Light-cleavable rapamycin dimer as an optical trigger for protein dimerization

Article abstract of DOI:10.1039/c4cc09442e

Rapamycin-induced protein heterodimerization of FKBP12 and FRB is one of the most commonly employed switches to conditionally control biological processes. We developed an optically activated rapamycin dimer that does not induce FKBP12-FRB dimerization un

Full text of DOI:10.1039/c4cc09442e

ChemComm
View Article Online
View Journal
COMMUNICATION
Light-cleavable rapamycin dimer as an optical
trigger for protein dimerization†
Cite this:DOI: 10.1039/c4cc09442e
Kalyn A. Brown,ab Yan Zou,b David Shirvanyants,c Jie Zhang,b Subhas Samanta,a
Pavan K. Mantravadi,a Nikolay V. Dokholyanc and Alexander Deiters*ab
Received 26th November 2014,
Accepted 17th February 2015
DOI: 10.1039/c4cc09442e
www.rsc.org/chemcomm
Rapamycin-induced protein heterodimerization of FKBP12 and FRB
The macrolide rapamycin (Rap) was isolated in the 1970s and has
is one of the most commonly employed switches to conditionally received substantial attention from the scientific community due to
control biological processes. We developed an optically activated its immunosuppressant activity.10 Naturally, rapamycin binds to the
rapamycin dimer that does not induce FKBP12-FRB dimerization FK506 binding protein, also known as FKBP12, with a Kd between
until exposed to light, and applied it to control kinase, protease, and 0.2 and 0.4 nM.11 The FKBP12-rapamycin complex then binds to the
recombinase function.
FRB (FKBP12-rapamycin binding) domain of mTOR (mammalian
target of rapamycin), forming a ternary complex consisting of the
Nature requires precise control of biological processes on the cellular two proteins and the small molecule.10,11 This rapamycin-induced
and multicellular level. Processes such as DNA replication,1 cell protein dimerization has been extensively exploited as a research
signaling,2 molecular transport,3 and protein biosynthesis4 tool to control a wide range of cellular processes.12–19 FKBP12 and
depend on the location of proteins, DNA, and other biomolecules FRB are modular proteins that have been fused to a variety of protein
and the timing of their activity and interactions. In trying to targets to conditionally regulate them, including transcription factor
elucidate the mechanisms of cellular processes, it is crucial to domains,12 kinases,20 split enzyme systems such as Tobacco Etch
examine them with protein switches that allow for precise control Virus Protease (TEVp)13 and Cre recombinase,14 as well as cellular
comparable to that of the natural system.5 Light is a unique localization domains21 in order to allow for control over protein
external regulatory element that can be used to probe the regula- function with the addition of rapamycin.
tion of protein switches that are linked to biological pathways.
Manipulating rapamycin to render it light-activatable through
The simplicity and accuracy of timing, location, and amplitude the installation of a caging group will provide precise control over
associated with light exposure makes it an ideal non-invasive triggering the aforementioned biological processes. Recently, we
technique to use in biology.6 Photochemical control of bio- reported a caged rapamycin in conjunction with an engineered
molecules can be achieved through the application of light- FKBP domain.20 We demonstrated that the caged Rap had a
removable protecting groups, so called ‘‘caging groups’’. The addi- similar activity as wild type rapamycin in mediating the dimeriza-
tion of a caging group to a critical position within the biomolecule tion of natural FKBP and FRB. Thus, a truncated mutant of
of interest renders the molecule inactive until the group is removed FKBP12, termed iFKBP, was developed in order to improve the
through light exposure, typically UV light of Z365 nm. Caging sensitivity of the interaction between caged Rap and FKBP.
groups can be installed on small molecule effectors,7 proteins,8
Since caged Rap was found to be still active in mediating the
and oligonucleotides.9 As presented here, the natural product heterodimerization of FKBP12 and FRB, we hypothesized that a
rapamycin was chemically modified with a caging group to allow larger caging group was required in order to fully abrogate its
for photochemical control of FKBP12 and FRB dimerization.
