Introduction

Highly sensitive method for the detection of DNA has attracted much attention due to its diverse applications including the identification of genetic diseases and disorders and the detection and characterization of genetic viruses, bacteria, and parasites [1]. Traditional methods for the detection of DNA on the basis of base-pair hybridization, such as gel electrophoresis or membrane blots, are slow and labor-intensive. DNA hybridization biosensors offer considerable promise for obtaining sequence-specific information in a simpler, faster and cheaper manner, compared to traditional hybridization assays [2]. Various DNA biosensors have been established including fluorescence imaging [3, 4], optical [5–7], electrochemical [8, 9], electrogenerated chemiluminescence (ECL) [10, 11], quartz-crystal microbalance [12], and surface plasmon resonance techniques [13]. In these DNA biosensors, ECL biosensors are very promising due to the combination of advantages of both electrochemical and chemiluminescent biosensors, such as high sensitivity and ease of control.

Much effort in the research of ECL DNA-based biosensors (DNA hybridization biosensors, aptasensor) has been made to improve the sensitivity and specificity. These include the employment of gold nanoparticles [14] and carbon nanotube [15] and dendrimers [16] as signal carriers for signal amplification, use of a hairpin DNA as a probe [17] for signal and recognition enhancement. In these ECL biosensors, the ss-DNA probe is generally labeled with an ECL reagent. The labeling process of the DNA probe, however, is time-consuming and reduces the hybridization efficiency. Thus, increasingly attractive attention is to develop the label-free ECL DNA biosensors, mainly using ECL reagents such as Ru(phen)3 2+ and Ru(bpy)3 2+ as intercalator. Yin et al. [18] developed an ECL aptamer-based biosensor for the detection of thrombin based on a Ru(phen)3 2+-double-strand DNA composite film. Plaxco et al. [19] firstly reported a novel supersandwich approach for electrochemical detection of target DNA in complex matrices. They employed a modified signal probe containing a methylene blue (a redox moiety) label and a “sticky end.” When the ss-DNA target hybridizes this signal probe, the sticky end remains free to hybridize another target leading to the creation of a supersandwich structure containing multiple labels. This leads to large signal amplification upon monitoring by voltammetry. Xu et al. [20] reported an ECL method for the detection of Hg2+ by employing a label-free supersandwich model based on T–Hg2+–T coordination and the intercalation of Ru(phen)3 2+. Zhuo et al. [21] reported an ECL aptasensor for the detection of thrombin based on DNA supersandwich structure, in which hollow gold nanospheres were employed as tag-carriers for the aptamer, and Ru(phen)3 2+ as an ECL signal reagent was intercalated into ds-DNA. Recently, Xiang et al. [22] reported an ECL method for the detection of DNA based on hybridization chain reaction for signal amplification and Ru(phen)3 2+ as ECL intercalator. To the best of our knowledge, ECL DNA hybridization biosensors on the basis of supersandwich model using linear ss-DNA have not been reported.

The aim of the present work is to develop a highly sensitive and label-free ECL method for the detection of HIV-1 gene based on a supersandwich DNA structure and Ru(phen)3 2+ as intercalator. A HIV-1 gene (38-mer ss-DNA from HIV-1 U5 long terminal repeat gene sequence) [23] was chosen as target analyte (See Table 1). Schematic diagram of the label-free ECL hybridization biosensing based on supersandwich DNA model is shown in Fig. 1. The thiolated ss-DNA as a capture probe is firstly self-assembled on the surface of a gold electrode. After the target HIV-1 gene is completely hybridized with the capture probe, previously hybridized two auxiliary probes are hybridized with the target HIV-1 gene to form long-range supersandwich ds-DNA polymers on the surface of the electrode. Finally, Ru(phen)3 2+ is intercalated into the supersandwich ds-DNA grooves, resulted in greatly increasing ECL intensity in tripropylamine solution since a great amount of Ru(phen)3 2+ are intercalated into supersandwich ds-DNA. In this paper, a label-free ECL supersandwich DNA hybridization biosensing for the detection of HIV-1 gene is fabricated and characterized, and the optimization of experimental conditions and analytical performance of this method are presented.

