Bisquinoline-based fluorescent cadmium sensors

Abstract

Rational molecular design afforded fluorescent Cd²⁺ sensors based on bisquinoline derivatives. Introduction of three methoxy groups at the 5,6,7-positions of the quinoline rings of BQDMEN (N,N′-bis(2-quinolylmethyl)-N,N′-dimethylethylenediamine) resulted in the reversal of metal ion selectivity in fluorescence enhancement from zinc to cadmium. Introduction of bulky alkyl groups and an N,N-bis(2-quinolylmethyl)amine structure, as well as replacement of one of the two tertiary amine binding sites with an oxygen atom and the use of a 1,2-phenylene backbone significantly improved the Cd²⁺ specificity. The fluorescent cadmium ion selectivity could be explained by the differential binding with Cd²⁺ and Zn²⁺, and the formation of a bis(μ-chloro) dinuclear cadmium complex in contrast to the mononuclear zinc complex.

---

1. Introduction

Detection and quantification of toxic heavy metal ions are currently important objectives in regulation of environmental pollutants. Fluorescence sensing provides high selectivity and sensitivity toward targeted metal ions via a rapid analytical protocol using relatively inexpensive equipment. One of the most important issues in the detection of toxic metal ions is discrimination of Cd²⁺ from naturally abundant Zn²⁺, both of which are group 12 elements in the periodic table and exhibit only a 21 pm difference in their ionic radii. Since recent human activities in industry increase the exposure of cadmium to air, water and soil, continuous development of fluorescent probes for strict detection of cadmium, especially in environmental water, is in high demand. Many research groups have been investigating quinoline-based molecules for this purpose considering that the long coordination distances and softness of the quinoline nitrogen atom in comparison with conventional pyridine ligands would be suitable for selective binding with Cd²⁺.^1–10 In this Frontier article, several modifications of the molecular structure of BQDMEN (N,N′-bis(2-quinolylmethyl)-N,N′-dimethylethylenediamine (1), Fig. 1)¹¹ for better Cd²⁺-selectivity in the fluorescence response are discussed (Table 1). A preceding Frontier article by the present author, dealing with a related hexadentate tetrakisquinoline ligand TQEN (N,N,N′,N′-tetrakis(2-quinolylmethyl)ethylenediamine, Fig. 1), contains other examples for fluorescence sensing of Zn²⁺, Cd²⁺, Hg²⁺ and phosphate species.¹²

Fig. 1 Structure of BQDMEN derivatives 1–3 and TQEN.

Table 1 Fluorescence properties of Cd²⁺ probes

	λ ex (nm)	λ em (nm)	I Cd/I0 (eq.)	I Zn/ICd (eq.)	K d (metal ion) (M)	ϕ Cd	Ref.
a In DMF-H2O (1 : 1). b In DMF-HEPES buffer (1 : 1).
TriMeOBQDMEN (3)^a	339	472	15 (1)	0.22 (1)	∼10^−8 (Cd)	0.29	25
					∼10^−7 (Zn)
3 ^b	335	458	25 (3)	0.24 (3)	3 × 10^−7 (Cd)	—	29
					3 × 10^−6 (Zn)
TriMeOBQDIEN (5)^a	339	459	23 (20)	0.04 (20)	4 × 10^−6 (Cd)	0.37	26
					5 × 10^−5 (Zn)
TriMeOBQDBEN (6)^a	337	459	27 (3)	0.20 (3)	∼10^−7 (Cd)	0.33	26
					8 × 10^−6 (Zn)
TriMeOBQDMPHEN (9)^b	341	467	51 (1)	0.26 (1)	8 × 10^−7 (Cd)	0.43	27
					6 × 10^−6 (Zn)
TriMeOBQDIPHEN (10)^b	343	480	427 (40)	0.05 (40)	1 × 10^−3 (Cd)	—	27
					4 × 10^−2 (Zn)
TriMeO-N,O-BQMAE (11)^b	338	470	10 (3)	0.34 (3)	1 × 10^−5 (Cd)	0.28	29
					8 × 10^−5 (Cd)
TriMeO-N,N-BQMAE (12)^b	340	465	12 (3)	0.26 (3)	∼10^−7 (Cd)	0.23	29
					2 × 10^−5 (Zn)
TriMeOBQMOA (14)^b	341	464	90 (5)	0.02 (5)	1 × 10^−5 (Cd)	0.18	30

---

2. Polymethoxy substitution on the quinoline ring

The BQDMEN 1 has been reported as a tetranitrogen ligand supporting mononuclear and dinuclear metal complexes including Fe²⁺, Co²⁺, Ni²⁺, Pd²⁺, Cu^2+/+, Ru²⁺, Cr⁰ and Mn^3+/2+ centres.^13–20 We utilized this skeleton for fluorescent Zn²⁺ sensors and improved the fluorescence quantum yield by introducing a methoxy group to the quinoline rings.¹¹ The fluorescence turn-on mechanism includes the inhibition of photo-induced electron transfer (PeT) and chelation enhanced fluorescence (CHEF).^21–23 Thus-obtained 6-MeOBQDMEN (2) (Fig. 1) exhibits sufficient fluorescence intensity and Zn²⁺ selectivity (ϕ_Zn = 0.28 and I_Zn/I_Cd = 2.2 in DMF–H₂O (1 : 1)†) for intracellular study. On the other hand, the unsubstituted BQDMEN 1 exhibits dim fluorescence (ϕ_Zn = 0.017) but slightly higher Zn²⁺ selectivity (I_Zn/I_Cd = 3.9) than 2. Although we have known that further methoxy substitution does not improve the quantum yield from our study on TQEN derivatives,²⁴ we expected that the fluorescence metal ion selectivity could shift (or even reverse) to the Cd²⁺ side. This estimation is indeed true, in which TriMeOBQDMEN 3 (Fig. 1) exhibits Cd²⁺ selectivity (I_Zn/I_Cd = 0.22) with similar fluorescence quantum yield (ϕ_Cd = 0.29) to the Zn²⁺ complex with 2 (Table 1).²⁵ Interestingly, the extension of the carbon chain in the molecular skeleton from ethylenediamine to propanediamine also shifted the fluorescence response toward the Cd²⁺ side.

The fluorescence titration revealed that TriMeOBQDMEN (3) binds an equimolar amount of metal ions. The Cd²⁺ complex with 3 exhibits significantly long fluorescence lifetime (τ = ∼30 ns) in comparison with the Zn²⁺ complex (∼20 ns), indicating the different complex structure and/or fluorescence pathway. X-ray crystallography revealed the structures of the bis(μ-chloro) dinuclear cadmium complex and mononuclear zinc complex with BQDMEN (1) (Fig. 2). ESI-MS also confirmed the formation of [(μ-Cl)₂Cd₂(3)₂](ClO₄)⁺ species under the spectral measurement conditions. The metal binding affinity of 3 is high enough to saturate the ligand in the presence of 1 equiv. of Cd²⁺ and Zn²⁺ under our experimental conditions (Table 1). These observations clearly indicate that the BQDMEN ligand discriminates Cd²⁺ and Zn²⁺ by differential complexation, and the introduction of three methoxy groups on the quinoline rings successfully highlights the difference in the complex structure via fluorescence outputs. The next step would be an improvement of the insufficient Cd²⁺ selectivity of 3. Several modifications of the molecular structure were explored to selectively reduce the Zn²⁺-induced fluorescence signal.

Fig. 2 Perspective view of (a) [(μ-Cl)₂Cd₂(1)₂]²⁺ (CSD refcode CIBKOR) and (b) [Zn(1)Cl]ClO₄ (CSD refcode CIBKUX). Reproduced from ref. 25 with permission from the Royal Society of Chemistry.

