Gene Therapy (1997) 4, 432–441
 1997 Stockton Press All rights reserved 0969-7128/97 $12.00
Positive and negative regulation of gene expression in
eukaryotic cells with an inducible transcriptional
regulator
Y Wang, J Xu, T Pierson, BW O’Malley and SY Tsai
Department of Cell Biology, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030, USA
To facilitate the understanding of the complex process of
target gene expression and its control, we report a modified
inducible system for activation or repression of target gene
expression in response to an exogenously administered
compound. The main component of this inducible system
is a chimeric transcriptional activator (GLVP) consisting of
an N-terminal VP16 transcriptional activation domain fused
to a yeast GAL4 DNA binding domain and a mutated
human progesterone receptor (hPR) ligand binding domain
(LBD). This chimeric regulator binds to a target gene containing the 17-mer GAL4 upstream activation sequence
(UAS) in the presence of anti-progesterone, RU486. We
showed that the combination of two different types of
domains (VP16 and poly-glutamine stretch) into one chimeric molecule could result in a further increase in transcriptional activation potency. Through mutational analysis,
we modified the original GLVP and generated a more
potent version of the RU486 inducible regulator GL914VPC9
with a 19 amino acid deletion of the hPR-LBD (DC19) and
a C-terminally located VP16 activation domain. More
importantly, this new chimeric regulator can effectively activate target gene expression at a much lower concentration
of RU486 (0.01 nM). The concept of RU486 regulatable
gene expression is not limited to gene activation. By
replacing the VP16 activation domain with a KRAB transcriptional repression domain, we are able to achieve
inducible repression of target gene expression. We also
present evidence that individual functional domains within
a chimeric protein could modulate each other’s function
depending on their relative positions within the molecule.
Using this potent regulator, we demonstrate that inducible
nerve growth factor (NGF) secretion into conditioned media
can elicit neurite outgrowth in co-cultured PC12 cells. This
new versatile inducible system can potentially be used to
control target gene expression in a mammalian system in
vivo.
Keywords: inducible system; gene activation and repression; RU486
Introduction
The expression of most mammalian genes is intricately
regulated in vivo in response to a wide range of stimuli,
including physical (pressure, temperature, light), electrical (eg motor and sensory neuron signal transmission) as
well as biochemical (ions, nucleotides, neurotransmitters,
steroids and peptides) in nature. While the mechanism
of transcriptional regulation of gene expression has been
extensively studied,1 progress on achieving target gene
regulation in mammalian cells, without interfering with
endogenous gene expression, has been limited. Currently, most strategies for target gene activation or
repression are performed in a constitutive manner. Such
uncontrolled regulation of gene expression is not ideal
physiologically, and can even be deleterious to cell
growth and differentiation.
Several inducible systems have been employed for controlling target gene expression. These inducible agents
include heavy metal ions, 2 heat shock,3 isopropyl b-dthiogalactoside (b-gal),4 and steroid hormones such as
estrogen5 and glucocorticoids. 6 However, many of these
Correspondence: SY Tsai
Received 5 September 1996; accepted 23 December 1996
inducers are either toxic to mammalian cells or interfere
with endogenous gene expression. 7 Utilizing a bacterial
tetracycline-responsive operon element, Gossen et al
developed an interesting model for controlling gene
expression with a tetracycline-controlled transactivator
(tTA and rtTA). 8,9 No et al recently reported a threecomponent system consisting of a chimeric GAL4-VP16ecdysone receptor, its partner retinoid X receptor (RXR),
and a target gene; they demonstrated its application in
activating reporter gene expression in an ecdysonedependent manner.10
Previously we have reported a novel RU486-inducible
system for controlling target gene expression.11 This
inducible system consists of a regulator and a target gene.
The regulator is a chimeric RU486-inducible transcriptional activator composed of a VP16 transcriptional activation domain in the N-terminus followed by a yeast
GAL4 DNA binding domain and a mutated human progesterone receptor (hPR) ligand binding domain (LBD).
This mutated PR, containing a 42 amino acid deletion in
the C-terminus (DC42) of the LBD, does not bind to progesterone, but binds the progesterone antagonist RU486
(mifepristone) and activates transcription.12 The advantage of using the yeast GAL4 DNA binding domain is its
specificity, since this chimeric regulator will only recognize target gene constructs containing the 17-mer (GAL4
Inducible regulation of gene expression
Y Wang et al
binding) sequence but not the endogenous mammalian
genes. Moreover, this system is only activated in the presence of an exogenous compound, RU486, but not by any
endogenous molecules present in the mammalian tissues
and organs. Furthermore, the regulator GLVP was activated by a very low concentration (0.1–1 nm) of a synthetic and orally active antiprogestin RU486. We have
demonstrated previously the successful use of this system both in vitro and ex vivo11 with different target gene
constructs harboring either the chloramphenicol acetyltransferase or the tyrosine hydroxylase gene.
While the original GLVP can efficiently activate target
gene expression containing stronger promoters such as
the thymidine kinase (tk) promoter, its activity on a minimal TATA promoter is limited. To enhance further the
transcriptional activity of GLVP, we report here the generation of a more potent RU486-inducible gene regulator.
Most importantly, this new gene regulator responds to
RU486 at a concentration even lower than that used by
the original GLVP. At this concentration, RU486 does not
have any anti-progesterone or anti-glucocorticoid
activity. The inducible system has been used successfully
to produce secreted NGF from a reporter gene in an
RU486 dependent manner to induce neurite outgrowth
in co-cultured PC12 cells (of rat adrenal pheochromocytoma). We also demonstrate that this RU486-controllable ligand binding domain can be converted to an
inducible repressor for shutting down target gene
expression. Collectively, these studies demonstrate that
individual domains within a chimeric fusion protein
influence each other’s function in a position-dependent
context.
