HAuCl₄ ^.4H₂O was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Cellulase, pepsin, trypsin and AA were obtained from Yuanye Biotechnology Co., Ltd. (Shanghai, China). Histidine, threonine, lysine, glycine, glutathione (GSH), maltose, sucrose, glucose and metal ions were acquired from Sangon Biotechnology Co., Ltd. (Shanghai, China). All reagents were of analytical purity and used directly. Milli-Q purified water prepared by the PR03200 ultra-pure water meter (Zhongshan Keningte Cleaning Supplies Co., Ltd.) was utilized in all experiments.

All glass containers in the laboratory were thoroughly washed with aqua regia, rinsed with ultrapure water and dried before use. UV-1800 spectrophotometer (Shimadzu, Japan), PF-5301PC fluorescence spectrophotometer (Shimadzu, Japan) and Spark-Multimode microplate reader (Tecan, Switzerland) were applied to measure the UV-vis absorption spectra, the fluorescence spectra, and the bacterial density, respectively. Transmission electron microscopy (TEM) images were collected on a JEOL 2010 LaB6 TEM (TECNAI G2, the Netherlands) at an acceleration voltage of 200 kV. Fourier transform infrared (FT-IR) spectra and X-ray photoelectron spectra (XPS) were separately detected by BW17-FTIR-650 spectrometer (Beijing, China) and X-ray photoelectron spectroscopy (Shimadzu, Japan). The fluorescence lifetime and quantum yield (QY) of the samples were recorded on an FLS920 fluorescence spectrometer (Edinburgh, UK). Zeta potential values were performed using the Malvern Zetasizer sizer NanoZS ZEM-3600 instrument (Malvern, UK). Furthermore, bacteria imaging was collected using the fluorescence microscope (Zeiss, Germany).

Typically, 0.16 mL of the HAuCl₄ solution (25 mM) and 9.84 mL of the cellulase solution (1 mM) were mixed thoroughly with a vortexer for 5 min. After adjusting pH to 12 with the addition of 1 M NaOH solution, the above mixture was reacted at 37°C for 12 h in the dark. Then the supernatant of the above mixture was collected by centrifugation at 8,000 rpm for 10 min, dialyzed to remove unreacted metal ions by a dialysis membrane (1, 000 MWCO) for 24 h, and placed at 4°C for future use.

For AA detection, the Cel-Au NCs (40 mg/mL, 50 μL), different concentrations of AA solutions (100 μL) and deionized water (850 μL) were mixed and incubated at 25°C for 5 min in a water bath. The fluorescence signal of the above mixture was then measured using an F-4500 fluorescence spectrophotometer by exciting at 560 nm. To evaluate the selectivity and specificity of Cel-Au NCs for AA, the fluorescence variations of Cel-Au NCs were investigated toward 16 kinds of compounds (histidine, threonine, lysine, glycine, GSH, maltose, sucrose, glucose, AA, KCl, NaCl, LiCl, ZnCl₂, CaCl₂, MgCl₂, MnCl₂). The as-prepared Cel-Au NCs were mixed with different compound solutions and measured under the same experimental condition as above. All experiments were performed three times in a parallel format.

To evaluate the applicability of the method, human serum samples provided from the Hospital of Traditional Chinese Medicine (Wuhu, China) were directly diluted 40 times with Milli-Q purified water before the experiment. Then, 50 μL of 40 mg/mL as-prepared Cel-Au NCs, 850 μL of diluted serum sample and 100 μL of different concentrations of AA solution were mixed and analyzed in accordance with the procedure mentioned above.

Bacillus subtilis (B. subtilis, gram-positive bacteria), Staphylococcus aureus (S. aureus, gram-positive bacteria) and Escherichia coli (E. coli, gram-negative bacteria) were separately cultured on Luria-Bertani (LB) agar plates at 37°C overnight. Subsequently, a single colony of the bacteria was separately picked and incubated in LB liquid culture medium with continuous shaking at 180 rpm at 37°C for another 16–24 h.

To estimate the biocompatibility of Cel-Au NCs, bacterial viabilities were measured by determining bacterial cell density at OD₆₀₀ on Spark-Multimode microplate reader. When OD₆₀₀ reached 0.6, the bacteria (B. subtilis, S. aureus, and E. coli) were seeded into a 96-well microplate at 1% inoculum. Then various concentrations of Cel-Au NCs (0, 10, 25, 50, 100 and 200 μg/mL) were separately added to the bacteria and cultured at 37°C and 180 rpm. The growth of organisms was observed by measuring OD₆₀₀ until 24 h and all of the experiments were executed three times in parallel. The percentage of bacterial density without adding Au NCs was taken as 100%.

