A ligation and photorelease strategy for the temporal and spatial control of protein function in living cells

Article abstract of DOI:10.1002/anie.200501244

(Graph Presented) Design and light: A photocleavable synthetic moiety is ligated to the C terminus of a recombinant protein. In the ligated state, the chemical and protein components influence the properties of one another. Upon irradiation with light, bo

Full text of DOI:10.1002/anie.200501244

Angewandte
Chemie
Photoactivatable Proteins
DOI: 10.1002/anie.200501244
A Ligation and Photorelease Strategy for the
Temporal and Spatial Control of Protein Function
in Living Cells**
Jean-Philippe Pellois and Tom W. Muir*
The temporal and spatial regulation of protein function is
central to biological processes. Hence, the ability to artificially
trigger molecular events in a biologically relevant context is
useful for the study of living organisms. By chemically
modifying a protein with groups that can respond to an
external input such as light, control of protein function inside
live cells can in principle be achieved. For instance, peptides
or proteins that contain photolabile protecting groups
appended to functionalities required for biological activity
can be delivered into live cells and activated by UV
irradiation at a desired time or location, and a cellular
response can subsequently be measured.[1] This can be
accomplished within single cells or in the context of whole
organisms.[2] Although protein photocaging is a powerful tool
to investigate biological systems, this approach has been
limited by the difficulties associated with the preparation of
the caged protein reagents.
To facilitate the semisynthesis of caged proteins and
expand the repertoire of suitable chemical reagents, we report
the design, synthesis, and use of a ligation and photorelease
(LPR) molecule suitable for the introduction of a light-
activatable cleavage site within a semisynthetic protein
(Scheme 1). In contrast to previously reported strategies
relying on site-specific photochemical cleavage,[3] our LPR
approach is based on the ligation of a photocleavable
synthetic moiety to the C terminus of a recombinant protein.
In principle, the chemical and protein components may
influence the properties of one another in the ligated state.
Irradiation with light and subsequent photorelease of each
element may then lead to the recovery of their respective
function. We demonstrate that such chimeric semisynthetic
proteins can be efficiently prepared and photocleaved in
vitro. We also demonstrate that photocleavage of protein
constructs microinjected in live cells can be triggered in a
[*] Dr. J.-P. Pellois, Prof. T. W. Muir
The Laboratory of Synthetic Protein Chemistry
The Rockefeller University
Box 223, 1230 York Avenue, New York, NY 10021 (USA)
Fax: (+1)212-327-7358
E-mail: muirt@mail.rockefeller.edu
[**] We gratefully acknowledge Dr. Alison North for her assistance at the
bio-imaging center of The Rockefeller University. We thank Dr.
Jianxin Shi, Edmund Schwartz, and Amy Tyszkiewicz for help with
cloning procedures. This work was supported by NIH grant
GM55843-07.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 5713 –5717
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5713

Communications
Scheme 1. Schematic representation of the semisynthesis and applications of protein constructs in which a photoactivatable cleavage site is
introduced. A synthetic moiety (e.g. peptide, small molecule) is ligated to a recombinant protein through molecule 1. The semisynthetic protein
is then delivered into living cells where the protein and chemical components can be photoreleased from one another with temporal and spatial
control.
dosable manner and that the subcellular localization of a
protein can be controlled with light.
obtained in near-quantitative yield in less than 1 h. Subse-
quent irradiation of 9 (in 1 mm in PBS, pH 7.5) at 325 nm with
a He–Cd laser for 2 s was sufficient for quantitative photolysis
and led to the formation of a single N-terminal product, Gly-
Lys(TAMRA)-Gly-NH2 (10). The LPR molecule was thus
shown to react rapidly with thioesters and to efficiently
generate the corresponding amide upon photolysis.[7]
The LPR molecule 1 was designed to combine a 2-
aminoethanethiol group required for native chemical ligation,
an o-nitrobenzyl chromophore required for bond cleavage,[4]
and a carboxylic acid tether for facile conjugation to amine-
containing compounds (Scheme 1).[5] For synthetic purposes,
the thiol and amine groups were protected as S-tert-butyl
disulfide and tert-butyl carbamate, respectively. The protected
molecule was synthesized in seven steps from vanillin (24%
overall yield), essentially according to the route developed by
Dawson and co-workers for the synthesis of a native chemical
ligation auxiliary.[6,7] The product is suitable for solid-phase
peptide synthesis (SPPS) using Boc (tert-butyloxycarbonyl) or
Fmoc (9-fluorenylmethoxycarbonyl) protecting groups. Once
deprotected, the 2-aminoethanethiol moiety should be com-
patible with native chemical ligation and expressed protein
ligation (EPL).[8]
The native chemical ligation and photocleavage reactions
were characterized using a model a-thioester peptide, Gly-
Lys(TAMRA)-Gly-SCH2CH2SO3Na (7; TAMRA = 5-car-
boxytetramethylrhodamine), and a peptide synthesized with
1 at its N terminus, 1-Lys-Lys(Fl)-NH2 (8; Fl = 6-carboxy-
fluorescein).[7] The two peptides were incubated in phos-
phate-buffered saline (PBS), 100 mm 2-mercaptoethanesul-
fonic acid sodium salt (MESNa; pH 7.5), and the expected
product, Gly-Lys(TAMRA)-Gly-1-Lys-Lys(Fl)-NH2 (9), was
The applicability of our caging strategy was next tested in
the context of EPL. The enhanced green fluorescent protein
(EGFP) containing the SV40 T-antigen Lys-Lys-Lys-Arg-Lys-
Val nuclear localization signal (NLS) at its C terminus was
used as a model protein and obtained as the recombinant
protein a-thioester (from transthioesterification of an EGFP-
NLS-GyrA intein fusion expressed in E. coli, see
ref. [11]).[9,10] The peptides 1-Lys-Lys(TR) (11) and 1-Lys-
Lys(Palm) (12; TR = Texas Red-X, Palm = palmitoyl) were
synthesized by SPPS and used as model small molecules. The
ligation reactions between the EGFP-NLS a-thioester and
peptides 11 and 12 were monitored by reverse-phase HPLC,
electrospray mass spectrometry (ES-MS), and sodium dode-
cyl sulfate/polyacrylamide gel electrophoresis (SDS-PAGE)
and were complete in less than 12 h (Figure 1a). The products,
EGFP-NLS-1-Lys-Lys(TR) (13) and EGFP-NLS-1-Lys-Lys-
(Palm) (14), were purified and their identity confirmed by ES-
MS. Protein 13 is labeled with a fluorophore that can undergo
intramolecular fluorescence resonant energy transfer (FRET)
with EGFP.[12] Protein 14 is modified with a non-hydrolyzable
5714
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 5713 –5717

Angewandte
Chemie
Figure 1. In vitro characterization of the semisynthetic proteins and their photocleaved products. a) SDS-PAGE analysis, Lanes M=marker;
I=EGFP-NLS-GyrA-CBD; II=EGFP-NLS a-thioester; III=EGFP-NLS-1-Lys-Lys(TR) (13); IV=13 + UV irradiation (15); V=EGFP-NLS-1-Lys-Lys-
(Palm) (14); VI=14 + UV irradiation (15). For all samples, UV irradiation was performed witha He–Cd laser at 325 nm for 2 s. Proteins were
detected by coomassie staining. b) Fluorescence spectra of protein 13 before and after (!15) UV irradiation. Fluorescence excitation was per-
formed at 488 nm. c) Images of test tubes containing protein 14, before and after UV irradiation, in a two-phase system of water and dichloro-
methane. An arrow indicates the interface of the two solvents.
lipid, thereby mimicking a widespread post-translational
phenomenon found in eukaryotic organisms.[13]
30% was observed accompanied with perfect colocalization
of the signals from the green (lex = 488 nm, lem = 525 nm) and
red (lex = 565 nm, lem = 620 nm) channels (Figure 2a). Upon
The photocleavage of 13 and 14 was first tested in vitro.
Both constructs were irradiated with UV light, and
the photolysis products were analyzed by HPLC, ES-
MS, and SDS-PAGE. Under the conditions tested, the
photolytic conversion was found to be quantitative,
with the EGFP-NLS amide 15 (from 13 + UV or 14 +
UV) identified as the only proteinogenic product
(Figure 1a). The fluorescence spectrum of 13 showed
FRET (lex = 488 nm, lem = 620 nm) between EGFP
and TR with an efficiency of 33%. Protein 13 was
used to determine the kinetics and quantum yield of
photolysis by using the decrease in FRET intensity as
a direct measure of photocleavage between the FRET
pair (Figure 1b).[14] Using irradiation at 365 nm and
1 mWcmꢀ2, the kinetics of photolysis was first-order
with a half-life of t1/2 = 178 s and an observed quantum
yield of 0.11. Photolysis was quantitative in less than
2 s when a high-intensity laser was used. The behavior
of 14 before and after photorelease of the lipid moiety
was tested by monitoring the affinity of the protein for
a hydrophobic environment. When dichloromethane
was added to 14 dissolved in PBS containing 10 mm 3-
[(3-cholamidopropyl)dimethylammonio]-1-propane-
sulfonate (CHAPS, pH 7.5), the protein localized
exclusively at the interface of the two solvents
(Figure 1c). Upon irradiation with UV light, the
EGFP-NLS moiety was released from the interface
and diffused in the aqueous phase.
Protein 13 was microinjected into Hela cells
cultured on glass slides and live-cell imaging was
performed on a confocal microscope. A FRET signal
Figure 2. FRET assay. Hela cells were microinjected with 13 and exposed to low-intensity
UV light (360 nm) at varying irradiation times. a) Pseudo-color images from confocal
microscopy. The fluorescence signal from EGFP (lex =488 nm, lem =525 nm), Texas Red
(lex =568, lem =620 nm), and FRET between EGFP and Texas Red (lex =488 nm,
lem =620 nm) were recorded before and immediately following UV irradiation. b) Time
course of photocleavage. The FRET intensity F (averaged over the cytoplasmic area and
normalized at t(irradiation)=0) is plotted versus time of irradiation, along withan expo-
nential curve of best fit, which yielded a first-order rate constant for photolysis of
1.3”10ꢀ2 sꢀ1. Error bars represent the standard deviation obtained from the averaging of
(lex = 488 nm, lem = 620 nm) with an efficiency of two experiments.
Angew. Chem. Int. Ed. 2005, 44, 5713 –5717
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5715

