Highly efficient catalytic RNA cleavage by the cooperative action of two Cu(II) complexes embodied within an antisense oligonucleotide

Abstract

Based on our recent studies of RNA cleavage by oligonucleotide–terpyridine·Cu(II) complex 5′- and/or 3′-conjugates, we designed 2′-O-methyloligonucleotides with two terpyridine-attached nucleosides at contiguous internal sites. To connect the 2′-terpyridine-modified uridine residue at the 5′-side to the 5′-O-terpyridyl nucleoside residue at the 3′-side, a dimethoxytrityl derivative of 5-hydroxypropyl-5′-O-terpyridyl-2′-deoxyuridine-3′-phosphoramidite was newly synthesized. Using this unit, we constructed two terpyridine conjugates, with either an unusual phophodiester bond or the bond extended by a propanediol(s)-containing linker. Cleavage reactions of the target RNA oligomer, under the conditions of conjugate excess in the presence of Cu(II), indicated that the conjugates precisely cleaved the RNA at the predetermined site and that one propanediol-containing linker was the most appropriate for inducing high cleavage activity. Furthermore, a comparison of the activity of the propanediol agent with those of the control conjugates with one complex confirmed that the two complexes are required for efficient RNA cleavage. The reaction of the novel cleaver revealed a bell-shaped pH–rate profile with a maximum at pH ∼7.5, which is a result of the cooperative action of the complexes. In addition, we demonstrated that the agent catalytically cleaves an excess of the RNA, with the kinetic parameter k_cat/K_m = 0.118 nM^–1 h^–1.

INTRODUCTION

Various types of modified oligonucleotides with reactive groups have been chemically synthesized and examined for their functions. One such area of interest today is the development of artificial RNases, which can cleave RNA in a sequence-specific manner and via the transesterification and/or hydrolysis of the target phosphodiester linkage(s) (1–4). Such RNA cleavers may be useful as nuclease mimics for studies on the structure–function relationships of native RNAs and are potentially applicable for antisense chemotherapy. They have been constructed by coupling an RNA cleavage catalyst, such as an oligoamine (5,6), an imidazole (7–10) or a metal complex (11–13), to an oligonucleotide (DNA, phosphate-modified DNA, sugar-modified RNA or PNA analog) via a linker arm.

One of the main goals in developing highly active RNA cleavers is to achieve the catalytic cleavage of an excess of RNA substrate, hopefully with an efficiency comparable to those of ribozymes or RNases. Currently, three groups have reported catalytic RNA cleavage, and in all cases metal complex-linked oligonucleotides have been used as the cleavers (14–16). Magda and co-workers demonstrated for the first time that a DNA oligomer with a texaphyrin· dysprosium(III) complex at the internal site can cleave RNA with multiple turnover under physiological conditions (14). The scheme of the catalytic RNA cleavage cycle, which is illustrated in their paper, is based on the following concept. If an antisense cleaver can cleave RNA at the internal site of the RNA hybrid-forming region, then release of the cleaved fragments from the hybrid will occur easily, as compared to product release in the case of the RNA cleavage of a single-stranded region adjoined to the hybrid. In addition, the freed cleaver will be able to bind and cleave additional RNA substrate. It is known that an RNA single strand is cleaved by some metal ions and complexes, but the RNA in a DNA·RNA hybrid is resistant to cleavage (17). These metal complex cleavers have been designed to overcome this problem.

Recently, we have shown that an antisense 2′-O-methyloligonucleotide, with a terpyridine·Cu(II) complex directly attached to the 5′-oxygen of the 5′-end nucleoside, cleaves RNA predominantly at the site opposite the 5′-end under physiological conditions (18–20). Figure 1 shows the structures of terpyridine-linked 2′-O-methyladenosine (tAm) and 2′-O-methyluridine (tUm) at the 5′-end. Furthermore, we have developed an effective approach to promote RNA cleavage by the 5′-conjugate. When a 3′-conjugate was used together with the 5′-conjugate in a tandem fashion, the RNA cleavage efficiency was greatly enhanced (20): the 3′-end nucleoside was a uridine derivative (Ut) with terpyridine linked to the 2′-oxygen via a short linker arm (Fig. 1A). Since no cleavage was observed using the 3′-conjugate alone, a cooperative cleavage action of the two complexes was strongly suggested (the full details of the RNA cleavage by the tandem cleavage system will be reported separately). In the tandem system, the forming RNA hybrid can bend or rotate at the phosphodiester of the nicked site, and this flexibility will affect the cleavage activity. Thus, we envisioned that if the two terpyridine·Cu(II) complexes were put into a relatively rigid system, then the system would gain higher cleavage activity by their more effective cooperative action.

Figure 1.

(A) Cleavage system used in our previous study and structures of terpyridine-linked nucleoside residues at the 3′- and 5′-ends of the conjugates, respectively. (B) Design for connecting the conjugates and the structure of the phosphoramidite unit (1) prepared for this purpose.

In this study, we have constructed a series of 2′-O-methyloligonucleotides with two contiguous terpyridine· Cu(II) complexes at the internal site and tested their ability to cleave the target RNA 24mers. These conjugates were designed to connect the aforementioned 3′- and 5′-conjugates with different lengths of a flexible linker and to target the central position of the duplex-forming region (Fig. 1B). To prepare these conjugates, we used a newly synthesized phosphoramidite unit of the deoxyuridine derivative 1. The most active conjugate in the series cleaved the target with multiple turnover and high site specificity. Combining these results with those from other investigations, we propose the possible roles of the two Cu(II) complexes in the RNA cleavage.

MATERIALS AND METHODS

General methods

¹H- and ³¹P-NMR spectra were recorded on a JEOL α-400 (400 MHz) spectrometer, with tetramethylsilane or trimethylphosphate as an internal standard, and are reported in p.p.m. from the standards, respectively. UV absorption spectra were recorded with a Beckman DU-640 spectrophotometer. HPLC analysis and purification of oligonucleotides were done on a Shimadzu HPLC system. 5-Iodo-2′-deoxyuridine was obtained from Yamasa Shoyu Co. The common solvents and reagents for the DNA/RNA synthesizer (Applied Biosystems 392) were purchased from Applied Biosystems and the phosphoramidite units for RNA and 2′-O-methyl RNA synthesis and the unit for the 1,3-propanediol linker were from Glen Research.

3′-O-acetyl-5′-O-t-butyldimethylsilyl-5-(3-hydroxy-1-propynyl)-2′-deoxyuridine (3)

The starting material, 3′-O-acetyl-5′-O-t-butyldimethylsilyl-5-iodo-2′-deoxyuridine 2, was easily obtained from 5-iodo-2′-deoxyuridine by common methods, which included the reaction with t-butyldimethylsilyl choride (TBDMSCl), followed by reaction with acetic anhydride. Compound 2 (5.0 g, 9.8 mmol) was dissolved in DMF (60 ml). To this solution, 2-propyn-1-ol (1.7 ml, 19.6 mmol), triethylamine (2.7 ml, 19.6 mmol), (Ph₃P)₂PdCl (387 mg, 0.49 mmol) and CuI (187 mg, 0.98 mmol) were added, and the mixture was stirred under an Ar atmosphere at room temperature. After 2.5 h, the solvent was evaporated and the residue was partitioned between CHCl₃ and an aqueous 2% Na₂ EDTA solution. The organic layer was washed with water, dried (Na₂SO₄) and concentrated to dryness. The residue was purified by silica gel column chromatography with 0–1.5% MeOH–CHCl₃ solutions to give a light yellow foam of compound 3 (2.8 g, 66% yield). FAB-MS (LR) m/z 439 (MH⁺); FAB-MS (HR) calculated for C₂₀H₃₁N₂O₇Si 439.1900, found 439.1899; ¹H NMR (DMSO-d₆) δ 11.61 (s, 1H, NH), 7.92 (s, 1H, 6-H), 6.13 (dd, 1H, 1′-H), 5.19–5.17 (2H, 3′-H and 5-C≡C-CH₂-OH), 4.21 (d, 2H, 5-C≡C-CH₂-OH), 4.10 (m, 1H, 4′-H), 3.83 (m, 2H, 5′-H), 2.38–2.19 (m, 2H, 2′-H), 2.06 (s, 3H, -COCH₃), 0.90 (s, 9H, TBDMS-t-butyl), 0.12, 0.11 (each s, each 3H, TBDMS-dimethyl).

3′-O-acetyl-5′-O-t-butyldimethylsilyl-5-(3-hydroxypropyl)-2′-deoxyuridine (4)

To a solution of 3 (7.0 g, 16.0 mmol) in AcOEt (280 ml), Pd/C (4.9 g) was added under an Ar atmosphere. The mixture was stirred for 5 h under a H₂ atmosphere at room temperature and was filtered to remove the Pd/C. The filtrate was concentrated in vacuo and the residue was purified by silica gel column chromatography with 0–6% MeOH–CHCl₃ solutions to give a yellow solid of compound 4 (5.93 g, 13.4 mmol, 83.8% yield). FAB-MS (LR) m/z 443 (MH⁺) FAB-MS (HR) calculated for C₂₀H₃₅N₂O₇Si 443.2213, found 443.2203; ¹H NMR (DMSO-d₆) δ 11.26 (s, 1H, NH), 7.41 (s, 1H, 6-H), 6.15 (dd, 1H, 1′-H), 5.18 (m, 1H, 3′-H), 4.33 (t, 1H, 5-(CH₂)₃-OH), 4.02 (m, 1H, 4′-H), 3.82 (m, 2H, 5′-H), 3.39 (dd, 2H, 5-CH₂CH₂CH₂OH), 2.27–2.21 (4H, 2′-H and 5-CH₂CH₂CH₂OH), 2.06 (s, 3H, -COCH₃), 1.57 (m, 2H, 5-CH₂CH₂CH₂OH), 0.89 (s, 9H, TBDMS-t-butyl), 0.09 (s, 6H, TBDMS-dimethyl).

5′-O-t-butyldimethylsilyl-3′-O-dimethoxytrityl-5-(3-dimethoxytrityloxy-propyl)-2′-deoxyuridine (5)

Compound 4 (5.93 g, 13.4 mmol) was dissolved in saturated methanolic ammonia (180 ml) and was stirred for 15 h at room temperature. The mixture was concentrated in vacuo and the residue, after adsorption to small amounts of silica gel, was purified by silica gel column chromatography with 0–6% MeOH–CHCl₃ solutions to give a light yellow gum of the deacetylated compound of 4 (5.09 g, 95% yield). Selected physical data for this compound: FAB-MS (LR) m/z 401 (MH⁺); FAB-MS (HR) calculated for C₁₈H₃₃N₂O₆Si 401.5590, found 401.2088. After successive co-evaporations with pyridine, the compound (5.08 g, 12.7 mmol) was dissolved in pyridine (130 ml). 4,4′-Dimethoxytrityl chloride (DMTrCl, 12.9 g, 38.1 mmol) was added to the solution and the mixture was stirred at 60°C for 24 h, and then MeOH (10 ml) was added. The solvent was evaporated to a quarter volume and, after the resulting residue was diluted with CHCl₃, the whole solution was washed with a saturated NaHCO₃ aqueous solution. The organic layer was dried (Na₂SO₄) and the solvent was evaporated in vacuo. The resulting mixture was purified by silica gel column chromatography with 10–40% AcOEt–hexane solutions to give a yellow foam of compound 5 (10.84 g, 10.8 mmol, 85% yield). FAB-MS (LR) m/z 1005 (MH⁺); FAB-MS (HR) calculated for C₆₀H₆₉N₂O₁₀Si 1005.4721, found 1005.4710; ¹H NMR (DMSO-d₆) δ 11.22 (s, 1H, NH), 7.42–6.82 (27H, DMTr-aromatic and 6-H), 6.16 (dd, 1H, 1′-H), 4.14 (m, 1H, 3′-H), 3.95 (m, 1H, 4′-H), 3.73, 3.72 (each s, each 6H, DMTr-dimethoxy), 3.52–3.27 (m, 2H, 5′-H), 2.92–2.89 (m, 2H, 5-CH₂CH₂CH₂O-), 2.17 (m, 2H, 5-CH₂CH₂CH₂O-), 1.68–1.52 (4H, 5-CH₂CH₂CH₂O- and 2′-H), 0.71 (s, 9H, TBDMS-t-butyl), –0.13, –0.15 (each s, each 3H, TBDMS-dimethyl).

3′-O-dimethoxytrityl-5-(3-dimethoxytrityloxy-propyl)-5′-O-terpyridyl-2′-deoxyuridine (6)

To a solution of 5 (10.8 g, 10.8 mmol) in THF (100 ml), tetrabutylammonium fluoride (1 M THF solution, 13.0 ml) was added. The mixture was stirred for 17 h at room temperature and then the solvent was evaporated in vacuo. The residue was partitioned between CHCl₃ and water and the organic layer was concentrated to dryness. The residue, after adsorption to small amounts of silica gel, was purified by silica gel column chromatography with 30–60% AcOEt–hexane solutions to give a light yellow foam of the desilylated compound of 5 (8.7g, 9.8 mmol, 91% yield). Selected physical data for this compound: FAB-MS (LR) m/z 891 (MH⁺); FAB-MS (HR) calculated for C₅₄H₅₅N₂O₁₀ 891.3856, found 891.3870. To a solution of the compound (3.1 g, 3.5 mmol) in DMSO (35 ml), finely crushed KOH (2.0 g) and 4′-chloro-2,2′:6′,2″-terpyridine (1.2 g, 4.2 mmol) were added and the mixture was stirred at room temperature. After 92 h, the mixture was filtered to remove the KOH and the filtrate was concentrated to a half volume and dissolved in AcOEt (350 ml). The solution was washed with a saturated aqueous NaCl solution and the organic layer was dried (Na₂SO₄) and concentrated to dryness. The residue, after adsorption to small amounts of alumina, was purified by alumina column chromatography with 20–50% AcOEt– hexane solutions and subsequent 0–3% MeOH–CHCl₃ solutions to give a white foam of 6 (3.92 g, 3.48 mmol, 99% yield). FAB-MS (LR) m/z 1122 (MH⁺); FAB-MS (HR) calculated for C₆₉H₆₄N₅O₁₀ 1122.4653, found 1122.4680; ¹H NMR (DMSO-d₆) δ 11.22 (s, 1H, NH), 8.60 (d, 2H, terpy-6-H and terpy-6″-H), 8.54 (d, 2H, terpy-3-H and terpy-3″-H), 8.30 (s, 2H, terpy-3′-H and terpy-5′-H), 7.98–7.94 (m, 2H, terpy-4-H and terpy-4″-H), 7.46–6.72 (29H, DMTr-aromatic, terpy-5-H, terpy-5″-H, and 6-H), 6.26 (dd, 1H, 1′-H), 4.40 (m, 1H, 3′-H), 4.12–3.83 (3H, 4′-H and 5′-H), 3.68, 3.66 (each s, each 6H, DMTr-dimethoxy), 2.75 (m, 2H, 5-CH₂CH₂CH₂O-), 2.16– 2.11 (m, 2H, 5-CH₂CH₂CH₂O-), 1.95–1.88 (m, 2H, 2′-H), 1.54 (m, 2H, 5-CH₂CH₂CH₂O-).

5-(3-Dimethoxytrityloxy-propyl)-5′-O-terpyridyl-2′-deoxyuridine (7)

A solution of 6 (90 mg, 0.080 mmol) in 80% aqueous acetic acid (10 ml) was stirred at room temperature. After 3 h, the solvent was evaporated and the residue was co-evaporated with water several times to remove the acetic acid, then dissolved in MeOH. The solution was concentrated to leave a light purple powder of the detritylated compound of 6 (27 mg, 65% yield). Selected physical data for this compound: FAB-MS (LR) m/z 518 (MH⁺); FAB-MS (HR) calculated for C₂₇H₂₈N₅O₆ 518.2039, found 518.2026. The title compound 7 was prepared as follows. Compound 6 (2.0 g, 1.8 mmol) was treated with 80% aqueous acetic acid for 24 h and was co-evaporated with EtOH several times. Without further purification, the residue was co-evaporated with pyridine three times and was dissolved in pyridine (30 ml). To this solution, DMTrCl (1.8 g, 5.4 mmol) was added and the mixture was stirred at room temperature. After 5 h, MeOH (5 ml) was added and the mixture was stirred for 10 min. The solution was concentrated to a quarter volume and was partitioned with AcOEt and a saturated aqueous NaHCO₃ solution. The organic layer was dried with Na₂SO₄ and was concentrated to dryness. The residue was purified by alumina column chromatography with 0–25% MeOH–CHCl₃ solutions to give a white foam of 7 (1.16 g, 1.4 mmol, 78% yield). FAB-MS (LR) m/z 820 (MH⁺); FAB-MS (HR) calculated for C₄₈H₄₆N₅O₈ 820.3346, found 820.3336; ¹H NMR (DMSO-d₆) δ 11.28 (s, 1H, NH), 8.61 (d, 2H, terpy-6-H and terpy-6″-H), 8.55 (d, 2H, terpy-3-H and terpy-3″-H), 8.01 (s, 2H, terpy-3′-H and terpy-5′-H), 7.98–7.94 (m, 2H, terpy-4-H and terpy-4″-H), 7.46–6.74 (16H, DMTr-aromatic, terpy-5-H, terpy-5″-H, and 6-H), 6.24 (dd, 1H, 1′-H), 5.41 (d, 1H, 3′-OH), 4.43 (m, 3H, 3′-H, and 5′-H), 4.16 (m, 1H, 4′-H), 3.68 (s, 6H, DMTr-dimethoxy), 2.80–2.76 (m, 2H, 5-CH₂CH₂CH₂O-), 2.24–2.15 (m, 4H, 5-CH₂CH₂CH₂O- and 2′-H), 1.61–1.57 (m, 2H, 5-CH₂CH₂CH₂O-).

5′-O-terpyridyl-5-(3-dimethoxytrityloxy-propyl)-2′-deoxyuridine 3′-O-(2-cyanoethyl) N,N-diisopropyl-phosphoramidite (1)

Compound 7 (150 mg, 018 mmol) was co-evaporated with pyridine twice and was dissolved in CH₂Cl₂ (5 ml). To this solution, N,N-diisopropylethylamine (75 µl, 0.43 mmol) and 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (50 µl, 0.22 mmol) were added and the mixture was stirred at room temperature under an Ar atmosphere. After 2 h, the mixture was diluted in CH₂Cl₂ (20 ml) and washed with a saturated aqueous NaHCO₃ solution twice and with a saturated aqueous NaCl solution. After the organic phase was filtered with 1ps filter paper (Whatman), the filtrate was concentrated to dryness. The residue was purified by alumina column chromatography with 0–1.0% MeOH–CHCl₃ solutions to give a white foam of 1 (122 mg, 67% yield). ³¹P-NMR (CDCl₃) δ 146.75, 146.43.

5′-O-t-butyldimethylsilyl-5-(3-dimethoxytrityloxy-propyl)-2′-deoxyuridine (8)

Compound 4 (1.4 g, 3.1 mmol) was co-evaporated with pyridine twice and was dissolved in pyridine (30 ml). To this solution, DMTrCl (1.3 g, 3.7 mmol) was added and the mixture was stirred at room temperature. After 3 h, MeOH was added and the solution was stirred for 10 min. The mixture was concentrated to a quarter volume, diluted with CHCl₃ (50 ml), washed with a 2% aqueous NaHCO₃ solution twice and then washed with water. The organic layer was dried (Na₂SO₄) and was concentrated to dryness. The residue, after adsorption to a small amount of silica gel, was purified by silica gel column chromatography with 30–65% AcOEt–hexane solutions containing 0.1% pyridine to give a white foam of the tritylated compound of 4 (1.8 g, 2.4 mmol, 79% yield). Selected physical data for this compound: FAB-MS (LR) m/z 744 (M⁺). The compound (1.8 g, 2.4 mmol) was dissolved in NH₃ saturated MeOH (50 ml) and was stirred at room temperature overnight. The solvent was evaporated and the residue was diluted in CHCl₃ (40 ml). The mixture was washed with water twice and with a saturated aqueous NaCl solution. The organic layer was dried with Na₂SO₄ and was concentrated to dryness. The residue was purified by silica gel column chromatography with 0–2% MeOH–CHCl₃ solutions to give a white foam of 8 (1.5 g, 2.2 mmol, 90% yield). FAB-MS (LR) m/z 725 (MNa⁺); FAB-MS (HR) calculated for C₃₉H₅₀N₂O₈SiNa 725.3234, found 725.3256; ¹H NMR (DMSO-d₆) δ 11.28 (s, 1H, NH), 7.37–6.85 (14H, DMTr-aromatic and 6-H), 6.15 (dd, 1H, 1′-H), 5.24 (d, 1H, 3′-OH), 4.17 (m, 1H, 3′-H), 3.78–3.66 (9H, DMTr-dimethoxy, 4′-H, and 5′-H), 2.94 (m, 2H, 5-CH₂CH₂CH₂O-), 2.16–2.13 (m, 2H, 5-CH₂CH₂CH₂O-), 2.07–1.91 (m, 2H, 2′-H), 1.74 (m, 2H, 5-CH₂CH₂CH₂O-), 0.83 (s, 9H, TBDMS-t-butyl), 0.03 (s, 6H, TBDMS-dimethyl).

