The Plant Journal (2002) 32, 1077–1086
TECHNICAL ADVANCE
Controlling gene expression in plants using synthetic zinc
finger transcription factors
Justin T. Stege1,y, Xuen Guan2, Thao Ho2, Roger N. Beachy3 and Carlos F. Barbas III1,
1
Department of Molecular Biology and The Skaggs Institute for Chemical Biology, Scripps Research Institute, 10550
North Torrey Pines Road, La Jolla, CA 92037,
2
Torrey Mesa Research Institute, 3115 Merryfield Row, San Diego, CA 92121, and
3
Donald Danforth Plant Science Center, 975 North Warson Road, St. Louis, MO 63132, USA
Received 3 September 2002; revised 30 September 2002; accepted 7 October 2002.
For correspondence (fax þ1 858 784 2583; e-mail carlos@scripps.edu).
y
Present address: Diversa Corporation, 4955 Director’s Place, San Diego, CA 92121, USA.
Summary
Synthetic zinc finger proteins can be fused to transcriptional regulatory domains to create artificial transcription factors that modulate the expression of a specific target gene. Recent studies have demonstrated
that synthetic zinc finger domains can be constructed to bind DNA sequences with a high degree of
specificity. To devise a general strategy for controlling plant gene expression with artificial transcription
factors, a rapid transient assay was developed to test the regulatory activity of synthetic zinc finger transcription factors (effectors) on target plasmids (reporters) in plant cells. Effective activation was demonstrated with zinc finger proteins fused to a derivative of the VP16 activation domain. The mSin3 interaction
domain (SID) of the human MAD1 protein provided moderate repression of target reporters. Unlike many
naturally occurring transcription factors, these synthetic effectors exhibit a strong dependence on binding
site position. Reporter genes that are stably integrated into plant cells responded similarly to transiently
transfected reporter plasmids, verifying that this assay accurately reflects the behavior of these transcription factors on an endogenous target within the context of chromosomal DNA. These results provide
evidence that synthetic zinc finger proteins can be used to manipulate the expression of endogenous genes
in plants.
Keywords: artificial transcription factors, gene regulation, transcriptional activation, transcriptional repression, zinc finger protein, plant gene expression.
Introduction
Studies of transcription factors have shown that they are
typically designed in modular fashion with a DNA binding
domain and a transcriptional regulatory domain (Ayer et al.,
1996; Margolin et al., 1994; Triezenberg et al., 1988). These
experiments also demonstrated that the regulatory
domains can function when fused to different DNA binding
domains to generate chimeric transcription factors with
altered gene specificity. In a typical transcription factor
protein, the DNA binding domain directs the activity of
the regulatory domain to the specific genetic locus to alter
the expression of the target gene. Thus, it is possible to
generate a synthetic transcription factor by fusing a genespecific DNA binding domain to a transcriptional regulatory
ß 2002 Blackwell Publishing Ltd
domain. Recent advances in synthetic zinc finger protein
(ZFP) research have led to the design of target-specific DNA
binding proteins and construction of transcription factors
with novel specificity (Beerli and Barbas, 2002). Once developed, this technology can be used to design artificial transcription factors to control the expression of any plant
gene. The application of this technology ranges from purely
scientific studies to the development of valuable agronomic
traits in crop plants.
Zinc finger proteins have been shown to bind specific
DNA sequences with high affinity and act as the DNA
binding domain of transcription factors (Cook et al., 1999;
Margolin et al., 1994). The Cys2His2 family of zinc finger
1077
1078 Justin T. Stege et al.
DNA binding domains has been shown to be particularly
amenable to rational manipulation to modify binding specificity and several methods have been developed to construct zinc finger DNA binding domains specific for a given
target sequence (Choo and Klug, 1994; Desjarlais and Berg,
1992; Dreier et al., 2001; Liu et al., 1997; Rebar and Pabo,
1994; Wu et al., 1995). Each finger motif is about 30 amino
acids (AA) long and contains cysteine and histidine residues that coordinate a zinc ion to stabilize the folding of this
simple bba domain. Using current approaches, it is possible
to construct 6-finger (polydactyl) proteins that specifically
bind targeted 18-bp DNA sequences (Beerli et al., 2000;
Beerli et al., 1998; Dreier et al., 2001; Liu et al., 1997; Ren
et al., 2002).
Studies of a ZFP constructed to bind to an 18-bp target
within the human erbB-2 promoter (Beerli et al., 1998) have
explored the efficacy of a variety of activator and repressor
domains in mammalian cells. These studies led to the
development of synthetic transcription factors that significantly increased or decreased the expression of the endogenous human erbB-2 and erbB-3 genes (Beerli et al., 2000).
Several other artificial transcriptional factors, constructed
by similar methods, have also been shown to regulate
endogenous genes (Dreier et al., 2001; Liu et al., 2001;
Ren et al., 2002; Zhang et al., 2000), demonstrating the
potential application of transcription factors customized
for a specific gene. However, before this technology can
be exploited as a robust tool for controlling gene expression in plants, a number of parameters need to be examined. Some of the factors that could influence the efficacy of
a transcription factor include: DNA binding site location
within the target promoter; choice of regulatory domain;
specificity of regulation; and non-target effects.
To date, the studies of regulation of gene expression by
synthetic ZFP transcription factor proteins have been performed exclusively in mammalian cells. The DNA binding
activity of the ZFP is expected to function in plants, but
interactions with endogenous transcriptional machinery
may be different than in mammalian cells. Because most
research on generating chimeric transcription factors has
been done in mammalian and yeast systems, it will be
necessary to identify plant-specific regulatory regions that
can function when fused to zinc finger DNA binding
domains. The herpes simplex virus VP16 domain
(Sadowski et al., 1988) and its derivatives have been shown
to function in plants and other systems to activate expression when fused to the Gal4 DNA binding domain (Estruch
et al., 1994). It is likely that the VP16 domain will function as
an effective transcriptional activation domain when fused
to ZFPs and expressed in plants.
The identification and characterization of repressors of
transcription has proven to be a more difficult task. To date,
no plant repressors have been identified that behave in a
modular fashion. However, it is possible that repression
domains already identified in other systems may function
well in plants. The mSin3 interaction domain (SID) (Ayer
et al., 1996) and the KRAB domain (KRAB) (Margolin et al.,
1994) have been fused to ZFPs to generate effective transcription factors for use in mammalian cells (Beerli et al.,
2000; Beerli et al., 1998). The SID domain acts by recruiting a
histone deacetylase complex that ultimately modifies local
chromatin structure (Eilers et al., 1999). The KRAB domain
acts by binding the KAP-1 co-repressor, which in turn
serves as a scaffold to recruit proteins for histone deacetylation and methylation (Schultz et al., 2002). These activities recruited by the KRAB domain are thought to modify
the chromatin structure in such a way as to repress transcription.
The goal of this study was to demonstrate the use of
synthetic ZFPs in artificial transcription factors to control
gene expression in plants. Furthermore, experiments were
performed to test the specificity of regulation and explore
the effects of binding site location, promoter strength and
different regulatory domains on gene regulation by chimeric transcription factors. A rapid assay was developed to
test the effects of zinc finger transcription factors (effectors)
on the expression of a target gene (reporter) in tobacco and
maize protoplasts by transient expression of an effector
protein and a reporter plasmid. In tobacco, the effector
protein was expressed using a TMV-based vector that
provided extremely high transfection efficiency and protein
expression. Various effector constructs were made with the
well-studied C7 ZFP (Liu et al., 1997; Wu et al., 1995). The C7
zinc finger contains three zinc finger domains and binds
nine contiguous bases. Linking two C7 proteins together
generated a 6-finger 2C7 protein, which has been shown to
bind specifically and tightly to 18 contiguous nucleotides
(Liu et al., 1997). The 2C7 zinc finger was used to construct
different effector proteins and tested against a variety of
reporter constructs to investigate the behavior of these
novel zinc finger transcription factors and to develop a
strategy for regulating targeted genes in plants.
Results
Transient activation of luciferase reporter in plant cells
Activation by chimeric transcription factors that contain the
VP16 activation domain has been previously demonstrated
in plants (Moore et al., 1998). In the present study, effector
proteins were constructed by fusing the 3-finger C7 ZFP
or the 6-finger 2C7 protein to four tandem repeats of
the minimal VP16 activation domain (Beerli et al., 1998)
(Figure 1b). These effectors were expressed using a transient expression vector derived from TMV. Effectors were
tested against a luciferase (LUC) gene reporter that contains
a minimal promoter with six repeats of the 2C7 binding site
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
Zinc finger transcription factors in plants
1079
Figure 1. Transient assay with reporter plasmids and effector proteins.
(a) A series of reporter constructs were constructed from derivatives of the CsVMV promoter or a minimal TATA-box containing promoter. The 6 2C7 binding
sites (small arrows) were inserted at different locations relative to the start of transcription (indicated above the promoter). Each of these versions of promoters
was used to drive the expression of luciferase (striped box). A control reporter plasmid was constructed using the cauliflower mosaic virus 35S promoter to drive
the expression of Renilla luciferase.
(b) Effector protein constructs were made by fusing ZFPs to activation or repression domains. The 3-finger C7 ZFP or the 6-finger 2C7 ZFP (angled stripes) or e2c
ZFP (vertical stripes) were fused to the VP64 (black box) activator or the SID (squares) and KRAB (checkered) repression domains and cloned into a TMV vector. A
C-terminal hemagglutinin (HA) epitope tag was added for immunodetection and affinity purification.
(c) The effector proteins were evaluated against the reporter plasmids in a transient assay. Viral RNA was made from the effector constructs and co-transfected
with the target reporter and control reporter plasmids into protoplasts. (1) Effector protein expressed from the viral RNA and (2) binds to target sequences on the
reporter plasmid, (3) affecting expression of the LUC reporter (4) while the RL reporter plasmid serves as an internal control for normalization of activity. The LUC
and RL activities resulting from different effector and reporter combinations are used to evaluate the effector activity (activity of reporter in the presence of
effector relative to the reporter activity without effector present).
(Figure 1a). Tobacco BY-2 protoplasts were co-transfected
with viral RNA encoding the effector proteins and reporter
plasmid DNA. To measure the activity of the effector proteins on reporter expression, the LUC activity is normalized
to Renilla luciferase (RL) activity that was elicited by the
control plasmid (pCAT-RL, Figure 1a). The relative activity is
calculated by dividing the normalized LUC activity obtained
with effector expression by the normalized LUC activity
without any effector expression.
The 2C7-VP64 (a 6-finger ZFP) activates this minimal
promoter over 20-fold relative to the unactivated level,
while the 3-finger C7-VP64 effector provides over 80-fold
activation (Figure 2a). There are two possible explanations
why the 3-finger construct promotes higher levels of
activation. First, the reporter construct contains 12 possible
binding places for the 3-finger ZFP C7-VP64 and only six
possible binding places for the 6-finger ZFP C7-VP64. Thus,
C7-VP64 may bring a greater number of activation domains
to the target and activate transcription more effectively.
Alternatively, the 3-finger ZFP C7-VP64 may accumulate to
higher levels than the 6-finger ZFP 2C7. It has been
observed that 3-finger ZFPs tend to accumulate to higher
levels of protein than 6-finger ZFP when expressed in
mammalian cells (C. Barbas, unpublished data).
We also performed similar activation experiments in
maize cells. In this system, the 2C7-VP64 activator was
expressed with the ubiquitin promoter (Cornejo et al.,
1993) on a plasmid construct. The 2C7-VP64 activator
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
1080 Justin T. Stege et al.
Figure 2. Activation of a reporter gene in tobacco BY2 cells with 3- or 6-finger ZFP.
(a) Activation of prbTATA and pC7rbTATA reporters (Figure 1a) following co-transformation
with C7-VP64 (black bar) or with 2C7-VP64
(white bar) activators (Figure 1b).
(b) Activation of p50 C7F and pC7rbTATA reporter plasmids (Figure 1b) in maize protoplasts
with 2C7-VP64 activator protein.
was co-transformed with two different reporters with minimal promoters containing the C7 binding sites (Figure 1a).
The 2C7-VP64 protein activated the p50 C7F reporter 146-fold
and the pC7rbTATA reporter 40-fold (Figure 2b). This indicates that ZFP transcription factors perform well in monocot as well as dicot cells.
of transcription at the very 30 end of the promoter (p30 C7A),
the 2C7 effector binding seemed to inhibit expression, even
when fused to the VP64 activator. The binding of multiple
effector molecules in this location may interfere with the
Position dependence of activation
To determine if the location of the zinc finger binding site
within the target promoter affects the function of these
effector proteins, a series of luciferase reporters were constructed using the constitutive Cassava Vein Mosaic Virus
(CsVMV) promoter (Verdaguer et al., 1996) with 6 2C7
binding sites located at different positions relative to the
transcriptional start site (TATA box) (Figure 1a). The activation of these reporters by the C7-VP64 and 2C7-VP64 activators was tested in tobacco BY-2 protoplasts (Figure 3). As
the C7 binding site is moved to a greater distance from the
TATA box, activation is significantly reduced. In the p50 C7A
reporter, the 6 2C7 binding sites are 443 bp away from the
start of transcription; however, there is very little activation
with this construct. In the p50 C7F promoter, the binding
sites are only 63 bp away from the TATA box and luciferase
expression is activated about fivefold by the 2C7-VP64
protein. When the binding site is positioned past the start
Figure 3. Position dependence of reporter activation on binding site location.
Activation of luciferase following co-transformation of each of the seven
CsVMV reporters (Figure 1a) with 3-finger activator C7-VP64 (black bar) or
with 6-finger activator 2C7-VP64 (white bar). Fold activation relative to the
activity observed with no effector protein is listed under each column.
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
Zinc finger transcription factors in plants
activity of the RNA polymerase complex and subsequently
reduce the expression of the reporter gene.
The inherent strength of the target promoter may affect
its regulation by exogenous transcription factors. p50 C7F
and pC7DE have the zinc finger binding site in the same
location relative to the TATA box in the promoter (Figure 1a), and pC7DE has an additional fragment of the
CsVMV promoter that is missing from the p50 C7F construct.
This results in much higher basal luciferase activity with the
pC7DE reporter than with the p50 C7F reporter (data not
shown). When activated by C7-VP64 or 2C7-VP64 effectors,
the stronger pC7DE reporter is activated to a lesser degree
than the weaker p50 C7F reporter (Figure 3). Furthermore,
when a minimal promoter (pC7rbTATA) with a very low
basal activity is used, very high level of activation was
observed (Figure 2a). The basal levels of p50 C7F and pC7DE
luciferase activity (LUC/protein) exhibited more than 100fold difference (data not shown), but the activated levels by
these reporters varied by less than fivefold. This may
indicate that there is a maximal level of expression for this
promoter/activator system.
Repression of the luciferase reporter
Unlike other systems, regulatory domains that repress
transcription in plants when fused to different DNA binding
domains have not been identified. However, the application
of foreign transcriptional activator domains in planta indicates that repression domains from other systems may
function in plants. We tested two repressor domains isolated from mammalian transcriptional regulators: SID (Ayer
et al., 1996) and KRAB (Margolin et al., 1994). Both domains
were fused to the N-terminus of the 2C7 ZFP (Figure 1b). As
a control for non-specific repression, a non-binding ZFP
(e2c) was fused to the SID domain. The repressor and
reporter constructs were then co-transformed into tobacco
BY2 cells to determine their effects on transcription. The
2C7-SID exhibited fivefold repression of the p50 C7F and
pC7rbTATA reporters (Figure 4). This level of repression
is approximately as effective in plant cells as in mammalian
cells (Beerli et al., 1998). The 2C7-KRAB effector shows no
specific repression. These results are contrary to the situation in mammalian cells where the KRAB domain proved to
be a more effective repressor than the SID domain (Beerli
et al., 1998). In our experiments, some degree of nonspecific repression was observed with the SID repressor.
The e2c-SID protein caused moderate levels of repression
with the p50 C7F and pC7rbTATA reporters, even though
there are no specific binding sites for the e2c protein in
these constructs (Figure 4).
As with activation, it appears that stronger promoters
may be more difficult to repress than weak promoters. The
strong pC7DE reporter is repressed at a lower level than the
weaker p50 C7F reporter (Figure 4).
1081
Figure 4. Repression of reporters with 6-finger ZFPs. The e2c-SID, 2C7-SID,
and 2C7-KRAB effector proteins (Figure 1b) were expressed in tobacco BY-2
protoplasts in combination with three reporter plasmids, pC7DE, pC7rbTATA, and p50 CF (Figure 1a). The levels of repression to reporter pC7DE, p50 CF,
and pC7rbTATA were represented as white, striped, and black bar, respectively. Fold activation relative to the activity observed with no effector
protein is listed under each column.
