Sequence-Recognition and Cleavage of DNA by a Netropsin

DNase I footprinting studies indicated that the conjugate interacts preferentially with AT-rich sequences, but the cleavage of DNA in the presence of ...
0 downloads 0 Views 291KB Size
Bioconjugate Chem. 2000, 11, 219−227

219

Sequence-Recognition and Cleavage of DNA by a Netropsin-phenazine-di-N-oxide Conjugate Philippe Helissey,† Sylviane Giorgi-Renault,*,† Pierre Colson,‡ Claude Houssier,‡ and Christian Bailly*,§ Laboratoire de Chimie The´rapeutique, Faculte´ des Sciences Pharmaceutiques et Biologiques, UMR CNRS-Universite´ Rene´ Descartes no. 8638, 4, Avenue de l’Observatoire, 75270 Paris Cedex 06, France, Laboratoire de Chimie Macromole´culaire et Chimie Physique, Universite´ de Lie`ge au Sart-Tilman 4000 Lie`ge, Belgium, and INSERM U-524 et Laboratoire de Pharmacologie Antitumorale du Centre Oscar Lambret, Place de Verdun, 59045 Lille, France. Received October 6, 1999; Revised Manuscript Received December 2, 1999

We report the synthesis, DNA-binding and cleaving properties, and cytotoxic activities of R-128, a hybrid molecule in which a bis-pyrrolecarboxamide-amidine element related to the antibiotic netropsin is covalently tethered to a phenazine-di-N-oxide chromophore. The affinity and mode of interaction of the conjugate with DNA were investigated by a combination of absorption spectroscopy, circular dichroism, and electric linear dichroism. This hybrid molecule binds to AT-rich sequences of DNA via a bimodal process involving minor groove binding of the netropsin moiety and intercalation of the phenazine moiety. The bidentate mode of binding was evidenced by linear dichroism using calf thymus DNA and poly(dA-dT)‚(dA-dT). In contrast, the drug fails to bind to poly(dG-dC)‚poly(dG-dC), because of the obstructive effect of the guanine 2-amino group exposed in the minor groove of this polynucleotide. DNase I footprinting studies indicated that the conjugate interacts preferentially with AT-rich sequences, but the cleavage of DNA in the presence of a reducing agent can occur at different sequences not restricted to the AT sites. The main cleavage sites were detected with a periodicity of about 10 base pairs corresponding to approximately one turn of the double helix. This suggests that the cleavage may be dictated by the structure of the double helix rather than the primary nucleotide sequence. The conjugate which is moderately toxic to cancer cells complements the tool box of reagents which can be utilized to produce DNA strand scission. The DNA cleaving properties of R-128 entreat further exploration into the use of phenazine-di-N-oxides as tools for investigating DNA structure.

INTRODUCTION

The antibiotics netropsin and distamycin bind with a high selectivity to AT-rich sequences of DNA (1). The crescent-shaped conformation of these two oligopyrrolecarboxamide-amidine antibiotics is ideally suited for fitting into the minor groove of the DNA double helix (2). They have long been used as prototypes for the design of sequence-specific drugs (3, 4). A large variety of conjugates containing pyrrole-amidine moieties attached to a polyamine, an alkylating agent, a metal complex, a enediyne structure, or intercalating chromophore have been synthesized and studied for their capacity to recognize and/or cleave specific DNA sequences (5). Some of these conjugates, such as the nitrogen mustarddistamycin conjugate FCE24517 (tallimustine), demonstrate significant anticancer activity (6, 7). These compounds provide an improved understanding of the mechanisms whereby small molecules can recognize and bind to specific nucleotide sequences in DNA (see ref 8 for a recent review). Netropsin and distamycin have also been coupled with photosensitive groups, including psoralen and coumarin derivatives, to induce light-dependent sequence-specific * To whom correspondence should be addressed. (S.G.-R.) Fax: (+33)143291403.E-mail: [email protected].(C.B.)Fax: (+33)320169229.E-mail: [email protected]. † UMR 8638 CNRS. ‡ Universite ´ de Lie`ge. § INSERM U-524.

reaction with DNA. In the same vein, minor groove binding oligopeptides have been attached to pyrene, cationic porphyrins or isoalloxazine photosensitizers (9). Upon photoactivation flavin-netropsin hybrids can generate oxy radicals capable of causing DNA breaks, preferentially at AT-rich sites (10, 11). We have also previously reported the DNA-binding properties of hybrids composed of a minor-groove binder covalently linked to anthraquinone derivatives. The conjugates were found to catalyze DNA cleavage after activation by irradiation (12). We have now extended our synthetic program with the design of oligopyrrolecarboxamide-phenazine-di-N-oxide hybrid ligands. Phenazine-di-N-oxide derivative can produce diffusible oxygen radicals responsible for DNA strand scission under physiological conditions (13). The proposed mechanism of action involves reductive activation of the phenazine moiety to produce O2-• or •OH-mediated DNA cleavage depending on aerobic or anaerobic conditions, in the absence of light or metal ions. Hecht and coworkers have also demonstrated that the cleavage activity of phenazine-di-N-oxide can be directed to specific sites in DNA via the tethering of oligonucleotides. Antisense oligonucleotides covalently linked to the phenazinedi-N-oxide prosthetic group via an aminopropyl chain can provoke DNA strand scission at the target DNA sequence (14). By analogy, we reasoned that the linkage of a phenazine-di-N-oxide group to a DNA minor groove binding element might plausibly lead to a synthetic endonuclease capable of inducing efficient DNA cleavage.

10.1021/bc990131t CCC: $19.00 © 2000 American Chemical Society Published on Web 02/23/2000

220 Bioconjugate Chem., Vol. 11, No. 2, 2000

Figure 1. Structure of the netropsin-phenazine-di-N-oxide conjugate R-128.

