Gene Therapy (1998) 5, 1656–1664
 1998 Stockton Press All rights reserved 0969-7128/98 $12.00
http://www.stockton-press.co.uk/gt
A strategy for enhancing the transcriptional activity of
weak cell type-specific promoters
DM Nettelbeck, V Jérôme and R Müller
Institut für Molekularbiologie und Tumorforschung (IMT), Philipps-Universität Marburg, Emil-Mannkopff-Strasse 2, D-35033
Marburg, Germany
Cell type- and tissue-specific promoters play an important
role in the development of site-selective vectors for gene
therapy. A large number of highly specific promoters has
been described, but their applicability is often hampered
by their inefficient transcriptional activity. In this study, we
describe a new strategy for enhancing the activity of weak
promoters without loss of specificity. The basic principle of
this strategy is to establish a positive feedback loop which
is initiated by transcription from a cell type-specific promoter. This was achieved by using a cell type-specific pro-
moter to drive the simultaneous expression of the desired
effector/reporter gene product and a strong artificial transcriptional activator which stimulates transcription through
appropriate binding sites in the promoter. Using a VP16LexA chimeric transcription factor, we show that this
approach leads to a 14- to ⬎100-fold enhancement of both
the endothelial cell-specific von Willebrand factor promoter
and the gastrointestinal-specific sucrase-isomaltase promoter while maintaining approximately 30- to ⬎100-fold
cell type specificity.
Keywords: tissue-specific promoters; endothelial cell-specific transcription; von Willebrand factor promoter; gastrointestinal-specific transcription; sucrase isomaltase promoter; transcriptional targeting
Introduction
The construction of vectors targeting gene expression to
specific sites in the organism is one of the current challenges in the field of gene therapy. In view of the fact
that the efficient targeted transduction of genes still represents a major obstacle, the possibility of restricting gene
expression to specific cell populations through specific
promoters is of particular importance.1–3 This not only
applies to the treatment of systemic diseases, such as
metastatic cancer, but also to local gene therapy, whose
efficacy and safety can be improved by the use of cell
type-specific promoters which keep the expression of the
therapeutic gene in non-target cells at a minimum.
For these reasons, many of the recently described
experimental gene therapy protocols make use of cell
type- or tissue-specific promoters.1–3 These include the
tyrosinase promoter for directing gene expression to
melanoma cells,4–6 the ␣-fetoprotein promoter for hepatoma cells,7 the mucin-1 promoter for mammary carcinoma,8 the prostate-specific antigen promoter for prostate
carcinoma9 or the neuron-specific enolase promoter for
neuronal cells.10 In other studies, transcriptional control
elements which can stimulate transcription in response
to disease-specific alterations, such as hypoxic conditions11 or the loss of cell cycle checkpoints in tumors,12
have also been used. In these scenarios, promoters containing binding sites for hypoxia-inducible factor-1 (HIF1) or the transcription factor E2F whose repression
through the retinoblastoma protein pathway is lost in
Correspondence: R Müller
Received 20 May 1998; accepted 15 July 1998
most, if not all, tumor cells have been successfully incorporated into experimental gene therapy strategies. Furthermore, regulatory sequences inducible by therapeutic
dose of ionizing radiation, such as the egr-1 promoter,13
or in response to oncogenic lesions, such as the erbB2
promoter,14 are worth mentioning in the same context. In
addition to these promoters, a plethora of other potentially useful transcriptional regulatory sequences with a
high degree of specificity has been described.1–3
Some of these promoters appear to be ideally suited
for gene therapy, because they combine strong transcriptional activity with a high degree of specificity. This is,
for instance, true for the melanocyte-specific tyrosinase
promoter.4–6,15 Frequently, however, specific promoters
are inefficient activators of transcription which severely
limits their applicability. A typical example is the von
Willebrand factor (vWF) promoter, which exhibits a
particularly high degree of endothelial cell (EC) specificity compared with other EC promoters, but is a poor
activator of transcription16–19 (see also Results below).
