Site-Specific Cleavage of RNA and DNA by Complementary

Thomas Jefferson University, Philadelphia, Pennsylvania 19107. Received August 23, 2003. Bleomycin displays clinical chemotherapeutic activity, but is...
1 downloads 0 Views 231KB Size
Bioconjugate Chem. 2003, 14, 1307−1313

1307

Site-Specific Cleavage of RNA and DNA by Complementary DNA-Bleomycin A5 Conjugates Pavel E. Vorobjev,† Janet B. Smith,‡ Inna A. Pyshnaya,† Asya S. Levina,† Valentina F. Zarytova,† and Eric Wickstrom§,* Laboratory of Nucleic Acid Chemistry, Institute of Bioorganic Chemistry, Russian Academy of Sciences, Novosibirsk 630090, Russia, and Department of Microbiology and Immunology, and Department of Biochemistry and Molecular Pharmacology, Kimmel Cancer Center, and Cardeza Foundation for Hematologic Research, Thomas Jefferson University, Philadelphia, Pennsylvania 19107. Received August 23, 2003

Bleomycin displays clinical chemotherapeutic activity, but is so nonspecifically toxic that it is rarely administered. It was therefore of interest to determine whether bleomycin could be directed to cleave RNA or DNA at a specific site by conjugation to a complementary oligonucleotide. A 15 nt MYC complementary oligodeoxynucleotide (HMYC55) bearing a 5′ bleomycin A5 (Blm) residue was designed to base-pair with nt 7047-7061 of human MYC mRNA. Reactivity of the Blm-HMYC55 conjugate (and mismatch controls) with a MYC mRNA 30-mer, a MYC DNA 30-mer, and a MYC 2′-O-methyl RNA 30-mer, nt 7041-7070, was analyzed in 100 µM FeNH4SO4, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5, at 37 °C. Cleavage of the substrate RNA or DNA occurred primarily at the junction of the complementary DNA-target RNA duplex, 18-22 nt from the 5′ end of the RNA. Reaction products with lower mobility than the target RNA or DNA also formed. Little or no reaction was observed with more than three mismatches in a Blm-oligodeoxynucleotide conjugate. Neither the short RNA or DNA cleavage fragments nor the low mobility products were observed in the absence of Fe(II), or the presence of excess EDTA. The target RNA was also cleaved efficiently by bleomycin within a hybrid duplex with a preformed single-nucleotide bulge in the RNA strand. New Blmoligodeoxynucleotide conjugates containing long hexaethylene glycol phosphate based linkers between oligodeoxynucleotide and bleomycin were designed to target this bulge region. These conjugates achieved 8-18% cleavage of the target RNA, depending on the length of the linker. Blmoligodeoxynucleotide conjugates thus demonstrated sequence specificity and site specificity against RNA and DNA targets.

INTRODUCTION

The ability to turn off individual genes at will in growing cells provides a powerful tool for elucidating the role of a particular gene and for therapeutic intervention when that gene is overexpressed or mutated. Following the first suggestion that a complementary oligonucleotide could be prepared to target a naturally occurring nucleic acid sequence (1), an antisense oligonucleotide was first utilized successfully against Rous sarcoma virus (2). Since then, complementary oligonucleotides were first used to inhibit the expression of an activated oncogene, MYC, in human HL-60 promyelocytic leukemia cells (3). Antisense DNA inhibition has been applied to a wide variety of target genes in cells (4), animals (5), and humans (6). Chemically reactive oligonucleotide derivatives on the complementarity principle underlie the design of a new generation of antiviral and antitumor therapeutic drugs (7). In particular, the chain scission antibiotic bleomycin A5 has been conjugated to complementary oligonucleotides to allow site-directed cleavage of DNA (8). Bleo* Corresponding author. Voice: 215-955-4578; fax: 215-9554580. E-mail: [email protected]. † Russian Academy of Sciences. ‡ Department of Microbiology and Immunology, Thomas Jefferson University. § Department of Biochemistry and Molecular Pharmacology, Thomas Jefferson University.

