Tài liệu Báo cáo khoa học: Improved ecdysone receptor-based inducible gene regulation system doc

Improved ecdysone receptor-based inducible gene regulation system
Subba R. Palli
1
, Mariana Z. Kapitskaya
2
, Mohan B. Kumar
2
and Dean E. Cress
2
1
Department of Entomology, College of Agriculture, University of Kentucky, KY, USA;
2
RHeoGene LLC, Spring House, PA, USA
To develop an ecdysone receptor (EcR)-based inducible
gene regulation system, several constructs were prepared by
fusing DEF domains of Choristoneura fumiferana EcR
(CfEcR), C. fumiferana ultraspiracle (CfUSP), Mus muscu-
lus retinoid X receptor (MmRXR) to either GAL4 DNA
binding domain (DBD) or VP16 activation domain. These
constructs were tested in mammalian cells to evaluate their
ability to transactivate luciferase gene placed under the
control of GAL4 response elements and synthetic TATAA
promoter. A two-hybrid format switch, where GAL4 DBD
was fused to CfEcR (DEF) and VP16 AD was fused to
MmRXR (EF) was found to be the best combination. It had
the lowest background levels of reporter gene activity in the
absence of a ligand and the highest level of reporter gene
activity in the presence of a ligand. Both induction and turn-
off responses were fast. A 16-fold induction was observed
within 3 h of ligand addition and increased to 8942-fold by
48 h after the addition of ligand. Withdrawal of the ligand

resulted in 50% and 80% reduction in reporter gene activity
by 12 h and 24 h, respectively.
Keywords: gene switch; ponasterone A; receptors; EcR;
RXR.
Twenty hydroxyecdysone (20E) is a steroid hormone that
regulates molting, metamorphosis, reproduction and vari-
ous other developmental processes in insects. Ecdysone
functions through a heterodimeric receptor complex com-
prised of ecdysone receptor (EcR) and ultraspiracle (USP).
Both EcR and USP cDNAs have been cloned from
Drosophila melanogaster and several other insects [1] and
were shown to be members of the steroid hormone receptor
superfamily. Members of this superfamily are characterized
by the presence of five modular domains, A/B (transacti-
vation), C (DNA binding/heterodimerization), D (hinge,
heterodimerization), E (ligand binding, heterodimerization,
transactivation) and F (transactivation). Crystallographic
studies on the E domain structures of several nuclear
receptors showed a conserved fold composed of 11 helices
(H1 and H3–H12) and two short strands (s1 and s2) [2].
Recently, the crystal structure of USP was solved by two
groups [3,4], both structures showed a long H1-H3 loop and
an insert between H5 and H6. These structures appear to
lock USP in an inactive conformation by displacing helix 12
from agonist conformation. In both crystal structures USP
had a large hydrophobic cavity, which contained phos-
pholipid ligands. The crystal structure of the EcR has yet
to be determined; however, homology models for CtEcR
(Chironomus tentans EcR) [5], and CfEcR (Choristoneura
fumiferana EcR) [6] have been generated [7,8].

Ecdysone receptors are found in insects and other related
invertebrates [1]. Ecdysteroids and related compounds have
been identified in plants, insects and other related inverte-
brates. EcR and its ligands are not detected in vertebrates
such as humans, therefore they are very good candidates for
developing gene regulation systems for use in vertebrates.
Insect EcR can heterodimerize with retinoid X receptor
(RXR) and transactivate genes that are placed under the
control of ecdysone response elements (EcRE) in various
cellular backgrounds including mammalian cells. The EcR-
based gene switch is being developed for use in various
applications including gene therapy, expression of toxic
proteins in cell lines as well as for cell-based drug discovery
assays [9–17].
After initial reports [18,19] on the function of EcR as an
ecdysteroid dependent transcription factor in cultured
mammalian cells, No et al. [20] used D. melanogaster EcR
(DmEcR) and human RXRa to develop an ecdysone
inducible gene expression system that can function in
mammalian cells and mice. Later, Suhr et al. [21] showed
that the nonsteroidal ecdysone agonist, tebufenozide,
induced high level of transactivation of reporter genes in
mammalian cells through Bombyx mori EcR (BmEcR) [22]
and endogenous RXR. Hoppe et al. [23] combined DmEcR
and BmEcR systems and created a chimeric Drosophila/
Bombyx EcR (DBEcR) that had combined positive aspects
of both systems, i.e. the chimeric receptor bound to
modified ecdysone response elements and functioned
without exogenous RXR. Recent improvements to the
EcR-based gene switch include expression of both EcR

and RXR in a bicistronic vector [24] and the discovery that
the RXR ligands enhance the ligand dependent activity of
the EcR-based gene switch [25].
An optimal gene regulation system should have the
following characteristics: (a) low or no basal expression in
the absence of an inducer (b) high induced expression in the
presence of a wide range of inducer concentration (c) rapid
Correspondence to S. R. Palli, Department of Entomology, College
of Agriculture, University of Kentucky, Lexington KY 40546.
Fax: + 1 859 3231120, Tel.: +1 859 2574962,
E-mail:
Abbreviations: 20E, twenty hydroxyecdysone; EcR, ecdysone receptor;
LBD, ligand binding domain; RXR, retinoid X receptor;
USP, ultrapiracle.
(Received 11 December 2002, revised 21 January 2003,
accepted 5 February 2003)
Eur. J. Biochem. 270, 1308–1315 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03501.x
induction response after addition of an inducer (d) rapid
switch-off response after removal of an inducer (e) repeated
on and off responses (f) specific response to inducer with no
pleiotropic effects and (g) the length of DNA constructs
should be smaller for convenient packaging into viruses for
in-vivo delivery.
The current versions of EcR-based gene switches do not
have some of the desirable characteristics of an optimal
gene regulation system described above. For example, the
background and induced levels of reporter gene activity are
higher and lower, respectively, resulting in lower-fold
induction. There was also no systematic analysis performed
to find the optimum length of EcR and RXR required for

