Förster Resonance Energy-Transfer-Based Ratiometric Fluorescent Indicator for Quantifying Fluoride Ion in Water and Toothpaste.

A Förster resonance energy-transfer (FRET)-based ratiometric fluorescent indicator Cou-FITC-Si toward fluoride ion has been designed and synthesized by combining coumarin unit and fluorescein derivative as energy donor and acceptor, respectively. The fluorescein unit is capped with tert-butyldiphenylchlorosilane. The indicator gives out emission responses based on switch-on of the FRET process that triggered by the desilylation mediated by the fluoride ion. The fluorescence emission spectrum of Cou-FITC-Si presents a significant bathochromic shift of 59 nm after the addition of fluoride ion with up to 180-fold increase of the fluorescence intensity ratio. The limit of detection of the Cou-FITC-Si indicator system toward fluoride ion was estimated to be 3.3 ppb. Furthermore, this indicator has been successfully applied for quantifying the fluoride ion of different concentrations from commercially available toothpaste.

Introduction

Numerous anions and cations are realized to play indispensable roles in human life, and considerable attention has been drawn to developing detecting systems toward them.[1] Among these anions and cations, fluoride ion is one of the most important anions because of its significance for health and environmental issues. Appropriate ingestion of fluoride is beneficial to our dental and skeletal health, but excessive intake of fluoride may cause fluorosis, nephrolithiasis, or even osteosarcoma.[2−4] The World Health Organization (WHO), European Union (EU), United States, China, and many other countries have established standards for fluoride in drinking water. The primary standard of 4 mg/L and secondary standard of 2 mg/L of fluoride in drinking water have been given to prevent osteofluorosis and to protect against dental fluorosis, respectively, by the Environmental Protection Agency (EPA) of United States. In addition, a recommended range of 0.7–1.2 mg/L of fluoride in drinking water has been proposed by the U.S. Department of Health and Human Services (HHS) which has been adjusted based on the most recent scientific data in 2011.[5] Accurate measurement of the levels of fluoride in water, environment, and biosystems is essential for scientific and healthy benefit. Although the ion-selective electrode, ion chromatography, and Willard–Winter methods are commonly applied for quantitative fluoride measurements,[6−8] however, some disadvantages such as high cost, complicated procedures, and low mobility still remain. Therefore, it is appealing to develop convenient and inexpensive methods for highly selective and sensitive detection of fluoride in aqueous environments.

Fluorescent chemosensors have been shown advantages for the detection of fluoride over the past decade because of their high sensitivity and specificity, fast response, and potential for real-time monitoring.[9] Various fluorescent chemosensors of fluoride based on supramolecular interactions such as anion−π interactions, hydrogen bonding, and Lewis acid coordination have been reported,[10−18] but a lot of them are not suitable for the detection of fluoride in aqueous solutions, which limits the application in aqueous medium or biological systems. To improve the performance of fluoride sensors, a strategy based on the fluoride-triggered desilylation has been developed for recognition of fluoride ion.[19] On the basis of this strategy, several fluoride chemosensors have been reported and applied successfully for the detection of fluoride ion in water and biologic targets.[20−22] However, most of them are intensity-based sensors, which is significantly affected by various external factors including microenvironment (pH value and polarity), sample thickness, excitation power, detector response, and so forth. In contrast, chemosensors based on ratiometric fluorescence can overcome the shortcomings of fluorescence intensity-based ones because they offer an internal calibration by recording the ratios of emission intensities of two different bands. Furthermore, the ratiometric fluorescence provides direct fluorescence color variations other than intensity, which is more convenient to distinguish F– from other anions by naked eyes.[23−26] Förster resonance energy transfer (FRET) is one of the most commonly applied mechanisms for ratiometric fluorescent chemosensors. The FRET process can be tuned or ceased by regulating the absorption of acceptor or/and enlarging the separation of donor and acceptor, resulting in emission spectra change and consequent reporting. However, FRET-based fluoride sensing is challenging because of the limitation of matching FRET chromophore pairs and fluoride-triggered spectra changes, and few examples were achieved.[27] In this work, a fluoride indicator (Cou-FITC-Si) with high sensitivity and selectivity is designed and created by taking advantages of ratiometric fluorescence provided by FRET and the selectivity of fluoride-triggered desilylation. A coumarin chromophore and a silyl-protected fluorescein derivative of spirolactam form are selected as the energy donor and the acceptor, respectively. They are linked covalently with a short linker to ensure that the donor and the acceptor are close enough for effective FRET. The coumarin unit and fluorescein chromophore is an elegant chromophore pair with FRET-on/off functions.[28] FRET is inefficient between the coumarin unit and the fluorescein of spirolactam form with silyl capped because of the negligible spectral overlap between the donor emission and the acceptor absorption. In the presence of fluoride, the fluorescein spirolactam unit proceeds desilylation and consequently transforms into a ring-open structure (fluorescein), leading to a large bathochromic shift of its absorption. The intramolecular FRET process is then activated between the coumarin and the ring-open structure of fluorescein (Scheme ). Therefore, the sensing system exhibits a dual fluorescence feature based on the fluoride-induced FRET, providing a measurement of fluoride with a ratiometric fluorescence. The indicator has been successfully applied in quantitative detection of fluoride in aqueous phase and toothpastes.

