A caged retinoic acid for one- and two-photon excitation in zebrafish embryos

Article abstract of DOI:10.1002/anie.200800037

(Chemical Equation Presented) Escaping the cage: Retinoic acid (RA), a crucial signaling molecule for the embryogenesis of vertebrates, can be photoreleased from a simple caged derivative (cRA) upon illumination. In zebrafish embryos, cRA causes RA-induced phenotypes with one- and two-photon excitation (see picture), which opens a route to the noninvasive generation of controlled RA concentration patterns in vivo.

Full text of DOI:10.1002/anie.200800037

Communications
DOI: 10.1002/anie.200800037
Embryogenesis
A Caged Retinoic Acid for One- and Two-Photon Excitation in
Zebrafish Embryos**
Pierre Neveu, Isabelle Aujard, Chouaha Benbrahim, Thomas Le Saux, Jean-François Allemand,
Sophie Vriz, David Bensimon, and Ludovic Jullien*
The embryogenesis of a multicellular organism relies on the
spatiotemporal control of the concentrations of signaling
molecules and the way cells respond to them.[1] One of these,
retinoic acid (RA), plays a key role in patterning the body axis
and in the formation of many organs.[2] RA concentration is
finely regulated both spatially and temporally and its alter-
ation results in developmental defects. An ability to artifi-
cially tune the RA concentration at the single-cell level would
bring new insight into these phenomena.
densation of RA with 7-dimethylaminocoumarin-4-yl meth-
anol[9] in the presence of DCC with 84% yield. We also
similarly synthesized from the corresponding 2-nitrobenzyl
alcohols[10] three other RA esters: cRA, cRABr, and cRACN
(see Scheme 1 and the Supporting Information).
The RA esters were preliminarily screened for their
potency by injection into zebrafish embryos. While none of
them was active in the dark, cRA was retained as the most
appropriate for further studies: 1) it was the most soluble
compound; 2) it produced the most pronounced teratogenic
effects upon UV (365 nm) illumination; and 3) it can be
readily synthesized from commercially available materials.
We first characterized the uncaging of cRA in vitro by
UV/Vis absorption and capillary electrophoresis (CE) to
estimate the typical time associated not only with uncaging
(t2) but also with photoisomerization (t1) and photodegrada-
tion (t3), processes known to occur in the parent compound
RA under UV illumination.[11] A drop in absorbance at
approximately 350 nm of a cRA solution after UV illumina-
tion for typically 20 s (Figure 1a) is also observed in RA
solution. This phenomenon is associated with relaxation
toward a photostationary state involving a dynamic exchange
between different RA stereoisomers, among which the trans-
RA isomer represents 20% (see Supporting Information).
In contrast, the blue shift of the absorbance at intermedi-
ate times (100–1000 s) significantly departs from the behavior
of RA (Figure 1b). The shift is associated with uncaging, as
Caged compounds,[3,4] which are molecules with a photo-
labile moiety, combined with two-photon excitation[5] provide
an attractive possibility to fulfill this goal, while the zebrafish
presents a particularly promising model organism:[6] its
embryo is transparent, its genome is known, and mutants
are available. Herein, we introduce a caged retinoic acid
(cRA) that can be photoactivated in zebrafish with both one-
and two-photon excitation. We characterize its two-photon
uncaging kinetics in vivo, which allows tuning of the RA
concentration at the single-cell level in a live organism.
In principle, a carboxylic acid like RA can be photo-
released from a carboxylic ester obtained after reaction with
an appropriate alcohol acting as a photolabile group. We
considered two putative series of photoactive RA esters
relying on benzylic coumaryl[7] or ortho-nitrobenzyl moiet-
ies.[8] Thus, we first prepared the 7-dimethylaminocoumarin-
4-ylmethyl all-trans retinoic acid ester (cRACoum) by con-
[*] Dr. P. Neveu, Dr. I. Aujard, C. Benbrahim, Dr. T. Le Saux,
Prof. Dr. L. Jullien
Ecole Normale SupØrieure, DØpartement de Chimie
UMR CNRS-ENS-UPMC Paris 06 8640 PASTEUR
24 rue Lhomond, 75231 Paris Cedex 05 (France)
Fax: (+33)1-4432-3325
E-mail: ludovic.jullien@ens.fr
Dr. P. Neveu, Dr. J.-F. Allemand, Dr. D. Bensimon
Ecole Normale SupØrieure
Laboratoire de Physique Statistique and DØpartement de Biologie
UMR CNRS-ENS 8550, 24 rue Lhomond, 75231 Paris (France)
Prof. Dr. S. Vriz
Inserm U770, HØmostase et Dynamique Cellulaire Vasculaire
80 rue du GØnØral Leclerc, 94276 Le Kremlin-BicÞtre (France)
Prof. Dr. S. Vriz
UniversitØ Paris 7 Denis Diderot
75005 Paris (France)
[**] We thank B. Lesaffre and F. Rosa for access to their fish data and
facility. This work was supported by the Association pour la
Recherche sur le Cancer (ARC, contract no. 3787). The authors
declare no competing financial interests.
Scheme 1. Synthesis and structures of the investigated caged esters
incorporating a photolabile protecting group (PLPG), obtained from
reaction of all-trans RA with the corresponding photolabile alcohols
HO-PLPG. DCC=N,N’-dicyclohexylcarbodiimide, DMAP=N,N’-dime-
thylamino-4-pyridine.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3744
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3744 –3746

Angewandte
Chemie
Figure 1. UV illumination (365 nm) of a cRA solution. a–c) In vitro irradiation of a 25 mm cRA solution in acetonitrile/20 mm Tris pH 9 buffer (1:1,
v/v) as a function of time (t=0, 20, 60, 100, 160, 280, 580, 1180, 1780 s). a) Evolution of the UV/Vis absorption spectra; b) the maximal
*
*
*
*
wavelength of absorption (cRA, ; RA, ; see Supporting Information, Figure 2Sa); and c) the absorbance at several wavelengths (364, ; 374, ;
384, ; 394, &; 404, ; 414 nm, ; lines, global biexponential fit A2e
+A3eÀt/t3). Using t3 =1700 s from the kinetics of RA photodegradation
Àt/t2
~
~
&
(see Supporting Information Figure 2Sb), we extracted t2 =400 s which yielded 35mÀ1 cmÀ1 for the uncaging action cross section for cRA in the
350-nm range. d) Photochemical processes occurring upon cRA photoactivation. e–h) In vivo experiments: 128-cell dechorionated embryos were
incubated at 258C for 90 min in 10 mm cRA. They were then washed with embryo medium and illuminated (or not). The embryos were
subsequently observed at 30 h post-fertilization. The nonilluminated embryos (e) develop normally as do the untreated ones (f). The embryos
illuminated for 80 s (g) exhibit strong teratogenic effects similar to those observed in embryos (h) incubated for 90 min in 1 mm RA.
expected from the longer wavelength of maximal absorption
in cRA (364 nm) than in RA (343 nm) and confirmed by
direct CE observations of several photoreleased intermedi-
ates: RA and also 13-cis- (C13RA) and 9-cis-RA (C9RA),
which are involved in RA photoisomerization. Eventually, the
drop in absorbance at long irradiation times (beyond 1000 s)
as a result of photodegradation is in line with similar
observations on RA (see Supporting Information).
The temporal evolution of the absorbance (between 364
and 414 nm) was analyzed with a simple kinetic model that
accounts for the observed photoinduced processes. We
deduced a typical cRA uncaging rate of (2.5 Æ 1) ” 10À3 sÀ1
(Figure 1c), which compares well with the value of (3.5 Æ
0.5) ” 10À3 sÀ1 obtained with a model compound and with
CE data (see Supporting Information). Hence, UV illumina-
tion of cRA causes three types of photochemical reactions
with well-separated timescales (Figure 1d): first, photoisome-
rization along the polyenic backbone (t1 ꢀ 20 s); then uncag-
ing (t2 ꢀ 400 s) to yield a mixture of continuously intercon-
verting RAs; and finally, photodegradation of the various
compounds (t3 ꢀ 2000 s). To manifest themselves, cRA-pho-
toinduced effects may thus require enough but not too many
photons.
to determine its effects upon UV (365 nm) illumination.
Dechorionated 128-cell embryos were incubated for 90 min in
a 10 mm cRA solution. Consistent with the injection experi-
ment, nonilluminated embryos developed normally (Fig-
ure 1e and f). In contrast, embryos illuminated for 80 s
exhibited the typical phenotype (Figure 1g and h) induced by
RA, C13RA or C9RA, at a concentration of about 1 mm (the
UV dose alone does not induce any defects, see the Support-
ing Information). This result shows that cRA passively
diffuses into dechorionated embryos.
From the in vitro experiments, which gave an uncaging
rate constant of 2.5 ” 10À3 sÀ1 for cRA, we evaluated the
extent of uncaging after 80 s of illumination to be about 20%.
As both cis-RAs produce phenotypes similar to those
observed with RA at the same concentration, we conclude
that the initial cRA concentration within the embryo before
illumination was close to the external incubating concentra-
tion of cRA (10 mm). Moreover, the teratogenic effects
disappeared after too long a UV illumination of cRA (see
Supporting Information), as expected from the in vitro
studies.
We then attempted to control the RA concentration at the
single-cell level using two-photon uncaging. To determine the
appropriate laser parameters and characterize the targeted
cells, a caged fluorophore sharing the same caging group as
cRA was used. As revealed by the rise in fluorescence
intensity, illumination at 750 nm (1.8 mW; below 5 mW, two-
photon illumination alone does not induce any defects, see
Supporting Information) results in uncaging within a typical
time t2 ꢀ 25 s. From the distribution of t2, we also computed
In view of biological application in a whole organism,
cRA has to fulfill several constraints. It should 1) enter the
cells passively, 2) be chemically stable under physiological
conditions, 3) be biologically inactive and nontoxic, and 4) be
uncaged at light levels that do not damage the cell. Having
already established during the screening step that cRA was
nonteratogenic during zebrafish development, we proceeded
Angew. Chem. Int. Ed. 2008, 47, 3744 –3746ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3745

Communications
the average size of a retina cell as (13.5 Æ 6) mm, consistent
induced by RA application on the dorsal part of the retina.[18]
Nondetrimental two-photon excitation (750 nm, 4.5 mW) was
applied in four dorsal cells of the retina of zebrafish incubated
in cRA (see Supporting Information). Figure 2d shows the
typical phenotype observed 15 h later. In comparison with the
nonilluminated control eye (Figure 2c), the illuminated retina
is larger, more elongated, and its dorsal part is slightly
invaginated (see reference [18]). Thus, two-photon uncaging
of cRA reproduces a known specific RA-induced phenotype.
In conclusion, we have shown that cRA exhibits attractive
features for use in a biological context: it permeates cells and
is stable in a zebrafish embryo where it is not biologically
active. Upon one- or two-photon excitation, RA can be
released in a controlled way using nondetrimental light
intensities. In particular, by confinement of two-photon
uncaging to a single cell, cRA could be a useful tool to alter
and investigate key developmental events, such as somito-
genesis, where comparison of models with experiments has
been hampered by the lackof means to precisely set RA
gradients.[19]
with values in the literature[12,13]. Equipped with a quantita-
tive characterization of two-photon uncaging of a model
compound, we investigated the kinetics of RA photorelease
in a retina cell and the photoinduction of a RA-induced
phenotype in the embryo retina (Figure 2a).
Although RA is almost nonfluorescent in solution, it
becomes fluorescent when bound to endogenous cytoplasmic
retinol or RA binding protein (CRBP or CRABP).[14,15] The
photorelease and complexation of RA can thus be monitored
through the change in fluorescence induced by its binding to
CRABP and CRBP as it competes with the more fluorescent
endogenous retinol. We did indeed observe a biphasic change
in fluorescence (Figure 2b). The first regime of decreasing
fluorescence was attributed to retinol unbinding from CRBP;
its rate constant koff = 0.06 sÀ1 is consistent with reported
values (0.01–0.1 sÀ1 [16,17]) and the 23% decrease in fluores-
cence is also in line with the 26% drop reported when RA
displaces retinol from its complex with CRBP.[15] The second
regime of fluorescence increase is caused by RA binding to
CRABP. It is limited by uncaging and gives an uncaging rate
constant of 0.015 sÀ1 at 1 mW power and 730 nm excitation,
akin to that measured on a similar caged fluorophore
(0.013 sÀ1). Thus, this experiment suggests that RA is released
upon two-photon cRA uncaging and remains confined to the
illuminated cell volume within the time span of the measure-
ment (about 3 min).
Received: January 4, 2008
Published online: April 10, 2008
Keywords: bioorganic chemistry · cage compounds ·
.
embryogenesis · photolysis · retinoic acid
Having estimated the cRA two-photon uncaging rate, we
examined whether two-photon cRA uncaging could cause a
RA-induced phenotype. We relied on retina malformations
[1] A. D. Lander, Cell 2007, 128, 245 – 256.
[2] L. J. Gudas, J. Biol. Chem. 1994, 269, 15399 – 15402.
[3] G. Mayer, A. Heckel, Angew. Chem. 2006, 118, 5020 – 5042;
Angew. Chem. Int. Ed. 2006, 45, 4900 – 4921.
[4] Dynamic Studies in Biology (Eds.: M. Goeldner, R. Givens),
Wiley-VCH, Weinheim, 2005.
[5] W. Denk, Proc. Natl. Acad. Sci. USA 1994, 91, 6629 – 6633.
[6] G. J. Lieschke, P. D. Currie, Nat. Rev. Genet. 2007, 8, 353 – 367.
[7] T. Furuta, S. S. Wang, J. L. Dantzker, T. M. Dore, W. J. Bybee,
E. M. Callaway, W. Denk, R. Y. Tsien, Proc. Natl. Acad. Sci.
USA 1999, 96, 1193 – 1200.
[8] C. G. Bochet, J. Chem. Soc. Perkin Trans. 1 2002, 125 – 142.
[9] V. R. Shembekar, Y. Chen, B. K. Carpenter, G. P. Hess, Bio-
chemistry 2005, 44, 7107 – 7114.
[10] I. Aujard, C. Benbrahim, M. Gouget, O. Ruel, J. B. Baudin, P.
Neveu, L. Jullien, Chem. Eur. J. 2006, 12, 6865 – 6879.
[11] R. W. Curley, Jr., J. W. Fowble, Photochem. Photobiol. 1988, 47,
831 – 835.
[12] Z. Li, N. M. Joseph, S. S. Easter, Jr., Dev. Dyn. 2000, 218, 175 –
188.
[13] Z. Li, M. Hu, M. J. Ochocinska, N. M. Joseph, S. S. Easter, Jr.,
Dev. Dyn. 2000, 219, 391 – 401.
[14] P. D. Fiorella, V. Giguere, J. L. Napoli, J. Biol. Chem. 1993, 268,
21545 – 21552.
[15] P. N. MacDonald, D. E. Ong, J. Biol. Chem. 1987, 262, 10550 –
10556.
Figure 2. Two-photon excitation in vivo. a) Experimental principle: a
single cell (b) or four cells (d) of the right retina of a manually
dechorionated 4–14 somite embryo previously incubated for 90 min in
10 mm cRA is (are) illuminated, and either the temporal variation of
the fluorescence signal (b) or the eye malformation phenotype 15 h
after treatment are observed. b) Variation in fluorescence intensity I(t)
upon cRA uncaging at a laser power of 1 mW. The changes are caused
by retinol unbinding from CRBP, which results in a decrease in
fluorescence, and by RA binding to CRABP, which increases the
fluorescence. Continuous line: double exponential fit
[16] F. M. Herr, E. Li, R. B. Weinberg, V. R. Cook, J. Storch, J. Biol.
Chem. 1999, 274, 9556 – 9563.
[17] N. Noy, W. S. Blaner, Biochemistry 1991, 30, 6380 – 6386.
[18] G. A. Hyatt, E. A. Schmitt, N. Marsh-Armstrong, P. McCaffery,
U. C. Drager, J. E. Dowling, Development 1996, 122, 195 – 204.
[19] A. Goldbeter, D. Gonze, O. Pourquiꢀ, Dev. Dyn. 2007, 236,
1495 – 1508.
off t
I(t)=I0ÀDIoff(1ÀeÀk )+DIon(1ÀeÀkon t), where koff =0.06 sÀ1 is the rate
of retinol unbinding and kon =0.015 sÀ1 is the uncaging rate at 1 mW
power, from which we deduce a two-photon action cross section of
25 mGM at 750 nm, compatible with reported values.[10] c) Control and
d) illuminated retina of the same embryo. The illuminated eye exhibits
an invagination (arrow) which is characteristic of a RA-induced
phenotype.
3746
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3744 –3746