Table 1 Sequence of oligonucleotides used in this work
Full size table
Fig. 1
figure 1

Schematic diagram of the label-free ECL DNA hybridization biosensing based on supersandwich ds-DNA structure

Full size image

Experimental

Reagents

Tris(1,10-phenanthroline) ruthenium(II) dichloride (Ru(phen)3Cl2 .H2O) and 6-mercapto-l-hexanol (MCH) were purchased from Sigma-Aldrich (St. Louis, MO, www.sigmaaldrich.com). Tripropylamine (TPA), potassium hexacyanoferrate (III) and other chemicals employed were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China, www.sinoreagent.com). All oligonucleotides used as listed in Table 1 were purchased from Sangon Biotechnology Co., Ltd (Shanghai, China, www.sangon.com). 0.10 M phosphate buffer saline (PBS, pH 7.4, 0.10 M NaCl + 0.10 M NaH2PO4/Na2HPO4) was used as hybridization buffer and washing solution. All ss-DNA solutions were prepared using 0.10 M PBS (pH 7.4). All reagents were of analytical grade. Millipore Milli-Q water (18 MΩ · cm) was used throughout.

Apparatus

ECL measurements were performed with a MPI-E ECL detector (Xi’an Remax Electronics, Xi’an, China, www.chinaremex.com). A commercial cylindroid glass cell was used as an ECL cell (diameter, 25 mm). A three-electrode system composed of a biosensor or a disk gold electrode (2.0 mm diameter with an effective electrode area of 3.42 mm2) [24] as working electrode, a platinum plate as counter electrode, and an Ag/AgCl (saturated KCl) as reference electrode, respectively. For detecting ECL, the cell was placed directly in front of a photomultiplier (PMT, operated at −900 V) and the PMT window was only opened toward the working electrode to eliminate the blank CL and the ECL from the counter electrode.

Immobilization of capture probe

A disk gold electrode was mechanically polished with 0.3 and 0.05 μm alumina slurry, ultrasonicated for 3 min in ultrapure water, ethanol, and ultrapure water respectively to remove any remaining polishing agent, and then electrochemically cleaned by a linearly scanning potential between −0.2 and +1.5 V in 0.10 M H2SO4 until a stable cyclic voltammogram was obtained[25], and finally washed with water, dried with nitrogen.

Ten microliters of 1 μM capture probe was drop-coated on the surface of the cleaned gold electrode, allowed to stay for 4 h to thiol-self-assembly the capture probe on the surface of the electrode. The resulted electrode was thoroughly washed with 0.10 M PBS to remove the unbound capture probe, and then immersed in 100 μL of 2.0 mM MCH prepared using 0.10 M PBS for 1 h to block the uncovered surface of the gold electrode, and washed and dried under a stream of nitrogen gas. The ECL biosensor obtained was stored in 0.10 M PBS.

ECL measurements

First, 10 μL of a fixed concentration of target HIV-1 gene was dropped onto the surface of the ECL biosensor fabricated above, and incubated for 90 min in a sealed humidity chamber at 37 ºC. Second, 10 μL of a freshly prepared auxiliary probe solution containing 1 μM auxiliary probe 1 and 1 μM auxiliary probe 2 (Table 1) was dropped onto the surface of the electrode and then incubated for 90 min in a sealed humidity chamber at 37 ºC to form long-range supersandwich ds-DNA polymers on the surface of the electrode. Finally, the resulted electrode was immersed in 100 μL of 0.1 mM Ru(phen)3 2+ for 90 min at room temperature. After each binding step, the resulted electrode was washed with 0.10 M PBS.

ECL measurements were performed at the resulted electrode prepared-above in 1 mL of 0.10 M TPA (pH 7.4) in the potential range from 0 to +1.3 V (versus Ag/AgCl) with a scan rate of 50 mV/s. The concentration of target HIV-1 gene was quantified by the increased ECL intensity (∆I = I − I 0), where I 0 is the ECL peak height in the absence of the target HIV-1 gene and I is the ECL peak height in the presence of the target HIV-1 gene.

