A thiosemicarbazone based chemo and fluorogenic sensor for Zn²⁺ with CHEF and ESIPT behaviour: computational studies and cell imaging application

Abstract

We report herein the development of a novel, diformyl-p-cresol (DFC)–thiosemicarbazide (TS) based sensor (DFC–TS) that selectively and sensitively recognizes Zn²⁺ by both UV-Vis and fluorescence methods. The gradual addition of Zn²⁺ to a solution of the ligand developed a new absorption band at 430 nm, while the bands at 370 and 316 nm gradually decrease generating one well defined isosbestic point at 390 nm exhibiting ∼17 fold turn-on fluorescent enhancement (FE). When we plot absorbance (at 430 nm) vs. [Zn²⁺] there is a gradual increase in absorption with [Zn²⁺], becoming saturated at ∼1 equivalent of Zn²⁺ and then again it increases with the increase in [Zn²⁺] and ultimately becomes saturated at ∼2 equivalents of added Zn²⁺. This clearly demonstrates that the Zn²⁺ binding event to the ligand occurs in two steps, one at a time. Non-linear least-squares computer-fitting of these data gives the parameters: K′_f1 = (9.70 ± 5.51) × 10⁵ M⁻¹, n = (1.28 ± 0.05) for the first step and K′_f2 = (1.11 ± 0.65) × 10⁵ M⁻¹ and n = (1.01 ± 0.06) for the second step. So far, this study provides the opportunity where we have successfully, for the first time, determined the stepwise formation constants; though they have values of the same order of magnitude. The ground state geometries of DFC–TS, both enol and keto forms and [Zn(DFC–TS)(OAc)], [Zn(DFC–TS)(OAc)]⁻, and [Zn₂(DFC–TS)(OAc)₂] were optimized using the Gaussian-03 suit program and bond distances of all species are in reasonable agreement with the reported values.

---

Introduction

The creation of new tools and approaches for optical imaging in biology is an important goal, and investigations focused on these objectives will ultimately increase our understanding of many biological phenomena. Zn²⁺ is one of the most important metal ions in biology that is highly regulated under normal physiological conditions.¹ The emerging importance of Zn²⁺ in neurological signaling and some proposed functions² in biological systems have generated an urgent demand for the development of Zn²⁺-specific molecular probes,³ and many Zn²⁺ fluorescent sensors have been reported, exhibiting high selectivity and sensitivity over other biologically essential metal ions in specific ranges of concentration.⁴

The mechanistic details of zinc homeostasis are not entirely clear,⁵ partly due to the difficulty in directly mapping the distribution and tracking the movement of the spectroscopically inert free zinc ions. Therefore, zinc ion-selective fluorescent probes have been the primary tools for the detection and quantification of biological free zinc. The past decade has witnessed the rapid progress in the development of zinc probes,⁶ which have enabled significant advances in the cell biology of zinc.⁷

Though the chemistry of thiosemicarbazones (Scheme 1) has been studied for a considerable period of time for their biological properties, the first report on their medicinal applications started to appear in the fifties as drugs against tuberculosis and leprosy.⁸ In the sixties their antiviral properties were discovered and a huge amount of research was carried out that eventually led to the commercialization of methisazone, Marboran® etc. to treat smallpox⁹ During this period the first antitumor activity results was published.¹⁰ Recently, triapine (3-aminopyridine-2-carboxaldehyde thiosemi-carbazone) has been developed as an anticancer drug for several cancer types.¹¹

Scheme 1

Understanding the excited-state proton/hydrogen-atom transfer Excited State Proton Transfer (ESPT) reaction is fundamental in chemistry and biology.¹² Taylor reported the first case of excited state biprotonic transfer in 7-azaindole¹³ and Tahara¹⁴ proved that this proton transfer mechanism is a concerted process with the initial formation of a dimer in solution. Villani¹⁵ recently proved that the trans-formation of the Watson–Crick (WC) DNA base pairs to their tautomers is achieved via a concerted double-hydrogen transfer process beginning with a hydrogen atom of a purinic base. The ESIPT reaction in green fluorescent protein (GFP) has also attracted much interest,¹⁶ because it is a major tool in many areas of the life science. ESIPT molecules have attracted attention for various applications including chemical sensors,¹⁷ luminescent materials¹⁸ and functional polymers.¹⁹ A distinctive feature for the ESIPT molecules is that their fluorescence is well separated from their absorption maxima (>100 nm). This resulting unusually large Stokes shift is an obvious advantage over other normal fluorophores, such as fluorescein, rhodamine or coumarin, as they can avoid self-absorption and thus the inner filter effect in fluorescence analysis.

In the present article we are going to disclose a simple and easily synthesizable DFC–TS based turn-on colori- and fluorimetric dual sensor for Zn²⁺.The promise of this new ligand as a quantitative probe for Zn²⁺ is demonstrated by its use in the estimation of affinities towards Zn²⁺ through absorption and fluorescence changes occurring mainly due to ESIPT blocked Chelation Enhanced Fluorescence (CHEF) effect. Most striking feature of this study is that it enables us, for the first time, the direct determination of stepwise formation constants (K′_f1 and K′_f2); though they are of same order of magnitude and otherwise, very difficult to determine by any suitable analytical techniques. In addition, the ligand displays ESIPT behavior with large Stokes shift (>155 nm) that can avoid self-absorption and thus the inner filter effect in fluorescence analysis.

Experimental section

Materials and methods

The starting materials such as thiosemicarbazide (Sigma Aldrich), 2,6-diformyl-p-cresol (DFC, prepared in the laboratory) were used for the preparation of ligands DFC–TS. Zn(OAc)₂·2H₂O was used to prepare Zn²⁺ complexes. Solvent like DMSO, MeCN etc. (Merck, India) were of reagent grade.

