A titer plate-based polymer microfluidic platform for

Biomed Microdevices (2008) 10:21–33 DOI 10.1007/s10544-007-9106-y

A titer plate-based polymer microfluidic platform for high throughput nucleic acid purification D. S.-W. Park & M. L. Hupert & M. A. Witek & B. H. You & P. Datta & J. Guy & J.-B. Lee & S. A. Soper & D. E. Nikitopoulos & M. C. Murphy

Published online: 21 July 2007 # Springer Science + Business Media, LLC 2007

Abstract A 96-well solid-phase reversible immobilization (SPRI) reactor plate was designed to demonstrate functional titer plate-based microfluidic platforms. Nickel, large area mold inserts were fabricated using an SU-8 based, UVLIGA technique on 150 mm diameter silicon substrates. Prior to UV exposure, the prebaked SU-8 resist was flycut to reduce the total thickness variation to less than 5 μm. Excellent UV lithography results, with highly vertical sidewalls, were obtained in the SU-8 by using an UV filter to remove high absorbance wavelengths below 350 nm. Overplating of nickel in the SU-8 patterns produced high quality, high precision, metal mold inserts, which were used

D. S.-W. Park : M. L. Hupert : M. A. Witek : B. H. You : J. Guy : S. A. Soper : D. E. Nikitopoulos : M. C. Murphy Center for Bio-Modular Multi-Scale Systems, Louisiana State University, Baton Rouge, LA 70803, USA M. L. Hupert : M. A. Witek : S. A. Soper Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803, USA B. H. You : D. E. Nikitopoulos : M. C. Murphy (*) Department of Mechanical Engineering, Louisiana State University, Baton Rouge, LA 70803, USA e-mail: [email protected] P. Datta Center for Advanced Microstructures and Devices (CAMD), Louisiana State University, Baton Rouge, LA 70806, USA J.-B. Lee Erik Jonsson School of Engineering and Computer Science, University of Texas at Dallas, Dallas, TX 75083, USA

to replicate titer plate-based SPRI reactors using hot embossing of polycarbonate (PC). Optimized molding conditions yielded good feature replication fidelity and feature location integrity over the entire surface area. Thermal fusion bonding of the molded PC chips at 150°C resulted in leak-free sealing, which was verified in leakage tests using a fluorescent dye. The assembled SPRI reactor was used for simple, fast purification of genomic DNA from whole cell lysates of several bacterial species, which was verified by PCR amplification of the purified genomic DNA. Keywords Titer plate . Multi-well microfluidic platform . Solid-phase reversible immobilization . UV-LIGA . Large area mold insert . Micro molding . Nucleic acid purification

1 Introduction The standard titer plate, with a 127.76 mm by 85.48 mm footprint (SBS 2004), is currently used to carry out multiple reactions in parallel in separate wells for high throughput applications. Rapid interrogation of many biochemical samples in the standard titer plate is needed for such diverse applications as high throughput screening for drug discovery, biological discovery in genomics and proteomics, medical diagnostics, forensics and homeland defense (Darvas et al. 2004; Honma et al. 2004). However, conventional titer plates typically serve only one function, containment of the sample and appropriate reagents in the wells with processing implemented by placing the titer plate in a bench top instrument, such as a thermal cycler, ultra-centrifuge, or capillary array. Microfluidic platforms produced via microelectromechanical system (MEMS) fabrication techniques, known

22

as micro total analysis systems (μTAS) or lab-on-a-chip systems, are expected to be alternatives to the bulky, bench top commercial instruments currently in use. They offer many advantages including reduced reagent consumption, shorter analysis times, the capability for integration of multiple microfluidic components, potential for automation, and significantly reduced manufacturing costs via mass production (Auroux et al. 2002; Erickson et al. 2004; Huang et al. 2002; Lee and Lee 2004; Reyes et al. 2002). Microfluidic devices have been widely used in standard analytical operations and applications such as cell separation, cell lysis, DNA preconcentration and purification, amplification using the polymerase chain reaction (PCR) and detection (Auroux et al. 2004; Huang et al. 2002; Liu et al. 2003; Obeid et al. 2003; Paegel et al. 2002; Pal et al. 2005). Most microfluidic platforms have been manufactured in Si or glass, because of their excellent mechanical and chemical properties, with the use of highly precise and reproducible microfabrication processes including thin film deposition, lithography, etching, and substrate bonding (Paegel et al. 2002; Ziaie et al. 2004). Microfluidic devices can also be realized in polymers from a large selection of different polymer materials with different combinations of ease of surface modification, biocompatibility, disposability, and mass producibility. Polymer microfluidic devices have been successfully demonstrated in materials such as poly(methyl methacrylate), PMMA (Qi et al. 2002), polycarbonate, PC (Barrett et al. 2004; Mitchell et al. 2003; Witek et al. 2006), cyclo olefin copolymer, COC (Kim et al. 2006), and poly(dimethylsiloxane), PDMS (Liu et al. 2003). These devices can be produced using micromolding techniques, such as injection molding or hot embossing (Heckele and Schomburg 2004; Kim et al. 2006). Incorporation of a microfluidic device at each well location in a standard titer plate format can significantly enhance the functional capability of the plate for the rapid analysis of a large number of biochemical samples at significantly lower costs. Microfabrication of 96-well capillary electrophoresis devices was demonstrated using micromilling and hot embossing (Gerlach et al. 2002a, b; Guber et al. 2004), but minimum feature dimensions and pattern densities were limited by the size of the finger mills used. Feature dimensions and pattern densities can be improved by using either the LIGA or UV-LIGA processes. Even though LIGA has the potential for use in applications over large areas, as would be required for a titer-plate format (10,920 mm2), using a step-and-repeat exposure would require the implementation of scanning protocols and the design of the fluidic networks that reduce the effects of stitching errors at exposure boundaries. An UV-LIGA technique with SU-8 UV lithography and nickel electroplating was adopted to realize such titer platebased microfluidic platforms. The SU-8 based UV-LIGA

Biomed Microdevices (2008) 10:21–33

process on 150 mm diameter silicon substrates enabled use of a single UV-exposure to make polymer electroplating templates, highly packed with small features (150°C resulted in some visible deformation of the microposts, so a bonding temperature of 150°C was selected for thermal fusion bonding of the 96-well SPRI chips. In order to assess whether there were any leaks between the cover plate and the hot embossed PC chip in the microfluidic channel network, fluorescein was pushed through the microfluidic channels and wells [see Fig. 7(a)]. A close-up view of one of the 96-wells at 20× magnification shows that no fluorescent signal was observed in the areas outside of the microfluidic channels indicating proper sealing between the hot embossed PC chip and the cover plate [see Fig. 7(b)]. The distribution channels were designed to be of the same length and have the same cross sectional dimensions in order to maximize the probability that the fluid will be

Fig. 6 (a) Post-processing sequence of a molded PC chip (R reservoir); and (b) SEM image of a laser-drilled hole

30

evenly distributed to each branch in terms of flow rate and have the same travel time. In reality, because of manufacturing tolerances including LAMI fabrication, micro molding, and thermal fusion bonding, this was not true. The flow in the entire chip is laminar, so the relationship between the pressure drop and flow rate is linear, with the proportionality factor representing the hydraulic resistance. The hydraulic resistance is very sensitive to the variability of the channel cross sectional area, and above all the narrowest dimension of the cross section. The manufacturing variability for the height (50 μm) of the distribution channels used was estimated, based on measurements, to be ±5%. For the width, which is larger, the percentage variation was smaller. A simple calculation of the hydraulic resistance using well-known laminar solutions (White 1974) and propagation of the dimensional variation leads to a corresponding variation in the hydraulic resistance of each channel branch is on the order of ±15%. The hydraulic resistance of the capture beds was estimated by treating each one of them as a porous medium and using a simple model based again on the laminar flow channel solution (Bear 1988). Using such a model, variability of the hydraulic resistance of the capture bed was most sensitive to the height of the flow path (the net flow width is much larger—on the order of 500 μm considering the 1 mm total width of the capture bed and assuming 50% average blockage). The expected variation of the hydraulic resisFig. 7 Photographs of the leakage test: (a) An overall view of an assembled PC chip with fluorescein-filled microchannels pushed through PMMA connectors; and (b) a close-up view at 20× magnification showing no leakage. Photographs of (c) a PPC SPRI chip; and (d) gDNA purification setup (P1 and P2: microfluidic control ports)


tance of the capture bed was of the same order as for the straight channels. Given the estimated variation of the hydraulic resistance of each element, and the number of channel splits in the distribution system network necessary to deliver fluid to the 96 different capture beds, an estimate of the variation of the flow rate (pulling or pushing) through a typical capture bed would be on the order of 37%; the hydraulic resistance variability multiplied by the square root of the number of splits. Through similar considerations, the variation of the linear velocity of the fluid in each capture bed was approximated as ±24%. Variability in the filling and emptying time should also be of that order. For a total flow rate of 2 ml/min, the estimated pressure drop through the distribution system up to a typical capture bed and under ideal conditions was ∼169 kPa, while through the well itself it was ∼10 kPa, for a total pressure drop of ∼179 kPa. These estimates were made using the laminar flow solutions mentioned above (White 1974) and include estimates of losses because of hydraulic elements such as tees and bends (White 1986). The chip is initially full of air, so when it is filled for the first time, the effect of the variation in the capillary resistance on the flow rates (and velocities) of the fluid in each branch should be considered. This is much more difficult to estimate, because it is sensitive to the variation of the smallest dimension of the cross section as well as the


wall-surface properties (e.g. roughness, wettability). The latter are very difficult to quantify. However, an estimate of the contributing capillary pressure inside the smaller distribution channels was ∼1.5–2.0 kPa, while in the capture bed between posts it was ∼8 kPa. These values, obtained by using the Young–Laplace equation, were quite small compared to the total pressure drop, which builds up as the vias fill, and should not affect the filling rate of the capture bed significantly. Experimental observations confirmed that the 96 capture beds had different filling times even though the nominal distances from port to reservoirs are designed to be all the same. For example, the measured fluid volume for deionized water pushed into the 96 beds was 15.4±5.2 μl, 34% RSD. The 96 capture beds also exhibited different pulling speeds (by vacuum pump) at different capture beds. The observed variation of the flow during initial aspiration of the nucleic acid samples through the immobilization beds did not influence the amount of nucleic acids captured. For a given sample and immobilization buffer, nucleic acid capture efficiency depends primarily on the capture bed characteristics, including the active surface area and the interpost distance. The uniformity of the flow during the final elution of nucleic acids influences, however, the final concentration of the nucleic acids (ng/ml) in the eluted solution, which can be adjusted if necessary for downFig. 8 Agarose gel electrophoresis images of the PCR products generated from 96 amplification reactions with purified (a) B. subtilis, (b) S. aureus, and (c) E. coli gDNA using the second version of the 96-well SPRI microfluidic platform. Lane m represents DNA sizing ladder (50, 150, 300, 500, 750, 1,000 bp)

31

stream applications. The observed nonuniformity of the flow for different capture beds does not affect the function of the device for high throughput purification of nucleic acids. The fluidic control ports P1 and P2 can be complementary to each other, so the syringe and vacuum pumps can be interchanged without altering the overall fluidic performance. A series of fluidic performance tests were carried out by interchanging the fluidic control ports. The variability in the filling and emptying times for each capture bed showed similar trends in either configuration. 3.3 Purification of genomic DNA from whole cell lysates The performance of the 96-well SPRI reactor using photoactivated polycarbonate (PPC) chips [see Fig. 7(c),(d)] was tested for the purification of genomic DNA (gDNA) from different bacterial species (Bacillus subtilis, Staphylococcus aureus, and Escherichia coli). A simple and fast purification protocol for the chip was developed and the quality of the purification verified with PCR and electrophoresis (Witek et al. 2006; Xu et al. 2003; Park et al. 2007). Whole cell lysates of several bacterial species (B. subtilis, S. aureus, and E. coli) in an immobilization buffer (3% PEG, 0.4 M NaCl) were introduced into each of the 96 sample reservoirs and drawn through the capture beds to

32

port P2 by using vacuum. Upon completion of sample immobilization, ethanol was pushed through the extraction beds from port P1 to the sample reservoirs and pulled out through port P2 to remove cell debris and proteins. Captured DNA was released from the PPC surface by pushing DI water from port P1 to the sample reservoirs, where it was collected in 96 micro-tubes. This unique operating procedure allowed effective gDNA purification in a high throughput, automated format with a closed architecture that could eliminate potential contamination. PCR was performed on the isolated gDNA samples using 1 μl (∼10 ng) of the SPRI/PPC purified material. The PCR products were electrophoresed on an ethidium bromide stained 3% agarose gel. Figure 8 shows the fluorescence image of the PCR amplicons generated during the amplification reaction with a DNA sizing ladder in lane m. Purified gDNA was clearly evident from the 159, 204, and 600 bp PCR amplicons in lanes a–c for the B. subtilis, S. aureus, and E. coli, respectively. Successful capture and purification of gDNA was obtained at all 96 capture beds in the 96-well PPC SPRI chip, demonstrating its high throughput purification capability.

4 Conclusions A titer plate-based polymer microfluidic platform was demonstrated by incorporating functional microfluidic devices at each well location to realize high throughput purification of nucleic acids. To demonstrate the potential of this format, a 96-well solid-phase reversible immobilization (SPRI) reactor was designed to fit the footprint of a titer plate; the prototype allowed simultaneous fluidic control of samples at 96 separate immobilization beds using only two control ports. By using the standard titer plate format, the system was compatible with multichannel pipettes or robotic equipment for loading samples and reagents. The device was replicated from a 150 mm diameter LAMI fabricated using a SU-8 based UV-LIGA technique. Excellent lithography results for the SU-8 electroplating templates, particularly the vertical sidewalls, were obtained over the entire 150 mm diameter surface area by using flycutting to reduce the SU-8 film total thickness variation to less than 5 μm and exposure through an UV filter to remove absorbance of UV wavelengths below 350 nm. A custom-designed electroplating apparatus was then used to effectively control overplating of nickel and to fabricate high quality, high precision, metal mold inserts. Hot embossing of PC was demonstrated resulting in good replication fidelity over the large surface area. The micro molded PC chips had high integrity in all feature locations, particularly for the required well-to-well spacing of 9 mm that would allow the platform to take advantage of


existing sample and reagent liquid handling technologies. Thermal fusion bonding of the hot embossed PC chips at 150°C yielded good sealing, which was verified by leakage testing using fluorescence microscopy. All 96 capture beds in the SPRI reactor successfully purified gDNA from whole cell lysates of bacterial species in a highly parallelized fashion. Efforts are currently underway to optimize the molding process parameters in order to achieve uniform filling of the 10 μm, or smaller, microcavity arrays over the complete surface area of the LAMI using extensive simulation and molding tool design (Worgull and Heckele 2004; Worgull et al. 2005). Alternative polymer materials with better flow characteristics are also being explored to address this issue. The 96-well photo-activated PC SPRI reactor in the titer plate-based polymer microfluidic platform will open up new avenues for low cost, disposable DNA/RNA sample purification and many other options for high throughput analysis once other micro-analytical systems are implemented in the multi-well format. Acknowledgements This work was supported by the National Science Foundation and the State of Louisiana Board of Regents Support Fund under grant number EPS-0346411, and the State of Louisiana Board of Regents Support Fund, Industrial Ties Program through grant number LEQSF(2005-08)-RD-B-04. The authors thank the staff of the Center for Advanced Microstructures and Devices (CAMD) at Louisiana State University for the microfabrication support. J. Guy was funded by a Louisiana Governor’s Biotechnology Initiative grant.

References P.-A. Auroux, D. Iossifidis, D.R. Reyes, A. Manz, Anal. Chem. 74, 2637 (2002) P.-A. Auroux, Y. Koc, A. deMello, A. Manz, P.J.R. Day, Lab Chip. 4, 534 (2004) D.O. Barrett, A. Maha, Y. Wang, S.A. Soper, D.E. Nikitopoulos, M.C. Murphy, in MicroFluidics, BioMEMS, and Medical Microsystems II, vol. 5345, ed. by P. Woias, I. Papautsky. (SPIE, Bellingham, WA, 2004), p. 78 J. Bear, Dynamics of Fluids in Porous Media (Dover Publications, New York, NY, 1988) F. Darvas, G. Dormán, P. Krajcsi, L.G. Puskás, Z. Kovári, Z. Lörincz, L. Ürge, Curr. Med. Chem. 11, 3119 (2004) P. Datta, F. Xu, S. Gurung, S.A. Soper, J. Goettert, in Microfluidics, BioMEMS, and Medical Microsystems IV, vol. 6112, ed. by I. Papautsky, W. Wang (SPIE, Bellingham, WA, 2006), p. 611208 M.M. DeAngelis, D.G. Wang, T.L. Hawkins, Nucleic Acids Res. 23, 4742 (1995) T.L. Edwards, S.K. Mohanty, R.K. Edwards, C.L. Thomas, A.B. Frazier, Microfluidic Devices and Systems III, vol. 4177, ed. by C.H. Mastrangelo, H. Becker, (SPIE, Bellingham, WA, 2000), p. 75 C.J. Elkin, P.M. Richardson, H.M. Fourcade, N.M. Hammon, M.J. Pollard, P.F. Predki, T. Glavina, T.L. Hawkins, Genome Res. 11, 1269 (2001)

Biomed Microdevices (2008) 10:21–33 C.J. Elkin, H. Kapur, T. Smith, D. Humphries, M. Pollard, N. Hammon, T. Hawkins, BioTechniques, 32(6), 1296 (2002) D. Erickson, D. Li, Anal. Chim. Acta. 507, 11 (2004) J.D. Gelorme, R.J. Cox, S.A.R. Guterrez, US Patent No. 4882245 (21 Nov 1989) A. Gerlach, G. Knebel, A.E. Guber, M. Heckele, D. Herrmann, A. Muslija, Th. Schaller, Microsyst. Technol. 7(5–6), 265 (2002a) A. Gerlach, G. Knebel, A.E. Guber, M. Heckele, D. Herrmann, A. Muslija, Th. Schaller, Sens. Mater. 14(3), 119 (2002b) A.E. Guber, M. Heckele, D. Herrmann, A. Muslija, V. Salie, L. Eichhorn, T. Gietzelt, W. Hoffmann, P.C. Hauser, J. Tanyanyiwa, A. Gerlach, N. Gottschlich, G. Knebel, Chem. Eng. J. 101, 447 (2004) T.L. Hawkins, T.O’Connor-Morin, A. Roy, C. Santillan, Nucleic Acids Res. 22, 4543 (1994) M. Heckele, W.K. Schomburg, J. Micromech. Microeng. 14, R1 (2004) K. Honma, T. Miyata, T. Ochiya, Curr. Drug Discov. Technol. 1, 287 (2004) Y. Huang, E.L. Mather, J.L. Bell, M. Madou, Anal. Bioanal. Chem. 372, 49 (2002) M.L. Hupert, W.J. Guy, S.D. Llopis, C. Situma, S. Rani, D.E. Nikitopoulos, S.A. Soper, in Microfluidics, BioMEMS, and Medical Microsystems IV, vol. 6112, ed. by I. Papautsky, W. Wang (SPIE, Bellingham, WA, 2006), p. 61220B K.S. Kim, High Aspect Ratio Microstructures and Their Applications to MEMS, Ph.D. dissertation (The University of Texas at Dallas, Richardson, TX, USA, 2004) D.S. Kim, S.H. Lee, C.H. Ahn, J.Y. Lee, T.H. Kwon, Lab Chip. 6, 794 (2006) S.J. Lee, S.Y. Lee, Appl. Microbiol. Biotechnol. 64, 289 (2004) J. Liu, C. Hansen, S. Quake, Anal. Chem. 75, 4718 (2003) H. Lorenz, M. Despont, N. Fahrni, J. Brugger, P. Vettiger, P. Renaud, Sens. Actuators, A, Phys. 64(1), 33 (1998) H. Lu, B. Pillans, J.-C. Lee, K. Kim, J.-B. Lee, Microsyst. Technol. 13 (3–4), 237 (2007) D.A. Mair, E. Geiger, A.P. Pisano, J.M.J. Fréchet, F. Svec, Lab Chip. 6, 1346 (2006) H. Mekaru, T. Yamada, S. Yan, T. Hattori, Microsyst. Technol. 10, 682 (2004) M.W. Mitchell, X. Liu, Y. Bejat, D.E. Nikitopoulos, S.A. Soper, M.C. Murphy, in MicroFluidics, BioMEMS, and Medical Microsystems, vol. 4982, ed. by H. Becker, P. Woias (SPIE, Bellingham, WA, 2003), p. 83

33 Y. Murakoshi, C.S. Xue, T. Shimizu, R. Maeda, in 2003 Proceedings of the Fifth Symposium on Design, Test, Integration and Packaging of MEMS/MOEMS, ed. by B. Courtois, K. Markus, J. Korvink, J.M. Karam, K. Bergman, B. Michel, p. 304 P.J. Obeid, T.K. Christopoulos, H.J. Crabtree, C.J. Backhouse, Anal. Chem. 75, 288 (2003) B.M. Paegel, S.H.I. Yeung, R.A. Mathies, Anal. Chem. 74, 5092 (2002) R. Pal, M. Yang, R. Lin, B.N. Johnson, N. Srivastava, S.Z. Razzacki, K.J. Chomistek, D.C. Heldsinger, R.M. Haque, V.M. Ugaz, P.K. Thwar, Z. Chen, K. Alfano, M.B. Yim, M. Krishnan, A.O. Fuller, R.G. Larson, D.T. Burke, M.A. Burns, Lab Chip. 5, 1024 (2005) D.S. Park, M. Hupert, J. Guy, P. Datta, J.-B. Lee, M. Witek, B.H. You, S.A. Soper, D.E. Nikitopoulos, M.C. Murphy, in 2006 Proceedings of the ASME International Mechanical Engineering Congress & Exposition (ASME, New York, NY, 2006), IMECE200615275 D.S. Park, M.L. Hupert, M. Witek, J. Guy, P. Datta, B.H. You, S.A. Soper, M.C. Murphy, in the 20th International Conference on Micro Electro Mechanical Systems (MEMS 2007), p. 433 S. Qi, X. Liu, S. Ford, J. Barrows, G. Thomas, K. Kelly, A. McCandless, K. Lian, J. Goettert, S.A. Soper, Lab Chip, 2(2), 88 (2002) D.R. Reyes, D. Iossifidis, P.-A. Auroux, A. Manz, Anal. Chem. 74 (12), 2623 (2002) Society for Biomolecular Screening (SBS), Microplate Standards Development Committee, ANSI/SBS 1-2004 (SBS, Microplate Standards Development Committee, Danbury, CT, 2004) F.M. White, Viscous Fluid Flow (McGraw-Hill, New York, 1974) F.M. White, Fluid Mechanics (McGraw-Hill, New York, 1986) M.A. Witek, S. Wei, B. Vaidya, A.A. Adams, L. Zhu, W. Stryjewski, R.L. McCarley, S.A. Soper, Lab Chip. 4, 464 (2004) M.A. Witek, S.D. Llopis, A. Wheatley, R.L. McCarley, S.A. Soper, Nucleic Acid Res. 34(10), e74 (2006) M. Worgull, M. Heckele, Microsyst. Technol. 10(5), 432 (2004) M. Worgull, M. Heckele, W.K. Schomburg, Microsyst. Technol. 12 (1–2), 110 (2005) Y. Xu, B. Vaidya, A.B. Patel, S.M. Ford, R.L. McCarley, S.A. Soper, Anal. Chem. 75(13), 2975 (2003) B. Ziaie, A. Baldi, M. Lei, Y. Gu, R.A. Siegel, Adv. Drug Deliv. Rev. 56, 145 (2004)