ABSTRACT Title of Dissertation: FUNCTIONAL STUDY OF SAPOSIN-LIKE PROTEINS IN ARABIDOPSIS THALIANA Changxu Pang, Doctor of Philosophy, 2020 Dissertation Directed By: Dr. Angus Murphy, Professor, Department of Plant Science & Landscape Architecture Dr. Wendy Ann Peer, Associate Professor, Department of Environmental Science & Technology Sphingolipids and microdomain-associated proteins that are associated with the plasma membrane and endomembrane system are important in plant growth and development. Elucidating functions of these proteins advance understanding of signal transduction from plasma membrane into cytosol and between different intracellular membrane compartments. Saposins and saposin-like proteins (SAPLIP) are among these proteins. In plants, two types of proteins contain saposin B-like domains (SapB- like domains): saposin-domain containing aspartic proteases (ASPAs) and prosaposin- like proteins (PSAPLIPs). Phenotypic analyses showed that single loss-of-function aspa2 showed delayed seed maturation. Seeds of aspa1-2 aspa2-1 aspa3-3 triple mutant (aspa1 is knock- down, aspa2 and aspa3 are knock-out alleles) showed delayed germination rates and delayed seed storage proteins degradation. Further, protein storage vacuolar fusion was also delayed in the mutant cotyledons. These results suggest that ASPAs process seed storage proteins during seed germination in vivo, and probably also involved in protein storage vacuolar fusion regulation. ASPAs also have a role in root architecture. Triple mutant showed longer primary root length under low nitrogen conditions. Further analysis suggested that the altered root architecture in the mutants may result from rates of tracheary element (TE) maturation in xylem tissues. Triple mutants were slightly delayed in TE maturation and the ASPA2 overexpression showed slightly early maturation. Together with the expression pattern of ASPA3, this indicates that ASPAs may take part in programmed cell death (PCD) in Arabidopsis. Further studies showed that ASPAs are involved in PCD execution. Results showed that the onset of PCD was not delayed in the triple mutant, but the execution time of PCD was extended. Membrane permeability increased more slowly in the triple mutants and faster in the overexpression plants. This reflects the role of ASPAs in membrane disturbance and permeability regulation during PCD. The prosaposin-like proteins (PSAPLIPs) have received little study. Sequence alignments identified that prosaposin-like proteins are ubiquitous in plant kingdom. Plant PSAPLIPs show highly conserved in secondary structure of SapB-like domains. This structural similarity was supported by glycosylation analyses of Arabidopsis thaliana AtPSAPLIP1 and AtPSAPLIP2. Both AtPSAPLIP1 and AtPSAPLIP2 traffic to vacuoles. Possible role of PSAPLIPs is facilitating target protein degradation. AtPSAPLIP1 was mainly expressed in inflorescence, especially in sepals, carpels and mature pollen grains, as well as leaves and roots. Young leaves had higher expression level than aged leaves. AtPSAPLIP2 was expressed in inflorescence too, but mainly in young anthers, petals, ovules and developing seeds. This result indicates function differentiation of PSAPLIPs in Arabidopsis. Both genes are important in male gametophyte development. The significance of this dissertation is that it demonstrates that ASPAs process seed storage proteins during seed germination in vivo for the first time. It also discovered a new role of ASPAs in regulating programmed cell death by promoting memberane permeability, and thus affecting root growth in Arabidopsis. The third is that this is also the first time to characterize the plant prosaposin-like proteins, which are important in male gametophyte development and provide novel sights on how plants regulate reproductive process. These results will broaden our understanding of the protein-lipid interaction in the cell and the biological functions of saposin-like proteins in plant growth and development. FUNCTIONAL STUDY OF SAPOSIN-LIKE PROTEIN IN ARABIDOPSIS THALIANA by Changxu Pang Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2020 Advisory Committee: Professor Angus Murphy, Chair Prof. Wendy Peer, Co-Chair Prof. Caren Chang Prof. Gary Coleman Prof. Zhongchi Liu ? Copyright by Changxu Pang 2020 Preface This dissertation is original, unpublished, independent work by the author, Changxu Pang. I am responsible for all major areas of data collection and analysis, and majority of figures. Ally Albers and Amani Perwaz Aulakh helped in genotyping the plants. This dissertation is composed of four chapters and five appendices. Each chapter is organized in a manuscript format (abstract, introduction, results, discussion, materials and methods). Materials and methods listed in the end of each chapter are used in this chapter only, and details in materials and methods are listed in appendix D and some are repeated. Supplementary figures and tables are included in the appendices. Supplementary figures for chapter 2 are in appendix A. Supplementary figures for chapter 3 are in appendix B and appendix C. Primer list and plant prosaposin-like protein gene list are in appendix E. Additional figures in other projects during PhD are listed in appendix F. I would like to give acknowledgement to Dr. Liwen Jiang who provided the GFP-FREE1 marker line in this dissertation. ii Acknowledgements First, I would like to show my gratitude to my advisor, Dr. Angus Murphy, for giving me the opportunity to work in this lab for all these years. He has been quite supportive both in this project and in my personal life. He is patient with all the questions that I asked and provides informative ideas for me. Every time that I encounter a hard situation, he encouraged me to overcome difficulties. I?m glad that I received academic trainings in this lab. Besides, I wish to pay my regards to my co-advisor Dr. Wendy Peer for academic advice (and cakes). Conversations with her have always been glad and helpful. Then I wish to express my sincere gratitude to the rest of my committee members, Dr. Caren Chang, Dr. Zhongchi Liu and Dr. Gary Coleman, for your perspectives on my project and great advice. I would also give my thanks to my colleagues Dr. Mark Jenness, Dr. Candace Prichard, Dr. Jun Zhang, Dr. Rueben Tayengwa, Dr. Doron Shkolnik, Dr. Wiebke Tapken, Sarah Turner, Ally Albers, Amani Perwaz Aulakh, Gabreille Bate and Juliane Henschel. I?m really appreciated for your advice whenever I needed and keep the lab operations running very well. Without your help, I could not go through all the work. I would like to recognize the invaluable assistance which you provided during my study. And last, I would like to express my greatest gratitude to my mother. Though we are apart for a long time, I still feel her deepest love and care, which is my greatest iii motivation for this venture. iv Table of Contents Preface ........................................................................................................................... ii Acknowledgements ....................................................................................................... iii Abbreviations .............................................................................................................. viii List of Figures ................................................................................................................. x Chapter 1: Literature review: Saposin-like proteins in plants ....................................... 1 Abstract ................................................................................................................................. 1 Introduction: Sphingolipids and microdomains in plants ..................................................... 1 Saposin-like Proteins (SAPLIPs) ............................................................................................. 5 Biological functions of SAPLIPs in animals............................................................................. 6 SAPLIPs in plants .................................................................................................................... 8 General structure features of SAPLIPs ................................................................................ 10 Structure and function of human saposins. ........................................................................ 11 Structure of and function of prosaposins ........................................................................ 11 Structure and function of mature saposins ..................................................................... 13 Structural features of saposin A ...................................................................................... 15 Structure feature of saposin C ......................................................................................... 15 Structure feature of saposin B ......................................................................................... 16 Structure feature of saposin D ........................................................................................ 17 Mechanistic model of saposin-lipid interactions ................................................................ 18 Structure of SAPLIPs in plants ............................................................................................. 21 Plant aspartic proteases ...................................................................................................... 23 General Information about plant aspartic protease ....................................................... 23 Aspartic proteases in Arabidopsis ................................................................................... 24 Biological functions of aspartic proteases in plants ........................................................ 27 Structure of plant aspartic protease ............................................................................... 33 Biological function of plant specific insert ...................................................................... 35 Perspective .......................................................................................................................... 38 Chapter 2 Functional study of saposin-like domain containing aspartic proteases in Arabidopsis thaliana .................................................................................................... 40 Abstract ............................................................................................................................... 40 Introduction ......................................................................................................................... 41 v Results ................................................................................................................................. 47 ASPAs function in seed development and germination .................................................. 47 Expression pattern of ASPA2 ........................................................................................... 58 Subcellular localization and trafficking of ASPA2 ............................................................ 62 ASPAs are involved in root architecture regulation ........................................................ 67 Transcriptional regulation of ASPA2 ................................................................................ 70 ASPAs are involved in programmed cell death ............................................................... 74 Discussion ............................................................................................................................ 80 Conclusion ........................................................................................................................... 87 Materials and Methods ....................................................................................................... 89 Plant materials ................................................................................................................. 89 Germination test ............................................................................................................. 90 Seed Protein extraction ................................................................................................... 91 SDS-PAGE ......................................................................................................................... 91 Coomassie blue staining .................................................................................................. 92 Glycosylation test ............................................................................................................ 92 Western blot .................................................................................................................... 92 Cloning and expression vector construction ................................................................... 93 Microscopy ...................................................................................................................... 94 Time-course image of PI (propidium iodide) and PI/FDA (fluorescein diacetate) double staining in lateral root cap ............................................................................................... 95 Chapter 3: Elucidating features and functions of plant prosaposin-like proteins (PSAPLIPs) .................................................................................................................... 96 Abstract ............................................................................................................................... 96 Introduction ......................................................................................................................... 97 Results ............................................................................................................................... 101 Phylogenic studies of PSAPLIPs in plants ....................................................................... 101 Structural features of AtPSAPLIPs ................................................................................. 106 Subcellular localization of Arabidopsis PSAPLIPs .......................................................... 124 Expression pattern of AtPSAPLIP1 and AtPSAPLIP2 ...................................................... 128 Discussion .......................................................................................................................... 132 vi Conclusion ......................................................................................................................... 137 Materials and Methods ..................................................................................................... 138 Primary and Secondary Structure Prediction ................................................................ 138 Sequence Alignment ...................................................................................................... 139 Phylogenetic tree construction ..................................................................................... 140 Preparation of Transgenic Plants .................................................................................. 140 Plant Materials and chemical treatment ....................................................................... 141 Protein extraction .......................................................................................................... 142 Glycosylation test .......................................................................................................... 142 SDS-PAGE ....................................................................................................................... 143 Western blot .................................................................................................................. 143 Microscopy .................................................................................................................... 144 Histochemistry ............................................................................................................... 145 Chapter 4: Conclusion and Perspective ..................................................................... 146 Appendix A Supplemental Figures for Chapter 2 ...................................................... 153 Appendix B Supplemental Figures for Chapter 3 ....................................................... 160 Appendix C Phylogenic tree of plant PSAPLIPs .......................................................... 216 Appendix D Materials and Methods .......................................................................... 224 Appendix E Supplemental Tables............................................................................... 242 Appendix F Additional Data ....................................................................................... 274 References ................................................................................................................. 292 vii Abbreviations Abbreviation and symbol Definition A alanine AA amino acid ABA abscisic acid Amp ampicillin ASPA aspartic protease aspa1-2/2-1/3-3 aspa1-2 aspa2-1 aspa3-3 triple mutant BFA brefeldin A bp base pairs Col-0 Columbia-0 CFP cyan fluorescent protein Conc A concanamycin A CRISPR clustered regularly interspaced short palindromic repeats CTAB cetyl trimethylammonium bromide D aspartic acid DAG days after germination DAP days after pollenation DMSO Dimethyl sulfoxide DPI diphenyleneiodonium DNA deoxyribonucleic acid EDTA ethylenediaminetetraacetic acid E glutamic acid E. coli Escherichia coli EE early endosome FDA fluorescein diacetate GA gibberellic acid Gent gentamicin GFP green fluorescent protein GUS ?-glucuronidase H2A histone 2A 10 H2O2 hydrogen peroxide HCl hydrochloric acid Hyg hygromycin K lysin Kan Kanamycin kb kilo base pairs kDa kilo Dalton KNO3 potassium nitrate L leucine LB Luria-Bertani viii LE late endosome MS Murashige and Skoog MVB multivesicular body N asparagine NaCl sodium chloride NADPH nicotinamide adenine dinucleotide phosphate NaOH sodium hydroxide NTPP N-terminal propeptide P proline PBS phosphate-buffered saline PCD programmed cell death PCR polymerase chain reaction PI propidium iodide PM plasma membrane PSAPLIP prosaposin-like protein PSI plant specific insert PSV protein storage vacuole PVC prevacuolar compartment PVDF Polyvinylidene fluoride Q glutamine R arginine Rif rifampicin RFP red fluorescent protein RNA ribonucleic acid rpm round per minute SapB saposin B SDS sodium dodecyl sulfate SDS PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SAPLIP saposin-like protein SP signal peptide Spec spectinomycin SSP seed storage protein TE tracheary element TEMED Tetramethyl ethylenediamine TGN trans-Golgi network TML transmitted light WT wild type X gluc 5-Bromo-4-chloro-3-indolyl-?-D-glucuronide YFP yellow fluorescent protein Zeo zeocin ix List of Figures Figure 1-01. Schematic depiction of the three main activities of SAPLIPs. Figure 1-02. Schematic comparison between animal aspartic proteases and plant aspartic proteases. Figure 1-03. Schematic comparison between saposin and swaposin. Figure 1-04. Schematic illustration of saposin domain of NK-lysin and the swaposin domain of prophytepsin from barley. Figure 1-05. Models of lipid activation by saposins. Figure 1-06. A model for the interaction with membranes and a general mechanism of action of the SAPLIP family. Figure 1-07. Plant PSI primary structure. Figure 1-08. Phylogenetic tree of A1 family proteases in Arabidopsis. Figure 1-09. Schematic illustration of the process steps for ASPA1. Figure 2-01. Characterization of ASPA T-DNA insertion mutants. Figure 2-02. ASPAs regulate seed maturation in Arabidopsis. Figure 2-03. Seed germination and seed storage protein degradation in aspa mutant. Figure 2-04. Germination rates in Col-0 and aspa triple mutant seeds with and without gibberellin acid 3, hydrogen peroxide and diphenylene iodonium treatments. Figure 2-05. Protein storage vacuole (PSV) fusion during germination in Col-0 and aspa triple mutant seeds. x Figure 2-06. ASPA2 expression during seed development. Figure 2-07. ASPA2 expression in vegetative and reproductive tissues. Figure 2-08. Intracellular trafficking pathway of ASPA2 in Arabidopsis roots. Figure 2-09. Glycosylation and vacuolar trafficking of ASPA2-CFP. Figure 2-10. Root architecture in Col-0 and ASPA mutants. Figure 2-11. Transcriptional regulation of ASPA1 and ASPA2 and root growth in responses to ABA and NaCl treatments. Figure 2-12. Propidium iodide (PI) staining in lateral root caps of Col-0 and ASPA mutants. Figure 2-13. Fluorescent diacetate (FDA) and propidium iodide (PI) double staining in lateral root cap cells in Col and ASPA mutants over time. Figure 3-01. Predicted structure of AtPSAPLIP1 and AtPSAPLIP2. Figure 3-02. Predicted structure of saposin B (SapB)-like domains in AtPSAPLIP1 and AtPSAPLIP2. Figure 3-03. Sequence alignment of plant PSAPLIPs from some angiosperms. Figure 3-04. Primary structure similarity between plant PSAPLIPs with human prosaposin. Figure 3-05. AtPSAPLIP1 promoter GUS staining. Figure 3-06. AtSAPLIP2 promoter GUS staining. Figure S01. Colocalization between autophagy marker ATG8a and ASPA2 Figure S02. Phenotype of 30 DAG plants of ASPAs overexpression. xi Figure S03. Phenotype of 40 DAG plants of ASPAs overexpression. Figure S04. 35S::ASPA2 D107A N404A-CFP (potential glycosylation site mutation) subcellular localization. Figure S05. ASPA1 expression in seedlings. Figure S06. ASPA3 promoter::YFP expression in lateral root cap. Figure S07. Sequence alignment of PSAPLIPs in green algae, liverwort, moss and gymnosperm. Figure S08. Sequence alignment of PSAPLIPs in angiosperms. Figure S09. Sequence alignment of PSAPLIPs which contain three SapB-like domains. Figure S10. Root growth in AT3g51730 overexpression plants. Figure S11. Root growth in AT5g01800 overexpression plants. Figure S12. Phenotype of 30 DAG Col and 35S::AtPSAPLIP1-CFP plants. Figure S13. Phenotype of 30 DAG Col and 35S::AtPSAPLIP2-CFP plants. Figure S14. Arabidopsis PSAPLIPs promoter :: GUS activity in seedlings. Figure S15. Phenotype of At5g01800 CRISPR mutant candidate. Figure S16. Phenotype of At3g51730 CRISPR mutant candidate. Figure S17. Phylogenetic tree of PSAPLIPs in plants xii Chapter 1: Literature review: Saposin-like proteins in plants Abstract The plasma membrane and endomembrane system is an essential component of all eukaryotic cells. Sphingolipids and microdomain-associated proteins that are associated with the plasma membrane and endomembrane system are important in plant growth and development. Elucidating functions of these proteins advances understanding of signal transduction from plasma membrane into cytosol and between different intracellular membrane compartments. Saposins and saposin-like proteins (SAPLIP) are among these proteins. SAPLIPs are a group of small proteins which usually consist of around eighty amino acids. Their main function is interacting with membranes. In plants, two types of proteins contain saposin B-like domains: aspartic proteases (ASPAs) and prosaposin-like proteins (PSAPLIPs). This review focuses on the functions of saposin-like proteins in animals, the reported saposin-like proteins in plants, and the knowledge gaps between plant saposin-like protein functions in vitro and in vivo. Introduction: Sphingolipids and microdomains in plants Sphingolipids are comprised of a lipid backbone and aromatic amino acid alcohol, 1 predominantly sphingosine, and there are at least 500 different molecular species of sphingolipids (Futerman et al., 2004). Sphingolipids are present in eukaryotes and some bacteria, and there are 168 different sphingolipids in Arabidopsis thaliana (Markham and Jaworsk, 2007). In general, plant sphingolipids can be classified into four groups: glycosyl inositol phosphoceramides (GIPCs), glycosylceramides (GlcCers), ceramides, free long-chain bases (LCBs) (Pata et al., 2010). GIPCs are the predominant forms of complex sphingolipids in fungi and plants, but not found in animals (Warnecke & Heinz, 2003; Worrall et al., 2003; Lynch & Dunn, 2004). Some sphingolipids are only found in certain species or tissues. For example, long chain base (LCB) d18:2 containing GlcCer is the most abundant GlcCer in tomato and soybean plants. In these two species, GlcCer distributes throughout the plant (Markham et al., 2006), while GlcCer is mainly found in flowers and pollens in Arabidopsis (Michaelson et al., 2009). Further, sphingolipid species and levels can change during development, such as in olive fruit where it increases at during fruit development and then decrease upon fruit ripening (Ines et al., 2008). This high diversity of the molecule and the regulation of its biosynthesis signifies its versatile functions in plant physiology. Sphingolipids are important components in the plant plasma membrane (PM)s and endomembrane system together with lipids, glycerolipids and sterols. In tobacco (Nicotiana tabacum) ?Bright Yellow 2? cells, GIPCs represent as high as 40% of the total PM lipids and 60% to 80% of total outer leaflet lipids (Cacas et al., 2016). Their structure contributes to the fluidity and biophysical properties of membranes (Huby 2 et al., 2019). For example, GlcCers have been implicated in chilling/freezing tolerance (Lynch and Steponkus, 1987; Imai et al., 1995; Minami et al., 2009; Takahashi et al., 2016). The Arabidopsis thaliana loss-of-function sphingolipid biosynthesis double mutant sld1sld2 (sphingolipid ? 8 long-chain base desaturases) is sensitive to cold (Chen et al., 2012). In addition to its role as a critical component in membranes, sphingolipids also show other biological functions. For example, sphingolipids are involved in programmed cell death (PCD) signaling transduction during plant development (Broderson et al., 2002; Liang et al., 2003) and immunity (Spassieva et al., 2002). Sphingolipids are also found to be necessary for sorting the membrane auxin carriers AUX1 and PIN1 from the trans-Golgi network (TGN) toward the plasma membrane (Markham et al., 2011) and glycosphingolipids with very long acyl chains stimulate lipid bilayer fusion during exocytosis and cytokinesis (Molino et al., 2014). The various species of sphingolipids may determine the biophysical and biochemical bases for microdomains formation in membranes, and these membrane microdomains may be the structural bases for the various functions of sphingolipids as a result. The microdomain model comes from experimental results such as biochemical definition of detergent resistant membranes as well as live cell imaging (Lagerholm et al., 2005; Day et al., 2009). Microdomains are lipid-ordered domains enriched in sterols and sphingolipids. They exhibit self-assembly and recruit specific proteins into their regions (Yu et al., 2020). Some sphingolipids are enriched in microdomains, such as polyphosphoinositides (PI4P and PI4,5P2) (Furt et al., 2010) 3 and some structural phospholipids are rarely found such as phosphatidylcholines and phosphatidic acids (Mongrand et al., 2004; Laloi et al., 2007). These heterogeneous membranes provide different environments for lipid-protein interactions. Some proteins can move in and out of microdomains such as aquaporin PIP2;1 (Li et al., 2011) and bacterial pathogen-associated molecular pattern flagellin (Flg22) (Keinath et al., 2010), while other proteins are very stable such as ATP Binding Cassette subfamily B (ABCB) transporters (Titapiwatanakun et al., 2008). The microdomains can integrate signals by recruiting specific proteins into the microdomains. Some proteins are exclusively localized in microdomains, such as flotillins (Borner et al., 2005) and remorins (Konrad et al., 2014). Their function may be recruit other proteins into microdomains required for physiological processes. Microdomains are important for biotic stress for aggregation of Flg22. During abiotic stress such as drought stress, AtFlot1 is involved in microdomain-mediated endocytosis of aquaporin PIP2;1 in Arabidopsis (Li et al., 2011). Microdomains also directly affect plant growth and development. For example, the auxin transporters ABCB1 and ABCB19 are localized in microdomains and these two proteins stabilize PIN1 auxin carrier in microdomains (Titapiwatanakun et al., 2008). In sphingolipid- defective mutants where microdomain structure is affected, ABCB19 is not properly trafficked or localized to the plasma membranes (Yang et al., 2012). In general, sphingolipids and microdomains are important for numerous physiological pathways in plant cells. Many of these functions are executed by the 4 interacting or associated proteins. One of these proteins is called Sphingolipid Activator PrO[S]teIN (SAPOSIN). These proteins are involved in sphingolipid metabolism in human cells. But in other species, the functions of saposin-like proteins (SAPLIPs) appear to have additional functions. Saposin-like Proteins (SAPLIPs) Saposin-like proteins (SAPLIPs) are named after saposins, which are four small proteins (Saposin A through D) derived from one single precursor called prosaposin. Saposins are important in cellular metabolism as cofactors in sphingolipid catabolism in human cells (Bruhn, 2005). SAPLIPs are found throughout eukaryotes from amoebozoans to mammals. They are not present in prokaryotes, except that three bacterial sequences have been assigned to this family in the InterPro database (http://www.ebi.ac.uk/interpro). However, these sequences lack the typical pattern of cysteine residues required to form the SAPLIP secondary and tertiary structures (accession number Q9FBA5 from Borrelia hermsii, accession number Q5XZA4 from Borrelia garinii and accession number Q5X236 from Legionella pneumophila) (Bruhn, 2005). From phylogenic data, sequence similarity among SAPLIPs is usually below 25%, which is the general threshold for a gene called a homolog (Bruhn, 2005). However, the data do show that SAPLIP domains evolved from an ancestral protein. Gene 5 duplication and subsequent mutations during evolution lead to functional versatility (Bruhn, 2005). The general features of a SAPLIP domain are six conservative cysteines and several conservative hydrophobic and polar charged residues (Bruhn, 2005). These cysteines form three pairs of disulfide bonds and together with the hydrophobic residues, forming a hydrophobic cave which allows lipid interaction. Biological functions of SAPLIPs in animals Human SAPLIPs are among the most well-studied SAPLIPs. From those studies, the function of SAPLIPs may be categorized into three types: (i) Membrane targeting (Figure 1-01A); (ii)membrane perturbation without lipid extraction (Figure 1-01B); (iii) membrane perturbation and lipid extraction (Figure 1-01C) (Bruhn, 2005). Figure 1-01. Schematic depiction of the three main activities of SAPLIPs. (A) Membrane targeting by the SAPLIP domain. (B) Presentation of lipids as substrate for an independent enzyme, either by extraction from the membrane or by disturbance of the well-packed lipid order. (C) Membrane permeabilization by perturbation owing to single molecules or by pore-formation of oligomeric proteins. Yellow and blue bars, 6 SAPLIP domain; green cloud, enzymatic domain; blue clouds, independent enzyme acting on lipids (arrows). Image is modified from Bruhn, 2005. SAPLIP domains may exist as an independent functional unit or as a part of a multidomain protein. In animals, SAPLIPs are found to participate in a variety of different functions. For example, they are co-factors of lipid-degrading enzymes (Kishimoto et al., 1992; Schuette et al., 2001), surfactant tension regulator surfactant protein B (Cochrane et al., 1991), and the antimicrobial effector NK-lysin (Pena et al., 1997). Studies show that saposins can extract lipids from membranes and load them on to the antigen-presenting molecules Cluster of Differentiation 1d (CD1d) (Zhou et al., 2004; Winau et al., 2004; Kang et al., 2004). Saposins A, B and C are implicated in various disease states whereas no known deficiency corresponding to loss of saposin D in humans has been documented. However, a saposin D mouse knockout resulted in deleterious effects (Matsuda, 2008). In general, defective saposin-disease states arise from the accumulation of ceramide derivatives in various tissues resulting in pathological states. NK-lysin is a member of the saposin-like protein family and an antimicrobial and antitumor polypeptide. It also has lytic activities against bacteria, fungi and protozoan parasites (Hong et al., 2008). Although most SAPLIP function is based on lipid binding property, recent studies found that some SAPLIP activities are independent of lipid interactions. One example is crystallin which functions in lens transparency in eyes. J3 crystallin, containing two 7 SAPLIP domains, is found in the transparent jellyfish Tripedalia cystophora (Piatigorsky et al., 1997). This raises the hypothesis that SAPLIPs are not only capable of lipid interactions, but also capable of protein-protein interactions in some cases. The function of SAPLIP multidomain proteins is less studied compared to the autonomous units. One example of a multidomain SAPLIP is the human acyloxy acylase. After proteolytic processing of precursor, the SAPLIP domain and the catalytic domain appear to be linked by a disulfide bond. The SAPLIP domain appears to be required for intracellular targeting and catalytic activity of the acylase (Staab et al., 1994), and therefore, the SAPLIP domain may contribute to the enzyme activation. The application of novel research methods will reveal more details to the function of SAPLIPs in multidomain proteins. SAPLIPs in plants SAPLIPs are found in from green algae to flowering plants. All reported plant SAPLIPs are characterized as domains or insertions in a subset of aspartic proteases. These insertions are often called plant specific inserts (PSI) because this is not found in animal aspartic proteases (Shown in Figure 1-02). 8 Figure 1-02. Schematic comparison between animal aspartic proteases and plant aspartic proteases that contain the saposin domain/plant specific insert. SP: signal peptide; NTPP: N-terminal propeptide; PSI: Plant specific insert. Interestingly, SAPLIPs in plants are the model of circular permutation: the orientation of helices is switched from N terminus to C terminus. As a result, they are sometimes called ?swaposins? (Blivem et al., 2012). The overall configuration of the secondary and tertiary structure is not affected. In addition to SAPLIPs contained in some aspartyl proteases, there are also independent SAPLIPs in plants. So far, there are no reports on either the structure or the biological functions of those independent plant SAPLIPs. 9 Figure 1-03. Schematic comparison between saposin and swaposin. The order of the helices in the swaposin is permuted relative to the saposin. However, the structure is conserved. Image is modified from Bliven et al. (2012). General structure features of SAPLIPs As mentioned above, SAPLIPs are highly diverse proteins with amino acid similarities below 30%. Although no common shared motif is found, there are conserved features: the distribution of hydrophobic amino acids forming the core and six conserved cysteine residues which form the disulfide bonds. Most reported SAPLIPs share a similar secondary structure (Bruhn, 2005). Using the NK-lysin as a typical example, 5 helices fold into two halves. The first half consists of helices 4 and 5 packed perpendicularly against helix 1. The other half contains helix 2 and 3 (Liepinsh et al. 1997) (Figure 1-04A). Saposin B is representative of other types of SAPLIPs. The two halves of saposin B crystallizes as a dimer. The saposin B monomer shows an open formation in a V shape. This has been proposed as the lipid binding position (Ahn et al., 2003). Studies from human saposins would provide information about how it functions in the cell. 10 Figure 1-04. Schematic illustration of saposin domain of NK-lysin and the swaposin domain of prophytepsin from barley. (A) The saposin domain of NK-lysin (Protein Data Bank 1NKL). (B) the swaposin domain of prophytepsin (Protein Data Bank ID 1QDM). Sequences were downloaded from Protein Data Bank. Structures were constructed in Phyre2 and illustrated in EzMol. Structure and function of human saposins. Structure of and function of prosaposins Saposins are processed from prosaposins. Prosaposins exist in three different forms - as a precursor for mature saposins, in a secreted form and as an integral membrane component (Hiraiwa et al., 1993). Prosaposin as a precursor for mature saposins is the well described but the latter two forms are not as well elucidated. Prosaposin precursor is found in two major forms, a 68 kDa intracellular form and a 73kDa extracellular form. Both forms are processed to a 50 kDa protein after 11 deglycosylation by N-glycosidase. In addition, prosaposin exists as a multimer at neutral pH and as a dimer in acidic pH (Hiraiwa et al., 1993). The prosaposin is biosynthesized, glycosylated, secreted extracellularly and it is proteolytically processed in the intracellular space of the lysosome to generate mature saposin A, B, C and D (Kishimoto et al., 1992). The signal peptide (first 16 amino acid residue) is cleaved from the preprosaposin to generate the prosaposin, as demonstrated in both rat and human milk (Kishimoto et al., 1992). Prosaposins can stimulate lysosomal ?-glucosidase and ?