A mitochondrion-targeted dual-site fluorescent probe for the discriminative detection of SO32− and HSO3− in living HepG-2 cells

Sulfur dioxide, known as an environmental pollutant, produced during industrial productions is also a common food additive that is permitted worldwide. In living organisms, sulfur dioxide forms hydrates of sulfite (SO2·H2O), bisulfite (HSO3−) and sulfite (SO32−) under physiological pH conditions; these three exist in a dynamic balance and play a role in maintaining redox balance, further participating in a wide range of physiological and pathological processes. On the basis of the differences in nucleophilicity between SO32− and HSO3−, for the first time, we built a mitochondrion-targeted dual-site fluorescent probe (Mito-CDTH-CHO) based on benzopyran for the highly specific detection of SO32− and HSO3− with two diverse emission channels. Mito-CDTH-CHO can discriminatively respond to the levels of HSO3− and SO32−. Besides, its advantages of low cytotoxicity, superior biocompatibility and excellent mitochondrial enrichment ability contribute to the detection and observation of the distribution of sulfur dioxide derivatives in living organisms as well as allowing further studies on the physiological functions of sulfur dioxide.


Introduction
Sulfur dioxide (SO 2 ), the most common and simple irritating gas, is one of the main pollutants in the atmosphere. 1 In recent years, increasing physiological functions of sulfur dioxide have been discovered in mammals. [2][3][4] How does SO 2 work in the internal environment of living organisms? It has been reported in numerous studies that SO 2 is not independent in action or directly affects, but dissociated to SO 3 2À and HSO 3 À (SO 2 derivatives) in neutral uid or plasma (HSO 3 À /SO 3 2À , 1 : 3 M/ M), 5,6 which mainly account for its toxicity. There is a dynamic conversion equilibrium between sulfur dioxide and sulte, which also exists in bisulte and sulte. 7 Among them, HSO 3 À /SO 3 2À at high concentrations are catalyzed to generate a variety of sulfur oxy radicals, which are known to give rise to negligible damage to the body. 8 Numerous studies have conrmed that abnormally high sulte levels are closely related to respiratory diseases, 9 cardiovascular diseases 10 as well as neurological diseases, such as migraine, stroke, brain cancer, 11 lung cancer 12 and liver cancer. 13 Besides, clinical studies suggest that the concentration of sulfur dioxide gas ranges from 1 to 2000 mM in living organisms, and the total concentration of serum sulte ranges from 0 to 10 mM in healthy donors. HSO 3 À /SO 3 2À also relax aortic rings in a dosedependent manner at high concentrations ranging from 0.5 to 12 mM. 10 However, whether HSO 3 À and SO 3 2À are independent or synergistic in action remains largely unknown. Therefore, the accurate and independent determination of the levels of SO 3 2À and HSO 3 À is fairly necessary and valuable for further investigating the physiological functions of sulfur dioxide in living organisms.
Over the past decades, uorescence imaging technology has drawn considerable attention beneting from its outstanding performances, such as eminent non-invasiveness, excellent signal-to-noise ratio, high sensitivity, extraordinary reliability, cheap availability and easy operation. 14

Materials and instruments
All the reagents were supplied by commercial suppliers and were directly used without further purication. Absorption spectra were recorded on a UNICO UV-4802 spectrophotometer. Fluorescence spectra were obtained on a uorescence spectrophotometer (Lengguang tech CO., Ltd. F97XP, China). 1 H NMR and 13 C NMR spectra were recorded on a Bruker AVANCE III 400 Nanobay at 500 MHz for 1 H NMR and 300 MHz for 13 C NMR (TMS as an internal standard). High-resolution mass spectra (HRMS) were recorded on a MicrOTOF Bruker. The pH values were measured with an acidity meter (alkalis, pH 400, China).