Other EC promoters are stronger, but less specific, such
as the promoters for PECAM-1/CD3120,21 or flk1/KDR.22,23 Another example is the promoter of the sucrase-isomaltase (SI) gene which is highly specific for
intestinal cells and gastrointestinal tumors,24,25 but again
is a poor transcriptional activator (see Results below). In
the present study, we have chosen the vWF and SI promoters as models for the establishment of a strategy that
would be suitable for enhancing the transcriptional
activity of weak promoters without affecting their specificity (self-enhancing promoters).26 This goal was achieved by incorporating into these promoters a positive
feedback loop26 provided by a chimeric transcription factor consisting of the strong herpes simplex virus VP16
Enhancing cell type-specific transcription
DM Nettelbeck et al
1657
transcriptional activation domain27,28 fused to the DNAbinding domain of LexA.29
Results
Comparative analysis of EC-specific promoters
A comparison of different EC promoters in various ECs
and non-endothelial cell lines was performed in order to
identify a suitable promoter for the transcriptional targeting of ECs. The promoters of the human vWF,17
PECAM-120,21 and KDR22 genes were analyzed in transient luciferase assays using human umbilical vein ECs
(HUVECs), bovine aortic ECs (BAECs), mouse NIH3T3
fibroblasts and the human melanoma cell line MeWo. As
shown in Figure 1a, a clear selectivity for ECs was found
with all three promoters, with the highest degree of cell
type specificity seen with the vWF promoter. This is
particularly well documented by the data in Figure 1b
which show a complete lack of activity in all non-endothelial cell lines tested (NIH3T3, MeWo, 10T1/2, HeLa),
ie RLU values similar to those obtained with the promoterless basic vector, while both the PECAM-1 and
KDR promoter showed readily detectable activities in the
non-endothelial cells (Figure 1a). We also tested other
fragments of the KDR and PECAM-1 promoters (see
Materials and methods for details), but these showed
even lower levels of activity and/or specificity (data not
shown). In contrast to its high degree of specificity, the
vWF promoter is endowed with very poor transcriptional
activation properties which severely limits its applicability (10–20% of SV40 promoter; Figure 1a). In a first
attempt to improve the activity of the vWF promoter we
tested a construct in which the SV40 enhancer was placed
upstream of the promoter. This construct (SV40EvWF)
showed strongly increased transcriptional activity, but
the cell type specificity was largely lost. We also tested
other EC promoters, including those of the human endoglin (kindly provided by W. Graulich, IMT, Marburg,
Germany) and human P-selectin30 genes, but these did
not provide the same high degree of specificity seen with
the vWF promoter (data not shown). We did not include
the preproendothelin promoter31 in these studies because
its activity has been reported to be restricted to bigger
blood vessels and is also active in some epithelial cells,32
which precludes its use for targeting the vasculature in
tumors or other proliferative diseases.
Establishment of a positive feedback loop for the vWF
promoter: proof of principle
In view of the data discussed above, we sought to establish a new strategy that would allow for an enhancement
of promoter activity by introducing a positive feedback
loop. The basic principle of this strategy is to use a cell
type-specific promoter to drive the simultaneous
expression of the desired effector/reporter gene product
and a strong artificial transcriptional activator, ie a chimeric protein consisting of the DNA-binding domain of
LexA29 and the strong herpes simplex virus VP16 transcriptional activation domain,27,28 which can stimulate
transcription through LexA binding sites introduced into
the promoter (Figure 2). To test the principle of this
approach, we first used the vWF promoter as a model
and performed transient luciferase assays after cotransfection of expression and reporter plasmids (see Figure 3a
Figure 1 (a)Activity of different promoters in ECs and non-endothelial
cells after transient transfection of luciferase reporter constructs. Values
were standardized individually for each cell line using a SV40 promoter
(SV40p) construct. Basic, promoterless pGL3 vector; SV40EvWF, human
vWF promoter with an upstream SV40 enhancer. (b) Activity of the vWF
promoter in ECs and non-endothelial cells after transient transfection of
luciferase reporter constructs. Values were standardized individually for
each cell line using the promoterless pGL3 vector (basic) which allows for
a clearer representation of the vWF promoter activity in the various cell
lines (especially the lack of activity in non-endothelial cells). The data
shown are averages of triplicates. The error bars indicate the standard
deviations.
for details) into HUVECs, BAECs, NIH3T3 cells and
MeWo cells. The data in Figure 3b (standardized to SV40
promoter activity (SP40p) for each cell line) and Figure 3c
demonstrate that (1) the insertion of the seven LexA sites
into the vWF promoter (LexvWFLuc) had no effect on
activity or specificity; (2) the cotransfection of a SV40 promoter-driven expression vector encoding a VP16-LexA
fusion protein (SV-VPLex) led to a dramatic increase in
activity (approximately 30- to 70-fold), but the promoter
also showed some activity in non-endothelial cells; (3) the
coexpression of the same protein from the EC-specific
vWF promoter resulted in a similar enhancement of transcription (23- to 66-fold), but expression was largely ECspecific (57- to 156-fold relative to NIH3T3; 29- to 78-fold
relative to MeWo); and (4) cotransfection of a control vec-
Enhancing cell type-specific transcription
DM Nettelbeck et al
1658
Figure 2 Schematic outline of the positive feedback loop. Transcription is initiated by the cell type-specific promoter in target cells (thin arrows) which
leads to the simultaneous expression of the reporter/effector gene and the VP16-LexA fusion protein. The VP16-LexA protein then interacts with the
LexA-bindings sites upstream of the 5′ cell type-specific promoter leading to transactivation and thus an enhancement of transcription (bold arrows).
The VP16-LexA fusion protein is expressed either via an IRES or a second cell type-specific promoter.
Figure 3 Establishment of a positive feedback loop for the EC-specific vWF promoter. (a) Structure of the constructs. NLS, nuclear localization signal.