Physical measurements

Elemental analyses were carried out using a Perkin–Elmer 240 elemental analyzer. Infrared spectra (400–4000 cm⁻¹) were recorded from KBr pellets on a Nickolet Magna IR 750 series-II FTIR spectrophotometers. ¹H-NMR was recorded in DMSO-d₆ on a Bruker 300 MHz NMR spectrometer using tetramethylsilane (δ = 0) as an internal standard. UV-Vis spectra were recorded on an Agilent diode-array spectrophotometer (Model, Agilent 8453), steady-state fluorescence were recorded on a Shimadzu spectrofluorimeter (Model RF-5301.) and PTI spectrofluorimeter (Model QM-40.), ESI-MS⁺ (m/z) of the ligand and Zn(II)-complex were recorded on a Waters' HRMS spectrometer (Model: XEVOG2 QTOF).

Preparation of DFC–TS

2,6-Diformyl-p-cresol (DFC) was prepared by following a literature procedure.²⁰ 2,6-Diformyl-p-cresol (DFC) (1.64 g, 10 mmol) was dissolved in 25 ml EtOH under nitrogen atmosphere. To this solution was added thiosemicarbazide (1.82 g, 20 mmol) and stirred at room temperature for 5 h and then the reaction mixture was filtered to remove any undissolved materials and the filtrate was kept at room temperature. After 1 day a crystalline product was deposited. (Yield, 85.57%). ¹H-NMR (in DMSO-d₆) (δ, ppm): 2.24 (s, 3H, methyl), 2.48 (2-H, NH), 7.6285 (s, 2H, azomethine), 8.31 (s, 2-H, aromatic), 8.06 and 8.16 (s, 4H, NH₂), 11.45 (s, 1-H, OH) (please see Fig. S1†). ESI-MS⁺ (m/z): 311.1094 (DFC–TS + H⁺), 316.0230 (DFC–TS + Li⁺), 333.0256 (DFC–TS + Na⁺). [Fig. S7†].

Synthesis of [Zn(DFC–TS)(OAc)] and [Zn₂(DFC–TS)(OAc)₂]

To a solution of DFC–TS (0.62 g, 2 mmol, dissolved in few drops of DMSO) in 20 ml MeCN was added Zn(OAc)₂·2H₂O (0.498 g, 2 mmol in 10 ml MeCN solution) dropwise with continuous stirring. After 1 hour the solution was filtered and the filtrate was kept aside undisturbed. After one day crystalline [Zn(DFC–TS)(OAc)] was precipitated out. It was collected by filtration, washed several times with MeCN and dried in air. [Zn₂(DFC–TS)(OAc)₂] was prepared in the same way as that of 1, only 2 equivalents of Zn(OAc)₂·2H₂O was used here. Several trials to grow single crystals were failed. ¹H-NMR (in DMSO-d₆) (δ, ppm): 2.19 (s, 3H, methyl), 2.48 (2-H, NH), 6.93 (s, 2H, azomethine), 8.07 (s, 2-H, aromatic), 6.24 and 6.42 (s, 4H, NH₂) for 1 : 1 Zn²⁺ complex and 2.19 (s, 3H, methyl), 2.52 (2-H, NH), 7.31 (s, 2H, azomethine), 8.31 (s, 2-H, aromatic), 6.43 (s, 4H, NH₂) for 1 : 2 Zn²⁺ complex, (please see Fig. S2 and S3†). ESI-MS⁺ (m/z): 372.9884, (DFC–TS + Zn²⁺) (1 : 1), 514.9107 ([Zn₂(DFC–TS)(OAc)(H₂O)]⁺) (please see Fig. S8 and S9†).

Solution preparation for UV-Vis and fluorescence studies

For both UV-Vis and fluorescence titrations, a stock solutions of 1.0 × 10⁻³ M of the probe DFC–TS was prepared by dissolving the DFC–TS in 9 ml DMSO and finally the volume was made upto 10 ml by DMSO. Similarly, another 1.0 × 10⁻³ M stock solution of Zn²⁺ was prepared in de-ionised H₂O. A solution of 10 mM HEPES buffer (9 : 1 DMSO : H₂O) was prepared and pH was adjusted to 7.2 by using HCl and NaOH. 2.5 ml of this buffer solution was pipetted out into a cuvette to which 20 μM of the probe was added and Zn²⁺ ion was added incrementally starting from 0 to 46 μM in a regular interval of volume and UV-Vis and fluorescence spectra were recorded for each solution. Path lengths of the cells used for absorption and emission studies were 1 cm. Fluorescence measurements were performed using 2 nm × 2 nm slit width.

For competition assays 1.0 × 10⁻³ M Na₂H₂EDTA solution was prepared in water. 2.5 ml of the buffered solution was pipetted out into a cuvette to which 20 μM of the DFC–TS was added and Zn²⁺ ion was added in 1 : 1 and 1 : 2 mole ratio separately and then Na₂H₂EDTA solution was added incrementally starting from 0 to 44 μM in a regular interval of volume and UV-Vis and fluorescence spectra were recorded for each solution.

K_comp = [M–L²][L¹]/[M–L¹][L²]; K_comp = K_L²/K_L¹

Emission spectra were recorded by excitation at 430 nm. Both excitation and emission slit widths were 2 nm. The equilibrium competition constant (K_fi-comp) (i = 1 and 2 for 1 : 1 and 1 : 2 L : M mole ratios respectively) were calculated based on the titration curve.

where, M = Zn²⁺, L¹ = DFC–TS, L² = H₂EDTA²⁻, M–L¹ = Zn²⁺–DFC–TS and M–L² = [Zn²⁺–EDTA]²⁻. Here, M_T is the total concentration of Zn²⁺, L¹_T is the total concentration of DFC–TS, L²_T is the total concentration of H₂EDTA²⁻ and [L²–M] is the concentration of the complex between H₂EDTA²⁻ and Zn²⁺. The fitting of the titration profiles with a non-linear least-squares procedure using a model provides the equilibrium competition constant (K_comp).²¹

Job's study

This method is based on the measurement of a series of solutions in which molar concentrations of two reactants vary but their sum remains constant. The absorbance of each solution is measured at a suitable wavelength and plotted against the mole fraction of one of the reactants. A maximum in absorbance occurs at the mole ratio corresponding to the combining ratio of the reactants.