Scheme 1

Structure and Proposed Sensing Mechanism of Cou-FITC-Si for the Detection of F–

Results and Discussion

Cou-FITC-Si was obtained by an efficient reaction of the amino group of 2-(aminoethyl)-carbamothioyl]-5-aminofluorescein (1) with 7-(diethylamino)-2-oxo-2H-chromene-3-succinimidyl ester (2) and a following silylation with tert-butyldiphenylchlorosilane. The synthetic route is depicted in Scheme , and the preparation of compounds 1 and 2 was synthesized according to the reported methods.[29,30] The structure of Cou-FITC-Si was characterized by means of 1H NMR and 13C NMR spectra, matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (MS), and IR spectra. The solubility of Cou-FITC-Si in water is poor. For application of Cou-FITC-Si in detection of fluoride in aqueous environment and oral care products, cetyltrimethylammonium bromide (CTAB) was used to solubilize the hydrophobic indicator molecule in the aqueous phase.

Scheme 2

Synthetic Route of Cou-FITC-Si

The sensing ability of the Cou-FITC-Si aqueous dispersion ([Cou-FITC-Si] = 2.5 μM, [CTAB] = 30 mM) for fluoride was examined by measuring the absorption and emission spectra in the presence of different amounts of sodium fluoride (Figure ). An absorption band with maximum at 420 nm assigned to the coumarin group dominates the absorption spectrum of Cou-FITC-Si in the absence of fluoride. The addition of fluoride ion to the dispersion of Cou-FITC-Si results in the appearance of an absorption band centered at ca. 500 nm in company with the absorption of the coumarin group. The absorption band is enhanced with increasing amounts of fluoride ion and is assigned to the fluorescein chromophore produced by the desilylation and the sequent ring-open transformation. Upon selective excitation of coumarin part with 420 nm light, a fluorescence characteristic of the coumarin with maximum at 464 nm was observed in the absence of fluoride ion, indicating that no detectable FRET process from the coumarin group to the fluorescein chromophore of spirolactam form occurs, which is rationalized by the lack of spectra overlap between the donor and acceptor. When the dispersion of Cou-FITC-Si was treated with sodium fluoride, selective excitation of the coumarin group results in not only the decrease of fluorescence emission of coumarin but also the appearance of a new emission band with maximum at ca. 523 nm which is assigned to the fluorescence of the fluorescein chromophore. The observation of fluorescence change of Cou-FITC-Si is caused by the occurrence of the intramolecular FRET process from coumarin unit to the fluorescein part. The fluoride ion induces the desilylation of the spirolactam form of fluorescein, which further transforms to the fluorescein structure. Because the absorption of fluorescein overlaps with the fluorescence emission of coumarin, the intramolecular FRET process is switched on, giving the green fluorescence of fluorescein. Upon increasing the concentrations of fluoride ion, the emission of the fluorescein chromophore increases, accompanying with a gradual decrease of the coumarin fluorescence around 464 nm. The inset of Figure b presents the fluorescence color change of the Cou-FITC-Si dispersion before and after the addition of 100 μM NaF, showing the visual observation of fluoride ion in water. The fluorescence quantum yield of Cou-FITC-Si aqueous suspension in the absence of fluoride is estimated to be 0.18, by using coumarin 343 in acetonitrile with a quantum yield of 0.75 as a reference. After desilylation by fluoride, the fluorescence quantum yield of the indicator reaches 0.45.

Figure 1

(a) Absorption and (b) fluorescence spectra of Cou-FITC-Si (2.5 μM) aqueous dispersions (CTAB 30 mM) upon treatment with F– of various concentrations (0–200 μM). λex = 420 nm.

The ratio of emission intensities at 523 and 464 nm (I523/I464) was used as the indicator for ratiometric quantification of fluoride ion. The relationship between the ratios of fluorescence intensities I523/I464 and the concentrations of fluoride ion is depicted in Figure and shows two distinct linear regions (≤25 and ≥50 μM), which may be caused by the different local concentrations and reaction dynamics at low and high concentrations of fluoride ion.[25,26] The linear regions suggest that Cou-FITC-Si is a potential chemodosimeter for quantitative determination of the levels of fluoride ion in drinking water. Importantly, up to 180-fold increase of the value of I523/I464 was observed after the addition of 200 μM (3.8 mg/L) F–, indicative of high potential for sensitive F– determination. The limit of detection (LOD) was then determined according to the 3σ standard: LOD = 3σ/k, where σ is the standard deviation of the intensity ratio (I523/I464) obtained by measuring the indicator solutions in the absence of fluoride ions (σ = 0.00145 in this work) and k is the slope of linear fitting of the working cure in the low F– concentration region (0.025 μM–1, Figure ). The LOD of Cou-FITC-Si indicator system toward F– was estimated to be 3.3 ppb (∼0.17 μM), which is one of the best sensitivities reported in water and is adequately sensitive enough for fluoride ion detection in various aqueous samples. Selected fluoride probes for fluoride detection in water based on desilylation are listed in Table for comparison.[22,31−33] The probe developed in this work provides the first successful example of FRET molecular probe for fluoride ion in water. The distinct spectral shift and the large ratiometric enhancement make Cou-FITC-Si one of the most sensitive indicators for fluoride ion, which extends the design way for fluorescence sensing fluoride ion.

Figure 2

Fluorescence intensity ratio (I523/I464) of Cou-FITC-Si vs F– concentrations (0–200 μM) and the linear fits of the two regions at concentrations below 25 μM and above 50 μM, respectively.

Table 1

Fluoride Detection Performances and Selected Literature Reportsa

TBDPS: tert-butyldiphenylsilyl; TBDMS: tert-butyldimethylsilyl.

