During their lifecycle, cells must produce forces which play important roles in several subcellular processes. Force-producing components are organized into macromolecular assemblies of proteins that are often dynamic, and are constructed or disassembled in response to various signals. The forces themselves may directly be involved in subcellular mechanics, or they may influence mechanosensing proteins either within or outside these structures. These proteins play different roles: they may ensure the stability of the force-producing structure, or they may send signals to a coupled process. The generation and sensing of subcellular forces is an active research topic, and this thesis focusses on the roles of these forces in two key areas: cell division and neurotransmitter release. The first part of the thesis deals with the effect of force on cell wall growth regulation during division in the fission yeast Schizosaccharomyces pombe, a cigar-shaped, unicellular organism. During cytokinesis, the last stage of cell division in which the cell physically divides into two, a tense cytokinetic ring anchored to the cellular membrane assembles and constricts, accompanied by the inward centripetal growth of new cell wall, called septum, in the wake of the inward-moving membrane. The contour of the septum hole maintains its circularity as it reduces in size—an indication of regulated growth. To characterize the cell wall growth process, we performed image analysis on contours of the leading edge of the septum obtained via fluorescence microscopy in the labs of our collaborators. We quantified the deviations from circularity using the edge roughness. The roughness was spatially correlated, suggestive of regulated growth. We hypothesized that the cell wall growers are mechanosensitive and respond to the force exerted by the ring. A mathematical model based on this hypothesis then showed that this leads to corrections of roughness in a curvature-dependent fashion. Thus, one of the roles of ring tension is to communicate with the mechanosensitive septum growth processes and coordinate growth to ensure the daughter cells have a functional cell wall. The second part of the thesis deals with how ring tension is produced and sustained, using experimentally measured ultrastructure of the cytokinetic ring itself. Recent super-resolution experiments have revealed that several key proteins of the fission yeast constricting ring are organized into membrane-anchored complexes called nodes. The force producing protein myosin-II in these nodes exerts pulling forces on polymeric actin filaments that are synthesized from polymerizers residing in the nodes. How these forces are marshalled to generate ring tension, and how such an organization maintains its stability is unclear. Using a mathematical model with coarse-grained representations of actin and myosin, we showed that such a node-based organization reproduces previously measured ring tension values. The model explains the origin of experimentally observed bidirectional motion of the nodes in the ring, and showed that turnover of the nodes rescues the ring from inherent contractile instabilities that would be expected when a force-producing structure is made up of small object that effectively attract one another. Finally, the third part of the thesis deals with the role of forces produced by SNARE proteins at synapses between two neurons during neurotransmission. A key step here is synaptic release, where inside a neuron, membrane-bound compartments called vesicles filled with neurotransmitter fuse with the membrane of the neuron forming a transient fusion pore, and release their contents to the outside of the cell. These neurotransmitter molecules are sensed by another neuron that is physically separate from the neuron in question and this neuron propagates the signal henceforth. Thus, regulation of neurotransmitter release is a key step in neurotransmission. A fusion machinery consisting of several proteins facilitates membrane fusion, and pore nucleation requires the formation of a SNARE protein complex in this machinery, whose role during pore dilation is unclear. Using electrophysiological measurements, our collaborators experimentally measured the statistics of the size of single fusion pores in vitro, and observed that average pore sizes increased with the number of SNARE proteins. Using mathematical modeling, we showed that this effect was due to an entropic crowding force that expands the pore and increases with the number of SNAREs, and counteracts the energy barrier to fusion pore expansion.

Intra-axonal protein synthesis is important for the proper wiring of the nervous system and can have restorative or pathogenic effects in response to nerve injury and neurodegenerative stimuli. The set of axonally translated transcripts, the axonal translatome, is regulated through the control of mRNA localization, stability, and translation. Targeting the axonal translatome could result in the development of novel therapies for the treatment of neurological disorders. Yet, there are gaps in our understanding of the selective mechanism regulating the specific localization of mRNAs into axons. Currently, axonal localization of transcripts is understood to be controlled by the presence of sequence elements that direct axonal transport. In an attempt to identify novel localization motifs, I found that a well-known motif corresponding to the Pumilio Binding Element (PBE) is significantly depleted in axonally enriched mRNAs. Moreover, I found this element to be highly informative of axonal mRNA localization and translation across different neuronal types and developmental stages suggesting that it is a highly conserved regulatory motif. I found Pum2 neuronal expression and subcellular localization to be highly consistent with the way the PBE predicts mRNA regulation. I then demonstrated that interfering with Pum2 function results in increased axonal localization of PBE containing mRNAs. Finally, Pum2 downregulation was associated with gross defects in axonal outgrowth, branching, and regeneration. Altogether, this data suggests that Pum2 regulates axonal mRNA localization through an exclusion mechanism that is important during neuronal development.

