Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Regulation of stomatal morphogenesis and lateral root development in Arabidopsis thaliana by the leucine-rich… Keerthisinghe, Sandra Roshini 2016

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

Item Metadata

Download

Media
24-ubc_2016_september_keerthisinghe_sandra.pdf [ 7.61MB ]
Metadata
JSON: 24-1.0303134.json
JSON-LD: 24-1.0303134-ld.json
RDF/XML (Pretty): 24-1.0303134-rdf.xml
RDF/JSON: 24-1.0303134-rdf.json
Turtle: 24-1.0303134-turtle.txt
N-Triples: 24-1.0303134-rdf-ntriples.txt
Original Record: 24-1.0303134-source.json
Full Text
24-1.0303134-fulltext.txt
Citation
24-1.0303134.ris

Full Text

  REGULATION OF STOMATAL MORPHOGENESIS AND LATERAL ROOT DEVELOPMENT IN ARABIDOPSIS THALIANA BY THE LEUCINE-RICH REPEAT RECEPTOR-LIKE KINASE MUSTACHES   by  Sandra Roshini Keerthisinghe BSc. The University of British Columbia, 2008 MSc.  The University of British Columbia, 2011  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Botany) THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver) May 2016 ©  Sandra Roshini Keerthisinghe, 2016    ii  Abstract  Shoot and root systems have evolved numerous specialized structures, including stomata and lateral roots.  Stomata, which are essential for mediating gas exchange across the shoot epidermis during photosynthesis, consist of two bilaterally symmetrical guard cells arranged around an epidermal pore.  The symmetry displayed by mature stomata is essential for stomatal function.   Stomata develop via a dedicated pathway defined by two key divisions, an asymmetric division, and a symmetric division which initiates stomatal bilateral symmetry.  The symmetric division is followed by pore and guard cell morphogenesis, which maintains the previously established bilateral symmetry.   Lateral roots increase the ability of root systems to acquire nutrients and water from the surrounding environment and lateral root development is also dependent on stage specific divisions.  Interestingly, several genes regulating symmetric divisions have also been shown to function in lateral root development.     The Leucine-Rich Repeat Receptor-Like Kinase, MUSTACHES (MUS), which belongs to a family of four closely related MUS-LIKE kinases (MUSLs), is required to enforce bilateral symmetry post-symmetric division.  Mutants in mus display pore and guard cell morphogenesis defects, as well as defects in microtubule array organization and the polarity of microtubule movement.  MUS is expressed in both stomatal lineage and root epidermal cells, suggesting that MUS also functions outside stomatal development.  Thus, MUS represents an ideal candidate through which mechanisms influencing pore and guard cell morphogenesis, as well as the impact of post-symmetric division stomatal genes on lateral root development, may be explored.     Mutant analysis and time-lapse studies demonstrated that MUS enforces bilateral symmetry by ensuring symmetrical positioning of microtubule organizing centres post-symmetric division.  Time lapse studies indicated that microtubule organizing centre delocalization occurred before, and is likely responsible for, the subsequently occurring alterations in microtubule polarity and guard cell morphogenesis defects observed in mus.  As well, this work revealed that MUS, the cytoplasmic kinase NIMA-RELATED KINASE 6 and iii  the microtubule associated protein CLIP-170 ASSOCIATED PROTEIN are required for stomatal microtubule organizing centre formation.  Additionally, a model where pore placement is regulated by opposing MUS and MUSL1 signaling pathways was developed.  Finally, a redundant role for MUS and MUSL1 in lateral root development was confirmed.                  iv  Preface  Portions of Chapter 1 and the bulk of Chapters 2, 3 and 4 will be submitted for publication as three primary research publications.   It is hoped that Chapters 2 and 4 will result in two primary research papers first authored by Sandra Keerthisinghe, while Chapter 3 will likely result in primary research article where Professor Tsuyoshi Nakagawa will be the first author, and Sandra Keerthisinghe will be the second author.    Experiments in Chapter 2 were designed by Fred Sack and Sandra Keerthisinghe, experiments in Chapter 3 were designed by Fred Sack, Geoffrey Wasteneys and Sandra Keerthisinghe, and experiments in Chapter 4 were designed by Fred Sack, Geoffrey Wasteneys, Abel Rosado and Sandra Keerthisinghe.   A portion of the materials utilized in Chapter 3 were generated by Jayme Kim (pro35S:MUSL3-GFP) and Dr.Eun-Kyoung Lee (pro35S:MUSL2-GFP and pro35S:MUSL4-GFP).   The remaining materials, experiments and analyses described in Chapters 2-4 were generated and performed by Sandra Keerthisinghe.             v  Table of contents   Abstract.…………………………………………………………………………………………………………………….….…....ii Preface………………………………………………………………………………………………………………………....…...iv Table of contents………………………………………………………………………………………………………………...v List of figures……………………………………………………………………………………………………………………...xi List of symbols and abbreviations…..………………………………………………………..…………….…….....xiii Acknowledgments …………………………………………………………………………………………………………….xv Dedication.………………………………………..…………………..............................................................xvii Chapter 1: Introduction…….………………………………………………………………………………………….…..…1 1.1 Stomatal function and development ......................................................................... 2 1.1.1  Importance and function .................................................................................... 2 1.1.2  Structural development ...................................................................................... 2 1.1.3  Genetic regulation of development .................................................................... 5 1.2  Lateral root function and development ..................................................................... 8 1.2.1  Role and function ................................................................................................ 8 1.2.2  Structural development ...................................................................................... 8 1.2.3  Genetic regulation of development .................................................................... 9 1.3  Auxin mediated regulation of development ............................................................ 11 1.3.1  Auxin transport and signaling ........................................................................... 11 1.3.2  Role of auxin in stomatal spacing ..................................................................... 11 1.3.3  Role of auxin in lateral root development ........................................................ 12 1.4   Role of the cytoskeleton in morphogenesis with particular reference on microtubules ....................................................................................................................... 12 1.4.1  Microtubule structure and function ................................................................. 12 vi  1.4.2   Mechanisms of microtubule array organization .............................................. 13 1.4.3  Visualization of microtubule organizing centers. ............................................. 14 1.4.4  Role of microtubules in stomatal and lateral root development ..................... 15 1.5  Receptor-like kinases ............................................................................................... 16 1.5.1  Roles, importance and function ........................................................................ 16 1.5.2  Structure ........................................................................................................... 17 1.5.3  Signal transduction ........................................................................................... 17 1.5.4  Receptor-like kinases in lateral root and stomatal development .................... 19 1.6   Research objectives .................................................................................................. 19 Chapter 2: The role of MUSTACHES in microtubule organizing center regulation and pore morphogenesis in stomata…………………………………………………………………………………………………21  2.1  Introduction.............................................................................................................. 21 2.2 Results ...................................................................................................................... 24 2.2.1 Translational GFP fusions of CLASP, GCP2, NEK6 and EB1 show similar distribution patterns in guard mother cells before these precursor cells divide ........... 24 2.2.2 CLASP, GCP2, NEK6 and EB1 are similarly (comparably) localized after guard mother cell divisions ........................................................................................................ 27 2.2.3 The loss of MUSTACHES (MUS) function alters the localization patterns of GCP2, CLASP, NEK6 and EB1 after guard mother cell symmetric divisions .................... 31 2.2.4 Morphogenetic defects in mus take place after CLASP, GCP2, and NEK6 and EB1 become delocalized during the formation of the pore site ..................................... 36 2.2.5 mus is associated with altered microtubule polarity as indicated by the direction of microtubule growth ..................................................................................... 42 2.2.6  The actin related protein BRICK1 localizes along nearly the entire length of the ventral wall, except for the pore site .............................................................................. 51 vii  2.2.7  Double mutants of mus and mutants of the microtubule organizing center markers clasp or nek6 display additional stomatal defects ............................................ 53 2.2.8 GCP2 does not localize to microtubule organizing centers in ‘capsule’ shaped stomata of mus nek6 and mus clasp double mutants ..................................................... 57 2.3 Discussion ...................................................................................................................... 58 2.3.1 GCP2, CLASP, EB1 and NEK6 localize to the cortical microtubule organizing centers in wild-type stomata ........................................................................................... 59 2.3.2 MUS is likely responsible for placement of cortical microtubule organizing centers in stomata ........................................................................................................... 61 2.3.3  Distribution of the actin related protein BRK1 is altered in mus ..................... 62 2.3.4  Distribution of the actin related protein BRK1 is altered in mus ..................... 64 2.3.5 NEK6 may be involved in the targeting of γ-TURCs to the stomatal microtubule organizing centers ........................................................................................................... 65 2.3.6 MUS NEK6 and CLASP likely regulate stomatal morphogenesis ...................... 66 2.4  Summary .................................................................................................................. 67 2.5  Materials and methods ............................................................................................ 68 2.5.1 Growth of plant materials ................................................................................. 68 2.5.2 Plant material .................................................................................................... 68 2.5.3 Sample preparation .......................................................................................... 69 2.5.4 Microscopy ........................................................................................................ 69 2.5.5 Counts and sampling .............................................................................................. 70 2.5.6 Statistical analysis ................................................................................................... 70 Chapter 3 : MUSTACHES, and the closely related kinase MUS-LIKE1 may regulate positioning of the stomatal pore…………………………………………………………………………………………………………..72 3.1  Introduction.............................................................................................................. 72 viii  3.2 Results ...................................................................................................................... 74 3.2.1 MUS and MUSL1 exhibit extensive structural symmetry ....................................... 74 3.2.2 musl mutants do not exhibit stomatal defects ................................................. 76 3.2.3 MUSL2, MUSL3 and MUSL4 over-expression lines do not exhibit distinct sub-cellular localization patterns or gain of function phenotypes ........................................ 78 3.2.4 pro35S:GFP-MUSL1 sub-cellular localization resembles proMUS:MUS-GFP localization ....................................................................................................................... 79 3.2.5   Prolonged localization of MUSL1 to the symmetric division likely causes the butterfly effect ................................................................................................................. 81 3.2.6  Prolonged localization of MUSL1 to the symmetric division likely causes the butterfly effect ................................................................................................................. 84 3.2.7 MUSL1 is expressed, and localizes, throughout the cotyledon epidermis ....... 86 3.3 Discussion ...................................................................................................................... 88 3.3.1 MUSL1 and MUSL3 may function in stomatal development ................................. 88 3.3.2 MUSL1 is neither functionally redundant to MUS, nor a MUS co-receptor, in stomata ……………………………………………………………………………………………………………………..89 3.3.3 An antagonistic relationship between MUS and MUSL1 likely regulates stomatal morphogenesis ................................................................................................. 91 3.4 Summary .................................................................................................................. 93 3.5  Materials and methods ............................................................................................ 94 3.5.1  Growth of plant materials ................................................................................. 94 3.5.2  Plant material .................................................................................................... 94 3.5.3 Expression constructs ....................................................................................... 94 3.5.4 Sample preparation .......................................................................................... 96 3.5.5 Microscopy ........................................................................................................ 96 ix  3.5.6 Counts and statistical analysis .......................................................................... 96 Chapter 4 : The role of MUSTACHES AND MUS-LIKE1 in lateral root development………..…..98 4.1  Introduction.............................................................................................................. 98 4.2 Results .................................................................................................................... 100 4.2.1 MUSTACHES is expressed in developing lateral roots. ................................... 100 4.2.2 Consistent MUSL1 localization is not present in lateral roots ........................ 106 4.2.3 mus mutants contain significantly fewer lateral root primordia.................... 107 4.2.4 mus, musl1 and mus musl1 double mutants contain significantly fewer emerged lateral roots .................................................................................................... 109 4.2.5 MUSL1 overexpression compliments the mus lateral root phenotype .......... 112 4.2.6 Stomatal density is not affected in mus, musl1¸ or mus musl1 backgrounds 112 4.2.7 Application of exogenous auxin rescues the mus lateral root phenotype ..... 112 4.2.8  Prolonged localization of MUSL1 to the symmetric division likely causes the butterfly effect ............................................................................................................... 114 4.3 Discussion ............................................................................................................... 115 4.3.1 MUS and MUSL1 likely function redundantly in lateral root formation ........ 116 4.3.2 MUS may regulate lateral root division and differentiation .......................... 117 4.3.3 The absence of MUSL1 fluorescence in lateral roots is likely the result of a weak promoter. ............................................................................................................. 118 4.3.4 MUS and MUSL1 expression are likely to be auxin dependent ...................... 118 4.4  Summary ................................................................................................................ 120 4.5  Materials and methods .......................................................................................... 120 4.5.1 Growth of plant materials ............................................................................... 120 4.5.2 Plant material .................................................................................................. 121 4.5.3 Expression constructs ..................................................................................... 121 x  4.5.4 Sample preparation and microscopy .............................................................. 122 4.5.5 Quantification of lateral root phenotypes ...................................................... 123 4.5.6  Statistical analysis of lateral root phenotypes ................................................ 123 4.5.7 Quantification and statistical analysis of pro35S:MUSL1-GFP over-expression lines…………………………………………………………………………………………………………………………..124 4.5.8 Comparison of stomatal density phenotypes ................................................. 125 4.5.9 Quantitative real-time PCR ............................................................................. 125 Chapter 5 : Conclusions and future directions………………………………………….………….…………..126 5.1  Major conclusions and significance of findings ..................................................... 126 5.2  Future directions .................................................................................................... 130 5.2.1 Chapter 2 :  the role of MUSTACHES in the regulation of microtubule organizing centers and pore morphogenesis in stomata .............................................. 130 5.2.2  Chapter 3 : MUSTACHES, and the closely related kinase MUS-LIKE1, may regulate positioning of the stomatal pore .................................................................... 133 5.2.3 Chapter 4 : the role of MUS and MUS-LIKE1 in lateral root development ..... 135 References……………………………………………………………………………………………………………………….137          xi  List of figures  Figure 1.1   Stages in stomatal development stages ...................................................................... 3 Figure 1.2   Stages in stomatal pore development ......................................................................... 4 Figure 1.3   Selected proteins involved in transitions between different developmental stages .............................................................................................................................................. 7  Figure 2.1 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in wild-type stomatal precursors prior to guard mother cell division .............................................................. 27 Figure 2.2 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in wild-type stomata after the symmetric division (post-symmetric division) ................................................. 30 Figure 2.3 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in mus stomatal precursors prior to guard mother cell division ............................................................................. 32 Figure 2.4  The mus mutation alters GCP2, CLASP, EB1 and NEK6 localization after the symmetric division (post-symmetric division) .............................................................................  35 Figure 2.5 Stomatal MTOC formation occurs within 2.5 hours of the symmetric division in wild-type stomata. ........................................................................................................................ 39 Figure 2.6 MTOC formation does not occur with 2.5 hours of the symmetric division in mus stomata. .......................................................................................................................................  43 Figure 2.7   Common categories of microtubule organizing center delocalization in mus.   ....................................................................................................................................................... 45 Figure 2.8 The mus mutation alters the direction of microtubule growth in developing guard cells, which may result in a switch from outward microtubule growth to inward microtubule growth. .........................................................................................................................................  50 Figure 2.9 The actin related protein BRK1 is excluded from the pore site during MTOC development. ................................................................................................................................ 52 Figure 2.10  mus clasp and mus nek6 mutants display synergistic phenotypes which generate a high abundance of ‘capsule-shaped’ stomata...........................................................  54 Figure 2.11 Quantification of mus nek6 and mus clasp double mutants ....................................  56 Figure 2.12 MTOC foci do not form in mus clasp and mus nek6 double mutants. .....................  58 Figure 3.1 MUS-LIKE 1 (MUSL1) ehxibits close structural similarity to MUS ...............................  76 xii  Figure 3.2 mus musl1 double mutants do not increase the severity of the mus single mutant phenotype. ...................................................................................................................................  77 Figure 3.3 MUSL2, MUSL3 and MUSL4 over expression (OX) lines do not display distinct sub-cellular localization patterns or gain of function phenotypes. ....................................................  78 Figure 3.4 MUSL1 sub-cellular localization resembles native MUS localization .........................  80 Figure 3.5 MUSL1 OX in a mus or mus musl1 background generates a ‘butterfly’ (bf) effect gain of function phenotype. ........................................................................................................  83 Figure 3.6 The bf phenotype may result from prolonged localization of bf to the symmetric division in the absence of mus. ....................................................................................................  85 Figure 3.7 MUSL1 native expression and localization .................................................................  87 Figure 3.8   Opposing MUS and MUSL1 signaling pathways may position the stomatal pore..... 92 Figure 4.1  MUS expression in lateral root primordia initiates approximately 2.5 days after germination  ................................................................................................................................ 102 Figure 4.2 MUS is expressed in lateral root primordia throughout all stages of lateral root primordia development .............................................................................................................. 104 Figure 4.3 MUS is expressed in 3 consecutive pericycle layers during early stages of lateral root primordia formation ........................................................................................................... 105 Figure 4.4   Patterns of MUS- LIKE1 (MUSL1) expression and subcellular distribution throughout the root system ....................................................................................................... 107  Figure 4.5 Lateral root density is reduced in mus and musl1 mutant backgrounds .................. 111  Figure 4.6   Exogenous auxin application complements the mus lateral root phenotype, but not the musl1 or mus musl1 Lateral root phenotype ................................................................. 115       xiii  List of symbols and abbreviations  AF  Actin filaments ARP  Actin related protein ANCOVA Analysis of covariance ANOVA Analysis of variance AuxRE Auxin responsive element bp  Base pairs EMS  Ethyl methanesulfonate DAG  Days after germination DW  Dorsal wall γ-TURC Gamma-tubulin ring complex γ-TUSC Gamma-tubulin small ring complex GCs  guard cells GMCs guard mother cell GDP  Guanosine diphosphate GTP  Guanosine triphosphate LBD  Lateral boundary domain LR  Lateral root LRP  Lateral root primordia LRR  Leucine rich repeat xiv  MAPK MAP kinase µm  micrometer MTOC Microtubule Organizing Center NA  Numerical aperture OX  Over-expression qPCR  quantitative real time polymerase chain reaction RLK  Receptor-like kinase RLP  Receptor-like protein                xv  Acknowledgments   I would like offer my sincerest thanks and gratitude to Professor Fred Sack (1947 – 2015), who supervised the majority of my PhD thesis.  Fred was a brilliant scientist, and a joyful, kind, patient, and encouraging mentor. I owe all that I have achieved in my academic career to Fred, and cannot adequately express the impact that he had on my development both as a scientist and as an individual. I miss Fred very much, and I will always be grateful that I had the chance to know and to be mentored by him.  I would like to thank Professor Geoffrey Wasteneys for kindly accepting me into his lab and for agreeing to be my supervisor for the remainder of my program.  I greatly appreciate Geoff’s kindness, encouragement and valuable advice which helped to further expand my experiments and finish this dissertation.     I thank the members of my PhD committee, Professor Abel Rosado and Professor George Haughn, for their support, advice and helpful insights, as well as their help in revising my thesis.    I acknowledge and thank all current and past members of Sack, Wasteneys and Rosado labs, and the staff and students of the Botany department for their help and friendship.  As well, I thank Kevin Hodgson, Garnet Martens and the Bio-Imaging Facility for training me in the use of their confocal microscopes, and their advice on imaging techniques.  