3. Introduction of bulky alkyl groups

Two methyl groups of TriMeOBQDMEN (3) attached to the aliphatic nitrogen atoms were replaced with isopropyl groups to afford TriMeOBQDIEN (N,N′-bis(5,6,7-trimethoxy-2-quinolylmethyl)-N,N′-diisopropylethylenediamine (5), Fig. 3).²⁶ This sterically hindered bisquinoline derivative exhibits weak metal binding affinity (Table 1) and shows characteristic fluorescence titration profiles responding to Cd²⁺ and Zn²⁺ (Fig. 4). The highly fluorescent Cd²⁺ complex with 5 (ϕ_Cd = 0.37) was formed by the addition of ∼5 equiv. of Cd²⁺ and was stable in the excess amount of metal ions; however, the gradual addition of Zn²⁺ to 5 exhibited small fluorescence enhancement around 5 equiv. of metal ions, and then, the fluorescence significantly decreased in the presence of an excess amount of zinc salt, Zn(ClO₄)₂. The monomethoxy derivative (4) (Fig. 3) also exhibits similar profiles. After careful investigations, this observation was explained by the selective decomplexation of the zinc complex by protons generated by excess Zn(ClO₄)₂ used in the titration experiment. Cd(ClO₄)₂ did not drop the pH significantly, and the use of Zn(NO₃)₂ or DMF-HEPES buffer (1 : 1) as a solvent did not disrupt the Zn²⁺-induced fluorescence even in the presence of a large amount of metal ions. Nevertheless, the excellent Cd²⁺ specificity in fluorescence enhancement achieved in 5 (I_Zn/I_Cd = 4% in the presence of 20 equiv. of metal ions) in DMF–H₂O (1 : 1) is of significant interest.

Fig. 3 Structure of BQDMEN derivatives 4–7.

Fig. 4 Fluorescence intensity plot of 5 with increasing amounts of (a) Cd²⁺ and (b) Zn²⁺.²⁶

Although X-ray crystallography employing 4 afforded only mononuclear Cd²⁺ complexes, the long fluorescence lifetime (τ = ∼30 ns) for the Cd²⁺ complex with 5 suggested the possible formation of bis(μ-chloro) dicadmium species assembled in the excited state even in the isopropyl derivative. The corresponding benzyl (TriMeOBQDBEN (6)) and phenyl (TriMeOBQDPEN (7)) derivatives (Fig. 3) resulted in similar metal ion selectivity and fluorescence stability with 3 and an extremely poor fluorescence response with any metal ions, respectively.

4. Introduction of a benzene skeleton

Instead of the introduction of bulky substituents, replacement of an ethylenediamine backbone of BQDMEN (1) with a 1,2-phenylenediamine also reduces metal binding affinity because the anilinic nitrogen atoms of BQDMPHEN (N,N′-bis(2-quinolylmethyl)-N,N′-dimethyl-1,2-phenylenediamine (8), Fig. 5) have weaker basicity than the aliphatic nitrogen binding site of 1.²⁷ This approach has been successful for the TQEN-based tetrakisquinoline compound, TQPHEN (N,N,N′,N′-tetrakis(2-quinolylmethyl)-1,2-phenylenediamine, Fig. 5), which scarcely binds to Zn²⁺ under the spectral measurement conditions and therefore exhibits Cd²⁺-specific fluorescence enhancement.²⁸ Upon addition of metal ions in DMF-HEPES buffer (1 : 1),‡ the ligand 8 exhibited similar fluorescence enhancement with Cd²⁺ and Zn²⁺ (I_Zn/I_Cd = 0.90 in the presence of 1 equiv. of metal ion), but apparent Cd²⁺ selectivity (I_Zn/I_Cd = 0.26 in the presence of 1 equiv. of metal ions) was achieved in the trimethoxy derivative (TriMeOBQDMPHEN (9), Fig. 5). The insufficient metal ion selectivity was further improved by introducing isopropyl groups as discussed above, affording high fluorescent Cd²⁺ specificity in TriMeOBQDIPHEN (10) (I_Zn/I_Cd = 0.05 in the presence of 40 equiv. of metal ions) (Fig. 5).

Fig. 5 Structure of BQDMPHEN derivatives 8–10 and TQPHEN.

Considering the extremely weak metal binding affinity of the isopropyl ligand 10 (K_d(Cd) = 1 × 10⁻³ M), the methyl counterpart 9 (K_d(Cd) = 8 × 10⁻⁷ M) is more suitable for practical use (Table 1). This high metal binding affinity causes diminished Cd²⁺/Zn²⁺ selectivity estimated from the ratio of binding constants (K_d(Zn)/K_d(Cd) = 8 and 32 for 9 and 10, respectively); however, the high fluorescence quantum yield (ϕ_Cd = 0.43) and wide pH window (pH = 4–10) of 9 are worth exploring. The structures of Cd²⁺ and Zn²⁺ complexes with 8 are both mononuclear but, here again, the long fluorescence lifetime (τ = ∼30 ns) for Cd²⁺ complexes with 9 and 10 suggests the excimer-like fluorescence mechanism including quinoline–quinoline interactions in the excited state.

