Simultaneous Detection of Adenosine Triphosphate and Glucose Based on the Cu-Fenton Reaction

Abstract

Both adenosine triphosphate (ATP) and glucose are important to human health, and their abnormal levels are closely related to angiocardiopathy and hypoglycaemia. Therefore, the simultaneous determination of ATP and glucose with a single test mode is highly desirable for disease diagnostics and early recognition. Herein, a new fluorescence on/off switch sensing platform is developed by carbon nanodots (CNDs) to detect ATP and glucose simultaneously. The fluorescence of CNDs can be quenched by Cu²⁺ and hydrogen peroxide (H₂O₂), due to the formation of hydroxyl radicals (·OH) produced in the Cu-Fenton reaction. Based on the high affinity of Cu²⁺ with ATP, the fluorescence of CNDs will recover effectively after adding ATP. Additionally, glucose can be efficiently catalyzed by glucose oxidase (GOx) to generate H₂O₂, so the platform can also be utilized to analyze glucose. Under optimum conditions, this sensing platform displays excellent sensitivity and the linear ranges are from 0.1 to 7 μM for ATP with a limit of detection (LOD) of 30.2 nM, and from 0.1 to 7 mM for glucose with a LOD 39.8 μM, respectively. Benefiting from the high sensitivity and selectivity, this sensing platform is successfully applied for simultaneous detection of ATP and glucose in human serum samples with satisfactory recoveries.

1. Introduction

As is well known, both adenosine triphosphate (ATP) and glucose are sources of energy and important for energy metabolism. ATP supplies energy for metabolic processes and glucose is a metabolic intermediate in biological systems. Furthermore, the ATP and glucose levels in blood are indicators of human health conditions. For example, the excessive production of ATP by creatine kinase is the pathogenesis of angiocardiopathy [1] and the aberrant concentration of ATP will cause energy disturbance, resulting in hypoglycaemia, ischemia, and Parkinsons’ disease [2]. On the other hand, the high glucose levels will produce diabetes mellitus, which damages the eyes, kidneys, feet, and heart, and also causes a series of angiocardiopathy [3], while hypoglycaemia is usually due to the low concentration of glucose. This means that both the levels of ATP and glucose are related to angiocardiopathy and hypoglycaemia by virtue of the coexistence of ATP and glucose in biological samples [4,5,6,7], thus, the simultaneous determination of ATP and glucose with a single test mode is highly desirable for disease diagnostics and early recognition.

A comparison with the traditional methods, such as chromatography and electrochemical techniques [8,9], fluorescence analysis receives a great deal of attention due to its simplicity, rapid response, and high selectivity. The fluorescence sensing strategy for ATP is usually developed according to the reaction between ATP and aptamer or between ATP and copper ions (Cu²⁺). For example, ATP-aptamer complexes provided greater protection for gold nanoparticles (AuNPs) against salt-induced aggregation than either aptamer or ATP alone, and the dispersive AuNPs rather than aggregation could efficiently quench the fluorescence of Tb ion-functionalized carbon dots due to the fluorescence resonance energy transfer [10]. In addition, compared with other cations, ATP has a strong affinity for Cu²⁺, which can completely quench the emission of perylene diimide functionalized with histidine. Thus, the fluorescence recovery positively correlated with ATP concentrations [11]. On the other hand, many fluorescence glucose sensors based on new fluorescence transduction schemes are already developed [12]. In these reports, the enzyme glucose oxidase (GOx) has been widely employed in glucose sensing because hydrogen peroxide (H₂O₂) is the main product of the reaction between glucose and GOx. For instance, H₂O₂ formed in the oxidization of glucose by GOx, resulting in the fluorescence quenching of carbon nanodots (CNDs) in the presence of Fe²⁺, and the concentration of glucose could be measured indirectly [13]. Although these reports showed good sensitivity and selectivity, these strategies could not incorporate both elements which responded, respectively, to Cu²⁺ and H₂O₂, so the simultaneous measurement of ATP and glucose could not be achieved. Up to now, only an amperometric biosensor was reported for the simultaneous determination of ATP and glucose [14], and other ways, especially fluorescence techniques, have not been presented.

