Introduction

Fluorescent glycoligands (FGs)1 are recently defined structural frameworks for the development of ion probes10,40. These compounds are structured by a sugar as the template upon which to install a diverse range of Lewis bases (as the metal chelation site) and fluorophores (as the optical reporter) by chemical modifications5,8,10,28,29. Employment of sugars, a cheap natural starting material, as the central platform lies on their structural diversity and rigidity1,2,3,4,5,6,7,8,9,10,11, high biocompatibility12,13 and good water solubility14,15,16,17.

The copper-catalyzed azide-alkyne 1,3-dipolar cycloaddition reaction (Cue-AAC)18,19,20, a prototype of the click chemistry21, has found wide applications in the construction of chemo-probes since the 1,4-disubstituted 1,2,3-triazole it forms represents a versatile ion coordination site8,10,22,23,24,41. We recently reported that Cue-AAC is also a promisingly suitable tool for fabrication of FGs8,15,24,25,26,27. By a twofold Cue-AAC between two azido fluorophores and a di-alkynyl glycoside, the resulting bis-triazolyl FGs exhibited distinct ion sensibilities in terms of the different fluorophores introduced10,24,25,26. Meanwhile, the epimeric identity of the glycosyl scaffold also impacts the ion coordination properties of the ligands15,27,28,29.

Owing to its potential toxicity but wide-range utility, much attention has been paid to the monitoring of silver (I) ions. In recent years, some fluorescent probes for Ag+ have been developed. However, due to the heavy metal effect of the ion, the majority of the probes suffered from FL quenching which is suboptimal for detection of biological samples30. Furthermore, some FL ‘turn-on’ probes reported are easily interfered with a competing metal or have limited solubility in aqueous media31,32,33,34.

Here, we unravel an interesting discovery that by appending coumarins to different substitution positions of a glucoside via the Cue-AAC, the fluorescence (FL) change of the resulting FGs is totally converse upon coordination with silver (I). A ‘turn-on’ probe that shows excellent sensitivity and selectivity for silver (I) in both aqueous and buffer media was determined to be of low toxicity and of applicability in FL imaging of silver ions internalized by cells. The ligand-ion complexation modes and the mechanism of the reversed FL alternation were proposed.

Results

As shown in Fig. 1, we have determined in a previous study that a C3,4-disubstituted bis-triazolyl coumarin glucoside (DT3) showed selective and remarkable FL quenching upon coordination with Ag+ in water15. The present study initiated with the interrogation of the metal-ion sensibility of two other structurally analogous coumarin glucosides, DT1 and DT2, synthesized by a microwave-assisted twofold Cue-AAC35. The only structural divergence is that two same triazolyl coumarin moieties were substituted on the different positions of the glucosyl platform to display these functional groups in diversely constrained manners5,8,10,28,29.

Figure 1
figure 1

Structure of the bis-triazolyl coumarin glucosides DT1, DT2 and DT3.

Full size image

FL spectroscopy was used to primarily test the FL change of the FGs in the presence of a range of metal cations. The FL intensity of both DT1 (Fig. 2a) and DT2 (Fig. 2b) enhanced evidently only in the presence of Ag+ in MeOH in a concentration-dependent manner. The quantum yields of DT1 and DT2 in MeOH were determined to be 0.06 and 0.12, respectively (reference compound: 9,10-diphenylanthracene). A red-shifted shoulder was observed while the ion solution (pre-dissolved in water) was added to the probe solution (Fig. 2c for DT1 and Fig. 2d for DT2), which might be caused by the presence of water that increases the polarity of the system. To test this, the FL of the more sensitive DT1 alone was measured in a series of premixed aqueous solvents (MeOH/H2O). As shown in Fig. S2, increasing the water content of the system resulted in gradual decrease of the original emission peak (λmax = 425 nm) and increase of the red-shifted peak (λmax = 475 nm); the new peak was found to predominate in a highly aqueous medium (H2O/MeOH = 4:1, V/V).

Figure 2
figure 2

FL change of 10 μM of (a) DT1 and (b) DT2 in the absence and presence of various metal cations (100 μM) in MeOH. FL titration of 10 μM of (c) DT1 and (d) DT2 in the presence of increasing Ag+ (0 to 55 μM for DT1 and 0 to 50 μM for DT2) in MeOH. (e) FL titration of 10 μM of DT1 in the presence of increasing Ag+ (0 to 84 μM) in H2O/MeOH = 4:1 (V/V). (f) FL change of 10 μM of DT1 in the absence and presence of various metal cations (100 μM) in H2O/MeOH = 4:1 (V/V). For all FL spectra, λex = 350 nm; for the original FL spectra of (a), (b) and (f) in the absence or presence of the cations, see Fig. S1a, Fig. S1b and Fig. S3a, respectively.

Full size image

With this aqueous medium we further measured the Ag+-sensing property of DT1. To our delight, results showed that the FL enhancement of the FG is similarly dependent on Ag+ concentration with a satisfactory linear range from 0 to 20 μM (Fig. S3b). The limit of detection of the probe was determined to be 1.7 μM (3σ/k) and the probe also showed excellent metal cation selectivity in this 80% aqueous solution (Fig. 2f). Further addition of increasing I− to the DT1-Ag+ complex recovered gradually the FL of the ligand, suggesting that the complexation is reversible (Fig. S3c and Fig. S3d). Additionally, co-existence of a series of competing ions did not impact the sensitivity of the probe for Ag+ (Fig. S3e).

Interestingly, we have determined previously that, in the presence of the same ion (Ag+), the FL of the C3,4-disubstituted coumarin FG, DT3 (Fig. 1), quenched sharply in water15. Since the only structural distinction between DT1/DT2 and DT3 lies in the substitution pattern of the bis-triazolyl coumarin moieties upon the glucosyl platform, we deduced that they probably adopt different complexation modes with the ion leading to the reversed FL variations observed.

To elaborate this intriguing observation, a series of additional spectroscopic analyses were performed. First, by measuring the response of DT1 and DT3 towards Ag+ in a uniform solvent system (10 μM each in H2O/MeOH = 4:1, V/V), we substantiated that their FL indeed changed conversely (Fig. 3a and Fig. 3b). The disassociation constants of DT1 and DT3 with Ag+ in this system were measured to be 2.2 × 10−4 M−1 and 5.2 × 10−5 M−1, respectively. The quantum yields of DT1 and DT3 in water (reference compound: 9,10-diphenylanthracene) were determined to be 0.04 and 0.40, respectively. After addition of 100 μM of Ag+, the quantum yield of DT1 increased (0.10) while that of DT3 decreased (0.23). A Job plot analysis suggested that the C2,3-substituted DT1 forms a 2:1 complex with Ag+ (Fig. 3c), whereas complexation of the C3,4-substituted DT3 follows a 1:1 stoichiometry (Fig. 3d).

Figure 3
figure 3

FL titration of 10 μM of (a) DT1 and (b) DT3 in the presence of increasing Ag+ (0 to 84 μM) in H2O/MeOH = 4:1 (V/V). Job plot of (c) DT1 and (d) DT3 in complexation with Ag+ (y = F − F0(1 − x), where F0 is the original FL intensity of probe and F that upon addition of Ag+). For all FL spectra, λex = 350 nm.

Full size image

Next, we resorted to 1H NMR titration of the two FGs for gaining a better understanding of the coordination modes. We used DT1′ and DT3′ (Fig. S5) which are the protected forms of DT1 and DT3, respectively, because of their better spectral resolution. Since DMSO-d6 was used as the deuterated solvent, we preliminarily confirmed that the trend in FL increase and decrease of DT1′ and DT3′ in the presence of Ag+ in DMSO (Fig. S4) accords with those of DT1 and DT3 in H2O/MeOH, respectively.

Fig. S5 displays the partial (aromatic part) spectral change of both glycoligands in the presence of increasing Ag+. From the spectra we noted that: 1) Complexation of DT1′-Ag+ and DT2′-Ag+ follows a 2:1 and 1:1 stoichiometry, respectively, which is in agreement with the result yielded by the Job plot analysis; 2) All aromatic protons (triazole-H and coumarin-H) of DT1′ shifted, whereas only the triazole protons and two pairs of coumarin protons adjacent to the triazole of DT3′ shifted. The second observation suggests that while both triazole and coumarin moieties of DT1′ participated in the ion coordination, probably only the N-atoms of triazole groups of DT3′ chelated a silver ion.

A quantum chemical calculation was further conducted to predict the binding motif of DT1 and DT3 with silver (I). After optimizations, it was simulated that, of the three possible binding motifs (Fig. 4, the upper diagram and Fig. S6), the energy of a DT3-Ag+ motif, where the two C3-triazolyl coumarin arms of two individual DT1 molecules together chelate an ion (Fig. 4), was the lowest. By contrast, a 1:1 DT3-Ag+ motif was simulated to be much less stable than the 2:1 counterpart (Fig. S7), which suggests that latter is optimal. The optimized motif of one molecule of DT3 complexed with silver (I) illustrated that only the triazole groups of the ligand are involved in chelation (Fig. 4, the lower diagram).

Figure 4
figure 4

The computationally simulated optimal complexation mode of DT1 (upper, the distance between triazole-N2 and coumarin-carbonyl-O1 and Ag+ is 2.4 Å and 2.5 Å, respectively) and DT3 (lower) with a silver ion (the green balls stand for carbon atoms, blue for nitrogen, red for oxygen, white for hydrogen and grey for silver).

Full size image

Based on the above data, we proposed a plausible explanation with respect to the FL changes of the FGs in the presence of Ag+ (Fig. 5). For DT1, two triazolyl coumarin arms that belong separately to two FG molecules coordinate with one silver ion through both the carbonyl groups of coumarin and nitrogen atoms of triazole, leading to a CHEF (chelation-enhanced fluorescence)-like mechanism. This is similar to a chelation mode recently described between a triazolyl coumarin-based chemo-probe and a heavy metal36, as well as to some other simulated motifs including constrained glycoligands in complex with heavy metals37,38. In contrast, as only the triazole groups are in coordination with silver, the FL quenching of DT3 could be possibly ascribed to a heavy metal effect15,39.

Figure 5
figure 5

The proposed coordination modes and mechanisms of the FL alternation of DT1 and DT3 in complexation with Ag+.

Full size image

To test the practicality of the FGs, a cell imaging assay was eventually performed with the ‘turn-on’ probe DT1 (Fig. 6). We first tested the sensitivity of the probe for Ag+ in HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid] buffer (pH 7.3) that will be used for the cellular assay. As shown in Fig. 6a, the FL of DT1 gradually enhanced with increasing Ag+ in the buffer, corroborating its good water solubility and potential utility for FL imaging in live cells under physiological conditions. An MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] cell-viability assay (3-day incubation of the probe at varying concentrations with the cells) then revealed that the probe was not toxic to Hep-G2 cells (human hepatoma) even at a relatively high concentration (500 μM, 50-fold the concentration used for sensing the ion in solution, Fig. 6b).

Figure 6
figure 6

(a) FL titration of 10 μM of DT1 in the absence (the bottom curve) and presence of increasing Ag+ (0–42 μM) in HEPES buffer (95% HEPES mixed with 5% DMSO, pH 7.3). (b) Cell viability of Hep-G2 in the presence of a control compound (doxorubicin) and DT1 with increasing concentrations. FL images of Hep-G2 cells in the presence of (c) 10 μM, (d) 20 μM and (e) 50 μM of DT1. FL images of Hep-G2 cells preincubated with 20 mM of Ag+ in the presence of (f) 10 μM, (g) 20 μM and (h) 50 μM of DT1.

Full size image

Next, by incubation of the probe alone with Hep-G2 cells at three different concentrations, only very weak FL was recorded (Fig. 6c, Fig. 6d and Fig. 6e). However, incubation of the cells with Ag+ prior to loading of the probe led to emergence of clearly intensified FL. This implies that DT1 could probably chelate silver ions internalized by the cells, thereby producing the enhanced FL (Fig. 6f, Fig. 6g and Fig. 6h). These results together support the promise of the coumarin-based FGs in monitoring silver ions in live cells.

Discussion

We unravelled with this research an unprecedented discovery that different substitution patterns of triazolyl coumarins upon an identical glucosyl platform could produce FGs with totally reversed optical response to a same heavy metal ion in an aqueous solution. By a series of analyses we determined that this interesting divergence was probably caused by the distinct coordination mode of the conformationally constrained glycoligands with the ion. A ‘turn-on’ C2,3-substituted FG that exhibited good water solubility and low cytotoxicity has proved suitable for imaging Ag+ in live cells. This study thereby paves the way for the design and development of sugar-based fluorescent chemo-probes with tuneable FL owing to the conformational constraint of fluorophore-receptor moieties modified on a rigid glycosyl platform.

Methods

General

All chemicals and regents are of high commercially available grade and were used as received.1H NMR spectra were recorded on a Bruker AM-400 spectrometer using tetramethyl silane (TMS) as the internal standard (δ = 0). LC-MS was performed on a Waters ACQUITY UPLC™ system with a Quattro Mirco MS (triple quadrupole MS).

Fluorescence spectroscopy

The fluorescence measurements were carried out on a Varian Cary Eclipse Fluorescence spectrophotometer by using a path length of 10 mm with excitation at 350 nm by scanning the emission spectra between 360 nm and 600 nm. The bandwidth for both excitation and emission spectra was 5 nm. All cations tested are perchlorate salts prepared in a stock solution of 10 mM in H2O and were diluted to the indicated concentrations for testing.

MTS cell viability assay

Hep-G2 cells were plated overnight on 96-well plates at 5000 cells per well in growth medium. After seeding, cells were maintained in growth media treated at increasing concentrations (5.12 μM, 12.8 μM, 32 μM, 80 μM, 200 μM and 500 μM) of DT1 (dissolved in DMSO, final concentration) for 72 h. 20 μL of MTS (Promega Corp) solution (2 mg/mL) was added to each well for 2 h at 37°C and then the absorbance was measured on a SpectraMax 340 microplate reader (Molecular Devices, USA) at 490 nm with a reference at 690 nm. The optical density of the result in MTS assay was directly proportional to the number of viable cells. Each experiment was done in triplicate.

Cell imaging

Hep-G2 cells were cultured in DMEM supplemented with 10% FBS. Cells (1.5 × 104/well) were seeded on a black 96-well microplate with optically clear bottom (Greiner bio-one, Germany) overnight. After pretreatment with 20 mM AgNO3 in 50 mM HEPES for 30 min, the cells were incubated with the probe in 50 mM HEPES at different concentrations for another 30 min. Then the cells on the microplate were rinsed in warm HEPES and fixed by 4% paraformaldehyde in HEPES for 15 min at room temperature. After rinsing in HEPES three times (5 min each time), the fluorescence was eventually detected and photographed with an Operetta high content imaging system (Perkinelmer, US).

Quantum chemical calculations

All the quantum chemical calculations were performed with Gaussian 09 software. The original geometries of all molecules were drawn using GaussView v5.08 program, which were further optimized by a density functional theory (DFT) with Becke's three-parameter hybrid exchange functional and the Lee-Yang-Parr correlation functional (B3LYP). In all calculations, a combined basis set was employed with LanL2DZ for silver and 6-31G(d) for other atoms.