I owe special thanks to Professor Chris Ambrose for plant materials, and for providing numerous invaluable insights and comments which lead to the initiation of the first project described in this thesis.  I would especially like to thank Dr. Eun-Kyoung Lee, whose friendship, happy nature and countless nuggets of invaluable advice and feedback have meant a great deal to me throughout my years as a graduate student.     Completion of this work was made possible by funding from NSERC and the Department of Botany.  Finally, I must thank my Mom and Dad, for all the sacrifices they have made, and for all the confidence, encouragement and support they have shown me throughout my studies and beyond.  I could never have achieved all that I have without them.   xvi  Dedication     Dedicated to the memory of Professor Fred David Sack             1  Chapter 1 : Introduction   The survival propagation and evolution of biological species is governed by their ability to extract and utilize energy sources from their surrounding environment.  The majority of organisms utilize energy sources ultimately derived from plants (Kingdom Plantae).  Plants, which are primary producers, generate consumable energy sources through the process of photosynthesis, where solar energy is converted into the simple sugar glucose.   As such, plants form the very base upon which all food chains are built upon.   Additionally, photosynthesis also releases the by-product of oxygen, which in turn is essential for respiration, the process which enables organisms to extract the energy stored in glucose.  Further, plants also play vital roles in global carbon cycles, global water cycles, maintaining soil health, and agriculture.  Thus, plants perform essential functions, which profoundly impact not only human society, but the entire global ecosystem as well.        Plantae is classified into two large classes: the non-vascular and vascular plants.   The majority of vascular plants display a standard body plan, comprised of a shoot and a root system.   The above-ground shoot system, which consists of leaves, flowers, anthers and stems, facilitates photosynthesis, reproduction and water transport.   The below ground root system, consisting of primary and secondary roots, enables extraction of nutrients and water from the soil, while also anchoring plants to the surface (Raven et al, 1999).    Shoot and root systems are comprised of a variety of specialized structures that greatly enhance their efficiency. Specialized cell morphologies enhance the ability of organisms to function adaptively, thereby conferring a significant evolutionary advantage.  As such, characterizing cellular mechanisms and genetic pathways governing cell morphogenesis is of great importance.  Two structures with cells having specialized morphologies that are especially integral to plant function are stomata and lateral roots. Stomata and lateral root development provide excellent model systems through which to study how developmental pathways mediate morphogenesis.  Stomatal and lateral root developmental pathways display several common characteristics, including regulation of 2  development by the phytohormone auxin, microtubule dependent morphogenesis, and mediation of the timing and positioning of key divisions through kinase signaling pathways.   1.1 Stomatal function and development  1.1.1  Importance and function    Stomata are specialized shoot epidermal structures consisting of two guard cells surrounding a pore.  As stomatal function is deeply tied to guard cell morphology, stomata provide a valuable model system through which to dissect molecular pathways regulating plant cell morphogenesis.  Stomata exhibit an extensive bilateral symmetry that is apparent both externally in the symmetrical positioning of the two guard cells (which mirror each other in shape) around the pore; and internally in the symmetrical placement of cellular components (including microtubule arrays, cellulose arrays and organelles).   This symmetry may enable guard cells to synchronize their movements in order for the stomata to co-ordinately open and close as a single turgor-mediated unit, thus efficiently permitting CO2 uptake for photosynthesis while also limiting water loss through evapotranspiration.   Terrestrial plants evolved stomata approximately four hundred million years ago.  Currently, stomata are found in a selection of non-vascular plants (Bryophyta), as well as in vascular plants (some Tracheophyta and all Spermatophyta).  However, in terms of both tissue and species diversity, stomata are most widely distributed in seed plants (Spermatophyta) (Willmer and Fricker, 1996).  The evolution of stomata and their wider distribution in seed plants represent major adaptations that enabled these plants to colonize more arid and diverse habitats (Peterson et al., 2010)   1.1.2  Structural development  Arabidopsis thaliana (Arabidopsis) stomata develop through a dedicated pathway, which initiates when an undifferentiated precursor cell, the Meristemoid Mother Cell undergoes an asymmetric division  that produces two daughter cells of unequal size.  The smaller of these daughter cells, the meristemoid, is a stomatal lineage-specific stem cell.  In contrast, the larger daughter cell either undergoes additional asymmetric divisions, or 3  transitions into a generic leaf epidermal cell (Pavement Cell).  The meristemoid may then undergo up to three further asymmetric divisions, before producing an oval guard mother cell.  Subsequently, guard mother cells develop two opposing end wall thickenings on each periclinal face of the guard mother cell.  Establishment of the end wall thickenings signifies the initiation of bilateral symmetry formation in the stomatal pathway.  Symmetry is further reinforced in the pathway when a single symmetric division in the centre of the guard mother cell generates the ventral wall, which extends from one end wall thickening to the other, thereby creating two young guard cells (Fig 1.1) (Bergmann and Sack, 2007; Zhao and Sack, 1999).      Figure 1.1  Stages in stomatal development    Bilateral symmetry is then maintained during the subsequently occurring stage of guard cell morphogenesis.  First, symmetrically placed, lens shaped, (pore) thickenings form 4  in the center of the ventral wall in each guard cell.  Finally, a combination of  hydrolytic enzymes and turgor mediated stretching of the pore thickenings is thought to initiate pore formation, ultimately resulting in a symmetrically placed pore within the mature stoma (Fig 1.2) (Galatis and Apostolakos, 2004; Zhao and Sack, 1999).   Stomata exhibit an extensive bilateral symmetry that is apparent both externally in the symmetrical positioning of the two guard cells (which mirror each other in shape) around the pore; and internally in the symmetrical placement of cellular components.   This symmetry enables guard cells to synchronize their movements in order for the stomata to co-ordinately open and close as a single turgor- mediated unit.   Thus, it has been suggested that disruptions in stomatal bilateral symmetry should markedly decrease the efficiency of shoot gas exchange required for photosynthesis and the regulation of evapotranspiration (Willmer and Fricker, 1996; Zhao and Sack, 1999).     Figure 1.2  Stages in stomatal pore development  5  Red ellipses  in diagram represent end wall thickenings.  Brackets in micrographs denote end wall thickenings, while the arrow denotes a pore thickening.    1.1.3  Genetic regulation of development   Stomatal development is governed by gene networks that act co-ordinately.  In order to ensure optimal function, mature stomata follow the one cell spacing rule, where stomata are spaced at least one cell apart from each other.   Tight control of the placement and formation of the asymmetric division, which initiate entry into the stomatal pathway, is essential for maintenance of one-cell spacing.  Thus, several sets of genes, consisting of both positive and negative regulators, promote the formation and placement of the asymmetric divisions (Fig. 1.3).  Asymmetric division initiation is positively regulated by dimers  consisting of the basic Helix Loop Helix (bHLH) transcription factors SPEECHLESS (SPCH) and two paralogous bHLH-Leucine zipper  proteins, INDUCER OF CBF EPRESSION 1 /SCREAM1 (ICE1/SCRM) and SCREAM2 (SCRM2),  which mediate transitions from meristemoid mother cells to meristemoids (Kanaoka et al., 2008; Serna, 2009).   Two additional SPCH paralogs, the bHLH transcription factors MUTE and FAMA, also dimerize with SCRM/SCRM2 to mediate the transition from meristemoid to guard mother cell (MUTE) and the transition from guard mother cell to mature guard cell (FAMA) (Kanaoka et al., 2008; Ohashi-Ito and Bergmann, 2006; Pillitteri et al., 2007).      SPCH enhances the expression of several key genes required for stomatal development, including the proteins BREAKING OF ASYMMETRY IN THE STOMATAL LINEAGE (BASL), POLAR and TOO MANY MOUTHS (TMM) (Lau et al., 2014).  BASL and POLAR are intrinsic polarity factors which facilitate asymmetric division placement.  BASL also appears to participate in the regulation of the one cell spacing rule, as seen in basl mutants which exhibit stomatal clusters  (Dong et al., 2009; Pillitteri et al., 2011).  Alternately, the Leucine-Rich Repeat Receptor-Like Protein (LRR-RLP) TMM enforces the one-cell spacing rule by negatively regulating stomatal formation (Nadeau and Sack, 2002).   TMM, along with LEUCINE-RICH REPEAT RECEPTOR-LIKE KINASES from the ERECTA (ER) and ER-LIKE (ERL) family, activates the MAP KINASE KINASE KINASE (MPKKK) YODA.  6  YODA initiates a MAP kinase cascade which leads  to phosphorylation and inactivation of SPCH, thereby preventing stomata from forming in contact (Bergmann et al., 2004; Shpak et al., 2005).   The balance between positive and negative regulatory pathways in stomatal development is achieved through the binding of several different small peptide ligands from the EPIDERMAL PATTERNING FACTOR (EPF) and EPF-LIKE (EPFL) families to TMM-ER complexes.   Binding of the negative regulators EPF1 and EPF2 maintains the one cell spacing rule and prevents meristemoid mother cell to meristemoid transitions respectively.   In contrast, binding of the positive regulator STOMAGEN/EPFL9 induces stomatal formation (Hara et al., 2007; Hunt and Gray, 2009; Sugano et al., 2010).    As in the case of asymmetric divisions, the number of Symmetric Division is also strictly regulated by a distinct gene network.   First, the R2R3 MYB transcription factors FOUR LIPS (FLP) and MYB88 act redundantly to limit the number of symmetric divisions to one.  FLP limits symmetric division number by repressing the cell cycle genes CYCLIN DEPENDENT KINASE B 1; 1 (CDKB1;1), and CYCLIN A2 (CYCA2) (Lai, 2005; Vanneste et al., 2011; Xie et al., 2010).   The bHLH FAMA also participates in regulating the number of symmetric divisions, but it does so via a pathway that is independent of FLP (Ohasi-Ito and Bergmann 2006, Lee et al., 2014b).  Further, FAMA interacts with RETINOBLASTOMA RELATED (RBR), a regulator of formative divisions, to control terminal differentiation of guard cells (Lee et al., 2014a; Matos et al., 2014).    In contrast to asymmetric and symmetric divisions, a detailed understanding of how gene networks impact the final stage of guard cell morphogenesis is lacking, and currently only three genes that function during this developmental stage have been identified.   Mutants in the  ADP-RIBOSYLATION FACTOR GUANINE-EXCHANGE FACTOR (ARF-GEF) GNOM display pore morphogenesis defects, while the Domain of unknown function (Dof) transcription factor  STOMATAL CARPENTER 1 (SCAP1) mutants exhibit bilateral symmetry defects.   In contrast, mutants in the LRR-RLK MUSTACHES (MUS) display both pore morphogenesis and bilateral symmetry defects  (Le et al., 2014a; Negi et al., 2013; Keerthisinghe et al., 2015). 7    Figure 1.3 Selected proteins involved in transitions between different developmental stages  A.  Overview of key genes known to regulate transitions between each stage of stomatal development. Dark blue denotes positive regulators of asymmetric divisions, red represents negative regulators of stomatal development, light blue denotes positive regulators of 8  guard mother cells and symmetric division development, orange represents proteins involved in pore and guard cell morphogenesis.  B. Depiction of the signaling pathway mediating asymmetric division placement between developing stomata and adjacent sister cells (inset).  The binding of EPF1 or EPF2 to the members of ERf kinase family and TMM activate a YODA mediated kinase cascade which in turn inactivates SPCH.  In contrast binding of STOMAGEN to ERf members prevents YODA activation, allowing SPCH to initiate transcription of downstream target genes required for stomatal development.      1.2  Lateral root function and development   1.2.1  Role and function    Root systems function to anchor plants, and to absorb nutrients and water from the surrounding soil environment.  The root system encompasses both the main (primary) root, and secondary roots, referred to as lateral roots.  Lateral roots form in response to varying environmental conditions, such as drought.  Thus, lateral roots greatly increase the capacity/efficiency of root systems to acquire resources from the surrounding environment (Malamy and Ryan, 2001).    1.2.2  Structural development  Lateral root formation involves the de novo generation of an entirely new organ from the primary root.  Primary roots exhibit three distinct growth zones - division, elongation and maturation.  The division zone forms near the tip of the root and contains the Root Apical Meristem, a region of active divisions which give rise to each of the distinct root tissues.  The elongation zone represents the region where cells elongate, thereby contributing to the overall growth of the root.  Finally, terminal cell differentiation occurs in the maturation zone (Raven et al, 1999).     Primary roots are comprised of three tissue systems, the epidermis, the cortex and the vascular tissue, arranged as nested and concentric cylinders.  The outermost tissue, the epidermis, forms a barrier between the external environment and the internal tissues.   As 9  well, the epidermis facilitates nutrient and water absorption.   The cortex, consisting of ground tissue, and the endodermis, which contains the Casparian Strip, regulates the diffusion of water and solutes into the central vascular tissue (stele).   Finally, the vascular tissue consists of the pericycle, which encompasses the stele, the phloem, which transports nutrients, and the xylem, which transports water (Raven et al, 1999)  Mature Arabidopsis lateral roots emerge from the maturation zone (Raven et al, 1999).  However, lateral root development commences in the basal meristem of the division zone, where a subset of pericycle cells that are directly adjacent to the xylem (xylem pole pericycle cells) acquire a lateral root founder cell fate (Lavenus, 2013).   Once the lateral root founder cell fate is established, mature lateral roots form through eight well-characterized lateral root primordia stages (Stages I-VIII).  Stage I lateral root primordia develop when the lateral root founder cells undergo two asymmetric divisions to generate a single layer primordium consisting of four cells.  Subsequently a second, symmetric, division results in the formation of a two layer Stage II lateral root primordia containing eight to twelve cells.  Stage III – VII lateral root primordia then continue to form additional cell layers through repeated symmetric divisions.  Finally, stage VI and VII lateral root primordia, which structurally resemble mature roots, form before the lateral root exits the main root    (Malamy and Benfey, 1997).     1.2.3  Genetic regulation of development Acquisition of lateral root founder cell fate (pericycle priming) predominately depends on the activation of the INDOLE-3-ACETIC-ACID28 (IAA28) – AUXIN RESPONSE FACTOR 7 (ARF7) signaling module, which regulates expression of the GATA23 transcription factor.  GATA23 is required for both lateral root founder cell specification in primed xylem pole pericycle cells (De Rybel et al., 2010b), and for initiating the asymmetric division that produces Stage I lateral root primordia (De Rybel et al., 2010b).  Hormone (auxin)-mediated activation of the SOLITARY ROOT / INDOLE-3-ACETIC-ACID14 (SLR/IAA14) - ARF5-ARF7 module, in turn activates downstream genes such as the LATERAL BOUNDARY DOMAIN (LBD) transcription factors, LBD16 and LBD29.  LBD16 and LBD29, which mediate migration of nuclei and the subsequent establishment of asymmetry in lateral root founder cells, are 10  also required for Stage I lateral root primordia formation (Goh et al., 2012; Lavenus et al., 2013). Several Leucine-Rich Receptor-Like Kinases (LRR-RLKs) also function in lateral root development.   Mutants in the RLK ACRINKLY4 (ACR4) exhibit a reduction in emerged lateral roots, multiple pericycle layers and incorrectly spaced lateral roots.   ACR4 facilitates lateral root formation by promoting asymmetric divisions in lateral root founder cells, while also restricting divisions in pericycle cells neighbouring the lateral root founder cell (De Smet et al., 2008).  Two additional LRR-RLKs, which are closely related, HAESA (HAE) and HAESA-LIKE 2 (HSL2) function during the latter stages (Stage VI – VII) of lateral root development.   HAE-HLS2 are required for the emergence of mature lateral roots from the main root (Kumpf et al., 2013). Interestingly, HAE and HLS2 induce the expression of the cell wall remodelling enzymes XTH23/ XYLOGLUCAN ENDOTRANSCLYCOSYLASE6 (XTR6) and EXPANSIN 17 (EXP17).  HAE-HSL2-mediated induction of XTR6 and EXP17 are required to thin the walls of the endodermis, thereby allowing for lateral root emergence (Kumpf et al., 2013; Vilches-Barro and Maizel, 2015).   Recently (Chen et al., 2015) demonstrated that FLP, in conjunction with ARF7, influences lateral root formation by regulating transcription of the auxin efflux carrier PIN3.  FLP, as mentioned previously, also negatively regulates CDKB1;1 and CYCA2;3 expression during stomatal development.  Correspondingly, cyca2;234 triple mutants also display reduced lateral root numbers (Chen et al., 2015).  Intriguingly, AURORA A (AURA) kinases, which regulate early mitotic events and spindle formation, also influence lateral root development.  The AURA subfamily consists of two kinases, AUR1 and AUR2, and aur1 aur2 double mutant’s exhibit reduced lateral root numbers.  This reduction in lateral root numbers is likely due to the presence of cytokinesis defects and mis-oriented cell plates in aur1 aur2 lateral root primordia (Van Damme et al., 2011).    Interestingly,  aur1-2 aur2-2 double mutants also displayed stomatal defects, which resemble the small cell islands, observed in basl2 mutants (Van Damme et al., 2011).   The aforementioned studies provide examples of genes which function in both lateral root and stomatal development, indicating 11  that stomatal developmental genes represent a potential pool of candidates that may function in lateral root development, and vice versa.     1.3  Auxin mediated regulation of development   1.3.1  Auxin transport and signaling  The phytohormone auxin regulates plant growth, patterning, plant cell morphogenesis and cell division.  Due to its wide-spread functions, auxin also mediates a variety of developmental processes in stomata and lateral roots.   Acropetal (shoot to root) auxin transport in the root is regulated, in part, by the AUXIN1/LIKE AUX1 (AUX1/LAX) and PIN-FORMED (PIN) auxin transport families.  AUX1, is an auxin influx carrier belonging to a family of four conserved proteins, which facilitate the transport of auxin into cells (Péret et al., 2012).  In contrast, the eight member PIN (PIN1 – PIN8) family of auxin efflux facilitates transport of auxin out of cells (Křeček et al., 2009).  Mutations in AUX1, as well as in multiple members of the PIN family, result in reduced lateral root numbers (Péret et al., 2009, 2012; Overvoorde et al., 2010). Once auxin accumulates in a cell, it binds to the F-box protein TRANSPORT INHIBITOR (TIR1), which functions as a cytoplasmic auxin receptor.  Once binding occurs, the auxin-TIR1 complex translocates into the nucleus, ultimately resulting in the de-repression of  numerous of auxin mediated genes characterized by the presence of  Auxin Response Elements (AuxREs) in their promoters (Dharmasiri et al., 2005).   AuxRE activity is regulated through opposing transcription factors belonging to the Auxin Response Factors (ARFs) activator and the AUXIN/INDOLE-3-ACETIC-ACID (AUX/IAA) repressor families.  ARFs dimerize with, and are repressed by, AUX/IAA transcription factors in the absence of auxin.  Degradation of AUX/IAA transcription factors in the presence of auxin, results in the de-repression of ARFs and gene transcription (Lavenus et al., 2013).  1.3.2  Role of auxin in stomatal spacing  Recent studies have demonstrated that auxin regulates the spacing of stomatal divisions.  Le et al (2014a) demonstrated that stomatal precursors (meristemoid- guard 12  mother cell) exhibit high concentrations of auxin, as well as the enrichment of PIN3.  In contrast, auxin and PIN3 levels are significantly down regulated post-symmetric division.  Additionally, pin2, 3, 4, 7 and pin1, 3, 4,7 quadruple mutants display stomatal patterning defects.   Thus, PIN3-mediated auxin depletion appears to be a pre-requisite for the transition of meristemoids to guard cells.   Additionally, mutants of the ARF-GEF GNOM, which is responsible for PIN recycling to the cell membrane, display guard cell patterning and guard cell morphogenesis defects, indicating GNOM functions in regulating both these processes (Le et al., 2014a).      1.3.3  Role of auxin in lateral root development As in stomata, initiation and formation of early lateral root developmental stages are auxin dependent.  The acquisition of lateral root founder cell fate is dependent on oscillating auxin concentrations that reach a maxima every 15 hours in a sub-set of undifferentiated pericycle cells in the basal meristem.  Cells exposed to the auxin maxima then become primed to form lateral root founder cells and subsequently lateral roots (De Smet et al., 2007, De Smet, 2012).  Pericycle priming predominately depends on the auxin-dependent activation of IAA28-ARF7 and SLR/IAA14-ARF7-ARF5 signaling modules (Lavenus et al., 2013).     1.4   Role of the cytoskeleton in morphogenesis with particular reference on microtubules   1.4.1  Microtubule structure and function  Application of auxin has been shown to influence morphogenesis by stimulating reorientation of the microtubules which, along with actin filaments, comprise the cytoskeleton (Chen et al., 2014).  In plants, microtubules play an integral role in cell morphogenesis, as they influence the deposition pattern and crystallinity of the Cellulose microfibrils that comprise the cell wall (Fujita et al., 2011).   Microtubules are rigid filaments comprised of repeating alpha-beta (αβ) tubulin dimers.  Both tubulin subunits are GTPases which hydrolyze GTP to GDP.  The α-tubulin subunit in one dimer associates with the β-13  tubulin subunit of another to form a proto-filament.  Thirteen proto-filaments then associate to form a hollow, cylindrical microtubule, 24 nm in width.  Microtubules display a structural polarity, whereby β-tubulin, which is able to rapidly hydrolyze GTP, is present at the rapidly growing plus (+) end of the microtubules.  In contrast, α-tubulin, which does not hydrolyze its GTP, is present at the slowly growing, minus (-) end of the microtubules (Kollman et al., 2011).     The (-) ends of microtubules are sometimes capped by Gamma (γ) Tubulin Ring Complexes (γ-TURCs), which facilitate efficient microtubule nucleation and formation.  γ-TURCs consist of γ-tubulin Small Complex (γ -TUSCs) and three to four additional accessory proteins, referred to as Gamma Complex Proteins (GCPs).  γ –TUSCs comprise the core of the y-TURC, and consists of a ring of γ-tubulin monomers attached to GCPC2 and GPC3 (Kollman et al., 2011).    1.4.2   Mechanisms of microtubule array organization   The majority of animal cells organize complex microtubule arrays through centrosomes localized near the nucleus in interphase cells, or to the edges of spindle poles during mitosis and meiosis.  γ-TURCs are embedded in the centrosomes, suggesting  that these structures function as Microtubule Organizing Centres (MTOCs).   In contrast, plant cells nucleate and organize microtubules through both MTOC-independent and MTOC-dependent mechanisms.  MTOC-independent mechanisms rely on the formation of new microtubules from dispersed cortical y-TURCs which are recruited to existing microtubule templates (microtubule dependent microtubule nucleation).  The formation of new microtubules is followed by their incorporation into higher order arrays through a variety of microtubule- microtubule encounters, such as bundling, cross-overs and severing, or catastrophe (Wasteneys and Ambrose, 2009).     Recent studies have demonstrated that several MTOC-dependent mechanisms of microtubule organization also function in plants.    Ambrose et al (2015) demonstrated that the nuclear envelope in root hairs, as well as young root and leaf epidermal cells, acts as an MTOC for perinuclear microtubules.   These perinuclear microtubules then contribute to the cortical microtubule array as the cell develops (Ambrose 2015).  Additionally, Ambrose and 14  Wasteneys (2011) also demonstrated that young cell edges emerging from recent divisions in roots and cotyledons display high levels of microtubule nucleation, indicating the presence of ‘edge-based’ MTOCs in young, newly divided cells.    1.4.3  Visualization of microtubule organizing centers.    Previous studies, completed over two decades ago, utilized γ-tubulin localization to characterize MTOCs in mature stomata (Marc, 1989; McDonald et al., 1993).   Further attempts to characterize stomatal MTOCs have not been reported, probably because of technical difficulties in applying immuno-localization techniques to stomata. Therefore, knowledge regarding MTOC formation during stomatal development is lacking.   However, a number of recent studies have described several fluorescent MTOC markers, which may also be useful for visualizing stomatal MTOCs.    Nakamura et al., (2010) demonstrated that the γ-TUSC components GCP2 (proGCP2:GCP2-GFP) localizes along cortical microtubules.   Subsequently, GCP2 was shown to localize to edge based MTOCs, along with a GFP marker for the Microtubule Associated Protein (MAP) CLIP-170 ASSOCIATED PROTEIN (CLASP).  CLASP acts as an edge-based rescue factor, which facilitates local microtubule organization, by aiding the movement of microtubules over the sharp edges present in young, recently divided, cells (Ambrose et al 2011).    