HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mischiati, C. Articles by Gambari, R. Search for Related Content PubMed PubMed Citation Articles by Mischiati, C. Articles by Gambari, R. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 46, 33114-33122, November 12, 1999

Interaction of the Human NF-B p52 Transcription Factor with DNA-PNA Hybrids Mimicking the NF-B Binding Sites of the Human Immunodeficiency Virus Type 1 Promoter^*

Carlo Mischiati^§, Monica Borgatti, Nicoletta Bianchi^¶, Cristina Rutigliano, Marina Tomassetti, Giordana Feriotto, and Roberto Gambari^**

From the  Department of Biochemistry and Molecular Biology and the  Biotechnology Center, Ferrara University, 44100 Ferrara, Italy

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We determined whether peptide nucleic acids (PNAs) are able to interact with NF-B p52 transcription factor. The binding of NF-B p52 to DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA hybrid molecules carrying the NF-B binding sites of human immunodeficiency type 1 long terminal repeat was studied by (i) biospecific interaction analysis (BIA) using surface plasmon resonance technology, (ii) electrophoretic mobility shift, (iii) DNase I footprinting, and (iv) UV cross-linking assays. Our results demonstrate that NF-B p52 does not efficiently bind to PNA-PNA hybrids. However, a DNA-PNA hybrid molecule was found to be recognized by NF-B p52, although the molecular complexes generated exhibited low stability. From the theoretical point of view, our results suggest that binding of NF-B p52 protein to target DNA motifs is mainly due to contacts with bases; interactions with the DNA backbone are, however, important for stabilization of the protein-DNA complex. From the practical point of view, our results suggest that DNA-PNA hybrid can be recognized by NF-B p52 protein, although with an efficiency lower than DNA-DNA NF-B target molecules; therefore, our results should encourage studies on modified PNAs in order to develop potential agents for the decoy approach in gene therapy.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

It is well established that both constitutive and tissue-specific regulation of gene expression is operated at the transcriptional level by the interaction between nuclear proteins (transcription factors) and promoter regions containing DNA elements (transcription signals) that exhibit specific nucleotide sequences (1-4). Several reviews reporting the nucleotide sequences of transcription signals and the relative binding proteins have been published (5-7). The requirement of protein-DNA interactions for the control of gene expression is clearly sustained by the finding that double-stranded DNA molecules could be used to alter gene transcription by the decoy approach (8). For instance, in vitro transfection of cis-element decoys against NF-B inhibits the expression of NF-B-regulated genes (major histocompatibility complex genes, interleukin-2 receptor , Igk, interleukin-6, -opioid receptor, preprogalanin, adhesion molecule-1) (8-12). Decoy molecules against other transcription factors (HNF-1, RFX1, NFYB, E2F) were found to alter specific functions in eukaryotic cells (13, 14). Therefore, these inhibitors have been proposed as a novel tool for the therapy of a variety of well characterized disorders (13-19).

A drawback of the decoy approach designed for the modulation of gene expression is the presence of intracellular DNases. Therefore, large amounts of DNA must be internalized by target cells in order to obtain biological responses leading to cytotoxic effects (20). On the other hand, modified oligonucleotides (either methylphosphonate or phosphorothioate) have been used by virtue of their resistance to DNase cleavage, but these molecules are highly toxic (20). A further problem of the decoy approach is the recently reported nonspecific activity of these molecules (9). For instance, dumbbell oligonucleotides to RFX1, in addition to a block of activation of RFX1-regulated genes, cause additional nonspecific effects, most likely via an interaction with the general transcription machinery (14).

Peptide nucleic acids (PNAs)¹ have recently been proposed as alternative reagents in experiments aimed at the control of gene expression (21-25). In PNAs, the pseudopeptide backbone is composed of N-(2-aminoethyl)glycine units (21). PNAs hybridize with high affinity to complementary sequences of single-stranded RNA and DNA, forming Watson-Crick double helices (22). In addition, PNAs form highly stable (PNA)₂-RNA triplexes with RNA targets (23). Therefore, antisense and antigene PNAs have been synthesized and characterized (24). For instance, PNA-RNA duplexes and (PNA)₂-RNA triplexes are powerful inhibitors of translation when they are designed to hybridize to targets overlapping AUG start sites (25). On the other hand, PNA-mediated inhibition of gene transcription is caused by a process that involves strand invasion of DNA with the generation of PNA-DNA-PNA triplexes (26).

With respect to the functional effects of the substitution of the phosphate backbone of DNA with the pseudopeptide backbone of PNAs, it should be emphasized that PNA-RNA hybrids are not a suitable substrate for RNase H (27); in addition, PNA-PNA duplexes were found to be completely unable to interact with minor groove binding drugs, including 4',6-diamidino-2-phenylindole and distamycin A (28). On the other hand, PNA-DNA chimeras lacking the phosphate backbone were recently found to be suitable primers for polymerase reaction catalyzed by DNA polymerases (29). Moreover, 4',6-diamidino-2-phenylindole and distamycin A were found to bind PNA-DNA duplexes by a sequence-selective interaction with the minor groove (28).

To date, no information is available on the possible use of PNAs as decoy molecules able to interact with DNA-binding proteins, including transcription factors (30). In this case, the biological effects of PNAs could be a significant reduction or even a block of the interactions of transcription factors with target promoter sequences. It is worth noting that PNAs should be designed to minimize difficulties in solubility (for instance, PNAs extremely rich in GC should be avoided) (31); in addition, palindromic DNA sequences (for instance, the symmetric GGGGATTCCCCT NF-B binding site of human p-selectin, human interleukin-2 receptor , mouse H2K, mouse major histocompatibility complex EA promoter regions) are probably not the first choice, due to the facility of single-stranded PNAs to self-hybridization or interstrand hybridization, possibly generating complexes exhibiting high stabilities (32).

