Gene Therapy (2008), 1–10
& 2008 Nature Publishing Group All rights reserved 0969-7128/08 $30.00
www.nature.com/gt
ORIGINAL ARTICLE
Drug-inducible and simultaneous regulation
of endogenous genes by single-chain nuclear
receptor-based zinc-finger transcription factor
gene switches
L Magnenat1,2,3,4, LJ Schwimmer1,2,3 and CF Barbas III1,2,3
1
The Skaggs Institute for Chemical Biology, La Jolla, CA, USA; 2Department of Molecular Biology, The Scripps Research Institute,
La Jolla, CA, USA and 3Department of Chemistry, The Scripps Research Institute, La Jolla, CA, USA
Chemically inducible gene switches that regulate expression
of endogenous genes have multiple applications for basic
gene expression research and gene therapy. Single-chain
zinc-finger transcription factors that utilize either estrogen receptor homodimers or retinoid X receptor-a/ecdysone
receptor heterodimers are shown here to be effective
regulators of ICAM-1 and ErbB-2 transcription. Using activator
(VP64) and repressor (Krüppel-associated box) domains
to impart regulatory directionality, ICAM-1 was activated
by 4.8-fold and repressed by 81% with the estrogen
receptor-inducible transcription factors. ErbB-2 was activated
by up to threefold and repressed by 84% with the retinoid
X receptor-a/ecdysone receptor-inducible transcription
factors. The dynamic range of these proteins was similar to
the constitutive system and showed negligible basal regulation when ligand was not present. We have also demonstrated that the regulation imposed by these inducible
transcription factors is dose dependent, sustainable for
at least 11 days and reversible upon cessation of drug
treatment. Importantly, these proteins can be used in
conjunction with each other with no detectable overlap
of activity enabling concurrent and temporal regulation of
multiple genes within the same cell. Thus, these chemically
inducible transcription factors are valuable tools for spatiotemporal control of gene expression that should prove
valuable for research and gene therapy applications.
Gene Therapy advance online publication, 5 June 2008;
doi:10.1038/gt.2008.96
Keywords: nuclear receptor and zinc-finger-based transcription factors; drug-inducible; single-chain gene switches
Introduction
Artificial zinc-finger transcription factors (TFZFs) have
been used to discover, regulate and study genes in vitro
and in vivo.1,2 Zinc-finger (ZF) domains are modular and
can be combined to create new proteins of desired DNAbinding specificity. By taking comprehensive approaches
and using phage display selection strategies, we and
others have successfully prepared ZF domains that target
almost all GNN, ANN and CNN DNA triplets.3–7
Typically, designer TFZFs are composed of up to six ZF
domains; a six-finger ZF binds to 18 bp of DNA; this
allows recognition of a unique sequence within the
human genome.8
When combined with repressor or activator domains,
ZF proteins can be used to regulate transcription of
genes. For instance, to regulate the ErbB-2/HER-2 gene,
Correspondence: Professor CF Barbas III, Department of Molecular
Biology and Department of Chemistry, The Scripps Research
Institute, 10550 N Torrey Pines Road, BCC-550, La Jolla, CA
92037, USA.
E-mail: carlos@scripps.edu
4
Current address: Merck Serono Geneva Research Center, Geneva,
Switzerland.
Received 4 December 2007; revised 6 March 2008; accepted 27 April
2008
the E2C polydactyl ZF was designed to bind to a specific
sequence within the promoter. This ZF domain has been
fused to the herpes simplex VP16 transcriptional activation domain (VP64), transcriptional repression domains
like the Krüppel-associated box (KRAB),9 mad mSIN3
interaction domain10 and a nuclear localization signal.11,12 As an alternative to individual rational design
of ZF proteins for each target sequence, high-throughput
and genome-wide approaches have been developed to
directly isolate functional TFZFs from large libraries of
randomly shuffled ZF domains.13,14 Using this type of
rapid selection in mammalian cells, a regulator
of intercellular adhesion molecule-1 (ICAM-1), CD5431-TFZF, was isolated. The TFZF binds with high affinity
and specificity to a unique site of the ICAM-1 promoter.15
Most TFZFs have been constructed for constitutive
regulation of natural promoters driving reporter
genes.11,16,17 TFZFs have been expressed from constitutive
promoters transiently, retrovirally or by generating
integrated stable cell lines18–21 and have also been
expressed in vivo.22
To effectively study gene function and develop gene
therapies, tightly controlled and highly inducible regulation of the target gene is desirable through direct control
of the TFZF with a cell-penetrating drug. Inducible
transcription regulatory systems, or gene switches, are
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
2
of stable cell lines, but did, however, require the delivery
and expression of the two separate components, ideally
at equimolar concentrations. We previously reported
inducible gene switches based on nuclear hormone
receptor LBDs and E2C-TFZFs to demonstrate the
upregulation of the ErbB2 promoter by ligand-dependent transcription regulators in transient reporter luciferase assays.12 Homodimeric single-chain gene switches
provide the advantage of requiring delivering of only
one expression cassette, but require two identical target
sequences at a certain distance and orientation to each
other in the promoter (Figure 1d). More sophisticated
versions are monomeric proteins composed of a TFZF
controlled by two linker-associated estrogen receptor
LBDs (scER) or by the retinoid X receptor-a linked to the
ecdysone receptor (RE) (Figure 1e).
