Rational engineering the DNA tetrahedrons of dual wavelength ratiometric electrochemiluminescence biosensor for high efficient detection of SARS-CoV-2 RdRp gene by using entropy-driven and bipedal DNA walker amplification strategy

Abstract

Fast and effective detection of epidemics is the key to preventing the spread of diseases. In this work, we constructed a dual-wavelength ratiometric electrochemiluminescence (ECL) biosensor based on entropy-driven and bipedal DNA walker cycle amplification strategies for detection of the RNA-dependent RNA polymerase (RdRp) gene of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The entropy-driven cyclic amplification reaction was started by the SARS-CoV-2 RdRp gene to generate a bandage. The bandage could combine with two other single-stranded S1 and S2 to form a bipedal DNA walker to create the following cycle reaction. After the bipedal DNA walker completed the walking process, the hairpin structures at the top of the DNA tetrahedrons (TDNAs) were removed. Subsequently, the PEI-Ru@Ti₃C₂@AuNPs-S7 probes were used to combine with the excised hairpin part of TDNAs on the surface of Au-g-C₃N₄, and the signal change was realized employing electrochemiluminescence resonance energy transfer (ECL-RET). By combining entropy-driven and DNA walker cycle amplification strategy, the ratiometric ECL biosensor exhibited a limit of detection (LOD) as low as 7.8 aM for the SARS-CoV-2 RdRp gene. As a result, detecting the SARS-CoV-2 RdRp gene in human serum still possessed high recovery so that the dual-wavelength ratiometer biosensor could be used in early clinical diagnosis.

1. Introduction

The outbreak of the new coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has endangered the health of people all over the world [1], [2]. Before an effective vaccine is developed, in order to prevent the spread of the new coronavirus disease, the first priority is to quickly and accurately diagnose the COVID-19, especially for asymptomatic patients [3]. At present, there are two main strategies for the detection of new coronavirus diseases in the market [4]. One strategy is to detect SARS-CoV-2 RNA by using polymerase chain reaction (PCR)-based technology [5], [6]. However, the operation process of PCR is complicated and different primers and temperatures need to be designed, and low accuracy usually results in false results. Another strategy is antibody-based serological and immunological testing [1]. However, after a patient being infected by SARS-CoV-2, the virus can not be detected immediately until antibodies are produced in the body. Therefore, a fast, accurate, simple, and reliable detection method is needed to detect SARS-CoV-2 to manage this epidemic. The target genomes of SARS-CoV-2 include the open reading frame 1 ab (ORF1ab) gene, nucleocapsid protein gene (N) sequence, envelope protein (E) gene, spike (S) gene, and RNA-dependent RNA polymerase (RdRp) gene [7]. These genomes play an essential role in maintaining coronavirus identity, participating in cross-reactivity with other coronaviruses, and distinguishing SARS-CoV-2 from other coronaviruses [8]. Among them, RdRp gene is a crucial catalytic component in the genomes, controlling viral RNA replication and transcription, and performs better in distinguishing SARS-CoV-2 from SARS-CoV [5]. Therefore, RdRp can be considered as an ideal target for detecting SARS-CoV-2 infection.

Compared with other detection techniques, electrochemiluminescence (ECL) is widely used for the analysis of various analytes due to its high sensitivity and simplicity of operation [9], [10]. The traditional ECL strategy relies on the intensity of a single signal to detect the analyte, which is easily affected by other interfering factors, such as the environment and the substrate [8], [11], [12]. However, ratiometric ECL using the signal ratio of the two emitters to detect the analyte can effectively avoid the interference of the environment and the substrate with improved assay accuracy [13], [14]. Compared with the dual-voltage ratiometric, the dual-wavelength ratiometric ECL biosensor has two advantages: multiple co-reactants and narrow voltage ranges [15], [16]. The former ECL biosensor mainly relies on the electrochemiluminescence resonance energy transfer (ECL-RET) between donors and acceptors with an improved dual signal ratio by overlapping donors’ ECL emission spectrum and the acceptors’ ultraviolet–visible (UV–Vis) absorption spectrum [17]. Au nanoparticles (AuNPs) have good biocompatibility and are widely used in the composite of functional nanomaterials [18]. Graphite-C₃N₄ (Au-g-C₃N₄) decorated with AuNPs has attracted attention as a donor material because of its good biocompatibility and excellent luminescence efficiency, but there is still a great lack of suitable acceptor materials [19]. Coincidentally, Ru(II) has a strong UV–vis absorption peak at 460–470 nm overlapping with the ECL emission peak of Au-g-C₃N₄. Feng et al. reported a dual-wavelength ratio ECL biosensor to detect microRNA, utilizing Au-g-C₃N₄ and Ru(bpy)₃ ²⁺ as acceptor and donor, respectively [16]. Therefore, it is feasible to introduce Au-g-C₃N₄ as the donor and Ru(II) materials as the acceptor in the dual-wavelength ratiometric ECL biosensors.

In order to improve the sensitivity of ECL biosensors, various cyclic amplification strategies have been introduced, including hybridization chain reaction (HCR), rolling circle amplification (RCA), catalytic hairpin assembly (CHA), polymerase chain reaction (PCR) and so on [5], [20], [21], [22], [23], [24], [25], [26], [27], [28]. These cyclic amplification strategies have complicated design processes and complicated secondary structures or require enzymes and variable temperature to participate in the reaction [29], [30]. As a simple, isothermal, enzyme-free, and fast detection strategy, the entropy-driven reaction has been proposed in recent years [31], [32]. An entropy-driven reaction is a kind of chain displacement reaction by acting on a toehold. Multiple hybridization and separation occur during the cycle amplification process followed by the formation of plenty of target DNAs [33], [34], [35]. In addition, the spontaneously, controllably, and continuously moveable enzyme-driven DNA walker on the track was introduced into biosensors as an amplification strategy [36]. Yuan’s research group used a single-step DNA walker to detect miRNA-141 [37]. Lei and his colleagues designed a single-step, two-site DNA walker to detect the target DNA [38]. However, walking efficiency cannot be improved due to the limited walking space of the single-step DNA walker. By contrast, the bipedal DNA walker has ample walking space and high efficiency [39]. Therefore, the simultaneous introduction of entropy-driven and bipedal DNA walker amplification strategies can significantly improve the sensitivity of the biosensor.

The electrode surface constructed by DNA tetrahedrons (TDNAs) can reduce the background signal and improve the hybridization efficiency of the target DNA [2]. One of the four single strands of each TDNA, the longest single strand can be designed to bind to the target DNA [40], possessing the advantages of easy synthesis, muscular mechanical rigidity, and uniform distribution of captured probes electrode surface [40]. Li et al. designed a multi-functional TDNA-based microarray platform that can be used to detect let-7a miRNA, prostate-specific antigen, and cocaine [41]. Lin et al. constructed a TDNAs sensor to detect Golgi protein 73 (GP73) [42]. The DNA tetrahedral sensor has a higher ECL signal compared to the single-stranded probe that detects GP73. The background signal of the TDNAs sensor is lower when there is no target present in the system, which is because the biosensor constructed from TDNAs can effectively prevent non-specific adsorption and significantly improve the detection sensitivity of the system.

In this work, we constructed a TDNAs-based dual-wavelength ratiometric biosensor combined with entropy-driven and bipedal DNA walker amplification strategies to detect the SARS-CoV-2 RdRp gene. First of all, the Au-g-C₃N₄ modified on the electrode surface possessed strong ECL emission at 460 nm. And a large number of TDNAs with a hairpin structure as the bipedal DNA walker’s trackies were modified on the surface of Au-g-C₃N₄. Secondly, by using the SARS-CoV-2 RdRp gene as a catalyst to trigger an entropy-driven reaction, the bandage was output, which could combine with two single-stranded S1 and S2 to form a bipedal DNA walker. It is worth mentioning that when the bandage is not introduced, even if 3′ terminus of the single-stranded S1 or S2 and the hairpin DNA are base-paired, their melting temperature (T_m) is too low to bind with each other. However, single-stranded S1 and S2 are induced by the bandage to form a bipedal DNA walker (S1 and S2 are used as the two feet of the bipedal DNA walker), resulting in a closer distance between single-stranded S1 and S2. This combination increased the T_m of the two feet of the bipedal DNA walker and the hairpin DNA [43]. Next, the bipedal DNA walker spontaneously moved on the hairpin structure at the top of the TDNAs with the help of Nb.BbvCI endonuclease, causing a large number of hairpin structures to be removed. Finally, a PEI-Ru@Ti₃C₂@AuNPs-S7 probe was synthesized to combine with the top of the TDNAs giving a strong ECL emission at 620 nm through the ECL-RET. The sensitivity of this ECL biosensor can be improved by the above-mentioned multi-step reaction. Experiments in human serum also demonstrated a high recovery rate, indicating that this biosensor could provide a reference value for early clinical diagnosis.

2. Experimental section

2.1. Chemicals and materials

Gold chloride trihydrate (HAuCl₄·3H₂O), sodium citrate, potassium persulfate (K₂S₂O₈) and sodium borohydride (NaBH₄) were obtained from Aladdin Biochemical Technology Co. Ltd. (Shanghai, China). Ti₃C₂ dispersion liquid and g-C₃N₄ were obtained from Jiangsu XFNANO Materials Tech, Co. Ltd. (Nanjing, China). Tris (4,4′-dicarboxylic acid-2,2′-bipyridyl)ruthenium (II) dichloride (Ru(dcbpy)₃Cl₂) was obtained from Suna Tech Inc. (Suzhou, China). Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) and 6-mercaptohexanol (MCH) were gained from Sigma-Aldrich (St Louis, MO, USA). Nb.BbvCI and 10 × NEB buffer were obtained from New England Biolabs (USA). The purified DNA sequences (Table S1) were synthesized by Genscript Bio-technology Co. Ltd. (Nanjing, China). Real human sera were collected from healthy volunteers at the Jiangyuan Hospital affiliated with Jiangsu Institute of Nuclear Medicine. And the medical ethical approval procedure was followed in the recovery experiment of SARS-CoV-2 RdRp assay, and consent was obtained from the subjects.

2.2. Instrumentation

Transmission electron microscopy (TEM) images were obtained from a JEM-2100F, JEOL (Japan). Cyclic voltammetry (CV) and electrochemical impedance spectroscopic (EIS) were measured by an electrochemical workstation (Shanghai Chenhua Instruments Co., China), performing in 5.0 mM [Fe(CN)₆]^3−/4− solution. ECL signals were captured by ECL-6B belonging to the State Key Laboratory of Analytical Chemistry for Life Sciences, Nanjing University. UV–Vis absorption spectrum was measured by Microplate Reader (Spectra Max M5e) with a range of 250 nm to 700 nm (Molecular Devices Co. Ltd, USA). Atomic force microscope (AFM) images were obtained from Dimension ICON (Bruker). The gel image was photographed by Bio-Rad ChemDoc XRS.

2.3. Preparation of the Au-g-C₃N₄

We successfully synthesized Au-g-C₃N₄ with reference to other people’s literature and made some changes [16]. In simple terms, 50 μL of HAuCl₄ (0.01 M) solution and 4 mL of g-C₃N₄ (0.15 mg mL⁻¹) were mixed and sonicated for 10 min, followed by stirring for 1 h. Then, 100 μL of fresh NaBH₄ (0.01 M) solution was injected into the mixed solution and treated in an ice bath for 20 min. Continuously, 50 μL of sodium citrate solution was added to the above-mixed solution, and the hybrid solution was continuously stirred at 25 °C for 20 min. Finally, to remove the excess AuNPs, sodium citrate and NaBH₄, the above-mixed solution was washed and sonicated three times to obtain pure Au-g-C₃N₄, which was redispersed in 1 mL of H₂O and stored at 4 °C for the next experiment.

2.4. Preparation of the PEI-Ru@Ti₃C₂@AuNPs-S7

The PEI-Ru@Ti₃C₂@AuNPs-S7 probe was successfully synthesized with slight modification by referring to others’ literature [44]. 10 mg of Ru(dcbpy)₃ ²⁺ and EDC (200 mM)/NHS (50 mM) were simultaneously dissolved in 5 mL of PBS (0.1 M, pH = 7.4) with continuous stirring for 2 h. The Ti₃C₂ suspension and PEI (1%, w/v) were mixed well and added to the PBS (0.1 M, pH = 7.4) of Ru(dcbpy)₃ ²⁺ for 1 h. Subsequently, 600 μL of HAuCl₄ (0.01 M) and 600 μL of fresh NaBH₄ (0.01 M) were added to the above solution and stirred for 1 h. The prepared PEI-Ru@Ti₃C₂@AuNPs were then centrifuged, washed three times to remove excess impurities and finally redispersed in 1 mL of PBS (0.1 M, pH = 7.4). For further preparing PEI-Ru@Ti₃C₂@AuNPs-S7 probes, 200 µL of S7 (10 μM) to 1 mL of Ru@Ti₃C₂@AuNPs was added and stored in a refrigerator at 4 °C for overnight. Finally, 100 μL of MCH (100 μM) was added to reduce non-specific adsorption. In this way, the preparation of the PEI-Ru@Ti₃C₂@AuNPs-S7 probe was made.

2.5. Preparation of TDNAs

The construction of TDNAs was based on previous literature with slight modifications[45]. Briefly, 10 μL of the four single-stranded DNAs (S3, S4, S5, S6) were dissolved in PBS buffer (pH = 7.4) containing 10 mM TCEP and 50 mM MgCl₂. Then, the four single-stranded DNAs were mixed and placed at 95 °C for 5 min, cooled to room temperature, and stored for 3 h in turn. Finally, sturdy TDNAs were obtained.

2.6. Polyacrylamide gel electrophoresis (PAGE) analysis

First of all, the prepared 20 % polyacrylamide gel run at 80 V for 40 min. Next, 10 μL of DNA sample (1 μM) and the loading buffer were mixed and injected into the gel sample tank and run at 110 V for 2.5 h. Then, 20 % of the gel was soaked in the Gel-Red diluent for 30 min. Finally, the stained gel was photographed by ChemiDoc MP Imager instrument. The gel electrophoresis experiments could prove the construction process of TDNAs, the entropy-driven reaction process, and the bipedal DNA walker.

2.7. Procedure of entropy-driven cycle amplification reaction

The SARS-CoV-2 RdRp gene initiated the reaction to generate a lot of bandage through the entropy-driven cyclic amplification reaction. Initially, 150 μL of three DNA strands, blocker (0.4 M), bandage (0.4 M), and scaffold (0.4 M) were placed at 95 °C for 5 min and slowly cooled to room temperature, so that a ternary structure of substrate could be formed. Then, various concentrations of the SARS-CoV-2 RdRp gene were added to the system and incubated at 25 °C. In this way, the early-formed intermediate 1 transformed to intermediate 2 while the bandage was released during the process. Simultaneously, 50 μL of the fuel chain (0.4 M) was added to the mixed solution and incubated at 25 °C for 40 min. As a result, a binary waste chain was formed and the SARS-CoV-2 RdRp gene was released to continually participate in the next cycle reaction.

2.8. Fabrication of the ratiometric ECL biosensor

At first, the glassy carbon electrode (GCE) was cleaned according to the previous method [46]. 10 μL of Au-g-C₃N₄ was imprisoned on GCE to obtain Au-g-C₃N₄/GCE, and dried at room temperature. Subsequently, the resulting Au-g-C₃N₄/GCE was immersed in 100 μL of TDNAs (1 μM) solution containing 10 mM TCEP for 8 h to reduce the S-S bond. Next, the modified GCE was immersed in 60 μL of MCH (100 μM) solution for one h to prevent non-specific adsorption. Then, the modified MCH/TDNA_S/Au-g-C₃N₄/GCE was incubated at 37 ℃ in the entropy-driven reaction mixture for 60 min (including different concentrations of SARS-CoV-2 RdRp gene). And the two single-stranded S1 (60 nM), S2 (60 nM) and Nb.BbvCI (10 U) was added into the mixed solution simultaneously, followed by the completed walking process of the DNA walker on the electrode surface. Finally, the synthesized PEI-Ru@Ti₃C₂@AuNPs-S7 probe was dropped on the surface of the modified electrode giving PEI-Ru@Ti₃C₂@AuNPs-S7/MCH/TDNAs/Au-g-C₃N₄/GCE, which was dried at room temperature. In each step mentioned above, PBS solution was required to wash off the unconnected substances. The constructed electrode was stored in a refrigerator at 4 °C for the next experiment.

2.9. Analysis procedure

The prepared electrode was placed in 5 mL of PBS (pH = 7.4) containing S₂O₈ ^2- (0.1 M) for the experiment, the ECL signal was captured by ECL-6B, and the scanning range was from −1.5–0 V. EIS and CV were measured by using the electrochemical workstation. A three-electrode system was used for testing, with modified 3 mm GCE as the working electrode, platinum wire electrode as the counter electrode and Ag/AgCl as the reference electrode.

3. Results and discussion

3.1. Mechanism of ratiometric ECL biosensor

We synthesized a ratiometric ECL biosensor to detect the SARS-CoV-2 RdRp gene based on entropy-driven and DNA walker’s cyclic amplification strategies. The principle of entropy-driven and DNA walker amplification strategy was shown in Scheme 1 . SARS-CoV-2 RdRp gene acts as a trigger to start cycle 1. With the help of the toehold of the substrate (consisting of the three chains of the scaffold, blocker, and bandage), SARS-CoV-2 RdRp can be combined with the scaffold to form an intermediate 1. At this time, since the binding force of the bandage and scaffold is weak, the bandage is released and the intermediate 1 transformed into intermediate 2. The exposed toehold of intermediate 2 is combined with fuel to produce double-stranded waste, blocker, and SARS-CoV-2 RdRp gene acting as a catalyst to stimulate the next cycle.

Scheme 1.

Schematic illustration of the entropy-driven and bipedal DNA walker amplification strategies based DNA tetrahedral ratiometric ECL biosensor for the assay SARS-CoV-2 RdRp gene.

After many cycles, a large amount of bandage is exported. Subsequently, Strand 1, Strand 2, and bandage combine to form a bipedal DNA walker in order to initiate cycle 2. Initially, the TDNAs were modified on the surface of Au-g-C₃N₄/GCE to obtain a strong ECL signal at 460 nm. In the presence of Nb.BbvCI endonuclease, the bipedal DNA walker can spontaneously walk on the hairpin at the top of the TDNAs. As a result, a large number of hairpins on top of the TDNAs are removed and bind specifically to the PEI-Ru@Ti₃C₂@AuNPs-S7 probe. Through ECL-RET, the ECL signal at 460 nm was quenched and at 620 nm was generated. The SARS CoV-2 RdRp gene can be quantitatively detected based on the degree of decrease in ECL signal at 460 nm and the increase in ECL signal at 620 nm. Combining with two signal amplification strategies, the ratiometric ECL biosensor has efficiently detected the SARS-CoV-2 RdRp gene.

3.2. Construction of the Dual-Wavelength ratiometric ECL-RET biosensor

The ratiometric ECL biosensor was designed based on the ECL-RET principle between Au-g-C₃N₄ and PEI-Ru@Ti₃C₂@AuNPs. As was shown in Fig. 1 A, the ECL emission spectrum of Au-g-C₃N₄ matched well with the UV–Vis absorption spectrum of PEI-Ru@Ti₃C₂@AuNPs. It was indicated that ECL-RET could be achieved between Au-g-C₃N₄ donor and PEI-Ru@Ti₃C₂@AuNPs acceptor. In addition, when the sensing system was scanned with a voltage from −1.5–0 V in the presence of S₂O₈ ^2-, there was an ECL emission peak on the Au-g-C₃N₄/GCE (Fig. 1B). However, the absence of ECL emission peak from PEI-Ru@Ti₃C₂@AuNPs/GCE indicated that the PEI-Ru@Ti₃C₂@AuNPs acceptor has not interfered with the ECL emission of Au-g-C₃N₄ donor at the cathode voltage. Fig. 1C showed the ECL intensity change curve before and after modification of the PEI-Ru@Ti₃C₂@AuNPs-S7 probe. A powerful ECL signal at 460 nm and a weak ECL signal at 620 nm appeared when the electrode was not modified with the PEI-Ru@Ti₃C₂@AuNPs-S7 probe. On the contrary, a weak ECL signal appeared at 460 nm and a powerful ECL signal appeared at 620 nm when the electrode was modified with the PEI-Ru@Ti₃C₂@AuNPs-S7 probe. It was further indicated that the ECL-RET between the Au-g-C₃N₄ donor and PEI-Ru@Ti₃C₂@AuNPs acceptor was reasonable. As displayed in Fig. 1D, a series of filters was used to explore the changes in ECL intensity with the increase of SARS-CoV-2 RdRp concentrations. When the SARS-CoV-2 RdRp was absent in the system, only the ECL emission peak appeared at 460 nm. However, as the concentration increased, the intensity of the ECL signal at 460 nm gradually decreased, and accordingly the ECL emission peak continued to increase at 620 nm. These results successfully demonstrated that the ratiometric biosensor of the ECL-RET could effectively detect the SARS-CoV-2 RdRp gene.

Fig. 1.

(A) UV–Vis absorption spectrum of PEI-Ru@Ti₃C₂@AuNPs (gray curve) and ECL spectrum of Au-g-C₃N₄ (red curve), ECL curves were obtained through a series of filters, measured in 0.1 M PBS (pH = 7.4) containing 0.1 M S₂O₈^2-. (B) ECL spectrum of GCE/Au-g-C₃N₄ (curve a) and GCE/PEI-Ru@Ti₃C₂@AuNPs (curve b), scanned at −1.5–0 V. (C) ECL signal intensity of (a) GCE/Au-g-C₃N₄/TDNAs/MCH, (b) GCE/Au-g-C₃N₄/TDNAs/MCH treated with 10 pM of SARS CoV-2 RdRp and further incubated with PEI-Ru@Ti₃C₂@AuNPs-S7. (D) Ratiometric ECL biosensor at different concentrations (0 aM, 10 aM, 10 fM, and 10 pM) of SARS CoV-2 RdRp was tested using a series of filters spaced 20 nm apart, measured in 0.1 M PBS (pH = 7.4) containing 0.1 M S₂O₈^2-. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.3. Possible luminescence mechanism about ECL biosensor

Referring to the previous luminescence mechanism of g-C₃N₄, the modified electrode was placed in the PBS solution containing 0.1 M S₂O₈ ^2-, and the scanning voltage was from −1.5–0 V. The possible luminescence mechanism is as follows equation (1) - (4). Firstly, g-C₃N₄ acquires an electron to become g-C₃N₄ ^•- (equation 1). Secondly, S₂O₈ ^2- is reduced to SO₄ ^2- and SO₄ ^•- (equation 2). Then electron transfer occurs between SO₄ ^•- and g-C₃N₄ ^•-, and then the g-C₃N₄ ^•- becomes excited state g-C₃N₄* (equation 3). Eventually, the excited state g-C₃N₄* becomes the ground state g-C₃N₄ to generate light (460 nm) (equation 4).

$$\begin{array}{c}\text{g}-{\text{C}}_3{\text{N}}_4+{\text{e}}^{-}\to \text{g}-{\text{C}}_3{\text{N}}_4^{·-}\end{array}$$ (1)

$$\begin{array}{c}{\text{S}}_2{\text{O}}_8^{2-}+{\text{e}}^{-}\to {\text{SO}}_4^{·-}+{\text{SO}}_4^{2-}\end{array}$$ (2)

$$\begin{array}{c}\text{g}-{\text{C}}_3{\text{N}}_4^{·-}+{\text{SO}}_4^{·-}\to \text{g}-{\text{C}}_3{\text{N}}_4^{\ast }+{\text{SO}}_4^{2-}\end{array}$$ (3)

$$\begin{array}{c}\text{g}-{\text{C}}_3{\text{N}}_4^{\ast }\to \text{g}-{\text{C}}_3{\text{N}}_4+hv(460\text{nm})\end{array}$$ (4)

3.4. Characterization of Au-g-C₃N₄

It can be seen from Fig. 2 A AuNPs with a diameter size of 8 ± 1 nm were uniformly dispersed on the surface of Au-g-C₃N₄. Fig. 2B was an energy dispersive X-ray (EDX) analysis. The results showed that the element peaks of C, N, and Au existed in the EDX spectrum of Au-g-C₃N₄, while the element peak of Au was not observed in the EDX spectrum of g-C₃N₄ (shown in the inset of Fig. 2B). The UV–Vis absorption spectrum was used to characterize Au-g-C₃N₄, as shown in Fig. 2C. Among them, g-C₃N₄ and AuNPs show characteristic peaks at around 360 nm and 520 nm, respectively. The prepared Au-g-C₃N₄ contained two characteristic peaks, which proved that Au-g-C₃N₄ was successfully synthesised. The ECL property of Au-g-C₃N₄ was verified in Fig. 2D. The ECL signal intensity of Au-g-C₃N₄ (curve b) was greater than that of g-C₃N₄ (curve a), which benefited from the outstanding conductivity of AuNPs. The CV curve also verified that Au-g-C₃N₄ (curve b’) possessed brilliant conductivity compared with single-component g-C₃N₄ (curve a’).

Fig. 2.

(A) TEM characterization of Au-g-C₃N₄, inset shows g-C₃N₄. (B) EDX spectrum of Au-g-C₃N₄, inset shows EDX of g-C₃N₄. (C) UV–vis absorption spectrum of g-C₃N₄ (curve a) and Au-g-C₃N₄ (curve b). (D) ECL intensity curves of g-C₃N₄ (curve a) and Au-g-C₃N₄ (curve b), cyclic voltammetry curves of g-C₃N₄ (curve a’) and Au-g-C₃N₄ (curve b’), modified on GCE tested in 0.1 M PBS (pH = 7.4) containing 0.1 M S₂O₈^2-.

3.5. Characterization of PEI-Ru@Ti₃C₂@AuNPs

Fig. 3A showed the TEM characterization of monolayer Ti₃C₂. From Fig. 3B, it can be seen that the AuNPs were uniformly dispersed on the surface of Ti₃C₂, and there was no aggregation, which indicated that Ti₃C₂ as an excellent carrier could prevent the accumulation of AuNPs. As seen in Fig. 3C, the diameter of AuNPs was 4.5 ± 1 nm. The PEI-Ru@Ti₃C₂@AuNPs nanocomposites characterized by EDX (see Fig. 3D) showed elemental peaks of C, N, O, Ti, Au, Ru, indicating that the PEI-Ru@Ti₃C₂@AuNPs nanocomposites were successfully synthesized. Besides, we verified the synthesis of PEI-Ru@Ti₃C₂@AuNPs with UV–vis characterization, as shown in Fig. S1. The characteristic absorption peaks of Au nanoparticles are around 525 nm, and the characteristic absorption peaks of Ru(dcbpy)₃ ²⁺ are around 300 nm and 475 nm. PEI-Ru@Ti₃C₂@AuNPs has both Au and Ru(dcbpy)₃ ²⁺ absorption peaks, which further verifies the successful synthesis of ECL acceptor materials.

Fig. 3.

(A) TEM characterization of Ti₃C₂. (B) TEM characterization of PEI-Ru@Ti₃C₂@AuNPs. (C) The corresponding size distribution diagram of (B). (D) The EDX spectrum of PEI-Ru@Ti₃C₂@AuNPs.

3.6. Characterization of TDNAs

Polyacrylamide gel electrophoresis (PAGE) was used to demonstrate the process of TDNAs construction, as shown in Fig. 4 A. Lane 1 to Lane 4 was the construction process from single-stranded S1 to four-stranded TDNAs. It can be seen from gel that as the number of DNA strands gradually increased, the corresponding band was further behind, which was consistent with the expected result. AFM was used to characterize TDNAs. In Fig. S2, the height of the TDNAs were 5 ± 0.2 nm, which is almost consistent with the theoretical value of 5.27 nm. This is further evidence that the TDNAs were successfully constructed.

Fig. 4.

Polyacrylamide gel electrophoresis analysis of (A) TDNAs construction process. Lane 1: S3, Lane 2: S3-S4, Lane 3: S3-S4-S5, Lane 4: S3-S4-S5-S6. (B) Entropy-driven cyclic amplification reaction, Lane 1: Scaffold, Lane 2: Blocker, Lane 3: Bandage, Lane 4: Substrate: Scaffold + Blocker + Bandage, Lane 5: Substrate + SARS-CoV-2 RdRp, Lane 6: Substrate + SARS-CoV-2 RdRp + Fuel. (C) Bipedal DNA walker, Lane 1: S1 probe, Lane 2: S2 probe, Lane 3: S1 + S2, Lane 4: S1 + S2 + Bandage.

3.7. Characterization of entropy-driven and bipedal DNA walker

PAGE was used to verify the entropy-driven reaction shown in Fig. 4B. The bands of scaffold, blocker and bandage were shown in Lane 1, Lane 2 and Lane 3, respectively. Lane 4 showed the band of the substrate formed by scaffold, blocker and bandage. By introducing SARS-CoV-2 RdRp, the substrate was converted into intermediate 2 and bandage with the help of toehold, and two new bands could be seen generated in Lane 5. Next, by introducing substrate, SARS-CoV-2 RdRp and fuel in the system, eventually generating binary waste, bandage and blocker could be observed in Lane 6, but the band of SARS-CoV-2 RdRp was not seen as involved in the next cycle.

Fig. 4C reflected the construction of the bipedal DNA walker. When there was no bandage in the system, S1 and S2 cannot combine with each other (Lane 3), and the position of the band was consistent with a single S1 (Lane 1) or S2 (Lane 2). However, when there was bandage in the system and S1 and S2 could form a bipedal DNA walker, the molecular weight was larger, and the band position was later in the gel. Therefore, it proved that by introducing bandage, bipedal DNA walker was successfully constructed.

3.8. Effects of material modifications on the performance of the electrodes

The CV was used to characterize the construction process of the biosensor in Fig. 5 A. The bare electrode (curve a) maintained the most prominent current peak. When Au-g-C₃N₄ (curve b) was deposited on the electrode surface, the current peak decreased. Then continued to modify the TDNAs (curve c) and MCH (curve d) on the electrode, the current peak continued to decline that the surface of the electrode adsorbed substances with poor conductivity. When SARS-CoV-2 RdRp and Nb.BbvCI were introduced (curve e), the walking process of DNA walker was completed on the electrode surface. As a result, TDNAs were excised and the peak current increased. Finally, the specific binding of the PEI-Ru@Ti₃C₂@AuNPs-S7 probe (curve f) was modified to the modified electrode caused the peak current to be increased. Meanwhile, the effect of material modifications on the CV characteristics of the electrode was also explored, as shown in Fig. S3. In the absence of the co-reactant S₂O₈ ^2- (curve a), the redox peak of the CV characteristic curve is very weak. However, when the co-reactant S₂O₈ ^2- is present (curve b), the curve shows a strong redox peak at −1.1 V, which indicates that S₂O₈ ^2- plays a significant role in the luminescence of C₃N₄. With the modification of TDNs (curve c) and MCH (curve d), the redox current gradually decrease. However, the peak of the electrode starts to increase when SARS CoV-2 RdRp and Nb.BbvCI (curve e) are involved in the reaction. In particular, the redox current further increases when PEI-Ru@Ti₃C₂@AuNPs (curve f) with excellent conductivity is modified to the electrode. The effect of material and DNA probe on the CV characteristics of the electrode is essentially the same as in the presence of 5 mM [Fe(CN)₆]^3-/4-, which mainly affects the electrode performance by influencing the electron transfer ability.

Fig. 5.

Characterization of the stepwise construction process of the biosensor with CV (A), EIS (B) and ECL (C). (a) bare GCE, (b) GCE/Au-g-C₃N₄, (c) GCE/Au-g-C₃N₄/TDNAs, (d) GCE/Au-g-C₃N₄/TDNAs/MCH, (e) GCE/Au-g-C₃N₄/TDNAs/MCH treated with SARS CoV-2 RdRp and Nb.BbvCI, (f) Futher incubation with PEI-Ru@Ti₃C₂@AuNPs-S7. The inset in Fig. 5B was the modified Randles circuit to simulate the electrochemical process during electrode modification.

The construction process of the biosensor was characterized by EIS shown in Fig. 5B. The classically modified Randles circuit (inset in Fig. 5B) can simulate the electrochemical process of the modified electrode. The resistance of the electrode surface, the electron transfer and matter transfer during the redox process are mainly expressed through the resistance to charge transfer (Rct), which is precisely quantified especially through the Nyquist diagram’s high-frequency semicircle [19]. The impedance value of the bare electrode (curve a) was the smallest. Continued to modify Au-g-C₃N₄ (curve b), TDNAs (curve c), MCH (curve d), and the impedance value gradually increased. When SARS-CoV-2 RdRp and Nb.BbvCI (curve e) presented, the TDNAs on the surface were clipped and the impedance value was slightly reduced. Finally, the suitable conductive PEI-Ru@Ti₃C₂@AuNPs-S7 probe (curve f) was modified on the electrode to obtain the impedance to be declined.

The ECL response to characterize the electrode modified with different substances was shown in Fig. 5C. The electrode was immersed in PBS solution containing 0.1 M S₂O₈ ^2- for measuring. The bare electrode (curve a) had no ECL response. Spreading Au-g-C₃N₄ (curve b) with good luminescence performance on the surface of the electrode presented a higher ECL signal. When TDNAs (curve c) with poor conductivity and MCH (curve d) were modified on the electrode surface, the ECL signal gradually decreased. Subsequently, SARS-CoV-2 RdRp (curve e) was introduced and the ECL signal recovered slightly. Finally, the PEI-Ru@Ti₃C₂@AuNPs-S7 probe (curve f) was connected to TDNAs, and the ECL signal was reduced. Through these three methods, we could confirm that the biosensor was successfully constructed.

3.9. Optimization of analytical conditions

In order to obtain the optimal performance of the biosensor, some experimental parameters were optimized, including the concentration of S1 and S2 of the DNA walker, the dosage of Nb.BbvCI, the digestion time of Nb.BbvCI, and the incubation time of the entropy-driven reaction. Fig. S4A depicted that as the concentrations of S1 and S2 increased, the ECL signal gradually decreased and tended to stabilize at 60 nM. Fig. S4B depicted that when the dosage of Nb.BbvCI reached 10 U, the ECL curve becomed stable. Fig. S4C showed that the optimal digestion time for Nb.BbvCI was 60 min. The optimal time for the entropy-driven reaction was 40 min as shown in Fig. S4D. Therefore, the concentration of S1 and S2 of DNA walker was selected to be 60 nM, 10 U and 60 min were used for the optimal dosage and optimal digestion time of Nb.BbvCI, and the optimal entropy-driven reaction time was 40 min.

3.10. Detection of SARS-CoV-2 RdRp with the ratiometric biosensor

Under the optimized conditions, the ratiometric ECL biosensor was used to detect SARS-CoV-2 RdRp gene. Fig. 6 A depicted that the ECL signal intensity decreased at 460 nm and the ECL signal intensity increased at 620 nm as the concentrations of SARS-CoV-2 RdRp increased from 10 aM to 10 pM. Fig. 6A also described that the change of ECL signal intensity at 460 nm possessed an excellent linear relationship of y = -502.82 lgC_(RdRp) − 986.17, R² = 0.9944. The change of ECL signal intensity at 620 nm also revealed an excellent linear relationship of y = 805.79 lgC_(RdRp) + 14704.28, R² = 0.9950. Here y represents the ECL intensity at 460 nm or 620 nm and lgC_(RdRp) represents the logarithmic value of RdRp concentration. In order to obtain higher reliability, we evaluated the logarithmic value of ECL (620 nm)/ECL (460 nm) as the concentration of SARS-CoV-2 RdRp increased, as shown in Fig. 6B. According to the limit of detection (LOD) = 3σ/k, the LOD was calculated as 7.8 aM. Compared with other methods in detecting DNA, our method possessed a lower detection limit, as shown in Table S2. Moreover, our method has a significant advantage in LOD, compared to other methods for SARS-CoV-2 detection (as shown in Table S3). The lower detection limits obtained result from our ratiometric ECL biosensing platform constructed on DNA tetrahedra, which combines the significant advantages of entropy-driven reactions and DNA walker cycling as signal amplification techniques. The conventional assays using molecular beacons (MBs) probes are shown schematically in Fig. S5. In the absence of target DNA, MBs exhibit a stem-loop structure that brings the attached fluorophore (Cy3) and quencher (BHQ-2) in close proximity, reducing the fluorescence emission. However, in the presence of target DNA, target DNA binds to MBs to form a hybrid complex, disrupting the stem-loop structure, which leads to the separation of Cy3 from BHQ-2, producing a large increase in fluorescence. By using the method with MBs, we obtained a LOD of 9.8 nM (see Fig. S6), which was much less sensitive than our method.

Fig. 6.

(A) Relationship between the ECL intensity of (a) Au-g-C₃N₄ at 460 nm and (b) PEI-Ru@Ti₃C₂@AuNPs at 620 nm and the logarithmic value of SARS-CoV-2 RdRp concentrations (10 aM, 100 aM, 1 fM, 10 fM, 100 fM, 1 pM, 10 pM). (B) The calibration curve for SARS-CoV-2 RdRp. (C) Stability of the ECL biosensor with 16 continuous measurements at 460 nm (left) and 620 nm (right), respectively, with a SARS-COV-2 RdRp concentration of 10 pM. (D) Selectivity of ratiometric ECL biosensor at (a) 1 pM for SARS-CoV-2 RdRp, (b) 100 pM for SARS-CoV RdRp, (c) 100 pM for random DNA, (d) blank solution. The ECL measurement used a series of filters spaced 20 nm apart, measured in 0.1 M PBS (pH = 7.4) containing 0.1 M S₂O₈^2-.

3.11. Stability and selectivity of the ratiometric biosensor

In order to evaluate the stability of the biosensor, the modified electrode was immersed in PBS containing 10 pM SARS-CoV-2 RdRp and scanned continuously for 16 cycles. It can see from Fig. 6C that the ECL signals remain stable at 460 nm with a relative standard deviation (RSD) of 2.34%, and at 620 nm with an RSD of 3.61%. In order to evaluate the selectivity of the ratiometric biosensor, some interferences were introduced. When the biosensor was immersed in 1 pM of SARS-CoV-2 RdRp, 100 pM of SARS-CoV RdRp, 100 pM of random DNA and blank. As shown in Fig. 6D, the biosensor processed by SARS-CoV-2 RdRp exhibited a significant distinction compared with the biosensor processed by other interferences. Therefore, the constructed biosensor possessed splendid selectivity.

3.12. Applicability of the biosensor in real sample analysis

It is necessary for the ratiometric biosensor to be used for early clinical detection. The recovery experiment was performed in 50-fold diluted human serum, shown in Table 1 . The biosensor was carried out in different concentrations of SARS-CoV-2 RdRp serum, and it was found that the recovery value was from 92.00% to 103.56%, and the RSD was from 1.51% to 4.30%. It exhibited that this biosensor could be applied to the detection of actual samples.

Table 1.

Recovery results for the assay of SARS-CoV-2 RdRp in 50-fold dilution of human serum.

Sample number	Added	Found	Recovery (%)	RSD (%, n = 3)
1	10 aM	9.45 aM	94.50	1.51
2	100 aM	101.35 aM	101.35	3.63
3	1 fM	0.97 fM	97.00	2.24
4	10 fM	10.25 fM	102.50	1.92
5	100 fM	103.56 fM	103.56	4.30
6	1 pM	0.92 pM	92.00	4.11
7	10 pM	9.79 pM	97.90	2.98

4. Conclusions

In summary, we reasonably designed a dual-wavelength ratiometer TDNAs ECL biosensor to detect SARS-CoV-2 RdRp gene combined with entropy-driven and bipedal DNA walker cycle amplification strategies. This ECL biosensor has the following advantages. First, in this ratiometric ECL biosensor, Au-g-C₃N₄ as an excellent donor and PEI-Ru@Ti₃C₂@AuNPs as an excellent acceptor were designed to achieve ECL-RET. Second, unlike the traditional DNA walker, the bipedal DNA walker walks on a walking track by bandage DNA combining two adjacent feet. However, two feet without the participation of bandage DNA cannot walk on the track because they are farther apart. Besides, TDNAs modified by hairpin structures act on the electrode surface as walking tracks for the bipedal DNA walker. Third, Ultra-sensitive detection was achieved by combining two amplification strategies, entropy-driven and bipedal DNA walker. The detection range of the biosensor can be from 10 aM to 10 pM, and the detection limit was as low as 7.8 aM. The dual-wavelength ratiometric ECL biosensor was used in human serum to exhibit an excellent recovery rate and could be used in early clinical diagnosis.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: