▎ 摘　　要

For the first time, we have successfully synthesized stable graphene nanosheets from graphite powder through sonication in the hemoglobin-capped gold nanoclusters (Hb@AuNCs) solution for biosensing application. This approach, as a simple method for the exfoliation and fragmentation of graphite in a nanocluster solution, enabled us to produce stable aqueous graphene dispersions at low cost and without the need for hazardous chemicals or tedious experimental procedures. In this method, Hb@AuNCs were used not only as stabilizing agent of graphene through non-covalent bonding, but also as dispersing agent of few-layer graphene nanosheets. The Hb@AuNCs stabilized graphene (Hb@AuNCs-G) was characterized by high resolution transmission electron microscopy (HRTEM), zeta-sizer and Raman spectroscopy. Then, the graphene nanosheets were applied as a novel versatile electrochemical platform for ultrasensitive biosensing of short DNA species of chronic myelogenous leukemia (CML) based on the "signal off" and "signal on" strategies. For this purpose, a single strand DNA (ssDNA) was immobilized on the Hb@AuNCs-G/AuNPs modified electrode surface and acted as the biorecognition element. Methylene blue (MB), as the signaling probe, was then intercalated into the ssDNA. The intercalated MB was liberated upon interaction with the synthetic complementary DNA (cDNA, target), thereby resulting in the apparent reduction of MB redox signal. This designed "signal off" sensing system enabled the voltammetric determination of the target cDNA over a dynamic linear range (DLR) of 0.1 fM to 10 pM with a limit of detection (LOD) of 0.037 fM. In the "signal on" strategy, the response to the cDNA was detected by monitoring the change in the electron transfer resistance (R-ct) using the ferro/ferricyanide system as a redox probe. The charge transfer resistance of the probe was found to increase linearly with increasing concentration of target cDNA in the range of 0.1 fM-10 pM with a limit of detection of 0.030 fM. Finally, the selectivity and feasibility of genosensor was evaluated by the analysis of derived nucleotides from mismatched sequences and the clinical samples of patients with leukemia as real samples, respectively.