Here, we describe the synthesis, DNA-binding and strand cleaving properties, and cytotoxic activities of the netropsin-phenazine-di-N-oxide hybrid ligand R-128, the structure of which is shown in Figure 1. EXPERIMENTAL PROCEDURES

Chemistry. Melting points were determined on a Maquenne apparatus and are uncorrected. The IR spectra were recorded on a Perkin-Elmer 1600 spectrophotometer, and only the principal sharply defined peaks are reported. The 1H NMR spectra were recorded on a Bruker AC-300 (300.13 MHz) MHz spectrometer. Mass spectra were determined by FAB (fast atom bombardment) on a VG 70-SEQ (VG Analytical Ldt, G. B.) apparatus, source FAB (positive ions) with direct introduction using glycerolthioglycerol (1:1) as a matrix. Thin-layer chromatography was carried out on Merck GF 254 silica gel plates. Flash chromatography was performed on silica gel (Silice 60 A C. C., SDS). Methyl 6-(Phenazine-2-carboxamido)hexanoate (3). A solution of phenazine-2-carboxylic acid (2) (15) (0.863 g, 3.85 mmol), EDC (0.886 g, 4.62 mmol), HOBT (0.520 g, 3.85 mmol), triethylamine (1.60 mL, 11.55 mmol), and methyl 6-aminohexanoate (0.840 g, 4.62 mmol) in dry DMF (60 mL) was stirred for 24 h at 20 °C and then further amounts of amine (0.086 g, 0.38 mmol) and EDC (0.089 g, 0.46 mmol) were added. One day later, after evaporation to dryness, the residue was dissolved in MeOH and precipitated by the addition of Et2O. The resulting solid was filtered off and purified by flash chromatography [solvents: CH2Cl2-MeOH (10:0) to (8.5: 1.5)] to give 0.660 g (49%) of pure compound 3: mp 144 °C; IR (KBr) 1638, 1733 [ν (CO)], 3302 [ν (NH)] cm-1; NMR (Me2SO-d6) δ 1.41 (m, 2, CH2CH2CH2CH2CH2), 1.62 (m, 4, CH2CH2CH2CH2CH2), 2.36 (t, 2, CH2CO), 3.38 (m, 2, NHCH2 ), 3.62 (s, 3, OCH3), 8.07 (m, 2, ArH), 8.34 (m, 4, ArH), 8.81 (m, 1, ArH), 8.96 (t, 1, NHCH2). Anal. (C20H21N3O3), C, H, N. Methyl 6-(5,10-Dioxyphenazine-2-carboxamido)hexanoate (4). An aqueous solution of 30% H2O2 (2 mL) was added to a solution of compound 3 (0.200 g, 0.57 mmol) in glacial acetic acid (20 mL). The mixture was stirred at 50 °C for 48 h. After elimination of the solvant in vacuo, the residue was purified by flash-chromatography [solvents CH2Cl2-AcOEt (10:0) to (4:6)] to afford 0.141 g (64%) of pure orange compound 4: mp 165 °C; IR (KBr) 1632, 1719 [ν (CO)], 3313 [ν (NH)] cm-1; NMR (Me2SO-d6) δ 1.34 (m, 2, CH2CH2CH2CH2CH2), 1.60 (m, 4, CH2CH2CH2CH2CH2), 2.34 (t, 2, CH2CO), 3.35 (m, 2, NHCH2 ), 3.59 (s, 3, OCH3), 7.99 (m, 2, ArH), 8.30 (m, 1, ArH), 8.69 (m, 3, ArH), 9.07 (m, 2, ArH and NHCH2). Anal. (C20H21N3O5), C, H, N. 6-(5,10-Dioxyphenazine-2-carboxamido)hexanoic acid (5). To a suspension of ester 4 (0.200 g, 0.52 mmol) in EtOH (5 mL) was added a solution of NaOH (0.104 g, 2.61 mmol) in H2O (5 mL). After refluxing for 5 min, the reaction mixture was acidified with dilute HCl until a pH of 2 was reached. The resulting precipitate was filtered off and recrystallized from MeOH to give

Helissey et al.

