Ti₃AlC₂ MAX was acquired from Aladdin (Shanghai, China). Tris (hydroxymethyl)aminomethane (Tris), and 6-Mercapto-1-hexanol (MCH) were acquired from Sigma. Tris (2-carboxyethyl) phosphine hydrochloride (TCEP) and chloroauric acid (HAuCl₄·4H₂O, ≥99.9%) were ordered from J&K scientific. Hydrofluoric acid was acquired from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). 0.1 M pH 7.4 Tris-HCl buffer (100 mM NaCl and 20 mM MgCl₂) was used for electrochemical measurement, and 10 mM pH 7.4 Tris-HCl buffer was used as washing buffer. All DNA strands were ordered from Sangon Biotech Co., Ltd. (Shanghai, China), and the DNA sequences were provided in Table 1. 18.2 MΩ cm ultrapure water was used in whole experiments.

Scanning electron microscopy (SEM) images and elemental mapping were characterized by SU8220 field-emission scanning electron microscope (Hitachi Ltd, Japan, 10.0 kV). X-ray diffraction (XRD) was characterized by X’pert Powder X-ray diffraction (Panalytical. Ltd, Netherlands). Transmission electron microscope (TEM) images were recorded on a JEM-2100F field emission electron microscope (JEOL, Japan) at an accelerating voltage of 200 kV.

Ti₃C₂ MXenes were prepared by etching the Al element in Ti₃AlC₂ MAX with HF solution referring to the previous literature with minor modifications (Lin et al., 2020) (Scheme 1A). Briefly, 1 g Ti₃AlC₂ was added to 50 mL 45% HF solution, and then etched at room temperature for 24 h. The precipitate was collected by centrifugation at 6,000 rpm for 15 min. Finally, the precipitate was washed several times with deionized water until the pH of the supernatant exceeded 6. The Ti₃C₂ MXenes were obtained.

AuNPs@Ti₃C₂ MXenes nanocomposites were synthesized referring to the previous literature with minor modifications (Mi et al., 2021; Song et al., 2022) (Scheme 1A). AuNPs were synthesized by in situ reduction on the surface of Ti₃C₂ MXenes and Ti₃C₂ MXenes were used as the reducing agent and support material. Firstly, 15 mg pre-prepared Ti₃C₂ MXenes were well-dispersed in 45 mL deionized water, and the Ti₃C₂ suspension (0.33 mg mL⁻¹) was prepared after sonicating for 4 h. Then, 600 μL of 1 g mL⁻¹ HAuCl₄ solution was slowly added to the Ti₃C₂ suspension drop by drop with gentle stirring. After the reaction at room temperature for 30 min, the resulting suspension was centrifuged at 8,000 rpm for 10 min and the precipitate was collected. The AuNPs@Ti₃C₂ two-dimensional nanocomposites were successfully prepared by washing with deionized water several times for subsequent studies.

The bare glassy carbon electrode (GCE) was polished with 0.05 μm alumina slurry for 5 min before modification and then sonicated in ethanol and deionized water for 2 min, respectively. Finally, the bare GCE was rinsed with deionized water and then dried with nitrogen gas for the following experiment. Subsequently, 1 mg mL⁻¹ AuNPs@Ti₃C₂ MXenes nanocomposites were dropped onto the GCE surface to obtain the AuNPs@Ti₃C₂ MXenes/GCE and dried naturally. 5 μM template DNA, 5 μM protected DNA and 5 μM auxiliary DNA were incubated at 90°C for 5 min, and then gradually cooled to room temperature to form stable DNA duplex structures. Then, a final concentration of 10 mM of TCEP was added to the DNA duplex for 1 hour at room temperature to break the disulfide bond. Subsequently, drop 10 μL 1 μM DNA double-strand probes onto the modified electrode AuNPs@Ti₃C₂ MXenes/GCE and incubate at room temperature for 3 h to form self-assembled monolayers. The electrodes were immersed in 1 mM MCH to block the unbound sites, then washed with Tris-HCl buffer and dried with nitrogen subsequently.

The electrodes of the modified DNA duplex probes were washed with buffer and then immersed in a solution containing 50 μL¹ μM probe DNA and various concentrations of target DNA KARS G12D. The TMSD recycling process took 2 hours to complete at room temperature. Subsequently, the electrodes were washed with 10 mM Tris-HCl buffer and were used for electrochemical measurement.

A CHI660E electrochemical workstation (Shanghai Chenhua Instrument Co. Ltd., China) was used for all electrochemical measurements which contained a three-electrode system consisting of a modified 3 mm GCE, a platinum auxiliary electrode, and a saturated Ag/AgCl reference electrode. Square wave voltammetry (SWV) experiments were carried out in 10 mM Tris–HCl buffer (100 mM NaCl and 20 mM MgCl₂, pH 7.4) using the following parameters: a potential range of −0.6 to 0 V, an amplitude of 25 mV, a frequency of 25 Hz, and a step potential of 4 mV. To prevent the interference of oxygen reduction during the electrochemical measurement, high-grade nitrogen should be used to purge the detection buffer for 30 min. Electrochemical impedance spectroscopy (EIS) was performed in 5 mM [Fe (CN)₆]^3−/4− contained 0.1 M KCl, in frequency range from 0.03 Hz to 10 kHz with the bias potential of 0.192 V and the amplitude of 5 mV, respectively.

The morphologies of Ti₃AlC_2, Ti₃C_2, and AuNPs@Ti₃C₂ MXenes were characterized by SEM and TEM. Ti₃AlC₂ without etching treatment exhibited an irregular morphology (Figure 1A). The Ti₃C₂ MXenes prepared by etching and ultrasound show a smooth, thin, and flake structure (Figure 1B) due to the disappearance of the Al layer after the HF etching. Figure 1C shows that the size of AuNPs synthesized by in situ reduction on the surface of Ti₃C₂ is about 125 nm. Because of the large surface area of Ti₃C₂, a lot of AuNPs were synthesized. By measuring the elemental properties of AuNPs@Ti₃C₂ MXenes, the results showed that the Ti, C, and Au elements of AuNPs@Ti₃C₂ MXenes were uniformly distributed and continuously (Figure 1E).

The XRD characterization further determined the composition and crystal structure of the nanocomposites (Figure 1D). The diffraction peaks of Ti₃AlC₂ at 9.5°, 19.06°, 38.7°, and 41.84° matched well with the diffraction peaks of the standard card (JCPDS card number: 52-0875) (curve a in Figure 1D). The XRD results of the Ti₃C₂ showed the disappearance of the 38.7° diffraction peak in the Al (104) plane and the left shift of the diffraction peak in the (002) plane of Ti₃C₂ from 9.5° to 8.27°, which indicated the successful preparation of Ti₃C₂ (curve b in Figure 1D) (Ran et al., 2017; Alhabeb et al., 2018; Meng et al., 2021). When AuNPs are generated by in situ reduction on the surface of Ti₃C₂, curve c in Figure 1D shows that AuNPs@Ti₃C₂ have the characteristic diffraction peaks of both Ti₃C₂ and AuNPs (Yang et al., 2022). The SEM images, TEM images, mapping, and XRD results indicated that the two-dimensional nanocomposite AuNPs@Ti₃C₂ MXenes were successfully synthesized.

Toehold-mediated strand displacement reaction was used in the development of many biosensors due to its high specificity and without the involvement of enzymes. In the toehold-mediated strand displacement reaction, one oligonucleotide hybridizes with the toehold domain of double-strand DNA resulting in the dissociation of the substrate strand from the double-strand DNA (Irmisch et al., 2020; Li et al., 2023). In this study, we developed two efficient and simple toehold-mediated strand displacement reactions (Scheme 1B). The first toehold-mediated strand displacement reaction was performed when target DNA KARS G12D hybridizes with the double-strand DNA probes. The second toehold-mediated strand displacement reaction was performed when the signal probes were present in the reaction system, and the dissociated target DNA KARS G12D entered the next cycle of reaction. The fabrication of the DNA biosensor based on toehold-mediated strand displacement reaction was illustrated in Scheme 1. First, the pre-prepared AuNPs@Ti₃C₂ MXenes were modified on the GCE. Then double-strand DNA probes were self-assembled onto the AuNPs surface via Au-S bonding. The toehold of strand displacement reaction was formed. Subsequently, the target DNA KARS G12D and the signal probe were added for strand displacement reactions to achieve circular hybridization of KARS G12D. Meanwhile, the signal probe was linked to the electrode by the second strand displacement reaction to generate an electrochemical signal.

The EIS measurement results were shown in Figure 2. The bare glassy carbon electrode shows a small Ret value (curve a). When the bare glassy carbon electrode was modified with AuNPs@Ti₃C₂ MXenes composites, the Ret value became smaller (curve b), indicating that AuNPs@Ti₃C₂ MXenes composites have excellent electrical conductivity. The Ret value gradually increased when the DNA duplex structure and MCH were gradually modified, which was caused by the low conductivity of the DNA duplex structure and MCH (curves c and d). When the TMSD occurred after adding probe DNA and KARS G12D, the Ret value became smaller, probably due to the reduction of a steric hindrance after the hybridization of probe DNA and template DNA (curve e). The results of EIS and PAGE (Supplementary Figure S1) demonstrate the successful modification of AuNPs@Ti₃C₂ MXenes composites on glassy carbon electrodes and the successful design of TMSD.

We thus used the sensor to detect a range of KARS G12D at concentrations from 10⁻⁵ nM–10 nM. As shown in Figure 3A, as the concentration of KARS G12D increased, the electrochemical signal values of methylene blue progressively increased, showing a typical concentration-dependent event. Further quantitative analysis exhibited a logarithmic relationship between the electrochemical signal values of methylene blue and the concentration of KARS G12D (Figure 3B). The calibrated regression equation was fitted to y (nA) = 110.6 log C_tDNA + 583.4 (R² = 0.9946), with a detection limit of 0.38 fM (S/N = 3). Such a low detection limit was equivalent to many enzyme-free sensors or even those sensors using enzymes. The constructed DNA sensor has high sensitivity and a wide linear range (Supplementary Table S1).

Since the sensor has good detection performance, we further test the specificity of the sensor. Point mutations increase the difficulty of gene detection because they cause changes in the characteristics of genes. Therefore, the ability to identify point mutations is an important parameter for gene detection technology (Chang et al., 2015; Yu et al., 2022). We verified the specificity of the DNA sensor using five types of DNA sequences, including target DNA KARS G12D (T), single-base mismatched DNA (1 M), double-base mismatched DNA (2 M), triple-base mismatched DNA (3 M) and random sequence DNA (R) (Figure 4). A comparison of the oxidation peak current values of methylene blue revealed the current value of the 1 M group was much smaller than the T group, only 57% of the T group, revealing the ability of the sensor to recognize single base mismatches effectively. The oxidation peak current values of 2 M and 3 M were only 41% and 17% of the target DNA, while the peak current values of the random sequence DNA were almost close to the background values. To further ensure the performance of the sensor, the reproducibility and stability were also tested. The relative standard deviation (RSD) of the current signal values for the four parallel electrodes was 3.59%. In addition to this, the current signal value of the treated electrodes after 7 days at 4°C was 96.4% of the original value. The above results indicate that the sensor has good specificity, reproducibility, stability, and has good potential for clinical applications.

The sensor has good sensitivity, specificity, and stability. We further evaluated its ability to be used in clinical practice with real samples. Different concentrations of KARS G12D were added to a 10-fold dilution of healthy human serum for recovery testing. The recoveries of different concentrations of KARS G12D were 109%, 100.1%, and 93.3%, respectively (Table 2). The results showed that the sensor has good potential for monitoring ctDNAs in complicated biological samples.