0 1992 by The American Society for Biochemistry and Molecular Biology, Inc. T H E JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 267, No. 5, Isme of February 15, pp. 3389-3395,1992

Printed in U.S.A.

A Triple Helix-forming Oligonucleotide-Intercalator Conjugate Acts as a Transcriptional Repressor via Inhibition of NF KB Binding to Interleukin-2 Receptor a-Regulatory Sequence*

(Received for publication, March 29, 1991)

Mikhail GrigorievSQ, Danielle Praseuthl, Philippe Robin$, Agnes Hemar (I, Tula Saison-Behmoarasl, Alice Dautry-Varsatll , Nguyen T. Thuong**, Claude Helenell, and Annick Harel-Bellan$ $$ From the SLaboratoire d’lmmunobwlogie, Unite de Recherche Associie 1156 Centre National de la Recherche Scientifique, Znstitut Gustave Roussy, 94800 Villejuih France, the llLaboratoire de Bwphysique, Znstitut National de la Santk et de la Recherche Medicale U.201, Museum National d’Histoire Naturelle, 75005 Paris, France, the (1 Unite de Genktique Somatique, Znstitut Pasteur, 75724 Paris, France, and the **Centre de Bwphysique MoEculaire du Centre National de la Recherche Scientifique, 45071 OrEans, France

Oligonucleotide-directed triplex formation within upstream regulatory sequences is envisioned as a po- tential tool for gene inhibition. However, this approach requires that triple helix-forming oligonucleotides are chemically modified, so that the triplex is stable under physiological conditions. Here, we have compared sev- eral chemical modifications of an oligonucleotide, targeted to a natural 15-base pair homopyrimi- dine homopurine sequence located in the upstream regulatory region of the gene encoding the interleukin- 2 receptor a chain (pS5, IL-2 Ra). Methylation of the cytosines strongly stabilized the triplex. Further at- tachment of an intercalating agent (acridine) dramat- ically increased the stability of the triplex, as assessed by T, measurements or by band shift assays. Further- more, the acridine-derivatized oligonucleotide was more efficient in competing away high affinity DNA- binding proteins, as assessed by restriction enzyme inhibition assays. Using a novel footprinting assay, we have further shown that the interaction of the meth- ylcytosine-substituted, acridine-derivatized oligonu- cleotide with a plasmidic target, harboring the IL-2 Ra regulatory region, remains highly sequence specific, occurs at physiological pH and is independent of the superhelicity of the plasmid. Acridine derivatization did not impair the exquisite target specificity of triplex formation, since the derivatized oligonucleotide inhib- ited the binding of nuclear proteins to the overlapping NF NB enhancer sequence on an IL-2 Ra target and not on the related human immunodeficiency virus long terminal repeat target. Finally, the oligonucleotide in- hibited the NF KB-dependent tax-induced transcrip- tional activation of the IL-2 Ra chloramphenicol ace- tyltransferase construct in live cells, whereas it did not have any effect on a human immunodeficiency virus long terminal repeat chloramphenicol acetyl-

* This work was supported by grants from the Agence Nationale de Recherches sur le SyndrBme d’Immunod6ficience Acquise and the Association pour la Recherche sur le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

8 Supported by a fellowship from the Agence Nationale de Re- cherches sur le Syndrijme d’Immunodeficience Acquise. $$To whom correspondence should be addressed Laboratoire

d’Immunobiologie, URA 1156 CNRS, Pavillon de Recherche No. 1, 39 rue Camille Desmoulins, 94805 Villejuif Cedex, France. Tel: 33-1- 45-59-45-15; Fax: 33-1-47-26-92-74.

transferase construct. We conclude that this modified oligonucleotide acts as a transcriptional repressor for the IL-2 Ra gene via triple helix formation with regu- latory sequences.

Homopurine-homopyrimidine sequences are frequently found in gene upstream regulatory regions, in which they are suspected to play a function in gene expression (Kedes, 1979; Cheng et al., 1982; Crabtree and Kant, 1982; Richards et al., 1983; Htun, et al., 1984; Htun et al., 1985; Boles and Hogan, 1987). In uitro and under stringent physicochemical condi- tions, i.e. low pH, high degree of superhelicity of the carrier plasmid, and high salt and magnesium, long enough homo- purine-homopyrimidine sequences are known to form intra- molecular triple-stranded structures (Kohwi-Shigematsu and Kohwi, 1985; Pulleyblank et aL, 1985; Evans and Efstratiadis, 1986; Hanvey et al., 1988; Htun and Dahlberg, 1988; Johnston, 1988; Kohwi and Kohwi-Shigematsu, 1988; Voloshin et al., 1988). Site-specific triple helix formation on large double- stranded target DNAs can be directed in uitro using short synthetic oligonucleotides (Le Doan et aL, 1987; Moser and Dervan, 1987; Cooney et al., 1988; Francois et aL, 1988,1989a; Lyamichev et aL, 1988; Praseuth et al., 1988; Vlassov et al., 1988; Rajagopal and Feigon, 1989; Birg et al., 1990; Perrouault et al., 1990; Strobe1 and Dervan, 1990), which raises the possibility of manipulating gene expression through artificial triple helix formation. A homopurine oligonucleotide, de- signed to form a Pur. Pyr . Pur’ colinear triplex, was shown to inhibit the in uitro transcription of the c-myc gene (Cooney et al., 1988). In an alternative approach, homopurine-homo- pyrimidine double-stranded DNA can be selectively recog- nized by homopyrimidine oligonucleotides that bind to the major groove and form local triple helices. Recognition in- volves the formation of Hoogsteen hydrogen bonds between the oligonucleotide bases and the purines of Watson-Crick base pairs (for a review, see H616ne and Toulm6 (1990)). Triple helix formation blocks restriction enzyme cleavage at sites located in the vicinity (Maher et al., 1989; Francois et al., 1989b; Hanvey et al., 1990). Furthermore, a homopyrimi-

The abbreviations used are: Pur, purine; Pyr, pyrimidine; IL-2 R, interleukin-2 receptor; HIV, human immunodeficiency virus; LTR, long terminal repeat; bp, base pair(s); CAT, chloramphenicol acetyl- transferase; TH, triple helix; Hepes, 4-(2-hydroxyethyl)-l-piperazine- ethanesulfonic acid.