For these reasons, we decided to perform experiments using the nonsymmetric NF-B binding site of the long terminal repeat (LTR) of the human immunodeficiency type 1 virus (HIV-1). In this respect, we emphasize that the decoy approach could be of interest in the experimental therapy of AIDS, since it has been demonstrated that transcription of HIV-1 depends on interactions between cellular transcription factors (including NF-B) and cis-acting elements present within the HIV-1 LTR (4, 33, 34).

The main issue of the present paper was to study the binding of the transcription factor NF-B p52 to DNA-DNA, DNA-PNA, PNA-DNA, or PNA-PNA molecules mimicking the NF-B binding sites present in the HIV-1 LTR.

The binding of NF-B p52 to these target molecules was studied by biospecific interaction analyses (BIA) using surface plasmon resonance and the BIAcore biosensor (35-38), and further characterized by UV cross-linking experiments (39). The effects of PNA-PNA and PNA-DNA hybrids on the molecular interactions between nuclear factors and oligonucleotides containing the NF-B binding sites were studied also by electrophoretic mobility shift assay and competitive DNase I footprinting (40, 41).

The results from these molecular analyses could have both theoretical and practical implications. For the theoretical point of view, the use of putative NF-B target molecules lacking a DNA backbone could help in understanding the role played by the DNA backbone in influencing either affinity of NF-B for DNA and/or stability of NF-B·DNA complexes. From the practical point of view, if PNA-DNA and/or PNA-PNA molecules are stably recognized by NF-B, PNAs could be proposed as potential agents for the decoy approach in gene therapy, due to their resistance to both DNases and proteinases.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Synthetic Oligonucleotides and Peptide Nucleic Acids-- The synthetic oligonucleotides used in this study were purchased from Amersham Pharmacia Biotech (Uppsala, Sweden). The PNAs were synthesized by ISOGEN Biosciences (Maarssen, The Netherlands).

BIA Technology-- Comparison of the binding kinetics of human NF-B p52 transcription factor with DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA target molecules was performed by surface plasmon resonance analysis on BIAcore-1000^TM (Amersham Pharmacia Biotech) (35-38). The sensor chip SA5 (research grade; Amersham Pharmacia Biotech) precoated with streptavidin was used. All procedures were performed at 25 °C and at a 5 µl/min flow rate (42, 43). The production of SA5 sensor chips containing NF-B target double-stranded DNA-DNA (42), DNA-PNA, and PNA-PNA molecules was achieved by (a) a 30-µl pulse (500 ng) injection of biotinylated NF-B oligomer (5'-TGGGGACTTTCCAG-3'), either biot(NF-B)DNA or biot(NF-B)PNA, to two different flow cells of the sensor chip, followed by (b) a 30-µl injection (500 ng) of the complementary NF-B oligomer (5'-CTGGAAAGTCCCCA-3'), either c(NF-B)DNA or c(NF-B)PNA, as required (Fig. 1 shows the location of NF-B binding sites within the HIV-1 LTR). Generation of double-stranded NF-B molecules was performed in HBS buffer (10 mM Hepes, pH 7.4, 150 mM NaCl, 3.4 mM EDTA, and 0.05% surfactant P2) in order to minimize triple-helix formation (43).

Preliminary experiments demonstrated that BIA technology allows highly reproducible hybridization between complementary DNA and PNA molecules and SA5 sensor chips carrying single stranded biot(NF-B)DNA or biot(NF-B)PNA molecules. In our experiments, we produced sensor chips carrying high (600-900 resonance units (RU)) or low (100-200 RU) amounts of biot(NF-B)DNA or biot(NF-B)PNA. Fig. 2A reports a representative example of the increase of RU bound to biot(NF-B)DNA sensor chips following injection of complementary DNA. Injection of complementary DNA to sensor chips carrying high amounts of biot(NF-B)DNA generated an increase of 492.7 ± 46.9 RU in 14 different experiments. Injection of complementary PNA to the same biot(NF-B)DNA sensor chips generated an increase of 621.6 ± 50.4 RU in six different experiments. Injection of complementary DNA to sensor chips carrying high amounts of biot-NF-B PNA generated an increase of 468.5 ± 87.5 RU in four different experiments. Finally, injection of complementary PNA to the same biot-NF-B PNA sensor chips generated an increase of 890 ± 125.9 RU in five different experiments. Regeneration of the sensor chips is obtained after a pulse with 5-10 µl of 50 mM NaOH without significant decrease in the capacity to bind either complementary PNA or complementary DNA. It should be pointed out that the obtained amounts of bound c(NF-B)DNA correspond to about 105 fmol/mm² of target DNA. Despite the fact that this amount of target NF-B DNA elements is similar to those used in a number of recent reports by research groups studying transcription factor/DNA interactions using BIAcore (44-47), we first determined whether these experimental conditions allow efficient binding of NF-B p52. The binding of human NF-B p52 transcription factor to double-stranded NF-B DNA target molecules was monitored after a 30-µl injection of purified NF-B (human p49; Promega Corp., Madison, WI) to SA5 sensor chips carrying high and low amounts of target DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA hybrid molecules. Injection was performed in binding buffer (buffer TF: 50 mM KCl, 20 mM Tris pH 7.5, 1 mM MgCl₂, 0.2 mM EDTA, 0.01% Triton X-100) and subsequent washing in the same buffer. Promega NF-B p49 is expressed in bacteria from a full-length cDNA encoding 447 amino acids (48). The concentration of NF-B p49 in the used stock solution was 300 ng/µl (3.4 gel shift units/µl). All buffers were filtered and degassed before the use. Sensorgrams analysis was performed by the BIA-evaluation software version 2.1 (Amersham Pharmacia Biotech). Suitable blank control injections with running and/or binding buffers, including control injections of p49 to blank sensor chips were performed, and, when appropriate, the resulting sensorgrams were subtracted from the experimental sensorgrams. The comparison of binding and dissociation curves in the sensorgram allows the determination of the stability of DNA-protein complexes (36).

