Materials, characterization methods and cytotoxicity to the fungus of the carbon dots are shown in the Supplementary Material.

The carbon dots were prepared by microwave assisted pyrolysis of L - glutamic acid and anhydrous ethylenediamine as raw materials (Lu et al., 2015; Li et al., 2016). Briefly, 500 mg of L - glutamic acid was dissolved in 20 mL ultrapure water to get homogeneous solution, into which 1.65 mL anhydrous ethylenediamine was slowly added. Subsequently, the resulting mixture was heated by a domestic microwave oven for 4 min. Then, another 10 mL of ultrapure water was added to dissolve the product after it naturally cooled down to room temperature. Thereafter, the obtained solution purified by dialyzing against ultrapure water through a dialysis membrane (MWCO 500) for 2 days. The purified solution was concentrated by means of rotary evaporation at 50 °C. Finally, the solid carbon dots were obtained by freeze drying under vacuum at ambient temperature for 48 h.

Typically, 0.15 mg/mL of the carbon dots were mixed with Fe³⁺ at various concentrations in Tris-HCl buffer (0.1 M, pH 7.0) solution, then their fluorescence spectra were collected at room temperature. The relative fluorescence intensity (F₀-F)/F₀ was calculated to be the signal output value, in which F₀ and F represent the fluorescence intensities at 460 nm of the carbon dots in the absence and presence of Fe³⁺, respectively.

To evaluate the feasibility of the carbon dots based sensor for the detection of Fe³⁺ in real samples, tap water samples obtained from our lab were analyzed using the present method. All water samples were spiked with Fe³⁺ at different concentrations without any pretreatment.

Fungal fermentation was carried out in liquid medium at 27 ± 2 °C for 3 days. Thereafter, the selected young vigorous mycelia washed by distilled water for three times were incubated in the Tris-HCl buffer solution containing the as-synthesized carbon dots at a certain concentration for another 36 h in a constant temperature incubator. The stained mycelia were imaged on a laser scanning confocal microscopy (Lieca, Germany) after washed three times with distilled water. Three drops of 0.005 M Fe³⁺, Fe(III)-EDTA, Cu²⁺, and Ca²⁺ were added from one side of the chink between the cover glass and the slide glass for intracellular imaging of metal ions.

TEM image of the as-synthesized carbon dots is shown in Figure 1A. The morphology of the carbon dots is nearly spherical, and their average diameter is 4.5 nm. The measured XRD pattern of the carbon dots showed an obvious peak centered at 2θ = 23.3° with d = 0.381 nm (Figure 1B), which is corresponding to the (002) lattice spacing of carbon based materials (Vikneswaran et al., 2014). And the functional groups on the surface of the carbon dots were identified by FTIR. As demonstrated in Figure 1C, the broad band centered at 3,263 cm⁻¹ was contributed by the stretching vibrations of –OH and –NH₂. The absorption at 3,074 cm⁻¹ was attributed to the stretching vibration of C=C (Dong et al., 2012). The peaks at 1,682, 1,585, 1,250, and 1,101 cm⁻¹ belonged to C=O stretching vibration, N-H bending vibration, C-OH stretching vibration and C-N stretching vibration. The FTIR information suggests that there are carboxyl and amine groups on the surface of the as-synthesized carbon dots.

As shown in Figure 2, an characteristic absorption band at 310–330 nm in the absorption spectrum of the carbon dots was attributed to n–π^* transition (Teng et al., 2014). The fluorescent spectra in Figure 2 exhibited an excitation-dependence feature (i.e., the emission was red-shifted with excitation wavelength increasing). The emission peak intensity firstly increased, and then decreased remarkably as excitation shifted to longer wavelength (see Figure S1). The optimum emission was revealed at 459 nm under excitation at 360 nm. Therefore, the emission of the carbon dots are controlled by surface states. And different emission centers on the surface result in excitation dependent emission (Chen et al., 2015; Dai et al., 2016; Kim et al., 2018). The fluorescent quantum yield of the carbon dots is 12.45% referencing to quinine sulfate (quantum yield 54.6%).

Photostability of the as-synthesized carbon dots was measured under continuous illumination at 360 nm with a 150 W Xe lamp in a fluorescence spectrophotometer. The fluorescence intensity exhibited negligible change within 50 min (Figure S2), suggesting that the as-synthesized carbon dots can keep photostable under light irradiation during detection and imaging.

