale using polymer nano-well chips, Biomed. Microdev. 9 (2007) 307–314.

H.-B. Liu et al. / Sensors and Actuators B 135 (2009) 671–677 [17] H.-B. Liu, H.-Q. Gong, N. Ramalingam, Y. Jiang, C.-C. Dai, K.M. Hui, Micro air bubble formation and its control during polymerase chain reaction (PCR) in polydimethylsiloxane (PDMS) microreactors, J. Micromech. Microeng. 17 (2007) 2055–2064.

Biographies Hao-Bing Liu received the B.Eng. Degree from Hubei University of Automotive Technology, China, in 1993, and M.Eng. degree from Huazhong University of Science and Technology, China, in 2000, in the area of automotive engineering. He received the Ph.D. degree from Nanyang Technological University (NTU), Singapore, in 2007, for his work in Microelectromechanical Systems (MEMS). He was a lecturer of mechanics and automotive engineering in Hubei University of Automotive Technology, China, in 1993–1997. From 2005 to 2007 he was a research engineer in National Cancer Centre Singapore, working on Bio-chip developing and application. He has been a Research Fellow with BioMEMS lab, School of Mechanical and Aerospace Engineering, Nanyang Technological University since Feb. 2008. He is pursuing research in micro fabrication, microfluidics, and Bio-MEMS. Naveen Ramalingam was born in India and was awarded the degree of Bachelors of Technology (B.Tech) Industrial Biotechnology by the Centre for Biotechnology, Anna University, India in 2001. He worked as a research assistant in Indian Institute of Science (IISc) in 2002. In 2003, he graduated Masters Molecular Engineering of Biological and Chemical Systems (MEBCS) from the Singapore-Massachusetts Institute of Technology (MIT) Alliance (SMA), National University of Singapore. In 2005, he joined National Cancer Centre, Singapore and developed microfluidic nanochip for predicting recurrence of human hepatocellular carcinoma. He is currently working as a Research Associate with BioMEMS lab at Nanyang Technological University

677

on a project to develop high-throughput biochips for parallel detection of multiple pathogens. His areas of research interest include design, fabrication and validation of MEMS devices for microfluidic applications in life sciences. Yu Jiang received the Bachelor of Science degree from Huazhong University of Science and Technology, China, in 1982 and received the Master of Science degree from Nanyang Technological University, Singapore in 2002. He is a Research Associate with the Nanyang Technological University, Singapore. Chang-Chun Dai received the M.S. degree from the University of Science and Technology, Chinese Academy of Sciences in 1990. He is a research associate at BioMEMs Lab in School of MAE, Nanyang Technological University, Singapore. Kam Man Hui received the B.Sc degree and M.Sc degree in Bacteriology & Public Health form Washington State University, Washington, USA, in 1976 and 1978, respectively. He received the Ph.D. degree in the area of Immunology from Northwestern University Medical School, Chicago, USA, in 1982. His current position is the head of Cellular and Molecular Research Division, National Cancer Centre, Singapore. He is a Fellow of The Royal College of Pathologists [FRCPath (UK)], and a Council Member of International Society of Cancer Gene Therapy. He has become an Adjunct Full Professor in Nanyang Technological University, Singapore since 2002. His research interests include Gene Therapy and the Immunotherapy of Cancer, and Genomics of human cancers Hai-Qing Gong received his bachelor degree from Wuhan University of Technology, China in 1982, and his master and Ph.D. degrees from University of Delaware, USA in 1987 and 1991, respectively, for his work in non-Newtonian fluid mechanics. He joined School of Mechanical and Aerospace Engineering at Nanyang Technological University, Singapore in 1991 and worked on electrorheological fluids and devices. From 1997, he started to work in MEMS and bioMEMS areas. He is currently an associate professor and the director of BioMEMS Laboratory at Nanyang Technological University, Singapore. His current research interests include biochips for genetic testing, microfluidics and nanofluidics.