The patterns and mechanisms by which eukaryotic cells regulate the expression of their genetic information are highly complex and intricate. The transmittance of this information from nuclear repository to cytoplasmic translation contains within it several steps, including the selective removal and concomitant joining of pieces of information in a process called alternative splicing. The projects detailed within this document describe the regulation of alternative splicing through the interaction of specific proteins with specific pre-mRNA transcripts.

The Rio lab has studied PSI, a protein involved in the regulation of the P element transposase transcript, for many years. It has since been shown to regulate the splicing of hundreds of other transcripts. The experiments described here look at the organization of PSI and other proteins on the P element transcript by site-specific labeling of the transcript using radioactive 32P. We also investigate two phosphorylation events of PSI, identifying the kinases responsible and demonstrate that these events may change the protein-protein interaction partners of PSI.

It has become increasingly apparent that alternative splicing may not only be regulated by protein/RNA interactions, but also by RNA/RNA interactions. To probe this, we designed experiments to test if some well-known small RNA-associated proteins are regulating alternative splicing. Using splice junction microarrays, we determined that Argonaute-2 (Ago-2) regulated the splicing of over 100 splice junctions, and further experiments using ChIP-seq and mRNA-seq of Ago-2 mutants revealed that Ago-2 also has a role in transcriptional repression, possibly through being incorporating in complexes composed of polycomb-group genes. We also used CLIP-seq to determine the RNA binding profile and preferences of Ago-2 in Drosophila tissue culture cells.

Finally, we characterized the functions of a Drosophila specific splicing factor called LS2. LS2 is orthologous to the highly conserved splicing factor dU2AF50, but its origin through retroduplication and subsequent divergence to acquire distinct sequence specificity, expression pattern, and function show it to be an interesting case in the evolution of alternative splicing regulation. This may be a mechanism that underlies the existence of some members of the large families of splicing factors, including hnRNP proteins and SR proteins. That is, by duplicating functional copies of genes, cellular systems create new proteins to tinker with and acquire new functions while keeping the former functionality and stability of the parent protein.

While these projects are essentially independent of each other, they all fall under the umbrella of protein regulation of RNA metabolism and hopefully contribute to a more complete understanding of the regulation of gene expression.

Efficient P element transposition is dependent on the expression of a transposon-encoded transposase and the recruitment of endogenous Drosophila DNA repair proteins to sites of transposase induced DNA breaks. Previous biochemical analysis of proteins bound to the 31 bp inverted repeats identified an 18 kDa basic leucine zipper protein (bZIP) (CG6272) termed Inverted Repeat Binding Protein 18 (IRBP18). IRBP18 forms a functional heterodimer with another bZIP protein Xrp1/CG17836, which is sufficient to bind the 31bp terminal inverted repeats in vitro. Together these proteins, in cell culture models, repress transcription from the P element promoter, as well as facilitate efficient DNA repair after transposase-mediated cleavage. Additionally flies homozygous for null mutations of IRBP18 show increased killing after somatic mobilization of P elements: consistent with an inability to repair post cleavage DNA breaks. In the absence of P elements, IRBP18-null flies are male sterile and display increased sensitivity to DNA damaging agents. IRBP18/Xrp1 heterodimer forms the site-specific binding core of a larger IRBP complex; of which several of the co-purifying proteins function in RNA interference (RNAi) pathways. Recently RNAi pathways have been shown to be integral components of general genome maintenance due to regulation of heterochromatin formation of repeated DNA elements, such as ribosomal DNA genes. In addition in this thesis Rm62, the novel role of the DEAD/H-box RNA helicase, in DNA repair is examined. Flies homozygous for a hypomorphic allele of Rm62 display similar sensitivity to DNA damaging agents as IRBP18 mutants. Interestingly, the majority of the DNA repair phenotype of Rm62 is caused by a nucleolar localized isoform of the protein. Furthermore RNAi-knockdown of the Rm62 interaction partner, Argonaute 2, also resulted in decrease in DNA repair efficiency. Taken together, these data implicate the IRBP complex as a critical component in maintaining genomic integrity and thereby connecting a new site-specific DNA binding protein complex to genome stability and DNA repair.

