A Naphthalimide–Sulfonylhydrazine Conjugate as a Fluorescent Chemodosimeter for Hypochlorite

: Hypochlorite anion (ClO − ) is a widely-used disinfectant and a microbicidal agent in the immune system. Accurate detection of ClO − in environmental and biological samples by simply prepared chemosensors / chemodosimeters is important. Herein, we report that a naphthalimide–sulfonylhydrazine conjugate with an imine (C = N) linker, prepared via simple condensation, acts as an e ﬀ ective ﬂuorescent chemodosimeter for ClO − . The molecule exhibits a weak emission, but ClO − -selective cleavage of its C = N bond creates a strong green emission. Ab initio calculation showed that the emission enhancement by ClO − originates from the suppression of intramolecular electron transfer from the photoexcited naphthalimide through the C = N linker. This response enables selective and sensitive detection of ClO − at physiological pH range (7–9) and allows ﬂuorometric ClO − imaging in the presence of cells.


Introduction
Hypochlorous acid (HClO) is one important reactive oxygen species (ROS). HClO is formed in vivo by the enzymatic reaction of hydrogen peroxide (H 2 O 2 ) with chloride anion (Cl -) in some specialized cells such as neutrophils [1]. Deprotonation of HClO at a physiological pH range produces hypochlorite anion (ClO − ) [2], which acts as an microbicidal agent in immune system [3]. The formed ClO − inhibits the invading microbes, but uncontrolled amounts of ClO − may result in several diseases such as arthritis [4], atherosclerosis [5], cancer [6], and Parkinson's disease [7] because ClO − reacts with several biomolecules [8]. Considerable amounts of HClO are used industrially or at home for sterilization, although high residual amounts of ClO − is hazardous [9]. A simple but inexpensive method to quantitatively detect small amount of ClO − is necessary.
One promising method for this purpose is fluorometric analysis using ClO − -selective molecular chemosensors/chemodosimeters because they facilitate simple quantification or imaging of ClO − with a common fluorescence spectrometer or microscope apparatus [10]. So far, various molecules based on several types of fluorescent dyes such as rhodamine, fluorescein, coumarin, diaminonaphthalene, and BODIPY have been proposed [11,12]. Several excellent molecules have successfully been applied to fluorescence imaging of ClO − in biological samples and quantification of very small amounts of ClO − in environmental samples. Challenges in the design and development of a new fluorescent chemosensors/chemodosimeters with higher reactivity toward ClO − are still ongoing. Among the fluorescent dyes, naphthalimide is one of the popular dyes often used in the synthesis of chemosensors/chemodosimeters because of its high chemical stability, high photostability, a large Stokes shift, and a strong fluorescence quantum yield [13]. There is, however, a few naphthalimide-based molecules for fluorometric detection of ClO − [14][15][16]. Therefore, creating a new naphthalimide-based chemosensor/chemodosimeter is an important issue.
Recently, an imine (C = N) moiety has been used as a linker to connect a fluorescent dye unit with an electron accepting unit in the design of chemosensors/chemodosimeters for ClO − detection [17][18][19]. This is because the C = N linker (i) facilitates an intramolecular electron transfer between the two units; (ii) behaves as a reaction site for ClO − and affects the electron transfer efficiency; and, (iii) can be created by a facile condensation of two reactive groups such as aldehydes and amines. The simple system using a C = N linker may become a versatile design of chemodosimeters for fluorometric detection of ClO − .
In the present work, we synthesized a simple naphthalimide-sulfonylhydrazine conjugate (1) through the C = N linker (Scheme 1) and studied the effects of the sulfonylhydrazine moiety on the reactivity of the C = N reaction site towards ClO − and the change in the emission/optical properties. This molecule, when dissolved in an aqueous solution, shows a weak fluorescence. Addition of ClO − to the solution selectively enhances fluorescence emission via an oxidative cleavage of the C = N moiety. This off-on fluorometric response facilitate rapid, selective, and sensitive detection of ClO − in neutral media with pH 7-9 and also allows ClO − imaging in the presence of cells.
Fluorescence spectra were measured on a JASCO FP-6500 fluorescence spectrophotometer (JASCO Corp., Tokyo, Japan) at 298 ± 1 K [22]. Absorption spectra were measured on an UV-visible spectrometer (Shimadzu Corp., Kyoto, Japan; UV-3600 plus) at 298 ± 1 K [23]. 1 H and 13 C NMR spectra were obtained on a JEOL JNM-ECS400 Spectrometer (JEOL Ltd., Tokyo, Japan). FAB-MS analysis was performed on a JEOL JMS 700 Mass (JEOL Ltd., Tokyo, Japan). Φ F was determined with rhodamine B (in ethanol) as a standard [24], using the refractive indexes for water and ethanol being 1.3334 and 1.3618, respectively [25]. Ab initio calculations were performed within the Gaussian 03 package using the B3LYP/6-3G+(d) basis set for all atoms. Cartesian coordinates of the models are summarized at the end of Supplementary Material.

Biological Experiments
Cell culture and fluorescence microscopy observations were carried out as follows. HeLa cells were grown in the Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum on a cover slip in 100 mm dishes. All of the cells were incubated at 37 • C in humidified air containing 5% CO 2 . The cells were treated with a phosphate buffered saline (PBS)/MeCN (8/2 v/v) mixture containing 1 (50 µM) followed by different concentrations of ClO − for 30 min, respectively. The cells were washed with PBS and subjected for observation on a fluorescence microscopy (BIOREVE BZ-9000, Keyence Corp., Osaka, Japan) [28]. Cell viability tests were performed as follows [29]: Trypsinized HeLa cells were incubated with PBS/MeCN (8/2 v/v) mixture containing 1 (50 µM) for 20 min at 37 • C in humidified air containing 5% CO 2 . After washing with PBS for three times, a Trypan blue solution was added to the cells. The respective live and dead cells were counted using the Countess II FL Automated Cell Counter.

Fluorescence Properties
Fluorescence spectra were obtained in a buffered water/MeCN mixture (8/2 v/v) with pH 7.0 (HEPES 0.1 M) at 25 • C (λ ex = 365 nm). As shown in Figure 1a, 1 exhibits only a weak emission (fluorescence quantum yield, Φ F = 0.01). Addition of 50 equiv of ClOto the solution creates a strong green fluorescence centred at 519 nm (Φ F = 0.15). However, addition of other anions (F − , Cl − , or reactive nitrogen species (RNS) [ONOO − ] barely change the spectrum of 1, indicating that ClO − selectively triggers fluorescence enhancement of 1. It must be noted that, as shown in Figure 1b, the 519 nm fluorescence enhanced by ClO − is unaffected by the other contaminants, suggesting that 1 selectively detects ClO − even in the presence of these contaminants. As shown in Figure 2a, the fluorescence titration of 1 with ClO − increases the intensity of the 519 nm fluorescence. A linear relationship between the fluorescence intensity and ClO − amount (Figure 2b) indicates accurate ClOsensing at~100 µM. The lower detection limit (DL) was determined based on the signal-to-noise ratio using the equation (DL = 3 × SD/S) [30], where SD is the standard deviation of blank measurement (SD = 0.061, n = 11) and S is the slope of the fluorescence intensity versus the ClO − concentration (S = 0.096 µM -1 ). This DL determined (1.9 µM) is below the physiological ClO − concentrations (5-25 µM) in neutrophils [31], suggesting that 1 also facilitates sensitive ClO − detection. The chemodosimeter 1, when left in a solution for 1 day, shows absorption and fluorescence spectra similar to those of the fresh molecule. In addition, the solution, when treated with ClO − , shows spectral responses similar to those of the fresh sample. This means that 1 is storable in solution and can stably be used for ClO − sensing.

