Angewandte
Chemie
to investigate the potential differences in the biological
activity of distinct DAG species. Here we introduce a set of
photoactivatable (caged) DAGs that feature different fatty
acid compositions in order to address this question.
The design of caged DAGs is straightforward as the
obvious attachment point of the cage is the single hydroxy
group (Scheme 1). In order to attach two different fatty acids,
we employed a recently developed synthetic route that
permits the stepwise introduction of different fatty acids
and features a final deprotection procedure that fully avoids
fatty acid migration, a common problem in DAG syntheses.[24]
We employed two different caging groups, a fluorescent
coumarinylmethylene variety as well as the more standard,
nonfluorescent 4,5-dimethoxy-2-nitrobenzyl (nitroveratroyl)
group, both attached through a carbonate linker.
The two variants were used in a complementary way, as
their intrinsic properties make them especially suited for
different experiments. The coumarinylmethylene group is
cleaved by 405 nm light and enables two-photon uncaging.
Furthermore, its intrinsic fluorescence can be used to quantify
compound loading into cells when the biological activity of
different compounds needs to be assessed. On the other hand,
the relatively broad emission spectrum of coumarin effec-
tively limits the available spectral range for multiparameter
live-cell imaging to the spectral region above 500 nm. In the
case of the nonfluorescent 4,5-dimethoxy-2-nitrobenzyl
(nitroveratroyl) group, the entire spectrum remains available,
but photoactivation has to be performed at 366–375 nm,
which might be phototoxic to cells. We prepared a set of five
different DAGs, three of them from DAGs with natural fatty
acid compositions (SAG = stearoyl-arachidonyl glycerol (1a),
SLG = stearyl-linoenyl (2a), POG = palmitoyl-oleyl (3a), as
well as the commonly used nonnatural 1,2-di-O-octanoyl
glycerol (1,2-DOG, 4a) and its inactive 1,3-isomer 5a which
serves as negative control throughout this study (Scheme 1,
see the Supporting Information for details). The photoacti-
vatable groups were attached through chloroformates in
acceptable to good yields and gave the respective DAGs as
nitroveratroyl (1b–5b) and coumarinylmethylene (1c–5c)
carbonate esters.
We chose HeLa cells as a model system for live cell
application of the caged DAGs and used a double scanner
confocal microscope (Olympus FV1000) in order to combine
uncaging through the microscope objective with simultaneous
live cell imaging. We quantified the compound uptake of the
coumarin-caged DAGs by measuring coumarin fluorescence
and devised a protocol that warranted comparable compound
loading for intracellular photoactivation experiments (see
Supporting Information for details). All caged compounds
were essentially inactive when applied to cells expressing
suitable biosensors. Successful photoreactions were con-
Figure 1. Compound release by local extracellular photoactivation of
1,2-DOG and 1,3-DOG monitored by translocation of the C1-GFP
translocation probe. The imaging medium contained 30 mmolLÀ1 of
the respective compounds 4b, 5b, 4c, and 5c. a) Time-lapse mon-
tages of photoactivation experiments with 1,2-DOG (4b, 5b, 375 nm)
and 1,3-DOG (4c, 5c, 405 nm). The indicated areas (white circles)
were irradiated for 10 s starting at t=15 s. The quantification of the
observed translocation events is described in the Supporting Informa-
tion.
photoactivation of the naturally occurring SAG (1b) fre-
quently resulted in spatially confined translocation events as
opposed to 1,2-DOG (4b) (Figure S4 in the Supporting
Information).
According to transcriptome analysis, the utilized HeLa
cells expressed human TRPC3 and TRPC6 channels, allowing
us to compare the potency of different DAG species. We
initially used the nitroveratroyl-caged DAGs in combination
with the genetically encoded ratiometric GEM-GECO sensor
for Ca2+ imaging.[26] We found a massive increase in intra-
cellular Ca2+ levels when SAG was photoactivated and an
intermediate response for SLG, whereas neither POG nor 1,2-
DOG displayed a significant potency for triggering Ca2+
transients (Figure 2a). Detailed analysis of individual traces
obtained by photoactivation of SAG and SLG showed that
SLG generally induced fewer, shorter, and less intense events.
SAG induced long-lasting Ca2+ elevations which were similar
in amplitude to that after ionomycin treatment, indicating
that the Ca2+ concentration was at least locally in the
micromolar range (Figure S5 in the Supporting Information).
To make sure that the observed Ca2+ transients were indeed
induced by SAG and not by its metabolites, we prepared
caged arachidonic acid and compared the Ca2+ transients
induced by photoactivation. We found that SAG was signifi-
cantly more potent, thereby emphasizing its role as initial
signaling molecule (Figure S6 in the Supporting Information).
To confirm the observed structure–activity relationship, we
used the coumarin-caged DAGs in combination with the
genetically encoded green-emitting G-GECO Ca2+ sensor
which does not interfere with coumarin emission.[26] Again,
SAG was identified as the by far most potent species, followed
by SLG, whereas the other three lipids were largely inactive
firmed by experiments using
a C1-GFP translocation
probe[25] and the caged 1,2-DOG (4b, 4c) or the inactive
1,3-DOG (5b, 5c) (Figure 1). Global plasma membrane
translocation was observed in each extracellular 1,2-DOG
photoactivation experiment for both nitroveratroyl- and
coumarin-caged variants but not for 1,3-DOG photoactiva-
tion. Intracellular photoactivation gave comparable results
(Figure S3 in the Supporting Information). Interestingly,
Angew. Chem. Int. Ed. 2013, 52, 6330 –6334
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6331