Search

N6-methyladenosine (m6A) is the most abundant internal messenger RNA (mRNA) modification. It affects nearly every aspect of mRNA metabolism such as splicing, transport and decay. m6A is a dynamic modification: a "writer" complex is responsible for its installation, whereas two "eraser" enzymes can remove this modification. m6A regulates gene expression by attracting effector or "reader" proteins that bind to it. YTH domain-containing family (YTHDF) proteins are a group of three cytoplasmic proteins YTHDF1, YTHDF2 and YTHDF3; that bind to m6A through the C-terminal YTH-domain. Among these proteins, YTHDF2 is the best characterised and it recruits the CCR4-NOT complex to mediate transcripts degradation by de-adenylation. The role of YTHDF2 in regulating the metabolism of transcripts is of central importance in several developmental, physiological and disease contexts. In zebrafish, YTHDF2 is required for the clearance of maternal transcripts during maternal to zygotic transition (MZT). Although in mice YTHDF2 is essential for oocyte maturation, it is still unknown whether YTHDF2 also has a key function during MZT. In this thesis I show that YTHDF2 is also required during MZT in mice. Furthermore, while it has been proved in vitro that YTHDF2 is an m6A reader, it has not been demonstrated that the functions attributed to YTHDF2 in vivo are dependent on m6A-binding or whether YTHDF2 has m6A-independent functions. Here I demonstrate that YTHDF2 binding to m6A-RNA is essential for YTHDF2 functions in vivo. Crystallography studies and in vitro assays have showed that YTHDF proteins all bind to m6A through a conserved aromatic cage and with similar binding affinity. Additionally, the YTHDF readers have a common RNA-binding motif. Therefore, it is debated whether YTHDF proteins function redundantly or whether they evolved context-dependent specific functions. Here I show that the activity of the YTHDF proteins in vivo is context-dependent. In summary, the key regulative role of YTHDF2 in different physiological contexts is linked to its role as an m6A reader, and, while YTHDF2 and YTHDF1 are redundant in certain contexts, they have evolved tissue-specific functions.

RNA binding proteins have essential roles in post-transcriptional gene regulation, evidenced by the fact that abnormal expression of these proteins is linked to many diseases. The main focus of this dissertation is the germ cell-restricted RNA binding protein DAZL (Deleted in Azoospermia Like), a member of the DAZ family (DAZ, DAZL, and BOULE). These proteins were first shown to be essential for spermatogenesis in humans and mice over two decades ago. Thus, while the biological importance of the DAZ proteins is clear, many critical questions have remained, including the identities of their direct RNA targets, how these RNAs are regulated, and why loss of this regulation results in germ cell defects.Here, we generated high-resolution, transcriptome-wide maps of DAZL-RNA interactions in mouse testes. These maps reveal Dazl binding to thousands of mRNAs, predominantly through sequence-specific interactions near the polyA tail. Using Cre-lox technology in transgenic mice and fluorescence activated cell sorting (FACS), we isolated Dazl knockout germ cells and used RNA-Seq to identify mRNAs sensitive to DAZL ablation. Intersection of the RNA-Seq and DAZL-RNA interaction datasets revealed that DAZL enhances expression of a subset of directly-bound transcripts, namely mRNAs for a network of essential cell cycle regulatory genes and a number of genes critical for spermatogenesis. We also demonstrate that DAZL’s polyA-proximal binding is facilitated by interactions between PABPC1 and the mRNA polyA tail, uncovering a novel mechanism for recruitment of an RNA-binding protein to mRNA. Our observations reveal the critical function of DAZL as a master regulator of a specific mRNA program necessary for germ cell survival in mammals. Given the functional conservation of DAZ family proteins, these findings also provide important insights into the molecular basis for azoospermia in 10-15% of infertile men with Y chromosome deletions affecting the DAZ gene. Overall, the work presented in this dissertation critically advances our understanding of the molecular mechanisms that control mammalian germ cell maintenance.

RNA, previously recognized merely as a messenger of genetic information, has been recently rediscovered as a versatile molecule with a central role in cellular regulation. These regulatory functions are enabled by its specific chemical makeup that allows it to fold into intricate and flexible structures. In stark contrast with DNA, RNA forms a variety of structural motifs that serve as efficient points of contact in molecular recognition. It is therefore clear, that dynamic RNA structures dictate the binding availability of interfaces that play important roles in molecular regulation inside living cells. As such, the need for tools that can accurately capture and predict RNA structure in vivo continues to be essential to understand RNA function. To this end, my dissertation focuses on the development of molecular tools to predict and characterize accessible RNA interfaces in their native environment. First, I established the usefulness of a fluorescence-based in vivo oligonucleotide hybridization approach to identify accessible interfaces by characterizing numerous RNA regions in several biologically relevant molecules in E. coli. I then described these RNA interactions using a biophysical model based on thermodynamic principles and incorporating large sets of data collected using this fluorescence-based system. This approach displayed improved prediction capabilities of RNA accessibility compared to un-optimized versions without incorporation of in vivo data. Finally, I detailed the development and application of a high throughput tool for the large-scale characterization of accessible interfaces within native RNAs in a single experiment. In this approach, in vivo oligonucleotide hybridization was coupled to transcriptional elongation control to allow analysis via next generation sequencing. This tool was used to obtain complete landscapes of functional structure for 72 regulatory molecules in a single experiment (>1000 regions). Altogether the results of this high throughput approach revealed a pattern indicating that RNA-RNA interaction sites are either highly accessible or highly protected, suggesting their binding status (e.g. actively bound or unbound). In addition, within bacterial small RNAs, our approached revealed the role of the global regulator Hfq as universal structural relaxer. The compendium of these tools provides a unique and fundamental perspective in the study of functional RNA structure, namely, the identification of dynamic structures. Furthermore, the information provided by these approaches significantly aids in the design of synthetic RNAs for a variety of purposes, including gene expression control.

Shigella bacteria cause 165 million annual cases of Shigellosis, resulting in 1.1 million fatalities each year. With growing antibiotic resistance and no vaccine, new methods must be developed to treat Shigellosis. Outer membrane protein A, or OmpA, is required for Shigella virulence and may be a good target for inhibitory drugs. Before these new treatments can be developed, we need to understand the regulation of ompA and its role in virulence. In a previous study, it was found that ompA mRNA (the molecular template for making protein) levels in E. coli are decreased at higher temperatures while OmpA protein levels remain the same. This suggests that translation efficiency is affected by temperature dependent regulation. Protein is more efficiently made from the lesser amount of mRNA present at higher temperatures. This project determined that S. dysenteriae ompA has a similar temperature-dependent regulation. Future studies will be conducted to determine if this temperature dependent-regulation is mediated by a putative FourU RNA thermometer in the 5’ UTR of ompA. Additionally, future studies will examine the contribution of this putative RNA thermometer to S. dysenteriae virulence.

Historically, the ribosome has been viewed as a ribozyme with constitutive rather than regulatory capacity in mRNA translation. However, emerging studies reveal that ribosome activity may be highly regulated to impart a new layer of specificity in control of gene expression and organismal biology. Heterogeneity in the composition of the ribosome may generate variations to the translation machinery that significantly impact on how the genomic template is translated into functional proteins. Studies in our laboratory have shown that ribosomal protein RPL38 confers specificity to the ribosome and is responsible for specifically translating a subset of Homeobox (Hox) mRNAs to regulate the mammalian body plan. In Rpl38 mutant embryos, global protein synthesis is unchanged while the translation of 8 of the 39 Hox mRNAs is specifically affected. Unexpectedly, Rpl38 transcripts are highly enriched in a tissue specific manner, most notably in somites and within the neural tube. Together, these studies led us to propose that ribosome in distinct regions of the embryo may be highly specialized to exert transcript-specific translational control. However how transcripts are recognized and selected by RPL38 remains unknown.Here I uncover a unique set of RNA regulons embedded in the 5'UTRs of Hox genes which are critical for ribosome-directed control of their spatiotemporal expression patterns. These structured RNA elements, resembling Viral Internal Ribosome Entry Sites (IRES), are positioned close to the AUG of subsets of Hox mRNA to interact with ribosomal proteins and directly recruit the ribosome. Using Hoxa9 as a model, I find that the IRES element folds into a stable structure in vitro and can recruit ribosomes directly to the mRNA. By generating the first targeted knock-out of an IRES element in the mouse, I show that the Hoxa9 transcript, despite numerous layers of transcriptional control, requires an IRES-element for accurate spatiotemporal expression and axial skeletal patterning. The unexpected requirement for specialized translation control of Hox transcripts is further explained by the activity of a novel Translational Inhibitory Element (TIE), which inhibits general cap-dependent translation, thereby converting a generic mode of mRNA translation into a highly specialized one. Together, these studies uncover a new paradigm of specialized translational regulation that is achieved through a combination of unique RNA regulons to facilitate ribosome-mediate control of gene expression.

Our ability to efficiently and predictably program cells is central to the fields of bioengineering and synthetic biology. Once thought to be a passive carrier of genetic information, RNA is now more appreciated as the main organizer of cellular networks. To harness the unique abilities of RNA molecules for programming cells, we show here how to rationally design novel synthetic RNA elements to recapitulate the functions of natural noncoding RNAs (ncRNAs), and how to assemble these synthetic elements into higher-order biological systems. To create synthetic RNA elements, we start with two primary types of ncRNA-mediated natural systems. Both modulate RNA-level regulatory signals encoded in the 5' untranslated region, and are mediated by ncRNAs. In the first system the ncRNA represses transcription elongation, whereas in the second system the ncRNA inhibits translation initiation. To create orthogonal RNA elements that work independently in the same cell, we systematically modify the RNA-RNA interaction in the natural systems. Our characterization results in families of orthogonally acting RNA elements for both transcription and translation controls. Furthermore, we develop mathematical thermodynamic models to predict new RNA elements in silico for translation controls. To engineer synthetic RNAs to sense and integrate cellular signals, we design allosteric RNA chimera molecules by fusing ncRNAs to RNA aptamers. We demonstrate the design principles for creating such chimeric RNA molecules that can sense proteins or small molecules and control transcription or translation. We show that the design strategy is modular, which allows us to reconfigure different ncRNA mutants and RNA aptamers to engineer orthogonal RNA chimeras that respond to different ligands and regulate different gene targets. We further show that multiple RNA chimeras allow logical integration of molecular signals in the same cell for cellular information processing. We assemble multiple synthetic RNA elements to create basic regulatory network motifs. These include independent control, logic control, and cascading control. We characterize the performance and properties of these engineered RNA circuits such as their time response, signal sensitivity, and noises across cell populations. We further explore a strategy that can effectively convert orthogonal translational regulators into orthogonal transcriptional regulators, which can be used to perform multi-input logic computation. In an effort to engineer feedback circuits, we demonstrate the use of translational repressor and activator based on RNA-binding proteins. The designed positive or negative feedback circuits form a basis for programming complex functions.To improve the predictability of engineered biological systems, we develop a synthetic RNA processing platform from the bacterial CRISPR genetic immune pathway. The synthetic RNA processing system can efficiently and specifically cleave desired precursor mRNAs at designed loci. Using this system, we show that transcript cleavage enables quantitative programming of gene expression by modular assembly of promoters, ribosome binding sites, cis regulatory elements, and riboregulators. These basic components can be grouped into multi-gene synthetic operons that behave predictably only after RNA processing. Physical separation of otherwise linked elements within biological assemblies allows design of sophisticated RNA-level regulatory systems that are not possible without it. Thus, our results exemplify a crucial design principle based on controllable RNA processing for improving the modularity and reliability of genetic systems. To sum, our work established bacterial ncRNAs as an intriguing engineering substrate for scalable genetic circuit design and for programming cells. We provide a set of engineering principles for designing synthetic RNA elements as well as using them to sense signals and form genetic circuits. Our RNA-based engineering platform provides a versatile and powerful strategy for designing higher-order cellular information processing and computation systems, which can be readily applied to practical applications including chemical production, environment remediation, and therapeutics.