0.152 g (78%) of pure compound 5: mp 216 °C; IR (KBr) 1630, 1698 [ν (CO)], 3276 [ν (NH) and ν (OH)] cm-1; NMR (Me2SO-d6) δ 1.38 (m, 2, CH2CH2CH2CH2CH2), 1.59 (m, 4, CH2CH2CH2CH2CH2), 2.26 (t, 2, CH2CO), 3.38 (m, 2, NHCH2 ), 8.00 (m, 2, ArH), 8.34 (m, 1, ArH), 8.65 (m, 3, ArH), 9.12 (m, 2, ArH and NHCH2), 12.07 (s, 1, CO2H). Anal. (C19H19N3O5), C, H, N. 3-[1-Methyl-4-[1-methyl-4-[5-[(5,10-dioxyphenazine2-carboxamido)]pentylcarboxamido]pyrrole-2-carboxamido] pyrrole-2-carboxamido]propionamidinium Chloride (R-128). A solution of acid 5 (0.111 g, 0.30 mmol), EDC (0.075 g, 0.39 mmol), HOBT (0.045 g, 0.33 mmol), and amine 6 (0.132 g, 0.36 mmol) in dry DMF (60 mL) was stirred for 24 h at 20 °C, and then further amounts of 6 (0.037 g, 0.10 mmol), EDC (0.019 g, 0.10 mmol), and HOBT (0.014 g, 0.10 mmol) were added. After 24 h, the solvent was removed under reduced pressure, the resulting residue was dissolved in MeOH and precipited by the addition of Et2O. The precipitate was filtered off and purified by flash chromatography [solvents CH2Cl2-MeOH (9.5:0.5) to (5:5)] to afford 0.108 g (50%) of pure R-128: mp 212-215 °C; IR (KBr) 1648, 1684 [ν (CO)], 3313 [ν (NH) and (NH2)] cm-1; NMR (Me2SO-d6) δ 1.38 (m, 2, CH2CH2CH2CH2CH2), 1.64 (m, 4, CH2CH2CH2CH2CH2), 2.28 (t, 2, CH2CO), 2.66 (m, 2, CH2C(d+NH2)NH2), 3.38 (m, 2, NHCH2 ), 3.52 (m, 2, CH2CH2C(d+NH2)NH2), 3.82 (s, 6, 2NCH3), 6.90, 6.94, 7.18, 7.21 (4s, 4 × 1, pyrrolic H), 7.98 (m, 2, ArH), 8.32 (m, 2, ArH and NHCH2), 8.62 (m, 3, ArH), 8.86 (broad s, 2, NH2), 9.13 (m, 4, ArH, NHCH2 and NH2), 9.94 (s, 2, 2NH). MS (FAB) m/z 683 [(M - Cl)+], 667.6 [(M - Cl-O)+], 651.3 [(M - Cl-2O)+]. Chemicals and Biochemicals. Ammonium persulfate, tris base, acrylamide, bis-acrylamide, ultrapure urea, boric acid, tetramethylethylenediamine, formic acid, piperidine, and formamide were from Sigma. The nucleoside triphosphate labeled with 32P (R-dATP) was obtained from Amersham. Unlabeled dATP, avian myeloblastosis virus reverse transcriptase, and restriction endonucleases EcoRI and PvuII were from Boehringer Mannheim (Germany). Bovine pancreatic deoxyribonuclease I (DNase I, Sigma Chemical Co.) was stored as a 7200 units/mL solution in 20 mM NaCl, 2 mM MgCl2, and 2 mM MnCl2, pH 8.0. The stock solution of DNase I was kept at -20 °C and freshly diluted to the desired concentration immediately prior to use. All other chemicals were analytical grade reagents, and all solutions were prepared using doubly deionized, Millipore filtered water. Absorption Spectra and Melting Temperature Studies. Melting curves were measured using an Uvikon 943 spectrophotometer coupled to a Neslab RTE111 cryostat. For each series of measurements, 12 samples were placed in a thermostatically controlled cell-holder, and the quartz cuvettes (10 mm path length) were heated by circulating water. The measurements were performed in BPE buffer pH 7.1 (6 mM Na2HPO4, 2 mM NaH2PO4, 1 mM EDTA). The temperature inside the cuvette was measured with a platinum probe; it was increased over the range 20-100 °C with a heating rate of 1 °C/min. The “melting” temperature Tm was taken as the midpoint of the hyperchromic transition. The Uvikon 943 spectrophotometer was also used to record the absorption spectra. Titrations of the drug with DNA, covering a large range of DNA phosphate/drug ratios (P/D), were performed by adding aliquots of a concentrated DNA solution to a drug solution at constant ligand concentration (20 µM). DNA blanks at the same nucleotide concentrations were prepared concomitantly and used as a reference in the recording of absorption spectra.

Netropsin-phenazine-di-N-oxide Conjugate

Circular Dichroism (CD). CD spectra were recorded from 250 to 500 nm on a Roussel-Jouan dichrograph Mark IV. Samples were placed in a quartz cell (3 mL, 1 cm path length) thermostated at 25 °C. The DNA (typically 70 µM) was dissolved in 10 mM sodium cacodylate buffer, pH 7, containing 100 mM NaCl. Concentrations of the ligands were in the range from 1 to 7 µM in a water-DMSO solution (0.1% maximum final DMSO concentration). Baseline corrected spectra were stored in an IBM/AT computer. Electric Linear Dichroism (ELD). ELD measurements were performed using a computerized optical measurement system built by C. Houssier and the procedures outlined previously were followed (16). All experiments were conducted at 20 °C with a 10 mm pathlength Kerr cell having 1.5 mm electrode separation, in 1 mM sodium cacodylate buffer, pH 7.0. The DNA samples were oriented under an electric field strength of 13 kV/cm, and the drug under test was present at 10 µM together with the DNA or polynucleotide at 100 µM unless otherwise stated. This electrooptical method senses only the orientation of the polymer-bound ligand: free ligand is isotropic and does not contribute to the signal (17). Purification and Radiolabeling of the DNA Substrates. The 265-mer fragment was rendered radioactive by 3′-32P-end labeling of the EcoRI-PvuII double digest of the plasmid pBS (Stratagene, La Jolla, CA) using [R-32P]dATP (6000 Ci/mmol) and avian myeloblastosis virus reverse transcriptase. The labeled digestion products were separated on a 6% polyacrylamide gel under nondenaturing conditions in TBE buffer (89 mM Trisborate, pH 8.3, 1 mM EDTA). After autoradiography, the requisite band of DNA was excised, crushed, and soaked in elution buffer (500 mM ammonium acetate and 10 mM magnesium acetate) overnight at 37 °C. This suspension was filtered through a Millipore 0.22 µm filter and the DNA was precipitated with ethanol. Following washing with 70% ethanol and vacuum-drying of the precipitate, the labeled DNA was resuspended in 10 mM Tris adjusted to pH 7.0 containing 10 mM NaCl. Footprinting Experiments. Cleavage reactions by DNase I were performed essentially according to the protocol previously described (18). Reactions were conducted in a total volume of 10 µL. Samples (3 µL) of the labeled DNA fragment were incubated with 5 µL of the buffer solution supplemented with the drug under test. After 30 min of incubation at 37 °C to ensure equilibration of the binding reaction, digestion was initiated by the addition of 2 µL of the endonuclease solution whose concentration had been adjusted to limit the enzyme attack to less than 30% of the starting material so as to minimize the incidence of multiple cuts in any strand (“single-hit” kinetic conditions). Optimal enzyme dilutions were established in preliminary calibration experiments. Typically, DNase I experiments included 0.01 unit/mL enzyme, 20 mM NaCl, 2 mM MgCl2, and 2 mM MnCl2, pH 7.3. At the end of the reaction time (routinely 4 min at room temperature), the digestion was stopped by freeze-drying. After lyophilization, each sample was washed once with deionized water and then lyophilized again prior to resuspending in 3 µL of an 80% formamide solution containing tracking dyes (purchased from Sigma). Cleavage of Plasmid DNA. Each reaction mixture contained 4 µL of supercoiled pUC12 DNA (∼0.5 µg), 5 µL of R-128 at concentrations varying from 1 to 500 µM, and 1 µL of 10 mM dithiothreitol. After 14 h of incubation at room temperature in the dark, 3 µL of loading buffer (0.25% bromophenol blue, 0.25% xylene cyanol, and 30%

