a Versatile Schiff Base Chemosensor for the Determination of Trace Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ in the Water and Its Bioimaging Applications

Abstract

In this work, a simple and versatile Schiff base chemosensor (L) was developed for the detection of four adjacent row 4 metal ions (Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺) through colorimetric or fluorescent analyses. L could recognize the target ions in solutions containing a wide range of other cations and anions. The recognition mechanisms were verified with a Job’s plot, HR-MS assays, and ¹H NMR titration experiments. Then, L was employed to develop colorimetric test strips and TLC plates for Co²⁺. Meanwhile, L was capable of quantitatively measuring the amount of target ions in tap water and river water samples. Notably, L was used for imaging Zn²⁺ in HepG2 cells, zebrafish, and tumor-bearing mice, which demonstrated its potential biological applications. Therefore, L can probably serve as a versatile tool for the detection of the target metal ions in environmental and biological applications.

Introduction

As is known to all, doubly charged metal ions such as Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ play important roles in the environment and biologic processes, especially in maintaining human nutrition and health.¹ For instance, an excess of Co²⁺ could lead to vasodilatation and cardiomyopathy,^1c,2 Similarly, an overload of Ni²⁺ would result in asthma, angina, or other cardiac symptoms.³ Daily intake of Cu²⁺ and Zn²⁺ is required to remain healthy.⁴ However, unregulated Cu²⁺ and Zn²⁺ may cause illness. For example, copper deficiency is associated with myelopathy.^4c,4d,5 Conversely, the excess intake of Cu²⁺ can adversely affect human health.^1e,2b,5 Non-regulatory Zn²⁺ levels could lead to Alzheimer’s disease, epilepsy, etc.⁶ Hence, the construction of a feasible method for the detection of the above four adjacent row 4 metal ions (Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺) is urgently needed . To date, principal component feed-forward neural networks (PCFFNNs) and feed-forward neural networks (FFNNs) are the only reported methods for the detection of multiple metal ions.⁷ However, these methods are based on a series of complexation reactions⁷ and therefore suffer from several limitations, such as lengthy sample preparations and complicated calculation procedures.

Among many analytical methods, colorimetric and fluorescent analyses with small molecular chemosensors that can recognize various metal ions simultaneously are of vital importance because of their readily available and rapid operations (Table S1).⁷ Compared with the detection of an individual target ion, the exploration of multiple toxic metal ions with a single chemosensor is time- and cost-effective,⁷ which maximizes the convenience and economy of an analytical technique.⁸ Nevertheless, the development of chemosensors with the ability to detect multiple metal ions is a challenge,^9,10 especially when those ions are adjacent in the periodic table.^9,10

Chemosensors containing julolidine are usually water-soluble.¹¹ 2-Hydrazinylpyridine is capable of recognizing metal ions because it can form coordinate bonds with metal ions via its hydrazine group. Therefore, in this work we describe a simple chemosensor (L) consisting of the julolidine and 2-hydrazinylpyridine moieties (Scheme 1). The recognition properties of L for Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ ions were evaluated by colorimetric or fluorescent analyses. Moreover, L is able to image Zn²⁺ in live cells, zebrafish, and tumor-bearing mice.

Scheme 1. Synthetic Procedure for L.

Results and Discussion

The synthetic procedure to target chemosensor L is outlined in Scheme 1. In brief, 8-hydroxyjulolidine-9-carboxaldehyde and 2-hydrazinylpyridine were efficiently condensed in methanol under microwave irradiation for 15 min. L was perfectly characterized by NMR (¹H and ¹³C), FT-IR, and mass spectrometry (MS) (Figures S1–S4).

The sensing behavior of chemosensor L was first investigated by absorption and fluorescence tests with various cations (Li⁺, Ag⁺, Cd²⁺, K⁺, Ca²⁺, Na⁺, Mg²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺, Pb²⁺, Al³⁺, Ba²⁺, Fe³⁺, Sr²⁺, and La³⁺) and different anions (CN^–, I^–, SCN^–, NO₃^–, HSO₄^–, SO₄^2–, Cl^–, HCO₃^–, CO₃^2–, ClO₃^–, PO₄^3–, HPO₄^2–, H₂PO₄^–, Cr₂O₇^–, S^2–, MoO₄^–, Br^–, and AcO^–) in a 50% ethanol/tris-HCl buffer solution (pH = 7.40) at 25 °C.

After adding different concentrations of cations and anions (detailed above) to the L solution, only Cu²⁺ caused a prominent blue-shift in the absorption spectrum. Instead, Co²⁺, Ni²⁺, and Zn²⁺ generated a remarkable red-shift in the absorption spectrum (Figures 1 and S5). Moreover, anions failed to induce any spectral response because of the absence of coordination between L and the anions (Figure S6). Besides, among the above four ions that showed obvious spectroscopic responses, only Co²⁺ caused the color of the solution to transform from colorless to wine-colored (Figures 1, S7 ,and S8). This is an increase from the longest spectral red-shift of L-Co²⁺ (among these four metal ions), which had a maximum absorption at 532 nm. The unique color transformation of L in response to Co²⁺ suggests that L could serve as a highly selective colorimetric chemosensor for Co²⁺ detection by the naked eye in aqueous solutions. Taken together, the above results demonstrate that L could be utilized as a versatile chemosensor for the synchronous measurement of target ions in aqueous solutions.

Figure 1.

Absorption spectra variation of chemosensor L (20 μM) before and after the addition of Co²⁺ (40 μM), Ni²⁺ (40 μM), Cu²⁺ (40 μM), and Zn²⁺ (40 μM) in a 50% ethanol/tris-HCl buffer solution (pH = 7.40). The inset shows the color change of L in response to Co²⁺.

The above solutions were also measured by fluorometry with an excitation wavelength of 413 nm. Only Zn²⁺ produced a distinct fluorescence intensity increase at 498 nm (ca. 10×) (Figure 2). By contrast, the fluorescence response for most of the other cations was negligible. The only change was Al³⁺, which gave an increase 3× the original intensity (Figure 2). Considering the abnormal accumulation of Cu²⁺ under certain illnesses conditions, L was also treated with high concentration of Cu²⁺ (200 μM). As shown in Figure S9, Cu²⁺ did not lead to a detectable fluorescence increase even at a high concentration. Moreover, in coexistence of Cu²⁺, the fluorescence induced by Zn²⁺ decreased slightly and remained constant over at least 1 h (Figure S10). Such a fluorescence decrease could be eliminated by the widely used chelation reagent resin Chelex-100 for Cu²⁺ (vide infra), implying that the chelated Cu²⁺ would not interrupt the fluorescence detection of Zn²⁺. In view of the fact that the intracellular Cu²⁺ is mainly the chelated form,¹² the above results further indicated that L is capable of serving as a usable sensor for the fluorescence detection of an abnormal increase in the free Zn²⁺ level in living systems. It is worth mentioning that under UV lamp irradiation at 365 nm, Zn²⁺ could selectively induce the strong blue-green fluorescence of L, which could thus be applied for the naked-eye recognition of Zn²⁺ (Figures 2 and S11).

Figure 2.

Fluorescence spectra variation of chemosensors L (20 μM) before and after the addition of the above-mentioned cations (40 μM), where λ_ex = 413 nm. The inset shows the visible fluorescence turn-on of L by Zn²⁺.

L can “turn on” fluorescence in response to Zn²⁺, which can be described as a result of the restricted Schiff base group (C=N) isomerization and the excited-state intramolecular proton transfer (ESIPT) mechanism.¹³ First, L exists as a free ligand in which the rotation of the C=N bond is not restricted. The rotation and isomerization of such an unrestricted C=N bond would act as a nonradiative inactivation process to quench the fluorescence of the fluorogen. Therefore, free L has no visible fluorescence. However, when Zn²⁺ is added to L, a chelate complex (with an enhanced rigid structure) may form between L with Zn²⁺ ion, resulting in the chelation-intensified fluorescence effect.¹³ Meanwhile, the intramolecular hydrogen bond (IMHB) between the phenolic O atom of julolidine shows ESIPT,^13a,14 suggesting that the phenol O atom of julolidine could be chelated with Zn²⁺.^13,14 Therefore, the chelation of Zn²⁺ by L may inhibit C=N isomerization and ESIPT, leading to an enhancement of the fluorescence (Scheme 2).

Scheme 2. Fluorescence Enhancement Mechanism of L by Zn²⁺ and the Proposed Structure of the L–Zn²⁺ Complex.

For further studies of the recognition responses between L and these four cations, absorption and fluorescence titration experiments were conducted. When L was added to increasing amounts of Co²⁺, Ni²⁺, Cu²⁺, or Zn²⁺, the initial absorption peak of L at 370 nm gradually decreased and new peak presented. For Co²⁺, a new absorption peak emerged at 532 nm following the increase of the Co²⁺ concentration from 0 to 20 μM (Figure 3). For Ni²⁺, a new peak presented at 438 nm, and the absorption intensity gradually increased with the increase of the Ni²⁺ concentration from 0 to 60 μM, reaching its maximum at 60 μM of Ni²⁺ (Figure S12). For Cu²⁺, three new absorption peak appeared at 306, 416, and 520 nm as the concentration of Cu²⁺ increased from 0 to 400 μM (Figure S13). For Zn²⁺, an absorption peak emerged at 413 nm as the Zn²⁺ concentration increased from 0 to 100 μM, reaching its maximum at 100 μM of Zn²⁺ (Figure S14). The blue-shift of L-Cu²⁺ might be explained by the push–pull effect of the internal charge transfer (ICT) effect.¹⁵ When Cu²⁺ was added to L, the chelate complex of L–Cu²⁺ was formed through the phenol O, the Schiff base group (C=N) N, and secondary amine (−N−) N atoms complexing with Cu²⁺, which may have resulted in a feeble push–pull electronic effect. Therefore, the variation of the ICT transition mechanism could be responsible for blue-shift in the absorption spectrum. Meanwhile, the red-shift of the absorption peak shown by L–Co²⁺, L–Ni²⁺, and L–Zn²⁺ could be ascribed to ligand-to-metal charge transfer (LMCT) and the ICT effect in the complex formation.¹⁴ The energy gap of the ICT band should decrease upon M²⁺ (M = Co, Ni, and Zn) binding to the electron-donating groups, such as the phenol O atom, the Schiff base group (C=N) N atom, and the secondary amine (−N−) N atom, resulting in the red-shift of L–M²⁺.¹⁶ Moreover, the prominent isoabsorptive points were discovered at 398 nm for Co²⁺, 397 nm for Ni²⁺, 328 nm for Cu²⁺, and 392 nm for Zn²⁺, which indicated that just one complex was produced from L upon binding to Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺, respectively. On the other hand, when fluorescence titration was conducted with Zn²⁺, the fluorescence intensity at 498 nm exhibited its maximum at 40 μM Zn²⁺ (Figure 4). The Job’s plots obtained from the absorption and fluorescence spectrum were applied to confirm the stoichiometry of the formed complexes. Figures 5, S15, and S16 demonstrate the 1:1 stoichiometry of L with Cu²⁺, Zn²⁺, Co²⁺, and Ni²⁺.

Figure 3.

Absorption spectra of chemosensor L (20 μM) in the presence of 0–20 μM Co²⁺ in a 50% ethanol/tris-HCl buffer solution (pH = 7.40).

Figure 4.

Fluorescence spectra of chemosensor L (20 μM) in the presence of 0–40 μM Zn²⁺ in a 50% ethanol/tris-HCl buffer solution (pH = 7.40).

Figure 5.

Job’s plots for metal ions with L. (a) Cu²⁺ (λ_max = 306 nm) and (b) Zn²⁺ (λ_em = 498 nm), presenting a 1:1 stoichiometry in a 50% ethanol/tris-HCl buffer solution (pH = 7.40).

