Doxycycline-dependent photoactivated gene
expression in eukaryotic systems
Sidney B Cambridge1,6, Daniel Geissler1,2, Federico Calegari3,6, Konstantinos Anastassiadis4,6, Mazahir T Hasan5,
A Francis Stewart4, Wieland B Huttner3, Volker Hagen2& Tobias Bonhoeffer1
High spatial and temporal resolution of conditional gene
expression is typically difficult to achieve in whole tissues
or organisms. We synthesized two reversibly inhibited,
photoactivatable (‘caged’) doxycycline derivatives with different
membrane permeabilities for precise spatial and temporal
light-controlled activation of transgenes based on the ‘Tet-on’
system. After incubation with caged doxycycline or caged
cyanodoxycycline, we induced gene expression by local
irradiation with UV light or by two-photon uncaging in diverse
biological systems, including mouse organotypic brain cultures,
developing mouse embryos and Xenopus laevis tadpoles. The
amount of UV light needed for induction was harmless as we
detected no signs of toxicity. This method allows high-resolution
conditional transgene expression at different spatial scales,
ranging from single cells to entire complex organisms.
High spatial and temporal resolution of transgene expression is
rarely obtained with the conditional gene expression systems
currentlyavailable.We therefore decidedtouse lightas aninducing
and precision with which it can be manipulated. Previously, light
had been used to regulate expression through caged transcription
factors1,2, caged RNA3, caged plasmids4, caged antisense oligos5or
other caged small-molecule inducers6–10. Even though all of these
attempts were successful in principle, there were alwayslimitations,
particularly with respect to the resolution of transgene expression
or difficulties in the delivery of the caged molecule.
We establishedan alternative method thatallowshigh-resolution
transgene expression by irradiation with light. By combining the
versatile Tet-on systemwiththe precision oflight irradiation, this is
a powerful tool for targeted transgene expression in almost any
biological system. Based on the Tet-on system11, we generated
photolabile (‘caged’) doxycycline derivatives for photoactivated
gene expression allowing for precise spatial and temporal control
of gene expression in various biological paradigms, including
single-cell induction of transgene expression in hippocampal slices
as well as topical activation in developing vertebrates.
Characterization of caged doxycycline derivatives
To produce a light-sensitive Tet-on system, we modified the potent
tetracycline analog, doxycycline, with a caging reagent to reversibly
inhibit its ability to induce tetracycline-dependent transcription12.
Inaddition,we alsosynthesizedasecond analog,2-decarbamoyl-2-
cyanodoxycycline(‘cyanodoxycycline’)with amuch reducedmem-
brane permeability to ensure longer retention inside the cells after
photoactivation. A nitrile modification on tetracycline had pre-
while actually increasing its induction efficiency14. We synthesized
cyanodoxycycline by reaction of the commercially available doxy-
cycline hyclate with N,N¢-dicyclohexylcarbodiimide15(Fig. 1a).
Doxycycline and cyanodoxycycline both contain several side
groups that can be derivatized. After the synthesis of caged doxy-
cycline12, reaction of cyanodoxycycline with 1-(4,5-dimethoxy-
2-nitrophenyl)diazoethane (DMNPE) generated a DMNPE ether
of cyanodoxycycline. NMR spectroscopy analysis revealed that the
DMNPE caging group was attached at the 12 position of cyano-
doxycycline, similar to the caging of doxycycline with DMNPE
(Supplementary Note 1). Some of the physicochemical properties
of caged doxycycline and caged cyanodoxycycline are listed in
Table 1. Both caged compounds are soluble in aqueous buffers,
display acceptable quantum yields and have long-wavelength
absorption maxima at 347 nm (Supplementary Fig. 1). Addition-
ally, both compounds are very stable and can be stored in aqueous
solution at ?20 1C for more than a year (data not shown).
To characterize the wavelength range of efficient uncaging,
we irradiated caged doxycycline at distinct wavelengths with a
defined number of photons. Uncaging was most efficient around
330–350 nm (Fig. 1b) and there was no photolysis at visible
wavelengths as irradiation with 436 and 488 nm did not lead to
detectable photorelease of doxycycline. Therefore, DMNPE-caged
doxycycline and cyanodoxycycline can also be used in combination
with GFP excitation.
We assessed the reduction of tetracycline-dependent expression
owing to the caging group by analyzing the tetracycline-dependent
RECEIVED 28 FEBRUARY; ACCEPTED 15 MAY; PUBLISHED ONLINE 7 JUNE 2009; DOI:10.1038/NMETH.1340
1Max Planck Institute of Neurobiology, Munich-Martinsried, Germany.2Leibniz Institute of Molecular Pharmacology, Berlin, Germany.3Max Planck Institute of
Molecular Cell Biology and Genetics, Dresden, Germany.4Genomics, BioInnovationsZentrum, Dresden University of Technology, Dresden, Germany.5Max Planck
Institute of Medical Research, Heidelberg, Germany.6Present addresses: Interdisciplinary Center for Neurosciences, University of Heidelberg, Heidelberg, Germany
(S.B.C.) and Deutsche Forschungsgemeinschaft Center for Regenerative Therapies, Dresden University of Technology, Dresden, Germany (F.C. and K.A.).