Bioconjugate Chem., Vol. 11, No. 2, 2000 221

glycerol in H2O) was added to each tube and the solution was applied to a 1% agarose gel. The electrophoresis was carried out for about 2 h at 100 V in TBE buffer. Gels were stained with ethidium bromide (1 µg/mL) then destained for 30 min in water prior to being photographed under UV light. Cleavage of Radiolabeled DNA. The reaction mixture contained 4 µL of 3′-end-labeled 265-mer DNA (∼500 cps), 5 µL of R-128 at concentrations varying from 1 to 500 µM, and 1 µL of 10 mM dithiothreitol. After 14 h of incubation at room temperature in the dark, the drugDNA samples were precipitated with 100 µL of cold ethanol and then resuspended in 4 µL of an 80% formamide solution containing tracking dyes. Samples were heated at 90 °C for 4 min and chilled in ice for 4 min prior to electrophoresis. Electrophoresis and Quantitation. DNase I- and DTT-mediated DNA cleavage products were resolved by polyacrylamide gel electrophoresis under denaturing conditions (0.3 mm thick, 8% acrylamide containing 8 M urea). Electrophoresis was continued until the bromophenol blue marker had run out of the gel (about 2.5 h at 60 W, 1600 V in TBE buffer, BRL sequencer model S2). Gels were soaked in 10% acetic acid for 15 min, transferred to Whatman 3MM paper, dried under vacuum at 80 °C, and subjected to the phosphorimager. A Molecular Dynamics 425E PhosphorImager was used to collect data from storage screens exposed to the dried gels overnight at room temperature. Baseline-corrected scans were analyzed by integrating all the densities between two selected boundaries using ImageQuant version 3.3 software. Each resolved band was assigned to a particular bond within the DNA fragment by comparison of its position relative to sequencing standards generated by treatment of the DNA with formic acid (G+A) followed by piperidine-induced cleavage at the modified bases. Cell Cultures and Survival Assay. Human HL-60 promyelocytic leukemia and Raji mammary carcinoma cells were obtained from the European Collection of Cell Cultures. Cells were grown at 37 °C in a humidified atmosphere containing 5% CO2 in RPMI 1664 medium, supplemented with 10% fetal bovine serum, glutamine (2 mM), penicillin (100 IU/mL), and streptomycin (100 µg/mL). The cytotoxicity of the drug was assessed using a cell proliferation assay developed by Promega (CellTiter 96 AQueous one solution cell proliferation assay, Promega). Briefly, 2 × 104 exponentially growing cells were seeded in 96-well microculture plates with various drug concentrations in a volume of 100 µL. After 72 h of incubation at 37 °C, 20 µL of MTS (19) was added to each well and the samples were incubated for a further 2 h at 37 °C. Plates were analyzed on a Labsystems Multiskan MS (type 352) reader at 492 nm. RESULTS

Chemistry. The 2-aminophenazine-di-N-oxide derivatives which have been conjugated to oligonucleotides (13, 14) were not optimized for reductive generation of oxygen free radicals. In our case, we decided to substitute the phenazine-5,10-dioxide nucleus at position 2 by a carboxamido function to facilitate its reduction and the radical formation. The hybrid molecule R-128 contains a phenazine-5,10-dioxide heterocycle linked to the pyrrolecarboxamide moiety via a pentamethylenic chain (Figure 1). The synthesis of the hybrid R-128 required the preparation of the 2-carboxamidophenazine-5,10dioxide moiety substituted at position 2 with a pentamethylenic carboxylic acid linker 5 and the aminooligopy-

222 Bioconjugate Chem., Vol. 11, No. 2, 2000

Helissey et al.

Scheme 1

rrole derivative bearing an amidinium function 6 (Scheme 1). Compound 6 was prepared from 1-methyl-4-nitropyrrole-2-carboxylic acid according to a procedure previously described (12). Three synthetic routes have been described in the literature for the preparation of phenazine-2-carboxylic acid (2) (15, 20, 21). The method reported by Kidani (15) was found to be the most efficient although we have not be able to reproduce the 75% yield claimed for the chromic oxidation of 2-methylphenazine (1) (22). Our first attempted route to oxidize the acid 2 into 5,10-dioxyphenazine-2-carboxylic acid by hydrogen peroxide in acetic medium according to ref 23 was unsuccessful whatever the experimental conditions. Therefore, the amide 3 was prepared by condensation of 2 with methyl 6-aminohexanoate. The highest yield (49%) was obtained using EDC, HOBT, and triethylamine as peptide coupling agents. Treatment of 3 with hydrogen peroxide in acetic medium at 50 °C afforded 4 in a 64% yield. After saponification of 4, a final peptide coupling between acid 5 and amine 6 gave the hybrid R-128 in 50% yield. DNA Affinity. The absorption spectrum of conjugate is largely modified in the presence of DNA (Figure 2). The two absorption bands centered at 288 and 477 nm, corresponding to the netropsin and phenazine moieties, respectively, are shifted by about 15 nm when the ligand is fully bound to DNA. The interaction of the drug with DNA also causes a marked hypochromism in the 290 nm band. The UV spectra suggested that the two moieties of the hybrid drug are engaged in the complex with DNA. The equilibrium constant measured for the binding of R-128 to calf thymus DNA is 1.2 × 106 (M-1, per nucleotide; n ) 3.5), i.e., about four times higher than that measured with netropsin (Ka ) 2.9 × 105 M-1, per nucleotide) under identical conditions. Therefore, we can

Figure 2. Absorption spectra of R-128 (15 µM) in the absence (full line) and presence (dashed line) of calf thymus DNA (500 µM) in 1 mM sodium cacodylate buffer, pH 7.0. The spectrum corresponding to the drug bound to DNA was referenced against a DNA solution of exactly the same concentration and was adjusted to a common baseline.

conclude that the linkage of the phenazine moiety reinforces significantly the interaction of the drug with DNA. We examined the ability of R-128 to affect the thermal denaturation profile of nucleic acids using three natural DNAs of different base pair composition [DNAs from calf thymus (42% GC), Clostridium perfringens (26% GC), and Micrococcus lysodeikticus (72% GC)] and two synthetic

Netropsin-phenazine-di-N-oxide Conjugate

Bioconjugate Chem., Vol. 11, No. 2, 2000 223

Figure 4. Reduced electric linear dichroism (∆A/A) spectra of (O) calf thymus DNA alone or in interaction with (0) netropsin or (b) R-128. ELD data were recorded in the presence of 100 µM DNA and 10 µM drug, in 1 mM sodium cacodylate buffer pH 7.0, under a field strength of 13 kV/cm.

Figure 3. (A) Variation in melting temperature (∆Tm in BPE buffer) of poly(dA-dT)‚(dA-dT) (open circles) and calf thymus DNA (filled circles) induced by the binding of R-128. The DNA concentration was fixed at 100 µM while the drug concentration varied from 1 to 20 µM. (B) Stabilizing effect of R-128 measured with various DNA and polynucleotides. In this case, the DNA and drug concentrations were 100 and 10 µM, respectively.

polynucleotides with different base pair arrangements poly(dA-dT)‚poly(dA-dT) and poly(dI-dC)‚poly(dI-dC). The stabilization of poly(dA-dT)‚poly(dA-dT) by R-128 against thermal denaturation is much higher than that observed with calf thymus DNA, which contains roughly equal proportions of AT and GC base pairs (Figure 3A). With the natural DNA, the measured ∆Tm values indicate that the extent of stabilization is related to the AT content of the DNA (Figure 3B). The thermal protection of the double helix induced by R-128 is much higher with the DNA from Clostridium than with the DNA from Microccocus. It was not possible to study the thermal stability with poly(dG-dC)‚poly(dG-dC) due to the very high stability of this polynucleotide, under the ionic conditions used in these experiments (Tm > 90 °C in BPE buffer), but we used the polynucleotide poly(dI-dC)‚poly(dI-dC), which contains inosine residues in place of guanosines. The effects observed with poly(dA-dT)‚poly(dA-dT) and poly(dI-dC)‚poly(dI-dC) can be compared because the two polynucleotides melt at about the same temperature (Tm ) 40 °C in BPE buffer). The stabilizing effect of R-128 was much higher with the AT polymer than with the IC polymer. DNA Binding Mode. Circular dichroism (CD) and electric linear dichroism (ELD) experiments were performed to define the orientation of the hybrid ligand with respect to the DNA helix. With the DNAs from Micrococcus and Clostridium, the CD signal monitored at 300320 nm increased with increasing DNA/drug ratios until a D/P value of 0.1 was reached (spectra not shown). Such an intense positive CD is characteristic of netropsin and is commonly observed with netropsin and distamycin conjugates. No CD signal was detected at 500 nm in the phenazine band, and therefore, it was not possible to define the orientation of the photosensitive group by this optical technique. In contrast, the electric linear dichroism (ELD) measurements provided very useful information. In the past, we have employed extensively this

electrooptical method to determine the orientation of drugs bound to DNA, including both minor groove binders and intercalators as well as a variety of hybrid molecules (17). The ELD spectrum of R-128 bound to calf thymus DNA (Figure 4) provides direct evidence that both linked functionalities of the hybrid are engaged in the binding reaction. Such an ELD spectrum is characteristic of a bimodal binding process, with the 400-500 nm negative band due exclusively to the phenazine moiety of the hybrid and the 300-330 nm positive band due to the bispyrrole moiety, with little or no contribution from the phenazine moiety. Two essential facts can be obtained from these ELD data. First, the positive dichroism at 320 nm is similar for the hybrid R-128 and its parent compound netropsin. Second, the reduced dichroism of the DNA bases at 260 nm in the absence of ligand (∆A/A ) -0.41 at 13 kV/cm) is the same as the reduced dichroism measured at 490 nm with the hybrid ligandDNA complex. This implies that (i) the bis-pyrrole moiety of R-128 is inserted into the minor groove of DNA, as is known to be the case with netropsin and (ii) the phenazine portion of the conjugate is oriented parallel to the DNA base pair, and this is perfectly consistent with an intercalation binding mode. It seems therefore that the linkage of the phenazine moiety does not impede minor groove binding of the attached bis-pyrrole unit, and conversely, the minor groove entity does not hinder intercalation of the attached planar chromophore. Thus, the ELD experiments strongly suggest that the hybrid molecule interacts with DNA in a geometrically welldefined fashion, placing its netropsin moiety in the minor groove and with the phenazine chromophore inserted between the base pairs. ELD was used further to compare the binding mode of R-128 to poly(dA-dT)‚poly(dA-dT) and poly(dG-dC)‚ poly(dG-dC) (Figure 5). The ELD spectrum of R-128 bound to the AT polymer is almost the same as the one obtained with calf thymus DNA. It shows a positive band at 320 nm and a negative band at 480 nm, in agreement with a bidentate mode of binding involving minor groove binding of the netropsin moiety and intercalation of the phenazine ring. In sharp contrast, the ELD measurements performed in the presence of the GC polymer and R-128 showed no specific signal, be it at 320 or 480 nm. Both the netropsin moiety and the appended phenazine ring apparently fail to bind to the GC sites. The removal of the 2-amino group of guanine protruding into the minor groove of DNA restored the dual binding mode. Indeed, with poly(dI-dC)‚poly(dI-dC), the ELD spectrum was again totally similar with those obtained with calf

224 Bioconjugate Chem., Vol. 11, No. 2, 2000

Helissey et al.

Figure 5. Reduced electric linear dichroism (∆A/A) spectra of R-128 bound to (O) poly(dA-dT)‚poly(dA-dT), (0) poly(dG-dC)‚ poly(dG-dC), or (b) poly(dI-dC)‚poly(dI-dC). ELD data were recorded in the presence of 100 µM polynucleotide and 10 µM drug, in 1mM sodium cacodylate buffer, pH 7.0, under a field strength of 13 kV/cm.

thymus DNA or poly(dA-dT)‚poly(dA-dT) (Figure 5). It is known that the guanine 2-amino group obstructs the insertion of netropsin into the minor groove (24), but it is surprising to observe that the intercalation of the phenazine ring is also directly dependent on the absence of that group. Sequence selectivity. The classical DNase I footprinting methodology was used to determine if the hybrid molecule can recognize preferentially certain sequences in DNA. Netropsin is reputed to be capable of selectively recognizing AT-rich sequences (25). The footprinting gel presented in Figure 6 shows that R-128 exhibits a moderate sequence selectivity. The radiolabeled 265 base pair fragment from plasmid pBS was incubated with graded concentrations of R-128 prior to limited cleavage by DNase I. Two marked footprints can be discerned around nucleotide positions 77 and 125, which both correspond to AT-rich sequences. The densitometric analysis of the gel revealed other weaker R-128 binding sites around positions 26, 55, 64, and 94 as presented in Figure 7. Apart from the binding site at position 24-28 (3′-GAGAT), all the other sites coincide with sequences composed of at least four contiguous A‚T base pairs. The hybrid has thus retained, at least partially, the AT selectivity of netropsin, but the footprints are less frequent and less pronounced than those obtained with the antibiotic on the same DNA fragment (26). DNA Cleavage. Prior to investigating the sequence specificity of the cleavage reaction with R-128, we optimized the procedure using a plasmid DNA analyzed on agarose gel. A typical gel obtained after treatment of a supercoiled plasmid for 12 h with increasing concentrations of R-128 is presented in Figure 8. The cleavage of DNA is relatively limited. Supercoiled DNA (form I) is gradually converted to nicked DNA (form II) in the presence of concentrations of R-128 ranging from 2 to 75 µM. At the highest drug concentration, about 80% of the DNA is under a nicked form but no linear DNA (form III) was detected. Therefore, the drug only produces single-stranded cleavage of DNA. Next, we repeated the cleavage experiments using the radiolabeled 265 bp substrate as used in the footprinting experiments. A typical gel is presented in Figure 9, and the positions of the most intense cleavage sites are represented by arrows in Figure 7. In the control lane with no drug present, the DNA remained intact after the incubation with DTT. With R-128, the cleavage of DNA is very clear. At the lowest drug concentration (1 µM), the cutting is almost uniform; all nucleotide positions are cleaved. When the drug concentration is raised, the patterns of cleavage

Figure 6. DNase I footprinting of R-128 bound to the 265mer EcoRI-PvuII restriction fragment cut out from plasmid pBS. The DNA was 3′-end labeled with [R-32P]dATP in the presence of AMV reverse transcriptase. The drug concentration (µM) is shown at the top of the appropriate gel lanes. The tracks labeled “Ct” contained no drug. The tracks labeled “GA” represent formic acid-piperidine markers specific for purines. Numbers on the left side of the gels refer to the numbering scheme of the fragment as indicated in Figure 7.

changed markedly. Certain sequences become poorly cut, whereas the extent of cleavage at other sites has increased considerably. The gel was analyzed by densitometry, and the positions of the most pronounced cleavage sites were mapped and compared with the positions of the drug-binding sites, inferred from the footprinting experiments (Figure 7). No obvious correlation can be

Netropsin-phenazine-di-N-oxide Conjugate

Bioconjugate Chem., Vol. 11, No. 2, 2000 225

Figure 7. Summary map representing the DNase I footprints and DNA strand cleavage sites produced by R-128 on the lower strand of the 265-mer DNA fragment. Only the region of the restriction fragment which was analyzed by densitometry is shown. The sequences underlined by black boxes represent the measured footprints, i.e., positions of inhibition of DNase I cutting by R-128 and, therefore, the putative reversible binding sites. The thickness of each box is roughly proportional to the relative intensity of the footprint at the indicated sequence. Arrows pointing toward the lettered sequence refer to the main sites of R-128-mediated cleavage, with small and large arrows denoting moderate and large variations in the amplitude of the cleavage, respectively.

Figure 8. Cleavage of closed circular DNA. Supercoiled DNA (0.5 µg) was incubated overnight at 20 °C with the drug at the indicated concentration in the presence of 1 mM dithiothreitol (DTT). The control lane (marked 0) refers to the plasmid DNA incubated without drug but in the presence of DTT. Forms I and II refer to the supercoiled and nicked DNA forms, respectively.

discerned. The cleavage can occur at sequences adjacent to the AT-binding sites (such as around positions 66 and 78) as well as at sequences distal from the binding sites (such as around positions 45 and 99 for examples). Therefore, it was not possible to correlate the cleavage experiments with the binding data. Cytotoxicity. The cytotoxic activities of the hybrid ligand R-128 were evaluated using human HL-60 leukemia cells and human Raji mammary carcinoma cells. After a 72 h incubation period, IC50 values of 0.75 and 0.08 µM were calculated with HL-60 and Raji cells, respectively. In the same cytotoxicity assay, netropsin proved totally nontoxic to both cell lines. DISCUSSION

