Bimetallic Au–Ag Nanoclusters as a Photo‐Responsive Oxidase‐Like Nanozyme for Antioxidant Detection and Intracellular Redox Analysis
Sanskruti Swain, I‐Hsuan Chou, Bikash C. Mallick, Shu‐Chen Liu, Gin‐Shin Chen, Hsing‐Ying Lin, Chen‐Han Huang
TL;DR
Gold-silver nanoclusters act as light-activated enzyme mimics to detect antioxidants and track redox changes in cells, offering a fast and reagent-free diagnostic tool.
Contribution
First demonstration of bimetallic Au–Ag nanoclusters as photo-responsive oxidase-like nanozymes for antioxidant detection and intracellular redox analysis.
Findings
AuAgNCs catalyze thiamine oxidation under visible light, generating peroxyl radicals for antioxidant detection.
The platform enables rapid, reagent-free detection of antioxidants and measurement of total antioxidant capacity in biofluids and products.
AuAgNCs allow dynamic tracking of intracellular redox changes in normal and steatotic hepatocyte models.
Abstract
Bimetallic nanoclusters (NCs) hold promise as catalytic materials, yet their potential as enzyme mimics remains largely unexplored. Here, we report the first demonstration that gold–silver nanoclusters (AuAgNCs) function as photo‐responsive oxidase‐like nanozymes. Under visible‐light irradiation, AuAgNCs catalyze thiamine oxidation via a Type I photosensitization pathway, predominantly generating physiologically relevant peroxyl radicals (ROO•). The AuAgNCs outperform monometallic analogs, achieving enhanced catalytic efficiency with lower Km and higher v max than previously reported nanozymes. Utilizing these capabilities, we established a rapid (<6 min), reagent‐free platform that enables: (i) selective detection of diverse physiological antioxidants, (ii) quantitative assessment of total antioxidant capacity (TAC) in simulated biofluids and consumer products, and (iii) dynamic…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7
FIGURE 8
FIGURE 9| Nanozyme type | Metal composition | Substrate |
|
| References |
|---|---|---|---|---|---|
| Monometallic | Ti | TMB | 0.107 | 1.56 × 10−7 | [ |
| V | OPD | 0.47 | 4.2 × 10−5 | [ | |
| Au | TH | 0.159 | 3.79 × 10−3 | [ | |
| C | TMB | 0.22 | — | [ | |
| Bimetallic MOF | Zn, Ru | TMB | 0.165 | 1.39 × 10−7 | [ |
| Bimetallic | Pd, Au | OPD | 0.34 | 2.97 × 10−7 | [ |
| Ti, Au | TMB | 0.026 | 1.36 × 10−8 | [ | |
|
|
|
|
|
|
- —the National Science and Technology Council10.13039/100020595
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Nanomaterials in Catalysis · Nanocluster Synthesis and Applications · Nanoplatforms for cancer theranostics
Introduction
1
Reactive oxygen species (ROS), including superoxide (O_2_•^−^), hydroxyl radicals (•OH), and peroxyl radicals (ROO•), are central regulators of cellular homeostasis. At physiological levels, ROS mediate essential signaling pathways, whereas excessive accumulation damages DNA, proteins, and lipids, driving cancer, neurodegeneration, and metabolic disorders such as non‐alcoholic fatty liver disease (NAFLD) [1, 2]. Accurate monitoring of redox balance is therefore critical for understanding disease progression and guiding therapeutic interventions. Low‐molecular‐weight antioxidants, such as uric acid, vitamin E, and glutathione, play key roles by directly scavenging ROS and interrupting the radical propagation cascade. These molecules are the primary contributors to the total antioxidant capacity (TAC), which is a widely used measure of overall antioxidant potential [3, 4]. Beyond biological systems, TAC measurement is commonly employed for quality control of the antioxidant content in consumer products, such as dietary supplements. However, conventional antioxidant assays such as Ferric reducing antioxidant power (FRAP), 2,2‐diphenyl‐1‐picrylhydrazyl (DPPH), Trolox equivalent antioxidant capaciy (TEAC), Folin–Ciocalteu (F–C), and oxygen radical absorbance capacity (ORAC) suffer from fundamental drawbacks like lack of specificity for physiological antioxidants, reliance on unstable reagents or synthetic radicals, long assay times, and high susceptibility to interference. For example, FRAP is insensitive to physiologically important thiol‐containing antioxidants like glutathione, while ORAC requires long incubation and suffers from probe instability. The F–C assay is highly susceptible to interference from many non‐antioxidant reducing agents such as creatinine and sugars. Similarly, DPPH and TEAC employ nonphysiological radicals absent in mammalian systems, limiting biological relevance [5]. These limitations highlight the need for platforms that combine biological relevance, rapid response, and broad applicability.
Nanozymes, that are nanomaterials with enzyme‐like catalytic activities, have recently emerged as promising tools for antioxidant detection. Their unique catalytic properties enable sensitive and rapid measurement of antioxidants, offering potential advantages over conventional assays [6, 7]. Among various types of nanozymes, peroxidase‐like nanozymes have been widely used; however, their reliance on unstable hydrogen peroxide (H_2_O_2_) remains a major limitation. In contrast, photo‐responsive oxidase‐like nanozymes offer notable advantages, as they utilize molecular oxygen and enable precise, light‐mediated spatiotemporal control of catalytic activity without the need for H_2_O_2_. Light, as an external stimulus, enables controllable activation under mild, reagent‐free conditions (i.e., without externally added reagents such as H_2_O_2_, as ROS are generated directly under light activation and thiamine conversion occurs in neutral medium), offering both spatiotemporal precision and a sustainable platform for catalytic applications [8, 9]. Previous studies have employed such types of nanozymes to detect antioxidants in a variety of contexts, such as glutathione detection in cells (e.g., photo‐sensitized metal organic framework [10]), TAC assessment in vitamin C tablets and commercial beverages (e.g., gold nanoclusters [11]), biofluids (e.g., silver nanoclusters together with graphene oxide [12]), fruit extracts (e.g., carbon dots [13]), and plant‐derived compounds like gallic acid, tannic acid and caffeic acid (e.g., 1D covalent organic frameworks [14]). However, these investigations typically focus on a single sample type per study, which may limit the broader applicability and translational potential of nanozyme‐based antioxidant detection platforms.
Nanoclusters represent an emerging class of ultra‐small materials (diameter below 2 nm) with unique optical and catalytic properties. Owing to their high surface‐to‐volume ratio, discrete energy states, and molecular‐like behavior, nanoclusters offer exciting potential for applications in catalysis, sensing, and biomedical research [15]. While gold nanoclusters (AuNCs) have been explored for their light‐activated oxidase‐like catalytic behavior [11], silver nanoclusters (AgNCs) typically require hybrid systems (e.g., with graphene oxide) to achieve comparable activity [12], likely because of their intrinsically lower catalytic efficiency. Although bimetallic AuAg nanoclusters (AuAgNCs) have been previously studied as fluorescent probes for sensing applications [16, 17], their catalytic potential as photo‐responsive oxidase‐like nanozymes remains unexplored. Here, we demonstrate for the first time that AuAgNCs function as photo‐responsive oxidase‐like nanozymes, catalyzing thiamine oxidation under visible light through a Type I photosensitization pathway dominated by peroxyl radicals (ROO•), as illustrated in Figure 1. The AuAgNCs exhibit synergistically enhanced catalytic performance over monometallic analogs, achieving lower K_m_ and higher v max than previously reported photo‐responsive oxidase‐like nanozymes. Building on this mechanistic insight, we establish a unified and reagent‐free platform for: (i) selective detection of multiple physiologically relevant antioxidants, (ii) quantitative TAC evaluation in simulated biofluids and commercial products, and (iii) tracking dynamic intracellular redox changes under oxidative stress, recovery, and disease‐mimicking steatosis, achieving a total assay time of under 6 min. Particularly, the cellular component extends beyond static antioxidant detection, exploring oxidative stress progression, recovery, antioxidant intervention (e.g., vitamin C), and disease‐relevant redox alterations, offering a broader biological perspective. This work expands the functional landscape of bimetallic nanoclusters and introduces a versatile redox‐active nanozyme system, offering new opportunities for antioxidant assessment, biomedical diagnostics, and redox biology research.
