Destabilizing domains derived from the human estrogen receptor

Article abstract of DOI:10.1021/ja209933r

Methods to rapidly and reversibly perturb the functions of specific proteins are desirable tools for studies of complex biological processes. We have demonstrated an experimental strategy to regulate the intracellular concentration of any protein of inter

Full text of DOI:10.1021/ja209933r

Communication
pubs.acs.org/JACS
Destabilizing Domains Derived from the Human Estrogen Receptor
Yusuke Miyazaki,‡ Hiroshi Imoto,†,‡ Ling-chun Chen, and Thomas J. Wandless*
Department of Chemical and Systems Biology, Stanford University, Stanford, California 94305, United States
S
* Supporting Information
successful both in cells and in mice. However, this approach
ABSTRACT: Methods to rapidly and reversibly perturb
the functions of specific proteins are desirable tools for
studies of complex biological processes. We have
demonstrated an experimental strategy to regulate the
intracellular concentration of any protein of interest by
using an engineered destabilizing protein domain and a
cell-permeable small molecule. Destabilizing domains have
general utility to confer instability to a wide range of
proteins including integral transmembrane proteins. This
study reports a destabilizing domain system based on the
ligand binding domain of the estrogen receptor that can be
regulated by one of two synthetic ligands, CMP8 or 4-
hydroxytamoxifen.
is presently limited to ATPases and GTPases, and thus a more
widely applicable method is desired. A temperature-sensitive
mutant of mammalian dihydrofolate reductase (DHFRts), a
destabilizing residue by the N-end rule, was found to be stable
at a permissive temperature but unstable at 37 °C.4,5 The
addition of methotrexate, a high-affinity ligand for mammalian
DHFR, to cells expressing DHFRts inhibited degradation of the
protein partially. This was an important demonstration that a
small molecule ligand can stabilize a protein otherwise targeted
for degradation in cells. Years later, a rapamycin derivative was
used to stabilize an unstable mutant of the FRB domain of
mTOR (FRB*) and restore the function of the fused kinase,
GSK-3β.6,7 This system demonstrated that ligand-dependent
stability represented an attractive strategy to regulate the
function of a specific protein in a complex biological
environment. The Muir group also developed a system to
control protein activity. In this system, the target protein
becomes functional when the ubiquitin complementation
occurs by rapamycin-induced dimerization of FK506-binding
protein and FKBP12.8
We previously developed a strategy in which a cell-permeable
ligand is used in conjunction with a genetically encoded protein
domain to regulate any protein of interest (Figure 1a).9,10
Mutants of the human FKBP12 or ecDHFR protein were
engineered to be metabolically unstable in the absence of their
high-affinity ligands, Shield-1 or trimethoprim (TMP),
respectively. We call these mutants destabilizing domains
(DDs) and observed that the instability of a DD conferred to
any fused partner protein results in degradation of the entire
fusion protein by the proteasome. Shield-1 and TMP bind to
and stabilize the DD in a dose-dependent manner. The genetic
fusion of the DD to the gene of interest ensures specificity, and
small-molecule control confers reversibility and dose depend-
ence to protein stability and function.
here are a number of perturbation techniques to study
T
gene and protein function in living biological systems.
RNA interference has been used to achieve posttranscriptional
gene silencing.1 This approach is widely used in cultured cells
for elucidating the function of the target protein or gene. In
animal studies the Cre recombinase is commonly used as a
gene perturbation method.2 By using tissue- or cell-specific
promoters, Cre expression can be controlled temporally and
spatially in order to disrupt targeted genes. However, these
approaches are neither reversible nor tunable. Tunable
regulation of protein function is desired to mimic pharmaceut-
ical inhibition for the purpose of target validation. All of the
techniques that target the DNA or mRNA precursor molecules
that encode a protein under study, since existing functional
protein molecules must be degraded, suffer from inevitable
delays following the perturbation.
Small molecule inhibitors are the most effective reagents to
target proteins directly and rapidly. Cell permeable small
molecules remain the most widely used inhibitors or activators
of specific proteins, and most drugs fall into this category of
molecules. For studies of biological processes, small molecules
are valued for their speed, dose-dependent manner, and
reversibility of their activities that provide a useful complement
to genetic techniques. However, the specificity of these agents
for their target proteins is always a concern. Small molecules
may bind not only to the target protein but also to one or more
off-target proteins.
To overcome the trade-off between genetic and chemical
perturbation, Shokat and co-workers developed a method by
which a specific kinase can be inhibited using a small molecule.3
They made mutations in the protein of the interest to modify
the binding pocket for the ligand. Furthermore, a known
chemical inhibitor was also redesigned to fit the modified
binding pocket in target kinase. The method has been
Based on our experiences with previous DDs, we chose the
estrogen receptor ligand binding domain (ERLBD, residues
305−549 of ERS1) as a candidate protein to engineer a novel
destabilizing domain. Since the estrogen receptor signaling
pathway is involved in a variety of diseases, such as breast
cancer, the pathway has been widely studied and numerous
agonist and antagonists of estrogen receptor have been
developed. Thus, compatible pairs of ERLBD and drugs are
known, and ERLBDs are often used as the base for making new
biological tools (Figure 1b).11 Furthermore, previous studies
disclosed ligands that bind to mutant but not wild-type forms of
the ERLBD.12 By using one of these mutant domains encoding
Received: October 21, 2011
Published: February 14, 2012
© 2012 American Chemical Society
3942
dx.doi.org/10.1021/ja209933r | J. Am. Chem. Soc. 2012, 134, 3942−3945