The selectivity of the sensing system was further assessed by measuring the fluorescence response of Cou-FITC-Si toward a series of anions (in the form of sodium salts), including Br–, H2PO4–, AcO–, NO3–, SO42–, Cl–, F–, HSO4–, CO32–, and HCO3–. The fluorescence spectra of the indicator dispersion were recorded after 30 min upon the addition of 40 equiv of the anions listed above. Only fluoride ion induce a remarkable red shift of the fluorescence maximum from 464 to 523 nm because of the formation of the open-ring form of fluorescein (Scheme ), giving a considerable change of the ratio of fluorescence intensities at 523 and 464 nm (I523/I464). Weak acid anions can induce slightly spectral changes, which may be attributed to the gradual hydrolysis of the siloxane unit induced by the basic species in aqueous solution with weak acid anions. The hydrolysis nature of siloxane in basic aqueous solution makes the indicator not to tolerate strong basicity environments, although the speed is much slower that the desilylation with fluoride. All other anions induce negligible changes in I523/I464, as shown in Figure . Thiourea was a common unit used for fluoride sensing especially in the organic medium. To evaluate if the thiourea connection unit is involved in the sensing process, the absorption and emission spectra of the aqueous indicator dispersion were examined after reaction with different amounts of fluoride for extended period (Figure S4). No substantial spectra difference and interference caused by the thiourea unit were found. These results indicate that the Cou-FITC-Si dispersion is highly selective for recognizing fluoride ion over other common anions, which is attributed to the trigger mechanism based on the specific reaction between fluoride and silyl.

Figure 3

(a) Emission intensity ratios (I523/I464) and (b) fluorescence spectra of Cou-FITC-Si (2.5 μM) aqueous dispersions (CTAB 30 mM) upon incubation with 40 equiv of Br–, H2PO4–, AcO–, NO3–, SO42–, Cl–, F–, HSO4–, CO32–, and HCO3– for 30 min (λex = 420 nm).

To evaluate the practical applicability of Cou-FITC-Si for quantitative determining fluoride ion from daily life samples, the content of fluoride in two representative and commercially available toothpastes (Crest and Darlie) was determined. A certain weight of the toothpastes was mixed with different volumes of deionized water; the liquid part was extracted from the suspension by centrifugation and filtration (three extracts of different concentrations for each toothpaste product). Upon addition of a certain amount of each extract to the Cou-FITC-Si dispersion, the ratiometric response of the indicator was recorded and the corresponding fluoride ion concentration was calculated according to the fitting cures in Figure . As listed in Table , the determined fluoride ion concentration was in good agreement with the expected fluoride ion concentrations with good recovery (97–100%), suggesting that Cou-FITC-Si can be used for quantifying the fluoride ion of different concentrations in many oral care products and aqueous samples.

Table 2

Analytical Results of the Determination of F– in Toothpaste Samples

samples	added [F–]/ppm	found [F–]/ppm	recovery/%
Darlie	0.51	0.51	100
	1.02	1.01	99
	2.03	1.97	97
Crest	0.52	0.52	100
	1.05	1.04	99
	2.09	2.07	99

Conclusions

In conclusion, a FRET-based ratiometric fluorescent indicator Cou-FITC-Si has been successfully developed for highly specific and sensitive detection of fluoride ion in aqueous. The indicator shows coumarin emission in the blue region without FRET occurring in the absence of fluoride ion. When fluoride ion is added, the desilylation reaction in the fluorescein part is triggered and the fluorescein of spirolactam form transforms into open-ring form, leading to bathochromic shift of the fluorescein absorption and the consequent switched-on of FRET process. As a result, the fluorescence spectra of Cou-FITC-Si presented a ratiometric fluorescence response and fluorescence color changed from blue to green can be observed by naked eyes. Cou-FITC-Si is selectively responsive toward fluoride, and the detection limit of Cou-FITC-Si dispersion toward F– is estimated to be 3.3 ppb. Practical application of the indicator for fluoride ion determination in toothpaste samples was demonstrated, and good recovery was obtained under various concentrations, showing the potential of the indicator in detecting F– in daily life and other aqueous samples.

Experimental Section

Materials

All chemicals and reagents used for reactions came from J&K Chemicals and Sigma-Aldrich without a process of distillation or purification except acetonitrile.

Instruments

1H NMR and 13C NMR spectra of intermediate and target compounds were performed on a Bruker Avance Π-400, and an internal standard was set by using tetramethylsilane. Fourier transform infrared (FT-IR) spectra were taken on a Varian Excalibur 3100 spectrometer. Mass spectra (MALDI-TOF) were recorded on a Bruker microflex mass spectrometer. UV–vis absorption and fluorescence spectra were measured on a Shimadzu UV-2550PC spectrometer and Hitachi F-4500 spectrometer, respectively.

Absorption and Emission Titrations

Cou-FITC-Si (2.5 μM) was solubilized in water with CTAB of 30 mM forming the homogeneous indicator solution. Fluoride anion was titrated into a 3 mL indicator solution by injecting certain amounts of the aqueous stock solutions of sodium fluoride. In addition, after stirring for 30 min, the absorption and emission spectra were recorded. Other anions and toothpaste samples were examined in the same way.

Synthesis and Characterization of Cou-FITC-Si

Synthesis of Compounds 1 and 2

The preparation of compounds 1 and 2 was synthesized according to the reported methods.[28,29]

Synthesis of Compound 3

Amino-functionalized FITC 1 (660 mg, 1.47 mmol) and 7-diethylaminocoumarin-3-carboxamide 2 (520 mg, 1.45 mmol) were dissolved in dimethylformamide (20 mL) in a 50 mL three-necked flask, and the reaction mixture was stirred at room temperature for 3 h under nitrogen. Then, the solvent was removed under reduced pressure, and the crude product was directly used in next step without further purification.

Compound 3 (352 mg, 0.5 mmol) and imidazole (82 mg, 1.2 mmol) were dissolved in dry CH3CN (20 mL) in a 50 mL three-necked flask. The mixture was stirred under nitrogen. Then, a solution of tert-butyldiphenoylchlorosilane (330 mg, 1.2 mmol) in dry CH3CN was injected into the above solution. The reaction mixture was stirred at room temperature for 2 h. Then, the reaction mixtures were filtered off, and the filtrate was evaporated under vacuum. After that, CH2Cl2 (20 mL) was added into the residue. The solution was washed with deionized water three times, and the organic phase was dried in anhydrous MgSO4. Finally, the solvent was removed under vacuum, and the residue was purified by silica-gel column chromatography (dichloromethane/ethyl acetate = 20:1) to afford a yellow solid (350 mg, 60%). 1H NMR (400 MHz, DMSO-d6): δ 9.97 (s, 1H, −CH2NHCSNHAr), 8.79 (s, 1H, −CONHCH2CH2NH−), 8.63 (s, 1H, ArH), 8.15 (br, 1H, −CH2NHCSNHAr), 8.10 (s, 1H, ArH), 7.60–7.66 (m, 10H, ArH), 7.42–7.49 (m, 12H, ArH), 7.10 (d, J = 8.2, 1H, ArH), 6.76 (d, J = 8.8, 1H, ArH), 6.59 (m, 3H, ArH), 6.55 (d, J = 2.0, 2H, ArH), 6.50 (dd, J = 8.8, 2.2, 2H, ArH), 3.68 (s, 2H, −CONHCH2CHNHCS−), 3.55–3.57 (m, 2H, −CONHCHCH2NHCS−), 3.44–3.49 (m, 4H, ArN(CHCH3)2), 1.13 (t, J = 7.2, 6H, ArN(CH2CH)2), 1.02 (s, 18H, −Si(CH)3). 13C NMR (100 MHz, CDCl3): δ 181.19, 162.84, 157.96, 157.63, 153.08, 152.49, 148.94, 135.52, 132.36, 131.80, 130.24, 129.15, 128.03, 116.42, 111.93, 111.89, 110.33, 108.76, 108.48, 107.58, 96.56, 83.76, 45.24, 38.92, 32.61, 31.10, 26.58, 19.58, 12.56. MS (MALDI-TOF) m/z: [M + H]+ calcd, 1169.44; found, 1169.9. FT-IR (KBr) ν (cm–1): 1140 (C=S), 1618 (C=C), 1639 (O=C).