(b) Activity of the constructs shown in panel (a) after transient transfection of ECs and non-endothelial cells. Details were as described in the legend
to Figure 1. (c) Quantitative evaluation of the data shown in panel (b).
tor coding for a protein lacking the VP16 activation
domain gave very low values similar to the transfection
of the reporter alone. These observations constitute clear
proof of principle. There are, however, several open questions that required further attention: (1) What is the reason for the observed EC selectivity after cotransfection
of the reporter plasmid with the SV40 promoter-driven
expression vector SV-VPLex (six- to 18-fold; see
Figure 3c), (2) would it be possible to endow the system
with an even higher specificity; and (3) is the same
approach also applicable to other cell type-specific promoters? All three issues will be addressed below.
Enhancing cell type-specific transcription
DM Nettelbeck et al
Mechanism of cell type specificity
The relatively good EC selectivity after cotransfection of
the reporter plasmid with the ubiquitous SV40 promoterdriven expression vector SV-VPLex points to the existence of a second mechanism conferring specificity on the
promoter enhancement system established above (in
addition to the cell type-specific feedback loop). This
could in principle be due to two mechanisms: (1) a cell
type-specific, promoter-independent function of the VP16
transactivation domain, which could be due to the existence of cell type-specific interacting modulators; or (2)
the suppression of VP16 activity in non-endothelial cells
in the context of a silent vWF promoter, eg due to the
assembly of a promoter-bound protein complex which is
unable to mediate, or actively represses, the activation
function of VP16. In order to distinguish between these
possibilities we constructed a reporter plasmid which,
instead of a cell-type specific promoter, only contained a
TATA-box and an initiator element (Inr)33 downstream
of seven LexA sites (LexTATA/ILuc; Figure 4a). As
expected, transfection of this construct on its own
resulted in hardly any luciferase activity, while the
cotransfection of SV-VPLex gave rise to a strong transcriptional activity (Figure 4b). Most importantly, high
activities were seen in both ECs and non-endothelial cells
(MeWo and LoVo colon carcinoma cells), which clearly
demonstrates that the activity of VP16 itself – at least in
the systems analyzed – is not subject to cell type-dependent regulation. We therefore conclude that the transcriptional activation function of VP16 is dependent on the
context of the promoter and suppressed through the vWF
promoter in non-target cells.
Establishment of a single-vector system
The cotransfection protocol described above has the
potential disadvantage that not all transduced cells
receive an optimum molar ratio of promoter and transactivator which might lead to a sub-optimal activity
and/or specificity. In addition, the requirement for two
different plasmids would make it difficult to use the sys-
Figure 4 Cell type specificity of the positive feedback loop is dependent
on a cell type specific promoter. (a) Structure of the non-specific promoter
construct LexTATA/ILuc. (b) Activity of SV-VPLex on LexTATA/ILuc
after transient transfection of BAECs and non-endothelial cells. Details
were as described in the legend to Figure 1.
tem with a vector suitable for in vivo gene therapy. We
therefore decided to introduce all necessary components
into a single plasmid vector using two different strategies. The first construct contains two copies of the vWF
promoter, one directing the expression of VP16-LexA, the
second being VP16-LexA-responsive and driving the
transcription of the luciferase gene (construct vWF-2P in
Figure 5a). The second construct contains a single VP16LexA-responsive promoter which directs the expression
of a bicistronic transcript containing both the luciferase
and VP16-LexA coding sequences separated by an
internal ribosome entry site (IRES; construct vWF-I in
Figure 5a).34 Both constructs were tested in HUVECs and
BAECs, as well as in NIH3T3 and MeWo cells. vWF-2P
was also analyzed in the non-endothelial cell lines PC-3
(human prostate carcinoma) and HepG2 (human
hepatoma). As negative controls, we also constructed
both vectors with a deletion of the VP16 domain (vWF2P⌬VP and vWF-I⌬VP). From the data in Figure 5b and
c, it is clear that the two promoter construct vWF-2P was
superior with respect to promoter strength compared
with the IRES-containing vector vWF-I. Thus, with
respect to the wild-type vWF promoter, vWF-2P showed
a 20- and 169-fold enhanced activity in BAECs and
HUVECs, respectively (Figure 5c). As expected, the
deletion of the VP16 moiety diminished the promoter
activities to very low levels in both cases (Figure 5b), proving the functionality of the VP16-dependent feedback
loop. While the data on promoter strength (at least for
vWF-2P) were similar to the results obtained by cotransfection (Figure 3c), a substantial improvement was seen
when the promoter specificities were compared. Thus,
both single-plasmid systems were practically silent in
non-endothelial cells, with values that were very similar
to those obtained with the empty reporter plasmid.
Although precise calculations are difficult, the promoter
specificity was estimated to be 47- to ⬎200-fold in BAECs
and approximately 100- to ⬎1000-fold in HUVECs compared with the other four cell lines (Figure 5c).