Conceptual scheme illustrating visible‐light‐induced ROS generation by photo‐responsive oxidase‐like AuAgNCs, leading to oxidation of non‐fluorescent thiamine into fluorescent thiochrome.
Experimental Section
2
Materials are described in the Supporting Information.
Preparation and Characterizations of BSA‐Stabilized Gold−Silver Bimetallic Nanoclusters
2.1
The BSA‐stabilized gold–silver bimetallic nanoclusters (AuAgNCs) were synthesized based on a previously reported method [16] with slight modifications reported in Supporting Information (Figure S1). For HRTEM analysis, multiple regions were examined across independently prepared grids to ensure reliability. HRTEM imaging was performed using a JEOL JEM‐2100 microscope operated at 80–200 kV with a LaB_6_ electron gun. Samples were prepared by drop‐casting the dispersion onto carbon‐coated copper grids, followed by overnight solvent evaporation in a dehumidifier. For XPS analysis, measurements were carried out using a Thermo VG‐Scientific Sigma Probe instrument. The system offers a detection limit of approximately 0.1 atomic % and an adjustable X‐ray spot size ranging from 15 to 400 μm. An Al Kα monochromatic X‐ray source (1486.6 eV) was used for excitation. All spectra were collected under high‐vacuum conditions, and only data with optimal signal‐to‐noise ratios were used for interpretation.
Photocontrolled Nanozyme Activity of AuAgNCs
2.2
To evaluate the photo‐responsive oxidase‐mimicking activity, a 500 μL reaction mixture was prepared containing 245 μL phosphate buffer (pH 7.4), 5 μL of AuAgNCs (final concentration: 4 μM), and 20 μL of TH (final concentration: 0.6 mM), and double‐distilled water (DDW). The mixture was irradiated for 5 min using a 530 nm LED light source (FWHM: 33 nm, 3.2 W), and fluorescence emission was measured at 440 nm using a Synergy H1 Microplate Reader (BioTek). All catalytic reactions were conducted in sealed vials under LED irradiation at room temperature. Unless otherwise specified, 200 μL of each reaction mixture was transferred to black 96‐well plates for fluorescence measurement.
Effect of ROS Scavengers on AuAgNCs Activity
2.3
To assess the involvement of reactive oxygen species (ROS) in the catalytic activity of AuAgNCs, the assay was conducted in the presence of various ROS scavengers: tryptophan (25 mM), catalase (100 U/mL), mannitol (5 mM), p‐benzoquinone (1 mM), EDTA‐2Na (1 mM), and Trolox (100 μM). Each scavenger (50 μL) was added to the standard reaction mixture, briefly vortexed, and irradiated under 530 nm LED for different time intervals. Fluorescence was subsequently recorded at 440 nm.
Steady State Kinetic Analysis of AuAgNC‐Catalyzed TH Oxidation
2.4
For kinetic analysis, reaction mixtures were prepared with 245 μL phosphate buffer (pH 7.4), 5 μL of AuAgNCs (4 μM), and varying concentrations of TH (0.02–0.6 mM), adjusted to 500 μL with DDW. Samples were mixed thoroughly and irradiated under 530 nm LED light. Fluorescence emission at 440 nm was recorded every 2 min. For each substrate concentration [S], the average rate of increase in fluorescence at 440 nm (v) was determined and plotted against [S] to generate a Michaelis–Menten curve. The data were fitted to the Michaelis–Menten equation as follows: v = (ν max × [S])/(K_m_ + [S]), where v is the initial reaction rate, ν max is the maximal reaction rate, and K_m_ represents the Michaelis constant. The kinetic parameters ν max and K_m_ were calculated according to the Lineweaver–Burk plot [18].
Detection of Physiological Antioxidants Using AuAgNC Nanozyme Assay
2.5
The photocontrolled nanozyme assay was used to evaluate the fluorescence response of six physiologically relevant antioxidants: ascorbic acid (AA), glutathione (GSH), uric acid (UA), cysteine (Cys), potassium thiocyanate (KSCN), and Trolox. Each antioxidant was tested independently at varying concentrations in a standard 500 μL assay mixture consisting of 245 μL phosphate buffer (pH 7.4), 5 μL of AuAgNCs (4 μM), 20 μL of TH (0.6 mM), and 10 μL of antioxidant solution. After mixing, samples were irradiated with 530 nm LED light for 5 min, and fluorescence was measured at 440 nm.
TAC Evaluation in Simulated Saliva and Human Saliva Spike‐and‐Recovery Experiments
2.6
To mimic the antioxidant composition of human saliva, six simulated saliva samples were formulated containing physiologically relevant concentrations of key antioxidants: potassium thiocyanate (KSCN), glutathione (GSH), uric acid (UA), ascorbic acid (AA), cysteine (Cys), homocysteine (Hcy), and Trolox. The concentrations of the constituents were proportionally varied around a central TAC reference value of ~1.19 mM based on a previously reported artificial saliva matrix [19]. The complete composition is reported in Table S1. Antioxidant activity was assessed using the standard assay protocol, with 10 μL of each simulated saliva sample serving as the antioxidant source. For spike‐and‐recovery experiments, human saliva samples were centrifuged at 13,000 rpm for 45 min to remove particulates and then diluted 100‐fold (v/v) with double‐distilled water (DDW). Known concentrations of ascorbic acid were spiked into the diluted saliva samples as well as into corresponding buffer (non‐saliva) controls. The antioxidant concentrations were quantified using an external calibration curve. Recovery (%) was calculated as the ratio of the measured antioxidant concentration in the saliva matrix to that obtained in buffer samples spiked at the same nominal concentration.
Cell Culture and Treatments
2.7
HepG2 cells were cultured in minimum essential medium (MEM), modified with 1% non‐essential amino acids (NEAA), 1% sodium pyruvate, 1% penicillin‐streptomycin (10,000 U/mL), and 10% FBS. The cell line was maintained in an incubator at 37°C in a humidified atmosphere containing 5% CO_2_. To induce mild oxidative stress, cells were treated with 5 or 10 µM hydrogen peroxide (H_2_O_2_) for 90 min. For direct analysis post‐stress, cells were lysed immediately. For recovery experiments, following H_2_O_2_ treatment, cells were incubated for an additional 24 h in complete medium containing 10% FBS prior to lysis. For L‐ascorbic acid 2‐phosphate trisodium (AAP) experiments, cells were grown to ~75% confluence and treated with AAP (50–200 µM) for 24 h, followed by H_2_O_2_ exposure or direct lysis, as per requirement.
Cytotoxicity Assay
2.8
Cell viability following H_2_O_2_ or AAP treatment was assessed using the WST‐1 assay. HepG2 cells were cultured in 24‐well plates and treated accordingly. WST‐1 reagent was added to the wells at a 1:10 dilution in culture medium and incubated at 37°C for 1 h. Absorbance of the resulting formazan product was measured at 440 nm.