Self-renewal mediates homeostasis across mammalian organ systems as the cellular components of mature tissues are continually replaced in the face of wear and tear, injury, infection, and malignancy. The hematopoietic and immune systems are crucial for organismal longevity and rely on the ability of progenitor cells to bifurcate in fate to produce mature terminally differentiated progeny while self-renewing to maintain more quiescent progenitors. Asymmetric cell division is associated with self-renewal of lymphocytes and hematopoietic progenitors, but the mechanisms underlying the cell biology of these processes remain incompletely understood. Here we show that metabolic signals in the form of differential anabolism and catabolism regulate asymmetric division and cell fate bifurcations. Key transcription factors, including TCF1 and IRF4 in lymphocytes and IRF8 in hematopoietic progenitors, occupy regulatory nodes where signals associated with metabolism and traditional cell fate determinants converge. Notably, anabolic PI3K/mTOR signaling was required for terminal differentiation of both lymphocytes and hematopoietic progenitors through the regulation of a constellation of nutrient uptake, mitochondrial turnover, reactive oxygen species production, and autophagy. Further, we found that antigen receptor signaling in lymphocytes organizes a cell-intrinsic polarity pathway of asymmetric intracellular membrane trafficking that is regulated by PI3K activity and associated with terminal differentiation. These results support a model wherein cell fate bifurcations are organized by metabolic signaling at the population and subcellular level to ensure self- renewal of progenitor and memory populations.

Human and murine diabetes is characterized by pancreatic β-cell dedifferentiation, a process in which β-cells lose expression of markers of maturity and gain those of endocrine progenitors. Failing β-cells inappropriately metabolize lipids over carbohydrates and exhibit impaired mitochondrial oxidative phosphorylation. Therefore, pathways involved in mitochondrial fuel selection and catabolism may represent potential targets for the prevention or reversal of dedifferentiation. In chapter I of this dissertation, we isolated and functionally characterized failing β-cells from various experimental models of diabetes. We found a striking enrichment in the expression of aldehyde dehydrogenase 1 isoform A3 (Aldh1a3) as β-cells become dedifferentiated. Flow-sorted Aldh1a3-expressing (ALDH+) islet cells demonstrate impaired glucose-induced insulin secretion, are depleted of Foxo1 and MafA, and include a Neurogenin3-positive subset. RNA sequencing analysis demonstrated that ALDH+ cells are characterized by: (i) impaired oxidative phosphorylation and mitochondrial complex I, IV, and V; (ii) activated RICTOR; and (iii) progenitor cell markers. We propose that impaired mitochondrial function marks the progression from metabolic inflexibility to dedifferentiation in the natural history of β-cell failure. In chapter II of this dissertation, we report that cytochrome b5 reductase 3 (Cyb5r3) is a FoxO1-regulated mitochondrial oxidoreductase critical to β cell function. Expression of Cyb5r3 is greatly decreased in multiple murine models of diabetes, and in vitro Cyb5r3 knockdown leads to increased ROS generation and impairment of respiration, mitochondrial function, glucose-stimulated insulin secretion, and calcium mobilization. In vivo, mice with β-cell-specific ablation of Cyb5r3 (B-Cyb5r3) display impaired glucose tolerance with decreased insulin secretion, and their islets have significantly lower basal respiration and glucose-stimulated insulin secretion. B-Cyb5r3 β-cells lose expression of Glut2, MafA, and Pdx1 expression despite a compensatory increase in FoxO1 expression. Our data suggest that Cyb5r3 is a critical mediator of FoxO1’s protective response in β-cells, and that loss of Cyb5r3 expression is an early event in β-cell failure.