As MTOCs anchor microtubule (-) ends, microtubules typically grow outwards from the MTOC.  Thus,  the presence of edge based MTOCs can be further visualized through observing  the MAP END BINDING PROTEIN 1 (EB1) , which localizes to growing (+) ends of microtubules.    The majority of EB1 marked microtubules grow outward from edge based MTOCs exhibiting CLASP and GCP2 localization, further corroborating the utility of GCP2, EB1 and CLASP as MTOC markers (Ambrose and Wasteneys, 2011; Ambrose et al., 2011).   Finally, recruitment of γ-TURCs to the centrosome and spindle in human cells has been shown to be mediated through the y-TURC-associated anchoring factor NEURALLY EXPRESSED DEVELOPMENTALLY DOWN-REGULATED1 (NEDD1) (Zeng et al., 2009).   Recently Motose et al., (2008) demonstrated that the Arabidopsis NEDD1 orthologue is positioned in close proximity to the cytoplasmic kinase NIMA-RELATED KINASE 6 (NEK6) in microtubule 15  branch site junctions, suggesting that NEK6 may be also involved in directing γ-TURC localization in plants (Motose et al., 2008, 2011).      The functions of GCP2, CLASP, EB1 and NEK6 have been well-characterized in non-stomatal tissues (hypocotyls and roots).  However, the localization and function of these proteins have not yet been thoroughly characterized in stomatal development.      1.4.4  Role of microtubules in stomatal and lateral root development   A variety of microtubule arrays form throughout stomatal development.   First, microtubules form mesh-like arrays in meristemoids and early guard mother cells.   Microtubule arrays then transition into longitudinal arrays in mature guard mother cells.  Once the guard mother cell undergoes a symmetric division, the microtubule array becomes radial, and displays intense microtubule foci concentrated on the center of the ventral wall.  The microtubule foci in the two guard cells are symmetrically placed (facing each other) across the ventral wall.  The symmetrically positioned, radial, microtubule arrays are subsequently maintained in mature stomata (Lucas et al., 2006). Interestingly, in stomata, γ-tubulin immuno-localization overlaps with the microtubule foci surrounding the developing and mature pore (Marc et al, 1989; McDonald et al., 1993).  Thus, the microtubule foci likely mark the location of a specialized site in the cortex which is required to organize the radial microtubule array, suggesting the presence of distinct, symmetrically positioned, cortical MTOCs in stomata.   In comparison to stomata, relatively few studies have characterized microtubule arrays, or microtubule associated proteins, in lateral root development.  However, (Abu-Abied et al., 2015) recently demonstrated that mutations in several microtubule associated proteins affected morphogenesis in adventitious roots.  Adventitious roots, which develop from shoot tissues (leaves, stems and hypocotyls), enhance the ability of plants to locate and absorb water and nutrients.  Although they differentiate from distinct tissues, adventitious roots and lateral roots display shared developmental characteristics and exhibit some similarities in auxin mediated signaling modules (Verstraeten et al., 2014).  Additionally, the presence of CLASP and GCPs at edge-based MTOCs in the primary roots (from which lateral roots originate) indicate that MAPs and γ-TURC components are 16  required for the maintenance of morphogenesis and the transition of cells from the division to elongation zone.  Thus, microtubule associated proteins and microtubule array organization, may also influence lateral root differentiation, development and morphogenesis.   1.5  Receptor-like kinases  1.5.1  Roles, importance and function  Plants, despite being sessile organisms comprised of non-motile cells, are capable of sophisticated and efficient responses to their environment. Much of this responsiveness may be due to the ability of plants to perceive extracellular signals through membrane-bound receptors, which rapidly transduce these signals into internal cellular responses. Two main categories of membrane bound receptors exist in in plants: Receptor Histidine Kinases and Receptor-Like Serine/Threonine Kinases (RLKs) (Shiu and Bleecker, 2003).   Shiu and Bleecker (2001) demonstrated that Arabidopsis/plant RLKs belong to the RLK/PELLE family, as they are monophyletic with Drosophila PELLE kinases.  The RLK/PELLE family encompasses more than 600 members (roughly 2.5% of the proteins encoded in the genome, in Arabidopsis), and is thus considered one of the largest protein families in Arabidopsis (Shiu and Bleecker, 2001, 2003).   Both membrane-bound receptor like kinases (RLKs) and non-membrane bound receptor-like cytoplasmic kinases (RLCKs) are contained in the RLK/PELLE family. Approximately two thirds of RLK/PELLE members are RLKs.    The importance of Arabidopsis RLKs is demonstrated by the diverse range of developmental processes they mediate in plants.   For instance, FLAGELLIN SENSITIVE 2 (FLS2) is involved in plant defense and immunity (Gómez-Gómez and Boller, 2000), while CLAVATA1 (CLV1) maintains the size of the shoot apical meristem, and  ERECTA regulates inflorescence structure respectively (Clark et al., 1993, 1997; Shpak et al., 2005).  Finally, BRASSINOSTEROID (BRI1) mediates Brassinosteroid (BR) responsive plant growth (Wang et al., 2001; Yokota, 1997).   17  1.5.2  Structure  The majority of RLKs exhibit a characteristic structure consisting of a N-terminal signal sequence, which is followed in consecutive order by three domains: (1) the extracellular domain, (2) the trans-membrane domain and (3) the Kinase Domain (Shiu and Bleecker, 2001; Wierzba and Tax, 2013).  The extracellular domain perceives and binds external signals, while the trans-membrane domain anchors the RLK to the cell membrane.  The intracellular kinase domain allows the transmission and amplification the external signal into the cell (Wierzba and Tax, 2013; Hanks and Hunter, 1995).   RLK/PELLE members exhibit a variety of extracellular domains, including Leucine-Rich Repeats (LRRs), Epidermal Growth Factors (EGF)-like repeats, Self-Incompatibility (S)-Domains and Lectin-domains (Goff and Ramonell, 2007).  LRRs are the most common extracellular domain, as they are present in approximately half of all RLKs (Shiu and Bleecker, 2001).    LRRs consist of twenty to thirty amino acids and are characterized by a high number of repeating leucine sub-units.  LRR domains typically form α/β horseshoe folds capable of mediating protein-protein interactions (Bella et al., 2008).  Shiu and Bleecker (2001) demonstrated that LRR-RLKs can be subdivided into 12 distinct families based on the amino acid similarity in their kinase domains.   However, in contrast to the diversity present in RLK extracellular domains, all members of the RLK/PELLE family exhibit Serine/Threonine specificity in their kinase domains.   Kinase domains contains kinase enzymes that catalyze the transfer of phosphates from ATP or GTP onto specific serine and threonine residues on interacting substrates.   The catalytic domain of the kinase consists of 250 – 300 amino acids, and contains 12 conserved sub-domains (I – XII), that are essential for substrate binding, phosphorylation and subsequent signaling transduction (Hanks and Hunter, 1995; Stone and Walker, 1995).     1.5.3  Signal transduction  Despite their structural and functional diversity, the majority of RLK signal transduction pathways exhibit a standard mode of action.  Signal transduction is initiated by binding of a signal molecule (ligand) to the extracellular domain.   Upon ligand binding, RLKs dimerize with a co-receptor, which results in a conformational change in the RLK that is 18  often followed by auto-phosphorylation of the kinase domain, which activates the RLK.  Subsequently, the kinase domain phosphorylates various downstream signaling components (substrates).   Thus, the initial extracellular signal is transduced and amplified into an intracellular signal that impacts immunity or development (Torii, 2008).   Additionally, there are many atypical kinases which contain non-functional kinase domains.   Atypical kinases transmit signals through a variety of alternate methods, including binding and activating RLKs with functional kinase domains, or acting as scaffolds for the attachment and activation of downstream signaling proteins. (Castells and Casacuberta, 2007)   Ligand binding to the extracellular domain is an integral part of RLK signaling, and numerous ligands, each mediating diverse signals, exist in Arabidopsis.   Well-characterized ligands include small peptide ligands,  phytohormones and Pathogen Associated Molecular Patterns (PAMPs)   (Torii, 2008).   PAMPs consist of structural components from various pathogens, such as parts of the cell wall from fungi, or components of  flagellin and peptides from bacteria (Torii, 2008).  The phytohormone BRASSINOSTEROID (BR) has been shown to activate the BRI1 signaling pathway.   Finally, small peptide ligands, including the EPF/EPFL family, are proteins consisting of 12 -200 amino acids that bind to extracellular domains (Gao and Guo, 2012,Torii, 2008).   Once the ligand binds to the extracellular domain, RLKs then dimerize with a co-receptor.  Two main categories of co-receptors exist: (1) Receptor-Like Kinases (RLKs), and (2) Receptor-Like Proteins (RLPs).   Co-receptors, which are also RLKs, may form homodimers consisting of two identical RLK partners, as in the case of CHITIN ELICITOR RECEPTOR KINASE 1 (CERK1).   Alternately, RLK co-receptors may form hetero-dimers consisting of two different RLK partners, as is the case of the LRR-RLK BRI1 and its co-receptor the LRR-RLK BAK1.  Often, co-receptor pairs are comprised of closely related kinases, as in the case of the LRR-RLKs BARELY ANY MERISTEM (BAM) and CLAVATA1 (CLV1) kinases which regulate shoot apical meristem maintenance (DeYoung et al., 2006).   Leucine-Rich Repeat Receptor-Like Proteins (LRR-RLPs) comprise the second co-receptor category.  LRR-RLPs exhibit  both an extracellular domain (LRRs) and trans-membrane 19  domain, but lack the kinase domain.   Well characterized LRR-RLK – LRR- RLP interactions include the LRR-RLK CLV1 and the LRR-RLP CLV2, as well as the LRR-RLK ERECTA and the LRR-RLP TMM (Jeong et al., 1999; Le et al., 2014b)     1.5.4  Receptor-like kinases in lateral root and stomatal development   As previously mentioned, LRR-RLKs significantly impact both stomatal and lateral root morphogenesis.  In lateral roots, ACR4 functions redundantly with RLKs from the CRR family, to initiate Stage I lateral root primordia formation (De Smet et al., 2008).   Additionally, the closely related HAE and HSL2 LRR-RLKs are required for emergence of the mature lateral root from primary roots (Kumpf et al., 2013).  In stomata, members of the ER/ERL and SERK LRR-RLK families influence stomatal formation and placement by regulating  asymmetric division formation (Meng et al., 2015).   Finally, MUSTACHES (MUS) regulates the critical post-symmetric division stage of pore morphogenesis (Keerthisinghe et al., 2015).    Mutants in MUS display abnormal guard cell and pore morphogenesis.  Additionally, mus mutants also display defects in both microtubule array organization and the direction of microtubule growth in mature stomata, indicating the presence of potential stomatal MTOC defects.  MUS exhibits an intriguing distribution pattern, whereby it localizes to almost all division sites in the leaf epidermis, with the notable exception of the symmetric division which forms the young stomata.  Instead, MUS localizes to the periphery (or dorsal walls) of young stomata, suggesting that MUS influences stomatal morphogenesis at the center of the stomata through a signal emitted from the stomatal periphery.    Arabidopsis also contains four MUS-LIKE kinases (MUSLs) which are closely related to MUS.  The MUSLs may represent potential MUS co-receptors (Keerthisinghe et al., 2015).     1.6   Research objectives  Although the latter stages of stomatal development (guard cell and pore morphogenesis) are vital for stomatal function, these stages currently remain under-studied.   As such, further characterization of several  intriguing research areas, including signaling pathways regulating microtubule associated  protein (MAPs and γ-TURCs) 20  localization and microtubule array organization during guard cell morphogenesis,  the presence of LRR-RLK networks during the latter part of stomatal morphogenesis and the role, if any, of late acting stomatal genes in developmental pathways outside stomatal development, is required.  The LRR-RLK MUS influences pore and guard cell morphogenesis (post-symmetric division) by regulating microtubule array organization, as well as the polarity of microtubule movement in stomata.   Additionally, MUS also belongs to a family of closely related LRR-RLKs.    As such, MUS represents an excellent tool through which to investigate the impact of LRR-RLKs, and LRR-RLK signaling networks, on the relationships existing between microtubule accessory proteins, microtubule array organization, and guard cell and pore morphogenesis.  Additionally, several proteins that act during the later stages of stomatal morphogenesis (such as FLP and CDKB1;1) also influence lateral root development.   Thus, MUS, which also acts post-symmetric division, may also participate in lateral root development.  Therefore, this thesis will characterize the role of MUS in stomatal and lateral root development by: (1) Exploring  the role of MUSTACHES (MUS), MTOCs and microtubule arrays in regulating stomatal bilateral symmetry, (2) Investigating  a possible role for MUS- LIKE KINASES (MUSLs) in the stomatal MUS signaling pathway, and (3)  Characterizing the role of MUS  and MUSL1 in lateral root development.                                                                                                                                           21  Chapter 2: The role of MUSTACHES in microtubule organizing center regulation and pore morphogenesis in stomata   2.1  Introduction   Specialized cell morphologies enhance the ability of organisms to function adaptively, thereby conferring a significant evolutionary advantage.  As such, characterizing genetic pathways and cellular mechanisms governing cell morphogenesis is of great importance.   Stomata are specialized shoot epidermal structures consisting of two guard cells surrounding a pore.  The guard cells function together to open and close the pore, thus controlling the amount of CO2 exchange while limiting transpiration. Stomata exhibit an extensive bilateral symmetry that is apparent both externally in the symmetrical positioning of the two guard cells (which mirror each other in shape) around the pore; and internally in the symmetrical placement of cellular components (including organelles, intracellular microtubule and actin filament arrays, and extracellular cellulose microfibrils).   This symmetry enables guard cells to synchronize their movements in order for the stomata to co-ordinately open and close as a single turgor-mediated unit, thus efficiently permitting CO2 uptake for photosynthesis while also limiting water loss through evapotranspiration (Willmer and Fricker, 1996; Zhao and Sack, 1999).  Since stomatal function is deeply tied to guard cell morphology, stomata provide a valuable model system through which to dissect molecular pathways regulating plant cell morphogenesis.     The stomatal development pathway has been well characterized, further adding to the value of the stomatal model system.  Stomata develop from an undifferentiated precursor cell, the Meristemoid Mother Cell (MMC), which undergoes an asymmetric division that produces two daughter cells of unequal size.  The smaller of these daughter cells, the Meristemoid (M), is a stomatal lineage-specific stem cell.   The meristemoid, after undergoing one to three further asymmetric divisions, then transitions into an oval guard mother cell and subsequently develops two opposing end wall thickenings  on each periclinal face of the guard mother cell.  Establishment of the end wall thickenings signifies the initiation of bilateral symmetry formation in the stomatal pathway.  Symmetry is further 22  enforced in the pathway when a single symmetric division of the guard mother cell generates the ventral wall and creates two young guard cells.  Bilateral symmetry is then maintained during the subsequent stage of guard cell morphogenesis via the formation of pad-shaped thickenings at the centre of the ventral wall in each guard cell.  As the stoma develops, the thickenings tear apart to create a symmetrically placed pore within a mature stoma. (Chapter 1 Fig 1.2) (Bergmann and Sack, 2007; Zhao and Sack, 1999).    Guard cells are characterized by their radial microtubule arrays, the precise arrangement of which is required for proper guard cell morphogenesis (Lucas et al., 2006).  microtubules nucleate from γ-Tubulin Ring Complexes (γ-TURCs), which consist of γ-tubulin and six associated gamma complex proteins (GCPs) (Kollman et al., 2011).  Recruitment of γ-TURCs to the centrosome and spindle in human cells has been shown to be mediated through the γ-TURC-associated anchoring factor NEURALLY EXPRESSED DEVELOPMENTALLY DOWN-REGULATED1 (NEDD1) (Zeng et al., 2009).   Recently Motose et al (2008) demonstrated that the Arabidopsis NEDD1 orthologue is positioned in close proximity to the cytoplasmic kinase NIMA-RELATED KINASE 6 (NEK6) in branch site junctions, suggesting that NEK6 may be also involved in directing γ-TURC localization in plants.    Each guard cell exhibits a distinct aggregation of microtubules (microtubule foci) concentrated on the centre of the ventral wall.  The microtubule foci in the two guard cells are symmetrically placed (facing each other) across the ventral wall.  The location of the microtubule foci in guard cells appears to overlap with γ-tubulin (Marc et al., 1989), indicating that the radial array is nucleated from the microtubule foci.  Thus, the microtubule foci likely mark the location of a specialized site in the cortex that is required to organize the radial microtubule array, suggesting the presence of distinct cortical  microtubule organizing centres (MTOCs) in stomata.  The majority of plant cells do not contain an equivalent to the cortical MTOCs (centrosomes) found in animal cells.  Instead, to date, three alternative mechanisms of cortical microtubule array organization have been uncovered in Arabidopsis. First, cortical microtubules have been shown to nucleate from γ-tubulin ring complexes dispersed throughout the cortical cytoplasm (Nakamura et al., 2010).  Recently, the nuclear envelope 23  in root hairs, as well as young root and leaf epidermal cells, has been shown to act as an MTOC for perinuclear microtubules.   These perinuclear microtubules then contribute to the cortical microtubule array as the cell develops (Ambrose and Wasteneys, 2014).  Finally, Ambrose and Wasteneys, (2011) have demonstrated that young cell edges emerging from recent divisions in roots and cotyledons display high levels of microtubule nucleation, indicating the presence of ‘edge-based’ MTOCs in young, newly divided cells.    Edge based MTOCs are characterized by aggregations of the γ-TURC component GAMMA COMPLEX PROTEIN 2 (GCP2) at newly formed cell edges. In addition to GCP2, the Microtubule Associated Protein (MAP) CLIP-170 ASSOCIATED PROTEIN (CLASP) also localizes to new cell edges, suggesting that CLASP is a component of edge- based MTOCs.  The presence of edge-based MTOCs is further validated by the observation that microtubules (visualized through a GFP fusion to END BINDING PROTEIN 1 , which marks the plus, or growing, end of microtubules )  grow outward from new cell edges when both GCP2 and CLASP localize to these new  edges (Ambrose et al., 2011; Ambrose and Wasteneys, 2011).   To our knowledge an understanding of genetic pathways and mechanisms regulating the relationship between stomatal MTOCs and stomatal morphogenesis is lacking.  To date, two classes of genes have been shown to regulate guard cell morphogenesis.  Class one genes  act on the symmetric division, and include a  set of co-ordinately functioning cell cycle genes, including CYCLINA 2.2 and 2.3, and the CYCLIN-DEPENDENT KINASE B1 (CDKB1;1), which regulate the symmetric division that produces the two young guard cells (Vanneste et al., 2011; Xie et al., 2010).   Class two genes are required for proper guard cell morphogenesis after the symmetric division stage, and include the domain of unknown function (Dof) transcription factor STOMTATAL CARPENTER 1 (SCAP1), and the leucine-rich repeat receptor-like kinase (LRR-RLK) MUSTACHES (MUS) (Keerthisinghe et al., 2015; Negi et al., 2013).   Of the genes known to regulate guard cell morphogenesis, MUS is of particular interest, as mus mutants display an assortment of abnormalities in guard cell morphogenesis ranging from skewed ventral walls and pores, misplaced pores, and abnormally expanded guard cells lacking pores.  Guard cells in mus mutants also display a 24  variety of abnormal microtubule arrays, and abnormal microtubule polarity, which reflect their abnormal cellular morphogenesis. For instance, skewed guard cells exhibit skewed microtubule foci and consequently also skewed microtubule arrays.  As MUS is an LRR-RLK that functions after the symmetric division, MUS may represent a key node in a developmental pathway regulating either the formation or placement of stomatal MTOCs (Keerthisinghe et al., 2015).  Here we demonstrate that MUS acts after the symmetric division to maintain symmetry during guard cell morphogenesis by regulating the symmetrical placement of the MTOCs during pore thickening formation.  We also show that CLASP and NEK6 influence guard cell morphogenesis by acting in pathways which likely function in parallel to the MUS pathway.  2.2 Results  2.2.1 Translational GFP fusions of CLASP, GCP2, NEK6 and EB1 show similar distribution patterns in guard mother cells before these precursor cells divide    Distribution patterns of the MTOC-associated proteins GCP2, CLASP, NEK6 and EB1 were characterized in wild-type and mus stomata to define the contribution of the stomatal MTOCs to guard cell morphogenesis, as well as to better resolve MUS function during the formation of stomatal symmetry.    Initial observations of successive stages of a sample of 15-20 stomata and a sample of wild-type Non-Stomatal Epidermal Cells (NSECs) in leaves indicate that the stomatal distribution of GCP2, CLASP and NEK6 resemble that of NSECs.  In NSECS, GCP2-GFP (proGCP2:GCP2-GFP), which marks the minus end of microtubules, localizes to the cortex region in the form of small, circular, punctae that are typically dispersed throughout the cortex and also aggregate at interfaces between newly divided epidermal cells, whereas GFP-CLASP (proCLASP:GFP-CLASP) localizes weakly along microtubules in the cortex and also concentrates at cell interfaces in newly divided NSECs (Ambrose et al., 2011).  NEK6-GFP (proNEK6:NEK6-GFP) localizes along cortical microtubules and cell interfaces as well.  However, while NEK6-GFP displays strong localization to NSEC interfaces, it does not form 25  aggregates at these interfaces. These findings are consistent with previous localization studies of GCP2-GFP, GFP-CLASP and NEK6-GFP in Arabidopsis (Ambrose et al., 2011; Ambrose and Wasteneys, 2011; Motose et al., 2008).  In stomata, GCP2-GFP and GFP-CLASP also localize to cortical regions, where GCP2-GFP is in the form of dispersed punctae as well as aggregates at the interfaces between cells in the stomatal lineage and their adjacent sister cells, while GFP-CLASP localizes weakly along microtubules in the cortical region and also concentrates at selected cell edges throughout stomatal development. NEK6-GFP localization largely resembles that of GFP-CLASP and GCP2-GFP throughout stomatal development, with the exception that NEK6-GFP localization along radial microtubules is either very weak, or absent.   The position of GFP-CLASP and GCP2-GFP aggregations vary according to the stage of stomatal development.   In meristemoids, GCP2-GFP punctae aggregate at the interface between the meristemoids and their adjacent sister cells (which usually develop into pavement cells) (Fig. 1 A).  Some meristemoids also exhibit weak aggregations of GCP2 that localize to the cell interface directly opposite the meristemoid/sister cell interface.  GCP2-GFP aggregations at the sister cell interface can also be found during early guard mother cell development (Fig. 2.1 B, J).   These GCP2-GFP aggregations (at the meristemoid-sister Cell interface, as well as the guard mother cell -sister cell interface), are limited to the cell cortex, and thus do not extend into the mid-plane.  Notably, as with GCP2-GFP, GFP-CLASP also aggregates at interfaces between meristemoid, as well as early guard mother cells and their respective sister cells (Fig. 2.1 E).  As in the case for GCP2-GFP, GFP-CLASP aggregations do not extend past the cell cortex during the meristemoid- guard mother cell stages of development. Although NEK6-GFP also localizes weakly to the interface between sister-cells and stomatal precursors, it does not form aggregations at these interfaces (Fig. 2.1 M).       During the developmental stages that immediately precede symmetric divisions (i.e. mid- to late-stage guard mother cells), GCP2-GFP, GFP-CLASP and NEK6-GFP distribution patterns are stage dependent.  In the cell cortices of middle-stage (mid-stage) guard mother cells, GCP2-GFP localizes to dispersed cortical punctae. Mid-stage guard mother cells 26  viewed in optical mid sections show that GCP2-GFP also localizes to punctae that are distributed discontinuously around the periphery of the guard mother cell cortex (Fig. 2.1 B).  GFP-CLASP localizes to cortical microtubules in mid-stage guard mother cells, as well as to discontinuous punctae located along the periphery of mid-stage guard mother cells viewed in their mid-section (Fig 2.1 F).  While NEK6-GFP localized to faint peripheral punctae in mid-stage guard mother cells, its localization to cortical microtubule arrays was either very weak, or absent.   In mature guard mother cells, GCP2-GFP and GFP-CLASP become redistributed to end wall thickenings, as well as to a narrow cortical band that forms between these end wall thickenings (Fig. 2.1 C,G).  However, the localization of GFP-CLASP to this cortical band is weaker than GCP2-GFP localization in mature guard mother cells.  In contrast to GFP-CLASP and GCP2-GFP, NEK6-GFP localization to end wall thickenings was weak, and at times inconsistent (Fig 2.1 N-O). During cytokinesis, the GCP2-GFP, GFP-CLASP and NEK6-GFP signals localize to the phragmoplast (Fig. 2.1 D-H, P), yet, GFP-CLASP and NEK6-GFP signals are much weaker than those  of GCP2-GFP.    