3389

3390 Inhibition of NF KB via Oligonucleotide-directed Triplex Formation

dine oligonucleotide has been used to inhibit binding of a eukaryotic transcriptional factor to a promoter in which a 21- bp target sequence had been artificially introduced (Maher et al., 1989). However, given the stringency of physiological conditions for triplex stability, designing efficient chemical modifications of the third strand oligonucleotide is the current limiting step for application of this technology in cell culture. Here, we demonstrate the use of such a homopyrimidine- derivatized oligonucleotide to direct formation of a Pyr . Pur. Pyr triplex on a natural target, the gene encoding the a chain of the interleukin-2 receptor (IL-2R).

High affinity IL-2 receptors are formed at the surface of T lymphocytes by at least two subunits, the p55 (a) and the p70 (p) chain, which are, independently, able to bind IL-2 with a low to intermediate affinity (Greene et al., 1985; Sharon et al., 1986; Tsudo et al., 1986; Kehrl et al., 1988; Hatakeyama et al., 1989). The induction of p55 (a) upon mitogenic or antigenic stimulation is an important step in T cell activation (Dukovich et al., 1987). Furthermore, the IL-2 receptor a chain is constitutively expressed in cells infected by the human T cell leukemia virus I (Leonard et al., 1984; Inoue et al., 1986; Maruyama et al., 1987) and is suspected to be a part of an autocrine pathway for malignant proliferation. It is therefore of considerable interest to set up powerful means designed to manipulate the expression of this gene. IL-2 Ra transcription is regulated through upstream sequences, which include an NF &-like element (Fig. l ) , homologous to the sequence responsible for kappa light chain modulation in B cells or human immunodeficiency virus (HIV) LTR-directed transcription in T lymphocytes (Bohnlein et al., 1988; Leung and Nabel, 1988; Cross et al., 1989; Lowenthal et al., 1989; Pomerantz et al., 1989; Toledano et al., 1990). Binding of the cognate factor to the NF KB-like site is inducible by various stimuli, including the T cell mitogen phytohemagglutinin and the tumor promoter phorbol myristate acetate (Lowenthal et al., 1988, 1989; Bohnlein et al., 1989). This NF KB site is indispensable for upregulation of the gene by the HTLV-I transactivator tax protein (Leung and Nabel, 1988; Ruben et al., 1988; Cross et al., 1989). The presence of a 15-bp homo- purine-homopyrimidine sequence overlapping factor-binding sites, including the NF KB element (Fig. 1) makes the IL-2 Ra chain an interesting model for oligonucleotide-directed triplex formation. We therefore used a 15-base homopyrimi- dine oligomer designed to form a triplex on this target, i.e. in a parallel orientation with respect to the target homopurine strand (Le Doan et al., 1987; Moser and Dervan, 1987; Vo- loshin et al., 1988; Praseuth et al., 1988). Pyr Pur Pyr triplex formation is highly pH-dependent, since-cytosines are re- quired to be protonated in order to form Hoogsteen pairs. Hence, this type of triple helix is not stable at neutral pH (Evans and Efstratiadis, 1986; Htun and Dahlberg, 1988). However, replacement of cytosines by 5-methylcytosines in the third strand oligonucleotide has been shown to reduce the stringency of this pH requirement (Lee et al., 1984; Strobe1 et al., 1988, Sun et al., 1989). Triple helices are also stabilized by covalent addition of an intercalating agent (Sun et al., 1989). In this paper, we have tested the effect of B-methylcy- tosine substitution and acridine derivatization on the stability and target specificity of the triplex formed on the IL-2 Ra natural target sequence. Furthermore, we show that triplex formation prevents the binding of regulatory factors to the NF KB regulatory sequence in vivo and inhibits tax-induced transcriptional activation of the IL-2 Ra CAT construct in live cells.

MATERIALS AND METHODS

Oligonucleotides and Plasmids-Unmodified oligonucleotides were synthesized on an Applied Biosystem DNA synthesizer, either at the Pasteur Institute (Paris, France) or at the Gustave Roussy Institute (Laboratoire d'oncologie Molkculaire). Methylcytosine-containing derivatives were synthetized at the Pasteur Institute. The double- stranded target sequence for T,,, measurements was 5' GGA ATC TCC CTC TCC TTT TAT GGG C 3'/3' CCT TAG AGG GAG AGG AAA ATA CCC G 5', and that used in enzyme inhibition

extension studies was: 5' AGT TCA ATT GCT GGA GGT GTG 3'. assays is shown in Fig. 1. The sequence of the primer used in primer

The sequence of the double-stranded HIV oligonucleotide is described in Harel-Bellan et al. (1989) and that of the double-stranded c-fos dyad symmetry element in Trouche et al. (1991). Acridine derivati- zation was performed according to Thuong and Chassignol(l988) by linking 2-methoxy-6-chloro-9-aminoacridine to the 5'-phosphate via a pentamethylene bridge.

The IL-2 Ra-CAT plasmid (the kind gift of Dr. W. Leonard, NIH, Bethesda, MD) harbors the IL-2 Ra promoter sequence from -352 to +110. The HIV LTR-CAT plasmid includes the HIV regulatory sequence from -160 to +80 and is the kind gift of Dr. R. Gaynor. For transfection experiments, both upstream regulatory sequences have been subcloned between the AccI and Hind111 (for IL-2 Ra) and AuaI and Hind111 (for HIV-LTR) sites of PGEM-CAT C (the kind gift of Dr. 0. Brison, IGR, Villejuif, France). The pMAX-neo, tax expression vector is the kind gift of Dr. Pavlakis, NCI, Frederick, MD (Felber et al., 1985), and pRSV2-neo is the kind gift of Dr. M. C. Dockhelar, IGR, Villejuif, France.

