Ultrasensitive Electrochemical DNA Biosensor Fabrication by Coupling an Integral Multifunctional Zirconia-Reduced Graphene Oxide-Thionine Nanocomposite and Exonuclease I-Assisted Cleavage

In this work, a simple but sensitive electrochemical DNA biosensor for nucleic acid detection was developed by taking advantage of exonuclease (Exo) I-assisted cleavage for background reduction and zirconia-reduced graphene oxide-thionine (ZrO2-rGO-Thi) nanocomposite for integral DNA recognition, signal amplification, and reporting. The ZrO2-rGO nanocomposite was obtained by a one-step hydrothermal synthesis method. Then, thionine was adsorbed onto the rGO surface, via π-π stacking, as an excellent electrochemical probe. The biosensor fabrication is very simple, with probe DNA immobilization and hybridization recognition with the target nucleic acid. Then, the ZrO2-rGO-Thi nanocomposite was captured onto an electrode via the multicoordinative interaction of ZrO2 with the phosphate group on the DNA skeleton. The adsorbed abundant thionine molecules onto the ZrO2-rGO nanocomposite facilitated an amplified electrochemical response related with the target DNA. Since upon the interaction of the ZrO2-rGO-Thi nanocomposite with the probe DNA an immobilized electrode may also occur, an Exo I-assisted cleavage was combined to remove the unhybridized probe DNA for background reduction. With the current proposed strategy, the target DNA related with P53 gene could be sensitively assayed, with a wide linear detection range from 100 fM to 10 nM and an attractive low detection limit of 24 fM. Also, the developed DNA biosensor could differentiate the mismatched targets from complementary target DNA. Therefore, it offers a simple but effective biosensor fabrication strategy and is anticipated to show potential for applications in bioanalysis and medical diagnosis.

