1. Introduction

The heavy and transition metals (HTMs) contamination has sparked interest in worldwide owing to their damage to environment and human health [1-7]. Among those HTMs, Hg²⁺ is one of the most dangerous and ubiquitous pollution in environment that the Environmental Protection Agency (EPA) standard for the maximum allowable level of inorganic Hg(Ⅱ) in drinking water is 2 ppb, because biomethylation of Hg²⁺ to MeHg⁺ occurs in ground-surface water bodies by various aquatic microorganisms (mainly sulfate-reducing bacteria) accumulated in the trophic chain [8-11]. Besides, Hg²⁺ could easily pass through biological membranes, skin and gastrointestinal tissues, and then accumulate in body leading to serious diseases of central nervous system, such as kidney failure, prenatal brain damage, cognitive and motion disorders, vision and hearing loss and even death [12-20].

Due to serious harm of Hg²⁺ to the environments and human health, it is quite necessary to develop a rapid and low-cost method for detection of Hg²⁺ with high sensitivity and selectivity. Traditional methods including atomic absorption/emission spectroscopy, high-performance liquid chromatography and inductively coupled plasma mass spectrometry remain limited by their expensive instrumentation and sophisticated sample preparation [21-25]. However, fluorescence detection with Hg²⁺-responsive probes not only meet the urgent demand for facile, sensitive, selective and cost-effective detection techniques of Hg²⁺, but also offer a promising approach for simple and rapid tracking of Hg²⁺ in biological, toxicological and environmental monitoring [26-32].

Generally, the strategy in the design of a fluorescent probe involves the incorporation of a recognition moiety into a fluorophore [33]. Among various fluorophores, rhodamine has been widely used as imaging reagents because of its high quantum yield, cell membrane permeability, nontoxic nature towards live cells [34-37]. Schiff base structural motif, an excellent coordination site, could be used in identification and quantitative analysis of metal ion on account of its electronegativity. Hydrazone compounds, a kind of Schiff base compound obtained by modified hydrazide compound, have better biological activity, lower toxicity to organisms [38-40]. It is generally known that Hg²⁺ is a representative example of soft acid and sulfur is a soft base based on the Hard-Soft-Acid-Base (HSAB) theory, which makes sulfurbased functional group could serve as good binding site for Hg²⁺ [41-44]. According to the above principle, our group designed a series of probes for specific detection of Hg²⁺ and investigated their structure-property relationship caused by steric-hindrance effect around targeting site [45]. Herein, in order to further figure out the impact on performance of fluorescence probes by the other crucial factor (electronic effect), we have introduced strong electron withdrawing group into the recognition moiety to gain other three probes. These probes had specific recognition to Hg²⁺, accompanied by high sensitivity, wide pH response range and strong antiinterference. More importantly, these probes exhibited ultralow limit of detection (LOD) than our previous probes (Fig. 1). It was also found that the fluorescence performance reduced by the electron withdrawing substituents and heavy atom effect, which provided a good guiding for the further screening of specific substituents and better design of synthetic probes.

All the reagent grade chemicals consumed in this work were procured from commercial sources and used as received. Hg²⁺ stock solution is prepared to be 0.1 mmol/L with ethanol using HgCl₂ as the source, and probe M1-M3 stock solution were prepared to be 0.1 mmol/L with ethanol and solubilized by little dichloromethane. The solutions of metal ions were prepared in ethanol for the selective and competitive experiments. To a 10 mL volumetric tube, different concentration of Hg²⁺, 5.0 mL 20 mmol/L PBS and 1.00 mL of 0.1 mmol/L solution of probe were added. The mixture was diluted to 10 mL with ethanol. Then, 3.0 mL each solution was transferred to a 1 cm quartz cell. The excitation and emission wavelength of slit width were both set as 5.0 nm and the excitation wavelength was set at 520 nm.

NMR spectra were performed on a Bruke-AVANCE Ⅲ 400 MHz spectrometer (at microTOF-Q Ⅱ ESI-Q-TOF LC/MS/MS Spectroscopy). IR spectra were recorded in KBr disks on a Bruker Tensor 27 spectrometer. The absorbance spectra were performed on a Shimadzu UV-1700 spectrophotometer. Fluorescent spectra measurements were collected by a Hitachi F-4500 fluorescence spectrophotometer equipped with a xenon discharge lamp and 1 cm quartz cell. Living mice imaging was performed on PerkinElmer Lumina LT Serios Ⅲ, 3% pentobarbital sodium normal saline solution were used for anesthesia.

Rhodamine 6G, hydrazine hydrate (80%), Lawesson's reagent, cinnamaldehyde, α-chlorocinnamaldehyde and α-bromocinnamaldehyde was obtained from Beijing InnoChem Science & Technology Co., Ltd. Analytical thin layer chromatography was performed using Merck 60 GF254 silicagel (precoated sheets, 0.25 mm thick). Silica gel (0.200–0.500 mm, 60 A, J&K Scientific Ltd.) was used for column chromatography. Probe M1-M3 stock solutions (100 μmol/L) were prepared in ethanol. Stock solutions of metal ions were prepared in ethanol. The solutions of metal ions were prepared with hydrochloride salts of K⁺, Li⁺, Na⁺, Ba²⁺, Ca²⁺, Cd²⁺, Co²⁺, Cu²⁺, Fe²⁺, Hg²⁺, Mg²⁺, Mn²⁺, Ni²⁺, Pb²⁺, Zn²⁺, Al³⁺, Cr³⁺, Fe³⁺, Sn⁴⁺ and the nitrate salt of Ag⁺. Double distilled water was used throughout the experiment. Chloride and nitrate salts of metal ions were all analytical reagent grade.

The synthetic route of the probe M1-M3 is depicted in Fig. 2. The probes were prepared through a three-step procedure with rhodamine 6G as the starting materials. We use Lawesson's reagent to give rhodamine 6G thio-hydrazide after the formation of lactam from rhodamine 6G and hydrazine hydrate in order to detect specifically Hg²⁺, which was further treated with cinnamyl aldehyde derivatives in EtOH solution at reflux temperature to yield the target probes. The structures of M1-M3 were all confirmed by IR, NMR, and MS spectra (Supporting information). The structural differences of the probes M1-M3 caused by introducing different halogen family, which were further investigated to have obvious effects on the fluorescent properties.

Rhodamine 6G (4.8 g, 0.01 mol) was dissolved in ethanol solution (50 mL) and stirred at room temperature, a solution of hydrazine hydrate (8.0 mL) was added dropwise over 30 min, and then the mixture was heated to reflux for 4 h and the color of solution changed. After the reaction, the reaction mixture was cooled to room temperature. The resulting precipitate was collected by filtration and washed with cold water, and then recrystallized from ethanol/water to afford rhodamine 6G hydrazine as a white crystal (3.2 g, yield 74.4%).

Newly prepared rhodamine 6G hydrazide (4.3 g, 0.01 mol) and Lawesson's reagent (4.1 g, 0.01 mol) and were dissolved in dry toluene, and the reaction mixture was stirred for 12 h under N₂ atmosphere. After removing the solvent under reduced pressure, the residue was purified by column chromatography on silica gel (eluent: dichloromethane) and give a white powder (1.4 g, yield 31.8%). Mp: 180–181 ℃. MS (ESI) (m/z): Calcd. for [C₂₆H₂₈N₄OS + H]⁺: 445.2057; Found: 445.2088. ¹H NMR (400 MHz, CDCl₃): δ 8.11 (d, 1H, J = 7.2 Hz), 7.53–7.42 (m, 2H), 7.07 (d, 1H, J = 7.2 Hz), 6.40 (s, 2H), 6.13 (s, 2H), 4.83 (s, 2H), 3.56 (m, 2H), 3.22 (q, 4H, J = 6.4 Hz), 1.90 (s, 6H), 1.32 (t, 6H, J = 6.8 Hz). ¹³C NMR (100 MHz, CDCl₃): δ 166.12, 152.14, 151.67, 147.46, 132.53, 129.76, 128.06, 127.60, 123.73, 122.95, 117.92, 104.76, 96.73, 65.98, 38.29, 16.68, 14.70. IR (KBr, cm^-1): 3415.44, 2966.31, 1622.06, 1517.69, 1374.26, 1452.25, 1422.87, 1363.86, 1343.13, 1276.29, 1199.53, 1160.38, 1097.41, 1012.47, 952.50, 871.36, 827.24, 774.96, 745.81, 642.68, 455.38.

Rhodamine 6G thio-hydrazine (4.5 g, 0.01 mol) and the cinnamyl aldehyde derivatives (0.01 mol) were mixed in 50 mL ethanol and the reaction mixture was heated to reflux. After reaction for 24 h, the reaction mixture was cooled to the room temperature, the resulting precipitate was collected by filtration and washed with cold ethanol. Column chromatography (eluent: DCM/PE = 2: 1) of the crude product over silica gel to give M1-M3.

M1: Yellow powder, yield 61.8%. Mp: 179–180 ℃. MS (ESI) (m/z): Calcd. for [C₃₅H₃₄N₄OS + H]⁺: 559.2526; Found: 559.2508. ¹H NMR (400 MHz, DMSO-d₆): δ 8.69 (d, 1H, J = 8.3 Hz), 7.90-7.86 (m, 1H), 7.58-7.49 (m, 4H), 7.33-7.25 (m, 3H), 7.00-6.96 (m, 1H), 6.84-6.70 (m, 2H), 6.32 (s, 2H), 5.06 (t, 2H, J = 5.3 Hz), 3.18-3.10 (m, 4H), 1.86 (s, 6H), 1.21 (t, 6H, J = 7.1 Hz). ¹³C NMR (100 MHz, DMSO-d₆): δ 163.86, 151.91, 150.81, 150.68, 147.67, 139.63, 135.62, 133.66, 128.82, 128.64, 128.50, 128.26, 127.14, 126.78, 126.38, 123.55, 122.89, 118.22, 104.99, 95.86, 65.45, 37.46, 16.96, 14.16. IR (KBr, cm^-1): 3408.53, 2966.51, 1620.58, 1596.55, 1511.76, 1476.98, 1449.64, 1417.34, 1346.08, 1263.87, 1214.70, 1155.60, 1012.61, 976.63, 944.16, 841.99, 810.99, 751.49, 732.94, 689.89.

M2: Yellow powder, yield 43.5%. Mp: 180–182 ℃. MS (ESI) (m/z): Calcd. for [C₃₅H₃₃ClN₄OS+]⁺: 593.2136; Found: 593.2126. ¹H NMR (400 MHz, CDCl₃): δ 8.37 (s, 1H), 8.13-8.07 (m, 1H), 7.84 (d, 2H, J = 2.0 Hz), 7.46-7.33 (m, 5H), 7.12 (s, 1H), 7.10-7.06 (m, 1H), 6.57 (s, 2H), 6.27 (s, 2H), 3.46 (s, 2H), 3.24-3.15 (q, 4H, J = 6.8 Hz), 1.91 (s, 6H), 1.30 (t, 6H, J = 6.8 Hz). ¹³C NMR (100 MHz, CDCl₃): δ 173.43, 157.47, 155.89, 150.20, 146.93, 136.24, 134.85, 134.00, 132.51, 130.35, 130.20, 129.49, 128.73, 128.48, 127.85, 127.15, 122.33, 117.92, 110.43, 96.27, 63.78, 38.39, 16.81, 14.78. IR (KBr, cm^-1): 3588.80, 3528.82, 3408.84, 2966.91, 1618.39, 1510.79, 1477.00, 1449.54, 1417.37, 1346.24, 1263.85, 1214.75, 1155.63, 1012.92, 976.68, 944.03, 841.92, 810.39, 751.50, 732.88, 689.84, 460.62.

M3: Yellow powder, yield 52.3%. Mp: 166–168 ℃. MS (ESI) (m/z): Calcd. for [C₃₅H₃₃BrN₄OS + H]⁺: 637.1631; Found: 637.1524. ¹H NMR (400 MHz, CDCl₃): δ 8.32 (s, 1H), 8.14-8.08 (m, 1H), 7.86 (d, 2H, J = 6.4 Hz), 7.46-7.31 (m, 6H), 7.10-7.08 (m, 1H), 6.56 (s, 2H), 6.27 (s, 2H), 3.46 (s, 2H), 3.19 (q, 4H, J = 7.2 Hz), 1.91 (s, 6H), 1.30 (t, 6H, J = 6.8 Hz). ¹³C NMR (100 MHz, CDCl₃): δ 173.60, 158.25, 155.89, 150.23, 146.94, 139.57, 134.88, 134.68, 132.52, 130.20, 130.04, 129.56, 128.35, 127.87, 127.17, 122.35, 119.91, 117.94, 110.48, 96.28, 63.75, 38.39, 16.82, 14.79. IR (KBr, cm^-1): 3416.38, 2926.42, 1619.52, 1601.26, 1516.06, 1471.10, 1453.60, 1418.96, 1346.30, 1268.84, 1216.57, 1155.01, 1009.92, 948.71, 815.76, 763.17, 695.23.

To test the validity of these probes, the spectral properties of M1-M3 were investigated with the common metal ions including Ag⁺, K⁺, Li⁺, Na⁺, Ba²⁺, Ca²⁺, Cd²⁺, Co²⁺, Cu²⁺, Fe²⁺, Hg²⁺, Mg²⁺, Mn²⁺, Ni²⁺, Pb²⁺, Zn²⁺, Al³⁺, Cr³⁺, Fe³⁺ and Sn⁴⁺ in EtOH-H₂O (5/5, v/v) solution (10 mmol/L PBS, pH 7.4). As expected, these solutions showed nearly no fluorescence change upon these analytes except Hg²⁺ under the laser irradiation of 520 nm (Fig. 1). It was obvious that these probes had excellent recognition performance and high specificity for Hg²⁺ with the concomitant color changes from colorless to orange. Besides, to value the effect of co-existing ions on this recognition process, competition experiments (Fig. 3) were then performed in the same testing environment. None of them could induce distinct fluorescent variation at 560 nm, revealing these probes have good anti-interference ability.

To better understand correlation of absorption/fluorescence intensity and the concentration of Hg²⁺, UV–vis and fluorescent responses of M1-M3 (10 μmol/L) to Hg²⁺ were measured systematically upon addition of Hg²⁺ with a gradually additive concentration range of 0 ~ 25 μmol/L to in EtOH-H₂O (5/5, v/v) solution (10 mmol/L PBS, pH 7.4). M1-M3 exhibited similar spectroscopic properties on Hg²⁺ detection, while the absorption and fluorescent intensity order is M1 > M2 ≈ M3, which might be caused by electron withdrawing effect and heavy atom effect (Fig. 4). In the absence of Hg²⁺, free M1-M3 exhibited virtually no absorption and emission. However, the addition of Hg²⁺ gave rise to evident enhancement of absorption and emission ascribed to their structural change from the spirocyclic to the ring-open form. An intense absorption peak progressively increased at 530 nm with a good linear (R = 0.9958, 0.9954, 0.9955) (Fig. 4, Fig. S1 in Supporting information). Concomitantly, it should be noted that a striking fluorescence amplification (overall 40-fold) at 560 nm emission band when the concentration of Hg²⁺ reached 2.0 equiv. (Fig. 4). The probes showed a nice linear relationship (R = 0.9984, 0.9945, 0.9974), which indicated that the probes could be potentially useful for the quantitation of Hg²⁺ (Fig. S2 in Supporting information). According to these linear relation, LOD of M1-M3 were determined to be 1.08 nmol/L, 1.24 nmol/L, 2.04 nmol/L. The fluorescence quantum yields of M1-M3 were calculated to be 0.53, 0.38, 0.37. The association constant of Hg²⁺ and M1-M3 were 4.8 × 10³ L/mol, 0.49 ×10³ L/mol, 2.0 ×10³ L/mol, respectively. The calculation details of LOD, fluorescence quantum yields and association constant were described in Supporting information.

The fluorescence intensity of M1-M3 were controlled by the electron-withdrawing groups and heavy atom effect in Hg²⁺ sensing. For one thing, when electron-withdrawing substituent (Cl and Br) were introduced to the recognition unit of probes, it is easy to find that fluorescent intensity of M2 and M3 were much lower than that of M1 in the Hg²⁺ sensing. Such result demonstrated electron-withdrawing effect resulted in the binding ability decreasing of probe and Hg²⁺, which could also be proven by the association constant. For another, approximate fluorescence intensity of M2 and M3 for Hg²⁺ is most likely attributable to coefficient of the fluorescence enhancement by the relatively higher association constant of M3 and fluorescence reduction by heavy atom effect (fluorescent intensity would be reduced with the atomic number of halogens increased).

To evaluate the pH tolerance of these probes, fluorescence response of M1-M3 to Hg²⁺ were also investigated between 1.7 to 13.0 (Fig. S3 in Supporting information). The fluorescence intensity increased when pH value was lower than 5, due to protonationinduce ring opening in these probes. The fluorescence intensity of M1-M3 with Hg²⁺ exhibited negligible changes in the pH range from 5.0 to 8.0 and decreased at pH > 8.0 (Fig. S3). The experimental results insured that probes could exhibit good fluorescence response towards Hg²⁺ in neutral pH range, which suggests that these probes could be promising in biological applications.

To verify whether the reaction of probes with Hg²⁺ were caused by the ring-opening mechanism, Job's method of fluorescence intensity was applied by keeping the sum of the concentration of Hg²⁺ and probes at 10 μmol/L (the molar ratio of [probe] to [probe + Hg²⁺] changing from 0 to 1). As showed in Fig. S4 (Supporting information), the fluorescence intensity of this system reached to maximum when the ratio of [probe]/[probe + Hg²⁺] arrived at 0.5, indicating that the binding stoichiometry of probe to Hg²⁺ is 1: 1. In order to further confirm this binding mode, mass spectrum of M1 + HgCl₂ was selected as a representative to be investigated. The cala. value for [M1 + Hg²⁺ + Cl^-]⁺ is 795.1848, the peak at m/z 795.1783 was assigned as [M1 + Hg²⁺ + Cl^-]⁺, providing a powerful evidence for the 1: 1 binding complex (Fig. S22 in Supporting information). The proposed binding mode of probes with Hg²⁺ was shown in Fig. 5, Hg²⁺ cooperated with sulfur atom firstly because of its thiophilic property, and then bound to N atom, which resulted in ring-open of thiolactam to generate quinoid structure accompanied by striking fluorescence appeared. In addition, the further mechanism study has been done by the addition of EDTA into the complex solution of M1 and Hg²⁺ (Fig. S5 in Supporting information). Upon addition of EDTA, it was observed that the fluorescence intensity decreased gradually, which manifested that chelating agents such as EDTA could make them recover the original state, further demonstrated the aforementioned mechanism that is chelation, not reaction.

To understand the bio-imaging application of the probes for Hg²⁺ better, the cytotoxicity of probes was investigated using a cell viability assay firstly. MCF-7 cells were treated with 5 mmol/L probe M1-M3 24 h, and then cell viability was evaluated using a MTT assay, which indicated that the probes M1-M3 had low cytotoxicity (Fig. S6 in Supporting information, ). We took probe M1 as a representative to detect the fluorescent imaging in living cells and living mice.

MCF-7 cells with the probes were cultured for 30 min at 37 ℃, and no fluorescence inside the living cells was observed. However, the cells displayed a significant increase in the fluorescence upon addition of Hg²⁺ (1.0 equiv.), which revealed that the fluorescence signals were localized in the perinuclear area of the cytosol, indicating a subcellular distribution of Hg²⁺ (Fig. 6). The experiments in MCF-7 cells demonstrated that the probes could respond to Hg²⁺ in biological system.

Then, Kunming mice were selected as the model to perform the living imaging experiment. And M1 was taken as a representative for this fluorescence imaging due to its prominent recognition performance for Hg²⁺ among these probes. The mice were given subcutaneous injection by M1 and then followed by different concentration (0.0 equiv., 1.0 equiv. and 2.0 equiv.) of HgCl₂ in order. As shown in Fig. 6, the mouse injected by pure M1 exhibited no fluorescence, and the mice treated with probe M1 + Hg²⁺ solution showed remarkable fluorescence enhancements, which manifested that the detection of Hg²⁺ by M1 could give a visual fluorescent signal through fur of mice. In addition, the detection of M1 to Hg²⁺ in vivo displayed certain fluorescent enhancement along with the increasing of Hg²⁺, which has the same trend as its vitro test. As time passed, the fluorescent intensity slowly diminished probably caused by the absorption of mice tissues.

In summary, this study set out to design three small molecule fluorescent probes M1-M3, composed of rhodamine 6G scaffold as a typical fluorescent reporter, C＝N Schiff base and C＝S structural motif as receptor, which is suitable for quantification of Hg²⁺ under physiological conditions with ultralow limit of detection. A series of experimental results indicated that different substituents of cinnamaldehyde had a significant effect on the recognition performance of the probes. Moreover, the bio-imaging experiment demonstrated that M1 holds great potential as a versatile molecular probe platform for quantitative mercury ion in vivo. This design and control strategy could be extended to construct practical fluorescent probes with precise and reliable measurement performance in biological systems. Therefore, we anticipate the design strategy could be extended to construct fluorescent probes with precise and reliable measurement function in biological systems.

Acknowledgments

We thank the National Natural Science Foundation of China (Nos. 21572177 and 21673173), the Shaanxi Provincial Natural Science Fund Project (No. 2016JZ004), the Xi'an City Science and Technology Project (No. CXY1529) for financial support.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cclet.2017.09.027.