Computational details

Ground state electronic structure calculations of DFC–TS (both enol and keto forms) and complexes (both mono and dinuclear species) have been carried out using Gaussian-03 suit program.²² Becke's hybrid function²³ with the Lee–Yang–Parr (LYP) correlation function²⁴ were used along with 6-31+G(d,p) basis set for H, C, N, O atoms and LANL2DZ effective core potentials and basis set for the Zn atom. The global minima of all these species were confirmed by the positive vibrational frequencies. Time dependent density functional theory (TDDFT)^25,26 with B3LYP density functional associated with the conductor-like polarizable continuum model (CPCM)²⁷ were applied to study the low-lying excited states of the ligand and the complexes in DMSO using the optimized geometry of the ground (S₀) state. The vertical excitation energies of the lowest 20 singlet states are also computed here (Table S4†).

Cell culture

HepG2 cell lines, human hepatocellular liver carcinoma cells, were procured from National Center for Cell Science, Pune, India, and used throughout the study. Cells were cultured in DMEM (Gibco BRL) supplemented with 10% FBS (Gibco BRL), and a 1% antibiotic mixture containing Penicillin, streptomycin and gentamicin (Gibco BRL) at 37 °C in a humidified incubator with 5% CO₂ and cells were grown to 60–80% confluence, harvested with 0.025% trypsin (Gibco BRL) and 0.52 mM EDTA (Gibco BRL) in phosphate-buffered saline (PBS), plated at the desired cell concentration and allowed to re-equilibrate for 24 h before any treatment.

Cell cytotoxicity assay

To test the cytotoxicity of DFC–TS assay was performed as per the procedure described earlier.²⁸ After treatment with DFC–TS at different doses of 1, 10, 20, 50 and 100 μM, respectively, for 12 h, 10 μl of MTT solution (10 mg ml⁻¹ PBS) was added to each well of a 96-well culture plate and again incubated continuously at 37 °C for a period of 3 h. All media were removed from wells and 100 μl acidic isopropyl alcohol was added into each well. The intracellular formazan crystals (blue-violet) formed were solubilized with 0.04 N acidic isopropyl alcohol and absorbance of the solution was measured at 595 nm with a microplate reader (Model: THERMO MULTI SCAN EX). The cell viability was expressed as the optical density ratio of the treatment to control. Values were expressed as mean ± standard errors of three independent experiments. The cell cytotoxicity was calculated as % cell cytotoxicity = 100% − % cell viability.

Cell imaging study

HepG2 Cells were incubated with 10 μM DFC–TS [the stock solution (1 mM) was prepared by dissolving DFC–TS to the mixed solvent (DMSO : water = 1 : 9 (v/v))] in the culture medium, allowed to incubate for 30 min at 37 °C. After incubation, cells were washed twice with phosphate-buffered saline (PBS). Bright field and fluorescence images of HepG2 cells were taken by a fluorescence microscope (Leica DM3000, Germany) with an objective lens of 40× magnification. Fluorescence images of HepG2 cells were taken separately from another set of experiment where cells incubated with 10 μM DFC–TS + 10 μM Zn²⁺ for 30 min. Similarly, in another set of experiment, cells were incubated with 10 μM DFC–TS + 10 μM Zn²⁺ for 30 min followed by addition of 100 μM EDTA for another 30 min and fluorescence images were taken. HepG2 cells showed almost complete quenching of fluorescence due to removal of Zn²⁺ from the DFC–TS–Zn²⁺ complex.

Results and discussion

2,6-Diformyl-p-cresol (DFC, 1)−thiosemicarbazide (TS, 2) based (DFC–TS) conjugate was synthesized by a simple Schiff base condensation in EtOH as outlined in Scheme 1. This ligand has potentially penta-coordinating N₂S₂O donor atoms that can bind to one or two metal ions leading to the formation of mono- and dinuclear complexes. The DFC–TS ligand and its Zn²⁺ complexes have been isolated and characterized by various physicochemical (NMR (Fig. S1–S5†), IR (Fig. S6†) and ESI-MS⁺ (Fig. S7–S9†) (m/z)) methods.

UV-Vis absorption studies

Intramolecular hydrogen bonded organic chromophores often exhibit the photophysical phenomenon like (ESIPT). The process of ESIPT is an intramolecular version of the chemistry of photo acids and photo bases whereby a molecule contains both acidic and basic groups in appropriate proximity and the acidic proton can migrate to the basic site upon excitation. Most ESIPT processes are aided by either six-or five-membered strong intramolecular hydrogen bonds in the ground state and usually require very little movement of nuclei. When solvent is not required to mediate such proton transfer process, it is described as intrinsic ESIPT and is barrier less in non-polar solvents occurring in the sub-picoseconds time region.²⁹ The first reported ESIPT reaction is observed in methyl salicylate.³⁰

The electronic absorption spectral properties of DFC–TS were investigated in 9 : 1 DMSO–H₂O at pH 7.2 HEPES buffer. The absorption spectrum of DFC–TS presents two important transitions with intense absorption bands at 316 and 370 nm. As Scheme 2 describes, there is an equilibrium between the two configurations, enol and keto tautomers of DFC–TS due to ESIPT. The band at 370 nm is assigned to the transition to quinoid (keto) form. On gradual addition of Zn²⁺, the absorption bands of DFC–TS at 370 and 316 nm decrease gradually; however, a new peak appears at 430 nm thereby generating one well defined isosbestic point at 390 nm indicating a clean transformation of free ligand to its metal bound state (Fig. 1, S10 and S11†). This clearly indicates the chelation of Zn²⁺ by the multidentate ligand moiety giving [[Zn(DFC–TS)]⁺ and [Zn₂(DFC–TS)]⁺]. The spectral change (decrease in absorbance at 370 nm and increase at 430 nm) clearly indicate that the addition of Zn²⁺ ion induces the decrease in concentration of free ligand both in keto and enol forms. Here the nitrogen atom of the azomethine group takes part in the coordination with Zn²⁺ and the formation of complexes Zn–DFC–TS and Zn₂–DFC–TS inhibit the ESIPT (Scheme 2).

