Towards superior mRNA caps accessible by click chemistry: synthesis and translational properties of triazole-bearing oligonucleotide cap analogs

Abstract

Messenger RNA (mRNA)-based gene delivery is a powerful strategy for the development of vaccines and therapeutics. Consequently, approaches that enable efficient synthesis of mRNAs with high purity and biological activity are in demand. Chemically modified 7-methylguanosine (m⁷G) 5' caps can augment the translational properties of mRNA; however, efficient synthesis of structurally complex caps, especially on a large scale, is challenging. Previously, we proposed a new strategy to assemble dinucleotide mRNA caps by replacing the traditional pyrophosphate bond formation by copper-catalyzed azide-alkyne cycloaddition (CuAAC). Here, we used CuAAC to synthesize 12 novel triazole-containing tri- and tetranucleotide cap analogs with the aim of exploring the chemical space around the first transcribed nucleotide in mRNA and overcoming some of the limitations previously reported for the triazole-containing dinucleotide analogs. We evaluated the efficiency of incorporation into RNA for these analogs and their influence on the translational properties of in vitro transcribed (IVT) mRNAs in rabbit reticulocyte lysate and JAWS II cultured cells. The incorporation of the triazole moiety within the 5',5'-oligophosphate of trinucleotide cap produced compounds that were well incorporated into RNA by T7 polymerase while replacing the 5',3'-phosphodiester bond with triazole impaired incorporation and translation efficiency, despite a neutral effect on the interaction with the translation initiation factor eIF4E. One of the compounds (m⁷Gppp-tr-C₂H₄pA_mpG), had translational activity and other biochemical properties comparable to natural cap 1 structure, thus being a promising mRNA capping reagent for potential in cellulo and in vivo applications in the field of mRNA-based therapeutics.

Fig. 1. Structures of previously synthesized phosphotriazole dinucleotide cap analogs which have been used as a template in this work.

Fig. 2. Structures of phosphotriazole trinucleotide cap analogs synthesized in this work.

Fig. 3. Structures of 5′-azido modified dinucleotides (compounds 12a, 12b, and 12c).

Scheme 1. Solid phase synthesis of 5′-azido modified dinucleotides. (A) Solid phase synthesis of compounds 12a and 12b; (i) (1) 3% trichloroacetic acid in dichloromethane, (2) 2 equivalents of phosphoramidites in acetonitrile 0.2 M and BTT Activator 0.3 M, (3) 50 mM iodine solution in pyridine/water (9 : 1, v/v); (ii) (1) 3% trichloroacetic acid in dichloromethane, (2) 20% (v/v) diethylamine in acetonitrile; (iii) 0.6 M (PhO)₃PCH₃I in anhydrous DMF, 15 minutes, rt; (iv) saturated solution of NaN₃ in DMF, 1 hour, 60 °C; (v) 33% ammonium hydroxide and 40% methylamine in water (1/1, v/v), 1 hour, 50 °C; (B) synthesis of compound 12c; (i) (1) 3% trichloroacetic acid in dichloromethane, (2) 2 equivalents of phosphoramidites in acetonitrile 0.2 M and BTT Activator 0.3 M, (3) 50 mM iodine solution in pyridine/water (9 : 1, v/v); (ii) saturated solution of NaN₃ in DMF, 1 hour, 60 °C; (iii) 33% ammonium hydroxide and 40% methylamine in water (1/1, v/v), 1 hour, 50 °C; (iv) TEA, TEA*3HF, DMSO, 3 h, 65 °C.

Scheme 2. Synthesis of tri- and tetranucleotide cap analogs by CuAAC. (A) Synthesis of trinucleotide cap analogs with phosphate bridge modification. (B) Synthesis of tri- and tetranucleotide cap analogs with phosphodiester modification. (C) Representative RP-HPLC chromatogram obtained for the synthesis of compound 9.

Fig. 4. (A) Cap analog pairing with the template sequence which initiates transcription. The particular type of analog incorporated into short RNA, i.e., di-, tri-, or tetranucleotide, defines the length of the final transcript after DNAzyme trimming as 26, 27, or 28 nucleotides, respectively. Control uncapped RNA is 25 nucleotides long. (B) Analysis of capping efficiency. Short RNAs obtained by in vitro transcription from a template containing φ6.5 promoter, capped with different analogs during the reaction (with ATP, CTP, UTP, 2 mM each; 0.5 mM GTP; 3 mM cap analog), were analyzed in 15% PAA after trimming by DNAzyme. Capping efficiency values based on densitometric quantification of bands corresponding to capped and uncapped RNAs in each sample are given above the gel.

Fig. 5. Translation efficiencies for mRNAs capped with novel triazole-bearing trinucleotide cap analogs from IVT reaction with ATP, CTP, UTP, 2 mM each; 0.5 mM GTP; 3 mM cap analog, measured in vitro in rabbit reticulocyte lysate system. Gaussia luciferase luminescence signal was measured as a function of mRNA concentration and the linear regression coefficients were obtained from those dependencies. The coefficients obtained for each capped-mRNA were normalized to the coefficient obtained for ARCA-mRNA to determine relative translational efficiencies. Four different concentrations of mRNA in translation reactions were tested (see Experimental section for details). mRNAs capped with ARCA and other unmodified m⁷G caps were used as positive controls and GpppG was used as a negative control. Data represent mean value ± SD calculated for three replications.

Fig. 6. Protein output from HPLC-purified mRNAs capped with various analogs. (A) Gaussia luciferase activity was measured in medium collected from JAWS II cultures at particular-time points: 16, 40, 64, and 88 h after cell transfection with luciferase mRNAs capped with various analogs. Medium was replaced with fresh one; therefore, results from each time point show activity of newly produced luciferase. Measurements for technical triplicates from single biological replicate are presented. (B) Total protein production (cumulative luminescence) over four days in JAWS II cells transfected with mRNA capped with various analogs (results from tetraplicate experiment). Bars represent mean value ± SEM normalized to ARCA-capped RNA. Statistical significance: ****P < 0.0001 (one-way ANOVA with Turkey's multiple comparison test).

Fig. 7. The affinities for eIF4E determined using FQT experiments. (A) Example curves obtained for triazole cap analogs. (B) Cumulative K_AS values for 2′-O-methyl triazole cap analogs.

Fig. 8. Susceptibility to degradation of capped RNAs by hDCP1/DCP2 complex. (A) Representative PAGE analysis; RNA₂₅: uncapped RNA. RNA₂₇ – RNA capped with m⁷GpppA_mpG, compound 4, or compound 9. RNA₂₆ – RNA₂₇ decapped by hDCP1/2. (B) Comparison of degradation of RNA capped with m⁷GpppA_mpG, 4, and 9 in time. Data represents mean value ± SD calculated for two replicates.