Preparation and characterization of BN-CDs
Our group previously reported that BN-CDs with high quantum yield and good fluorescence stability were synthesized by hydrothermal method, and it had been successfully applied to microbial labeling. In this study, BN-CDs were synthesized by microwave hydrothermal method. The advantage of microwave hydrothermal method is that it can shorten the time of CDs synthesis, and the size of CDs is more uniform and smaller, which is more conducive to microbial imaging .
The morphology of BN-CDs was characterized using the transmission electron microscopy (TEM). As shown in Fig. 1A, BN-CDs were uniformly dispersed, and the particle diameter was concentrated around 4 nm. In the UV spectrum (Fig. 1B), there were two absorption bands at 238 nm and 336 nm, which was due to the π = π* electronic transition of C = C and the n–π* transition of surface groups (C = O and C = N) . In addition, BN-CDs possessed optimal excitation and emission wavelengths at 350 nm and 446 nm (Fig. 1C). As shown in Fig. 1D, an intense and broad peak at about 3350 cm−1 was attributed to stretching vibration such as O–H/N–H , and stretching vibrations of C–H (3250 cm−1), C = O (1654 cm−1)  bonds were observed. The peak at 1554 cm−1 corresponded to C–N stretching vibration, which indicated that N atom has been successfully introduced into BN-CDs as passivator . The peak at 1165 cm−1 corresponded to the C-B vibration, indicating that B atom was successfully doped into BN-CDs .
X-ray photoelectron spectroscopy (XPS) was used to further study the surfaces of BN-CDs. The full spectrum (Additional file 1: Fig. S2) showed four typical peaks: B 1 s, C 1 s, N 1 s and O 1 s. In the high-resolution spectrum (Fig. 2), the C 1 s band was convoluted into three peaks, corresponding to sp2 carbon (C = C, 283.5 eV), sp3 carbon (C–N, 284.35 eV) and carbonyl carbon (C = O, 286.1 eV), respectively . The N 1 s band was convoluted into two peaks at 399.1 and 398.35 eV, representing N–H and N–C, respectively. O 1 s XPS spectrum was decomposed into peaks at 530.6 and 529.5 eV, corresponding to O = C and O–C, respectively . The spectrum of B 1 s exhibited two peaks at 190.8 and 190.2 eV, which could be assigned to B–C and B–O, respectively . The fluorescence quantum yields (QYs) of BN-CDs was calculated to be 66.59% using quinine sulfate as a standard. And their fluorescence lifetime was 125.63 ns (Additional file 1: Fig. S1), indicating that BN-CDs have better fluorescence stability.
Turbidity method is a conventional method to detect the growth of bacteria according to the turbidity of bacteria suspension. In order to determine whether BN-CDs can be used as probes for yeast imaging, we used the turbidimetric method to detect the toxicity of BN-CDs to yeast. It was found that when the concentration of BN-CDs reached 400 mg/mL, the growth curve of yeast was not significantly affected, which indicated that the cytotoxicity of BN-CDs was low (Additional file 1: Fig. S3), and confirmed that BN-CDs could act as eco-friendly biological fluorescent-labeling probes for yeast imaging.
Yeast can be clearly labeled with BN-CDs in only 1 min.(Fig. 3) Further observation showed that there were two status in the image of yeast cell, one was bright blue, and the other was dark blue with a bright halo, which was related to the life and death of yeast cell  (Fig. 3a). Yeast cell wall has a certain thickness, and cell membrane also has selective permeability, so BN-CDs was difficult to enter the living yeast cell in a short time. As a result, the fluorescence in living yeast cell was weak and dark blue. For dead yeast cell, the structure of protein and phospholipid bilayer in cell membrane is destroyed, resulting in the increase of cell permeability. Therefore, BN-CDs could quickly enter the cell, making the whole cell bright blue.