5′-O-t-butyldimethylsilyl-5-(3-dimethoxytrityloxy-propyl)-2′-deoxyuridine 3′-O-(2-cyanoethyl N,N-diisopropyl-phosphoramidite) (9)

Compound 8 (150 mg, 0.21 mmol) was co-evaporated with pyridine twice and was dissolved in CH₂Cl₂ (2 ml). To this solution, N,N-diisopropylethylamine (89 µl, 0.51 mmol) and 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (58 µl, 0.26 mmol) were added and the mixture was stirred at room temperature under an Ar atmosphere. After 30 min, the mixture was diluted with CH₂Cl₂ (15 ml) and was washed with a 2% aqueous NaHCO₃ solution twice and with a saturated aqueous NaCl solution. The organic layer was dried with Na₂SO₄ and was concentrated to dryness. The residue was purified by silica gel column chromatography with 50% AcOEt–hexane containing 0.1% pyridine to give a white foam of compound 9 (0.11 g, 0.12 mmol, 58% yield). ³¹P-NMR (CDCl₃) δ 145.96, 146.01.

Synthesis of oligonucleotides

Oligonucleotides were synthesized on the 1 µmol scale. The coupling time for the phosphoramidite derivatives of the unnatural nucleosides and the propanediol linker was 25 min (10 min for 2′-O-methyl RNA). Fully protected oligonucleotides were removed from the CPG supports with 28% ammonium hydroxide and were heated at 60°C for 8 h. The solutions were concentrated in vacuo and were purified on a C-18 open column (Preparative C18; Waters) (50 mM triethylammonium acetate with a linear gradient of 5–40% acetonitrile). The fractions containing the oligonucleotides with the dimethoxytrityl group were concentrated and then the protecting group was removed with 80% acetic acid. Finally the oligonucleotides were purified by reversed phase HPLC using an ODS column (Puresil C18 5µ or µBondasphere C18 5µ; Waters). The structures of the series of PL₁-conjugates (Fig. 5A) were confirmed by electrospray ionization mass spectrometry using a Quattro mass spectrometer (Micromass). The agent for Entry 1, calculated mass 6690.74 (C₂₇₀H₂₈₀N₇₈O₁₂₇P₁₈), observed mass 6690.48; the agent for Entry 2, calculated mass 6429.45 (C₂₁₁H₂₆₉N₇₅O₁₂₆P₁₈), observed mass 6429.26; the agent for Entry 3, calculated mass 6459.48 (C₂₁₂H₂₇₁N₇₅O₁₂₇P₁₈), observed mass 6459.18.

Figure 5.

(A) Reaction of the labeled RNA 24mer with the PL₁-conjugate (Entry 1) and the related reactions as controls (Entries 2–4). The reaction conditions are described in the legend to Figure 3, except 1 µM CuCl₂ was used for Entries 2 and 3. Arrows indicate the cleavage sites and the cleavage yields are also shown. (B) Autoradiogram of products from the reactions of Entries 1–4. Lanes 1 and 2, bicarbonate and RNase T1 sequencing reactions, respectively; lanes 3 and 4, incubation of the RNA in the absence or presence of CuCl₂ (2 µM), respectively; lanes 5–8, the reactions of Entries 1–4, respectively.

RNA cleavage with an excess of conjugate(s)

5′-End-labeling of RNA substrates (20 pmol) was carried out using T4 polynucleotide kinase and [γ-³²P] ATP (ICN) for 1.5 h at 37°C, in a buffer containing 5 mM Tris–HCl (pH 7.6), 1 mM MgCl₂ and 1 mM 2-mercaptoethanol (final concentrations). After purification with a disposable reversed phase column (YMC Dispo SPE), the labeled RNA was mixed with non-labeled RNA (180 pmol) for the cleavage reaction. A solution of the RNA substrate in water was heated at 85°C for 2 min and then was cooled in an ice bath. A mixture of the solution of RNA (0.1 µM) and the cleavage agent(s) (1 µM), in a reaction buffer containing 100 mM NaClO₄ and 20 mM HEPES–NaOH (final concentrations), was preincubated at 37°C for 10 min. After the addition of CuCl₂ (1 or 2 µM final concentration) in water to start the cleavage reaction, the mixture was incubated at 37°C. Aliquots (2 µl) were taken at specific times and loading buffer was added to quench the reaction. The mixtures, along with the products of bicarbonate and RNase T1 sequencing reactions, were fractionated by PAGE on a denaturing 20% gel containing 8 M urea. Quantification of the radioactivity was performed with a bioimaging analyzer (Fujix BAS1000). Autoradiography of the gel was done at –80°C using Fuji Medical X-ray Film (HR-S 30) with an intensifying screen.

RNA cleavage with an excess of RNA substrate

The RNA substrate in water was heated at 85°C for 2 min and then was cooled in an ice bath. The cleavage reaction was initiated by mixing the solution of RNA (500 nM) with the cleavage agent(s) (50 nM), which were premixed with CuCl₂ (50 or 100 nM) in a reaction buffer containing 100 mM NaClO₄ and 20 mM HEPES–NaOH (final concentrations). The mixture was incubated at 37°C and aliquots (2 µl) were taken at specific times. The subsequent experiments were done under conditions similar to those for the cleavage reactions using excess conjugate(s).

RESULTS AND DISCUSSION

Synthesis of two terpyridine-linked 2′-O-methyloligonucleotides

For the preparation of the two terpyridine-attached 2′-O-methyloligonucleotides, we newly synthesized a dimethoxytrityl derivative of the 5-hydroxypropyl-5′-O-terpyridyl-2′-deoxyuridine-3′-phosphoramidite 1. The synthetic route of the phosphoramidite unit 1 is shown in Scheme 1. The starting material 2, easily prepared from 2′-deoxy-5-iodouridine, was converted to the 5-propynol derivative 3 by the palladium-catalyzed coupling of 2-propyn-1-ol (21). Catalytic hydrogenation of 3 gave the 5-propanol derivative 4, which was deacetylated and di-dimethoxytritylated to afford the fully protected compound 5. Desilylation and subsequent reaction with 4′-chloro-2,2′:6′,2″-terpyridine under basic conditions (22) gave the 5′-O-terpyridyl derivative 6. Deprotection of the two dimethoxytrityl groups of 6, followed by selective mono-dimethoxytritylation, gave the compound 7 with a free 3′-hydroxyl group. Phosphitylation of 7 by the standard method afforded the desired compound 1. In addition, we also synthesized the phosphoramidite unit 9, lacking the terpyridine group, from the intermediate 4, and used it for the preparation of control oligonucleotides.

Scheme 1. Synthetic route of the phosphoramidite units 1 and 9 (all reactions were carried out at room temperature unless otherwise noted). (a) 2-propyn-1-ol, (Ph₃P)₂PdCl₂, CuI, Et₃N, DMF, 2.5 h; (b) H₂, Pd/C, CH₃COOEt, 5 h; (c) NH₃/MeOH, 15 h; (d) DMTrCl (3 equiv.), pyridine, 60°C, 24 h; (e) n-Bu₄NF, THF, 17 h; (f) 4′-chloro-2,2′:6′,2″-terpyridine, KOH, DMSO, 92 h; (g) 80% CH₃COOH aqueous solution, 3 h; (h) DMTrCl (3 equiv.), pyridine, 5 h; (i) i-Pr₂NP(Cl)OCH₂CH₂CN, i-Pr₂NEt, CH₂Cl₂, 2 h; (j) DMTrCl (1.2 equiv.), pyridine, 3 h; (k) NH₃/MeOH, overnight; (l) i-Pr₂NP(Cl)OCH₂CH₂CN, i-Pr₂NEt, CH₂Cl₂, 30 min.

Using the phosphoramidite unit 1 and the phosphoramidite unit for the introduction of the Ut residue, we synthesized 2′-O-methyloligonucleotides with a Ut residue at the 5′-side and a 5-hydroxypropyl-5′-O-terpyridyl-2′-deoxyuridine (tU^L) residue at the 3′-side. Figure 2 shows the sequences of the conjugates and the target RNA 24mers (5′-end-labeled with ³²P) and the partial structures of the conjugates complexed with Cu(II). The propanediol(s) linker (PL_n) was inserted into the unusual phosphodiester bond between Ut and tU^L. The RNA substrates were the RNAs containing a U or A residue at position 12 (RNA-U12 or RNA-A12), and the RNA cleavage reactions were mainly carried out using RNA-U12.

Figure 2.

Sequences and structures of 2′-O-methyloligonucleotides with two terpyridine·Cu(II) complexes at the internal site and sequences of the target RNA 24mers.

RNA cleavage using an excess of the complex(es) conjugates and determination of the pH dependence of the activities

Four conjugates with PL_n (n = 0–3) were used to investigate the effect of the linker length on the RNA cleavage activity. Reactions of the target RNA-U11 (0.1 µM) with a 10-fold excess of the PL_n-conjugates, in the presence of an equivalent molar amount of Cu(II) ion to each terpyridine group, were carried out at pH 7.5 and 37°C for 5 h.

An analysis of the cleavage products by denaturing PAGE is shown in Figure 3A. As compared with the bicarbonate and RNase T1 sequencing reactions, the cleavages occurred between U12 and A13, where U12 was opposite the tU^L residue, regardless of the PL_n length. On the other hand, the cleavage yields (%) depended on the PL length: ∼80% yields for PL_0–2 (lanes 3–5); 41.5% yield for PL₃ (lane 6). The time course of the cleavage yields indicated that the PL₁-conjugate cleaved the RNA most rapidly, and the yield reached a plateau by 3 h (Fig. 3B). This high activity was confirmed as follows. From calculations based on the amounts of the remaining substrate, we determined the pseudo-first order rate constants (k_obs) for these reactions: 0.386 h^–1 for PL₀; 0.833 h^–1 for PL₁; 0.385 h^–1 for PL₂; 0.107 h^–1 for PL₃.

Figure 3.

(A) Autoradiogram of products from the reactions of the labeled target RNA 24mer with the PL_n-conjugates after 20% denaturing PAGE and (B) time course for cleavage reactions. Each reaction was carried out at 37°C for 5 h in a total volume of 30 µl containing: 20 mM HEPES–NaOH buffer (pH 7.5), 100 mM NaClO₄, 0.1 µM RNA, 1 µM conjugate and 2 µM CuCl₂. Lanes 1 and 2, bicarbonate and RNase T1 sequencing reactions, respectively; lanes 3–6, reactions with the PL₀-, PL₁-, PL₂- and PL₃- conjugates, respectively.

We used 2 molar equivalents of Cu(II) ions to the conjugates in the above cleavage experiments. This was based on the result of the following experiment. Examination of the optimal ratio of Cu(II) ions to the PL₁-conjugate for the activity revealed that ∼2 equiv. of Cu(II) to the conjugate were needed to obtain the highest cleavage yield, which remained almost unchanged up to 20 equiv. (Fig. 4). The activities shown by the cleavage yields of the 2.5 h reaction may exclusively reflect the high activity of the PL₁-conjugate with the two metal complexes, because the PL₁-conjugate with the mono-metal complex of Ut or tU^L was almost inert even in a 5 h reaction, as shown in Figure 5. It is interesting that the activity of the agent did not appear until ∼0.75 equiv. of Cu(II) were present. Although the exact origin of the sigmoid curve remains unclear, we speculate that the curve was obtained by increasing the amounts of the PL₁-conjugate with the uniform (or non-uniform) mono-metal complex, which preferentially forms, after which the gradual formation of the PL₁-conjugate with two Cu(II) ions probably occurs.

Figure 4.

(A) Autoradiogram of products from the reactions of the labeled RNA 24mer with the PL₁-conjugate in the presence of 0–20 equiv. of Cu(II) to the conjugate and (B) plot of the cleavage yield (%) at each Cu(II) concentration, [Cu²⁺]. The reaction conditions are described in the legend to Figure 3, except the reactions were carried out for 2.5 h. The concentration of the PL₁-conjugate, [PL₁-conjugate], was 1 µM and [Cu²⁺] ranged from 0 to 20 µM.

Next, to clarify the possible roles of the two metal complexes within the PL₁-conjugate, we examined the effects of the attachment position and the number of metal complexes on the cleavage activity. Entry 1 in Figure 5A shows the reaction of the PL₁-conjugate with RNA-U12, and the control reactions are shown in Entries 2–4. All of the reactions were performed under the conditions described above. An autoradiogram of the cleavage products of the reactions shows that the observed cleavage site was between U12 and A13 (Fig. 5B). Entry 1 gave a high cleavage yield (81.9%), whereas the single terpyridine agent with tU^L in Entry 2 was much less active. In the case of Entry 3, using one terpyridine agent with Ut, no cleavage was observed. Although the cleavage yields of Entries 2 (3.7%) and 3 (0%) for the 5 h reaction were too low to be compared with each other, for a 25 h reaction the yields were 14.4 and <1%, respectively. These data and the previous results for the tandem cleavage system (described in the Introduction) would lead to the conclusion that in the two metal complexes systems the tUm or tU^L complex is essential for the cleavage, and is assisted by the Ut complex. Thus, the cooperative action of the two complexes should certainly be involved in the cleavage reactions. On the other hand, Entry 4, which contained tandem terpyridine-linked 2′-O-methyloligonucleotides, gave a cleavage yield of 58.5%. This result demonstrates that the PL₁-conjugate with the two metal complexes augments the cleavage efficiency of the tandem system. The comparison of the activities of Entries 1, 2 and 4 is based on their kinetic data (at pH 7.5), provided in Figure 6, which indicated that the PL₁-conjugate with two Cu(II) ions was about 130 times and 3 times more reactive than the mono-metal agent and the tandem systems, respectively.

Figure 6.

pH–rate profile for the cleavage reactions of Entries 1, 2 and 4. The reaction conditions are described in the legend to Figure 3, except 1 µM CuCl₂ was used for Entry 2. The buffers (20 mM) were MES–NaOH (pH 6.0 or 6.5), HEPES–NaOH (pH 7.0, 7.5 or 8.0) and TAPS–NaOH (pH 8.5 or 9.0).

We also tested RNA-A12 as a substrate in the above reaction. In this case, the A12 residue was opposite tU^L in the forming hybrid, and thus base pairing was possible. Under the conditions shown in Figure 3, cleavage of RNA-A12 with the PL₁-conjugate occurred at the A12–A13 site, and the yield was 15.9%. RNA-U12 may be more susceptible to cleavage due to the absence of base pairing at U12, which may relax the structure around the active site in the forming hybrid, thus allowing the spatial orientations of the two complexes to the target phophodiester to become more suitable. As an intrinsic property of RNA, the U–A sequence is more susceptible to cleavage via transesterification than the A–A sequence (23,24). This property may also be an additional factor.

We next investigated the pH dependence of the cleavage activities of the agents used in Entries 1, 2 and 4, to clarify the cooperative action of the two metal complexes. The cleavage reactions were carried out within the pH range 6.0–9.0 and at 37°C and the pseudo-first order rate constants were determined. The pH–rate profiles, as shown in Figure 6, reveal that Entry 1, with the PL₁ conjugate, gave a typical bell-shaped curve with a maximum rate constant at approximately pH 7.5, which was 2.3 and 3.4 times larger than those at pH 6.0 and 9.0, respectively. On the other hand, the cleavage activity for Entry 2 was enhanced as the pH value increased, but was still much lower than that of Entry 1. The profile for Entry 4 was not a bell-shaped curve, but a somewhat hill-shaped curve. Since the rate constants around pH 7.5 were higher than those at pH 6.0 and 9.0, we suppose that the reaction occurs by a cleavage mechanism similar to that for the reaction of Entry 1.

A bell-shaped profile is a salient feature of the cooperative action of general acid and base catalysts. For example, the catalysis by RNase A and its homolog with two histidine-imidazole residues as the catalysts exhibits this profile (25). Thus, we propose a transesterification mechanism for RNA cleavage of the target phosphodiester moiety by the PL₁-conjugate complexed with Cu(II). Specifically, a hydroxide ion coordinated to the Cu(II) complex of the tU^L residue and a H₂O coordinated to the corresponding Ut residue work cooperatively, as general base and acid catalysts, respectively. It was proposed that a hydroxide ion coordinated to a terpyridine·Cu(II) complex acts as a general base catalyst for the deprotonation of the sugar 2′-hydroxyl in the initial transesterification step (26). The pK_a value for the H₂O-bound terpyridine·Cu(II) is 8.08 (27). Although we have not determined the pK_a values of the alkoxyterpyridine complexes in the PL₁-conjugate, the values for each complex seemed to be around 7.5, as deduced from the bell-shaped profile.

The role of the tU^L complex as the base is derived from the pH–rate profile for Entry 2: the higher pH afforded higher cleavage activity. On the other hand, to clarify the role of the Ut complex, we examined the pH dependence of the rate enhancement by the Ut complex in the PL₁-conjugate complexed with two Cu(II) ions relative to the conjugate with the tU^L complex alone. As shown in Figure 7, the data indicate that the higher pH afforded lower cleavage activity, supporting acid catalysis of the Ut complex. Thus, we suppose that the acid catalyst delivers a proton to the phosphodiester (e.g. the 5′-oxygen) and this event facilitates the departure of the 5′-hydroxyl group. Although our mechanism follows the well-known mechanism for RNase A, the possibility that the Cu(II) complex(es) may also participate as a Lewis acid(s) to the phosphodiester cannot presently be ruled out.

Figure 7.

pH dependence of the rate enhancement by the Cu(II) complex of the Ut residue. The relative value at each pH was obtained from the equation for the rate constant for Entry 1/the rate constant for Entry 2: the rate constants at various pH values are shown in Figure 6.

Cleavage reactions with an excess of the RNA substrate and determination of the kinetic parameters

As mentioned above, we demonstrated that the PL₁-conjugate complexed with Cu(II) has high RNA cleavage activity. To test the catalytic cleavage activity of this agent, the reactions of Entry 1 and the related Entries 2 and 4 were carried out under conditions using a 10-fold excess of the target RNA (500–50 nM conjugates), at pH 7.5 and 37°C for 24 h. The time course of the reactions (Fig. 8) indicated that Entry 1 cleaved an equimolar amount of the RNA by as early as 1 h, and the cleavage yield achieved 88% after 15 h. In the cases of Entries 2 and 4, however, the cleavage yields did not reach 10%, even after 24 h. Thus, RNA cleavage with catalytic turnover was observed only with Entry 1.

Figure 8.

Time course for cleavage reactions under the conditions of an excess of the RNA. Each reaction was carried out at 37°C for 24 h in a total volume of 30 µl containing 20 mM HEPES–NaOH buffer (pH 7.5), 100 mM NaClO₄, 500 nM RNA, 50 nM conjugate(s) and 100 nM CuCl₂ (for Entry 2, 50 nM CuCl₂).

We further examined the cleavage efficiency of Entry 2 under different substrate excess conditions, in which the concentration of the conjugate was maintained at 2 nM and that of the target RNA was changed from 2 to 20 nM. The initial rates of each reaction gave a Michaelis–Menten constant K_m of 5.49 (nM) and a turnover rate constant k_cat of 0.647 h^–1, and thus the second order rate constant, k_cat/K_m = 0.118 nM^–1 h^–1, was obtained.

There are three precedents for RNA cleavage with multiple turnover by fully synthetic antisense cleavers, which all contained one metal complex. Magda’s group (14) and Bashkin’s group (16) used DNA oligomers with a texaphyrin· dysprosium(III) complex and a terpyridine·Cu(II) complex at the central position, respectively: the metal complexes were linked to an abasic site analog within the DNA. On the other hand, the RNA cleaver used by Häner’s group was a 2′-O-(methoxyethyl) RNA oligomer with a macrocyclic lanthanide complex at the 5′-end (15). In this case, the target RNA formed a bulge upon hybrid formation with the cleaver and the bulge was the predetermined site for cleavage. We compared the cleavage efficiency of our Cu(II) complex of the PL₁-conjugate with those of the other three cleavers. Only Magda’s group furnished the kinetic parameters, and an examination of these and other data published by the three groups indicated that the cleavage efficiency of our designed conjugate was the highest, even considering that the reaction conditions and the target RNA sequences differed from each other. The turnover numbers (TN) observed under conditions using a 10-fold excess of substrates are as follows: for our agent, TN = 8.8 (pH 7.5, 37°C, 15 h); for Magda’s agent, TN = 6.7 (pH 7.5, 37°C, 24 h); for Bashkin’s agent, TN = 6.7 (pH 7.4, 45°C, 40 h); for Häner’s agent, TN = 8.5 (pH 7.2, 37°C, 64 h). A comparison of k_cat/K_m values revealed that the value of our agent was ∼8-fold greater than the value (0.0143 nM^–1 h^–1) of Magda’s agent. In addition, our agent cleaved the RNA at one site, whereas the other agents, which had a relatively long linker and/or a flexible abasic site, cleaved the RNA at a few or several contiguous sites. The enhanced site specificity of our agent may be a consequence of each metal complex being attached to a nucleoside residue either directly or via a short linker, thus allowing the cooperative action of the metal complexes to occur with the restricted spatial orientation of each complex.

In conclusion, we constructed an oligonucleotide with two metal complexes at an internal site and demonstrated that the agent cleaves RNA at one site with high efficiency, by the cooperative action of the complexes. A dinuclear Cu(II) complex containing terpyridine(s) is known to cleave RNA efficiently (28,29). Thus, our agent could be regarded as an oligonucleotide with a dinuclear complex at the internal site. An antisense agent with a dinuclear complex at the 5′-end has been described, but its RNA cleavage activity is low (30). Furthermore, our agent cleaves RNA with catalytic turnover, and its ability is higher than those of the agents reported thus far. Our RNA cleaver design and its modifications will be useful for the construction of a very active artificial RNase. In addition, the strategy of allowing two metal complexes to access the target phosphodiester may be applicable to the design of site-specific DNases. We are now studying the structure–activity relationship of our RNase, including the temperature dependence of the activity and the effects of other metal ions, such as zinc ions, on the activity.