Development consists of complex morphogenetic movements that shape individual tissues as well as the embryo. Tissues must interact both physically and through signaling molecules to coordinate the formation of various organs and the entire body plan. My thesis work sought to characterize tissue interactions that govern muscle formation during vertebrate embryo development using the African clawed frog Xenopus laevis as a model system.

Chapter 1 of this dissertation provides a general introduction to the morphogenetic events and molecular regulation leading to muscle formation. Chapter 2 presents two previously-undiscovered morphogenetic phenomena involving the muscle tissue as well as individual muscle fibers. My experiments confirmed the function and mechanisms of muscle tissue unfolding as well as the cell rearrangements that ultimately place muscle fibers into organized arrays for proper functioning. Chapter 3 summarizes my work to further analyze a previously-unreported reciprocal relationship between the neural tissue and the prospective muscle tissue. My work shows that the neural plate influences the size of the underlying muscle. Enlarging the neural plate results in a corresponding increase in the size of the muscle tissue. Here, I dissect the source and fate of ectopic muscle tissue and determine the effect of tissue size on muscle morphogenesis. As a whole, this work summarizes my research to characterize the molecular and morphological events necessary for proper muscle formation during embryo development.

The development of multicellular organisms relies on extracellular signaling pathways regulating the patterning and specification of tissues, with the same pathway often being deployed to serve many different functions. The development of Xenopus laevis has served as a classic model for identifying context specific transcriptional targets of signaling pathways and the regulatory loci responsible for their precise expression. The focus of this dissertation was to understand the regulation of the Wnt/ -catenin transcriptional target genes such that the correct patterns of expression are formed during gastrulation. During gastrulation Wnt, along with FGF, function to form the anterior to posterior axis of the embryo by inducing expression of target genes necessary for posterior identity. Target genes of the Wnt signaling pathway during gastrulation are not the same as those induced at early and later times and in addition, form different patterns. I begin in Chapter I by summarizing the early inductive events during Xenopus development that form the proper tissues for Wnt to act on as well as provide an overview of our current understanding of gene regulation and methods used to discover gene regulatory networks. In Chapter II, I explain the methods used in my dissertation. Then in Chapter III, I identify direct transcriptional targets of the Wnt signaling pathway. I found that this set of target genes are expressed around the blastopore during gastrulation, contain -catenin bound regions within 50kb of the gene, and are reduced in expression as a result of Wnt signaling knockdown. Next, in Chapter IV, I focus on two transcription factors, Sox2 and Zic1/2/3, to explore whether they are co-regulating a subset of Wnt target genes to provide context dependent activation. Lastly, in Chapter V, I aim to understand if FGF directly posteriorizes the embryos by activating either all or a subset of Wnt target genes identified in Chapter III. I found that FGF directly regulates all Wnt target genes during gastrulation. In addition, I found -catenin preferentially binds at the promoter of a subset of genes and that negative regulation in the absence of Wnt signaling may be responsible for the expression patterns of this subset of genes. Together, these results reveal the complex mechanisms of gene regulation that are necessary to insure the precise expression of Wnt and FGF target genes such that they form the anterior to posterior axis.

Over the course of development, cells and tissues of the embryo must take on the correct fates and morphologies to produce a functioning organism. The patterning events and morphogenetic processes that accomplish this task have been the subject of decades of research, the consequence of which has been a detailed comprehension of the molecular mechanisms that regulate each. Equally important is an understanding of the mechanisms that coordinate patterning with morphogenesis, such that they occur with the correct relative spatiotemporal dynamics. My thesis work sought to characterize such co-regulation in the context of two developmental events in a vertebrate model, the African clawed frog Xenopus laevis: induction of bottle cell formation at the onset of gastrulation after germ layer induction, and regulation of the morphogenetic movements of neurulation in relation to neural plate patterning.

