An Acid-triggered Reactive and Enhanced Fluorescent Probe for Selective Detection of Al3+/H+ and its Application in Real Water Samples and Living Cells

A reactive fluorescent “turn-on” probe (di-PIP) with imine-linked dual phenanthro[9,10-d]imidazole luminophore have been conveniently prepared as an Al3+ and H+ dual functional receptor. di-PIP displayed high selectivity and sensitivity towards Al3+ ion in DMF/HEPES accompanied by fluorescence blue-shift and a good linear relationship as well as a low detection limit of 30.5 nmol·L–1, which can root from the synergetic functions of the decomposition reaction of di-PIP promoted by acidic Al3+ and the coordination effect between decomposition product and Al3+. Intriguingly, it was found that hydrogen ion H+ can be sufficient for simulating the fluorescence enhancing of di-PIP. 1H NMR titration and MS analyses for elucidation of the intermediate structure further revealed that the acid-triggered decomposition reaction resulted in the rapid, and sensitive sensing to Al3+ and H+. In addition, the probe di-PIP could be successfully applied to the detection of Al3+ in real water samples, and also utilized to visualize Al3+ and H+ in the living cells.


Introduction
As the third most prevalent (8.3% by weight) metal element on the earth's crust, aluminum is gradually being widely used in industrial development due to its unique nature and ease of production [1][2][3]. It is normal that aluminum is daily exposed to the human body and natural ecology. Despite the fact that aluminum element is relative stable, aluminum can exist and accumulate in its ionic form Al 3+ in biological and environmental systems [4,5] due to its excess usage, which have not only also caused major health difficulties but also disturbed the plants and aquatic ecosystem. It also has been reported that the tolerable daily and weekly aluminum dietary intake in the human body is about 3 ~ 10 mg/day and 7 mg/kg body weight [6]. A large number of proofs show that excessive intake of Al 3+ might cause some diseases in human beings, including Alzheimer's disease [7], Parkinson's disease [8], encephalopathy [9], cardiac arrest [10], kidney stone [11], and so on [12][13][14]. In addition, the toxicants of Al 3+ have also impacted adversely on the environmental systems. For example, the accumulation of excess Al 3+ in its ionic state would lead to acidification of soil; the high levels of Al 3+ in irrigation water can caused the poisoning of growing plants [15]. Therefore, it has been of great significance to effectively detect Al 3+ in various environments and accurately mark the whereabouts of Al 3+ in organisms.
Generally, the fluorescence analysis method is popular because of its high sensitivity and super anti-interference ability compared with other detection methods [16,17]. To date, many fluorescence sensors for the detection of Al 3+ are altered by the heteroatoms (N and/or O atoms) as the donor sites such as Schiff bases [18,19], flavone bases [20] and peptide bases [21], all of which are almost designed based on the incorporation and chelation using the reported sensing mechanism such 1 3 as photoinduced electron transfer (PET) [22], fluorescence resonance energy transfer (FRET) [23], chelation-enhanced fluorescence (CHEF) [24], aggregation-induced emission (AIE) [25], and many more [26,27]. Moreover, most of Al 3+ sensors have exhibited tremendous potential ascendent in physiological environmental and actual environmental. However, the fluorescence response of Al 3+ has been abstracted owing to its strong hydration enthalpy in the aqueous circumstance and the relatively poor coordination ability [28,29]. Moreover, the reported fluorescence sensors for Al 3+ are generally subjected to the fluorescence quenching, the relative higher detection limit and longer response time. Thus, it is still a challenge to develop new type of the fluorescent probes that can favorably display the sensing ability towards Al 3+ in complex matric.
pH is one of the most essential and necessarily controlled parameters and acts a vital role in biology, environment, and medicine [30]. Exiguous changes of intracellular pH usually signify cell dysfunction, which results in abnormal physiological processes and even diseases, including stroke, Alzheimer's disease, cancer, inflammation, cardiovascular diseases, and so on. Therefore, it is indispensable to monitor the acid − base balance in living cells to understand physiological and pathological processes. Although some approaches for the pH test have been well set up in the literatures [31,32], the optical technology with pH sensors could adopt as an alternative to conventional methods in particular cases, for example, the high throughput screening, the remote sensing, the real-time analysis. Recently, a huge number of the organic sensors for pH have been multifariously modified based on BODIPY [33], rhodamine [34], some cyanine dyes [35], etc. Derivatization is usually necessary for binding with proton. However, few reported sensors have abilities to response another specific analyte besides pH. This is a universal issue encountered during the fluorescence detection processes. Therefore, it is imminently demanded to develop the multi-targeted fluorescent sensors for two or more analytes based on a single molecular platform, which must availably avoid the excessive consumption of the resource and ameliorate the detecting efficiency of the target analytes.
In this contribution, we intend to report a facile synthesis of symmetrically organic molecule of phenanthro [9,10d]imidazole linked by Schiff base (di-PIP) as a novel multi-targeted fluorescent sensor. In the backbone of di-PIP, phenanthro [9,10-d]imidazole was used as a fluorescent indicator. Nicely, we have found that the reported phenanthro [9,10-d]imidazole derivatives had excellent aggregation-active in the aqueous medium, which indicates that the fluorescence emission of di-PIP might employ as a water trace indicator in water solvents. In addition, the large amounts of literatures have revealed that the unit of Schiff base came into play as one of the donor locations for selective detection of Al 3+ . Meanwhile, in view of the Lewis acid characterization of Al 3+ , the response ability of di-PIP towards H + is also expectant. Encouraged with above information, we organically integrate two segments into a whole fluorescent sensor di-PIP that can simultaneously detect Al 3+ and H + circumstance on account of the aggregation state, which successfully circumvents the demand that a sensor platform can detect the multi-targeted samples in the aqueous medium.

Materials and Instrumentation
All chemicals were commercially purchased as analytical grade from Shanghai Aladdin Biochemical Technology Co., Ltd (Shanghai, China) without further purification. All solvents were bought from local reagent stores without further purification. HeLa cells and transferrin were obtained from Gaining Biological Co., Ltd (Shanghai, China). Water used in spectral experiments was distilled twice. Melting points were recorded on an X4 melting apparatus without correction. IR spectra were measured as KBr disks with a Perkin Elmer Spectrum one FT-IR. 1 H and 13 C NMR spectra were recorded on a Bruker Avance 600 and 150 MHz spectrometer with tetramethyl silane as internal standard. UV absorption and fluorescence spectra were performed on a TU-1901 and a Perkin Elmer LS55 Luminescence Spectrometer, respectively. MS spectra were collected using a Xevo G2-XS QTof. Elemental analysis was implemented with a 4100 elemental analyzer. Images of scan electron microscope (SEM) were obtained on an S-3400 SEM (HITACHI, Japan).

Synthesis of di-PIP
The synthesis method of compound di-PIP was described in Scheme 1. Compound (1H-phenanthro[9,10-d]imidazol-2-yl)aniline (PI-aniline) was prepared according to the Scheme 1 Synthetic method for compound di-PIP reported procedure in the literature [36,37]. The mixture of PI-aniline (2.5 mmol, 0.78 mg) and 1,4-phthalaldehyde (1.0 mmol, 0.13 g) in acetic acid (20 mL) was stirred at room temperature for 6 h. After that, 200 mL water was added into the above mixture. The resulting precipitate was filtered and recrystallized from the mixture of ethyl acetate and petroleum ether as a yellow solid (0.67 g, 93% yield).  (Fig. S2). 13 S4).

Spectroscopy Measurements
The stock solution (0.1 mol·L -1 ) of all organic compounds were prepared at absolute DMF and the corresponding tested solution (10 μmol·L -1 ) was obtained through diluting the stock solutions by HEPES buffering solution (10 μmol·L -1 ). All the inorganic salts used in the measurements were dissolved in twice-distilled water. The fluorescence emission spectra were recorded at an excitation wavelength of 310 nm. Both excitation and emission slit widths were set at 10 nm.

Application in the Real Samples
Lake water was captured from local river (Nenjiang River, Heilongjiang province, China), which filter the impurities such as float and precipitate. Other water samples were purchased from local markets and analyzed before expire dates. All the samples were diluted 100 folds with 400 mmol·L -1 HEPES buffer solution (pH = 7.4) prior to the preparation of spiked solutions with different concentrations of Al 3+ , after which the fluorescence intensities of samples were checked to determine the concentrations of Al 3+ . All the measurements were done in triplicate throughout the experiments.

Cell Cytotoxicity and Images
The procedures for the cytotoxicity test and the cell culture were proceeded according to the literatures [38,39]. The standard MTT assay was used to detect the compatibility of di-PIP with HeLa cells. The cells were first cultured in Dulbecco's modified Eagle medium (DMEM, high glucose: 4.5 g·L -1 ) supplemented with 10% fetal bovine serum (FBS) in an atmosphere of 5% CO 2 and 95% air at 37 °C. After that, the cells were incubated to di-PIP at various concentrations (0, 25, 50, 75, 100, 150, and 200 μg·mL) for 24 h. Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay and the cytotoxicity of di-PIP was evaluated by the ratio of the absorbance of the sample treated with di-PIP to the control sample.
The cells at a concentration of 3 × 10 3 cells/mL were seeded in 24 well plate for 24 h and then were incubated with the concentrations of Al 3+ (10 μmol·L -1 ) for extra 2 h. The cells under the conditions of the different pH values were incubated in an incubator for 1 h. Before the confocal imaging, the cells were washed off with PBS solution (10 μmol·L -1 ) to remove di-PIP, Al 3+ , and environmental media. Before the confocal imaging, the cells were fixed with 4% paraformaldehyde for half an hour. After washing and drying, the fluorescence images of the cells were captured using n a laser scanning confocal microscopy.

Design and Synthesis of the Probe
As outlined in Scheme 1, the targeted compound di-PIP was facilely prepared by PI-aniline and p-phthalaldehyde as starting materials in the solvent of acetic acid. The reaction between PI-aniline and p-phthalaldehyde proceed smoothly at room temperature, giving the product di-PIP that was composed of double 1H-phenanthro[9,10-d]imidazole moieties linked through covalent conjugation by phenyl imine. This structural characterization containing π sufficient moieties that might freely rotate in the single-molecule state can lead to energy consumption of the excited state through non-radiative pathways, thus ensuring that these compounds are weakly emissive in solution.

Behavior of Aggregates
The aggregation behavior of compound di-PIP was firstly examined by its UV-vis and fluorescence spectra in the DMF/water cosolvent system with varied water fraction (from 0 to 90%). As shown from the UV-vis absorption spectrum of di-PIP (Fig. 1a), in the pure DMF solution, di-PIP possessed two absorption bands centered at about 315 nm and 405 nm, respectively, with the molar extinction coefficients (ε) of 4.85 × 10 4 and 4.32 × 10 4 M −1 ·cm −1 . The ultraviolet band at 315 nm could be attributed to the π → π* transition of the isolated phenanthro[9,10-d]imidazole unit; whereas the visible one near 405 nm was likely due to the π → π* and n → π* electronic transition between two parts of phenanthro [9,10-d]imidazole unit due to the extended conjugated system in the molecular di-PIP. The results of UV-vis spectrum was compared with other reported phenanthro [9,10-d]imidazole compounds in the some literatures [40,41] those had only single absorption at the ultraviolet area, which indicated that di-PIP might be seized of much predominant properties in spectrum. Upon the addition of water into DMF solution, there was no obvious changes observed at the lower water fraction (below 30%). However, when the water fraction exceeded over 30%, the absorption intensities for both two bands conspicuously descended with the significant redshifts of 15 and 45 nm, respectively. At the same time, two absorption regions became broader than those in the neat DMF and low water fraction. Above spectra phenomena of the lower intensity, broader and red-shifted wavelength were ascribed to the formation of the aggregated states of the molecular di-PIP in the highly soluble solvent system.
On the basis of the absorbance data, the photoluminescence property of di-PIP in a series of DMF/water cosolvent was enumerated by fluorescence spectrum and the result was depicted in Fig. 1b. Compound di-PIP could well dissolve in DMF solution with a weak emission at 460 nm upon excitation at 311 nm. Once the water was gradually induced, a growing trend towards the fluorescence intensity occurred and quickly reached the maximum when the water fraction reached 30%. Nevertheless, with the continual increment of water content, the fluorescence intensity of di-PIP sharply declined and quenched the minimum upon 90% of water fraction. The phenomena of the fluorescence intensity for first arising and then failing could be explained by the two effects of aggregation-induced emission (AIE) and aggregation-caused quenching (ACQ). That is to say, the changes of water fractions in the solution of di-PIP might cause the transformation of aggregated states for di-PIP which led to the alteration of the fluorescence intensity.
To intuitively validate the behavior of aggregates, we observed the color changes of compound di-PIP in the DMF/water solution with the increasement of water fraction by both the "naked-eye" and the Tyndall phenomenon. As depicted in Fig. 1a inset, in neat DMF without water, the di-PIP solution with the color of light yellow was homogenous with well dissolution. Whereas, with increasing of the water fraction, the color of the di-PIP solution gradually deepened from light yellow to yellowish-brown. Moreover, a cloudy uniform dispersion was gradually produced and a clear "Tyndall Effect" was observed under the irradiation by the red laser-pen at dark field ( Fig. 1b inset), which proved the degree of the aggregated di-PIP deepened along with the increment of water content.
In order to intuitively acquire the evidence for the aggregation, the changes of the morphology of di-PIP depending on the changes of water fraction were recorded by SEM experiments. As result in Fig. 2, the incompact strip-like morphology (Fig. 2a) was observed with the size of about  , and 90% (c) volume percentages of water 0.3 μm × 2 μm in the 10% water medium; and the concordant strip-type morphology (Fig. 2b) was much densely arranged on the carrier in 30% water solution, which clearly indicated that the aggregated degree of di-PIP deepened gradually along with the increasing content of water. Under the condition of 90% water fraction, there was no regularly zonal arrangement (Fig. 2c) to be observed any more hinting that the disordered arrangement of di-PIP caused by deep aggregation have been formed in the nearly aqueous medium, which might be responsible for the phenomenon of the quenching fluorescence intensity.

Fluorescent Sensing Performance of di-PIP for Al 3+
The spectral behavior of di-PIP persuaded us to study its photoluminescence properties for ion fluorescence sensors. The metal-binding performance of di-PIP (10 μmol·L -1 ) was inspected by UV-vis and fluorescence spectra toward Co 2+ , Fe 3+ , Cr 3+ , Cd 2+ , Zn 2+ , Cu 2+ , Hg 2+ , Ni 2+ , Ba 2+ , Na + , K + , Ca 2+ , Mg 2+ , Ag + , Pb 2+ and Al 3+ in DMF/HEPES buffer solution (v/v, 7:3, 10 mmol·L -1 , pH = 7.4). The UV-vis spectra of di-PIP was measured immediately after the addition of metal ions (50 μmol·L -1 ) and the results were represented in Fig. 3a. No remarkable changes in intensity and position of the absorption peaks were glanced after the addition of other metal ions. Notably, the addition of Al 3+ resulted in a significant alteration. The original two absorption at 315 and 405 nm for free di-PIP disappear simultaneously and was merged into a strong and wide adsorption band centered at 340 nm, which indicated that the electronic transition among the conjugative structure in free di-PIP might have been destroyed because of the existence of Al 3+ . The titration experimental of the concentration of Al 3+ was further carried out by UV-vis spectra (Fig. 3b). It was obviously seen that the absorption band at 405 nm vanished little by little along with the augmenting of the concentration of Al 3+ and the absorption band at 315 nm gradually shifted toward the longer wave. At the same time, an isosbestic point at 365 nm was noticed distinctly. When the Al 3+ concentration reached 20 μmol·L -1 , the absorption band in visual region completely perished and a new absorption band at 340 nm appeared. At this moment, the color of the di-PIP solution containing 20 μmol·L -1 Al 3+ was colorless which was compared to that of free di-PIP in light yellow (Fig. S5). Above result initially demonstrated that di-PIP had ability to respond Al 3+ .
The fluorescent response of di-PIP to Al 3+ was subsequently assessed based on mildly aggregated condition in DMF/HEPES buffer solution (v/v, 7:3, 10 mmol·L -1 , pH = 7.4), under which the free di-PIP expressed the moderate fluorescence intensity at 460 nm. Excitingly, the fluorescent intensity of the mixture of di-PIP and Al 3+ significantly sprang up after two equivalents of Al 3+ was introduced into the solution of di-PIP (Fig. 3c). Simultaneously, a bright blue emission was readily observed by naked eye in the presence of Al 3+ with a sharp contrast to its original unobservable luminescence (Fig. 3c inset). Figure 3d gave a gradual fluorescence increment (~ 10 folds) as the increasing concentration of Al 3+ from 0.0 to 20 μmol·L -1 with the fluorescence quantum yield of 0.35. Interestingly, Al 3+ -induced enhancement was accompanied with a visible blue shift from 460 to 429 nm, which also meant the fact that the large conjugation platform of di-PIP had been crushed by the introduction of Al 3+ . The intense emission might be responsible for the restriction of the rotation of the molecule di-PIP and chelation-enhanced fluorescence behavior (CHEF) in presence of Al 3+ . Based on the changes of the intensity related with the amount of Al 3+ , the Job's plot measurement was sketched with the 1:2 complex between compound di-PIP and Al 3+ (Fig. S6) and the association constant K a of di-PIP was found to be 4.6227 × 10 6 , which illustrated the formation of complex behavior between di-PIP and Al 3+ .
Furthermore, the sensitivity of the response of di-PIP towards Al 3+ was fully evaluated by the fluorescence spectra. A trace amount of Al 3+ concentration (0.1 equivalent) could effectively stimulate the promotion of the emission by around twofold, which meant that di-PIP could sensitively respond to Al 3+ . When the coordination equilibrium was reached, the ratio of Al 3+ to di-PIP was 2, which was according with the result of the Job's plot. In addition, a good linear relationship based on the equation of y = 0.9983x + 0.9661 existed between the fluorescence intensity and the Al 3+ concentration in the range of 0 ~ 1.0 μmol·L -1 with a correlation coefficient (R 2 ) of 0.9960 (Fig. 3d inset). According to this calibration curve, the detection limit of Al 3+ was calculated to be 30.5 nmol·L -1 , which was comparable with those of reported Al 3+ probes [42,43], suggesting that di-PIP was enough sensitive to determine Al 3+ . Besides, the possible interferences by other metal ions were justified through competitive experiments. 40 μmol·L -1 other competitive ions followed by 20 μmol·L -1 Al 3+ was added into the tested solution di-PIP (10 μmol·L -1 ) and the results of the fluorescence spectra were recorded in Fig. S7. The competitive ions produced negligible change for the detection of Al 3+ , which further manifested that compound di-PIP was highly selective for Al 3+ as indicated in a previous absorbance study and could be applied to determine Al 3+ quantitatively.

The Fluorescence Sensing of di-PIP Toward H +
Notably, the associations and dissociations of host − guest systems were also controlled by acid − base treatments. Therefore, the pH effects on the fluorescence properties of di-PIP and the recognition of di-PIP to Al 3+ were evaluated for potential application. The pH dependent on the fluorescent intensity of di-PIP before and after the addition of Al 3+ with various pH condition (pH = 2.0 ~ 9.0) was illustrated in Fig. 4. For the rare di-PIP, titrating the DMF/ water buffer solution (v/v, 7:3, pH = 9.0) of di-PIP with 0.1 mol·L -1 HCl, the no significant fluorescence change at 450 nm was observed when the pH values were adjusted in the range of 8.0 ~ 9.0, indicating that di-PIP could maintain the stability in the alkalescent phenomena. Along with the decreasing of pH value from 8.0 to 2.5, the steady enhancement in intensity was clearly monitored (Fig. 4a), and the maximum fluorescence intensity at pH = 2.5 was 12 times higher than that at pH = 8.0 (Fig. 4b). The stronger emission under acid conditions might be because of the protonation of N atom in molecule di-PIP, which destroyed the PET effect from N atom to the luminophore phenanthro [9,10-d]imidazole unit. In addition, the recycling fluorescence performance caused by the conversion between the protonation and deprotonation was fully tested at pH 3.0 and 8.0, which could be repeated four times with little errors upon alternate addition of HCl (0.1 mol·L -1 ) and NaOH (0.1 mol·L -1 ) solutions (Fig. S8). Furthermore, the linear response between the emission intensity and pH values was obtained with a linear coefficient (R 2 ) of 0.9931 in the ranges of 3.0 ~ 8.0 according to the linear regression equation (I = 870.5332-109.5119 pH) (Fig. 4b  inset). Accordingly, the pK a value was estimated to be 5.21 based on the emission intensity ratios versus pH values (Fig. S9). All results of pH response supported that di-PIP was sensitive to H + concentration and could be used as a pH sensor.

Sensing Mechanism
Aiming to decipher the specific interaction mode between di-PIP and Al 3+ , 1 H NMR titration experiments in DMSOd 6 /D 2 O (7/3, v/v) was measured and the results were illustrated in Fig. 5. It was obvious that the 1 H NMR spectrum of the free di-PIP (Fig. 5a) was clear, distinct and simple in pure DMSO-d 6 . Two single signals at δ 13.53 and 8.19 ppm were attributed to the -NH protons in the imidazole ring and -CH = N imine proton, respectively; other unsaturated H protons emerged in the 7.6 ~ 8.8 ppm range. With the increasing addition of Al 3+ (0.5 ~ 2.0 equivalents), the more intricate and attractive spectra were noticed. As shown in Fig. 5b ~ e, two identical single peaks emerged at 10.14 (H a1 ) and 10.12 (H b1 ) ppm, respectively, which was typically emblemized as the signals of formyl groups, hinting that two different aldehydes might engender at the same time by the existence of Al 3+ ; in addition, the areas of two single peaks (H a1 and H b1 ) changed accordingly with the addition amount of Al 3+ . Furthermore, a characteristic aromatic protons (H a2 ) was found at δ 8.13 ppm where the integral area of aromatic protons was twice as much as that of H a1 proton when the addition amount of Al 3+ was 2.0 euqiv. Coincidently, the combining the characterizations of the formyl proton H a1 with the aromatic protons H a2 exactly deserved as the characteristic peaks of p-phthalaldehyde, which suggested that it was possible that di-PIP have been decomposed into the fragment of p-phthalaldehyde in the presence of Al 3+ . Encouraging by this envisaging, we examined carefully the 1 H NMR spectra of di-PIP-Al 3+ (2.0 equivalents) with that of the pure PI-aniline (Fig. S10). As expected, the chemical shift values (8.85, 8.55, 7.99, 7.70, 7.59, 6.73 and 5.55 ppm), the splitting patterns as well as the numbers of proton signals were almost according with those of the pure PI-aniline. Especially, -NH 2 amino signals produced due to the decomposing of di-PIP was distinctly observed at around 5.55 ppm in Fig. 5b ~ d. These results favorably versified the above hypothesis that the symmetrical and larger conjugated system of di-PIP was wrecked into two parts of p-phthalaldehyde and PI-aniline based on the broken of the imine C = N bonds by the assistant of Al 3+ . Interestingly, when the addition amount of Al 3+ reached 2.0 equivalent, the amino NH 2 and the imidazole NH protons disappeared simultaneously in the NMR spectra (Fig. 5e), which seriously signified the coordination behavior between Al 3+ and the N atoms in PI-aniline followed by the decomposing process of di-PIP. Besides, the appearance of another formyl proton H b1 manifested that the fracture of the two imine C = N bonds in the symmetrical di-PIP was proceeded inch by inch; in other word, one imine C = N bond in di-PIP was initially wrecked into one molecule PI-aniline and an intermediate containing phenanthro [9,10-d]imidazol-2-yl aldehyde called as PI-C = N-aldehyde which then absolutely decomposed into two smaller fragments including p-phthalaldehyde and another molecule PI-aniline. The results of the 1 H NMR titration illustrated that the exhaustive decomposition reaction of di-PIP happened by the promotion of Al 3+ and the symmetrical and larger conjugated system of di-PIP collapsed into conjugated fragment, which effectively led to the changes of UV-vis and fluorescence spectra. Actually, Al 3+ served as the role of Lewis acid during the procedure of the decomposition reaction, which could be supported by the changes of the H + -induced fluorescence spectra.
In succession, to validate the rationality, we attempted to synthesize the compounds PI-aniline and PI-C = N-aldehyde (the 1 H NMR spectra of PI-C = N-aldehyde was provided in Fig. S11) and their fluorescence spectra before and after the addition of 2.0 equivalent Al 3+ were contrasted with those of di-PIP as represented in Fig. 6a. Evidently, the enhancement of the emission intensity and the hypsochromic shift of the emission wavelength were distinctly perceived when Al 3+ was injected into the solutions of PI-aniline and PI-C = N-aldehyde, respectively, and the changing tendency were closely analogous to that of di-PIP, which deeply attested the decomposition of di-PIP induced by Al 3+ . In addition, the m/z signal in the ESI-MS spectra (Fig. S12) were observed at 310.1338 which was belonged to [PI-aniline + H] + (cal. 310.1306), which clearly showed the existence of PI-aniline. All results confirmed that the decomposition reaction of di-PIP happened in presence of Al 3+ . According to the above analysis, the sensing mechanism of di-PIP towards Al 3+ /H + was proposed based on the decomposition reaction followed by the coordination as shown in Scheme 2.
In addition, the response times and the fluorescence stabilities of di-PIP, PI-aniline and PI-C = N-aldehyde towards Al 3+ were also investigated in DMF/HEPES buffer solution (v/v, 7:3, 10 mmol·L -1 , pH = 7.4). As disclosed in Fig. 6b, three compounds could quickly respond Al 3+ and achieve the corresponding emission intensity at 3 min. After that, the stable intensities were retained in the testes time (120 min). Therefore, this work exploits a fluorescent sensor for highly stable, rapid and sensitive analysis of Al 3+ base on the decomposed reaction.

