An azido-BODIPY probe for glycosylation: Initiation of strong fluorescence upon triazole formation

Article abstract of DOI:10.1021/ja5010174

We have designed a low fluorescent azido-BODIPY-based probe AzBOCEt (Az10) that undergoes copper(I)-catalyzed 1,3-dipolar cycloadditions with alkynes to yield strongly fluorescent triazole derivatives. The fluorescent quantum yield of a triazole product T

Full text of DOI:10.1021/ja5010174

Article
pubs.acs.org/JACS
An Azido-BODIPY Probe for Glycosylation: Initiation of Strong
Fluorescence upon Triazole Formation
Jiun-Jie Shie,† Ying-Chih Liu,† Yu-Ming Lee,‡ Carmay Lim,‡ Jim-Min Fang,†,§ and Chi-Huey Wong*,†
†The Genomics Research Center, Academia Sinica, Taipei 11529, Taiwan
‡Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan
§Department of Chemistry, National Taiwan University, Taipei 106, Taiwan
S
* Supporting Information
ABSTRACT: We have designed a low fluorescent azido-BODIPY-
based probe AzBOCEt (Az10) that undergoes copper(I)-catalyzed
1,3-dipolar cycloadditions with alkynes to yield strongly fluorescent
triazole derivatives. The fluorescent quantum yield of a triazole
product T10 is enhanced by 52-fold as compared to AzBOCEt
upon excitation at a wavelength above 500 nm. Quantum
mechanical calculations indicate that the increase in fluorescence
upon triazole formation is due to the lowering of the HOMO
energy level of the aryl moiety to reduce the process of acceptor
photoinduced electron transfer. AzBOCEt is shown to label alkyne-
functionalized proteins in vitro and glycoproteins in cells with excellent selectivity, and enables cell imaging and visualization of
glycoconjugates in alkynyl-saccharide-treated cells at extremely low concentration (0.1 μM). Furthermore, the alkyne-tagged
glycoproteins from cell lysates can be directly detected with AzBOCEt in gel electrophoresis.
glycans on the cell surface. The subsequent labeling with proper
azido-containing probes via CuAAC gives the fluorescent
triazole-tagged glycans that are easily visualized and profiled on
a proteomic scale. We have also designed the fluorogenic 1,8-
naphthalimide-based probes6a and trifunctional coumarin-based
probes6d for in vitro and in vivo labeling of glycans. Several
fluorescence-forming probes upon CuAAC reactions have also
been utilized by other groups to label specific biomolecules.7
The distinct fluorescence enhancement induced by efficient
triazole formations would have broad applications in the
emerging field of cell biology and functional proteomics.
However, these azido- and alkynyl-functionalized probes
usually require excitation in the UV region and emit blue
light with poor quantum yield in aqueous solution; such optical
properties are not ideal for biological applications. Therefore,
there is a need to develop azido- or alkynyl-fluorogenic probes
with excitation and emission wavelengths in the visible region.
In addition to a fluorescein-based probe,8 Wang and co-workers
have designed an azido-containing BODIPY-based probe.9 The
azido group is located at the 3-position of the BODIPY core, so
that the electron-donating effect of the α-nitrogen of the azido
group can quench fluorescence via an internal charge transfer
(ICT) process. Upon formation of the triazole derivative
through a CuAAC reaction with alkyne, fluorescence is
switched on due to the decreased electron density at the 3-
position of the BODIPY core. However, the azido-BODIPY
INTRODUCTION
■
Glycosylation is an important co/post-translational modifica-
tion that has been shown to be crucial for the structure and
function of many proteins. Aberrant glycosylation on the
surface of malignant cells is often observed in pathological
conditions, such as inflammation and cancer metastasis.1,2 In
particular, altered terminal sialylation and fucosylation, which
presumably result from changes of expression locations and
levels of sialyltransferases and fucosyltransferases, are associated
with tumor malignancy.1,3 Exploring the biological content of
glycans attached to proteins or lipids as cancer biomarkers has
become a major subject of research. Utilizing bioorthogonal
chemical reporter is a successful strategy to study the role of
glycans in numerous physiological and pathological processes as
well as to visualize the process of glycosylation in vivo. In this
approach, metabolic oligosaccharide engineering (MOE) is
employed to modify glycans with an unnatural monosaccharide
precursor containing a unique bioorthogonal chemical reporter
group as substrate for the biosynthetic pathway in a living
system (cell or whole organism) for incorporation into a target
glycan, which then react with a complementary bioorthogonal
functional group covalently linked to a set of probes, including
biotin and fluorescence tags.4
Research for specific glycan tagging in our laboratory has
benefited greatly from Cu(I)-catalyzed azide−alkyne cyclo-
addition (CuAAC)5 optimized for biological conjugation. We
have previously demonstrated that peracetylated alkynyl
derivatives of monosaccharides, for example, Fucyne,6a,b
ManNAcyne,6b−d and alkyne-hinged 3-fluorosialyl fluoride
(DFSA),6e are incorporated into sialylated and fucosylated
Received: January 29, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Fluorogenic CuAAC reaction of azido-BODIPY with 4-pentyn-1-ol. Electron flow is involved in the a-PeT process from the aryl moiety to
the excited BODIPY as shown in the molecular orbital diagram.
