Since the discovery of transfer RNAs (tRNAs) as decoders of the genetic code, life science has transformed. Particularly, as soon as the importance of tRNAs in protein synthesis has been established, researchers recognized that the functionality of tRNAs in cellular regulation exceeds beyond this paradigm. A strong impetus for these discoveries came from advances in large-scale RNA sequencing (RNA-seq) and increasingly sophisticated algorithms. Sequencing tRNAs is challenging both experimentally and in terms of the subsequent computational analysis. In RNA-seq data analysis, mapping tRNA reads to a reference genome is an error-prone task. This is in particular true, as chemical modifications introduce systematic reverse transcription errors while at the same time the genomic loci are only approximately identical due to the post-transcriptional maturation of tRNAs. Additionally, their multi-copy nature complicates the precise read assignment to its true genomic origin. In the course of the thesis a computational workflow was established to enable accurate mapping of tRNA reads. The developed method removes most of the mapping artifacts introduced by simpler mapping schemes, as demonstrated by using both simulated and human RNA-seq data. Subsequently, the resulting mapping profiles can be used for reliable identification of specific chemical tRNA modifications with a false discovery rate of only 2%. For that purpose, computational analysis methods were developed that facilitates the sensitive detection and even classification of most tRNA modifications based on their mapping profiles. This comprised both untreated RNA-seq data of various species, as well as treated data of Bacillus subtilis that has been designed to display modifications in a specific read-out in the mapping profile. The discussion focuses on sources of artifacts that complicate the profiling of tRNA modifications and strategies to overcome them. Exemplary studies on the modification pattern of different human tissues and the developmental stages of Dictyostelium discoideum were carried out. These suggested regulatory functions of tRNA modifications in development and during cell differentiation. The main experimental difficulties of tRNA sequencing are caused by extensive, stable secondary structures and the presence of chemical modifications. Current RNA-seq methods do not sample the entire tRNA pool, lose short tRNA fragments, or they lack specificity for tRNAs. Within this thesis, the benchmark and improvement of LOTTE-seq, a method for specific selection of tRNAs for high-throughput sequencing, exhibited that the method solves the experimental challenges and avoids the disadvantages of previous tRNA-seq protocols. Applying the accurate tRNA mapping strategy to LOTTE-seq and other tRNA-specific RNA- seq methods demonstrated that the content of mature tRNAs is highest in LOTTE-seq data, ranging from 90% in Spinacia oleracea to 100% in D. discoideum. Additionally, the thesis addressed the fact that tRNAs are multi-copy genes that undergo concerted evolution which keeps sequences of paralogous genes effectively identical. Therefore, it is impossible to distinguish orthologs from paralogs by sequence similarity alone. Synteny, the maintenance of relative genomic positions, is helpful to disambiguate evolutionary relationships in this situation. During this thesis a workflow was computed for synteny-based orthology identification of tRNA genes. The workflow is based on the use of pre-computed genome-wide multiple sequence alignment blocks as anchors to establish syntenic conservation of sequence intervals. Syntenic clusters of concertedly evolving genes of different tRNA families are then subdivided and processed by cograph editing to recover their duplication histories. A useful outcome of this study is that it highlights the technical problems and difficulties associated with an accurate analysis of the evolution of multi-copy genes. To showcase the method, evolution of tRNAs in primates and fruit flies were reconstructed. In the last decade, a number of reports have described novel aspects of tRNAs in terms of the diversity of their genes. For example, nuclear-encoded mitochondrial-derived tRNAs (nm-tRNAs) have been reported whose presence provokes intriguing questions about their functionality. Within this thesis an annotation strategy was developed that led to the identification of 335 and 43 novel nm-tRNAs in human and mouse, respectively. Interestingly, downstream analyses showed that the localization of several nm-tRNAs in introns and the over-representation of conserved RNA-binding sites of proteins involved in splicing suggest a potential regulatory function of intronic nm-tRNAs in splicing.

RNA molecules are diverse in form and function. They include messenger RNAs (mRNAs) that are templates for proteins, splice products such as introns that can generate functional noncoding RNAs, and a slew of smaller RNAs such as transfer RNAs (tRNAs) that help decode mRNAs into proteins. RNAs can show distinct patterns of subcellular localization that play an important role in protein localization. However, RNA distribution in cells is incompletely understood, with prior studies focusing primarily on RNAs that are long (>200 nucleotides), fully processed, and polyadenylated. We examined the distribution of RNAs in neurons. Neuronal compartments can be separated by long distances and play distinct roles, raising the possibility that RNA localization is especially overt and functionally meaningful in these cells. In our exploration, we physically dissected projections from cell bodies of neurons from the rat brain and sequenced total RNA. We describe two main findings. First, we identified excised introns that are enriched in neuronal projections and confirmed their localization by single- molecule fluorescence in situ hybridization. These are a previously unknown set of circular RNAs in neuronal projections: tailless lariats that possess a non- canonical C branchpoint. Second, we observed a highly abundant population of small (20-150 nucleotide) RNAs in neuronal projections, most of which are tRNAs. For both circular introns and tRNAs, we did not observe known RNA localization signals. Thus, many types of RNA, if sufficiently stable, appear free to diffuse to distant locations, their localization perhaps aided by the movement of large organelles in the confines of neuronal projections. Our survey of RNA molecules across subcellular compartments provides a foundation for investigating the function of these molecules and the mechanisms that localize them.

Ancestral environmental conditions can instruct offspring development, although the mechanism(s) underlying such transgenerational epigenetic inheritance is unclear. In murine models focused on paternal dietary effects, we and others have identified tRNA fragments (tRFs) in mature sperm as potential carriers of epigenetic information. In our search for molecular targets of specific tRFs, we observed that altering the level of 5’-tRF Glycine-GCC (tRF-GG) in mouse embryonic stem cells (mESCs) and preimplantation embryos modulates the expression of the endogenous retrovirus MERV-L and genes regulated by MERV-L. Intriguingly, transient derepression of MERV-L is associated with totipotency of two-cell stage embryos and a subset of two-cell-like mESCs. Here, I reveal the mechanistic basis for tRF-GG regulation of MERV-L. I show that tRF-GG supports the production of numerous small nuclear RNAs associated with the Cajal body, in mouse and human embryonic stem cells. In particular, tRF-GG modulates the levels of U7 snRNA to ensure an adequate supply of histone proteins. This in turn safeguards heterochromatin-mediated transcriptional repression of MERV-L elements. Importantly, tRF-GG effects on histone mRNA levels, activity of a histone 3’UTR reporter, and expression of MERV-L associated transcripts can all be suppressed by appropriate manipulation of U7 RNA levels. I also show that hnRNPF and H bind directly to tRF-GG, and display a stark overlap of in vivo functions to tRF-GG. Together, this data uncovers a conserved mechanism for a 5’ tRNA fragment in the fine-tuning of a regulatory cascade to modulate global chromatin organization during pre-implantation development.

No description available.

No description available.

There is strong crystallographic evidence that the 23S rRNA is the only catalytic entity in the peptidyl transferase center. Various mechanisms for the catalysis of peptidyl transfer have been proposed. Recently, attention has been given to the idea that the 23S rRNA simply acts to position the tRNA for spontaneous peptidyl transfer and that chemical catalysis may play only a secondary role. Conserved nucleotides U2585 and U2506 are thought to be involved in positioning the 3’ ends of A- and P-site substrates based on the crystallographic evidence, and because mutagenesis at these sites severely impairs peptide bond formation. In this study, pure populations of ribosomes with either U2585A or U2506G mutations in the 23S rRNA were analyzed to test the hypothesis that substitutions at nucleotides U2585 and U2506 in the peptidyl transferase center impair peptide bond formation by altering the position of the 3’ end of P-site tRNA relative to the 23S rRNA. Pure populations of mutant or wild-type ribosomes were obtained by an affinity tagging system and probed with 32P-labeled [2N3A76]tRNAPhe to determine how the 3’ end of tRNA interacts with the ribosomal proteins and 23S RNA at the peptidyl transferase center. Some of the data for the ribosomes with a G at position 2506 are consistent with a model suggested by Schmeing and coworkers in which nucleotide U2506 breaks from its original wobble base pair with nucleotide G2583 during A-site tRNA binding and swings towards the 3’ end of P-site tRNA, while nucleotide U2585 simultaneously moves away from the 3’ end of P-site tRNA.

No description available.

No description available.

No description available.

No description available.