Scheme 2

Fig. 1 (a) Absorption titration of DFC–TS (20 μM) with Zn²⁺ (0–46 μM) in DMSO–H₂O (9 : 1, v/v, pH 7.20, 10 mM HEPES buffer). Inset shows (b) absorption vs. [Zn²⁺] plot; (c) Job's plot; (d) non-linear least squares curve-fit of absorption titration data for the first and second formation steps.

The cation binding affinities of sensor toward Zn²⁺ was investigated by UV-Vis absorption studies and when a change in absorbance at 430 nm was plotted against [Zn²⁺], there is a gradual increase in absorbance with [Zn²⁺], gets saturated at ∼1 equivalent of Zn²⁺ and then it again increases with the increase in [Zn²⁺] and finally get saturated at ∼2 equivalent of added Zn²⁺ (inset (b), Fig. 1). This clearly demonstrates that the Zn²⁺ binding event to the ligand occurs in two steps, one at a time. Non-linear least-squares fit of the data to eqn (1)

y = (a + b × c × xⁿ)/(1 + c × xⁿ),²¹ (1)

where a and b are the absorbance in the absence and presence of excess metal ions, c (=K′) is the apparent formation constant and n is the stoichiometry of the reaction gives the parameters: K′_f1 = (9.70 ± 5.51) × 10⁵ M⁻¹, n = (1.28 ± 0.05) for the first step, and K′_f2 = (1.11 ± 0.65) × 10⁵ M⁻¹ and n = (1.01 ± 0.06) for the second step of complex formation, (Fig. 1 and Table 1).

Table 1 Formation constant values in different spectroscopic methods

Formation constant	Spectrophotometrically	Fluorometrically
K′f1	(9.7 ± 5.51) × 10^5 M^−1	(7.17 ± 0.2) × 10^4 M^−1 [(1.14 ± 0.47) × 10^5 M^−1; λex = 430 nm]
K′f2	(1.11 ± 0.65) × 10^5 M^−1	—
K′f1 (competition)	(2.28 ± 0.32) × 10^5 M^−1	(1.38 ± 0.29) × 10^5 M^−1
K′f2 (competition)	(1.64 ± 0.17) × 10^5 M^−1	(7.2 ± 2.3) × 10^4 M^−1

---

We have also attempted to determine the formation constants by competition method by reacting the in situ generated complexes like [Zn(DFC–TS)] and [Zn₂(DFC–TS)] with Na₂H₂EDTA and equilibrium competition constants (K′_f1 and K′_f2) were calculated from the plots of (A_max − A)/A_max as a function of H₂EDTA²⁻ concentration by adopting displacement model. The evaluated apparent formation constants are: K′_f1 = (2.28 ± 0.32) × 10⁵ M⁻¹ and K′_f2 = (1.64 ± 0.17) × 10⁵ M⁻¹ and corresponding formation constants K_f1 and K_f2 were calculated from the relation K_fi = K′_EDTA/K′_fi where K′_EDTA = K_EDTA × α₄ = 2.55 × 10¹³ and α₄ = K₁K₂K₃K₄/([H⁺]⁴ + K₁[H⁺]³ + K₁K₂[H⁺]² + K₁K₂K₃[H⁺] + K₁K₂K₃K₄) = 7.98 × 10⁻⁴ at pH 7.20 with K₁ = 1.02 × 10⁻², K₂ = 2.14 × 10⁻³; K₃ = 6.92 × 10⁻⁷ and K₄ = 5.50 × 10⁻¹¹ and K_Zn-EDTA = 3.2 × 10¹⁶.³¹ So the calculated values are K_fi = 1.55 × 10⁸ and K_f2 = 1.11 × 10⁸. The apparent formation constant values (K′_f1 = (2.28 ± 0.32) × 10⁵ M⁻¹ and K′_f2 = (1.64 ± 0.17) × 10⁵ M⁻¹) are found to be very close to the values obtained by non-linear fitting of absorption data as a function of [Zn²⁺] (Fig. S12†).

The composition of the complex was determined by Job's method and it was revealed that an overall composition is 2 : 1 with respect to the Zn²⁺ ion.

We have also spectroscopically determined the acid dissociation constants of the free ligand (DFC–TS) and its Zn²⁺ complex with slightly excess over two equivalents of Zn²⁺ ions by monitoring the absorption of free ligand at 430 nm as a function of pH spanning between 2.0 and 12.0 using the eqn (2)

A = (A_AH [H⁺] + A_A⁻K_a)/(K_a + [H⁺]),^32,33 (2)

where, A = absorbance at any pH, A_AH and A_A⁻ are the absorbance of the fully protonated and fully deprotonated species. The evaluated K_a values are: (3.90 ± 0.64) × 10⁻¹¹ M (pK_a = 10.41) for the free ligand and (1.13 ± 0.10) × 10⁻⁵ M (pK_a = 4.93) for the Zn^II-complex (Fig. 2). The decrease in pK_a value by ∼5.48 may be attributed to the strong complexation of DFC–TS with Zn²⁺ ion and similar order of pK_a decrement was observed previously.³⁴ It is to be mentioned here that the pK_a of phenolic group of the present ligand is quite low (10.41, 9 : 1 DMSO–water, v/v) compared to the free 4-methylphenol (19.1 and 10.2 in DMSO and pure water respectively).³⁵

Fig. 2 Determination of acid dissociation constant of (a) DFC–TS (free) and (b) DFC–TS in presence of 2.1 equivalents of Zn²⁺ in 9 : 1 DMSO–water (v/v) at 25 °C.