Myelination of axons in the CNS by oligodendrocytes (OLs) is critical for the rapid and reliable conduction of action potentials down neuronal axons, as evidenced by the severe disabilities associated with myelin loss in multiple sclerosis and other diseases of myelin. The specification, differentiation, and maturation of OLs along with myelin formation by OLs have been thoroughly characterized. How myelin is turned over, however remains unclear. It is unsurprising that little is known about myelin turnover considering that for decades following their discovery, myelin and OLs were considered static elements in the adult nervous system. Recent evidence, however, shows that myelin in the CNS is actually plastic. Moreover, myelin remodeling in humans has been suggested to be mediated by mature OLs. As mature OLs have limited capacity to generate new myelin sheaths, we must ask whether mature OLs can remodel the myelin at preexisting myelin sheaths. One intriguing but unproven possibility is that myelin at individual internodes may be remodeled cell-autonomously by mature OLs to modulate neuronal circuit function. Macroautophagy (MA) is responsible for the lysosome-mediated elimination of cytosolic proteins, lipids, and organelles. MA achieves this by capturing cargo in bulk or selectively in a transient, multilamellar structure known as an autophagosome (AP). In this study, we used a combination of in vivo and cellular approaches to test the hypothesis that MA in OLs may be important for myelin remodeling in the adult CNS. We establish that myelin of individual internodes is remodeled, and does so through the coordinated efforts of endocytosis and MA. We found that autophagy protein Atg7 is essential for myelin remodeling in vivo: loss of Atg7 in OLs leads to an age-dependent increase of myelin at the internode and the formation of aberrant myelin structures, most notably myelin outfoldings. In addition, we find that MA has the potential to occur throughout the mature OL, and examination of OLs in culture suggests that formation of a mature AP structure, the amphisome, is required to facilitate the efficient degradation of myelin-containing endocytic structures. Together, we propose that myelin is a dynamic structure that is regularly remodeled through the cooperative efforts of MA and endocytosis. These findings raise the possibility that myelin remodeling is involved in neural plasticity and the tuning of neural circuits.

The decisions stem cells make impact both the development of adult vertebrates and systems within the body that require cellular replenishment to sustain life. Regardless whether a stem cell remains quiescent, divides, differentiates, or undergoes apoptosis—these processes are precisely controlled by internal gene regulatory networks that are instructed by external stimuli. The exact mechanisms governing stem cell fate are not completely understood. These studies explore new ways in which cell fate is mediated. Through a study of mitochondrial content in human embryonic stem cells (hESCs) and their differentiated progeny, we discovered differences in mitochondrial morphologies. Mitochondria began as elongated and networked structures in self-renewing conditions and changed their shape after differentiation. The addition of external growth factors that direct hESCs toward the definitive endoderm (DE) lineage promoted mitochondrial fragmentation, which was mediated by the mitochondrial fission machinery. Globular, punctate mitochondria were observed prior to the induction of the DE-specific transcriptional program. Differentiation of hESCs to other lineages did not result in any mitochondrial shape changes. Thus, mitochondrial fission in differentiating hESCs, an internal cellular process, is induced by DE-inducing external stimuli, an effect that was lineage specific. In a second study, we investigated the role of the splenic environment in the development of the blood system—during hematopoiesis. The spleen made a distinct contribution to hematopoiesis, a process predominantly attributed to the bone marrow. We discovered a previously unidentified population of cells, uniquely represented in the mouse spleen that could develop into erythrocytes, monocytes, granulocytes, and platelets. These multipotent progenitors of the spleen (MPPS) expressed higher levels of the transcription factor, NR4A1 compared to their bone marrow counterparts and relied on NR4A1 expression to direct their cell fate. The activation of NR4A1 in MPPS biased their production of monocytes and granulocytes in vitro whereas NR4A1-deficient MPPS over-produced erythroid lineage cells in vivo. Together, these data suggest the splenic niche supports distinct myeloid differentiation programs of multi-lineage progenitors cells. Both studies identify new mechanisms by which external stimuli regulate internal mechanisms of cell fate. These insights provide a better understanding of stem and progenitor cell differentiation that have the potential to impact cellular replacement therapies.