Communications
irradiation of the whole cell at 360 nm (5 mW), the FRET
EGFP was excluded from the nucleus in the absence of light
but underwent free nucleocytoplasmic shuttling after irradi-
ation. Hence, these results raise the possibility of the photo-
control of a proteinꢀs activity not by the caging of a critical
group required for function but rather by control of its
location.[19]
intensity diminished and EGFP fluorescence inversely
increased. The observed kinetics of photolysis was first-
order with a half-life of t1/2 = 53 s (Figure 2b). This demon-
strates that photolysis of a protein in a cell can be achieved in
a dosable manner with low-intensity UV light.[15] This is
especially relevant if different concentrations of a photo-
cleaved species would be expected to trigger different
responses. Interestingly, the red and green signals accumu-
lated in different subcellular locations after UV irradiation
(Figure 2). These results demonstrate that a protein and a
small molecule can be photoreleased from one another,
thereby triggering changes in their respective function or
localization.
The possibility of photocontrolled cellular localization
was further explored with protein 14. It was expected that the
modification of EGFP with a lipid would target the protein to
cellular membranes.[16] Indeed, when microinjected in Hela
cells, 14 localized with great specificity at the plasma
membrane, with less than 6 and 2.5% of the total signal
present in the cytoplasm and nucleus, respectively (Figure 3).
In summary, we have reported the synthesis of a LPR
molecule and demonstrated its utility in the synthesis of caged
proteins. We demonstrated that a photoactivatable cleavage
site can be introduced in a semisynthetic protein containing
complex chemical modifications. Furthermore, we showed
that photocleavage can be achieved in a dosable manner to
generate a controlled amount of cleaved product inside a
living cell. Finally, we demonstrated that the localization of an
EGFP-NLS construct can be controlled by photochemically
releasing a lipid moiety from its C terminus. The chemical
diversity that can be introduced by this approach should
permit the manipulation of a protein in a variety of ways and
the generation of a variety of responses upon photocleavage.
Such caging strategies hold potential for the characterization
of complex biological processes in their native context, single
cells or whole organisms. Direct applications might involve
the photoregulation of proteins containing photocleavable
localization signals and post-translational modifications.
Received: April 8, 2005
Published online: August 1, 2005
Keywords: bioorganic chemistry · drug delivery ·
.
photoactivation · proteins
[1] a) K. Curley, D. S. Lawrence, Curr. Opin. Chem. Biol. 1999, 3, 84;
b) S. N. Cook, W. E. Jack, X. Xiong, L. E. Danley, J. A. Ellman,
P. G. Schultz, C. J. Noren, Angew. Chem. 1995, 107, 1736; Angew.
Chem. Int. Ed. Engl. 1995, 34, 1629; c) K. D. Philipson, J. P.
Gallivan, G. S. Brandt, D. A. Dougherty, H. A. Lester, Am. J.
Physiol. Cell Physiol. 2001, 281, C195; d) K. Zou, W. T. Miller,
R. S. Givens, H. Bayley, Angew. Chem. 2001, 113, 3139; Angew.
Chem. Int. Ed. 2001, 40, 3049; e) M. Ghosh, X. Song, G.
Mouneimne, M. Sidani, D. S. Lawrence, J. S. Condeelis, Science
2004, 304, 743; f) A. Nguyen, D. M. Rothman, J. Stehn, B.
Imperiali, M. B. Yaffe, Nat. Biotechnol. 2004, 22, 993; g) J. P.
Pellois, M. E. Hahn, T. W. Muir, J. Am. Chem. Soc. 2004, 126,
7170; h) M. E. Hahn, T. W. Muir, Angew. Chem. 2004, 116, 5924;
Angew. Chem. Int. Ed. 2004, 43, 5800.
Figure 3. Representative images illustrating the photocontrol of the
subcellular localization of a protein. Hela cells were microinjected with
14 and Dextran-TMR (70 kDa). Dextran-TMR was used as a red fluores-
cent marker that is excluded from the nucleus when injected in the
cytoplasm. The images were acquired on a confocal microscope with
green (lex =488 nm, lem =525 nm) and red (lex =568, lem =620 nm)
channels and pseudo-colored. The fluorescence signals were recorded
before and immediately following UV irradiation (360 nm).
[2] S. B. Cambridge, R. L. Davis, J. S. Minden, Science 1997, 277,
825.
[3] a) P. M. England, H. A. Lester, N. Davidson, D. A. Dougherty,
Proc. Natl. Acad. Sci. USA 1997, 94, 11025; b) M. Endo, K.
Nakayama, Y. Kaida, T. Majima, Angew. Chem. 2004, 116, 5761;
Angew. Chem. Int. Ed. 2004, 43, 5643.
[4] a) A. Patchornik, B. Amit, R. B. Woodward, J. Am. Chem. Soc.
1970, 92, 6333; b) S. R. Adams, R. Y. Tsien, Annu. Rev. Physiol.
1993, 55, 755.
[5] C. P. Holmes, J. Org. Chem. 1997, 62, 2370.
[6] C. Marinzi, J. Offer, R. Longhi, P. E. Dawson, Bioorg. Med.
Chem. 2004, 12, 2749.
[7] See the Supporting Information for experimental details.
[8] T. W. Muir, Annu. Rev. Biochem. 2003, 72, 249.
[9] R. Y. Tsien, Annu. Rev. Biochem. 1998, 67, 509.
[10] M. R. Fontes, T. Teh, B. Kobe, J. Mol. Biol. 2000, 297, 1183.
When the cells were irradiated with UV light, the cleaved
product 15 diffused in the cytoplasm and nucleus with
respective 11- and 15-fold increases in signal. Even though
an NLS was present in the construct, the protein did not
accumulate in the nucleus presumably because its low
molecular weight (28 kDa) allowed it to diffuse through the
nuclear pore complex (NPC).[17] It is therefore interesting to
note that the photocaging of a functional lysine present in the
NLS sequence might not be sufficient to achieve photocontrol
of the nuclear import of a small protein such as EGFP.[18]
Indeed, a protein with a molecular weight below the NPC
diffusion limit ( ꢁ 40 kDa) might diffuse in and out of the
nucleus in both its caged-NLS and uncaged-NLS states. In
contrast, using a non-hydrolyzable but photocleavable lipid,
5716
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 5713 –5717

Angewandte
Chemie
[11] The fusion protein was cloned into the pTXB1
bacterial expression vector (New England Biolab,
Beverly, MA) upstream of the Mycobacterium
xenopi GyrA intein and a chitin binding domain.
Following overexpression in E. coli, the soluble
fraction of the cell lysate was purified on chitin
resin. The EGFP-NLS moiety was cleaved from the
resin by incubation of the beads with MESNa to
yield an EGFP-NLS with a C-terminal a-thioester.
[12] P. R. Selvin, Methods Enzymol. 1995, 246, 301.
[13] M. D. Resh, Cell. Signalling 1996, 8, 403.
[14] Irradiation was performed with either a He–Cd
laser (325 nm and 4.74 Wcmꢀ2) or a Hg lamp
equipped with a narrow-bandwidth filter (365 nm
and 1 mWcmꢀ2).
[15] Rapid (< 2 s) and quantitative photocleavage was
observed when irradiation with high-intensity light
(365 nm, 50 mW) was performed.
[16] a) D. Kadereit; J. Kuhlmann, H. Waldmann, ChemBioChem
2000, 1, 144; b) R. Reents, M. Wagner, S. Schlummer, J.
Kuhlmann, H. Waldmann, ChemBioChem 2005, 6, 86.
[17] a) O. Rosorius, P. Heger, G. Stelz, N. Hirschmann, J. Hauber,
R. H. Stauber, BioTechniques 1999, 27, 350; b) D. Gꢁrlich, U.
Kutay, Annu. Rev. Cell Dev. Biol. 1999, 15, 607.
[18] Y. Watai, I. Sase, H. Shiono, Y. Nakano, FEBS Lett. 2001, 488, 39.
[19] P. J. Belshaw, S. N. Ho, G. R. Crabtree, S. L. Schreiber, Proc.
Natl. Acad. Sci. USA 1996, 93, 4604.
Angew. Chem. Int. Ed. 2005, 44, 5713 –5717
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5717