Electrophoretic Mobility Shift Assay-- The electrophoretic mobility shift assay (41) was performed by using the double-stranded synthetic oligonucleotides mimicking the NF-B (the nucleotide sequences are given under "BIA Technology" and in Fig. 1) or the Sp1 (the plus strand sequence was 5'-GAGGCGTGGC-3') binding sites. The synthetic oligonucleotides were 5'-end-labeled using [-³²P]ATP and T4 polynucleotide kinase (MBI Fermentas, Milan, Italy). Binding reactions were set up as described elsewhere (41) in a total volume of 25 µl containing buffer TF plus 5% glycerol, 1 mM dithiothreitol, 10 ng of human NF-B p52 protein, or 10 ng of human purified Sp1 protein (Promega) and 0.25 ng of ³²P-labeled oligonucleotides. When 2 µg of crude nuclear extracts isolated from human lymphoid Raji and Jurkat cells were used instead of purified NF-B or Sp1 factors, the binding reaction was carried out in the presence of 1 µg of the nonspecific competitor poly(dI-dC)·poly(dI-dC) (41). After 5 min of binding at room temperature, the samples were electrophoresed at constant voltage (200 V) under low ionic strength conditions (0.25× TBE buffer: 22 mM Tris borate, 0.4 mM EDTA) on 6% polyacrylamide gels. Gels were dried and subjected to standard autoradiographic procedures (49). In competition experiments, the competitor molecules (DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA) were preincubated for 20 min with purified NF-B p52 protein, purified Sp1 factor, or nuclear extracts, before the addition of labeled target DNA. Nuclear extracts were prepared according to Dignam et al. (50). The nucleotide sequences of competitor target DNAs used as control were 5'-TAATATGTAAAAACATT-3' (sense strand, NFIL2A), 5'-AGCATGAGTCAGACAC-3' (sense strand, AP1), and 5'-GAACATGTCCCAACATGTTG-3' (sense strand, p53).

UV Cross-linking Assay-- UV cross-linking (39) was performed by using ³²P-labeled DNA-DNA or DNA-PNA molecules containing the NF-B binding site. In this case, the synthetic oligonucleotides of the hybrid molecules were 5'-end-labeled using [-³²P]ATP and T4 polynucleotide kinase (MBI Fermentas, Italy). Binding conditions were similar to those described above for the electrophoretic mobility shift assay. Briefly, 10 ng of human NF-B p52 protein were incubated in the presence or in the absence of 100 ng of DNA-DNA or DNA-PNA molecules for 20 min at room temperature in a total volume of 25 µl containing buffer TF plus 5% glycerol and 1 mM dithiothreitol, and, after this time, 2 ng of labeled probe (about 200,000 cpm) were added. After a further 5 min of binding at room temperature, the mixture was irradiated for 30 min using a UV transilluminator (254 nm; 7000 milliwatts/cm²) at a distance of 2 cm from the UV light source. The UV cross-linking reaction was blocked by the addition of 5 µl of 6× Laemmli gel loading buffer and electrophoresed at constant voltage (200 V) on a polyacrylamide-SDS gel (10% running gel, 4% stacking gel) (39). One lane was for the prestained protein molecular weight marker (broad range, 6-175 kDa, New England Biolabs Inc., Beverly, MA). After electrophoresis, the gel was fixed and dried under vacuum, and the proteins directly involved in the DNA interaction were identified by autoradiography.

Competitive DNase I Footprinting Assay-- 150 ng of the reverse HIV-1-R primer (5'-GGCAAGCTTTATTGAGGCT-3') were 5'-end-labeled by T4 polynucleotide kinase and [-³²P]ATP, purified by Sephadex G50 chromatography, precipitated, resuspended, and used together with the HIV-1-F primer (5'-ATTTCATCACATGGCCCGAG-3') to amplify a polymerase chain reaction fragment of 259 base pairs mimicking the region of the HIV-1 LTR depicted in Fig. 1. The ³²P-labeled polymerase chain reaction fragment was resuspended to have 100,000 cpm/footprinting reaction. Before incubation with the ³²P-labeled HIV-1 LTR polymerase chain reaction fragment, 1 footprinting unit of purified p52 NF-B (Promega) was incubated in 50 µl of buffer TF plus 5% glycerol, 1 mM dithiothreitol, for 20 min in the absence or in the presence of 200 ng of cold NF-B DNA-DNA, DNA-PNA, or PNA-PNA molecules. After this step, the binding reactions were carried out for an additional 5 min in the presence of 5'-end-labeled DNA. Fifty µl of 10 mM MgCl₂, 5 mM CaCl₂ were added 1 min before the addition of DNase I (Promega) diluted in 10 mM Tris-HCl, pH 8, to working concentrations immediately before use. After a 2-min reaction with DNase I, the footprinting reactions were blocked at room temperature by adding 95 µl of 200 mM NaCl, 30 mM EDTA, 1% SDS, 100 µg/ml yeast tRNA. Reactions were phenol-extracted and precipitated by adding 2.5 volumes of ethanol. The pellets were washed in 70% ethanol, air-dried, resuspended in 5 µl of loading dye (96% formamide, 6 mg/ml bromphenol blue), denatured for 2 min at 90 °C, ice-cooled for 1 min, and layered onto a 6% polyacrylamide (acrylamide/bisacrylamide ratio 19:1), 7 M urea sequencing gel. Control footprinting experiments were performed in the absence of NF-B p52 factor. Molecular weight markers were obtained by G + A sequencing reactions of the footprinting probes (49).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Design of PNAs and Synthetic Oligonucleotides-- The nucleotide sequence corresponding to a single asymmetric NF-B binding site of the HIV-1 LTR (14 base pairs) was chosen in order to maximize solubility of synthesized PNAs. In addition, unlike symmetric NF-B binding sites, possible problems related to self-hybridization and/or interstrand hybridization are expected to be minimal in the case of asymmetric NF-B binding site. For these reasons, the experiments were performed with two oligonucleotides and two PNAs carrying the HIV-1 LTR asymmetric NF-B binding site in both sense and antisense orientations (see Fig. 1). In addition to unmodified molecules, the sense strand PNA and the corresponding oligonucleotide were biotinylated, in order to prepare SA5 sensor chips suitable for generation of target DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules for BIA studies employing surface plasmon resonance and biosensor technologies.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Structure of the HIV-1 genome, location of NF-B and Sp1 binding sites, sequences of the oligonucleotides (ODN) and PNAs used, and location of the HIV-1-F and HIV-1-R polymerase chain reaction primers used for the production of the DNase I footprinting probe.

BIA of NF-B p52 Binding to Target DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA Molecules-- A typical sensorgram of NF-B p52 binding to a DNA-DNA target hybrid is depicted in Fig. 2B. Injection of 200 ng of purified p52 protein determined a sharp and easily detectable increase of resonance units (Fig. 2B, b), due (i) to a bulk contribution (374 RU) caused by the protein itself and the solution in which it was stored (10 mM Hepes, pH 7.9, 50 mM KCl, 0.2 mM EDTA, 2.5 mM dithiothreitol, 10% glycerol, 0.05% Nonidet P-40) and (ii) to its binding to immobilized double-stranded DNA. From the sensorgram depicted in Fig. 2B (step c) it could be concluded that p52·DNA-DNA complexes are stable. These data are in agreement with observations by Galio et al. (47) demonstrating high levels of stability of molecular interactions between NF-B p50 and target NF-B DNA elements from different promoters. In order to make the sensor chip available for other binding experiments, a regeneration step was performed by injection of 50 mM NaOH (Fig. 2B, d) and equilibration in binding buffer (Fig. 2B, e). In the experiment reported in Fig. 2C, the same amount of p52 (300 ng/50 µl of binding buffer) was injected to two different SA-5 sensor chips, one carrying low amounts (dotted line), the other carrying high amounts (continuous line) of target NF-B DNA-DNA molecules. Blank controls were performed by injecting p52 to empty flow cells, and the resulting sensorgrams were subtracted from the experimental sensorgrams. The results obtained in this experiment demonstrate that in our experimental conditions 2088 and 4549 RU (RU_fin  RU_i) of NF-B p52 were found bound to 113 RU and 297 RU (RU_i  RU_o) of target NF-B DNA-DNA, respectively. In both cases, the values of RU_res  RU_i were found to be similar to those of RU_fin  RU_i, suggesting that NF-B p52·DNA-DNA complexes are stable. In Fig. 2, D and E, we report the RU_fin  RU_i values when SA5 sensor chip flow cells carrying different amounts of target NF-B DNA-DNA molecules were employed (Fig. 2D) or when different amounts of p52 NF-B were injected to the same sensor chip (Fig. 2E). The expected RU_max obtained after injection of p52 NF-B could be obtained by the equation RU_max = (analyte M_r × ligand RU)/ligand M_r (35, 36). Accordingly, 100% saturation of binding sites after injection of p52 NF-B to a SA5 sensor chip containing 100 fmol of double-stranded NF-B target DNA sequence will be obtained with an increase of approximately 9800 RU. Therefore, when results from Fig. 2, C and D, are taken together, it is evident that the best experimental results (at least 89.9% of saturation of binding sites) are obtained with the SA-sensor chip carrying low amounts of NF-B target molecules, probably due to mass transport effects that can take place at a very high density of ligand molecules (35, 36). Therefore, sensor chips carrying low amounts of target DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules were used.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Surface plasmon resonance analysis of NF-B p52 binding to HIV-1 NF-B target DNA-DNA site. A, binding of c(NF-B)DNA to a SA5 streptavidin-coated sensor chip flow cell carrying biot(NF-B)DNA. Binding was performed by a 30-µl injection of HBS buffer containing 500 ng of c(NF-B)DNA (a). In order to study stability of the generated DNA-DNA complexes, a 30-µl injection of binding buffer was performed (b). Running and binding buffers were HBS. B, NF-B p52 binding to NF-B DNA-DNA molecules. The flow cell of the sensor chip prepared as described for A was used in this experiment. Double-stranded DNA-DNA molecules carrying the NF-B HIV-1 binding site were obtained by injection of 30 µl of HBS buffer containing 500 ng of the c(NF-B)DNA oligonucleotide (a). C, NF-B p52 binding to flow cells to which 113 (dotted lines) and 297 (continuous lines) RU of complementary c(NF-B)DNA have been attached. In B and C, binding of NF-B p52 was performed by injection of 30 µl of binding buffer containing 200 ng of the transcription factor (b) followed by a 30-µl injection of binding buffer (c). Regeneration of the sensor chip was performed by injection of 5 µl of 50 mM NaOH (d), followed by a 20-µl injection of binding buffer (e). Binding buffer (buffer TF) was used as running buffer. Sensorgrams reported in C have been blank-subtracted using the BIAevaluation 2.1 software. D, relationship between the level of bound c(NF-B)DNA and the amount of NF-B p52 bound to the SA5 sensor chip (RUfin  RUi). E, relationship between the amount of injected NF-B p52 (ng/30 µl) and NF-B p52 bound after 300-s injection onto SA5 sensor chips containing DNA-DNA NF-B target sites (bound NF-B = (RUfin  RUi), expressed as resonance units).

