Tris base was obtained from Promega (USA), and sodium acetate was obtained from Sigma (USA). UltraPure distilled water was purchased from Invitrogen (USA), and reagents were of analytical grade. All DNA oligonucleotides (probe, target and non-complementary sequences) were synthesized by Integrated DNA Technologies (Brazil). Probe: Dengue virus type specific 1 (TS-1P)_ 5′CGTCTCAGTGATCCGGGG 3′; Target: Dengue virus type specific 1 (TS-1T)_ 5′ CCCCGGATC ACTGAGACG 3′; Non-complementary: Dengue virus type specific 2 (TS-2 NC)_ 5′ CTGTTCATGG CCCTTGTGG CG 3′; Dengue virus type specific 3 (TS-3 NC)_ 5′ GCTCTGTCTCATGATGATGTT A 3′; PolyG-NC_ 5′ GGGGGGGGGGGGGGGGGGGG 3′. The oligonucleotide stock solutions and dilute solutions were prepared with 0.5 M acetate buffer (pH 4.8) and kept frozen.

The complete genomes of dengue virus types 1, 2 and 3, corresponding to GenBank_ accession numbers U88536.1, U87411.1 and AY099336.1, respectively, were obtained from the National Center for Biotechnology Information (NCBI) database. The probe specific for dengue type 1 (TS-1P) was constructed based on Duebel [20], and the sequence was aligned using the CLC Combined Workbench v 3.6.1 software. Target sequence (TS-1T) and non-complementary (TS-2 NC and TS-3 NC) sequences were also constructed using this software.

Electrochemical analysis was carried out with the Autolab PGSTAT apparatus (Metrohm Autolab, The Netherlands). Voltammetric signals were measured using a system consisting of two electrodes [21] Pencil graphite (type 4B) was used as the working electrode, and the reference electrode was screen-printed using Ag/AgCl ink (Electrodag-Acheson, USA) under gold wire and then dried at 60 °C. The experiments were performed in triplicate.

Pencil lead (type 4B, total length of 3 cm and diameter of 2.5 mm) typically composed of natural graphite was used as the pencil graphite electrode (PGE). The graphite was polished with an emery-impregnated disc to obtain a smooth surface. The body of the pencil was coated well with silicone, resulting in a free area for immobilization of 28.50 mm². Afterwards, the graphite electrode was washed with ultra-pure water to remove possible contaminants on the surface of the electrode. The polished surface of the working electrode was then activated by applying a potential of 1.8 V for 5 min [22]. The working electrode was fixed vertically and immersed in a solution of 0.5 M acetate buffer (pH 4.8) containing the TS-1P probe.

The TS-1P probe was immobilized on the activated electrode by applying 0.5 V [20] to the electrode for 1–10 min in 0.5 M acetate buffer solution (pH 4.8) containing oligonucleotides at different concentrations. The electrode was then rinsed with Tris-HCl (pH 7.0).

The working electrode with the immobilized probe was immersed in a solution containing the target oligonucleotides in acetate buffer, pH 4.8, and incubated at 57 °C for 3 min. The electrode was then washed with Tris-HCl (pH 7.0) to remove non-hybridized sequences. The same protocol was applied for the interaction of the probe with non-complementary sequences.

To obtain electrochemical signal data, the biosensor surface with hybridized probe and target nucleic acid was immersed in an electrolytic cell containing Tris-HCl (pH 7.0), and attached to the Autolab potentiostat, equipped with GPES 4.9 software. Differential pulse voltammetry was applied to detect electrochemical signals, whereby a potential sweep was applied between 0.5 and 1.2 V, at a pulse amplitude of 50 mV and scan rate of 20 mV/s. The raw data were treated using the GPES software with a moving average baseline correction using a “peak width” of 0.01.

Experimental data were analyzed using the Statistica 8 software by non-parametric tests. To evaluate statistical significance, the Mann–Whitney U test was used for the comparison of two unpaired groups. The Kruskal-Wallis test was used for the comparison of multi independent group data. A level of p < 0.05 was considered significant.

Research has demonstrated the utility of graphite electrodes [23] in the construction of electrochemical biosensor utilizing oligonucleotides [12,15]. The good sensitivity of the carbon electrodes in detecting the oxidation of nucleic acids makes them widely useful in DNA research, especially when inexpensive pencil graphite electrode is used [24]. The immobilization of nucleic acids on the surface of the electrode occurs by adsorption, which is considered the simplest method, and it does not require any special reagents or nucleic acid modifications. The smoothed surface of the pencil graphite was pretreated applying a potential (vs. Ag/AgCl reference electrode) of 1.8 V for 5 min. The pretreatment of the carbon surface increases its roughness and hydrophilicity, facilitating the adsorption of the probe on the electrodes [23].

The covalent immobilization of the probe is more stable, but the electrode surface may contain groups capable of reacting with and fragmenting the oligonucleotide molecule, damaging and changing its original structure [25]. The immobilization and hybridization are based on the electrochemical oxidation of guanine and adenine groups, which occurs with different electron transfer reaction rates [26,27]. We detect the oxidation of guanine that it is the most redox active nitrogenous base in DNA strands [12], which has been shown in some works as peaks of +0.93–1.0 V [26,28].

The influence of the pretreatment of the PGE at 1.8 V was investigated. Figure 1 shows no significant difference between the current peaks of activated PGE and TS-1P (1 μM) non-activated PGE. This observation indicates that the adsorption of nucleic acid did not occur. While the signal of the TS-1P (1 μM) on activated PGE were significantly higher than the signals of three other electrodes, due to the immobilization of the oligonucleotides containing guanines. These results demonstrated that applying a potential (1.8 V) on electrodes allows the adsorption of oligonucleotides.

The immobilization of the probe on the activated PGE involved the application of a potential of 0.5 V for 5 min [12,23,29,30]. This potential enhances the stability of the immobilized probe through the electrostatic attraction between the positively charged carbon surface and the negatively charged hydrophilic sugar-phosphate backbone with the bases oriented toward the solution ready to hybridize with the target [27].

The influence of the number of guanine in the probe was investigated immobilizing 1 μM of TS-1P (with seven guanines) and TS-1T (with five guanines) on different working electrodes. Figure 2 shows that the TS-1P probe (206 ± 4.3 nA) has a current peak two times larger than that of TS-1T (84.7 ± 5.0 nA). These results were similar those obtained by Pournaghi-Azar et al. [31] with a biosensor for interleukin-2. They also observed that current peak was higher for the sequence with seven guanines compared to that with one guanine.

The TS-1T sequence was chosen as a target sequence, because it is present in the dengue virus genome. TS-1P is the sequence complementary to the RNA present in the virus, and thus used as the probe for construction of the biosensor. The data obtained were analyzed using the CLC Combined Workbench v 3.6.1 and NCBI Software.

Probe immobilization is a crucial step for fabricating an electrochemical biosensor [18]. In Figure 3, the results obtained with anodic differential pulse (ADP) show the guanine oxidation signal increasing with time up to 5 min and decreasing significantly after that. This decrease can be attributed to the massive accumulation of the probe [29] on the graphite electrode, leading to an overlapping of the probes and lower availability of guanine bases. These results showed that probe immobilization could be achieved within 1 to 10 min. However, five minutes was suggested as the optimal time for probe immobilization since the electrode achieved a higher peak current. This result also was found in the systems for gene detection of interleukin-2 [29], hepatitis B [15] and electrochemical detection for human papilloma virus [23].

The effect of probe concentration is shown in Figure 4, it was carried out by differential pulse voltammogram after 5 min of oligonucleotides (TS-1P) immobilization using a potential of 0.5 V in acetate buffer (pH 4.8) containing different concentrations of probe (0.1 to 10 μM).

The results showed that oxidation peak currents increased from 0.1 μM up to 1.0 μM with TS-1P. However, since the 1 μM the signal is stabilized, it was the chosen concentration to carry on further experiments. The results, analyzed with STATISTICA 8.0 using non-parametric tests (Kruskal-Wallis), showed that 1 μM–10 μM results were statistically identical. Thus, the concentration of the 1 μM and 5 min were considered the best experimental conditions for electrode preparation. Next, the study sought to build a working electrode, with an area completely filled with the oligonucleotide sequence (probe).

The electrodes modified with nucleic acid (polynucleotide) identify the sequence of complementary bases, through the formation of a double helix. This identification is effective and specific showed in the presence of other non-complementary sequences. When the target oligonucleotide sequence corresponds to the probe (based on the complementarily principle stating that G pairs with C and A with T or U), a hybrid is formed (probe-target).

The development of biosensor DNA or RNA consists of three steps: (a) adsorption of the oligonucleotide probe; (b) hybridization between the probe and its complementary sequence (the target) to form the hybrid; and (c) transduction [27]. The hybridization devices for biological detection of specific oligonucleotides sequences can be fabricated by applying a constant potential [32], incubating at a specific annealing temperature and at room temperature [12,33].

There have been various reports on label-free oligonucleotide biosensors for hybridization detection [12,22]. The oligonucleotide-modified electrode is dipped into a solution of target (DNA or RNA) to test its nucleotide sequence. When hybridization occurs, there is a decrease in electrochemical signals due to the interaction of free guanines of the probe with complementary cytosine bases present in the target sequence [22,27], the principle was used in this work. This is because of less guanine bases are availabile in the hybrid form for oxidation. Another strategy for the detection of guanine oxidation is the use of immobilized inosine-substituted (for guanine) or guanine-free probes [22], where there is a direct detection of oligonucleotide hybridization by the appearance of the oxidation signal due to the presence of guanines in the target sequence [27].

The hybridization experiments were performed in a microtube (0.2 mL), containing the probe-modified PGE, which were immersed in acetate buffer containing a known amount of complementary target oligonucleotides at a specific annealing temperature. Detection of target oligonucleotides was monitored with the formation of a guanine current peak on the probe immobilized on PGE with or without the target sequence.

The results are shown in the voltammogram for guanine oxidation (0.97 V in 20 mM Tris-HCl buffer solution, pH 7) on the probe-modified and activated PGE, before and after hybridization in solutions of complementary and non-complementary oligonucleotides (Figure 5). The current peak generated by guanine oxidation was monitored by the detection of the target oligonucleotides. The probe contains seven guanine bases, and after hybridization with the target sequence it contains only five guanines, thus there is a significant reduction in the guanine oxidation peak.

The current peak of the hybridization at the concentrations 0.5, 1, 10, 20, 30, 40 and 100 nM was statistically different from that observed with the TS-1P sequence at the concentration of 1.0 μM. The results showed that the complementary sequence could produce hybridization, causing a decrease in the guanine oxidation signal.

The highest guanine oxidation signal was observed with TS-1P immobilized on PGE, which has seven guanine bases present in the oligonucleotides. Experiments with non-complementary oligonucleotides including TS-2 NC, TS-3 NC and PolyG-NC (Figure 5) were performed to evaluate the selectivity of electrochemical nucleic acid biosensor. The interaction between these non-complementary oligonucleotides and probe did not lead to a significant decrease in the guanine oxidation due to absence of entire hybridization. So, the current peak of hybridized with a non-complementary sequence showed values similar to that observed following TS-1P. The selectivity demonstrates that the sequence (TS-1T) could form a double helix with the complementary oligonucleotides, causing a significant decrease in the guanine oxidation current peak.

These results confirmed that the electrochemical detection in the present study successfully distinguished complementary sequences from non-complementary sequences, using the annealing temperature to perform the hybridization. The selectivity of the biosensor was also examined in a sample containing both complementary and non-complementary sequences in similar proportions. The mixture showed a peak current similar to immobilized TS-1P after hybridization with complementary sequence. This demonstrates that the presence of non-complementary samples did not interfere in the specificity of the biosensor. These results are important for the construction of the device, because the system can be better controlled by regulating temperature than by other means such as the application of a potential.

The detection limit was determined using the synthetic target sequence. The difference in the guanine oxidation signal of the probe-modified PGE in the absence and presence of complementary sequence (ΔI) increased when the target concentration increased and leveled off at a concentration of 40 nM. As seen in the inset of Figure 6, the signal was linear between 1 and 40 nM with a correlation coefficient of 0.9856 for complementary target.

The detection limit and quantitation limit values were calculated using the following equations: y_LOD = 3s/m and y_LOQ = 10 s/m, respectively [34], where s is signal of the standard deviation of the blank (0.5 nM), and m is the slope (1.6154) of the related calibration line. The regression equation was ΔI = 1.6154C + 92.248 (C, nM; ΔI, A), thus the detection limit (0.92 nM) and quantitation limit (3.09 nM) were measured, and both values confirmed the sensitivity of the device. The within day and between day reproducibility (R.S.D.%) results of the signal (three independently probes) measured at 40 nM of target were 0.61% and 1.72%, respectively, indicating a remarkable reproducibility of the detection method.

Compared with other nucleic acid biosensors obtained under the same immobilization conditions (adsorption), ours displayed good sensitivity with a hybridization time of 3 minutes. In Table 1, the biosensor proposed in the present work has a low detection limit compared with the others. High-sensitivity nucleic acid detection is essential for clinical diagnosis, pathology, and genetics [35]. The use of an electroactive indicator can provide a slightly lower detection limit when compared with guanine oxidation. However, label-free monitoring based on guanine moiety oxidation signal of target in biosensors are easily manipulated, have fast responses and are inexpensive. Some studies using electroactive molecules showed higher detection limits than our results, such as for the detection of human hepatitis B (7.19 nM) [36] and detection of HPV (3.8 nM) [37].