Results
Role of additional activation domains in the chimeric
regulator
Several different functional domains have been characterized in transcription factors; they can be either acidic
(VP16, GAL4), glutamine-rich (SP1, Oct-1, Oct2A),
proline-rich (Oct3/4), or serine- and threonine-rich
(Pit1).13 It is known that different types of transcriptional
activation domains interact with different coactivators of
the general transcriptional machinery. When different
activation domains are fused together in a transactivator,
they can synergize with each other to increase its transcriptional potential. Recently, Schaffner and colleagues14
demonstrated that insertion of either a poly-glutamine
(poly-Q) or poly-proline (poly-P) stretch within the
GAL4-VP16 enhances the activation of GAL4-VP16. In
order to increase the potency of the GLVP regulator, we
inserted varying lengths of poly-Q stretches encoded by
the triplet repeats (CAG)n, into the N-terminus of the
GLVP regulator (Figure 1a). Transactivation analysis of
the various sizes of poly-Q insertions in the GLVP indicate that addition of 10–34Q increases transcriptional
activity of the regulator on the reporter gene (17 × 4TATA-hGH), while further extension of poly-Q from a
66 Q oligomer to a 132 Q oligomer results in decreased
activation of the target gene (Figure 1b). These experiments demonstrated that a combination of different types
of functional domains of appropriate strength further
improves the activation potential of the GLVP chimeric
regulator.
To understand whether additional activation domains
of the same type would also increase the activation
potential of the chimeric regulator, we constructed
GLVP × 2 with two copies of VP16 activation domain at
the N-terminus (Figure 1a). As shown in Figure 1c,
further addition of the same type of transactivation
domain (VP16) did not increase the activation potential
of the regulator.
Extension of the ligand binding domain of hPR further
improves the responsiveness of GLVP to RU48
The original chimeric regulator GLVP contains a Cterminally truncated hPR-LBD at amino acid (aa) position
891, which disrupted the activation function (AF2) as
well as the dimerization ability of the hPR. We asked
whether an increase in the length of hPR-LBD in the
GLVP might enhance its transcriptional activity in
response to the antiprogestin RU486. For this purpose, a
series of deletion mutants which extend the C-terminus
of the LBD from aa 879 to aa 928 were generated (Figure
2a). These mutants were examined for their ability to confer stronger RU486, but not progesterone, inducible
activity in transient cotransfection assay with a 17 × 4TATA-CAT reporter construct. By lengthening the C-terminal ligand binding domain from 879 to 914 (Figure 2a),
we observed a gradual increase in RU486 induced activation of target gene expression (Figure 2b). Importantly,
these mutants responded specifically to RU486, but not
to the progesterone agonist R5020. Further extension of
the C-terminal LBD beyond aa 914 resulted in a decrease
of GLVP response to RU486 (Figure 2b).
Location of the transcriptional activation domain VP16
within the chimeric regulator affects its transactivation
potential
The original chimeric regulator GLVP contained the VP16
transcriptional activation domain located at the Nterminus of the molecule. To address the possibility
whether positioning of the VP16 activation domain
within the chimeric molecule would affect its transcription activation potential, we constructed a series of chimeric regulators with the VP16 activation domain located
at the C-terminus. As illustrated in Figure 3, the regulator
with a C-terminally located VP16 is more potent than its
N-terminal counterpart (compare GLVP and GL891VPC9 in
Figure 3a and GL914VP and GL914VPC9 in Figure 3b). In
addition, extension of the C-terminal LBD from aa 879 to
aa 914 further increased transactivational activity of the
regulator in this C-terminally located VP16 chimera
(compare GL879VPC9 with GL914VPC9 in Figure 3a). Thus,
extension of the LBD to aa 914 further enhances the
RU486-dependent
transactivation,
irrespective
of
whether VP16 is located in the N- or C-terminus, suggesting the existence of a weak dimerization and activation
function between aa 879 and 914 of the PR-LBD. By transferring the VP16 activation domain from the N-terminus
to the C-terminus, we generated a much more potent
transactivator GL914VPC9 . In order to exclude the possibility that this significant difference in the capacity for
gene activation was not due to different protein
expression levels of the two chimeric regulators, we performed Western blot analysis of extracts from HeLa cells
transiently transfected with either GL914 VP or GL914VPC9
(duplicates) and confirmed that the two regulator proteins are expressed at a similar level (Figure 3c, compare
433
Inducible regulation of gene expression
Y Wang et al
434
Figure 1 Effect of additional transcriptional activation domain on GLVP activity. (a) Diagrams of GLVP and its derivatives containing an additional
transactivation domain. In GLVP-nQ, a poly-glutamine (n-oligomer) stretch was fused directly to the N-terminal part of the GLVP. GLVP × 2 has
tandem copies of VP16 activation domain (ac 411–487) fused together at the N-terminus. The reporter plasmid, 17 × 4-TATA-hGH,40 contains four
copies of 17-mer GAL4 binding sequence, adenoviral E1B minimal promoter (TATA box) and the human growth hormone gene. (b) Effect of various
lengths of poly-Q insertion on GLVP transactivation potential. HepG2 cells (of liver origin) were transfected with 2 mg of expression plasmid (in pCEP4
vector) and 10 mg of reporter plasmid 17 × 4-TATA-hGH. Cells were treated with either 1 nm RU486 or vehicle control (80% ethanol). Aliquots (20
ml) of the cell culture media were taken at different time intervals as indicated and diluted 1:5 for measurement of hGH. (c) An additional copy of the
VP16 activation domain into GLVP does not further increase its transactivation potential. HepG2 cells were cotransfected with 0.5 mg expression
plasmid (RSV) and 5 mg of reporter 17 × 4-TATA-hGH. Cells were incubated with 10 nm RU486 (filled bar) or vehicle control (open bar) for 36 h.
