Phenothiazine and semi-cyanine based colorimetric and fluorescent probes for detection of sulfites in solutions and in living cells

Four hemicyanine probes for selectively detecting sulfites (HSO3−/SO32−) have been constructed by the condensation reaction of 7-substituted (CN, Br, H and OH) phenothiazine aldehyde with 1-ethyl-2,3,3-trimethylindolium iodide. All four probes show a fast and sensitive response to HSO3−/SO32−via a Michael addition, with a detection limit lower than 40 nM based on monitoring their UV/vis absorption changes. Although all four probes display an increase in fluorescence when responding to HSO3−/SO32−, the increment is larger for the probe with an electron-withdrawing group than the probe with an electron-donating group, except for Br. Thus, among four probes the 7-cyano probe (PI-CN) possesses the largest fluorescent response to HSO3−/SO32−, and the lowest detection limit (7.5 nM). More expediently and easily, a film and a test paper with PI-CN have been prepared to detect HSO3−/SO32− in a sample aqueous solution selectively. Finally, the detection of HSO3−/SO32− by PI-CN in biological environments has been demonstrated by cell imaging.


Introduction
Sulfur dioxide (SO 2 ) is both a main atmospheric pollutant and a valuable commercial reagent, and human exposure to SO 2 has become increasingly widespread due to the combustion of fossil fuels and industrial manufacture, for example paper pulp manufacturing and metal processing. More and more medical studies have conrmed that exposure to SO 2 may not only cause respiratory responses, 1 but also lead to lung cancer, cardiovascular diseases, and some neurological diseases, such as migraine headaches and brain cancer. 2 SO 2 dissolves in water to form a pH dependent equilibrium mixture between bisulte and sulte (HSO 3 À /SO 3 2À ) with a molar ratio of about 3 : 1 in neutral aqueous solution. Bisulte and sulte are widely used as preservatives for food and beverages to prevent oxidation and bacterial growth. 3 However, sulte is toxic in high doses, which is associated with allergic reaction and food intolerance symptoms, such as mild to severe skin allergy, asthma, or gastrointestinal diseases. 4 Hence, the Joint FAO/WHO Expert Committee on Food Additives has issued that an acceptable daily intake should be lower than 0.7 mg kg À1 of body weight; the U.S. Food and Drug Administration (FDA) has required the labeling of products containing no more than 10 ppm (125 mM) sulte in foods or beverages. 5 Endogenous HSO 3 À and SO 3 2À can be metabolically generated from thiol-containing amino acids, such as cysteine and glutathione. 6 The studies have showed that HSO 3 À /SO 3 2À have an endothelium-dependent vasorelaxing effect at low concentrations (<450 mM) and also function as messengers in cardiovascular systems. 7 For this reason, the development of new detection methods for HSO 3 À /SO 3 2À is important for environmental security and human health. Due to the advantages of simplicity, sensitivity, nontoxicity, and ease of operation, uorescent probes have been recognized as efficient molecular tools for visualizing anions in living systems. 8 In recent years, some excellent uorescent probes have been reported by various effective reactions of HSO 3 À /SO 3 2À including nucleophilic addition to an aldehyde, 9 mediated levulinate cleavage, 10 nucleophilic addition to an unsaturated bond, 11 and others. 12 However, there are still some drawbacks in reported probes such as poor selectivity over biothiols or H 2 S for most aldehydeor levulinate-based uorescent probes, long response time (>5 min for most probes), high detection limits (>1 mM for half of probes). Hence, it is still a challenge to develop more convenient and quick response probes for HSO 3 À /SO 3

2À
. Additionally, further application of lm or test paper detection based on PI-CN and bioimaging in living cells were performed.

Reagents and instrumentation
All the chemicals for synthesis were purchased from commercial suppliers and were used as received without further puri-cation. 1 H and 13 C NMR spectra were measured in CDCl 3 or DMSO-d 6 with a Bruker AV spectrometer operating at 400 MHz and 100 MHz, respectively and chemical shis were reported in ppm using tetramethylsilane (TMS) as the internal standard. Mass spectra were obtained with a Thermo LTQ Orbitrap mass spectrometer. UV-vis absorption and uorescence emission spectra were recorded with a Shimadzu UV-2450 UV/vis spectrometer and a Shimadzu RF-5301pc luminescence spectrometer, respectively.

Preparation of sample solutions
Sample solutions for the measurement were EtOH/PBS (v/v 1 : 3) solvent mixtures, the concentration of PBS is 10 mM. All concentrations of the probes for test were 15 mM in above solution. Water for sample solution preparation was puried with a Millipore water system. All pH values were measured on a MQK PHS-3C pH meter.

Cell culture and cell images
HeLa cells were seeded on the coverslips in 24-well plates at a density of 50 000 cells per well and incubated in a humidied 5% CO 2 atmosphere for 24 h with the complete Dulbecco's modied eagle's medium (DMEM) containing 10% fetal calf serum at 37 C. 10 mL of DMSO solution of PI-CN (1 mM) was added to a well to give concentration of 10 mM and volume ratio of DMSO/culture medium is 1 : 100. Aer incubation for 30 min, and then the cells were washed three times with PBS buffer aer culture medium was removed. The cells were further treated with 0.1 mM NaHSO 3 for 30 min, then removed the culture medium and washed three times with PBS buffer. The cellular localization was visualized under a laser scanning confocal microscope (LSM 710 Meta, Carl Zeiss Inc., Thornwood, NY). The green uorescence of cells was collected with 476-526 nm channel under excitation at 405 nm.

Synthesis of probes
A mixture of aldehyde and 1-ethyl-2,3,3-trimethyl-3H-indol-1ium iodide (1 mmol, 315 mg) was dissolved in EtOH (5 mL). The reaction mixture was stirred for 12 h at 80 C under an N 2 atmosphere. The mixture was cooled to room temperature, solvent was evaporated in vacuo and compound was puried by silica gel using DCM : MeOH (v/v 25 : 1) as the eluent. Product probe was obtained as dark purple solid.
Structures of four probes were fully characterized by 1 H NMR, 13 C NMR and HRMS analyses.

Photophysical property of probes
As optical probes for HSO 3 À /SO 3

2À
, photophysical properties of four hemicyanines were rst investigated. The four 7-substituents at phenothiazine cover various electronic properties, electron-withdrawing (CN), unsubstituted (H) and electrondonating (OH). Fig. 1 shows UV/vis absorption spectra of four probes in the buffer solution (EtOH/PBS v/v 1 : 3, pH 7.4). As shown in Fig. 1, the absorption maxima red shi gradually from electron-withdrawing group (CN) to electron-donating group (OH), with the corresponding values from 520 nm to 570 nm. However, all four probes have no uorescence emission, which may ascribe to effective photoinduced cis-trans isomerism of the double bond or the formation of a TICT state. When nucleophilic addition of HSO 3 À at the double bonds of probes occurs the phenothiazine moieties could emit uorescence. The photophysical and sensing properties of four probes provided in Table S1. † Spectral response of probes to HSO 3 À /SO 3

2À
Aer addition of 1.0 equiv. of HSO 3 À , the absorption spectra of 15 mM probes in EtOH/PBS (v/v 1 : 3, pH 7.4) decrease rapidly in the long-wavelength region (310-700 nm) and increase in shortwavelength region (<310 nm) (Fig. 2 for PI-CN, Fig. S1 for other three probes provided as ESI †). The four probes exhibit "turnon" uorescent response to HSO 3 À , with the most obvious for PI-CN, and the weakest for PI-Br. The latter weak uorescence should ascribe to the heavy-atom effect of bromine. The scaffold phenothiazine is an electron rich chromophore, and modied by an electron-withdrawing group (EWG) to form a molecule with ICT character, which would reveal a longer-wavelength uorescence emission. As shown in Fig. 2   four probes (inset of Fig. S1 †) show that the sensing reaction of PI-CN could is the fastest. Hence, based on the fact that PI-CN can sense HSO 3 À /SO 3 2À by remarkable changes by in both absorption and uorescence, PI-CN was selected as a representative in next experiments. Above results reveal 7-substituent has effect on the photophysical property, the sensing reactivity and the corresponding spectral response of four probes. Among them, remarkable effects are the inuence of 7-substituents to the absorption spectra of probes and luminescence property of sensing products. However, there is small difference in the sensing reactivity and no observable effect on their luminescence.
The spectral response of four probes can be observed directly by naked eyes. The solutions of all four probes display dark purple, and no observable uorescence emission. Aer addition of HSO 3 À , the color of all probe solutions change from red or blue to colorless (Fig. 3 le), and only the solution of PI-CN shows bright uorescence under portable UV lamp (Fig. 3 right).

The sensing mechanism
To verify the sensing mechanism, 1 H NMR titration of PI-CN with NaHSO 3 was performed. As shown in Fig. 4, the chemical shis at 8.33 ppm and 7.54 ppm were assigned to the proton H B and H A in PI-CN, respectively. Aer addition of NaHSO 3 , the proton signals of H A and H B disappear gradually and two groups of new peaks at 4.87 ppm and 4.72 ppm emerged which are assigned to H b and H a . Meanwhile, other proton signals of the product appear and increase such as methylene at m 0 and n 0 sites.
Moreover, the formation of the adduct PI-CN-NaHSO 3 was conrmed by high-resolution mass spectroscopy, where   a dominant peak at m/z value of 582.1860 (calcd 582.1861) corresponds to [PI-CN + NaHSO 3 ] + provided in Fig. S2. † Therefore, the sensing reaction was conrmed to be the nucleophilic addition of the probe with HSO 3 À .

Selectivity
To evaluate the selectivity of the probe for HSO 3 À /SO 3 2À , we measured the UV/vis absorption (Fig. 5a) and uorescence spectra (Fig. 5b) of PI-CN before and aer the addition of various species, respectively. The absorption and uorescence spectra of PI-CN displayed a large change only in the presence of HSO 3 À , and little change for HS À and CN À . However, other species such as F À , NO 2 À , AcO À , ClO 4 À and biothiols caused no signicant change in both UV/vis absorption and uorescence    spectra of PI-CN. The uorescence proles at 499 nm of the probe showed a remarkably higher selectivity for HSO 3 À over the other species (inset of Fig. 5b). The uorescence increment of PI-CN to HSO 3 À /SO 3 2À is more than 110-fold. Moreover, the anti-interference ability of PI-CN has been studied (Fig. 5c). HSO 3 À induced a great uorescence change, whereas, other species did not result in obvious uorescence change. The coexistence of other species didn't affect the probe's sensing behaviour to HSO 3 À . These results show that the tested species do not interfere with HSO 3 À detection.
Moreover, to measure the detection limit of PI-CN, a titration experiment of HSO 3 À was performed (Fig. 6). With increasing amounts of HSO 3 À (0-15 mM), spectral change of PI-CN is similar to time-dependent spectral change of PI-CN aer addition of 1 equiv. HSO 3 À . From the plot of the orescence intensity at 499 nm vs. the concentration of HSO 3 À , the detection limit of PI-CN was obtained to be 7.5 nM in terms of the signalto-noise ratio (S/N ¼ 3). Moreover, detection limits of the four probes to HSO 3 À were measured by colorimetric analysis, 22 nM for PI-CN, 28 nM for PI-Br, 27 nM for PI-H and 37 nM for PI-OH (Fig. S3 †). These values are much lower than the standard of 10 ppm (125 mM) required by the U.S. Food and Drug Administration.
The kinetic reaction of PI-CN with NaHSO 3 was further investigated. The time-dependent uorescence response of PI-CN with NaHSO 3 recorded the change of the intensity at 499 nm with time shown in Fig. 7. Aer addition of NaHSO 3 , the uorescence intensity increased rapidly and reached a plateau within 1 min. The observed rate constant (k obs ) of the reaction between PI-CN and NaHSO 3 was obtained, with a high value of, 8.2 Â 10 À2 s À1 .
The sensing behavior can be easily observed by the naked eyes from both the color change and uorescence emission of solutions. As shown in Fig. 8

pH effects and MTT analysis
To assess the function of PI-CN under physiological conditions, the absorption and uorescence spectra of PI-CN before and aer addition of HSO 3 À were recorded under different pH values. The pH-dependent absorption and uorescence responses of PI-CN to HSO 3 À reveal a remarkable change of absorbance and signicant uorescence enhancements at 499 nm under physiological conditions (pH 6-9) (Fig. S4 †).
These results indicate that PI-CN could be used as uorescent probe in a biological system In order to detect HSO 3 À /SO 3 2À in living cells, a MTT analysis was performed to assess the cytotoxicity of the probe. In the MTT assays, HeLa cells were dealt with PI-CN at different concentrations from 10 to 50 mM for 24 h. The results show low toxicity to cultured cells under the experimental condition, and the cell viability is up to 80% for PI-CN at 50 mM (Fig. 11). This result shows that PI-CN is of very low cytotoxicity.

Cell imaging
Finally, the probe PI-CN was utilized for imaging in HeLa cells. HeLa cells were seeded on a 24-well plate in a culture medium for 24 h. The HeLa cells were incubated with PI-CN (10 mM) for   min, followed by PBS washing for three times. Confocal uorescence exhibited no emission for the cells incubated with the probe at respective channels with excitation at 405 nm ( Fig. 12a-c). This implies that PI-CN has no response toward biomolecules such as biothiols in living cells. In contrast, aer further incubated with 0.1 mM NaHSO 3 for 30 min, the HeLa cells emit bright green uorescence (Fig. 12d-f). It shows that the probe PI-CN is a cell membrane permeable uorescent probe and can be achieved to detect HSO 3 À /SO 3 2À in living cells.

Conclusions
In conclusion, four 7-substituted phenothiazine-hemicyanine dyes (PI-CN, PI-Br, PI-H, PI-OH) were synthesized as colorimetric and uorescent probes for the detecting of HSO 3 À /SO 3 2À from the condensation reactions of four different substituted phenothiazine aldehydes with 1-ethyl-2,3,3-trimethylindolium iodide. All probes can sensitively respond HSO 3 À /SO 3 2À via a Michael addition to cause remarkable color change. The sensing reactions displays remarkable substituent effect on uorescence property of their sensing products. Among them, PI-CN can detect HSO 3 À /SO 3 2À fast as both a colorimetric and uorescent probe with the lowest detection limit (7.5 nM). Both acylamide lms and the test papers of PI-CN can detect HSO 3 À / SO 3 2À in solutions fast and selectively, and observed conveniently by naked eyes. Furthermore, cell-imaging experiments reveal that PI-CN can detect HSO 3 À /SO 3 2À selectively in biological environment.

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