Results and discussion

Characteristics of the electrode assembling process

Electrochemical impedance spectroscopy (EIS) technique has proven to be one of the most powerful tools for the characteristics of interfacial properties in the fabrication process of the biosensors and the hybridization process [26, 27]. Figure 2 shows the Nyquist plots of the gold electrode subjected to the step-by-step modification process, using [Fe(CN)6]4-/3- as the redox indicator. The inset shows the equivalent circuit applied to fit the impedance spectra, where Rs is the solution resistance, R et is the interfacial electron transfer resistance, ZW is a Warburg impedance, C1 and C2 are the double layer capacitances, CPE1 and CPE2 are the constant phase elements, and R2 is the resistance that develops between the MCH layer and the electrode surface. It can be seen that the simulating curves (solid line) are quite well fitted with the experimental data (dot). After the capture probe was self-assembled on the surface of gold electrode, the interfacial electron transfer resistance (R et) value increases from 260 Ω to 5537 Ω (Fig. 2 curve a and curve b). This is attributed to the fact that the self-assembled capture probe having a single strand nucleic acid with negative charges on its phosphate backbone makes an electrostatic repulsive force to [Fe(CN)6]4-/3- [28, 29]. This indicates that the capture probe is successfully assembled onto the surface of gold electrode. After blocking with 2 mM MCH, the R et value increases from 5,537 Ω to 12,700 Ω (Fig. 2 curve c), indicating that the capture probe-uncovered surface of the gold electrode is blocked. After the resulted electrode is hybridized with the target HIV-1 gene, the R et value slightly increases from 12,700 Ω to 13,490 Ω (Fig. 2 curve d), indicating that the HIV-1 gene is hybridized with the capture probe on the surface of the gold electrode. A relatively large increase of the R et (16,310 Ω, Fig. 2 curve e) is observed after two hybridization auxiliary probes prepared are introduced into the HIV-1 gene hybridized electrode. This indicates that long-range supersandwich ds-DNA polymers are formed on the surface of the electrode. This coincides with the characterization obtained by atomic force microscopy [23]. The EIS results indicate that the desired immobilization and hybridization are successfully accomplished on the electrode surface.

Fig. 2
figure 2

Nyquist plots of electrochemical impedance spectra obtained in 0.10 M PBS (pH 7.4)−5.0 mM K4[Fe(CN)6]−5.0 mM K3[Fe(CN)6]−0.10 M NaCl. Experimental data (dot). Fitted data (solid line). Electrodes used from a to e were: (a) bare gold electrode, (b) capture probe modified gold electrode, (c) capture probe modified gold electrode after blocking with 2 mM MCH, (d) capture probe/MCH modified gold electrode after interaction with 1.0 × 10−9 M target ss-DNA, (e) capture probe/MCH modified gold electrode after interaction with 1 μM auxiliary probe 1 and auxiliary probe 2. Experimental conditions: The biased potential of 0.225 V, the frequency from 100 kHz to 0.1 Hz, and the amplitude of 5.0 mV

Full size image

Feasibility of supersandwich ECL approach

Figure 3 shows ECL intensity-potential profiles at different electrodes. From Fig. 3 (line a), it can be seen that a small ECL peak (1,000 a. u.) appears. This is attributed to the fact that the small amount of Ru(phen)3 2+ is electrostatically bound to the negative charged capture probe immobilized on the gold electrode [30]. Compared Fig. 3 line b with line a, the ECL peak height increases from 1,000 a. u. to 2,500 a. u. after target HIV-1 gene is hybridized with the capture probe. This is attributed to the fact that the ECL reagent Ru(phen)3 2+ is intercalated to the double strands formed by the capture probe and the target HIV-1 gene. It was found that the ECL peak height greatly increases from 2,500 a. u. to 6,100 a. u. after two auxiliary probes were further hybridized with the target HIV-1 gene on the surface of the electrode (Fig. 3, line c). This is attributed to the fact that the supersandwich ds-DNA is formed and a large number of Ru(phen)3 2+ is intercalated to the long-range supersandwich ds-DNA. This indicates that a remarkable amplification is achieved. A supersandwich ECL DNA hybridization biosensing method is, therefore, feasible.

Fig. 3
figure 3

ECL intensity-potential profiles at different electrodes. (a) capture probe-assembled gold electrode; (b) capture probe/MCH modified gold electrode after interaction with 5.0 × 10−11 M HIV-1 gene; (c) capture probe/MCH modified gold electrode after interaction with 1 μM auxiliary probe 1 and auxiliary probe 2 after interaction with 5.0 × 10−11 M HIV-1 gene. The intercalation time of Ru(phen)3 2+ was 90 min. The ECL measurements were performed with a scan rate of 50 mV/s in 0.10 M PBS (pH 7.4) containing 0.10 M TPA