Emitting Species
To identify the green emitting species formed by the reaction of 1 with ClO − , we tried to isolate the species from the resulting solution. Concentration of the solution by evaporation followed by a thin-layer chromatography (TLC) analysis (n-hexane/EtOAc = 1/1 v/v) of the resultant ( Figure S8, Supplementary Material) shows a single emitting spot (R f = 0.70), where the R f of 1 is 0.25. This suggests that the reaction of 1 with ClO − produces a single emitting species. 1 H NMR analysis (DMSO-d 6 ) of the fraction eluted using a silica gel column containing the green emitting species showed several impurity protons, where strong protons appear at~2.4 ppm,~7.2 ppm, and~7.6 ppm, respectively, with the integration ratio being 3:2:2, which are assigned to the aromatic protons of the p-toluenesulfonic acid moiety. This indicates that the sulfonylhydrazine moiety of 1 is decomposed during the reaction with ClO − , and the formed fragments elute with the green emitting species owing to their similar high polarities. As a result of this, the isolation and 1 H NMR analysis of the emitting species were unsuccessful.
The crude product was then analyzed by FAB-MS. As shown in Figure S7 (Supplementary Material), 1 exhibits a clear molecular ion at m/z 466.2 (FAB(+) mode), assigned to the [1 + H + ] + species. In contrast, as shown in Figure S9 (Supplementary Material), the crude product does not show a clear molecular ion, but a very strong ion appears at m/z 296.2 (FAB(+)). ClO − oxidizes the C = N moieties and produces C = N cleaved aldehydes or carboxylic acids by hydrolysis [17][18][19]. Therefore, as proposed in Scheme 2 (right top), the m/z 296.2 ion (FAB(+)) may be assigned to the species containing an oxetane moiety ([2 − H 2 O + H + ] + ), which may be produced by intramolecular dehydration of the compound 2 during the FAB ionization. In addition, as shown in Figure S10 (Supplementary Material), the chart in the FAB(-) mode shows a strong molecular ion at m/z 312.08, which is assigned to the [2 − H + ] − species formed by deprotonation of 2 during the FAB ionization (Scheme 2, right bottom). This clearly suggests that the C = N cleaved carboxylic acid (2) is the green emitting species. It is reported that some C = N-based chemodosimeters produce aldehyde as the product by the reaction with ClO − [17,19]. However, as shown in Figure S3

Reaction Mechanism
The proposed sequence (Scheme 2) is confirmed by the fluorescence and absorption spectra of 1 upon reaction with ClO − . As shown in Figure 3a, addition of ClO − increases the 519 nm fluorescence. As shown in the inset, the intensity increases with time and becomes plateaus after 600 s (10 min). During the measurement, the emission wavelengths barely change, suggesting the formation of a single emitting species (2), which is supported by the single emitting spot observed in TLC analysis ( Figure S8, Supplementary Material). As shown in Figure 3b, the ClO − addition immediately changes the absorption spectrum of 1 within 1 min (blue → red) with an isosbestic point at 311 nm. The spectrum further changes with a decrease in the absorbance at 327 nm with an isosbestic point at 454 nm, finally exhibiting a peak at 429 nm. As shown by the inset, the time profile for the decrease in the 327 nm absorbance is consistent with the profile for the fluorescence intensity increase (inset of Figure 3a). These findings indicate that, as shown in Scheme 2, the reaction of 1 with ClO − produces some C = N cleaved intermediates, and further oxidation by ClO − with hydrolysis produces the green emitting species (2). The hydrolysis is indeed involved in the reaction mechanism (Scheme 2). Figure 4 shows the change in fluorescence intensity of 1 after addition of ClO − in the MeCN solutions with different water amounts. Without water, 1 shows only a small fluorescence enhancement by ClO − . Increasing the water amount in the solution increases the enhancement by ClO − . This indicates that, as shown in Scheme 2, hydrolysis of the intermediates produces the carboxylic acid emitting species (2). It must also be noted that the fluorescence enhancement by ClO − is reduced in solution with >90% water. This is probably because the decreased solubility of 1 suppresses the reaction with ClO − , suggesting that sufficient dissolution of 1 in solution is necessary for the fluorometric ClO − detection.
It must be noted that 1 H NMR analysis of the product indicates that the reaction of 1 with ClO − produces p-toluenesulfonic acid as a byproduct. This means that 1, when used for biological experiments, leaves p-toluenesulfonic acid in the body or cell systems. It has been clarified that sulfonic acid compounds have a low systemic toxicity and are neither mutagenic nor carcinogenic regardless of the structure and the route or duration of the application [32]. In addition, some sulfonic acid-containing molecules have successfully been used for cell imaging [33,34]. These findings indicate that the p-toluenesulfonic acid byproduct may not have a negative effect on the biological applications.

Ab Initio Calculations
To further confirm 2 as the green emitting species, structures and optical properties of 1 and 2 were calculated by the density functional theory (DFT) and the time-dependent DFT (TD-DFT), respectively [35]. As shown in Table S1 (Supplementary Material), singlet electronic transition of 1 mainly consists of HOMO → LUMO+1 (S 0 → S 4 ) transition. Its calculated transition energy (4.03 eV, 308 nm) is close to the absorption maximum (λ max ) of 1 at 327 nm ( Figure 3b). As shown in Figure 5 (left), π-electrons of the main orbitals of 1 such as HOMO, LUMO, and LUMO+1 are located on the naphthalimide moiety and the C = N moiety. This indicates that photoexcited electrons on the naphthalimide moiety are transferred to the p-toluenesulfonyl moiety via the C = N linker. This may result in almost no fluorescence of 1.  Table S1 (Supplementary Material), the electronic excitation of 2 consists of HOMO → LUMO (S 0 → S 1 ) transition. Its energy (2.98 eV, 416 nm) is consistent with the absorption band of 1 at 429 nm after addition of ClO − (Figure 3b). As shown in Figure 5 (right), π-electrons of both HOMO and LUMO are located on the naphthalimide moiety, indicating that the C = N cleavage of 1 suppresses electron transfer from the photoexcited naphthalimide moiety. This may facilitate strong fluorescence from the carboxylic acid (2). These data support the proposed reaction mechanism of 1 with ClO − (Scheme 2).

Effect of pH
The chemodosimeter 1 facilitates ClO − sensing at neutral pH (7)(8)(9). Figure 6 shows the fluorescence intensity of 1 at different pH, with and without ClO − . Also shown are the mole fractions of HClO and ClO − , calculated based on their equilibrium in water [36]:
(1) ClO − enhances the intensity of 1 at a neutral pH (7)(8)(9). At pH < 7, protonation of ClO − (HClO formation) cancels the basicity of ClO − and inhibits the reaction with the C = N moiety of 1. At pH > 9, ClO − is stabilized in solution, but the fluorescence enhancement decreases significantly. This is because, as reported for some ClO − chemodosimeters [37][38][39], the oxidation ability of ClO − decreases in basic media and inhibits the reaction. Therefore, 1 facilitates fluorometric sensing of ClO − in physiological pH range (7-9).  To clarify the toxicity of the cell treatment, viability of the HeLa cells before and after treatment with a PBS/MeCN (8/2/v/v) solution containing 1 was determined. A Trypan blue solution was added to the cells, and the respective live and dead cells were counted. The cell viability after the treatment was only ca. 11%, although the cell viability before the treatment was ca. 96%. This means that the cell treatment with the PBS/MeCN solution containing 1 leads to a loss of the functionality for almost all of the cells. In the present case, the use of MeCN with high cytotoxicity is necessary for permeation of the chemodosimeter into the cells due to its low solubility in water. Living cell imaging is therefore not facilitated by the present chemodosimeter.

Conclusions
We demonstrated that a simple naphthalimide-sulfonylhydrazine dye (1) acts as a fluorescent chemodosimeter for ClO − . 1 shows almost no fluorescence, but ClO --selective oxidative cleavage of its C = N bond creates a strong green fluorescence owing to the suppression of intramolecular electron transfer from the photoexcited naphthalimide moiety through the C = N bond. 1 facilitates relatively rapid (<10 min), selective, and sensitive ClOdetection (detection limit: 1.9 µM) at pH 7-9. The C = N linker may therefore become a versatile design for highly efficient fluorescent chemodosimeters for ClO − . However, fluorometric imaging of ClOin living cells was unsuccessful because use of organic solvent is necessary due to the low solubility of 1. In addition, the imaging at physiologically relevant ClO − concentrations cannot be performed due to the low reactivity of 1 with ClO − . Replacement of p-toluenesulfonyl hydrazine moiety to other moieties may affect the solubility to aqueous solution and the reactivity towards ClO − . These improvements are considered essential for the living cell imaging.  Table S1: TD-DFT calculation results. Figure S1: 1 H NMR chart of 3. Figure S2: 13 C NMR chart of 3. Figure S3: FAB(+)-MS chart of 3. Figure S4: 1 H NMR chart of 1. Figure S5: 13 C NMR chart of 1. Figure S6: 1 H-1 H COSY chart of 1. Figure S7: FAB(+)-MS chart of 1. Figure S8: TLC results. Figure