The urothelium is a stratified epithelium that serves as a barrier between the urinary tract and blood. It consists of terminally differentiated umbrella cells, which are specialized for synthesizing and assembling uroplakins into a tough apical plaque and responsible for the barrier function; intermediate cells which are few in number but serve as umbrella cell progenitors; and unipotent basal cells, which populate the majority of the urothelium. The urothelium is one of the most quiescent epithelia in the body but can rapidly regenerate in response to damage. The urothelium is also a source of cells that give rise to bladder cancer. Patients with chronic inflammation caused by indwelling catheters or repeated urinary tract infections have a higher risk of developing bladder cancer. Bladder cancers with squamous histological features are considered to be more aggressive with poor prognosis and the majority are categorized as basal subtype. The expression of Peroxisome Proliferator-Activated Receptor-gamma (PPARG) is strongly down regulated in the basal subtype of bladder cancer, suggesting that its removal might be essential in tumorigenesis. PPARG is a nuclear hormone receptor that was originally described as a master regulator of adipogenesis but could also promote cellular differentiation in a number of epithelium. PPARG also serves as an important regulator in anti-inflammatory activity after a variety of injuries, acting in part by antagonizing the NF-B pathway. In urothelial cells, it has been shown that PPARG promotes urothelial differentiation in vitro, but its function in vivo remains unexplored. To determine the role of PPARG in vivo, we used Cre-Lox recombination to conditionally delete the Pparg gene in the mouse urothelium using the ShhCre driver, which drives recombination in basal and intermediate cells, and their respective daughters. Interestingly, ShhCre;Ppargfl/fl mutants lack umbrella and intermediate cells which normally populate the luminal and sub-luminal layers, and are instead populated with an abnormal cell population negative for classical urothelial markers. The basal compartment, which in wild type mice is largely populated by P63+ KRT5+ basal cells with a small sub-population expressing KRT14; has an increased number of KRT14-expressing cells in the mutants and exhibits squamous features that are not present in the normal urothelium. In wild type animals, urinary tract infection (UTI) with uropathogenic E.coli results in a transient innate immune response, followed by proliferation and repair, which is largely complete within 2 weeks. When ShhCre;Ppargfl/fl mutants were challenged with urinary tract infection, the innate immune response was not resolved even after several weeks, as characterized by persistent NF-B activity, excessive influx of neutrophils and macrophages, and massive granulation tissue in the stroma. In addition, the Pparg-knockout urothelium exhibited squamous metaplasia. The Krt14+ basal cell population, which is considered to be the cells of origin of bladder cancer, greatly expanded in the Pparg-deleted urothelium after infection, and some lesions progressed to acquire invasive features. Together these findings suggest that PPARG is essential for the normal differentiation of the urothelium and is a potent regulator of the inflammatory response after UTI. Understanding the link between the loss of PPARG, chronic inflammation and tumorigenesis in the urothelium could shed light on the urothelial differentiation network and pave the way for the development of therapeutic approaches to various urinary diseases.

Phospholipids are the major components of cell membranes and have a wide variety of structures, shapes and properties. Different ratios of phospholipid species confer different properties to membranes and contribute to the normal function of organelles. We have previously shown that acute phosphatidylcholine (PC) biosynthesis inhibition leads to a severe form of lipid imbalance that disrupts ER morphology and structure. Furthermore, our previous studies also revealed a mechanism for ER proteostasis under conditions of lipid-imbalance-induced ER stress in yeast, whereby unfolded ER proteins are removed by lipid droplets (LDs) and targeted to the vacuole for degradation by microlipophagy. Here, we find that LDs also contribute to ER proteostasis during chemically induced ER stress. Furthermore, we find that ER stress results in an increase in ubiquitinated proteins in LDs as well as recruitment of cytosolic and ER heat shock proteins, as well as ER proteins to LDs. ER stress-induced microlipophagy does not require core ATG genes and can occur in the absence of lipid ordered microdomains (Lo) in the vacuolar membrane. Instead, we find that the ESCRT machinery is up-regulated and localizes to the vacuolar membrane in response to ER stress induced microlipophagy and that ESCRT I, II and III complexes are required for microlipophagy in response to each of these stressors. Similar to yeast, we find that lipid imbalance in skeletal muscle from CHKB CMD, new autosomal recessive CMD (Congenital Muscular Dystrophy) caused by a mutation of choline kinase beta (CHKB), results in abnormal SR/ER morphology. CHKB is the first enzyme in the de novo PC biosynthesis pathway and causes phospholipid imbalance in cell membranes similar to that observed in yeast. Besides the disruption of SR morphology, we also detect a dysfunction of the Ryanodine Receptor (RyR), the Ca2+ channels responsible for initiating muscle contraction. Specifically, we observe abnormal RyR morphology and increased association of RyRs with lipid droplets (LD) in muscle fibers from a CHKB CMD patient. Finally, we detect ER stress and pronounced UPR activation in rmd mice, a mouse model of CHKB CMD. Given these results, we propose that inhibition of PC biosynthesis leads to phospholipid imbalance in the SR, which in turn, causes RyR leakage and ER stress which lead to mitochondrial dysfunction and dystrophy in CHKB CMD.

