Protein-Assisted Stretching and Immobilization of DNA Molecules in a

Venkat ram Dukkipati,†,§ Ji Hoon Kim,‡,§ Stella W. Pang,† and Ronald G. Larson*,‡. Department of Electrical Engineering and Computer Science...
0 downloads 5 Views 381KB Size
Download PDF
NANO LETTERS

Protein-Assisted Stretching and Immobilization of DNA Molecules in a Microchannel

2006 Vol. 6, No. 11 2499-2504

Venkat ram Dukkipati,†,§ Ji Hoon Kim,‡,§ Stella W. Pang,† and Ronald G. Larson*,‡ Department of Electrical Engineering and Computer Science and Department of Chemical Engineering, UniVersity of Michigan, Ann Arbor, Michigan 48109 Received July 26, 2006; Revised Manuscript Received September 12, 2006

ABSTRACT We demonstrate a novel technique we call “protein-assisted DNA immobilization” (PADI), to immobilize and stretch, but not overstretch, DNA molecules inside a micro/nanochannel with limited surface interactions while maintaining continuous hydration at physiological pH. The biological activity of the immobilized DNA molecules is confirmed by digesting the DNA with restriction enzymes in the microchannel. Single-molecule transcription, which has stringent requirements on the immobilized DNA with respect to surface interactions and stretched lengths, is also successfully demonstrated on DNA molecules immobilized by PADI. In addition to arraying DNA molecules for study of DNA−protein interactions, the immobilization method could be used to construct DNA-templated nanoelectronic devices.

Studies of DNA-protein interaction at the single molecule level are increasingly used to quantify distributions of molecular mechanical properties, transient intermediates, and reaction pathways. Such studies require stretching and immobilization of DNA molecules to suppress Brownian motion so that protein motion along the DNA contour can be followed. Often in single-molecule experiments, DNA molecules are immobilized at one end to a substrate while the other end is manipulated with optical1-3 or magnetic tweezers,4,5 laminar flows,6-8 or electric fields.9,10 DNA molecules can also be stretched and permanently immobilized to the surface along their backbone by the motion of an airwater interface.11,12 This so-called “molecular combing” method is known to produce highly overstretched DNA molecules in which the bases are unstacked into a flat parallel ladder.13-16 Apart from arraying DNA molecules for studying DNAprotein interactions at the single-molecule level, DNA stretching and immobilization is also a critical step toward constructing DNA-templated nanostructures17 and manipulation of nucleotides in single DNA molecules for genetic applications.18 DNA, with its 2 nm diameter and length on the order of micrometers, is becoming an important component in molecular nanotechnology, such as in DNAtemplated nanowires for applications in sensors, electronics, and optoelectronics.19,20 The ability to interrogate single DNA * To whom correspondence should be addressed. Phone: (734) 9360772. Fax: (734) 763-0459. E-mail: [email protected]. † Department of Electrical Engineering and Computer Science. ‡ Department of Chemical Engineering. § The first two authors contributed equally to this work. 10.1021/nl0617484 CCC: $33.50 Published on Web 09/27/2006

© 2006 American Chemical Society

molecules also may lead to applications in sequencing technology, potentially eliminating costly and often problematic procedures such as cloning and PCR amplification.21 One such single-molecule technology is optical mapping, which is based on the measurement of fragment lengths of surface-deposited DNA molecules after digestion with restriction enzymes.22,23 In another recently demonstrated application, single-base-pair mismatch was detected by the change in conduction of a stretched double-stranded DNA (dsDNA) molecule immobilized across two gold electrodes.24 These single-molecule technologies could benefit from high throughput alignment and immobilization of DNA molecules. Of the many existing methods of stretching and anchoring DNA molecules, molecular combing is widely used to align large numbers of DNA molecules on a hydrophobic surface. However, several drawbacks of the method include the aforementioned overstretching of the molecule, potential interference created by the surface to interactions of the DNA with other molecules, dehydration of the DNA as the meniscus moves across it, and the low pH (∼5.5 to 6.6) of the DNA solution necessary to promote efficient DNA binding to the surface. It has been reported that in some cases proteins show limited, if any, catalytic activity on overstretched DNA molecules.16 In the present study, we report a novel technique for DNA immobilization that we call “protein-assisted DNA immobilization” (PADI), which consists of the following. DNA binding proteins such as restriction enzymes or RNA polymerases (RNAP) are allowed to bind to the DNA

Figure 1. (A) Schematic cut-away representation of the cross-sectional area of the microchannel (not drawn to scale). (B) Schematic showing top and bottom surfaces and objective used for imaging (not drawn to scale). (C) Images of λ-DNA (5.5 pM) stretched onto a PMMA-coated glass in the microfluidic device (100 µm width, 1 µm depth) in the presence of T7 RNAP (10 nM) at 200 µm from the inlet and (D) near the inlet. The direction of flow is from right to left.

molecule in bulk solution at nonspecific segments, whereupon they diffuse along the DNA in search of target sequences at a rate faster than in a three-dimensional diffusion-limited search.25,26 Whether or not the protein finds a specific binding site on the particular DNA strand used, the result is a DNA-protein complex with multiple proteins of ∼5 nm in size wrapped around each DNA molecule. When this DNA-protein complex is subjected to a hydrodynamic flow inside a channel shown in Figure 1A, it stretches out as depicted in the enlargement in Figure 1B. When the DNA-protein complex reaches the channel surface, the proteins on the DNA adsorb to the surface, resulting in immobilization of the stretched DNA molecule inside the channel. While protein adsorption onto surfaces is a complex phenomenon,27,28 it has been reported that hydrophobic surfaces adsorb more proteins than hydrophilic ones.29 We use a hydrophobic polymethylmethacrylate (PMMA) surface inside our microchannel, leading to a large number of immobilized DNA molecules (Figures 1C and 1D). Several advantageous features of PADI are as follows: (1) overstretching of DNA molecules is avoided; (2) the degree of attachment of DNA to the substrate can be controlled by changing the protein concentration without changing the substrate material; (3) the number of DNA molecules immobilized onto the substrate is time and concentration 2500

dependent and can be controlled simply by varying the pumping time as well as the concentration; and (4) the stretching and immobilization is achieved at physiological pH. Here we demonstrate the PADI technique in a microfluidic device. Microfluidic devices reduce the cost of running assays, decrease procedural times, and limit the required concentration and hands-on manipulation of samples, as compared to conventional flow cells.30,31 It has been reported that laminar flow in microchannels offers improved enzymatic reactions, PCR, and hybridization.32 The PADI technique preserves biological properties of stretched DNA molecules which can be used in the field of self-assembly based biomolecular nanotechnology, where specific recognition of DNA sequences is used to assemble different structures. For example, using bottom-up nanotechnology, a carbon nanotube field-effect transistor can be constructed by using stretched DNA as a template.33 Our microfluidic system consists of microchannels etched in Si and bonded to glass using PMMA as an adhesive (Figure 1A). A channel in Si wafer is dry etched to a width of 100 µm and depth of 1 µm. An opening of 300 µm in diameter and 550 µm in depth is etched at each side of the straight microchannel using a Si plasma etching tool. The Si wafer with etched channels and inlet/outlet openings is cleaned in 1:1 H2O2:H2SO4 before bonding. The cleaning Nano Lett., Vol. 6, No. 11, 2006

Figure 2. (A) Sequence of images of DNA stretching and immobilization by PADI. T7 DNA molecules (6.7 pM) were mixed with T7 RNAP (5 nM) and then introduced into the microchannel. The direction of fluid flow is indicated by an arrow. (B) A close-up showing a DNA molecule adsorbing at one point (indicated by white arrows) followed by complete adsorption of the molecule. Scale bar ) 5 µm.

step forms hydroxyl groups on the Si surface, making the channel more hydrophilic and less likely to adsorb proteins. To adsorb proteins and create an adhesion layer for bonding, PMMA of 600 nm thickness is spin-coated onto the 100 µm thick glass. Due its high viscosity, PMMA does not flow into the Si channels during the bonding process. The bonding is performed at 110 °C for 15 min at a pressure of 0.4 MPa. The sealed channels are 4 mm long with three sidewalls made of hydrophilic Si surfaces with the hydroxyl groups and the fourth sidewall consisting of the hydrophobic PMMA surface. The DNA-protein complex is introduced at one end of the channel using a pipet. The fluid is immediately sucked into the channel by capillary action (Figure 2A). The inlet end of the channel acts as a reservoir for fluid entering into the channel and the outlet end is open to the atmosphere for fluid to evaporate (Figure 1A). The continuous evaporation of the fluid from the outlet opening drives the fluid from the reservoir into the channel, which results in a continuous flow of the DNA-protein complex into the channel.34 The evaporation-driven flow inside the microchannel stretches the DNA-protein complex and transports it to the PMMA surface where it initially adsorbs at a single point followed by complete attachment (Figure 2B). The DNA is initially out of focus because it is attached at one end to the PMMA surface while the rest of the backbone is in the solution (second image in Figure 2B), where it cannot be seen easily with total internal reflection fluorescence microscopy (TIRFM) (TIRF imaging relies on the evanescent wave that decays exponentially with distance, limiting the observation region to ∼100 nm from the surface). The DNA comes into focus once it is completely adsorbed to the surface (third image in Figure 2B). We have successfully stretched and immobilized DNA molecules in the presence of RNA polymerase, DNA polymerase, topoisomerase, and restriction enzymes includNano Lett., Vol. 6, No. 11, 2006

ing EcoR I, EcoR V, Hind III, BamH I, and Sma I. Figure 3A shows T7 DNA molecules immobilized to the PMMAcoated glass in the microchannel in the presence of T7 RNAP labeled with fluorescent antibodies. We note that the DNA does not need to have a target sequence that the protein recognizes for the immobilization to occur; λ-DNA, which lacks a T7 promoter sequence, can be stretched and immobilized in the presence of T7 RNAP (Figures 1B and 1C). We could not immobilize DNA molecules with non-DNA binding proteins such as bovine serum albumin (BSA) and secondary antibodies, or without any proteins. The number of molecules adsorbed onto the PMMAcoated glass surface was counted at various positions in the microchannel (Figure 3B). Note that approximately 40% of adsorbed DNA molecules are found within 86 µm (the width of field of view when a 100× objective is used) from the inlet for the 1 µm deep channel. Such a high percentage of molecules is found near the inlet because the channel is initially dry and the number of DNA molecules adsorbed after the fluid is introduced into the channel exhibits an exponential decay with distance (Figure 3B), presumably because the molecules are depleted from the solution by adsorption near the entrance region. Using diffusion-limited adsorption to the surface, the distance downstream that the DNA molecules will travel before being adsorbed should be of order Vd2/2D, where V is the fluid velocity, d is the depth of the channel (1 µm), and D is the DNA diffusivity (∼0.5 µm2/s). Thus at the fluid velocity of 200 µm/s, which is measured by tracking 20 nm fluorescent nanoparticles near the center of the channel, the DNA adsorption decay distance is on the order of 200 µm. This agrees well with the decay distance of 183 µm that we have obtained by fitting the number of adsorbed DNA molecules as a function of distance to a single-exponential fit (Figure 3B). Since the fluid velocity scales with d-1 at constant evaporation rate, the DNA adsorption decay distance is directly proportional to 2501

Figure 3. (A) T7 DNA molecules (green, 0.7 pM) stretched and immobilized with the assistance of T7 RNAP (red, 3 nM). Arrows indicate positions of bound T7 RNAPs. Scale bar ) 2.5 µm. (B) The number of λ-DNA (9.2 pM) adsorbed in the presence of T7 RNAP (1.7 nM) as a function of distance from the channel inlet is fitted to a single exponential (n ) N0e-x/D), yielding an adsorption decay distance of D1 ) 183 ( 18 µm in a 1 µm deep channel and D3 ) 619 ( 59 µm in a 3 µm deep channel (inset). The flow was stopped 2 min after the introduction of the DNA solution by hydrating the other end of the microchannel. (C) λ-DNA molecules (6.7 pM) stretched and immobilized with the assistance of T7 RNAP (0.5 nM) in a series of parallel 100 nm deep and 350 nm wide channels. Dashed lines approximately represent channel sidewalls which are omitted in other areas in the figure for clear representation. (D) The mean DNA stretch ratio is plotted as a function of distance in a 1 µm deep channel.

d. The DNA adsorption decay distance increases to 619 µm when the depth of the channel is tripled (Figure 3B inset), suggesting that the scaling is valid within the experimental error. We also have immobilized DNA molecules in channels whose dimensions (100 nm depth, 350 nm width) are smaller than the radius of gyration of λ-DNA (∼500 nm). The number of DNA molecules immobilized is less in these channels than those found in 1 µm and 3 µm deep channels because the DNA molecules are entropically blocked from entering the channel, due to the small channel size (Figure 3C). By using channels with step changes in height, one could exclude large molecules from the channel, and in a downstream process, bleed away small DNA molecules, and thereby concentrate molecules of some desired size.35 As mentioned above, we rarely observe overstretched DNA molecules with the PADI technique. The maximum value of the stretch ratio, which is defined as the stretch length divided by the contour length (x/L), is 1.2, and only ∼2% of the adsorbed DNA molecules had a stretch ratio greater than unity. The mean stretch ratio of DNA molecules 2502

adsorbed within 86 µm (the width of field of view) from the inlet is 0.56 for 1 µm deep channel (Figure 3D). At this stretch ratio, the force (F) stretching the DNA is expected to be ∼0.1 pN according to the worm-like chain model by Marko and Siggia,36 for which FP/kBT ) 1/4(1 - x/L)-2 - 1/4 + x/L where P is the persistence length (∼66 nm for YOYO-stained λ-DNA). This is much weaker than the force needed to overstretch DNA molecules (∼65 pN).15,16 With the relaxation time of ∼0.1 s for DNA in water and the average fluid velocity of 200 µm/s, the DNA chains should relax significantly by the time they travel 20 µm into the channel, assuming that they are most strongly stretched by the extensional flow at the entrance to the channel. After entering the channel, the extensional flow is replaced by simple shear, which is not as effective in stretching the DNA.37 Therefore the mean stretch ratio should decrease with distance from the inlet, as we observed (Figure 3D). Nano Lett., Vol. 6, No. 11, 2006

Figure 4. T7 DNA molecules were immobilized at (A) 5 nM and (B) 0.5 nM T7 RNAP followed by DNA photocleavage by exposure to illumination. After photocleavage, DNA fragments coil back to the many or few attachment points provided by the adsorbed proteins; the coiled DNA fragments can be observed as bright spots on the surface in B.

The degree of attachment along the DNA backbone to the surface can be controlled by varying the protein concentration. Increasing the protein concentration increases the number of proteins interacting with the DNA within its radius of gyration, leading to more attachment points on the PMMA-coated glass. Figure 4 shows the result of DNA attachment in a microchannel for two different RNAP concentrations. For an RNAP concentration of 5 nM, the DNA remains attached at multiple points after photocleavage for 120 s (Figure 4A), indicating that the DNA was firmly fixed to the PMMA-coated glass at multiple places. In the case of 0.5 nM RNAP, there are fewer attachment points (Figure 4B). Depending on the application, different degrees of DNA attachment can be obtained. For example in optical mapping, the DNA has to be tightly bound to the surface so that the restriction fragments are retained on the surface. In case of study of DNA-protein interactions with proteins that translocate along the DNA contour as they catalyze a biochemical reaction such as transcription, it is preferable to have the DNA attached at two points only (preferably the molecular extremities, if possible) with the rest of the backbone free from the surface so that the interaction is not hindered by the substrate. As mentioned above, the PADI technique can be used to generate simultaneously many DNA templates for optical mapping applications (Figure 5A). We first immobilized DNA molecules with T7 RNAP and then introduced type II restriction enzyme Sma I into the microchannel. The concentration of T7 RNAP was adjusted to provide enough protein molecules to hold the DNA at several points along the backbone so that the molecule remains anchored to the surface after the cleavage by Sma I, yet the attachment points were not so numerous as to sterically hinder the enzymatic digestion. We note that a double digestion by two different restriction enzymes may be observed if the DNA is immobilized with one restriction enzyme such as EcoR I rather than RNAP, followed by cleavage by another restriction enzyme such as Hind III. The first restriction enzyme that assists the immobilization needs to bind to the target sequence in the absence of a divalent metal ion to prevent the cleavage prior to the immobilization. This DNA restriction enzyme complex is then immobilized in the microchannel by PADI, followed by the introduction of the second Nano Lett., Vol. 6, No. 11, 2006

Figure 5. (A) λ-DNA molecules were stretched and immobilized with T7 RNAP followed by enzymatic cleavage by Sma I. DNA molecules were incubated in the dark with Sma I at room temperature for 2 h before imaging. The location of predicted cleavage sites for Sma I on λ-DNA is shown on the right. Scale bar ) 2.5 µm. (B) Alexa Fluor 546-UTP labeled RNA transcripts (red) formed along YOYO stained T7 DNA (green). Scale bar ) 5 µm.

restriction enzyme along with the divalent metal ion so that both first and second restriction enzymes can cleave the DNA. We also note that the method works over a range of RNAP concentrations of 0.5 to 17 nM under the conditions of our experiments. At RNAP concentrations much below 0.5 nM, we do not observe enough adsorption of DNA molecules, and at concentrations well above 17 nM, the time it takes for DNA molecules to imbibe into the microchannel dramatically increases, presumably due to clogging of the channel near the inlet by the protein molecules. There seems to be an optimum protein concentration of around 1 nM at which DNA is immobilized at multiple points, but not so many as to hinder further enzymatic reaction. The optimum concentration may vary from one DNA-binding protein to another, depending on the protein properties. While the number of attachment points could easily be controlled with RNAP by adjusting the concentration to the optimum value, we experienced a limited control over it with commercially purchased restriction enzymes, apparently due to the presence of highly concentrated BSA, which is mixed in the solution as a stabilizer. Typically in an undiluted stock solution, ∼10 nM of restriction enzyme is mixed with 3 µM of BSA, whereas 2 µM of RNAP is readily available commercially without BSA. Therefore, when a restriction enzyme solution is diluted to the optimum concentration (∼1 nM), it is BSA, which is present in the solution at the concentration 3002503

fold higher than the restriction enzyme, that primarily adsorbs on the PMMA-coated surface, presumably leaving little room for proteins on DNA to be adsorbed. The presence of such a high concentration of BSA in the solution also considerably increases the time for imbibition, hindering further increase in the protein concentration. We have also investigated whether transcription can be observed on DNA molecules immobilized by proteins. Transcription on single combed T7 DNA molecules that are not overstretched has been previously demonstrated.16 Newly synthesized RNA transcripts can be detected with TIRF microscopy as bright dots, when enough fluorescent uridine triphosphates (UTPs) are incorporated into the growing RNA chain. To verify whether this in vitro transcription can be observed on DNA molecules immobilized in the microchannel by the PADI technique, we stretched and immobilized T7 DNA molecules with the assistance of 1 nM of T7 RNAP, followed by introduction of transcription buffer containing nucleotriphosphates (200 µM of NTPs, 8 µM Alexa Fluor 546-UTP), and 100 nM of fresh T7 RNAP. We observe bright dots showing RNA transcripts synthesized along the DNA (Figure 5B). In summary, we have developed a novel DNA stretching method that utilizes the specific or nonspecific interactions between the DNA and DNA binding proteins to immobilize large numbers of DNA molecules onto a substrate at physiological hydration and pH conditions. In this way, we found that DNA molecules can reliably be stretched and immobilized inside a microchannel when introduced by a simple capillary force. Many different DNA binding proteins such as restriction enzymes, DNA polymerase, and RNAP can be used to achieve DNA immobilization irrespective of binding sequence specificity. The number of attachment points along the DNA contour can easily be controlled by varying the protein concentration. The biological activity of the immobilized DNA molecules was tested by digesting the DNA with restriction enzymes in the microchannel. Singlemolecule transcription, which places stringent requirements on the immobilized DNA with respect to surface interactions and stretched lengths, is also successfully demonstrated in the microchannel. Acknowledgment. This material is based upon work supported by the National Science Foundation (NSF) under grant No. NSF-NIRT 0304316. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF. Supporting Information Available: Detailed experimental procedures and movie clips of PADI are available free of charge via the Internet at http://pubs.acs.org. References (1) Wang, M. D.; Yin, H.; Landrick, R.; Gelles, J.; Block, S. M. Biophys. J. 1997, 72, 1335-1346.

2504

(2) Davenport, R. J.; Wuite, G. J. L.; Landick, R.; Bustamante, C. Science 2000, 287, 2497-2500. (3) Wuite, G. J. L.; Smith, S. B.; Young, M.; Keller, D.; Bustamante, C. Nature 2000, 404, 103-106. (4) Strick, T. R.; Allemand, J.-F.; Bensimon, D.; Bensimon, A.; Croquette, V. Science 1996, 271, 1835-1837. (5) Maier, B.; Bensimon, D.; Croquette, V. Prod. Natl. Acad. Sci. U.S.A. 2000, 97, 12002-12007. (6) Perkins, T. T.; Smith, D. E.; Larson, R. G.; Chu, S. Science 1995, 268, 83-87. (7) van Oijen, A. M.; Blainey, P. C.; Crampton, D. J.; Richardson, C. C.; Ellenberger, T.; Xie, X. S. Science 2002, 16, 2479-2484. (8) Crut, A.; Lasne, D.; Allemand, J.-F.; Dahan, M.; Desbiolles, P. Phys. ReV. E 2003, 67, 051910. (9) Kabata, H.; Kurosawa, O.; Arai, I.; Washizu, M.; Margarson, S. A.; Glass, R. E.; Shimamoto, N. Science 1993, 262, 1561-1563. (10) Namasivayam, V.; Larson, R. G.; Burke, D. T.; Burns, M. A. Anal. Chem. 2002, 74, 3378-3385. (11) Bensimon, A.; Simon, A.; Chiffaudel, A.; Croquette, V.; Heslot, F.; Bensimon, D. Science 1994, 265, 2096-2098. (12) Michalet, X.; Ekong, R.; Fougerousse, F.; Rousseaux, S.; Schurra, C.; Hornigold, N.; van Slegtenhorst, M.; Wolfe, J.; Povey, S.; Beckmann, J. S.; Bensimon, A. Science 1997, 277, 1518-1523. (13) Bensimon, D.; Simon, A. J.; Croquette, V.; Bensimon, A. Phys. ReV. Lett. 1995, 74, 4754-4757. (14) Cluzel, P.; Lebrun, A.; Heller, C.; Lavery, R.; Viovy, J.-L.; Chatenay, D.; Caron, F. Science 1996, 271, 792-794. (15) Smith, S. B.; Cui, Y.; Bustamante, C. Science 1996, 271, 795799. (16) Gueroui, Z.; Place, C.; Freyssingeas, E.; Berge, B. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6005-6010. (17) Dittmer, W. U.; Simmel, F. C. Appl. Phys. Lett. 2004, 85, 633635. (18) Ashworth, J.; Havranek, J. J.; Duarte, C. M.; Sussman, D.; Monnat, R. J.; Stoddard, B. L.; Baker, D. Nature 2006, 441, 656-659. (19) Gu, Q.; Cheng, C.; Gonela, R.; Suryanarayanan, S.; Anabathula, S.; Dai, K.; Haynie, D. T. Nanotechnology 2006, 17, R14-R25. (20) Wanekaya, A. K.; Chen, W.; Myung, N. V.; Mulchandani, A. Electroanalysis 2006, 18, 533-550. (21) Shendure, J.; Mitra, R. D.; Varma, C.; Church, G. M. Nature ReV. Genet. 2004, 5, 335-344. (22) Schwartz, D. C.; Li, X.; Hernandez, L. I.; Ramnarian, S. P.; Huff, E. J.; Wang, Y. K. Science 1993, 262, 110-114. (23) Reihn, R.; Lu, M.; Wang, Y. M.; Lim, S. F.; Cox, E. C.; Austin, R. H. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 10012-10016. (24) Hihath, J.; Xu, B.; Zhang, P.; Tao, N. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 16979-16983. (25) von Hippel, P. H.; Berg, O. G. J. Biol. Chem. 1989, 264, 675-678. (26) Shimamoto, N. J. Biol. Chem. 1999, 274, 15293-15296. (27) Norde, W. AdV. Colloid Interface Sci. 1986, 25, 267-340. (28) Nakanishi, K.; Sakiyama, T.; Imamura, K. J. Biosci. Bioeng. 2001, 91, 233-244. (29) Sigal, G. B.; Mrksich, M.; Whitesides, G. M. J. Am. Chem. Soc. 1998, 120, 3464-3473. (30) Hong, J. W.; Quake, S. R. Nat. Biotechnol. 2003, 21, 1179-1183. (31) Dittrich, P. S.; Manz, A. Nature ReV. Drug DiscoV. 2006, 5, 210218. (32) Yamashita, K.; Yamaguchi, Y.; Miyazaki, M.; Nakamura, H. Chem. Lett. 2004, 33, 628-629. (33) Keren, K.; Berman, R. S.; Buchstab, E.; Sivan, U.; Braun, E. Science 2003, 302, 1380-1382. (34) Geodecke, N.; Eijkel, J.; Manz, A. Lab Chip 2002, 2, 219-223. (35) Han, J.; Turner, S. W.; Craighead, H. G. Phys. ReV. Lett. 1999, 83, 1688-1691. (36) Marko, J. F.; Siggia, E. D. Macromolecules 1995, 28, 8759-8770. (37) Smith, D. E.; Babcock, H. P.; Chu, S. Science 1999, 283, 17241727.

NL0617484

Nano Lett., Vol. 6, No. 11, 2006