Constant efforts have been made to develop new method to realize sequence-specific RNA degradation, which could cause inhibition of the expression of targeted gene. Herein, by using an unmodified short DNA oligonucleotide for sequence recognition and endogenic small molecue, vitamin B2 (riboflavin) as photosensitizer, we report a simple strategy to realize the sequence-specific photocleavage of targeted RNA. The DNA strand is complimentary to the target sequence to form DNA/RNA duplex containing a G•U wobble in the middle. The cleavage reaction goes through oxidative elimination mechanism at the nucleoside downstream of U of the G•U wobble in duplex to obtain unnatural RNA terminal, and the whole process is under tight control by using light as switch, which means the cleavage could be carried out according to specific spatial and temporal requirements. The biocompatibility of this method makes the DNA strand in combination with riboflavin a promising molecular tool for RNA manipulation.
Site-selective RNA cleavage to inhibit expression of special genes is consistently pursued by scientist for the promising application in biotechnology and therapy as well as biology. Previous research focused on gene suppression strategies through RNA degradation involving antisense oligonucleotides and artificial ribonucleases1,2,3,4,5. More efforts are still need to put into these methodologies to improve the efficiency before they could be broadly adapted as genomic or therapeutic tools. Since the first description of RNA interference (RNAi), there has been rapid progress towards its application as a therapeutic modality6. The mechanism of the RNAi is that the RNAi (the length of RNA must be longer than 21 bases), generated from double-stranded RNA, leads to specific degradation of complementary mRNA sequences7, triggers silencing of homologous gene expression8. Therefore, RNAi requires cell’s own machinery to fulfill its job. Ribozymes found in nature9,10 and deoxyribozyme obtained from in vitro selection11,12 could high-efficiently cleave the targeted RNA sequence-specifically through formation the right tertiary structure, which usually comprises of a catalytic domain which are flanked by two substrate-recognition domains.
Dilara Crate and Charles Wilson reported a site-specific inactivation of RNA transcripts through Laser-induced hydrolysis of the malachite green (MG)-binding RNA, which extend the RNA cleavage into a new scenario, namely under tight spatial and temporal control13. The application of this method is facing several challenges: (1) How to Introduction of the MG-binding motif into the 5′ untranslated region (UTR) of target RNA. (2) Toxicological effects of malachite green when applied in therapy. Herein, by using an unmodified short DNA oligonucleotide as binder and endogenic small molecue, vitamin B2 (riboflavin) as photosensitizer, we report a sample strategy to realize the sequence-specific photocleavage of targeted RNA.
Riboflavin, as the most efficient natural photoshensitizer, has been shown to induce photooxidative DNA damage via the generation of an 8-hydroxydeoxyguanosine intermediate14,15,16,17. The flavin-oligonucleotide conjugates have been successfully applied in the photocleavage of double stranded DNA target. In 1997, Famulok group reported that riboflavin could cleave RNA molecules with overwhelming specificity at G•U wobble via a photo-induced mechanism when they study an RNA aptamer for isoalloxazine derivatives which was obtained from in vitro selection18,19,20. As shown in Fig. 1A, they found the riboflavin could specifically recognize the G•U wobble in RNA duplex and induce the oxidative cleavage reaction through photosensitization. The Cleavage occurred independently of the residue located at the 3′ of the U (red, Fig. 1A), and the nucleoside downstream of the uracil (U) of the G•U base pairing was verified to be removed through oxidative mechanism. Their work inspires us to ask questions: How about the RNA/DNA duplex containing G•U wobble pair? Can any RNA sequence be photocleaved by DNA strand in presence of riboflavin?
As shown in Fig. 1B, a DNA strand was designed to complimentary to RNA target sequence except for forming a G•U wobble pair at the 17 nt position. The photocleavage of RNA target was verified by the isotope-labeled experiments (Fig. 1C). The 5′ 32P-labeled RNA target with 26 nt in length was incubated with the DNA sequence in presence 200 μM of riboflavin at 25 °C. After 4 h of irradiation with 45 W household light, a clear cleaving fragment with 17 nt in length was observed based on the PAGE gel analysis. Meanwhile, no cleavage could be observed when the DNA strand, riboflavin or light irradiation was absent in the cleaving reaction (Lane 1–3, Fig. 1C). To verify the importance of G•U wobble pair for the cleavage, the full complementary DNA sequence of RNA target was applied into the same reaction (Fig. 1B). As illustrated in the lane 4 of PAGE analysis in Fig. 1C, 5′ 32P-labeled RNA target kept intact after 4 h irradiation under the same reaction condition in presence of full complementary DNA strand, which indicates riboflavin could recognize the G•U wobble in RNA/DNA duplex to induce the site-selective photocleavage of RNA target.
The chemical requirements for the RNA photocleavage were investigated. As shown in Fig. 2A, the DNA strand worked efficiently even without magnesium ion in presence of 200 μM riboflavin (lane 1), and the cleavage efficiency did not change too much with different concentrations of Mg2+ (Lane 2–4). However, the catalytic activity of the DNA strand strongly depended on the concentration of riboflavin (Fig. 2B). The photocleavge of RNA target by the DNA strand increased along with the increase of riboflavin concentration, reaching the maximal activity in around 50 μM (Fig. 2B). Moreover, the temperature has little influence to the cleavage efficiency based on the DNA strand in combination with riboflavin from 20 to 40 oC (data no shown). Under optimal condition, the DNA strand in combination with riboflavin catalyzed the photocleavage of target RNA with a kobs as high as 0.25 h−1, the half life being around 6 hours, which is much faster than most of reported artificial ribonucleases2. The RNA substrate containing XUY for this strategy has been carefully studied (Supplementary Fig. S1). The nucleoside X located at the 5′ of the uracil of the G•U base pair could affect the cleavage efficiency of RNA substrates (G ≈ A > C > U). The nucleoside Y RNA in substrate that will be removed from 3′ of U revealed the tread of C > U > A > G in terms of cleavage efficiency (Supplementary Fig. S1). The RNA substrate for the sequence-specific photocleavage has the frequency of one repeat in every 4 bp, which reveals the broad generality of this molecular tool with respect to different RNA sequence. The length of two binding arms of the DNA strand is variable and the catalytic efficiency would not be affected as long as the total length of two binding arms are longer than 12 nt (Supplementary Fig. S2).
As comparison, the RNA hairpin (35 nt in length) that encodes the same sequence of RNA target at 5′end was applied to self-cleavage under the same conditions. Both the intramolecular and intermolecular reaction provided the cleaved products with the same length (Lane 2 and Lane 4, Fig. 3C), revealing riboflavin could bind to the G•U wobble base pair in RNA/RNA as well as RNA/DNA duplexes and photocleave the RNA target sequence with similar efficiency. However, the cleaved product has slightly faster mobility than the 5-32P-labeled RNA marker (17 nt), revealing the presence of one more phosphate group at the 3′-end (Lane 1 and Lane 2, Fig. 3C). Therefore, T4 polynucleotide kinase (PNK) that could remove phosphate group from 3′-end of RNA has been added to the reaction mixture. The treatment with PNK didn’t change the mobility of cleaved fragment on gel (Lane 2 and Lane 4, Fig. 3C), indicating that the cleavage reaction could go through oxidative elimination mechanism to form unnatural 3′-phosphate RNA terminal. Moreover, the photocleavage activity reduce 74% in the absence of oxygen (Fig. 3D), revealing the excited photosensitizer interacts with ground state oxygen to generate a singlet oxygen molecule, which is the same as what Famlok group has observed in the photocleavage of RNA duplex19.
To validate the scope of this method, mircoRNA21 with 22 nt in length was applied as the target to verify whether our photocleavage method based on the DNA strand in combination with riboflavin can work or not. An uracil (U) in the middle of mircoRNA21 was selected as the cleaving site, and the DNA strand was designed to complement to the target sequence except for the formation of a G•U wobble (Fig. 4A). As shown in the isotope-labeled experiment, a cleaved fragment with expected length (11 nt) was detected based on the PAGE analysis of the photo-induced cleavage reaction (Lane 3 Fig. 4B). For the longer RNA target, we can design multiplex cleaving-sites to improve the cleavage efficiency and specificity based on our method. As a proof-of-concept, three uracil (U) of RNA (Long-RNA, Fig. 4C) has been targeted by DNA strand, and any of them once cleaved by DNA strands could result in the disfunction of RNA target. As comparison, RNA (Short-RNA, Fig. 4C) with only one cleaving-site was cleaved by one DNA strand under same reaction condition. After 2 h of photocleavage in presence of riboflavin, more than 53% of the longer RNA containing three cleaving-sites was cleaved, while only 28% of cleavage obtained for the reaction of the shorter RNA (Lane 3 and Lane 6, Fig. 4D).
To further confirm the DNA stand in combination with riboflavin as tool to hinder the RNA function through cleavage, RNA mimic of GFP (Green Fluorescent Protein) was chosen as the object of study. Recently, Jaffrey group has described a complex of RNA aptamer (24–2) and a derivative of the GFP fluorophore, 3,5-difluoro-4-hydroxybenzylidene imidazolinone (DFHBI), termed as “Spinach” that emits fluorescence upon transformation in specific conformation21 (Fig. 5A). This RNA-fluorophore complex displays fluorescence properties similar to GFP and is unlike most fluorescent proteins, resistant to photo-bleaching and conditionally fluorescent. We designed DNA strand (24–2-Cat) to cleave RNA 24–2 through forming G•U wobble base pair at 31 nt position of the uracil (U) (Fig. 5B), and the cleavage activity of 24–2-Cat to RNA target in presence of riboflavin was verify by the following isotop-labeled experiment and PAGE analysis (Lane 3 Fig. 5C). The fluorescence intensity of the complexes (24–2 + DFHBI) was significant (blue curve, Fig. 5D). However, by adding 24–2-Cat into the reaction, the fluorescence intensity dropped to 44.61% after 5 h photocleavage in comparison with the reaction without the DNA strand (green curve, Fig. 5D). As a control, full complementary DNA sequence (24–2-NC, Fig. 5B) was introduced into the same reaction as a comparison, showing only 7.59% fluorescence intensity decrease (red curve, Fig. 5D). Those results indicates the photocleavage can directly hinder the function of RNA target through cleavage other than block the RNA target by forming DNA/RNA duplex.
The established sequence-specific photocleavge of RNA described here appears to have several promising features to become a new molecular tool for research and application as follows: (1) the designation of DNA strand to realize the seqence specific cleavage of RNA target is very easy to fulfill only if there are uracil nucleotides in the RNA sequence and the DNA strand is complementary to the target sequence expect to form a G•U wobble in between. (2) Though the minimal size of DNA strand is 13 nt in length, high sequence specificity could be achieved by increase the length of DNA stand to form the longer DNA/RNA duplex with target. The efficiency of the cleavage reaction based on the DNA strand in combination with riboflavin is very high. And the cleaving specificity and efficiency could be further improved by introducing more strands to cleave the same RNA at multiplex cites. (3) The cleavage of RNA by DNA strand is under tight control by using light as switch, which means the cleavage could be carried out according to specific spatial and temporal requirements. Moreover, the photosensitizer, riboflavin (vitamin B2), is endogenic coenzyme existing in the cell of most living creature, which makes this method possible to apply in vivo. (4) Unlike antisense oligonucleotide, the DNA strand in our method could interfere with the function of RNA through the irreversible photocleavge, forming cleaved fragments with unnatural 3′ -phosphate RNA terminal.
In summary, we developed a new RNA-cleaving DNA strand that catalyzes the sequence-specific photo-induced cleavage of RNA at the 3′ of the uracil of G•U wobble base pair in the presence of rioflavin as cofactors. Unlike most artificial ribonuclease that need conjugate functional chemical group to DNA sequence to fulfill the cleavage of RNA molecules specifically, the DNA strand described here is unmodified DNA sequence, cooperating with riboflavin to realize the photocleavage of RNA target through oxidative elimination mechanism. This method has broad substrate scope if the targeted RNA contains required uracil nucleotide that has high frequency of one repeat in every 4 bp. The function of RNA molecules were successfully hinder by photocleavage based on our method in vitro, which was proved to proceed through the cleavage reaction other than the hybridization with target. The biocompatibility of this method makes the DNA strand in combination with riboflavin a promising manipulable molecular tool in vivo. The direct conjunction the DNA strand with riboflavin and the application our method in the cell experiment is underway in our lab.
How to cite this article: Zhao, Y. et al. Sequence-specific RNA Photocleavage by Single-stranded DNA in Presence of Riboflavin. Sci. Rep. 5, 15039; doi: 10.1038/srep15039 (2015).
This study was supported by National Sciences Foundation of China (Grant No. 21172215, 21102140, 21402189 and 21322208).
About this article
Organic & Biomolecular Chemistry (2017)