Celf1 (Cugbp1) mediated post-transcriptional gene regulation in vertebrate lens development
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
RNA-binding proteins (RBPs) control the post-transcriptional fate of mRNAs at various stages in their life, beginning at initial stages of transcript formation, in their modification and processing, translation into protein and eventually their degradation. Advances in high-throughput approaches over the past few years have identified ~1500 RBPs encoded in vertebrate genomes. However, their significance in mediating post-transcriptional control of gene expression in vertebrate organ development is not well defined. ☐ Oculogenesis, or development of the eye, involves coordination of a myriad of regulatory events in a spatiotemporal manner the net of which is to allow genetic information to be expressed into the characteristic cellular proteomes necessary for successful execution of this complex process. While we have a comprehensive understanding of the signaling factors and transcriptional control mechanisms in eye development, our knowledge on the mechanism of post-transcriptional control of gene expression in this process is limited. Studies from different vertebrate species have identified ~40 RBPs to be expressed in the eye suggesting that they may have functional significance in ocular development. In support of this hypothesis, recent findings from our laboratory have led to the characterization of two RBPs that exhibit conserved expression and function in vertebrate eye development, and whose deficiency results in early onset ocular defects. ☐ To further investigate the importance of post-transcriptional gene expression control in lens development, I applied a systems-based bioinformatics tool iSyTE (integrated Systems Tool for Eye gene discovery) to identify a new RBP Celf1 (Cugbp1) that exhibits high expression and enrichment in the lens throughout mouse embryonic development. Significantly, celf1 is highly expressed in zebrafish (Danio rerio) and frog (Xenopus laevis) lens development, suggesting a conserved expression pattern in vertebrate lens. I sought to understand whether the conserved expression of Celf1 in the developing lens was reflective of its important function in this process. To do so, in this research dissertation, I have performed a detailed investigation of the function of Celf1 in mouse lens development using germline and lens-specific conditional Celf1 gene knockout mouse models. Further, in collaboration with Dr. Jeffrey Gross and Dr. Luc Paillard, celf1 knockdown zebrafish and Xenopus morphants were generated and characterized. Together, these findings have led to a comprehensive understanding of Celf1’s function in vertebrate eye development. ☐ I have performed a detailed phenotypic characterization of Celf1 deficient mice using a variety of approaches such as light and dark-field microscopy, grid imaging, histological analysis, and scanning electron microscopy. These efforts have demonstrated that Celf1 germline or lens-specific conditional knockout mouse mutants exhibit severe lens defects that present with refractive errors, as well as lens fiber cell organization and nuclear degradation defects. Furthermore, celf1 knockdown zebrafish and Xenopus embryos also exhibit eye and lens defects. Together, these findings demonstrate that Celf1 is necessary for lens development in vertebrates. ☐ To gain insight into the molecular implications of Celf1 deficiency on the mouse lens, I performed genome-level expression profiling of newborn Celf1 lens-specific knockout lenses and identified 102 differentially regulated genes in the mutant lens compared to the control. Among the genes down-regulated in Celf1 mutant lens, I identified Dnase2b as a high-priority candidate because it encodes a lysosomal enzyme necessary for fiber cell nuclear degradation and Dnase2b knockout mice have been described to develop cataracts. Therefore, understanding the relationship between Celf1 and Dnase2b in the lens would offer a molecular explanation for a subset of the lens defects, namely the abnormal retention of nuclei in fiber cells, observed in Celf1 mutant lenses. Therefore, to investigate if Celf1 protein directly binds to Dnase2b transcripts, RNA immunoprecipitation (RIP) and Crosslinking immunoprecipitation (CLIP) analysis was performed on wild-type lens tissue samples. These assays identify Dnase2b as a direct target of Celf1 and suggest that the specific RBP-RNA (Celf1-Dnase2b RNA) interaction may be necessary for maintaining the required levels of Dnase2b in the lens. Strikingly, celf1 knockdown zebrafish morphants also displayed fiber nuclear degradation defects, suggesting the conservation of this aspect of Celf1 function in the lens across diverse vertebrate species. ☐ The above findings indicated that reduction in Dnase2b levels contribute to the lens defects observed in Celf1 deficient animals. However, the access of Dnase2b to nuclear DNA is a further requirement for proper execution of the nuclear degradation program in normal differentiating lens fiber cells. To investigate whether the lens defects observed in Celf1 mutants are exacerbated by an additional defect in the access to nuclear DNA of the already lowered amounts of Dnase2b, I investigated the cell biological aspects