Chapter 1 of this dissertation provides a general introduction to the patterning and morphogenetic events of early development relevant to my thesis. Chapter 2 presents a discussion of my work to characterize the function of two signaling pathways, namely Nodal signaling and Wnt/Planar Cell Polarity, in the induction of bottle cells. My experiments confirm the requirement for Nodal signaling in bottle cell induction, but do not support a role for the individual transcriptional targets of Nodal signaling tested here or for Wnt/PCP. Chapter 3 summarizes my work to address the function of two transcription factor-encoding genes, sall1 and sall4, in neural development, including their roles in anteroposterior neural patterning, neural tube morphogenesis, and neural differentiation. My work shows that both sall1 and sall4 are required for all three processes, and supports the hypothesis that their key role in this context is to transcriptionally repress stem cell factors of the pou5f3 family, allowing progression through neural development. As a whole, this work summarizes my research to characterize molecules that co-regulate early patterning and morphogenetic events in the X. laevis embryo.

RNA-binding proteins as regulators of transcription and axial patterning during Xenopus embryogenesis

by Caitlin Suzanne DeJong

Doctor of Philosophy in Molecular and Cell Biology

University of California, Berkeley

Professor Richard M. Harland, Chair

The over-arching goal of this thesis is to expand our knowledge of the mechanisms by which one cell, a fertilized egg, develops into an organism composed of multiple cell types, each with different functions and behaviors. RNA-binding proteins have been identified as potent regulators of development and embryogenesis. The studies presented in this thesis illustrate the pleiotropic effects of RNA-binding proteins in Xenopus development and will focus specifically on two RNA-binding proteins that are maternally deposited and zygotically transcribed: TAF15 and DGCR8.

TATA-binding protein-associated factor 15 (TAF15) belongs to the FET family of atypical RNA-binding proteins, which also includes Fused in sarcoma (Fus) and Ewing’s sarcoma (EWS). FET proteins were originally discovered as components of fusion oncogenes and are most noted for their implication in various cancers and neuromuscular degenerative diseases. However, little is known of the endogenous function of FET proteins. The diverse biological activities of the FET family proteins can be likened to a biological Swiss army knife; as these proteins contain domains for transcriptional activation, RNA-binding, DNA-binding, and function in both RNA Polymerase II-mediated transcription and pre-mRNA splicing. An exciting possibility is that the FET proteins may function to connect transcription and splicing. By employing the bioinformatics approach of RNA-sequencing, I generated a list of significant genes that are differentially expressed between uninjected and taf15 depleted embryos. From this analysis I found that TAF15 regulates target genes at both the transcriptional and post-transcriptional level. The studies that focus on the role of TAF15 in Xenopus development are described in chapters two and three of this thesis.

In the second chapter of this thesis I describe studies that illustrate the novel concept that a protein can regulate the same set of target genes but through different molecular mechanisms. Both maternal and zygotic TAF15 regulate the expression of the transcripts fgfr4, isl1, and pax8. Interestingly, maternal TAF15 is required for the post-transcriptional regulation of fgfr4, isl1, and pax8, regulating the splicing of single introns within these transcripts, whereas zygotic TAF15 is required for the transcriptional regulation of these genes. Therefore, the studies described in chapter two demonstrate, for the first time, that a single protein can utilize a different molecular mechanism to control the same target genes and the use of these different mechanisms of action appears to be dependent on whether the protein is maternally deposited or zygotically transcribed. Single intron retention is a known mechanism to retain transcripts in the nucleus, preventing their translation. In chapter two of this thesis I provide evidence for the following model: in the absence of genome activation, before the zygotic genome is transcribed, maternal TAF15 cooperates with a splicing factor, the RNA-binding protein SRSF4, to regulate the splicing of single introns from transcripts. As a result, TAF15 and SRSF4 control the splicing of target genes and thus control the timing of transcript maturation and subsequent translation. This mechanism is logical as it provides a mechanism by which to spatially and temporally regulate gene expression in the absence of the ability to transcriptionally regulate genes. I further show evidence that following zygotic genome is activation, zygotic TAF15 activates target gene transcription, regulating genes at the transcriptional level, likely associating with the core promoter. The findings described in chapter two of this thesis are the first to show that a single protein can regulate the same gene targets but depending on the milieu of maternal of zygotic cofactors, regulates these targets via different underlying mechanisms. The variety of functional domains intrinsic to TAF15 supports the hypothesis that this atypical RNA-binding protein could operate as part of both a splicing and transcriptional complex.

In the third chapter of this thesis I describe studies that illustrate the novel finding that TAF15 is required for dorsoventral patterning via the repression of ventx2.1. Ventx2 and BMP4 function in an autocatalytic positive feedback loop to specify ventral tissue and antagonize organizer function. Following taf15 depletion, ventx2.1 expression is expanded in the neural ectoderm and embryos exhibit a BMP overexpression phenotype: reduction in head, and dorsal, and posterior fin structures, with an increase in ventral tissue. Unlike the findings in chapter two, in this study, both maternal and zygotic TAF15 function to suppress ventx2.1 expression. These findings place TAF15 in the regulatory network of dorsoventral patterning and suggest that maternal and zygotic TAF15 control expression of ventx2.1 in a similar manner but do not rule out differential mechanisms of this control. Currently, it is unknown if TAF15 represses ventx2.1 expression directly or if TAF15 is required to activate a repressor of ventx2.1.

In the fourth chapter of this thesis I describe studies that serve as a resource for future investigations into the role of microRNAs (miRNAs) in Xenopus development. DiGeorge syndrome critical region 8 (DGCR8) is a subunit of the microprocessor complex required for miRNA biogenesis. Unlike most members (e.g. Dicer, Argonaute2) of the RNA interference biogenesis pathway, DGCR8 is required specifically for miRNA biogenesis. Furthermore, unlike previous studies in mice and zebrafish that have depleted maternal dgcr8 throughout oogenesis to look at the role of miRNAs during embryogenesis, the antisense oligodeoxynucleotide (ODN) that I have designed can be used in host transfer assays to assess the effects of maternal dgcr8 depletion once oogenesis is complete, specifically during embryogenesis. Additionally, I have designed a splice-blocking morpholino (MO) antisense oligonucleotide that targets zygotic dgcr8 for depletion. Using these two tools (ODN and MO), the first studies can be performed that tease apart the role of maternal versus zygotic DGCR8 during embryogenesis.

The work presented in this thesis exemplifies the value of carefully assessing biological functions of genes that are both maternally deposited and zygotically transcribed. The surprising finding that TAF15 utilizes distinct molecular mechanisms to control conserved target genes depending on whether this protein is maternally deposited or zygotically expressed demonstrates a new level of molecular complexity that future studies must address. Additionally, these studies further support the motivation to investigate RNA-binding proteins in development and disease as they continually prove to be multifaceted players in molecular biology.

In this thesis, I describe the transcriptional regulation of neural crest throughout neurulation of Xenopus embryos. This work focuses on the global expression dynamics of genes involved in the specification of the ectodermal territories, i.e., neural plate, pre-placodal, neural border, and epidermis. In particular, I study the function of tfap2e, a member of the tfap2 family of transcription factors that is expressed in the neural crest of Xenopus.

In the first chapter, I provide an overview of the morphogenetic movements and signaling pathways that contribute to adoption of the ectodermal fates in Xenopus embryos, and in particular, to the process of specification of the neural crest. Furthermore, I describe what is currently known about the role of Tfap2 transcription factors in vertebrate neural crest development.

In the second chapter, I characterize the functional role of tfap2e in Xenopus development. I identified tfap2e in a screen for genes that are expressed exclusively in the neural crest. Through gain and loss of function analyses, I concluded that tfap2e is required for the early processes of neural crest specification, and it is also necessary and sufficient to promote neural crest delamination, a mesenchymal property essential for neural crest cell behaviors. To better understand why tfap2e is important for these processes, I developed a transcriptomic analysis strategy to identify genes whose expression is either positively or negatively correlated with Tfap2e activity. These genes are likely to be regulated directly or indirectly by Tfap2e. The results of this study contribute to a further understanding of the roles of Tfap2 factors within the neural crest module of the gene regulatory network.

The third chapter summarizes the results of collaborative work between the Harland and Eisen labs, at U.C. Berkeley, and the Monsoro-Burq lab at the Curie Institute in Orsay, France. I describe the expression dynamics of genes known to specify the different ectodermal fates in Xenopus embryos. In addition, I describe how we obtained and analyzed the expression profiles of manually dissected ectodermal tissues throughout neurulation. The dissection dataset was utilized as a reference to evaluate how tfap2e knockdown altered the fate of the neural plate border. In agreement with the analyses from the second chapter, we conclude that loss of tfap2e delays neural crest specification, and consequently, neural plate border cells seem to adopt an expression profile that is similar to the anterior placodes.

In the fourth and final chapter, I describe the syntenic relationships between the different tfap2 clusters. The analysis provides insights into the genomic rearrangements that have occurred during the evolution of vertebrates, which may have contributed to alterations in gene expression patterns after the divergence between the common ancestor of vertebrates and tunicates. Finally, I summarize the results of chromatin immunoprecipitation to identify putative cis-regulatory elements of genes that are downstream of Tfap2e. This protein may have both activator and repressor functions that help maintain the neural crest fate over other neural plate border fates.

The work described here has contributed to our understanding of how the development the neural crest is regulated, both through focusing on a single transcription factor, tfap2e, and more broadly by characterizing gene expression dynamics in this unique cell population and important vertebrate innovation.

MicroRNAs are approximately 22-nucleotide non-coding RNAs that are important regulators of diverse biological processes. I have developed and adapted methods to study the precise function of individual microRNAs during embryonic development of the African frog, Xenopus. I began by developing an in situ hybridization protocol to study the spatiotemporal expression patterns of microRNAs during embryogenesis. Using digoxigenin-labeled probes complementary to the primary microRNA sequence I was able to determine tissue-specific expression patterns. These included numerous conserved expression patterns between Xenopus and other vertebrates, including miR-9 and miR-124 (central nervous system), miR-1 and miR-133 (muscle), and miR-10c (posterior mesoderm), as well as novel expression patterns for conserved microRNAs such as miR-23b and miR-24a. This method can also distinguish the unique expression patterns for different members of a microRNA family.

From this analysis, I selected two microRNAs for further functional studies. miR-24a is expressed in the neural retina during its development, and blocking the function of miR-24a with an antisense morpholino results in a small eye phenotype. I show that this reduction in eye size is not due to changes in patterning, specification, differentiation, or proliferation, but is due instead to an increase in programmed cell death (apoptosis). I have identified two genes important for apoptosis, caspase9 and apaf1, that are regulated targets of miR-24a. Caspase9 protein levels are increased when miR-24a is knocked down, caspase9 inhibitors can specifically rescue the knockdown phenotype, and miR-24a is able to rescue caspase9-induced apoptosis. These data strongly suggest that miR-24a is required in the developing neural retina to repress apoptosis by regulating caspase9 and apaf1.

The second microRNA that I have done extensive functional studies on is miR-133b, which is expressed in somitic mesoderm and developing hypaxial myoblasts. Knockdown of miR-133b causes a reduction in markers of hypaxial muscle differentiation without affecting specification, migration, or proliferation. At late stages, embryos lacking miR-133b function have increased levels of apoptosis in hypaxial domains and a severe reduction in body wall and head muscles derived from hypaxial myoblasts. Overexpression of miR-133b causes premature differentiation but has no effect on myoblast proliferation. Animal caps injected with the myogenic factor myoD or activin can induce miR-133b and may be useful as a secondary experimental system to identify miR-133b targets. Several computationally predicted targets are identified and discussed.

While computational predictions have become the standard method for identifying potential microRNA targets, the algorithms used to make these predictions are flawed in many ways. To ensure that I was investigating biologically relevant targets, I have attempted several different biochemical purification techniques aimed at isolating microRNA:mRNA duplexes in vivo. I have had little success with microRNA-directed RT-PCR or digoxigenin-labeled pre-microRNA immunopurification, but I am encouraged by results using biotin-labeled mature microRNAs for purification. Future refinements to this protocol must be completed before a screen for miR-133b targets is initiated.

Xenopus has provided a powerful system to study cellular, developmental, and neuro-biology. The availability of their embryos and the advent of modern molecular techniques allowed investigators to revisit the observations of classical embryologists and begin to determine the molecular mechanisms underlying germ layer formation and axis induction. My thesis work took advantage of the frog Xenopus to first address the developmental role of Noggin, a Bone morphogenic protein (Bmp) antagonist, and then to determine the mechanism of Wnt-induced anterior-posterior patterning of the neural plate.

The frog Xenopus, an important research organism in cell and developmental biology, currently lacks tools for targeted mutagenesis. In the first part of this work, I address this problem by genome editing with zinc finger nucleases (ZFNs). ZFNs directed against an eGFP transgene in X. tropicalis induced mutations that are consistent with results of non homologous end joining at the target site, resulting in mosaic loss of fluorescence phenotype at high frequencies. ZFNs directed against the noggin gene produced tadpoles and adult animals carrying up to 47% disrupted alleles. Founder animals yielded progeny that carry insertions and deletions in the noggin gene with no indication of off-target effects. Furthermore, functional tests demonstrated an allelic series of activity among three germline mutant alleles. Breeding an identified null allele to homozygosity resulted in tadpoles with deformaties in the cranial skeleton. Anatomical analysis revealed severe reductions in Meckel's cartilage with joint fusions. Gene expression analysis via in situ hybridization for chondrogenesis regulating factors in noggin mutants revealed a reduction in sox9 and col2a expression domains. Analysis of Bmp targets showed an expansion of hand2, edn1, and msx2 in the pharygeal arches (PAs) of mutants. This suggested a mechanism whereby incresed Bmp signaling inhibits chondrogenesis and ventralizes the PAs resulting in the jaw deformities observed in mutants.

Neural development in amphibians occurs as a two-step process. First, ectodermal precursors adopt a neural fate in the absence of Bmp signaling. A second signal is then required to pattern the anterior posterior neuraxis. Signaling through Fibroblast growth factor (Fgf), retinoic acid (RA), and Wnt have each been demonstrated to be both necessary and sufficient for inducing posterior fates in undifferentiated neural tissue. Wnt signaling in particular has been closely studied. However, the mechanism by which this pathway induces posterior fates remains unclear. To address this question, I used RNA-Seq to identify direct transcriptional targets in neural tissue by activating Wnt signaling in Xenopus neural explants pretreated with the translation inhibitor cycloheximide. Wnt-activated neural tissue resulted in over 200 genes with expression increased greater than 2-fold when compared to anterior neural tissue. in situ hybridization analysis of highly expressed transcription factors and RNA-binding proteins showed posterior expression. Of particular interest, the transcription factor sal-like 1 (Sall1) and sal-like 4 (Sall4) showed specific posterior neural expression suggesting a role in Wnt-induced neural patterning.

The RNA-Seq screen found sall1 and sall4 expression to be induced by canonical Wnt signaling in the presence of cycloheximide, and TCF/LEF sites present in the first intron of sall4 were enriched in β-catenin chromatin imunoprecipitations. Knockdown of Sall4 resulted in the loss of spinal cord marker expression and an increase in the expression of pou25, pou60 and pou91 (pouV genes), the three Xenopus homologs of the stem cell factor pou5f1/Oct4. Overexpression of the pouV genes resulted in the loss of spinal cord identity, and knockdown of pouV function restored spinal cord marker expression in Sall4 morphants. Finally, knockdown of Sall4 blocked the posteriorizing effects of Fgf and retinoic acid signaling in the neurectoderm. These results suggest that Sall4, activated by Wnt signaling, represses the pouV genes to provide a permissive environment that allows for additional Wnt/Fgf/RA signals to posteriorize the neural plate.

Abstract

Developmental genetics of Xenopus

by

Jessica Burnham Lyons

Doctor of Philosophy in Molecular and Cell Biology

University of California, Berkeley

Professor Richard Harland, Chair

The frog Xenopus laevis has been studied for over 200 years. Its advantages as a robust and easily manipulable model organism have been complemented more recently by genetic and genomic studies in Xenopus tropicalis. My thesis work utilized established and cutting-edge techniques to advance our knowledge of developmental genetics in both Xenopus systems.

I used an embryological approach to investigate the roles of the Fgf receptors (Fgfrs) during development. Knockdown and overexpression studies suggested that each Fgfr plays a different role in the specification of mesoderm, and my results are consistent with Fgfr4 playing a role in dampening the Fgf signal. I also showed that X. tropicalis Fgfrs 1-3 are alternatively spliced in D3, the extracellular immunoglobulin domain important for ligand specificity. These isoforms exhibit different temporal and spatial expression patterns, suggesting that control of this alternative splicing plays a role in regulating development.

My thesis work has also harnessed the power of the X. tropicalis system to understand development using a forward genetic approach. Tadpoles homozygous for the recessive lethal mutation curly exhibit ventral edema and curled tails. I used classical genetics to map curly to a 1.9 Mb window on X. tropicalis chromosome 4. The pteg gene, which lies in this region, is misspliced in curly embryos. Isabelle Philipp and I used next-generation (nextgen) sequencing technology to identify differences between the curly mutant DNA and that of the reference genome. Focusing on the region around the pteg gene, we are currently evaluating these differences to find the lesion that causes the curly phenotype.

The efficient application of genetic techniques requires a high-quality genetic map and reference genome, and those available for X. tropicalis were flawed (Wells et al., 2011) (Hellsten et al., 2010). Thus, I developed a genotyping by sequencing (GBS) technique using reduced-representation multiplex nextgen sequencing, and my collaborators in the Rokhsar lab have generated a high-quality SNP map for X. tropicalis based on my data from 192 F2 individuals. This map is being used to construct an improved X. tropicalis genome assembly. By finding regions of correlation for the pigmentation mutation gray, as well as the X. tropicalis sex locus, we demonstrated that this technique can be used for genetic mapping. I applied my GBS method to X. laevis as well, and the SNP map thus generated will facilitate the assembly of a reference genome for this allotetraploid species.

By combining the classical advantages of the Xenopus system with modern techniques, my thesis work has contributed to our understanding of the development, genetics, and genomics of vertebrate biology.

The neural crest (NC) is a vertebrate-specific population of multipotent cells, often referred to as the fourth germ layer because of its unique ability to migrate to all parts of the body and contribute to a variety of tissues. Positional information allows neural crest cells to determine their location relative to each other and in complex 3D environments, and to respond accordingly. A central question in developmental biology is how these positional cues are imparted to cells, and how the cells interpret them for the appropriate response. The work in this thesis sets out to determine 1) how Wnt and BMP gradients define cells as neural crest, 2) how the position along the AP axis determines neural crest fate and 3) how abolishing one positional cue results in a craniofacial phenotype.

Precise control of BMP and Wnt signaling in both space and time is necessary for proper neural crest induction. While the Wnt signaling pathway has been extensively studied, it is not clear how the cells interpret and respond to the Wnt signals. Knowing the importance of these pathways in setting up the embryo axes, I wanted to further our understanding of the precise way these pathways initiate the neural crest program. An important first step toward determining how the intersection of the Wnt and BMP signaling pathways set up the neural crest cell program in space and time was to develop a method to rapidly and uniformly manipulate the signaling pathways in Xenopu laevis embryos. I was able to validate that small molecules can replace the use of mRNA injections in order to manipulate signaling pathways.

The next chapter of my thesis determines how neural crest cells are specified at different axial levels. The derivatives and migratory paths of neural crest cells in the cranial and non-cranial regions are very distinct, so the gene regulatory networks are likely to differ as well. Previous work has focused on the cranial neural crest cells, and therefore, regulation of the non-cranial neural crest program remains poorly understood. By taking advantage of transgenic zebrafish lines, I show that a neural crest cell’s position along the AP axis during early development determines its fate.

In the final chapter of this thesis, I examine how abolishing one cue (Wnt signaling) affects patterning of neural crest-derived cells without affecting earlier aspects of neural crest formation, such as proliferation and differentiation. Analysis of the craniofacial structures in mutants for multiple components of non-canonical Wnt signaling provides evidence that each component is required for proper formation in different axes.