Zn2+ detection of a benzimidazole 8-aminoquinoline fluorescent sensor by inhibited tautomerization

A new fluorescent chemosensor based on 8-aminoquinoline L1 bearing a benzimidazole moiety was synthesized, which exists as two predominant tautomers L1A and L1B in diluted DMSO-d6 solution. Among various metal ions, L1 showed a highly selective and sensitive turn-on fluorescence response to the presence of Zn2+ ions in methanol. The detection limit for Zn2+ by L1 was calculated to be 1.76 × 10−7 M. The 1 : 1 complexation ratio of the L1–Zn complex was confirmed through Job plot measurements. Complexation studies were performed by FT-IR, NMR and HR-ESI MS measurements and DFT calculations. With the gained insight, it was possible to successfully apply L1 in water sample analysis.


Introduction
Zinc is known to be a common essential element to all living organisms. It has proven to be of vital importance to various biological processes such as enzyme activity or DNA protection. 1 Though mostly present in complexed form, 2 a lack or excess of zinc causes disturbances in biological systems and is oen linked to a series of diseases, e.g. epilepsy, Alzheimer's disease or hypoxia ischemia. 3 Apart from biological processes, excess zinc concentrations cause environmental problems. Thus, soil microbial activities are unfavorably affected which subsequently leads to phytotoxicity and, in general, to lowered crop quality. 4 This is particularly problematic in former mining regions such as the Harz mountains (Lower Saxony, Germany) where we are located. Increased zinc concentrations are not only caused by contaminated sites as a result of centuries of mining activity, but also unintentionally by the local metal industry. 5 Zinc is difficult to detect spectroscopically due to its 3d 10 electron conguration. Even though various methods are known for its detection, most of them have not proven to be eld-operational. Oentimes, expensive equipment and extensive sample preparations are necessary for obtaining results. 6 Therefore, the demand for low cost and easy monitoring systems has been high. Fluorescent chemosensors have proven to be a valuable asset in the sensing and monitoring of heavy and transition metal ions.
Numerous uorescent chemosensors for Zn 2+ -sensing have been designed over the past decades. 7,8 Among many trace metal ion sensors, quinoline is a common uorophore used as a backbone in zinc sensing structures. 9,10 The application of various 8-hydroxyquinolates in the uorometric detection of zinc dates back as far as the 1960s. 11 8-Aminoquinoline based Zn 2+ -chemosensors are also known. 10,12,13 However, uorogenic Zn 2+ -sensors are not limited to quinoline structures only. Contemporary examples range from systems based on e.g. coumarin, 14 uorescein, 15 benzimidazole 16,17 or silsesquioxane. 18 Further contemporary examples can be found in Table S1  (ESI †).
Unfortunately, the interference of same group metals such as Hg or Cd can cause severe problems in Zn 2+ detection due to their similar properties. 19,20 Therefore, there is a great interest in the design of chemosensors that are easily synthesized, have a high sensitivity, a high response and can discriminate between Zn 2+ and Cd 2+ /Hg 2+ in real time.
The implementation of additional heteroatom containing fragments have been reported to enhance chelation abilities. 21 Intramolecular hydrogen bonds are oen observed between binding units present in aminoquinoline chemosensors. Upon addition of metal ions, these hydrogen bonds can be broken by chelation resulting in a uorescence emission due to stronger ICT processes. 12,22 On the other hand, it has been reported that the inhibition of prototropic tautomerization phenomena in benzimidazole fragments can be accompanied by a uorescence response. 16,17 In continuation of our interest in metal adducts and metal complexes of heterocycles such as mesomeric betaines and Nheterocyclic carbenes 23 in catalysis, 24 metal recovery and recycling, 25 we report here on a new 8-aminoquinoline based chemosensor L1 for the Zn 2+ detection. L1, bearing a benzimidazole moiety, was acquired through a simple two-step synthesis and exhibits a prototropic tautomerization, which was spectroscopically proven to be inhibited by Zn 2+ ions. Showing a highly selective and sensitive turn-on uorescence in the presence of Zn 2+ , L1 was examined by UV-vis, IR, 1 H NMR, high resolution electrospray ionization mass spectrometry (HRESIMS), and uorescence spectroscopy. DFT calculations have been carried out. Apart from the fact that L1 could successfully distinguish Zn 2+ from Cd 2+ and Hg 2+ , its potential use in water sample analysis is shown.