Cell culture medium was then harvested and diluted 1:4 for analysis of hGH content by radioimmunoassay. The standard error bar represents variations
of the mean from three individual transfection experiments.
lanes 2 and 3 with lanes 4 and 5). Together, these results
suggested that through modification of the PR-LBD
within the chimeric regulator we could further improve
its response to a ligand by approximately one order of
magnitude.
The new regulator GL914VPC9 activates target gene
expression potently at a lower concentration of RU486
Since RU486 has been known to act as an antagonist of
progesterone and glucocorticoid when used at a high
concentration (100 nm), it would be desirable for the
chimeric regulator GLVP to activate target gene
expression at a substantially lower concentration of
RU486. We examined the optimal concentration of RU486
for these two transcriptional activators GL914VPC9 and
GL914VP. As shown in Figure 4, GL914VP activity
occurred at an RU486 concentration of 0.1 nm and
reached a maximal level at 1 nm. The result is similar to
what we have observed previously in a stable cell line
harboring the GLVP and reporter 17 × 4-tk-TH (tyrosine
hydroxylase).11 In contrast, GL914VPC9 increased reporter
gene expression at an RU486 concentration 10-fold lower
Inducible regulation of gene expression
Y Wang et al
435
Figure 2 Extension of the ligand binding domain enhances RU486-inducible transcriptional activation. (a) Diagram of the original chimeric GLVP and
its C-terminally extended derivatives. VP16 activation domain contains aa 411–489 of the HSV VP16, the GAL4 DNA binding domain was derived
from residue 1 to 93. Different amino acid segments of the human progesterone receptor (hPR) ligand binding domain (LBD) were used in the chimeric
protein. (b) Various C-terminal truncations of progesterone receptor ligand binding domain affect the transactivation potential of chimeric regulator.
Transient transfection assay shows the activity of various GLVP derivatives on activation of target gene 17 × 4-TATA-CAT. HeLa cells were transfected
with 4 mg of chimeric regulators (driven by RSV promoter) as indicated and 10 mg of reporter 17 × 4-TATA-CAT. After transfection, cells were
incubated with either 10 nm RU486 (RU) or 10 nm progesterone agonist R5020 (P). Ethanol (85%) was used as vehicle control (−).
(0.01 nm) than that of GL914VP. Thus, the modified
GL914VPC9 is not only more potent but also activates the
reporter gene at a lower concentration of ligand as compared with GL914VP. This newly discovered character of
GL914VPC9 is important for its use in inducible target gene
expression, since it would allow the use of a concentration which has no anti-progesterone or anti-glucocorticoid activity. This represents a significant advantage
when the inducible system is applied in in vivo situations,
as exemplified by transgenic mice and potentially by
gene therapy.
Inducible repression of target gene expression
To explore the possibility of converting a transactivator
GLVP to a regulatable repressor, we replaced the VP16
activation domain with a repression domain, the
Krüppel-associated box-A (KRAB), from the zinc finger
protein Kid-1. Kid-1 was identified as a kidney-specific
transcription factor and was shown to be regulated during renal ontogeny and injury. 16,17 The KRAB domain (aa
1–70) was inserted in either the N- or C-terminus of GL914
(Figure 5a) and inducible repression by RU486 was analyzed in cotransfection experiments using HeLa cells. The
reporter plasmids 17 × 4-tk-CAT (with the thymidine kinase promoter) and 17 × 5-SV-CAT (containing the SV40
enhancer) were selected in order to study the repression
of basal promoter and enhancer-mediated transcriptional
activity, respectively. Both reporters contain multiple
copies of the GAL4 binding site (17-mer UAS sequence)
and exhibit constitutive basal expression from either a tk
promoter or SV40 enhancer by themselves. As shown in
Figure 5b, the chimeric regulator GL914KRAB, with the
KRAB repression domain inserted in the C-terminus,
strongly repressed expression (six- to eight-fold) of both
reporters in an RU486-dependent manner. However, the
N-terminally located KRAB repression domain
(KRABGL914) did not repress target gene expression in
the presence of RU486 to the degree of that achieved with
KRAB located in the C-terminus (GL914 KRAB). This
observation is similar to what we have observed previously with the positioning of the VP16 activation
domain.
Inducible neurite outgrowth in PC12 cells with regulated
expression of nerve growth factor
To demonstrate the use of the inducible system in a biological situation, we designed a regulatable expression
model for nerve growth factor (NGF). NGF has been
shown to stimulate neurite (axon) outgrowth of PC12
cells (from rat adrenal pheochromocytoma) when added
to the cell culture medium.18 We have established a stable
rat fibroblast cell line (C4FRNGF) that possesses both
regulator GL914VPC9 and reporter 17 × 4-TATA-NGF.
These cells were grown in the presence or absence of
RU486 (10 nm) for 2 days and the conditioned medium
was collected and assayed for NGF-stimulated neurite
outgrowth of PC12 cells. When conditioned medium
(from C4FRNGF cells treated with RU486) was added to
PC12 cells, we observed strong neurite outgrowth from
PC12 cells after 48 h of incubation (Figure 6C and D).
Little if any neurite outgrowth was observed in PC12
cells incubated with the conditioned medium that was
collected from stable cells treated with vehicle only (85%
ethanol) (Figure 6 A and B). These results demonstrate
that the inducible system can be used to study a particular biological phenomenon in a controllable fashion.