The budding pattern of the yeast cell can be clearly observed by using BN-CDs. There are two ways of yeast cell reproduction: sexual reproduction and asexual reproduction. Budding is the most common way of asexual reproduction of yeast . As shown in Fig. 3b, it was the budding status of yeast cell. The surface of yeast cell protruded outward and sprouted. When the buds grew to normal size, they would separate from the mother and become independent cells. Each mature yeast cell could sprout at one or more places (Fig. 3c). According to Fig. 3b, c; a bright blue color appeared inside the yeast cell during budding, indicating that the BN-CDs entered the budding yeast cell. During the budding process of yeast cell, the hydrolase decomposed cell wall polysaccharides to make cell wall thinner, so BN-CDs could enter cell in a short time . In addition to budding, when the nutritional status of yeast cell was not good, some cells could carry out sexual reproduction, formed spores (generally four spores), and germinated when the conditions were suitable (Fig. 3d), while the original vegetative cell became ascospores (Fig. 3e) . However, with the gradual aging of yeast cell, the protein expression was abnormal, and finally part of the aging cell appeared lysis (Fig. 3f). It can be concluded that BN-CDs can label yeast in a short time and clearly observe the status of yeast cell, which plays a positive role in understanding the growth process of yeast. BN-CDs are very different from the fluorescent CDs currently reported [25,26,28]. For those CDs, it needed a long co-incubation time for labeling yeast, and the growth status of yeast cell could not be observed. For BN-CDs, it only took one minute to label yeast, and the image was not only clear, but also showed the growth stage of yeast cell. BN-CDs staining can be used as a rapid screening method to monitor yeast during the fermentation process, and it is good for adjusting fermentation conditions in time.
Yeast imaging with different lethal methods
In order to further confirm that BN-CDs can quickly identify the dead and living yeast cells, six different methods were used to kill yeast. Methyl blue staining and pyridine iodide (PI)/fluorescein diacetate (FDA) staining were used as control. The dehydrogenases in the active yeast cell promote the reduction of methylene blue to a colorless substance, while dead yeast cell remains blue. PI can pass through the dead cell membrane and bind with DNA to emit red light. As shown in Fig. 4, the yeast cells stained with methylene blue were all blue, and those stained with PI/FDA were red, which indicated that they were all dead. Using BN-CDs to label the same yeast sample, as shown in Fig. 3, the whole yeast cell was bright blue, which was significantly different from the image of living yeast cell (Fig. 3a). It further confirmed that BN-CDs can identify the viability of yeast cell, which is consistent with the report of Ma .
It was worth noting that there was no obvious difference in the imaging when using methylene blue to stain yeast killed in different ways. However, when PI/FDA and BN-CDs staining were used, the images of cell killed by ultrasound and tobramycin sulfate were significantly different from those of other methods. The cell stained with PI/FDA was darker red, while yeast cell stained with BN-CDs, its surroundings and itself were bright blue. Water bath heating, microwave heating, formaldehyde and ethanol treatment will denature the protein in yeast cell, thereby killing the cells. Using these four methods to kill yeast there is no significant difference in imaging. However, the mechanical shearing force generated by ultrasound breaks cell membrane, which caused the cell contents to overflow. PI staining requires repeated washing, which reduces the binding of PI and cells, resulting in weaker fluorescence. When yeast was stained with BN-CDs, there was no need to wash, so BN-CDs marked the substances leaked out of the cell, causing bright blue light around the cell. Similarly, tobramycin sulfate could increase the permeability of cell membrane, leading to the leakage of potassium ions, adenine nucleotides, enzymes and other important substances in cell, so its image was similar to that of yeast cell killed by ultrasound.
In summary, the advantage of using BN-CDs to stain yeast is that there is no need to repeatedly wash, which avoids the experimental error caused by washing off cells during operation. BN-CDs staining can also quickly identify the death of yeast, and speculate whether the cause of death is related to cell membrane destruction.
