Self-Assembly of Highly Stable Zirconium(IV) Coordination Cages with Aggregation Induced Emission Molecular Rotors for Live-Cell Imaging

Article abstract of DOI:10.1002/anie.201915199

The self-assembly of highly stable zirconium(IV)-based coordination cages with aggregation induced emission (AIE) molecular rotors for in vitro bio-imaging is reported. The two coordination cages, NUS-100 and NUS-101, are assembled from the highly stable

Full text of DOI:10.1002/anie.201915199

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Self-Assembly of Highly Stable Zirconium (IV) Coordination
Cages with Aggregation Induced Emission (AIE) Molecular
Rotors for Live-Cell Imaging
Authors: Jinqiao Dong, Yutong Pan, Heng Wang, Kuiwei Yang,
Lingmei Liu, Zhiwei Qiao, Yi Di Yuan, Shing Bo Peh, Jian
Zhang, Leilei Shi, Hong Liang, Yu Han, Xiaopeng Li, Jianwen
Jiang, Bin Liu, and Dan Zhao
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201915199
Angew. Chem. 10.1002/ange.201915199
Link to VoR: http://dx.doi.org/10.1002/anie.201915199
http://dx.doi.org/10.1002/ange.201915199

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
Self-Assembly of Highly Stable Zirconium (IV) Coordination Cages
with Aggregation Induced Emission (AIE) Molecular Rotors for
Live-Cell Imaging
Jinqiao Dong,+[a] Yutong Pan,+[a] Heng Wang,[b] Kuiwei Yang,[a] Lingmei Liu,[c] Zhiwei Qiao,[d] Yi Di
Yuan,[a] Shing Bo Peh,[a] Jian Zhang,[a] Leilei Shi,[a] Hong Liang,[d] Yu Han,[c] Xiaopeng Li,[b] Jianwen
Jiang,[a] Bin Liu,*[a] and Dan Zhao*[a]
This paper is dedicated to the 20 Anniversary of Aggregation Induced Emission (AIE)
Abstract: Although supramolecular coordination complexes have
demonstrated outstanding host-guest chemistry in solution or solid
state, the low structural stability, weak fluorescence emission, and cell
viability in aqueous solutions have limited their applications in real
cellular environment. Herein, we report the self-assembly of highly
stable zirconium (IV)-based coordination cages with aggregation
induced emission (AIE) molecular rotors for in-vitro bio-imaging. The
two coordination cages, viz. NUS-100 and NUS-101, are assembled
from the highly stable trinuclear zirconium vertices and two flexible
carboxyl-decorated tetraphenylethylene (TPE) spacers. Extensive
experimental and theoretical results show that the emissive intensity
of the coordination cages can be controlled by restricting the
dynamics of AIE-active molecular rotors though multiple external
stimuli. Because the two coordination cages have excellent chemical
stability in aqueous solutions (pH stability: 2 – 10) and impressive AIE
characteristics contributed by the molecular rotors, they can be
employed as novel biological fluorescent probes for in-vitro live-cell
imaging.
shapes,[1] as well as functionality for mimicking host-guest
recognition in living organisms.[2] SCCs have garnered great
interest in the past two decades as enzyme mimetics, with
advanced functions and potential applications in anion binding
and
recognition,[3]
chemical
sensing,[4]
light-emitting,[5]
supramolecular catalysis,[6] synthetic ion channels,[7] molecular
recognition and separation,[8] etc. Compared with other molecules
or materials, the advantage of discrete SCCs such as
coordination cages is their ability to form host−guest complexes
through their finely tunable hollow cavities,[9] and they can be
potentially used as nanocarriers for biological applications.
However, there remain critical gaps in applying SCCs in real
biological applications such as live-cell imaging, which is
significant for the efficient and early diagnosis of cancer.[10]
Basically, to our knowledge, there are still several issues to be
overcome: (1) The limited structural stability of SCCs in harsh
cellular environment. For example, most of the coordination
cages are constructed by metal–ligand bonds (N→metal or O→
metal). As a result, these cages have poor water stability, making
them difficult to survive in living cellular environment.[11] (2) The
weak fluorescence emission of SCCs in cellular environment.
There are three possible reasons: (i) The aggregation caused
quenching (ACQ) phenomenon[12] of SCCs strongly hampers the
fluorescence when those SCCs aggregate together in the cellular
environment. (ii) Some metal ions such as Fe (II), Cu (II), Co (II),
and Ni (II) in the SCCs can easily self-quench the fluorescence
due to their exceptionally short (sub-picosecond) ligand-to-metal
charge transfer (LMCT) excited-state lifetimes.[13] (iii) Their
fluorescence may be easily quenched by various biomolecules
and components, such as amino acids, peptides, proteins, and
serums in the complicated cellular environment. (3) The cell
viability of SCCs is also a crucial problem for in-vitro biological
applications.[14] Therefore, self-assembly of highly stable and
strong emissive SCCs with low cytotoxic is still a synthetic
challenge that, if overcome, could significantly expand the utility
of SCCs in many biological applications.
Introduction
Coordination-driven self-assembly has emerged as a powerful
bottom-up approach to construct artificial supramolecular
coordination complexes (SCCs) with well-defined sizes and
[a]
Dr. J. Dong,[+] Y. Pan,[+] Dr. K. Yang, Y. D. Yuan, S. B. Peh, Dr. J.
Zhang, Dr. L. Shi, Prof. J. Jiang, Prof. B. Liu, Prof. D. Zhao
Department of Chemical and Biomolecular Engineering, National
University of Singapore, 4 Engineering Drive 4, Singapore 117585
E-mail: chezhao@nus.edu.sg (D.Z.); cheliub@nus.edu.sg (B.L.)
Dr. H. Wang, Prof. X. Li
Department of Chemistry, University of South Florida, Tampa, USA
33620
Dr. L. Liu, Prof. Y. Han
[b]
[c]
King Abdullah University of Science and Technology (KAUST),
Advanced Membranes and Porous Materials Center, Physical
Sciences and Engineering Division, Thuwal, Saudi Arabia 23955-
6900
We have shown previously that coordination cages built by
trinuclear zirconium clusters are highly stable due to the high Zr-
O bond energy (776 kJ mol-1), showing excellent stability in
aqueous solutions with acidic, neutral, and weak basic
conditions.[15] We believe that this building unit can solve the
stability problem of coordination cages in the cellular environment.
For the second issue, Tang and co-workers have developed
species that exhibit the opposite phenomenon called aggregation-
induced emission (AIE),[16] which could offer strong fluorescence
[d]
[+]
Prof. Z. Qiao, Prof. H. Liang
Guangzhou Key Laboratory for New Energy and Green Catalysis,
School of Chemistry and Chemical Engineering, Guangzhou
University, Guangzhou, P. R. China 510006
These authors contributed equally to this work.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
upon molecular aggregation. Based on the AIE mechanism, we
recently found that the dynamic phenyl rings of TPE core can work
as AIE molecular rotors in porous materials,[17] which could be a
good strategy to overcome the ACQ behavior in the cellular
environment. Compared to our previous studies in which the AIE
molecular rotors are anchored inside the cavities of porous
organic frameworks (POFs),[17b-d] such molecular rotors exposed
on the external surface of coordination cages would further
manifest the expected AIE effect. Additionally, the TEP molecules
contain four hydrophobic phenyl rings have limited interactions
with biomolecules, leading to inhibition of quenching in the
complicated cellular environment. For the third issue, we try to
avoid toxic and self-quenching metal ions for the self-assembly of
targeted coordination cages. With these design targets in mind,
we envision that trinuclear zirconium (IV) cluster-based
supramolecular coordination cages containing free rotary AIE
molecular rotors would fulfill the above-mentioned requirements
for biological applications.
Figure 1. (a) Self-assembly of Zr6L3 coordination cage without AIE molecular
rotors using 4,4’-sulfonyldibenzoic acid ligands. (b) Self-assembly of Zr6L3
coordination cages with AIE molecular rotors using flexible carboxyl-
decorated TPE ligands. (c) The chemical structures of TPE ligands and
trinuclear zirconium clusters used in this study.
Scheme 1. Self-assembly of highly stable zirconium (IV) coordination cages
containing AIE molecular rotors for live-cell imaging (red colour groups
represent the AIE molecular rotors in the coordination cages).
In an effort to prove the above hypotheses, we herein
demonstrate the first zirconium (IV) coordination cages with AIE
molecular rotors that break the above limitations and yield novel
biological probes for live-cell imaging (Scheme 1). Our strategy
leverages the well-established TPE ligands with free phenyl rings
as molecular rotors and carboxyl-decorated phenyl rings as
stators to self-assemble highly stable Zr6L3 coordination cages.
Our investigations show that the phenyl rings on the external
surface of the coordination cages can function as AIE molecular
rotors, leading to impressive AIE characteristics as proved by
theoretical calculations and extensive experimental studies
including temperature-responsive, viscosity-responsive, and
guest size-responsive effects, as well as control experiments.
Eventually, in-vitro cell culture assays suggest that these AIE
molecular rotor-containing coordination cages are promising
materials for live-cell imaging.
Figure S2 and S3b). The single crystals of self-assembled
coordination cages named NUS-100 (H2L0 ligand) and NUS-101
(H2L1
ligand)
were
obtained
by
heating
bis(cyclopentadienyl)zirconium dichloride (ZrCp2Cl2) and H2L0 or
H2L1 in N,N-dimethylformamide (DMF) at 65 °C for 12 h (Figure
1b-c), with the yields of 76% for NUS-100 and 64% for NUS-101.
High-resolution electrospray ionization time-of-flight mass
spectrometry (ESI-TOF-MS) was used to ascertain the correct
molecular geometry of the self-assembled coordination cages. As
expected, highly pure Zr6L3 supramolecular structures were
identified as the target assemblies with the measured molecular
weights of 2326 Da [experimental data at 1164.04 Da for (M +
2H)2+] and 2782 Da [experimental data at 1392.11 Da for (M +
2H)2+] for NUS-100 (Figure 2a) and NUS-101 (Figure 2b),
respectively. In addition, the experimental isotopic patterns of two
charge state of NUS-100 and NUS-101 are in good agreement
with the simulated isotopic distributions (Inset of Figure 2a-b).
Furthermore, the results of traveling wave ion mobility-mass
spectrometry (TWIM-MS) show only one narrowly distributed
signal band assigned to the [M + 2H]2+ species of NUS 100-101,
indicating the high purity of the two cages with no existence of
other isomers or conformers. Moreover, we can observe that the
drift time of NUS-101 (8.49 ms) is slightly longer than that of NUS-
100 (6.28 ms) (Table S1), suggesting that the molecular size and
Results and Discussion
We began our studies with the design and synthesis of TPE
ligands. As shown in Figure 1, the ligand H2L0 was synthesized
according to the reported procedures (Figure S1 and S3a),[17a]
and the longer ligand H2L1 was synthesized by the Suzuki-
Miyaura
couplings
of
4,4'-(2,2-diphenylethene-1,1-
diyl)bis(bromobenzene) and 4-(methoxycarbonyl)phenylboronic
acid, followed by hydrolysis reaction (see SI for synthetic details,
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
mass of NUS-101 is larger than that of NUS-100 (Figure S4). The
cage for control experiments was self-assembled by 4,4’-
sulfonyldibenzoic acid (H2SDB) and ZrCp2Cl2 in DMF/H2O at
65 °C (Figure 1a and Figure S5), which was previously reported
as UCM-1.[18] In this study, we denoted it as non-rotor-based
coordination cage (NR-cage), which has the same molecular
geometry (Zr6L3) as that of NUS 100-101 (Figure S6-S7), with the
difference that the six phenyl rings in NUS-100 or NUS-101 are
replaced by six S=O groups in NR-cage. Such an isostructural
Zr6L3 cage is beneficial to understanding the function of AIE
molecular rotors in the coordination cages.
respect to the free ligands. For example, the proton resonance
signals of phenyl rings at coordination part (Ha) of NUS-100
undergo a remarkable upfield shift (from 7.71 to 7.48 ppm, ∆Ha =
– 0.23 ppm) with respect to H2L0. Such an upfield shift suggests
that the free ligands are confined within the interlinked cage
structure and thus experience
a
shielded magnetic
environment.[5a] A similar upfield shift was observed in NUS-101.
The quantitative formation of the two coordination cages was
further confirmed via two-dimensional (2D) 1H–1H correlation
spectroscopy (COSY) (Figure 3e-f) and 1H–1H nuclear
Overhauser effect spectroscopy (NOESY) experiments (Figure
S8), which show a single vertical trace in all cases. 2D diffusion-
ordered spectroscopy (DOSY) experiments also confirm the
formation of a single species with diffusion coefficients of 8.09 ×
10-11 m2 s-1 for NUS-100 (Figure 3g and Figure S9a) and 8.76 ×
10-11 m2 s-1 for NUS-101 (Figure 3h and Figure S9b), further
suggesting that the Stokes−Einstein hydrodynamic radius of
NUS-101 is larger than that of NUS-100, which is consistent with
their geometry. Taken together, all the above NMR data indicate
that the C2-symmetry of TPE ligands are preserved in the Zr6L3
coordination cages.
Figure 2. ESI-TOF-MS peaks of discrete NUS-100 (a) and NUS-101 (b). Inset:
experimental and calculated isotope ESI-TOF-MS peaks. TWIM-MS plots of
discrete NUS-100 (c) and NUS-101 (d). Minimum energy structures of NUS-100
(e) and NUS-101 (h) based on molecular dynamics simulations. Isodensity plots
of the LUMO and HOMO of NUS-100 (f-g) and NUS-101 (i-j) based on TD-DFT
calculations.
Figure 3. Partial 1H NMR spectra of H2L0 (a), NUS-100 (b), H2L1 (c), and NUS-
101 (d). 2D COSY spectra of NUS-100 (e) and NUS-101 (f). 2D DOSY spectra
of NUS-100 (g) and NUS-101 (h).
The formation of NUS 100-101 was confirmed by 1H nuclear
magnetic resonance (NMR) spectroscopy. Compared with the
sharp 1H-NMR signals of the TPE ligands (Figure 3a and 3c), the
1H-NMR spectra of NUS 100-101 display slightly broadened
peaks of all protons, due to their much slower tumbling motion on
the NMR time scale. In addition, each cage displays only one set
of NMR resonances in solution, suggesting the formation of
discrete and highly symmetric assemblies (Figure 3b and 3d).
However, most of these signals have slight upfield shifts with
Unfortunately, the weak X-ray diffraction limited the crystal
structure determination in both cases. In this study, a low-dose
(ca. 5 e-/Å2) high-resolution transmission electron microscope
(HR-TEM) technique[19] was employed to characterize the two
coordination cages for avoiding electron beam-induced structural
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
damage (Figure 4c and 4f). Impressively, selected area electron
diffraction (SAED) patterns show the rectangular patterns with
strong diffraction in NUS-100 (Figure S10b), further confirming
the high crystallinity of the cage. Moreover, the lattice fringes can
be clearly observed from the obtained HR-TEM images in both
cages, and the d-spacing values of around 1.47 nm for NUS-100
(Figure 4d) and 1.09 nm for NUS-101 (Figure 4g) match well with
the XRD data (Figure S11). In addition, field-emission scanning
electron microscopy (FE-SEM) show that the bulk crystals of
these cages exhibit sheet-like morphology (Figure S12), and can
show that NUS-100 and NUS-101 have blue-shifts of around 25
and 22 nm compared to H2L0 and H2L1 in the solid-state (Figure
S19-S20), respectively. UV-Vis spectra show that NUS-100 and
NUS-101 have 31 and 3 nm blue-shifts than H2L0 (λmax = 309 nm)
and H2L1 (λmax = 319 nm) (Figure S21-S22), respectively. To gain
insight into the electronic property of the two cages, time
dependent density functional theory (TD-DFT) calculations were
performed to understand their transition from the highest energy
occupied molecular orbital (HOMO) to the lowest unoccupied
molecular orbital (LUMO) (Figure 2f-g, Figure 2i-j, and Figure
be easily exfoliated into thin nanosheets in n-hexane (Figure S13). S23-S25). TD-DFT results indicate that when the ligand
Quantitative measurements of their thickness were performed
using atomic force microscopy (AFM), with the thicknesses of
around 20 nm for NUS-100 (Figure 4e) and 15 nm for NUS-101
(Figure 4h).
coordinates with metal to form cage complexes, the interplay
becomes complicated. The transition of HOMO→LUMO only
contributes about 20% and 10% to the absorption peak of NUS-
100 (λmax = 301 nm) and NUS-101 (λmax = 332 nm), respectively,
which are much smaller than their ligands (see more details and
explanations in Table S2-S5). Other information such as Fourier
transform infrared spectra (FT-IR), thermogravimetric analyses
(TGA), and corresponding optical band gaps (Eg) are shown in the
supporting information (Figure S26-S28).
Figure 4. Fluorescence images of NUS-100 crystals (a) and NUS-101 crystals
(b). Inset: optical images of NUS-100 and NUS-101 crystals, both scale bars
are 500 μm. HR-TEM images of NUS-100 crystals (c) and NUS-101 crystals (f).
HR-TEM images of NUS-100 crystals (d) and NUS-101 crystals (g) featuring the
planar lattice structure. Inset: the fast Fourier transformation of Figure (d) and
(g). AFM images of as-exfoliated nanosheets of NUS-100 (e) and NUS-101 (h).
NUS 100-101 were further characterized by various
spectroscopic methods. X-ray photoelectron spectroscopy (XPS)
spectra show the presence of Zr (Zr 3d: 182 eV) and O (O 1s: 532
eV) elements in both cages (Figure S14). In addition, the
corresponding energy-dispersive X-ray (EDX) spectroscopy
mapping indicates that there are homogeneous distributions of Zr
and O elements in the crystals (Figure S15-S16), which is
consistent with XPS data. Fluorescence images show that the two
cage crystals exhibit cyan emission (Figure 4a-b). Furthermore,
we performed 3D confocal laser scanning microscopy (CLSM) to
observe the anisotropic behavior of the fluorescence of the two
cage crystals, unveiling the homogeneous strong emissions in the
whole crystals (Figure S17-S18). Fluorescence emission spectra
Figure 5. (a) Fluorescence emission spectra of NUS-100 versus n-hexane
fraction in CH2Cl2/n-hexane mixtures. (b) Plot of relative emission intensity (I/I0)
of NUS 100-101 in CH2Cl2/n-hexane mixtures. (c) Fluorescence emission
spectra of NUS-100 in solutions of various viscosities. (d) Plot of relative
emission intensity (I/I1cp) of NUS 100-101 with various viscosities. (e)
Fluorescence emission spectra of NUS-100 in various VOC solutions with
different molecular sizes. (f) Relative fluorescence intensity (I/Ibenzene) of NUS
100-101 in various VOC solutions. The concentrations of NUS 100-101 are 1.0
× 10−5 M, and the excitation wavelength is 360 nm in the experiments.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
To have a better understanding on the dynamic behavior of
the phenyl ring rotors located on the external surface of the two
coordination cages, the AIE characteristics of NUS 100-101 were
studied in mixtures of dichloromethane (CH2Cl2) and n-hexane.[5a]
As expected, both cages exhibit typical AIE characteristics. As
shown in Figure 5a, the fluorescence intensity of NUS-100 is
substantially enhanced with increasing n-hexane fraction (fh). The
highest fluorescence intensities were obtained at fh of 100% for
NUS-100 and NUS-101, and are 4.51- and 3.18-fold higher than
that in CH2Cl2 solutions (Figure S29a and Figure 5b), respectively.
However, the NR-cage exhibits a slightly reduced fluorescence
intensity at fh of 100%, with is only 0.90-fold of that in CH2Cl2
solution (Figure S30a). Such a typical AIE phenomenon of NUS
100-101 can be attributed to the restriction of phenyl ring rotors of
the cages in their aggregated state. Next, various stimuli were
applied to evaluate the sensitivity of AIE molecular rotors of the
two cages. First of all, it is known that fluorescent molecular rotors
with the nonradiative decay of their excited state is influenced by
the viscosity of the medium.[20] Therefore, fluorescence
measurements of the two cages were made in DMF/glycerol
mixtures with different viscosities. As expected, when the
viscosity is increased to 950 cp (99% glycerol solution), the
emission intensity of NUS-100 greatly enhances by 12.40-fold (IE
= I950cp/I1cp) compared to that in DMF solution (1 cp) (Figure 5c-d).
A similar trend of IE was found in NUS-101 with a value of 8.68-
fold (Figure S29b and Figure 5d). The control experiments show
that the IE value (1.76) of NR-cage is much lower from DMF to
glycerol solution (Figure S31 and Figure 5d). These results
unambiguously indicate that such highly-stable coordination
cages equipped with AIE molecular rotors are highly sensitive to
environmental viscosity, which can be contributed to the restricted
rotation of the AIE molecular rotors preventing relaxation via
populating of the dark excited state.
that in other VOC solutions, suggesting the size-dependent
restricted effect for the molecular rotors. Our previous study
indicates that the emission intensities of POFs NUS 20-23 in
1,3,5-triisopropylbenzene are much lower than that in mesitylene
because the pore sizes of these POFs are too small for the
incorporation and interaction of 1,3,5-triisopropylbenzene.[17b]
Surprisingly, we observed the sustained increase of emission
intensity in 1,3,5-triisopropylbenzene solution in this study (Figure
5f). This is because the molecular rotors located at the external
surface of the cages have free access to analytes, which is
different from the confined molecular rotors located inside the
pores of those POFs. In addition, the IR value of NR-cage1,3,5-
triisopropylbenzene is only 1.28 (Figure S30b), which is much
smaller than that of the two cages. Our findings unambiguously
indicate that the dynamic molecular rotors of the cages can be
restricted by guest molecules with size-dependent effect.
Figure 6. Temperature-responsive emissions of NUS-100 (a) and NUS-101 (b)
in THF solutions recorded between −10 and 60 °C. (c) Relative fluorescence
intensity (I/I-10 °C) of NUS 100-101, TPE ligands, and NR-cage under various
temperatures. (d) Cyclic switching (10 times) of fluorescence intensity upon
heating and cooling for NUS-100 (top) and NUS-101 (bottom) in THF solutions.
The concentrations of NUS 100-101 are 5.0 × 10−5 M, and the excitation
wavelength is 360 nm in the experiments.
Second, the two cages exhibit different fluorescence
emissions after soaking in various volatile organic compounds
(VOCs) with size-dependent effect. Using the emission intensity
of cages being soaked in benzene (Ibenzene, 4.3 × 5.0 Å) as the
reference, we evaluated the relative intensity (IR = I/Ibenzene) of the
cages in different analytes, such as toluene (4.3 × 5.9 Å), o-xylene
(5.5 × 5.9 Å), m-xylene (5.0 × 6.7 Å), p-xylene (4.3 × 6.8 Å),
mesitylene (5.8 × 6.7 Å), and 1,3,5-triisopropylbenzene (8.4 × 9.0
Å).[17c] A clear positive correlation between IR values versus
molecular size of the analytes can be found (Figure 5e and Figure
S29c). Taking NUS-100 as an example, the IR values of NUS-
100benzene, NUS-100toluene, NUS-100p-xylene, and
NUS-100mesitylene are 1.00, 1.32, 1.52, and 3.57 (Figure 5f),
respectively. A similar trend was also observed in NUS-101 (1.00,
1.37, 2.05, and 2.83, respectively). We speculated that the size-
dependent turn-on fluorescence of the two cages is due to the
different restriction degrees of the phenyl ring rotors. To prove this
point, we used 1,3,5-triisopropylbenzene as a large analyte to see
if the phenyl ring rotors of the cages can be further restricted for
higher fluorescence emission. As shown in Figure 5f, the IR values
of NUS-1001,3,5-triisopropylbenzene (IR = 6.48) and NUS-
1011,3,5-triisopropylbenzene (IR = 3.36) are much higher than
Third, temperature-responsive fluorescence is also a typical
characteristic of molecular rotors.[21] In order to prove this, we
tested the temperature-responsive fluorescence of the two cages.
As shown in Figure 6a-b, upon increasing temperature from -10
to 60 °C, the emission intensities of NUS 100-101 significantly
decrease. This is because rotations of the phenyl ring rotors are
expected to speed up upon increasing temperature, which can
lead to the disruption of charge-transfer and consumption of
energy with reduced fluorescence emissions.[21b] In particular, the
emission intensities of NUS 100-101 drop by 68.9% and 59.9%,
respectively, over this temperature range. Furthermore, the
temperature-dependent emission responses of NUS 100-101 can
be fitted with linear functions, with average slopes of -0.96% and
-0.87% per °C, respectively. Moreover, the control experiment
shows that the decrease in emission intensity of NR-cage is only
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
24.5% with an average slope of -0.35% per °C over this
temperature range (Figure S32 and Figure 6c), which is also
much smaller than that of NUS 100-101, highlighting the
sensitivity of phenyl ring rotors toward temperature. In addition,
although we observed the same trend in the ligands upon
increasing temperature, the drop degrees of H2L0 and H2L1 are
25.0% and 21.3%, respectively, Similarly, the average slopes of
the linear functions of the two ligands are also much smaller (-
0.40% and -0.31% per °C for H2L0 and H2L1, respectively) than
that of the cages (Figure S33 and Figure 6c). These results reveal
that the dynamic molecular rotors in the cages are more sensitive
than in the free ligands. Impressively, the temperature cycling
experiments indicate highly reversible rotation of phenyl ring
rotors under temperature control over ten runs attempted (Figure
6d and Figure S34-S35), highlighting that the dynamic AIE
molecular rotors can operate effectively in the highly stable
coordination cages.
simulations to estimate the rotational energy barriers of phenyl
ring rotors of NUS 100-101. Figure 7c-d show the potential energy
of one phenyl ring rotor of each cage as a function of rotor angle
varying from 0 to 360° with an interval of 15°, in which two stable
conformations were observed at around 135 and 315°. On the
contrary, less favorable conformations could be observed at
around 45 and 225° with relative higher energy barriers. The
rotation energy barriers were calculated to be 9.66 kcal mol-1 in
NUS-100 and 9.42 kcal mol-1 in NUS-101 at 25 °C, which are
lower than that of the confined molecular rotors in our previous
POFs NUS-24 (14.0 kcal mol-1)[17c] and NUS-25 (12.1 kcal mol-
1).[17d] Moreover, we performed the potential energy calculations
for one phenyl ring rotor under 60 °C, and obtained the similar
trend of the potential energy with 25 °C. However, the rotation
energy barriers decrease to 7.54 kcal mol-1 in NUS-100 and 7.73
kcal mol-1 in NUS-101 (Figure 7c-d), indicating that the phenyl ring
rotors are sensitive to temperature. Therefore, the computational
and experimental results are consistent and reasonable to
interpret the nonradiative energy release pathways of the cages
via multiple external stimuli-responsive environments.
To further understand the temperature response of AIE
molecular rotors in the two cages, we performed the variable-
temperature 1H-NMR experiments to test any possible proton
signal change for the phenyl ring rotors of NUS 100-101. As
expected, we can clearly observe that the proton signals of the
phenyl ring rotors (He in NUS-100 and Hg in NUS-101) exhibit
downfield shifts from 25 to 65 °C, with ∆He of 0.04 ppm in NUS-
100 and ∆Hg of 0.02 ppm in NUS-101 (Figure 7a-b), confirming
the faster rotary speed of the phenyl ring rotors under higher
temperatures.[21b] Combined with viscosity-responsive, guest
size-dependent-responsive,
and
temperature-responsive
fluorescence change behaviors, the above results strongly
confirm the existence of dynamic molecular rotors in the
coordination cages.
Figure 8. ESI-TOF-MS spectra of NUS-100 (a) and NUS-101 (b) in
methanol/water solutions with different pH values. The cytotoxicity assay results
of NUS-100 and NUS-101 in NIH-3T3 cells (c) and HeLa cells (d).
The suitable molecular size and AIE fluorescence behavior
of NUS 100-101 prompt us to evaluate their feasibility as
molecular fluorescent probes in live-cell imaging. We initially
studied their structural stability, fluorescence emission, and
cytotoxicity in aqueous solutions, which are the prerequisites for
in-vitro bio-imaging. Impressively, NUS 100-101 are highly stable
in aqueous solutions with acidic, neutral, and weak basic
conditions, confirmed by the ESI-TOF-MS data of their aqueous
solutions with pH values from 2 to 10 (Figure 8a-b). Furthermore,
fluorescence emission spectra of the two cages keep unchanged
under similar conditions (Figure S36-S37). In addition, time-
Figure 7. Partial 1H NMR spectra of NUS-100 (a) and NUS-101 (b) at different
temperatures. The theoretical calculations of energy barriers of phenyl ring
rotors (red color in molecular models) in NUS-100 (c) and NUS-101 (d) at 25 or
60 °C.
Given the significance of the above findings, it is essential
to elucidate why the phenyl ring rotors have such a dynamic
behavior. Accordingly, we performed molecular dynamics (MD)
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 9. The CLSM imaging in HeLa cells using NUS-100 (a), H2L0 (b), NUS-101 (c), and H2L1 (d) compared with Lysotracker Red after 24 h incubation. (i)
HeLa cells CLSM images. (ii) Lysotracker Red images. (iii) Colocalized images. (iv) Overlay images with bright field. (v) Scatter plots and Pearson’s R values.
(e) The CLSM imaging in four cancerous cells and four non-cancerous cells using NUS-100 under the same conditions. All scale bars are 20 μm.
resolved photoluminescence measurement (TRPL) results show
that the lifetimes of NUS-100 and NUS-101 are 0.97 and 1.03 ns
at neutral aqueous solutions (Figure S38), respectively, and the
quantum yields (QY) of NUS-100 and NUS-101 under neutral
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
aqueous solutions were determined to be 19.0 ± 1% and 28.8 ± Conclusion
1%, respectively. Importantly, cell viability assay of NUS 100-101
In summary, we have designed and synthesized highly stable
zirconium (IV)-based coordination cages containing AIE
molecular rotors for live-cell imaging. The phenyl ring rotors
located on the external surface of the coordination cages can be
effectively restricted by various stimuli, including temperature,
viscosity, and guests with various molecular sizes, leading to
strong AIE characteristics. Such highly stable and fluorescent
supramolecular coordination cages show the following
advantages over other supramolecular coordination assemblies
in biological applications: (1) high stability in cellular environment,
(2) bright fluorescence based on the impressive AIE
characteristics, and (3) low cytotoxicity suitable for biological
applications. As a proof-of-concept study, we have successfully
applied the two cages for universal live-cell imaging. We envision
that such zirconium (IV)-based fluorescent coordination cages
have potential applications in various biological applications
including diagnosis of living cancer cells, drug delivery, and
photodynamic therapy.
was conducted in NIH-3T3 and HeLa cell lines by incubating for
24 h at concentrations ranging from 1 μM to 20 mΜ by
methylthiazolyldiphenyltetrazolium bromide (MTT) method.[14] As
shown in Figure 8c-d, no significant toxicity was noted in both two
cell lines with viabilities higher than 80% under all conditions. All
these data indicate that the two coordination cages have high
stability, strong fluorescence, and low cytotoxicity, making them
suitable candidates for in-vitro bio-imaging applications.
Inspired by the above results, NUS 100-101 were further
applied into HeLa cell line to evaluate their imaging capabilities.
After incubating for 24 h at 10 μM, strong signals from inside
cytoplasm region were observed in CLSM images (Figure 9a and
9c, Figure S39). Notably, non-specific targeting nanomaterials
tend to be internalized through endocytosis, and lysosomes are
the last compartment of the endocytic pathway, leading to
distribution of nanomaterials in lysosomes. The location of signal
was further studied by colocalization analysis with Lysotracker™
Red DND-99. The Pearson’s R values for NUS-100 and NUS-101
are 0.84 and 0.76, respectively, indicating the major distribution
in lysosomes. The internalization process was captured by time
lapse imaging using live-imaging chamber. Signal intensity
gradually increased from 6 to 24 h with very few aggregates
outside cells, showing excellent imaging ability (Figure S40-S41).
Additionally, H2L0 and H2L1 were also incubated with HeLa cells
for 24 h to compare the imaging capability between ligands and
cages (Figure 9b and 9d). CLSM images were taken under the
same conditions followed by colocalization analysis, showing
much weaker fluorescence and lower Pearson’s R values for H2L0
(Figure 9b) and H2L1 (Figure 9d). Moreover, NUS-100 and NUS-
101 can also be applied into imaging of MCF-7 and MDA-MB-231
cancerous cells, as well as NIH-3T3 normal cells (Figure S42-
S43). In order to illustrate the universal live-cell imaging ability of
the cages, we took NUS-100 as an example and investigated a
broader live-cell imaging screening for four cancerous cells
including HeLa, MCF-7, MDA-MB-231 and C6, and four non-
cancerous cells such as COS-7, NIH-3T3, HEK-293T and one
macrophage cells RAW 264.7 under the same conditions (Figure
S44). As shown in Figure 9e, CLSM images showed that both
cancerous cells and non-cancerous cells are bright, indicating
that the AIE-active cages are potential fluorescent probes to
universally label a wide range of cell lines. However, due to the
difference in endocytosis abilities and the different intracellular
environment (i.e., different viscosity), the fluorescence intensity
varies in different cell lines. Furthermore, we also performed the
control experiment of HeLa cells imaging using NR-cage, which
exhibited much weaker imaging than that of NUS 100-101 (Figure
S45-S46), further highlighting that the AIE molecular rotors in
NUS 100-101 are beneficial for live-cell imaging. Our findings
have successfully demonstrated the universal live-cell imaging
using highly stable and fluorescent zirconium (IV) coordination
cages with AIE molecular rotors, paving the way for the biological
applications of these coordination cages.
Experimental Section
Experimental details on the synthesis and characterization of H2L0,
H2L1, and NUS 100-101, AIE experiments, temperature-response,
viscosity-response, and size-effect of guest molecules, in-vitro
cancer cell imaging, MD simulations and TD-DFT calculations are
provided in the Supplementary Information.
Acknowledgements
This work was supported by the National Research Foundation
Singapore (NRF2018-NRF-ANR007 POCEMON), the Ministry of
Education - Singapore (MOE AcRF Tier 1 R-279-000-540-114, R-
279-000-482-133, Tier 2 MOE2018-T2-2-148), the Agency for
Science, Technology and Research (IRG A1783c0015, IAF-PP
A1789a0024), National Natural Science Foundation of China (No.
21676094 and 21576058), Fundamental Research Funds for the
Central Universities (2017MS083), and the technical support from
National Supercomputer Center in Guangzhou (Tianhe-2).
Keywords: self-assembly • coordination cage • aggregation
induced emission • molecular rotor • live-cell imaging
[1]
a) T. R. Cook, Y.-R. Zheng, P. J. Stang, Chem. Rev. 2013, 113, 734-777.
b) T. R. Cook, P. J. Stang, Chem. Rev. 2015, 115, 7001-7045. c) W. Wang, Y.-
X. Wang, H.-B. Yang, Chem. Soc. Rev. 2016, 45, 2656-2693.
[2]
2148-2167. b) K. L. Haas, K. J. Franz, Chem. Rev. 2009, 109, 4921-4960.
[3] a) D. Zhang, T. K. Ronson, J. Mosquera, A. Martinez, L. Guy, J. R.
a) L. Xu, Y.-X. Wang, L.-J. Chen, H.-B. Yang, Chem. Soc. Rev. 2015, 44,
Nitschke, J. Am. Chem. Soc. 2017, 139, 6574-6577. b) A. B. Grommet, J. B.
Hoffman, E. G. Percástegui, J. Mosquera, D. J. Howe, J. L. Bolliger, J. R.
Nitschke, J. Am. Chem. Soc. 2018, 140, 14770-14776.
[4]
a) M. Zhang, M. L. Saha, M. Wang, Z. Zhou, B. Song, C. Lu, X. Yan, X.
Li, F. Huang, S. Yin, P. J. Stang, J. Am. Chem. Soc. 2017, 139, 5067-5074. b)
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
J. Dong, C. Tan, K. Zhang, Y. Liu, P. J. Low, J. Jiang, Y. Cui, J. Am. Chem. Soc.
2017, 139, 1554-1564. c) J. Dong, Y. Zhou, F. Zhang, Y. Cui, Chem. - Eur. J.
2014, 20, 6455-6461.
Zhou, H. Wang, X. Li, M. L. Saha, Y.-W. Zhong, P. J. Stang, J. Am. Chem. Soc.
2018, 140, 7723-7729.
[5]
a) X. Yan, T. R. Cook, P. Wang, F. Huang, P. J. Stang, Nat. Chem. 2015,
7, 342-348. b) X. Yan, M. Wang, T. R. Cook, M. Zhang, M. L. Saha, Z. Zhou, X.
Li, F. Huang, P. J. Stang, J. Am. Chem. Soc. 2016, 138, 4580-4588.
[6]
a) Y. Ueda, H. Ito, D. Fujita, M. Fujita, J. Am. Chem. Soc. 2017, 139,
6090-6093. b) H. Takezawa, T. Kanda, H. Nanjo, M. Fujita, J. Am. Chem. Soc.
2019, 141, 5112-5115. c) C. J. Brown, F. D. Toste, R. G. Bergman, K. N.
Raymond, Chem. Rev. 2015, 115, 3012-3035. d) J. Jiao, C. Tan, Z. Li, Y. Liu,
X. Han, Y. Cui, J. Am. Chem. Soc. 2018, 140, 2251-2259.
[7]
a) N. Sakai, S. Matile, Angew. Chem., Int. Ed. 2008, 47, 9603-9607;
Angew. Chem. 2008, 120, 9745-9749. b) M. Jung, H. Kim, K. Baek, K. Kim,
Angew. Chem. Int. Ed. 2008, 47, 5755-5757; Angew. Chem. 2008, 120, 5839-
5841.
[8]
a) G. Li, W. Yu, Y. Cui, J. Am. Chem. Soc. 2008, 130, 4582-4583. b) K.
Wu, K. Li, Y.-J. Hou, M. Pan, L.-Y. Zhang, L. Chen, C.-Y. Su, Nat. Commun.
2016, 7, 10487.
[9]
a) P. Mal, B. Breiner, K. Rissanen, J. R. Nitschke, Science 2009, 324,
1697-1699. b) B. Chen, J. J. Holstein, S. Horiuchi, W. G. Hiller, G. H. Clever, J.
Am. Chem. Soc. 2019, 141, 8907-8913. c) K. Yazaki, M. Akita, S. Prusty, D. K.
Chand, T. Kikuchi, H. Sato, M. Yoshizawa, Nat. Commun. 2017, 8, 15914. d) X.
Jing, C. He, Y. Yang, C. Duan, J. Am. Chem. Soc. 2015, 137, 3967-3974.
[10] G. Feng, B. Liu, Acc. Chem. Res. 2018, 51, 1404-1414.
[11] Y. Fang, X. Lian, Y. Huang, G. Fu, Z. Xiao, Q. Wang, B. Nan, J.-P. Pellois,
H.-C. Zhou, Small 2018, 14, 1802709.
[12] R. Hu, N. L. C. Leung, B. Z. Tang, Chem. Soc. Rev. 2014, 43, 4494-4562.
[13] a) C. Creutz, M. Chou, T. L. Netzel, M. Okumura, N. Sutin, J. Am. Chem.
Soc. 1980, 102, 1309-1319.
[14] J. Zhou, Y. Zhang, G. Yu, M. R. Crawley, C. R. P. Fulong, A. E. Friedman,
S. Sengupta, J. Sun, Q. Li, F. Huang, T. R. Cook, J. Am. Chem. Soc. 2018, 140,
7730-7736.
[15] G. Liu, Y. Di Yuan, J. Wang, Y. Cheng, S. B. Peh, Y. Wang, Y. Qian, J.
Dong, D. Yuan, D. Zhao, J. Am. Chem. Soc. 2018, 140, 6231-6234.
[16] a) J. Luo, Z. Xie, J. W. Y. Lam, L. Cheng, H. Chen, C. Qiu, H. S. Kwok,
X. Zhan, Y. Liu, D. Zhu, B. Z. Tang, Chem. Commun. 2001, 1740-1741. b) Y.
Hong, J. W. Y. Lam, B. Z. Tang, Chem. Soc. Rev. 2011, 40, 5361-5388. c) J.
Mei, N. L. C. Leung, R. T. K. Kwok, J. W. Y. Lam, B. Z. Tang, Chem. Rev. 2015,
115, 11718-11940.
[17] a) M. Zhang, G. Feng, Z. Song, Y.-P. Zhou, H.-Y. Chao, D. Yuan, T. T.
Y. Tan, Z. Guo, Z. Hu, B. Z. Tang, B. Liu, D. Zhao, J. Am. Chem. Soc. 2014,
136, 7241-7244. b) J. Dong, A. K. Tummanapelli, X. Li, S. Ying, H. Hirao, D.
Zhao, Chem. Mater. 2016, 28, 7889-7897. c) J. Dong, K. Zhang, X. Li, Y. Qian,
H. Zhu, D. Yuan, Q.-H. Xu, J. Jiang, D. Zhao, Nat. Commun. 2017, 8, 1142. d)
J. Dong, X. Li, K. Zhang, Y. Di Yuan, Y. Wang, L. Zhai, G. Liu, D. Yuan, J. Jiang,
D. Zhao, J. Am. Chem. Soc. 2018, 140, 4035-4046.
[18] S. Lee, J. H. Lee, J. C. Kim, S. Lee, S. K. Kwak, W. Choe, ACS Appl.
Mater. Inter. 2018, 10, 8685-8691.
[19] D. Zhang, Y. Zhu, L. Liu, X. Ying, C.-E. Hsiung, R. Sougrat, K. Li, Y. Han,
Science 2018, 359, 675-679.
[20] a) M. K. Kuimova, G. Yahioglu, J. A. Levitt, K. Suhling, J. Am. Chem. Soc.
2008, 130, 6672-6673. b) L. Wang, Y. Xiao, W. Tian, L. Deng, J. Am. Chem.
Soc. 2013, 135, 2903-2906.
[21] a) M. Jin, T. S. Chung, T. Seki, H. Ito, M. A. Garcia-Garibay, J. Am. Chem.
Soc. 2017, 139, 18115-18121. b) J.-H. Tang, Y. Sun, Z.-L. Gong, Z.-Y. Li, Z.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915199
Angewandte Chemie International Edition
RESEARCH ARTICLE
RESEARCH ARTICLE
We report the self-assembly of highly
stable coordination cages from the
trinuclear zirconium vertices and
Dr. Jinqiao Dong, Yutong Pan, Dr. Heng
Wang, Dr. Kuiwei Yang, Dr. Lingmei Liu,
Prof. Zhiwei Qiao, Yi Di Yuan, Shing Bo
Peh, Dr. Jian Zhang, Dr. Leilei Shi, Prof.
Hong Liang, Prof. Yu Han, Prof. Xiaopeng
Li, Prof. Jianwen Jiang, Prof. Bin Liu,* and
Prof. Dan Zhao*
flexible
carboxyl-decorated
tetraphenylethylene (TPE) spacers.
The two water-stable cages exhibit
impressive
emission
aggregation
(AIE) characteristics
induced
Page No. – Page No.
contributed by the molecular rotors,
and can be employed as novel
biological fluorescent probes for live-
cell imaging.
Self-Assembly of Highly Stable
Zirconium (IV) Coordination Cages
with Aggregation Induced Emission
(AIE) Molecular Rotors for Live-Cell
Imaging
This article is protected by copyright. All rights reserved.