In Fig. 3, we compared the binding of NF-B p52 to DNA-DNA (A), DNA-PNA (B), PNA-DNA (C), and PNA-PNA (D) hybrid molecules by 30-µl injection of this protein (200 ng) into the appropriate flow cells. Comparison of the sensorgrams shown in Fig. 3 allows us to identify differences in both binding and dissociation kinetics. In particular, NF-B p52 binding to PNA-PNA (Fig. 3D, b) is much less efficient than binding to DNA-DNA (Fig. 3A, b), and the assembled NF-B p52·PNA-PNA complex is unstable with respect to the NF-B p52·DNA-DNA complex, as suggested by dissociation kinetics (compare Fig. 3D, c, and Fig. 3A, c). In addition, comparison of the sensorgrams obtained shows that the binding of NF-B p52 to the double-stranded DNA-PNA hybrid (Fig. 3B, b) occurs, although with lower efficiency with respect to the binding of NF-B p52 to DNA-DNA (Fig. 3A, b). Dissociation kinetics are also different, since the stability of NF-B p52·DNA-PNA complex is lower than that of the NF-B p52·DNA-DNA complex. The binding of NF-B p52 to PNA-DNA hybrids (Fig. 3C, b) exhibits intermediate levels of efficiency when comparison is performed with binding to DNA-PNA (Fig. 3B, b) and PNA-PNA (Fig. 3D, b) molecules. The stability of NF-B p52·PNA-DNA complexes is lower than that of NF-B p52·DNA-PNA complexes.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Binding of NF-B p52 to biotDNA-DNA (A), biotDNA-PNA (B), biotPNA-DNA (C), and biotPNA-PNA (D) hybrid molecules. In order to obtain the required hybrid molecules, 500 ng of c(NF-B)DNA (A (a) and C (a)) or c(NF-B)PNA (B (a) and D (a)) were injected (30-µl injection in HBS) to flow cells carrying biot(NF-B)DNA (A and B) or biot(NF-B)PNA (C and D). Following generation of the required hybrid molecules, 200 ng of NF-B p52 protein were injected in 30 µl of binding buffer (b). Subsequently, a washing step was performed by injecting 30 µl of binding buffer (c). Binding buffer (buffer TF) was also used as running buffer. Sensorgrams reported have been blank-subtracted using the BIAevaluation 2.1 software.

Table I contains information deduced from the sensorgrams depicted in Fig. 3. When the data corresponding to bound NF-B p52 (RU_fin  RU_i) are related to the amount of bound PNA or DNA (RU_i  RU_o) it could be concluded that the binding of NF-B p52 to the target molecules is as follows: DNA-DNA > DNA-PNA > PNA-DNA > PNA-PNA.

View this table: [in this window] [in a new window]   Table I Binding of p52 NF-B to DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules mimicking NF-B binding sites of HIV-1 LTR

Dissociation rates were also compared. The complete set of the results obtained is shown in Table II, which includes the results from 11 independent experiments, reporting the percentage of residual NF-B p52 bound to DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA target molecules after washing with binding buffer, calculated by the formula ((RU_res  RU_i)/(RU_fin  RU_i)) × 100. Higher values were consistently obtained in the cases of NF-B p52·DNA-DNA and NF-B p52·DNA-PNA complexes; lower values were found in the case of NF-B p52·PNA-DNA and NF-B·PNA-PNA complexes. Taken together, these results indicate that NF-B p52 does not efficiently bind to PNA-PNA hybrids mimicking the HIV-1 LTR NF-B binding sites, due also to instability of the generated NF-B p52·PNA-PNA complex. Among the two possible DNA-PNA hybrids (biotDNA-PNA or biotPNA-DNA), the biotDNA-PNA was found to be more efficiently recognized by NF-B p52, although the stability of the NF-B p52·DNA-PNA complexes was found to be lower with respect to that exhibited by NF-B p52·DNA-DNA complexes.

View this table: [in this window] [in a new window]   Table II Stability of p52 NF-B binding to NF-B DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules

DNA-PNA Hybrid Molecules Mimicking the HIV-1 LTR NF-B Binding Sites Inhibit the Interactions between Human Purified NF-B p52 and Target DNA-DNA Molecules-- The binding of NF-B to DNA-DNA molecules mimicking the HIV-1 NF-B target sites was analyzed by electrophoretic mobility shift assay (Fig. 4) (41). In preliminary experiments, 10 ng of human NF-B p52 were incubated for 20 min in the presence of double-stranded oligonucleotides mimicking the binding sites of NF-B, AP1, NFIL2A, and p53. After this binding period, a 5-min incubation was performed in the presence of the ³²P-labeled NF-B DNA-DNA probe, and the samples were analyzed by electrophoresis on native 6% polyacrylamide gels. In our hands, one retarded band was generated by the addition of labeled DNA-DNA to purified NF-B (Fig. 4, arrows). The results obtained clearly demonstrate that under these experimental conditions, only NF-B DNA-DNA molecules efficiently compete for the binding of NF-B p52 to the ³²P-labeled NF-B DNA-DNA probe (Fig. 4, left). When 10 ng of human NF-B p52 were incubated for 20 min in the presence of NF-B DNA-DNA, DNA-PNA, PNA-DNA, or PNA-PNA molecules before the addition of labeled NF-B DNA-DNA, only the NF-B DNA-DNA and DNA-PNA were found to efficiently compete (Fig. 4, middle and right). As expected, the NFIL2A DNA-DNA does not efficiently compete (Fig. 4, right). These results are fully in agreement with data obtained by surface plasmon resonance analysis and suggest that DNA-PNA hybrid molecules carrying HIV-1 NF-B binding sites could function as "decoy" molecules.