of nuclear degradation in these mutants. It is well defined that the phosphorylation of the nuclear membrane protein Lamin A/C is required for the disassembly of the nuclear membrane, which represents an initial critical step for access of nucleases to the genomic DNA within the nucleus. Using immunostaining, I showed that phosphorylation of Lamin A/C is reduced and its distribution around the nucleus is abnormal in Celf1 knockout mouse lenses. This finding suggests that the observed nuclear degradation defects in Celf1 mutant mouse lenses are likely due to the combinatorial effect of reduced Dnase2b levels and defective phosphorylation of Lamin A/C. ☐ Interestingly, it has been recently shown that phosphorylation of Lamin A/C is dependent on the strict spatio-temporal inhibition of a cyclin dependent kinase inhibitor, p27Kip1 protein in normal development. Specifically, p27Kip1 protein levels are up-regulated in the transition zone of the lens to commit epithelial cells to fiber differentiation and they remain high in early differentiating fibers. However, in older differentiating and maturing fiber cells, p27Kip1 protein is precipitously reduced and remains so. It is considered that in normal lens development, the reduction of p27Kip1 protein in maturing fiber cells is necessary for the activity of CDK1 (cyclin dependent kinase 1), which is required for the phosphorylation of Lamin A/C, and thereby the breakdown of fiber nuclear membrane. Therefore, to investigate if Lamin A/C phosphorylation defect in Celf1 mutant mouse lenses was reflective of an underlying defect in the levels or expression pattern of p27Kip1 protein, I examined the status of p27Kip1 protein in these mutants. Using immunostaining, Western blotting and RT-qPCR, I found that p27Kip1 protein – but not transcript – levels were abnormally high in Celf1 mutant mouse lenses. Furthermore, in addition to the early differentiating fiber cells, the p27Kip1 protein was also abundant in late differentiating (maturing) fiber cells. These findings suggest a potential defect in the post-transcriptional control of p27Kip1 in Celf1 knockout lenses. ☐ To gain insight into the mechanism of Celf1-based control of the precise spatio-temporal expression of p27Kip1 protein in lens cells, I performed luciferase reporter assays in Celf1-knockdown lens cell culture. These experiments reveal that Celf1 functions to inhibit translation of the p27Kip1 mRNA by interacting with its 5’ UTR sequence. Together, these comprehensive set of experiments aimed to understand the phenotypic, cellular, molecular and functional characterization have led to the identification of a new mechanism of an RBP-mediated control of the cellular proteome in lens development, which is likely conserved across 500 million years of vertebrate evolution. This study presents us with a new regulatory model wherein Celf1-mediated post-transcriptional regulation in the lens positively regulates Dnase2b transcripts and negatively regulates p27Kip1 translation to orchestrate the fiber cell nuclear degradation program and achieve lens transparency. ☐ While the above findings satisfactorily address an important aspect of Celf1 function in lens development, I sought to comprehensively elucidate all the molecular alterations that Celf1 deficiency may cause in the mouse lens tissue. My motivation to do so stemmed from the understanding that in addition to mediating translational control of its target transcripts, Celf1 is also known to be involved in regulating their splicing and RNA decay in other tissue and cell types. Therefore, to gain further insights into the molecular changes and also to identify the differential expression of transcripts at the isoform level in Celf1 knockout lenses, I performed high-throughput RNA sequencing (RNA-seq) on newborn mutant lens tissue. The RNA-seq analysis led to the identification of ~900 differentially regulated genes (p-value 0.05, 1.5-fold cut off, 10 FPKM minimum expression in lens tissue) and 16 genes with differential isoform abundance in Celf1 conditional lens-specific knockout mouse mutants. Further analysis using the iSyTE lens expression database indicated that majority of the down-regulated genes in Celf1 knockout lenses exhibit highly enriched expression in normal mouse lens development, while up-regulated genes in Celf1 mutant lenses had no such pattern, suggesting that Celf1 activity is necessary for gaining molecular characteristics of fiber cells. Functional annotation clustering of the down-regulated genes identified that Celf1 deficiency resulted in the mis-expression of several cytoskeletal candidate genes such as Actn2 (encodes α-actinin, which has a function in regulating F-actin), Sptb (functions in the cellular cytoskeleton), Sorbs1, and Arhgaps. To gain further insights into Celf1-mediated regulation of cytoskeletal genes and the impact of Celf1 deficiency on the lens fiber cell cytoskeleton, I investigated F-actin disposition along the fiber cells using lens whole mount and section assays. I find a significant reduction and alteration of F-actin deposited along the fiber cells in Celf1 mutant lenses, which suggests that Celf1 also functions to mediate control over the fiber cell cytoskeleton potentially by regulating Actn2. ☐ Thus, the above findings demonstrated that RNA-seq