Acridino-Diaza-20-Crown-6 Ethers: New Macrocyclic Hosts for Optochemical Metal Ion Sensing

Acridino-diaza-20-crown-6 ether derivatives as new turn-on type fluorescent chemosensors with an excellent functionality and photophysical properties have been designed and synthesized for metal ion-selective optochemical sensing applications. Spectroscopic studies revealed that in an acetonitrile-based semi-aqueous medium, the sensor molecules exhibited a remarkable fluorescence enhancement with high sensitivity only toward Zn2+, Al3+ and Bi3+, among 23 different metal ions. Studies on complexation showed a great coordinating ability of logK > 4.7 with a 1:1 complex stoichiometry in each case. The detection limits were found to be from 59 nM to micromoles. The new ionophores enabled an optical response without being affected either by the pH in the range of 5.5–7.5, or the presence of various anions or competing metal ions. Varying the N-substituents of the new host-backbone provides diverse opportunities in both immobilization and practical applications without influencing the molecular recognition abilities.


Introduction
Acridine and its derivatives have been extensively studied. Researchers typically pay attention to the favorable fluorescent properties of these heterocyclic dyes. Unsurprisingly, acridines have for many years also become widespread in the field of supramolecular chemistry. They usually play an important role as optochemically active units of sensor molecules by providing the signaling function in the sensing process. In addition to the receptor unit responsible for molecular recognition, the most advanced direct-type optochemical sensor molecules also contain a signaling unit-such as an acridine fluorophore-and are able to fulfill these two basic functions of chemical sensors together [1]. They can be exemplified by a large number of recently developed optochemical ionophores containing a 4,5-dimethyleneacridine unit [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18], which is shown in Figure 1.

Introduction
Acridine and its derivatives have been extensively studied. Researchers typically pay attention to the favorable fluorescent properties of these heterocyclic dyes. Unsurprisingly, acridines have for many years also become widespread in the field of supramolecular chemistry. They usually play an important role as optochemically active units of sensor molecules by providing the signaling function in the sensing process. In addition to the receptor unit responsible for molecular recognition, the most advanced direct-type optochemical sensor molecules also contain a signaling unit-such as an acridine fluorophore-and are able to fulfill these two basic functions of chemical sensors together [1]. They can be exemplified by a large number of recently developed optochemical ionophores containing a 4,5-dimethyleneacridine unit [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18], which is shown in Figure 1. Regarding the soft nucleophilic character of the heteroaromatic nitrogen, the application of these sensor molecules is mostly based on their selectivity for typically soft cations [2][3][4][5][6][7][8] such as Hg 2+ , Cd 2+ , Zn 2+ , Fe 3+ , Ni 2+ , but their representatives also include anion- Regarding the soft nucleophilic character of the heteroaromatic nitrogen, the application of these sensor molecules is mostly based on their selectivity for typically soft cations [2][3][4][5][6][7][8] such as Hg 2+ , Cd 2+ , Zn 2+ , Fe 3+ , Ni 2+ , but their representatives also include a pH of about 7.5. In contrast, low pH can also significantly limit the applications, thus proper pH-range-tolerating sensors are in the highest demand.
Herein we report the synthesis, photophysical characterization and complexation properties of acridino-diaza-20-crown-6 ethers as a new class of fluorescent chemosensors for metal ion detection. Based on the present work, the new sensor molecules hold great promise for the future development of ion-selective optochemical sensing methods and devices.

Synthesis
For the initial step of the synthesis, the bromomethylation of acridine (1) at positions 4 and 5 was carried out based on a reported method [55] using bromomethyl methyl ether (BMME) as a reagent. Since, in our experience, the commercially available reagent did not have the proper quality, the alkylating agent was freshly prepared according to a previously reported procedure [56] as outlined in Scheme 1.
Molecules 2021, 26, x 3 of 21 about 7.5. In contrast, low pH can also significantly limit the applications, thus proper pHrange-tolerating sensors are in the highest demand.
Herein we report the synthesis, photophysical characterization and complexation properties of acridino-diaza-20-crown-6 ethers as a new class of fluorescent chemosensors for metal ion detection. Based on the present work, the new sensor molecules hold great promise for the future development of ion-selective optochemical sensing methods and devices.

Synthesis
For the initial step of the synthesis, the bromomethylation of acridine (1) at positions 4 and 5 was carried out based on a reported method [55] using bromomethyl methyl ether (BMME) as a reagent. Since, in our experience, the commercially available reagent did not have the proper quality, the alkylating agent was freshly prepared according to a previously reported procedure [56] as outlined in Scheme 1. Bis(bromomethyl)acridine 5 is an important starting material for the subsequent synthetic steps, thus its preparation has been optimized. The yields of the reactions were examined as a function of three conditions. Experimental data can be found in Table 1. It can be seen that raising the temperature and increasing the reaction time resulted in moderately increased yields. The optimization resulted in a 7% higher yield than that which was reported in the literature (described in [55]), after a chromatographic purification. Further studies were carried out to improve the yield by modifying the work-up procedure. When the reported chromatographic purification step [55] was omitted and the crude product was first triturated with propan-2-ol and then recrystallized twice from methanol, the yield increased to 79%. In this case, the recommended chromatographic Bis(bromomethyl)acridine 5 is an important starting material for the subsequent synthetic steps, thus its preparation has been optimized. The yields of the reactions were examined as a function of three conditions. Experimental data can be found in Table 1. It can be seen that raising the temperature and increasing the reaction time resulted in moderately increased yields. The optimization resulted in a 7% higher yield than that which was reported in the literature (described in [55]), after a chromatographic purification. Further studies were carried out to improve the yield by modifying the work-up procedure. When the reported chromatographic purification step [55] was omitted and the crude product was first triturated with propan-2-ol and then recrystallized twice from methanol, the yield increased to 79%. In this case, the recommended chromatographic purification was not necessary. The use of a larger excess of reagent and a temperature above 70 • C Molecules 2021, 26, 4043 4 of 20 resulted in more by-products. In cases when the times of the reactions were longer than 48 h, no higher conversions were observed.
Bis(bromomethyl)acridine 5 was reacted with three different primary amines to prepare the new secondary amines required for the macrocyclization reactions as outlined in Scheme 2.
Molecules 2021, 26, x 4 of 21 purification was not necessary. The use of a larger excess of reagent and a temperature above 70 °C resulted in more by-products. In cases when the times of the reactions were longer than 48 h, no higher conversions were observed. Bis(bromomethyl)acridine 5 was reacted with three different primary amines to prepare the new secondary amines required for the macrocyclization reactions as outlined in Scheme 2. Scheme 2. Preparation of new secondary amines (7, 11 and 14) and their macrocyclization with tetraethylene glycol diiodide (8).
Diiodo-compound 8 was prepared from its dichloro-analogue according to a reported procedure [57]. It is noteworthy to mention, that diiodo-compound 8 undergoes rapid decomposition in light or above 70 °C; therefore, it is recommended that it should be stored below −30 °C and in the dark until use. The first reaction gave the expected secondary amine 7, but the macrocyclization failed, probably due to the steric repulsion of the bulky tert-butyl groups. The macrocyclization with diiodide 8 and diamine 11 gave the expected macrocycle (12), but with a poor yield. The reaction of diamine 14 with dibromo-compound 5, and the subsequent macrocyclization, gave a N-protected azacrown ether derivative 15.
The structure of the obtained macrocycles is advantageous from several aspects. The heteroaromatic unit provides sufficient rigidity to the host to render the required selectivity, while the flexible methylene groups attached to positions 4 and 5 of the heterocycle allow a rapid conformational change, thus allowing the easy uptake of the active ion-trap conformation. Since the fluorophore unit is a part of the coordination sphere, the complexation of a guest ion induces photophysical changes directly. While the macrocycle
Diiodo-compound 8 was prepared from its dichloro-analogue according to a reported procedure [57]. It is noteworthy to mention, that diiodo-compound 8 undergoes rapid decomposition in light or above 70 • C; therefore, it is recommended that it should be stored below −30 • C and in the dark until use. The first reaction gave the expected secondary amine 7, but the macrocyclization failed, probably due to the steric repulsion of the bulky tert-butyl groups. The macrocyclization with diiodide 8 and diamine 11 gave the expected macrocycle (12), but with a poor yield. The reaction of diamine 14 with dibromo-compound 5, and the subsequent macrocyclization, gave a N-protected azacrown ether derivative 15.
The structure of the obtained macrocycles is advantageous from several aspects. The heteroaromatic unit provides sufficient rigidity to the host to render the required selectivity, while the flexible methylene groups attached to positions 4 and 5 of the heterocycle allow a rapid conformational change, thus allowing the easy uptake of the active ion-trap conformation. Since the fluorophore unit is a part of the coordination sphere, the complexation of a guest ion induces photophysical changes directly. While the macrocycle containing benzyl groups (12) have a highly lipophilic character (logP = 6.5, logD pH=7.0 = 3.8 predicted by ChemAxon), its analogue 15 with allyl groups is suitable for direct polymerization.
The former property facilitates the physical immobilization, the latter one enables the covalent immobilization of the sensor molecule, paving the way for a wide range of future applications.
Activating the amino function of N-benzyl-protected macrocycle 12 was carried out by catalytic hydrogenation followed by oxidation of the unstable intermediate 16 as outlined in Scheme 3. Moreover, macrocycle 17 containing secondary amine groups can also be used as a partially water-soluble sensor molecule (logD pH=7.0 = −1.3 predicted by ChemAxon), a property which can be exploited in bioimaging.
Molecules 2021, 26, x 5 of 21 containing benzyl groups (12) have a highly lipophilic character (logP = 6.5, logDpH=7.0 = 3.8 predicted by ChemAxon), its analogue 15 with allyl groups is suitable for direct polymerization. The former property facilitates the physical immobilization, the latter one enables the covalent immobilization of the sensor molecule, paving the way for a wide range of future applications. Activating the amino function of N-benzyl-protected macrocycle 12 was carried out by catalytic hydrogenation followed by oxidation of the unstable intermediate 16 as outlined in Scheme 3. Moreover, macrocycle 17 containing secondary amine groups can also be used as a partially water-soluble sensor molecule (logDpH=7.0 = −1.3 predicted by Che-mAxon), a property which can be exploited in bioimaging.

Scheme 3.
Removal of the benzyl groups to gain substitutable amine functions.
In addition to the deprotection, the reduction of the acridine unit to acridane 16 is inevitable in these conditions. Acridane 16 was oxidized by elemental oxygen to acridine 17 in an acetic acid-ethanol mixture.
In the case of diallylamino-macrocycle 15, deprotection was carried out catalytically in mild conditions based on reported analogies [58]. This is particularly important because hard conditions and acids can induce polymerization of the allyl groups. The catalyst was tetrakis(triphenylphosphine)palladium(0), which promoted the intermolecular trans-allylation reaction between the tertiary amine and 1,3-dimethylbarbituric acid via redoxprocesses as shown in Scheme 4. Macrocycle 17 is a stable compound due to the benzylamine nature of its amino groups. Furthermore, it has a widely modifiable synthetic receptor backbone due to its secondary amine functionalities. This new macrocyclic derivative will hopefully serve as a precursor for a number of acridino-20-crown-6 ether-type fluorescent chemosensors.

Spectral Properties
Determination of quantum yields was carried out according to the method based on the comparison with a reference [59]. Acridine (1) was selected as a reference compound In addition to the deprotection, the reduction of the acridine unit to acridane 16 is inevitable in these conditions. Acridane 16 was oxidized by elemental oxygen to acridine 17 in an acetic acid-ethanol mixture.
In the case of diallylamino-macrocycle 15, deprotection was carried out catalytically in mild conditions based on reported analogies [58]. This is particularly important because hard conditions and acids can induce polymerization of the allyl groups. The catalyst was tetrakis(triphenylphosphine)palladium(0), which promoted the intermolecular transallylation reaction between the tertiary amine and 1,3-dimethylbarbituric acid via redoxprocesses as shown in Scheme 4. containing benzyl groups (12) have a highly lipophilic character (logP = 6.5, logDpH=7.0 = 3.8 predicted by ChemAxon), its analogue 15 with allyl groups is suitable for direct polymerization. The former property facilitates the physical immobilization, the latter one enables the covalent immobilization of the sensor molecule, paving the way for a wide range of future applications. Activating the amino function of N-benzyl-protected macrocycle 12 was carried out by catalytic hydrogenation followed by oxidation of the unstable intermediate 16 as outlined in Scheme 3. Moreover, macrocycle 17 containing secondary amine groups can also be used as a partially water-soluble sensor molecule (logDpH=7.0 = −1.3 predicted by Che-mAxon), a property which can be exploited in bioimaging.

Scheme 3.
Removal of the benzyl groups to gain substitutable amine functions.
In addition to the deprotection, the reduction of the acridine unit to acridane 16 is inevitable in these conditions. Acridane 16 was oxidized by elemental oxygen to acridine 17 in an acetic acid-ethanol mixture.
In the case of diallylamino-macrocycle 15, deprotection was carried out catalytically in mild conditions based on reported analogies [58]. This is particularly important because hard conditions and acids can induce polymerization of the allyl groups. The catalyst was tetrakis(triphenylphosphine)palladium(0), which promoted the intermolecular trans-allylation reaction between the tertiary amine and 1,3-dimethylbarbituric acid via redoxprocesses as shown in Scheme 4. Macrocycle 17 is a stable compound due to the benzylamine nature of its amino groups. Furthermore, it has a widely modifiable synthetic receptor backbone due to its secondary amine functionalities. This new macrocyclic derivative will hopefully serve as a precursor for a number of acridino-20-crown-6 ether-type fluorescent chemosensors.

Spectral Properties
Determination of quantum yields was carried out according to the method based on the comparison with a reference [59]. Acridine (1) was selected as a reference compound

Scheme 4. Removal of the allyl groups in an intermolecular trans-allylation reaction in mild conditions.
Macrocycle 17 is a stable compound due to the benzylamine nature of its amino groups. Furthermore, it has a widely modifiable synthetic receptor backbone due to its secondary amine functionalities. This new macrocyclic derivative will hopefully serve as a precursor for a number of acridino-20-crown-6 ether-type fluorescent chemosensors.

Spectral Properties
Determination of quantum yields was carried out according to the method based on the comparison with a reference [59]. Acridine (1) was selected as a reference compound (Φ = 3.50 × 10 −4 in acetonitrile [60]) and an excitation wavelength of 249 nm was chosen. The UV/Vis absorption and fluorescence emission spectra of the new crown compounds are shown in Figure 2.  The quantum yields of sensor molecules 12, 15, and 17 were calculated to be 2.60 × 10 −4 , 2.21 × 10 −3 , and 2.06 × 10 −3 , respectively, indicating a relatively poor fluorescence in the form of free host molecules. The absorption peak wavelengths were 253-254 nm, while the emission peak wavelengths were 424-437 nm, which resulted in extremely high Stokes shifts of 171-184 nm. This is an advantageous spectral property as it helps to eliminate autofluorescence, thus facilitates the sensitive analysis of real samples.
(Φ = 3.50 × 10 −4 in acetonitrile [60]) and an excitation wavelength of 249 nm was chosen. The UV/Vis absorption and fluorescence emission spectra of the new crown compounds are shown in Figure 2. The quantum yields of sensor molecules 12, 15, and 17 were calculated to be 2.60 × 10 −4 , 2.21 × 10 −3 , and 2.06 × 10 −3 , respectively, indicating a relatively poor fluorescence in the form of free host molecules. The absorption peak wavelengths were 253-254 nm, while the emission peak wavelengths were 424-437 nm, which resulted in extremely high Stokes shifts of 171-184 nm. This is an advantageous spectral property as it helps to eliminate autofluorescence, thus facilitates the sensitive analysis of real samples.

Studies on Metal Ion Complexation
Studies on metal ion-selectivity were carried out by adding 23 different metal salts (with carbonate counterions: Rb 2+ , Li + , Cs + ; sulfate counterions: Mn 2+ , Fe 2+ ; a hydroxide counterion: Ba 2+ ; chloride counterions: Sr 2+ , Al 3+ , Hg 2+ , Bi 3+ ; an iodide counterion: Cd 2+ ; acetate counterions: K + , Ni 2+ , Co 2+ , Na + , Cu 2+ , Ag + , Ca 2+ , Zn 2+ , Mg 2+ , and nitrate counterions: Cr 3+ , Pd 2+ , Pb 2+ ) in 50 mM aqueous solutions in 10 molar equivalent amounts (regarding the host molecule) separately to the solution of the macrocycle in acetonitrile. Among the 23 metal salts, only Zn 2+ , Al 3+ and Bi 3+ caused a detectable change in the fluorescence signal as shown in Figure 3.  In the cases of the other metal salts, no spectral change was observed, indicating that no complexation took place with these cations. To the best of our knowledge, macrocycle 15 is the first one to show selectivity toward Bi 3+ . Furthermore, complexed cations rarely occur as contaminants of each other in practical analysis. Although these macrocycles tend to complex three different ions, they can be considered selective in practice, as interference with competing contaminants is not expected in real samples. In the cases of macrocycles 12 and 17, it was found that the different N-substituents did not affect their complexing properties (the spectra can be found in Supplementary Materials). This is not surprising as the substituents are not parts of the coordination sphere of the macrocycle. This property allows the new host analogues to be diversely N-substituted without altering their coordination ability, resulting in an excellent functionality of this macrocyclic backbone.
In order to determine the stability constants (K) of the ligand-metal ion complexes (detailed calculation method can be found in Section 4.2) the acetonitrile solution of the host molecule was titrated with aqueous solutions of the preferred three metal salts. The observed titration series in the case of macrocycle 15 with Zn 2+ , Al 3+ and Bi 3+ are shown in Figure 4.
In the cases of the other metal salts, no spectral change was observed, indicating that no complexation took place with these cations. To the best of our knowledge, macrocycle 15 is the first one to show selectivity toward Bi 3+ . Furthermore, complexed cations rarely occur as contaminants of each other in practical analysis. Although these macrocycles tend to complex three different ions, they can be considered selective in practice, as interference with competing contaminants is not expected in real samples. In the cases of macrocycles 12 and 17, it was found that the different N-substituents did not affect their complexing properties (the spectra can be found in Supplementary Materials). This is not surprising as the substituents are not parts of the coordination sphere of the macrocycle. This property allows the new host analogues to be diversely N-substituted without altering their coordination ability, resulting in an excellent functionality of this macrocyclic backbone.
In order to determine the stability constants (K) of the ligand-metal ion complexes (detailed calculation method can be found in Section 4.2) the acetonitrile solution of the host molecule was titrated with aqueous solutions of the preferred three metal salts. The observed titration series in the case of macrocycle 15 with Zn 2+ , Al 3+ and Bi 3+ are shown in Figure 4. To determine the complex stability constants, a nonlinear regression curve was globally fitted on titration data based on the least squares method, which is shown in Figure 5 in the case of titration with Zn 2+ . To determine the complex stability constants, a nonlinear regression curve was globally fitted on titration data based on the least squares method, which is shown in Figure 5 in the case of titration with Zn 2+ .
Based on the calculations, 1:1 complex stoichiometry is preferred with every metal ion. Since studies showed no significant dependence of the molecular recognition ability on the different N-substituents of the macrocycles, experimental results for macrocycle 15 were only reported here (further spectroscopic examinations on compounds 12 and 17 can be found in the Supplementary Materials). The calculated logK and LOD values are summarized in Table 2. Based on the calculations, 1:1 complex stoichiometry is preferred with every metal ion. Since studies showed no significant dependence of the molecular recognition ability on the different N-substituents of the macrocycles, experimental results for macrocycle 15 were only reported here (further spectroscopic examinations on compounds 12 and 17 can be found in the Supplementary Materials). The calculated logK and LOD values are summarized in Table 2. The results also showed that there were no significant differences among the recognition abilities of macrocycles (12, 15 and 17) toward the investigated metal ions. The stability constants were above 4.7 in all cases, which indicate the formation of stable complexes among crown derivatives. However, these stability constants are not too high to inhibit the reversible operation of the sensor molecules, thus they are suitable candidates for optochemical sensing.
Reversibility and regenerability are important features for the practical application of the chemosensors. Therefore, studies on complexation were carried out several times after removing the solvents and extracting the dichloromethane solutions of the host molecules with distilled water. After regeneration, an identical enhancement of the emission intensity could be observed upon addition of 10 molar equivalents of metal salts (except in the case of macrocycle 17 due to its partial water-solubility), indicating that the new macrocycles show a reversible complexation. The regeneration could be carried out effectively by using only distilled water for extraction. No additional chelating agents, like  The results also showed that there were no significant differences among the recognition abilities of macrocycles (12, 15 and 17) toward the investigated metal ions. The stability constants were above 4.7 in all cases, which indicate the formation of stable complexes among crown derivatives. However, these stability constants are not too high to inhibit the reversible operation of the sensor molecules, thus they are suitable candidates for optochemical sensing.
Reversibility and regenerability are important features for the practical application of the chemosensors. Therefore, studies on complexation were carried out several times after removing the solvents and extracting the dichloromethane solutions of the host molecules with distilled water. After regeneration, an identical enhancement of the emission intensity could be observed upon addition of 10 molar equivalents of metal salts (except in the case of macrocycle 17 due to its partial water-solubility), indicating that the new macrocycles show a reversible complexation. The regeneration could be carried out effectively by using only distilled water for extraction. No additional chelating agents, like EDTA were required for decomplexation, which makes the new sensor molecules easy to regenerate.

Acid-Base Properties
In order to determine the limits of application, it is very important to study the proton association ability of the new macrocycles and the effects of protonation on spectral properties. It is known that acridine is a weak base [61] with a pK a (conjugate acid at Molecules 2021, 26, 4043 9 of 20 20 • C) of 5.58 in water. Consequently, different ionization states are present upon acidifying the aqueous medium, which strongly influence molecular recognition. The pK a values of the protonated aliphatic amine units of the macrocycle are 9.0 ± 0.5 in water (predicted by ChemAxon), which suggest that these nitrogens are mostly protonated in a neutral aqueous medium. The pK a of the N-protonated heteroaromatic unit of macrocycle 15 was also determined in acetonitrile due to its poor solubility in water. Nitric acid dissolved in acetonitrile was gradually added to the solution of macrocycle 15. The series of fluorescence emission spectra are shown in Figure 6.
EDTA were required for decomplexation, which makes the new sensor molecules easy to regenerate.

Acid-Base Properties
In order to determine the limits of application, it is very important to study the proton association ability of the new macrocycles and the effects of protonation on spectral properties. It is known that acridine is a weak base [61] with a pKa (conjugate acid at 20 °C) of 5.58 in water. Consequently, different ionization states are present upon acidifying the aqueous medium, which strongly influence molecular recognition. The pKa values of the protonated aliphatic amine units of the macrocycle are 9.0 ± 0.5 in water (predicted by ChemAxon), which suggest that these nitrogens are mostly protonated in a neutral aqueous medium. The pKa of the N-protonated heteroaromatic unit of macrocycle 15 was also determined in acetonitrile due to its poor solubility in water. Nitric acid dissolved in acetonitrile was gradually added to the solution of macrocycle 15. The series of fluorescence emission spectra are shown in Figure 6.  The calculated pKa of the N-protonated acridine unit of macrocycle 15 in acetonitrile was 9.6 ± 0.1. Compared to the aqueous medium, the pKa showed an expected increase. A molar proportional increase in fluorescence intensity was observed during the titration. A large bathochromic shift of the emission maximum indicated the appearance of a new molecular form. The pK a was calculated according to the mathematical method detailed in Section 4.2. The applied globally fitted nonlinear regression curve is shown in Figure 7.
EDTA were required for decomplexation, which makes the new sensor molecules easy to regenerate.

Acid-Base Properties
In order to determine the limits of application, it is very important to study the proton association ability of the new macrocycles and the effects of protonation on spectral properties. It is known that acridine is a weak base [61] with a pKa (conjugate acid at 20 °C) of 5.58 in water. Consequently, different ionization states are present upon acidifying the aqueous medium, which strongly influence molecular recognition. The pKa values of the protonated aliphatic amine units of the macrocycle are 9.0 ± 0.5 in water (predicted by ChemAxon), which suggest that these nitrogens are mostly protonated in a neutral aqueous medium. The pKa of the N-protonated heteroaromatic unit of macrocycle 15 was also determined in acetonitrile due to its poor solubility in water. Nitric acid dissolved in acetonitrile was gradually added to the solution of macrocycle 15. The series of fluorescence emission spectra are shown in Figure 6.  The calculated pKa of the N-protonated acridine unit of macrocycle 15 in acetonitrile was 9.6 ± 0.1. Compared to the aqueous medium, the pKa showed an expected increase. The calculated pK a of the N-protonated acridine unit of macrocycle 15 in acetonitrile was 9.6 ± 0.1. Compared to the aqueous medium, the pK a showed an expected increase.
Regarding structural analogy of the new macrocycles, further studies on protonation and complexation were carried out only in the case of macrocycle 15. Since the basicity of the complex of crown compounds differs from that of the free host, the pK a of the protonated complex was also determined by acidic titration after adding 10 molar equivalents of Zn 2+ to the macrocycle in acetonitrile (Figure 8).
Regarding structural analogy of the new macrocycles, further studies on protonation and complexation were carried out only in the case of macrocycle 15. Since the basicity of the complex of crown compounds differs from that of the free host, the pKa of the protonated complex was also determined by acidic titration after adding 10 molar equivalents of Zn 2+ to the macrocycle in acetonitrile (Figure 8). The pKa of the conjugate acid of the complex was 8.7 ± 0.1, which is a decrease of about one unit compared to the proton dissociation constant of the N-protonated acridine unit of the uncomplexed ligand, thus protonation equilibrium is suppressed by complex formation. In the complex form, no proton-induced change in emission took place in the pH range of 5.5-7.5 in acetonitrile-water mixtures, allowing the selective metal ion detection in a wide range of the biologically most relevant pH values.
It was found that the presence of protonated aliphatic amine units of the crown compound did not affect the stable coordination of cations; however, the doublepositively charged parts carry potential for complexing anions. In order to study the interference of different counterions used in metal ion-selectivity studies, solutions of tetrabutylammonium salts of the corresponding anions (tetrabutylammonium cations cannot be complexed) were added to macrocycle 15 in acetonitrile.

Coordination Ability toward Anions
Various tetrabutylammonium salts (H2PO4 − , NO3 − , HSO4 − , CH3COO − , F − , Cl − , Br − , I − in 50 mM solutions) in distilled water were added to macrocycle 15 in 10 molar equivalent amounts with regard to the host molecule. No significant spectral change was observed in any of the cases (Figure 9). Thus, the double-positively charged macrocycle in a neutral water-acetonitrile medium did not coordinate the studied anions. The pK a of the conjugate acid of the complex was 8.7 ± 0.1, which is a decrease of about one unit compared to the proton dissociation constant of the N-protonated acridine unit of the uncomplexed ligand, thus protonation equilibrium is suppressed by complex formation. In the complex form, no proton-induced change in emission took place in the pH range of 5.5-7.5 in acetonitrile-water mixtures, allowing the selective metal ion detection in a wide range of the biologically most relevant pH values.
It was found that the presence of protonated aliphatic amine units of the crown compound did not affect the stable coordination of cations; however, the double-positively charged parts carry potential for complexing anions. In order to study the interference of different counterions used in metal ion-selectivity studies, solutions of tetrabutylammonium salts of the corresponding anions (tetrabutylammonium cations cannot be complexed) were added to macrocycle 15 in acetonitrile.

Coordination Ability toward Anions
Various tetrabutylammonium salts (H 2 PO 4 − , NO 3 − , HSO 4 − , CH 3 COO − , F − , Cl − , Br − , I − in 50 mM solutions) in distilled water were added to macrocycle 15 in 10 molar equivalent amounts with regard to the host molecule. No significant spectral change was observed in any of the cases (Figure 9). Thus, the double-positively charged macrocycle in a neutral water-acetonitrile medium did not coordinate the studied anions. The reported macrocycles are weak organic bases, thus they are able to accep protons in acidic medium. Not only the double-positively charged macrocycles, but als the triple-positively charged ones can have molecular recognition abilities different fro those of the corresponding neutral forms, thus studies on anion-complexation were als carried out after acidifying macrocycle 15.
A solution of nitric acid in acetonitrile was added to macrocycle 15 in an amou The reported macrocycles are weak organic bases, thus they are able to accept protons in acidic medium. Not only the double-positively charged macrocycles, but also the triplepositively charged ones can have molecular recognition abilities different from those of the corresponding neutral forms, thus studies on anion-complexation were also carried out after acidifying macrocycle 15.
A solution of nitric acid in acetonitrile was added to macrocycle 15 in an amount corresponding to the end point of its acid titration (800 eq. H + ). Then, 50 mM solutions of tetrabutylammonium salts with a 10-fold excess in distilled water, were similarly added to a solution of the macrocycle 15 protonated at all nitrogens ( Figure 10). The reported macrocycles are weak organic bases, thus they are able to acce protons in acidic medium. Not only the double-positively charged macrocycles, but a the triple-positively charged ones can have molecular recognition abilities different fro those of the corresponding neutral forms, thus studies on anion-complexation were al carried out after acidifying macrocycle 15.
A solution of nitric acid in acetonitrile was added to macrocycle 15 in an amou corresponding to the end point of its acid titration (800 eq. H + ). Then, 50 mM solutions tetrabutylammonium salts with a 10-fold excess in distilled water, were similarly add to a solution of the macrocycle 15 protonated at all nitrogens ( Figure 10). No spectral change was observed in this case either, thus the new macrocycle do not form complexes with anions even in the fully-protonated state.

Conclusions
The parent fluorescent macrocycle (17) with easily functionalizable secondary ami units and its two analogues (12 and 15) were reported, covering the synthetic chemi background of these new hosts. The benzylated derivative (12) is preferred for physi immobilization, while the allyl analogue (15) is suitable for covalent attachme Moreover, the improved water solubility of the parent macrocycle (17) can make selecti imaging in living organisms possible. The new sensor molecules showed favorab spectral features-i.e., large Stokes shifts of 171-184 nm and a weak fluorescen background signal of Φ < 3 × 10 −3 in absence of the preferred ions-for optical detectio Complexation studies were carried out with 23 metal ions. The sensor molecules show turn-on fluorescence responses in the presence of Zn 2+ , Al 3+ and Bi 3+ , which generally No spectral change was observed in this case either, thus the new macrocycle does not form complexes with anions even in the fully-protonated state.

Conclusions
The parent fluorescent macrocycle (17) with easily functionalizable secondary amine units and its two analogues (12 and 15) were reported, covering the synthetic chemical background of these new hosts. The benzylated derivative (12) is preferred for physical immobilization, while the allyl analogue (15) is suitable for covalent attachment. Moreover, the improved water solubility of the parent macrocycle (17) can make selective imaging in living organisms possible. The new sensor molecules showed favorable spectral features-i.e., large Stokes shifts of 171-184 nm and a weak fluorescence background signal of Φ < 3 × 10 −3 in absence of the preferred ions-for optical detection. Complexation studies were carried out with 23 metal ions. The sensor molecules showed turn-on fluorescence responses in the presence of Zn 2+ , Al 3+ and Bi 3+ , which generally do not occur as contaminants of each other. The limits of detection were from 59 nM to micromoles for each crown derivatives. Studies on complexation showed a 1:1 complex stoichiometry and large stability constants of logK > 4.7 for each preferred metal ion in an acetonitrile-water medium. Reversibility and regenerability were proved by the simple and effective decomplexation and metal ion removal by extraction with distilled water. Studies showed a weak influence of the N-substituents on complexation, basicity and signaling. The new chemosensors did not tend to coordinate various types of anions even in their protonated forms. The pK a of fluoroionophore 15 protonated on its acridine nitrogen atom was 9.6 in acetonitrile, which decreased about one unit upon metal ion complexation and allowed pH-independent chemosensing in the range of 5.5-7.5 in an acetonitrile-water medium. The reported new macrocycles have promising signaling properties and enable both diverse immobilization techniques and several post-synthetic modifications for future development, especially for optical sensor applications.

Chemicals, Apparatus and Measurements
Starting materials and reagents were purchased from Sigma-Aldrich Corporation (USA, owned by Merck, Darmstadt, Germany) and used without further purification unless otherwise noted. Solvents were dried and purified according to well established methods [62]. Silica Gel 60 F254 (Merck, Germany) and aluminum oxide 60 F254 neutral type E (Merck, Germany) plates were used for thin-layer chromatography (TLC). All reactions were monitored by TLC and visualized by UV-lamp. Aluminum oxide (neutral, activated, Brockman I) and Silica Gel 60 (70-230 mesh, Merck) were used for column chromatography. Purifications by preparative thin-layer chromatography (PTLC) were carried out using Silica gel 60 F254 (Merck, Germany) plates of 2 mm layer thickness (art No.: 1.05744) or aluminum oxide 60 F254 neutral type E (Merck, Germany) plates of 0.25 mm layer thickness (art No.: 1.05727). Ratios of solvents for the eluents are given in volumes (mL/mL). Evaporations were carried out under reduced pressure unless otherwise stated.
The new compounds were characterized by their physical constants such as melting point, thin-layer chromatography retention factor (R f ), infrared, 1 H-NMR and 13 C-NMR spectroscopies and HRMS spectrometry. Melting points were taken on a Boetius micromelting point apparatus and are uncorrected. Infrared spectra were recorded on a Bruker Alpha-T FT-IR spectrometer (Bruker Corporation, Billerica, MA, USA) using KBr pastilles. NMR spectra were recorded on a Bruker 300 Avance spectrometer (Bruker Corporation, USA; at 300 MHz for 1 H and at 75.5 MHz for 13 C spectra). HRMS analyses were carried out on a Thermo Velos Pro Orbitrap Elite (Thermo Fisher Scientific, Dreieich, Germany) system. The ionization method was ESI and was operated in positive ion mode. The protonated molecular ion peaks were fragmented by CID at a normalized collision energy of 35-45%. The samples were dissolved in methanol. Data acquisition and analysis were accomplished with Xcalibur software version 2.2 (Thermo Fisher Scientific, Germany).
UV/Vis spectra were recorded on a UNICAM UV4-100 spectrophotometer controlled by VIZION 3.4 software (ATI UNICAM, Hatley Saint George, UK). Fluorescence emission spectra were recorded on a Perkin-Elmer LS 50B luminescent spectrometer (PerkinElmer Inc., Waltham, MA, USA) and were corrected by FL Winlab 3.0 spectrometer software (PerkinElmer Inc., USA). Quartz cuvettes with a path length of 1 cm were used in all cases. Spectroscopic measurements were carried out at room temperature (25 ± 1 • C). During spectrophotometric titrations, the solutions were added with a Hamilton syringe to the acetonitrile solutions of the ligands. As macrocycles (except 17) showed poor water solubility, their acetonitrile solutions were used for spectrophotometric studies in all cases. The reported spectra were corrected in each case with the background signal of the added solutions and concentration values were also corrected corresponding to the caused dilution.

Evaluation of the Results
OriginPro 8.6 (OriginLab Corp., Northampton, MA, USA) software was used for evaluation and visualization of the spectroscopic results.
Relative quantum yields were determined in acetonitrile according to a literature method [59] based on a comparison with acridine as a standard [60]. The excitation and emission spectra were recorded in the same conditions and instrument settings as in the case of the standard. The following equation was used for calculations: where subscript i refers to the sample of the initial investigated compound, while subscript r refers to the reference. The Φ is the quantum yield, n is the respective refractive index of the solvents, I is the fluorescence intensity, λ ex is the excitation wavelength, λ em is the emission wavelength and A is the absorbance. The stability constants of the complexes were determined by global non-linear regression analysis. For determination of the complex stability constant based on the observed fluorescence enhancement upon complexation, the following equation was used [63]: where F is the measured fluorescence intensity, I 0 is the intensity of the emission, Φ is the fluorescence quantum yield, ε is the molar absorption coefficient, b is the optical path length, [X] is the molar concentration of the X species and k X is a constant referring to the optical properties of the X species.
In the case of complexes with 1:1 stoichiometry, the association constant can be calculated by the following equation: where the ratios of k parameters and K were left as floating parameters during the fitting method. Parameters F and F 0 are wavelength-dependent variables and [G] was set as a variable, too. F 0 refers to the initial fluorescence intensity of the free host molecule, k H is a constant referring to the optical properties of the free host molecule, k 0 H is a constant referring to the optical properties of the free host in the presence of preferred guest molecules, constant k HG describes the photophysical features of the complex, K is the association constant and [G] is the concentration of the initial guest molecules.
Global non-linear fitting was carried out similarly in the case of complexes with 1:2 (host:guest) stoichiometry based on the following equation: where ∆F obs is the change in fluorescence during titration steps, k ∆HG = k HG − k H , [H] 0 is the initial concentration of the host, K 1 is the association constant of the first step of the complex formation equilibrium, while K 2 is the association constant of the second step of the complexation. The described method for the complexes with a 2:1 (host:guest) stoichiometry was carried out based on the following mathematical formula: where [G] 0 is the initial concentration of the guest molecule and [H] is the concentration of the free hosts. Standard errors were calculated analytically from the regression and indicated as uncertainties of the logK determination. The investigations of the complex stoichiometries were also carried out by applying the described global non-linear fitting methods. These results were compared in terms of the quality of fit indicators. The choice of the model was made by a statistical F-probe, which aimed to test the sum of squares from each model fitting. In the cases of formulas for 1:1 stoichiometry, the observed F-values supported the null hypothesis that the fitted model described an appropriate relationship for characterizing the population of experimental data at a confidence level of 95%. In contrast, in the cases of models for other stoichiometries, the test implies that the applied models did not fit the experimental data. (Job plot was not used for stoichiometry estimations as reported mathematical algorithms can solve these nonlinear polynomials by exactly expressing the photophysical changes during titration experiments. Studies on every crown analogue confirm the determined 1:1 stoichiometry of complexes since the one macrocycle ring is the only coordination sphere of the hosts.) Titration experiments were carried out with careful consideration of the relevant recommendations [64].
Limits of detection were also calculated based on the data of fluorescence titrations. To determine the signal-to-noise ratio, the fluorescence intensity of the sensor molecule was measured nine times and the standard deviation of these blank measurements was determined. Three separate measurements were carried out in the presence of the metal ions and a linear regression on the average of the measured intensities was fitted as a function of concentration of the initial metal ions to determine the slope. The limit of detection was calculated using the following equation [65]: where d is the standard deviation of the optical signal of the free host and s is the slope of the emission intensities as a function of the concentration of the guest. Determination of pK a in a nonaqueous medium was carried out based on the following equation [66]: where F is the measured fluorescence intensity, F max is the fluorescence intensity at the starting point of acid titration, [H + ] refers to the proton concentration, n shows the number of associated protons/molecules, F min is the fluorescence intensity at the end point of acid titration, K acid is the acid dissociation constant of the investigated compound. During the fitting method, the n and the K acid were defined as floating parameters in the equation. The value of n proved to be close to 1, thus it was set as a constant. Based on the known values of variable [H + ] and wavelength-dependent variables F, F max , F min , parameter K acid can be determined.
In the absence of water, the spectral change is in a direct relationship with the protonation of the sensor molecule in acetonitrile, thus Equation (7) contains the concentration of H + as a variable instead of the pH. During the calculation it was considered that the proton dissociation of nitric acid is strongly reduced in acetonitrile compared to the estimated total dissociation of protons in water. The pK a for nitric acid in acetonitrile is 10.6 [67]. The [H + ] values in Equation (7) were corrected with the degree of dissociation (α) corresponding to the concentration of nitric acid using the Ostwald's dilution law: where c is the corresponding molar concentration of nitric acid and K s is the dissociation constant of the acid derived from its pK a determined in acetonitrile. A mixture of acridine (1) (1.00 g, 5.58 mmol) and concentrated H 2 SO 4 (98%, 12.5 mL) was stirred under argon at room temperature and BMME (4) (3.04 g, 22.32 mmol) was added to it in one portion. The temperature of the mixture was raised to 65 • C, when intense formation of reddish-brown gas was observed. The mixture was maintained at this temperature for 48 h, then it was poured onto 300 g of crushed ice and stirred slowly for 1 h. The precipitate was filtered off and dissolved in chloroform. The organic phase was dried over MgSO 4 , filtered and the solvent was removed. The resulting yellow solid was purified by chromatography on silica gel, using a hexane/chloroform (2:1) mixture as eluent to give 5 (1.12 g, 55%) as a bright yellow powder.

Synthesis of the New Compounds
When the crude product was first triturated with propan-2-ol and then recrystallized twice from methanol, the purification by chromatography was unnecessary and the yield increased to 79% (1.60 g).
R f = 0.47 (SiO 2 TLC, hexane/chloroform 2:1). All other physical and spectroscopic data of 5 concurred with those reported in the literature [55]. 4.3.2. Preparation of N,N -(Acridine-4,5-diylbis(methylene))bis(2-methylpropan-2-amine) (7) Freshly distilled tert-butyl amine (29 mL, 274 mmol) was added dropwise to 4,5bis(bromomethyl)acridine 5 (500 mg, 1.37 mmol) under an argon atmosphere at −75 • C using an acetone-dry ice cooling bath. The reaction mixture was stirred at this temperature for 1 h, then the temperature was gradually increased by 30 • C per hour while the reaction was continuously monitored by TLC. After reaching room temperature, the reaction mixture was stirred for an additional 1 h to reach complete conversion. The excess of amine was removed by distillation and the crude product was purified without any work-up procedure by column chromatography on neutral aluminum oxide using a methanol/dichloromethane (1:20) mixture as an eluent. The product was further purified by PTLC on aluminum oxide using dichloromethane as an eluent to give 7 (244 mg, 51%) as yellow crystals. Freshly distilled benzyl amine (30 mL, 274 mmol) was added dropwise to 4,5-bis-(bromomethyl)acridine 5 (500 mg, 1.37 mmol) under an argon atmosphere at −75 • C using an acetone-dry ice cooling bath. The reaction mixture was stirred at this temperature for 1 h, then the temperature was gradually increased by 30 • C per hour while the reaction was continuously monitored by TLC. After reaching room temperature, the reaction mixture was stirred for an additional 1 h to reach complete conversion. The excess of amine was removed by distillation and the crude product was purified without any work-up procedure by column chromatography on neutral aluminum oxide using a methanol/dichloromethane (1:20) mixture as an eluent. The product was further purified by PTLC on aluminum oxide using dichloromethane as an eluent to give 11 (520 mg, 91%) as a brown viscous oil. Freshly distilled allylamine (21 mL, 274 mmol) was added dropwise to 4,5-bis-(bromomethyl)acridine 5 (500 mg, 1.37 mmol) under an argon atmosphere at −75 • C using an acetone-dry ice cooling bath. The reaction mixture was stirred at this temperature for 1 h, then the temperature was gradually increased by 30 • C per hour while the reaction was continuously monitored by TLC. After reaching room temperature, the reaction mixture was stirred for an additional 1 h to reach complete conversion. The excess of amine was removed by distillation and the crude product was purified without any work-up procedure by column chromatography on neutral aluminum oxide using a methanol/dichloromethane (1:50) mixture as an eluent. The product was further purified