Characterization of the extreme 5' ends of RNA molecules by RNA ligation-PCR.

1Permanent address: P..O. Box 677, Nedlands, Western Australia, 6009. Despite major recent advances in PCRbased methodology, characterization of the extreme terminal sequences of RNA molecules may be difficult and time consuming, particularly when the 3' end is not polyadenylated.~l~ The usual approach to defining the 5' end involves cDNA synthesis using gene-specific downstream priming of the reverse transcription (RT) reaction, followed by the addition of a homopolymer tail by terminal transferase and single-sided PCR using a homopolymeric dN-based oligonucleotide to prime the 5' PCR reaction (the RACE protocols). Verification of the extreme 5' end of the RNA molecule requires two separate reactions in which different homopolymer tails are added by terminal deoxytransferase. (2~ However, premature termination of the cDNA reaction cannot be excluded using this technique and would result in undetected and foreshortened cDNA clones, even if dual tailing reactions are used. Additionally, the presence or absence of a 5'-terminal structure, such as a methylated cap or a covalently bound VpG protein, cannot be deduced by homopolymer tailing methodology. Both 5' hairpin loops (3~ and terminal cap structures ~4~ are known to exist at the 5' end of many RNA viruses. Recently, head-to-tail ligation of viral molecules with PCR amplification across the 3' to 5' junction has been advocated as a method of defining the terminal sequences and/or structure of RNA molecules. (s,6~ However, this method requires knowledge of the complete sequences of one end or the other before the junctional sequence can be defined fully. Furthermore, variable-length homopolymer tails on the 3' end of many RNA molecules prevent application of direct PCRbased sequence strategies to the resulting PCR products. ~6~ The 3' of nonpolyadenylated RNA molecules or molecules of unknown sequence is more problematic. Using cobait-based buffers, RNA is a suitable substrate for the addition of a homopolymeric tail by terminal transferase. It is conceivable that a strategy based on tailing followed by 3' single-sided PCR using a 5' gene-specific primer and an oligo(dN)-based 3' primer would work. However, definit ion of the extreme 3' nucleotide would require two separate tailing experiments.