Specificity of regulation
In developing a system to regulate gene expression, it is
important that the desired effects are specific to the target
gene. The specificity of each of these effector constructs
was investigated using an internal control plasmid. In these
experiments, a control plasmid expressing the Renilla luciferase (pCAT-RL) was co-transfected with the effector and
reporter plasmids. Because the pCAT-RL control plasmid
has no 2C7 zinc finger binding sites, the RL activity should
be unaffected by the expression of 2C7 transcription factors. Thus, the pCAT-RL was co-transfected along with the
target reporter plasmid to serve as an internal control for
transformation efficiency. However, if there is non-specific
activity of these effectors, the RL activity may be affected.
Instead of normalizing luciferase activity to RL activity, the
LUC and RL activities were normalized to protein levels in
the cell extracts. These protein-normalized activities were
then divided by the activities obtained in the no effector
condition to generate a measure of relative activity. A
perfectly specific activator should have a relative RL/protein
of 1.0 and the relative LUC/Protein and LUC/RL should be
the same. For an activator with non-specific activity, LUC/
protein normalization should reflect the total level of activation (non-specific and specific), RL/protein should reflect
non-specific activation, and LUC/RL should measure only
specific activation.
When expressed with the p50 C7F reporter, all of the
effectors tested in this study had some level of non-specific
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
1082 Justin T. Stege et al.
Table 1 Specificity of zinc finger transcription factors
Controls
LUC/Protein
RL/Protein
LUC/RL
Repression
Activation
No effector
GFP control
C7-GFP
e2c-SID
2C7-SID
2C7-KRAB
e2c-VP64
2C7-VP64
C7-VP64
1.00
1.00
1.00
1.38
0.98
1.44
2.01
1.62
1.55
0.83
1.31
0.63
0.20
1.00
0.22
0.52
0.47
1.18
1.96
0.84
2.44
8.11
1.99
5.25
13.88
1.68
10.11
activity (Table 1). Even the expression of C7-GFP (C7 ZFP
fused to GFP) appeared to affect a non-specific activation of
38%. It is suggested that binding of the ZFP upstream of this
promoter causes modification of DNA structure, allowing
more efficient initiation of transcription. For some of the
effector constructs, the e2c (Beerli et al., 1998) zinc finger
was used as a control. The e2c zinc finger binds specifically
to the erbB-2 promoter and should not bind in these reporter plasmids. In our experiments, the e2c-SID effector
caused modest repression (17%) of the target LUC activity
(Table 1). However, the 2C7-SID effector represses the luciferase target reporter but not the RL control reporter fivefold, indicating that it is a specific repressor. The 2C7-KRAB
effector causes significant (twofold) repression of both the
luciferase target reporter and the RL control reporter indicating that it is non-specific. In Figure 3, repression by the
2C7-KRAB effector protein was calculated by normalizing
LUC to RL activity. From these calculations, one would
conclude that there was no repression by the 2C7-KRAB.
However, analyzing both the LUC and RL activities separately indicates that there is a non-specific repression of
both reporters. These calculations illustrate the need to
accurately characterize the specificity of transcription factors for non-target effects.
Effector constructs designed to activate transcription
used the VP64 activation domain. The non-target e2cVP64 effector exhibits some non-specific activation of the
target reporter and the 2C7-VP64 effector activates the RL
expression to a low degree (Table 1). However, the strong
activation of p50 C7F luciferase activity by both the 2C7-VP64
and the C7-VP64 effector protein indicates that this is a
relatively specific activator.
Regulating expression of integrated versus
non-integrated genes
One concern of using a transient assay to test synthetic ZFPbased artificial transcription factor is that the results may
not be applicable to endogenous gene targets. It can be
argued that the chromatin structure of a transiently transfected plasmid will be different from that of an endogenous
sequence located on a chromosome. Thus, the binding and
regulation by these effector proteins may behave differently on endogenous targets. This is especially true for
some repression domains that are thought to function by
altering the structure of the chromatin around their target
gene. To address this issue, transgenic tobacco BY-2 cell
lines were established with the pC7DE target reporter genes
stably integrated into the plant genome. Transient expression assays were performed with protoplasts derived from
these cell lines to test the activation and repression of an
integrated target by synthetic ZFP-based artificial transcription factors expressed from the TMV vector. The integrated
reporter appears to respond to the effector constructs in
essentially the same manner as when the cells are cotransfected with the plasmid reporter (Figure 5). Thus, a
transiently transfected plasmid target behaves similarly to
an integrated target. This finding suggested that this rapid
co-transformation assay can be reliably used to characterize new effector designs and putative regulatory domains.
Discussion
The experiments described in this paper demonstrate that
artificial transcriptional factors constructed with synthetic
ZFPs can activate and repress expression of a target gene.
Combined with the ability to create ZFP domains to bind
any DNA sequence, it may be possible to manipulate the
expression of endogenous plant genes with this approach.
However, there are many issues to be resolved before this
technology can be used as a general method to alter the
expression of any given target gene in plants. The selection
of DNA binding domains and transcriptional regulatory
domains will influence the efficacy, specificity, and reliability of the artificial ZFP transcription factors. Thus, it is
important to evaluate the behavior of different effector
constructs in studies to regulate endogenous genes.
In the many published studies of chimeric transcription
factors, very little has been done to address the issue of
non-specific regulation of non-target genes (Ayer et al.,
1996; Estruch et al., 1994; Hsieh et al., 1999; Jones et al.,
1998; Margolin et al., 1994; Nan et al., 1998). The specificity
of any effector protein that is used to regulate gene expression is of significant concern if they are to be used in gene
therapy or in agriculture. Therefore, it will be very important
to identify regulatory proteins and effector designs that
provide effective regulation of the target gene without
significantly affecting expression of non-target genes.
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
Zinc finger transcription factors in plants
1083
Figure 5. Repression and activation of a cotransformed plasmid reporter versus integrated
reporter.
A BY-2 cell line containing the reporter gene
was generated with the pC7DE reporter stably
integrated into the genome. Protoplasts were
prepared and transfected in the same manner as
in the co-transformation condition except no
luciferase reporter plasmid was added to the
transgenic cells. C7-GFP, e2c-SID, and e2c-VP64
effectors (Figure 1b) were used as controls that
should have little or no regulatory activity. The
levels of activation or repression achieved
through co-transformation of the reporter gene
are shown as white bars and the values with the
integrated reporter are shown as black bars.
Fold activation relative to the activity observed
with no effector protein is listed under each
column. All assays were performed in triplicate.
The specificity of DNA binding for synthetic ZFP has been
shown to be very high in vitro (Segal et al., 1999). It remains
to be seen how in vitro DNA binding specificity correlates
with in vivo regulatory behavior. A 6-finger polydactyl ZFP
binds a contiguous 18-bp DNA sequence and if the ZFP is
specific for those 18 bp, the chances of another target site
existing elsewhere in the genome is low, even in the large
genomes of humans and higher plants. In our studies, a 3finger ZFP effector protein provided more activation than a
similar 6-finger protein. However, a 3-finger ZFP will bind a
9-bp DNA target, a factor that is expected to increase the
likelihood of binding to multiple genomic locations, and
increases the likelihood of non-specific effects of the ZFP.
Clearly, the specificity of binding of the ZFP to the target
DNA will be a major determinant of the specificity of the
transcriptional regulation. However, it is possible that the
choice of regulatory domains could also affect specificity. In
these experiments, almost all of the effector proteins,
including proteins that lack regulatory domains, exhibited
some level of activity. It is possible that the level of effector
protein expression could influence the degree of regulation. All the experiments in this study used the TMV viral
vector for expression of the effector proteins that can lead
to high levels of protein production in plant cells (Padgett
et al., 1996). Lowering the level of effector expression may
result in more specific regulation. Therefore, it may be
important to tightly control effector expression to achieve
optimum regulation with minimal non-target effects.
The optimal location of the site for ZFP binding within
a target gene is not well characterized. While naturally
occurring transcription factors can act over long or
short distances (Gallie, 1998; Stein et al., 2000), the experiments described here indicate that ZFP-based artificial
transcription factors are highly dependent on the distance
of the binding site from the start of transcription. Maximum
regulation of a strong viral promoter (CsVMV) was obtained
when the binding site was located as close to the start of
transcription as possible. However, when the binding site
was placed downstream of the transcriptional start site,
regulation was not effective.
It is possible that modification of transcription factor
design could affect position dependence. For example,
inserting a long flexible linker of amino acids between
the DNA binding domain and the regulatory domain may
allow the regulatory domain more freedom to influence
transcription from a greater distance. Also, other activation
domains may function differently and have better activity
from a distance. Position dependence may seriously limit
the possibilities when designing new ZFP-based artificial
transcriptional factors. On a positive note, this may imply
that zinc finger effectors will not affect transcription of
endogenous genes that are adjacent to the target gene.
The effect of promoter strength on regulatory activity was
examined in this study by using a number of derivatives of
the CsVMV promoter. These results indicate that stronger
promoters are harder to regulate than weaker ones. To
demonstrate that the ZFP technology can be used as a
general method for modifying gene expression in plants,
the promoter dependence of effector proteins needs to be
evaluated. We are currently targeting several different promoters since every gene is likely to have a unique set of
control elements that respond differently to certain regulatory domains. If the regulatory domains currently being
used exhibit strong promoter dependence, it may be necessary to search for other domains with more general regulatory activity.
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
1084 Justin T. Stege et al.
When fused to the 2C7 zinc finger, the VP64 regulatory
domain provides effective, specific activation of target
genes in both monocot and dicot plant cells. This is not
too surprising since VP16 activation domains have been
successfully demonstrated to function in plant systems
(Moore et al., 1998). Thus, the VP64 is a suitable choice
for constructing artificial transcriptional activators based
on synthetic ZFPs. However, there may be other plantspecific activation domains that work as well or better than
VP64.
Identifying effective repression domains for use in constructing artificial transcriptional repressors based on synthetic ZFP is a greater challenge. To date, no plant proteins
have been identified that contain a domain that can serve as
a modular transcriptional repressor when fused to a DNA
binding domain. Two mammalian repression domains, SID
and KRAB, showed only limited function in plants. The
KRAB domain produced moderate, non-specific repression
in plants cells, while the SID domain provided moderate,
but specific repression in plant cells. It is possible that lower
levels of KRAB effector protein expression could increase
the specific repression. Homology searches reveal no close
KAP-1 homologs in public plant sequence databases (data
not shown). Because the KRAB domain mediates repression by binding the KAP-1 co-repressor in mammalian cells,
the lack of a similar protein will prevent KRAB from functioning in plant cells. Recent studies have identified a
number of plant proteins capable of repressing transcription in plants (Jin et al., 2000; Raventos et al., 1998). The
transient assay developed here could be used to quickly
identify and characterize any modular repression domains
in these proteins that can function in a chimeric transcription factor. Basic research and genome sequencing efforts
may reveal new proteins and homologs of existing repressors that could also be evaluated in this assay system.
Clearly, much information is required in order to define
an effective strategy for modifying gene transcription in
plants and there is a need to develop a rapid assay to test
these various parameters. This study describes the development of a transient assay in plant cells to test the function
of effector proteins on the transcription of target reporter
genes. Different effectors and reporter configurations were
tested in this assay to investigate the behavior of these
novel zinc finger transcription factors and to develop a
strategy for targeted gene regulation in plants. In this study,
reporter target genes transiently transfected as a plasmid
into plant cells were shown to respond to effector proteins
in a similar manner to target genes integrated into chromosomal DNA. Two recent reports (Guan et al., 2002; Ordiz
et al., 2002) demonstrated the stable, heritable regulation of
endogenous plant genes by ZFP transcription factors and
found that transient expression of effector proteins provides similar regulation as stable expression. Combined
with the experiments described here, these results support
the idea that effector regulatory function observed in this
transient assay are predictive of the behavior of the ZFP
transcription factor in stably transformed plants.
Continuing efforts using transient assays like the one
developed in this study will identify the regulatory domains
and effector designs that provide the optimal combination
of activity, specificity, position and promoter dependence.
Combined with inducible or tissue specific effector expression, this technique could provide highly precise control
over the expression of a target gene. The results of these
experiments indicate that this technology will be a powerful
tool for basic plant research and agricultural biotechnology.
Experimental Procedures
Reporter constructs
The promoter from CsVMV (Verdaguer et al., 1996; Verdaguer et al.,
1998) and its derivatives were used to express a reporter gene
encoding the luciferase enzyme. Six copies of the binding site
(6 2C7) for the C7 ZFP were inserted at various locations within
the promoter. Plasmid pAluc contains the full length CsVMV
promoter with no C7 binding sites. p50 C7A has the 6 2C7 binding
site at the 50 end of the full-length CsVMV promoter. p30 C7A has the
6 2C7 binding site at the 30 end of the full length CsVMV promoter
near the start of translation for the luciferase gene. pC7DE contains
the full length CsVMV promoter with 6 2C7 binding site replacing
nucleotides 112 to 63. The 6 2C7 binding site was placed at the
50 end of CsVMV promoter fragments of different lengths to generate p50 C7C (222 to þ72), p50 C7D (178 to þ72), and p50 C7F
(112 to þ72). prbTATA is a minimal promoter with a TATA box
but without a ZFP binding site (actagtgctagcccgggtatataataagcttggcattccggtactgttggtaaagccaccatg). pC7rbTATA is identical
to prbTATA except that there is a 6 2C7 binding site at the 50 end
of the minimal promoter. pCAT-RL was constructed as a control,
which contains the 35S promoter driving expression of RL.
Activator and repressor constructs
Effector proteins were constructed in the same manner used by
Beerli et al. (1998) and blunt end cloned into a TMV expression
vector under the control of the coat protein subgenomic promoter
(Padgett et al., 1996). Plasmid p2C7-SID consists of the 6-finger 2C7
ZFP fused to the C-terminus of the Sin3 interaction domain (SID,
amino acid sequence of 1–36) (Ayer et al., 1996). Plasmid p2C7KRAB contains the 6-finger 2C7 ZFP fused to the C-terminus KRAB
domain (amino acid sequence of 1–97) (Margolin et al., 1994).
Plasmid p2C7-VP64 contains the 6-finger 2C7 ZFP fused to the
N-terminus of four tandem repeats of the minimal VP16 activation
domain (VP64) (Beerli et al., 1998). Plasmid pC7-VP64 is 3-finger C7
ZFP fused to the VP64 activation domain. The control vector pC7GFP contains the 3-finger ZFP C7 fused to the green fluorescent
protein (GFP). The 2C7-VP64 coding sequence was also cloned 50 of
the maize ZmUbi (Cornejo et al., 1993) promoter for expression in
maize protoplasts.
Tobacco BY2 and maize HE89 cell line maintenance
Cell lines were maintained at 288C in the dark in shaking flasks
with MS-based tissue culture medium (2.2 g l1 Murashige and
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
Zinc finger transcription factors in plants
Skoog Plant Salt Base - Gibco BRL #11117–074, 0.1 g l1
myo-inositol, 1 mg l1 thiamin-HCl, 0.2 mg l1 2,4-D, 10 g l1
sucrose, 0.4 M mannitol, pH 5.8) and subcultured about every
7 days.
Transient transfection of BY2 and HE89 protoplasts
BY2 cells were transfected with purified DNA or RNA by electroporation essentially as described by Watanabe et al. (1987) and
incubated at 288C in dark for 24–72 h. Two reactions were performed for each treatment. Cells were collected by centrifugation
and lysed by a freeze-thaw method in 80 ml of 1.2x Passive Lysis
Buffer (Promega). HE89 protoplasts were prepared using methods
similar to those for BY-2 cells and were transfected by PEG precipitation (Chourey and Zurawski, 1981). About half million protoplasts were used for each treatment. Ten micrograms plasmid
DNA or viral RNA was used in transient transfection assays of
tobacco BY-2 protoplasts.
Viral RNA transcription
The effector proteins pC7-GFP, p2C7-SID, p2C7-KRAB, pC7-VP64,
and p2C7-VP64 were cloned into a TMV expression plasmid (Padgett et al., 1996). One microgram of plasmid DNA was digested
with KpnI and cleaned with phenol:chloroform or via Qiaquick PCR
purification (Qiagen). RNA was transcribed from the template DNA
with the T7 Megascript Kit (Ambion). The concentration of resulting RNA was adjusted to 4 mg ml1 with MOPS (10 mM, pH 7.0) and
stored at 808C.
Luciferase assay
The luciferase assay was performed using 20 ml of cell lysate with
100 ml of luciferin assay solution (Promega) using an EG&G Wallac
luminometer with a 4-sec delay and a 15-sec integration of emitted
light. In the dual luciferase assays, 20 ml of cell lysate was assayed
for LUC and RL using the Dual Luciferase System (Promega). One
hundred microliters of LARII luciferin buffer was injected to a well
followed by a 1.6-sec delay and 10-sec integration of emitted light.
This measures the LUC activity. The LUC activity is quenched by
addition of 100 ml Stop and Glow Buffer (Promega). After a 4-sec
delay, the RL activity is measured by a 10-sec integration of emitted
light.
Generation of stable reporter cell lines
BY2 protoplasts were prepared for electroporation as described above and 2 106 cells were electroporated with 1–10 mg
of the pBin19 (Bevan, 1984) binary transformation vector
(pbinC7DE). Protoplasts were incubated at 288C in 15 ml protoplast culture media (Watanabe et al., 1987) for 48 h. Cells
were collected by centrifugation at 300 g for 2 min and plated
on BY-2 media (2.2 g l1 Murashige and Skoog Plant Salt Base Gibco BRL #11117–074, 0.1 g l1 myo-inositol, 1 mg l1 thiaminHCl, 0.2 mg l1 2,4-D, 10 g l1 sucrose, 0.4 M mannitol, 0.6%
Agar, pH 5.8) containing 50 mg ml1 Kan and incubated at 288C
in dark. Clumps of transformed cells were visible in 3–4 weeks
and were transferred to fresh plates. When clumps were sufficiently large, they were assayed for luciferase activity and for
the presence of plasmid integration by PCR. Suspension cell
cultures were developed from positive events for use in transient
assays.
1085
Acknowledgements
This study was supported by the National Institutes of Health
grants (CA27489 and DK61803) and funding from the Torrey Mesa
Research Institute to C.F.B. and by a grant from the Department of
Energy to R.N.B. (DE-FG02–99ER20355). J.T.S. was supported from
a postdoctoral NRSA post graduate fellowship from the National
Institutes of Health (1 F32 G19830-01).
References
Ayer, D.E., Laherty, C.D., Lawrence, Q.A., Armstrong, A.P.
and Eisenman, R.N. (1996) Mad proteins contain a dominant transcription repression domain. Mol. Cell. Biol. 16,
5772–5781.
Beerli, R.R. and Barbas, C.F., III (2002) Engineering polydactyl zincfinger transcription factors. Nat. Biotechnol. 20, 135–141.
Beerli, R.R., Dreier, B. and Barbas, C.F., III (2000) Positive and
negative regulation of endogenous genes by designed transcription factors. Proc. Natl Acad. Sci. USA, 97, 1495–1500.
Beerli, R.R., Segal, D.J., Dreier, B. and Barbas, C.F., III (1998)
Towards 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, 95, 14628–14633.
Bevan, M. (1984) Binary Agrobacterium vectors for plant transformation. Nucleic Acids Res. 12, 8711–8721.
Choo, Y. and Klug, A. (1994) Toward a code for the interactions of
zinc fingers with DNA: selection of randomized fingers displayed
on phage [published erratum appears in Proc. Natl Acad. Sci.
USA 1995, January 17; 92 (2): 646]. Proc. Natl Acad. Sci. USA, 91,
11163–11167.
Chourey, P.S. and Zurawski, D.B. (1981) Callus formation from
protoplasts of a maize cell culture. Theor. Appl. Genet. 59, 341–
344.
Cook, T., Gebelein, B., Belal, M., Mesa, K. and Urrutia, R. (1999)
Three conserved transcriptional repressor domains are a defining feature of the TIEG subfamily of Sp1-like zinc finger proteins.
J. Biol. Chem. 274, 29500–29504.
Cornejo, M.J., Luth, D., Blankenship, K.M., Anderson, O.D. and
Blechl, A.E. (1993) Activity of a maize ubiquitin promoter in
transgenic rice. Plant Mol. Biol. 23, 567–581.
Desjarlais, J.R. and Berg, J.M. (1992) Toward rules relating zinc
finger protein sequences and DNA binding site preferences.
Proc. Natl Acad. Sci. USA, 89, 7345–7349.
Dreier, B., Beerli, R.R., Segal, D.J., Flippin, J.D. and Barbas, C.F., III
(2001) 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. 276,
29466–29478.
Eilers, A.L., Billin, A.N., Liu, J. and Ayer, D.E. (1999) A 13-amino
acid amphipathic alpha-helix is required for the functional interaction between the transcriptional repressor Mad1 and mSin3A.
J. Biol. Chem. 274, 32750–32756.
Estruch, J., Crossland, L. and Goff, S. (1994) Plant activating
sequences: positively charged peptides are functional as transcriptional activation domains. Nucl Acids Res. 22, 3983.
Gallie, D.R. (1998) Controlling gene expression in transgenics.
Curr. Opin. Plant Biol. 1, 166–172.
Guan, X., Stege, J., Kim, M., Dahmani, Z., Fan, N., Heifetz, P.,
Barbas, C.F., III Briggs, S.P. (2002) Heritable endogenous
gene regulation in plants with designed polydactyl zinc
finger transcription factors. Proc. Natl Acad. Sci. USA, 99,
13296–13301.
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086
1086 Justin T. Stege et al.
Hsieh, J., Zhou, C., Young, D. and Hayward, S. (1999) CIR, a corepressor linking the DNA binding factor CBF1 to the histone
deacetylase complex. Proc. Natl Acad. Sci. USA, 96, 23–28.
Jin, H., Cominelli, E., Bailey, P., Parr, A., Mehrtens, F., Jones, J.,
Tonelli, C., Weisshaar, B. and Martin, C. (2000) Transcriptional
repression by AtMYB4 controls production of UV-protecting
sunscreens in Arabidopsis. EMBO J. 19, 6150–6161.
Jones, P.L., Veenstra, G.J., Wade, P.A., Vermaak, D., Kass, S.U.,
Landsberger, N., Strouboulis, J. and Wolffe, A.P. (1998) Methylated DNA and MeCP2 recruit histone deacetylase to repress
transcription. Nat. Genet. 19, 187–191.
Liu, P.Q., Rebar, E.J., Zhang, L. et al. (2001) Regulation of an
endogenous locus using a panel of designed zinc finger proteins
targeted to accessible chromatin regions. Activation of vascular
endothelial growth factor A. J. Biol. Chem. 276, 11323–11334.
Liu, Q., Segal, D.J., Ghiara, J.B. and Barbas, C.F., III (1997) Design of
polydactyl zinc finger proteins for unique addressing within
complex genomes. Proc. Natl Acad. Sci. USA, 94, 5525–5530.
Margolin, J.F., Friedman, J.R., Meyer, W.K., Vissing, H., Thiesen,
H.J. and Rauscher, F.J.R. (1994) Kruppel-associated boxes are
potent transcriptional repression domains. Proc. Natl Acad. Sci.
USA, 91, 4509–4513.
Moore, I., Galweiler, L., Grosskopf, D., Schell, J. and Palme, K.
(1998) A transcription activation system for regulated gene
expression in transgenic plants. Proc. Natl Acad. Sci. USA, 95,
376–381.
Nan, X., Ng, H.H., Johnson, C.A., Laherty, C.D., Turner, B.M.,
Eisenman, R.N. and Bird, A. (1998) Transcriptional repression
by the methyl-CpG-binding protein MeCP2 involves a histone
deacetylase complex [see comments]. Nature, 393, 386–389.
Ordiz, M.I., Barbas, C.F., III Beachy, R.N. (2002) Regulation of
transgene expression in plants with polydactyl zinc finger transcription factors. Proc. Natl Acad. Sci. USA, 99, 13290–13295.
Padgett, H.S., Epel, B.L., Kahn, T.W., Heinlein, M., Watanabe, Y.
and Beachy, R.N. (1996) Distribution of tobamovirus movement
protein in infected cells and implications for cell-to-cell spread of
infection. Plant J. 10, 1079–1088.
Raventos, D., Skriver, K., Schlein, M., Karnahl, K., Rogers, S.W.,
Rogers, J.C. and Mundy, J. (1998) HRT, a novel zinc finger,
transcriptional repressor from barley. J. Biol. Chem. 273,
23313–23320.
Rebar, E.J. and Pabo, C.O. (1994) Zinc finger phage: affinity selection of fingers with new DNA-binding specificities. Science, 263,
671–673.
Ren, D., Collingwood, T.N., Rebar, E.J., Wolffe, A.P. and Camp,
H.S. (2002) PPARgamma knockdown by engineered transcription factors: exogenous PPARgamma2 but not PPARgamma1
reactivates adipogenesis. Genes Dev. 16, 27–32.
Sadowski, I., Ma, J., Triezenberg, S. and Ptashne, M. (1988) Gal4Vp16 is an unusually potent transcriptional activator. Nature
(London), 335, 563–564.
Schultz, D.C., Ayyanathan, K., Negorev, D., Maul, G.G. and
Rauscher, F.J. (2002) SETDB1: a novel KAP-1-associated histone
H3, lysine 9-specific methyltransferase that contributes to HP1mediated silencing of euchromatic genes by KRAB zinc finger
proteins. Genes Dev. 16, 919–932.
Segal, D.J., Dreier, B., Beerli, R.R. and Barbas, C.F., III (1999)
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, 96,
2758–2763.
Stein, G.S., van Wijnen, A.J., Stein, J.L., Lian, J.B., Montecino, M.,
Choi, J., Zaidi, K. and Javed, A. (2000) Intranuclear trafficking of
transcription factors: implications for biological control. J. Cell
Sci. 113 (14), 2527–2533.
Triezenberg, S.J., Kingsbury, R.C. and McKnight, S.L. (1988)
Functional dissection of VP16, the trans-activator of herpes
simplex virus immediate early gene expression. Genes Dev. 2,
718–729.
Verdaguer, B., de Kochko, A., Beachy, R.N. and Fauquet, C. (1996)
Isolation and expression in transgenic tobacco and rice plants of
the cassava vein mosaic virus (CVMV) promoter. Plant Mol. Biol.
31, 1129–1139.
Verdaguer, B., De Kochko, A., Fux, C.I., Beachy, R.N. and Fauquet,
C. (1998) Functional organization of the cassava vein mosaic
virus (CsVMV) promoter. Plant Mol. Biol. 37, 1055–1067.
Watanabe, Y., Meshi, T. and Okada, Y. (1987) Infection of tobacco
protoplasts with in vitro transcribed tobacco mosaic virus RNA
using an improved electroporation method. FEBS Lett. 219,
65–69.
Wu, H., Yang, W.P. and Barbas, C.F.R. (1995) Building zinc fingers
by selection: towards a therapeutic application. Proc. Natl Acad.
Sci. USA, 92, 344–348.
Zhang, L., Spratt, S.K., Liu, Q., Johnstone, B., Qi, H., Raschke, E.E.,
Jamieson, A.C., Rebar, E.J., Wolffe, A.P. and Case, C.C. (2000)
Synthetic zinc finger transcription factor action at an endogenous chromosomal site. Activation of the human erythropoietin
gene. J. Biol. Chem. 275, 33850–33860.
ß Blackwell Publishing Ltd, The Plant Journal, (2002), 32, 1077–1086