Zhixue
Liu
,
Weilei
Zhou
,
Jingjing
Li
,
Haoyang
Zhang
,
Xianyin
Dai
,
Yaohua
Liu
and
Yu
Liu
*
State Key Laboratory of Elemento-Organic Chemistry, College of Chemistry, Nankai University, Tianjin 300071, P. R. China. E-mail: yuliu@nankai.edu.cn
First published on 22nd April 2020
A unique fluorescent supramolecular assembly was constructed using coumarin-modified β-cyclodextrin as a reversible ratiometric probe and an adamantane-modified cyclic arginine–glycine–aspartate peptide as a cancer-targeting agent via host–guest inclusion complexation. Importantly, the coumarin-modified β-cyclodextrin not only showed higher sensitivity than the parent coumarin derivatives owing to the presence of numerous hydroxyl groups on the cyclodextrin but also provided a hydrophobic cavity for encapsulation of a cancer-targeting agent. The assembly showed a reversible and fast response to biothiols with a micromolar dissociation constant, as well as outstanding cancer cell permeability, which can be used for high-efficiency real-time monitoring of biothiols in cancer cells. This supramolecular assembly strategy endows the fluorescent probe with superior performance for dynamic sensing of biothiols.
Efforts have been made to develop fluorescent probes for this purpose because they are non-invasive and simple for rapid detection.7,8 The sensing mechanisms of these probes can be classified as disulfide exchange,9 cleavage of sulfonamide and sulfonate ester,10–12 Michael addition,13,14 cyclization with aldehydes,15 thiol-halogen nucleophilic substitution,16 metal complex–displace coordination,17etc. However, for real-time observation of concentration changes of biothiols, the Michael addition–elimination mechanism is the optimum one because of its reversibility. To date, a number of researchers have focused on modifying Michael addition acceptors to obtain highly sensitive and accurate reversible probes.18,19 For example, Wang and co-workers reported the first reversible fluorescent probe with a large dissociation constant (Kd = 1.6 mM) to quantify intracellular GSH concentrations.20 Then they introduced the cyano group adjacent to the double bond to improve the sensitivity and accuracy of the probe for quantitative real-time imaging of GSH in the cytoplasm (Kd = 3.7 mM),21 mitochondria (Kd = 1.0 mM) and other organelles.22–24 Furthermore, Urano and co-workers used an improved strategy to devise two novel reversible fluorescent probes (Kd = 3 and 0.6 mM) for live-cell imaging and quantification of rapid changes in GSH concentrations.25 Yoon and co-workers also reported a cyano-modified reversible fluorescent probe for real-time quantitative monitoring of cellular GSH (Kd = 2.59 mM).26 Likewise, similar strategies were introduced for designing reversible fluorescent probes by Jiang,27 Zhang,28 Yin,29 and Guo30etc., separately. Even though the above-described fluorescent probes are revolutionary tools for exploring biothiol concentration dynamics under physiological conditions, their Kd values are in the millimolar range; therefore, an interesting approach is to develop reversible fluorescent probes with higher sensitivity (Kd at the micromolar level).
The most common method for enhancing the biothiol-detection efficiency of a fluorescent probe that acts as a Michael acceptor is to introduce an electron-withdrawing substituent adjacent to the α-carbon of the unsaturated double bond, which renders the β-carbon more susceptible to attack by the nucleophilic thiol, thereby accelerating the addition reaction.31,32 The above-described fluorescent probes were in fact modified in this way; however, methods for increasing the nucleophilicity of the thiol group have not been thoroughly explored, despite the fact that the efficiency of a Michael addition depends on the nature of both the acceptor and the nucleophile. In addition, currently available fluorescent probes could be improved by endowing them with cancer-targeting ability and better bioavailability.33 Therefore, the development of methods for enhancing the biothiol-detection efficiency of fluorescent probes and simultaneously endowing them with cancer-targeting ability is urgently needed.
One possible strategy for accomplishing these tasks is to take advantage of supramolecular macrocyclic compounds such as cyclodextrins (CDs). CDs, which are enzymatically modified forms of starch and are composed of six to eight D-glucose units linked by α-1,4-glycosidic bonds,34 can improve the solubility, stability, and bioavailability of organic and inorganic materials.35,36 Hence, CD-based catalysts,37,38 molecular-recognition systems,39,40 metal–organic frameworks,41,42 room-temperature-phosphorescent materials,43–45 artificial actuators,46,47 drug delivery and controlled-release systems,48–50 and light-harvesting systems51 have proved to be fertile research areas in chemistry and the life sciences. Under appropriate conditions, CDs and their derivatives can act as selective catalysts by means of the formation of hydrogen bonds between the hydroxyl groups of the CD and a guest molecule, which can drive a reaction in the desired direction.52 Sakuraba et al. reported that CD derivatives can accelerate aqueous Michael addition reactions involving thiols by means of host–guest interactions.53,54 In addition, Suresh and Pitchumani reported that β-CD-catalyzed Michael reactions involving thiols can take place readily at room temperature without the need for hazardous external acids and bases and that the catalyst can be reused.55 Hydrogen bonding between thiols and the hydroxyl groups of β-CD can weaken the S–H bond, thus enhancing the nucleophilicity of the sulfur atom and facilitating addition reactions with electron-deficient alkenes.56
Taken together, these previously reported results suggest that covalently linking a CD to a fluorescent probe might be a simple and effective method for improving biothiol-detection sensitivity while simultaneously endowing the probe with a cavity for host–guest complexation of a cancer-targeting agent, thus realizing a method for dynamic, targeted imaging of biothiols in cancer cells. In this study, we constructed a novel cancer-targeting fluorescent supramolecular assembly based on coumarin-modified β-cyclodextrin (rCP-βCD) as a reversible ratiometric probe and adamantane-modified cRGD (cRGD-ADA, where cRGD = cyclic arginine–glycine–aspartate motif) as a cancer-targeting agent (Scheme 1). rCP-βCD exhibited a high affinity for biothiols, with a small dissociation constant (Kd) and a short half-life (t1/2). Formation of hydrogen bonds involving the CD hydroxyl groups played a crucial role in improving the biothiol-detection sensitivity, and the hydrophobic cavity could encapsulate cRGD-ADAvia host–guest inclusion complexation to form a supramolecular conjugate denoted as rCP-βCD@cRGD-ADA. This conjugate exhibited outstanding cancer cell permeability; it showed a faster reverse reaction rate for GSH than rCP-βCD, and its biothiol-detection sensitivity was on par with that of rCP-βCD, making it useful for high-efficiency real-time monitoring of dynamic changes of biothiols in cancer cells.
Scheme 1 Construction of rCP-βCD@cRGD-ADA and the reversible Michael addition reactions between rCP-βCD@cRGD-ADA and biothiols. |
To assess the stability of rCP-βCD and its selectivity toward biothiols, we measured its UV-vis and fluorescence spectra under physiological conditions. rCP-βCD showed no fluorescence response to pH variations and exhibited excellent responsivity to biothiols at pH 5–9 (Fig. S4 and S5†). In addition, it also showed good selectivity for biothiols with an enhancement of the fluorescence ratio (F463 nm/F560 nm) of at least 2700-fold relative to that of various other intracellular amino acids, as well as a tolerance to various reactive oxygen and nitrogen species, accompanied by negligible interference of reactive sulfur (Fig. 1i, S6 and S7†), suggesting that rCP-βCD was suitable for qualitative detection of biothiols in living cells. To test the reversibility of the reaction of rCP-βCD with biothiols, we performed dilution experiments (Fig. S8†). Partial recovery of rCP-βCD was observed by UV-vis spectroscopy during 10-fold dilution experiments. This result confirms that the reversible reaction between rCP-βCD and biothiols was fast and depended mainly on the concentration of biothiols in the reaction system. The irreversible thiol scavenger N-methylmaleimide (NMM) was also used to explore the reversibility of the reaction. Addition of NMM caused a rapid decrease in the fluorescence at 405 nm (Fig. 1j). However, the response times in the reverse direction were different, especially for GSH, the recovery percentage from rCP-βCD + GSH to rCP-βCD was only 20%, perhaps as a result of differences in the degree of interaction between the biothiols and β-CD, such as hydrogen bonding and complexation.57,58 The result demonstrated that rCP-βCD was suitable for real-time monitoring of intracellular Cys and Hcy, but not of GSH.
Next, we focused on elucidating the high efficiency of the reaction between rCP-βCD and the biothiols. We began by synthesizing coumarin derivatives bearing an alkyl group (rCP-Al) and a pyridinium moiety (rCP-Py) and then evaluated the efficiency of their respective reactions with the three thiols (Scheme S1†). When the ratio of the two main absorption peaks (A405 nm/A485 nm) was plotted as a function of biothiol concentration, different linear relationships were observed (Fig. S9†). rCP-βCD gave the largest slope among them, which exhibited higher sensitivity toward biothiols relative to the compounds that were not conjugated to β-CD (Fig. 2b). Calculation of the dissociation constants (Kd) by means of the reported procedures20,21,25 revealed that rCP-βCD showed a much lower Kd (micromolar level) than rCP-Al and rCP-Py (millimolar level) (Table S1†). The smaller Kd reflects the stronger affinity between the double bond and mercapto group, indicating that rCP-βCD showed a remarkably high affinity for the biothiols and that the β-CD moiety played an important role in enhancement of the reaction efficiency.
To verify the spatial configuration of coumarin and β-CD, we measured the circular dichroism spectra of rCP-βCD and rCP-βCD-biothiols (Fig. S10†). The spectra of rCP-βCD and rCP-βCD-biothiols displayed no Cotton effects at the testing concentration (10 μM), which suggests that the coumarin was not encapsulated in the hydrophobic cavity of the β-CD. We suggest that self-inclusion was hampered by the relatively low concentration used to measure the spectra and by the relatively long and rigid triazole linker. On the basis of these results, we deduced that hydrogen bonding between β-CD hydroxyl groups and the coumarin moiety improved the efficiency of the reactions with the biothiols.
In order to exclude the influence of inclusion and to confirm the effects of hydrogen bonding, we filled the rCP-βCD cavity with adamantane sodium formate (ADA), and we also generated a permethyl-β-cyclodextrin decorated coumarin rCP-βCD(OMe) to block the effects of the hydroxyl groups (Fig. 2a). Reactions of rCP-βCD@ADA and rCP-βCD(OMe) with the biothiols showed that occupation of the cavity by ADA caused only a slight change in Kd relative to that of rCP-βCD (Fig. 2c, S11a–c and Table S1†), whereas methylation of the β-CD hydroxyl groups resulted in an obvious increase in Kd (Fig. 2d, S11d–f and Table S1†). In addition to confirming that the hydroxyl groups played a decisive role in the enhancement of reaction efficiency, these results also suggested that rCP-βCD could host a cancer-targeting agent without any interference with its ability to react with biothiols. After this, to evaluate the effect of the covalent bond between the probe and β-CD in rCP-βCD, we prepared nonbonded complexes between β-CD and rCP-Al and between β-CD and rCP-Py with β-CD, and evaluated their reactions with the biothiols (Fig. S12†). Under these conditions, no obvious spectral change was observed upon addition of the biothiols in the absence or presence of β-CD, which demonstrated that simply mixing a fluorescent probe with β-CD did not improve the reaction efficiency and that the covalent connection between β-CD and the probe was necessary.
To identify the mechanisms of the reaction between rCP-βCD and biothiols, DFT calculations were performed (Fig. S13, S14 and Table S2†). We found that there are two steps including dehydrogenation of biothiols and nucleophilic attack in the process of C–S bond formation but usually not synchronous addition (Fig. 2e and S13†). Of note, the water molecule complexed β-CD plays an important role in the process. When the mercapto group loses the H to a water molecule, an intermediate cation and anion pair form and the Gibbs free energy increases up to 45.05 kJ mol−1. The Gibbs free energy barrier is 33.07 kJ mol−1 for the formation of the C–S bond to form INT2via a transition state (TS) (Fig. 2f), which means that it is very easy for this step of the reaction to occur. At the same time, cooperative proton-transfer occurs at the hydroxyl of β-CD. After this, the carbanion of INT2 accepts a proton to form the stable product P. On the whole, the changes of the Gibbs free energy are acceptable in the mechanism, so that it is reasonable and confirms the reason for efficient detection of rCP-βCD. Therefore, on the basis of the experimental results and DFT calculation study, we propose a mechanism for the enhanced Michael addition reactions through supramolecular interaction between rCP-βCD and the biothiols. First, the rigid triazole linker and the hydrogen bonding between the hydroxyl group of β-CD and the unsaturated carbonyl group of the conjugated coumarin derivative jointly stabilizes the configuration of rCP-βCD. Second, the biothiols could be fixed by the hydroxyl group of β-CD and ester carbonyl, as well as one H2O which facilitates the nucleophilic addition reaction of thiol groups.
To target cancer cells via ligand–receptor-mediated endocytosis, we chose cRGD because it targets integrin αvβ3, which is overexpressed on tumor cells.59,60 The cRGD was modified with adamantane (cRGD-ADA), which could bind to rCP-βCD by means of adamantane–β-CD interactions and lead to cell-specific internalization. 1H NMR titration was used to investigate the complexation behavior of rCP-βCD and cRGD-ADA (Fig. S15†). When rCP-βCD was added to cRGD-ADA (0.5 mM), the proton signals of cRGD-ADA gradually shifted downfield as the rCP-βCD concentration increased from 0 to 3.0 mM, indicating inclusion of the ADA moiety in the rCP-βCD cavity. By nonlinear least-squares fitting of the titration data, we determined the binding constant (KS) for the complex of rCP-βCD@cRGD-ADA to be 1.1 × 104 M−1 (Fig. 3a). In addition, the stoichiometry of the complex was calculated to be 1:1 from Job's plot, in which the maximum was observed at a molar fraction of 0.5 (Fig. 3b and S16†). Meanwhile, the fluorescence quantum yields were measured to be 1.81% for rCP-βCD@cRGD-ADA at 560 nm (Fig. S17†).
Next, we confirmed the ability of rCP-βCD@cRGD-ADA to respond to biothiols. Upon addition of biothiols to the complex, it showed a detection performance similar to that of rCP-βCD (Fig. 3c and S18a–c†). In addition, excellent ratiometric fluorescence responses were observed under the same conditions (Fig. S18d–i†). In the dose–response plots for the reactions of rCP-βCD with biothiols, we observed obvious changes in the biothiol concentration range of 0–0.8 mM (Fig. 3d). By plotting absorption peak ratios (A405 nm/A485 nm) as a function of biothiol concentration, we determined Kd(Cys), Kd(Hcy), and Kd(GSH) to be 0.252, 0.077, and 0.122 mM, respectively. In addition, we tested the detection performance of rCP-βCD@cRGD-ADA towards biothiols in FBS and culture medium. As shown in Fig. S19,† the assembly exhibited excellent ratiometric fluorescence responses towards biothiols, which means that the FBS and culture medium shows no interference in the detection. Next, we tracked and tested the stability of the supramolecular assembly by means of fluorescence spectra (Fig. S20†). Within 24 h, no obvious fluorescence quenching occurred, demonstrating that the assembly has a good stability in the complex environment. The above results demonstrated that the rCP-βCD@cRGD-ADA was stable and could be used to sense biothiols in a complex environment, such as FBS and culture medium. Calculation of the rate constants showed that the reactions of rCP-βCD@cRGD-ADA with biothiols were very fast (t1/2(Cys) = 3.02 s, t1/2(Hcy) = 10.49 s, and t1/2(GSH) = 57.55 s; Fig. 3e). Taken together, these results demonstrate that the rCP-βCD@cRGD-ADA supramolecular conjugate detected biothiols with an efficiency on par with that of rCP-βCD.
After this, NMM was added to demonstrate the reversibility of the reactions. The addition of NMM caused a rapid decrease in the fluorescence generated by excitation at 405 nm, and rCP-βCD@cRGD-ADA exhibited a faster reverse response rate for GSH than rCP-βCD which attained full-recovery in 16 min (Fig. 3f and S21†). This difference probably arose from occupation of the β-CD cavity by ADA and the interaction between β-CD and cRGD, which would have reduced the interaction between β-CD and GSH and thus accelerated the depletion of GSH. However, 1H NMR spectroscopy did not capture obvious changes in the chemical shifts of the proton signals of β-CD or the biothiols, possibly owing to the presence of water (Fig. S22†). In addition, the addition of H2O2 also confirmed the reversibility of rCP-βCD@cRGD-ADA towards biothiols with the fluorescence spectra (Fig. S23†). In summary, the t1/2 values for rCP-βCD@cRGD-ADA with respect to reactions with the biothiols were very small, indicating that the reaction kinetics were fast and that this system could be used to reliably monitor the biothiol concentration dynamics.
To explore the utility of the rCP-βCD@cRGD-ADA complex for monitoring intracellular biothiol concentration dynamics in living cells, we carried out real-time imaging using live MCF-7 breast cancer cells. First, to clearly compare the internalization of rCP-βCD@cRGD-ADA between normal cells and cancer cells, MCF-7 and 293T cells were incubated separately with rCP-βCD and rCP-βCD@cRGD-ADA for 2 h (Fig. 4a and S24†). Compared with 293T cells, rCP-βCD@cRGD-ADA emitted bright visible fluorescence in MCF-7 cells in both the blue and green channels. Conversely, rCP-βCD exhibited weak fluorescence in both the MCF-7 and 293T cells. Notably rCP-βCD@cRGD-ADA gave brighter blue and green fluorescence than rCP-βCD did, confirming that rCP-βCD@cRGD-ADA was internalized in larger quantities and reacted with intracellular biothiols because of the presence of the cRGD cancer-targeting moiety. These results demonstrated that the cancer cell targeting ability is from the RGD. Statistical analysis of the fluorescence intensity data for the two channels confirmed the results indicated by the images (Fig. 4c). In addition, similar cell uptake experiments with rCP-βCD@cRGD-ADA in A549, KYSE-150 and COS-7 cells were also performed (Fig. S25†). The results demonstrated that rCP-βCD@cRGD-ADA was internalized in larger quantities than in normal cells. To monitor the biothiol concentration dynamics in the cells, we added NMM to perturb the intracellular redox homeostasis. Upon treatment of MCF-7 cells in PBS with a bolus of NMM, ratiometric imaging of the cells showed that the blue/green ratio initially slightly decreased (within 0.75 min) and then increased back to an equilibrium value after 3.75 min (Fig. 4b [upper panels], 4d and S26†). However, when the cells were treated with NMM in culture medium rather than in PBS, a different response was observed; specifically, the ratio initially increased over the course of 3.75 min and then remained elevated (Fig. 4b [lower panels], 4e and S27†). In contrast, there was no obvious change in the blue/green ratio when cells were incubated in NMM-free culture medium for 10.5 min (Fig. S28†). These results demonstrate that the rCP-βCD@cRGD-ADA could be used to monitor various biothiols. Importantly, because rCP-βCD@cRGD-ADA detected the biothiols with high sensitivity, we were still able to capture the rapid increase in the biothiol concentration necessary to maintain intracellular redox homeostasis upon depletion of biothiols by treatment with NMM. Moreover, real-time biothiol imaging with rCP-βCD@cRGD-ADA in A549 and KYSE-150 cells upon NMM treatment was also explored (Fig. S29a and S30a†). Upon treatment with a bolus of NMM, ratiometric imaging of the cells showed the decrease of the blue/green ratio which confirmed the reduction of biothiols. The related statistical analysis of the fluorescence intensity data is displayed in Fig. S29b and S30b.† From representative blue/green ratio images, obvious biothiol concentration variations in A549 and KYSE-150 cells were observed (Fig. S31†). These experiments demonstrate that rCP-βCD@cRGD-ADA could reversibly react with intracellular biothiols and sense intracellular biothiol concentration dynamics.
Cell culture and fluorescence imaging experiment. MCF-7, 293T, A549, KYSE-150 and COS-7 cells were grown in 1640 and F12 media containing 10% FBS, 1% penicillin, and 1% streptomycin at 37 °C in 5% CO2. Cells were plated on confocal dishes and allowed to adhere for 24 hours. For the confocal fluorescence imaging experiment, cells were washed with PBS and then incubated with rCP-βCD and rCP-βCD@cRGD-ADA in culture medium for another 2 h at 37 °C. The cell staining experiment was performed after washing with PBS 3 times. Experiments to assess reversible variation were performed under different conditions (culture medium or PBS) supplemented with NMM. Emission was collected at 430–510 nm (excited at 405 nm) for the blue channel and at 525–625 nm (excited at 480 nm) for the green channel.
Footnote |
† Electronic supplementary information (ESI) available: Experimental procedures and characterization. See DOI: 10.1039/d0sc00414f |
This journal is © The Royal Society of Chemistry 2020 |