Communication
Antibacterial Therapy
Metal–Organic-Framework-Assisted In Vivo Bacterial
Metabolic Labeling and Precise Antibacterial Therapy
Duo Mao, Fang Hu, Kenry, Shenglu Ji, Wenbo Wu, Dan Ding, Deling Kong,* and Bin Liu*
Bacterial infection is one of the most serious physiological conditions threat-
ening human health. There is an increasing demand for more effective bacterial
diagnosis and treatment through noninvasive theranostic approaches. Herein,
a new strategy is reported to achieve in vivo metabolic labeling of bacteria
through the use of MIL-100 (Fe) nanoparticles (NPs) as the nanocarrier for
precise delivery of 3-azido-d-alanine (d-AzAla). After intravenous injection,
MIL-100 (Fe) NPs can accumulate preferentially and degrade rapidly within the
high H2O2 inflammatory environment, releasing d-AzAla in the process. d-AzAla
is selectively integrated into the cell walls of bacteria, which is confirmed by
fluorescence signals from clickable DBCO-Cy5. Ultrasmall photo­sensitizer NPs
with aggregation-induced emission characteristics are subsequently designed
to react with the modified bacteria through in vivo click chemistry. Through
photodynamic therapy, the amount of bacteria on the infected tissue can be sig-
nificantly reduced. Overall, this study demonstrates the advantages of metal–
organic-framework-assisted bacteria metabolic labeling strategy for precise
bacterial detection and therapy guided by fluorescence imaging.
Bacterial infection is one of the most prominent underlying
causes of many severe diseases, such as septic arthritis, derma-
tosis, and inflammatory bowel disease, which have increasingly
raised medical and public concerns across the world. Up to
now, rapid and effective in vivo bacterial detections and treat-
ments have hardly been realized due to lack of well-developed
methods. Clinical detections of most bacteria are still based
on conventional tissue biopsies and cultures, which are time-
consuming and highly inefficient. Recently, antibiotics such
as vancomycin were reported as targeting ligands to help
selectively label and kill different kinds of bacteria.[1] Unfor-
tunately, due to the abuse of antibiotics, incidence of multire-
sistant bacterial infections continuously increases. For example,
methicillin-resistant Staphylococcus aureus (MRSA) has given
Dr. D. Mao, Dr. F. Hu, Dr. Kenry, Dr. W. Wu, Prof. B. Liu
Department of Chemical and Biomolecular Engineering
National University of Singapore
4 Engineering Drive 4, Singapore 117585, Singapore
E-mail: cheliub@nus.edu.sg
S. Ji, Prof. D. Ding, Prof. D. Kong
State Key Laboratory of Medicinal Chemical Biology
Key Laboratory of Bioactive Materials
Ministry of Education and College of Life Sciences
Nankai University
Tianjin 300071, China
E-mail: kongdeling@nankai.edu.cn
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.201706831.
DOI: 10.1002/adma.201706831
Adv. Mater. 2018, 1706831 1706831  (1 of 7)
www.advmat.de
rise to several difficult-to-treat infections
in recent years.[2] As such, there is an
urgent need for alternative methods that
can specifically and noninvasively detect,
target, and kill bacteria without concern
about antibiotic resistance. Monoclonal
antibody and antibacterial peptide have
been demonstrated as effective bacteria
recognition systems via surface antigen
binding or electrostatic interaction.[3]
However, high cost and fast body clear-
ance are impeding the development of
specific recognition and bacteria killing
in vivo based on antigen and peptide.[4] In
this regard, effortless and reliable strate-
gies are highly desirable to identify and
kill infection-causing bacterial strains.
Metabolic biomolecular labeling tech-
nology has emerged as a highly powerful
tool with great potentials for cell and
pathogen targeting due to its capability to
decorate the surface of cells and pathogens
with chemical functional groups.[5] The technique makes use of
the delivery and uptake of synthetic sugars or amino acids to
attach chemical functional groups onto cell membrane, which
in turn provides connectors on the cell surface for further con-
jugation of fluorescent dyes or drugs. However, it is still chal-
lenging to achieve specific in vivo metabolic labeling due to
the difficulty in specific delivering of functionalized synthetic
sugars or amino acids into targeted cells in vivo. To realize
this goal, azide-functionalized sugar molecules were protected
with acetyl groups, and the resulting hydrophobic molecules
were loaded into polymer carriers, which could target tumor
via enhanced permeation and retention effect.[6] However, up
to date, no example of bioorthogonal reaction for in vivo bacte-
rial detection and treatment has been reported. Although click-
able group (e.g., azide and cyclooctyne) modified d-alanines are
able to be metabolically expressed on bacterial peptidoglycan
for subsequent bioorthogonal labeling, the high water solubility
of d-alanines makes them easy to be metabolized in vivo and
therefore very difficult to be delivered to inflamed lesions. It
is of great importance to develop different strategies to deliver
azide d-alanines to inflamed lesions.