4. Replacement of a nitrogen atom with an oxygen atom

Another strategy to reduce the metal binding affinity of the tetranitrogen ligand TriMeOBQDMEN (3), aiming at the improvement of fluorescent Cd²⁺ selectivity, is the replacement of one of the two aliphatic ethylenediamine nitrogen atoms of 3 with an oxygen atom. Thus-designed 2-aminoethanol-based N3O1 ligand, TriMeO-N,O-BQMAE (N,O-bis(5,6,7-trimethoxy-2-quinolylmethyl)-2-methylaminoethanol (11), Fig. 6) exhibited two order lower metal binding affinity and slightly decreased Cd²⁺ preference in fluorescence enhancement in comparison with 3 (I_Zn/I_Cd = 0.34 and 0.24 for 11 and 3, respectively, in the presence of 3 equiv. of metal ions in DMF-HEPES buffer (1 : 1)) (Table 1).²⁹ Interestingly, TriMeO-N,N-BQMAE (12) (Fig. 6), the regioisomer of 11, in which both quinoline moieties are attached to the same aliphatic nitrogen atom of the aminoethanol skeleton, exhibits one order higher metal binding affinity than 11 and similar fluorescent Cd²⁺ selectivity to 3 (I_Zn/I_Cd = 0.26 for 12). The fluorescence quantum yield of 12 (ϕ_Cd = 0.23) is slightly smaller in comparison with those of 3 (0.29) and 11 (0.28) under the same experimental conditions. This difference is also reflected by the fluorescence lifetimes, where long fluorescence lifetimes (τ = ∼30 ns) with a possible intramolecular excimer emission for the Cd²⁺ complex were observed only for 3 and 11. For the N,N-isomer 12, no difference in fluorescence lifetimes of Zn²⁺ and Cd²⁺ complexes was detected (τ = ∼18 ns), indicating the monomer emission for both metal complexes with this ligand. The ligand structure of 12 strictly controls the metal binding affinity, possibly due to the difference in the complex structures, affording the fluorescent Cd²⁺ selectivity. It is important to note that the two order difference in dissociation constants between Cd²⁺ and Zn²⁺ complexes was achieved in the N,N-bisquinoline structure of 12 (Table 1). The corresponding N,N-isomer for 3 has not been examined due to problems in its synthesis.

Fig. 6 Structure of BQMAE derivatives 11 and 12.

5. Use of a benzene skeleton with an oxygen binding site

Finally, the N3O1 ligand with a 1,2-phenylene skeleton was explored. Considering the synthetic accessibility, only the N,N-isomer was examined here. Thus-designed TriMeOBQMOA (N,N-bis(5,6,7-trimethoxy-2-quinolylmethyl)-2-methoxyaniline (14), Fig. 7) exhibits excellent Cd²⁺ specificity in fluorescence enhancement (I_Zn/I_Cd = 0.02 in the presence of 5 equiv. of metal ions) with moderate metal binding affinity (K_d(Cd) = 1 × 10⁻⁵ M) and fluorescence quantum yield (ϕ_Cd = 0.18) (Table 1).³⁰ The binding affinity of 14 with Zn²⁺ is extremely weak. The LOD (limit of detection) for Cd²⁺ was estimated to be 25 nM, which is lower than the environmental limit of water in Japan (3 ppb, 27 nM).

Fig. 7 Structure of BQMOA and BQDMPHEN derivatives 13–15.

Both Cd²⁺ and Zn²⁺ complexes with BQMOA (13) (Fig. 7) were structurally characterized by X-ray crystallography (Fig. 8) where the bis(μ-chloro) dinuclear cadmium complex ([(μ-Cl)₂Cd₂(13)₂]²⁺) and mononuclear zinc complex ([Zn(13)Cl₂]) with an uncoordinated quinoline moiety were revealed. These complex structures clearly explain the reason for the short fluorescence lifetime (τ = ∼13 ns) for the Cd²⁺ and Zn²⁺ complexes with 14 because of the lack of interquinoline stacking interaction even in the bis(μ-chloro) dicadmium complex. The fluorescent Cd²⁺ specificity is a result of the extremely weak binding affinity of this ligand with Zn²⁺.

Fig. 8 Perspective view of (a) [(μ-Cl)₂Cd₂(13)₂]²⁺ (CSD refcode YOMQOK) and (b) [Zn(13)Cl₂] (CSD refcode YOMQUQ). Adapted from ref. 30 with permission from American Chemical Society, copyright 2024.

The importance of chloride ions in HEPES buffer was also disclosed in this study. As shown in Fig. 9, chloride ions exclusively enhance the fluorescence intensity of 14 in the presence of Cd²⁺via enhanced complexation, but no such effect was observed for Zn²⁺ with 14 and dimethylamino derivative TriMeO-N,N-BQDMPHEN (N,N-bis(5,6,7-trimethoxy-2-quinolylmethyl)-N′,N′-dimethyl-1,2-phenylenediamine (15), Fig. 7) with Cd²⁺ and Zn²⁺. Not only the moderate metal binding affinity but also the appropriate size and coordination environment of the metal binding cavity of the methoxy derivative 14 in the presence of chloride ions are necessary for fluorescence discrimination of Cd²⁺ from Zn²⁺.

Fig. 9 Fluorescence intensity plot of (a) 14 and (b) 15 in the presence of Cd²⁺ (blue) or Zn²⁺ (red) with increasing amount of KCl. Adapted from ref. 30 with permission from American Chemical Society, copyright 2024.

6. Conclusions

Based on the structure of BQDMEN (1), a set of fluorescent Cd²⁺ sensors were rationally designed (Fig. 10). The extensive use of polymethoxy-substituted quinolines as a metal binding motif and chromophore provides several unique characteristics suitable for strict discrimination of Cd²⁺ from Zn²⁺via fluorescence signals. Unprecedented fluorescence enhancement mechanisms including intramolecular excimer emission from the bis(μ-chloro) dinuclear cadmium complex were clarified by X-ray crystallography. Tiny changes in the molecular structure of the fluorescent probe largely affect the structure and stability of the resulting metal complexes, affording significant improvement in fluorescent metal ion selectivity and optical properties. In particular, the size and flexibility of the metal binding pocket with a carefully organized coordination geometry are crucial for selective accommodation of the target metal ion. This Frontier article may provide a useful dataset for future molecular design of new fluorescent sensors targeting a wide variety of metal ions.

Fig. 10 Summary of the structural modifications of Cd²⁺ probes.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The author expresses sincere thanks to his students and collaborators for their valuable help in the exploration of BQDMEN-based ligands. This work was also supported by the JSPS KAKENHI Grant Number JP23K04808 and the Nara Women's University Intramural Grant for Project Research.

References

1. X.-L. Tang, X.-H. Peng, W. Dou, J. Mao, J.-R. Zheng, W.-W. Qin, W.-S. Liu, J. Chang and X.-J. Yao, Org. Lett., 2008, 10, 3653–3656.
2. L. Xue, G. Li, Q. Liu, H. Wang, C. Liu, X. Ding, S. He and H. Jiang, Inorg. Chem., 2011, 50, 3680–3690.
3. M. Mameli, M. C. Aragoni, M. Arca, C. Caltagirone, F. Demartin, G. Farruggia, G. De Filippo, F. A. Devillanova, A. Garau, F. Isaia, V. Lippolis, S. Murgia, L. Prodi, A. Pintus and N. Zaccheroni, Chem. – Eur. J., 2010, 16, 919–930.
4. X. Zhou, P. Li, Z. Shi, X. Tang, C. Chen and W. Liu, Inorg. Chem., 2012, 51, 9226–9231.
5. H. Tian, B. Li, J. Zhu, H. Wang, Y. Li, J. Xu, J. Wang, W. Wang, Z. Sun, W. Liu, X. Huang, X. Yan, Q. Wang, X. Yao and Y. Tang, Dalton Trans., 2012, 41, 2060–2065.
6. Y. Li, H. Chong, X. Meng, S. Wang, M. Zhu and Q. Guo, Dalton Trans., 2012, 41, 6189–6194.
7. Y. Ma, F. Wang, S. Kambam and X. Chen, Sens. Actuators, B, 2013, 188, 1116–1122.
8. X.-J. Jiang, M. Li, H.-L. Lu, L.-H. Xu, H. Xu, S.-Q. Zang, M.-S. Tang, H.-W. Hou and T. C. W. Mak, Inorg. Chem., 2014, 53, 12665–12667.
9. E. Hrishikesan, R. Manjunath and P. Kannan, J. Solution Chem., 2016, 45, 907–919.
10. D. Udhayakumari, J. Mol. Struct., 2023, 1287, 135715.
11. Y. Mikata, A. Yamashita, A. Kawamura, H. Konno, Y. Miyamoto and S. Tamotsu, Dalton Trans., 2009, 3800–3806.
12. Y. Mikata, Dalton Trans., 2020, 49, 17494–17504.
13. B. Rieger, A. S. Abu-Surrah, R. Fawzi and M. Steiman, J. Organomet. Chem., 1995, 497, 73–79.
14. A. S. Abu-Surrah, U. Thewalt and B. Rieger, J. Organomet. Chem., 1999, 587, 58–66.
15. A. S. Abu-Sarrah, Asian J. Chem., 2002, 14, 1251–1256.
16. Y. Mikata, H. So, A. Yamashita, A. Kawamura, M. Mikuriya, K. Fukui, A. Ichimura and S. Yano, Dalton Trans., 2007, 3330–3334.
17. L. S. Morris, M. P. Girouard, M. H. Everhart, W. E. McClain, J. A. van Paridon, R. D. Pike and C. Goh, Inorg. Chim. Acta, 2014, 413, 149–159.
18. M. Saga, T. Anamushi, W. Miyahara, S. Yamazaki and K. Saito, Anal. Sci., 2015, 31, 185–189.
19. N. Singh, J. Niklas, O. Poluektov, K. M. Van Heuvelen and A. Mukherjee, Inorg. Chim. Acta, 2017, 455, 221–230.
20. R. V. Ottenbacher, A. G. Medvedev, A. A. Nefedov and K. P. Bryliakov, Inorg. Chem. Commun., 2023, 156, 111282.
21. A. P. de Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev., 1997, 97, 1515–1566.
22. B. Daly, J. Ling and A. P. de Silva, Chem. Soc. Rev., 2015, 44, 4203–4211.
23. D. Escudero, Acc. Chem. Res., 2016, 49, 1816–1824.
24. Y. Mikata, M. Wakamatsu, A. Kawamura, N. Yamanaka, S. Yano, A. Odani, K. Morihiro and S. Tamotsu, Inorg. Chem., 2006, 45, 9262–9268.
25. Y. Mikata, M. Tanaka, S. Yasuda, A. Tsuruta, T. Hagiwara, H. Konno and T. Matsuo, Dalton Trans., 2023, 52, 7411–7420.
26. Y. Mikata, S. Yasuda, T. Hagiwara, H. Konno and S. Shoji, Inorg. Chim. Acta, 2024, 571, 122218.
27. Y. Mikata, K. Kawakami, M. Nagaoka, S. Shoji, H. Konno and T. Matsuo, Inorg. Chim. Acta, 2024, 565, 121968.
28. Y. Mikata, A. Kizu and H. Konno, Dalton Trans., 2015, 44, 104–109.
29. Y. Mikata, A. Tsuruta, H. Koike, S. Shoji and H. Konno, Molecules, 2024, 29, 369.
30. Y. Mikata, N. Tosaka, S. Yasuda, Y. Sakurai, S. Shoji, H. Konno and T. Matsuo, Inorg. Chem., 2024, 63, 8026–8037.

---

Footnotes

† The DMF-H₂O (1 : 1) was used as a solvent for spectroscopic measurements in the first half (Sections 2 and 3) of this article.

‡ The DMF-HEPES buffer (1 : 1, 50 mM HEPES, 0.1 M KCl, pH = 7.5) was used as a solvent for spectroscopic measurements in the second half (Sections 4–6) of this article, considering the pH change in the presence of large amount of metal salts.

---