Fluorescence strategy for evaluating the effect of traditional Chinese medicine and its active ingredients on CD73 activity using supramolecular host-guest reporter pair

2.1. Materials

CD73 and albiflorin were obtained from Titan Scientific Co., Ltd. (Shanghai, China). Salvianic acid A and senkyunolide I were obtained from Aladdin Reagent Co., Ltd. (Shanghai, China). Hydroxysafflor yellow A was purchased from Standard Co., Ltd. (Shanghai, China). Paeoniflorin and AB were purchased from bidepharm Co., Ltd. (Shanghai, China). GMP, G, and ATP were purchased from Energy Chemical Co., Ltd. (Beijing, China). TCM decoction pieces were obtained from Hunan Hengdong County TCM Piece Factory (Hunan, China). TCM patent medicines were obtained from The First Affiliated Hospital, Hengyang Medical School, University of South China (Hunan, China). TM was synthesized as previously reported [24].

2.2. Instrumentation

Fluorescence spectra were obtained on a Hitachi F-7100 spectrometer. UV-vis absorption spectra were obtained on a Shimadzu UV-2600i spectrophotometer. Constant temperature condition is provided by the DFY-5/20 vertical low-temperature constant temperature stirring reaction bath (Gongyi Yuhua Instrument Co., Ltd., Henan, China).

2.3. Preparation of solutions

Tris-HCl buffer solution (25 mM, pH = 7.4) was obtained by diluting commercially available reagent. TM (1 mM) and AB (0.1 mM) solutions were stored in a refrigerator at 4°C. G (10 mM) and GMP (10 mM) solutions were stored at room temperature. CD73 (0.2 mg/mL) enzyme solution was stored in a refrigerator at -20°C to avoid reduction of enzyme activity caused by repeated freezing and thawing. All solutions were freshly prepared with Tris buffer before use to avoid affecting the experimental results.

2.4. Handling of TCMs

TCM decoction pieces: after grinding the TCM decoction pieces, 10 g of the resulting powder was boiled with 100 mL of ultrapure water for 1 h. This process was repeated twice, and the resulting aqueous solutions were combined. The solution was then filtered and concentrated to a final concentration of 1 mg/mL to obtain the water extract of the Chinese herbal decoction pieces.

TCM patent medicines: liquid preparations are not processed; tablets were ground to create a 0.5 g/mL solution; granules were directly dissolved in ultrapure water to form a 1 g/mL solution; and the powder of the capsule was dissolved in ultrapure water to achieve a 1 g/mL solution.

During the experiments, the solutions were diluted to the appropriate concentration prior to use. For TCM decoction pieces, the concentrations of Taraxacum officinale L., Millettia, and Tripterygium wilfordii were 50 μg/mL, while the concentrations of the other ingredients were 250 μg/mL. For TCM patent medicines, the amount used in the liquid formulations was 1 μL, while the concentration of the solid formulation solutions was 50 μg/mL, except for Luohua Zizhu Dispersible (Tablet) at 10 μg/mL. All TCM solutions were stored in a refrigerator at 4°C.

2.5. Fluorescence spectroscopy

In the direct titration method, the sequential change in fluorescence intensity at various host concentrations of the host was used to determine the host-guest association constant (K_a). Competitive fluorescence titrations were carried out by gradually adding known concentrations of the substrate (GMP) and the product (G) to a solution containing the appropriate concentrations of the host and the dye. This induced a change in the fluorescence intensity. The host and dye concentration ratios were taken from http://supramolecular.org/. The enzyme dosage was determined by measuring the extent to which the AB@TM IDA sensor’s fluorescence intensity recovered at different concentrations of CD73 enzyme (0.01-1.2 μg/mL). The Michaelis-Menten constants (K_M) were obtained by non-linear fitting of the Michaelis-Menten equation via GraphPad Prism 10 software after enzymatic reactions at different substrate concentrations GMP (7.5 - 30 μM). All experiments were performed in 25 mM Tris-HCl buffer solution (pH = 7.4) at 37°C (λ_ex = 590 nm, λ_em = 610 nm).

