Molecular interactions between nuclear factor kB (NF-kB) transcription factors and a PNA–DNA chimera mimicking NF-kB binding sites Alessandra Romanelli 1 , Carlo Pedone 1 , Michele Saviano 1 , Nicoletta Bianchi 2 , Monica Borgatti 2 , Carlo Mischiati 2 and Roberto Gambari 2,3 1 Biocrystallography Research Center, CNR, Napoli; 2 Department of Biochemistry and Molecular Biology, and 3 Biotechnology Centre, Ferrara University, Ferrara, Italy The decoy approach against nuclear factor kB (NF-kB) is a useful tool to alter NF-kB dependent gene expression using synthetic oligonucleotides (ODNs) carrying NF-kB specific cis-elements. Unfortunately, ODNs are not stable and need to be extensively modified to be used in vivo or ex vivo.We have previously evaluated the possible use of peptide nucleic acids (PNAs) as decoy molecules. The backbone of PNAs is composed of N-(2-aminoethyl)glycine units, rendering these molecules resistant to both nucleases and proteases. We found that the binding of NF-kB transcription factors to PNAs was either very low (binding to PNA–PNA hybrids) or exhibited low stability (binding to PNA–DNA hybrids). The main consideration of the present paper was to determine whether PNA–DNA chimeras mimicking NF-kB binding sites are capable of stable interactions with proteins belonging to the NF-kB family. Molecular modeling was employed for the design of PNA–DNA chimeras; prediction of molecular interactions between chimeras and NF-kB nuclear proteins were investigated by molecular dynamics simulations, and interactions between PNA–DNA chimeras and NF-kB proteins were studied by gel shifts. We found significant differences between the structure of duplex NF-kB PNA–DNA chimera and duplex NF-kBDNA– DNA. However, it was found that these differences do not prevent the duplex PNA–DNA chimera from binding to NF-kB transcription factors, being able to suppress the molecular interactions between HIV-1 LTR and p50, p52 and nuclear factors from B-lymphoid cells. Therefore, these results demonstrate that the designed NF-kBDNA–PNA chimeras could be used for a decoy approach in gene therapy. Keywords: peptide nucleic acids; PNA–DNA chimeras; AIDS; NF-kB; transcription factors. In vitro transfection of cis element decoys against nuclear factors leads to the alteration of gene expression, and was recently proposed as a novel molecular medicine tool for possible use in therapy of a variety of well-characterized disorders [1 –9]. Decoy molecules against HNF-1, RFX1, NFYB, E2F, CRE and Sp1 were found to alter specific functions in eukaryotic cells [7 –11]. One of the most effective decoy approaches so far described involves nuclear proteins belonging to the NF-kB superfamily. Decoy molecules against NF-kB inhibit the expression of NF-kB regulated genes (e.g. genes coding for MHC, IL2 receptor a, Igk, IL6, d opioid receptor, preprogalanin and adhesion molecule-1) [12–20]. More recently, dumbell DNA decoy elements against NF-kB were demonstrated to be active in inhibiting ex vivo transcription driven by the long-terminal repeat (LTR) of human immunodeficiency type-1 virus (HIV-1) [21]. A drawback of the decoy approach designed for the modulation of gene expression is the presence of intracellular DNases [1– 7]. Therefore, large amounts of DNA must be internalized by target cells in order to obtain biological responses leading to alteration of gene expression [2]. In contrast, modified oligonucleotides (either methyl- phosphonate or phosphorothioate) have been used by virtue of their resistance to DNase cleavage, but these molecules are highly toxic [22]. A further problem of the decoy approach is the recently reported nonspecific activity of these molecules. For example, dumbbell oligonucleotides to RFX1, in addition to blocking activation of RFX1 regulated genes, cause additional nonspecific effects most likely via an interaction with the general transcription machinery [2]. In a recent paper, we investigated the possible use of peptide nucleic acids (PNAs) [23–27] as alternative reagents in experiments aimed at the control of gene expression involving the decoy approach [28]. In PNAs, the pseudo- peptide backbone is composed of N-(2-aminoethyl)glycine units [23]. PNAs hybridize with high affinity to comple- mentary sequences of single-stranded RNA and DNA, forming Watson–Crick double helices [23,24] and are resistant to both nucleases and proteases [29]. We demon- strated that NF-kB p52 is able to bind to both NF-kBDNA– DNA and DNA– PNA hybrids mimicking the NF-kB target sites present in the HIV-1 LTR. Low binding of NF-kB p52 to PNA–PNA hybrids was found [28]. We have also reported a conformational study to explain these binding Correspondence to R. Gambari, Department of Biochemistry and Molecular Biology, Via L. Borsari n.46, 44100 Ferrara, Italy. Fax: 1 39 532 202723, Tel.: 1 39 532 291448, E-mail: gam@dns.unife.it (Received 25 May 2001, revised 19 September 2001, accepted 23 September 2001) Abbreviations: NF-kB, nuclear factor kB; Sp1, promoter-specific transcription factor Sp1; AIDS, acquired immunodeficiency syndrome; HIV-1, human immunodeficiency virus type 1; LTR, long-terminal repeat; PNA, peptide nucleic acids; PDP, PNA–DNA–PNA chimera; HATU, O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate; DIPEA, N,N 0 -diisopropylethylamine; MD, molecular dynamics; ODN, synthetic oligonucleotides; cvff, consistent valence force field. Eur. J. Biochem. 268, 6066–6075 (2001) q FEBS 2001 data using a molecular dynamics approach. These data have underlined that the loss of charged phosphate groups and the different shape of helices in PNA–DNA and PNA–PNA hybrids drastically reduce binding efficiency to NF-kB transcription factor [30,31]. In order to develop PNA-based molecules able to stably interact with transcription factors, in the present paper we investigated whether PNA–DNA chimeras mimicking the NF-kB binding sites are able to interact with both purified NF-kB p52 and p50, as well as nuclear factors from B-lymphoid cells. In should be noted that PNA–DNA chimeras are com- pounds of great interest [32– 34], as they are more water- soluble than PNAs [33] and are far more resistant to enzymatic degradation than oligonucleotides [32–34]. In addition, they were recently found to be suitable primers for DNA polymerases [32]. Finally, PNA–DNA chimeras generated with comple- mentary RNA hybrid molecules were recognized by RNase H [33]. However, PNA–RNA hybrids are not recognized by RNase H [33]. In spite of these very promising results, no data on the possible recognition of double stranded PNA–DNA chimeras by transcription factors are currently available in the literature. In the present paper, we designed, synthesized and tested complementary PNA–DNA–PNA (PDP) chimeras poten- tially able to interact with nuclear proteins belonging to the NF-kB family. With respect to the choice of the target sequence, we decided to perform experiments using the nonsymmetric NF-kB binding site of HIV-1 LTR in order to maximize solubility (PNAs extremely rich in GC should be avoided) and minimize the possibility to generate self- or inter-strand hybridization, possibly forming highly stable complexes [23–27]. In this respect, palindromic DNA sequences (for example the symmetric GGGGATTCCCCT NF-kB binding site of human p-selectin, human IL2Ra, mouse H2K, mouse MHC EA promoter regions) are not the first choice [28]. Molecular modeling was employed for the design of the NF-kB PNA–DNA–PNA chimeras, prediction of the molecular interactions between a double-stranded PNA– DNA–PNA chimera and nuclear proteins belonging to the NF-kB family was performed by energy minimization and molecular dynamics simulations, and interactions between PNA–DNA chimeras and NF-kB proteins were studied by electrophoretic mobility shift assays [28]. MATERIALS AND METHODS Synthetic oligonucleotides and peptide nucleic acids The synthetic oligonucleotides used in this study were purchased from Pharmacia (Uppsala, Sweden). HPLC- purified PNAs were purchased from ISOGEN Biosciences (Maarssen). Synthesis of NF-kB PNA– DNA chimeras Tetrabutylammonium N-[2-[(4-methoxytrityl)amino]ethyl] ethyl]-N-[thymin-1-yl-acetyl] glycinate, tetrabutyl ammo- nium N-[(N 6 -benzoyladenin-9-yl)acetyl]-N-[2-[(4-meth- oxytrityl) amino] ethyl]glycinate, tetrabutylammonium N-[(N 2 -isobutyrylguanin-9-yl)acetyl]-N-[2-[(4-methoxy- trityl) amino]ethyl] glycinate, tetrabutylammonium N-[(N 4 - benzoylcytosine-1-yl)acetyl]-N-[2-[(4-methoxytrityl)ami- no]ethyl]glycinate PNA monomers were synthesized in the laboratories of J H. Van Boom [35,36] (Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, the Netherlands); DNA monomers were obtained from Persep- tive Biosystems. Methanol (Rathburn, HPLC grade) was stored over molecular sieves (3 A ˚ ) and used without further purification. All the other solvents (Biosolve, synthesis grade DNA) were used as received. Automatized syntheses of the chimeras were performed on a Pharmacia Gene Assembler, using highly cross-linked polystyrene (loading 26–28 mmol : g 21 ) as the solid support on a 1-mmol scale. The support was functionalized with a Fmoc-glycine via a 4-hydroxymethylbenzoic acid linker. Assembly of the PNA parts was realized using 0.3 M solutions of the monomers in acetonitrile/dimethylformamide 1 : 1 (v/v) (containing 25% of dimethylsulfoxide in the case of pyrimidine build- ing blocks), 0.3 M N,N 0 -diisopropylethylamine (DIPEA) in acetonitrile/dimethylformamide (1 : 1, v/v) and 0.3 M O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate (HATU) in acetonitrile/dimethylform- amide (1 : 1, v/v). PNA monomers (15 equivalents per mmol of resin) were preactivated for 1 min by mixing with equal amounts of DIPEA and HATU, before coupling. The protocol for the PNA oligomer synthesis on a 1 mmol scale consisted of a cycle of the following steps. (a) Washing with 2.5 mL of acetonitrile/dimethylformamide (1 : 1, v/v); (b) coupling with the preactivated solutions of PNA, DIPEA and HATU for 15 min in acetonitrile/dimethylformamide (1 : 1, v/v); (c) washing with 2.5 mL of acetonitrile/dimethyl- formamide (1 : 1, v/v); (d) capping with Ac 2 O/2,6-lutidine/ N-methylimidazole/tetrahydrofuran (1 : 1 : 1 : 7, v/v/v/v), 2.0 mL; (e) washing with 2.5 mL of acetonitrile and then 3.5 mL of dichloromethane; (f) detritylation with a solution of trichloroacetic acid (2%) in dichloromethane for 3 min; (g) washing with 2.5 mL of dichloromethane and 5 mL of acetonitrile. DNA tract chain elongation was carried out using 2-cyanoethyl-phosphoramidite-2 0 -ribonucleoside building blocks (15 equivalents). Two successive couplings were used to assure a high yield when obtaining the PNA-3 0 - DNA junction via a phosphoramidate bond. 5-(O-nitrophe- nyl)-tetrazole was used as activator. Standard DNA capping, oxidation and detritylation cycles were used. Coupling yields were gauged spectrophotometrically at 254 nm by the absorption of the released trityl cation after each deprotec- tion step. Finally, a DNA elongation step was performed using a monomethoxytrityl protected 5 0 -amino-5 0 -deoxythy- midine (T) phosphoramidite as the linker between the DNA and PNA sections. The amidic bond between the 5 0 -amino- 5 0 -deoxythymydine phosphoramidite was realized using two successive coupling cycles for the first PNA unit. PNA chain synthesis was carried out following the above described procedure. The yield of each PNA coupling was in the range 95–99 %, and the DNA couplings were quantitative. After the last elongation step, the oligomers were cleaved from the solid support and deprotected by treatment with 1.5 mL methanolic ammonia at 50 8C for 16 h [34]. The samples were filtered and then purified by RP-HPLC on a LiChrosphere 100 RP-18 endcapped column (4 Â 250 mm) on a Jasco HPLC system. Gradient elution was performed at 40 8C, building up gradient starting with q FEBS 2001 Binding of NF-kB proteins to PNA–DNA chimeras (Eur. J. Biochem. 268) 6067 buffer A (50 mM triethylammonium acetate in water) and applying buffer B (50 m M triethylammonium acetate in acetonitrile/water, 75 : 25, v/v, with a flow rate of 1mL : min 21 . Chimera 1: HPLC purity 100%, t R ¼ 18 min (gradient 3–20% B in 25 min); chimera 2: HPLC purity 100%, t R ¼ 16 min (gradient 5–25% B in 25 min). HPLC- MS analysis was carried out on a Jasco LCMS system equipped with a LiChrosphere 100 RP-18 endcapped column (4 Â 250 mm) using a gradient of acetonitrile in 10 m M ammonium acetate buffer with mass detection on a Perkin Elmer Sciex API 165 equipped with an electrospray interface (ESI). Chimera 1: t R ¼ 7 min (gradient 5–20% acetonitrile in 20 min); ESI-MS: [M 1 4H] 41 ¼ 1438.2, [M 1 5H] 51 ¼ 1150.5, calculated. for C 193 H 245 N 90 O 95 P 13 5748.26. Chimera 2: t R ¼ 8 min (gradient 0–20% aceto- nitrile in 20 min); ESI-MS: [M 1 4H] 41 ¼ 1443.6, [M 1 5H] 51 ¼ 1154.9, calculated for C 194 H 247 N 86 O 99 P 13 5770.26. The chimera sequences were Gly-ccg-5 0 -TGGAAAGTCC CCA-3 0 -gcg-Ac (1) and Gly-cgc-5 0 -TGGGGACTTTCCA-3 0 - cgg-Ac (2). Molecular dynamics simulations All calculations and graphical analyses were run on a Silicon Graphics O2 R10000 workstation. The package INSIGHTDISCOVER (Biosym Technologies) was used to per- form energy minimization and molecular dynamics simu- lations (MD) in vacuo at 300 K, with the consistent valence force field (cvff), setting a pH 7 for all simulations. In all simulations, the Arg, Glu, Gln, His, Lys, Asp and Asn side chains carry a full charge, in agreement with the pH value. The starting structures used in structural analysis and simulations were those obtained from the Protein Data Bank (http://www.rcsb.org/pdb/). Computational conditions were chosen to avoid boundary effects [37]. The preparation of the starting models was performed in agreement with other MD studies on this class of com- pounds [38]. Using the solid state coordinates of the com- plexes NF-kB p50/p65 homodimer with DNA, the PNA duplex was generated by replacing the backbones of the DNA strands of the DNA–DNA duplex with the PNA backbone atoms. In all models, the coordinates of PNA backbone atoms were generated by geometrical calculations on their local topology and coordinates of their nearest connected atoms and literature structural data. The coordi- nates were then minimized, keeping the bases in a fixed position. Then, the restraints were removed and further energy minimization was performed. These resulting struc- tures were then used for subsequent MD simulations. The simulation was performed with a time step of 1.0 fs at 300 K and the system was equilibrated for 80 ps. After this first step, an additional 80 ps of simulation without rescaling were carried out, as energy conservation was observed and the average temperature remained essentially constant around the target values. Coordinates and velocities for the four simulations were dumped to a disk every 10 steps during the last 80 ps of the simulation. Circular dichroism spectra CD spectra were recorded at 20 8C on a Jasco model J-700 spectropolarimeter. The data were collected at 0.2 nm intervals, with a 20-nm : min 21 scan rate, 1 nm band with a 16-s response, from 400 to 200 nm. Five scans were performed for each sample, the CD spectra were obtained as an average of the scans. The solutions were prepared with concentration of 3.55 Â 10 27 M for PDP–PDP, 1.52 Â 10 26 M for DNA–DNA in 10 mM phosphate buffer at pH 7. CD spectra are reported in molar ellipticity vs wavelength. Single strand concentrations were determined by UV. Double strands were annealed by warming up at 80 8C and cooling down at 4 8C before recording CD spectra. Melting experiments were also performed on the duplexes in 10 m M phosphate buffer at pH 7. Electrophoretic mobility shift assay The electrophoretic mobility shift assay (EMSA) [39] was performed by using the double-stranded synthetic oligo- nucleotides mimicking the NF-kB (the nucleotide sequences have been reported above and are shown in Fig. 1). The synthetic oligonucleotides were 5 0 end-labelled using [g- 32 P]ATP and T4 polynucleotide kinase (MBI Fermentas). Binding reactions were set up as described elsewhere [39] in a total volume of 25 mL containing buffer TF plus 5% glycerol, 1 m M dithiothreitol, 10 ng of human NF-kB p52 protein (Promega Corporation, Madison, WI) and 0.25 ng of 32 P-labelled oligonucleotides. When 2 mg of crude nuclear extracts isolated from human cell lines were used instead of purified NF-kB p50 and p52 factors, the binding reaction was carried out in the presence of 1 mg of the nonspecific competitor poly(dI-dC) : poly(dI-dC) [40]. After 20 min binding at room temperature, the samples were electro- phoresed at constant voltage (200 V) under low ionic strength conditions (0.25 Â Tris/borate/EDTA buffer; 22 m M Tris/borate, 0.4 mM EDTA) on 6% polyacrylamide gels. Gels were dried and subjected to standard autoradiographic procedures [39]. In competition experiments, the competitor Fig. 1. Structure of the HIV-1 genome, location of NF-kB and Sp1 binding sites, sequences of the ODNs, PNAs and PNA–DNA–PNA (PDP) chimeras used. 6068 A. Romanelli et al. (Eur. J. Biochem. 268) q FEBS 2001 molecules carrying HIV-1 NF-kB binding sites (DNA– DNA, PDP–PDP and PNA–PNA) were preincubated for 20 min with purified NF-kB p52 protein, purified NF-kB p50 factor or nuclear extracts, before the addition of labelled target DNA. Nuclear extracts were prepared according to Dignam et al. [40]. The nucleotide sequences of competitor double stranded target DNAs used as controls were 5 0 - TAATATGT AAAAACATT-3 0 (sense strand, NF-IL2A), 5 0 - CACTTGAT AACAGAAAGTGATAACTCT-3 0 (sense strand, GATA-1) and 5 0 -CATGTTATGCATATTCCTGTA- AGTG-3 0 (sense strand, STAT-1). Stability of decoy molecules The stability of decoy molecules was evaluated after incu- bation of DNA and PNA–DNA–PNA based decoys with 3 0 !5 0 exonuclease III, 5 0 !3 0 lambda exonuclease and DNase I. ExoIII and lambda exonuclease were purchased from MBI Fermentas and DNase I from Promega Corporation, Madison, WI, USA. After incubation with increasing amounts of the enzymes (for 10 min in the case of ExoIII, for 30 min in the case of lambda exonuclease and DNase I), the decoy molecules were layered on the top of a 2% agarose gel and detected by ethidium bromide staining. Disappearence of the decoy molecule was considered as an evidence of degradation by the employed enzymes. Results were presented as percentage of recovery with respect to control untreated reaction mixtures. RESULTS Design of synthetic oligonucleotides, PNAs and PNA–DNA chimeras The nucleotide sequence corresponding to a single asym- metric NF-kB binding site of the HIV-1 LTR was chosen in order to maximize solubility of synthetized PNAs and PNA–DNA chimeras. In addition, unlike symmetric NF-kB binding sites, possible problems related to self and/or inter- strand hybridization are expected to be minimal in the case of asymmetric NF-kB binding sites. For these reasons, the experiments were performed with synthetic molecules carrying the HIV-1 LTR asymmetric NF-kB binding site in both sense and antisense orientations (see Fig. 1). The rational design of the NF-kB DNA–PNA chimera was carried out, taking into account previous computational analysis reported by us [30,31] and the solid state NF-kB p50/p65 complex structure [41]. To preserve all protein– DNA contacts the chimera was designed keeping the nonsymmetric HIV-1 NF-kB binding site (5 0 -GGGGACT TTCC-3 0 ) and linking to this DNA core a T base in the 5 0 position for synthetic reasons and an A at the 3 0 end for the Watson–Crick base pair interactions. At both the 5 0 and 3 0 ends we have added three PNA monomers to ensure a highly stable duplex. Computational analysis The MD simulation was performed on the complex between nuclear factors and the double stranded PNA–DNA–PNA chimera mimicking the HIV-1 LTR NF-kB binding sites, in order to investigate the molecular interactions between nuclear factors and the PDP– PDP hybrid molecule. The comparison between the complex structure between NF-kB p50/p65 bound to DNA–DNA, previously reported [41], and the complex structure between NF-kB p50/p65 bound to PDP–PDP, as obtained by MD simulation, demonstrates that in both cases the central DNA duplex cores show a comparable conformation (Fig. 2). In fact, the rmsd of all atoms DNA of 5 0 -GGGACTTTC-3 0 /5 0 -GAAAGTCC-3 0 duplex from the canonical B-DNA is 2.2 A ˚ . This fragment is slightly unwound with an overall twist of 11.2 bp per turn (canonic DNA in B form has 10.0 bp per turn). On the contrary, a large distortion of DNA due to PNA duplex is present in the first (T) and last (CA) two DNA bp. In fact the PNA duplexes have a conformation similar to that observed in the solid state structure of PNA–PNA duplex [42]. The PNA base pairs present a wide helix with an average calcu- lated pitch of < 14 base pairs. The PNA helix conformation could be described as comparable to a P-form duplex, and this P-form would appear to be the natural conformation for PNA [31]. The DNA bp near the PNA duplex in 5 0 and 3 0 show a distortion from the canonic DNA B-form. In par- ticular, the sugar ring dihedral angles have values com- parable to those observed in B-form DNA, but an average overall twist of 12.5 bp per turn. These results underline that PNA prefers a helical structure that is different from that of DNA–DNA helices, indicating that PNA seems to have clear conformational preferences that are the driving force that leads to the modification of the less flexible DNA backbone during the formation of the duplex. It is worth noting that the PNAs in Fig. 2. Average structure of NF-kB PDP–PDP duplex as obtained from MD simulation in vacuo at 300 K of NF-kB p50/p65 heterodimer-PDP–PDP complex. q FEBS 2001 Binding of NF-kB proteins to PNA–DNA chimeras (Eur. J. Biochem. 268) 6069 chimeric strands are able to modify the DNA conformation of 2 or 3 bp starting from the PNA–DNA junction without influencing the conformation of the remaining base pairs. Computational analysis of the interaction of p50 NF-kB with the PDP-PDP hybrid molecule In the complexes between NF-kBPDP–PDPhybrid molecule, the structural analysis reveals that the chimera duplex has all base–specific interactions. In addition, the system presents similar energy stabilization with respect to the DNA transcription factor complex. Finally, the compari- son of the relative orientation of the two subunits, both in the solid-state structures and in the MD average models, reveals that the NF-kB residues of both subunits are in optimal positions to bind to PNA. In particular, for p50/p65 hetero- dimer, the base–specific interactions mediated by Arg54, Arg56, Glu60 and His64 of the p50 subunit and mediated by Arg34, Arg35 and Arg187 of the p65 subunit are present, and are comparable to the NF-kB DNA– DNA hybrid molecule [41]. CD analysis and melting experiments CD spectra of the duplex PDP–PDP chimera were recorded and compared to the full DNA duplex sequence (Fig. 3). All the comparisons were made on spectra normalized with respect to the concentration. The analysis of the chimera spectrum as compared to the DNA double-strand spectrum [43] suggested that PDP –PDP tends to adopt a RNA A like conformation. A-form RNA usually has a maximum close to 260 nm, a minimum close to 210 nm and a small negative CD between 290 and 300 nm [43]. One maximum was found at 267 nm in the chimera, the value observed in the DNA analogue is 274 nm. Also, a shallow negative band was at 294 nm. According to the conformational results from the chimera duplex models, the amount of RNA A form in the chimera, calculated considering a shift from the DNA B form to the RNA A form, is < 50%. The intensity of the bands was higher in the full DNA duplex than in the chimera. This can be attributed to a bigger contribution of stacking in the full DNA duplex than in the chimera duplex. It could be speculated that adding PNA units at both termini of the DNA strands causes a distortion in the double strand. The rigid junction between the PNA and the DNA at the 5 0 end of the DNA might play an important role in the conformation of the chimeric oligomer. UV melting experiments were carried out on the same duplex samples. A comparison between the obtained data revealed that duplexes have similar melting temperatures within the experimental error range, the chimeric duplex being the less stable. The melting temperatures, calculated assuming a two state process, were 53 ^ 1 8C for PDP– PDP and 55 ^ 1 8C for DNA duplex (DNA). The double stranded PDP–PDP chimera inhibits the interactions between NF-kB transcription factors and target DNA–DNA molecules When 10 ng of human NF-kB p52 and p50 proteins were incubated for 20 min in the presence of the cold double- stranded PDP–PDP chimera, it was found that the 32 P-labelled NF-kB DNA–DNA probe was not efficiently recognized by the NF-kB proteins (Fig. 4A). The results obtained strongly suggest that, under these experimental conditions, the double stranded PDP–PDP chimera efficiently binds to p52 and p50 NF-kB transcription factors. In addition, Fig. 4A shows that, unlike NF-kB PDP–PDP molecules, NF-kB PNA–PNA hybrids do not affect the binding of p50 and p52 NF-kB transcription factors to 32 P-labelled NF-kB DNA–DNA target. As expected [28], competition performed with GATA-1 and NF-IL2A control oligonucleotides was also uneffective in inhibiting interactions of NF-kB factors with the target DNA. Figure 4B shows a detailed study of the relationship between the amount of competitor added and the inhibitory effects observed. In this experiment, increasing amounts of NF-kB DNA–DNA or PDP–PDP molecules were incu- bated for 20 min in the presence of 10 ng of human NF-kB p50 (Fig. 4B) and p52 (Fig. 4C) proteins; after this binding period, a further 20 min incubation was performed in the presence of the 32 P-labelled NF-kB DNA–DNA probe and the samples were analysed by electrophoresis on native 6% polyacrylamide gels. These results demonstrate that the NF-kB PDP –PDP hybrid does act as a competitive inhibi- tor, despite having a lower efficiency than NF-kBDNA– DNA hybrid in binding to purified p52 and p50 NF-kB proteins. The double stranded PDP–PDP chimera mimicking the HIV-1 NF-kB binding sites inhibits the interactions between crude nuclear extracts and target NF-kB DNA–DNA molecules In order to determine the activity of the double stranded PDP–PDP chimera carrying NF-kB binding sites on a more complex protein context, we repeated the experiments reported in Fig. 4A by using, instead of purified NF-kB p50 and p52 proteins, crude nuclear extracts from B-lymphoid Raij cells. In the experiment shown in Fig. 5, the double stranded NF-kB PDP– PDP chimera was preincubated with nuclear extracts from Raji cells, and processed as described for the experiments shown in Fig. 4A. The obtained data confirm that the double stranded PDP– PDP chimera inhibits the binding of protein factors to the 32 P-end-labelled NF-kB DNA–DNA target molecule. Control oligonucleotides Fig. 3. CD spectra of NF-kB PDP–PDP (- - -) and of DNA–DNA (– – –) duplexes. 6070 A. Romanelli et al. (Eur. J. Biochem. 268) q FEBS 2001 (NF-IL2A and GATA-1) were found to be inactive. In agreement with data reported elsewhere by our research group [28] the corresponding PNA–PNA hybrid molecule was unable to inhibit the interactions between 32 P-labelled DNA–DNA target and p52 or p50 NF-kB factors. The double stranded PDP–PDP chimera mimicking the HIV-1 NF-kB binding sites does not inhibit the interactions between NFIL2, GATA-1 and STAT-1 transcription factors to the relative target DNA–DNA sequences The experiment reported in Fig. 6 was performed using 32 P-end-labelled NF-IL2A, STAT-1 and GATA-1 DNA– DNA target molecules and nuclear factors isolated from Raji and K562 cell lines. The results obtained firmly establish that the effects of PDP– PDP chimera are sequence-specific. In fact, while NF-IL2A, STAT-1 and GATA-1 cold oligo- mers suppress the binding of nuclear factors to the relative 32 P-end-labelled DNA–DNA target molecules, no inhibitory activity was determined by addition of double stranded PDP–PDP chimera mimicking the HIV-1 NF-kB binding sites. Differential effects of the HIV-1 NF-kB PDP–PDP chimera on binding of NF-kB p52 and p50 to NF-kB binding sites of HIV-1 and IgK gene The experiment reported in Fig. 7 demonstrates that DNA– DNA and double stranded PDP –PDP chimeras mimicking the HIV-1 NF-kB binding sites, while effective inhibitors of binding of NF-kB p52 and p50 to HIV-1 LTR sequences, do not efficiently inhibit the binding of the same factors to the palindromic GGGGATTCCCCT NF-kB IgK DNA sequences. This result is of some relevance and is probably due to the well-known differential affinity of NF-kB p52 and p50 to HIV-1 or Igk NF-kB binding sites [44]. The data shown in Fig. 5 demonstrate that the NF-kB PDP–PDP duplex chimera exhibits biological effects very similar to those of NF-kBDNA–DNA. Stability of the decoy molecules based on PNA–DNA–PNA chimeras The stability of the decoy molecules based on PNA– DNA– PNA chimeras mimicking the NF-kB binding sites was evaluated after incubation with 3 0 !5 0 exonucleases, 5 0 !3 0 Fig. 4. Hybrid effects. (A) Effects of DNA–DNA, PNA–PNA and PDP–PDP hybrids, carrying the target sites of HIV-1 NF-kB, on the interaction between purified NF-kB p50 (upper part of the panel) or p52 (lower part of the panel) and 32 P-labelled HIV-1 NF-kB DNA–DNA target molecules. A total of 10 ng of NF-kB factors were incubated for 20 min in binding buffer in the absence (–) or in the presence of 100 ng of DNA–DNA, PNA–PNA and PDP–PDP molecules, as indicated. After this incubation period, a further 20 min incubation step was performed in the presence of the 32 P-labelled HIV-1 NF-kB DNA–DNA target molecule. Protein – DNA complexes are indicated by an arrow. Asterisks indicate the free 32 P-labelled NF-kB DNA–DNA. Control DNA–DNA competitors carrying binding sites for GATA-1 and NF-IL2A were used, as indicated. (B,C) Effects of increasing amounts of DNA–DNA and PDP – PDP hybrids carrying the target sites of HIV-1 NF-kB, on the interaction between purified NF-kB p50 (B) or p52 (C) and 32 P-labelled HIV-1 NF-kB DNA–DNA target molecules. A total of 10 ng of NF-kB factors were incubated for 20 min in binding buffer in the absence (–) or in the presence of the indicated concentrations of NF-kB DNA–DNA and NF-kB PDP–PDP molecules, as indicated. After this incubation period, a further 20 min incubation step was performed in the presence of 32 P-labelled HIV-1 NF-kBDNA–DNA target molecule. Protein–DNA complexes are marked with an arrow. Asterisks indicate the free 32 P-labelled NF-kB mer. Lane ‘b’ ¼ free 32 P-labelled HIV-1 NF-kB DNA–DNA target molecule, no NF-kB protein added. q FEBS 2001 Binding of NF-kB proteins to PNA–DNA chimeras (Eur. J. Biochem. 268) 6071 exonucleases, endonucleases, cellular extracts and serum. After incubations, the decoy molecules were isolated, layered on top of an agarose gel, electrophoresed and stained with ethidium bromide (results of a typical experiment are shown in Fig. 8A). Disappearance of the ethidium bromide stained bands gives evidence for degradation of the decoy molecules under the stated experimental conditions. The observed stabilities of PDP –PDP chimeras were compared to those of DNA–DNA decoy molecules. Examples of the results obtained are depicted in Fig. 8B –D, which clearly demonstrates that the decoy molecules based on PNA–DNA–PNA chimeras are resistant to Exo III 3 0 !5 0 exonucleases (Fig. 8B) and 5 0 !3 0 lambda exonuclease (Fig. 8C), unlike the corresponding DNA–DNA hybrid. Interestingly, also when experiments were performed employing DNase I, higher stability of the PNA–DNA– PNA based decoys was obtained (Fig. 8D). Taken together, these data suggest that PDP–PDP chimeras exhibit higher levels of resistance to nucleases with respect to decoy molecules based on DNA–DNA hybrids. When cytoplasmic extracts from human leukemic K562 cells or human serum was employed, we obtained results conferming the increasing stabilities of PDP– PDP chimeras when com- pared to DNA–DNA hybrid molecules (data not shown and M. Borgatti, C. Mischiati, N. Bianchi and R. Gambari, unpublished results). DISCUSSION The NF-kB/Rel family of transcription factors is involved in the control of the expression of a number of mammalian genes, such as those encoding for major histocompatibility complex (MHC) proteins, interferons and growth factors [12– 20]. In addition, transcription factors belonging to the NF-kB/Rel family are involved in the transactivation of viral genomes, such as HIV-1 [21]. In fact, it has been demonstrated that HIV-1 transcription depends on interactions between cellular transcription factors of the NF-kB/Rel family and two target sites (5 0 -GGGGACT- TTCC-3 0 ) present within the long terminal repeat [45]. Accordingly, biomolecular approaches able to inhibit NF-kB activity could be of interest for the experimental therapy of AIDS. For example, triple-helix-forming oligonucleotides are able to inibit HIV-1 LTR-directed transcription [46]. With respect to gene therapy, the decoy approach against NF-kB has been proposed as a useful tool to alter NF-kB dependent gene expression [12–20]. This was achieved by using ODNs as decoy molecules, carrying NF-kB specific cis-elements. Unfortunately, synthetic ODN are not stable and therefore should be extensively modified in order to be used in vivo or ex vivo [1 –7]. Fig. 5. Effects of NF-kB DNA–DNA, PNA–PNA and PDP–PDP and GATA-1 and NF-IL2 DNA–DNA hybrids on the interaction between crude nuclear extracts from B-lymphoid Raij cells and 32 P-labelled HIV-1 NF-kB DNA–DNA target molecules. A total of 1 mg of nuclear factors were incubated for 20 min in binding buffer in the absence (–) or in the presence of 50–200 ng of competitor molecules, as indicated. After this incubation period, a further 20 min incubation step was performed in the presence of 32 P-labelled HIV-1 NF-kB DNA – DNA target molecule. Protein–DNA complexes are marked with an arrow. Asterisks indicate the free 32 P-labelled NF-kBmer. Fig. 6. Effects of DNA–DNA and PDP – PDP hybrids carrying the target sites of HIV-1 NF-kB, on the interaction between crude nuclear extracts from B-lymphoid Raij or human leukemic K562 cells, as indicated, and 32 P-labelled NF-IL2A, STAT-1 and GATA-1 DNA–DNA target molecules. A total of 1 mg of nuclear factors were incubated for 20 min in binding buffer in the absence (–) or in the presence 100 ng of DNA–DNA and PDP – PDP molecules, as indicated. After this incubation period, a further 20 min incubation step was performed in the presence of 32 P-labelled DNA–DNA target molecules. Protein–DNA complexes are marked with an arrow. Asterisks indicate the 32 P-labelled NF-IL2, STAT-1 and GATA-1mers. Control DNA–DNA competitors carrying binding sites for STAT-1, GATA-1 and NF-IL2Awere used, in order to verify the specificity of protein–DNA interactions observed. 6072 A. Romanelli et al. (Eur. J. Biochem. 268) q FEBS 2001 In a recent paper, we have proposed PNAs as alternative reagents in experiments aimed at the control of gene expression involving the decoy approach [28]. In PNAs, the pseudopeptide backbone is composed of N-(2-aminoethyl)- glycine units [23,24]. PNAs hybridize with high affinity to complementary sequences of single-stranded RNA and DNA, forming Watson– Crick double helices [23,24] and are resistant to both nucleases and proteases [29]. We demonstrated that NF-kB p52 is able to bind to both NF-kB DNA–DNA and DNA–PNA hybrid mimicking the NF-kB target sites present in the HIV-1 LTR. However, the binding of the NF-kB DNA–PNA to NF-kB transcription factors was found to exhibit low stability, and therefore this reagent is expected to be not suitable for a decoy approach [28]. The main issue of the present paper was to determine whether PNA–DNA chimeras mimicking the NF-kB bind- ing sites are capable of stable interactions with both purified NF-kB p52 and p50, as well as nuclear factors from Fig. 7. Effects of increasing amounts of DNA–DNA and PDP–PDP hybrids carrying the target sites of HIV-1 NF-kB or Igk NF-kB, as indicated, on the interaction between purified NF-kB p50 (A) or p52 (B) and 32 P-labelled IgK NF-kB DNA – DNA target molecules. A total of 10 ng of NF-kB factors were incubated for 20 min in binding buffer in the absence (–) or in the presence of the indicated concentrations of DNA–DNA and PDP–PDP molecules, as indicated. After this incubation period, a further 20 min incubation step was performed in the presence of 32 P-labelled IgK NF-kB DNA – DNA target molecule. Protein–DNA complexes are marked with an arrow. Asterisks indicate the free 32 P-labelled NF-kB IgK DNA. Lane ‘b’ ¼ free 32 P-labelled Igk DNA–DNA target molecule, no NF-kB protein added. Fig. 8. Stability of decoy molecules. (A) Preliminary experiment showing the effects of ExoIII on DNA–DNA and PDP–PDP decoy molecules. A total of 250 ng of NF-kB PDP–PDP and DNA–DNA decoys were incubated for 10 min in the absence (a) or in the presence of 0.001 (b), 0.01 (c), 0.1 (d), 1 (e), 10 (f) and 100 (g) U of ExoIII in 20 mL reaction mixture. After incubation the decoy molecules were layered on the top of a 2% agarose gel and detected by ethidium bromide staining. B–D. Differential effects of ExoIII (B), lambda exonuclease (C) and DNase I (D) on DNA (open symbols) and PNA–DNA–PNA (closed symbols) based decoys. Length of incubation was 10 min with ExoIII, 30 min with lambda exonuclease and 30 min with DNase I. Disappearence of the decoy molecules was considered as an evidence of degradation by the employed enzymes. Results shown in panels B–D are presented as percentage of recovery with respect to control untreated reaction mixtures. q FEBS 2001 Binding of NF-kB proteins to PNA–DNA chimeras (Eur. J. Biochem. 268) 6073 B-lymphoid cells. DNA– PNA chimeras were originally designed to improve the poor cellular uptake and solubility of PNAs. More recently, they were found to exhibit biological properties typical of DNA, such as the ability to stimulate RNaseH activation and to act as substrate for cellular enzymes (for example DNA polymerases) [32,33]. No information is available in the literature on the possible use of double stranded PNA–DNA chimeras as target molecules of transcription factors. This is not an unexpected result, as the PNA–PNA hybrid structure is considerably different compared to the DNA–DNA double helix and therefore could alter the molecular structure of double- stranded PNA–DNA chimeras, perturbing the interactions with specific transcription factors. Therefore, molecular modeling was firstly employed for the design of the NF-kB PNA–DNA chimera [30] and prediction of molecular interactions between the PNA–DNA chimera and nuclear proteins belonging to the NF-kB family was performed by energy minimizations and molecular dynamics simulations [30,31]. Furthermore, the conforma- tional behaviour and the thermal stability of the PDP– PDP duplex chimera were studied by circular dichroism analysis and melting experiments, respectively. The results obtained with these independent approaches convergently demon- strated significant differences between the duplex NF-kB PDP–PDP chimera and the duplex NF-kBDNA–DNA. However, when interactions between the PNA– DNA chimeras and NF-kB proteins were studied by electrophoretic mobility shift assay, it was clearly demonstrated that the differences in molecular structure and conformation do not prevent the PDP–PDP chimera from binding to NF-kB transcription factors. We found indeed that the double stranded PDP –PDP chimeras mimicking the HIV-1 NF-kB binding sites are able to suppress the molecular interactions between HIV-1 LTR and p50, p52 and nuclear factors from B-lymphoid cells. Therefore, the results obtained conclu- sively demonstrate that the designed NF-kBDNA–PNA chimeras could be proposed as powerful decoy molecules. To our knowledge, this is the first report indicating that double stranded PNA–DNA chimeras are target molecules for transcription factors. In addition, we hope our results will have practical implications. The finding that DNA–PNA chimeras stably interact with NF-kB transcription factors encourages further experiments focused on the possible use of these molecules for the development of potential agents for a decoy approach in gene therapy. In this respect, the finding that PDP-based decoy molecules are more resistant than DNA–DNA hybrids to enzymatic degradation (Fig. 8 and data not shown) appears to be of great interest. Furthermore, their resistance can be improved further after complexation with cationic lipo- somes or microspheres (M. Borgatti, C. Mischiati, N. Bianchi and R. Gambari, unpublished results) to which PDP–PDP chimeras are able to bind in virtue of their internal DNA structure. ACKNOWLEDGEMENTS This work was supported by Istituto Superiore di Sanita ` (AIDS/1998– 99), CNR-PF Biotecnologie, MURST-PRIN-98 and Finalized Research funds (year 2001) from the Italian Ministry of Health. Mr Giuseppe Perretta is acknowledged for technical assistance. The authors thank Prof. J. H. van Boom and J. C. Verhejien for giving the possibility of synthesizing the chimeras in their laboratory. REFERENCES 1. Lim, C.S., Jabrane-Ferrat, N., Fontes, J.D., Okamoto, H., Garovoy, M.R., Peterlin, B.M. & Hunt, C.A. (1997) Sequence-independent inhibition of RNA transcription by DNA dumbbells and other decoys. Nucleic Acids Res. 25, 575–580. 2. Kozlov, I.A., Kubareva, E.A., Ivanovskaya, M.G. & Shabarova, Z.A. (1997) Design of new reagents on the base of DNA duplexes for irreversible inhibition of transcription factor NF-kappa B. Antisense Nucleic Acid Drug Dev. 7, 279–289. 3. Iwase, R., Namba, M., Yamaoka, T. & Murakami, A. (1997) Gene regulation by decoy approach (I): synthesis and properties of photo- crosslinked oligonucleotides. Nucleic Acids Symp. Series 37, 203–204. 4. Yamashita, J., Yoshimasa, T., Arai, H., Hiraoka, J., Tayaya, K., Miyamoto, Y., Ogawa, Y., Itoh, H. & Nakao, K. (1998) Identifi- cation of cis-elements of the human endothelin-A receptor gene and inhibition of the gene expression by the decoy strategy. J. Biol. Chem. 273, 15993 –15999. 5. Clusel, C., Meguenni, S., Elias, I., Vasseur, M. & Blumenfeld, M. (1995) Inhibition of HSV-1 proliferation by decoy phosphodiester oligonucleotides containing ICP4 recognition sequences. Gene Expr. 4, 301–309. 6. Morishita, R., Gibbons, G.H., Horiuchi, M., Ellison, K.E., Nakama, M., Zhang, L., Kaneda, Y., Ogihara, T. & Dzau, V.J. (1995) A gene therapy strategy using a transcription factor decoy of the E2F binding site inhibits smooth muscle proliferation in vivo. Proc. Natl Acad. Sci. USA 92, 5855–5859. 7. Cho-Chung, Y.S., Park, Y.G., Nesterova, M., Lee, Y.N. & Cho, Y.S. (2000) CRE-decoy oligonucleotide-inhibition of gene expression and tumor growth. Mol. Cell. Biochem. 212, 29 – 34. 8. Ishibashi, H., Nakagawa, K., Onimaru, M., Castellanous, E.J., Kaneda, Y., Nakashima, Y., Shirasuna, K. & Sueishi, K. (2000) Sp1 decoy transfected to carcinoma cells suppresses the expression of vascular endothelial growth factor, transforming growth factor beta1, and tissue factor and also cell growth and invasion activities. Cancer Res. 60, 6531–6536. 9. Kawauchi, M., Suzuki, J., Morishita, R., Wada, Y., Izawa, A., Tomita, N., Amano, J., Kaneda, Y., Ogihara, T., Takamoto, S. & Isobe, M. (2000) Gene therapy for attenuating cardiac allograft arteriopathy using ex vivo E2F decoy transfection by HVJ-AVE- liposome method in mice and nonhuman primates. Circ. Res. 87, 1063–1068. 10. Lee, Y.N., Park, Y.G., Choi, Y.H., Cho, Y.S. & Cho-Chung, Y.S. (2000) CRE-transcription factor decoy oligonucleotide inhibition of MCF-7 breast cancer cells: cross-talk with p53 signaling pathway. Biochemistry 39, 4863 – 4868. 11. Mischiati, C., Feriotto, G., Borgatti, M., Giacomini, P. & Gambari, R. (2000) Characterization of a major histocompatibility complex class II X-box-binding protein enhancing tat-induced transcription directed by the human immunodeficiency virus type 1 long terminal repeat. J. Virol. 74, 8989–9001. 12. Morishita, R., Sugimoto, T., Aoki, M., Kida, I., Tomita, N., Moriguchi, A., Maeda, K., Sawa, Y., Kaneda, Y., Higaki, J. & Ogihara, T. (1997) In vivo transfection of cis element ‘decoy’ against nuclear factor-kappaB binding site prevents myocardial infarction. Nat. Med. 3, 894–899. 13. Khaled, A.R., Butfiloski, E.J., Sobel, E.S. & Schiffenbauer, J. (1998) Functional consequences of the SHP-1 defect in motheaten viable mice: role of NF-kappa B. Cell. Immunol. 185, 49–58. 14. Ono, S., Date, I., Onoda, K., Shiota, T., Ohmoto, T., Ninomiya, Y., Asari, S. & Morishita, R. (1998) Decoy administration of NF-kappa B into the subarachnoid space for cerebral angiopathy. Hum. Gene Ther. 9, 1003–1011. 6074 A. Romanelli et al. (Eur. J. Biochem. 268) q FEBS 2001 15. Sharma, H.W., Perez, J.R., Higgins-Sochaski, K., Hsiao, R. & Narayanan, R. (1996) Transcription factor decoy approach to decipher the role of NF-kappa B in oncogenesis. Anticancer Res. 16, 61–69. 16. Tomita, N., Morishita, R., Tomita, S., Yamamoto, K., Aoki, M., Matsushita, H., Hayashi, S., Higaki, J. & Ogihara, T. (1998) Transcription factor decoy for nuclear factor-kappaB inhibits tumor necrosis factor-alpha-induced expression of interleukin-6 and intracellular adhesion molecule-1 in endothelial cells. J. Hypertens. 16, 993–1000. 17. Tomita, N., Morishita, R., Lan, H.Y., Yamamoto, K., Hashizume, M., Notake, M., Toyosawa, K., Fujitani, B., Mu, W., Nikolic- Paterson, D.J., Atkins, R.C., Kaneda, Y., Higaki, J. & Ogihara, T. (2000) In vivo administration of a nuclear transcription factor- kappaB decoy suppresses experimental crescentic glomerulone- phritis. J. Am. Soc. Nephrol. 11, 1244 – 1252. 18. Suzuki, J., Morishita, R., Amano, J., Kaneda, Y. & Isobe, M. (2000) Decoy against nuclear factor-kappa B attenuates myocardial cell infiltration and arterial neointimal formation in murine cardiac allografts. Gene Ther. 7, 1847–1852. 19. Tomita, N., Morishita, R., Tomita, S., Gibbons, G.H., Zhang, L., Horiuchi, M., Kaneda, Y., Higaki, J., Ogihara, T. & Dzau, V.J. (2000) Transcription factor decoy for NF kappaB inhibits TNF- alpha-induced cytokine and adhesion molecule expression in vivo. Gene Ther. 7, 1326–1332. 20. Vos, I.H., Govers, R., Grone, H.J., Kleij, L., Schurink, M., De Weger, R.A., Goldschmeding, R. & Rabelink, T.J. (2000) NF kappaB decoy oligodeoxynucleotides reduce monocyte infiltration in renal allografts. FASEB J. 14, 815–822. 21. Hosoya, T., Takeuchi, H., Kanesaka, Y., Yamakawa, H., Miyano- Kurosaki, N., Takai, K., Yamamoto, N. & Takaku, H. (1999) Sequence-specific inhibition of a transcription factor by circular dumbbell DNA oligonucleotides. FEBS Lett. 461, 136–140. 22. Uhlman, E. & Peyman, A. (1990) Antisense oligonucleotides: a new therapeutic principle. Chem. Rev. 90, 543–549. 23. Egholm, M., Buchardt, O., Nielsen, P.E. & Berg, R.H. (1992) Peptide nucleic acids (PNA): oligonucleotide analogues with an achiral peptide backbone. J. Am. Chem. Soc. 114, 1895–1897. 24. Egholm, M., Buchardt, O., Christiensen, L., Behrens, C., Freier, S.M., Driver, D.A., Berg, R.H., Kim, S.K., Norden, B. & Nielsen, P.E. (1993) PNA hybridizes to complementary oligonucleotides obeying the Watson –Crick hydrogen-bonding rules. Nature 365, 566–568. 25. Betts, L., Josey, J.A., Veal, J.M. & Jordan, S.R. (1995) A nucleic acid triple helix formed by a peptide nucleic acid–DNA complex. Science 270, 1838 – 1841. 26. Larsen, H.J. & Nielsen, P.E. (1996) Transcription-mediated binding of peptide nucleic acid (PNA) to double-stranded DNA: sequence- specific suicide transcription. Nucleic Acids Res. 24, 458–463. 27. Hanvey, J.C., Peffer, N.J., Bisi, J.E., Thomson, S.A., Cadilla, R., Josey, J.A., Ricca, D.J., Hassman, C.E., Bonham, M.A., Au, K.G., Carter, S.G., Bruckenstem, D.A., Boyd, V.L., Noble, S.A. & Babiss, L.E. (1992) Antisense and antigene properties of peptide nucleic acids. Science 258, 1481–1485. 28. Mischiati, C., Borgatti, M., Bianchi, N., Rutigliano, C., Tomassetti, M., Feriotto, G. & Gambari, R. (1999) Interaction of the human NF-kappaB p52 transcription factor with DNA–PNA hybrids mimicking the NF-kappaB binding sites of the human immuno- deficiency virus type 1 promoter. J. Biol. Chem. 274, 33114–33122. 29. Demidov, V.V., Potaman, V.N., Frank-Kamenetsk, M.D., Egholm, M., Buchard, O., Sonnichsen, S.H. & Nielsen, P.E. (1994) Stability of peptide nucleic acids in human serum and cellular extracts. Biochem. Pharmacol. 48, 1310 – 1313. 30. Saviano, M., Romanelli, A., Bucci, E., Pedone, C., Mischiati, C., Bianchi, N., Feriotto, G., Borgatti, M. & Gambari, R. (2000) Computational procedures to explain the different biological activity of DNA/DNA, DNA/PNA and PNA/PNA hybrid molecules mimicking NF-kappaB binding sites. J. Biomol. Struct. Dyn. 18, 353–362. 31. Saviano, M., Bucci, E., Romanelli, A. & Pedone, C. (1999) Conformational feature of Peptidic Nucleic Acids for the rational design of biologically active PNAs and hybrid PNA-oligonucleo- tides. Minerva Biotecnologica 11, 175–185. 32. Misra, H.S., Pandey, P.K., Modak, M.J., Vinayak, R. & Pandey, V.N. (1998) Polyamide nucleic acid-DNA chimera lacking the phosphate backbone are novel primers for polymerase reaction catalyzed by DNA polymerases. Biochemistry 37, 1917–1925. 33. Uhlmann, E., Peyman, A. & Will, D.W. (1997) Antisense: chemical modifications. In Encyclopedia of Cancer (Bertino, J.R., ed.) Vol. 1, pp. 64–81. Academic Press, San Diego, CA 34. Vinayak, R., Van der Laan, A.C., Brill, R., Otteson, K., Andrus, A., Kuyl-Yeheskiely, E. & Van Boom, J.H. (1997) Automated chemical synthesis of PNA and PNA–DNA chimera on a nucleic acid synthesizer. Nucleosides Nucleotides 16, 1653 – 1656. 35. van der Laan, A.C., Meeuwenoord, N.J., Kuyl-Yeheskiely, E., Oosting, R.S., Brands, R. & van Boom, J.H. (1995) Solid support synthesis of a PNA–DNA hybrid. Recl. Trav. Chim. Pays-Bas 114, 295–297. 36. Will, D.W., Breipohl, G., Langner, D., Knolle, J. & Uhlmann, E. (1995) The synthesis of polyamide nucleic acids using a novel monomethoxytrityl protecting-group strategy. Tetrahedron 51, 12069–12082. 37. Saviano, M., Aida, M. & Corongiu, G. (1991) Molecular dynamics simulation in vacuo and in solution of cyclolinopeptide A: a conformational study. Biopolymers 31, 1017 – 1024. 38. Sen, S. & Nilsson, L. (1998) Molecular dynamics of duplex systems involving PNA: structural and dynamical consequences of the nucleic acid backbone. J. Am. Chem. Soc. 120, 619 – 631. 39. Feriotto, G., Mischiati, C. & Gambari, R. (1994) Sequence-specific recognition of the HIV-1 long terminal repeat by distamycin: a DNAase I footprinting study. Biochem. J. 299, 451–458. 40. Dignam, J.D., Lebovitz, R.M. & Roeder, R.G. (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11, 1475–1489. 41. Chen, F.E., Huang, D., Chen, Y. & Ghosh, G. (1998) Crystal structure of p50/p65 heterodimer of transcription factor NF-kappaB bound to DNA. Nature 391, 410–413. 42. Rasmussen, H., Kastrup, J.S., Nielsen, J.N., Nielsen, J.M. & Nielsen, P.E. (1997) Crystal structure of a peptide nucleic acid (PNA) duplex at 1.7 A ˚ resolution. Nat. Struct. Biol. 4, 98–101. 43. Bloomfield, V.A., Crothers, D.M. & Tinoco, I.T. Jr (1999) Nucleic Acids: structures, properties and functions. University Science Book, Sausalito, CA. 44. Nelsen, B. & Sen, R. (1992) Regulation of immunoglobulin gene transcription. Int. Rev. Cytol. 133, 121–149. 45. Gaynor, R. (1992) Cellular transcription factors involved in the regulation of HIV-1 gene expression. AIDS 6, 347–363. 46. McShean, W.M., Rossen, R.D., Laughter, A.H., Trial, J., Kessler, D.J., Zendegui, J.G., Hogan, M.E. & Orson, F.M. (1992) Inhibition of transcription of HIV-1 in infected human cells by oligodeoxy- nucleotides designed to form DNA triple helices. J. Biol. Chem. 267, 5712–5721. q FEBS 2001 Binding of NF-kB proteins to PNA–DNA chimeras (Eur. J. Biochem. 268) 6075 . Molecular interactions between nuclear factor kB (NF -kB) transcription factors and a PNA DNA chimera mimicking NF -kB binding sites Alessandra Romanelli 1 ,. stranded target DNAs used as controls were 5 0 - TAATATGT AAAAACATT-3 0 (sense strand, NF-IL 2A) , 5 0 - CACTTGAT AACAGAAAGTGATAACTCT-3 0 (sense strand, GATA-1)