Ultrasensitive fluorescent aptasensor for MUC1 detection based on deoxyribonuclease I-aided target recycling signal amplification

A novel sensing strategy for sensitive detection of mucin 1 protein (MUC1) based on deoxyribonuclease I-aided target recycling signal amplification was proposed. In this paper, in the absence of MUC1, the MUC1 aptamer is absorbed on the surface of graphene oxide (GO) via π-stacking interactions. This results in quenching of the fluorescent label and no fluorescence signal is observed. Upon adding MUC1, the probe sequences could be specifically recognized by MUC1, leading to an increase in the fluorescence intensity. The detection limit is as low as 10 pg mL−1, and a linear range from 50 pg mL−1 to 100 ng mL−1. The assay is specific and sensitive, and successfully applied to the determination of MUC1 in spiked human serum, urine and saliva. Importantly, the proposed aptasensing strategy has great potential in detecting various protein and even cancer cells.


Introduction
At present, cancer is one of the most life-threatening diseases in the worldwide. 1 The latest WHO gures show that 8.8 million people worldwide die from cancer every year, accounting for nearly 1/6 of deaths a year, and the gure is expected to increase to more than 21 million by 2030. Thus, it is important to develope effective methods for cancer therapy. Among them, the early identication and quantitative analysis of tumor biomarkers could provide a simpler and more effective method for monitoring tumor progression. It is of particular importance in early clinical diagnosis and treatment of cancer.
The mucin1 protein (MUC1), is a cell-surface associated glycoprotein of the mucin family, 2 which contains a hydrophobic membrane-spanning domain of 31 amino acids, a cytoplasmic domain of 69 amino acids, and an extracellular domain consisting of a region of nearly identical repeats of 20 amino acids per repeat. 3,4 MUC1 presents in a variety of malignant tumors and is over-expressed in almost all human epithelial, including breast, gastric, colorectal, lung, prostate, ovarian, pancreatic and bladder carcinomas; 5 but, low levels or no expression in normal healthy tissues in normal cells. It is considered as a potential cancer biomarker. Thus, an sensitive detection of MUC1 is great importance in preliminary diagnosis of cancer. To date, several techniques including ELISA, 6 surface plasmon resonance spectroscopy (SPR), 7 colorimetric, 8 electrochemistry, [9][10][11] photoelectrochemical, 12 electrochemical biosensors, 13-16 electrochemiluminescence [17][18][19] and uorescence biosensors [20][21][22][23] have been applied in the detection of MUC1. Among them, uorescence assay is a well-known detection method with acknowledged advantages such as high sensitivity, easy readout, simplicity and the feasibility of quantication. 22,24 Despite many advances in this eld, these strategies still have shortcomings including the costly modication of uorophores and potential false positive signals. Moreover, the sensitivity needs to be further improved. Therefore, it's important to develop an efficient and highly sensitive uorescent strategy to detect MUC1 in the fundamental research and practical applications.
In recent years, there have some reports of mucin 1 glycoprotein based on MUC-1 aptamer. Aptamers are nucleic acid ligands (DNA or RNA) that can specically bind protein, which are articially synthesized and selected in vitro through exponential enrichment (SELEX). 25,26 In comparison with other targeting agents, aptamers possess unique advantages such as small size, design exibility, stability in harsh biological environments, easy for chemical modication as well as high specicity and high purity. 27,28 These special characteristics will enable them be widely used biomarker discovery and biosensors. 29,30 To date, a series of aptamer-based biosensors have been applied to detect VEGF 165 , 31 EpCAM, 32 Salmonella paratyphi A, 33 cocaine 34 and so on. There have also been reported for detecting MUC1 using electrochemical, 35 photoelectrochemical, 12 colorimetric 8 and uorescent. 22 But, among them, no standard method has been established. Therefore, a novel, highly sensitive and selective approaches for MUC1 detection need to be designed.
Graphene oxide (GO), a one-atom-thick two-dimensional carbon nanomaterial with a honeycomb structure. It's owing to unique features such as high mechanic strength, good water dispersibility and facile surface modication. 36,37 GO has a high specic surface are a for binding plenty of molecules, such as nucleic acids, proteins, liposomes and pharmaceutical molecules. [38][39][40] Moreover, GO is capable of quenching uorescence with high efficiency, and can be used to construct GObased sensors based on the principle of uorescence resonance energy transfer (FRET). 41 More importantly, GO can improve their stability in the biological environments, due to its protection capacity for biomolecules binded to its surface from enzymatic cleavage. 42 Consequently, GO, as a superb nanomaterial, would be easily used in biosensors for biochemical analysis. On the other hand, to overcome these limitations, signal amplication strategies, enzyme-aided signal amplication have recently been developed to achieve the sensitive detection of biomolecules, including endonuclease, deoxyribonuclease I (DNase I), exdonuclease and DNA polymerase. 43 Among them, DNase I, a common endonuclease, can digest single-and double-stranded DNA (dsDNA) into oligo-and mononucleotides. DNase I can cleave phosphodiester bonds to product many polynucleotides with 5 0phosphate and 3 0 -OH groups, but cannot digest RNA sequences. Thus, these makes it suitable for RNA analysis based on DNase I-assisted target recycling and signal ampli-cation. 44 Based on these ndings, He et al., described a approach nanographite-based uorescent biosensor and DNase I for the amplied detection of Salmonella enteritidis, 43 and Yan et al., described a highly sensitive uorescent aptasensor for Salmonella paratyphi A via DNase I-mediated cyclic signal amplication 33 and so on. DNase I-aided signal ampli-cation has the advantages of sensitive, reliable and specic. To meet the demands of the specicity and sensitivity of MUC1 detection, the development of an DNase I-aided signal amplication strategy is extremely urgent.
Hence, in this work, a very sensitive and specic aptasensor was designed to detect MUC1 based on DNase I-aided target recycling signal amplication. The MUC1 aptamer 5 0 terminal was labeled with an FAM that amplies uorescence signals by cleaving the aptamer randomly. In the absence of MUC1, the MUC1 aptamer is strongly adsorbed onto GO, results in quenching of the uorescent label and no uorescence signal. Upon adding MUC1, it can recognize aptamer specically and form aptamer/MUC1 complexes, resulting in the aptamer being away from GO. And then, when DNase I is introduced, enzymatic digestion reaction that releases the MUC1 for the next round occurs, leads to the accumulation a large amount of FAM uorophores. The quantities of MUC1 can be achieved by uorescence increment. Thus, a novel, sensitive, and specic aptasensor is obtained for assaying MUC1. Furthermore, it's successfully applied for detection of MUC1 in various spiked biological samples. One may believe that this sensing system possesses great potential for the monitoring of proteins, clinical diagnosis, and treatment.

Reagents and materials
The MUC1, EpCAM, BSA, PSA and VEGF were purchased from Cusabio Biotech Co. Ltd. (http://www.cusabio.cn/). Graphene oxide was purchased from XFNANO Co. Ltd. (Nanjing, China, http://www.xfnano.com). The deoxyribonuclease I (DNase I) were obtained from solarbio (Beijing, China, http:// www.solarbio.com). The MUC1 (5 0 -FAM-CCCGTCTTCCAGA-CAAGAGTGCAGGG-3 0 ) aptamer were synthesized by Sangon Biological Engineering Technology Co., Ltd. (Shanghai, China, http://www.sangon.com) and puried using high performance liquid chromatography. All of the reagents were diluted to the required concentration with working buffer (10 mM Tris, 100 mM NaNO 3 pH 7.4) before usage. Healthy human serum, urine and saliva were provided by Affiliated Dongfeng Hospital, Hubei University of Medicine, and approved by Hospital's Ethics Committee. The other reagents employed were of analytical grade and used without further purication. Ultrapure water obtained from a Millipore water purication system (18.2 MU cm resistivity, Milli-Q Direct 8) was used in all runs.

Optimization of sensing conditions
Firstly, the 40 nM FAM labeled aptamer (probe) and a desired concentration of MUC1 was rst mixed and kept at 37 C for 15 min, followed by adding 15 mg mL À1 of GO. 5 min later, the reaction solution was added with 2 U mL À1 DNase I and incubated at 37 C for 15 min, then the solution was diluted to 1 mL. Finally, the uorescence of the mixture were carried out on a Hitachi F-4600 spectrophotometer (Hitachi Co. Ltd, Japan, http://www.hitachi.co.jp) equipped with a xenon lamp excitation source at room temperature. The excitation was set at 495 nm and the emission spectra were collected from 510 nm to 600 nm. The slits of the excitation and emission were both set at 5 nm. The uorescence intensity at 519 nm was used to choose the optimal experimental conditions and evaluate the performance of the proposed sensing system. In control experiments, the measurement process was all the same with the above except the addition of MUC1. Unless otherwise noted, each uorescence measurement was repeated three times, and the standard deviation was plotted as the error bar.

Principle of design
In the present study, the principle of uorescent aptasensor detection of MUC1 based on DNase I-aided target recycling signal amplication is illustrated in Fig. 1. In the absence of MUC1, the GO can strongly adsorb single-stranded nucleic acids, thus, the MUC1 probe is adsorbed onto the surface of GO via p-p stacking between DNA bases and GO, leading to the uorescence is quenched, resulting in no uorescence signal. Upon adding the MUC1, the probe sequences could be specic recognized by MUC1 to form probe/MUC1 complexes, leading to increase in the uorescence intensity. Then, when the DNase I is added, all DNA sequences are degraded, resulting in the release of MUC1 and FAM uorophores. The released MUC1 is recycled repeatedly, leading to the accumulation of free FAM uorophores. The quantities of MUC1 can be achieved by uorescence increment. Thus, the uorescence signal is signicantly amplied by DNase I-mediated target recycling process.

Feasibility analysis of the developed method for MUC1 detection
To further verify the feasibility of DNase I-aided target recycling signal amplication strategy, Fig. 2 shows the uorescence emission spectra under different conditions. The uorescence signal produced by mixture solution DNase I and GO (curve c) is relatively weak in the absence of MUC1, indicating that GO can protect the probe from digesting by DNase I, resulting in the no FAM uorophores are produced. And then, when the MUC1 is added, the uorescence intensity do not obviously increase (curve b) in the absence of probe, thus, the MUC1 cannot specically bind to probe, lead to the no FAM uorophores are produced.
However, upon adding MUC1 and probe, signicant enhancement of the uorescence intensity was observed (curve a), as a result of the specic binding of probe to MUC1, lead to the probe/MUC1 complex formation, and keep the FAM uorophores away from GO. More importantly, the DNase Iaided target recycling results in the signicant uorescence amplication. Taking these results together, the feasibility of the proposed aptasensor for MUC1 detection by our design.

Optimization of reaction conditions
To achieve optimal sensing performance, several reaction conditions such as the concentration of DNase I, the concentration of GO, the enzyme digestion time and reaction temperature were optimized. While the F 1 and F 0 were the uorescence intensities in the presence and absence of MUC1, respectively. The uorescence intensity and the value of F 1 /F 0 are selected to evaluate the effects of the reaction conditions on the sensing performance of the method. As shown in Fig. 3A, Maximum F 1 /F 0 value is observed when the concentration of DNase I was 2 U mL À1 , however, F 1 /F 0 value decreases obviously along with the further increasing of DNase I concentration. Thus, a DNase I concentration of 2 U mL À1 was conrmed as the optimized concentration. At the same time, the concentration of GO is another important factor affecting uorescence intensity. As depicted in Fig. 3B, with the increase in the concentration of GO, the uorescence intensity initially increased and then gradually decrease, and maximum F 1 /F 0 values is observed when the concentration of GO is 15 mg mL À1 . Thus, 15 mg mL À1 GO are used in the subsequent experiments.
In addition, the enzyme digestion time and reaction temperature are another important reaction condition affecting uorescence intensity for this sensor. The uorescence intensity and the value of F 1 /F 0 change is related to the enzyme digestion reaction time. As shown in Fig. 4A, maximum F 1 /F 0 values is observed when enzyme digestion time is 15 min. Thus, 15 min of enzyme digestion time was selected for the rest of the experiments. Fig. 4B shows that the enzyme digestion temperature could obviously affect the sensitivity; the F 1 /F 0 value reached a maximum when enzyme digestion temperature was 37 C and then decreased gradually. Therefore, 37 C was conrmed as the optimized enzyme digestion temperature.
To obtain high sensitivity, the concentrations of aptamer were also investigated. As shown in Fig. S1, † the concentration of aptamer was selected in the following experiments; the F 1 /F 0 value reached a maximum when the concentration of aptamer     4 The effect of enzyme digestion reaction time and reaction temperature on the fluorescence intensity at the emission wavelength of 519 nm. The concentrations of Dnase I, MUC1, probe and GO were 2 U mL À1 , 5 ng mL À1 , 40 nM and 15 mg mL À1 , respectively. Error bars: SD, n ¼ 3. Fig. 5 (A) Fluorescence emission spectra of the biosensor in the presence of MUC1 with different concentrations: from bottom to top: 0, 10 pg mL À1 , 50 pg mL À1 , 100 pg mL À1 , 500 pg mL À1 , 1 ng mL À1 , 5 ng mL À1 , 10 ng mL À1 , 50 ng mL À1 , 100 ng mL À1 , 200 ng mL À1 and 300 ng mL À1 , respectively. (B) The relationship curve of the fluorescence intensity as a function of MUC1 concentration. While shows the relationship between fluorescence intensity and MUC1 concentration, both experimental conditions: Dnase I, 2 U mL À1 ; MUC1, 5 ng mL À1 ; probe, 40 nM; GO, 15 mg mL À1 and the emission wavelength of 519 nm. Error bars: SD, n ¼ 3. was 40 nM. Therefore, a concentration of 40 nM was selected as the optimal concentration of aptamer.

Sensitivity and specicity
Under the optimal reaction conditions, the sensitivity of the sensor for detection of MUC1 is investigated. Fig. 5A shows the uorescence emission spectra of the biosensor incubated in different concentrations of MUC1. We found that the uorescence dramatically enhance with the increasing concentration of MUC1 from 0 to 300 ng mL À1 . It illustrates a highly concentration dependence of the sensor for detection of MUC1. At the same time, Fig. 5B shows a good linear correlation between the uorescence intensity and the concentration of MUC1 in the range from 50 pg mL À1 to 100 ng mL À1 . The calibration plot of the linear equation is given as y ¼ 115.62 lg c À 91.48 (R 2 ¼ 0.9898), where lg c is logarithm of MUC1 concentration and y is the uorescence intensity. Furthermore, the detection limit is at 10 pg mL À1 . In comparison, it is much lower than other sensors reported in literatures listed in Table  1. In addition, for the specicity study, adding different control proteins was investigated, including EpCAM (50 ng mL À1 ), BSA (50 ng mL À1 ), PSA (50 ng mL À1 ) and VEGF (50 ng mL À1 ). As shown in Fig. 6, in the presence of other control proteins (50 ng mL À1 ), the signicant increase of uorescence signal is observed in the presence of the MUC1 (5 ng mL À1 ), indicating that this proposed strategy exhibited good specicity for MUC1 detection.

Determination of MUC1 in real samples
To further verify the potential applicability of this present strategy in biological samples, the detection of MUC1 in biological samples by spiking MUC1 to human urine, saliva and serum diluted to 10% with buffer solution with the nal concentration of 5 ng mL À1 were performed. As shown in Fig. 7, a signicant increase in various biological samples is observed, compared with no spiking biological samples. These results clearly demonstrate that this sensor can be a potential analytical method to detect MUC1 in real samples sensitively. The recoveries for the various concentrations of spiked MUC1 in  6 Fluorescence intensity (at the emission wavelength of 519 nm) of the sensor in the presence of MUC1 (5 ng mL À1 ), EpCAM (50 ng mL À1 ), BSA (50 ng mL À1 ), PSA (50 ng mL À1 ), VEGF (50 ng mL À1 ) and black, respectively. Error bars: SD, n ¼ 3. Fig. 7 Fluorescence intensity of the sensor for detection of MUC1 in buffer and blank biological samples (human urine, saliva and serum, respectively). Error bars: SD, n ¼ 3.
human serum were in the range of 108.4-111.2%, with the relative standard deviations (RSDs) of 7.9%, 9.8% and 8.9% at 0.1, 5.0 and 50.0 ng mL À1 of MUC1, respectively. This indicates an acceptable precision and reproducibility of the present approach for detecting MUC1 in real samples (n ¼ 3) (Table 2).

Conclusions
In summary, we have successfully developed an amplied GObased uorescence aptasensor for detecting MUC1. In this design, because of the deoxyribonuclease I-aided target recycling signal amplication, and based on GO efficient uorescence quencher, selectivity and sensitivity for MUC1 were achieved. As to MUC1, this sensor can detect MUC1 with a linear range from 50 pg mL À1 to 100 ng mL À1 , and the detection limit was low to 10 pg mL À1 . In addition, the uorescence signals are also obtained when this sensor is used for biological samples. More importantly, this biosensor achieved a high sensitivity by DNase I-aided target recycling and accumulated plenty of FAM uorophores. Thereby, this design provides an efficient platform for MUC1 detection, may be a potential method for tumor diagnosis, bioanalysis, and clinical analysis.

Conflicts of interest
There are no conicts to declare. a The average value of three measurements.