Herein, it is found that the complex of Cu²⁺ and H₂O₂, rather than either Cu²⁺ or H₂O₂ alone, can quench the fluorescence of CNDs due to the hydroxyl radicals (·OH) produced in the Cu-Fenton reaction [15,16,17]. By virtue of the binding of ATP to Cu²⁺ and arising H₂O₂ in the reaction between glucose and GOx, ATP and glucose can be simultaneously detected. The linear range of ATP is 0.1 to 7 µM and the linear response of glucose is from 0.1 to 7 mM. Benefiting from the superior sensitivity and selectivity, this sensing platform is successfully applied for simultaneous detection of ATP and glucose in human serum samples with satisfactory recoveries.

2. Materials and Methods

2.1. Chemicals and Materials

Ascorbic acid (AA), hydrogen peroxide (H₂O₂), glucose oxidase (GOx), glucose, galactose, fructose, sucrose, mannose, lactose, sodium citrate, adenosine triphosphate (ATP), cytidine 5′-triphosphate disodium salt (CTP), adenosine 5′-diphosphate (ADP), uridine-5′-triphosphate (UTP), guanosine triphosphate (GTP) were purchased from Aladdin (Shanghai, China). CH₃COONa, HCl, KCl and CuSO₄ were obtained from Shanghai Shenbo Chemical Co., Ltd., Shanghai, China. All reagents were of analytical grade. All the solutions were prepared using ultrapure water produced with a Millipore-Q water system.

2.2. Instruments

The fluorescence spectra were recorded with a Hitachi F-7000 fluorescence spectrophotometer (Tokyo, Japan). Transmission electron microscopy (TEM) images were recorded on a JEM-2100PLUS (JEOL, Japan). The Fourier transform infrared (FT-IR) spectra of the samples were analyzed using Thermo Nicolet Nexus 470 FT-IR ESP spectrometer (Nicolet, WI, USA). The ultraviolet-visible (UV-vis) absorption spectra were obtained on a Cary 300 Bio UV-vis spectrophotometer (Varian, Palo Alto, CA, USA) and the pH values of solutions were measured using a pH meter (Mettler Toledo FE20, Zurich, Switzerland).

2.3. Synthesis of CNDs

Typically, 0.8 g AA was dissolved in 20 mL water. Then the above solution was stirred thoroughly and then transferred into a 50 mL autoclave reactor. The reactor was heated to 160 °C in a constant temperature drying oven for 70 min. After cooling to room temperature, a clear yellow aqueous dispersion containing CNDs was gained. Subsequently, the resulting yellow solution (20 mL) was dried in a drying oven at 60 °C to constant weight (for about 14 h). Then, 39 mg powder of carbon nanodots was weighed and diluted in 1 mL water to obtain CNDs solution (39 mg/mL). The prepared solution of CNDs was diluted 100 times before use, and the concentration was 0.39 mg/mL.

2.4. Fluorescence Responses of Cu²⁺

The fluorescence detection of Cu²⁺ was performed as follows: briefly, 35 µL CNDs (0.39 mg/mL), 50 µL H₂O₂ (0.1 M), different concentrations of Cu²⁺, 200 µL HCl-CH₃COONa buffers (pH = 5) and water were mixed together, and the final volume was 1 mL. After incubation for 10 min at room temperature, the fluorescence of the mixture was measured at 380 nm with an excitation of 315 nm. With increasing the concentration of Cu²⁺, the fluorescence of CNDs decreased linearly and the fluorescence differences were expressed as F₀-F, in which F₀ and F represented the fluorescence intensities of CNDs-H₂O₂ in the absence and presence of Cu²⁺, respectively.

Additionally, the LOD is the limit of detection, which is defined as the concentration of analyte that corresponds to three times the signal-to-noise ratio (S/N = 3) [18]. LOD is calculated according to the expression LOD = 3σ/K, where σ is the standard deviation for the blank solution (n = 10), and K is the slope of the calibration curve.

2.5. Fluorescent H₂O₂ Assay

A typical H₂O₂ detection procedure was conducted as follows: Firstly, 35 µL CNDs (0.39 mg/mL), 60 µL Cu²⁺ (0.01 mM), 200 µL HCl-CH₃COONa buffers (pH = 5), various concentrations of H₂O₂ and water were mixed thoroughly. The final volume of the mixture was 1 mL. After incubation for 10 min at ambient temperature, the concentration of H₂O₂ linearly responded to the fluorescence differences of F₁-F₂ (F₁ and F₂ represented the fluorescence intensity of CNDs-Cu²⁺ in the absence and presence of H₂O₂).

2.6. Fluorescence Detection of ATP

The assay was carried out as follows: (i) 200 µL HCl-CH₃COONa buffers (pH = 5), 60 µL Cu²⁺ (0.01 mM) and ATP with different concentrations were added to a certain amount of water and reacted for 10 min at room temperature; (ii) 35 µL CNDs (0.39 mg/mL) and 50 µL H₂O₂ (0.1 M) were injected to the mixture and the volume of final solution was 1 mL. After the mixture reacted for 10 min at room temperature, a linear correlation was found between the concentration of ATP and the fluorescence differences of F₄ − F₃, in which F₃ was the fluorescence intensity of CNDs-Cu²⁺-H₂O₂ system and F₄ was the fluorescence intensity of the system with the addition of various concentrations of ATP.

2.7. Fluorescence Measurement of Glucose

Firstly, the mixture containing 13 μL GOx (765 U/mL) and different concentrations of glucose was incubated at 37 °C for 60 min, and then 200 µL HCl-CH₃COONa buffers (pH = 5), 60 μL Cu²⁺ (0.01 mM) and 35 µL CNDs (0.39 mg/mL) were added to the mixture. Ten minutes later, their fluorescence spectra were recorded. The concentration of glucose linearly responded to the fluorescence differences of F₅ − F₆, where F₅ and F₆ were the fluorescence of CNDs-Cu²⁺-GOx system in the absence and presence of glucose.

2.8. Real Sample Detection

The real sample detection in our assay was carried out as follows: Typically, a 1.5 mL blood sample was kept at 37 °C in a water bath for about 0.5 h. Then, 3.0 mL acetonitrile was added and followed by centrifugation at 4000 rpm for 20 min to remove large molecules and proteins. Subsequently, the supernatant was removed and kept in a water bath (60 °C) for about 1 h to evaporate acetonitrile. The supernatant of the serum sample was diluted 10 times and measured by the present sensing system to detect ATP and glucose. Then, the standard ATP and glucose solutions with different concentrations were spiked into the human serum samples, respectively, to investigate the reliability of this assay.

2.9. Validation Test of TMB

The validation of the Cu-Fenton activity was performed as follows: the experiments were performed in 200 µL HCl-CH₃COONa buffers (pH = 5) containing 5 mM TMB, 500 mM H₂O₂ and differing concentrations of Cu²⁺ followed by the addition of water to a final volume of 1 mL. After reaction for 10 min, the colour was present in these samples.

3. Results and Discussion

3.1. Characterization of CNDs

As shown in the TEM image (Figure 1a), the CNDs exhibit an approximately spherical shape and are well-dispersed. These nanoparticles have a size distribution from 1 to 3.4 nm and the average diameter of CNDs is about 1.8 nm by estimating the size of 100 dots in the TEM image. The structure and composition of CNDs are characterized by FT-IR spectroscopy (Figure 1b). The CNDs have a main absorption of C-H stretching vibration at 2900 cm⁻¹, the C=O stretching at 1740 cm⁻¹, and the peak at 3452 cm⁻¹ is related to the C-OH bond stretching vibrations. It demonstrates the presence of oxygen-containing functional groups such as hydroxyl and carbonyl groups on the surface of particles. Additionally, these nanodots show a strong peak at 260 nm (Figure 1c), which is ascribed to π-π* transition of carbon [19]. Next, the emission is found to be excitation-dependent, including wavelength and intensity in a well-regulated mode within the exciting range of 285–345 nm (Figure S1). Although the emission peak shifts from 375 to 390 nm with increasing excitation wavelengths, its intensity reaches a maximum at 315 nm excitation (Figure 1c). The excitation-dependent emission spectra have been reported by various works involving CNDs [20,21]. Meanwhile, the CNDs solution under UV light exhibits bright blue fluorescence, whereas it is yellow under daylight.

3.2. Fluorescence Quenching of CNDs

According to Figure 2, when only H₂O₂ or Cu²⁺ is mixed with CNDs, the emission at 380 nm changes inconspicuously. However, by both adding H₂O₂ and Cu²⁺, the fluorescence of CNDs is quenched effectively. While the amount of H₂O₂ is fixed, the emission of CNDs sensitively responds to the concentration of Cu²⁺ in the range from 1 nM to 2 μM with LOD 0.65 nM (Figure S2), which is lower than most works [22,23]. Additionally, it should be noted that when the concentration of Cu²⁺ is higher than 4 μM (Figure S3), it can directly quench the fluorescence of CNDs without the addition of H₂O₂, which is consistent with previous literature [24], but Cu²⁺ ions with a low amount (especially below 2 μM) is invalid. Additionally, with the unchanged amount of Cu²⁺, the fluorescence of these nanoparticles also decreases linearly when the concentration of H₂O₂ increases from 25 µM to 5 mM with LOD 7.8 µM (Figure S4). Like the Fenton reaction between Fe²⁺ and H₂O₂, Cu²⁺ and H₂O₂ can also produce the reactive ·OH. In this process, Cu²⁺ oxidizes H₂O₂ to generate O₂^•-and Cu⁺, and Cu⁺ is able to react with excess H₂O₂ to form ·OH [25,26,27,28], which can catalyse the oxidation of TMB to produce the blue colour reaction (Figure S5). TMB oxidation-induced colour development is positively correlated with Cu²⁺ concentrations. Therefore, the ·OH is believed to quench the fluorescence of CNDs, and the quenching efficiency reaches the maximum at pH 5 (Figure S6). Cu-Fenton reaction is different from the Fe-Fenton reaction, and it usually requires a small amount of Cu²⁺ and a large amount of H₂O₂, which just correspond to the low concentration of ATP and high concentration of glucose in human serum. Based on the binding of ATP to Cu²⁺ and arising H₂O₂ in the reaction between glucose and GOx, ATP, and glucose can be simultaneously detected with a single test mode (Scheme 1).

3.3. Fluorescence Recovery of CNDs-Cu²⁺-H₂O₂ System by ATP

As shown in Figure 3, when the ATP is introduced in the CNDs-Cu²⁺-H₂O₂ system, the quenched fluorescence recovers effectively because Cu²⁺ ions exhibit a strong binding affinity to ATP with respect to multiple phosphates, while ATP alone will not influence the fluorescence of CNDs, CNDs-Cu²⁺, and CNDs-H₂O₂. All of results indicate that ATP indeed binds with Cu²⁺ ions, which cannot react with H₂O₂ to form ·OH. At the same time, the reaction between ATP and Cu²⁺ ions can complete within 10 min, suggesting a rapid reaction rate (Figure S7). The enhanced fluorescence positively correlates with the concentration of ATP in the linear range from 0.1 to 7 µM with LOD 30.2 nM (Figure 4). Comparing with the previous reports (

Table 1. Comparison of analysis methods for the detection of ATP.

Method	Probe	Linear Range	LOD	Ref.
Fluorescence	Ru complex	0–0.1 µM	20 nM	29
Fluorescence	DNA template Ag nanoclusters (AgNCs)	0.5–8.0 µM	91.6 nM	30
Colorimetric detection	Gold nanoparticles	4.4–132.7 µM	0.6 µM	31
Fluorescence	Cysteamine capped CdS quantum dots (QDs)	20–80 µM	17 µM	32
Fluorescence	DNA binding dye berberine	0.5–50 µM	140 nM	1
Fluorescence	QD-tagged aptamer	0.1–1 mM	24 µM	33
Electrochemical	labeled anti-ATP aptamer	10 nM–1 mM	-	34
Fluorescence and Colorimetric detection	cationic polythiophene derivative	10−8–10−4 M	-	35
Fluorescence	CNDs	0.1–7 µM	30.2 nM	This work

), our method is not inferior to others [29,30,31,32,33,34,35] and also exhibits relatively wide linear ranges and a low limit of detection of ATP.

3.4. Detection of Glucose Based on CNDs-Cu²⁺ System

H₂O₂ is the main product of the reaction between glucose and GOx, so CNDs-Cu²⁺-H₂O₂ system can be utilized to analyse glucose. As illustrated in Figure 5, there is no significant difference when glucose or GOx is added in CNDs and CNDs-Cu²⁺ system. When glucose is oxidized by GOx, the fluorescence of CNDs-Cu²⁺ platform was quenched effectively. The fluorescence difference was proportional to the concentration of glucose in the range from 0.1 to 7 mM with LOD 39.8 µM (Figure 6). The fluorescence response of this platform to the glucose ranging from 0 to 1.0 mM is depicted in Figure S8, where the slope of the linear equation is similar to that of glucose with concentration from 0.1 to 7.0 mM, indicating that the glucose with low concentrations can be also sensitively detected. In comparison with the reported studies (

Table 2. Comparison of analysis methods for the detection of glucose.

Method	Probe	Linear Range	LOD	Ref.
Colorimetric detection	MoS2 nanosheets	5–150 µM	1.2 µM	36
Fluorescence	boronic acid modified carbon dots	9–900 µM	1.5 µM	19
Colorimetric detection	Gold nanoparticles	0–5 mM	-	37
Colorimetric detection	ZnS nanoparticles	0.05–0.5 mM	36 µM	38
Fluorescence	Copper nanoclusters	10–100 µM	8 µM	39
Colorimetric detection	Prussian blue nanoparticles	0.1–50 µM	0.03 µM	40
Fluorescence	carbon nitride dots	1–100 µM	0.4 µM	41
Fluorescence	CNDs	0.1–7 mM	39.8 µM	This work

), this new strategy is demonstrated to not be inferior to other works [36,37,38,39,40,41]. Additionally, CNDs involved in this assay, as a novel probe, are easily synthesized, experimentally convenient, and low-cost, and can be used to detect ATP and glucose simultaneously.

3.5. Selectivity for ATP and Glucose

To examine the selectivity of this CNDs-Cu²⁺-H₂O₂ system for ATP, other analogous molecules, including CTP, ADP, UTP, and GTP are investigated under the optimized conditions. As shown in Figure 7a, only the addition of ATP produces an obvious fluorescence recovery, suggesting the highly selectivity for ATP. Additionally, some interference substances, such as galactose, fructose, sucrose, mannose, lactose, sodium citrate, AA, and KCl, are evaluated. Due to the high specificity of GOx for glucose, other compounds cannot induce the effective fluorescence quenching of CNDs, indicating a high selectivity of our proposed sensing system for glucose.

3.6. Detection of ATP and Glucose in Human Serum Samples

To evaluate the applicability of this sensing system, the proposed strategy is used to determine ATP and glucose in human serum samples. The results are shown in

Table 3. Determination of ATP in serum samples.

Sample	Standard Added µM	Found Value µM	Recovery (%)	RSD (n = 3, %)
1	0.0	0.433 *	-	1.43
2	0.1	0.537	104.0	1.68
3	1.0	1.497	106.4	1.16
4	5.0	5.553	102.4	1.61

* The concentration of ATP in human serum sample with diluted 10 times measured by our method is 0.43 µM, which is in agreement with the values detected by HPLC (4.1 µM without dilution).

, the concentration of ATP in human serum sample (without dilution) measured by our method is about 4.3 µM, which is consistent with the value detected by the high-performance liquid chromatograph (HPLC). With a standard addition method, the real samples are spiked with certain amounts of ATP and the satisfactory recoveries of ATP in the range from 102.4 to 106.4% are reached. On the other hand, the concentration of glucose in human serum sample (without dilution) detected by our method is about 3.41 mM, which is similar to the clinic value provided by local hospital (3.78 mM). Next, the recoveries in the ranges from 98.0 to 104.8% are obtained by spiking different concentrations standard glucose solutions into human serum samples (

Table 4. Determination of glucose in serum samples.

Sample	Standard Added mM	Found Value mM	Recovery (%)	RSD (n = 3, %)
1	0	0.341 *	-	1.38
2	0.1	0.439	98.0	1.02
3	1.0	1.389	104.8	1.52
4	3.0	3.284	98.1	1.83

* The concentration of glucose in human serum sample with diluted 10 times measured by our method is 0.341 mM, which is consistent with the clinical data provided by a local hospital (3.78 mM without dilution).

). All results demonstrate that this proposed method is practicable, reliable, and can be used for simultaneous determination of ATP and glucose in human serum samples.

4. Conclusions

Herein, a fluorescence sensing platform based on CNDs for the detection of ATP and glucose with superior sensitivity and selectivity has constructed. The sensing mechanism is that Cu²⁺ can react with H₂O₂ to generate ·OH, which quenches the fluorescence of CNDs. Incorporating both ATP and glucose, which respond respectively to Cu²⁺ and H₂O₂, the measurements of ATP and glucose can be simultaneously achieved in human serum samples with satisfactory recoveries. Due to the abnormal levels of ATP and glucose closely related to some diseases, this assay has been proposed in an effort to find a proper candidate for use in routine clinical practice.