2.6. Screening the inhibitory activity of TCM and its active ingredients on CD73

The AB@TM IDA sensor could detect changes in GMP conversion rates using fluorescence spectroscopy and thus reveal CD73 enzyme activity. In a solution of 25 mM Tris-HCl buffer (pH = 7.4) containing AB@TM (1/23 μM; λ_ex = 590 nm, λ_em = 610 nm), GMP (20 μM), 0.25 mM MgCl₂ (an activator of CD73), and CD73 (0.3 μg/mL) the inhibitory effect of the TCMs to be tested on CD73 was examined by comparing the degree of fluorescence recovery at 37°C. The inhibitory effect of the active ingredients against CD73 was evaluated using the same method. The active ingredients dose-response curve and associated plots for the inhibition of CD73 activity were analyzed by a continuous fluorescent assay.

2.7. Malachite green assay

The reaction of malachite green molybdate with inorganic phosphate under acidic conditions is capable to generate a green complex that can be detected by a UV spectrophotometer at 620 nm. Since CD73 catalyzes the degradation of GMP to G and inorganic phosphate, the malachite green assay can be used to validate the effect of the screened active ingredients on the enzymatic activity of CD73. Sodium molybdate and malachite green reagents were added to 25 mM tris-HCl buffer (pH = 7.4) containing GMP (20 μM), paeoniflorin (50 μM), and CD73 (0.3 μg/mL), and the change in absorbance was detected. The experiment was performed in a 37°C water bath.

2.8. Molecular docking simulation

Molecular docking simulations were performed using AutoDock (a suite of automated docking tools) to analyze the potential binding modes of the active ingredients to CD73. Firstly, the crystal structure of CD73 (PDB code: 6XUE) was prepared using PyMOL (a molecular visualization system), involving the addition of hydrogen atoms and the removal of water and heteromolecules. Next, the active ingredients were set as ligands and protein structure as a rigid receptor and Autodock was run to obtain the docking results. Those with the lowest docking binding energy were finally selected for molecular docking analysis.

3.1. Optimization of IDA sensor and CD73 catalytic system

To enhance the accuracy and reliability of CD73 inhibitor screening, we optimized the sensitivity of the IDA sensor towards the CD73 enzymatic reaction. Building upon our previous study [24], we evaluated two host-guest reporter pairs: AB@TM and TM complexed with tolonium chloride (TC@TM), for responsiveness to the enzymatic product G. As shown in Figure S1, AB@TM demonstrated significantly higher fluorescence response to G at low concentrations (0.5 µM) than TC@TM. The performance gap between AB@TM and TC@TM increased with rising G concentrations, with AB@TM showing significantly higher signal amplitude. The formation of the AB@TM complex was confirmed by fluorescence quenching of AB upon the addition of TM. Nonlinear fitting of fluorescence titration data revealed a high binding affinity between AB and TM (K_a = 4.26 × 10⁵ M^-1) (Figure S2). Titration with the substrate GMP induced negligible change in AB@TM fluorescence intensity (using AB as the fluorescent guest) (Figure 2a). In contrast, titration with the product G caused a significant fluorescence change in AB@TM under identical conditions (Figure 2b). Based on its strong host-guest affinity and specific response to G (but not GMP), AB was selected as the optimal fluorescent indicator for CD73 activity assays.

Close

Figure 2.

Fluorescence spectra of the AB@TM reporter pair (1 μM/23 μM) in the presence of different concentrations of (a) substrate (GMP) and (b) product (G). Different colored curves represent the fluorescence response of the AB@TM reporter pair to varying concentrations of the guest molecule.

Export to PPT

The association of AB with TM was weaker than that of G with TM, suggesting that the dye within the host can be replaced by the product. To determine the optimal binding ratio of the macrocyclic host to the dye, the AB@TM reporter pair with different binding ratios (70%, 80%, 90% and 95%) was titrated with 1 mM GMP or G, respectively. As shown in Figure S3, the fluorescence response difference to G versus GMP was maximal at a 90% host-guest binding ratio (1 μM/23 μM), demonstrating optimal detection efficacy. Hence, we constructed the AB@TM IDA sensor with AB and TM concentrations fixed at 1 μM and 23 μM, respectively. We examined sensor performance at varying CD73 concentrations. The results showed a gradual increase in fluorescence intensity change magnitude with rising enzyme levels (Figure S4). At 0.6 μg/mL CD73, fluorescence intensity plateaued, indicating saturating substrate conversion capacity. To minimize enzyme usage while ensuring accuracy, we selected 0.3 μg/mL as the optimal CD73 concentration.

3.2. Validation of AB@TM sensor validity and reliability

The kinetic parameters of the CD73 for GMP were calculated using the Michaelis-Menten equation, yielding a K_M value of 24.54 ± 1.85 μM (Figure S5). This value is consistent with those obtained using other assay methods (1-50 μM) [25]. In addition, to accurately monitor CD73’s enzymatic activity, a linear correlation between enzyme concentration and fluorescence intensity is required. As shown in Figure S6, fluorescence signal amplification occurred proportionally to increasing G concentration. A strong linear relationship (R² = 0.9911) was observed at 634 nm within the range of 0-25 μM, confirming the quantitative capability of the assay (limit of detection = 2.53 μM, 3σ/k). The potential of the AB@TM sensor as a screening tool for enzyme inhibitors was further evaluated using the known inhibitor ATP. As shown in Figure S7(a), without CD73, ATP alone induced negligible change in AB@TM fluorescence. When CD73 was present, ATP dose-dependently suppressed enzymatic activity, causing significant fluorescence reduction even at 1 μM (Figure S7b). These experiments demonstrate the sensor’s reliability and sensitivity, supporting its use in screening and evaluating CD73 inhibitors from TCM.

3.3. Effects of 19 types of TCMs on CD73 activity

Firstly, to assess the applicability of the AB@TM sensor in evaluating the effects of TCMs on CD73 activity, it is necessary to detect any interference from the TCM solution being tested on AB@TM fluorescence. We selected TCMs that have been shown to have anti-tumor effects in clinical trials. These include 10 types of TCM decoction pieces (Panax ginseng, Radix Paeoniae Alba, Curcuma longa, Portulaca oleracea, Taraxacum officinale L., Sophora tonkinensis, Smilax glabra Roxb., Saponaria vaccaria, Millettia, and Tripterygium wilfordii) and 10 types of TCM patent medicines (Fufang Banmao (Capsule), Chaihu (Injection), Luohua Zizhu (Dispersible Tablet), YLRK; Dengtaiye (Granules), Rupi Sanjie (Granules), Zhenqi Fuzheng (Granules), Banlangen (Granules), XBJ, and Shuanghuanglian (Oral Liquid)). As shown in Figure S8, all tested TCMs except Taraxacum officinale L. induced negligible AB@TM fluorescence changes, confirming minimal interference and validating the sensor’s reliability for CD73 activity screening in TCM matrices.

The inhibitory effects of these 19 kinds of TCMs on CD73 enzyme activity were further evaluated by AB@TM IDA sensor. Compared with the blank group, Curcuma longa (26.90%), YLRK (17.52%), and XBJ (30.31%) exhibited significant inhibition of fluorescence recovery and demonstrated excellent inhibitory activity (Figure 3). Previous research demonstrated that curcumin, a polyphenolic compound derived from Curcuma longa, significantly reduced CD73 protein expression at specific concentrations [26]. Our results corroborate Curcuma longa’s CD73-inhibitory effect, validating the sensor’s reliability. Consequently, we focused subsequent analysis on TCM patent medicines. Among these, XBJ offers distinct advantages for IDA sensing due to its well-defined composition of clinically validated bioactive compounds. We therefore selected XBJ as a model system to investigate CD73 inhibition by its bioactive components.

Close

Figure 3.

Heat map of CD73 inhibition by 9 kinds of TCM decoction pieces and 10 kinds of patent medicines.