Detection of Al 3+ in Real Water Sample
To assess the practical applications of di-PIP, the detection of Al 3+ in the samples of the tap water, lake water and canned drinks (canned beverages, canned beer and canned milk) were performed, respectively. Each sample was spiked with the different concentrations of Al 3+ (1.0, 2.0 and 5.0 μmol·L -1 ), respectively, according to the standard addition method. As listed in Table 1, it could be found that the concentrations of Al 3+ in the canned drinks were much higher than those in tap water and lake water as expected. At the same time, the relative standard deviations (RSDs) and the recovery was conducted to evaluate the feasibility of di-PIP to detect the actual sample. According to the actual measured recovery, RSDs of detection results from di-PIP range from 0.12 to 0.41% and the spiked recoveries of Al 3+ in water samples were in the range of 95.8-102.8%, hinting

Application in Living Cells
The in vitro cytotoxicity of di-PIP was estimated for checking the biocompatibility of living HeLa cells line. The cells were treated with five different concentrations (0, 10, 20, 30, 40 and 50 μmol·L -1 ) of compound di-PIP for 24 h and after that the cell viability was assessed by an MTT assay. As shown in Fig. S13, compound di-PIP exhibited the good cell viability (around 86%) even at the highest concentration of 50 μmol·L -1 , which hinted that compound di-PIP had good biocompatibility and was advantageous for biological applications. Fluorescence microscopic study using compound di-PIP was performed to envisage the cellular uptake of the intracellular Al 3+ . The solution of Al 3+ (20 μmol·L -1 ) was injected into the cells pretreated with compound di-PIP (10 μmol·L -1 ). After the cells were treated with standard methods, fluorescence images were immediately captured with a digital camera. HeLa cells with only incubation of compound di-PIP initially showed a faint fluorescence image (Fig. 7b). After the addition of Al 3+ , the blue fluorescence signal for HeLa cells increased immediately (Fig. 7e), which was consistent with the response of the fluorescence spectral.
The intracellular H + imaging behavior of compound di-PIP was further studied on HeLa cells by fluorescence microscopy. At the acid cellular condition (pH = 4.0), the strongest blue fluorescence was observed (Fig. 7h); while, when the pH value of the cell environment was adjusted to the neutral condition (pH = 7.0), the blue fluorescence   (Fig. 7k) compared with the acid condition. However, the blue fluorescence signal was nearly invisible (Fig. 7n) under the condition of the basic cellular circumstance (pH = 9.0). The images under the acid condition displayed a strong fluorescence emission from the HeLa cell, which meant that compound di-PIP as a H + sensor had the ability to decide whether a cell had a lesion to abnormal physiological conditions.

Conclusion
In summary, an extremely sensitive fluorescence probe (di-PIP) was developed for the responding Al 3+ and H + via a collaborating mode of the decomposition reaction and the coordination effect, together with distinct aggregation behavior in DMF/HEPES. In particular, the aggregated di-PIP can simultaneously accomplish the outstanding performances to Al 3+ (LOD = 30.5 nmol·L -1 ) and H + (pK a = 5.21) with the significantly enhanced fluorescence signal (I/I 0 = 10). The sensing mechanism promoted by acid-triggered was further ascertained by 1 H NMR and MS spectra in detail, respectively. Moreover, the probe di-PIP was demonstrated to be capable to detect Al 3+ concentration in real water samples and monitor Al 3+ and H + changes in living HeLa cells with well cell permeability and low toxicity.