Anal. Chem. 2009, 81, 2904–2911
Microsystem for Field-Amplified Electrokinetic Trapping Preconcentration of DNA at Poly(ethylene terephthalate) Membranes Thomas Hahn,†,‡ Ciara K. O’Sullivan,*,‡,§ and Klaus S. Drese*,† Fluidics and Simulation, Institut fu¨r Mikrotechnik Mainz GmbH, Carl Zeiss Strasse 18-20, 55129 Mainz, Germany, Nanobiotechnology and Bioanalysis Group, Department of Chemical Engineering, University of Rovira i Virgili, Av. Paisos Catalan, 26, 43007, Tarragona, Spain, and Institucio´ Catalana de Recerca i Estudis Avanc¸ats, Passeig Lluı´s Companys 23, 08010 Barcelona, Spain In electrokinetic trapping (EKT), the electroosmotic velocity of a buffer solution in one area of a microfluidic device opposes the electrophoretic velocity of the analyte in a second area, resulting in transport of DNA to a location where the electrophoretic and electroosmotic velocities are equal and opposite and DNA concentrates at charged nanochannels. The method does not require an optical plug localization, a considerable advantage as compared to preconcentration techniques previously presented. In the work reported here, the trapping process is preceded by a field-amplification in the sample reservoir to reduce trapping time, as field-amplified EKT is shown to be an effective technique to preconcentrate samples from larger volumes. A theoretical model explaining the principle of field-amplified EKT considers different ionic strengths and cross-sectional areas in the microchip segments. The model is supported by experimental data using nucleic acids and fluorescein as sample analytes. An incorporated poly(ethylene terephthalate) (PET) membrane provides anion exclusion due to a negatively charged surface. A fluidic counter flow supports the trapping process in 100 nm pores due to anion exclusion. An analysis of Joule heating gives evidence that temperature gradient focusing effects are negligible and charge exclusion is responsible for trapping. The theoretical model developed and experimentally demonstrated can be exploited for the preconcentration of cell free fetal DNA circulating in maternal plasma and other rare nucleic acid species present in large sample volumes. Developments in the field of microsystems over the past decade, to meet the requirements of medical applications1 in terms of sensitivity, reliability, and cost, have positioned them as highly promising tools for point-of-care (POC) diagnostics.2-4 The * To whom correspondence should be addressed. E-mail: drese@ imm-mainz.de (K.S.D.);
[email protected] (C.K.O’S.). † Institut fu ¨ r Mikrotechnik Mainz GmbH. ‡ University of Rovira i Virgili. § Institucio´ Catalana de Recerca i Estudis Avanc¸ats. (1) Verpoorte, E. Electrophoresis 2002, 23, 677–712. (2) Gulliksen, A.; Solli, L. A.; Drese, K. S.; Sorensen, O.; Karlsen, F.; Rogne, H.; Hovig, E.; Sirevag, R. Lab Chip 2005, 5, 416–420. (3) Lagally, E. T.; Medintz, I.; Mathies, R. A. Anal. Chem. 2001, 73, 565–570.
2904
Analytical Chemistry, Vol. 81, No. 8, April 15, 2009
increasing demands of POC diagnostics, where ever decreasing amounts of target analyte are required to be detected, dictates that miniaturization to a micro/nanoliter scale necessitates efficient preconcentration prior to analysis. Electrokinetic-driven methods dominate in the field of preconcentration in microchips.5 The most frequently used techniques are field-amplified sample stacking (FASS),6 isotachophoresis (ITP),7 electrokinetic trapping (EKT),8-11 and temperature gradient focusing (TGF).12,13 In FASS, a sample is prepared in an electrolyte with a lower ionic strength than the surrounding electrolyte. The gradient in electric conductivity accelerates the sample zone to stack in front of the surrounded electrolyte. In ITP, the sample ions are surrounded by a trailing electrolyte and a leading electrolyte. The sample co-ions of the trailing electrolyte have a lower electrophoretic mobility than the charged sample molecules. The sample coions of the leading electrolyte have a higher electrophoretic mobility than the sample molecules. This leads to an alignment of sample ions in the zone of their specific electrophoretic mobility.7 FASS and ITP require a localization of sample molecules, while EKT does not, as preconcentration is achieved in front of a fixed nanostructure. Moreover ITP and single FASS applications need to employ long channels to preconcentrate larger volumes, which necessitate narrower bending structures to reduce band broadening.14 TGF and EKT achieve preconcentration in a defined zone and can preconcentrate larger sample volumes. The disadvantage of TGF is the difficulty of the required use of heating (4) Tu ¨ dos, A. J.; Besselink, G. A.; Richard, B. M.; Schasfoort, R. B. Lab Chip 2001, 1, 83–95. (5) de Mello, A. J.; Beard, N. Lab Chip 2003, 3, 11N–19N. (6) Jung, B.; Bharadwaj, R.; Santiago, J. G. Electrophoresis 2003, 24, 3476– 3483. (7) Timerbaev, A. R.; Hirokawa, T. Electrophoresis 2006, 27, 323–340. (8) Dai, J.; Ito, T.; Sun, L.; Crooks, R. M. J. Am. Chem. Soc. 2003, 125, 13026– 13027. (9) Dhopeshwarkar, R.; Crooks, R. M.; Hlushkou, D.; Tallarek, U. Anal. Chem. 2008, 80, 1039–1048. (10) Hlushkou, D.; Dhopeshwarkar, R.; Crooks, R. M.; Tallarek, U. Lab Chip 2008, 8, 1153–1162. (11) Zhou, K.; Kovarik, M. L.; Jacobson, S. C. J. Am. Chem. Soc. 2008, 130, 8614–8616. (12) Hoebel, S. J.; Balss, K. M.; Jones, B. J.; Malliaris, C. D.; Munson, M. S.; Vreeland, W. N.; Ross, D. Anal. Chem. 2006, 78, 7186–7190. (13) Munson, M. S.; Danger, G.; Shackman, J. G.; Ross, D. Anal. Chem. 2007, 79, 6201–6207. (14) Griffiths, S. K.; Nilson, R. H. Anal. Chem. 2000, 72, 5473–5482. 10.1021/ac801923d CCC: $40.75 2009 American Chemical Society Published on Web 03/18/2009
blocks.12,13 Moreover, TGF is restricted to the preconcentration of analytes by changing their effective electrophoretic mobility with pH and does not preconcentrate nucleic acids. Taking this into consideration, EKT is the most relevant choice for microchip preconcentration of samples larger than a few microliters that contain nucleic acids. A multitude of reports regarding EKT phenomena15-17 offering explanations of the hindrance of charged molecules entering nanometer-sized pores or channels by charge exclusion have appeared. Sample ions have been preconcentrated in front of nanopores of a hydrogel plug,9 an etched nanochannel,18 or a membrane8,11 where a concentration polarization in the electrolyte is exploited.19 The pore sizes used range between 2 to 200 nm and are always larger than the sample molecule, thus excluding a mechanical trapping phenomenon. The concentration polarization zone which causes a localized increase in the specific conductivity on the cathodic side of a cation-selective membrane imbedded in a microchannel decelerates the sample ions supporting preconcentration.19 Vice versa, a depletion of anions occurs on the anodic side of the membrane resulting in an accompanied depletion of cations. The stability of EKT is dependent on various factors such as pore size, charge density of the nanostructure, and a counteracting flow.9 Usually, etched nanochannels or membranes with >50 nm pore size only offer incomplete charge exclusion requiring a fluidic counterflow. The principal problem in stabilizing a preconcentration zone using EKT has been an insuperable obstacle when using sample volumes greater than a few microliters. In the work reported here, we take a rational approach to enhance the efficiency of EKT using PET membranes, where we develop a theoretical model to identify the critical parameters involved in EKT at 100 nm pores. A number of reports have demonstrated the negative charge of PET, measuring the ζ potential using streaming potential measurements, and this counterintuitive observation has been explained by several groups.20-27 Using the identified critical parameters, we experimentally demonstrate the efficiency of trapping DNA and fluorescein at a PET membrane using EKT within a polymeric microfabricated device and explain the different phenomena observed with each type of molecule. Additionally, we analyze the conductance regarding different applied potentials to exclude a temperature gradient focusing effect. The developed microdevice has tremendous application for the concentration of DNA/mRNA, particularly for the preconcentration of cell-free circulating nucleic (15) (16) (17) (18)
(19) (20) (21) (22) (23) (24) (25) (26) (27)
Plecis, A.; Schoch, R. B.; Renaud, P. Nano Lett. 2005, 5, 1147–1155. Pu, Q.; Yun, J.; Temkin, H.; Liu, S. Nano Lett. 2004, 6, 1099–1103. Ho ¨ltzel, A.; Tallarek, U. J. Sep. Sci. 2007, 30, 1398–1419. Pennathur, S.; Baldessari, F.; Kattah, M.; Utz, P. J.; Santiago, J. G. 2006 ASME Joint U.S.-European Fluids Engineering Summer Meeting, Miami, FL, July 17-20, 2006. Leinweber, F. C.; Tallarek, U. J. Phys. Chem. B 2005, 109, 21481–21485. Benavente, J.; de Lara, R. Port. Electrochim. Acta 2007, 25, 79–88. Berezkin, V. V.; Volkov, V. I.; Kiseleva, O. A.; Mitrofanova, N. V.; Sobolev, V. D. Adv. Colloid Interface Sci. 2003, 104, 325–331. Bianchi, F.; Wagner, F.; Hoffmann, P.; Girault, H. H. Anal. Chem. 2001, 73, 829–836. Kirby, B. J.; Hasselbrink, E. F. Electrophoresis 2004, 25, 203–213. Tatchou-Nyamsi-Ko ¨nig, J.-A.; Dague, E.; Mullet, M.; Duval, J. F. L.; Gaboriaud, F.; Block, J.-C. Water Res. 2008, 42, 4751–4760. Jones, C. R.; Adams, M. R.; Zhdan, P. A.; Chamberlain, A. H. L. J. Appl. Microbiol. 1999, 86, 917–927. Beattie, J. K. Lab Chip 2006, 6, 1409–1411. Zangi, R.; Engberts, B. F. N. J. Am. Chem. Soc. 2005, 127, 2272–2276.
Figure 1. Schematic illustration of the microchip for preconcentration showing the cathode reservoir (1), anode reservoir (2), sample chamber (3), membrane chamber (4), sample inlet (5), and anode channels (6).
acids, which are being demonstrated to have increasing clinical relevance for early cancer, and noninvasive prenatal, diagnosis.28 EXPERIMENTAL SECTION Material and Reagents. Poly(ethylene terephthalate) membranes of 100, 200, 800, and 5000 nm pore size were purchased from Sterlitech Corp. (WA). DNA fragments of 230-510 bp were generated by PCR using commercially available pGEM Vector as template DNA (Promega GmbH, Germany). PCR reagents were purchased from Qiagen (Hilden, Germany) and single-stranded nucleotides (nt) from Metabion GmbH (Germany). The PCR mixture was prepared with the following concentrations of reagents: 1× PCR buffer, 200 µM of each dNTP, 0.2 µM of each primer, 0.2 units Taq polymerase in 25 µL total volume. The salt concentration in the PCR products has been declined by ethanol precipitation (70% v/v) at -20 °C overnight, centrifuging, and resuspending in deionized water. SYBRGreen I purchased from Invitrogen served for DNA-staining in EKT experiments and used at concentrations as recommended by the supplier. Fluorescein was prepared in deionized water to 0.138 ng µL-1 from fluorescein disodium salt of normal grade (Fluka, Germany). 2-Amino-2(hydroxymethyl)-1,3-propanediol (Tris) and thymol blue was distributed by Sigma Aldrich (Germany) in pure grade. Glycine (g99%) was purchased from Carl Roth GmbH (Germany). Chip Fabrication and Assembly. Poly(methyl methacrylate) (PMMA) 64 × 43 × 2 mm3 templates were micromachined and cleaned once with isopropanol, 0.5 M sodium hydroxide, and deionized water for 3 min each in ultrasound. PET membranes with 100-5000 nm pore size were incorporated into the chamber (Figure 1, part 4) by thermal bonding using a dedicated welding apparatus. The PET membrane was heated with a ring structure by applying 160 °C for 1-2 s, which offered advantages over etching methods or gel slugs in its simplicity of use.9,18 Moreover, PET had the advantage of exhibiting a highly electronegative character providing exclusion of anions.20,22,23 A perforation to the other side of the chip connected the membrane chamber with the anode channels. A second perforation guided the anode channels back to the previous chip side (Figure 1, part 6), and the chip was sealed on both sides with adhesive polymer foil. (28) Bianchi, D. W. Placenta 2004, 18, S93-S101.
Analytical Chemistry, Vol. 81, No. 8, April 15, 2009
2905
The process of micromachining facilitated a high accuracy of channel depth but still required a verifying measurement of the anodic compartment to obtain equal counterflow velocities in the anode channels (100 × 350 µm2). Current measurement was a reliable parameter for measuring the reproducibility of the chip dimensions and a reproducible counterflow induced by the negative wall potential of the used PMMA material. Plastic polymers provided a ζ potential23 lower in its absolute value compared to, for example, silica, but these polymers are easily endued with channels on both chip sides. This advantage provided a useful solution for the incorporation of a membrane that was welded onto the chamber. The anode channels were flushed with 5 mM Tris, 27.5 mM glycine (pH of 8.5 at 25 °C) from the anode reservoir (Figure 1, part 2). During the filling process, the buffer entered the sample reservoir and was released through the cathode reservoir (Figure 1, part 1). The sample reservoir was emptied and subsequently filled with 80 µL of deionized water containing nucleic acids of a 230 bp PCR product or fluorescein from an inlet (Figure 1, part 5). After the filling process, the sample contained