In this work, a simple but sensitive electrochemical DNA biosensor for nucleic acid detection was developed by taking advantage of exonuclease (Exo) I-assisted cleavage for background reduction and zirconia-reduced graphene oxide-thionine (ZrO 2 -rGO-Thi) nanocomposite for integral DNA recognition, signal amplification, and reporting. The ZrO 2 -rGO nanocomposite was obtained by a one-step hydrothermal synthesis method. Then, thionine was adsorbed onto the rGO surface, via π-π stacking, as an excellent electrochemical probe. The biosensor fabrication is very simple, with probe DNA immobilization and hybridization recognition with the target nucleic acid. Then, the ZrO 2 -rGO-Thi nanocomposite was captured onto an electrode via the multicoordinative interaction of ZrO 2 with the phosphate group on the DNA skeleton. The adsorbed abundant thionine molecules onto the ZrO 2 -rGO nanocomposite facilitated an amplified electrochemical response related with the target DNA. Since upon the interaction of the ZrO 2 -rGO-Thi nanocomposite with the probe DNA an immobilized electrode may also occur, an Exo I-assisted cleavage was combined to remove the unhybridized probe DNA for background reduction. With the current proposed strategy, the target DNA related with P53 gene could be sensitively assayed, with a wide linear detection range from 100 fM to 10 nM and an attractive low detection limit of 24 fM. Also, the developed DNA biosensor could differentiate the mismatched targets from complementary target DNA. Therefore, it offers a simple but effective biosensor fabrication strategy and is anticipated to show potential for applications in bioanalysis and medical diagnosis.
Keywords: electrochemical DNA biosensor, reduced graphene oxide, zirconia, signal amplification, exonuclease I INTRODUCTION Highly sensitive detection of nucleic acid analytes is always pursued by researchers to accommodate the ever-increasing demands in disease diagnosis, environmental and food monitoring, and forensic identification (Debouck and Goodfellow, 1999;Liu et al., 2008;Zhang et al., 2013;Chen et al., 2019a). Polymerase chain reaction is the most commonly used method for the detection of low amounts of nucleic acids (Asiello and Baeumner, 2011;Xu et al., 2017), but the delicate temperature control by a relatively expensive thermal cycler and the complex primer design limit its wide applications, especially in some resource-constrained regions. Compared with it, biosensor technology is especially attractive for bioanalysis due to its potential advantages including economy, portability, and facile operation. Until now, various techniques such as electrochemistry, spectroscopy, surface plasmon resonance, electrochemiluminescence, and photoelectrochemistry have been well-explored for nucleic acid biosensor fabrication (Hu et al., 2014;Shi et al., 2016;Ding et al., 2017;Feng et al., 2017;Li et al., 2018). Among them, the electrochemical method possesses some inherent advantages such as simple instrumentation, signal stability, flexible operation, and easy integration and miniaturization Wang et al., 2015;Liu et al., 2017). In order to upgrade the DNA detection sensitivity to satisfy the requirement of profiling trace amounts of nucleic acids, various signal amplification strategies have been explored for biosensor fabrication, for example, nuclease-based target recycling or enzyme-free DNA assembly strategies have been well-developed Chen et al., 2019b). By virtue of various nucleases such as nucleic acid endonuclease, exonuclease, and polymerase, the target DNA amount could be indirectly amplified owing to target cycling or generation of target analogs. Enzyme-free nucleic acid assembly strategies such as catalytic hairpin assembly, hybridization chain reaction, and DNA-fueled target recycling are usually based on the cascade toehold-mediated strand displacement reactions for signal amplification toward target DNA recognition events (Lv et al., 2015;Ding et al., 2018;Karunanayake Mudiyanselage et al., 2018). Although these nucleases or enzyme-free DNA strategies could substantially improve the detection limit of the target DNA, the relatively complex or rigorous sequence design or the use of too much DNA fragments or various nucleases for signal amplification increases the assay cost and also the risk for error readouts. Nanomaterial or enzyme-based post-amplification means could be considered as another kind of widely used strategies for the amplified detection of nucleic acid. In these strategies, a "sandwich-type" detection mode is usually adopted with the first immobilization of capture probe DNA on the electrode surface. Upon hybridization with the target DNA, the bioprobe-labeled nanomaterial or enzyme was then recognized for signal amplification. Until now, various nanomaterials such as noble metal nanoparticles, carbon nanotube, graphene, and semiconductor nanoparticles have been fully explored for biosensor fabrication (Swain et al., 2008;Wu et al., 2009;Nie et al., 2012;Benvidi et al., 2015;Jahanbani and Benvidi, 2016;Baluta et al., 2018;Yan et al., 2018;Yu et al., 2019). However, the careful and tedious modification, control, or labeling of biomolecules or reporters onto nanomaterials increases the assay complexity. Thus, development of simple and effective signal amplification means for electrochemical nucleic acid detection is still in high demand.
Herein a simple and sensitive electrochemical nucleic acid biosensor was developed by coupling a zirconia-reduced graphene oxide-thionine (ZrO 2 -rGO-Thi) nanocomposite and exonuclease I (Exo I)-assisted cleavage. The ZrO 2 -rGO nanocomposite was obtained by a one-step hydrothermal synthesis method. Then, a large amount of thionine molecules were adsorbed onto the rGO surface, via π-π stacking, as an excellent electrochemical probe. Graphene or graphene oxide-based two-dimensional materials have been widely explored for application in biosensor and biomedicine, owing to their fascinating electronic, thermal, mechanical, and chemical properties (Dong et al., 2012;Chen et al., 2016;Cao et al., 2018;Campos et al., 2019). The obtained ZrO 2 -rGO-Thi nanocomposite was used as an integral multifunctional sensing element for DNA recognition, signal amplification, and reporting. The ZrO 2 could recognize the phosphate group on the DNA skeleton via a multicoordinative interaction. The adsorbed abundant thionine molecules onto the ZrO 2 -rGO nanocomposite are responsible for signal amplification and reporting toward a DNA recognition event. The unique electronic properties of rGO would be also beneficial for improved electrochemical response. Furthermore, it avoids the relatively complex modification, control, or labeling procedures for most of the nanomaterial-based signal amplification strategies. Considered that the ZrO 2 -rGO-Thi nanocomposite may also interact with probe DNA immobilized electrode, a simple Exo I-assisted cleavage was further combined to achieve the discrimination of hybridized from unhybridized DNA and contribute to background reduction. The target DNA related with P53 gene was sensitively analyzed. The P53 gene has been well-known as the guardian of the genome that can code and express p53 protein to suppress the malignant transformation of cells. The p53 gene has been associated with the occurrence of many human tumors such as liver, breast, bladder, and stomach cancers and so on (Hasanzadeh and Shadjou, 2017;Shen et al., 2018). Therefore, the current electrochemical nucleic acid biosensing strategy not only would pave a new avenue for biosensor fabrication by nanomaterials but also is anticipated to show potential for applications in bioanalysis and medical diagnosis.