Metal–organic frameworks (MOFs) have been considered as
a promising candidate as drug carriers because of their excel-
lent storage capacities, easy body clearance, and low cytotox-
icity.[7] More importantly, the versatile chemical properties of
MOFs endow them with the capability to regulate the delivery
of drugs, as triggered by various external stimuli.[8] Of all
responsive MOFs, the pH-responsive MOFs have been widely
investigated for antitumor therapy due their sensitivity to the
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

low-pH tumor environment.[9] In addition, various MOFs that
can be triggered by other stimuli factors (e.g., magnetic field,
ions, temperature, and light) have also been developed for regu-
lating the drug delivery process.[8] Therefore, it is promising to
design a delivery system based on MOFs to realize the
responsive release of metabolic labeling molecule in
the bacteria-infected tissues. In this study, we report a novel
bacterial detection and therapeutic strategy through the cou-
pling of MOFs with metabolic labeling technique. MIL-100 (Fe)
nanoparticles (NPs) were chosen as a carrier for the delivery
of metabolic labeling molecule 3-azido-d-alanine (d-AzAla) in
vivo. MIL-100 (Fe) is composed of iron(III) metal centers and
trimesic acid (TMA) ligands. The nontoxic nature of the com-
ponents endows it excellent biocompatibility. It also owns good
loading capacity for different kinds of drugs regardless of their
water solubility.[7b] In addition, iron(III) in MIL-100 (Fe) can
catalyze the decomposition of H2O2. When participating in the
reaction, the coordination between TMA and iron(III) breaks,
and the frameworks of MIL-100 (Fe) could be easily dissociated.
In this way, the dissociated MIL-100 (Fe) would specifically
release the encapsulated d-AzAla in H2O2 oversecreted tissues,
such as the inflammation site.
Recently, photodynamic therapy (PDT) has attracted much
attention in diagnosis and treatment of diseases. Under light
irradiation, most of the organic photosensitizers (PSs) can
generate singlet oxygen (1O2) through type II photochemical
reaction processes,[10] which can further cause bacterial lethal
injury through damage of DNA and cytoplasmic membrane of
bacteria.[11] Therefore, PDT is emerged as a novel approach for
non-antibiotic bacteria therapy in clinic.
The proposed strategy is shown in Scheme 1. In the first
step, nanoscale MIL-100 (Fe) is synthesized to load d-AzAla
through a simple absorbing process. Polymer pluronic F-127 is
further introduced to form d-AzAla@MIL-100 (Fe) NPs in order
to improve the dispersity of MOFs in physiological conditions
(Scheme 1A). Similar to enhanced permeation and retention
effect (EPR) of tumor, the blood vessels are also permeable in
the inflammation region due to mediation of various inflam-
matory factors including nitric oxide and peroxy­nitrite, leading
to passive targeting of NPs to bacteria-infected region.[12] After
intravenous (i.v.) injection, d-AzAla@MIL-100 (Fe) NPs will be
preferentially accumulated within the infected region of the
tissue of the mice due to the EPR effect, and the frameworks of
MIL-100 (Fe) are expected to be selectively damaged by H2O2,
which is highly secreted by immune cells in the infected region.
Subsequently, d-AzAla will be robustly released from the decom-
posed MIL-100 (Fe) and specifically taken up by the bacteria at
the infected region. During this process, unnatural azide groups
will be expressed on the bacterial wall (Scheme 1B). Thanks to
the elevated permeability of blood vessels in the inflammation
region, once dibenzocyclooctyne (DBCO)-modified PS NPs
are administrated, they are supposed to gradually accumulate

Scheme 1.  Schematic illustration of the proposed strategy of bacteria diagnosis and therapy by the H2O2-responsive MOFs assisted in vivo metabolic
labeling of bacteria. A) d-AzAla@MIL-100 (Fe) NPs are synthesized by using pluronic F-127 as a matrix to encapsulate the d-AzAla-loaded MIL-100 (Fe).
B) d-AzAla@MIL-100 (Fe) NPs accumulate at the site of the infected tissue and are decomposed in the presence of H2O2. Meanwhile, invading bacteria
internalize these released d-AzAla and express azide group on their cell wall. C) Ultrasmall US-TPETM NPs with dibenzocyclooctyne (DBCO) group bind
with bacteria through click reaction, and specific tracking and effective photodynamic therapy (PDT) of bacteria could be achieved in the infected tissue.

Adv. Mater. 2018, 1706831 1706831  (2 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

around bacteria. Subsequently, selective fluorescence labeling
and precise bacteria killing can be achieved via bioorthogonal
reaction and PDT. In particular, we propose to use ultrasmall
2-(1-(5-(4-(1,2,2-tris(4-methoxyphenyl)­vinyl)­phenyl)thiophen-
2-yl)ethylidene)malononitrile NPs (US-TPETM NPs) with
aggregation-induced emission (AIE) characteristics as the labe-
ling reagent, in which DBCO-DSPE-PEG2000 forms an outer
shell in contact with water, and DBCO groups are exposed to
the surface of NPs. Hydrophobic TPETM molecules are encap-
sulated in the interior, which can not only specifically track
MRSA in the infected skin tissues, but also precisely eradicate
bacteria and reduce acute inflammation upon light irradiation
(Scheme 1C). Other unattached DBCO–TPETM NPs will be
gradually cleared by body clearance process.
MIL-100 (Fe) was synthesized based on a hydrothermal
process modified from literatures,[7b] and was further charac-
terized by transmission electron microscopy (TEM), dynamic
light scattering (DLS), and X-ray diffraction (XRD). As shown
in Figures S1 and S2 in the Supporting Information, MIL-
100 (Fe) has a clear crystal morphology and the XRD pattern
matches well with the simulated spectrum. After loading with
d-AzAla, the obtained d-AzAla@MIL-100 (Fe) exhibits similar
morphology to MIL-100 (Fe) (Figure 1A; Figure S1, Supporting
Information). According to thermogravimetric analysis (TGA),
the d-AzAla encapsulation loading in MIL-100 (Fe) was calcu-
lated to be 18.8 wt% in ethanol solution (Figure 1B). To ensure
the stability of NPs under physiological conditions, the surface
of d-AzAla@MIL-100 (Fe) was coated with polymer pluronic
F-127 through microemulsion. The TEM image shows that
MIL-100 (Fe) NPs possess round morphology and are well
dispersed in aqueous media, with an average diameter of 120 nm
(Figure 1C,F). After treatment with H2O2, the crystalline mor-
phology of d-AzAla@MIL-100 (Fe) NPs was destroyed over time
(Figure 1E), indicating the activation of d-AzAla@MIL-100 (Fe)
NPs in the presence of H2O2 solution. The same result was also
observed by the XRD analysis (Figure S2A, Supporting Infor-
mation). Since the reaction between d-AzAla@MIL-100 (Fe)
and H2O2 is accompanied by the degradation of MIL-100 (Fe),
the reaction can be studied by monitoring the release of TMA.
As shown in Figure S2B in the Supporting Information, in the
presence of H2O2 at pathological concentration (50 × 10−6 m),[13]
69.1% of d-AzAla@MIL-100 (Fe) was decomposed after 12 h
incubation in phosphate buffered saline (PBS) solution
according to the released TMA monitored by high-performance
liquid chromatography. However, when the concentration of
incubated H2O2 was decreased to normal physiological condi-
tion (0.5 × 10−6 m), only 12.3% of d-AzAla@MIL-100 (Fe) was
decomposed. These results indicate that d-AzAla@MIL-100 (Fe)
can be selectively activated in the presence of H2O2 at patholog-
ical concentration rather than at physiological condition. After
MIL-100 (Fe) activation, d-AzAla was robustly released. Since
d-AzAla can hardly be quantified directly, the release kinetic of
d-AzAla was measured by the ninhydrin assay, in which process
ninhydrin turns to a detectable purple color after reaction with
the amino acid. After incubation with H2O2 at 50 × 10−6 m, 60%
of d-AzAla was released from the activated MIL-100 (Fe) NPs
within 10 h, which was significantly higher than that incubated
with H2O2 at 0.5 × 10−6 m, indicating effective release of d-AzAla

Figure 1.  Characterization, activation, and release of d-AzAla@MIL-100 (Fe) NPs. A) TEM image of d-AzAla@MIL-100 (Fe). B) TGA curves of d-AzAla,
MIL-100 (Fe), and d-AzAla@MIL-100 (Fe). C) TEM image of d-AzAla@MIL-100 (Fe) NPs coated with pluronic F-127. Inset in panel (C) shows
the magnified TEM image of a single d-AzAla@MIL-100 (Fe) NP. D) d-AzAla release from d-AzAla@MIL-100 (Fe) NPs in the presence of H2O2 at
50 and 0.5 × 10−6 m, as measured by ninhydrin assay. E) TEM image of d-AzAla@MIL-100 (Fe) NPs after H2O2 treatment. F) DLS size distributions of
d-AzAla@MIL-100 (Fe) NPs before/after coating with pluronic F-127 and incubation with H2O2.

Adv. Mater. 2018, 1706831 1706831  (3 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

in the bacteria-infected environment (Figure 1D). In addition,
the water dispersity and framework damage of pluronic F-127-
coated d-AzAla@MIL-100 (Fe) NPs were also studied by DLS
(Figure 1F). Compared to the large hydrodynamic size of MIL-
100 (Fe) due to self-aggregation in water, the smaller average
size of pluronic F-127-coated d-AzAla@MIL-100 (Fe) NPs indi-
cated a better dispersity of the surface-coated NPs. After H2O2
treatment, the increase in NP size illustrated the damaged
structure of NPs, correlating well with the TEM images.
After demonstrating the negligible in vivo toxicity of the MIL-
100 (Fe) NPs by blood chemistry test for liver function (Figure
S3, Supporting Information), their in vivo delivery performance
was examined. A bacterial infection model was established by
subcutaneous inoculation of MRSA into the left side of the
backs of mice. To monitor the in vivo distribution of MOFs,
Cy5 was loaded into MIL-100 (Fe) to form Cy5@MIL-100 (Fe)
NPs. As shown in Figure S4 in the Supporting Information,
after i.v. injection of Cy5-loaded MIL-100 (Fe) NPs into the bac-
teria-bearing mice, the fluorescence signals indicate preferen-
tial accumulation of NPs within the inflammation region and
reticuloendothelial system organs such as liver, which is largely
different from the mice treated with free Cy5. This indicates an
effective target and delivery effect of MIL-100 (Fe) NPs to the
bacteria-infected region.
We next explored whether MIL-100 (Fe) NPs could assist the
hydrophilic d-AzAla to metabolically label bacteria in vivo. It is
well known that a large amount of H2O2 (≈500 nmol min−1) is
typically released by immune cells during acute bacterial infec-
tions.[14] As such, we hypothesized that the elevated level of
H2O2 in the infected tissue would activate MIL-100 (Fe) NPs
and, subsequently, release d-AzAla. To validate this hypothesis,
d-AzAla@MIL-100 (Fe) NPs were first injected subcutaneously
into the bacteria-infected site of the tissue from the mice. After
24 h, DBCO-Cy5 was subsequently i.v. injected into the bodies
of the mice and then monitored using fluorescence imaging.
As shown in Figure S5 in the Supporting Information, with
the clearance of DBCO-Cy5, a fluorescence signal was clearly
observed in the infected skin of mice treated with d-AzAla@
MIL-100 (Fe) NPs. However, the infected skin treated with
dimethylthiourea (a well-known H2O2 scavenger)[15] prior to
the injections of d-AzAla@MIL-100 (Fe) NPs and DBCO-Cy5
hardly yielded any fluorescence signal in the infected region.
These results indicated that the activation of MIL-100 (Fe) NPs
was indeed related to H2O2 level.
Encouraged by the excellent H2O2 responsiveness in vivo,
we further tested the systematic delivery performance of MIL-
100 (Fe) NPs through i.v. injection of d-AzAla@MIL-100 (Fe)
NPs into the MRSA-bearing mice, followed by i.v. injection of
DBCO-Cy5. As depicted in Figure 2A and Figure S6 (Supporting
Information), the infected tissue of mice treated with d-AzAla@
MIL-100 (Fe) NPs showed a significant DBCO-Cy5 accumula-
tion as compared to those of mice treated with free d-AzAla or
saline. In addition, the excellent bacteria labeling capability of
d-AzAla@MIL-100 (Fe) NPs was further confirmed by ex vivo
fluorescence imaging of different organs (Figure 2B). The results
show that the average fluorescence intensity of the infected skin
is at least 3.2-fold higher than that of other tissues (Figure S7,
Supporting Information). Due to the relatively low H2O2 level in

Figure 2.  In vivo metabolic labeling of bacteria using MIL-100 (Fe) NPs. A) Time-dependent in vivo fluorescence images of bacteria-bearing mice
pretreated with d-AzAla@MIL-100 (Fe) NPs, free d-AzAla or saline, followed by intravenous (i.v.) injection of DBCO-Cy5. Bacteria-infected regions are
marked with blue circles. B) Ex vivo fluorescence images of different tissues of mice at 24 h postinjection of DBCO-Cy5. 1: liver, 2:kidney, 3: spleen,
4: heart and lung, 5: intestine, 6: normal skin, 7: skin with bacterial infection. Subparts (a)–(c) represent mice pretreated with d-AzAla@MIL-100 (Fe)
NPs, free d-AzAla or saline, respectively. C) Fluorescence images of the infected skin slice of mice successively i.v. injected with d-AzAla@MIL-100 (Fe)
NPs and DBCO-Cy5. The image on the right indicates the enlarged view of the box shown on the left image.

Adv. Mater. 2018, 1706831 1706831  (4 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

normal tissue, no obvious fluorescence signal was observed in
other organs, suggesting that d-AzAla@MIL-100 (Fe) NPs were
selectively activated in the infected tissues. Moreover, the fact
that d-AzAla hardly exists in mammalian cells and can only be
incorporated by bacteria ensures a high infected skin-to-organ
fluorescence intensity ratio. The infected skin was also sliced,
stained and imaged using a confocal microscope. As shown in
Figure 2C and Figure S8 (Supporting Information), a substan-
tial amount of 4′,6-diamidino-2-phenylindole, dihydrochloride
(DAPI)-positive tiny dots that represent the DNA of MRSA could
be clearly observed around the normal cell nuclei. In fact, these
bacteria were specifically labeled with bright red color and could
be easily distinguished from other normal tissue regions with a
high imaging contrast.
Due to the excellent bacteria targeting and labeling perfor-
mance, we then sought to assess if efficient and selective in
vivo bacteria killing could be achieved by replacing DBCO-Cy5
with DBCO-modified antibacterial agents. Recently, we deve­
loped a series of PSs with AIE characteristics.[16] Different from
traditional PSs that exhibit quenched fluorescence and reduced
photosensitizing activities in aggregate state, AIE PSs exhibit
high brightness and strong 1O2 generation capability under
light irradiation when they are encapsulated into NPs.[17] With
excellent optical properties and 1O2 generation capability, AIE
PS NPs have been increasingly considered as a promising
candidate for tumor ablation and fluorescence contrast rea-
gents.[18] Motivated by this, herein, TPETM was developed as a
new far-red/near-infrared (FR/NIR) PS. Its chemical structure

Figure 3.  In vivo bacteria killing using ultrasmall TPETM nanoparticles (US-TPETM NPs) by a metabolic labeling strategy. A) Chemical structure of
TPETM. B) Size distribution of US-TPETM NPs as measured by dynamic light scattering (DLS). C) Measurement of 1O2 production of US-TPETM
NPs by ADBA (black) under white light irradiation. ABDA (red) solutions under white light irradiation were used as the control, where A0 and A are
the absorbance of ABDA at 378 nm. D) Time-dependent in vivo fluorescence images of bacteria-bearing mice pretreated with d-AzAla@MIL-100 (Fe)
NPs (Nano-Click group) or saline (Nano-Only group), respectively, followed by i.v. injection of US-TPETM NPs. E) Fluorescence images of the infected
skin slices of mice pretreated with d-AzAla@MIL-100 (Fe) NPs or saline, respectively, followed by i.v. injection of US-TPETM NPs. The areas within the
dotted lines indicate bacterium aggregates on the infected skin. F) Bacteria colony-forming unit (CFU) recovered from the infected skin (average ± the
standard error of the mean (SEM), *p < 0.05 analysis of variance (ANOVA)). G) Hemotoxylin and eosin stain of the infected skin slices subjected to
different treatment.

Adv. Mater. 2018, 1706831 1706831  (5 of 7) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
is shown in Figure 3A, and the detailed synthetic route and
characterization results are presented in Scheme S1 and
Figures S9–S14 in the Supporting Information.
The absorption spectrum of TPETM is located at 400–600 nm
with a molar absorption coefficient of 3.39 × 104 L mol−1 cm−1
at the absorption maximum of 460 nm (Figure S15A, Sup-
porting Information), which is redshifted as compared to most
AIE PSs. The fluorescence spectra of TPETM are shown in
Figure S15B (Supporting Information), and negligible fluo-
rescence is observed when TPETM is dissolved in dimethyl-
sulfoxide (DMSO). After gradual addition of water into the
DMSO solution, the fluorescence with an emission maximum
at 680 nm gradually increases, demonstrating the AIE charac-
teristic of TPETM. To achieve a better in vivo biocompatibility
and blood circulation, US-TPETM NPs were formulated by
using DBCO-DSPE-PEG2000 as a polymer matrix, which show
negligible in vitro dark toxicity (Figure S16, Supporting Infor-
mation). The average diameter of US-TPETM NPs is around
10 nm (Figure 3B; Figure S17A, Supporting Information),
which endows them with rapid body clearance (Figure S18,
Supporting Information). Meanwhile, US-TPETM NPs inherit
the broad absorption (400–600 nm; Figure S17B, black curve,
Supporting Information) and red fluorescence with an emis-
sion maximum of 680 nm (Figure S17B, red curve, Supporting
Information) of TPETM, which are beneficial for in vivo
imaging and PDT. In addition, the US-TPETM NPs showed
strong 1O2 generation with a 1O2 quantum yield of 10.1%
(Rose bengal as reference), as verified through the change in
the UV–vis spectrum of 9,10-anthracenediyl-bis(methylene)
dimalonic acid (ABDA, an 1O2 indicator) under white light
irradiation (Figure 3C). The ex vivo antibacterial effect of
US-TPETM NPs was also confirmed before their in vivo appli-
cation. In these experiments, MRSA was first incubated with
d-AzAla for 12 h, followed by the labeling with US-TPETM NPs
for 30 min. The imaging results showed that bacteria could be
easily detected due to their specific binding to US-TPETM NPs.
In addition, after white light irradiation, the bacteria-killing effi-
ciency of the NPs reaches more than 75%, indicating their great
potentials in in vivo bacteria ablation (Figure S19, Supporting
Information).
The in vivo bacteria-killing capability of US-TPETM NPs was
further evaluated based on MIL-100 (Fe) NP-assisted metabolic
labeling of bacteria. d-AzAla@MIL-100 (Fe) NPs were first i.v.
injected into the MRSA-bearing mice. After 24 h, US-TPETM
NPs were i.v. injected and monitored using fluorescence imaging
immediately. Sequential fluorescence images of the mice treated
with d-AzAla@MIL-100 (Fe) NPs and US-TPETM NPs exhibited
more intense fluorescence signals with longer half-life as com-
pared to those obtained from the mice treated with US-TPETM
NPs only (Figure 3D; Figure S20, Supporting Information).
Meanwhile, the biodistribution of the US-TPETM NPs in dif-
ferent organs also reflected this trend (Figure S21, Supporting
Information). The fluorescence images of infected tissue sec-
tions of mice treated with d-AzAla@MIL-100 (Fe) NPs revealed
a large amount of US-TPETM NP aggregates around MRSA as
compared to those of mice treated with US-TPETM NPs only.
This suggests the pivotal role of MIL-100 (Fe) NPs in d-AzAla
delivery as well as successful in vivo bacteria labeling in the
inflammation environment (Figure 3E; Figure S22, Supporting
Adv. Mater. 2018, 1706831 1706831  (6 of 7)
www.advmat.de
Information). Furthermore, we investigated whether these US-
TPETM NP-labeled bacteria could be killed through photody-
namic therapy in vivo. After the in vivo labeling, the infected skin
of bacteria-bearing mice was further irradiated by white light for
10 min. The mice treated with US-TPETM NPs only or saline
only were used as the control groups. After 7 d, the infected
skins were excised and the number of bacteria on the skin was
tittered. As expected, the number of bacteria on the infected skin
of mice treated with d-AzAla@MIL-100 (Fe) NPs and US-TPETM
NPs was significantly lower than those in the control groups
(Figure 3F). The skins were also collected and sliced at day 14
for histological analysis. Hemotoxylin and eosin staining results
showed that the infected skin of mice with d-AzAla@MIL-100
(Fe) NPs and US-TPETM NPs had less inflammatory cell infil-
tration and higher skin recovery as compared to the mice in the
control groups. These results clearly demonstrate that MOF-
assisted administration approach is effective in in vivo antibac-
terial therapy (Figure 3G; Figure S23, Supporting Information).
In summary, we designed a novel two-step strategy for spe-
cific in vivo imaging of bacteria based on the metabolic labeling
technique for image-guided antibacterial therapy. MIL-100 (Fe)
NPs with a diameter of 120 nm were first formulated as d-AzAla
carriers, which could be degraded in the presence of H2O2 and,
consequently, the loaded molecules were gradually released
over time. Using the H2O2-responsive carrier, d-AzAla could be
effectively delivered into bacteria-infected tissues and they could
be specifically taken up by bacteria and lightened up by DBCO-
Cy5 for fluorescence imaging. To the best of our knowledge,
it is the first time that incorporation of unnatural functional
group into bacteria has been realized in vivo, which provides a
unique platform for subsequent antibacterial treatment. Using
ultrasmall US-TPETM NPs which exhibit intense red/NIR fluo-
rescence and strong photosensitizing capability as an example,
we developed a non-antibiotic tool for image-guided antibacte-
rial therapy. After successive injection of d-AzAla@MIL-100
(Fe) NPs and US-TPETM NPs, followed by white light illumi-
nation, the number of bacteria in the infected tissue decreased
significantly and the inflammation was successfully relieved.
Considering the unique advantages of simultaneous bacterial
detection and treatment, we anticipate that our strategy will
find wide applications in preclinical research.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
D.M. and F.H. contributed equally to this work. The authors thank the
Singapore NRF Competitive Research Program (Grant No. R279-000-
483-281), NRF Investigatorship (Grant No. R279-000-444-281), National
University of Singapore (Grant No. R279-000-482-133), the National
Science Foundation of China (Grants Nos. 81220108015 and 51622305),
and the Program for Changjiang Scholars and Innovative Research Team
in University (IRT13023) for financial support. All animal studies were
performed in compliance with the guidelines set by Tianjin Committee
of Use and Care of Laboratory Animals and the overall project protocols
were approved by the Animal Ethics Committee of Nankai University.
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
Conflict of Interest
The authors declare no conflict of interest.
Keywords
aggregation-induced emission, antibacterial therapy, bacteria imaging,
metabolic labeling, metal–organic frameworks
Received: November 22, 2017
Revised: December 22, 2017
Published online:
[1] M. Van Oosten, T. Schäfer, J. A. Gazendam, K. Ohlsen,
E. Tsompanidou, M. C. De Goffau, H. J. Harmsen, L. M. Crane,
E. Lim, K. P. Francis, Nat. Commun. 2013, 4, 2584.
[2] B. Spellberg, J. G. Bartlett, D. N. Gilbert, New Engl. J. Med. 2013,
368, 299.
[3] a) G. Roscia, C. Falciani, L. Bracci, A. Pini, Curr. Protein Pept.
Sci. 2013, 14, 641; b) K. C. Black, T. S. Sileika, J. Yi, R. Zhang,
J. G. Rivera, P. B. Messersmith, Small 2014, 10, 169; c) R. Chen,
M. D. Willcox, K. K. K. Ho, D. Smyth, N. Kumar, Biomaterials 2016,
85, 142; d) S. M. Lehar, T. Pillow, M. Xu, L. Staben, K. K. Kajihara,
R. Vandlen, L. DePalatis, H. Raab, W. L. Hazenbos, J. H. Morisaki,
Nature 2015, 527, 323.
[4] P. Sahariah, K. K. Sørensen, M. A. Hjálmarsdóttir,
Ó. E. Sigurjónsson, K. J. Jensen, M. Másson, M. B. Thygesen,
Chem. Commun. 2015, 51, 11611.
[5] a) E. Kuru, S. Tekkam, E. Hall, Y. V. Brun, M. S. Van Nieuwenhze,
Nat. Protoc. 2015, 10, 33; b) P. Shieh, M. S. Siegrist, A. J. Cullen,
C. R. Bertozzi, Proc. Natl. Acad. Sci. USA 2014, 111, 5456; c) H. Wu,
J. Yang, J. Šečkutė, N. K. Devaraj, Angew. Chem., Int. Ed. 2014, 53,
Adv. Mater. 2018, 1706831 1706831  (7 of 7)
www.advmat.de
5805; d) Y. Yuan, S. Xu, X. Cheng, X. Cai, B. Liu, Angew. Chem., Int.
Ed. 2016, 55, 6457.
[6] a) H. Wang, M. Gauthier, J. R. Kelly, R. J. Miller, M. Xu,
W. D. O’Brien, J. Cheng, Angew. Chem., Int. Ed. 2016, 128, 5542;
b) H. Wang, L. Tang, Y. Liu, I. T. Dobrucka, L. W. Dobrucki, L. Yin,
J. Cheng, Theranostics 2016, 6, 1467.
[7] a) P. Horcajada, C. Serre, M. Vallet-Regí, M. Sebban, F. Taulelle,
G. Férey, Angew. Chem., Int. Ed. 2006, 118, 6120; b) P. Horcajada,
T. Chalati, C. Serre, B. Gillet, C. Sebrie, T. Baati, J. F. Eubank,
D. Heurtaux, P. Clayette, C. Kreuz, Nat. Mater. 2010, 9, 172.
[8] M. X. Wu, Y. W. Yang, Adv. Mater. 2017, 29, 1606134.
[9] I. A. Lázaro, S. Haddad, S. Sacca, C. Orellana-Tavra,
D. Fairen-Jimenez, R. S. Forgan, Chem 2017, 2, 561.
[10] W. Fan, P. Huang, X. Chen, Chem. Soc. Rev. 2016, 45, 6488.
[11] a) S. Menezes, M. Capella, L. Caldas, J. Photochem. Photobiol., B
1990, 5, 505; b) B. S. Hass, R. B. Webb, Mutat. Res., Fundam. Mol.
Mech. Mutagen. 1979, 60, 1.
[12] H. Maeda, Proc. Jpn. Acad., Ser. B 2012, 88, 53.
[13] R. Weinstain, E. Savariar, C. Felsen, R. Tsien, J. Am. Chem. Soc.
2014, 136, 874.
[14] a) P. A. Hyslop, D. B. Hinshaw, I. U. Scraufstatter, C. G. Cochrane,
S. Kunz, K. Vosbeck, Free Radical Biol. Med. 1995, 19, 31;
b) R. Weinstain, E. N. Savariar, C. N. Felsen, R. Y. Tsien, J. Am.
Chem. Soc. 2014, 136, 874; c) D. Blake, R. Allen, J. Lunec, Br. Med.
Bull. 1987, 43, 371.
[15] K. M. Toth, J. M. Harlan, C. J. Beehler, E. M. Berger, N. B. Parker,
S. L. Linas, J. E. Repine, Free Radical Biol. Med. 1989, 6, 457.
[16] a) G. Feng, R. T. Kwok, B. Z. Tang, B. Liu, Appl. Phys. Rev. 2017,
4, 021307; b) W. Wu, D. Mao, F. Hu, S. Xu, C. Chen, C. J. Zhang,
X. Cheng, Y. Yuan, D. Ding, D. Kong, Adv. Mater. 2017, 29,
1700548; c) D. Wang, H. Su, R. T. Kwok, G. Shan, A. Leung, M. Lee,
H. H. Sung, I. D. Williams, J. W. Lam, B. Z. Tang, Adv. Funct. Mater.
2017, 27, 1704039.
[17] Y. Yuan, G. Feng, W. Qin, B. Z. Tang, B. Liu, Chem. Commun. 2014,
50, 8757.
[18] S. Chen, H. Wang, Y. Hong, B. Z. Tang, Mater. Horiz. 2016, 3, 283.
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim