Mitochondria are two-membrane-bound sub-organelles surrounded by an outer and an inner smooth membrane, which is folded to form the cristae (Sasaki, 2010). The inner mitochondrial membrane encircles a space identified as the matrix. The membrane potential difference of mitochondria tends to attract positively charged molecules to accumulate in its interior (Leung et al., 2013). Based on this, we modified the coumarin group with pyridine to make it positively charged. Under the attraction of mitochondrial membrane potential difference, CPE can accurately target mitochondria (Figures 1A,B) (Zhao et al., 2013; Zhao et al., 2014; Zhao et al., 2015; Shi et al., 2016; Gui et al., 2017). In addition, the fluorophores possess two emission peaks (∼500 and 660 nm) (Figure 1H), which provides them the chance to label the mitochondrial membrane and matrix. To indicate this point, we incubated CPE in vitro with lecithin for 1 h, and then imaged using SIM with 405 and 561 channels emitted. As expected, lecithin that loaded CPE showed green fluorescence at the excitation of 405 nm and red fluorescence at the excitation of 561 nm (Supplementary Figure S1), which suggests that CPE is a double-labeled probe. In addition, the fluorophores can react with RSS such as H₂S and SO₂ (Figures 1C,F). It is shown that the color of CPE changed from mauve to light yellow when it reacted with Na₂SO₃ in vitro experiments (Supplementary Figure S2). To measure the degree of response to RSS, we used Na₂S as an H₂S donor and Na₂SO₃ as an SO₂ donor to simulate the environment rich in RSS of mitochondria. As observed, the fluorescence of CPE decreased distinctly with the increase of RSS (Figures 1D,G). The 561 excited light was quenched at 10 equivalent Na₂SO₃, while the 405 excited light did not quench and remained at a certain fluorescence intensity still. Moreover, the probe did not respond to other biologically relevant species such as H₂O₂, ClO⁻, and F⁻ (Figure 1E). These results suggest that CPE can not only respond to active sulfur and detect the active sulfur content but also have the potential to label mitochondria.

To verify whether CPE could target organelles in living cells or not, HeLa cells were incubated with CPE and imaged under SIM using a dual-channel mode with excitation at 405 and 561 nm. As shown in SIM images, green fluorescence excited at 405 nm stained the outer membrane of mitochondria, which revealed fibrous, rod-like, and punctate morphology (Figures 2C,D), consistent with previous literature reports (Sasaki, 2010). Red particles or fibers with weaker fluorescence excited at 561 were encased in a green membrane, illustrating it targets the mitochondrial matrix (Figure 2A). Here is the surprise, CPE could show the crest line distribution of mitochondria (Figure 2B), which is closely associated with the pathology of cancer, osteoarthritis, and AIDS (Guarani et al., 2015; Blanco et al., 2011). In addition, this scene in live cells could only be captured by SIM compared to other reported methods of imaging (Shao et al., 2020). Therefore, the combination of SIM can take advantage of the probe CPE, which suggests its potential for use in the diagnosis of mitochondrial diseases. Next, we used the length-to-width ratio (L/W) to quantitatively analyze the distribution of mitochondria and found that various morphologies could be assigned into four groups as follows: hyperfused (L/W ≥ 5.0), tubular (2.0 ≤ L/W < 5.0), intermediate (1.5 ≤ L/W < 2.0), and round or nearly round (1.0 ≤ L/W < 1.5) (Figures 2C–E) (Cao et al., 2017). The fibrous, rod-shaped, and spotted once mentioned earlier might be classified as hyperfused, tubular, and round (Figure 2F), and then, we used CCK-8 assay to evaluate the cytotoxicity of CPE to HeLa cells (Qin et al., 2015). No cytotoxicity was shown at the range of 0–20 μM to HeLa cells during 24 h, demonstrating that 10 μM is a relatively safe working concentration for CPE with no interference in mitochondrial imaging under SIM in living cells. Finally, different temperatures and endocytosis levels of CPE incubated to HeLa cells were observed to define the uptake properties of CPE. The cells showed weaker fluorescence when incubated with CPE at 4°C (Supplementary Figure S3) or with an endocytosis inhibitor (NH₄Cl) (Supplementary Figure S4) than that of cells incubated with CPE at 37°C (Fang et al., 2019), whether it is red or green fluorescence. These results strongly indicate that CPE enters cells through energy-dependent endocytosis. Therefore, we conclude that CPE stains organelles in living cells with low toxicity and good cell permeability.