Journal of the American Chemical Society
Communication
CMP8. Mutants displaying low basal fluorescence levels and
also high dynamic range were then sequenced (Figure 2, Table
Figure 1. (a) Strategy for conferring ligand-dependent stability to a
protein of interest (POI) by fusing the DNA sequence encoding DD
to the gene encoding the POI. (b) Structure of wild-type ERLBD and
the ligands, CMP8 and 4OHT.
three mutations (L384M, M421G, G521R),12 we anticipated
that it would be possible to regulate the stability of an ERLBD-
derived DD using a ligand that does not perturb endogenous
estrogen-sensitive networks. An additional mutation (Y537S)
was introduced to further destabilize the ERLBD and to
configure it as a potential DD candidate.13 This tetra mutant
represented a promising starting point for DD development, as
our experiments showed a seven-fold dynamic range regulated
by ligand [Figure S1, Supporting Information (SI)].
Figure 2. NIH 3T3 cells stably transduced with the indicated YFP-
ERLBD fusions derived from error-prone PCR were either mock-
treated or treated with 3 μM CMP8 for 24 h and YFP expression was
monitored by analytical flow cytometry.
S1, SI). From the screen using ERLBD fused to the C-terminus
of YFP, we chose clone no. 50 with the largest dynamic range
(∼80 fold) for further analysis (ER50 with 6 mutations: T371A
L384M M421G N519S G521R Y437S). Treating cells with
various concentrations of CMP8 caused YFP levels to increase
in a dose-dependent manner as measured by analytical flow
cytometry (Figure S3b, SI). DDs fused to the C-terminus of
YFP were fully stabilized by 3 μM CMP8 and display wider
dynamic range relative to the parental sequence (Figure 2). The
enhanced dynamic range is the result of more effective
destabilization in the absence of ligand as well as enhanced
stabilization after binding to the ligand. Fusing these mutants to
the N-terminus of YFP also conferred ligand-dependent
stability (Figure S3a, SI). YFP fluorescence, measured by
analytical flow cytometry, was nearly as low as the
autofluorescence of untransduced NIH 3T3 cells. Ligand for
the wild-type estrogen receptor, estradiol and tamoxifen, were
also tested, but these did not have any affect on DD stability as
predicted (Figure S3c,d, SI).
Next, we investigated the kinetics of DD rescue and
degradation. Upon treatment with CMP8 or 4-hydroxytamox-
ifen (4OHT), cells stably expressing DD-YFP became
fluorescent at a faster rate than former DDs, with YFP levels
reaching nearly 100% in 12 h, whereas previous DDs took twice
as long (Figure 3a). The rate of protein degradation was
measured after removing the stabilizing ligand, and YFP
fluorescence reached basal levels in few hours as anticipated
from studies involving previous DDs (Figure 3b).
To engineer mutants that display drug-dependent stability,
we designed a cell-based screen using yellow fluorescent
protein (YFP) as a reporter for ERLBD stability. The strategy
was designed to identify mutants possessing the desired
characteristics of a destabilizing domain: low fluorescence
signal levels in the absence of ligand, large dynamic range,
robust and predictable dose−response behavior, and rapid
kinetics of degradation. We used error-prone PCR to generate
libraries of ERLBD mutants based on the four parental
mutations (L384M M421G G521R Y537S) sequence.14,15
For the ligand, we synthesized a known compound, CMP8, that
has been reported to bind only to the mutant ERLBD.16 The
library was prepared with the ERLBD mutants cloned in-frame
at the 3′-end of the YFP gene, and a retroviral expression
system was used to stably transduce the library into NIH 3T3
fibroblasts. Fluorescence activated cell sorting (FACS) was
used to screen libraries of candidate DDs. First, we sorted the
cells treated with 5 μM CMP8 for 43 h, at which point YFP-
expressing cells were isolated by FACS. Next, this population
was sorted in the absence of CMP8, where we selected cells
that exhibited low levels of YFP expression. Cells were dosed
with 5 μM CMP8 again, after which point a third sort was
performed to select bright cells (Figure S2, SI). This sorted
population of cells showed a significant increase in YFP signal
compared to parental cells (Figure S2, SI). After three rounds
of sorting, the cells were allowed to recover, their genomic
DNA was extracted, and candidate DDs were amplified by PCR
and isolated.
To gain insight into the cellular mechanism by which these
new DDs were degraded, cells stably transduced with the ER50
fusion proteins were cultured in the presence and absence of
either CMP8 or 4OHT to stabilize the YFP-DD fusion
proteins. Cells were then treated with a proteasome inhibitor
(MG132 or bortezomib) or a lysosome inhibitor (chloroquine).
The stabilizing ligand was then removed, and YFP levels were
Genes encoding individual YFP-ERLBD fusions were stably
transduced back into NIH 3T3 cells, and YFP fluorescence
levels were measured in both the absence and presence of 5 μM
3943
dx.doi.org/10.1021/ja209933r | J. Am. Chem. Soc. 2012, 134, 3942−3945

Journal of the American Chemical Society
Communication
Figure 3. Flow cytometry data of the kinetics of ERLBD-DD. (a) NIH
3T3 cells stably expressing YFP-ER50 fusions were either treated with
3 μM CMP8 or 10 μM 4OHT, and increases in fluorescence were
monitored. (b) Cells were treated with ligand for 24 h, at which point
the cells were washed with media to remove ligand, and decreases in
fluorescence were monitored.
measured after four hours. Cell lysates were resolved by sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-
PAGE) and immunoblotted with an anti-YFP antibody (Figure
S4, SI). CMP8- or 4OHT-treated cells showed strong
expression of the expected fusion proteins. To further elucidate
the pathway of degradation, we collected cells and quantified
the fluorescence of the YFP by analytical flow cytometry
(Figure 4). The data showed that the proteasome inhibitors
MG132 and bortezomib slowed degradation of YFP-DD fusion
protein switching to nonpermissive conditions, indicating that
proteasome is involved in the degradation of the DD. However,
findings also suggest that other pathways may be involved in
the degradation of the DD, since YFP levels dropped 2-fold
even with proteasome inhibitors present. These results are
consistent with those of other DDs.
In addition, to investigate the generality of the new DD
system, the ER50 DD was fused to H-ras as well as the cell
cycle regulatory protein, p21. Cells that stably express these
fusion proteins were cultured in the presence and absence of
CMP8. Cell extracts were resolved by SDS-PAGE and
immunoblotted with either an anti-Hras or anti-p21 antibody
(Figure S5, SI). As anticipated, these fusion proteins exhibit
ligand-dependent stability, indicating that, like the other DD
systems, this DD can be used to regulate a variety of proteins
other than YFP.
Figure 4. NIH 3T3 cells stably expressing YFP-ER50 fusions were
treated with 10 μM 4OHT or 3 μM CMP8 for 24 h. Cells were then
washed with media and treated with 10 μM MG132 (M), 2 μM
bortezomib (B), or 100 μM chloroquine (C) for 4 h. Fluorescence was
monitored using flow cytometry.
In conclusion, we have developed an additional general
method to regulate the stability of specific proteins in a rapid,
reversible, and tunable manner using a small molecule, CMP8
or 4OHT. Mutants of human ERLBD protein were engineered
to be strongly unstable when expressed in mammalian cells in
the absence of a stabilizing ligand. When this new DD is fused
to a protein of interest, its instability is conferred to the protein
of interest, resulting in rapid degradation of the entire fusion
protein. CMP8 or 4OHT are high-affinity ligands for mutant
ERLBD that stabilizes fusion proteins in a dose-dependent
manner, and protein levels in the absence of ligands are
negligible. This ERLBD-derived DD system is orthogonal to
the existing FKBP- and DHFR-based DD systems, in the sense
that the stabilizing ligands do not affect the stability of their
noncognate DDs. This novel DD system will enable researchers
to simultaneously and independently regulate three proteins in
biological studies. For example, the factors for induced
pluripotent stem cells could be interesting targets.17 Another
important advantage of using the ERLBD-derived DD system is
that one of the ligand demonstrated here, 4OHT, is
commercially available, which is significant when using the
DD system in animals. Additionally, this novel DD system
strongly implies that one will be able to develop any new DD
systems by following our method with a high-affinity ligand and
its ligand binding domain.
ASSOCIATED CONTENT
■
S
* Supporting Information
A detailed description of experimental procedures and controls
and complete synthesis of CMP8. This material is available free
of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
wandless@stanford.edu
Present Address
†Takeda Pharmaceutical Company Ltd. 2-26-1, Muraoka-
Higashi, Fujisawa, Kanagawa, 251−8555, Japan.
3944
dx.doi.org/10.1021/ja209933r | J. Am. Chem. Soc. 2012, 134, 3942−3945

Journal of the American Chemical Society
Communication
Author Contributions
‡These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by grants from the NIH (GM068589
and GM073046). Y.M. is a recipient of a fellowship from the
Nakajima Foundation.
REFERENCES
■
(1) Fire, A.; Xu, S.; Montgomery, M. K.; Kostas, S. A.; Driver, S. E.;
Mello, C. C. Nature 1998, 391, 806−811.
(2) Gaver
619−634.
́
iaux-Ruff, C.; Kieffer, B. L. Pharmacol. Ther. 2007, 113,
(3) Shah, K.; Liu, Y.; Deirmengian, C.; Shokat, K. M. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 3565−3570.
́
(4) Levy, F.; Johnston, J. A.; Varshavsky, A. Eur. J. Biochem. 1999,
259, 244−252.
(5) Dohmen, R. J.; Wu, P.; Varshavsky, A. Science 1994, 263, 1273−
1276.
(6) Stankunas, K.; Bayle, J. H.; Gestwicki, J. E.; Lin, Y.-M.; Wandless,
T. J.; Crabtree, G. R. Mol. Cell 2003, 12, 1615−1624.
(7) Liu, K. J.; Arron, J. R.; Stankunas, K.; Crabtree, G. R.; Longaker,
M. T. Nature 2007, 446, 79−82.
(8) Pratt, M. R.; Schwartz, E. C.; Muir, T. W. Proc. Natl. Acad. Sci.
U.S.A. 2007, 104, 11209−11214.
(9) Banaszynski, L. A.; Chen, L.-C.; Maynard-Smith, L. A.; Ooi, A. G.
L.; Wandless, T. J. Cell 2006, 126, 995−1004.
(10) Iwamoto, M.; Bjorklund, T.; Lundberg, C.; Kirik, D.; Wandless,
̈
T. J. Chem. Biol. 2010, 17, 981−988.
(11) Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.; Kushner, P. J.;
Agard, D. A.; Greene, G. L. Cell 1998, 95, 927−937.
(12) Gallinari, P.; Lahm, A.; Koch, U.; Paolini, C.; Nardi, M. C.;
Roscilli, G.; Kinzel, O.; Fattori, D.; Muraglia, E.; Toniatti, C.; Cortese,
R.; De Francesco, R.; Ciliberto, G. Chem. Biol. 2005, 12, 883−893.
(13) Nettles, K. W.; Bruning, J. B.; Gil, G.; Nowak, J.; Sharma, S. K.;
Hahm, J. B.; Kulp, K.; Hochberg, R. B.; Zhou, H.; Katzenellenbogen, J.
A.; Katzenellenbogen, B. S.; Kim, Y.; Joachmiak, A.; Greene, G. L. Nat.
Chem. Biol. 2008, 4, 241−247.
(14) Atwood-Moore, A.; Ejebe, K.; Levin, H. L. J. Virol. 2005, 79,
14863−14875.
(15) Zaccolo, M.; Williams, D. M.; Brown, D. M.; Gherardi, E. J. Mol.
Biol. 1996, 255, 589−603.
(16) Kinzel, O.; Fattori, D.; Muraglia, E.; Gallinari, P.; Nardi, M. C.;
Paolini, C.; Roscilli, G.; Toniatti, C.; Gonzalez Paz, O.; Laufer, R.;
Lahm, A.; Tramontano, A.; Cortese, R.; De Francesco, R.; Ciliberto,
G.; Koch, U. J. Med. Chem. 2006, 49, 5404−5407.
(17) Takahashi, K.; Yamanaka, S. Cell 2006, 126, 663−676.
3945
dx.doi.org/10.1021/ja209933r | J. Am. Chem. Soc. 2012, 134, 3942−3945