Figure 2. Structures of azido-substituted BODIPY derivatives Az1−Az8 used in the fluorescence screening through the CuAAC reactions.
compound is unstable and fails to react with alkynyl
biomolecules under cellular conditions.
PAGE by using AzBOCEt labeling for direct detection of the
probe-labeled glycoproteins.
Intramolecular photoinduced electron transfer (PeT) is a
well-known mechanism through which the fluorescence of a
fluorophore is quenched by transfer of electron between
nonplanar parts of a fluorescent molecule.10 Nagano and co-
workers have demonstrated a PeT-based fluorescence ON−
OFF switching mechanism is applicable to fluorescein
derivatives11 as well as to BODIPY molecules.12 The first
attempt toward this concept upon triazole formation was
recently reported by Bertozzi’s group,8 who designed
fluorogenic azidofluorescein derivatives by addition of the
corresponding aryllithium compound to the xanthone deriva-
tive. Actually, BODIPY fluorophores are more photostable than
fluorescein13 and have relative insensitivity in the physiological
pH range.14 We believe that the development of BODIPY-
based fluorogenic probe upon triazole formation by PeT
mechanism is valuable in this stage. We present herein a
rational design and experimental validation of new azido-
functionalized fluorogenic probes, for example, AzBOCEt
(Az10), based on the green-emitting BODIPY scaffold via an
intramolecular PeT pertinent to cell environments. Our present
study indicates that AzBOCEt is a selective fluorogenic labeling
reagent for alkyne-functionalized proteins, and is suitable for
visualizing the localization of alkyne-tagged glycosyl conjugates
in cells by confocal microscopy. Furthermore, the alkynyl-
saccharide modified cells can be lyzed and analyzed on SDS-
RESULTS AND DISCUSSION
■
4,4-Difluoro-4-bora-3a,4a-diaza-s-indacene (better known as
BODIPY)14 dyes are a type of popular fluorophores for many
biological applications. BODIPY dyes have numerous advan-
tages including great chemical and photophysical stability,
relatively high molar absorption coefficients and fluorescence
quantum yields (Φfl), excitation/emission wavelengths in the
visible spectral region (longer than 500 nm), and narrow
emission bandwidths with high peak intensities. We chose the
BODIPY scaffold as a starting module for its appealing
synthetic and fluorescent features. BODIPYs are easily modified
at the 8-position. Arylation at this position has no substantial
influence on absorption and emission wavelengths because the
aryl moiety and BODIPY core are twisted and conjugation
uncoupled. Nagano and co-workers have demonstrated that a
pendant aryl moiety can quench the fluorescence of BODIPY
via an acceptor-excited PeT (a-PeT) process.12 This quenching
mechanism can be modulated by using various substituents to
alter the electron density of the aryl ring. We thus developed a
series of azido-aryl BODIPY fluorescence-forming probes.
Upon CuAAC reaction, formation of the triazole derivative
would relieve the a-PeT12a by lowering the HOMO energy level
of the aryl moiety, relative to that of the parental azido-
substituted aryl substrate, to turn on the fluorescence (Figure
1).
B
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
As the first step, we investigated the structures of various
BODIPYs Az1−Az8 (Figure 2) by introducing an azido group
and other substituents into the pendant benzene moieties of
BODIPYs to modulate the electron density of the aryl rings.
The acid-catalyzed condensation of 2,4-dimethyl-3-ethylpyrrole
with substituted nitrobenzaldehydes, followed by oxidation with
DDQ in mild conditions, gave dipyrromethene intermediates,
which were treated with BF3·OEt2 to yield the corresponding
nitro-BODIPYs 1−8 (Scheme S1 in the Supporting Informa-
tion). According to the previously reported method,15 the
amino-BODIPYs Am1−Am8 were obtained in reasonable
yields by reduction of the nitro-BODIPYs with hydrazine in
the presence of 10% Pd/C (Scheme S2 in the Supporting
Information). On treatment with triflyl azide (TfN3) in mild
conditions, the amino-BODIPYs were converted to the target
azido-BODIPYs Az1−Az8.16
To test the fluorogenic CuAAC reaction of azido-BODIPYs,
Az1−Az8 were reacted with 4-pentyn-1-ol, as a model alkyne,
to give the corresponding 1,3-dipolar cycloaddition products
T1−T8, in a microtiter plate containing CuSO4, sodium
ascorbate, and a tris-triazole ligand7c prepared from tripropar-
gylamine and ethyl azidoacetate. The highly efficient CuAAC
reactions allowed direct fluorescence measurements (λex = 488
nm) using a plate reader. The formation of fluorescent or
nonfluorescent triazole compounds could be easily monitored
upon irradiation at 365 nm with a UV lamp (Figure S1 in the
Supporting Information). In the preliminary fluorescence
screening experiments, the parental azido-BODIPY Az1 (R =
H) has modest fluorescence, and no obvious change of
fluorescence intensity was observed upon the CuAAC reaction.
The azido-BODIPYs Az6−Az8 and their CuAAC products
T6−T8 showed almost no fluorescence, presumably because
the excited BODIPY core was quenched by a-PeT process due
to the strong electron-donating effect of the amino groups in
the aryl moieties. Compounds Az3−Az5 and T3−T5
containing two methoxy substituents still showed weak
fluorescence, albeit the fluorescence intensity was slightly
enhanced upon triazole formation. In contrast, an appreciable
fluorescence enhancement was observed upon the CuAAC
reaction of azido-BODIPY Az2 (Φfl = 0.28) containing a
methoxy substituent to form the triazole T2 (Φfl = 0.34). This
result implied that the HOMO energy level of BODIPY core
might be too high to accept electrons donated from 4-
azidoanisole or 4-triazolylanisole moieties, thus inhibiting the
fluorescence quenching in the a-PeT process. Our calculations
(Table S1 in the Supporting Information) supported that 4-
azidoanisole (−5.75 eV) and 4-triazolylanisole (−6.13 eV) have
low-lying HOMO energy levels as compared to that of 2,6-
diethyl-1,3,5,7-tetramethyl BODIPY (−5.25 eV).
Figure 3. Structures of amino-BODIPY Am10, azido-BODIPYs Az9−
Az11, and the corresponding triazolyl-BODIPYs T9−T11 obtained by
CuAAC reactions with 4-pentyn-1-ol.
the Supporting Information). Using 4-pentyn-1-ol as a model
alkyne for CuAAC reactions (Scheme S5 in the Supporting
Information), the triazolyl-BODIPY products T9−T11 showed
significant increases of fluorescence quantum yields by 3−52-
fold as compared to the parental azido-BODIPYs (Figure 4 and
Table 1). For example, azido-BODIPY Az10 in ethanol
solution produced a weak emission band centered at 517 nm
with a low quantum yield (Φfl = 0.011, λex = 507 nm), whereas
triazolyl-BODIPY T10 exhibited a strong fluorescence at 515
nm with a 52-fold enhancement in quantum yield (Φfl = 0.572,
λex = 501 nm).
To support the a-PeT-controlled fluorescence switching
upon CuAAC reaction, we carried out density functional theory
calculations for the HOMO energy levels of the 4-azidoanisole,
4-triazolylanisole, and the BODIPY cores (Table S1 in the
Supporting Information). The orbital energies were calculated
using the Gaussian 09 program17 at the B3-LYP/6-31G level of
theory.12a,18−20 The calculations indicated that conversion of 4-
azidoanisole to the corresponding triazole results in a significant
decrease in the HOMO energy level of the aryl moiety (from
−5.75 to −6.13 eV); the changes of the HOMO energy levels
of BODIPY cores due to the substituents at the C-2 and C-6
positions were calculated as −5.25, −5.43, −5.88, and −6.45 for
ethyl, hydrogen, ethoxycarbonyl, and cyano groups, respec-
tively. In the case of ethoxycarbonyl-substituted BODIPY, the
HOMO energy level of BODIPY core was located between that
of 4-azidoanisole and 4-triazolylanisole (Figure 5). The
fluorescence could be restored by the triazole formation due
to the lowering of the HOMO energy level of the aryl moiety to
reduce the a-PeT process.
We have clearly demonstrated that the fluorescence of
AzBOCEt (Az10) could be efficiently switched on through the
CuAAC reaction. For biological applications, Az10 might be
reduced by protein or endogenous thiols to generate a side
product amino-BODIPY Am10, which showed almost no
fluorescence (Φfl = 0.007). Even if this side reaction were to
occur, it would not contribute to the background fluorescence
(Table 1 and Figure 4). Thus, Az10 would act as a pertinent
fluorogenic probe for labeling various alkyne-containing
biomolecules. For example, bovine serum albumin (BSA) was
modified with an alkyne moiety by reaction of its lysine residues
with 4-pentynoyl N-hydroxysuccinimide (NHS) ester. The
alkyne-functionalized BSA and unmodified BSA were incubated
with Az10, CuSO4, tris-triazole ligand,7c and sodium ascorbate
in a cosolvent system of PBS buffer (pH 7.4) and DMSO (9:1)
for 1 h. The samples were analyzed by SDS-PAGE, followed by
fluorescence imaging and staining with Coomassie blue for
protein detection (Figure 6). The fluorescence scanning
showed that only the alkyne-functionalized BSA was selectively
For practical uses, an azido-BODIPY fluorogenic probe is
better in the dark state and only becomes bright after CuAAC
reaction. Accordingly, the BODIPY core should have a HOMO
energy level higher than 4-triazolylanisole for fluorescence
emission, but lower than 4-azidoanisole for PeT quenching.
With these considerations, we thus designed three azido-
BODIPY derivatives Az9−Az11 (Figure 3), which still have 4-
azidoanisole as the aryl moiety, but have the ethyl groups at the
C-2 and C-6 positions of the BODIPY core replaced by
hydrogen atoms or the electron-deficient ethoxycarbonyl and
cyano groups to reduce the HOMO energy level of the
fluorophore. The synthesis of Az9−Az11 was smoothly carried
out via intermediacy of amino-BODIPYs Am9−Am11 by a
procedure similar to that for Az1−Az8 (Schemes S3 and S4 in
C
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. Absorption and normalized emission spectra of triazolyl-BODIPY T10, azido-BODIPY Az10, and amino-BODIPY Am10 in ethanol
solution (12 μM) at 25 °C. Inset: Images of T10, Az10, and Am10 in ethanol solution (120 μM). The change of yellow Az10 solution to green T10
solution was apparent.
Table 1. Fluorescence Properties of Amino-BODIPY Am10, Azido-BODIPYs Az2, and Az9−Az11, As Well As Triazolyl-
BODIPYs T2 and T9−T11 in Ethanol at 25 °C
a
compound
λex (nm)
ε (M−1 cm−1
)
λem (nm)
Φfl
folds of Φfl enhancement
b
b
c
Az2
T2
522
525
504
507
507
501
497
502
503
31 800
35 700
27 300
36 200
35 700
48 400
15 100
40 800
48 400
541
545
513
517
517
515
512
516
519
0.28 0.01
0.34 0.02
0.17 0.01
0.55 0.02
1.2×
3.2×
Az9
T9
c
c
c
c
c
c
Az10
T10
Am10
Az11
T11
0.011 0.002
0.572 0.002
0.007 0.001
0.014 0.002
0.043 0.002
52×
0.6×
3.1×
b
a
Calculation based on dividing the Φfl of triazole-BODIPY (or amino-BODIPY) by the Φfl of the corresponding azido-BODIPY. Quantum yields
were determined using Rhodamine 6G (Φfl = 0.95) in ethanol as a standard. Quantum yields were determined using fluorescein (Φfl = 0.85) in 0.1
c
M NaOH as a standard.
N-acetylglucosamine (GlcNAc)-modified proteins via the N-
acetylgalactosamine salvage pathway.22 As a negative control,
CL1−5 cells were grown in the presence of peracetylated N-
acetylmannosamine (Ac4ManNAc) or peracetylated N-galactos-
amine (Ac4GalNAc). After 3 days, the cells were fixed with
paraformaldehyde, followed by permeation with 0.2% Triton X-
100 and incubation with Az10 for CuAAC reaction. The
Ac4ManNAl- and Ac4GalNAl-treated cells showed intense
fluorescence labeling (Figure 7B), whereas the control cells
exhibited almost no green fluorescence staining. Although slight
green autofluorescence signals are indeed observed in the
control cells (Figure S2 in the Supporting Information), the
potential background staining we attribute to unidentified
aberrant reactivity of the probe in cells.
Interestingly, the fluorescence-labeled glycosyl conjugates
generated from Ac4ManNAl and Ac4GalNAl occurred with
Figure 5. Frontier orbital energy diagram. Illustration of the
different distributions in cells. To visualize the location of the
triazole-tagged glycosyl conjugates, the cells were stained with
anti-GRASP65 followed by Cy3 conjugated antirabbit for
marking Golgi, and Hoechst for marking nucleus. Figure 7C
shows the imaging results by confocal fluorescence microscopy
obtained from cells grown in the presence of different peracetyl
glycoses. When CL1−5 cells were incubated with Ac4ManNAl,
the expressed sialylated glycosyl conjugates were clearly
visualized in the cytosol using AzBOCEt probe (green
fluorescence), and significantly overlapped with the Golgi
apparatus (red staining), but not in the nucleus (blue staining).
In Ac4GalNAl-treated cells, the green fluorescence signals were
also observed mainly in the cytosol and partially in the Golgi
thermodynamic simulation of the fluorescence OFF/ON switch by
the a-PeT process.
labeled by Az10 in a concentration-dependent manner (4−100
μM).
Next, we sought to evaluate the performance of Az10 in cell
imaging. For this purpose, highly invasive lung cancer cells,
CL1−5, were cultured in the presence of peracetylated alkynyl-
N-acetylmannosamine (Ac4ManNAl) or peracetylated alkynyl-
N-acetylgalactosamine (Ac4GalNAl). The modified saccharide
Ac4ManNAl will metabolically lead to the alkyne-tagged sialic
acid expressed on cells (Figure 7A), whereas Ac4GalNAl will be
incorporated into mucin-type O-linked glycans21 and O-linked
D
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 6. Alkyne-functionalized BSA labeling with AzBOCEt (Az10). The gel was analyzed by fluorescence imaging (λex = 488 nm; λem = 526 nm).
The total protein content was revealed by Coomassie blue stain.
apparatus. Thus, AzBOCEt (Az10) was successfully used as a
fluorogenic probe to image the locations of two different
biosynthetically incorporated sugars in cells. The punctate
staining in the Ac4ManNAl treated cells may be related to
recycling of the surface SiaNAl-terminated glycoproteins to
cytosol and Golgi complex during incubation with Az10 dye.23
GalNAc is a core residue often found in O-glycans of proteins.
Incorporation of unnatural GalNAl to glycoproteins may affect
further glycosylations due to the specificity and efficiency of
glycosyltransferase. Thus, the immature aberrant glycoproteins
may be transported to cytosol for degradation. Incubation of
peracetylated alkynyl-N-acetylglucosamine (Ac4GlcNAl) in
CL1−5 cells only showed weak fluorescence after staining
with Az10 (Figure S3 in the Supporting Information), although
GlcNAc is abundant within the N- and O-linked glycans
produced in the endoplasmic reticulum and Golgi. This
phenomenon has been related to a metabolic bottleneck,22 so
that treating cells with Ac4GlcNAl could not give an optimal
fluorescence labeling of the GlcNAc-modified proteins.
We next examined whether AzBOCEt of CuAAC reaction
could be utilized for cell imaging to track the movement of the
sialylated glycoconjugates. To this end, CL1−5 cells were
incubated with 500 μM Ac4ManNAl for 1 h, and excess
Ac4ManNAl was subsequently removed. We then performed
the imaging to monitor the trafficking of the sialylated
glycoconjugates. As shown in Figure S4 in the Supporting
Information, the sialylated glycoconjugates were readily imaged
by using AzBOCEt and significantly overlapped with the Golgi
apparatus in initial stage. However, after 7 h, we observed the
appearance of a green fluorescence signal in the cell membrane
in cells. This signal’s intensity increased over time and then
reached saturation for 14 h. These findings show that the use of
AzBOCEt upon CuAAC provides a chemical tool for evaluating
the movement of the alkyne-bearing glycoconjugates in cells.
In comparison, the CuAAC nonactivatable BODIPY (Az2)
and another organic dye azido-TAMRA were used to label the
Ac4ManNAl-treated cells. As shown in Figure S5 in the
Supporting Information, both of the fluorescence signals on
Az2 and azido-TAMRA staining could not clearly define the
location of the alkyne-bearing glycoconjugates within cell. The
obvious background staining was observed in the control
experiment with Ac4ManNAc on Az2 labeling, thereby
obscuring the intracellular staining of alkynyl sugars. On the
other hand, although azido-TAMRA can also label the cells
treated with alkyne-containing sugar, the relatively weak
fluorescence output we attribute to low reactivity of the
probe under the same CuAAC conditions. Thus, Az10 is
superior to Az2 and azido-TAMRA in probing alkyne-
containing biomolecules by confocal microscopy.
We further examined whether AzBOCEt (Az10) could be
used for direct in-gel fluorescence detection, without enrich-
ment process, and for proteomic analyses of the glycosylated
proteins from cell lysates. The cell extracts (20 μg) harvested
from CL1−5 cells by treatment with 100 μM of Ac4ManNAc or
Ac4ManNAl for 3 days were incubated with the indicated
concentration of Az10 using CuAAC chemistry. The labeled
cell extracts were further separated by gel electrophoresis, and
visualized by in-gel fluorescence imaging. As shown in Figure 8,
fluorescence signals were only detected in the protein lysates
obtained from the cells treated with Ac4ManNAl. By contrast,
the Ac4ManNAl-treated cell lysates with Az2 staining showed
fluorescent bands similar to those with Az10 with comparable
background in SDS-PAGE, but obscure fluorescent signals
when labeled with azido-TAMRA (Figure S6 in the Supporting
Information). Interesting, Az10 staining revealed some back-
ground staining that we attribute to nonspecific binding with
protein lysates as it is also visible in the control experiments
without addition of Ac4ManNAl. The reason for the
fluorescence enhancement upon nonspecific binding to
proteins is presumably a decrease of PeT pathway due to the
hydrophobic environment.12c
To characterize the nature of the glycoproteins derived from
Ac4ManNAl treatment, the cell lysates were treated with
sialidase, and then incubated with Az10. As expected, sialidase
E
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 7. continued
F
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 7. Cell fluorescence labeling with AzBOCEt (Az10) and imaging by confocal microscopy. (A) Illustration of the cell labeling experiments
using Ac4ManNAl, Ac4GalNAl, and Az10. CL1−5 cells were incubated with 100 μM Ac4ManNAl, Ac4GalNAl, or control sugars (Ac4ManNAc and
Ac4GalNAc) for 3 days, and then treated with 0.1 μM Az10 for 1 h under CuAAC conditions. (B) Fluorescence, bright field, and overlaid images.
Scale bar: 75 μm. (C) Localization of the expressed glycosyl conjugates in CL1−5 cells. These glycosyl conjugates were labeled with fluorogenic
probe Az10 (green), anti-GRASP65 (a Golgi marker) followed by Cy3 conjugated antirabbit (red), and Hoechst (blue, a nucleus marker). Scale bar:
10 μm.
Figure 8. Direct in-gel fluorescence detection of alkyne-tagged glycoproteins using CuAAC with AzBOCEt (Az10) from cell lysates. The gel was
analyzed by fluorescence imaging (λex = 488 nm; λem = 526 nm) and Coomassie blue staining to reveal the total protein content.
treatment reduced the fluorescence in the fluorescent bands
due to cleavage of sialic acid residues by sialidase (Figure S7 in
the Supporting Information). Accordingly, 12 selected
fluorescent-labeled bands were excised from gel by proteolytic
digestion, and then subjected to MS (LTQ-FT) analysis. The
data acquisition and database searching methodology employed
G
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
are detailed in the Supporting Information. We thus identified
27 putative sialic acid-modified glycoproteins as listed in Table
S2 of the Supporting Information. ERLIN1 is one of the
putative sialic acid-modified glycoproteins, which was pre-
viously identified as an N-linked glycoprotein.24 To validate
whether ERLIN1 can be recognized by sialic acid-specific
lectins (Sambucus nigra (SNA) recognizes α2,6-linked sialic
acids, while Maackia amurensis lectin II (MALII) binds α2,3-
linked sialic acid), the cell lysates were incubated with
biotinylated lectins (SNA and MALII). Associated proteins
were precipitated with NeutrAvidin agarose beads. As shown in
Figure S8 in the Supporting Information, ERLIN1 was detected
in SNA and MALII affinity precipitates by immunoblot analysis
using anti-ERLIN1. It could be confirmed that ERLIN1 is a
sialic acid-modified glycoprotein by our methodology. Our
study demonstrated the feasibility of fluorescence labeling using
AzBOCEt to identify glycoproteins.
AUTHOR INFORMATION
Corresponding Author
chwong@gate.sinica.edu.tw
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Academia Sinica and the National Science Council
for financial support.
REFERENCES
■
(1) Dube, D. H.; Bertozzi, C. R. Nat. Rev. Drug Discovery 2005, 4,
477−488.
(2) Ohtsubo, K.; Marth, J. D. Cell 2006, 126, 855−867.
(3) Fuster, M. M.; Esko, J. D. Nat. Rev. Cancer 2005, 5, 526−542.
(4) (a) Prescher, J. A.; Bertozzi, C. R. Cell 2006, 126, 851−854.
(b) Laughlin, S. T.; Baskin, J. M.; Amacher, S. L.; Bertozzi, C. R.
Science 2008, 320, 664−667. (c) Sletten, E. M.; Bertozzi, C. R. Angew.
Chem., Int. Ed. 2009, 48, 6974−6998. (d) Friscourt, F.; Ledin, P. A.;
Mbua, N. E.; Flanagan-Steet, H. R.; Wolfert, M. A.; Steet, R.; Boons,
G.-J. J. Am. Chem. Soc. 2012, 134, 5381−5389. (e) Niederwieser, A.;
CONCLUSIONS
■
We have designed a new CuAAC-activatable probe AzBOCEt
(Az10), which showed little fluorescence (Φfl = 0.011).
However, upon triazole formation with 4-pentyn-1-ol, a strong
fluorescence is induced with excellent enhancement of
fluorescence quantum yield (52-fold increase) at a long
excitation wavelength above 500 nm. Quantum mechanical
calculations indicate that conversion of 4-azidoanisole to the
corresponding triazoles results in a significant decrease of the
HOMO energy level of the triazole-substituted aryl moiety,
supporting that the fluorescence enhancement upon CuAAC
reaction is due to inhibition of an a-PeT process. As compared
to previously reported fluorogenic azidofluoresceins, AzBOCEt
appears to have a lower fluorescence intensity of unreacted
reagent and a higher signal-to-noise ratio upon triazole
formation, indicating that use of AzBOCEt can more efficiently
inhibit the background noise for labeling of biomolecules and
materials. AzBOCEt probe is utilized in selective labeling of
alkyne-tagged proteins in vitro, and incorporation of alkyne-
tagged glycosyl conjugates in cultured cells can also be
visualized with an extremely low concentration (0.1 μM) of
AzBOCEt. Furthermore, we have shown the use of AzBOCEt
for direct in-gel detection of the alkyne-tagged glycoproteins
from cell lysates after SDS-PAGE. We believe that AzBOCEt is
useful for imaging alkynyl-modified biomolecules for better
understanding of biological processes. In addition, AzBOCEt is
a methoxy- and ester-functionalized azido-BODIPY, which can
be easily converted by hydrolysis to the corresponding
carboxylate probe for improvement of water solubility. On
the other hand, biotin-conjugated azido-BODIPY can be
synthesized by incorporating a biotin derivative in the benzene
moiety. We believe that the biotinylated BODIPY probe should
be a powerful tool for use in profiling and comparing low-
abundant glycoconjugates at different stages for identification of
glycan-related biomarkers.
Spate, A.-K.; Nguyen, L. D.; Jungst, C.; Reutter, W.; Wittmann, V.
̈
̈
Angew. Chem., Int. Ed. 2013, 52, 4265−4268.
(5) (a) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596−2599. (b) Tornøe, C. W.;
Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057−3064.
(6) (a) Sawa, M.; Hsu, T.-L.; Itoh, T.; Sugiyama, M.; Hanson, S. R.;
Vogt, P. K.; Wong, C.-H. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
12371−12376. (b) Hsu, T.-L.; Hanson, S. R.; Kishikawa, K.; Wang, S.-
K.; Sawa, M.; Wong, C.-H. Proc. Natl. Acad. Sci. U.S.A. 2007, 104,
2614−2619. (c) Hanson, S. R.; Hsu, T.-L.; Weerapans, E.; Kishikawa,
K.; Simon, G. M.; Cravatt, B. F.; Wong, C.-H. J. Am. Chem. Soc. 2007,
129, 7266−7267. (d) Tsai, C.-S.; Liu, P.-Y.; Yen, H.-Y.; Hsu, T.-L.;
Wong, C.-H. Chem. Commun. 2010, 46, 5575−5577. (e) Tsai, C.-S.;
Yen, H.-Y.; Lin, M.-I.; Tsai, T.-I.; Wang, S.-Y.; Huang, W.-I.; Hsu, T.-
L.; Cheng, Y.-S. E.; Fang, J.-M.; Wong, C.-H. Proc. Natl. Acad. Sci.
U.S.A. 2013, 110, 2466−2471.
(7) (a) Le Droumaguet, C.; Wang, C.; Wang, Q. Chem. Soc. Rev.
2010, 39, 1233−1239. (b) Sivakumar, K.; Xie, F.; Cash, B. M.; Long,
S.; Barnhill, H. N.; Wang, Q. Org. Lett. 2004, 6, 4603−4606. (c) Zhou,
Z.; Fahrni, C. J. J. Am. Chem. Soc. 2004, 126, 8862−8863. (d) Xie, F.;
Sivakumar, K.; Zeng, Q. B.; Bruckman, M. A.; Hodges, B.; Wang, Q.
Tetrahedron 2008, 64, 2906−2914. (e) Qi, J.; Han, M.-S.; Chang, Y.-
C.; Tung, C.-H. Bioconjugate Chem. 2011, 22, 1758−1762. (f) Herner,
́ ́
A.; Nikic, I.; Kallay, M.; Lemke, E. A.; Kele, P. Org. Biomol. Chem.
2013, 11, 3297−3306.
(8) Shieh, P.; Hangauer, M. J.; Bertozzi, C. R. J. Am. Chem. Soc. 2012,
134, 17428−17431.
(9) Wang, C.; Xie, F.; Suthiwangcharoen, N.; Sun, J.; Wang, Q. Sci.
China: Chem. 2012, 55, 125−130.
(10) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. Rev. 1997,
97, 1515−1566.
(11) (a) Tanaka, K.; Miura, T.; Umezawa, N.; Urano, Y.; Kikuchi, K.;
Higuchi, T.; Nagano, T. J. Am. Chem. Soc. 2001, 123, 2530−2536.
(b) Miura, T.; Urano, Y.; Tanaka, K.; Nagano, T.; Ohkubo, K.;
Fukuzumi, S. J. Am. Chem. Soc. 2003, 125, 8666−8671. (c) Ueno, T.;
Urano, Y.; Setsukinai, K.-i.; Takakusa, H.; Kojima, H.; Kikuchi, K.;
Ohkubo, K.; Fukuzumi, S.; Nagano, T. J. Am. Chem. Soc. 2004, 126,
14079−14085. (d) Urano, Y.; Kamiya, M.; Kanda, K.; Ueno, T.;
Hirose, K.; Nagano, T. J. Am. Chem. Soc. 2005, 127, 4888−4894.
(12) (a) Gabe, Y.; Urano, Y.; Kikuchi, K.; Kojima, H.; Nagano, T. J.
Am. Chem. Soc. 2004, 126, 3357−3367. (b) Ueno, Y.; Urano, Y.;
Kojima, H.; Nagano, T. J. Am. Chem. Soc. 2006, 128, 10640−10641.
(c) Sunahara, H.; Urano, Y.; Kojima, H.; Nagano, T. J. Am. Chem. Soc.
2007, 129, 5597−5604. (d) Matsumoto, T.; Urano, Y.; Shoda, T.;
Kojima, H.; Nagano, T. Org. Lett. 2007, 17, 3375−3377.
ASSOCIATED CONTENT
■
S
* Supporting Information
Supplemental figures and tables, synthetic and experimental
procedures, materials, and NMR spectra of synthesized
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
H
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(13) Hinkeldey, B.; Schmitt, A.; Jung, G. ChemPhysChem 2008, 9,
2019−2027.
(14) (a) Loudet, A.; Burgess, K. Chem. Rev. 2007, 107, 4891−4932.
(b) Ulrich, G.; Ziessel, R.; Harriman, A. Angew. Chem., Int. Ed. 2008,
47, 1184−1201. (c) Boens, N.; Leen, V.; Dehaen, W. Chem. Soc. Rev.
2012, 41, 1130−1172. (d) Kamkaew, A.; Lim, S. H.; Lee, H. B.; Kiew,
L. V.; Chung, L. Y.; Burgess, K. Chem. Soc. Rev. 2013, 42, 77−88.
(15) Li, L.; Han, J.; Nguyen, B.; Burgess, K. J. Org. Chem. 2008, 73,
1963−1970.
(16) Liu, Q.; Tor, Y. Org. Lett. 2003, 5, 2571−2572.
(17) Frisch, M. J.; et al. Gaussian 09, revision A.02; Gaussian, Inc.:
Wallingford, CT, 2009. See the Supporting Information for the full
reference.
(18) Kennedy, D. P.; Kormos, C. M.; Burdette, S. C. J. Am. Chem.
Soc. 2009, 131, 8578−8586.
(19) Krumova, K.; Cosa, G. J. Am. Chem. Soc. 2010, 132, 17560−
17569.
(20) Wang, T.; Douglass, E. F., Jr.; Fitzgerald, K. J.; Spiegel, D. A. J.
Am. Chem. Soc. 2013, 135, 12429−12433.
(21) Dube, D. H.; Prescher, J. A.; Quang, C. N.; Bertozzi, C. R. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 4819−4824.
(22) Boyce, M.; Carrico, I. S.; Ganguli, A. S.; Yu, S.-H.; Hangauer, M.
J.; Hubbard, S. C.; Kohler, J. J.; Bertozzi, C. R. Proc. Natl. Acad. Sci.
U.S.A. 2011, 108, 3141−3146.
(23) Huang, K. M.; Snider, M. D. J. Biol. Chem. 1993, 268, 9302−
9310.
(24) Chen, R.; Jiang, X.; Sun, D.; Han, G.; Wang, F.; Ye, M.; Wang,
L.; Zou, H. J. Proteome Res. 2009, 8, 651−661.
I
dx.doi.org/10.1021/ja5010174 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX