Label-free electrical quantification of the dielectrophoretic response of DNA
© Henning et al. 2008
Received: 03 July 2008
Accepted: 05 November 2008
Published: 05 November 2008
A purely electrical sensing scheme is presented that determines the concentration of macromolecules in solution by measuring the capacitance between planar microelectrodes. Concentrations of DNA in the ng/mL range have been used in samples of 1 μL volume. The method has been applied to the characterisation of the dielectrophoretic response of DNA without the need for any chemical modifications. The influence of electrical parameters like duty cycle, voltage and frequency has been investigated. The results are in good agreement with data from dielectrophoretic studies on fluorescently labelled DNA. Extension of the method down to the single molecule level appears feasible.
PACS: 87.50.ch, 87.80.Fe, 87.85.fK
For the construction of systems on the nanometre scale there is a growing need for alternatives to classical photolithography. A promising approach for this is the exploitation of the self-organising properties of biological macromolecules, in particular DNA (deoxyribonucleic acid) [1, 2]. Double-stranded DNA consists of two DNA single strands which form the double helix. It is stabilised by hydrogen bonds between complementary purine and pyrimidine bases. This coupling is very specific and allows to address distinct sites on the DNA at a resolution of 0.34 nm (for B-DNA), i.e. the distance between neighbouring bases. Addressing can easily be accomplished by chemical means in arbitrary volumes, hence in an extremely parallelised manner. Most methods for the synthesis and modification of DNA are well established in molecular biology. However, the characterisation of these constructs and their connection to the macroscopical world are still demanding. For optical detection fluorescent markers are common. Still, labelling of the molecules is necessary, and bleaching of the fluorophores leads to artefacts and limits the possible observation time. Electrochemical sensing calls for chemical modifications, too, either of the target molecules or of the electrodes [3–5]. Label-free characterisation on surfaces is possible e.g. by scanning probe microscopy [6, 7] and optical methods like surface plasmon resonance and grating couplers [8, 9]. Highly desirable would be a purely electrical detection scheme. This is because such a principle could be well integrated into lab-on-a-chip systems, neither optical nor mechanical access would be necessary, and geometrical resolution would principally not be restricted as it is the case with optical methods. Here we present a purely electrical sensing scheme based on the measurement of capacitance changes between microelectrodes caused by DNA concentration changes.
These variations in local DNA concentration are also achieved by electrical means applying dielectrophoresis (DEP). Here an inhomogeneous electrical AC field exerts forces onto macromolecules like DNA towards the electrode edges [10–12]. This method is increasingly exploited for the concentration and alignment of nano-objects like DNA, proteins, nano wires and carbon nanotubes [13, 14]. Whilst it is widely applied as a micro- and nano-tool [15–17] there are only few studies aimed at a fundamental understanding of molecular DEP [18–21]. Therefore we have used the presented sensing scheme for quantifying the dielectrophoretic response of DNA. In contrast to all other known studies on molecular dielectrophoresis of DNA there is no need for any fluorescent labelling of the sample.
Dielectrophoresis and impedance measurements have first been combined by Milner et al.  and Suehiro et al.  for the characterisation of bacteria. They applied a lock-in amplifier or an oscilloscope for the determination of phase and amplitude and, hence, impedance. Arnold  used an impedance analyser for studying the DEP behaviour of yeast cells. In the simplest case the DEP field signal also served as the source signal for the impedance measurement. Alternatively, two signal sources were used in order to choose the properties of both signals independently. This made it possible to select the measuring frequency for optimal sensitivity. However, this led to the need for additional electronic circuitry in order to combine both signals.
As DNA sample the phagemid pBluescript was used. It has a length of 2961 base pairs corresponding to 1.0 μm contour length. That means that it did not bridge the electrode gap of 1.7 μm. pBluescript was prepared from a transformed E. coli culture and linearised by digestion with the restriction enzyme Eco RI. It was purified using an Invisorb Spin PCRapid kit (Invitek) and diluted with deionised water to final concentrations ranging from 18 pM to 18 nM.
Results and discussion
The time constant of the decrease in capacitance due to diffusion of the DNA after DEP application was found to amount to 30 s at 18 nM DNA concentration (Fig. 3, t = 120 s...240 s). This decrease is much slower than the data acquisition speed which is 200 ms or less for each capacitance data. However, this variation in capacitance still limits the resolution of the presented method. This is because the actual capacitance changes take place on a shorter time scale than the bridge's automatic balancing procedure for complete balancing. The observed capacitance decay of 80 pF is an order of magnitude larger than the typical capacitance change recorded for DEP quantification. This means that by far most of the measured signal is a consequence of reversible DNA attraction and that most of the DNA does not adhere to the electrodes.
This is inconsistent with the observations of Washizu et al.  who report spontaneous permanent fixation of DNA to aluminium. On the other hand this result agrees well with the work of Kabata et al.  who deliberatly functionalised the ends of DNA with avidin in order to achieve permanent adherence to the aluminium electrodes.
The values for capacitance change at 100% duty cycle increased in the course of the experiment by 9% (t1 = 100 s, t2 = 800 s). That means that there is a systematic error introduced, presumably by the gradual concentration increase close to the electrodes due to DEP action, which also is reflected by the increase in the absolute average capacitance by 4% within this period. Errors are also introduced by aggregation of DNA molecules [19, 29] which leads to variations of the local DNA concentration in the electrode gap . As a consequence of the DEP duration being much shorter than the time needed to reach equilibrium concentration the determined capacitance changes reflect the initial dielectrophoretic collection rate as described by Bakewell and Morgan , whilst in other studies steady-state DNA concentrations were monitored [18, 31].
In the typical combination of impedance measurement and dielectrophoresis biological cells are usually studied, which have to be agitated after or during each measurement to achieve an even distribution for the following measurement. For this end flow through systems are used [23, 24]. When extending the method onto molecules mixing already occurs by thermally driven stochastic fluctuations (Brownian motion). Consequently a batch system is also suitable. Such a static system is not only simpler, it also requires much smaller sample volumes. Therefore it is possible to reduce the current volume of about 1 μL even further.
Zheng et al.  investigated the suitability of DNA and proteins for the manufacturing of electronic devices. For this purpose they measured the electrical resistance between narrow electrode tips during and after dielectrophoretic manipulation of DNA and of the protein BSA, however, without any success. Probably this was mainly a consequence of the small interaction length of their tip electrodes of only about 10 μm as compared to the interdigitated electrode arrangement of this work, which extends over 55 mm with only a fifth of the gap width. Above this, we observed that concentration changes of DNA led to a more distinct response in the capacitive part of the impedance than in its resistive part.
The electrical conductivity of the sample solution strongly influences the dielectrophoretic response. Due to the small volumes involved and, hence, high surface-to-volume ratio it is problematic to presume equal conductivities in the stock solution and in the actual sample in situ. Even the stock solution itself is too small to be measured with a standard conductivity probe. We therefore used the resistive part of the bridge's output to deduce the electrical conductivity of the actual sample volume after calibration with solutions of known conductivity. From this an upper bound of 5 mS/m for the electrical conductivity in all experiments follows. From AC electrokinetic studies on similarly small sample volumes [32, 33] a lower bound of 1 mS/m can be deduced.
The stability of the DNA double helix is influenced by the solution's salt content. At low salt concentrations double stranded DNA tends to disintegrate into its composing single strands. This is reflected by the melting point of hybridised DNA, which can be approximately calculated from its base composition and the salt concentration . The latter can be estimated from the sample's electrical conductivity to lie in the range between 0.08 mM and 0.4 mM. This results in a melting point of pBluescript between 34°C and 45°C. The melting point is also influenced by the solution's pH-value. For deionised water it is around pH 5 due to atmospheric CO2. This will lower the melting point by about 2°C  to 32°C to 43°C. This is well above the experimental temperature of 22 (± 2) °C and in accordance with other dielectrophoretic studies on double stranded DNA having been performed in deionised water [21, 36]. Still, it cannot fully ruled out for this work as well as for other dielectrophoretic studies that a minor portion of DNA is present as single strands.
It is of interest to estimate whether the present setup should be capable of detecting single macromolecules. The capacitance bridge used in this work is specified by the manufacturer with a reportable resolution of 10-7 pF under optimal conditions. If one considers a maximal measuring voltage of 30 mVRMS as used in this work the resolution should be reduced to about 10-4 pF. This is three orders of magnitude better than has been achieved in this work and is mainly a consequence of the fast concentration changes after dielectrophoretic attraction. Additionally, the shielding of the cables currently used is not better than 60 dB and the contact-to-contact capacitance of the relays amounts to about 1 pF. Both factors allow interference by external noise sources. From the present data it follows, that under optimal conditions and if one considers electrodes of e.g. 1 μm width and a mutual distance of 1 μm a single pBluescript molecule would result in a capacitance change of more than 10-2 pF, which is still two orders of magnitude above the calculated resolution. Therefore it appears quite feasible to further develop the presented apparatus towards an electronic labelfree single molecule detector. Selectivity would be readily achieved by chemical modifications of the electrodes or of the gap by e.g. complementary DNA or antibodies as it is common practise in other labelfree detection schemes like surface plasmon resonance and grating couplers [8, 9]. Additionally, such a detector could be used in combination with a dielectrophoretic single molecule trap  allowing for an automatic consecutive investigation of large numbers of single molecules.
A system has been developed for the measurement of the concentration of macromolecules by monitoring the capacitance between interdigitated electrodes. It has been applied to the determination of the dielectrophoretic response of DNA without the need for any chemical modification of the analyte. Being purely electronic the method can be easily integrated into lab-on-a-chip systems. Neither optical nor mechanical access to the sample is needed in the course of the measurement. An improvement of the temporal resolution by about an order of magnitude appears rather straighforward. Some modifications of the experimental design will allow for a downscaling of the actual sample volume to a few μm3 or even less. In this case the resolution of the instrumentation will be adequate to automatically detect and possibly characterise single macromolecules.
We would like to thank Andreas Calender for assistance in software design and Mandy Lorenz for her help concerning DNA preparations. Financial support by the European Communities within the project Nucan (STRP 013775) is gratefully acknowledged.
- Carbone A, Seeman NC: Proc Nat Acad Sci (USA). 2002, 99: 12577-12582.View ArticleADSMathSciNetGoogle Scholar
- Yan H, Park SH, Finkelstein G, Reif JH, LaBean TH: Science. 2003, 301: 1882-1884.View ArticleADSGoogle Scholar
- Drummond TG, Hill MG, Barton J: Nature Biotechnol. 2003, 21: 1192-1199.View ArticleGoogle Scholar
- Park S-J, Taton TA, Mirkin CA: Science. 2002, 295: 1503-1506.View ArticleADSGoogle Scholar
- Xiao Y, Lubin AA, Baker BR, Plaxco KW, Heeger AJ: Proc Nat Acad Sci (USA). 2006, 103: 16677-16680.View ArticleADSGoogle Scholar
- Hörber JKH, Miles MJ: Science. 2003, 302: 1002-1005.View ArticleADSGoogle Scholar
- Kada G, Kienberger F, Hinterdorfer P: Nano Today. 2008, 3: 12-19.View ArticleGoogle Scholar
- Homola J: Anal Bioanal Chem. 2003, 377: 528-539.View ArticleGoogle Scholar
- Gauglitz G: Anal Bioanal Chem. 2005, 381: 141-155.View ArticleGoogle Scholar
- Washizu M, Kurosawa O: IEEE Trans Ind Appl. 1990, 26: 1165-1172.View ArticleGoogle Scholar
- Hölzel R, Bier FF: IEE Proc Nanobiotechnol. 2003, 150: 47-53.View ArticleGoogle Scholar
- Morgan H, Green NG: AC electrokinetics: colloids and nanoparticles. 2003, Baldock, Research Studies PressGoogle Scholar
- Lapizco-Encinas BH, Rito-Palomares M: Electrophoresis. 2007, 28: 4521-4538.View ArticleGoogle Scholar
- Stokes P, Khondaker SI: Nanotechnology. 2008, 19: 175202-View ArticleADSGoogle Scholar
- Kabata H, Kurosawa O, Arai I, Washizu M, Margarson SA, Glass RE, Shimamoto N: Science. 1993, 262: 1561-1563.View ArticleADSGoogle Scholar
- Krupke R, Hennrich F, v Löhneysen H, Kappes MM: Science. 2003, 301: 344-347.View ArticleADSGoogle Scholar
- Liu X, Spencer JL, Kaiser AB, Arnold WM: Curr Appl Phys. 2006, 6: 427-431.View ArticleADSGoogle Scholar
- Du M-L, Bier FF, Hölzel R: AIP Conf Proc. 2006, 859: 65-72.View ArticleADSGoogle Scholar
- Bakewell DJ, Morgan H: IEEE Trans Nanobioscience. 2006, 5: 1-8.View ArticleGoogle Scholar
- Salonen E, Terama E, Vattulainen I, Karttunen M: Eur Phys J E. 2005, 18: 133-142.View ArticleGoogle Scholar
- Zheng L, Brody JP, Burke PJ: Biosens Bioelectron. 2004, 20: 606-619.View ArticleGoogle Scholar
- Bier FF, Gajovic-Eichelmann N, Hölzel R: AIP Conf Proc. 2002, 640: 51-59.View ArticleADSGoogle Scholar
- Milner KR, Brown AP, Allsopp DWE, Betts WB: Electronics Lett. 1998, 34: 66-68.View ArticleGoogle Scholar
- Suehiro J, Yatsunami R, Hamada R, Hara M: J Phys D: Appl Phys. 1999, 32: 2814-2820.View ArticleADSGoogle Scholar
- Arnold WM: IEEE Conf Proc Electrical Insul Diel Phenom. 2001: 40-43.Google Scholar
- Beck JD, Shang L, Marcus MS, Hamers RJ: Nanolett. 2005, 5: 777-781.View ArticleADSGoogle Scholar
- Hölzel R, Bier FF: AIP Conf Proc. 2004, 725: 77-83.View ArticleADSGoogle Scholar
- Washizu M, Kurosawa O, Arai I, Suzuki S, Shimamoto N: IEEE Trans Ind Appl. 1995, 231: 447-456.View ArticleGoogle Scholar
- Washizu M, Suzuki S, Kurosawa O, Nishizaka T, Shinohara T: IEEE Trans Ind Appl. 1994, 30: 835-843.View ArticleGoogle Scholar
- Hölzel R: J Electrostat. 2002, 56: 435-447.View ArticleGoogle Scholar
- Tuukkanen S, Kuzyk A, Toppari JJ, Häkkinen H, Hytönen VP, Niskanen E, Rinkiö M, Törmä P: Nanotechnol. 2007, 18: 295204-View ArticleGoogle Scholar
- Hölzel R: Biochim Biophys Acta. 1999, 1450: 53-60.View ArticleGoogle Scholar
- Hölzel R: Biophys J. 1997, 73: 1103-1109.View ArticleGoogle Scholar
- Sambrook J, Russell DW: Molecular cloning: A laboratory manual. 2001, New York, Cold Spring Harbor Laboratory PressGoogle Scholar
- Privalov PL, Ptitsyn OB, Birshtein TM: Biopolymers. 1969, 8: 559-571.View ArticleGoogle Scholar
- Asbury CL, Engh van den G: Biophys J. 1998, 74: 1024-1030.View ArticleGoogle Scholar
- Pohl HA: Dielectrophoresis. 1978, Cambridge, Cambridge University PressGoogle Scholar
- Wang X-B, Huang Y, Hölzel R, Burt JPH, Pethig R: J Phys D: Appl Phys. 1993, 26: 312-322.View ArticleADSGoogle Scholar
- Asbury CL, Diercks AH, Engh van den G: Electrophoresis. 2002, 23: 2658-2666.View ArticleGoogle Scholar
- Green NG, Ramos A, Gonzalez A, Morgan H, Castellanos A: Phys Rev E. 2000, 61: 4011-4018.View ArticleADSGoogle Scholar
- Green NG, Ramos A, Gonzalez A, Morgan H, Castellanos A: Phys Rev E. 2002, 66: 026305-View ArticleADSGoogle Scholar
- Rothemund PWK: Nature. 2006, 440: 297-302.View ArticleADSGoogle Scholar
- Kuzyk A, Yurke B, Toppari JJ, Linko V, Törmä P: Small. 2008, 4: 447-450.View ArticleGoogle Scholar
- Chou C-F, Tegenfeldt JO, Bakajin O, Chan SS, Cox EC, Darnton N, Duke T, Austin RH: Biophys J. 2002, 83: 2170-2179.View ArticleGoogle Scholar
- Macanovic A, Marquette C, Polychronakos C, Lawrence MF: Nucl Acid Res. 2004, 32: e20-View ArticleGoogle Scholar
- Hölzel R, Calander N, Chiragwandi Z, Willander M, Bier FF: Phys Rev Lett. 2005, 95: 128102-View ArticleADSGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.