Application of iridium(III) complex in label-free and non-enzymatic electrochemical detection of hydrogen peroxide based on a novel “on-off-on” switch platform

We herein report a label-free and non-enzymatic electrochemical sensor for the highly sensitive detection of hydrogen peroxide (H2O2) based on a novel “on-off-on” switch system. In our design, MB was used as an electron mediator to accelerate the electron transfer while AuNPs was used to amplify the electrochemical signal due to its excellent biocompatibility and good conductivity. The “switch-off” state was achieved by introducing the guanine-rich capture probe (CP) and an iridium complex onto the electrode surface to form a hydrophobic layer, which then hinders electron transfer. Upon addition of H2O2, fenton reaction occurs and produces OH• in the presence of Fe2+. The OH• cleaves the CP into DNA fragments, thus resulting in the release of CP and iridium complex from the sensing interface, recovering the electrochemical signal to generate a “switch-on” state. Based on this novel switch system, a detection limit as low as 3.2 pM can be achieved for H2O2 detection. Moreover, satisfactory results were obtained by using this method for the detection of H2O2 in sterilized milk. To the best of our knowledge, this is the first G-quadruplex-based electrochemical sensor using an iridium(III) complex.

G-quadruplex-based sensing platforms are convenient, low-cost and relatively quick. Previously, our group has synthesized different iridium(III) complexes and discovered that some of them have highly specific binding properties to G-quadruplex, and a number of detection platforms were subsequently constructed based on the specific interaction between iridium(III) complexes and G-quadruplex DNA [32][33][34][35][36][37] , through detecting the luminescent enhancement of iridium(III) complexes.
Recently, we discovered that some of the iridium(III) complexes can hinder the electron transfer between the electrode surface and the electrolyte mainly because of the hydrophobicity and the large molecular structure of them. Based on of this unique property, we designed the application of one type of the iridium(III) that specific to G-quadruplex (the structure of it were shown in Fig. 1) in label-free and non-enzymatic electrochemical sensing platform for H 2 O 2 by using a novel "on-off-on" switch system. As shown in Fig. 1, the first "switch on" state was achieved based on the immobilize of Nafion (Nf) and methylene blue (Nf@MB) composite membrane as inner-layer, and self-assembled positively charged gold nanoparticles ((+ )AuNPs) as outer-layer. Successively, the capture probe (CP, guanine-rich nucleic acid) was modified onto the (+ )AuNPs surface through an Au-S bond, and the CP would change to G-quadruplex structure in the presence of K + , followed by the specific binding with an iridium(III) complex using a simple and label-free method [38][39][40][41][42][43][44] . As a consequence, a hydrophobic and large molecule structural layer wad formed on the electrode surface, and resulted in a "switch off " state. Then, upon incubation of the sensor with H 2 O 2 and Fe 2+ , Fenton reaction would happen and produced tremendous OH • to cleave CP into DNA fragments, leading to the release of such fragments and the iridium(III) complex from the sensing interface, recovering the electrochemical signal and achieving a "switch on" state. Importantly, the electrochemical signal change was proportional to H 2 O 2 concentration. Thus, H 2 O 2 could be detected by using such proposed method based on monitoring the signal change of MB. There are several advantages for such a proposed method: Firstly, due to the good stability of negatively charged Nf membrane on the electrode surface, abundant positively charged and highly stable Nf@MB membrane could formed on the electrode, which can greatly improve the stability of the sensor. In addition, MB is highly conductive as a redox indicator, which can great enhance the sensitivity of the sensor. Meantime, about 4 nm of (+ )AuNPs, with excellent biocompatibility and good conductivity, was successfully self-assembly onto the Nf@MB membrane, which can greatly enlarge the electrode surface for more CP immobilization. Moreover, to the best of our knowledge, this is the first application of the iridium(III) complex and (+ )AuNPs in electrochemical sensor for H 2 O 2 analysis. We envision that this "on-off-on" platform could also provide new opportunities for biosensor development.

Results and Discussion
To investigate the conductivity of the iridium(III) complex, several modified electrodes were characterized by cyclic voltammetry experiments (CVs) in the presence of 5 mM Fe(CN) 6 3−/4− . As seen from curve a in Fig. 2A, a stable and well-defined redox peak was obtained when the bare gold electrode was scanned in 5 mM of Fe(CN) 6 3−/4− . After the immobilization of CP on the electrode surface, an obvious decrease of the peak current was observed, which was attributed to the fact that DNA hinders electron transfer (curve b) 45,46 . Then, G-quadruplex structures were formed after the CP-modified electrode was incubated in tris-buffer containing 100 mM of K + . Subsequently, the interaction of iridium(III) complex with the G-quadruplex led to a further decrease in peak current, because of the high resistance of the electrode interface induced by the adsorption of iridium(III) complex and the formation of a hydrophobic layer on the electrode surface (curve c). Since other organic dyes such as crystal violet (CV) and rhodamine can also specifically bind to G-quadruplex 47,48 , we also investigated the conductivity of CV and rhodamine after they bound to the G-quadruplex that immobilized on the electrode. The results demonstrated that CV and rhodamine would accelerate the electron transfer (curve d and e). Therefore, to obtain the step of the "switch off " state, we selected iridium(III) complex as the hydrophobic layer.
In order to confirm the successful fabrication of the proposed sensing platform, the fabrication process of the sensor was characterized by DPV experiments. As shown in Fig. 2B, no oxidation peak was observed at the bare gold electrode while an obvious oxidation peak appeared after electro-polymerization of MB on the electrode surface, which is mainly due to the high conductivity of MB as an electron mediator (curve b). Then, the peak current increased after the electro-deposition of AuNPs onto the MB film on account of the excellent conductivity of AuNPs (curve c). However, after the interaction of iridium(III) complex with G-quadruplex form of CP in the presence of K + , the peak current obviously decreased, since iridium(III) complex would form a hydrophobic and large molecule structural layer on the electrode surface and hinder the electron transfer (curve d). However, upon incubation of the sensor with 1.5 nM of H 2 O 2 and Fe 2+ , a dramatic increase of the peak current was observed (curve e). This increase could be ascribed to the release of iridium(III) complex and CP fragments from the electrode surface after the cleavage of CP by Fenton reaction in the presence of Fe 2+ and H 2 O 2 .
To validate the mechanism of the assay, circular dichroism (CD) spectroscopy was performed. As shown in Fig. 2C, there was no obvious peak for CP alone, while a positive Cotton effect peak at around 264 nm and a negative Cotton effect peak at around 236 nm appeared after CP was incubated with tris-buffer containing 100 mM of K + , indicating the formation of the classic G-quadruplex structure in accordance with literature 49 . No significant change was observed upon the selective interaction of the iridium(III) complex with the G-quadruplex structure of CP (Fig. 2D, curve a and b). However, the intensity at the peak of 264 nm and 236 nm decreased dramatically when the CP was incubated with H 2 O 2 , these results might be attributed to the cleavage of CP (curve c).
Experimental conditions including the self-assembly time of (+ )AuNPs, the concentration of Fe 2+ and the cleavage time of CP by H 2 O 2 were optimized. The self-assembly time of (+ )AuNPs is an important factor that might affect the properties of the sensor. As shown in Fig. 3A, the current intensity was proportional to the self-assembly time of (+ )AuNPs from 0 to 30 min and then reached a plateau. In view of the sensitivity of the sensor, an electro-deposition time of 30 min was chosen for all the experiments.
The role of Fe 2+ in the Fenton reaction is to catalyze the cleavage of DNA. Thus, Fe 2+ concentration is an important factor that can affect the degree of CP cleavage. Figure 3B shows that the current signal increased along with the increase of Fe 2+ concentration up to 1.5 μ M in the presence of 1.2 nM H 2 O 2 , indicating that CP could be DNA structure depends on the pH environment, so the cleavage activity of H 2 O 2 and Fe 2+ may depend on the pH as well. As shown in Fig. 3C, the current signal increased with the pH over the range from 5.0 to 6.5 in the absence of H 2 O 2 , and reached a plateau over the range of 6.5-7.4. However, when the pH was higher than 7.4, the current signal decreased dramatically, such result might be due to the fact that an acidic environment can stabilize the Fe 2+ ions, and higher pH can induce the oxidation of ferrous ions to ferric ions under atmosphere (O 2 ). Considering the sensitivity of the sensor, we selected 7.4 for further experiments.
Since the second "switch-on" state depends on the cleavage of CP, the cleavage time of CP in the presence of 1.2 nM H 2 O 2 was investigated in Fig. 3D. It was found that the current signal increased when the cleavage time increased from 0 to 25 min, and then reached a plateau. To achieve the effective cleavage of CP, 30 min was chosen as the cleavage time of CP.
To evaluate the effect of H 2 O 2 concentration on current signal, the proposed sensor was incubated with tris-buffer that contained different concentration of H 2 O 2 . As shown in Fig. 4A, the DPV signal increased along with the increase of H 2 O 2 concentration. According to the calibration plot of Fig. 4B, the linear regression equation was I = 11.18 + 7.87c (c: nM) in a dynamic range from 8.0 pM to 2.0 nM with a correlation coefficient of R 2 = 0.994. The detection limit was 3.2 pM calculated by the three-signal method. Here, the electro-deposition of AuNPs onto MB film can greatly enlarge the electrode surface for CP immobilization, and accordingly improve the sensitivity of the sensor. The detection limit of our "on-off-on" detection platform for H 2 O 2 detection was lower than other AuNPs 49,50-53 or MB 54-57 based electrochemical methods, which can satisfy the demand for H 2 O 2 detection in real samples such as food or industrial products. Thus, our proposed sensor could be used to quantify H 2 O 2 for food safety and clinical diagnosis.
