Investigation of a halloysite-based fluorescence probe with a highly selective and sensitive “turn-on” response upon hydrogen peroxide

Inorganic halloysite nanotubes (HNTs) were modified with an organic fluorescein derivative (PA) to prepare HNTs-based hybrid fluorescence probe (HNTs-PA). Thermogravimetric analysis (TGA), Fourier transform infrared spectroscopy (FTIR), solid-state C NMR spectroscopy, X-ray photoelectron spectroscopy (XPS), energy dispersive spectrum (EDS) and transmission electron microscopy (TEM) confirmed that PA was successfully grafted onto the lumen of HNTs to obtain HNTs-PA with a grafting degree of 6.0%. The established B–C bond endows a selective fluorescence response toward hydrogen peroxide (H2O2) to HNTs-PA. The new hybrid probe exhibited a highly specific “turn-on” fluorescence response to H2O2 over other reactive oxygen species (ROS) and common ions owing to their chemoselective boronate-tophenol switch. The “turn-on” response could even be tracked when the additional amount of H2O2 was limited to 1 10 7 mol. Moreover, human lung cancer cells (A549 cells) were successfully stained and the staining intensity enhanced as time prolongs, which can be due to the overexpressed H2O2 of cancer cells. Thus, the as-prepared organic–inorganic hybrid fluorescence probe can have a broad range of applications for identification and diagnosis.


Introduction
Reactive oxygen species (ROS) oen play important roles in cell signaling and are also regarded as biomarkers for some pathological changes in human beings. 1,2 H 2 O 2 , one of the representative ROS, is oen involved in inammation, hemorrhage, tumor proliferation and so on. [3][4][5][6] The accumulation of H 2 O 2 can also destruct various cellular oxidants. Hence, it is important to trace the H 2 O 2 at low levels in order to prevent and diagnose the abovementioned diseases, which presents the need for developing H 2 O 2 -sensitive uorescent probes (Schemes 1 and 2).
Recent studies showed that the H 2 O 2 -triggered deprotection of arylboronates to phenols provided a reaction-based approach to the specic detection of H 2 O 2 , which results in the exploration of arylboronate-based uorescent probes for H 2 O 2 . 3,7 Fluorescein and its derivates have attracted much attention primarily due to their high quantum yields, desirable photostability, and long emission wavelengths. Chang et al. reported a series of arylboronate-containing uorescein derivatives. The H 2 O 2 -triggered deprotection of arylboronates to phenols can cause a dramatic increase in uorescent intensity, where the "turn on" effect makes the synthesized uorescein derivatives potential H 2 O 2 -sensitive uorescent probes. 3,7,8 However, the Scheme 1 Synthesis of boronic acid-contained fluorescein derivative (PA) by treating PF with diethanolamine and HCl. undesirable solubility in aqueous system restricts their applicability to a broader range of applications.
Halloysite nanotubes (HNTs) are a type of natural tubular nanoparticles with a broad range of length to diameter ratios. 9,10 The exterior and interior surfaces of HNTs are composed of siloxane (Si-O-Si) groups and gibbsite-like arrays of aluminol (Al-OH) groups, respectively, which results in a negatively charged outer surface and a positively charged inner lumen. 11,12 The peculiar composition allows selective functionalization at the inner or outer side, leading to the formation of a composite with a hierarchical nanostructure. 13,14 Their attractive properties, including non-toxicity, harmlessness, high mechanical strength, and good biological adaptability, 15-20 make HNTs nd a variety of applications in elds such as electronics, catalysis, separation, and functional materials. [21][22][23][24][25] Environment friendliness and biocompatible nature enable halloysite nanotubes (HNTs) to be a promising nanomaterial for developing organic/ inorganic composites in the biomedicine eld. 16,20,[26][27][28][29][30][31][32] The surface modication of HNTs would create new types of smart functional materials. [33][34][35][36][37][38][39][40] The selective modication methods have been extensively summarized in a latest review by our group. 41 In our previous studies, pyrenylboronic acid (PBA)-modied HNTs were prepared and the established Al-O-B linkage provided a "turn off" uorescence response for H 2 O 2 . 42 In addition, a H 2 O 2 -responsive drug delivery system was also constructed using the established Al-O-B linkage. 43 In this study, the uorescein derivative bearing arylboronic acid was conjugated to the inner surface of HNTs using the previously reported method. Moreover, the established Al-O-B linkage can endow a "turn on" uorescence response for H 2 O 2 toward the uorescein-modied HNTs. The good water-dispersity and biocompatibility properties of uorescein-modied HNTs make them more suitable for imaging H 2 O 2 signaling in living cells.

Materials
HNTs were obtained from Guang Zhou Shinshi Metallurgy and Chemical Company Ltd (Guangzhou, China). Fluorescein was purchased from J&K Chemical Technology. Dimethyl sulfoxide (DMSO) was dried and distilled from CaH 2 under vacuum before use. Distilled water was used throughout the study. The reactions were monitored by thin layer chromatography (TLC) with silica gel 60 F254 (Merck 0.2 mm). Column chromatography was carried out on silica gel (200-300 mesh). Peroxyuor-1 (PF1) was synthesized according to the literature. 8

Purication of HNTs
HNTs were puried according to the method described in our previous study. 44 Synthesis of boronic acid-contained uorescein derivative (PA) PF1 (940 mg, 1.7 mmol) and diethanolamine (400 mg, 3.8 mmol) were dissolved into an ether solution and then stirred at room temperature. Aer 50 min, the precipitate was ltered, washed with ether, and dried in vacuum to afford a brown powder. Then, the obtained solid was treated with 0.1 M HCl. Aer ultrasonication for about 30 min at room temperature, a yellow precipitate was obtained, which was centrifuged and washed with 0.1 M HCl to afford the product PA (435 mg, 66%) (Scheme S1 †). 1

Preparation of HNTs-PA
A mixture of puried HNTs (120 mg) and PA (40 mg) in anhydrous DMSO was carefully degassed. The system was heated at 80 C for 12 h under stirring in nitrogen. The mixture was cooled to room temperature and washed by diethyl ether. The residue was collected by centrifugation. Aer vacuum-drying, the HNTs-PA was obtained as a faint grey solid.

Cell cultures
Human lung cancer cells (A549 cells) and Hela cells were cultured in Dulbecco's modied Eagles medium (DMEM, Beijing North TZ-Biotech Develop, Co. Ltd.) with 10% fetal bovine serum and 100 U mL À1 of 1% antibiotics (penicillin/ streptomycin). A549 cells were placed on a confocal dish and the number of cells on each confocal dish was about 5.0 Â 10 3 . Then the cells were maintained at 37 C for 12 h in a 100% humidied atmosphere containing 5% CO 2 for further use.

MTT assay
A549 cells or Hela cells were seeded in a 96-well plate at a density of 4 Â 10 3 per well. The cells were provided by the Chinese Academy of Medical Sciences, Peking Union Medical College. Accurately weighed HNTs-PA probes were dissolved in phosphate buffered saline (PBS) to achieve mother liquors. Following this, the cells were further incubated in media containing various concentrations diluted from the mother liquors. The plates were maintained at 37 C in a 5% CO 2 /95% air incubator for 24 h. Aer 24 h, the medium was removed and washed with PBS. MTT solution (0.5 mg mL À1 , 100 mL) was then added to each well. Aer 4 h, the remaining MTT solution was removed and 100 mL of DMSO was added to each well to dissolve the formazan crystals. Absorbance was set at 520 nm by a Triturus microplate reader. Wells containing no samples were used as blank control. The cell viability (%) was calculated as the ratio of the number of surviving cells in the test samples to the control groups.

Cell imaging
Cells were grown in Minimum Essential Medium Eagle (MEM medium, Sigma) in a watch glass bottom poly-D-lysine coated Petri-dish for at least 24 h to enable adherence to the bottom. Accurately weighed amounts of HNTs-PA suspension or Mitotracker Red were added to the medium at 37 C in 5% CO 2 . The cells were maintained in an incubator for another 30 min to achieve a sample loaded MEM medium. Then, the cells were washed three times with DPBS buffer to remove the samples that were not taken up into the cells. Finally, DMEM was added.
A Hitachi F-4500 spectrouorometer (Tokyo, Japan) equipped with a xenon lamp was used to obtain the uorescence spectra. All experimental parameters (the laser intensity, exposure time, objective lens) were stationary when the different uorescence images were captured. The excitation wavelength was xed at 488 nm and uorescent signals were collected from 500 to 600 nm.

Characterizations
Microwave reactions were performed using a Discover (PE-CEM) reactor. UV-visible absorption spectra were obtained on a Shimadzu UV-visible spectrometer model UV-2550. Fluorescence spectra were recorded on a Shimadzu RF-5301PC. TGA was performed on Perkin-Elmer Pyris 6 under a nitrogen ow. Accurately weighed amounts of samples were heated at a scanning rate of 10 C min À1 from 40 C to 800 C. The morphological characterizations were performed using a Tecnai G 2 F20 S-TWIN transmission electron microscope (TEM) with an accelerating voltage of 200 kV. FTIR spectra were recorded in the region of 400-4000 cm À1 for each sample on a Varian-640 spectrophotometer. The samples were previously grounded and mixed thoroughly with KBr. The spectrum for each sample was obtained by averaging 32 scans over the selected wave number range. 13 C solid-state NMR spectra were recorded on a Bruker Advance III spectrometer. X-ray photoelectron spectroscopy (XPS) was carried out on a Thermo Scientic ESCALab 250Xi using 200 W monochromated Al Ka radiation. The 500 mm X-ray spot was used for XPS analysis. The base pressure in the analysis chamber was about 3 Â 10 À10 mbar. Typically the hydrocarbon C 1s line at 284.8 eV from adventitious carbon was used for energy referencing. The energy dispersive spectrum (EDS) was determined by a Hitachi S-4800 scanning electron microscope.

Results and discussion
Previous studies have demonstrated that arylboronic acid can react with the hydroxyls spread on the surface of HNTs. 42,43 In this study, a uorescein derivative bearing an arylboronic acid group was synthesized and then bound to the interior surface of HNTs according to the above-mentioned method. The obtained nanocomposite (HNTs-PA) showed desirable dispersibility in aqueous solution. The differences between HNTs-PA and HNTs were revealed by FTIR, XPS and solid-state 13 C NMR. Fig. 1 shows the solid-state 13 C NMR spectra of HNTs and HNTs-PA. HNTs show no visible resonance in the region from 0 to 200 ppm, implying the absence of carbon units in HNTs. Moreover, remarkable resonances ranging from 180 to 40 ppm can be clearly observed in the solid-state 13 C NMR spectrum of HNTs-PA. The peaks at 170.7 and 83.8 are attributed to the carbon in -C]O and the quaternary carbon in the PA units. The resources from 157.1 to 116.9 can be assigned to the aromatic carbons in the PA units. The remarkable difference in the 13 C NMR spectra between HNTs and HNTs-PA indicated the presence of uorescein groups in HNTs-PA.
The FTIR spectra of HNTs and HNTs-PA are shown in Fig. 2. In the spectrum of HNTs, two distinct peaks located at 3696 and 3622 cm À1 are attributed to the inner hydroxyl groups lying between the tetrahedral and octahedral sheets and the surface hydroxyl groups in the lumen of the nanotubes, respectively. The band at 1034 cm À1 is assigned to the stretching vibration of Si-O-Si. The FTIR spectrum matches well with the composition of HNTs.
As for HNTs-PA, the distinct band corresponding to the surface hydroxyl groups in the lumen at 3696 cm À1 displays a signicant decrease, suggesting the consumption of surface hydroxyl groups in lumen during the modication. Simultaneously, some new peaks at 3525, 3025, 2921, 1765 and 1410 cm À1 can be clearly observed, which can be attributed to the -OH stretching, Ar-H group, C-H bond, C]O bond and aromatic ring group, respectively. The FTIR results indicate that the uorescein derivative was graed on the inner surface of HNTs via the condensation reaction.
The quantity of the uorescein derivative graed to the nanotube was calculated to be 6.0 wt% by thermogravimetric analysis (TGA, Fig. 3). The XPS spectra are presented in the ESI. † Original HNTs and HNTs-PA all show the presence of aluminium (Al 2s and Al 2p) and silicon (Si 2p), in accordance The presence of carbon units (atomic% ¼ 11.25%) revealed that the uorescein groups were immobilized onto the HNTs, which matched well with the solid-state 13 C NMR results. The TEM images of original HNTs and HNTs-PA are shown in Fig. 4. Fig. 4A shows that the original HNTs possess a cylindrical shape and contain a transparent central area that runs longitudinally along the cylinder. The nanotubular particles are hollow and open ended, with a length of 0.5-1.0 mm, an inner diameter of 20-30 nm and an outer diameter of 40-60 nm, which can be untaken or phagocytized in cell cultures.
Aer modication, the tubes retain their smooth outermost surface. Moreover, the lumen is almost completely lled, suggesting that the uorescein derivative was successfully graed onto alumina at the tube lumen and did not bind the outer siloxane surface of the tube. The overall morphologies of these nanotubes were similar to those of the original nanotube, indicating that the nanotubes underwent slight damage during the graing reaction.
The established B-C link in the modied HNTs can be degraded into B-OH and Ar-OH in the presence of H 2 O 2 as well as the B-C link in the uorescein derivative. The quantum yield (F f ) of HNTs-PA was calculated as ca. 0.01, with quinine sulfate (F f ¼ 0.55 in 0.05 M H 2 SO 4 ) as the reference standard. Moreover, the degradation product, uorescein, possesses a signicantly higher uorescence quantum yield (F f ¼ 0.94 in 0.1 M NaOH 3 ) than modied HNTs and uorescein derivative bearing arylboronic acid groups. Therefore, the degradation from HNTs-PA to uorescein can dramatically increase the  uorescence intensity, which can be called a "turn-on" response. The "turn-on" response in uorescence intensity toward H 2 O 2 can endow a H 2 O 2 -detecting ability of modied HNTs. Fig. 5 shows the uorescence responses of HNTs-PA to H 2 O 2 . When 5 Â 10 À6 mol H 2 O 2 was added, the HNTs-PA aqueous solution displayed a drastic increase in uorescence intensity in 30 min. A 50-fold difference in uorescence intensity can even be detected when the amount of added H 2 O 2 decreased to 1 Â 10 À7 mol (shown in ESI †), where the concentration of H 2 O 2 is 5 Â 10 À5 M. Previously, we developed a HNTs-based H 2 O 2sensitive uorescence probe, for which the uorescence response could not be tracked when the concentration of H 2 O 2 decreased to 5 Â 10 À5 M. Therefore, the HNTs-PA shows a prominent advantage as compared to the previous probe. 42 Owing to their chemoselective boronate switch, HNTs-PA probes retain highly specic responses to H 2 O 2 . Other groups did not show the "turn-on" uorescence when treated with other ROS agents, including NOc and ClO À (shown in the ESI † To evaluate the cytotoxicity of the modied HNTs, the growth inhibitory activity against A549 cells was calculated based on a MTT assay. The concentration expressed in the Fig. 6 represents the concentration of HNTs-PA in aqueous suspension. The results indicate that HNTs-PA shows low cytotoxicity (over 80% cell viability) within 24 h of incubation time, in which the concentration of HNTs-PA ranges from 0.05 to 1.0 mg mL À1 , which makes the modied HNYs desirable biological materials at cellular levels. In addition, the cytotoxicity of HNTs-PA was also evaluated using Hela cells, where the IC 50 was calculated to be 1.11 mg mL À1 .
It is well-known that the overexpression of H 2 O 2 usually occurs around the tumor cells. We next assessed the ability of HNTs-PA to stain within human lung cancer cells (A549 cells). A549 cells were incubated with 0.5 mg mL À1 HNTs-PA for 1, 2, 4, 8 and 12 h at 37 C (shown in Fig. 7). The excitation wavelength was xed at 488 nm and uorescent signals were collected from 500 to 600 nm. A reference test was also made by incubating A549 cells with Mitotracker Red, which is a commercially available uorescent dye for cell imaging. The images indicated that the cells can be stained under the endogenous levels of    2 O 2 derived from the cancer cells, where the uorescence intensity increased as time extends. Moreover, the uorescence intensity of the cells treated by Mitotracker Red remained almost unchanged aer 4 h. The difference can be due to the fact that the uorescein unit gradually released from the lumen as the B-C bond broke into B-OH and Ar-OH in the presence of H 2 O 2 , which corresponds to the "turn on" response of HNTs-PA on H 2 O 2 . We also assessed the ability of HNTs-PA to stain within Hela cells. The results indicated that the cells can also be stained and the uorescence intensity increased as time extends, which can also be due to the endogenous levels of H 2 O 2 . The H 2 O 2 concentration in mammals commonly ranges from 1 nM to 0.5 mM and is usually overexpressed in cancer cells and tissues. Sometimes the H 2 O 2 concentration in cancer cells can exceed 50 mM. 46 The H 2 O 2 in A549 cells and Hela cells should also be in this range, implying the HNTs-PA is capable of detecting the endogenic H 2 O 2 at this level.

Conclusions
In summary, we have described the design and synthesis of a uorescent probe HNTs-PA, which can be used for the detection of H 2 O 2 . The nanocomposite showed regular morphology and desirable dispersibility in aqueous solution. The established B-C linkage provides a highly specic H 2 O 2sensitivity to HNTs-PA; as a result, the probe exhibits a highly specic "turn-on" uorescence response for hyperoxide. Moreover, HNTs-PA was found to be capable to monitor H 2 O 2 derived from A549 cells under physiological conditions. The prepared HNTs-based uorescent probe can open new applications in the detection of cancer cells and thus serve as a practical tool for cancer diagnosis.

Conflicts of interest
The authors declare that they have no conict of interest.