General
All chemicals used were purchased and used as received unless noted otherwise. NMR spectra were taken on a BRUKER Avance FT-NMR AVANCE III (600 MHz). DMSO-d 6 was used as NMR solvent with chemical shis (d) being reported in ppm. IR spectra (ATR-IR) were recorded on a BRUKER Alpha T in a range of 400-4000 cm À1 . Mass spectra were recorded on a BRUKER Impact II mass spectrometer. UV-vis measurements were performed on a JASCO V-550 spectrophotometer. Fluorescence measurements were performed on a JASCO FP-8500 spectro-uorometer using a prismatic cell to avoid inner-eld effects. All measurements were conducted at room temperature. The precursor 2-chloro-N-(quinolin-8-yl)acetamide was synthesized according to known literature procedures. 26 Preparation of 2-((5-methoxy-1H-benz[d]imidazol-2-yl)thio)-N-(quinolin-8-yl)acetamide L1 A sample of 127 mg (0.58 mmol) of 2-chloro-N-(quinolin-8-yl) acetamide, 80 mg (0.58 mmol) of potassium carbonate and 104 mg (0.58 mmol) of 5-methoxy-2-mercaptobenzimidazole was dissolved in 5 mL of acetone and reuxed for three hours. Upon completion, monitored by tlc, the reaction mixture was ltered and the solvent was removed in vacuo to afford 191 mg of a light brown solid in 91% yield, mp 180 C. 1 13   Fluorescence experiments with various metal ions 9 mL of a 10 mM solution of L1 (0.01 mmol in 1 mL of MeOH) were added to 2.991 mL of MeOH to make a nal concentration of 30 mM. Aerwards, 30 mL of a 30 mM MCl x -solution (M ¼ K + , Na + , Ba 2+ , Mg 2+ , Hg 2+ , Cu 2+ , Ca 2+ , Co 2+ , Cd 2+ , Ni 2+ , Al 3+ , Zn 2+ , 0.03 mmol in 1 mL of H 2 O) were titrated to the aforementioned solution of L1. Aer shaking the sample for a couple of seconds, the uorescence spectra were measured.  Competition experiments with various metal ions 9 mL of a 10 mM solution of L1 (0.01 mmol in 1 mL of MeOH) were added to 2.991 mL of MeOH to make a nal concentration of 30 mM. Aerwards, 30 mL of a 30 mM MCl x -solution (M ¼ K + , Na + , Ba 2+ , Mg 2+ , Hg 2+ , Cu 2+ , Ca 2+ , Co 2+ , Cd 2+ , Ni 2+ , Al 3+ , 0.03 mmol in 1 mL H 2 O) were titrated to the aforementioned solution of L1 followed by the addition of 30 mL of a ZnCl 2 solution. Aer shaking the sample for a couple of seconds, the uorescence spectra were taken.

UV-vis titration experiments
Job plot measurement 90 mL of a 10 mM solution of L1 were added to 29.91 mL of MeOH to make a nal concentration of 30 mM. This procedure was repeated for ZnCl 2 . Then, 2.7, 2.4, 2.1, 1.8, 1.5, 1.2, 0.9, 0.6, and 0.3 mL of L1 were transferred to individual vials. Aerwards, 0.3, 0.6, 0.9, 1.2, 1.5, 1.8, 2.1, 2.4 and 2.7 mL of the Zn 2+ solution were added separately to yield a total volume of 3 mL. Aer shaking the sample for a couple seconds, the uorescence spectra were taken.

NMR experiments
Samples of L1 in presence of different equivalents of anhydrous ZnCl 2 (0.5, 1.0, 2.5 eq.) were dissolved in DMSO-d 6 . Aerwards their 1 H NMR spectra were measured.

pH experiments
A series of MeOH : H 2 O (95 : 5, v/v) samples at different pH values were prepared by addition of dilute NaOH or HCl. Aer the desired pH value was set, 9 mL of L1 were added to 2.991 mL of pH-adjusted MeOH : H 2 O to make a 30 mM concentration. Aerwards, 6.0 mL of a 10 mM ZnCl 2 -solution (0.03 mmol in 1 mL H 2 O) were added to the aforementioned sample. Aer shaking the sample for a couple of seconds, the uorescence spectra were taken.

Theoretical calculations
DFT calculations were performed using ORCA 5 of Neese and co-workers. [27][28][29] This DFT package was run on a MS Windows 10 Pro based (Version 21H1) PC system equipped with an AMD Ryzen Threadripper 3970X 32-Core and 128 GB RAM in combination with the appropriate message passing interface MS-MPI 10.0.12498.5. MMFF optimized structures were used as starting geometries for the geometry optimizations with the recently published robust "Swiss army knife" composite method r 2 SCAN-3c of Grimme and co-workers 30 with D4 dispersion correction and geometrical counter poise correction applying the modied triple-zeta basis set def2-mTZVPP. Subsequent frequency calculation of the nal structure evidenced the absence of imaginary frequencies and thus the presence of true minima on the potential energy surface. In case of calculations that include a solvent, the Conductor-like Polarizable Continuum Model (CPCM) implemented in ORCA 5 was applied.

NMR calculations
Additionally, DFT calculated anisotropic NMR shis of tautomers L1A and L1B were obtained by means of SPARTAN'20 (www.wavefun.com) with the implemented NMR calculation method based upon the hybrid density functional with dispersion correction uB97X-D by Chai and Head-Gordon 31 and the standard basis set 6-31G*. The calculation soware was run on the abovementioned MS Windows 10 Pro PC system equipped with the AMD Ryzen Threadripper 3970X 32-core and 128 GB RAM.

Determination of Zn 2+ in water samples
An articially polluted water sample was added to a 30 mM solution of L1 in MeOH, which was prepared as aforementioned. Aer shaking the sample for a couple of seconds, the uorescence spectra were taken.

Tautomerism and hydrogen bonds of L1
The structure of L1 enables the formation of tautomers such as L1A-L1C (Scheme 1). In order to elucidate the structure of the predominant tautomers in DMSO, 1 H NMR and 13 C NMR studies were performed with concentrated as well as diluted solutions of L1 in DMSO-d 6 . In concentrated solutions, the N-H resonance frequencies of the amide as well as of H-N1 0 /H-N3 0 were observed as extremely broadened signals in the 1 H NMR spectra. Under these conditions, the chemical shi of 2-H of the quinoline ring of L1 resonates at d ¼ 8.80 ppm in DMSO-d 6 so that the contribution of the zwitterionic tautomer L1C can be neglected under these conditions. The hydrogen 2-H of quinolinium salts usually resonates at lower elds which was proven by signals of H-2 at d ¼ 8.93 ppm aer the addition of gaseous HCl to a solution of L1 in DMSO-d 6 . In accordance with the fact that tautomerization of imidazoles commonly leads to very broad and weak signals in the 13 C NMR spectra which cannot be detected under standard measurement conditions, 32 the detection of the 13 C NMR resonance frequencies of the benzimidazole carbon atoms C-3a 0 , C-7a', C-7 0 and C-4 0 required longterm measurements (Fig. S5 †). The predominant formation of the two tautomers L1A and L1B was then proven by NMR experiments with diluted solutions in DMSO-d 6 . Under these conditions two distinct sets of benzimidazole protons in addition to the NH resonance frequencies were detectable (Fig. S10-S12 †). Full assignment of both tautomers was possible by means of 1 H, 13 C-HMBC measurements, especially based on the remarkable carbon shi differences between the adjacent quaternary carbon atoms 3a' and 7a' of the benzimidazole unit. Thus, the signals of tautomer L1A shows a larger shi difference Dd between its C-3a 0 and C-7a 0 atoms (Dd ¼ 14.2 ppm) in comparison to L1B (Dd ¼ 1.6 ppm). The structure elucidation was strongly supported by DFT NMR shi calculations using the 6B97X-D functional and the 6-31G* standard basis set within the concurrent Spartan'20 soware. 31 An additional shi prediction 33 also promoted the structural assignment (Table  S2 † (ACD) were in very good agreement. Contrary to these interesting shi differences in the benzimidazole subunit, the corresponding NMR resonances of the quinoline part of the tautomers were virtually isochronous. The calculated structure of L1 in DMSO shows transoid amide bonds with respect to NH-C]O which are almost coplanar with the quinoline rings, respectively. The conformer of tautomer L1B is calculated to be 1.1 kJ mol À1 more stable than the corresponding tautomer L1A. This small difference is reected experimentally by an almost equalized tautomer ratio (43% L1A : 57% L1B) in the diluted DMSO-d 6 solution (Fig. 1).

Selectivity of L1 for metal ions
To assess the selectivity of chemosensor L1, uorescence spectra in the presence of various metal ions were taken. Herein, K + , Na + , Ba 2+ , Mg 2+ , Hg 2+ , Cu 2+ , Ca 2+ , Co 2+ , Cd 2+ , Ni 2+ , Al 3+ and Zn 2+ were examined ( Fig. 2 and 3). The measurements were conducted at an excitation wavelength of l ex ¼ 291 nm. For a better comparability, this average value was determined from isobestic points and absorbance maxima of various structural analogues that we conduct research on. It is evident that chemosensor L1 shows no visible uorescence in methanol under the measurement conditions. However, upon addition of ten equivalents of Zn 2+ ions a broad uorescence band at 510 nm was observed. The narrow peak at 582 nm results from light reected from the hypotenuse of the prismatic cell with twice the wavelength of the excitation light. In contrast to Zn 2+ ions, no signicant changes in the uorescence behavior were observed when other metal ions were present. This indicated that L1 is not only suitable as a selective turn-on detector for Zn 2+ ions, but also suitable for distinguishing Zn 2+ from metal ions of the same group, i.e. Hg 2+ and Cd 2+ .

Binding properties of L1
To examine the binding properties of L1, titration experiments were conducted via uorescence and UV-vis spectroscopy. Fluorescence titration experiments have shown, that upon incremental addition of Zn 2+ , an increasing turn-on uorescence was observed at 510 nm (l ex ¼ 291 nm, Fig. 4). Chemosensor L1 showed no uorescence at 510 nm, which might be due to a photo-induced electron transfer (PET) to the quinoline moiety induced by benzimidazole nitrogen atoms. 34 As shown in Fig. 2, a strong uorescence enhancement was observed when Zn 2+ was added which can be attributed to a chelationinduced enhanced uorescence (CHEF). 35 Further addition of Zn 2+ past 1 eq. did not cause signicant changes regarding the uorescence intensity.
Additionally, UV-vis experiments have been conducted to further examine the binding properties. The UV-vis absorbance spectra of L1 (30 mM) in methanol display two distinct absorption bands at 241 and 300 nm, respectively (Fig. 5). These bands have been assumed to be due to p-p* and n-p* transitions of aminoquinolines. 35 Additionally, these absorption bands redshied to 256 and 360 nm accompanied by three isobestic points at 247, 286 and   337 nm. The spectral response suggests the formation of only one L1-Zn-complex. 12,21,35 Furthermore, the incremental addition of Zn 2+ (0-2 eq.) showed saturation at a L1-Zn-ratio of 1 : 1, as the absorption ratio A 360 /A 300 did not change signicantly aer 1 eq. (Fig. 5, inset). The results derived from the titration experiments indicated that formation of a 1 : 1 complexation must be the case.

Job plot and Benesi-Hildebrand analysis
In order to verify the stoichiometry, a Job plot analysis was performed. 36 As seen in Fig. 6, the emission maximum was observed at a molar fraction of 0.5. 8,34 This indicated that a 1 : 1 complex was formed, which is also visible in the HR-ESI mass spectra (Fig. S1 †) (Fig. 7). 37 Plotting 1/DF against 1/[Zn 2+ ] yielded a linear regression. Using titration data, the Benesi-Hildebrand equation for 1 : 1 complexes is dened as follows: 38 The binding constant was calculated to be K b ¼ 2.16 Â 10 3 M À1 for the L1-Zn-complex and is in accordance to expected values, according to literature (1-10 12 ). 8,39 Detection limit The limit of detection (LOD) was calculated by the equation 3s/ s. 40 Herein, s represents the standard deviation of blank measurements and s is the slope between the uorescence intensity and Zn 2+ concentration (Fig. S3 †). The standard deviation s over six blank measurements was calculated to be 1.8678.    According to the equation, the detection limit of L1 was found to be 1.76 Â 10 À7 M, which proved to be much lower than the WHO guideline (76 mM) for Zn 2+ ions in drinking water. 41 In comparison to other studies, our determined LOD appeared to be lower than reported Zn 2+ chemosensors (Table  S1 †). Additionally, the reversibility of the L1-Zn complex was examined. Upon addition of excess EDTA, the uorescence emission of the L1-Zn complex was successfully reverted. This proved the reversible use of the synthesized chemosensor L1 (Fig. S2 †).

Competition experiments
In order to examine the effect of other cations on the uorescence emission of the L1-Zn complex, competition experiments were conducted (Fig. 8). In presence of 10 equivalents of Zn 2+ cations various metal ions have been added to the L1 sample. Ba 2+ , Ca 2+ , Co 2+ and K + ions have proven to show no effect, whereas Na + and Ni 2+ caused negligible uorescence quenching to the L1-Zn complex. The presence of same group metal ions, Cd 2+ and Hg 2+ , caused no interference to the uorescence emission induced by Zn 2+ . This additionally proved that chemosensor L1 can easily distinguish Zn 2+ from Cd 2+ and Hg 2+ . However, a strong quenching phenomenon was observed in the presence of both Al 3+ and Cu 2+ ions. It is known that Zn 2+ detection can be quenched in the presence of Cu 2+ and that a cation-exchange reaction between zinc and metal ions such as Al 3+ can take place. 19,42 The Zn-selective behaviour can be explained through Pearson's HSAB model. 43 Due to the harder nature of the incorporated oxygen and nitrogen atoms, it is evident that Zn 2+ , Cu 2+ and Al 3+ , as harder metal centres, preferably interact with these receptor sites. Furthermore, it has been reported that the incorporation of nitrogen and oxygen atoms into ligand systems has proven to favour the complexation of Zn 2+ ions in contrast to other metal ions. 44 pH experiments The pH-dependence of various quinoline chemosensors has been reported. 12,45 To assess the photophysical properties, the uorescence emission was examined at different pH values in MeOH : H 2 O (95 : 5, v/v). As seen in Fig. 9, L1, in presence of Zn 2+ , exhibits the strongest uorescence emission at a pH value of 8. In contrast, under strongly acidic or basic conditions a considerable uorescence quenching is observed. At low pH values, this might be attributed to possible protonation of nitrogen sites in quinoline or benzimidazole moieties. 46 Fluorescence quenching at higher pH values might be due to the deprotonation of NH fragments resulting in a stronger PET towards the uorophore. Nevertheless, in the pH range from 4 to 10 L1 exhibits a satisfactory uorescence response with a peak at a pH value of 8, thus demonstrating that the detection of Zn is possible under physiological pH conditions.

Complexation studies
NMR spectroscopic analyses were nally conducted to investigate the binding behaviour of L1 in presence of Zn 2+ ions (Fig. 10). Upon addition of Zn 2+ to the concentrated solution of L1 in DMSO-d 6 , two distinct sharp -NH signals were observed at 12.61 and 11.17 ppm, respectively, indicative of an inhibited tautomerism. Whereas the signals of H-4, H-5, H-6 and H-7 of L1 shied only slightly on addition of Zn 2+ , the signals of H-2 and H-3 were considerably broadened, hinting at a complexation through the quinoline N-atom. As the benzimidazole protons H-4 0 and H-7 0 showed signicant upeld shis on complexation with Zn 2+ , one of its N-atoms obviously is involved in complexation [e.g. Dd(H-4 0 ) ¼ 0.11 ppm; e.g. Dd(H-7 0 ) ¼ 0. 16 ppm]. The third complexation site can be identied by 13 C NMR spectroscopy. Thus, the addition of Zn 2+ ions induced a significant shi of the 13 C NMR resonance frequencies of the carbonyl carbon atom to higher elds [Dd(C]O) ¼ 0.159 ppm, Fig. S8 †], accompanied by a considerably enlarged peak width at half-height. Signicant changes were also observed in case of the signals of the benzimidazole carbon atoms C-5 0 and C-6 0 Fig. 8 Competition studies of L1 (30 mM) toward Zn 2+ (10 eq.) in the presence of various metal ions (10 eq.) in MeOH (l ex ¼ 291 nm). ( Fig. S9 †). IR-spectroscopic investigations unambiguously support the participation of the carbonyl oxygen upon zinc complexation in the solid state. The carbonyl stretching vibration of the free ligand L1 appears at 1661 wavenumbers (ATR IR), whereas this band is shied to 1595 cm À1 in the zinc complex (Fig. S4b †). Actually, in the DFT calculated IR of this complex (functional r 2 SCAN-3c; def2-mTZVPP basis set) this crucial band is found with excellent congruence at 1598 cm À1 (Fig S4c †).
We performed comparative DFT calculations of the Zn complex with the recently published r 2 SCAN-3c method of Grimme and co-workers 30 recently implemented in ORCA 5 [27][28][29] and additional consideration of DMSO by the CPCM solvent model supported by ORCA.
The calculations resulted in a structure of the complex which is in total accordance with the experimental data and which is shown in Fig. 11. Utilizing Pearson's HSAB model, 43 the sulfur atom was readily ruled out as a potential complexation site. As a borderline metal ion, Zn 2+ shows a greater affinity towards harder oxygen and nitrogen atoms as opposed to the soer sulphur centre which is in accordance with previous studies. 47 The zinc complex based on ligand tautomer L1A is energetically favoured with a difference of 3.7 kJ mol À1 in comparison to L1B. This is probably caused by the positive mesomeric effect of the methoxy group in 5-position of the benzimidazole moiety that supports the nitrogen donor centre of the benzimidazole moiety. DFT calculations predict that the tautomer L1A is xed upon addition of Zn 2+ and the PET to the uorophore deriving from the nitrogen atom of the benzimidazole moiety is inhibited, resulting in a chelation-induced enhanced uorescence (CHEF) of L1A-Zn (Fig. S17 †).

Determination of Zn 2+ in water samples
Finally, the applicability of L1 was tested with water samples. We created articially polluted water samples by the addition of Ca 2+ , Na + , K + , Mg 2+ aside metal ions of the same group, Cd 2+ and Hg 2+ , to water. Plotting the uorescence intensity against the Zn 2+ concentration yielded a linear calibration plot (Fig. S3 †) which was used to determine the Zn 2+ content in given water samples. Table 1 shows that L1 was successfully able to recover the given Zn 2+ concentrations even in the presence of various metal ions. Therefore, it is safe to assume that L1 could potentially be used for Zn 2+ detection in real water samples.

Conclusions
To summarize, we have designed and synthesized a new chemosensor L1 based on 8-aminoquinoline bearing a benzimidazole moiety. L1 showed a high selectivity and sensitivity towards Zn 2+ in methanol, which was accompanied by a distinct green uorescence emission. Moreover, L1 was capable of distinguishing Zn 2+ from same group metal ions Cd 2+ and Hg 2+ . The LOD was determined to be 0.176 mM, which proved to be lower than the WHO standard (76 mM). Spectroscopic studies have shown that a 1 : 1 complexation takes place, which upon addition of EDTA showed the possible reversibility of the L1-Zn complex. The prototropic tautomerism exhibited by the benzimidazole moiety was used as proof to successfully identify the binding sites. Furthermore, the capability of L1 to quantify Zn 2+ in water samples was shown. Hence, we believe that L1 shows a great potential for use in both biological and environmental applications.

Conflicts of interest
There are no conicts to declare. Fig. 10 Aromatic proton shift of L1 (blue, concentrated solution in DMSO-d 6 ), upon addition of 1 eq. ZnCl 2 (red). The addition past 1 eq. of Zn 2+ yielded no further changes (Fig. S7 †).