-galactosidase activities bound to gangliosides, which is like mature saposins. However, prosaposins cannot stimulate hydrolysis of sulfatides, while mature saposin B have this activity (Hiraiwa et al., 1993). A thiol protease was shown to catalyze the proteolysis of the recombinant prosaposin to subsequent mature saposin domains (A, B, C and D) by cleavage at the peptide linkages (Kishimoto et al., 1992). Two fragments, 39 kDa and 26 kDa, from partially purified samples of recombinant prosaposin cross-reacted with the anti-saposin C antibody. Through N- terminal sequencing, the 39 kDa protein was found to be produced by cleavage between leucine179 and phenylalanine180, corresponding to the linkage between saposin A and B, leaving a tri-saposin composed of B, C and D. The 26 kDa protein was produced by the cleavage between glutamic acid297 and leucine298 between saposins B and C, leaving a di-saposin of C and D (Hiraiwa et al., 1993; Kishimoto et al., 1992). In human cells, trisaposin B-C-D can also be processed into saposin D between 12 leucine 387 and cysteine 388, and produce the disaposin B-C. It is unable to distinguish between the pathway that liberates saposin D or saposin B from the trisaposin quantitatively (Leonova et al., 1996). The cleavage site is different for human seminal plasma prosaposin and insect prosaposin and that the cleavage of saposin A generates a derivative with 20 extra residues from the N-terminus. This suggests that post proteolysis activities are required to generate mature saposin A (Kishimoto et al., 1992). In insect cells, prosaposins are predominantly processed into A-B and C-D disaposins, and only small amount of mature saposins could be detected (Leonova et al., 1996). These findings indicate that the mature saposins come from cleavage between saposin A and B, B and C, and C and D. Structure and function of mature saposins Mature saposins A, B, C and D are structurally similar and are composed of six cysteines forming three intramolecular disulfide bonds, a glycosylation site and conserved prolines in identical positions (Kishimoto et al., 1992). Saposins are considered highly dense and firmly disulfide-linked molecules due to their high heat stability, extensive disulfide linkages and resistance to many proteases (O?Brien and Kishimoto, 1991). In addition, saposins are glycoproteins with high levels of carbohydrates. 13 Approximately 40% of total glycosylation events are found in saposin A and about 20% were present in saposins B, C and D. Saposin A is also found to have two N-linked chains, whereas in saposins B, C and D, only one N-linked chain is present (Yamashita et al., 1990). However, the carbohydrate moiety is not essential for the activation of glucosylceramides (Sano and Radin, 1988). Each saposin is composed of approximately 80 amino acid residues. Saposin A is between amino acids 60 and 143 in the prosaposin and activates ?-glucosylceramidase, ?-glucosidase and ?-galactosylceramidase (Fabbro and Grabowski, 1991). Saposin B is between amino acids 195 and 275 in the prosaposin and activates arylsulfatase A, ?- galactosidase A, GM1-?-galactosidase and various other enzymes. Saposin C is between amino acids 311 and 390 in the prosaposin and activates ?- glucosylceramidase, ?-glucosidase and ?-galactosylceramidase. Lastly, saposin D between amino acids 405 and 487 in the prosaposin is responsible for the activation of sphingomylinase (O?Brien and Kishimoto, 1991). Proteolytic cleavage is critical for the formation of these mature saposins. The four saposins differ in their hinge regions and in the alpha helix-3 sites which may allow conformational changes during association with lipids or the lipid bilayer. 15N labeled NMR spectroscopy of all four human saposins at both neutral and acidic pH showed that the mature saposins were highly unstable at pH 4.0 (John et al., 2006), but exhibited maximal ?-helical stability at pH 4.5, the optimal pH for most lysosomal hydrolases. This finding suggested that their ?-helical structures were important for 14 their physiological functions. Structural features of saposin A Under neutral conditions, saposin A is monomeric in a closed conformation. At lysosomal pH 4.8 saposin A forms a dimer, but remains in the closed conformation (Hill et al., 2015). Only in the presence of lipids or detergents does it undergo conformational change into the open state, forming lipo-protein particles with a variety of lipids (Ahn et al., 2006; Popovic et al., 2012; Hill et al., 2015). The crystal structure of open saposin A dimer with detergent lauryldimethylamine-N-oxide (LDAO) shows that 40 LDAO molecules are enclosed in the two open chains. This suggests that the dimer configuration shields the hydrophobic surface sides of monomers (Ahn et al., 2006). Structure feature of saposin C Saposin C shows similar pH- and detergent-induced oligomerization (Ahn et al., 2006). Saposin C is monomeric under neutral conditions. It has been reported to be dimeric (John et al., 2006; Rossmann et al., 2008) or trimeric (Ahn et al., 2006) in solution at low pH. Saposin C remains the closed and compact conformation which shields hydrophobic residues in the absence of lipids. In the presence of SDS micelles, it changes to an open V-shaped conformation (Haukins et al., 2005). 15 Saposin C interaction with membranes might be facilitated by neutralization of acidic residues. Negatively charged surfaces might create electrostatic repulsion from negatively charged groups of membrane lipids (De Alba et al., 2003; Hawkins et al., 2005). However, about 50% of the glutamates were neutralized at pH 5 in saposin C by pH titration measurements, although no conformational change occurred between pH 5 and 7 (Hawkins et al., 2005). As a result, several lysines in saposin C are proposed to contribute to interactions with membranes (Hawkins et al., 2005). Structure feature of saposin B Saposin B is slightly different from saposin A and C. Saposin B has been reported as the primary saposin facilitating lipid binding to CD1d molecules (Yuan et al., 2007), although all saposins promote lipid binding to CD1d. The first 24 N-terminal amino acids residues of saposins B appear to form ?-sheet configurations, while in saposins A, C and D, the helical structures are predominant (Chou and Fasman, 1978). Circular dichroism analysis has also shown that sapsoin B has high ?-sheet content (O?Brien and Kishimoto (2001). Unlike saposin A and C whose dimerization requires the presence of lipid or detergents, saposin B dimerizes at neutral and low pH, either with or without detergents (Ahn et al., 2006; Popovic and Prive, 2008). The notable feature of the dimer is the V-shaped hydrophobic open cave formed by clasping monomers. The dimers may bind one or more lipid molecules where lipid polar headgroups 16 remains in the solvent (Ahn et al.,2003; Ciaffoni et al., 2006). The saposin B pH optimum is 6, which is higher than lysosomal pH. The affinity for phospholipid membranes of saposins A, C, and D depends on low pH, in contrast to saposin B. This suggests that saposin B might facilitate lipid binding to CD1d throughout the endomembrane system (Yuan et al., 2007). Saposin B may bind, transport and transfer a large variety of membrane sphingolipids and phospholipids to lysosomes (Ciaffoni et al., 2006). In general, Sapisin B seems function as a lipid extractor and solubilizer that interacts transiently with membranes. Reports show that saposin B extracts target lipids from membranes and forms soluble protein-lipid complexes in open conformation dimeric state (Ahn et al., 2003). Structure feature of saposin D Saposin D is the least studied saposin compared to the other three. Saposin D is a ceramide activator protein involved in activation of hydrolysis of ceramide to fatty acids and sphingosines by acid ceramidases (Klein et al., 1994; Linke et al., 2001). Unlike to other saposins, saposin D (SapD) crystal structures show a compact closed monomeric form both as neutral and acid pH (Rossmann et al., 2008; Popovic et al., 2008). However, there is still the possibility that SapD forms dimers (Popovic et al., 2008). At low pH and in the presence of phospholipids, saposin D shows lipid binding activity and sphingolipid activation function (Ciaffoni et al., 2001; Linke et al., 2001; 17 Popovic et al., 2008). Mechanistic model of saposin-lipid interactions Rossmann et al. (2008) proposed a lipid solubilizer model for saposin D. Before the interactions with membranes, saposin D is in a monomer-dimer equilibrium in a closed, compact configuration. The low pH in the lysosomes neutralizes negatively charged glutamates and possible other residues, and thereby makes saposin D more hydrophobic and reduces repulsion of saposin D by negatively charged membrane surface. The ?bottom? of saposin D, which contains the positively charged amino acids, likely interacts on the intralysosomal membranes that are enriched with negatively charged lipids (Figure 1-05A). On the other side, nonpolar residues are enriched on the ?top? of the proposed saposin D dimer. This interaction may lead to moving towards hydrophobic environment by rolling the dimer by 180? around its long axis on the membrane surface. Then the hydrophobic residues are brought into membrane bilayer and the positively charged residues are exposed to the solvent. During the initial interaction with the membrane, monomer-monomer interactions in the dimer are possibly weakened by structural rearrangements in saposin D. The interaction of carbohydrate moieties with the gatekeeper amino acids in saposin D (Phe50, Phe4 and Tyr54) hide the hydrophobic interior. This initiates a hinge-bending and opening of saposin D allowing hydrophobic surface of the ? helices to insert into the membrane. Thus, the membrane structure is perturbed. 18 Saposin C functions in vesicle fusion and destabilization after the initial binding to negatively charged membranes in a manner similar to saposin D before gate opening, membrane insertion and transition to an open configuration (Figure 1-05C). Helix pairs ?1/?4 at both ends of saposin C dimers clip to opposing liposomal vesicles. As a result, the vesicles are brought close enough for fusion to occur (Rossmann et al., 2008). The schematic summary is illustrated in Figure1-06. Figure 1-05. Models of lipid activation by saposins. (A) Schematic model for saposin D (SapD)-stimulated lipid activation. Step 1: water-soluble SapD monomers and dimers bind to the negatively charged membrane surface. Step 2: SapD rotates such that the hydrophobic ??top?? of the dimer faces the membrane surface. Step 3: SapD changes configuration into a boomerang shape also found in saposin C (SapC)(below), and amphipathic ? helices stretch parallel to the lipid bilayer, exposing polar residues to the solvent. The hydrophobic surface of the saposin dips into the membrane and 19 perturbs its structure. Step 4: SapD changes configuration in the closed form, lifts a lipid molecule out of the membrane, and may leave the membrane with bound lipid. (B) Clip-on model for SapC-induced vesicle fusion proposed by Wang et al., 2003. SapC molecules anchored to phospholipid bilayers of vesicles clip to adjacent membrane layer, bringing the vesicles close enough for fusion. The size of the vesicle and saposins are not on the same scale. Image and descriptions are adopted from Rossmann et al. (2008). In summary, the general working model for SAPLIP family would be like the following: the soluble, monomeric form of SAPLIP holds a closed conformation with the hydrophobic surface hidden in the cavity. Charged residues mediated the initial contact with the negatively charged lipid membrane surface by electrostatic interactions. Then the protein change into open conformation. This change would lead to dimerization or oligomerization. This is speculated that a deeper perturbation of the membrane by interaction between the cavity and the lipid acyl chains. The membrane-embedded oligomer is hypothesized to form a pore in the membrane allowing presentation to the hydrolytic enzymes. 20 Figure 1-06. A model for the interaction with membranes and a general mechanism of action of the SAPLIP family. Figures inside the dotted green box are entirely speculative configuration that could be adopted by members of this family. (1) Soluble monomer in the solution (2) Monomer associates with membrane surface (3) Dimerization occurs on membrane surface (4) Dimers form into oligomers (5) helices insert into the membrane, and probably create a pore structure on membrane (top) (6) Saposin-like proteins loaded with lipid molecues (7) SAPLIP leaves membrane with lipid molecules. Image is adopted from Olmeda et al. (2012). Structure of SAPLIPs in plants The structure of PSIs in plants are also reported, such as cardosin A in cadoon Cynara cardunculus (Frazao et al., 1999; Egas et al., 2000), StAP in potato Solanum tuberosum (Bryska et al., 2011), phytepsin in barley Hordeum vulgare (Kervinen et al., 1999). Together with Arabidopsis aspartic proteases, these four proteins are often studied as models of SAPLIPs in plants (Figure 1-07A). 21 Figure 1-07. Plant PSI primary structure. Alignments with Clustal MUSCLE in Mega X. (A) Multiple sequence alignments for PSI from AtAP (1), phytepsin (2), cardosin A (3) and StAP (4). Red diamonds: charge differences; orange diamonds: charge inversions; green diamond: additional Trp exclusive to StAP PSI. (B) Multiple sequence alignments for human saposin B (1), human saposin C (2) and StAP PSI (3) colored based on sequence conservation. StAP PSI is presented with its N-and C-terminal halves swapped to align with the saposins. Image and descriptions are adopted from Bryksa et al., 2017. In general, plant PSIs are similar with human saposins in terms of sequence features and overall structure (Figure1-07B). The six conserved cysteines also appear in plant PSIs. PSI from phytepsin shows similarity with NK-lysin (Kervinen et al., 1999). PSI from cardoon shows high similarity to human saposins C and it can also activate human glucosylceramidase in vitro (Brodelius et al., 2005). PSI of StAP is also reported to show similarity to human saposin C and it also induces vesicle disruption in vitro. The secondary conformation is pH-dependent, which is similar to human saposins (Bryksa et al., 2011). A recent study shows high sequence similarity and conservation among these 22 four PSI. They all show leakage activity in bilayer composed of a vacuole-like phospholipid mixture and membrane fusion activity in vitro. This activity is pH- dependent. The leakage activity is higher at pH 4.5 and requires the presence of acidic phospholipids such as phosphatidylserine. Low pH results in dimerization of potato PSI, and the monomer is prevalent under neutral pH. All the studies support that plant PSIs are similar to mammalian saposins in terms of structure and molecular activities. As mentioned above, low pH activates bilayer membrane leakage activity. Conformation change is likely the molecular basis of this. A recent study found a novel 6-residue motif in H3 helix [N/Q]-[N/Q]-[N/Q]-[A/L/I/V]-[K/R]-[N/Q] which may contribute to this configuration change. A point mutation K83Q in this motif in helix H3 blocks the response to low pH activation with respect to conformation change (Bryksa et al., 2017). This motif may be responsible for lipid-interactions as this motif is also found in several other membrane-interacting proteins (Bryksa et al., 2017). This motif is not seen in human saposins. As PSI is part of the aspartic protease, elucidating functions of aspartic proteases helps the better understanding the role of PSI in plant cells. Plant aspartic proteases General Information about plant aspartic protease Proteases are an important for physiological processes and in commercial 23 applications. Proteases are one of the most important type of industrial use enzymes and they comprise approximately 60% of all commercial enzymes on the market (Feijoo-Siota and Villa, 2011). The diverse applications include food science and technology, the pharmaceutical industry, and detergent manufacturing. (Feijoo-Siota and Villa, 2011). Aspartic proteases in Arabidopsis In Arabidopsis thaliana genome, there are over 550 protease sequences categorized into five types: serine, cysteine, aspartic acid, metallo and threonine (MEROPS peptidase database, http://merops.sanger.ac.uk/). This reflects a wide variety of biological functions. (Beers et al., 2004). Aspartic proteases (family A1) are characterized by a common bilobal tertiary structure containing two catalytic aspartic acid residues. They are found in all higher organisms. The most noticeable feature of A1 family is that there are two conservative aspartyl sites. They are believed to be involved in the processing of propeptides in various plant tissues, such as in the breakdown of storage protein in seed germination (Belozersky et al. 1989) and the proteolytic processing and maturation of storage proteins (Hiraiwa et al. 1997b, Runeberg-Roos et al. 1994). They have also been shown to be involved in the turnover of pathogenesi-related proteins in tobaccos induced by stress (Rodrigo et 24 al. 1991), plant senescence and programmed cell death (Chen and Foolad 1997, Cordeiro et al. 1994). Plant aspartic proteinases are also used by man for food processing. Protein extracts of Cynara cardunculus are used for cheese manufacturing (Cordeiro et al. 1992), and the aspartic proteinases from cocoa are important in the fermentation process of the beans for generation of flavor peptides from storage proteins (Biehl et al. 1985). There are 59 annotated Arabidopsis A1 proteases identified. Predicted aspartic proteases from other families include the A11 family (approximately 45 members) retrotransposon endopeptidases and two presenilin-like proteins from A22 family (AAL24266 and AAD23630) (Beers et al., 2004). According to the sequence similarity, they are divided to five subfamilies, A1-1 (35 members), A1-2 (17 members), A1-3 (2 members), A1-4 (3 members) and A1-5 (2 members). The main difference between each subfamily is the number and distribution of exons and introns (Beers et al., 2004). 25 Figure 1-08. Phylogenetic tree of A1 family proteases in Arabidopsis. Neighbor-jointing tree based on point accepted mutation (PAM) distances generated from an alignment of protein sequences in ClustalW. Five putative groups are indicated with colors. At5G33350 is mentioned in Beers et al., 2004 as a member in A1 family, but it doesn?t contain the first conservative aspartic site. It is outside the five subgroups in this tree. ASPA1 (At1g11910), ASPA2 (At1g62290) and ASPA3 (At4g04460) are indicated in purple characters. The subfamilies are based on Beers et al., 2004. Several members in A1-1 subfamily are annotated as ?chloroplast nucleoid DNA 26 binding protein-like? (CND41-llike), originally identified from tobacco CND41(T01996). CND41 exhibits both proteolytic (Murakami et al., 2000) and chloroplast DNA-binding (Nakano et al., 1997) capabilities in vitro. No biological functions are reported for A1- 3 and A1-5 subfamilies. A1-4 subfamily is the only one that contains the SAPLIP domain. Biological functions of aspartic proteases in plants The biological function of the A1-4 subfamily in several plant species have been reported. Cardosin A and cardosin B accumulate during seed maturation, and cardosin A is synthesized de novo at the time of radicle emergence. This suggests that cardosins are involved in storage protein processing during seed development and protein degradation during seed germination (Pereira et al., 2008). Cardosin B gene expression was also observed in pistils and ovules and is proposed to be involved in programmed cell death dependent degeneration of nucellus in cardoon (Figueiredo et al., 2006; Pereira et al., 2008). An aspartic protease in leaves of common bean (Phaseolus vulgaris) was found in a screen for drought tolerance (Cruz de Carvalho et al., 2001), and is upregulated by water stress in beans. A typical aspartic protease in pinapple fruit Ananas comosus was found to be upregulated by chilling treatment, which suggests a role in chilling stress resistance (Raimbault et al., 2013). Overexpression of a sweet potato aspartic protease SpAP1 promotes ethephon-induced leaf senescence (Chen et al., 2015). In the pitcher plant Nepenthes alata, of the five aspartic proteases 27 identified and four of them contained the SAPLIP domain. NaAP2 and NaAP4 transcripts were detected in the digestive glands which suggests that they are associated with secretion (An et al., 2002). Arabidopsis has three typical aspartic proteases and they are also called phytepsins due to their similarity to mammalian enzymes pepsin and cathepsin D. The homolog in yeast Saccharomyces cerevisiae PEP4 is required for activation of several vacuolar zymogens (Van den Hazel et al., 1992). The biological functions of the A1-4 subfamily in Arabidopsis is not well studied, but the expression patterns may provide information about their biological functions. There are three members of A1-4 subfamily in Arabidopsis, named ASPA1 (At1g11910), ASPA2 (At1g62290) and ASPA3 (At4g04460) respectively. ASPA1 mRNA is detected in all tissues and is abundant in leaves during daytime. ASPA3 is primarily expressed in flowers and ASPA2 is primarily expressed in seeds. (Chen et al., 2002). All three genes can be detected in developing siliques and seeds, which suggests their roles in seed development. Their expression patterns also suggest that they have multiple roles in Arabidopsis development. All three proteases are targeted in the vacuoles (Otegui et al., 2006; Figure 2-09; Figure S07). In plants, there are two kinds of vacuoles: the central lytic vacuole and the protein storage vacuole. The central vacuole resembles animal and yeast vacuoles, while the protein storage vacuole is plant specific. Protein storage vacuoles are found in germinating seedlings (Paris et al. 1996; Swanson et al. 1998), nutrition storage 28 tissues such as tubers, leaves and tree bark (M?ntz and Muntz 1998; Zouhar et al. 2010). The structure of protein storage vacuoles is complex: there is a ?globoid cavity? exhibits an acidic environment similar to the central vacuole, and it is partitioned within a more neutral protein storage vacuole lumen (Jiang et al. 2001; Tse et al. 2007). The neutral lumen allows the storage proteins to stay for longer time from degradation. As a result, protein storage vacuoles have dual functions (Xiang et al. 2013). The intracellular trafficking pathway is similar between the lytic vacuole destination and the protein storage vacuole destination, but there are specific receptors in protein storage vacuoles (Hinz et al. 1999). Protein storage vacuoles reserve nitrogen in the form of storage proteins during seed maturation. Seed storage proteins are the major component of many agriculturally crops, such as legume seeds (40% dry weight) (Bradford and Bowley 2003; Atta et al. 2004; Gottschalk and Muller 2012). During germination, protein storage vacuoles fuse to form a single central vacuole, and seed storage proteins are catabolized to provide amino acids for protein biosynthesis (Jiang et al., 2001). The primary storage proteins found in mature seeds are classified based on solubility as albumins (water-soluble), globulins (salt-soluble), prolamins (alcohol- soluble) or glutelins (weak-acid/weak-base soluble) (Ferreira et al. 1999). In Arabidopsis, the predominant seed storage proteins are the 12S legumin-type globulins and the 2S napin-type albumins (Gruis et al. 2002). Seed storage proteins need to be post-translationally modified for stable and dense package. Both the 2S and 12S proteins are translated as long precursors, inserted into the ER lumen, and 29 undergo post-translational cleavage en route to the protein storage vacuoles (Paris et al. 1996; Hara-Nishimura et al. 1998a; Swanson et al. 1998; Gruis et al. 2002; Otegui et al. 2006). For the 12S globulin precursors, the ER signal peptide is cleaved after insertion into the ER lumen, and disulfide bonds form and the proteins assemble into trimers (Muntz, 1998), then transported via the Golgi body to the protein storage vacuoles. In the multivesicular body or prevacuolar compartment, 12S are processed by the enzymes known as Vacuolar Processing Enzymes (VPEs) which produce mature disulfide-linked ?- and ?- chains (Otegui et al., 2006; Baud et al., 2008). The VPEs are a family of asparagine-specific cysteine endopeptidases which cleave seed storage proteins in vitro and in vivo (Hara-Nishimura et al. 1991; Gruis et al 2002; Gruis et al. 2004). These proteases were identified in the seed, the leaf and the root. They are involved in senescence, programmed cell death and biotic defences (Hara-Nishimura et al. 1991; Kinoshita et al. 1995; Misas-Villamil 2013). The expression pattern of ASPA1 is in parallel with VPEs and it is believed that it is also involved in seed storage protein procession. Reports have shown that ASPA1 was highly expressed during embryo development, accumulated in protein storage vacuoles, and has been shown to cleave 2S seed storage protein napins in vitro (D?Hondt et al. 1993a; Mutlu et al. 1999; Otegui et al. 2006). The involvement of VPEs in seed germination is not documented, but the detection of ASPA in seed germination is reported (Pereira et al., 2008). It is possible that ASPAs are involved in seed germination regulation and possibly degrade seed 30 storage proteins during germination. Among the three ASPAs in Arabidopsis, the most studied gene is ASPA3 due to its specific expression pattern. The ASPA3 promoter-reporter constructs showed signals in almost all tissues that undergo programmed cell death (PCD), such as lateral root caps, tracheary elements in proxylem, fading petals, tapetum in stamens and endosperm in developing seeds (Fendrych et al., 014; Olvera-Carrillo et al., 2015). In plants, PCD can be categorized into two types: developmental PCD or environmental PCD (Daneva et al., 2016). Developmental PCD occurs in specific cell types such as tracheary elements in xylem, lateral root caps and tapetum in stamens, in order to facilitate normal growth and development. Developmental PCD can also be triggered conditionally by cell signaling, which could be seen in self-incompatibility responses (Wilkins et al. 2014; Petrov et al. 2015). Developmental PCD also occurs in all types of aging cells in the end of plant senescence (Klime?ov? et al. 2015). Environmental PCD occurs in response to stresses such as irradiation or pathogens (Wu et al. 2014). A portion of cells sacrifice in order to protect the remaining tissues. Comparative studies showed that these two types are different in transcriptional signaling (Olvera-Carrillo et al. 2015). However, the executing components downstream of many PCD-related transcriptional factors are shared in different cell types (Olvera-Carrillo et al, 2015; Huysmans et al., 2018), such as ASPA3, BIFUNCTIONAL NUCLEASE1 (BFN1), RIBONUCLEASE3 (RNS3), CYSTEIN ENDOPEPDITASE1 (CEP1), DOMAIN OF UNKNOWN FUNCTION679 MEMBRANE 31 PROTEIN4 (DMP4) (Olvera-Carrillo et al., 2015; Ye et al., 2020). In terms of morphology, PCD can be divided into three types: apoptosis, autophagic cell death, and necrosis (Lockshin and Zakeri, 2004; Bras et al., 2005). In necrosis, organelles swell up and the plasma membrane ruptures to release the components. This type is less studied and is believed less controlled by genetic programming. In apoptosis, the cells shrink, DNA is fragmented into small pieces and cell component is compacted into small vesicles. In autophagic type, autophagic vacuoles are formed for degradation of cell components, but the cell doesn?t necessarily die, which distinguish from apoptosis (Theresa et al., 2008). The early events of apoptosis process include caspase signaling (Danon et al., 2004) and proteases synthesis in preparation for protein degradation and nucleases synthesis for DNA fragmentation (Fendrych et al., 2014). Take PCD in lateral root caps as an example, the events during PCD include a decrease in cytoplasmic pH, plasma membrane permeabilization, vacuolar collapse, and final degradation cell materials (Fendrych et al. 2014). ASPA3 is believed to one of the proteases in this process, although the single mutant of ASPA3 doesn?t show a PCD-related phenotype in lateral root caps (Fendrych et al., 2014). A recent study showed a potential role of ASPA1 in drought tolerance by overexpression in Arabidopsis (Sebasti?n et al., 2020). Overexpression lines of ASPA1 had longer primary root length under drought conditions. The overexpressors also had reduced stomata index, reduced stomata density and a smaller stomatic aperture 32 compared to wild type plants. Higher expression levels of genes related to ABA signaling and biosynthesis were also found in ASPA1 overexpression lines. ASPA1 promoter-GUS activity showed that ASPA1 was induced by ABA in leaves. These results indicate multiple roles of ASPAs in plant growth and development. Structure of plant aspartic protease Most known aspartic proteases are from a single chain proenzyme which are then proteolytically processed and then form either a homomeric or heterodimeric mature enzyme (Laloi et al., 2002; Simoes and Faro, 2004).The proenzyme is characterized by a hydrophobic signal peptide, an N-terminal propeptide of approximately 40 amino acid residues, and the mature protein region composed of an N-terminal domain and a C-terminal domain, separated by the plant specific insert (PSI, SAPLIP) of approximately 100 amino acids (Koelsch et al., 1994; Asakura et al., 1995). Phytepsin undergoes several proteolytic cleavage steps to produce the two-chain form of the mature enzyme (Figure 1-09). Early processing involves insertional removal of the signal peptide at the endoplasmic reticulum (ER), followed by removal of the N- terminal propeptide (Glathe et al. 1998), and then PSI cleavage, followed/accompanied by cleavage 5 kDa upstream of the PSI cleavage point (Ala-378) gives a heavy chain and a light chain. This process is BFA sensitive, indicating that it occurs after the peptide has passed the Golgi. Finally, these chains are processed to mature forms by removal of the remaining PSI; this occurs only 24h after synthesis in 33 vivo (Glathe et al. 1998). This step is critical for activation of the enzyme. Proteolytic processing of procardosin A to its mature form is highly pH sensitive in vitro, with optimal processing at pH 4, but retarded processing at pH 3 or 5 (Castanheira et al. 2005). This processing is unlikely to be totally autocatalytic; although processing was active in vitro and inactivated by pepstatin A and cleavage patterns were different from those observed in vivo (Glathe et al. 1998). One interesting example is found in an aspartic protease cirsin from Cirsium vulgare. The procirsin was expressed in Escherichia coli and shown to be active without autocatalytically cleaving its propeptide domain (Lufrano et al., 2012). This contrasts with the acid-triggered autoactivation by pro-segment removal. Recombinant procirsin displayed all typical proteolytic features of aspartic proteinases such as optimum acidic pH, inhibition by pepstatin, cleavage between hydrophobic amino acids and strict dependence on two catalytic Asp residues for activity. In general, complete activation of typical aspartic proteases requires correct processing at a pH conducive to efficient cleavage. The cDNA encoding the precursor of AtASPA1 was expressed as a functional protein using the yeast Pichia pastoris. The mature form of the recombinant AtASPA1 was found to be a heterodimeric glycosylated protein with a molecular mass of 47 kDa consisting of heavy and light chain components, approximate 32 and 16 kDa, respectively, linked by disulfide bonds. Glycosylation occurred via the plant specific insert in the light chain. The catalytic properties of the recombinant AtASPA1 were similar to other plant aspartic 34 proteinases with activity in acid pH range, maximal activity at pH 4.0, Km of 44 ?M, and kcat of 55 s?1 using a synthetic substrate. The enzyme was inhibited by pepstatin A (Miguel et al., 2008). Figure 1-09. Schematic illustration of the protein processing steps for ASPA1. NTPP: N- terminal propeptide; SP: signal peptide; PSI: plant specific insert. First, the signal peptide (SP) is removed, then N-terminal propeptide (NTPP) is trimmed from the propeptide. Then the plant specific insert (PSI) is cleaved after transporting from Golgi body. The heavy chain (27kDa) and the light chain (10kDa) assemble into the mature protease. Destination of PSI is still unclear. Cleavage sites are predicted from alignment with cardosin A. Image is modified from Miguel et al., 2008. Biological function of plant specific insert The plant specific insert (PSI) distinguishes plant aspartic proteases from animal ones, and the biological functions of PSI have been a subject of study for a long time. Studies show that PSI is likely to be critical for vacuolar targeting (Kervinen et 35 al.,1999; Terauchi et al., 2006). The intact proteases are targeted to vacuoles across the plant kingdom, from moss to seed plants (Schaaf et al., 2004; Kervinen et al., 1999; Teraochi et al., 2006). The deletion of PSI in soyAP2 resulted in the retention of peptides in ER (Teraochi et al., 2006). The PSI brings phytepsin into contact with membranes, possibly with membrane- binding receptors proteins in Golgi apparatus. The prophytepsin is then trafficked to vacuoles and activated by proteolytic cleavage and the PSI is subsequently removed which breaks the interaction of the membrane receptor or the membrane itself (Kervinen et al., 1999). The PSI also influences the route that phytepsin takes after leaving the endoplasmic reticulum, in addition to the vacuolar sorting function. It is currently unclear whether the PSI has two sorting signals: one for endoplasmic reticulum export or another one for vacuolar sorting or if the vacuolar sorting motif is recognized at the endoplasmic reticulum export site (Tormakangas et al., 2001). Studies show that removing the plant specific domain has no effect on phytepsin activity; however, it does cause an accumulation of phytepsin in the extracellular space of the plant (Tormakagas et al., 2001). These findings support the role in vacuolar target and sorting, and the default secretion pathway proceeds without the PSI. Another possible function of PSI is in pathogen defense pathway. Solanum tuberosum aspartic protease (StAP) PSI in vitro is able to kill spores of two potato pathogens in a dose-dependent manner without any deleterious effect on plant cells 36 (Mu?oz et al., 2010). The StAP-PSI ability to kill microbial pathogens is dependent on the direct interaction of the protein with the microbial cell wall/or membrane, leading to increased fungi or bacteria cell permeability and lysis. StAP-PSI is able to kill human pathogenic bacteria in a dose dependent manner as well, but it is not toxic to human red blood cells at the concentrations and times assayed. Minimal bactericidal concentration (MBC) values determined for StAPs and StAP PSI are in the same order of magnitude as those previously reported for NK-lysin and granulysin (Mu?oz et al., 2010). The constitutive expression of StAP-PSI induces defense genes in Arabidopsis and enhances Arabidopsis resistance against Botrytis cinerea infection. The StAP-PSI domain exerts cytotoxic activity toward Botrytis cinerea, and the constitutive expression of StAP-PSI increases growth in A. thaliana (Frey et al., 2018). In a similar study, Pagano et al. (2007) investigated the importance of glycosylation, to the Solanum tuberosum aspartic protease. It was observed that aspartic protease accumulation into the apoplast of tubers and leaves after wounding required glycosylation (Pagano et al., 2007). This suggests that glycosylation may be necessary for Solanum tuberosum aspartic protease membrane and/or protein interactions. In a more recent study, the role of glycosylation was thought to affects intracellular trafficking of the aspartic protease. The recombinant cardosin B PSI undergoes the conventional route from ER, the Golgi and the prevacuolar compartment to the vacuole if glycosylated. The non- glycosylated proteins entered vacuoles directly from ER and bypassed the Golgi bodies 37 (Vieira et al., 2019). This study suggests that there are unconventional trafficking pathways in the plant cell. It also elicits a more complex question: what?s the exact function of glycosylation in PSI? In general, most of reported data was based on the hypothesis that PSI is released during aspartic protease procession and secreted to the extracellular space. This needs more experimental data to support. Perspective Since the amino acid sequences of plant aspartic proteinases were described in the 1990s, a wide range of studies have explored their function in development largely and defense responses. In vitro biochemical and biophysical methods have been used to elucidate the function of the plant specific insert, but corresponding in vivo studies are largely lacking. There are some studies on the expression pattern of aspartic proteases, but there are few in vivo studies, especially the genetic studies. So far, no mutants of these aspartic proteases with growth and development phenotypes have been reported yet. A better understanding of these unique proteases could eventually be applied to agricultural production as a tool to manipulate plant growth and development. For SAPLIPs in aspartic proteases, most studies are in vitro, and the in vivo studies are not with the whole length of the protein in context. This leads to a question: does 38 the PSI function independent from the protease or does it function in coordination with the protease? Or does it simply function as a signal for trafficking towards vacuoles? And what about the SAPLIPs which are independent units in plants? These SAPLIPs are called prosaposin-like proteins (PSAPLIPs) in this dissertation, and there are no published articles regarding these proteins to date. The biological functions of these PSAPLIPs also need to be explored. This will contribute to the current knowledge of plant cellular biology. With the in vitro studies of PSI and mammalian saposins, and the relatively small size of this protein family, this will also produce biological tools for cell biology in both scientific and industrial areas. As a result, three questions are raised and will be discussed in this dissertation: (1) What?s the biological function of the typical aspartic protease and its PSI in Arabidopsis? Is there an independent role of PSI from the proteolytic domains? (2) How these aspartic proteases function in Arabidopsis cells? Are they involved in programmed cell death as proposed? (3) What?s the feature of plant prosaposin-like proteins? What are the biological functions of the plant prosaposin-like proteins in growth and development? 