The effect of pH on the fluorescence of the carbon dots was shown in Figure 3A. Fluorescence intensity decreased significantly as pH increased in the range of 5–7 and pH > 10. But fluorescence intensity of the carbon dots was independent on pH in the ranges of 2–4 and 7–9. Protonation–deprotonation occurring on the surface of the carbon dots with pH variation induced the changes in the surface charges, resulting in pH-dependence (Hu et al., 2014; Yuan et al., 2016; Liu et al., 2017). Therefore, if the carbon dots are used as bioimaging agents under the physiological pH condition, fluorescence intensity will hardly suffer from the pH variation.

Nineteen kinds of cations, including Li⁺, Na⁺, K⁺, NH4+, Al³⁺, Ag⁺, Ca²⁺, Fe³⁺, Ba²⁺, Co²⁺, Cr³⁺, Cu²⁺, Zn²⁺, Hg²⁺, Mg²⁺, Mn²⁺, Ni²⁺, Pb²⁺, and Yb³⁺, were selected as representatives to evaluate the effects of cations on the fluorescence of the carbon dots. As exhibited in Figure 3B, the fluorescence emission of the carbon dots could be significantly quenched by Fe³⁺ while the other cations exerted negligible effects. Therefore, the carbon dots respond to Fe³⁺ selectively in the system containing cations. In order to test the selectivity of the carbon dots to Fe³⁺ in complex environment, the other potential interferences from anions, biomolecules, and ionic strength were further tested. Figure S3 shows the results of the interferences of anions (Ac⁻, Br⁻, Cl⁻, CO32-, F⁻, HCO3-, NO3-, PO43-, and SO42-) on fluorescence intensity of the carbon dots, from which no obvious changes were found in the presence of different anions by comparison with the blank. Furthermore, negligible influences of amino acids and glucose on fluorescence intensity of the carbon dots were found (see Figure S4). Additionally, an increase in ionic strength would not fluctuate the fluorescence of the carbon dots (Figure S5), suggesting that the detection of Fe³⁺ will not be affected by the variation of ionic strength. Overall, the carbon dots showed an excellent selectivity toward Fe³⁺ under complex condition.

The responsive time of the carbon dots toward Fe³⁺ ions was measured and the result is shown in Figure S6, which indicates that the response the carbon dots toward Fe³⁺ ions is balanced within 4 min.

The absorption spectra of the carbon dots in the presence of Fe³⁺ ions were collected in Figure S7A, which exhibits that absorbance at about 320 nm increases upon addition of Fe³⁺. Combined with the FT-IR results that there are hydroxyl, carboxyl, and amine groups on the surface of the as-synthesized carbon dots, it can infer that carbon dots can form stable complex with Fe³⁺ ions through carboxyl, resulting in fluorescence quenching due to photo-induced electron transfer and high selectivity of the carbon dots to Fe³⁺ (Chen et al., 2016; Li et al., 2017). The fluorescence lifetimes of the carbon dots in absence and presence Fe³⁺ ions were measured for the investigation of the mechanism of fluorescence quenching of Fe³⁺ ions toward carbon dots. As shown in Figure S7B, the fluorescence lifetime of the carbon dots ions nearly did not change upon addition of Fe³⁺, suggesting that the mechanism of fluorescence quenching of Fe³⁺ ions toward carbon dots can be attributed to static quenching (Jiang et al., 2019).

The fluorescence intensities of the carbon dots in Tris-HCl buffer solutions at pH = 7.0 containing Fe³⁺ at various concentrations were measured to study the detection sensitivity. As shown in Figure 4A, the emission of the carbon dots showed a decrease in intensity upon addition of Fe³⁺ with concentration increasing from 0 to 800 μM. A good linear relationship (Figure 4B) between (F₀-F)/F₀ and the concentration was found in the range of 8–80 μM (R² = 0.9936). The linear equation was fitted as (F₀-F)/F₀ = 0.00358 c−0.00875 (c representing the concentration of Fe³⁺), and the limit of detection (LOD) was calculated to be 3.8 μM based on three times the standard deviation rule (LOD = 3Sd/s).

A comparison between the proposed method based on the carbon dots with the previously reported methods are listed in Table S1, which suggests that this method based on the carbon dots for the detections of Fe³⁺ is comparable or even better than those previously reported in literatures.

The carbon dots with high sensitivity and selectivity to Fe³⁺ might be directly applied for the detection of Fe³⁺ ions in real water samples. Therefore, quantitative detection of Fe³⁺ in tap water was conducted for evaluation the practical application of the carbon dots as a Fe³⁺ sensor. The tap water samples which were taken from our lab and spiked with Fe³⁺ at different concentrations without any pre-treatment. Then, the spiked Fe³⁺ was analyzed by the proposed method. As presented in Table 1, the recoveries of the spiked water samples are in the range from 90.6 to 107%, suggesting that the as-synthesized carbon dots can be employed as a promising sensor for Fe³⁺ in real water samples.

The cytotoxicity of the as-synthesized carbon dots to C. gloeosporioides fungus was tested by the method of growth rate. As shown in Figure 5A, fungal mycelia growing on the PDA media containing the as-synthesized carbon dots at the concentrations of 0.2, 0.5, 0.8, 1.0, 1.2 mg/mL, respectively, show no obvious difference compared with the control. The fungal viability is slightly reduced at the concentration of 1.2 mg/mL (Figure 5B). Therefore, the as-synthesized carbon dots are a kind of excellent biocompatible label for bioimaging of C. gloeosporioides fungus.

The potential of the as-synthesized carbon dots to be label agents for fluorescently imaging pathogenic fungal cells (C. gloeosporioides) on a confocal microscopy was explored. The isolated mycelia from liquid medium were incubated in the Tris-HCl buffer solutions in the presence of the as-synthesized carbon dots at different concentrations. After 36 h incubation at the room temperature, brightly illuminating fungal cells were observed by the confocal microscopy under excitation at 405 nm, which suggests that the carbon dots can readily enter the fungal cells and stain mycelia for imaging (see Figure S8). By analysis of the fluorescent intensity of the fluorescent images of the fungal cells, the internalization of the carbon dots into fungal cells exhibited dose dependence (Figure S8). The strongest fluorescent intensity of the fluorescent images was observed for the fungal cells incubated in the Tris-HCl buffer solution containing the as-synthesized carbon dots in the concentration range from 0.4 to 0.5 mg/mL. Thus, 0.4 mg/mL was chosen as the optimum concentration for subsequent cell labeling.

In order to evaluate the feasibility of the as-synthesized carbon dots as fluorescent agent for detecting Fe³⁺ inside fungal cells, imaging was carried out on a confocal microscopy with the carbon dots labeled mycelia being incubated in Tris-HCl buffer solution containing 0.005 M Fe³⁺, and in Tris-HCl buffer solution containing 0.005 M EDTA-Fe(III) complex, Cu²⁺ and Ca²⁺ as comparison. As shown in Figure 6, the fungal cells incubated in blank Tris-HCl buffer solution showed strong intracellular fluorescence (control). But the intracellular fluorescence was significantly reduced in its intensity while the fungal cells were incubated in Tris-HCl buffer solution containing 0.005 M Fe³⁺ for 2 and 5 min. And incubation time had no obvious effect on the intracellular fluorescence intensity of the fungal cells (confocal photo taken at 5 min is not significantly different with that taken at 2 min), indicating that free Fe³⁺ can readily enter the fungal cells and effectively quench the fluorescence of the internalized carbon dots. Hence, the carbon dots show selective fluorescent response not only to free Fe³⁺ but also to the Fe(III) complex. However, the intracellular fluorescence intensity of the fungal cells incubated in Tris-HCl buffer solution containing the Fe(III) complex at a concentration of 0.005 M had not obvious change compared with that incubated in blank Tris-HCl buffer solution (Figure 6). Thus it can be speculated that EDTA may hinder Fe(III) to enter the fungal cells because iron uptake of fungal cells involves the use of iron siderophores secreted by fungi (David et al., 1992).

As discussed above, Cu²⁺ and Ca²⁺ have negligible effect on the fluorescence of the carbon dots in the Tris-HCl buffer solution. Herein, Cu²⁺ and Ca²⁺ were selected as the representative metal ions to evaluate the selectivity of the carbon dots for intracellular imaging of Fe³⁺ inside fungal cells. As presented in Figure 6, there was not obvious change in intracellular fluorescence intensity of the fungal cells incubated in Tris-HCl buffer solutions containing Cu²⁺ and Ca²⁺ compared with that in blank Tris-HCl buffer solution. Therefore, the other metal ions, such as Cu²⁺ and Ca²⁺, will not exert effect on the fluorescence of the internalized carbon dots in fungal cells during intracellular imaging of Fe³⁺. The as-synthesized carbon dots with high selectivity to Fe³⁺ have the potential to be developed as fluorescent probe for detection of Fe³⁺ inside fungal cells.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.