analysis have the potential to provide new insights into the nature of Celf1 function in the lens. To gain further insights using this resource, I proceeded to further investigate the RNA-seq data obtained on Celf1 mutant lenses by performing an in silico approach to identify GU-rich Celf1-binding motifs within the 3’UTRs of the up-regulated genes. I hypothesized that this approach will serve to identify the potential Celf1 decay targets in the lens. This analysis identified an enrichment of the Celf1 preferred binding motif in 24 of the 50 up-regulated genes thus providing initial evidence for Celf1’s function in mediating mRNA decay in the lens. Among the genes of interest is Ell2, an RNA polymerase II elongation factor that is identified as a Celf1 direct target in other systems through meta-analysis and is mis-regulated in the Celf1 mutant lens, suggesting that Cefl1 may control Ell2 mRNA decay in the lens. ☐ Next, to investigate if Celf1 impacts splicing in the lens, I examined the RNA-seq data and identified 16 genes that have alterations in their isoform levels in Celf1 mutant lenses. This is an important finding given that the functional significance of genes is thoroughly comprehended only by understanding the expression of their specific isoforms in a tissue of interest. I identified differential expression of Sptb (β-spectrin ) isoforms between mutant and control lenses, which may be relevant to a subset of defects observed in the Celf1 knockout lens. This finding generates a specific hypothesis that Celf1 regulated stoichiometric expression of Sptb isoforms may have implications on its specific interactions with ligands and that in turn may impact its downstream function in maintaining mechanical stability in the fiber cell cytoskeletal membrane. ☐ In addition to making these original contributions to the understanding of RBP-mediated post-transcriptional control of gene expression in fiber cells during lens development, I also contributed to the understanding the function of the small Maf family proteins Mafg and Mafk in the lens and their impact on transcriptional control of fiber cell gene expression. Using iSyTE, a previous graduate student in the Lachke laboratory Ms. Smriti Agrawal had identified that Mafg and Mafk are expressed in the lens and have a critical function in maintaining its transparency. She had demonstrated that Mafg−/−:Mafk+/− compound mouse mutants progressively exhibit severe lens defects with age. Further, genome-level expression profiling coupled with the bioinformatics-based integrative analysis had identified among the differentially expressed genes in these mutants, rather surprisingly, largely non-crystallin genes that function in pathways such as oxidative stress and sterol synthesis. However, the biochemical impact of these changes on the Mafg−/−:Mafk+/− compound mouse mutant lenses was not investigated, and the finding that known markers of the lens are unperturbed had also not been validated. To address these knowledge gaps, I performed immunostaining on adult Mafg−/−:Mafk+/− compound mouse lens and control tissues and demonstrated that indeed known epithelial cell markers such as E-cadherin and Foxe3, as well the fiber marker γ-crystallin were unaltered in mutant lenses, validating the findings from the microarray analysis that genes commonly linked to cataractogenesis were not impacted by Mafg and Mafk deficiency. Further, I performed glutathione assays and demonstrated that Mafg−/−:Mafk+/− compound mouse lenses have elevated oxidative stress. Together, the new findings described in my dissertation have confirmed the function of these newly identified transcription factors Mafg and Mafk in controlling largely non-crystallin genes in the lens, and have presented the biochemical evidence for oxidative stress as a potential pathobiological mechanism linked to the cataracts observed in the Mafg−/−:Mafk+/− compound mouse mutant lenses. ☐ In sum, the research described in this dissertation has focused on unraveling new mechanisms of molecular control on gene expression in the lens. A major part of this study involves the identification of a new RBP Celf1 in the lens, and the molecular characterization of its multiple, distinct functions in the development of this tissue. Significantly this work shows that Celf1 is required in diverse vertebrates such as fish, frog and mouse for proper lens development, the perturbation of which results in the lens defect called as cataract. Furthermore, this work has shed light on the surprisingly non-crystallin based pathological mechanism of cataract in mice deficient for the transcription factor Mafg and Mafk. Elucidating the function of these molecules in the context of the lens has provided significant new insights into the growing knowledge of post-transcriptional gene regulation in ocular development, while also making significant new inroads on transcriptional regulation in this process. Finally, the molecular knowledge gained on the function of these molecules in the lens will likely have wider application in understanding the formation and pathology of other tissues and organs, where Celf1 and Mafg and Mafk functions are implicated.
Description
Keywords
Biological sciences, Cataract, Celf1, Lens development, Post-transcriptional regulation, RNA binding proteins