The development of fluorescence turn-on probe for Al(III) sensing and live cell nucleus-nucleoli staining

The morphology of nucleus and nucleolus is powerful indicator of physiological and pathological conditions. The specific staining of nucleolus recently gained much attention due to the limited and expensive availability of the only existing stain “SYTO RNA-Select”. Here, a new multifunctional salen type ligand (L1) and its Al3+ complex (1) are designed and synthesized. L1 acts as a chemosensor for Al3+ whereas 1 demonstrates specific staining of nucleus as well as nucleoli. The binding of 1 with nucleic acid is probed by DNase and RNase digestion in stained cells. 1 shows an excellent photostability, which is a limitation for existing nucleus stains during long term observations. 1 is assumed to be a potential candidate as an alternative to expensive commercial dyes for nucleus and nucleoli staining.

Recent advancement and development in fluorescence imaging techniques has witnessed tremendous upsurge in the field of metal ion sensing and live cell imaging [1][2][3][4][5] . The sensitive bio-imaging of Al 3+ in various cell lines can be helpful for understanding the fundamental mechanism of aluminium-induced human diseases. The widespread prevalence of Aluminium and its serious health threats such as Parkinson's disease, Kidney damage and Alzheimer's disease makes it imperative to develop Al(III) sensors [6][7][8][9][10] . Tolerable limit of Al 3+ in human is estimated to be around 7 mg Kg −1 per week 11 . The conventional methods for Al 3+ detection like inductively coupled plasma-atomic emission spectrometry (ICP-AES), atomic absorption spectrometry (AAS) and inductively coupled plasma-mass spectrometry(ICP-MS) are relatively expensive and suffers from lack of selectivity, sensitivity and interferences generated from the matrix 12 . However, spectrofluorimetry technique is superior in terms of quick analysis, high selectivity, sensitivity and ease of operation 13 . On the other hand, the critical role of cell nucleus in various cellular events like metabolism, reproduction and heredity highlights the importance of nucleus staining agents for determining the morphology and functionality of nucleus 14 . The existing nuclear staining agents like DAPI and Hoechst suffers from photo-bleaching and self-quenching which limits their applicability for long term observations 15,16 . The nucleolus is an important sub-nuclear structure and its dynamic morphology is indicative of pathological and physiological situations 17 . The inability of DAPI and Hoechst to stain nucleolus and limitations of extremely expensive existing nucleolus stains emphasize the requirement of a total nucleus staining dye. Hence, the development of new generation of low cytotoxic and high photostable nucleus staining agent 16 as well as selective and sensitive sensor for intracellular detection of trace amount of Al 3+ accorded significant attention [18][19][20] .
There are few reports on specific staining of nucleus by employing nano-materials 21 and Ir(III) 14 / Ru(II) 15 complexes to different cell lines. Specific staining of nucleolus is also achieved using carbon based nano-materials 22 , nanoclusters 23 , coumarin and pyronin based moieties 24,25 , Eu(III) 26 /Ir(III) 27 complexes and semiconducting quantum dots 28 . Two water soluble molecules containing N-methyl benzothiazolium moiety are recently reported which individually stains nucleus and nucleolus 29 . There exists only few commercial dye(s) which specifically stains nucleoli but has some limitations like specific storage requirements, toxicity, rare availability and expensiveness 14 . However, to the best of our knowledge, so far no report is available on specific nucleus as well as nucleoli staining using 1.
Herein, we report synthesis of multifunctional ligand L 1 and its Al 3+ complex (1) for specific nucleus and nucleoli staining along with sensitive intracellular Al 3+ detection.

Results
We recently reported synthesis of a highly selective, sensitive and reversible symmetric chemosensor H 2 L for Zn 2+ ion ( Supplementary Fig. S1) 30 . Encouraged by the results, we have slightly altered the ratio of 2,4,6-trimethylbenzene-1,3-diamine and 2-Hydroxy-1-napyhaldehyde to 1:1 ( Fig. 1) instead of 1:2 ratio, which results in another chemosensor (L 1 ) for selective and sensitive detection of Al 3+ . L 1 is characterized by elemental analysis, HRMS, NMR and further authenticated by single crystal X-ray studies. Furthermore, the reaction of L 1 with Al(NO 3 )3.9H 2 O in presence of methanol at room temperature yields complex 1 (Fig. 1). L 1 crystallize in monoclinic P2 1 /n space group (Supplementary Fig. S2 and Supplementary Table S1). The packing of L 1 reveals intermolecular H-bonding interaction and C-H···π interaction 31 forming hydrogen bonded 2D-network (Supplementary Fig. S3 and Supplementary Table S2).
Photophysical properties. The absorption spectra of L 1 shows absorption peaks at 309 nm and 360 nm corresponding to n-π * and at 419 nm corresponding to π -π * transition. Upon subsequent addition of increasing concentration of Al 3+ ions, the absorbance intensity gradually enhanced at 419 nm and decreased at 309 and 360 nm having one isosbestic point at 391 nm ( Fig. 2(a)) with change in color from yellow to blue under UV light. However, addition of varying concentrations of other metal ions viz.; Ag + , Na + , K + , Li + , Mn 2+ , Cd 2+ , Hg 2+ , Mg 2+ , Ca 2+ , Cr 3+ , Ni 2+ , Co 2+ , Cu 2+ , Zn 2+ , Pb 2+ , Fe 2+ and As 3+ to the L 1 solution results in no significant change in the absorbance and color ( Fig. 2(b), Supplementary Fig. S4).
The emission studies of L 1 at excitation wavelength 309 nm, shows weak fluorescence at 341 nm which is attributed to the isomerization of the imine bond (− HC= N) and excited state intramolecular proton transfer (ESIPT) from the −OH group to the imine nitrogen 30 . The titration of L 1 with Al 3+ shows gradual fluorescence enhancement with a bathochromic shift (red shift) of 2 nm having maximum fluorescence intensity at 343 nm. Also, emergence of two more emission peaks at 351 and 448 nm is observed (Fig. 2(c)).
The probable reason for the fluorescence switch on towards Al 3+ may be due to (i) the photoinduced electron transfer (PET) suppressed by Al 3+ coordination and (ii) new band in absorption spectrum 32 . PET shows changes in emission intensities with some or no spectral shifts. In our case, the spectral shift is of 2 nm in emission maxima during the fluorescence titration. The shorter fluorescence lifetime and lower quantum yield also support the operation of PET process 33,34 . The schematic representation of working mechanism is shown in Supplementary  Fig. S5. Thus, the results obtained from absorption and emission studies validate that the sensing behavior of L 1 is specific to Al 3+ ions. The binding affinity of L 1 with Al 3+ is found to be 1.18 × 10 5 M −1 and 1.21 × 10 5 M −1 by absorption and emission methods, respectively by using Benesi-Hildebrand equation 35,36 ( Supplementary  Fig. S6). The limit of detection (LOD) is found to be 86 nM. In general, chemosensor suffer with long response time [37][38][39] . In our case, the binding of Al(III) towards L 1 is found to be very fast. After adding the aqueous solution of Al(III) to L 1 , the fluorescence intensity at 343 nm reached maximum within 120 s ( Supplementary Fig. S7). Selectivity, reversibility and mode of binding. Further, the selectivity test is performed using various metal ions with L 1 which shows that only Al 3+ ion exhibits fluorescence turn on ( Fig. 2(d) and Supplementary  Fig. S8(a)). A competitive selectivity test was also performed for various metal ions in presence of Al 3+ in 1:4 ratio which shows that the presence of other metal ion do not affect the selective Al 3+ sensing behavior of L 1 (Supplementary Fig. S8(b)). The interference of DNA in the probe's ability to detect Al(III) was examined and it was found that the DNA does not interfere with the sensing property of L 1 (Supplementary Fig. S9). For validating the reversibility, the chemosensor L 1 is studied by employing EDTA as a chelating agent. The increased fluorescence intensity of L 1 with Al 3+ ion was quenched by addition of EDTA ( Fig. 3 and Supplementary Fig. S10).
Further, to understand the mode of complexation of L 1 with Al 3+ ion, 1 H NMR titration is performed in DMSO-D6 and D 2 O solvents. On subsequent addition of Al 3+ to L 1 , the napthyl -OH peak (15.34 ppm) decreases  and finally at 0.8 eq. of Al 3+ , − OH peak was completely diminished which indicates complex formation of L 1 with Al 3+ (Supplementary Fig. S11). The stoichiometry of 1 is also confirmed by Job's method which indicates 1:1 complex formation of L 1 with Al 3+ (Supplementary Fig. S12).
The Cytotoxicity assessment. To understand the practical ability of L 1 and 1 as a sensor for biological systems, it is obvious to study the cytotoxicity. The cytotoxicity studies are carried out using MTT assay 40 . Human prostate cancer cell line DU145 is incubated with L 1 and 1 (0.01-100 μ M) for 24 h and the cell viability is quantified. The results summarized in Supplementary Fig. S15, indicates the cytocompatibility of the probe L 1 and 1. The cell viability is more than 80% at a high concentration of 100 μ M which represent very low cytotoxicity of L 1 and 1.
Intracellular Al 3+ sensing. The excellent sensing response and low cytotoxicity of L 1 , prompted us to explore for intracellular Al 3+ sensing using confocal laser scanning microscopy. In this regard, two cell lines, DU145 and MCF-7 were first treated with 10 μ M of L 1 and incubated for 45 minutes at 37 °C. These L 1 treated cells are subjected to confocal microscopy and were excited using lasers wavelength of 405nm and 488nm. As depicted in Fig. 4, the cells treated with L 1 resulted in a very weak intracellular fluorescence with blue and green colored signal at different excitation with uneven distribution in the cells. On addition of 10 μ M of Al 3+ to these cells, a considerable enhancement in blue and green colored fluorescent signal was observed. This confirms the ability of L 1 to probe intracellular Al 3+ and a possible intracellular multicolor bioimaging potential of 1. Moreover, to our surprise, it is observed that the fluorescence signal becomes confined to the nuclei after some time (within 10 minutes), which hinted towards the possible nuclei specificity of 1, which was further explored. The wavelength tuned emission spectra of 1 was recorded, which shows bathochromic shift with increase in excitation wavelength ( Supplementary Fig. S16) resulting into multicolor bioimaging property. The similar wavelength tuned emission behavior of fluorophores is shown in previous reports 5, 41-43 . Nucleus and nucleolus specific staining. The interesting intracellular sensing response by L 1 and hint of nucleus targeting by 1, provoked us to study the bioimaging potential of 1. The breast cancer cell line MCF-7, cervical cancer cell line HeLa and skin melanoma cell line A375 is used for comparison between the staining ability of a standard nucleus stain and 1. These cells are first stained with Hoechst 33342 and the emission is recorded at λ ex = 405 nm Further, 20 μ M of 1 is added and after 5 minutes incubation, the intracellular emission is recorded at λ ex = 488 nm. The blue fluorescence from nucleus specific Hoechst dye is completely overlapped by the nucleus staining green fluorescence of 1 (Fig. 5, Supplementary Fig. S17). The strong overlap validated the potential of 1 as a specific nucleus stain. The Pearson's colocalisation coefficient was obtained to be R r = 0.75 ( Supplementary  Fig. S18), which is in close agreement with recently reported nucleic acid selective probes 29 .
It is worth to mention that the dual staining experiment was performed in serial "staining and imaging" mode. First, the cells were stained with Hoechst and the fluorescence emission was recorded in blue channel in absence of 1 to avoid excitation of 1 in blue channel. Further, 1 was added without disturbing the cell position and the emission was recorded in green channel to avoid any interference with Hoechst fluorescence.
A closer look at the comparative nucleus stain of Hoechst and 1, exhibits the dominance of 1 in not only staining nucleus but also nucleolus, which remains unstained in case of Hoechst as demonstrated in Fig. 6. To study intracellular photostability of 1, the MCF-7 cells are stained with 1 and a dual channel emission video is recorded upto 450 scans, where laser of 405 and 488 nm is triggered at every scan (Supplementary Video 1). The video shows no obvious photobleaching, which suggests high photostability and feasibility of 1 as an efficient and suitable candidate for total nucleus stain. The fluorescence stability of 1 is also tested under continuous UV illumination up to two hours and varied temperature in range of (− 80 °C to 90 °C), which shows high photostability  Fig. S19(a,b)). The effect of pH values on the fluorescence intensity of L 1 and 1 is also examined and shown in Supplementary Fig. S19(c,d)).

DNase and RNase digestion test.
After exploring the specific nucleus and nucleolus staining capability of 1, it is anticipated that 1 must specifically bind to nucleic acid. The mechanistic pathways and reason of specific nucleus and nucleolus staining by 1 is hypothesized by binding of 1 with DNA and RNA in the nucleus and nucleolus, respectively. The intracellular nucleic acid binding is explored by a deoxyribonuclease (DNase) and ribonuclease (RNase) digestion test 44 and the results can be visualized in Fig. 7. HeLa and MCF-7 cell lines were stained with 1 and subsequently treated with DNase and RNase. Since, DNase hydrolyze only DNA substrate in the cells,  the fluorescence of the nucleus is vanished while the nucleolus is found glowing in DNase treated cells, which confirmed the interaction of 1 with DNA present in the cell nucleus as evident in Fig. 7. Similarly, the RNase hydrolyze only the RNA substrate in the cells, resulting in no fluorescence from the nucleolus and the weak cytoplasmic fluorescence has also vanished while the nucleus was glowing in RNase treated cells, which approved the interaction of 1 with RNA present in the cells. The DNA binding of 1 is also validated by EtBr displacement assay and the results are shown in Supplementary Fig. S20. These observations supported the hypothesis of nucleic acid binding of 1 resulting into specific cellular localization and nucleus and nucleoli staining.
The photostability of 1 was studied in live cells and was compared with Hoechst and FITC. The intensity of FITC and Hoechst was reduced to 30% of the initial intensity after 1500 scans while the fluorescence from 1 was almost consistent and was clearly visible upto 3600 scans studied. (Supplementary Fig. S21).
The existence of RNA selective compounds is limited. One of the reasons is better affinity of small molecules towards double stranded DNA than single stranded RNA 45 . Also, the complicated environment of living cells, sometimes results in non-specific binding and imaging. SYTO RNA-Select is the only commercially available live cell RNA staining dye. Another existing limitation comes from photo-bleaching and self-quenching of nucleus and nucleolus staining agents. The present work successfully overcomes these limitations and reports a nucleic acid specific compound which stains both the nucleus and the nucleolus with better photostability.

Discussion
In summary, we for the first time report the specific staining of nucleus and nucleoli by 1, which is prepared in a facile manner. The ligand L 1 involved in synthesis of 1 also promotes selective and sensitive sensing of Al 3+ . 1 is explored for specific nucleic acid binding using DNA and RNA digestion and employed as a specific nucleus and nucleolus staining agent. Our results are expected to inspire replacement of existing expensive and sensitive nucleic acid specific dyes. Use of 1 for both nucleus and nucleoli staining can be of great scientific interest.

Methods
Synthesis of chemosensor (L 1 ). 2,4,6-trimethylbenzene-1,3-diamine (0.189 g, 1mmol) dissolved in dry methanol (10ml) was added to a solution of 2-Hydroxy-1-napthaldehyde (0.172 g, 1mmol) in methanol (15ml). The content of the flask heated under reflux for 3 h. A yellow oily compound was formed which is monitored by TLC. Solvent was evaporated on vacuo and washed with hexane. A yellow solid was obtained which was dried under vacuum. Needle shaped yellow crystals were suitable for single crystal X-ray analyses were obtained by recrystallization from methanol-chloroform (2:1) within two days. Characterization data. 1   Nucleus specific staining. Breast cancer cell line MCF-7, cervical cancer cell line HeLa and skin melanoma A375 were seeded in confocal dishes. Adherent cells were stained with Hoescht 33342 (50μ g/mL) and incubated for 15 minutes at at 37 °C in 5% CO 2 . The cells were rinsed with PBS thrice and was visualised under confocal microscope at an excitation of λ ex . = 405nm and λ ex . = 488nm. Without changing the positioning of cells under microscope, cells were stained with 1 (30 μ M) and kept for 10 minutes. Further the cells were visualised using confocal microscope at λ ex . = 488nm.
DNase and RNase digestion tests. MCF-7 and HeLa cell lines were first seeded in confocal dishes. The cells were fixed by cold methanol (HPLC grade) for 1 min. Further, the cells were treated with 1% Triton X-100 to make the cell membrane permeabilized. The cells were rinsed thrice with 1X PBS to remove the leftover detergent. Afterwards, cells were stained with 30 μ M 1. After staining cells were rinsed with PBS thrice. Blank PBS (as control experiment), 50 mg/ml DNase or 75 mg/ml RNase was added into different confocal dishes containing cells and kept in an incubator at 37 °C in 5% CO 2 for 3 hr. Further, the cells were subjected to confocal microscopy. Physical measurements. 1 H NMR (400 MHz), and 13 C NMR(100 MHz) spectra were recorded on the Bruker Avance (III) instrument by using CDCl 3 . 1 H NMR chemical shifts are reported in parts per million (ppm) relative to the solvent residual peak (CDCl 3 , 7.26 ppm). 13 C NMR chemical shifts are reported relative to the solvent residual peak (CDCl 3 , 77.36 ppm). IR spectra [4000− 400 cm − 1 ] were recorded with a Bio-Rad FTS 3000MX instrument on KBr pellets. Elemental analyses were carried out with a Thermo-Flash 2000 elemental analyser. Spectrophotometric measurement were performed on a varian UV-vis spectrophotometer (model: Cary 100) using a quartz cuvette with a path length of 1cm. The excitation and emission slits were 5/5 nm for the emission measurements. The mass spectra were recorded on Brucker-Daltonics, micrOTOF-QII mass spectrometer. The life time measurement recorded on TCSPC system from Horiba Yovin (Model: Fluorocube-01-NL). The samples were excited at 375 nm using a picosecond diode laser (Model: Pico Brite-375L) with the repetition rate of 5 MHz. The signals were collected at magic angle (54.70) polarization using a photomultiplier tube (TBX-07C) as Scientific RepoRts | 6:34807 | DOI: 10.1038/srep34807 detector, which has a dark counts less than 20 cps. The instrument response function of our setup was 150 ps. The data analysis was performed using IBH DAS (version 6, HORIBA Scientific, Edison, NJ) decay analysis software. NMR spectra of L 1 and 1 are depicted in Supplementary Figs S22-S25.