Triplex Melting Experiment-Equimolar amounts (1.2 WM each) of double-stranded 25-bp target and 15-mer third strand oligonucleo- tides were mixed in a buffer containing 100 mM NaC1, 10 mM cacodylate, pH 6.5, supplemented, when indicated, with 0.4 mM spermine. Experiments were performed as described in Sun et al. (1989). Melting curves were obtained by substracting the absorbance (measured at 260 nM) of duplex alone from the absorbance of the three-strand mixture at each temperature. The T, was determined as the temperature of half-dissociation of the third strand from the duplex.

Gel Shift-The purine strand of the 47-mer duplex (IL-2 Ra promoter sequence from -290 to-244; see Fig.1) was end-labeled using T4 polynucleotide kinase and [y3'P]ATP. The duplex was formed by heating to 90 "C the labeled strand (10 pmol) with a 5-fold molar excess of the complementary strand and cooling slowly to room temperature (in these conditions, more than 95% of the labeled strand was incorporated in the duplex). After incubation with increasing amounts of 15-methylcytosine TH or 15-methylcytosine acridine T H for 30 min in 40 mM Tris acetate (pH 7), 1 mM EDTA (TEA buffer) supplemented with 0.4 mM spermine, the products were analyzed on a 20% non-denaturing acrylamide gel in the same buffer, with recir- culation of the buffer, at 4 "C (duration of the run, 18 h).

Restriction Experiments-The 32P-end-labeled 47 bp duplex was incubated with increasing concentrations of 15-methylcytosine TH or 15-methylcytosine acridine TH, in a buffer containing 100 mM NaC1, 10 mM MgC12, 12 mM Tris-HC1 at pH 6.5 for HinfI and pH 7.5 for MnlI, 5 mM 2-mercaptoethanol, 100 pg/ml bovine serum albumin, and 0.4 mM spermine. After digestion at 30 "C for 30 min, restriction products were analyzed on a denaturing 20% acrylamide gel. Bands were excised and counted, and the percentage of inhibition was calculated.

DNase Z Footprinting-Plasmids (3 pg) were incubated with in- creasing concentrations of 15-methylcytosine acridine TH or 11- methylcytosine oligonucleotides in 25 mM Tris acetate (pH 5.1 or 6.8) or Hepes (pH 7.8), 70 mM NaCl, 20 mM MgClz (total volume, 4 pl) for 1 h at room temperature. DNase I digestion was performed for 1 min at room temperature. DNase I degradation products were dena- tured in 0.1 N NaOH, 0.1 mM EDTA for 5 min at room temperature. After neutralization and precipitation, the degradation products were annealed with 32P-end-labeled primer for 1 h at 37 'C and extended using Sequenase enzyme (U. S. Biochemical Corp.) and the extending nucleotide mix, according to the recommendation of the manufac- turer. Sequencing reactions were run in parallel in each experiment using the same primer. Extension products were analyzed on a se- quencing gel.

Protein Gel Shift-The IL-2 Ra CAT construct was cleaved with ClaI, end-labeled using the Sequenase enzyme and [a-32P]dCTP (5 min, 37 "C), and cleaved with XmnI, and the labeled fragment (in- cluding the IL-2 Ra upstream sequence from -347 to -226) was

Inhibition of NF KB via Oligonucleotide-directed Triplex Formation 3391

purified on a 4% nondenaturing acrylamide gel. The HIV LTR fragment was prepared by polymerase chain reaction amplification of the -160 to +10 sequence of the corresponding plasmid, using a forward 32P-end-labeled primer.

The end-labeled fragment was incubated with 15-methylcytosine acridine TH or 11-methylcytosine acridine oligonucleotides (20 p M final concentration; total volume, 4 pl in Tris acetate (pH 6.8) supplemented with 4 mM spermidine) for 1 h at room temperature. Protein nuclear extracts (prepared according to Dignam et al., 1983) from phorbol myristate acetate-treated (4 h) IARC 301 or Jurkat cells (both are human tumor T cell lines) were then added (20 pg in 10 p1, supplemented with 0.6 mg/ml salmon sperm DNA and 4 mM sper- midine). After 10 min at room temperature, the products were ana- lyzed on a nondenaturing 4% acrylamide gel.

2 Ra CAT or HIV LTR CAT) were preincubated for 2 h with 15- Transfections-Before transfection, 2 pg of reporter plasmids (IL-

methylcytosine acridine TH oligonucleotide (total volume, 10 pl in 10 mM potassium/dipotassium phosphate buffer supplemented with 2.5 mM EDTA and 0.5 M NaCI). HSB2 cells (human tumor T cell line, 10' cells) were incubated on ice with the reporter plasmid and 5 pg of pMAX-neo or pRSV2-neo (5 min in 150 pl of RPMI 1640 culture medium (5% fetal calf serum) supplemented with 2 mM spermidine) and transfected by electroporation (250 V, 960 microfar- ads) using a Bio-Rad gene pulser (Cann et aL, 1988). Cells were harvested 24 h later, and CAT assays were performed according to Gorman et al. (1982).

Transfection efficiency was measured by direct estimation of in- tracellular plasmid according to McIntyre and Stark (1988), using Zeta-probe blotting membranes (Bio-Rad). The probe was a HindIII- EcoRI fragment of the CAT gene, 32P-labeled using a random priming labeling kit (Amersham Corp.). Within an experiment, variability was less than 20%.

RESULTS

Effect of Various Chemical Modifications on Triple Helix Melting Temperature-As an affinity index, we first compared the melting temperatures (T,) of triplexes formed by modified oligonucleotides on a 25-mer double-stranded synthetic se- quence as a target. This sequence extends from positions -264 to -240 of the IL-2 Ra promoter region (see Fig. 1). Triplex stability was measured by the hyperchromic effect induced upon dissociation of the third strand from the double- stranded 25-mer target (results of a typical experiment are shown in Fig. 2). The T,,, values were calculated as the temperature of half-dissociation, and the data are summarized in Table I. The unmodified oligonucleotide showed a T, value of 10 "C in the absence of polyamines, which was increased to 26 "C by the addition of spermine to 0.4 mM. Substitution of 5-methylcytosines for cytosines (oligonucleotide 15-meth- ylcytosine TH) resulted in a substantial increase of the T,,, both in the presence and absence of spermine (7-8 "C), as previously shown (Lee et al., 1984; Pulleyblank et al., 1985; Strobe1 et al., 1988; Sun et al., 1989). The effect of acridine substitution was as dramatic, with a further increase of 8 "C, shifting the T,,, to a temperature range compatible with phys- iological conditions (Sun et al., 1989).