The emission spectra of DFC–TS and its fluorescence titration with Zn²⁺ were recorded in DMSO–H₂O (9 : 1, v/v, pH 7.20, HEPES buffer) (Fig. 3) at λ_ex = 390 nm. The binding between the free DFC–TS and Zn²⁺ leads to chelation enhanced fluorescence (CHEF) effect causing ∼17 fold fluorescence intensity enhancement due to increased structural rigidity of the formed complexes (Fig. 3).³⁶ On excitation at 390 nm, the isosbestic point in absorption spectra, there is gradual increase in fluorescence intensity (FI) at λ_em = 500 nm, with the increase in [Zn²⁺] upto 1 : 1 mole ratio, after that on increasing [Zn²⁺] there is a blue shift in λ_em from 500 to 490 nm (Fig. 3). This observation clearly indicates the stepwise uptake of Zn²⁺ ions in the ligand framework. The linear dependence of FI as function of [Zn²⁺] upto 1 : 1 mole ratio was analyzed with the help of eqn (1) in the form y = a + b × c × x, under the conditions 1 ≫ c × x and n = 1. It is interesting to note that linear least-squares analysis of fluorescence titration data gives K′_f (7.17 ± 0.20) × 10⁴ M⁻¹. In addition, at λ_ex = 430 nm there is a gradual blue shift of λ_em from 520 nm for pure ligand to 500 nm on complexation with Zn²⁺ and may be a consequence of slight involvement of Intramolecular Charge Transfer (ICT).³⁷ It is interesting to note that non-linear least-squares analysis of fluorescence titration data gives K′_f (1.14 ± 0.47) × 10⁵ M⁻¹ and n = (1.03 ± 0.04) and it is obvious that on gradual addition of Zn²⁺ fluorescence intensity at λ_em 500 nm increases and gets saturated at ∼DFC–TS : Zn²⁺ = 1 : 1 when exited at 430 nm (Fig. S13†). On further addition of Zn²⁺ there is no visible increase/decrease in fluorescence intensity.

Fig. 3 Fluorescence titration of DFC–TS (20 μM) with Zn²⁺ (0–48 μM) in DMSO–H₂O (9 : 1, v/v, pH 7.20, 10 mM HEPES buffer), λ_ex = 390 nm, R² = 0.99.

Again, when we added Na₂H₂EDTA to an in situ generated [Zn(DFC–TS)] (1 : 1 L : M) and [Zn₂(DFC–TS)] (1 : 2 L : M) complexes the observations are same as in case of absorption titration i.e. there is a gradual decrease in absorbance with the increase in Na₂H₂EDTA concentration (Fig. 4 and S14†). However, here the phase separation (2 : 1 and 1 : 1 complexes) is quite distinct thereby facilitating the determination of K_f1 and K_f2 unambiguously which were evaluated to be K_f1 = 1.03 × 10⁸ and K_f2 = 0.92 × 10⁸. These values are in good agreement to those obtained by absorption titration data.

Fig. 4 (a) Plot of FI as a function of [Na₂H₂EDTA ] for the competitive displacement of Zn²⁺ from [Zn₂(DFC–TS)] ensemble (20 μM with respect to DFC–TS) by Na₂H₂EDTA (0–44 μM) in DMSO-H₂O (9 : 1, v/v) at pH 7.2 (HEPES buffer) and (b) non-linear fitting of data using eqn (1).

The stoichiometries of the Zn²⁺ complexes of DFC–TS as delineated by UV-VIS and fluorescence titrations were further confirmed by ESI-MS⁺ (m/z) mass spectrometry. The prominent ESI-MS⁺ peak of the pure ligand at 311.1094, 316.0230 and 333.0256 corresponds to [DFC–TS + H⁺], [DFC–TS + Li⁺] and, [DFC–TS + Na⁺], respectively. The mass spectrum of the isolated [Zn(DFC–TS)]⁺ complex at m/z = 372.9884 corresponds to [Zn(DFC–TS)]⁺ indicating a 1 : 1 complexation in the initial stage. However the presence of a very small peak for 1 : 2 complex [Zn₂(DFC–TS)(OAc)(H₂O)]⁺ in solution is also apparent (Fig. S8†). When we prepared the solid product by reacting DFC–TS and Zn²⁺ in 1 : 2 mole ratio the a prominent ESI-MS⁺ peak appears at 514.9107 (m/z) that indicating the presence of complex species [Zn₂(DFC–TS)(OAc)(H₂O)]⁺ in solution. However no peaks appear for 1 : 1 complex.

The coordination modes were further supported by ¹H-NMR studies. ¹H-NMR peaks appear at 7.68, 8.31 and 11.45 ppm, correspond to azomethine proton (a), aromatic proton (b) and phenolic proton (c) and also peaks at 8.07 and 8.16 correspond to NH₂ proton. The appearance of two separate peaks for NH₂ proton may be due to the presence of two NH₂ groups in two different chemical environments. The appearance of OH proton signal at >10.0 ppm signifies the presence of H-bonding. This peak intensity decreases significantly for [Zn(DFC–TS)]⁺ (1) and vanishes completely in case of dinuclear complex (2) indicating that phenolate group in DFC–TS participates in complexation with Zn²⁺ but in solution there is certain extent of free ligand in the case of complex 1. The ¹H-NMR peak of azomethine proton (a) shifted to high field region (lower ppm, 6.93 for complex 1 and 7.31 for complex 2) reflecting the fact that the azomethine nitrogen participated in complexation. The aromatic protons remain almost invariant in the three species. Similar is the case for methyl proton. The coordination through sulphur atom to the Zn²⁺ is confirmed by IR spectra, where free ligand gives IR strectching frequency centered at 1110 cm⁻¹ for C=S bond (normal range 1050–1200 cm⁻¹) (Fig. 5) which on complexation splits into two peaks appearing at 1110 and 1090 cm⁻¹. The shifting to smaller frequency clearly points out the lowering in electron density around the C–S bond and corresponds to the existence of thiolate ion.

