Visualizing Microglia with a Fluorescence Turn-On Ugt1a7c Substrate

Article abstract of DOI:10.1002/anie.201903058

Microglia, the brain-resident macrophage, are involved in brain development and contribute to the progression of neural disorders. Despite the importance of microglia, imaging of live microglia at a cellular resolution has been limited to transgenic mice.

Full text of DOI:10.1002/anie.201903058

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201903058
German Edition: DOI: 10.1002/ange.201903058
Fluorescent Probes Very Important Paper
Visualizing Microglia with a Fluorescence Turn-On Ugt1a7c Substrate
Beomsue Kim+, Masahiro Fukuda+, Jung-Yeol Lee, Dongdong Su, Srikanta Sanu,
Aymeric Silvin, Audrey T. T. Khoo, Taejoon Kwon, Xiao Liu, Weijie Chi, Xiaogang Liu,
Sejong Choi, Diana S. Y. Wan, Sung-Jin Park, Jin-Soo Kim, Florent Ginhoux, H. Shawn Je,* and
Young-Tae Chang*
Abstract: Microglia, the brain-resident macrophage, are
involved in brain development and contribute to the progres-
sion of neural disorders. Despite the importance of microglia,
imaging of live microglia at a cellular resolution has been
limited to transgenic mice. Efforts have therefore been
dedicated to developing new methods for microglia detection
and imaging. Using a thorough structure–activity relationships
study, we developed CDr20, a high-performance fluorogenic
chemical probe that enables the visualization of microglia both
in vitro and in vivo. Using a genome-scale CRISPR-Cas9
knockout screen, the UDP-glucuronosyltransferase Ugt1a7c
was identified as the target of CDr20. The glucuronidation of
CDr20 by Ugt1a7c in microglia produces fluorescence.
protein (TSPO) has been also utilized as a pharmacological
target to label activated microglia.[7] However, chemical
ligands of TSPO are not ideal for labeling pan-microglial
populations, as TSPO is ubiquitously expressed in the brain
under normal conditions.[7]
Therefore, we previously developed CDr10 using diversity-
oriented fluorescent molecule libraries (DOFLs).[8] However,
CDr10 failed to label microglia in brain tissues due to the less
stringent conditions of cell-based screening using purified
microglia.[8] In this study, we applied a focused fluorescent
library to confluent cultures of mouse primary glial populations
containing microglia, astrocytes, and oligodendrocytes with
neurons,[9] which provided a more stringent condition to
identify suitable microglia probes that might work in vivo
(Figure 1A). We screened 880 chemical probes derived from
monostyryl-containing asymmetric BODIPY backbones
(BDs), from which CDr10 was derived.[10] Using unbiased
library screening, we identified one derivative named CDr20
(cell designation red 20; Abs/Emi: 568/600 nm in MeOH; e:
53547mÀ1 cmÀ1) that clearly overlapped with Csf1r–EGFP
signals; EGFP is expressed in microglia (Figure 1B).[11]
Because CDr20 has a 3-hydroxy-4-methoxystyryl moiety,
we evaluated the combinatorial structure–activity relationship
profiles of 36 related styryl structures using epi-fluorescence
imaging and flow cytometry analysis (Supporting Information,
Table S1 and Figures S1 and S2). A hydroxyl group in any
position of the benzene moiety (compounds 2–4) was indis-
pensable for labeling microglia by increasing the fluorescence
M
icroglia, the brain-resident macrophage, are known to be
involved in brain development and to contribute to the
progression of neural disorders, such as stroke, autism, and
Alzheimerꢀs disease.[1,2] Despite growing interest, microglia
examinations have been restricted to either postmortem CNS
tissue histopathology or fluorescent imaging in transgenic
mice that express fluorescent proteins from microglia line-
ages.[3] Therefore, efforts have been dedicated to developing
methods that can identify microglia populations.
Isolectin B4 (IB4), the plant lectin from Griffonia sim-
plicifolia, has been widely used to identify microglia in both
live and fixed brain tissues.[4] However, due to its strong
affinity for perivascular cells and some neurons, IB4-based
staining is not very specific for microglia.[5,6] The translocator
[*] Dr. J.-Y. Lee, X. Liu, Prof. Y.-T. Chang
Department of Chemistry,
X. Liu, Prof. Y.-T. Chang
Center for Self-assembly and Complexity,
Institute for Basic Science (IBS), Pohang 37673 (Korea)
Pohang University of Science and Technology, Pohang 37673 (Korea)
E-mail: ytchang@postech.ac.kr
Dr. B. Kim,[+] Dr. J.-Y. Lee, Dr. D. Su, Dr. S. Sanu, D. S. Y. Wan,
Dr. W. Chi, Prof. X. Liu
Singapore University of Technology and Design
Singapore 487372 (Singapore)
Dr. S.-J. Park, Prof. Y.-T. Chang
Singapore Bioimaging Consortium,
Agency for Science Technology and Research (A*STAR)
S. Choi
Department of Chemistry, Seoul National University
Seoul 08826 (Korea)
Singapore 138667 (Singapore)
Dr. M. Fukuda,[+] A. T. T. Khoo, Prof. H. S. Je
Program in Neuroscience and Behavioral Disorders,
Duke-NUS Medical School, Singapore 169857 (Singapore)
E-mail: shawn.je@duke-nus.edu.sg
Prof. J.-S. Kim
Center for Genome Engineering, IBS, Daejeon 34047 (Korea)
Dr. J.-Y. Lee
Present address: New drug discovery center, DGMIF
Daegu 41061 (Korea)
Dr. A. Silvin, Prof. F. Ginhoux
Singapore Immunology Network, A*STAR
Singapore 138648 (Singapore)
[+] These authors contributed equally to this work.
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.201903058.
Prof. T. Kwon
Department of Biomedical Engineering,
Ulsan National Institute of Science and Technology
Ulsan 44919 (Korea)
Angew. Chem. Int. Ed. 2019, 58, 1 – 5
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
position of CDr20-stained live cells on Iba-1+-microglia
(97.3%) and GFAP+- (0.4%) or Aldh1l1+-astrocytes (0.0%
of live cells overlaid to the immunostained cells; Figure 1C and
Supporting Information, Figure S3). Interestingly, CDr20 did
not emit a fluorescence signal in an aqueous environment,
which enabled time-lapse imaging of microglia without remov-
ing CDr20 from the culture media (Supporting Information,
Figure S4A and Video S1).
Due to its high S/N ratio in fluorescent imaging, we
evaluated CDr20-based fluorescent-activated cell sorting
(FACS). Isolated cell populations based on CDr20 fluores-
cence expressed high levels of microglia gene markers, such as
P2ry12, Csf1r, and Cx3cr1, suggesting that single-probe
staining is sufficient to isolate microglia from heterogeneous
cell populations (Supporting Information, Figure S4B,C and
Data S1). Using acutely isolated mouse brain cells, we also
compared the microglia sorting efficiency of CDr20 to that of
other conventional methods (Figure 1D,E and Supporting
Information, Table S2). We used Csf1r–EGFP-positive,[11]
CD11b+/CD45low-gated,[12] and fluorescence-dye-conjugated-
IB4-stained[4] cells as experimental controls, whereas CDr10-
stained[8] and unsorted total live cells were used as the
comparison groups (Figure 1D). Gene expression analyses
revealed that IB4- and CDr10-stained sorted cells and total
live cells were more heterogeneous because IB4 and CDr10
exhibited affinity for endothelial cells and other glial cells
(Figure 1E and Supporting Information, Figure S5A,C). In
contrast, cells sorted on the basis of CDr20 fluorescent signal
exhibited similar transcriptome profiles as those of microglia
from Csf1r–EGFP mice and CD11b+/CD45low-gated cells
(Figure 1E and Supporting Information, Figure S5B,C), indi-
cating that CDr20-based FACS can be used to isolate
microglia populations.
Next, we tested whether CDr20 could label microglia
populations in brain tissues using live-cell, confocal imaging.
Compared to IB4-based microglia staining, CDr20 did not
stain the blood vessels in the brain slices (Figure 2A and
Supporting Information, Figure S6), and its intracellular
fluorescence signal was clear in identifying microglia (Fig-
ure 2B). A total of 94.1% of Csf1r–EGFP+ mouse microglia
in the brain slices of adult mice were labeled by CDr20
(Figure 2B and Supporting Information, Figure S7A). To test
its ability to label microglia in an intact brain, a whole, live
Csf1r–EGFP day 13 mouse embryo was incubated with
CDr20 (Figure 2C). CDr20 noninvasively labeled microglia
in the live embryo (Figures 2C); 99.2% of EGFP+ microglia
were labeled by CDr20 (Supporting Information, Fig-
ure S7B). To further test whether intravenous administration
of CDr20 could label microglia in vivo, we performed
a craniotomy above the primary somatosensory cortex in
mice (Supporting Information, Figure S8). Intravenous infu-
sion of CDr20 resulted in CDr20 fluorescence signals on the
EGFP-expressing microglia of neurological models, manni-
tol-injected wild-type (Figure 2E and Supporting Informa-
tion, Figure S9), and on the microglia of aged Alzheimerꢀs
disease mice (Supporting Information, Video S2), suggesting
that CDr20 can label microglia in vivo.
Figure 1. Identification of CDr20 as a microglia-specific fluorescent
probe. A) Left: Experimental setup. Right: Chemical structure of CDr20
with the structure–activity relationships. B) Densitometry of non-
stained or CDr20-stained Csf1r–EGFP brain cells. C) Superimposed
images of CDr20 live-cell labeling and the indicated immunolabeling in
primary cultured glial cells. Scale bars=100 mm. D) Experimental
scheme of the RNA-Seq analysis using the five different microglia
sorting methods. Total, acutely isolated brain cells were used. E) Spear-
man’s rank correlation coefficients of the gene expression values
between the independent replicates of (D). Gating details are in
Figure S16 in the Supporting Information. CD11b+/45low, CD11b-pos-
itive, and CD45-low. All experiments were using live cells, except for
the immunolabeling shown in (C).
signal-to-noise (S/N) ratios. Other compounds without a hy-
droxyl group did not show any specificity to microglia. More
hydroxyl groups may not be essential for labeling microglia.
For instance, compounds 11 and 12 exhibited better specificity
in labeling microglia, while compounds 13–15 and 29 did not.
As demonstrated by compounds 16–20, a methoxy group with
an additional hydroxyl group improved the specificity for
labeling microglia. Based on these findings, we selected CDr20
as the best microglia probe due to its stability (Figure 1A). The
specificity of CDr20 was further confirmed by the super-
To identify proteins that confer the functionality of CDr20
in microglia, we performed a genome-scale CRISPR-Cas9
2
www.angewandte.org
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2019, 58, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
Figure 2. Two-photon ex vivo and in vivo microscopy of live microglia
using CDr20. A) Left: Experimental setup. Right: Difference between
Isolectin B4 (IB4)–Alexa Fluor 488 (AF488) and CDr20 in labeling
blood vessels and microglia in an adult mouse brain slice (live).
B) CDr20-labeled microglia colocalized with Csf1r–EGFP+ (Top) or
IB4–AF488+ (Bottom) microglia (Arrowheads). C) Labeling of embry-
onic microglia by simple incubation of a whole live mouse embryo
(day 13) with CDr20. Top: A low-magnification epi-fluorescence image.
Bottom: Colocalization of CDr20-labeled cells with EGFP+ microglia.
D) Scheme of in vivo labeling. E) In vivo microglia imaging with CDr20
intravenous injections after systemic inflammation induced by intra-
peritoneal lipopolysaccharide injection. Left: 3D view, Right: Single-
channel images. Scale bars=50 mm in (A) and (C-bottom), 500 mm in
(C-top), and 10 mm in (B) and (E).
Figure 3. Identification of the turn-on fluorescence mechanism of
CDr20 by Ugt1a7c. A) Schematic steps for the identification of CDr20
target genes with genome-scale knockout (KO) microglia. B) Images of
CDr20-stained control and six-times enriched CDr20+ and CDr20À
populations. C) Dot plots of the read counts of the 78,637 sgRNAs in
the library (x-axes) and CDr20À populations (y-axes). Number: library
ID; 88071539 is enriched in outliers (left, 14000 k; right, 13200 k).
Repeatedly observed sgRNAs are indicated with their target(s). D) The
relative expression levels of the three candidate genes in CDr20+
populations from other glia. n.d.: Not detected. E) Inhibition of micro-
glial CDr20 staining by treatment of microglia with four Ugt inhibitors
at 0.1, 1, and 10 times their IC50 concentrations. Bold lines: mean
intensities. Dots: each data points (N=4–6). F) Generation of the
CDr20+ population by Ugt1a7c–FLAG overexpression in Neuro2a
(N2A) cells. Epi-fluorescence images. Arrows: Ugt1a7c–FLAG-express-
ing and CDr20-stained cells. G) LC-MS spectra of CDr20 (m/z 353.0)
and CDr20–Gluc converted by a Ugt1a7c-containing microsome (m/z
531.1). H) The proposed fluorescence turn-on process of CDr20. ESPT:
excited-state proton transfer. Scale bars=100 mm (B) and 20 mm (F);
**: p<0.01.
knockout screening by transducing SpCas9-expressing BV2
microglial cells with the Brie library (Figure 3A).[13] The
resulting single gene-knockout populations were subse-
quently labeled with CDr20, and cells with different CDr20-
derived fluorescence intensities were further enriched by
multiple rounds of FACS to separate CDr20-negative
(CDr20À) from CDr20-positive (CDr20+) cells (Figure 3A).
The appearance of CDr20À cell populations indicates that
a gene is responsible for generating the fluorescence signal for
CDr20 in microglia (Figure 3B and Supporting Information,
Figure S10). The abundant single guide RNAs (sgRNAs)
from the enriched CDr20À populations were then identified
by comparing the counts of each sgRNA with those from the
unselected library.[13] Three sgRNAs (88071539, 88095234,
and 21658821) targeting the Ugt1a9/Ugt1a7c, Ugt1a7c, and
Olfr1361 genes, respectively, were repeatedly accumulated in
CDr20À populations in two independent experiments (Fig-
ure 3C). Only Ugt1a7c had substantial expression levels
(greater than 1 fragment per kilobase of exon model per
million reads mapped, FPKM) and exhibited enhanced
expression in CDr20+ microglia among the candidate genes
(Figure 3D). Interestingly, in the Ugt gene family, Ugt1a7c
was the only member enriched in microglia (Supporting
Information, Table S3). Previous transcriptomic analyses
Angew. Chem. Int. Ed. 2019, 58, 1 – 5
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
have confirmed that the Ugt1a7c gene was expressed in
a microglial-specific manner (accession number: GSE43808
and GSE52564) in the brain.[14,15]
M.F.), and a Start-up Research Grant (SUTD-T1SRCI17126
to X.G.L.). F.G. is an EMBOYIP awardee and is supported by
Singapore Immunology Network (SIgN) core funding and
a Singapore National Research Foundation Senior Investiga-
torship (NRFI) NRF2016NRF-NRFI001-02. Flow cytometry
core is part of the SIgN Immunomonitoring platform
(supported by a BMRC IAF 311006 grant and BMRC
transition funds #H16/99/b0/011).
Ugt1a7c protein is a UDP-glucuronosyltransferase with
phenol-induced glucuronidase activity.[16] Since CDr20 con-
tains a phenolic moiety essential for microglia labeling
(Figure 1A), we tested whether the fluorogenicity of CDr20
in microglia was caused by the enzymatic activity of Ugt1a7c.
Microglia displayed decreased CDr20-derived fluorescence
upon either magnolol or silybin treatment, selective inhibitors
of the Ugt1a7c human isoform,[16,17] but not upon ritonavir or
diclofenac treatment, inhibitors of other Ugt1 family mem-
bers (Figure 3E).[16,18] More importantly, ectopic overexpres-
sion of Ugt1a7c in a nonmicroglial cell line, Neuro2a (N2A),
was sufficient to elicit CDr20 fluorescence (Figure 3F). Using
LC-MS/HRMS, we observed that the glucuronidated form of
CDr20 (CDr20–Gluc) in both CDr20 solutions, which had
been treated with microsomal Ugt1a7c (Figure 3G and
Supporting Information, Figure S11), and cell lysates from
CDr20-treated Ugt1a7c-overexpressing N2A cells (Support-
ing Information, Figure S12).
Lastly, we examined how the conversion from CDr20 to
CDr20–Gluc generated strong fluorescence signals in micro-
glia. Quantum chemical calculations showed that CDr20
emission can be quenched through excited-state proton
transfer (ESPT)[19] by losing one proton at the hydroxyl
group upon photoexcitation (Supporting Information, Fig-
ure S13). However, ESPT can be effectively suppressed after
CDr20 is converted to CDr20–Gluc (Figure 3H). Conse-
quently, the fluorescence signal was enhanced due to in vitro
conversion of CDr20 (quantum yield, F = 0.007) into CDr20–
Gluc (F = 0.611) by Ugt1a7c (Supporting Information, Fig-
ure S14). In summary, we have discovered CDr20, a fluoro-
genic probe that labels microglia in both cell cultures and live
rodent brains through its Ugt1a7c-mediated fluorescence
turn-on process. CDr20 could be a useful tool for identifying
microglia in neural disorders.
Conflict of interest
The authors declare no conflict of interest.
Keywords: biological activity · fluorescent probes ·
imaging agents · microglia · structure–activity relationships
[1] M. Prinz, J. Priller, Nat. Rev. Neurosci. 2014, 15, 300 – 312.
[2] M. W. Salter, B. Stevens, Nat. Med. 2017, 23, 1018 – 1027.
[3] A. Chow, B. D. Brown, M. Merad, Nat. Rev. Immunol. 2011, 11,
788 – 798.
[4] W. J. Streit, G. W. Kreutzberg, J. Neurocytol. 1987, 16, 249 – 260.
[5] F. Boscia, C. L. Esposito, A. Casamassa, V. De Franciscis, L.
Annunziato, L. Cerchia, J. Neurochem. 2013, 126, 428 – 436.
[6] T. Wꢁlchli, J. M. Mateos, O. Weinman, D. Babic, L. Regli, S. P.
Hoerstrup, H. Gerhardt, M. E. Schwab, J. Vogel, Nat. Protoc.
2015, 10, 53 – 74.
[7] G. J. Liu, R. J. Middleton, C. R. Hatty, W. W. Y. Kam, R. Chan,
T. Pham, M. Harrison-Brown, E. Dodson, K. Veale, R. B. Banati,
Brain Pathol. 2014, 24, 631 – 653.
[8] C. Leong, S. C. Lee, J. Ock, X. Li, P. See, S. J. Park, F. Ginhoux,
S.-W. Yun, Y.-T. Chang, Chem. Commun. 2014, 50, 1089 – 1091.
[9] J. Saura, J. Neuroinflammation 2007, 4, 26.
[10] J. S. Lee, N. Y. Kang, K. K. Yun, A. Samanta, S. Feng, K. K.
Hyeong, M. Vendrell, H. P. Jung, Y. T. Chang, J. Am. Chem. Soc.
2009, 131, 10077 – 10082.
[11] P. Wieghofer, K. P. Knobeloch, M. Prinz, Glia 2015, 63, 1 – 22.
[12] J. D. Sedgwick, S. Schwender, H. Imrich, R. Dçrries, G. W.
Butcher, V. ter Meulen, Proc. Natl. Acad. Sci. USA 1991, 88,
7438 – 7442.
[13] J. G. Doench, N. Fusi, M. Sullender, M. Hegde, E. W. Vaimberg,
K. F. Donovan, I. Smith, Z. Tothova, C. Wilen, R. Orchard,
H. W. Virgin, J. Listgarten, D. E. Root, Nat. Biotechnol. 2016, 34,
184 – 191.
[14] C. Beutner, B. Linnartz-Gerlach, S. V. Schmidt, M. Beyer, M. R.
Mallmann, A. Staratschek-Jox, J. L. Schultze, H. Neumann, Glia
2013, 61, 1429 – 1442.
Acknowledgements
We thank Prof. Hwan Myung Kim (Ajou University, South
Korea) and Prof. Zeng Li (NNI, Singapore) for providing
animals. We thank Drs. Yong-An Lee, Jong-Jin Kim, and
Nam-Young Kang for valuable discussions. We thank the
SBIC-Nikon Imaging Centre, Advanced Bioimaging Core of
DUKE-NUS Medical School, and Singapore Health Services.
This work was supported by intramural funding from
A*STAR (Agency for Science, Technology and Research,
Singapore) Biomedical Research Council and the Joint
[15] Y. Zhang, K. Chen, S. A. Sloan, M. L. Bennett, A. R. Scholze, S.
OꢀKeeffe, H. P. Phatnani, P. Guarnieri, C. Caneda, N. Ruderisch,
S. Deng, S. A. Liddelow, C. Zhang, R. Daneman, T. Maniatis,
B. A. Barres, J. Q. Wu, J. Neurosci. 2014, 34, 11929 – 11947.
[16] B. T. Gufford, G. Chen, P. Lazarus, T. N. Graf, N. H. Oberlies,
M. F. Paine, Drug Metab. Dispos. 2014, 42, 1675 – 1683.
[17] L. Zhu, G. Ge, H. Zhang, H. Liu, G. He, S. Liang, Y. Zhang, Z.
Fang, P. Dong, M. Finel, L. Yang, Drug Metab. Dispos. 2012, 40,
529 – 538.
Council
Office
Development
Program,
A*STAR
(1334k00083 to Y.T.C), Institute for Basic Science (IBS)
(IBS-R007-A1 to X.L. and Y.T.C.), the Singapore National
Research Foundation Competitive Research Program (NRF-
CRP17-2017-04 to H.S.J.), a National Medical Research
Council Open-Fund Individual Research Grant (NMRC/
[18] S. Algeelani, N. Alam, M. A. Hossain, G. Mikus, D. J. Green-
blatt, Xenobiotica 2017, 47, 1 – 6.
[19] L. M. Tolbert, K. M. Solntsev, Acc. Chem. Res. 2002, 35, 19 – 27.
OFIRG/0050/2017 to H.S.J.),
a Duke-NUS Signature
Research Program Block Grant (to H.S.J.), a Khoo Postdoc-
toral Fellowship Award (Duke-NUS-KPFA/2016/0007 to
Manuscript received: March 12, 2019
Version of record online: && &&, &&&&
4
www.angewandte.org
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2019, 58, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Fluorescent Probes
Lightbulb moment: The fluorogenic
probe, CDr20, specifically labels mouse
microglia both in vitro and in vivo, ena-
bling their visualization through a fluo-
rescence turn-on response, mediated by
the UDP-glucuronosyltransferase,
B. Kim, M. Fukuda, J.-Y. Lee, D. Su,
S. Sanu, A. Silvin, A. T. T. Khoo, T. Kwon,
X. Liu, W. Chi, X. Liu, S. Choi, D. S. Y. Wan,
S.-J. Park, J.-S. Kim, F. Ginhoux, H. S. Je,*
Y.-T. Chang*
&&&& — &&&&
Ugt1a7c. CDr20 could be useful for the
characterization of the role of microglia
during brain development and the
pathogenesis of neural disorders.
Visualizing Microglia with a Fluorescence
Turn-On Ugt1a7c Substrate
Angew. Chem. Int. Ed. 2019, 58, 1 – 5
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!