After centrifuging at 8,000 rpm for 5 min, the above cultured bacterial cells were collected, washed with PBS, and incubated in the mixture of the prepared Cel-Au NCs (0.1 mL) and PBS (0.4 mL) in a shaker at 37°C for 15 min. The bacterial cultures were examined on a Zeiss upright fluorescence microscope under 605 nm.

The red-emitting Cel-Au NCs were firstly prepared via a facile and green one-step biomineralization method based on the reduction of cellulase provided by sulfur-containing cysteines and methionines, which made the Au-S band formed between cellulase and Au atom (Balu et al., 2019; Wang et al., 2019). To obtain the optimal conditions of the synthesized Cel-Au NCs, the molar ratio (cellulase/HAuCl₄) and reaction pH were conducted in Supplementary Figure S1. The molar ratio of 2.5:1 and the reaction pH of 12 served as optimal conditions were selected for further study.

Initially, UV-vis absorption spectra and fluorescence spectroscopy were employed to identify related optical properties of Cel-Au NCs. The UV-vis spectrum showed that Cel-Au NCs had a shoulder peak in the region range of 300–400 nm with a continuous rise and a distinct peak at 350 nm attributed to oxidation between cellulase and Au atoms, whereas the spectrum of cellulose showed no peak in these ranges, signifying the Cel-Au NCs were fabricated (Figure 1A). As shown in Figure 1B; Supplementary Figure S2, the red-emitting cellulase protected Au NCs displayed an emission peak maximum at 680 nm upon 560 nm excitation with a marked Stokes shift of 120 nm. Additionally, the QY of Cel-Au NCs in aqueous solution was determined to be 10.19% using Rhodamine 6 G as a reference (Supplementary Figure S3).

The morphology of the prepared Cel-Au NCs was characterized by TEM, revealing that the Cel-Au NCs had a good dispersion and the average size was 1.68 nm by counting 146 samples (Figures 1C, D), which was consistent with the diameter of metal NCs prepared in previous studies (Wei et al., 2010; Bhamore et al., 2019). Subsequently, FT-IR was used to characterize the chemical composition of Cel-Au NCs. As shown in Figure 1E, the peaks of pure cellulase and Cel-Au NCs for O-H stretching, C-H stretching, C=O stretching, C-H bending, N-H stretching and C=C bending were separately observed at 3,400 cm^-1, 2,940 cm^-1, 1,655 cm^-1, 1,415 cm^-1, 1,250 cm^-1 and 1,025 cm^-1, whereas a distinct peak in the spectrum of Cel-Au NCs was observed at 1,580 cm⁻¹ ascribing to the formation of a bond between Au and cellulose.

XPS was used to measure the oxidation states of gold in Au NCs and it showed the peaks of Au, S, C, N and O in the XPS spectra (Supplementary Figure S4). Two peaks centered at 88.0 and 84.3 eV were separately ascribed to 4f_5/2 and 4f_7/2 for Au (Figure 1F). The peak of 4f_5/2 of the prepared Cel-Au NCs was further deconvoluted into two different components, one at 88.05 eV corresponding to Au (0), and the second one at 88.60 eV attributed to Au (I). Also, the two peaks of 4f_7/2 assigning to 84.32 and 84.99 eV showed the simultaneous presence of Au (0) and Au (I) in Cel-Au NCs. The spectra of Au 4f_7/2 showed a binding energy of > 84.0 eV, indicating both Au (0) and Au (I) existed in Cel-Au NCs and the presence of Au-S complexes formed by the formation of charge transfer bands (Bothra et al., 2017).

When the addition of AA was increased from 10 μM to 800 μM, a corresponding reduction in the fluorescent signal of Cel-Au NCs was examined (Figure 2A). Figure 2B depicted the relationship between the fluorescent intensity of the Cel-Au NCs and the different concentrations of AA, and showed a good linear correlation over a range of 10–800 µM with a LOD of 2.5 µM (R ² = 0.99134), indicating that the detection system possessed superior sensitivity. Simultaneously, the fluorescent intensity of Cel-Au NCs was correspondingly reduced with the increasing concentration of AA by UV light (Figure 2C). Furthermore, the specificity of the Cel-Au NCs for AA was conducted by testing the response of the biosensor prepared against other compounds. Interestingly, the fluorescent intensity of Cel-Au NCs was extremely decreased just after adding AA, whereas there were barely any changes in the presence of the other compounds (Figures 2D, E; Supplementary Figure S5). Compared with the published methods in AA detection (Table 1 and Supplementary Table S1), the proposed method displayed a wider detection range and an appreciable detection limit, which is simplicity, rapidity, efficiency and economics. Thus, as an alternative biosensor, it is potential for AA detection in the biological environment using Cel-Au NCs.

To elucidate the quenching mechanism of AA on the Cel-Au NCs, fluorescence resonance energy transfer (FRET), inner filter effect (IFE), dynamic and static quenching as well as photoinduced electron transfer had been investigated. As depicted in Supplementary Figure S6, AA displayed a strong absorption peak at 245 nm, which did not overlap with the Cel-Au NCs emission spectrum (600–800 nm), demonstrating that the mechanism of quenching mechanism caused by AA was not FRET and IFE (Fan et al., 2022). Notably, the fluorescent lifetimes were 10.27 μs and 9.60 μs for Cel-Au NCs before and after the addition of AA, separately (Figure 2F). The noticeable change in the fluorescence lifetime of Cel-Au NCs upon the addition of AA indicated that the quenching mechanism might be dynamic quenching rather than static quenching. Similarly, the fluorescence quenching of LDH-GQD caused by Fe³⁺ was determined to be dynamic quenching due to the reduction of fluorescence lifetime from 6.45 ns to 1.21 ns (Shi et al., 2021). Furthermore, the zeta potential of Cel-Au NCs increased from −15.2 mV to −13.3 mV after adding AA (Supplementary Figure S7). The negative zeta potential of the Cel-Au NCs is attributed to the presence of carboxylic groups with negative charges on the surface of cellulase, while the apparent increase in the zeta potential of Cel-Au NCs after the addition of AA confirms that the positively charged AA was attached to the surface of the negatively charged Cel-Au NCs. Additionally, the reducing power of AA caused the alteration in the oxidation state of Au (I), localized on the surface of the Au (0) core, further leading to the fluorescence quenching of Cel-Au NCs. (Li et al., 2015; Li et al., 2017). Hence, the quenching mechanism of Cel-Au NCs might be attributed to photoinduced electron transfer and dynamic quenching mechanism.

For assessing the practicality of the method in actual samples, the detection of AA in serum samples was carried out. As depicted in Table 2, the recovery rates of AA in actual samples were in the range of 98.76%–104.83%, and the relative standard deviations (RSD) ranged from 1.05% to 5.04%. Furthermore, to demonstrate the practicability and accuracy of this biosensor, diverse concentrations of AA in serum samples were analyzed by the commercial HPLC method (Supplementary Table S2). The recoveries of AA were between 94.24% and 102.24% with RSD of 0.13%–3.07%. These results illustrated that this developed biosensor was applicable for the detection of AA in biological samples in comparison with the HPLC method.

The biocompatibility of Cel-Au NCs was evaluated by measuring the bacterial density at OD600. The assay was conducted on three kinds of bacteria including B. subtilis (gram-positive bacteria), S. aureus (gram-positive bacteria) and E. coli (gram-negative bacteria). As evidenced by Supplementary Figure S8, Cel-Au NCs exhibited a negligibly inhibitory effect on bacterial cell proliferation within the range of 0–100 μg/ml and had a slight inhibitory on bacterial cell proliferation at 200 μg/ml, indicating low cytotoxicity of the Cel-Au NCs to bacteria.

To verify the bacterial labeling ability of Cel-Au NCs, bacterial cells incubated with Cel-Au NCs were observed under a fluorescence microscope. B. subtilis (gram-positive bacteria, Figures 3A, D), S. aureus (gram-positive bacteria, Figures 3B, E), and E. coli (gram-negative bacteria, Figures 3C, F) stained by Cel-Au NCs were respectively shown in the bright field and the dark field with strong red emission when excitation at 605 nm. In light of this, we hypothesized that Cel-Au NCs with ultra-small size might be absorbed by bacteria and interact with multiple proteins in the bacteria. In our previous study, S. aureus, B. subtilis as well as Microbacterium incubated with papain-Pt NCs could emit distinct green fluorescence (Chang et al., 2021). Besides, in the latest research, Li’s group used red-fluorescent cBSA-AuAgNCs with an average diameter of 1.80 nm to label E. coli (Li et al., 2022). Therefore, Cel-Au NCs with satisfactory fluorescence characteristics could be explored as a bioprobe that effectively labels the microorganism cells.