766
Bioconjugate Chem. 2008, 19, 766–777
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides Kaisa Ketomäki and Pasi Virta* Department of Chemistry, University of Turku, FIN-20014 Turku, Finland. Received November 20, 2007; Revised Manuscript Received January 4, 2008
Aminoglycoside conjugates of 2′-O-methyl oligoribonucleotides have been synthesized entirely on a solid phase using conventional phosphoramidate chemistry. For this purpose, appropriately protected neamine-derived phosphoramidites, viz., 2-cyanoethyl [6,3′,4′-tri-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl)neamine-5-O-ethyl] N,N-diisopropylphosphoramidite, 1, and 2-cyanoethyl [6,3′,4′,2″,3″-penta-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl) ribostamycin-5″-yl] N,N-diisopropylphosphoramidite, 2, have been prepared and attached via phosphodiester linkage to an appropriate 2′-O-methyl oligoribonucleotide. Levulinoyl esters are used to cap the hydroxyl groups of the aminoglycoside moieties, since they may be selectively removed prior to ammonolysis. In this manner, the potential OfN acyl migration is excluded. Applicability of the strategy has been demonstrated by the synthesis of eight different aminoglycoside conjugates, in which 1 and 2 are attached directly to the 5′-end (6 and 10) or, alternatively, to an inserted non-nucleosidic hydroxyalkyl armed branching unit (3, 4, or 5), which results in intrachain conjugates (7–9, 11–13). The potential of these conjugates to act as a sequence-selective artificial nuclease has been studied.
INTRODUCTION Aminoglycosides are well-known antibiotics having affinity for structural elements of RNA in which the usual base pairing is disrupted. Their bactericidal effect is based on binding to the ribosomal decoding site (A-site), which reduces the fidelity of protein translation (1–4), but they additionally have comparable affinities to other structurally related RNA targets. Aminoglycosides are, e.g., able to inhibit several ribozymes (5, 6) and, perhaps most interesting for the general public, they bind to important HIV regions, TAR,1RRE (7, 8), and DIS (9, 10), resulting in inhibition of viral replication. However, the known natural aminoglycosides are very diverse RNA binders, which is mainly a consequence of their polycationic nature. To overcome this promiscuity, numerous tailored aminoglycosides have been synthesized in recent years. One approach is conjugation of aminoglycosides to nucleobases (9, 11–14), oligonucleotides (15), or PNA (16), in which affinity to numerous RNA targets may be whittled down by increased sequence selectivity. In addition to the binding properties discussed above, aminoglycosides show potential for chemical cleavage of RNA. This feature is common to diamines and polyamines in general (17, 18), but with aminoglycosides, the more defined spatial positioning of the amino groups, their individual pKa values, and lowered conformational freedom may further accelerate the hydrolysis. As an example, neomycin hydrolyzes simple adenylyl(3′-5′) adenosine more effectively than an unstructured diamine (19). This metal-ion-independent catalytic activity expectedly remains at a marginal level, but remarkable rate acceleration may be obtained by participation of metal *
[email protected]. 1 Abbreviations: All, Allyl; Boc, tert-butoxycarbonyl; CPG, controlled pore glass; DCA, dichloroacetic acid; DMAP, 4-dimethylaminopyridine; DIS, dimerization initiation site; DMTr, 4,4′-dimethoxytrityl; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; iPr, 1-methylethyl; Lev, levulinoyl (4-oxopentanoyl); MMTr, 4-methoxytrityl; RRE, revresponse element, TAR, trans-acting response element, TBDMS, tertbutyldimethylsilyl; Tfa, trifluoroacetyl, TMS, tetramethylsilane, TMSTf, trimethylsilyl triflate.
cations. Cu2+-complexes of neamine and kanamycin have been reported to be efficient cleaving agents of RNA (20), and even of DNA (21). These observations give an insight that both requirements (the RNA recognition and the cleavage) for an artificial sequence-specific nuclease may already exist in aminoglycosides (22–25), but conjugation to a complementary oligonucleotide or PNA may be a respectable option. A PNA-neamine conjugate has been reported not only to bind to the HIV-1 TAR region, but also to catalyze its hydrolysis (26, 27). The conjugate results in relatively fast metal-ion-independent cleavage, enhances cellular uptake, and inhibits viral replication. With the above goals in mind, aminoglycoside-conjugated oligonucleotides are an attractive subject in organic chemistry, but from the synthetic point of view, they may still need some consideration. Ligation of those two classes of biomolecules in solution has advantages, such as asimple protecting group scheme, but it may be time-consuming to carry out. Fully automated synthesis may, hence, be a more attractive choice. In that case, however, several amino and hydroxyl groups of an appropriate aminoglycoside have to be masked with easily removable protecting groups, which should be compatible with the standard oligonucleotide synthesis and, preferably, be baselabile. Furthermore, the amino groups should not be exposed when the adjacent hydroxyl groups still bear acyl-type protection, since OfN-acyl migration may take place. Base-modified 2′-deoxyuridine phosphoramidites, conjugated with 2,6-diaminoglucose (28) and neomycine units (29) have been used for the synthesis of aminoglycoside-modified oligonucleotides. The aminoglycoside moieties have been capped with protecting group combinations, such as acetyl esters with trifluoroacetamides and trifluoroacetylesters with Boc-carbamates. On using the former combination (28), a substantial amount of acetamide byproducts have been found in the product mixture. This has been attributed to acetyl capping of the trifluoroacetamide functions upon oligonucleotide assembly, but the possibility of acyl migration has not been excluded. The latter combination (29) overcomes acyl migration, but the strategy suffers from potential depurination, when expanded to adenine- and guaninecontaining oligonucleotides.
10.1021/bc7004279 CCC: $40.75 2008 American Chemical Society Published on Web 02/19/2008
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides
In the present study, an entirely solid phase synthesis of aminoglycoside conjugates of 2′-O-methyl oligoribonucleotides (6–13) using standard phosphoramidate chemistry has been described. Appropriately tethered neamine-based phosphoramidites, 1 and 2, were first synthesized using cheap neomycin trisulfate as an initial starting material. The resulted phosphoramidites 1 and 2 were then attached to resin-bound oligonucleotides via a phosphodiester linkage. Inserted non-nucleosidic branching units (3–5) were utilized to obtain intrachain conjugates (7–9, 11–13), or alternatively, direct coupling to the 5′end was performed (6 and 10). Prior to ammonolysis, the resins were treated with a mixture of hydrazinium acetate to orthogonally remove the levulinoyl protections. In this manner, acyl migration to the adjacent amino groups on the aminoglycoside moieties (1 and 2) was eliminated. Applicability of the strategy was demonstrated by the synthesis of eight different conjugates (6–13), which all were obtained with acceptable 41–65% purity. A strategy to obtain intrachain conjugates (7–9, 11–13) was devised to be expanded to the library synthesis, in which various conjugates may readily be obtained by mixing different building blocks, the aminoglycosides (1 and 2) with the branching units (3–5), at an appropriate site in the oligonucleotide sequence. Since a neamine PNA conjugate has been reported to be an efficient artificial nuclease (26, 27), a small library of aminoglycosides (6–13) bearing the native negatively charged phosphodiester backbone may advisably give further information about the general nuclease activity of these attractive biomolecules. In the present study, conjugates (6–13) were targeted to chimeric ribo/2′-O-methyl oligonucleotides (14 and 15), which upon hybridization (excluding experiments with 6 and 10) constitute triand pentanucleotide bulges (A3 and A5) opposite the aminoglycosidic regions. One of the conjugates (9) resulted in metal-ion-independent cleavage of the targets [14 and 15, k ) (2.7 ( 0.6) × 10-7 s-1 and (2.3 ( 0.6) × 10-7 s-1, respectively], but usually the cleavage obtained by the conjugates (6–8, 10–13) remained at a marginal level or the conjugates were fully inactive.
EXPERIMENTAL PROCEDURES General Methods and Materials. The NMR spectra of 1–3, 19–31, and 33–36 were recorded at 400 and 500 MHz. The chemical shifts are given in ppm from internal TMS and coupling constants in hertz. Peak assignments were supported by appropriate 2D NMR methods, e.g., COSY, HSQC, and HMBC. The mass spectra of 1–3, 6–13, 20–31, and 33–36 were recorded by EI and ESI ionization methods. The 2′-O-methyl oligoribonucleotide conjugates (6–13) and the chimeric target oligonucleotides (14 and 15) were assembled by conventional phosphoramidate strategy using a 1 µmol scale and following the standard RNA coupling protocol of Appplied Biosystems 342 DNA synthesizer. LiChroCART Hypersil ODS column (250 × 10 mm, 5 µm) at a flow rate of 3 mL min-1 was used for RP HPLC purifications of the 2′-O-methyl oligoribonucleotide conjugates (6–13). In each case, a gradient elution from 20% to 70% acetonitrile in 0.1 mol L-1 Et3N+OAc-buffer over 30 min was used. 6,3′,4′-Tri-O-acetyl-5-O-allyl-1,3,2′,6′-tetraazido Neamine (20). 6,3′,4′-tri-O-acetyl-1,3,2′,6′-tetraazido neamine (19) was synthesized from neomycine trisulfate (16) according to the literature (30–32). Improvements to the selective acylation of 18 may be mentioned. Acetic anhydride (0.50 mL, 51 mmol) in pyridine (2.0 mL) was dropwise added to a mixture of 18 (0.68 g, 16 mmol) and a catalytic amount of DMAP in pyridine (1.0 mL). The reaction was allowed to stir overnight at ambient temperature, quenched by addition of methanol, and evaporated to dryness. The residue was purified by silica gel chromatography (40% EtOAc in petroleum ether) to yield 0.64 g (73%)
Bioconjugate Chem., Vol. 19, No. 3, 2008 767
of the product (19) as white foam. All side products, mainly 5,6,3′,4′-tetra-O-acetyl-1,3,2′,6′-tetraazido neamine (0.16 g, 17%), were also isolated, combined, and restored to 18 (acetyl groups were removed by 0.1 mol L-1 NaOMe in MeOH, 1 h at 25 °C, cf. synthesis of 26 below). The regenerated crude 18 may again be used for the selective acylation, with the same yield. 19: 1H NMR (500 MHz, CDCl3) δ 5.50 (dd, 1H, J ) 10.3 Hz, 9.5 Hz), 5.37 (d, 1H, J ) 3.6 Hz), 5.06 (dd, 1H, J ) 10.0 Hz, 9.5 Hz), 4.95 (dd, 1H, J ) 10.0 Hz, 9.9 Hz), 4.35 (m, 1H), 3.70–3.66 (m, 2H), 3.62 (d, 1H, J ) 3.6 Hz), 3.55 (ddd, 1H, J ) 12.5 Hz, 10.1 Hz, 4.6 Hz), 3.47–3.37 (m, 3H), 3.34 (dd, 1H, J ) 13.5 Hz, 5.1 Hz), 2.42 (ddd, 1H, J ) 13.3 Hz, 4.5 Hz, 4.5 Hz), 2.19, 2.11, and 2.07 (each s, each 3H), 1.64 (ddd, 1H, J ) 12.9 Hz, 12.9 Hz, 12.9 Hz); 13C NMR (CDCl3, 250 MHz) δ 170.5, 170.2, 169.7, 98.8, 83.7, 75.0, 74.4, 71.2, 69.5, 69.3, 61.7, 58.3, 57.9, 50.8, 31.9, 20.8, 20.7, 20.6. A mixture of 19 (1.1 g, 20 mmol), Pd(OAc)2 (9 mg, 0.04 mmol), Ph3P (53 mg, 0.20 mmol), and allyl methyl carbonate (0.46 mL, 41 mmol) in THF (5.0 mL) was refluxed for 15 min and evaporated to dryness (33). The residue was purified by silica gel chromatography (5% EtOAc in CH2Cl2) to yield 0.50 g (42%) of the desired product 20 as white foam and 0.21 g (17%) of 6-Oallyl-1,3,2′,6′-tetraazido-5,3′,4′-tri-O-acetyl neamine as yellowish oil. 20: 1H NMR (500 MHz, CDCl3) δ 5.88 (m, 1H), 5.61 (d, 1H, J ) 3.9 Hz), 5.50 (dd, 1H, J ) 10.6 Hz, 9.4 Hz), 5.26 (m, 1H), 5.17 (m, 1H), 5.04 (dd, 1H, J ) 10.0 Hz, 9.5 Hz), 5.02 (dd, 1H, J ) 10.0 Hz, 9.9 Hz), 4.51 (m, 1H), 4.34 (m, 1H), 4.17 (m, 1H), 3.63 (dd, 1H, J ) 9.6 Hz, 9.6 Hz), 3.51 (dd, 1H, J ) 9.5 Hz, 9.4 Hz), 3.50–3.45 (m, 2H), 3.41–3.36 (m, 2H), 3.31 (dd, 1H, J ) 13.5 Hz, 5.1 Hz), 2.40 (ddd, 1H, J ) 13.3 Hz, 4.6 Hz, 4.6 Hz), 1.63 (ddd, J ) 12.7 Hz, 12.7 Hz, 12.7 Hz); 13C NMR (CDCl3, 250 MHz) δ 170.1, 169.7, 169.6, 133.7, 117.0, 97.6, 82.2, 78.0, 75.5, 74.2, 70.2, 69.3, 69.3, 60.9, 58.7, 58.0, 50.8, 50.8, 31.7, 20.9, 20.7, 20.7; HRMS (ESI) [M + Na]+ C21H28N12NaO9 requires 615.1994, found 615.2005. 6-O-allyl1,3,2′,6′-tetraazido-5,3′,4′-tri-O-acetyl neamine: 1H NMR (500 MHz, CDCl3) δ 5.60 (m, 1H), 5.16 (dd, 1H, J ) 10.8 Hz, 9.2 Hz), 5.01–4.91 (m, 2H), 4.92 (d, 1H, J ) 3.9 Hz), 4.88 (dd, 1H, J ) 9.7 Hz, 9.7 Hz), 4.75 (dd, 1H, J ) 10.1 Hz, 9.4 Hz), 4.19 (ddd, 1H, J ) 10.3 Hz, 5.0 Hz, 2.7 Hz), 4.01 (m, 1H), 3.82 (m, 1H), 3.30 (m, 1H), 3.28 (dd, 1H, J ) 9.6 Hz, 4.4 Hz), 3.22 (ddd, 1H, J ) 12.1 Hz, 9.9 Hz, 4.4 Hz), 3.10 (dd, 1H, J ) 13.5 Hz, 2.7 Hz), 3.07 (dd, 1H, J ) 10.8 Hz, 3.9 Hz), 3.02 (dd, 1H, J ) 13.5 Hz, 5.0 Hz), 2.93 (dd, 1H, J ) 9.8 Hz, 9.8 Hz), 2.11 (ddd, 1H, J ) 13.4 Hz, 4.6 Hz, 4.6 Hz), 1.86, 1.80, 1.78 (each s, each 3H), 1.23 (ddd, 1H, J ) 12.4 Hz, each); 13C NMR (CDCl3, 250 MHz) δ 169.5, 169.5, 168.7, 133.3, 117.7, 98.3, 81.6, 78.3, 74.1, 74.0, 69.4, 69.2, 69.1, 60.3, 59.2, 58.3, 50.4, 31.7, 20.7, 20.4, 20.4; HRMS (ESI) [M + Na]+: C21H28N12NaO9 requires 615.1994, found 615.1966. 1,3,2′,6′-Tetraazido-5-O-(2-hydroxyethyl)neamine (21). Osmium tetroxide (2.5 wt % solution in 2-methyl-2-propanol, 0.20 mL, 16 µmol) was added to a mixture of 20 (0.50 g, 0.84 mmol) in water-dioxane (1:4 v/v, 5.0 mL) (34). After 1 h stirring, the mixture was cooled to 0 °C, the solvent volume was increased to 20 mL, and NaIO4 (0.36 g, 17 mmol) was slowly added. The reaction was then allowed to warm up, stirred at room temperature for 5 h, and the resulted aldehyde intermediate was then extracted with ethyl acetate (3 × 10 mL). The organic layers were combined, dried with Na2SO4, and evaporated to dryness. The residue was dissolved in a mixture of dichloromethane-ethanol (1:2, v/v, 15 mL), and NaBH4 (64 mg, 1.7 mmol) was added. After 5 h stirring, the reaction was quenched by addition of saturated ammonium chloride, and the crude product intermediate was extracted with ethyl acetate (3 × 10 mL). The organic layers were combined, dried with Na2SO4, and evaporated to dryness. The residue was dissolved in 0.1
768 Bioconjugate Chem., Vol. 19, No. 3, 2008
mol L-1 methanolic sodium methoxide (5.0 mL). The mixture was stirred at ambient temperature for 1 h, neutralized by addition of strongly acidic cation-exchange resin, and filtered. The filtrate was evaporated to dryness and purified by silica gel chromatography (40% EtOAc in CH2Cl2) to yield 0.21 g (53%) of the product (21) as a colorless oil. 1H NMR (500 MHz, CDCl3) δ 5.59 (d, 1H, J ) 3.7 Hz), 4.24 (m, 1H), 4.04–3.96 (m, 2H), 3.90 (dd, 1H, J ) 9.9 Hz, 9.5 Hz), 3.78 (ddd, 1H, J ) 11.6 Hz, 7.5 Hz, 3.1 Hz), 3.72 (ddd, 1H, J ) 11.5 Hz, 3.7 Hz, 3.4 Hz), 3.55–3.33 (m, 8H), 3.16 (dd, 1H, J ) 10.6 Hz, 3.8 Hz), 2.26 (ddd, 1H, J ) 12.4 Hz, 4.2 Hz, 4.2 Hz), 1.42 (ddd, 1H, J ) 12.2 Hz, 12.2 Hz, 12.2 Hz); 13C NMR (CDCl3, 250 MHz) δ 97.9, 85.8, 77.6, 76.9, 74.4, 71.8, 71.2, 70.8, 62.9, 61.7, 60.0, 59.5, 51.2, 31.8; HRMS (ESI) [M + Na]+ C14H22N12NaO7 requires 493.1627, found 493.1615. 1,3,2′,6′-Tetraazido-5-O-[2-(4-methoxytrityloxy)ethyl]neamine (22). Compound 21 (0.20 g, 0.43 mmol) was dissolved in pyridine (2.0 mL), and 4-methoxytrityl chloride (0.20 g, 0.64 mmol) was added to the mixture. Reaction mixture was stirred overnight at ambient temperature, quenched by addition of methanol, and evaporated to dryness. The residue was dissolved in ethyl acetate and washed with saturated NaHCO3. The organic layer was separated, dried with Na2SO4, and evaporated to dryness. The residue was purified by silica gel chromatography (20% EtOAc in CH2Cl2) to yield 0.32 g (100%) of the product 22 as a white foam. 1H NMR (500 MHz, CDCl3): δ 7.47 (d, 4H, J ) 7.6 Hz), 7.35–7.29 (m, 6H), 7.23 (dd, 2H, J ) 7.4 Hz, 7.3 Hz), 6.87 (d, 2H, J ) 8.9 Hz), 5.94 (d, 1H, J ) 3.8 Hz), 4.25 (m, 1H), 4.15 (m, 1H), 4.02 (m, 1H), 3.91 (dd, 1H, J ) 10.3 Hz, 10.1 Hz), 3.79 (s, 3H), 3.63 (dd, 1H, J ) 9.6 Hz, 9.5 Hz), 3.57–3.33 (m, 9H), 3.27 (m, 1H), 2.94 (dd, 1H, J ) 10.5 Hz, 3.8 Hz), 2.23 (ddd, 1H, J ) 12.3 Hz, 4.3 Hz, 4.3 Hz), 1.39 (ddd, 1H, J ) 12.3 Hz, each). 13C NMR (CDCl3, 250 MHz) δ 158.8, 144.6, 144.5, 135.5, 130.3, 128.4, 127.4, 126.6, 112.7, 97.4, 86.6, 85.2, 77.1, 76.8, 72.1, 71.8, 71.3, 70.8, 63.6, 62.9, 60.5, 60.2, 59.7, 54.3, 51.3, 31.7; HRMS (ESI) [M + Na]+ C34H38N12NaO8 requires 765.2828, found 765.2804. 6,3′,4′-Tri-O-levulinoyl-5-O-[2-(4-methoxytrityloxy)ethyl]N1,N3,N2′,N6′-tetra(trifluoroacetyl) Neamine (23). Trimethyl phosphine (1 mol L-1 Me3P in toluene, 1.41 mL, 14 mmol) was added to a mixture of 22 (0.21 g, 0.28 mmol) in water-dioxane (1:4, v/v, 2.5 mL) under nitrogen. The reaction was stirred for 4 h at ambient temperature, and then concentrated ammonia (1.0 mL) was added. After overnight stirring, volatiles were removed, and the residue was dissolved in dry methanol (2.0 mL). Triethylamine (1.0 mL) and freshly distilled methyl trifluoroacetate (0.23 mL,) were then added to the mixture, the reaction was stirred overnight, and evaporated to dryness. The resulted crude trifluoroacetylated neamine derivative was dissolved in ethyl acetate and washed with saturated NaHCO3. The organic layer was separated, dried with Na2SO4, and evaporated to dryness. The residue was dissolved in pyridine (2.0 mL), and then levulinic anhydride and a catalytic amount of DMAP were added to the mixture. After 2 h stirring, the reaction was quenched by addition of methanol. Ethyl acetate was added, and the crude product was washed with saturated NaHCO3. The organic layer was separated, dried with Na2SO4, and evaporated to dryness. The crude final product was purified by silica gel chromatography (20% petroleum ether in EtOAc) to yield 0.28 g (74%) of 23 as a white foam. 1H NMR (500 MHz, CDCl3) δ 8.30 (t, 1H, J ) 5.5 Hz), 7.75 (d, 1H, J ) 9.3 Hz), 7.46 (d, 1H, J ) 9.4 Hz), 7.42 (d, 4H, J ) 7.5 Hz), 7.32–7.2 (m, 9H), 7.14 (d, 1H, J ) 7.8 Hz), 6.85 (d, 2H, J ) 9.0 Hz), 5.51 (d, 1H, J ) 3.3 Hz), 5.13 (dd, 1H, J ) 10.3 Hz, 9.9 Hz), 5.00 (dd, 1H, J ) 10.1 Hz, 9.9 Hz), 4.98 (dd, 1H, J ) 9.8 Hz, 9.8 Hz), 4.23–4.07 (m, 3H), 4.01 (ddd, 1H, J ) 10.1 Hz, 4.5 Hz, 4.5 Hz), 3.91 (dd, 1H, J ) 9.9 Hz, 9.6 Hz), 3.86 (m, 1H), 3.81 (s,
Virta and Ketomäki
3H), 3.74 (dd, 1H, J ) 9.2 Hz, 9.2 Hz), 3.73 (m, 1H), 3.60–3.48 (m, 2H), 3.29 (m, 2H), 2.93 (ddd, 1H, J ) 18.9 Hz, 9.5 Hz, 4.2 Hz), 2.77–2.45 (m, 8H), 2.42–2.22 (m, 4H), 2.18 (s, 3H), 2.11 (s, 3H), 2.09 (s, 3H), 1.87 (ddd, 1H, J ) 12.7 Hz, each); 13C NMR (CDCl3, 250 MHz) δ 208.5, 206.5, 206.1, 173.7, 172.3, 172.2, 158.6, 144.3, 144.3, 135.4, 130.5, 128.5, 127.7, 126.9, 112.1, 96.5, 86.7, 81.3, 76.8, 75.1, 72.5, 70.0, 69.9, 67.9, 63.0, 55.2, 52.7, 49.4, 48.6, 41.7, 38.0, 37.8, 37.7, 31.8, 29.6, 29.5, 29.4, 28.0, 27.7, 27.6 (Tfa related peaks not listed); HRMS (ESI) [M + Na]+ C57H60F12N4NaO18 requires 1339.3603, found 1339.3653. 5-O-(2-Hydroxyethyl)-6,3′,4′-tri-O-levulinoyl-N1,N3,N2′,N6′tetra(trifluoroacetyl)neamine (24). Compound 23 was dissolved in a mixture of dichloroacetic acid in dichloromethane (3:97, v/v, 6.0 mL). The reaction was stirred for 1 h at ambient temperature, quenched by addition of methanol, and then saturated NaHCO3was added to the mixture. The resulted organic layer was separated, dried with Na2SO4, and evaporated to dryness. The residue was purified via silica gel chromatography (4% MeOH in CH2Cl2) to yield 0.21 g (97%) of the product (24) as white foam. 1H NMR (500 MHz, CDCl3) δ 8.74 (d, 1H, J ) 6.6 Hz), 8.18–8.15 (b, 2H), 7.62 (d, 1H, J ) 9.1 Hz), 5.38 (d, 1H, J ) 3.4 Hz), 5.26 (dd, 1H, J ) 10.4 Hz, 10.0 Hz), 5.20 (dd, 1H, J ) 9.9 Hz, 9.8 Hz), 4.96 (dd, 1H, J ) 9.9 Hz, 9.8 Hz), 4.37 (m, 1H), 4.24–4.06 (m, 5H), 3.99 (dd, 1H, J ) 9.6 Hz, 9.1 Hz), 3.90 (m, 1H), 3.78–3.75 (m, 3H), 3.60 (m, 1H), 3.43 (m, 1H), 2.84–2.37 (m, 12H), 2.17 (s, 3H), 2.15 (s, 3H), 2.14 (s, 3H), 2.00 (m, 1H), 1.42 (m, 1H); 13C NMR (CDCl3, 250 MHz) δ 208.9, 207.2, 207.1, 174.3, 172.9, 172.0, 97.9, 81.8, 78.9, 77.6, 74.2, 69.8, 69.2, 67.9, 62.0, 53.0, 50.7, 48.7, 47.9, 38.6, 37.7, 37.7, 37.4, 31.6, 29.6, 29.6, 29.4, 28.0, 27.9, 27.8 (Tfa related peaks not listed); HRMS (ESI) [M + Na]+ C37H44F12N4NaO17 requires 1067.2402, found 1067.2416. 2-Cyanoethyl [6,3′,4′-tri-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl)neamine-5-O-ethyl] N,N-diisopropylphosphoramidite (1). 2-Cyanoethyl N,N-diisopropylphosphonamidic chloride (53 µL, 0.24 mmol) was added to a mixture of 24 (0.19 g, 0.18 mmol) and triethylamine (0.13 mL, 0.90 mmol) in dichloromethane (2.5 mL) under nitrogen. The mixture was stirred for 1 h at ambient temperature and then subjected directly to a silica gel column. Elution with 2% triethylamine and 20% petroleum ether in ethyl acetate yielded 0.21 g (93%) of the product (1) as white foam. 1H NMR (500 MHz, CDCl3) δ 8.16 (t, 0.56H, J ) 5.4 Hz), 8.11 (t, 0.44H, J ) 5.5 Hz), 7.87 (b, 1H), 7.45 (d, 0.44H, J ) 10.0 Hz), 7.42 (d, 0.56H, J ) 9.6 Hz), 7.11 (b, 1H), 5.59 (d, 0.56H, J ) 3.2 Hz), 5.55 (d, 0.44H, J ) 3.4 Hz), 5.22 and 5.21 (both dd, sum 1H, J ) 10.7 Hz, 2.5 Hz, and J ) 10.5 Hz, 2.4 Hz), 5.10 (dd, 0.56H, J ) 9.9 Hz, 9.8), 5.05 (dd, 0.44H, J ) 9.9 Hz, 9.8 Hz), 4.99 and 4.97 (both dd, sum 1H, J ) 10.1 Hz, 3.9 Hz, and J ) 10.1 Hz, 4.3 Hz), 4.40–4.32 (m, 1H), 4.24–4.06 (m, 3H), 3.95–3.71 (m, 8H), 3.64–3.55 (m, 4H), 3.44–3.39 (m, 1H), 2.88 (ddd, 1H, J ) 18.8 Hz, 8.9 Hz, 4.5 Hz), 2.79–2.41 (m, 14H), 2.29–2.25 (m, 1H), 2.19 and 2.18 (both s, sum 3H), 2.17 and 2.16 (both s, sum 3H), 2.13 (s, 3H), 1.94–1.86 (m, 1H), 1.20, 1.19, 1.18, 1.17, and 1.15 (each s, sum 12H); 13C NMR (CDCl3, 250 MHz) δ 208.0, 207.9, 206.7, 206.7, 206.4, 206.3, 173.58, 173.56, 172.38, 172.36, 172.2, 172.1, 118.25, 118.22, 96.9, 81.38, 81.34, 75.0, 74.9, 77.4, 75.0, 74.9, 72.7, 72.7, 72.9, 72.8, 70.1, 69.7, 69.5, 68.0, 67.9, 62.6, 62.5, 62.3, 62.2, 58.5, 58.3, 58.3, 58.2, 52.8, 52.8, 49.3, 48.8, 43.0, 42.9, 41.2, 38.0, 37.9, 37.7, 31.6, 29.6, 29.5, 29.5, 28.1, 27.7, 27.6, 27.6, 24.7, 24.6, 24.6, 24.6, 24.5, 24.5, 24.5, 21.0, 20.3, 20.3 (Tfa related peaks not listed); 31P NMR (CDCl3, 200 MHz) δ 149.2, 148.2; HRMS (ESI) [M + Na]+ C46H61F12N6NaO18P requires 1267.3481, found 1267.3432.
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides
1,3,2′,6′-Tetraazido-5″-O-(4-methoxytrityl)ribostamycin (26). 6,3′,4′, 2″, 3″, 5″-Hexa-O-acetyl-1,3,2′,6′-tetraazidoribostamycin (23) (35) was first synthesized from the neamine derivative (19). 19 (0.30 g, 0.54 mmol) and 1,2,3,5-tetra-O-acetyl-β-D-ribofuranose (0.35 g, 1.1 mmol) were dissolved in dry dichloromethane (2.0 mL). The mixture was cooled to 0 °C, trimethylsilyl trifluoromethanesulfonate (0.20 mL, 1.1 mmol) was slowly added, and then the reaction was allowed to stir for 3 h under nitrogen. Saturated NaHCO3 was added to the mixture, and the resulted organic phase was separated, dried with Na2SO2, and evaporated to dryness. The residue was purified by silica gel chromatography (from 30% to 40% EtOAc in petroleum ether) to yield 0.26 g (59%) of the product (25) as white foam. 1H NMR (500 MHz, CDCl3): δ 5.91 (d, 1H, J ) 3.8 Hz), 5.46 (dd, 1H, J ) 10.7 Hz, 9.3 Hz), 5.43 (d, 1H, J ) 3.9 Hz), 5.20 (dd, 1H, J ) 4.9 Hz, 4.8 Hz), 5.10 (dd, 1H, J ) 4.9 Hz, 4.1 Hz), 5.02 (dd, 1H, J ) 10.0 Hz, 9.6 Hz), 4.97 (dd, 1H, J ) 9.9 Hz, 9.9 Hz), 4.49 (ddd, 1H, J ) 10.1 Hz, 5.3 Hz, 2.7 Hz), 4.44 (dd, 1H, J ) 12.2 Hz, 3.1 Hz), 4.26 (m, 1H), 4.16 (dd, 1H, J ) 12.2 Hz, 3.6 Hz), 3.88 (dd, 1H, J ) 9.3 Hz, 9.1 Hz), 3.71 (dd, 1H, J ) 9.8 Hz, 9.0 Hz), 3.54 (ddd, 1H, J ) 12.5 Hz, 10.0 Hz, 4.6 Hz), 3.45 (ddd, 1H, J ) 12.5 Hz, 10.2 Hz, 4.4 Hz), 3.38 (dd, 1H, J ) 13.5 Hz, 2.8 Hz), 3.32 (dd, 1H, J ) 13.5 Hz, 5.4 Hz), 3.24 (dd, 1H, J ) 10.7 Hz, 3.8 Hz), 2.40 (ddd, 1H, J ) 13.3 Hz, 4.6 Hz, 4.6 Hz), 2.23, 2.12, 2.11, 2.10, 2.07, 2.07 (each s, each 3H), 1.63 (ddd, 1H, J ) 13.0 Hz each); 13C NMR (CDCl3, 250 MHz) δ 170.6, 170.1, 169.7, 169.6, 169.5, 169.5, 106.3, 96.4, 80.9, 79.7, 76.4, 74.7, 74.1, 70.9, 70.0, 69.2, 69.2, 63.1, 61.0, 59.1, 58.1, 51.0, 31.6, 20.9, 20.7, 20.7, 20.7, 20.5, 20.4. Compound 25 (0.98 g, 1.2 mmol) was dissolved in 0.1 mol L-1 methanolic sodium methoxide (10 mL). The mixture was stirred at ambient temperature for 1 h, neutralized by addition of strongly acidic cation-exchange resin, filtered, and evaporated to dryness. The residue was dissolved in pyridine (3.0 mL), and then a mixture of 4-methoxytrityl chloride (0.49 g, 1.6 mmol) in pyridine (3.0 mL) was slowly added. The reaction was allowed to stir overnight at ambient temperature, quenched by addition of methanol, and evaporated to dryness. The residue was purified by silica gel chromatography (5% MeOH in CH2Cl2) to yield 0.71 g (71%) of the product (26) as white foam. 1H NMR (500 MHz, CDCl3): δ 7.50 (d, 4H, J ) 7.6 Hz), 7.37–7.30 (m, 6H), 7.23 (m, 1H), 6.89 (d, 2H, J ) 8.9 Hz), 5.87 (d, 1H, J ) 3.8 Hz), 5.35 (d, 1H, J ) 1.5 Hz), 4.26–4.22 (m, 2H), 4.16 (ddd, 1H, J ) 9.9 Hz, 5.7 Hz, 2.0 Hz), 4.11 (m, 1H), 3.83 (dd, 1H, J ) 10.4 Hz, 9.1 Hz), 3.80 (s, 3H), 3.67–3.61 (m, 2H), 3.53–3.39 (m, 5H), 3.33–3.28 (m, 4H), 2.92 (dd, 1H, J ) 10.4 Hz, 3.9 Hz, 3.8 Hz), 2.22 (m, 1H), 1.38 (ddd, 1H, J ) 12.2 Hz, each); 13C NMR (CDCl3, 250 MHz) δ 158.8, 144.6, 144.5, 135.3, 130.4, 128.3, 128.3, 127.5, 127.5, 126.5, 126.5, 112.8, 109.3, 96.7, 86.4, 84.3, 82.2, 76.3, 76.0, 75.3, 71.5, 71.3, 71.2, 71.0, 64.4, 63.6, 60.3, 59.7, 54.4, 51.3, 31.7; HRMS (ESI) [M + Na]+ C37H42N12NaO11 requires 853.2988, found 853.2982. 6,3′,4′, 2″, 3″-Penta-O-levulinoyl-5″-O-(4-methoxytrityl)N1,N3,N2′,N6′-tetra(trifluoroacetyl)ribostamycin (27). Compound 27 was synthesized from 26 using the three-step procedure, as described for 23, but 10 molar equiv levulinic anhydride was used. After silica gel chromatography (2% MeOH in CH2Cl2), product (27) was obtained as white foam in a 79% yield (0.45 g). 1H NMR (500 MHz, CDCl3) δ 7.98 (t, 1H, J ) 5.4 Hz), 7.68 (d, 1H, J ) 8.5 Hz), 7.43–7.40 (m, 5H), 7.32–7.21 (m, 9H), 6.88 (d, 2H, J ) 9.0 Hz), 5.80 (b, 1H), 5.36 (d, 1H, J ) 4.1 Hz), 5.37–5.32 (m, 2H), 5.12 (dd, 1H, J ) 10.3 Hz, 9.9 Hz), 5.00 (dd, 1H, J ) 10.2 Hz, 8.6 Hz), 4.93 (dd, 1H, J ) 9.9 Hz, 9.8 Hz), 4.26–4.15 (m, 2H), 4.09 (m, 1H), 4.05–3.98 (m, 4H), 3.81 (s, 3H), 3.46–3.35 (m, 2H), 3.29 (dd, 1H, J ) 10.5 Hz, 4.0 Hz), 3.24 (dd, 1H, J ) 10.5 Hz, 3.1 Hz), 2.89
Bioconjugate Chem., Vol. 19, No. 3, 2008 769
(ddd, 1H, J ) 18.9 Hz, 9.0 Hz, 4.6 Hz), 2.83–2.30 (m, 21H), 2.21, 2.20, 2.17, 2.11, 2.09 (each s, each 3H), 1.89 (ddd, J ) 12.7 Hz each); 13C NMR (CDCl3, 250 MHz) δ 208.3, 206.8, 206.8, 206.6, 206.2, 174.6, 172.1, 171.8, 171.8, 171.7, 158.7, 144.1, 144.0, 134.9, 130.7, 130.6, 128.6, 127.8, 127.8, 127.0, 127.0, 113.2, 107.4, 95.5, 86.7, 80.7, 74.7, 73.5, 71.5, 70.4, 69.4, 67.7, 62.7, 55.2, 53.5, 51.9, 49.9, 48.6, 41.7, 38.2, 38.0, 37.8, 37.7, 37.6, 31.4, 29.8, 29.7, 29.7, 29.4, 29.3, 27.9, 27.8, 27.6, 27.5, 27.5 (Tfa related peaks not listed); HRMS (ESI) [M + Na]+ C70H76F12N4NaO25 requires 1623.4499, found 1623.4434. 6,3′,4′, 2″, 3″-Penta-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl)ribostamycin (28). Compound 28 was synthesized from 27 as described for 24 above. After silica gel chromatography (4% MeOH in CH2Cl2), product 28 was obtained as white foam in a 91% yield (0.34 g). 1H NMR (500 MHz, CDCl3) δ 8.10–8.06 (m, 2H), 7.74 (b, 1H), 7.50 (b, 1H), 5.61 (b, 1H), 5.27 (b, 1H), 5.21–5.17 (m, 2H), 5.12 (m, 1H), 4.99 (dd, 1H, J ) 9.5 Hz, 9.5 Hz), 4.94 (dd, 1H, J ) 9.8 Hz, 9.8 Hz), 4.39 (ddd, 1H, J ) 10.4 Hz, 10.0 Hz, 3.8 Hz), 4.26–4.15 (m, 2H), 4.07–4.03 (m, 2H), 4.00 (dd, 1H, J ) 8.7 Hz, 8.6 Hz), 3.91 (dd, 1H, J ) 9.1 Hz, 8.9 Hz), 3.78 (m, 1H), 3.65 (m, 1H), 3.56 (m, 1H), 3.41–3.38 (m, 2H), 2.88–2.40 (m, 21H), 2.20, 2.19, 2.16, 2.14, 2.13 (each s, each 3H), 1.79 (ddd, 1H, J ) 12.7 Hz, each); 13C NMR (CDCl3, 250 MHz) δ 207.9, 207.3, 206.9, 206.8, 206.7, 173.8, 172.4, 172.2, 172.1, 172.0, 107.2, 96.4, 81.8, 80.1, 78.0, 74.6, 73.7, 69.8, 69.7, 69.5, 68.5, 60.7, 52.1, 49.9, 48.6, 40.4, 38.1, 37.8, 37.7, 37.6, 31.6, 29.7, 29.6, 29.6, 29.5, 29.3, 28.0, 27.7, 27.6, 27.6, 27.4 (Tfa related peaks not listed); HRMS (ESI) [M + Na]+ C50H60F12N4NaO24 requires 1351.3298, found 1351.3300. 2-Cyanoethyl[6,3′,4′,2″,3″-penta-O-levulinoyl-N1,N3,N2′,N6′tetra(trifluoroacetyl)ribostamycin-5″-yl] N,N-diisopropylphosphoramidite (2). Compound 2 was synthesized from 28 as described for 1 above. After silica gel chromatography (2% Et3N in EtOAc), product 2 was obtained as white foam in a 92% yield (0.35 g). 1H NMR (500 MHz, CDCl3): δ 8.49 (b, 0.54H), 8.25 (t, 0.46H, J ) 5.9 Hz), 7.64 (d, 0.46H, J ) 9.3 Hz), 7.53 (d, 0.54H, J ) 9.3 Hz), 7.26 (d, 0.46H, J ) 9.3 Hz), 7.21 (d, 0.54H, J ) 9.3 Hz), 7.05 (m, 1H), 6.00 (d, 0.54H, J ) 3.1), 5.90 (d, 0.46H, J ) 3.5 Hz), 5.34 (d, 0.54H, J ) 4.7 Hz), 5.33 (d, 0.46H, J ) 4.6 Hz), 5.29–5.25 (m, 1H), 5.18–4.95 (m, 4H), 4.44 (ddd, 0.46H, J ) 10.2 Hz, 10.2 Hz, 3.8 Hz), 4.38 (ddd, 0.56H, J ) 10.1 Hz, 10.1 Hz, 3.9 Hz), 4.34–4.25 (m, 1H), 4.18–3.79 (m, 7H), 3.72–3.41 (m, 6H), 2.98–2.32 (m, 23H), 2.21, 2.21, 2.20, 2.19, 2.18, 2.17, 2.14, 2.14 (each s, 12H), 1.86–1.76 (m, 1H), 1.22, 1.21, 1.21, 1.20, 1.18 (each s, 12H); 13 C NMR (CDCl3, 250 MHz) δ 208.6, 208.2, 206.7, 206.7, 206.4, 206.4, 206.3, 206.2, 174.3, 174.2, 172.5, 172.2, 171.8, 171.7, 171.7, 171.7, 171.6, 118.8, 118.4, 106.9, 106.5, 95.5, 95.1, 81.4, 81.4, 75.5, 75.3, 74.6, 74.4, 74.1, 74.1, 71.0, 70.8, 70.8, 70.3, 69.8, 66.8, 61.8, 61.6, 58.5, 58.4, 51.9, 51.8, 49.7, 49.6, 48.0, 47.9, 43.0, 42.9, 38.2, 38.0, 38.0, 37.8, 37.7, 37.6, 31.8, 31.6, 29.8, 29.8, 29.8, 29.7, 29.7, 29.6, 29.4, 28.1, 28.0, 27.7, 27.6, 27.5, 27.3, 24.7, 24.6, 24.6, 20.4, 20.3, 20.1, 20.0 (Tfa related peaks not listed); 31P NMR (CDCl3,200 MHz) δ 148.9 and 148.7; HRMS (ESI) [M + Na]+ C59H77F12N6NaO25P requires 1551.4377, found 1551.4329. 6,3′,4′-Tri-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl)neamine (30). N1,N3,N2′,N6′-tetra(trifluoroacetyl)neamine (29) was first synthesized from neamine tetrahydrochloride (17) according to the literature (28). Levulinic anhydride (13.3 g, 62 mmol) was slowly added to a mixture of 29 (14.6 g, 21 mmol) and a catalytic amount of DMAP in pyridine (60 mL). The mixture was stirred overnight at ambient temperature and evaporated to dryness. The residue was dissolved in ethyl acetate and washed with saturated NaHCO3. The organic layer was separated, dried with Na2SO4, and evaporated to dryness. The
770 Bioconjugate Chem., Vol. 19, No. 3, 2008
crude product was purified by silica gel chromatography (50% EtOAc in CH2Cl2) to yield 13.8 g (67%) of 30 as white foam. 1 H NMR (500 MHz, CDCl3) δ 8.39 (d, 1H, J ) 8.3 Hz), 8.09 (b, 1H), 7.69 (d, 1H, J ) 8.4 Hz), 7.65 (d, 1H, J ) 7.9 Hz), 5.37 (d, 1H, J ) 3.6 Hz), 5.27 (dd, 1H, J ) 10.4 Hz, 9.9 Hz), 5.16 (dd, 1H, J ) 10.1 Hz, 10.1 Hz), 4.95 (dd, 1H, J ) 10.0 Hz, 9.9 Hz), 4.29 (m, 1H), 4.21 (dd, 1H, J ) 9.7 Hz, 9.6 Hz), 4.12–4.05 (m, 2H), 3.97–3.94 (m, 2H), 3.83 (ddd, 1H, J ) 9.4 Hz, 9.3 Hz, 3.2 Hz), 3.48 (m, 2H), 2.92 (ddd, 1H, J ) 19.0 Hz, 9.0 Hz, 4.5 Hz), 2.87–2.40 (m, 12H), 2.34 (ddd, 1H, J ) 13.0 Hz, each), 2.20 (s, 3H), 2.19 (s, 3H), 2.18 (m, 1H), 2.14 (s, 3H); 13C NMR (CDCl3, 250 MHz) δ 209.2, 208.4, 206.8, 174.1, 172.7, 172.5, 98.1, 80.0, 75.4, 74.6, 70.3, 69.6, 68.4, 52.9, 49.9, 48.4, 40.4, 38.2, 38.1, 37.7, 30.3, 29.6, 29.5, 29.4, 28.5, 27.9, 27.7 (Tfa related peaks not listed); HRMS (ESI) [M + Na]+ requires 512.1106, found 512.1122. 2-Cyanoethyl [6,3′,4′-tri-O-levulinoyl-N1,N3,N2′,N6′-tetra(trifluoroacetyl)neamine-5-yl] N,N-diisopropylphosphoramidite (31). Compound 31 was synthesized from 30 as described for 1, but 3.6 molar equiv 2-cyanoethyl N,Ndiisopropylphosphonamidic chloride and a prolonged reaction time (3 h) were required. After silica gel chromatography (1% Et3N and 30% petroleum ether in EtOAc), product 31 was obtained as white foam in a 90% yield (0.54 g). 1H NMR (500 MHz, CDCl3) δ7.81 (d, 0.55H, J ) 8.8 Hz), 7.76 (d, 0.45H, J ) 9.2 Hz), 7.70–7.66 (m, 1H), 7.56 (d, 0.45H, J ) 9.7 Hz, 7.50–7.48 (m, 1H), 7.41 (d, 0.55H, J ) 9.5 Hz), 5.81 (d, 0.55H, J ) 3.7 Hz), 5.45 (b d, 0.45H, J ) 2.9 Hz), 5.26 (dd, 0.45H, J ) 10.4 Hz, 9.1 Hz), 5.22 (dd, 0.55H, J ) 10.2 Hz, 9.4 Hz), 5.11–4.97 (m, 2H), 4.47–4.39 (m, 1H), 4.31–3.34 (m, 11H), 2.80–2.36 (m, 14H), 2.15–2.08 (m, 1H), 2.18, 2.10, 2.10 (each s, sum 9H), 1.97–1.87 (m, 1H); 13C NMR (CDCl3, 250 MHz) δ 207.0, 206.9, 206.6, 206.6, 206.4, 206.3, 172.6, 172.5, 172.1, 172.0, 172.0, 171.9, 118.6, 118.3, 96.3, 94.6, 78.8, 76.5, 76.3, 75.0, 74.7, 73.5, 69.9, 69.8, 68.9, 68.7, 68.6, 68.0, 58.9, 58.8, 58.7, 58.6, 43.2, 43.1, 39.5, 39.5, 37.4, 37.3, 37.3, 31.1, 30.7, 28.8, 28.8, 28.7, 28.6, 28.6, 28.1, 28.1, 27.7, 27.6, 27.6, 23.9, 23.9, 23.8, 23.8, 23.8, 23.7, 23.6, 22.2, 22.2, 22.2, 19.8, 19.8, 19.7, 19.7 (Tfa related peaks not listed); 31P NMR (CDCl3, 200 MHz) δ 150.8, 150.5; HRMS (ESI) [M + Na]+ C44H57F12N6NaO17P requires 1223.3218, found 1223.3191. Ethyl 2-{(2R,3S,5R)-3-tert-butyldimethylsilyloxy-5-[(4,4′dimethoxytrityl)oxymethyl] Tetrahydrofuran-2-yl}acetate (33). Compound 32 was synthesized as described earlier (36, 37). tert-Butyldimethylsilyl chloride (1.43 g, 9.5 mmol) was added to a mixture of 32 (2.4 g, 4.7 mmol) and a catalytic amount of imidazole in pyridine (20 mL). The reaction was stirred overnight at 40 °C, quenched by addition of methanol, and evaporated to dryness. The residue was dissolved in dichloromethane and washed by saturated NaHCO3. The organic phase was separated, dried with Na2SO4, and evaporated to dryness. The product was purified by silica gel chromatography (0.1% Et3N and 3% MeOH in CH2Cl2) to yield 2.9 g (60%) of 33 as a colorless oil. 1H NMR (400 MHz, CDCl3) δ 7.47–7.44 (m, 2H), 7.37–7.17 (m, 7H), 6.84–6.82 (m, 4H), 4.56 (m, 1H), 4.18 (m, 1H), 4.17 (m, 2H), 3.93 (m, 1H), 3.08 (m, 2H), 2.68 (dd, 1H, J ) 15.3 Hz, 7.0 Hz), 2.52 (dd, 1H, J ) 15.3 Hz, 6.2 Hz), 1.96 (ddd, 1H, J ) 12.6 Hz, 5.4 Hz, 2.2 Hz), 1.72 (ddd, 1H, J ) 12.6 Hz, 9.8 Hz, 5.8 Hz), 1.27 (t, 3H, J ) 7.3 Hz), 0.86 (s, 9H), 0.07 (s, 3H), 0.02 (s, 3H); 13C NMR (CDCl3,100 MHz) δ 171.2, 158.4, 144.9, 136.2, 136.2, 130.1, 128.2, 127.7, 126.6, 113.0, 86.6, 85.9, 74.6, 74.2, 64.1, 60.4, 55.2, 41.1, 40.8, 25.8, 14.2, -4.7, -4.8; MS (EI) M+ requires 620.3, found 620. (2R,3S,5S)-3-tert-Butyldimethylsilyloxy-2-[(4,4′-dimethoxytrityl)oxymethyl]-5-(2-hydroxyethyl)tetrahydrofuran (34). Lithium aluminum hydride (0.16 g, 4.3 mmol) was slowly added to a cooled (0 °C) mixture of 33 in diethyl ether (20 mL). The
Virta and Ketomäki
mixture was allowed to warm to room temperature, and then the stirring was continued overnight at room temperature. The excess hydride was destroyed by careful addition of water, and then the crude product was extracted with dichloromethane (3 × 20 mL). The organic layers were combined, dried with Na2SO4, and evaporated to dryness. The residue was purified by silica gel chromatography (0.1% Et3N and 3% MeOH in CH2Cl2) to yield 1.4 g (83%) of 34 as yellowish oil. 1H NMR (400 MHz, CDCl3) δ 7.50 (d, 2H, J ) 8.1 Hz), 7.40–7.24 (m, 7H), 6.88 (d, 4H, J ) 8.8 Hz), 4.38 (m, 1H), 4.28 (m, 1H), 3.95 (m, 1H), 3.87 (m, 2H), 3.84 (s, 6H), 3.14 (d, 2H, J ) 4.7 Hz), 2.75 (b, 1H), 1.98–1.69 (m, 4H), 0.90 (s, 9H), 0.06 (s, 3H), 0.04 (s, 3H); 13C NMR (CDCl3, 100 MHz) δ 158.4, 144.9, 136.1, 136.1, 130.0, 128.2, 127.7, 126.7, 113.1, 86.9, 86.0, 78.6, 74.0, 64.1, 61.6, 55.2, 41.6, 37.3, 25.8, -4.7, -4.8; MS (EI) M+ requires 578.3, found 578. (2R,3S,5S)-3-tert-Butyldimethylsilyloxy-2-[(4,4′-dimethoxytrityl)oxymethyl]-5-(2-levulinoyloxyethyl)tetrahydrofuran (35). Levulinic anhydride (1.5 g, 7.1 mmol) was added to a mixture of 34 (1.3 g, 2.4 mmol) in pyridine (20 mL). The mixture was stirred at ambient temperature for 3 h and evaporated to dryness. The residue was dissolved in dichloromethane and washed with saturated NaHCO3. The organic phase was dried with Na2SO4 and evaporated to dryness. The crude product was purified by silica gel chromatography (0.1% Et3N and 3% MeOH in CH2Cl2) to yield 1.5 g (86%) of 35 as colorless oil. 1H NMR (500 MHz, CDCl3) δ 7.49–7.46 (m, 2H), 7.37–7.20 (m, 7H), 6.85–6.83 (m, 4H), 4.30–4.18 (m, 4H), 3.90 (m, 1H), 3.81 (s, 6H), 3.12–3.05 (m, 2H), 2.76 (m, 2H), 2.60 (m, 2H), 2.20 (s, 3H), 1.93–1.85 (m, 3H), 1.66 (ddd, 1H, J ) 12.6 Hz, 9.8 Hz, 5.8 Hz), 0.86 (s, 9H), 0.02 (s, 3H), 0.00 (s, 3H);13C NMR (CDCl3, 100 MHz) δ 172.7, 158.3, 144.9, 136.2, 136.1, 130.0, 128.2, 127.7, 126.6, 113.0, 86.3, 85.9, 75.2, 74.3, 64.3, 62.1, 55.2, 41.4, 37.9, 34.8, 29.9, 28.0, 25.8, -4.7, -4.8; MS (EI) M+ requires 676.4, found 676. (2R,3S,5S)-2-[(4,4′-Dimethoxytrityl)oxymethyl]-5-(2-levulinoyloxyethyl)tetrahydrofuran-3-ol (36). Tetrabutylammonium fluoride monohydrate (1.5 g, 5.7 mmol) was added to a mixture of 35 (1.2 g, 1.8 mmol) in tetrahydrofurane (15 mL). The mixture was stirred for 2 h at room temperature and then volatiles were removed. The residue was dissolved in dichloromethane and washed with water. The organic phase was dried with Na2SO4 and evaporated to dryness. The crude product was purified by silica gel chromatography (0.1% Et3N and 3% MeOH in CH2Cl2) to yield 0.90 g (87%) of 36 as colorless oil. 1 H NMR (500 MHz, CDCl3) δ 7.47–7.45 (m, 2H), 7.36–7.21 (m, 7H), 6.86–6.83 (m, 4H), 4.32 (b, 1H), 4.29–4.17 (m, 3H), 3.92 (m, 1H), 3.81 (s, 6H), 3.24 (dd, 1H, J ) 9.6 Hz, 4.6 Hz), 3.10 (dd, 1H, J ) 9.6 Hz, 5.9 Hz), 2.76 (m, 2H), 2.58 (m, 2H), 2.20 (s, 3H), 2.01 (b, 1H), 2.00 (ddd, 1H, J ) 13.0 Hz, 7.6 Hz, 2.1 Hz), 1.96–1.89 (m, 2H), 1.79 (ddd, 1H, J ) 13.0 Hz, 9.8 Hz, 6.3 Hz); 13C NMR (CDCl3, 250 MHz) δ 206.8, 172.7, 158.5, 144.9, 136.1, 130.1, 128.2, 127.9, 127.8, 126.8, 113.1, 86.1, 85.8, 75.3, 74.7, 64.7, 62.0, 55.3, 55.2, 40.9, 38.0, 34.7, 29.9, 28.0; MS (EI) M+ requires 562.3, found 562. 2-Cyanoethyl {(2R,3S,5S)-2-[(4,4′-Dimethoxytrityl)oxymethyl]-5-(2-levulinoyloxyethyl)tetrahydrofuran-3-yl} N,NDiisopropylphosphoramidite (3). Compound 3 was synthesized from 36 as described for 1, above. After silica gel chromatography (2% Et3N and 48% petroleum ether in EtOAc), product 3 was obtained as white foam in an 82% yield (0.54 g). 1H NMR (500 MHz, CD3CN) δ 7.49–7.46 (m, 2H), 7.38–7.33 (m, 4H), 7.32–7.28 (m, 2H), 7.24–7.20 (m, 1H), 6.86–6.82 (m, 4H), 4.46–4.42 (m, 1H), 4.32–4.21 (m, 3H), 4.09–4.08 (m, 1H), 3.86–3.53 (m, 4H), 3.82 (s, 3H), 3.81 (s, 3H), 3.19–3.09 (m, 2H), 2.78–2.75 (m, 2H), 2.64–2.58 (m, 3H), 2.48–2.45 (m, 1H), 2.20 (s, 3H), 2.20–2.07 (m, 1H), 1.97–1.93 (m, 2H), 1.82–1.75
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides
Figure 1. Aminoglycoside phosphoramidites (1 and 2) applicable to the automated oligonucleotide synthesis.
Figure 2. Branching units (3–5) used for the intrachain conjugation.
(m, 1H), 1.20, 1.19, 1.17, 1.17, 1.15, 1.10, 1.08 (each s, sum 12H); 13C NMR (CD3CN, 250 MHz) δ 206.8, 206.8, 172.6, 172.6, 158.7, 145.3, 136.1, 136.1, 136.1, 136.0, 130.1, 130.0, 128.1, 128.1, 127.8, 126.8, 126.8, 118.6, 118.4, 113.1, 85.9, 85.3, 85.3, 85.0, 85.0, 75.7, 75.7, 75.6, 75.5, 75.4, 75.2, 64.4, 64.3, 61.6, 58.5, 58.4, 54.9, 43.0, 42.9, 40.0, 40.0, 40.0, 39.9, 37.5, 34.5, 34.4, 29.0, 27.8, 24.0, 24.0, 23.9, 23.9, 20.1, 20.1, 20.0, 20.0; 31P NMR (CDCl3, 200 MHz) δ 148.0, 147.8; HRMS (ESI) [M + Na]+ C42H55N2NaO9P requires 785.3537, found 785.3542. Synthesis of the Aminoglycoside Conjugates of the 2′O-Methyl Oligoribonucleotides (6–13). The 2′-O-methyl oligoribonucleotide conjugates (6–13) were assembled on an Applied Biosystems 392 DNA synthesizer in 1.0 µmol scale using conventional phosphoramidite chemistry and following the standard RNA coupling protocol. The coupling efficiency of the branching units 3–5 was equal to that of the commercially available 2′-O-methylribonucleoside phosphoramidites (0.1 mol L-1 solution of 3, 4, or 5 in acetonitrile, 4,5-dicyanoimidazole activation, 600 s coupling time, standard oxidation cycle: 0.02 mol L-1 I2 in mixture of pyridine, H2O, and THF, 0.4/9.1/90.5 v/v/v, 30 s) (42). Once the 2′-O-methyl oligoribonucleotide chains were assembled, the orthogonal levulinoyl protections of 40–42 were manually removed by conventional hydrazinium acetate treatment [i.e., resins are suspended in a mixture of NH2NH2 · H2O, pyridine, and AcOH (0.124/4/1 v/v/v, 1 h at 25 °C), washed with pyridine, MeOH, and acetonitrile, dried, and reset to the DNA synthesizer]. The aminoglycoside phosphoramidites (1 and 2) were then coupled to the hydroxyalkyl arms of 43–45 or to the exposed 5′-OH group of 37, using two consecutive standard couplings (0.1 mol L-1 solution of 1 or 2 in dry acetonitrile, 4,5-dicyanoimidazole activation, 600 s coupling time, twice, followed by standard oxidation cycle, as above). In order to avoid potential acyl migration, the levulinoyl protections of the amino glycoside moieties of 38, 39, and 46–51 were selectively removed using the same hydrazinium acetate treatment, as above. Finally, the resins were subjected to concentrated ammonia (15 h at 55 °C) to release the crude 2′O-methyl oligoribonucleotide conjugates (6–13). The product mixtures were filtered, the filtrates were evaporated to dryness, the residues were dissolved in water, and then the conjugates (6–13) were subjected to RP HPLC purification (Figure 1, 2, 3, 4). Purity of the crude conjugates ranged from 41% to 66%. Isolated yields of the homogenized conjugates were 15% to 26% (according to UV absorbances of dissolved 6–13). The authenticity of the conjugates was verified by MS(ESI) spectroscopy (properties of the 6–13; see Table 1). Kinetic Measurements. The reactions were carried out in Eppendorf tubes immersed in a water bath, the temperature
Bioconjugate Chem., Vol. 19, No. 3, 2008 771
being maintained at 35.0 ( 0.1 °C. pH of the samples was adjusted to 7.3 with a HEPES buffer (0.1 mol L-1) prepared in sterilized water. For the metal-ion-dependent experiments, Zn2+ ion was added as a nitrate salt to give the overall metal ion concentration of 90 µmol L-1. The ionic strength was adjusted to 0.1 mol L-1 with sodium nitrate. The concentrations of the aminoglycoside conjugates (6–13) and the targets (14 and 15) were 18 µmol L-1. 4-Nitrobenzenesulfonate ion was used as an internal standard. The total volume of the reaction mixture was 200 µL in each kinetic run. Aliquots of 20 µL were withdrawn at suitable intervals, cooled to 0 °C, and the reaction quenched by adding of 1.0 mol L-1 aqueous hydrogen chloride (1.0 µL). The composition of the samples was quantified by capillary zone electrophoresis (Beckman Coulter P/ACE MDQ CE System) using a fused silica capillary (50 µm, 50 cm). Inverted polarity, citrate buffer (0.2 mol L-1, pH 3.1), and –30 kV voltage were used. The temperature of the capillary was kept at 25 °C. Hydrodynamic injection at 2 psi for 8 s was applied. Between each analytical run, the capillary was flushed for 3 min with water, 10 mmol L-1 aqueous hydrogen chloride, and the background electrolyte buffer. The quantification of the target oligonucleotides (14 and 15) was based on comparison of their UV absorption at 254 nm to that of internal standard. The peak area was first normalized by dividing it by the migration time and then by the similarly normalized peak area of the internal standard. First-order rate constants for the cleavage of the target oligonucleotides by the conjugates (6–13) were obtained by applying the integrated first-order rate law to the disappearance of the target oligonucleotide (14 and 15).
RESULTS AND DISCUSSION Synthesis of the Aminoglycoside Phosphoramidites (1, 2, and 31). For the synthesis of aminoglycoside conjugates of 2′-O-methyl oligoribonucleotides, three different neamine-based phosphoramidites (1, 2, and 31) were initially synthesized (outlined in Schemes 1, 2, 3). Commercially available neamine and ribostamycin are rather expensive starting materials to be used for the multistep synthesis. Neomycin trisulfate (16) may readily be hydrolyzed to neamine tetrahydrochloride (17) (30), and, hence, it served as a cheap starting material for each compound. In 1 and 2, hydroxyethyl and ribose residues were introduced between the phosphoramidite and neamine moieties. In 31, the O5 was directly phosphitylated. Synthesis of this reagent was straightforward, but in contrast to 1 and 2, it turned out be incompatible with the automated RNA synthesis. Introduction of hydroxyethyl tether to 1 involves ether bond formation, which requires complete protection of the amino functions. In contrast to carbamates or amides (11, 38), azide masking of 18 meets this requirement. This was carried out according to literature by Cu2+-catalyzed diazo transfer (31). Three of the four hydroxyl groups were then selectively acetylated with acetic anhydride in pyridine to give 19 (71%) (31). In this reaction, all byproducts may be combined and 18 may be regenerated by conventional sodium methoxide treatment (39). The reaction may be repeated with crude regenerated 16 in the same yield. The hydroxyethyl moiety was introduced to the O5-position of neamine via an allyl ether intermediate. The free OH-group of 19 was allylated under neutral condition with allyl methylcarbonate in the presence of Pd(OAc)2 and triphenylphosphine resulting in 20 in 42% yield (33). During this reaction, some acyl migration occurred giving 6-O-allyl1,3,2′,6′-tetraazido-5,3′,4′-tri-O-acetyl neamine (17%) as a side product. OsO4 hydroxylation and concomitant NaIO4 oxidation converted the allyl group of 20 to an oxoethyl group, which was subsequently reduced to the hydroxyethyl group with NaBH4 (34). Removal of the acetyl protections by sodium methoxide treatment (39) gave 21 in 53% overall yield. MMTr-
772 Bioconjugate Chem., Vol. 19, No. 3, 2008
Virta and Ketomäki
Figure 3. Synthesized neamine (6–9) and ribostamycin (10–13) conjugates of 2′-O-methyl oligoribonucleotides and the target oligonucleotides (14 and 15). Table 1. Properties of the Synthesized 2′-O-Methyl Oligoribonucleotide Conjugates (6–13) 2′-O-methyl observed oligoribonucleotide purity of the required entry conjugate crude conjugatea % massb g mol-1 mass 1 2 3 4 5 6 7 8
6 7 8 9 10 11 12 13
64 66 51 61 65 65 44 41
3727.8 5868.9 6372.9 6372.9 3815.7 5956.8 6550.2 6550.2
3726.8 5867.1 6371.9 6371.9 3814.8 5955.2 6547.4 6547.4
a Purities according to RP HPLC chromatograms of crude mixtures of 6–13. b Masses calculated from [(M - 3H)/3]-3.
Figure 4. An example of RP HPLC chromatogram of crude product mixture (conjugate 8).
capping of the primary OH group, in the presence of the secondary ones, gave quantitatively 22. Staudinger reaction with Me3P released the amino groups, which were subsequently protected by trifluoroacetyl groups. These two steps may be performed without chromatographic purification, and synthesis was directly continued to the levulinoylation resulting in 23 in 74% overall yield. Finally, the MMTr group was removed with 3% dichloroacetic acid (24, 97%) and the exposed primary hydroxyl group was phosphitylated (1, 93%). The azide masked neamine (19) was a useful starting material also for the ribostamycin derived phosphoramidite (2). The 5-OH of protected neamines is usually only weakly reactive, like in acylations (cf. 16 and 27), but 17 has turned out to be a
good glycosyl acceptor (35, 40, 41). Analogously to the previously published furanosylation of paromamine (40), 1,2,3,5tetra-O-acetyl-β-D-ribofuranose was simply used as a donor. The protected ribostamycin (25) was obtained in 59% yield. PerO-acetylated phenylthioribofuranoside has recently been used for the same reaction to give 25 in 37% yield (35). After deacetylation of 25, synthesis to obtain 2 was continued analogously to the described preparation of 1 from 21. The primary hydroxyl group was protected with an MMTr group (26, 71%), the azide masks were removed, the trifluoracetyl and levulinoyl protections were introduced (27, 79%), the primary hydroxyl group was released (28, 91%) and, finally, the phosphitylation was performed to give 2 (92%). The phosphoramidite 31 was synthesized from the neamine tetrahydrochloride (17). The amino groups were converted to the trifluoroacetamides according to the literature (29, 91%) (28). The hydroxyl groups were then selectively esterified as levulinoylates (30, 67%), and the remained free 5-OH was phosphitylated to give 31 (90%). Synthesis of Branching Units 3–5. Preparation of the ribosebased branching units 4 and 5, and their applicability to the automated RNA synthesis have recently been described (42). Previously prepared ethyl 2-[(2R,3S,5R)-5-(4,4′-dimethoxytrityloxymethyl)-3-hydroxytetrahydrofuran-2-yl] acetate (32) (36, 37)
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides
Bioconjugate Chem., Vol. 19, No. 3, 2008 773
Scheme 1. Syntheses of the Aminoglycoside Phosphoramiditesa
a Reagents and conditions: (i) aq HCl, MeOH, reflux; (ii) TfN3, K2CO3, CuSO4, H2O, MeOH, CH2Cl2; (iii) Ac2O, DMAP, pyridine; (iv) allyl methyl carbonate, Pd(OAc)2, PPh3, THF, reflux; (v) OsO4, iPrOH, NaIO4, H2O, dioxane; (vi) NaBH4, EtOH, CH2Cl2; (vii) NaOMe, MeOH; (viii) MMTrCl, pyridine; (ix) Me3P, toluene, H2O, dioxane, N2 atm.; (x) TfaOMe, Et3N, MeOH; (xi) levulinic anhydride, DMAP, pyridine; (xii) DCA, CH2Cl2; (xiii) 2-cyanoethyl N,N-diisopropylphosphonamidic chloride, Et3N, dichloromethane; (xiv) 1,2,3,5-tetra-O-acetyl-β-ribofuranose, TMSTf, CH2Cl2, 0°C.
Scheme 2. Synthesis of the Branching Unit 3a
a Reagents and conditions: (i) TBDMSCl, imidazole, pyridine, 40 °C; (ii) LiAlH4, THF; (iii) Lev2O, pyridine; (iv) Bu4N+F- · H2O, THF.
was used as a starting material for the branching unit 3. The secondary hydroxyl group of 32 was tert-butyldimethylsilylated (33, 60%) and then the ethyl ester was reduced by lithium aluminum hydride to give 34 (86%). The resultant primary alcohol was capped as a levulinoyl ester (35, 86%), the TBDMS protection was removed (36, 87%), and the exposed hydroxyl group was phosphitylated to give 3 (82%). Synthesis of the Aminoglycoside Conjugates of 2′-OMethyl Oligoribonucleotides. To verify and optimize the applicability of the aminoglycoside phosphoramidites (1, 2, and 31) to the automated oligonucleotide synthesis, they were first coupled to the 5′-end of a simple solid-supported T5 homooligodeoxyribonucleotide. Coupling efficiencies were evaluated by RP HPLC analysis of the released product mixtures. Use of standard RNA coupling protocol gave ca. 80% coupling yield
for 1 and 2. These couplings were improved to an acceptable 95% yield by using of two consecutive coupling cycles (i.e., 0.1 mol L-1 solution of 1 or 2 in dry acetonitrile, 4,5dicyanoimidazole activation, 600 s coupling time, twice, followed by standard oxidation cycle: 0.02 mol L-1 I2 in mixture of pyridine, H2O, and THF, 0.4/9.1/90.5 v/v/v, 30 s). This procedure turned out to also give an acceptable yield for the conjugates bearing two aminoglycoside ligands (cf. 8, 9, 12, and 13). In contrast to 1 and 2, coupling of 31 remained at a poor level and, hence, this building block was excluded from the synthesis of final conjugates. Coupling of the hydroxyalkylarmed branching units (4 and 5) has recently been shown to be as efficient as that of the commercially available 2′-O-methylribonucleoside building blocks; the couplings were quantitative using a standard RNA coupling protocol (42). The branching unit 3 behaved similarly. In order to demonstrate the importance of orthogonally removable levulinoyl protections on the aminoglycoside moieties, direct ammonolysis (15 h, 55 °C) without hydrazinium acetate treatment was used for release of the fully protected ribostamycin conjugate (2-T5). In this manner, the yield of the desired conjugate decreased by 22% and N-acyl side products appeared (according to RP HPLC chromatograms; see Supporting Information). Thus, the OfN acylation would be a remarkable problem, especially if two aminoglycoside ligands are to be attached (cf. 8, 9, 12, and 13). One hour treatment with a mixture of hydrazinium acetate (NH2NH2 · H2O,
774 Bioconjugate Chem., Vol. 19, No. 3, 2008
Virta and Ketomäki
Scheme 3. Synthesis of the Neamine (6–9) and Ribostamycin (10–13) Conjugates of the 2′-O-Methyl Oligoribonucleotidesa
a Reagents and conditions: (i) DCA, CH2Cl2; (ii) 1 or 2, 4,5-dicyanoimidazole, acetonitrile; (iii) I2, pyridine, H2O, THF; (iv) standard oligoribonucleotide synthesis; (v) hydrazinium acetate, AcOH, pyridine; (vi) aq NH3, 15 h at 55 °C.
pyridine, AcOH, 0.124/4/1, v/v/v at 25 °C) eliminated acyl migrated products completely. Traces of a defective Tfa removal were found in the product mixure. This small amount of unexpectedly stable Tfa-amides may be slightly decreased (6% to 4% of total absorbance in RP HPLC) by using prolonged ammonolysis (48 h, at 55 °C). With these remarks, conjugates 6–13 were synthesized. The 2′-O-methyl oligoribonucleotide chains 37, and 40–42 were assembled on a 1 µmol scale following the standard RNA coupling protocol. Levulinoyl protections of the inserted branching units (40-42) were removed by hydrazinium acetate treatment (NH2NH2 · H2O, pyridine, AcOH, 0.124/4/1, v/v/v, 1 h at 25 °C), and then the aminoglycoside phosphoramidites (1 and 2) were coupled to the hydroxyalkyl arms of 43-45 or to the exposed 5′-end of 37. Two consecutive standard RNA couplings (0.1 mol L-1 solution of 1 or 2 in dry acetonitrile, 600 s coupling time, twice) were used in each case. Levulinoyl protections of the aminoglycoside moieties (38, 39, 46–51) were orthogonally removed by the same hydrazinium acetate treatment as used for the hydroxyalkyl arms above, and then the conjugates were released by ammonia (conc aq NH3, 15 h at 55 °C). Purity of the conjugates (6, 7, 9, and 10) having one aminoglycoside unit was ca. 65%. Attachment of two aminoglycosidic ligands lowered the purity (8, 9, 12, and 13) to 41–61% (see Figure 4 and Table 1). Each conjugate was easily purified by RP HPLC (Figure 5). The isolated yield ranged from 15% to 26%. Authenticity of the conjugates 6–13 was verified by MS(ESI) (see Table 1). MS(ESI) also showed that the desired conjugates were free from other closely related oligonucleotide derivatives, such as OfN-acyl migrated side products.
Figure 5. RP HPLC chromatograms (i-viii) of the homogenized conjugates (6–13), respectively.
RNA Cleavage by the Aminoglycoside Conjugates (6– 13). Cleavage activity of the conjugates (6–13) was studied by using chimeric ribo/2′-O-methylribo oligonucleotides (14 and 15) as targets. Upon hybridization (excluding experiments with 6 and 10), a tri- and pentanucleotide bulge (A3 and A5), serving as potential cleaving sites, is formed opposite the aminoglycoside regions. Each aminoglycoside conjugate (6–13) was
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides
incubated with the target (14 or 15) at 18 µmol L-1 equimolar concentration (35 °C, pH ) 7.3, I ) 0.1 mol L-1). Aliquots were withdrawn at appropriate intervals, and their composition was determined by zone capillary electrophoresis. In these metalion-independent cleavage studies, one of the conjugates (i.e., conjugate 9) accelerated the hydrolysis of target oligonucleotides (14 and 15), the first-order rate constants being (2.7 ( 0.6) × 10-7 s-1 and (2.3 ( 0.6) × 10-7 s-1, respectively, while the other conjugates turned out to be almost inactive. Cleavage of 14 by 11 [k ) (1.0 ( 0.3) × 10-7 s-1] may still be noteworthy, but, e.g., target 14 in the presence of conjugates 6, 7, 10, and 12 remained fully intact (100%, kcat ) kuncat ) 0, an aliquot withdrawn after three weeks incubation). Similarly, target 15 remained intact (99–100%) in the presence of 6-8 and 10, and negligible cleavage (under the limits of experimental error) was observed with the other experiments (14 in the presence of 8 and 13 and 15 in the presence of 11, 12, and 13). All the conjugates contain the same active core, i.e., neamine, but site within the sequence selective motif (6 and 10 vs 7–9, 11, 12, and 13), number of the cores (6, 7, 10, and 11 vs 8, 9, 12, and 13), structural orientations and lengths of the spacers are different. Comparing 9 to 8, both conjugates have similarly orientated neamine in the 1-O-position (branching unit 4 vs 5), but in contrast to 9, conjugate 8 is inactive. Comparing 9 to 13, both conjugates bear the same branching unit (5), but ribostamycine (13) instead of ethylene glycol derived neamine (9) reduces the activity. Hence, the 3-O-tethered orientation for the neamine (4) together with an appropriate spacer (ethylene glycol in 9) probably is favorable for the cleavage. We recently described cleavage properties of di(azacrown) conjugated oligonucleotides with the same targets (14 and 15). Interestingly, the 3-O ribose (4) branched azacrown conjugate was even in that case superior to other closely related conjugates (42). Metalion-independent cleavage of HIV-1 TAR RNA by neamine-PNA conjugates has recently been reexamined by Chaubey et al. (27). Consistent with the foregoing, they noted that small structural differences perturbed the activity of the neamine-PNA conjugates; only one of the conjugates resulted in a remarkable cleavage. Chaubey et al. proposed the following mechanism for the neamine catalyzed hydrolysis of RNA (27). The deprotonated amino group (pKa ) 6.9) at the 3-position of ring I extracts a hydrogen atom from the 2′-OH group of the ribose moiety of the RNA, while the protonated amino group (pKa ) 9.2) at the 6′-position of ring II polarizes the phosphate group by electrostatic interaction. These two interactions, together, facilitate intramolecular nucleophilic attack of 2′-OH to the phosphorus atom, which leads to the formation of 2′,3′-cyclic phosphate and cleavage of RNA. In comparison to metal chelate-based nucleases, metal-ion-independent activity, which is based on unprotonated and protonated amino groups in a negatively charged environment, understandably may be perturbed more dramatically by structural differences. In order to study metalion dependence of the cleavages by 6–13, each experiment was additionally performed in the presence of Zn2+ ion [90 µmol L-1 Zn(NO3)2] keeping the other conditions unchanged. Somewhat increased activities were now found. However, only in one case (conjugate 11 vs target 14) was the first-order rate constant (k ) 9.0 × 10-7 s-1) markedly different from that of the nonspecific cleavage by Zn2+ complexes. Otherwise, the accelerations [k ) (1.5–7.0) × 10-7 s-1] may be considered to be more or less marginal compared to the background obtained with the targets (14 and 15) in 90 µmol L-1 Zn2+ concentration alone (k < 1 × 10-6 s-1) (42, 43). It is worth noting that cleavage activity of conjugate 9 was not markedly accelerated by addition of the Zn2+ ion. The cleavage of 14 by 9 in the presence of the Zn2+ ion remained unchanged [k ) (2.7 ( 0.6)
Bioconjugate Chem., Vol. 19, No. 3, 2008 775
× 10-7 s-1] and slight Zn2+ promoted acceleration to cleavage of 15 by 9 [k ) (6.0 ( 0.3) × 10-7 s-1] was found. In the present study, equimolar concentrations of the conjugates (6–13) with the targets (14 and 15) have been used, in which the rate constants for the metal-ion-independent cleavages of 14 and 15 were as low as k e 2.7 × 10-7 s-1. Real catalytic activity would require further identification (including experiments, in which a catalytic amount of the conjugate and excess of the targets are used), but as may already be noticed by these rate constants, none of the conjugates met requirements for a real catalyst. An obvious explanation to the modest cleavage activities is restricted access of the neamine cores to the phosphodiester bonds of target RNA bulges. However, this is not necessarily completely caused by structural reasons, such as an inappropriate spacer between the neamine and the 2′-Omethyl oligoribonucleotides or the structure of the branching units, but competitive binding of the positively charged aminoglycoside to phosphodiester groups on the same molecule (i.e., intramolecular electrostatic interaction) may also play a role. This explanation is partly consistent with findings that a polyamine-PNA conjugate provides higher catalytic activity compared to its oligonucleotide counterpart (44, 45). Among the metal-ion-independent nucleases, bearing a 2′-deoxy oligoribonucleotide or 2′-O-methyl oligoribonucleotide as a sequence-selective motif, only a tris(2-aminobenzimidazole)-based nuclease has been reported to have marked cleaving activity so far (46). It should also be noted that the chimeric oligonucleotides 14 and 15 are relatively stable as targets. For example, the Zn2+complexes of the di(azacrown) conjugates of 2′-Omethyl oligoribonucleotides, which in an optimal case have equal catalytic activity to the most efficient Zn2+ and Cu2+based artificial nucleases (47, 48), also have moderate cleavage ability [k ) (0.9–1.7) × 10-6 s-1] to these targets (14 and 15) (42). Thus, the real cleavage potential of the aminoglycoside conjugates (6–13) may still remain obscure, and further studies with different targets are required.
CONCLUSION In the present study, aminoglycoside conjugates of 2′-Omethyl oligoribonucleotides (6–13) have been successfully synthesized entirely on a solid support. The O-levulinoyl and N-trifluoroacetyl protecting group combination has been used for the aminoglycoside moieties, which enables elimination of OfN-acyl migration. The potential of these conjugates to act as a sequence-selective and metal-ion-independent artificial nuclease has been studied with chimeric ribo/2′-O-methyl oligonucleotides (14 and 15). With these experiments, one of the conjugates (9) resulted in metal-ion-independent cleavage of the targets [14 and 15, k ) (2.7 ( 0.6) × 10-7 s-1 and (2.3 ( 0.6) × 10-7 s-1, respectively], but usually the cleavage obtained by the conjugates (6–8, 10–13) remained at a marginal level or the conjugates were fully inactive.
ACKNOWLEDGMENT The authors thank Professor Harri Lönnberg for advice considering the cleavage studies and corrections in the preparation of the manuscript. The financial support from the Academy of Finland is gratefully acknowledged. Supporting Information Available: NMR and MS spectral data for 1–3, 6–13, 20–31, and 33–36 and RP HPLC chromatographic data as evidence for the OfN-acyl migration. This material is available free of charge via the Internet at http:// pubs.acs.org.
776 Bioconjugate Chem., Vol. 19, No. 3, 2008
LITERATURE CITED (1) Davies, J., Gorini, L., and Davis, B. D. (1965) Misreading of RNA codewords induced by aminoglycoside antibiotics. Mol. Pharmacol. 1, 93–106. (2) Moazed, D., and Noller, H. F. (1987) Interaction of antibiotics with functional sites in 16S ribosomal RNA. Nature 327, 389– 394. (3) Von Ahsen, U., and Noller, H. F. (1993) Footprinting the sites of interaction of antibiotics with catalytic group I intron RNA. Science 260, 1500–1503. (4) Francois, B., Russell, R. J. M., Murray, J. B., Aboul-ela, F., Masquida, B., Vicens, Q., and Westhof, E. (2005) Crystal structures of complexes between aminoglycosides and decoding site oligonucleotides: role of the number of rings and positive charges in the specific binding leading to miscoding. Nucleic Acid Res. 33, 5677–5690. (5) Von Ahsen, U., Davies, J., and Schroeder, R. (1991) Antibiotic inhibition of group I ribozyme function. Nature 353, 368–370. (6) Herman, T. (2005) Drugs targeting the ribosome. Curr. Opin. Struct. Biol. 15, 355–366. (7) Mei, H.-Y., Galan, A. A., Halim, N. S., Mack, D. P., Morelan, D. W., Sanders, K. B., Truong, H. N., and Czarnik, A. W. (1995) Inhibition of an HIV-1 Tat-derived peptide binding to TAR RNA by aminoglycoside antibiotics. Bioorg. Med. Chem. Lett. 5, 2755– 2760. (8) Zapp, M. L., Stern, S., and Green, M. R. (1993) Small molecules that selectively block RNA binding of HIV-1 rev protein inhibit rev function and viral production. Cell 74, 969– 978. (9) Tam, V. K., Kwong, D., and Tor, Y. (2007) Fluorescent HIV-1 dimerization initiation site: design, properties, and use for ligand discovery. J. Am. Chem. Soc. 129, 3257–3266. (10) Ennifar, E., Paillart, J.-C., Bodlenner, A., Walter, P., Weibel, J.-M., Aubertin, A.-M., Pale, P., Dumas, P., and Marquet, R. (2006) Targeting the dimerization initiation site of HIV-1 RNA with aminoglycosides: from crystal to cell. Nucleic Acid Res. 34, 2328–2339. (11) Liu, M., Haddad, J., Axucena, E., Kotra, L. P., Kirzhner, M., and Mobashery, S. (2000) Tethered bisubstrate derivatives as probes for mechanism and as inhibitors of aminoglycoside 3′phosphotransferases. J. Org. Chem. 65, 7422–7431. (12) Yajima, S., Shionoya, H., Akagi, T., and Hamasaki, K. (2006) Neamine derivatives having a nucleobase with a lysine or an arginine as a linker, their synthesis and evaluation as potential inhibitors for HIV TAR-Tat. Bioorg. Med. Chem. Lett. 14, 2799– 2809. (13) Blount, K. F., and Tor, Y. (2006) A tale of two targets: Differential RNA selectivity of nucleobase-aminoglycoside conjugates. ChemBioChem 7, 1612–1621. (14) Cai, L., Li, Q., Ren, B., Yang, Z.-J., Zhang, L.-R., and Zhang, L.-H. (2007) Synthesis of aminodisaccharide-nucleoside conjugates for RNA binding. Tetrahedron 63, 8135–8144. (15) Charles, I., Xue, L., and Arya, D. P. (2002) Synthesis of aminoglycoside-DNA conjugates. Bioorg. Med. Chem. Lett. 12, 1259–1262. (16) Hyun, S., Lee, K. H., and Yu, J. (2006) A strategy for the design of selective RNA binding agents. Preparation and RRE RNA binding affinities of a neomycin-peptide nucleic acid heteroconjugate library. Bioorg. Med. Chem. Lett. 16, 4757–4759. (17) Yoshinary, K., Yamasaki, K., and Komiyama, M. (1991) Oligoamines as simple and efficient catalyst for RNA hydrolysis. J. Am. Chem. Soc. 113, 5899–5901. (18) Oivanen, M., Kuusela, S., and Lönnberg, H. (1998) Kinetics and mechanism for the cleavage and isomerization of the phosphodiester bonds of RNA by Brθnsted acids and bases. Chem. ReV. 98, 961–990. (19) Kirk, S. R., and Tor, Y. (1998) Hydrolysis of an RNA dinuclear monophosphate by neomycin B. Chem. Commun. 147– 148.
Virta and Ketomäki (20) Sreedhara, A., Patwardhan, A., and Cowan, J. A. (1999) Novel reagents for targeted cleavage of RNA sequences: towards a new family of inorganic pharmaceuticals. Chem. Commun. 1147– 1148. (21) Sreedhara, A., Freed, J. D., and Cowan, J. A. (2000) Efficient inorganic deoxyribonucleases. Greater than 50 million-fold rate enhancement in enzyme-like DNA cleavage. J. Am. Chem. Soc. 122, 8814–8824. (22) Niittymäki, T., and Lönnberg, H. (2006) Artificial ribonucleases. Org. Biomol. Chem. 4, 15–25. (23) Cowan, J. A. (2001) Chemical nucleases. Curr. Opin. Chem. Biol. 5, 634–642. (24) Komiyama, M., Sumaoka, J., Kuzuya, A., and Yamamoto, Y. (2001) Sequence-selective artificial ribonucleases. Methods Enzymol. 341, 455–468. (25) Travick, B. N., Daniher, A. T., and Bashkin, J. K. (1998) Inorganic mimics of ribonucleases and ribozymes: from random cleavage to sequence-specific chemistry to catalytic antisense drugs. Chem. ReV. 98, 939–960. (26) Riguet, E., Tripathi, S., Chaubey, B., Désire, J., Pandey, V. N., and Décout, J.-L. (2004) A peptide nucleic acid-neamine conjugate that targets and cleaves HIV-1 TAR RNA inhibits viral replication. J. Med. Chem. 47, 4806–4809. (27) Chaubey, B., Tripathi, S., Désire, J., Baussanne, I., Décout, J.-L., and Pandey, V. (2007) Mechanism of RNA cleavage by sequence specific polyamide nucleic acid-neamine conjugate. Oligonucleotides 17, 302–313. (28) Tona, R., Bertolini, R., and Hunziker, J. (2000) Synthesis of aminoglycoside-modified oligonucleoties. Org. Lett. 2, 1693– 1696. (29) Charles, I., Xi, H., and Arya, D. P. (2007) Sequence-specific targeting of RNA with an oligonucleotide-neomycin conjugate. Bioconjugate Chem. 18, 160–169. (30) Park, W. K. C., Auer, M., Jaksche, H., and Wong, C.-H. (1996) Rapid combinatorial synthesis of aminoglycoside antibiotic mimetics: Use of a polyethylene glycol-linked amine and a neamine-derived aldehyde in multiple component condensation as a strategy for the discovery of new inhibitors of the HIV RNA rev response element. J. Am. Chem. Soc. 118, 10150–10155. (31) Alper, P. B., Hung, S.-C., and Wong, C.-H. (1996) Metal catalyzed diazo transfer. Tetrahedron Lett. 37, 6029–6032. (32) Van den Broek, S. A. M. E., Gruijters, B. W. T., Rutjes, F. P. J. T., van Delft, F. L., and Blaauw, R. H. (2007) A Short and scalable route to orthogonally O-protected 2-deoxystreptamine. J. Org. Chem. 72, 3577–3580. (33) Guibe, F., and Saint M’Leux, Y. (1981) The allyloxycarbonyl group for alcohol protection: quantitative removal or transformation into allyl protecting group via π-allyl complexes of palladium. Tetrahedron Lett. 22, 3591–3594. (34) Buskas, T., Söderberg, E., Konradsson, P., and Fraser-Reid, B. (2000) Use of n-pentenyl glycosides as precursor to various spacer functionalities. J. Org. Chem. 65, 958–963. (35) Li, J., Chiang, F., Chen, H.-N., and Chang, C.-W. T. (2007) Investigation of the regioselectivity for the Staudinger reaction and its application for the synthesis of aminoglycosides with N-1 modification. J. Org. Chem. 72, 4055–4066. (36) Hovinen, J., and Salo, H. (1997) C-glycoside phosphoramidite building block for versatile functionalization of oligodeoxyribonucleotides. J. Chem. Soc., Perkin Trans. 1, 3017–3020. (37) Niittymäki, T., Kaukinen, U., Virta, P., Mikkola, S., and Lönnberg, H. (2004) Preparation of azacrown-functionalized 2′O-methyl oligoribonucleotides, potential artificial Rnases. Bioconjugate Chem. 15, 174–184. (38) Haddad, J., Kotra, L. P., Llano-Sotelo, B., Kim, C., Azucena, E. F., Jr., Liu, M., Vakulenko, S. B., Chow, C. S., and Mobashery, S. (2002) Design of novel antibiotics that bind to the ribosomal acyltransfer site. J. Am. Chem. Soc. 124, 3229– 3237. (39) Zemplen, G., and Kuntz, A. (1923) The sodium compounds of glucose and the saponification of the acylated sugars. Brit. 56B, 1705–1710.
Synthesis of Aminoglycoside Conjugates of 2′-O-Methyl Oligoribonucleotides (40) Ding, Y., Swayze, E. E., Hofstadler, S. A., and Griffey, R. H. (2000) Efficient synthesis of neomycin B related aminoglycosides. Tetrahedron Lett. 41, 4049–4052. (41) Nudelman, I., Rebibo-Sabbah, A., Shallom-Shezifi, D., Hainrichon, M., Stahl, I., Ben-Yosef, T., and Baasov, T. (2006) Redesign of aminoglycosides for treatment of human genetic diseases caused by premature stop mutations. Bioorg. Med. Chem. Lett. 16, 6310–6315. (42) Niittymäki, T., Virta, P., Ketomäki, K., and Lönnberg, H. (2007) Di(azacrown) conjugates of 2′-O-methyl oligoribonucleotides as sequence-selective artificial ribonucleases. Bioconjugate Chem. 18, 1583–1592. (43) Niittymäki, T., and Lönnberg, H. (2004) Sequence-selective cleavage of oligoribonucleotides by 3d transition metal complexes of 1,5,9-triazacyclododecane-functionalized 2′-O-methyl oligoribonucleotides. Bioconjugate Chem. 15, 1275–1280. (44) Komiyama, M., Inokawa, T., and Yoshinary, K. (1995) Ethylenediamine-oligo DNA hybrid as sequence-selective artificial ribonuclease. J. Chem. Soc., Chem. Commun. 77–78.
Bioconjugate Chem., Vol. 19, No. 3, 2008 777 (45) Verheijen, J. C., Deiman, B. A. L. M., Yeheskiely, E., van der Marel, G. A., and van Boom, J. H. (2000) Efficient hydrolysis of RNA by a PNA-diethylenetriamine adduct. Angew. Chem., Int. Ed. 39, 369–372. (46) Gnaccarini, C., Peter, S., Scheffer, U., Vonhoff, S., Klussmann, S., and Göbel, M. W. (2006) Site-specific cleavage of RNA by a metal-free artificial nuclease attached to antisense oligonucleotides. J. Am. Chem. Soc. 128, 8063–8067. (47) Putnam, W. C., Daniher, A. T., Trawick, B. N., and Bashkin, J. K. (2001) Efficient new ribozyme mimics: direct mapping of molecular design principles from small molecules to macromolecular, biomimetic ctalysis. Nucleic Acid Res. 29, 2199–2204. (48) Åström, H., Williams, N. H., and Strömberg, R. (2003) Oligonucleotide based artificial nuclease (OBAN) systems. Bulge size dependence and positioning of catalytic group in cleavage of RNA-bulges. Org. Biomol. Chem. I, 1461–1465. BC7004279