Synthesis of probe Mito-CDTH-CHO
The probe was synthesized according to the reported literature via a facile two-step reaction. 21,22 The synthesis routes of the probe are depicted in Scheme S1, † and it was characterized via highresolution mass spectrometry, 1 H NMR and 13 C NMR (see ESI †). Synthesis of Mito-CDTH. Freshly distilled cyclohexanone was added dropwise to a solution of concentrated H 2 SO 4 cooled down to 0 C in advance. To a concentrated H 2 SO 4 solution of 2-(4diethylamino-2-hydroxybenzoyl), benzoic acid was added dropwise in freshly distilled cyclohexanone at 0 C. Further, heating up to 90 C, the mixture was vigorously stirred for 2 h, poured into ice, the perchloric acid (70%) was added, the supernatant was ltered off, and the residue was washed with cold water for three times. The residue was dried under vacuum and further puried via silica gel column chromatography (CH 2 Cl 2 : MeOH ¼ 20 : 1, v/ v) to afford Mito-CDTH as a bright red solid (372 mg, 68%). ESI-MS calcd for C 24  Synthesis of Mito-CDTH-CHO. To an acetic acid solution (30 ml) of Mito-CDTH (376 mg, 1 mmol), terephthalaldehyde (268 mg, 2 mmol) was added. The reaction solution was stirred at 110 C for 3 h, and the solvent was evaporated under a reduced pressure. The crude product was extracted with CH 2 Cl 2 (100 ml) and water (300 ml), washed with a saturated ammonium chloride solution and dried over anhydrous sodium sulfate. Aer evaporating using a rotary evaporator, the purple target compound (300 mg, 61%) was obtained via silica gel column chromatography (CH 2 Cl 2 : MeOH ¼ 50 : 1, v/v). 1  , and HSO 3 À ) were prepared by the dissolution of 10 mmol solid in puried water, and diluted to the desired concentrations when needed.

pH value adjustment
The pH values of the solutions were directly obtained by preparing a series of specic pH buffers, including acetate buffer, phosphate buffer, and sodium hydroxide/potassium chloride/boric acid buffer.

Spectral analysis
1 mM of the probe stock solution was prepared by dissolving 1 mg Mito-CDTH-CHO in 2 ml anhydrous ethanol, and diluted with PBS solution (10 mM, pH ¼ 7.4, 6.0 or 8.0, containing 2% EtOH) for nal test solutions.

Cell cytotoxicity assay
The cytotoxicity was measured using a CCK-8 kit. Hela cells were cultured in Dulbecco's modied Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) and 1% antibiotics at 37 C under 5% CO 2 for 24 h. Hela cells were cultured with a fresh medium containing various concentrations of Mito-CDTH-CHO (0-40 mM) for another 12 h. Next, Hela cells were washed three times with PBS and incubated with diluted CCK-8 reagent for 1 h, and then the cell viability was determined by a microplate reader. The procedure was repeated three times for each concentration.

Cell culture and imaging
HepG-2 cells were cultured in a DMEM medium (containing 1% penicillin/streptomycin and 10% FBS) under an air condition at 37 C under 5% CO 2 . HepG-2 cells at the logarithmic growth phase were implanted into 25 mm glass-bottomed dishes and incubated overnight. Aer the attachment of cells, the cells were treated with different pH values (pH ¼ 6.0, 7.4 and 8.0) of the DMEM medium for 3 h. The pH of the DMEM medium was adjusted by adding a specic concentration of hydrochloric acid or sodium hydroxide. 23 HepG-2 cells were washed with PBS three times and incubated with Mito-CDTH-CHO (20 mM) in an untreated DMEM medium. Confocal uorescence images were recorded using a Zeiss LSM 800 confocal laser scanning microscope. The green channel was collected at 460-520 nm at an excitation of 390, and the blue uorescence channel was covered over the range of 420-470 nm at an excitation of 370 nm.

Design and synthesis of Mito-CDTH-CHO
According to the previous reports on the response mechanism of detecting sulfur dioxide type uorescent probes, 24-30 we proposed that (Scheme 1), on the one hand, the oxygen positive ion on the benzopyran ring acts as a strong electronwithdrawing group, which reduces the electron cloud density of the C]C double bond and enable the C]C double bond strong electrophilicity. On the other hand, the C]O double bond conjugated to the benzene ring also possesses weaker  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 26349-26357 | 26351   ) in PBS buffer. There were mainly two absorption peaks at 300-700 nm, centered at 330 nm and 553 nm, respectively. Mito-CDTH-CHO exhibited similar UV spectra changes aer it reacted with SO 3 2À (Fig. 1a) and HSO 3 À (Fig. 1b). The absorption intensity at 553 nm sharply decreased, which means for the breaking of the conjugate system, and a new absorption at 365 nm was elevated.
To explain the uorescence response distinction, the dualsite uorescence response of probe toward SO 3 2À and HSO 3 À was investigated. As shown in Fig. 1c, the probe Mito-CDTH-CHO (20 mM) had almost no uorescence in the absence of SO 3 2À /HSO 3 À at 400-600 nm. However, aer the reaction with 50 of mM Na 2 SO 3 , the strong green uorescence was emitted (l ex ¼ 390 nm, l em ¼ 492 nm). When at the same concentration in the case of NaHSO 3 , the solution exhibited luminous blue emission at a shorter wavelength (l ex ¼ 370 nm, l em ¼ 456 nm). These results provide a preliminary proof that the probe Mito-CDTH-CHO conjugate structure is destroyed by SO 3 2À and HSO 3 À . As shown in Fig. 1d and 2a, for SO 3 2À , with the addition of the Na 2 SO 3 (0-600 mM), the emission intensity at 492 nm increased signicantly (pH ¼ 7.4, 37 C, l ex ¼ 390 nm), and an excellent linear relationship (R 2 ¼ 0.992) was obtained in the range of 10-100 mM. Moreover, the detection limit was calculated to be 100 nM. For HSO 3 À , Fig. 2b and c revealed that the uorescence intensity at 456 nm constantly increased aer the addition of 0-1000 mM NaHSO 3 in the phosphate buffer solution (pH ¼ 6.0, 37 C, l ex ¼ 370 nm), in a wide linear range (40-200 mM). The detection limit was 80 nM (Fig. 2c).
Aerward, we evaluated the selectivity and pH stability of the probe Mito-CDTH-CHO towards sulfur dioxide derivatives. As described in Fig. 2d    , the probe had little uorescence and was unaffected by the variation of pH values (Fig. S5 †). When Na 2 SO 3 or NaHSO 3 was added, the uorescence intensity changed with the mutual conversion balance between SO 3 2À and HSO 3 À in the range of pH 4 to 10 ( Fig. S6 †). In the range of acidic pH (4-6), HSO 3 À ion dominates, thus Mito-CDTH-CHO exhibited stronger uorescence at 456 nm than in neutral and weak basic pH ranges (Fig. S7 †). In basic pH ranges (7-10), SO 3 2À accounts for the main part, so the uorescence intensity increased with the pH value increase (Fig. S6 †).
In general, compared to other uorescence probes sensing SO 2 , the most evident superiority of Mito-CDTH-CHO is selectivity for SO 3 2À and HSO 3 À . Most uorescent probes for detecting SO 2 are not selective toward SO 3 2À and HSO 3 À due to their very similar chemical properties, showing that the same response toward SO 3 2À (HSO 3 À ) when detecting HSO 3 À (SO 3 2À ).
In addition, superior water solubility, suitable detection limit, and accurate mitochondrial targeting performance also indicate that Mito-CDTH-CHO is a fairly qualied uorescent probe for the accurate detection of sulfur dioxide derivatives (see Table S1 †).

The proposed mechanism of Mito-CDTH-CHO for SO 2 derivatives detection
In order to illuminate the reaction mechanism between Mito-CDTH-CHO and SO 2 derivatives, the NMR titration experiments in DMSO-d 6 /D 2 O (8 : 2) were performed. As shown in Fig. 3, a proton at 6.34 ppm represents the double bond conjugated to benzopyrone of the probe Mito-CDTH-CHO. As the concentration of Na 2 SO 3 increased from 1 to 10 eq., the proton peak disappeared, while the singlet at 4.89 ppm appeared. The response mechanism of the probe Mito-CDTH-CHO to NaHSO 3 was also conrmed (Fig. 4), except the singlet at 4.89 ppm, an additional hydroxyl proton peak at 5.12 ppm emerged, which is attributed to the difference in nucleophilicity between SO 3 2À and HSO 3 À . These results further veried that the dual-site sensing of Mito-CDTH-CHO toward Na 2 SO 3 and NaHSO 3 via different double bond nucleophilic addition reactions.

Cellular imaging of Mito-CDTH-CHO
In view of the excellent performances of the probe Mito-CDTH-CHO in vitro, the capability of the discriminative detection of SO 3 2À and HSO 3 À was investigated. Prior to bioimaging experiments, the cytotoxic assay was carried out by a CCK-8 method in HepG-2 cells, the results indicated that Mito-CDTH-CHO had low cytotoxicity (Fig. 5).
According to the literature, 31-33 cationic small molecules could enter into mitochondria and interact with anionic species via the electrostatic interaction. The design of Mito-CDTH-CHO is based on our considerations that the benzopyran cation (containing oxygen positive ions) can act as a mitochondriontargeting moiety. The positive charge and hydrophobic properties of the benzopyran cation are supposed to mediate the localization of Mito-CDTH-CHO inside the mitochondrial membrane. Thus, we speculate that Mito-CDTH-CHO will mainly distribute in the mitochondria. In order to verify our hypothesis, the mitochondrial colocalization experiment was carried out. The commercial mitochondrion tracker (Mito-tracker) and Mito-CDTH-CHO were co-incubated in HepG-2 cells. The uorescence imaging from Mito-CDTH-CHO in the blue channel ( Fig. 6c and g) overlapped well with the Mitotracker in the green channel ( Fig. 6b and f), resulting in the Pearson's correlation coefficient of 0.98. Furthermore, the region of interest (ROI) is illustrated in Fig. 6j, and the normalized uorescence intensity of Mito-CDTH-CHO changed in coordination with the normalized uorescence intensity of Mito-Tracker Green. These results suggested that Mito-CDTH-CHO possesses the excellent ability to target mitochondrial of subcellular organelle in HepG-2 cells.
For the sake of better experimental results, we conducted the control experiments with a lyso-tracker. As illustrated in Fig. 7, the green uorescence of Mito-CDTH-CHO is not overlapped at all with the red uorescence of LysoTracker RED with the Pearson's correlation coefficients (R r ) of 0.3308 and an overlap coefficient (R) of 0.6277 (Fig. 7e). The green uorescence of Mito-CDTH-CHO and red uorescence of LysoTracker RED changes in the intensity proles of ROIs are not synchronized at all (Fig. 7f). The result further indicates that Mito-CDTH-CHO mainly localizes in the mitochondria of living cells.
HepG-2 cells were pre-treated with a probe (20 mM) in the DMEM medium and then incubated with Na 2 SO 3 (200 mM) for 30 min. As shown in Fig. 8, due to the equilibrium conversion between SO 3 2À and HSO 3 À in a neutral uid, HepG-2 cells exhibited distinct uorescence (a 1 ) in the green channel and weak uorescence (a 2 ) in the blue channel. In contrast, aer incubating the HepG-2 cells with NaHSO 3 (200 mM) for 30 min, clear uorescence in the blue channel (b 1 ) and weak uorescence (b 2 ) in the green channel were observed. Inspired by the above experimental results, HepG-2 cells were cultured with NaHSO 3 (200 mM) for 30 min, it is worth noting that the uorescence in the green channel disappeared (d 1 ), whereas the uorescence in the blue channel enhanced (d 2 ). Similarly, in the HepG-2 cells incubated with Na 2 SO 3 in the DMEM medium for 30 min, there was uorescence enhancement in the green channel (c 1 ) and almost no uorescence in the blue channel (c 2 ). Therefore, probe Mito-CDTH-CHO can detect the intracellular SO 3 2À and HSO 3 À levels with different uorescence signals.

Conclusions
In short, to the best of our knowledge, for the rst time, a dualsite uorescence probe for HSO 3 À and SO 3 2À with two different emission signals was designed and synthesized. Mito-CDTH-CHO can distinguishingly sense the levels of HSO 3 À and SO 3 2À with different uorescence signals under separate pH conditions in living biological systems and possesses low cytotoxicity, excellent biocompatibility and outstanding mitochondrial targeting. In the meantime, the sensing mechanism of the double bond nucleophilic addition was successfully validated using the NMR titration experiments. We envision that Mito-CDTH-CHO could provide a deeper insight into and a better understanding of the physiological and pathological processes of SO 2 derivatives.

Conflicts of interest
There are no conicts to declare.