function. Furthermore, we speculated that an efficient way to
substantially increase the size of the caging group is by using a
second rapamycin molecule as a sterically demanding structure
that is able to recruit an even more sterically demanding second
FKBP protein. Here we present a new caged rapamycin, the light-
cleavable rapamycin dimer dRap (Fig. 1). This analog is fully
compatible with wild-type FKBP12, thus obviating the need for
protein engineering of the many established rapamycin-controlled
processes.12–19 The rapamycin dimer dRap was synthesized from
rapamycin in just two steps (Fig. S1, ESI†).
a Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260, USA.
E-mail: deiters@pitt.edu
b Department of Chemistry, North Carolina State University, Raleigh,
NC 27695-8204, USA
c Department of Biochemistry and Biophysics, University of North Carolina,
Chapel Hill, NC 27599, USA
† Electronic supplementary information (ESI) available: Molecular dynamics
simulations, synthetic and biological protocols. See DOI: 10.1039/c4cc09442e
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
Fig. 2 K-LISA mTOR assay of Rap and dRap. (a) Cartoon representation of
K-LISA mTOR Activity Assay. (b) In the absence of Rap or dRap, mTOR is
functional, leading to high luminescence. dRap (50 nM) is blocked from
inducing the heterodimerization of FKBP12 and mTOR, as shown by the
high levels of luminescence before irradiation. Irradiation with UV light at
365 nm for 3 minutes converts dRap into Rap, which leads to formation of
the FKBP12-Rap-mTOR complex and to the same low levels of phosphorylation
and thus luminescence as observed in the case of Rap.
Fig. 1 The structure of the photo-cleavable rapamycin dimer dRap.
Fluorescence polarization was used to investigate the capability
of dRap to homodimerize a fluorescein-labeled FKBP12 (via
thiourea formation through reaction with FITC) (Fig. S4, ESI†).
Indeed, the addition of dRap to a solution of FKBP12-FITC led
to an increase in fluorescent polarization, indicating homo-
dimerization of FKBP12-FITC. Moreover, further addition of a
10-fold excess of dRap led to a decrease in fluorescence rapamycin as a result of the mTOR-Rap-FKBP12 complex for-
polarization due to the saturation of FKBP12 binding sites with mation. The decaging of dRap after exposure to UV light of
dRap molecules, effectively preventing protein dimer formation. 365 nm releases native rapamycin and FKBP12-mTOR hetero-
Thus, we speculated that the FKBP12-dRap-FKBP12 complex dimerization is observed by a decrease in luminescence to the
could prevent binding of FRB, until it is fragmented through same level as in case of rapamycin. This shows that only upon
photochemical cleavage of dRap. This hypothesis is supported by irradiation and photolysis, but not before, dRap successfully
the comparison of complex formation energies calculated in dimerizes FKBP12-mTOR in vitro.22
extensive molecular dynamics simulations (see the ESI† for a
The ability to tightly control dRap makes it well suited for
detailed discussion). Several different scenarios were computationally light-regulation of biological functions in live cells, as demonstrated
explored to explain the experimentally seen FKBP12 homodimers by two important biological processes – protein cleavage and DNA
instead of FKBP12-FRB heterodimers (see mTOR assay below). First, recombination – that were investigated in tissue culture.
we excluded that steric hindrance imposed by the second rapamycin Tobacco Etch Virus NIa protease (TEVp) is a widely used
molecule may block complex formation, since it appears that FRB protease that recognizes a specific seven amino acid sequence,
binding would be allowed in an outstretched linker conformation. ENLYFQG. Cleavage occurs between the glutamine and glycine
Next, the binding poses of rapamycin and dRap to FKBP12 were residues, and this recognition sequence has been genetically
examined to determine if it is possible that dRap is found in a position engineered into various proteins to either activate or deactivate
that in unfavorable for FRB binding. However, it was found that dRap their function.13 TEVp has been used to intracellularly cleave
frequently assumes FRB favorable binding positions. The final set of proteins from purification tags,23 to cleave stabilization proteins
simulations determined the energies necessary for formation of from bioactive glycopeptides,24 to aid in the discovery of protein–
the dRap mediated FKBP12 homodimer and the dRap mediated protein interactions,25 and in the inactivation of proteins in vivo.26
FKBP12-FRB heterodimer. These showed that the formation of dRap Conditional control of TEVp activity was achieved through the
mediated FKBP12 homodimers is more favorable than FKBP12-FRB application of an inducible TEV expression system23 and the
heterodimers. The large free energy of dRap-mediated homo- development of a split TEVp system.13 In the latter case, TEVp
dimerization produces highly stable FKBP12-dRap-FKBP12 com- was divided into N-(amino acids 1–118) and C-(amino acids
plexes that are virtually inert to FRB, until activated by dRap 119–241) terminal fragments and each fragment was fused to FRB
photolysis. This tightly controlled dRap complex formation and FKBP12, respectively. Co-expression of both fusion proteins did
makes it well suited for light-regulation of biological functions. not display any protease activity until the addition of rapamycin,
In order to validate dRap as an effective light-activation which induces dimerization and restores activity of TEVp.
tool, a K-LISA mTOR Activity Assay (Calbiochem), Fig. 2a, was
To investigate the photochemical control of protease activity
performed. mTOR phosphorylates the target protein p70S6K. (Fig. 3a), dRap was tested in HEK293T cells expressing the
However, when mTOR forms a ternary complex with FKBP12 FKBP12-TEV C-terminus and the FRB-TEV N-terminus,13 as well
and rapamycin, it cannot maintain kinase activity. The presence as a circularly permuted luciferase reporter containing a TEVp
of phosphorylation was monitored through an a-phosphorylation- cleavage site (GloSensor, Promega).27 Active TEVp will cleave the
HRP conjugated antibody. As seen in Fig. 2b, phosphorylation GloSensor protein, causing a conformational change that leads
is indicated by high levels of luminescence in the absence to an activation of the luciferase reporter and luminescence. As
of rapamycin as well as in the presence of dRap before UV seen in Fig. 3b, addition of rapamycin to the transfected cells led
irradiation. This is due to the inability of dRap to dimerize to the generation of a luminescence signal, while in the absence
FKBP12 and mTOR, allowing mTOR to retain kinase activity. of rapamycin only a very low background level of luminescence
Phosphorylation is reduced by nearly 80% in the presence of was detected. Importantly, only background luminescence was
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
Genetic engineering in mice is widely applied and has allowed
for the conditional control of genes by a loxP flanked stop
region,32 as well as specific chromosomal rearrangements to
mimic human diseases.33 One of the limitations of Cre is the
inability to tightly control the recombination process with
external regulatory elements. Methodologies have been developed
for the spatiotemporal control of Cre recombinase, including
the regulation of Cre catalysis with a photocaged tamoxifen,34
by caging the catalytic tyrosine residue within the active site,35
and by fusing each fragment of a split Cre enzyme to a blue
light-absorbing photoreceptor and its binding partner.36 However,
there are limitations to each of the current methodologies.
When using photocaged tamoxifen there is need for multiple
irradiations and limited restoration of Cre activity is seen.34
There is a requirement of continuous pulses of light for 24
hours when using the split-Cre fused to a natural photoreceptor,36
and in case of the active-site caged Cre recombinase genetic
encoding in mammalian cells has not been achieved.35
Fig. 3 (a) Photochemical activation of split TEV using dRap. FKBP12-TEV
C-terminus (blue) and FRB-TEV N-terminus (red) dimerize only after
irradiation of dRap and release of native rapamycin. Active TEVp then
cleaves the TEVp recognition site of a circularly permutated luciferase
protein, inducing catalytic activity and thus luminescence. (b) Lumines-
cence data from the split TEVp/GloSensor assay using dRap. HEK293T
cells expressing FKBP12-TEV C-terminus, FRB-TEV N-terminus, and the
GloSensor reporter were treated with rapamycin (100 nM) or dRap (50 nM)
and irradiated for 5 minutes at 365 nm. Luminescence readings were taken
24 hours after irradiation.
Here, we are reporting a solution to these problems by
utilizing a split Cre system composed of amino acids 19–59
observed when the light-cleavable rapamycin dimer dRap was (N-terminus) and amino acids 60-343 (C-terminus) that was
added to the media in the absence of UV irradiation. This fused to FKBP12 and FRB,14 respectively, enabling the photo-
demonstrates the inability of dRap to heterodimerize FRB and chemical control of Cre-catalyzed DNA recombination by dRap
FKBP12. Upon a 5 minute irradiation with UV light of 365 nm, (Fig. 4a). To examine the ability of dRap to photochemically
the luminescence signal is restored to the same level as rapamycin control Cre-catalyzed DNA recombination, HEK293T cells expressing
induction. Thus, irradiation efficiently cleaves the light-sensitive a split Cre system and the Cre Stoplight reporter were used (Fig. 4b).
linker, releasing active rapamycin molecules from dRap. The The Cre Stoplight plasmid expresses dsRed when Cre recom-
released rapamycin stimulates dimerization of FBKP12 and binase is absent or non-functional. In the presence of active
FRB, leading to TEVp activity. Importantly, in addition to being Cre, the gene coding for dsRed is excised from the plasmid due
fully inactive in vitro, the rapamycin dimer is also fully inactive to flanking loxP sites and CMV-driven EGFP expression is
in cell culture and a brief UV irradiation restores TEVp activity activated.37 In the presence of dRap but the absence of UV
to the same level as the addition of rapamycin itself.
irradiation, only dsRed is expressed, an indicator that dRap is
Light-activation of dRap, and thus photochemical control of not capable of heterodimerizing FKBP12 and FRB leading to an
TEVp, allows for non-invasive activation of site-specific protein inactive Cre recombinase (as seen by red fluorescence present
cleavage in live cells. In addition to protein manipulation, in Fig. 5a). This is in perfect agreement with the results
the application of the light-activated rapamycin dimer was also obtained from cells that are not treated with rapamycin. DNA
investigated in the manipulation of DNA, specifically DNA recombinase is active through the decaging of dRap with a brief
recombination (Fig. 4). Cre recombinase is an important tool 5 minute irradiation at 365 nm producing native rapamycin
in cell and developmental biology, as it recognizes palindromic molecules to participate in the FKBP12-Rap-FRB ternary complex
sequences of DNA, known as loxP sites, and is able to delete, and thus complementation of Cre fragments. The reconstituted
insert, or invert any DNA sequence that is located between
those sites, depending on their orientation.14 Cre-catalyzed
DNA recombination has been extensively used for a wide range
of applications starting from simple DNA recombination in
E. coli,28 to genetic engineering in yeast,29 plants30 and mice.31
Fig. 5 Light-activated Cre DNA recombination. Plasmids encoding each
fragment of the split Cre enzyme and the Cre Stoplight reporter were
Fig. 4 Split Cre induction using dRap. (a) FRB-Cre C-terminus (red) and
FKBP12-Cre N-terminus (blue) are dimerized after irradiation of dRap.
(b) The Cre stoplight reporter expresses dsRed before irradiation, but in the
presence of reconstituted, active Cre, the dsRed gene is excised and EGFP
is expressed.
transfected into HEK293T cells. (a) Cells were treated with rapamycin
(5 nM) or dRap (10 nM) and imaged for fluorescence. (b) Micrographs
were quantified by integrating randomly selected areas using Zen
imaging software. Error bars are standard deviations generated from three
independent integrations.
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
8 C. Chou, D. D. Young and A. Deiters, ChemBioChem, 2010, 11,
972–977.
9 A. Prokup, J. Hemphill and A. Deiters, J. Am. Chem. Soc., 2012, 134,
3810–3815.
Cre enzyme excises the dsRed gene and induces EGFP expression
(Fig. 5a). Quantification of the micrographs through fluorescent
cell counting further confirms these results (Fig. 5b).
10 S. Huang, M. Bjornsti and P. Houghton, Cancer Biol. Ther., 2003, 2,
222–232.
11 J. Choi, J. Chen, S. Schreiber and J. Clardy, Science, 1996, 273,
239–242.
12 R. Pollock and T. Clackson, Curr. Opin. Biotechnol., 2002, 13,
459–467.
13 D. J. Williams, H. L. Puhl and S. R. Ikeda, PLoS One, 2009, 4, e7474.
14 N. Jullien, F. Sampieri, A. Enjalbert and J. P. Herman, Nucleic Acids
Res., 2003, 31, e131.
15 S. Zhu, H. Zhang and M. Matunis, Exp. Cell Res., 2006, 312,
1042–1049.
16 S. Ho, S. Biggar, D. Spencer, S. Schreiber and G. Crabtree, Nature,
1996, 382, 822–826.
17 T. Komatsu, I. Kukelyansky, J. McCaffery, T. Ueno, L. Varela and
T. Inoue, Nat. Methods, 2010, 7, U206–U261.
18 M. Pratt, E. Schwartz and T. Muir, Proc. Natl. Acad. Sci. U. S. A., 2007,
104, 11209–11214.
19 M. Putyrski and C. Schultz, Fed. Eur. Biochem. Soc., Lett., 2012, 586,
2097–2105.
20 A. V. Karginov, Y. Zou, D. Shirvanyants, P. Kota, N. V. Dokholyan,
D. D. Young, K. M. Hahn and A. Deiters, J. Am. Chem. Soc., 2011, 133,
420–423.
21 K. Boeckeler, C. Rosse, M. Howell and P. Parker, J. Cell Sci., 2010,
123, 2725–2732.
22 J. Gore, F. Ritort and C. Bustamante, Proc. Natl. Acad. Sci. U. S. A.,
2003, 100, 12564–12569.
23 R. B. Kapust and D. S. Waugh, Protein Express Purifi., 2000, 19,
312–318.
We have successfully developed a light-cleavable rapamycin
dimer (dRap) that does not induce FKBP12 and FRB hetero-
dimerization until it is activated through UV exposure. Previously
reported, caged rapamycin analogues have been applied to photo-
chemically control protein phosphorylation;20,38,39 however, these
molecules have limitations. They require protein engineering of
the FKBP domain or mislocalization of the rapamycin molecule
outside of the cell, since the presence of the synthetic caging
groups alone did not prevent rapamycin-induced FKBP12-FRB
dimerization. Thus, the caged rapamycin analogues could
not be directly interfaced with the wide range of biological
processes that have been placed under conditional control
through the construction of FKBP12-Rap-FRB based biological
switches in cells and organisms. We envisioned developing an
optochemical approach that utilizes recruitment of an entire
protein40 in order to dramatically increase the steric hindrance
imposed by the caging group installed at position C-40 of
rapamycin. Moreover, we planned to keep the design as simple
as possible and to utilize a naturally occurring, endogenous
protein for this purpose. Thus, the light-cleavable rapamycin
dimer dRap was synthesized and we found that dRap was
completely inactive as a heterodimerizer, until irradiated.
Because of the modularity of FRB and FKBP12, dRap has
implications in the light-regulation of protein dimerization
and protein activity in biological processes that have already
been rendered responsive to rapamycin, but could not be
readily placed under optochemical control until now.15–20
We would like to thank Dr Stephen Ikeda (NIH) and Dr Jean-Paul
24 T. Tolbert, D. Franke and C. Wong, Bioorg. Med. Chem., 2005, 13,
909–915.
25 M. Knuesel, Y. Wan, Z. Xiao, E. Holinger, N. Lowe, W. Wang and
X. Liu, Mol. Cell. Proteomics, 2003, 2, 1225–1233.
26 X. Yang, J. Gregan, K. Lindner, H. Young and S. Kearsey, BMC Mol.
Biol., 2005, 6, DOI: 10.1186/1471-2199-6-13.
27 S. S. Wigdal, J. L. Anderson, G. J. Vidugiris, J. Shultz, K. V. Wood and
F. Fan, Curr. Chem. Genomics, 2008, 2, 16–28.
28 D. Adams, J. Bliska and N. Cozzarelli, J. Mol. Biol., 1992, 226,
661–673.
´
Herman (CNRS et Aix-Marseille Universite) for plasmids. This work
29 J. Laplaza, B. Torres, Y. Jin and T. Jeffries, Enzyme Microb. Technol.,
2006, 38, 741–747.
was funded in part by the NSF (CHE-1404836).
30 L. Lyznik, W. Gordon-Kamm and Y. Tao, Plant Cell Rep., 2003, 21,
925–932.
31 A. Nagy, Genesis, 2000, 26, 99–109.
Notes and references
32 M. Lakso, B. Sauer, B. Mosinger, E. Lee, R. Manning, S. Yu,
K. Mulder and H. Westphal, Proc. Natl. Acad. Sci. U. S. A., 1992,
89, 6232–6236.
1 S. Bell and A. Dutta, Annu. Rev. Biochem., 2002, 71, 333–374.
2 R. Csepanyi-Komi, M. Levay and E. Ligeti, Mol. Cell. Endocrinol.,
2012, 353, 10–20.
33 L. van der Weyden and A. Bradley, Annu. Rev. Genomics Hum. Genet.,
2006, 7, 247–276.
3 R. Vale, Cell, 2003, 112, 467–480.
4 W. Merrick, Microbiol. Rev., 1992, 56, 291–315.
34 K. Link, Y. Shi and J. Koh, J. Am. Chem. Soc., 2005, 127, 13088–13089.
35 W. Edwards, D. Young and A. Deiters, ACS Chem. Biol., 2009, 4,
441–445.
5 A. R. Buskirk and D. R. Liu, Chem. Biol., 2005, 12, 151–161;
M. Golynskiy, M. Koay, J. Vinkenborg and M. Merkx, ChemBioChem,
2011, 12, 353–361.
36 M. Kennedy, R. Hughes, L. Peteya, J. Schwartz, M. Ehlers and
C. Tucker, Nat. Methods, 2010, 7, 973–975.
37 Y. S. Yang and T. E. Hughes, Biotechniques, 2001, 31, 1036.
38 N. Umeda, T. Ueno, C. Pohlmeyer, T. Nagano and T. Inoue, J. Am.
Chem. Soc., 2011, 133, 12–14.
39 O. Sadovski, A. Jaikaran, S. Samanta, M. Fabian, R. Dowling,
N. Sonenberg and G. Woolley, Bioorg. Med. Chem., 2010, 18,
7746–7752.
6 L. Fenno, O. Yizhar, K. Deisseroth, S. Hyman, T. Jessell, C. Shatz,
C. Stevens and H. Zoghbi, Annu. Rev. Neurosci., 2011, 34, 389–412;
H. M. Lee, D. R. Larson and D. S. Lawrence, ACS Chem. Biol., 2009, 4,
409–427; G. Mayer and A. Heckel, Angew. Chem., Int. Ed., 2006, 45,
4900–4921; Q. Liu and A. Deiters, Acc. Chem. Res., 2014, 47, 45–55;
A. S. Baker and A. Deiters, ACS Chem. Biol., 2014, 9, 1398–1407;
L. Gardner and A. Deiters, Curr. Opin. Chem. Biol., 2012, 16, 292–299.
7 D. D. Young and A. Deiters, Angew. Chem., Int. Ed., 2007, 46,
4290–4292.
40 K. Harwood and S. Miller, ChemBioChem, 2009, 10, 2855–2857.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015