39 Chapter 2 Functional study of saposin-like domain containing aspartic proteases in Arabidopsis thaliana Abstract Aspartic proteases (ASPAs) are important in plant growth and development. They are also one of the most important commercial proteases. Aspartic protases containing a plant specific insert (PSI) sequence have been the subject of numerous studies. Several studies have reported the properties of PSI in vitro. However, few have reported the function of PSI or aspartic proteases that contain them in vivo. Here molecular genetic analysis has revealed that ASPAs were involved in seed maturation and regulate seed germination, root morphology in response to nitrogen supply, and were involved in programmed cell death (PCD) execution in Arabidopsis. Triple mutant aspa1-2 aspa2-1 aspa3-3 showed delayed seed germination. Protein storage vacuoles fusing into the central vacuole was also delayed in the triple mutant during seed germination. ASPA2 was expressed throughout the plant similar with ASPA1. ASPA2 expression was responsive to environment stresses while ASPA1 expression was stable. ASPA2 was first imported into the endoplasmic reticulum and then through the endomembrane system: trans-Golgi network (TGN) then multivesicular bodies (MVB), and finally was transported to vacuoles. PSI was important for vacuolar localization, but site-directed mutagenesis revealed that the lipid binding motif inside 40 the PSI didn?t seem to be required. ASPA2 colocalized with early endosome (EE) marker RabA3, which indicates the possibility for plasma membrane protein degradation. Root morphology was affected in aspa triple mutant. Primary root length was longer for triple mutant under low nitrogen levels. This is likely to result from the delayed programmed cell death of tracheary elements in the mutant xylem. Further analysis revealed that membrane permeability increased more slowly in the triple mutant lateral root cap cells during PCD. These results indicate that ASPA2 is involved in membrane disturbance during programmed cell death and modulate the rate of membrane permeability increase and therefore the rate of programmed cell death. These results indicate that ASPAs function in proteolytic activities in bulk and membrane disturbance. The independent role of PSI was not supported in this dissertation. This is the first time showing that ASPAs process seed storage in vivo and have impacts on seed maturation and seed germination. This is also the first time showing that ASPAs are involved in programmed cell death by promoting membrane permeability. Introduction Aspartic proteases are important in commercial application such as milk clotting for cheese making (Heimgartner et al., 1990). Aspartic proteaes are also important in plant growth regulation. For example, cardosin A in cardoons is involved in storage protein processing during seed development and protein degradation during seed 41 germination (Figueiredo et al., 2006; Pereira et al., 2008). Aspartic proteases are also believed to be associated with drought tolerance in the common bean (Phaseolus vulgaris), chilling response in pineapple (Ananas comosus) and senescence in sweet potato leaves (Cruz de Carvalho et al., 2001; Raimbault et al., 2013; Chen et al., 2015). Aspartic proteases are also found in digestive ligands in pitcher plants (Nepenthes alata) (An et al., 2002). The typical ASPA aspartic proteases contain an N-terminal propeptide and a plant specific insert (PSI) inside the protease peptide. Animal aspartic proteases lack the PSI sequence. The PSI resembles human saposin proteins in primary and secondary structures. Saposin-like proteins in animals have been well-studied and their major function is interacting with membrane lipids (Bruhn, 2005). In plants, the activity of the plant specific insert from the aspartic proteases have also been studied. This plant specific insert is critical for vacuolar targeting of the protease (Kervinen et al.,1999; Terauchi et al., 2006). The intact proteases are trafficked to vacuoles in the moss nad soybean (Schaaf et al., 2004; Kervinen et al., 1999; Teraochi et al., 2006), and deletion of PSI in soyAP2 result in the retain of peptides in the endoplasmic reticulum (ER) (Teraochi et al., 2006). Studies also show that removing the plant specific insert has no effect on phytepsin activity; however, it does cause an accumulation of phytepsin in the extracellular space of the plant (Tormakagas et al., 2001). Other possible function of plant specific insert is in pathogen defense pathway. Solanum tuberosum aspartic protease (StAP) PSI in vitro is able to kill spores of two potato pathogens in a dose- 42 dependent manner without any deleterious effect on plant cells (Mu?oz et al., 2010). The constitutive expression of StAP-PSI induces defense genes expression in Arabidopsis and enhances Arabidopsis resistance against Botrytis cinerea infection (Frey et al., 2018). The StAP-PSI domain exerts cytotoxic activity toward Botrytis cinereal (Frey et al., 2018). These findings suggest a role of PSI in plant defense, and it is assumed that this plant specific insert functions independently from the protease. A recent study indendified a six amino acid motif in plant specific insert which accounts for conformation change and lipin bilayer fusion activity in vitro (Bryska et al., 2017). This motif is unique to plant saposin-like domains, which indicates that plant specific insert functions in a way different from animal saposin-like proteins. In plant cells, ASPAs undergo several proteolytic cleavage steps to produce the two-chain form of mature enzyme during which the PSI is released (Glathe et al., 1998). The destination of the released PSI is still not clear. This leads to the hypothesis that PSI function independently in vivo for normal plant growth and development. In Arabidopsis, there are 59 annotated Arabidopsis A1 aspartic proteases identified (Beers et al., 2004). Only the A1-4 subfamily contains the saposin-like domain or plant specific insert. A1-4 subfamily has 3 members named ASPA1 (At1g11910), ASPA2 (At1g62290) and ASPA3 (At4g04460), and they are also called phytepsins due to their similarity to mammalian enzymes pepsin and cathepsin D. The biological functions of the A1-4 subfamily in Arabidopsis is not well studied. Their homolog in yeast Saccharomyces cerevisiae PEP4 is reported for activation of several 43 vacuolar zymogens (Van den Hazel et al., 1992). This suggests that Arabidopsis ASPAs may also be activatice in the vacuoles. The expression patterns may provide information about their biological functions. ASPA1 mRNA is detected in all tissues. ASPA3 is primarily expressed in flowers and ASPA2 is primarily expressed in seeds. (Chen et al., 2002). These expression patterns suggest that they have multiple roles in Arabidopsis development. ASPA1 is a well-known marker for prevacuolar compartment in developing seeds (Otegui et al., 2006) and believed to take part in seed storage proteins processing. In Arabidopsis, the predominant seed storage proteins are the 12S legumin-type globulins and the 2S napin-type albumins (Gruis et al. 2002). Seed storage proteins need to be post-translationally modified for stable and dense package. Both the 2S and 12S proteins are translated as long precursors, inserted into the ER lumen, and undergo post-translational cleavage en route to the protein storage vacuoles (Paris et al. 1996; Hara-Nishimura et al. 1998a; Swanson et al. 1998; Gruis et al. 2002; Otegui et al. 2006). In the prevacuolar compartment, seed storage proteins are processed by the enzymes known as Vacuolar Processing Enzymes (VPEs) which produce mature disulfide-linked ?- and ?- chains (Otegui et al., 2006; Baud et al., 2008). The VPEs are a family of asparagine-specific cysteine endopeptidases which cleave seed storage proteins in vitro and in vivo (Hara-Nishimura et al. 1991; Gruis et al 2002; Gruis et al. 2004). The expression pattern of ASPA1 is in parallel with these VPEs and it may also process seed storage proteins, as studies showed that ASPA1 accumulated in protein 44 storage vacuoles, and ia able to cleave 2S seed storage protein napins in vitro (D?Hondt et al. 1993a; Mutlu et al. 1999; Otegui et al. 2006). The involvement of VPEs in seed germination is not documented, but the detection of ASPA in seed germination is reported (Pereira et al., 2008). It is possible that ASPAs are primary proteases in seed germination for degradation of seed storage proteins rather than the VPEs. This leads to another hypothesis that ASPAs regulate seed storage proteins procession during seed germination. Amont the three ASPAs in Arabidopsis, the most studied gene is ASPA3 due to its specific expression pattern in tissues that undergo programmed cell death (PCD), such as lateral root caps, tracheary elements in proxylem, fading petals, tapetum in stamens and endosperm in developing seeds (Fendrych et al., 2014; Olvera-Carrillo et al., 2015). In plants, PCD can be categorized into two types: developmental PCD or environmental PCD (Daneva et al., 2016). Developmental PCD occurs in specific cell types such as tracheary elements in xylem, lateral root caps and tapetum in stamens, in order to facilitate normal growth and development. Developmental PCD also occurs in all types of aging cells in the end of plant senescence (Klime?ov? et al. 2015). Environmental PCD occurs in response to stresses such as irradiation or pathogens (Wu et al. 2014). Comparative studies showed that these two types are different in transcriptional signaling (Olvera-Carrillo et al. 2015). However, the executing components downstream of many PCD-related transcriptional factors are shared in different cell types (Olvera-Carrillo et al, 2015; Huysmans et al., 2018), such as ASPA3, 45 BIFUNCTIONAL NUCLEASE1 (BFN1), RIBONUCLEASE3 (RNS3), CYSTEIN ENDOPEPDITASE1 (CEP1), DOMAIN OF UNKNOWN FUNCTION679 MEMBRANE PROTEIN4 (DMP4) (Olvera-Carrillo et al., 2015; Ye et al., 2020). But the expression time is different for these genes. In Arabidopsis stigmas, the expression order is CEP1 first, then ASPA3, and BFN1 is the last (Gao et al., 2018). CEP1 functions in the cytosol, and it is likely to participate in the signaling transduction. BFN1 functions in the nuclei for the final degradation of DNA. This result indicates that ASPA3 may function after the upstream singaling events and is involved in bulk proteolytic activity of cell components. In lateral root cap PCD, the final events include a decrease in cytoplasmic pH, plasma membrane permeabilization, vacuolar collapse, and final degradation cell materials (Fendrych et al. 2014). It is likely that ASPA3 is one of the crucial proteases in these final steps, although the single mutant of ASPA3 doesn?t show a PCD-related phenotype in lateral root caps (Fendrych et al., 2014). It is likely that ASPA1 and ASPA2 have reduncancy roles, and the third hypothesis can be drawn that ASPAs regulate programmed cell death in Arabidopsis. A recent study showed another potential role of ASPA1 in drought tolerance by overexpression in Arabidopsis (Sebasti?n et al., 2020). Overexpression lines of ASPA1 had longer primary root length under drought conditions. The overexpressors also had reduced stomata index, reduced stomata density and a smaller stomatic aperture compared to wild type plants. Higher expression levels of genes related to ABA signaling and biosynthesis were also detected in ASPA1 overexpression lines. ASPA1 46 promoter-GUS activity showed that ASPA1 was induced by ABA in leaves. These results indicate multiple roles of ASPAs in stress reponses. ASPAs have been inferred as important proteases in plant growth and development, yet few studies have reported their activities in vivo. In this dissertation, three hypotheses mentioned above were tested, and the results showed that ASPAs function in seed storage protein processing in vivo during seed germination. ASPAs also regulate programmed cell death by membrane disterbation to increase membrane permeability in lateral root caps. These results indicate that ASPAs may regulate nitrogen storage and recycle in the plants. Results ASPAs function in seed development and germination To explore the possible biological functions of ASPAs in Arabidopsis, T-DNA insertion mutants were used for phenotypic studies. There are three ASPA genes containing the saposin-like domain in Arabidopsis. Two alleles of ASPA2 were obtained and characterized. The T-DNA insertions were verified by PCR. The aspa2-1 (SALK097505) harbors a T-DNA insertion in the promoter and aspa2-2 (SALK021601) has an insertion in the 5? untranslated region (Figure 2-01A). Quantitative real-time PCR revealed that both mutant lines are null alleles (Figure 2-01B). The single aspa2 mutants showed delayed seed maturation reflected by the delayed seed size increase and delayed color change (Figure 2-02A). The delay corresponds to the heart stage of 47 development (Figure 2-02B). However, the phenotype was subtle, which suggests the possibility that other two ASPAs might have redundancy roles. T-DNA insertion mutants were also screened for ASPA1 and ASPA3. T-DNA insertion lines for ASPA1 and ASPA3 were obtained. The aspa1-1 (SALK092586) has an insert in the 5? untranslated region, while aspa1-2 (SALK041027) has an insertion in the second intron. The allele aspa3-3 (SALK056711) has an insertion in the fifth intron, 5? of the PSI domain (Figure 2-02A). The aspa1 alleles are both knock-down alleles with approximately 40% expression for wild type, and aspa3-3 is a null allele (Figure 2-02B). The triple mutant aspa1-2 aspa2-1 aspa3-3 (aspa1-2/2-1/3- 3 for short) was generated. A triple knockout could not be obtained, since aspa1 alleles are knockdowns. Seeds of triple mutants were slightly larger and weighted almost twice as much compared to the wild type seeds (Figure 2-02C, D, E). This may result from delayed seed development allowing accumulation of more storage materials in the seeds. Considering the function of ASPAs in other species, ASPA2 is likely involved in seed storage protein processing. 48 Figure 2-01. Characterization of ASPA T-DNA insertion mutants. (A) Schematic diagrams of the gene models and the T-DNA insertion sites in ASPA mutants. The inverted triangles represent the T-DNA insertion sites in the genomic DNA. Boxes represent the exons in the genomic DNA. Black boxes represent translated regions. White boxes represent untranslated regions. Intervening lines between black boxes represent the introns. (B) Quantitative real-time PCR of in T-DNA insertion mutants. 49 The inflorescence and opening flowers were harvested to extract RNA and quantitative real-time PCR was conducted to detect ASPA gene transcription level. ACTIN2 was chosen as an internal standard. Three biological replicates are represented in quantitative real-time PCR experiments. N.D.: none detected. Seed germination was also affected in aspa1-2 aspa2-1 aspa3-3 mutant. Seeds germinated more slowly than wild type seeds (Figure 2-03A), and this delay could not be rescued by gibberellin acid treatment (Figure 2-04A). This suggests that the delay may not result from transcriptional events. This also indicates that the major source of ASPAs during germination is synthesized during seed development, not newly synthesized after imbibition. Stratification resulted in alleviation in germination delay of mutant seeds (Figure 2-03A). This is probably because seed storage protein degradation was slower in the mutant due to lack of proteases and seed growth was slower as a result. To test this hypothesis, total proteins from imbibed seeds were extracted in a time course. By SDS-PAGE and Coomassie Blue staining, seed storage proteins were degraded faster in wild type and the protein levels decreased more slowly in mutant seeds (Figure 2-03B, C). 50 Figure 2-02. ASPAs regulate seed maturation in Arabidopsis. (A) Seed maturation was delayed in size increasing and color change in aspa2 mutant seeds. Representative images show the relative seed size and color of wild type and mutant seeds for 6 DAP 51 (top) and 15 DAP (bottom). DAP: days after pollination. Bar=1mm. (B) Rate of seed size increases over time in wild type and aspa2 mutant. Five individual plants were selected and for each plant at least ten developing seeds were measured at each timepoint for statistical analysis. (C) Representative images of wild type and aspa1-2 aspa2-1 aspa3- 3 seeds. Bar=1mm. Seeds were freshly harvested and put in drying oven at room temperature for at least three days, within a month. (D) Weight of wild type and aspa mutant fully mature seeds. Three biological replicates were represented and for each replicate, 200-300 seeds were weighed. Asterisks indicate statistical sinigicance p<0.05 for Student? t-test. (E) Seed length and (F) seed width of wild type and aspa mutants. N=50 seeds were selected for three replicate each. Asterisks indicate statistical sinigicance p<0.05 for Student? t-test. 52 Figure 2-03. Seed germination and seed storage protein degradation in aspa mutant. (A) Germination rates of Col-0 and aspa1-2/2-1/3-3 seeds germination rate with and without stratification. The delayed germination in aspa mutant seeds could be rescued by stratification for 2 days. For germination experiments, three biological replicates and N=150 fresh mature seeds were selected for each line for each replicate. (B) Coomassie blue staining of seed storage proteins from imbibed aspa1-2/2-1/3-3 and wild type seeds. Total proteins of approximately 20 seeds were loaded for each lane. 53 (C) Relative intensity of staining in B was measured with ImageJ. Time 0 was chosen as baseline and set to 1. In addition to new protein synthesis, cell expansion is also an important aspect in increasing cell volumes during seed germination. A major factor contributing to cell expansion is the central vacuole fusion and expansion. During seed germination, the small protein storage vacuoles fuse with each other and form the large central vacuole. By absorbing lots of water during imbibition, central vacuoles increase in volume, the embryo rapidly expands, the radicle breaks the seed coat and grows into the soil. To test whether storage protein vacuolar fusion is also delayed in the mutants, the embryos were dissected from the imbibed seeds over a time course. The embryo cotyledons were visualized by autofluorescence with confocal microscopy (Figure 2- 05A). The protein storage vacuoles fusion was slower in the mutant cells compared to the wild type. This indicates that ASPAs may also regulate membrane disturbance for vacuolar fusion during seed germination. Total protein was also extracted from imbibed seeds and analyses via SDS-PAGE followed by Coomassie blue staining. The protein gel analysis showed that the protein degradation was slower in the mutant seeds compared to the wild type (Figure 2-03B, C). 54 Figure 2-04. Germination rates in Col-0 and aspa triple mutant seeds with and without gibberellin acid 3, hydrogen peroxide and diphenylene iodonium treatments. (A) Germination rates of Col-0 and aspa mutant seeds with and without gibberellin acid 3 55 (GA3) treatment. Seeds were sown on 1/4MS media supplemented with 1?M GA3. Three biological replicates with N=150 seeds for each line for each replicate. (B) Germination rates of Col-0 and aspa triple mutant seeds with and without hydroden peroxide (H2O2) or dipenylene iodonium treatment. Seeds were sown on 1/4MS media supplemented with 10mM H2O2, 10?M diphenylene iodonium (DPI) and 0.2% DMSO (solvent) for control. 56 Figure 2-05. Protein storage vacuole (PSV) fusion during germination in Col-0 and aspa triple mutant seeds. (A) Morphology of protein storage vacuoles in imbibed wild type and aspa1-2/2-1/3-3 mutant seeds. Autofluorescence in cotyledons was imaged at 57 488 nm excitation. Bar=20?m. (B) Protein storage vacuole sizes in Col-0 and aspa triple mutant seeds over time. N=14-20 for each line for each replicate at each time point. Asterisks indicate statistical significance, p< 0.05, ANOVA followed by Tukey post-hoc test with six replicates. These results suggest that ASPAs are directly involved in seed storage proteins processing, and function downstream of signaling during seed maturation and germination. They take part in metabolism rather than signaling. If this is the case, the mutant seeds should be more sensitive to environment stress that affects metabolism. To test this hypothesis, seeds were treated with hydrogen peroxide to mimic reactive oxygen species stress in overly active metabolic state in the cell, and with NADPH synthase inhibitor diphenyleneiodonium to reduce hydrogen peroxide production, to mimic inhibited metabolic state in the cell. The mutant seeds were more sensitive these treatment as the germination rate was even slower (Figure 2-04B). In summary, ASPAs are involved in seed maturation and seed germination by processing seed storage proteins. The sensitivity to environmental stress may be disadvantage trait in evolution selection. Expression pattern of ASPA2 As previously published, ASPA1 mRNA is detected in all tissues and more abundant in leaves during daytime. ASPA3 is primarily in flowers and ASPA2 is primarily 58 in seeds. (Chen et al., 2002; Sebasti?n et al., 2020). The ASPA3 promoter-reporter constructs showed signals in almost all tissues that undergo programmed cell death (PCD), such as lateral root caps, tracheary elements in proxylem, fading petals, tapetum in stamens and endosperm in developing seeds (Fendrych et al., 014; Olvera- Carrillo et al., 2015). In terms of ASPA2, no reports have been shown about its expression in other tissues except seeds. To test the expression pattern of ASPA2, 2kb promoter was cloned and incorporated into an expression construct with the reporter HISTONE 2A 10 (H2A) fused to a YFP tag. The expression appeared in the suspensor at the globule stage (Figure 2-06B). In developing embryos, ASPA2 was expressed beginning at heart stage (Figure 2-06C) throughout seed maturation (Figure 2-06D, E). It was also expressed in integuments and endosperms (Figure 2-06D). 59 Figure 2-06. ASPA2 expression during seed development. Embryo is depicted with red outlines. ASPA2 promoter:: H2A-YFP reporter lines in Col-0 background were imaged using confocal laser scanning microscopy. (A) Early globular stage. Bar=20?m. (B) 60 Globular stage. Bar=50?m. (C) Early heat stage. Bar=50?m. (D) Torpedo stage. Bar=50?m. (E) Mature stage. Bar=100?m. Embryo is outlined in red. In seedling and vegetative tissues, the expression was detected in almost all tissues: roots, hypocotyls, cotyledons, true leaves (Figure 2-07A to D). In reproductive tissues, signals were detected in stems, sepals, stamens including filaments and the anther epidermis (but not pollen), carpels, stigma, transmission tissues and ovules (Figure 2-07 E to J). In general, ASPA2 was ubiquitously expressed in Arabidopsis plants, similar to ASPA1. This suggests that ASPA2 is likely to be redundant with ASPA1. It also indicates that ASPAs have other functions besides seed maturation and germination. The functions in other tissues remains unclear and need to be explored. 61 Figure 2-07. ASPA2 expression in vegetative and reproductive tissues. ASPA2 promoter:: H2A-YFP reporter lines in Col-0 background were imaged using confocal laser scanning microscopy. (A) Cotyledon. The yellow arrow points to nucleus in the epidermis. Bar=20?m. (B) First pair of true leaves. Bar=20?m. (C) Hypocotyl. Bar=20?m. (D) Root. Bar=20?m. (E) Sepal. Bar=20?m. (F) Petal. Bar=50?m. (G) Anther. Bar=50?m. (H) filament. Bar=20?m. (I) Stigma and carpel. Bar=50?m. (J) Transmission tissue. Bar=50?m. For all images, left: YFP; middle: TML, transmitted light; right: merge. Subcellular localization and trafficking of ASPA2 To further explore the functions of ASPA2 in other tissues, overexpression plants 62 were generated. Coding sequence (CDS) of ASPA1, ASPA2 and ASPA3 were cloned and inserted into plant expression vector driven by 35S promote fused with either the CFP- HA tag or the RFP tag on the C-terminus. The constructs were transformed into wild type background. Full length proteins were detected by Western blotting (Figure 2- 09A). The subcellular localization and trafficking of ASPA2 was analyzed. ASPA2 was trafficked to the vacuoles (Figure 2-08B). ASPA1 and ASPA3 were also trafficked to vacuoles (Figure S07). With brefeldin A (BFA) treatment, a fungal inhibitor which blocks trafficking between endoplasmic reticulum (ER) and Golgi complex, ASPA2 colocalized with trans-Golgi body network (TGN) marker SYP61 (Figure 2-08A) and TGN/early endosome (EE) marker RabA3 (Figure 2-08B). With concanamycin A (conc A) treatment, which inhibits the vacuolar type H-ATPase and further inhibits fusion with vacuoles, ASPA2 colocalized with the multivesicular body (MVB)/prevacuolar compartment (PVC) marker RabF1/ARA6 (Figure 2-08C). These results showed that ASPA2 is first synthesized on ER, transported to TGN and then trafficking to MVB/PVC, finally to the vacuoles. The route passing through the TGN and fuse with EE compartments suggests that proteins on plasma membrane may also contact with ASPA2. 63 Figure 2-08. Intracellular trafficking pathway of ASPA2 in Arabidopsis roots. Colocalization of 35S:: ASPA2-RFP in Col-0 background with TGN marker SYP61-YFP (A), TGN/EE marker RabA3-GFP (B) and MVB marker RabF1/ARA6-GFP (C). Images were taken after one hour incubation with 10?M brefeldin A (BFA) or 100nM concanamycin A (Conc A). ASPA-RFP did not colocalize with RabF1/ARA6-GFP with BFA treatment. Bar=10?m. Pearson?s correlation (r) range is listed in each panel with three seedlings. 64 The saposin-like domain, or plant specific insert (PSI) has been suggested as vacuolar trafficking signal (Kervinen et al.,1999; Terauchi et al., 2006). The PSI deletion version of ASPA2 was generated in this dissertation, which was 35S promoter::ASPA2- 321AA-CFP-HA construct. This deleted ASPA2 contains 1st to 321st amino acid residues. Result showed that there were signals failing to traffic to vacuoles (Figure 2-09B). This indicates the PSI is required in vacuolar targeting. ASPAs are processed to produce mature enzymes. To test whether self-catalytic activities are required for vacuolar trafficking, the first conserved aspartic site was mutated to the alanine in this dissertation, which was the 35S promoter::ASPA2- D107A-CFP-HA construct. This mutation abolishes protease activity, and no self- proteolytic activity occurs. Result showed that ASPA2-D107A-CFP were trafficked to vacuoles. This indicates that proteolytic activity is not required for vacuolar trafficking. This catalytic inactive protease version keeps the intact PSI, and it could be regarded as a PSI overexpression in plants for further analysis. Studies show that conformation change is important for saposin-like proteins interacting with lipids. A novel six-amino acid-motif [N/Q]-[N/Q]-[A/L/I/V]-[K/R]-[N/Q] in helix H3 of the saposin-like domain from the potato aspartic protease StAP appeared to be responsible for interaction with membrane lipids, as a point mutation blocked the conformational change and abolished the membrane fusion ability in vitro (Bryksa et al., 2017). Conformation change has been reported for saposin-like proteins interacting with membrane lipids for human saposin C and D (Rossmann et al., 2008). 65 It could be speculated that this motif affects PSI function in vacuolar targeting. If conformation change is required for vacuolar targeting, a point mutation in this motif may block aspartic protease trafficking to vacuoles. To test whether the abolishment of this motif in saposin-like domain of ASPA2 affects vacuolar targeting, the point mutation was generated in this motif (R402Q) in the ASPA2-D107A context. Which was 35S promoter:: ASPA2 D107A R402Q-CFP-HA in this dissertation. Results showed that both versions showed the vacuolar subcellular localization (Figure 2-09B). This result indicated that this motif is not related to vacuolar targeting of ASPA2. This also suggests that vacuolar targeting role of PSI doesn?t require conformation change in the cell. The modification of PSI was also investigated. There?s only one potential glycosylation site (N404) in ASPA2, which is just after the six-amino acid motif. Glycosylation may also affect the interaction between the motif in PSI and the membrane lipids. To test whether this glycosylation affects ASPA2 vacuolar targeting, the point mutation version 35S promoter:: ASPA2 D107A N404A-CFP-HA was generated and transformed into Arabidopsis. Results showed that ASPA2 D107A N404A-CFP targeted in the vacuoles. This suggests that glycosylation doesn?t affect vacuolar targeting either (Figure S04). 66 Figure 2-09. Glycosylation and vacuolar trafficking of ASPA2-CFP. (A) Western blot analysis of total proteins from 30mg mature leaves for each lane. Samples were treated with/without (+/-) Endo Hf to evaluate glycosylation of ASPA2-CFP: native ASPA2, deletion of PSI and C-terminus version (ASPA2 321AA), single point mutation version in conserved aspartyl site (ASPA2 D107A) and double point mutations in conserved aspartyl site and mutation in lipid binding motif (ASPA2 D107A R402Q). (B) Confocal laser scanning images of the subcellular localization of these ASPA2 mutations in cotyledons of 14-day-old seedlings. Bar=100?m. 10-20 seedlings were tested. ASPAs are involved in root architecture regulation To further elucidate the biological functions of ASPAs in other tissues, the triple mutants and overexpression lines were grown for phenotypic analysis. The primary root of seedlings was slightly shorter in 35S ::ASPA2-RFP overexpression lines than wild 67 type (Figure 2-10A, B). The overexpression level of ASPAs were not verified in this dissertation, but the full-length protein and fuorescent tag was verified (Figure 2-09). While there?s no significant difference in primary root growth length between triple mutant and wild type, more lateral roots formed in the mutants (Figure2-13C). Since ASPAs are likely to involved in seed storage protein processing, this suggests that ASPAs are important in nitrogen metabolism in Arabidopsis. To test whether ASPAs are also involved in nitrogen metabolism in vegetative tissues, seedlings were transferred to low nitrogen media. The primary root length in the mutants was longer than wild type, and the mutant roots were insensitive to the low nitrogen treatment (Figure 2-10B). The overexpression lines did not show significant differences from the wild type (Figure 2-10B). This suggests that ASPAs affect root architecture in Arabidopsis with respect to integration of nutritional signals. 68 Figure 2-10. Root architecture in Col-0 and ASPA mutants. (A)-(B) Primary root growth in response to low nitrogen in wild type, aspa1-2/2-1/3-3 mutant and 35S:: ASPA2-RFP seedlings. Bar=5cm. Seedlings were grown on regular 1/4MS media for 4 days and then 69 transferred to media with low nitrogen (10?M KNO3) or sufficient nitrogen (5mM KNO3) afor additional 4 days. Black dots mark the position of 4DAG seedling root tip at the time of transfer. (B) Statistics of primary root growth length in Col-0, aspa1-2/2-1/3-3 mutant and 35S:: ASPA2-RFP seedlings. Different letters indicate statistical significance, p<0.05 ANOVA followed by Tukey post-hoc test. (C)-(D) Treachery elements maturation in Col-0, aspa1-2/2-1/3-3, 35S::ASPA2-RFP seedlings. (C) Representative image of 7 DAG seedlings stained with 4?M propidium iodide. Yellow arrows indicate the first matured treachery elements. Bar=500?m. (D) Distance between root tip and the first matured treachery elements.N=10-15 seedlings. Different letters indicate statistical significance, p<0.05 ANOVA followed by Tukey post-hoc test. Transcriptional regulation of ASPA2 To test whether ASPA1 and ASPA2 are transcriptionally regulated by environmental signals, the promoters of ASPA1 and ASPA2 were cloned and incorporated in expression constructs with reporter HISTONE 2A 10 (H2A) fused with either mCherry or YFP. The reporter lines were treated with low nitrogen, abscisic acid (ABA) and sodium chloride (NaCl). There are ABA responsive cis elements in ASPA2 promoter, while none is found in ASPA1 promoter. ASPA1 expression was not changed with these treatments (Figure 2-11A, B). ASPA2 expression was slightly higher (approximately 1.5 times higher) with low nitrogen treatment, highly upregulated (approximately 3 times higher) by ABA, and downregulated (approximately 60% lower) 70 by NaCl. This suggests that ASPA2 is responsive to different environmental signals, while ASPA1 functions like a housekeeping gene. However, the primary root length was affected to the same extent with either ABA or NaCl treatment compared to the wild type, except that ASPA1 overexpression plants showed slightly shorter roots with ABA treatment. This might be the artificial effects of ASPA1 overexpression in the plants, or the genes were not highly overexpressed in the plants. 71 Figure 2-11. Transcriptional regulation of ASPA1 and ASPA2 and root growth in responses to ABA and NaCl treatments. (A) Confocal laser microscopy images of ASPA1::H2A-mCherry (Top) and ASPA2:: H2A-YFP (bottom) reporter lines in responses to low nitrogen (10?M KNO3), ABA (2?M) and NaCl (75mM) treatment. Seedlings were growing on 1/4MS media for 5 days and then transferred to new plates containing the 72 corresponding chemicals for another 2 days. 0.2% ethanol (solvent) as the control. Bar=100?m. N=50 nuclei from 5 seedlings in each line, each treatment. (B) Statistics of the fluorescent intensity based on (A). (C) Primary root growth length of Col-0, ASPA1 overexpression lines and ASPA2 overexpression lines with ABA (2?M) and NaCl (75mM) treatments. Seedlings were growing on 1/4MS media for 4 days and then transferred to new plates containing the corresponding chemicals for another 4 days. 0.2% ethanol (solvent) as the control. N=10 seedlings with three replicates. To further examine how root architecture was affected, the position of the first mature tracheary element was measured, since disturbance of xylem maturation affects root growth in plants. 7DAG seedlings were treated with propidium iodide (PI) to visualize the spiral pattern of tracheary elements, and the distance between the first tracheary element and root tip was measured. This distance was slightly longer in triple mutant roots, and slightly shorter in ASPA2 overexpression plant roots (Figure 2- 10C, D). These results suggest that xylem maturation was slightly slower in triple mutant and slightly faster in overexpression plants. When the first conservative aspartic site in the protease was mutated (D107A), the distance was not affected. This indicates that the proteolytic activity is necessary for xylem maturation. Since ASPA3 has been believed in take part in programmed cell death (PCD) in Arabidopsis, although aspa3-3 single mutant doesn?t show PCD related phenotype in lateral root cap cells in the published work (Fendrych et al., 2014), it is likely that all three ASPAs 73 are involved in regulation of PCD of tracheary elements and thus affect root morphology. Autophagy pathway is activated under nutrient deficient condition and autophagy is also involved in programmed cell death (PCD). It is possible that ASPAs are associated with autophagy pathway in response to low nitrogen supply and PCD in TE. To test whether ASPA2 is involved in autophagy pathway, colocalization between autophagy marker ATG8a and ASPA2-RFP was imaged. Colocalization was not found with concanamycin A treatment (Figure S01A). This suggests that ASPA2 is not associated with autophagy pathway. A dual functional endosomal sorting complex required for transport (ESCRT) machinery associated protein FREE1 is reported to regulate both autophagy pathway and MVB formation (Gao et al., 2015). Colocalization results showed partial colocalization between GFP-FREE1 and ASPA2- RFP (Figure S01B). This indicates that ASPA2 colocalized with FREE1 in MVB, not in the autophagy compartments. These results also indicate that the PCD type in which ASPA2 is associated with is apoptosis type rather than autophagy type. ASPAs are involved in programmed cell death To test whether ASPAs might be involved in programmed cell death, the lateral root cap was chosen as a model system to study because it is easier to observe and the whole process can be monitored. The dead cells were identified by propidium iodide staining of the nuclei. Two measurements will indicate the initiation/onset of 74 PCD and the rate of PCD. The distance between root tip and all stained nuclei indicates the onset of PCD. The distance between root tip and all stained nuclei was measured (Figure 2-12 A). There was no difference in this distance distribution between the triple mutant and wild type. This suggests that the onset of PCD was not different between the mutant and wild type. Then the distance between the distal stained nucleus and root tip was measured to determine the rate of PCD. If the cell collapses and peels off, there would be no signal. Since cell division continues at root tip at the same rate, if the cell collapse is slower and delayed in peeling off, then this distance would be longer. The results showed that this distance was longer in triple mutant (Figure 2-12B, C). This suggests that PCD execution process was slower in triple mutant. By monitoring the appearing and disappearing PI signals, the triple mutants showed a longer sustained PI signal over the time period, which suggests that cell death was slower in the triple mutant (Figure 2-12D). The PCD onset in the mutant and wild type were not different but the rate of cell death was different. This indicates that ASPAs may function in degrading cell components rather than signaling transduction during PCD. 75 Figure 2-12. Propidium iodide (PI) staining in lateral root caps of Col-0 and ASPA mutants. (A) Distance between the root tip and all stained nuclei in Col-0 and aspa1- 2/2-1/3-3. 