Discussion
Transcriptional regulation of gene expression has been
intensively studied over the past decade.1,19,20 It is generally believed that transcription factors selectively bind to
their recognition sequences on DNA (promoters and
enhancers) and directly interact with the TBP-associated
factors (TAFs), coactivators, or corepressors to activate or
repress transcriptional activity. Nuclear hormone recep-
Inducible regulation of gene expression
Y Wang et al
436
Figure 3 C-terminally located VP16 activation domain is more potent in activating target gene expression. (a) Transcriptional activation of GLVP
versus its C-terminally located VP16 activation domain and various extensions of the hPR-LBD. HeLa cells were cotransfected with regulator plasmid
(4 mg) and reporter 17 × 4-TATA-CAT (10 mg). RU486 (+) or vehicle control (−) was added to the cells as indicated. Results were from three separate
transfection experiments and experimental variations are illustrated with error bars. Numbers above the bar are fold induction for each regulator in the
presence of RU486 (10 nm). (b) C-terminally located VP16 domain is a more potent transcriptional activator. CAT assay showing activity of Nterminally located VP16 regulator pGL 914VP and the C-terminally located VP16 regulator pGL 914VPC9 . HeLa cells were cotransfected with the expression
plasmid (4 mg) and reporter 17 × 4-TATA-CAT (10 mg) and 100 mg of cell extracts were assayed for 1.5 h. RU486 (+, 10 nm) or vehicle control (−,
85% ethanol) was added to the cells as indicated. (c) Expression level of GLVP and GL914VP C9 regulators is comparable in the HeLa cells. Western blot
of protein extracts (20 mg) from HeLa cells transfected with 10 mg of GLVP (lanes 2 and 3) or GL 914VPC9 (lanes 4 and 5) and pCEP4 control vector
(lane 1). Each lane represents an individual transfection of HeLa cells. The blot was probed with anti-GAL4 (1–147) antibody (Clontech) and developed
with ECL staining (Amersham).
tors, such as steroid, thyroid, retinoid and orphan receptors, are a unique class of inducible transcription factors
that can modulate their respective target genes in
response to their cognate ligands. Recently, we and
others have identified several coactivators (SRC-1),21
CBP,22 and corepressors (N-CoR, SMART) 23,24 that
mediate nuclear hormone receptor activation of target
genes. These studies suggest that multiple protein factors
are involved in the complex process of transcriptional
regulation of gene expression.
To understand further this complex process of gene
regulation, we have utilized an inducible chimeric regulator to study: (1) how mutations in the hPR ligand binding domain affect its response to ligand; (2) how different
types of activation domains function together in a chimeric molecule; and (3) whether the positioning of a transactivation domain might affect its activation potential.
Mutagenesis studies of the hPR ligand binding domain
have demonstrated that extension of the LBD deletion
from aa position 891–914 increases the activation potential of the chimeric regulator. We propose that the
addition of this short stretch of 23 aa increases the PRLBD’s dimerization potential and subsequent binding to
its response element. During the course of this study, we
also discovered that further extension of the hPR-LBD
from residue 917 to 928 results in a decrease of transactivation, suggesting that this region may serve as a repressor
interacting domain. In fact, when this 12 aa stretch was
ligated to the GAL4 DNA binding domain, it was sufficient to confer transcriptional repression of a target
gene, suggesting that these 12 aa might interact with a
yet unidentified cellular co-repressor.25
Many chimeric proteins have been constructed in
recent years in order to combine different functional
domains of various proteins into one versatile chimera.
While it is clear that each protein domain can function
independently, relatively little is known about how individual domains modulate each other’s function within a
chimeric protein. In this study, we demonstrated that the
activation potential of VP16 is influenced by its relative
position within the chimeric regulator. The C-terminally
located VP16 chimeric regulator GL914VPC9 effectively
Inducible regulation of gene expression
Y Wang et al
437
Figure 4 Dose–response of regulator GL 914VP vs GL914 VPC9 to RU486
induced gene activation. HeLa cells were transfected with 4 mg of regulators and 10 mg of reporter 17 × 4-TATA-CAT. Cells were incubated
with various concentrations of RU486 or vehicle control (85% ethanol)
as indicated. The CAT activities were assayed using 100 mg of protein
extracts from transfected cells with overnight incubation at 37°C.
activates target gene expression containing a minimal
promoter at an RU486 concentration 10-fold lower than
its N-terminally located VP16 counterpart, GL914VP. At
this concentration, RU486 is expected to have no interference with endogenous gene expression. Therefore, this
new inducible system will afford an improved margin of
safety over our previously reported GLVP and further
contribute to its application for gene regulation in vivo.
Figure 6 Control of neurite outgrowth with regulatable expression of
nerve growth factor (NGF). Morphology of PC12 cells grown in medium
collected from the C4FRNGF2 stable cell line culture in the presence of
10 nm RU486 or vehicle control (85% ethanol). The stable rat fibroblast
cell line (C4FRNGF2) was established by transfecting GL914VP C9 (pCEP4
vector containing the hygromycin resistance gene) and 17 × 4-TATANGF reporter gene (containing the neomycin resistance gene) and selected
with hygromycin and G418. C4FRNGF2 cells were incubated with either
RU486 or vehicle control for 48 h and morphological phenotypes of the
cells were studied. (A and B) vehicle control. (C and D) RU486-treated
stable cell culture medium stimulates neurite outgrowth of PC12 after 48
h incubation.
In all, mutational studies revealed that the chimeric regulator GL914VPC9 is about eight to 10 times more potent
than our originally described regulator GLVP and
responds at a lower ligand concentration. Our results also
demonstrated that within a chimeric protein, individual
Figure 5 Inducible repression of target gene expression. (a) Diagram of inducible repressor and reporter constructs. Kid-1 KRAB domain containing
amino acid residues 1–70 was fused to the N- or C-terminus of the GAL4-PRLBD (640–914) chimera. (b) CAT assay demonstrating inducible repression
of target gene expression in the presence of RU486. HeLa cells were cotransfected with 4 mg of expression plasmid containing the inducible repressor
constructs and 10 mg of reporter plasmid. pCEP4 parental vector was used as the control. After transfection, cells were incubated with either 10 nm
of RU486 (+) or vehicle control (−). Numbers indicate the percentage of acetylated chloramphenicol (% conversion) and fold repression in the presence
of RU486.
Inducible regulation of gene expression
Y Wang et al
438
functional domains, such as those involved in transactivation, DNA binding and ligand binding, can modulate
each other’s function, depending on their relative
positions.
Protein–protein interaction studies suggest that different types of transactivation or transrepression domains
interact with their respective TAFs or coactivator, corepressor molecules within the RNA polymerase II preinitiation complex to alter gene transcription. 20,26 Glutamine-rich stretches have been identified in various
transcriptional factors (SP1, Oct-1 and androgen receptor)
although their precise function is unknown.13,14
Expanded regions of triplet CAG repeats have been
implicated in several neurodegenerative diseases such as
Huntington’s, Kennedy’s, dentatorubral-pallidoluysian
atrophy (DRPLA), and hereditary spinocerebellar ataxias
(SCA1).27–29 Recently, several groups have isolated proteins responsible for the above mentioned neurodegenerative diseases and confirmed that they indeed contain
long poly-glutamine (Q) stretches encoded by the
expanded CAG repeats.30–32 To understand further the
role of poly-Q stretches in transcriptional regulation, we
inserted various lengths of poly-Q in the N-terminus of
GLVP. Our results demonstrated that addition of a 10–34
oligomer of poly-Q results in synergistic transcriptional
activation, while expanded CAG triplet repeats beyond
66 oligomeric glutamines do not increase further the
transactivation potential of chimeric regulator GLVP.
These observations suggest that structural and conformational changes might be involved in proteins encoded
by the expanded CAG triplet repeat as compared with
the regular length poly-Q which is encoded by 10–30
repeats of CAG in normal protein. These results suggest
that a neurological disease with expanded CAG repeats
(.40 mer) may not be due to aberrant high transcriptional potential but rather due to an influence on other
aspects of cell function.33
A transcription factor can either activate or repress
gene expression depending on the promoter/enhancer
context of its particular target DNA and the coregulator
proteins with which it interacts.34 For example, in the
absence of thyroid hormone (T3), the thyroid hormone
receptor (TR) normally binds to its recognition sequence
on DNA and represses target gene activation through
interactions with corepressors (Refs 24, 35 and H Shibata,
SY Tsai, M-J Tsai, BW O’Malley (unpublished data)). In
the presence of T3, the corepressor is released from the
receptor and coactivators are recruited to enhance gene
expression. Many transcription factors, such as p53, WT1, YY1, Rel, can also act as dual activators and repressors
depending on the DNA template and protein cofactors
with which they interact. The Drosophila zinc finger transcription factor, Krüppel, is encoded by a gap gene and
is essential for organogenesis during later stages of the
development. Through in vitro protein–protein interaction studies, Sauer et al36 have demonstrated that the
Krüppel protein can act as a transcriptional activator at
a low protein concentration (monomeric form) by interacting with TFIIB. However, at a higher protein concentration, Krüppel forms a dimer and directly interacts with
TFIIEb resulting in transcriptional repression. Several
Krüppel-related proteins have recently been identified in
mammalian cells.16,17,37 One of them, Kid-1, was isolated
from rat kidney and contains a highly conserved region
of approximately 75 aa at the N-terminus termed
Krüppel-associated box (KRAB). It has been shown that
the KRAB domain can act as a potent repressor when
fused to a yeast GAL4 DNA binding domain or TetR.38
In the present study, we demonstrated that by replacing
the VP16 transcriptional activation domain with this Kid1 KRAB repression domain, we could convert a regulatable transactivator into a regulatable repressor. We postulate that by exchanging the GAL4 DNA binding
domain with the DNA binding domain of another protein, repression of a target gene (eg tumor proliferation
gene) could be achieved in response to ligand RU486.
Recently, Deuschle et al38 reported that the KRAB domain
isolated from Kox1 zinc finger protein, which shares
extensively homology with that of Kid-1, interacts with a
110 kDa adaptor protein termed SMP1 (silencingmediating protein 1). The characteristics and mechanism
of this adaptor protein have yet to be determined.
Using the newly modified GL914VPC9, we have successfully achieved regulation of neurite outgrowth in PC12
cells via RU486 controllable expression of NGF. Our studies show that this novel inducible system can be
employed to analyze biological function in a temporal
manner. For example, the role of a growth factor could
be assessed at a particular stage of development and the
sequential relationship of in vivo cell death and proliferation could be delineated in a manner not possible with
constitutive expression of the test gene.
Recently, we have successfully demonstrated tissuespecific regulation of gene expression in transgenic mice
utilizing this inducible system.39 We also foresee the use
of our RU486 inducible regulator to create an inducible
gene knockout (either temporal and/or spatial) in transgenic mice which could circumvent an embryonic lethality resulting from the use of current gene knockout techniques. We envision that combinatorial inclusion of other
inducible systems such as the tetracycline or ecdysone
system with the RU486 inducible system might allow
biologists one day to modulate complex biological processes which involve multiple levels of control.
Materials and methods
Plasmid constructs
Construction of poly-glutamine stretch insertion into
GLVP: The poly-glutamine stretch containing multiple
repeats of CAG was constructed by a method developed
by Seipel et al15 utilizing multimerization of DNA fragment (BsaI and BbsI digested) coding glutamine repeats
leading to poly-Qn. Plasmid pBluescript-KS(II) was
digested with Acc65I and SacI, the linearized vector was
gel purified and ligated with the annealed oligonucleotide pair R3/R4 to create plasmid pPAP. The oligonucleotide sequence for R3 (upper strand) is: 5′-GTACGTTTAA
ACGCGGCGCGCCGTCGACCTGCAGAAGCTTACTAG
TGGTACCCCATGGAGA
TCTGGATCCGAATTCAC
GCGTTCTAGATTAATTAAGC-3′ and the sequence for
R4 (lower strand) is: 5′-GGCCGCTTAATTAATCTAGAA
CGCGTGAATTCGGATCCAGAT CTCCATGGGGTAC
CACTAGTAAGCTTCTGCAGGTCGACGGCGCGCCGC
GTTTAAAC-3′. The following restriction sites are
incorporated into pPAP as the multiple cloning sites
(from T3 to T7): PmeI, AscI, SalI, PstI, HindIII, SpeI, Acc65I,
NcoI, BglII, BamHI, EcoRI, MluI, XbaI, PacI, NotI, SacI. Oli-
Inducible regulation of gene expression
Y Wang et al
gonucleotides coding for 10 glutamines were annealed
and subcloned into the BglII and BamHI site of plasmid
pPAP. The sequence for the upper and lower strand
oligonucleotide is, 10QU 5′-GATCTCGGTCTCCAACAG
CAACAGCAACAGCAACAGCAACAGGGTCTTCTG-3′
and 10QL: 5′-GATCCAGAAGACCCTGTTGCTGTTGCT
GTTGCTGTTGCTGTTGGAGACCGA-3′,
respectively.
The insert was confirmed by restriction digestion and
sequencing. The plasmid with 10Q insert (pPAP-10Q)
was digested with BsaI and BbsI (New England Biolab,
Beverly, MA, USA) overnight and precipitated. Onetenth of the precipitated DNA (containing both vector
and fragment) was re-ligated to create plasmid pPAP18Q. Each ligation step results in pAP-2(n-1)Q from the
previous vector pPAP-nQ. In this way various expansions of poly-Q were achieved and the resulting plasmids
pPAP-34Q, pPAP-66Q and pPAP132Q were created and
confirmed by sequencing. The BglII and BamHI fragments
(coding for poly-Q stretch) from these plasmids were
purified and cloned into the BglII site of pRSV-GLVP to
generate GLVP with various poly-Q inserts at the N-terminus. These GLVP-nQ were reinserted into the pCEP4
expression vector (containing CMV immediate–early
gene enhancer/promoter) creating pCEP4-GLVP-nQ.
Chimeric fusion protein with various C-terminus deletions:
To construct GLVP chimeras with various C-terminal
deletions of the human progesterone receptor ligand
binding domain, the HindIII to BamHI fragment containing these various deletions in pRSV-hPR plasmids25 was
gel purified with QIAEX II gel extraction kit (Qiagen,
Chatworth, CA, USA). The purified fragments were subcloned into HindIII and BamHI sites of pRSV-GLVP11
replacing the aa region 610–891 of the GLVP.
GLVPC9 chimeras with VP16 activation at the C-terminus:
We used two-step clonings to move VP16 activation to
the C-terminus of the chimeric fusion protein. First, the
hPR-LBD region (from aa 800 to various C-termini) was
amplified using 5′ primer (5′-TATGCCTTACCATGT
GGC-3′) with a different 3′ primer as a pair and digested
with HindIII to SalI to prepare the fragment for ligation.
For a different position of amino acid truncation, the 3′
primers incorporating the SalI site are: P3S–879: 5′-TTGG
TCGACAAGATCATGCATTATC-3′; P3S–891: 5′-TTGTC
GACCCGCAGTACAGATGAAGTTG-3′, and P3S–914: 5′TTGGTCGACCCAGCAATAACTTCAGACATC-3′. The
DNA fragment containing the VP16 activation domain
(aa 411–490) was isolated from pMSV-VP16-D3′-b58N′
(from A Friedman, Johns Hopkins Oncology Center, Baltimore, MD, USA) with SalI and BamHI. The digested
PCR fragment and VP16 activation were ligated together
into the HindIII and BamHI sites of expression vector
pCEP4 (Invitrogen, San Diego, CA, USA). The ligated
vector pCEP4-PV (LBD 810–879 and VP16), -C3 (LBD
810–891 and VP16), -C2 (LBD 810–914 and VP16), respectively, now contain C-terminal fragments of hPR-LBD
from the HindIII site (amino 810) to various truncations
of LBD fused 3′ to the VP16 activation domain with
BamHI after the termination codon of VP16. We then
replaced the HindIII–BamHI fragment from pGL (in pAB
vector) with PV, C3, and C2 fragments, respectively, to
yield pGL879VPC9, pGL891VPC9, and pGL914VPC9. These
chimeric fusion proteins were then subcloned into Acc65I
and BamHI sites of pCEP4 expression vector and were
named as pCEP4-GL879VPC9, pCEP4-GL891VPC9, pCEP4GL914VPC9.
Inducible repressor containing the Kid-1 KRAB domain:
The Kid-1 gene (cDNA kindly provided by Dr JV Bonventre, Massachusetts General Hospital, Charlestown,
MA, USA) containing the KRAB domain (aa 1–70) was
amplified with two sets of primers for insertion into the
N- or C-terminus of GL914, respectively. For the KRAB
domain to be inserted at the N-terminus of the fusion
protein, the Kid-1 cDNA was amplified with a set of primers as follows: Kid3: 5′-CGACAGATCTGGCTCCTGAG
CAAAGAGAA-3′, Kid4: 5′-CCAGGGATCCTCTCCTTGC
TGCAA-3′. The PCR products were digested with BglII
and BamHI and subcloned into pRSV-GL891 to create
pRSV-KRABGL891. The KpnI–SalI fragment of KRABGL891
was then purified and subcloned into KpnI–SalI sites in
pRSV-GL914VP to create pRSV-KRABGL 914. The entire
KRABGL914 fragment (KpnI–BamHI) was then inserted
into the KpnI and BamHI-digested pCEP4 generating
pCEP4-KRABGL914. For C-terminally located KRAB
domain, the Kid-1 gene was amplified with the following
set of primers: Kid1: 5′-TCTAGTCGACGATGGCTCCTG
AGCAAAGAGAAG-3′, Kid2: 5′-CCAGGGATCCTATCC
TTGCTGCAACAG. The primer Kid2 also contains a termination codon (TAG) after aa 70. The PCR products
were digested with SalI and BamHI and purified using
QIAEX II gel extraction kit (Qiagen). The HindIII and SalI
fragment (317 bp) from pBS-GL 914VPC9, was isolated as
was the vector fragment of pCEP4-GL914VPC9 digested
with HindIII and BamHI. These three piece fragments
were ligated to create pCEP4-GL914KRAB.
Transient transfection, CAT assay, hGH assay and
Western blot
HeLa and CV1 cells were transfected with the described
amount of DNA using the polybrene-mediated Ca2PO4
precipitation method and CAT assay was performed and
quantified as described previously.11 HepG2 cells (106)
were grown in DMEM with 10% fetal bovine serum and
1 × penicillin–streptomycin–glutamine (GIBCO-BRL,
Gaithersburg, MD, USA) and transfected with the polybrene-mediated Ca2 PO4 precipitation method. Aliquots
of the cell culture media were taken at different time
intervals and hGH production was measured using the
hGH clinical assay kit (Nichols Institute Diagnostics, San
Juan Capistrano, CA, USA) according to the manufacturer’s instructions. For Western blot analysis, protein
extracts (20 mg) were prepared from transiently transfected HeLa cells, separated on SDS polyacrylamide gel
and trans-blotted on to nylon membrane as previously
described. The blot was probed with anti-GAL4-DBD (aa
1–147) monoclonal antibody (Clontech, Palo Alto, CA,
USA) and developed with an ECL kit (Amersham,
Arlington Heights, IL, USA).
Stable cell line generation and neurite outgrowth assay
Rat FR cells, derived from rat fetal skin cells (CRL 1213;
American Type Culture Collection, Rockville, MD, USA)
were transfected with pCEP4-GLVP914VPC9 by the Ca2PO4
method as described previously.11 Cells were grown in
DMEM with 10% fetal bovine serum and selected with
50 mg/ml hygromycin-B (Boehringer Mannheim, Indianapolis, IN, USA). After 2–3 weeks, colonies were picked
and subsequently expanded. Each clone was then transi-
439
Inducible regulation of gene expression
Y Wang et al
440
ently transfected with 2 mg of the p17 × 4-TATA-CAT
plasmid utilizing Lipofectin (GIBCO-BRL). Twenty-four
hours later, the cells were treated with either RU486
(10−8 m) or 80% ethanol vehicle. Cells were harvested 48
h later and CAT activity was measured using 50 mg of
cell extracts. Clones showing RU486-inducible CAT
activity were subsequently transfected with the vector
p17 × 4-TATA-rNGF (Neo). Stable cells containing both
genes were selected with hygromycin (50 mg/ml) and
G418 (100 mg/ml) for 2–3 weeks and subsequently
expanded. Each colony was then seeded into a 10 cm culture dish and treated with 10−8 m RU486 or vehicle control (80% ethanol). After 48 h, the conditioned medium
was collected and frozen. Subsequently, the conditioned
medium was thawed and diluted two-fold in DMEM
with 10% horse serum and 5% fetal bovine serum. The
diluted conditioned medium was then placed on PC12
cells, with new diluted conditioned medium added every
2 days. After 5–7 days, PC12 cells were observed for
neurite outgrowth.
12
13
14
15
16
17
18
Acknowledgements
We thank Dr JV Bonventre for Kid-1 cDNA and 17 × 5SV-CAT plasmids and Dr SR Whittemore for rat NGF
cDNA. We are also grateful to Drs Tom Spencer and Neil
McKenna for comments on the manuscript. We especially
thank Dr Ming-Jer Tsai for helpful suggestions and discussions. This work was supported by National Institute
Health grants to Sophia Y Tsai and Bert W O’Malley. The
RU486 regulatable gene switch system has been licensed
by Baylor College of Medicine to Gene Medicine Inc.,
Woodlands, Texas.
References
1 McKnight SL. Transcription revisited: a commentary on the 1995
Cold Spring Harbor Laboratory meeting, ‘Mechanisms of Eukaryotic Transcription’. Genes Dev 1996; 10: 367–381.
2 Mayo KE, Warren R, Palmiter RD. The mouse metallothioneinI gene is transcriptionally regulated by cadmium following
transfection into human or mouse cells. Cell 1982: 29: 99–108.
3 Nover L (ed). Heat Shock Response. CRC Press: Boca Raton, 1991,
pp 167–220.
4 Baim SB, Labow MA, Levine AJ, Shenk T. A chimeric mammalian transactivator based on the lac repressor that is regulated
by temperature and isopropyl b-d-thiogalactoside. Proc Natl
Acad Sci USA 1991; 88: 5072–5076.
5 Braselmann S, Graninger P, Busslinger M. A selective transcriptional induction system for mammalian cells based on Gal4estrogen receptor fusion proteins. Proc Natl Acad Sci USA 1993;
90: 1657–1661.
6 Lee F, Mulliagen R, Berg P, Ringold G. Glucocorticoids regulate
expression of dihydrofolate reductase cDNA in mouse mammary tumour virus chimaeric plasmids. Nature 1981; 294: 228–
232.
7 Figge J et al. Stringent regulation of stably integrated chloramphenicol acetyl transferase genes by E. coli lac repressor in monkey cells. Cell 1988; 52: 713–722.
8 Gossen M, Bujard H. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl
Acad Sci USA 1992; 89: 5547–5551.
9 Gossen M et al. Transcriptional activation by tetracyclines in
mammalian cells. Science 1995; 268: 1766–1769.
10 No D, Yao T-P, Evans RM. Ecdysone-inducible gene expression
in mammalian cells and transgenic mice. Proc Natl Acad Sci USA
1996; 93: 3346–3351.
11 Wang Y, O’Malley BW Jr, Tsai SY, O’Malley BW. A regulatory
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
system for use in gene transfer. Proc Natl Acad Sci USA 1994;
91: 8180–8184.
Vegeto E et al. The mechanism of RU486 antagonism is dependent on the conformation of the carboxy-terminal tail of the
human progesterone receptor. Cell 1992; 69: 703–713.
Wegner M, Drolet DW, Rosenfeld MG. POU-domain proteins:
structure and function of developmental regulators. Curr Opin
Cell Biol 1993; 5: 488–498.
Gerber H-P et al. Transcriptional activation modulated by homopolymeric glutamine and proline stretches. Science 1994; 263:
808–811.
Seipel K, Georgiev O, Gerber H-P, Schaffner W. C-terminal
domain (CTD) of RNA-polymerase II and N-terminal segment
of the human TATA binding protein (TBP) can mediate remote
and proximal transcriptional activation, respectively. Nucleic
Acids Res 1993; 21: 5609–5615.
Witzgall R et al. Kid-1, a putative renal transcription factor: regulation during ontogeny and in response to ischemia and toxic
injury. Mol Cell Biol 1993; 13: 1933–1942.
Witzgall R et al. The Krüppel-associated Box-A (KRAB-A)
domain of zinc finger proteins mediates transcriptional
repression. Proc Natl Acad Sci USA 1994; 91: 4514–4518.
Greene LA, Tischler AS. Establishment of a noradrenergic clonal
line of rat adrenal pheochromocytoma cells which respond to
nerve growth factor. Proc Natl Acad Sci USA 1976; 73: 2424–2428.
Goodrich JA, Tjian R. TBP-TAF complexes: selectivity factors for
eukaryotic transcription. Curr Opin Cell Biol 1994; 6: 403–409.
Pugh BF. Mechanisms of transcription complex assembly. Curr
Opin Cell Biol 1996; 8: 303–311.
Oñate SA, Tsai SY, Tsai M-J, O’Malley BW. Sequence and
characterization of a coactivator for the steroid hormone receptor superfamily. Science 1995; 270: 1354–1357.
Kamei Y et al. A CBP integrator complex mediates transcriptional activation and AP-1 inhibition by nuclear receptors. Cell
1996; 85: 403–414.
Hörlein AJ et al. Ligand-independent repression by the thyroid
hormone receptor mediated by a nuclear receptor co-repressor.
Nature 1995; 377: 397–404.
Chen JD, Evans RM. A transcriptional co-repressor that interacts
with nuclear hormone receptors. Nature 1995; 377: 454–457.
Xu J et al. The extreme C-terminus of progesterone receptor contains a transcriptional repressor domain that functions through
a putative corepressor. Proc Natl Acad Sci USA 1996; 93:
12195–12199.
Goodrich JA et al. Drosophila TAFII40 interacts with both a VP16
activation domain and the basal transcription factor TFIIB. Cell
1993; 75: 519–530.
Kuhl DPA, Caskey CT. Trinucleotide repeats and genome variation. Curr Opin Gen Dev 1993; 3: 404–407.
Ross CA, McInnis MG, Margolis RL, Li S-H. Genes with triplet
repeats: candidate mediators of neuropsychiatric disorders.
Trends Neurosci 1993; 16: 254–260.
Ashley CT Jr, Warren ST. Trinucleotide repeat expansion and
human disease. Ann Rev Genet 1995; 29: 703–728.
Servadio A et al. Expression analysis of the ataxin-1 protein in
tissues from normal and spinocerebellar ataxia type 1 individuals. Nat Genet 1995; 10: 94–98.
Yazawa I et al. Abnormal gene product identified in hereditary
dentatorubral-pallidoluysian atrophy (DRPLA) brain. Nat Genet
1995; 10: 99–103.
Trottier Y et al. Cellular localization of the Huntington’s disease
protein and discrimination of the normal and mutated form. Nat
Genet 1995; 10: 104–110.
Burke JR et al. Huntingtin and DRPLA proteins selectively interact with the enzyme GAPDH. Nat Med 1996; 2: 347–350.
Kingston RE, Bunker CA, Imbalzano AN. Repression and activation by multiprotein complexes that alter chromatin structure.
Genes Dev 1996; 10: 905–920.
Baniahmad A et al. The tau4 activation domain of the thyroid
hormone receptor is required for release of a putative
Inducible regulation of gene expression
Y Wang et al
corepressor(s) necessary for transcriptional silencing. Mol Cell
Biol 1995; 15: 76–86.
36 Sauer F et al. Control of transcription by Krüppel through interactions with TFIIB and TFIIEb. Nature 1995; 375: 162–164.
37 Margolin J et al. Krüppel-associated boxes are potent transcriptional repression domains. Proc Natl Acad Sci USA 1994; 91:
4509–4513.
38 Deuschle U, Meyer WK-H, Thiesen H-J. Tetracycline-reversible
silencing of eukaryotic promoters. Mol Cell Biol 1995; 15:
1907–1914.
39 Wang Y, DeMayo F, Tsai SY, O’Malley BW. Ligand-inducible
and liver-specific target gene expression in transgenic mice. Nat
Biotechnol 1997 (in press).
40 Wang Y, O’Malley BW, Tsai SY. Inducible system designed for
future gene therapy. In: Tuan R (ed). Methods in Molecular
Biology, Vol. 63: Recombinant Proteins: Detection and Isolation Protocols. Humana Press: New Jersey, 1997 (in press).
441