A new fluorescent probe for colorimetric and ratiometric detection of sulfur dioxide derivatives in liver cancer cells

A new ratiometric fluorescent probe was constructed with hemicyanine and 7-nitrobenzofurazan for detection of sulfur dioxide derivatives (HSO3−/SO32−). The ratiometric response mode could be attributed to the efficient FRET (Förster resonance energy transfer) platform. The probe exbihited some desirable properties including fast response (within 2 minutes), good selectivity and high sensitivity. Moreover, the probe could detect endogenous HSO3− in liver cancer cells rather than normal liver cells, implying the diagnosal potential of the probe.

Sulfur dioxide is one of the major origins of acid rain, putting serious impacts on water and soil 1 . People who ever exposed to high levels of sulfur dioxide might suffer respiratory and cardiovascular diseases, lung cancer or neurological disorders 2 . However, research revealed that sulfur dioxide could be produced in cytosols and mitochondria of cells 3 . The main raw materials for sulfur dioxide generation in cells are hydrogen sulfide (H 2 S) and sulfur-containing amino acids 3 . On account of its functions in many physiological processes and physiopathology, sulfur dioxide was recognized as a new gaseous transmitter 4 . However, study of sulfur dioxide in living systems is still in its infancy. Lacking reliable analytical methods is one of the main bottlenecks.
Sulfur dioxide can easily dissolve in water (94 mg mL −1 , 25 °C) to form sulfite (SO 3 2− ) and bisulfite (HSO 3 − ) anions. So the physiological effects of sulfur dioxide can be attributed to its derivatives (HSO 3 − /SO 3 2− ). Traditionally, many methods such as titrimetry, chromatography, electrochemistry and capillary electrophoresis are available for the detection of HSO 3 − /SO 3 2− 5-8 . However, these methods usually suffered long operation time, low sensitivity, tedious operation and no applications inside living cells. Alternatively, fluorometric analysis has emerged in recent years, which could be used to detect analytes with high sensitivity and high accuracy 9,10 . More importantly, in situ and real-time imaging of analytes could be carried out by non-destructive fluorescent probes 11,12 .
Here a new ratiometric fluorescent probe (HCy-NBD) was constructed by connecting hemicyanine and 7-nitrobenzofurazan with a piperazine moiety as the non-conjugate "bridge". Both the good overlap of the two bands (fluorescence emission band of 7-nitrobenzofurazan fluorophore and absorption band of hemicyanine fluorophore) and the proper space distance between the two fluorophores benefit the FRET process in HCy-NBD. As a result, the 7-nitrobenzofurazan fluorophore may emit very weak fluorescence while the hemicyanine fluorophore may emit strong fluorescence. Upon nucleophilic addition of HSO 3 Scientific RepoRts | 7:45294 | DOI: 10.1038/srep45294 fluorophore, the conjugated system was broken and the FRET process was blocked, restoring the fluorescence of the 7-nitrobenzofurazan fluorophore. The proposed sensing mechanism was shown in Fig. 1.