Throughout stomatal development EB1b-GFP (pro35S:EB1b-GFP) localizes to mobile “comets” seen as “dots” in cell cortices, which  represent the growing (plus) ends of cortical microtubules.  EB1 distribution does not closely resemble that for GCP2-GFP and GFP-CLASP during early stomatal development.  In contrast to GFP- CLASP and GCP2-GFP, on the whole, EB1b-GFP does not concentrate at specific locations during early stomatal development (from the early meristemoid to late guard mother cell stages) (Fig. 2.1 I-J).  However, EB1b-GFP aggregations were observed in end wall thickenings  in a very small proportion of guard mother cells (approximately 2 out of 20 checked). EB1b-GFP was also detected in well-defined cortical microtubule bands that emerged from end wall thickenings  in guard mother cells just before their symmetric division (Fig. 2.1 K).  During their symmetric division, EB1b-GFP became strongly localized to the phragmoplast (Fig. 2.1 L).  Despite the absence of EB1b-GFP concentrations during early meristemoid-to-late guard mother cell development), mid-plane sections from these stages revealed the presence of dispersed points of fluorescence that were localized along the periphery of these cell types.   27    Figure 2.1 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in wild-type stomatal precursors prior to guard mother cell division 28  Green fluorescence represents proGCP2:GCP2-GFP (A –D),   proCLASP:CLASP-GFP  (E-H), pro35S:EB1b-GFP  (I – L) and proNEK6:NEK6-GFP (M-P) respectively.   Meristemoids  (M),  early and mature stage guard mother cells  (GMC), dividing guard mother cell (DIV).  All scale bars are 5 µm.  (A – D) proGCP2:GCP2-GFP distribution patterns prior to the symmetric division in wild-type.  Meristemoids (A).  Young guard mother cells (B).  GCP-GFP localizes to end wall thickenings  in mature guard mother cells (C).  Intense GCP2-GFP localization is present in the phragmoplast of dividing guard mother cells (D) (E - H) proCLASP:GFP-CLASP localization prior to the symmetric division in wild-type.  GFP-CLASP localizes to  Meristemoids (E), young guard mother cells (F),  and end wall thickenings  in mature guard mother cells (G).  GFP-CLASP  end wall thickening localization is less intense than GCP2-GFP localization at comparable stages of development.  Finally, GFP-CLASP localizes to the phragmoplast of dividing guard cells (H).   (I – L)  pro35S:EB1b-GFP localization prior to the symmetric division in wild-type.  Meristemoids, (I).   Young guard mother cells (J).  EB1b-GFP does not localize to end wall thickenings , but does localize to the band that forms between them in mature guard mother cells (K). Intense EB1b-GFP distribution is visible in the phragmoplast of dividing guard mother cells (L). (M- P)  proNEK6:NEK6-GFP localization prior to the symmetric division in wild-type.   NEK6-GFP localization in  meristemoids   (M), young guard mother cells (N), the end wall thickenings  in mature guard mother cells (O) and to the phragmoplast in dividing guard mother cells (P).    2.2.2 CLASP, GCP2, NEK6 and EB1 are similarly (comparably) localized after guard mother cell divisions   At the pore thickening phase, which occurs shortly after the guard mother cell divides to form young guard cells, GCP2-GFP, GFP- CLASP and NEK6-GFP become discontinuously localized along the length of the formed ventral wall (Fig. 2.2 A, E, I, M).   As stomata develop, GCP2-GFP, GFP-CLASP and NEK6-GFP then aggregate into a narrow circular domain, or focus, in the middle of the maturing ventral wall (Fig.2.2 B-C, F-G, J-K, and N-O).  The GCP2-GFP aggregations were only seen in cortical face views, while the GFP-CLASP and NEK6-GFP aggregations appeared to extend deeper into the stoma.  Intriguingly, in contrast 29  to the weak localization observed during earlier developmental stages, NEK6-GFP displays especially strong localization to the circular foci in developing guard cells.  In addition, observations of mid-plane sections in young guard cells indicate that GCP2-GFP, GFP-CLASP and NEK6-GFP also localize discontinuously around the periphery of the guard cell cortex. The peripheral distribution of GFP-CLASP in young guard cells appears to be more pronounced than that of GCP2-GFP and NEK6-GFP at comparable stages of development.  At the pore formation stage, GCP2-GFP and GFP-CLASP localization persists at the pore site.  In addition, a new site of ‘cortical’ aggregation develops during pore maturation.  This site consists of GCP2-GFP and GFP-CLASP punctae arranged in a semi-elliptical array, which is symmetrically positioned around the developing pore in each facing guard cell (Fig. 2.2 C, G, K, O).  GCP2-GFP and GFP-CLASP remain localized to the cortex around the pore in mature stomata.  However, in contrast to developing stomata, the distribution of GCP2-GFP and GFP-CLASP is less concentrated around the mature pore (Fig. 2.2 D, H).  Observations of mid-plane sections demonstrate that GCP2-GFP and GFP-CLASP are also irregularly localized to the stomatal periphery in mature guard cells, with GFP-CLASP appearing to be higher in intensity.  NEK6-GFP localization was not observed at the pore site, or the periphery, in mature stomata (Fig 2.2 P).  During later stages of development, EB1b-GFP localization begins to resemble that of GCP2-GFP and GFP-CLASP.   Just after the symmetric division, EB1b-GFP localizes discontinuously along the length of the newly divided wall (Fig. 2.2 I).  EB1b-GFP then forms a distinct aggregation at the centre of the newly formed ventral wall in young guard cells (Fig. 2.2 J-K).  EB1b-GFP continues to mark the pore as stomata mature, although EB1b-GFP localization is weaker here compared earlier stages (Fig.2.2 L).   In summary, all four markers show similarities distributions patterns, and these similarities are especially evident after the symmetric division. GCP2-GFP,GFP-CLASP and EB1b-GFP localize to the end wall thickenings, and the cortical band forming between the end wall thickenings . All four markers can then be detected in the phragmoplast. The markers then align discontinuously along the young wall formed by a new symmetric 30  division.  Finally, the markers aggregate into a condensed circular domain that appears to predict the site where pore formation will occur.          31  Figure 2.2 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in wild-type stomata after the symmetric division (post-symmetric division) Green fluorescence represents proGCP2:GCP2-GFP (A –D),   proCLASP:GFP-CLASP  (E-H), pro35S:EB1b-GFP  (I – L) and proNEK6:NEK6-GFP (M-P) respectively.  Developing guard cells (Young GCs)  and   mature stomata  (Mature GCs).  All scale bars are 5 µm.   (A – D) proGCP2:GCP2-GFP localization post-symmetric division in wild-type.  Young, recently divided, guard cells exhibit GCP2-GFP localization along the newly formed ventral wall (A).  GCP2 forms a condensed aggregate in the centre of the ventral wall in developing stomata (B-C).   GCP2 remains localized around the pore in mature stomata (D).  (E - H) proCLASP:GFP-CLASP  localization post-symmetric division in wild-type.  GFP-CLASP localizes along the new ventral wall in young guard cells (E), GFP-CLASP forms a condensed aggregate in the centre of the ventral wall in developing stomata (F-G).   GFP-CLASP remains localized around the pore in mature stomata (H).   (I – L)  pro35S:EB1b-GFP  localization post-symmetric division in wild-type.  In young, recently divided guard cells, EB1b-GFP localizes along the new ventral wall (I).  EB1b-GFP then forms a condensed aggregate in the centre of the ventral wall in developing stomata (J-K).   EB1b-GFP remains localized around the pore in mature stomata (L)   (M- P)  proNEK6:NEK6-GFP localization post-symmetric division in wild-type.  NEK6-GFP localizes along the newly formed ventral wall in recently divided guard cells (M). NEK6-GFP displays intense localization to condensed aggregates in the centre of ventral wall in developing guard cells (N-O).   NEK6-GFP does not localize to mature guard cells (P).     2.2.3 The loss of MUSTACHES (MUS) function alters the localization patterns of GCP2, CLASP, NEK6 and EB1 after guard mother cell symmetric divisions  Prior to the symmetric division during stomatal development, the localization of GCP2-GFP, GFP-CLASP, NEK6-GFP and EB1b-GFP in mus stomata mirrors that of the wild type at comparable stages of development. All four markers localize to the interface between the meristemoid and sister cell as well as to that of the early guard mother cell -sister cell interface.  In mid-stage guard mother cells, all three markers are then peripherally localized.  In mature guard mother cells, these markers are then expressed in end wall thickenings. GCP2-GFP, EB1b-GFP, and GFP-CLASP can then be detected in the 32  phragmoplast.  Finally, all three markers align discontinuously along the newly formed wall (Fig. 2.3 A-P, Fig. 2.4 A).     Figure 2.3 GCP2, CLASP, EB1 and NEK6 display similar localization patterns in mus stomatal precursors prior to guard mother cell division 33  Green fluorescence represents proGCP2:GCP2-GFP (A –D),   proCLASP:GFP:-CLASP  (E-H), pro35S:EB1b-GFP  (I – L) and proNEK6:NEK6-GFP (M-P) respectively.   Meristemoids (M). Early and Mature stage guard mother cells  (GMC) dividing guard mother cell  (DIV).  All scale bars are 5 µm.  (A – D) proGCP2:GCP2-GFP distribution patterns prior to the symmetric division in mus.  Meristemoids  (A).  Young guard mother cells (B).  GCP2-GFP localizes to end wall thickenings in mature guard mother cells (C).  Intense GCP2-GFP localization is present in the phragmoplast of dividing guard mother cells (D). (E - H) proCLASP:GFP-CLASP localization prior to the symmetric division in mus.  GFP-CLASP localizes to meristemoids (E), around the cortical periphery in young guard mother cells (F), and to end wall thickenings in mature guard mother cells (G). GFP-CLASP end wall thickening localization is less intense than GCP2-GFP localization at comparable stages of development.   Finally, GFP-CLASP localizes to the phragmoplast of dividing guard cells (H).   (I – L)  pro35S:EB1b-GFP localization prior to the symmetric division in mus.  EB1b-GFP localization in  meristemoids (I) and  young guard mother cells (J).  EB1b-GFP does not localize to end wall thickenings, but does localize to the band which forms between the in mature guard mother cells (K). Intense EB1b-GP localization is visible in the phragmoplast of dividing guard mother cells (L). (M- P)  proNEK6:NEK6-GFP localization prior to the symmetric division in mus.   NEK6-GFP localization in to meristemoids  (M-N).  NEK6-GFP localizes to the periphery in young guard mother cells (N), to the end wall thickenings  in mature guard mother cells (O) and to the phragmoplast in dividing guard mother cells (P).     In contrast to these general distribution patterns in wild-type, after the symmetric guard mother cell division in mus, a subset of young guard cells do not exhibit the strong GFP-CLASP, GCP2-GFP, NEK6-GFP and EB1b-GFP foci that are normally located in the centre of the ventral wall in young wild-type stomata.   Instead, in some young mus stomata, a variety of delocalization effects occur, including the skewing of the markers around the future pore site, or a shift of the markers towards the top or bottom of the stoma (in periclinal views) (Fig. 2.4 B-C, F-G, J-K, N-O).  Also, in contrast to stomata that are wild-type, or those in mus that resemble the wild-type, distinguishing young from mid-stage guard cells is difficult, since fan-like microtubule arrays around forming pores were not often 34  detected in mus stomata.  Mature mus stomata display a range of delocalization effects for GFP-CLASP, GCP2-GFP, NEK6-GFP and EB1b-GFP,  which include the skewing in the positioning of  each of the reporter proteins around the mature pore site, or their movement to the top or bottom edges of the cell (in periclinal views). Mature mus stomata also display a range of shape defects, such as skewed stomata and stomata in which one or both guard cells are abnormally expanded. GFP-CLASP and GCP2-GFP also localize discontinuously to the stomatal periphery in mature mus stomata.   As in wild-type stomata, NEK6-GFP localization was not observed at the pore site or the periphery in mature mus stomata (Fig 2.4 D, H, L, P).  The extent of this peripheral localization does not appear to differ from wild-type stomata.    35    Figure 2.4 The mus mutation alters GCP2, CLASP, EB1 and NEK6 localization after the symmetric division Green fluorescence represents proGCP2:GCP2-GFP (A –D),   proCLASP:GFP-CLASP  (E-H), pro35S:EB1b-GFP  (I – L) and proNEK6:NEK6-GFP (M-P) respectively.  Young Guard Cells 36  (GCs)     Mature stomata (mature GCs).  All scale bars are 5 µm. Arrows mark locations of delocalized GCP2-GFP, GFP-CLASP, EB1b-GFP or NEK6-GFP aggregates  (A – D) proGCP2:GCP2-GFP localization post- symmetric division in mus.  Young, recently divided, guard cells display GCP2-GFP localization along the new ventral wall (A).  Young guard cell displaying  GCP2-GFP foci which have shifted to one end of the cell (B) Young guard cell displaying  skewed GCP2 foci (C).   Mature, abnormal stomata displaying ectopic GCP2-GFP foci at the top edge of the guard cell (D)  (E - H) proCLASP:GFP-CLASP localization post- symmetric division in mus.  GFP-CLASP localizes along the new ventral wall in young guard cells (A).  Delocalization of GFP-CLASP foci, including shifting of the GFP-CLASP foci to the bottom edge of the cell (F) and skewing of the foci (G) are present in a sub-set of developing mus guard cells.  Mature mus stomata displaying abnormal pore and ventral wall skewing, as well as GFP-CLASP foci skewing (H).  (I - L)  pro35S:EB1b-GFP  localization post- symmetric division  in mus.  In, young, recently divided guard cells EB1b-GFP localizes along the new ventral wall (I).  A subset of developing mus guard cells exhibit delocalized EB1b-GFP foci, including shifting of the EB1b-GFP foci to one end of the stoma (J) and retention of the EB1b-GFP foci at the former site of the end wall thickenings  (K).  Mature, abnormal, mus stoma displaying ectopic EB1b-GFP foci and pore opening at the top edge of the stoma (L).  (M- P)  proNEK6:NEK6-GFP localization post- symmetric division.  NEK6-GFP localizes along the new ventral wall of recently divided guard cells (M).  NEK6-GFP foci display various delocalization defects in a subset of mus stomata, including skewing of the NEK6-GFP foci (N) and delocalization of the NEK6-GFP foci to one end of the stoma (O). Mature mus stomata do not display NEK6-GFP localization.   2.2.4 Morphogenetic defects in mus take place after CLASP, GCP2, and NEK6 and EB1 become delocalized during the formation of the pore site  Recent studies demonstrate the importance of post-symmetric division stages in regulating the final bilateral symmetry of stomata.  To date, only two genes have been shown to regulate the generation of stomatal bilateral symmetry, SCAP1 and MUS, and both appear to act after the symmetric division (Keerthisinghe et al., 2015; Negi et al., 2013).  Thus, time-lapse analysis of a sample of 3-5 stomata in MUS mutants was used to monitor the stages following symmetric divisions to define when the first defect occurs in symmetry disruption.  37   Preliminary observations of developmental stages closely following the symmetric division indicated that a time window of approximately 24 hours following the symmetric division was sufficient for defining the functions of MUS as well as details of pore site formation.  Time-lapse images of wild-type and mus stomata were acquired at 2-2.5 h intervals post-symmetric division over about a 24 hour period.  Early-stage guard cells (which had just undergone a symmetric division) were identified by either the presence of phragmoplasts, partially formed ventral walls, the presence of a very thin and young ventral wall, or the small size of young, unexpanded guard cells just after division.    The symmetric division of wild-type stomatal precursors (Guard Mother Cell stage) occurs within 2-2.5 hours of phragmoplast formation. Shortly after this division, the localizations of GCP2-GFP, GFP-CLASP, EB1b-GFP and NEK6-GFP are often weak and discontinuous along the newly formed (ventral) wall.  All four markers also localize in discontinuous puncta around the mid-plane of the stomatal periphery, when viewed in paradermal section (Fig. 2.5 A, G, N, T).  About 2- 2.5 hours after symmetric division, GCP2-GFP, GFP-CLASP, EB1b-GFP and NEK6-GFP concentrate into spherical aggregations in the top (ventral) and bottom (dorsal) cortices of the of developing stoma (Fig. 2.5 B, H, O, U).   The markers then retain their centrally located distributions during pore and stomatal maturation (a period that lasts about 12 - 15 hours) (Fig. 2.5 C-E, I-J, P-R, V-X).  NEK6-GFP is not present in mature stomata.  However, GCP2-GFP, GFP-CLASP, and EB1b-GFP remain centrally-localized in mature stomata, though at a lower intensity compared to younger stages (Fig 2.5 F, L, S, Y).  This reduced intensity might result from sample bleaching, since micrographs of mature stomata from samples that did not undergo time lapse analysis displayed a high intensity of fluorescence.     38     39   Figure 2.5  Stomatal microtubule organizing center formation occurs within 2.5 hours of the symmetric division in wild-type stomata.  Green fluorescence represents proGCP2:GCP2-GFP ( A-F) and proCLASP:GFP-CLASP (G – L), pro35S:EB1-GFP (N – S) and proNEK6:NEK6-GFP (T-Y) respectively.    All scale bars are 5 µm.  Each row contains three images : (1) the original micrograph.  (2) A tracing of the original micrograph - dark green represents GFP distribution : light green represents pore location  (3) Intensity plot – X-axis depicts the length ventral wall in µm (X –axis) ; Y-axis depicts the corresponding  fluorescence intensity values.  Left most part of plot represents the top of the stoma ; Rightmost part of the plot represents the bottom of the stoma.  Blue lines depict raw intensity values ; red lines represent a moving average of the raw values. (A-X) Post-symmetric division localization of GCP2, CLASP, EB1 & NEK6 ( t =hours) 40  (A-F)   GCP2-GFP localization  :   t = 0.0 hours - GCP2-GFP localizes along the newly formed ventral wall  s.  Corresponding Intensity plots show a wide intensity distribution (A).   t = 2.0 hours - GCP2-GFP forms a tighter focus in the centre of the ventral wall, as indicated by narrower peak intensities in the intensity plot (B).  After GCP2-GFP foci formation, GCP2-GFP remains centrally localized, as indicated by the intensity plot peaks.  Slight bleaching from consecutive exposures to the laser is apparent and likely causes the noise represented by the smaller peaks in D and E (C-E).  t= 24 hours -  a symmetrical mature wild-type stoma is formed.  The lack of a distinct focus in the intensity plot is likely due to sample bleaching. The strong and continuous fluorescence present around the pore is auto-fluorescence (F).  (G-L) GFP-CLASP localization  : t = 0.0 hours – GFP-CLASP localizes along the newly formed ventral wall.  Corresponding intensity plots show a wide intensity distribution (G).    t = 2.5 hours - a tight CLASP distribution pattern has formed in the centre of the ventral wall, as indicated by narrower peak intensities (H). Following GFP-CLASP foci formation, GFP-CLASP remains centrally localized, as indicated by the intensity plot peaks (I-K).  t= 24 hours - a symmetrical, mature wild-type stoma is formed.  As the sample did not bleach, a distinct peak representing the GFP-CLASP foci is visible in the intensity plot (L).   (M-R) EB1b-GFP localization : t =0.0 hours – EB1b-GFP localizes along the newly formed ventral wall.  Corresponding intensity plots show a wide intensity distribution (M).  t = 2.5 hours - EB1b-GFP has formed a tight focus in the centre of the ventral wall, as indicated by narrower peak intensities (N).  Post formation of EB1 foci, EB1 remains centrally localized, as indicated by the intensity plot peaks (O-Q).  t=24 hour - formation of a wild-type, symmetrical, mature stoma.  As the sample did not bleach, a distinct EB1b-GFP focus is visible in the intensity plot peak (R).   (S-X) t = 0.0 hours - NEK6-GFP localizes along the newly formed ventral wall.  Corresponding Intensity plots show wide intensity distribution (S).   t = 2.5 hours - NEK6-GFP forms a tighter focus in the centre of the ventral wall, as indicated by narrower peak intensities.  NEK6-GFP localization was weak, background fluorescent signals (chloroplast auto-fluorescence) are much more visible in NEK6 lines.  This background noise may have caused the smaller intensity peak visible in the intensity plot (T - U).   Post formation of the NEK6-GFP foci, NEK6-GFP remains centrally localized, as indicated by the intensity plot peaks (V).  t = 24.0 hours - formation of a mature, symmetrical, wild-type stoma.  NEK6-GFP is not expressed in mature stomata, and the strong continuous fluorescence visible around the pore is auto-fluorescence (X).   41   In the mus mutant, the symmetric division of the guard mother cell occurs within 2- 2.5 hours of phragmoplast formation, a period comparable to wild-type stomata.   As in wild-type stomata, GCP2-GFP, GFP-CLASP and EB1b-GFP also localize unevenly along the newly formed ventral wall (Fig. 2.6  A, G, M, S).  But in contrast to the wild-type, approximately 2-2.5 hours after the symmetric division, GFP-CLASP, GCP2-GFP and EB1b-GFP in mus do not form spherical aggregates in the centre of a sub-set of mus stomata.  Instead, all three markers exhibit varying but overall comparable degrees of mis-localization (Fig. 2.6 B-E, H-K, N-Q, and T-V) that can be broken down into several categories in mus.  Ultimately the type of mis-localization influences the final shape of the mature mus guard cell.    Three main categories of mis-localization were observed in mus.  The categories were largely based on observations from GFP-CLASP, EB1b-GFP and GCP2-GFP time-lapse data sets, as rapid bleaching often made prolonged observation of NEK6-GFP difficult (Fig. 2.7).  In Category 1 stomata, the foci of all three MTOC markers (CLASP-GFP, EB1b-GFP and GCP2-GFP) were skewed around the future pore site.   Although the symmetric division itself was initially straight, the skewing of MTOC markers around the future pore site in young mus guard cells eventually leads to the ventral wall itself becoming skewed, a twisting retained in mature mus stomata.   In Category 2 mus stomata GFP-CLASP, EB1b-GFP, and GCP2-GFP move to either the top or bottom edge of the ventral wall (Fig. 2.11).  Here, the initial symmetric division of the guard mother cell, as well as guard cell shape, are normal.  However, the shifting of the positions of GCP2-GFP, EB1b-GFP, and GFP-CLASP, from the centre of the stoma to the top or bottom edge of the ventral wall in mus, results in the formation of a pore at just one edge of the stoma.  This, in turn, induces the formation of an irregularly expanded, abnormal, and roughly heart-shaped guard cell.   In Category 3 mus stomata GFP-CLASP, EB1b-GFP, and GCP2-GFP are primarily localized at the former site of the guard mother cell end wall thickenings.  Here CLASP and GCP2 do not re-localize to the ventral wall, and as these stomata mature, they develop irregularly expanded, roughly cylindrically shaped, guard cells. These time-lapse analyses show that MTOC components become delocalized in mus within 2-2.5 hours after symmetric division formation.    42         43    Figure 2.6  Microtubule organizing center formation does not occur with 2.5 hours of the symmetric division in mus stomata.   Green fluorescence represents proGCP2:GCP2-GFP ( A-F) and proCLASP:GFP-CLASP (G – L), pro35S:EB1-GFP (N – S) and proNEK6:NEK6-GFP (T-Y) respectively.    All scale bars are 5 µm.  Each row contains three images: (1) the original micrograph;  (2) a tracing of the original micrograph (dark green represents GFP distribution and light green represents pore location); (3) intensity plot- (X –axis) the length ventral wall in µm (X –axis) : (Y-axis)  the corresponding fluorescence intensity values.  Leftmost part of the plot represents the top of the stoma, while the rightmost part of the plot represents the bottom of the stoma.  Blue lines depict raw intensity values, while red lines represent a moving average of the raw values. 44  (A-X) Post-symmetric division localization of GCP2, CLASP ( t = hours).   (A-F)   t =0.0 hours - GCP2-GFP localizes along the newly formed ventral wall.  Corresponding intensity plot displays wide elongated GCP2-GFP distribution (A).   t = 2.0 hours - GCP2-GFP does not form a focus in the centre of the ventral wall, but appears to shift to the bottom of the stoma.  This shift is also apparent in the corresponding intensity plot (B).   GCP2-GFP localization continues to shift towards the bottom of the stoma, which is again mirrored by continuing, right-ward shift of peak intensities in the intensity plots.  Although the GCP2-GFP foci is shifting, the young guard cell does not display morphogenesis defects.(C – E).  t = 24.0 hours - the development of an abnormal, ‘heart-shaped’ mus stoma.  The abnormal stoma displays ectopic GCP2-GFP localization, as well as an ectopic opening (tear) at the bottom of the stoma (F). (G-L) t =0.0 hours – GFP-CLASP localizes along the newly formed ventral wall.  Corresponding Intensity plots exhibit wide CLASP distribution (G).   t = 2.5 hours – GFP-CLASP has not formed a distinct focus, but appears to have shifted towards the top of the stoma (H).  t = 5.0 hours - the fluorescence appears to be absent from the ventral wall, and instead localizes on either side of the ventral wall at the top of the stomata (I).  GFP-CLASP re-localizes to the top half of the ventral wall and remains localized to the top half of the stomata (J-L)  24 hours - the formation of an abnormal heart shaped stomata, with an opening towards the top of the stoma.  The GFP-CLASP foci remains at the top edge of the young guard cell, as indicated by the leftward shift in the intensity plot (L).   (M-R) t = 0.0 hours -  EB1b-GFP localizes along the newly formed ventral wall at.   t = 2.5 hours - EB1 foci appear to skew, as indicated by the presence of two distinct peaks of approximately equally size the intensity plot (N-O)  t = 5.0 h - EB1b-GFP foci appear to move closer together, as is also indicated by a smaller separation between the peaks in the intensity plots (Q).  t=24 h-  the formation of an abnormal stomata exhibiting  skewed ventral wall, a skewed pore, and skewed EB1b-GFP foci. The skewed EB1b-GFP foci is corroborated by the presence of two clear intensity peaks in the intensity plot (R)  (S-X) t = 0.0 and t = 2.0 hour NEK6-GFP  localizes along the ventral wall.  Correspondingly Intensity plots show a wide/elongated NEK6 distribution (S-T).  After the t=  2.0 hour time-point distinct NEK6 fluorescence appeared to disappear.   NEK6 localization in mus was difficult to observe due to the faintness of the marker and rapid bleaching upon multiple laser exposures.   Large fluorescent structures do not represent NEK6 foci, but are instead auto-fluorescent chloroplasts.  The high levels of auto-fluorescent background noise may have caused the smaller intensity peak visible in the intensity plot (U-V).   t=24 hour the formation an abnormally elongated, pore-less stomata.   An aggregation of auto-fluorescent chloroplasts near the ventral wall likely cause the fluorescence peak in the intensity plot (X). 45   Figure 2.7  Common categories of microtubule organizing center delocalization in mus.   Green circles represents MTOC localization.   A.  In wild-type stomata MTOCs localize to the end wall thickenings  in guard mother cells, which is followed by MTOCs localization along the ventral wall in recently divided guard cells.  MTOCs then form a focus at the centre of the ventral wall in developing guard cells.  Subsequently, a mature stoma, exhibiting a symmetrically placed pore surrounded by symmetrically placed MTOCs, is formed. B.  In Category 1 mus stomata, MTOCs localize to end wall thickenings  and along the recently formed ventral wall.  However, instead of forming a focus at the centre of ventral wall, the MTOCs fail to position themselves symmetrically and skew along the ventral wall, resulting in the formation of a mature stoma with a skewed pore, ventral walls and MTOCs. 46  C. In Category 2 mus stomata, the MTOCs form at the end wall thickenings  and along the recently formed ventral wall.  However,  the MTOCS then shift towards the top or bottom of the stoma (bottom-ward shift shown) resulting in a depression or partial pore forming at the site of MTOC delocalization, resulting in a heart shaped mature stomata. D.  In category 3 mus stomata, the MTOCs do no localize to the ventral wall and instead remaining at the site of the end wall thickenings, resulting in an abnormally expanded mature stoma, which lacks a pore.  2.2.5 mus is associated with altered microtubule polarity as indicated by the direction of microtubule growth  Mature wild-type stomata exhibit an outbound microtubule growth polarity, where the majority of microtubules grow outwards from the pore to the stomatal periphery.  In contrast, mature mus stomata exhibit an altered microtubule growth polarity, where a significant percentage of microtubules display inbound growth (microtubules growth from the stomatal periphery inwards towards the pore) (Keerthisinghe et al., 2015).  Short-term time-lapses (captured over a two hour period starting just after the occurrence of a symmetric division) of microtubule dynamics were obtained  in order to determine whether the microtubule polarity defects observed in mus are a consequence of MTOC displacement.     The direction of microtubule growth, as indicated by the direction of EB1 comet movement, was categorized as either outbound (away from ventral wall) or inbound (towards ventral wall).  A count of outbound and inbound comets was taken from a total of 12 random comets from each of 5 randomly selected wild-type and mus stomata at 0, 30, 60, 90, and 120 minute intervals following the completion of the symmetric division (Fig  2.8  A-D). The proportion of outbound microtubules was significantly higher in wild-type compared to mus as shown by 2 way ANOVA of raw data (F =36.67; p < 0.05), as well as arcsine-transformed data (F=35.91 p <0.05), indicating that the mus mutation may be significantly associated with a lower outbound microtubule growth.  However, no significant difference in the proportion of outbound microtubules between the 5 different time points was found for raw ( F = 0.3099 : p > 0.05 ) or arcsine-transformed ( F = 0.3571 : p > 0.05) 47  proportions.  Thus, although the proportion of outbound microtubules was higher in wild-type compared to mus, the change in the proportion of outbound microtubules over time was not significantly different for either wild-type or mus, indicating that time may not be a significant factor regulating the direction of microtubule growth in mus in the first 2 hours following the symmetric division. The interaction between time and genotype was also not significant for both raw ( F = 1.117  p= 0.3419 ) and arcsine-transformed ( F=0.9143 : P > 0.05) proportions indicating that time and genotype did not influence each other in regard to change in the direction of microtubule growth .  Chi Square Contingency tables for direction of microtubule growth (outbound and inbound) vs time ( 0, 30, 60, 90 and 120 minutes after symmetric division) indicate that for both wild-type (Chi=2.4854 : P > 0.05)  and mus ( chi=2.8579 : P > 0.05), the direction of microtubule growth was independent of time, which h confirmed the findings of ANOVA that time had no significant effect on the direction of microtubule growth.  However, 2x2 tables for direction of microtubule growth vs genotype for each time point indicated that direction of microtubule growth was dependent on genotype for all time points (0 min (Chi 4.48 ;p<0.05), 30 min (chi 10.47 ;p<0.05), 90 min ( chi =13.37;p<0.05) and 120 min (chi =11.25; p<0.05) except the 60 minute time point (chi = 1.7759 : p> 0.05) , which again confirms the results of the ANOVA analysis that direction of microtubule growth was significantly associated with genotype, indicating that the mus mutation is significantly associated with change in the direction of microtubule growth.          As outbound to inbound movement of microtubules can be considered a binary variable, the probability ratio of an outward bound microtubule and an inward bound microtubules (referred to as the odds of finding an outbound microtubule) as the dependent variable was regressed on genotype (wild-type=0 : mus=1) and time (0, 30, 60, 90 and 120 minutes) as Independent variables, using the Logistic Regression program provided in the Real Statistics and solver packages in Excel. Once again, confirming the results of ANOVA and Chi Square analyses, the regression of the odds of finding an outbound microtubule on time was not significant but the regression of the odds of finding an outbound microtubule on genotype was highly significant (b = -1.0737 : P < 0.01), 48  indicating that mus has about a 34% probability of outbound microtubule growth compared to wild-type. The odds ratio between wild-type and mus was 1:0.342 indicating that the odds of finding an outbound microtubule in mus is approximately 0.34, or 1/3rd, of finding an outbound microtubule in wild-type.           The time-series trend line for the ‘outbound to inbound microtubule’ ratio indicated that wild-type started with a higher outbound to inbound microtubule ratio than mus immediately following the symmetric division (at 0 mins), and continued to maintain a higher outbound to inbound microtubule ratio than mus for the next 120 minutes (Fig 8E). In addition, the outbound to inbound microtubule ratio for wild-type displayed a slight increasing trend through the 120 min period following the symmetric division as opposed to mus, which displayed a slight decreasing trend.           Similarly, time series trend-lines of the proportion of outbound microtubule growth indicate that wild-type cells started with a higher proportion of outbound microtubules than mus cells immediately following the symmetric division (at 0 mins) and continued to maintain a higher proportion of outbound microtubules than mus for the next 120 minutes. In addition, the proportion of outbound microtubules in wild-type displayed an increasing trend through the 120 min period following the symmetric division as opposed to mus which displayed a decreasing trend.               Although neither the regression nor the correlation statistics were significant for the ratio or the proportion data, these results indicate that, immediately following the symmetric division, the number of outbound microtubules in mus start lower than that of wild-type and gradually decreases over time.   A plot of the outbound to inbound microtubule ratio of wild-type vs that of mus indicates that, although the regression is not significant (F = 1.241 : P > 0.05) and there was no significant correlation ( r=- 0. 541 : P> 0.05) a trend can be observed where for every one unit increase in the outbound to inbound microtubule ratio of wild-type, there is a corresponding decrease of approximately 0.16 (16%) in the outbound to inbound ratio of mus.  These findings reinforce the suggestion that the mus mutation causes a lower outbound to inbound microtubule ratio than wild-type from the very onset of guard cell 49  morphogenesis (0 minutes after the symmetric division), and that this ratio continues to remain lower with a gradually decreasing trend through the next 120 minutes of guard cell morphogenesis.               While both wild-type and mus display a higher proportion of outbound microtubules compared to inbound microtubules during early guard cell morphogenesis stages (the first 90 minutes after the symmetric division), the proportion of outbound microtubules in mus stomata, ranging from 50 to 58%, is only marginally higher than the break-even point of 50%, indicating that the number of outbound microtubules is only just slightly higher than the number of inbound microtubules.  By contrast, in wild-type that proportion, ranging from 70 to 80%, is much higher, indicating that a much higher number of microtubules in wild-type are outbound than inbound.                                                                                                         Also, as opposed to wild-type, which displays a steady proportion of outbound microtubules with time  (b= 0.0001), the outbound proportion in mus shows a minute decreasing trend with time (b=-0.0008), which ends in a higher inbound proportion of microtubules around 120 minutes after the symmetric division, indicating that an actual switch from a higher number of outbound microtubules to a higher number of inbound microtubules may occur in mus around 120 minutes following a symmetric division.           Time series plot results indicate that in contrast to wild-type stomata, in which the proportion of outbound microtubules remain steadily higher than inbound microtubules from early guard cell stages up to the mature guard cell stage, mus stomata start off with a marginally higher proportion of outbound microtubules but undergo a switch in direction around 2 hours after the symmetric division and end up with a higher proportion of inbound microtubules (Fig 2.8 E).   50    Figure 2.8 The mus mutation alters the direction of microtubule growth in developing guard cells, which may result in a switch from outward microtubule growth to inward microtubule growth.   Green fluorescence represents pro35S:EB1-GFP localization.  All scale bars are 5 µm. (A-D) The direction of microtubule movement in developing mus and wild-type guard cells were observed short term time-lapse imaging (120 minutes). Only t = 0 min and t=120 min time-points are shown for wild-type (A, C) and mus (B, D).  Randomly selected microtubules, marked by pro35S: EB1b-GFP were categorized as either outward growing (yellow) or inward growing (red).   The data collected was utilized to construct a time-series plot (E), where blue points represent the ratio of OUT:IN movement in wild-type (X-axis) plotted against the ratio of OUT:IN movement in mus (Y-axis) for each of the five time-point observed, while the red line represents the line of best fit.  Raw values used to create the trend-line are visible in the table nested in the lower left hand corner of the time-series plot.   The time-series analysis reveals that mus guard cells display a slight trend towards a lower OUT:IN ratio (more microtubules moving inward than outward) at the 120 min time-point, which roughly corresponds to the developmental stage at which the MTOCs delocalize in mus.  51  2.2.6 The actin related protein BRICK1 localizes along nearly the entire length of the ventral wall, except for the pore site  Pavement cell morphogenesis is largely governed by antagonistic interactions between the actin filament (AF) and microtubule cytoskeletons (Fu et al., 2005).  The possible existence of a similar mechanism in guard cell morphogenesis was explored by characterizing the distribution of BRICK1 (BRK1), a component of the SCAR/WAVE AF nucleating complex, in wild-type and mus stomata.    In wild-type stomata, BRK1-YFP (proBRK1:BRK1-YFP) localizes to junctions between cells in the early stages (meristemoid to mid-stage guard mother cells) of stomatal development.  BRK1-YFP does not localize to the phragmoplast during cytokinesis of the guard mother cell.  However, just after the symmetric division, BRK1-YFP localizes weakly along the entire length of ventral wall.  Time-lapse analysis demonstrated that approximately 2-2.5 hours after a symmetric division, BRK1-YFP is excluded from the pore site and instead localizes intensely along the ventral wall on either side of the pore.  BRK1-YFP remains on either side of the pore in mature guard cells, where it is also present in dispersed punctae in the cortical plasma membrane (Fig 2.9 A - F).       In mus, BRK1-YFP also appears to localize to the ventral wall just after a symmetric division.  In developing mus guard cells where the pore site is displaced, the domain of BRK1-YFP takes up the space that remains after the pore delocalizes (Fig 2.9 M-R).   52   Figure 2.9 The actin related protein BRK1 is excluded from the pore site during microtubule organizing center development.   Yellow fluorescence represents proBRK1:BRK1-YFP localization. All scale bars are 5 um.     Yellow fluorescence represents proBRK1:BRK1-YFP localization. All scale bars are 5 µm.     (A- F) BRK1-YFP in wild-type.  At t = 0 h weak BRK1-YFP localization is visible along the ventral wall (A).  Within 2.5 hours of the symmetric division BRK1-YFP is excluded  from the pore site and localizes intensely to the ventral wall on either side of the forming pore (B).  BRK1-YFP ventral wall localization is maintained in both the developing and mature stoma  (C-F).    (G-L) BRK1-YFP in mus.  Weak BRK1-YFP ventral wall localization is visible in the developing  guard cell (G).  Although movement of BRK1-YFP away from the pore site is visible from  t = 2 to 8 h,   this movement appears to occur less rapidly  than  in wild-type stomata  at comparable stages of development.  The ventral wall remains  symmetrically placed, and the developing guard cell does not display morphogenesis defects (H-K).   t = 24 h time-53  point demonstrating that the mature stomata displays a skewed pore and ventral wall (outlined by brackets) (L). (M-R) BRK1-YFP in mus.  BRK1-YFP localizes along the ventral wall in recently divided guard cells, t = 0 h, (M).   BRK1-YFP localization appears to have shifted towards the top of the stoma (N-P).  The leaf being observed has shifted such that the entire stoma is not visible at t = 8.5 h.  However, BRK1 localization to the top of the stoma is still visible (Q).   Mature stoma displaying an ectopic pore at the bottom of the pore (arrow), from which BRK1 was excluded during guard cell development.   BRK1 is no longer visible at the ventral wall (R.)  2.2.7  Double mutants of mus and mutants of the microtubule organizing center markers clasp or nek6 display additional stomatal defects  The distributions of the MTOC-associated proteins (GCP2-GFP, GFP-CLASP, NEK6-GFP and EB1b-GFP) as well as BRK1-YFP are altered in mus, suggesting that these proteins may represent potential targets of the MUS signaling pathway. Single and higher order mutants (double and quadruple mutants with mus) were generated to establish whether these proteins are MUS targets.      Stomatal morphogenesis is normal in single mutants of gcp2-1, clasp-1, nek6-1, and brk1. Similarly, stomatal morphogenesis in the eb1a-1 eb1b-1 eb1c-1 triple mutant is also normal.  We found that a putative quadruple mutant that harbors eb1a-1 eb1b-1 eb1c-1 and mus, as well as mus brk-1 and mus1 gcp2-1 double mutant, were additive, since they exhibit traits from both respective parental lines (Fig 2.10 G -J).   In contrast, mus clasp-1 and mus nek6 double mutants appeared to display a more severe stomatal phenotype than would be predicted by independent combination of the parental line phenotypes.  More ‘capsule’-shaped stomata were observed in first leaves of the double mutants as compared to mus.  Capsule-shaped stomata retain their bilateral symmetry, but contain abnormally expanded guard cells, which appear to lack a pore (Fig 2.10  C-D, E-F ; Fig 2.11 C-D).  Despite the increase in severity of the stomatal phenotype, double mutant plants resembled clasp-1 in their over-all stature and growth habit.    The mus-1 nek6-1 double mutants display traits observed in both mus (skewed stomata, or stomata with abnormal morphogenesis) and nek6-1 (hypocotyl protrusions)  54   55  Figure 2.10  mus clasp and mus nek6 mutants display synergistic phenotypes which generate a high abundance of ‘capsule-shaped’ stomata.   Cell walls were visualized by propidium iodide staining.  All images have been gray scaled and inverted.  All scale bars are 20 µm.  (A) wild-type cotyledon displaying symmetrical stomata (B) mus cotyledon displaying stomata with abnormal morphogenesis  (C)  clasp cotyledon displaying symmetrical, wild-type stomata (D) mus clasp stomata displaying ‘capsule-shaped’ stomata (E)  nek6 stomata displaying symmetrical, wild-type stomata (F) mus nek6 stomata displaying ‘capsule-shaped stomata (G) gcp2-1 mutant displaying symmetrical, wild-type stomata (H) gcp2-1 mus mutants display both mus and wild-type stomata, but do not display ‘capsule-shaped’ stomata. (I) eb1 triple mutant displaying symmetrical, wild-type stomata (J) eb1 triple mus putants display both mus and wild-type stomata, but do not display ‘capsule’ shaped stomata   single mutants.   However, these double mutants also exhibit a phenotype that is much more severe than either parental line singly. Double mutant seedlings (at about 15 DAG), were severely dwarfed compared to the wild-type, as well as each of the mus, and nek6-1 single mutants from seedlings of an equivalent age.  Once planted on soil, the mus nek6-1 seedlings survived, and produced small stems and siliques.  However, the double mutant seedlings did not produce any viable seeds.   In mus nek6-1 cotyledons, almost all stomata (98.8% ± SE) displayed abnormal morphogenesis (n=11 cotyledons, 5 fields per seedling).  Of the 98.8% of abnormal stomata, 93.7% ± SE exhibit an elongated ‘capsule-like’ shape and appear to lack a functional pore.  The remaining abnormal stomata in this double mutant resemble ‘mus-type’ stomata that 56  are either skewed, or exhibit abnormal morphogenesis, such as heart-shaped or abnormally expanded guard cells).  The majority of the ‘mus-like’ stomata appeared to contain partial pores (Fig 2.11 G).      Three-dimensional (3D) reconstructions were performed, in order to verify whether the ‘capsule’ stomata observed in mus1 clasp1 and mus1 nek6-1 truly lacked a pore. The 3D reconstructions confirmed that the capsule-like stomata lacked centrally positioned pores on the ventral wall.   Interestingly, many of the capsule-like stomata contained ectopic wall thickenings, or papilla, located on the dorsal surface exposed to the environment in each guard cell.  Many of these papillae exhibited pseudo-pores, which do not extend through the depth of the guard cell (Fig 2.11 E-F).   Figure 2.11 Quantification of mus nek6 and mus clasp double mutants Cell walls were visualized by propidium iodide staining.  All images have been gray scaled and inverted (A-D).  All scale bars are 20 µm.  57  (A-D) The mus nek6 double mutant epidermis consists almost entirely of capsule shaped stomata.  wild-type epidermis  (A) mus epidermis (B)  nek6 epidermis (C) and mus nek6 epidermis (D) (E-F)  Maximum Z-projections of capsule shaped stomata in mus nek6 double mutants.  Though these mutants lack pores, they display ectopic wall thickenings (papillae) on their dorsal (exposed) surfaces (E).  Papillae often also exhibit breaks, or openings (pseudo-pores)  (F). (G)  Quantification of stomatal phenotypes in wild-type, mus, mus clasp and mus nek6 double mutants in cotyledons (C) and Rosette leaves (R). mus clasp rosette leaves, as well as mus nek6 cotyledons and rosette leaves contain a large number of stomata, which lack a pore (capsule shaped),  compared to the mus single mutant.    2.2.8 GCP2 does not localize to microtubule organizing centers in ‘capsule’ shaped stomata of mus nek6 and mus clasp double mutants   proGCP2:GCP2-GFP was introgressed into mus clasp and mus nek6 double mutant backgrounds to investigate the effect of the double mutants on MTOC formation.    GCP2-GFP localization in the double mutant resembled that of wild-type stomata during early stages of development (meristemoid to guard mother cell).  In contrast, GCP2-GFP localization differed greatly from wild-type after the symmetric division stage, as the majority of recently divided guard cells in the double mutant did not exhibit GCP2-GFP along the ventral wall.  Further, GCP2-marked foci were largely absent in young guard cells.   Subsequently, the mature ‘capsule’-shaped stomata in double mutants also lacked GCP2 foci in the centre of in the ventral wall.   Despite the absence of ventral wall associated GCP2-GFP in the double mutant, GCP2-GFP did localize weakly to discontinuous punctae in the periphery of young guard cells and mature stomata.   Intriguingly, when compared to wild-type and mus lines, the intensity of GCP2-GFP fluorescence appeared to be noticeably lower in double mutants (Fig. 2.12). 58   59  Figure 2.12 Microtubule organizing center foci do not form in mus clasp and mus nek6 double mutants.   Green fluorescence represents proGCP2:GCP2-GFP localization.  Red boxes in C and D denote stomata that have been enlarged in E-H.  Scale bars are 20 µm (A – D) and 5 µm (E-H).   (A) GCP2-GFP localization in the wild-type epidermis.  Intense GCP2-GFP localization can be observed in the MTOC of developing stomata (arrows) and stomatal precursors (asterisks). (B) GCP2-GFP localization in the mus epidermis.  Bright GCP2-GFP localization can be observed in both developing stomata (arrows) and stomatal precursors (asterisks) (C)  GCP2-GFP localization in mus clasp mutant.  Intense GFP localization cannot be observed in ‘capsule’ stomata (red boxes).  However, localization in stomatal precursors is maintained (asterisks) (D) GCP2-GFP localization in mus nek6 mutant.  Intense GFP localization cannot be observed in ‘capsule’ stomata (red boxes).  However, localization in stomatal precursors is maintained (asterisks).  As well some recently divided guard cells appear to exhibit GCP2-GFP ventral wall localization. (E-F) GCP2-GFP localization in mus clasp mutant.  Micrographs are enlarged images of the boxed region in C (marked with an E or F).  Although MTOCs do not form in capsule stomata, a few punctae of GCP2-GFP are visible along the ventral wall, and a small number of GCP2 punctae are visible in the stomatal periphery. (G-H) GCP2-GFP localization in mus nek6 mutant.  Micrographs are an enlarged image of the boxed region in C (marked with a G or H ).  A small number of discontinuous GCP2-GFP punctae are visible along the ventral wall.  GCP2-GFP does not appear to localize to the periphery in (G), some peripheral GCP2-GFP is visible in (H).  2.3 Discussion  2.3.1 GCP2, CLASP, EB1 and NEK6 localize to the cortical microtubule organizing centers in wild-type stomata The overlap in distribution patterns between GCP2-GFP, GFP-CLASP, EB1b-GFP and NEK6-GFP is especially apparent after a symmetric division occurs, as all four proteins 60  aggregate into circular foci.  These foci appear to predict the site of the stomatal MTOC, as well as where pore formation will occur in the mature stomata.  Of the characterized proteins, GCP2-GFP localization is the most likely to reflect γ-TURC localization, as GCP2 is a member of the γ-TUBULIN SMALL complex (γ-TUSC).  γ-TUSCs form the core of the γ-TURC, and consist of γ-TUBULIN in direct contact with GPC2 and GCP3 (Kollman et al., 2011).  Additionally, proGCP2:GCP2-GFP concentrates around the mature pore site, thereby emulating previously observed γ -tubulin localization patterns detected through immuno-localization (Marc et al., 1989).  As such, GCP2 represents an ideal candidate for visualizing sites of γ -TURC aggregation, including the MTOC.  The overlap of GCP2-GFP distribution with GFP-CLASP, NEK6-GFP and EB1b-GFP distribution indicates that these three proteins may associate with stomatal MTOCs, and also that they can be utilized to visualize MTOC placement in stomata.   Organization of plant cortical microtubule arrays occurs through multiple mechanisms.  One recently characterized mechanism involves the nucleation of microtubules from MTOCs situated at the edges of newly divided, unexpanded, root and shoot epidermal cells (Ambrose and Wasteneys, 2011).   The presence of both CLASP and GCP2 at edge based MTOCs is necessary for the predominantly outward growth of microtubules (visualized through pro35:EB1b-GFP), from these MTOCs (Ambrose and Wasteneys, 2011).  Ambrose et al (2011) demonstrated that edge-based MTOCs dissipate as root epidermal cells mature.  MTOC dissipation may precipitate the formation of transverse microtubule arrays, thereby allowing for longitudinal root cell elongation.   In contrast to edge-based MTOCs seen in root and shoot epidermal cells, stomatal MTOCs remain localized to the ventral wall in mature stomata.  GCP2-GFP, GFP-CLASP and EB1b-GFP localize to mature stomatal MTOCs while NEK6-GFP does not, suggesting that NEK6 may only function during early stages of guard cell morphogenesis.  Although EB1b-GFP localizes to the MTOC, it is likely not required for MTOC function, and instead the MTOC may represent a loading site where EB1 attaches to growing microtubule (+) ends.  The radial microtubule array is essential for stomatal function, and the persistence of 61  MTOCs in mature stomata may be related to the importance of maintaining the integrity of the radial array.  2.3.2 MUS is likely responsible for placement of cortical microtubule organizing centers in stomata  Although mus mutants undergo normal symmetric divisions that produce young guard cells that initially exhibit bilateral symmetry, a sub-set of these guard cells also display mis-placed MTOCs.  Often the type of MTOC mis-placement predicts the category of stomatal defect observed in mature mus stomata.  