Fig. 5 IR spectra of DFC–TS ligand (shown in black), (Zn–DFC–TS) complex (shown in green), (Zn₂–DFC–TS) complex (shown in red) in solid KBr-pellets.

The detection of Zn²⁺ was not perturbed by biologically abundant Na⁺, K⁺, Ca²⁺, Mg²⁺, Al³⁺ etc. metal ions. Several transition metal ions, namely Cr³⁺, Mn²⁺, Fe²⁺, Fe³⁺, Co²⁺, Ni²⁺, Cu²⁺, and heavy metal ions like Cd²⁺, Pb²⁺, and Hg²⁺, also caused no interference (Fig. 6). The sensor was found to bind Zn²⁺ reversibly as tested by reacting with excess EDTA in intra-cellular conditions and also with EDTA in extra-cellular conditions. Moreover, no significant change in emission spectra was observed even on excess addition of the above mentioned interfering metal ions.

Fig. 6 Fluorescence and histogram plots to indicate the selectivity of DFC–TS toward Zn²⁺ over other biologically relevant metal ions, λ_ex = 430 nm.

For practical application, the appropriate pH conditions for successful operation of the sensor were evaluated. The probe DFC–TS fluoresces rather very weakly between pH 2 and 8 while Zn–DFC–TS complexes fluoresce extensively between pH 5 and 8 in 10 mM HEPES buffer, clearly indicating that this pH range is suitable for fluorescence studies for the complexation with Zn²⁺ (Fig. 7). The fluorescence of free ligand DFC–TS increases enormously beyond pH 10.0 and may be due to the acid dissociation of the thioamide proton.

Fig. 7 pH depence of FI of the free ligand, DFC–TS and [Zn₂(DFC–TS)]⁺ with DFC–TS : Zn = 1 : 2.15 in 10 mM HEPES buffer in 9 : 1 DMSO : H₂O (v/v) solvent system, λ_ex = 430 nm.

The intracellular Zn²⁺ imaging behaviour of DFC–TS was studied on HepG2 cells with the aid of fluorescence microscopy which displayed intensive fluorescence when cells pre-incubated with DFC–TS were exposed to Zn²⁺ ion (10.0 μM), (Fig. 8) which, however, strongly suppressed when Na₂H₂EDTA (100 μM) was added (Fig. 8). Therefore, DFC–TS renders a confirmatory evidence of the intracellular monitoring of Zn²⁺ without interference by biologically abundant other metal ions. We have also determined quantum yields of the ligand, [Zn₂(DFC–TS)] complex, which showed higher quantum yield for Zn²⁺complex (0.143) than the pure ligand (0.0074).

Fig. 8 The phase contrast and fluorescence images of HepG2 cells captured after incubation for 30 min at 37 °C with (a) DFC–TS, (b) DFC–TS + Zn² and (c) DFC–TS + Zn²⁺ followed by addition of 100 μM EDTA.

Geometry optimization and electronic structure

Geometries of DFC–TS and mono [Zn(DFC–TS)(OAc)] (1) and dinuclear [Zn₂(DFC–TS)(OAc)₂] (2) complexes were fully optimized using B3LYP functional as implemented in Gaussian 03.²² The nature of all the stationary points was confirmed by carrying out a normal mode analysis, where all vibrational frequencies were found to be positive. Some selected optimized geometrical parameters of DFC–TS for both keto and enol forms and complexes 1 and 2 are listed in Table S1 and S2 (ESI†) and geometry optimized structures are given in Fig. 9. The HOMO–LUMO gap for keto (2.87 eV) and enol (3.39 eV) decidedly indicates that the later is more stable than the former. The modelled geometry of complex 1 showed a penta-coordination around the metal centre in a distorted trigonal-bipyrimidal fashion. Here, phenoxo-O, azomethyne-N, keto-S and two oxygen atoms from acetate group participate in binding to Zn²⁺ giving a neutral species while other thiosemicabazone group rather remains dangling without participating in bonding to the metal center.

Fig. 9 DFT-optimized molecular structures DFC–TS and complexes 1 and 2.

We have also optimized the geometry of complex 1 as an anionic species where sulphide (S⁻) from C–SH (enolic form) coordinates to the metal center, other coordinating atoms remain unchanged. It is to be mentioned here that neutral species is more stable than the anionic species as indicated by the HOMO → LUMO energy gap (Fig. 10). The complex 2 is a dinuclear neutral species where one DFC–TS with N₂S₂O donor set of atoms and 4 oxygen atoms from two acetate groups involved in bonding. The phenoxo-O atom acts as a bridge between two metal centers along with two bridging acetate groups to hold the Zn²⁺ atoms firmly resulting distorted TBP geometries around them. The idea of such geometry as a dinuclear complex comes from our previous structural studies with analogous ligand. The bond distances and bond angles (Table S1 and S2 in ESI†) around the metal center(s) are very similar to those found in analogous complexes. In 1 and 2 all the calculated Zn–N/Zn–O distances fall in the range 2.054–2.319 Å and comparable to the reported values in the analogous complexes.^36,38

Fig. 10 HOMO–LUMO energy gaps for DFC–TS, [Zn(DFC–TS)(OAc)] and [Zn₂(DFC–TS)(OAc)₂].

