8400
Langmuir 2005, 21, 8400-8408
Nanomechanical Sensing of DNA Sequences Using Piezoresistive Cantilevers R. Mukhopadhyay,*,†,‡ M. Lorentzen,†,‡ J. Kjems,†,§ and F. Besenbacher†,‡ Interdisciplinary Nanoscience Center, Department of Physics and Astronomy, Department of Molecular Biology, University of Aarhus, 8000 A° rhus C, Denmark Received April 30, 2005. In Final Form: June 23, 2005 A microfabricated cantilever with an internal piezoresistive component has been sensitized with thiol tethered ss-DNA strands and utilized for an in situ, label-free, highly specific, and rapid DNA detection assay. The generation of a differential surface stress onto the functionalized cantilever surface upon target recognition has allowed nanomechanical identification of 12-nucleotide complementary DNA probes with single base mismatch discrimination (sensitivity of 0.2 µM). Interestingly, utilization of an overhang extension distal to the surface enhanced the sensitivity to the 0.01 µM level. The cantilever was functionalized by inkjet printing technology. Replacing the capture probe with locked nucleic acid (LNA) resulted in a faster target probe capture kinetics compared to DNA-DNA hybridization. The capabilities of the piezoresistive cantilever indicate future ergonomic convenience via miniaturization alternative to the conventional laserbased detection method for portable on-site applications.
Introduction The development of biosensors that can be used for diagnostics is important for health-related industries, and several novel biophysical approaches have been considered toward this end. There has been considerable focus on developing biosensors that quickly, reproducibly, and with high sensitivity can detect DNA sequences by monitoring the DNA-DNA hybridization process and especially detect the single nucleotide polymorphism (SNP) in DNA. SNPs are frequent in human genomes, occurring on average once every thousand nucleotides, and they provide the functional variation that is associated to the human diversity and diseases. In recent years, methods for SNP detection have received considerable attention because SNP studies can help to identify genes that contribute to the diagnosis of common disorders with complex genetic causality and allow the development of improved therapeutic drugs. DNA microarrays based on a fluorescence detection method have proven to be indispensable and very versatile for a high throughput identification of DNA-DNA hybridization in general and the SNPs in DNA. However, it requires fluorescent or radioactive labeling of the probes that makes the method time-consuming and expensive. Alternatively, surface plasmon resonance (SPR) and quartz crystal microbalance (QCM) biosensors can provide information on the kinetics for the net mass uptake/loss of nonlabeled molecules on a functionalized surface, but both methods are difficult to conduct in a high throughput manner. Recently, it has been demonstrated that microfabricated functionalized cantilevers are capable of detecting nonlabeled, specific DNA sequences.1-4 The nanomechanical motion of the cantilever is induced by a change in the * To whom correspondence should be addressed. Telephone: 0045-8942-3691. Fax: 0045-8612-0740. E-mail:
[email protected] and
[email protected]. † Interdisciplinary Nanoscience Center. ‡ Department of Physics and Astronomy. § Department of Molecular Biology. (1) Fritz, J.; Baller, M. K.; Lang, H. P.; Rothuizen, H.; Vettiger, P.; Meyer, E.; Gu¨ntherodt, H.-J.; Gerber, Ch.; Gimzewski, J. K. Science 2000, 288, 316-318.
differential surface stress upon molecules binding to the functionalized gold surface at the top and the absence of molecules on the nonfunctionalized silicon nitride surface at the bottom side of the cantilever. The nanomechanical bending signal of the cantilever is registered either optically using a laser deflection system1-4 or electrically using a piezoresistive readout.5 The optical readout method has proven to be very successful for the cantileverbased DNA sensors,1-4 but it suffers from the limitation that measurements can be difficult in opaque liquids (e.g., blood) because of absorption of the laser light. The optical readout system is ergonomically complicated for a cantilever array set up for fluid-phase experiments. On the contrary, for the piezoresistive cantilever, the operation can be relatively convenient in any liquid environment. The integrated readout from the piezoresistive method makes it useful for constructing a compact sensor device. A small amount of target probe (10-100 µL) is required, and the readout can be obtained from an in situ analysis within a few minutes. It offers a noninvasive method compared to the laser-based technique for detection of biological interactions. It is potentially inexpensive because only a voltage measurement system is necessary. Moreover, these nanoactuators offer the advantage of laser-independent sensor parallelization into arrays,6 thereby enabling a high throughput detection, multiplexing, and fabrication in bulk quantities based on low-cost semiconductor technology. In the present report, the piezoresistive cantilever sensor has been tested for the detection of SNP by monitoring the complementarity between immobilized 12nucleotide ss-DNA probes and 12- or 29-nucleotide target (2) Wu, G.; Ji, H.; Hansen, K.; Thundat, T.; Datar, R.; Cote, R.; Hagan, M. F.; Chakraborty, A. K.; Majumdar, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1560-1564. (3) Hansen, K.; Ji, H.; Wu, G.; Datar, R.; Cote, R.; Majumdar, A.; Thundat, T. Anal. Chem. 2001, 73, 1567-1571. (4) McKendry, R. A.; Zhang, J.; Arntz, Y.; Strunz, T.; Hegner, M.; Lang, H. P.; Baller, M. K.; Certa, U.; Meyer, E.; Gu¨ntherodt, H.-J.; Gerber, Ch. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 9783-9788. (5) Boisen, A.; Thaysen, J.; Jensenius, H.; Hansen, O. Ultramicroscopy 2000, 82, 11-16. (6) Lutwyche, M. I.; Despont, M.; Drechsler, U.; Durig, U.; Ha¨berle, W.; Rothuizen, H.; Stuz, R.; Widmer, R.; Binnig, G. K.; Vettiger, P. Appl. Phys. Lett. 2000, 77, 3299-3301.
10.1021/la0511687 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/04/2005
Nanomechanical Sensing Using Piezoresistive Cantilevers
Langmuir, Vol. 21, No. 18, 2005 8401
Figure 1. (a) Four cantilever array before gold coating. (b) Close up SEM view of the piezoresistive cantilever after gold coating. The rectangular brighter area is the gold-coated region used for functionalization. (c) Location of the piezoresistive element inside the cantilever. (d) Schematic of a Wheatstone bridge with four piezoresistors, with two being placed on the cantilevers and two on the main body of the chip. Table 1. Base Sequences of the Nucleic Acid Probes sensor probes S-oligo-12-mer-1: 5′-HS-(CH2)6-CTA TGT CAG CAC-3′ S-oligo-12-mer-2: 5′-HS-(CH2)6-CTA TGT AAG CAC-3′ S-oligo-12-mer-noncomp: 5′-HS-(CH2)6-CGA TCT GCT AAC-3′ (fully noncomplementary sensor probe) (point mismatch locations are shown in bold and italic) S-oligo-12-mer-1-rev: 5′-CTA TGT CAG CAC-(CH2)6-SH-3′ S-oligo-12-mer-1-long-spacer: 5′-HS-(CH2)6-P-(CH2)12-P-CTA TGT CAG CAC-3′ S-oligo-12-mer-1-LNA: 5′-HS-(CH2)6-CTA TGT CAG CAC-3′ S-oligo-12-mer-2-LNA: 5′-HS-(CH2)6-CTA TGT AAG CAC-3′ S-oligo-12-mer-LNA-noncomp: 5′-HS-(CH2)6-CAC GAC TGT ATC-3′ (fully noncomplementary sensor probe) (LNA monomers are shown in bold, and point mismatch locations are in bold and italic) target probes T-oligo-12-mer-1: 5′-GTG CTG ACA TAG-3′ T-oligo-12-mer-2: 5′-GTG CTT ACA TAG-3′ T-oligo-12-mer-noncomp: 5′-CGA TCT GCT AAC-3′ (fully noncomplementary control probe) T-oligo-29-mer-ext: 5′-GAT ACA GTC GTG CAATC GTG CTG ACA TAG-3′ (the region similar to T-oligo-12-mer-1 is underlined; the rest of the sequence is an overhang extension) T-oligo-29-mer-bridge: 5′-GTG CTG ACA TAG ACTCT GTG CTG ACA TAG-3′ (the regions similar to T-oligo-12-mer-1 connected by a bridging linker are underlined) T-oligo-39-mer-bridge: 5′-GTG CTG ACA TAG ACTCTACTCTACTCT GTG CTG ACA TAG-3′ (the regions similar to T-oligo-12-mer-1 connected by a long bridging linker are underlined)
ss-DNA strands (Table 1) in different electrolytic conditions. Locked nucleic acid (LNA)7 that has the advantages of high thermal stability and improved discrimination is used as an alternative sensor construct. We show that the piezoresistive cantilever approach offers a method for determining both the affinity and the kinetics of DNA hybridization interactions, besides being highly sensitive of base sequences to the level of detection of SNPs in DNA. Experimental Procedures The Cantilever System. The cantilever signal measurement system (CantiLab) and the cantilever chips (CantiChip) were supplied by Cantion A/S,8 Denmark. The complete experimental setup8 consists of a flow-regulating syringe pump (Kent Scientific (7) Singh, S. K.; Nielsen, P.; Koshkin, A. A.; Wengel, J. Chem. Commun. 1998, 455-456.
Corporation), a 6-port valve (Vici) attached with a 100 µL sample loop, a cantilever chip placed in the signal readout unit (CantiLab), and a waste bottle. The change in resistance is measured as an electrical signal by placing the piezoresistor in a Wheatstone bridge configuration. The bridge voltage is 2.5 V for all of the experiments discussed in this paper. The valve and the CantiLab unit are exclusively controlled by software. An electrode at ground potential is placed in the solution in a waste bottle. In the “closed valve” condition, the solution from the pump is pumped directly to the sensor placed in the instrument, and in the “open valve” condition, the solution is pumped through the sample loop. Hybridization experiments were carried out in a liquid cell of volume ∼1 µL containing the cantilever assembly immersed in the buffer solution. (8) Cantion A/S, Lyngby, Copenhagen, Denmark (http://www.cantion.com).
8402
Langmuir, Vol. 21, No. 18, 2005
Two or four parallel gold-coated silicon nitride cantilevers (Figure 1a) were functionalized with a selection of thiol-containing DNA sensor probes (Table 1). Each of the cantilevers was 120 µm long, 50 µm wide, and 480 nm thick (including a 30 nm gold layer), and they were separated by 932 µm (for the two-cantilever array) or 466 µm (for the four-cantilever array). The area of the gold-coated region (Figure 1b) is 105 × 46 µm2. The spring constant and the resonance frequency of these cantilevers are ∼0.139 N/m and ∼43 kHz, respectively. The piezoresistor is made of silicon, because it exhibits a large piezoresistive effect,9 and it was placed above the neutral axis of the cantilever (Figure 1c). The sensor cantilevers were integrated into the Wheatstone bridge configuration (Figure 1d). The microcantilever was clamped at one end to the main body of the chip, and the other end was suspended in the flow channel (Figure 1a). Cantilever Functionalization. Functionalization was carried out with an inkjet printing spotter (CantiSpot, Cantion A/S, Denmark)8 that can selectively coat the top (gold) side of the cantilever and requires less material compared to the microcapillary method or using pipets.10,11 It was fitted with a 100 pL spotter head that deposited droplets of 50-70 µm diameter onto the gold-coated cantilever surface when an electrical pulse was provided from the actuator module. The vertical separation between the orifice of the spotting nozzle and the cantilever surface was approximately 0.5 mm. The functionalization of the two cantilevers was performed under identical conditions of concentrations, volume of sample deposited (10 droplets), buffer, and temperature. The cantilevers were routinely cleaned by UVozone treatment for 20 min prior to functionalization using an UV-tip cleaner (Bioforce Laboratories). After spotting, the sensor chips were kept in a humid environment (generated with a saturated solution of NaCl) for 18-24 h to allow the formation of self-assembled monolayers onto the cantilever surface and to prevent dehydration of the DNA probes. The functionalized arrays were stored in the humidity chamber and were used within 3-4 days without any significant loss of performance. The cantilevers were treated with the Na-phosphate buffer under constant flow conditions (flow rate of 10 µL min-1) for a few hours to remove the nonspecifically adsorbed materials and expose the grafted sensor probe layer for effective interaction with the target probes in subsequent experiments. The cantilevers could be cleaned after an experiment by removing the gold layer with aqua regia (concentrated hydrochloric acid and concentrated nitric acid in the ratio of 3:1, v/v) and reused by depositing a fresh gold layer (30 nm thick on top of 3 nm thick chromium layer). Sensor Probes and Target Probes. The thiol-containing 12-nucleotide ss-DNA sensing probes (Table 1), 5′-HS-(CH2)6CTA TGT CAG CAC-3′ (S-oligo-12-mer-1), 5′-HS-(CH2)6-CTA TGT AAG CAC-3′ (S-oligo-12-mer-2), and 5′-HS-(CH2)6-CGA TCT GCT AAC-3′ (S-oligo-12-mer-noncomp), were obtained tritylprotected from DNA Technology, Denmark, to prevent thiol oxidation to the -S-S- dimeric form. The trityl deprotection was carried out using a deprotection protocol.12 The detritylated sample was dissolved in phosphate buffer (1 M Na-phosphate and 100 mM NaCl at pH 7.0) to ∼50 µM concentration. The deprotected sample was purged with pure N2 gas to create an inert atmosphere, stored in aliquot quantities of 50 µL at -50 °C, and used within 1 week. S-oligo-12-mer-1-rev, 5′-CTA TGT CAG CAC-(CH2)6-S-S-(CH2)6-OH-3′, and S-oligo-12-mer-1-longspacer, 5′-HO-(CH2)6-S-S-(CH2)6-P-(CH2)12-P-CTA TGT CAG CAC-3′ (P is phosphate), were obtained from DNA Technology, (9) Kovacs, G. T. A. Micromachined Transducers Sourcebook, McGraw-Hill, New York, 1998. (10) Bietsch, A.; Zhang, J.; Hegner, M.; Lang, H. P.; Gerber, Ch. Nanotechnology 2004, 15, 873-880. (11) Bietsch, A.; Hegner, M.; Lang, H. P.; Gerber, Ch. Langmuir 2004, 20, 5119-5122. (12) Deprotection protocol: Suspend the oligo in 0.1 M triethylammonium acetate (TEAA) at pH 6.5 and a concentration of approximately 100 OD/mL. Add 0.15 volumes of 1 M aqueous silver nitrate solution, mix thoroughly, and leave to react at room temperature for 30 min. Add 0.20 volumes of 1 M aqueous dithiothreitol (DTT) solution, mix thoroughly, and leave at room temperature for 5 min. Centrifuge the suspension to remove the silver DTT complex. Remove the supernatant. Wash the precipitate with 1 volume of 0.1 M TEAA. Centrifuge and combine the supernatant with the first volume.
Mukhopadhyay et al. Denmark, and were activated by Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) treatment13 prior to functionalization. The LNA sensor probes (Table 1), S-oligo-12-mer-1-LNA, 5′HS-(CH2)6-CTA TGT CAG CAC-3′; S-oligo-12-mer-2-LNA, 5′HS-(CH2)6-CTA TGT AAG CAC-3′; and S-oligo-12-mer-LNAnoncomp, 5′-HS-(CH2)6-CAC GAC TGT ATC-3′ (noncomplementary sensor probe), were obtained in disulfide form from Exiqon A/S, Denmark. The probes were activated by TCEP treatment13 and exchanged in the phosphate buffer (1 M Na-phosphate and 100 mM NaCl at pH 7.0) prior to functionalization. The 12-nucleotide ss-DNA target probes (Table 1), T-oligo12-mer-1, 5′-GTG CTG ACA TAG-3′ (complementary to S-oligo12-mer-1 and S-oligo-12-mer-1-LNA); T-oligo-12-mer-2, 5′-GTG CTT ACA TAG-3′ (complementary to S-oligo-12-mer-2 and S-oligo-12-mer-2-LNA); T-oligo-12-mer-noncomp, a noncomplementary 12-nucleotide strand 5′-CGA TCT GCT AAC-3′; T-oligo29-mer-ext, 5′-GAT ACA GTC GTG-CAATC-GTG CTG ACA TAG3′ (a 29-nucleotide sequence with a 12-nucleotide match with S-oligo-12-mer-1 and the rest an overhang extension); T-oligo29-mer-bridge, 5′- GTG CTG ACA TAG-ACTCT-GTG CTG ACA TAG-3′ (29-nucleotide long DNA with a 5-nucleotide linker between the two 12-nucleotide repeat motifs responsible for binding with S-oligo-12-mer-1); and T-oligo-39-mer-bridge that is similar to T-oligo-29-mer-bridge except for a 15-nucleotide long intermediate linker, were used as received from DNA Technology, Denmark. These samples were diluted to suitable concentrations (1 µM or lower) before each experiment with the Na-phosphate buffer (1 M Na-phosphate and 100 mM NaCl at pH 7.0). The chemicals used for the buffer preparation (Na2HPO4 from J. T. Baker Chemical Co.; NaH2PO4 from Sigma Aldrich, Denmark, and NaCl from Merck, Germany) and deprotection of the trityl group (reagents provided by DNA Technology, Denmark) were of reagent grade and used as obtained. Ultrapure water (Milli-Q) of resistivity 18.2 MΩ cm and the buffer solutions were autoclaved, filtered through 0.2 µm filters, and degassed before use. Data Acquisition and Analysis: Cantilever Experiments. All of the experiments were performed in situ at room temperatures (23 ( 1.0 °C) and constant flow conditions (flow rate of 10 µL min-1). The functionalized cantilevers were equilibrated for a few hours until a stable baseline was obtained. In the following step, the cantilevers were exposed to the complementary strands for 10 min followed by flow continued in the buffer environment. Unless specifically mentioned, the buffer used for all of the experiments was 1 M Na-phosphate and 100 mM NaCl at pH 7.0. The injection points are shown with an arrow in all of the diagrams. The time lag between the injection and the sample entering the chamber is ∼120 s. A total of 3-8 signals from each type of experiment were considered and averaged to obtain the mean readout values. SPR Experiments. SPR experiments were carried out using commercial SPR equipment (Biacore X from Biacore, Sweden) at a constant temperature of 25 °C. Gold chips were obtained from Biacore and were cleaned by UV-ozone treatment for 20 min immediately before the experiment. They were then equilibrated in a buffer environment (1 M Na-phosphate and 100 mM NaCl at pH 7.0) at the flow rate of 10 µL min-1 for a few hours. A total of 100 µL of the thiolated DNA sample was injected and kept in contact with the gold surface for 19-20 h under static conditions. The chip was subsequently flushed with the buffer solution before the recording of the final reading. The difference between the final reading and that prior to injection of the sample was considered to obtain the net mass uptake.
Results and Discussion The real time nanomechanical response of 12-nucleotide target (T) ss-DNA, T-oligo-12-mer-1 and T-oligo-12-mer2, upon binding to cantilevers functionalized with the 12nucleotide ss-DNA sensor (S) probes, S-oligo-12-mer-1 and S-oligo-12-mer-2, are presented in parts a and b of Figure (13) TCEP treatment protocol: Mix 100 µL of TCEP (50 mM at pH 7.0) to 20 µL of DNA (∼80 µM/µL) and keep the solution at room temperature for 30 min. Remove excess TCEP by running this solution through a sephadex G-25 microspin column by centrifugation. For exchange in a buffer, elute in the buffer using a similar column.
Nanomechanical Sensing Using Piezoresistive Cantilevers
Langmuir, Vol. 21, No. 18, 2005 8403
Figure 2. (a and b) Single base mismatch detection at 1 µM target DNA concentration. Control response from the (c) noncomplementary target DNA and the hybridization buffer using the same sensor construct.
Figure 3. Variation in the differential readout values at different concentrations (0.2-1 µM) of the target DNA T-oligo-12-mer-1.
2. T-oligo-12-mer-1 and T-oligo-12-mer-2 are related to each other by a single mismatch at position 6 from the 5′ end. A strong differential signal for both the complementary target probes was observed reproducibly. In a twocantilever sensor configuration, the preferential binding of a target probe onto one of the cantilevers can generate a positive or negative signal depending on the location of the cantilever in the bridge configuration. The signals obtained for T-oligo-12-mer-1 and T-oligo-12-mer-2 were positive and negative, respectively, indicating that hybridization of complementary oligonucleotides onto the piezoresistive cantilevers generates a clear sequencespecific response and that a single base mismatch between
two 12-nucleotide sequences is detectable using the piezoresistive cantilever assay. No signal was observed in the control experiments with the fully noncomplementary target DNA T-oligo-12-mer-noncomp and the hybridization buffer (parts c and d of Figure 2, respectively). The signals obtained for a range of concentrations (Figure 3) show that the cantilever response is dose-dependent. At low concentrations (