Establishment of a positive feedback loop for the SI
promoter
Finally, we sought to verify the positive feedback loop
system by applying it to a second cell type-specific promoter. Towards this end, we chose the promoter of the
SI gene which is specifically active in crypt cells of the
gut and in gastrointestinal tumor cells.24,25,35 Appropriate
plasmids were constructed (Figure 6a) analogous to the
vWF promoter system and tested by cotransfection into
the colon carcinoma cell lines LoVo and CaCo-2, as well
as the non-target cells NIH3T3 and MeWo. The following
results were obtained with this system (Figure 6b and c):
(1) the SI promoter was highly specific for colon carcinoma cells, but showed a very weak transcriptional
activity (approximately 20% the activity of the SV40
promoter); (2) the insertion of the seven LexA sites
upstream of the SI promoter (LexSILuc) had no significant effect on either activity or specificity; (3) the cotransfection of the SV40 promoter-driven expression vector
SV-VPLex led to a dramatic increase in activity
(approximately 30- to 40-fold), however, the promoter
was also weakly activated in non-target cells; (4) the coexpression of the same protein from the cell specific SI promoter also resulted in a strong enhancement of transcription (14- to 37-fold), but in this case transcription was
1659
Enhancing cell type-specific transcription
DM Nettelbeck et al
1660
Figure 5 Establishment of a one-vector system conferring an enhancement of EC-specific transcription. (a) Structure of the constructs. (b) Activity of
the constructs shown in panel (a) after transient transfection of ECs and non-endothelial cells. ⌬VP, VP16 moiety deleted. Details were as described
in the legend to Figure 1. (c) Quantitative evaluation of the data shown in panel (b).
specific for the colon carcinoma cells (⬎200-fold); and (5)
cotransfection of a VP16-deleted control vector gave rise
to similar values as seen with the reporter alone. These
observations therefore clearly show that the colon carcinoma-specific activity of the SI promoter can be substantially increased by the system established in the present
study. As in the case of the vWF promoter, the higher
transcription in colon carcinoma cells observed with the
SV40 promoter-driven expression vector SV-VPLex (see
(3) above) appears to be due to promoter-specific suppression of the VP16 activation function in non-target
cells, since no specificity was seen in LoVo cells with the
TATA-Inr driven reporter plasmid LexTATA/ILuc (see
Figure 4b).
Discussion
A frequent problem with the generation of tissue- or cell
type-specific vectors is the lack of promoters suitable for
transcriptional targeting. A major reason lies in the fact
that specific promoters are often relatively weak inducers
of transcription. We have therefore decided to establish
a strategy for the enhancement of the transcriptional
activity of such promoters without losing the specificity.
In the present study, we describe such a strategy and
demonstrate its applicability to enhance the transcriptional efficiency of two weak cell type-specific promoters,
the EC-specific vWF promoter17 and the intestinal cellspecific SI promoter.24 The principle of this strategy is the
implementation of a positive auto-regulatory loop provided by a VP16-LexA fusion protein which is expressed
simultaneously with the effector/reporter gene product
and can bind to LexA sites inserted 5′ to the cell typespecific promoter (Figure 2). The feasibility of this
approach was shown in cotransfection experiments,
which showed an enhancement of transcriptional activity
in the range of 23- to 67-fold for the vWF promoter, and
14- to 37-fold for the SI promoter. In spite of this dramatic
increase in promoter activity, specificity was completely
retained, and was in the order of 30- to 147-fold for the
Enhancing cell type-specific transcription
DM Nettelbeck et al
Figure 6 Establishment of a positive feedback loop for the intestine-specific
SI promoter. (a) Structure of the constructs. (b) Activity of the constructs
shown in panel (a) after transient transfection of colon carcinoma cell lines
(LoVo and CaCo-2) and non-target cells. Details were as described in
the legend to Figure 1. (c) Quantitative evaluation of the data shown in
panel (b).
vWF promoter (Figure 3) and ⬎200-fold for the SI promoter when compared to the activities in non-target
cells (Figure 6).
Some of the data obtained in the experiments shown
in Figures 3 and 6 suggest a second mechanism to be
involved in conferring cell type specificity. Thus,
expression of the VP16-LexA activator protein from the
ubiquitous SV40 promoter resulted in a clearly cell type
specific auto-regulatory loop. In order to address the
basis of this observation we implemented the identical
VP16-LexA-mediated feedback loop system into a ubiquitous basal promoter, TATA-Inr. In this case, the SV40driven expression of VP16-LexA did not result in any cell
type specificity, indicating that the VP16 protein itself is
not subject to cell type-dependent modulatory mechanisms. This finding rather suggests that the vWF and SI
promoters are able to suppress the action of a heterologous activator, such as VP16, in non-target cells where the
promoter is silent. This scenario is supported by the
observation that repression apparently plays a role in
silencing the vWF promoter in non-endothelial cells.19 A
similar mechanism may operate in the case of the SI promoter, since we reproducibly observed a decrease in
luciferase activities in non-target cells upon insertion of
the SI promoter into the promoterless reporter plasmid
pGL3.
We also investigated the effect of a positive auto-regulatory loop on a strong tissue-specific promoter, the melanocyte-specific tyrosinase promoter.15 In this case, we
also observed an increase in activity, but this enhancement was relatively modest (approximately two-fold;
data not shown). Melanocyte-specific expression was not
affected. These findings indicate that the promoter
enhancement system in principle also works with the
tyrosinase promoter, but that the increase in activity is
rather poor presumably due to its strong activity in the
wild-type form. We therefore anticipate that our strategy
is particularly useful for enhancing the activity of weak
promoters. In this context it will be interesting to investigate in future studies whether the same approach can
also be used to increase the activity of promoters
endowed with other kinds of specificity, such as
cell cycle-controlled, mitogen-regulated, oncoproteininducible or hypoxia-responsive promoters.
The enhancement of transcription and promoter specificity were particularly impressive when all components
of the positive feedback loop system were cloned into the
same plasmid vector, which was demonstrated for the
EC-specific activity of the vWF two-promoter construct
vWF-2P (approximately 50- to ⬎1000-fold relative to nonendothelial cells; Figure 5). We ascribe this result to an
equal ratio of the transgene products in all transduced
cells, a situation that cannot be achieved in cotransfection
experiments. Surprisingly, a considerably reduced
activity was seen when an IRES was used instead of the
second promoter. A possible reason for this observation
might be a less efficient expression of the VP16-LexA
gene downstream of the IRES. We did not pursue this
issue in view of the high activity and pronounced specificity of the two-promoter construct.
The functionality of the single plasmid system is of
particular relevance when viral vectors are to be used for
transduction. Viral vectors are currently the preferred
means to achieve efficient transduction efficiencies in
vivo. We are therefore currently in the process of generat-
1661
Enhancing cell type-specific transcription
DM Nettelbeck et al
1662
ing replication-deficient adenoviruses with a vWF promoter feedback loop to be able to investigate the activity
and specificity of the positive feedback loop system in
animal experiments. These experiments will also have to
address the critical question as to whether the viral
sequences will interfere with the transcriptional feedback
loop of the inserted transgene.
Previous studies have shown that the natural vWF
promoter is suitable to transduce a suicidal signal to
ECs using the herpes simplex virus thymidine
kinase/ganciclovir system after selection for transduced
cells.36 In view of the strong increase in promoter activity
achieved by the system established in the present study
it is very likely that the efficacy of such gene therapeutic
approaches can be dramatically increased, especially if
one takes into account the poor transfection efficiencies
in vivo and that other effector systems can also be successfully used. In this context it is also relevant that a VP16based transcriptional activator has been shown to be
functional in vivo,37 albeit in a different context. We therefore feel that the system established in the present study
could represent an important step in the development of
efficient genetic therapies.
Materials and methods
Cell culture
The mouse fibroblast cell lines NIH3T3 and 10T., and the
human cell lines LoVo (ATCC CCL 229; colon
carcinoma), CaCo-2 (ATCC HTB 37; colon carcinoma),
HeLa (ATCC CCL 2; cervix carcinoma), PC-3 (ATCC CRL
1435; prostate carcinoma), HepG2 (ATCC HB 8065;
hepatoma), and MeWo38 (melanoma) were grown at 37°C
in a humidified atmosphere of 5% CO2. PC-3 cells were
maintained in RPMI 1640 (Boehringer Ingelheim BioProducts, Heidelberg, Germany) and the other cell lines
in Dulbecco-Vogt modified Eagle’s medium (DMEM;
Boehringer Ingelheim BioProducts). RPMI 1640 and
DMEM were supplemented with 10% fetal bovine serum
(FBS; Life Technologies, Eggenstein, Germany) and 2 mm
l-glutamine. MeWo cells were obtained from I Hart
(London, UK) and LoVo cells from Hoechst Marion Roussell (Marburg, Germany). BAECs were kindly provided
by W Risau (Bad Nauheim, Germany) and maintained
in DMEM supplemented with 10% inactivated FBS (PAA
Laboratories, Coelbe, Germany), and 2 mm l-glutamine.
HUVECs were prepared by the method of Jaffe,39 as
modified by Thornton et al,40 and cultivated in EGM-2
medium (Boehringer Ingelheim BioProducts).
Plasmids
pGL3basic (basic), pGL3promoter (SV40p), and pGL3
control were purchased from Promega (Heidelberg,
Germany). The basic construct lacking eukaryotic promoter and enhancer sequences was used for determining
the vector-related background activity. The SV40p construct was used for standardizing transfection
efficiencies. The luciferase constructs containing the
human vWF promoter17 (−487/+247), the human
PECAM-1 promoter20 (−50/+469 in Figure 1, we also
cloned the −939/+469, −228/+469 and +232/+469
fragments), the KDR promoter22 (−225/+240 in Figure 1,
we also cloned the −780/+240 and −115/+240 fragments),
and the SI promoter24 (−175/+70) were cloned by PCR
amplification using the listed oligonucleotides. The PCR
products were digested with the corresponding restriction enzymes and ligated into the pGL3basic vector. The
KDR promoter (−115/+240) construct was cloned using
the intrinsic KpnI restriction site. The SV40EvWF construct containing the vWF promoter and SV40 enhancer
was obtained by replacing the SV40 promoter of the
pGL3control plasmid with the vWF promoter.
LexvWFLuc, LexSILuc and LexTATA/ILuc contain
seven LexA binding sites (5′-GTCGAGTACTGTATGTACATACAGTAC-3′)29 inserted upstream of the promoters
in the vWF and SI constructs or upstream of a synthetic
TATA-box/initiator element33 in pGL3. SV-VPLex was
constructed by replacing the luciferase gene of pGL3 promoter with a fusion of the sequences encoding the
nuclear localization signal (NLS) of the SV40 large T antigen and the transcriptional activation domain of the herpes virus protein VP16 (aa 411–455) to a cDNA fragment
of the E. coli repressor LexA (aa 2–202) encoding its DNA
binding domain.29 PCR amplifications were performed
using the oligonucleotides nVP16–5′, nVP16–3′ (template:
pVP16; Clontech, Heidelberg, Germany), LexA-5′, and
LexA-3′ (see below). The constructs vWF-VPLex and SIVPLex were obtained by replacing the SV40 promoter of
SV-VPLex with the vWF promoter or the SI promoter.
vWF-Lex and SI-Lex were cloned by inserting a PCRamplified fragment encoding the SV40 large T NLS fused
to LexA (aa 2–202). For PCR amplification the
oligonucleotides nLexA-5′ and LexA-3′ were used.
For cloning of vWF-2P and vWF-2P⌬VP, NotI/SalI
digests of vWF-VPLex or vWF-Lex were converted to
blunt ends and ligated into blunted, NotI-digested
LexvWFLuc in a head-to-tail orientation. Two cloning
steps were necessary to generate the vWF-I and vWFI⌬VP constructs. Firstly, the IRES sequence of pIRESEGFP (Clontech) was PCR amplified using the oligonucleotides IRES-5′ and IRES-3′ and cloned into the XbaI
site of LexvWFLuc (the initiation ATG was reverted to
the original position of the encephalomyocarditis virus).34
Then, the NcoI/SalI fragments of the vWF-VPLex and
vWF-Lex plamids were introduced into the IRESNcoI/SalI sites to give, respectively, the vWF-I and the
vWF-I⌬VP constructs.
All plasmid DNAs were amplified in E. coli and purified according a standard protocol (Qiagen, Hilden,
Germany). Plasmids obtained by PCR amplification were
verified by DNA sequencing. The following oligonucleotides were used for PCR amplification (numbers refer to
nucleotide positions, restriction sites are shown in italics):
vWF promoter17
5′-primer: 5′-AGTCGCTAGCATCTTTAGCCGATC
CATTCAA (−487/−467)
3′-primer: 5′-AGTCCTCGAGCCCCTGCAAATGAG
GGCT (+247/−230)
PECAM-1 promoter20
5′-primer:
3′-primer:
5′-GATCGCTAGCGAAAAGTCAGCCC
CAAGAAAG (−50/−30)
5′-GTAGCTCGAGACTGCAGGCACAG
TTAGTTCTGC (+491/+469)
Other PECAM-1 promoter constructs:
5′-primer (−939/+469): 5′-GATCGGTACCGGGTTCA
AGCGATTCTCCTGCCT
(−939/−917)
Enhancing cell type-specific transcription
DM Nettelbeck et al
5′-primer (−228/+469): 5′-GATCGGTACCTTATTTT
GTATGTTCTTCTTCTAT
(−228/−205)
5′-primer (+50/+469): 5′-GATCGCTAGCGGCCGCG
TTCGGTGGCCCCTCAG
(+50/+72)
KDR promoter22
5′-primer: 5′-GATCGCTAGCGTTGTTGCTCTGG
GATGTTCTC (−225/−205)
3′-primer: 5′-GTACCTCGAGTCTAGAGAAGGA
GGCGCGG (+240/+211)
KDR promoter −780/+240 construct:
5′-primer (−780): 5′-GATCGCTAGCCCTCCTTC
CCCTGGGCCT (−780/−763)
SI promoter24
5′-primer: 5′-TGCATGCATGCAGGATCCCAATT
ACTAATTAACTTAGA (−175/−156)
3′-primer: 5′-ACGTACGTACGTAAGCTTATGAG
CCATAGACTACCTTA (+70/+61)
nVP16–5′: 5′-GATCAAGCTTCCACCATGGGCCC
TAAAAAGAAGCGT
nVP16–3′: 5′-CTAGGTCGACCCCACCGTACTCGT
CAATTCC
LexA-5′:
LexA-3′:
nLexA-5′:
IRES-5′:
IRES-3′:
5′-GATCGTCGACCTCGAGGCGCCAA
GAAGAA
5′-ACGTTCTAGAGTTACAGCCAGTCG
CCGTTG
5′-GATCAAGCTTCCACCATGGGCCC
TAAAAAGAAGCGTAAAGTCAAAG
CGTTAACGGCCAGGCAA
5′-GATCGCTAGCGCCAATTCCGCCC
CTCTCCCT (1287/1307)
3′-GATCTCTAGATCACCATGGTATCA
TCGTGTTTTTCAAAGGAA (1869/1847)
Transfections and luciferase assays
NIH3T3, LoVo, CaCo-2, PC-3, HepG2, MeWo cells and
BAECs (passage 10 to 14) were transfected at 50% confluency in six-well plates using the cationic liposome
DOTAP as described by the manufacturer (Boehringer
Mannheim, Mannheim, Germany). Three micrograms of
plasmid (for cotransfection experiments 1 ␮g of reporter
plasmid and 2 ␮g of activator plasmid or pUC19) were
used per well and mixed with 6 ␮l of DOTAP in transfection buffer. The transfection mixture was mixed with
OptiMEM (Life Technologies) or EGM-2 (for BAECs) and
added to the cells, followed by a 6-h incubation. 10T. and
HeLa cells were transfected at 80% confluency by the
DEAE-dextran technique.41 HUVECs (passage 4 to 6)
were transfected at 80% confluency with Superfect
(Qiagen) following the manufacturer’s protocol. 1.5 ␮g
plasmid (or 0.5 ␮g reporter plasmid + 1 ␮g activator plasmid or pUC19) and 9 ␮l Superfect per well were mixed
with serum-free DMEM and incubated for 2 h in EGM-2
on the cells. Luciferase activities were determined 40 h
after replacement of the transfection medium with normal growth medium as decribed.42 Transfections were
performed in triplicates and repeated at least once using
an independent plasmid preparation. The standard deviations were in the range of 5–20%. We did not use a
second cotransfected plasmid for standardization,
because in previous experiments we found that all
ubiquitous promoters (including SV40 and CMV) are
more or less influenced by cotransfected plasmids, in
particular those encoding (domains of) transcription factors, and can sometimes also affect the activity of the
regulatory sequences to be tested. This may result in an
incorrect standardization. We therefore feel that it is more
appropriate to transfect the reference plasmid separately
and to perform each assay several times to control for
experimental variation.
Acknowledgements
We are grateful to Dr S Brüsselbach, Prof W Risau and
Dr M Clauss for HUVECs and BAECs, to Profs I Hart
and HH Sedlacek (HMR) for tumor cell lines and to
M Krause for oligonucleotide synthesis. This work was
supported in part by a grant from the BMBF.
References
1 Sikora K. Gene therapy for cancer. Trends Biotech 1993; 11:
197–201.
2 Miller N, Vile RG. Targeted vectors for gene therapy. FASEB J
1995; 9: 190–199.
3 Miller N, Whelan J. Progress in transcriptionally targeted and
regulatable vectors for genetic therapy. Hum Gene Ther 1997; 8:
803–815.
4 Vile RG, Hart IR. Use of tissue-specific expression of the herpes
simplex virus thymidine kinase gene to inhibit growth of established murine melanomas following direct intratumoral injection of DNA. Cancer Res 1993; 53: 3860–3863.
5 Vile RG, Hart IR. In vivo and in vitro targeting of gene expression
to melanoma cells. Cancer Res 1993; 53: 962–967.
6 Siders WM, Halloran PJ, Fenton RG. Transcriptional targeting
of recombinant adenoviruses to human and murine melanoma
cells. Cancer Res 1996; 56: 5638–5646.
7 Arbuthnot PB et al. In vitro and in vivo hepatoma cell-specific
expression of a gene transferred with an adenoviral vector. Hum
Gene Ther 1996; 7: 1503–1514.
8 Chen L et al. Breast cancer selective gene expression and therapy
mediated by recombinant adenoviruses containing the
DF3/MUC1 promoter. J Clin Invest 1995; 96: 2775–2782.
9 Lee C-H, Liu M, Sie K-L, Lee M-S. Prostate-specific antigen
driven gene therapy targeting DNA-polymerase ␣ and topoisomerase IIa in prostate cancer. Anticancer Res 1996; 16: 1805–1812.
10 Abdallah B et al. A powerful nonviral vector for in vivo gene
transfer into the adult mammalian brain: polyethylenimine.
Hum Gene Ther 1996; 7: 1947–1954.
11 Dachs GU et al. Targeting gene expression to hypoxic tumor
cells. Nature Med 1997; 3: 5515–5520.
12 Parr MJ et al. Tumor-selective expression in vivo mediated by
an E2F-responsive adenoviral vector. Nature Med 1997; 3:
1145–1149.
13 Hallahan DE et al. Spatial and temporal control of gene therapy
using ionizing radiation. Nature Med 1995; 1: 786–791.
14 Sikora K, Harris J, Hurst H, Lemoine N. Therapeutic strategies
using c-erbB-2 promoter-controlled drug activation. Ann NY
Acad Sci 1994; 716: 115–124.
15 Shibata K et al. Identification of a cis-acting element that
enhances the pigment cell-specific expression of the human
tyrosinase gene. J Biol Chem 1992; 267: 20584–20588.
16 Ferreira V et al. The role of the 5′-flanking region in the cellspecific transcription of the human von Willebrand factor gene.
Biochem J 1993; 253: 641–648.
17 Jahroudi N, Lynch DC. Endothelial-cell-specific regulation of
von Willebrand factor gene expression. Mol Cell Biol 1994; 14:
999–1008.
1663
Enhancing cell type-specific transcription
DM Nettelbeck et al
1664
18 Jahroudi N, Ardekani AM, Greenberger JS. Ionizing irradiation
increases transcription of the von Willebrand factor gene in
endothelial cells. Blood 1996; 88: 3801–3814.
19 Jahroudi N, Ardekani AM, Greenberger JS. An NF1-like protein
functions as a repressor of the von Willebrand factor promoter.
J Biol Chem 1996; 271: 21413–21421.
20 Almendro N et al. Cloning of the human platelet endothelial cell
adhesion molecule-1 promoter and its tissue-specific expression.
Structural and functional characterization. J Immunol 1996; 157:
5411–5421.
21 Gumina RJ, Kirschbaum NE, Piotrowski K, Newman PJ. Characterization of the human platelet/endothelial cell adhesion molecule-1 promoter: identification of a GATA-2 binding element
required for optimal transcriptional activity. Blood 1997; 89:
1260–1269.
22 Patterson C et al. Cloning and functional analysis of the promoter for KDR/flk-1, a receptor for vascular endothelial growth
factor. J Biol Chem 1995; 270: 23111–23118.
23 Rönicke V, Risau W, Breier G. Characterization of the endothelium-specific murine vascular endothelial growth factor
receptor-2 (Flk-1) promoter. Circ Res 1996; 79: 277–285.
24 Wu GD, Wang W, Traber PG. Isolation and characterization of
the human sucrase-isomaltase gene and demonstration of intestine-specific transcriptional elements. J Biol Chem 1992; 267:
7863–7870.
25 Traber PG, Wu GD, Wang W. Novel DNA-binding proteins
regulate intestine-specific transcription of the sucrase-isomaltase
gene. Mol Cell Biol 1992; 12: 3614–3627.
26 Müller R, Sedlacek HH. Selbstverstärkende, pharmakologisch
kontrollierbare Expressionssysteme. Patent Application: 1997;
DE 19651443.6.
27 Sadowski I, Ma J, Triezenberg S, Ptashne M. GAL4-VP16 is an
unusually potent transcriptional activator. Nature 1988; 335:
563–564.
28 Triezenberg SJ, Kingsbury RC, McKnight SL. Functional dissection of VP16, the trans-activator of herpes simplex virus
immediate–early gene expression. Genes Dev 1988; 2: 718–729.
29 Brent R, Ptashne M. A eukaryotic transcriptional activator bear-
30
31
32
33
34
35
36
37
38
39
40
41
42
ing the DNA specificity of a prokaryotic repressor. Cell 1985; 43:
729–736.
Pan J, McEver RP. Characterization of the promoter for the
human P-selectin gene. J Biol Chem 1993; 268: 22600–22608.
Kawana M, Lee ME, Quertermous EE, Quertermous, T. Cooperative interaction of GATA-2 and AP1 regulates transcription of
the endothelin-1 gene. Mol Cell Biol 1995; 15: 4225–4231.
Harats D et al. Targeting gene expression to the vascular wall
in transgenic mice using the murine preproendothelin-1 promoter. J Clin Invest 1995; 95: 1335–1344.
Kaufmann J, Smale ST. Direct recognition of initiator elements
by a component of the transcription factor IID complex. Genes
Dev 1994; 8: 821–829.
Rees S et al. Bicistronic vector for the creation of stable mammalian cell lines that predisposes all antibiotic-resistant cells to
express recombinant protein. Biotechniques 1996; 20: 102–104.
Rodolosse A et al. A limited upstream region of the human sucrase-isomaltase gene confers glucose-regulated expression on a
heterologous gene. Biochem J 1996; 315: 301–306.
Ozaki K et al. Use of von Willebrand factor promoter to transduce suicidal gene to human endothelial cells, HUVEC. Hum
Gene Ther 1996; 7: 1483–1490.
Fang B, Ji L, Bouvet M, Roth JA. Evaluation of GAL4/TATA in
vivo. Induction of transgene expression by adenovirally
mediated gene codelivery. J Biol Chem 1998; 273: 4972–4975.
Bean MA et al. Cell-mediated cytotoxicity for bladder carcinoma:
evaluation of a workshop. Cancer Res 1975; 35: 2902–2913.
Jaffe E (ed). Culture and identification of large vessel endothelial
cells. In: Biology of Endothelial Cells. Martinus Nijhoff: Boston,
1986, pp. 1–13.
Thornton SC, Mueller SN, Levine EM. Human endothelial cells:
use of heparin in cloning and long-term serial cultivation.
Science 1983; 222: 623–625.
Pollock R, Treisman R. Human SRF-related proteins: DNA-binding properties and potential regulatory targets. Genes Dev 1991;
5: 2327–2341.
Lucibello FC et al. Periodic cdc25C transcription is mediated by
a novel cell cycle-regulated repressor element (CDE). EMBO J
1995; 14: 132–142.