View larger version (62K): [in this window] [in a new window]   Fig. 4.   Effects of DNA-DNA (D/D), DNA-PNA (D/P), PNA-DNA (P/D), and PNA-PNA (P/P) hybrids, carrying the target sites of AP1, NFIL2A, p53, and HIV-1 NF-B, on the interaction between purified NF-B p52 and 32P-labeled NF-B DNA-DNA target molecules. 10 ng of NF-B factor were incubated for 20 min in binding buffer in the absence (none) or in the presence of 100 ng of DNA-DNA, DNA-PNA, PNA-DNA, or PNA-PNA molecules, as indicated. After this incubation period, a further 5-min incubation step was performed in the presence of 32P-labeled NF-B DNA-DNA target molecule. Protein-DNA complexes are indicated with arrows at the top of the panels. The free 32P-labeled NF-B oligomer is indicated with an arrow at the bottom of the panels. In the experiment reported in the middle panel, PNA-PNA molecules caused (i) a retardation in the electrophoretic migration of the gel shift probe, presumably due to strand invasion of 32P-labeled NF-B oligomer by PNA, as elsewhere reported (26), and (ii) the appearance of high molecular weight complexes unable to enter the polyacrylamide gels (see the band at the level of the wells), probably due to the formation of scarcely soluble complexes between PNA-PNA and the 32P-labeled NF-B mer. In the lower part of the figure, a quantitative presentation of the data reported is shown after densitometric analysis of the autoradiograms shown in the upper part.

Lack of Recognition by Transcription Factor Sp1 of NF-B DNA-PNA Molecules-- In order to further test the specificity of the putative NF-B DNA-PNA decoy molecule, the binding of purified transcription factor Sp1 was evaluated. In agreement with recent published observations (51), we reproducibly find, as reported in Fig. 5, that purified factor Sp1 efficiently recognizes NF-B binding sites. From the results shown in Fig. 5A, it is evident that similar amounts of Sp1 and NF-B cold DNA-DNA competitors are able to inhibit the interactions between purified Sp1 factor and ³²P-labeled Sp1 oligomer. In the experiment shown in Fig. 5B, 10 ng of purified human Sp1 transcription factor were incubated in the absence or in the presence of 100 ng of DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules mimicking the HIV-1 NF-B binding sites and then incubated with labeled Sp1 DNA-DNA molecules mimicking the HIV-1 Sp1 binding sites. After 5 min incubation, the samples were analyzed as already described. As expected, the formation of Sp1·DNA-DNA complexes was inhibited by Sp1 DNA-DNA molecules. In agreement with the data shown in Fig. 5A, the formation of the Sp1·DNA-DNA complexes was also inhibited by NF-B DNA-DNA, indicating that these molecules are recognized by the human Sp1 factor. On the contrary, DNA-PNA, PNA-DNA, and PNA-PNA molecules mimicking the HIV-1 NF-B binding sites did not inhibit the Sp1·DNA-DNA complex assembly. This finding suggests that DNA-PNA hybrid molecules could be considered in decoy experiments as "specific competitors" instead of DNA-DNA molecules.

View larger version (56K): [in this window] [in a new window]   Fig. 5.   A, effects of cold DNA-DNA hybrids containing the NF-B binding site, NF-B(D/D), and the Sp1 binding site, SP1(D/D), on the interaction between purified transcription factor Sp1 and 32P-labeled Sp1 DNA-DNA target molecules. 10 ng of purified Sp1 transcription factor were incubated for 20 min in binding buffer in the absence (0) or in the presence of the indicated amounts of the competitor oligonucleotides. After this incubation period, a further 5-min incubation step was performed in the presence of 32P-labeled Sp1 DNA-DNA target molecules. Protein-DNA complexes are indicated by arrows in the upper part of the gel. The free 32P-labeled Sp1 mer is indicated by an arrow in the lower part of the gel. The densitometric analysis of the autoradiograms is shown in the bottom of the panel. B, effects of DNA-DNA (D/D NFKB), DNA-PNA (D/P NFKB), PNA-DNA (P/D NFKB), and PNA-PNA (P/P NFKB) hybrids containing the NF-B binding site and DNA-DNA (D/D SP1) hybrid containing the Sp1 binding site on the interaction between purified transcription factor Sp1 and 32P-labeled Sp1 DNA-DNA target molecules. 10 ng of purified Sp1 transcription factor were incubated for 20 min in binding buffer in the absence (none) or in the presence of 100 ng of the indicated competitor molecules. After this incubation period, a further 5-min incubation step was performed in the presence of 32P-labeled Sp1 DNA-DNA target molecules. Protein-DNA complexes are indicated with an arrow in the upper part of the gel. The free 32P-labeled Sp1 oligomer is indicated by an arrow in the lower part of the gel. The densitometric analysis of the autoradiogram is shown at the bottom of the panel.

DNA-PNA Hybrid Molecules Mimicking the HIV-1 NF-B Binding Sites Inhibit the Interactions between Nuclear Proteins of Crude Nuclear Extracts and Target NF-B DNA-DNA Molecules-- Before determining the effects of NF-B DNA-PNA molecules on the molecular interactions between ³²P-end-labeled double-stranded NF-B (DNA-DNA) binding sites and crude nuclear extracts from eukaryotic cells, two further experimental approaches were undertaken using purified p52 NF-B: UV cross-linking (Fig. 6) and competitive DNase I footprinting (Fig. 7) assays.

View larger version (51K): [in this window] [in a new window]   Fig. 6.   UV cross-linking assay. 10 ng of human NF-B p52 protein were incubated in the presence or in the absence (none) of competitor DNA-DNA (D/D) or DNA-PNA (D/P) molecules carrying the HIV-1 NF-B binding sites. After 20 min of binding at room temperature, 32P-labeled DNA-DNA or DNA-PNA molecules containing the NF-B binding sites were added for 5 min, and the mixtures were irradiated for a further 30 min using a UV transilluminator and loaded onto a 10% polyacrylamide-SDS gel. Autoradiography is shown, demonstrating UV cross-linking of NF-B p52 protein to both NF-B DNA-DNA and DNA-PNA 32P-labeled probes. Origin of migration, NF-B p52·DNA-DNA, and NF-B p52·DNA-DNA complexes and free probes are indicated with arrows.

