7',5'-alpha-bicyclo-DNA: new chemistry for oligonucleotide exon splicing modulation therapy

Abstract

Antisense oligonucleotides are small pieces of modified DNA or RNA, which offer therapeutic potential for many diseases. We report on the synthesis of 7',5'-α-bc-DNA phosphoramidite building blocks, bearing the A, G, T and MeC nucleobases. Solid-phase synthesis was performed to construct five oligodeoxyribonucleotides containing modified thymidine residues, as well as five fully modified oligonucleotides. Incorporations of the modification inside natural duplexes resulted in strong destabilizing effects. However, fully modified strands formed very stable duplexes with parallel RNA complements. In its own series, 7',5'-α-bc-DNA formed duplexes with a surprising high thermal stability. CD spectroscopy and extensive molecular modeling indicated the adoption by the homo-duplex of a ladder-like structure, while hetero-duplexes with DNA or RNA still form helical structure. The biological properties of this new modification were investigated in animal models for Duchenne muscular dystrophy and spinal muscular atrophy, where exon splicing modulation can restore production of functional proteins. It was found that the 7',5'-α-bc-DNA scaffold confers a high biostability and a good exon splicing modulation activity in vitro and in vivo.

Graphical Abstract

General scheme illustrating the synthesis, base-pairing characterization, modeling and potential therapeutic application of 7?,5?-?-bc-oligonucleotides.

Figure 1.

Schematic comparison of 7′,5′-α-bc-DNA with DNA analogs. Chemical structures of DNA, α-DNA and bc-DNA (top), the diastereomeric 7′,5′-β-bc-DNA (center left) and 7′,5′-α-bc-DNA (center right) as well as the 3D structural relationship arising from a 180° rotation around the pseudo C2-axis running through the centers C1′ and C6′ (bottom).

Figure 2.

Synthesis of pyrimidine building blocks. (a) Thymine, BSA, TMSOTf, MeCN, rt, 18 h, 82%; (b) TBAF, THF, 2 h, 75%; (c) DMTr-Cl, pyridine, rt, 24 h, 96%; (d) K₂CO₃, MeOH, 3 h, 86%; (e) 2-cyanoethyl N,N,N′,N′-tetraisopropylphosphordiamidite, ETT, DCM, rt, 1 h, 81% for 6, 30 min, 80% for 8; (f) (i) BSA, 1,2,4-triazole, POCl₃, Et₃N, MeCN, rt, 7 h, (ii) 1,4-dioxane/NH₄OH, rt, 3 h, (iii) Bz₂O, Et₃N, DMF, rt, 18 h, 83%.

Figure 3.

Synthesis of purine building blocks. (a) N⁶-Benzoyladenine, BSA, TMSOTf, MeCN, 70°C, 20 min, 64%; (b) NaOH, THF/MeOH/H₂O, 0°C, 20 min, 51% α-anomer, 18% β-anomer; (c) Ac₂O, DMAP, DCM, rt, 18 h, 90%; (d) TBAF, THF, rt, 3.5 h, 90%; (e) DMTr-Cl, pyridine, rt, 24 h, 89%; (f) NaOH, THF/MeOH/H₂O, 0°C, 30 min, 94%; (g) 2-cyanoethyl N,N,N′,N′-tetraisopropylphosphordiamidite, ETT, DCM, rt, 1 h, 77% for 15, 50 min, 67% for 23; (h) 2-amino-6-chloropurine, BSA, TMSOTf, MeCN, 55°C, 50 min, 77%; (i) NaOH, THF/MeOH/H₂O, 0°C, 20 min, 85%; (j) TBD, 3-hydroxypropionitrile, DCM, 48 h, 87%; (k) Ac₂O, DMAP, DCM, rt, 48 h, 76%; (l) TBAF, THF, rt, 4 h, 87%; (m) DMTr-Cl, pyridine, rt, 48 h, 99%; (n) (i) K₂CO₃, MeOH, rt, 7 h, (ii) N,N-dimethylformamide dimethylacetal, DMF, 55°C, 2 h, 77%.

Figure 4.

X-ray structure. (A) 5′-O-p-nitrobenzoyl-7′,5′-α-bc-T. (B) 5′-O-acetyl-7′,5′-α-bc-G^Ac. Non-polar hydrogen atoms are omitted for clarity.

Figure 5.

Oligonucleotide design. (A) Insertion of 7′,5′-α-bc-DNA with polarity reversal inside β-DNA. (B) Representation of the palmitic acid conjugate. (C) Representation of the GalNAc conjugate.

Figure 6.

UV-melting curves (260 nm) of ON6 with fully modified parallel (ON7) and antiparallel (ON8) complement, parallel DNA and parallel RNA, in comparison with the corresponding natural DNA and RNA duplexes. Total strand conc. 2μM in 10 mM NaH₂PO₄, 150 mM NaCl, pH 7.0

Figure 7.

CD-spectra at 20°C of ON6 with parallel DNA, parallel RNA and fully modified parallel complement (ON7), in comparison with the corresponding natural DNA·RNA duplex. Experimental conditions: Total strand conc. 2 μM in 10 mM NaH₂PO₄, 150 mM NaCl, pH 7.0.

Figure 8.

Conformational analyses. The four minimum energy conformers of 7′,5′-α-bc-DNA-G nucleoside: (A) conformer a, (B) conformer b, (C), conformer c, (D) conformer d. Potential energy surface starting from (E) C6′-endo conformation and (F) C6′-exo conformation.

Figure 9.

Comparison between X-ray structure and molecular model. (A) X-ray structure of the 5′-O-p-nitrobenzoyl-7′,5′-α-bc-T. (B) Minimal conformation calculated for the 5′-O-p-nitrobenzoyl-7′,5′-α-bc-T and (C) its corresponding PES.

Figure 10.

Last frame of the 60 ns simulation for (A) 7′,5′-α-bc-DNA homo-duplex and (B) 7′,5′-α-bc-DNA/RNA duplex.

Figure 11.

Plots of the phase angle versus γ angle for each snapshot during the last 50 ns of the simulation for (A) ON6 in the context of 7′,5′-α-bc-DNA homo-duplex and (B) ON6 in the context 7′,5′-α-bc-DNA/RNA duplex.

Figure 12.

(A) Cropped agarose gel for mouse exon 23 and exon 22 + 23 skipping efficacy after transfection detected by nested RT-PCR. (B) Combined agarose gel for mouse exon 23 and exon 22 + 23 skipping efficacy after gymnosis detected by nested RT-PCR. (C) Combined agarose gels for mouse exon 23 and exon 22 + 23 skipping efficacy after intramuscular injections in gastrocnemius right (GR) and left (GL) and the triceps right (TR) and left (TL) to mice detected by nested RT-PCR. Quantification of antisense activity. (D) Quantification (n= 2) with ImageJ software for mouse exon 23 and exon 22 + 23 skipping efficacy after transfection detected by nested RT-PCR. (E) Quantification (n= 2) with ImageJ software for mouse exon 23 and exon 22 + 23 skipping efficacy after gymnosis detected by nested RT-PCR. (F) Quantification of skipped products with a lab-on-a-chip for intramuscular injections in mdx mice (n= 4 for ON10-12 and positive controls, n= 3 for ON9). *FA denotes a conjugation to a fatty acid (palmitic acid). The skipping efficiencies are reported as the average of the quantified values and the error bars represent the standard deviations.

Figure 13.

Quantification by ddPCR of the level of SMN transcript including exon 7 (SMN-fl) or excluding exon 7 (SMN-Δ7), after subcutaneous injection with ON13 or MOE1 at 10, 33 or 100 mg/kg, or with a saline solution (n= 3 for each cohort, all ddPCR experiment were performed in duplicate). The transcripts levels are reported as the average of the quantified values and the error bars represent the standard deviations. The transcript levels were normalized by taking the negative control cohort as reference.