Synthesis of ZrO 2 -rGO and ZrO 2 -rGO-Thi NAnocomposites
The ZrO 2 -rGO and the ZrO 2 -rGO-Thi nanocomposites were prepared according to our reported method . Simply mixed into 20 ml deionized water were 0.004 g of ZrOCl 2 ·8H 2 O and 0.004 g of GO, and the mixture was ultrasonically treated for 1 h. The hydrothermal synthesis process for the above mixture was operated at 160 • C for 10 h. After the centrifugal collection of the precipitate and freeze-drying, the ZrO 2 -rGO nanocomposite was obtained. The ZrO 2 -rGO-Thi nanocomposites were prepared by mixing 2.5 mg of ZrO 2 -rGO into 10 ml of thionine solution (1 mM) and agitating for 1 h.

DNA Immobilization on the Electrode Surface
The gold electrodes (2 mm in diameter) were pretreated prior to probe DNA immobilization. The electrodes were polished with 0.3 and 0.05 µm alumina oxide slurries for 5 min, respectively, followed by ultrasonic cleaning in acetone and water to remove residual alumina powder on the electrode surface. Then, the electrodes were electrochemically scanned in 0.5 M H 2 SO 4 solution, with a potential window ranging from−0.3 to +1.5 V at 100 mV/s for 25 cycles. Finally, the electrodes were rinsed with water and dried with nitrogen. The immobilization of probe DNA on the electrode was performed by immersing into 50 µl of 10 mM Tris-HCl (pH 7.4, 0.1 M NaCl, and 10 mM TCEP) buffer containing 1 µM probe DNA for 4 h at room temperature. The electrode was then rinsed with 20 mM Tris-HCl (pH 7.4, 0.1 M NaCl) buffer solution. After DNA immobilization, the electrode was immersed into 50 µl of MCH solution (1 mM) for 30 min to remove some non-specifically adsorbed DNA strands.

Target DNA Recognition, Exonuclease I (Exo I) Cleavage, and Binding With ZrO 2 -rGO-Thi Nanocomposite
Target DNA recognition was operated by immersing the probe DNA modified electrode into 50 µl of 10 mM Tris-HCl (pH 7.4, 0.1 M NaCl, and 1 mM MgCl 2 ) buffer containing different concentrations of target DNA for 50 min at 37 • C. Then, the electrodes were simply rinsed with 20 mM Tris-HCl (pH 7.4, 0.1 M NaCl) buffer solution. After the target DNA recognition, the unhybridized probe DNA was removed by Exo I, which was operated in 1 × exonuclease I reaction buffer (pH 9.5, 67 mM glycine-KOH, 6.7 mM MgCl 2 , and 10 mM 2-thioethanol) containing 5 U of Exo I at 37 • C for 60 min. After that, the electrode was thoroughly rinsed with 20 mM Tris-HCl (pH 7.4, 0.1 M NaCl) buffer solution. Then, the electrode was recognized by 0.5 mg/ml ZrO 2 -rGO-Thi nanocomposites for 60 min at room temperature. After the recognition of ZrO 2 -rGO-Thi complex with the electrodes, the electrodes were thoroughly rinsed with 20 mM Tris-HCl (pH 7.4, 0.1 M NaCl) buffer solution to remove some possible non-specifically bounded ZrO 2 -rGO-Thi nanocomposites. Then, the obtained electrode was interrogated for electrochemical measurement.

Electrochemical Measurements
After binding with the ZrO 2 -rGO-Thi nanocomposite, differential pulse voltammetry (DPV) and cyclic voltammetry (CV) measurements were performed in 10 mM phosphatebuffered saline (PBS; pH 7.4, 0.2 M KNO 3 ). The DPV was operated under the following experimental parameters: scanning potential from 0.1 to−0.4 V, amplitude of 50 mV, pulse period of 0.1 s, and sampling width of 16.7 ms. The CV was recorded between 0.2 and−0.6 V, with a scan rate of 100 mV s −1 . Electrochemical impedance spectroscopy (EIS) was used for biosensor fabrication process characterizations, which was carried out in 10 mM PBS (pH 7.4, 5 mM [Fe(CN) 6 ] 3−/4− and 1 M KCl), with a scan frequency ranging from 0.1 to 10 kHz and an AC amplitude of 5 mV. Corresponding CV characterizations were also conducted by scanning the potential between−0.1 and 0.6 V at 100 mV s −1 . Prior to the electrochemical measurement, the electrolyte solution should be purged with high-purity nitrogen for 20 min to avoid interference from oxygen.

Apparatus
All electrochemical experiments were performed with a CHI 660E electrochemical workstation (CH Instruments, Shanghai, China) at room temperature by using a three-electrode system consisting of a modified gold electrode, a platinum wire auxiliary electrode, and a Ag/AgCl reference electrode. The morphology of the nanocomposite was characterized by a field emission scanning electron microscope (SEM, JSM 7500F, JEOL, Japan) and a transmission electron microscope (TEM, JEM-F200, Japan). Gel images were captured on a FR-980A gel image analysis system (Shanghai, China).

Detection Principle of the Developed Nucleic Acid Biosensor
The fabrication principle of the electrochemical nucleic acid biosensor by using ZrO 2 -rGO-Thi nanocomposites is illustrated in Scheme 1. The zirconia-reduced graphene oxide (ZrO 2 -rGO) nanocomposites were firstly prepared by a hydrothermal synthesis method, with the use of GO and ZrOCl 2 ·8H 2 O as reactants. After ZrOCl 2 hydrolysis and Zr 4+ adsorption onto the GO surface, the ZrO 2 nanoparticles were obtained under hydrothermal conditions. Simultaneously, GO was reduced into rGO to obtain ZrO 2 -rGO nanocomposites. Then, thionine, as a well-characterized electrochemical probe, was conjugated onto the rGO surface via π-π interaction to obtain ZrO 2 -rGO-Thi nanocomposite (Li et al., 2012;Zhou et al., 2016). The obtained ZrO 2 -rGO-Thi nanocomposite was used as an integral sensing element for DNA recognition, signal amplification, and reporting. The ZrO 2 nanoparticles could recognize DNA via its multicoordinative interaction with the phosphate group of the DNA skeleton. The multicoordinative interaction of ZrO 2 with the phosphate group has been widely employed for the enrichment and the analysis of phosphorylated substrates (Monot et al., 2008;Pang et al., 2011). A large amount of thionine molecules adsorbed onto the rGO surface could provide an amplified signal response toward the recognized DNA information. Since the ZrO 2 -rGO-Thi nanocomposite may interact with both the single-stranded DNA and the doublestranded DNA, thus a simple Exo I-assisted cleavage was adopted to achieve DNA discrimination and background reduction. The probe DNA was modified with a sulfhydryl group at the 5 ′ -terminus. The sulfhydryl group could form an Au-S bond with the gold electrode to achieve probe DNA immobilization.
In the presence of the target DNA, it can hybridize with the immobilized probe DNA on the electrode. The DNA hybrids formed could resist the cleavage from Exo I, owing to the formation of the blunt 3 ′ -terminus. However, the unhybridized probe DNA would be digested from 3 ′ to 5 ′ by Exo I. Then, the remaining hybridized dsDNA onto the electrode could bind with the ZrO 2 -rGO-Thi nanocomposite via the multi-coordination between ZrO 2 and the phosphate group on the DNA skeleton, generating an amplified electrochemical response related with the target DNA recognition event. In the absence of the target DNA, the immobilized probe DNA would be digested by Exo I, and then the ZrO 2 -rGO-Thi nanocomposite could not be effectively captured onto the electrode for an electrochemical response.

Characterization of ZrO 2 -rGO-Thi Nanocomposites
The morphology of the GO and ZrO 2 -rGO nanocomposites was characterized by SEM and TEM. GO exhibited a typical folded morphology at the edges, confirming the layered structure of GO (Figures 1A,D). After the hydrothermal process for the mixed GO and ZrOCl 2 ·8H 2 O (1:1), it could be clearly seen that the ZrO 2 nanoparticles were uniformly grown on the rGO surface (Figures 1B,E). The interplanar spacing of a single crystalline ZrO 2 nanoparticle was measured as about 0.294 nm, suggesting a monoclinic phase of the ZrO 2 nanoparticles (inset in Figure 1E). With the increase of mass ratio between GO and ZrOCl 2 ·8H 2 O (1:5), the density of the ZrO 2 nanoparticles onto the rGO surface was significantly increased (Figure 1C). SCHEME 1 | Schematic illustration of electrochemical nucleic acid fabrication by ZrO 2 -rGO-Thi nanocomposite.
Frontiers in Chemistry | www.frontiersin.org It could be further seen from the elemental mapping images of ZrO 2 -rGO ( Figure 1F) that C, O, and Zr elements were uniformly distributed throughout the nanocomposite. More detailed experimental characterizations toward the ZrO 2 -rGO and the ZrO 2 -rGO-Thi nanocomposites could be referred to our previous work .

Electrochemical Characterization of Fabricated DNA Biosensor
The EIS was firstly used to follow the nucleic acid biosensor fabrication process by using 5  , suggesting the resistance of the DNA hybrids with the blunt 3 ′end against the digestion from Exo I. Then, after the ZrO 2 -rGO-Thi nanocomposites' binding onto the DNA modified electrode, the Rct value was decreased to 3,212 (curve f). It was speculated that some of the bound ZrO 2 -rGO-Thi nanocomposites may be close enough to the electrode surface for the enhanced electron transfer activity. In the absence of the target DNA, after the treatment of the probe DNA modified electrode by Exo I, the Rct value was evidently decreased to 2,500 (curve g) compared with that of the probe DNA and the MCH modified electrode, suggesting the effective digestion of unhybridized probe DNA by Exo I for background reduction. The corresponding cyclic voltammetric characterizations for the differently modified electrodes are shown in Figure 2B. After probe DNA immobilization (curve b), MCH assembly (curve c), and hybridization with the target DNA (curve d), the redox peak current of [Fe(CN) 6 ] 3−/4− decreased and the peak-to-peak potential increased sequentially compared with that The cyclic voltammetric responses for the differently assembled electrodes were basically in accordance with the EIS results, indicating the successful fabrication of an electrochemical nucleic acid biosensor.
We then conducted the detection feasibility of the fabricated biosensor toward the target DNA by using DPV and CV measurements based on the electrochemical responses of the ZrO 2 -rGO-Thi nanocomposites. The results are shown in Figures 2C,D. It could be seen that, in the presence of 100 nM target DNA, a distinct DPV response of thionine was observed, at a reduction potential of−0.185 V, for the hybridized electrode after Exo I treatment (curve d in Figure 2C). Also, a pair of large redox peaks of thionine, with oxidation and reduction potentials of−0.14 and−0.18 mV, respectively, was obtained in the CV results (curve d in Figure 2D). Thus, the ZrO 2 -rGO-Thi nanocomposites have been effectively combined with the hybridized DNA for an electrochemical response. For the probe DNA modified electrode with no Exo I treatment, an obvious electrochemical response could also be observed (curve c in Figures 2C,D). Thus, the ZrO 2 -rGO-Thi nanocomposites could also be conjugated with the probe DNA modified electrode via the multicoordinative interaction for an electrochemical response. However, after Exo I treatment toward the probe DNA modified electrode, only a weak background response could be obtained (curve b in Figures 2C,D). Such a background response was only slightly higher than that of the MCH modified electrode (curve a in Figures 2C,D). Furthermore, an electrochemical response comparison by using Exo I cleavage or not is shown in Figure 2E. The signal-to-background ratio in the presence of Exo I was about 7, but it was only about 2.14 in the absence of Exo I for the detection of 100 nM target DNA. Thus, Exo I could digest the unhybridized probe DNA for background reduction, and the DNA hybrids could resist the cleavage against Exo I and be captured by the ZrO 2 -rGO-Thi nanocomposite for an amplified electrochemical response related with target DNA recognition information. Exo I cleavage toward the probe DNA, but not DNA hybrids, was also confirmed by gel electrophoresis experiments. It could be seen from Figure 2F that the probe DNA (lane 2) and the target DNA (lane 3) showed the corresponding migration bands. The hybridization of the probe DNA and the target DNA showed the band with a lower migration rate (lane 5). After the Exo I treatment, the band related with the probe DNA disappeared (lane 4), suggesting the cleavage of probe DNA by Exo I, but the DNA hybrids with the blunt 3 ′ -end could not be digested by Exo I, and the band related with the DNA hybrids could be still clearly observed (lane 6).

Optimization of Experimental Conditions
In order to demonstrate the best detection capability of the developed biosensor, some critical experimental parameters were optimized, including the immobilization concentration of probe DNA, reaction time, reaction temperature, and amount of Exo I. A proper immobilization concentration of probe DNA would benefit for DNA hybridization recognition and then for electrochemical response. It could be seen from Figure 3A that the DPV current change I ( I = I -I 0 , I, and I 0 represent the current response in the presence and in the absence of the target DNA, respectively), increased with increasing probe DNA immobilization concentrations. It could be easily explained that a higher concentration of probe DNA would increase the assembly amount of probe DNA on the electrode to generate more amounts of DNA hybrids on the electrode for a larger electrochemical response. However, with further increasing probe DNA concentration over 1 µM, the I decreased gradually. Since the background response in the absence of the target DNA changed slightly at different probe DNA concentrations, the decreased I value obtained at a higher probe DNA concentration may be caused by the decreased recognition efficiency at a higher probe DNA concentration or coverage. Thus, 1 µM was selected as the optimized value for the immobilization concentration of the probe DNA. The surface density of the probe DNA on the electrode was further calculated by the chronocoulometry method with the use of Ru(NH 3 ) 3+ 6 as a redox indicator (Steel et al., 1998). It was 3.27 ± 0.18 × 10 12 molecules/cm 2 at an immobilization probe concentration of 1 µM. The hybridization time was also optimized and is shown in Figure 3B. It could be seen that the current response toward 100 nM target DNA increased stepwise with hybridization time extension and almost reached the saturation value at 50 min. Also, in the absence of the target DNA, no evident current change was observed at the studied hybridization time range. Thus, the hybridization time of 50 min was recommended for performance determination of the DNA biosensor. The reaction temperatures for Exo I cleavage and Exo I dosage were also optimized and are shown in Figures 3C,D. It could be seen that the best electrochemical response could be obtained at the reaction temperature of 37 • C for Exo I cleavage. It could be explained that 37 • C was the suitable temperature to maintain Exo I activity. An elevated temperature would be not beneficial for the stability of the DNA hybrids between the target DNA and the probe DNA, inducing a decreased electrochemical response toward the target DNA. The electrochemical response toward the target DNA increased with the employed Exo I amount, and a maximum value could be reached at an amount of 5 U ( Figure 3D). A further increase in Exo I amount would induce a slightly weakened electrochemical response, which might be due to the non-specific cleavage of Exo I toward some hybridized DNA for signal response decrease. Thus, the optimized reaction temperature and the Exo I amount chosen were 37 • C and 5 U, respectively.

Detection Performance of the Developed Electrochemical DNA Biosensor
The analytical capability of the constructed electrochemical biosensor was investigated by using different concentrations of the target DNA under optimal experimental conditions. As shown in Figure 4A, the electrochemical response signal increased gradually with increasing concentrations of the target DNA, suggesting a concentration-dependent response behavior. A weak background response in the absence of the target DNA could be observed, which might be explained as follows: the immobilized probe DNA may be not completely digested by Exo I, inducing the adsorption of a small amount of ZrO 2 -rGO-Thi  nanocomposites for electrochemical response; also, the ZrO 2 -rGO-Thi nanocomposites may be non-specifically adsorbed onto the electrode for electrochemical responses. A linear relationship between the DPV current and the logarithm value of the target DNA concentration, ranging from 100 fM to 10 nM, could be obtained. The corresponding regression equation is expressed as Y = −0.1909 + 0.2674 logX (Y and X represent the DPV peak current and the target DNA concentration, respectively), with a correlation coefficient of 0.9943. The detection limit toward the target DNA was calculated as about 24 fM, according to the classic method of LOD = 3β / k (β represents the blank response and k is the slope of the calibration curve). Such a detection limit was superior or comparable with some reported methods ( Table 2), but it only needs a simple Exo I cleavage for background reduction and eliminates the complex DNA designs or too much DNA fragments used by some reported DNA-based signal amplification strategies. The detection reproducibility of the fabricated DNA biosensor toward the target DNA was checked. The relative standard deviations for the detection of 1 pM, 100 pM, and 100 nM were obtained as 6.9, 6.3, and 5.8%, respectively, based on five replicated experiments by different electrodes, suggesting an acceptable detection reproducibility of the fabricated biosensor for the target DNA.

Selectivity and Stability of Fabricated Electrochemical DNA Biosensor
The selectivity of the proposed electrochemical biosensor was evaluated by using different DNA sequences, including fully complementary target DNA, single-base mismatched target (1MT), two-base mismatched target (2MT), and noncomplementary DNA (NC) with the same concentration. As can be seen in Figure 5A, the DPV current response of NC was almost the same with the blank solution. The mismatched DNA showed the decreased electrochemical responses compared with the complementary target DNA, owing to the decreased hybridization stability with the probe DNA. The DPV responses for 1MT and 2MT were about 40 and 13.5% of that for the fully complementary target DNA, respectively, demonstrating the potential for base mutation analysis. The stability of the probe DNA modified electrode was then investigated. After its storage at 4 • C for 14 days, the electrochemical response toward 100 nM target DNA could still remain over 90% of its original response, suggesting the robust stability of the fabricated biosensor. To demonstrate the applicative potential of the fabricated DNA biosensor in the relatively complex biological samples, detections of target DNA spiked in 10 and 25% diluted fetal calf serum and 10% diluted human serum were conducted. It could be seen in Figure 5B that the diluted serum could exert some influences on the background and the signal responses, but the electrochemical responses were still dependent on the spiked target DNA concentration in the diluted serum. The background responses in the diluted serum are slightly larger than that in the buffer. It might be due to the non-specific adsorption of some biological molecules in the serum on the electrode, further inducing the possible adsorption of the ZrO 2 -rGO-Thi nanocomposite for increased background response. The signal response in the diluted serum was mostly lower than that in the buffer. It might be explained that the non-specific adsorbed biological molecules on the electrode influenced the DNA hybridization recognition to some extent for the observed signal decrease. These results suggested an applicative potential of our fabricated DNA biosensor in relatively complex biological matrices.

CONCLUSIONS
In conclusion, a simple electrochemical biosensor was developed for sensitive DNA detection by coupling the ZrO 2 -rGO-Thi nanocomposite and Exo I-assisted cleavage. The ZrO 2 -rGO-Thi nanocomposite was explored as an integral sensing nanomaterial, including DNA recognition (multi-coordinative interaction of ZrO 2 with the phosphate group of the DNA skeleton), signal amplification and reporting (abundant thionines onto the rGO surface for direct electrochemical response), and excellent conductivity of rGO. With the aid of Exo I-based cleavage for background reduction, the sensitive and the selective detection of target DNA related with the P53 gene could be achieved with a wide linear detection range (100 fM to 10 nM) and a low detection limit of 24 fM. Although the detection limit by the current biosensor is still higher compared with those most sensitive methods reported, it does not involve complex DNA designs or operations or too much DNA fragments for signal amplification. It can also be applied for DNA detection in a relatively complex biological matrix such as serum. Therefore, it offers a simple but effective biosensor fabrication strategy and is anticipated to show potential for applications in bioanalysis and medical diagnosis.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/supplementary material.

AUTHOR CONTRIBUTIONS
ZC and SL: designed the work and wrote the manuscript. ZC, XL, DL, FL, and LW: carried out the experiments. ZC and DL: performed the statistical analysis. FL and SL: revised and edited the manuscript. All authors reviewed the manuscript and have agreed to its publication.