6 days after germination (DAG) Col-0 and aspa1-2/2-1/3-3 seedlings were stained with 4?M PI. N=20-40 nuclei from 5-8 seedlings per line were imaged. P>0.05 by Student? t-test. (B) Representative images of 7 DAG seedlings stained with 4?M 76 propidium iodide. White arrows point to the distal cell stained with 4?M PI. Bar=20?m. (C) The distance between root tip and the distal nucleus stained with PI. 8-12 seedlings per line were imaged and measured. Different letters indicate statistical significance, p<0.05 by Tukey post-hoc test with ANOVA. (D) PI stained nucleus in Col-0, aspa1-2/2- 1/3-3 and 35S::ASPA1-RFP lateral root caps over time. Bar=20?m. 6DAG seedlings were stained with 4?M PI and imaged every five minutes. To further investigate the possible roles of ASPAs in PCD processes, fluorescein diacetate (FDA) was used to stain living cells. The dye emits green fluorescence in the cytosol under neutral pH, and the intensity drops dramatically with decreasing pH. During PCD there is a pH drop in the cytosol, and this is considered as one of the first events of PCD. Then cell membrane permeability increases, and DNA is fragmented, cell components are compartmented afterwards (Fendrych et al., 2014). Therefor PCD could be indicated by the disappearance of FDA signals. Then with the increasing permeability of cell membranes, PI enters cell and stains the nucleus. The time period between the pH decrease and PI staining indicates the rate of disruption of the cell membrane system and increasing cell membrane permeability. This time period was monitored by FDA and PI double staining in a time course (Figure 2-13A). The results showed that this period for loss of FDA signal and increased PI signal was longer in the triple mutant (Figure 2-13B). ASPA2 overexpression lines shortened this time period and catalytic inactive protease overexpression lines did not show change this time 77 period (Figure 2-13B). This means that it takes a longer time for the triple mutant cells to exhibit an increase in cell membrane permeability. This suggests that when these proteases are insufficient, the digestion of membrane components is slower, and the membrane system remains intact and ordered for a longer time in the mutant cells. Reports showed that the lateral root caps serve as an auxin sink. Disturbing PCD in lateral root caps affect the auxin distribution along the roots and thus affect lateral root formation (Xuan et al., 2016). With sufficient nitrogen supply, the triple mutant roots showed more lateral roots (Figure 2-13C). This result indicates that ASPA2 may regulate root morphology through regulating PCD in the root. 78 Figure 2-13. Fluorescent diacetate (FDA) and propidium iodide (PI) double staining in lateral root cap cells in Col and ASPA mutants over time. (A) Representative image of time course of fluorescein diacetate (FDA) and propidium iodide (PI) double staining in 6DAG Col seedling lateral root cap. Bar = 20?m.(B) Time between disappearing of FDA signal and appearing of PI signal in Col-0, aspa1-2/2-1/3-3, 35S::ASPA2-RFP, 35S:: 79 ASPA2 D107A-RFP, 35S:: ASPA2 D107A R402Q-RFP in lateral root caps. N=13-30 cells for each line. Different letters indicate statistical significance, p<0.05 ANOVA followed byTukey post-hoc test. (C) Lateral root number in Col-0, aspa1-2/2-1/3-3 and 35S::ASPA2-RFP under different nitrogen conditions. seedlings were grown on 1/4MS media for 4 days and then transferred to media with low nitrogen (10?M KNO3) or sufficient nitrogen (5mM KNO3) for additional 4 days. Different letters indicate statistical significance, p<0.05 ANOVA followed byTukey post-hoc test. Discussion Aspartic proteases (ASPAs) in seed development and germination Aspartic proteases have long been believed to function in seed maturation due to their high expression levels in seeds in several plant species, such as cardoon and barley (Pereira et al., 2008; Sarkkinen et al., 1992). They have also been found to function in both seed maturation and seed germination processes (Wrobel et al., 1992). However, due to the lack of loss-of-function mutants, the role of aspartic proteases during these two processes in vivo was still unclear. To test this hypothesis, by molecular genetic study by using the Arabidopsis mutants, this dissertation shows that insufficient aspartic protease activity led to delayed seed maturation and delayed germination. ASPA1 was proposed to first function in MVB for seed storage processing at torpedo stage (Otegui et al., 2006), and ASPA1 was frequently chosen as the marker in seed development as well. ASPA1 may have other targets such as other proenzymes, 80 but so far there are no further reports on this. Without enough ASPA proteases, the accumulation of seed storage proteins slows down, and the developing seeds show delay desiccation. The extended time for seed storage protein accumulation compensates for this and thus the total amount of storage proteins is higher. This might be the reason why the mutant seeds were larger and weighed near twice as the wild type. Seed storage protein processing may not be the only function in seed maturation. Some proteases may not be active or not directly involved in proteolytic activity. They may be packaged during seed dormancy, then are activated during seed imbibition and germination. This possibility for the ASPAs could be inferred by the impact on seed germination in the mutant seeds. Though there is indeed ASPA2 expression during germination, the major source of proteases in imbibed seeds are likely to be the ones stored during seed maturation. The reason is that if most proteases were newly synthesized, gibberellin treatment should have rescued the phenotype to some extent as other proteases may compensate for some of the missing ASPAs. The major function of ASPAs during seed germination is likely to be degradation of seed storage proteins for the growing young seedling. On the other hand, the delayed fusion of seed storage vacuoles may suggest its second role during seed germination in membrane disturbance. This comes from the structural aspect that the PSI in aspartic proteases is quite unique from other proteases. Besides the fact that this PSI is cleaved from the protease, there is a 81 hypothesis that this PSI may function as an independent protein in membrane disturbance. The logic supports this hypothesis is that in MVBs, there are smaller compartments surrounded by intact membranes, and therefore the contents inside these compartments are not released for degradation. A protein that disrupts the membrane structure under low pH may help the proteases interact with their targets. And aspartic proteases are good candidates for these functions. The vacuole fusion during seed germination supports the role of membrane disturbance for aspartic proteases. However, this result did not show whether this function comes from the PSI or the proteolytic domains. Thus, ASPAs are important for plant seed development and germination. Delayed seed development and germination are not advantageous traits in evolution, because the plants may not respond to the proper time for seed maturation and seedling growth. ASPAs function in tissues other than seeds Though ASPAs were first found in seeds, ASPA1 and ASPA2 were expressed almost throughout the whole plant, but their functions in these tissues haven?t been reported yet. The role in seed maturation and germination suggests the hypothesis that ASPAs process the bulk of the targets and regulate nitrogen supplies by proteolytic processing of aged or broken proteins in the cell. The reduced response to low nitrogen in both mutant seedlings and overexpression plants support this hypothesis. In both seed 82 maturation and low nitrogen conditions, ABA is an important signaling component. The transcriptional results show that ASPA2 expression was slightly upregulated to low nitrogen, and highly upregulated by ABA. ASPA2 promoter region contains ABI binding elements. In contrast, ASPA1 do not have these elements in the promoter, and the expression level remained at a constant level. This is consistent with the reported results that ASPA1 expression is relative stable (Endo et al., 2014). The differences between ASPA1 and ASPA2 indicate that ASPA1 is more likely a housekeeping gene and ASPA2 is responsive to environment stresses. Another ABA regulated physiological process is stomata opening and drought tolerance. ASPA1 overexpression has been reported to enhance drought tolerance in Arabidopsis by regulate stomata opening (Sebasti?n et al, 2020). However, this is more likely to mimic another aspartic protease ASPG1 (ASPARTIC PROTEASE IN GUARD CELL1) function in guard cells which does not contain a saposin-like domain (Yao et al., 2012). The plant specific insert is important for the aspartic protease vacuolar targeting, and this is also the case for ASPA2. The vacuolar targeting of the catalytic inactive form of ASPA2 suggests that the self-catalytic activity is not required for vacuolar targeting either. The processing of ASPA2 is likely to occur via other proteases. The primary function of PSI is interacting with lipids, and PSI from potato StAP shows the lipid interaction activity in vitro. The newly identified six amino acids motif in potato StAP PSI shows its important role in conformation change. Abolishment of this motif blocks the conformation change and blocks interactions with lipids. As a result, it was 83 hypothesized that mutation in this motif in ASPA2 PSI would also block conformation change, and thus block its function in the vacuolar targeting. However, the results showed that the mutated ASPA2 was still trafficked to the vacuole. This result suggests that this motif in PSI is not responsible for vacuolar targeting. It could be possible that the PSI interaction with lipid membranes doesn?t require conformational change in Arabidopsis for vacuolar targeting, or there is another novel mechanism for PSI function in vivo. The catalytically inactive form of ASPA2 (ASPA2-D107A) could be regarded as a ?native? version of PSI. In this dissertation, no significant phenotype was observed in this mutated ASPA2 overexpression lines. One possible reason is that PSI is not directly involved in the normal plant growth. Potato PSI form StAP show anti- bacteria activity (Mu?oz et al., 2010; Fery et al., 2018). Overexpression potato PSI in Arabidopsis led to enhanced resistance to Botrytis cinereal, and the plants were taller than wild type (Frey et al., 2018). Overexpression of the catalytically inactive ASPA2 D107A in Arabidopsis, the plants did not show a higher height in this dissertation. This anti-bacterial activity for PSI requires that the PSI is secreted to the extracellular space. However, no reports have been shown that PSI is able to traffic to the extracellular space. The artificial recombinant PSI constructs show that PSI still traffics to the vacuole (Vieira et al., 2019). However, it is possible that under certain circumstances, PSI is secreted in plant defense responses, which needs further studies. In terms of plant defense, another interesting thing is, the mature ASPAs (which does not have PSI) show structural similarity with the anti-pathogen protein xylanase 84 inhibitor 1 from wheat (Fierens et al., 2003; Sansen et al., 2004). Xylanase is synthesized and secreted from the pathogens and digest plant cell walls to attack the plants. Xylanase inhibitor 1 binds and inactivate xylanases (Fierens et al., 2003). Xylanase inhibitor 1 lacks essential catalytical residuals and is proteolytically nonfunctional (Sansen et al., 2004). This provides another possible function of ASPAs in plant defense. If ASPAs are secreted to the extracellular space, they might have the ability to inactive xylanases, and this function is independent of PSI. It is interesting that a single peptide encodes two independent functional units functioning in plant defense. Further studies will explore whether the ASPAs are able to traffic to extracellular space in plant defense responses. ASPAs in programmed cell death (PCD) The properties of ASPAs in vitro is well-studied. However, from these studies, these proteins seem to be simply a tool for proteolytic activity without any specificity. If this is the case, it seems that PSI is only necessary for vacuolar targeting. However, the PSI structural feature is conservative in plants, which presupposes to a function that is important yet known in plant growth and development. The ASPA3 expression pattern seems to provide hints on their functions. The roles of ASPA3 in PCD has long been proposed due to its restricted expression pattern. But the lack of PCD-related phenotypes in single knockout mutant makes the exact role remained unclear. This may partially result from the redundancy of ASPA1 and ASPA2 in those tissues. This 85 leads to another hypothesis that ASPAs function in regulating PCD in Arabidopsis. In PCD cells, a set of proteins are usually co-expressed such as CEP1, ASPA3 and BFN1. These proteins function in the last stages of PCD for nutrient recycling and cell components disruption. While most upstream transcriptional factors show tissue specific expression pattern, the set of CEP1, ASPA3 and BFN1 is expressed in almost all the PCD tissues. But the expression time is different for these genes. In Arabidopsis stigmas, the expression order is CEP1 first, then ASPA3, and BFN1 is the last (Gao et al., 2018). CEP1 functions in the cytosol, and it is likely to participate in the signaling transduction. BFN1 functions in the nuclei for the final degradation of DNA. ASPA3 may be the primary protease that is involved in bulk proteolytic activity of proteins for recycling nitrogen for other tissues. The membrane disturbance ability is also an advantage during this process. The results here showed that insufficient ASPAs reduced the rate of membrane permeability increase in lateral root cap, and delayed xylem maturation in the root. A delay in PCD processes may impact on plant growth and development, such as root architecture since PCD occurs in root caps, tracheary elements and the base of lateral roots. For example, lateral root caps are believed to be an auxin sink and the peeling off affects the release of auxin. Therefore, the position of lateral root cap PCD affects the auxin distribution in the root tip, and thus affect the auxin distribution along the root (Xuan et al., 2016). The distribution of auxin, or the maximum auxin sites along the root, is associated with lateral root formation (Wei et al., 2016). As a result, PCD is associated with root architecture in Arabidopsis. ASPAs 86 regulate the rate of PCD in lateral root caps and affect root architecture. Conclusion In summary, the biological functions of ASPAs might be bulk proteolytic activity in vacuoles for nitrogen recycling. The membrane disturbance activity promotes interaction between proteases and substrates. ASPA1 appears to be a housekeeping protease and ASPA2 functions in response to environmental stresses. These two proteases are expressed in most plant tissues. ASPA3 functions in PCD tissues as an additional contributor to the degradation events. ASPAs are involved in seed development and germination, as well as programmed cell death in the plants. The independent function of PSI was not found in this dissertation. There are still some unresolved questions. First, the triple mutants do not show a dramatic phenotype in older seedlings or adult plants (Figure S2 and S3). This may be due to compensation from a low level of ASPA1 activity in the triple mutants. As ASPA1 expression is normally relatively high throughout the plants. The remaining protease activity may still be enough for the basic metabolic requirements. As a result, the aspa1 knockout mutant is preferred for studying the biological functions of ASPAs in vivo. Generating the aspa1 knockout mutant by CRISPR is a good choice. One of the future directions is to create a knockout triple mutant for phenotypic studies and compensating this knockout mutant to determine the function of these proteins in vivo. 87 Second, the biological role of PSI other than vacuolar targeting remains unclear. Overexpression of the catalytic inactive ASPA2 D107A in Arabidopsis did not affect plant growth. The root growth and PCD in lateral root caps were not affected either in ASPA2 D107A overexpression lines. These results suggest that PSI does not function independently from the proteolytic domain. PSI overexpression does not promote PCD in lateral root caps (Figure 11) or enhance plant growth like StAP PSI (Figure S2 and S3). The major function of PSI is associating the protease domain with membranes, bringing it to vacuoles. Glycosylation may be a signal for ASPA transport to TGN so that vesicles containing ASPAs could fuse with early endosomes for plasma membrane protein digestion. One of the future aspects is to find whether ASPAs or PSIs are secreted to the extracellular space. This will provide information on whether they may be involved in plant defense response. Third, the role of ASPAs in programmed cell death needs further studies. The difference between wild type and the triple mutant was subtle, and the only phenotypes found were in tracheary element maturation and lateral root cap turnover. In other PCD tissues such as the tapetum, PCD related phenotypes were not detected. The knockout triple mutants will also help with this question. Another direction is to further mutate co-expressed genes such as CEP1 and BFN1 and explore how these genes affect PCD in combination. This dissertation demonstrated the role of ASPAs processing seed storage proteins in seed germination in vivo for the first time. And this dissertation also 88 provides evidence of the involvement of ASPAs in programmed cell death by promoting membrane permeability. These results broaden the knowledge of the multiple roles of aspartic proteases in plant growth and development. Materials and Methods Plant materials All the Arabidopsis thaliana plants are in the Columbia-0 (Col-0) ecotype genetic background. T-DNA insertional mutants (aspa2-1 SALK097505; aspa2-2 SALK021601; aspa1-1 SALK092586; aspa1-2 SALK041027; aspa3-3 SALK056711) were sourced from The Arabidopsis Biological Resource Center, The Ohio State University (ABRC; www.abrc.osu.edu). T-DNA insertions were confirmed by PCR with the primer in T-DNA sequence (LBb1.3) and the primer in the flanking gene regions (primer RP). The primers used in genotyping are listed in Table S01. The expression level of each ASPA was detected by real-time PCR and the primers for real-time PCR are listed in Table S01. Details on methods for genotyping and real-time PCR are described in appendix D. For germination on solid media, Arabidopsis seeds were surface sterilized by soaking in 20% bleach (containing sodium hypochlorite) for 15 minutes with agitation. Seeds were then rinsed 3-5 times in sterile water. Seeds were sown on 1/4 Murashige and Skoog (MS) medium (RPI Corp.) media containing 0.5% sucrose with 0.8% agar. Seeds were stratified at 4?C for 2 days in the dark and then placed in growth chamber at 22?C, with 24 hr continuous white light at 100 ?mol m-2 s -1. For chemical treatment, 89 seedlings were first grown for 4 days on regular 1/4MS media, then transferred to new media containing the corresponding chemicals. The working concentrations of chemicals used in this research were: 1?M Gibberellin acid; 10mM hydrogen peroxide (H2O2); 10?M diphenyleneiodonium (DPI); 10?M brefeldin A (BFA) ; 100nM concanamycin A (conc A); 4?M propidium iodide (PI); 5?g/ml fluorescein diacetate (FDA); 2?M abscisic acid (ABA); 75mM sodium chloride (NaCl). Solvent (ethanol or DMSO) was added as the control, and the concentration was the same with the corresponding chemical concentration in each experiment. Low nitrogen media was prepared by adding 10?M potassium nitrate (KNO3) in 1/4MS without nitrogen (MS w/o nitrogen) media. Sufficient nitrogen media was prepared by adding 5mM KNO3 in 1/4MS w/o nitrogen media. Adult plants were grown in growth chamber at 22?C with 16 hr light and 8 hr darkness cycles. Light intensity was 100 ?mol m-2 s -1 with a mixture of fluorescent and incandescent bulbs. Relative humidity was 50%. For seed weight measurement, seeds were harvested and stored in drying chamber containing drierite for at least three days. Germination test Seeds were harvested and stored in drying chamber for at least three days. Only seeds harvested within a month were used. Each time, 3 biological replicates were measured. Each replicate contained around 150 seeds. Seeds were sterilized and sowed as mentioned above. Seeds were placed at 22?C, with 24 hr continuous light at 90 60 ?mol m-2 s -1. Germinated seeds were counted every 12 hours. Germination was defined as radicle emergence. Seed Protein extraction For each time point, 200 Arabidopsis seeds were accurately counted and imbibed in sterilized water. Seeds were ground with a grind stick in Eppendorf tubes on ice. The ground seeds were immediately resuspended in 100 ?L protein extraction buffer (50 mM sodium citrate, pH 5.5; 5% SDS (w/v); 0.01% BSA (w/v); 150 mM NaCl; 2% (v/v) ?- mercaptoethanol and 1 ?L of protease inhibitor cocktail (Genesee Scientific). The mixture was incubated for 60 minutes at 100? C. Samples were centrifuged at 4? C, 14,000g for 30 minutes and the supernatant was collected. The samples were stored in -80? C if not used immediately. SDS-PAGE Total proteins were separated by SDS polyacrylamide gel electrophoresis (SDS- PAGE). 10?L samples were prepared by adding 2?L of 6X SDS (sodium dodecyl sulfate) loading buffer (1.2g SDS, 0.01% bromophenol blue, 4.7ml glycerol, 1.2ml Tris 0.5M pH=6.8, 2.1ml ddH2O). Samples were loaded onto 12% polyacrylamide 0.75mm 10- well or 15-well gel (Bio-Rad?). Precision Plus Protein Dual Color Standards (Bio Rad) was used as marker size. Electrophoresis was carried out in 1X Running Buffer (3g of Tris base, 14.4g of glycine, and 1g of SDS in 1000 ml water) at 120V for approximately 4 hours or until the dye front reached the front of the gel. Seed proteins were 91 visualized by staining with Coomassie Blue. Coomassie blue staining For visualization of seed storage proteins, the gel was stained by incubating overnight in 20ml Coomassie staining solution (0.1% Coomassie bright blue in 50% methanol, 10% acetic acid). The gel was de-stained for 3 hours with de-staining solution (10% acetic acid, 50% methanol) with at least two changes of this solution until the background was nearly clear. Glycosylation test 300mg Arabidopsis leaves were harvested and then ground with a grind stick in the Eppendorf tube with liquid nitrogen. The ground tissues were resuspended in 300 ?L protein extraction buffer, incubated for 60 minutes at 100? C, centrifuged at 4? C, 14,500g for 30 minutes and the supernatant was collected. Glycosylation was detected by Endo Hf (New England BioLabs) digestion according to the manufacturer?s instruction. Briefly, 17?L of the extracted protein sample was added with 2?L 10xGlycoBuffer 3 and 1?L Endo Hf. The sample was incubated at 37?C for 1 hour. Then the sample was analyzed by SDS-PAGE and Western blot. Western blot For immunoblotting, proteins were transferred to polyvinylidene difluoride (PVDF) 92 membrane in Tris-glycine-methanol transfer buffer (2.9g glycine, 5.8g Tris, 0.37g SDS 100mL methanol, 900mL water) at 120V for 80 minutes at 4?C and then rinsed briefly in 1xPBS. Membranes were blocked overnight at 4?C in blocking buffer (5% non-fat milk in 1xPBS with 0.02% Tween20) or 1.5 hours at room temperature. The membrane was rinsed gently with washing buffer (1% non-fat milk in 1x PBS with 0.02% Tween20) for three times, 15 minutes each. The membrane then was incubated with primary antibody (anti-HA) in blocking buffer overnight at 4?C or 1.5 hours at room temperature. The membrane was rinsed with washing buffer for three times and each time for 15 minutes. Then the membrane was incubated with secondary antibody (anti-rabbit digoxigenin) at room temperature for 1.5 hours. The membrane was rinsed with washing buffer for three times and each time for 15 minutes. Proteins were visualized using a SuperSignal West Femto Kit (Thermo Scientific). Images were taken by C-DiGit Blot Scanner (LI-COR). Cloning and expression vector construction ASPA1, ASPA2 and ASPA3 coding sequences (CDS) were cloned without the stop codon from seedling cDNA. Fragments were inserted into pDONR/Zeo by BP reaction. The entry clones were comfirmed by sequencing. Primers for cloning, site-direct mutagenesis and sequencing are listed in Table S01. The entry clones were incorporated into pH7RGW2 (35S promoter, RFP tag fused on C-terminus) and pEarleyGate102 (35S promoter, CFP and HA tag fused on C-terminus) by LR reactions. 93 The expression contructs were confirmed by sequencing and the corresponding primers are listed in Table S01. Transgenic plants were created by floral dipping method (Clough and Bent,1998) with agrobacterium strain GV3101. For promoter fusions with the histone tag, first the promoters of ASPA1 (1.9kb), ASPA2 (1.9kb) and ASPA3 (2.0kb) were cloned and inserted into pGEM-T-Easy vector. Entry clones were confirmed by sequencing. Primers for cloning and sequencing are listed in Table S01. Gateway vector pUBC::RFP-Dest (Grefen et al., 2010) was digested by SpeI and PsiI and ligated with mCherry. Then the modified pBUC::mCherry vector was digested by SacI and PspXI and ligated with ASPA1 promoter. The gateway vector pUBC::YFP-Dest was digested by SacI and PspXI and ligated with ASPA2 promoter. The gateway vector pUBN::YFP-Dest was digested by SacI and PspXI and ligated with ASPA3 promoter. Histone 2A 10 CDS was cloned from seedling cDNA with and without the stop codon and inserted into pDONR/Zeo vector. The final constructs ASPA1 promoter::H2A-mCherry and ASPA2 promoter:: H2A-YFP were created by LR reaction. The expression contructs were confirmed by sequencing and the corresponding primers are listed in Table S01. Details on methods of molecular cloning are described in appendix D. Microscopy Confocal microscopy was carried out using a Zeiss LSM 710 Confocal laser scanning microscope (Carl Zeiss, Germany) with Axio Imager 2. Pixel dwell time was 94 0.01 ms. The master gain was always set to less than 893, with a digital gain of 1.5. For RFP/mCherry acquisition: 594 nm (5%) excitation and 588-696 nm emission. For YFP acquisition: 514 nm (5%) excitation and 519-560 nm emission. For GFP: 488 nm (5%) excitation and 493-598 nm emission. For CFP: 458 nm (5%) excitation and 453-580 nm emission. For PI: 543 nm (5%) excitation and 583-718 nm emission. For FDA: 488 nm (5%) excitation and 493-583 nm emission. Quantification of florescence intensity was analyzed using ZEN Lite 2012. All images were processed with ZEN Lite 2012 (Zeiss) and ImageJ. Time-course image of PI (propidium iodide) and PI/FDA (fluorescein diacetate) double staining in lateral root cap To keep the roots alive for a long time while imaging, seedlings were imaged in the 35 mm petri dish with high precision 1.5 coverslip on the bottom (14 mm glass diameter, MatTek). Seedlings were placed with 1/4MS (0.5% sucrose, 1% agar) containing PI only (4?M) or PI (4?M) and FDA (5?g/ml). Images were taken by every minute or every five minutes as noted. All images were processed with ZEN Lite 2012 (Zeiss) and ImageJ. 95 Chapter 3: Elucidating features and functions of plant prosaposin-like proteins (PSAPLIPs) Abstract Saposin-like proteins have been well studied in animals. In plants, the only reported saposin-like structure is the plant specific insert in some aspartic proteases. Another type of saposin-like (SapB-like) proteins, the prosaposin-like proteins (PSAPLIP) have been paid less attention. These proteins are ubiquitously present across the plant kingdom from green algae to flowering plants, indicating their importance in plant growth and development. Here alignment and comparison of protein sequences among different species revealed the high similarity between plant prosaposin-like proteins and human prosaposin in primary and secondary structure. Unique features of prosaposin-like proteins in plants were also identified. PSAPLIPs contain two SapB- like domains in angiosperms, while in gymnosperm, moss, liverwort and green algae, most PSAPLIPs contain three SapB-like domains. In most species, there are 1-4 PSAPLIPs encoded genes in their genomes, and Arabidopsis thaliana has two PSAPLIPs, AtPSAPLIP1 (At3g51730) and AtPSAPLIP2 (At5g01800). Both AtPSAPLIP1 and AtPSAPLIP2 were targeted to vacuoles and both proteins were sensitive to concanamycin A treatment. However, AtPSAPLIP1 was sensitive to brefeldin A treatment while AtPSAPLIP2 was not. The promoter reporter activity results showed 96 that AtPSAPLIP1 was primarily expressed in sepals and pollen grains, while AtPSAPLIP2 was expressed in petals and young anthers. These results suggest the important role of prosaposin-like proteins in reproductive organ development, especially in male gametophyte development. They may facilitate degradation of target signaling proteins in the cell. This dissertation characterized the plant prosaposin-like proteins for the first time and provided insights on a new class of proteins regulating male gametophyte development in plant reproductive process. Introduction Saposin-like proteins (SAPLIP) are named after saposins, which contain four small proteins derived from one single precursor called prosaposin. Saposins are important in cellular metabolism as cofactors in sphingolipid catabolism (Bruhn, 2005). SAPLIPs are found throughout eukaryotes from amoebozoans to mammals. SAPLIPs exhibit low sequence similarities among different species, but they are conserved in the six conserved cysteines and several conserved hydrophobic and polar charged residues which enable the protein folding into the conformation to interact with lipids (Bruhn, 2005). In animals, SAPLIPs are found to participate in a variety of different pathways, such as co-factors of lipid-degrading enzymes (Kishimoto et al., 1992; Schuette et al., 2001), surface tension regulator (surfactant protein B) (Cochrane et al., 1991), antimicrobial effector (Pena et al., 1997). Some SAPLIPs activities are independent of lipid interactions, such as J3 crystallin found in jellyfish Tripedalia cystophora 97 (Piatigorsky et al., 1997). Loss of SAPLIP function in mammals is associated with diseases states, such as deficiency in saposin C leading to Gaucher disease which is a type of lysosomal storage disorder (Tamargo et al., 2012), and deficiency in saposin A leading to Krabbe disease which is a disorder that the protective coat in nerve system is defective (Spiegel et al., 2005). Structural similarity is high among SAPLIPs. There two types of conformation reported, and they show slightly different. NK-lysin is one type and the SAPLIP domain contains 5 helices fold into two halves. The first half consists of helices 4 and 5 packed perpendicularly against helix 1. The other half contains helix 2 and 3 (Liepinsh et al. 1997). Saposin B is representative of other types of SAPLIPs. The two halves of saposin B crystallizes as a dimer into a shell shape. The saposin B monomer has four helices and shows an open formation in a V shape. This has been proposed as the lipid binding position (Ahn et al., 2003). Both types of conformation are in favor of lipid interaction. The soluble, monomeric form of SAPLIP holds a closed conformation with the hydrophobic surface hidden in the cavity. Charged residues mediated the initial contact with the negatively charged lipid membrane surface by electrostatic interactions. Then the protein change into open conformation, probably associated with dimerization or oligomerization. The membrane-embedded oligomer is hypothesized to form a pore in the membrane allowing presentation to the hydrolytic enzymes (Rossmann et al., 2008; Olmeda et al., 2012). And in the end, either the lipids are extracted or two adjacent memrbanes fused 98 by saposin-like proteins. Animal SAPLIPs have been extensively studied on the structures and functions (Azuma et al., 1994; Ciaffoni et al., 2001; Ahn et al., 2003; De Alba et al., 2003; Kang et al., 2004; Hawkins et al., 2005; Hill et al., 2006; Popovic et al., 2008; Olmeda et al., 2013). However, plant SAPLIPs are not studies as extensively as animals. Plant SAPLIPs are generally referred as a domain called plant specific inserts (PSI) in aspartic proteases (Brodelius et al., 2005). In general, plant PSIs are similar with human saposins in terms of sequence features and the overall structure. The PSI from phytepsin in barley shows highly structural similarity with NK-lysin (Kervinen et al., 1999). PSI from cardoon also shows high similarity to human saposins C and it is able to activate human glucosylceramidase in vitro (Brodelius et al., 2005). PSI of StAP in potato is able to induce vesicle disruption in vitro, similar with human saposin C and the secondary conformation is pH-dependent which is similar to human saposins (Bryksa et al., 2011). Sequences are highly similar and conserved among plant PSIs (Bryksa et al., 2017). They all exhibit leakage activity in bilayer composed of a vacuole-like phospholipid mixture and membrane fusion activity in vitro. This activity is pH-dependent and the optimal pH is 4.5 and requires the presence of acidic phospholipids such as phosphatidylserine. Low pH results in dimerization of potato PSI, and the monomer is prevalent under neutral pH. All these behaviours are similar to mammalian saposins Athough there are a lot of similarities between plant PSI and mammalian saposin- 99 like proteins, there are some features unique to plants. A recent study showed that conformation change is the molecular basis of bilayer membrane leakage at low pH. A novel six-residue motif in H3 helix ([N/Q]-[N/Q]-[N/Q]-[A/L/I/V]-[K/R]-[N/Q]) was identified which accounts for this configuration change. A point mutation K83Q in this motif in helix H3 blocks the response to low pH activation with respect to conformation change (Bryksa et al., 2017). This motif may be responsible for lipid-interactions as this motif is also found in several other membrane-interacting proteins in different plant species (Bryksa et al., 2017). But This motif is not seen in human and other mammalian saposins. Another difference between PSI and mammalian saposin is that the orientation of helices is switched from N terminus to C terminus (Bliven et al., 2012). The overall configuration of the secondary structure is not affected. This could be evolved from gene duplication and subsequent deletion event in plant evolution history. The third difference is that the PSI in aspartic proteases is cleaved off to produce the mature protease, while in mammalians, saposin-like domains are still linked to the mature proteins (Bruhn, 2005). PSI is important for the aspartic protease vacuolar targeting in vivo (Kervinen et al., 1999; Terauchi et al., 2006). Overexpression of PSI from the potato aspartic protease in Arabidopsis enhances the plant resistance against Botrytis cinerea (Frey et al., 2018). However, the independent function of PSI in vivo still lacks experimental supports. This leads to one hypothesis: there are other SAPLIPs in the plants which are not characterized yet. 100 With information from Uniprot and other protein databases, there are a group of uncharacterized saposin-like domain containing proteins in plant genomes. Almost no reports studied on the structural features or functional analysis of these proteins in the plants, so little information is available. In this dissertation, the primary and secondary structures of these proteins from across the plant kingdom were predicted and analyzed to understand the distribution and diversity of these proteins. By sequence screening with the keyword search in Uniprot, sequence alignments and comparisons, this group of proteins showed a high similarity with the human prosaposin. As a result, they are named prosaposin-like proteins (PSAPLIPs) in this dissertation. This analysis showed that PSAPLIPs are found in all plant phyla and the number of genes varies. With structural prediction in Phyre2, saposin-like domain from Arabidopsis PSAPLIPs showed high similarity to human saposins. Arabidopsis AtPSAPLIP1 and AtPSAPLIP2 were analyzed spatiotemporal expression and subcellular targeting. The results suggest that PSAPLIPs are important in male gametophyte development in Arabidopsis, possibly by facilitating target signaling proteins trafficking to the vacuole for degradation. Results Phylogenic studies of PSAPLIPs in plants More than 160 PSAPLIP genes from 67 species have been identified in higher plants via pairwise ortholog predictions (EggNOG, eggnogdb.embl.de). In Uniprot 101 database, more than 2000 proteins are annotated as containing the saposin-like domains. Some of them are aspartic proteases, which contains the saposin-like domain called the plant specific insert (PSI). The remaining 459 proteins were uncharacterized in 152 plant species from green algae to flowering plants. In angiosperms, there are 417 sequences annotated as containing at least one saposin B like domain (SapB-like domain). These sequences can be divided into two groups depending on their predicted sequence structures. The first one contains the N-terminal signal peptide and SapB-like domain of 80-82 amino acid residues, which is the