Cellular uptake kinetics
It can be found that with the increase of BN-CDs concentration, the imaging of yeast becomes clearer, fluorescence inside the cell is also gradually enhanced (Fig. 5), and the average fluorescence intensity of cell increases significantly (Fig. 6a). All these indicate that the cellular uptake of BN-CDs has a dose-dependent characteristic. As shown in Fig. 5, the image of yeast cell is the clearest when the concentration of BN-CDs is 200 μg/mL, so BN-CDs (200 μg/mL) were used to label yeast in subsequent experiments. According to Fig. 7a, b, yeast was incubated with BN-CDs for a short period of time, the outer wall of some yeast cells had a bright light halo, while the inside was dark blue, and the cell wall and cytoplasm were clearly distinguished. With the extension of co-incubation time of BN-CDs and yeast, the fluorescence intensity in cells gradually increased, and the distribution of BN-CDs within cells became more and more uniform. As shown in Fig. 7, with the increase of incubation time, the fluorescence intensity inside cells became stronger and stronger., which showed that the cellular uptake of BN-CDs was time-dependent.
Endocytosis pathway of BN-CDs into yeast
Endocytosis, also known as transcytosis, is the process of transporting extracellular material into the cell through the deformation movement of plasma membrane. The endocytosis of nanoparticles is not only related to their size, surface potential, shape and surface chemical modification, but also depends on different cell types . Low-temperature, chlorpromazine and genistein were used to inhibit cell energy-dependent, clathrin-mediated and caveolin-mediated endocytosis, respectively.
Yeast was incubated with BN-CDs for 8 h at 26 ℃, BN-CDs entered yeast cell smoothly (Fig. 8a), and the average fluorescence intensity in cells was 160.62 ± 4.67 a.u (Fig. 9). As we all know, low temperature will reduce the activity of intracellular enzymes, resulting in a decrease in mitochondrial energy production [55, 56]. After BN-CDs and yeast were incubated at 4 ℃ for 8 h, the fluorescence intensity in cells decreased significantly to 129.55 ± 5.32 a.u (P < 0.05) (Fig. 9). The cellular uptake of BN-CDs at low temperature suggested that there were also non-energy dependent uptake pathways. Passive diffusion is a simple mode of transport without energy consumption. Ultra-small nanoparticles can enter cells through passive diffusion, such as metal nanoparticles less than 10 nm  and gadolinium nanoparticles less than 5 nm . Therefore, it isuggested that the cellular uptake of BN-CDs (4 nm) was partially energy-dependent and passive diffusion also participated in the process. The mechanism of clathrin-mediated endocytosis is that extracellular macromolecules are packaged into clathrin cavities and taken up by cells in the form of clathrin-coated vesicles. Chlorpromazine is used to inhibit clathrin-mediated endocytosis by transferring clathrin and its connexin from the plasma membrane to endosomes, thereby inhibiting the formation of clathrin-coated cavities . In the experiment, 200 μg/mL BN-CDs, 5 μg/mL chlorpromazine and yeast cells were incubated at 26 ℃ for 8 h. As shown in Fig. 8c, under the inhibition of chlorpromazine, the yeast still emitted bright blue light, and the average fluorescence intensity was 158.49 ± 5.69 a.u (P > 0.05). It showed that endocytosis mediated by grid was not the main way for BN-CDs to enter yeast cell.
Caveolae is a small concave structure that exists on the surface of a variety of cells, and the small concave protein plays a central regulatory role in signal transduction . Genistein is a tyrosine kinase inhibitor that can block caveolin-1 phosphorylation and is often used to inhibit caveolin-mediated endocytosis . It can be seen from Fig. 8d that after 200 μg/mL BN-CDs, 40 μg/mL genistein and yeast cells were incubated at 26 ℃ for 8 h, the inside of yeast was obviously dark blue, and the cell wall and cytoplasm could be clearly distinguished. It was significantly different from the previous three components. Its average fluorescence intensity was 106.23 ± 3.87 a.u, which is significantly lower than the other three groups (Fig. 9). Therefore, it can be clarified that BN-CDs enter yeast cell mainly by caveolae mediated endocytosis.