For instance, long-term time lapse imaging showed that young guard cells exhibiting skewed MTOCs develop into mature stomata containing skewed ventral walls and pores. These results suggest that morphogenesis defects in mus occur as a consequence of the initial MTOC and microtubule-array defects.  Cortical microtubule arrays coordinate cell morphogenesis through their influence on cellulose microfibril (CMF) orientation, length and mechanical strength (Fujita et al., 2011).  Defective microtubule arrays in mus may adversely impact CMF organization and/or mechano-chemical properties, thus impacting the guard cells ability to successfully withstand turgor pressure, which would consequently lead to the guard cell deformations and morphogenesis defects present in mature mus guard cells.  Time-lapse imaging of GCP2-GFP, GFP-CLASP, NEK6-GFP and EB1b-GFP localization in wild-type stomata revealed that symmetrically placed MTOCs form in the centre of the new ventral wall within 2.5 hours of symmetric division.   Additionally, as in root cells, microtubules grow predominately away from stomatal MTOCs in both developing and mature wild-type stomata.  MTOCs in mus also form within 2.5 hours of symmetric division.   However, a subset of mus stomata display mis-localized MTOCs which either lack bilateral symmetry, or form away from the centre of the ventral wall.  Though MTOCs are misplaced in mus, the majority of abnormal mus stomata still appear to exhibit some form of an MTOC.  Consequently, MUS may have a primary function in MTOC placement and a secondary, less essential, role in MTOC formation.   Further studies are required to determine whether or not the misplaced mus MTOCs are completely functional.   62  2.3.3 mus is associated with altered microtubule polarity   In addition to delocalization of the MTOC, a further defect noticeable in mus stomata is the disruption of microtubule polarity, as indicated by changes in the direction of microtubule growth. Two Way ANOVA as well as CHI-SQUARE analysis indicate that the mus mutation maybe associated with a significant decrease in outward growth of microtubules compared to wild-type.  Additionally, time series plots indicate that mus may also be associated with a reversal, or switch, in the overall direction of microtubule growth from a higher outbound proportion to a higher inbound proportion at some stage in guard cell morphogenesis. Two Way ANOVA as well as CHI-SQUARE results indicate that time did not affect direction of microtubule growth in any significant manner which is also confirmed by time series results.  However, time series plots provided useful information which indicate that during the first two hours of guard cell morphogenesis mus is associated with a slight downward trend in the proportion of outbound microtubules which later transforms into a marginally higher proportion of inbound microtubule growth towards the end of the two hour period Starting immediately after the completion of the symmetric division, mus stomata display a significantly lower proportion of outbound microtubules compared to that of wild-type stomata. While both wild-type and mus display a higher proportion of outbound microtubules compared to inbound microtubules during early guard cell morphogenesis stages (the first 90 minutes after the symmetric division), the proportion of outbound microtubules in mus stomata is only marginally higher than inbound, whereas in wild-type that proportion is much higher than inbound. Also, in contrast to wild-type, which displays a steady trend in the proportion of outbound microtubules with time, the outbound proportion in mus shows a decreasing trend over time, which results in a higher inbound proportion around 120 minutes after the symmetric division. These results indicate that, in contrast to wild-type stomata in which the proportion of outbound microtubules consistently remains much higher than inbound microtubules from early guard cell stages up to the mature guard cell stage, mus stomata start off with a proportion of outbound microtubules that is marginally higher than that of inbound microtubules, and undergo a 63  switch in direction around 2 hours after the symmetric division and end up with a higher proportion of inbound microtubules.   Results of the present study appear to agree with those of previous studies (Keerthisinghe et al., 2015), which showed that when mus stomata were categorized according to the degree of phenotypic disruption as wild-type-like, mild, high and disrupted, mature stomata in the two highly disrupted stomatal phenotypes (high mus and disrupted mus), displayed higher proportions of inbound microtubules (57% & 80% respectively), whereas the two mildly disrupted categories (wild-type-like mus and mild mus) still displayed a higher proportion of outbound microtubules (70 and 61% respectively). Therefore it is possible that the switch from a higher outbound to a higher inbound proportion of microtubules may occur only in more severely disrupted mus stomata and not in mildly disrupted stomata. This could not be confirmed in the present study for two reasons: (1) all mus phenotypes were not separated into the four categories described above; (2) only the first 2 hours of guard cell morphogenesis were monitored due sample bleaching, whereas guard cell morphogenesis from start to mature guard cell stage requires approximately 12 hours. As a result, a precise stage at which the switch may occur could not be specified for each category. In general, however, the present study points to a switch from higher outbound to higher inbound around two hours after the completion of the symmetric division in mus.  Whether such a switch is common to all four mus stomatal phenotypes or confined only to the highly disrupted mus phenotypes requires further investigation. It is noted that time series trend-lines of outbound to inbound microtubule ratios as well as the proportion of outbound microtubules (both of which are indicators of microtubule direction) did not indicate any significant correlation between time and either the outbound to inbound ratios or the outbound proportions.  As indicated by ANOVA, CHI-SQUARE analysis and logistic regression, however, there was no significant relationship between time and microtubule direction during the first two hours of guard cell morphogenesis that was monitored. Hence, the absence of a significant correlation 64  between these indicators and time was to be expected, and does not take away the observational value of the trends indicated by the time series plots Collectively, the results establish that MTOC formation and placement, as well as microtubule growth polarity may play a role in ensuring proper stomatal morphogenesis. Precisely co-ordinated MTOC formation and placement as well as a properly maintained ratio of outbound to inbound microtubules appear to be related to the successful establishment of bilateral symmetry in stomata. As well, the results demonstrate that the primary function of MUS may be to ensure that stomatal MTOCs are symmetrically positioned to the centre of the ventral wall in developing stomata, thus ensuring bilateral symmetry is carried forward into the last stage of morphogenesis. 2.3.4 Distribution of the actin related protein BRK1 is altered in mus The impact of MUS on the actin cytoskeleton was observed utilizing the Actin Related Protein (ARP) BRICK1 (BRK1).  BRK1 is a member of the SCAR/WAVE complex that activates the actin nucleating factor ARP2/3 (Le et al., 2006).  The movement of BRK1 away from centre of the ventral wall in wild-type stomata occurs approximately 2.5 hours after the symmetric division, which is within the same time frame as MTOC formation at this site.  BRK1 movement also occurs in response to MTOC formation in mus stomata.  In mus, the BRK1 distribution domain extends to encompass the space that was left vacant along the ventral wall as a result of the MTOC shift, implying that MTOC formation leads to the exclusion of BRK1, and likely actin, from the pore site.   F-actin arrays, which are restricted to pavement cell lobes, and microtubules arrays, which are restricted to pavement cell  indentations, have been shown to mutually exclude each other during pavement cell  morphogenesis (Fu et al., 2005).  RHO-LIKE GTPASE6 (ROP6) and ROP-INTERACTIVE CRIB MOTIF-CONTAINING1 (RIC1) organize microtubule arrays in pavement cell  indentations, while also inhibiting ROP2-mediated F-actin formation in these indentations.  In rop6-1 mutants, which display microtubule array organization defects, F-actin is able to extend its distribution domain into pavement cell  indentations (Fu et al., 2009).  Similar to observations in pavement cells, abnormal mus stomata exhibit extension of the BRK1 distribution domain, mutual exclusion of the MTOCs 65  and BRK1-YFP, as well as microtubule array organization defects (Keerthisinghe et al., 2015).  As such, several of the key mechanisms of pavement cell  morphogenesis may also be conserved during guard cell morphogenesis.  Intriguingly, PIN1, which is enriched in pavement cell  lobes along with F-actin, is also a regulator of pavement cell  morphogenesis (Xu et al., 2010).   In guard cells, PIN3 appears to be excluded from the pore site, and PIN3 also localizes along the ventral wall with BRK1 (Le et al., 2014a).  It may be of interest to further explore whether PIN3, F-actin and microtubules act together to regulate guard cell morphogenesis.   2.3.5 NEK6 may be involved in the targeting of γ-TURCs to the stomatal microtubule organizing centers Observation of NEK6 throughout stomatal development revealed that NEK6-GFP localization is strongest in the MTOCs of young guard cells.  The high expression and focussed distribution of NEK6-GFP in developing guard cells indicates that it may play a central role in stomatal MTOC formation or placement.   Transient expression of NEK6 in tobacco leaves revealed that NEK6 localizes to the junction of microtubule branches.  Moreover, transient co-expression of NEK6 and the γ-TURC associated protein NEDD1/GCP-WD established that both these proteins are in close proximity at microtubule branch sites (Motose et al., 2008).   Walia et al (2014) utilized stable expression lines to demonstrate that NEDD1 also co-localizes with GCP2 at the junctions of microtubule branches.   Several studies have revealed that orthologues of NEDD1/WD-GCP in Xenopus and human cells are responsible for directing γ-TURCs to spindle poles and centrosomes respectively (Sdelci et al., 2012).  NEDD1-mediated targeting of γ-TURCs to the centrosome in human cells is activated when NEDD1 is sequentially phosphorylated by the cytoplasmic serine threonine cyclin dependent kinase 1 (CDK1) and POLO-LIKE KINASE 1 (PLK1) (Zhang et al., 2009).   PLK1 has recently been shown to indirectly regulate NEDD1 through phosphorylation of the human NEK orthologue, NEK9.  NEK9 then subsequently phosphorylates and activates NEDD1, resulting in the movement of γ-TURCs to the centrosome (Sdelci et al., 2012).       66  NEK kinases are conserved in fungi, animals and plants.  The human genome encodes thirteen NEK homologs, while Arabidopsis has seven NEK-encoding genes (NEK1 – NEK7).  All Arabidopsis NEKs are thought to have evolved from a NEK6-like ancestor, suggesting that NEK6 is the most basal member of the NEK family in Arabidopsis.  In contrast to the NEK family, the PLK family has been lost in all higher plants.  Takatani et al (2015) propose that the loss of PLK1 in Arabidopsis may have been compensated for by the expansion and neo-functionalization of other serine/threonine kinase families, including the CDK1, ARORA and NEK families of kinases.   Although interactions between PLK1, NEK9 and NEDD1 have been investigated largely in organisms outside the plant lineage, it may be possible that some aspects of the established metazoan pathways have been conserved in plant lineages.  As well, Arabidopsis may also employ additional, novel, pathways involving plant-specific partners for NEK6, such as MUS.     2.3.6 MUS NEK6 and CLASP likely regulate stomatal morphogenesis Mutants in mus display a range of stomatal defects.  The most severe of these defects are rarely observed in mus, and encompass phenotypes in which stomata exhibit abnormally expanded guard cells and/or lack a pore.  Double mutants of mus clasp and mus nek6 appear to contain an increased percentage of stomata that both lack a pore and are abnormally expanded, indicating that the double mutant phenotypes, particularly the mus nek6 double mutant, enhance the severity of the mus phenotype.   The double mutants significantly affect pore formation and guard cell morphogenesis, indicating that CLASP and NEK6 may also be involved in the regulation of stomatal morphogenesis.   However, clasp-1 and nek6-1 single mutants do not display stomatal defects, and the combination of either of these mutants with the mus mutation is necessary to reveal the involvement of CLASP and NEK6 in stomatal morphogenesis.  The wild-type stomata observed in clasp-1 and nek6-1 may be attributed to the presence of functionally redundant proteins in the signaling pathway.  Although the Arabidopsis genome does not contain any genes redundant to CLASP, it does contain several genes closely related to MUS and seven genes are in the Arabidopsis NEK family.   The presence of redundant proteins in the pathway raises the possibility that a single mutation in MUS, 67  NEK6 or CLASP can be fully compensated, or partially compensated in the case of mus, for by the one or more of the redundant proteins.  In contrast the mutation of two proteins in the pathway cannot be compensated for by redundancy, leading to the increased severity observed in double mutants.  NEK6 dimerizes with NEK4 and NEK5 to negatively regulate hypocotyl cell expansion (Motose et al., 2011).   Mutations in nek6-1 may consequently alter the activity of NEK4 and NEK5, thereby adversely affecting the function of multiple redundant proteins, and greatly exacerbating the mus phenotype in the mus nek6-1 double mutant as a consequence.   The current data points towards two potential scenarios that may explain how MUS, CLASP and NEK6 co-ordinately regulate stomatal morphogenesis.  In the first scenario, CLASP and NEK6 are downstream components of the MUS signaling pathway.  In this pathway peripherally localized MUS may represent the apex member of a pathway that integrates spatial information from both guard cells, ultimately resulting in the phosphorylation of CLASP and NEK6.  NEK6 may then phosphorylate NEDD1, which in turn directs the aggregation of γ-TURCs at the stomatal MTOC.  In the second, more likely, scenario CLASP and NEK6 may function independently of the MUS signaling pathway.  Here, the CLASP-NEK6 pathway may initiate MTOC formation, while signaling through the MUS pathway may ensure or enforce symmetrical MTOC placement.    2.4  Summary In Arabidopsis, stomatal MTOCs form at the cortex and can be easily visualized, in contrast to nuclear MTOCS.  In addition, unlike centrosomes in human/metazoan cells, which have been difficult to characterize due to their low abundance (Schatten, 2012), stomatal MTOCs are numerous and can be found throughout the shoot epidermis.   Therefore stomata provide a valuable model system which can be utilized not only to examine mechanisms of MTOC formation in plants, but also to explore the existence of conserved MTOC formation mechanisms between plants and other organisms.  Time-lapse analysis revealed that the LRR-RLK MUSTACHES maintains stomatal bilateral symmetry by enforcing the placement of MTOCs at the centre of the ventral wall, and that MUS may also regulate the polarity of microtubule growth, thereby ensuring the 68  establishment of a bilaterally symmetrical shape that is required for optimal function in mature stomata.  Further, we confirmed the existence of a stomatal-specific cortical MTOC which begins to form after the symmetric division takes place, and predicts the site of pore formation.  We also show that formation and placement of the stomatal MTOC is dependent on MUS, CLASP and NEK6.  Further studies will be required to elucidate whether NEK6 and CLASP are components of the MUS signaling pathway, or whether these proteins function in pathways that work independently of the MUS pathway.  Additionally, as human NEKs are involved in NEDD1 phosphorylation and γ-TURC targeting to the maturing centrosome, it may be of interest to determine whether AtNEDD1 is a direct target of the MUS pathway.  To our knowledge, this study is the first to describe potential mechanisms that influence stomatal MTOC formation.  Although these mechanisms require further refinement and study, our results provide a base from which a greater understanding of the genetic regulation of stomatal MTOC formation can be achieved.      2.5  Materials and methods  2.5.1 Growth of plant materials Arabidopsis thaliana seeds of glabrous1 ( gl1) in the Col-0 ecotype background were sterilized in a solution of four parts 87.5% ethanol and one part 30% hydrogen peroxide for thirty-five seconds.  Sterilized seeds were then dried and planted on ½ strength Murashige and Skoog (MS) media containing 0.8% w/v agar.  Seeds were germinated in a 21°C incubator under a 16 h light/8 h dark photoperiod.   2.5.2 Plant material Mutants: Double and quadruple mutants were generated by introgressing mus-1 (Keerthisinghe et al., 2015) into spr-3 (Nakamura and Hashimoto, 2009), clasp-1 (Ambrose et al., 2007), nek6-1  (Motose et al., 2011), brk1-1 (Dyachok et al., 2011) and the eb1-a eb1-b eb1-c triple mutant (Bisgrove et al., 2008).  The resulting double and quadruple mutants were selected based on their respective phenotypes.   69  Marker lines: proCLASP:CLASP-GFP (Ambrose et al., 2011), proGCP2:GCP2-GFP (Hashimoto,),  pro35S:EB1b-GFP (Takashi Hashimoto) , proNEK6:NEK6-GFP (Motose et al., 2011) and proBRK1:BRK1-YFP (Dyachok et al., 2011) were introgressed into mus-1.    2.5.3 Sample preparation  Staining: To visualize cell walls in the shoot epidermis, young cotyledons and first leaves were immersed in an aqueous solution of 5 mg/ml propidium iodide (PI) for 10 min, and then washed with distilled water for approximately 30 sec.   Time-lapse: Seedlings were prepared for time-lapse imaging as in Peterson and Torii (2012), but with the following modifications:  All seedlings were grown vertically on ½ MS media containing 0.8% w/v agar.   When seedlings reached 1- 1.5 days after germination (DAG), three to four seedlings were transferred into each chamber of a 2-well chambered cover-glass system (Lab-Tek II Chambered Cover-glass, Cat.No. 12565336 ).   The chambers were then covered by an approximately 5-7 mm slice of preset  ½ MS containing 0.8% agar w/v.  Seedlings prepared in the manner described above remained viable for 2-3 days after preparation.    2.5.4 Microscopy  Image acquisition:  Images were obtained from a Nikon (www.nikon.com) C1 confocal laser scanning microscope and a Perkin Elmer Spinning Disk Confocal microscope.  Both microscopes were equipped with 488 nm and 543 nm lasers.  FIJI was used to analyze all images and to generate intensity plots, while figures were prepared on Adobe Photoshop. Single frame images:  Images were obtained on either a 63X HCX Plan Apochromat glycerol objective (NA 1.30) on the Perkin Elmer Spinning Disk, or on a 60 X Plan Apo VC oil objective (NA 1.40) on the Nikon C1 confocal. Time-lapse: All time-lapse Images were obtained on a 63X glycerol objective (NA.1.30) mounted on the Perkin Elmer Spinning Disk microscope.   Long term time-lapse images were taken at approximately 2-2.5 hour intervals for a total of 8 h.  In order to obtain an image of mature guard cell morphology, the final time-lapse image was taken 70  approximately 16-17 h after the 8 hour time-point, resulting in a 24 hour time-point.   Short term time-lapse images were taken at approximately 30 min intervals for a total of 2-2.5 h.  The final time point was taken roughly 20-22 hours after the 2 h time point, resulting in an approximately 24 hour time point, thus again providing an image of the mature guard cell.      2.5.5 Counts and sampling  Double Mutant Counts: Cotyledons were counted 15 DAG.  5 cotyledons and rosette leaves each were randomly selected from wild-type, mus, mus clasp and mus nek6.  30 fields from each cotyledon were randomly scored, and counts of abnormal and normal stomata were obtained.  Microtubule polarity in young  guard cells:  Samples of 5 wild-type stomata and 5 mus stomata containing pro35S:EB1-GFP were randomly selected from several randomly selected cotyledons 1 – 1.5 DAG.   Twelve comets per stoma (6 comets per each of the 2 guard cells) were monitored for direction of microtubule growth, as indicated by the direction of the (+) end of each microtubule (also referred to as a comet).  The direction of microtubule growth, as indicated by the direction of EB1 comet movement, was categorized as either outbound (away from the ventral wall ) or inbound (towards the ventral wall ).  A count of outbound and inbound comets was computed for the 12 random comets from each of the 5 wild-type and mus stomata at 0, 30, 60, 90, and 120 min intervals following the completion of the symmetric division. The proportion of outbound comets per 12 comets per stoma at 0, 30,60,90 and 120 min time points was then computed for wild-type and mus.  2.5.6 Statistical analysis  ANOVA: The raw proportion data was analyzed for assumptions of normality using residual analysis with Shapiro Wilks test and homoscedasticity using Levenes test. The raw proportions were then transformed into arcsine in both degree and radian format. Two-way ANOVA with replication was conducted for the effect of genotype and time on direction using PAST 3 and Excel statistic packages. Post-hoc tests were considered only if significant 71  differences were indicated between Factor levels numbering more than two (genotype had only 2 levels and time had 6 ).  Chi Square Analysis: As normality tests of small samples do not necessarily yield accurate results, chi square analysis, which is distribution independent, was conducted. Contingency tables for Direction (Inbound & outbound)  vs Time (six time points) as well as 2 x 2 tables for Genotype (wild-type & mus ) vs time (six time points) were  carried out using the PAST3 Statistical package Logit Regression: Although 2-WAY ANOVA determines the significance of the effect of genotype and time on microtubule direction it does not quantify the level of relationship between direction and genotype or time. Therefore, in order to quantify the relationship between genotype and direction, logistic regression of microtubule direction, which is a binary variable (inbound and outbound), on Genotype (wild-type and mus) was conducted using the EXCEL Real Statistics and solver packages.  Time Series Plots:  The proportion of outbound microtubules were computed as ‘# outbound microtubules / total # microtubules,’ and the ratio of outbound microtubules to inbound microtubules was computed as total # outbound microtubules : total # inbound microtubules, from raw counts of the two samples of 5 stomata each obtained from wild-type and mus.  The proportion of outbound microtubules and ratios of outbound microtubules to inbound microtubules were plotted against six time points 0,30,60,90 and 120 min following the symmetric division for both wild-type and mus (Figs. 1a, 1b. 2a and 2b). Residuals of the dependent variable for both plots were analysed for normality for regression purposes. Both the proportion of outbound microtubules and the ratio of outbound microtubules to inbound microtubules for wild-type were plotted against those of mus for each of the six time points 0, 30, 60, 90, and 120 min after the symmetric division.  Regression and correlation statistics were computed for all plots.     72  Chapter 3 : MUSTACHES, and the closely related kinase MUS-LIKE1 may regulate positioning of the stomatal pore   3.1  Introduction  Plants, despite being sessile organisms comprised of non-motile cells, are capable of sophisticated and efficient responses to their environment.   Much of this responsiveness may be due to the ability of plants to perceive extracellular signals through membrane bound Receptor-Like Kinases (RLKs), which rapidly transduce these signals into internal cellular responses.  Over the past decade, numerous studies have shown that RLK signaling pathways regulate crucial developmental processes, including plant immunity morphogenesis, and growth.   Receptor-Like Kinases (RLKs)  in Arabidopsis thaliana                ( Arabidopsis) belong to the large RLK/PELLE kinase family which encompass approximately 2.5% of the protein-coding genes (600 RLKs) encoded in the Arabidopsis genome (Shiu and Bleecker, 2001).  The isolation and functional characterization of individual RLKs from the RLK/PELLE family will be essential for uncovering and understanding many vital plant developmental pathways.  The majority of plant RLKs exhibit a characteristic structure consisting of three main domains.  First, the extracellular domain perceives external signal molecules, such as small peptide ligands or hormones, released from the surrounding environment.  Second, the trans-membrane domain, which follows the extracellular domain, embeds the RLK into the cellular membrane.   Finally, the intracellular kinase domain transmits signals received by the extracellular domain into the cell via phosphorylation of downstream targets (Wierzba and Tax, 2013).  Plant RLKs are orthologues of the Drosophila PELLE family and, as such, their kinase domains belong to the serine/threonine class of kinases.   In contrast to the conserved kinase domain, RLKs exhibit a variety of extracellular domains, including Leucine-Rich Repeats (LRRs), Epidermal Growth Factor (EGF)-like repeats, Self-Incompatibility (S)-Domains and Lectin-domains (Goff and Ramonell, 2007).  However, LRRs are the most common extracellular domain, as they are present in approximately one third of receptors 73  (Shiu and Bleecker, 2001).  LRRs consist of twenty to thirty amino acids and are characterized by a high number of repeating leucine sub-units.  LRR domains are curved structures capable of mediating protein-protein interactions (Bella et al., 2008)   LRR-RLKs often interact with a partner to efficiently facilitate signal transmission.  Common RLK partners include other LRR-RLKs, or Leucine-Rich Repeat Receptor-Like proteins (LRR-RLPs).  LRR-RLPs contain extracellular domains , consisting of LRRs, and trans-membrane domains, but they lack kinase domains.  The LRR-RLK CLAVATA1 (CLV1) interacts with both the LRR-RLP CLAVATA2 (CLV2) and closely related kinases of the BARELY ANY MERISTEM (BAM) family to maintain shoot apical meristem size (DeYoung et al., 2006).   