Chemical synthesis of U1 snRNA derivatives.

U1 snRNA is an interesting biological tool for splicing correction and regulation of gene expression. However, U1 snRNA has never been chemically synthesized. In this study, the first chemical synthesis of U1snRNA and its analogues was carried out. Moreover, it was found that the binding affinity of the modified U1 snRNA with an ethylene glycol linkage to snurportin 1 (nuclear import adaptor) was as high as that of the unmodified RNA.

Uridine-rich small nuclear RNAs (U snRNAs) are involved in very important pre-mRNA splicing events in the nucleus of eukaryotic cells.[1] Among them, U1snRNA can recognize the 5′-splice site of pre-mRNA. At the first splicing step, U1snRNA directly binds to the splice site by hybridization between the 5′-terminal eight nucleotides of U1snRNA and the complementary sequence of pre-mRNA.[2]

Splicing disorders are estimated to account for about 15% of disease-causing mutations in humans, and the majority of genetic mutations are point mutations within the sequences of splice sites.[3] A subset of these defects should result from mismatched base pairs between U1snRNA and mutant pre-mRNAs. It was reported in a recent study that mutant pre-mRNA splicing could be promoted by modified U1snRNAs that were prepared using the biological technique in the cell and could hybridize with the mutant pre-mRNA.[4] These results indicate that modified U1snRNA could be useful for splicing correction in gene therapy.

Additionally, a modified U1snRNA containing the sequence complementary to the 3′-terminal exon of the targeted gene can downregulate the targeted gene expression via U1 interference (U1i).[5] This downregulation results from inhibition of mRNA polyadenylation by the U1 snRNA-binding protein (U1–70K protein). However, U1 snRNA has never been chemically synthesized because U1 snRNA is very long (164 nt) and contains a complicated 5′-terminal 2,2,7-N-trimethylguanosine (m32,2,7G) cap (Figures 1 and 2a).[6] The m32,2,7G cap can bind to the nuclear import adaptor snurportin 1 (SPN1) and plays an important role in transporting UsnRNAs between the cytoplasm and the nucleus. In this study, we carried out the chemical synthesis of U1snRNA and modified U1snRNAs containing the m32,2,7G cap analogues for the first time, as shown in Figure 2.

Figure 1

Structure of U1snRNA.

Figure 2

Chemical structure of (a) m32,2,7G cap, (b) pyrophosphate analogue, and (c) ethylene glycol analogue.

Figure 3 depicts the strategy for synthesizing U1snRNA with a fluorescent label at the 3′-terminus. First, we synthesized 5′-short RNA containing the m32,2,7G cap by chemical reactions on polymer supports (Schemes 1 and 2) and 3′-long RNA by biochemical methods using T7 RNA polymerase and a chemical reaction in solution phase (Scheme 3). We subsequently generated the complex formed between 5′-short RNA, 3′-long RNA, and the complementary splint DNA. After hybridization, the ligation reactions were carried out using T4 DNA ligase.[7]

Figure 3

Strategy for U1snRNA synthesis.

Scheme 1

Synthesis of 5′-Short RNA I Containing an m32,2,7G Cap

The sequence of oligonucleotide is AmUmACΨΨAC.

Scheme 2

Synthesis of 5′-Short RNA III Containing an m32,2,7G Cap Analogue

The oligonucleotide sequence is AmUmACΨΨAC.

Scheme 3

Synthesis of 3′-Long RNA

In the synthesis of 5′-short RNAs, we used the highly cross-linked polystyrene resin[8]1 that has a silyl-type linker,[9] which can be cleaved under neutral conditions, phosphoryl[10] and phosphityl reagents 2–4 (Figure 4). The synthesis of 5′-short RNA I is shown in Scheme 1. After chain elongation of the RNA sequence using 2′-O-TBDMS RNA phosphoramidite units on the resin by using an RNA synthesizer, the treatment of resin 5 with 100 equiv of diphenyl phosphonate in pyridine was followed by hydrolysis mediated by a H-phosphonate compound. The compound was subsequently oxidized using (2R,8S)-(+)-(camphorsulfonyl)oxaziridine in the presence of N,O-bis(trimethylsilyl)acetamide (BSA). The 5′-terminal pyrophosphate bond was formed by the reaction of the protected oligonucleotide immobilized on the resin with compound 2 for 15 min, as shown in Scheme 1. After deprotection of the 2-cyanoethyl, sulfonylethyl, and N-acyl groups, the 5′-terminal capping reaction was performed by treating residue 6 with m32,2,7G phosphorimidazolide unit 7.[11] The 2′-TBDMS groups of the oligomer were subsequently deprotected, and the oligomer was released from the resin by treatment with 1 M tetrabutylammonium fluoride (TBAF) in the presence of 0.5 M AcOH for 24 h. Although the only 2′-TBDMS groups of pseudouridine residues could not be cleaved under the deprotection conditions using TBAF, the TBDMS groups were easily removed by treatment with a 20% acetic acid aqueous solution for 3 h.[12] The target 5′-short RNA I was isolated by anion-exchange HPLC in 2% yield and characterized by MALDI-TOF mass spectrometry. Additionally, the 5′-short RNA II having a pyrophosphate group was obtained as a byproduct in 3% isolated yield.

Figure 4

Structure of resin 1, phosphoryl and phosphityl reagents 2–4.

In the synthesis of 5′-short RNA III containing an m32,2,7G cap analogue, the resin 8 was prepared by chain elongation using phosphityl reagents 3 and 4 on resin 1. Resin 8 was treated with DBU and MeNH2 to obtain resin 9. After capping using phosphorimidazolide unit 7, the oligomer was deprotected and released from the resin in 12% isolated yield. The yield of 5′-short RNA III was much higher than that of RNA with an unmodified m32,2,7G cap.

The 3′-long RNA was synthesized by RNA transcription using T7 RNA polymerase, as shown in Scheme 3. In this RNA transcription, we combined the desired RNA sequence and the hammerhead ribozyme sequence because efficient synthesis requires the RNA to have a purine-rich sequence at the 5′-terminus.[13] The 5′-sequence of hammerhead ribozyme was efficiently eliminated under the transcription conditions in the presence of Mg2+. After PAGE purification, we introduced a fluorescent label into the 3′-terminus of the RNA by treatment with NaIO4 and fluorescein-5-thiosemicarbazide in 65% isolated yield. For the final step, 5′-phosphorylation was carried out using T4 polynucleotide kinase.

The 5′-short RNAs I–III were ligated with the 3′-long RNA by T4 DNA ligase to obtain the U1 snRNA I and its analogues (U1snRNA 2 with a pyrophosphate linkage and U1 snRNA III with an ethylene glycol linkage) with a fluorescent label in the presence of the complementary splint DNA. The target U1 snRNAs I–III were isolated by PAGE in 14, 28, and 30% yields, respectively, and characterized by MALDI-TOF mass spectrometry. These results are the first demonstration of the chemical synthesis of U1 snRNA and their analogues.

Moreover, we examined the binding affinities of U1 snRNA I–III to the nuclear import adaptor SPN 1, as shown in Figure 5. The RNA without an m32,2,7G cap could not specifically bind to SPN 1, though nonspecific binding was observed in 2 μM SPN1 due to electrostatic interaction between the phosphate groups of the RNA and lysine (or arginine) residues of SPN1 (Figure 5a). On the other hand, U1 snRNA I with an m32,2,7G cap could completely bind to SPN1 in 1 μM SPN1 in Figure 5b. Figure 5c and 5d show the results of U1snRNAs with m32,2,7G cap analogues. Interestingly, the binding affinity of the analogue with an ethylene glycol linkage was as high as that of U1 snRNA I with an m32,2,7G cap, as shown in Figure 5d, though the binding affinity of the U1 snRNA II with a pyrophosphate linkage was lower than that of U1 snRNA I. Ficner et al. previously reported that the m32,2,7G interacts with another nucleobase and Trp276 of SNP1 in the crystal structure of the complex formed between SPN 1 and the m32,2,7GpppG dimer.[6e] Therefore, the shortage of the atom length of the pyrophosphate linkage in U1 snRNA II might decrease the binding affinity to SPN 1. On the other hand, a longer linkage such as that afforded by ethylene glycol in U1 snRNA III might not affect the binding affinity to SPN 1.

Figure 5

Electrophoretic mobility shift assay of the complexes formed between snurportin 1 (0–2 μM) and U1 snRNA I-III (50 nM) on a 16% nondenaturing polyacrylamide gel at pH 7.4. (a) U1snRNA without an m32,2,7G cap, (b) unmodified U1 snRNA I, (c) modified U1 snRNA II, and (d) modified U1 snRNA III with an ethylene glycol linkage.

In summary, we carried out a first chemical synthesis of U1snRNA and its analogues by a combination of solid-phase and enzymatic reactions. The synthesis of U1 snRNA III with an m32,2,7G cap analogue was easier and more efficient than that of U1 snRNA I with an m32,2,7G cap. Moreover, we found that the binding affinity of U1 snRNA III was as high as that of U1snRNA I. These results should be very useful for investigating splicing correction and U1i by an artificial U1 snRNA in gene therapy. Further studies on these issues are currently in progress.