Programmed ribosomal frameshifting (PRF) is a translational recoding mechanism used by many viral and cellular mRNAs. PRF occurs at a heptanucleotide slippery sequence and is stimulated by a downstream RNA structure, most often in the form of a pseudoknot. The utilization of −1/ +1 PRF to produce proteins encoded by the −1/+1 reading frame is wide-spread in RNA viruses, but relatively rare in cellular mRNAs. In human, only three such cases of −1 PRF events have been reported, all involving retroviral-like genes and protein products. To evaluate the extent of PRF utilization in the human transcriptome, we have developed a computational scheme for identifying putative pseudoknot-dependent −1 PRF events and applied the method to a collection of 45,000 human mRNAs in the NCBI RefSeq database. Using core program PKscan, we have performed a large-scale search for putative pseudoknot dependent PRF. In addition to the three reported cases, our study identified more than two dozen putative −1 PRF cases. The genes involved in these cases are genuine cellular genes without a viral origin. Moreover, in more than half of these cases, the frameshift site locates far upstream from the stop codon of the 0 reading frame, which is nonviral-like.Using dual luciferase assays in HEK293FT cells, we confirmed that the −1 PRF signals in the mRNAs of CDK5R2 and SEMA6C are functional in inducing efficient frameshifting. We also present evidence to show that the mRNA of human inorganic pyrophosphatase 1 (PPA1) harbors functional cis-acting signals for +1 PRF. The consequence of frameshifting is the production of a longer PPA1 protein in which the C-terminal 25 residues of normal PPA1 are replaced by 68 residues translated from the +1 reading frame. To the best of our knowledge, the human PPA1 mRNA is the only other mammalian cellular mRNA known to date to utilize the +1 PRF mechanism, besides the antizyme mRNAs. Results from the studies on the involvement of PPA1 in tumorigenesis suggest that PPA1 is a potential prognostic biomarker for certain cancers, and strategies for PPA1 down-regulation may have therapeutic potential for the treatment of cancers. This study also recognized a new pseudoknot structure in SARS-CoV that has been validated as a cis-acting regulation factor in viral frameshifting event, referred to 'intertwined double pseudoknots’. They are present at the −1 PRF site in SARS-CoV-1/2 and many other coronaviruses. An even larger scale analysis on transcriptome-wide study in all available human mRNAs identified many of them have the potential to form the same structure, which may involve in regulation of translation initiation, or mRNA stability. Finally, preliminary design of CRISPR-inspired strategy to induce -1 ribosomal frameshifting by trans-acting factors was verified to be functional. Our findings have significant implications in expanding the repertoire of the PRF phenomenon in both human and viral mRNAs and the protein-coding capacity of the human transcriptome.

Pseudoknots are a special kind of RNA structures that play functional roles in a wide variety of biological processes. Pseudoknots are best known for their involvement in the −1 programed ribosomal frameshifting (−1 PRF) and stop codon readthrough translational recoding events as the stimulatory structures. In this dissertation, three large scale bioinformatics investigations were carried out on the roles of pseudoknots in the −1 PRF, as well as stop codon readthrough, recoding mechanisms in viral and human mRNAs. To meet the specific needs of the bioinformatics investigations, a new algorithm and method for the detection of RNA pseudoknots has been developed. The new approach differs from all existing pseudoknot detection programs in that it is capable of identifying all potential pseudoknots in any given RNA sequence with no length limitation, in a time efficient manner. This capability is essential for large scale applications in which large datasets of long RNA sequences are analyzed. The algorithm and method have been implemented, with different flavors, in three large scale sequence analysis investigations. The three datasets of mRNA sequences are: 1) full-length genomic mRNA sequences of all animal viruses known or expected to use the −1 PRF and stop codon readthrough recoding mechanisms for viral protein production; 2) full-length genomic mRNA sequences of 4000 plus different strains of human immunodeficiency virus type-1 (HIV-1); 3) 34,000 plus full-length human mRNA sequences. Results from systematic sequence analysis on these three datasets prove the usefulness and robustness of the newly developed pseudoknot detection approach. A large number of previously unknown potential pseudoknots were detected in the viral and human mRNA sequences under investigation. Post detection analysis leads to new mechanistic insights and hypotheses of pseudoknot dependent translational recoding. Some unifying themes of RNA pseudoknot structures in general are also uncovered. The results provide solid basis for further experimental and bioinformatics studies in the future.