Export to PPT

3.4. Effects of the active ingredients of XBJ on CD73 activity

We selected the five most prominent active ingredients from XBJ based on pharmacological effects, pharmacokinetics, and chemical properties [27,28]. These ingredients are hydroxysafflor yellow A, paeoniflorin, albiflorin, salvianic acid A, and senkyunolide I. No significant disturbance to the fluorescence intensity of the active ingredients at varying concentrations, except at high concentrations of hydroxysafflor yellow A and salvianic acid A (Figure S9). We further screened these components for CD73 inhibitory activity. Salvianolic acid A showed virtually no inhibitory effect (Figure 4a). High concentrations of albiflorin and senkyunolide I, and a medium concentration of hydroxysafflor yellow A, inhibited fluorescence recovery (Figures 4b-d). Increasing paeoniflorin concentration significantly decreased fluorescence recovery (Figure 4e). Thus, paeoniflorin exhibited the strongest CD73 inhibitory effect among the components (Figure 4f). A continuous fluorescent assay confirmed paeoniflorin’s dose-dependent CD73 inhibition, with an IC₅₀ value of 0.2 ± 0.025 μM (Figure 5).

Close

Figure 4.

(a-e) CD73 activity inhibition by active ingredients at different concentrations (n= 3).**p<0.05, *p< 0.1, “ns” represents “no statistically ignificant difference”; (f) Inhibitory effects of the active ingredients.

Export to PPT

Close

Figure 5.

(a) Continuous fluorescent assay for the inhibition of CD73 activity by Paeoniflorin (0.01 - 50 μM). (b) Dose-response curve and associated plot analysis for CD73 activity inhibition by Paeoniflorin were generated.

Export to PPT

3.5. Validation of the inhibitory activity of Paeoniflorin

The malachite green assay, which is a standard method for determining inorganic phosphate levels in solution, was used to evaluate CD73 activity. UV-vis spectroscopic analysis revealed that both the low- and high-concentration paeoniflorin groups exhibited significant increases in absorbance at 620 nm compared to the blank group. Critically, the absorbance values for both paeoniflorin concentrations also exceeded those of the low-concentration ATP group, which was used as a positive control (Figure S10). Reduced phosphate release (implying lower GMP conversion) confirmed paeoniflorin’s CD73 inhibition, validating the AB@TM IDA sensor’s reliability. Paeoniflorin, a monoterpene glycoside from Paeonia species (family Paeoniaceae), exhibits anti-inflammatory, anti-tumor, and immunomodulatory activities [29]. However, its anti-tumor mechanisms remain incompletely characterized [30,31]. To our knowledge, this is the first report demonstrating paeoniflorin’s CD73 inhibitory effect. This offers a novel perspective for elucidating its anti-tumor mechanisms and supports paeoniflorin’s potential as a CD73-targeting therapeutic.

3.6. Molecular docking analysis of each active ingredient with CD73

Molecular docking is a computational technique that predicts ligand-receptor binding conformations and affinities. In this proof-of-concept study, we performed molecular docking using CD73 as the receptor protein and the five most significant active ingredients from XBJ (paeoniflorin, senkyunolide I, hydroxysafflor yellow A, albiflorin, and salvianic acid A) as ligand small molecules (Table S1). Analysis of Figure S11 indicates that hydrogen bonding is the main non-covalent interaction between these ligands and CD73, with hydrophobic and van der Waals interactions also potentially contributing. The docking results revealed varied binding affinities to CD73. While paeoniflorin and senkyunolide I exhibited some binding activity, hydroxysafflor yellow A, albiflorin, and salvianic acid A displayed high binding energies, which is indicative of poor docking interactions with CD73. Specifically, our analysis showed that paeoniflorin forms four hydrogen bonds with the key amino acid residues (SER244, ASN245, PHE247, and ARG356) within the CD73 binding site (Figure 6). This molecular interaction profile demonstrates paeoniflorin’s ability to bind CD73. These findings theoretically validate the paeoniflorin-CD73 ligand-receptor interaction. Consistency between docking predictions and experimental inhibition data confirms the AB@TM sensor’s efficacy.

Close

Figure 6.

3D structure simulated by molecular docking of paeoniflorin-CD73.

Export to PPT