Results and Discussion
Response of probe HCy-NBD toward bisulfite in aqueous solution. NaHSO 3 was used as donor of In the presence of 25 equiv. of NaHSO 3 , the emission band peaked at 595 nm decreased gradually, while a new band peaked at 535 nm gradually increased. The intensity ratios of the two emission bands (I 535 /I 595 ) changed by a 61-fold from 0.024 to 1.47. The space between the two bands was much wider to avoid the overlap (Supplementary Table 1) 32 . The reaction could complete rapidly (in 2 min), which was suitable for real-time detection (Fig. 2, Supplementary Fig. 1). Upon addition of NaHSO 3 (0-40 equiv.) to the buffer solution of the probe, I 535 /I 595 changed quantitatively depending on NaHSO 3 concentrations (Fig. 2, Supplementary Fig. 2). A good linear relationship between I 535 /I 595 and the concentrations of NaHSO 3 (0-18 equiv.) was observed. Based on the linearity, the detection limit was determined to be 68 nM (S/N = 3), which was superior to many reported probes [33][34][35] .
Among various anions and biothiols, only NaHSO 3 and Na2SO 3 could lead to naked-eye changes in the colour of probe solutions ( Supplementary Fig. 3). This implied the potential of probe HCy-NBD for colorimetric and selective detection of HSO 3 − /SO 3 2− . The good selectivity was further verified by fluorescence measurements (Fig. 3, Supplementary Fig. 4). Common biological relevant anions and small molecules including CN − , HS − and biothiols hardly brought about significant fluorescence changes. The good selectivity could be well understood  by the strong nucleophilic ability of HSO 3 − /SO 3 2− under neutral conditions. It was reported that HS − could react with 7-nitrobenzofurazan moiety to quench fluorescence 36 . So compound Donor (Supplementary Scheme 1) was used to reveal the reactivity toward HS − . Results showed that the fluorescence of compound Donor was almost unaffected, even with 100 equiv. HS − for up to 1 h ( Supplementary Fig. 5). The reactivity of probe HCy-NBD toward NaHSO 3 was almost unaffected in the presence of various species (Supplementary Fig. 6).
Mechanisms. The solution pH could usually put large impacts on the detection by affecting the state of reactants (NaHSO 3 and HCy-NBD) or the reaction efficiency between them 37 . The fluorescence of probe HCy-NBD was almost constant in the range of pH 4-8. The stronger the basicity of the solution, the easier the reaction between probe HCy-NBD and NaHSO 3 . (Supplementary Fig. 7). So it was the nucleophilicity of NaHSO 3 that was mainly affected by changing its existence forms 38 .
Hemicyanine moiety is electrophilic and is liable to react with nucleophilic HSO 3 − /SO 3 2− . Upon addition of NaHSO 3 , the fluorescence emission band of the Acceptor (Supplementary Scheme 1) diminished but that of the Donor was unaffected (Fig. 4). Furthermore, the UV-vis absorption spectra of probe HCy-NBD in the presence of NaHSO 3 (0-300 μ M) were recorded (Fig. 5). The absorption band peaked at 530 nm which was assigned to hemicyanine moiety decreased and shifted to 490 nm which should belong to the absorption band of the 7-nitrobenzofurazan moiety (Fig. 5b). Meanwhile, a new band centred at 288 nm appeared. High-resolution mass spectroscopy revealed two dominant peaks at m/z 509.2304 and 591.2103, which could be ascribed to [HCy-NBD + H] + and [HCy-NBD + HSO 3 − ] + , respectively. (Supplementary Fig. 8). 1 H NMR titration experiment also revealed the proposed reaction mechanism. The two peaks at about 8.3 and 7.3 ppm, which could be attributed to protons on the C= C double bond of the probe, shifted to 4.8 and 4.9 ppm after reaction with Na 2 SO 3 (Supplementary Scheme 2). Thus HSO 3 − /SO 3 2− was captured by the hemicyanine moiety via Michael addition reaction, which was in consistence with literatures 27,39 .  Absorption spectra of compound Acceptor overlapped greatly with the fluorescence spectra of compound Donor (Fig. 6a). Thus the fluorescence of the 7-nitrobenzofurazan moiety would be quenched because of the energy transfer to the hemicyanine moiety via FRET process 40,41 . The FRET efficiency was determined to be 0.72. Interruption of the conjugated system in the hemicyanine moiety led to its absorption band blue-shifted greatly. So the abovementioned essential spectral overlap disappeared, and 7-nitrobenzofurazan moiety would emit fluorescence as a result. From the view of fluorescence intensity, the fluorescence of probe HCy-NBD was much greater than that of compound Acceptor excited at 345 nm, which further confirmed the FRET process in HCy-NBD (Fig. 6b). Fig. 9) and ignorable cytotoxicity (Supplementary Fig. 10). Thus probe HCy-NBD could be suitable for HSO 3 − /SO 3 2− detection in living cells with minimum interference. Exogenous HSO 3 − imaging experiments were primarily conducted in Hela cells. The ratios of fluorescence intensity (I green /I red ) increased with the gradual increase of exogenous HSO 3 − (Supplementary Fig. 11). L-02 cells treated with NaHSO 3 (100 μ M) and HCy-NBD (5 μ M) showed 8-fold enhancement in the fluorescence intensity ratio compared with control group, which was much more sensitive than the previous reported probe HCy-D (Fig. 7) 32 .

