A Fluorescence Inner-Filter Effect Based Sensing Platform for Turn-On Detection of Glutathione in Human Serum

A novel turn-on fluorescence assay was developed for the rapid detection of glutathione (GSH) based on the inner-filter effect (IFE) and redox reaction. Molybdenum disulfide quantum dots (MoS2 QDs), which have stable fluorescent properties, were synthesized with hydrothermal method. Manganese dioxide nanosheets (MnO2 NSs) were prepared by exfoliating the bulk δ-MnO2 material in bovine serum albumin (BSA) aqueous solution. The morphology structures of the prepared nanoparticles were characterized by transmission electron microscope (TEM). Studies have shown that the fluorescence of MoS2 QDs could be quenched in the presence of MnO2 NSs as a result of the IFE, and is recovered after the addition of GSH to dissolve the MnO2 NSs. The fluorescence intensity showed a good linear relationship with the GSH concentration in the range 20–2500 μM, the limit of detection was 1.0 μM. The detection method was applied to the analysis of GSH in human serum samples. This simple, rapid, and cost-effective method has great potential in analyzing GSH and in disease diagnosis.


Introduction
Glutathione (GSH), the most abundant intracellular low-molecular-weight biothiol, is a tripeptide consisting of cysteine, glutamic acid, and glycine. As an essential endogenous antioxidant, GSH plays crucial roles in biological systems and many cellular processes, such as maintaining the appropriate intracellular redox status, protecting against many toxins, and promoting metabolism [1]. Abnormal GSH levels in the blood are closely related with some clinical diseases including leucocyte loss, inflammation, liver damage, psoriasis, Parkinson's, Alzheimer's, and cancer [2,3]. It has been found that the GSH level is significantly elevated in cancer cells [4]. Therefore, developing a simple and sensitive method for GSH detection in human serum is critically important for understanding GSH homeostasis in clinical settings and in effective cancer diagnosis at an early stage.
To date, a number of analytical methods have been established for GSH measurement, including high-performance liquid chromatography [5], mass spectrometry [6], surface enhanced Raman scattering [7], enzyme-linked immunosorbent assay [8], electrochemical analysis [9], and fluorescence spectroscopy [10]. Fluorescence spectrometry has been found to be a powerful tool for GSH detection because of its simplicity, high sensitivity, good flexibility, and nondestructive readout [11,12]. Several fluorescence assays have been developed for estimating GSH levels in biological samples [13,14]. Most (Hitachi, Tokyo, Japan). High-resolution transmission electron microscopic (HRTEM) imaging was performed on a JEOL 2100F field emission transmission electron microscope (JEOL, Tokyo, Japan). The dynamic light scattering (DLS) experiment was performed on a Zetasizer Nano-ZS90 (Malvern, Worcestershire, UK). The XRD pattern was obtained by a D8 Advance X-ray diffractometer (Bruker, Karlsruhe, Germany). Absolute fluorescence quantum yields (QY) measurements were conducted on an Edinburgh FLS 980 spectrometer with an integrating sphere (Edinburgh, Livingston, UK).

Preparation of Bulk δ-MnO 2 and MnO 2 NSs
Bulk δ-MnO 2 was prepared as reported previously [32]. Typically, 20 mL of a mixture containing TMA·OH (0.6 M) and H 2 O 2 (3 wt %) was added into 10 mL of MnSO 4 (0.3 M) aqueous solution within 15 s. The as-prepared dark brown solution was stirred vigorously overnight at room temperature. Then, the resulting precipitate was removed by centrifugation at 10,000 rpm for 10 min, and washed three times with ultrapure water. Finally, the precipitate was dried in a vacuum oven at 30 • C.
The MnO 2 NSs were prepared by exfoliating the bulk δ-MnO 2 in BSA aqueous solution [33]. Briefly, 10 mg bulk δ-MnO 2 was added into a 10 mL BSA (1 mg/mL) aqueous solution and the mixture was sonicated for 12 h. Then, the suspension was centrifuged at 1000 rpm for 20 min to remove the unexfoliated MnO 2 . The supernatant, which contained the MnO 2 NSs, was collected and stored in a refrigerator at 4 • C.