Another well characterized LRR-RLK signaling pathway  regulates asymmetric division placement in the stomatal lineage by the ERECTA family of LRR-RLKs, along with their partner the LRR-RLP TOO MANY MOUTHS (TMM) (Le et al., 2014b).   Stomata are specialized shoot epidermal structures consisting of two kidney-shaped guard cells surrounding a pore.  Movement of the guard cells opens and closes the pore, which regulates the exchange of CO2 and O2 for photosynthesis and respiration, while also limiting evapotranspiration.   Stomatal development initiates when an asymmetric division generates a dispersed stem cell, the meristemoid.   Soon after an asymmetric division establishes the meristemoid, the meristemoid transitions into a guard mother cell.  The guard mother cell then undergoes a single symmetric division  which generates two young guard cells. The new wall generated by the symmetric division, is referred to as the ventral wall, and the ventral wall is opposed on either side by the external, anticlinal, dorsal walls.   The completion of pore formation in the centre of the ventral wall denotes the generation of a mature stoma and the termination of stomatal differentiation (Willmer and Fricker, 1996; Zhao and Sack, 1999).  Stomata follow the one cell spacing rule, whereby the majority of asymmetric divisions, and subsequently mature guard cells, are spaced at least one cell apart.   The one cell spacing rule ensures efficient guard cell movement and pore opening.  Mutations in TMM lead to a clustering phenotype, where stomata form in contact.   In contrast to tmm, mutants in the ERECTA (ER), ERECTA-LIKE 1 (ERL1) and ERECTA-LIKE 2 (ERL2) LRR-RLKs do 74  not display clustering phenotypes.  However, higher order mutants between er and one or more of the erls, exhibit sequentially more severe clustering phenotypes, with the er erl1 erl2 triple mutant clustering phenotype resembling tmm in severity (Shpak et al., 2005). These findings demonstrate that kinases in the ER/ERL family function redundantly, and that the ER/ERL-TMM mediated pathways enforce the one-cell spacing rule.  Stomatal development concludes upon the completion of guard cell and pore morphogenesis.  A second LRR-RLK, MUSTACHES (MUS) has been shown to regulate guard cell and pore morphogenesis post- symmetric division (Keerthisinghe et al., 2015).  Mutants of MUS exhibit abnormal guard cell morphogenesis leading to ectopic, or skewed, pores.  mus mutants also display defects in microtubule array organization, as well as alterations in the direction of microtubule growth.  Intriguingly, only 20% of the stomata in mus display abnormal phenotypes, suggesting that the mus mutation may not completely eliminate MUS function.   As in the case of the asymmetric division placement, it is possible that guard cell morphogenesis is regulated through multiple redundant kinases.  The Arabidopsis genome encodes four MUS-LIKE KINASES (MUSLs).  The MUSLS may represent potential MUS co-receptors or members of the MUS pathway (Keerthisinghe et al., 2015). Here we show that over-expression (OX) of one of the MUSLs, MUSL1, generates a gain of function phenotype, which inverts the polarity of the guard cell structure.  After synthesizing data from numerous sources (35S over-expression lines, transcriptional fusions and translational fusions), we propose that guard cell morphogenesis is established through the antagonistic interactions of the MUS and MUSL1 pathways.     3.2 Results 3.2.1 MUS and MUSL1 exhibit extensive structural symmetry The Arabidopsis genome contains four kinases closely related to MUS, referred to here as MUS-LIKE KINASE (MUSL) one through four (MUSL1 – MUSL4).  At4g36180 (MUSL1) includes 3757 base pairs (bp) and contains one intron, At5g58150 (MUSL2) includes 3572 bp and contains no introns, At2g24230 (MUSL3) includes 2996 bp and contains one intron, and finally At5g45800 (MUSL4) includes 3572 bp and contains no introns.   75  The amino acid similarity between MUS and each of the four MUSLs was characterized in order to determine the degree of structural similarity within the MUS kinase family.   Of the four MUSLs, MUSL1 displays the closest similarity to MUS (Fig. 3.1 A).  MUSL1 and MUS align over their entire lengths, and they share 74% total amino acid similarity over their respective lengths.  In contrast, MUS and the other three MUSLs (MUSL2-MUSL4) do not align over their entire lengths, and these kinases only share  46% amino acid identity with MUS.   Comparison of the individual RLK domains (extracellular domain, trans-membrane domain and kinase domain) between MUS and MUSL1 demonstrated that the MUS and MUSL1 extracellular domains share 70% similarity, the trans-membrane domains exhibit 71% similarity, and finally the MUS and MUSL1 kinase domains share 89% similarity.   In contrast, comparisons of individual RLK domains between MUS and MUSL2-MUSL4  demonstrate that the extracellular domains exhibit roughly 50% similarity, between 0% - 61% similarity between the trans-membrane domains, and approximately 50% amino acid similarity between the kinase domains .  76    Figure 3.1 MUS-LIKE 1 (MUSL1) ehxibits close structural similarity to MUS  (A)  Map signifying the percentage of amino acid similarity between MUS and each of the MUSLs (MUSL1 – MUSL4).  Each domain is represented by 10 cells and one cell represents 10 % of the domain.   Blue Cells – percent similarity between amino acid in the kinase domains, Red cells – similarity between  amino acids in trans-membrane domains, Green Cells – percents similarity between amino acids in the Extracellular Domain (ECD), and Grey Cells – percent unaligned amino acids in all domains.   (B)  T-DNA alleles utilized in this study. Insertion sites are denoted by red triangles.    77  3.2.2 musl mutants do not exhibit stomatal defects  Single musl mutants were examined to determine whether the MUSL kinases participate in stomatal development (Fig 3.1 B).  Abnormal stomata were not observed in any of the four single musl mutant lines.   As the single mutants did not display stomatal phenotypes, double mutants of mus and each of the musls were examined.   mus musl double mutants displayed both normal, wild-type, stomata as well as abnormal, mus-type, stomata.  The percentage of abnormal stomata in mus single mutants was 21.6% ± 1.0 % (S.E.).  mus musl1 and mus musl2 double mutants displayed similar percentages of abnormal stomata (23.1% ± 1.0 % and 19.5% ± 0.78%  respectively) compared to the mus single mutant (Fig  3.2).    In contrast, mus musl3 and mus musl4 double mutants exhibited slightly lower percentages of abnormal stomata (16.4% ± 1.1% and 16.2% ± 0.74% respectively) in comparison to mus.  However, the percentage of mus type stomata in the double mutants was not significantly different from the percentage of abnormal stomata in mus single mutants (ANOVA, F=1.448, p=0.264, n = 30 fields, R= 5 cotyledons).     78  Figure 3.2 mus musl1 double mutants do not increase the severity of the mus single mutant phenotype.   Black bars represent wild-type stomata, while white bars represent mus-type stomata.  Standard Error (S.E) bars are shown.   3.2.3 MUSL2, MUSL3 and MUSL4 over-expression lines do not exhibit distinct sub-cellular localization patterns or gain of function phenotypes   35S GFP over-expression (OX) lines of each of the four MUSLs (MUSL1 - MUSL4) were generated to identify and isolate potential gain of function mutants.  MUSL2-MUSL4 OX lines were transformed into wild-type and mus backgrounds.  MUSL2 and MUSL4 OX lines exhibited broad sub-cellular localization patterns in both wild-type and mus, as GFP was observed in the plasma membrane of both stomatal lineage cells and pavement cells throughout the cotyledon epidermis (Fig 3.3 A - B).  pro35S:MUSL3-GFP also exhibited sub-cellular localization to the plasma membrane of all epidermal cell types in wild-type and mus.  However, MUSL3 appeared to be brighter (enriched) in the cell membranes of stomatal precursor cells (meristemoids and young guard mother cells) in OX lines (Fig 3.3 C-D).  Gain of function phenotypes were not observed in MUSL2 - MUSL4 OX lines in either wild-type or mus. Additionally, no obvious non-stomatal phenotypes were observed.      Figure 3.3 MUSL2, MUSL3 and MUSL4 over expression lines do not display distinct sub-cellular localization patterns or gain of function phenotypes. 79  Green fluorescence represents pro35S:MUSL2-GFP (A), pro35S:MUSL3-GFP (B) and pro35S:MUSL4-GFP  respectively.   All scale bars are 20 µm.  (A) pro35S:MUSL2-GFP over-expression (OX) lines display ubiquitous sub-cellular localization throughout the wild-type cotyledon epidermis (B) pro35S:MUSL4-GFP OX lines display ubiquitous sub-cellular localization throughout the wild-type cotyledon epidermis (C – D) pro35S:MUSL3-GFP OX lines display ubiquitous sub-cellular localization throughout both the wild-type (C) and mus (D) cotyledon epidermises.  However, pro35S driven MUSL3-GFP appears to be especially enriched in stomatal precursor cells (asterisks).  3.2.4 pro35S:GFP-MUSL1 sub-cellular localization resembles proMUS:MUS-tripleGFP localization    As MUSL1 displays the closest amino acid similarity to MUS the MUSL1 OX construct was transformed into wild-type, mus, and musl1 and mus musl1 backgrounds.  Interestingly, in contrast to the other MUSLs, MUSL1 OX in wild-type results in a sub-cellular localization pattern that resembles proMUS:MUS-tripleGFP localization (Fig. 3.4).  Sub-cellular localization of MUSL1 is visible throughout the cotyledon epidermis.  However, MUSL1 localization is fainter in cell membranes and brighter in cell plates and new divisions. In stomatal lineage cells, as with the proMUS:MUS-tripleGFP reporter construct, GFP-MUSL1 driven with the 35S promoter (pro35S:GFP-MUSL1) in wild-type was first observed at the cell plates and recently divided walls of young meristemoids.  As meristemoids transition to guard mother cells, GFP-MUSL1 re-localized to the periphery of the guard mother cells.   Peripheral GFP-MUSL1 was then maintained once the guard mother cell underwent the symmetric division, which generates the two young guard cells.  However, in contrast to proMUS driven MUS-GFP, peripheral pro35S driven GFP-MUSL1 localization in young guard cells is weak.   Strikingly, and in direct contrast to MUS-GFP localization during comparable developmental stages, pro35S-driven GFP-MUSL1 displays intense localization to the symmetric division in newly divided guard cells.   Finally, as with MUS-GFP, peripheral pro35S driven GFP-MUSL1 persists in mature stomata (Fig 3.4 A, D –J). 80  Sub-cellular localization of the pro35S-driven GFP-MUSL1 in the mus background also resembles that of MUS-GFP driven with its endogenous promoter.  First, GFP-MUSL1 localizes along new divisions in young meristemoids.  However, localization of GFP-MUSL1 to young meristemoids in both mus and mus musl1 appears to be less frequent than that of GFP-MUSL1 in wild-type meristemoids.  GFP-MUSL1 then becomes peripherally distributed as the meristemoids transition into guard mother cells.   In contrast to both MUS-GFP and GFP-MUSL1 localization in wild-type, GFP-MUSL1 does not consistently localize to the periphery of young mus (or mus musl1) guard cells.   Instead, GFP-MUSL1 frequently localizes only to new symmetric divisions in young mus guard cells.   Finally, peripheral GFP-MUSL1 localization can be observed in mature mus stomata (Fig 3.4 B).  In contrast to wild-type, mus and mus musl1 backgrounds, GFP-MUSL1 localization in musl1 cotyledons appeared to be either very weak, or entirely absent. GFP-MUSL1, when visible in musl1, exhibited faint sub-cellular localization to cell membranes in pavement cells, while GFP-MUSL1 localization appeared to be slightly more intense in stomatal lineage cells (Fig. 3.4  C).    Interestingly, the pro35S – driven GFP-MUSL1 distribution patterns in the musl1 root epidermis mirrored proMUS:MUS-GFP localization in wild-type roots (Keerthisinghe et al., 2015).        81  Figure 3.4 MUSL1 sub-cellular localization resembles native MUS localization Red fluorescence indicates cell walls visualized with Propidium Iodide (PI) stain, while green fluorescence represents pro35S:GFP-MUSL1.  The brightness and contrast of the green channel in a subset of micrographs has been enhanced to ensure the visibility of weak GFP localization (C, G-H)   Scale bars are 20 µm (A-C) and 5µm (D – J) respectively.  Asterisks represent GFP-MUSL1 localization to stomatal lineage cells, while arrows represent GFP-MUSL1 localization to symmetric divisions. (A - C) As observed for MUS-GFP localization, GFP-MUSL1 localizes to new divisions and cell plates in asymmetric divisions as well as to the periphery of stomatal precursor cells.  wild-type (A), mus (B) and musl1 (C) (D – J) Sub-cellular localization of pro35S:GFP-MUSL1 throughout wild-type stomatal development. First, GFP-MUSL1 localizes to new cell plates and recent divisions sites of the asymmetric divisions which generate meristemoids (D).  GFP-MUSL1 distribution then becomes peripheral in guard mother cells (E-F).   Strikingly, in contrast to MUS-GFP,  pro35S driven GFP-MUSL1 localizes to cell plates and new divisions, in addition to cell peripheries, in guard mother cells (G-H).  Initially, GFP-MUSL1 localizes to the periphery of developing stomata, but then GFP-MUSL1 localization gradually diminishes as the stomatal completely matures (I – J).    3.2.5 Over-expression of GFP-MUSL1 in mus and mus musl1 double mutants results in a gain of function ‘butterfly’ effect  MUSL1 OX in both mus and mus musl1 backgrounds resulted in the development of an additional category of abnormal stomata, which is defined by an inversion, or flipping of guard cells, such that the dorsal walls of each guard cell are in contact while the ventral walls face outwards.  As these abnormal, inverted stomata, resemble a butterfly, we refer to this gain of function phenotype as the ‘butterfly ’ effect (bf).  ‘Semi-butterfly (sbf)’ phenotypes, where only one guard cell in the stoma was inverted while the other guard cell remained un-inverted, were also observed (Fig 3.5 A, B-E).  Control lines (untransformed mus and mus musl1 mutants) did not exhibit bf stomata.  However, 0.13% ± 0.08% of stomata in mus and 0.50% ± 0.14% of stomata in mus musl1 control lines did exhibit the sbf 82  phenotype. Neither the bf nor the sbf phenotype was observed in OX lines of the other MUSLs (MUSL2 – MUSL4).  In order to fully characterize the bf gain of function phenotype, two independent lines containing pro35S:GFP-MUSL1 in mus , which will be termed ‘bf’ lines for the remainder of the text, were scored.  Single mus mutants contain 21% abnormal stomata, while both bf lines exhibited approximately 30%- 40% abnormal stomata.  Of the abnormal stomata in the bf  lines, approximately 10.1% ± 0.85%  to 15.7% ± 1.20%  of  stomata exhibited a bf or sbf phenotype, while the remainder of the abnormal stomata were more typical of those previously observed in mus single mutants (n=30 fields R = 5 cotyledons) (Fig 3.5 F).    Abnormal stomata were also observed in wild-type MUSL1 OX lines, which contained 4.76% ± 0.54% abnormal stomata.  These abnormal stomata exhibited a variety of phenotypes, including mildly skewed stomata, stomata with abnormally expanded guard cells and, on occasion, bf or sbf stomata.  Intriguingly, OX of GFP-MUSL1 in the musl1 background did not result in abnormal stomatal phenotypes.    83   Figure 3.5 MUSL1 over-expression in a mus or mus musl1 background generates a ‘butterfly’ (bf) effect gain of function phenotype. Red fluorescence indicates cell walls visualized with Propidium Iodide (PI) stain, while green fluorescence represents pro35S:GFP-MUSL1  Scale bars are 20 µm (A) and 5µm (C – E).   84  (A)  OX of pro35S:GFP-MUSL1 in a mus  or mus musl1 background (only mus shown), results in the generation of a bf phenotype, which is characterized by the inversion of guard cells, such that the dorsal walls  touch and the ventral walls face outwards.     (B-E) pro35S:GFP-MUSL1 OX lines in a mus background exhibit a variety of stomatal phenotypes ranging from wild-type (B), mus type (C), butterfly (D), and semi-butterfly (E).   (F) Percentages of wild-type (blue), mus-type (pink) and bf + sbf type (green) stomata in wild-type, mus, wild-type pro35S:GFP-MUSL1 OX and mus pro35S:GFP-MUSL1 OX lines.   3.2.6 Prolonged localization of MUSL1 to the symmetric division likely causes the butterfly effect  Time-lapse imaging of OX GFP-MUSL1 in mus was undertaken to determine whether the bf phenotype resulted from ectopic localization of MUSL1 to the symmetric division.  Thirteen stomata were observed and, of these, five developed into bf or sbf stomata, while eight developed into wild-type or mus-type stomata.  Time-lapse observations demonstrated that GFP-MUSL1 localized to the symmetric division in wild-type as well as bf and sbf stomata.  However, in bf (and sbf) stomata GFP-MUSL1 appeared to localize to the symmetric division for a prolonged period of time, as MUSL1 localized to the symmetric division for 8 hours or more in 4/5 (80%) of bf and sbf stomata.   In comparison, only 2/8 (25%) of wild-type or mus-type stomata displayed GFP-MUSL1 localization to the symmetric division for 8 hours or more (Fig 3.6).  85   Figure 3.6 The bf phenotype may result from prolonged localization of MUSL1-GFP to the symmetric division in the absence of mus. Green fluorescence represents pro35S:GFP-MUSL1.  Time-lapses were obtained in two hour intervals, starting at the recently divided guard cell stage,  The final time-points were taken at 24 and 48 hours respectively.   All time-lapses were obtained from pro35S:GFP-MUSL1 OX lines mus background.  Scale bars are 5µm (C – E).   (A – G) Time-lapse of a young guard cell which developed into a wild-type stomata.  GFP-MUSL1 localization appears to diminish after 8 hours. (H – N) Time-lapse of a young guard cell which developed into a bf stomata. GFP-MUSL1 localization persists past 8 hours. (O)  The majority of cells developing into bf stomata exhibit localization to the symmetric division for more than 8 hours, while the majority of cells that develop into non-bf stomata (mus or wild-type) do not.  86  3.2.7 MUSL1 is expressed, and localizes, throughout the cotyledon epidermis Native MUSL1 expression patterns were characterized in wild-type and mus cotyledons expressing proMUSL1:GFP in order to assess the MUSL1 promoter activity.   MUSL1 expression was only visible in young cotyledons (0.5 – 1 days after germination).  In both wild-type and mus, GFP was localized to nuclei, as well as in the cytoplasm, of stomatal lineage cells and pavement cells.  In wild-type lines, MUSL1 is expressed during all stages of stomatal development (meristemoid to mature guard cell stage).  MUSL1 expression appears to be stronger in the early stages of stomatal development (meristemoid to young guard cell stages) as compared to the later stages (mature guard cell stages) of development.  Strong MUSL1 expression is visible in young, unexpanded pavement cells , and this strong expression persists as they expand and mature.  Expression of MUSL1 in mus largely resembled that of wild-type (Fig 3.7 A – D).   Localization of the native MUSL1 protein was analyzed through a proMUSL1:MUSL1-GFP translational fusion, transformed into both wild-type and mus backgrounds.  However, only wild-type localization is described here, as mus transformants were not recovered.   GFP Fluorescence in proMUSL1:MUSL1-GFP lines was very weak, and often difficult to visualize.  Although MUSL1-GFP localized to cell membranes throughout the epidermis, it appeared at times to be enriched in mature meristemoids and young guard mother cells.  In young, recently divided guard cells, MUSL1-GFP localizes weakly to both the cell periphery and the new symmetric division (ventral wall)  (Fig 3.7 E – F, I – J).   In rare cases, MUSL1-GFP also appeared to be enriched in the periphery of mature stomata.    87    88  Figure 3.7 MUSL1 native expression and localization patterns  Red fluorescence indicates cell walls visualized with Propidium Iodide (PI) stain, while green fluorescence represents  proMUSL1-GFP (A-D) and proMUSL1:MUSL1-GFP (E – J) respectively. The brightness and contrast of the green channel in each micrograph has been enhanced to ensure the visibility of weak GFP localization.   Asterisks indicate stomatal lineage cells, while arrows represent  symmetric divisions.   Scale bars are 20 µm (A) and 5µm (C – E).   (A - B) proMUSL1-GFP in a wild-type background.   proMUSL1-GFP is expressed throughout the shoot epidermis in both stomatal (A) and non-stomatal, or pavement cell (B), lineage cells.  (C – D) proMUSL1:GFP expression in a mus background resembles proMUSL1:GFP expression in wild-type.   MUSL1 is expressed throughout the shoot epidermis in both stomatal (C) and non-stomatal, or pavement cell (D), lineage cells. (E – F)  proMUSL1:MUSL1-GFP in a wild-type background resembles proMUSL1:GFP expression in wild-type.   MUSL1 native expression was often weak and difficult to observe.  However, when it was possible to visualize MUSL1-GFP,  it was localized weakly throughout the epidermis (A), and it also appeared to be enriched in a subset of stomatal precursor cells (B).  (G – H)  proMUSL1:GFP expression is similar in wild-type (G) and mus (H) young guard cells.   (I – J)  MUSL1-GFP localizes to symmetric divisions  in young guard cells. Weak peripheral MUSL1-GFP localization was also visible (I).     3.3 Discussion 3.3.1 MUSL1 and MUSL3 may function in stomatal development  Characterization of 35S OX lines for each of the four MUSLs (MUSL1 – MUSL4) revealed that MUSL3 OX lines exhibited a unique localization pattern, while MUSL1 OX lines displayed both a gain of function phenotype and an intriguing localization pattern.   Although MUSL3 OX lines did not display any gain of function or loss of function phenotypes, MUSL3 appears to be enriched in stomatal precursor cells (meristemoids and 89  young guard mother cells) in both wild-type and mus cotyledons.   MUSL3 sub-cellular localization patterns coincide with the expression and localization of several early-stage stomatal development genes including the Basic Helix Loop Helix (bHLH) transcription factors SPEECHLESS (SPCH) and MUTE, as well as the LRR-RLP TMM.  SPCH regulates the transition of undifferentiated precursor cells into meristemoids, while MUTE regulates the transition of meristemoids into guard mother cells (Le et al., 2014b).  In contrast, TMM inhibits cells neighbouring existing stomata from acquiring a stomatal fate.  Recently, Meng et al. (2015) demonstrated that TMM is component of a multiprotein ‘recepterosome’ which includes members of the ER/ERL and SERK LRR-RLK families.   The enrichment of MUSL3 in stomatal precursors, along with the complexity of existing interactions between LRR-RLKs known to regulate stomatal patterning, may indicate that MUSL3 represents an additional LRR-RLK that functions during the early stages of stomatal development.   In contrast to MUSL3, MUSL1 OX lines exhibited the bf gain of function phenotype, where the structural polarity of a subset of guard cells was inverted.  bf phenotypes were also present in MUSL1 mis-expression lines generated using a variety of promoters which act during early stages of stomatal development, such as proTMM:MUSL1-GFP.  MUSL1-GFP also localized to the symmetric division in all of the MUSL1 mis-expression lines (Tsuyoshi Nakagawa, personal communication).    Despite the absence of stomatal phenotypes in musl1 single mutants, the inversion of guard cells in MUSL1 OX and mis-expression lines indicates that MUSL1 may participate in the regulation of stomatal morphogenesis during later stages of stomatal development.  3.3.2 MUSL1 is neither functionally redundant to MUS, nor a MUS co-receptor, in stomata Of the four MUSLs, MUSL1 displays the highest overall amino acid similarity to MUS.   Additionally, characterization of proMUSL1:MUSL1-GFP, and pro35S:GFP-MUSL1 localization patterns, as well as the bf phenotype, indicate that MUSL1, like MUS, acts during the later stages of stomatal development.  Thus, MUSL1 and MUS may be functionally redundant, or   alternately, MUSL1 may function as a MUS co-receptor.   Closely related kinases often function redundantly in Arabidopsis.  The LRR-RLK BRASSINOSTEROID INSENSITIVE 1 (BRI1) is responsible for mediating a variety of processes 90  regulated by the phytohormone Brassinosteroid (BR), including shoot and hypocotyl expansion, stomatal patterning, and xylem differentiation (Li and Chory, 1997).  BRI1 is closely related to its homologue BRASSINOSTEROID INSENSITIVE-LIKE 1 (BRL1).  Although BRI1 and BRL1 only share 45% amino acid similarity, all amino acids that are essential for BRI1 function are conserved in BRL1.   35S:BRL1 OX lines are able to partially complement the bri1-4 mutant and, although brl1 single mutants do not display any phenotypes, bri1-5 brl1-1 double mutants display increased severity compared to the bri1 single mutant.   The OX and double mutant phenotypes demonstrate that BRL1 and BRI1 are functionally redundant (Zhou et al., 2004).   Unlike BRL1 OX, MUSL1 OX does not complement the mus stomatal phenotype, and in fact MUSL1 OX increases the severity of the mus phenotype.  As well, the mus musl1 double mutant did not increase the severity of the mus phenotype.  Thus, MUSL1 and MUS are not functionally redundant.  Although MUS and MUSL1 do exhibit some over-lap in localization during early stages of stomatal development (meristemoid – guard mother cell), MUSL1 localization differs from that of MUS during the post-symmetric division stage (young guard cell – mature guard cell).  The post symmetric division stage is crucial for stomatal development as this is the stage at which MUS functions.   During this stage, MUS does not localize to the symmetric division that produces the guard cells (Keerthisinghe et al., 2015).  In contrast, MUSL1 does localize to the symmetric division in both proMUSL1:MUSL1-GFP lines and pro35S:GFP-MUSL1 OX lines.   Further, the bf phenotype that arises when MUSL1 is overexpressed demonstrates that MUSL1 is able to function in mus mutants, which likely lack functional MUS protein.   The difference between MUS and MUSL1 localization, in addition to the ability of MUSL1 to function in the absence of MUS, indicates that MUSL1 is not a MUS co-receptor.   Instead, MUSL1 and MUS may function in independent but inter-related pathways that act in conjunction to co-ordinately regulate stomatal morphogenesis.    91  3.3.3 An antagonistic relationship between MUS and MUSL1 likely regulates stomatal morphogenesis proMUSL1:MUSL1-GFP and pro35S:GFP-MUSL1 localize to dorsal wall  and ventral walls in wild-type stomata, whereas proMUS:MUS-GFP only localizes to the guard cell periphery in wild-type stomata (Keerthisinghe et al., 2015).   