typical length of a saposin-like protein. The second group contains an N-terminal signal peptide, an annotated saposin-like domain around 130 amino acid residues, following by disordered regions in C terminus, usually with a polyampholyte region. From amino acid alignments, these sequences have clearly diverged from the other SapB-like domain containing proteins. The sequence features of typical SAPLIPs, such as the distribution of the six conserved cysteines, and conserved hydrophobic and polar residues, were not found in this group of proteins. This second group is more similar with human nucleophosmin than to saposins and this group is classified as nucleophosmin family by gene ontogeny (PANTHER, pantherdb.org). The reason that members of this group were predicted to be saposin-like proteins is likely due to the prediction algorithm is based on the six conserved cysteines in saposin-like proteins, and plant nucleophosmins happen to contain several cysteines in their sequences. As a result, these sequences are likely incorrectly auto-predicted by Uniprot database, 102 and this group should be considered and annotated as nucleophosmin-like proteins rather than saposin-like proteins. Among the 417 uncharacterized angiosperm sequences, 73 belongs to nucleophosmin family. The remaining 344 should be considered the PSAPLIPs in plants. Then the sequences annotated as incomplete were also excluded. The remaining sequences were used for the following analyses. Typical PSAPLIPs contain an N-terminal signal peptide and two saposin-B like domains. PSAPLIPs are predicted to be in the vacuole. However, possibility of secretion to extracellular space or other compartments could not be excluded. The Arabidopsis AtPSAPLIP1 is annotated in both the cytosol and the vacuole. Several proteins have no N-terminal signal peptide (SP) prediction. In most sequences, the signal peptide is around 18-35 amino acid residues. In those gene apparently missing SP sequences, there are indeed sequences long enough to be a signal peptide. Novel types of signal peptide may exist in these species. Only six sequences show an N-terminal domain less than 10 amino acid residues (Ananas comosus ACMD2_06262; Arundo donax no gene ID, protein ID A0A0A9V254; Dichanthelium oligosanthes BAE44_0009052; Panicum miliaceum C2845_PM06G26640; Helianthus annuus HannXRQ_Chr10g0286291; Dorcoceras hygrometricum F511_29468) (Figure S08). However, these annotations were auto- predicted from the genomic DNA sequence and It is possible that the start codon was predicted incorrectly. If these genes indeed lack a signal peptide, this suggests that these PSAPLIPs may either function in other cellular compartments or traffic to vacuole 103 by other facilitating proteins. Human prosaposin is processed into four mature saposins, and human saposin B remains dimerizes under most conditions (Hiraiwa et al., 1993; Kishimoto et al., 1992; Leonova et al., 1996). It can be inferred that Sap-B like domains may also function as dimers, and this may be the reason that most plant PSAPLIPs contain two Sap-B like domains. In some genes, only one Sap-B domain is found (36 sequences among all 344 angiosperm PSAPLIPs), but it is more likely a prediction error because all these predictions are from genomic DNA sequence. Some sequences appear to lack the N- terminal part of the SapB-like domains and some seem to lack C-terminal part. This suggests that they may be fragments but incorrectly annotated as complete ones, or they may be complete sequences. This would need to be confirmed in these species by further studies. Overall, most identified angiosperm PSAPLIPs contains two SapB- like domains, and it can be concluded that this is the prevalent form in angiosperms (Figure 3-02). In green algae, liverworts, mosses and gymnosperms, the copy number of SapB- like domain varies. The gymnosperm, Picea sitchensis contains a protein with two SapB-like domains (Uniprot protein ID A9P228) and a protein with three SapB-like domains (Uniprot protein ID A9P283). While PSAPLIPs from Araucaria cunninghamii (Uniprot protein ID AOAOD6R2G) and Wollemia nobilis (Uniprot protein ID A0AOC9RXJ5) contain three copies of Sap-B like domains in PSAPLIPs and no PSAPLIP was identified with two SapB-like domains in these two species. Due to the limited 104 data for ferns, no PSAPLIPs were found in ferns. In liverworts and mosses, only three PSAPLIPs were found and they contain three SapB like domains (Chara braunii CBR_g3540; Physcomitrella patens subsp. patens PHYPA_022478; Physcomitrella patens subsp. patens PHYPA_018982). In green algae PSAPLIPs, the number of SapB-like domains varies from one to three. However, most sequences are derived from whole genome shotgun (WGS) entries (an EMBL/GenBank/DDBJ), therefore this is an initial analysis based on preliminary data. Three SapB-like domains still appear to be the major form of PSAPLIPs in green algae (Figure S09). The trend of evolution of plant PSAPLIPs can be depicted as following: in green algae, genes with three SapB-like domains is the prevalent form. In liverworts and mosses, only genes with three SapB-like domains are found. The supports the hypothesis that land plants evolved from single origin that contained only this type of PSAPLIP genes. In gymnosperms Picea sitchensis, genes with two and three SapB-like domains are found. This suggests that genes with two SapB-like domains first evolved in gymnosperms, and probably evolved from deletion of one of the SapB-like domains after a gene duplication event. In angiosperms, no PSAPLIP genes were found with three SapB-like domains. PSAPLIPs with two SapB-like domains are found in all angiosperms reported in Uniprot. This supported the single origin for all angiosperms. PSAPLIPs with only one SapB-like domain may exist in some angiosperm species, like the grape, but this hypothesis is not supported by current information. Sequence data from more species, and more experimental sequence data are needed to obtain a 105 better understanding of the sequence features of PSAPLIPs and the evolution of PSAPLIP family in plants. With the data currently available, only one or two PSAPLIPs genes occur in the genome in most species. However, some species may contain more copies, but all are less than 10 copies, such as rice and bananas. This may result from genome duplication during evolution or artificial selection in agriculture. And in terms of unique sequences, in most plant species, there are one to three PSAPLIPs genes across the genomes. This suggests that this family does not expand during evolution, but still persists in the genome, which indicates that this family is important in plant growth and development. The redundant alleles may disappear in natural selections. Structural features of AtPSAPLIPs Like PSAPLIPs in animals, plant PSAPLIPs also exhibit highly variable protein sequences, but the conservative cysteines remain the same and the distribution of cysteines and polar residues are well aligned with mammalian saposins. The high divergence can be visualized in the phylogenetic tree (Figure S15). Although most proteins are clustered into the major plant groups, many protein positions in the tree do not correspond to phylogenetic relationships between or among different plant groups. However, the aligned conservative cysteines and hydrophobic sites indicate that these SapB-like domains evolved from a single ancestor rather than independently. In most PSAPLIPs, all six cysteines can be found, and the distribution is conserved 106 (Figure 3-03). There are some sequences with only five conserved cysteines, but this is less likely to affect the overall structure. Cysteines are important for disulfide bonds, but they are not likely to determine the folding of the protein. The impact on mutation in conserved cysteines in different SAPLIPs is still unknown. As a result, the secondary structure can be predicted for these SapB-like domains. With the sequence information, structures of Arabidopsis PSAPLIPs were constructed in Phyre2 and visualized in EzMol (Figure 3-01 and Figure 3-02). The overall secondary structure is highly similar with mammalian saposins and plant specific insert (PSI) in aspartic proteases. From the study of mammalian saposins, the primary function of these cysteines is forming disulfide bonds, which provide extra stability of protein configuration. This can be seen in the predicted structure of AtPSAPLIP1 (At3g51730) and AtPSAPLIP2 (At5g01800). Three pairs of conservative cysteines are shown in a direction advantaged for forming disulfide bonds (Figure 3- 02). Overall, the highly divergent primary sequence and highly similar secondary structure is a feature among all saposin-like proteins in eukaryotes. 107 Figure 3-01. Predicted structure of AtPSAPLIP1 and AtPSAPLIP2. (A) Predicted secondary structure of AtPSAPLIP1 and AtPSAPLIP2. Alignment for the first saposin B (SapB)-like domain of AtPSAPLIP1 with chosen template is d1n69a. Confidence 99.66%. 108 Alignment for the second SapB-like domain of AtPSAPLIP1 with chosen template d1nkla. Confidence 99.71%. Alignment for the first SapB-like domain of AtPSAPLIP2 with chosen template is d1n69a. Confidence 99.79%. Alignment for the second SapB- like domain of AtPSAPLIP2 with chosen template d1nkla. Confidence 99.70%. Analysis was done by Phyre2. (B) Hydropathy plot for AtPSAPLIP1 and AtPSAPLIP2. Plots were created in ExPASy. Window size was 9 with linear weight variation model. Figure 3-02. Predicted structure of saposin B (SapB)-like domains in AtPSAPLIP1 and AtPSAPLIP2. Image colored by rainbow N to C terminus with EzMol. Conserved cysteines are highlighted by brown sticks and the potential glycosylation site is highlighted by pink stick. In surface view, the negative region is colored by blue and positive region is colored by red. The distribution of some of the conserved residues was slightly different between two SapB-like domains in the single PSAPLIP. For example, in the second SapB like 109 domain, the conserved aspartic site (D190 in AtPSAPLIP1) is close to the fifth cysteine (C192 in AtPSAPLIP1), while a conservative aspartic site (D86 in AtPSAPLIP1) is closer to the fourth cysteine (C81 in AtPSAPLIP1) and relative away from the fifth cysteine (C105 in AtPSAPLIP1) (Figure 3-02). The positively charged lysine (K155 in AtPSAPLIP1) is close to the third and fourth cysteines (C157 and C167 in AtPSAPLIP1)in the second SapB like domain but this does not present in the first one (Figure 3-03). As a result, conformation of these two SapB-like domains is likely to be slightly different. And it is also possible that two SapB-like domains are processed into two mature forms like human saposins. Based on mammalian saposins working model, saposins can form and function as dimers (Rossmann et al., 2008; Olmeda et al., 2012). In plant cells, the PSAPLIP might be processed into individual mature saposins like human prosaposin, or function on its own by bending and forming a ?self-dimer? with two lobes or forming a true dimer with another PSAPLIP molecule. Post-transcriptional processing into individual mature saposins is more likely in gymnosperms, liverworts, mosses and algae because there are three SapB-like domains in a single gene. Unlike PSI in aspartic proteases, the direction of SapB-like domains in plant PSAPLIPs is not permuted. The PSI in aspartic proteases occur in green algae, which suggests that the saposin-like domains in aspartic proteases are very ancient and the relationship between PSI in modern aspartic proteases and modern PSAPLIPs is not close. PSI is likely to have evolved from an ancient duplication and deletion event. Therefore, aspartic proteases containing PSI and PSAPLIPs should be considered as two 110 different groups of proteins. 111 112 113 114 115 116 117 118 119 120 121 122 123 Figure 3-03. Sequence alignment of plant PSAPLIPs from several selected angiosperms. One or two representative sequences were selected from each species. Alignment was conducted with Clustal MUSCLE and image was produced by JalView. Color method was Taylor with a conservation level 85%. Annotation was calculated automatically. Conserved cysteines were highlighted in yellow. Subcellular localization of Arabidopsis PSAPLIPs Sequence alignment of Arabidopsis AtPSAPLIP1 and AtPSAPLIP2 with human prosaposin showed alignments with human saposin B, saposin C and saposin D. In human saposins, there are two glycosylation sites in saposin A (N80 and N101), and one glycosylation sites in saposin B (N215), saposin C (N332) and saposin D (N426). The glycosylation sites in saposin B, saposin C and saposin D are in the same position in alignments. No post-translational modifications are predicted for Arabidopsis PSAPLIPs in Uniprot. However, sequence alignment results provide putative glycosylation information in Arabidopsis PSAPLIPs. in AtPSAPLIP1, the N57 in the first SapB-like domain is aligned with human saposin B glycosylation site (Figure 3-04A), therefore it is speculated that this site might be glycosylated in the plants. The in the second SapB-like domain, P143 is aligned with the corresponding glycosylation site in human saposin C (Figure 3-04A), and therefore this site. This suggests that this site is unlikely to be glycosylated in the plants. And in total there?s only one putative site in AtPSAPLIP1 might be glycosylated in the cells. In AtPSAPLIP2, the first aligned site in 124 the first SapB-like domain is Y corresponding to the human saposin B glycosylation site and the site in the second SapB-like domain is P (Figure 3-04A). This suggests AtPSAPLIP2 is not glycosylated in the plant cells. Both genes were cloned and overexpressed using the 35S promoter in Arabidopsis. Total protein was extracted and digested with Endo-Hf to release the N-linked glycosylation from the proteins. The results showed that AtPSAPLIP1 was glycosylated and AtPSAPLIP2 was not glycosylated (Figure 3-04B). Together, these results support the sequence and structural similarity between Arabidopsis PSAPLIPs and human prosaposin. As a further expectation, they may share a similar molecular mechanism in lipid interactions. 125 Figure 3-04. Glycosylation and subcellular localization of AtPSAPLIP1 and AtPSAPLIP2. (A) Sequence alignment between human prosaposin and Arabidopsis PSAPLIPs. Red boxes mark the positions of N-glycosylation in human saposins and the corresponding 126 sites in Arabidopsis PSAPLIPs. Green diamonds mark the conserved cysteines. (B) Western blot of AtPSAPLIP1-CFP and AtPSAPLIP2-CFP with and without Endo Hf digestion. Blot was probed with anti-HA. (C) Confocal laser microscopy images of AtPSAPLIP1-CFP and AtPSAPLIP2-CFP. 7 days after germination seedlings were treated with 10?M brefeldin A (BFA) or 100nM concanamycin A (conc A). Bar=100?m. Among other species, this putative glycosylation site corresponds to the same glycosylation site as human saposin B. However, this site can also be K, E, H or Y (Figure 3-03; Figure S09), and this suggests that this site is not conserved. This may reflect that glycosylation is not likely to be essential for saposin activities in plants, and the glycosylation may be a trait inherited from ancestors. In Chlamydomonas reinhardtii, the corresponding sites are T and E, which also suggests that these two PSAPLIPs may not be glycosylated. The second SapB-like domain corresponding glycosylation site in AtSAPLIP2 is also P, and this site is relatively conserved across plant species. This suggest that plant PSAPLIPs may function in a way different from human saposins. Since human prosaposins are processed into four mature saposins, it is possible that plant PSAPLIPs may also be like mammalian counterparts. However, whole protein extraction and Western blotting analysis of AtPSAPLIP1 and AtPSAPLIP2 overexpression showed that the amount of full-length protein (58kDa) was much greater than predicted processed single domain (46kDa) (Figure 3-03B). This result suggests that Arabidopsis PSAPLIPs might not be processed into single saposin-like 127 proteins. A ?self-dimer? or dimerization with another molecule are likely to occur in the cell. This would function differently from human saposins. Both AtPSAPLIP1 and AtPSAPLIP2 were targeted to the vacuole (Figure 3-04C). To explore the trafficking pathway of PSAPLIP1 and PSAPLIP2, brefeldin A (BFA) treatment, a fungal inhibitor which blocks trafficking between endoplasmic reticulum (ER) and Golgi complex, were applied. Accumulated signals of PSAPLIP1 appeared after 30 minutes, while PSAPLIP2 did not accumulate after 30 minutes or one hour (Figure 3- 04C). These results suggest that PSAPLIP1 passes Golgi body while PSAPLIP2 does not. To test whether PASPLIP1 and PSAPLIP2 traffic to the vacuole, concanamycin A (conc A) treatment, which inhibits the vacuolar type H-ATPase and further inhibits fusion with vacuoles, were applied. Both proteins were affected and accumulated outside the vacuole after 30 minutes treatment (Figure 3-04C). This result indicates that both proteins traffic to the vacuoles. Expression pattern of AtPSAPLIP1 and AtPSAPLIP2 To find the possible roles of PSAPLIPs in Arabidopsis growth and development, the promoters were cloned and fused with the GUS reporter to elucidate the expression pattern of these two genes. Both genes were highly expressed in floral organs (Figure 3-05 and Figure 3-06). In floral organs, AtPSAPLIP1 was primarily expressed in inflorescences, pedicels, receptacles, sepals and the mature pollen, with weak expression in carpels, filaments 128 and petals (Figure 3-05A, B, C). No expression was detected in stigma or ovules (Figure 3-05D, E). Expression in petals and filaments showed increasing signals with developmental stages (Figure 3-05B). No other expressions were found in anthers except the mature pollen (Figure 3-05C). Signals in germinated pollens were detected on stigmas (Figure 3-05D). AtPSAPLIP1 expression was previously shown to be upregulated during leaf senescence (Gepstein et al., 2003). However, the promoter GUS results showed that expression was higher in young leaves, while in senescent leaves the staining was weaker (Figure 3-05F). Expression was also detected in the roots (Figure S14). Figure 3-05. AtPSAPLIP1 promoter GUS staining. (A) Inflorescence and young flower buds. (B) Flowers. From left to right, young to old flowers, with increasing staining in petals and filaments. (C) Mature pollen. (D) 2DAP silique. (E) Developing seeds. (F) Emerging leaf (left) and senescent first true leaf (right). DAP: days after pollination. Bar=1mm. 129 AtPSAPLIP2 was primarily expressed in inflorescences, pedicels, receptacles, petals, anthers, carpels and ovules, with weak expression in sepals and filaments (Figure 3-06). Expression of AtSAPLIP2 in petals and anthers showed an interesting pattern: expression was high in anthers and low in petals in younger stages (Figure 3- 06B), and decreased in anthers till around flower stage 9 and increased in petals with developmental stage starting from flower stage 8 (Figure 3-06B). Expression in the pollen was not detected (Figure 3-06C). Expression in ovules was present before fertilization (Figure 3-06C). In developing seeds, the expression was detected in young siliques, especially in integuments (Figure 3-06D, E, F, G). The signal decreased rapidly, and approximately 10 days after pollination, no signals were detected in seeds (Figure 3-06H). Developmental stage of embryos approximately matches linear to early mature stages around 10 days after pollination. From Seedgenenetwork.net, PSAPLIP2 expression is moderate in chalazal seed coat and general seed coat, low in chalazal endosperm at linear stage. Expression is low in embryos before linear stage and not detected in mature embryos. From the GUS staining (Figure 3-06F), the chalaza was stained darker. In vegetative tissues, root tips and leaf veins were also stained (Figure S14). 130 Figure 3-06. AtSAPLIP2 promoter:: GUS staining. (A) Inflorescence and young flower buds. (B) Flowers. From left to right, young to old flowers, with the changing staining in anthers and petals with different development stages. (C) Flowers at stage 15. Carpels was opened to show the unfertilized ovules. (D) Fertilized ovules. (E) 2DAP 131 silique. (F) Developing seeds and the integument. (G) 5 DAP silique and the inter integument and endosperms. (H) 10DAP silique. DAP: days after pollination. Bar=1mm. It appears that one of the most important function of AtPSAPLIP1 is in pollen maturation, while the function of AtPSAPLIP2 is in anther development, most likely in pollen formation. AtPSAPLIP2 is also involved in regulation of early seed development. The involvement in reproductive processes may explain why PSAPLIPs are ubiquitous across the plant kingdom, and the gene copies did not expand since reproductive regulation is a complicated and delicate. Discussion PSAPLIPs are ubiquitous in the plant kingdoms The plant specific insert (PSI) in plant aspartic proteases was the only saposin-like protein found in the plants. The PSAPLIPs in plants remained uncharacterized for a long time. Although the PSAPLIP family is not a large gene family, it is ubiquitous in the plant kingdom. This suggests its important role in plant growth and development. However, the function of PSAPLIPs in plant remains unclear. To elucidate the functions of plant PSAPLIPs, phylogenetic studies were conducted. By searching in protein database Uniprot, around 340 protein sequences were identified as plant PSAPLIPs. While PSAPLIPs containing three SapB-like domains are prevalent from green algae to gymnosperms, the major type of PSAPLIPs in 132 angiosperms contains only two SapB-like domains. This type first shown in gymnosperms, and this supports the single origin of angiosperms from gymnosperms that containing two SapB-like domains form of PSAPLIPs. In most plant species, there are one to four members in this family in the genome. Similar to animal PSAPLIPs, plant PSAPLIP also show diverse protein sequences. This was reflected in the phylogenetic tree: the relationship between branches was not always consistent to the phylogenetic relationship between different plant groups. However, the secondary structures of saposin-like domains in plants are highly similar to the mammalian saposins. This was reflected by the comparison between predicted structure of SapB-like domains from Arabidopsis PSAPLIPs and mammalian saposin- like protein structure. This highly similar secondary structure may help understanding the molecular function of PSAPLIPs in plant cells. Plant PSAPLIPs may also function as a membrane interactor, similar to human saposins, either by promoting membrane disruption or interaction with other proteins on membranes. Plant PSAPLIPs are similar to human prosaposin Some basic features were concluded from the results in Arabidopsis PSAPLIPs AtPSAPLIP1 and AtPSAPLIP2. Similar to human saposins in lysosomes, AtPSAPLIP1 and AtPSAPLIP2 were localized to vacuoles. The trafficking of both proteins was sensitive to concanamycin A, which supports that they are targeted to the vacuoles. The difference between these two proteins is PSAPLIP1 was sensitive to BFA treatment while PSAPLIP2 was not. This suggests that AtPSAPLIP1 is transported to trans-Golgi 133 netword first before trafficking to vacuoles, and AtPSAPLIP2 is directly trafficked to vacuoles. Unconventional trafficking directly towards vacuoles from ER has been previously reported, and it is dependent on post-translational modification, glycosylation, of the PSI in aspartic proteases (Vieira et al., 2019). Experimental results showed that the AtPSAPLIP1 was glycosylated and AtPSAPLIP2 was not. This may explain the difference between PSAPLIP1 and PSAPLIP2 in trafficking routes. However, how glycosylation affects the trafficking route choice, and whether glycosylation affects saposin-like protein activity are still unclear. It seems that the glycosylation does not significantly affect the saposin activity (Rossman et al., 2008). In the working model of human saposins, glycosylation may help hide the hydrophobic cavity, not it does not appear to be necessary for interaction with lipids (Rossmann et al., 2008). In plant PSAPLIPs, the putative glycosylation site is not conserved either across different species. This also supports the hypothesis that glycosylation seems beneficial but not essential to saposin-like protein activity. The primary form of mature Arabidopsis PSAPLIPs proteins was found to contain two SapB-like domains. The single saposin-like protein versions were much less than the full-length form. This reflects plant PSAPLIPs function in a different way from human prosaposin. Human prosaposin is processed into individual saposins, while in insect cells, di-saposins are the major products of processed prosaposins (Leonova et al., 1996). This indicates that prosaposins are not necessarily processed into individual saposin-like domains to function, and di-saposins are also able to function in the cell. 134 In plants, di-saposins may be the primary form as the functional unit in the cell. It may form a ?self-dimer? by interaction between the two SapB-like domains or interact with another PSAPLIP molecule in the cell. PSAPLIPs are likely involved in regulating plant reproductive processes Human saposins are co-factors for lipid-degradation enzymes (Kishimoto et al., 1992; Schuette et al., 2001). Little is known about degradation of sphingolipids in plants. Overexpression of PSAPLIPs in Arabidopsis did not affect plant growth and development (Figure S10-S13). One explanation is that PSAPLIPs are not involved in sphingolipid metabolism in plants. However, PSAPLIPs are not the lipid degradation enzymes themselves, and overexpression is not likely to disturb the metabolic levels without additional hydrolases. The proteomic studies on PSAPLIP interacting proteins may help in identify the sphingolipid hydrolases in the plants. The expression pattern of AtPSAPLIP1 and AtPSAPLIP2 suggest that they are important in reproduction processes. The differential expression indicates functional differentiation between AtPSAPLIP1 and AtPSAPLIP2. Animal PSAPLIP biological function is usually identified with the corresponding diseases and it is hard to infer the biological functions in the plants by sequences or structural features. The interaction between PSAPLIPs and other target proteins may help to explore the biological functions of plant PSAPLIPs. Expression and interacting databases provide clues for SAPLIP interacting partners. In the BioGrid database, AtPSAPLIP1 is annotated as interacting with AtRACK1A(Receptor for Activated C Kinase 1 A). RACK1A is a member 135 of the tryptophan-aspartate repeat (WD-repeat) family of proteins, function in shuttling proteins around the cell, anchoring proteins at particular locations and in stabilizing protein activity (Adams et al., 2011). RACK1A was reported involved in ABA signaling and may be required for production of ribosomes complex (Guo et al., 2011). This suggests that PSAPLIP1 may function in interaction between the RACK1A and the membrane system. The expression pattern of AtPSAPLIP1 also suggests a role in pollen maturation. AtPSAPLIP2 is annotated as interacting with SYNTAXIN-23 (SYP23), WAVY GROWTH 2 (WAV2) and EXCESS MICROSPOROCYTES1 (EMS1). WAV2 is primarily expressed in the roots. It encodes a protein belonging to the BUD EMERGENCE 46 family of proteins with a transmembrane domain at the N terminus and an ?/?-hydrolase domain at the C terminus (Mochizuki et al.,2005). This may be one of the lipid hydrolase candidates. The SNARE protein SYP21, SYP22 and SYP23 all localize on vacuolar membrane (Shirakawa et al., 2010). They may be the interactor for facilitating target protein interactions with the vacuolar membrane. EMS1 is expressed in tapetum, inner integument and chalaza. This expression pattern overlaps with AtPSAPLIP2 expression pattern (Figure 3-06). EMS1 is a leucin-rich receptor-like kinase which is localized on plasma membrane. It can interact with the ligand TAPETUM DEVELOPMENT 1 (TPD1) in regulation tapetum development (Huang et al., 2016). It is likely that AtPSAPLIP2 interacts with endocytosed EMS1 and transports it to vacuoles to terminate signal transduction. The primary biological function of AtPSAPLIP1 and AtPSAPLIP2 may be 136 in tapetum development and pollen maturation. The knockout mutant analysis is essential to further explore the role of PSAPLIPs in plants. Conclusion Plant PSAPLIPs are a small gene family in the plant kingdom. It shows similarity to the human prosaposin, in terms of protein structures, post-translational modification and subcellular localization. Plant PSAPLIPs may share a similar molecular mechanism of lipid interaction to human saposins. The expression pattern suggests their important role in flower development, but the exact role remains unresolved. Till now no T-DNA insertional lines are available for both genes. Generation mutants with CRISPR is one of the alternative choices. Second, the proteomic studies help identifying the biological pathways that AtPSAPLIPs are involved in, such as lipid metabolism or male gametophyte development signaling. Third, structural studies and lipid interaction assays in vitro will explore how plant PSAPLIPs interact with lipids and elucidate the biophysical and biochemical properties of these proteins. It is likely that Arabidopsis PSAPLIPs from a ?self-dimer? structure with the two SapB-like domains. Whether there is oligomerization is also of interest. Trafficking of PSAPLIPs in the cell is also important, since PSAPLIP2 appear to adopt an unconventional trafficking pathway in the cell. Whether the trafficking pathway affects the function of PSAPLIPs remains unclear. Some of the potential PSAPLIPs interactors such as EMS1, are targeted to the plasma membranes. Where the PSAPLIPs and the target proteins meet 137 in the trafficking pathways need further explore. Other directions also deserve attention, such as plant defense response. PSI from aspartic proteases is shown to have anti-bacteria activity in vitro and enhance the plants resistance to bacteria pathogens (Mu?oz et al., 2010; Frey et al., 2018). However, the independent function of PSI from the aspartic protease in vivo is not reported yet. Plant PSAPLIPs may also show anti-bacteria activity and function in plant defense responses. If PSAPLIPs are involved in anti-bacterial activity in vivo, then it could be considered a good candidate for enhancing resistance to a broad spectrum of bacteria pathogens. The study of PSAPLIPs from other plant species will also broaden the understanding of their biological functions in plants. For those species lacking floral structures, whether PSAPLIPs take part in reproductive process needs to be explored. This might contribute to our understanding that how reproduction system changes during evolution. In summary, this dissertation first characterized some feasures of plant prosaposins-like proteins and provides new insights on how plants regulate reproductive process. Materials and Methods Primary and Secondary Structure Prediction Hydropathy plot was drawn in ExPASy with Kyte and Doolittle method. Window size was 9 with the linear weight variation model. Structure prediction was conducted 138 in Phyre2. Each SapB-like domain was predicted separately. Predicted structure of AtPSAPLIP1 and AtPSAPLIP2. Final image was visualized with EzMol. Sequence Alignment PSAPLIPs protein sequences were selected in EggNOG and Uniprot. In EggNOG, sequences were found by screening orthologs with AT3G51730. 167 sequences from 67 species were outputs. In Uniprot, sequences were screened by the keyword saposin. Only sequences in Viridiplantae were chosen for further screening. The sequences annotated as fragments were removed. Aspartic proteases were removed as well. For those without a gene ID, if sequence similarity was above 95%, the longer one was kept. If the annotated SapB-like domains length was below 50 amino acid residues, the corresponding sequences were also removed. Some sequences were removed because they belong to the neucleophosmin family. This may result from incorrect auto-prediction in Uniprot. The remaining sequences were considered valid PSAPLIP proteins in plants and were selected for further analysis. Alignment was conducted in MegaX with Clustal MUSCLE method. The parameters were as following: gap open -2.9, gap extend 0, hydrophobicity multiplier 1.2, max memory in MB 2048, max iterations 16, cluster method UPGMA, cluster method UPGMA, min diag length 24. Some manual adjustments were applied for gap positions for better alignments. 139 To better search for conserved positions, the sequences that only contain one SapB-like domain were removed. Sequences in green algae, liverworts, mosses and gymnosperms were aligned separately due to their variable number of copies of SapB- like domains. Human prosaposin and Arabidopsis thaliana PSAPLIPs were chosen as the outlier in this case. Images were processed with JalView. Color was added by Taylor method with conservation level 85%. Annotation was calculated automatically. Phylogenetic tree construction Phylogenetic tree of plant PSAPLIPs were constructed in MegaX with maximum likelihood method. Phylogeny test was bootstrap method, with 2000 bootstrap replications. Substitutions type was amino acid with WAG model. Rates among sites were uniform. All sites were considered. ML heuristic method was nearest neighbor interchange method. No branch swap filter. Number of threads was 3. Preparation of Transgenic Plants AtPSAPLIP1 and AtPSAPLIP2 coding sequences (CDS) were cloned from Arabidopsis flower cDNA. The fragments were incorporate into pDONR/Zeo by BP reaction according to the manufacturer?s instructions. The entry clones were confirmed by sequencing and the primers for cloning and sequencing are listed in Table S01. The entry clones were then incorporated into pEarleyGate102 (35S promoter, with CFP and HA tag on C-terminus) by LR reaction. The expression constructs were 140 confirmed by sequencing and the sequencing primers are listed in Table S01. The expression constructs were transformed into agrobacterium strain GV3101, and floral dipped into Arabidopsis flowers. The positive seedlings were screened by hygromycin selection and confirmed by confocal microscopy. Promoter GUS reporter lines were constructed in a similar way. For AtPSAPLIP1, the promoter includes 3 prime UTR of previous gene and 5 prime UTR of PSAPLIP1, with a total length 451bp. For AtPSAPLIP2, the promoter includes the 5 prime UTR with a total length 1.5kb. The fragments were also inserted into pDONR/Zeo by BP reaction and incorporated into pGWB3 by LR reaction. Both entry clones and expression constructs were confirmed by sequencing. The cloning and sequencing primers are listed in Table S01. Transgenic plants were screened by hygromycin selection. Details on methods of molecular clonings are described in appendix D. Plant Materials and chemical treatment All the Arabidopsis thaliana plants were in Columbia-0 (Col-0) ecotype genetic background. For germination on solid media, Arabidopsis seeds were surface sterilized by soaking in 20% bleach (containing sodium hypochlorite) for 15 minutes with agitation. Seeds were then rinsed three to five times in sterile water. Seeds were sowed on 1/4 Murashige and Skoog (MS) medium (RPI Corp.) media containing 0.5% sucrose with 0.8% agar. Seeds were stratified at 4?C for 2 days in the dark and then placed in growth chamber at 22?C, with 24 hr continuous white light at 100 ?mol m-2 s -1. For 141 chemical treatment, seedlings were first grown for 4 days on regular 1/4MS media, then transferred to new media containing the corresponding chemicals. The working concentrations if chemicals used in this research were: 10?M brefeldin A (BFA); 100nM concanamycin A (conc A); 4?M propidium iodide (PI); 5?g/ml fluorescein diacetate (FDA); 2?M abscisic acid (ABA); 75mM sodium chloride (NaCl). Low nitrogen media was prepared by adding 10?M potassium nitrate (KNO3) in 1/4MS without nitrogen (MS w/o nitrogen) media; sufficient nitrogen media was prepared by adding 5mM KNO3 in 1/4MS w/o nitrogen media. Ethanol or DMSO was added as the control, with the concentration same as the corresponding chemicals. Adult plants were grown in growth chamber at 22?C with 16 hr light and 8 hr darkness cycles. Protein extraction 300mg Arabidopsis seedling tissues were ground with a grind stick in Eppendorf tubes with liquid nitrogen. The ground tissues were resuspended in 300 ?L protein extraction buffer (50 mM sodium citrate, pH 5.5; 5% SDS (w/v); 0.01% BSA (w/v); 150 mM NaCl; 2% (v/v) ?-mercaptoethanol and 1 ?L of protease inhibitor cocktail (Genesee Scientific). The mixture was incubated for 60 minutes at 100? C. Samples were centrifuged at 4? C, 14,500g for 30 minutes and the supernatant was collected. The samples were stored in -80? C if not used immediately. Glycosylation test Glycosylation was detected by Endo Hf (New England BioLabs) digestion 142 according to manufacturer?s instruction. Briefly, 17?L extracted protein sample was added with 2?L 10xGlycoBuffer 3, 1?L Endo Hf. Samples were incubated at 37?C for 1 hour. Then the sample was for SDS-PAGE and Western blot. SDS-PAGE Total proteins were separated by SDS polyacrylamide gel electrophoresis (SDS- PAGE). 