mycin (Blm) is an antibiotic derived from the soil bacterium Streptomyces verticillus (9). The nonspecific cleavage of double stranded DNA or higher ordered tRNA by bleomycin results in antibacterial as well as antitumor activity (10). The bleomycins are glycopeptides with a molecular weight of approximately 1500 Da. Differences exist among bleomycin molecules due to differences in the composition of their terminal amines. There are four functional domains within the bleomycin structure (9) (Figure 1). Bleomycins are classified according to the different terminal amines attached to the bithiazole group at the C-terminus. It is the bithiazole residue and the C-terminal amine that are responsible for binding to DNA. Thus, the differences in the C-terminal amines in the various classes of bleomycin result in different binding affinities, with subsequent differences in cleavage efficiencies. Bleomycins contain metal binding domains at the N-terminus, and binding of a divalent metal cation is thought to be necessary for this domain to become functional. Although Fe(II) is one of the most efficient ions that serve this function, other divalent cations may also form a complex with the N-terminal domain (10). Cleavage of yeast tRNAPhe by free bleomycin has been observed, dependent on the secondary and tertiary structure, as well as the primary sequence, of the tRNA, resulting in cleavage at numerous positions; the percentage of cleavage at each site was quite low (10). It is known that some RNAs serving different biological functions are

10.1021/bc034148u CCC: $25.00 © 2003 American Chemical Society Published on Web 11/04/2003

1308 Bioconjugate Chem., Vol. 14, No. 6, 2003

Vorobjev et al.

model of B-cell lymphoma (12), in agreement with earlier observations of inhibition of human MYC expression in human HL60 cells by the corresponding HMYC55 antisense 3′-N-phosphoramidate (13). EXPERIMENTAL PROCEDURES

Figure 1. Bleomycin A5-oligonucleotide. The metal-binding domain (1) at the amino terminus of bleomycin is responsible for cleavage and also interacts with the nucleic acid target. The disaccharide domain (2) favors the binding of oxygen, which is necessary for oxidative cleavage of the nucleic acid. The structure of spacer region (3) plays a significant role in defining optimal conformation of the molecule, thus influencing the efficiency of DNA cleavage. The DNA-binding domain (4) at the C-terminal end of bleomycin forms a strong noncovalent complex with the nucleic acid.

efficiently cleaved by free bleomycin. In particular, the cleavage of 5S yeast ribosomal RNA occurs at U nucleotide residues in the 5′-GUA-3′ sequences preceding single-nucleotide bulges (11). Apparently, these regions of the RNA structure serve as sites of the preferential binding of bleomycin in the conformation favorable for selective oxidation of the ribose residue. We hypothesized that the cleavage activity of bleomycin could be targeted to a specific DNA or mRNA sequence by conjugation of bleomycin to the 5′ end of a complementary oligodeoxynucleotide. We also hypothesized that the cleavage of the target RNA could be increased significantly by formation of one-nucleotide bulge motifs in the RNA:antisense duplex. Previously we reported that a 15 nt antisense oligonucleotide, MMYC55, targeted to a C-terminal site in murine c-myc mRNA inhibited tumor growth in a mouse

Oligonucleotides. The oligonucleotide portion of BlmHMYC55 conjugate was designed to base-pair with the sense strand of the human MYC mRNA corresponding to those same nucleotides (Figure 2). The target sequence, nucleotides 7041-7070 (14), encodes 30 nucleotides encoding residues 382-391 of the C-terminal domain of Myc protein. Below the target RNA in Figure 2 are the three Blm-oligonucleotide reagents that were used in the experiments described. Three internal mismatches (underlined sequence) in HMYC56 disrupt continuous base-pairing to the target, allowing six base-pairs to form on either side of the mismatch zone. The random reagent includes 15 nucleotides that are unable to base-pair stably with the target RNA. We hypothesized that cleavage of the target RNA could occur at the junction of the 5′ end of the Blm-HMYC55 antisense reagent and the RNA target, 21 nucleotides from the 5′ end of the RNA target (blue arrow in Figure 2). Alternatively, we hypothesized that cleavage of the target RNA could occur at a single nucleotide bulge (blue arrow in Figure 3) created near the 5′ end of the RNA upon formation of bulged hybrid duplexes with 3′-Blm-linker-oligonucleotide 16-mers (Figure 3). Bleomycin A5 was obtained from the Institute of Organic Synthesis (Riga, Latvia). The MYC sense DNA and RNA 30-mer targets, DNA 15-mers HMYC55, HMYC56, HMYC 57, MMYC57 (Figure 2), and DNA 16mer B1, and its hexaethylene glycol-containing analogues (Figure 3) were synthesized by the phosphoramidite approach, then purified by ion exchange and reversedphase HPLC. The hexaethylene glycol phosphoramidite synthon for Blm-B2-B4 oligonucleotides was synthesized as described previously (15). Bleomycin A5 was conjugated with oligonucleotides as described (16, 17). Oligonucleotides were activated with a mixture of triphenylphosphine and 2,2′-dipyridyl disulfide in the presence of (dimethylamino)pyridine N-oxide. Bleomycin A5 was attached to the activated 5′-phosphate of oligonucleotides through the terminal spermidine amino group. Analytical denaturing gel electrophoresis of Blm-DNA conjugates revealed single bands. MALDI-TOF mass

Figure 2. MYC RNA: Blm-DNA schematic. a. MYC RNA target, 30 nucleotides derived from the human MYC exon 3, nucleotides 7041-7070 (14); b. antisense Blm-oligodeoxynucleotide 15-mer; c, d. mismatch Blm-oligodeoxynucleotide 15-mers; e. random Blmoligodeoxynucleotide 15-mer.

Figure 3. Bulged MYC RNA: Blm-DNA schematic. a. [5′-32P] MYC RNA target with bulge at U7; b. B1, antisense oligodeoxynucleotide 16-mer lacking complement for U7; c. Blm-Bn, antisense Blm-linker-oligodeoxynucleotide 16-mer with 2, 3, or 4 hexaethylene glycol phosphate linkers.

RNA and DNA Cleavage by Bleomycin−DNA

Bioconjugate Chem., Vol. 14, No. 6, 2003 1309

Figure 4. MYC RNA reactions with 10 µM Blm-HMYC55 vs 10 µM Blm-HMYC56. Each 10 µL reaction mixture also contained 1 µM MYC target RNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. Bleomycin reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2, and were incubated at 37 °C. Lane C: MYC RNA in reaction conditions for 120 min; lane OH-: 0.2 M NaHCO3, pH 8.9, 30 min. at 37 °C.

spectroscopy of the antisense Blm-linker-oligodeoxynucleotide 16-mer with 2, 3, or 4 hexaethylene glycol phosphate linkers (Figure 3c) was consistent with prediction: for n ) 2, calculated 6985.8 Da, measured 6982.6 Da; for n ) 3, calculated 7330.1 Da, measured 7330.7 Da; for n ) 4, calculated 7674.4 Da, measured 7688.0 Da. Concentrations of oligonucleotides were measured spectrophotometrically. Calculated molar extinction coefficients at 260 nm were 149 300/M‚cm for Blm-HMYC55, 156 000/M‚cm for Blm-HMYC56, 163 700/M‚cm for BlmHMYC57, 159 600/M‚cm for Blm-MMYC57, 157 400/M‚ cm for Blm-Bn, 305 500/M‚cm for 30-mer target DNA, and 307 700/M‚cm for 30-mer target RNA. The 5′-biotinyl MYC sense RNA 30-mer and homologous 5′-biotinyl DNA were obtained from CyberSyn (Aston PA) following synthesis by the phosphoramidite approach and purified by gel electrophoresis. PAGEpurified RNA and DNA were desalted on a C18 SepPak (Whatman). The MYC sense RNA and DNA 30-mer targets were also 5′-labeled with [γ-32P]ATP catalyzed by T4 polynucleotide kinase (20). Cleavage Reactions. Unless otherwise noted, cleavage reactions contained 1 µM target RNA or DNA, 10 µM Blm-DNA, 100 µM Fe(NH4)2(SO4)2, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The total volume of each reaction was adjusted to 10 µL with argon-saturated diethylpyrocarbonate-treated doubledeionized water. The reaction was incubated at 37 °C for 2 h, unless otherwise noted. In some cases, reaction mixtures were treated subsequently with base (0.25 M hydrazine‚HCl, pH 8.0, for RNA target, or 0.1 M nbutylamine, for DNA target) to hydrolyze unstable intermediate products and resolve low mobility complexes. RNA was incubated 30 min. at 37 °C in 0.2 M NaHCO3, pH 8.9, to generate alkaline hydrolysis ladders. Electrophoretic Analysis. Each reaction was terminated by addition of four volumes of gel loading buffer II (Ambion, Austin TX) (95% formamide, 18 mM EDTA, 0.025% NaDodSO4, 0.025% xylene cyanol FF, and 0.025% bromophenol blue) and stored at -20 °C until electrophoresis on a 20% denaturing polyacrylamide gel at 50 V/cm until the bromophenol blue dye reached 3/4 of the way into the gel. 5′-Biotinyl-RNA and DNA gels were electroblotted onto Zeta-probe nylon membrane (Bio-Rad) at 12 V for 33 min. The RNA or DNA was immobilized by UV-cross-linking using UV-transilluminator. Detection of the RNA and DNA, which were labeled at the 5′ end with biotin, was performed using the Ambion Bright-

Star BioDetect Kit. The resulting chemiluminescence was recorded onto Kodak XOMAT film with a typical exposure time of 15 min. Quantitative estimations were made using [32P]RNA and [32P]DNA targets. Products of cleavage were separated by electrophoresis on 20% denaturing polyacrylamide gel. Gels were dried on a Fisher Biotech FB GD 45 gel drier and autoradiographed with Kodak XOMAT film. Developed films were quantitated on a Kodak 440CF imaging system. RESULTS

Sequence Dependence of RNA Reactions. The time course of the reaction of MYC RNA 30-mer with Blm-HMYC55 vs Blm-HMYC56 was analyzed (Figure 4). The reactions were carried out as described in Materials and Methods above, initiated by the simultaneous addition of Fe(II) and β-mercaptoethanol. In the presence of conjugated Blm-HMYC55, cleavage of MYC RNA was observed, 18-22 nucleotides from the 5′ end of the RNA. These sites correlated with the positioning of bleomycin by the antisense portion of the reagent (blue arrow in Figure 2). The same bands were observed in the presence of Blm-HMYC56 (3 nt mismatch) at much lower intensities. The reactions with Blm-HMYC55, but not BlmHMYC56, also produced products that migrated more slowly than the original RNA target, suggesting an additional reaction mechanism more complex than cleavage. No reaction products were observed in the presence of HMYC55 plus free Blm, or Blm conjugated to a noncomplementary MMYC57 15-mer (not shown). Treatment of reacted RNA samples with 0.25 M hydrazine‚HCl, pH 8.0, for 30 min. at 37 °C did not change the band intensities (not shown). Sequence Dependence of DNA Reactions. When a MYC DNA 30-mer was also reacted with BlmHMYC55, and the time course compared for both the RNA and DNA targets, the same analogous low mobility products were seen, along with 18-22 nt DNA fragments (Figure 5). The cleavage product bands reached a plateau early in the 50 min incubation, but the slowly moving products increased linearly throughout the course of the reaction. Lower levels of the reaction products were also observed, in the case of the DNA target (Figure 6), with Blm-HMYC56 (3 mismatches), Blm-HMYC57 (2 × 2 mismatches), but not with random Blm-MMYC57 (12 mismatches). Blm-MMYC57 produced a 5 nt DNA fragment which can be observed also in reaction with free

1310 Bioconjugate Chem., Vol. 14, No. 6, 2003

Vorobjev et al.

Figure 5. MYC RNA and MYC DNA reactions with 10 µM Blm-HMYC55. Each 10 µL reaction mixture also contained 1 µM MYC target RNA 30-mer or DNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2. Lane C: MYC target in reaction conditions for 50 min.; lane OH-: 0.2 M NaHCO3, Pu, Py: Maxam-Gilbert reactions at purine and pyrimidine residues, respectively.

Figure 6. MYC DNA reactions with 10 µM Blm-DNA conjugates. Each 10 µL reaction mixture also contained 1 µM MYC target DNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2, and were incubated at 37 °C for 2 h. Lanes 1, 7: MYC DNA in reaction conditions; lanes 2, 8: Blm-HMYC55; lanes 3, 9: Blm-HMYC56; lanes 4, 10: Blm-HMYC57; lanes 5, 11: Blm-MMYC57; lanes 6, 12: Blm. Before analysis, the reactions in lanes 7-12 were treated with 0.1 M butylamine, 8 min., 95 °C. Lanes Pu, Py: Maxam-Gilbert reactions at purine and pyrimidine residues, respectively.

Figure 7. MYC RNA and DNA reactions with varying concentrations of Blm-HMYC55. Each 10 µL reaction mixture also contained 1 µM MYC target RNA 30-mer or DNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2, and were incubated at 37 °C for 50 min. Lane OH-: 0.2 M NaHCO3; lanes Pu, Py: Maxam-Gilbert reactions at purine and pyrimidine residues, respectively.

bleomycin. Generally, the alkaline treatment reveals additional products of DNA cleavage by bleomycin. Samples of 2 h. DNA reactions were treated with 0.1 M n-butylamine for 8 min. at 95 °C before electrophoresis (Figure 6). Treatment with n-butylamine eliminated the slowly moving bands and increased the intensities of the cleavage bands.

Concentration Dependence. To examine the dependence of product formation on Blm-HMYC55 concentration, a ramp of concentrations from 0.1 µM to 10 µM, reacting with 1 µM MYC RNA or DNA, was studied (Figure 7). This experiment revealed that as the BlmHMYC55 concentrations increased, the extent of cleavage products and slowly migrating products after 50 min also

RNA and DNA Cleavage by Bleomycin−DNA

Figure 8. MYC [32P]RNA and [32P]DNA reactions with BlmHMYC55. Each 10 µL reaction mixture also contained 1 µM MYC target RNA 30-mer or DNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 mM Fe(NH4)2(SO4)2, and were incubated at 37 °C for 50 min. Lane 1: MYC DNA in reaction conditions; lane 2: 1 µM of Blm-HMYC55 added; lane 3: 10 µM of Blm-HMYC55 added; lane 4: MYC RNA in reaction conditions; lane 5: 10 µM of Blm-HMYC55 added.

increased, reaching a plateau at 5 µM for the RNA target, and 1 µM for the DNA target. However, product yield remained low; in no case was all target RNA or DNA consumed. Quantitative estimations were made using [32P]RNA and [32P]DNA targets. For both targets the amount of slowly moving products did not exceed 3% (Figure 8). The maximum amount of rapidly moving products formed with the RNA target by 10 µM BlmHMYC55 was two times higher (6%). Under the same conditions, 25% of the DNA target was cleaved. Control reactions extended to 7 h revealed no slowly migrating products without bleomycin, with 1 µM free bleomycin, with 1 µM Blm-HMYC56, with 1 µM Blm-HMYC55 but without mercaptoethanol, or with a mixture of 1 µM free bleomycin and 1 µM HMYC55 (not shown). Fe(II) Dependence. The appearance of tRNAPhe cleavage products in the presence of free bleomycin,

Bioconjugate Chem., Vol. 14, No. 6, 2003 1311

without free Fe(II) (10), invited a similar experiment with Blm-HMYC55 (Figure 9). Preincubation of RNA or DNA with 10 mM EDTA prior to adding Blm-DNA, without any subsequent addition of Fe(II), did not result in any of the slowly moving products or specific cleavage products seen before. 2′-OH Dependence. Considering the cleavage of RNA by Blm-HMYC55, one possible mechanism characteristic of RNA is self-hydrolysis by a 2′-OH, mediated by some general base moiety of adjacent bleomycin. Substitution of the MYC RNA target with MYC 2′-O-methyl RNA, suggested by Dr. Serge Beaucage, provided a simple test of that model. Without free Fe(II), no products were observed upon incubation of 1 µM MYC 2′-O-methyl RNA with Blm-HMYC55. Both rapidly migrating and slowly migrating products were observed in the presence of Fe(II) (Figure 10). Cleavage of the RNA by Free Bleomycin at a Bulged Site. In the case of imperfect complementation between the 16-base oligdeoxynucleotide B1 (Figure 3) and the [32P]RNA target, an imperfect duplex with a onenucleotide bulge at U7 in the RNA chain was presumably formed. Efficient cleavage of the [32P]RNA target by free bleomycin, 30% in 2 h, was observed at U6 adjacent to the presumed bulge at U7 (Figure 11). Cleavage of the RNA by Bleomycin-DNA Conjugate at a Bulged Site. Analogous bulged duplexes of [32P]RNA with Blm-(hexaethylene glycol phosphate)nB1 DNA conjugates were studied. For the bleomycin residue that was attached to the 3’-terminus of B1 to reach the desired cleavage site, flexible linkers consisting of two, three, or four residues of hexaethylene glycol phosphate, respectively, were included in the Blm-DNA conjugates (Figure 3). Molecular modeling suggested that the lengths of all three linkers are larger than the distance between the 3′-terminal phosphate group of oligonucleotide B1 in conjugates Blm-B2, Blm-B3, and Blm-B4 and the cleaved U6 residue in the RNA target involved in the hybrid duplexes. In the presence of conjugates Blm-B2, Blm-B3, and Blm-B4, cleavage of the RNA target occurred at U6 (Figure 11). The efficiency of cleavage in the complex depended on the length of the linker between the bleo-

Figure 9. MYC RNA and DNA reactions with 10 µM Blm-DNA conjugates preincubated with EDTA. Each 10 µL reaction mixture also contained 1 µM MYC target RNA (lanes 1-6) or DNA (lanes 7-12) 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5, 10 mM EDTA. Reactions were initiated by the addition of the Blm-DNA conjugate, and were incubated at 37 °C for 2 h. Lane 1: MYC RNA, lane 7: MYC DNA in reaction conditions; lanes 2, 8: Blm-HMYC55; lanes 3, 9: Blm-HMYC56; lanes 4, 10: Blm-HMYC57; lanes 5, 11: Blm-MMYC57; lanes 6, 12: Blm; lane OH-: 0.2 M NaHCO3; lanes Pu, Py: Maxam-Gilbert reactions at purine and pyrimidine residues, respectively.

1312 Bioconjugate Chem., Vol. 14, No. 6, 2003

Vorobjev et al.

Figure 10. MYC 2′-O-methyl RNA reactions with increasing concentrations of Blm-HMYC55. Each 10 µL reaction mixture contained 1 µM MYC target 2′-O-methyl RNA 30-mer, Blm-HMYC55 at the concentrations shown, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2, and were incubated at 37 °C for 2 h.

Figure 11. MYC [32P]RNA reactions with 10 µM Blm(hexaethylene glycol phosphate)n-B1 DNA conjugates that induce a bulge in the RNA at U7. Each 10 µL reaction mixture also contained 1 µM MYC target RNA 30-mer, 50 mM β-mercaptoethanol, 200 mM LiCl, 10 mM Tris-HCl, pH 7.5. The reactions were initiated by the addition of 100 µM Fe(NH4)2(SO4)2, and were incubated at 37 °C for 2 h. Lane 0: 10 µM free bleomycin in the presence of 10 µM B1 oligonucleotide; OH-: 0.2 M NaHCO3.

mycin residue and the oligonucleotide: 8% with BlmB2, 12% with Blm-B3, and 18% with Blm-B4. DISCUSSION

These results reveal that targeted cleavage of RNA or DNA by antisense-conjugated bleomycin is possible. Bleomycin conjugated to HMYC55 cleaved the complementary MYC RNA or DNA target at the junction between the DNA:RNA duplex and the RNA single strand, the site to which the oligodeoxynucleotide portion of the reagent directed the bleomycin molecule (Figure 2). It was essential for the antisense oligodeoxynucleotide to be conjugated to bleomycin and fully base-paired in order for cleavage of the RNA target to occur (Figures 4, 6). In addition, reaction of MYC RNA or DNA with BlmHMYC55 yielded slowly migrating products. Considering the extent of formation of the slowly migrating products, it would be valuable to identify their composition and structure. Mild treatment of reacted RNA samples with hydrazine‚HCl did not convert much of the slowly migrating bands to high mobility RNA fragments. On the other hand, more vigorous treatment of reacted DNA

samples with n-butylamine fully converted slowly migrating bands to high mobility cleavage fragments. The slowly migrating products included less than 1% of total band intensity in those reactions that included the Blm-HMYC56 3-mismatch control, and none were detected in the presence of the Blm-HMYC57 4-mismatch control, Blm-random, or free Blm. Reactions in the absence of Fe(II) (Figure 9) did not result in any of the slowly moving products or cleavage products, implying the absence of an Fe(II)-independent cleavage mechanism. One of the conceivable routes for the formation of such cross-links is the interaction of primary or secondary amines on bleomycin with base labile lesions, which are formed on DNA or RNA as a consequence of oxidation by activated bleomycin (18). The low extent of target RNA or DNA cleavage by Blm-HMYC55 might reflect a requirement for a rarely occurring conformation. In light of these considerations, it is possible that greater reactivity might result from the use of short tandem oligonucleotides (19), or some other oligonucleotide design, instead of a single completely complementary 15-nt oligonucleotide. The experiments with Blm-(hexaethylene glycol phosphate)n-B1 DNA conjugates provided an interesting test of the latter suggestion. Considerably greater cleavage of the target RNA was achieved using conjugates that induce a single nucleotide bulge in the RNA strand of the hybrid duplex. The increase in the extent of cleavage with the lengthening of the linker probably indicates that a particular conformation is favored for efficient cleavage. Further investigations will be necessary to determine whether complementary bleomycin-oligodeoxynucleotide conjugates exhibit greater antisense efficacy in cells than do complementary oligodeoxynucleotides alone and whether bleomycin-oligodeoxynucleotide conjugates are less toxic than free bleomycin in normal cells. ACKNOWLEDGMENT

We thank Dr. Serge Beaucage for suggesting the 2′O-methyl RNA experiment. This work was supported by RFBR grant 99-04-49731a and 02-04-49597a to V.F.Z., Fogarty International Research Collaboration Award TW01094 to V.F.Z. and E.W., and NIH grant CA42960 to E.W. LITERATURE CITED (1) Belikova, A. M., Zarytova, V. F., and Grineva, N. I. (1967) Tetrahedron Lett. 37, 3557-62.

RNA and DNA Cleavage by Bleomycin−DNA (2) Zamecnik, P. C., and Stephenson, M. L. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, 280-4. (3) Wickstrom, E. L., Bacon, T. A., Gonzalez, A., Freeman, D. L., Lyman, G. H., and Wickstrom, E. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 1028-32. (4) Wickstrom, E. (1991) Prospects for Antisense Nucleic Acid Therapy of Cancer and AIDS, Wiley-Liss, New York. (5) Agrawal, S. (1996) Antisense Therapeutics, Humana Press, Totowa, NJ. (6) Wickstrom, E. (1998) Clinical Trials of Genetic Therapy with Antisense DNA and DNA Vectors, Marcel Dekker, New York. (7) Knorre, D. G., Vlassov, V. V., Zarytova, V. F., Lebedev, A. V., and Fedorova, O. S. (1994) Design and Targeted Reactions of Oligonucleotide Derivatives, CRC Press, Boca Raton, FL. (8) Sergeyev, D. S., and Zarytova, V. F. (1996) Russian Chemical Reviews 65, 355-378. (9) Umezawa, H. (1976) Prog. Biochem. Pharmacol. 11, 1827. (10) Keck, M. V., and Hecht, S. M. (1995) Biochemistry 34, 12029-37. (11) Holmes, C. E., and Hecht, S. M. (1993) J. Biol. Chem. 268, 25909-13.

Bioconjugate Chem., Vol. 14, No. 6, 2003 1313 (12) Smith, J. B., and Wickstrom, E. (1998) J. Natl. Cancer Inst. 90, 1146-54. (13) Gryaznov, S., Skorski, T., Cucco, C., Nieborowska-Skorska, M., Chiu, C. Y., Lloyd, D., Chen, J. K., Koziolkiewicz, M., and Calabretta, B. (1996) Nucleic Acids Res. 24, 1508-14. (14) Gazin, C., Dupont de Dinechin, S., Hampe, A., Masson, J. M., Martin, P., Stehelin, D., and Galibert, F. (1984) EMBO J. 3, 383-7. (15) Pyshnaya, I. A., Pyshnyi, D. V., Ivanova, E. M., Zarytova, V. F., Bonora, G. M., Scalfi-Happ, C., and Seliger, H. (1998) Nucleosides Nucleotides 17, 1289-1297. (16) Zarytova, V. F., Sergeyev, D. S., and Godovikova, T. S. (1993) Bioconjugate Chem. 4, 189-93. (17) Sergeyev, D. S., Godovikova, T. S., and Zarytova, V. F. (1995) Nucleic Acids Res. 23, 4400-6. (18) Gavin, I. M., Melnik, S. M., Yurina, N. P., Khabarova, M. I., and Bavykin, S. G. (1998) Anal. Biochem. 263, 26-30. (19) Sergeyev, D. S., Vorobjev, P. E., and Zarytova, V. F. (1997) Nucleosides Nucleotides 16, 1575-1577. (20) Berkner, K. L., and Folk, W. R. (1977) J. Biol. Chem. 252, 3176-84.

BC034148U