Discrimination between Hard Metals with Soft Ligand Donor Atoms: An On-Fluorescence Probe for Manganese(II)

Manganese is an essential metal in all forms of life.[1] It participates as a co-factor in diverse classes of enzymes and the photosynthetic machinery[2] and is used widely as a versatile tool for biological studies. For example, high spin Mn²⁺ is an excellent MRI relaxation agent that has been used in clinical diagnosis and is of widespread interest as a tool in neurobiological research.[3] However, chronic over-exposure can result in movement disorders and mental disturbances and other brain-related toxicities.[4] Fluorescent probes would be useful for detection and quantification of Mn²⁺, as this method offers high efficiency, high sensitivity and easy operation[5] among available methods of detection.[6] However, development of an effective fluorescent probe for Mn²⁺ faces several challenges: 1) Unlike diamagnetic metal ions such as Zn²⁺, paramagnetic Mn²⁺ can quench fluorescence. Although chelation-induced fluorescence quenching (CHEQ) is the most commonly used method of paramagnetic metal ion detection,[7] “on-fluorescence” probes for Mn²⁺ are preferred. 2) Mn²⁺ selectivity over abundant cellular metal ions is required, especially Ca²⁺ (up to high μM).[8] Mn²⁺ and Ca²⁺ share many common properties, underscored by the fact that Mn²⁺ can enter cells using some of the same transport systems as Ca²⁺.[9] 3) Mn²⁺ probes must be compatible with biological environments, including water solubility, biological inertness, long-wavelength excitation and emission profiles to minimize sample damage and native cellular autofluorescence.[10] 4) To visualize Mn²⁺ in living cells or tissues, membrane permeability is important.[10]

Several commercially available chelating dyes produce strong fluorescence enhancement upon binding Mn²⁺.[11] However, the fluorescence of available dyes such as calcium green is also enhanced in the presence of Ca²⁺. BAPTA (1,2-bis(o-aminophenoxy)ethane-N,N,N,N-tetraacetic acid) is a known Ca²⁺-selective ligand that serves as the chelating moiety of calcium green.[12] We undertook to modify the BAPTA unit in such a way as to achieve adequate Mn/Ca selectivity.

Optimizing stereoelectronic complementarity between host and guest to achieve efficient complexation is a long-held axiom of supramolecular and coordination chemistry. However, while optimizing a receptor for a substrate leads to strong binding, it may not result in good selectivity over competing substrates. In the case of metal ion complexation, selection of ligand donor atom can sometimes be used to advantage. For example, choosing soft donor atoms may improve selectivity for relatively soft metal ions in competition with hard ones. Both Mn²⁺ and Ca²⁺, however, are generally considered to be hard and show maximal stability with hard oxygen donors. Thus, superficial considerations would lead away from hard/soft donor atom considerations as a strategy for achieving Mn²⁺ selectivity.

However, although both Mn²⁺ and Ca²⁺ are classified as “hard” metals and therefore form stronger complexes with oxygen donors, Mn²⁺ appears to be more tolerant of softer atom donors than Ca²⁺.[13,14] More recently, the relative softness of Mn²⁺ compared with Mg²⁺ has been debated as the basis for Mn²⁺ rescue of activity in dialkyl-thiophosphate RNAzymes[15-17]. Our hypothesis was thus to replace two or more carboxylate groups in BAPTA with softer ligating moieties. Since Mn²⁺ is a hard metal ion, weaker binding might be expected from such a change, but since Ca²⁺ is an even harder metal ion, the effect on Ca²⁺ should be more pronounced, resulting in a net increase in selectivity. We chose nitrogen atom donors from pyridine, which is a common binding group in transition metal ion ligands considered to be borderline but softer than oxygen.[14] To evaluate the feasibility of our strategy, a prototype ligand 1 was synthesized, which has one carboxylate group of each dicarboxymethylamino moiety of BAPTA replaced by a pyridine (Scheme 1). The chemical synthesis of ligands 1-3 is included as supporting information.

Scheme 1.

Structures of ligands and probes.

UV titrations were carried out by addition of MnCl₂ to MOPS buffered aqueous solution (pH=7.2) of 1 (SI). In the absence of Mn²⁺, the spectrum of 1 showed a maximum at 256 nm with a shoulder at 286 nm, similar to BAPTA. Mn²⁺ complexation caused significant hypsochromic shifts towards a limiting spectrum with a small maximum at 278 nm surrounded by shoulders. Absorbance at 256 nm was plotted as a function of Mn²⁺ concentration and the minimum level of absorbance was reached upon addition of 1 equivalent of Mn²⁺, suggesting 1:1 metal-ligand complex. The same analysis was applied to determine 1:1 complexation of BAPTA to Mn²⁺. The binding constants were obtained by titration in pH- and Mn²⁺- buffered aqueous media. The plot of absorbance as a function of free Mn²⁺ produced a sigmoidal curve. Nonlinear fitting analysis[18] gave association constants (logK) of 8.62 for ligand 1 and 9.14 for BAPTA (Table 1). These results indicate that substitution of two carboxylate groups of BAPTA with two pyridines impairs binding affinity to Mn²⁺ to a surprisingly small extent.

Table 1.

Mn²⁺ and Ca²⁺ affinities and spectroscopic properties of ligands.

	logK				selectivity log(KMn/KCa)	λ max				ϕ f
ligands	Mn2+	Ca2+	Mg2+	Zn2+		Free ligand	Mn2+ complex	Ex	Em	Free ligand	Mn2+ Complex
BAPTA	9.14	6.89	1.69	9.41	2.25	290(4.8)[a]	278 (4.5)	--	--	--	--
1	8.62	3.79	0.76	9.19	4.83	286 (6.3)	283 (4.1)	--	--	--	--
2	7.01	2.96	0.42	6.16	4.05	493 (84)	493 (100)	493	519	0.10	0.37
3	8.00	3.33	0.93	7.09	4.68	507 (80)	507 (92)	505	530	0.13	0.49

^[a]In parentheses is the corresponding molar absorption coefficient ε in units of 10³ L mol⁻¹ cm⁻¹.

Calcium binding to ligand 1 was investigated in a similar manner. A hypsochromic shift was again observed, but even excess Ca²⁺ did not saturate the UV absorption indicating weak affinity. The association constant of 1-Ca²⁺ was determined to be log(K_a) = 3.79, about 4.83 log units weaker than for Mn²⁺. The log(K_a) of BAPTA-Ca²⁺ was 6.89,[11] only 2.25 log units lower than that of BAPTA-Mn²⁺ complex. Therefore, ligand 1 indeed shows much higher selectivity for Mn²⁺ over Ca²⁺ as compared to BAPTA. The significant improvement in Mn²⁺/Ca²⁺ selectivity validates our “selective poisoning” strategy.

To realize the goal of a fluorescent Mn²⁺ probe, compound 1 was further functionalized to include a chromophore similar to that present in calcium green. Thus, amino groups installed para-to the N-atoms of both aniline moieties were coupled with fluorescein 5-isothiocyanate in high yield followed by hydrolysis afforded fluorescent probe 2. Probe 3, most similar in structure to calcium green-2, was prepared containing chloro-substituents and an amide linkage between chelating unit and fluorophore.

The binding properties of fluorescent probe 2 were first investigated by titration with Mn²⁺ (Figure 1a). When excited at 493 nm, 2 showed an emission maximum at 519 nm. Upon addition of Mn²⁺, an enhanced fluorescence was observed until saturation after one equivalent. The quantum yields of the free probe and Mn²⁺-bound complex were determined to be 0.10 and 0.37, respectively. The association constant with Mn²⁺ was logK = 7.01. Similar to the prototype ligand 1, only a large excess of Ca²⁺ ion (mM) caused fluorescence enhancement indicating weak association of 2 to Ca²⁺ (logK = 2.96). In the cellular environment, calcium concentration is generally lower than 100 μM. Therefore, probe 2 has the potential to detect Mn²⁺ in the presence of calcium ion interference in biological systems. Screening for selectivity against other metal ions (SI) Na⁺, Mg²⁺, Ba²⁺ and K⁺ showed no effect on fluorescence intensity of 2. However, transition metal ions, Ni²⁺ and Cu²⁺ quench the fluorescence, while Fe²⁺, Co²⁺, Zn²⁺, Cd²⁺ and Hg²⁺ may interfere with Mn²⁺ binding. Among these, Zn²⁺ (Table 1) is probably the most significant concern for likely applications. Dissociation constants determined for Co²⁺, Ni²⁺, and Cd²⁺ for BAPTA (9.13, 10.51, 13.38) and 1 (9.22, 10.23, 13.58) together with data for Mn²⁺ and Zn²⁺ (Table 1) indicate that the two ligands follow the Irving-Williams series very similarly, and that the discrimination against Ca²⁺ is much greater than the effect observed on soft metals.

Figure 1.

Fluorescence titration of 2 (5 μM, λ_exc=493 nm) with MnCl₂ and CaCl₂ (50 mM HEPES, 0.1 M KNO₃, pH 7.2): a) 0 to 200 μM Mn²⁺; b) 0 to 10 mM Ca²⁺. Inserts: λ_em=519 nm.

Compound 3 showed a longer λ_exc (505 nm) and λ_em (530 nm) due to the two incorporated chlorine atoms on the fluorophore (SI). The Mn²⁺ complex showed enhanced fluorescence. Probe 3 maintained high selectivity for Mn²⁺ over Ca²⁺ with logK of 8.00 for Mn²⁺ and 3.33 for Ca²⁺. The quantum yield of free probe 3 was 0.13, while that for the Mn²⁺ complex was 0.49. Probe 3 shared similar binding properties to probe 2 to other metal ions (SI). Probe 2 was responsive over physiologically relevant pH 6.8 to 8.2. Interestingly, probe 3 functions well in basic conditions: From pH 8.3 to pH 12.2, the high pH limit of our measurement, the fluorescence showed nearly full response. To our knowledge, probes 2 and 3 are the first selective “on-fluorescence” Mn²⁺ probes reported.

To confirm that the probes would function in a biological environment, intracellular Mn²⁺-sensing was performed with the Hela cell line. For cell membrane permeability, the ester precursor of 3 (SI) was used for in vitro Mn²⁺ detection, as it has been reported that permeable ester probes can be hydrolyzed in the intracellular environment.[19] Hela cells were first incubated with Mn²⁺, followed by fluorescent probe treatment. A 2.4-fold enhanced fluorescent signal was observed in the presence of the probe and Mn²⁺ (Fig. 2), consistent with the measurements performed in solution.

Figure 2.

Detection of Mn²⁺ in Hela cells. a) 5 μM 3 ethyl ester; b) 200 μM added MnCl₂; c) Mn²⁺ supplemented cells treated with 2 mM TPEN for 5 min at room temperature.

In conclusion, ligand 1 was rationally designed from BAPTA using a “soft atom poisoning” strategy to differentiate binding affinities to Mn²⁺ and Ca²⁺. Binding preferences were tuned by substitution of carboxylate groups of BAPTA with pyridines, resulting in much stronger Mn²⁺ selectivity over Ca²⁺. Fluorescent probes based on ligand 1 were synthesized. Solution and in vitro properties demonstrated that these sensing compounds have good selectivity towards Mn²⁺ with “on” fluorescence response.