Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic analysis of the role of cellulose in Arabidopsis seed coat development and mucilage adherence Griffiths, Jonathan Stewart 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_fall_griffiths_jonathan.pdf [ 144.25MB ]
Metadata
JSON: 24-1.0073854.json
JSON-LD: 24-1.0073854-ld.json
RDF/XML (Pretty): 24-1.0073854-rdf.xml
RDF/JSON: 24-1.0073854-rdf.json
Turtle: 24-1.0073854-turtle.txt
N-Triples: 24-1.0073854-rdf-ntriples.txt
Original Record: 24-1.0073854-source.json
Full Text
24-1.0073854-fulltext.txt
Citation
24-1.0073854.ris

Full Text

 Genetic analysis of the role of cellulose in Arabidopsis seed coat development and mucilage adherence  by   Jonathan Stewart Griffiths   B.Sc., Carleton University, 2005 M.Sc., The University of Western Ontario, 2007   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  The Faculty of Graduate Studies (Botany)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  May 2013  © Jonathan Stewart Griffiths, 2013  ii Abstract Primary plant cell walls are comprised largely of the polysaccharides cellulose, hemicellulose, and pectins, and can also contain up to 10% protein. These cell wall components interact non-covalently and covalently to form a functional cell wall. Interactions between cellulose and pectins are poorly understood, and are the focus of this research. Arabidopsis seed coat epidermal cells produce three distinct types of cell walls: an outer primary wall; mucilage, a specialized wall composed primarily of pectins; and a rigid columella. When seeds are hydrated, mucilage expands rapidly, breaking the outer wall, to form a mucilage halo that surrounds and remains adherent to the seed. The columella appears to be composed primarily of cellulose, and is therefore an excellent model for investigating cellulose biosynthesis. Cell wall biosynthesis and polysaccharide interactions were examined during seed coat development and in mucilage adherence to better understand cell wall assembly and function. Cellulose is synthesized by the CELLULOSE SYNTHASE A (CESA) family of glucosyltransferases. It has been proposed that at least three different CESAs are required to form a functional Cellulose Synthase Complex (CSC). I investigated the contribution of CESA2, CESA5 and CESA9 in cellulose biosynthesis during seed coat development. Based on seed coat epidermal cell morphology and cellulose quantification, all three CESAs have non-redundant roles in secondary wall biosynthesis, while CESA5 specifically functions in mucilage biosynthesis. CESA3 is expressed in the seed coat during mucilage biosynthesis and missense mutations in CESA3, isoxaben resistant 1 (ixr1-1 and ixr1-2), result in altered mucilage structure and pectin distribution, and reduced cellulose amounts in seeds. The mechanism of mucilage adherence was examined by comparing two loss of function mutants that disrupt adherence, cesa5-1 and sos5-2. SOS5 encodes an arabinogalactan protein  iii hypothesized to influence adherence through CESA5. However, the phenotype of each single mutant differs and a cesa5 sos5 double mutant has an enhanced phenotype. Therefore, it is unlikely that SOS5 promotes mucilage adherence through CESA5. SOS5 may influence mucilage structure through galactans, as it is required for the proper function of the !- galactosidase, MUM2. This demonstrates a role for AGPs in galactan metabolism and cell wall polysaccharide interactions.   !  iv Preface  This thesis contains 4 research chapters intended to be published in peer-reviewed journals. Chapter 3 is published in two separate manuscripts. Stork, J., Harris, D., Griffiths, J., Williams, B., Beisson, F., Li-Beisson, Y., Mendu, V., Haughn, G., & DeBolt, S. (2010). CELLULOSE SYNTHASE9 serves a nonredundant role in secondary cell wall synthesis in Arabidopsis epidermal testa cells. Plant Physiol., 153(2), 580-9. Mendu, V, Griffiths, J.S., Persson, S., Stork, J., Downie, A.B., Voiniciuc, C., Haughn, G.W., & DeBolt, S. (2011). Subfunctionalization of cellulose synthases in seed coat epidermal cells mediates secondary radial wall synthesis and mucilage attachment. Plant Physiol., 157(1), 441-53.  Chapter 3: Jonathan Griffiths designed parts of the research project, performed the research, analyzed data and helped prepare two manuscripts (Stork et al., 2010; Mendu et al., 2011). This research was performed in collaboration with the DeBolt lab at the University of Kentucky. For the analysis of CESA9 (Stork et al., 2010), Jonathan Griffiths performed the histochemical analysis of developing cesa9 seeds, while Jozsef Stork and Darby Harris and Seth DeBolt analyzed the surface of seeds using SEM and performed chemical analysis. The manuscript was prepared and edited by Jozsef Stork, Darby Harris, Jonathan Griffiths, George Haughn and Seth DeBolt. For the analysis of CESA2, CESA5 and CESA9 (Mendu et al., 2011), Jonathan Griffiths designed the research with Venugopal Mendu, George Haughn and Seth DeBolt. Staffan Persson, Jozsef Stork, Bruce Downie and Catalin Voiniciuc assisted in performing the research.  v The manuscript was prepared by Jonathan Griffiths, Venugopal Mendu, George Haughn and Seth DeBolt. Published data including figures and tables are reprinted with the permission of the American Society of Plant Biologists.  Chapter 4: Jonathan Griffiths performed the research. Jonathan Griffiths and George Haughn designed the research project, analyzed the data and prepared the manuscript.  Chapter 5: Jonathan Griffiths performed the research. Catalin Voiniciuc assisted in performing the research. Jonathan Griffiths and George Haughn designed the research project, analyzed the data and prepared the manuscript.  Chapter 6: Jonathan Griffiths performed the research. Jonathan Griffiths and George Haughn designed the research project, analyzed the data and prepared the manuscript.   vi Table of contents Abstract ................................................................................................................... ii! Preface ..................................................................................................................... iv! Table of contents .................................................................................................... vi! List of tables ............................................................................................................. x! List of figures .......................................................................................................... xi! List of abbreviations ............................................................................................ xiv! Acknowledgements ............................................................................................... xvi! Dedication .......................................................................................................... xviii! 1! Introduction ........................................................................................................ 1! 1.1! The cell wall ...................................................................................................................... 1! 1.1.1! Composition and structure of the primary cell wall ................................................... 3! 1.1.2! Composition and structure of secondary cell walls .................................................... 6! 1.2! Structure and biosynthesis of cell wall polysaccharides .............................................. 6! 1.2.1! Cellulose ..................................................................................................................... 6! 1.2.2! Pectins ....................................................................................................................... 11! 1.2.3! Hemicelluloses .......................................................................................................... 14! 1.2.4! Proteins in the cell wall ............................................................................................ 15! 1.3! The Arabidopsis seed coat ............................................................................................ 19! 1.3.1! Development and structure of the seed coat ............................................................. 19! 1.3.2! Seed coat epidermal cell development ..................................................................... 21! 1.4! The structure and composition of seed coat mucilage ................................................ 22! 1.4.1! Mucilage composition .............................................................................................. 23! 1.4.2! Genetic analysis of mucilage biosynthesis and structure ......................................... 25! 1.5! Research goals and hypothesis ..................................................................................... 29! 1.5.1! Chapter 3: CESA2, CESA5 and CESA9 are involved in secondary cell wall biosynthesis, yet only CESA5 is required for mucilage biosynthesis .................................. 30! 1.5.2! Chapter 4: CESA3 participates in mucilage cellulose biosynthesis ......................... 31! 1.5.3! Chapter 5: CESA5 and SOS5 mediate mucilage adherence independently ............. 31! 1.5.4! Chapter 6: SOS5 is required for MUM2 function in mucilage ................................ 32! 1.6! Summary ........................................................................................................................ 32! 2! Materials and methods .................................................................................... 33! 2.1! Plant materials and growth conditions ........................................................................ 33! 2.1.1! Genotypic and phenotypic screening of mutant lines ............................................... 33!  vii 2.2! Statistical analysis .......................................................................................................... 35! 2.3! Developmental analysis of the seed coat epidermis .................................................... 36! 2.4! Microscopy ..................................................................................................................... 37! 2.4.1! Transmitted light microscopy ................................................................................... 37! 2.4.2! Confocal microscopy ................................................................................................ 37! 2.4.3! Scanning electron microscopy .................................................................................. 40! 2.5! Chemical analysis .......................................................................................................... 40! 2.5.1! Determination of monosaccharide composition by HPAEC .................................... 40! 2.5.2! Crystalline cellulose quantification .......................................................................... 42! 2.6! Bioinformatic analysis ................................................................................................... 43! 2.6.1! Analysis of gene and protein structure ..................................................................... 43! 2.6.2! Transcript analysis .................................................................................................... 43! 3! CESA2, CESA5 and CESA9 are involved in secondary cell wall biosynthesis in seed coat epidermal cells, yet only CESA5 is required for mucilage biosynthesis ............................................................................................ 44! 3.1! Introduction ................................................................................................................... 44! 3.1.1! Cellulose biosynthesis .............................................................................................. 45! 3.2! Results ............................................................................................................................. 48! 3.2.1! Phenotype of CESA9 T-DNA insertion lines ........................................................... 48! 3.2.2! Epidermal cell morphology is altered in cesa9 mutant seeds ................................... 48! 3.2.3! Isolation of CESA2, CESA5, and CESA9 mutant lines for the analysis of cellulose biosynthesis during seed coat development. ......................................................................... 54! 3.2.4! Radial walls are absent or reduced in cesa2-1 cesa5-1 cesa9-1 seeds ..................... 55! 3.2.5! Crystalline cellulose is reduced in CESA6-like mutants ........................................... 60! 3.2.6! Reduced mucilage adherence in cesa5-1 seeds ........................................................ 62! 3.2.7! Reduced mucilage adherence and cellulose staining in cesa5-1 seeds ..................... 62! 3.2.8! CESA5 is required for mucilage adherence .............................................................. 64! 3.3! Discussion ....................................................................................................................... 66! 3.3.1! CESA6-like CESAs involvement during columella biosynthesis ............................. 66! 3.3.2! CSC organization during mucilage and columella biosynthesis .............................. 68! 3.3.3! Role of cellulose in mucilage ................................................................................... 71! 4! Chapter 4: CESA3 participates in mucilage cellulose biosynthesis ............ 74! 4.1! Introduction ................................................................................................................... 74! 4.1.1! Primary wall CESAs ................................................................................................. 74! 4.2! Results ............................................................................................................................. 78! 4.2.1! CESA3 and CESA10 are highly expressed during mucilage biosynthesis ................ 78! 4.2.2! CESA5::GFP and CESA3::GFP are localized in the cytoplasmic column during mucilage biosynthesis ........................................................................................................... 81!  viii 4.2.3! ixr1-1 and ixr1-2 seeds have altered mucilage structure and adherent halo properties  .................................................................................................................................. 81! 4.2.4! ixr1-1 and ixr1-2 mutations result in increased disorganization of mucilage structure and cell wall polysaccharides ............................................................................................... 88! 4.2.5! Mucilage hydration properties in ixr1 mutants modified by calcium availability .... 96! 4.2.6! ixr1-1 and ixr1-2 have reduced monosaccharide amounts in mucilage and whole seeds  .................................................................................................................................. 99! 4.2.7! Reduced crystalline cellulose amounts in ixr1-1 and ixr1-2 whole seeds .............. 101! 4.3! Discussion ..................................................................................................................... 103! 4.3.1! CESA expression during seed coat development .................................................... 103! 4.3.2! The role of other CESAs during seed coat development ........................................ 104! 4.3.3! Effects of ixr1-1 and ixr1-2 missense mutations on cellulose biosynthesis ........... 105! 4.3.4! ixr1-1 and ixr1-2 mutations alter mucilage structure ............................................. 107! 5! CESA5 and SOS5 function independently to mediate mucilage adherence.  ......................................................................................................................... 110! 5.1! Introduction ................................................................................................................. 110! 5.1.1! Arabinogalactan proteins ........................................................................................ 110! 5.2! Results ........................................................................................................................... 113! 5.2.1! sos5 and cesa5 have distinct and additive effects on seed mucilage adherence ..... 113! 5.2.2! Mucilage structure is altered in cesa5-1 and sos5-2 seeds ..................................... 116! 5.2.3! CESA5 and SOS5 are both required for normal columella deposition .................. 123! 5.2.4! SOS5 influences columella morphology through the cytoplasmic column ............ 125! 5.2.5! SOS5 does not affect total glucose amounts in seeds ............................................. 130! 5.3! Discussion ..................................................................................................................... 137! 5.3.1! Mechanism of SOS5 function in mucilage ............................................................. 137! 5.3.2! The function of SOS5 in determining columella structure. .................................... 139! 6! Role of galactans in mucilage adherence ..................................................... 141! 6.1! Introduction ................................................................................................................. 141! 6.1.1! The functions of galactans in mediating adherence ................................................ 141! 6.2! Results ........................................................................................................................... 143! 6.2.1! CESA5 has no affect on the mucilage phenotype of mum2 ................................... 143! 6.2.2! SOS5 is required for MUM2 function .................................................................... 145! 6.2.3! Quantification of mucilage release in cesa5-1 mum2-1 and sos5-2 mum2-1 double mutants ................................................................................................................................ 146! 6.3! Discussion ..................................................................................................................... 150! 6.3.1! Relationship between MUM2 and CESA5 ............................................................. 151! 6.3.2! Relationship between SOS5 and MUM2 ................................................................ 152! 6.3.3! Galactan function in adherence .............................................................................. 152! 6.3.4! Mechanism of mucilage extrusion and adherence .................................................. 154!  ix 7! Conclusions and future directions ................................................................ 156! 7.1! Overview ....................................................................................................................... 156! 7.1.1! Cellulose in the seed coat ....................................................................................... 156! 7.1.2! Two independent cell wall networks ...................................................................... 160! 7.2! Future directions .......................................................................................................... 161! 7.2.1! Cellulose biosynthesis in the Arabidopsis seed coat .............................................. 161! 7.2.2! Transcriptional regulation of cellulose biosynthesis .............................................. 162! 7.2.3! Function of SOS5 ................................................................................................... 162! 7.3! Summary ...................................................................................................................... 162! References ............................................................................................................ 164!     x List of tables  Table 2.1: List of mutant lines, genomic locus, and genotyping primers used in this study 34 Table 2.2: List of washes and antibodies used in immunolabelling ........................................ 39 Table 3.1: HPLC analysis of mucilage and cell wall monosaccharide composition from wild type and cesa mutant seeds ......................................................................................................... 65   xi List of figures  Figure 1.1: Primary cell wall architecture, structure and biosynthesis.  .................................. 4 Figure 1.2: Cellulose biosynthesis and CSC assembly ............................................................... 8 Figure 1.3: Mucilage secretory cell development, structure and composition. ...................... 20 Figure 1.4: Regulation of mucilage biosynthesis and seed coat development. ....................... 28 Figure 3.1: Diagram of Arabidopsis CESA relationship and functional grouping ............... 46 Figure 3.2: Wild type and cesa9-1 SEM images of seed coat epidermal cell morphology. ... 49 Figure 3.3: Wild type and cesa9-1 seed epidermal cell size and columella area .................... 51 Figure 3.4. Wild type and cesa9 seed coat epidermal cell development and radial wall height. ............................................................................................................................................ 52 Figure 3.5: Epidermal cell shape and morphology of cesa mutants ....................................... 56 Figure 3.6: Developmental analysis of wild type and cesa mutant seed coats ....................... 57 Figure 3.7: Radial wall height of cesa mutant lines. ................................................................. 59 Figure 3.8: Acid insoluble cellulose content of cesa mutant seeds. ......................................... 61 Figure 3.9: Mucilage hydration and adherence phenotype of cesa5, cesa2 cesa9, and cesa2 cesa5 cesa9 seeds .......................................................................................................................... 63 Figure 4.1: CESA3 protein structure and amino acid changes caused by the ixr1-1 and ixr1-2 point mutations. ................................................................................................................ 76 Figure 4.2: Expression of the CESA gene family during seed development .......................... 79 Figure 4.3: Localization of CESA3, CESA5 and CESA6 in seed coat epidermal cells ......... 82 Figure 4.4: Ruthenium red staining of cesa mutant seeds. ...................................................... 84  xii Figure 4.5: Initial mucilage hydration halo and unstained mucilage halo ............................. 85 Figure 4.6: S4B stained cellulose in cesa5-1, ixr1-1 and ixr1-2 adherent mucilage ............... 87 Figure 4.7: Ray length in S4B stained mucilage ....................................................................... 89 Figure 4.8: CBM3a labelling of adherent mucilage.  ............................................................... 91 Figure 4.9: CBM28 labelling of adherent mucilage .................................................................. 92 Figure 4.10: JIM5 labelling of adherent mucilage .................................................................... 94 Figure 4.11: CCRC-M36 labelling of adherent mucilage ........................................................ 95 Figure 4.12: Effects of calcium on ixr1-1 and ixr1-2 mucilage hydration .............................. 98 Figure 4.13: Monosaccharide composition of non-adherent mucilage and whole seed AIR   ............................................................................................................................................ 100 Figure 4.14: Crystalline cellulose amounts in ixr1-1 and ixr1-2 whole seeds ....................... 102 Figure 5.1: Diagram of the protein structure of SOS5/FLA4 and a phylogenetic tree illustrating the relation of all AGP proteins in the FLA family.  .......................................... 112 Figure 5.2: Mucilage phenotype of cesa5-1, sos5-2 and cesa5 sos5 double mutants ........... 115 Figure 5.3: CBM3a labelling of adherent mucilage.  ............................................................. 118 Figure 5.4: CBM28 labelling of adherent mucilage.  ............................................................. 120 Figure 5.5: JIM5 labelling of adherent mucilage.  ................................................................. 121 Figure 5.6: CCRC-M36 labelling of adherent mucilage. ....................................................... 122 Figure 5.7: Columella formation is affected in cesa5 sos5 double mutants  ......................... 124 Figure 5.8: Live cell imaging of columella development in cesa5-1, sos5-2 and cesa5 sos5 double mutant seed coat epidermal cells ................................................................................. 127 Figure 5.9: Columella development as determined by fixed, embedded seeds sections ...... 128  xiii Figure 5.10: Monosaccharide composition of non-adherent mucilage and whole seeds  .... 131 Figure 5.11: Crystalline cellulose amounts in cesa5-1, sos5-2 and cesa5 sos5 double mutant whole seeds ................................................................................................................................. 136 Figure 6.1: Mucilage phenotype of ruthenium red stained cesa5 mum2 and sos5 mum2 double mutants.  ......................................................................................................................... 144 Figure 6.2: Monosaccharide content of non-adherent mucilage from Na2CO3 treated seeds   ............................................................................................................................................ 147 Fig 6.3: Monosaccharide analysis of whole seed AIR ............................................................. 149 Figure 7.1: Diagram of mucilage structural networks, organization and biosynthesis. ..... 157    xiv List of abbreviations  ABRC Arabidopsis Biological Resource Center AG Arabinogalactan AGP Arabinogalactan protein AIR Alcohol insoluble residue Ca2+ Calcium CBM Carbohydrate binding module CESA Cellulose Synthase A Col-0 Columbia ecotype 0 Col-2 Columbia ecotype 2 CSC Cellulose Synthase Complex CSL CELLULOSE SYNTHASE-LIKE DP Degree of polymerization DPA Days Post Anthesis EDTA Ethylene diamine tetraacetic acid, disodium salt dehydrate FLA Fasciclin-like arabinogalactan protein GAUT GALACTURONOSYLTRANSFERASE GFP Green fluorescent protein GPI glycosylphophatidylinositil HG Homogalacturonan HYP hydroxyproline irx irregular xylem ixr Isoxaben resistant MUM Mucilage-modified  xv n Sample size PME Pectin methylesterase PMEI Pectin methylesterase inhibitor QUA QUASIMODO RG I Rhamnogalacturonan I RG II Rhamnogalacturonan II RHM Rhamnose biosynthesis RSW Radial swelling RW Radial wall S4B Pontamine Fast Scarlet S4B SBT1.7 Subtilisin-like serine protease1.7 SEM Scanning electron microscopy SOS Salt-overly sensitive TAIR The Arabidopsis Information Resource T-DNA Transfer DNA of Agrobacterium tumefaciens XXT Xyloglucan xylosyltransferases    xvi Acknowledgements  I would like to sincerely thank Dr. George Haughn for accepting me into his lab, for his tutelage and instruction, and most importantly, his boundless patience. Thank you for supporting me, listening, and being critical. I would also like to thank Dr. Shawn Mansfield for his mentorship, detailed instructions on the analysis of cell walls and for constantly offering his different and unique point of view. Thank you to all of my committee members, and other faculty members of the Botany department for their support and guidance during my studies. Dr. Geoff Wasteneys for providing valuable scientific insight and support as a wonderful supervisory committee member. Dr. Reinhart Jetter for chemical tutelage and scientific insight as a great supervisory committee member. Dr. Fred Sack for an understanding ear and scientific guidance as a supervisory committee member. Dr. Ljerka Kunst as an unofficial mentor and constructive critic. Dr. Lacey Samuels for support and positive attitude. All of your guidance and support were absolutely required during the course of this degree. I would like to thank members of the Haughn and Kunst labs, past, present, and future, for scientific discussions, technical advice, and camaraderie. Patricia Lam for all the coffee and technical advice. Catalin Voiniciuc for our successful collaborations and discussions on cell wall structure and function. Dr. Gill Dean for coming back and providing excellent scientific advice and guidance. Dr. Robin Young for expert advice in microscopy and working in the Haughn lab.  xvii Thank you to Heather McFarlane for always being willing to discuss her past results and in providing advice on future experiments. Thank you to all of the staff in the UBC Bioimaging Facility. Thank you to Dr. Jennifer Klenz helping me along the way. I would also like to thank the UBC Bioimaging Facility and its staff, Garnet Martens, Brad Ross, Derrick Horne, and Kevin Hodgson for the many different micrographs used in this thesis. Lastly, thank you to my family for helping me be who I am. Thank you to my parents, Colin and Peggy Griffiths, for their support (spiritual and financial). Thank you to Lindi Masur for being the most wonderful person on this planet and for loving me.  xviii Dedication         For Lindi,  With Love and Squalor   1 1 Introduction  The cell wall is vital for the life of any plant. Cellulosic cell walls are a defining characteristic of plants, and are essential for many different aspects of plant form and function. Primary plant cell walls are composed of three major classes of macromolecules: cellulose, hemicelluloses and pectins (McNeil et al., 1984). The prevailing model for a type 1 primary cell wall states that there is a hemicellulose-cellulose network embedded in an independent pectin matrix (Albersheim et al., 2011). Polysaccharides interactions are complex and often difficult to analyze and understand. This research is focused on cell wall biosynthesis and structure, using the mucilage secretory cells of the Arabidopsis seed coat as a model. More specifically, this research will focus on the biosynthesis, function, and interactions of cellulose and pectins in the cell wall. The broad goal of this thesis is to understand how cell wall polymers form a cell wall, and how their interactions can affect cell wall physiology. 1.1 The cell wall Primary plant cell walls are defined as an extracellular matrix composed predominantly of polysaccharides. There are two classes of cell walls that are differentiated on the basis of when they are synthesized (Buchanan et al., 2000). Primary cell walls are synthesized during cell growth, while secondary cell walls are deposited once expansion has ceased. Cell walls have many functions during the life cycle of a plant, including structural support, protection from the environment, and cell-cell adhesion. Primary cell walls are generally flexible and extensible, allowing cells to expand, while still providing structural support and resisting turgor pressure  2 that would otherwise cause the plant cell to burst. Secondary cell walls are deposited once cell expansion has ceased, and often provide structural support. Cell walls are found surrounding every plant cell, and are extremely diverse in terms of composition and structure. The composition of a plant cell and the interactions between the cell wall polymers in the wall are of the utmost importance to cell wall function. Cell walls are composed primarily of three classes of polysaccharides, and other minor phenolic compounds and proteins (McNeil et al., 1984). The interactions between these polysaccharides in the wall determine the properties, and therefore the function of the wall. Based on the variety and different glycosidic linkages found in plant cell walls, it has been estimated that 10% of a plant’s genome encodes proteins, mostly uncharacterized, that are required for the biosynthesis and maintenance of plant cell walls (McCann and Rose, 2010; Harholt et al., 2010). Cell wall polysaccharides backbones are connected through glycosidic linkages (Reviewed in Harhohlt et al., 2010). Cell wall proteins can also be glycosylated, and can be covalently linked to cell wall polysaccharides (Fry, 1982; Held et al., 2004; Tan et al., 2013). Beside these covalent bonds, other non-covalent interactions exist between cell wall polysaccharides, including ionic bonds between pectic domains (Caffall and Mohnen, 2009), and hydrogen bonds between cellulose fibrils and between hemicellulose and cellulose (Pauly et al., 1999). Transient interactions between pectins and cellulose have been proposed that can also contribute to cell wall integrity (Zykwinska et al., 2005; Zykwinska et al., 2007a; Zykwinska et al., 2008; Wang et al., 2012). All of these interactions between wall polymers provide considerable strength to the cell wall.  3 1.1.1 Composition and structure of the primary cell wall As mentioned above, the three major classes of carbohydrates in primary walls are cellulose, hemicelluloses, and pectins (Fig. 1.1A). Cellulose is a major structural component of plant cell walls composed of 1,4-!-linked glucan chains (reviewed in Somerville, 2006). Pectins are found throughout the primary cell wall, concentrated in the middle lamella where they mediate the junctions between cells (Willats et al., 2001). Pectins are defined as acidic-complex heteropolymers that contain galacturonic acid (Fig. 1.1B; Mohnen, 2008). There are three different types of pectic polysaccharides: homogalacturonan (HG), rhamnogalacturonan I (RG I), and rhamnogalacturonan II (RG II; Fig. 1.1B; Willats et al., 2001). Hemicellulose is a term defining a broad class of heteropolymeric polysaccharides, including xylans, xyloglucans, glucomannans and mannans (Hayashi et al., 2011). In many primary cell walls, cellulose, hemicelluloses, and pectins are present in equal proportions, although relative amounts of each polysaccharide can be highly variable depending on the tissue type, and also varies widely depending on the species of plant (Burton et al., 2010). Cellulose is the prime structural component of primary cell walls, which is cross-linked by hemicelluloses and embedded in a matrix of pectins (Willats et al., 2001). The cellulose-hemicellulose network is believed to be the main load-bearing structure in the cell wall (Hayashi, 1989; Cosgrove, 2005). In primary walls, cellulose microfibrils are cross-linked by hemicelluloses through hydrogen bonding (Hayashi, 1989; Cosgrove, 2005). This cross-linking serves to strengthen and add integrity to the cell wall. Cross-linking networks between cell wall polymers can be altered during cell development to accommodate growth (Cosgrove, 2005). In a few cases, covalent  4  Figure 1.1: Primary cell wall architecture and pectin structure. A) Generalized diagram of a primary cell wall indicating various components of the wall in relation to the plasma membrane (From Sticklen, 2008. Reprinted with permission). B) Structure and composition of homogalacturonan and rhamnogalacturonan (Adapted from Harholt et al., 2010. Reprinted with permission. www.plantphysiol.org. Copyright American Society of Plant Biologists). A B  5 bonding of hemicelluloses to pectins has been described (Thompson and Fry, 2000; Popper and Fry, 2008). The reducing end of xyloglucan is can be linked through a glycosidic bond to side chains of RG I (Keegstra et al., 1973; Popper and Fry, 2005; Popper and Fry, 2008). Such linkages are expected to add considerable complexity and structural integrity to cell walls (Scheller and Ulvskov, 2010). It has been thought that pectins only play a space-filling role in the cell wall, but more recently they have been ascribed a more active role in interactions with other polymers (Willats et al., 2001; Vincken et al., 2003). New evidence indicates that it is possible for pectins to cross-link cellulose microfibrils, independently of hemicelluloses (Zykwinska et al., 2005; Zykwinska et al., 2007a; Zykwinska et al., 2007b; Dick-Perez et al., 2011). The exact details of how these cell wall components are organized and how they interact are not well understood. At least five different ways of cross-linking cellulose microfibrils have been proposed (Cosgrove, 2005). Cellulose is synthesized separately from matrix polysaccharides like pectins and hemicelluloses. Cellulose is synthesized directly into the apoplast by plasma membrane bound complexes (Pear et al., 1996; Arioli et al., 1998), while pectins and hemicellulose are synthesized in the Golgi apparatus, and then secreted to the apoplast (Reviewed in Caffal and Mohnen, 2009). Any linkages that occur between these polymers must be formed in muro. These cell wall polysaccharides are then linked together by weak physical forces and covalent cross- linking reactions. A better understanding of the relationship between pectins and cellulose should help us determine how the cell wall functions to accommodate growth and expansion, while withstanding internal pressures and maintaining an intact extracellular layer.  6 1.1.2 Composition and structure of secondary cell walls  Secondary cell walls are deposited after cell expansion has ceased, and are located in the apoplast, between the primary wall and the plasma membrane (Albersheim et al., 2011). Secondary cell walls are often associated with thick and rigid structures that are highly cellulosic in nature, and can also be lignified. They are typically composed primarily of cellulose, xylans (hemicelluloses), and lignin, with a reduced pectin component. 1.2  Structure and biosynthesis of cell wall polysaccharides 1.2.1 Cellulose 1.2.1.1 Structure of cellulose  Cellulose is the most abundant biopolymer on earth. It is a linear polymer composed of chains of 1,4 !-linked glucose molecules. Multiple chains associate with each other through hydrogen bonding to form microfibrils. It is thought that roughly 36 ! -glucan chains associate with each other to create one cellulose microfibril, although the exact number of individual polymers that form a microfibril varies between plant species and within plant species in different cell wall types (reviewed in Delmer & Armor, 1995). The length of this glucose polymer, i.e. its degree of polymerization (DP), is thought to range from 250 DP in primary walls, up to 15,000 DP in secondary walls (Reviewed in Brett, 2000; Albersheim et al., 2011). The straight, rigid structure of cellulose results in high tensile strength.  Cellulose microfibrils are composed of highly ordered crystalline regions and less crystalline amorphous regions (Updegraff, 1969). The degree of crystallinity is influenced by hydrogen bonds and Van der Waals forces between associated chains (Nishiyama et al., 2002; Nishiyama et al., 2003). These amorphous zones are possibly the site of association between hemicelluloses and cellulose, since there is greater surface area on the microfibril for  7 hemicelluloses to interact with, and hemicelluloses can also embed themselves inside a more amorphous microfibril (Nishiyama et al., 2002). Crystallization appears to be influenced by the geometry of the Cellulose Synthase Complex (CSC), by the speed of the CSC trafficking through the plasma membrane, and by the rate of glucan chain polymerization (Benziman et al., 1980; Jarvis, 2003; Harris et al., 2012). 1.2.1.2 Biosynthesis of cellulose Cellulose is synthesized by plasma-membrane-embedded CSC that appear as a rosette (Fig. 1.2; Mueller & Brown, 1980). Rosettes appear as 6-lobed structures embedded in the plasma membrane, and are often associated with the terminal ends of microfibrils (Fig. 1.2C; Mueller & Brown, 1980). CSCs are thought to be composed of subunits of multiple CELLULOSE SYNTHASE A (CESA) polypeptides (Fig. 1.2A). Each lobe of the rosette is thought to be composed of six CESAs, creating a total of 36 CESAs in one CSC (Fig. 1.2B; reviewed in Somerville, 2006). A CESA polypeptide has eight transmembrane domains, a catalytic active site, and a pair of N-terminal zinc finger domains thought to be involved in dimerization (Kurek et al., 2002). The transmembrane domains form a pore in the plasma membrane, through which one !-glucan chain is secreted into the cell wall (reviewed in Delmer & Armor, 1995). For simplicity, this thesis focuses on CESAs encoded by Arabidopsis thaliana, and the research reviewed here uses the CESA nomenclature from that species.  There are ten CESA genes in Arabidopsis, the expression of which are differentially regulated both temporally and spatially (Richmond & Somerville, 2000; Fagard et al., 2000; Beeckman et al., 2002; Somerville, 2006; Carroll and Specht, 2011). Different combinations of CESA polypeptides can form CSC with different roles. For example, the synthesis of cellulose in primary cell walls has been associated with CESA1, CESA3 and CESA6 (Desprez et al., 2007;  8  Figure 1.2: Cellulose biosynthesis and CSC assembly. A) Diagram of a CESA synthesizing !- (1,4)- glucan, and assembling into higher ordered rosette subunits and rosettes (From Doblin et al., 2002. Reprinted with permission). B) Putative rosette CESA composition. Six CESAs form one rosette, and 36 CESAs form a CSC (Scheible et al., 2001; Doblin et al., 2002). C) CSC rosette visualized in the plasma membrane by electron microscopy of freeze-fractured plasma membranes (From Kimura et al., 1999. Reprinted with permission. www.plantcell.org. Copyright American Society of Plant Biologists). !! "! #!C  9 Persson et al., 2007). CESA6 is functionally redundant with other closely related CESAs: CESA2, CESA5, and CESA9 (Desprez et al., 2007; Persson et al., 2007). Genetic evidence suggests that only three CESAs form one functional CSC comes from genetic evidence that showed groups of three cesa mutants having similar phenotypes for either primary (Desprez et al., 2007; Persson et al., 2007) and secondary cell walls (Taylor et al., 2000; Taylor et al., 2003). Null mutations in CESA1 and CESA3 are embryo lethal consistent with their role in primary cell wall biosynthesis (Persson et al., 2007). Vascular secondary cell walls are synthesized by CESA4, CESA7, and CESA8 (Taylor et al., 2000; Taylor et al., 2003). Mutations in CESA4, CESA7 and CESA8 were all identified through a screen for irregular xylem mutants (irx), indicating a role in secondary wall thickening in the vasculature (Taylor et al., 2000; Taylor et al., 2003). CESA4, CESA7 and CESA8 are capable of forming homodimers, but in order to form a higher order oligomer, all three CESAs appear to be required (Atanassov et al., 2009). CESA1 and CESA3 form a complete CSC with CESA6, or one of a group of CESA6-like proteins that are partially redundant (Desprez et al., 2007; Persson et al., 2007). CESA5 and CESA6 are the most closely related CESA6-like proteins (Desprez et al., 2007; Carroll & Specht, 2011). Until recently, cesa5 plants had no reported phenotype, while cesa5 cesa6 double mutants are seedling lethal, indicating CESA5 is redundant to CESA6 (Desprez et al., 2007). cesa2 cesa6 cesa9 triple mutant plants are male gametophytic lethal, indicating functional redundancy in pollen development (Persson et al., 2007). It has been proposed that different cell types require cell-specific cellulose deposition (Carpita & Vergara, 1998). The functional redundancy between the CESA6-like CESAs represents an interesting adaptation that allows for cell type specific regulation of cellulose biosynthesis. While the genetic evidence that groups CESA1, CESA3 and one of the CESA6-like CESAs in primary wall cellulose synthesis, and  10 CESA4, CESA7 and CESA8 in secondary wall synthesis is compelling, there are still many questions remaining regarding the specific composition and function of the CSC. The precise interactions between individual CESA polypeptides and between CSC polypeptides and other proteins involved in cellulose synthesis are unclear. There are a number of proteins in addition to CESAs that have clear roles in cellulose biosynthesis. KORRIGAN encodes a !-1,4 endoglucanase, and mutations in this gene cause severely stunted plants and altered cellulose qualities (Lane et al., 2001; Szyjanowicz et al., 2004). COBRA encodes a protein essential for microfibril orientation (Roudier et al., 2005). The orientation of cellulose microfibrils can determine the orientation of cellular expansion (reviewed in Baskin, 2001; Cosgrove, 2005). Cellulose microfibrils align with cortical microtubules (Green, 1962; Ledbetter & Porter, 1963). CESAs movement in the plasma membrane is guided by cortical microtubules (Paradez et al., 2006). Microtubules are required for the correct positioning of CESAs in the plasma membrane, which in turn affects microfibril orientation and cell wall structure (Baskin, 2001; Bichet et al., 2001; Baskin et al., 2004; Wasteneys & Fujita, 2006; Chan et al., 2010; Fujita et al., 2011; Bringmann et al., 2012). Microtubules organization and orientation can also affect cellulose crystallinity (Fujita et al., 2011). POM-POM2/CELLULOSE SYNTHASE INTERACTING1 interacts with CESAs to guide their movement, and is required for aligning CESA movement in the plasma membrane with microtubules, thereby controlling the orientation of cellulose synthesis (Gu et al., 2010; Bringmann et al., 2012).  11 1.2.2 Pectins 1.2.2.1 Structure and composition of pectins  Pectins are defined as acidic heteropolysaccharides, and all types of pectins are composed in parts by the acid sugar galacturonic acid. There are three types of pectins: homogalacturonan (HG), rhamnogalacturonan I (RG I), and rhamnogalacturonan II (RG II). HG is comprised solely of repeating units of α-1,4 -linked galacturonic acid molecules (reviewed in Caffall & Mohnen, 2009). The backbone of RG I is composed of alternating ["-galacturonic acid-"-1,2-rhamnose -"-1,4], and can be substituted on rhamnose molecules with side chains composed of galactose and arabinose. The backbone of RG II is composed of "-1,4-linked galacturonic acid molecules similar to HG, but it is decorated with complex side chains. RG II side chains can form dimers mediated by borate molecules, which can effectively cross-link pectin molecules through the galacturonic acid backbone (reviewed in Caffal & Mohnen, 2009). RG II is fairly well understood, has not been shown to be involved in mucilage, and is not the main focus of this research. In the cell wall, these pectin polymers can be linked together in many different ways to form a complex three-dimensional matrix (Cosgrove, 2005). The current model of cell wall structure indicates that pectins and hemicelluloses can be linked, and are synthesized in the Golgi and then secreted (Keegstra et al., 1973; Goubet and Mohnen, 1999; Sterling et al., 2001; Willats et al., 2001; Cumming et al., 2005; Popper and Fry, 2005; Popper and Fry, 2008; Atmodjo et al., 2011). Both HG and RG I can be altered by a number of different backbone modifications, and the addition of side chains. Galacturonic acid molecules of HG are secreted fully methyl esterified, and these methyl groups can be removed in the apoplast (reviewed in Caffal and Mohnen, 2009). By removing methyl groups, galacturonic acid molecules become negatively  12 charged, and can then freely associate with positively charged divalent cations. Molecules like calcium (Ca2+) can cross link different molecules of HG in cell walls, providing strength and structure to the cell wall (Micheli, 2001; Wolf et al., 2009; Peaucelle et al., 2012). In contrast, many plant pectin methyl esterases (PMEs) can remove methyl groups in a blockwise fashion, so that multiple adjacent galacturonic acids become charged and can associate with other stretches of non-methyl esterified HG through Ca2+ bridges (Goldberg et al., 1996; Limberg et al., 2000). In addition, galacturonic acid molecules can also be acetylated, however, this modification is not as well understood (Ishii, 1997). The structure and composition of pectins is further complicated by side chains of arabinans and/or galactans bound to rhamnose molecules of the RG I backbone. Three types of RG I side chains are known, those composed of arabinan, galactan, or arabino-galactans, which are a mix of both. The size and composition of side chains can be quite heterogeneous, depending on which tissue, and at what point during development they are isolated from the cell wall. Side chains can be very short, composed of only one or two molecules of arabinose or galactose, or hundreds of residues long. Arabinan and galactan side chains are modified during cell elongation; glycosyl hydrolases can cleave these side chains, which modulate their role in the wall (Labacitch and Ray, 1974; Willats et al., 1999; Fujino et al., 2000). The exact role of side chains in cell wall fluidity and strength is not entirely clear, however, they are thought to be able to cross-link independent molecules of RG I in the cell wall (Jones et al., 2003). It is also possible that longer side chains could associate with cellulose, increasing the integrity, or interconnectedness, of the cell wall (Zykwinska et al., 2005; Zykwinska et al., 2007a).  13 1.2.2.2 Biosynthesis of pectins  Pectins are synthesized in the Golgi apparatus, and then secreted to the apoplast. Many genes associated with pectin biosynthesis are localized to the Golgi. Nucleotide sugars are synthesized in the cytosol and transported into the Golgi where they undergo polymerization to form pectins and hemicelluloses. Given the inherent structural complexity of pectins, it is estimated that at least 51 different glycosyl transferases are required to fully synthesize all types of pectins (Ridley et al., 2001). However, only a few of these genes have been identified. The following are some examples of genes involved in pectin biosynthesis in Arabidopsis (reviewed in Mohnen, 2008; Caffal and Mohnen, 2009; Harholt et al., 2010): QUASIMODO1 (QUA1) is believed to be a HG galacturonic acid transferase (Bouton et al., 2002), while QUA2 is thought to be a pectin methyl transferase responsible for proper methylation of HG (Mouille et al., 2007). GAUT1 was identified as a galacturonic acid transferase using a biochemical approach, and is the only pectin biosynthetic gene shown to have galacturonic acid transferase activity (Sterling et al., 2006). A number of other GAUT genes (GAUT2-12) have been identified based on homology to GAUT1, but their roles in cell wall biosynthesis are less clear (Sterling et al., 2006; Driouich et al., 2012). GAUT7 is also required for HG biosynthesis, where it anchors two GAUT1 subunits to the Golgi membrane forming a core HG biosynthesis complex (Atmodjo et al., 2011). The role of three genes was identified in RG I biosynthesis through screens for mutants that modify seed coat mucilage, which is composed primarily of RG I (explained in more detail later). MUM4/RHM2 encodes a rhamnose synthase which is involved in providing rhamnose for RG I (Western et al., 2001; Usadel et al., 2004; Oka et al., 2006). MUM2 encodes a !- galactosidase which is hypothesized to modify galactan side chains on RG I (Dean et al., 2007).  14 Both mum2 and mum4 mutants fail to extrude mucilage when hydrated in water. BXL1 encodes an arabinofuranosidase/xylosidase that is also thought to modify RG I side chains in seed coat mucilage by modifying arabinan side chains in mucilage (Arsovski et al., 2009a). 1.2.3 Hemicelluloses 1.2.3.1 Biosynthesis and structure of hemicellulose  Hemicelluloses are a diverse class of polysaccharides that include xyloglucans, xylans, mannans, glucomannans and other glucans (reviewed in Scheller & Ulvskov, 2010). Xyloglucan is the most prominent hemicellulose in dicot primary cell walls (Scheller & Ulvskov, 2010). Xyloglucans have a similar backbone structure to cellulose, composed of repeating 1,4-!-glucose molecules; however, it is often substituted with side chains, such as xylose, galactose and fucose. A family of CELLULOSE SYNTHASE-LIKE (CSL) genes (Cocuron et al., 2007) synthesizes the glucan backbone of xyloglucans. The CSL-C family encodes proteins with ! -glucan synthase activity that catalyzes the polymerization of the glucan backbone (Cocuron et al., 2007). Another family of glycosyl transferases, called xyloglucan xylosyltransferase (XXT), are responsible for transferring xylosyl residues onto the glucan backbone (Faik et al., 2002; Cavalier and Keegstra, 2006). Hemicelluloses are thought to interact with cellulose, creating an interconnected matrix, and increasing the structural integrity of the cell wall (Levy et al., 1991). Xyloglucans are thought to be able to bind cellulose microfibrils in two ways: by being incorporated internally into a microfibril during crystallization, and surface binding by hydrogen bonds (McCann & Roberts, 1991). Xyloglucans therefore can link cellulose microfibrils together, but can also be covalently bound to pectins, creating a highly structured cell wall matrix. However, the exact role of hemicelluloses has recently been challenged by a xyloglucan xylosyltransferase triple  15 mutant (xxt1 xxt2 xxt5) that lacks all detectable xyloglucan, but has no serious effect on plant growth and development (Cavalier et al., 2008; Zabotina et al., 2012). 1.2.4 Proteins in the cell wall 1.2.4.1 Enzymatic extracellular proteins  In addition to copious amounts of polysaccharides, cell walls also contain small amounts of proteins. Cell wall proteins are involved in the modification of wall components, providing wall structure, mediation of signalling events and other interactions with the plasma membrane and the protoplast (reviewed in Jamet et al., 2006). Two major classes of cell wall-located proteins are the pectin methyl esterases (PMEs), and pectin methyl esterase inhibitors (PMEIs; reviewed in Jolie et al., 2010). These related proteins are involved in a complex relationship that modifies the methyl esterification of HG. HGs with a high degree of methylesterification are thought to be able to form gels through noncovalent hydrophobic interactions and hydrogen bonds between two carboxylic acids (Jolie et al., 2010). Alternatively, non-esterified HG is thought to facilitate Ca2+ binding between two unesterified HGs molecules, thereby forming gels (Yoo et al., 2003). PMEs remove methyl groups. PMEIs prevent this action by binding to PMEs and preventing their function (reviewed in Jolie et al., 2010). This relationship is further complicated by the fact that the activity of PMEIs can be altered through PMEI posttranslational modification in the wall. Proteases are also present in the cell wall. One subtilisin-like serine protease, AtSBT1.7, is thought to modulate PMEs or PMEI function in seed coat mucilage, which in turn affects cell wall properties (Rautengarten et al., 2008). Cell walls also contain many glycosyl hydrolases like MUM2 and BXL1 that have previously been discussed (Dean et al., 2007; Arsovski et al., 2009a). Other proteins also modify hemicellulose crosslinks, such as xyloglucan endotransglucosylase/hydrolases that catalyze the cleavage of xyloglucans, followed  16 by ligation of the cleaved polysaccharide to a different xyloglucan molecule (Fry et al., 1992; Nishitani & Tominaga, 1992; Yokoyama & Nishitani, 2001). Expansins are a family of secreted proteins that enable pH dependent wall loosening and growth (Cosgrove et al., 2005). Expansins are thought to affect cell wall expansion by influencing the dissociation of the polysaccharide complex that links cellulose microfibrils (McQueen-Mason & Cosgrove, 1995; Cosgrove, 2005). 1.2.4.2 Structural cell wall proteins Many other cell wall localized proteins are often highly glycosylated and can have various functions in cell wall structure and modification. Two major classes of cell wall structural proteins are extensins and arabinogalactan (AGP). The extension family contains 20 genes that encode hydroxyproline rich proteins (HRGP). HRGPs can be glycosylated on Hyp residues with one to four arabinose molecules (Held et al., 2004; Cannon et al., 2008). Null mutants of EXT3 have a lethal phenotype due to defects in the assembly of primary cell walls (Cannon et al., 2008). Extensins are also play a structural role in wound regeneration and formation of the cell plate (Stafstron & Staehelin, 1986; Cannon et al., 2008). AGPs are another secreted wall protein whose functions are poorly understood (Seifert and Roberts, 2007). AGPs are small glycoproteins that are highly glycosylated by Type II arabino-galactans (AG; reviewed in Seifert and Roberts, 2007). AGPs have been implicated in many biological processes, including cell division, programmed cell death, biological pattern formation, pollen tube guidance, pollen incompatibility, cell growth, secondary wall deposition, abscission, and intercellular signaling (Seifert and Roberts, 2007). AGPs are inherently difficult to study due to diversity and complexity in their glycosyl side chains, in muro modification, and possible pleiotropic functions (Shultz et al., 2004; Seifert & Roberts, 2007). The glycosyl side chains attached to AGP protein are highly diverse and contain many different sugar linkages  17 (Tan et al., 2004). The attached carbohydrates, which are primarily O-linked to hydroxyproline (Hyp) residues on the protein backbone, can constitute 90-98 % of the weight of a mature AGP (Tan et al., 2003; Tan et al, 2004). The carbohydrate polymer is primarily composed of a branched type II arabino-3,6-galactan polysaccharide. While these polysaccharides are comprised primarily of arabinose and galactose residues, studies have also identified rhamnose, glucuronic acid, fucose, and xylose amongst the isolated sugar residues (Clarke et al., 1979; Nothnagel, 1997; Tan et al., 2004). AGPs are secreted, and can remain attached to the plasma membrane in the apoplastic space through glycosylphophatidylinositol (GPI) anchors (Youl et al., 1998). Great progress has been made in elucidating the structure of the glycan side chains on AGPs, and in determining different protein domains in the individual domains. Previous research has highlighted that AGPs vary significantly in terms of the type of glycan chains, organization of proteins, and the presence or absence of a GPI anchor (Schultz et al., 2000; Seifert and Roberts, 2007). The role of the GPI anchor is complicated by the fact that these anchors can be cleaved by phospholipases that release the AGP from the plasma membrane into the apoplast, which might alter the function of AGPs during development (Shultz et al., 1998). AGP function can be further modified by glycosylases that degrade the polysaccharide side chains in muro (Sekimitata et al., 1989; Kotake et al., 2005; Leonard et al., 2009). Degradation of AGP glycans stimulates pollen tube elongation in tobacco stylar transmitting tissue (Cheung et al., 1995; Wu et al., 1995). Arabinofuranosidases can cleave terminal L-arabinofuranosyl residues on AGP side chains, which can make the remaining glycans more susceptible to ! -galactosidases (Sekimitata et al., 1989; Kotake et al., 2005). The glycosyl side chains of AGPs are important for the function of AGPs in muro. AGP glycosyl  18 side chains can also bind to calcium, which has many implications for the function of AGPs in signalling, and its interactions with other cell wall polymers (Lamport and Varnai, 2013). AGPs have been proposed to act as pectin plasticizers through interactions with AGP glycans, thereby modulating the ability of pectins to form gels (Lamport & Kieliszewki, 2005). Ultimately, how AGPs modulate cell wall interactions is unclear and is an active area of research.  Fasciclin-like AGPs 1.2.4.3  Fasciclin-like AGPs are a distinct subclass of AGPs that have putative cell-cell adhesion domains known as fasciclin (FLA) domains (Johnson et al., 2003). Proteins containing FLA domains from many different organisms can act as adhesion molecules (Elkins et al., 1990; Huber and Sumper, 1994; Kim et al., 2000). In total, twenty-one genes have been identified in Arabidopsis that contain FLA domains (Gaspar et al., 2001; Schultz et al., 2002). FLA proteins are one of the many classes of proteins that contain AG glycosyl side chains (Borner et al., 2003). Many FLAs have specific functions in regulating cell wall structure and function (Johnson et al., 2011).  FLA11 and FLA12 are important for maintaining cell wall integrity and elasticity (MacMillan et al., 2010). Mutation of both FLA11 and FLA12 resulted in altered stem biomechanics, and reduced arabinose, galactose, and cellulose content. The authors suggest that these FLAs affect cellulose deposition, and the integrity of the cell wall matrix (MacMillan et al., 2010). Other FLAs have also been linked to cellulose biosynthesis: FLA3 is required for normal microspore development in Arabidopsis (Li et al., 2010). FLA3-RNA interference lines showed abnormal pollen grains, which were attributed to defects in the intine layer cell wall deposition (Li et al., 2010). These FLA3-RNA interference lines also show reduced calcofluor white staining, a dye that can bind to cellulose, again implicating FLAs in cellulose deposition and cell  19 wall integrity (Li et al., 2010). SALT-OVERLY SENSITIVE 5/FASCICLIN-LIKE ARABINOGALACTAN-PROTEIN 4 (SOS5/FLA4) encodes another FLA AGP, which was initially identified in a screen for salt sensitivity in roots (Shi et al., 2003). Mutations of SOS5 result in reduced cell wall organization, specifically in the pectin-rich middle lamella (Shi et al., 2003). Later research linked SOS5 to cellulose biosynthesis, based on similarities in seed mucilage hydration and adherence phenotypes to cesa5-1 seeds (Harpaz-Saad et al., 2011). The link between FLAs and cellulose biosynthesis, and specifically between SOS5 and CESA5, is investigated in greater detail in Chapter 5. 1.3  The Arabidopsis seed coat  In Arabidopsis, fertilization of an ovule induces a developmental process that culminates in the production of an embryo, endosperm and the seed coat (Haughn & Chaudhury, 2005). The seed coat is uniquely designed to protect the embryo until suitable conditions for germination exist. The seed coat consists of five maternally derived cell layers that undergo dramatic cellular changes during seed development (reviewed in Haughn & Chaudhury, 2005). Seed coat cells are derived from the two maternal ovule integuments, the inner and outer integuments. A number of unique cell walls are synthesized to protect the embryo from the environment, while mediating information about the environmental conditions and promoting germination (Haughn and Chaudhury, 2005). These cell walls have proven to be useful systems in the investigation of cell wall biosynthesis and structure (Fig. 1.3; Western et al., 2000; Haughn & Chaudhury, 2005; Haughn & Western, 2012). 1.3.1 Development and structure of the seed coat  The innermost cell layer of seed coat, the pigmented layer, synthesizes proanthrocyanidin flavonols that form the tannins, which are thought to protect the seed from pathogens and, when  20  Figure 1.3: Mucilage secretory cell development, structure and composition. A) Seed coat epidermal cells at 7 DPA, 10 DPA, and Maturity, stained with Toluidine Blue. Mucilage pockets (labeled M) are stained light purple while the columella (labeled S) is stained dark purple (From Mendu et al., 2011). B) Seed coat mucilage stained with ruthenium red. C) Adherent mucilage of seed shaken in water, then stained with ruthenium red. D) Cellulose in mucilage labeled with CBM3a (From Blake et al., 2006). E) Cellulose in mucilage stained with S4B (Mendu et al., 2011). F) Scanning electron micrograph of the surface of Arabidopsis seeds. Arrow indicates the columella in the center of the cell. Bar A = 10 µm, B-E = 50 µm, F = 25 µm.  21 mature, impart the brown colour to seeds (Debeaujon et al., 2003; Dixon et al., 2005; Haughn & Chaudhury, 2005). The next two layers of seed coat cells undergo early programmed cell death and are crushed during the development of the seed coat. The outer two cell layers of the seed coat are characterized by the presence of secondary cell walls. The inner layer of the outer integument produces a secondary cell wall on the basal face of the cell. Cells of the outer (epidermal) layer have two complex secondary cell walls that give seeds their distinctive shape: mucilage, that consists primarily of pectins (Fig. 1.3A, labeled M), and the columella, that is thought to be composed primarily of cellulose and hemicellulose (Fig. 1.3A, labeled S). Because the seed coat epidermal cells undergo a developmental timeline where vast amounts of cell wall material are synthesized and secreted during a short period of time, this cell type provides an excellent system to study cell wall biosynthesis (Haughn & Western, 2012). While both the outer and inner cell layer of the outer integument synthesize cellulosic secondary cell walls, this research focuses primarily on the epidermal cell layer of the seed coat. 1.3.2 Seed coat epidermal cell development The seed coat epidermal cells differentiate from cells of the ovule integument. Ovule epidermal cells are initially surrounded by a primary cell wall, which is composed of pectins, hemicelluloses, and cellulose. Following fertilization, all cells of the integument grow, undergoing cell division and expansion. At roughly 4 days post-fertilization, cells are filled with a large central vacuole, with a few amyloplasts visible at the extremities of the cells (Western et al., 2000). By approximately 6 days post-fertilization, the epidermal cells of the seed coat have ceased expansion and begun secondary cell wall biosynthesis (Western et al., 2000). Seed coat epidermal cells deposit two discreet types of secondary cell wall. The first secondary cell wall is composed primarily of pectinaceous mucilage (Fig. 1.3A). This mucilage is secreted to the  22 apoplast at the junction of the outer tangential and radial primary cell walls, forming a mucilage pocket (Fig. 1.3A, labeled M). Secretion of mucilage continues until a large donut-shaped pocket of mucilage is formed, leaving a cytoplasmic column in the center of the cell, while the bottom half of the cell contains a vacuole and the remainder of the cytoplasm (Fig. 1.3A). By roughly nine days post-fertilization, seed coat epidermal cells begin to synthesize a thick cellulosic secondary cell wall along the outer tangential and radial sides, which fills the entire cell by maturity (Fig. 1.3A, labeled S). Upon maturity, seed coat epidermal cells undergo apoptosis leaving a dehydrated mucilage pocket and the columella enveloped by the primary cell wall. The radial walls of the seed coat epidermal cells form a unique hexagonal cell shape when viewed from the outer surface (Fig. 1.3F). In the center of each cell, the top of the columella protrudes, since at maturity the mucilage is dehydrated and is recessed compared to the radial wall and the top of the columella (see arrow, Fig. 1.3F). 1.4  The structure and composition of seed coat mucilage The epidermal seed coat cells and the mucilage pocket are dehydrated at maturity. When mature seeds are hydrated, the mucilage swells rapidly and extrudes from the mucilage pocket by bursting the primary cell wall below the junction of the radial wall, and the outer tangential wall. The hydrated mucilage forms two distinct layers: inner adherent mucilage and outer non- adherent mucilage (Fig. 1.3B). By approximately 1 minute after hydration, the adherent layer has fully expanded and remains attached to the seed coat. In contrast, the non-adherent layer continues to expand for a long period post-hydration. The non-adherent layer is more diffuse and less structured than the adherent layer. If seeds are shaken under aqueous conditions, the non- adherent layer quickly dissociates from the seed and dissipates, breaking into small pieces (Fig. 1.3C). The different properties of adherent and non-adherent mucilage are likely due to  23 differences in the polysaccharide composition of the two fractions, and the interactions between these polysaccharides. 1.4.1 Mucilage composition  Our understanding of mucilage composition comes from quantitative chemical analyses of the monosaccharide composition and linkages within the mucilage polysaccharides, as well as from qualitative cytological analysis using carbohydrate specific antibodies (reviewed in Haughn & Western 2012). These data help identify different cell wall polymers in mucilage and their arrangement within and between cells. Both types of analysis indicate that mucilage is composed predominantly of RG I that contains a few short arabinan, galactan, or arabinan-galactan side chains (Western et al., 2000; Macquet et al., 2007a). In addition to RG I, mucilage contains small amounts of HG, cellulose, and hemicelluloses. 1.4.1.1 Quantitative analysis of mucilage composition  Non-adherent mucilage is easily released from the seed and quantified, whereas adherent mucilage must be released by a strong acid or base, which breaks or weakens the association between the adherent mucilage and the seed. The literature reports the percentage of non- adherent mucilage that is RG I ranges from 80% (Huang et al., 2011) to 90% (Macquet et al., 2007a) of total recovered mucilage. In addition, linkage analyses indicates that there are very few side chains present in non-adherent mucilage, with only two side chain sugars present for every ~50 RG I backbone sugars. Side chains are typically small, usually only one or two sugars long, and composed of arabinans and galactans (Dean et al., 2007; Huang et al., 2011). The other ~20% of non-adherent mucilage is composed of approximately equal proportions of HG, cellulose, and hemicelluloses (Huang et al., 2011). It is not clear if the glucose present in mucilage is derived from hemicelluloses, or from cellulose microfibrils; however, attempts at  24 isolating cellulose microfibrils from non-adherent mucilage have been unsuccessful (Mendu et al., 2011).  Similarly, adherent mucilage is primarily composed of unbranched RG I. Seeds treated with 2 M NaOH show no ruthenium red staining, and no remaining mucilage is visible under any circumstances. The monosaccharide content of adherent mucilage is roughly 75-85% RG I (includes 5-15% arabinan and galactan side chains), 5-20% HG, and 5-10% hemicellulose (Macquet et al., 2007a; Huang et al., 2011). Linkage analysis indicates that the amount of side chains per backbone RG I sugar is higher in adherent mucilage versus non-adherent mucilage. For every 50-backbone sugar residues, there are roughly 10 side chain sugars, and side chains are slightly longer in adherent mucilage (Fig. 1.3; Arsovski et al., 2009a; Huang et al., 2011). Cellulose is also thought to be present in adherent mucilage, but again, no crystalline cellulose has been chemically detected in this layer of mucilage. Amounts of glucose detected by monosaccharide analysis increases drastically from non-adherent mucilage to adherent mucilage (Macquet et al., 2007a; Arsovski et al., 2009a). 1.4.1.2 Qualitative analysis of mucilage composition  The adherent mucilage halo stains intensely with pectin specific dyes and cellulose specific dyes. Ruthenium red binds to HG and intensely stains the adherent mucilage halo (Sterling, 1970; Western et al., 2000). Cellulose staining fluorescent dyes like calcofluor white (Herth & Schneph, 1980) and pontamine fast scarlet S4B (S4B; Anderson et al., 2010; Mendu et al., 2011) and fluorescently labeled antibodies directed against carbohydrate binding domains specific for cellulose also stain mucilage (Fig. 1.2D and E; Willats et al., 2001a; Blake et al., 2006; Young et al., 2008; Sullivan et al., 2011). Staining seeds with S4B demonstrates two major areas within the adherent mucilage halo, a region of diffuse staining, and a more concentrated  25 region centered above the columella called a ray (Mendu et al., 2011; Sullivan et al., 2011; Harpaz-Saad et al., 2011). Rays are also labelled with antibodies that detect certain polysaccharide epitopes (Sullivan et al., 2011; Harpaz-Saad et al., 2011). Immunolabelling can provide qualitative information regarding the location of specific cell wall polymer domains in adherent mucilage. JIM5 and JIM7 are two antibodies that recognize low and highly methylated homogalacturonan, respectively (Willats et al., 2001b; Willats et al., 2001c). Neither antibody binds to the entire mucilage halo: JIM7 binding is primarily localized to the outer region of the adherent layer, while JIM5 localizes to the primary cell wall remnants attached to the top of the columella, as well as the ray structures that form above the columella (Macquet et al., 2007a). Carbohydrate binding modules derived from cellulases and fused to fluorescent reporters have been reported to be specific for cellulose microfibrils also show binding specifically to the ray structures present in mucilage and more diffusely in the whole mucilage halo, indicating some co-localization of unmethyl esterified pectins and cellulose microfibrils (Blake et al., 2006; Young et al., 2008; Sullivan et al., 2011). Xylogalacturonan-specific antibodies bind to the majority of the mucilage pocket, however, there is much stronger signal from the area surrounding the columella (Young et al., 2008). Antibodies specific for the RG I backbone bind the entire mucilage halo, consistent with the hypothesis that mucilage is composed primarily of RG I, accentuated with other cell wall polymers. 1.4.2 Genetic analysis of mucilage biosynthesis and structure Arabidopsis seed coat epidermal cells represent an excellent system for using genetics to study cell wall structure and biosynthesis. Since cell walls are essential, one of the main problems involved in isolating cell wall mutants is lethality that might result from cell wall  26 modification. However, seed coat epidermal cells themselves are not essential. Mutations in the APETALA2 gene, that encodes a transcription factor required for seed coat epidermal cell differentiation, result in seeds that lack differentiated epidermal cells, yet the seeds are able to germinate under laboratory conditions (Jofuku et al., 1994; Western et al., 2001). Thus, none of the features of the seed coat epidermal cells, including mucilage and the columella, are required for viability. Therefore, under laboratory conditions it is possible to isolate mutants involved in cell wall biosynthesis that might be otherwise impossible to study in other cell types. Mucilage and columella also provide unique and consistent biological features that are highly amenable to forward and reverse genetic screens (Arsovski et al., 2010; Haughn & Western, 2012). The adherent mucilage halo is very consistent in terms of size and composition, allowing for easy identification of mutants that have altered characteristics. Indeed, mucilage has been used extensively to identify multiple genes involved in cell wall biosynthesis and regulation (Western et al., 2001; Arsovski et al., 2009b; Arsovski et al., 2010). 1.4.2.1 Genes involved in mucilage biosynthesis The first screen for mucilage altered mutants isolated five independent mutants termed MUCILAGE MODIFIED 1-5 (Western et al., 2001). Four of these mutants have now been cloned and their role in mucilage biosynthesis characterized. MUM4 was the first gene cloned from this screen, and encodes a rhamnose synthase, RHM2 (Western et al., 2004; Usadel et al., 2004). Mutant mum4 seeds fail to extrude mucilage when treated with water (Western et al., 2004). Chemical analysis identified an ~90% reduction in RG I (Western et al., 2004). MUM2 encodes a !-galactosidase that is proposed to act on galactan side chains found in mucilage. mum2 mutant seeds synthesize normal amounts of mucilage, but does not release mucilage when treated with water. Some mucilage is released when treated with Na2CO3 (Dean et al., 2007;  27 Macquet et al., 2007b). Chemical analysis indicates that this mutant is not deficient in mucilage biosynthesis, rather that its hydration properties are altered. MUM1 has a very similar phenotype to MUM2, and encodes a transcription factor, LEUNIG HOMOLOGUE, which regulates the expression of MUM2 and other mucilage modifying enzymes (Huang et al., 2011; Walker et al., 2011). Recently, it has been shown that MUM3 encodes CESA5; loss of CESA5 function results in loss of mucilage adherence (Sullivan et al., 2011; Mendu et al., 2011). CESA5’s role in mucilage was also identified in this thesis using a reverse genetics screen (Mendu et al., 2011; Sullivan et al., 2011; Harpaz-Saad et al., 2011). cesa5-1 seeds show ample mucilage release in water, however, mucilage is no longer adherent to the parent seed, and is discussed at length in Chapter 3. Many other genes are also involved in mucilage biosynthesis. BXL1 encodes a bi- functional "-arabinofuranosidase/xylosidase that also affects mucilage hydration and adherence similar to mum2, but is less severe (Arsovski et al., 2009a). SBT1.7 encodes a subtilisin-like serine protease that does not release mucilage when hydrated with water (Rautengarten et al., 2008). This mutant does release mucilage when treated with EDTA and shows a unique primary cell wall detachment phenotype (Rautengarten et al., 2008). SBT1.7 is thought to act by modifying PME or PMEI activity in mucilage such that the number of methyl esters in HG is reduced (Rautengarten et al., 2008). MUM2, BXL1 and SBT1.7 all appear to be regulated by MUM1 (Fig. 1.4; Huang et al., 2011). In addition to MUM1, many other transcription factors have been implicated in mucilage biosynthesis (Fig 1.4). 1.2.4.2 Transcriptional regulation of mucilage biosynthesis As described previously, AP2 appears to be a master regulator of seed coat epidermal cell development. ap2 mutants do not synthesize mucilage, nor do they develop a columella. Other  28  Figure 1.4: Regulation of mucilage biosynthesis and seed coat development. Two major pathways are known to regulate mucilage biosynthesis. A ternary complex composed of TTG1, TT8/EGL3, and MYB5 regulated the expression of GL2 and TTG2, which in turn regulate mucilage biosynthesis and seed coat development. The second pathway involves LUH/MUM1 regulation of MUM2, BXL1 and SBT1.7 to modify the pectin component of mucilage (From Huang et al., 2011, www.plantphysiol.org. Copyright American Society of Plant Biologists).  29 transcription factors are involved in more specific aspects of differentiation. Mucilage biosynthesis appears to be regulated by a complex consisting of at least three transcription factors: TTG1, MYB5, and one of TT8 or EGL3, transcription factors which are partially redundant (Western et al., 2001; Gonzalez et al., 2009; Li et al., 2009; Huang et al., 2011). This TTG1 complex is thought to directly regulate the expression of GL2 and TTG2, which in turn regulates mucilage biosynthetic genes (Fig. 1.4; Szymanski et al., 1998; Johnson et al., 2002; Western et al., 2004; Ishida et al., 2007). Mutant ttg1, ttg2 and gl2 seeds do not appear to release mucilage in water, and have flattened columella (Koornneef, 1981; Rerie et al., 1994; Western et al., 2001; Johnson et al., 2002). MYB5 also regulates mucilage biosynthesis, and myb5-1 mutant seeds show altered mucilage hydration in water (Li et al., 2009; Gonzalez et al., 2009). MYB5 is thought to participate in a ternary complex with TTG1 and TT8/EGL3 to regulate mucilage biosynthesis and seed coat development (Fig. 1.4; Gonzalez et al., 2009; Li et al., 2009). MYB61 and KNAT7 also regulate mucilage biosynthesis, however, it is not clear how they fit into the main mucilage regulatory pathway (Penfield et al., 2001; Romano et al., 2012). 1.5  Research goals and hypothesis  The cell walls of seed coat epidermal cells represent an excellent system to study cell wall structure and biosynthesis. The main goal of my research is to increase our understanding of cellulose biosynthesis and function in the cell wall using the seed coat epidermis as a model. This goal was separated into two distinct studies: First the analysis of cellulose biosynthesis and function during seed coat development in Chapters 3, 4, and the analysis of the role of SOS5 in relation to cellulose during seed coat development in Chapters 5 and 6. Using genetics, the redundancy between the CESA6-like CESAs function in seed coat development was investigated using a number of different mutant lines (Chapter 3). This initial characterization led to the  30 identification of the central role of CESA5 in mucilage adherence and new questions regarding the biosynthesis and role of cellulose in mucilage. I sought to identify other CESAs that participate in cellulose biosynthesis with CESA5 during mucilage deposition, and identified a role for CESA3 in mucilage biosynthesis (Chapter 4). The mechanism of mucilage adherence was examined by analyzing the genetic interactions between CESA5 and SOS5 (Chapter 5). This study identified independent roles in mucilage adherence for CESA5 and SOS5, while providing novel evidence for the function of SOS5 in cell wall biology and development. Finally, I was able to provide indirect evidence that MUM2’s role in cell wall modification is dependent of SOS5, providing indirect evidence that these two proteins are interacting in the cell wall (Chapter 6). The focus of each chapter is described in more detail here. 1.5.1 Chapter 3: CESA2, CESA5 and CESA9 are involved in secondary cell wall biosynthesis, yet only CESA5 is required for mucilage biosynthesis  The complex redundancy between CESA2, CESA5, CESA6 and CESA9 has been a curious and interesting aspect of the biology of cellulose biosynthesis. The goal of this research was to use the seed coat to examine the redundancy between the CESA6-like genes in the seed coat, and to understand how individual CESAs contribute to cellulose biosynthesis. I hypothesized that each CESA has a unique role in cellulose biosynthesis and that the redundancy between the CESA6-like CESAs is not complete. Mutations in CESA2, CESA5 and CESA9 all showed similar reductions in cell wall biosynthesis in seed coat epidermal cells, and this phenotype was enhanced in a cesa2 cesa9 double mutant and a cesa2 cesa5 cesa9 triple mutant. Additionally, I identified a unique role for CESA5 in mucilage cellulose biosynthesis providing clear evidence it is not redundant to CESA2 or CESA9 during mucilage biosynthesis.  31 1.5.2 Chapter 4: CESA3 participates in mucilage cellulose biosynthesis  The identification of CESA5 having a unique role in mucilage biosynthesis challenged the generally accepted hypothesis that three independent CESAs are required to form a functional CSC. In order to identify other CESAs involved in mucilage biosynthesis, I examined the expression of all 10 CESAs during seed coat development using bioinformatics. CESA3 and CESA10 were highly expressed during mucilage biosynthesis and selected as good candidates. CESA5::GFP and CESA3::GFP are present in seed coat epidermal cells during mucilage biosynthesis. Two mutant alleles of cesa10 showed no significant differences from wild type seeds in terms of mucilage hydration and structure. Two mutant alleles of cesa3, ixr1-1 and ixr1- 2, show reduced cellulose amounts in seeds, altered adherent mucilage structure and pectin distribution, demonstrating a role for CESA3 in mucilage biosynthesis 1.5.3 Chapter 5: CESA5 and SOS5 mediate mucilage adherence independently  During the progress of this work, another gene was identified that had a nearly identical mucilage adherence defects to cesa5-1. SOS5 encodes a fasciclin-like AGP, and was hypothesized to mediate mucilage adherence through CESA5. I tested this hypothesis by examining the phenotype of both mutants, identifying clear differences in their phenotypes. The creation of a double cesa5 sos5 double mutant also provided evidence that the two genes mediate adherence through different mechanisms. SOS5 does not appear to be involved in cellulose biosynthesis, and instead may have a role in organizing and providing structure to mucilage independent of cellulose. Detailed analysis of this double mutant and each single mutant provided insights into the structure of mucilage, the role of cellulose in adherence, and the function of SOS5.  32 1.5.4 Chapter 6: SOS5 is required for MUM2 function in mucilage Finally, I examined the function of MUM2 in mucilage adherence, hypothesizing that its function to cleave galactans could be affecting mucilage adherence by reducing interactions between galactan side chains and cellulose, or by modifying the function of SOS5 by degrading the side chains present on SOS5. Double mum2 cesa5 and mum2 sos5 mutants were isolated to investigate how the function of MUM2 on galactans affects the mucilage adherence mediated by CESA5 and SOS5. Double mum2 cesa5 mutants appeared phenotypically similar to mum2-1 mutants, indicating no interaction between MUM2 and CESA5. However, double mum2 sos5 mutants appeared phenotypically similar to sos5-2 mutants, indicating that SOS5 is required for proper MUM2 function. 1.6  Summary Taken as a whole, this thesis provides novel information on the biosynthesis and function of cellulose. I have shown (1), evidence for incomplete redundancy between the CESA6-like CESAs during seed coat development; (2), a novel role for cellulose in mucilage biosynthesis; (3), that CESA5 and CESA3 contributes to mucilage biosynthesis; (4), new information on the nature of mucilage adherence and its underlying mechanisms; and (5), new evidence on the function of SOS5, the relationship between SOS5 and MUM2, and on the role of AGPs in cell wall biology and plant physiology in general.   33 2 Materials and methods 2.1  Plant materials and growth conditions Seeds were germinated on AT media (Haughn and Somerville, 1986) plates with 6% (w/v) agar, and seedlings were transferred to soil (Sunshine Mix 4; SunGro, Kelowna, British Columbia) after 6 days. Plants were grown in chambers with continuous fluorescent illumination of 80-140 µEm-2s-1 at 20-22ºC. Seed were either harvested from individual plants, or bulked for chemical analysis. For microscopy, chemical analysis, and mucilage hydration assays, all seeds from each genotype for a given experiment were harvested at the same time and treated similarly. 2.1.1 Genotypic and phenotypic screening of mutant lines 2.1.1.1 Seed lines Arabidopsis plants used in this study are derived from the Columbia-0 (Col-0) ecotype, except for mum2-1, which comes from the Col-2 line (Dean et al., 2007). Col-2 was developed by Shauna Somerville by propagating Col-0 5 times through single seed descent. All SALK T- DNA insertion lines, and individual point mutations were obtained from the Arabidopsis Biological Resource Center (ABRC) in Columbus, Ohio (Table 2.1;Alonso et al., 2003) and from Persson et al. (2007), Stork et al. (2010), Mendu et al. (2011) and Harpaz-Saad et al. (2011) as follows: CESA2 (cesa2-1, SAIL_400_D01] and cesa2-2, SALK_096542; Persson et al., 2007), CESA5 (cesa5-1, SALK_118491), CESA6 (procuste-ixr2-1, R1064W; Desprez et al., 2002), and CESA9 (cesa9-1 SALK_107750C, and cesa9-2 SALK_046455; Persson et al., 2007; Stork et al., 2010). Additionally, cesa2-1 cesa9-1 double mutants were obtained (Persson et al., 2007) and a cesa2-1 cesa5-1 cesa9-1 triple mutant was generated by crossing the cesa2-1 cesa9-  34 Table 2.1: List of mutant lines, genomic locus, and genotyping primers used in this study. Allele Locus Mutation line Forward primer Reverse primer cesa9- 1 AT2G21770 SALK_107750 C CGTAGGGAAAGCACATTCGTTG CACATACAGGTGGACACATAGAGGG cesa9- 2 AT2G21770 SALK_046455 CAGACCCATATAGCAAACCAAAG CAGACCCATATAGCAAACCAAAG cesa2- 1 AT4G39350 SAIL_400_D01 TCATTCCAGGATGGTCAGGG TTGTTGTTTTCAGCACAGGCG cesa2- 2 AT4G39350 SALK_096542 AAAGCGGTTTTTGTCCTCTTC TCTCCAATTTCAATTGTGTTGC cesa5- 1 AT5G09870 SALK_118491 GGGAACCGGGTGTGTTTTTAGGAGA ATTTCGCCGCATCATTAGTGCCTTTGGT prc1-1 AT5G64740 CS297 CTTCAAGTAGTCCATTTTATGGCAG GGTACAGGTTGTGTTTTCAGGA sos5-2 AT3G46550 SALK_125874 GAAACTGGGAATAACCTTCGG AGCTTCTCGAGACCAAACCTC mum2- 1 AT5G63800  MUM2F - CGTCAACAATGCACTAGAAG mum2F CGTCAACAATGCACTAGGAA CTAACTTTCTCTCCAAGCAAAC ixr1-1 AT5G05170 CS6201 GAGAGGTTTGCGTATGTGAAC CTCAACAGTTGATTCCACATTCC ixr1-2 AT5G05170 CS6202 GAGAGGTTTGCGTATGTGAAC CTCAACAGTTGATTCCACATTCC cesa10 -1 AT2G25540 SALK_052533 AACCCCTTCAAGTTGATCTGC TACCCAGTGGATAAAGTTGCG cesa10 -2 AT2G25540 SALK_150533 CTCTAGGGTTGTTACCAGGCC AATGGTGACGAACTCCAAATG any1-1 AT4G32410 irx1-5 AT4G18780 SALK_046685 GAAATTTGTGTCGAGACCAGC TACAGTCCACCTTCAAAACCG  35 1 double mutant with cesa5-1 (genotyping primers for each allele are documented in Table 2.1). sos5-2/mum2-1/cesa5-1 double mutants were by crossing the single mutants together. CESA5::GFP expressing plants were obtained from Bischoff et al. (2011), CESA6::GFP and CESA3::GFP were obtained from (Desprez et al., 2007). CESA3 mutants, ixr1-1 (G998A) and ixr1-2  (T942I) were obtained from the ABRC, and were first identified by Scheible et al. (2001). ixr1-1 and Genotyping primers for all lines are listed below (Table 2.1). The CESA1 mutant, anisotropy1 (any1-1) was isolated by Miki Fujita in the Wasteneys lab and is an unpublished allele of cesa1. 2.1.1.2 Genotyping of SALK T-DNA insertions and point mutations Left (LP) and right (RP) gene specific primers were selected using the SALK T-DNA Primer Design tool (http://signal.salk.edu/tdnaprimers.2.html). The LBb1.3 primer (5’- ATTTTGCCGATTTCGGAAC-3’) was used as the insert-specific primer. PCR reactions were set up with FP+LP, and LP+LBb1.3 primer combinations. Wild type amplicons of roughly 1kb were expected from RP + LP, while insert-specific DNA fragments of 500-800bp were expected from RP+LBb1.3. ixr1-1 and ixr1-2 were identified by sequencing a region flanking both mutations with primers in Table 2.1. mum2-1 was identified using differential PCR with mutation specific forward primers. 2.2  Statistical analysis Mutant segregation was analyzed by PCR identification of T-DNA insertions, and point mutation identification. Single and double mutant segregation patterns were determined by Chi squared (!2) test, with significance determined at P < 0.05. Comparison of columella size, and monosaccharide amounts were performed by a Students T-test, with values being significant at P < 0.05. Monosaccharide and crystalline cellulose amounts were analyzed using a two-way  36 ANOVA. Values were also analyzed with Tukey’s post-hoc comparison to determine differences using cesa5-1 and sos5-2 as factors. ANOVA and Tukey’s were performed using JMP statistical discovery software (SAS Institute Inc., North Carolina, US). 2.3  Developmental analysis of the seed coat epidermis Seed stages were determined by marking flowers with nontoxic, water-soluble paint at 0 days post-anthesis (DPA) just as the flowers start opening and long stamens grow over the gynoecium (Western et al., 2001). The embryo was also dissected to confirm similar seed stages between genotypes. Staged siliques were dissected using a sharp razor blade. Seeds were removed from the siliques and preserved using high pressure freezing and freeze substitution, embedded in Spurr’s resin and sectioned as previously described (Mendu et al., 2011). The seed coat was punctured with a razor blade to facilitate resin infiltration. Seeds were transferred onto copper hats (Ted Pella; Redding, California) containing 1-hexadecene, and fixed using a Leica EM HPM 100 High Pressure Freezer (Leica; Germany). The copper hats were then transferred to cryovials containing freeze substitution medium [2 % (w/v) osmium tetroxide with 8% (v/v) dimethoxypropane in acetone]. Freeze substitution was performed at -80°C for 5 days in a Leica EM AFS chamber (Leica; Germany), followed by -20°C for 20 h to allow for reaction of the fixatives. Following substitution, samples were removed from the copper hats and rinsed in anhydrous acetone several times and slowly infiltrated and embedded in Spurr’s epoxy resin over a period of four days (Canemco; Lakefield, Quebec; Spurr, 1969). Seeds embedded in resin were sectioned (50 !m) using a Reichert Ultracut E microtome (Reichert; Seefeld, Germany) and stained with 1 % (w/v) toluidine blue O in 1X (w/v) sodium borate (pH 11). Seed coat epidermal cell development was also analyzed using propidium iodide stained seeds using confocal light  37 microscopy Seeds were incubated for 10–15 min with a 10 µg/ml propidium iodide solution to visualize cell walls. 2.4  Microscopy 2.4.1 Transmitted light microscopy Mature dry seeds, hydrated in water or other solutions, were stained with 0.01% (w/v) ruthenium red (Sigma-Aldrich; USA) for 1 h while shaking on a rotator. Brightfield micrographs of stained samples were taken with QCapture software and digital camera (QImaging; Surrey, British Columbia) equipped on a Zeiss AxioSkop 2 upright light microscope (Carl Zeiss AG; Germany). 2.4.1.1 CaCl2, EDTA and Na2CO3 mucilage treatments Seeds were shaken in 50mM CaCl2, 50mM EDTA, or 100mM Na2CO3 for 1 h while shaking on a rotator. Seeds were then rinsed twice with distilled water and stained with ruthenium red. 2.4.2 Confocal microscopy 2.4.2.1 Whole seed staining with calcofluor white and pontamine S4B Seeds were mixed with water on an orbital shaker for 1-2 hr, then stained with 0.01 % (w/v) pontamine fast scarlet S4B (S4B; Sigma-Aldrich Rare Chemical Library, #S479896; USA) and 100mM NaCl for 1 h (Anderson et al., 2010; Mendu et al., 2011). Seeds were then rinsed twice with distilled water and imaged using a 561 nm laser on either a Zeiss 510 Meta Laser Scanning Confocal Microscope (Carl Zeiss AG; Germany) or a PerkinElmer Ultraview VoX Spinning Disk Confocal system (PerkinElmer; Waltham Massachusetts). Calcofluor staining and imaging was carried out using a previously described method (Willats et al., 2001). All confocal micrographs were processed and measured with ImageJ (Abramoff et al., 2004).  38 2.4.2.2 Whole seed immunolabelling The immunochemistry techniques used closely resemble previously described protocols (Blake et al., 2006; Young et al., 2008; Harpaz-Saad et al., 2011). The specificities of the four monoclonal antibodies and two CBMs used have been extensively described (Knox et al., 1990; Knox, 1997; Willats et al., 2001b; Blake et al., 2006; Macquet et al., 2007a; Young et al., 2008; Pattathil et al., 2010;). For CBM3a, CBM28, CCRC-M36, and JIM5 (PlantProbes, Leeds, England; CarboSource; Athens, Georgia) immunolabelling, seeds were sequentially washed with the following solutions while rotating on an orbital shaker at room temperature (Table 2.2). For CBMs, CBMs were used as the primary antibody, followed by a PB wash and the Anti-HIS secondary body, followed by PB washes, then the Alexafluor488 tertiary antibody. The secondary antibodies used against JIM5 and JIM7 were goat-anti-rat conjugated to AlexaFluor488 secondary antibodies, and against CBM3a, CBM28 and CCRC-M36 was goat- anti-mouse conjugated to AlexaFluor488 (Molecular Probes, Invitrogen; Carlsbad, California). CBM3a and CBM28 (PlantProbes; Leeds, England) were used as a primary probe and were detected by Mouse anti-Tetra-his antibody (Qiagen, USA). The immunolabelling method was carried out without primary antibody as a negative control. Seeds were imaged using a 488 nm laser for antibody fluorescence and 561 nm laser for S4B and seed intrinsic fluorescence on a Zeiss 510 Meta Laser Scanning Confocal Microscope (Carl Zeiss AG; Germany) or a PerkinElmer Ultraview VoX Spinning Disk Confocal system (PerkinElmer; Waltham Massachusetts).   39 Table 2.2: List of washes and antibodies used in immunolabelling  Step Solution Volume (!L) Duration (min) 1 Water 1000 30 2 Phosphate buffer (PB, pH 7.4) 800 30 3 PB with 5% (w/v) Bovine Serum Albumin (BSA) 100 30 4 Primary antibody (CBM3a, CBM28, M36, JIM5 or JIM7) diluted 1/10 in 1% BSA in PB 50 90 5 PB (this step is repeated 5 times in total) 800 10 (each wash) 6 Secondary antibody diluted 1/100 in 1% BSA in PB (incubation in dark) 100 90 7 5X PB wash 800 10 (each wash) 8 0.01% S4B with 100mM NaCl 1000 30    40 2.4.2.3 Expression and subcellular localization Dissected seeds were imaged using a 488nm laser (GFP fluorescence) and transmitted light on a PerkinElmer Ultraview VoX Spinning Disk Confocal system (PerkinElmer; Waltham Massachusetts). 2.4.3 Scanning electron microscopy Dry seeds were mounted on stubs and coated with gold-palladium in a SEMPrep2 sputter coater (Nanotech; Worcester, Massachusetts). Images were taken with a Hitachi S4700 scanning electron microscope (Hitachi High-Technologies; Canada). Columella and cell area were analyzed in ImageJ (Abramoff et al., 2004). 2.5 Chemical analysis 2.5.1 Determination of monosaccharide composition by HPAEC The protocol described here is a modified version of previously published procedures (Dean et al., 2007; Mendu et al., 2011). The average seed weight of wild type and mutant seeds was determined by averaging the weight of three replicates of 100 seeds for each genotype. To prepare the mucilage extraction sample for High-Performance Anion-Exchange Chromatography (HPAEC), four technical replicates of 20-25mg wild type or mutant seeds (exact weight recorded) were mixed with 1.4 mL of distilled water and 10 !L of 5 mg/mL D-erythritol (internal standard). The samples were lightly shaken using a tube rotator speed for 1 h. For mum2-1 sodium carbonate (Na2CO3) analysis, seeds were mixed with 50 mL of 1 M Na2CO3 and left to stand at room temperature for 15 min before adding 1.2 mL of water and 10 !L of 5 mg/mL D-erythritol was added as an internal standard and the samples vortexed for 2 h on medium speed. The mucilage in the supernatant was transferred (1mL) to a glass tube and dried at 60ºC under nitrogen gas. Serial dilutions (1 mM, 0.5 mM, 0.25 mM,  and 0.125 mM) of  41 neutral sugar standards (fucose, arabinose, rhamnose, galactose, glucose, and xylose) and acid sugar standards (galacturonic acid) in distilled water or Na2CO3 were transferred (0.5 mL) to glass tubes, mixed with 10 !L of 5 mg/mL D-erythritol (internal standard), and dried under nitrogen gas at 60ºC. All mucilage samples and sugar standards were hydrolyzed simultaneously. After drying under nitrogen gas, samples were treated with 17.4 !L of 72% (w/v) sulphuric acid for 2 h and shaken vigorously every 30 min with a vortex mixer. 482.6 !L of distilled water was then added to each glass tube to give a "nal concentration of 2.5% sulphuric acid, and all samples and standards were autoclaved for 60 min at 121ºC before being "ltered through 0.45- mm nylon syringe "lters. For the analysis of whole seeds, tubes were filled with 4-5 mg of seeds (exact weight recorded), frozen in liquid nitrogen, and ground using pestles. The powder was resuspended in 1 mL of 70% (w/v) ethanol and heated at 65ºC for 10 min. Three 30 min washes in 1 mL 70% ethanol were then performed on an orbital shaker in order to remove small soluble sugars. The alcohol-insoluble residue (AIR) was collected with a microcentrifuge between washes, and was then dried under nitrogen gas. The dried AIR was weighed, transferred to glass tubes containing 10 mL of 5 mg/mL D-erythritol, and dried once again. After 2 h in 70 µL 72 % (w/v) sulphuric acid at 40C, 1.93 mL water was added to each sample to give a final concentration of 2.5% sulphuric acid. The concentration and identity of neutral monosaccharide sugars was determined using HPAEC analysis on a DX-600 BioLC chromatograph (Dionex). Sugars were separated on a CarboPac PA1 anion exchange column (Dionex), and detected via pulsed amperometry across a gold electrode. A 10 !L volume of sample was injected. The column was equilibrated with 250 mM NaOH and eluted with deionized water at a flow rate of 1.0 mL min. Detection of  42 carbohydrates was facilitated with a postcolumn addition of 200 mM NaOH at 0.5 mL min. The concentration of acid sugars in the filtrate was determined with the identical HPLC system and column, with a different elution gradient: sugars were eluted in 100 mM NaOH with a linear gradient of 0 to 400 mM sodium acetate from 5 to 40 min, followed by a 10 min 300 mM NaOH wash. The column was then equilibrated with 100 mM NaOH between injections. Monosaccharides were quantified with the D-erythritol internal standard after correction of response factors with monosaccharide standards of different concentrations. Sugars from mucilage were normalized to the mass of seed extracted, and sugars from whole-seed AIR were normalized to the mass of AIR. 2.5.2 Crystalline cellulose quantification Crystalline cellulose amounts were determined based on a microscale modification of Updegraff (1969), developed by Shawn Mansfield. 10-20 mg of seeds were frozen in liquid nitrogen, ground using mortal and pestle, and then dried at 500C overnight. The exact dry weight was then treated with 2 mL acetic 150 mL 80% acetic acid (30 mL H20 + 120 mL glacial acetic) + 15 mL conc. nitric acid (70%) and vortexed. Samples were heated at 1000C for 1 h, centrifuged for 5 min at 3000 rpm, and washed. Samples were then treated with 1 mL 72% (w/v) H2SO4 vortex, and incubated at room temperature for 90 min, centrifuged 5 min, 10 000 rpm and then diluted 10X in distilled water (Duplicated samples). 100 µL of diluted sample was treated with 200 µL cold (40C) anthrone reagent and vortexed. Anthrone-mixtures were incubated for 15 min at 1000C, and samples were measured twice for absorbance at 620 nm in a spectrophotometer. A standard curve was prepared from a standard dilution of 1.2 mg/mL Avicel microcrystalline cellulose that was treated identically to samples. Total amounts of cellulose were calculated per weight of dry seed mass.  43 2.6 Bioinformatic analysis 2.6.1 Analysis of gene and protein structure All genes analysed in this study were first investigated using the Arabidopsis Information Resource (TAIR; http://arabidopsis.org; Swarbreck et al., 2007). 2.6.2 Transcript analysis Expression patterns in specific Arabidopsis organs and cell types were visualized with the eFP browser (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi; Winter et al., 2007) and corroborated with GENEVESTIGATOR (https://www.genevestigator.com/gv/plant.jsp; Hruz et al., 2008). Expression analysis during seed development was calculated by comparing total expression values and error from the seed specific data set on the eFP browser (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi?dataSource=Seed) and on the seed expression data from the developmental map data set on the eFP (http://bar.utoronto.ca/efp/cgi- bin/efpWeb.cgi?dataSource=Developmental_Map).   44 3 CESA2, CESA5 and CESA9 are involved in secondary cell wall biosynthesis in seed coat epidermal cells, yet only CESA5 is required for mucilage biosynthesis 3.1 Introduction  Cellulose is the most abundant cell wall polymer in plants, and is attributed with providing mechanical strength and structure to plant cell walls (reviewed in Somerville, 2005). Cellulose is required for providing strength to the seed coat; cellulose deficient mutants in Arabidopsis are more susceptible to tetrazolium salt uptake (Stork et al., 2010). Despite its importance, little is known regarding the synthesis of cellulose in this tissue. The seed coat epidermal cells have three distinct types of cell walls that all contain cellulose: the outer primary wall, mucilage, and the columella (Western et al., 2000; Macquet et al., 2007a). Cellulose has been identified in the adherent layer of mucilage through immunolabelling and staining with cellulose specific dyes (Blake et al., 2006; Macquet et al., 2007a; Young et al., 2008). Seed coat development is tightly regulated, and the timing of mucilage production and columella biosynthesis are well defined (Western et al., 2000). Mucilage is deposited from 4 DPA through 9 DPA. It is secreted to the apoplast at the junction of the radial and tangential walls, on the outer side of seed coat epidermal cells beneath the primary cell wall. Mucilage deposition coincides with the contraction of the large central vacuole toward the bottom of the cell, and formation of a large cytoplasmic column in the center of each epidermal cell. A large apoplastic mucilage pocket surrounds the cytoplasmic column. Following mucilage deposition, a secondary cell wall, the columella, is deposited from 9 DPA through 12 DPA. The volcano- shaped columella is formed beneath the mucilage pocket, displacing the cytoplasmic column  45 (Western et al., 2000). Secondary wall thickening also occurs at the radial walls of epidermal cells. While the chemical composition of the columella and radial walls has not been determined, cellulose most likely plays an important role in seed coat secondary walls. Lignins have not been identified in the Arabidopsis seed coat (C. Douglas and G. Haughn, Unpublished results). Despite the importance of the seed coat, very little is known about the biosynthesis of cellulose during seed coat development. When mature seeds are hydrated, the mucilage bursts from the pocket, and separates into two distinct fractions: a non-adherent layer that is easily separated from the seed by mild shaking and an adherent layer that remains attached. How mucilage remains adherent to the parent seed is not clear, however, two mechanisms are possible: first, mucilage could be covalently linked to the parent seed through the secondary wall of the epidermal cells; or, possibly the mucilage creates a cohesive carbohydrate network that remains adherent with no direct attachment. It is also unclear how the two layers of mucilage separate from each other. 3.1.1 Cellulose biosynthesis There are 10 cellulose synthases in the Arabidopsis genome. Generally, three different CESAs are believed to be required to produce cellulose in a single cell type (Desprez et al., 2007; Persson et al., 2007). CESA1 and CESA3 are normally responsible for primary cell wall biosynthesis, and interact with CESA6 or a CESA6-like CESA. CESA2, CESA5 and CESA9 comprise a group of CESAs that share high sequence identity with CESA6 (Fig. 3.1; Desprez et al., 2002). CESA2, CESA5 and CESA9 are fully redundant to CESA6 in many different plant tissues, and partially redundant to CESA6 (Desprez et al., 2002; Desprez et al., 2007; Persson et al., 2007; Carroll & Specht, 2009). These CESA6-like CESAs are thought to be involved in primary wall synthesis  46  Figure 3.1: Diagram of Arabidopsis CESA relationship and functional grouping (adapted from a phylogeny in Desprez et al., 2002; Carroll & Specht, 2009).  CESA9 CESA5 CESA2 CESA6 CESA1 CESA3 CESA4 CESA7 CESA8 ! ! ! CESA6-like Primary wall Secondary wall CESA10  47 based on genetic data (Desprez et al., 2007; Persson et al., 2007). Double mutants for cesa5 cesa6 are seedling lethal, and CESA5 is thought to be redundant to CESA6 because of this (Desprez et al., 2007). Double cesa2 cesa6 mutants show enhanced seedling growth phenotypes compared to the cesa6 mutant procuste1-1 (prc1-1), indicating partial redundancy between these two genes (Desprez et al., 2007; Persson et al., 2007). Triple cesa2 cesa6 cesa9 mutants are pollen lethal, indicating a role for all three in pollen development (Persson et al., 2007). The biosynthesis of cellulose in the embryo has been investigated (Beeckman et al., 2002); however, no studies have examined the role of any CESAs in seed coat development. Despite the importance of the seed coat in plant reproductive cycles, and the importance of cellulose in secondary wall deposition, the role of CESAs in seed coat epidermal development has not been investigated. Since there are large amounts of secondary wall synthesized in a short time period, seed coat epidermal cells were used to more fully understand the redundancy between CESA2, CESA5 and CESA9. A reverse genetics approach was used to examine seed coat development in mutant lines for these genes. T-DNA insertion lines were isolated first in CESA9 (Stork et al., 2010), then in CESA2 and CESA5 (Mendu et al., 2011). These mutants were then screened for defects in secondary cell wall biosynthesis and mucilage hydration properties. We show that these CESAs are not completely redundant in cellulose biosynthesis within the seed coat epidermal cells, with each appearing to contribute to columella synthesis.  Finally, cellulose synthesized by CESA5 is required for mucilage adherence, demonstrating a unique role for CESA5 and an important role for cellulose in mucilage adherence.  48 3.2 Results 3.2.1 Phenotype of CESA9 T-DNA insertion lines Little is known regarding the specific role of CESA9. CESA9 is expressed in the embryo (Beeckman et al., 2002), however, seed coat expression was not investigated. Using the GENEVESTIGATOR gene expression profiling tool (Zimmerman et al., 2004), CESA9 was shown to be expressed during seed development, and specifically up-regulated after stage three of seed development (Stork et al., 2010). Two independent cesa9 T-DNA insertion lines were isolated (cesa9-1: Persson et al., 2007; cesa9-2: Stork et al., 2010) and examined for defects in seed coat epidermal development. Both lines showed no obvious morphological changes in other plant tissues (Persson et al., 2007; Stork et al., 2010). Examination of the embryo of cesa9 insertion lines did not show any significant changes from wild type embryo development (not shown). Mutant cesa9 seeds were hydrated in water and stained with ruthenium red, however, no change in mucilage hydration properties, or the size of the mucilage halo was observed for either mutant line (not shown). 3.2.2 Epidermal cell morphology is altered in cesa9 mutant seeds  Since CESA9 is expressed in the seed, but the cesa9 mutant has no obvious embryo defects, we next examined the ultrastructure of the cesa9 mature seed coat cells. Scanning electron microscopy (SEM) was used to examine the structure of epidermal seed coat cells (imaging and analysis performed by Jozsef Stork; Fig. 3.2; Stork et al., 2010). Compared to wild type, cesa9-1 and cesa9-2 seeds appeared to have an altered seed coat epidermal cell shape (Fig. 3.2). Most notably, the radial walls between cells appeared reduced in the cesa9 mutant lines. In some cases for cesa9 lines the radial border between cells was not observable at all, whereas  49  Figure 3.2: Wild type and cesa9-1 SEM images of seed coat epidermal cell morphology. Different regions of epidermal cells are highlighted for easier identification of radial walls (highlighted in purple), the mucilage pocket (highlighted in light blue) and the columella (highlighted in green). A) Col-0; B) cesa9-1. Bar = 30 µm. From Stork et al., 2010. www.plantphysiol.org. Copyright American Society of Plant Biologists.  ! !  50 wild type radial walls are almost always present. This made the epidermal cells of cesa9 insertion lines appear to be fused. In addition, the cesa9 seed coat epidermal cells had an altered cell shape compared to the fairly uniform hexagonal shape of wild type cells. In order to quantify these changes in cell morphology, the SEM images of the seed coat were used to measure the area of the cell, not including the radial wall and the columella, nor the area of the columella for wild type and cesa9 mutant lines (analysis performed by Joseph Stork; Stork et al., 2010). A total of 30 cells from each of 10 different seeds of each genotype were used for the calculations of seed coat epidermal cell area (Fig. 3.3). In total, wild type cells had a greater area (830 µm2) compared to cesa9-1 (717 µm2), while the columella area of wild type was smaller (114 µm2) compared to cesa9-1 (140 µm2).  Based on these results, cell wall deposition appears to be reduced in cesa9-1 seed coat epidermal cells. In order to investigate this possibility further, the development of cesa9-1 and cesa9-2 seeds were examined by sectioning fixed seeds at different stages of development. Seeds at 4 DPA, 7 DPA and 10 DPA were cryo-fixed, embedded in resin, sectioned, and stained with Toluidine blue (Fig. 3.4A). Mature seeds were fixed in glutaraldehyde prior to resin embedding, sectioning and Toluidine blue staining (Fig. 3.4B). At 4 DPA, the primary cell wall can be observed surrounding the large central vacuole for all three genotypes. By 7 DPA, the mucilage pockets have begun to form, and the cytoplasmic column is filled with amyloplasts. No major differences were observed between cesa9-1, cesa9-2 and wild type cells. By 10 DPA, the columella has begun to displace the cytoplasmic column as it is synthesized. In all three genotypes, a small (1-2 !m) secondary wall, which stains a lighter purple compared to the darker staining mucilage, can be observed lining the underneath of the mucilage pocket, the top of the  51   Figure 3.3: Wild type and cesa9-1 seed epidermal cell size and columella area. A) Seed coat epidermal cell area for Col-0 and cesa9-1. B) Surface columella area for Col-0 and cesa9-1. From Stork et al., 2010. www.plantphysiol.org. Copyright American Society of Plant Biologists. A re a (μ m 2 )   52   Figure 3.4. Wild type and cesa9 seed coat epidermal cell development and radial wall height. A) Development of cryo-fixed seed coat epidermal cells at 4, 7 and 10 DPA. Arrows in 10 DPA images indicate the location of the radial wall. B) Mature seeds fixed in glutaraldehyde. C) Radial wall (RW) height (n = 20). Asterisks indicate significant difference from Col-0, P < 0.05. D) Radial wall (RW) width (n = 20). Bar = 10 µm. From Stork et al., 2010. www.plantphysiol.org. Copyright American Society of Plant Biologists.   53 cytoplasmic column, and the sides of the radial walls. In contrast to wild type Columbia cells, cesa9-1 and cesa9-2 showed a reduction in the amount of secondary cell wall material deposited in the radial walls and beneath the mucilage pocket (Fig. 3.4A, arrow). The columella of cesa9 seeds appeared to have an increased amount of cell wall material, which could be due to differences in the staging of cell development. To further analyze these differences, mature seeds were fixed in glutaraldehyde, an aqueous fixative that allows the mucilage pocket to burst, and removes the mucilage (Fig. 3.4B). In the sections of resin embedded, glutaraldehyde-fixed seeds, the major features of mature epidermal seed coat cells can clearly be observed. The mucilage pocket has burst at the radial wall, leaving the remnants of the tangential primary wall attached to the columella, radiating upwards away from the parent seed. The columella itself is very large in the center of the cell, and the smaller radial walls can be seen on either side of the columella, extending to almost the same height as the columella (Fig. 3.4B). Changes in secondary cell wall deposition were quantified by measuring the radial wall height and width for Col-0 and cesa9 seeds. Radial walls were chosen as the most representative measure of secondary wall deposition because the radial walls are quite distinctive, and almost all planes of section will cut through the radial wall. Secondly, compared to the columella, the radial wall is much smaller, and changes in the amount of secondary cell wall should be more readily observed. The radial wall height of the cesa9 mutant lines appeared shorter than wild type Columbia cells (Fig. 3.4B; no major difference was observed in the columella). The radial wall height of wild type and cesa9 mutant lines was quantified by measuring from the top of the radial wall to the base of the mucilage pocket (Fig. 3.4B). A total of 20 radial walls from at least 10 cells were measured for each genotype. The results indicated a significant reduction in radial  54 wall height in cesa9-1 (6.7 ± 0.33 µm) and cesa9-2 (4.9 ± 0.13 µm) compared to Col-0 (8.3 ± 0.41 µm; Fig. 3.4C). The differences in wall heights between Col-0 and cesa9 lines were significant based on a student t-test between each individual mutant line and wild type (P < 0.05). The radial wall width was also measured at roughly halfway up the height of the radial wall, however, no significant difference was observed between mutant seeds and wild type (Fig. 3.4D). 3.2.3 Isolation of CESA2, CESA5, and CESA9 mutant lines for the analysis of cellulose biosynthesis during seed coat development.  Having identified a role for CESA9 in secondary wall biosynthesis, we sought to further investigate the role of the CESA6-like CESAs in seed coat development. The reduction in radial wall height in cesa9 mutants was only ~20%, indicating other factors are involved in this process. CESA2, CESA5 and CESA9 are all highly homologous, and are partially redundant with CESA6. T-DNA insertions and point mutations were isolated by Staffan Persson in CESA2 (cesa2-1, SAIL_400_D01; cesa2-2 SALK_096542), CESA5 (cesa5-1, SALK_118491), CESA6 (procuste1-1; Fagard et al., 2000), and CESA9 (cesa9-1, SALK_107750C; cesa9-2 SALK_046455; Persson et al., 2007; Stork et al., 2010). Double cesa2-1 cesa9-1 mutants were obtained (Persson et al., 2007), and a triple cesa2-1 cesa5-1 cesa9-1 mutant was created by Jozsef Stork through crossing cesa5-1 with the cesa2-1 cesa9-1 double mutant (Stork et al., 2010). In all mutant lines, plant growth morphology was comparable to wild type Col-0, however, there was slight reduction in stem height for the cesa2-1, cesa2-1 cesa9-1 double mutant, and cesa2-1 cesa5-1 cesa9-1 triple mutant (data not shown).   55 3.2.4 Radial walls are absent or reduced in cesa2-1 cesa5-1 cesa9-1 seeds  Similarly to what was previously done for the cesa9-1 and cesa9-2 mutants, Arabidopsis seed coat epidermal cells were examined using SEM for altered morphological shapes in the cesa2 cesa5 cesa9 mutant lines by Joseph Stork. Much like cesa9 mutants, cesa2 and cesa5 single mutants showed changes in seed coat cell morphology (Fig. 3.5). The cesa2 cesa9 double mutant was more severe than any single mutant, and the triple cesa2 cesa5 cesa9 mutant showed the most prominent changes, with the radial wall completely unobservable for most cells (Fig. 3.5). To quantify radial cell reduction, seeds were cryo-fixed during development at 7 DPA and 11 DPA, embedded in resin and sectioned (Fig. 3.6). Additionally, mature seeds were again fixed under aqueous conditions (3% glutaraldehyde), embedded in resin and sectioned (Fig. 3.6).  No major differences in the seed coat epidermal cells were observed at 7 DPA between any of the mutants and Col-0 seeds (Fig. 3.6, left panel). However, at 11 DPA there was a visible reduction in secondary wall material beneath the mucilage pocket and at the radial wall (Fig. 3.6, center panel). Both mutant alleles of cesa2 showed a reduction in secondary wall material. The cesa5-1 mutant line looked similar to both cesa9 mutant alleles in terms of reduction in secondary wall thickening at the radial wall and beneath the mucilage pocket. The cesa2 cesa9 double mutant showed a stronger reduction in wall thickening compared to cesa9 mutants. In addition, it appeared as though there was a reduction in the amount of secondary wall deposited in the center of the columella, however, this could result from differences in plane of section. The cesa2 cesa5 cesa9 triple mutant showed the most severe reductions in secondary wall deposition (Fig. 3.6, center panel). Only a thin line of secondary wall material was observed underneath the mucilage pocket and at the radial wall. In addition, the columella were less developed, showing large gaps of cytoplasm in the center of the columella, which was not  56  Figure 3.5: Epidermal cell shape and morphology of cesa mutants. Surface SEM micrographs of mature seeds coat epidermal cells (left). Bar = 30 µm. Individual cell area measured for 300 cells (right). Asterisks indicate significant difference from wild type (P < 0.01). From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.  57  Figure 3.6: Developmental analysis of wild type and cesa mutant seed coats. Epidermal cell morphology is shown for all lines in Toluidine blue-stained sections of cryofixed 7 DPA (left column), 10 DPA (middle column), and glutaraldehyde-fixed mature seeds (right column). Mucilage stains dark purple, while the columella stains a lighter purple at 10 DPA. Bar = 10 µm. From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.  58 observed in the sections from any other mutant line or in Col-0. The nature of deposition of secondary wall material in the columella also appeared to be affected. The front of the extending secondary wall that was deposited in the cytoplasmic column appeared lumpy and irregular, when all other columella from other lines appeared smooth. Mature seeds were fixed, embedded and sectioned under aqueous conditions to allow an unrestricted view of mature radial walls and secondary cell wall deposition (Fig. 3.6, right panel). It was readily apparent that all mutant lines showed some decrease in radial wall integrity. The two cesa2 mutant lines showed stunted radial walls. In cesa5-1, the radial walls appeared much thinner, and shorter than wild type. The double cesa2 cesa9 mutant had a stronger reduction in radial wall height than all the single mutants. The triple cesa2 cesa5 cesa9 mutant had extremely stunted radial walls, which were not observable for some cells (Fig. 3.6, right panel). Even though the triple mutant showed hollow columella at 11 DPA, mature seeds had solid columella, indicating that secondary cell wall deposition is delayed, but not arrested.  In addition to these visual observations of reduced secondary wall deposition, radial wall height was quantified (Fig. 3.7). Previously, for the analysis of cesa9 radial wall height, height was measured from the base of the mucilage pocket to the top of the radial wall. For the analysis of cesa2, cesa5, and cesa9 mutant seeds, radial wall height was measured from the base of the secondary wall, which proved to be a more accurate and consistent measure of wall height. A total of 20 radial walls from various numbers of seed were measured from three separate resin blocks (n = 60). Radial walls for wild type Col-0 were the highest (8.14 ± 0.27 µm), while all cesa single mutants were reduced to a similar height (cesa2-1, 6.51 ± 0.17 µm; cesa2-2, 5.71 ± 0.15 µm; cesa5-1, 6.27 ± 0.18 µm; cesa9-1, 6.64 ± 0.22 µm; cesa9-2, 5.78 ± 0.16 µm). The double mutant radial walls were reduced more than any single mutant, and the triple mutant was  59  Figure 3.7: Radial wall height of cesa mutant lines. Asterisks indicate significant difference from wild type (P < 0.05 by Student’s t-test). From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.   60 severely reduced compared with all other mutant lines (cesa2-1 cesa9-1, 5.49 ± 0.15 mm; cesa2- 1 cesa5-1 cesa9-1, 4.87 ± 0.12 mm). Radial wall heights for cesa9-1 and cesa9-2 were similar to previous measurements of these genotypes, if differences in the measurement procedure are taken into account (Stork et al., 2010). Based on the wall height measurements, all cesa mutants were significantly different from the wild type based on Student’s t-tests between each single mutant and Col-0 (P < 0.05). In addition, the triple cesa2 cesa5 cesa9 radial walls were further reduced compared to cesa single mutant lines, and the cesa2-1 cesa9-1 double mutant. Results of the developmental series indicate that reduced radial wall height is consistent with the decreased radial wall ridges observed by SEM. 3.2.5 Crystalline cellulose is reduced in CESA6-like mutants  The observed changes in radial wall integrity are presumably a result of reduced cellulose biosynthesis in seed coat epidermal cells. Cellulose content in seeds from CESA6-like mutants was determined by Venugopal Mendu (Mendu et al., 2011). Differences in acid-insoluble glucose are a quantitative estimate of crystalline cellulose (Updegraff, 1969). Significantly lowered cellulose contents were identified in the cell walls of single, double, and triple mutants seeds [Fig. 3.8; 114 ± 4 (Col-0), 93 ± 3 (cesa2-1), 91 ± 2 (cesa2-2), 91 ± 2 (cesa5-1), 92 ± 4 (cesa5-2), 96 ± 1 (cesa9-1), 92 ± 3 (cesa2-1 cesa9-1), 63 ± 4 (cesa2-1 cesa5-1 cesa9-1) mg cellulose/g seed; P < 0.05 by Student’s t-test]. When comparing the single cesa2 or cesa9 mutants to the double cesa2 cesa9 mutant, no significant difference was observed. However, when the triple cesa2 cesa5 cesa9 mutant is compared to cesa2 cesa9, and to all three single mutants, a cumulative reduction in cellulose in whole seeds is observed based on individual Student’s T-Test between single mutants and wild type, and between the triple mutant and all  61  Figure 3.8: Acid insoluble cellulose content of cesa mutant seeds.. Error bars are SE calculated from 3 technical replicates from three independent batches of seeds as biological replicates. From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.  !" #!" $!" %!" &!" '!!" '#!" '$!" !" ##$ #% &" '( %) *" )* '+, -. -' (" ##' / 0# #1 '  62 other mutant lines (P < 0.05). Quantification of total seed crystalline cellulose amounts also includes cellulose present in the embryo, and is not specific to the seed coat. 3.2.6 Reduced mucilage adherence in cesa5-1 seeds  Mucilage biosynthesis is followed closely by secondary cell wall columella biosynthesis. In order to determine if there is any relation between the function of CESAs in mucilage and in the columella, seed mucilage extrusion was examined in all cesa mutant lines. Seeds were hydrated directly in ruthenium red and viewed within one minute of hydration. No difference was observed between wild-type and cesa mutant lines. Both wild type and cesa mutants had large mucilage halos (Fig. 3.9). Shaking wild-type seeds in water causes the non-adherent layer to dissociate, leaving only the adherent layer. When mutant seeds were hydrated in water and shaken for ~1 hr prior to ruthenium red staining, stark differences between lines homozygous for the cesa5 mutation, and all other lines tested were observed (Fig. 3.9, second row). Unlike the other lines, the cesa5-1 mutant and the cesa2 cesa5 cesa9 triple mutant had no adherent mucilage staining (Fig. 3.9). These observations indicate that seed mucilage in cesa5-1 mutant lines have lost mucilage adherence to the parent seed. 3.2.7 Reduced mucilage adherence and cellulose staining in cesa5-1 seeds  Cellulose has been shown to be present in hydrated mucilage using staining techniques (Willats et al., 2001; Blake et al., 2006; Macquet et al., 2007; Young et al., 2008). Therefore, seeds were examined after shaking in water, and then stained with two cellulose dyes, calcofluor white and pontamine fast scarlet S4B (S4B; Fig. 3.9). Calcofluor white binds cellulose, but also binds to other glucan molecules (Anderson et al., 2010). S4B has been shown to highly be specific to cellulose (Anderson et al., 2010). S4B staining of wild type Col-0, cesa2-1 and cesa9- 1 mutant lines showed no changes in the S4B staining pattern (not shown). A strong reduction in  63  Figure 3.9: Mucilage hydration and adherence phenotype of cesa5, cesa2 cesa9, and cesa2 cesa5 cesa9 seeds. Top row: Seeds hydrated directly in ruthenium red and imaged within 1 minute of hydration. Second row: Seeds shaken in water for one hour prior to straining. Third row: Seeds hydrated in water and then stained with S4B. Bottom row: Seeds hydrated in water, then stained with calcofluor white. Single cesa2 and cesa9 mutant seeds appeared similar to cesa2-1 cesa9-1 double mutant seeds (not shown). Bar = 100 µm. From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.  64 S4B staining was observed for cesa5-1 and the cesa2-1 cesa5-1 cesa9-1 triple mutant (Fig. 3.9). No difference was observed for the cesa2-1 cesa9-1 double mutant, therefore only CESA5 is required for mucilage cellulose biosynthesis. Closer examination of wild-type Col-0 S4B-stained seeds illustrates two staining patterns, a diffuse S4B staining pattern above the mucilage pocket and more concentrated staining regions that appear as “rays” extending from the columella in wild-type seeds. The cesa5-1 mutant line shows no diffuse staining of S4B, however, the rays are still present extending from columella (Fig. 3.9). Calcofluor white stained seeds showed a similar fluorescence pattern to S4B, with diffuse staining between stronger staining rays. Again, cesa5-1 lines lost the diffuse staining mucilage. These results support a hypothesis that cellulose synthesized by CESA5 is required for mucilage adherence. 3.2.8 CESA5 is required for mucilage adherence  In order to support our observations that cellulose synthesized by cesa5 is responsible for mediating mucilage attachment, soluble mucilage monosaccharide content and whole seed monosaccharide content was examined to look for potential reductions in total sugar content, or changes in the monosaccharide composition that might affect staining, as opposed to a loss of adherence. Non-adherent mucilage was isolated, hydrolyzed and analyzed using High Performance Liquid Chromatography (HPLC). I isolated whole seed and mucilage monosaccharides, which were then separated on an HPLC by Bruce Downie. I processed and analyzed the raw data from the HPLC. Non-adherent mucilage from cesa5-1 lines showed a ~20% increase in RG I sugars, rhamnose and galacturonic acid (Table 3.1). This increase in RG I sugars was significant (Student’s t-test, P < 0.05). Other monosaccharides showed no significant changes from wild-type monosaccharide compositions. Whole seeds were ground and washed with ethanol and acetone, and treated with amylase to yield a cell wall enriched fraction, which  65 Table 3.1: HPLC analysis of mucilage and cell wall monosaccharide composition from wild type and cesa mutant seeds. Values shown are the averages of three replicates, with standard error. N.d., not determined. From Mendu et al., 2011. www.plantphysiol.org. Copyright American Society of Plant Biologists.  66 was hydrolyzed and analyzed with HPLC. Whole seed enriched cell wall content showed a significant reduction in rhamnose and glucose for cesa5-1 lines. The cause of rhamnose reductions in whole seeds is not clear, but could be caused by the loss of mucilage adherence in cesa5-1 lines. Treatment of seeds with ethanol, acetone and amylose could have resulted in increased loss of cell wall polymers, especially in cesa5-1 lines. Reductions in glucose amounts could be attributed to reduced cellulose biosynthesis, however, we were unable to quantify crystalline cellulose in mucilage. Xylose levels were also reduced in cesa5-1 seeds. Overall, these results suggest that CESA5 has a unique role of the CESA6-like CESAs in mucilage CESAs in mucilage cellulose biosynthesis, and that cellulose synthesized by CESA5 is involved in mediating mucilage adherence. 3.3 Discussion  During seed coat development, three different types of cell walls are synthesized, a primary cell wall, mucilage, and a volcano-shaped secondary cell wall. Here we investigated the role of the CESA6-like CESAs in cellulose biosynthesis in mucilage and the columella. Using reverse genetics, non-redundant roles for CESA2, CESA5 and CESA9 were identified in seed coat cellulose biosynthesis. All three CESAs appear to contribute to radial wall biosynthesis, while CESA5 shows a unique role of the three CESAs in mucilage cellulose biosynthesis. While all three subunits are present during seed coat development, we have shown functional specificity for CESAs in different cell walls in a single cell type. 3.3.1 CESA6-like CESAs involvement during columella biosynthesis  By analyzing double and triple mutants of CESA2, CESA5 and CESA9, the cumulative reduction in secondary wall biosynthesis indicates that all three subunits play a similar role, and yet are not completely redundant. Comparing the double cesa2 cesa9 mutants to either single  67 mutant showed no significant difference in radial wall height (Fig. 3.4, Fig. 3.6, Fig. 3.7) or in total cellulose content (Fig. 3.8). However, the triple mutant was significantly different from the double cesa2 cesa9 line, as well as all single mutant lines for radial wall height and cellulose content. This difference indicates that while all three CESAs participate in seed coat cellulose biosynthesis, their contributions are not all equal. CESA5 has a unique role in mucilage cellulose biosynthesis, which accounts for the difference in total cellulose content. It is not clear why CESA2 and CESA9 do not show a cumulative difference in cellulose content and radial wall height. It is possible that different populations of CSCs can exist in a single cell type, and the contribution of each is not completely redundant. Mutation of all three CESAs results in an enhanced reduction in wall height and cellulose content, indicating all three CESAs are involved in radial wall cell wall biosynthesis, and yet are not completely redundant. Quantification of total seed crystalline cellulose amounts also includes cellulose present in the embryo, and is not specific to the seed coat. However, crystalline cellulose amounts were reduced in similar amounts to the quantitative measurement of radial wall height, suggesting these CESAs are specifically involved in seed coat development. CESA1, CESA2, and CESA3 are expressed during embryo development (Beeckman et al., 2002). Dissection of the seed coat from the embryo would allow for a more specific determination of the impact of CESA2, CESA5 and CESA9 in the seed coat, however obtaining enough tissue would be difficult and laborious. Current theories on the mechanisms of cellulose biosynthesis suggest that three different CESA subunits are required (Taylor et al., 2003; Desprez et al., 2007). In primary cell walls, it is believed that the CSC is composed of CESA1, CESA3, and CESA6 or one of the CESA6-like CESAs (Desprez et al., 2007; Persson et al., 2007). Secondary cell wall biosynthesis has been shown to require CESA4, CESA7 and CESA8 (Taylor et al., 2003). In order to explain all three  68 CESA6-like CESAs having a functional role in columella biosynthesis, I propose that CESA2, CESA5 and CESA9 are all present in seed coat epidermal cells, creating a pool of CESAs that can all be included into multiple CSC that are composed of other core CESAs (CESA1, CESA3; or potentially two of CESA4, CESA7 and CESA8). Given the similarities between the CESA6- like CESAs and CESA6, it is more likely that the CESA2, CESA5 and CESA9 can replace CESA6 in a complex with CESA1 and CESA3. The role of CESA6 in seed coat is unclear and difficult to investigate since double cesa5 cesa6 and triple cesa2 cesa6 cesa9 mutants are not viable (Persson et al., 2007). In addition, we were unable to identify any changes to primary cell wall biosynthesis in epidermal seed coat cells, indicating that CESA1, CESA3 and CESA6 sufficient for this purpose. I have shown that mutation of all three CESA6-like CESAs results in a substantial reduction in radial wall biosynthesis. However, columella biosynthesis appears delayed but ultimately is complete, indicating that other CESAs can complete its biosynthesis. Additionally, radial wall height, and total seed cellulose content is only reduced by ~40%, also indicating the involvement of other CESAs in this process. Which CESAs are involved is not clear, and close examination of mutations in other CESAs, like CESA6 and CESA10, or the secondary wall CESAs, CESA4, CESA7 and CESA8, could help identify which genes are required for the process. 3.3.2 CSC organization during mucilage and columella biosynthesis  The results of this study that demonstrate that the CESA6-like CESAs participate in cellulose biosynthesis in seed coat epidermal cells. CESA5 has a unique function in mucilage cellulose biosynthesis. This indicates that CESAs can have specific roles in a single cell type. It is not clear how CESA5 is regulated to synthesize mucilage cellulose, and whether specific subsets of CESAs can form specific CSCs in one cell. Expression analysis of the CESA6-like  69 family of CESAs indicated that CESA5 is specifically expressed in epidermal cells of the seed coat, while CESA2 and CESA9 were expressed throughout all the seed coat cell layers (Mendu et al., 2011). Differential transcriptional regulation of CESAs is one mechanism cells can use to fine-tune their cell wall composition. However, all three CESAs are presumably present while synthesizing mucilage yet only cesa5-1 shows a mucilage phenotype. It is possible that CESA2 and CESA9 are excluded from forming a functional CSC through an as-yet-unidentified mechanism, since no differences in the mucilage phenotype between cesa5-1 and the cesa2 cesa5 cesa9 triple mutant were observed.  Previously, it has been shown that GLABRA2 (GL2) transcriptionally regulates CESA5 in roots (Tominaga-Wada et al., 2009). GL2 is also known to regulate other genes involved in mucilage biosynthesis such as MUM4/RHM2. MUM4 is specifically up regulated in the seed during mucilage production, between 4 and 10 DPA. Based on publicly available microarray databases, GL2, MUM4, and CESA5 show very similar expression profiles in the seed (not shown; Winter et al., 2007; Bassel et al., 2008; Dean et al., 2011). While CESA5 is repressed by GL2 in roots, it would have to be positively regulated by GL2 during seed development (Tominaga-Waga et al., 2009). Based on CESApromoter::GFP expression results (Mendu et al., 2011), CESA5 is not as strongly up regulated at 10 DPA compared to CESA2 and CESA9, consistent with an earlier expression profile, and a primary role in mucilage biosynthesis. In addition, the expression profile for CESA5 appears more restricted to the outer epidermal layer, where mucilage production is occurring (Mendu et al., 2011). The control of CESA5 by GL2 in different tissues represents an interesting model to investigate cell type specific expression and functions of CESA subunits.  70 It has been shown that CESA proteins can also be regulated directly (Bischoff et al., 2011). New evidence supports a role for CESA5 in fine-tuning cellulose biosynthesis. Mutation of phosphorylation sites present on CESA5 indicates that the phosphorylation state of CESA5 can control CSC velocity through interactions with microtubules (Bischoff et al., 2011). PHYTOCHROME B (PHYB) is a red light photoreceptor that has been shown to regulate cellulose biosynthesis through phosphorylation of CESAs (Bischoff et al., 2011). It is not clear how, or if, PHYB is involved in seed coat mucilage biosynthesis, but it does provide evidence that CESAs can be regulated by protein phosphorylation that could contribute to specific functions within a single cell type (Bischoff et al., 2011). Additionally, this study provides a model for cellulose synthesis where multiple CSC populations can exist, indicating that CESA6- like CESAs can exist in a cell, competing for a position in the CSC. Bischoff et al. (2011) propose that CESA1 and CESA3 are non-redundant and present in all rosette subunits, while CESA5 and CESA6 can compete for the third position (Fig. 3.9). In seed coat epidermal cells, a similar mechanism of CSC formation could be present. During columella biosynthesis, CESA2, CESA5, and CESA9 may all be present and compete for the third position on the CSC. Cellulose deposition might be influenced by the composition of the CSC. In addition, by eliminating one or more of CESA2, CESA5 and CESA9 will reduce the amount of CSC present during columella biosynthesis. It is possible that completely different populations of CESAs are available during mucilage biosynthesis. Alternatively, CESA2 and CESA9 may be excluded from the CSC during mucilage biosynthesis by an unknown mechanism, which only allows CESA5 to be utilized. Given the recent evidence that PHYB can control the phosphorylation state of CESA5, which in turn influences the CSCs interaction with  71 microtubules to control cellulose biosynthesis, it is possible that a similar mechanism is also involved in seed coat epidermal cells (Bischoff et al., 2011). It is not clear which other CESA subunits interact with CESA5 to synthesize cellulose in mucilage. Based on current theories of cellulose biosynthesis, it may be that CESA5 forms a rosette with CESA1 and CESA3. No null mutations have been isolated in either CESA1 or CESA3 due to lethality, which makes it increasingly difficult to determine if they are functional partners with CESA5 in epidermal cellulose synthesis in vivo. Alternatively, CESA5 could form a functional CSC with secondary wall CESAs. Identifying which other CESAs participate in mucilage cellulose biosynthesis would provide insight into how CSCs are formed, and provide further tools for the analysis of cell wall biosynthesis in mucilage secretory cells. 3.3.3 Role of cellulose in mucilage  Hydrated mucilage separates into two discreet layers: a non-adherent layer and an adherent layer. Both layers are composed predominantly by RG I that contains few small side chains composed of arabinan and/or galactan. Xyloglucans have also been identified in mucilage, along with cellulose and HG. Based on many different publications that have analyzed mucilage content, it can be estimated that non-adherent mucilage is composed of 80-90% RGI, 5-10% HG, with the remainder as hemicelluloses (Macquet et al., 2007a; Dean et al., 2007; Arsovski et al., 2009a; Huang et al., 2011; Mendu et al., 2011). The adherent layer shows a similar monosaccharide composition, however, there are increases in the amount of galacturonic acid, arabinose, galactose and glucose (Macquet et al., 2007a; Huang et al., 2011). The amount of glucose is higher in adherent mucilage, which could be explained by inferring that there is a higher proportion of cellulose in adherent mucilage compared to non-adherent mucilage (Macquet et al., 2007a; Huang et al., 2011).  72  Cellulose has been identified in adherent mucilage using multiple kinds of stains and carbohydrate binding modules specific for cellulose (Willats et al., 2001b; Blake et al., 2006; Macquet et al., 2007a; Young et al., 2008; Sullivan et al., 2011; Mendu et al., 2011; Harpaz-Saad et al., 2011). It has been suggested that cellulose could be involved in maintaining connections of mucilage to the parent seed (Macquet et al., 2007a), however, until now, there has been no genetic evidence to confirm this claim. Here I show that CESA5 is required for mucilage adherence, identifying a central role for cellulose in mucilage structure and adherence. In support of this hypothesis, the temperature sensitive mutant, mor1, which is responsible for maintaining microtubule organization, also displays a loss of mucilage adherence (McFarlane et al., 2008). The authors demonstrated that the mucilage defect is not due to defective secretion of mucilage, and suggest that the mutant phenotype is related to microtubules guiding CESAs during cellulose biosynthesis in the plasma membrane (McFarlane et al., 2008; Li et al., 2011). This theory is consistent with cellulose organization and biosynthesis being required for mucilage adherence and structure. It is not clear how cellulose mediates mucilage attachment, but two competing hypotheses can explain the loss of adherence. First, cellulose forms a matrix, interconnected by hemicelluloses that remain attached to the columella. Pectic RG I, comprising the majority of mucilage, is not directly linked to cellulose, but is trapped in the cellulose-hemicellulose matrix of the adherent layer. A second hypothesis involves direct linkages between cellulose and RG I side chains (arabinans and/or galactans), most likely through hydrogen bonding (Zykwinska et al., 2005; Zykwinska et al., 2007a). In support of this second hypothesis, treatment of seeds with dilute hot acid or base does not completely liberate the adherent layer, suggesting that RG I pectins are anchored to the parent seed by direct linkages (Macquet et al., 2007).  73 Generally, pectins are thought to provide a space-filling and water-retaining role in cell wall structure. However, recent in vitro evidence supports the possibility of direct linkages between RG I side chains (arabinans and galactans) and cellulose (Zykwinska et al., 2005; Zykwinska et al. 2007a; Zykwinska et al., 2007b; Zykwinska et al., 2008). It has been suggested that pectins, specifically arabinan and galactan side chains, can bind to cellulose especially in xyloglucan poor cell walls. Seed coat mucilage does not contain much xyloglucan, however, the monosaccharide xylose is present in seed mucilage at roughly equal concentrations to galactose. Images of seed sections immunolabeled withα-xyloglucan antibodies indicate that xyloglucans are present near the columella, the primary cell wall, and diffusely present in the mucilage pocket (Young et al., 2008). The anti-xyloglucan antibodies did not fluoresce as intensely as the RG I specific antibody, CCRC-M36 (Young et al., 2008). Interestingly, there exists a striking difference between xyloglucans and galactans in the different layers of mucilage. Xylose is easily extracted in soluble mucilage, while galactans are largely absent. Alternatively, galactans are present at much higher concentrations in adherent mucilage (Macquet et al., 2007a; Arsovski et al., 2009a; Huang et al., 2011). Mutation of MUM2, a β-galactosidase thought to cleave RG I galactan side chains, alters mucilage hydration properties, so that mucilage can only be released using Na2CO3 (Dean et al., 2007). MUM2 could decrease the interactions between cellulose and pectins, resulting in mucilage extrusion and the separation of the non-adherent faction from the adherent fraction. Ultimately, the identification of CESA5 as a key factor in mediating mucilage attachment is an interesting discovery that will improve our understanding of pectin-cellulose interactions.   74 4 Chapter 4: CESA3 participates in mucilage cellulose biosynthesis 4.1 Introduction  The Arabidopsis genome encodes 10 different CESAs (Delmer, 1999; Richmond and Somerville, 2000). Multiple lines of evidence suggest that three different CESAs are required to form one active cellulose synthase complex (CSC; reviewed in Somerville, 2006.) CESAs have been classified into three groups based on their role in primary or secondary cell wall biosynthesis (Taylor et al., 2003; Persson et al., 2007; Desprez et al., 2007). CESA1, CESA3 and CESA6 are considered the core components of primary wall cellulose synthesis (Persson et al., 2007; Desprez et al., 2007). CESA2, CESA5 and CESA9 are thought to be partially redundant to CESA6 in primary wall biosynthesis, and genetic evidence suggests that each of these CESA can form a functional CSC with CESA3 and CESA1 (Persson et al., 2007; Desprez et al., 2007). CESA4, CESA7 and CESA8 are considered to be the core components of secondary cell wall cellulose biosynthesis (Taylor et al., 2003). CESA10 is the least understood CESA and has no association with any specific type of cell wall, or any other CESA. The discovery that CESA5 is required for cellulose biosynthesis during mucilage production (chapter 3; Sullivan et al., 2011; Mendu et al., 2011; Harpaz-Saad et al., 2011) suggests that other CESAs may also be required. Here, the role of other CESAs during mucilage biosynthesis is investigated further using a reverse genetics approach. 4.1.1 Primary wall CESAs Cellulose biosynthesis in the primary cell wall is required for normal plant development. CESA1 and CESA3 are considered to be key factors in primary cell wall biosynthesis, and are required for normal plant development. As such, null mutations in CESA1 and CESA3 are male gametophytic lethal (Persson et al., 2007; Daras et al., 2009). CESA6, and the CESA6-like group  75 of CESAs (CESA2, CESA5 and CESA9) are also involved in primary wall cellulose. CESA6 and the CESA6-like group of CESAs are all believed to be able to form a functional CSC with CESA1 and CESA3 (Persson et al., 2007; Desprez et al., 2007; Bischoff et al., 2011). However, due to redundancy, null mutations in these 4 CESAs are not lethal (Persson et al., 2007; Desprez et al., 2007). Double cesa5 cesa6 mutants are seedling lethal, while triple cesa2 cesa6 cesa9 mutants are male gametophytic lethal (Persson et al., 2007; Desprez et al., 2007). Many missense mutations are available in CESA1 and CESA3 that are still capable of synthesizing cellulose. One such allele, radial swelling1 (rsw1-1), is a temperature sensitive allele that at the non-restrictive temperature has many phenotypes, including swollen roots (Williamson et al., 2001). Two mutant cesa3 alleles, isoxaben resistant1 (ixr1-1) and ixr1-2, were isolated in a screen for resistance to the herbicide isoxaben (Scheible et al., 2001). Isoxaben inhibits the incorporation of glucose into cell walls, and is believed to be a potent and specific cellulose biosynthesis inhibitor (Heim et al., 1990). Homozygous ixr1-1 and ixr1-2 lines show increased resistance to the actions of the herbicide and were mapped to the genomic locus of CESA3 (Heim et al., 1990; Scheible et al., 2001). Both mutations lie in the C-terminal end of CESA3, with ixr1-1 causing a guanine to adenine transition (G11204A) that results in a Glycine to Asparagine (G998A) substitution located in a transmembrane domain, while ixr1-2 contains a cytosine to thymine (C11372T) transition that results in a threonine to isoleucine substitution (T942I) in an apoplastic span of the protein between two transmembrane domains (Fig. 4.1; Scheible et al., 2001; Harris et al., 2012). Plants homozygous for either mutation are fully viable (Scheible et al., 2001; Harris et al., 2012), however, their effects on plant growth and cellulose biosynthesis have not been fully investigated (Scheible et al., 2001).  76   Figure 4.1: CESA3 protein structure and amino acid changes caused by the ixr1-1 and ixr1-2 point mutations. Purple areas indicate Zing finger homology. Red indicates the approximate location of the catalytic domain. Teal indicates transmembrane domains.   77  Recently, the ixr1-2 allele has been shown to affect the velocity of CSCs through the plasma membrane, which consequently modifies the degree of crystallinity of the cellulose produced (Harris et al., 2012). ixr1-2 was compared to a novel mutant allele of CESA1 termed aegeus, and both mutants show reduced cellulose crystallinity in hypocotyls and stems (Harris et al., 2012). The reduction in cellulose crystallinity is presumably due to changes in the speed of the CESA rosette in the plasma membrane (Harris et al., 2012). Using GFP Fusions of CESA3 and CESA6, researchers were able to show increases in the speed of the CESA particle velocity when ixr1-2 was included in the CSC (Harris et al., 2012). It is not clear exactly how the ixr1-1 mutation affects cellulose crystallinity, or CESA velocities. While the CESA1 through CESA9 genes have been assigned specific functions, including tissue and cell wall specific cellulose biosynthesis functions, no biological role has been identified for CESA10. It has been reported that CESA10 is expressed in seeds and in the base of the rosette leaves, at the side of the attachment of the inflorescence (Doblin et al., 2002). In order to identify other CESAs involved in mucilage cellulose biosynthesis with CESA5, the expression pattern of all 10 CESAs was examined during seed coat development. Plants homozygous for mutations in the CESA genes most highly expressed during mucilage biosynthesis were identified and examined for defects in mucilage biosynthesis. CESA3 is expressed during seed coat development, and both ixr1-1 and ixr1-2 mutant alleles of CESA3 shows defects in mucilage adherence and increased density in the ray structures, which is due to changes in cellulose biosynthesis. These changes also altered the distribution of pectin components in the mucilage halo. In whole seeds, crystalline cellulose and glucose amounts were reduced in ixr1-1 and ixr1-2 seeds, indicating a general reduction in  78 crystalline cellulose similar to cesa5-1 seeds. These results indicate that CESA3 is involved in mucilage biosynthesis. 4.2 Results 4.2.1 CESA3 and CESA10 are highly expressed during mucilage biosynthesis  During seed coat development, mucilage is actively synthesized in large amounts at approximately 6-8 days post anthesis (DPA). Thus during this period there is a large increase in synthesis of cell wall components, and therefore requires increased expression of genes required for cell wall biosynthesis. MUCILAGE MODIFIED 4 (MUM4) encodes a rhamnose synthase required for mucilage biosynthesis, and it is strongly up regulated from 4 DPA to 7 DPA (Western et al., 2004). As such, I looked for CESA genes that have increased expression from 4 DPA to 7 DPA. Using publically available microarray data and transcriptomics resources available for Arabidopsis, the expression of all 10 CESAs was examined during seed development (Fig. 4.2A; Arabidopsis seed expression browser, Winter et al., 2007; Bassel et al., 2008). I focused on CESA genes expressed most highly at 7 DPA, the linear cotyledon stage, since mucilage production is at its highest at this stage (Western et al., 2001; Haughn and Chaudhury, 2005). The maturation green stage is roughly analogous to 10 DPA, when mucilage production has ceased and columella biosynthesis is underway (Western et al., 2000; Haughn and Chaudhury, 2005).   Based on the microarray-based expression results for seed coat specific expression taken from the seed tissue specific data set (Fig. 4.2A; Le et al., 2010), CESA3 showed the highest expression values at the linear cotyledon stage, followed by CESA10, CESA2 and CESA5 (Fig. 4.2A). The expression of CESA3, CESA10, and CESA5 was highest at 7 DPA. CESA2 was highly expressed at the heart embryo stage, approximately 4 DPA, but expression was reduced  79 Figure 4.2: Expression of the CESA gene family during seed development. A) Seed coat specific expression of CESA1 through CESA10 during seed coat development. Expression data is derived from Le et al., 2010. B) Whole seed expression of CESA1 through CESA10. Expression data derived from Winter et al., 2007. 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Pre-gloular globular (3 DPA) heart linear (7 DPA) mature green (10 DPA) R el at iv e e xp re ss io n va lu e Seed stage CESA1 CESA2 CESA3 CESA4 CESA5 CESA6 CESA7 CESA8 CESA9 CESA10 A B 0 200 400 600 800 1000 1200 Torpedo Walking-stick Curled cotyledon Green cotyledon Mature green R el at iv e e xp re ss io n va lu e Seed stage CESA1 CESA2 CESA3 CESA4 CESA5 CESA6 CESA7 CESA8 CESA9 CESA10 MUM4 (4 DPA) (5-6 DPA) (7 DPA) (8-9 DPA) (10 DPA)  80 from 4 DPA to 7 DPA. CESA8 also showed a slight increase in expression at 7 DPA compared to other time points, however, its expression was much lower compared to CESA3, CESA10, CESA5 and CESA2. CESA1 and CESA6 displayed relatively constant expression levels throughout seed coat development. The expression pattern of MUM4 was also examined in this data set as a positive control, however, the data wasn’t included in the graph since its expression was over 4 fold higher than any CESA. The expression of MUM4 increased drastically in the seed coat from the heart stage (~4 DPA; 1018.76 ± 33.19) to the linear cotyledon stage (~7 DPA; 7764.91 ± 503.37; data not shown), confirming that this time point is representative of mucilage biosynthesis.  Expression of CESAs was also analyzed in whole seed expression data taken from the data set in Winter et al., 2007. Results observed (Fig. 4.2B) were similar to those of the seed coat specific data set. Seed stage 7 is the closest time point to 7 DPA, when mucilage production is highest. Seed stage 10 is roughly analogous to 11 DPA, which is near the end of columella biosynthesis. Similar to the seed coat specific data set, CESA3 showed the highest expression of all CESAs during seed development. At seed stage 7 (7 DPA), the next highest expressed CESA was CESA2, CESA6, CESA10, CESA1 and CESA5, respectively. CESA2, CESA10 and CESA5 show a sharp decrease in expression from seed stage 7 to seed stage 8. This expression pattern is similar to MUM4 in this data set.  CESA3, CESA1 and CESA6 did not show the same decrease. The whole seed expression data set includes expression data from the seed coat and the embryo, indicating that CESA3 is also expressed during embryo development. When comparing the seed coat specific expression data set to the whole seed expression data set, CESA5 and CESA10 appear to be more specifically involved in seed coat cellulose biosynthesis than embryo cell wall  81 biosynthesis. From these results, CESA3 and CESA10 appear to be the best candidates for involvement in cellulose biosynthesis during mucilage deposition. 4.2.2 CESA5::GFP and CESA3::GFP are localized in the cytoplasmic column during mucilage biosynthesis  In addition to analyzing the expression of CESAs during seed coat development, the localization of GFP::CESA5 and GFP::CESA3 was examined at 7 DPA during mucilage biosynthesis (Fig. 4.3). The localization of GFP::CESA5 has previously been examined, and shown to be present during all stages of seed coat development (Sullivan et al., 2011). At 7 DPA, GFP::CESA5 can be observed surrounding the columella, localized to the plasma membrane (Fig. 4.3A). Similarly GFP::CESA3 is also visible in the columella at 7 DPA, however, the signal was much weaker compared to GFP::CESA5 (Fig. 4.3B). The localization of GFP::CESA6 was used as a negative control (Fig. 4.3C). In contrast to GFP::CESA5 and GFP::CESA3, no fluorescence signal from GFP::CESA6 was detected in seed coat epidermal cells at 7 DPA when imaged under similar conditions as CESA3. This result shows that CESA3, and confirms that CESA5, are expressed in seed coat epidermal cells during mucilage biosynthesis, and are localized to the cytoplasmic column when mucilage is being synthesized. 4.2.3 ixr1-1 and ixr1-2 seeds have altered mucilage structure and adherent halo properties  Having identified that CESA3 and CESA10 are highly expressed in the seed coat during mucilage biosynthesis and that CESA3 is localized to the cytoplasmic column, a reverse genetics approach was used to identify cesa mutants with altered seed coat mucilage. Plants homozygous for mutations in CESA1 (any1-1; Fujita & Wasteneys, unpublished results), CESA3 (ixr1-1; ixr1- 2; Scheible et al., 2001), CESA6 (prc1-1; Fagard et al., 2000), CESA8 (irx1-5) and CESA10 (cesa10-1 SALK_052533; cesa10-2 SALK_150533) were obtained. CESA2 and CESA9 were   82 ! Figure 4.3: Localization of CESA3, CESA5 and CESA6 in seed coat epidermal cells. A) GFP::CESA5. B) GFP::CESA3. C) GFP::CESA6. Top row: merged DIC and GFP channels. Middle row: DIC channel. Bottom row: GFP channel. Bar = 10 µm.  83 shown previously to not be involved in mucilage biosynthesis (Mendu et al., 2011). Seeds from each mutant were shaken in water, stained with ruthenium red, and compared to Col-0 and cesa5-1 seeds (Fig. 4.4). Similar to previous mucilage staining experiments, Col-0 seeds show a large mucilage halo that stains bright red with ruthenium red (Fig. 4.4A). Mutant cesa5-1 seeds show a significant decrease in the size of the mutant halo (Fig. 4.4B). ixr1-1 and ixr1-2 seeds showed reduced ruthenium red staining compared to Col-0 (Fig. 4.4C and D, respectively), however, this phenotype was not nearly as severe as cesa5-1 seeds (Compare Fig. 4.4B to C and D). Ruthenium red staining of ixr1-1 and ixr1-2 seeds is variable and the mucilage halo often seemed disorganized compared to Col-0 seeds. Some seeds appear very similar to cesa5-1 seeds, while other show increased staining, almost similar to Col-0 seed. Mutations in other CESAs, including cesa10-1, cesa10-2 (not shown), any1-1, prc1-1 and irx1-5 did not show any significant differences Col-0 seeds (Fig. 4.4E-H). Homozygous ixr1-1 plants were outcrossed to Col-0 plants. The altered mucilage phenotype of ixr1-1 co-segregated with the ixr1-1 point mutation in the F2 progeny. From a population of 40 F2 individuals, seven contained the ixr1-1 phenotype (Chi squared = 1.55 > 3.84). Four individuals displaying the ixr1-1 mucilage phenotype were sequenced and shown to have the ixr1-1 point mutation, while two individuals that did not have the phenotype also lacked the point mutation.  The initial mucilage hydration properties of ixr1-1 and ixr1-2 seed were examined to determine if there are any changes in mucilage release and non-adherent mucilage amounts (Fig. 4.5). Col-0 and cesa5-1 seeds show a large mucilage halo that is released almost immediately when hydrated (Fig. 4.5A and B, respectively. ixr1-1 and ixr1-2 seeds showed no major differences to Col-0 seeds and cesa5-1 seeds in terms of mucilage release (Fig. 4.5C and D,  84  Figure 4.4: Ruthenium red staining of cesa mutant seeds. Seeds hydrated and shaken in water for 1 hr, followed by ruthenium red staining. A) Col-0, B) cesa5-1, C) ixr1-1, D) ixr1-2, E) cesa10-1, F) any1-1, G) prc1-1, and H) irx1-5. Bar = 50 µm.  !! "! #! $! %! &! '! (!  85  Figure 4.5: Initial mucilage hydration halo and unstained mucilage halo of cesa5-1, ixr1-1 and ixr1-2 seeds. A-D) Seeds hydrated in ruthenium Red. E-F) Unstained seeds hydrated in water. A and E) Col-0, B and F) cesa5-1, C and G) ixr1-1, D and H) ixr1-2. Bar = 50 µm. !! "! #! $! %! &! '! (!  86 respectively). However, ixr1-1 and ixr1-2 have ray-like structures in the adherent mucilage halo that are more visible than those of wild type and cesa5. These ray-like structures extend perpendicularly from the seed surface above the columellae (Fig. 4.5C and D) and are even more pronounced when viewing mucilage without staining (Fig. 4G-H). Unstained Col-0 seeds show very faint rays extending from the seed (Fig. 4.5E), while cesa5-1 seeds show no visible structure in what remains of the mucilage halo (Fig. 4.5F). The distribution of cellulose in the mucilage halo of ixr1 seeds was also investigated using pontamine fast S4B (S4B), a stain that fluoresces more intensely in the presence of cellulose (Anderson et al., 2010). Similar to previous analyses, Col-0 mucilage and the primary cell wall, which remains attached to the parent seed at the top of the columella (arrow) have intense fluorescence when stained with S4B (Fig. 4.6A). Two distinct domains can be observed in the adherent mucilage halo when stained with S4B: intensely staining rays above the columella and the diffuse mucilage above the mucilage pocket. S4B stained cesa5-1 mucilage showed drastically reduced staining of the rays and an almost complete loss of the diffuse staining (Fig. 4.6B). S4B stained ixr1-1 mucilage showed a similar loss of the diffuse mucilage but the rays were more strongly labelled, longer and with increased diameter when stained with S4B (Fig. 4.6C) than that of wild type or cesa5-1. ixr1-2 seed mucilage had a more intensely staining halo than ixr1-1 seeds, especially focused on the rays. ixr1-2 seeds appear to have a shorter halo size than Col-0 mucilage when stained with S4B, and the rays appear much larger in diameter than Col-0 rays (Fig. 4.6D). The loss of the diffuse staining does not appear as severe in ixr1-2 mucilage.   87  Figure 4.6: S4B stained cellulose in cesa5-1, ixr1-1 and ixr1-2 adherent mucilage. A) Col-0, B) cesa5-1, C) ixr1-1, D) ixr1-2. Bar = 10 µm. !! "! #! $!  88 The length of the rays in ixr1-1 and ixr1-2 seeds was compared to Col-0 and cesa5-1 seeds. In total, 20 rays were measured from the top of the columella to the end of the ray, from 4 different seeds, for three biological replicates (Fig. 4.7). Col-0 seeds had a range of ray lengths from 69 to 83 µm. cesa5-1 seeds had on average much shorter rays, ranging from 49 to 56 µm. ixr1-1 seeds had on average much longer rays, and were much more variable in length, ranging from 87 to 110 µm. Finally, ixr1-2 seeds showed extremely variable ray length in the three biological replicates, ranging from 54 µm, roughly the size of cesa5-1 rays, to 98 µm which is longer than the length of Col-0 rays. These results demonstrate the variability of mucilage structure in ixr1-1 and ixr1-2 seeds, consistent with ruthenium red staining. Additionally, ixr1-1 and ixr1-2 seeds have increased cell wall density in rays that corresponds with increased cellulose staining. 4.2.4 ixr1-1 and ixr1-2 mutations result in increased disorganization of mucilage structure and cell wall polysaccharides To further characterize the distribution of cellulose in adherent mucilage, seeds were immunolabeled with two carbohydrate-binding modules (CBM) and two pectin specific antibodies (JIM5, CCRC-M36). CBM3a binds specifically to crystalline cellulose structures while CBM28 is more specific to amorphous cellulose structures (Blake et al., 2006; Dagel et al., 2011). Both are labeled with a 6X histidine tag that allows for indirect immunological detection of cellulose epitopes (Blake et al., 2006). JIM5 is a rat monoclonal antibody that binds specifically to partially methyl-esterified homogalacturonan (HG; Knox et al., 1990; Willats et al., 2001). Finally, CCRC-M36 is a mouse-derived monoclonal antibody specific to the un- branched backbone of rhamnogalacturonan I (RG I; Pattathil et al., 2010). Negative controls lacking the primary antibody showed no detectable fluorescence in the mucilage halo when viewed under similar imaging conditions (not shown). Similar results were observed for two  89  Figure 4.7: Ray length in S4B stained mucilage. For evaluation of variation within biological replicats, each bar represents one biological replicate for each genotype. Error bars are standard error measurements for each replicate.  !"#$%&&&&&&&&&&!"#$%&'(((((((()*+'&'(((((((((()*+'&,! '()&#*+ ,-.&/01 2! !" #!" $!" %!" &!" '!!" '#!"  90 biological replicates for CBM3 and JIM5. Seeds were first hydrated in water for one hour before immunolabelling. Following CBM3a immunolabelling, seed mucilage was also stained with S4B (Fig. 4.8). In Col-0 adherent mucilage, there was an obvious separation of the epitopes bound by S4B and with CBM3a (Fig. 4.8A). As previously described, S4B stains the adherent layer, highlighting the rays above the columella and the diffuse mucilage (Fig. 4.8E). CBM3a strongly labelled the outer periphery of the adherent mucilage halo, which was not stained by S4B (Fig. 4.8A). Both CBM3a and S4B appear to label the inner adherent mucilage halo (Fig. 4.8I). In addition both CBM3a and S4B appear to label the rays above the columella, with S4B labelling the base of the ray, while CBM3a labels top of the ray at the periphery of the mucilage halo (See arrow, Fig. 4.8I). CBM3a labelling of cesa5-1 seeds shows a reduction in halo size and a more disorganized mucilage halo, consistent with a general reduction in cellulose biosynthesis and a loss of mucilage adherence (Fig. 4.8B). S4B staining is reduced, but still stains the rays above the columella, while CBM3a labels the ends of the rays above the S4B staining (Fig. 4.8I). CBM3a labelling of ixr1-1 seeds appeared similar to Col-0 seeds, however, the mucilage halo appears slightly larger than wild-type and is less labelling of the rays (Fig. 4.8C). In ixr1-2 seeds, some areas of the CBM3a labelled mucilage halo were much longer that other areas (Fig. 4.8D). This labelling pattern was not consistent, with some regions of the ixr1-2 adherent halo appearing similar to the size of the Col-0 adherent halo. CBM28 labelling was reduced in Col-0 seeds compared to CBM3a such that the laser intensity and detection had to be greatly increased when imaging CBM28. In Col-0 seeds, CBM28 appears to label the outer periphery of the mucilage halo and also appeared most intense around the ray structures (Fig. 4.9A). In cesa5-1 seeds, there appeared to be less specific  91 Figure 4.8:!CBM3a labelling of adherent mucilage. Top row, CBM3a channel, Middle row: S4B channel, Bottom row: Merged image. A, E, I) Col-0. B, F, J) cesa5-1. C, G, K) ixr1-1. D, H, L) ixr1-2. Bar = 10 µm. !"#$% ! &'"! #()*(! !+,-.! !"#$%&'! ()*'&'! ()*'&+!/! !!"! 0! 1! 2! 3! 4! 5! 6! 7! 8!  92 l  Figure 4.9: CBM28 labelling of adherent mucilage. Top row, CBM28 channel, Middle row: S4B channel, Bottom row: Merged image. A, E, I) Col-0. B, F, J) cesa5-1. C, G, K) ixr1-1. D, H, L) ixr1-2. Bar = 10 !m. !"#$% ! &'"! #()*(! !+,-.! !"#$%&'! ()*'&'! ()*'&+!/! !!"! 0! 1! 2! 3! 4! 5! 6! 7! 8!  93 labelling of the rays (Fig. 4.9B). In both ixr1-1 and ixr1-2 seeds, CBM28 appeared to label less of the mucilage halo, with very little fluorescence detected (Fig. 4.9C and D, respectively). CBM28 did not strongly label rays in ixr1-1 adherent mucilage, however, ixr1-2, mucilage showed stronger labelling of the rays (Fig. 4.9C and D respectively, top row).  When examining the distribution of low methyl esterified pectins using JIM5, strong differences were observed between ixr1-1 and ixr1-2 mutant seeds and Col-0 seeds. Col-0 seed mucilage shows a clear JIM5 labelling pattern, with labelling throughout the adherent layer and some specific labelling localized to the rays above the columella (Fig. 4.10A). This labelling pattern is almost completely absent in cesa5-1 seeds (Fig. 4.10B). Some labelling is still observable at the base of the rays. In both ixr1- 1 and ixr1-2 seeds, the labelling with JIM5 was drastically altered. ixr1-1 seeds show a clustering of JIM5 labelling localized around the columella and on the rays (Fig. 4.10C). JIM5 labelling of ixr1-2 seeds appears similar to Col-0 seeds, however, there appears to be an increase in labelling around the rays, with the labelling appearing to surround the ray (Fig. 4.10D). The labelling pattern also appeared to be more robust as well, with a very large increase in the diameter of the rays labeled with JIM5 (See arrow, Fig 4.10L). The distribution of RG I was also examined in adherent mucilage using CCRC- M36. In Col-0 adherent mucilage, CCRC-M36 labels the entire adherent mucilage halo, with increased labelling located at the periphery (Fig. 4.11A). In contrast, cesa5-1 seeds show reduced labelling in the adherent halo, with strong labelling localized around the base of the rays (Fig. 4.11B). In ixr1-1 mucilage, CCRC-M36 labels the top of the rays in a very distinct way, however, overall labelling appears to be reduced compared to Col-0  95  Figure 4.11: CCRC-M36 labelling of adherent mucilage. Top row, CCRC-M36 channel, Middle row: S4B channel, Bottom row: Merged image. A, E, I) Col-0. B, F, J) cesa5-1. C, G, K) ixr1-1. D, H, L) ixr1-2. Bar = 10 µm. !!"!#$ %&! '()! $*+,*! !-.#/! !"#$%&'! ()*'&'! ()*'&+! 0! 1!2! 3! 4! 5!6! 7! 8! !!)! 9!  96 seeds (Fig. 4.11C). In ixr1-2 seeds, labelling is observable at the outer edge of the mucilage halo, and appears more disorganized and reduced compared to Col-0 seeds (Fig. 4.11D). Overall, I observed many changes in the distribution of the epitopes of cell wall polymers ixr1-1 and ixr1-2 seed mucilage. The size of the mucilage halo based on CBM3a labelling of crystalline cellulose appears increased, while amorphous cellulose epitopes are reduced in ixr1-1 and ixr1-2 seeds. Besides changes in cellulose epitopes, pectin epitopes were also altered. Most notably, low-methyl-esterified HG epitopes were strongly increased specifically around the rays. Immunolabelling also demonstrates changes to the width of the mucilage halo when labelled with CBM3a or CCRC-M36. These changes in mucilage polysaccharide organization and distribution observed in ixr1-1 and ixr1-2 seeds result in the increased visibility of the rays seen in unstained seeds. 4.2.5 Mucilage hydration properties in ixr1 mutants modified by calcium availability  Based on the changes in the JIM5 label distribution in ixr1-1 and ixr1-2 seeds, I hypothesized that changes in the distribution of HG could be affecting the mucilage hydration of these mutants. HG polymers can be cross-linked through calcium bridges (Ca2+). The positive charges of Ca2+ are thought to balance the charges of de-methylated galacturonic acid molecules present on HG, therefore the methyl esterification state of HG affects this crosslinking. Therefore ixr1-1 and ixr1-2 mutant seeds were hydrated in a solution of CaCl2, which is expected to increase cross-linking of mucilage, and with ethylenediaminetetraacetic acid (EDTA), which sequesters divalent cations such as Ca2+ and is expected to reducing cross-linking. If the increased JIM5 epitopes are indicative of changes in the cohesiveness of mucilage in ixr1-1 and  97 ixr1-2 seeds, then CaCl2 should enhance any mucilage release phenotypes, while EDTA should rescue these phenotypes.  In Col-0 seeds shaken in water and stained with ruthenium red, a standard mucilage halo is observable (Fig. 4.12A). In cesa5-1 seeds, the mucilage falls off, as previously reported (Fig. 4.12B). In ixr1-1 and ixr1-2, the phenotype resembles cesa5-1 in that the mucilage fails to stain well with ruthenium red and the halo is not as large as wild type (Fig. 4.12C and D, respectively). When seeds are treated with EDTA, Col-0 seeds have a larger mucilage halo, presumably due to reduced Ca2+ crosslinking between HG, allowing increased expansion (Fig. 4.12E). cesa5-1 seeds appear to have a much smaller mucilage halo than Col-0 seeds, although some mucilage stained with ruthenium red is still observable (Fig. 4.12F). Both ixr1-1 and ixr1-2 seeds show very large mucilage halos, closely resembling Col-0 seeds (Fig. 4.12G and H, respectively) suggesting that the observed differences in mucilage between irx1 seeds and wild type are due to increased cation cross-linking of HG in the mutants.  When seeds are treated with CaCl2, Col-0 seeds have smaller mucilage halos than water treated seeds (Fig. 4.12I). cesa5-1 seeds also show a small mucilage halo stained with ruthenium red, which appears slightly smaller than the width of Col-0 halos (Fig. 4.12J). In ixr1-1 and ixr1- 2 seeds, some seeds appear not to release any mucilage at all (Fig. 4.12K and L, respectively). ixr1-1 seeds showed patchy mucilage release, with some areas of the seed releasing small amounts of mucilage, while most did not release any mucilage. ixr1-2 seeds were more variable than ixr1-1 seeds, with some seeds not releasing any mucilage, some seeds showing a patchy mucilage release, and other seeds displaying large, complete mucilage halos, (Fig. 4.12L). Overall, these results demonstrate that ixr1-1 and ixr1-2 mucilage has altered HG distribution that affects mucilage hydration and response to calcium  98  Figure 4.12: Effects of calcium on ixr1-1 and ixr1-2 mucilage hydration. Top row: shaken in water, Middle row: shaken in 0.05 M EDTA, Bottom row: shaken in 0.1 M CaCl2. A) Col-0. B) cesa5-1. C) ixr1-1. D) ixr1-2. Bar = 50 µm.   !"#!! !"#$%!! ! ! ! ! !"#$!! !"#$!! !"!# ! !"#$ !"#$%  99 4.2.6 ixr1-1 and ixr1-2 have reduced monosaccharide amounts in mucilage and whole seeds  In order to understand the changes in mucilage caused by ixr1-1 and ixr1-2, soluble mucilage and whole seed monosaccharide compositions were analyzed.  Seeds were shaken in water, and the non-adherent mucilage fraction was isolated, hydrolyzed and analyzed using High Performance Anion Exchange Chromatography (HPAEC; Fig. 4.13A). Results are from one biological replicate consisting of four technical replicates. Compared to Col-0 seeds, cesa5-1 displayed significantly increased amounts of rhamnose, and galacturonic acid in the non- adherent mucilage fraction (Col-0, rhamnose, 45.85  ± 0.51 ng sugar/mg seed, galacturonic acid, 47.10 ± 0.64 ng sugar/mg seed; cesa5-1, rhamnose 56.29 ± 0.58 ng sugar/mg seed, and galacturonic acid, 58.19 ± 0.24 ng sugar/mg seed), consistent with a loss of mucilage adherence. Galactose amounts were largely unchanged in cesa5-1 seeds while glucose showed a significant decrease in cesa5-1 seeds (glucose, Col-0, 2.93 ± 0.27 ng sugar/mg seed; cesa5-1, 1.00 ± 0.08 ng sugar/mg seed; P < 0.01) compared to Col-0 non-adherent mucilage, consistent with a role for CESA5 in cellulose biosynthesis. Both ixr1-1 and ixr1-2 mutant lines showed a significant decrease in rhamnose amounts (ixr1-1, 34.97 ± 0.84 ng sugar/mg seed, P < 0.001; ixr1-2, 28.49 ± 1.33 ng sugar/mg seed, P < 0.001) and galacturonic acid amounts (ixr1-1, 38.24 ± 0.45 ng sugar/mg seed, P < 0.005; ixr1-2, 30.20 ± 0.28 ng sugar/mg seed, P < 0.001) in non-adherent mucilage. Non-adherent mucilage from ixr1-1 seeds showed similar amounts of glucose compared to Col-0 seeds (4.49 ± 1.88 ng sugar/mg seed), while ixr1-2 non-adherent mucilage showed a significant decrease in amounts of glucose (1.23 ± 0.06 ng sugar/mg seed, P < 0.05). The reduction in non-adherent mucilage suggest increased mucilage adherence in ixr1-1 and ixr1-2 seeds, consistent with larger mucilage halos observed through staining and immunolabelling.  100 Figure 4.13: Monosaccharide composition of non-adherent mucilage and whole seed AIR. A) Non-adherent mucilage fraction B) Whole seeds AIR. (a) indicates significantly increased compared to wild type, (b) indicates significantly decreased compared to wild type, Students t-test, P < 0.05. Unless indicated, no significant difference from wild type was observed. Error bars indicate standard error, n= 4. !" #!" $!" %!" &!" '()*+,-." /)0)1234,+51" 6157" !" 8" #" 9" $" :" %" ;" /)0)12,-." /031,-." <=0,-." >,0?!" !"#$%&'( )*+'&'( )*+'&,( !" :!" 8!!" 8:!" #!!" #:!" 9!!" 9:!" $!!" '()*+,-." /)0)1234,+51" )157" 64)@5+,-." /)0)12,-." /031,-." <=0,-."  N on -a dh er en t m uc ila ge  fr ac tio n (! g su ga r/m g se ed )! W ho le  se ed  A IR  m on os ac ch rid e co nt en t ( !g  su ga r/m g se ed )! A! B! a a b b b b b b b b b b b b b a b b b b b b b b b  101 It is possible that these changes are due to reduced cell wall biosynthesis, therefore the monosaccharide content of whole seeds was determined to examine total cell wall content in whole seeds (Fig. 4.12B). When comparing amounts of rhamnose, cesa5-1 shows no significant difference from Col-0 seeds (Col-0 188.15 ± 5.67 ng sugar/mg seed; cesa5-1, 192.8 ± 8.9 ng sugar/mg seed). Surprisingly, both ixr1-1 and ixr1-2 mutant lines showed a strong decrease in amounts of rhamnose (ixr1-1, 155.36 ± 6.0, P < 0.001; ixr1-2, 147.82 ± 2.14, P < 0.005). Amounts of galacturonic acid were decreased in all mutant lines analyzed (Col-0, 318.56 ± 30.32 ng sugar/mg seed; cesa5-1, 204.12 ± 18.28 ng sugar/mg seed, P < 0.05; ixr1-1, 195.66 ± 7.70 ng sugar/mg seed, P < 0.01; ixr1-2, 178.48 ± 7.62 ng sugar/mg seed, P < 0.01). Glucose amounts were significantly decreased in cesa5-1 and ixr1-1 seeds (Col-0, 108.81 ± 4.22 ng sugar/mg seed; cesa5-1, 77.08 ± 2.70, P < 0.01; ixr1-1, 79.48 ± 5.06 ng sugar/mg seed, P < 0.01) compared to Col-0, while ixr1-2 seeds were not significantly different (ixr1-2, 95.41 ± 5.23 ng sugar/mg seed, P > 0.05). Arabinose levels were largely unchanged in cesa5-1, ixr1-1 and ixr1-2 seeds compared to Col-0 seeds. Galactose levels were reduced in ixr1-1 and ixr1-2 seeds, while xylose levels were decreased in cesa5-1, ixr1-1 and ixr1-2 compared to Col-0 seeds. The reduction in many cell wall components indicates an overall reduction in cell wall content in ixr1-1 and ixr1-2 seeds. 4.2.7 Reduced crystalline cellulose amounts in ixr1-1 and ixr1-2 whole seeds  Since ixr1-1 and ixr1-2 are mutant alleles of a CESA subunit, total crystalline cellulose amounts in whole seeds were quantified using the Updegraff method (Fig. 4.14; Updegraff, 1969). Crystalline cellulose amounts were significantly reduced in cesa5-1, and ixr1-1 seeds (Col-0, 110.05 ± 12.15 mg cellulose/g seed; cesa5-1, 76.04 ± 5.57 mg cellulose/g seed, P < 0.05;  102  Figure 4.14: Crystalline cellulose amounts in ixr1-1 and ixr1-2 whole seeds. Asterisks indicate significantly different from wild type (Students t-test, P < 0.05). Error bars indicate standard error, n=4. Results are from one representative biological replicate, from two total biological replicate that all showed the same trends. !" #!" $!" %!" &!" '!!" '#!" '$!" Col-0        cesa5-1       ixr1-1         ixr1-2  !  !"#$%&' '()*+,*' '-'.$*+/ 01+,*'' -'.$*21 +$**3+ ,*''+4& ''5!  103 ixr1-1, 81.76 ± 5.17 mg cellulose/ g seed, P < 0.05). However, ixr1-2 seeds did not show an amount of crystalline cellulose significantly different from Col-0 seeds (ixr1-2, 88.16 ± 6.54 mg cellulose/g seed, P > 0.05). These results are consistent for two independent biological replicates, with one representative replicate shown. These reductions in crystalline cellulose amounts are consistent with mutations affecting cellulose biosynthesis in mucilage and whole seeds. 4.3 Discussion Cellulose plays an important role in the structure of mucilage, and is required for the retention of the adherent layer.  In the previous chapter, CESA5 was shown to be necessary for proper cellulose biosynthesis during mucilage deposition, and is required for mucilage adherence. While CESA5 has a clear role in mucilage biosynthesis, it is not obvious what other CESAs, if any, are also involved. In order to determine whether other CESAs are involved in mucilage biosynthesis, the expression of all CESAs was examined during seed coat development. CESA3 and CESA10 showed increased expression in seed coats at 7 DPA, when mucilage biosynthesis is at its peak. Two mutant alleles of CESA3, ixr1-1 and ixr1-2, showed distinct changes in the cellulose structure of mucilage, while two mutant alleles of cesa10 showed no obvious mucilage defects. In ixr1-1 and ixr1-2 mucilage, changes in the mucilage structure and pectin distribution were also observed. Monosaccharide content and crystalline cellulose amounts were reduced in mucilage and whole seeds. Altogether, these results demonstrate a clear role for CESA3 in mucilage biosynthesis. 4.3.1 CESA expression during seed coat development Seed development includes many different aspects of cell wall biosynthesis and modification. Following fertilization, the seed undergoes distinct developmental pathways to produce a mature seed coat, endosperm and embryo (reviewed in Haughn and Chaudhury, 2005).  104 Seed coat epidermal cells undergo expansion during the early stages of seed development, followed by mucilage biosynthesis and columella biosynthesis. During this time, the embryo is also undergoing cell expansion, requiring the expression of cell wall biosynthetic genes. CESA3 and CESA10 both appeared to be highly expressed during seed development (Fig. 4.2). Their expression is increased specifically at 7 DPA in seed coats and reduced in expression from 7 DPA to 10 DPA (Fig. 4.2A) similar to the previously characterized MUM4 and CESA5 suggesting a specific involvement in mucilage biosynthesis. Other CESAs also appeared to be expressed in the seed coat, yet did not show the same increases in expression at 7 DPA. In whole seeds, CESA3 does not have as drastic a reduction in expression later in seed development (Fig. 4.2B). This increase in expression relative to the seed coat expression data could be due to CESA3 expression in the embryo, which is still undergoing cell expansion from 7 DPA to 10 DPA.  The localization of GFP::CESA3 and GFP::CESA5 in the cytoplasmic column of seed coat epidermal cells during mucilage biosynthesis (Fig. 4.3; Sullivan et al., 2011) demonstrates that both proteins are found in a time and place consistent with a role in mucilage cellulose synthesis. 4.3.2 The role of other CESAs during seed coat development A number of different mutant alleles were isolated in many different CESAs subunits, however, only cesa5 and cesa3 showed any changes in seed coat mucilage structure. It has previously been shown that CESA1 is not involved in mucilage biosynthesis, as the rsw1-1 mutant shows no change in mucilage at the non-restrictive temperature (Burn et al., 2002). Similarly, CESA2, CESA6 and CESA9 do not appear to have a role either (Mendu et al, 2011). Secondary cell wall CESAs have not been shown to be involved in mucilage biosynthesis;  105 CESA4, CESA7 and CESA8 show very low expression in the seed coat. CESA8 showed the highest expression of the three secondary cell wall CESAs during mucilage biosynthesis, yet the irx1-5 mutant showed no obvious seed mucilage phenotype (Fig.4.4H). Assuming that cellulose biosynthesis requires three different CESAs, another CESA besides CESA5 and CESA3 must be required for mucilage biosynthesis. CESA10 appears to be specifically expressed during seed development. Analysis of whole plant microarray data sets indicates that CESA10 expression is only detected in seeds. Previous studies have analyzed the expression of all 10 CESAs during embryo development, and found that CESA10 was not expressed in the embryo (Beeckman et al., 2002). Other studies have not found CESA10 expression in seedling development, although it did appear to be expressed in flowers, and up-regulated when treated with brassinosteroids (Xie et al., 2011). CESA10 is most closely related to CESA1, with over 86% amino acid similarity. It is possible that CESA10 is partially redundant to CESA1 during mucilage biosynthesis. This would be expected based on a CSC that requires 3 different CESAs for cellulose biosynthesis. This would also explain the lack of any adherent mucilage phenotypes for both rsw1-1 and cesa10-1. Isolation of a cesa10 cesa1 double mutant would provide more insight into the potential redundancy between these two genes. Additionally, the expression and localization of CESA10 during seed coat development remains to be determined. Construction of a GFP::CESA10 transgenic construct in still in progress. 4.3.3 Effects of ixr1-1 and ixr1-2 missense mutations on cellulose biosynthesis Quantitative analysis of total glucose amounts and crystalline cellulose amounts in seeds suggest reduced crystalline cellulose in ixr1-1 and ixr1-2 seeds. However, immunolabelling of ixr1-1 and ixr1-2 mucilage with CBM3a and CBM28 demonstrates increased labelling of  106 crystalline cellulose and reduced labelling of amorphous cellulose, potentially due to increased size of the adherent halo. Overall, ixr1-1 and ixr1-2 result in some drastic changes in the distribution of cell wall polysaccharides. These changes could result in modifications in the availability of the epitope recognized by CBM3a. Also, CBM3a labelling of ixr1-1 and ixr1-2 mucilage show a larger mucilage halo, indicating increased mucilage adherence, which could also lead to increased CBM3a labelling. Immunolabelling is useful in observing changes to the distribution of cell wall epitopes, but it is not quantitative. Changes in cellulose crystallinity could also result in altered interactions with other cell wall polymers, potentially resulting in increased epitope availability for CBM3a. The main conclusion from theses experiments is that CESA3 appears to be involved in mucilage biosynthesis. Considering all the data, the most plausible explanation of the effects of ixr1-1 and ixr1-2 on seed coat mucilage is that they result in reduced crystalline cellulose amounts in mucilage, which result in altered adherence properties and cell wall polysaccharide distribution. The ixr1-1 and ixr1-2 missense mutations are located near the C-terminus of CESA3; ixr1-1 lies in a transmembrane domain while ixr1-2 is located between two transmembrane domains, in an extracellular loop (Fig. 4.1). ixr1-2 was shown to increase the speed of CSCs in the plasma membrane, which resulted in reduced cellulose crystallinity (Harris et al., 2012). In total, CESAs contain 8 transmembrane helices that form a pore in the plasma membrane through which the newly synthesized cellulose passes through, into the apoplast (Carpita, 2011). It has been suggested that the mutated residue in ixr1-2, located within a transmembrane domain, is required for the correct formation of the pore, which is required for correct glucan chain alignment into a crystalline microfibril (Harris et al., 2012). The quantitative data on ixr1-1 and ixr1-2 seeds is consistent with a reduction in crystalline cellulose. Both ixr1 point mutations are  107 located either in the transmembrane domain, or in an extracellular loop, which can potentially affect the formation of the pore, and the polymerization and crystallization of cellulose. Unfortunately null alleles of CESA3 are gametophytic lethal. Other alleles of cesa3 might provide more information on how cellulose is synthesized during mucilage biosynthesis. Given that the seed coat is not required for plant viability under laboratory conditions, it would be interesting to examine the effects of a CESA3 silenced line using a seed coat specific promoter to drive the expression of a RNA interference (RNAi) construct directed against CESA3. Similarly, an RNAi construct could be designed to reduce the expression of all 10 CESAs in the seed coat. Such constructs would be useful not only for understanding the roles of CESA in mucilage, and in the columella. Additionally, this research indicates that mucilage can be used as a system to understand how cellulose properties affect pectin distribution. CESA3 and CESA5 could be used for targeted mutagenesis in key regions of the CESA protein to identify changes in cellulose crystallinity and the subsequent effects on pectin distribution. 4.3.4 ixr1-1 and ixr1-2 mutations alter mucilage structure CESA3 is clearly involved in mucilage cellulose biosynthesis, but understanding the effects of the ixr1-1 and ixr1-2 missense mutations on mucilage structure is difficult to interpret. While the cesa5-1 seeds show a loss of adherence due to a reduction in cellulose biosynthesis, ixr1-1 and ixr1-2 show clear changes in mucilage structure. Overall we can observe that compared to Col-0 mucilage, ixr1-1 and ixr1-2 mucilage has altered cell wall polymer distribution (Fig. 4.8 - 4.11) resulting in altered ruthenium red staining, (Fig. 4.4, 4.5), increased density and length of the rays (Fig. 4.5, 4.6, and 4.7), and altered hydration and gelling (Fig. 4.12). Presumably, all of these changes result from altered cellulose crystallinity in ixr1-1 and ixr1-2 seed mucilage. How mutations in CESA3 result in changed pectin distribution and/or  108 biosynthesis is not clear. Multiple lines of evidence suggest a close association between pectins and cellulose (Iwai et al., 2001; Oechslin et al., 2003; Vignon et al., 2004; Zykwinska et al., 2005). Recently, changes in the degree of pectin esterification have been shown to stabilize the cellulosic network (Chebli et al., 2012). Changes in cellulose crystallinity could influence the interactions between pectins and cellulose, resulting in altered pectin distribution in ixr1 seed mucilage. The phenotype of ixr1-1 and ixr1-2 seed mucilage appears similar to that of cesa5-1, however, some clear differences are observable. Ruthenium red staining is reduced in ixr1-1 and ixr1-2 seeds (Fig. 4.4C and D). Reduced staining in ixr1 seeds could result from loss of mucilage adherence similar to cesa5-1 seeds, or altered ability of ruthenium red to bind to pectins, due to changes in the mucilage cell wall network. Multiple lines of evidence support the postulate that mucilage adherence is increased in ixr1-1 and ixr1-2 seeds. The length of rays as determined by S4B staining, and in mucilage halo width is slightly larger as determined by CBM3a staining in ixr1-1 and ixr1-2 seeds (Fig. 4.7 – 4.11). EDTA completely rescues the altered mucilage staining phenotype of ixr1-1 and ixr1-2 seeds, implying that the structure of mucilage and its ability to form a cohesive network is increased, not reduced (Fig 4.12). Finally, non-adherent mucilage amounts are reduced in ixr1-1 and ixr1-2 seeds (Fig. 4.13). However, whole seed cell wall monosaccharide amounts are also reduced, indicating reduced cell wall biosynthesis, which make interpretation of the quantitative monosaccharide data difficult. It is not clear how ixr1-1 or ixr1-2 could result in reduced cell wall biosynthesis, but it is possible that the changes in the cell wall cohesiveness caused by the ixr1 point mutations is compensated for by reduced cell wall biosynthesis. Previously, ixr1-2 was also shown to have reduced cell wall monosaccharide content in stems (Harris et al., 2012). CESA3 is also involved  109 in cell wall biosynthesis in the embryo. Changes in embryo cell wall biosynthesis would also be detected in whole seed monosaccharide quantification. When considering the qualitative labelling and staining of mucilage with the quantitative monosaccharide data, the simplest interpretation is that ixr1-1 and ixr1-2 mutations result in reduced cell wall biosynthesis in addition to increased mucilage. In summary, this research demonstrates that CESA3 is involved in mucilage biosynthesis. The affects of the ixr1-1 and ixr1-2 mutations on the function of CESA3 do not appear to result in complete loss of function, and rather appear to modify cellulose biosynthesis in mucilage. These mutations results in reduced cellulose crystallinity that affect the distribution of pectins in the adherent mucilage halo. Mucilage adherence is increased in ixr1 seeds as a result from increased interactions between cellulose and pectins. This research provides a good background for understanding how cellulose crystallinity affects cell wall plasticity and interactions between cell wall polymers.  110 5 CESA5 and SOS5 function independently to mediate mucilage adherence. 5.1 Introduction  The identification of CESA5 as being required for mucilage adherence provided the first genetic evidence for the presence and function of cellulose in seed coat mucilage. A simple hypothesis for the role of CESA5 in mucilage adherence is that it synthesizes the cellulose in the mucilage pockets, which interacts with pectin to form an adherent mucilage layer. Loss of CESA5 function reduces cellulose biosynthesis and results in a less cohesive mucilage matrix (Mendu et al., 2011; Sullivan et al., 2011). Interestingly, salt overly sensitive 5 (sos5)/fasciclin- like arabinogalactan-protein 4 (fla4) mutants cause mucilage defects that resemble those of cesa5 seeds (Shi et al., 2003; Harpaz-Saad et al., 2011). Although its exact function is not clear, SOS5 is hypothesized to mediate mucilage adherence through interactions with CESA5 (Harpaz- Saad et al., 2011). In order to more clearly understand the role of SOS5 in mucilage adherence, the phenotype of cesa5-1 and sos5-2 were examined in great detail, and a double mutant was constructed to analyze their genetic interactions. 5.1.1 Arabinogalactan proteins Arabinogalactan proteins (AGPs) are a small family of evolutionarily conserved, secreted proteins that are highly glycosylated with Type II arabinogalactans, and can be localized to the cell surface by a C-terminal glycophosphatidylinositol (GPI) lipid anchor at the plasma membrane (Fig. 5.1A; reviewed in Showalter et al., 2001; Seifert and Roberts, 2007; Ellis et al., 2010). AGPs can be extensively modified in the cell wall, and many glycosyl hydrolases have been shown to act on AGP glycosyl side chains (Sekimitata et al., 1989; Cheung et al., 1995; Wu et al., 1995; Kotake et al., 2005). In addition, AGPs can be isolated in soluble cell wall fractions,  111 indicating that their GPI anchor can be cleaved (Samson et al., 1984). Although AGP proteins may have a variety of functions, from intracellular signaling to cell wall organization (reviewed in Seifert and Roberts, 2007; Ellis et al., 2010), their exact role in the cell wall is poorly understood. FLAs are a subclass of AGPs, distinguished by the presence of fasciclin-like (FAS) domains (Schultz et al., 2000; Gaspar et al., 2001; Johnson et al., 2003), which are known to facilitate cell adhesion in animals, algae and microbes (Kawamoto et al., 1998). There are 21 genes in the Arabidopsis genome that contain FAS domains (Johnson et al., 2003). SOS5 contains two highly glycosylated AGP domains and two FAS domains that have low sequence similarity to other known FLA proteins (Fig. 5.1; Johnson et al., 2003). SOS5 was initially identified in a screen for salt sensitivity in roots (Shi et al., 2003). The sos5 roots display increased salt sensitivity coupled with an overall loss of cell wall organization. This evidence led to the proposal that SOS5 functions as a cell adhesion molecule, and suggests that SOS5 could interact with other cell wall components to form a network (Shi et al., 2003). The recent discovery that sos5 mutants also display a loss of seed mucilage adherence has led to the hypothesis that SOS5 functions to modify cellulose biosynthesis, through interactions with CESA5 (Harpaz-Saad et al., 2011). However, aside from the similar mucilage phenotypes of sos5 and cesa5 seeds, there is little evidence to support this hypothesis. It is not clear how SOS5 mediates mucilage adherence and if SOS5 can interact with CESA5. In this chapter, I conduct a detailed investigation of the roles of CESA5 and SOS5 in seed coat epidermal cells.  112  Figure 5.1: Diagram of the protein structure of SOS5/FLA4 and a phylogenetic tree illustrating the relation of all AGP proteins in the FLA family. A) Diagram of a pre- processed and mature SOS5 protein structure illustrating the N-terminal signal secretion sequence, two Fasciclin-like domains, two AGP domains that are the sites of glycosylation and the GPI anchor signal (Shi et al., 2003).  B) Phylogeny of FLA AGPs including FLA4/SOS5, From MacMillan et al., 2010. Copyright John Wiley & Sons, reprinted with permission. !! "!  113 5.2 Results 5.2.1 sos5 and cesa5 have distinct and additive effects on seed mucilage adherence  Both cesa5-1 and sos5-2 seeds show a similar loss of mucilage adhesion following seed hydration. In order to better understand the functions of CESA5 and SOS5, cesa5-1 sos5-2 double mutants were isolated from an F2 population of a cross between the two single mutants. The alleles used in all studies found here are derived from the cesa5-1 and sos5-2 mutant lines; often the specific allele is omitted for simplicity. Plants homozygous for both mutations were identified using a PCR-based assay to detect T-DNA insertions. Three double mutant individuals were isolated from an F2 population of 72 plants, consistent with two independently segregating recessive mutant alleles. Two of these double mutants, cesa5 sos5 1.1 and cesa5 sos5 3.7, were selected for further analysis. These double mutant lines displayed similar mucilage phenotypes, indicating that the observed defects are not linked with other background mutations. When relevant, both independently isolated double mutant phenotypes are shown, however the lines were largely identical in all aspects of the phenotype, so often only one line is shown. When cesa5-1 and sos5-2 seeds are hydrated directly in ruthenium red, mucilage extrudes rapidly and is clearly present in all mutant seeds similar to Col-0 seeds (Fig. 5.2A-C). However, cesa5 sos5 double mutant seeds do not release mucilage as quickly, or as uniformly as Col-0, cesa5-1 and sos5-2 seeds (Fig. 5.2D, E). This indicates that mucilage hydration is not as efficient in cesa5 sos5 seeds.  After 1 h of shaking in water, no ruthenium red stained mucilage remains attached to cesa5-1 and cesa5 sos5 seeds (Fig. 5.2G, H). Some residual staining is still observable in the sos5-2 single mutant (Fig. 5.2M). When seeds were shaken in water for 4 hours, sos5-2 seeds  114 lose all observable ruthenium red stained mucilage (not shown). Double mutant seeds have two novel phenotypes: the outer primary cell wall does not appear to rupture normally and can either remain intact on the parent seed or released as a “sheet” of primary cell wall material (see arrow Fig. 5.2N), and some cells have smaller or absent columella (Fig. 5.2O). Mucilage staining phenotypes for cesa5 sos5 seeds were variable in different regions of a seed and between different seeds of the same genotype. The differences between cesa5 and sos5 seeds are most clear when seeds are stained with pontamine S4B (S4B), a fluorescent dye that binds preferentially to cellulose (Fig. 5.2P-T; Anderson et al., 2011). Wild-type seeds have a strong S4B staining halo, and distinct regions of staining in the mucilage halo can be observed (Fig. 5.2P). The primary wall remnants, which remain attached to the columella, stain very brightly (see arrow, Fig 5.2P). Above each columella, cellulosic ‘rays’ are clearly observable, extending to the periphery of the mucilage halo. S4B also stains a less-intense (diffuse) staining region between the rays (Mendu et al., 2011). cesa5-1 seeds lack the diffuse S4B staining region, yet retain the rays that are located above the columella (see arrow, Fig. 5.2Q). Rays appear slightly reduced compared to Col-0 seeds (compare Fig. 5.2 P and Q). Unlike cesa5-1 seeds, sos5-2 seeds lack distinct rays but instead retain the diffuse staining pattern (Fig. 5.2R). These data indicate differences in the effect of cesa5-1 and sos5-2 on the pattern of cellulose distribution in mucilage. S4B-stained cesa5 sos5 seeds appear to have more severe defects than either single mutant (Fig. 5.2S, T). Double mutant seeds display no S4B-labelled rays or diffuse mucilage, and have abnormal primary wall attachment. Although some cesa5 sos5 seed coat epidermal cells have primary walls attached to the top of columellae (Fig. 5.2T), the outer tangential primary walls that are often detached from multiple columellae as a large sheet (Fig. 5.2S).  115  Figure 5.2: Mucilage phenotype of cesa5-1, sos5-2 and cesa5 sos5 double mutants. A-E) Seeds stained with ruthenium red, unshaken. F-J) Seeds shaken in water and stained with ruthenium red. K-O) Higher magnification of ruthenium red stained seeds, showing the mucilage halo and the columella. P-T). S4B stained seeds illustrating the cellulosic ray above the columella and the diffuse staining mucilage. U-Y) Seeds shaken in 0.05M EDTA and then stained with ruthenium red. Col-0 A, F, K, P, U; cesa5-1 B, G, L, Q, V; sos5-2 C, H, M, R, W; cesa5 sos5 1.1 D, I, N, S, X; cesa5 sos5 3.7 E, J, O T, Y. A-J, U-Y Bar = 200 !m. K-T Bar = 10 !m. Col-0! A! B! C! D! E! F! G! H! I! J! K! L! M! N! O! P! Q! R! S! T! U! V! W! X! Y! cesa5-1! sos5-2! cesa5 sos5 1.1! cesa5 sos5 3.7!  116 Columellae also appear to be smaller when visualized with S4B staining, and fewer columellae were observed in the double mutant compared to Col-0, cesa5-1, or sos5-2 single mutant seeds. Treatment of wild type seeds with EDTA, a divalent cation chelator, results in a larger mucilage halo that shows reduced ruthenium red staining compared to the water hydration (Fig.5.2U). Unlike the loss of mucilage adherence observed in water, EDTA-treated cesa5 and sos5 seeds have surprisingly large ruthenium red-stained mucilage halos (Fig. 5.2V, W). However, there is still a reduction in the intensity of staining and size of the adherent mucilage layer compared to wild type. The cesa5 sos5 double mutant seeds show a more severe defect when compared to either single mutant, suggesting that CESA5 and SOS5 function independently to mediate adherence (Compare Fig. 5.2X and Y to Fig. 5.2V and W). Despite the EDTA treatment, cesa5 sos5 seeds show very little mucilage extrusion and ruthenium red staining (Fig. 5.2X, Y). 5.2.2 Mucilage structure is altered in cesa5-1 and sos5-2 seeds In an attempt to further characterize the distribution of cellulose and pectins in adherent mucilage, water-hydrated seeds were immunolabeled with two 6x histidine tagged carbohydrate binding module (CBM) proteins and two pectin specific antibodies (JIM5, CCRC-M36; VandenBosch et al., 1989; Blake et al., 2006; Pattathil et al., 2010). CBM3a is highly specific for crystalline cellulose structures and CBM28 preferentially binds amorphous cellulose epitopes (Blake et al., 2006; Dagel et al., 2011). JIM5 is a Rat monoclonal antibody that is specific for partially methyl-esterified HG (Knox et al., 1990; Willats et al., 2000). CCRC-M36 is a mouse monoclonal antibody specific for unsubstituted RG-1 (Pattathil et al., 2010) that strongly labels Arabidopsis seed mucilage (Young et al., 2008). Many other antibodies specific to arabinans  117 (LM6), galactans (LM5), and arabinogalactan proteins (MAC207) did not show any binding in wild type or mutant mucilage (Data not shown). Seeds were initially hydrated in water, immunolabeled, and finally stained with S4B. In wild-type adherent mucilage, there are obvious differences in the position of epitopes bound by S4B and CBM3a (Fig. 5.3). As previously described, S4B stains the adherent layer, highlighting the cellulosic rays above columellae and some mucilage emanating from the mucilage pocket (Fig. 5.3E; Mendu et al., 2011). CBM3a strongly labeled the entire adherent mucilage halo with increased labelling along the rays (Fig. 5.3A). S4B and CBM3a labels appeared to overlap in the middle of the mucilage halo, but their primary labelling pattern is clearly different. This labelling pattern indicates that while both CBM3a and S4B are specific to cellulose, S4B binds a distinct subset of cellulose epitopes (Fig. 5.3I). In cesa5-1 mutant seeds, a strong reduction in S4B staining is observed, similar to previous reports (Fig. 5.2Q; Mendu et al., 2011). As in wild type, CBM3a labels more broadly within the mucilage halo relative to S4B (compare Fig. 5.3B with Fig. 5.3F). The overall radius of the mucilage halo labeled by CBM3a is slightly reduced in cesa5-1 seeds compared to wild type, consistent with a general reduction in cellulose biosynthesis and loss of mucilage adherence. CBM3a intensely labelled the top of the rays, yet the staining pattern is less intense between the rays in cesa5-1 mucilage, similar to S4B staining (compare Fig. 5.3A with Fig. 5.3B). CBM3a labelling of sos5-2 seeds appeared more similar to wild type than to cesa5-1, however, compared to wild-type seeds, sos5-2 seeds had reduced structure and organization of the mucilage and lacked any observable ray structure (Fig. 5.3C). This CBM3a labelling pattern is clearly different from cesa5-1 seeds (compare Fig. 5.3J with 5.3K). The cesa5 sos5 mutant  118  Figure 5.3:!CBM3a labelling of adherent mucilage. Seed mucilage structure of crystalline cellulose in the adherent layer of Col-0 (A, E, and I), cesa5-1 (B, F, and J), sos5-2 (C, G and K) and cesa5 sos5 (D, H and L) seeds. Crystalline cellulose was immunolabeled with CBM3a (Green, A-D), and then stained with S4B (Magenta, E-H). Green and Magenta images were then merged (I-J). Bar =10 µm. !"#$% ! &'"! #()*(! +! !!"! ,! -! .!/! 0! 1! 2!3! 4! !5678! !"#$%&'! #(#%&)! !"#$%*#(#%!  119 seeds show an almost complete loss of labelling for both CBM3a and S4B (Fig. 5.3D and H, respectively). Some areas of seed mucilage show no CBM3a labelling at all, while other areas show small-disorganized tufts of labelling. There is also much stronger labelling of the primary cell wall, likely due to a reduced amount of CBM3a epitopes in cesa5 sos5 mucilage.  CBM28 showed very little mucilage labelling compared to CBM3a. Laser intensities and sensitivity settings for CBM28 were significantly increased compared to CBM3a settings. Wild- type seeds displayed diffuse, punctate CBM28 labelling in the mucilage halo (Fig. 5.4A), and in the ray regions (see arrow, Fig. 5.4I). CBM28 labelling of cesa5-1 seed mucilage appeared reduced and more disorganized than wild type (compare Fig. 5.4A to Fig. 5.4B), however, there was still some labelling of the ray. Both sos5-2 and cesa5 sos5 seeds lacked almost all CBM28 signal in adherent mucilage (Fig. 5.4C and D). cesa5 sos5 seeds did have some faint CBM28 labelling immediately above the epidermal cells (See arrow, Fig. 5.4D).  JIM5 was also used to analyze the distribution of HG with a low degree of methylesterification in mucilage (Fig. 5.5). In Col-0 seeds, JIM5 labels the rays and the adherent mucilage layer (See arrow, Fig. 5.5I). JIM5 only labels the base of the rays directly above the columella in cesa5-1 mucilage compared to Col-0 and diffuse signal is absent in the adherent mucilage layer (compare Fig. 5.5A to Fig. 5.5B). sos5-2 JIM5 labelling is further reduced compared to cesa5-1 seeds, and is only labelled at the primary walls that remain attached to columellae (Fig. 5.5C). Finally, the cesa5 sos5 double mutant had no visible staining in the mucilage halo of either S4B or JIM5 (Fig. 5.5L). JIM5 appears to more strongly label the remnants of the primary cell wall more intensely than sos5-2 seeds, most likely because of increased accessibility of antibody to the primary cell wall.   Finally, CCRC-M36 was used to probe the distribution of unsubstituted RG I in adherent    120  Figure 5.4:!CBM28 labelling of adherent mucilage. Seed mucilage structure of amorphous cellulose in the adherent layer of Col-0 (A, E, and I), cesa5-1 (B, F, and J), sos5-2 (C, G and K) and cesa5 sos5 (D, H and L) seeds. Amorphous cellulose was immunolabeled with CBM28 (Green, A-D), and then stained with S4B (Magenta, E-H). Green and Magenta images were then merged (I-L). Bar =10 µm. !"#$% ! &'"! #()*(! !+,-.! !"#$%&'! #(#%&)! !"#$%*#(#%!/! !!"! 0! 1! 2!3! 4! 5! 6!7! 8!    121  Figure 5.5:!JIM5 labelling of adherent mucilage. Seed mucilage structure of homogalacturonan in the adherent layer of Col-0 (A, E, and I), cesa5-1 (B, F, and J), sos5-2 (C, G and K) and cesa5 sos5 (D, H and L) seeds. Homogalacturonan was immunolabeled with JIM5 (Green, A-D), and then stained with S4B (Magenta, E-H). Green and Magenta images were then merged (I-L). Bar =10 µm.  . !"#$! %&'! #()*(! +,-./! !"#$%&'! #(#%&)! !"#$%*#(#%!0! +!'! 1! 2! 3!4! 5! "! 6!!! 7!    122 Figure 5.6:!CCRC-M36 labelling of adherent mucilage. Seed mucilage structure of rhamnogalacturonan I in the adherent layer of Col-0 (A, E, and I), cesa5-1 (B, F, and J), sos5-2 (C, G and K) and cesa5 sos5 (D, H and L) seeds. Unsubstituted RG I was immunolabeled with CCRC-M36 (Green, A-D), and then stained with S4B (Magenta, E-H). Green and Magenta images were then merged (I-L). Bar =10 µm. !!"!#$ %&! '()! $*+,*! -! !!)! .! /! 0!1! 2! 3! 4!5! 6! !78#9! !"#$%&'! #(#%&)! !"#$%*#(#%!  123 mucilage (Fig 5.6). Col-0 adherent mucilage is strongly labelled, with some increased labelling of the ray structures at the periphery of the mucilage halo (see arrow, Fig. 5.6A). cesa5-1 adherent mucilage shows a unique labelling pattern, with CCRC-M36 signal concentrated around the rays labelled by S4B (Fig. 5.6J). Besides the intense labelling of rays, cesa5-1 seeds show a general reduction of diffuse CCRC-M36 labelling in adherent mucilage compared to Col-0 (Compare Fig. 5.6A to Fig. 5.6B). sos5-2 seeds have reduced CCRC-M36 labelling compared to Col-0 seeds (compare Fig. 5.6A to Fig. 5.6C), and sos5-2 seeds show a complete loss of ray labelling by CCRC-M36, which is in strong contrast to cesa5-1 seeds (compare Fig 5.6B. to Fig. 5.6C). However, sos5-2 seeds do show some diffuse labelling (Fig. 5.6C). Adherent mucilage in cesa5 sos5 seeds failed to show any labelling by CCRC-M36 (Fig. 5.6D). Immunolabelling illustrates key differences between cesa5-1 and sos5-2 adherent mucilage phenotypes, and demonstrates the complete lack of all adherent mucilage in cesa5 sos5 seeds. 5.2.3 CESA5 and SOS5 are both required for normal columella deposition  While visualizing the mucilage phenotype of S4B stained seeds, I observed that columella appeared smaller and were often completely absent in the cesa5 sos5 double mutant. Therefore, I examined the surface of seed epidermal cells using Scanning Electron Microscopy (SEM) of gold-palladium coated seeds. Col-0 seed coat epidermal cells appear hexagonal in shape and are separated by the radial wall, which appears as an elevated ridge between all cells (Fig. 5.7). The columella is present in the center of each cell, and is also elevated relative to the mucilage pocket that surrounds the columella. Mutant cesa5-1 and sos5-2 seeds appear similar to wild-type seeds. Many cesa5 sos5  124 Figure 5.7: Columella formation is affected in cesa5 sos5 double mutants. A) SEM of the seed coat surface of cesa5-1, sos5-2 and cesa5 sos5 double mutants. N= 80, Bar = 25 µm. B) Quantification of columella area. One asterisk: significantly different from wild type, P < 0.05, two asterisks significance: P < 0.001 based on Student’s t-test. C) Cell surface area of Col-0, cesa5-1, sos5-2 and cesa5 sos5 seeds showed no significant change in ell area, n= 20. Error bars are standard error measurements. Col-0  ! A! 120! 100! 80! 60! 40! 20! 0! C ol um el la  a re a (µ m 2 ) !B! C! 780! 740! 700! 660! 620! C el l a re a (µ m 2 ) ! cesa5-1  ! sos5-2! cesa5 sos5! Col-0  ! cesa5-1  ! sos5-2! cesa5 sos5! Col-0  ! cesa5-1  ! sos5-2! cesa5 sos5!  125 epidermal cells lack observable columella, and when columella are present, they appear much smaller (Fig. 5.7A). To quantify this phenotype, I measured the surface area of 80 columellae, from at least 6 different seeds for each genotype (Fig 5.7B). cesa5-1 seeds showed a minor reduction in columella area (~10%), consistent with previous results (Mendu et al., 2011). Mutant sos5-2 seed showed a more severe decrease in columella area (~30%). cesa5 sos5 seeds showed a strong reduction in columella area (~70%). All three mutant lines were significantly different from wild type based on Students T-test (Col-0, 94.26 ± 4.20 µm2; cesa5-1, 82.63 ± 3.37 µm2, P < 0.05; sos5-2, 63.90 ± 2.57 µm2, P < 0.001; cesa5 sos5, 25.60 ± 1.62 µm2, P < 0.001). This result demonstrates that both genes are involved in columella biosynthesis. Since the columella is deposited once mucilage biosynthesis has ceased, reduced columella area could result from a smaller cytoplasmic column, reduced secondary cell wall deposition, and/or reduced cell size. Therefore, I also measured total cell area for all mutant lines (Fig. 5.7C). While there was a slight reduction in total cell area for cesa5-1, sos5-2, and cesa5 sos5 double mutant seeds, it was not significantly different from wild- type cell areas based on Student’s t-test (P > 0.05). Since both single mutants show a reduction in the size of the top of the columella, and the double mutant phenotype is more severe than either single, both CESA5 and SOS5 may contribute independently to controlling columella size. 5.2.4 SOS5 influences columella morphology through the cytoplasmic column Two hypotheses could explain the defects in columella morphology in mature cesa5 sos5 seeds. First, the shape of the mucilage pocket could be altered, which would  126 lead to deformed columella. The second hypothesis would involve defects in the cell wall biosynthesis of the columella. While CESA5 has a known role in both mucilage biosynthesis and secondary cell wall biosynthesis (Mendu et al., 2011), it was not clear how the columella deposition was affected by the loss of SOS5. Therefore, I examined developing seeds to investigate the relation between the mucilage pocket and the formation of the columella. I chose to examine seed coat epidermal cell development at 8, 10 and 12 DPA in order to determine when the defects in columella deposition first occur. I conducted live cell imaging with propidium iodide (PI, Fig. 5.8) and Toluidine blue staining of cryo-fixed seeds (8 and 10 DPA), and glutaraldehyde-fixed mature seeds (Fig. 5.9). Col-0 seeds stained with PI at 8 DPA, when mucilage biosynthesis is nearly complete, display cytoplasmic columns filled with amyloplasts (Fig. 5.8A). Amyloplasts appear to be labeled by PI, which is excluded from the surrounding membranes. Some staining is also observed in the radial wall, and although PI is reported to bind pectin, no mucilage labelling was observed (Fig 5.8A; Rounds et al., 2011). sos5-2 and cesa5-1 single mutants show normal columella development (Fig. 5.8B and C), while the cesa5 sos5 double mutant shows a smaller cytoplasmic column, and amyloplasts appear more frequently in the basal regions of the cell (Fig. 5.8D). By 10 DPA, mucilage biosynthesis is complete and columella deposition is well underway in wild-type cells. In Col-0 epidermal cells, a thick secondary wall displaces the cytoplasmic column (Fig. 5.8E). cesa5-1 columellae frequently appear delayed, with gaps in the center of the columella, and some amyloplasts were still observable similar to previous reports (Fig. 5.8F; Mendu et al., 2011). Columella of sos5-2 seed coat epidermal cells appears similar to Col-0 seeds  127  Figure 5.8: Live cell imaging of columella development in cesa5-1, sos5-2 and cesa5 sos5 double mutant seed coat epidermal cells. DIC images of developing seed coat epidermal cells overlayed with green fluorescence derived from propidium iodide staining. A-D, 8 DPA; E-H, 10 DPA, I-L, 12 DPA. Bar = 10 µm.   Col-0! 12  D PA ! 10  D PA ! 8 D PA ! cesa5-1! sos5-2! cesa5 sos5!!! "!#! $! %! &!'! (! )! *!+! ,!  128  Figure 5.9: Columella development as determined by fixed, embedded seeds sections. Columella development in Cryo-fixed (A- H) and glutaraldehyde fixed (I-J) resin embedded seed coat epidermal cells stained with toluidine blue. A-D, 8 DPA; E-H, 10 DPA, I- L, Mature seeds. Bar = 10 µm. M at ur e! 10  D PA ! 8 D PA ! Col-0! cesa5-1! sos5-2! cesa5 sos5! !! "!#! $! %! &!'! (! )! *!+! ,!  129 (Fig. 5.8G). Columella deposition is largely complete, and appears more advanced than cesa5-1 seeds (Compare Fig. 5.8F with 5.8G). Double cesa5 sos5 mutants show much smaller columella, and some columella appeared absent (Fig. 5.8H). However, compared to cesa5-1 seeds, no delay in columella deposition was observed. By 12 DPA columella deposition is complete in wild type, cesa5-1 and sos5-2 epidermal cells (Fig. 5.8I-K). The cells of cesa5 sos5 mutants often lack columella entirely, while other columellae are noticeably smaller (Fig. 5.8L). Some columellae appear stunted, narrower or absent compared to Col-0 epidermal cells. Similar results were observed with cryo-fixed, toluidine blue stained developing seeds (Fig. 5.9). At 8 DPA, Col-0, cesa5-1 and sos5-2 seed coat epidermal cells have produced a large mucilage pocket that stains light purple with toluidine blue, with a central cytoplasmic column filled with amyloplasts (Fig. 5.9A, B and C, respectively). A large vacuole is present in the basal portion of each epidermal cell. cesa5 sos5 double mutant seeds show normal mucilage pockets with a clear cytoplasmic column, however, the column is often much narrower than Col-0 seeds or either single mutant (See arrow, Fig. 5.9D). By 10 DPA, wild-type Col-0 seeds have nearly completed columella deposition (Fig. 5.9E). While mucilage stains a light purple with toluidine blue, the columella stains a darker purple. Both cesa5-1 and sos5-2 seeds also show substantial secondary wall deposition in the cytoplasmic column, forming a large columella (Fig. 5.9F and G, respectively). In contrast, in cesa5 sos5 seeds, some columella are either narrower or entirely absent at 8 DPA (See arrow, Fig 5.9H). I next examined mature whole seeds fixed in glutaraldehyde, resin embedded and stained with toluidine blue (Fig. 5.9I-L). In mature Col-0 seed epidermal cells, the mucilage pocket has burst; all that remains is the secondary cell wall including the columella and the radial walls (Fig. 5.9I). The primary cell wall can be observed attached to the top of the columella. Epidermal  130 cells of cesa5-1 and sos5-2 seeds also appear normal, except for a reduction in radial wall height for cesa5-1 lines, similar to previous reports (Fig. 5.9J and K, respectively; Mendu et al., 2011; Harpaz-Saad et al., 2011). In cesa5 sos5 double mutant seed coat epidermal cells, many cells lack columella entirely (Fig. 5.9L). Some cells do have columella, and large amounts of secondary wall deposited along the basal portion of the cell. Surprisingly, despite the clear presence of radial walls in SEM images of mature cesa5 sos5 seeds, they appear to be completely absent from fixed and embedded seed coat sections. In summary, analysis of developing and mature seeds indicates that cesa5 sos5 seed coat epidermal cells have reduced or absent columella that result from abnormal morphology of the cytoplasmic column, not due to reduced secondary wall biosynthesis. 5.2.5 SOS5 does not affect total glucose amounts in seeds I quantified whole seed monosaccharide amounts to investigate if cesa5 sos5 has altered cell wall biosynthesis, and non-adherent mucilage amounts to quantify the adherence defect (Fig. 5.10A). First, I examined the amounts of rhamnose and galacturonic acid, as these two sugars comprise ~90% of total mucilage content and are representative of total mucilage adherence. Both rhamnose and galacturonic acid in non-adherent mucilage isolated from cesa5-1 and sos5-2 seeds showed significant increases of roughly 20% compared to wild-type seeds, similar to previously published reports (determined by two way ANOVA, with the two factors being the presence or absence of cesa5-1 and sos5-2; Col-0, rhamnose 45.85 ± 0.51 ng sugar/mg seed, galacturonic acid 48.64 ± 0.64 ng sugar/mg seed; cesa5-1, rhamnose 56.29 ± 0.57 ng sugar/mg seed; ANOVA, cesa5-1, P < 0.01, galacturonic acid 58.19 ± 0.24 ng sugar/mg seed, P < 0.01; sos5-2, rhamnose 56.68 ± 0.73 ng sugar/mg seed, ANOVA, sos5-2, P < 0.01, galacturonic acid 56.09 ± 1.11 ng  131 Figure 5.10: Monosaccharide composition of non-adherent mucilage and whole seed AIR. A) Non-adherent mucilage fraction B) Whole seeds AIR. Letters indicates different levels of values based on a two-way ANOVA with Tukey’s post-hoc analysis. Genotypes connected by the same letter show no significant difference, while those with different letters are significantly different, P < 0.05. Error bars indicate standard error, n=4. !" #!" $!" %!" &!" '!" (!" )!" *+,-./01" 2,3,4567/.84"948:" !" #" $" %" &" '" (" )" 2,3,45/01" 2364/01" ;<3/01" =/3>!" !"#$%&'( #)#%&*( !"#$%(#)#%('+'( !"#$%(#)#%(,+-(  N on -a dh er en t m uc ila ge  fr ac tio n (! g su ga r/m g se ed )! !" '!" #!!" #'!" $!!" $'!" %!!" %'!" &!!" *+,-./01" 2,3,4567/.84" ,48:" 97,?8./01" 2,3,45/01" 2364/01" ;<3/01" =/3>!" !"#$%&'( #)#%&'( !"#$%(#)#%('+'( !"#$%(#)#%(,+-( W ho le  se ed  A IR  m on os ac ch rid e co nt en t ( !g  su ga r/m g se ed )! A! B! b a b b b b b b b b a b b b a b a c ab a ab a c b bc b a b a ab a ab a b b a ab b b b  132 sugar/mg seed, ANOVA, sos5-2, P < 0.01; Mendu et al., 2011; Sullivan et al., 2011). cesa5 sos5 double mutant seeds also showed significant increases compared to Col-0 seeds in the amount of non-adherent rhamnose and galacturonic acid based on a two-way ANOVA (cesa5 sos5 1.1, rhamnose 57.03 ± 0.56 ng sugar/mg seed , P < 0.01, galacturonic acid, 56.26 ± 0.69 ng sugar/mg seed , P < 0.01; cesa5 sos5 3.7, rhamnose 55.13 ± 2.68, P < 0.05, galacturonic acid 60.85 ± 0.40 ng sugar/mg seed , P < 0.01). Amounts of rhamnose and galacturonic acid were not different between the double mutant and either single mutant (Tukey’s HSD, P > 0.05). Despite reduced mucilage staining compared to cesa5-1 and sos5-2 seeds, the cesa5 sos5 double mutant showed no significant change in the amount mucilage extracted compared to cesa5-1 and sos5-2 seeds single mutant. This indicates that, both CESA5 and SOS5 are required independently for adherence. Next, I examined other cell wall sugars present in mucilage, in order to try to understand the mechanisms mediating adherence. Mucilage is composed primarily of rhamnose and galacturonic acid found in the backbone of RG I. Other sugars found as minor components of mucilage include galactose, glucose and xylose. Galactose residues are found in arabinogalactan side chains that are likely present on SOS5 and also on side chains of RG I. Glucose may be derived from cellulose or hemicellulose found in mucilage. Xylose residues could be derived from hemicellulose. Due to sensitivity issues, arabinose was below the detection limit, and not quantified. Amounts of galactose were higher in sos5-2 compared to Col-0 seed mucilage (Col-0, 1.47 ± 0.10 ng sugar/mg seed; sos5-2, 2.09 ± 0.16 ng sugar/mg seed), while cesa5-1 and cesa5 sos5 seeds had similar amounts of released galactose to Col-0 seeds (cesa5-1, 1.53 ± 0.40 ng sugar/mg seed;  133 cesa5 sos5 1.1 1.41 ± 0.07 ng sugar/mg seed; cesa5 sos5 3.7 2.45 ± 0.82 ng sugar/mg seed). Glucose amounts were highly variable in non-adherent mucilage, with cesa5-1 seeds showing a significant reduction compared to Col-0 seeds based on a two-way ANOVA (Col-0, 2.93 ± 0.27 ng sugar/mg seed; cesa5-1, 1.00 ± 0.40 ng sugar/mg seed, P < 0.005) and sos5-2 seeds showing a significant increase in non-adherent glucose (5.26 ± 0.47 ng sugar/mg seed, P < 0.005). The cesa5 sos5 double mutant had intermediate amounts of non-adherent glucose amounts (cesa5 sos5 1.1, 3.41 ± 0.68 ng sugar/mg seed; cesa5 sos5 3.7, 2.19 ± 0.27 ng sugar/mg seed) when compared to cesa5-1 and sos5-2 seeds. Levels of glucose were significantly reduced in cesa5-1 seeds compared to all other genotypes, while sos5-2 was increased (Tukey’s HSD, P < 0.05). Amounts of xylose were increased in all mutant lines analyzed (Col-0, 3.23 ± 0.06 ng sugar/mg seed; cesa5-1, 3.52 ± 0.02 ng sugar/mg seed; sos5-2, 4.04 ± 0.07 ng sugar/mg seed; cesa5 sos5 1.1, 3.63 +/- 0.06 ng sugar/mg seed; cesa5 sos5 3.7, 3.78 ± 0.19). Overall, non-adherent mucilage amounts were increased in cesa5-1, sos5-2 and cesa5 sos5 seeds. cesa5-1 seeds showed a significant reduction in non-adherent amounts of glucose, while sos5-2 seeds showed significant increase.  Whole seed alcohol insoluble residue (AIR) monosaccharide content was also quantified in order to determine if the changes seen in non-adherent mucilage are due to reduced adherence or changes in the amount of mucilage synthesized (Fig. 5.10B). Analysis of whole seed monosaccharide amounts can also provide information pertaining to the role of SOS5 in columella development. No significant changes in amounts of rhamnose were detected (Col-0, 188.15 ± 5.67 ng sugar/mg seed; cesa5-1, 192.80 ± 8.90 ng sugar/mg seed; sos5-2, 203.83 ± 8.80 ng sugar/mg seed; cesa5 sos5 1.1, 184.99 ± 4.33  134 ng sugar/mg seed; cesa5 sos5 3.7 188.47 ± 7.20 ng sugar/mg seed; Two-way ANOVA, P > 0.05). Very surprisingly, significant decreases in the amounts of galacturonic acid relative to wild type were identified in all mutant lines (Col-0, 318.56 ± 30.32 ng sugar/mg seed; cesa5-1, 204.13 ± 18.27 ng sugar/mg seed; sos5-2, 228.05 ± 41.56 ng sugar/mg seed; cesa5 sos5 1.1, 187.86 ± 34.00 ng sugar/mg seed; cesa5 sos5 3.7, 184.81 ± 35.25 ng sugar/mg seed; Two-way ANOVA, P < 0.05). Arabinose amounts were largely unchanged between Col-0 and the mutant lines. Interestingly both cesa5 sos5 double mutants showed decreases in the amount of galactose in whole seeds (P < 0.05). Finally cesa5-1, sos5-2 and cesa5 sos5 seeds showed decreases in total amounts of xylose isolated from whole seeds (Col-0, 84.84 ± 6.61 ng sugar/mg seed; cesa5-1, 63.16 ± 3.32 ng sugar/mg seed; sos5-2 68.40 ± 2.50 ng sugar/mg seed; cesa5 sos5 1.1, 53.60 ± 1.00 ng sugar/mg seed; cesa5 sos5 3.7 62.78 ± 2.86 ng sugar/mg seed). Both cesa5-1 and sos5-2 seeds showed decreases in the total amount of glucose, however, cesa5-1 seeds had significantly less glucose amounts than sos5-2 seeds based on a two-way ANOVA (Col-0, 108.81 ± 4.22 ng/mg seed; cesa5-1, 77.08  ± 2.70 ng/mg seed; sos5-2, 97.40 ± 3.64 ng /mg seed; P < 0.05). Both were significantly different from wild type seeds (P < 0.005 for cesa5-1, P < 0.05 for sos5-2). Both cesa5 sos5 lines showed similar reductions in the amount of glucose to cesa5-1, compared to Col-0 and sos5-2 seeds (cesa5 sos5 1.1 70.55 ± 2.17 ng/mg seed, P < 0.005; cesa5 sos5 3.7 71.32 ± 3.33 ng/mg seed P < 0.005; Tukey’s HSD, P < 0.05). By comparing the amounts of glucose in the double mutants to each single mutant indicates that both double mutants are significantly different from sos5-2 lines (cesa5 sos5 1.1 P < 0.01; cesa5 sos5 3.7 P < 0.001; Tukey’s HSD, P < 0.05). The difference in the amount of glucose between cesa5-1  135 seeds and double mutant seeds is less than 10% for each individual. These results indicate that SOS5 likely does not have any major effects on the amount of glucose present in cell walls, as cesa5-1 seeds show a substantial decrease in glucose amounts, while sos5-2 seeds are not decreased as significantly. Additionally, cesa5 sos5 double mutant lines showed similar amounts of glucose to cesa5-1 seeds. 5.2.6 SOS5 does not affect crystalline cellulose amounts in whole seeds  In order to more precisely determine the role of SOS5 on cellulose biosynthesis, crystalline cellulose content was quantified using the Updegraff method (Updegraff, 1969; Fig. 5.11). Cellulose amounts were significantly reduced in cesa5-1 and cesa5 sos5 seeds, but not in sos5-2 single mutant seeds, compared to Col-0 seeds as determined by a two-way ANOVA, with the two factors being the presence or absence of cesa5-1 and sos5-2 (Col-0, 125 ± 13.87 mg cellulose/g seed; cesa5-1, 63.26 ± 12.61, P < 0.05; sos5-2 99.52 ± 11.29; cesa5 sos5 69.53 ± 5.26, P < 0.05). Therefore, single cesa5-1 mutants produced significantly less cellulose than control or sos5-2 single mutant (ANOVA, sos5- 2, P < 0.01; Tukey’s HSD, P < 0.05). No significant effect of the sos5 mutation on cellulose production was observed (ANOVA, P = 0.6). Additionally, there was no significant interaction between the two mutations (ANOVA, cesa5-1 x sos5-2, P = 0.12) and cellulose production in the single mutant (cesa5-1) and the double mutant were not different (Tukey’s HSD, P > 0.05). The same trend was observed in three independent biological replicates, and is consistent with immunolabelling and S4B staining. This indicates that SOS5 has no significant role in cellulose biosynthesis and is mediating its influence on mucilage adherence and columella morphology through other mechanism(s).  136  Figure 5.11: Crystalline cellulose amounts in cesa5-1, sos5-2 and cesa5 sos5 double mutant whole seeds. Letters indicates different levels of values based on a two-way ANOVA with Tukey’s post-hoc analysis. Genotypes connected by the same letter show no significant difference, while those with different letters are significantly different, P < 0.05.  Error bars indicate standard error measurements, n=4. !" #!" $!" %!" &!" '!!" '#!" '$!" '%!" ()*+!" ,-./0+'" .).0+#" ,-./0".).0" !"#$%&' '()*+,*' '-'.$*+/ 01+,*'' -'.$*21 + $**3+,* ''+4&''5 ! ab a b ab  137 5.3 Discussion Previous studies have identified a role for CESA5 in cellulose biosynthesis and in mediating mucilage adherence (Mendu et al., 2011; Sullivan et al., 2011) and suggested that SOS5 mediates mucilage adherence through CESA5 (Harpaz-Saad et al., 2011). Indeed, given their pleiotropic phenotypes, AGPs are frequently implicated in cellulose biosynthesis, yet very few publications have shown a direct link between cellulose biosynthesis and AGPs (Girault et al., 2000; Seifert and Roberts, 2007; Li et al., 2010; MacMillan et al., 2010). I investigated the nature of the relationship between CESA5 and SOS5 by examining the phenotypes of cesa5 and sos5 mutants and the cesa5 sos5 double mutant. Three lines of evidence suggest that SOS5 has a unique role in mucilage adherence that is relatively independent of CESA5. First, unlike cesa5, sos5 does not affect the amount of glucose in seeds coat epidermal cells or the amount of crystalline cellulose in whole seeds. Secondly, detailed histologic analysis of cesa5 and sos5 mucilage revealed major differences in mucilage structure, including altered distribution of cellulose and pectins in the adherent layer, which may lead to loss of adherence. Finally, the cesa5 sos5 double mutant has a more severe mucilage adherence phenotype than either single mutant, as well as a more severe columella defects. The fact that SOS5 appears to act to increase mucilage adherence independently of CESA5 implies that SOS5 acts to increase the structural integrity of mucilage independently of cellulose. Therefore, a simple hypothesis is that SOS5 acts as a structural agent in mucilage. 5.3.1 Mechanism of SOS5 function in mucilage Understanding the functions of AGPs in plant cell walls has been challenging. Since SOS5 is not mediating these changes in cell wall structure through cellulose  138 biosynthesis, it must function through other mechanisms. Ray structures observed in wild type mucilage are completely absent in sos5-2 seeds. Further experiments have indicated that rays fail to form at any point during mucilage hydration (Data now shown). Immunolabelling indicates that rays are composed of both cellulose and pectins. SOS5 obviously has a major role in the formation of rays, but exactly how is unclear. I have ruled out that SOS5 is involved in cellulose biosynthesis amounts, but it could also be involved in the orientation of cellulose deposition. Further experiments examining the distribution of GFP::CESA5 in a sos5-2 mutant background are underway to determine if SOS5 is required for correct CESA patterning in the cytoplasmic column during mucilage biosynthesis. One final possibility is that SOS5 influences the cellulose microfibril angle in mucilage independently of the localization of CESA5. A more plausible hypothesis for the function of SOS5 in mucilage structure and adherence is through pectins. Mucilage is composed primarily of relatively unsubstituted RG I. sos5-2 mucilage shows enhanced loss of pectin immunolabelling epitopes in adherent mucilage (Fig. 5.5, 5.6) and increased amounts of galactose in the non-adherent mucilage fraction (Fig. 5.10A). Multiple lines of evidence show associations between AGPs and pectins. Pectins frequently co-localize with AGPs in pollen tubes (Jauh and Lord, 1995; Li et al., 1995; Mollet et al., 2002). Treatment of cell wall fractions with pectin degrading enzymes allows for the increased release of AGPs (Immerzeel et al., 2006; Lamport et al., 2006). AGPs have also been shown to bind specifically to pectins, in a calcium-dependent manner (Baldwin et al., 1993). While interactions between pectins and AGPs are well documented, the effect of this interaction is not well understood (Serpe and Nothnagel, 1999; Lamport; 2001; Lamport et al., 2006). SOS5  139 could be involved in providing structure and organization to the mucilage cell wall through interactions with pectins, which in turn affect the distribution of cellulose. 5.3.2 The function of SOS5 in determining columella structure. One of the most interesting aspects of the cesa5 sos5 double mutant phenotype is the drastic change observed in columella structure. While CESA5 has previously been implicated in columella and secondary cell wall deposition (Mendu et al., 2011), SOS5 was previously thought to not be involved (Harpaz-Saad et al., 2011). However, my SEM analysis of mature seeds does show a reduction in columella area in sos5-2 seeds, similar to the reduction seen in cesa5 seeds. Indeed, cesa5 sos5 seed coat epidermal cells often fail to develop any columella at all. The drastic changes in columella morphology show a clear role for SOS5 in columella development, which is not redundant with CESA5. It is unlikely that the changes in columella size and shape result from abnormal cellulose deposition because differences in cytoplasmic column shape were observed well before columella formation. The cytoplasmic column at 8 DPA, prior to secondary cell wall biosynthesis, is often noticeably smaller or absent in cesa5 sos5 seeds. The timing of columella biosynthesis appears normal. Thus, the defects in columella morphology arise from differences in the shape of the cytoplasmic column at the time of secondary cell wall deposition. The size and shape of the mucilage pocket is directly connected to the shape of the cytoplasmic column. The mucilage pocket may be affected in turn by the ability of the mucilage to expand prior to columella synthesis. Assuming the mucilage in cesa5, sos5 and cesa5 sos5 mutants is less cohesive, the mucilage pocket could expand more  140 than in wild-type cells, causing the cytoplasmic column to be reduced in size and thus distorting the shape of the columella deposited afterward. In conclusion, I have shown clear independent functions in mucilage adherence for CESA5 and SOS5. SOS5 does not affect the amount of cellulose synthesized, and it acts relatively independent of CESA5. These results indicate two separate structural elements that together establish mucilage adherence, one requiring SOS5 and another requiring cellulose synthesized by CESA5. These two genes are also both required for normal columella shape probably through the influence of the mucilage pocket on the shape of the cytoplasmic column.  141 6 Role of galactans in mucilage adherence 6.1 Introduction The plant cell wall requires interactions between polysaccharides for the formation of a cohesive network. Current models of primary cell wall structure suggest a load-bearing cellulose-hemicellulose network that is embedded in a network of pectins that provide considerable tensile strength on their own (reviewed in Cosgrove, 2005). How pectins interact with each other, and how pectins interact with the cellulose- hemicellulose network remains poorly understood. As was described in Chapter 5, the SOS5 AGP and the cellulose synthesized by CESA5 are both required for mucilage adherence, but appear to do so through distinct mechanisms. In this chapter, I explore the role of pectic galactan side chains in the mechanism of adherence mediated by SOS5 and the cellulose synthesized by CESA5. 6.1.1 The functions of galactans in mediating adherence Mucilage is composed primarily of relatively unbranched RG I. Adherent mucilage has increased amounts of HG, cellulose, and RG I side chains relative to non- adherent mucilage (Macquet et al., 2007a). RG I side chains are composed of linear or branched arabinans, galactans, and arabinogalactans. These glycosyl side chains are important for mediating interactions between pectic polysaccharides, and in mediating cell-to-cell adhesion (Orfila et al., 2001). Recently, interactions between pectins and cellulose have received much attention. Arabinans and galactans can bind cellulose in vitro (Zykwinska et al., 2005; Zykwinska et al., 2007a; Zykwinska et al., 2008), and in vivo (Oechslin et al., 2003). Pectins can be closely associated with cellulose in the cell  142 wall (Dick-Perez et al., 2010). These reports suggest a more direct role for pectins in mediating cell wall strength and integrity than current models predict. The use of Arabidopsis seed coat mucilage as a model to investigate cell wall properties has led to the identification of several genes encoding proteins that are required for mucilage release and adherence properties (Western et al., 2004; Dean et al., 2007; Macquet et al., 2007a; Rautengarten et al., 2008; Arsovski et al., 2009; Mendu et al., 2011). One such gene, MUM2, encodes a !-galactosidase that is required for mucilage release (Dean et al., 2007). MUM2 is an exo-acting enzyme that cleaves terminal galactose molecules, and does not appear to cleave internal glycosidic bonds in a polymer (Dean et al., 2007). When hydrated in water, mum2-1 seeds fail to release mucilage, and when treated with Na2CO3 mum2-1 seeds show limited mucilage release, and increased adherent mucilage (Dean et al., 2007; Macquet et al., 2007b). Therefore, MUM2 cleaves galactans, which reduces the adherence of seed coat mucilage. MUM2 has been hypothesized to act on RG I galactan side chains in mucilage (Dean et al., 2007; Macquet et al., 2007b), but may also act on AGP galactan side chains (Dean et al., 2007). One hypothesis for the role of MUM2 in seed mucilage extrusion is that it loosens mucilage by reducing the ability of galactans to bind to cellulose by degrading galactan side chains (Oechslin et al., 2003; Zykwinska et al., 2005; Zykwinska et al., 2007a; Dean et al., 2007; Zykwinska et al., 2008). Alternatively, MUM2 may loosen mucilage by modifying the function of AGPs, specifically SOS5, provided that SOS5 functions as a structural component of mucilage (Dean et al., 2007; Harpaz-Saad et al., 2011). I tested these hypotheses by constructing cesa5-1 mum2-1 and sos5-2 mum2-1 double mutants. The phenotype of both double mutants was compared to their respective single mutant  143 parents to determine the role of galactan side chains in CESA5 and SOS5 mediated mucilage adherence. 6.2 Results 6.2.1 CESA5 has no affect on the mucilage phenotype of mum2  If the failure of mum2 seed mucilage to extrude and expand is due to galactan interaction with cellulose, a double mutant with reduced cellulose in mucilage should be able to at least partially rescue the mum2 phenotype. To test this hypothesis, homozygous mum2-1 lines were crossed with homozygous cesa5-1 mutant lines and 3 double mutants homozygous for both the cesa5-1 T-DNA insertion and the mum2-1 point mutation were identified in an F2 population of 72 plants using PCR screening. Two of these cesa5 mum2 double mutants were selected for further analysis, and backcrossed to each parent to confirm their genotype. No major differences in the mucilage phenotype were detected between two independently isolated double mutants, indicating no other mutations affecting mucilage properties were present in the background. When shaken in water, Col-0 seeds show a large mucilage halo (Figure 6.1A), cesa5-1 seeds show no ruthenium red staining due to loss of mucilage adherence (Figure 6.1B) and mum2-1 mutants fail to extrude mucilage (Figure 6.1D). cesa5-1 mum2-1 double mutant seeds failed to release mucilage when treated with water, contrary to the original hypothesis (Fig 6.1C). Seeds were also treated with 0.1 M Na2CO3, a salt that weakens cell wall structure (Thimm et al., 2009) and has been shown to facilitate mucilage extrusion in mum2 mutants (Dean et al., 2007; Macquet et al., 2007b). Col-0 seeds shaken in 0.1 M Na2CO3 appear to have a larger mucilage halo than water treated seeds (Figure 6.1G), while  144  Figure 6.1: Mucilage extrusion and halo size of ruthenium red stained cesa5 mum2 and sos5 mum2 double mutants. Col-0: A, G, M; cesa5-1: B, H, N; cesa5-1 mum2-1: C, I, O; mum2-1: D, J, P; sos5-2 mum2-1: E, K, Q; sos5-2: F, L, R. A-F seeds shaken in water; G-R seeds shaken in Na2CO3. Bar: A-L 200 µm; M-R, 25 µm. ! "! #! $! %! &! '! (! )*! +! ,! -! .! /! 0! 1! 2! 3!  145 cesa5-1 0.1 M Na2CO3 treated seeds also appear to have increased ruthenium red staining compared to water treated seeds (Figure 6.1H). When treated with 0.1 M Na2CO3, cesa5- 1 mum2-1 and mum2-1 seeds showed limited mucilage release, and have very similar mucilage halo sizes (Figure 6.1I, J and O, P). cesa5-1 mum2-1 seeds treated with Na2CO3 have a reduced mucilage halo size compared to cesa5-1 seeds (compare Fig. 6.1O with P). Together, these data imply that the failure of mum2 seed mucilage to expand and extrude is not dependent on the cellulose produced by CESA5. 6.2.2 SOS5 is required for MUM2 function If the failure of mum2 seed mucilage to extrude and expand requires SOS5, then the sos5-2 mutation should be able to rescue the mum2-1 mucilage phenotype. To test this hypothesis, homozygous mum2-1 lines were crossed with homozygous sos5-2 mutant lines. In the F2 population, both genes segregated independently. Four independent double mutants were isolated from a population of 72 F2 individuals, as determined through PCR screening for the sos5-2 T-DNA insertion and the mum2-1 point mutation. One sos5-2 mum2-1 double mutant was selected for further analysis and was backcrossed to the single mutant parents to confirm their genotype. Similar to the cesa5-1 mum2-1 double mutant, sos5-2 mum2-1 seeds failed to release mucilage when treated with water (Fig 6.1E, F). However, when treated with Na2CO3, sos5-2 mum2-1 seeds release large mucilage halos (Figure 6.1K, Q) similar to sos5-2 but unlike mum2-1 seeds. Thus, sos5-2 is able to, at least partially, rescue the mum2-1 phenotype in terms of mucilage expansion and halo size.  146 6.2.3 Quantification of mucilage release in cesa5-1 mum2-1 and sos5-2 mum2-1 double mutants  The affect of cesa5-1, sos5-2, and mum2-1 mutations on the adherence of mucilage was quantified by extracting mucilage with 0.1 M Na2CO3 and analyzing the mucilage monosaccharide composition using High performance Anion-exchange chromatograph (HPAEC). The most abundant monosaccharides in mucilage, rhamnose and galacturonic acid (galacturonic acid), were used to quantify non-adherent mucilage amounts. As expected, cesa5-1 and sos5-2 mutant seeds show an increase in non- adherent mucilage amounts relative to wild type, while mum2-1 mutant seeds have reduced non-adherent mucilage amounts (Figure 6.2A). Similar to mum2-1 seeds, cesa5- 1 mum2-1 double mutant seeds released less non-adherent mucilage than wild type, indicating that a decrease in cellulose had no effect on the release of mum2-1 mucilage. In contrast, sos5-2 mum2-1 double mutants showed roughly wild type amounts of non- adherent rhamnose and galacturonic acid, demonstrating increased mucilage release compared to mum2-1 seeds, yet reduced release compared to sos5-2 seeds. These data confirms the observation that sos5-2 mum2-1 seeds have increased non-adherent mucilage amounts compared to mum2-1 seeds. When analyzing other minor components of non-adherent mucilage, insights into the mechanism of mucilage adherence can be gleaned (Figure 6.2B). Arabinose amounts were largely equal in Col-0, mum2-1, cesa5-1 and cesa5-1 mum2-1 seeds; however, small increases detected in sos5-2 and sos5-2 mum2-1 seeds. Galactose amounts were relatively consistent between Col-0, cesa5-1, mum2-1, and cesa5-1 mum2-1 double mutant seeds. Surprisingly, sos5-2 and sos5-2 mum2-1 double mutant seeds showed increases in the amount of non-adherent galactose amounts compared to wild type and  147  Figure 6.2: Monosaccharide content of non-adherent mucilage from Na2CO3 treated seeds. A) Major components: rhamnose and galacturonic acid. B) Minor components: arabinose, galactose, glucose and xylose. Error bars are standard error measurements, n=4 !" #!" $!" %!" &!" '!" (!" )*+,-./0" 1+2+3456.-73"8379"N a 2 C O 3 t re at ed , n on -a dh er en t m uc ila ge  fr ac tio n (! g su ga r/m g se ed )! !! "! 0 1 2 3 4 5 6 7 8 Arabinose Galactose Glucose Xylose Col-0 sos5-2 mum2 sos5 mum2-1 cesa5 mum2 cesa5-1  148 mum2-1 seeds (Col-0, 2.40 ± 0.59 ng/mg seed; sos5-2, 3.91 ± 0.57 ng/mg seed; sos5-2 mum2-1, 3.79 ± 0.16 ng/mg seed). Glucose levels were also increased in the non-adherent mucilage fractions of sos5-2 and sos5-2 mum2-1 seeds compared to Col-0 seeds (Col-0, 2.52 ± 0.30 ng/mg seed; sos5-2, 6.50 ± 0.80 ng/mg seed; sos5-2 mum2-1, 4.45 ± 0.61 ng/mg seed). These results demonstrate that sos5-2 and sos5-2 mum2-1 lines have increased galactose and glucose amounts in the non-adherent mucilage fraction, similar to results observed for rhamnose and galacturonic acid. SOS5 is therefore required for the adherence of galactose containing polysaccharides, which are most likely linked to RG I. Whole seed monosaccharide concentrations were also quantified in order to determine if the defects seen in the mucilage fraction are due to changes in mucilage adherence, or changes in whole seed cell wall biosynthesis (Figure 6.3). Relative to wild type, only minor reductions in monosaccharide amounts were observed in mutant seeds, except for a strong reduction in glucose in cesa5-1 lines (Col-0, 108.81 ± 4.22 ng sugar/mg seed; cesa5-1, 77.08 ± 2.70 ng sugar/mg seed; cesa5-1 mum2-1, 52.51 ± 5.34 ng sugar/mg seed). Concentrations of galacturonic acid were also reduced in all mutant lines. In summary, the monosaccharide analysis of the non-adherent mucilage fraction indicates reduced mucilage adherence in sos5-2 mum2-1 seeds compared to mum2-1 seeds, while loss of CESA5 has no major effects on the phenotype of mum2-1 mucilage. The major changes seen in non-adherent mucilage concentration of these mutants are not reflected in whole seed monosaccharide amounts, indicating they are specific changes to the properties of mucilage adherence.  149 Fig 6.3: Monosaccharide analysis of whole seed AIR. W ho le  se ed  A IR  m on os ac ca hr id e co nt en t ( !g  su ga r/m g se ed )! 0 50 100 150 200 250 300 350 400 Rhamnose Galacturonic Acid Arabinose Galactose Glucose Xylose Col-0 sos5-2 sos5 mum2 mum2-1 cesa5 mum2 cesa5-1  150 6.3 Discussion  As discussed in Chapter 5, CESA5 and SOS5 mediate mucilage adherence through independent mechanisms. However, it is not clear how either protein functions to organize the mucilage so that it is adherent to the parent seed. MUM2 is a !-galactosidase that functions in the apoplast to allow mucilage to extrude and expand upon hydration (Dean et al., 2007). Loss-of-function mum2 mutations result in a phenotype where mucilage fails to expand and extrude from the seed coat. Because this phenotype is distinct from the effects of CESA5 and SOS5, I constructed cesa5-1 mum2-1 and sos5-2 mum2-1 double mutants to determine whether CESA5 or SOS5 functions antagonistically to MUM2 to mediate mucilage adherence. My results show that cellulose is not required for the function of MUM2 in mucilage adherence and release. However, MUM2 function requires SOS5, demonstrating an important role for galactans in mediating mucilage release and adherence. Seed mucilage contains only a small amount of galactose; monosaccharide and linkage analysis indicates that the galactans are very short (1-3 sugar units long) and are fairly rare in mucilage (Dean et al., 2007). In fact, mucilage RG I has only one branch point for every ~50 rhamnose molecules (Dean et al., 2007). However, in mum2-1 mucilage, there is one branch for every 25 rhamnose molecules, and galactans are slightly longer (Dean et al., 2007). It was not clear how the loss of function of MUM2 could have such a drastic effect on mucilage given that galactans are such a small proportion of mucilage. The primary hypothesis concerning the function of MUM2 was that it shortened galactan side chains found on RG I (Dean et al., 2007; Macquet et al., 2007b;  151 Huang et al., 2011). However, linkage analysis of wild type and mum2-1 mucilage is more consistent with short galactans found within the glycosyl side chains found on AGPs than with the normally longer RG I galactan side chains (Churns et al., 1984; Bacic et al., 1987; Dean et al., 2007; Huang et al., 2011). Specifically, mum2-1 mucilage contains increased amounts of terminal galactose molecules, 1,6 galactose, and 3,6- galactose linkages, consisted with arabinogalactan II molecules found on RG I side chains and AGP glycosyl side chains (Dean et al., 2007). Therefore, MUM2 could be affecting mucilage hydration and adherence through removal of galactose side chains from RG I, SOS5, or both. Longer galactan side chains, present on either RG I or SOS5 could be involved in increasing adherence by mediating interactions with other cell wall polymers. 6.3.1 Relationship between MUM2 and CESA5  Galactans have been reported to be able to bind to cellulose (Oechslin et al., 2003; Zykwinska et al., 2005; Zykwinska et al., 2007a). Assuming that loss of MUM2 function results in longer galactan side chains, creating a cesa5-1 mum2-1 double mutant tested the potential interactions between galactans and cellulose. In this scenario, RG I galactan side chains would be longer due to a loss of MUM2 activity, increasing interactions with cellulose (Oechslin et al., 2003; Zykwinska et al., 2005; Zykwinska et al., 2007a; Zykwinska et al., 2008). However, total amounts of cellulose would be reduced due to the cesa5-1 mutant background, which should partially rescue the mum2-1 phenotype. No evidence of such a rescue was observed in a cesa5-1 mum2-1 double mutant, indicating that the role of MUM2 is relatively independent of CESA5 function. These results are not consistent with the hypothesis that MUM2 degrades galactans to reduce interaction with  152 cellulose. 6.3.2 Relationship between SOS5 and MUM2 AGP contain galactans, and have been hypothesized to be involved in MUM2’s function in mucilage modification (Dean et al., 2007). Thus, it is possible that MUM2 and SOS5 functionally interact to mediate mucilage adherence. In stark contrast to the cesa5-1 mum2-1 double mutant, when hydrated in Na2CO3, sos5-2 mum2-1 seeds appeared phenotypically similar to sos5-2, suggesting that mutation of SOS5 can rescue the mum2-1 phenotype. This provides indirect evidence that SOS5 is required for the function of MUM2. The dependence of MUM2 on SOS5 could be explained in one of two ways. MUM2 acts directly on SOS5 to remove galactose side-chains or alternatively, MUM2 acts on RG I side chains that are required to interact with SOS5. Interestingly, non- adherent mucilage in sos5-2 and sos5-2 mum2-1 shows significant increases in galactose amounts (Fig. 6.2). This increase in galactans indicates that SOS5 mediates the adherence of galactan containing polymers. MUM2 acts to degrade galactan polysaccharides that are required for SOS5 mediated mucilage adherence. MUM2 can either directly degrade SOS5 glycans, or SOS5 could be mediating the access of MUM2 to RG I galactans. Degradation of both targets may result in reduced mucilage adherence. With either hypothesis, I can conclude that both RG I galactans, and SOS5 are important for mucilage extrusion and adherence, and that the role of galactans is dependent on SOS5. 6.3.3 Galactan function in adherence Exactly how galactans function in adherence is not clear. Many studies have identified other glycosyl hydrolases that can act on RG I side chains and AGP glycans.  153 For example, in tobacco, pollen tube elongation is stimulated by degradation of AGP glycans in the stylar transmitting tract (Cheung et al., 1995; Wu et al., 1995). !- Galactosidases can also cleave 1-4 linked galactans, typical of RG I galactan side chains (Pressey et al., 1983; Martin et al., 2005). Inhibition of AGP function by Yariv reagent treatment (Kitazawa et al., 2013) affected the transient appearance of pectic galactans in radish roots, demonstrating a role for AGPs in regulating RG I galactan metabolism (McCartney et al., 2003). !-Galactosidases have also been demonstrated to degrade AGP galactan side chains (Ichinose et al., 2006; Takata et al., 2010). Arabinogalactan side chains are frequently decorated with terminal arabinose molecules that can protect internal galactan residues from further cleavage. However, arabinofuranosidases can cleave terminal L-arabinofuranosyl residues on AGP side chains, which can make the remaining glycans more susceptible to !-Galactosidases (Sekimitata et al., 1989; Kotake et al., 2005). The complex structure of arabinogalactan side chains presents an interesting mechanism to modify AGP and/or RG I side chain function through sequential degradation. This research emphasizes the importance of galactans in cell wall organization, and their modifications by glycosyl hydrolases. It is apparent that SOS5 is required for polymer interactions and structural organization of mucilage, and that hydrolyases can alter AGP function to reduce these interactions and loosen the cell wall. Based on my results I propose that SOS5 acts to bind and cross link different pectin components, which serves to increase wall strength and reduce plasticity. The glycosyl side chains present on AGPs, which can be degraded by hydrolases that will then increase wall plasticity, mediate this function. This balance between increased wall strength mediated by AGPs, followed by side chain degradation to reduce wall strength  154 could explain how mucilage separates into two distinct fractions: a low-molecular weight, RG I rich non-adherent fraction and a higher molecular weight (indicating increased polymer connections) RG I rich adherent fraction (Macquet et al., 2007a). The hypothesis that MUM2 acts on SOS5, and that SOS5 crosslinks pectic polysaccharides deserves to be fully explored. In order to test this hypothesis, SOS5 could be tagged and purified from developing seeds, or mature mucilage. All attempts at cloning SOS5 with an epitope tag has thus far failed. However, if a tagged SOS5 protein were to become available, SOS5 could be purified from developing seeds and the activity of MUM2 on SOS5 galactan side chains could be directly tested in vitro with a purified recombinant MUM2 protein. Additionally, purification of SOS5 could allow for identification of specific linkages with pectins using multiple different approaches such as mass spectrometry, nuclear magnetic resonance, and linkage analysis, which could elucidate any potential interactions between SOS5 and pectins. 6.3.4 Mechanism of mucilage extrusion and adherence The fact that SOS5 suppression of MUM2 requires Na2CO3 and that, based on monosaccharide analysis, the rescue is not complete even in Na2CO3 (Figure 6.2A) suggest that MUM2 may have functions that are not dependent on SOS5. This observation can be explained by MUM2 having a role in the primary cell wall that does not require SOS5. Preventing mucilage extrusion through the outer primary cell wall could result in increased mucilage adherence. The exact function of Na2CO3 in loosening the cell wall is poorly understood, however, it has been used extensively in the extraction of pectins (Selvendran and Ryden, 1990; Fry, 2000; McCartney and Knox; 2002; Thimm et al., 2009).  155 In summary, the role of galactans was investigated in mucilage adherence mediated by CESA5 and SOS5. Galactans can bind to cellulose and potentially mediate mucilage adherence in association with CESA5. However, reduced cellulose in cesa5-1 mucilage did not have any affect on mum2-1 mucilage, indicating that cellulose synthesized by CESA5 is not interacting with galactans to mediate adherence. SOS5 appears to be required for MUM2 function, indicating that MUM2 removes galactose from SOS5 and/or from RG I galactans that interacts with SOS5. This illustrates the importance of SOS5 in galactan mediated mucilage adherence. While the exact mechanism of mucilage adherence mediated by SOS5 remains elusive, this research provides useful information regarding the further determination of the function of SOS5.   156 7 Conclusions and future directions 7.1 Overview  Plant cell walls are complex structures that can rapidly expand to accommodate growth, while maintaining strength and rigidity required to withstand internal osmotic pressures. The plasticity of the cell wall is mediated by the composition of cell wall polymers, and their interactions. Despite many years of research, our understanding of plant cell polysaccharide interactions and remodelling remain poor. In this thesis, the Arabidopsis seed coat was used as a model system to study cellulose biosynthesis and polysaccharide interactions. Two independent polysaccharide networks mediate seed coat mucilage cell wall integrity (Fig. 7.1). First, cellulose mediates adherence through either direct or indirect interactions with RG I. Secondly, SOS5 is required for mucilage adherence, which is mediated by galactan metabolism. 7.1.1 Cellulose in the seed coat 7.1.1.1 Cellulose biosynthesis during seed coat development  Cellulose is an important part of plant cell walls and understanding how it is synthesized is a major research goal. Previously, CESA2 and CESA9 were shown to be redundant to CESA6 in pollen, while CESA5 is also redundant to CESA6 in hypocotyl elongation (Desprez et al., 2007; Persson et al., 2007; Bischoff et al., 2011). However, in seed coat secondary wall biosynthesis, the contribution of CESA2, CESA5 and CESA9 is additive (Chapter 3), indicating incomplete redundancy. Columella and radial wall biosynthesis requires that a vast amount of cellulose must be synthesized in a short time period. The simplest interpretation of the role of CESA2, CESA5 and CESA9 during  157   Figure 7.1: Diagram of mucilage structural networks, organization and biosynthesis. CESA5 and CESA3 synthesize cellulose in mucilage, which is required for adherence. Independently of cellulose, SOS5 is required for proper galactan-mediated adherence, which involves MUM2, and is independent of cellulose. CESA2, CESA5 and CESA9 are required for proper secondary cell wall biosynthesis.        !"#$%! !"#$&!                                   #'#%!             ()(*!  !"#$%!  !"#$*! !"#$+!  Legend:! Rhamnogalacturonan I! Cellulose! Galactans!  158 columella biosynthesis is that these CESAs form a pool from which they can be recruited for secondary wall biosynthesis. With the loss of each CESA in single, double or triple mutants, this pool of CESAs is reduced, which in turn reduces the amount of cellulose synthesized in the secondary cell wall. In current models of CSC assembly, a fully functional CSC requires only three CESAs, yet 10 CESAs are present in the Arabidopsis genome (Taylor et al., 2003; Desprez et al., 2007). The CESA gene family presents an interesting case of functional specificity and redundancy. Previous analysis of the CESA6-like CESAs indicated that they were fully redundant to CESA6 in many tissues such as hypocotyls and pollen (Persson et al., 2007; Desprez et al., 2007). This research demonstrates that CESA2, CESA5 and CESA9 have functional significance in secondary cell wall deposition. The simplest explanation is that all three CESAs are present in a pool of CESAs available for cellulose biosynthesis. However, a plausible alternative hypothesis explaining this observation is that these three CESAs could interact to form a functional CSC. Examining the potential biochemical interactions between CESA2, CESA5 and CESA9 would help our understanding of CSC formation.  While all three CESAs are required for columella and radial wall synthesis, CESA5 has a unique role during mucilage biosynthesis. The identification of the role of CESA3 in mucilage biosynthesis aids our understanding of the composition of the CSC during mucilage biosynthesis. It is curious that cesa1 mutant lines do not show any obvious mucilage phenotypes (Chapter 4; Burn et al., 2002; Sullivan et al., 2011). Given that three CESAs are required for CSC formation, and that CESA1 and CESA3 are  159 considered core components of primary wall cellulose synthesis, it is possible that CESA10 is redundant with CESA1 during mucilage biosynthesis. CESA1 and CESA10 are closely related, and may exhibit functional redundancy similar to the CESA6-like CESAs, which would explain the lack of any visible mucilage phenotype in cesa1 or cesa10 mutant seeds. A cesa1 cesa10 double mutant would clarify their role during seed coat development. 7.1.1.1 The function of cellulose in mucilage adherence  The presence of cellulose in mucilage has been clearly documented (Blake et al., 2006; Macquet et al., 2007a; Young et al., 2008;). However, until this research the functional significance of cellulose in mucilage was largely unknown. Here, I show that cellulose is required for mucilage adherence. However, how cellulose mediates mucilage adherence is still not clear. Interactions between pectins and cellulose are well documented (Iwai et al., 2001; Oechslin et al., 2003; Vignon et al., 2004; Zykwinska et al., 2005; Chebli et al., 2012). Given that mucilage is composed primarily of RG I, cellulose could be mediating adherence through pectins, or hemicelluloses which are in turn bound to pectins (Thompson and Fry, 2000; Popper and Fry, 2008). Xyloglucans have been identified in adherent mucilage, however, their functional role remains to be established (Young et al., 2008).  The ixr1-1 and ixr1-2 point mutations also help provide useful information regarding the role of cellulose in mucilage adherence as described in Chapter 4. The degree of cellulose crystallinity appears reduced in ixr1 seed mucilage, which caused drastic rearrangement of mucilage structure. This further establishes the link between cellulose and pectins in maintaining cell wall integrity. Using mucilage as a cell wall  160 model could be useful in determining the affects of the degree of cellulose crystallinity in interactions with other cell wall polymers. 7.1.2 Two independent cell wall networks  As discussed in Chapter 5, SOS5 mediates mucilage adherence independently of CESA5. The phenotype of the double mutant is additive, implying two independent cell wall networks that are both required for mucilage adherence. As mentioned above, the first network is composed of cellulose and pectin/hemicellulose, while SOS5 functions to mediate adherence through galactans. Primary plant cell walls are often described as a combination of the cellulose-hemicellulose network and the pectin network. Cell wall modeling research has suggested that hemicellulose tethering of cellulose microfibrils is insufficient to maintain cell wall integrity, and that other mechanisms must be present to increase wall strength (Yi & Puri, 2012). I propose two independent cell wall matrices are required for mucilage adherence. First, cellulose, potentially linked by hemicellulose forms a scaffold required to keep adherent mucilage in place. Secondly, RG I and other pectins, in association with SOS5, form a matrix that is embedded in cellulose scaffold. SOS5 is required for pectin organization in this matrix, and potentially is involved in crosslinking pectin with other pectin molecules, and/or cellulose.  Recently, an AGP was shown to be covalently linked to RG I through galactan side chains (Tan et al., 2013). Through the identification of multiple different covalent bonds, the authors were able to conclusively show for the first time that an AGP can be directly linked to pectins and other cell wall polysaccharides (Tan et al., 2013). They suggest that other AGPs may serve as structural agents, cross-linking some pectin and hemicellulose polysaccharides with proteins, forming a continuous matrix (Tan et al.,  161 2013). A direct linkage between SOS5 and RG I in mucilage remains to be identified. However, a direct linkage between SOS5 and RG I seems likely, and could explain how SOS5 mediates adherence independently of cellulose. 7.2 Future directions 7.2.1 Cellulose biosynthesis in the Arabidopsis seed coat  The Arabidopsis seed coat is a proven system for the analysis of cellulose biosynthesis. This system is now ripe for further exploitation. The most obvious questions remaining are: is a third CESA required to form a CSC; and what is the identity of this third CESA. As mentioned above, a cesa1 cesa10 double mutant might provide some answers to this question. Additionally, identification of specific protein-protein interactions between CESAs might shed light on the redundancy between CESA2, CESA5 and CESA9. Furthermore, mucilage and columella development can be used to examine the effects of cellulose crystallinity on cell wall structure and function. It would be interesting to express different kinds of cellulases that can specifically target crystalline or amorphous regions of cellulose, such as endoglucanase or cellobiohydrolases, using a seed coat specific promoter (Esfandiari et al., 2013). Finally, microtubules are required for the proper deposition of cellulose (Green, 1962; Paradez et al., 2006; Guiterrez et al., 2009; Bischoff et al., 2011; Fujita et al., 2011). Proper microtubule assembly is also required for normal mucilage development (McFarlane et al., 2008). Further investigation of the relationship between microtubules, CESA velocities, and cellulose crystallinity in seed coat mucilage would provide a complete story of cellulose biosynthesis in mucilage.  162 7.2.2 Transcriptional regulation of cellulose biosynthesis Seed coat development has proven to be a useful tool in identifying highly expressed genes involved in cell wall biosynthesis. Additionally, many different transcription factors are known to regulate seed coat epidermal cell development (Fig 1.4). Our understanding of the regulation of primary wall CESA expression is sorely lacking. The identification of CESA3 and CESA5 in mucilage, and CESA2, CESA5 and CESA9 during columella biosynthesis represents an excellent opportunity to investigate what transcription factors control their expression. For example, GL2 represses CESA5 expression in roots (Tominaga-Wada et al., 2009). GL2 up regulates MUM4 during mucilage biosynthesis and would also be required as a transcriptional activator of CESA5 during seed coat development (Western et al., 2001; Western et al., 2004). 7.2.3 Function of SOS5  While SOS5 is required MUM2 function in mucilage, its exact role is still unclear. The most pressing question is the subcellular localization and expression pattern of SOS5. Determining the exact composition of the glycosyl side chains of SOS5 could provide further clues as to how it is involved in galactan metabolism. Given that AGP glycans are frequently decorated with other sugars such as arabinans, glucuronic acid and rhamnose, elucidation of the glycosyl structure would be very informative. It is also possible that SOS5 interacts with other proteins in mucilage. Examining any protein- protein interactions in mucilage would be a useful experiment. 7.3 Summary  Cell walls are the culmination of complex interactions between the different cell wall components. Understanding how these components are synthesized and assembled  163 into a mature cell wall is a primary goal of plant research. In Arabidopsis seed coat epidermal cells, CESA2, CESA5 and CESA9, previously identified as primary wall CESAs, are involved in secondary cell wall biosynthesis. CESA5 has a distinct role from the other CESA6-like CESAs in mucilage biosynthesis. CESA3 is also involved in mucilage biosynthesis, and the ixr1 mutations result in reduced cellulose crystallinity that modifies cell wall polymer distribution. CESA5 and SOS5 mediate mucilage adherence independently, indicating two independent cell wall networks that are both required for mucilage adherence. SOS5 is also required for proper galactan metabolism mediated by MUM2. The research in this thesis has led to conclusions that have enhanced our understanding of cell wall biosynthesis, assembly, and polysaccharide interactions required for proper cell wall function    164 References Abdel-Massih, R.M., Baydoun, E.A., & Brett, C.T. (2003). In vitro biosynthesis of 1,4- beta-galactan attached to a pectin-xyloglucan complex in pea. Planta, 216(3), 502- 11. Abramoff, M.D., Magalhaes, P.J., & Ram, S.J. (2004). Image processing with ImageJ. Biophotonics Int. 11, 36–42. Albersheim, P., Darvill, A., Roberts, K., Sederoff, R., & Staehelin, A. (2011). Plant cell walls. Garland science. Alonso, J.M., Stepanova, A.N., Leisse, T.J., Kim, C.J., Chen, H., Shinn, P., Stevenson, D.K., Zimmerman, J., Barajas, P., Cheuk, R., et al. (2003). Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science, 301, 653–657. Anderson, C.T., Carroll, A., Akhmetova, L., & Somerville, C. (2010). Real-time imaging of cellulose reorientation during cell wall expansion in Arabidopsis roots. Plant Physiol., 152, 787–796. Arioli, T., Peng, L., Bretzner, A.S., Burn, J., Wittke, W., Herth, W., Camilleri, C., Hofte, H., Plazinski, J., Birch, R., Cork, A., Glover, J., Redmond, J., & Williamson, R.E. (1998). Molecular analysis of cellulose biosynthesis in Arabidopsis. Science, 279(5351), 717-20. Arsovski, A.A., Haughn, G.W., & Western T.L. (2010). Seed coat mucilage cells of Arabidopsis thaliana as a model for plant cell wall research. Plant Signal Behav., 5, 796–801 . Arsovski, A.A., Popma, T.M., Haughn, G.W., Carpita, N.C., McCann, M.C., & Western, T.L. (2009a). AtBXL1 encodes a bifunctional !-D-Xylosidase/"-L- arabinofuranosidase required for pectic arabinan modification in Arabidopsis  165 mucilage secretory cells. Plant Physiol., 150, 1219 –1234. Arsovski, A.A., Villota, M.M., Rowland, O., Subramaniam, R., & Western, T.L. (2009b). MUM ENHANCERS are important for seed coat mucilage production and mucilage secretory cell differentiation in Arabidopsis thaliana. J. Exp. Bot. 60, 2601–2612. Atmodjo, M. A., Sakuragi, Y., Zhu, X., Burrell, A. J., Mohanty, S. S., Atwood, J. A., Orlando, R., Scheller, H.V., & Mohnen, D. (2011). Galacturonosyltransferase (GAUT)1 and GAUT7 are the core of a plant cell wall pectin biosynthetic homogalacturonan:galacturonosyltransferase complex. Proc. Natl. Acad. Sci. USA, 3-8. Atanassoc, I.I., Pittman, J.K., & Turner, S.R. (2009). Elucidating the mechanism of assembly and subunit interaction of the cellulose synthase complex of Arabidopsis secondary cell walls. J. Biol. Chem., 284(6), 3833-41. Baldwin, T.C., McCann, M.C., & Roberts, K. (1993) A novel hydroxyproline-deflcient arabinogalactan protein secreted by suspension-cultured cells of Daucus carota. Purification and partial characterization. Plant Physiol., 103, 115–123. Bacic, A., Churms, S.C., Stephen, A.M., Cohen, P.B. & Fincher, G.B. (1987). Fine structure of the arabinogalactan-protein from Lolium multiflorum. Carbohydr. Res., 162, 85–93. Baskin, T.I. (2001). On the alignment of cellulose microfibrils by cortical microtubules: A review and a model. Protoplasma, 215, 150-171. Baskin, T.I., Beemster, G.T.S., Judy-March, J.E., & Marga, F. (2004). Disorganization of cortical microtubules stimulates tangential expansion and reduces the uniformity of cellulose microfibril alignment among cells in the root of Arabidopsis. Plant Physiol., 135, 2279-2290.  166 Bassel, G.W., Fung, P., Chow, T.F., Foong, J.A., Provart, N.J., & Cutler, S.R. (2008). Elucidating the germination transcriptional program using small molecules. Plant Physiol., 147(1), 143-55. Beeckman, T., Przemeck, G.K.H., Stamatiou, G., Lau, R., Terryn, N., De Rycke, R., Inze, D., & Berleth, T. (2002). Genetic complexity of cellulose synthase A gene function in Arabidopsis embryogenesis. Plant Physiol., 130, 1883–1893. Benziman, M., Haigler, C.H., Brown, R.M., White, A.R., & Cooper, K.M. (1980). Cellulose biogenesis: Polymerization and crystallization are couple processes in Acetobacter xylinum. Proc. Natl. Acad. Sci. USA, 77(11), 6678-82. Bichet, A., Desnos, T., Turner, S., Grandjean, O., & Hofte, H. (2001). BOTERO1 is required for normal orientation of cortical microtubules and anisotropic cell expansion in Arabidopsis. Plant J., 25, 137-148. Bischoff, V., Desprez, T., Mouille, G., Vernhettes, S., Gonneau, M., & Hofte, H. (2011). Phytochrome regulation of cellulose synthesis in Arabidopsis. Curr. Biol., 21(21), 1822-7. Blake, A.W., McCartney, L., Flint, J.E., Bolam, D.N., Boraston, A.B., Gilbert, H.J., & Knox, J.P. (2006). Understanding the biological rationale for the diversity of cellulose-directed carbohydrate-binding modules in prokaryotic enzymes. J. Biol. Chem., 281(39), 29321-9. Borner, G.H., Lilley, K.S., Stevens, T.J., & Dupree, P. (2003). Identification of glycosylphosphatidylinositol-anchored proteins in Arabidopsis. A proteomic and genomic analysis. Plant Physiol., 132(2), 568-77. Bouton, S., Leboeuf, E., Mouille, G., Leydecker, M.T., Talbotec, J., Granier, F., Lahaye, M., Hofte, H., & Truong, H.N. (2002). QUASIMODO1 encodes a putative membrane-bound glycosyltransferase required for normal pectin synthesis and  167 cell adhesion in Arabidopsis. Plant Cell, 14(10), 1577-90. Brett, C.T. (2000). Cellulose microfibrils in plants: biosynthesis, deposition, and integration into the cell wall. Int. Rev. Cytol., 99, 161–199. Bringmann, M., Li, E., Sampathkumar, A., Kocabek, T., Hauser, M.T., & Persson, S. (2012). POM-POM2/cellulose synthase interacting1 is essential for the functional association of cellulose microtubules in Arabidopsis. Plant Cell, 24(1), 163-77. Buchanan,B.B., Gruissem, W., & Jones, R.L, editors. (2000). Biochemistry & molecular biology of plants. John Wiley & Sons. Burn, J.E., Hurley, U.A., Birch, R.J., Arioli, T., Cork, A., & Williamson, R.E. (2002). The cellulose-deficient Arabidopsis mutant rsw3 is defective in a gene encoding a putatitve glucosidase II, an enzyme processing N-glycans during ER quality control. Plant J. 32(6), 949-60. Burton, R.A., Gidley, M.J., & Fincher, G.B. (2010). Heterogeneity in the chemistry, structure and function of plant cell walls. Nat. Chem. Biol., 6, 724-732. Caffal, K.H., & Mohnen, D. (2009). The structure, function, and biosynthesis of plant cell wall pectic polysaccharides. Carbohydr. Res., 344(14), 1879-900. Cannon, M.C., Terneus, K., Hall, Q., Tan, L., Wang, Y., Wegenhart, B.L., Chen, L., Lamport, D.T., Chen, Y., & Kieliszewski, M.J. (2008). Self-assembly of the plant cell wall requires an extension scaffold. Proc. Natl. Acad. Sci. USA, 105(6), 2226-31. Carroll, A., & Specht, C.D. (2011) Understanding plant cellulose synthases through a comprehensive investigation of the cellulose synthase family sequences. Front. Plant Sci., 2(5), 1-11. Carpita, N.C. (2011). Update on mechanisms of plant cell wall biosynthesis: how plants  168 make cellulose and other (1->4)-b-d-glycans. Plant Physiol., 155(1), 171-84. Carpita, N., & Vergara, C. (1998). A recipe for cellulose. Science, 279, 672–673. Cavalier, D.M., & Keegstra, K. (2006). Two xyloglucan xylosyltransferases catalyze the addition of multiple xylosyl residues to cellohexaose. J. Biol. Chem., 281(45), 34197-207. Cavalier, D.M., Lerouxel, O., Neumetzler, L., Yamauchi, K., Reinecke, A., Freshour, G., Zabotina, O.A., Hahn, M.G., Burgert, I., Pauly, M., Raikhel, N.V., & Keegstra, K. (2008). Disrupting two Arabidopsis thaliana xylosyltransferase genes results in plants deficient in xyloglucan, a major primary cell wall component. Plant Cell, 20(6), 1519-37. Chan, J., Crowell, E., Eder, M., Calder, G., Bunnewell, S., Findlay, K., Vernhettes, S., Hofte, H., & Lloyd, C. (2010). The rotation of cellulose synthase trajectories is microtubule dependent and influences the texture of epidermal cell walls in Arabidopsis hypocotyls. J. Cell Sci, 123(20), 3490-5. Chebli, Y., Kaneda, M., Zerzour, R., & Geitmann, A. (2012). The cell wall of the Arabidopsis pollen tube-spatial distribution, recycling, and network formation of polysaccharides. Plant Physiol., 160(4), 1940-55. Cheung, A.Y., Wang, H., & Wu, H.M. (1995). A floral transmitting tissue-specific glycoprotein attracts pollen tubes and stimulates their growth. Cell, 82(3), 383-93. Churms, S.C., & Stephen, S.C. (1984). Structural studies of an arabinogalactan-protein from the gum exudate of Acacia robusta. Carbohydr. Res., 133 (1984), pp. 105– 123. Clarke, A.E., Anderson, R.L., & Stone, B.A. (1979). Form and function of arabinogalactans and arabinogalactan-proteins. Phytochemistry, 18, 521-540. Cocuron, J.C., Lerouxel, O., Drakakaki, G., Alonso, A.P., Liepman, A.H., Keegstra, K.,  169 Raikhel, N., & Wilkerson, C.G. (2007). A gene from the cellulose synthase-like C family encodes a beta-1,4 glucan synthase. Proc. Natl. Acad. Sci. USA, 104(20), 8550-5. Cosgrove, D. J. (2005). Growth of the plant cell wall. Nat. Rev. Mol. Cell Biol., 6(11), 850-61. Cumming, C.M., Rizkallah, H.D., McKendrick, K.A., Abdel-Massih, R.M., Baydoun, E.A., & Brett, C.T. (2005). Biosynthesis and cell-wall deposition of a pectin- xyloglucan complex in pea. Planta, 222(3), 546-55. Dagel, D.J., Liu, Y.-S., Zhong, L., Luo, Y., Himmel, M.E., Xu, Q., Zeng, Y., Ding, S.-Y., & Smith, S. (2011). In situ imaging of single carbohydrate-binding modules on cellulose microfibrils. J. Phys. Chem. B., 115, 635–641. Daras, G., Rigas, S., Penning, B., Milioni, D., McCann, M.C., Carpita, N.C., Fasseas, C., & Hatzopoulos, P. (2009). The thanatos mutation in Arabidopsis thaliana cellulose synthase 3 (AtCesA3) has a dominant-negative effect on cellulose synthesis and plant growth. New Phytol., 184(1), 114-26. Dean, G., Cao, Y., Xiang, D., Provart, N. J., Ramsay, L., Ahad, A., White, R., Selvaraj, G. Datla, R., & Haughn, G. (2011). Analysis of gene expression patterns during seed coat development in Arabidopsis. Molecular Plant, 4(6), 1074-91. Dean, G.H., Zheng, H., Tewari, J., Huang, J., Young, D.S., Hwang, Y.T., Western, T.L., & Haughn, G. (2007). The Arabidopsis MUM2 gene encodes a beta-galactosidase required for the production of seed coat mucilage with correct hydration properties. Plant Cell, 19(12), 4007-21. Debeaujon, I. (2003). Proanthocyanidin-accumlating cells in Arabidopsis testa: regulation of differentiation and role in seed development. Plant Cell, 15, 2514-2531. Delmer, D.P., & Armor, Y. (1995). Cellulose biosynthesis. Plant Cell, 7(7), 987-1000.  170 Delmer, D.P. (1999). Cellulose biosynthesis: exciting times for a difficult field of study. Annu. Rev. Plant Physiol. Plant Mol. Biol., 50, 245-276. Desprez, T., Juraniec, M., Crowell, E. F., Jouy, H., Pochylova, Z., Parcy, F., Höfte, H., Gonneau, M., & Vernhettes, S. (2007). Organization of cellulose synthase complexes involved in primary cell wall synthesis in Arabidopsis thaliana. Proc. Natl. Acad. Sci. USA, 104(39), 15572-7. Desprez, T., Vernhettes, S., Fagard, M., Refre, G., Desnos, T., Py, N., Pelletier, S., & Höfte, H. (2002). Resistance against herbicide isoxaben and cellulose deficiency caused by distinct mutations in same cellulose synthase isoform CESA6. Plant Physiol., 128(2), 482-490. Dick-Perez, M., Zhang, Y., Hayes, J., Salazar, A., Zabotina, O.,A., & Hong, M. (2011). Structure and interactions of plant cell-wall polysaccharides by two- and three- dimensional magic angle solid-state NMR. Biochemistry, 50(6), 989-1000. Dixon, R.A., (2005). Proanthocyanidins – a final frontier in flavonoid research? New Phytol. 165, 9-28. Doblin, M.S., Kurek, I., Jacob-Wilk, D., & Delmer, D.P. (2002). Cellulose biosynthesis in plants: from genes to rosettes. Plant Cell Physiol. 43(12), 1407-20. Driouich, A., Follet-Gueye, M.L., Bernard, S., Kousar, S., Chavalier, L., Vicre-Gibouin, M., & Lerouxel, O. (2012). Golgi-mediated synthesis and secretion of matrix polysaccharides of the primary cell wall of higher plants. Front. Plant Sci., 3, 79. Elkins, T., Hortsch, M., Bieber, A.J., Snow, P.M., & Goodman, C.S. (1990) Drosophila fasciclin I is a novel homophilic adhesion molecule that along with fasciclin III can mediate cell sorting. J. Cell Biol., 110, 1825–1832. Ellis, M., Egelund, J., Schultz, C.J., & Bacic, A. (2010). Arabinogalactan-proteins: key regulators at the cell surface? Plant Physiol., 153(2), 403-19.  171 Esfandiari, E., Jin, Z., Abdeen, A., Griffiths, J.S., Western, T.L., & Haughn, G.W. (2013). Identification and analysis of an outer-seed-coat-specific promoter from Arabidopsis thaliana. Plant Mol. Biol., 81, 93-104. Fagard, M., Desnos, T., Desprez, T., Goubet, F., Refregier, G., Mouille, G., McCann, M., Rayon, C., Vernhettes, S., & Höfte, H. (2000). PROCUSTE1 encodes a cellulose synthase required for normal cell elongation specifically in roots and dark-grown hypocotyls of Arabidopsis. Plant Cell, 12(12), 2409-2424. Faik, A., Price, N.J., Raikhel, N.V., & Keegstra, K. (2002). An Arabidopsis gene encoding an alpha-xylosyltransferase involved in xyloglucan biosynthesis Proc. Natl. Acad. Sci. USA, 99(11), 7797-802. Fry, S.C., Smith, R.C., Renwick, K.F., Martin, D.J., Hodge, S.K., & Matthews, K.J. (1992). Xyloglucan endotransglucosylase, a new wall-loosening enzyme activity from plants. Biochem. J., 282, 821-8. Fry, S.C. (2000). The Growing Plant Cell Wall: Chemical and Metabolic Analysis. (Caldwell, NJ: The Blackburn Press). Fry, S.C. (1982). Phenolic components of the primary cell wall: feruloylated disaccharide of D-galactose and L-arabinose from spinach polysaccharide. Biochemical Journal, 203, 493–504. Fujino, T., Sone, Y., Mitsuishi, Y., & Itoh, T. (2000). Characterization of cross-links between cellulose microfibrils, and their occurrence during elongation growth in pea epicotyl. Plant Cell Physiol., 41(4), 11-5. Fujita, M., Himmelspach, R., Hocart, C.H., Williamson, R.E., Mansfield, S.D., & Wasteneys, G.O. (2011). Cortical microtubules optimize cell-wall crystallinity to drive unidirectional growth in Arabidopsis. Plant J, 66(6), 915-28. Gaspar, Y.M., Johnson, K.L., McKenna, J.A., Bacic, A., & Schultz, C.J. (2001). The  172 complex structures of arabinogalactan-proteins and the journey towards a function. Plant Mol. Biol., 47, 161–176. Girault, R., His, I., Andeme-Onzighi, C., Driouich, A., & Morvan, C. (2000). Identification and partial characterization of proteins and proteoglycans encrusting the secondary cell walls of flax fibers. Planta, 211, 256-264. Goldberg, R., Morvan, C., Jauneau, A., & Jarvis, M.C. (1996). Methyl-esterification, de- esterification and gelation of pectins in the primary cell wall. In Pectins and Pectinases Proceedings of an International Symposium (Elsevier), 151–172. Gonzalez, A., Mendenhall, J., Huo, Y., & Lloyd, A. (2009). TTG1 complex MYBs, MYB5 and TT2 control outer seed coat differentiation. Developmental biology, 325(2), 412-421. Goubet, F., & Mohnen, D. (1999). Subcellular localization and topology of homogalacturonan methyltransferase in suspension-cultured Nicotiana tabacum cells. Planta, 209(1), 112-117. Green, P.B. (1962). Mechanism for plant cellular morphogenesis. Science, 138(3548), 1404-5. Gu, Y., Kaplinsky, N., Bringmann, M., Cobb, A., Carroll, A., Sampathkumar, A., Baskin, T.I., Persson, S., & Somerville, C.R. (2010). Identification of a cellulose synthase-associated protein required for cellulose biosynthesis. Proc. Natl. Acad. Sci. USA 107, 12866-71. Gutierrez, R., Lindeboom, J. J., Paredez, A. R., Emons, A. M. C., & Ehrhardt, D. W. (2009). Arabidopsis cortical microtubules position cellulose synthase delivery to the plasma membrane and interact with cellulose synthase trafficking compartments. Nature Cell Biology, 11(7), 797-806. Hayashi, T. (1989). Xyloglucans in the primary cell wall. Ann. Rev. Plant Physiol. Plant  173 Mol. Biol. 40, 139-168. Hayashi, T., & Kaida, R. (2011). Functions of xyloglucan in plant cells. Mol. Plant, 4(1), 17-24. Harholt, J., Suttangkakul, A., & Scheller, H.V. (2010). Biosynthesis of pectin. Plant Physiol., 153, 384–395. Harpaz-Saad, S., McFarlane, H.E., Xu, S., Divi, U.K., Forward, B., Western, T.L., & Kieber, J.J. (2011). Cellulose synthesis via the FEI2 RLK/SOS5 pathway and cellulose synthase 5 is required for the structure of seed coat mucilage in Arabidopsis. Plant J., 68, 941–953. Harris, D.M., Corbin, K., Want, T., Gutierrez, R., Bertolo, A.L., Petti, C., Smilgies, D.M., Estevez, J.M., Bonetta, D., Urbanowicz, B.R., Ehrhardt, D.W., Somerville, C.R., Rose, J.K., Hong, M., & Debolt, S. (2012). Cellulose microfibril crystallinity is reduced by mutating C-terminal transmembrane region residues CESA1A903V and CESA3T942I of cellulose synthase. Proc. Natl. Acad. Sci. USA, 109(11), 4098-103. Haughn, G., & Chaudhury, A. (2005). Genetic analysis of seed coat development in Arabidopsis. Trends Plant Sci., 10(10), 472-7. Haughn, G.W., & Western, T.L. (2012). Arabidopsis seed coat mucilage is a specialized cell wall that can be used as a model for genetic analysis of plant cell wall structure and function. Front. Plant Sci., 3, 64. Heim, D.R., Skomp, J.R., Tschabold, E.E., & Larrinua, I.M. (1990). Isoxaben inhibits the synthesis of acid insoluble cell wall materials in Arabidopsis thaliana. Plant Physiol., 93(2), 695-700. Held, M.A., Tan, L., Kamyab, A., Hare, M., Shpak, E., and Kieliszewski, M.J. (2004). Di-isodityrosine is the intermolecular cross-link of isodityrosine-rich extensin  174 analogs cross-linked in vitro. J. Biol. Chem. 279, 55474–55482. Herth, W., & Schneph, E. (1980). The fluorochrome, calcofluor white, binds oriented to structural polysaccharide fibrils. Protoplasma, 105(1-2), 129-133. Hruz, T., Laule, O., Szabo, G., Wessendorp, F., Bleuler, S., Oertle, L., Widmayer, P., Gruissem, W., & Zimmermann, P. (2008). Genevestigator v3: a reference expression database for the meta-analysis of transcriptomes. Adv. Bioinformatics 2008, 420747. Huang, J., DeBowles, D., Esfandiari, E., Dean, G., Carpita, N. C., & Haughn, G.W. (2011). The Arabidopsis transcription factor LUH/MUM1 is required for extrusion of seed coat mucilage. Plant Physiol., 156(2), 491-502. Huber, O., & Sumper, M. (1994). Algal-CAMs: isoforms of a cell adhesion molecule in embryos of the alga Volvox with homology to Drosophila fasciclin I. EMBO J 13, 4212–4222. Ichinose, H., Kuno, A., Kotake, T., Yoshida, M., Sakka, K., Hirabayashi, J., Tsumuraya, Y., & Kaneko, S. (2006). Characterization of an exo-beta-1,3-galactanase from Clostridium Thermocellum. Appl Environ Microbiol, 72(5), 3515-23. Immerzeel, P., Eppink, M.M., de Vries, S.C., Schols, H.A., & Voragen, A.G.J. (2006). Carrot arabinogalactan proteins are interlinked with pectins. Physiol. Plant, 128, 18–28. Ishida, T., Hattori, S., Sano, R., Inoue, K., Shirano, Y., Hayashi, H., Shibata, D., Sato, S., Kato, T., Tabata, S., Okada, K., & Wada, T. (2007). Arabidopsis TRANSPARENT TESTA GLABRA2 is directly regulated by R2R3 MYB transcription factors and is involved in regulation of GLABRA2 transcription in epidermal differentiation. Plant Cell, 19(8), 2531-43. Ishii, T. (1997). O-acetylated oligosaccharides from pectins of potato tuber cell walls.  175 Plant Physiol., 113(4), 1265-72. Iwai, H., Ishii, T., & Satoh, S. (2001). Absence of arabinan in the side chains of the pectic polysaccharides strongly associated with cell walls of Nicotiana plumbaginifolia non-organogenic callus with loosely attached constituent cells. Planta, 213, 907– 15. Jamet, E., Canut, H., Boudart, G., & Pont-Ezica, R.F. (2006). Cell wall proteins: a new insight through proteomics. Trends Plant Sci. 11(1), 33-9. Jarvis, M. (2003). Chemistry: cellulose stacks up. Nature, 426(6967), 611-2. Jauh, G.Y., & Lord, E.M. (1996). Localization of pectins and arabinogalactan-proteins in lily (Lilium longiflorum L.) pollen tube and style, and their possible roles in pollination. Planta, 199, 251–261. Jofuku, K.D., den Boer, B.G., Van Montagu, M., & Okamuro, J.K. (1994). Control of Arabidopsis flower and seed development by the homeotic gene APETALA2. Plant Cell, 6, 1211–1225. Johnson, K.L., Jones, B.J., Bacic, A., & Schultz, C.J. (2003). The Fasciclin-like arabinogalactan proteins of Arabidopsis. A multigene family of putative cell adhesion molecules. Plant Physiol., 133, 1911-1925. Johnson, K.L., Kibble, N.A., Bacic, A., & Schultz, C.J. (2011). A fasciclin-like arabinogalactan-protein (FLA) mutant of Arabidopsis thaliana, fla1, shows defects in shoot regeneration. PLoS One, 6(9), e25154. Johnson, C.S., Kolevski, B., & Smyth, D.R. (2002). TRANSPARENT TESTA GLABRA2, a trichome and seed coat development gene of Arabidopsis, encodes a WRKY transcription factor. Plant Cell, 14(6), 1359-75. Jolie, R.P., Duvetter, T., Van Loey, A.M., & Hendrickx, M.E. (2010). Pectin methylesterase and its proteinaceous inhibitor: a review. Carbohydr Res. 345(18),  176 2583-95. Jones, L., Milne, J.L., Ashford, D., & McQueen-Mason, S.J. (2003). Cell wall arabinan is essential for guard cell function. Proc. Natl. Acad. Sci. USA, 100(20), 11783-8. Kawamoto, T., Noshiro, M., Shen, M., Nakamasu, K., Hashimoto, K., Kawashima-Ohya, Y., Gotoh, O., & Kato, Y. (1998). Structural and phylogenetic analyses of RGD- CAP/beta ig-h3, a fasciclin-like adhesion protein expressed in chick chondrocytes. Biochimica et Biophysica Acta 1395, 288–292. Keegstra K, Talmadge KW, Bauer WD, & Albersheim P. (1973). The structure of plant cell walls. III. A model of the walls of suspension-cultured sycamore cells based on the interconnections of the macromolecular components. Plant Phys., 51, 188– 196. Kim, J.-E., Kim, S.-J., Lee, B.-H., Park, R.-W., Kim, K.-S., & Kim, I.-S. (2000). Identification of motifs for cell adhesion within the repeated domains of transforming growth factor--induced gene, big-h3. J. Biol. Chem., 275, 30907– 30915. Kimura, S., Laosinchai, W., Itoh, T., Cui, X., Linder, C.R., & Brown, R.M. (1999). Immunogold labeling of rosette terminal cellulose-synthesizing complexes in the vascular plant Vigna angularis. Plant Cell, 11(11), 2075-86. Kitazawa, K., Tryfona, T., Yoshimi, Y., Hayashi, Y., Kawauchi, S., Antonov, L., Tanaka, H., Takahashi, T., Kaneko, S., Dupree, P., Tsumuraya, Y., & Kotake, T. (2013). !-Galactosyl Yariv reagent binds to the "-1,3-galactan of arabinogalactan- proteins. Plant Physiol., Jan 7, Epub. Knox, J.P. (1997). The use of antibodies to study the architecture and developmental regulation of plant cell walls. Int. Rev. Cytol., 171, 79–120. Knox, J.P., Linstead, P.J., King, J., Cooper, C., & K. Roberts. (1990). Pectin  177 esterification is spatially regulated both within cell walls and between developing tissues of root apices. Planta, 181, 512-521. Koornneef, M. (1981). The complex syndrome of TTG mutants. Arabid. Inf. Serv., 18, 45–51. Kotake, T., Dina, S., Konishi, T., Kaneko, S., Igarashi, K., Samejima, M., Watanabe, Y., Kimura, K., & Tsummuraya, Y. (2005). Molecular cloning of a (beta)- galactosidase from radish that specifically hydrolyzes (beta)-(1-3)- and (beta)-(1- 6)- galactosyl residues of Arabinogalactan protein. Plant Physiol, 138(3), 1563- 76. Kurek, I., Kawagoe, Y., Jacob-Wilk, D., Doblin, M., & Delmer, D. (2002). Dimerization of cotton fiber cellulose synthase catalytic subunits occurs via oxidation of the zinc-finger domains. Proc. Natl. Acad. Sci. USA, 99(17), 11109-14. Labavitch, J.M., & Ray, P.M. (1974). Turnover of cell wall polysaccharides in elongating pea stem segments. Plant Physiol., 53(5), 669-73. Lamport, D.T.A., & Kieliszewski, M.J. (2005). Stress upregulates periplasmic arabinogalactanproteins. Plant Biosyst., 139, 60–64 Lamport, D.T., Kieliszewski, M.J., & Showalter, A.M. (2006). Salt-stress upregulates arabinogalactan-proteins: using salt-stress to analyse AGP function. New Phytol., 169, 479–492. Lamport, D.T., & Varnai, P. (2013). Periplasmic arabinogalactan glycoproteins act as a calcium capacitor that regulates plant growth and development. New Phytol., 197(1), 58-64. Lane, D.R., Wiedemeier, A., Peng, L., Hofte, H., Vernhettes, S., Desprez, T., Hocart, C.H., Birch, R.J., Baskin, T.I., Burn. J.E., Arioli, T., Betzner, A.S., & Williamson, R.E. (2001). Temperature-sensitive alleles of RSW2 link the KORRIGAN endo-  178 1,4-beta-glucanase to cellulose synthesis and cytokinesis in Arabidopsis. Plant Physiol., 126(1), 278-88. Le, B.H., Cheng, C., Bui, A.Q., Wagmaister, J.A., Henry, K.F., Pelletier, J., Kwong, L., Belmonte, M., Kirkbride, R., Horvath, S., Drews, G.N., Fischer, R.L., Okamuro, J.K., Harada, J.J. & Goldberg, R.B. (2010). Global analysis of gene activity during Arabidopsis seed development and identification of seed-specific transcription factors. Proc. Natl. Acad. Sci. USA, 107(18), 8063-70. Ledbetter, M.C., & Porter, K.R. (1963). A “microtubule” in plant cell fine structure. J. Cell Biol., 19(1), 239-50. Leonard, R., Strasser, R., & Altmann, F. (2009). Plant glycosidases acting on protein- linked oligosaccharides. Phytochemistry, 70, 318-324. Levy, S., York, W.S., Stuikeprill, R., Meyer, B., & Staehelin, L.A. (1991). Xyloglucan- the role of the fucosylated side-chain in surface-specific side-chain folding. Plant J., 1, 195-215. Limberg, G., Korner, R., Buchholt, H.C., Christensen, T.M., Roepstorff, P., & Mikkelsen, J.D. (2000). Analysis of different de-esterification mechanisms for pectin by enzymatic fingerprinting using endopectin lyase and endopolygalacturonase II from A. niger. Carbohydr. Res., 327(3), 293-307.  Li, Y.-Q., Faleri, C., Geitmann, A., Zhang, H.Q., & Cresti, M. (1995). Immunogold localization of arabinogalactan proteins, unesterified and esterified pectins in pollen grains and pollen tubes of Nicotiana tabacum L. Protoplasma, 189, 26–36 Li, J., Yu, M., Geng, L.-L., & Zhao, J. (2010). The fasciclin-like arabinogalactan protein gene, FLA3, is involved in microspore development of Arabidopsis. Plant J., 64, 482-497. Li, S.F., Milliken, O.N., Phan, H., Seyit, R., Napoli, R., Preston, J., Koltunow, A.M., &  179 Parish, R.W. (2009). The Arabidopsis MYB5 transcription factor regulates mucilage synthesis, seed coat development, and trichome morphogenesis. Plant Cell, 21(1), 72-89. Li, S., Lei, L., Somerville, C.R., & Gu, Y. (2011). Cellulose synthase interactive protein 1 (CSI1) links microtubules and cellulose synthase complexes. Proc. Natl. Acad. Sci. USA, 109(1), 185-190. MacMillan, C.P., Mansfield, S.D., Stachurski, Z.H., Evans, R. & Southerton, S.G. (2010). Fasciclin-like arabinogalactan proteins: specialization for stem biomechanics and cell wall architecture in Arabidopsis and Eucalyptus. Plant J., 62(4), 689-703. Macquet, A, Ralet, M.-C., Kronenberger, J., Marion-Poll, A., & North, H.M. (2007a). In situ, chemical and macromolecular study of the composition of Arabidopsis thaliana seed coat mucilage. Plant Cell Physiol., 48(7), 984-99. Macquet, A., Ralet, M.-C., Loudet, O., Kronenberger, J., Mouille, G., Marion-Poll, A., & North, H.M. (2007b). A naturally occurring mutation in an Arabidopsis accession affects a beta-D-galactosidase that increases the hydrophilic potential of rhamnogalacturonan I in seed mucilage. Plant Cell, 19, 3990–4006. Martin, I., Dopico, B., Munoz, F.J., Esteban, R., Oomen, R.J., Driouich, A., Vincken, J.P., Visser, R., & Labrador, E. (2005). In vivo expression of a Cicer arietinum beta-galactosidase in potato tubers leads to a reduction of the galactans side- chains in cell wall pectin. Plant Cell Physiol., 46(10), 1613-22. McCartney, L., & Knox, J.P. (2002). Regulation of pectic polysaccharide domains in relation to cell development and cell properties in the pea testa. J. Exp. Bot. 53, 707–713. McCartney, L., Steele-King, C.G., Jordan, E., & Knox, J.P. (2003). Cell wall pectic (1- 4)- !-D-galactan marks the acceleration of cell elongation in the Arabidopsis  180 seedling root meristem. Plant J., 33, 447-54. McCann, M.C., & Roberts, K. (1991). Architecture of the primary cell wall. In The Cytoskeletal Basis of Plant Growth and Form, ed. CW Lloyd, pp. 109–29. New York: Academic McCann, M., & Rose, J. (2010). Blueprints for building plant cell walls. Plant Physiol., 153(2), 365. McFarlane, H.E., Young, R.E., Wasteneys, G.O., & Samuels, A.L. (2008). Cortical microtubules mark the mucilage secretion domain of the plasma membrane in Arabidopsis seed coat cells. Planta, 227(6), 1363-75. McNeil, M., Darvill, A.G., Fry, S.C. & Albersheim, P. (1984) Structure and function of the primary cell walls of plants. Annu. Rev. Biochem.53, 625-663. McQueen-Mason, S.J., & Cosgrove, D.J. (1995). Expansin mode of action on cell walls. Analysis of hydrolysis, stress relaxation and binding. Plant Physiol., 107(1), 87- 100. Mendu, V, Griffiths, J.S., Persson, S., Stork, J., Downie, A.B., Voiniciuc, C., Haughn, G.W., & DeBolt, S. (2011). Subfunctionalization of cellulose synthases in seed coat epidermal cells mediates secondary radial wall synthesis and mucilage attachment. Plant Physiol., 157(1), 441-53. Micheli, F. (2001). Pectin methylesterases: cell wall enzymes with important roles in plant physiology. Trends Plant Sci., 6, 414-9. Mohnen, D. (2008). Pectin structure and biosynthesis. Curr. Opin. Plant Biol., 11(3), 266- 77. Mollet, J.C., Kim, S., Jauh, G.Y., & Lord, E.M. (2002). Arabinogalactan proteins, pollen tube growth, and the reversible effects of Yariv phenylglycoside. Protoplasma.  181 219(1-2), 89-98. Mouille, G., Ralet, M.C., Cavelier, C., Eland, C., Effroy, D., Hematy, K., McCartney, L., Truong, H.N., Gaudon, B., Thibault, J.F., Marchant, A., & Hofte, H. (2007). Homogalacturonan synthesis in Arabidopsis thaliana requires a Golgi-localized protein with a putative methyltransferase domain. Plant J., 50(4), 605-14. Mueller, S.C., & Brown, R.M. Jr. (1980). Evidence for an intramembrane component associated with a cellulose microfibril synthesizing complex in higher plants. J. Cell Biol., 84(2), 315-26. Nishitani, K., & Tominaga, R. (1992). Endo-xyloglucan transferase, a novel class of glycosyltransferase that catalyzes transfer of a segment of xyloglucan molecule to another xyloglucan molecule. J. Biol. Chem., 267, 21058-64. Nishiyama Y, Langan P, & Chanzy H. (2002). Crystal structure and hydrogen-bonding system in cellulose 1! from synchrotron X-ray and neutron fiber diffraction. J. Am. Chem. Soc. 124, 9074–82 Nishiyama Y, Sugiyama J, Chanzy H, & Langan P. (2003). Crystal structure and hydrogen bondingsystem in cellulose 1" from synchrotron X-ray and neutron fiber diffraction. J. Am. Chem. Soc., 125, 14300–6. Nothnagel, E.A. (1997). Proteoglycans and related components in plant cells. Int. Rev. Cytol. 174, 195-91. Oechslin, R., Lutz, M.V., & Amado, R. (2003). Pectic substances isolated from apple cellulosic residue: structural characterization of a new type of rhamnogalacturonan I. Carbohydr. Polym., 51, 301–310. Oka, T., Nemoto, T., & Jigami, Y. (2007). Functional analysis of Arabidopsis thaliana RHM2/MUM4, a multidomain protein involved in UDP-D-glucose to UDP-L- rhamnose conversion. J. Biol. Chem., 282(8), 5389-403.  182 Orfila, C., Seymour, G.B., Willats, W.G., Huxham, I.M., Jarvis, M.C., Dover, C.J., Thompson, A.J., & Knox, J.P. (2001). Altered middle lamella Homogalacturonan and disrupted deposition of (1!5)-alpha-L-arabinan in the pericarp of Cnr, a ripening mutant of tomato. Plant Physiol., 126(1), 210-21. Paradez, A.R., Somerville, C.R., & Ehrhardt, D.W. (2006). Visualization of cellulose synthase demonstrates functional association with microtubules. Science, 312, 1491-1495. Pattahil, S., Avi, U., Baldwin, D., Swennes, A.G., McGill, J.A., Popper, Z., Bootten, T., Albert, A., Davis, R.H., Chennaredd, C., Dong, R., O’Shea, B., Rossi, R., Leoff, C., Freshour, G., Narra, R., O’Neil, M., York, W.S., & Hahn, M.G. (2010). A comprehensive toolkit of plant cell wall glycan-directed monoclonal antibodies. Plant Physiol., 153(2), 514-25. Pauly, M., Albersheim, P., Darvill, A., & York, W.S. (1999). Molecular domains of the cellulose/xyloglucan network in the cell walls of higher plants. Plant J. 20, 629– 39 Pear, J.R., Kawagoe, Y., Schreckengost, W.E., Delmer, D.P., & Stalker, D.M. (1996). Higher plants contain homologs of the bacterial celA genes encoding the catalytic subunit of cellulose synthase. Proc. Natl. Acad. Sci. USA, 93(22), 12637-42. Peaucelle, A., Bradybook, S.A, & Hofte, H. (2012). Cell wall mechanics and growth control in plants: the role of pectins revisited. Front. Plant Sci., 3, 121. Penfield, S., Meissner, R.C., Shoue, D.A., Carpita, N.C., & Bevan, M.W. (2001). MYB61 is required for mucilage deposition and extrusion in the Arabidopsis seed coat. Plant Cell, 13, 2777-2791. Persson, S., Paredez, A., Carroll, A., Palsdottir, H., Doblin, M., Poindexter, P., Khitrov, N., Auer, M., & Somerville, C.R. (2007). Genetic evidence for three unique components in primary cell-wall cellulose synthase complexes in Arabidopsis.  183 Proc. Natl. Acad. Sci. USA, 104(39), 15566-71. Popper, Z.A., & Fry, S.C. (2005). Widespread occurrence of a covalent linkage between xyloglucan and acidic polysaccharides in suspension-cultured angiosperm cells. Annals of Botany, 96, 91-99. Popper, Z.A., & Fry, S.C. (2008). Xyloglucan-pectin linkages are formed intra- protoplasmically, contribute to wall assembly, and remain stable in the cell wall. Planta, 277(4), 781-94. Pressey, R. (1983). Beta-Galactosidase in ripening tomatoes. Plant Physiol., 71(1), 132-5. Rautengarten, C., Usadel, B., Neumetzler, L., Hartmann, J., Bussis, D., & Altmann, T. (2008). A subtilisin-like serine protease essential for mucilage release from Arabidopsis seed coats. Plant J., 54(3), 466-480. Rerie, W.G., Feldmann, K.A., & Marks, M.D. (1994). The GLABRA2 gene encodes a homeo domain protein required for normal trichome development in Arabidopsis. Genes Dev, 8(12), 1388-99. Richmond, T.A., & Somerville, C.R. (2000). The cellulose synthase superfamily. Plant Physiol., 124, 495–498. Ridley, B. L., Neill, M. A. O., & Mohnen, D. (2001). Pectins: structure, biosynthesis, and oligogalacturonide-related signaling. Phytochemistry, 57(6), 929-67. Romano, J.M., Dubos, C., Prouse, M.B., Wilkins, O., Hong, H., Poole, M., Kang, K.-Y., Li, E., Douglas, C.J., Western, T.L., Mansfield, S.D., & Campbell, M.M. (2012). AtMYB61, an R2R3-MYB transcription factor, functions as a pleiotropic regulator via a small gene network. New Phyt., 195(4), 774-786. Roudier, F., Fernandez, A.G., Fujita, M., Himmelspach, R., Borner, G.H., Schindelman, G., Song, S., Baskin, T.I., Dupree, P., Wasteneys, G.O., & Benfey, P.N. (2005). COBRA, an Arabidopsis extracellular glycosyl-phosphatidyl inositol-anchored  184 protein, specifically controls highly anisotropic expansion through its involvement in cellulose microfibril orientation. Plant Cell, 17(6), 1749-63. Rounds, C.M., Lubeck, E., Hepler, P.K., & Winship, L.J. (2011). Propidium iodide competes with Ca(2+) to label pectin in pollen tubes and Arabidopsis root hairs. Plant Physiol,. 157(1), 175-87. Samson, M.R. Jongeneel, R., & Klis F.M. (1984). Arabinogalactan protein in the extracellular space of Phaseolus vulgaris hypocotyls. Phytochemistry, 23, 493– 496. Scheible, W.R., Eshed, R., Richmond, T., Delmer, D., & Somerville, C. (2001). Modifications of cellulose synthase confer resistance to isoxaben and thiazolidininone herbicides in Arabidopsis ixr1 mutants. Proc. Natl. Acad. Sci. USA, 98(18), 10079-84. Scheller, H.V., & Ulvskov, P. (2010). Hemicelluloses. Annu. Rev. Plant Biol., 61, 263- 89. Schultz, C.J., Ferguson, K.L., Lahnstein, J., & Bacic, A. (2004). Post-translational modifications of arabinogalactan-peptides of Arabidopsis thaliana. Endoplasmic reticulum and glycosylphosphatidylinositol-anchor signal cleavage sites and hydroxylation of proline. J.Biol. Chem., 279, 45503–11 Schultz, C.J., Rumsewicz, M.P., Johnson, K.L., Jones, B.J., Gaspar, Y.M., & Bacic, A. (2002). Using genomics resources to guide research directions: the arabinogalactan protein gene family as a test case. Plant Physiol., 129, 1448–63 Schultz, C.J., Johnson, K., Currie, G., & Bacic, A. (2000). The classical arabinogalactan protein gene family of Arabidopsis. Plant Cell 12, 1–18 Seifert, G.J., and Roberts, K. (2007). The biology of arabinogalactan proteins. Annu. Rev. Plant Biol., 58, 137–61.  185 Sekimata, M., Ogura, K., Tsumuraya, Y., Hashimoto, Y., & Yamamoto, S. (1989). A beta-Galactosidase from Radish (Raphanus sativus, L.) seeds. Plant Physiol., 567- 74. Selvendran, R.R., & Ryden, P. (1990). Isolation and analysis of plant cell walls. In Methods in Plant Biochemistry, Vol. 2: Carbohydrates, P.M. Dey and J.B. Harbourne, eds (San Diego, CA: Academic Press), pp. 549–575. Serpe, M.D., & Nothnagel, E.A. (1999). Arabinogalactan-proteins in the Multiple Domains of the Plant Cell Surface. Adv. in Bot. Res., 30, 207-289. Showalter, A.M. (2001). Arabinogalactan-proteins: structure, expression and function. Cell Mol Life Sci. 58(10), 1399-417. Shi, H., Kim, Y., Guo, Y., Stevenson, B., & Zhu, J. (2003). The Arabidopsis SOS5 locus encodes a putative cell surface adhesion protein and is required for normal cell expansion. Plant Cell, 15(1), 19–32. Somerville, C. (2006). Cellulose synthesis in higher plants. Ann. Rev. Cell and Dev. Biol., 22, 53-78. Spurr, A.R. (1969). A low-viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res., 26, 31–43. Sullivan, S., Ralet, M.-C., Berger, A., Diatloff, E., Bischoff, V., Gonneau, M., Marion- Poll, A., & North, H.M. (2011). CESA5 Is Required for the Synthesis of Cellulose with a Role in Structuring the Adherent Mucilage of Arabidopsis Seeds. Plant Physiol., 156(4), 1725-39. Stafstrom, J.P., & Staehelin, L.A. (1986). Cross-linking patterns in salt-extractable extension from carrot cell walls. Plant Physiol., 81, 234–241. Sterling, C. (1970). Crystal-structure of ruthenium red and stereochemistry of its pectic stain. Am. J. Bot. 57, 172–175.  186 Sterling, J. D., Atmodjo, M. A., Inwood, S. E., Kumar Kolli, V. S., Quigley, H. F., Hahn, M. G., & Mohnen, D. (2006). Functional identification of an Arabidopsis pectin biosynthetic homogalacturonan galacturonosyltransferase. Proc. Natl. Acad. Sci. USA, 103(13), 5236-41. Sterling, J.D., Quigley, H.F., Orellana, A., & Mohnen, D. (2001). The catalytic site of the pectin biosynthetic enzyme alfpha-1,4-galacturonosyltransferase is located in the lumen of the Golgi. Plant Physiol, 127(1), 360-71. Sticklen, M.B. (2008). Plant genetic engineering for biofuel production: towards affordable cellulosic ethanol. Nat. Rev. Genetics, 9(6), 433-43. Stork, J., Harris, D., Griffiths, J., Williams, B., Beisson, F., Li-Beisson, Y., Mendu, V., Haughn, G., & DeBolt, S. (2010). CELLULOSE SYNTHASE9 serves a nonredundant role in secondary cell wall synthesis in Arabidopsis epidermal testa cells. Plant Physiol., 153(2), 580-9. Swarbreck, D., Wilks, C., Lamesch, P., Berrardini, T.Z., Garcia-Harnandez, M., Foerster, H., Li, D., Meyer, T., Muller, T., Ploetz, L., Radenbaugh, A., Singh, S., Swing, V., Tissier, C., Zhang, P., & Huala, E. (2007). The Arabidopsis Information Resource (TAIR); gene function and annotation. Nucleic Acids Res.. 36, D1009- 14. Szyjanowicz, P.M., McKinnon, I., Taylor, N.G., Gardiner, J., Jarvis, M.C., & Turner, S.R. (2004). The irregular xylem 2 mutant is an allele of korrigan that affects the secondary cell wall of Arabidopsis thaliana. Plant J., 37(5), 730-40. Szymanski, D.B., Jilk, R.A., Pollock, S.M., & Marks, M.D. (1998). Control of GL2 expression in Arabidopsis leaves and trichomes. Development, 125(7), 1161-71. Takata, R., Tokita, K., Mori, S., Shimoda, R., Harada, N., Ichinose, H., Kaneko, S., Igarashi, K., Samjima, M., Tsumuraya, Y., & Kotake, T. (2010). Degradation of carbohydrate moieties of arabinogalactan proteins by glycoside hydrolases from  187 Neurospora crassa. Carbohydr. Res., 245(17), 2516-22. Tan, L., Eberhard, S., Pattathil, S., Warder, C., Glushka, J., Yang, C., Hao, Z., Zhu, X., Avci, U., Miller, J.S., Baldwin, D., Pham, C., Orlando, R., Darvill, A., Hahn, M.G., Kieliszewski, M.J. & Mohnen, D. (2013). An Arabidopsis cell wall proteoglycan consists of pectin and arabinoxylan covalently linked to an Arabinogalactan protein. Plant Cell, Jan 31, Epub. Tan, L., Leykan. J.F., Kieliszewski, M.J. (2003). Glycosylation motifs that direct arabinogalactan addition to arabinogalactan proteins. Plant Physiol., 132, 1362- 69. Tan, L., Qiu, F., Lamport, D.T., & Kieliszewski, M.J. (2004). Structure of a hydroxyproline (Hyp)-arabinogalactan polysaccharide from repetitive Ala-Hyp expressed in transgenic Nicotiana tabacum. J. Biol. Chem., 279(13), 13156-65. Taylor, N.G., Howells, R.M., Huttly, A.K., Vickers, K., & Turner, S.R. (2003). Interactions among three distinct CesA proteins essential for cellulose synthesis. Proc. Natl. Acad. Sci. USA, 100(3), 1450-5. Taylor, N. G., Laurie, S., & Turner, S. R. (2000). Multiple cellulose synthase catalytic subunits are required for cellulose synthesis in Arabidopsis. Plant Cell, 12(12), 2529-2540. Thimm, J.C., Burritt, D.J., Ducker, W.A., & L.D. Melton. (2009). Pectins influence microfibril aggregation in celery cell walls: An atomic force microscopy study. J Structural Biol. 168, 337-44. Thompson, J.E., & Fry, S.C. (2000). Evidence for covalent linkage between xyloglucan and acidic pectins in suspension cultured rose cells. Planta, 211(2), 275-86. Tominaga-Wada, R., Iwata, M., Sugiyama, J., Kotake, T., ishida, T., Yokoyama, R., Nishitani, K., Okada K., & Wada, T. (2009). The GLABRA2 homeodomain  188 protein directly regulates CESA5 and XTH17 gene expression in Arabidopsis roots. Plant J., 60(3), 564-74. Updegraff, D.M. (1969). Semimicro determination of cellulose in biological materials. Anal. Biochem., 3, 420–424. Usadel, B., Kuschinsky, A.M., Rosso, M.G., Eckermann, N., & Pauly, M. (2004). RHM2 is involved in mucilage pectin synthesis and is required for the development of the seed coat in Arabidopsis. Plant Physiol., 134(1), 286-95. VandenBosch, K.A., Bradley, D.J., Knox, J.P., Perotto, S., Butcher, G.W., & Brewin, N.J. (1989). Common components of the infection thread matrix and the intercellular space identified by immunocytochemical analysis of pea nodules and uninfected roots. EMBO J. 8, 335–341. Vincken, J.P., Schols, H.A., Oomen, R.J., McCann, M.C., Ulvskov, P., Voragen, A.G., & Visser, R.G. (2003). If Homogalacturonan were a side chain of rhamnogalacturonan I. Implications for cell wall architecture. Plant Physiol., 132(4), 1781-9. Vignon, M.R., Heux, L., Malainine, M.E., & Mahrouz, M. (2004). Arabinan cellulose composite in Opuntia ficus-indica prickly pear spines. Carbohydr. Res, 339, 123– 31. Walker, M., Tehseen, M., Doblin, M.S., Pettolino, F.A., Wilson, S.M., Bacic, A., & Golz, J.F. (2011). The transcriptional regulator LEUNIG_HOMOLOG regulates mucilage release from the Arabidopsis testa. Plant Physiol., 256(1), 46-60. Wang, T., Zabotina, O., & Hong, M. (2012). Pectin-cellulose interactions in the Arabidopsis primary cell wall from two-dimensional magic-angle-spinning solid- state nuclear magnetic resonance. Biochemistry, 51(49), 9846-56. Wasteneys, G.O., & Fujita, M. (2006). Establishing and maintaining axial growth: wall  189 mechanical properties and the cytoskeleton. J. Plant Res., 119, 5-10. Western, T.L., Burn, J., Tan, W.L., Skinner, D.J., Martin-McCaffrey, L., Moffatt, B.A., & Haughn, G.W. (2001). Isolation and Characterization of Mutants Defective in Seed Coat Mucilage Secretory Cell Development in Arabidopsis. Plant Physiol., 127, 998-1011. Western, T.L., Skinner, D.J., & Haughn, G.W. (2000). Differentiation of mucilage secretory cells of the Arabidopsis seed coat. Plant Physiol., 122(2), 345-56. Western, T. L., Young, D. S., Dean, G. H., Tan, W. L., Samuels, A. L., & Haughn, G. W. (2004). MUCILAGE-MODIFIED4 Encodes a Putative Pectin Biosynthetic Enzyme Developmentally Regulated by APETALA2 , TRANSPARENT TESTA GLABRA1, and GLABRA2 in the Arabidopsis Seed Coat. Plant Physiol., 134, 296-306. Willats, W.G.T., McCartney L., Mackie, W., & Knox, J.P. (2001a). Pectin: cell biology and prospects for functional analysis. Plant Mol. Biol., 47, 9–27 Willats, W.G.T., McCartney, L., & Knox, J.P. (2001b). In-situ analysis of pectic polysaccharides in seed mucilage and at the root surface of Arabidopsis thaliana. Planta, 213, 37–44. Willats, W.G.T., Orfila, C., Limberg, G., Buchholt, H.C., van Alebeek, G.-J.W.M., Voragen, A.G.J., Marcus, S.E., Christensen, T.M.I.E., Mikkelsen, J.D., Murray, B.S., & Knox, J.P. (2001c). Modulation of the Degree and Pattern of Methyl- esterification of Pectic Homogalacturonan in Plant Cell Walls. J. Biol. Chem., 276, 19404 –19413. Williamson, R.E., Burn, J.E., Birch, R., Baskin, T.I., Arioli, T., Betzner, A.S., & Cork, A. (2001). Morphology of rsw1, a cellulose-deficient mutant of Arabidopsis thaliana. Protoplasma, 215(1-4), 116-27.  190 Winter, D., Vinegar, B., Nahal, H., Ammar, R., Wilson G.V., & Provart, N.J. (2007). An “Electronic Fluorescent Pictograph” browser for exploring and analysing large- scale biological data sets. PLoS One, 2(8), e718. Wolf, S., Mouille, G., & Pelloux, J. (2009). Homogalacturonan methyl-esterification and plant development. Mol. Plant, 2, 851-60. Wu, H.M., Wang, H., & Cheung, A.Y. (1995). A pollen tube growth stimulatory glycoprotein is deglycosylated by pollen tubes and displays a glycosylation gradient in the flower. Cell, 82(3), 395-403. Xie, L., Yang, C., & Wang, X. (2011). Brassinosteroids can regulate cellulose biosynthesis by controlling the expression of CESA genes in Arabidopsis. J. Exp. Bot., 62(13), 4495-506. Yi, H., & Puri, V.M. (2012). Architecture-based multiscale computational modeling of plant cell wall mechanics to examine the hydrogen-bonding hypothesis of cell wall network structure model. Plant Physiol., 160(3), 1281-92. Yokoyama, R., & Nishitani, K. (2001). A comprehensive expression analysis of all members of a gene family encoding cell-wall enzymes allowed us to predict cis- regulatory regions involved in cell-wall construction in specific organs of Arabidopsis. Plant Cell Physiol., 42, 1025-33. Yoo, S.H., Fishman, M.L., Savary, B.J., & Hotchkiss, A.T. Jr. (2003). Monovalent salt- induced gelation of enzymatically deesterified pectin. J. Agric. Food Chem., 51(25), 7410-7. Youl, J.J., Bacic, A., & Oxley, D. (1998). Arabinogalactan-proteins from Nicotiana alata and Pyrus communis contain glycosylphosphatidylinositol membrane anchors. Proc. Natl. Acad. Sci. USA, 95, 7921-26. Young, R.E., McFarlane, H.E., Hahn, M.G., Western, T.L., Haughn, G.W., & Samuels,  191 A.L. (2008). Analysis of the Golgi apparatus in Arabidopsis seed coat cells during polarized secretion of pectin-rich mucilage. Plant Cell, 20(6), 1623-38. Zabotina, O.A., Avci, U., Cavalier, D., Pattathil, S., Chou, Y.H., Eberhard, S., Danhof, L., Keegstra, K., & Hahn, M.G. (2012). Mutations in multiple XXT genes of Arabidopsis reveal the complexity of xyloglucan biosynthesis. Plant Physiol., 159(4), 1367-84. Zimmermann, P., Hirsch-Hoffmann, M., Hennig, L., & Gruissem, W. (2004). GENEVESTIGATOR: Arabidopsis microarray database and analysis toolbox. Plant Physiol., 136, 2621–2632. Zykwinska, A., Gaillard, C., Buleon, A., Pontoire, B., Carnier, C., Thibault, J.-F., & Ralet, M.-C. (2007b). Assessment of in vitro binding of isolated pectic domains to cellulose by adsorption isotherms, electron microscopy, and X-ray diffraction methods. Biomacromolecules, 8(1), 223-32. Zykwinska, A.W., Ralet, M.-C., & Garnier, C. D. (2005). Evidence for in vitro binding of pectin side chains to cellulose. Plant Physiol., 139, 397-407. Zykwinska, A., Thibault, J.-F., & Ralet, M.-C. (2008). Competitive binding of pectin and xyloglucan with primary cell wall cellulose. Carbohydr. Polymers, 74(4), 957- 961. Zykwinska, A., Thibault, J.-F., & Ralet, M.-C. (2007a). Organization of pectic arabinan and galactan side chains in association with cellulose microfibrils in primary cell walls and related models envisaged. J. Exp. Bot., 58(7), 1795-802. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0073854/manifest

Comment

Related Items