Introduction

The detection of DNA sequences has recently evolved into an important issue due to its wide applicability to a range of different fields. Classical techniques used for DNA analysis are based on direct DNA sequencing [1] or DNA hybridization [2]. Though the former is indispensable in some cases, it is often considered to be too time-consuming [3]. The hybridization of two complementary DNA single strands to form a double helix is the phenomenon exploited in biosensor analysis in order to detect DNA sequences [4]. Hybridization is thus rapidly becoming the default gene assay procedure; labeled oligomers are normally used in such procedures in order to achieve spectrophotometric transduction properties. Furthermore, hybridization can be coupled to selective amplification by using stages from the polymerase chain reaction process. Hybridization procedures can be used to create DNA biosensors—genosensors—which are currently undergoing rapid development as a tool for the diagnosis of infectious or genetic diseases, environmental screening, various analyses performed in the food industry, forensic investigations, etc. The use of genosensors yields some advantages over classical DNA analysis, i.e., direct sequencing by chromatography or electrophoresis after the fragmentation of DNA [5, 6], including their low cost, the simple equipment involved, the possibility of miniaturization and analysis in situ, and the speed of analysis. Genosensors can be classified, depending on the technique employed for transduction, into optical, piezoelectric and electrochemical types. Electrochemical genosensors make use of electrochemical transduction for hybridization detection. Several variants of this type of sensor can be distinguished, such as those based on the use of electroactive indicators called intercalators [7, 8], techniques employing DNA sequences labeled with enzymes or different ad hoc electrochemically functional groups [9, 10], or procedures based on the guanine oxidation signal [11–13]. Recently, among the different electrochemical techniques available, electrochemical impedance spectroscopy (EIS) [14, 15] has been used in numerous studies on DNA hybridization detection [16, 17]. This technique is very sensitive to changes in the interfacial properties of modified electrodes caused by biorecognition events at the electrode surface [18, 19]. For this reason, EIS is becoming an attractive electrochemical tool for numerous applications such as immunosensing [20–22] and genosensing [23–25], enzyme activity determinations [26, 27], and studies of corrosion [28, 29] or other surface phenomena [8]. Using this technique, each spectrum obtained can be associated with an equivalent electrical circuit that represents the experimental electrochemical system [14, 15]. Each electrical element in the circuit is directly correlated to a simple process such as electron transfer resistance, diffusion, double layer formation or capacitance change. Many studies on the application of EIS to DNA hybridization detection have been reported [30]. Some protocols are based on direct double-layer capacitance measurements (without using any redox label) before and after hybrid formation [31–37]. Other variants add a redox marker (i.e., an electroactive species) for the transduction [24, 38–43]. The application of this technique in conjunction with DNA hybridization detection, also allows the investigation and characterization of the single layer formed after probe immobilization [35, 37, 44] or the determination of single-nucleotide polymorphisms [42, 45, 46]. This detection technique is now starting to be applied to different fields of genosensing, such as the detection of the genes involved in various diseases [47], species differentiation [48], or investigations into food and water quality [49].

Over the years, one key issue for biosensors employed for DNA hybridization detection has been the choice of the immobilization protocol used to fix the DNA probe onto the electrode surface. Various options have been described [50]. Some of them are based on simple physical adsorption of the probe onto the electrode surface [50–52], while others make use of covalent bond formation between the electrode surface and a labeled probe [53–55]. Recently, in more innovative protocols, the probe has been immobilized onto the sensor surface by means of a conductive polymer, such as polypyrrole [23, 56, 57] or polyaniline [58, 59], or by taking advantage of the strong interactions between biomolecules, such as those involved in biotin–avidin complex formation. This interaction, due to its strength (association constant K a = 1015 M), has been studied extensively and exploited by several authors for the immobilization of several types of biomolecules, such as enzymes, proteins, antibodies and oligonucleotides [60–64].