Correspondence should be addressed to S.B.C. (firstname.lastname@example.org).
NATURE METHODS | VOL.6 NO.7 | JULY 2009 | 527
© 2009 Nature America, Inc. All rights reserved.
luciferase expression in HeLa cells16after administration of
doxycycline or caged doxycycline. This assay revealed that the
modification with the DMNPE caging group yielded a caged
doxycycline that only induced transcription if used in amounts at
least 20-fold greater than the amount of unmodified doxycycline
normally used. The dose-response curve of caged doxycycline was
markedly shifted to higher concentrations (Fig. 1c) because of its
reduced ability to induce tetracycline-dependent transcription.
This assay showed that a final concentration of 2 mM doxycycline
induced a full transcriptional response but the same amount of
caged doxycycline did not lead to any detectable luciferase expres-
sion and thus no background in the absence of photoinduction.
When we measured the activity of 2 mM caged doxycycline after
photoinduction, we saw 73% of gene expression activity of 2 mM
unmodified doxycycline (data not shown). Thus, there is a useful
window of caged doxycycline concentrations, 0.5–5 mM, which
produce no detectable background transcriptional activity and the
equivalent concentration of unmodified doxycycline induced a full
Photoactivation with caged doxycycline
To achieve widespread tetracycline-dependenttransgene expression,
we usedatransgenic mouse line inwhichthe potentandubiquitous
CAGGS promoter17controls the expression of a gene encoding a
codon-optimized version of rtTA (irtTA) fused to a mutated
glucocorticoid receptor ligand-binding domain (GBD*)18(Supple-
mentary Note 2 and Supplementary Fig. 2). To release irtTA
from GBD and allow translocation into the nucleus, we adminis-
tered 25 mM dexamethasone in all mouse
experiments including doxycycline controls
as well as irradiated and nonirradiated
caged doxycycline- or cyanodoxycycline-
We used the hippocampi of double-
lacZ mice (CIG-tGFP) to obtain 300 mm
thick long-term organotypic slice cultures.
We performed all subsequent photoactiva-
tion experiments using the following proto-
col: we added either caged doxycycline or
caged cyanodoxycyclinedirectlytothe med-
ium, incubated the samples for specified amounts of time, washed
them in culture medium, irradiated the samples where appropriate
and quickly returned them to culture medium for at least 12 h.
Irradiation was done with an upright epifluorescence microscope
equipped with a standard DAPI filter set (band pass excitation filter
290–370 nm). For all hippocampal cultures, a single 0.5, 1 or 2 s
pulse of DAPI excitation light (?40, ?20 or ?10 air objectives,
respectively) was sufficient for photolysis of caged doxycycline or
caged cyanodoxycycline and the ensuing release of active doxycy-
cline or cyanodoxycycline. Irradiated control slices without caged
doxycycline did not have any GFP fluorescence (Supplementary
Fig. 3a). In general, there was diffuse background fluorescence that
also varied across different regions within the hippocampal slices.
However, this diffuse fluorescence was clearly discernible from the
more intense GFP signal.
We found a scattered distribution of GFP-positive cells in brain
slicesafter4 mMdoxycycline treatment for 16h (Fig.2a).The same
concentration of caged doxycycline did not induce any transgene
expression over the same time period (Fig. 2b). However, exposure
to caged doxycycline followed by photoactivation of the entire slice
produced expression of GFP (Fig. 2c), which was similar to
treatment with unmodified doxycycline. Also, in all experimental
photoactivation paradigms tested, the time course of transgene
expression was identical to that after induction with unmodified
doxycycline (data not shown).
To limit transgene expression to a patch of cells, we reduced the
areaofirradiationbyconfiningthe aperture ofthe microscope field
stop (Fig. 2d). Pseudocoloring and quantitative assessment of
DMNPE caging compound
2 20 200
300 350 400
Photolysis side product
Table 1 | Properties of caged cyanodoxycycline and caged doxycycline
quantum yieldb, j
aIn HEPES buffer solution (pH 7.2).bIn HEPES buffer solution (pH7.2) with acetonitrile (80:20; vol/vol).cThe caging reaction produced two
diastereomers, compounds 2a and 2b (Supplementary Note 1 and Supplementary Table 1), but we used only compound 2a for
photoactivated gene expression because it produces a low transcriptional background activity and higher photosensitivity compared to that
of compound 2b.
Figure 1 | Caged doxycycline or caged
cyanodoxycycline and photoactivated gene
expression. (a) Synthesis and photoactivation of
caged cyanodoxycycline (Supplementary Note 1).
Nitrile group is shown in blue and 1-(4,5-
dimethoxy-2-nitrophenyl)diazoethane in red.
DCC, N,N¢-dicyclohexylcarbodiimide. RT, room
temperature, 20–24 1C. (b) Wavelength
dependence of the uncaging of caged doxycycline.
Numbers above bars indicate irradiation
wavelengths. (c) Transcriptional activity in the
presence of caged, unmodified or no doxycycline
was quantified with a luciferase assay. The dashed
line represents the projected continuation of the
dose-response curve for concentrations that could
not be measured. From the inflection points of the
sigmoidal fits, a 58-fold reduced affinity of the
caged doxycycline compared to doxycycline
528 | VOL.6 NO.7 | JULY 2009 | NATURE METHODS
© 2009 Nature America, Inc. All rights reserved.
the staining intensities confirmed the specificity of photo-
activated transgene expression versus tissue autofluorescence
(Supplementary Fig. 3b). Also, we adjusted the autofluorescence
in the images shown in Figure 2d,e to aid visualization of the
hippocampal slice and the GFP-positive cells. We restricted GFP
expression to a single cell by switching to a ?40 objective and by
irradiating a spot of about 2–3 cells in diameter (Fig. 2e). Thus,
photoactivated gene expression with caged doxycycline allows
induction of transgenes with high spatial resolution, including
A major concern regarding the photoactivation approach is that
either the UV-light irradiation itself and/or the side product of the
uncaging reaction may be harmful to the cells. We evaluated cell
viability by assessing the extent of apoptosis 24 h after photoactiva-
tion of hippocampal slice cultures using a fluorescent terminal
deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)
reaction. The extent ofapoptosis in a photoactivated culture and in
a nonirradiated control was similar (Fig. 2f,g). TUNEL stain
quantitation revealed no significant difference between the two
conditions (n ¼ 6, non-paired student’s t-test, P ¼ 0.348; Supple-
mentary Fig. 4a). Measuring the extracellular field potentials as
well as the (resting) membrane potential of neurons before and
after irradiation also revealed no physiological differences (Supple-
mentary Fig. 4) suggesting that the amount of UV light needed for
uncaging is probably innocuous.
Photoactivation with caged cyanodoxycycline
We also tested caged cyanodoxycycline in hippocampal cultures
to determine its photoactivation efficiency. Because caged cyano-
doxycycline has a low membrane permeability like its parent
Figure 2 | Photoactivation of caged doxycycline in hippocampal cultures of double-transgenic CIG-tGFP mice.
(a–c) Immunohistochemistry analysis of GFP expression in a culture after 12-h incubation with 4 mM doxycycline (a; fluorescence
mainly in astrocytes), with 4 mM caged doxycycline (b) and with 4 mM caged doxycycline after a 2-s photoactivation (c).
(d) Immunohistochemistry analysis of GFP expression in a group of cells (d) and in a single cell (e). The approximate areas
of irradiation are circled. Inset in e, high-magnification confocal image of the GFP-positive cell. Other local fluorescent spots
are staining artifacts. (f,g) The fluorescent TUNEL staining of the slices shown in b and c, respectively, showing dying cells
in unirradiated and irradiated cultures. Scale bars, 500 mm (a,e) and 125 mm (d).
Figure 3 | Photoactivation of caged cyanodoxycycline in hippocampal cultures.
(a,b) Fluorescence images of cultures of CIG-tGFP mice after photoactivation of
2.4 mM caged cyanodoxycycline (a) and without irradiation (b). (c) ‘Smiley face’
pattern of photoactivated GFP expression in culture of CIG-tGFP mice; GFP
expression, red, and DAPI staining, blue. (d) Locally restricted photoactivated
GFP expression in the dentate gyrus of a culture of CIG-tGFP mice. The circled
area indicates the approximate area of irradiation. (e–g) Confocal fluorescence
images of hippocampal slices of wild-type mice; slices were doubly transfected
with rAAV viruses carrying the gene encoding human synapsin-rtTA,
tet-dependent tdTomato red fluorescence was pseudocolored in green) after
48 h incubation with 3.1 mM cyanodoxycycline (e), after 24 h incubation with
3.1 mM caged cyanodoxycycline followed by global photoactivation (f), and
after local photoactivation (in a single neuron) (g). Red fluorescence was
pseudocolored in green. (h) Immunohistochemistry analysis in a slice showing
local GFP expression after two-photon irradiation in two planes in a culture
incubated with 1.9 mM caged cyanodoxycycline. Inset, approximate location
of the planes (red lines) close to the surface of the culture. (i) ‘Side view’
of a confocal z-dimension stack of the entire slice culture in h shows
GFP expression (red) restricted to the upper irradiated part of the culture.
(j) Immunohistochemistry analysis of GFP expression after a single pulse
UV-light irradiation of the entire culture. (k) ‘Side view’ of a confocal
z-dimension stack of the entire culture in j reveals GFP expression (red)
throughout the tissue (green background fluorescence). Cultures of CIG-tGFP
mice are shown in h–k. Scale bars, 500 mm (a,c,d,j), 125 mm (h), 60 mm (e)
and 40 mm (f).
NATURE METHODS | VOL.6 NO.7 | JULY 2009 | 529
© 2009 Nature America, Inc. All rights reserved.
compound cyanodoxycycline, it required about 5–10 times longer
incubation times for full transgene expression after irradiation
(data not shown).
widespread transgene expression compared to doxycycline whereas
the background expression in the absence of light was similar
(Fig. 3b). Presumably, the nitrile modification of the cyanodoxy-
cycline increased the induction efficiency similarly to what has been
reported for ‘cyanotetracycline’14. To demonstrate the enhanced
photoactivation efficiency, we used a custom-made microscope
insert with the shape of a smiley face to irradiate a hippocampal
culture. The shape of the smiley face was reproduced by the GFP
transgene expression (Fig. 3c) across various subregions of the slice
as indicated by the DAPI counterstain. With photoactivated gene
expression, the three main subregions, dentate gyrus (Fig. 3d), CA3
and CA1, could also easily be targeted (Supplementary Fig. 5a,b).
To circumvent the need of having to use genetically manipulated
with virus-mediated delivery of transgenes. We specifically directed
using two adeno-associated viruses (rAAV) including one driving
rtTA expression using the human synapsin promoter19. Many
doubly transfected neurons showed strong fluorescence upon
administration of 3.1 mM cyanodoxycycline for 48 h (Fig. 3e).
Photoactivation after incubation with 3.1 mM caged cyanodoxy-
cycline also produced many fluorescent neurons (Fig. 3f) whereas
cultures that were not irradiated displayed no fluorescence
(Supplementary Fig. 5c). Local irradiation of only a small spot
limited expression to a single neuron (Fig. 3g) or two adjacent
neurons (Supplementary Fig. 5d).
Two-photon mediated uncaging of caged cyanodoxycycline
Theuseoftwo-photonmicroscopy forphotoactivated geneexpres-
sion would extend transgene expression to less-transparent tissues
with particular emphasis on in vivo applications. After incubation
with 1.9 mM caged cyanodoxycycline, we irradiated hippocampal
cultures in a plane of 100 ? 100 mm with a wavelength of 710 nm.
Originally, we scanned several stacks of 50–100 planes (z-dimension
steps of 1 mm and scan time of 1 s), which did not produce any
photoactivated transgene expression (data not shown). Instead, if
only two planes were irradiated (about 5–8 mm separation between)
with 500 scans each, transgene expression was robustly induced
(Fig. 3h). A ‘side-view’ of a z-dimension stack through the entire
slice demonstrated that photoactivated transgene expression only
occurred in the upper, irradiated parts of the culture, whereas the
(Fig. 3i). For comparison, irradiation of the entire culture with a
single pulse of UV light produced widespread transgene expression
throughout the slice and penetrated all layers (Fig. 3j,k).
Photoactivated gene expression in living organisms
For photoactivationin entire organisms, we combined photoactiva-
tion with the mouse whole-embryo culture system, which repro-
duces in utero development of postimplantation mouse embryos
at the morphological, cellular and molecular level20–22. For these
experiments, we used the faster diffusing caged doxycycline because
diffusion in the compact embryo is expected to be more limited.
To characterize the extent of inducible transgene expression in
embryos in the presence of 20 mM doxycycline for 24 h (Fig. 4a).
Irradiation was done with a ?4 air objective (15 s) using a DAPI
similar to that seen with unmodified doxycycline although the
expressionwas somewhat reduced. Notably, nonirradiated embryos
showed no detectable GFP fluorescence except for the trigeminal
ganglia, in which we observed low fluorescence (Fig. 4b,c).
To determine whether transgene expression could be enhanced
by repeated photoactivation, we performed each step of incubation
with caged doxycycline and irradiationtwice. In fact, double photo-
activation produced GFP fluorescence that markedly increased
compared to singly irradiated embryos (Fig. 4d). Finally, in one
embryo, we limited UV-light irradiation to about three dorsal root
ganglia, resulting in highly localized GFP fluorescence restricted to
the photoactivated area (Fig. 4e). Pseudocoloring and quantitative
assessment of the staining intensities confirmed the specificity of
photoactivated transgene expression versus tissue autofluorescence
(Supplementary Fig. 3c).
Figure 4 | Photoactivated gene expression in
living organisms. (a) Fluorescence image of a
control embryonic day 10.5 (E10.5) embryo (CIG-
tGFP) incubated for 24 h with 20 mM doxycycline,
displaying widespread GFP fluorescence particularly
in neurons of the spinal cord dorsal root ganglia,
the developing heart and in trigeminal ganglia of
the head. (b) Fluorescence image of an E10.5
embryo (CIG-tGFP) incubated with 2.6 mM caged
doxycycline without irradiation. (c,d) Fluorescence
in an E10.5 embryo (CIG-tGFP) after a single 15-s
pulse of UV-light irradiation (c) and after a second
pulse of irradiation 3 h after the first pulse (d).
(e) Highly restricted fluorescence in three dorsal
root ganglia of an E10.5 embryo (CIG-tGFP) after
localized irradiation at the tip of the tail with
a spot size of about three ganglia (circle).
Inset, enlargement of the irradiated area
(box). (f) Brightfield image of a tadpole.
(g–i) Fluorescence micrographs of the dorsal region of interest (boxed region in f) after induction with 100 mM unmodified doxycycline in the dorsal region (g),
with 65 mM caged doxycycline after photoactivation of the whole tadpole for 15 s (h) and with caged doxycycline in the absence of UV-light irradiation (i).
Scale bars, 1 mm (a,g).
530 | VOL.6 NO.7 | JULY 2009 | NATURE METHODS
© 2009 Nature America, Inc. All rights reserved.
Overall, we observed no anatomical differences between
untreated control embryos, nonirradiated and photoactivated
embryos as all embryos had normal somite numbers, size and
shape of developing organs, and limb formation (data not
shown). Therefore, high-resolution photoactivated gene expres-
sion by itself does not interfere with the complex developmental
programs of living organisms.
Photoactivated gene expression in Xenopus tadpoles
To determine whether photoactivation could be used to induce
transgenes in other organisms, we tested our method in Xenopus
tadpoles. The transgenic tadpoles contained a construct with the
collagen promoter ubiquitously driving rtTA and a tetracycline-
dependent nuclear GFP23. Doxycycline-dependent GFP fluores-
cence was most readily induced in a dorsal structure that appeared
to be part of the spinal cord (Fig. 4f). We observed abundant
nuclear GFP fluorescence in a 2-week-old tadpole that had been
incubatedwith 100 mMdoxycycline for 12h (Fig.4g).We observed
the same pattern of GFP fluorescence after whole-tadpole photo-
activation with 65 mM caged doxycycline (Fig. 4h). Control
tadpoles that we only incubated with caged doxycycline but did
not irradiate did not show any GFP fluorescence (Fig. 4i). Irra-
any obvious developmental defects.
The advantage of using light as the inducing agent is its ease of
manipulation and the ability to identify target cellsby, forexample,
GFP fluorescence and to then photoactivate them by simply
switching excitation filters. We had previously demonstrated loca-
lizedphotoactivatedtransgeneexpression in cellcultureand plants,
and have now extended this paradigm to mice and frogs, in
addition demonstrating single-cell induction of transgene expres-
sion in mouse brain cultures. Our studies indicate that the photo-
activated Tet system will work wherever the Tet system works.
However, Tet-responder transgenic mice do not allow doxycycline-
induced, rtTA-dependent gene activation in a majority of post-
mitotic neurons19. To circumvent this problem, we introduced a
Tet-responder in neurons via rAAV and successfully applied this
methodology for photoactivated gene expression in neurons of
wild-type hippocampal cultures19. In all systems tested so far, a
single pulse of photoreleased doxycycline was sufficient for sub-
stantial transgene expression. In addition, photoactivation can be
repeated at least once for enhanced induction. We foresee a wide
uncaging to achieve high-resolution in vivo transgene expression
deepinsidethe brain orotherorgans. Oneof themain applications
will be in developmental biology, an area of research in which
regional and pattern-specific regulation of gene expression is
central to many questions. We believe that this will be a powerful
approach for elucidating and manipulating the individual contri-
butions of cells in the complex network of cell-cell interactions.
Methods and any associated references are available in the online
version of the paper at http://www.nature.com/naturemethods/.
Note: Supplementary information is available on the Nature Methods website.
We thank I. Mansuy and O. Griesbeck for critically reading the manuscript,
D. Brown and members of his laboratory for help with the photoactivation of the
Xenopus tadpoles, V. Staiger for conducting the electrophysiological measurements,
C. Huber and F. Voss for technical assistance, and P. Schmieder and M. Beerbaum
for recording the NMR spectra. This work was supported by the Max Planck Society.
Additional support came from the Ford Foundation, the Minorities in Neuroscience
Fellowship Program and a grant of the Volkswagen-Stiftung to S.B.C., F.C. and
W.B.H were supported by the Federal Ministry of Education and Research (BMBF)
in the framework of the National Genome Research Network, Systematic
Methodological Platform RNAi, Fo ¨rderkennzeichen 5 (NGFN-2).
Published online at http://www.nature.com/naturemethods/
Reprints and permissions information is available online at
1. Cambridge, S.B., Davis, R.L. & Minden, J.S. Drosophila mitotic domain boundaries
as cell fate boundaries. Science 277, 825–828 (1997).
2. Minden, J., Namba, R., Mergliano, J. & Cambridge, S. Photoactivated gene
expression for cell fate mapping and cell manipulation. Sci. STKE 2000, PL1
3. Ando, H., Furuta, T., Tsien, R.Y. & Okamoto, H. Photo-mediated gene activation
using caged RNA/DNA in zebrafish embryos. Nat. Genet. 28, 317–325 (2001).
4. Monroe, W.T., McQuain, M.M., Chang, M.S., Alexander, J.S. & Haselton, F.R.
Targeting expression with light using caged DNA. J. Biol. Chem. 274,
5. Tang, X., Swaminathan, J., Gewirtz, A.M. & Dmochowski, I.J. Regulating gene
expression in human leukemia cells using light-activated oligodeoxynucleotides.
Nucleic Acids Res. 36, 559–569 (2008).
protein expression. Chem. Biol. 9, 1347–1353 (2002).
7. Link, K.H. et al. Photo-caged agonists of the nuclear receptors RARgamma and
TRbeta provide unique time-dependent gene expression profiles for light-
activated gene patterning. Bioorg. Med. Chem. 12, 5949–5959 (2004).
8. Shi, Y. & Koh, J.T. Light-activated transcription and repression by using
photocaged SERMs. ChemBioChem 5, 788–796 (2004).
9. Hayashi, K. et al. Caged gene-inducer spatially and temporally controls gene
expression and plant development in transgenic Arabidopsis plant. Bioorg. Med.
Chem. Lett. 16, 2470–2474 (2006).
10. Young, D.D. & Deiters, A. Photochemical activation of protein expression in
bacterial cells. Angew. Chem. Int. Edn. Engl. 46, 4290–4292 (2007).
11. Gossen, M., Bonin, A.L., Freundlieb, S. & Bujard, H. Inducible gene expression
systems for higher eukaryotic cells. Curr. Opin. Biotechnol. 5, 516–520 (1994).
12. Cambridge, S.B., Geissler, D., Keller, S. & Curten, B. A caged doxycycline analogue
for photoactivatedgene expression. Angew. Chem. Int. Edn. Engl. 45, 2229–2231
13. Sigler, A., Schubert, P., Hillen, W. & Niederweis, M. Permeation of tetracyclines
through membranes of liposomes and Escherichia coli. Eur. J. Biochem. 267,
14. Lederer, T. et al. Tetracycline analogs affecting binding to Tn10-encoded Tet
repressor trigger the same mechanism of induction. Biochemistry 35, 7439–7446
in the tetracycline series. J. Antibiot. (Tokyo) 34, 34–39 (1981).
16. Baron, U., Gossen, M. & Bujard, H. Tetracycline-controlled transcription in
eukaryotes: novel transactivators with graded transactivation potential. Nucleic
Acids Res. 25, 2723–2729 (1997).
17. Niwa, H., Yamamura, K. & Miyazaki, J. Efficient selection for high-expression
transfectants with a novel eukaryotic vector. Gene 108, 193–199 (1991).
18. Anastassiadis, K. et al. Apredictable ligandregulatedexpressionstrategy for stably
integrated transgenes in mammalian cells in culture. Gene 298, 159–172 (2002).
19. Zhu, P. et al. Silencing and un-silencing of tetracycline-controlled genes in
neurons. PLoS ONE 2, e533 (2007).
20. Osumi, N. & Inoue, T. Gene transfer into cultured mammalian embryos by
electroporation. Methods 24, 35–42 (2001).
21. Calegari, F. & Huttner, W.B. An inhibition of cyclin-dependent kinases that
lengthens, but does not arrest, neuroepithelial cell cycle induces premature
neurogenesis. J. Cell Sci. 116, 4947–4955 (2003).
22. Cockroft, D.L. Dissection and culture of postimplantation mouse embryos. In
Postimplantation Mammalian Embryos. A Practical Approach. (eds., Rickwood,
D.A.H. and Hames, B.D.) 15–40 (Oxford University press, Oxford, 1990).
23. Das, B. & Brown, D.D. Controlling transgene expression to study Xenopus laevis
metamorphosis. Proc. Natl. Acad. Sci. USA 101, 4839–4842 (2004).
NATURE METHODS | VOL.6 NO.7 | JULY 2009 | 531
© 2009 Nature America, Inc. All rights reserved.
Irradiation of caged doxycycline at different wavelengths. Caged
doxycycline (25 mM) in HEPES buffer (10 mM HEPES, 120 mM
KCl, pH 7.2) with acetonitrile (80:20, vol/vol) was irradiated
monochromatically at wavelengths of 297, 334, 365, 405, 436 or
488 nm (cuvette, 3.8 cm ? 1 cm ? 1 cm; filling volume, 2.8 ml;
and irradiated area, 2.8 cm2). The monochromatic light was
provided by a 500 W mercury arc lamp (Lot Oriel) equipped
with metal interference filters (Lot Oriel). Using a power meter
(model 1815-C, detector 818-UV, Newport Corporation), the
irradiance (mW cm?2) of the monochromatic light was measured
at the incidence of the cuvette to determine the number of
photons per second and square centimeter. Samples of caged
doxycycline were exposed to 1 or 10 mmol of photons at different
wavelengths. During irradiation, the solution in the cuvette was
stirred by a small magnetic stir bar. The irradiated and control
nonirradiated samples were analyzed by HPLC (HP 1100 series,
Hewlett-Packard; column, EC 250/4 Nucleodur 100-5 C18 ec,
Macherey-Nagel; eluents: 0.1% aqueous TFA (solution A) with
acetonitrile (solution B); flow rate: 1 ml min?1; gradient: 5–95% B
in 25 min; UV-light detection wavelengths: 254 and 350 nm), and
the extent of photolysis was calculated from the decrease in the
peak area upon irradiation.
Luciferase assay. The luciferase assay was performed according to
the manufacturer’s protocol (Promega). Stably transfected HeLa
cells expressing tet-dependent luciferase (line X1.6; provided by
H. Bujard) were incubated with the respective concentrations of
caged or unmodified doxycycline for 24 h, the cells were then
washed, lysed, and finally, the lysate was mixed with the luciferase
assay reagent for immediate measurements. The values represent
the average of three independent samples. The luminescence was
read at 555 nm. Caged doxycycline was irradiated for 1 h with a
low-power hand-held UV lamp before adding it to the cells.
Photoactivation of biological specimen. Irradiations were done
on standard upright fluorescent microscopes (Zeiss and Olympus)
equipped with a standard DAPI fluorescent filter set using the
DAPI excitation light for photoactivation of the samples. For the
electrophysiological measurements, photoactivation was per-
formed on an inverted microscope using a FURA filter set for
excitation. Using 100 W mercury lamps, the shutter-controlled
irradiation times were 0.5 s (?40 air objective), 1 s (?20), 2 s
(?10) or 15 s (?4).
Mice. The CAGGS-irtTA-GBD* (CIG) line was obtained by con-
ventional transgenesis of ES cells (Anastassiadis, K. et al., unpub-
lished data). All protocols related to animal experiments were
performed according to the German Animal Welfare legislation,
and the Animal Welfare Officers appointed for the facility oversaw
them. The CAGGS-irtTA-GBD* mice were crossed with a GFPte-
tO7lacZ mouse (G3, kindly provided by R. Sprengel). Double
transgenic mice were identified by PCR genotyping using the
following primers: irtTA1, 5¢-AGTTGGCATTGAGGGCTTGACC-
3¢ and irtTA2, 5¢-CAACTTGGTGCTCCTGGTCCTC-3¢ for the
irtTA-GBD* and primer hGFP12 and hGFP13 for GFPtetO7lacZ24
(Supplementary Table 2). Expression of the irtTA-GBD* fusion in
different organs was analyzed by northern blots and RT-PCRusing
standard protocols. To monitor GFP expression in adult animals,
mice were injected intraperitoneally with a daily dose of 0.5 mg
doxycycline (50 mg ml?1in 50% EtOH) and 0.2 mg dexametha-
sone (10 mg ml?1in 50% EtOH, freshly prepared each day) for
four consecutive days. Mice were sacrificed after injections and
expression of GFP in different organs was analyzed by RT-PCR.
Hippocampal cultures. Postnatal day 2–7 hippocampal cultures
were prepared according to standard procedures25. Hippocampi
were excised from the brain, cut into 300 mm thick slices with a
Tissue Chopper (McIlwain) and placed on a membrane filter.
Photoactivation experiments were performed after 1–5 d in vitro.
For the electrophysiological recordings, wild-type cultures were
prepared from postnatal day 3 pups and the experiments were
performed at 5–6 d in vitro.
Photoactivation procedure and staining of hippocampal cul-
tures. Double transgenic cultures were first incubated in stan-
dard culture medium25with 50 mM Tris pH 7.0 plus the caged
doxycycline or cyanodoxycycline, briefly (15–30 min) trans-
ferred to medium containing 50 mM Tris pH 7.0, 25 mM
dexamethasone, 0.5 mM octanol, irradiated, and were then
quickly returned to medium containing only 25 mM dexa-
methasone. The octanol was included to block gap junctions,
the dexamethasone to allow translocation of irtTA-GBD* from
the cytoplasm to the nucleus. GFP expression could be detected
immunohistochemically within 3 h, but longer incubation of
cultures (12–48 h) increased signal-to-noise. Primary antibo-
dies: GFP 1:1,000 (RBD), glial fibrillary acidic protein 1:100
monoclonal antibody (Sigma), microtubule-associated protein
2 1:10,000 monoclonal antibody (Sigma). Double immunos-
tainings to identify GFP positive cells were done with cultures
aged 20–22 d in vitro, which were continuously exposed to
2 mM doxycycline and 2.5 mM dexamethasone. TUNEL stains
(cell death detection kit, TMR red; Roche) were performed
subsequent to the immunostaining procedure. Membranes with
the antibody stained cultures were completely dried, then
incubated at 37 1C for 1 h with the TUNEL reaction mixture,
and finally washed twice in PBS. TUNEL stained cultures were
analyzed in a blind fashion.
Two-photon illumation–mediated photoactivation of hippo-
campal cultures. Cultures were incubated in caged cyanodoxycy-
cline and prepared the same way as with UV photoactivation.
Cultures were irradiated while on the interface-membrane through
a ?20, 0.4 NA air objective (Zeiss). Irradiations were done on
a custom-built two-photon microscope with excitation light
(l ¼ 710 nm) from a mode-locked Ti-sapphire laser (Mai Tai;
Spectra-Physics). The average light power after the objective
was set to 6 mW. Scanning was controlled by custom-made
software written in LabView (National Instruments). The scan
parameters were the following: 100 ? 100 mm area, 1-s duration,
1 mm pixel size.
Recombinant adeno-associated viruses (rAAV) equipped with
tetracycline-controlled genetic modules. The plasmids rAAV-
Phsyn-rtTA-nM2, rAAV-Ptetbi-Cre/Venus and rAAV-Ptetbi-Cre/
tdTomato are based on the AAV expression plasmid26. High titer
rAAVs with serotypes 1 and 2 were prepared by transfecting one of
the plasmids, rAAV-Phsyn-rtTA-nM2, rAAV-Ptetbi-Cre/Venus, or
© 2009 Nature America, Inc. All rights reserved.
rAAV-Ptetbi-Cre/tdTomato, together with helper plasmids, pDp1
(serotype 1), and pDp2 (serotype 2) in a ratio of 2:1:1. Titers for
these viruses were almost 1011genomic particles per milliliter and
viruses were prepared as previously described19.
Whole-embryo culture and photoactivation. For whole-embryo
culture, embryos are typically collected between embryonic day 7
(E7) to E11 followed by removal of their decidua and Reichert’s
membranes. Embryos are then allowed to develop in vitro for up
to 48 h in continuously oxygenated medium containing immedi-
ately centrifuged mouse serum21,22. Culture of mouse embryos was
performed as previously described21. E10.5 embryos were removed
from the uteri, the yolk sac was opened, and normal development
was allowed to continue in a whole embryo culture incubator
(Ikemoto) in 1.5 ml of a 2:1 mixture of immediately centrifuged,
heat-inactivated mouse serum (Harlan): DMEM, and 0.1 mM
octanol in a continuous-flow (50 cm3min?1) of 60% O2, 5% CO2,
35% N2for 12 h and 95% O2, 5% CO2for the following 12 h. To
identify transgenic CIG-tGFP embryos, tail buds of each embryo
were removed, incubated in 200 mM and 25 mM of doxycycline
and dexamethasone, respectively, and analyzed for GFP expression
3 h later. rtTA/tet-GFP positive embryos were cultured 3 h in
medium containing 2.6 mM caged doxycycline, washed briefly in
PBS, irradiated for 15 s through a ?4 objective and were finally
transferred to medium containing 25 mM dexamethasone. For
double irradiations, the induction protocol was repeated. Thus,
following 3 h in dexamethasone containing medium, the embryos
were transferred back to 2.6 mM caged doxycycline (3 h), washed,
irradiated and finally incubated again in dexamethasone. GFP
fluorescence was assessed the following day.
Xenopus tadpole experiments. Tadpoles contained a rtTA-M2
construct and a tet-dependent GFP fusion with a dominant
negative form of the thyroid hormone receptor a, which localizes
the fusion protein to the nucleus23. For photoactivation, tadpoles
were incubated 24 h in standard 0.1x MMR containing 65 mM
caged doxycycline at room temperature. Tadpoles were then
briefly washed in 0.1? MMR and were subsequently irradiated
for 15 s with an ?4 air objective. GFP expression was assessed
12–16 h after irradiation.
Electrophysiology. Electrophysiological recordings were per-
formed at 32 1C in an external solution (ASCF) containing:
125 NaCl, 2.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 1.25 mM
NaH2PO4, 26 mM NaHCO3, 25 mM glucose saturated with 95%
O2and 5% CO2.The recording chamber was constantly super-
fused with a flow rate of 1 ml min?1. Data were collected at a
sample frequency of 10 kHz, low-pass filtered at 3 kHz for the
patch recordings and band-pass filtered between 1–3 kHz for the
extracellular recordings. Data were analyzed by programs written
in LabView (National Instruments).
Whole-cell patch-clamp. Recording electrodes (B8 MO) were
filled with internal solution containing: 125 mM potassium
gluconate, 5 mM KCl, 10 mM Hepes, 1 mM EGTA, 2 mM Na-
ATP, 2 mM Mg-ATP, 10 mM sodium phosphocreatine. Recordings
were carried out with an Axoclamp 2B (Axon Instruments)
amplifier in current clamp mode. Cells in the CA1 region were
patched in a blind fashion in voltage clamp mode. After establish-
ing a stable patch, the recording mode was switched to current
clamp. Before and after 10-s irradiation of the entire culture using
a ?10 objective, a current–voltage characteristic (I/V curve) was
recorded by injecting current pulses (250 ms, 20 pA steps, ?240 pA
to 160 pA) through the recording electrode.
Extracellular recordings. Field EPSPs were recorded in the CA1
region with a glass electrode filled with 3 M NaCl (8–15 MO). The
stimulating electrode was placed in the Schaffer collateral region.
Stimuli were delivered at a frequency of 0.5 Hz (50–100 mA,
duration 0.1 ms). After baseline recording for 15 min, a 2 s UV-
light pulse (same conditions as for patch experiments) was applied
and after another 15 min of recordings, a second pulse of 10 s was
applied. At the end of the experiment, the light was applied
continuously to test whether the signal eventually breaks down.
The amplitude of the signal was measured as the difference
between the minimum and maximum of the excitatory post-
synaptic potential including population spikes.
24. Krestel, H.E., Mayford, M., Seeburg, P.H. & Sprengel, R. A GFP-equipped
gene expression in mouse. Nucleic Acids Res. 29, E39 (2001).
25. Stoppini, L., Buchs, P.A. & Muller, D. A simple method for organotypic cultures of
nervous tissue. J. Neurosci. Methods 37, 173–182 (1991).
26. Shevtsova, Z., Malik, J.M., Michel, U., Bahr, M. & Kugler, S. Promoters and
serotypes: targeting of adeno-associated virus vectors for gene transfer in the
rat central nervous system in vitro and in vivo. Exp. Physiol. 90, 53–59
© 2009 Nature America, Inc. All rights reserved.