10?L samples were prepared by adding 2?L of 6X SDS (sodium dodecyl sulfate) loading buffer (1.2g SDS, 0.01% bromophenol blue, 4.7ml glycerol, 1.2ml Tris 0.5M pH=6.8, 2.1ml water). Samples were loaded onto 12% polyacrylamide 0.75mm 10-well or 15-well gel (Bio-Rad?). Precision Plus Protein Dual Color Standards (Bio Rad) was used as size markers. Electrophoresis was carried out in 1X Running Buffer (3g of Tris base, 14.4g of glycine, and 1g of SDS in 1000 ml water) at 120V for approximately 4 hours or until the dye front reached the front of the gel. Western blot For immunoblotting, proteins were transferred to polyvinylidene difluoride (PVDF) membrane in Tris-glycine-methanol transfer buffer (2.9g glycine, 5.8g Tris, 0.37g SDS 100mL methanol, 900mL water) at 120V for 80 minutes at 4?C and then rinsed briefly in 1xPBS. Membranes were blocked overnight at 4?C in blocking buffer (5% non-fat milk in 1xPBS with 0.02% Tween20) or 1.5 hours at room temperature. The membrane was rinsed gently with washing buffer (1% non-fat milk in 1x PBS with 0.02% Tween20) for three times each for 15 minutes. The membrane then was incubated with primary 143 antibody (anti-HA) in blocking buffer overnight at 4?C or 1.5 hours at room temperature. The membrane was rinsed with washing buffer for three times and each for 15 minutes. Then the membrane was incubated with secondary antibody (anti- rabbit digoxigenin) at room temperature for 1.5 hours. The membrane was rinsed with washing buffer for three times and each for 15 minutes. Proteins were visualized using a SuperSignal West Femto Kit (Thermo Scientific). Images were taken by C-DiGit Blot Scanner (LI-COR). Microscopy Confocal microscopy was carried out using a Zeiss LSM 710 Confocal laser scanning microscope (Carl Zeiss, Germany) with Axio Imager 2. Pixel dwell time was 0.01 ms. The master gain was always set to less than 893, with a digital gain of 1.5. For RFP/mCherry acquisition: 594 nm (5%) excitation and 588-696 nm emission. For YFP acquisition: 514 nm (5%) excitation and 519-560 nm emission. For GFP: 488 nm (5%) excitation and 493-598 nm emission. For CFP: 458 nm (5%) excitation and 453-580 nm emission. For PI: 543 nm (5%) excitation and 583-718 nm emission. For FDA: 488 nm (5%) excitation and 493-583 nm emission. Quantification of florescence intensity was analyzed using ZEN Lite 2012. Briefly, representative images from 10 nuclei in each of the 5 Seven to 10 anthers from stage 13-14 were selected for imaging. All images were processed with ZEN Lite 2012 (Zeiss) and ImageJ. 144 Histochemistry For GUS staining detection, plant tissues were fixed in cold 90% acetone for 30 minutes, then washed twice in GUS buffer before staining. Samples were infiltrated with GUS buffer under vacuum for 10 minutes, then incubated at 37?C for 48 hours. Tissue was cleared in 70% ethanol overnight and repeated several times until the tissue becomes clean and clear. The sample was mounted on microscope slides for visualization. 145 Chapter 4: Conclusion and Perspective Plants have two types of proteins contain saposin B-like domains: aspartic proteases with the plant specific insert (PSI) and prosaposin-like proteins (PSAPLIPs). In this dissertation, three main questions were addressed. What are the biological functions of these aspartic proteases in plants? What is the role of the saposin-like domain (plant specific insert) in those processes? What are the biological functions of prosaposin-like proteins in plants? Using molecular genetic analysis, I conclude that the typical aspartic proteases function in bulk proteolytic activity such as seed storage protein processing/degradation and programmed cell death (PCD). The plant specific insert guides the protease towards the vacuole and perhaps also facilitates membrane disturbance. There is no evidence supporting that this PSI could function independently from the protease. From phylogenetic analysis and studies in Arabidopsis AtPSAPLIPs, I conclude that PSAPLIPs are important in reproductive processes especially in male gametophyte development. First, I began phenotypic studies in the least studied Arabidopsis aspartic protease ASPA2. ASPA2 is expressed throughout the plant, which is similar with the reported expression pattern of ASPA1. Single loss-of-function aspa2 mutant showed delayed seed maturation. As the delay in maturation is subtle, I suspected that there was redundancy in three ASPAs in Arabidopsis. Then I tested phenotype of aspa1-2 aspa2- 1 aspa3-3 triple mutant (ASPA1 is knock-down, ASPA2 and ASPA3 are knock-out alleles). Triple mutant seeds showed delayed germination in terms of germination rate and 146 seed storage proteins degradation. The fusion of small vacuoles to form the central vacuole was also delayed in the mutant cotyledons. This result suggests that ASPAs involved in not only proteolytic activity but also membrane disturbance. To explore ASPAs function in other tissues, I compared root growth in the wild type, triple mutant and ASPAs overexpression plants. Root architecture was different in response to nitrogen supply in that the triple mutant root showed more lateral roots and primary root growth was relatively insensitive to nitrogen levels compared to wild type. Further analysis suggested that the altered root architecture may result from tracheary element (TE) maturation in xylem tissues. The triple mutant showed slight delay TE maturation and the ASPA2 overexpression showed slightly earlier maturation. Together with the expression pattern of ASPA3, this indicates that ASPAs may regulate the rate of programmed cell death in Arabidopsis. Then I monitored PCD in lateral root cap with propidium iodide (PI) staining or propidium iodide/ fluorescein diacetate (PI/FDA) double staining. The distance between stained nuclei and root tip was not different between triple mutant and wild type. This suggests that onset of PCD was not delayed in the mutant. But the distance between the distal nucleus and root tip was longer in the mutant which indicated a longer execution time of PCD. This was confirmed by time-course imaging with PI staining. To test the mechanism of ASPAs in PCD, PI/FDA double staining was applied to monitor the time from cytosolic pH drop to PI signal appearance in nucleus. This time was longer in the triple mutant and shorter in overexpression plants. This 147 indicates that membrane permeability increased more slowly in the mutants and faster in the overexpression plants. This reflects the role of ASPAs in the rates of membrane permeability regulation during PCD. The ASPA promoters were cloned for constructs with a HISTONE 2A 10 reporter tag constructs to detect transcriptional responses to stress signals. ASPA1 expression remained constant while ASPA2 expression was upregulated by low nitrogen and ABA treatment but downregulated by salt stress. This result indicated that ASPA1 is more like a housekeeping gene and ASPA2 is more responsive to different signals for fine tuning plant growth and development to stress signals. ASPA3 expression was confined to tissues that undergo PCD, and most likely to function in protein degradation and nitrogen recycling to fine-tune the rate of PCD. The independent function of PSI was not detected in this dissertation. No phenotype was found in triple mature mutant or catalytic inactive ASPA2 overexpression plants. For future studies, it would be good to create a knock-out triple mutant for analysis. Since the ASPA1 allele is knock-down, the remaining ASPA1 activity may compensate of the missing ones and thus the phenotype of aspa1-2/2-1/3-3 may be weak. Second, one would complement the phenotype of knock-out plants. If there is a triple knock-out mutant, the catalytic inactive version ASPA2 D107A would also be used to determine if the PSI itself might rescue the phenotype. The reason may be that the phenotypes of the knock-down mutant was weak, and it was not powerful to test the difference when ASPA2 D107A was overexpressed in plants. Third, one would 148 screen for the substrates of APSAs to discover if membrane proteins were targets. The PSI allows the protease to associate with membranes, and this close interaction brings the protease and substrates together and makes it possible for membrane protein degradation. Through sequence screening and alignment, I found that prosaposin-like proteins (PSAPLIPs) are ubiquitous throughout plant kingdom. This family did not disappear, nor did it expand in evolution either. In most species, there are 1-4 genes in the genome. In angiosperms, there is an N-terminal signal peptide and two saposin B (SapB)-like domains. In gymnosperms, liverworts, mosses and green algae, PSAPLIPs contain three SapB-like domains. Plant PSAPLIPs show low sequence similarity but high similarity in secondary structure of SapB-like domains. This structural similarity was indicated by glycosylation analysis of Arabidopsis AtPSAPLIP1 and AtPSAPLIP2, and AtPSAPLIP1 was glycosylated and AtPSAPLIP2 was not. Both AtPSAPLIP1 and AtPSAPLIP2 were targeted in vacuoles. However, trafficking of AtPSAPLIP1 was sensitive to BFA while AtPSAPLIP2 was not. The differences in glycosylation and response to trafficking inhibitors indicate different trafficking routes to the vacuole. Although they are trafficked in different routes, both are in the vacuole, and this indicates that PSAPLIPs function in facilitating degradation of specific proteins. Then the promoter GUS reporter constructs were created for expression analysis. AtPSAPLIP1 was primarily expressed in inflorescence, especially in sepals, carpels and mature pollen. AtPSAPLIP2 was expressed in inflorescence too, but primarily in young 149 anthers, petals and ovules. These results suggest differential functions of PSAPLIPs in Arabidopsis. Since both are expressed in stamens, the results indicate the role in male gametophyte development. This may explain why this family is widely spread in the plant kingdom but did expand during evolution. Future studies include the molecular genetic analysis using loss-of-function mutants. So far no T-DNA insertional mutants are available for those two genes, CRISPR would be employed to create mutants. The potential male sterility phenotype would be the focus of the biological observation. Preliminary data suggest that there is a possible AtPSAPLIP2 mutant which shows male sterility phenotype. Conformation of the mutation via sequencing is needed. Second is the proteomic study to find protein-protein interactions. One possible target for AtPSASLIP2 is EXCESS MICROSPOROCYTES1 (EMS1) due to annotation in plat protein- protein interaction data base BioGrid and the overlapping expression patterns in flowers. The third, in plants where floral structures do not exist, such as liverworts and mosses, the functions of PSAPLIPs need to be explored, and this may also provide information about the reason why there is one SapB-like domain missing in angiosperms. Summary This dissertation was divided into two parts. Part one is the first in vivo study of ASPA biological functions in the plants. Besides demonstrating ASPAs role in both seed development and seed germination in vivo for the first time, this is also the first 150 time showing that ASPAs are involved in programmed cell death in plants. Part two characterized some features of the plant prosaposin-like proteins (PSAPLIPs) for the first time. The potential role in male gametophyte development may contribute to our knowledge of the regulation of plant reproductive processes. In the first part, three hypotheses were made: ASPA2 was involved in regulation of seed development; ASPAs were involved in regulation of programmed cell death; the plant specific insert (PSI) in ASPAs has an independent biological function in vivo. By molecular genetic studies, results showed that ASPAs regulated seed development and seed germination. ASPAs also regulated programmed cell death in lateral root caps, and they are likely to promote membrane permeability in this process. By studying the catalytic inactive form of ASPA2 overexpression plants, the supporting results for an independent role of PSI were not obtained. Further research directions would include proteomic studies on whether ASPAs prefer degradation of membrane targets. Phenotypic studies on the impact of PCD defects in other tissues such as the tapetum, ideally with the knockout triple mutants. Cellular and molecular studies on whether there are other trafficking pathways for ASPAs such as secretion to the extracellular space, and how glycosylation affects ASPAs in the cell. And in other species, whether ASPAs have other biological functions is another important direction. In the second part, plant PSAPLIPs were characterized for the first time. Some features, such as the distribution in the plant kingdom, the sequence structures, the 151 predicted secondary structures were described. By molecular genetic studies of Arabidopsis PSAPLIP1 and PSAPLIP2, their functions were proposed as interaction with and facilitating target proteins for degradations. They were important in male gametophyte development. Further directions of studies include phenotypic studies on the knockout mutants, elucidating the biological functions in different tissues. Proteomic studies are also helpful for screening the interactors for potential targets and help to understand in which pathways are PSAPLIPs in the cells. The structural studies will increase our understanding on how these proteins function, especially the interaction with lipids. This may help to explore the common and unique features of plant PSAPLIPs compared from animal PSAPLIPs. 152 Appendix A Supplemental Figures for Chapter 2 Figure S01. Colocalization between ASPA2 and autophagy marker ATG8a, and endosomal sorting complex required for transport (ESCRT) machinery associated protein FREE1. (A) Colocalization between YFP-ATG8a and ASPA2-RFP. Pearson? 153 correlation r range: -0.25 ? 0.03 (B) Colocalization between GFP-FREE1 and ASPA2-RFP. Pearson?s correlation r range: 0.17 ? 0.26. Yellow arrows point to the overlapped signals. 5 DAG seedlings were treated with 100nM conc A for 1 hour and imaged. Bar=10?m. 154 Figure S02. Phenotype of 30 DAG plants of ASPA overexpression lines. (A) Col (B) aspa1-2/2-1/3-3 (C) 35S::ASPA1-RFP (D) 35S::ASPA2-RFP (E) 35S::ASPA2 D107A-RFP (F) 35S::ASPA2 D107A R402Q-RFP (G) 35S::ASPA2 321AA-RFP (H) 35S::ASPA3-RFP. 155 Bar=2cm. Figure S03. Phenotype of 40 DAG plants of ASPA overexpression plants. (A) Col (B) aspa1-2/2-1/3-3 (C) 35S::ASPA1-RFP (D) 35S::ASPA2321AA-RFP (E) 35S::ASPA2-RFP (F) 35S::ASPA2-D107A-RFP (G) 35S::ASPA2 D107A R402Q-RFP (H) 35S::ASPA3-RFP. 156 Bar=5cm. Figure S04. 35S promoter::ASPA2 D107A N404A-CFP (potential glycosylation site mutation) subcellular localization. (A) Control. (B) BFA treatment for 30 min. (C) Conc A treatment for 30min. Bar=20?m. Figure S05. ASPA1 expression in seedlings. Hypocotyls (top) and cotyledon (bottom). 157 Yellow arrows point to the signals in nuclei. Bar=200?m. Figure S06. ASPA3 promoter::YFP expression in lateral root cap. Bar=50?m. 158 Figure S07. Subcellular localization of ASPA1 and ASPA3 in mature leaves. (A) ASPA1- CFP. (B) ASPA3-CFP. Bar=50?m. 159 Appendix B Supplemental Figures for Chapter 3 160 161 162 163 164 165 166 167 168 169 Figure S07. Sequence alignment of PSAPLIPs in green algae, liverwort, moss and gymnosperm. Left: Species names with gene ID after the vertical line. If gene ID was not available, protein is was annotated. Conserved sites were shaded with colors in JalViews. Conservation, quality, consensus, and occupancy were calculated and visualized in JalViews by default. 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 Figure S08. Sequence alignment of PSAPLIPs in angiosperms. Left: Species names with gene ID after the vertical line. If gene ID was not available, protein is was annotated. Conserved sites were shaded with colors in JalViews. Conservation, quality, consensus, and occupancy were calculated and visualized in JalViews by default. 202 203 204 205 206 Figure S09. Sequence alignment of PSAPLIPs which contain three SapB-like domains. Human prosaposin and Arabidopsis PSAPLIPs as outliers. Left: Species names with gene ID after the vertical line. If gene ID was not available, protein is was annotated. Conserved sites were shaded with colors in JalViews. Conservation, quality, consensus, and occupancy were calculated and visualized in JalViews by default. 207 Figure S10. Root growth in AT3g51730 overexpression plants. 4 DAG seedlings were transferred to media containing ABA (2?m) or NaCl (75mM) for another 4 days. Black dots marked the root tip position of 4 DAG seedlings. Figure S11. Root growth in AT5g01800 overexpression plants. 4 DAG seedlings were transferred to media containing ABA (2?m) or NaCl (75mM) for another 4 days. Black dots marked the root tip position of 4 DAG seedlings. 208 Figure S12. Phenotype of 30 DAG Col and 35S::AtPSAPLIP1-CFP plants. Left: Col; Middle: 35S::AtPSAPLIP1-CFP Line 1; Right: 35S::AtPSAPLIP1-CFP Line 2. Bar=5cm. 209 Figure S13. Phenotype of 30 DAG Col and 35S::AtPSAPLIP2-CFP plants. Left: Col; Middle: 35S::AtPSAPLIP2-CFP Line 1; Right: 35S::AtPSAPLIP2-CFP Line 2. Bar=5cm. 210 Figure S14. Arabidopsis PSAPLIPs promoter::GUS activity in seedlings. (A) PSAPLIP1. (B) PSAPLIP2. 2-week-old seedlings were stained. Bar=0.5cm. 211 Figure S15. Candidate of At5g01800 CRISPR mutant. (A) Col plant (B) Possible At5g01800 CRISPR plant (C) Inflorescence of Col (left) and possible mutant (right). (D) Silique length and seed number in Col (left) and possible mutant (right). (E) DIC image of anther from stage 14 flower in Col (left) and At5g01800 CRISPR candidate (right). Yellow arrows show the wrinkled pollens. Statistics shown in (F). P<0.05 by Student? t- 212 test (G) propidium iodide (PI)/ fluorescein diacetate (FDA) double staining of the pollens in At5g01800 CRISPR candidate. Pollens from stage 14 flowers. From left to right: PI, FDA, TML, merge. PI staining indicates dead pollens. (H)Portions of small and wrinkled pollens in Col and At5g01800 CRISPR candidate. P<0.05 by Student? t-test. Bar=2cm in (A)(B)(C), 2mm in (D), 100?m in (E), 50?m in (G). 213 Figure S16. Candidate of At3g51730 CRISPR mutant. (A) Col plant (B) Possible At3g51730 CRISPR plant (C) Stage 15 flower of Col and (D) stage 15 flower of CRISPR candidate. Yellow arrows indicate the surface of the anther and released pollens. (E)Inflorescence Influorescence of Col and (F) Inflorescence possible mutant. Red 214 arrows indicate the fertile siliques. (G) DIC image of anther from stage 14 flower in Col and (H) At3g51730 CRISPR candidate. Blue arrows show the wrinkled pollens. Statistics shown in (I). P<0.05 by Student? t-test (J) Portions of small and wrinkled pollens in Col and At5g01800 CRISPR candidate. P<0.05 by Student? t-test. (K) propidium iodide (PI)/ fluorescein diacetate (FDA) double staining of the pollens in At3g51730 CRISPR candidate. Pollens from stage 14 flowers. From left to right: PI, FDA, TML, merge. PI staining indicates dead pollens. Blue arrows show the wrinkled pollens. Bar=2cm in (A)(B)(E)(F), 2mm in (C)(D), 100?m in (G)(H), 50?m in (K). 215 Appendix C Phylogenic tree of plant PSAPLIPs 216 217 218 219 220 221 222 Figure S17. Phylogenetic tree of PSAPLIPs in plants. Phylogenetic tree was constructed in MegaX with maximum likelihood method. Phylogeny test was bootstrap method, with 2000 bootstrap replications. Substitutions type was amino acid with WAG model, which was chosen by Richard et al., 2007. Rates among sites were uniform. All sites were considered. ML heuristic method was nearest neighbor interchange method. No branch swap filter. Number of threads was 3. 223 Appendix D Materials and Methods Plant materials Wild-type Arabidopsis plants (Arabidopsis thaliana ecotype Columbia-0) and T- DNA insertion mutants were sourced from the Arabidopsis Biological Resource Center, The Ohio State University (ABRC; www.abrc.osu.edu). For germination on soil, seeds were evenly distributed directly on seed germination potting mix. For BASTA selection, BASTA were diluted in sterile water with a concentration of 120mg/ml and sprayed every other day. Green seedlings were selected for further study. For germination on solid 1/4MS media, Arabidopsis seeds were surface sterilized by soaking in 20% bleach (containing sodium hypochlorite) for 15 minutes with agitation. Seeds were then rinsed three to five times in sterile water. Seeds were sowed on 1/4MS media supplemented with the appropriate antibiotics and/or chemicals. Seedlings were transferred to soil if needed. Seeds were stratified for 2 days (4?C, dark) then transferred to continuous white light (100?mol m-2 s -1) at 22?C conditions for germination. For germination experiments, Arabidopsis seeds (harvested within a month and dried for at least three days) were surface sterilized and plated to 1/4MS. With or without stratification, plates were transferred to continuous white light (60?mol m-2 s -1) conditions and assessed every 12 hours for radical emergence. Nucleic acid isolation Genomic DNA isolation 224 Genomic DNA was extracted from Arabidopsis seedlings or inflorescence as indicated. Approximately 100mg tissues were collected and grinded in 200?l CTAB extraction buffer and heated at 65?C for at least 30 minutes. Then add 200?l chloroform and isopropanol mix (v: v=24:1) and vortex for 30 seconds. The mixture was centrifuged at 12000rpm for 3 minutes. The supernatant was transferred to new tubes. Add 2?l glycogen (5%) and 167?l isopropanol and mixed thoroughly. The samples were placed at -20?C overnight. Then samples were centrifuged with 12,000rpm at 4?C for 15 minutes. Dispose the supernatant and wash the pellet with 500?l CTAB washing buffer and re-centrifuge with the same setting. Dispose the supernatant, dry the samples with blowing air and dissolve the samples with 500?l sterile water. CTAB Extraction Buffer 100 mM Tris-HCl (pH 7.5), 25 mM EDTA, 1.5 M NaCl, 2% (w/v) CTAB CTAB Wash Buffer 70% Ethanol, 10mM Ammonium acetate Genomic DNA was subsequently used for sequence amplification, or for genotyping by PCR. Total RNA isolation Total RNA was isolated from 50-100 mg fresh tissue using ZR plant RNA MiniPrep kit (Zymo Research) according to the manufacturer?s instructions. To remove 225 contaminating genomic DNA, an on-column DNA digest was performed at the time of RNA extraction (DNase I, New England BioLabs) according to manufacturer?s instructions. Plasmid isolation and purification 1.5mL overnight Escherichia coli culture (A600=2-4) containing the plasmid of interest were pelleted by centrifugation at 10,000 g for 1 minute. Cells were resuspended with 100?l solution I. Add 200?l solution II and mix well and sit for 1 minute. Mix with 150?l solution III and centrifuge at max speed for 15 minutes. Transferred the supernatant to new tubes and add 1ml ethanol. Set the samples on ice for 5 minutes and then centrifuged at 12,000rpm for 15 minutes. Wash the pellet with 1ml 70% ethanol and centrifuge again. Dissolve the pellet with 100?l sterile water for further use. Solution I 50 mM glucose, 10 mM EDTA and 25 mM Tris-HCl, pH 8.0 Solution II 0.2 N NaOH, 1% SDS Solution III 3 M potassium acetate, 2 M acetic acid Nucleic acid manipulations 226 Agarose gel electrophoresis Nucleic acids were prepared and analyzed using a 1% (w/v) agarose gel made with 1X TAE buffer and molecular grade agarose (Dot Scientific). Nucleic acids were visualized by staining with 0.5% ethidium bromide. Voltage was applied in BioRad submerged horizontal gel. Gels were visualized using Gel DocTM XR system (BioRad). 50xTAE buffer Tris base 0.04M, disodium EDTA 0.002M, acetic acid 0.02M. PCR For genotyping and colony-PCR, Taq DNA Polymerase with standard Taq buffer (New England BioLabs) was used according to manufacture instructions. A typical reaction contains components as following: 10x Standard Taq Reaction Buffer 2.5?l 10mM dNTPs 0.5?l 10?M Forward Primer 0.5?l 10?M Reverse Primer 0.5?l Template variable Taq DNA Polymerase 0.125?l Nuclease-Free Water to 25?l Typical PCR program includes steps as the following: initial denaturation (95?C) for 3 minutes, denaturation (95?C), 30s for 35 cycles 227 annealing (55?C - 60?C), 30s extension (72?C), 1kb/60s final extension, 5 minutes (72?C). final step, 12?C For promoters, genomic sequence and CDS sequence cloning, Herculase II Fusion DNA polymerase was used. According to the manufacturer?s instructions, a standard reaction contains 5xHerculase II reaction buffer 10?l dNTP mix (25mM each dNTP) 0.5?l Template variable 10?M Forward Primer 1.25?l 10?M Reverse Primer 1.25?l Herculase II DNA polymerase 0.5?l Nuclease-Free Water to 50?l Typical PCR program includes steps as the following: initial denaturation (95?C) for 2 minutes, denaturation (95?C), 15s annealing (40?C - 60?C depends on the sequence), 20s extension (72?C), 1kb/30s final extension, 3 minutes (72?C). final step, 12?C 228 PCR genotyping PCR genotyping was used to confirm the identity of T-DNA insertion mutants, or putative crosses (F1s). The presence or absence of alleles of interest was determined using diagnostic PCR primer pairs. For known T-DNA insertion mutants, including for F1s, the wild-type allele was identified using primer pairs which spanned the insertion site; the mutant allele was identified using one of the wild-type primers in combination with a T-DNA specific primer (LBb1.3). Homozygous T-DNA insertion mutants were identified by the presence of the allele, and absence of the wild-type allele. Quantitative RT-PCR Gene transcript levels were analyzed from total RNA by quantitative RT-PCR using SYBR Green PCR Master Mix kit (Thermo Fisher Scientific) in a BioRad thermo-Cycler, according to manufacturer?s instructions. RT-PCR was performed at 95?C for 2 minutes, then 35 cycles of 95?C for 20 s and 54?C for 20 s and 72?C 20s. Melt curve analysis was performed to ensure specificity of the reaction. Threshold values were determined by the CFX manager software (BioRad) and the relative mRNA levels were determined by the 2-??CT method (Pfaffl 2004), using ACTIN 2(ACT2) as a reference gene. Site-directed mutagenesis PCR Herculase II was used with the same protocol. After PCR the mixture was digested with DpnI (New England BioLabs) at 37?C for 1 hour. 2?l were used for transformation 229 into E.coli. Plasmids were extracted and sequenced for confirmation. Purification of PCR products PCR products were purified using silica. Bands of interest were excised after separation on an agarose gel and two volume: weight ratio of 6M NaI was added. Incubate the agarose in 6M NaI at 55?C for 5-10 min with occasional mixing. Add 10?l of the silica suspension. Vortex gently. Stand for 5 min at room temperature with occasional mixing. One mg of the silica (=10?l of the silica suspension) binds 3-4.5?g of DNA. Spin for 1 min at 12,000rpm. Discard the supernatant and carefully remove residual liquid. Suspend the pellet in 500?l of Solution E. Spin for 1 min at 12000rpm. Discard the supernatant and wash the pellet again. Allow the pellet to air-dry for 10 min. Add an appropriate volume (at least one pellet volume) of sterile water. Vortex gently to resuspend the pellet. Stand for 3 min at 70?C and Spin for 1 min. Transfer the supernatant into a new microfuge tube. Solution E 50 mM NaCl, 10 mM Tris-HCl pH 7.5, 2.5 mM EDTA, 50%(v/v) ethanol. Preparation of Silica Suspend 5 g of silica (Sigma, S-5631) in 50 ml of sterile water. Allow the silica to settle for 2h. Discard the supernatant containing fine particles. Resuspend the pellet with sterile water and re-settle for 2hr. After discarding the supernatant, the packed silica was resuspended in 50ml sterile water to make a final concentration of approximately 230 100mg/ml. Digest and ligation reactions Restriction digests and ligation reactions were carried out as per manufacturer?s instructions using 1?L enzyme per 50?L reaction (New England BioLabs). Reactions were incubated overnight at 37?C. Ligation reactions were carried out using T4 ligase (New England BioLabs). Enzymes used for digestion are described in primers. DNA sequencing Sequencing of purified DNA was performed by the Eurofins Genomics. Concentrations of primer and purified DNA were as recommended by Eurofins Genomics. Sequence analysis was performed using the VectorNTI? software (Life TechnologiesTM). Gateway Cloning Gateway Cloning Binary vectors for in planta genetic modification were constructed using Gateway technology (InvitrogenTM) as follows. TOPO reaction TOPO of entry clones Gateway? compatible entry clones are prepared according to manufacturer?s instructions. TOPO-compatible overhang was incorporated into the 231 fragment of interest during PCR amplification by including a CACC at the beginning of the forward primer. Mix the following components for reaction: Fresh PCR product 0.5?4 ?l, Salt Solution 1 ?l, Water add to a total volume of 5 ?l, pENTR/D-TOPO? vector 1 ?l. Final Volume 6 ?l. The mixture was incubated at least 30 minutes at room temperature. 2?l of this reaction was transformed into Mach1-T1 (InvitrogenTM) chemically competent E. coli cells. Kanamycin was added to the media for selection. Positive clones were identified by colony PCR, in which a small amount of bacterial colony was incorporated directly into a PCR reaction. The plasmids were sequenced using M13F and M13R primers and additional internal primers where necessary. BP reaction BP entry clones are prepared according to manufacturer?s instructions. BP- compatible overhang was incorporated into the fragment of interest during PCR amplification by including attB1 and attB2 at the beginning of the forward and reverse primers. Mix the following components for reaction: Fresh PCR products 1 ?l, BP Clonase II enzyme mix 0.5?l, 232 pDONR/Zeo vector 1?L, water add to a total volume 10?L. The mixture is incubated at least 3 hours at room temperature. 4?L of this reaction was transformed into Mach1-T1 (InvitrogenTM) chemically competent E. coli cells. Zeocin was added to media for selection. Positive clones were identified by colony PCR. The plasmids were sequenced using M13F and M13R primers and additional internal primers where necessary. LR reactions Gateway destination vectors contain attR Gateway compatible sites flanking a ccdB death gene. These plasmids were cultured in DB3.1 ccdB survival E. coli cells (InvitrogenTM). Destination vectors used in this project included the pEARLEYGATE102 (pEG; Earley et al. 2006), the pUBC series (Curtis and Grossniklaus 2003), pH7WGC2, pH7WGR2, and pGBW3 (for promoter analysis, Karimi et al. 2002). Expression clones were prepared using LR Clonase II Gateway kit (InvitrogenTM) according to the manufacturer?s instructions. Molar ratios were carefully balanced. LR reactions were incubated for 3 hours at room temperature. 2 ?l was transformed into E. coli cells. Proper antibiotics are added in the media for selection depending on the destination vectors. Positive clones were identified by colony PCR. Transformation of bacteria 233 Mach1-T1 chemically competent E. coli cells were prepared by Mix and Go! Transformation Buffer Set (Zymo Research) according to the manufacturer?s instruction. Briefly, E. coli cells were thawed on ice, then incubated on ice with 2 ?l of plasmids cloning mixture for 30 minutes. 300 ?l room temperature LB was added. Transformed cells were shaken horizontally at 37?C, 200rpm for 1 hour, then inoculated onto LB agar plates containing the relevant antibiotic and incubated at 37?C overnight. For transformation of Arabidopsis, vectors of interest were transformed into electrocompetent Agrobacterium cells by electroporation. GV3101 agrobacterium cells were thawed on ice, then incubated on ice with 1?l plasmid DNA for 30 minutes. Cells were transferred to a chilled 1 mm electroporation cuvette and electroporated using a Bio-Rad Micropulser? (?AGR? settings). Transformed cells were shaken horizontally at 28?C, 200rpm for 3 hours, then inoculated onto LB agar plates containing the relevant antibiotics and incubated at 28?C for two days. Transformation of Arabidopsis plants by floral dipping A modified version of the floral dip method (Clough and Bent 1998) was used for agrobacterium -mediated transformation of Arabidopsis plants. Briefly, agrobacterium carrying the desired construct was streaked to LB agar plates (with selection) and incubated at 28?C for 24 hours. Agrobacterium was resuspended in 80 ml fresh LB media to an OD600 of 2-2.5. 234 The bacteria were harvested by centrifuging with 3000g at room temperature and then suspended in 5% sucrose solution containing 0.02% Silvet-77. Arabidopsis floral buds were dipped in this solution and wrapped with plastic wraps to keep the humidity. The plants were covered in large plastic bags in the dark overnight. Positive transformants were identified by germinating on agar plates supplemented with hygromycin, or by germinating on soil and treating seedlings with a selective herbicide BASTA. Then selected seedlings were confirmed by LSM confocal microscope Crossing Arabidopsis genetic lines Suitable inflorescences containing healthy flower clusters were chosen. Elongating siliques and open flower buds were removed under a dissecting microscope. Ideal flower buds (large, but without an exposed stigma) were carefully emasculated, avoiding damage to the stigma, style petals and sepals. The other flower buds were removed to avoid confusion. The emasculated flowers were ready after 24 hours for pollination. Mature anthers from the paternal parent were collected and used to spread pollens onto the exposed stigmatic region. Cross-pollinated flowers were label in a piece of paper tape. Successful crosses were identified in the F2 generation. Protein assays Total protein extraction from imbibed seeds Protein extraction 235 300mg Arabidopsis tissues were ground with a grind stick in Eppendorf tubes with liquid nitrogen. The ground tissues were resuspended in 300 ?L protein extraction buffer (50 mM sodium citrate, pH 5.5; 5% SDS (w/v); 0.01% BSA (w/v); 150 mM NaCl; 2% (v/v) ?-mercaptoethanol and 1 ?L of protease inhibitor cocktail (Genesee Scientific). The mixture was incubated for 60 minutes at 100? C. Samples were centrifuged at 4? C, 14,500g for 30 minutes and the supernatant was collected. The samples were stored in -80? C if not used immediately. Glycosylation test Glycosylation was detected by Endo Hf (New England BioLabs) digestion according to manufacturer?s instruction. Briefly, 17?L extracted protein sample was added with 2?L 10xGlycoBuffer 3, 1?L Endo Hf. The samples were incubated at 37?C for 1 hour. Then the sample was used for SDS-PAGE and Western blot. SDS-PAGE Total proteins were separated by SDS polyacrylamide gel electrophoresis (SDS- PAGE). 10?L samples were prepared by adding 2?L of 6X SDS (sodium dodecyl sulfate) loading buffer (1.2g SDS, 0.01% bromophenol blue, 4.7ml glycerol, 1.2ml Tris 0.5M pH=6.8, 2.1ml water). Samples were loaded onto 12% polyacrylamide 0.75mm 10-well or 15-well gel (Bio-Rad?). Precision Plus Protein Dual Color Standards (Bio Rad) was used to mark band size. Electrophoresis was carried out in 1X Running Buffer (3g of 236 Tris base, 14.4g of glycine, and 1g of SDS in 1000 ml water) at 120V for approximately 4 hours or until the dye front reached the front of the gel. Western blot For immunoblotting, proteins were transferred to polyvinylidene difluoride (PVDF) membrane in Tris-glycine-methanol transfer buffer (2.9g glycine, 5.8g Tris, 0.37g SDS 100mL methanol, 900mL water) at 120V for 80 minutes at 4?C and then rinsed briefly in 1xPBS. Membranes were blocked overnight at 4?C in blocking buffer (5% non-fat milk in 1xPBS with 0.02% Tween20) or 1.5 hours at room temperature. The membrane was rinsed gently with washing buffer (1% non-fat milk in 1x PBS with 0.02% Tween20) for three times each for 15 minutes. The membrane then was incubated with primary antibody (anti-HA) in blocking buffer overnight at 4?C or 1.5 hours at room temperature. The membrane was rinsed with washing buffer for three times each for 15 minutes. Then the membrane was incubated with secondary antibody (anti-rabbit digoxigenin) at room temperature for 1.5 hours. The membrane was rinsed with washing buffer for three times and each time for 15 minutes. Proteins were visualized using a SuperSignal West Femto Kit (Thermo Scientific). Images were taken by C-DiGit Blot Scanner (LI-COR). Coomassie blue staining For visualization of seed storage proteins, the gel stained by incubating overnight 237 in 20ml Coomassie staining solution (0.1% Coomassie bright blue in 50% methanol, 10% acetic acid). The gel was de-stained for 3 hours with de-staining solution (10% acetic acid, 50% methanol). At least two changes of this solution until the background was nearly clear. Histochemistry Promoter GUS activity was visualized in planta using a GUS or histone10 2A (H2A) tagged with fluorescent protein reporter system. Promoter fragments were amplified from wild type genomic DNA. Promoters were inserted upstream of ?-glucuronidase (GUS) in the pGBW3 destination vector for transformation into wild type Arabidopsis. For ASPAs, target promoters were replaced for the UBQ10 promoter in pUBC::YFP-Dest or pUBC::mCherry-Dest vectors. Then H2A was incorporated by LR reaction. For all constructs, putative transformants were identified by hygromycin (GUS constructs) or BASTA (H2A-YFP/mCherry construct) and confirmed by genomic DNA PCR using promoter-specific forward GUS reverse primers. For GUS staining detection, plant tissues were fixed in cold 90% acetone for 30 minutes, then washed twice in GUS buffer before staining. Samples were infiltrated with GUS buffer under vacuum for 10 minutes, then incubated at 37?C for 48 hours. Tissue was cleared in 70% ethanol overnight and repeated several times until the tissue becomes clean and clear. The sample was mounted on microscope slides for visualization. 238 GUS staining buffer Sodium phosphate buffer (pH=7) 100mM, EDTA 10mM, Triton X-100 (w/v) 0.1%, potassium ferrocyanide 2mM, potassium ferricyanide 2mM, X-glucuronide 0.5mg/ml. Microscopy and imaging Microscopy Confocal microscopy was carried out using a Zeiss LSM 710 Confocal laser scanning microscope (Carl Zeiss, Germany) with Axio Imager 2. Pixel dwell time was 0.01 ms. The master gain was always set to less than 893, with a digital gain of 1.5. For RFP/mCherry acquisition: 594 nm (5%) excitation and 588-696 nm emission. For YFP acquisition: 514 nm (5%) excitation and 519-560 nm emission. For GFP: 488 nm (5%) excitation and 493-598 nm emission. For CFP: 458 nm (5%) excitation and 453-580 nm emission. For PI: 543 nm (5%) excitation and 583-718 nm emission. For FDA: 488 nm (5%) excitation and 493-583 nm emission. Quantification of florescence intensity was analyzed using ZEN Lite 2012. Image production Post-processing of microscopy images was performed using Fuji/ImageJ and associated plugins (www. http://fiji.sc/; Schneider et al. 2012; Schindelin et al. 2012), or Zeiss ZEN Black v10.0 (Carl Zeiss, Germany; http://www.zeiss.com/microscopy/). Image quantification was carried out using ImageJ. 