The conjugate described here was designed to contain a bispyrrolecarboxamide moiety covalently linked to a 2-carboxamidophenazine-di-N-oxide chromophore capable of triggering reductive cleavage of DNA. The complementary biophysical and footprinting data show that substituting the phenazine chromophore for the guanidinium terminal residue of netropsin does not impede the AT sequence preference of the parent antibiotic. The conjugate can still recognize AT-rich sequences in DNA selectively. Conversely, the cleavage experiments concur that the introduction of the minor groove binding element does not abolish the capacity of

the prosthetic DNA-cleaving group to induce strand scission in the presence of a reducing agent such as dithiothreitol. The relative geometrical requirements of the two moieties of the hybrids were satisfied. The linker between the two drug moieties seems to present the correct length and/or flexibility to ensure the proper binding to DNA of the two parts of the molecules. It is interesting to observe that the oligopyrrolecarboxamide moiety is located in the minor groove of the double helix while the appended phenazine chromophore behaves as a typical intercalating agent. Our study confirms that double-stranded DNA is highly adaptable and can accommodate a minor groove binding process in close proximity to a more invasive type of binding such as intercalation. Interestingly, we noticed that our previously synthesized anthraquinone-netropsin hybrid AQ(CN)Net (12), which differs from R-128 only by the nature of the heterocyclic nucleus (replacement of the 2-carboxamidophenazine-di-N-oxide by a 2-carboxamidoanthraquinone), was found not to present such a bimodal interaction with DNA. In this case, the netropsin element prevented intercalation of the anthraquinone nucleus. At first sight, we were surprised to find the phenoxazine-di-N-oxide moiety intercalated into DNA. Intercalation was not reported in the previous cleavage studies (13, 14). Recently, 2-aminophenazine-oligonucleotide conjugates have been synthesized (27, 28). The fluorescence studies suggested that the phenazine moiety could not intercalate or bind to the minor groove of DNA but interact with the outside of the double helix. However, we found in the literature evidences that structurally related phenazine derivatives (3-hydroxyethylamino-5phenazinium and a phenazine annelated to an imidazoglycoside) can intercalate into DNA (29). Nevertheless, it was proposed that the phenazines intercalate preferentially between G‚C base pairs compared to A‚T base pairs. Here, we observe that the tethered phenoxazinedi-N-oxide element intercalates into poly(dA-dT)‚poly(dAdT), but not into poly(dG-dC)‚poly(dG-dC). The sequence selectivity and the binding geometry are most likely imposed by the netropsin element. A similar molecular arrangement has been previously observed with oligopyrrolecarboxamide-anilinoacridine conjugates (30). It is manifest that the hybrid ligand R-128 binds preferentially to AT-rich tracts and cleaves DNA in the presence of a reducing agent. However, it does seem clear that there is no simple correlation between the binding data and the cutting profile for the drug. Sequence

226 Bioconjugate Chem., Vol. 11, No. 2, 2000

Helissey et al.

recognition of DNA by R-128 and cleavage of DNA likely correspond to two unrelated events. The cleavage of DNA proved to be nonspecific at low drug concentrations (e.g., 1 µM). At higher concentrations (g5 µM) the cleavage profile becomes clearly nonuniform, with regions of attenuated cleavage separated by sequences where the scission is much more pronounced. Interestingly, the most pronounced cleavage sites are separated one from another by approximately 10 base pairs (Figure 7). The cutting is more pronounced at certain sites, such as 3′ATC (positions 65-67) and 3′-ACT (positions 119-121), but on average a peak of cleavage appears every 9-10 base pairs. The reason for such a periodic cleavage remains obscure at present, but it is plausible to envisage that it is related to the helical periodicity of B-DNA. Perhaps the cleavage process is dictated by the structure of the double helix rather than the primary nucleotide sequence. Further experiments will be needed to verify this hypothesis. The conjugate which is moderately toxic to cancer cells complements the tool box of reagents which can be utilized to produce DNA strand scission. ACKNOWLEDGMENT

This work was supported by grants (to C. B.) from the Association pour la Recherche sur le Cancer and the Ligue Nationale Franc¸ aise Contre le Cancer and (to C. H. and P. C.) from the Actions de Recherches Concerte´es Contract 95/00-93. Support by the “convention INSERMCFB” is acknowledged. S. G. R. thanks Dr. S. Fermandjian for the use of the CD equipment and Dr. A. Deroussent for mass spectra recording. LITERATURE CITED

Figure 9. Cleavage of the DNA restriction fragment. The 265mer DNA fragment labeled at the 3′-end of the EcoRI site was reacted with graded concentrations of R-128 (µM as indicated) in the presence of 1 mM dithiothreitol overnight at 20 °C. The DNA was subsequently precipitated with cold ethanol and then electrophoresed on an 8% denaturating polyacrylamide gel. Lane marked control corresponds to the DNA subjected to the same treatment in the absence of drug but in the presence of DTT. Purine-specific sequence markers obtained by treatment of the DNA with formic acid-piperidine were run in the lane marked G+A. Numbers between on the left side of the gel refer to the nucleotide sequence of the DNA fragment as indicated in Figure 7.

(1) Zimmer, C., and Wa¨hnert, U. (1986) Nonintercalating DNAbinding ligands: specificity of the interaction and their use as tools in biophysical, biochemical and biological investigations of the genetic material. Prog. Biophys. Mol. Biol. 47, 31-112. (2) Kopka, M. L., and Larsen, T. A. (1992) Netropsin and the lexitropsins. The search for sequence-specific minor-groovebinding ligands. In Nucleic Acid Targeted Drug Design (C. L. Propst and T. J. Perun, Eds.) pp 303-374. Marcel Dekker, Inc., New York. (3) Zunino, F., Animati, F., and Capranico, G. (1995) DNA minor groove binding drugs. Curr. Pharm. Des. 1, 83-94. (4) Geierstanger, B. H., and Wemmer, D. E. (1995) Complexes of the minor groove of DNA. Annu. Rev. Biophys. Biomol. Struct. 24, 463-493. (5) Bailly, C. (1998) Sequence-specific recognition and modification of double-helical DNA by minor groove binding conjugates structurally related to netropsin and distamycin. In Advances in DNA Sequence Specific Agents (M. Palumbo, Ed.) Vol. 3, pp 97-156, JAI Press. Inc. (6) Broggini, M., Erba, E., Ponti, M., Ballinari, D., Geroni, C., Spreafico, F., and D’Incalci, M. (1991) Selective DNA interaction of the novel distamycin derivative FCE 24517. Cancer Res. 51, 199-204. (7) Pezzoni, G., Grandi, M., Biasoli, G., Capolongo, L., Ballinari, D., Giuliani, F. C., Barbieri, B., Pastori, A., Pesenti, E., and Mongelli, N. (1991) Biological profile of FCE 24517, a novel benzoyl mustard analogue of distamycin A. Br. J. Cancer 64, 1047-1050. (8) Bailly, C., and Chaires, J. B. (1998) Design of sequencespecific DNA minor groove binders. Netropsin and distamycin analogues. Bioconjugate Chem. 9, 513-538. (9) Herfeld, P., Helissey, P., Giorgi-Renault, S., Goulaouic, H., Pager, J., and Auclair, C. (1994) Poly(pyrrolecarboxamides) linked to photoactivable chromophore isoalloxazine. Synthesis, selective binding, and DNA cleaving properties. Bioconjugate Chem. 5, 67-76. (10) Bouziane, M., Ketterle´, C., Helissey, P., Herfeld, P., Le Bret, M., Giorgi-Renault, S., and Auclair, C. (1995) Sequence-

Netropsin-phenazine-di-N-oxide Conjugate directed single-strand cleavage of DNA by a netropsin-flavin hybrid molecule. Biochemistry 34, 14051-14058. (11) Ketterle´, C., Gabarro-Arpa, J., Ouali, M., Bouziane, M., Auclair, C., Helissey, P., Giorgi-Renault, S., and Le Bret, M. (1996) Binding of Net-Fla, a netropsin-flavin hybrid molecule, to DNA: Molecular mechanics and dynamics studies in vacuo and in water solution. J. Biomol. Struct. Dyn. 13, 963-977. (12) Helissey, P., Bailly, C., Vishwakarma, J. N., Auclair, C., Waring, M. J., and Giorgi-Renault, S. (1996) DNA minor groove cleaving agents: synthesis, DNA binding and DNA cleaving properties of anthraquinone-oligopyrrolecarboxamide hybrids. Anti-Cancer Drug Des. 11, 527-551. (13) Nagai, K., Carter, B. J., Xu, J., and Hecht, S. M. (1991) DNA cleavage by oxygen radicals produced in the absence of metal ions or light. J. Am. Chem. Soc. 113, 5099-5100. (14) Nagai, K., and Hecht, S. M. (1991) Site-specific DNA cleavage by antisense oligonucleotides covalently linked to phenazine di-N-oxide. J. Biol. Chem. 266, 23994-24002. (15) Kidani, Y. (1959) Studies on phenazines. Synthesis of phenazine-carboxylic acids. Chem. Pharm. Bull. 7, 88-90. (16) Houssier, C. (1981) Investigating nucleic acids, nucleoproteins, polynucleotides, and their interactions with small ligands by electrooptical systems. In Molecular Electrooptics (S. Krause, Ed.) pp 363-398, Plenum Publishing Corporation, New York. (17) Colson, P., Bailly, C., and Houssier, C. (1996) Electric linear dichroism as a new tool to study sequence preference in drug binding to DNA. Biophys. Chem. 58, 125-140. (18) Bailly, C., and Waring, M. J. (1995) Comparison of different footprinting methodologies for detecting binding sites for a small ligand on DNA. J. Biomol. Struct. Dyn. 12, 869-898. (19) Cory, A. H., Owen, T. C., Barltrop, J. A., and Cory, J. G. (1991) Use of an aqueous soluble tetrazolium/formazan assay for cell growth assays. Cancer Commun. 3, 207-212. (20) Ko¨gl, F., and To¨nnis, B. (1932) Uber Chlororaphin und Xanthoraphin, ein Beitrag zur Chemie der Chinhydrone. Liebigs Ann. 497, 265-289.

Bioconjugate Chem., Vol. 11, No. 2, 2000 227 (21) Vivian, D. L., Hartwell, J. L., and Waterman H. C. (1955) Phenazine syntheses. IV. Nitriles. J. Org. Chem. 20, 797802. (22) Rozum, Y. S. (1955) Some carbonyl derivatives of phenazine. J. Gen. Chem. 25, 583-587. (23) Maffei, S., and Pietra, S. (1952) La sintesi dell’acido 2-fenazincarbonico. Ann. Chim. 42, 519-525. (24) Bailly, C., and Waring, M. J. (1995) Transferring the purine 2-amino group from guanines to adenines in DNA changes the sequence-specific binding of antibiotics. Nucleic Acids Res. 23, 885-892. (25) Portugal, J., and Waring, M. J. (1987) Comparison of binding sites in DNA for berenil, netropsin and distamycin. A footprinting study. Eur. J. Biochem. 167, 281-289. (26) Plouvier, B., Bailly, C., Houssin, R., Rao, K. E., Lown, J. W., He´nichart, J.-P., and Waring, M. J. (1991) DNA-sequence specific recognition by a thiazole analogue of netropsin: a comparative footprinting study. Nucleic Acids Res. 19, 58215829. (27) Zamaratski, E., and Chattopadhyaya, J. (1998) Synthesis of phenazine-tethered arabino and xylofuranosyl oligonucleotide conjugates: the thermal stability and fluorescence properties of their duplexes (DNA-DNA & DNA-RNA) and triplexes. Tetrahedron 54, 8183-8206. (28) Ossipov, D., and Chattopadhyaya, J. (1998) Synthesis of 1′-phenazine-tethered psicofuranosyl oligonucleotide conjugates: the thermal stability and fluorescence properties of their duplexes and triplexes. Tetrahedron 54, 5667-5682. (29) Zozulya, V., Blagoi, Y., Lo¨ber, G., Voloshin, I., Winter, S., Makitruk, V., and Shalamay, A. (1997) Fluorescence and binding properties of phenazine derivatives in complexes with polynucleotides of various base compositions and secondary structures. Biophys. Chem. 65, 55-63. (30) Bourdouxhe-Housiaux, C., Colson, P., Houssier, C., Waring, M. J., and Bailly, C. (1996) Interaction of a DNA-threading netropsin-amsacrine combilexin with DNA and chromatin. Biochemistry 35, 4251-4264.

BC990131T