Transposable elements are mobile genetic sequences that are found in the genomes of nearly all organisms. DNA transposons constitute a major class of transposable elements and can move throughout the host genome by a cut-and-paste mechanism catalyzed by an encoded transposase protein. Although transposase activity can be detrimental to the host, numerous examples of host benefit have been documented. Over evolutionary time transposon-related sequences and proteins have been adapted to serve a wide range of cellular functions, a process termed transposon domestication.

The Drosophila P element is one well-studied example of a eukaryotic DNA transposable element. Although the encoded P element transposase protein has been biochemically characterized, it exhibits several features that distinguish it from the other characterized DNA transposases. Namely, P element transposase requires a guanosine triphosphate (GTP) cofactor and generates unusually long 17 nucleotide staggered DNA breaks at the transposon ends during transposition. To gain insight into the molecular basis of these distinguishing features we determined the cryo-EM structure of the Drosophila P element transposase strand transfer complex (STC) to 3.6 Å - a nucleoprotein complex in which the transposase protein is bound to P element donor DNAs covalently joined to a target DNA. Our structure reveals that the STC is dimeric, the P element donor DNAs adopt a highly unusual DNA geometry and further reveals a function for GTP in positioning the P element ends into the transposase active site for catalysis. This structure provides the first view of the P element superfamily of eukaryotic DNA transposases, offers new insights in P element transposition and implies a transposition pathway that is mechanistically distinct from other cut-and-paste DNA transposases.

Furthermore, bioinformatic and biochemical analysis have identified C2CH DNA binding domain termed the THAP domain. This novel and evolutionarily conserved domain is found across a wide range of animal genomes, including vertebrates, invertebrates, Drosophila P element transposase, in primates and in 12 human genes. Of the 12 THAP domain containing genes in humans, THAP9 is homologous to the entirety of Drosophila P element transposase, still has DNA transposase activity, but lacks the hallmarks of an active DNA transposable element. maintained. The evidence implies that THAP9 has likely been domesticated/adapted by the cell in early chordates from an ancient THAP9-like P element transposon, such as those found in Ciona. However, a cellular function for THAP9 has not been identified. In an attempt to elucidate a cellular function for THAP9, we carried out genome-editing in human embryonic stem cells (hESCs) to either knockout or epitope tag the endogenous THAP9 gene. Disruption of THAP9 did not produce overt phenotypic changes in hESCs and did not affect differentiation into fibroblast-like cells, indicating that THAP9 is likely not required for the hESC maintenance. However, endogenously epitope tagged THAP9 is translated, can be immunoprecipitated and localizes to the nucleus in hESCs. To determine potential THAP9 human genome cleavage and binding sites, we raised an antibody to purified, recombinant human THAP9 protein, performed direct in situ breaks labeling, enrichment on streptavidin and next-generation sequencing, or BLESS, to detect potential DNA cleavage site, a method used successfully to find Cas9 off-target genomic cleavage sites and ChIP-Nexus experiment, a chromatin immunoprecipitation method similar to ChIP-Exo. The ongoing analysis and comparison of both the BLESS and ChIP-Nexus sequencing data should identify genomic binding sites, potential genomic DNA cleavage sites, motifs associated with human THAP9 DNA binding and cleavage and should uncover a cellular function for the human THAP9 gene.

While these projects are essentially independent of one another, they all relate to P element DNA transposases. Together, they hopefully contribute to a deeper understanding of the mechanisms of P element transposition and the expanding roles that transposase-related proteins play in the context of cellular function in human cells.

Alternative splicing plays a crucial role in increasing the amount of protein diversity and in regulating gene expression at the post-transcriptional level. In humans, almost all genes produce more than one mRNA isoform and, while the fraction varies, many other species also have a substantial number of alternatively spliced genes. Alternative splicing is regulated by splicing factors, often in a developmental time- or tissue-specific manner. Mis-regulation of alternative splicing, via mutations in splice sites, splicing regulatory elements, or splicing factors, can lead to disease states, including cancers. Thus, characterizing how alternative splicing shapes the transcriptome will lead to greater insights into the regulation of numerous cellular pathways and many aspects of human health.

A critical tool for investigating alternative splicing is high-throughput mRNA sequencing (RNA-seq). This technology produces hundreds of millions of short (~100bp) sequencing reads from mRNA molecules and can be used to both discover novel transcripts and to quantify the expression of transcripts. While short read length is a limitation of the technology in its current form, RNA-seq has resulted in the discovery of hundreds of thousands of new transcripts and revealed an increased complexity of the transcriptome via alternative splicing, particularly in human. Here, I used RNA-seq analysis to investigate the global effect of post-transcriptional regulation via alternative splicing coupled to nonsense-mediated mRNA decay and to examine natural human variation in alternative splicing, particularly in genes associated with differential therapeutic drug response.

The nonsense-mediated mRNA decay pathway (NMD), which degrades transcripts containing a premature termination codon, plays an important role in post-transcriptional gene regulation when coupled to alternative splicing. If a gene produces an alternative isoform that is targeted by NMD, the mRNA abundance of the protein-producing transcripts can be post-transcriptionally regulated at the alternative splicing level. This has been shown to be important in the regulation of a number of genes, including many of the splicing factors themselves. I have used RNA-seq analysis on cells where NMD has been inhibited to discover alternative isoforms that are NMD targets on a genome-wide scale in human and a number of diverse other eukaryotic species. I found that around 20% of expressed human genes are potentially regulated by alternative splicing coupled to NMD and that they fall into many different functional categories. I also found that hundreds to thousands of genes produce NMD-targeted alternative isoforms in each of frog, zebrafish, fly, fission yeast, and plant, highlighting the prevalence of this relatively under-studied method of gene regulation across the three major branches of eukaryotic organisms. I also gained insight into the features that define NMD targets, which are thought to vary between species although the field is still unclear. I find that an exon-exon junction downstream of the termination codon is a much stronger predictor of NMD than 3’ UTR length in every species except yeast.

I also used RNA-seq to investigate alternative splicing in genes of pharmacologic importance. Natural human variation in the expression level and activity of genes involved in drug disposition and action (“pharmacogenes”) can affect drug response and toxicity. Previous studies have relied primarily on microarrays to understand gene expression differences, or have focused on a single tissue or small number of samples. Here, we used RNA-seq to determine the expression levels and alternative splicing of 389 selected pharmacogenes across four pharmacologically relevant tissues (liver, kidney, heart and adipose) and lymphoblastoid cell lines (LCLs), which are used widely in pharmacogenomics studies. Analysis of data from 18 different individuals for each of the 5 tissues (90 samples in total) revealed substantial variation in both expression levels and splicing across samples and tissue types. Comparison with an independent RNA-seq dataset yielded a consistent picture. This in-depth exploration also revealed 183 splicing events in pharmacogenes that were previously not annotated. Overall, this study serves as a rich resource for the research community to inform biomarker and drug discovery and use.

In conclusion, the roles of alternative splicing and NMD in the regulation of cellular processes and in human health are wide-open but critical fields of study. Advancements in sequencing technologies have had and will continue to have a huge impact on the studies of these mechanisms. New long-read technologies will likely soon be readily available and promise to greatly increase our ability to accurately interpret RNA-seq results. As the cost of sequencing continues to decrease, more and more data will be generated, allowing for a better view of how the transcriptome varies between individuals and shapes differential disease risks and drug responses.