Multicellular organisms arise from a single fertilized zygote, which must contain all information necessary to develop. As the embryo divides, the cells adopt distinct functional characteristics, and as they do so, they become committed to these fates, unable in most cases to convert from one identity to another. Though it has been well known and described for over a century now this year, this latter process, in this work referred to as cell fate restriction, is not well understood. In this thesis, I aim to contribute to the understanding of this developmental phenomenon. The tool I use is the ectopic expression of a terminal fate specifying transcription factor, CHE-1. This transcription factor normally functions to specify the fate of a pair of gustatory neurons in the nematode Caenorhabditis elegans. If ectopically expressed early in development, it is able to induce expression of its target genes, but by adulthood, most cells are refractory to its transcriptional activation, evidence of developmental cell fate restriction in most tissues of the animal. I first describe the work of Tulsi Patel to which I contributed, in which an RNAi screen revealed that PRC2 complex is responsible for preventing CHE-1 activity in the germline cells of C. elegans. I then describe a semi-clonal genetic screen in which I found many more mutants with a similar phenotype affecting germline cells, and cloned another gene that is able to induce expression specifically in the epidermis of the animals: usp-48, a highly conserved ubiquitous nuclear deubiquitinating enzyme. Next, I describe another screen where I ectopically express CHE-1 specifically in the adult epidermis, in which I found and cloned an additional six mutants: ogt-1, dot-1.1, pmk-1, sek-1, nhr-48, and C08A9.6, here named epco-1. In this screen I also isolated but was unable to clone an additional four mutants that likely represent an additional four genes. I discuss the nature of these genes and their potential roles in restricting cell fate. Lastly, I describe the optimization of a tissue-specific transcriptional profiling protocol, INTACT, for use in the characterization of the mutants. With this optimized protocol, I was able to perform detailed RNAseq on two individual neuron types from the animal, as well as wild-type epidermis. This optimized protocol will be used to characterize the mutants in the future. Together, these results tie unexpected genes to the function of cell fate restriction in the C. elegans epidermis, which will aid in our understanding of this fundamental developmental phenomenon.

Fung Ka Leung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (leaves 83-93). / Abstracts in English and Chinese. / Thesis Committee --- p.ii / Statement --- p.iii / Acknowledgements --- p.iv / Abstract --- p.v / 摘要 --- p.vii / Table of Contents --- p.viii / List of Tables --- p.xii / List of Figures --- p.xiii / List of Abbreviations --- p.xv / Chapter Chapter 1 --- General Introduction --- p.1 / Chapter 1.1 --- An overview of the secretory pathway in eukaryotic cells --- p.2 / Chapter 1.2 --- The secretory pathway in plants --- p.4 / Chapter 1.2.1 --- Plant cells contain two functionally distinct vacuoles --- p.4 / Chapter 1.2.2 --- Three vesicular pathways to two vacuole --- p.6 / Chapter 1.2.3 --- Transport vesicles in the three vesicular pathways --- p.9 / Chapter 1.2.4 --- Vacuolar sorting determinants (VSDs) --- p.10 / Chapter 1.2.5 --- Vacuolar sorting receptors (VSRs) --- p.12 / Chapter 1.3 --- The PSVs in mature seeds --- p.15 / Chapter 1.3.1 --- Biogenesis of PSV --- p.15 / Chapter 1.3.2 --- The two chimeric integral membrane reporters --- p.16 / Chapter 1.3.3 --- Subcellular localization of the two chimeric integral membrane reporters in PSVs of mature tobacco seeds --- p.17 / Chapter 1.4 --- Project objectives --- p.19 / Chapter Chapter 2 --- Materials and Methods --- p.20 / Chapter 2.1 --- Construction of the YFP-BP-80 and the YFP- a -TIP reporters --- p.21 / Chapter 2.1.1 --- The pYFP-BP-80-K construct --- p.21 / Chapter 2.1.2 --- The pYFP- a -TIP-K construct --- p.22 / Chapter 2.2 --- Construction of GFP-RMR reporter --- p.23 / Chapter 2.2.1 --- Cloning of pGFP-RMR --- p.23 / Chapter 2.2.2 --- Cloning of pGFP-RMR-K --- p.23 / Chapter 2.3 --- Construction of pGONST1-YFP construct --- p.26 / Chapter 2.3.1 --- The pGONSTl-YFP construct --- p.26 / Chapter 2.4 --- Transformation of Agrobacterium by electroporation --- p.27 / Chapter 2.5 --- Tobacco transformation and selection --- p.28 / Chapter 2.5.1 --- Plant materials --- p.28 / Chapter 2.5.2 --- Tobacco transformation --- p.28 / Chapter 2.6 --- Screening of transgenic tobacco plants expressing YFP fusion proteins --- p.30 / Chapter 2.6.1 --- Kanamycin screening --- p.30 / Chapter 2.6.2 --- Extraction of genomic DNA from leaves --- p.30 / Chapter 2.6.3 --- PCR of genomic DNA --- p.31 / Chapter 2.7 --- Southern blot analysis of genomic DNA --- p.32 / Chapter 2.8 --- Western blot analysis of transgenic tobacco plants --- p.33 / Chapter 2.8.1 --- Extraction of total protein from tobacco leaves or seeds --- p.33 / Chapter 2.8.2 --- Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot analysis --- p.34 / Chapter 2.9 --- Confocal immunofluorescence studies --- p.35 / Chapter 2.9.1 --- Preparation of sections --- p.35 / Chapter 2.9.2 --- Single labeling --- p.35 / Chapter 2.9.3 --- Double labeling with one polyclonal and one monoclonal antibodies --- p.36 / Chapter 2.9.4 --- Double labeling with two polyclonal antibodies --- p.36 / Chapter 2.9.5 --- Collection of images --- p.37 / Chapter 2.10 --- Chemicals --- p.38 / Chapter 2.11 --- Primers --- p.38 / Chapter 2.12 --- Bacterial strain --- p.38 / Chapter 2.13 --- Antibodies --- p.39 / Chapter 2.14 --- Growing condition of transgenic plants and determining the developmental stage of tobacco flowers --- p.39 / Chapter Chapter 3 --- Results --- p.41 / Chapter 3.1 --- Generation of transgenic tobacco plants --- p.42 / Chapter 3.2 --- PCR screening of transgenic tobacco plants --- p.46 / Chapter 3.3 --- Southern blot analysis --- p.48 / Chapter 3.4 --- Detection of the YFP fusion proteins in transgenic tobacco plants by western blot analysis --- p.50 / Chapter 3.4.1 --- Detection of the YFP fusion proteins in leaves --- p.50 / Chapter 3.4.2 --- Western blot analysis of vegetative tissues --- p.57 / Chapter 3.4.3 --- Western blot analysis of mature seeds --- p.59 / Chapter 3.5 --- Confocal immunofluorescence studies --- p.61 / Chapter 3.5.1 --- Detection of YFP signals in root tip cells --- p.61 / Chapter 3.5.2 --- Detection of YFP signals in developing seeds --- p.65 / Chapter 3.5.3 --- Subcellular localization of the YFP fusion proteins in mature seeds --- p.67 / Chapter Chapter 4 --- Discussion --- p.72 / Chapter Chapter 5 --- Summary and Future Perspectives --- p.77 / Chapter 5.1 --- Summary --- p.78 / Chapter 5.1.1 --- Generation of transgenic tobacco plants expressing the YFP fusion proteins --- p.78 / Chapter 5.1.2 --- Full-length fusion proteins and cleaved soluble YFP were detected in vegetative tissues --- p.79 / Chapter 5.1.3 --- Only cleaved soluble YFP was detected in mature seeds --- p.79 / Chapter 5.1.4 --- The two fusion proteins might localized in different compartments in developing seeds --- p.79 / Chapter 5.1.5 --- Both fusion proteins were localized within the PSVs of mature seeds --- p.80 / Chapter 5.2 --- Future perspectives --- p.81 / References --- p.83