Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
10.1002/anie.201904700
Angewandte Chemie International Edition
The image part with relationship ID rId2 was not found in the
file.
COMMUNICATION
acids was not observed because water is a competitive nucleophile
that is present in large excess (Figure S14). Therefore, the ketene
intermediate of compound 2 (Figure 3A) can only be trapped by
nucleophiles that are present in close proximity immediately after
photoactivation. We hypothesized that if the photochemical reaction
occurs immediately after enzymatic conversion, the nitroreductase
itself might be able to trap the ketene. To test this hypothesis, NTR
from E. coli was incubated with 1 and immediately irradiated with 405
nm light. After purification by gel electrophoresis, the band
corresponding to NTR displays fluorescence at 602 ± 50 nm, whereas
a non-irradiated control sample remains completely non-fluorescent.
These data confirm that photoactivation immediately after the
enzymatic transformation of 1 to 2 results in covalent modification of
the activating enzyme with a fluorescent label. This approach to label
enzymes was further tested in live HEK 293 cells that were incubated
with 1 and exposed to 350 nm light (8 lamps of 8 W each). Gel
electrophoresis of lysates revealed a few distinct fluorescent bands
indicating enzymes with potential nitroreductase activity (Figure S15).
Although the probe is able to label the nitroreductases
covalently, many molecules might still escape the proximity of the
protein, react with water, and become a diffusible small molecule
(Figure 3A). We noticed, however, that this small molecule has a very
low quantum yield of emission in water. In contrast, the quantum yield
of emission in apolar solvents is much higher (Figure S16). This
fluorogenicity provides additional contrast for specific imaging of
fluorophores covalently linked to proteins that provide an apolar
environment.[19,20]
Encouraged by these results, we tested whether this covalent
labeling also occurs in nitroreductases in live HEK 293 cells. We used
405 nm light to irradiate a small region of interest (ROI) within a cell
that had been incubated with 1 and observed the diffusion of
fluorescent molecules over a course of 10 min (Figure S17). Analysis
of the signal revealed that the fluorescence stayed within the ROI,
indicating minimal diffusion of the fluorophores formed by
photoactivation of 2 within the cell. Lipophilic organic molecules
carrying a permanent positive charge are known to be taken up and
retained by mitochondria.[21,22] The fluorophore generated by
photolysis of 2 carries a permanent positive charge, which makes it
amenable for uptake and retention in mitochondria without covalent
labeling. To exclude this possibility, we used the protonophore
carbonyl cyanide m-chlorophenyl hydrazine (CCCP) to depolarize
mitochondria, which induces dissipation of positively charged
species.[23] Cells pretreated with 1 and the cationic dye 4, which also
localizes to mitochondria (Figure S18), were exposed to 405 nm light.
Fluorescence images taken at 515 nm and 640 nm revealed that the
signals of the two fluorophores colocalize with each other and with
mitochondria (Figure 3D). Whereas addition of CCCP (20 μM)
induced dissipation of compound 4 after 30 min, photoactivated
compound 1 remained in mitochondria, resulting in a drastic decrease
in the colocalization between compounds 1 and 4 (Figure 3E).
Mitochondria of cells that were not treated with CCCP retained both
compounds equally, resulting in no change in colocalization (Figure
3F). These data, together with the results of the previous experiments,
confirm that 1 can be used to covalently label proteins in cells upon
irradiation.
Photoactivatable probes that label proteins selectively are
ideally suited for single-molecule localization microscopies such as
STORM. Employing probe 1, we set out to map the distribution of
active nitroreductases within the mitochondria of live HEK 293 cells.
In order to label mitochondria evenly and non-specifically, we
prepared compound 5 (Figure 4A), a mitochondria-targeting,[24] far-
red-emitting photoactivatable probe (Figure S19).[17] Using this
compound and probe 1, we performed two-color, three-dimensional
STORM imaging. To minimize diffusion of compound
1 after
Figure 3. A) Proposed mechanism of photolysis of compound 2, ketene
formation and nucleophilic attack. B) Sodium dodecyl sulfate – polyacrylamide
gel electrophoresis of NTR from E. coli crosslinked with fluorophores imaged at
602 ± 50 nm (dark lanes) and as stain-free gel image (white lanes). C)
Structures of compound 4 and CCCP. D) Fluorescence images of live HEK 293
cells treated with compound 1 (10 μM, 10 min) imaged with 515 nm excitation
light (magenta) and compound 4 (0.1 μM, 30 min) imaged with 640 nm excitation
light (yellow) after exposure to 405 nm light (~80 mW, 10 s) before and 30 min
after addition of CCCP (20 μM). Selected ROIs (dotted cyan squares) are
displayed enlarged as insets. E) Colocalization between the signals obtained
from compounds 1 and 4 in untreated cells before and 30 min after addition of
CCCP (20 μM). F) Colocalization between the signals obtained from
compounds 1 and 4 in control cells without addition of CCCP over a course of
30 min. Scale bars: 5 μm (panel D, insets: 1 μm). Graphs display individual
enzymatic conversion and obtain super-imposable signals of the two
imaging channels, we applied a 405 nm photoactivation pulse (100
μs), followed by 40 alternating readout frames with 514 nm (50 ms)
and 647 nm excitation (50 ms), respectively. This sequence was
repeated every 4 s. From this acquisition, a super-resolved image of
mitochondria as well as a super-resolved image of nitroreductase
activity were reconstructed (Figure 4B) with average localization
precisions of 40 ± 17 nm and 63 ± 18 nm, respectively. It must be
mentioned that typical STORM experiments usually achieve better
localization precisions (<20 nm), but most of these experiments
employ fixed cells and do not report on the activity of the enzyme. In
this case, imaging enzymatic activity requires live-cell STORM
imaging and the localization precision in these experiments is
values, averages, and standard deviations for
N = 7 independent
measurements from three biological replicates. Nu = nucleophile.
This article is protected by copyright. All rights reserved.