The stability of the sensor was investigated by CV. After continuous scanning for 20 cycles, the sensor retained 98.1% of its initial response, indicating acceptable stability. Meantime, the selectivity of such sensor for H 2 O 2 was evaluated against other interferences including ascorbic acid, purine trione, glucose and mixed metal ions (contained Pb 2+ , Cu 2+ , K + , Ca 2+ and Mg 2+ ). From the results in Fig. 5, it could be seen that the current intensity change (Δ I) upon addition of ascorbic acid (150 nM), purine trione (150 nM), glucose (150 nM) and metal ions (150 nM of Pb 2+ , Cu 2+ , K + , Ca 2+ and Mg 2+ ) were much lower than that of H 2 O 2 (1.5 nM). Moreover, the Δ I of the sensor after the incubation of it with a solution containing both H 2 O 2 and the interferences was almost the same with that of H 2 O 2 (1.5 nM) only. Such high selectivity could be attributed to the highly specific cleavage of DNA based on the Fenton reaction in the presence of H 2 O 2 and Fe 2+ . H 2 O 2 is used as a stabilizer in milk in some European and American countries. However, residual H 2 O 2 may cause adverse effects to the human body. Thus, the sensitive and rapid detection of H 2 O 2 in milk samples is of great importance. To evaluate the robustness of the system for H 2 O 2 detection in milk, milk samples were collected and centrifuged at 13600 rpm to remove fat. After corroborating the absence of hydrogen peroxide, such milk samples were diluted ten-fold with tris-buffer (pH 7.4) for recovery experiments. As shown in Table 1, a good recovery ranging from 96.0% to 109% and a relative standard deviation (RSD) between 1.38% and 3.16% were obtained (each sample was tested 3 times). This result indicated that our sensor has potential application for the quantitative determination of H 2 O 2 level in real samples.
In conclusion, a novel and label-free "on-off-on" switch system was successfully developed for the sensitive detection of H 2 O 2 . The first "switch on" state was achieved based on the immobilization of Nf@MB and (+ ) AuNPs. Here, MB was used as the current signal and (+ )AuNPs were used to enlarge the surface of the electrode and improve the sensitivity of the sensor. Then, the "switch off " state was obtained based on the formation of a massive hydrophobic layer that contained G-quadruplex structures and iridium(III) complex on the electrode surface. After the incubation of the sensor with H 2 O 2 , CP would be cleaved into DNA fragments and released from the electrode surface, followed by signal recovering as the second "switch on" state. The novel concept of this "on-off-on" platform was successfully used for H 2 O 2 detection with a detection limit down to 3.2 pM. Moreover, experiments proved that such strategy could be applied effectively for H 2 O 2 detection in milk samples.   Apparatus. Electrochemical measurements were monitored by a CHI 630C electrochemical workstation (CH Instruments, Inc. U.S.A.). A conventional three-electrode system consisted of a modified working electrode, a platinum wire counter electrode and an Ag/AgCl reference electrode was used in the experiment. CD spectroscopy measurements were constructed by using an Olis 17/UV/VIS/NIR spectropolarimeter at room temperature.
H 2 O 2 sensing protocol. Gold electrodes were pretreated according to our previous method. Then, 0.5 mg of MB was added into 2 mL of 0.5% Nf ethanol solution and sonicated to obtain a homogeneous suspension contained Nf@MB. Next 5 μ L of Nf@MB was cast on the pretreated electrode and dried in the air to obtain an Nf@MB film. After that, 4 nm of (+ )AuNPs prepared according to literature 59 was self-assembly onto the Nf@ MB film for 30 min, based on the electrostatic adsorption to enlarge the electrode surface for CP immobilization, Subsequently, the modified electrodes were incubated in 10 mM of tris-buffer (pH 7.4) that contained 2.0 μ M of CP for 12 h at room temperature (prior to modification, the disulfide bond at the 3′ end of CP was cleaved with tris(2-carboxyethyl)phosphine (TCEP)), followed by reaction with 2.0 μ M of MCH for 1 h to block nonspecific sites, and incubated in 10 mM of tris-buffer (pH 7.4) that contained 100 mM of K + and 3.0 μ M of iridium(III) complex for another 1 h to induce the formation of a hydrophobic layer on the electrode surface. Then, the prepared sensor was incubated with different concentrations of H 2 O 2 at 37 °C for 30 min in the presence of Fe 2+ . Finally, the electrochemical characteristics of the sensor were investigated in tris-buffer by using differential pulse experiments (DPV) from − 500 mV to 0.0 mV at room temperature.