Fluorescent dyes have gained increasing attention over the past few decades due to their varieties of application in the fields of nanoscience [1], optoelectronic materials [2], and biological chemistry etc. [3]. The great progress made the areas of bioimaging and sensing have created a demand for biocompatible fluorescent dyes with absorption and/or emission in the far red and nearinfrared (NIR) region (650-900 nm) due to that NIR dyes enable lower photodamage, minimum fluorescence background, deep photon penetration in tissue and less light scattering [4]. Among the various of fluorescent dyes, polymethine cyanine dyes have played a vital role in the field of photography and other sophisticated arenas of dye application since 1856 [5]. Heptamethine cyanine as a frequently used NIR dyes, employed as fluorescent probes, labels of biomolecules in vivo as well as photosensitizer for photodynamic therapy and photothermal therapy have attracted immense interest in recent years [6].

However, most of the heptamethine cyanine dyes show the persistent and remarkable disadvantage that their Stokes shift is less than 25 nm. A small Stokes shift causes self-quenching and measurement error due to the interference from excitation light and scattered light [7]. Thus, heptamethine cyanines with large Stokes shift are attractive and valuable contrast agents. A great deal of efforts has been paid for developing heptamethine cyanine dyes with a larger Stokes shift. For example, Peng and co-workers developed a type of heptamethine cyanine dyes containing robust C-N bond with a Stokes shift around 140 nm and strong fluorescence [8]. However, it is still a challenge to explore the novel cyanine backbone with larger Stokes shift.

It is found that the hemicyanines like the structure of O-Hcy in Scheme 1a were obtained after heptamethine cyanine IR-780 was decorated with phenol derivatives [9, 10]. As the visible and NIR emission fluorophores, these xanthene-based hemicyanines have been widely applied in bioimaging [11-20]. Owing to the influence of functional groups (FG) on the backbone of O-Hcy, these hemicyanines showed broader ranges of UV-vis absorption (around 580-770 nm) and fluorescent emission (around 640-780 nm), leading to more obvious Stokes shifts of 20-125 nm. However, the dye O-Cy with the backbone of IR-780 only displayed the Stokes shift of 20-25 nm (Scheme 1a) [21]. Unexpectedly, the N-Hcy derivatives in which the xanthene moiety was replaced by tetrahydroacridine ring showed a Stokes shift of around 110 nm despite it possessed a shorter absorption and fluorescent emission in Scheme 1b [9]. In respect of the donating electron capability of NH atom, it is assumed that incorporating the tetrahydroacridine ring with IR-780 to form N-Cy probably led to a unique optical property (Scheme 1b). Although xanthene-fused heptamethine cyanine derivative (O-Cy) had a smaller Stokes shift, it is possible that the tetrahydroacridine-conjugated heptamethine cyanine (NCy) possesses a larger Stokes shift when an additional moiety is installed on the R' site of N-Cy.

To ensure that the absorption and emission bands are in NIR region, phenylethynyl naphthalimide would be introduced to prolong the conjugation of N-Cy [22]. Furthermore, by attaching a targeting moiety, fluorescent dyes are able to recognize specific acceptors in cells [23-26]. Herein, a morpholine group was employed to target lysosome by conjugating a tetrahydroacridine with heptamethine cyanine dye (N-CyNp). In which, the morpholine-modified naphthalimide linked the tetrahydroacridine moiety via a conjugated acetylene bridge. As shown in Scheme 2, the commercially available IR-780 iodide was employed to label with morpholine-containing phenylethynyl naphthalimide that was reported by our group, to afford the targeting probe N-CyNp.

The optical properties of dye N-CyNp (10 μmol/L) in HEPES buffer (10 μmol/L, pH 7.4) containing 10% DMSO was investigated by UV-vis absorption and fluorescence spectroscopies. As shown in Fig. 1a, N-CyNp displayed well-one sharp absorption peak at 369 nm (ε = 2.20 ×10⁴ mol L^-1 cm^-1) and a broad absorption peak at around 620 nm (ε = 1.08 × 10⁴ mol L^-1 cm^-1). The fluorescence spectrum of N-CyNp (Φ_f = 0.01, using ICG as a fluorescence standard) demonstrated that N-CyNp had a Stokes shift of 165 nm [10]. The heptamethine cyanine dyes with such a large Stokes shift are very rare. The fluorescence changes of N-CyNp were also detected in buffer with different pH values from 2 to 12 (Fig. S1 in Supporting information). The result showed that N-CyNp had a stable fluorescence in a wide pH range (3-9), which was showed in Fig. 1b. This pH range covered the pH condition of lysosome, inside, indicating that N-CyNp was a promising lysosome-targeting indicator in bioimaging.

To obtain better insight into the relationship of structure and excited states, the time-dependent density functional theory (TD-DFT) calculations at the B3LYP/6-31G^* level were carried out using the Gaussian 09 program (Fig. S2 in Supporting informaiton). The optimized conformation of N-CyNp showed that the cyanine moiety possesses a twist conformation while the naphthalimide moiety was coplanar with tetrahydroacridine component. An intense transition for N-CyNp was predicted at about 570 nm with larger oscillator strength of 0.5639 (HOMO-1 → LUMO). The HOMO-1 was mainly assigned to the heptamethine cyanine backbone while the electron density of LUMO orbital was mainly located the phenylethynyl naphthalimide, implying that dye NCyNp was fluorescent as a result of an obvious electron transfer. The result was well in agreement with the experimental measurement. Moreover, the lower band gap of HOMO and LUMO energy orbitals indicated that N-CyNp had an emission at long wavelength. Accordingly, the introduction of the naphthalimide moiety can change the distribution of orbit coefficient, which may be one of the reasons of large Stokes shift for N-CyNp.

As described above, N-CyNp had good water solubility and favorable optical properties under physiological conditions, which enables its further application in living cells. HeLa cells and U87 cells were cultivated in DMEM medium supplemented with 10% (v/v) calf serum, penicillin (100 U/mL) and streptomycin (100 μg/mL) and maintained in a humidified atmosphere containing 5% CO2 at 37 ℃. As shown in Fig. 2, HeLa cells and U87 cells were incubated with N-CyNp (10 μg/mL) at 37 ℃ for 1 h, respectively. A weak red emission was monitored by laser confocal fluorescence microscopy (LCFM). The result indicated that N-CyNp was capable of permeating into cells [27-33]. Moreover, it caused a remarkable strong red fluorescence when the cells were incubated with N-CyNp (10 μg/mL) at 37 ℃ for 2 h. The bioimaging experiment strongly implied that N-CyNp could be applied in the living cells.

HeLa cells were incubated with N-CyNp (10 μg/mL) for 0.5 h, followed by incubation with Lyso-Tracker Blue (10 μmol/L) for additional 30 min. The fluorescence change was monitored by LCFM. The morpholine moiety on N-CyNp can recognize lysosomes in living cells, which was validated by the co-localization experiment with a commercially available lysosome probe Lyso-Tracker blue. As shown in Fig. 3, the cell image with N-CyNp (Ex = 655 nm) in red channel was well in agreement with that of the Lyso-Tracker Blue (Ex = 405 nm) in blue channel. After incubated for another 1 h, the cell images still matched very well. These results greatly implied that N-CyNp can efficiently localize in lysosome of living Hela cells.

To validate the efficiency of N-CyNp as a NIR contrast agent for in vivo imaging, the mice were intravenously injected with N-CyNp buffer solution (100 μL, 10 μg/mL) and then observed by NIR fluorescence imaging system. As shown in Fig. 4, the strong fluorescence signal appeared in the whole body of the mice after 0.5 h of the injection and last for 2 h, which indicated that N-CyNp was an effective NIR contrast agent for in vivo fluorescence imaging. As the detection time prolonged, it was found that the fluorescence signal decreased dramatically after 4 h and disappeared completely in 24 h. After 24 h, no obvious fluorescence signals was observed in the main organs such as intestine, liver, spleen, kidney, heart and lung. These results suggested that the NCyNp can be utilized as a promising fluorescence probe for NIR fluorescence imaging.

In summary, a new lysosome-targeting near-infrared fluorescent dye with large stokes shift based on heptamethine cyanine backbone was reported. The dye was successfully applied in bioimaging of living cells, mice and tissues. Introducing phenylethynyl naphthalimide moiety to tetrahydroacridine-fused cyanine played an important role for increasing the Stokes shift. The decoration based on tetrahydroacridine-fused heptamethine cyanine probably provided a new strategy for the design of cyanine dyes with large Stokes shift.

Acknowledgments

The authors acknowledge financial support from National Natural Science Foundation of China (Nos. 21676113, 21402057, 21472059, 81671803), Youth Chen-Guang Project of Wuhan (2016070204010098), the 111 Project B17019 and the MinistryProvince Jointly Constructed Base for State Key Lab-Shenzhen Key Laboratory of Chemical Biology, Shenzhen. This work is also supported by Self-determined Research Funds of CCNU from the colleges' basic research and operation of MOE (No. CCNU16A02004).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cclet.2017.07.004.