Applications of probe HCy-NBD in living cells. Probe HCy-NBD showed excellent photostability in living cells (Supplementary
Endogenous HSO 3 − could be produced with the help of TST enzyme 42 , which is widespread in nature and especially abundant in human liver cells 43 . HepG2 cells and L-02 cells were separately incubated with HCy-NBD (5 μ M) for 1 h, followed by incubation with GSH (500 μ M) and Na 2 S 2 O 3 (250 μ M) for another 0.5 h. Remarkable fluorescence change was observed only in HepG2 cells (Figs 7 and 8). By contrast, no significant flu rescence change was observed in HepG2 cells incubated with HCy-NBD and GSH. Moreover, HepG2 cells pre-treated with HCy-NBD (5 μ M) and TNBS (10 mM, 2,4,6-trinitrobenzenesulphonate, known as a TST inhibitor) then with GSH (500 μ M) and Na 2 S 2 O 3 (250 μ M) showed no significant fluorescence change (Fig. 8). These results indicated that the endogenous HSO 3 − produced enzymatically in HepG2 cells was responsible for the fluorescence change.   On the whole, endogenous HSO 3 − in liver cancer cells rather than in normal liver cells could be detected, which revealed a new diagnostic feature of liver cancer cells. Therefore, the new way based on cellular level could be promising in liver cancer diagnosis and pathogenesis study of liver cancer.

Conclusions
A new fluorescent probe based on an FRET platform was reported. The probe could detect HSO 3 − /SO 3 2− rapidly, sensitively and selectively. The probe also showed high energy transfer efficiency, good biocompatibility and high reactivity toward HSO 3 − /SO 3 2-. Endogenous bisulfite was successfully detected in liver cancer cells by the probe, which might pave a new way for liver cancer diagnosis and pathogenesis study of liver cancer.

Methods
Apparatus and chemicals. 1 H NMR (300 or 400 MHz) and 13 C NMR (100 MHz) spectra were recorded on a Bruker Avance 300 or 400 spectrometer using CDCl 3 , DMSO-d 6 or D 2 O as solvent and tetramethylsilane (TMS) as an internal standard. HR-MS spectra were recorded on a Q-TOF6510 spectrograph (Agilent). IR spectra were recorded by use of the IR spectrophotometer VERTEX 70 FT-IR (Bruker Optics). Melting points were measured on an XD-4 digital micro-melting point apparatus. Thin-layer chromatography (TLC) was conducted on silica gel 60F 254 plates (Merck KGaA) and column chromatography was conducted over silica gel (mesh 200-300). Fluorescence measurements were conducted on a Perkin-Elmer LS-55 luminescence spectrophotometer, and UV-vis spectra were recorded on a U-4100 UV-Vis-NIR Spectrometer (Hitachi). Quartz cuvettes with a 1 cm path length and 3-mL volume were involved in fluorescence and UV-vis absorption measurements. The pH values were measured by use of a PHS-3C digital pH-meter (YouKe, Shanghai). All reagents were purchased from J&K, Aladdin and Sinopharm Chemical Reagent Co. and used without further purification.

Preparation for UV-vis absorption and fluorescence spectral measurements. Tris-HCl buffer
(50 mM, pH 7.4) was used throughout. HCy-NBD was dissolved in DMSO to get the stock solution (1 mM). Distilled water was used to prepare stock solutions (1 mM) of NaF, NaCl, NaBr, KI, NaHCO 3 , KNO 3 , NaClO, Na 2 SO 4 , KSCN, Na 2 S 2 O 3 , NaHS, Na 2 SO 3 , NaHSO 3 ,(n-Bu) 4 CN, cysteine, homocysteine and glutathione. Stock solutions of NaHSO 3 and Na 2 SO 3 were freshly prepared each time before use. Test solutions were prepared by placing the stock solution of HCy-NBD (100 μ L) and an appropriate aliquot of each testing species solution into a 10-mL volumetric flask, and the solution was diluted to 10 mL with Tris-HCl buffer (50 mM, pH 7.4) containing 40% ethanol (v/v).
Cell imaging. L-02 cells or HepG2 cells were cultured in a 6-well plate in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum in an atmosphere of 5% CO 2 and 95% air at 37 °C. HCy-NBD was dissolved in DMSO to get the stock solution (10 mM) and diluted to 5 μ M each time before use. L-02 cells or HepG2 cells were incubated with HCy-NBD (5 μ M) for 1 h, then treated with exogenous substances. Subsequently, the cell images were taken under a confocal microscope (LSM 700) at emission channels of 405-555 nm (green channel) and 560-700 nm (red channel), respectively.