Preparation of MoS 2 QDs
MoS 2 QDs were synthesized by a hydrothermal method according to the literature [34] with some modifications. Concisely, 0.25 g of Na 2 MoO 4 ·2H 2 O was dissolved in 25 mL of water and subjected to ultrasonication for 5 min. Then, the solution was adjusted to pH 6.5 with 0.1 M HCl. After that, 0.5 g Cys and 50 mL water were added into the solution followed by ultrasonication for 10 min. The mixture was then heated in a Teflon-lined stainless steel autoclave at 200 • C for 36 h. After cooling to room temperature, the supernatant was collected after being centrifuged at 12,000 rpm for 30 min. Finally, the stable MoS 2 QDs were obtained after dialyzed for 19 h to remove the impurity using dialysis bag (3500 Da).

Detection of GSH
Firstly, 5 µL of GSH aqueous solution with different concentrations was added into 100 µL of 0.2 M acetate buffer (pH = 5.0) in a low adsorption tube. Then, 15 µL of MnO 2 NSs (1.0 mM) was added and the mixture was shaken and reacted for 10 min. Next, 80 µL of MoS 2 QDs (1.6 mg/mL) was added into the as-prepared solution, shaking well to ensure complete reaction. Finally, the fluorescence intensity of the whole reaction solution was detected at room temperature (Ex: 320 nm, Em: 340-600 nm, the peak was about 400 nm). The specific experiments were conducted in the same way, except using other amino acid instead of GSH solution.

Analysis of GSH in Human Serum Samples
The standard addition method was used to analyze the GSH level in human serum. Briefly, 95 µL of 0.2 M acetate buffer (pH = 5.0), 5 µL different concentrations of GSH aqueous solution, 5 µL human serum, and 15 µL of MnO 2 NSs were added into a low-adsorption tube. After being reacted for 10 min, 80 µL of MoS 2 QDs was added into the above mixture, shaking well to ensure complete reaction. Finally, the fluorescence intensity of the resulting solution was measured. The whole process was carried out at room temperature.

Fluorescence Sensing Principle of GSH
The sensing principle of GSH is shown in Scheme 1. The synthesized MoS 2 QDs emitted blue fluorescence under irradiation at 320 nm. MnO 2 NSs possessed a broad absorption spectrum and large molar extinction coefficient, making them a promising quencher in the IFE. After the addition of MnO 2 NSs, the fluorescence of MoS 2 QDs was effectively quenched as a result of the IFE. In the presence of reductive GSH, the MnO 2 NSs (strong oxidation ability) can be decomposed through the redox reaction. The released Mn 2+ lost its quenching ability, thus the fluorescence of MoS 2 QDs was recovered. On the basis of this mechanism, a "turn-on" fluorescent detection method was developed for the rapid sensing of GSH in human serum.

Fluorescence Sensing Principle of GSH
The sensing principle of GSH is shown in Scheme 1. The synthesized MoS2 QDs emitted blue fluorescence under irradiation at 320 nm. MnO2 NSs possessed a broad absorption spectrum and large molar extinction coefficient, making them a promising quencher in the IFE. After the addition of MnO2 NSs, the fluorescence of MoS2 QDs was effectively quenched as a result of the IFE. In the presence of reductive GSH, the MnO2 NSs (strong oxidation ability) can be decomposed through the redox reaction. The released Mn 2+ lost its quenching ability, thus the fluorescence of MoS2 QDs was recovered. On the basis of this mechanism, a "turn-on" fluorescent detection method was developed for the rapid sensing of GSH in human serum.