Effect of Acridine on Triplex Stability as Assessed by Gel Shift Assays-Triple-stranded structures can be detected by gel shift assays performed at neutral pH (Cooney et al., 1988). These conditions are relatively stringent since the triplex needs to remain stable throughout the duration of the migra- tion. In these conditions (Fig. 3), a significant difference was observed between the methylated (15-methylcytosine TH) and the fully modified (15-methylcytosine acridine TH) oli- gonucleotides. A small fraction of the double-stranded target shifted to a position assigned to the triple helix in the presence of 20 p~ of methylcytosine-substituted oligonucleotide, and the fraction of the triplex became significant only with doses of methylcytosine-substituted oligonucleotide as high as 100 pM. In contrast, the acridine-derivatized, methylcytosine-sub- stituted, oligonucleotide shifted a large majority of the duplex

A

C h I BnN I

B

OLIGONUCLEOTIDES USED

15CTH "T CTCCCT CTCC Till ' 15M.CTH T CTCCCT CTCC T I T

. . . . . . . I 5 M.Crn2 m T ;T%p ar;; l"r . ... . .. 15 M . C C TH T CTCCCT CTCC llTl m,Ac

11 M.C.ACEO#WOI T CTCC- -TCC . l T m 5 A c . L. ..

FIG. 1. 11-2 receptor a chain promoter ( A ) and oligonucleo- tides used in this study (23). A, the NF &-binding site is indicated by a shaded box. Restriction enzyme recognition sites are marked by thick broken lines, and cutting sites are indicated by arrows; the double-stranded sequence corresponds to the 47-mer used as a target in this study; the triple helix-forming region is boxed, together with the third strand oligonucleotide. B, C, cytosine; MeC, methylcytosine; Ac, acridine; m, methyl; !! is 5-methylcytosine; m2, mutant 2.

J 3 8 0,24 I2

Temp. ("C) FIG. 2. Melting curve of the triplex formed by 15 cytosine

TH ( I ) , 15-methylcytosine TH (Z), and 15-methylcytosine acridine TH (3) oligonucleotide in the presence of 0.4 mM spermine. The absorbance of the 25-bp double helix was subtracted from that of the 1:l mixture with the 15-mer oligonucleotides and normalized for initial OD value at 10 "C (for conditions, see "Mate- rials and Methods").

TABLE I Melting temperatures for the triplexes formed by the

three 15-mer oligonucleotides The target was a 25-bp synthetic fragment (position -264 to -240

of the IL-2 Ra regulatory sequence; see Fig. 1). For conditions see "Materials and Methods."

Oligonucleotide T,

0.4 mM spermine No spermine "C

15-Cytosine TH 26 20 15-Methylcytosine TH 33 28 15-Methvlcvtosine acridine TH 42 36

when used at the lowest concentration assayed (20 p ~ ) . A deletion mutant of this sequence (see Fig. l), the ll-methyl- cytosine acridine control oligonucleotide, did not have any effect at any concentration used (up to 100 p ~ ; data not

3392 Inhibition of NF KB via Oligonucleotide-directed Triplex Formation

"- IS MeC-ACTA 15 MeCTA

Oligo luM) 0 20 40 60 20 40 60 100 - -

t r i p l e h e l i x

DS +

ss 1 2 3 1 5 0 1 8 9

FIG. 3. Triplex stabilization by acridine, as assessed by band shift assays. The pyrimidine strand of the 47-mer target was end-labeled and incubated alone (lane I) , with an excess of the unlabeled purine strand (lane 2). The 47-bp double helix was then incubated with increasing concentrations of the 15-methylcytosine acridine (15 MeC-Ac) TH (lanes 3-5) or 15-methylcytosine (15 MeC)TH (lanes 6-9) oligonucleotide (20 pM, lanes 3 and 6; 40 pM, lanes 4 and 7; 60 phi, lanes 5 and 8; 100 p ~ , lane 9). ds, double- stranded; s ~ , single-stranded.

shown). These results indicate that acridine derivatization strongly increased the half-life of the triplex, allowing its detection under these stringent conditions.

Acridine Substitution Strongly Increases the Inhibition of Restriction Enzyme Cleauage-The stabilization of triplex structure by acridine substitution was further confirmed by studies on restriction enzyme inhibition. According to pre- vious reports (Francois et al., 1989b; Maher et al., 1989; Hanvey et al., 1990), triplex formation prevents restriction enzyme cutting, provided that the recognition sites of the third strand oligonucleotide and of the enzyme overlap over a few base pairs. In this assay, oligonucleotides are in com- petition with high affinity proteins for binding to the target. This provides an alternate means to estimate, through ID50 determination, the benefit of different modifications. Two restriction sites overlap the homopurine-homopyrimidine tar- get sequence, HinfI and MnlI, whose recognition and cutting sites are shown in Fig. 1. Using a 47-bp double-stranded DNA (Fig. l), end-labeled with 32P, as a target, we measured the cleavage products after preincubation with increasing concen- trations of the various oligonucleotides. HinfI was used to probe for triplex formation at pH 6.5, and MnlI was used at pH 7.5. Results of a typical experiment are shown in Fig. 4, and the data are summarized in Table 11. Results from the assays paralleled data obtained from direct T,,, measurement of the triple-stranded structures. The oligonucleotide effi- ciency at inhibiting the cutting by HinfI was increased by cytosine methylation; 5 times less of the 15-methylcytosine TH oligonucleotide was required to observe 50% inhibition. One order of magnitude in efficiency was gained by further addition of an acridine substituent, since the fully modified oligonucleotide (15-methylcytosine acridine TH) showed an IDs0 value of 0.5 p ~ . When the triplex was probed using MnlI inhibition at pH 7.5, further confirmation of the stabilizing effect of acridine was obtained, since acridine derivatization shifted the IDso value from 5 to 1 p ~ . Polyamines stabilized the triplex and, in the absence of spermine, 10-20-fold higher ID, values were obtained in both cases, but a similar advan- tage for the acridine-derivatized oligonucleotide was retained (results not shown). Taken together, these results confirm the data from gel shift assays and indicate that both 5-methyl- cytosine and acridine modifications are contributing to sta- bilization of the triplex when assayed under stringent condi- tions or when competing with high affinity binding proteins.

Acridine Substitution Does Not Impair Sequence Specificity for Triplex Formation-Since acridine is a nonspecific inter- calating agent, we next investigated sequence and target spec- ificity of triplex formation by the fully modified oligonucleo- tide. In order to assess triplex formation in conditions relevant to cell physiology, we used a footprinting assay on a plasmidic

1 5MeC-Ac TH 15MeC TH

I 1

I e

1 2 3 4 5 6 7 0 9 1 0 1 1 1 2 1 3 1 4

" . 0 0.5 1 1.5

oligonucleotide (I"

FIG. 4. Triplex stabilization as assessed by HinfI and MnlI restriction cleavage inhibition. A, the 47-mer double-stranded target was end-labeled, incubated with increasing concentrations of 15-methylcytosine acridine (15MeC-Ac)TH (lanes 1-7) or 15-meth- ylcytosine (15MeC)TH (lanes 8-14) in the presence of spermine (0.4 mM), cleaved with HinfI at pH 6.5, and analyzed on a denaturing 20% polyacrylamide gel. Oligonucleotide concentrations were 0 (lanes 1 and 8), 0.02 pM (lanes 2 and 9), 0.05 p~ (lanes 3 and lo), 0.1 pM (lanes 4 and 11), 0.3 p~ (lanes 5 and 12), 0.5 pM (lanes 6 and 13), and 1.5 p~ (lanes 7 and 14). B, HinfI cleavage efficiency as a function of oligonucleotide concentration (from A). 0, 15-methylcytosine ac- ridine TH; ., 15-methylcytosine TH.

TABLE I1 Concentrations and oligonucleotides giving 50% inhibition of cleavage

by restriction enzymes HinfI and MnlI See Fig. 1 for nomenclature. Conditions were as described under

"Materials and Methods." IDm

Hint7 Mnll P M

Oligonucleotide

15-Cytosine TH 25 ND" 15-Methylcytosine TH 5 5 15-Methylcytosine acridine 0.5 1

15-Methylcytosine-mutant >loo ND TH

2 TH ' ND, not determined.

target (Fig. 5A) harboring the native IL2-Fta promoter (from -352 to +110), including the Pur-Pyr triple helix target site and the NF KB regulatory protein-binding site. In order to assess any influence of the degree of superhelicity of the target, this assay was designed in such a way that it could be used with native plasmids (bacterial superhelicity), as well as with linearized fragments (Fig. 5A). Native or linearized plasmids were incubated with increasing doses of either the 15-methylcytosine acridine TH oligonucleotide or the deletion mutant of this sequence with identical chemical modifications

Inhibition of NF KB via Oligonucleotide-directed Triplex Formation 3393

B pH 6.8

ollponudeotlde 11 MeC-Ac 15 MeC-Ac 11 M e - A C 15Me-AC

uM G A T && 12550110.5 2.6 25 d coni m

" " .. . " d, r """ r

0 " " -

- . . . . . . . : ; I : : &

G A T C

- -

". 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20

C 19 2'

plalmld supercoiled llnearlred

ollponudsotlde L z L z '&? nn n n GATc

-> . . C." .. I." 9 ; i i ;tt ;:>

~ " 1.. - h&i- " " 5

-e- - - -

- ";ir , - . -., " -., --;

. I C " "" * - 4 . . "I. 0- - ~ -

; * && ;; ,j; - - c

r ? T tt= :- m

0 "

9 * ? t ) Z tt -5 :

1 2 3 4 5 0 7 8 9 10 12 14 16 18

=a

11 13 15 17 19

FIG. 5. Triplex formation as assessed by DNase I footprint- ing on plasmidic targets. A, description of the plasmid (1 ) and method (2); the position of the triplex target site, the primer, and the ClaI cutting site are indicated. B, analysis of DNase I cleavage products of a superhelical target a t pH 5.1 (lanes 4-9) or 6.8 (lanes

(11-methylcytosine acridine control; Fig. 1). Plasmids were then treated with DNase I, and the reaction products were revealed by primer extension using a "P-end-labeled primer located 60 bp downstream from the 15-methylcytosine acri- dine TH-binding site (Fig. 5A). These experiments showed a dose-dependent footprint on the Pur.Pyr target sequence (Fig. 5B) when preincubation of the oligonucleotide with the target was conducted at pH 5.1 or 6.8 (Fig. 5 B ) or pH 7.8 (Fig. 5C). The sequence protected, using this assay, extended beyond the Pur. Pyr sequence a few nucleotides on both sides of the triple-stranded structure. An enhancement of DNase I cleavage was observed a few nucleotides 5' of the triple helix- forming oligonucleotide. The footprint was not due to inter- ference of the third strand oligonucleotide with the primer extension process, since it disappeared when the incubation mixture was heated 10 min at 65 "C prior to DNase I treat- ment and subsequent extension (not shown). Furthermore, a t pH 7.8, the addition of a stabilizing agent, such as spermidine, during the preincubation of the oligonucleotide with the target sequence was required to observe the footprint (not shown). No footprints were observed with the control oligonucleotide (11-methylcytosine acridine control, Fig. 5, B and C), or an unrelated 16-mer sequence with identical chemical modifica- tions (not shown) used a t concentrations up to 50 PM. Super- helicity had little or no effect on triplex formation in these conditions, since at pH 7.8 in the presence of polyamines, triple helix formed at the same doses and in the same manner on supercoiled or linearized targets (Fig. 5C). Taken together, these data indicate that acridine substitution did not impair the exquisite sequence specificity of triple helix formation.

Inhibition of N F KB Transcription Factor Binding-The effect of triplex formation on nuclear protein binding to the overlapping NF KB site was assessed by gel shift assay, using a restriction fragment that includes the IL-2 Rcr upstream regulatory region from-352 to -153 as a probe. In the con- ditions used, two complexes were detected (Fig. 6A, lane 1 ), one of which corresponds to N F KB binding (marked by an arrow in the figure), since it was inhibited by an excess of a double-stranded oligonucleotide, including a dimer of NF KB site from the HIV LTR (lane 2) and not by an excess of an oligonucleotide harboring the c-fos dyad symmetry element (lane 3). Preincubation of the DNA fragment with the 15- methylcytosine acridine T H oligonucleotide strongly inhib- ited the appearance of this NF KB specific complex (lane 4 ) , whereas the 11-methylcytosine acridine control oligonucleo- tide did not have any effect (lane 5). These results indicate that triplex formation inhibited the binding of nuclear protein to the NF &-binding site. As a control, the HIV LTR, which

10-16) after preincubation with the 11-methylcytosine acridine (I1 MeC-Ac) control oligonucleotide (lanes 4, 5, IO, and 1 1 ) or the 15- methylcytosine acridine (15 MeC-Ac) T H oligonucleotide (lanes 6-9 and 12-16); doses of oligonucleotides were 0.5 p~ (lanes 6 and 12). 2.5 pM (lanes 7 and 13), 5 p~ (lanes 8 and 14), 25 p M (lanes 4, 9, 10, and 15) , 50 p~ (lanes 5, 11, and 16). Lanes 1-3 and 18-21 show dideoxy sequencing reactions performed in parallel in each experi- ment, using the same labeled primer. Lane 17, extension in the absence of DNase I treatment (extension control). C, analysis of DNase I cleavage products of a superhelical target ( l a n e 1-8) or a linearized target (lanes 9-15) after preincubation with the Il-meth- ylcytosine acridine control oligonucleotide (lanes 2, 3, 10, and 11) or the 15-methylcytosine acridine T H oligonucleotide (lanes 4-8 and 12-15) a t pH 7.8 in the presence of spermidine 4 mM. Oligonucleotide doses were 0.2 p~ (lanes 4 and 12), 0.8 pM (lanes 5, 6, and 13). 1 pM (lanes 2, 7, 10, and 14), 6 p~ (lanes 3, 8, 11, and 15). Lanes 16-19 show dideoxy sequencing reactions performed in parallel in each experiment, using the same labeled primer. Lanes I and 9, extension in the absence of oligonucleotides or DNase I treatment (extension control).

3394 Inhibition of NF KB via Oligonuck A dsHlV - + e - -

dsDSE - - + - *

15MeCAcTH - - * + - 11 MeC-AcCont - - - - +

rotide-directed Triplex Formation

reporter

tax - + + - + + - + +

IL-2 R alpha CAT HN -LTR CAT I 1-

15MeCAcTH 0 0 5 0 0 10 0 0 10 OJM 1

1 2 3 4 5

dsHlV dsDSE

- + " " + -

15 MeCAcTH - - - +

B

4

1 2 3 4 5 6 7 8 9

FIG. 7. Inhibition of tax-induced transcriptional activation of the IL-2 Ra CAT constructs in HSB2 cells. IL-2 Rcu CAT (lanes 1-6) or HIV LTR CAT (lanes 7-9) reporter constructs were preincubated with 15-methylcytosine acridine TH oligonucleotide when indicated (lane 3,5 PM; lanes 6 and 9,lO PM) and co-transfected into HSB2 cells together with a tax expression vector (lanes 2, 3, 5, 6, 8, and 9) or a control vector without tax (lanes 1, 4, and 7). Cell lysates were assayed for CAT activity 24 h later. These data are representative of five independent experiments.

1 2 3 4

FIG. 6. Effect of the oligonucleotides on NF KB binding. A, an end-labeled fragment including from -349 to -150 of the IL-2 Rcu promoter was preincubated alone (lanes 1 3 ) , with 15-methylcytosine acridine (15 MeC-Ac) TH (lane 4 ) or 11-methylcytosine acridine control (11 MeC-Ac Cont) (lane 5) oligonucleotides (20 J I M ) and further incubated with nuclear proteins in the absence of double- stranded competitors (lanes 1, 4, and 5) or in the presence of a 50- fold molar excess of a double-stranded oligonucleotide including the c-fos dyad symmetry element (dsDSE) (lane 3) or the HIV LTR tandem of NF KB sites (ds HZV) (lane 2). Incubation mixtures were analyzed on a nondenaturing acrylamide gel. The solid arrow indi- cates the KB-like complex. B, an end-labeled fragment including from -160 to +10 of the HIV LTR was preincubated alone (lanes 1- 3) or with the 15-methylcytosine acridine oligonucleotide (20 PM; lane 4 ) and further incubated with nuclear extracts in the absence of double-stranded competitor (lanes 1 and 4 ) , in the presence of a 50- fold molar excess of a double-stranded oligonucleotide including the c-fos dyad symmetry element (lane 3) or including the HIV LTR tandem of NF KB sites (lane 2).

contains two NF &-binding sites, was used as a target. The 15-methylcytosine acridine TH oligonucleotide did not have any effect on NF KB binding in this sequence environment (Fig. 6B, lane 3), indicating that the 15-methylcytosine acri- dine oligonucleotide did not trap the NF KB protein and did not have any artifactual effect. Taken together, these results indicate that triplex formation prevents binding of NF KB to the IL-2 Ra sequence and that this inhibition requires the presence of the overlapping homopurine-homopyrimidine tar- get sequence.

Inhibition of NF KB-dependent Transcription in Live Cells- We next investigated whether the acridine-derivatized oligo- nucleotide had any effect on the NF &-dependent transcrip- tional response to the human T cell leukemia virus-I transac- tivator tax in live cells. Reporter constructs were preincubated with the oligonucleotide and transfected, together with a tax expression vector, into human tumor T cells (HSB2). As shown in Fig. 7, a dose-dependent inhibition of tax-induced CAT activity was observed with 15-methylcytosine acridine T H on the IL-2 Ra CAT construct, whereas, used at the same

doses, it did not have any effect on the HIV-LTR CAT construct. Both constructs responded to tax induction, indi- cating that the inhibitory effect of the oligonucleotide was not due to trapping of tax or of any activating factor, such as NF KB. A deletion mutant lacking both the NF KB and the TH target site (including from -177 to +lo9 of the IL-2 Ra promoter) and therefore responding only minimally to tax was not affected by the 15-methylcytosine acridine TH oli- gonucleotide (data not shown). This last result eliminates any antisense effect of the 15-methylcytosine acridine TH oligo- nucleotide at the translational level (it has to be noted that we did not detect any complementary sequence on the CAT transcript, making an antisense effect very unlikely). These cellular experiments parallel the in vitro inhibition of NF KB binding by the 15-methylcytosine acridine TH oligonucleotide (described above) and demonstrate that NF KB inhibition by the oligonucleotide leads to the repression of an NF KB- dependent enhancement of transcription in live cells.

DISCUSSION

Oligonucleotide-directed triple helix formation is envi- sioned as a possible tool for artificial gene modulation through the inhibition of transcription factor binding. However, the use of such a strategy in vivo requires an adaptation to natural systems. The IL-2 Ra target sequence for triple helix forma- tion is only 15 nucleotides in length. Such a length should be sufficient for target specificity in the genome. However, the size of this target site raised the problem of triplex stability, given the high stringency of physiological conditions and, in particular, taking into account that the triple strand-forming oligonucleotide must compete with high affinity proteins binding to overlapping sites. Therefore, the key factor for success in these conditions is likely to be the design of efficient chemical modifications of the third strand oligonucleotide. Our data suggest that the simultaneous attachment of an intercalating agent and replacement of cytosines by 5-meth- ylcytosines confer triplex stability under physiological condi- tions, even in the case of a relatively short target sequence.

One of the main potential advantages of oligonucleotide- directed triplex formation for gene inhibition is the theoretical target specificity that could be achieved, since the target is not the more or less ubiquitous enhancer sequence itself, but an overlapping, gene-specific sequence. Here, we show that acridine derivatization did not impair sequence and target specificity of the third strand oligonucleotide. Whereas the 15-methylcytosine acridine TH oligonucleotide was able to

Inhibition of NF KB via Oligonuck

direct triplex formation, we were not able to detect any binding of an identically modified deletion mutant of this sequence to the IL-2 Ra target. Furthermore, the cognate TH oligonucleotide did not bind to the related HIV LTR sequence, as assessed by footprinting experiments (not shown). Like- wise, the fully modified TH oligonucleotide inhibited the binding of NF KB to its cognate sequence on the IL-2 Ra target and not on an HIV LTR. Finally, the fully modified oligonucleotide inhibited the tax-induced NF KB-dependent transcriptional activation of IL-2 R a CAT constructs in live cells. This did not occur through NF KB or tax alteration, since inhibition was not observed with an HIV LTR CAT construct (Fig. 7) or a Fos-CAT construct (not shown), whereas both promoters respond to tax transactivation. Our results therefore suggest that such a modified oligonucleotide behaves as a transcriptional repressor through triple helix formation with a regulatory sequence and could be used as a tool to inhibit the IL-2 Ra chain gene in uiuo.

Acknowledgments-We thank M. C. Dokhelar and S. Chouaib for helpful discussions, G. Trinquenaux for oligonucleotide synthesis, and A. PBle for expert secretarial assistance.

REFERENCES Birg, F., Praseuth, D., Zerial, Z., Thuong, N. T., Asseline, U., Le

Doan, T., and Helene, C. (1990) Nucleic Acids Res. 18, 2901-2908 Bohnlein, E., Lowenthal, J. W., Siekevitz, M., Ballard, D., Franza, B.

R., and Greene, W. C. (1988) Cell 53,827-836 Bohnlein, E., Ballard, D. W., Bogerd, H., Peffer, N. J., Lowenthal, J.

W., and Greene, W. C. (1989) J. Biol. Chem. 264,8475-8478 Boles, T. C., and Hogan, M. E. (1987) Biochemistry 26,367-376 Cann, A. J., Koyanagi, Y., and Chen, I. S. Y. (1988) Oncogene 3,123-

Cheng, H.-L., Blattner, F. R., Fitzmaurice, L., Mushinski, J. F., and

Cooney, M., Czernuszewicz, G., Postel, E. H., Flint, S. J., and Hogan,

Crabtree, G. R., and Kant, J. A. (1982) Cell 31 , 159-166 Cross, S. L., Halden, N. F., Lenardo, M. J., and Leonard, W. J. (1989)

Science 244,466-469 Dignam, J. D., Lebowitz, R. M., and Roeder, R. G. (1983) Nucleic

Acids Res. 11, 1474-1486 Dukovich, M., Wano, Y., Thuy, L. B., Katz, P., Cullen, B., Kehrl, J.,

and Greene, W. C. (1987) Nature 327,518-522 Evans, T., and Efstratiadis, A. (1986) J. Bwl. Chem. 261, 14771-

14780 Felber, B. K., Paskalis, H., Kleinman-Ewing, C., Wong-Staal, F., and

Pavlakis, G. N. (1985) Science 229 , 675-679 Francois, J. C., Saison-Behmoaras, T., and Helene, C. (1988) Nucleic

Acids Res. 16 , 11431-11440 Francois, J. C., Saison-Behmoaras, T., Barbier, C., Chassignol, M.,

Thuong, N. T., and Helene, C. (1989a) Proc. Natl. Acad. Sci. U. S.

Francois, J. C., Saison-Behmoaras, T., Thuong, N. T., and Helen, C. (1989b) Biochemistry 28 , 9617-9619

Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2 , 1044-1051

Greene, W. C., Robb, R. J., Svetlik, P. B., Rusk, C. M., Depper, J. M., and Leonard, W. J. (1985) J. Exp. Med. 162 , 363-368

Hanvey, J. C., Shimizu, M., and Wells, R. D. (1988) Proc. Natl. Acad. Sci. U. S. A. 86,6292-6296

Hanvey, J. C., Shimizu, M., and Wells, R. D. (1990) Nucleic Acids Res. 18, 157-161

Harel-Bellan, A., Korner, M., Brini, A. T., Ferris, D., and Farrar, W. L. (1989) Biochem. Biophys. Res. Commun. 162 , 238-243

Hatakeyama, M., Tsudo, M., Minamoto, S., Kono, T., Doi, T., Miyata, T., Miyasaka, M., and Tanigushi, T. (1989) Science 2 4 4 , 551-556

128

Tucker, P. W. (1982) Nature 296,410-415

M.E. (1988) Science 241, 456-459

A. 86,9702-9706

?otide-directed Triplex Formation 3395 HBlcine, C., and ToulmB, J.-J. (1990) Biochim. Biophys. Acta 1049,

Htun, H., and Dahlberg, J. E. (1988) Science 241, 1791-1796 Htun, H., Lund, E., and Dahlberg, J. E. (1984) Proc. Natl. Acad. Sci.

Htun, H., Lund, E., Westin, G., Pettersson, U., and Dahlberg, J. E.

Inoue, J., Seiki, M., Tanigichi, T., Tsuru, S., and Yoshida, M. (1986)

Johnston, B. H. (1988) Science 241 , 1800-1804 Kedes, L. (1979) Annu. Rev. Biochem. 48,837-870 Kehrl, J. H., Dukovich, M., Whalen, G., Katz, P., Fauci, A. S., and

Greene, W. C. (1988) J. Clin. Inuest. 8 1 , 200-205 Kohwi, Y., and Kohwi-Shigematsu, T. (1988) Proc. Natl. Acad. Sci.

Kohwi-Shigematsu, T., and Kohwi, Y. (1985) Cell 43 , 199-206 Le Doan, T., Perrouault, L., Praseuth, D., Habboub, N., Decout, J.,

Thuong, N. T., Lhomme, J., and Hblcine, C. (1987) Nucleic Acid Res. 19 , 7749-7760

Lee, J. S., Woodsworth, M. L., Latimer, L. J. P., and Morgan, A. R. (1984) Nucleic Acids Res. 12,6603-6614

Leonard, W. J., Depper, J. M., Crabtree, G. R., Rudikoff, S., Pum- phrey, J., Robb, R. J., Kronke, M., Svetlik, P. B., Peffer, N. J., Waldmann, T. A., and Greene, W. C. (1984) Nature 311,626-631

Leung, K., and Nabel, G. J. (1988) Nature 333 , 776-778 Lowenthal, J. W., Bohnlein, E., Ballard, D. W., and Greene, W. C.

(1988) Proc. Natl. Acad. Sci. U. S. A. 85,4468-4472 Lowenthal, J. W., Ballard, D. W., Bohnlein, E., and Greene, W. C.

(1989) Proc. Natl. Acad. Sci. U. S. A. 8 6 , 2331-2335 Lyamichev, V. I., Mirkin, S. M., Frank-Kamenetskii, M. D., and

Cantor, C. R. (1988) Nucleic Acids Res. 16, 2165-2178 Maher, L. J., 111, Wold, B., and Dervan, P. B. (1989) Science 245 ,

Maruyama, M., Shibuya, H., Harada, H., Hatakeyama, M., Seiki, M., Fujita, T., Inoue, J., Yoshida, M., and Taniguchi, T. (1989) Cell

99-125

U. S. A. 8 1 , 7288-7292

(1985) EMBO J. 4, 1839-1845

EMBO J. 5,2883-2888

U. S. A. 85,3781-3785

725-730

48,343-350 McIntyre, P., and Stark, G. (1988) Anal. Biochem. 174 , 209-214 Moser, H. E., and Dervan, P.B. (1987) Science 238 , 645-650 Perrouault, L., Asseline, U., Rivalle, C., Thuong, N. T., Bisagni, E.,

Giovannangeli, C., Le Doan, T., and Helene, C. (1990) Nature 344 ,

Pomerantz, J . L., Mauxion, F., Yoshida, M., Greene, W. C., and Sen, R. (1989) J. Zmmunol. 143,4275-4281

Praseuth, D., Perrouault, L., Le Doan, T., Chassignol, M., Thuong, N., and Helene, C. (1988) Proc. Natl. Acad. Sci. U. S. A. 8 5 , 1349- 1353

Pullevblank. D. E.. Haniford. D. B.. and Moraan. A. R. (1985) Cell

358-360

42; 271-280 '

- ,

RaiaeoDal. P.. and Feizon. J. (1989) Nature 339.637-640 Richards, J. E., Gilliam, A. C., Shen, A., Tucker, P . W., and Blattner,

Ruben, S., Poteat, H., Tan, T.-H., Kawakami, K., Roeder, R., Has-

Sharon, M., Klausner, R. D., Cullen, B. R., Chizzonite, R., and

Strobel, S. A., and Dervan, P. B. (1990) Science 249 , 73-75 Strobel, S. A,, Moser, H. E., and Dervan, P. B. (1988) J. Am. Chem.

Sun, J. S., Francois, J. C., Montenay-Garestier, T., Saison-Beh-

Natl. Acad. Sci. U.S.A. 86,9198-9202 moaras, T., Roig, V., Thuong, N. T., and Helene, C. (1989) Proc.

Thuong, N. T., and Chassignol, M. (1988) Tetrahedron Lett. 29,

Toledano, M. B., Roman, D. G., Halden, N. F., Lin, B. B., and Leonard, W. J. (1990) Proc. Natl. Acad. Sci. U.S. A. 8 7 , 1830-1834

Trouche, D., Robin, P., Sassone-Corsi, P., Farrar, W. L., and Harel- Bellan, A. (1991) Blood 77 , 55-63

Tsudo, M., Kozak, R. W., Goldman, C. K., and Waldmann, T. A. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 9694-9698

Vlassov, V. V., Gaidamakov, S. A., Zarytova, V. F., Knorre, D. G., Levina, A. S., Nikonova, A. A., Podust, L. M., and Fedorova, 0. S. (1988) Gene (Amst.) 72, 313-322

Voloshin, 0. N., Mirkin, S. M., Lyamichev, V. I., Belotserkovskii, P., and Frank-Kamenetskii, M. D. (1988) Nature 333,475-476

. ,

F. R. (1983) Nature 306,483-487

eltine, W., and Rosen, C. A. (1988) Science 241,89-92

Leonard, W. J. (1986) Science 234,859-863

SOC. 110,7927-7929

5905-5908