ABSTRACT
In our previous work we have shown that the oligonucleotide 5
'
-GGGGAGGGGGAGG-3
'
gives a very stable and specific triplex with the promoter of the murine c-
pim
-1 proto-oncogene
in vitro
[Svinarchuk,F., Bertrand,J.-R. and Malvy,C. (1994)
Nucleic Acids Res.,
22, 3742-3747]. In the present work, we have tested triplex formation with some
derivatives of this oligonucleotide which are designed to be degradation-resistant inside the cells, and we show that phosphorothioate and the
oligonucleotide with a 3
'
terminal amino group are still able to form triplexes. Moreover these
oligonucleotides, like the 13mer oligonucleotide of similar composition
[Svinarchuk,F., Paoletti,J., and Malvy,C. (1995)
J. Biol. Chem.,
270, 14068-14071], are able to stabilize the targeted duplex.
In vivo
DMS footprint analysis after electroporation of the pre-formed triplex into the cell have shown the presence of the triple helix
inside the cells. This triplex structure partially blocks c-
pim
-1 promoter activity as shown by transient assay with a c-
pim
-1 promoter-luciferase gene construct. To our knowledge these data are the first
direct evidence that conditions inside cells are favorable for triplex
stability with non-modified oligonucleotides. However we were unable to show triplex
formation inside living cells using various methods of oligonucleotide
delivery. We suppose that this may be due to the oligonucleotide being
sequestered by cellular processes or proteins. Further work is needed to find
oligonucleotide derivatives and ways of their delivery to overcome the problem
of triplex formation inside the cells.
Homopurine-homopyrimidine regions in DNA have attracted a great deal of attention in
connection with their possible role in gene regulation in eukaryotes (
1
,
2
). These regions raise the possibility of manipulating gene expression through
artificial triple helix formation (
3
,
4
). Displacement of DNA-bound regulatory proteins from their recognition sites by triplex-forming oligonucleotides might provide a general strategy for the
alteration of sequence-specific function in eukaryotes (
5
). While the formation of an oligodeoxyribonucleotide-directed triple helix as well as its stability were carefully investigated
in vitro
, little is known about its
in vivo
behavior (
3
,
6
). We also have not found any information about triplex formation with non-modified oligonucleotides inside the cells. In the present work, we have
investigated the triplex complex formation with a 13mer polypurine
oligonucleotide 5'-GGGGAGGGGGAGG-3' targeted to the promoter region of c-
pim
-1 gene. This proto-oncogene encodes a highly conserved serine/threonine phosphokinase
which is predominantly expressed in hematopoietic organs and gonads in mammals
(
7
,
8
). Its over-expression is linked to lymphomagenesis in mice (
9
), and this gene can act as an apoptotic inhibitor (
10
). In our previous work (
11
), we have shown that oligonucleotide 5'-GGGGAGGGGGAGG-3' binds to its target sequence in the presence of 50 mM
Na
+
or K
+
, 10 mM MgCl
2
and 20 mM Tris-acetate, pH 7.5. This oligonucleotide is bound in an
antiparallel orientation with respect to the homopurine sequence. As was shown
by co-migration assay, the triplex is stable up to 65oC. At 37oC it was practically irreversible: after 24 h of co-migration assay there was no trace of triplex
dissociation. As we have already shown, 13mer oligonucleotide with the similar
composition even stabilizes its target double-stranded DNA (
12
). In the present work we have characterized triplex formation with
phosphorothioate and 3' end amino-protected oligonucleotide 5'-GGGGAGGGGGAGG-3'
in vivo
and
in vitro
. Melting temperature studies have shown that all the derivatives as well as the
phosphodiester stabilize their double-stranded DNA target. Once pre-formed, triple helix structures are stable for at least 24 h after
electroporation into the cells as judged by
in vivo
DMS footprint studies. Formation of the triplex leads to a decrease in c-
pim
-1
promoter activity in the transient assay with a c-
pim
-1
promoter-luciferase construct. In contrast, when electroporated first into a cell,
target DNA is inaccessible for subsequent triplex formation using various
methods of oligonucleotide delivery (passive addition, cationic liposomes or
electroporation). This work presents direct evidence that the conditions inside
the cell are favorable for the existence of the triplex with non-modified oligonucleotides.
Oligonucleotides were synthesized using the Applied Biosystems 391A DNA
synthesizer and purified by electrophoresis in 20% polyacrylamide denaturating
gels. Oligonucleotides, bearing an
amino group on the 3' end were synthesized on 3'-amino-Mod-C3-CPG columns (GLEN Research Corporation).
Phosphorothioate oligonucleotides were purchased from GENSET.
Construction of the plasmid pPimN1 which contains mouse c-
pim
-1 promoter region (-905 to -14) has been previously described (
11
). Plasmid pPim-Luc was made by inserting the
Bam
HI-
Kpn
I fragment of pPimN1 (after filling-in of the
Bam
HI site with Klenow fragment in the presence of dATP and dGTP) into the
Sal
I-
Kpn
I sites of the vector pcLUC/[Delta]
Kpn
I (after filling-in of the
Sal
I site with Klenow fragment in the presence of dTTP and dCTP). Plasmid pcLUC/[Delta]
Kpn
I is a derivative of the plasmid pcLUC (
13
) with deleted second
Kpn
I site near 3' end of the luciferase gene. As an internal control in transfection
experiments, we used plasmid pCMV-[beta]Gal (Clontech). All plasmids were grown in bacterial strain XL Blue
1 (Stratagene) and purified by CsCl gradient methods (
14
).
A cat fibroblast cell line (G355-5), kindly provided by Dr Thierry Heidmann, was maintained in DMEM
supplemented with 10% foetal calf serum (FCS), 100 [mu]g/ml streptomycin and 100 U/ml penicillin, routinely passaged every 3 days.
Cells which were maintained confluent for 2 days were used for these
experiments. The cells were trypsinized and washed with 10 ml DMEM supplemented
with 10% FCS. After centrifugation the pellets were resuspended in RPMI medium
containing 10 mM MgCl
2
in a volume to give 10
8
cells/ml. Cells (150 [mu]l) were put in a 4 mm electroporation cuvette (Eurogentec), and 20 [mu]l of DNA solution was added. (The quantity of DNA is specified in figure
legends and in the text). After 30 min of incubation on ice, the cells were
electroporated with a single pulse (200 V, 960 [mu]F) with a BioRad electroporator system. Immediately after the pulse, the
clumps of dead cells were removed and the remaining cells were washed three
times with 12 ml of warm (37oC) RPMI medium containing 3 mM EDTA. The cells were then resuspended with
DMEM containing 10% FCS and incubated for 10 or 24 h before treatment.
To check the efficiency of transfection, 5 [mu]g of pCMV-[beta]Gal plasmid was used for electroporation; after 2 days, [beta]-galactosidase activity was checked by a standard X-gal coloration assay (
15
).
Preparation of labeled DNA fragment.
The pPimW2 plasmid (50 [mu]g, 20 pmol) (
11
) was cut with
Cla
I restriction enzyme, 3'-labeled with Klenow fragment of DNA polymerase I in the presence of
80 pmol [[alpha]-
32
P]dCTP (3000 Ci/mmol), and digested with
Bam
HI restriction enzyme. The smaller labeled fragment
Bam
HI-
Cla
I (445 bp) was used for
in vitro
and
in vivo
footprint experiments.
Triple-helix formation.
For
in vivo
assay, 100 pmol of the oligonucleotide or its derivative was added to 1-1.5 pmol of the radiolabeled fragment in 20 [mu]l buffered solution containing 10 mM MgAc
2
, 50 mM NaAc and 30 mM MOPS, pH 7.5. After 1 h of incubation at 37oC, the DNA was used to electroporate 1.5 * 10
7
cells.
Probing with DMS in vitro.
This procedure was performed as described in (
11
). The prepared fragment (~0.5 pmol) was dissolved in 20 [mu]l of the buffer: 50 mM MOPS, pH 7.2, 50 mM NaAc and 10 mM MgAc
2
. Then 20 pmol of the oligonucleotide 5'-GGGGAGGGGGAGG-3' was added. The mixture was incubated 1 h at room
temperature. Then 2 [mu]l of 5% DMS was added and the reaction was performed for 4 min at 24oC. The reaction was stopped by the addition of a 5 [mu]l solution containing 10% [beta]-mercaptoethanol, 1 mM EDTA and 0.1 M NaAc. After double
precipitation in ethanol, the samples were treated with 50 [mu]l of 10% piperidine at 95oC for 20 min and the cleavage products were separated in 6%
polyacrylamide denaturating gels.
Probing with DMS in vivo.
Twelve or 24 h after electroporation, cells were washed five times with a
solution of 0.9% NaCl and 2 mM of EDTA. Then 5 ml of 0.5% DMS was added in a
buffered solution containing 0.9% NaCl, 10 mM MgAc
2
, 50 mM MOPS pH 7.5, and the reaction was performed for 4 min at room
temperature. The reaction was stopped by brief washing with 0.9% NaCl solution
followed by the addition of the same solution with 1% [beta]-mercaptoethanol for 2 min. To extract the DNA fragment, cells were
lysed in 5 ml of solution: 20 mM EDTA, 0.5% SDS and 50 mM Tris-HCl, pH 7.5. Cellular DNA was precipitated by centrifugation at 30 000
g
for 30 min after 2 h of incubation of the cell lysate with 2.5 ml of a 3 M
solution of NaAc, pH 5.0, on ice. The DNA fragment was precipitated by
addition of an equal volume of isopropanol to the cell supernatant. The
precipitate was dissolved in 200 [mu]l water, extracted once with an equal volume of phenol:chloroform (1:1) and
precipitated with 2.5 vol of ethanol at -20oC for 2 h. The DNA was collected by centrifugation for 10 min,
washed once with 75% ethanol and allowed to dry for 5 min at room temperature.
The samples were treated with 50 [mu]l 10% piperidine at 95oC for 20 min, and the products of cleavage were separated in 6%
polyacrylamide denaturating gels.
The level of guanine protection in DMS footprinting experiments was estimated
using a Fujix Bas 1000 PhosphoImager. This value for each type of the
oligonucleotide (x
i
) was calculated according to the equation:
x
i
= (1 - I
cc
/I
tc
* I
ti
/I
ci
) * 100%
where I
c
indicates the sum of the intensity of the five bands located near the site of
protection, and I
t
is the sum of the intensity of the five bands inside the triplex forming
region. Additionally, `c' (for the control oligonucleotide) and `i' (for the
triplex-forming oligonucleotides) indicate the line in Figure
4
(and correspondingly the type of oligonucleotide we have used in the
experiment). Labels I
cc
and
I
tc
indicate control oligonucleotide (lines 4 and 7 for the
in vivo
and
in vitro
experiments correspondingly); I
ti
and I
ci
indicate triplex-forming oligonucleotides (
in vivo
experiments: lines 1, 2 and 3 for phosphorothioate, oligonucleotide with a 3' terminal group and phosphodiester correspondingly;
in vitro
experiments: line 6 for phosphodiester).
Absorbance of the oligonucleotide mixture was measured at 258 nm as a function
of temperature with an Uvicon 941 spectrophotometer equipped with Huber
cryotermostat and Huber PD410 temperature programmer through software developed
for
T
m
recording. The rate of temperature increase/decrease was 0.5oC/min. The buffer contained 10 mM MgAc
2
, 50 mM NaAc and 20 mM Tris-acetate, pH 7.5, with each oligonucleotide at a
concentration of 1.15 [mu]M. Before melting studies all samples were heated to 80oC for 15 min and then allowed to return slowly to room temperature.
Passive addition.
After electroporation of the duplex DNA, the cells were incubated in 10 ml DMEM
supplemented with 10% FCS which contained 5 [mu]M phosphorothioate or 3'-amino group-protected oligonucleotide.
Electroporation
. After electroporation of the duplex, the cells were incubated in 10 ml DMEM
supplemented with 10% FCS for 2 or 10 h. After washing twice with PBS to remove
the dead cells, the cells were trypsinized and electroporated, as describe
above, in the presence of 15 [mu]M oligonucleotide with either an amino or phosphorothioate group at the 3' end. After a second electroporation, the cells were washed twice to
remove non-internalized oligonucleotide.
Cationic liposomes. A
fter DNA electroporation, the cells were incubated for 2 h until attachment of
the cells to the plate; then they were washed twice and incubated with 10 ml
DMEM without serum containing 100 [mu]l of DOTAP:DOPE liposomes (dioleoyl phosphatidyl ethanolamine, 1:1,
concentration 69 [mu]M in lipids; kindly provided by Dr F. Adnet) (
16
) which had been previously incubated for 10 min with 2 nM oligonucleotide.
To estimate the ability of triplex formation to down-regulate the c-
pim
-1
promoter activity, we performed a co-transfection assay with the plasmids pPim-Luc and pCMV-[beta]Gal (5:1). For each electroporation we used 5 [mu]g (2 pmol) of pPim-Luc plasmid after pre-incubation with 200 pmol of the triplex-forming or control oligonucleotide
using the same conditions as for the
in vivo
footprint experiments. For these studies, 48 h after electroporation cells were
washed five times with 0.9% NaCl and harvested in 5 ml 0.9% solution of NaCl
with 2 mM EDTA with the help of a cell lifter. After centrifugation, cells
were lysed in 150 [mu]l buffered solution containing 0.25 M Tris-phosphate, pH 7.8, 1 mM EDTA, 1%
Triton X-100, 15% glycerol, 1 mM DTT and 0.4 mM phenylmethyl-sulfonyl fluoride (PMSF) by three freeze-thaw cycles using dry ice/ethanol and a 25oC bath. The extracts were then centrifuged in a
microtube to pellet cell debris. Luciferase activity was assayed on the
biocounter M2010 Lumac/3M in buffer containing 10 mM MgCl
2
, 0.2% BSA, 0.1 mM D-luciferin (Sigma), 1 mM ATP, 0.25 M Tris-phosphate, pH 7.8, 1 mM EDTA, 1% Triton X-100, 15% glycerol, 1 mM DTT and 0.4 mM PMSF, integration time
60 s. Cellular extract (20 [mu]l) from 3 * 10
6
cells was used for each determination. The [beta]-galactosidase activity was measured by the method described in (
17
). Cellular extract (2 [mu]l) was used for each determination; [beta]-galactosidase activity was expressed as OD units after 30 min of
incubation (linear part of the coloration reaction). As a measure for the efficiency of down-regulation through triplex formation, we compared the ratio of luciferase:[beta]-galactosidase activity of cells which had been co-electroporated with the plasmid pPim-Luc preincubated with triplex-forming oligonucleotide, or a control
oligonucleotide and plasmid pCMV-[beta]Gal. Errors in the percent inhibition were calculated as the maximum
deviation from the average value in three independent transfection experiments.
In previous work (
11
), we have shown that the oligonucleotide 5'-GGGGAGGGGGAGG-3' forms a very stable and specific triple helix with
its target DNA in the c-
pim
-1
promoter region, and we have studied
in vitro
triplex formation with this oligonucleotide in physiological buffer conditions.
In the present work, we investigate the stability of this triplex directly
inside cells. Because oligonucleotides are normally quickly degraded in cell
culture medium and in cells, we first tried different oligonucleotide
derivatives, known to be more resistant to nuclease digestion, for their
ability to form triplex
in vitro
. As shown in Figure
1
, either non-modified phosphodiester oligonucleotide, oligonucleotide bearing an amino
group on its 3' end or phosphorothioate equally protect guanines in the purine strand of
the target, indicating triplex formation under experimental conditions. In
order to quantify the stability of the triplex structures with the
oligonucleotide derivatives, we have performed UV spectroscopic melting
temperature studies (Fig.
2
). In all cases (duplex or duplex plus the derivative of the third strand
oligonucleotide) we observed only one transition, and the melting temperature
of the three-strand structure is always higher than for the duplex. (The small
transition at the lower temperature could be explained by formation of a
partial duplex between the third strand oligonucleotide 5'-GGGGAGGGGGAGG-3' and the pyrimidine strand of the duplex. This is in
agreement with our unpublished observation, that the third strand
oligonucleotide 5'-GGGGTGGGGGTGG-3', which gives the triplex the same stability, at the
same time does not give additional transitions at lower temperature.) As we
have already shown (
12
), the third strand purine oligonucleotide 5'-GGAGGGGGAGGGG-3' binds to the targeted DNA in an antiparallel
orientation and stabilizes the targeted duplex under the same conditions.
In vitro
pre-formed triplex, between oligonucleotide 5'-GGGGAGGGGGAGG-3', or its derivatives, and the radioactively
labeled fragment
Cla
I-
Bam
HI of the c-
pim
-1 promoter, was delivered into the cells by electroporation. Because we
cannot exclude that the triplex might also be present on the surface of the
cells after electroporation, we have determined the conditions for triplex
stability in order to dissociate the triplex outside the cells. As shown in
Figure
3
, addition of EDTA completely dissociates the pre-formed triplex. This is in agreement with our own results (
11
,
12
) and those of others (
18
,
19
) that divalent cations are crucial for purine(purine-pyrimidine) triplex formation. This fact makes it possible to dissociate
an external triplex by washing the cells with EDTA solution immediately after
electroporation.
To estimate stability of the pre-formed triplex inside the cells, we used
in vivo
DMS footprinting. In these experiments, pre-formed triplex made with the radiolabeled
c-
pim
-1 promoter fragment with either the oligonucleotide or its derivatives was
electroporated into the cells, which were subjected to DMS treatment 12 or 24 h
later. As a control, we used the same DNA fragment after pre-incubating with the phosphorothioate oligonucleotide 5'-TGAACACGCCATGTC-3' which does not form triplex
in vitro
. It appears that there is protection of the guanines in the target sequence,
which indicates presence of the triplex inside the cells (Fig.
4
). Additional proteinase K and DNase treatment of the cells (10 min with
proteinase K 1 mg/ml in 0.9 M NaCl, followed by subsequent washing with DMEM
containing 10% serum, DMEM without serum and treatment with DNase 0.1 mg/ml)
does not change the DMS footprint picture. This confirms the intracellular
triplex localization. No protection is detected in the experiment with the
control oligonucleotide, which indicates that protection of the target sequence
is really due to the triplex formation and does not reflect specific protein(s)
binding to the target sequence. At the same time, the level of protection is
lower than in the
in vitro
experiments. Estimates made with the help of the PhosphoImager give 60%
protection for
in vivo
experiments, compared with 90% protection obtained
in vitro
. This level of protection is consistent for all of the derivatives and is the
same for non-modified phosphodiester, indicating protection against nucleases of the
non-modified oligonucleotide in triple helix structure. After 12 or 24 h, the
level of protection was nearly the same. Possible mechanisms for this
phenomenon are discussed below.
To evaluate the formation of triplex inside the cells, we used the following
methods of oligonucleotide delivery: (i) passive addition of the
phosphorothioate to the culture medium at a concentration of 5 [mu]M 1 h after DNA electroporation; (ii) cationic liposomes with the
oligonucleotide bearing an amino group at the 3' end; (iii) second electroporation of the cells in the presence of 15 [mu]M oligonucleotide either bearing an amino group or phosphorothioate
at the 3' end; (iv) co-electroporation of the radiolabeled fragment (2 pmol) and the
triplex-forming oligonucleotide (2500 pmol). Even in the last case we did not
detect any triplex formation inside the cells. (Cell were electroporated
without Mg
2+
to avoid any triplex pre-formation.) Details of the experiments are described in Material and
Methods. No triplex formation in these experiments was observed (results not
shown).
To estimate the influence of the triplex formation on the c-
pim
-1 promoter activity, we performed a co-transfection assay with the plasmids Pim-luc and pCMV-[beta]Gal. In the case of the pre-formed triplex with 3'-amino-protected oligonucleotide
the ratio luciferase activity:[beta]-galactosidase activity was 5870 +- 6%; in the presence of the control 3'-amino-protected oligonucleotide 5'-GGcGAGGcGGAGG-3' it was 7360 +- 5%.
When the triplex-forming oligonucleotide or the control one were added after cell
transfection this ratio was 7460 +- 7% and 7480 +- 6% correspondingly. These data are taken from three independent
experiments and they clearly show reproducible down-regulation of the c-
pim
-1
promoter activity in the case of pre-formed triplex (20% inhibition). In accordance with the data from the
footprint assay, there is no inhibition of promoter activity when triplex-forming oligonucleotide is added to the cells after electroporation.
In our previous work, we found that the oligonucleotide 5'-GGGGAGGGGGAGG-3' gives a very particular triplex structure with its
target sequence in the c-
pim
-1 proto-oncogene promoter (
11
).
In vitro
parameters of the triplex formation are very different from those measured by
other authors (
3
,
20
): (i) the triplex is stable at 65oC in the co-migration assay; (ii) one mismatch in the target sequence completely
abolishes triplex formation as judged by the co-migration assay; (iii) the value of the dissociation constant can be
estimated to be <1.5 * 10
-9
M; (iv) once formed, the triplex is practically irreversible at 37oC: after 24 h of co-migration assay, there was no trace of triplex dissociation.
Following the
in vitro
data it was very tempting to test this oligonucleotide for its ability to form
triplex inside the cells.
As a first step, we have investigated
in vitro
triplex formation with different oligonucleotide derivatives which are known to
be more resistant to nucleases: phosphorothioates and oligonucleotides with an
amino group at the 3' end. DMS footprint and melting temperature studies have shown that all
tested oligonucleotide derivatives form triplexes with the same affinity as non-modified phosphodiester oligonucleotide. The behavior of the
phosphorothioate oligonucleotide is in agreement with previously published data
that purine-rich oligonucleotides containing phosphorothioate linkages form triplexes
with affinities similar to those of the corresponding natural phosphodiester
oligonucleotides (
21
). However the fact that phosphorothioates and oligonucleotides with an amino
group at the 3' end stabilize their target double-stranded DNA is new.
In a second round we have studied the stability of the pre-formed triplexes inside the cells. As a measure of the triplex stability
we have used our modification of the
in vivo
DMS footprint method. The main feature of this approach is electroporation of
the 3' labeled DNA fragment into the cells. Experiments have shown that even 24
h after electroporation the DNA is stable enough to obtain a readable
footprint. Experiments performed with pre-formed triplexes have shown that all types of oligonucleotides
(phosphodiester, phosphorothioate and oligonucleotide with 3' amino group) do protect targeted DNA inside the cells. In all of our
experiments, we obtained a lower level of the guanine protection than in
in vitro
experiments. One can suggest two possibilities to explain this phenomenon.
First, it might be due to specific protein(s) which recognize the triple-stranded DNA and make the purine strand of the duplex more accessible to
DMS treatment. Triplex DNA-binding proteins have already been described (
22
). Secondly, we can not exclude a strong attachment of a portion of the DNA to
the cell's surface proteins after electroporation which makes the triplex
accessible to disruption by EDTA treatment.
The reported results are the first direct proof to our knowledge that conditions
inside the cell are favorable for the existence of the triplex with non-modified oligonucleotides. If the pre-formed triplexes are stable inside the cell, one expects that they
would form inside the cell when an appropriate delivery route is used for the
oligonucleotide.
In a third round of our experiments our aim was to know whether triplex
formation is possible inside the cells. With this purpose in mind, we used
three different methods of oligonucleotide delivery.
The phosphorothioate oligonucleotide was added to the culture medium after 1.5 h
of electroporation of the labeled c-
pim
-1 promoter fragment. After 12 h of cell growth in the presence of 5 [mu]M oligonucleotide, there was no detectable protection of the target
sequence. This was not due to phosphorothioate degradation. In agreement with
the published data (
23
) we observed full-length oligonucleotide in the culture medium and in cells lysate even
after 24 h of cell culture (data not shown). This result may be explained by
different localizations of the electroporated duplex and the oligonucleotide:
electroporated DNA is going to the nucleus and the localization of passively
added oligonucleotides is predominantly cytoplasmic (
24
).
The delivery of the oligonucleotide protected by an amino group on its 5' end with the help of cationic liposomes, which are very efficient in
gene transfection experiments, also does not lead to triplex formation inside
the cells. Efficiency of delivery with this type of liposome was tested with
pCMV-[beta]Gal plasmid in a transient assay (typically more than 90% of
transfected cells) and with fluorescein-labeled oligonucleotides in confocal microscopy experiments (all
oligonucleotide bound to the cells after a few hours was located in the
nucleus) (Dr F. Adnet, personal communication).
No triplex formation was obtained in co-electroporation experiments or after electroporation of the triplex-forming oligonucleotide either 2 or 10 h after electroporation of
the radiolabeled c-
pim
-1 promoter fragment.
We suggest that the failure to form triplex inside the cells was not due to the
high K
+
concentration inside the cells: we were able to see ~50% protection of the targeted site in the
in vitro
DMS footprinting experiments in the buffer containing 10 mM MgAc
2
, 40 mM MOPS pH 7.5, 150 mM KCl and 2.5 [mu]M of the triplex-forming oligonucleotide after 2 h incubation with 0.5 pmol of the
fragment (2.5 [mu]M oligonucleotide corresponds to the concentration after electroporation of
the cells in the presence of 2500 pmol oligonucleotide). Probably, this failure
is due to quick trapping of the oligonucleotides by cellular proteins or
different compartmentalization of the oligonucleotide and double-stranded DNA inside the cells. Some proteins which specifically bind
oligonucleotides have been described (
25
). Another possibility is that while being electroporated, plasmid DNA quickly
adopts a nucleosomal structure (
26
), which prevents triplex formation (
27
).
The oligonucleotide 5'-GGGGAGGGGGAGG-3' is targeted to the PPF 348 site (
28
) in the c-
pim
-1
promoter. As was shown in transient assay, deletion of the region of ~50 bp which contains the AP2 and the PPF 348 sites leads to a 60% decrease
in human c-
pim
-1 promoter activity (
28
).
In vitro
footprint experiments have shown the existence of a nuclear factor specifically
interacting with the PPF 348 site (
28
). High levels of homology between mouse and human c-
pim
-1
promoters in this region (only four mismatches between mouse and human sequences
at the length of 70 bp) underlines the importance of the PPF 348 site for c-
pim
-1 promoter function. In our hands transient transfection experiments with
pre-formed triplex show a reproducible down-regulation of the c-
pim
-1 promoter activity (20% inhibition). This value looks reasonable compared
with 60% inhibition obtained in deletion analysis. We suggest that this down-regulation might be due to prevention of the regulatory protein(s) binding
to the PPF 348 site by the triplex-forming oligonucleotide.
In complete agreement with the result of the experiment of
in vivo
DMS footprint, c-
pim
-1
promoter down-regulation by triplex-forming oligonucleotide is efficient only in the case of pre-formed triplex, and passive addition of the 3' end protected oligonucleotide does not influence c-
pim
-1
promoter activity.
Our data on the inhibition of c-
pim
-1
promoter activity in the transient assay are in a good agreement with those of
Grigoriev
et al
. (
3
,
5
), who were able to inhibit promoter activity of IL2R only after pre-formation of the triplex
in vitro
or
in situ
with acridine or psoralene-conjugated oligonucleotides.
Specific inhibition of endogenous genes (
4
,
29
) or virus propagation (
30
) through triplex formation inside the cells has also been demonstrated. The
data suggesting triplex formation inside living cells in these works are
indirect, and the growing number of publications demonstrating `specific'
inhibition of genes by oligonucleotides with non-related sequences (
31
-
36
) shows that direct proofs concerning the mechanisms of oligonucleotides action
are of great importance.
Our data present direct proof that the conditions inside the cell are favorable
for the existence of a triplex with non-modified short oligonucleotides. However, in each study the possibility to
extrapolate triplex formation in cells from
in vitro
experiments should be carefully investigated. Despite the failure to form
triplexes inside the cells after different routes of oligonucleotide delivery,
these data offer the possibility that triplex formation could play a role in
natural genetic processes (gene regulation, recombination events). Further work
is needed to discover new oligonucleotide derivatives and new ways of delivery
to overcome the problem of triplex formation inside the cells.
We thank Dr E. Lescot for oligonucleotide synthesis and Dr U. Hazan for
providing the pcLUC vector. We are very grateful to Dr J. Paoletti who gave us
the software for
T
m
recording and to Drs J. Tanaka and M. Lee for fruitful discussion. This work
was supported by an Agence Nationale de Recherches sur le SIDA research
fellowship to F.S., an IFSBM fellowship to A.D. and by INSERM grant no. 94 EO
08.
REFERENCES
Return