Anal. Chem. 2001, 73, 1043-1047
Development of A Microchamber Array for Picoliter PCR Hidenori Nagai,* Yuji Murakami, Yasutaka Morita, Kenji Yokoyama, and Eiichi Tamiya
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1, Asahidai, Tatsunokuchi, Ishikawa 923-1292, Japan
A microchamber array for PCR was developed by semiconductor microfabrication technology. The microchambers were designed to be of picoliter quantity. To optimize fluid retention, the surface states of the substrate and the inner walls were examine for four different types of microchamber. The substrate was silicon, while silicon dioxide was selected for the inner walls. PCR was performed in the microchamber array, and the amplification of DNA was detected using a technique based on the energy transfer of fluorescent dyes. The lower volume limit for PCR was investigated using various sizes of microchambers. Microchambers with volume greater than 86 pL gave successful PCR. In addition, the system was improved in order to take up the PCR product. To prevent mixing of the samples, the samples were dried after PCR using a membrane that permeates only vapor. Miniaturization concepts have recently been brought to the forefront of the analytical sciences. This paradigm shift toward micro- and nanoscale experiments is driven partly by the increasing costs of samples and reagents, but mostly by the constantly improving sensitivity of modern analytical equipment. Moreover, miniaturization of devices facilitates high integration, which has significantly increased the sample throughput. Presently, there are numerous laboratories that have utilized some form of parallel synthetic processes to create new chemicals.1-4 This powerful, synthetic methodology allows chemically diverse sample sets (i.e., libraries) to be created in order to facilitate the discovery of new materials, catalysts, and even pharmaceutical agents.5,6 Furthermore, there is great interest in applying this synthetic technique to the screening of biological materials, such as genome, RNA, peptides, enzymes, receptors, and antibodies, because there is a huge number of these biological materials. For example, the human genome is ∼3 billion base pairs, in which it is estimated that the number of significant genes is 4.5 × 105.7 These biological * To whom correspondence should be addressed: (phone) +81-761-51-1663; (fax) +81-761-51-1665; (e-mail)
[email protected]. (1) Lam, K. S.; Lebl, M.; Krchna´k, V. Chem. Rev. 1997, 97, 411-448. (2) Vardine, G. L. Nature 1996, 384, 11-13. (3) Hogan, J. C., Jr. Nature 1996, 384, 17-19. (4) Ramsay, G. Nature 1998, 16, 40-44. (5) Xiang, X. D.; Sun, X.; Briceno, G.; Lou, Y.; Wang, K. A.; Chang, H.; WallaceFreedman, W. G.; Chen, S. W.; Schultz, P. G. Science 1995, 268, 17381740. (6) Thompson, L. A.; Ellman, J. E. Chem. Rev. 1996, 96, 555-600. (7) Dunhan, I.; Shimizu, N.; Roe, B. A.; Chissoe, S.; et al. Nature 1999, 402, 489-495. 10.1021/ac000648u CCC: $20.00 Published on Web 12/08/2000
© 2001 American Chemical Society
materials have many special functions (i.e., genetic information, catalytic activity, molecule recognition function, and antibody diversity) compared with inorganic materials. In the Human Genome Project, an effort is being made to accumulate genetic sequence data by many organizations, where hundreds of polymerase chain reaction (PCR)8 instruments are used continuously. The efficiency and cost of PCR is therefore very important. The PCR is an enzymatic process that amplifies a specific DNA target sequence in response to temperature cycling among denaturation (94 °C), annealing (e.g., 55 °C), and extension (e.g., 72 °C). The PCR has advanced DNA amplification dramatically. However, the acquisition time often plays a critical role in the overall throughput of a given technique. However, large-scale integration of PCR affords an advantage in terms of sample throughput that allows synthesizing of a huge number of samples simultaneously. For the past number of years, many researchers have reported on high-throughput PCR equipment. Sasaki and co-workers9 developed a PCR instrument with four plates attached to 384 reaction chambers. Taylor et al.10 fabricated a silicon microdevice with 48 chambers in an area of ∼2 mm2 and demonstrated PCR amplification in 0.5 µL. Furthermore, a number of researchers have investigated the microminiaturization of the PCR process for potential integration in the microfabricated chip format. Woolley et al.11 accomplished PCR amplification of DNA in a novel microfabricated PCR device that could be directly interfaced with an electrophoretic chip for PCR product analysis. Northrup and co-workers12 designed a portable system containing a miniature analytical thermal cycling instrument in which a silicon-micromachined reaction chamber with integrated heaters and optical windows was used to conduct PCR. This particular device was not interfaced with an electrophoretic separation channel. Instead, characterization of the amplification product was accomplished by real-time fluorescence monitoring. Thus, the chamber density on PCR array has increased in order to elevate the throughput. The silicon-based microfabrication techniques can fabricate micrometer-sized chamber and realize higher chamber density. (8) Mullis, K.; Faloona, F.; Scharf, S.; Saiki, R.; Horn, G.; Erlich, H. Cold Spring Harbor Symp. Quantum Biol. 1986, 51, 263-273. (9) Sasaki, N.; Izawa, M.; Shimojo, M.; Shibata, K.; Akiyama, J.; Itoh, M.; Nagaoka, S.; Carninci, P.; Okazaki, Y.; Moriuchi, T.; Muramatsu, M.; Watanabe, S.; Hayashizaki, Y. DNA Res. 1997, 4, 387-391. (10) Taylor, T. B.; Winn-Deen, E. S.; Picozza, E.; Woudenberg, T. M.; Albin, M. Nucleic Acids Res. 1997, 25, 3164-3168. (11) Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086. (12) Northrup, M. A.; Bennet, B.; Hadley, D.; Landre, P.; Lehew, S.; Richards, J.; Stratton, P. Anal. Chem. 1998, 70, 918-922.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001 1043
However, an overminiaturized PCR chamber may prevent the amplification reaction in itself. So we examined the volume limit of PCR in this report. The challenges associated with highly integrating PCRs are not insignificant. The miniaturization of the reaction into a picoliter volume results in peculiar phenomena. In such a volume, we have to consider the number of reactant molecules. For example, when a test tube contains low-concentration solution, which means the number of reactants is extremely small, the concentration increases by enclosing them in a picoliter chamber. The kinetics depends on the concentration of reactant. Therefore, the efficiency of reaction in the microchamber is higher than one in an ordinal vessel, such as test tube, when the reactant quantity is infinitesimal.13,14 In general, the template concentration is low in the PCR solution. Furthermore, a 1 pL (1 pL ) 1000 µm3) chamber filled with 1 pM solution will be occupied with only one molecule on the average. A single molecule of the template can be isolated briefly by enclosing the PCR mixture into a 1-pL chamber.15 Most importantly, it is possible to amplify only one template in mixtures containing many kinds of templates (i.e., cDNA library) without interference by other templates. PCR is a typical competitive reaction.16 For example, a cDNA encoding a valuable molecule, such as interferon and growth factors, is infinitesimal in quantity in a cDNA library. To screen such an infinitesimal quantity of cDNA, it is necessary to investigate from 100 to 1000 clones. Such work is of enormous expense and therefore only possible for the largest organizations. However, it is expected that the methodology by means of a picoliter chamber will facilitate the screening of infinitesimal biomaterials. In this report, we fabricated a large-scale integrated picoliter chamber array and describe PCR amplification in this device. Furthermore, we improved the extraction of products from the microchambers by using a special membrane. EXPERIMENTAL SECTION Materials. The Si 〈100〉 wafers were purchased from ShinEtsu, Japan. The primers for the pGFP template (Clontech, Palo Alto, CA) were purchased from International Reagent. The sequences of the primers are as follows: forward primer, d(GTGCTGCCAT AACCATGAGT G); reverse primer, d(ATCCCCCATG TTGTGCAAAA). The probe and the PCR reagents were purchased from Perkin-Elmer. The sequence of the probe is as follows: FAM-CGGCCAACTT ACTTCTGACA ACGATCG-TAMRA. Tetramethylammonium hydroxide (TMAH) was kindly provided by Toyo Gosei. Nuclease-free BSA and all other chemicals were obtained from Wako Pure Chemical. Microchamber Array Fabrication and Examination of Optimal Surface Condition. The microchamber array used to run PCR was fabricated on a Si 〈100〉 wafer using photolithography and anisotropic etching. Prior to processing, SiO2 was grown on the undoped Si 〈100〉 wafer using wet thermal oxidation. The oxide surface was then coated with OMR-83 photoresist (Tokyo Ohka (13) Masuhara, H.; Kitamura, N.; Misawa, H.; Sasaki, K.; Koshioka, M. J. Photochem. Photobiol. A: Chem. 1992, 65, 235-247. (14) Ekstro ¨m, S.; O ¨ nnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; MarkoVerga, G. Anal. Chem. 2000, 72, 286-293. (15) Jackman, R. J.; Duffy, D. C.; Ostuni, E.; Willmore, N. D.; Whitesides, G. M. Anal. Chem. 1998, 70, 2280-2287. (16) Lo, D. Y. M., Ed. Clinical Applications of PCR; Humana Press Inc.: Totowa, NJ, 1998.
1044 Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
Figure 1. Two scanning electron microscopy images of microchambers. (A) Expanded view of 80 µm × 80 µm microchamber. (B) Top view of a part of the microchamber array.
Kogyo, Tokyo, Japan), exposed to UV light in a Karl Suss MA-6 photolithography instrument. This process transferred the microchamber pattern (arrays). The various sizes of microchamber were designed in 2 × 2 cm. The oxide was removed using a HF/NH4F (1:6) solution, and the remaining photoresist was then stripped using acetone. Anisotropic etching was performed using a 25% (w/v) TMAH in water solution. This procedure resulted in microchambers with a well-defined geometry,17 therefore facilitating the determination of the internal volume using a Sloan DEKTAK 3030 profilemeter and scanning electron microscope (SEM). An expanded view of a single microchamber is shown in Figure 1A. This scanning electron micrograph clearly illustrates the tapered walls of the chambers. Each microchamber has a reversed-pyramidal cross section, and the sizes range from 2500 to 400 mm2,which corresponds to a volume range of 1.3 pL to 32 µL, respectively. A top view of a small portion of the arrayed microchambers is shown in Figure 1B. Part of the 100 × 100 chamber arrays is shown in this SEM image. Detection of PCR Products. Part of the pGFP fragment was amplified by PCR. The PCR mixture contained 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 3.5 mM MgCl2, 200 mM each dATP, dTTP, (17) Bean, K. E. IEEE Trans. Electron Devices 1978, 25, 1185-1193.
dCTP, and dGTP, 300 nM each primer, 200 nM probe, and 0.025 unit/µL AmpliTaq Gold DNA polymerase. BSA (0.05% w/v) was also added in the PCR mixture. Initial assessment of PCR performance was determined using electrophoresis. Separations were performed using 2% agarose gel. PCR performances between polypropylene tubes and a 32-µL silicon microchamber array were compared. Various sizes of the microchambers were employed to evaluate the limit of volume in PCR, ranging in volume from 1.3 pL to 32 µL. The amplification of a fragment of gfp was monitored by a technique using energy transfer of the fluorescent dyes.18 This technique employs a dye-labeled probe technology and 5′exonuclease activity of DNA polymerase. The probe is labeled with a fluorescent reporter dye and a fluorescent quencher dye. The proximity of the quencher to the reporter results in efficient quenching of fluorescence. The polymerase degrades the bond between the reporter and the quencher only if the probe hybridizes to the target. The probe is displaced from the target and the strand polymerization continues. Separation of the reporter and the quencher in solution results in an increase of fluorescence. This fluorescence is proportional to the amount of PCR product generated. The PCR mixture was dropped onto the microchamber array. The glass cover slip was put on so that it would cover all chambers. Overflowing of the solution was eliminated, and the substrate was sealed by varnish. The device was put onto the heat sink of a thermal cycler (model 9600, PE Applied Biosystems). The thermal cycling conditions were controlled in the following sequence: preheat at 95 °C for 1 min; 40 cycles consisting 95 °C for 1 min, 55 °C for 1 min, and 72 °C for 1 min; hold at 4 °C continuously. The fluorescence following the amplification of the gfp fragment was observed by a CCD camera through a microscope. Improvement for Taking Out the Sample. To take out PCR products from the microchambers, the glass cover slip must first be removed from the microchamber array. However, crosscontamination of the PCR products will occur due to a capillary effect if the cover is removed. To overcome this problem, a HIPORA membrane (Kolon Co.) was employed. The membrane contains as many as 9 billion micropores/1 in2 and prevents water permeation while allowing the passage of vapor. PCR was performed on a microchamber array following the above procedure with the membrane which was held between the microchamber array and the glass cover slip. After PCR, the glass cover slip was removed and the microchamber array was heated on a hot plate until dry. After drying was completed, the membrane was removed. To check the PCR amplification, DNA in microchambers was stained with YOYO-1. The fluorescence intensity in each microchamber was compared before and after PCR. RESULTS AND DISCUSSION Examination of Surface Conditions. By deposition or removing of the SiO2 layer of the silicon substrate surface, the hydrophilic property of the surface was controlled. The inner walls of the chambers and the rest of the substrate were separately processed by wet thermal oxidation in order to optimize the retaining ability of the solution in the microchambers. Four kinds (18) Holland, P. M.; Adramson, R. D.; Watson, R.; Gelfand, D. H. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 7276-7280.
Table 1. Relationship between Surface State of the Microchamber Array and Retention Time of Watera hydrophilicity
a
inner walls
substrate
retention time/min
X X O O
O X O X
13 180
(Ο) indicates hydrophilicity and (×) indicates hydrophobicity.
of microchamber arrays were produced with a combination of hydrophilic and hydrophobic surfaces (Table 1). The retaining ability of the solution was examined by the following procedure: the water was dropped on the microchamber, then a glass cover slip was placed over the microchamber array, and the extra water was removed. The water evaporated from the edge of the cover glass by capillary effect and the speed of the disappearance of water was compared. The ability to retain the solution in a microchamber strongly depended on the hydrophilic property of the inner wall and the rest of substrate. Table 1 shows the relationship between the surface condition of the array and the retention time of the water. In cases where the whole surface is bare silicon (hydrophobic), the water leaked out in several tens of seconds around the glass cover slip. On one hand, when only the inner walls are hydrophilic, the microchamber array can retain the water for over 3 h. This is ∼3 times that of the microchamber array whose surface contains silicon dioxide (hydrophilic). The microchamber with hydrophilic inner walls on a hydrophobic substrate has excellent water retention abilities. Furthermore, the hydrophobicity of the SiO2 stripped silicon layer remained for over one week. Hence, it has been adopted as the best microchamber array for PCR. PCR in the Microchamber Array. The amplification of DNA was examined in the PCR using the microchamber array. The amplification was estimated using the technique based on the energy transfer of the fluorescent dyes. When the PCR mixture was of ordinary composition, a change of fluorescence intensity was hardly observed before and after the PCR (Figure 2A). Oda et al.19 reported that the DNA polymerase adsorbs to the metal and is unable to extend DNA. In addition, DNA is also easily adsorbed to glass, and it seems to be absorbed to SiO2 similarly. At pH 7.0, the SiO2 surface has a negative charge, but near the surface, there is a positive charge as a result of an electric double layer. DNA has a negative charge under the same conditions. Consequently, a DNA strand sticks to the surface easily. Hence, it is with difficulty that the DNA polymerase binds to the DNA strands geometrically. To improve the efficiency of PCR, BSA was added to the PCR mixture. When BSA is present in the PCR mixture, BSA sticks to the SiO2 surface competitively with DNA polymerase and DNA because BSA also has a negative charge at pH 7.0. By adding BSA to the PCR mixture, the DNA was amplified efficiently. As the result, a difference in the fluorescence intensity appeared between before and after the PCR, as shown (19) Oda, R. P.; Strausbauch, M. A.; Huhmer, A. F. R.; Borson, N.; Jurrens, S. R.; Craighead, J. P.; Wettstein, J.; Eckloff, B.; Kline, B.; Landers, J. P. Anal. Chem. 1998, 70, 4361-4368.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
1045
Figure 3. Relationship between volume of PCR solution and the difference in fluorescence intensity before and after PCR. The lower limit of volume for PCR in the silicon-based microchamber was 85 pL.
Figure 2. Change of the fluorescence intensity by PCR. In the upper four pictures, the left side is the expanded view of the microchamber (32 µL) and the right side is the rest of the substrate. Lower graph shows the profile of the fluorescence intensity. (A) Prepared PCR solution without BSA. (B) Added 0.2% w/v BSA into PCR solution.
Figure 2B. The peak height was 2 times as large as the one before PCR. The amplified rate of the fluorescence intensity was the same as the result in a 100-mL tube. Moreover, using electrophoresis, the amplification of the gfp fragment was estimated in a 32-µL silicon chamber. Consequently, a single band appeared at ∼200 bp which corresponds to the gfp fragment. The lower limit of the volume for PCR was examined using microchambers of various volumes. The chambers were prepared with 21 kinds with capacities from 1.3 pL to 32 µL, and PCR was performed in each chamber. Figure 3 shows the relationship between the volume and the difference in fluorescence intensity before and after PCR. Though the gfp fragment was amplified in the 86-pL capacity, it hardly increased in 23-pL chamber. There are two possibilities for insufficient increase of fluorescent intensity. One is the lack of fluorescent dye in each microchamber. However, when the 23-pL chambers filled with the PCR product amplified in a 100-µL tube, the fluorescence intensity was higher than the PCR product processed in the 23-pL chamber. The other is a lack of the contents in the PCR mixture. Although the reagents were estimated as a high enough quantity for PCR, it seems that the total activity of DNA polymerase was not enough by adsorption onto the inner walls. Moreover, the possibility of cross talk of solutions between each microchamber was examined. We used glass beads (GlassMilk, Bio 101) to dispense pGFP templates into individual microchambers. The pGFP-absorbed beads were prepared by the usual method for gene purification. The pGFP-absorbed beads and the beads without pGFP absorption were dispensed into 1046 Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
Figure 4. Delivery of the pGFP-absorbed beads and the beads without pGFP adsorption into individual microchambers by a glass micropipet and the difference in the fluorescent intensity before and after PCR. Black arrows indicate the microchambers containing pGFP-absorbed beads and white arrows indicate the microchambers containing the beads without pGFP absorption. Unfilled arrows indicate the microchambers containing only PCR mixture without both kinds of beads.
selected microchambers by using a glass micropipet (Figure 4). A glass cover slip was applied and PCR mixture without template was poured between the microchamber array and the cover slip, and then the cover slip was glued by vanishing. After PCR on this microchamber array, fluorescence intensity was increased only in microchambers containing the pGFP-absorbed beads. In the microchambers containing the beads without pGFP absorption, there were fluorescence changes caused by light scattering on the surface of the beads. However, the fluorescence changes were very small, compared with the microchambers containing the pGFP-absorbed beads. Thus, the fluorescent molecules produced by PCR were kept in the microchamber. This result suggested that there was no cross talk between each solution. Improvement Using an HIPORA Membrane. In the case of PCR on the microchamber array directly sealed with a glass
cover slip, it is difficult to open the cover slip after PCR. This is because the sample solutions are mixed together by capillary action between the substrate and the glass cover slip while it is removed. The HIPORA membrane is an excellent water repellent and allows permeability of vapor through many micropores. Therefore, the liquid is unable to penetrate this membrane but the water vapor can leak out. The membrane was held between the microchamber array and a glass cover slip. During the PCR, the sample solutions never evaporated when covered. Only when the cover slip was removed after the PCR did the water in the microchambers evaporate by heating through the HIPORA membrane. After the evaporation was completed, the membrane was removed. By this improvement, it is possible to remove the glass cover slip without contamination and the sample can be utilized. However, HIPORA membrane is not transparent. The membrane prevents fluorescence observation in the microchambers. To observe the amplification of DNA, the dried PCR products were stained by YOYO-1. This chemical was not mixed during the PCR because it is inhibits PCR. There was a little fluorescence, which was caused by a few template genes, in the microchambers without the PCR process. After PCR, the fluorescence was observed along the ridge lines on the pyramidal shape of the microchamber (data not shown). We considered that the PCR products in the microchambers were concentrated and adsorbed onto the ridge lines during the drying process with the HIPORA membrane. It was obvious that the DNA fragments were amplified by PCR and enough PCR products were retained in the microchambers dried with the HIPORA membrane. Hence, the HIPORA membrane is useful for taking up the samples after PCR from the microchamber array. For our experiments, it is difficult to manipulate multiple samples into each microchamber individually. Of course, excellent microdispensers15,20-22 have been developed by many researchers; these microdispensers are able to dispense samples and reagents individually. However, the objects of our research are new types
of PCR, such as single-cell PCR and single-molecule PCR. In these PCRs, we will use a diffusion of cells and cDNA molecules on the microchamber array with the following Poisson distribution.23 The microchamber array is well suited for these applications and will be useful to discover new functional biomaterials.
(20) Morozov, V. N.; Morozova, T. Y. Anal. Chem. 1999, 71, 3110-3117. (21) Lemmo, A. V.; Fisher, J. T.; Geysen, H. M.; Ross, D. J. Anal. Chem. 1997, 69, 543-551. (22) Martin, B. D.; Gaber, B. P.; Patterson, C. H.; Turner, D. C. Langmuir 1998, 14, 3972-3975.
AC000648U
CONCLUSION A microchamber array for PCR was developed by semiconductor microfabrication techniques; the volume of each chamber was 85 pL. Amplification of pGFP was confirmed using a technique based on energy transfer of fluorescent dyes. PCR efficiency was increased by mixing BSA in the PCR mixture before the reaction. Furthermore, it is possible to pick up the PCR products as a result of using a HIPORA membrane. Recently, the DNA sequence of human chromosome 21 and 22 was reported in the Human Genome Project;7,24 the whole human genome sequence will soon be determined. At “postgenome”, all genes will need to be experimentally identified by the preparation of cDNA libraries and the determination of the function of all genes. This picoliter chamber array for PCR will provide a strong tool in such experiments, as new applications of PCR are considered using this device. A 1-pL chamber filled with 1 pM solution will contain, on average, one molecule. Since template concentration in the PCR solution is generally low, one molecule of the template can be simply isolated by the microchamber array. One-molecule PCR will be easily performed, which can amplify only one template in the mixtures containing various kinds of templates (i.e., cDNA library) without interference by the other templates. ACKNOWLEDGMENT This work was partially supported by research for the future program of JSPS. Special thanks are given to Toyo Gosei Kogyo Co., Ltd. for kindly providing the TMAH solution. Received for review June 6, 2000. Accepted October 19, 2000.
(23) Oldenburg, K. R.; Zhang, J. H.; Chen, T.; Maffia, A. M., III; Blom, K. F.; Combs, A. P.; Chung, D. Y. J. Biomol. Screen. 1998, 3, 55-62. (24) Reeves, R. H. Nature 2000, 405, 283-284.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
1047