the best performance of the switch. The studies presented
here are designed to overcome some of the drawbacks
associated with the EcR-based gene switch. We have
constructed several gene switch plasmids by fusing C. fumi-
ferana EcR (CfEcR) [6,26], C. fumiferana ultraspiracle
(CfUSP) [27], Mus musculus RXR (MmRXR) [28], to
either GAL4 DNA binding domain or VP16 activation
domain. Combinations of these receptor constructs were
analyzed in several mammalian cell lines using transient
transactivation assays and selected a switch that had very
low basal expression in the absence of a ligand and high-
induced expression in the presence of a ligand. Both
induction and turn-off of reporter gene in response to
addition and withdrawal of ligand, respectively, were rapid.
This CfEcR-based switch also showed differential sensiti-
vity to steroids and nonsteroidal ligands. We have also
identified that DEF domains of CfEcR and EF domains of
MmRXR are required for the best performance of this
gene switch.
Materials and methods
Constructs
VgRXR (DmVgRXR) receptor plasmid and pIND
b-galactosidase reporter plasmid were purchased from
Invitrogen (Invitrogen Corporation, Carlsbad, CA, USA).
This switch contains two receptors, DmEcR(CDEF) fused
to VP16 activation domain [(V:DmE(CDEF)] and
expressed under CMV promoter and full-length HsRXRa
[HsR (A/BCDEF)] expressed under RSV promoter. In
addition, the P-box in the DmEcR C (DNA binding)
domain was altered to resemble that of glucocorticoid

receptor (GR) recognizing a hybrid of EcR and GR
response elements (E/GRE). These two receptors hetero-
dimerize and bind to ligand and regulate b-galactosidase
reporter placed under the control of 4X E/GRE-DMTV
promoter (pIND b-galactosidase). CfVgRXR plasmid
was constructed by replacing CDEF domains of DmEcR
of DmVgRXR with CDEF domains of CfEcR. First
the CfEcR fragment containing DNA binding domain
C-terminal to the P-box and complete DEF domains
was amplified using primers containing BamHI and
BstXI sites on 5¢ and 3¢ ends, respectively. The PCR
product was cloned into BamHI and BstXI digested
DmVgRXR. CfVgRXRDel plasmid was constructed by
removing HsRXRa from CfVgRXR. The CfVgRXR DNA
was digested with EcoRV and NotI and the recessed 3¢ ends
of NotI-digested fragments were filled using Klenow
fragment of DNA polymerase I, ligated and transformed
into E. coli.
GAL4:CfEcR(DEF) [G:CfE(DEF)] and various trunca-
tions of CfEcR were constructed by amplifying defined
regions of CfEcR (NCBI accession number AAC36491)
using primers containing a BamHI or EcoRI site on the
5¢ end and a XbaIsiteonthe3¢ end. The PCR products were
cloned into BamHI/EcoRI and XbaI sites of pM vector
(Clontech Inc. Palo Alto, CA, USA). VP16: CfEcR (CDEF)
[V:CfE(CDEF)] and VP16:CfEcR(DEF) [V:CfE(DEF)]
were constructed by transferring BamHI and XbaI frag-
ments from G:CfE(CDEF) and G:CfE(DEF) to BamHI
and XbaI digested pVP16 vector (Clontech Inc. Palo Alto,
CA, USA). VP16:MmRXR (DEF) [V:MmR(DEF)] and

various truncations of MmRXRa were constructed by
amplifying defined regions of MmRXR (NCBI accession
number NP035435) with primers containing EcoRI site on
the 5¢ end and XbaIsiteonthe3¢ end. The PCR products
were cloned into EcoRI and XbaI digested pVP16 vector.
GAL4:MmRXR(DEF) [G:MmR(DEF)] was constructed
by ligating EcoRI and XbaI digested fragment of MmRXR
from V:MmR(DEF) to EcoRI and XbaIdigestedpM
vector. V:MmR[DEF(H 4–12] was constructed by deleting
BamHI fragment containing helices (H) 1–3 from
V:MmR(EF). V:MmR(EF) was digested with BamHI and
the larger fragment containing vector plus helices 4–12 of
MmRXR was isolated, ligated and transformed.
VP16:CfUSP(DEF) [V:CfU(DEF)] was constructed by
amplifying DEF domains of CfUSP (NCBI accession
number AAC31795) using primers containing EcoRI and
BamHI sites in the forward and reverse primer, respectively,
followed by cloning of the PCR product into EcoRI and
BamHI digested pVP16 and pM vectors. pFRLUC reporter
plasmid (5· GAL4 response element was fused to synthetic
GGGTATATAAT sequence) was purchased from Strata-
gene Cloning Systems (La Jolla, CA, USA). pFRLUCE-
cRE was constructed by replacing 5· GAL4 response
elements of pFRLUC with 7X EcRE. The 7X EcRE
fragment was amplified from pMK43.2 [29] using primers
containing PstIandXmaI sites in forward and reverse
primers, respectively. The TATAA sequence was included
in the reverse primer. The PCR products were cloned into
PstIandXmaI digested pFRLUC. The pINDSEAP
reporter vector was constructed by replacing b-galactosidase

