- Email: [email protected]
DNA hybridization detection on electrical microarrays using coulostatic pulse technique
Biosensors and Bioelectronics 22 (2006) 744–751
DNA hybridization detection on electrical microarrays using coulostatic pulse technique V. Dharuman a , E. Nebling a,∗ , T. Grunwald b , J. Albers a , L. Blohm b , B. Elsholz a , R. W¨orl a , R. Hintsche a,b a
Department of Biotechnical Microsystems, Fraunhofer Institute for Silicon Technology, Fraunhoferstrasse 1, D-25524 Itzehoe, Germany b Department of Biotechnical Microsystems, eBiochip Systems GmbH, Fraunhoferstrasse 1, D-25524 Itzehoe, Germany Received 20 December 2005; received in revised form 14 February 2006; accepted 20 February 2006 Available online 29 March 2006
Abstract We demonstrated a novel application of transient coulostatic pulse technique for the detection of label free DNA hybridization on nm-sized gold interdigitated ultramicroelectrode arrays (Au–IDA) made in silicon technology. The array consists of eight different positions with an Au–IDA pair at each position arranged on the Si-based Biochip. Immobilization of capture probes onto the Au–IDA was accomplished by self-assembling of thiol-modified oligonucleotides. Target hybridization was indicated by a change in the magnitude of the time dependant potential relaxation curve in presence of electroactive Fe(CN)6 3− in the phosphate buffer solution. While complementary DNA hybridization showed 50% increase in the relaxation potential, the non-complementary DNA showed a negligible change. A constant behaviour was noted for all positions. The dsDNA specific intercalating molecule, methylene blue, was found to be enhancing the discrimination effect. The changes in the relaxation potential curves were further corroborated following the ELISA like experiments using ExtraAvidine alkaline phosphatase labelling and redox recycling of para-aminophenol phosphate at IDAs. The coulostatic pulse technique was shown to be useful for identifying DNA sequences from brain tumour gene CK20, human herpes simplex virus, cytomegalovirus, Epstein–Barr virus and M13 phage. Compared to the hybridization of short chain ONTs (27 mers), the hybridization of long chain M13 phage DNA showed three times higher increase in the relaxation curves. The method is fast enough to monitor hybridization interactions in milli or microsecond time scales and is well suitable for miniaturization and integration compared to the common impedance techniques for developing capacitative DNA sensors. © 2006 Elsevier B.V. All rights reserved. Keywords: Coulostatic pulse technique; DNA hybridization; Label free detection; IDA electrodes; Si-technology; Impedance; Electrical array biochips
1. Introduction Microarray technology is a powerful tool for measuring DNA hybridization in medical diagnostics and biotechnology (Ramsay, 1998; Niemeyer and Blohm, 1999; Berney et al., 2000; Albers et al., 2003; Wie et al., 2003). At present, detection of DNA hybridization events on electrochemical transducers attracts paramount interest particularly due to the feasibility of producing miniaturized, portable and cost effective measuring devices and integrated sensors with high degree of parallelism (Wang, 2000; Albers et al., 2003). Because the sensors based on the change in dielectric properties of electrode/film interface
∗
Corresponding author. Tel.: +49 4821 174312; fax: +49 4821 174250. E-mail address: [email protected] (E. Nebling).
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.02.014
could be more versatile and convenient, the application of capacitance and impedance measurements receives a constant interest (Katz and Willner, 2003; Guan et al., 2004). Impedance spectroscopy was shown to be an effective method for the detection of biomolecular affinity interactions at different electro sensors (Montelius et al., 1995; Paeschke et al., 1996a; Ehret et al., 1997; Saum et al., 1998; Van Gerwen et al., 1998; Jacobs, 1998; Laureyn et al., 2000; Souteyrand et al., 2000; Lillie et al., 2001; Cloarec et al., 2002; Frace et al., 2002; Katz and Willner, 2003; Wie et al., 2003; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Menikh et al., 2004; Guiducci et al., 2004; Guan et al., 2004; Yang et al., 2004a,b; Dharuman et al., 2005; Estrela et al., 2005; Peng et al., 2005; Bart et al., 2005) for the reason that it informs directly about the changes occurring at the electrode/film/electrolyte interface. The molecular hybridization interaction at the sensor surface induces a
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
shift in the total impedance, or capacitance of the bulk electrode. Because the IDAs are more sensitive than the conventional electrodes, several types of interdigitated microelectrodes manufactured in silicon technology on either glass or SiO2 substrates (Montelius et al., 1995; Ehret et al., 1997; Saum et al., 1998; Van Gerwen et al., 1998; Laureyn et al., 2000; Frace et al., 2002; Katz and Willner, 2003; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Dharuman et al., 2005) have been reported to be useful for this purpose. This includes the detection of cells (Ehret et al., 1997), enzyme activity (Saum et al., 1998; Katz and Willner, 2003) and immuno interactions (Lillie et al., 2001). However, the direct detection avoiding the enzyme labelled targets on IDAs employing EIS is rather limited (Paeschke et al., 1996a; Jacobs, 1998; Frace et al., 2002; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Dharuman et al., 2005). These studies indicate that the DNA hybridization is optimal to be observed in the low frequency region of the impedance spectrum. Recently, we (Dharuman et al., 2005) have shown that the impedance detection at the IDA structures is influenced by the diffusion layers behaviour arising from the parallel arrangement of the IDAs (microelectrode edge effects) (Stulik et al., 2000) and molecular layer defects existing in the self-assembled capture mono layer (Finklea et al., 1993). Further, the transport of electroactive Fe(CN)6 3− ions is strongly influenced by the frequency dependent behaviour of these diffusion layers for DNA-modified surfaces as well. In other words, the stability of the data observed at the low frequency region is affected by several diffusion processes and the precise evaluation of impedance parameters is still difficult. The other problem associated with using the frequency response analyzer is that more than one ac cycle is necessary (at least n > 5) for the electrochemical system to attain the equilibrium prior to measurements. This consumes time for scanning wide frequency spectrum and may result in electrode deactivation. Owing to these reasons, the low frequency data below 100 Hz were omitted in the study of molecular affinity binding at IDA electrodes embedded on SiO2 substrate by Van Gerwen et al. (1998). These difficulties could be overcome with transient pulse impedance technique such as coulostatic (Reinmuth, 1962; Delahay, 1962; Pilla, 1970). The coulostatic pulse technique has been widely used to study metal corrosion and inhibition effects (Glass et al., 1998; Danilov et al., 2003; Zhao et al., 2004). The method involves an application of current (charge) pulse of very short duration (e.g., 0.1–1 s) to the electrochemical cell. This results in perturbation of the electrode/film interface of the cell inducing a change in the equilibrium potential of the interface equivalent to the double layer capacitance. The relaxation of the induced potential is then measured as a function of time under open circuit conditions. The charging pulse is generated at a constant frequency and potential amplitude, and the electrochemical cell was perturbed by triggering the signal. The length of the pulse is such that even very fast electron transfer reaction does not proceed to an appreciable extent during this charging process. Thus eliminates the interferences from diffusion layers and ohmic resistance. The advantages of this method including impedance data extraction by Fourier and/or Laplace transformations were recently disclosed (Yoo and Park, 2000; Alvarez
745
et al., 2002; Park and Yoo, 2003; Jurczakowski and Lasia, 2004) showing that the impedance data over a wide frequency range can be obtained in a matter of few milliseconds. It was also shown that the method yields impedance data quite similar to the conventional frequency dependent impedance. Although it is quite simple and fast to perform, to the best of our knowledge, there was no application of this technique in DNA hybridization sensing reported in the literature. Therefore, this work aimed to explore the capability of this technique for the detection of DNA hybridization affinity interactions at the electrode/film interface of Au–IDAs embedded in SiO2 . The hybridization is monitored just by observing the change in the signal magnitude of capture probe upon hybridization with the target DNA. The 27 mer synthetic oligomers of tumour marker CK20 base sequence is demonstrated as a model system. This is followed by qualitative detection of oligonucleotide sequences from human virus DNAs such as Herpes simplex Virus (HSV), Cytomegalovirus (CMV) and Epstein–Barr virus (EBV) on the multiple arrays formats. The efficiency of this method in discriminating larger DNAs was demonstrated using M13 phage DNA hybridization. We have used thiolated oligonucleotides for preparing the target recognition layers on Au–IDA surface as it was shown to be the most suitable and fast method (Albers et al., 2003; Nebling et al., 2004) of producing orientation controlled, robust and reusable DNA films (Peterson et al., 2001). The coulostatic measurements were made in presence of electroactive Fe(CN)6 3− redox probe before and after the dsDNA specific intercalation of methylene blue (Boon et al., 2000) for enhancing the discrimination effect. The development of electrochemical DNA sensors based on thiol self-assembling is rather limited by the electrode surface activity which determines the capture probe layer density and the hybridization efficiency (Huang et al., 2001; Peterson et al., 2001). This is addressed by following the target labeling with ExtrAvidine® alkaline phosphatase (Ex-aP) and chronoamperometric redox recycling of para-aminophenylphosphate (p-APP) in ELISA like experiments (Nebling et al., 2004) after the coulostatic measurements were completed. This corroborates the observed signal changes from coulostatic measurements in the absence of enzyme labeled targets. This report presents systematically the observed preliminary results by coulostatic technique for the first time in the biosensor field in relation to DNA microarray development and miniaturization technology. 2. Experimental 2.1. Materials K3 Fe(CN)6 , NaCl, NaOH, were obtained from Fluka while H2 SO4 was purchased from Merck (Darmstadt, Germany). Phosphate buffered saline (PBS), para-aminophenol (p-AP), methylene blue, trizma base (Tris), sodium citrate, Bovine serum albumin (BSA) (30% solution), 6-mercapto-1-hexanol and ExtrAvidin® alkaline phosphatase (Ex-aP) were obtained from Sigma, Deisenhofen, Germany. para-Aminophenyl phosphate (p-APP) was purchased from ICN Biomedicals Inc. (Aurora, IL, USA). All chemicals were used without any further purification. The thiol labelled capture ONT strands used for
746
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
immobilization and biotinylated target ONTs used for hybridization were all short chain synthetic oligomers, synthesized by Interactiva Thermohybaid (Ulm, Germany). Buffer solutions used were: (i) PBS: 120 mM sodium chloride; 2.7 mM potassium chloride; 10 mM sodium dihydrogen phosphate; pH 7.4 (for coulostatic measurements) and (ii) 4×SSC: 300 mM sodium chloride; 30 mM sodium citrate; pH 7.0 (for hybridization reaction). The pH of the freshly prepared buffer solutions was adjusted using dilute NaOH and HCl solutions. Sequences of the used oligonucleotides are grouped in Table 1. The used target ONTs were biotinylated at the 5 -end to enable the control experiments using Ex-aP enzymatic redox recyling in parallel to coulostatic measurements following the established procedure (Nebling et al., 2004).
Table 2 Physical parameters of 8-position Au–IDA microarray Number of positions Each Position Diameter Distance between each position Number of IDA and Au digits in each position Width of digits, w Interdigit spacing, d Total area of available surface for reaction
8 900 m 340 m 2 × 371 800 nm 400 nm 420000 m2
positions. The details of the chip fabrication are depicted in Fig. 1. Fig. 1 presents the complete optical view of a 8 position chip developed (A), details of chip fabrication process steps (B), interdigitated arrangement of Au–IDA (C), REM photograph of IDAs (D) and cross-sectional view of the IDAs (E). The REM photographs reveal the interelectrode distances and the crosssectional view of the chip.
2.2. IDA structures and Au array construction Thin film gold microelectrode arrays were manufactured in standard silicon technology (Hintsche et al., 1991, 1994). The chips were fabricated on 6 in. thermally oxidized silicon wafers in a semiconductor fabrication environment in this institute. One hundred and twenty nanometer thick gold electrodes were deposited onto a titanium adhesion layer through vapour deposition technique followed by interdigitization using liftoff technique to generate IDA structures. Physical dimensions of one chip are given in Table 2. Insulation was achieved by chemical vapour deposition of 500 nm thick silicon oxinitride, excluding the active electrode areas and connecting pads. Each chip consists of 8 active positions with a pair of IDA electrodes on each position whose physical dimensions are also included in Table 2. The hydrophilic Au array sensor at each position is protected by using the hydrophobic polymer BCB surrounding each position to prevent cross-reaction and contamination among the
2.3. Capture DNA immobilization, hybridization and surface control For the first set of experiments with tumor CK20 sequences, 10 M Capture ONTs I and II (500 nl) were immobilized onto the Au surface at different spot positions on the same chip using thiol-alkane spacer via covalent linkage through –SH groups. The spotting on the different chip positions was done using a 10 l Hamilton syringe with a specially shaped stainless steel cannular (300 m in diameter, flat). The syringe was positioned over each position by a micromanipulator 5171 system from Eppendorf AG (Hamburg, Germany). Optical control was done with a stemi 2000-C microsope from Zeiss (Oberkochen, Germany). Finally the microspotted chip was left in a humid chamber to prevent droplet evaporation for 2 h followed by extensive washing with PBS buffer pH 7.4 to remove the non-specifically bound capture ONTs.
Table 1 Used ONT sequences for constructing electrochemical DNA array Name of the sequence Tumour CK20
Human virus
M13 phage DNA
Captures
Targets 5 -mercapto-1-hexane-TATAATTCT-
Non-complementary: CATCTCTGAAAACTTCCG-3 (I) Complementary: 5 -mercapto-1-hexane-CGATCTGTTTTATGTAGGGTTAGGTCA 3 (II) Herpes simplex virus (HSV): 5 -mercapto-1-hexane-TTTTTGTGCGCCACTGCGTCGGCCCTCAGGGAGAG-3 (III) Cytomegalovirus (CMV): 5 -mercapto-1-hexane-TTTTTGGGTCCACAGGGTACTCGCCACCCGGCACC-3 (IV) Non-complementary control Epstein–Barr virus (EBV): 5 -mercapto-1-hexane-TTTTTCGGGCGCAGGCCGGCTAGCCTGTGCTCTTC-3 (V) Capture (II) mentioned for tumour CK20 system is acting as one of the non-complementary control here Complementary: 5 -mercapto-1-hexane-GCCGCCACCAGAACCACCACCAGAG-3’ (VI) Non-complementary: The capture (I) given under tumour CK20 is also used as non-complementary capture for M13 KO7 helper Phage DNA
5 -biotin-TGACCTAACCCTACATAAAACAG-3 (VII) Complementary to capture II Complementary HSV: 5 -biotin-CTCTCCCTGAGGGCCGACGCAGTGGCGCAC-3 (VIII) Complementary CMV: 5 -biotin-GGTGCCGGGTGGCGAGTACCCTGTGGACCC-3 (IX)
Complementary Target (for coulostatic measurements): M13 K07 helper phage DNA with nearly 5000bp was obtained from Amersham Pharmacia biotech (X) 5 -biotin-GCCCCCTTATTAGCGTTTGCCATCTT-3 (XI)
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
747
Fig. 1. Au–IDA microarrays on SiO2 substrate. (A) Photograph of the 8 position Au–IDA microarray, (B) scheme of chip layers, (C) interdigitated arrangement of Au–IDA and (D and E) the REM photographs top-view and cross-sectional views, respectively.
The immobilized capture ONTs I (non-complementary, positions 1–4) and II (complementary, positions 5–8) were hybridized with target ONT (VII) (10 l of 0.1 M) in 4×SSC buffer for 90 min in a humidity chamber. Both the immobilization and hybridization processes were done outside of the flow cell. At each step of the surface modification, the surface state of the Au electrodes was monitored by coulostatic technique in presence of 5 M K3 Fe(CN)6 in PBS solution. Finally, after all coulostatic measurements were completed, the presence of the hybrid layer on the sensor surface was confirmed by coupling the enzyme Ex-aP with biotinylated dsONT (if formed) and its reaction with the substrate p-APP in pH 8.0 under stopped flow mode (Hintsche et al., 1994; Nebling et al., 2004). The change in current–time (i–t) curve slope was taken as the measurement of the layer density and hybridization efficiency at the electrode surface. A similar approach was adopted to detect human virus DNA sequences and M13 phage DNA.
2.5. Coulostatic instrumentation The used instrument set-up was a normal coulostatic pulse technique described in the literature (Reinmuth, 1962; Delahay, 1962; Pilla, 1970; Glass et al., 1998). The instrumental setup consists of a charging pulse generator (Hewlett Packard, 33120A) and a cathode ray oscilloscope (Textronix, TDS520B). As it has been proved that transient pulse coulostatic and the conventional frequency dependent impedance measurements both yield similar quantitative data (Glass et al., 1998; Yoo and Park, 2000; Park and Yoo, 2003) for charge transfer reactions, the same equivalent circuit, recently evaluated from independent impedance measurements (Dharuman et al., 2005), may be applicable to this system as well. Hence, Fourier or Laplace transformation of the data was not attempted here, and only change in the magnitude of potential relaxation curve with time was taken for qualitative analysis in this report. 3. Results and discussions
2.4. Flow through system A polycarbonate flow through cell with an internal volume of 4 l sealed by an O-ring onto the silicon chip was used as a reaction cell for all measurements. A needle type electrical connector was integrated with the flow cell. This allows both individual and parallel electrical measurements. While an individual position was monitored by coulostatic technique, the parallel measurements were made chronoamperometrically using a multi-channel potentiostat (Paeschke et al., 1996b) for surface layer stability monitoring. The flow rate was controlled by a peristaltic pump (Ismatec Glattbrugg-Z¨urich, CH) at 100 l/min.
3.1. Detection of tumour CK20 ONT sequence hybridization The capture probes I and II (10 M of 500 nl volume) are spotted onto an 8 position IDA shown in Fig. 1A, such that the capture I is immobilized on positions 1–4 and the capture II is immobilized on positions 5–8. The potential relaxation of these positions after capture probe immobilization was measured in presence of 5 M Fe(CN)6 3− . The charging pulse was generated at a frequency of 20 kHz and at potential amplitude of 9 V. All positions showed similar coulostatic behaviour indicating a constant coverage for ssDNA within the experi-
748
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
mental error of 5–10%. This variation might be caused by slight differences in the immobilized layer density at different positions. Therefore, prior to the immobilization and hybridization processes, the coulostatic behaviour of the bare positions was studied with the aim of characterizing the surface for parallel measurements in biosensor applications. The relaxation potentials measured at different positions arranged on the same and different chips are compared at 0.2 ms at a time constant of 50 s. It is nearly constant at 20 mV and varies from 2 to 10% between the positions. Next to this, the positional dependent behaviour of the array with a monolayer of 6-mercapto-1hexanol (MCH, 1 mM) was made without and with ssONT layer. Without ssONT, the MCH covered surface showed nearly three times increase (∼70 mV) in relaxation potential compared to the bare Au surface (20 mV), suggesting that the MCH layer behaves as an effective barrier at the sensor surface resulting in high impedance to the diffusion of the Fe(CN)6 3− ions (data not shown). Identical to the bare Au surface, the diffusion of Fe(CN)6 3− at different positions is not the same but varies to a negligible amount. The CV measurements conducted in parallel using a multi-channel potentiostat (Hintsche et al., 1994; Nebling et al., 2004) showed a suppressed quasi reversible peak for Fe(CN)6 3− in pH 7.4 for MCH layer immobilized Au surface than that of the bare Au array positions. The mixed monolayer was prepared in two steps. First, the ssONT was immobilized for 2 h and washed extensively with PBS buffer, followed by incubation of 1 mM MCH for 1.5 h. Comparison of these different monolayer behaviours suggests that the monolayer of ssONT covers nearly 60–70% while the mixed monolayer covers 90–95% (n = 3). These results are in corroboration with the literature (Finklea et al., 1993) that the diffusion limitation to Fe(CN)6 3− is increased for the mixed monolayer compared to the lone ssONT layer. That is, the diffusion of Fe(CN)3− to free sensor surface through the existing monolayer defects such as pin holes and collapsed sites causes small increase in capacitance of the electrodes observed for mixed monolayer.
The arrays carrying a ssONT monolayer are now allowed to hybridize with a target ONT (VII), which is complementary to the capture II at positions 5–8 and non-complementary to the capture I immobilized at positions 1–4. Fig. 2 compares the coulostatic relaxations of the complementary II (Fig. 2A) and non-complementary I (Fig. 2B) captures before (curve a) and after (curve b) hybridization in presence of Fe(CN)6 3− measured at two different positions. In order to examine the effect of measuring time constant on the relaxation behaviour of hybrids present on the surface, the potential relaxation is measured at six different time constants ranging from 5 s to 500 ms at every position. Significant increase in the relaxation potential curve after hybridization of ssONT is observed only in the microsecond time constant region and not in the millisecond regions. The insignificant signal shift for hybridized positions in the absence of Fe(CN)6 3− ions in the buffer indicates that the observed signal shift in the microsecond region is due to the diffusion of charged species. The charge or potential relaxation is not expected to be affected by the electroactive ions diffusion in the smaller time constant region such as microsecond time scale, while this can be observed only at longer time constants (Perverelli and Van Leeuwen, 1980). However, the slight influence by the diffusion of charged species during the relaxation cannot be eliminated completely even at smaller time constants (Reinmuth, 1962; Delahay, 1962; Pilla, 1970). Comparison of the complementary and non-complementary hybridized positions suggests that the diffusion of electro active Fe(CN)6 3− is affected by the formation of double strands effectively in the microsecond region (Fig. 2A, curves a and b) and a maximum discrimination effect was observed at 50 s. Therefore, for further studies, the relaxation potential magnitude change at 0.2 ms at 50 s time constant is taken as a qualitative indicator of hybrid formation. The results of five repeated measurements on the same position showed a constant behaviour indicating a satisfactory reproducibility within 5% for a given capture and target interaction. The impedance increase to the relaxation potential after hybridization is highly specific for complementary hybrid posi-
Fig. 2. Comparative coulostatic response curves observed for (A) complementary hybridization (at position 5) and (B) non-complementary hybridization (at position 3) measured on 8 position Au–IDA chip in presence of 5 M K3 Fe(CN)6 in PBS buffer. Curve a, ssDNA; curve b, dsDNA measured without methylene blue intercalation. Curve c, effect of methylene blue intercalation after hybridization. The charging pulse generated at 20 kHz and potential amplitude 9 V. The potential decay measured at a time constant of 50 s.
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
tions (Fig. 2A, curves a and b) than to the non-complementary positions (Fig. 2B). In other words, the Fe(CN)6 3− diffusion is hindered by the formation of dsONT on the electrode surface. Similar to these observations, the experiments with mixed mono layers also revealed identical behaviours. This clearly suggests that the layer defects are blocked restricting the free diffusion of Fe(CN)6 3− to the sensor surface. Another reason for the increased impedance due to the dsONT layer formation is the increased electrostatic repulsion for Fe(CN)6 3− due to the increased negative charge density on DNA layer (Finklea et al., 1993). Similar measurements were made on all other six different positions and compared for both complementary and non-complementary hybrid positions, Fig. 3A. The figure plots the potential relaxation signal magnitude difference observed before and after hybridization in presence of Fe(CN)6 3− . The results demonstrate a clear discrimination of ssONT and dsONT at all positions effectively by the coulostatic technique. Repeated measurements on different chip surfaces showed reproducibility within the relative standard deviation of 10–15% (n = 7) when the array was constructed in a similar fashion. However, we noticed positional differences as indicated in Fig. 3A among the hybridized positions. For example, the positions at the middle of the chip, namely 5 and 6 showed the signal difference of 70 mV while the corner positions, 7 and 8 are showing 35–40 mV. The near 50% activity difference observed may be due to difference in the activities of the
749
inner and cornered positions. But the behaviour is not observed with the lone MCH layer present at the positions. Therefore, the difference in the accessibility of Fe(CN)6 3− by middle and cornered positions when hybrids present on the surface may depend on the nature of the interactions of diffusion profile, the charged species and the solution flow mode. In order to increase the discrimination signal difference further we used dsDNA specific intercalator, methylene blue (MB) and Barton’s principle of mediated charge transport along dsDNA (Kelley and Barton, 1997; Boon et al., 2000). This was done by incubating the hybridized surface with 2 M MB in PBS buffer for 10 min. The coulostatic measurements were repeated on all positions in presence of 5 M K3 Fe(CN)6 . The results are included in both Fig. 2 (curve c) and Fig. 3B for comparison. It is noted that the charge transfer due to Fe(CN)6 3− after MB intercalation was greatly hindered for ssONT compared to dsONT for which the observed signal displacement is very small. This is in similar trend to the expected operation of Barton’s principle of enhancing the charge transport by dsONT intercalated MB in presence of solution borne Fe(CN)6 3− . The strong increase for the noncomplementary hybrid (ssONT, 140 mV) by MB interaction compared to the complementary hybrid (dsONT, 20–40 mV) indicates severe interaction of MB with non-hybridized ssONT. This may be due to high reactivity of MB for free guanine base (Kara et al., 2002) and negatively charged phosphate by charge compensation (Park and Hahn, 2004) which are said to be not
Fig. 3. Detection of tumour CK20 sequences DNA hybridization. Differences in relaxation potential (in volts) of ssDNA before and after hybridization with target DNA measured at 0.2 ms for non-complementary (at positions 1–4) and complementary (at positions 5–8) captures immobilised measured in presence of 5 M K3 Fe(CN)6 . (A) Without methylene blue intercalation. (B) With methylene blue intercalation. Other conditions are the same as in Fig. 2. (C) Chronoamperometric slopes observed from ELISA experiment.
750
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
assisting the mediated catalytic process of Fe(CN)6 3− (Boon et al., 2000) in solution. But in the present case, among the two interactions, the ionic interaction of positively charged intercalator dominates the intercalative effects. This means that the hybridization may be very less efficient so that some of ssONTs may be left unhybridized even on the complementary positions and therefore, the intercalative assisted charge transfer rate is not more effective to be observed. These facts necessitate the confirmation of dsONT formation and stability on the surface by other techniques. We have utilised ELISA like experiments by target labelling with Ex-aP and p-APP redox recycling principle at IDAs described in our earlier reports (Hintsche et al., 1994; Nebling et al., 2004). The hybrid formation is indicated by increase in the slope of i–t curve while non-hybrid positions showed negligible change. Current–time curve (i–t) slopes observed for both non-complementary and complementary positions are presented in Fig. 3C. That is, the slope of the i–t curves raised steeply at the hybridized positions (∼250 nA/min) due to the Ex-aP catalyzed conversion of p-APP → p-AP, and p-AP diffuses to the IDA surface where it undergoes the reversible reaction, p-AP ↔ quinoimine, at constant applied potentials +300 (at anode) and −50 mV (at cathode) (Ag/AgCl). The same reaction was hindered at non-hybridized positions due to the absence of Ex-aP indicated by small values around 20 nA/min. The small increase of i–t slope for non-complementary positions may be due to the presence of a very small quantity of electroactive pAP in the commercial sample which can diffuse to the electrode surface through the monolayer defects. Various other experimental parameters studied are ionic concentration of the buffer electrolyte, different capture and target concentrations. The minimum concentration of oligonucleotides related to tumor CK20 sequences that can be detected by this method is 0.01 M under optimal conditions. 3.2. Detection of Viral and M13 phage DNAs The detection of human viral infectious DNAs like herpes simplex virus (HSV), cytomegalovirus (CMV) and Epstein–Barr virus (EBV) is vital in clinical diagnosis and drug monitoring therapy of psoriasis (Mehraein et al., 2004; Ibrahim et al., 2005) and in earlier acute coronary syndromes (Miller et al., 2005). Onto the eight-position array chip shown in Fig. 1A, the complementary captures HSV (III), CMV (IV) and the noncomplementary captures EBV (V) and CK20 (II) (this capture is non-complementary to CMV and HSV targets) are microspotted on four different positions. Hybridization was made using a mixture of complementary targets (10 l) HSV (VIII) and CMV (IX) (each 5 l of 1 M concentration). The comparative coulostatic results observed in presence of 5 M K3 Fe(CN)6 after intercalation with 2 M MB showed ∼80 mV (measured at 0.2 ms of coulostatic curve) for complementary positions and ∼50 mV for non-complementary positions for the reason discussed before. However, compared to tumour CK20 dsONT (∼20–40 mV), the MB signal for viral DNA (∼50 mV) is increased probably due to different hybridization efficiency to form dsONT. The i–t slopes from enzyme Ex-aP labelling showed 531.47
and 262.60 nA/min for complementary positions and 12.42 and −23.92 nA/min for non-complementary positions confirming the perfect discrimination by the coulostatic technique. Because the hybridization efficiency is strongly dependent on the length of the capture and target molecules as well (Okahata et al., 1992; Huang et al., 2001), we have examined the detection of M13 phage DNA hybridization with its capture probe (VI) immobilized on two positions. Here, the ONT capture (I) was used as negative control to M13 KO7 target ONT (XI) and immobilized onto two other positions on the same chip surface. The detection scheme involves immobilization of M13 capture (VI) followed by its sequential hybridization with M13 phage DNA natural sample (X) and detection target DNA (XI). The coulostatic measurements were made only to know the effect of hybridization of M13 KO7 helper phage with its complementary capture probe (VI) in presence of 5 M K3 Fe(CN)6 . The MB intercalation study was neglected due to its strong interaction with free guanine bases and phosphate backbone of ssDNA which makes data processing difficult, if the target causes unspecific binding (Kara et al., 2002). The biotinylated target was used just to ensure the confirmation test following the enzymatic conjugate Ex-aP and p-APP redox recycling method after coulostatic technique, similar to the tumour and viral DNA tests. Compared to the tumour CK20 and human viruses (HSV, CMV, EBV) systems (∼40 mV), the observed signal difference before and after hybridization with M13 KO7 phage is nearly three times (∼120 mV) higher in presence of Fe(CN)6 3− alone, for hybrid positions (115 and 110 mV) and 25–35 mV for non-complementary positions. This is attributed to the large number of base pairs that are electrostatically repelling the Fe(CN)6 3− ion diffusion. The i–t slopes from ELISA measurements showed 259.78 and 274.86 nA/min for hybrid positions and 11.58 and 4.87 nA/min for non-complementary positions. This suggests the easy, simple discrimination and detection of natural DNA hybridization on these chip platforms very well. These experiments have been reproduced three times, with R.S.D. of 6%. 4. Conclusions The application of coulostatic technique in biomolecular affinity interaction was explored for the first time in biosensor research. The comparison of the signal change magnitudes of short chain ONTs hybridization and larger base pair DNAs such as M13 phage DNAs indicates the technique’s sensitiveness to the molecular environment change at the electrode film/solution surface. Further, the results suggested its effective sensing ability in the microsecond time constant scale which is quite fast compared to the conventional impedance technique. The tandem use of coulostatic and chronoamperometric redox recycling principle following the enzyme Ex-aP labelling, corroborates the surface coverage as well as the hybridization efficiency on the electrode surface. Further advantage of this system is the elimination of the reference electrode and the use of a two pole measurement. This is more convenient for the development of miniaturized devices with array sensor integration. With the preliminary results observed in the present study and the literature
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
evidences (Yoo and Park, 2000; Alvarez et al., 2002; Park and Yoo, 2003; Jurczakowski and Lasia, 2004) on the potential use of transient pulse techniques to replace the time consuming impedance technique to overcome the difficulty especially in the low frequency region, we expect a major breakthrough in label free sensing in relation to fastness and simplicity in data acquisition. It may be recalled here that hybridization interaction was monitored just by negotiating the change in the signal magnitude of the ssONT upon its hybridization with its target; and therefore, future attention will be given to the optimization of the accuracy of the method with reference to the conventional impedance method. Acknowledgement We thank the Fraunhofergesellschaft and e-BiochipSystems GmbH for the financial support for this project. References Albers, J., Grunwald, T., Nebling, E., Piechotta, G., Hintsche, R., 2003. Anal. Bioanal. Chem. 377, 521–527. Alvarez, A., Sanchez, M., Pinhole, C.A., 2002. Corros. Rev. 20, 69–93. Bart, M., Stigter, E.C.A., Stapert, H.R., de Jong, G.J., van Bennekom, W.P., 2005. Biosens. Bioelectr. 21, 49–59. Boon, E.M., Ceres, D.M., Drummond, T.G., Hill, M.G., Barton, J.K., 2000. Nat. Biotechnol. 18, 1096–1100. Berney, H., West, J., Haefele, E., Alderman, J., Lane, W., Collins, J.K., 2000. Sens. Actuators B 68, 100–108. Cloarec, J.P., Deligianis, N., Martin, J.R., Lawrence, I., Souteyrand, E., Polychronakos, C., Lawrence, M.F., 2002. Biosens. Bioelectr. 17, 405–412. Danilov, F., Obraztsov, V., Kapitonov, A., 2003. J. Electroanal. Chem. 552, 69–76. Delahay, P., 1962. Anal. Chem. 34, 1267–1272. Dharuman, V., Grunwald, T., Nebling, E., Albers, J., Blohm, L., Hintsche, R., 2005. Biosens. Bioelectr. 21, 645–654. Ehret, R., Baumann, W., Brischwein, M., Schwinde, A., Stegbauer, K., Wolf, B., 1997. Biosens. Bioelectr. 12, 29–41. Estrela, P., Migliorato, P., Takiguchi, H., Fukshima, H., Nebashi, S., 2005. Biosens. Bioelectr. 20, 1580–1586. Finklea, H., Sinder, D.A., Fedyk, J., Sabatani, E., Gafni, Y., Rubinstein, I., 1993. Langmuir 9, 3660–3667. Frace, G., Lillie, G., Hianik, T., Payne, P., Vadgama, P., 2002. Bioelectrochemitsry 55, 1–3. Glass, G.K., Hassanien, A.M., Buenfeld, N.R., 1998. Electrochim. Acta 43, 1863–1871. Guiducci, C., Stagni, C., Zuccheri, G., Bogliolo, A., Benini, L., Samori, D., Ricco, R., 2004. Biosens. Bioelectr. 19, 781–787. Gheeorghe, M., Elie, A.G., 2003. Biosens. Bioelectr. 19, 95–107. Guan, J.G., Miao, Y., Zhang, Q., 2004. J. Biosci. Bioeng. 97, 219–226. Hang, T.C., Elie, A.G., 2004. Biosens. Bioelectr. 19, 1537–1548. Hintsche, R., M¨oller, B., Dransfeld, I., Wollenberger, U., Scheller, F., Hoffmann, B., 1991. Sens. Actuators B 4, 287–291. Hintsche, R., Paeschke, M., Wollenberger, U., Schnakenberg, U., Wagner, B., Lisec, T., 1994. Biosens. Bioelectr. 9, 697–705.
751
Huang, E., Satjapipat, M., Han, S., Zhou, F., 2001. Langmuir 17, 1215– 1224. Ibrahim, A.L., Obeid, M.T., Jouma, M.J., Moasis, G.A., Al-Richane, W.L., Kindermann, I., Boehm, M., Roemer, K., Lantzch, N.M., Gartner, B., 2005. J. Clin. Virol. 32, 29–35. Jacobs, P., 1998. Proceedings of the Second International Conference on Microreaction technology, New Orleans, Louisiana, USA, March 8–12. Jurczakowski, R., Lasia, A., 2004. Anal. Chem. 76, 5033–5038. Katz, E., Willner, I., 2003. Electroanalysis 15, 913–947. Kelley, S.O., Barton, J.K., 1997. Bioconju. Chem. 8, 31–37. Kara, P., Kerman, K., Ozkan, D., Meric, B., Erdem, A., Ozkan, Z., Ozsoz, M., 2002. Electrochem. Commun. 4, 705–709. Laureyn, W., Nelis, D., Van Gerwen, P., Baert, K., Hermans, L., Magnee, R., Pireaux, J.-J., Maes, G., 2000. Sens. Actuators B 68, 360–370. Lillie, G., Payne, P., Vadgama, P., 2001. Sens. Actuators B 78, 249–256. Mehraein, Y., Lennerz, C., Ehlhardt, S., Zhang, K.D., Madry, H., 2004. J. Clin. Virol. 31, 25–31. Miller, G.E., Freedland, K.E., Duntley, S., Carney, R.M., 2005. Am. J. Cardiol. 95, 317–321. Montelius, L., Tegenfeldt, J.O., Ling, T.G., 1995. J. Vac. Sci. Technol. A 13, 1755–1760. Menikh, A., Mickan, S.P., Liu, H., MacColl, R., Zhang, X.C., 2004. Biosens. Bioelectr. 20, 658–662. Nebling, E., Grunwald, T., Albers, J., Sch¨afer, J., Hintsche, R., 2004. Anal. Chem. 76, 689–696. Niemeyer, C.M., Blohm, D., 1999. Angew. Chem. Int. Ed. 38, 2865–2869. Okahata, Y., Matsunobu, Y., Ijiro, K., Mukae, M., Murakami, A., Makino, K., 1992. J. Am. Chem. Soc. 114, 8299–8300. Paeschke, M., Bucmann, L.M., Seitz, R., Hintsche, R., 1996a. Microsystems Technologies, vol. 96. VDE Verlag, pp. 399. Paeschke, M., Dietrich, F., Uhlig, A., Hintsche, R., 1996b. Electroanalysis 8, 891–898. Park, S.M., Yoo, J.S., 2003. Anal. Chem. 75, 455a–461a. Park, N., Hahn, J.H., 2004. Anal. Chem. 76, 900–906. Peng, H., soeller, C., Vigar, N., Kilmartin, P.A., Cannell, M.B., Bowmaker, G.A., Cooney, R.P., Sejdic, J.T., 2005. Biosens. Bioelectr. 20, 1821–1828. Peterson, A.W., Heaton, R.J., Georgiadis, R.M., 2001. Nucl. Acids Res. 29, 5163–5168. Perverelli, K.J., Van Leeuwen, H.P., 1980. J. Electroanal. Chem. 110, 119–135. Pilla, A.A., 1970. J. Electrochem. Soc. 117, 467–477. Ramsay, G., 1998. Nat. Biotechnol. 16, 40–44. Reinmuth, W.H., 1962. Anal. Chem. 34, 1272–1276. Saum, A.E.G., Cumming, R.H., Rowell, F.J., 1998. Biosens. Bioelectr. 13, 511–518. Souteyrand, E., Chen, C., Cloarec, J.P., Nesme, X., Simonet, P., Navarro, I., Martin, J.R., 2000. Appl. Biochem. Biotechnol. 89, 195–207. Stulik, K., Amatore, C., Holub, K., Marecek, V., Kutner, W., 2000. Pure Appl. Chem. 72, 1483–1492. Wang, J., 2000. Nucl. Acids Res. 28, 3011–3016. Wie, F., Sun, B., Guo, Y., Zhao, X.S., 2003. Biosens. Bioelectr. 18, 1157– 1163. Van Gerwen, P.V., Laureyn, W., Laureys, W., Huyberechts, H., Op De Beeck, M., Baert, K., Suls, J., Sansen, W., Jacobs, P., Hermans, L., Martens, R., 1998. Sens. Actuator B 49, 73–80. Yang, L., Li, Y., Griffis, C.L., Johnson, G.M., 2004a. Biosens. Bioelectr. 19, 1139–1147. Yang, L., Li, Y., Erf, G.F., 2004b. Anal. Chem. 76, 1107–1113. Yoo, J.S., Park, S.M., 2000. Anal. Chem. 72, 2035–2041. Zhao, Y., Guo, X., Dong, Z., 2004. Corros. Eng. Sci. Tech. 39, 245–249.
DNA hybridization detection on electrical microarrays using coulostatic pulse technique V. Dharuman a , E. Nebling a,∗ , T. Grunwald b , J. Albers a , L. Blohm b , B. Elsholz a , R. W¨orl a , R. Hintsche a,b a
Department of Biotechnical Microsystems, Fraunhofer Institute for Silicon Technology, Fraunhoferstrasse 1, D-25524 Itzehoe, Germany b Department of Biotechnical Microsystems, eBiochip Systems GmbH, Fraunhoferstrasse 1, D-25524 Itzehoe, Germany Received 20 December 2005; received in revised form 14 February 2006; accepted 20 February 2006 Available online 29 March 2006
Abstract We demonstrated a novel application of transient coulostatic pulse technique for the detection of label free DNA hybridization on nm-sized gold interdigitated ultramicroelectrode arrays (Au–IDA) made in silicon technology. The array consists of eight different positions with an Au–IDA pair at each position arranged on the Si-based Biochip. Immobilization of capture probes onto the Au–IDA was accomplished by self-assembling of thiol-modified oligonucleotides. Target hybridization was indicated by a change in the magnitude of the time dependant potential relaxation curve in presence of electroactive Fe(CN)6 3− in the phosphate buffer solution. While complementary DNA hybridization showed 50% increase in the relaxation potential, the non-complementary DNA showed a negligible change. A constant behaviour was noted for all positions. The dsDNA specific intercalating molecule, methylene blue, was found to be enhancing the discrimination effect. The changes in the relaxation potential curves were further corroborated following the ELISA like experiments using ExtraAvidine alkaline phosphatase labelling and redox recycling of para-aminophenol phosphate at IDAs. The coulostatic pulse technique was shown to be useful for identifying DNA sequences from brain tumour gene CK20, human herpes simplex virus, cytomegalovirus, Epstein–Barr virus and M13 phage. Compared to the hybridization of short chain ONTs (27 mers), the hybridization of long chain M13 phage DNA showed three times higher increase in the relaxation curves. The method is fast enough to monitor hybridization interactions in milli or microsecond time scales and is well suitable for miniaturization and integration compared to the common impedance techniques for developing capacitative DNA sensors. © 2006 Elsevier B.V. All rights reserved. Keywords: Coulostatic pulse technique; DNA hybridization; Label free detection; IDA electrodes; Si-technology; Impedance; Electrical array biochips
1. Introduction Microarray technology is a powerful tool for measuring DNA hybridization in medical diagnostics and biotechnology (Ramsay, 1998; Niemeyer and Blohm, 1999; Berney et al., 2000; Albers et al., 2003; Wie et al., 2003). At present, detection of DNA hybridization events on electrochemical transducers attracts paramount interest particularly due to the feasibility of producing miniaturized, portable and cost effective measuring devices and integrated sensors with high degree of parallelism (Wang, 2000; Albers et al., 2003). Because the sensors based on the change in dielectric properties of electrode/film interface
∗
Corresponding author. Tel.: +49 4821 174312; fax: +49 4821 174250. E-mail address: [email protected] (E. Nebling).
0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.02.014
could be more versatile and convenient, the application of capacitance and impedance measurements receives a constant interest (Katz and Willner, 2003; Guan et al., 2004). Impedance spectroscopy was shown to be an effective method for the detection of biomolecular affinity interactions at different electro sensors (Montelius et al., 1995; Paeschke et al., 1996a; Ehret et al., 1997; Saum et al., 1998; Van Gerwen et al., 1998; Jacobs, 1998; Laureyn et al., 2000; Souteyrand et al., 2000; Lillie et al., 2001; Cloarec et al., 2002; Frace et al., 2002; Katz and Willner, 2003; Wie et al., 2003; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Menikh et al., 2004; Guiducci et al., 2004; Guan et al., 2004; Yang et al., 2004a,b; Dharuman et al., 2005; Estrela et al., 2005; Peng et al., 2005; Bart et al., 2005) for the reason that it informs directly about the changes occurring at the electrode/film/electrolyte interface. The molecular hybridization interaction at the sensor surface induces a
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
shift in the total impedance, or capacitance of the bulk electrode. Because the IDAs are more sensitive than the conventional electrodes, several types of interdigitated microelectrodes manufactured in silicon technology on either glass or SiO2 substrates (Montelius et al., 1995; Ehret et al., 1997; Saum et al., 1998; Van Gerwen et al., 1998; Laureyn et al., 2000; Frace et al., 2002; Katz and Willner, 2003; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Dharuman et al., 2005) have been reported to be useful for this purpose. This includes the detection of cells (Ehret et al., 1997), enzyme activity (Saum et al., 1998; Katz and Willner, 2003) and immuno interactions (Lillie et al., 2001). However, the direct detection avoiding the enzyme labelled targets on IDAs employing EIS is rather limited (Paeschke et al., 1996a; Jacobs, 1998; Frace et al., 2002; Gheeorghe and Elie, 2003; Hang and Elie, 2004; Dharuman et al., 2005). These studies indicate that the DNA hybridization is optimal to be observed in the low frequency region of the impedance spectrum. Recently, we (Dharuman et al., 2005) have shown that the impedance detection at the IDA structures is influenced by the diffusion layers behaviour arising from the parallel arrangement of the IDAs (microelectrode edge effects) (Stulik et al., 2000) and molecular layer defects existing in the self-assembled capture mono layer (Finklea et al., 1993). Further, the transport of electroactive Fe(CN)6 3− ions is strongly influenced by the frequency dependent behaviour of these diffusion layers for DNA-modified surfaces as well. In other words, the stability of the data observed at the low frequency region is affected by several diffusion processes and the precise evaluation of impedance parameters is still difficult. The other problem associated with using the frequency response analyzer is that more than one ac cycle is necessary (at least n > 5) for the electrochemical system to attain the equilibrium prior to measurements. This consumes time for scanning wide frequency spectrum and may result in electrode deactivation. Owing to these reasons, the low frequency data below 100 Hz were omitted in the study of molecular affinity binding at IDA electrodes embedded on SiO2 substrate by Van Gerwen et al. (1998). These difficulties could be overcome with transient pulse impedance technique such as coulostatic (Reinmuth, 1962; Delahay, 1962; Pilla, 1970). The coulostatic pulse technique has been widely used to study metal corrosion and inhibition effects (Glass et al., 1998; Danilov et al., 2003; Zhao et al., 2004). The method involves an application of current (charge) pulse of very short duration (e.g., 0.1–1 s) to the electrochemical cell. This results in perturbation of the electrode/film interface of the cell inducing a change in the equilibrium potential of the interface equivalent to the double layer capacitance. The relaxation of the induced potential is then measured as a function of time under open circuit conditions. The charging pulse is generated at a constant frequency and potential amplitude, and the electrochemical cell was perturbed by triggering the signal. The length of the pulse is such that even very fast electron transfer reaction does not proceed to an appreciable extent during this charging process. Thus eliminates the interferences from diffusion layers and ohmic resistance. The advantages of this method including impedance data extraction by Fourier and/or Laplace transformations were recently disclosed (Yoo and Park, 2000; Alvarez
745
et al., 2002; Park and Yoo, 2003; Jurczakowski and Lasia, 2004) showing that the impedance data over a wide frequency range can be obtained in a matter of few milliseconds. It was also shown that the method yields impedance data quite similar to the conventional frequency dependent impedance. Although it is quite simple and fast to perform, to the best of our knowledge, there was no application of this technique in DNA hybridization sensing reported in the literature. Therefore, this work aimed to explore the capability of this technique for the detection of DNA hybridization affinity interactions at the electrode/film interface of Au–IDAs embedded in SiO2 . The hybridization is monitored just by observing the change in the signal magnitude of capture probe upon hybridization with the target DNA. The 27 mer synthetic oligomers of tumour marker CK20 base sequence is demonstrated as a model system. This is followed by qualitative detection of oligonucleotide sequences from human virus DNAs such as Herpes simplex Virus (HSV), Cytomegalovirus (CMV) and Epstein–Barr virus (EBV) on the multiple arrays formats. The efficiency of this method in discriminating larger DNAs was demonstrated using M13 phage DNA hybridization. We have used thiolated oligonucleotides for preparing the target recognition layers on Au–IDA surface as it was shown to be the most suitable and fast method (Albers et al., 2003; Nebling et al., 2004) of producing orientation controlled, robust and reusable DNA films (Peterson et al., 2001). The coulostatic measurements were made in presence of electroactive Fe(CN)6 3− redox probe before and after the dsDNA specific intercalation of methylene blue (Boon et al., 2000) for enhancing the discrimination effect. The development of electrochemical DNA sensors based on thiol self-assembling is rather limited by the electrode surface activity which determines the capture probe layer density and the hybridization efficiency (Huang et al., 2001; Peterson et al., 2001). This is addressed by following the target labeling with ExtrAvidine® alkaline phosphatase (Ex-aP) and chronoamperometric redox recycling of para-aminophenylphosphate (p-APP) in ELISA like experiments (Nebling et al., 2004) after the coulostatic measurements were completed. This corroborates the observed signal changes from coulostatic measurements in the absence of enzyme labeled targets. This report presents systematically the observed preliminary results by coulostatic technique for the first time in the biosensor field in relation to DNA microarray development and miniaturization technology. 2. Experimental 2.1. Materials K3 Fe(CN)6 , NaCl, NaOH, were obtained from Fluka while H2 SO4 was purchased from Merck (Darmstadt, Germany). Phosphate buffered saline (PBS), para-aminophenol (p-AP), methylene blue, trizma base (Tris), sodium citrate, Bovine serum albumin (BSA) (30% solution), 6-mercapto-1-hexanol and ExtrAvidin® alkaline phosphatase (Ex-aP) were obtained from Sigma, Deisenhofen, Germany. para-Aminophenyl phosphate (p-APP) was purchased from ICN Biomedicals Inc. (Aurora, IL, USA). All chemicals were used without any further purification. The thiol labelled capture ONT strands used for
746
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
immobilization and biotinylated target ONTs used for hybridization were all short chain synthetic oligomers, synthesized by Interactiva Thermohybaid (Ulm, Germany). Buffer solutions used were: (i) PBS: 120 mM sodium chloride; 2.7 mM potassium chloride; 10 mM sodium dihydrogen phosphate; pH 7.4 (for coulostatic measurements) and (ii) 4×SSC: 300 mM sodium chloride; 30 mM sodium citrate; pH 7.0 (for hybridization reaction). The pH of the freshly prepared buffer solutions was adjusted using dilute NaOH and HCl solutions. Sequences of the used oligonucleotides are grouped in Table 1. The used target ONTs were biotinylated at the 5 -end to enable the control experiments using Ex-aP enzymatic redox recyling in parallel to coulostatic measurements following the established procedure (Nebling et al., 2004).
Table 2 Physical parameters of 8-position Au–IDA microarray Number of positions Each Position Diameter Distance between each position Number of IDA and Au digits in each position Width of digits, w Interdigit spacing, d Total area of available surface for reaction
8 900 m 340 m 2 × 371 800 nm 400 nm 420000 m2
positions. The details of the chip fabrication are depicted in Fig. 1. Fig. 1 presents the complete optical view of a 8 position chip developed (A), details of chip fabrication process steps (B), interdigitated arrangement of Au–IDA (C), REM photograph of IDAs (D) and cross-sectional view of the IDAs (E). The REM photographs reveal the interelectrode distances and the crosssectional view of the chip.
2.2. IDA structures and Au array construction Thin film gold microelectrode arrays were manufactured in standard silicon technology (Hintsche et al., 1991, 1994). The chips were fabricated on 6 in. thermally oxidized silicon wafers in a semiconductor fabrication environment in this institute. One hundred and twenty nanometer thick gold electrodes were deposited onto a titanium adhesion layer through vapour deposition technique followed by interdigitization using liftoff technique to generate IDA structures. Physical dimensions of one chip are given in Table 2. Insulation was achieved by chemical vapour deposition of 500 nm thick silicon oxinitride, excluding the active electrode areas and connecting pads. Each chip consists of 8 active positions with a pair of IDA electrodes on each position whose physical dimensions are also included in Table 2. The hydrophilic Au array sensor at each position is protected by using the hydrophobic polymer BCB surrounding each position to prevent cross-reaction and contamination among the
2.3. Capture DNA immobilization, hybridization and surface control For the first set of experiments with tumor CK20 sequences, 10 M Capture ONTs I and II (500 nl) were immobilized onto the Au surface at different spot positions on the same chip using thiol-alkane spacer via covalent linkage through –SH groups. The spotting on the different chip positions was done using a 10 l Hamilton syringe with a specially shaped stainless steel cannular (300 m in diameter, flat). The syringe was positioned over each position by a micromanipulator 5171 system from Eppendorf AG (Hamburg, Germany). Optical control was done with a stemi 2000-C microsope from Zeiss (Oberkochen, Germany). Finally the microspotted chip was left in a humid chamber to prevent droplet evaporation for 2 h followed by extensive washing with PBS buffer pH 7.4 to remove the non-specifically bound capture ONTs.
Table 1 Used ONT sequences for constructing electrochemical DNA array Name of the sequence Tumour CK20
Human virus
M13 phage DNA
Captures
Targets 5 -mercapto-1-hexane-TATAATTCT-
Non-complementary: CATCTCTGAAAACTTCCG-3 (I) Complementary: 5 -mercapto-1-hexane-CGATCTGTTTTATGTAGGGTTAGGTCA 3 (II) Herpes simplex virus (HSV): 5 -mercapto-1-hexane-TTTTTGTGCGCCACTGCGTCGGCCCTCAGGGAGAG-3 (III) Cytomegalovirus (CMV): 5 -mercapto-1-hexane-TTTTTGGGTCCACAGGGTACTCGCCACCCGGCACC-3 (IV) Non-complementary control Epstein–Barr virus (EBV): 5 -mercapto-1-hexane-TTTTTCGGGCGCAGGCCGGCTAGCCTGTGCTCTTC-3 (V) Capture (II) mentioned for tumour CK20 system is acting as one of the non-complementary control here Complementary: 5 -mercapto-1-hexane-GCCGCCACCAGAACCACCACCAGAG-3’ (VI) Non-complementary: The capture (I) given under tumour CK20 is also used as non-complementary capture for M13 KO7 helper Phage DNA
5 -biotin-TGACCTAACCCTACATAAAACAG-3 (VII) Complementary to capture II Complementary HSV: 5 -biotin-CTCTCCCTGAGGGCCGACGCAGTGGCGCAC-3 (VIII) Complementary CMV: 5 -biotin-GGTGCCGGGTGGCGAGTACCCTGTGGACCC-3 (IX)
Complementary Target (for coulostatic measurements): M13 K07 helper phage DNA with nearly 5000bp was obtained from Amersham Pharmacia biotech (X) 5 -biotin-GCCCCCTTATTAGCGTTTGCCATCTT-3 (XI)
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
747
Fig. 1. Au–IDA microarrays on SiO2 substrate. (A) Photograph of the 8 position Au–IDA microarray, (B) scheme of chip layers, (C) interdigitated arrangement of Au–IDA and (D and E) the REM photographs top-view and cross-sectional views, respectively.
The immobilized capture ONTs I (non-complementary, positions 1–4) and II (complementary, positions 5–8) were hybridized with target ONT (VII) (10 l of 0.1 M) in 4×SSC buffer for 90 min in a humidity chamber. Both the immobilization and hybridization processes were done outside of the flow cell. At each step of the surface modification, the surface state of the Au electrodes was monitored by coulostatic technique in presence of 5 M K3 Fe(CN)6 in PBS solution. Finally, after all coulostatic measurements were completed, the presence of the hybrid layer on the sensor surface was confirmed by coupling the enzyme Ex-aP with biotinylated dsONT (if formed) and its reaction with the substrate p-APP in pH 8.0 under stopped flow mode (Hintsche et al., 1994; Nebling et al., 2004). The change in current–time (i–t) curve slope was taken as the measurement of the layer density and hybridization efficiency at the electrode surface. A similar approach was adopted to detect human virus DNA sequences and M13 phage DNA.
2.5. Coulostatic instrumentation The used instrument set-up was a normal coulostatic pulse technique described in the literature (Reinmuth, 1962; Delahay, 1962; Pilla, 1970; Glass et al., 1998). The instrumental setup consists of a charging pulse generator (Hewlett Packard, 33120A) and a cathode ray oscilloscope (Textronix, TDS520B). As it has been proved that transient pulse coulostatic and the conventional frequency dependent impedance measurements both yield similar quantitative data (Glass et al., 1998; Yoo and Park, 2000; Park and Yoo, 2003) for charge transfer reactions, the same equivalent circuit, recently evaluated from independent impedance measurements (Dharuman et al., 2005), may be applicable to this system as well. Hence, Fourier or Laplace transformation of the data was not attempted here, and only change in the magnitude of potential relaxation curve with time was taken for qualitative analysis in this report. 3. Results and discussions
2.4. Flow through system A polycarbonate flow through cell with an internal volume of 4 l sealed by an O-ring onto the silicon chip was used as a reaction cell for all measurements. A needle type electrical connector was integrated with the flow cell. This allows both individual and parallel electrical measurements. While an individual position was monitored by coulostatic technique, the parallel measurements were made chronoamperometrically using a multi-channel potentiostat (Paeschke et al., 1996b) for surface layer stability monitoring. The flow rate was controlled by a peristaltic pump (Ismatec Glattbrugg-Z¨urich, CH) at 100 l/min.
3.1. Detection of tumour CK20 ONT sequence hybridization The capture probes I and II (10 M of 500 nl volume) are spotted onto an 8 position IDA shown in Fig. 1A, such that the capture I is immobilized on positions 1–4 and the capture II is immobilized on positions 5–8. The potential relaxation of these positions after capture probe immobilization was measured in presence of 5 M Fe(CN)6 3− . The charging pulse was generated at a frequency of 20 kHz and at potential amplitude of 9 V. All positions showed similar coulostatic behaviour indicating a constant coverage for ssDNA within the experi-
748
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
mental error of 5–10%. This variation might be caused by slight differences in the immobilized layer density at different positions. Therefore, prior to the immobilization and hybridization processes, the coulostatic behaviour of the bare positions was studied with the aim of characterizing the surface for parallel measurements in biosensor applications. The relaxation potentials measured at different positions arranged on the same and different chips are compared at 0.2 ms at a time constant of 50 s. It is nearly constant at 20 mV and varies from 2 to 10% between the positions. Next to this, the positional dependent behaviour of the array with a monolayer of 6-mercapto-1hexanol (MCH, 1 mM) was made without and with ssONT layer. Without ssONT, the MCH covered surface showed nearly three times increase (∼70 mV) in relaxation potential compared to the bare Au surface (20 mV), suggesting that the MCH layer behaves as an effective barrier at the sensor surface resulting in high impedance to the diffusion of the Fe(CN)6 3− ions (data not shown). Identical to the bare Au surface, the diffusion of Fe(CN)6 3− at different positions is not the same but varies to a negligible amount. The CV measurements conducted in parallel using a multi-channel potentiostat (Hintsche et al., 1994; Nebling et al., 2004) showed a suppressed quasi reversible peak for Fe(CN)6 3− in pH 7.4 for MCH layer immobilized Au surface than that of the bare Au array positions. The mixed monolayer was prepared in two steps. First, the ssONT was immobilized for 2 h and washed extensively with PBS buffer, followed by incubation of 1 mM MCH for 1.5 h. Comparison of these different monolayer behaviours suggests that the monolayer of ssONT covers nearly 60–70% while the mixed monolayer covers 90–95% (n = 3). These results are in corroboration with the literature (Finklea et al., 1993) that the diffusion limitation to Fe(CN)6 3− is increased for the mixed monolayer compared to the lone ssONT layer. That is, the diffusion of Fe(CN)3− to free sensor surface through the existing monolayer defects such as pin holes and collapsed sites causes small increase in capacitance of the electrodes observed for mixed monolayer.
The arrays carrying a ssONT monolayer are now allowed to hybridize with a target ONT (VII), which is complementary to the capture II at positions 5–8 and non-complementary to the capture I immobilized at positions 1–4. Fig. 2 compares the coulostatic relaxations of the complementary II (Fig. 2A) and non-complementary I (Fig. 2B) captures before (curve a) and after (curve b) hybridization in presence of Fe(CN)6 3− measured at two different positions. In order to examine the effect of measuring time constant on the relaxation behaviour of hybrids present on the surface, the potential relaxation is measured at six different time constants ranging from 5 s to 500 ms at every position. Significant increase in the relaxation potential curve after hybridization of ssONT is observed only in the microsecond time constant region and not in the millisecond regions. The insignificant signal shift for hybridized positions in the absence of Fe(CN)6 3− ions in the buffer indicates that the observed signal shift in the microsecond region is due to the diffusion of charged species. The charge or potential relaxation is not expected to be affected by the electroactive ions diffusion in the smaller time constant region such as microsecond time scale, while this can be observed only at longer time constants (Perverelli and Van Leeuwen, 1980). However, the slight influence by the diffusion of charged species during the relaxation cannot be eliminated completely even at smaller time constants (Reinmuth, 1962; Delahay, 1962; Pilla, 1970). Comparison of the complementary and non-complementary hybridized positions suggests that the diffusion of electro active Fe(CN)6 3− is affected by the formation of double strands effectively in the microsecond region (Fig. 2A, curves a and b) and a maximum discrimination effect was observed at 50 s. Therefore, for further studies, the relaxation potential magnitude change at 0.2 ms at 50 s time constant is taken as a qualitative indicator of hybrid formation. The results of five repeated measurements on the same position showed a constant behaviour indicating a satisfactory reproducibility within 5% for a given capture and target interaction. The impedance increase to the relaxation potential after hybridization is highly specific for complementary hybrid posi-
Fig. 2. Comparative coulostatic response curves observed for (A) complementary hybridization (at position 5) and (B) non-complementary hybridization (at position 3) measured on 8 position Au–IDA chip in presence of 5 M K3 Fe(CN)6 in PBS buffer. Curve a, ssDNA; curve b, dsDNA measured without methylene blue intercalation. Curve c, effect of methylene blue intercalation after hybridization. The charging pulse generated at 20 kHz and potential amplitude 9 V. The potential decay measured at a time constant of 50 s.
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
tions (Fig. 2A, curves a and b) than to the non-complementary positions (Fig. 2B). In other words, the Fe(CN)6 3− diffusion is hindered by the formation of dsONT on the electrode surface. Similar to these observations, the experiments with mixed mono layers also revealed identical behaviours. This clearly suggests that the layer defects are blocked restricting the free diffusion of Fe(CN)6 3− to the sensor surface. Another reason for the increased impedance due to the dsONT layer formation is the increased electrostatic repulsion for Fe(CN)6 3− due to the increased negative charge density on DNA layer (Finklea et al., 1993). Similar measurements were made on all other six different positions and compared for both complementary and non-complementary hybrid positions, Fig. 3A. The figure plots the potential relaxation signal magnitude difference observed before and after hybridization in presence of Fe(CN)6 3− . The results demonstrate a clear discrimination of ssONT and dsONT at all positions effectively by the coulostatic technique. Repeated measurements on different chip surfaces showed reproducibility within the relative standard deviation of 10–15% (n = 7) when the array was constructed in a similar fashion. However, we noticed positional differences as indicated in Fig. 3A among the hybridized positions. For example, the positions at the middle of the chip, namely 5 and 6 showed the signal difference of 70 mV while the corner positions, 7 and 8 are showing 35–40 mV. The near 50% activity difference observed may be due to difference in the activities of the
749
inner and cornered positions. But the behaviour is not observed with the lone MCH layer present at the positions. Therefore, the difference in the accessibility of Fe(CN)6 3− by middle and cornered positions when hybrids present on the surface may depend on the nature of the interactions of diffusion profile, the charged species and the solution flow mode. In order to increase the discrimination signal difference further we used dsDNA specific intercalator, methylene blue (MB) and Barton’s principle of mediated charge transport along dsDNA (Kelley and Barton, 1997; Boon et al., 2000). This was done by incubating the hybridized surface with 2 M MB in PBS buffer for 10 min. The coulostatic measurements were repeated on all positions in presence of 5 M K3 Fe(CN)6 . The results are included in both Fig. 2 (curve c) and Fig. 3B for comparison. It is noted that the charge transfer due to Fe(CN)6 3− after MB intercalation was greatly hindered for ssONT compared to dsONT for which the observed signal displacement is very small. This is in similar trend to the expected operation of Barton’s principle of enhancing the charge transport by dsONT intercalated MB in presence of solution borne Fe(CN)6 3− . The strong increase for the noncomplementary hybrid (ssONT, 140 mV) by MB interaction compared to the complementary hybrid (dsONT, 20–40 mV) indicates severe interaction of MB with non-hybridized ssONT. This may be due to high reactivity of MB for free guanine base (Kara et al., 2002) and negatively charged phosphate by charge compensation (Park and Hahn, 2004) which are said to be not
Fig. 3. Detection of tumour CK20 sequences DNA hybridization. Differences in relaxation potential (in volts) of ssDNA before and after hybridization with target DNA measured at 0.2 ms for non-complementary (at positions 1–4) and complementary (at positions 5–8) captures immobilised measured in presence of 5 M K3 Fe(CN)6 . (A) Without methylene blue intercalation. (B) With methylene blue intercalation. Other conditions are the same as in Fig. 2. (C) Chronoamperometric slopes observed from ELISA experiment.
750
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
assisting the mediated catalytic process of Fe(CN)6 3− (Boon et al., 2000) in solution. But in the present case, among the two interactions, the ionic interaction of positively charged intercalator dominates the intercalative effects. This means that the hybridization may be very less efficient so that some of ssONTs may be left unhybridized even on the complementary positions and therefore, the intercalative assisted charge transfer rate is not more effective to be observed. These facts necessitate the confirmation of dsONT formation and stability on the surface by other techniques. We have utilised ELISA like experiments by target labelling with Ex-aP and p-APP redox recycling principle at IDAs described in our earlier reports (Hintsche et al., 1994; Nebling et al., 2004). The hybrid formation is indicated by increase in the slope of i–t curve while non-hybrid positions showed negligible change. Current–time curve (i–t) slopes observed for both non-complementary and complementary positions are presented in Fig. 3C. That is, the slope of the i–t curves raised steeply at the hybridized positions (∼250 nA/min) due to the Ex-aP catalyzed conversion of p-APP → p-AP, and p-AP diffuses to the IDA surface where it undergoes the reversible reaction, p-AP ↔ quinoimine, at constant applied potentials +300 (at anode) and −50 mV (at cathode) (Ag/AgCl). The same reaction was hindered at non-hybridized positions due to the absence of Ex-aP indicated by small values around 20 nA/min. The small increase of i–t slope for non-complementary positions may be due to the presence of a very small quantity of electroactive pAP in the commercial sample which can diffuse to the electrode surface through the monolayer defects. Various other experimental parameters studied are ionic concentration of the buffer electrolyte, different capture and target concentrations. The minimum concentration of oligonucleotides related to tumor CK20 sequences that can be detected by this method is 0.01 M under optimal conditions. 3.2. Detection of Viral and M13 phage DNAs The detection of human viral infectious DNAs like herpes simplex virus (HSV), cytomegalovirus (CMV) and Epstein–Barr virus (EBV) is vital in clinical diagnosis and drug monitoring therapy of psoriasis (Mehraein et al., 2004; Ibrahim et al., 2005) and in earlier acute coronary syndromes (Miller et al., 2005). Onto the eight-position array chip shown in Fig. 1A, the complementary captures HSV (III), CMV (IV) and the noncomplementary captures EBV (V) and CK20 (II) (this capture is non-complementary to CMV and HSV targets) are microspotted on four different positions. Hybridization was made using a mixture of complementary targets (10 l) HSV (VIII) and CMV (IX) (each 5 l of 1 M concentration). The comparative coulostatic results observed in presence of 5 M K3 Fe(CN)6 after intercalation with 2 M MB showed ∼80 mV (measured at 0.2 ms of coulostatic curve) for complementary positions and ∼50 mV for non-complementary positions for the reason discussed before. However, compared to tumour CK20 dsONT (∼20–40 mV), the MB signal for viral DNA (∼50 mV) is increased probably due to different hybridization efficiency to form dsONT. The i–t slopes from enzyme Ex-aP labelling showed 531.47
and 262.60 nA/min for complementary positions and 12.42 and −23.92 nA/min for non-complementary positions confirming the perfect discrimination by the coulostatic technique. Because the hybridization efficiency is strongly dependent on the length of the capture and target molecules as well (Okahata et al., 1992; Huang et al., 2001), we have examined the detection of M13 phage DNA hybridization with its capture probe (VI) immobilized on two positions. Here, the ONT capture (I) was used as negative control to M13 KO7 target ONT (XI) and immobilized onto two other positions on the same chip surface. The detection scheme involves immobilization of M13 capture (VI) followed by its sequential hybridization with M13 phage DNA natural sample (X) and detection target DNA (XI). The coulostatic measurements were made only to know the effect of hybridization of M13 KO7 helper phage with its complementary capture probe (VI) in presence of 5 M K3 Fe(CN)6 . The MB intercalation study was neglected due to its strong interaction with free guanine bases and phosphate backbone of ssDNA which makes data processing difficult, if the target causes unspecific binding (Kara et al., 2002). The biotinylated target was used just to ensure the confirmation test following the enzymatic conjugate Ex-aP and p-APP redox recycling method after coulostatic technique, similar to the tumour and viral DNA tests. Compared to the tumour CK20 and human viruses (HSV, CMV, EBV) systems (∼40 mV), the observed signal difference before and after hybridization with M13 KO7 phage is nearly three times (∼120 mV) higher in presence of Fe(CN)6 3− alone, for hybrid positions (115 and 110 mV) and 25–35 mV for non-complementary positions. This is attributed to the large number of base pairs that are electrostatically repelling the Fe(CN)6 3− ion diffusion. The i–t slopes from ELISA measurements showed 259.78 and 274.86 nA/min for hybrid positions and 11.58 and 4.87 nA/min for non-complementary positions. This suggests the easy, simple discrimination and detection of natural DNA hybridization on these chip platforms very well. These experiments have been reproduced three times, with R.S.D. of 6%. 4. Conclusions The application of coulostatic technique in biomolecular affinity interaction was explored for the first time in biosensor research. The comparison of the signal change magnitudes of short chain ONTs hybridization and larger base pair DNAs such as M13 phage DNAs indicates the technique’s sensitiveness to the molecular environment change at the electrode film/solution surface. Further, the results suggested its effective sensing ability in the microsecond time constant scale which is quite fast compared to the conventional impedance technique. The tandem use of coulostatic and chronoamperometric redox recycling principle following the enzyme Ex-aP labelling, corroborates the surface coverage as well as the hybridization efficiency on the electrode surface. Further advantage of this system is the elimination of the reference electrode and the use of a two pole measurement. This is more convenient for the development of miniaturized devices with array sensor integration. With the preliminary results observed in the present study and the literature
V. Dharuman et al. / Biosensors and Bioelectronics 22 (2006) 744–751
evidences (Yoo and Park, 2000; Alvarez et al., 2002; Park and Yoo, 2003; Jurczakowski and Lasia, 2004) on the potential use of transient pulse techniques to replace the time consuming impedance technique to overcome the difficulty especially in the low frequency region, we expect a major breakthrough in label free sensing in relation to fastness and simplicity in data acquisition. It may be recalled here that hybridization interaction was monitored just by negotiating the change in the signal magnitude of the ssONT upon its hybridization with its target; and therefore, future attention will be given to the optimization of the accuracy of the method with reference to the conventional impedance method. Acknowledgement We thank the Fraunhofergesellschaft and e-BiochipSystems GmbH for the financial support for this project. References Albers, J., Grunwald, T., Nebling, E., Piechotta, G., Hintsche, R., 2003. Anal. Bioanal. Chem. 377, 521–527. Alvarez, A., Sanchez, M., Pinhole, C.A., 2002. Corros. Rev. 20, 69–93. Bart, M., Stigter, E.C.A., Stapert, H.R., de Jong, G.J., van Bennekom, W.P., 2005. Biosens. Bioelectr. 21, 49–59. Boon, E.M., Ceres, D.M., Drummond, T.G., Hill, M.G., Barton, J.K., 2000. Nat. Biotechnol. 18, 1096–1100. Berney, H., West, J., Haefele, E., Alderman, J., Lane, W., Collins, J.K., 2000. Sens. Actuators B 68, 100–108. Cloarec, J.P., Deligianis, N., Martin, J.R., Lawrence, I., Souteyrand, E., Polychronakos, C., Lawrence, M.F., 2002. Biosens. Bioelectr. 17, 405–412. Danilov, F., Obraztsov, V., Kapitonov, A., 2003. J. Electroanal. Chem. 552, 69–76. Delahay, P., 1962. Anal. Chem. 34, 1267–1272. Dharuman, V., Grunwald, T., Nebling, E., Albers, J., Blohm, L., Hintsche, R., 2005. Biosens. Bioelectr. 21, 645–654. Ehret, R., Baumann, W., Brischwein, M., Schwinde, A., Stegbauer, K., Wolf, B., 1997. Biosens. Bioelectr. 12, 29–41. Estrela, P., Migliorato, P., Takiguchi, H., Fukshima, H., Nebashi, S., 2005. Biosens. Bioelectr. 20, 1580–1586. Finklea, H., Sinder, D.A., Fedyk, J., Sabatani, E., Gafni, Y., Rubinstein, I., 1993. Langmuir 9, 3660–3667. Frace, G., Lillie, G., Hianik, T., Payne, P., Vadgama, P., 2002. Bioelectrochemitsry 55, 1–3. Glass, G.K., Hassanien, A.M., Buenfeld, N.R., 1998. Electrochim. Acta 43, 1863–1871. Guiducci, C., Stagni, C., Zuccheri, G., Bogliolo, A., Benini, L., Samori, D., Ricco, R., 2004. Biosens. Bioelectr. 19, 781–787. Gheeorghe, M., Elie, A.G., 2003. Biosens. Bioelectr. 19, 95–107. Guan, J.G., Miao, Y., Zhang, Q., 2004. J. Biosci. Bioeng. 97, 219–226. Hang, T.C., Elie, A.G., 2004. Biosens. Bioelectr. 19, 1537–1548. Hintsche, R., M¨oller, B., Dransfeld, I., Wollenberger, U., Scheller, F., Hoffmann, B., 1991. Sens. Actuators B 4, 287–291. Hintsche, R., Paeschke, M., Wollenberger, U., Schnakenberg, U., Wagner, B., Lisec, T., 1994. Biosens. Bioelectr. 9, 697–705.
751
Huang, E., Satjapipat, M., Han, S., Zhou, F., 2001. Langmuir 17, 1215– 1224. Ibrahim, A.L., Obeid, M.T., Jouma, M.J., Moasis, G.A., Al-Richane, W.L., Kindermann, I., Boehm, M., Roemer, K., Lantzch, N.M., Gartner, B., 2005. J. Clin. Virol. 32, 29–35. Jacobs, P., 1998. Proceedings of the Second International Conference on Microreaction technology, New Orleans, Louisiana, USA, March 8–12. Jurczakowski, R., Lasia, A., 2004. Anal. Chem. 76, 5033–5038. Katz, E., Willner, I., 2003. Electroanalysis 15, 913–947. Kelley, S.O., Barton, J.K., 1997. Bioconju. Chem. 8, 31–37. Kara, P., Kerman, K., Ozkan, D., Meric, B., Erdem, A., Ozkan, Z., Ozsoz, M., 2002. Electrochem. Commun. 4, 705–709. Laureyn, W., Nelis, D., Van Gerwen, P., Baert, K., Hermans, L., Magnee, R., Pireaux, J.-J., Maes, G., 2000. Sens. Actuators B 68, 360–370. Lillie, G., Payne, P., Vadgama, P., 2001. Sens. Actuators B 78, 249–256. Mehraein, Y., Lennerz, C., Ehlhardt, S., Zhang, K.D., Madry, H., 2004. J. Clin. Virol. 31, 25–31. Miller, G.E., Freedland, K.E., Duntley, S., Carney, R.M., 2005. Am. J. Cardiol. 95, 317–321. Montelius, L., Tegenfeldt, J.O., Ling, T.G., 1995. J. Vac. Sci. Technol. A 13, 1755–1760. Menikh, A., Mickan, S.P., Liu, H., MacColl, R., Zhang, X.C., 2004. Biosens. Bioelectr. 20, 658–662. Nebling, E., Grunwald, T., Albers, J., Sch¨afer, J., Hintsche, R., 2004. Anal. Chem. 76, 689–696. Niemeyer, C.M., Blohm, D., 1999. Angew. Chem. Int. Ed. 38, 2865–2869. Okahata, Y., Matsunobu, Y., Ijiro, K., Mukae, M., Murakami, A., Makino, K., 1992. J. Am. Chem. Soc. 114, 8299–8300. Paeschke, M., Bucmann, L.M., Seitz, R., Hintsche, R., 1996a. Microsystems Technologies, vol. 96. VDE Verlag, pp. 399. Paeschke, M., Dietrich, F., Uhlig, A., Hintsche, R., 1996b. Electroanalysis 8, 891–898. Park, S.M., Yoo, J.S., 2003. Anal. Chem. 75, 455a–461a. Park, N., Hahn, J.H., 2004. Anal. Chem. 76, 900–906. Peng, H., soeller, C., Vigar, N., Kilmartin, P.A., Cannell, M.B., Bowmaker, G.A., Cooney, R.P., Sejdic, J.T., 2005. Biosens. Bioelectr. 20, 1821–1828. Peterson, A.W., Heaton, R.J., Georgiadis, R.M., 2001. Nucl. Acids Res. 29, 5163–5168. Perverelli, K.J., Van Leeuwen, H.P., 1980. J. Electroanal. Chem. 110, 119–135. Pilla, A.A., 1970. J. Electrochem. Soc. 117, 467–477. Ramsay, G., 1998. Nat. Biotechnol. 16, 40–44. Reinmuth, W.H., 1962. Anal. Chem. 34, 1272–1276. Saum, A.E.G., Cumming, R.H., Rowell, F.J., 1998. Biosens. Bioelectr. 13, 511–518. Souteyrand, E., Chen, C., Cloarec, J.P., Nesme, X., Simonet, P., Navarro, I., Martin, J.R., 2000. Appl. Biochem. Biotechnol. 89, 195–207. Stulik, K., Amatore, C., Holub, K., Marecek, V., Kutner, W., 2000. Pure Appl. Chem. 72, 1483–1492. Wang, J., 2000. Nucl. Acids Res. 28, 3011–3016. Wie, F., Sun, B., Guo, Y., Zhao, X.S., 2003. Biosens. Bioelectr. 18, 1157– 1163. Van Gerwen, P.V., Laureyn, W., Laureys, W., Huyberechts, H., Op De Beeck, M., Baert, K., Suls, J., Sansen, W., Jacobs, P., Hermans, L., Martens, R., 1998. Sens. Actuator B 49, 73–80. Yang, L., Li, Y., Griffis, C.L., Johnson, G.M., 2004a. Biosens. Bioelectr. 19, 1139–1147. Yang, L., Li, Y., Erf, G.F., 2004b. Anal. Chem. 76, 1107–1113. Yoo, J.S., Park, S.M., 2000. Anal. Chem. 72, 2035–2041. Zhao, Y., Guo, X., Dong, Z., 2004. Corros. Eng. Sci. Tech. 39, 245–249.