Time dependent density functional theory (TDDFT)^23–25 with B3LYP density functional was applied to study the low-lying excited states of the complex in DMSO using the optimized geometry of the ground (S₀) state. The vertical excitation energies of the lowest 20 singlet states are also computed here. The UV-Vis spectra computed from TDDFT calculations in DMSO show two important peaks in the range 250–600 nm (see Fig. S6 and S7†) for DFC–TS. For complex 1, the band around 370 nm is dominated by the HOMO → LUMO excitations, while the band around ∼320 nm is mainly due to HOMO−1 → LUMO, HOMO−1 → LUMO+1 and HOMO → LUMO+1 transition. The corresponding experimental absorption bands appear at 370 nm and 317 nm respectively. In case of [Zn(DFC–TS)(OAc)] the absorption band around ∼420 nm is dominated by the HOMO → LUMO excitations, while the band around ∼340 nm it is mainly due to HOMO−1 → LUMO, and HOMO → LUMO+1 transitions. The absorption band at ∼300 nm arises mainly due to HOMO−3 → LUMO transition. The corresponding experimental absorption bands appear at 430 nm and 300 nm. Similarly, in case of [Zn₂(DFC–TS)(OAc)₂] the absorption bands at∼ 420 nm is dominated by the HOMO → LUMO excitations, while the band around ∼340 nm is mainly due toHOMO−1 → LUMO, and HOMO → LUMO+1 transitions. The TDDFT calculation on the mono- and dinuclear complexes reveal that there is no basic differences in the pattern of absorptions spectra of these species which is also evidenced from the experimental spectra where the λ_max (absorption) remains almost unaltered for these two species but absorption continues to increase on adding Zn²⁺ beyond 1 : 1 mole ration, until it gets saturated at 2 : 1 mole ratio with respect to Zn²⁺. The details of the vertical excitation energies, oscillator strengths, and nature of excitations are shown in Table S3 (ESI†). HOMO–LUMO diagrams of all the optimized species along with the energy gaps are depicted in Fig. 10.

A list of 2,6-diformyl-p-cresol based Schiff base ligands^36,39–43 that showed Zn²⁺ sensing is given in Scheme 3.

Scheme 3

A quick inspection of these studies reveal that all these are turn on Zn²⁺ sensor with moderate LOD and find application towards intracellular Zn²⁺ monitoring.^36,39–43

Conclusion

In summary, we have been successful to design and synthesize a DFC–TS based penta-coordinating N₂S₂O donor ligand that selectively recognizes Zn²⁺ colorimetrically as well as fluorimetrically with 17 fold fluorescence enhancement. The fluorescence enhancement occurs through ESIPT blocked chelation enhanced fluorescence effect (CHEF). Though, previously it was found that dithiosemicabazone based probes have very poor aqueous solubility and form complexes with Zn²⁺ very slowly⁴⁴ precluding the direct determination of formation constants, the present Zn²⁺ sensor provides opportunity for the first time, to the direct determination of step-wise formation constants of similar order of magnitude by UV-Vis titration. This is possibly due to the reorganization of the ligand environment of the mononuclear complex to accommodate the second Zn²⁺ ion to form the dinuclear species. This probe has cell penetrating ability thereby providing opportunity for in vitro/in vivo cell imaging of Zn²⁺ ion without much worry about cytotoxicity.

Acknowledgements

Financial support from DST (Ref. SR/S1/IC-35/2006) and CSIR (ref. 02(2490)/11/EMR-II) New Delhi, are gratefully acknowledged.

Notes and references

1. B. L. Vallee and K. H. Falchuk, Physiol. Rev., 1993, 73, 79–118.
2. C. J. Frederickson, J. Y. Koh and A. I. Bush, Nat. Rev. Neurosci., 2005, 6, 449–462.
3. (a) S. C. Burdette and S. J. Lippard, Coord. Chem. Rev., 2001, 333, 216–217; (b) S. Sreejith, K. P. Divya and A. Ajayaghosh, Chem. Commun., 2008, 2903–2905; (c) K. P. Divya, S. Sreejith, B. Balakrishna, P. Jayamurthy, P. Aneesa and A. Ajayaghosh, Chem. Commun., 2010, 46, 6069–6071; (d) S. Sreejith, K. P. Divya, P. Jayamurthy, J. Mathew, V. N. Anupama, D. S. Philips, P. Aneesa and A. Ajayaghosh, Photochem. Photobiol. Sci., 2012, 11, 1715–1723; (e) K. P. Divya, S. Sreejith, P. Ashokkumar, K. Yuzhan, Q. Peng, S. K. Maji, Y. Tong, H. Yu, Y. Zhao, P. Ramamurthy and A. Ajayaghosh, Chem. Sci., 2014, 5, 3469–3474; (f) A. Ajayaghosh, P. Carol and S. Sreejith, J. Am. Chem. Soc., 2005, 127, 14962–14963; (g) E. L. Que, R. Bleher, F. E. Duncan, B. Y. Kong, S. C. Gleber, S. Vogt, S. Chen, S. A. Garwin, A. R. Bayer, V. P. Dravid, T. K. Woodruff and T. V. O'Halloran, Nat. Chem., 2015, 7, 130–139.
4. E. M. Nolan, J. Jaworski, K. I. Okamoto, Y. Hayashi, M. Sheng and S. J. Lippard, J. Am. Chem. Soc., 2005, 127, 16812–16823.
5. D. H. Nies, Science, 2007, 317, 1695–1696.
6. (a) K. Kikuchi, H. Komatsu and T. Nagano, Curr. Opin. Chem. Biol., 2004, 8, 182–191; (b) S. Goswami, A. K. Das, K. Aich, A. Manna, S. Maity, K. Khanrab and N. Bhattacharyya, Analyst, 2013, 138, 4593–4598; (c) S. Goswami, S. Paul and A. Manna, RSC Adv., 2013, 3, 25079–25085; (d) S. Goswami, A. Manna, S. Paul, A. K. Maity, P. Saha, C. K. Quah and H. K. Fun, RSC Adv., 2014, 4, 34572–34576.
7. C. J. Chang and S. J. Lippard, Life Sci., 2006, 1, 321–370.
8. E. M. Bavin, R. J. W. Rees, J. M. Robson, M. Seiler, D. E. Seymour and D. J. Suddaby, J. Pharm. Pharmacol., 1950, 2, 764–772.
9. G. A. Kune, Br. Med. J., 1964, 2, 621–622.
10. A. C. Sartorelli and B. A. Booth, Cancer Res., 1967, 27, 1614–1619.
11. C. M. Nutting, C. M. L. vanHerpen, A. B. Miah, S. A. Bhide, J. P. Machiels, J. Buter, C. Kelly, D. de Raucourt and K. J. Harrington, Ann. Oncol., 2009, 20, 1275–1279.
12. J. Waluk, Acc. Chem. Res., 2003, 36, 832–838.
13. C. A. Taylor, M. A. El-Bayoumi and M. Kasha, Proc. Natl. Acad. Sci. U. S. A., 1969, 63, 253–260.
14. S. Takeuchi and T. Tahara, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 5285–5290.
15. G. Villani, J. Phys. Chem. B, 2010, 114, 9653–9662.
16. S. R. Meech, Chem. Soc. Rev., 2009, 38, 2922–2934.
17. (a) J. S. Wu, W. M. Liu, J. C. Ge, H. Y. Zhang and P. F. Wang, Chem. Soc. Rev., 2011, 40, 3483–3495; (b) S. Goswami, A. Manna, S. Paul, A. K. Das, P. K. Nandi, A. K. Maity and P. Saha, Tetrahedron Lett., 2014, 55, 490–494; (c) S. Goswami, A. K. Das, A. Manna, A. K. Maity, H. K. Fun, C. K. Quah and P. Saha, Tetrahedron Lett., 2014, 55, 2633–2638; (d) S. Goswami, A. Manna, S. Paul, A. K. Das, K. Aich and P. K. Nandi, Chem. Commun., 2013, 49, 2912–2914; (e) S. Goswami, A. Manna, M. Mondal and D. Sarkar, RSC Adv., 2014, 4, 62639–62643; (f) A. Manna and S. Goswami, New J. Chem., 2015, 39, 4424–4429.
18. S. Park, J. E. Kwon, S. H. Kim, J. Seo, K. Chung, S. Y. Park, D. J. Jang, B. M. Medina, J. Gierschner and S. Y. Park, J. Am. Chem. Soc., 2009, 131, 14043–14049.
19. S. Park, S. Kim, J. Seo and S. Y. Park, Macromol. Res., 2008, 16, 385–395.
20. R. R. Gagne, C. L. Spiro, T. J. Smith, C. A. Hamann, W. R. Thies and A. K. Schiemke, J. Am. Chem. Soc., 1981, 103, 4073–4081.
21. C. R. Lohani, J. M. Kim, S. Y. Chung, J. Yoon and K. H. Lee, Analyst, 2010, 135, 2079–2084.
22. M. J. Frisch, et al., Gaussian 03, Revision C.02, ed. C. Gonzalez and J. A. Pople, Gaussian Inc., Wallingford, CT, 2004.
23. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
24. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299–310.
25. V. Barone, M. Cossi and J. Tomasi, J. Comput. Chem., 1998, 19, 404–417.
26. M. Cossi, N. Rega, G. Scalmani and V. Barone, J. Comput. Chem., 2003, 24, 669–681.
27. T. Liu, H. X. Zhang and B. H. Xia, J. Phys. Chem. A, 2007, 111, 8724–8730.
28. T. Mossman, J. Immunol. Methods, 1983, 65, 55–63.
29. S. Scheiner, J. Phys. Chem. A, 2000, 104, 5898–5909.
30. A. Weller, Elektrochemie, 1956, 60, 1144–1147.
31. Fundamentals of Analytical Chemistry, ed. D. A. Skoog, D. M. West and F. J. Holler, Harcourt College Publishers, 7th edn, 1996.
32. C. F. Bernasconi, M. Ali and J. C. Gunter, J. Am. Chem. Soc., 2003, 125, 151–157.
33. C. F. Bernasconi, M. Ali and F. Lu, J. Am. Chem. Soc., 2000, 122, 1352–1359.
34. E. Kimura, I. Nakamura, T. Loike, M. Shionoya, Y. Kodama, T. Ikeda and M. Shiro, J. Am. Chem. Soc., 1994, 116, 4764–4771.
35. F. G. Bordwell, R. J. McCallum and W. N. Olmstead, J. Org. Chem., 1984, 49, 1424–1427.
36. T. Mistri, M. Dolai, D. Chakraborty, A. R. Khuda-Bukhsh, K. K. Das and M. Ali, Org. Biomol. Chem., 2012, 10, 2380–2384.
37. J. Y. Choi, D. Kimb and J. Yoon, Dyes Pigm., 2013, 96, 176–179.
38. X. Zhou, B. Yu, Y. Guo, X. Tang, H. Zhang and W. Liu, Inorg. Chem., 2010, 49, 4002–4007.
39. S. Anbu, R. Ravishankaran, M. F. C. G. da Silva, A. A. Karande and A. J. L. Pombeiro, Inorg. Chem., 2014, 53, 6655–6664.
40. P. Roy, K. Dhara, M. Manassero, J. Ratha and P. Banerjee, Inorg. Chem., 2007, 46, 6405–6412.
41. R. Alam, T. Mistri, A. Katarkar, K. Chaudhuri, S. K. Mandal, A. R. Khuda-Bukhsh, K. K. Das and M. Ali, Analyst, 2014, 139, 4022–4030.
42. B. K. Datta, D. Thiyagarajan, S. Samanta, A. Ramesh and G. Das, Org. Biomol. Chem., 2014, 12, 4975–4982.
43. A. Jana, P. K. Sukul, S. K. Mandal, S. Konar, S. Ray, K. Das, J. A. Golen, A. L. Rheingold, S. Mondal, T. K. Mondal, A. R. Khuda-Bukhsh and S. K. Kar, Analyst, 2014, 139, 495–504.
44. G. Buncic, P. S. Donnelly, B. M. Paterson, J. M. White, M. Zimmermann, Z. Xiao and A. G. Wedd, Inorg. Chem., 2010, 49, 3071–3073.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra25653d

---