Browsing by Subject "RNA binding protein"

Germ cells represent the immortal line: they are guardians of a totipotent genome and are essential for the genetic survival of an individual organism and ultimately a species. An error at any stage in development (specification, migration, colonization, differentiation, adult maintenance) can lead to one of two disastrous outcomes: (1) germ cell death or (2) unchecked growth and proliferation leading to tumorigenesis. The work in this dissertation utilizes a classic mouse model (Ter) resulting in both of these phenotypes to further explore the molecular mechanisms important for development of germ cells.

A homozygous nonsense mutation (Ter) in murine Dnd1 (Dnd1Ter/Ter) results in a significant (but not complete) early loss of primordial germ cells (PGCs) prior to colonization of the gonad in both sexes and all genetic backgrounds tested. The same mutation also leads to testicular teratomas only on the 129/SvJ background. Male mutants on other genetic backgrounds ultimately lose all PGCs with no incidence of teratoma formation. It is not clear how these PGCs are lost, develop into teratomas, or what factors directly control the strain-specific phenotype variation.

Work here demonstrates that Dnd1 expression is restricted to germ cells and that the Ter mutant defect is cell autonomous. The early loss of germ cells is due in part to BAX–mediated apoptosis which also affects the incidence of tumorigenesis on a mixed genetic background. Moreover, tumor formation is-specific to the male developmental pathway and not dependent on sex chromosome composition of the germ cell (XX vs. XY). Despite normal initiation of the male somatic pathway, mutant germ cells fail to differentiate as pro–spermatogonia and instead prematurely enter meiosis.

Results here also reveal that, on a 129/SvJ background, many mutant germ cells fail to commit to the male differentiation pathway, instead maintain expression of the pluripotency markers, NANOG, SOX2, and OCT4, and initiate teratoma formation at the stage when male germ cells normally enter mitotic arrest. RNA immunoprecipitation experiments reveal that mouse DND1 directly binds a group of transcripts that encode negative regulators of the cell cycle, including p27Kip1, which is not translated in Dnd1Ter/Ter germ cells. Additionally, overexpression of DND1 in a teratocarcinoma cell line leads to significant alteration of pathways controlling the G1/S checkpoint and the RB tumor suppressor protein. This strongly suggests that DND1 regulates mitotic arrest in male germ cells through regulation of cell cycle genes, serving as a gatekeeper to prevent the activation of a pluripotent program leading to teratoma formation. Furthermore, strain–specific morphological and expression level differences possibly account for sensitivity to tumor development.

RNA localization provides the blueprint for compartmentalized protein synthesis in eukaryotic cells. Current paradigms indicate that RNAs encoding secretory and membrane proteins are recruited to the endoplasmic reticulum (ER), via positive selection of a `signal peptide' tag encoded in the protein. Thus RNA sorting to the ER follows protein sorting and the RNA is considered a passive player. However, RNAs have been shown to access the ER independent of the signal peptide and display a wide range of affinities to the ER that does not correlate with signal peptide strength. How and why mRNAs localize to the ER to varying extents and whether such localization serves a purpose besides protein sorting is poorly understood. To establish the cause and consequence of RNA binding to the ER membrane, I pose three primary questions: 1. How are mRNAs targeted to the ER? 2. Once targeted, how are mRNAs anchored to the ER membrane? 3. Are ER localized mRNAs subject to transcript-specific regulation?

I address cytosolic mRNA targeting to the ER by comparing the partitioning profiles of cytosolic/nuclear protein-encoding mRNA population (mRNACyto) to that of mRNAs encoding a signal peptide (mRNAER). I show that, at a population level, mRNACyto display a mean ER enrichment that is proportional to the amount of ER-bound ribosomes. Thus, I propose that targeting of mRNACyto to the ER is stochastic and over time, the specific interactions engaged by an individual mRNACyto with the ER determines its steady state partitioning profile between the cytoplasm and the ER.

To address the modes of direct binding of mRNA to the ER, I examined the association of various RNA populations with the ER after disrupting membrane-bound ribosome's interaction with its ER receptor. mRNACyto and most of mRNAs encoding secretory proteins (mRNACargo) are released upon disruption of ribosome-receptor interactions, indicating no direct mRNA-ER interactions. However, the population of mRNAs that encode resident proteins of the endomembrane organelles such as the ER, lysosome, endosome and the Golgi apparatus (mRNARes) maintain their association with the ER despite the disruption of ribosome-receptor interactions. These results indicate direct binding of mRNARes to the ER, further suggesting that the function of the encoded proteins dictates the mode of association of corresponding mRNA with the ER.

To uncover the mode of mRNARes binding directly to ER, I performed differential proteomic analysis of cytosolic and membrane bound RNA-protein complexes, which revealed a network of RNA binding proteins that interact uniquely with the ER-anchored mRNAs. The anchoring of endomembrane resident protein-encoding RNAs to the ER through these RNA binding proteins may reflect an imprinting of the ER with the information necessary for the continued biogenesis of the endomembrane organelle system even in situations where translation-dependent ER targeting of an mRNA is compromised.

Finally, I address whether ER-bound mRNAs can be regulated differentially by comparing the fates of two signal peptide-encoding RNAs, B2M and GRP94, during the unfolded protein response (UPR). I show that in response to ER stress, GRP94 mRNA, but not B2M, relocates to stress-induced RNA granules, thus escaping an RNA decay program that operates at the ER membrane during the UPR. Hence, I propose that the mode of RNA association to the ER is subject to regulation and influences the fate of RNAs during cellular stress. Thus, by demonstrating diverse modes of mRNA localization to the ER and differential regulation of ER bound mRNAs during cellular stress, my work has helped establish an emerging role for the ER as a post-transcriptional gene regulatory platform.

RNA-binding proteins (RBPs) play an important role in post-transcriptional gene regulation by binding mRNAs to regulate splicing, stability, nuclear export, and target them for translation at specific subcellular compartments. While the role of RBPs in localization of mRNA to different subcellular compartments has been expanding in recent years, little has been done to understand their role in targeting mRNA for translation at the ER. The recent introduction of UV-based RNA-centric methods for proteome-wide identification of RNA-interactors has revealed numerous candidate RBPs on the ER. However, validation for most of these candidate integral ER RBPs is lacking, and the biological function of their RNA-binding activity is unknown. To address this, we introduce signal-to-noise (S:N)-based analytical approaches for direct validation of RNA-binding and reveal previously unrecognized sources of experimental noise inherit to UV-based RNA-centric methods which confound meaningful interpretation and introduce false positives. To overcomes these issues, we devise a novel biochemical method termed LEAP-RBP (Liquid-Emulsion-Assisted-Purification of RNA-Bound Protein) for the selective isolation of UV-crosslinked RNA-protein adducts and introduce new signal-to-noise (S:N)-based metrics for distinguishing relevant vs non-specific RNA-interactors. Using this approach, we identify integral ER RBPs with potential RNA-regulatory roles and provide a solid foundation for the proper identification and characterization of RNA-protein interactomes.

SR proteins are a family of splicing factors involved in the regulation of both constitutive and alternative splicing of pre-mRNAs. Despite years of studies, several big questions still remain: how the expression levels of SR proteins are regulated; what are the underlying mechanisms responsible for SR proteins-mediated gene regulation; what are the physiological targets of SR proteins in vivo. In my dissertation study, I am focusing on two members of the family, SF2/ASF and SRp20, to study their functional involvement in regulating microRNA/mRNA biogenesis and their own expression.

Negative feedback regulation is a common mechanism maintaining the steady-state level of SR proteins (i.e. SC35 and SRp20), and several mechanism may be involved. In order to test if miRNAs are also involved in such negative feedbacks, small RNA sequencing was used to identify differentially expressed miRNAs after SF2/ASF overexpression in an inducible stable cell line system. Among the 40 differentially expressed miRNAs, miR-7 is particularly interesting, because it is also predicted to target SF2/ASF, which forms a negative feedback regulation. This is indeed the case as shown by luciferase reporter assay and overexpression/knocking down of miR-7 in vivo. To our knowledge, this is the first identified negative feedback circuit between a SR protein and a miRNA, which may be a general mechanism in regulating SR protein homeostasis.

To characterize the mechanism underlying SF2/ASF-enhanced miRNA biogenesis, I have employed a series of molecular and biochemical approaches to pinpoint the key molecular interactions in a minigene system, which is consist of miR-7 embedded intron and the flanking exons of its host gene. By manipulating the splicing pattern of such minigene, I have uncovered a splicing-independent function of SF2/ASF in regulating miRNA biogenesis. Directly binding between SF2/ASF protein and pri-miR-7 was demonstrated by Cross-linking and immunoprecipitation assay (CLIP) and RNA affinity purification. The precise binding site was then pinpointed by combining computational prediction and mutagenesis assay. Finally, by using in vitro pri-miRNA processing assay, I showed that SF2/ASF can promote the Drosha cleavage step of pri-miR-7 through directly association with the predicted binding site. So far, this is the first SR protein discovered, which is directly involved in miRNA biogenesis. Moreover, our preliminary data also suggested that SF2/ASF may promote miRNA biogenesis in other steps after Drosha cleavage; and different SR proteins can regulate miRNA biogenesis in a substrate-specific manner. Taken together, SR family of splicing factors may be broadly involved in miRNA biogenesis through direct interactions.

In order to study the general involvement of SR proteins in RNA biogenesis, one important step stone is to have a better profile of their targets in vivo. To achieve this, I focused on SRp20, another classic SR protein. Photoactivatable-Ribonucleoside-Enhanced Cross-linking and immunoprecipitation assay combined with deep sequencing (PAR-CLIP-seq) was used to identify the binding partners of SRp20 globally, which is subsidized by candidate gene validations. Consistent with the literature, I found that SRp20 primarily targets exonic regions for splicing regulation, and such interactions are likely to be sequence dependent on the CWWCW motif. Surprisingly, I also observed extensive binding between SRp20 and the 3' UTRs of mRNA, which may affect the choice of alternative polyadenylation sites. The underlying mechanisms are being investigated by a variety of molecular methods.

In summary, I have identified a subset of miRNAs, the expression of which can be regulated by SF2/ASF in a splicing independent manner. This is the first SR protein identified in regulating miRNA biogenesis. One of the upregulated miRNAs, miRNA-7 can form a negative feedback with SF2/ASF by negatively regulating the expression of SF2/ASF on translational level. By using PAR-CLIP method, I have identified the genome-wide binding partners of SRp20 in vivo. When SRp20 binds to the exonic regions, it potentially affects the alternative splicing patterns of nearby introns. Interestingly, the 3' end choices for a subset of genes may be regulated by SRp20 through directly binding, which may be a new mechanism for the regulation of 3' end processing.