The stoichiometry results were further verified by high-resolution mass spectrometry (HR-MS). The HR-MS spectrum of the mixture of L with Co²⁺ (Figure 6) shows a peak at m/z 364.2441, which is assigned to [Co(L)]²⁺. Meanwhile, the HR-MS results of the mixtures of L with Ni²⁺, Zn²⁺, and Cu²⁺ did not show any homologous peaks for [Ni(L)]²⁺, [Zn(L)]²⁺, or [Cu(L)]²⁺. However, the peaks observed in Figures S17–19 at m/z 417.2923, 436.1867, and 438.2 could be attributed to [Ni(L)(Cl)(H₂O)]²⁺, [Zn(L)(CH₃CH₂OH)(H₂O)]²⁺, and [Cu(L) (CH₃CH₂OH)+Na⁺] (calcd. for 417.0628, 436.1265, and 438.1004), respectively. These study confirmed that L formed 1:1 complexes with Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺.

Figure 6.

HR-MS spectrum of L-Co²⁺ (1.0 eqv) in a 50% ethanol/tris-HCl buffer solution (pH = 7.40).

To further survey the bonding of chemosensor L to Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺, ¹H NMR titration tests were executed in a 10% D₂O/DMSO-d₆ mixed solvent (Figures 7 and S20–S22). As the concentrations of these metal ions increased, the phenolic −OH proton (Ha, δ 11.19 ppm) and −NH (Hb, δ 10.56 ppm) gradually weakened. The results indicated the complexation of the four metal ions with the phenolic O and N two atoms.^11a,17 Meanwhile, when 1.0 equiv of the metal ion was added into the L solution, the phenolic −OH proton and −NH proton almost completely disappeared, indicating 1:1 complexation for [M(L)]²⁺ (M = Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺). In addition, the imine proton CH=N (Hc, δ 8.09 ppm) and benzene protons (Hd–Hi, δ 6.63–7.99 ppm) were disturbed and formed broader peaks due to the asymmetry of the [M(L)]²⁺ complexes.^11a,17 Besides, the paramagnetic nature of Cu²⁺, Co²⁺, and Ni²⁺ is also a main reason for the quenching of the NMR signals. Therefore, ¹H NMR titration experiments not only verified the binding sites of these four metal ions with L but also further confirmed the 1:1 complexation of [M(L)]²⁺.

Figure 7.

¹H NMR tests of L with different concentrations of Zn²⁺ in a mixture of D₂O/DMSO-d₆ (v/v, 1:9).

To study the detection ability of L after being recycled for several measurements, the solution of L was alternately treated with metal ions and the strong chelating agent EDTA (Figures 8, S23, and S24). The initial absorbances of L-Co²⁺, L-Ni²⁺, L-Cu²⁺, and L-Zn²⁺ were recorded at 532, 438, 306, and 413 nm, respectively. After 1.0 equiv of EDTA was added to the L–M²⁺ solution, the absorbance of the system prominently decreased. Then, as more metal ions were added to the resulting solution of [L–(Co²⁺/EDTA)], [L–(Ni²⁺/EDTA)], [L–(Cu²⁺/EDTA)], or [L–(Zn²⁺/EDTA)], a recovery of the absorbance or fluorescence was observed. Furthermore, repeated experiments also yielded the same results for at least five cycles. Therefore, these competitive experiments with EDTA further verified the complexation durability of L with these four metal ions.

Figure 8.

Durability researches of L by the alternate treatment of (a) Ni²⁺ or (b) Zn²⁺ and EDTA in 50% ethanol/tris-HCl buffer solution (pH = 7.40).

Based on the above-mentioned experiments, the 1:1 complexation structures of L with four target ions are proposed in Scheme 3.

Scheme 3. Proposed Coordination Mechanism of L–M²⁺ (M²⁺ = Co²⁺, Ni²⁺, Cu²⁺, or Zn²⁺).

The linear relationships between absorption and the concentration of the four metal ions are shown in Figures S25–S28. Meanwhile, the association constants (K_a) of L with the four target ions were calculated using the Benesi–Hildebrand equation (Figures S29–S32).^11a,18 The K_a values with L were calculated as 1.02 × 10⁵, 1.35 × 10⁵, 1.16 × 10⁵, and 2.10 × 10⁷ M^–1 for Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺, respectively. Furthermore, using 3σ/K (where σ is the standard deviation of 11 blank absorbances at the detection wavelength shown in Figures S25–S28 and K is the slope of the relative linear equation), the detection limits of L were calculated to be 3.20 × 10^–8, 6.00 × 10^–8, 6.23 × 10^–7, and 1.09 × 10^–7 M for Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺, respectively.

Next, the competitive tests were also performed by adding 2 equiv of 16 cations and 17 anions (as mentioned above) to the L–M²⁺ (M = Co²⁺, Ni²⁺, Cu²⁺, or Zn²⁺) system. The results showed that some metal ions would interfere with the absorption or fluorescence detection of the four metal ions (Figures 9 and S33–S38). For example, in the L–Co²⁺ and L–Ni²⁺ systems, the coexistence of other metal ions or anions did not influence the response to Co²⁺ and Ni²⁺ by chemosensor L expect for Cu²⁺ (Figures 9a and S33–35). However, the application of resin Chelex-100, a widely used chelation reagent for Cu²⁺, can eliminate the interference of Cu²⁺ for Co²⁺ and Ni²⁺, which is probably due to the stronger affinity between resin Chelex-100 and Cu²⁺ compared to Co²⁺ and Ni²⁺ (Figures S39 and S40, respectively).¹⁹ Therefore, L can serve as a selective chemosensor for the detection of Co²⁺ and Ni²⁺ even with the coexistence of 16 metal ions or 17 anions. For Cu²⁺, the absorption peak of L at 306 nm was not obviously interfered with by the other metal ions or anions except Fe³⁺ and Pb²⁺ (Figures S36 and S37, respectively). However, the interferences from Fe³⁺ and Pb²⁺ can be successfully inhibited using F^– and EDTA, respectively (Figures S41 and S42, respectively). It should be noted that EDTA has comparable affinity to Pb²⁺ and Cu²⁺²⁰ but may prefer to form a 1:1 complex with Pb²⁺ to remove its interference in the current system of L + Cu²⁺ + Pb²⁺ + EDTA, which is supported by our results. In the case of Zn²⁺, the fluorescence intensity of L–Zn²⁺ suffered from 60% or 90% quenching with the addition of Cu²⁺ or Co²⁺, respectively (Figure 9b). Fortunately, applications of triethanolamine and the chelating resin Chelex-100 were able to overcome the interferences from Co²⁺ and Cu²⁺ ions, respectively (Figure S43 and S44, respectively).

Figure 9.

Anti-interference experiments of L for metal ions. (a) Absorbance change of L–Co²⁺ before and after the addition of other different metal ions (2.0 equiv), where [L] = [Co²⁺] = 20 μM and λ_max = 532 nm. (b) Fluorescence intensity of L–Zn²⁺ before and after the addition of other different metal ions (2.0 equiv), where [L] = [Zn²⁺] = 20 μM and λ_em = 498 nm.

In addition, we also tested the response time of chemosensor L to Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ (Figures S45–S47). The results show that L can rapidly recognize the four target ions and obtain a stable respond readout in a long time period.

To develop the practical application of L, test strips or TLC plates were coated with L for the naked-eye detection of Co²⁺ in water. Among the examined metal ions, a prominent color transformation from gray to wine red can be produced by only in the case of Co²⁺ (Figure 10a), implying the selective naked-eye detection of Co²⁺. In additional, the concentration of Co²⁺ can be quantitatively measured by naked eye using the test strips, and the detection limit was estimated to be 1.0 × 10^–5 M. Furthermore, the L-coated TLC plates were also sprayed with different concentrations of Co²⁺. The results show that the L-coated TLC plates present an obvious color change from gray to wine, with an estimated detection limit of 1.0 × 10^–4 M. Therefore, both L-coated test strips and L-coated TLC plates could be used for the rapid and easy detection of hazardous Co²⁺ by the naked eye in a real water sample (Figure 10b and c), which may position them as portable tools to be carried and used.

Figure 10.

Applications of the 20 μM L-coated test strips and TLC plates for the detection of Co²⁺. (a) The L-coated test strips sprayed with 40 μM solutions of different metal ions. (b) The L-coated test strips and (c) TLC plates sprayed with different concentrations of Co²⁺ (from left to right: 0.0, 1.0 × 10^–6, 1.0 × 10^–5, 1.0 × 10^–4, 1.0 × 10^–3, 1.0 × 10^–2, and 1.0 × 10^–1 M).

Next, L was utilized for the quantitative detection of Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ in a variety of water samples such as tap water and river water (Table S2–S5, respectively). At the same time, the detection of Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ in environmental water was also performed by ICP-MS assays, which obtained results comparable with those of L (Table S6–S9). These results further demonstrate that chemosensor L is highly usable in multiple fields such as environmental chemistry and analytical chemistry for the detection of the hazardous Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺.

To apply the chemosensor L in living cells for the fluorescence imaging detection of Zn²⁺, we first tested the cytotoxicity of L against HepG2 cells using the MTT assay.²¹ The results showed that L had a low cell inhibitory rate (Figure S48) and could be used as a biocompatible probe for the cellular imaging of Zn²⁺. As shown in Figure 11, inappreciable fluorescence was found when HepG2 cells were incubated with L for 30 min. However, after HepG2 cells were incubated with exogenous Zn²⁺ for other 30 min, strong fluorescence was observed under a fluorescence microscope (IX73). These experimental results indicate that chemosensor L is cell-penetrable and can serve as a valid candidate for the detection of Zn²⁺ in living cells by fluorescence imaging.

Figure 11.

Fluorescence imaging of Zn²⁺ in HepG2 cells with L. (a and b) Bright field images. (c and d) Fluorescence images of HepG2 cells treated with (c) L or (d) L + Zn²⁺. (e and f) Merged images of (e) panels a and c (f) panels d and b. [L] = [Zn²⁺] = 5 μM, λ_ex = 425–485 nm, and λ_em = 515 nm.

To explore the possibility of detecting Zn²⁺ with L in living organisms, we performed fluorescence imaging in zebrafish or tumor-bearing mice. As shown in Figure 12, after being cultured with L for 30 min, the zebrafish was observed to have negligible fluorescence. However, the addition of Zn²⁺ to the L-loaded zebrafish led to a strong fluorescence enhancement under the fluorescence microscope. These results indicate that L could potentially be applied as a chemosensor for monitoring Zn²⁺ accumulation in living organisms.

Figure 12.

Fluorescence imaging of Zn²⁺ in zebrafish with L. (a and b) Bright field images. (c and d) Fluorescence images of zebrafish treated with (c) L or (d) L + Zn²⁺. (e and f) Merged images of (e) panels a and c and (f) panels b and d. [L] = [Zn²⁺] = 5 μM, λ_ex = 425–485 nm, and λ_em = 515 nm.

To further expand the new application field of L for imaging Zn²⁺ in live animals, we also established a two-tumor model in C57 mice (one on each side) with B16F10 cells for the fluorescence imaging of Zn²⁺ in the tumors. As shown in Figure 13, it can be seen that the fluorescence signal was not detectable whether the mice were intratumorally injected with PBS (Figure 13a, right) or without PBS (Figure 13a, left), followed by intratumoral injection of Zn²⁺. Similarly, when only L was injected into the tumor, no fluorescence signal was observed (Figure 13b, left). However, the injection of Zn²⁺ resulted in a significant fluorescence increase within 10 min (Figure 13b, right). Overall, these studies indicate that L can be applied in live animals to image Zn²⁺.

Figure 13.

Fluorescence imaging of Zn²⁺ in B16F10 tumor-bearing C57 mice. (a) Fluorescence images of the mice intratumorally injected with only PBS for 10 min (left) and without PBS for 10 min (right), followed by an intratumoral injection of Zn²⁺ on both sides. (b) Fluorescence images of the mice intratumorally injected with L on both sides for 10 min, followed by an intratumoral injection of Zn²⁺ on the right side.

Conclusions

In summary, in this work we reported a simple and versatile chemosensor L, which can be used to recognize Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ ions. As for as we know, L is one of only a few cases of a simple chemosensor capable of sensing four metal ions (Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺) that are adjacent on the fourth row of the periodic table. L could be used for the quantitative detection of trace amounts of Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ in tap water and river water. Particularly, the selective colorimetry determination capacity of L for Co²⁺ enables it to be used for the naked-eye monitoring of Co²⁺ levels. More importantly, L can also be utilized to image Zn²⁺ in live cells, zebrafish, and tumor-bearing mice, demonstrating its potential application in biological systems. Therefore, L is expected to be employed as a simple and versatile chemosensor for the selective detection of environmental and biological traces of Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ ions.

Experimental Section

All chemicals reagents were of analytical grade and applied without treatment. Double-distilled water was applied for all tests. The NMR tests were performed on an Agilent-400 DD2 spectrometer (Agilent, Palo Alto, CA) in DMSO-d₆ with tetramethylsilane (TMS) as the internal standard. The chemical shifts (δ) were expressed in units of parts per million (ppm) relative to TMS. High-resolution mass spectrometry (HR-MS) experiments were conducted on a time-of-flight Micromass LCT Premier XE spectrometer (McKinley Scientific, Sparta Township, NJ). Fluorescence spectra were recorded on a Cary Eclipse spectrophotometer (Agilent, Palo Alto, CA). UV–vis absorption spectra were recorded on a TU-1901 UV–vis spectrophotometer (PuXi Science and Technology Development Co. Ltd., Beijing, China). IR spectra were recorded on a Vary 1000 FT-IR spectrophotometer (Varian, Palo Alto, CA). Fluorescence images of the cells were obtained by an Olympus IX73+DP73 inverted microscope. Small-animal living imaging was performed on IVIS Lumina LT Series III instrument (Caliper Life Sciences, Hopkinton, MA). Additional experiment procedures are found in the Supporting Information.

Synthesis of Chemosensor L

A mixture of 8-hydroxyjulolidine-9-carboxaldehyde (0.23 g, 1 mmol) and 2-hydrazinylpyridine (0.10 g, 1 mmol) in MeOH (25 mL) was irradiated of 300 W at 60 °C for 15 min. After cooling the reaction mixture to an ambient temperature, the formed solid was filtered and washed with cold MeOH and diethyl ether. The crude product was purified by recrystallization from methanol to give 0.32 g of L as a yellow solid (85.1%). ¹H NMR (DMSO-d₆, 400 MHz) δ (ppm): 1.83(s, 4H, −CH₂−), 2.59(s, 4H, −CH₂−), 3.11(s, 4H, −CH₂−), 6.63(s, 1H, Ar–H), 6.70–6.71 (m, 1H, Ar–H), 6.74–6.76 (d, J = 8.4 Hz, 1H, Ar–H), 7.58 (s, 1H, Ar–H), 7.99 (s, 1H, Ar–H), 8.09(s, 1H, CH=N), 10.58 (s, 1H, NH), 11.20 (s, 1H, OH). ¹³C NMR (DMSO-d₆, 100 MHz) δ (ppm): 156.20, 153.51, 144.34, 138.21, 127.40, 114.34, 138.21, 127.40, 114.34, 112.55, 107.01, 106.70, 105.53, 26.72, 21.82, 21.06, 20.54. IR (KBr, cm^–1) ν: 3445, 2953, 1620, 1545, 1428, 1368, 1307, 1272, 1171, 967, 845, 775, 729. MS(ESI) calcd for C₁₉H₂₀N₄O₂ (M + H⁺): 310.2. Found: 310.2.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.1c05960.

• Additional experiment procedures, spectra (UV–vis absorption, fluorescence, NMR, ESI-MS, and XRD), imaging data, and full reference information(PDF)

The authors declare no competing financial interest.