To determine whether CPE can specifically label mitochondria, we co-stained the cells with a commercial probe, PKMTDR, for 1 h. The following merged SIM images revealed that the green fluorescence signal of CPE colocalized well with the red fluorescence signal of PKMTDR (Figures 3A,B), whose Pearson colocalization coefficient (PCC) was as high as 0.72 with PKMTDR (Figure 3C). It is clear that CPE has high specificity for mitochondrial attachment. Next, we observed whether CPE-labeled mitochondria depends on mitochondrial membrane potential (MMP) or not. To damage the membrane potential of mitochondria, HeLa cells were treated with 10 μM carbonyl cyanide m-chlorophenyl hydrazone (CCCP), which was used as a common mitophagy inducer (Chen et al., 2020a). After that, we re-stained the cells with both CPE and PKMTDR for co-location imaging (Figures 3D,E), while most did not attach to broken mitochondria. This indicates that labeled mitochondria depends on MMP. These results show that CPE could not only label mitochondria specifically but also provide references for measuring MMP.

To confirm whether CPE can detect the content of active sulfur in mitochondria, we processed mitochondria with CCCP of 50 μM for 30 min. Mitochondria were broken into round shapes (Figure 4A), consistent with previous literature reports (Chen et al., 2020b). Then, L/W was used to quantitatively analyze the distribution of mitochondrial morphology as before. After that, we used automatic analysis software (ImageJ) to calculate the distribution of individual mitochondria in CCCP-treated HeLa cells and found that the ratio of round structures accounts for nearly a half (Figure 4D), which indicates the mitochondria were in an unhealthy status. In addition, we found that red fluorescent excited by 561 was quenched in round shape mitochondria (Figure 4C), which was due to CCCP treatment increased the concentrations of ROS (Kane et al., 2018), high concentrations of ROS then increased RSS levels (Tabassum and Jeong, 2019). However, there was still some red fluorescent outside mitochondria, thus we hypothesized that it was caused by the destruction of mitochondrial outer membrane and the outflow of mitochondrial matrix (Figure 4B). Meanwhile, the content of active sulfur flowing out of the matrix was not enough to quench red fluorescent. All these indicate that CPE has the potential of detecting active sulfur in mitochondria and judging the status of mitochondria, which can provide a powerful reference value for the diagnosis of mitochondrial diseases.

A mixture of 7-(diethylaMino)-2-oxo-2H-chromene-3-carbaldehyde (0.26 g, 1 mmol) and 4-pyridineacetonitrile (0.12 g, 1 mmol) refluxed in dry ethanol (15 ml). A large amount of brown powder solids precipitated after reacted overnight. After cooling to room temperature, the crude product was filtered and washed with cool acetonitrile. Subsequently, a mixture of the previous product (0.172 g, 0.5 mmol) and iodoethane (0.47 g, 3 mmol) refluxed in acetonitrile (3 ml). A large amount of purple powder solids precipitated after reacted for 24 h. Then, the crude product was obtained by filtering, followed by washing with cold ethanol and ethyl ester. A purple compound CPE (0.11 g, 0.3 mmol, 60%) was obtained after drying. ¹H NMR (600 MHz, DMSO-d ₆ ) δ (ppm) (Supplementary Figure S5): 9.05 (d, J = 7.1 Hz, 2H), 8.84 (s, 1H), 8.39 (s, 1H), 8.32 (d, J = 7.1 Hz, 2H), 7.66 (d, J = 9.1 Hz, 1H), 6.88 (dd, J = 9.1, 2.4 Hz, 1H), 6.68 (d, J = 2.2 Hz, 1H), 4.62 (q, J = 7.3 Hz, 2H), 3.56 (q, J = 7.0 Hz, 4H), 1.55 (t, J = 7.3 Hz, 3H), and 1.18 (t, J = 7.1 Hz, 6H). ¹³C NMR (151 MHz, DMSO-d ₆ ) δ (ppm) (Supplementary Figure S6): 160.46, 158.10, 154.22, 150.01, 145.28, 144.98, 144.82, 132.97, 123.23 116.91, 111.58, 111.31, 108.85, 102.17, 97.29, 56.04, 45.28, 16.61, and 12.94. HRMS m/z (Supplementary Figure S7): calculated for C₂₃H₂₄N₃O₂ ⁺ [M]⁺: 374.1863, found 374.1775.

Dulbecco’s modified Eagle’s medium (#11965118, DMEM), phenol-free medium (#1894117), penicillin–streptomycin (#15140163, 10,000 units/ml), trypsin-EDTA (#25200–072), and other reagents for cell culture were obtained from Gibco BRL (Grand Island, NY, United States). Fetal bovine serum (FBS) was obtained from VivaCell (Shanghai, China). HeLa cells were gifted from the Chunyan Liu’s lab (Shandong First Medical University).

HeLa cells were cultured in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum, penicillin (100 μg/ml), and streptomycin (100 μg/ml) in a 5% CO₂ humidified incubator at 37°C.

CPE (10 μM) was put into a color dish and allowed to react with different concentrations such as 0, 2, 4, 6, and 8 eq., Na₂SO₂ or H₂S in order to complete the reactive sulfur concentration response experiment, and then the fluorescence spectra of CPE at different concentrations were detected. Next, we incubated CPE with lecithin for 1 h and imaged under SIM to search for the luminescence properties of CPE.

HeLa cells were seeded on 35 mm glass-bottom micro dishes at a density of 1×10⁵ and incubated with 2 ml of DMEM medium supplemented with 10% FBS for 24 h. After that, cells were incubated with 10 μM CPE for 1 h and washed with fresh DMEM for five times. At last, the cells with no phenol in culture medium were imaged under an OMX 3D-SIM extended-resolution microscope. Images were acquired at 512 × 512, with a step size of 0.125 μm, and CPE was excited at 405 and 561 nm.

The Cell Counting Kit-8 (CCK-8) assay was used to measure the cytotoxicity assay. HeLa cells at a density of 8×10³ every well were seeded in a 96-well plate in DMEM with 10% FBS at 37°C for 24 h. Then, the original medium was replaced with 100 μL fresh medium, each well containing CPE with the concentrations of 0, 1, 5, 10, and 20 μM. After 24 h incubation, 10 μL CCK-8 solution was added to each well, and the plate was incubated in the incubator for 1 h. Finally, the absorbance at 450 nm was determined by enzyme-linked immunosorbent assay.

Cells at a density of 1×10⁵ were seeded on 35 mm glass-bottom culture dishes and incubated with 2 ml of DMEM medium supplemented with 10% FBS. After 24 h incubation, cells were incubated with 100 nM PKMTDR and 10 μM CPE for 1 h. Finally, the cells were cultured in a phenol-free medium and imaged under an OMX 3D-SIM. PKMTDR was excited at 561 nm, and CPE was at 405 nm. The images were analyzed using ImageJ.

HeLa cells were stained with 10 μM CPE under different conditions. 37 °C: the cells were stained with CPE at 37 °C for 1 h. 4 °C: the cells were stained with CPE at 4 °C for 1 h. NH₄Cl: the cells were pre-incubated with NH₄Cl (50 mM) in FBS-free DMEM at 37 °C for 2 h, and then incubated with CPE at 37 °C for 1 h.

Statistical analysis was performed with Prism 9 (GraphPad) and ImageJ. Statistical significances and sample sizes in all graphs are indicated in the corresponding figure legends.