15 min Read

Custom circRNAs can be generated in living cells by over expression from specialized plasmid constructs which can utilize different circularization strategies as discussed in the below sections. These constructs can be cloned into non-viral vectors for transient expression or can be cloned and packaged into AAV for non-integrating stable expression. Notably circRNA constructs are largely incompatible with lentiviral vectors and virus. This is mainly due to the circRNA construct containing complementary sequences which can be recognized during the packaging process resulting in viral RNA splicing and removal of the circRNA construct from the viral particles.

1. Introns with Complementary Sequences

RNA-seq studies have revealed that most endogenous circRNAs are flanked by introns which contain reverse complementary sequences. These flanking regions are thought to base pair together and in turn bring splice donor and acceptor sites in close proximity, which ultimately promotes back splicing into circRNA (12).

1.1 Introns with repetitive complementary sequences

These introns are derived from endogenous sequences which contain the repetitive DNA element Alu. Alu elements are the most common repetitive DNA element in humans and primates and are classified as short interspersed repeat elements (SINEs). Notably Alu elements are unique to humans and primates, thus their application to mimic endogenous non-human circRNA generation is uncertain. Currently there are several well documented repetitive introns that have been utilized for the generation of circRNAs including ZKSCAN1, HIPK3 and Laccase2. One disadvantage of the complementary sequences system (repetitive or non-repetitive) includes the possible generation of undesirable byproducts arising from competing splice reactions.

1.2 Introns with non-repetitive complementary sequences

Similarly circRNAs can also be generated using non-repetitive introns lacking Alu elements. These introns can either be derived from endogenous sequences or can be partially designed in silico.

abm utilizes a non-repetitive complementary sequence strategy for its circRNA expression vectors.

2. Introns with RNA Binding Protein Motifs

Several RNA binding proteins (RBP) have been determined to play a role in enhancing back splicing in vivo. A secondary strategy for circularization involves incorporating the specific binding sites for these proteins into the over expression plasmid.

2.1 Muscleblind

Muscleblind (Mbl) is a RBP present in Drosophila and humans. Its binding sites are highly conserved and are often found in flanking intronic regions. The Mbl protein binds to its conserved sites and dimerizes, effectively bringing the flanking introns in close proximity. These binding sites can be incorporated into the over expression plasmid to promote back splicing into circRNA (13). It is also noteworthy that this strategy may not be applicable for expressing circRNAs in cells other than Drosophila or humans as an Mbl ortholog may be absent or Mbl expression may be insufficient thus hampering circularization.

2.2 Quaking

Quaking (QKI) is another RBP that is known to facilitate back splicing (14). The mode of mechanism is similar to that of Muscleblind as seen above.

3. PIE System

The permuted introns-exon (PIE) system is an adaptation of the bacterial group I self-splicing intron. These elements can essentially act as ribozymes and thus have been adapted to yield circRNAs (15). The PIE mechanism involves the 5’ half of the group I tRNA intron transferring to the tail of the tRNA exon, and lastly the remaining 3’ half of the intron becomes positioned at the head of the same exon. Following transcription of this construct results in spontaneous circularization of the intervening exonic sequence. The main drawback of PIE includes the fact that the final circRNA will retain the terminal sequences of the originating exons at the ligation junction point. Additionally the system has been widely used in E. coli and fungi, but its application in mammalian cells is still unclear.

4. tRNA Introns

This novel system was adapted from the metazoan tRNA splicing mechanism (16). A polIII promoter drives the expression of a custom RNA sequence which is flanked by intronic cleavage sites. Following transcription the pre-tRNA is cleaved by conserved host enzymes and the excised introns are then ligated by a host RctB ligase resulting in a “tricRNA.” The system is unique in that it does not utilize the spliceosome machinery. However, disadvantages of this system include the retention of a ligation sequence in the final circRNA and the fact that transcription is driven by a polIII promoter. PolIII promoters efficiently drive expression of small RNAs, but are not efficient when transcribing larger RNAs or RNAs which contain internal U rich sequences. This may occlude its utility at producing larger multi-exon circRNAs or circRNAs that naturally contain U rich sequences which could be recognized as an early termination sequence.

1. Chemical Synthesis of Linear RNA Precursor

This strategy requires a supplier or a synthesizer machine to chemically synthesize the desired linear RNA precursor, which is subsequently ligated together to form circular RNA. The main advantage of chemical synthesis is high yield and purity of the RNA precursor. However, it also has its limitations as large RNAs cannot be synthesized efficiently and often costs are quite prohibitive.

2. Enzymatic synthesis of linear RNA precursor

Enzymatic synthesis utilizes in vitro transcription of the RNA precursor via phage RNAP such as T7 (17). T7 RNAP can transcribe RNA from a DNA template and is able to produce much longer RNAs when compared to chemical synthesis. However this strategy is prone to producing non-uniform transcripts which often contain an extra 1-3nt at the 3’ end and additionally it is known that the enzyme can sometimes completely fail to fully complete transcription.

3. Chemical Ligation

The synthesized linear RNA precursor can undergo circularization by condensation reagents such as cyanogen bromide (BrCN) or 1-ethyl-3-(3’-dimethylaminopropyl) carbodiimide (EDC) which facilitate an intermolecular reaction between the first and last nucleotides of the RNA (18). Alternative strategies involve substituting the linear RNA 5’ phosphate and 3’ hydroxyl groups with other functional groups which can undergo bond formation with high specificity and efficiency. However the latter strategies would lack a circularizing phosphodiester bond resulting in a circRNA which is dissimilar to endogenous circRNAs. Disadvantages of chemical ligation include a propensity to form a competing concatenated linear byproduct instead of the desired circular product.

4. Enzymatic ligation

DNA and RNA ligases derived from T4 bacteriophage can also facilitate RNA ligation into a circular species. These enzymes include DNA ligase I, RNA ligase I and II, all of which specifically require the RNA to contain a 5’ monophosphate instead of triphosphate in order to catalyze circularization. Regardless of which ligase is utilized, intermolecular ligations have a propensity to form RNA concatemers which is a major competing byproduct to the desired circular RNA.

4.1 DNA ligase I catalyzes the repair of nicks within double-stranded DNA substrates. Applied to circularization, the enzyme requires the addition of a short DNA oligo splint which can hybridize to the ssRNA and subsequently be recognized by the enzyme (18). This enzyme exhibits high ligation efficiency when there is perfect complementarity to the DNA splint and thus is the preferable method of ligation following in vitro transcription which often results in 3’ heterogeneity.

4.2 RNA ligase I utilizes only single-stranded RNA as substrate material. The enzyme joins RNA termini when they are in close proximity - this can be achieved by using an RNA splint to bind the termini ends as well as leave a couple unpaired nucleotides available at the ligation junction (19). In contrast to DNA ligase I, the enzyme has been implicated with lower reaction specificity.

4.3 RNA ligase II can be applied to circularization as its substrates include both double-stranded and single stranded RNA (20). Similarly, intermolecular ligation can be favoured by using either a RNA or DNA splint. Notably the enzyme joins nicks within double-stranded RNA with the highest efficiency.