239 Bioinformatics Primary and Secondary Structure Prediction Hydropathy plot was drawn in ExPASy with Kyte and Doolittle method. Window size was 9 with the linear weight variation model. Structure prediction was conducted in Phyre2. Each SapB-like domain was predicted separately. Predicted structure of AtPSAPLIP1 and AtPSAPLIP2. Final images were visualized with EzMol. Sequence Alignment PSAPLIPs protein sequences were selected in EggNOG (http://eggnog5.embl.de/) and Uniprot (www.uniprot.org). In EggNOG, sequences were identified via pairwise ortholog predictions with AT3G51730. 167 sequences from 67 species were outputs. In Uniprot, sequences were screened by searching keyword saposin. Only sequences in Viridiplantae were chosen for further screening. The sequences which were annotated as fragments were removed. Aspartic proteases were removed as well. For those sequences without the gene ID, if sequences similarity was above 95%, the longer one was kept. If the annotated SapB-like domain length was below 50 amino acid residues, the corresponding sequences were also removed. After first try of alignment, the sequences belonging to the neucleophosmin family were removed. The remaining sequences were considered valid PSAPLIP proteins in plants and for further analysis. 240 Alignment was conducted in MegaX with Clustal MUSCLE method. The parameters were as following: gap open -2.9, gap extend 0, hydrophobicity multiplier 1.2, max memory in MB 2048, max iterations 16, cluster method UPGMA, cluster method UPGMA, min diag length 24. Some manual adjustments were applied for gap positions for better alignments. To search for conservative positions, the sequence that only contain one SapB- like domains were removed because they may be incomplete sequences if there are errors in predictions. Sequences in green algae, liverworts, mosses and gymnosperms were aligned separately due to their variable number of copies of SapB-like domains. Human prosaposin and Arabidopsis PSAPLIPs were chosen as the outlier. Images were processed with JalView. Color was added by Taylor method with conservation level 85%. Annotation was calculated automatically. Phylogenetic tree construction Phylogenetic tree of plant PSAPLIPs were constructed in MegaX with maximum likelihood method. Phylogeny test was bootstrap method, with 2000 bootstrap replications. Substitutions type was amino acid with WAG model. Rates among sites were uniform. All sites were considered. ML heuristic method was nearest neighbor interchange method. No branch swap filter. Number of threads was 3. Statistical analysis 241 All means and standard errors were calculated using Microsoft Excel 2013. Where indicated, statistical significance was determined using a Student?s t-test, with tails=2 and type=3 (independent samples of unequal variance; Microsoft Excel 2013) unless otherwise indicated. Pearson?s chi square analyses were performed to determine the segregation ratios for single insertion segregation where mentioned. Appendix E Supplemental Tables Table S01. Primer List in this dissertation. Primer Name Direction Use Sequence aspa2-1 Forward Genotyping for T-DNA TTTTTGGAGCATTATTGCGAC 242 SALK_097505 LP insertion aspa2-1 Reverse Genotyping for T-DNA AATTCGAATGTGTGACAAATCG SALK_097505 RP insertion aspa2-2 Forward Genotyping for T-DNA TTTTTGGAGCATTATTGCGAC SALK_021601 LP insertion aspa2-2 Reverse Genotyping for T-DNA ATTGATCCTGAGCCGTAATGG SALK_021601 RP insertion aspa1-1 Forward Genotyping for T-DNA GTCTTGGTGCAATTGAGATT SALK_092586 LP insertion aspa1-1 Reverse Genotyping for T-DNA AATAGCATTTTGATGATGGC SALK_092586 RP insertion aspa1-2 Forward Genotyping for T-DNA ATGAAGATATACTCTAGAAC SALK_041027 LP insertion aspa1-2 Reverse Genotyping for T-DNA ATACCAAACAGGAGCAGCTT SALK_041027 RP insertion aspa3-1 CS330614 RP Forward Genotyping for T-DNA ATGGGAACTAGGTTCCAATC insertion aspa3-1 CS330614 LP Reverse Genotyping for T-DNA ACATCATCATTGCTAAAGTA insertion aspa3-2 SK36621 LP Forward Genotyping for T-DNA CTATTTGGATGCTCAATACT insertion 243 aspa3-2 SK36621 RP Reverse Genotyping for T-DNA GGAGATCACCCATGTCAAAC insertion aspa3-3 Forward Genotyping for T-DNA CATAAAGGTTACTGGCAGTT SALK_056711C LP insertion aspa3-3 Reverse Genotyping for T-DNA TACTGCAGACAGACATGAAT SALK_056711C RP insertion LBb1.3 Forward Genotyping for T-DNA ATTTTGCCGATTTCGGAAC insertion ASPA2 qRT F Forward Quantitative PCR TTGAGGCAGAACATGACTCA ASPA2 qRT R Reverse Quantitative PCR CACGGCTTCTGCGAAGCCAA ASPA1 qRT F Forward Quantitative PCR GAGCGCATATTGAACTACGT ASPA1 qRT R Reverse Quantitative PCR GGCTGCCTCTGCAAACCCGA ASPA3 qRT F Forward Quantitative PCR ACACAAGAACGCATACTCGC ASPA3 qRT R Reverse Quantitative PCR AGCAGCTTTGGCGAATCCAA ACT2 qRT F Forward Quantitative PCR ACACTGTGCCAATCTACGAGGGT T ACT2 qRT R Reverse Quantitative PCR ACAATTTCCCGCTCTGCTGTTGTG ASPA2 F attB1 Forward CDS cloning GGGGACAAGTTTGTACAAAAAA GCAGGCTCCATGTCCCCTATAGAT CC ASPA2 R NS attB2 Reverse CDS cloning GGGGACCACTTTGTACAAGAAA 244 GCTGGGTCCACGGCTTCTGCGAA GCCAA ASPA1 F attB1 Forward CDS cloning GGGGACAAGTTTGTACAAAAAA GCAGGCTCCATGAAGATATACTCT AGAAC ASPA1 R ns attB2 Reverse CDS cloning GGGGACCACTTTGTACAAGAAA GCTGGGTCGGCTGCCTCTGCAAA CCCGA ASPA3 F attB1 Forward CDS cloning GGGGACAAGTTTGTACAAAAAA GCAGGCTCCATGGGAACTAGGTT CCAATC ASPA3 R ns attB2 Reverse CDS cloning GGGGACCACTTTGTACAAGAAA GCTGGGTCAGCAGCTTTGGCGA ATC ASPA2 400 SEQ F Forward Primer for sequencing CAATCTTGGTGGTGATTCTG ASPA2 800 SEQ F Forward Primer for sequencing CTGGCAGTTCGACATGGGTG ASPA2 1200 SEQ F Forward Primer for sequencing TTGAGGCAGAACATGACTCA ASPA2 1400 SEQ F Forward Primer for sequencing ACAATGTATTAGCGGCTTTA ASPA1 400 SEQ F Forward Primer for sequencing AGAAGAATGGAAAAGCTGCC ASPA1 800 SEQ F Forward Primer for sequencing TGTTCTTATTGGCGGTGCAC ASPA1 1200 SEQ F Forward Primer for sequencing GAGCGCATATTGAACTACGT 245 ASPA1 1400 SEQ F Forward Primer for sequencing TGCTCTTGACGTTGCTCCAC ASPA3 400 SEQ F Forward Primer for sequencing AGTCATCGTCATATAGAAAG ASPA3 800 SEQ F Forward Primer for sequencing GTTTGACATGGGTGATCTCC ASPA3 1200 SEQ F Forward Primer for sequencing ACACAAGAACGCATACTCGC ASPA3 1400 SEQ F Forward Primer for sequencing TTTCACGGCAATGGATATTG ASPA2 promoter PstI Forward Promoter cloning GGGCTGCAGATCTGATGCAAAGA F CGTGAC ASPA2 promoter SalI Reverse Promoter cloning GGGGTCGACTTTGACCTACAAAA R TCAAAG ASPA1 PRO PmeI SacI Forward Promoter cloning GAGTGTTTAAACGAGCTCAGTAA F GCTTGGAATGTCTTG ASPA1 PRO SalI R Reverse Promoter cloning GAGTGTCGACTTTACCTATTCATT GACAAC ASPA3 PRO SacI F Forward Promoter cloning CACCGAGCTCGGAAACGTATGCT TATGGGT ASPA3 PRO XhoI R Reverse Promoter cloning GGGGCTCGAGTTTTACCTGTCAT CAAAAAC ASPA2 PRO 500 SEQ F Forward Primer for sequencing CTCAAATCCTTATTTTTGGA ASPA2 PRO 1000 SEQ Forward Primer for sequencing AAACCTTTAGCCTATTAAAT F ASPA2 PRO 1500 SEQ Forward Primer for sequencing TCATGATGACACTTTTGTTC 246 F ASPA2 PRO 1900 SEQ Forward Primer for sequencing TCGAGGAACAGTTGTCTTAG F ASPA1 PRO 500 SEQ F Forward Primer for sequencing CTCAATCCAACGGTTAGTAT ASPA1 PRO 1000 SEQ Forward Primer for sequencing TTAGGTAAGAGTTTTGTTAC F ASPA1 PRO 1500 SEQ Forward Primer for sequencing TAGCAAAAGAAGTCTTTAGT F ASPA1 PRO 1800 SEQ Forward Primer for sequencing GGTATGGTTCTCTGCTTTTT F ASPA3 PRO 500 SEQ F Forward Primer for sequencing GTACCTAATGCTAAACAAAC ASPA3 PRO 1000 SEQ Forward Primer for sequencing CATCCTAGAAGATATCTTAA F ASPA3 PRO 1500 SEQ Forward Primer for sequencing TGTGAGTGTTCTTTTATACT F ASPA3 PRO 2000 SEQ Forward Primer for sequencing TCTTAGTCTAATAGTCTTCA F mCherry SpeI F Forward mCherry cloning GGGGACTAGTATGGTGAGCAAG GGCGAGGA mCherry PsiI R Reverse mCherry cloning GGGGTTATAATTACTTGTACAGCT CGTCCAT 247 ASPA2 D107A F Forward Site-directed CTGTCATTTTTGCTACCGGAAGCT mutagenesis CTAACC ASPA2 D107A R Reverse Site-directed GAGCTTCCGGTAGCAAAAATGAC mutagenesis AGTGAAC ASPA2 R402Q F Forward Site-directed GATACAGAGCCAATTGCAGCAGA mutagenesis ACATGACT ASPA2 R402Q R Reverse Site-directed CTTGAGTCATGTTCTGCTGCAATT mutagenesis GGCTCTG ASPA2 N404A F Forward Site-directed GAGCCAATTGAGGCAGGCCATG mutagenesis ACTCAAGAG ASPA2 N404A R Reverse Site-directed TCCTCTCTTGAGTCATGGCCTGCC mutagenesis TCAATTG attB1 SAPOSIN A3 F Forward Cloning GGGGACAAGTTTGTACAAAAAA GCAGGCTAAATGGGTGATCTCCA AATTGCT attB2 SAPOSIN A3 R Reverse Cloning GGGGACCACTTTGTACAAGAAA ns GCTGGGTAAGCAGCTTTGGCGA ATCCAAC attB1 AT3G51730 CDS Forward CDS cloning GGGGACAAGTTTGTACAAAAAA F GCAGGCTAAATGGGTCTTAAAGC TGGAAC 248 attB2 AT3G51730 CDS Reverse CDS cloning GGGGACCACTTTGTACAAGAAA R ns GCTGGGTAAGAATCAGCCAACTC CGGCT attB1 AT5G01800 CDS Forward CDS cloning GGGGACAAGTTTGTACAAAAAA F GCAGGCTAAATGGGCGGTAGATT TGGAGT attB2 AT5G01800 CDS Reverse CDS cloning GGGGACCACTTTGTACAAGAAA R ns GCTGGGTACGAATCTGCCAATGA CTCCAC attB1 AT3G51730 Forward Promoter cloning GGGGACAAGTTTGTACAAAAAA PROMOTER SacI F GCAGGCTAAGAGCTCAAGAGTG ATTGAAATGGTCT attB2 AT3G51730 Reverse Promoter cloning GGGGACCACTTTGTACAAGAAA PROMOTER XhoI R GCTGGGTACTCGAGGATTCCTGA TAAAGAAAAAAAG attB1 AT5G01800 Forward Promoter cloning GGGGACAAGTTTGTACAAAAAA PROMOTER SacI F GCAGGCTAAGAGCTCAAGGCAAT AACCACTCGATG attB2 AT5G01800 Reverse Promoter cloning GGGGACCACTTTGTACAAGAAA PROMOTER XhoI R GCTGGGTACTCGAGGTTTCCTCG TGAGATCTATA 249 AT5G01800 Forward Primer for sequencing CTCATCAGAATTTACATCTC PROMOTER 500 SEQ F AT3G51730 guideRNA Forward Primer for guide RNA ATTGAGACGTTTGCACTCTGTGT 1 F in CRISPR G AT3G51730 guideRNA Reverse Primer for guide RNA AAACCACACAGAGTGCAAACGTC 1 R in CRISPR T AT5G01800 guideRNA Forward Primer for guide RNA ATTGCCGATTCTTCTCGAACCATT 1 F in CRISPR AT5G01800 guideRNA Reverse Primer for guide RNA AAACAATGGTTCGAGAAGAATCG 1 R in CRISPR G ATG8a_CACC_F Forward Genomic sequence CACCATGATCTTTG CTTGCTTGAA cloning ATG8a_R Reverse Genomic sequence TCAAGCAACGGTAAGAGATC cloning M13 Forward Forward Primer for sequencing GTAAAACGACGGCCAG M13 Reverse Reverse Primer for sequencing CAGGAAACAGCTATGAC 35S SEQ F Forward Primer for sequencing GACGCACAATCCCACTATCCTTCG pUBC::CFP SEQ F Forward Primer for sequencing CTCGAGTGCGGGATCCTCTA Table S02. List of Plant PSAPLIPs. Data were screened from Uniprot. Protein ID was added if no gene ID was available in the Gene ID column. Number of SapB-like domains 250 was auto-predicted by Uniprot. If other type of domains were also predicted, the names of domains were indicated. After alignments, some results were added with a question mark which indicates the uncertainty of SapB-like domain numbers due to mutated or missing conserved cysteines. The inferred incomplete SapB-like domain was also indicated as incomplete? In the column. The order of domain annotations was from left to right: from N to C. Signal peptide was auto-predicted by Uniprot. NA: none available. Species Gene ID Number of Signal SapB-like peptide domains prediction Chloropocon primus A3770_07P47130 2 YES Chloropocon primus A3770_02p14820 2? YES Chloropocon primus A3770_04p29840 2 YES Chloropocon primus A3770_04p29830 2? YES Tetradesmus obliquus BQ4739_LOCUS15020 1 NA (Acutodesmus obliquus) Raphidocelis subcapitata Rsub_10640 3 YES Monoraphidium neglectum MNEG_12603 2? YES Coccomyxa subellipsoidea COCSUDRAFT_45864 3 YES (strain C-169) Chlorella variabilis CHLNCDRAFT_58828 3 YES 251 Chlorella sorokiniana C2E21_8413 3 YES Auxenochlorella APUTEX25_001631 3 YES protothecoides (Chlorella protothecoides) Tetraselmis sp. GSL018 TSPGSL018_26319 PPIase FKBP- YES type+2 Micromonas commoda MICPUN_62224 1 YES (strain RCC299 / NOUM17 / CCMP2709) Micromonas commoda MICPUN_105899 2 YES (strain RCC299 / NOUM17 / CCMP2709) Micromonas commoda MICPUN_98458 1+disordered NA (strain RCC299 / NOUM17 / region CCMP2709) Ostreococcus tauri BE221DRAFT_194138 2 YES Bathycoccus prasinos Bathy10g00200 1 YES Bathycoccus prasinos Bathy07g01820 2 YES Gonium pectorale GPECTOR_69g440 3 YES Tetrabaena socialis TSOC_008198 3? YES Chlamydomonas CHLRE_05g235700v5 3 YES 252 reinhardtii (Chlamydomonas smithii) Chlamydomonas CHLRE_02g105200v5 3 YES reinhardtii (Chlamydomonas smithii) Chlamydomonas eustigma CEUSTIGMA_g11715.t1 3 YES Klebsormidium nitens KFL_001110040 3 YES (Ulothrix nitens) Chara braunii CBR_g3540 3 YES Physcomitrella patens PHYPA_022478 3 YES subsp. patens Physcomitrella patens PHYPA_018982 3 YES subsp. patens Wollemia nobilis NA 3 YES Araucaria cunninghamii NA|A0A0D6R2G8_ARACU 3 YES Picea sitchensis NA|A9NUE1_PICSI 3 YES Picea sitchensis NA|A9P228_PICSI 2 YES Amborella trichopoda AMTR_s00007p00225690 2 YES Amborella trichopoda AMTR_s00062p00198130 2 YES Cinnamomum micranthum CKAN_01065200 2 YES f. kanehirae 253 Cinnamomum micranthum CKAN_00757300 1+incomplete YES f. kanehirae ? Cinnamomum micranthum CKAN_00308200 1 NA f. kanehirae Anthurium amnicola Psapl1_1 2 YES Anthurium amnicola Sftpb_0 2 YES Anthurium amnicola Psapl1_2 2 YES Anthurium amnicola PSAP_6 2 YES Anthurium amnicola PSAP_15 2 YES Anthurium amnicola PSAPL1_3 1 YES Anthurium amnicola mglC_0 1 YES Zostera marina ZOSMA_381G00120 2 YES Zostera marina ZOSMA_56G01350 2 YES Apostasia shenzhenica AXF42_Ash004723 2 YES Apostasia shenzhenica AXF42_Ash015547 2 YES Dendrobium catenatum MA16_Dca011512 2 YES Dendrobium catenatum MA16_Dca015668 2 YES Dendrobium catenatum MA16_Dca010547 2 YES Dendrobium catenatum MA16_Dca020165 2 YES Dendrobium catenatum MA16_Dca009510 incomplete?+ YES 1 254 Ensete ventricosum (Musa B296_00015606 2 YES ensete) Ensete ventricosum (Musa B296_00023675 2 YES ensete) Ensete ventricosum (Musa GW17_00023743 2 YES ensete) Ensete ventricosum (Musa B296_00030464 2 YES ensete) Musa acuminata subsp. 103971073 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. NA 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. 103974546 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. 103970701 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. 103995409 2 YES 255 malaccensis (Musa malaccensis) Musa acuminata subsp. 103992043 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. NA|M0RGN3 2 YES malaccensis (Musa malaccensis) Musa acuminata subsp. NA|M0REL0 2 NA malaccensis (Musa malaccensis) Musa balbisiana C4D60_Mb06t00990 2 NA Musa balbisiana C4D60_Mb08t21690 2 YES Musa balbisiana C4D60_Mb11t07550 2 YES Musa balbisiana C4D60_Mb02t19540 2 YES Musa balbisiana C4D60_Mb10t19630 2 YES Musa balbisiana C4D60_Mb10t28080 2 YES Musa balbisiana C4D60_Mb07t12400 1 YES Ananas comosus (Ananas ACMD2_06213 2 YES ananas) Ananas comosus (Ananas ACMD2_06262 2 NO 256 ananas) Phoenix dactylifera LOC103721950 2 YES Phoenix dactylifera LOC103702109 2 YES Phoenix dactylifera LOC103718784 2 YES Phoenix dactylifera LOC103713171 2 YES Phoenix dactylifera LOC103704544 1 NA Leersia perrieri NA|A0A0D9UX66 2 NA Leersia perrieri NA|A0A0D9WF85 2 YES Leersia perrieri NA|A0A0D9UX65 2 NA Oryza barthii NA|A0A0D3HQL1 2 YES Oryza barthii NA|A0A0D3EK02 2 YES Oryza barthii NA|A0A0D3G678 2 YES Oryza brachyantha 102703271 2 YES Oryza brachyantha NA|J3M622 2 NA Oryza brachyantha 102702884 2 YES Oryza glaberrima NA|I1QX61 2 YES Oryza glaberrima NA|I1NKJ7 2 YES Oryza glaberrima NA|I1PUJ1 2 YES Oryza glumipatula NA|A0A0D9Y3R0 2 NA Oryza glumipatula NA|A0A0D9ZXK7 2 YES Oryza glumipatula NA|A0A0E0BMU7 2 NA 257 Oryza meridionalis NA|A0A0E0DQ22 1+mutated 1? YES Oryza meridionalis NA|A0A0E0BXM2 2 YES Oryza punctata NA|A0A0E0JEI8 1+mutated 1? YES Oryza punctata NA|A0A0E0L159 2 YES Oryza rufipogon NA|A0A0E0RCU4 2 YES Oryza rufipogon NA|A0A0E0MRU5 2 YES Oryza rufipogon NA|A0A0E0PKX1 2 YES Oryza sativa subsp. indica OsI_00546 2 YES Oryza sativa subsp. indica OsI_34843 2 YES Oryza sativa subsp. indica OsI_37293 2 YES Oryza sativa subsp. indica OsI_19500 2 YES Oryza sativa subsp. Os12g0112200 2 YES japonica Oryza sativa subsp. P0028E10.2 2 NA japonica Oryza sativa subsp. Os01g0166700 2 NA japonica Oryza sativa subsp. Os05g0334400 2 NA japonica Brachypodium distachyon BRADI_4g44500v3 2 YES Brachypodium distachyon BRADI_4g25580v3 2 YES 258 Brachypodium distachyon BRADI_2g31070v3 2 YES Brachypodium distachyon BRADI_2g04110v3 2 YES Hordeum vulgare subsp. NA|A0A287NI79 2 NA vulgare Hordeum vulgare subsp. NA|A0A287R2L5 2 NA vulgare Hordeum vulgare subsp. NA|F2DBE9 2 YES vulgare Hordeum vulgare subsp. NA|F2CQA9 2 YES vulgare Hordeum vulgare subsp. NA|A0A287KA24 2 YES vulgare Aegilops tauschii subsp. NA|A0A453HNI0 2 YES strangulata Aegilops tauschii subsp. F755_31720 2 YES strangulata Aegilops tauschii subsp. NA|A0A453E3V5 2 YES strangulata Aegilops tauschii subsp. NA|A0A453ZQF0 2 YES strangulata Triticum aestivum NA|A0A3B6JIC3 2 YES 259 Triticum aestivum NA|A0A3B6KEQ2 2 YES Triticum aestivum NA|A0A3B6LJX9 2 YES Triticum aestivum NA|A0A3B6MMT5 2 YES Triticum aestivum NA|A0A3B6IRE3 2 YES Triticum aestivum NA|A0A3B6EBA4 2 YES Triticum aestivum NA|A0A3B5Y5L6 2 YES Triticum aestivum NA|A0A3B6FHG9 2 YES Triticum aestivum NA|A0A3B5Z461 2 NA Triticum aestivum NA|A0A3B6A1D1 2 NA Triticum turgidum subsp. TRITD_1Av1G205520 2 YES durum Triticum turgidum subsp. TRITD_4Bv1G048790 2 NA durum Triticum turgidum subsp. TRITD_4Av1G152380 2 YES durum Triticum turgidum subsp. TRITD_3Av1G029160 2 YES durum Triticum turgidum subsp. TRITD_5Av1G112490 2 YES durum Triticum turgidum subsp. TRITD_5Bv1G093140 2 YES durum 260 Triticum turgidum subsp. TRITD_1Bv1G194670 2 NA durum Triticum turgidum subsp. TRITD_3Bv1G033240 2 YES durum Triticum urartu TRIUR3_03527 2 YES Triticum urartu TRIUR3_22517 2 YES Triticum urartu TRIUR3_29270 2 YES Triticum urartu TRIUR3_22718 2 YES Arundo donax (Donax NA|A0A0A9R7P1 2 NA arundinaceus) Arundo donax (Donax NA|A0A0A9RV12 2 YES arundinaceus) Arundo donax (Donax NA|A0A0A9V0R9 2 YES arundinaceus) Arundo donax (Donax NA|A0A0A9V254 2 NO arundinaceus) Arundo donax (Donax NA|A0A0A9QNN3 1 NA arundinaceus) Arundo donax (Donax NA|A0A0A9LQW1 1 NA arundinaceus) Eragrostis curvula EJB05_31312 1+1 mutated? YES 261 Eragrostis curvula EJB05_03028 2 YES Eragrostis curvula EJB05_03037 2 YES Eragrostis curvula EJB05_29979 2 YES Eragrostis curvula EJB05_34950 2 YES Eragrostis curvula EJB05_29935 1 YES Eragrostis curvula EJB05_31440 1 NA Sorghum bicolor SORBI_3008G032600 2 YES Sorghum bicolor SORBI_3003G055700 2 YES Sorghum bicolor SORBI_3009G097200 1 YES Zea mays Zm00014a_038950 2 YES Zea mays Zm00014a_044659 2 YES Zea mays ZEMMB73_Zm00001d023371 2 YES Zea mays ZEMMB73_Zm00001d042734 2 YES Zea mays ZEMMB73_Zm00001d039719 2 YES Dichanthelium BAE44_0015216 2 YES oligosanthes Dichanthelium BAE44_0008708 2 YES oligosanthes Dichanthelium BAE44_0009052 2? NO oligosanthes Panicum hallii var. hallii GQ55_8G009100 2 YES 262 Panicum hallii var. hallii GQ55_5G490800 2 YES Panicum hallii var. hallii GQ55_3G007700 2 YES Panicum hallii var. hallii GQ55_3G333500 2 YES Panicum miliaceum C2845_PM08G04430 2 NA Panicum miliaceum C2845_PM17G00420 2 YES Panicum miliaceum C2845_PM05G21750 2 NA Panicum miliaceum C2845_PM07G00670 2 NA Panicum miliaceum C2845_PM06G26640 2 NO Setaria italica SETIT_7G327400v2 2 YES Setaria italica SETIT_5G117400v2 2 YES Setaria italica SETIT_8G015000v2 2 YES Setaria italica SETIT_3G284400v2 2 YES Setaria viridis SEVIR_7G337200v2 2 YES Setaria viridis SEVIR_5G113900v2 2 YES Setaria viridis SEVIR_8G013800v2 2 YES Setaria viridis SEVIR_3G292600v2 2 YES Aquilegia coerulea AQUCO_00400489v1 2 YES Macleaya cordata BVC80_1837g47 1 NA Macleaya cordata BVC80_9017g10 2 YES Papaver somniferum C5167_002404 2 NA Nelumbo nucifera LOC104597199 2 YES 263 Nelumbo nucifera LOC104602464 2 YES Spinacia oleracea SOVF_050110 2 YES Actinidia chinensis var. CEY00_Acc01858 2 YES chinensis Actinidia chinensis var. CEY00_Acc00072 2 YES chinensis Actinidia chinensis var. CEY00_Acc08725 2 YES chinensis Actinidia chinensis var. CEY00_Acc01859 2 YES chinensis Camellia sinensis var. TEA_000122 1 NA sinensis Davidia involucrata Din_006700 2 YES Davidia involucrata Din_026378 2 YES Nyssa sinensis F0562_017856 2 YES Nyssa sinensis F0562_015152 1? NA Artemisia annua CTI12_AA282550 2 YES Artemisia annua CTI12_AA349490 2 NA Helianthus annuus HannXRQ_Chr10g0286291 2 no Cynara cardunculus var. Ccrd_003008 2 YES scolymus 264 Lactuca sativa LSAT_9X38061 2 YES Daucus carota subsp. DCAR_010960 2 YES sativus Daucus carota subsp. DCAR_018655 2? NA sativus Dorcoceras hygrometricum F511_29468 2 NO Erythranthe guttata MIMGU_mgv1a013247mg 2 YES (Mimulus guttatus) Genlisea aurea M569_00799 2 YES Handroanthus CDL12_11605 2 YES impetiginosus Striga asiatica (Buchnera STAS_21183 2 YES asiatica) Striga asiatica (Buchnera STAS_33432 2 NA asiatica) Coffea canephora GSCOC_T00023234001 2 NA Cuscuta australis DM860_002763 2 YES Cuscuta campestris CCAM_LOCUS31065 2 YES Cuscuta campestris CCAM_LOCUS32789 2 YES Nicotiana attenuata A4A49_38798 2 YES Nicotiana attenuata A4A49_19559 2 YES 265 Nicotiana sylvestris LOC104216406 2 YES Nicotiana sylvestris LOC104224609 2 YES Nicotiana tabacum LOC107812754 2 YES Nicotiana tabacum LOC107816607 2 YES Nicotiana tabacum LOC107792809 2 YES Nicotiana tabacum LOC107777346 2 YES Capsicum annuum LOC107843427 2 YES Capsicum annuum LOC107851224 1 NA Capsicum baccatum CQW23_24170 2 YES Capsicum baccatum CQW23_32279 1 YES Capsicum baccatum CQW23_29496 1 YES Capsicum chinense BC332_26027 2 YES Capsicum chinense BC332_31415 1 NA Solanum chacoense NA|A0A0V0I0V1 2 YES Solanum chacoense NA|A0A0V0HIM7 2 YES Solanum tuberosum 102602502 2 YES Solanum lycopersicum NA|A0A3Q7I0I0 2 YES Vitis vinifera VIT_08s0058g01030 2 NA Vitis vinifera VITISV_040420 2 NA Vitis vinifera Psapl1_1 2 NA Vitis vinifera VITISV_040421 1 NA 266 INCOMPLETE ? + 1 Vitis vinifera CK203_030312 1 NA Vitis riparia NA|Q9M614 1 NA INCOMPLETE ? + 1 Arachis hypogaea Ahy_B04g070055 1 NA Arachis hypogaea Ahy_B04g071408 2 NA Arachis hypogaea Ahy_A02g006469 2 YES Arachis hypogaea Ahy_B02g061501 2 YES Arachis hypogaea Ahy_A04g018839 2 YES Lupinus angustifolius TanjilG_22489 1 YES Lupinus angustifolius TanjilG_19378 1 YES Cicer arietinum LOC101491522 2 NA Cicer arietinum LOC101508260 2 YES Medicago truncatula MTR_7g072560 2 YES Medicago truncatula MtrunA17_Chr4g0013141 2 NA Medicago truncatula MTR_4g029040 2 YES Trifolium pratense L195_g026334 2 NA Trifolium subterraneum TSUD_160400 1? YES Trifolium subterraneum TSUD_160390 1 YES 267 Trifolium subterraneum TSUD_266660 2 YES Lotus japonicus NA|I3S9R9 2 YES Cajanus cajan KK1_035920 2 YES Cajanus cajan KK1_029931 2 YES Mucuna pruriens CR513_52785 2 NA Mucuna pruriens CR513_55238 2 NA Phaseolus vulgaris PHAVU_008G084800g 2 YES Phaseolus vulgaris PHAVU_008G0847000g 2 YES Glycine max GLYMA_18G212100 2 YES Glycine max GLYMA_19G111400 1 YES Glycine max GLYMA_09G277100 2 YES Glycine max GLYMA_18G211900 2 NA Glycine max GLYMA_01G131400 2 YES Glycine max GLYMA_09G277200 2 YES Glycine max GLYMA_04G159500 1 NA Glycine max GLYMA_19G111500 1 YES Glycine soja D0Y65_025469 2 NA Glycine soja D0Y65_001396 2 NA Glycine soja D0Y65_025468 2 YES Glycine soja D0Y65_049180 2 NA Vigna angularis var. VIGAN_04117700 2 NA 268 angularis Vigna angularis var. VIGAN_04117800 2 YES angularis Vigna angularis var. VIGAN_09109000 2 YES angularis Vigna radiata var. radiata LOC106758717 2 YES Vigna radiata var. radiata LOC106754929 2 YES Vigna radiata var. radiata LOC106758948 2 YES Vigna unguiculata DEO72_LG10g3244 2 YES Vigna unguiculata DEO72_LG10g3245 2 YES Vigna unguiculata DEO72_LG8g1152 2 YES Citrus unshiu CUMW_001140 1 NA Acer yangbiense EZV62_016774 1 YES Eucalyptus grandis EUGRSUZ_K01273 2 YES Eucalyptus grandis EUGRSUZ_A00687 2 YES Punica granatum CRG98_041613 2 YES Punica granatum CRG98_041612 2 YES Punica granatum CRG98_016680 2 YES Corchorus capsularis CCACVL1_28877 2 YES Corchorus olitorius COLO4_30004 2 YES Gossypium arboreum F383_27015 2 YES 269 Gossypium arboreum F383_21360 2 YES Gossypium australe EPI10_020460 2 NA Gossypium barbadense GOBAR_AA23056 1 YES Gossypium barbadense GOBAR_AA12144 2 YES Gossypium barbadense GOBAR_AA02853 2 YES Gossypium barbadense ES319_D10G128500v1 2 YES Gossypium barbadense ES319_A10G160500v1 2 YES Gossypium barbadense ES319_D02G005800v1 2 YES Gossypium barbadense ES319_A02G004900v1 2 YES Gossypium darwinii ES288_D10G136900v1 2 YES Gossypium darwinii ES288_A10G179500v1 2 YES Gossypium darwinii ES288_A02G005100v1 2 YES Gossypium darwinii ES288_D02G001700v1 2 YES Gossypium hirsutum LOC107896756 2 YES Gossypium hirsutum LOC107914554 2 YES Gossypium hirsutum LOC107935966 2 YES Gossypium hirsutum LOC107903579 2 YES Gossypium mustelinum E1A91_D10G132600v1 2 YES Gossypium mustelinum E1A91_A10G165200v1 2 YES Gossypium mustelinum E1A91_D02G005900v1 2 YES Gossypium mustelinum E1A91_A02G005100v1 2 YES 270 Gossypium raimondii B456_011G129400 2 YES Gossypium raimondii B456_005G005800 2 YES Gossypium tomentosum ES332_D10G139500v1 2 YES Gossypium tomentosum ES332_A10G178500v1 2 YES Gossypium tomentosum ES332_A02G005200v1 2 YES Gossypium tomentosum ES332_D02G005900v1 2 YES Theobroma cacao TCM_019744 2 YES Arabis alpina AALP_AA5G141700 2 YES Arabis nemorensis ANE_LOCUS23250 2 YES Arabis nemorensis ANE_LOCUS15826 2 YES Arabis nemorensis ANE_LOCUS15790 2 YES Brassica rapa subsp. NA|M4D8N0 2 YES pekinensis Brassica rapa subsp. NA|M4CRM9 2 YES pekinensis Brassica napus BnaC07g32480D 2 YES Brassica napus BnaA03g57960D 2 YES Brassica napus BnaCnng40660D 2 YES Brassica oleracea var. NA|A0A0D3DSC3 2 YES oleracea Brassica oleracea var. NA|A0A0D3DE13 2 YES 271 oleracea Brassica oleracea var. NA|A0A0D3EIH8 2 YES oleracea Arabidopsis lyrata subsp. ARALYDRAFT_486888 2 YES lyrata Arabidopsis lyrata subsp. ARALYDRAFT_666001 2 YES lyrata Arabidopsis thaliana AT5G01800 2 YES Arabidopsis thaliana AT3G51730 2 YES Capsella rubella CARUB_v10001904mg 2 YES Capsella rubella CARUB_v10019623mg 2 YES Eutrema halophilum NA|E4MWI5 2 YES Eutrema salsugineum EUTSA_v10010716mg 2 YES Eutrema salsugineum EUTSA_v10014673mg 2 YES Noccaea caerulescens LE_TR12690_c0_g1_i1_g.41286 2 YES Noccaea caerulescens LC_TR4311_c0_g1_i1_g.15690 2 YES Noccaea caerulescens GA_TR12421_c0_g1_i1_g.39801 2 YES Noccaea caerulescens MP_TR8698_c0_g1_i1_g.27351 2 YES Noccaea caerulescens MP_TR15565_c0_g1_i1_g.44534 2 YES Noccaea caerulescens LC_TR7688_c0_g1_i1_g.27127 2 YES Noccaea caerulescens GA_TR10503_c0_g1_i1_g.34365 2 YES 272 Noccaea caerulescens LE_TR17411_c0_g1_i1_g.56298 2 YES Rosa chinensis RchiOBHm_Chr3g0460961 2 YES Prunus persica PRUPE_6G290000 2 YES Prunus dulcis ALMOND_2B028996 2 YES Malus domestica DVH24_036312 2 NA Malus baccata C1H46_040009 2 YES Trema orientale TorRG33x02_098860 2 YES Parasponia andersonii PanWU01x14_361630 2 YES Rhizophora mucronata NA|A0A2P2JI44 2 YES Populus alba D5086_0000056270 2 YES Populus trichocarpa POPTR_016G133400 2 NA Populus trichocarpa POPTR_006G107300 2 YES Juglans regia LOC108989981 2 YES Juglans regia LOC109019257 2 YES Fagus sylvatica FSB_LOCUS40270 2 NA Cucumis sativus Csa_4G331080 2 YES Cucumis melo var. makuwa E5676_scaffold127G001120 2 YES Cucumis melo var. makuwa E6C27_scaffold1166G00310 2 YES Cucumis melo LOC103502188 2 YES 273 Appendix F Additional Data Figure F-01. Subcellular localization of YFP tagged ABCB4. (A) N-terminal fusion (left) version YFP-ABCB4 and C-terminal fusion (right) version ABCB4-YFP. (B) Point mutation version ABCB4-Y1094A-YFP. Bar=20?m. 274 Figure F-02. ABCB4 responses to salt treatment. (A) ABCB4 promoter::ABCB4-CFP responses to mock treatment, 100mM NaCl, 100mM KCl and 400mM mannitol for 20min. Arrows point to the internalized signals. Bar=20?m. (B) Colocalization between ABCB4 promoter::ABCB4-YFP and fluorescent membrane dye FM4-64. Arrows point to the colocalized signals. Bar=20?m. (C) ABCB4 promoter::ABCB4-YFP responses to low concentration of salt. 5DAG seedlings were treated with 25mM NaCl. Bar=20?m. 275 Figure F-03. Internalization of ABCB4 in response to other chemical treatments. (A) ABCB4 promoter::ABCB4-YFP with glucose treatment for 20min. (B) ABCB4 promoter::ABCB4-YFP with mannitol treatment for 20min. (C) Chloride inhibitor 5- nitro-2-(3-phenylpropyl-amino) benzoic acid (NPPB) pre-treated ABCB4 promoter::ABCB4-YFP seedlings in response to salt treatment. Bar=20?m. 276 Figure F-04. ABCB4 responses to hydrogen peroxide treatment. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were treated for 20min. Bar=20?m. 277 Figure F-05. Internalization of ABCB4-YFP in response to NaCl treatment. (A) Heterogeneous responses to NaCl in different individuals. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were treated with 150mM NaCl for 20min. Bar=50?m. (B) Cell size distribution in each section in the 5DAG seedling root tip. 278 Figure F-06. ABCB4-YFP internalization events in different sections in the root tip in response to different salt concentration. (A) Numbers of internalized vesicles in each cell. (B) Density of internalized vesicles. Data from one 5 days after germination ABCB4 promoter::ABCB4-YFP seedling was shown. Similar patterns also showed in other four seedlings. 279 Figure F-07. NaCl pre-treated seedlings in response to salt treatment. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were treated with 150mM NaCl for 2hr (right) or no pretreatment (left), then treated with 150mM NaCl for 20min. 280 Figure F-08. Some plasma membrane markers in response to salt treatment. 5 days after germination seedlings were treated with 150mM NaCl for 20min. Bar=20?m. 281 Figure F-09. Colocalization between ABCB4 and some markers associated with clathrin-mediated endocytosis. (A) Colocalization between ABCB4 promoter::ABCB4- YFP with CLC-CFP. (B) Colocalization between ABCB4 promoter::ABCB4-YFP with DRP1c-CFP. (C) Colocalization between ABCB4 promoter::ABCB4-CFP with clathrin heavy chain HUB domain YFP-HUB1. (D) Clathrin-mediated endocytosis inhibitor 282 Tyrphostin A23 (TyrA23) pre-treated seedlings in response to salt treatment. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were pre-treated with TyrA23 for 60min and transferred to 100mM NaCl for 30min. Bar=20?m. Figure F-10. ABCB4 promoter::ABCB4-YFP seedlings pretreated with Latrunculin B (LatB) in response to salt treatment. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were pretreated with actin inhibitor LatB for 60min then treated with 100mM NaCl for 15min. Bar=20?m. 283 Figure F-11. ABCB4-YFP in response to gravitropism. (A) 5 days after germination vertical growing ABCB4 promoter::ABCB4-YFP seedlings were rotated for 90 degree in the dark (top) or remained vertical (bottom) for 6 hours. Bar=50?m. (B) Fluorescent intensity differences between two sides of the root (top/bottom or left/right). N=5. P>0.05 by Student? t-test. 284 Figure F-12. Upregulation of ABCB4-YFP by IAA and ACC. (A) 5 days after germination (DAG) ABCB4 promoter::ABCB4-YFP seedlings treated with IAA for 2 hr. (B) 5DAG ABCB4 promoter::ABCB4-YFP seedling treated with IAA or ACC for 5 hr. (C) Fluorescent intensity statistics in (B). Bar=20?m. N=5. P<0.05 by Student? t-test. 285 Figure F-13. ABCB4 in response to small organic acids. 5 days after germination ABCB4 promoter::ABCB4-YFP seedlings were treated with the corresponding chemicals for 1 hr. Concentration = 1?M. Bar=50?m. 286 Figure F-14. ABCB4 expression in response to ABA treatment. (A) 5 days after germination (DAG) ABCB4 promoter::ABCB4-YFP seedlings were pretreated with 50?M protein synthesis inhibitor cycloheximide (CHX) for 30min. Followed by 10?M ABA treatment for 1 hr. (B) ABCB4 in response to high concentration of ABA. 5DAG ABCB4 promoter::ABCB4-YFP seedlings were treated with 50?M or 1mM ABA. Bar=20?m. (C) RT-PCR of ABCB4 expression in Col, abi2-1 and abi5-1 in response to ABA treatment. 5DAG seedlings were treated with 2?M ABA for 2 days. ACTIN2 was chosen as the reference gene. Three biological replicates. 287 Figure F-15. ABCB4 promoter::ABCB4-GFP in abi4-1. 5 days after germination seedlings were imaged. Fluorescent intensity was summarized in the right chart. Bar=20?m. N=5. P<0.05 by Student? t-test. 288 Figure F-16. ABCB4, PIN2 and AUX1 in response to ABA. (A) Confocal microscopy images of 5 days after germination seedlings treated with 1?M ABA for 2 hr. (B) Root growth in response to ABA treatment in pgp4-1, pin2-1, aux1-7. 5 days after germination seedlings were transferred to new media supplemented with 1?M ABA. The ratio is root length in the corresponding day / root length in day 0. N=10. 289 Figure F-17. Colocalization between ABCB4 and some cellular markers. 5DAG seedlings were treated with 150mM NaCl for 30min. (A) Colocalization between ABCB4 promoter::ABCB4-YFP and trans-Golgi network marker SYP61-CFP. (B) colocalization between ABCB4 promoter::ABCB4-CFP and autophagy marker YFP- ATG8a. (C) Colocalization between ABCB4 promoter::ABCB4-YFP and plasma membrane associated SNARE protein CFP-SYP121. Arrows point to the colocalized signals. Bar=20?m. 290 Figure F-18. ABCB4 in response to nitrogen starvation. (A) ABCB4 promoter::ABCB4- YFP in response to low nitrogen treatment. ABCB4 promoter::ABCB4-YFP seedlings were grown on 1/4MS media for 5 days and then transferred to new media with the corresponding nitrogen supplies for another 2 days. Bar=20?m. (B) RT-PCR for ABCB4 transcript in response to low nitrogen. Wild type seedlings were grown on 1/4MS media for 5 days and then transferred to new media with the corresponding nitrogen supplies for another 2 days and harvested for RNA extraction. Three biological replicates. ACTIN2 as the reference gene. 291 References Adams, D.R., Ron, D. & Kiely, P.A. (2011). RACK1, A multifaceted scaffolding protein: Structure and function. Cell Commun Signal, 9: 22. Ahn, V.E., Faull, K.F., Whitelegge, J.P., Fluharty, A.L., Prive G.G. (2003). Crystal structure of saposin B reveals a dimeric shell for lipid binding. Proc. Natl. Acad. Sci. U.S.A., 100: 38?43. Ahn, V.E., Leyko, P., Alattia, J.R., Chen, L., Priv?, G.G. (2006). Crystal structures of saposins A and C. Protein Sci., 15: 1849?1857. An, C., Fukusaki, E. & Kobayashi, A. (2002). Aspartic proteinases are expressed in pitchers of the carnivorous plant Nepenthes alata Blanco. Planta, 214(5): 661-667. Asakura, T., Watanabe, H., Abe, K. and Arai, S. (1995). Rice aspartic proteinase, oryzasin, expressed during seed ripening and germination, has a gene organization distinct from those of animal and microbial aspartic proteinases. European Journal of Biochemistry, 232: 77?83. Atta, S., Maltese, S. and Cousin, R. (2004). Protein content and dry weight of seeds from various pea genotypes. Agronomie, 24(5): 257?266. Azuma, N., O'Brien, J. S., Moser, H. W., Kishimoto, Y. (1994). Stimulation of acid ceramidase activity by saposin D. Arch. Biochem. Biophys., 311: 354-357. Baud, S., Dubreucq, B., Miquel, M., Rochat, C. and Lepiniec, L. (2008). Storage reserve accumulation in Arabidopsis: metabolic and developmental control of seed filling. The Arabidopsis book, 6, e0113. 292 Berent, S. L., Radin, N. S. (1981). Mechanism of activation of glucocerebrosidase by co- ?-glucosidase (glucosidase activator protein). Biochim. Biophys. Acta, 664: 572-582. Borner, G.H., Sherrier, D.J., Weimar, T., Michaelson, L.V., Hawkins, N.D., Macaskill A., Napier, J.A., Beale, M.H., Lilley, K.S., Dupree P. (2005). Analysis of detergent-resistant membranes in Arabidopsis. Evidence for plasma membrane lipid rafts. Plant Physiol, 137:104?116. Bowman, J.L., Drews, G.N., Meyerowitz, E.M. (1991). Expression of the Arabidopsis floral homeotic gene AGAMOUS is restricted to specific cell types late in flower development. The Plant Cell, 3 (8):749-758. Bradford, K. and Bowley, J. (2003). Seeds: biology, technology and role in agriculture. In M. J. Chrispeels and D. E. Sadava, eds. Plants, Genes, and Crop Biotechnology. Sadbury: Jones & Bartlett Publishers, pp. 2112?2239. Bras, M., Queenan, B., Susin, S.A. (2005). Programmed cell death via mitochondria: different modes of dying. Biochemistry (Moscow), 70:231-239. Brodelius, M., Hiraiwa, M., Marttila, S., Al Karadaghi, S., Picaud, S. and Brodelius, P.E. (2005). Immunolocalization of the saposin-like insert of plant aspartic proteinases exhibiting saposin C activity. Expression in young flower tissues and in barley seeds. Physiologia Plantarum, 125: 405-418. Brodersen, P., Petersen, M., Pike, H.M., Olszak, B., Skov, S., Odum, N., Jorgensen, L.B., Brown, R.E., Mundy, J. (2002). Knockout of Arabidopsis accelerated-cell-death11 encoding a sphingosine transfer protein causes activation of programmed cell death 293 and defense. Genes Dev., 16 (4) :490-502. Bruhn, H. (2005). A Short-guided tour through functional and structural features of saposin-like proteins. Biochemical Journal, 389: 249-257. Bryksa, B.C., Bhaumik, P., Magracheva, E., De Moura, D.C., Kurylowicz, M., Zdanov, A., Dutcher, J.R., Wlodawer, A., Yada, R.Y. (2011). Structure and mechanism of the saposin- like domain of a plant aspartic protease. J. Biol. Chem.,286: 28265-. Bryksa, B.C., Grahame, D.A., Yada, R.Y. (2017). Comparative structure-function characterization of the saposin-like domains from potato, barley, cardoon and Arabidopsis aspartic proteases. Iochimica et Biophysica Acta (BBA) - Biomembranes, 1859: 1008-1018. Buono, A., Hudecek, R., Nowack, M. (2019). Plant proteases during developmental programmed cell death. JOURNAL OF EXPERIMENTAL BOTANY, 70(7): 2097?2112. Cacas, J.L., Bure, C., Grosjean, K., Gerbeau-Pissot, P., Lherminier, J., Rombouts, Y., Maes, E., Bossard, C., Gronnier, J., Furt, F., et al. (2016). Revisiting plant plasma membrane lipids in tobacco: a focus on sphingolipids. Plant Physiology, 170: 367-384. Castanheira, P., Samyn, B., Sergeant, K., Clemente, J.C., Dunn, B.M., Pires, E., Van Beeumen, J. and Faro, C. (2005). Activation, proteolytic processing, and peptide specificity of recombinant cardosin A. The Journal of biological chemistry, 280(13): 13047?13054. Chen, F., Foolad, M.R. (1997). Molecular organization of a gene in barley which encodes a protein similar to aspartic protease and its specific expression in nucellar 294 cells during degeneration. Plant Molecular Biology 35: 821-831. Chen, H., Huang, Y., Huang, G., Huang, S., Chow, T., Lin Y. (2015). Sweet potato SPAP1 is a typical aspartic protease and participates in ethephon-mediated leaf senescence. Journal of Plant Physiology, 180:1-17. Chen, M., Markham, J.E., Cahoon, E.B. (2012). Sphingolipid Delta8 unsaturation is important for glucosylceramide biosynthesis and low-temperature performance in Arabidopsis. The Plant Journal 69: 769-781. Chen, X., Pfeil, J.E. and Gal, S. (2002). The three typical aspartic proteinase genes of Arabidopsis thaliana are differentially expressed. European Journal of Biochemistry, 269: 4675-4684. Ciaffoni, F., Salvioli, R., Tatti, M., Arancia, G., Crateri, P., Vaccaro, A.M. (2001). Saposin D solubilizes anionic phospholipid-containing membranes. J. Biol. Chem., 276: 31583? 31589. Ciaffoni, F., Tatti, M., Boe, A., Salvioli, R., Fluharty, A., Sonnino, S., Vaccaro, A.M. (2006). Saposin B binds and transfers phospholipids. J. Lipid Res. 47: 1045?1053. Cochrane, C.G., and Revak, S.D. (1991). Pulmonary surfactant protein B (SP-B): structure?function relationships. Science, 254: 566?568. Cruz de Carvalho, M.H., d'Arcy-Lameta, A., Roy-Macauley, H., Gareil, M., El Maarouf, H., Pham-Thi, A. and Zuily-Fodil, Y. (2001). Aspartic protease in leaves of common bean (Phaseolus vulgaris L.) and cowpea (Vigna unguiculata L. Walp): enzymatic activity, gene expression and relation to drought susceptibility. FEBS Letters, 492. 295 D?Hondt, K., Bosch, D., Van Damme, J., Goethals, M., Vandekerckhove, J. and Krebbers, E. (1993). An aspartic proteinase present in seeds cleaves Arabidopsis 2S albumin precursors in vitro. The Journal of biological chemistry, 268(28): 20884?20891. D?Ipp?lito, S., Fiol, D.F., Daleo, G.R., Guevara, M.G. (2020). Overexpression of Arabidopsis aspartic protease APA1 gene confers drought tolerance. Plant Science, 292. Daneva, A., Gao, Z., Durme, M.V., Nowack, M.K. (2016). Functions and regulation of programmed cell death in plant development. Annual Review of Cell and Developmental Biology, 32: 441-468. Daniluk, P., Lesyng, B. (2011). A novel method to compare protein structures using local descriptors. BMC Bioinformatics 12: 344. Danon, A., Rotari, V.I., Gordon, A., Mailhac, N., Gallois, P. (2004). Ultraviolet-C overexposure induces programmed cell death in Arabidopsis, which is mediated by caspase-like activities and which can be suppressed by caspase inhibitors, P35 and Defender against Apoptotic Death. The Journal of Biological Chemistry, 279: 779-787. Day, C.A., Kenworthy, A.K. (2009). Tracking microdomain dynamics in cell membranes. Biochim. Biophys. Acta., 1788:245?253 De Alba, E., Weiler, S., Tjandra, N. (2003). Solution structure of human saposin C: pH- dependent interaction with phospholipid vesicles. Biochemistry, 42: 14729?14740. De Moura, D. C., Bryksa, B. C., & Yada, R. Y. (2014). In silico insights into protein-protein interactions and folding dynamics of the saposin-like domain of Solanum tuberosum aspartic protease. PloS one, 9(9). 296 Devaraj, K.B., Gowda, L.R., Prakash, V. (2008). An unusual thermostable aspartic protease from the latex of Ficus racemosa L.. Phytochemistry, 69(3): 647-655. Duarte, P., Pissarra, J., Moore, I. (2008). Processing and trafficking of a single isoform of the aspartic proteinase cardosin A on the vacuolar pathway. Planta, 227: 1255?1268 Egas, C., Lavoura, N., Resende, R., Brito, R. M., Pires, E., de Lima, M. C., Faro, C. (2000). The saposin-like domain of the plant aspartic proteinase precursor is a potent inducer of vesicle leakage. J. Biol. Chem., 275: 38190-38196. Egas, C., Lavoura, N., Resende, R., Brito, R.M.M., Pires, E., Pedroso de Lima, M.C., and Faro, C. (2000). The saposin-like domain of the plant aspartic proteinase precursor is a potent inducer of vesicle leakage. J. Biol. Chem., 275: 38190-. Egas, C., Lavoura, N., Resende, R., Brito, R.M., Pires, E., de Lima, M.C., Faro, C. (2000). The saposin-like domain of the plant aspartic proteinase precursor is a potent inducer of vesicle leakage. The Journal of biological chemistry, 275(49): 38190-38196. Endo, M., Shimizu, H., Nohales, M. et al. (2014). Tissue-specific clocks in Arabidopsis show asymmetric coupling. Nature, 515: 419?422. Fendrych, M., Hautegem, T.V., Durme, M.V., Olvera-Cariilo, Y., Huysmans, M., Karimi, M., Lippens, S., J.Guerin, C., Krebs M., Schumacher, K., Nowack, M. (2014). Programmed cell death controlled by ANAC033/SOMBRERO determines root cap organ size in Arabidopsis. Current Biology, 24: 931-940. Fendrych, M., Hautegem, TV., Durme, MV., Olvera-Cariilo, Y., Huysmans, M., Karimi, M., Lippens, S., J.Guerin, C., Krebs, M., Schumacher, K., Nowack, M. (2014). Programmed 297 cell death controlled by ANAC033/SOMBRERO determines root cap organ size in Arabidopsis. Current Biology, 24: 931-940. Fierens, K., Brijs, K., Courtin, C.M., Gebruers, K., Goesaert, H., Raedschelders, G., Robben, J., Van Campenhout., Volckaert, G. and Delcour, J.A. (2003). Molecular identification of wheat endoxylanase inhibitor TAXI-I1, member of a new class of plant proteins. FEBS Letters, 540 Figueiredo, R., Duarte, P., Pereira S., Pissarra J. (2006). The embryo sac of Cynara cardunculus: Ultrastructure of the development and localisation of the aspartic proteinase cardosin B. Sexual Plant Reproduction. 19: 93-101. Fraz?o, C., Bento, I., Costa, C., Soares, C.M., Verissimo, P., Faro, C., Pires, E., Cooper J., Carrondo, M.A. (1999). Crystal structure of cardosin A, a glycosylated and Arg-Gly-Asp- containing aspartic proteinase from the flowers of Cynara cardunculus L.. J. Biol. Chem., 274: 27694?27701. Frey, M.E., D'Ippolito, S., Pepe, A., Daleo, G.R., Guevara, M.G. (2018). Transgenic expression of plant-specific insert of potato aspartic proteases (StAP-PSI) confers enhanced resistance to Botrytis cinerea in Arabidopsis thaliana. Phytochemistry, 149: 1-11. Furt, F., Konig, S., Bessoule, J.J., Sargueil, F., Zallot, R., Stanislas, T., Noirot, E., Lherminier, J., Simon-Plas, F., Heilmann, I., et al. (2010). Polyphosphoinositides are enriched in plant membrane rafts and form microdomains in the plasma membrane. Plant Physiol, 152:2173?2187 298 Futerman, A.H., Hannun, Y.A. (2004). The complex life of simple sphingolipids. EMBO Reports, 5: 777?782. Gao, C., Zhuang, X., Cui, Y., et al. 2015. Dual roles of an Arabidopsis ESCRT component FREE1 in regulating vacuolar protein transport and autophagic degradation. Proc Natl Acad Sci U S A. ,112(6):1886?1891. Gao, Z., Daneva, A., Salanenka, Y. et al. (2018). KIRA1 and ORESARA1 terminate flower receptivity by promoting cell death in the stigma of Arabidopsis. Nature Plants, 4: 365? 375. Gepstein, S., Sabehi, G., Carp, M.-J., Hajouj, T., Nesher, M.F.O., Yariv, I., Dor, C. and Bassani, M. (2003). Large-scale identification of leaf senescence-associated genes. The Plant Journal, 36: 629-642. Glathe, S., Kervinen, J., Nimtz, M., Li, G.H., Tobin, G.J., Copeland, T.D., Ashford, D.A., Wlodawer, A. and Costa, J. (1998). Transport and activation of the vacuolar aspartic proteinase phytepsin in Barley (Hordeum vulgare L.). Journal of biological chemistry, 273(47): 31230?31236. Gottschalk, W. and Muller, H.P. (2012). Seed proteins: biochemistry, genetics, nutritive value. W. Gottschalk and H. P. Muller, eds., Springer Science & Business Media. Gruis, D., Schulze, J. and Jung, R. (2004). Storage protein accumulation in the absence of the vacuolar processing enzyme family of cysteine proteases. The Plant cell, 16(1): 270?290. Gruis, D., Selinger, D.A., Curran, J.M. and Jung, R. (2002). Redundant proteolytic 299 mechanisms process seed 184 storage proteins in the absence of seed-type members of the vacuolar processing enzyme family of cysteine proteases. The Plant cell, 14(11): 2863?2882. Guo, J., Wang, S., Valerius, O., Hall, H., Zeng, Q., Li, J.-F., Weston D.J., Brian, E. Ellis, B.E, Chen, J.-G. (2011). Involvement of Arabidopsis RACK1 in Protein Translation and Its Regulation by Abscisic Acid. Plant Physiology Jan, 155 (1): 370-383. Hara-Nishimura, I. and Hatsugai, N. (2011). The role of vacuole in plant cell death. Cell death and differentiation, 18(8): 1298?304. Hawkins, C.A., De Alba, E., Tjandra, N. (2005). Solution structure of human saposin C in a detergent environment. J. Mol. Biol., 346: 1381?1392. Hecht, O., Van Nuland, N. A., Schleinkofe,r K., Dingley, A. J., Bruhn, H., Leippe, M. and Grotzinger, J. (2004). Solution structure of the pore forming protein of Entamoeba histolytica. J. Biol. Chem., 279: 17834?17841 Hill, C.H., Read, R.J., Deane, J.E. (2015). Structure of human saposin A at lysosomal pH. Acta Crystallogr. F, 71: 895?900. Hinz, G., Hillmer, S., Baumer, M. and Hohl, I. (1999). Vacuolar storage proteins and the putative vacuolar sorting receptor BP-80 exit the Golgi apparatus of developing pea cotyledons in different transport vesicles. The Plant cell, 11(8): 1509?1524. Hiraiwa, M., O?Brien, J., Kishimoto, Y., Galdzicka, M., Fluharty, A.L., Ginns, E.I., Martin, B.M. (1993). Isolation, characterization, and proteolysis of human prosaposin, the precursor of saposins (Sphingolipid Activator Proteins). Archives of Biochemistry and 300 Biophysics, 304 (1): 110-116. Hong, Y.H., Lillehoj, H.S., Siragusa, G.R., Bannerman, D.D., Lillehoj, E.P. (2008). Antimicrobial activity of chicken NK-lysin against Eimeria sporozoite. Avian Disease, 52: 302305. Huang, J, Zhang, T, Linstroth, L, Tillman, Z, Otegui, M.S., Owen, H.A., Zhao, D. (2016). Control of anther cell differentiation by the small protein ligand TPD1 and its receptor EMS1 in Arabidopsis. PLoS Genet. 12(8):e1006147. Huby, E., Napier, J.A., Baillieul, F., Michaelson, L.V. and Dhondt-Cordelier, S. (2020). Sphingolipids: towards an integrated view of metabolism during the plant stress response. New Phytol, 225: 659-670. Huysmans, M., Buono, R.A., Skorzinski, N., Radio, M.C., Winter, F.D., Parizot, B., Mertens, J., Karimi, M., Fendrych, M., Nowack, M.K. (2018). NAC transcription factors ANAC087 and ANAC046 control distinct aspects of programmed cell death in the Arabidopsis columella and lateral root cap. The Plant Cell, 30(9): 2197-2213. Huysmans, M., Lema, S.A., Coll, N.S., Nowack, M.K. (2017). Dying two deaths - programmed cell death regulation in development and disease. Current Opinion in Plant Biology, 35: 37-44. Huysmans, M., Buono, RA., Skorzinski, N., Radio, MC., Winter FD., Parizot, B., Mertens, J., Karimi, M., Fendrych, M., Nowack, MK. (2018). NAC transcription factors ANAC087 and ANAC046 control distinct aspects of programmed cell death in the Arabidopsis columella and lateral root cap. The Plant Cell, 30(9): 2197-2213. 301 Imai, H., Ohnishi, M., Kinoshita, M., Kojima, M., Ito, S. (1995). Structure and distribution of cerebroside containing unsaturated hydroxy fatty acids in plant leaves. Bioscience, Biotechnology, and Biochemistry, 59: 1309?1313. Ines, C., Parra-Lobato, M.C., Paredes, M.A., Labrador, J., Gallardo, M., Saucedo-Garcia, M., Gavilanes-Ruiz, M., Gomez-Jimenez, M.C. (2018). Sphingolipid distribution, content and gene expression during olive-fruit development and ripening. Frontiers in Plant Science, 9: 28. Jiang, L., Phillips, T.E., Hamm, C.A., Drozdowicz, Y.M., Rea, P.A., Maeshima, M., Rogers, S.W. and Rogers, J.C. (2001). The protein storage vacuole: a unique compound organelle. The Journal of cell biology, 155(6): 991?1002. John, M., Wendeler, M., Heller, M., Sandhoff, K., Kessler, H. (2006). Characterization of human saposins by NMR spectroscopy. Biochemistry, 45: 5206?5216. Kang, S. J. and Cresswell, P. (2004). Saposins facilitate CD1d-restricted presentation of an exogenous lipid antigen to T cells. Nat. Immunol. 5: 175?181. Keinath, N.F., Kierszniowska, S., Lorek, J., Bourdais, G., Kessler, S.A., Shimosato-Asano, H., Grossniklaus, U., Schulze, W.X., Robatzek, S., Panstruga, R. (2010). PAMP (pathogen- associated molecular pattern)-induced changes in plasma membrane compartmentalization reveal novel components of plant immunity. J. Biol. Chem., 285:39140?39149. Kerr, J.F., Wyllie, A.H., Currie, A.R. (1972). Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. British Journal of Cancer, 26: 239-257. 302 Kervinen, J., Tobin, G.J., Costa, J., Waugh, D.S., Wlodawer, A., Zdanov, A. (1999). Crystal structure of plant aspartic proteinase prophytepsin: inactivation and vacuolar targeting. The EMBO journal, 18(14): 3947?3955. Kervinen, J., Tobin, G.J., Costa, J., Waugh, D.S., Wlodawer, A., Zdanov, A. (1999). Crystal structure of plant aspartic proteinase prophytepsin: inactivation and vacuolar targeting. EMBO. J., 18: 3947?3955. Kinoshita, T., Nishimura, M. and Hara-nishimura, I. (1995). Homologues of a vacuolar processing enzyme that are expressed in different organs in Arabidopsis thaliana. Plant molecular biology, 29: 81?89. Kishimoto, Y., Hiraiwa, M. and O?Brien, J. S. (1992). Saposins: structure, function, distribution, and molecular genetics. Journal of Lipid Research, 33: 1255-1268. Klein, A., Henseler, M., Klein, C., Suzuki, K., Harzer, K., Sandhoff, K. (1994). Sphingolipid activator protein D (sap-D) stimulates the lysosomal degradation of ceramide in vivo. Biochem. Biophys. Res. Commun., 200:1440-1448. Klime?ov?, J., Nobis, M.P., Herben, T. (2015). Senescence, ageing and death of the whole plant: morphological prerequisites and constraints of plant immortality. New Phytol. 206: 14?18. Koelsch, G., Mares, M., Metcalf, P., and Fusek, M. (1994). Multiple functions of pro- parts of aspartic proteinase zymogens. Federation of European Biochemical Society, 343: 6?10. Konrad, S.S.A., Popp, C., Stratil, T.F., Jarsch, I.K., Thallmair, V., Folgmann, J., Mar?n, M. 303 and Ott, T. (2014). S-acylation anchors remorin proteins to the plasma membrane but does not primarily determine their localization in membrane microdomains. New Phytol, 203: 758-769. Lagerholm, B.C., Weinreb, G.E., Jacobson, K., Thompson, N.L. (2005). Detecting microdomains in intact cell membranes. Annu. Rev. Phys. Chem., 56:309?336 Laloi, M., McCarthy, J., Morandi, O., Gysler, C. and Bucheli, P. (2002). Molecular and biochemical characterization of two aspartic proteinases TcAP1 and TcAP2 from Theobroma cacao seeds. Planta, 215: 754?762. Laloi, M., Perret, A.M., Chatre, L., Melser, S., Cantrel, C., Vaultier, M.N., Zachowski, A., Bathany, K., Schmitter, J.M., Vallet, M. et al. (2007). Insights into the role of specific lipids in the formation and delivery of lipid microdomains to the plasma membrane of plant cells. Plant Physiol, 143:461?472. Leippe, M., Bruhn, H., Hecht, O., Grotzinger, J. (2005). Ancient weapons: the three- dimensional structure of amoebapore A. Trends Parasitol 21:5?7. Lenarcic, T., Albert, I., Bohm, H., Hodnik, V., Pirc, K., Zavec, A.B., Podobnik, M., Pahovnik, D., Zagar, E., Pruitt R,., et al. (2017). Eudicot plant-specific sphingolipids determine host selectivity of microbial NLP cytolysins. Science, 358: 1431-1434. Leonova, T., Qi, X., Bencosme, A., Ponce, E., Sun, Y., and Grabowski, G.A. 1996. Proteolytic Processing Patterns of Prosaposin in Insect and Mammalian Cells. J. Biol. Chem., 271: 17312-. Li, X., Wang, X., Yang, Y., Li, R., He, Q., Fang, X., Lu, D.T., Maurel, C., Lin, J. (2011). Single- 304 molecule analysis of PIP2;1 dynamics and partitioning reveals multiple modes of Arabidopsis plasma membrane aquaporin regulation. The Plant Cell, 23:3780?3797. Liang, H., Yao, N., Song, J.T., Luo, S., Lu, H., Greenberg, J.T. (2003). Ceramides modulate programmed cell death in plants. Genes and Development, 17: 2636?2641. Liepinsh, E., Andersson, M., Ruysschaert, J. et al. Saposin fold revealed by the NMR structure of NK-lysin. Nat. Struct. Mol. Biol. 4: 793?795. Linke, T., Wilkening, G., Sadeghlar, F., Mozcall, H., Bernardo, K., Schuchman, E., Sandhoff, K. (2001). Interfacial regulation of acid ceramidase activity. Stimulation of ceramide degradation by lysosomal lipids and sphingolipid activator proteins. J. Biol. Chem., 276: 5760?5768. Lockshin, R.A., Zakeri, Z. (2004). Apoptosis, autophagy, and more. The International Journal of Biochemistry and Cell Biology, 36:2405-2419. Lufrano, D., Faro, R., Castanheira, P., Parisi, G., Ver?ssimo, P., Vairo-Cavalli, S., Sim?es, I., Faro, C. (2012). Molecular cloning and characterization of procirsin, an active aspartic protease precursor from Cirsium vulgare (Asteraceae). Phytochemistry, 81: 7- 18. Lufrano, D., Faro, R., Castanheira, P., Parisi, G., Ver?ssimo, P., Vairo-Cavalli, S., Sim?es, I., Faro, C., Heimgartner, U., Pietrzak, M., Geertsen, R., Brodelius, P., da Silva Figueiredo, A.C., Pais, M.S.S. (1990). Purification and partial characterization of milk clotting proteases from flowers of Cynara cardunculus. Phytochemistry, 29(5): 1405-1410. Lynch D.V., Dunn T.M. (2004). An introduction to plant sphingolipids and a review of 305 recent advances in understanding their metabolism and function. New Phytologist, 161: 677?702. Lynch, D.V., Steponkus, P.L. (1987). Plasma membrane lipid alterations associated with cold acclimation of winter rye seedlings (Secale cereale L. cv Puma). Plant Physiology, 83: 761?767. Markham, J.E., Jaworski, J.G. (2007). Rapid measurement of sphingolipids from Arabidopsis thaliana by reversed-phase high-performance liquid chromatography coupled to electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 21: 1304?1314. Markham, J.E., Li, J., Cahoon, E.B., Jaworski, J.G. (2006). Separation and identification of major plant sphingolipid classes from leaves. Journal of Bioogical Chemistry, 281: 22684?22694. Markham, J.E., Molino, D., Gissot, L., Bellec, Y., Hematy, K., Marion, J., Belcram, K., Palauqui, J.C., Satiat-Jeunemaitre, B., Faure, J.D. (2011). Sphingolipids containing very- long-chain fatty acids define a secretory pathway for specific polar plasma membrane protein targeting in Arabidopsis. The Plant Cell, 23 (6):2362-2378. Matsuda, J. (2008). Sphingolipid Activator Proteins. In: Taniguchi N., Suzuki A., Ito Y., Narimatsu H., Kawasaki T., Hase S. (eds) Experimental Glycoscience. Springer, Tokyo. pp. 125-129. Matsuda, J., Vanier, M.T., Saito, Y., Tohyama, J., Suzuki, K., Suzuki, K. (2001). A mutation in the saposin A domain of the sphingolipid activator protein (prosaposin) gene results 306 in a late-onset, chronic form of globoid cell leukodystrophy in the mouse. Human Molecular Genetics, 10(11): 1191-1199. Mazorra-Manzano, M.A., Yada, R.Y. (2008). Expression and characterization of the recombinant aspartic proteinase A1 from Arabidopsis thaliana. Phytochemistry, 69(13): 2439-2448 Mendieta, J.R., Pagano, M.R., Munoz, F.F., Daleo, G.R. & Guevara, M.G. (2006). Antimicrobial activity of potato aspartic proteases (StAPs) involves membrane permeabilization. Microbiology, 152: 2039-2047. Michaelson, L.V., Zauner, S., Markham, J.E., Haslam, R.P., Desikan, R., Mugford, S., Albrecht, S., Warnecke, D., Sperling, P., Heinz, E., et al. (2009). Functional characterization of a higher plant sphingolipid Delta4-desaturase: defining the role of sphingosine and sphingosine-1-phosphate in Arabidopsis. Plant Physiology, 149: 487- 498. Minami, A., Fujiwara, M., Furuto, A., Fukao, Y., Yamashita, T., Kamo, M., Kawamura, Y., Uemura, M. (2009). Alterations in detergent-resistant plasma membrane microdomains in Arabidopsis thaliana during cold acclimation. Plant and Cell Physiology, 50: 341-359. Misas-Villamil, J.C., Toenges, G., Kolodziejek, I., Sadaghiani, A.M., Kaschani, F., Colby, T., Bogyo, M. and Van der Hoorn, R.A.L. (2013). Activity profiling of vacuolar processing enzymes reveals a role for VPE during oomycete infection. The Plant journal, 73(4): 689?700. 307 Mochizuki, S., Harada, A., Inada, S., Sugimoto-Shirasu, K., Stacey, N., Wada, T., Ishiguro, S., Okada, K., Sakai, T. (2005). The Arabidopsis WAVY GROWTH 2 Protein Modulates Root Bending in Response to Environmental Stimuli. The Plant Cell Feb, 17 (2): 537- 547. Molino, D., Van der Giessen, E., Gissot, L., Hematy, K., Marion, J., Barthelemy, J., Bellec Y., Vernhettes, S., Satiat-Jeunemaitre, B., Galli, T., Tareste, D., Faur,e J.D. (2014). Inhibition of very long acyl chain sphingolipid synthesis modifies membrane dynamics during plant cytokinesis. Biochim. Biophys. Acta., 1842 (10):1422-1430. Mongrand, S., Morel, J., Laroche, J., Claverol, S., Carde, J.P., Hartmann, M.A., Bonneu, M., Simon-Plas, F., Lessire, R., Bessoule, J.J. (2004). Lipid rafts in higher plant cells: purification and characterization of Triton X-100-insoluble microdomains from tobacco plasma membrane. J. Biol. Chem., 279:36277?36286. Morimoto, S., Kishimoto, Y., Tomich, J., Weiler, S., Ohashi, T., Barranger, J. A., Kretz, K. A., O'Brien, J. S. (1990). Interaction of saposins, acidic lipids, and glucosylceramidase. J. Biol. Chem., 265:1933-1937 Morimoto, S., Martin, B. M., Yamamoto, Y., Kretz, K. A., O'Brien, J. S., Kishimoto, Y. (1989). Saposin A: second cerebrosidase activator protein. Proc. Natl. Acad. Sci. U.S.A., 86: 3389-3393. Mu?oz, F., Mendieta, J., Pagano, M., Paggi, R., Daleo, G., Guevara, M. (2010). The swaposin-like domain of potato aspartic protease (StAsp-PSI) exerts antimicrobial activity on plant and human pathogens. Peptides. 31: 777-85. 308 Mu?oz, F., Palomares-Jerez, M.F., Daleo, G., Villala?n, J., Guevara, M.G. (2011). Cholesterol and membrane phospholipid compositions modulate the leakage capacity of the swaposin domain from a potato aspartic protease (StAsp-PSI). Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, 1811(12): 1038-1044. Mu?oz, F. F., Mendieta, J. R., Pagano, M. R., Paggi, R. A., Daleo, G. R. & Guevara, M. G. (2010). The swaposin-like domain of potato aspartic protease (StAsp-PSI) exerts antimicrobial activity on plant and human pathogens. Peptides, 31(5): 777-785. Mu?oz, F., Palomares-Jerez, M. F., Daleo, G., Villala?n, J. & Guevara, M. G. (2014). Possible mechanism of structural transformations induced by StAsp-PSI in lipid membranes. Biochimica et Biophysica Acta - Biomembranes, 1838(1, Part B): 339-347. M?ntz, K. and Muntz, K. (1998). Deposition of storage proteins. Plant molecular biology, 38(1-2): 77?99. Murakami, S., Kondo, Y., Nakano, T., (2000). Protease activity of CND41, a chloroplast nucleoid DNA-binding protein, isolated from cultured tobacco cells. FEBS Letters 468: 15-18. Mutlu, A., Chen, X., Reddy, S.M. and Gal, S. (1999). The aspartic proteinase is expressed in Arabidopsis thaliana seeds and localized in the protein bodies. Seed science research, 9(1): 75?84. Mutlu, A. and Gal, S. (1999). Plant aspartic proteinases: enzymes on the way to a function. Physiologia Plantarum, 5: 569?576. Nakano, T., Murakami, S., Shoji, T., Yoshida, S., Ymada, Y., Sata, F., (1997). A novel 309 protein with DNA binding activity from tobacco chloroplast nucleoids. The Plant Cell 9: 1673-1682. Nakaune, S., Yamada, K., Kondo, M., Kato, T., Tabata, S. and Nishimura, M. (2005). A vacuolar processing enzyme, ?VPE , is involved in seed coat formation at the early stage of seed development. The Plant cell, 17(3): 876?887. O?Brien, J.S., Kishimoto, Y. (1991). Saposin protein: structure, function, and role in human lysosomal storage disorder. Official Journal of the Federation of American Societies for Experimental Biology, 5: 301-308. Olmeda, B., Garc?a-?lvarez, B. & P?rez-Gil, J. (2013). Structure?function correlations of pulmonary surfactant protein SP-B and the saposin-like family of proteins. Eur. Biophys. J. 42: 209?222. Olvera-Carrillo, Y., Bel, M.V., Hautegem, T.V., Fendrych, M., Huysmans, M., Simaskoya M., Durme, M.V., Buscaill, P., Rivas, S., Coll, N.S., Coppens, F., Maere, S., Nowack, M.K. (2015). A conserved core of programmed cell death indicator genes discriminates developmentally and environmentally induced programmed cell ceath in plants. Plant Physiology, 169 (4): 2684-2699. Olvera-Carrillo, Y., Van Bel, M., Van Hautegem, T., Fendrych, M., Huysmans, M., et al. (2015). A conserved core of programmed cell death indicator genes discriminates developmentally and environmentally induced programmed cell death in plants. Plant Physiol. 169: 2684-2699. Olvera-Carrillo, Y., Bel, MV., Hautegem, TV., Fendrych, M., Huysmans, M., Simaskoya, 310 M., Durme, MV., Buscaill, P., Rivas, S., Coll, NS., Coppens, F., Maere, S., Nowack, MK. (2015). A conserved core of programmed cell death indicator genes discriminates developmentally and environmentally induced programmed cell death in plants. Plant Physiology, 169 (4): 2684-2699. Otegui, M.S., Herder R., Schulze, J., Jung, R. and Staehelin, L.A. (2006). The proteolytic processing of seed storage proteins in Arabidopsis embryo cells starts in the multivesicular bodies. The Plant Cell, 18(10): 2567?2581. Otegui, MS., Herder, R., Schulze, J., Jung, R., Staehelin, LA. (2006). The proteolytic processing of seed storage proteins in Arabidopsis embryo cells starts in the multivesicular bodies. The Plant Cell, 18(10):2567-2581. Pagano, M. R., Mendieta, J. R., Mu?oz, F. F., Daleo, G. R. & Guevara, M. G. (2007). Roles of glycosylation on the antifungal activity and apoplast accumulation of StAPs (Solanum tuberosum aspartic proteases). International Journal of Biological Macromolecules, 41(5): 512-520. Paris, N., Stanley, C.M., Jones, R.L. and Rogers, .JC. (1996). Plant cells contain two functionally distinct vacuolar compartments. Cell, 85(4): 563?572. Pata M.O., Hannun Y.A., Ng C.K. (2010). Plant sphingolipids: decoding the enigma of the Sphinx. New Phytologist, 185: 611?630. Pena, S.V., and Krensky, A. M. (1997). Granulysin, a new human cytolytic granuleassociated protein with possible involvement in cell-mediated cytotoxicity. Semin. Immunol. 9: 117?125. 311 Pena, S.V. and Krensky, A.M. (1997). Granulysin, a new human cytolytic granuleassociated protein with possible involvement in cell-mediated cytotoxicity. Semin. Immunol., 9: 117?125 Pereira, C., Pereira S., Satiat-Jeunemaitre, B. and Pissarra, J. (2013). Cardosin A contains two vacuolar sorting signals using different vacuolar routes in tobacco epidermal cells. The Plant Journal, 76(1): 87-100. Pereira, C.S., de Costa, D.S., Pereira, S., Nogueira, F.M. Albuquerque, P.M., Teixeira, J., Faro, C., Pissarra, J. (2008). Cardosins in postembryonic development of cardoon: towards an elucidation of the biological function of plant aspartic proteinases. Protoplasma 232: 203?213. Petrov, V., Hille J., Mueller-Roeber, B., Gechev, T.S. (2015). ROS-mediated abiotic stress?induced programmed cell death in plants. Front. Plant Sci. 6: 69. Piatigorsky, J., Norman, B., Dishaw, L. J., Kos, L., Horwitz, J., Steinbach, P. J. and Kozmik, Z. (2001). J3-crystallin of the jellyfish lens: similarity to saposins. Proc. Natl. Acad. Sci., 98: 12362?12367. Popovic, K., Holyoake, J., Pom?s, R., Priv?, G.G. (2012). Structure of saposin a lipoprotein discs. Proc. Natl. Acad. Sci. USA, 109: 2908?2912. Popovic, K., Priv?, G.G. (2008). Structures of the human ceramide activator protein saposin D. Acta Crystallogr. D, 64: 589?594. Popovic, K., Priv?, G.G. (2008). Structures of the human ceramide activator protein saposin D. Acta Crystallogr. D, 64: 589?594. 312 Raimbault, A., Zuily-Fodil, Y., Soler, A., Cruz de Carvalho, M.H. (2013). A novel aspartic acid protease gene from pineapple fruit (Ananas comosus): Cloning, characterization and relation to postharvest chilling stress resistance. Journal of Plant Physiology, 170(17): 1536-1540. Reape, T.J., Molony, E.M., McCabe, P.F. (2008). Programmed cell death in plants: distinguishing between different modes. Journal of Experimental Botany, 59(3):435? 444. Richardt, S., Lang, D., Reski, R., Frank, W., & Rensing, S.A. (2007). PlanTAPDB, a phylogeny-based resource of plant transcription-associated proteins. Plant physiology, 143(4): 1452?1466. Rossmann, M., Schultz-Heienbrok, R., Behlke, J., Remmel, N., Alings, C., Sandhoff, K., Saenger, W., Maier, T. (2008). Crystal Structures of Human Saposins C and D: Implications for Lipid Recognition and Membrane Interactions. Structure, 16(5): 809- 817. Rossmann, M., Schultz-Heienbrok, R., Behlke, J., Remmel, N., Allings, C., Sandhoff, K., Saenger, W., Maier, T. (2008). Crystal structures of human saposins C and D: Implications for lipid recognition and membrane interactions. Structure, 16: 809?817. Sansen, S., De Ranter, C.J., Gebruers, K., Brijs, K., Courtin, C.M., Delcour, J.A., Rabijns, A. (2004). Structural Basis for Inhibition of Aspergillus niger Xylanase by Triticum aestivum Xylanase Inhibitor-I J. Biol. Chem., 279: 36022-. Sarkkinen, P., Kalkkinen, N., Tilgmann, C. et al. (1992). Aspartic proteinase from barley 313 grains is related to mammalian lysosomal cathepsin D. Planta 186: 317?323. Sarkkinen, P., Kalkkinen, N., Tilgmann, C. et al. (1992). Aspartic proteinase from barley grains is related to mammalian lysosomal cathepsin D. Planta, 186: 317?323. Sarmento, A.C., Lopes, H., Oliveira, C.S., Vitorino R., Samyn B., Sergeant K., Debyser G., Beeumen, J.V. Domingues, P., Amado, F., Pires, E., Domingues, M.R.M., Barros, M.T. (2009). Multiplicity of aspartic proteinases from Cynara cardunculus L.. Planta 230: 429?439. Schaaf, A., Reski, R., Decker, E.L. (2004). A novel aspartic proteinase is targeted to the secretory pathway and to the vacuole in the moss Physcomitrella patens. European Journal of Cell Biology, 83(4):145-152 Schuette, C. G., Pierstorff, B., Huettler, S. and Sandhoff, K. (2001). Sphingolipid activator proteins: proteins with complex functions in lipid degradation and skin biogenesis. Glycobiology, 11: 81R?90R. Shirakawa, M., Ueda, H., Shimada, T., Koumoto, Y., Shimada, T.L., Kondo, M., Takahashi, T., Okuyama, Y., Nishimura, M. and Hara-Nishimura, I. (2010). Arabidopsis Qa-SNARE SYP2 proteins localized to different subcellular regions function redundantly in vacuolar protein sorting and plant development. The Plant Journal, 64: 924-935. Simoes, I., Faro, C. (2004). Structure and function of plant aspartic proteinases. European Journal of Biochemistry, 271: 2067-2075 Simoes, I. & Faro, C. (2004). Structure and function of plant aspartic proteinases. The FEBS Journal, 271(11): 2067-2075. 314 Spassieva, S.D., Markham, J.E., Hille, J. (2002). The plant disease resistance gene Asc- 1 prevents disruption of sphingolipid metabolism during AAL-toxin-induced programmed cell death. Plant Journal, 32: 561-572. Spiegel, R., Bach, G., Sury, V., Mengistu, G., Meidan, B., Shalev, S., Shneor, Y., Mandel, H., Zeigler, M. (2005). A mutation in the saposin A coding region of the prosaposin gene in an infant presenting as Krabbe disease: first report of saposin A deficiency in humans. Molecular Genetics and Metabolism, 84(2):160-166. Staab, J. F., Ginkel, D. L., Rosenberg, G. B. and Munford, R. S. (1994). A saposin-like domain influences the intracellular localization, stability, and catalytic activity of human acyloxyacyl hydrolase. J. Biol. Chem., 269: 23736?23742 Stael, S., Breusegem, F.V., Gevaert, K., Nowack, M.K. (2019). Plant proteases and programmed cell death. Journal of Experimental Botany, 70:1991?1995. Swanson, S., Bethke, P. and Jones, R. (1998). Barley aleurone cells contain two types of vacuoles: characterization of lytic organelles by use of fluorescent probes. The Plant cell, 10(5): 685?698. Takahashi, D., Imai, H., Kawamura, Y., Uemura, M. (2016). Lipid profiles of detergent resistant fractions of the plasma membrane in oat and rye in association with cold acclimation and freezing tolerance. Cryobiology, 72: 123?134. Takahashi, D., Imai, H., Kawamura, Y., Uemura, M. (2016). Lipid profiles of detergent resistant fractions of the plasma membrane in oat and rye in association with cold acclimation and freezing tolerance. Cryobiology, 72: 123?134. 315 Tamargo, R.J., Velayati, A., Goldin, E., Sidransky, E. (2012). The role of saposin C in Gaucher disease. Molecular Genetics and Metabolism, 106(3):257-263. Tayama, M., Soeda, S., Kishimoto, Y., Martin, B. M., Callahan, J. W., Hiraiwa, M., O'Brien, J. S. (1993). Effect of saposins on acid sphingomyelinase. Biochem. J., 290: 401-404 Terauchi, K., Asakura, T., Ueda, H., Tamura, T., Tamura, K., Matsumoto, I., Misaka, T., Hara-Nishimura, I., Abe, K. (2006). Plant-specific insertions in the soybean aspartic proteinases, soyAP1 and soyAP2, perform different functions of vacuolar targeting. Journal of Plant Physiology, 163(8): 856-862. Titapiwatanakun, B., Blakeslee, J.J., Bandyopadhyay, A., Yang, H., Mravec, J., Sauer, M., Cheng, Y., Adamec, J., Nagashima, A., Geisler, M., Sakai, T., Friml, J., Peer, W.A. and Murphy, A.S. (2009). ABCB19/PGP19 stabilizes PIN1 in membrane microdomains in Arabidopsis. The Plant Journal, 57: 27-44. Tormakangas, K., Hadlington, J.L., Pimpl, P., Hillmer, S., Brandizzi, F., Teeri, T.H. & Denecke, J. (2001). A vacuolar sorting domain may also influence the way in which proteins leave the endoplasmic reticulum. The Plant Cell, 13(9): 2021-2032. Tse, Y.C., Wang, J., Miao, Y. and Jiang, L. (2007). Biogenesis of the compound seed protein storage vacuole. In Seeds: biology, development and ecology. pp. 112?119. Vaccaro, A.M., Ciaffoni, F., Tatti, M., Salvioli, R., Barca, A., Tognozzi, D., Scerch, C. (1995). pH-dependent conformational properties of saposins and their interactions with phospholipid membranes. J. Biol. Chem., 27: 30576?30580. Vieira, M., Pissarr, J., Ver?ssimo, P., Castanheira, P., Costa, Y., Pires, E., Faro, C. (2001). 316 Molecular cloning and characterization of cDNA encoding cardosin B, an aspartic proteinase accumulating extracellularly in the transmitting tissue of Cynara cardunculus L.. Plant Mol Biol 45: 529?539. Vieira, M., Pissarra, J., Ver?ssimo, P., Castanheira, P., Costa, Y., Pires, E. & Faro, C. (2001). Molecular cloning and characterization of cDNA encoding Cardosin B, an aspartic proteinase accumulating extracellularly in the transmitting tissue of Cynara cardunculus L. Plant Molecular Biology, 45(5): 529-539. Vieira, V., Peixoto, B., Costa, M., Pereira, S., Pissarra, J., Pereira, S. (2019). N-linked glycosylation modulates Golgi-independent vacuolar sorting mediated by the plant specific insert. Plants, 8(9):312. Vieira, V., Peixoto, B., Costa, M., Pereira, S., Pissarra, J., Pereira, C. (2019). N-Linked glycosylation modulates Golgi-independent vacuolar sorting mediated by the plant specific insert. Plants, 8: 312. Vogel, A., Schwarzmann, G., Sandhoff, K. (1991). Glycosphingolipid specificity of the human sulfatide activator protein. Eur. J. Biochem., 200: 591-597 Warnecke, D., Heinz, E. (2003). Recently discovered functions of glucosylceramides in plants and fungi. Cellular and Molecular Life Sciences, 60: 919?941. Wenger, D. A., Sattler, M., Roth, S. (1982). A protein activator of galactosylceramide ?- galactosidase. Biochim. Biophys. Acta, 712: 639-649 Wilkins, K.A., Poulter, N.S., Franklin-Tong, V.E. (2014). Taking one for the team: self- recognition and cell suicide in pollen. J. Exp. Bot. 65: 1331?1342. 317 Winau, F., Schwierzeck, V., Hurwitz, R., Remmel, N., Sieling, P. A., Modlin, R. L., Porcelli S. A., Brinkmann, V., Sugita, M., Sandhoff, K. et al. (2004). Saposin C is required for lipid presentation by human CD1b. Nat. Immunol. 5: 169?174. Winter, D., Vinegar, B., Nahal, H., Ammar, R., Wilson, G.V. and Provart, N.J. (2007). An ?Electronic Fluorescent Pictograph? browser for exploring and analyzing large-scale biological data sets. PloS ONE, 2(8), e718. Worrall, D., Ng, C.K.Y., Hetherington, A.M. (2003). Sphingolipids, new players in plant signaling. Trends in Plant Science, 8: 317?320. Wrobel, R., Jones, B.L. (1992). Appearance of Endoproteolytic Enzymes during the Germination of Barley. Plant Physiology, 100 (3): 1508-1516. Wrobel, R., Jones, B.L. (1992). Appearance of endoproteolytic enzymes during the germination of barley. Plant Physiology, 100 (3) :1508-1516. Wu, L., Chen, H., Curtis, C., Fu, Z.Q. (2014). Go in for the kill: how plants deploy effector- triggered immunity to combat pathogens. Virulence, 5: 710?721. Xiang, L., Etxeberria, E. and Van den Ende, W. (2013). Vacuolar protein sorting mechanisms in plants. FEBS journal, 280(4), 979?993. Xuan, W., Band, L.R., Kumpf, R.P., Van Damme, D., Parizot, B., De Rop, G., Opdenacker, D., M?ller, B.K., Skorzinski, N., Njo, M.F., De Rybel, B., Audenaert, D., Nowack, M.K., Vanneste, S., Beeckman, T. (2016). Cycles of programmed cell death establish the developmental clock in plant roots. Science, 351: 384-387. Yang, H., Richter, G.L., Wang, X., M?odzi?ska, E., Carraro, N., Ma, G., Jenness, M., Chao, 318 D.-Y., Peer, W.A. and Murphy, A.S. (2013). Sterols and sphingolipids differentially function in trafficking of the Arabidopsis ABCB19 auxin transporter. Plant. J., 74: 37-47. Yao X., Xiong W., Ye T., Wu Y. (2012). Overexpression of the aspartic protease ASPG1 gene confers drought avoidance in Arabidopsis. Journal of Experimental Botany, 63(7):2579?2593. Ye, H., Ren, F., Guo, H., Guo, L., Bai, J., Wang, Y. (2020). Identification of key genes and transcription factors in ageing Arabidopsis papilla cells by transcriptome analysis. Plant Physiology and Biochemistry, 147:1-9. Ye, H., Ren, F., Guo, H., Guo, L., Bai, J., Wang, Y. (2020). Identification of key genes and transcription factors in ageing Arabidopsis papilla cells by transcriptome analysis. Plant Physiology and Biochemistry, 147:1-9. Yu, M., Cui, Y., Zhang, X. et al. (2020). Organization and dynamics of functional plant membrane microdomains. Cell. Mol. Life Sci., 77: 275?287. Yuan, W., Qi, X., Tsang, P., Kang, S.J., Ilarionov, P.A., Besra, G.S., Gumperz, J., Cresswell, P. (2007). Saposin B is the dominant saposin that facilitates lipid binding to human CD1d molecules. Proc. Natl. Acad. Sci. USA, 104: 5551?5556. Zhou, D., Cantu 3rd, C., Sagiv, Y., Schrantz, N., Kulkarni, A. B., Qi, X., Mahuran, D. J., Morales, C. R., Grabowski, G. A., Benlagha, K. et al. (2004). Editing of CD1d-bound lipid antigens by endosomal lipid transfer proteins. Science, 303: 523?527. Zouhar, J., Munoz, A. and Rojo, E. (2010). Functional specialization within the vacuolar sorting receptor family: VSR1, VSR3 and VSR4 sort vacuolar storage cargo in seeds and 319 vegetative tissues. Plant J. 64: 577?588. 320