View larger version (45K): [in this window] [in a new window]   Fig. 7.   Competitive DNase I footprinting assay. Purified NF-B p52 was incubated for 30 min in the absence (none) or in the presence of 200 ng of cold NF-B DNA-DNA, DNA-PNA, and PNA-PNA molecules. After this step, the 32P-labeled HIV-1 LTR footprinting probe (labeled on the antisense strand) was added and the binding reactions were carried out for an additional 5 min. After DNase I digestion, phenol extraction, and ethanol precipitation, the obtained DNA fragments were layered onto a 6% polyacrylamide, 7 M urea sequencing gel. Control footprinting experiments were performed in the absence of NF-B p52 factor. Molecular weight markers were obtained by G + A sequencing reactions of the footprinting probe. The nucleotide sequences corresponding to NF-B and Sp1 binding sites are indicated.

Direct interaction of NF-B p52 with target NF-B DNA-DNA and DNA-PNA molecules was demonstrated by a UV cross-linking assay. In this assay, the binding of NF-B p52 to DNA-PNA hybrid molecules was fixed by exposing the protein·DNA-DNA or protein·DNA-PNA complexes to short wave UV light for 30 min (Fig. 6). The specificity of binding of NF-B p52 to DNA-DNA or DNA-PNA molecules was assessed by the addition of an excess amount of cold NF-B DNA-DNA or DNA-PNA molecules, respectively, before the UV cross-linking step. The results obtained demonstrate that NF-B p52 is able to interact efficiently with ³²P-labeled NF-B DNA-PNA target molecules.

The results of the competitive DNase I footprinting experiment are shown in Fig. 7. The ³²P-labeled HIV-1 LTR footprinting probe was resuspended to 100,000 cpm/footprinting reaction. Before incubation with the footprinting probe, 1 footprinting unit of purified p52 NF-B was incubated in 50 µl of binding buffer for 20 min in the absence or in the presence of 200 ng of cold NF-B DNA-DNA, DNA-PNA, or PNA-PNA molecules. After this step, the binding reactions were carried out for an additional 5 min in the presence of the 5'-end-labeled footprinting probe, and DNase I was added for 2 min. In the presence of a decoy effect, NF-B p52 should not be able to generate the expected footprints (40). Fig. 7 clearly demonstrates that, as expected, NF-B p52 generates a footprint that is competed by cold NF-B DNA-DNA molecules. Similarly, NF-B DNA-PNA molecules are able to interfere with the generation of this footprint. On the contrary, an NF-B p52-generated footprint was still detectable when the NF-B p52 protein factor was incubated in the presence of NF-B PNA-PNA molecules. The DNase I digestion pattern is not affected when the Sp1 binding sites region is considered. Taken together, the data reported in Figs. 6 and 7 conclusively demonstrate that NF-B p52 is able to recognize NF-B DNA-PNA molecules (Fig. 6) and that these molecules are able to act, unlike PNA-PNA hybrids, as in vitro decoy molecules (Fig. 7).

In order to determine the activity of the DNA-PNA molecules carrying NF-B binding sites in a more complex protein context, we repeated the experiments reported in Fig. 4 by using, instead of purified NF-B p52, crude nuclear extracts from lymphoid cells. In the experiment reported in the left panel of Fig. 8, DNA-DNA, DNA-PNA, PNA-DNA, and PNA-PNA molecules mimicking the HIV-1 NF-B binding sites were preincubated with nuclear extracts from Raji B-lymphoid cells and processed as described for the experiments shown in Fig. 4. The data obtained confirm that the DNA-PNA hybrid, but not the PNA-PNA molecule, inhibits the binding of protein factors to the ³²P-end-labeled NF-B DNA-DNA target molecule. Control oligonucleotides (NFIL2A, AP1, p53, and Sp1) were also inactive.

View larger version (70K): [in this window] [in a new window]   Fig. 8.   Effects of DNA-DNA (D/D NF-B, D/D AP1, D/D NFIL2A, D/D p53), DNA-PNA (D/P NFKB), and PNA-PNA (P/P NFKB) hybrids on the interaction between nuclear factors from Raji (left) and Jurkat (right) cells and the 32P-labeled NF-B (left) and NFIL2A (right) DNA-DNA probes. 2 µg of crude nuclear extracts were incubated for 20 min in binding buffer in the absence (none) or in the presence of 200 ng of DNA-DNA, DNA-PNA, PNA-DNA, or PNA-PNA hybrid molecules, as indicated, and then incubated with radiolabeled DNA-DNA probes for 5 min. Nuclear factor·DNA-DNA complexes are indicated with arrows in the upper part of the gel. The free 32P-labeled probe is indicated with an arrow in the lower part of the gel. The densitometric analyses of the autoradiograms are shown at the bottom of the figure.

Further control experiments, performed with ³²P-end-labeled NFIL2A DNA-DNA target molecules demonstrated that this interference is specific. In fact, while NFIL2A cold oligomer suppresses the binding of nuclear factors from Jurkat T-lymphoid cells to ³²P-end-labeled NFIL2A DNA-DNA target molecules, no inhibitory activity was found for control oligonucleotides (AP1 and p53) as well as for NF-B DNA-DNA, DNA-PNA, and PNA-PNA molecules.

These results, when considered together with the data shown in Fig. 4, suggest that the NF-B DNA-PNA hybrid molecules, although less efficient than NF-B DNA-DNA in the binding to the NF-B transcription factor, could be considered as a potential decoy molecule.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The NF-B/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 proteins, interferons, and growth factors (34, 52-54); in addition, transcription factors belonging to the NF-B/Rel family are involved in the transactivation of viral genomes, such HIV-1 (4, 33). In fact, it has been demonstrated that HIV-1 transcription depends on interactions between cellular transcription factors of the NF-B/Rel family and two target sites (5'-GGGGACTTTCC-3') present within the long terminal repeat (4). Accordingly, biomolecular approaches able to inhibit NF-B activity could be of interest in the experimental therapy of AIDS. For instance, triple helix-forming oligonucleotides are able to inhibit HIV-1 LTR-directed transcription (55).

With respect to gene therapy, the decoy approach against NF-B has been proposed as a useful tool to alter NF-B-dependent gene expression (8-12). This was achieved by using as decoy molecules synthetic oligonucleotides carrying NF-B specific cis-elements. Unfortunately, synthetic oligonucleotides are not stable and therefore should be extensively modified in order to be used in vivo or ex vivo (14).

The main issue of the present paper was to determine whether PNAs are able to interact with the NF-B p52 transcription factor. The reason for studying PNA is related to their low toxicity and to their stability when exposed to DNases and proteinases (21-25). The reason for studying interactions of PNAs with NF-B is in our opinion of some interest, since no information is available in the literature on the possible interaction between transcription factors and PNAs.

Our results clearly demonstrate that NF-B p52 is able to bind to both NF-B DNA-DNA and DNA-PNA hybrid mimicking the NF-B target sites present in the HIV-1 LTR. Low binding of NF-B p52 to PNA-PNA hybrids was found. In addition, deep differences of stabilities of the generated molecular complexes were found. BIA using the BIAcore-1000 demonstrated indeed that, unlike NF-B p52·DNA-DNA complexes, the NF-B p52·PNA-PNA complexes are highly unstable. Interestingly, we found that NF-B p52 was able to recognize one DNA-PNA target (DNA: 5'-TGGGGACTTTCCAG-3'), although the stability of this NF-B·DNA-PNA complex was lower than that of the NF-B·DNA-DNA complex. The other PNA-DNA target (DNA: 5'-CTGGAAAGTCCCCA-3') generates complexes with NF-B p52 with lower efficiency. The binding of NF-B p52 to NF-B DNA-PNA hybrids was further confirmed by three additional experimental approaches, competitive electrophoretic mobility shift assay (Figs. 4 and 5), UV cross-linking (Fig. 6), and competitive DNase I footprinting (Fig. 7).

This information has both theoretical and practical implications. From the theoretical point of view, our results suggest that the binding of NF-B p52 to target DNA motifs is mainly due to contacts with bases; interactions with the DNA backbone are, however, required for stabilization of the protein-DNA complex. Accordingly, we found that one DNA-PNA hybrid generates, with NF-B p52, molecular complexes exhibiting appreciable stability. We emphasize that our experiments do not clarify the biochemical basis of the interaction of NF-B p52 to the DNA-PNA hybrid. In particular, we have no data on the capacity of the DNA-PNA NF-B hybrid to generate a molecular structure exhibiting a major groove similar to that of a DNA-DNA hybrid molecule. Interestingly, however, the DNA portion of the DNA-PNA target hybrid contains the sequence (5'-GGGGAT-3') known to interact with NF-B p52 at the level of both DNA backbone and nucleotide bases (57). The results presented in this paper are consistent with the hypothesis that molecular contacts between NF-B p52 and the DNA backbone of the target cis-elements play a role in determining the stability of NF-B p52·DNA-DNA complexes.

From the practical point of view, the finding that DNA-PNA hybrids can be recognized by NF-B p52 allows us to propose these molecules for the development of potential agents for a decoy approach in gene therapy. Accordingly, gel shift data (Figs. 4 and 8) demonstrate that the DNA-PNA hybrid molecules are capable to interfere with the binding of NF-B to the HIV-1 LTR. In addition, our data suggest that the NF-B DNA-PNA hybrid, although less efficient in binding NF-B p52, could be more specific than the NF-B DNA-DNA hybrid, since it does not cause any inhibitory effects on Sp1·DNA-DNA interactions (Fig. 5). We emphasize that the specificity of decoy molecules is crucial, since dumbbells and other decoys inhibit transcription most likely via an interaction with the general transcription machinery (14). In addition, we want to point out, on one hand, that DNA-PNA hybrid molecules are expected to exhibit very low levels of molecular interactions with RNA and DNA. On the other hand, it should be mentioned that further experiments are necessary to determine whether DNA-PNA hybrids could be employed as potential decoy molecules for other transcription factors.

In conclusion, although the NF-B DNA-PNA molecule described in the present paper is expected to be less efficient then that dumbbell NF-B DNA molecules in performing NF-B decoy (due to the described lower binding efficiency and instability of NF-B p52·DNA-PNA complexes), the data reported in the present paper should encourage further experiments in order to find whether the described DNA-PNA hybrid is able to inhibit HIV-1 transcription. Modified PNAs should be designed, synthesized, and tested to find DNA-PNA and PNA-PNA hybrids able to generate highly stable complexes with NF-B, in order to obtain efficient decoy activity. Of interest, from this point of view, are the recently described PNA-DNA chimeras (58) or the structural variant PHONA, in which the peptide bond is replaced by a phosphonic acid ester bridge (56).

	ACKNOWLEDGEMENT

The BIAcore-1000 was obtained from "Commissione Grandi Attrezzature" of Ferrara University.

	FOOTNOTES

^* This work was supported by ISS (AIDS 1998), by the Consiglio Nazionale delle Ricerche Target Project on Biotechnology, by PRIN-1998, and by Progetto per la Ricerca Finalizzata 1998 (Ministero della Sanità).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Recipient of an Associazione Italiana per la Ricerca sul Cancro fellowship.

^¶ Recipient of an Federazione per la Ricerca sul Cancro fellowship.

^** To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, Via L. Borsari 46, 44100 Ferrara, Italy. Tel.: 39-532-291448; Fax: 39-532-202723; E-mail: gam@dns.unife.it.

	ABBREVIATIONS

The abbreviations used are: PNA, peptide nucleic acid; HIV-1, human immunodeficiency virus type 1; LTR, long terminal repeat; BIA, biospecific interaction analysis; RU, resonance units; RU_o, resonance units before injection of c(NF-B)DNA; RU_i, resonance units before NF-B p52 binding; RU_fin, resonance units after the binding step; RU_res, resonance units after the washing step.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES