An efficient strategy for the synthesis of circular RNA molecules.

Abstract

We have developed a fast and simple one test-tube procedure for synthesizing large amounts of pure, homogeneous circular RNA molecules of any sequence. This new strategy is based on the exploitation ofRNA internal secondary structure to position the 5' and 3' termini such that they will be single stranded, but held in close proximity to each other for subsequent ligation by T4 RNA ligase. This method consists of five simple steps (Fig. IA): (i) use of the circularly permuted RNA strategy (1) to introduce any desired mutations and to position the ligation site; (ii) run-off transcription [using a highly improved commercial kit, RiboMax System, Promega (2)]; (iii) replacement of the ensuing transcript's 5' triphosphate by a monophosphate (e.g. dephosphorylation-phosphorylation); (iv) ligation of the RNA molecules by T4 RNA ligase (3) (e.g. circularization) and (v) gel purification. The critical step in this procedure is the use of the circularly permuted RNA strategy (1) to produce the desired transcriptional template. This strategy involves a PCR amplification of a unit-length gene from a tandemly duplicated gene template, and is used to position the ligation site and to introduce any desired mutations into the subsequentRNA molecule. The amplification is performed using a forward primer with T7 RNA polymerase promoter sequences at its 5' end, followed by two or three guanosine residues for more efficient transcriptional initiation, and ending with the 5' terminal sequence desired for the subsequent RNA molecule. The 3' terminal of this RNA is similarly determined by the reverse primer utilized. By tandemly moving both primers along the template, effectively any position can be engineered to become the ligation site. The use of a highly processive 5'-3' DNA polymerase that possesses a 3'-5' exonuclease activity to effect the amplification ensures the elimination of polymerase errors and, therefore, sequence homogeneity Conventional recombinant RNA techniques may be used for synthesis of circular molecules; however, they require polyacrylamide gel purification of the products at almost every step resulting in a laborious and inefficient synthesis. Here, we described a procedure that may be performed in <3 days, and requires only a single, terminal gel purification. Moreover, this method complements the T4 DNA ligase strategy for joining RNAs (4), a strategy which we have shown to be unsuccessful for the preparation of large amounts of model viroid using either a complementary oligodeoxyribonucleotide, when the ligation site was located in a loop due to inefficient RNA-DNA heteroduplex formation attributable to the stable secondary structure, or when the ligation site was located in a native stable helix because the 5' end dephosphorylation-phosphorylation reactions of the previous step were sterically inefficient. Using the strategy described here, the introduction of a mutation in a double-stranded region of the molecule is not limiting because the ligation site is selected to be in a single-stranded region. The limiting factor to the application of this method is some knowledge of the secondary structure in the region of the ligation site. However, this restriction may be easily circumvented by a fast study of the secondary structure near the ligation site, for example by RNase TI partial hydrolysis or by computer prediction. Use of this simple strategy permits either the circularization ofan RNA, or the ligation oftwo RNAs of any sequence by using primers leading to production of RNA whose extremities fulfil these requirements. A synthesis of mutated circular transcripts of sequences derived from peach latent mosaic viroid (PLMVd) is described here as an example. Since viroids are composed of a series of singleand doublestranded regions, they offer several potential T4 RNA ligase ligation sites. The ligation site chosen in the current example is located in the loop at the end of the left arm region so as to ensure efficient ligation by T4 RNA ligase (Fig. IA, enlarged circle). This loop has been selected since several positions are mutated as compared with the sequence variants published (5) and, consequently, the replacement of three nucleotides (UCA) by three guanosine residues should allow for an efficient initiation of transcription without interfering in the ensuing studies. All enzymatic reactions within the method have been optimized to a preparative scale. The enzymatic steps (PCR-amplification, run-off transcription, dephosphorylation, phosphorylation and ligation) are linked sequentially by simple manipulations. Both phenol (1 vol) and phenol-chloroform (0.5 vol:0.5 vol) extractions are performed after each step prior to proceeding. After both the PCR-amplification and transcription reactions, the large nucleic acid products are separated from any remaining nucleotides and primers by isopropanol precipitation; whereas the products resulting from the dephosphorylation, phosphorylation and ligation reactions are ethanol precipitated. All precipitations are followed by 70% ethanol washes and pellet lyophilyzation. Hence, it is therefore possible to interrupt the synthesis at any step.