Full size image

In order to understand the hybridizations among the capture probe and the target HIV-1 gene as well as auxiliary probes, the changes of Gibbs free energy (∆G) were predicted with the computer program RNA Structure version 4.5 at 25 °C [31]. The absolute values of the predicted ∆G from the capture probe/HIV-1 gene (-24.9 kcal/mol), HIV-1 gene/auxiliary probe 1 (−23.1 kcal/mol), and auxiliary probe 1/auxiliary probe 2 (−23.0 kcal/mol) are much higher than those of capture probe/auxiliary probe 1(−4.1 kcal/mol) and HIV-1 gene/auxiliary probe 2 (−4.8 kcal/mol). From the changes of Gibbs free energy, it is anticipated that the capture probe can be hybridized with HIV-1 gene and HIV-1 gene with auxiliary probe 1, and auxiliary probe 1 with auxiliary probe 2. However, the capture probe could be not hybridized with auxiliary probe 1, and HIV-1 gene could be not hybridized with auxiliary probe 2. The predicted results with the computer program RNA Structure can support our design and the preliminary ECL results described above.

Optimization of experimental conditions

Important experimental parameters including intercalation time of Ru(phen)3 2+, the concentration of auxiliary probes and the hybridization time were optimized according to the experiment protocol described in ECL Measurements.

As expected, both the concentration of Ru(phen)3 2+ and intercalation time of Ru(phen)3 2+ influence the sensitivity due to their affecting the intercalation quantity of Ru(phen)3 2+ into the grooves of the supersandwich ds-DNA. Fig. S1 (Electronic Supplementary Material, ESM) shows dependence of the intercalation time on the increased ECL intensity (∆I) for the detection of 5.0 × 10−11 M target HIV-1 gene. From Fig. S1 can be seen that the ∆I lineally increases about 333 % with increasing intercalation time from 15 min to 60 min, and then slowly increases about 105 % with increasing intercalation time from 60 min to 90 min. Therefore, 90 min was chosen as the intercalation time of 0.1 mM Ru(phen)3 2+ in following experiments. This is quite shorter than that reported in ECL aptasensors using Ru(phen)3 2+ intercalated into ds-DNA as ECL indicator [18, 21].

Fig. S2 (ESM) shows effect of the concentrations of two auxiliary probes on the ECL intensity for the detection of 1.0 × 10−11 M target HIV-1 gene. It should be noted that the same concentrations of the two auxiliary probes were chosen since the hybridization ratio of the two auxiliary probes is one to one and the concentration of the target HIV-1 gene is much lower than that of the auxiliary probe. From Fig. S2, it can be seen that the ECL intensity sharply increases from 2,318 to 5,433 about 234 % as the concentrations of auxiliary probes increases from 0 to 1.0 μM. This is attributed to the fact that amount of the long-range hybridized ds-DNA polymers increases and thus amount of Ru(phen)3 2+ intercalated into ds-DNA increases. This indicates that the long-range supersandwich ds-DNA is formed. And the ECL intensity slightly increases from 5,433 to 5,739 about 106 % as the increase of the concentrations of auxiliary probes from 1.0 to 2.0 μM. This is attributed to long range electron transfer and the steric hindrance of the long-range hybridized ds-DNA polymers [32]. On consideration of the sensitivity and saving the reagents, therefore, the concentrations of auxiliary probes were chosen as 1.0 μM in the subsequent experiments. These results indicate that in the case of tested experiment, the ECL intensity is 2.34 fold of that in the absence of the auxiliary probes. The amplification of supersandwich model is, therefore, evident. The signal amplification of supersandwich model coincides with those reported previously in the literatures [22, 23, 33].

Fig. S3 (ESM) shows effect of super-hybridization time of auxiliary probes on the increased ECL intensity for the detection of 1.0 × 10−11 M target HIV-1 gene. It can be seen that the increased ECL intensity nearly lineally increases from 540 to 2,908 with increasing the hybridization time from 10 min to 60 min, and slowly increases from 2,908 to 3,306 with increasing the hybridization time from 60 min to 150 min. This change tendency coincides with the report by electrochemical method using [Ru(NH3)6]3+ as intercalator [23]. This suggests that the positive charge of the intercalator may be dominated rather than the size of intercalator. In subsequent experiments, 90 min was employed as the super-hybridization time to obtain high sensitivity.

Analytical performance

Figure 4 shows the ECL profiles of the ECL biosensor fabricated after interaction with different concentrations of target HIV-1gene under the optimized conditions. From Fig. 4, it can be seen that the ECL intensity increases with an increase of the concentration of HIV-1 gene. The inset in Fig. 4 shows the quantitative relationship between the increased ECL intensity and the logarithm of the concentration of HIV-1 gene. The increased ECL intensity was directly related to the logarithm of the concentrations of target HIV-1 gene in the range from 1.0 × 10−13 to 1.0 × 10−10 M. The linear regression equation was ΔI = 14126.5 + 1038.2lgC (unit of C is M) with a correlation coefficient of 0.9921. The detection limit was calculated to be 2.2 × 10−14 M HIV-1 gene (S/N = 3). Comparison of this method with some reported genosensors with the detection limit and the signal amplification strategy is listed in Table 2. The detection limit obtained is considerably lower than that of the genosensors reported in label free model [12, 33, 34] and labeled model [3, 14, 17, 19, 35]. It is higher than that of the genosensors reported with nanomaterials amplification [6, 15, 36] or chain reaction amplification [22] or rolling circle amplification DNAzyme amplification [37]. Although the detection limit obtained by differential pulse voltammetry with a label free supersandwich model is quite lower than that obtained in our work, it needs degassing tested solution with nitrogen for 15 min [23]. This is time-consuming and there may present unstable problems from Ru(NH3)2+. Therefore, an excellent detection limit obtained is considerably satisfactory.

Fig. 4
figure 4

The ECL profiles of the ECL biosensor after interaction with different concentrations of HIV-1 gene under the optimized conditions. Inset: the linear relationship between increased ECL intensity (ΔI) and the logarithm of HIV-1 gene sequence concentration. The concentrations of HIV-1 gene: (a) 0 M; (b) 1.0 × 10−13 M; (c) 3.0 × 10−13 M; (d) 5.0 × 10−13 M; (e) 1.0 × 10−12 M; (f) 5.0 × 10−12 M; (g) 1.0 × 10−11 M; (h) 5.0 × 10−11 M; (i) 1.0 × 10−10 M. ECL measurement was as the same as Fig. 3. The illustrated error bars represent the standard deviation of three repetitive measurements

Full size image
Table 2 Comparison of the proposed method with some reported genosensors
Full size table

To assay the specificity of this method, the ECL response of the DNA biosensor fabricated was examined under the optimized conditions for 1.0 × 10−11 M target HIV-1 gene, 1.0 × 10−9 M two-base mismatched target HIV-1 gene, and 1.0 × 10−9 M completely mismatched HIV-1 gene, respectively. As shown in Fig. 5, the increased ECL intensity from completely mismatched HIV-1 gene is only 5.6 % of that from the perfect-matched HIV-1 gene while the increased ECL intensity from two-base mismatched HIV-1 gene is only 44.9 % of that from the perfect-matched HIV-1 gene although the concentrations of completely mismatched HIV-1 gene and two-base mismatched HIV-1 gene are higher than two orders of the perfect-matched target HIV-1 gene. These results indicate that the label-free supersandwich ECL method has a highly selectivity to effectively discriminate the perfect-matched HIV-1 gene from the two-base mismatched HIV-1 gene and completely mismatched HIV-1 gene.

Fig. 5
figure 5

The ECL response for different gene sequences (a) blank (without target HIV-1 gene on the electrode); (b) completely mismatched HIV-1 gene (1.0 × 10−9 M); (c) two-base mismatched HIV-1 gene (1.0 × 10−9 M); (d) perfect-matched HIV-1gene (1.0 × 10−11 M). ECL measurement was as the same as Fig. 3

Full size image

Conclusions

A label-free ECL hybridization biosensing for the detection of target HIV-1 gene with high sensitivity and selectivity has been designed on the basis of supersandwich ds-DNA structure and ruthenium complex as an intercalated signal-producing compound. The detection limit of 2.2 × 10−14 M was achieved. The perfect-matched HIV-1 gene can be effectively discriminated from the two-base mismatched HIV-1 gene. Two ss-DNA auxiliary probes used can form a long-range self-assembled ds-DNA on the basis of themself hybridization and target HIV-1 gene hybridization, and achieve 2.34 fold signal amplification. This work demonstrates that the high sensitivity and selectivity of an ECL DNA biosensing method could be greatly improved using supersandwich ds-DNA and ECL indicators. The strategy presented could be easily extended to various analytical platforms including ECL and electrochemical biosensors for the detection of many kinds of analytes or their interactions such as DNA/RNA, DNAzyme/target, and aptamer/target.