Preparation of Cell Lysates and Cellular Antioxidant Detection
2.9
Culture medium was aspirated, and cells were washed with phosphate‐buffered saline (PBS). Cold lysis buffer (0.15 M NaCl, 5 mM EDTA, 1% Triton X‐100, 10 mM Tris‐HCl, 1% protease inhibitor cocktail, and DDW) was added, and cells were scraped on ice. Cell disruption was performed by low‐power ultrasonication for 1 min at low temperature, followed by centrifugation at 12,000 rpm for 20 min at 4°C. Supernatants were collected as cell lysates and kept on ice throughout the experiment. For antioxidant detection, 100 µL of cell lysate was mixed with 5 µL of AuAgNCs (final concentration: 4 µM), 20 µL of TH (final concentration: 0.6 mM), and phosphate buffer (pH 7.4) to a final volume of 500 µL. The mixture was irradiated with a 530 nm LED for 5 min, and the fluorescence intensity of thiochrome was measured at 440 nm.
In Vitro Steatosis Model and Treatments
2.10
Oleic acid (OA) and palmitic acid (PA) were each dissolved separately in isopropanol to 100 mM and then combined at a 2:1 (v/v) ratio to prepare a 100 mM mixed fatty acid (FA) mixture stock solution. The FA mixture (oleic acid and palmitic acid, 2:1 molar ratio) was then diluted in serum‐free medium containing 1% isopropyl alcohol (IPA), to a final FA concentration of 0.4 mM and warmed at 37°C up to 30 min before use. HepG2 cells were seeded and allowed to attach for at least 36 h, then treated with the FA mixture for 24 h to induce lipid accumulation. Lipid accumulation was assessed by Oil Red O (ORO) staining. Cells were washed with PBS, fixed with 10% formalin, and stained with ORO working solution. Excess stains were removed by washing with DDW, and lipid‐rich regions were observed under a microscope. Detailed information regarding the ORO staining protocol and preparation is provided in the Supporting Information. Following steatosis induction, cells were treated with H_2_O_2_ for 90 min, followed by 24 h incubation in either 10% serum medium or serum‐free medium, as required, before lysis.
Statistical Analysis
2.11
Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using OriginPro 8.5 (OriginLab Corporation, USA). Statistical significance between two independent groups was assessed using an unpaired two‐tailed Student's t‐test, where applicable, with significance defined as p < 0.05. Significance levels are indicated in the corresponding figure captions (***p < 0.001). All experiments were independently repeated at least three times to ensure reproducibility.
Results and Discussion
3
Synthesis of AuAg Nanoclusters and Selection of the Optimal Nanozyme Candidate
3.1
To investigate the potential nanozyme activity of bimetallic AuAg nanoclusters, a series of compositions were synthesized by systematically varying the Au:Ag molar ratio (1:0, 8:1, 6:1, 4:1, 2:1, 1:1, 1:2, and 0:1) through controlled adjustment of Au and Ag precursor concentrations. Gold and silver were specifically chosen because both metals are well known for their strong fluorescence properties, yet their combination has not been previously explored in the context of nanozyme activity. As shown in Figure 2A, the nanocluster with a 4:1 Au:Ag ratio exhibited the highest photocatalytic activity, which was enhanced relative to that of monometallic AuNCs (1:0). The 8:1 and 6:1 compositions also showed superior activity compared to AuNCs, while clusters with higher Ag content (1:2 and 0:1) exhibited the lowest activity, likely due to the lower stability and reduced surface reactivity of AgNCs under catalytic conditions [20, 21]. The enhanced performance at the 4:1 ratio indicates an optimal synergy between Au and Ag atoms; however, further mechanistic studies would be required to confirm the exact underlying mechanism. Similar enhancements have also been reported in other Au‐based bimetallic nanozymes, where alloying modulates surface charge distribution and electronic structure, increasing catalytic efficiency [22]. Au‐rich AuAgNCs (1:0–4:1) exhibited red fluorescence (~680–690 nm), while increasing Ag content decreased emission intensity, consistent with Ag‐induced quenching [16, 23] (Figure S2). These fluorescence changes are independent of the catalytic activity reported above. Based on this initial screening of nanozyme activity across various Au:Ag molar ratios, the 4:1 Au:Ag composition was identified as the most catalytically active; hence, it was selected for detailed characterization and subsequent applications in this work.
(A) Photocatalytic screening of AuAgNCs prepared with different Au:Ag molar ratios (1:0 to 0:1) under LED light. (B) HRTEM image of monodispersed AuAgNCs with a 4:1 Au:Ag composition. (C) Size distribution histogram of the AuAgNCs showing an average particle diameter of 1.17 nm. (D–F) High‐resolution XPS spectra of the AuAgNCs: (D) Ag 3d region showing peaks at 367.11 and 373.34 eV, and at 368.85 and 374.78 eV; (E) Au 4f region showing peaks at 83.19 and 87.02 eV, and at 84.82 and 88.63 eV; (F) S 2p region showing peaks at 161.19, 162.84, and 165.91 eV.
Characterization of the AuAgNCs
3.2
The high‐resolution transmission electron microscopy (HRTEM) image of the selected AuAgNCs (4:1 Au:Ag) indicates monodispersed nanoclusters with an ultra‐small size and a mean diameter of 1.17 nm (Figure 2B,C). No core–shell structure or distinct FCC lattice fringes were observed in the AuAgNCs. The absence of FCC lattice fringes is likely due to their ultra‐small size, which results in loss of long‐range crystallinity. This loss of crystalline or amorphous structure was also supported by the XRD results which displayed a broad peak around 18°, characteristic of ultra‐small size nanoclusters presented in Figure S3. FTIR analysis revealed slight amide band shifts and decreased band intensities, indicating subtle secondary‐structure changes in BSA during AuAgNC formation. These changes support its stabilizing role, in line with previous reports on protein–metal nanoclusters [24]. X‐ray photoelectron spectroscopy (XPS) was employed to investigate the surface elemental composition and the chemical states of the constituent elements in the AuAgNCs. The XPS survey scan (Figure S5) confirmed the presence of O, N, C, and S originating from BSA, along with gold and silver signals from the metal core, indicating successful formation of BSA‐stabilized AuAgNCs. High‐resolution XPS analysis provided insights into the oxidation states and binding environments of the metal components in AuAgNCs. The Ag 3d spectrum (Figure 2D) exhibited peaks at 367.11 eV (Ag^0^) and 368.85 eV (Ag^+^) for Ag 3d_5/2_ and at 373.34 eV (Ag^0^) and 374.78 eV (Ag^+^) for Ag 3d_3/2_, indicating the coexistence of zero‐valent and oxidized silver states. Similarly as shown in Figure 2E, the Au 4f spectrum displayed two sets of peaks: Au 4f_7/2_ with a dominant peak at 83.19 eV (Au^0^) and a minor peak at 84.82 eV (Au^+^) and Au 4f_5/2_ with corresponding peaks at 87.02 eV (Au^0^) and 88.63 eV (Au^+^), confirming the presence of both metallic and partially oxidized gold species [25]. The high‐resolution S 2p XPS spectrum (Figure 2F) exhibited multiple peaks: 161.19 and 162.84 eV, attributed to S 2p_3/2_ components of sulfur bound to gold (Au–S) and silver (Ag–S), respectively [26, 27]. The higher‐energy peak at 165.91 eV likely corresponds to unbound thiol groups (–SH) [25], indicating the partial involvement of BSA thiols in metal coordination. UV–Vis absorption analysis showed a slight redshift of the characteristic BSA peak upon AuAgNC formation, suggesting subtle with the metal core [28]. Importantly, no surface plasmon resonance peak was observed, confirming the absence of larger nanoparticles.
Photocatalytic Activity of AuAgNCs and Reaction Optimization
3.3
As shown in Figure 3A, visible‐light irradiation significantly enhanced the catalytic conversion of non‐fluorescent thiamine (TH) to its fluorescent oxidized form, thiochrome (TC), in the presence of AuAgNCs, as indicated by a distinct emission peak near 440 nm. In contrast, no catalytic activity was observed when the reaction was conducted without the light irradiation, despite the presence of both TH and AuAgNCs, highlighting the essential role of light in the reaction. Furthermore, when TH was irradiated in the absence of AuAgNCs, no fluorescence response was detected, confirming the necessity of the nanozyme catalyst (AuAgNCs). Given this light dependence, we systematically investigated the effect of irradiation wavelength on the photo‐induced activity of AuAgNCs by employing LEDs emitting at wavelengths across the UV, visible, and NIR regions (365 to 940 nm). As shown in Figure 3B, the catalytic activity was notably higher in the 420–530 nm range, with 530 nm producing the highest catalytic efficiency; hence, it was selected for all experiments. Next, to assess the light‐dependent regulation of catalytic activity, the irradiation source was periodically switched on and off every 2 min. This led to a staircase‐like increase in fluorescence, with each “light on” phase triggering a rise in signal, while no further change was observed during “light off” intervals, indicating a photo‐controlled catalytic response from the AuAgNCs as displayed in Figure 3C. These results demonstrate that the catalytic activity of AuAgNCs can be externally regulated with high precision through light, offering spatiotemporal control over the reaction process. Additionally, the effect of irradiation time (5–90 min) was also evaluated (Figure 3D), showing a steady increase in fluorescence intensity with a strong linear correlation to irradiation duration (R^2^ = 0.99; inset). This linearity indicates proportional TC formation with light exposure, demonstrating robustness, stability, and sustained photoactivation of the nanozyme without catalytic deactivation. An irradiation time of 5 min was chosen to ensure reliable signal within the linear range while minimizing assay duration.
(A) Fluorescence emission peak at 440 nm indicates catalytic conversion of non‐fluorescent thiamine (TH) to fluorescent thiochrome (TC) in the presence of AuAgNCs and light (red). No fluorescence peak observed in the absence of light (blue) or in the absence of AuAgNCs (black). (B) Photocatalytic activity of AuAgNCs under irradiation at wavelengths ranging from 365 to 940 nm, with maximum catalytic efficiency observed at 530 nm. Error bars represent standard deviations (n > 3). (C) Light‐controlled catalytic activity of AuAgNCs under 2 min light on/off cycles. Fluorescence intensity shows a staircase‐like increase with signal rising during “light on” phases and remaining stable during “light off” intervals, indicating precise photo‐regulation of the reaction. (D) Fluorescence intensity as a function of irradiation time (5–90 min), showing a strong linear increase (R2 = 0.99), indicating stable and sustained photoactivation of AuAgNCs without catalyst deactivation. Error bars represent standard deviations (n > 3).
To evaluate the influence of substrate and catalyst concentrations on photocatalytic performance, varying concentrations of TH (0.02‐2 mM) and AuAgNCs (4‐40 µM) were examined. As shown in Figure S7, catalytic activity increased with TH concentration up to 0.6 mM, beyond which it declined, suggesting “substrate inhibition” at higher levels [29]; hence, 0.6 mM was selected for subsequent experiments. In contrast, at the fixed TH concentration of 0.6 mM, the catalytic activity increased steadily with rising AuAgNC concentration (Figure S8), and the lowest effective concentration (4 µM) was chosen to minimize nanozyme consumption and improve cost‐effectiveness. Then, the effects of temperature, pH, and various buffer medium were tested (Figure S9–S11). As expected, when the pH of the reaction mixture was increased from acidic to alkaline, the fluorescence intensity increased because TH is easily oxidized in alkaline conditions [30] (Figure S9). Also, when the reaction temperature was increased, catalytic activity followed an upward trend with increasing temperature (5–65°C), indicating enhanced reaction efficiency at elevated temperatures (Figure S10). However, neutral pH (7) and room temperature (~25°C) were selected for all subsequent experiments to simplify the procedure, maintain operational consistency, and ensure compatibility with ambient conditions relevant for practical applications. As shown in Figure S11, various buffer systems, including acetate (HAc), sodium acetate (NaAc), Tris‐acetate (Tris‐HAc), and PBS were evaluated for their influence on catalytic activity. PBS provided the highest activity and, being a physiologically relevant medium, was selected as the reaction medium throughout the study. To further elucidate the underlying catalytic mechanism, we investigated key factors influencing the reaction, including reactive intermediates, oxygen dependence, and kinetic parameters.
Mechanistic and Kinetic Studies of AuAgNC Catalysis
3.4
To assess the role of oxygen in photocatalytic TH oxidation, AuAgNC activity was measured under varied oxygen conditions. Nitrogen purging reduced activity by ~50%, while oxygen supplementation enhanced it up to twofold, confirming that oxygen is essential as a reactant and precursor for ROS generation (Figure 4A). To identify the reactive oxygen species (ROS) driving TH oxidation, targeted scavengers were used to selectively quench possible ROS involved. Trolox for peroxyl radicals (ROO•), mannitol for hydroxyl radicals (•OH), catalase for hydrogen peroxide (H_2_O_2_), tryptophan for singlet oxygen (^1^O_2_), p‐benzoquinone for superoxide anions (O_2_•^−^), and EDTA for photogenerated holes (h^+^) were added individually to the reaction system. As shown in Figure 4B, Trolox completely suppressed TH oxidation, confirming ROO• as the primary reactive species. Marked inhibition by EDTA and p‐benzoquinone indicates the involvement of h^+^ and O_2_•^−^, while mannitol and catalase showed moderate effects, suggesting auxiliary roles for •OH and H_2_O_2_. Minimal inhibition by tryptophan indicates negligible involvement of ^1^O_2_. Based on these observations, we propose that photoexcited AuAgNCs abstract a hydrogen atom from TH at the thiazole C2 position [31], generating a carbon‐centered radical ((Equation 1)). This behavior is supported by a recent report demonstrating that photoactivated metal nanoclusters can abstract hydrogen atoms [32]. The resulting radical reacts with molecular oxygen to produce a peroxyl radical (ROO•), consistent with the fundamental principle that peroxyl radicals are formed when carbon‐centered organic radicals react with molecular oxygen [33]. The ROO• species were identified as the primary oxidants responsible for converting TH to fluorescent TC ((Equation 2)).
(A) Fluorescence intensities representing the photocatalytic activity of AuAgNCs under different oxygen conditions: ambient conditions (blue), nitrogen purging (gray), and oxygen supplementation (green). Error bars represent standard deviations (n = 3). (B) Fluorescence responses obtained in the presence of various reactive oxygen species (ROS) scavengers during the photocatalytic oxidation of thiamine (TH) by AuAgNCs under light irradiation. Error bars represent standard deviations (n = 3). (C) Michaelis–Menten plot for the photocatalytic oxidation of thiamine by AuAgNCs under optimized conditions. The inset shows the corresponding Lineweaver–Burk plot. Kinetic parameters determined from the fit include Km = 0.095 mM and v max = 4.76 × 10−³ M s−¹. Error bars represent standard deviations (n = 3). (D) Inter‐batch reproducibility across five independently synthesized batches (upper panel) and intra‐batch repeatability from five replicate measurements on the same batch (lower panel), showing minimal variability in the catalytic activity of the AuAgNCs nanozyme. Error bars represent standard deviations (n = 5). Statistical significance was determined using an unpaired two‐tailed Student's t‐test. (E) Proposed mechanism for the oxidase‐like photocatalytic oxidation of thiamine by AuAgNCs under visible‐light irradiation. Under visible light irradiation, photoexcited AuAgNCs abstract a hydrogen atom from thiamine, generating a carbon‐centered radical, which reacts with molecular oxygen to form a peroxyl radical (ROO•) converting TH to TC.
Consistent with classical photocatalytic systems, where photogenerated electrons (e^−^) reduce molecular oxygen to superoxide (O_2_•^−^), and photogenerated holes (h^+^) oxidize water to form hydroxyl radicals (•OH) [34], our observation of O_2_•^−^ and •OH involvement, although playing auxiliary roles, aligns with these reaction pathways. Collectively, these results support a Type I photosensitization pathway, commonly observed in photoactivated systems [35, 36]. These findings explain the oxidase‐like photocatalytic behavior of AuAgNCs under visible‐light irradiation. The proposed mechanistic principle is illustrated in Figure 4E. Importantly, ROO• identified here as key species followed by O_2_•^−^ are physiologically relevant radicals implicated in oxidative stress and various disease states unlike the synthetic radicals used in conventional antioxidant assays (such as DPPH• and ABTS•^+^). This highlights the AuAgNC‐based system as a more biologically relevant platform for studying antioxidant activity.
Next, to quantitatively evaluate the catalytic performance of the AuAgNCs, steady‐state kinetic parameters were determined by fitting the experimental measurements to the Michaelis–Menten model (Figure 4C), followed by double‐reciprocal Lineweaver–Burk analysis (Figure 4C; inset) under optimized conditions. The key kinetic parameters obtained include a Michaelis–Menten constant (K_m_) of 0.095 mM, indicating strong substrate affinity, as K_m_ represents the substrate concentration at which the reaction rate reaches half of its maximum, and a maximum reaction velocity (v max) of 4.76 × 10^−3^ M s^−1^, reflecting high catalytic turnover at saturating substrate levels. A comparative analysis with previously reported photo‐responsive oxidase nanozyme systems (Table 1) highlights the improved performance of AuAgNCs, as evidenced by its lower K_m_ and higher v max, indicating enhanced substrate binding and catalytic turnover. To assess the reproducibility and repeatability of the photocatalytic system, both inter‐batch and intra‐batch precision studies were performed. Five independently synthesized AuAgNCs batches exhibited consistent catalytic performance with minimal variation, demonstrating excellent inter‐batch reproducibility (Figure 4D; upper panel). Five replicate measurements of the same batch within a single day yielded minimal variability, indicating high intra‐batch repeatability (Figure 4D; lower panel). These results highlight the robustness, and batch‐to‐batch reliability of the synthesized AuAgNCs nanozyme.
TABLE 1: Comparison of the kinetic parameters (Km and ν max) of the present nanozyme and previously reported nanozymes, along with their metal composition and substrate.
Versatile and Rapid Detection of Key Physiological Antioxidants
3.5
To evaluate the versatility and physiological relevance of our assay, we tested a panel of representative antioxidants and antioxidant‐like compounds naturally present in human body, including glutathione, cysteine, trolox (a water‐soluble vitamin E analog), homocysteine, ascorbic acid, thiocyanate, and uric acid. As shown in Figure 5A–H, all tested antioxidants exhibited a clear dose‐dependent decrease in fluorescence intensity, indicating effective radical scavenging. The corresponding fluorescence spectra and inset calibration plots revealed linear responses within the tested ranges. The linear ranges and the limit of detection (LOD) for each antioxidant are reported in Figure 5H.
Fluorescence responses of the AuAgNC‐based assay to varying concentrations of (A) glutathione (GSH), (B) cysteine, (C) Trolox, (D) homocysteine, (E) ascorbic acid (AA), (F) thiocyanate, and (G) uric acid. Insets show the corresponding calibration curves for each antioxidant within the tested concentration ranges. Error bars represent standard deviations (n = 3). (H) Summary of linear ranges and limits of detection (LOD) for all tested antioxidants.
These results confirmed the assay's broad reactivity and quantitative capability toward physiologically relevant antioxidants. To assess the assay's selectivity and resistance to interference, a wide range of biologically relevant non‐antioxidant species were tested at physiological concentrations (Table S1), as shown in Figure 6. These included common ions (Na^+^, K^+^, Ca^2+^, Mg^2+^, Cl^−^, HCO_3_ ^−^, NO_3_ ^−^, Fe^3+^, CH_3_COO^−^, NH_4_ ^+^), amino acids (arginine, lysine, glutamic acid, glycine), and metabolites (glucose, urea, creatinine, lactate, citric acid, cholesterol), along with mixtures representing metal ions (Ca^2+^, Mg^2+^, Fe^3+^), electrolytes (Na^+^, K^+^, Ca^2+^, Mg^2+^, Cl^−^, HCO_3_ ^−^, PO_4_ ^3−^), and metabolic intermediates (glucose, lactate, urea, creatinine). None of these interferents, tested individually or in combination, produced significant changes in fluorescence, confirming the assay's high selectivity toward antioxidants. To further assess robustness under complex conditions, glutathione was examined in three settings: alone, with a single interferent (Na^+^), and with multiple interferents (Na^+^, K^+^, Cl^−^, PO_4_ ^3−^, glucose, and urea). In all cases, glutathione was consistently detected, and its fluorescence‐quenching response remained unaffected by the presence of interfering species signifying the assay's high specificity, operational stability, and suitability for use in chemically diverse environments. The high selectivity of the proposed AuAgNC‐based assay is rooted in its specific reaction mechanism. The observed suppression in the presence of antioxidants directly supports the proposed ROS‐mediated reaction pathway, in which co‐existing antioxidants scavenge ROS and hinder TH oxidation to TC. Analytically, this results in a lower TC readout in highly reducing matrices. This effect is well‐known in oxidation‐based assays and can be managed experimentally by sample dilution or standard‐addition calibration. The extent of suppression depends on the antioxidant type and concentration, as also reflected in Figure 5.
Selectivity and interference study of the AuAgNC‐based antioxidant assay. Blank, individual non‐antioxidant species: sodium, potassium, calcium, iron, magnesium, ammonium, bicarbonate, nitrate, acetate, arginine, lysine, glutamic acid, glycine, cholesterol, glucose, creatinine, lactate, urea, citric acid, (M1) electrolyte mix (Na+, K+, Ca2+, Mg2+, Cl−, HCO3−, PO4 3−), (M2) metal ions mix (Ca2+, Mg2+, Fe3+), (M3) Mmetabolite mix (glucose, lactate, urea, creatinine), (A1) single antioxidant (glutathione), (A2) antioxidant (glutathione) + single interferent (sodium), and (A3) antioxidant (glutathione) + multiple interferents (Na+, K+, Cl−, PO4 3−, glucose, urea). Error bars represent standard deviations (n > 3). Statistical significance was determined using an unpaired two‐tailed Student's t‐test *** indicates p < 0.001.
To compare against standard methods, we independently tested Trolox and glutathione using two widely used antioxidant assays, FRAP and ORAC (Figure S12 and S13). As expected, FRAP showed a strong response to Trolox but failed to detect glutathione, consistent with its known limitation in measuring thiol‐based antioxidants [5]. ORAC assay detected both compounds; however, the response for glutathione was substantially weaker than that of Trolox, consistent with previous reports highlighting variability and underestimation of thiol‐containing antioxidants. In contrast, our system reliably detects both thiol and non‐thiol antioxidants under mild conditions, highlighting its practical advantages for physiologically relevant antioxidant screening.
Validation of TAC Assay in Simulated, Biological, and Commercial Samples
3.6
To demonstrate the practical versatility of our assay, we evaluated total antioxidant capacity (TAC) in two sets of real‐world samples: (1) synthetic saliva matrices and (2) commercial antioxidant‐containing products. Six synthetic saliva formulations were prepared to simulate the physiological antioxidant profile of human saliva (Table S2). As shown in Figure 7A, our nanozyme‐based assay produced a clear concentration‐dependent response across the six formulations exhibiting excellent linearity (R^2^ = 0.99). To validate the assay's accuracy, the same samples were also analyzed using the standard ORAC assay which also showed strong linearity (R^2^ = 0.99) (Figure 7B). As shown in Figure 7C, a strong correlation (R^2^ = 0.99) between both methods confirmed the quantitative reliability of our assay in biologically relevant matrices. We further extended the application to commercial antioxidant products, including three vitamin C supplement tablets and three packaged fruit juice samples. As shown in Figure 7D, both products yielded comparable TAC values, expressed in vitamin C equivalents, which were consistent with their reported antioxidant content. Here, ascorbic acid was selected as the reference antioxidant due to its extensive use as a calibration standard in antioxidant assays and its well‐defined redox properties. Using “ascorbic acid equivalents” enables consistent comparison across different systems and studies. This standardization improves both the analytical relevance and the biological interpretability of the antioxidant results. Collectively, these results demonstrate that our assay enables reliable, quantitative assessment of antioxidant capacity across diverse complex samples, from simulated biological fluids to consumer nutritional products, under mild, rapid, and reagent‐free conditions.
(A) Dose‐dependent fluorescence responses of the nanozyme‐based assay measured across six synthetic saliva formulations. Linear fits are shown (R² = 0.99). (B) ORAC assay results for the same synthetic saliva samples with corresponding linear fits (R² = 0.99). (C) Correlation plot comparing values obtained from the nanozyme‐based assay and the ORAC method (R² = 0.99). Error bars represent standard deviations (n > 3). (D) Total antioxidant capacity (TAC) of three commercial vitamin C tablets and three packaged fruit juices measured using the nanozyme‐based assay.
To evaluate the accuracy and matrix tolerance of the proposed nanozyme‐based assay under physiologically relevant conditions, spike‐and‐recovery experiments were conducted using human saliva and buffer controls. Near‐quantitative recoveries (98.6–100.88%) were observed in saliva relative to buffer, indicating negligible matrix interference (Table S3). The low relative standard deviation (RSD < 4%, n = 4) further demonstrates the high precision and robustness of the assay in complex biological samples. Antioxidant capacities measured by the nanozyme‐based assay exhibited an excellent linear correlation with the standard FRAP assay (R^2^ = 0.99; Figure S14), providing independent validation of its quantitative reliability for saliva analysis. Collectively, these results establish a clear proof‐of‐concept that the nanozyme‐based strategy enables reliable evaluation of antioxidants in complex biofluids such as human saliva, with satisfactory matrix tolerance and selectivity under realistic conditions. This performance supports the practical potential of the approach for bioanalytical applications. Future work will focus on expanding validation across larger and more diverse clinical cohorts to further define generalizability and clinical utility.
Tracking Redox Dynamics during Oxidative Stress, Recovery, and Antioxidant Intervention
3.7
We previously demonstrated that intracellular antioxidants quench thiochrome (TC) fluorescence in the presence of photoactivated AuNCs [39]. Expanding on this, we developed an enhanced AuAgNC‐based assay capable of sensitively tracking intracellular redox dynamics during oxidative stress, recovery, and disease‐relevant conditions. We hypothesized that mild, non‐cytotoxic oxidative stress would transiently deplete endogenous antioxidants, with subsequent recovery in fresh medium restoring redox homeostasis. Hydrogen peroxide (H_2_O_2_), at low doses, is known to induce oxidative stress without causing cytotoxicity [43]. HepG2 cells were first treated with 5 and 10 µM H_2_O_2_ for 90 min, and cytotoxicity assays confirmed both doses were nonlethal (Figure 8A), making them suitable for subsequent oxidative stress induction. As shown in Figure 8B, treatment with 5 and 10 µM H_2_O_2_ led to a dose‐dependent increase in TC fluorescence in cell lysates, indicating reduced intracellular antioxidant levels under oxidative stress. Upon 24 h recovery in fresh complete medium (10% FBS), TC fluorescence decreased, indicating replenishment of intracellular antioxidants. These results confirm that the AuAgNC‐based assay reliably monitors redox dynamics during both oxidative challenge and recovery phases, providing a robust platform for further antioxidant modulation studies. Heat inactivation of cell lysates led to a progressive increase in TC fluorescence, indicating that suppression of TC fluorescence in untreated lysates is primarily due to thermolabile, biologically active intracellular antioxidants (Figure S15).
(A) WST‐1 assay results showing cell viability of HepG2 cells after treatment with 5 and 10 µM H2O2 for 90 min. (B) Fluorescence responses of cell lysates measured using the AuAgNC‐based assay after 90 min H2O2 treatment (second and fourth panels) and following 24 h recovery in fresh medium (thirrd and fifth panels). (C) Fluorescence intensities obtained from cell lysates after treatment with different concentrations of AAP. (D) Fluorescence intensities of AAP‐treated cells pre‐exposed to H2O2. Error bars represent standard deviations (n > 3). (E) Schematic illustration of the AuAgNC‐based cell assay workflow. The upper panel depicts ROS‐mediated oxidation of thiamine (TH) to thiochrome (TC) under light irradiation in the presence of AuAgNCs, and the influence of intracellular components on the resulting fluorescence signal. The lower panel shows the four experimental treatment conditions used in the study.
Next, we tested whether exogenous antioxidant supplementation could modulate TC fluorescence under both basal and oxidative stress conditions. Cytotoxicity assays confirmed that 0–200 µM 2‐Phospho‐L‐ascorbic acid trisodium salt (AAP), a stable and cell‐compatible derivative of vitamin C [44], was non‐lethal (Figure S16). As shown in Figure 8C, AAP treatment led to a dose‐dependent decrease in TC fluorescence, indicating an increase in intracellular antioxidant levels. When AAP was applied to cells pretreated with H_2_O_2_ (mild oxidative stress), a similar dose‐dependent reduction in TC fluorescence was observed, although the magnitude of decrease was comparatively lower, consistent with partial antioxidant depletion under stress conditions (Figure 8D). Together, these results further validate the ability of the AuAgNC‐based assay to sensitively detect intracellular antioxidant fluctuations under both normal and oxidative stress conditions. A schematic illustration summarizing the cell‐based assay principle, treatment conditions and corresponding outcomes is shown in Figure 8E. The figure outlines how intracellular antioxidant levels influence TC fluorescence generation by AuAgNCs and highlights the redox status under different conditions.
Intracellular Redox Alterations in a Disease‐Mimicking Steatosis Model
3.8
To extend our study beyond normal oxidative stress and recovery, we examined the performance of the AuAgNC‐based assay in a disease‐relevant context. Oxidative stress is recognized as a hallmark of non‐alcoholic fatty liver disease (NAFLD) or steatosis, with elevated oxidative stress markers observed even in early stages [45]. In vitro modeling of NAFLD is well established through exposure of liver cells to fatty acids (FAs), which induces intracellular lipid accumulation, mimicking early‐stage disease pathology [46]. Accordingly, an in vitro steatosis model was developed using HepG2 cells treated with FAs. Successful lipid accumulation was confirmed using Oil Red O staining, with optical microscopy images (Figure S17) showing lipid‐rich regions stained red. Before introducing oxidative stress, we assessed the cumulative cytotoxicity of FA and H_2_O_2_ co‐treatment. As shown in Figure 9A, combined exposure did not induce cytotoxicity, validating the use of these conditions in subsequent assays. Treatment with FAs alone led to an increase in TC fluorescence (Figure 9B, panel 1), suggesting decreased antioxidant capacity. This aligns with prior studies showing that fatty acid‐induced in vitro steatosis triggers oxidative stress [47, 48], as lipid accumulation increases ROS generation, leading to oxidative stress and depletion of intracellular antioxidants, indicating reduced antioxidant levels. Co‐treatment with FAs and H_2_O_2_ further elevated TC fluorescence (Figure 9B, panel 2), indicating an increased oxidative burden. Notably, when steatotic cells were allowed to recover in fresh complete medium, a decrease in TC fluorescence was observed (Figure 9B, panel 3), suggesting restoration of antioxidant capacity. However, the recovery was less pronounced than in nonsteatotic cells, consistent with elevated basal oxidative stress in the steatotic condition. A schematic illustration of the model development and the treatment outcomes in terms of antioxidant fluctuations is presented in Figure 9C. Together, these results demonstrate that the AuAgNC‐based assay can sensitively track redox fluctuations in a disease‐relevant in vitro steatosis model, highlighting its potential utility for monitoring oxidative stress in pathophysiological contexts.
(A) Cumulative cytotoxicity assessment of fatty acid (FA) and H2O2 co‐treatment using the WST‐1 assay. Error bars represent standard deviations (n > 3). (B) TC fluorescence measurements under different treatment conditions: lipid treatment, lipid treatment followed by H2O2 exposure, and post‐treatment recovery in fresh medium. Error bars represent standard deviations (n > 3). (C) Upper panel: schematic representation of the in vitro steatosis (NAFLD) model. A 1% isopropyl alcohol (IPA) solution was prepared in serum‐free medium, mixed with a fatty acid (FA) solution (oleic acid: palmitic acid = 2:1), and warmed in a water bath for 30 min before cell treatment. Lower panel: schematic illustration depicting the experimental conditions corresponding to lipid loading, lipid loading followed by H2O2 treatment, and recovery in fresh medium.
Overall, this work presents a novel and comprehensive application of a nanozyme‐based antioxidant assay across diverse cell models, including oxidative stress, recovery, and steatosis conditions, an approach not previously reported in antioxidant detection platforms.
Conclusion
4
We developed a bimetallic Au‐Ag nanoclusters (AuAgNCs) nanozyme that exhibits significantly enhanced catalytic efficiency compared to its monometallic counterparts. The photocatalytic process follows a Type I photosensitization pathway, with peroxyl radicals (ROO•) serving as the dominant oxidizing species. The system demonstrated excellent catalytic kinetics, with favorable ν max and K_m_ values that outperformed previously reported mono‐ and bimetallic photo‐responsive oxidase‐like nanozymes. Compared with existing antioxidant assays, this nanozyme platform offers superior physiological compatibility, uses biologically relevant free radicals, enables rapid detection, and operates under simple, reagent‐free conditions. It retained high selectivity in the presence of biologically abundant interferents and required only low micromolar catalyst concentrations, supporting its cost‐effectiveness. Functionally, the assay enabled detection of a wide range of physiologically relevant antioxidants including thiols. Practical applicability was demonstrated through accurate total antioxidant capacity (TAC) measurements in synthetic saliva formulations and commercial products such as vitamin C tablets and fruit juices. Additionally, the system successfully tracked intracellular redox dynamics during oxidative stress, recovery, antioxidant intervention, and in a fatty acid‐induced steatosis (NAFLD) model, highlighting its utility for biological redox sensing. Collectively, this work presents a versatile, rapid, and physiologically relevant nanozyme platform with strong potential for antioxidant quantification in consumer products, redox biology studies, and disease‐related oxidative stress monitoring.
Supporting Information
Additional supporting information can be found online in the Supporting Information section. Supporting Fig. S1: Schematic illustration of the synthesis of BSA‐stabilized AuAgNCs. 5.0 mL of BSA solution (50 mg/mL) was mixed with 4.0 mL of HAuCl_4_ solution (10 mM) under vigorous stirring. Subsequently, 1.0 mL of AgNO_3_ solution (at varying concentrations) was added, followed by 1.0 mL of NaOH solution (1.0 M), with continuous stirring. BSA acts as a stabilizing and capping agent, while HAuCl4 and AgNO3 provide Au and Ag ions. NaOH induces nanocluster formation, and the reaction is maintained at 37°C for 12 h. The resulting nanoclusters were purified by ultrafiltration using 10 kDa centrifugal filters to remove unreacted metal ions, NaOH, and low‐molecular‐weight impurities. The retained nanoclusters were diluted and stored at 4°C, and a portion was freeze‐dried for long‐term storage and further characterization. Supporting Fig. S2: (a) Fluorescence spectra of AuAgNCs synthesized at different Au:Ag molar ratios (1:0 to 0:1). A distinct emission peak around 680–690 nm is observed for Au‐rich compositions (1:0 to 4:1). (b) Photographs of AuAgNCs under UV (365 nm) and visible light. The upper panel shows red fluorescence under UV illumination, while the lower panel shows a dark brown color under ambient daylight. Supporting Fig. S3: XRD pattern of AuAgNCs showing a broad peak around 18°, indicating an amorphous or disordered structure characteristic of ultrasmall nanoclusters. Supporting Fig. S4: FTIR spectra of native BSA (black) and AuAgNCs (red). Characteristic BSA bands are observed at ~1650 cm^−1^ (amide I, C=O stretching), 1518 cm^−1^ (amide II, N–H/C–N bending and stretching), and 1240 cm^−1^ (amide III). Upon AuAgNC formation, the amide I band shows a slight increase in intensity, the amide II band shifts to 1536 cm^−1^, and the amide III band is flattened, indicating backbone perturbation and secondary‐structure changes due to metal coordination. Additional changes in the C–H stretching region (2870–2960 cm^−1^) and the broad O–H/N–H band (3295 cm^−1^) suggest enhanced hydrogen bonding and conformational reorganization. Supporting Fig. S5: XPS survey spectrum of AuAgNCs confirming the presence of C, N, O, and S from BSA, along with Au and Ag signals from the metal core, indicating successful formation of BSA‐stabilized AuAgNCs. Supporting Fig. S6: UV–Vis