Angewandte
Chemie
GAI1–92 with GA3-PP (1 mM) for 3 h did not result in
observable co-localization suggesting that the photo-caged
GA3-esters are poor substrates for endogenous esterases.
Moreover, we can exclude the possibility that spontaneous
hydrolysis results in significant amounts of GA3 within the
timescale of our experiments (Figure S3).
Most notably, the bioactivity of GA3 in plants is regulated
by a number of different pathways among which is the
methylation of the C-6 carboxylic acid function.[17] It has been
shown that Arabidopsis thaliana encode gibberellin methyl-
transferases (GAMTs) catalyze the methylation of the C-6
carboxy group of gibberellins[18] to convert gibberellic acid
into the biologically inactive C-6 methyl ester. By caging the
C-6 carboxy group we created a photo-activatable CID simply
by utilizing the naturally evolved structure–activity relation-
ship of gibberellins.
As we did not observe co-localization of GID1 and GAI1–
upon addition of the photo-caged gibberellic acid deriva-
92
tives, we tested whether we can recover the GA3ꢀs bioactivity
by irradiation with light. Therefore we incubated COS-7 cells
expressing EGFP-GID1 and TOM20-mCherry-GAI1–92 with
the DMNPP caged pcGA3-1 for 5 min and irradiated for 3 s
with standard DAPI (DAPI = 4’,6-diamidino-2-phenylindole)
illumination (377/50 nm). Promptly after irradiation we
observed translocation of EGFP-GID1 to the mitochondria
(Movie S1, Figure S4) and complete sequestration was ach-
ieved within seconds. The co-localization of the EGFP and
mCherry signals at the mitochondria clearly indicates that we
efficiently recovered substantial amounts of biological active
GA3.
After having shown the general applicability of pcGA3-
1 as a photo-activatable CID in live cells, we wanted to
demonstrate the spatial resolution of this method by activat-
ing pcGA3-1 in a single cell using a 405 nm laser for photo-
activation in a defined region of interest (ROI). We chose two
positively transfected cells in the field of view, first irradiating
a ROI within only one of the cells, resulting in translocation of
EGFP-GID1 to the mitochondria in this specific cell but not
in the neighboring one. Subsequent activation of pcGA3-1 in
the second, neighboring cell likewise led to sequestration of
EGFP-GID1 to the mitochondria which demonstrates the
high spatial control of the photo-activatable CID (Figure S5).
Next, we tested the photo-caged gibberellins pcGA3-2 and
pcGA3-3 in the same experimental setup. Both compounds
can be efficiently activated with epifluorescent DAPI illumi-
nation within seconds (Figure 2b–i). However, both com-
pounds have red-shifted absorption spectra compared to
pcGA3-1 (Figure S1) and we noted that even irradiation with
GFP-filtered light for 10 s resulted in noticeable photo-
mediated release of GA3 (Figure S6). This process has to be
taken into account when using these compounds in combina-
tion with a green fluorescent protein (GFP) read out.
Nevertheless, the red-shifted absorption is of advantage, as
it allows chromatically orthogonal photo-activation of
pcGA3-2 or pcGA3-3 when combined with other photo-
activatable effectors which are insensitive to light of
470 nm.[6,19] Finally we wanted to utilize the excellent 2P-
absorption properties of pcGA3-3 to unleash the active GA3
in live cells by near-infrared laser irradiation at 800 nm. For
Figure 2. Photo-induced protein dimerization using the photo-caged
gibberellic acid derivative pcGA3-3. a) Schematic presentation of the
fluorescence readout for photo-induced protein dimerization. Addition
of caged gibberellic acid remains without effect until light of a suitable
wavelength is used to release the active dimerizer GA3. b–i) Epifluor-
escence microscopy images of photo-induced protein dimerization in
COS-7 cells expressing EGFP-GID1 and TOM20-mCherry-GAI1–92 incu-
bated with pcGA3-3 for 15 min with (1 mm, in complete medium
(0.001% DMSO) at 378C). Images show the localization of EGFP-
GID1 after incubation with pcGA3-3 before and after 10 s irradiation
with DAPI filtered light (377/50 nm). b) Image of GFP-channel after
incubation with pcGA3-3 before irradiation with DAPI-filtered light and
c) after irradiation with DAPI filtered light. d),f),h) Magnification of
the white box in (b); d) green channel, before irradiation, f) red
channel, before irradiation, and h) both channels merged, before
irradiation. e),g),i) Magnification of the white box in (c); e) green
channel, after irradiation, g) red channel, after irradiation and i) both
channels (e) and (g) merged, after irradiation.
can be observed by co-localization of the fluorescent signals
from EGFP and mCherry at the mitochondria (Figure 2a).
Addition of GA3-AM, the cell-permeable acetoxymethyl
derivative, led to efficient co-localization within seconds
(even using concentrations as low as 100 nm in complete
medium) as shown previously. GA3 however, did not result in
co-localization at the mitochondria, indicating that GA3 itself
is not able to pass the cell membrane.[2b] This result led us to
the assumption that, in general, esters of GA3 possess good
cell permeability and we expected similar good cell perme-
ability for the photo-caged esters. GA3-AM is converted into
the active GA3 by endogenous esterases within the cell on
a very short timescale. To investigate the esterase-mediated
release of GA3 from the caged GA3-esters, we synthesized the
2-phenylpropan-1-ol ester of gibberellic acid (GA3-PP, Fig-
ure S3). The compoundꢀs structure is closely related to the
structure of pcGA3-1, missing the o-nitro group thus render-
ing the molecule GA3-PP light-insensitive. Incubation of
COS-7 cells expressing EGFP-GID1 and TOM20-mCherry-
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!