Characterization of MoS2 QDs and MnO2 NSs
The morphologies of MoS2 QDs and MnO2 NSs were characterized by transmission electron microscopy (TEM). TEM ( Figure 1A) and DLS ( Figure 1C) analysis results show that the synthesized MoS2 QDs are well dispersed with the particle size of about 3 nm. The HRTEM image shows that the MoS2 QDs has obvious lattice fringes ( Figure 1B). Figure 1D is the XRD spectrum of MoS2 QDs. Three major diffraction peaks attributed to (100), (102), and (103) planes of MoS2 can be observed [34]. The fluorescence properties of the MoS2 QDs were further investigated. From Figure 2, it was clearly observed that the most intense peak of MoS2 QDs appeared at 320 nm in the excitation spectrum (curve a). Under this excitation light, a well distinguishable peak with maximum emission intensity at 400 nm appeared (curve b). The MoS2 QDs solution emitted blue fluorescence when exposure to a 365 nm UV lamp (inset Figure 2). As shown in Figure 3A, an ultrathin lamellar structure with an irregular edge of MnO2 NSs was observed. Figure 3B shows the HRTEM image of MnO2 NSs with the diffraction pattern. The UV-Vis spectrum shows that the prepared MnO2 NSs possessed a broad absorption in the range 200-600 nm ( Figure 3C).

Characterization of MoS 2 QDs and MnO 2 NSs
The morphologies of MoS 2 QDs and MnO 2 NSs were characterized by transmission electron microscopy (TEM). TEM ( Figure 1A) and DLS ( Figure 1C) analysis results show that the synthesized MoS 2 QDs are well dispersed with the particle size of about 3 nm. The HRTEM image shows that the MoS 2 QDs has obvious lattice fringes ( Figure 1B). Figure 1D is the XRD spectrum of MoS 2 QDs. Three major diffraction peaks attributed to (100), (102), and (103) planes of MoS 2 can be observed [34]. The fluorescence properties of the MoS 2 QDs were further investigated. From Figure 2, it was clearly observed that the most intense peak of MoS 2 QDs appeared at 320 nm in the excitation spectrum (curve a). Under this excitation light, a well distinguishable peak with maximum emission intensity at 400 nm appeared (curve b). The MoS 2 QDs solution emitted blue fluorescence when exposure to a 365 nm UV lamp (inset Figure 2). As shown in Figure 3A, an ultrathin lamellar structure with an irregular edge of MnO 2 NSs was observed. Figure 3B shows the HRTEM image of MnO 2 NSs with the diffraction pattern. The UV-Vis spectrum shows that the prepared MnO 2 NSs possessed a broad absorption in the range 200-600 nm ( Figure 3C).

Feasibility Analysis of GSH Detection
To investigate the ability of GSH to dissolve MnO 2 NSs, different concentrations of GSH were added into the MnO 2 NSs, and then the solutions were analyzed by UV-Vis spectrophotometer. As shown in Figure 4, the MnO 2 NSs has a characteristic absorption peak at about 380 nm (curve a), which is consistent with the literature [35]. With the increase of GSH concentration, the absorption intensity decreased (curve b-e) and the color of the solution became gradually shallower (inset Figure 4). These results indicated that the MnO 2 NSs can be decomposed by GSH through the redox reaction [36]. According to the Beer-Lambert law (A = kbc, k = 9.6 × 10 3 M −1 cm −1 , b = 1 cm), the concentration of the synthetic MnO 2 NSs was calculated to be about 1 mM.
In order to verify the fluorescence sensing feasibility of GSH based on the probe of MoS 2 QDs and MnO 2 NSs, serials experiments were performed ( Figure 5). It can be seen that MoS 2 QDs emitted strong fluorescence (bar graph a). The added GSH did not greatly change the fluorescence intensity of MoS 2 QDs (bar graph b). After addition of MnO 2 NSs, the fluorescence of MoS 2 QDs was significantly quenched (bar graph c). In the presence of GSH, the MnO 2 NSs was dissolved and thus recovered the fluorescence (bar graph d). Therefore, it can be proved that the proposed fluorescence assay was feasible for the "turn-on" detection of GSH. The QY of the MoS 2 QDs probe and for the restored fluorescence in the presence of GSH were measured to be 2.13% and 2.39% by the absolute quantum yield measurement method, respectively.

Feasibility Analysis of GSH Detection
To investigate the ability of GSH to dissolve MnO2 NSs, different concentrations of GSH were added into the MnO2 NSs, and then the solutions were analyzed by UV-Vis spectrophotometer. As shown in Figure 4, the MnO2 NSs has a characteristic absorption peak at about 380 nm (curve a), which is consistent with the literature [35]. With the increase of GSH concentration, the absorption intensity decreased (curve b-e) and the color of the solution became gradually shallower (inset Figure 4). These results indicated that the MnO2 NSs can be decomposed by GSH through the redox reaction [36]. According to the Beer-Lambert law (A = kbc, k = 9.6 × 10 3 M −1 cm −1 , b = 1 cm), the concentration of the synthetic MnO2 NSs was calculated to be about 1 mM. In order to verify the fluorescence sensing feasibility of GSH based on the probe of MoS2 QDs and MnO2 NSs, serials experiments were performed ( Figure 5). It can be seen that MoS2 QDs emitted strong fluorescence (bar graph a). The added GSH did not greatly change the fluorescence intensity of MoS2 QDs (bar graph b). After addition of MnO2 NSs, the fluorescence of MoS2 QDs was significantly quenched (bar graph c). In the presence of GSH, the MnO2 NSs was dissolved and thus recovered the fluorescence (bar graph d). Therefore, it can be proved that the proposed fluorescence assay was feasible for the "turn-on" detection of GSH. The QY of the MoS2 QDs probe and for the restored fluorescence in the presence of GSH were measured to be 2.13% and 2.39% by the absolute quantum yield measurement method, respectively.

Feasibility Analysis of GSH Detection
To investigate the ability of GSH to dissolve MnO2 NSs, different concentrations of GSH were added into the MnO2 NSs, and then the solutions were analyzed by UV-Vis spectrophotometer. As shown in Figure 4, the MnO2 NSs has a characteristic absorption peak at about 380 nm (curve a), which is consistent with the literature [35]. With the increase of GSH concentration, the absorption intensity decreased (curve b-e) and the color of the solution became gradually shallower (inset Figure 4). These results indicated that the MnO2 NSs can be decomposed by GSH through the redox reaction [36]. According to the Beer-Lambert law (A = kbc, k = 9.6 × 10 3 M −1 cm −1 , b = 1 cm), the concentration of the synthetic MnO2 NSs was calculated to be about 1 mM. In order to verify the fluorescence sensing feasibility of GSH based on the probe of MoS2 QDs and MnO2 NSs, serials experiments were performed ( Figure 5). It can be seen that MoS2 QDs emitted strong fluorescence (bar graph a). The added GSH did not greatly change the fluorescence intensity of MoS2 QDs (bar graph b). After addition of MnO2 NSs, the fluorescence of MoS2 QDs was significantly quenched (bar graph c). In the presence of GSH, the MnO2 NSs was dissolved and thus recovered the fluorescence (bar graph d). Therefore, it can be proved that the proposed fluorescence assay was feasible for the "turn-on" detection of GSH. The QY of the MoS2 QDs probe and for the restored fluorescence in the presence of GSH were measured to be 2.13% and 2.39% by the absolute quantum yield measurement method, respectively.

Optimization of Experimental Conditions
The concentration of MoS 2 QDs and MnO 2 NSs, used as the fluorescence and quencher probe, respectively, have a great effect on the detection sensitivity. The fluorescence intensities of different concentrations of MoS 2 QDs solutions were determined. As shown in Figure 6A, the fluorescence intensity decreased as the dilution multiple of the MoS 2 QDs increased, then the decreasing trend gradually tended to be gentle. Considering the sensitivity and fluorescence intensity, the final concentration of 2.5-fold diluted MoS 2 QDs solution was used in the following experiments. As the amount of MnO 2 NSs increased, the fluorescence of MoS 2 QDs sharply decreased ( Figure 6B). Excessive MnO 2 NSs reduces the GSH detection sensitivity. After a comprehensive analysis of the fluorescence inhibition degree and sensitivity, 15 µL of MnO 2 NSs was finally selected. The reaction time between GSH and MnO 2 NSs was also investigated. The fluorescence intensity increased gradually with the time prolonging from 0 to 10 min, then it tended to be stable ( Figure 6C). Therefore, the reaction time was selected to be 10 min.

Optimization of Experimental Conditions
The concentration of MoS2 QDs and MnO2 NSs, used as the fluorescence and quencher probe, respectively, have a great effect on the detection sensitivity. The fluorescence intensities of different concentrations of MoS2 QDs solutions were determined. As shown in Figure 6A, the fluorescence intensity decreased as the dilution multiple of the MoS2 QDs increased, then the decreasing trend gradually tended to be gentle. Considering the sensitivity and fluorescence intensity, the final concentration of 2.5-fold diluted MoS2 QDs solution was used in the following experiments. As the amount of MnO2 NSs increased, the fluorescence of MoS2 QDs sharply decreased ( Figure 6B). Excessive MnO2 NSs reduces the GSH detection sensitivity. After a comprehensive analysis of the fluorescence inhibition degree and sensitivity, 15 μL of MnO2 NSs was finally selected. The reaction time between GSH and MnO2 NSs was also investigated. The fluorescence intensity increased gradually with the time prolonging from 0 to 10 min, then it tended to be stable ( Figure 6C). Therefore, the reaction time was selected to be 10 min.

GSH Detection Sensitivity
To investigate the capability of the fluorescence assay for quantitative measurement of GSH, different concentrations of GSH were added to the system and the fluorescence responses were recorded under optimal conditions. The fluorescence of MoS2 QDs was found to be gradually restored with an increasing concentration of GSH ( Figure 7A). The more GSH added, the more MnO2 NSs decomposition, which resulted in a restoration of the fluorescence. The standard curve was obtained between the fluorescence intensity and the GSH concentration ( Figure 7B). The inset in Figure 7B shows that the fluorescence intensity was linearly related to the GSH concentration in a range from 0.02 to 2.5 mM with a correlation coefficient of 0.996. The linear equation is Y = 301.94X + 702.72, where Y is the fluorescence intensity and X is the concentration of GSH (mM). The detection limit was calculated to be 1.0 μM according to the three-fold standard deviation of the blank signal (3σ). As shown in Table 1, the detection sensitivity was comparable to that of some reported methods and could meet the requirement of GSH detection in biological and clinical applications (mM levels) [7].

GSH Detection Sensitivity
To investigate the capability of the fluorescence assay for quantitative measurement of GSH, different concentrations of GSH were added to the system and the fluorescence responses were recorded under optimal conditions. The fluorescence of MoS 2 QDs was found to be gradually restored with an increasing concentration of GSH ( Figure 7A). The more GSH added, the more MnO 2 NSs decomposition, which resulted in a restoration of the fluorescence. The standard curve was obtained between the fluorescence intensity and the GSH concentration ( Figure 7B). The inset in Figure 7B shows that the fluorescence intensity was linearly related to the GSH concentration in a range from 0.02 to 2.5 mM with a correlation coefficient of 0.996. The linear equation is Y = 301.94X + 702.72, where Y is the fluorescence intensity and X is the concentration of GSH (mM). The detection limit was calculated to be 1.0 µM according to the three-fold standard deviation of the blank signal (3σ). As shown in Table 1, the detection sensitivity was comparable to that of some reported methods and could meet the requirement of GSH detection in biological and clinical applications (mM levels) [7].

Interference Detection Results
To assess the selectivity of GSH detection based on the MoS2 QDs-MnO2 NSs system, the influence of some amino acids, metal ions, and proteins which possibly exist in human serum was studied. As shown in Figure 8, the system exhibited a remarkable increase in fluorescence in the presence of GSH. On the contrary, there were no obvious changes in fluorescence intensity with the other molecules. It was also found that when GSH was mixed with some other amino acids, the fluorescent intensity was similar to GSH alone. Although a high concentration of cysteine (Cys) and homo-cysteine (H-Cys) can cause a measurable fluorescence response to this system, their contents (μM levels) are much lower than that of GSH (mM levels) in biological systems [41,42]. These results show that the system displays a highly selective response of fluorescence enhancement towards GSH over other molecules. Figure 8. Selectivity investigation of the system for GSH detection. The concentration of GSH, oxidized glutathione (GSSG), and other amino acid solutions were 2 mM, bovine serum albumin (BSA) was 1 mg/mL, and the metal ions concentration was 5 mM. Bar graph "X" was the fluorescent recovery of the mixture of several amino acids (Glu, His, Phe, Gly, Met, Ala, and Dith with a concentration of 2 mM) with 2 mM GSH.

GSH Detection in Human Serum
Human serum experiments were further performed to investigate the feasibility of the MoS2 QDs-MnO2 NSs system in the analysis of GSH in actual samples. Different concentrations of standard GSH were spiked in the human serum and the samples were measured in parallel three times. As shown in Table 2, the recovery rate of the method was found to be 99.2%-104%, and the

Interference Detection Results
To assess the selectivity of GSH detection based on the MoS 2 QDs-MnO 2 NSs system, the influence of some amino acids, metal ions, and proteins which possibly exist in human serum was studied. As shown in Figure 8, the system exhibited a remarkable increase in fluorescence in the presence of GSH. On the contrary, there were no obvious changes in fluorescence intensity with the other molecules. It was also found that when GSH was mixed with some other amino acids, the fluorescent intensity was similar to GSH alone. Although a high concentration of cysteine (Cys) and homo-cysteine (H-Cys) can cause a measurable fluorescence response to this system, their contents (µM levels) are much lower than that of GSH (mM levels) in biological systems [41,42]. These results show that the system displays a highly selective response of fluorescence enhancement towards GSH over other molecules.

Interference Detection Results
To assess the selectivity of GSH detection based on the MoS2 QDs-MnO2 NSs system, the influence of some amino acids, metal ions, and proteins which possibly exist in human serum was studied. As shown in Figure 8, the system exhibited a remarkable increase in fluorescence in the presence of GSH. On the contrary, there were no obvious changes in fluorescence intensity with the other molecules. It was also found that when GSH was mixed with some other amino acids, the fluorescent intensity was similar to GSH alone. Although a high concentration of cysteine (Cys) and homo-cysteine (H-Cys) can cause a measurable fluorescence response to this system, their contents (μM levels) are much lower than that of GSH (mM levels) in biological systems [41,42]. These results show that the system displays a highly selective response of fluorescence enhancement towards GSH over other molecules. Figure 8. Selectivity investigation of the system for GSH detection. The concentration of GSH, oxidized glutathione (GSSG), and other amino acid solutions were 2 mM, bovine serum albumin (BSA) was 1 mg/mL, and the metal ions concentration was 5 mM. Bar graph "X" was the fluorescent recovery of the mixture of several amino acids (Glu, His, Phe, Gly, Met, Ala, and Dith with a concentration of 2 mM) with 2 mM GSH.

GSH Detection in Human Serum
Human serum experiments were further performed to investigate the feasibility of the MoS2 QDs-MnO2 NSs system in the analysis of GSH in actual samples. Different concentrations of standard GSH were spiked in the human serum and the samples were measured in parallel three times. As shown in Table 2, the recovery rate of the method was found to be 99.2%-104%, and the

GSH Detection in Human Serum
Human serum experiments were further performed to investigate the feasibility of the MoS 2 QDs-MnO 2 NSs system in the analysis of GSH in actual samples. Different concentrations of standard GSH were spiked in the human serum and the samples were measured in parallel three times. As shown in Table 2, the recovery rate of the method was found to be 99.2-104%, and the RSD was within 10%. It is shown that this method can be used for the detection of GSH in human serum with acceptable reproducibility and accuracy.

Reaction Mechanism Discussion
The fluorescence internal filter effect refers to the phenomenon that occurs when the fluorophor coexists with other absorbents, and the fluorescence is weakened due to the absorption of the excitation or emission light by the absorbent. The broad absorption of MnO 2 NSs possesses a peak at about 380 nm ( Figure 3C), which overlaps both with the emission and excitation wavelength of MoS 2 QDs. Therefore, when both of them exist in the system, the fluorescence of MoS 2 QDs will be quenched due to the IFE ( Figure 9, curve a). Interestingly, we found that after high-speed centrifugation of the reaction solution, the fluorescence of the supernatant was restored ( Figure 9, curve b), possibly as a result of the suspended MnO 2 NSs in the solution being separated from the MoS 2 QDs and precipitated in the bottom of the centrifuge tube. Thus, the fluorescence of MoS 2 QDs in the supernatant cannot be quenched by MnO 2 NSs through the IFE. The experimental results confirmed that the detection of GSH was based on the IFE, which is different from the reported FRET mechanism of single-layer MnO 2 nanosheets [36,43]. RSD was within 10%. It is shown that this method can be used for the detection of GSH in human serum with acceptable reproducibility and accuracy.

Reaction Mechanism Discussion
The fluorescence internal filter effect refers to the phenomenon that occurs when the fluorophor coexists with other absorbents, and the fluorescence is weakened due to the absorption of the excitation or emission light by the absorbent. The broad absorption of MnO2 NSs possesses a peak at about 380 nm ( Figure 3C), which overlaps both with the emission and excitation wavelength of MoS2 QDs. Therefore, when both of them exist in the system, the fluorescence of MoS2 QDs will be quenched due to the IFE ( Figure 9, curve a). Interestingly, we found that after high-speed centrifugation of the reaction solution, the fluorescence of the supernatant was restored ( Figure 9, curve b), possibly as a result of the suspended MnO2 NSs in the solution being separated from the MoS2 QDs and precipitated in the bottom of the centrifuge tube. Thus, the fluorescence of MoS2 QDs in the supernatant cannot be quenched by MnO2 NSs through the IFE. The experimental results confirmed that the detection of GSH was based on the IFE, which is different from the reported FRET mechanism of single-layer MnO2 nanosheets [36,43].

Stability Investigation
The stability of the prepared MoS2 QDs with and without dialysis was investigated. We found that after being placed at room temperature for 42 days, the synthesized MoS2 QDs after dialysis retained about 96% of their original fluorescence intensity. However, the fluorescence of MoS2 QDs decreased sharply without dialysis, being only 61% of the previous measurement. It shows that the stability of MoS2 QDs is greatly improved by dialysis treatment, which is a benefit for practical applications.

Stability Investigation
The stability of the prepared MoS 2 QDs with and without dialysis was investigated. We found that after being placed at room temperature for 42 days, the synthesized MoS 2 QDs after dialysis retained about 96% of their original fluorescence intensity. However, the fluorescence of MoS 2 QDs decreased sharply without dialysis, being only 61% of the previous measurement. It shows that the stability of MoS 2 QDs is greatly improved by dialysis treatment, which is a benefit for practical applications.

Conclusions
In summary, a signal enhanced fluorescence assay was developed for the rapid detection of GSH in human serum. The fluorescence of MoS 2 QDs can be effectively quenched by MnO 2 NSs as a result of the IFE. Interestingly, the quenched fluorescence was restored when reductive GSH was introduced to dissolve MnO 2 NSs through the redox reaction. The final fluorescence intensity was linearly correlated with the concentration of GSH. The whole detection process can be completed within 15 min. Notably, the fluorescence response was selective toward GSH without interference from other substances in human serum. Thus, the assay was successfully used for the analysis of GSH in human serum samples. We anticipate that this simple, stable, selective, and cost-effective "turn-on" fluorescent assay has a good prospect of application for GSH analysis in biomedical fields.
Author Contributions: S.T. and X.Y. conceived and designed the experiments; Q.F. and X.L. performed the experiments; X.Y. and G.L. have helped analyze the results of the measured data; S.T. wrote the paper. W.C. and J.C. have proposed valuable suggestions on the revise of the manuscript. All coauthors reviewed and revised the paper.

Conflicts of Interest:
The authors declare no conflict of interest.