In this work, an avidin bulk-modified graphite–epoxy composite electrode (Av-GEB) was used. The presence of avidin molecules as anchorage points on the conductive biopolymer allows the easy immobilization of a biotin–modified DNA molecules onto the electrode surface. Graphite–epoxy composite (GEC) has been extensively studied and used by our research group due to its improved electrochemical performance and versatility as transducer material [65, 66]. The bulk modification [67] of this conductive paste has been already reported for specific applications—immunosensing [68, 69] and genosensing [60, 70]—by the same group. These sensors present exceptional properties that make them good transducers and reservoirs of biological materials, thus allowing their regeneration and repeated use [71]. Moreover, biocomposites can be rapidly prepared using dry-conditions chemistry [67], hence avoiding the need for more time-consuming and expensive immobilization procedures. One of the main advantages of using this material is that the sensor surface can be easily renewed after each analysis by simply polishing with abrasive paper, and performing a one-step immobilization–hybridization procedure [72].

This work is novel in that it represents the application of electrochemical impedance spectroscopy to DNA hybridization detection in a biocomposite sensor. EIS has been already employed by the same authors for genosensing with a different platform. In the present paper it is applied for the first time to DNA analysis performed with Av-GEB biocomposites. Using this technique, it is possible to perform very rapid and simple determinations since it is not necessary to use labeled DNA for detection purposes. Moreover, by employing a redox marker, a very clear and well-defined change in the experimental signal is achieved. The study first focuses on the detection of hybridization of homooligonucleotides, and it is then extended to the specific detection of a sequence related to Salmonella spp [73]. Insertion sequence IS200 has already been employed in different works from the same research group using amperometric [74] and voltammetric [13, 75] detection. This sequence is a transportable element some of 700 bp in length, and it is present in more than 90% of the pathogenic or food poisoning isolates of Salmonella spp [76], and can thus be used to detect outbreaks of infection in food as well as in clinical fields.

Experimental

Materials

Reagents were commercially available and were all of analytical reagent grade. All solutions were made up using doubly distilled water.

Tris–HCl buffer, avidin (ref. A9275), biotin-4-fluorescein (ref. B9431), potassium ferricyanide K3[Fe(CN)6] and potassium ferrocyanide K4[Fe(CN)6] were purchased from Sigma (St. Louis, MO, USA). The various oligonucleotides used in the study were prepared by TIB MOLBIOL (Berlin, Germany). Their sequences and modifications are shown in Table 1. Oligonucleotide stock solutions were diluted with sterilized and deionized water and stored at a temperature of −20 °C until required.

Table 1 Sequences of oligonucleotides used and their modifications
Full size table

The following buffers were employed: Tris (0.01 M Tris-HCl, pH 7.5), 0.1 M PBS (0.1 M NaCl, 0.01 M sodium phosphate buffer, pH 7.0), TSC1 (0.75 M NaCl, 0.075 M trisodium citrate, pH 7.0), TSC2 (0.30 M NaCl, 0.030 M trisodium citrate, pH 7.0).

Graphite–epoxy composite electrodes (GECs) were prepared using graphite powder (50 μm particle size; BDH Laboratory Supplies, Poole, UK) and Epotek H77 resin and hardener (both from Epoxy Technology Inc., Billerica, MA, USA). Avidin was added to graphite–epoxy conductive paste to prepare avidin-modified biocomposites (Av-GEB).

Equipment

AC impedance measurements were performed with an IM6e Impedance Measurement Unit (BAS-Zahner, Kronach, Germany). Thales software was used to acquire the data and control the experiments.

A three-electrode cell was used to perform the impedance measurements. The cell comprised an Ag/AgCl reference electrode (an AgCl-covered silver wire), a ring-platinum electrode (Crison 52-67-1, Barcelona, Spain), and the constructed graphite–epoxy composite (GEC) or avidin-modified biocomposite (Av-GEB) working electrodes.

Temperature-controlled incubations were performed in an Eppendorf (Hamburg, Germany) 5436 Thermomixer.

A Leica (Wetzlar, Germany) TCS SP2 AOBS microscope was used to take confocal laser scanning microphotographs of the electrodes (laser excitation: 568 nm; voltage: 352 V).

Preparation of GEC and Av-GEB electrodes

All solid-state working electrodes were prepared using a PVC tube body (A) (20 mm long, 6 mm i.d.) and a small copper disk soldered to the end of an electrical connector. The conductive part of the GEC was an epoxy–graphite conductive paste, formed from graphite (20%) and epoxy resin (80%), which was deposited into the cavity in the plastic body, filling it (as shown in Fig. 1). The composite material was cured at 80 °C for two days. Av-GEB electrodes were prepared using the same conductive paste; for every gram of graphite–epoxy mixture, the proper amount of avidin was added, resulting in a 2% (w/w) avidin–graphite epoxy biocomposite. The biocompostite material was cured at 40 °C for one week, and then stored at 4 °C before and after use. The geometrical conducting area of the electrode was thus 28.26 mm2.

Fig. 1
figure 1

Genosensor assembly: a, electrode body; b, attachment to the copper disk; c, mounting of the PVC body; d, introduction of the graphite–epoxy paste in order to obtain a GEC electrode; e, introduction of the avidin-modified graphite–epoxy paste to obtain an Av-GEB electrode

Full size image

Each time before it was used, the surface electrode was wetted with doubly distilled water and then thoroughly smoothed with abrasive paper and then with alumina paper (polishing strips 301044-001, Orion, Boston, MA, USA). The reproducibilities of the construction and polishing processes [74] for sensors based on graphite–epoxy composites, as well as the stability and optimization of Av-GEB biocomposites, have been reported previously [60].

Procedures

Two different procedures were employed in this work. The first one is called “one step” immobilization–hybridization. This means that DNA is immobilized onto the electrode surface when the hybrid between the probe and the target has already been formed, and EIS measurement is performed only once, after DNA double-strand immobilization. This procedure was employed during DNA testing.

The second one, used when constructing a calibration curve, involves different steps (probe immobilization, blocking step, and hybridization with target). EIS measurements were taken after each step.

DNA testing

The first step of the protocol was DNA hybrid formation in a solution containing the biotinylated probe and its complementary target. The bulk-modified graphite–epoxy biocomposites (Av-GEB) were then employed to immobilize the readily obtained double-stranded biotinylated DNA onto the electrode surface (in a one-step immobilization hybridization procedure).

The procedure is described in detail below:

  • Biotinylated double-stranded DNA (bio-dsDNA) formation: 30 pmol of biotinylated DNA probe and 30 pmol of its complementary target in TSC1 buffer were mixed in an eppendorf tube, as shown in Fig. 2a; hybridization was encouraged with gentle stirring in a thermomixer at 42 °C for 30 min.

  • Bio-dsDNA immobilization onto the Av-GEB surface: an avidin-modified electrode was dipped into an eppendorf containing biotinylated hybrid, as shown in Fig. 2b. The tube with the sensor was gently shaken for 20 min at 42 °C. The previously formed bio-dsDNA then became immobilized onto the electrode surface due to the high affinity between the avidin in the conductive paste and the biotin of the biotinylated hybrid. This was followed by two washing steps with Tris at 42 °C for 10 min.

Fig. 2a, b
figure 2

a Hybridization in solution between biotinylated-DNA probe and a DNA complementary target. b Immobilization strategy

Full size image

For all incubations at controlled temperature in the thermomixer, the stirring speed employed was 600 rpm. The oligonucleotide probe concentration was optimized in previous studies. An amount of 30 pmol (in 140 μL of TSC1 buffer) corresponded to the complete saturation of the Av-GEB platform with biotinylated DNA.

The oligonucleotide target concentration used during DNA testing was fixed at 30 pmol (in 140 μL of TSC1 buffer) after observing the results obtained during the construction of the calibration curve.

Calibration curve

The analytical procedure used during the construction of the calibration curve consisted of immobilizing the biotinylated DNA probe on the electrode surface followed by the hybridization of the former with the complementary target at different concentration values. A blocking step was performed between immobilization and hybridization to block the adsorption of the target onto the free electrode surface.

More details are given below:

  • Biotinylated DNA probe (bio-ssDNA) immobilization: an avidin-modified electrode was dipped into an eppendorf tube containing 30 pmol of biotinylated DNA probe in TSC1 buffer solution. The total volume of the solution in each tube was 140 μL. This volume was chosen because it allowed the complete coverage of the electrode surface when dipped in the solution. The tube was then incubated at 42 °C with gentle stirring for 20 min. Due to the high affinity between the avidin contained in the conductive paste and the biotin in the probe, the bio-ssDNA was fixed onto the electrode surface. The incubation was followed by two washing steps with Tris buffer in the thermomixer at 42 °C for 10 min, thus eliminating any unbound DNA probe.

  • Blocking step: the probe-modified electrode was dipped into an eppendorf tube containing a solution with a certain concentration of a noncomplementary target sequence, thus blocking the free electrode surface. The tube was then incubated at 42 °C with gentle stirring for 20 min. This step was followed by two gentle washing steps in TSC2 buffer for 5 min at 42 °C.

    The possibility of using a more classical blocking agent like bovine serum albumin (BSA) was discarded due to the high impedance change caused by the adsorption of this protein onto the electrode surface. This leads to decreased signal variation after hybridization, thus making it more difficult to interpret the results. Considering that at the working pH both the protein and oligomer are negatively charged, the different behavior during impedance analysis must be due to the increased size [77, 78] of the considered BSA molecules.

  • Target hybridization: after the blocking step, the probe-modified electrode was incubated in a solution containing a certain concentration of complementary target sequence and maintained at 42 °C with gentle stirring for 30 min, thus allowing hybridization. The incubation was followed by two washing steps with Tris buffer at 42 °C for 10 min. Any nonspecific adsorption of target onto the electrode surface was avoided by using the blocking step.

    For all incubations performed at controlled temperature in the thermomixer, the stirring speed employed was 600 rpm.

EIS detection

All measurements were performed in 0.1 M PBS buffer solution containing a 0.01 M K3[Fe(CN)6]/K4[Fe(CN)6] (1:1) mixture, used as a redox marker [38]. Impedance experiments were carried out at an applied potential of 0.17 V (vs. Ag/AgCl reference electrode), with a range of frequency of 50 KHz to 0.05 Hz and an AC amplitude of 10 mV. The obtained spectra were represented as Nyquist plots (−Z im vs. Z re) in the complex plane. The theoretical curve used to fit the experimental data corresponds to an equivalent circuit consisting of a resistance in series with a capacitor element (i.e., a capacitor and a resistor in parallel). The chi-square goodness of fit was calculated for each fitting by the FRA software employed (Eco Chemie, Utrecht, The Netherlands).

The procedures followed during the EIS measurements were different when constructing the calibration curve and when testing the DNA.

DNA testing

In this case, the sequence followed for spectral registration was: (1) bare Av-GEB electrode; (2) biotinylated hybrid modified electrode. Noncomplementary target sequences as well as the buffer solution alone were used as negative controls in the hybridization step. A further negative control was performed during these experiments: a nonbiotinylated DNA sequence was used during the incubation step to confirm that the signal variation is attributable only to the biotinylated hybrid immobilization due to the avidin–biotin interaction, and not to any nonspecific adsorption onto the electrode surface. With the same aim, all of the experiments were also repeated employing GEC electrodes without any avidin in the bulk of the composite.

Different DNA sequences have been tested, as explained below.

  • Homooligonucleotides detection: d(pA)20-bio was used as biotinylated probe and d(pT)20 as its complementary target; d(pC)20 was employed as a noncomplementary target and d(pA)20 as a nonbiotinylated probe for negative controls.

  • Salmonella spp detection: the IS200 probe sequence was used as the biotinylated probe and the IS200 target sequence as its complementary target; d(pT)20 was employed as a noncomplementary target for negative controls.

Calibration curve

In these experiments, the sequence of EIS spectra recorded was: (1) bare Av-GEB electrode; (2) probe-modified electrode before blocking step; (3) probe-modified electrode after blocking step; (4) hybrid-modified electrode. Noncomplementary target sequences as well as the buffer solution alone were used as negative controls in the hybridization step in order to confirm that the signal variation was only due to hybrid formation.

For this study, the oligonucleotide sequences employed were as follows: d(pA)20-bio as the biotinylated probe; d(pC)20 as the blocking agent; d(pT)20 as the complementary target.

Microscopic characterization of Av-GEB electrodes

The Av-GEB electrode surface was studied with a confocal laser scanning microscope to confirm the presence of active binding sites due to the avidin groups on the surface. Biotin-4-fluorescein was thus immobilized onto the electrode surface using the same procedure previously reported for DNA. The electrode was then thoroughly washed several times with Tris buffer to avoid any nonspecific adsorption of the molecule on the surface.

The modified electrode surface was examined with the microscope: only if avidin is present, biotin-4-fluorescein is immobilized through its interaction with the avidin, thus producing fluorescence. A negative control was performed using an Av-GEB electrode treated with biotin-4-fluorescein after the electrode had been exposed to a biotin solution, thus blocking all of the active sites of avidin with biotin.

Images were taken by integrating the fluorescence in five different planes in which the signal was detected (the distance between each section corresponded to 2 μm).

Results and discussion

The modification of the electrode surface with DNA sequences as well as the DNA hybridization phenomenon can be extensively studied by electrochemical impedance spectroscopy using a redox marker such as K3[Fe(CN)6]/K4[Fe(CN)6] in the bulk solution.

A typical spectrum obtained in these experiments is shown in Fig. 3. The equivalent circuit R1(R2Q1), consisting of one resistor/capacitor element in series with a resistance, was found to give the best fit to the experimental data. The chi-square goodness-of-fit, calculated for each fit by the FRA software employed, was systematically checked to validate all of the calculations performed. For all of the cases studied in this work, calculated values of chi-square for the mentioned equivalent circuit were in the range 0.0001–0.1, much lower than the tabulated value for 50 degrees of freedom (67.505 at the 95% confidence level), thus demonstrating the high significance of the final fits.

Fig. 3
figure 3

Typical Nyquist diagram obtained with the Av-GEB tests and the equivalent circuit used to fit the obtained data

Full size image

In the equivalent circuit shown in Fig. 3, the parameter R1 corresponds to the resistance of the solution, R2 is the charge transfer resistance (also called Ret) between the solution and the electrode surface, whilst Q1 is associated with the double-layer capacitance (due to the interface between the electrode surface and the solution). The use of a constant phase element (Q1) instead of a capacitor is required to optimize the fit to the experimental data, and this is due to the nonideal nature of the electrode surface [14, 18]. The parameter of interest in our case is represented by the charge transfer resistance R2. This value in the spectrum corresponds to the diameter of the semicircle. When the electrode surface is modified with a DNA sequence such a ssDNA or a dsDNA, the value of R2 is altered. This is because the charge transfer process for the redox marker in the bulk solution is greatly influenced by any surface modification. In our case, we record an increase in R2 due to some hindrance to the arrival of marker species at the electron transfer sites at the electrode surface [38]. By comparing the spectra obtained before and after the different modifications of the electrode surface, it was possible to detect DNA hybridization.

To validate the immobilization of DNA through biotin–avidin interactions, the experiments were compared with negative controls, where nonbiotinylated DNA sequences or unmodified avidin electrodes were employed.

DNA biosensing

Characterization using homooligonucleotides

Typical EIS spectra obtained in these experiments are shown in Fig. 4. The EIS signal was recorded before (bare electrode) and after the immobilization of biotinylated-dsDNA onto the electrode surface (hybrid-modified electrode). The first graph (Fig. 4a) illustrates that the charge transfer resistance, corresponding to the diameter of the semicircle in the spectrum, increases after sensor surface modification. This is because the DNA layer formed on the electrode surface hinders the electron transfer process between the solution and the transducer [31]. Since the nucleic acids and the resulting nucleic acid/DNA complex are oligoanionic polymers, their immobilization onto the electrode surface generates a negative charged interface. A negatively charged redox marker such as the [Fe(CN)6]3−/4− used in this work is then strongly repelled from such an interface. The effect of this repulsion is an inhibition of the interfacial electron transfer process, thus increasing Ret [16].

Fig. 4a, b
figure 4

a Nyquist diagrams for EIS measurements of (circles) bare Av-GEB electrode, (triangles) biotinylated-d(pA)20-d(pT)20 hybrid-modified electrode; b Nyquist diagrams for EIS measurements of (circles) bare Av-GEB electrode, (triangles) biotinylated-d(pA)20 probe-modified electrode (corresponding to a noncomplementary target experiment). All hybridization experiments were performed in TSC1 buffer solution; all EIS measurements were performed in 0.1 M PBS buffer solution containing 0.01 M K3[Fe(CN)6]/K4[Fe(CN)6]

Full size image

A smaller increase is observed when a noncomplementary target is employed during the hybridization step (Fig. 4b). This is because the two ssDNA are noncomplementary, so hybrid formation does not occur in the hybridization solution, thus leading to the immobilization of just biotinylated probe onto the electrode surface. The increase in Ret due to the immobilization of the probe alone is clearly smaller than the increase in the same parameter due to the biotinylated-d(pA)20−d(pT)20 immobilization.

The results are expressed as as increase in Ret between the bare electrode and the DNA-modified electrode (Δ). This data processing was required to correct the results obtained for any variations due to transducer construction or to the polishing procedure performed after each use. Actually, the Ret value is different for each bare electrode, and it can even change for the same sensor after a surface polishing treatment. In fact, the electrode surface had to be renewed after use each time, thus producing superficial differences which could modify the bare electrode’s Ret value. For this reason, in order to compare results from different experiments, the increase in Ret was used instead of the absolute Ret value.

In this experiment, nonspecific adsorption was avoided due to the one-step immobilization/hybridization of dsDNA onto the electrode surface. In this procedure, only the biotinylated hybrid produced previously in the hybridization solution could be immobilized due to the presence of avidin molecules on the Av-GEB platform. At the same time, variations in Ret due to the possible adsorption of biotinylated hybrid onto the electrode surface were not significant. This meant that a blocking step is not useful in this procedure. In Table 2, the results obtained for hybridization plus various negative controls are shown.

Table 2 Results obtained for EIS measurements performed during homooligonucleotide characterization
Full size table

It is clear that the variation in Ret was almost double for hybrid immobilization compared to the DNA probe alone. The low signal obtained for the negative control (no. 4), where a nonbiotinylated hybrid was used in experiments with Av-GEB electrodes, demonstrates that when the avidin–biotin interaction does not occur, the variation in Ret is not significant. The adsorption of biotinylated DNA onto electrodes that are not modified with avidin (GEC electrodes, experiment no. 3) cannot be considered to be completely negligible. In fact, as already reported in previous works [13, 39], the platform was successfully used in DNA physical adsorption experiments; however, the signal obtained in this experiment can be considered to be sufficiently different from the one obtained in the Av-GEB experiments.

Good reproducibility was attained for the measurements, as also shown in the table. In fact, for the hybridization experiments, the RSD value was around 4%.

A Student’s t-test was carried out to compare results obtained from the hybridization and nonspecific hybridization assays for homooligonucleotide characterization (experiments 1 and 2 in the table). The difference between the mean values was highly significant, t calc (9.99) > ttab (2.13), at the 95% confidence level, for four degrees of freedom. This confirms that there is a statistically significant difference between the signal obtained during the hybridization experiments and the one obtained during nonspecific hybridization experiments. The data used to perform the Student’s t-test are shown in Table 3.

Table 3 Triplicate measurements that were used to perform the Student’s t-test for homooligonucleotides and Salmonella spp detection (Δ = Ret (sample) − Ret (blank))
Full size table

Salmonella spp detection

The same types of experiments were performed for the IS200 gene sequence analysis. In this case, the EIS signal was recorded before (bare electrode) and after the immobilization of IS200 double-stranded DNA onto the electrode surface (hybrid-modified electrode). Like before, the value of the charge transfer resistance increased after sensor surface modification, as explained above. A slightly lower signal for the negative controls was also observed in these experiments. Figure 5 summarizes all of the results and the various negative controls performed in this case.

Fig. 5
figure 5

Results obtained from EIS measurements during Salmonella spp detection. 1, Av-GEB electrode modified with biotinylated IS200 hybrid; 2, Av-GEB electrode modified with biotinylated IS200 probe (corresponding to a noncomplementary target experiment); 3, GEC electrode modified with biotinylated IS200 hybrid; 4, Av-GEB electrode modified with nonbiotinylated IS200 hybrid. (Δ = Ret (sample) − Ret (blank)). The reproducibility shown corresponds to triplicate experiments. All hybridization experiments were performed in TSC1 buffer solution; all EIS measurements were performed in 0.1 M PBS buffer solution containing 0.01 M K3[Fe(CN)6]/K4[Fe(CN)6]

Full size image

The first bar in the histogram represents the hybridization experiment with IS200 complementary target (hybrid immobilization), whilst the second represents the experiment where a noncomplementary target is employed (this corresponds to the immobilization of the probe alone). In this case, the variation in the signal achieved for the hybrid was almost double the signal variation obtained for the probe. The third bar represents the results obtained when a GEC electrode without avidin is used as a platform to immobilize biotinylated DNA. As already explained, the signal variation obtained in this case can be considered to be sufficiently different from the one obtained in the experiment using Av-GEB electrodes (first bar in the histogram). The fourth bar corresponds to an experiment when an Av-GEB is used with nonbiotinylated DNA sequences. The signal obtained confirmed that the nonspecific adsorption of nonbiotinylated DNA is not significant here.

As we can see from the histogram, the reproducibility of the results is good. In fact, the RSD value is around 7% in the hybridization experiments.

Like before, a Student’s t-test was carried out to compare the results obtained from the hybridization and nonspecific hybridization assays (experiments 1 and 2 in the histogram). The difference between the mean values was highly significant, t calc (10.37) > ttab (2.13), at the 95% confidence level, for four degrees of freedom. The data used to perform the Student’s t-test are shown in Table 3.

Calibration curve

The reasons for constructing a calibration curve were twofold: to evaluate the proper target concentration for DNA biosensing with the Av-GEB platform, and to establish the detection limit for the genosensor.

Different Av-GEB sensors modified with the same quantity of DNA probe (30 pmol) and different amounts of DNA target were tested. A noncomplementary sequence was used as a negative control during the hybridization step.

As shown in Fig. 6, increasing the target concentration led to a higher analytical signal due to the resulting increase in the charge transfer resistance, thus achieving a linear range between 5–30 pmol. After this point, a plateau was reached and increasing the amount of target further did not result in any signal amplification. This was obviously due to the fact that all of the immobilized DNA probe became hybridized with the complementary target and so there were no free probe left for further hybridization. Hence a concentration of 30 pmol is recommended as a compromise between the sensitivity of the technique and the need to use a small amount of DNA. The limit of detection achieved for homooligonucleotide sequences was 5.1 pmol (corresponding to three times the standard deviation of the value obtained in negative control experiments).

Fig. 6
figure 6

Calibration curve obtained with an Av-GEB electrode modified with 30 pmol of DNA probe and different amounts of DNA target (filled circles) and negative control experiments (unfilled circles). All DNA sequences were diluted in 140 μL of buffer solution. (Δ = Ret (sample) − Ret (blank)). Error bars correspond to triplicate experiments. All hybridization experiments were performed in TSC1 buffer solution; all EIS measurements were performed in 0.1 M PBS buffer solution containing 0.01 M K3[Fe(CN)6]/K4[Fe(CN)6]

Full size image

The results obtained for negative control experiments, performed with the same amount of noncomplementary target, are also shown in the same figure. In this case, data processing was required to express the results as an increase in Ret.

A calibration curve was also constructed for Salmonella spp which showed the same type of trend, so the same concentration of 30 pmol was chosen for such experiments. The limit of detection achieved in this case was 6.5 pmol.

Confocal microscopy inspection

Characterization of the Av-GEB surface by confocal laser scanning microscopy was realized in order to investigate the presence of active avidin binding sites on the electrode surface. The images obtained are shown in Fig. 7. The first image (Fig. 7a) corresponds to the electrode modified with the biotinylated florescein. The strong fluorescence observed in this case is due to the presence of the fluorescent marker immobilized on the electrode surface through biotin–avidin interactions. The intensity of the fluorescence is directly correlated with the presence of active avidin binding sites on the electrode surface. The weak fluorescence observed in the second image (Fig. 7b) corresponds to the negative control, where the electrode surface was saturated with biotin molecules before being incubated with biotinylated florescein. This blocked all of the binding sites on the electrode surface, so the low-level fluorescence observed is due to the nonspecific adsorption of biotinylated fluorescein onto the electrode surface. Comparison of the two images indicates that the Av-GEB platform exposes active binding sites for biotin, which acts as an affinity matrix, and confirms again that the nonspecific adsorption of biotinylated molecules is not significant here.

Fig. 7a, b
figure 7

Images obtained with a confocal laser scanning microscope. a Av-GEB placed in a 80 pmol biotinylated fluorescein solution (with the same volume, 140 μl, used in previous experiments). b Negative control. Laser excitation: 568 nm. Voltage: 352 V. Images were taken by integrating the fluorescence in five different planes in which signal was detected (the distance between each section corresponds to 2 μm)

Full size image

Since graphite–epoxy is a porous material [9], the fluorescence obtained is also due to the diffusion of the small biotinylated marker into the pores of the material in order to reach avidin binding sites. Therefore, confocal microscopy shows the sum of the fluorescence from various different planes that constitute the transducer. Similar behavior is expected for small and flexible biotinylated oligonucleotide sequences.

Conclusions

The detection of DNA hybridization by EIS is reported in this work. Av-GEB electrodes were used for the first time for impedimetric transduction. The combination of this biocomposite transducer with electrochemical impedance allows the realization of rapid and simple analyses without the need for DNA labeling or protocols with several steps.

The utilization of the Av-GEB platform permits the simple and rapid immobilization of biotinylated DNA as well as the formation of a strong interaction between the electrode surface and oligonucleotide sequences.

As well as its good electrochemical performance, the Av-GEB transducer provides other advantages, such as ease of preparation, low cost, robustness, fast response and the possibility of repeated use. The limit of detection for the conditions used was found to be ~5.1 pmol DNA, low enough for many screening applications.

Nonspecific adsorption can be minimized by performing the incubations with slow stirring during DNA immobilization, since this encourages specific avidin–biotin binding instead of weak adsorption interactions [78]. Moreover, the one-step immobilization–hybridization procedure greatly reduces the duration of analysis, making the assay extremely simple.

Another important feature, the reproducibility of the measurements, is clearly improved through the use of this affinity immobilization procedure, especially when compared to previously reported results on the use of adsorption procedures to immobilize DNA. In the same way, good reproducibility was achieved for results obtained during Salmonella spp sequence detection (RSD ~7%).