In this study, we demonstrate endogenous gene
regulation by monocistronically encoded single-chain
gene switches that avoid the problems inherent to the coexpression of heterodimers or transgenes and alleviate
the specific constraints that homodimers place on the
nature of the target site. By combining constitutive
retroviral expression in bulk cell populations and
single-chain gene switches targeting 18 unique base
pairs within DNA promoter sequences of endogenous
genes, we demonstrated simultaneous positive and
chimeric transregulators typically consisting of a DNAbinding domain (DBD) fused to a transcription effector
domain (ED) and a ligand-binding domain (LBD). These
chimeric transregulators dimerize upon drug addition
and eventually undergo conformational changes that
result in control of a target promoter (Figure 1). As these
systems have been based on naturally existing DBDs
of definite DNA sequence specificity, use has been
restricted to regulation of exogenously delivered transgenes or reporter genes23 (Figure 1a). Use of ZF-based
domains as the DBD of the chimeric transregulator
allows targeting of virtually any sequence in a promoter,
minimizing endogenous cross-reactivity and potentially
allowing regulation of any desired chromosomal promoter. The first ZF-based gene switches were under the
control of a tetracycline- or ecdysone-inducible response
element. However, this required co-expression of the
LBD transregulators and imposed the use of preengineered host cells or the generation of stable cell lines
for the regulation of endogenous genes18,19,24 or transgenes6,25 (Figure 1b). An advance came with the
development of systems that fused split TFZFs with
LBDs or dimerizer domains. This is typically achieved by
the drug-dependent reconstitution of functional transregulatory heterodimers that connect the DBD to the
ED12,26 (Figure 1c). These systems did not require use
Inducible TFZF expression
LB
D
LB
D
Inducible transgene expression
DBD
DBD
+inducer
+inducer
ED
ZFP
D
LB
D
LB
DBD
transgene
ED
Inducible TFZF heterodimer
Inducible TFZF homodimer
ED
3-ZFP
+inducer
ED
6-ZFP
+inducer
ED ED
+inducer
ED
3-ZFP
D
LB
D
LB
D
LB
ZFP
Inducible TFZF monomer
ED
LB
D
+
ZFP
LB
D
LB
D
ED
endogene
LB
D
ZFP
LB
D
DBD
3-ZFP
6- ZFP
Figure 1 Evolution of inducible transcription regulatory systems, also called gene switches. (a) Directly controlled inducible expression
systems require a specific DNA element (indicated by thick DNA helix) located upstream of a transgene of interest and engineered cells lines
that stably express a ligand-inducible regulatory protein, such as the tetracycline repressor. (b) Engineered zinc-finger transcription factors
(TFZFs) placed under the regulation of an inducible DNA element offer the opportunity for control of endogenous genes. (c) Heterodimericinducible systems provide drug-inducible regulation of the activity of split TFZFs fused to ligand-binding domains (LBDs) or dimerizer
domains without the use of a stable cell line. (d) Homodimeric inducible TFZF (iTFZF) gene switches, delivered by a single-chain protein, are
useful when two contiguous target sites are available. (e) Monomeric iTFZF-inducible systems conveniently associate the desired benefits of
the other systems, including the inducible regulation of endogenous genes using single-chain six-fingered iTFZF gene switches with unique
genome specificity and without cell-line restrictions.
Gene Therapy
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
negative regulation of transcription. Regulation was
dependent on the dose of hormone analog, was sustainable and reversible, and had a low basal and highly
inducible activity. We propose that these inducible TFZFs
(iTFZFs) will be valuable for use in vivo as spatiotemporal
endogenous gene regulation is key in treatment of
human diseases such as metastasized cancer.
Results
Constitutive ICAM-1 and ErbB2 endogenous gene
regulation
To establish a generic gene switch system for endogenous gene regulation, previously described designer sixZF proteins were used as DBD models. When fused to
VP64 activation or KRAB-repression domains, CD54-31
and E2C regulate the promoters of the ICAM-115 and
ErbB218 genes, respectively, with exquisite specificity.
Constitutive expression of the TFZFs from the pMX
retroviral vector in HeLa and A431 cells allowed for the
almost complete knockdown and a four- to sixfold
overexpression of each gene product in the whole-cell
populations; results were quantified using specific antitarget antibodies with either flow cytometry or western
blotting (Figure 2).
Inducible regulation of endogenous gene expression
by single-chain iTFZFs gene switches
Because the control of endogenous gene expression with
cell permeable drugs is highly desirable, iTFZFs endogenous gene switches were generated by subcloning the
TFZFs fused with single-chain hormone receptor LBDs
into the pMX vector. The N-terminal ZF DBD and the
C-terminal transcriptional ED were separated by dimeric
hormone-binding domains, consisting of two estrogen
receptor LBDs joined by an 18 amino-acid linker (scER)
or of the retinoid X receptor-a linked to the ecdysone
receptor with an 18 or a 30 residue long peptide (RE and
RLE, respectively) (Figure 1e). By fusion of the hormone
receptor into the TFZF, the activity of the artificial
transcription factors and consequently, the regulated
expression of the target genes were rendered druginducible. When expressed transiently, the prototype
ligand-dependent transcription factors E2C-scER-VP64,
E2C-RE-VP64 and E2C-RLE-VP64 were previously described to upregulate an ErbB2 promoter driving a
luciferase reporter gene in presence of 4-hydroxytamoxifen (4-OHT) and ponasterone-A (ponA).12 All combinations of CD54-31 and E2C six-ZFs, coupled with the
scER-, RE- and RLE-inducible domains to the VP64 and
KRAB EDs, were cloned into the retroviral pMX vector.
The resulting constructs (pMX-CD54-31-scER-VP64,
pMX-CD54-31-scER-KRAB,
pMX-CD54-31-RE-VP64,
pMX-CD54-31-RE-KRAB,
pMX-CD54-31-RLE-VP64,
pMX-CD54-31-RLE-KRAB, pMX-E2C-scER-VP64, pMXE2C-scER-KRAB, pMX-E2C-RE-VP64, pMX-E2C-REKRAB, pMX-E2C-RLE-VP64 and pMX-E2C-RLE-KRAB)
were transiently co-delivered into the 293-GagPol packaging cell line with a plasmid encoding the vesicular
stomatitis virus (VSV) pseudotyping viral envelope. The
cell supernatants containing the viral particles were
applied to HeLa or A431 cells for infection.
Whereas no induction was observed when the mocktransduced cells were treated with 4-OHT or ponA
(not shown), retrovirally delivered iTFZFs efficiently upand downregulated the expression of the chromosomal
ICAM-1 and the ErbB2 genes in presence of inducers
(Figure 3). Notably, the estrogen receptor-based CD54-31
scER iTFZFs controlled the ICAM-1 gene with virtually no
alteration of expression in the absence of drug and with
an average of 4.8-fold activation and 81% percent of
repression equivalent to or better than the constitutive
forms in the presence of drug (compare Figures 3a and b
with Figure 2a). Although the retinoid X/ecdysone
receptor gene switches regulated ICAM-1 in HeLa cells
less efficiently than the scER iTFZFs, these iTFZFs
increased ICAM-1 expression up to sixfold in A431 cells
in the presence of drug (data not shown). The E2C-RE
and E2C-RLE endogenous gene switches controlled the
ErbB2 target gene with low basal activity and good
inducibility, whereas the E2C-scER switches were ineffective (Figures 3c and d).
3
Reversible, sustainable and latent iTFZFs activity. To
investigate the robustness and the reversibility of the
inducible activity, cell-surface expression of ICAM-1 was
determined in HeLa cells transduced with CD54-31scER, -RE and -RLE iTFZFs fused with the VP64
transactivation domain (Figure 4a). If no drug was
applied, ICAM-1 expression was the same as in
untransduced cells. When cells were treated with
chemical inducer for 2 days and grown for 7 more days
in the presence of drug treatment, expression was
upregulated. When cells were treated with chemical
inducer for 2 days and grown for 7 more days in the
absence of drug treatment, expression was initially
upregulated and then returned to basal levels without
observable changes in cell morphology and proliferation.
When transduced cells were grown for 11 days and
then drug was added for 2 days, the iTFZF activity and
ICAM-1 upregulation were similar to that observed
when drug was added immediately upon infection
(Figure 4b, compare with Figure 3a). Accordingly,
the E2C-RE-VP64 and the E2C-RLE-VP64 iTFZFs
demonstrated a similar robustness in latent inducible
activity, regardless of the time of drug addition (that is,
after 1 or 11 days post-transduction) (Figure 4d, compare
to Figure 3c). These experiments demonstrated that,
when linked to the VP64 transactivation domain,
the scER-, RE- and RLE-based gene switches can be kept
in inactive or active states or reversed simply by
changing drug content in the cell culture media. ICAM1 repression in cells expressing CD54-31-KRAB iTFZFs
was almost fully reversible and showed similar repression after 9 days as after 2 days of drug stimulation
(data not shown; Figure 3b). A notable difference was
that for the most potent ICAM-1-inducible repressor,
CD54-31-scER-KRAB, only a portion (23%) of the
bulk-transduced cells were still expressing green fluorescent protein (GFP) and thus exhibiting complete
ICAM-1 repression (data not shown). Despite the
shortfall of cells expressing IRES-linked GFP over time,
CD54-31-scER-KRAB expressing cells were not presenting any ICAM-1 repression after an 11-day period
without drug, and after that, were still inducible to
reach 60% ICAM-1 repression (Figure 4c), with the most
GFP-positive cells displaying the most ICAM-1 repression (data not shown).
Gene Therapy
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
4
88%(+/-12) 3.4x(+/-0.4)
93%(+/-2.2) 4.1x(+/-1.4)
200
94%(+/-0.6)
0
HeLa
0
100
101
102 103
ICAM-1
104
6x(+/-0.4)
500
A431
HeLa
200
0
100
101
Key
102 103
ICAM-1
104
100
101
102 103
ErbB2
104
autofluorescence
mock transduced
KRAB
VP64
KRAB
VP64
ICAM-1 protein expression
(normalized to beta-actin)
HeLa
ICAM-1
b-actin
KRAB VP64 -
4.2x
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
97%
HeLa
OTF1
RVF
Ap-2
Sp1
E1A
Sp1
KRAB
Sp1
VP64
E2C
ErbB2
Ap-1
Ap-1/
Ap-1/
Ets-ARE Ets-ARE
C/EBPNF-κB AP-3
Ap-1 RARE
NF-κB
GAS/
IRE Ap-2
Ets-1
ICAM-1
C/EBP
CD54-31
Ets-1
Sp1
Figure 2 Positive and negative regulation of endogenous genes using constitutively expressed zinc-finger transcription factors (TFZFs).
Endogenous protein expression levels were measured in cells that expressed six zinc-finger proteins that bind to the promoters of
intercellular adhesion molecule-1 (ICAM-1) (CD54-31) or ErbB2 (E2C) fused to the Krüppel-associated box (KRAB) repressor or the VP64
transactivator domains. (a) Fluorescence-activated cell sorting (FACS) analysis of cell-surface expression of ICAM-1 (in HeLa and A431 cells)
and ErbB2 (in HeLa cells). The percentage of repression (black overlay) and fold activation (thick black overlay) using TFZFs compared to
mock-transduced cells (gray overlay) is calculated from background-subtracted mean fluorescence intensities (MFI) (dashed gray overlay).
The FACS histograms are representative of 3–4 independent experiments. Means of calculated fold activation and percent repression are
presented with standard deviation (in parenthesis). Statistical significance of observed differences with mock-transduced cells was
determined using the Student’s t-test, Po0.02. (b) Western analysis of ICAM-1 regulation in HeLa cells. Total protein extracts (125 mg) were
separated on denaturing SDS–polyacrylamide gel electrophoresis (PAGE) gel, blotted to a polyvinylidene fluoride (PVDF) membrane,
immunostained using specific antibodies and chemiluminescence detection. ICAM-1 expression in mock-transduced cells (HeLa) is
compared to CD54-31 repressor (KRAB)- and activator (VP64) TFZF-expressing cells, relative to constitutive protein (b-actin) and retroviral
TFZF expression. This western blot is a representative of two independent experiments. (c) Relative quantification of ICAM-1 bands
normalized to b-actin levels. Fold activation and percentage repression are indicated on top of overlays and bar graphs. (d) Diagrams of
ErbB2 and ICAM-1 promoters with locations of endogenous transcription factor-binding sites (black boxes) and the binding sites for the
corresponding zinc-fingers (striped boxes). Arrows indicate translation start sites and triangles indicate transcription start sites.
4-OHT and ponA dose-dependent endogenous gene
regulation. It is important that the endogenous gene
expression in a gene switch system depends on the drug
dose. To test this, HeLa cells retrovirally transduced with
iTFZF were treated with a range of concentrations of
inducer drugs and levels of ICAM-1 protein expression
were measured on the cell surface (by fluorescenceactivated cell sorting, FACS) and in extracts (by western
blotting) (Figure 5). The level of ICAM-1 expression on
cells that expressed CD54-31-scER-VP64 increased as the
concentration of 4-OHT was increased from 4 to 500 nM
Gene Therapy
to reach a maximal sixfold induction. Downregulation in
cells that expressed CD54-31-scER-KRAB was barely
measurable at 0.8 nM 4-OHT; at 500 nM 4-OHT, ICAM-1
expression was repressed by 73% and barely detectable
by western blot. Although less potent, the retinoid X/
ecdysone receptor-based gene switches were also dosedependent within the range of 100 nM to 5 mM ponA.
Simultaneous and independent inducible co-regulation
of ICAM-1 and ErbB2. When the regulation of two
endogenous genes is necessary to achieve a complete
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
5
500
100 101 102 103 104
100 101 102 103 104
100 101 102 103 104
ICAM-1
ICAM-1
ICAM-1
ICAM-1
4.8x
ICAM-1 repression (MFI)
100 101 102 103 104
ICAM-1
ICAM-1
6.0
5.0
- inducer
+ inducer
3x
4.0
2.6x
3.0
2.0
100
29%
80
35%
60
- inducer
+ inducer
40
81%
20
0
RE-VP64
scER-VP64
RE-VP64
RLE-VP64
0
1
2
3
4
10 10 10 10 10
ErbB2
0
0
1
2
3
4
10 10 10 10 10
3.5
ErbB2
- inducer
+ inducer
\
2.0
1.5x
1.5
1.0
ErbB2 repression (MFI)
3.0
2.2x
0
100 101 102 103 104
100 101 102 103 104
ErbB2
ErbB2
100
3x
2.5
0
100 101 102 103 104
ErbB2
ErbB2
RLE-KRAB
500
500
HeLa
0
0
1
2
3
4
10 10 10 10 10
RLE-KRAB
RE-KRAB
500
HeLa
HeLa
HeLa
0
0
RE-KRAB
scER-KRAB
500
500
500
scER-KRAB
RLE-VP64
HeLa
scER-VP64
HeLa
ICAM-1 activation (MFI)
0
100 101 102 103 104
120
1.0
ErbB2 activation (MFI)
0
100 101 102 103 104
7.0
500
500
0
0
0
RLE-KRAB
RE-KRAB
HeLa
500
HeLa
HeLa
0
scER-KRAB
RLE-VP64
HeLa
500
HeLa
500
RE-VP64
HeLa
scER-VP64
80
28%
60
- inducer
+ inducer
40
83%
83%
RE-KRAB
RLE-KRAB
20
0
scER-VP64
RE-VP64
RLE-VP64
scER-KRAB
Key
autofluorescence
mock transduced
without drug
with drug
Figure 3 Inducible up- and downregulation of endogenous genes by single-chain inducible zinc-finger transcription factors (iTFZFs) gene
switches containing VP64 transactivator or Krüppel-associated box (KRAB) repressor domains. HeLa cells retrovirally expressing monomeric
intercellular adhesion molecule-1 (ICAM-1) and ErbB2-specific estrogen (scER) and retinoid X/ecdysone receptor with a short or a long linker
(RE or RLE respectively) nuclear receptor-based TFZFs were treated with 4-hydroxytamoxifen (4-OHT) (100 mM) and ponasterone-A (ponA)
(5 mM), respectively. Cell-surface expression of target gene products without (dashed black overlay) or after drug induction (thick black
overlay) were measured by flow cytometry using specific antibodies and compared to normal protein levels of mock-transduced cells (gray
overlay) and cell autofluorescence (dashed gray overlay). The fluorescence-activated cell sorting (FACS) histograms are representative of 2–7
independent experiments. Bar graphs show the average and standard deviations of inducible up- and downregulation of the target gene
product by single-chain iTFZFs gene switches in the absence (white bars) or presence (black bars) of inducer. The calculated mean fluorescence
intensity (MFI) is presented in comparison to normal expression levels of mock-transduced cells, normalized to 1 in VP64-activator graphs
(a, c) and to 100 in KRAB-repressor graphs (b, d). Average of inducible fold activation and percentage repression are indicated on top of
bar graphs. Statistical significance of observed differences with mock-transduced cells was determined using the Student’s t-test, Po0.005.
(a) Inducible ICAM-1 upregulation with CD54-31 iTFZFs: scER-VP64, RE-VP64 and RLE-VP64. (b) Inducible ICAM-1 downregulation
with CD54-31 iTFZFs: scER-KRAB, RE-KRAB and RLE-KRAB. (c) Inducible ErbB2 upregulation with E2C iTFZF transactivators: scER-VP64,
RE-VP64 and RLE-VP64. (d) Inducible ErbB2 downregulation with E2C iTFZF repressors: scER-KRAB, RE-KRAB and RLE-KRAB.
phenotypic or therapeutic effect, it would be valuable to
combine two target-specific iTFZF with dependence on
different drugs. This would allow for either activation or
repression in a simultaneous or staggered fashion. To
demonstrate these possibilities, HeLa cells were cotransduced with the CD54-31 scER iTFZF and either
E2C-RE-KRAB or E2C-RLE-VP64 and then stimulated
with 4-OHT and ponA simultaneously (Figure 6). All
combinations of ICAM-1 and ErbB2 co-regulations were
achieved: Both genes could be chemically repressed or
overexpressed, one gene could be downregulated,
whereas the other was upregulated, and vice versa.
When only one drug was used, only the expected gene
was regulated (data not shown).
Gene Therapy
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
6
ICAM-1 activation (MFI)
4.0
3.5
3.5
2.9
3.0
- inducer
+ inducer
+/- inducer
2.5
2.5
2.0
1.4
1.4
1.3
1.5
1
1
1.1
RE-VP64
RLE-VP64
1.0
ICAM-1 activation (MFI)
scER-VP64
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
- inducer
+ inducer
scER-VP64
RE-VP64
4.0
ErbB2 activation (MFI)
ICAM-1 repression (MFI)
120
RLE-VP64
100
80
60
40
20
0
scER-KRAB
3.5
3.0
- inducer
+ inducer
2.5
2.0
1.5
1.0
RE-VP64
RLE-VP64
Figure 4 Reversible, sustained and latent induction of intercellular
adhesion molecule-1 (ICAM-1) and ErbB2 regulation by inducible
zinc-finger transcription factors (iTFZFs). (a) HeLa cells retrovirally
expressing CD54-31-scER, -RE and -RLE iTFZFs with VP64
transactivation domains were treated with inducer for 2 days,
grown for 7 more days with drug treatment retained (+ inducer) or
removed (± inducer) or were not treated at all for the equivalent
period of time ( inducer). Cell-surface expression of ICAM-1 was
measured by flow cytometry using an anti-ICAM-1 antibody, and
mean fluorescence intensity (MFI) was normalized to protein levels
of mock-transduced cells and cell autofluorescence (representative
of two independent experiments). (b–d) Latent induction of gene
regulation using iTFZFs. ICAM-1 iTFZFs were retrovirally expressed
into HeLa cells for 11 days without drug (), then drug was added
(+) for 2 days. At day 13 post-infection, the cell-surface expression
of the target gene products was monitored by fluorescenceactivated cell sorting (FACS). (b) ICAM-1 upregulation with
CD54-31-scER, -RE and -RLE with VP64 transactivator domain.
(c) ICAM-1 repression with CD54-31-scER-KRAB and (d) ErbB2
upregulation with E2C-RE-VP64 and E2C-RLE-VP64. The calculated MFI is presented in comparison to normal expression levels of
mock-transduced cells, normalized to 1 in VP64-activator graphs
and to 100 in Krüppel-associated box (KRAB)-repressor graphs.
Discussion
In this era of genomics and proteomics, the ability to
regulate gene expression with the flick of a switch would
provide a valuable tool for discovering novel gene
function and pathways. A toolbox of chemically inducible gene switches that provide tunable levels of gene
expression, which is both sustainable and reversible,
would ensure the necessary control needed for basic
biological studies, as well as, for gene therapy applications. Previous studies demonstrated that chemically
iTFZFs, prepared using single-chain estrogen receptor
Gene Therapy
homodimers and retinoid X/ecdysone receptor heterodimers with engineered LBDs that are not sensitive to
endogenous hormones, worked to regulate transcription
of reporter genes in response to bioavailable ligand
homologs that are considered safe.12 Dent et al.27 used the
progesterone receptor in conjunction with a ZF and the
p65 activation domain to regulate endogenous VEGF-A.
Here we have expanded the gene switch toolbox to show
that iTFZFs prepared using engineered estrogen receptor
and retinoid X/ecdysone receptor can both activate and
repress the transcription of endogenous genes.
Key features for inducible gene switches include
negligible basal regulation, a dynamic range of regulation that approaches the constitutive system, and
especially for gene therapy applications, dose-dependent
levels of regulation. These features are necessary to
ensure tight control of target gene transcription to more
closely mimic natural transcriptional control. For ICAM1 regulation, the scER iTFZF achieves these goals with 4.8fold activation, 81% repression and dose dependency
over a 1000-fold range of ligand concentration for an
overall ICAM-1 expression dynamic range of 25-fold
from down- to upregulation (Figures 3–5). Although the
results are not as dramatic for ErbB2 regulation, the RE
and RLE receptors show up to threefold activation, 84%
repression for a bidirectional induction ratio of 18-fold,
and dose dependency over a 50-fold range of ligand
concentration (Figures 3–5) The difference observed in
the inducible activation of the ICAM-1 and ErbB2 genes
with their respective iTFZFs might reflect different
efficacies and mechanisms of action between the scER
and RE switches. Different overexpression potential of
the two genes with constitutively expressed TFZFs was
also observed. Besides the ErbB2 gene amplification in
some tumor cell lines, the position and the affinity of the
regulators upstream or downstream of the starts of
mRNA transcription, the post-translational limitations in
the expression of the respective genes or their different
levels of basal expression could possibly result in
different inducibility potentials.
Sustainability and reversibility of gene regulation is
also important for effective inducible gene switches.
Stable, long-term expression of a therapeutic protein or
repression of a faulty gene is essential for gene therapy
applications and when studying the effects of gene
expression in tissue culture or animal models. Consistent
gene expression should be able to be maintained for the
duration of the treatment, through multiple passages of
tissue culture, or through the lifetime of the animal. Full
and rapid reversibility and a fine-tuning rheostat of gene
regulation is an important safety feature in the event of
adverse reactions in gene therapy. For animal models,
reversibility would be advantageous for investigating the
effects of activating or repressing a gene at a specific
stage of development. All of the iTFZFs showed reversible and sustained activation of ICAM-1 and the RE and
RLE iTFZFs also showed sustained activation for ErbB2.
Sustained repression of ICAM-1 was also effective with
repression after 9 days of drug stimulation similar to
repression after 2 days and 60% drug-inducible repression obtainable after 11 days latency without drug. The
observed loss of GFP expression could be caused by
retroviral expression silencing or negative selection due
to an antiproliferative effect of ICAM-1 repression or
iTFZF expression.
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
7
120
ICAM-1 repression (MFI)
ICAM-1activation (MFI)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.1
1
10
100
Inducer (nM)
1000
10000
100
scER
RE
RLE
80
60
40
20
0
0.1
scER-VP64 (nM 4-OHT)
HeLa
0
0.8
4
20
1
10
100
Inducer (nM)
1000
10000
scER-KRAB (nM 4-OHT)
100
500
HeLa
0
0.8
4
20
100
500
ICAM-1
β-actin
RE-VP64 (uM ponA)
HeLa
0
0.05
0.1
0.5
RLE-VP64 (uM ponA)
1
5
HeLa
0
0.05
0.1
0.5
1
5
ICAM-1
β-actin
Figure 5 Dose-dependent control of intercellular adhesion molecule-1 (ICAM-1) expression. CD54-31-scER-, -RE- and -RLE inducible zincfinger transcription factors (iTFZFs) with either -VP64 and -KRAB (Krüppel-associated box) domains were retrovirally expressed in HeLa cells
for 2 days. 4-Hydroxytamoxifen (4-OHT) was added at concentrations ranging from 0.8 to 500 nM or ponasterone-A (ponA) was added at
concentrations ranging from 0.05 to 5 mM. ICAM-1 expression was measured on the cell surface by fluorescence-activated cell sorting (FACS)
or in cell extracts by western blot (representative of 2–3 independent experiments). (a) Mean fluorescence intensity (MFI) quantification of
ICAM-1 activation (normalized to 1 for mock-transduced cells) and repression (normalized to 100) relative to drug dose. (b) Equal total
protein amounts of whole-cell extracts were separated on denaturing SDS–polyacrylamide gel electrophoresis (PAGE) with reducing
conditions and immunoblotted on polyvinylidene fluoride (PVDF) membranes for the detection of ICAM-1 regulation in comparison to
constitutive b-actin expression using specific antibodies. Protein expression levels in cells expressing CD54-31 iTFZF (-scER-VP64, -scERKRAB, -RE-VP64 and -RLE-VP64) treated with drug are compared to uninduced mock-transduced cells (HeLa).
In addition, we have shown that the scER and RE
iTFZFs can work in conjunction with each other to
independently control endogenous expression of two
genes (Figure 6). When both transcription factors and
ligands are present, the levels of up- and downregulation
of the target genes are equivalent to regulation levels
seen when they are used separately. When a single
ligand is present, only the target iTFZF shows gene
regulation. This control opens the possibility to regulate
transcription of two or more genes simultaneously and
independently, which could be useful when studying
protein–protein interactions or biological pathways, as
well as for experimental approaches in systems biology.
When combined with large polydactyl-ZF libraries such
as those used recently,13 iTFZF inducibility would be an
important asset for in vivo screening of new gene
modulators at a certain time in living organisms. It could
also prove advantageous to treat diseases by repressing
transcription of a detrimental protein and activating
transcription of a therapeutic protein.
The modularity of the ZFs, nuclear receptor ligandbinding domains and EDs lend themselves to a flexible
system in which the parts can be interchanged. Our
results show that the efficacy of iTFZFs varies depending
on target promoter and cell line (Figure 3). For example
in HeLa cells, the scER iTFZF works very well when
targeting the ICAM-1 promoter for both activation and
repression, however they were virtually ineffective at
regulating ErbB2 expression, whereas the RE- and RLEbased iTFZFs worked in the opposite manner. The scER
may not work with the ErbB2 promoter due to interactions between the estrogen receptor and ErbB2 signaling
pathways. Tamoxifen, which in an endogenous ER
antagonist, may be affecting AP-2 regulation of ErbB2
transcription.28 For the RE and RLE iTFZFs, there is a
retinoic acid-responsive element close to the CD54-31Gene Therapy
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
8
120
HeLa
120
0
0
100 101 102 103 104
ICAM-1
100 101 102 103 104
ICAM-1
120
HeLa
120
0
0
100
101
102
103
100 101 102 103 104
ErbB2
104
ErbB2
120
HeLa
120
0
0
100 101 102 103 104
ICAM-1
120
HeLa
120
100 101 102 103 104
ICAM-1
0
0
0
10
101
102
103
104
ErbB2
100 101 102 103 104
ErbB2
also observed different levels of inducible regulation in
A431 cells than in HeLa cells, indicating that CD54-31
iTFZFs with retinoid X/ecdysone-inducible domains
were functionally able to regulate ICAM-1 as well as
CD54-31-scER iTFZFs (data not shown). Cell-type specific
and TFZF-dependent activity has also been seen with
constitutive regulators of this type,15 and thus might
reflect differences with endogenous regulators. This
variation of regulatory ability demonstrates the need
for a variety of ZFs, ligand-inducible receptors and EDs
that can be easily assembled in different combinations to
optimize performance for the specific application.
Like the system developed by Dent et al.,27 the iTFZFs
presented in this paper are expressed as a single fusion
protein enabling simple gene delivery, which is ideal for
gene therapy applications. In addition to allowing for
inducible repression and using more specific six-fingered
DBDs, another advantage of our system is the formation
of intramolecular dimers, making dimerization with
endogenous nuclear receptors unlikely, although still a
possibility when response elements of endogenous
nuclear receptors are in close proximity to the iTFZF
target site. The single-chain construction also eliminates
the response element repeat requirement of natural
nuclear receptors. However, the effect that the ligands
and iTFZFs have on the expression and function of other
endogenous genes in various cell and tissue types will
have to be studied to identify the iTFZF with the least
potential for side effects.
In conclusion, we have created inducible transcription
factors that can activate and repress the expression of
endogenous genes. Regulation was shown to be sustainable, reversible, dose dependent, have a dynamic range
similar to constitutive regulators, have negligible basal
regulation and can be used in conjunction with each
other. These iTFZFs are valuable tools for genomic and
proteomic research and gene therapy applications, where
spatiotemporal control of gene expression is essential.
Key
autofluorescence
no drugs
only the drug relevant to the other target
both drugs
Figure 6 Simultaneous inducible positive and negative control of
two genes within the same cells. HeLa cells co-transduced with
different retroviral combinations of estrogen-receptor-based CD5431 and retinoid X/ecdysone receptor-based E2C-inducible zincfinger transcription factors (iTFZFs) were induced with both drugs
(thick black overlay) or with only the drug relevant to the iTFZF of
the other target gene (black overlay), and protein expression levels
were compared to cells not treated with either drug (gray overlay)
and cell autofluorescence (dashed gray overlay) (representative of
two independent experiments). (a) Up and down: CD54-31-scERVP64 and E2C-RE-Krüppel-associated box (KRAB). (b) Double
down: CD54-31-scER-KRAB and E2C-RE-KRAB. (c) Double up:
CD54-31-scER-VP64 and E2C-RLE-VP64. (d) Down and up: CD5431-scER-KRAB and E2C-RLE-VP64.
binding site in the ICAM-1 promoter.29–32 Interaction
with the retinoic acid receptor that binds here may affect
regulation of ICAM-1 expression by the RE and RLE
switches. On the other hand, compared to untreated
cells, 4-OHT and ponA treatment did not influence the
expression of the ErbB2 and ICAM-1 promoters, respectively (compare gray and black overlays in Figure 6). We
Gene Therapy
Materials and methods
Human cell lines
A431 epidermoid carcinoma cells and HeLa cervix
carcinoma cells were obtained from the American Type
Culture Collection (Manassas, VA, USA). The 293-GagPol
packaging cell line was obtained from I Verma (Salk
Institute, San Diego, CA, USA). Cells were cultured in
Dulbecco’s modified Eagle’s medium supplemented
with 10% fetal calf serum and antibiotics.
Construction of retroviral plasmids for the expression
of iTFZFs
The expression cassettes of pcDNA E2C monomeric
switch constructs, those with C-terminal EDs used in the
transient assays,12 were subcloned into the pMX-IRESGFP retroviral vector33 using the following restriction
digest and ligation strategy. The single-chain estrogen
receptor constructs pcDNA E2C ScER/30-VP64 and
-KRAB (30 aa linker) were digested and ligated into
pMX E2C-VP64 vector using BamH1-Pac1 restriction
enzymes. The single-chain retinoid/ecdysone receptor
constructs pcDNA E2C-RE-VP64 and KRAB (18 residue
linker), and E2C-RLE-VP64 and -KRAB (30 residue
linker) were digested with Fse1-Pac1 and assembled in
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
a one-step two-way ligation with a BamH1-Fse1 E2C
fragment into the pMX E2C-VP64 vector previously
digested with BamH1 and Pac1. The ICAM-1 gene switch
constructs were created by replacing the E2C by the CD5431 six-finger coding DNA using the Sfi1 restriction enzyme.
Retroviral delivery and inducible iTFZFs assay
Retroviral transductions were performed using 293GagPol packaging cells and VSV envelope pseudotyping
to produce retrovirus as previously described15 with the
following modifications. With constitutive TFZFs, target
gene expression was measured 72 h after the retrovirus
was applied to the cells. Unless otherwise stated in the
figure legends, the host cells transduced with the
inducible iTFZFs were split in two dishes 24 h after
transduction, and while one-half of the cells was kept
uninduced to measure the basal activity of the iTFZFs, the
other half was supplemented with 5 mM ponA (Invitrogen, Carlsbad, CA, USA) or 100 nM 4-OHT (Sigma,
St Louis, MO, USA) for the induction of the iTFZFs based on
the estrogen and retinoid X/REs, respectively. After 48 h,
the cells were harvested for analysis of the target gene
expression by FACS or western blotting. Transduction
efficiency and iTFZFs expression was verified by flow
cytometry with the percentage of cells indicating GFP
fluorescence, which is co-expressed on the same bicistronic transcript as the iTFZFs. Using these conditions, the
bulk (480%) of the A431 or HeLa cells were expressing
GFP after 3 days transduction, which persisted up to 19
days (not shown). The pcDNA3.1 vector (Invitrogen) was
used as mock transduction control for the determination
of the basal target gene expression. The addition of drug
inducer to the negative control cell sample did not alter
the cell-surface expression of the targets (data not
shown).
Flow cytometric analysis of inducible target gene
regulation
At day 3 after cell transduction, cell-surface expression of
the target genes was measured with anti-ICAM-1 (clone
HA58; BD Pharmingen, San Diego, CA, USA) and antiErbB2 (clone FSP77, gift from Nancy E Hynes, Friedrich
Miescher Institute, Basel34) monoclonal antibodies at
5 mg ml1 and with a phycoerythrin-labeled donkey antimouse immunoglobulin Gsecondary antibody (Jackson
Immunochemicals, Westgrove, PA, USA) using a flow
cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).
Comparative analysis of mean fluorescence intensities
(MFI) was determined using CELLQuest software
(Becton Dickinson, Franklin Lakes, NJ, USA). MFI was
corrected for background by subtraction of the autofluorescence of cells incubated with the secondary
antibody only and normalized to the expression value
of mock-transduced cells.
Western blotting
For measuring protein expression by western blotting,
whole-cell extracts were prepared with cell lysis buffer
(50 mM Tris HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100)
containing 1 protease inhibitor cocktail (Roche, Basel,
Switzerland). For each sample, the total protein concentration was determined with a Bradford assay (Bio-Rad,
Hercules, CA, USA). Equal amounts of protein were
separated on reducing 4–12% bis-tris polyacrylamide
gels with 2-(N-morpholinolethane sulfonic acid (MES)
SDS buffer and transferred to a Hybond-P polyvinylidene fluoride (PVDF) membrane (Amersham, Piscataway, NJ, USA) according to the instructions of the
manufacturer (Invitrogen). For normalization with a
constitutive protein, membranes were treated in blocking
buffer (PBS, 0.1% Tween 20 (Sigma) with 5% blocking
agent (Amersham), incubated with anti-b-actin (clone
AC-15; Sigma), washed in phosphate-buffered saline
tween-20 (PBST) and detected with a secondary antibody
conjugated to horseradish peroxidase. For the detection
of the target protein and TFZF, the membranes were
stripped in 62.5 mM Tris-HCl (pH 6.8), 2% SDS and
100 mM b-mercaptoethanol for 30 min at 65 1C, washed
in PBST and treated with blocking buffer containing
10 mg ml1 extravidin and 1 mg ml1 a-biotin (both
Sigma). ICAM-1 was immunostained with a biotinylated
anti-ICAM-1 (BAF720; R&D Systems, Minneapolis, MN,
USA) and TFZF proteins with biotinylated anti-influenza
hemagglutinin tag (clone 3F10; Roche), respectively.
Membranes were then washed and incubated with
streptavidin conjugated to horseradish peroxidase
(Amersham) in PBST. Alternatively, for the western blots
of the dose–response experiments, the anti-ICAM-1
primary antibody was detected with a peroxidasecoupled donkey anti-sheep secondary antibody (Jackson
Immunochemicals). The blots were revealed with the
ECL Plus kit according to the manufacturer’s instructions (Amersham) and exposed to an X-OMAT film
(Kodak) or quantified using a Storm 860 phosphorimager
(Molecular Dynamics) using the blue chemifluorescence
scanning mode. The intensity of the bands corresponding to the respective size of the proteins on reducing
gel was quantified using the ImageQuant software
(Molecular Dynamics).
9
Acknowledgements
This study was supported in part by the National
Institute of Health Grant GM065059 to CFB. LM was
the recipient of postdoctoral fellowships from the Swiss
National Science Foundation. LJS is supported by
The American Cancer Society Illinois Division—Linda
M Campbell Postdoctoral Fellowship.
References
1 Beerli RR, Barbas III CF. Engineering polydactyl zinc-finger
transcription factors. Nat Biotechnol 2002; 20: 135–141.
2 Jamieson AC, Miller JC, Pabo CO. Drug discovery with
engineered zinc-finger proteins. Nat Rev Drug Discov 2003; 2:
361–368.
3 Segal DJ, Dreier B, Beerli RR, Barbas III CF. Toward controlling
gene expression at will: selection and design of zinc finger
domains recognizing each of the 50 -GNN-30 DNA target
sequences. Proc Natl Acad Sci USA 1999; 96: 2758–2763.
4 Dreier B, Segal DJ, Barbas III CF. Insights into the molecular
recognition of the 50 -GNN-30 family of DNA sequences by zinc
finger domains. J Mol Biol 2000; 303: 489–502.
5 Dreier B, Beerli RR, Segal DJ, Flippin JD, Barbas III CF.
Development of zinc finger domains for recognition of the 50 ANN-30 family of DNA sequences and their use in the
construction of artificial transcription factors. J Biol Chem 2001;
276: 29466–29478.
Gene Therapy
Endogenous regulation of multiple genes with TFZFs
L Magnenat et al
10
6 Liu Q, Xia Z, Zhong X, Case CC. Validated zinc finger protein
designs for all 16 GNN DNA triplet targets. J Biol Chem 2002;
277: 3850–3856.
7 Dreier B, Fuller RP, Segal DJ, Lund CV, Blancafort P, Huber A
et al. Development of zinc finger domains for recognition of the
50 -CNN-30 family DNA sequences and their use in the construction of artificial transcription factors. J Biol Chem 2005; 280:
35588–35597.
8 Liu Q, Segal DJ, Ghiara JB, Barbas III CF. Design of
polydactyl zinc-finger proteins for unique addressing
within complex genomes. Proc Natl Acad Sci USA 1997; 94:
5525–5530.
9 Margolin JF, Friedman JR, Meyer WK, Vissing H, Thiesen HJ,
Rauscher III FJ. Kruppel-associated boxes are potent transcriptional repression domains. Proc Natl Acad Sci USA 1994; 91:
4509–4513.
10 Ayer DE, Laherty CD, Lawrence QA, Armstrong AP, Eisenman
RN. Mad proteins contain a dominant transcription repression
domain. Mol Cell Biol 1996; 16: 5772–5781.
11 Beerli RR, Segal DJ, Dreier B, Barbas III CF. Toward controlling
gene expression at will: specific regulation of the erbB-2/HER-2
promoter by using polydactyl zinc finger proteins constructed
from modular building blocks. Proc Natl Acad Sci USA 1998; 95:
14628–14633.
12 Beerli RR, Schopfer U, Dreier B, Barbas CFI. Chemically
regulated zinc finger transcription factors. J Biol Chem 2000;
275: 32617–32627.
13 Blancafort P, Magnenat L, Barbas CF. Scanning the human
genome with combinatorial transcription factor libraries. Nat
Biotechnol 2003; 21: 269–274.
14 Park KS, Lee DK, Lee H, Lee Y, Jang YS, Kim YH et al.
Phenotypic alteration of eukaryotic cells using randomized
libraries of artificial transcription factors. Nat Biotechnol 2003; 21:
1208–1214.
15 Magnenat L, Blancafort P, Barbas III CF. In vivo selection of
combinatorial libraries and designed affinity maturation of
polydactyl zinc finger transcription factors for ICAM-1 provides
new insights into gene regulation. J Mol Biol 2004; 341: 635–649.
16 Bartsevich VV, Juliano RL. Regulation of the MDR1 gene by
transcriptional repressors selected using peptide combinatorial
libraries. Mol Pharmacol 2000; 58: 1–10.
17 Corbi N, Libri V, Fanciulli M, Tinsley JM, Davies KE, Passananti C.
The artificial zinc finger coding gene ‘Jazz’ binds the utrophin
promoter and activates transcription. Gene Therapy 2000; 7:
1076–1083.
18 Beerli RR, Dreier B, Barbas III CF. Positive and negative
regulation of endogenous genes by designed transcription
factors. Proc Natl Acad Sci USA 2000; 97: 1495–1500.
19 Zhang L, Spratt SK, Liu Q, Johnstone B, Qi H, Raschke EE et al.
Synthetic zinc finger transcription factor action at an endogenous chromosomal site. Activation of the human erythropoietin
gene. J Biol Chem 2000; 275: 33850–33860.
20 Liu PQ, Rebar EJ, Zhang L, Liu Q, Jamieson AC, Liang Y et al.
Regulation of an endogenous locus using a panel of designed
Gene Therapy
21
22
23
24
25
26
27
28
29
30
31
32
33
34
zinc finger proteins targeted to accessible chromatin regions.
Activation of vascular endothelial growth factor A. J Biol Chem
2001; 276: 11323–11334.
Ren D, Collingwood TN, Rebar EJ, Wolffe AP, Camp HS.
PPARgamma knockdown by engineered transcription factors:
exogenous PPARgamma2 but not PPARgamma1 reactivates
adipogenesis. Genes Dev 2002; 16: 27–32.
Rebar EJ, Huang Y, Hickey R, Nath AK, Meoli D, Nath S et al.
Induction of angiogenesis in a mouse model using engineered
transcription factors. Nat Med 2002; 8: 1427–1432.
Toniatti C, Bujard H, Cortese R, Ciliberto G. Gene therapy
progress and prospects: transcription regulatory systems. Gene
Therapy 2004; 11: 649–657.
Xu D, Ye D, Fisher M, Juliano RL. Selective inhibition of
P-glycoprotein expression in multidrug-resistant tumor cells by
a designed transcriptional regulator. J Pharmacol Exp Ther 2002;
302: 963–971.
Kang JS, Kim JS. Zinc finger proteins as designer transcription
factors. J Biol Chem 2000; 275: 8742–8748.
Pollock R, Giel M, Linher K, Clackson T. Regulation of
endogenous gene expression with a small-molecule dimerizer.
Nat Biotechnol 2002; 20: 729–733.
Dent CL, Lau G, Drake EA, Yoon A, Case CC, Gregory PD.
Regulation of endogenous gene expression using small molecule-controlled engineered zinc-finger protein transcription
factors. Gene Therapy 2007; 14: 1362–1369.
Perissi V, Menini N, Cottone E, Capello D, Sacco M,
Montaldo F et al. AP-2 transcription factors in the regulation
of ERBB2 gene transcription by oestrogen. Oncogene 2000; 19:
280–288.
Chadwick CC, Shaw LJ, Winneker RC. TNF-alpha and 9-cisretinoic acid synergistically induce ICAM-1 expression: evidence
for interaction of retinoid receptors with NF-kappa B. Exp Cell
Res 1998; 239: 423–429.
Cilenti L, Toniato E, Ruggiero P, Fusco C, Farina AR, Tiberio A et
al. Transcriptional modulation of the human intercellular
adhesion molecule gene I (ICAM-1) by retinoic acid in
melanoma cells. Exp Cell Res 1995; 218: 263–270.
Janssens S, Bols L, Vandermeeren M, Daneels G, Borgers M,
Geysen J. Retinoic acid potentiates TNF-alpha-induced ICAM-1
expression in normal human epidermal keratinocytes. Biochem
Biophys Res Commun 1999; 255: 64–69.
Aoudjit F, Bosse M, Stratowa C, Voraberger G, Audette M.
Regulation of intercellular adhesion molecule-1 expression by
retinoic acid: analysis of the 50 regulatory region of the gene. Int J
Cancer 1994; 58: 543–549.
Liu X, Constantinescu SN, Sun Y, Bogan JS, Hirsch D, Weinberg
RA et al. Generation of mammalian cells stably expressing
multiple genes at predetermined levels. Anal Biochem 2000; 280:
20–28.
Harwerth I-M, Wels W, Marte BM, Hynes NE. Monoclonal
antibodies against the extracellular domain of the erbB-2
receptor function as partial ligand agonists. J Biol Chem 1992;
267: 15160–15167.
Gene Therapy (2008) 15, 1246
& 2008 Macmillan Publishers Limited All rights reserved 0969-7128/08 $30.00
www.nature.com/gt
ERRATUM
Drug-inducible and simultaneous regulation
of endogenous genes by single-chain nuclear
receptor-based zinc-finger transcription factor
gene switches
L Magnenat, LJ Schwimmer and CF Barbas III
Gene Therapy (2008) 15, 1246; doi:10.1038/gt.2008.136
Correction to: Gene Therapy (2008) 15, 1223–1232;
doi:10.1038/gt.2008.96
Following the advanced online publication of this article,
the authors noticed that one of the concentrations
was printed incorrectly in the legend of Figure 3.
The correct concentration used for the treatment with
4-hydroxytamoxifen (4-OHT) is (100 nM).
The publisher apologizes for this error.