Charge control of fluorescent probes to selectively target the cell membrane or mitochondria: theoretical prediction and experimental validation

Xiaoyan Zheng*a, Dong Wang*b, Wenhan Xuc, Siqin Caod, Qian Peng*e and Ben Zhong Tang*c
aBeijing Key Laboratory of Photoelectronic/Electrophotonic Conversion Materials, Key Laboratory of Cluster Science of Ministry of Education, School of Chemistry and Chemical Engineering, Beijing Institute of Technology, Beijing 100081, China. E-mail: xiaoyanzheng@bit.edu.cn
bCenter for AIE Research, College of Materials Science and Engineering, Shenzhen University, Shenzhen 518060, China. E-mail: wangd@szu.edu.cn
cDepartment of Chemistry, Hong Kong Branch of Chinese National Engineering Research Center for Tissue Restoration and Reconstruction, Institute of Molecular Functional Materials, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China. E-mail: tangbenz@ust.hk
dDepartment of Chemistry, The Hong Kong University of Science and Technology, Hong Kong, P. R. China
eKey Laboratory of Organic Solids, Beijing National Laboratory for Molecular Science (BNLMS), Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, P. R. China. E-mail: qpeng@iccas.ac.cn

Received 13th June 2019 , Accepted 16th July 2019

First published on 17th July 2019


To achieve efficient and precise design of fluorescent probes, unraveling the intrinsic mechanism of their selectivity on different subcellular organelles in cell imaging is important. Using a theoretical protocol combining large-scale molecular dynamics simulations and the hybrid QM/MM model, we explored the mechanism of the strikingly different fluorescence imaging behaviors of two amphiphilic AIEgens with quite similar chemical structures. We proposed that the hydrophobic moiety and the pyridine group of AIEgens manipulate the emission behaviors, and the charged headgroups control the permeation ability through the cell membrane. Therefore, we proposed a molecular design strategy of AIEgen-based fluorescent probes to selectively target the cell membrane or mitochondria, and designed and synthesized four AIEgens. The cell imaging experimental results successfully confirmed the theoretical prediction, which would open an efficient shortcut to design AIEgen-based fluorescent probes to selectively target the cell membrane or mitochondria and lays a solid foundation for the rational design of advanced cell imaging materials.



New concepts

Fluorescent probe discovery still proceeds largely through expensive trial-and-error experiments, because the intrinsic mechanism of fluorescent probes at the molecular level is not unraveled. Exploring material design strategies by computational methods to complement and assist experimental material design and synthesis is an effective way. In this manuscript, we report a strategy for designing novel fluorescent probes which can selectively target the cell membrane or mitochondria, by multiscale computational modeling. The innovative concept is realized by combining large-scale molecular dynamics simulations and the hybrid QM/MM model protocol, firstly to demonstrate the conformations of fluorescent probes in a complex cell membrane environment and then to characterize the fluorescent emission of the probes in this environment. We identify the specific roles (permeation ability control or fluorescent emission control) of different functional groups of the fluorescent probes in their targeting process, and then we design and synthesize fluorescent probes to selectively “light-up” the cell membrane or mitochondria. Validated by experiments, the designed fluorescent probes show excellent bioimaging behaviour with high selectivity on different subcellular organelles. This demonstrates the power of the proposed concept, which should be applicable to the design of a wide range of clinical materials.

Visualizing subcellular organelles with high resolution in living cells is important for characterising biological processes, diagnosing pathogenesis and investigating pharmacodynamics. Fluorescent technology is a good choice for realizing this task and is nowadays undergoing flourishing development, because of its super sensitivity, fast response, simple operation, and in situ workability.1–5 Of particular interest is luminogens with aggregation-induced emission (AIE) characteristics, which open an avenue to an array of possibilities for high-tech innovations of fluorescent probes,6,7 mainly benefiting from their superior features, such as great tolerance for any concentrations, high fluorescence efficiency in the aggregate state, extraordinary photostability, large Stokes shift and great potential as “light-up” probes.8,9 AIE luminogens (AIEgens) have been utilized as probes for detecting various subcellular organelles, including the cell membrane,10 cytoplasm,11 lysosomes,12 lipid droplets,13 mitochondria,14 RNA,15,16 and so on.17–20

Nonetheless, the development of AIEgen-based fluorescent probes for cell imaging requires the optimization of many factors, such as fluorescence emission wavelength, membrane permeability, solubility and stability and so on. These factors could not be optimized separately, because they interplay with each other, which makes the rational design of fluorescent probes a great challenge. From the experimental point of view, designing new AIEgen-based fluorescent probes nowadays is achieved by trial-and-error. No clear interpretation of the intrinsic working mechanism for fluorescent probes targeting specific subcellular organelles has been found, because of the limited spatial-temporal resolution in experiments and the high sensitivity of chemical structures of fluorescent probes to biological surroundings. This can be exemplified by two representative amphiphilic AIEgens, which have quite similar chemical structures but feature distinguished cell imaging characteristics.14,21 The two AIEgens are (E)-4-(2-(5-(4-(diphenylamino)phenyl)thiophen-2-yl)vinyl)-1-methyl-pyridin-1-ium (TTPy) and (E)-4-(2-(5-(4-(diphenylamino)phenyl)-thiophen-2-yl)vinyl)-1-(3-(trimethylammonio)propyl)pyridin-1-ium (TTVP). As shown in Fig. 1a, both TTPy and TTVP are amphiphilic and consist of the same hydrophobic moiety, including a triphenylamine group, thiophene group and ethylenic bond, termed as the TTV moiety (green). The only difference in the chemical structures between TTPy and TTVP is in their hydrophilic part: TTPy contains only one positively charged pyridine group (yellow), while TTVP bears two positively charged groups, including a pyridine group (yellow) and a quaternary 3-(trimethylammonio)propyl group (purple). Strikingly, the seemingly subtle difference in the chemical structures of TTPy and TTVP leads to dramatically different behaviours in cell imaging.14,21 That is, TTPy can penetrate the cell membrane and target mitochondria with excellent fluorescence imaging,14 while TTVP is able to specifically target and “light-up” the cell membrane.21 Purely considering the chemical structure of a single AIEgen, it seems difficult to explain why TTPy exhibits totally different targeting behaviour from TTVP, because the elaborate interplay of different intermolecular interactions among the AIEgen, lipid membrane and solvents should play important roles in the cell imaging process. In addition, the permeability coefficient of the AIEgen through the cell membrane directly determines the selectivity of the fluorescent probe to the cell membrane or other subcellular organelles, because passing through the cell membrane is a prerequisite for targeting other subcellular organelles.


image file: c9mh00906j-f1.tif
Fig. 1 (a) Chemical structures of two AIEgens TTPy and TTVP, with key dihedral angles labelled. (b) Snapshot and the partial density profiles of the equilibrated lipid membrane. The lipid membrane system is divided into four regions and labelled by I, II, III and IV, respectively. The phosphate and choline groups of lipids are shown as blue and orange spheres, respectively. The carbonyl group and the hydrocarbon tails of the lipids are shown as green and cyan cylinders, respectively. The water molecules are shown in red (oxygen) and white (hydrogen) lines. (c) The calculated permeability coefficients of two AIEgens. (d) Free energy profiles of TTPy and TTVP. The standard derivations are also shown. The free energy was set to zero in bulk water. Four divided Regions (I–IV) of a lipid membrane were labelled by dashed grey lines.

In this contribution, we adopt a theoretical protocol combining large-scale molecular dynamics (MD) simulations and the hybrid QM/MM model to unravel the intrinsic mechanism of the strikingly different fluorescence imaging behaviours of two amphiphilic AIEgens (TTPy and TTVP) with quite similar chemical structures. Our protocol successfully maps the interactions between the AIEgen and lipid membrane at the microscopic level to the cell imaging selectivity of AIEgens at the mesoscopic level. Based on the theoretical calculation, we proposed a rational design strategy of AIEgen-based fluorescent probes to selectively target the cell membrane or mitochondria, and further designed and synthesized four AIEgens. The obtained cell imaging experimental outcomes of four newly designed AIEgens are in good accordance with the theoretical prediction.

To assist in the description of free energy profiles, we set up a four-region membrane model22 according to the partial density profiles of the lipid membrane (Fig. 1b). The membrane center is taken as a reference with z-position defined by z = 0 Å. Considering the symmetric property of the lipid membrane, the discussion is focused on half of the system. Region I (z = 0–8 Å) only contains the hydrophobic lipid tails and has the lowest partial density of the whole system. Region II (z = 8–16 Å) is the interfacial area between the hydrophobic tails and the initial part of the hydrophilic headgroups, in which the density of the lipid tails drops remarkably while the density of the carbonyl group, phosphate group, choline and water molecules increases obviously. In this region, the density of the whole system reaches the maximum. Region III (z = 16–25 Å) is the interfacial area between the lipid headgroups and the bulk water, including negatively charged phosphate groups, positively charged choline and the penetrated water molecules. Region IV (z > 25 Å) is primarily composed of bulk water and a small portion of lipid headgroups.

The free energy profiles of TTPy and TTVP through the lipid membrane are illustrated in Fig. 1d. The two profiles exhibit similar shapes. Upon entry to the lipid membrane, the free energy profiles drop sharply and reach the minima in Region II. TTPy has a global minimum with free energy −28.73 kJ mol−1 at z = 12 Å, while for TTVP the corresponding free energy is −15.63 kJ mol−1 at z = 14 Å, indicating that both AIEgens prefer the amphiphilic environment in Region II, which is well in accordance with their amphiphilic nature. Crossing the hydrophobic core of the lipid membrane, TTPy needs to overcome a free energy barrier of 19.04 kJ mol−1, while for TTVP the free energy barrier is 30.44 kJ mol−1. Remarkably, at the membrane center, the free energy barrier of TTVP is significantly larger than that of TTPy, indicating that TTPy would have better translocation ability than TTVP. To characterize the overall translocation ability of the two AIEgens through the lipid membrane more directly, we further calculate the permeability coefficients of the two AIEgens, shown in Fig. 1c and Fig. S3, S4 in ESI. It is clear that the permeability coefficient of TTPy (7.685 cm s−1) is ca. 213 times larger than that of TTVP (3.612 × 10−2 cm s−1), implying that TTPy could translocate through the lipid membrane and target mitochondria, while TTVP can be buried in and target the cell membrane, consistent with the experimental observation.14,21

To unravel the intrinsic mechanism of the different translocation ability of TTPy and TTVP with similar chemical structures, we deeply analyzed the intermolecular interactions among the AIEgen, lipid membrane and water molecules, and the deformation of the membrane structure caused by AIEgen translocation. Fig. 2 illustrates the snapshots of two AIEgens at both the global minima and the lipid membrane center respectively. Here we define the distance between two peaks of the phosphate group distribution profile as the thickness of the lipid membrane in each system, termed as dP–P.


image file: c9mh00906j-f2.tif
Fig. 2 Snapshots of TTPy and TTVP in the lipid membrane: (a) TTPy at window z = 12 Å, (d) TTPy at the membrane center (z = 0 Å); (g) TTVP at window z = 14 Å and (j) TTVP at the membrane center. The TTV moiety, the pyridine group and quaternary 3-(trimethylammonio)propyl group of AIEgens are shown in green, yellow and purple spheres, respectively. (b, e, h and k) The zoomed-in snapshots at each z-position and water molecules within 10 Å around each AIEgen. (c, f, i and l) Distribution of the phosphate groups of the pure membrane and the membrane with AIEgen inserted at the selected z-position. The membrane thickness is labelled by dP–P.

At the global minima of the free energy profiles, the positively charged headgroups of both AIEgens locate in the headgroup region of the lipid membrane and are surrounded by the negatively charged phosphate groups, while the hydrophobic TTV moieties of the two AIEgens lie inside the hydrophobic core of the lipid membrane (Fig. 2b and h). Due to the higher positive charges of the TTVP headgroup, it attracts more negatively charged phosphate groups around itself than TTPy to keep stable in the lipid membrane. This causes the membrane thickness dP–P with TTVP embedded increase by 1.9 Å (Fig. 2i), while the increment for TTPy is only 0.8 Å (Fig. 2c). Thus, TTVP at the minimum has higher free energy than TTPy (Fig. 1d). At the membrane center, both AIEgens are buried in the hydrophobic core of the lipid membrane (Region I). The positively charged headgroups of AIEgens are unfavorable in the hydrophobic environment and attract negatively charged phosphate groups approaching themselves by electrostatic interactions to keep stable, which increases the membrane thickness dP–P by 2.0 Å for TTVP (Fig. 2l) and 0.4 Å for TTPy (Fig. 2f). The induced membrane deformation requires additional entry of water molecules for charge compensation; thus, we observe that water defects accompany AIEgen permeation (Fig. 2e and k). The number of water molecules that enter into Region I for TTVP is two times larger than that of TTPy, significantly resulting in the much higher free energy barrier of TTVP at the membrane center than TTPy. As shown in Fig. S5 in the ESI, there are 4 and 8 water molecules accompanying TTPy and TTVP entry into the hydrophobic core of the membrane respectively, to compensate for the large energetic penalty of letting charged AIEgens access the purely hydrophobic core (Region I). In addition, the water defects are always close to the polar charged headgroups of both AIEgens (Fig. S6, ESI).

Above all, for both AIEgens, entry into the lipid membrane from bulk water is favorable, not only because of the attractive electrostatic interactions between the positively charged headgroups of the AIEgens and the negatively charged phosphate groups of the lipid membrane, but also due to the hydrophobic interactions between the TTV moiety of the AIEgens and the hydrophobic tail groups of the lipid membrane. Moving closer to the membrane center, contacts with charged groups of the AIEgens become much more difficult. The induced negatively charged phosphate groups from the lipid membrane surface and the accompanying water defects largely deform the membrane structures and lead to the sharp increase of the free energy of the two AIEgens at the membrane center. In particular, the free energy increment at the membrane center of TTVP (with two positive charges) is significantly larger than that of TTPy (with only one positive charge), and therefore the resulting free energy barrier of TTVP at the membrane center is much larger than that of TTPy; this is the intrinsic reason which causes the much lower permeability coefficient of TTVP through the lipid membrane.

The sensitivity of the fluorescent probe in both fluorescence color and quantum efficiency at different environments is the most key factor for its targeting application in living cells. The emission spectra and excited state radiative rate constants of TTVP in the lipid membrane were performed by the QM/MM model based on the randomly extracted conformations from MD trajectories at the window (z = 14 Å) with the largest partition (Fig. 3a and Fig. S7b, ESI). Because of the densely packed nature of the lipid membrane, the embedded TTVP is quite rigid, reflected by the well-aligned 40 conformations randomly extracted from the MD trajectory at window z = 14 Å, (see Fig. S7b and S8 in ESI) and their small fluctuations in the backbone length (less than 1 Å, see Fig. S9, ESI). Here five conformations were chosen to investigate the photophysical properties of TTVP in the lipid membrane. The calculated geometrical modifications, electronic transition properties and reorganization energies of TTVP in the excited-state decay processes are given in Fig. 3b–g and Tables S1–S5 in the ESI, as well as those of TTVP in dilute THF solution. The backbone of TTVP becomes more linear when moving from the solution into the lipid membrane, therefore, both the electron and hole natural transition orbitals (NTOs) of TTVP in the solution are localized in the central well-conjugated region, while the electron NTO of TTVP in the lipid membrane can delocalize over the whole TTV moiety and pyridine group (see Fig. 3b and c), which leads to the slightly larger transition dipole moment of 16.9–17.2 Debye in the lipid membrane than that of 16.2 Debye in solution (Table S3, ESI). Besides, in both cases, the quaternary 3-(trimethylammonio)propyl group does not contribute to NTO; thus, there is no effect on the emission property (Fig. 3b and c). Relative to TTVP in solution, the TTVP in the lipid membrane shows stronger rigidity with small geometric changes between the S1 and S0 states (Fig. 3d, e and Tables S1, S2, ESI). For example, the largest modification of the dihedral angles (ΔD4) is decreased from ∼24.8° to less than 4.7°, and the largest change of bond length (ΔB7) is reduced from 0.07 Å to 0.02 Å. These certainly result in the much smaller reorganization energies of ca. 116 meV in the lipid membrane, compared with 633 meV in dilute solution (Fig. 3f and Table S4, ESI). Interestingly, as shown in Fig. 3g, the emission spectra of TTVP in the lipid membrane (∼630 nm) are about 20 nm blue-shifted compared to that in the dilute solution (652 nm), because of the less reorganization energy in the lipid membrane than that in the THF solution (Fig. 3f),23,24 which is in good agreement with the experimental results.21


image file: c9mh00906j-f3.tif
Fig. 3 (a) The QM/MM model of TTVP in the lipid membrane. (b and c) Calculated NTOs for TTVP at S1 states in both the lipid membrane and dilute THF solution. (d and e) Superposition of optimized structures at both S0 and S1 states for TTVP in both the lipid membrane and dilute THF solution. Calculated (f) total reorganization energies and (g) fluorescence emission spectrum of TTVP in both the lipid membrane and dilute THF solution.

The radiative and nonradiative decay rate constants of TTVP in dilute THF solution are calculated to be 3.3 × 108 s−1 and 2.4 × 1010 s−1, respectively, which results in a low fluorescence quantum yield of 1.3% (Table S5, ESI). This is well consistent with the non-emissive phenomenon of TTVP in dilute THF solution.21 The radiative rate constant of TTVP in the lipid membrane is 3.6 × 108 to 3.7 × 108 s−1, which is a little bit larger than that in solution, because of the larger transition dipole moment and the higher emissive energy of TTVP in the lipid membrane as discussed above (Table S3, ESI). Far beyond the computational ability, the vibrational frequencies of TTVP in the lipid membrane could not be calculated by using the QM/MM model, and the accurate nonradiative decay rate constants could also not be obtained consequently. As known from previous work,23,25–28 the decrease in reorganization energy can sharply retard the electron-vibration coupling caused nonradiative decay rate constant. Based on the calculated adiabatic and vertical excitation energies of the optimized geometries at both S0 and S1 states (Table S3, ESI), the total reorganization energies can be evaluated (Fig. 3f and Table S4, ESI). Accordingly, the nonradiative decay rate constants of TTVP in the lipid membrane are sure to be very small, because of their much smaller reorganization energies than the isolated molecule in dilute THF solution. Similar to AIE, the embedded TTVP in a densely packed lipid membrane starts to emit strong fluorescence and can “light-up” the cell membrane of living HeLa cells.21 It can be safely inferred that the fluorescent dye TTPy with the TTV moiety and pyridine group also would emit bright light once it enters into the densely packed mitochondria through the cell membrane.14

Based on the above discussions, the fluorescent detection mechanism of AIEgens (TTVP and TTPy) that selectively target the cell membrane or mitochondria is revealed. That is, the hydrophobic TTV moiety and pyridine group of AIEgens control the emission property, while the permeation ability of the AIEgens through the cell membrane is determined by their charged headgroups, especially, the 3-(trimethylammonio)propyl group (see Fig. 4a). Therefore, we proposed a rational design strategy of AIEgen-based fluorescent probes, which should be amphiphilic AIEgens, including a hydrophobic emissive group with AIE characteristics and a hydrophilic headgroup with different positive charges, to selectively target the cell membrane or mitochondria. According to this strategy, we designed four amphiphilic AIEgens, TPEPy, TPy, TPEVP and TVP, with the chemical structures shown in Fig. 4b–e. Both TPEPy and TPy are composed of the AIE-core and pyridine group, indicating their TTPy-analogue characteristics; meanwhile, both TPEVP and TVP are structurally similar to TTVP, considering the existence of the moiety with two positive charges. We facilely synthesized them through the condensation reaction between the aldehyde derivative and pyridinium (Scheme S1, ESI) and studied their optical properties by UV-Vis and photoluminescence (PL) spectroscopies (Fig. S10, ESI). In addition, the typical AIE features were confirmed for all four compounds (Fig. S11 and Table S6, ESI). The cell imaging experiments are further proceeded by using HeLa cells as a cell model. For both TPEPy and TPy, after incubating HeLa cells with 1 μM of AIEgens for 10 minutes, the reticulum-like mitochondria can be clearly visualized with an excellent image contrast to cell background (Fig. 5a and e). To assess the staining specificity of these two AIEgens to mitochondria, the colocalization experiment proceeded through co-staining with MitoTracker Green, which is a commercially available mitochondria-specific biomarker. It is observed that the cell imaging of AIEgens and MitoTracker Green are well merged, and the Pearson correlation coefficients are determined to be 91% and 94%, respectively, indicating excellent mitochondria-specific staining properties of TPEPy and TPy (Fig. 5c and g). Moreover, both TPEVP and TVP can specifically “light-up” the cell membrane showing high signal-to-noise ratio of cell imaging after incubation of HeLa cells with 1 μM of AIEgen for 1 minute (Fig. 5i and m). Colocalization investigation involving a cell membrane-specific bio-probe Cellmask Green demonstrates that the cell imaging outputs of TPEVP and TVP can perfectly overlap with that of Cellmask Green, strongly suggesting their prominent cell membrane specific targeting capability (Fig. 5k and o). Overall, as a proof-of-concept, two series of TTPy and TTVP analogues are prepared and evaluated in a cell imaging study, and the results are in good accordance with the theoretical prediction. In addition, a recently discovered amphiphilic tetraphenylethene-based pyridinium salt belonging to the TTVP analogues also showed high cell membrane specificity,29 and this is a further strong evidence supporting our theoretical prediction.


image file: c9mh00906j-f4.tif
Fig. 4 (a) Schematic illustration of the design principle of AIEgens. (b–e) Four rationally designed AIEgens.

image file: c9mh00906j-f5.tif
Fig. 5 Colocalization test of AIEgens. Confocal images of HeLa cells stained with (a) TPEPy, (b and f) MitoTracker Green, (e) TPy, (i) TPEVP, (j and n) Cellmask Green, (m) TVP, and (c, g, k and o) Merged images of (a) and (b), (e) and (f), (i) and (j), and (m) and (n). (d, h, l and p) Bright field. Concentrations: TPEPy and TPy (1 μM), TPEVP and TVP (1 μM), MitoTracker Green (50 × 10−9 M), and Cellmask Green (2.5 μg mL−1). λex (all AIEgens): 405 nm, λex (MitoTracker Green or Cellmask Green): 488 nm; the emission filter of all AIEgens: 600–744 nm, Cellmask Green: 490–600 nm, and MitoTracker Green: 490–580 nm. Scale bar = 20 μm.

Conclusions

By combining large-scale molecular dynamics simulations and the hybrid QM/MM model, we systematically perform a detailed investigation of the overall translocation ability through the lipid membrane and the derivation of the structure–property relationship of two representative AIEgens and reveal the intrinsic mechanism of their striking difference in targeting specificity to the cell membrane or mitochondria at the atomic level. The free energy profiles reveal that the free energy barrier of TTVP at the membrane center is significantly higher than that of TTPy, which directly decreases its permeability coefficients and reduces the translocation ability of TTVP through the lipid membrane. We find that the large difference in translocation ability of the two AIEgens is caused by their different intermolecular interactions between the AIEgens and the lipid membrane, due to the different number of charges in their headgroups. The electronic transitions of TTVP mainly happen on the central TTV moiety and pyridine group and the charged quaternary 3-(trimethylammonio)propyl headgroups do not contribute in both dilute solution and lipid membrane environments. When the TTVP is transferred from solution into the lipid membrane, it shows stronger rigidity with much smaller structural changes between the S1 and S0 states. The densely packed environment in the lipid membrane makes TTVP much more rigid and largely restricts its nonradiative decay rate constants and induces its strong fluorescence, with sharp contrast to the non-emissive nature of TTVP in dilute THF solution; these results are in good agreement with the observations in experiments.14,21 So far, the fluorescent detection mechanism of TTPy and TTVP in living HeLa cells is disclosed. That is, the hydrophobic TTV moiety and pyridine group of AIEgens manipulate the emission behaviors, and the charged headgroups are in charge of the permeation ability of the fluorescent probe through the cell membrane. Naturally, the molecular design strategy is proposed to achieve excellent AIEgen-based fluorescent probes to selectively target the cell membrane or mitochondria through combining a hydrophobic emissive group with AIE characteristics and a hydrophilic headgroup with different positive charges. Finally, four AIEgen-based fluorescent probes were designed and synthesized. The cell imaging experimental results reveal that TPEPy and TPy, with TTPy-analogue characteristics, exhibit excellent mitochondria-specific staining properties, while as TTVP analogues, both TPEVP and TVP could specifically “light-up” the cell membrane, which perfectly verifies the theoretical prediction. This would provide a universal blueprint and theoretical direction for designing the next generation of mitochondria and cell membrane-specific biomarkers, and stimulate the development of fluorescent bio-probes in clinical applications.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank the financial support by the National Natural Science Foundation of China (Grants No. 21803007, 21801169) and the Beijing Institute of Technology Research Fund Program for Young Scholars.

Notes and references

  1. B. N. G. Giepmans, S. R. Adams, M. H. Ellisman and R. Y. Tsien, Science, 2006, 312, 217 CrossRef CAS PubMed.
  2. R. Weissleder and M. J. Pittet, Nature, 2008, 452, 580 CrossRef CAS PubMed.
  3. M. Nuriya, S. Fukushima, A. Momotake, T. Shinotsuka, M. Yasui and T. Arai, Nat. Commun., 2016, 7, 11557 CrossRef CAS PubMed.
  4. M. Grossi, M. Morgunova, S. Cheung, D. Scholz, E. Conroy, M. Terrile, A. Panarella, J. C. Simpson, W. M. Gallagher and D. F. O’Shea, Nat. Commun., 2016, 7, 10855 CrossRef CAS PubMed.
  5. L. Wang, M. S. Frei, A. Salim and K. Johnsson, J. Am. Chem. Soc., 2019, 141, 2770–2781 CrossRef CAS PubMed.
  6. J. Luo, Z. Xie, J. W. Y. Lam, L. Cheng, H. Chen, C. Qiu, H. S. Kwok, X. Zhan, Y. Liu, D. Zhu and B. Z. Tang, Chem. Commun., 2001, 1740–1741 RSC.
  7. J. Mei, N. L. C. Leung, R. T. K. Kwok, J. W. Y. Lam and B. Z. Tang, Chem. Rev., 2015, 115, 11718–11940 CrossRef CAS PubMed.
  8. Y. Hong, J. W. Y. Lam and B. Z. Tang, Chem. Soc. Rev., 2011, 40, 5361–5388 RSC.
  9. H. Shi, J. Liu, J. Geng, B. Z. Tang and B. Liu, J. Am. Chem. Soc., 2012, 134, 9569–9572 CrossRef CAS PubMed.
  10. Y. Li, Y. Wu, J. Chang, M. Chen, R. Liu and F. Li, Chem. Commun., 2013, 49, 11335–11337 RSC.
  11. G. Liang, J. W. Y. Lam, W. Qin, J. Li, N. Xie and B. Z. Tang, Chem. Commun., 2014, 50, 1725–1727 RSC.
  12. M. Gao, Q. Hu, G. Feng, B. Z. Tang and B. Liu, J. Mater. Chem. B, 2014, 2, 3438–3442 RSC.
  13. E. Wang, E. Zhao, Y. Hong, J. W. Y. Lam and B. Z. Tang, J. Mater. Chem. B, 2014, 2, 2013–2019 RSC.
  14. D. Wang, M. M. S. Lee, G. Shan, R. T. K. Kwok, J. W. Y. Lam, H. Su, Y. Cai and B. Z. Tang, Adv. Mater., 2018, 30, 1802105 CrossRef PubMed.
  15. D. Li, X. Tian, A. Wang, L. Guan, J. Zheng, F. Li, S. Li, H. Zhou, J. Wu and Y. Tian, Chem. Sci., 2016, 7, 2257–2263 RSC.
  16. Z. Feng, D. Li, M. Zhang, T. Shao, Y. Shen, X. Tian, Q. Zhang, S. Li, J. Wu and Y. Tian, Chem. Sci., 2019 10.1039/C9SC01705D.
  17. C. Zhang, Y. Zhao, D. Li, J. Liu, H. Han, D. He, X. Tian, S. Li, J. Wu and Y. Tian, Chem. Commun., 2019, 55, 1450–1453 RSC.
  18. J. Mei, Y. Huang and H. Tian, ACS Appl. Mater. Interfaces, 2018, 10, 12217–12261 CrossRef CAS PubMed.
  19. D. Wang, M. M. S. Lee, W. Xu, G. Shan, X. Zheng, R. T. K. Kwok, J. W. Y. Lam, X. Hu and B. Z. Tang, Angew. Chem., Int. Ed., 2019, 58, 5628–5632 CrossRef CAS PubMed.
  20. Y. Chen, W. Zhang, Z. Zhao, Y. Cai, J. Gong, R. T. K. Kwok, J. W. Y. Lam, H. H. Y. Sung, I. D. Williams and B. Z. Tang, Angew. Chem., Int. Ed., 2018, 57, 5011–5015 CrossRef CAS PubMed.
  21. D. Wang, H. Su, R. T. K. Kwok, X. Hu, H. Zou, Q. Luo, M. S. Lee, W. Xu, J. W. Y. Lam and B. Z. Tang, Chem. Sci., 2018, 9, 3685–3693 RSC.
  22. S.-J. Marrink and H. J. C. Berendsen, J. Phys. Chem., 1994, 98, 4155–4168 CrossRef CAS.
  23. X. Zheng, Q. Peng, L. Zhu, Y. Xie, X. Huang and Z. Shuai, Nanoscale, 2016, 8, 15173–15180 RSC.
  24. Q. Wu, T. Zhang, Q. Peng, D. Wang and Z. Shuai, Phys. Chem. Chem. Phys., 2014, 16, 5545–5552 RSC.
  25. T. Zhang, Q. Peng, C. Quan, H. Nie, Y. Niu, Y. Xie, Z. Zhao, B. Z. Tang and Z. Shuai, Chem. Sci., 2016, 7, 5573–5580 RSC.
  26. Y. Xie, T. Zhang, Z. Li, Q. Peng, Y. Yi and Z. Shuai, Chem. – Asian J., 2015, 10, 2154–2161 CrossRef CAS PubMed.
  27. T. Zhang, Y. Jiang, Y. Niu, D. Wang, Q. Peng and Z. Shuai, J. Phys. Chem. A, 2014, 118, 9094–9104 CrossRef CAS PubMed.
  28. Q. Peng, Y. Yi, Z. Shuai and J. Shao, J. Am. Chem. Soc., 2007, 129, 9333–9339 CrossRef CAS PubMed.
  29. W. Zhang, Y. Huang, Y. Chen, E. Zhao, Y. Hong, S. Chen, J. W. Y. Lam, Y. Chen, J. Hou and B. Z. Tang, ACS Appl. Mater. Interfaces, 2019, 11, 10567–10577 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available: Experimental and computational details; additional experimental and computational results. See DOI: 10.1039/c9mh00906j

This journal is © The Royal Society of Chemistry 2019