gene of pINDLaZ with SEAP from pSEAP2-basic vector
(Clontech Inc. Palo alto, CA, USA) at HindIII and XbaI
restriction enzyme sites. The pFRSEAP reporter vector was
constructed by replacing luciferase of pFRLUC with SEAP
at KpnIandXbaI restriction sites.
Ligands
Ponasterone A and Muristerone A were purchased from
Alexis Corporation (San Diego, CA, USA). RG-102240
also known as GS
TM
-E [N-(1,1-dimethylethyl)-N¢-(2-ethyl-
3-methoxybenzoyl)-3,5-dimethylbenzohydrazide] and
RG-102317 [N-(1,1-dimethylethyl)-N¢-(5-methyl-2,3-dihydro-
benzo-1,4-dioxine-6-carbonyl)-3,5-dimethylbenzohydrazide]
are synthetic stable bisacylhydrazine ecdysone agonists
synthesized at Rohm and Haas Company. All ligands were
supplied in dimethylsulfoxide and the final concentration of
dimethylsulfoxide was maintained at 0.1%.
Ó FEBS 2003 EcR-based gene switch (Eur. J. Biochem. 270) 1309
Cells and transfections and reporter assays
CHO and A549 cells were grown in F12 medium containing
2m
ML
-glutamine and 10% bovine calf serum. 3T3, 293 and
CV1 cells were grown in Dulbecco’s modified Eagle’s
medium containing 4 m
ML
-glutamine, 1.5 gÆL
)1
sodium

bicarbonate, 4.5 gÆL
)1
glucose and 10% bovine calf serum.
All media and serum were purchased from Life Techno-
logies, Rockville, MR, USA. One hundred thousand CHO
or 293 cells or 50 000 of 3T3 or CV1 or A549 cells were
plated per well of 12-well plates. The following day the cells
were transfected with 0.25 lg of receptor(s) and 1.0 lgof
reporter constructs using 4 lL of LipofectAMINE 2000
(Life Technologies, Rockville, MR, USA) for CHO and 293
cells, LipofectAMINE plus (Life Technologies, Rockville,
MR, USA) for CV1 cells or SuperFect (Qiagen Inc.
Valencia, CA, USA) for 3T3 cells or A549 cells. After
transfection, the cells were grown in the medium containing
ligands for 24–48 h. A second reporter, Renilla luciferase
(0.1 lg), expressed under a thymidine kinase constitutive
promoter was cotransfected into cells and was used for
normalization. The cells were harvested, lyzed and the
reporter activity was measured in an aliquot of lyzate. All
transfection experiments were performed in triplicate and
the experiments were repeated at least three times.
Luciferase was measured using Dual-luciferase
TM
repor-
ter assay system from Promega Corporation (Madison, WI,
USA). b-Galactosidase was measured using Gal-ScreenÒ
system from Applied Biosystems (Foster City, CA, USA).
The SEAP activity in the medium was quantified using
Phospha-Light
TM

System from Applied Biosystems.
Results
The fold inductions were lower for VgRXR-based
switch formats
We first tested DmVgRXR, CfVgRXR and CfVgRXRDel
switches (Fig. 1A) for their ability to transactivate pIND-
b-galactosidase reporter gene in 3T3 cells. CfVgRXR,
DmVgRXR and CfVgRXRDel switches showed dose
dependent induction of reporter gene activity upon addition
of RG-102240 and supported maximum induced levels of
26-fold, 21-fold and sixfold, respectively (Fig. 1B). The
lower fold inductions observed were mainly due to high
background levels of reporter gene activity in the absence
of ligand. The CfVgRXRDel switch, where no exogenous
RXR was added, showed both higher background levels
and lower induced levels of reporter gene activity and as a
result the fold induction was lower for this switch when
compared to DmVgRXR and CfVgRXR switches. Similar
results were also observed in CHO, 293 and CV1 cells
(data not shown). In all these cell lines, a maximum of
100-fold induction and an average of 25-fold induction were
observed for CfVgRXR and DmVgRXR switches.
CfVgRXR and DmVgRXR switches performed better than
the CfVgRXRDel switch in all four cell lines tested.
Two-hybrid switch formats showed high fold induction
In order to develop a switch that has lower background
and higher induced levels, we prepared receptor constructs
where DEF domains of CfEcR, CfUSP and MmRXR were
fused to either GAL4 DNA binding domain or VP16
activation domain. Different combinations of a GAL4

fusion receptor, a VP16 fusion receptor (Fig. 2A) and
pFRLUC reporter were tested in 3T3 cells. Out of the four
combinations tested, the G:CfE(DEF) + V:MmR(DEF)
switch showed the highest level of induction (1014-fold;
Fig. 2B). The reporter gene induction was RG-102240 dose-
dependent and significant levels of reporter gene induction
were observed at 1 l
M
or higher concentration of ligand.
The G:MmR(DEF) + V:CfE(DEF) switch format also
showed ligand-dependent induction of reporter gene acti-
vity, but the induction was lower at 80-fold when compared
to 1014-fold observed for the G:CfE(DEF) + V:MmR
(DEF) switch. Use of CfUSP in place of MmRXR resulted
in high background levels of reporter gene activity in the
absence of ligand, as a result the induction was only twofold
(Fig. 2B). These four switches performed in a similar way in
CHO, CV1, 293 and A549 cells (data not shown).
Fig. 1. Induction of b-galactosidase reporter gene by CfVgRXR,
DmVgRXR and CfVgRXRdel switches. (A) Schematic diagram of
constructs used in the experiment. (B) Plasmid DNA samples of
CfVgRXR or DmVgRXR or CfVgRXRDel and pINDLacZ were
transfected into 3T3 cells using Superfect (Qiagen Inc., Valencia, CA)
lipid reagent. The transfected cells were grown in the medium con-
taining0,0.1,1,5,10and50m
M
concentration of RG-102240. The
cells were harvested at 48 h after adding ligand and the reporter
activity was measured using the Gal-ScreenÒ system (Applied Bio-
systems. Total relative light units (RLU) presented are mean ± SD

(n ¼ 3). Numbers above the bars show the maximum fold induction
observed for that particular combination. Fold induction was calcu-
lated by dividing total RLUs in the presence of ligand by total RLUs in
the absence of ligand.
1310 S. R. Palli et al.(Eur. J. Biochem. 270) Ó FEBS 2003
The G:CfE(DEF) + V:MmR(DEF) switch performs better
through nonsteroidal ligands when compared
to steroids
We tested dose–response of two nonsteroids (RG-102240
and RG-102317) and two steroids (PonA and MurA) for
the G:CfE(DEF) + V:MmR(DEF) switch. This switch
induced the luciferase gene expression at 1 l
M
or higher
concentration of RG-102240, 0.04 l
M
or higher concentra-
tion of RG-102317, 5 l
M
or higher concentration of PonA
and 25 l
M
or higher concentration of MurA (Fig. 3A). The
G:CfE(DEF) + V:MmR(DEF) switch seems to be more
sensitive to nonsteroidal ligands when compared to steroids.
Similar differential sensitivity between nonsteroidal ligands
and steroids was also observed in CHO, 293 and CV1 cells
(data not shown). To determine whether this difference in
ligand sensitivity is due to the two-hybrid switch format or
due to CfEcR itself, we have evaluated the dose–response of

RG-102240 and PonA to the V:CfE(CDEF) switch. In this
switch format, only V:CfE(CDEF) and pFRLUCEcRE
reporter were transfected and V:CfE(CDEF) heterodimer-
izes with endogenous RXR. As shown in Fig. 3B, the
V:CfE(CDEF) switch induced the reporter gene activity by
45-fold in the presence of 5 l
M
concentration of RG-102240
and by threefold in the presence of 5 l
M
concentration of
PonA. These results suggest that CfEcR is the most likely
contributor to the differences in dose–response of non-
steroidal ligands and steroids.
Truncation analysis of MmRXR
The optimum fragment of RXR required for a two-hybrid
switch was identified by preparing VP16 activation domain
fusions of MmRXR A/BCEDF, CDEF, DEF, EF, DEF
(H4–12), DEF (H1–11) (Fig. 4A) and analyzing them in 3T3
cells in combination with G:CfE(CDEF) or G:CfE(DEF)
and pFRLUC. As shown in Fig. 4(B), the V:MmR(EF) +
G:CfE(CDEF) combination showed the highest fold induc-
tion (13 881). Deleting the first three helices or the 12th helix
of V:MmR(EF) reduced its activity significantly. A similar
pattern was observed when G:CfE(DEF) was used as a
partner for MmRXR truncations. Out of all truncations
tested, V:MmR(EF) was the best partner for G:CfE(DEF)
Fig. 3. The G:CfE(DEF) + V:MmR(EF) switch works better through
nonsteroidal ligands than steroids. (A) Dose–response of the two-hybrid
switch to four ligands. 3T3 cells were transfected with G:CfE(DEF),

V:MmR(EF), pFRLUC and pTKRL. The transfected cells were
grown in the medium containing 0, 0.04, 0.2, 1, 5 and 25 l
M
concen-
tration of RG-102240 or RG-102317 or PonA or MurA. (B) Dose–
response of V:CfE(CDEF) switch to two ligands. 3T3 cells were
transfected with V:CfE(CDEF), pFRLUCEcRE and pTKRL. The
transfected cells were grown in the medium containing 0, 0.04, 0.2, 1. 5
and 25 l
M
concentration of RG-102240 or PonA.
Fig. 2. Induction of luciferase reporter gene by two-hybrid switches. (A)
Schematic diagram of constructs used in the experiment. (B) Plasmid
DNA samples of pFRLUC, pTKRL and G:CfE(DEF) + V:MmR
(DEF) or G:CfE(DEF) + V:CfU(DEF) or G:MmR(DEF) + V:CfE
(DEF) or G:CfU(DEF) + V:CfE(DEF) were transfected into 3T3
cells using Superfect lipid reagent. The transfected cells were grown in
the medium containing 0, 0.1, 1, 5, 10 and 50 m
M
concentration of
RG-102240. The cells were harvested at 48 h after adding ligand and
the reporter activity was measured using a dual luciferase assay kit
from Promega Corporation (Madison, WI, USA). Total relative light
units (RLU) presented are mean ± SD (n ¼ 3). Numbers above the
bars show the maximum fold induction observed for that particular
combination.
Ó FEBS 2003 EcR-based gene switch (Eur. J. Biochem. 270) 1311
(Fig. 4C). In this case also deleting the first three helices or
the 12th helix of V:MmR(EF) reduced the performance of
the switch significantly.

Truncation analysis of CfEcR
To identify the optimum fragment of CfEcR required for
the best performance of the two-hybrid switch, we con-
structed GAL4 DNA binding domain fusions of CfEcR A/
BCDEF, CDEF, 1/2CDEF (half of the DNA binding
domain containing second zinc finger was included), DEF,
EF and DE(H1-11) domains and assayed them in 3T3 cells
in the presence of V:MmR(EF) and pFRLUC. Among the
truncations tested (Fig. 5A), G:CfE(CDEF) + V:MmR
(EF) showed the highest fold induction (Fig. 5B). The
G:CfE(DEF) + V:MmR(EF) was the most sensitive com-
bination (Fig. 5B). Deleting the D domain or the 12th helix
and F domain reduced the activity of receptor gene
significantly. Thus, the CfE(DEF) truncation showed the
maximum ligand sensitivity and the CfE(CDEF) truncation
showed the maximum induction.
Rapid induction and turn off of reporter gene activity
through the G:CfE(DEF) + V:MmR(EF) switch
The best two-hybrid switch combination, G:CfE(DEF) +
V:MmR(EF), was used to study the time-course of induc-
tion and subsequent decline of reporter gene activity in 3T3
cells. Increase in reporter gene activity was observed one
hour after adding ligand and the reporter activity increased
steadily until 72 h after the addition of ligand (Fig. 6A). To
study the time-course of decrease in reporter gene activity
after withdrawal of ligand, G:CfE(DEF) + V:MmR(EF)
and pFRLUC were transfected into 3T3 cells and the cells
were grown in the presence of 1 l
M
RG-102240 for 24 h.

Then the cells were washed with medium containing no
ligand and were grown in the same medium for 72 h. As
shown in Fig. 6B a 50% and 80% decrease in reporter gene
activity was observed by 12 h and 24 h, respectively, after
withdrawal of ligand. Thus, both the induction and decline
of reporter gene activity are rapid for this switch.
Comparison of the performance of the
G:CfE(DEF) + V:MmR(EF) switch with other EcR-based
switches
To compare the performance of the two-hybrid switch
developed with previous versions of EcR-based gene
Fig. 4. Truncation analysis on MmRXR. (A) Truncations of MmRXR
tested. The numbers above the horizontal bars indicate amino acid
boundaries between domains of RXR. Ligand-binding domain and
locations of helices were identified based on Egea et al.[37].(B)VP16
fusions of six MmRXR truncations were transfected into 3T3 cells
along with G:CfE(CDEF), pFRLUC and pTKRL. The transfected
cells were grown in the medium containing 0, 1, 5 and 25 m
M
RG-
120240. The cells were harvested at 48 h after adding ligands and the
reporter activity was quantified. The numbers shown above the zero
for each panel represent the mean RLUs observed in DMSO-treated
cells. (C) Same as in B except G:CfE(DEF) was used in place of
G:CfE(CDEF).
Fig. 5. Truncation analysis on CfEcR. (A) Truncations of CfEcR tes-
ted. The numbers on the top of horizontal bars indicate amino acid
boundaries between domains of EcR. Helices within the LBD were
identified based on CtEcR [7] and CfEcR [8] homology models. (B)
GAL4 fusions of six CfEcR truncations were transfected into 3T3 cells

along with V:MmR(EF), pFRLUC and pTKRL. The transfected cells
were grown in the medium containing 0, 1, 5 and 25 m
M
RG-102240.
The cells were harvested at 48 h after adding ligands and the reporter
activity was quantified.
1312 S. R. Palli et al.(Eur. J. Biochem. 270) Ó FEBS 2003
switches, we modified previous versions of EcR-based
switches so that our comparisons are carried out with the
same receptor (CfEcR) and reporter (SEAP). As shown in
Table 1, the G:CfE(DEF) + V:HsR(EF) version of the
switch being reported here performed better than the two
previous versions of switches. The fold induction with this
new switch is higher when compared to the fold inductions
observed for CfVgRXR and CfVgRXRdel versions of
switches. The lower fold induction in the case of earlier
versions of switches is mainly due to the higher background
levels of reporter activity in the absence of ligand.
Discussion
The most significant contribution of this study is the
development of an EcR-based gene switch that has over-
come most of the drawbacks associated with the earlier
versions [18,20,21,23]. This two-hybrid format EcR-based
gene switch showed the lowest levels of background reporter
gene activity in the absence of ligand and the highest levels
of induced reporter gene activity in the presence of ligand,
resulting in a strikingly high fold induction. There are three
major differences between the two-hybrid switch developed
in this study and the previous versions of EcR-based
switches [18,20,21,23]. First, in the two-hybrid switch, we

used the heterologous GAL4 DNA binding domain in place
of the EcR DNA binding domain or its modified form used
in the previous EcR-based switches. Second, DNA binding
and activation domains were placed on two different
proteins instead of on a single protein as carried out for
previous versions of the EcR-based switch. Third, in the
reporter construct, we have used a synthetic TATAA
element in the place of minimal promoters used in the
previous versions of EcR-based switches.
We have constructed some switches where GAL4 DNA
binding and VP16 activation and EcR ligand binding
domains were placed in the same molecule. These switches
in combination with pFRLUC showed higher background
reporter activity in the absence of a ligand and as a result the
fold induction was lower (data not shown). These results
indicate that changing GAL4 DNA binding domain alone
would not have significantly improved the performance of
the EcR-based switch. The V:CfE(CDEF) switch that used
EcRE and synthetic TATAA showed lower fold induction
(maximum 45-fold; Fig. 3B) when compared to the
G:CfE(DEF) + V:MmR(EF) switch that used GALRE
synthetic TATAA (maximum 1014-fold; Fig. 2B), indica-
ting that the use of synthetic TATAA in the reporter
Fig. 6. Time-course of induction (A) and turn-off (B) of two-hybrid
switch. 3T3 cells were transfected with G:Cf(DEF), V:MmR(EF),
pFRLUC and pTKRL. The transfected cells were grown in the
medium containing 1 l
M
concentration of RG-102240. For the data
showninA,cellswerecollectedat0,1,3,6,12,24,48and72hafter

adding ligand. The reporter activity was quantified and plotted. For
the data shown in B and 24 hours after adding ligand, the cells were
washed with ligand-free medium, grown in the medium containing
dimethylsulfoxide for 0, 1, 3, 6, 12, 24, 48 and 72 h, then harvested and
reporter activity was quantified and plotted.
Table 1. Comparison of performance of CfVgRXR, CfVgRXRdel and G:CfE(DEF) + V:HsR(EF) switches. Plasmid DNAs of pINDSEAP and
CfVgRXR or CfVgRXRdel, pFRSEAP and G:CfE(DEF) + V:HsR(EF) constructs were transfected into 3T3 cells plated into 96-well plates. After
transfection, the cells were exposed to 0, 0.2, 1 and 5 l
M
RG-102240 and 1, 5 and 25 l
M
PonA for 48 h. The SEAP activity in the medium was
quantified using the Phospha-Light
TM
System from Applied Biosystems. FI, fold induction.
Ligand
CfVgRXR CfVgRXRdel G:Cf(DEF) + V:HsR(EF)
RLU ± SD FI ± SD RLU ± SD FI ± SD RLU ± SD FI ± SD
Dimethylsulfoxide 55 ± 10 1 59 ± 2 1 9 ± 1 1
RG-102240 0.2 l
M
108 ± 10 2 53 ± 22 1 9 ± 1 1
RG-102240 1 l
M
1230 ± 89 29 ± 9 136 ± 24 2 525 ± 164 62 ± 10
RG-102240 5 l
M
1478 ± 249 41 ± 21 355 ± 106 6 ± 2 2356 ± 73 288 ± 48
RG-102240 25 l
M

2572 ± 470 47 ± 2 713 ± 138 12 ± 3 2582 ± 149 316 ± 49
PonA 1 l
M
253±73 5±1 60±16 1 76±69 8±6
PonA 5 l
M
185 ± 16 3 ± 1 70 ± 12 1 194 ± 90 23 ± 6
PonA 25 lm 771 ± 86 14 ± 1 77 ± 12 1 2540 ± 187 313 ± 70
Ó FEBS 2003 EcR-based gene switch (Eur. J. Biochem. 270) 1313
construct alone would not have improved the performance
of the EcR-based switch to the extent observed in these
studies.
We have also tested switch formats using either
CfE(A/BCDEF) or CfE(CDEF) in combination with
V:MmR(A/BCDEF) or V:MmR(CDEF) or V:MmR(DEF)
or V:MmR(EF) and pFRLUCEcRE. None of these switch
formats supported the high fold inductions observed for
two-hybrid format switches (data not shown) indicating
that merely separating the DNA binding domain and the
activation domain onto two molecules is not sufficient to
improve the performance of this switch. It appears that a
combination of several different factors contributed to the
dramatic improvement in the performance of this two-
hybrid switch.
The mechanism of action of this two-hybrid format
switch is not entirely understood. Truncation analyses
showed that the helix 12 of both CfEcR and MmRXR are
required for efficient transactivation. Deletion of either of
these domains resulted in drastic reduction in transactiva-
tion of reporter genes through these receptors indicating

that C-terminal activation domains of these receptors are
involved in transactivation through this switch. Previous
studies showed that neither ligand nor F domain of CfEcR
is required for heterodimerization of CfEcR with CfUSP
[26,30]. Taken together, these studies indicate that this new
two-hybrid format switch functions through heterodimeri-
zation and ligand binding followed by conformational
change in both receptors resulting in transactivation of
genes placed under the control of this switch.
It is interesting that RXR-based two-hybrid switches are
highly inducible because of their low background reporter
activity in the absence of ligand. On the other hand, USP-
based switches are not inducible mainly because of high
levels of reporter activity in the absence of ligand. We have
observed similar results in yeast, where expression of full-
length C. fumiferana EcR and USP led to transactivation of
reporter gene in the absence of a ligand [31]. Deletion of A/B
domains from both EcR and USP abolished the constitutive
activation in this assay. Recently, Lezzi et al. [32] reported
ligand-induced heterodimerization between the ligand bind-
ing domains of D. melanogaster EcR and USP in yeast. The
differences in performance of RXR- and USP-based
switches are most likely due to the differences in the
requirement of ligand for formation of heterodimers with
CfEcR. Previous studies showed that CfEcR and CfUSP
[26,30], BmEcR and D. melanogaster USP (DmUSP) can
form heterodimers in the absence of ligand [21], whereas
both DmEcR and BmEcR require the presence of ligand for
formation of heterodimers with RXR.
We observed differential sensitivity of the CfEcR-based

two-hybrid switch to steroids and nonsteroidal ligands.
Dose–response studies using two steroids (PonA and
MurA) and two nonsteroids (RG-102240 and RG-
102317) in four cells lines (3T3, CHO, 293 and CV1)
showed that the CfEcR-based two-hybrid switch is more
sensitive to nonsteroidal ligands when compared to steroids.
Previous studies also showed similar differences in binding
of steroid and nonsteroidal ligands to CfEcR and CfUSP.
RH-5992 and RH-2485 (bisacylhydrazines) bound to
CfEcR and CfUSP at 10-fold higher affinity than the
steroids, PonA and MurA [33]. Earlier versions of EcR-
based gene switches also showed higher activity with
nonsteroidal ligands when compared to the activity with
steroids [23,25,34].
Previous published versions of EcR-based gene switches
used CDEF domains of EcR and full-length RXR. In this
study, we performed a systematic analysis and identified
regions of both EcR and RXR required for optimum
performance. The two-hybrid version of the CfEcR-based
gene switch uses only 1072 nucleotides of CfEcR and 725
nucleotides of MmRXR when compared to 1973 nucleo-
tides of DmEcR and 1388 nucleotides of RXR used in the
commercially available version of EcR-based gene switch
(Invitrogen Corporation, Carlsbad, CA, USA). The size
of receptors used in gene switches are very important
due to size limitations in packaging gene regulation and
target gene constructs into various viruses for in vivo
delivery.
In transactivation assays, the CfEcR-based two-hybrid
format switch showed very low reporter gene activity in the

absence of ligand and high reporter gene activity in the
presence of ligand. Both induction and switch-off responses
were rapid. Recently, our collaborators confirmed the
performance of this switch in stable cell lines as well as in
mouse tumors [35]. Bisacylhydrazine nonsteroidal ligands
have undergone an extensive battery of toxicology tests for
EPA registration as commercial insecticides [33]. These
chemicals are classified as green chemistry by EPA because
of their favorable environmental and toxicology profiles
[36]. Thus, this new CfEcR-based switch has most of the
desirable properties of an optimal gene regulation system
and is currently being evaluated for in vivo efficacy. One
limitation of the current version of this switch is the
requirement of slightly higher concentration of ligands for
maximum induction. Experiments are in progress to
improve the sensitivity of this switch by modifying both
EcR and RXR.
Acknowledgements
We thank M.Padidam, D.W.Potter, P.White and P.Kumar for
critical reading of the manuscript and M. R. Koelle of Stanford
University for the gift of pMK43.2 reporter vector.
References
1. Riddiford, L.M., Cherbas, P. & Truman, J.W. (2000) Ecdysone
receptors and their biological actions. Vitam. Horm. 60, 1–73.
2.Bourguet,W.,Ruff,M.,Chambon,P.,Gronemeyer,H.&
Moras, D. (1995) Crystal-structure of the ligand-binding
domain of the human nuclear receptor RXR-alpha. Nature 375,
377–382.
3. Billas, I.M., Moulinier, L., Rochel, N. & Moras, D. (2001) Crystal
structure of the ligand-binding domain of the ultraspiracle protein

USP, the ortholog of retinoid X receptors in insects. J. Biol. Chem.
276, 7465–7474.
4. Clayton, G.M., Peak-Chew, S.Y., Evans, R.M. & Schwabe, J.W.
(2001) The structure of the ultraspiracle ligand-binding domain
reveals a nuclear receptor locked in an inactive conformation.
Proc.NatlAcad.Sci.USA98, 1549–1554.
5. Imhof, M.O., Rusconi, S. & Lezzi, M. (1993) Cloning of a
Chironomus tentans cDNA encoding a protein (cEcRH) homo-
logous to the Drosophila melanogaster ecdysteroid receptor
(dEcR). Insect Biochem. Mol. Biol. 23, 115–124.
1314 S. R. Palli et al.(Eur. J. Biochem. 270) Ó FEBS 2003
6. Kothapalli, R., Palli, S.R., Ladd, T.R., Sohi, S.S., Cress, D.,
Dhadialla, T.S., Tzertzinis, G. & Retnakaran, A. (1995) Cloning
and developmental expression of the ecdysone receptor gene from
the spruce budworm, Choristoneura fumiferana. Dev. Genet. 17,
319–330.
7. Wurtz, J.M., Guillot, B., Fagart, J., Moras, D., Tietjen, K. &
Schindler, M. (2000) A new model for 20-hydroxyecdysone and
dibenzoylhydrazine binding: a homology modeling and docking
approach. Protein Sci. 9, 1073–1084.
8. Kumar, M., Fujimoto, T., Potter, D.W., Deng, Q. & Palli, S.R.
(2002) A single point mutation in ecdysone receptor leads to
increased ligand specificity: implications for gene switch applica-
tions. Proc. Natl Acad. Sci. USA 99, 14710–14715.
9. Albanese,C.,Reutens,A.T.,Bouzahzah,B.,Fu,M.,D’Amico,
M., Link, T., Nicholson, R., Depinho, R.A. & Pestell, R.G. (2000)
Sustained mammary gland-directed, ponasterone A-inducible
expression in transgenic mice. Faseb. J. 14, 877–884.
10. Zhou, Y. & Ratner, L. (2001) A novel inducible expression system
to study transdominant mutants of HIV-1 Vpr. Virology 287,

133–142.
11. Fussenegger, M. (2001) The impact of mammalian gene regulation
concepts on functional genomic research, metabolic engineering,
andadvancedgenetherapies.Biotechnol. Prog. 17, 1–51.
12. Yarovoi, S.V. & Pederson, T. (2001) Human cell lines expressing
hormone regulated T7 RNA polymerase localized at distinct
intranuclear sites. Gene 275, 73–81.
13. Yam, J.W., Chan, K.W. & Hsiao, W.L. (2001) Suppression of the
tumorigenicity of mutant p53-transformed rat embryo fibroblasts
through expression of a newly cloned rat nonmuscle myosin heavy
chain-B. Oncogene 20, 58–68.
14. Sparacio, S., Pfeiffer, T., Schaal, H. & Bosch, V. (2001) Genera-
tion of a flexible cell line with regulatable, high-level expression of
HIV Gag/Pol particles capable of packaging HIV-derived vectors.
Mol. Ther. 3, 602–612.
15. Patrick, C.W. Jr, Zheng, B., Wu, X., Gurtner, G., Barlow, M.,
Koutz, C., Chang, D., Schmidt, M. & Evans, G.R. (2001) Muri-
sterone a-induced nerve growth factor release from genetically
engineered human dermal fibroblasts for peripheral nerve tissue
engineering. Tissue Eng. 7, 303–311.
16. Pacchia, A.L., Adelson, M.E., Kaul, M., Ron, Y. & Dougherty,
J.P. (2001) An inducible packaging cell system for safe, efficient
lentiviral vector production in the absence of HIV-1 accessory
proteins. Virology 282, 77–86.
17. Lin, G. & Stern, R. (2001) Plasma hyaluronidase (Hyal-1) pro-
motes tumor cell cycling. Cancer Lett. 163, 95–101.
18. Christopherson, K.S., Mark, M.R., Bajaj, V. & Godowski, P.J.
(1992) Ecdysteroid-dependent regulation of genes in mammalian
cells by a Drosophila ecdysone receptor and chimeric transactiva-
tors. Proc. Natl Acad. Sci. USA 89, 6314–6318.

19. Yang, G., Hannan, G., Lockett, T. & Hill, R. (1995) Functional
transfer of an elementary ecdysone gene regulatory system to
mammalian cells-transient transfections and stable cell-lines. Eur.
J. Entomol. 92, 379–389.
20. No, D., Yao, T.P. & Evans, R.M. (1996) Ecdysone-inducible gene
expression in mammalian cells and transgenic mice. Proc. Natl
Acad. Sci. USA 93, 3346–3351.
21. Suhr, S.T., Gil, E.B., Senut, M.C. & Gage, F.H. (1998) High level
transactivation by a modified Bombyx ecdysone receptor in
mammalian cells without exogenous retinoid X receptor. Proc.
NatlAcad.Sci.USA.95, 7999–8004.
22. Swevers, L., Drevet, J.R., Lunke, M.D. & Iatrou, K. (1995) The
silkmoth homolog of the Drosophila ecdysone receptor (B1
isoform): cloning and analysis of expression during follicular cell
differentiation. Insect Biochem. Mol. Biol. 25, 857–866.
23. Hoppe, U.C., Marban, E. & Johns, D.C. (2000) Adenovirus-
mediated inducible gene expression in vivo by a hybrid ecdysone
receptor. Mol. Ther. 1, 159–164.
24. Wyborski, D.L., Bauer, J.C. & Vaillancourt, P. (2001) Bicistronic
expression of ecdysone-inducible receptors in mammalian cells.
Biotechniques 31, 618–620,622,624.
25. Saez, E., Nelson, M.C., Eshelman, B., Banayo, E., Koder, A.,
Cho, G.J. & Evans, R.M. (2000) Identification of ligands and
coligands for the ecdysone-regulated gene switch. Proc. Natl Acad.
Sci. USA. 97, 14512–14517.
26. Perera, S.C., Ladd, T.R., Dhadialla, T.S., Krell, P.J., Sohi, S.S.,
Retnakaran, A. & Palli, S.R. (1999) Studies on two ecdysone
receptor isoforms of the spruce budworm, Choristoneura fumi-
ferana. Mol. Cell Endocrinol. 152, 73–84.
27. Perera, S.C., Palli, S.R., Ladd, T.R., Krell, P.J. & Retnakaran, A.

(1998) The ultraspiracle gene of the spruce budworm, Chori-
stoneura fumiferana: cloning of cDNA and developmental
expression of mRNA. Dev. Genet. 22, 169–179.
28. Leid, M., Kastner, P., Lyons, R., Nakshatri, H., Saunders, M.,
Zacharewski, T., Chen, J.Y., Staub, A., Garnier, J.M., Mader, S.
et al. (1992) Purification, cloning, and RXR identity of the HeLa
cell factor with which RAR or TR heterodimerizes to bind target
sequences efficiently. Cell 68, 377–395.
29. Koelle, M.R., Talbot, W.S., Segraves, W.A., Bender, M.T.,
Cherbas, P. & Hogness, D.S. (1991) The Drosophila EcR gene
encodes an ecdysone receptor, a new member of the steroid
receptor superfamily. Cell 67, 59–77.
30. Perera, S.C., Sundaram, M., Krell, P.J., Retnakaran, A.,
Dhadialla, T.S. & Palli, S.R. (1999) An analysis of ecdysone
receptor domains required for heterodimerization with ultra-
spiracle. Arch. Insect Biochem. Physiol. 41, 61–70.
31. Tran, H.T., Askari, H.B., Shaaban, S., Price, L., Palli, S.R.,
Dhadialla, T.S., Carlson, G.R. & Butt, T.R. (2001) Reconstruc-
tion of ligand-dependent transactivation of Choristoneura fumi-
ferana ecdysone receptor in yeast. Mol. Endocrinol. 15, 1140–1153.
32. Lezzi, M., Bergman, T., Henrich, V.C., Vogtli, M., Fromel, C.,
Grebe, M., Przibilla, S. & Spindler-Barth, M. (2002) Ligand-
induced heterodimerization between the ligand binding domains
of the Drosophila ecdysteroid receptor and ultraspiracle. Eur. J.
Biochem. 269, 3237–3245.
33. Dhadialla, T.S., Carlson, G.R. & Le, D.P. (1998) New insecticides
with ecdysteroidal and juvenile hormone activity. Ann. Rev.
Entomol. 43, 545–569.
34. Martinez, A., Scanlon, D., Gross, B., Perara, S.C., Palli, S.R.,
Greenland, A.J., Windass, J., Pongs, O., Broad, P. & Jepson, I.

(1999) Transcriptional activation of the cloned Heliothis virescens
(Lepidoptera) ecdysone receptor (HvEcR) by muristeroneA.
Insect Biochem. Mol. Biol. 29, 915–930.
35. Karns, L.R., Kisielewski, A., Gulding, K.M., Seraj, J.M. &
Theodorescu, D. (2001) Manipulation of gene expression by an
ecdysone-inducible gene switch in tumor xenografts. BMC Bio-
technol. 1,11.
36. Carlson, G.R., Dhadialla, T.S., Hunter, R., Jansson, R.K., Jany,
C.S., Lidert, Z. & Slawecki, R.A. (2001) The chemical and bio-
logical properties of methoxyfenozide, a new insecticidal ecdy-
steroid agonist. Pest Manag. Sci. 57, 115–119.
37. Egea, P.F., Mitschler, A., Rochel, N., Ruff, M., Chambon, P. &
Moras, D. (2000) Crystal structure of the human RXRalpha
ligand-binding domain bound to its natural ligand: 9-cis retinoic
acid. EMBO J. 19, 2592–2601.
Ó FEBS 2003 EcR-based gene switch (Eur. J. Biochem. 270) 1315