absorption spectra of native BSA (black) and AuAgNCs (blue). The characteristic BSA peak exhibits a characteristic peak at ~278 nm, corresponding to π–π transitions of aromatic amino acids (tryptophan, tyrosine, and phenylalanine). Upon formation of AuAgNCs, this peak shifts slightly to 280 nm and flattens upon nanocluster formation, indicating conformational changes and interaction with the metal core. Absence of surface plasmon resonance confirms no larger nanoparticles are present. Supporting Fig. S7: Effect of thiamine (TH) concentration (0.02–2 mM) on the photocatalytic activity of AuAgNCs with maximum catalytic activity at 0.6 mM TH. Error bars represent standard deviations (n = 3). Supporting Fig. S8: Effect of AuAgNCs concentration (4‐40 µM) on the photocatalytic activity at 0.6 mM TH. Activity increased steadily with increasing nanozyme concentration, indicating enhanced catalytic performance. Error bars represent standard deviations (n = 3). Supporting Fig. S9: Effect of pH on the photocatalytic activity of AuAgNCs. Fluorescence intensity increased from acidic to alkaline conditions due to enhanced oxidation of thiamine at higher pH. Error bars represent standard deviations (n = 3). Supporting Fig. S10: Effect of temperature (5°C–65°C) on photocatalytic activity. Catalytic efficiency increased with temperature, indicating improved reaction efficiency at raised temperatures. Error bars represent standard deviations (n = 3). Supporting Fig. S11: Influence of buffer system on AuAgNC catalytic activity. Catalytic activity was highest in PBS compared to HAc, NaAc, and Tris‐HAc buffers. Error bars represent standard deviations (n = 3). Supporting Fig. S12: FRAP assay response for Trolox and glutathione. Trolox showed strong antioxidant activity, while glutathione exhibited negligible response. Supporting Fig. S13: ORAC assay response for Trolox and glutathione. Both Trolox and glutathione were detected; however, glutathione showed a notably weaker response. Supporting Fig. S14: Correlation analysis between the antioxidant capacities of saliva samples measured using the nanozyme‐based fluorescence assay and the FRAP assay. Antioxidant capacities were determined for saliva samples spiked with known concentrations of ascorbic acid (0.5, 1.0, and 2.0 μM). Linear regression analysis was performed to evaluate the agreement between the two methods. The solid red line represents the linear fit. Error bars represent standard deviation from triplicate measurements; error bars for the FRAP assay are smaller than the symbol size and therefore not visible. Supporting Fig. S15: Effect of heat inactivation on antioxidant activity in cell lysates. TC fluorescence increased with rising temperature (30°C to 70°C), indicating progressive loss of antioxidant activity with comparable quenching effects observed in 50°C and 70°C, suggesting contribution from thermolabile intracellular antioxidants. Supporting Fig. S16: Cytotoxicity assessment of 2‐Phospho‐L‐ascorbic acid trisodium salt (AAP) (50‐200 µM) on HepG2 cells using WST‐1 assay, showing no significant cytotoxicity, confirming their suitability for cell study on exogenous antioxidant supplementation. Supporting Fig. S17: Oil Red O (ORO) staining of HepG2 cells. a) Control cells without fatty acid (FA) induction, showing slight red staining due to intrinsic (basal) lipid content b) FA‐ treated cells with markedly increased lipid accumulation, observed as intense‐red stained areas. Scale bar: 50 µm. Supporting Table S1: Concentrations of interferents used in the selectivity of antioxidants detection assay. Supporting Table S2: Concentration matrix of various antioxidants and related compounds used for total antioxidant capacity (TAC) profiling in the photocontrolled nanozyme assay. Each row corresponds to a specific TAC level based on Trolox equivalents, with matched concentrations of KSCN, GSH, uric acid (UA), ascorbic acid (AA), cysteine (Cys), and homocysteine (Hcy) used for comparative analysis. Supporting* Table S3: Recovery and precision analysis of the proposed assay for ascorbic acid determination in saliva samples. Known concentrations of ascorbic acid were spiked into saliva and non‐saliva (buffer) matrices. Recovery (%) was evaluated to assess matrix effects, and precision was expressed as relative standard deviation (RSD, %, n = 4).
Author Contributions
Sanskruti Swain: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), validation (lead), visualization (lead), writing – original draft (lead), writing – review and editing (lead). I‐ Hsuan Chou: data curation (supporting), methodology (supporting), writing – review & editing (supporting), Bikash C. Mallick: data curation (supporting), methodology (supporting), writing – review and editing (supporting). Shu‐Chen Liu: resources (supporting), supervision (supporting), writing – review and editing (supporting). Gin‐Shin Chen: resources (supporting), supervision (supporting), writing review & editing (supporting). Hsing‐Ying Lin: funding acquisition (supporting), resources (supporting), supervision (supporting), writing – review and editing (equal). Chen‐Han Huang: conceptualization (lead), data curation (lead), formal analysis (lead), funding acquisition (lead), investigation (lead), project administration (lead), resources (lead), supervision (lead), validation (lead), visualization (lead), writing – review and editing (lead).
Funding
This work was supported by the National Science and Technology Council (grant nos. 111‐2112‐M‐008 ‐021 ‐MY3, 114‐2640‐B‐008‐001, and 112‐2221‐E007‐019‐MY3).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplementary Material
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1W. Zhao , P. Zhuang , Y. Chen , Y. Wu , M. Zhong , and Y. Lun , “Double‐Edged Sword” Effect of Reactive Oxygen Species (ROS) in Tumor Development and Carcinogenesis,” Physiological Research 72, no. 3 (2023): 301.37449744 10.33549/physiolres.935007 PMC 10669002 · doi ↗ · pubmed ↗
- 2H. Cichoż‐Lach and A. Michalak , “Oxidative Stress as a Crucial Factor in Liver Diseases,” World Journal of Gastroenterology 20, no. 25 (2014): 8082.25009380 10.3748/wjg.v 20.i 25.8082 PMC 4081679 · doi ↗ · pubmed ↗
- 3A. Silvestrini , E. Meucci , B. M. Ricerca , and A. Mancini , “Total Antioxidant Capacity: Biochemical Aspects and Clinical Significance,” International Journal of Molecular Sciences 24, no. 13 (2023): 10978.37446156 10.3390/ijms 241310978 PMC 10341416 · doi ↗ · pubmed ↗
- 4S. Chevion , E. M. Berry , N. Kitrossky , and R. Kohen , “Evaluation of Plasma Low Molecular Weight Antioxidant Capacity by Cyclic Voltammetry,” Free Radical Biology and Medicine 22, no. 3 (1997): 411–421.8981032 10.1016/s 0891-5849(96)00337-1 · doi ↗ · pubmed ↗
- 5R. L. Prior , X. Wu , and K. Schaich , “Standardized Methods for the Determination of Antioxidant Capacity and Phenolics in Foods and Dietary Supplements,” Journal of Agricultural and Food Chemistry 53, no. 10 (2005): 4290–4302.15884874 10.1021/jf 0502698 · doi ↗ · pubmed ↗
- 6J. Fang , Y. Wang , Y. Jiang , T. Li , and X. Qiu , “Advances in Total Antioxidant Capacity Detection Based on Nanozyme,” Talanta 292 (2025): 127941.40088770 10.1016/j.talanta.2025.127941 · doi ↗ · pubmed ↗
- 7X. Cao , T. Liu , X. Wang , Y. Yu , Y. Li , and L. Zhang , “Recent Advances in Nanozyme‐Based Sensing Technology for Antioxidant Detection,” Sensors 24 (2024): 6616.39460096 10.3390/s 24206616 PMC 11511242 · doi ↗ · pubmed ↗
- 8C. Wang , H. Wang , B. Xu , and H. Liu , “Photo‐Responsive Nanozymes: Mechanism, Activity Regulation, and Biomedical Applications,” View 2, no. 1 (2021): 20200045.