Similarly, in mus and mus musl1 backgrounds, pro35S:MUSL1-GFP  localizes to both the guard cell periphery and symmetric division, though localization is stronger in the symmetric division.  The differences in MUS and MUSL1 localization patterns indicate that the signaling pathways mediated by these two kinases may function antagonistically in order to maintain stomatal morphogenesis.     MUS and MUSL1 localization patterns suggests that two opposing signaling pathways may be present in wild-type stomata.  The first pathway, mediated by MUSL1, emits a signal from the symmetric division and the second pathway, mediated by MUS, emits an opposing signal from the guard cell periphery.   Correct guard cell morphogenesis is dependent, in part, on the correct placement of stomatal pore.  The net balance in signal strengths between the MUS and MUSL1 pathways may ensure correct pore placement in wild-type stomata.   MUS and MUSL1 kinase domains share 89% amino acid similarity, indicating that both these kinases may have the ability to interact with the same substrates.   As described in chapter 2, MUS enforces stomatal symmetry by regulating localization of microtubule organizing centres (MTOCs) to the centre of the ventral wall.   Therefore, MUSL1 may also be able to influence the localization of either the MTOC, or MTOC components.   When MUSL1 is over-expressed in mus1 or mus musl1 backgrounds, which lack functional MUS protein, the signal emitted from the MUSL1 pathway is emitted without the opposing, balancing, signal.  The unopposed MUSL1 signal may then influence MTOCs to re-locate to the dorsal wall, thereby leading to the formation of the bf phenotype (Fig. 3.8).  Re-location of the MTOCs to the dorsal wall was corroborated in bf lines utilizing the microtubule marker END BINDING 1 (EB1) to visualize microtubule growth and movement.   pro35S:EB1b-GFP marked microtubules in bf lines grew from the center of the dorsal wall towards the internal ventral wall (Tsuyoshi Nakagawa, personal communication).    92    Figure 3.8 Opposing MUS and MUSL1 signaling pathways may position the stomatal pore  (A)  In wild-type, MUS signals from the periphery to the center of the stomata, while MUSL1 signals from the center to the periphery.  The presence of both signals may allow for the correct, symmetrical, placement of γ-TURCs, thereby ensuring the development of a bilaterally symmetrical stoma. (B)  In wild-type OX lines both signals (MUS and MUSL1) are still present, resulting in the formation of a majority of wild-type stomata.  However, in OX lines, MUSL1 expression and localization is more intense, thereby, in some cases, allowing the MUSL1 signal to over-come the MUS signal, resulting in misplaced γ-TURCs and the small percentage of mus-type stomata observed in this background.    93  (C)  In mus OX lines, as MUS is likely non-functional, the MUSL1 signal is emitted unopposed, leading to γ-TURC mis-localization to the dorsal wall in a subset of cells, resulting in the inverted bf phenotype.     3.4 Summary   The current study demonstrates that mus musl1 double mutants do not increase the severity of the mus phenotype, and that MUSL1 OX does not complement the mus stomatal phenotype, which indicates that MUS and MUSL1 do not function redundantly during stomatal development.   Further, the bf phenotype observed when MUSL1 is over-expressed in a mus background suggests that MUSL1 is able to function in a mus background.  Thus, MUSL1 and MUS may not function as co-receptors.  Interestingly, MUS and MUSL1 display opposing localization patterns, where MUS localizes to the young guard cell periphery, while MUSL1 localizes predominately to the symmetric division.  The opposing MUS and MUSL1 localization patterns,  in addition to the bf  OX phenotype, suggests  the presence of a mechanism where opposing signals emitted from MUS and MUSL1 regulate pore and guard cell morphogenesis during stomatal development.     The final stage of stomatal development, which establishes correct pore and guard cell morphogenesis, is vital for stomatal function.  However, despite the importance of this developmental stage, few studies have attempted to define genes and mechanisms that regulate pore morphogenesis.  Therefore, the current study is significant as it reveals a previously undescribed mechanism, in which two closely related LRR-RLKs appear to co-ordinately regulate pore placement, and pore morphogenesis.   Finally, the MUS-MUSL1 signaling module provides a valuable tool that can be implemented to further uncover mechanisms and processes mediating a vital, yet under-characterized, stage of stomatal development.    94  3.5  Materials and methods 3.5.1  Growth of plant materials  Arabidopsis thaliana seeds in a wild-type, glabrous1 background (Col-0 gl1 ecotype) were sterilized in a solution of four parts 87.5% ethanol and one part 30% hydrogen peroxide for thirty-five seconds.  Sterilized seeds were then dried and planted on ½ strength Murashige and Skoog (MS) media containing 0.8% w/v agar.  Seeds were germinated in a 21°C incubator under a 16 h light/8 h dark photoperiod.   3.5.2  Plant material Mutants: The mus-1 allele (referred to as mus in the text) contains a point mutation in the kinase domain, suggesting that mus-1 encodes a kinase-inactive protein (Keerthisinghe et al., 2015).  SALK insertions for each of the musls were obtained from TAIR (https://www.arabidopsis.org/about/citingtair.jsp), and the presence of insertions were verified through genotyping.   The following primers were used :   LB (5’-ATT TTG CCG ATT TCG GAA C-3’);  MUSL1, (FWD – 5’-AGA CTT GCT AAA GCG AGT CCC-3’,  REV – 5’- GAA TGG TTC TCT CTA TCC GCC-3’);  MUSL2,  (FWD -5’-GGA AGG CAA GTC TGG TCC TAC-3’, REV – 5’-TCA TAC GAG GCA TGG TAG GAC-3’);  MUSL3 (FWD – 5’-TCC AAA GAT TGA AGC TTG CTG-3’, REV- 5’- TGC ACA CTC TCC TTT AAT GGC-3’) and MUSL4, (FWD -,5’-ACA CCG TTC AGC ATT TAA TGC-3’; REV-5’- TTG AAT TTA TCT GCA AAC CGG-3’).   Double mutants were generated by introgressing mus (Keerthisinghe et al., 2015) into musl1 – musl4.  Subsequently, double mutants were selected by phenotyping for mus and genotyping for the respective musl mutants.    3.5.3 Expression constructs Translational and Transcriptional fusions: The MUSL1 promoter was cloned using the following primers: proMUSL1-TOPO-F (5’-CACCTTCTCGCCGTGAAAATCGATC-3’) and  proMUSL1-TOPO-R (5’- AGATCTGTGGCAGTGAACAAA-3’).  The amplified fragment was cloned into a pENTR D-TOPO ENTRY vector (Thermo Fisher Scientific), according to the manufacturer’s instructions, to generate a pENTR-proMUSL1 construct.  LR clonase II (Thermo Fisher Scientific) was then utilized to recombine pENTR-proMUSL with the 95  pMDC107 DESTINATION (pDEST) vector to generate the proMUSL1-pMDC107 (proMUSL1-GFP) transcriptional fusion.  The MUSL1 Open Reading Frame (ORF) was cloned using the following primers:  F (5’-GGCGCGCCATGGCCATGGATATATCTCTG-3’) and R (5’-GGCGCGCCTGGCCGGTGAAGGCTG AGAGGT-3’).   Both proMUSL1-pMDC107 and the MUSL1 ORF were then digested with ASCI  (NEB Biolabs), and the digested proMUSL1-pMDC107 vector then underwent Calf  Intestinal  Phosphatase (CIP) treatment (NEB  biolabs).  T4 DNA ligase (NEB biolabs) was then utilized to combine the MUSL1 ORF with proMUSL1-pMDC107 to generate the proMUSL1:MUSL1-GFP translational fusion.    proMUSL1-GFP and proMUSL1:MUSL1-GFP  constructs were introduced into Agrobacterium tumefaciens strain GV3101, and lines resistant to 50 µg/µl kanamycin and 50 µg/µl hygromycin were used for transformation of wild-type and mus1 plants via the floral dip procedure (Clough and Bent, 1998).   Over-expression (OX) lines :  RNA was extracted from seedlings 15 DAG utilizing the Plant RNAEasy extraction Kit (Qiagen).  5 µg/mL of RNA and OligodT12-18 primers (Thermo Fisher Scientific) were used to generate cDNA.  Subsequently, this cDNA was utilized to clone the  intron containing MUSLs (MUSL1 and MUSL3 ) utilizing the following gene specific primers:  MUSL1  (F-5’-CAC CAT GGC CAT GGA TAT ATC TCT-3’; R- 5’-GGC CGG TGA AGG CTG AGA GGT-3’) and MUSL3  (F-5’-CAC CAT GGG TCT TGG TCT TGG TTT T-3’; R-5’- CTG GTT AGA TTT TGG TTC AAT-3).   MUSL2, MUSL4 and MUS, which lack introns, were cloned using genomic DNA and the following gene specific primers: MUSL2 (F- 5’-CAC CAT GTT TCT AAA GTT GTT TCT TCT CTT ATC TTT-3’; R- 5’- ACG TGA AAC CGT ACG TAT GTC TTT AAG AAG A-3’), and MUSL4 (F-5’-CAC CCA TGA GGT TGT CTC TAT GGG GAT CTC T-3’ ; R-5’-GTA GTT GGG AGA AAT GTC TTT GAG AAG TCC A-3’).  Purified DNA was recombined into the TOPO PCR 2.1 ENTRY vector (Thermo Fisher Scientific), according to manufacturer’s instructions, to generate pENTR-MUSL vectors.  LR clonase (Thermo Fisher Scientific) was utilized to recombine the pENTR-MUSL vectors into the pDEST vectors pMDC83 (MUSL1),  pGWB4 (MUSL2 - MUSL4).   All constructs were then placed into Agrobacterium strain GV3101 and transformed into wild-type, mus, musl1 and the mus musl1 double mutant through the floral dip procedure (Clough et al., 1998).   96  3.5.4 Sample preparation  Staining: To visualize cell walls in the shoot epidermis, young cotyledons and first leaves were immersed in an aqueous solution of 5 mg/ml propidium iodide (PI) for 10 minutes, and then washed with distilled water for approximately 30 seconds.   Time-lapse: Seedlings were prepared for time-lapse as in (Peterson and Torii, 2012), but with the following modifications:  All seedlings were grown vertically on ½ MS media containing 0.8% w/v agar.   When seedlings reached 1- 1.5 days after germination (DAG), three to four seedlings were transferred into each chamber of a 2-well chambered cover-glass system (Lab-Tek II Chambered Cover-glass, Cat.No. 12565336 ).   The chambers were then covered by an approximately 5-8 mm slice of preset  ½ MS containing 0.8% agar w/v.  Seedlings prepared in the manner described above remained viable for 2-3 days after preparation.    3.5.5 Microscopy  Image acquisition:  Images were obtained from a Nikon (www.nikon.com) C1 confocal laser scanning microscope and a Perkin Elmer Spinning Disk Confocal microscope.  Both microscopes were equipped with 488 nm and 543 nm lasers.  FIJI was used to analyze all images, while figures were prepared on Adobe Photoshop. Single frame images:  Images were obtained on a 60 X Plan Apo VC 60X objective (NA 1.40) on the Nikon C1 confocal. Time-lapse: All time-lapse images were obtained on a 63X glycerol objective (NA.1.30) mounted on the Perkin Elmer Spinning Disk.   Long term time-lapse images were taken at approximately 2-2.5 hour intervals for a total of 8 hours.  In order to obtain an image of mature guard cell morphology, the final time-lapse images were taken approximately 16-17 hours and 31 hours after the 8 hour time-point, resulting in the 24 hour and 48 hour time-points.     3.5.6 Counts and statistical analysis  Counts:  Cotyledons were counted 15 days after germination (DAG).  Five cotyledons each were randomly selected from wild-type, mus, mus musl1, mus musl2, mus musl3 and 97  mus musl4 .  30 fields from each cotyledon were randomly scored, and counts of abnormal and normal stomata were obtained. Sampling and counts for wild-type, mus, mus MUSL1 OX and wild-type MUSL1 OX lines were obtained utilizing the method described above.  Statistical Analysis: The data for the mus musl1 double mutant counts was analyzed using single factor ANOVA with replication utilizing the Real Statistics program in Excel.                           98  Chapter 4 : The role of MUSTACHES AND MUS-LIKE1 in lateral root development   4.1  Introduction Plants are characterized by their developmental plasticity, which enables new secondary stems and roots to be initiated in response to shifting environmental cues.  Secondary roots, also referred to as lateral roots,  enable plants to efficiently locate and absorb nutrients and water (Malamy and Ryan, 2001).   As lateral roots comprise the bulk of mature roots in fibrous root systems, they are vital for maintaining plant health as well as the adaptation of plants to environmental changes.  Identifying genes influencing lateral root development may assist in improving crop tolerance to varying environmental conditions.  As well, lateral roots, provide an excellent model system for the study of the genetic and cell biological mechanisms regulating organogenesis in plants.   Lateral roots develop through a well characterized pathway that commences when a subset of xylem pole pericycle cells acquire a lateral root founder cell fate.   Once lateral root founder cell fate is established, mature lateral roots form through eight distinct lateral root primordia stages (Stages I-VIII).  Stage I lateral root primordia develop when lateral root founder cells undergo two asymmetric divisions generating single layer primordia consisting of four cells.  Subsequently a second, symmetric, division results in the formation of two layer Stage II lateral root primordia containing eight to twelve cells.  Stage III – VII lateral root primordia then continue to form additional cell layers through repeated symmetric divisions.  Finally, stage VI and VII lateral root primordia, which structurally resemble mature roots, form just before the lateral root exits the primary root (Malamy and Benfey, 1997).    Both lateral root founder cell specification and early lateral root primordia formation are highly auxin-dependent processes.  Auxin responsive genes are characterized by the presence of Auxin Response Elements (AuxREs) in their promoters.  AuxRE activity is regulated by two antagonistic transcription factor families, whereby the Auxin Response Factor (ARF) family activates gene transcription and the AUXIN/INDOLE-3-ACETIC-ACID (AUX/IAA) family represses transcription.  In the absence of auxin, ARFs dimerize with, and 99  are repressed by, AUX/IAA transcription factors.  Degradation of AUX/IAA transcription factors in the presence of auxin results in the de-repression of ARFs, which subsequently results in downstream gene transcription. Mutants in several ARFs and AUX/IAA proteins, including the IAA14/SOLITARY ROOT gain of function mutant and arf7 arf19 double mutants, display severe lateral root defects (Okushima et al., 2007; Lavenus et al., 2013).  Auxin responsive gene modules are essential for the progression of early lateral root development.  Acquisition of lateral root founder cell fate is dependent on oscillating auxin concentrations, which reach a maximum every 15 hours in a sub-set of undifferentiated pericycle cells located in the basal meristem.  Cells exposed to auxin maxima become primed to form lateral root founder cells and subsequently lateral roots (De Smet et al., 2007; Smet, 2012).  Pericycle priming predominately depends on the activation of the IAA28-ARF7 signaling module, which regulates expression of the GATA23 transcription factor.  GATA23 is required for both lateral root founder cell specification in primed xylem pole pericycle cells (De Rybel, 2010), and for initiating the asymmetric division that produces Stage I lateral root primordia (De Rybel et al., 2010b).  High auxin levels also activate the SLR/IAA14-ARF5-ARF7 module, which in turn activates downstream genes such as the LATERAL BOUNDARY DOMAIN (LBD) transcription factors, LBD16 and LBD29.  LBD16 and LBD29, which mediate migration of nuclei and the subsequent establishment of asymmetry in lateral root founder cells, are also required for Stage I lateral root primordia formation (Lavenus et al., 2013; Goh et al., 2012). Genes outside the ARF and IAA families, such as the Receptor-Like Kinase (RLK) ACRINKLY4 (ACR4), have also been shown to participate in lateral root formation.   ACR4 facilitates lateral root formation by promoting asymmetric divisions in the lateral root founder cell, while also restricting divisions in pericycle cells neighbouring the lateral root founder cell (De Smet et al., 2008).    As in lateral root development, asymmetric divisions are also required for the development of stomata, which are epidermal structures that facilitate gas exchange across the shoot epidermis.  Optimal stomatal function requires stomata to be spaced at least one cell apart in the epidermis. One cell spacing is achieved by controlling the initiation and 100  placement of asymmetric divisions.  The ERECTA (ER) family of LRR-RLKs, along with their co-receptor the Leucine-Rich Repeat Receptor-Like Protein (LRR-RLP) TOO MANY MOUTHS (TMM), regulates asymmetric division placement in the stomatal lineage. Several small peptide ligands, including STOMAGEN, bind to ER-TMM complexes to facilitate maintenance of the one cell spacing rule (Shpak et al., 2005; Nadeau, 2009; Sugano et al., 2010; Le et al., 2014b).   Intriguingly, maintenance of one-cell spacing in stomata is an auxin-dependent process (Le et al., 2014a). A number of genes involved in stomatal development have also been shown to participate in additional developmental processes.  The ER kinase family, along with STOMAGEN, mediate root cortex proliferation (Cui et al., 2014).  The Basic Helix Loop Helix (bHLH) transcription factor FAMA regulates terminal differentiation of stomata, while also limiting symmetric divisions during stomatal development.  Recent studies have demonstrated that FAMA is also involved in myrosin idioblast development (Li and Sack, 2014; Shirakawa et al., 2014).  Further, in stomata, the MYB transcription factor, FOUR LIPS (FLP) ensures that young stomata only undergo one symmetric division.  FLP also functions in ovule development, female gametophyte fertility and lateral root  formation (Makkena et al., 2012; Wang et al., 2015).   I previously demonstrated that the Leucine-Rich Repeat Receptor-Like Kinase (LRR-RLK) MUSTACHES (MUS) localizes to stomata, where it enforces bilateral symmetry maintenance (Keerthisinghe et al., 2015, Chapter 2).  MUS also localizes to cell plates and newly divided cells in the root epidermis, suggesting that MUS may have additional functions outside the stomatal pathway.   Here we demonstrate that MUS localizes to lateral roots, and that mus and the closely related LRR-RLK mus-like1 (musl1) exhibit lateral root defects.    4.2 Results   4.2.1 MUSTACHES is expressed in developing lateral roots. Publically available expression databases indicated that MUSTACHES (MUS) is expressed in developing lateral roots. In order to confirm MUS expression and translation in 101  lateral roots, the translational reporter proMUS:MUS-triple GFP (MUS-GFP) was observed in living roots at approximately 1 day intervals, from 0.5 to 4 days after germination (DAG).  Approximately 0.5 DAG, MUS-GFP was visible throughout the shoot epidermis, as well as in the new epidermal divisions in the division zone of the primary root (Fig 4.1 A - B).  MUS-GFP was consistently present in the shoot and primary root epidermis throughout the entire developmental period observed.  Approximately 1.5 DAG, semi-continuous bands of MUS-GFP, expressed in parenchyma tissue surrounding the xylem, were visible in areas above the division zone (Fig 4.1 D).  By 2.5 DAG, strong MUS-GFP was evident in lateral root founder cells and lateral root primordia (Fig 4.1 E-F).  Additionally, MUS-GFP was still weakly expressed in the parenchyma.   At 4 DAG, MUS-GFP continued to display strong lateral root founder cell and lateral root primordia expression, as well as occasional and weak expression in the parenchyma.  MUS-GFP was also co-expressed with the lateral root founder cell and lateral root primordia markers proGATA23:nuclearGFP and proDR5:venusYFP (Fig. 4.1 G-H). 102     103  Figure 4.1  MUS expression in lateral root primordia initiates approximately 2.5 days after germination  Cell walls were visualized by Propidium Iodide (PI) staining (red fluorescence).  MUS was visualized through proMUS:MUS-tripleGFP (Green fluorescence in lateral root primordia and stomatal lineage divisions and cell peripheries).  The lateral root founder cell markers GATA23 and DR5 were visualized through proGATA23:nuclearGFP and proDR5:venusYFP respectively (Green nuclear fluorescence).  All scale bars are 20 µm.  (A - B)  proMUS:MUS-tripleGFP is expressed in the division zone of the root epidermis (A) and in the stomatal lineage cells (B) in seedlings (approximately 0.5 DAG).   (C)  proMUS:MUS-tripleGFP expression was rarely detected in areas outside the division zone at 0.5 DAG.  (D)  proMUS:MUS:tripleGFP is expressed at high levels in the parenchyma surrounding the vascular tissues in seedlings at 1.5 DAG.  Specific proMUS:MUS-tripleGFP expression in lateral root primordia was not detected. (E-F) Intense proMUS:MUS-tripleGFP expression in lateral root primordia becomes increasingly visible 2.5 DAG.  As proMUS:MUS-tripleGFP expression in lateral root primordia increases, its expression the parenchyma surrounding vascular tissue concurrently decreases.   (G-H) proMUS:MUS-tripleGFP is co-expressed with the lateral root founder cell markers proGATA23:nuclearGFP (G) and proDR5:venusYFP (H).  As preliminary observations verified the presence of MUS in lateral roots, a detailed characterization of MUS subcellular distribution patterns during lateral root development was performed using the proMUS-driven MUS-tripleGFP translational reporter, which rescues the mus mutant.  MUS-GFP was first observed in both the cytoplasm and periphery of lateral root founder cells, suggesting that MUS expression and MUS-GFP localization to the plasma membrane initiates during the earliest stage of lateral root development (Fig. 4.2 A - B). As lateral root founder cells transitioned into Stage I lateral root primordia, MUS-GFP was distributed at the periphery of lateral root founder cells, and also to the newly formed asymmetric division sites.  Once Stage I lateral root primordia were formed, MUS-104  GFP localization was consistently retained in asymmetric divisions, while weak peripheral MUS localization was occasionally observed in a subset of Stage I lateral root primordia.  MUS localized to the sites of both pre-existing divisions, as well as to the new symmetric division  sites, in Stage II lateral root primordia. MUS-GFP localization to the general cell periphery was largely absent in Stage II lateral root primordia.   Similarly, MUS localized to all division planes in Stage III – VII lateral root primordia, while general peripheral MUS-GFP localization was largely absent (Fig 4.2 C - F).    Figure 4.2  MUS is expressed in lateral root primordia throughout all stages of lateral root primordia development  Cell walls were visualized through Propidium Iodide (PI) staining (red fluorescence).  MUS was visualized with the proMUS:MUS-tripleGFP translational reporter (green fluorescence).  All scale bars are 20 µm.  (A - B)  MUS-GFP in lateral root founder cells.  MUS-GFP localizes to the periphery and cytoplasm of an undivided lateral root founder cell (A).   MUS-GFP localizes to both the cell periphery and the division plane in a lateral root founder cell transitioning to a Stage I lateral root primordia (B).  (C)  MUS-GFP is confined to the division site,  and not the general cell periphery, in a Stage I lateral root primordium. (D - E)  All divisions sites exhibit MUS-GFP in Stage II  (D) and VI (E) lateral root primordia.   (F)  MUS-GFP continues to be associated with division sites in a mature lateral root, which is exiting the parent root.    105  Lateral root primordia are derived from three pairs of pericycle cells, which develop from three adjacent cell files (Lucas et al., 2013).  MUS-GFP localized intensely to the cytoplasm and periphery of all three pericycle cells during the early stages of lateral root primordia development (Stage I – II). proMUS:MUS-GFP was also co-expressed with pGATA23:nuclearGFP in the three pericycle cells (Fig 4.3).     Figure 4.3 MUS is expressed in 3 consecutive pericycle layers during early stages of lateral root primordia formation  Cell walls visualized through Propidium Iodide (PI) staining (red fluorescence).  MUS was visualized with the pMUS:MUS:tripleGFP translational reporter (green fluorescence).  The lateral root founder cell marker GATA23 was visualized through pGATA23:nuclearGFP transcriptional reporter (green nuclear fluorescence).  All scale bars are 20 µm.  (A - E)  Z-stacks reveal that pGATA23:nuclearGFP and proMUS:MUS:triple GFP co-express in three consecutive pericycle cell layers (denoted by 1,2,3 for GATA23 and a,b,c for MUS).  Single Z-stack frame -7.0 µm depth (A).  Single Z-stack frame -12.0 µm depth (B). Single Z-stack frame - 22.0 µm depth (C).  Orthogonal XZ slice demonstrating that GATA23 is expressed in the nucleus of three consecutive pericycle cell layers (1,2,3) (D).   Orthogonal XZ slice demonstrating that MUS localizes to recent division sites in consecutive pericycle cell layers (a,b,c) (E).  (F - I)  Z-stacks reveal that early stage lateral root primordia exhibit peripheral MUS-GFP distribution in three consecutive pericycle cell layers (denoted by 1, 2, and 3). Single Z-stack frame - 5.1 µm depth (F) Single Z-stack frame - 12.9 µm depth (G) Single Z-stack frame - 19.5 106  µm depth (H).  Orthogonal YZ slice demonstrating that MUS is situated at the cell peripheries in three consecutive pericycle cell layers (I).   4.2.2 Consistent MUSL1 localization is not present in lateral roots  The LRR-RLK MUS-LIKE1 (MUSL1) is closely related to MUS. MUS and MUSL1 appear to have distinct signaling pathways which co-ordinately regulate stomatal morphogenesis (Chapter 3).  Based on their structural similarity and their function in stomatal development, it is possible that MUS and MUSL1 also co-ordinately regulate lateral root development. proMUSL1:MUSL1-GFP and  proMUSL1-GFP were utilized to observe MUSL1 localization and expression respectively.   proMUSL1-GFP expression was characterized in both wild-type and mus backgrounds.  The GFP signal indicated that MUSL1 was consistently expressed throughout the shoot epidermis (Fig 4.4 A, C).  In contrast, MUSL1 expression was rarely observed in primary and lateral roots, as a small sub-set of roots displayed fluorescence, while the majority did not (Fig 4.4 B, D).  Although proMUSL1:GFP appeared to be occasionally expressed in the lateral roots of T1 transformants (Fig 4.3 E), lateral root expression was not visible in the T2 generations of these same lines.   proMUSL1:MUSL1-GFP (MUSL1-GFP) translational fusions constructs were also transformed into wild-type and mus backgrounds.  Fluorescent lines were not, however, recovered in mus, and only one fluorescent line was recovered in a wild-type background.  Approximately 25% of seedlings from this wild-type line exhibited weak proMUSL1:MUSL1-GFP fluorescence throughout the shoot epidermis, whereas, only approximately 4% of roots exhibited proMUSL1:MUSL1-GFP  (n=23).  In seedlings expressing MUSL1-GFP in the root, very faint MUSL1-GFP expression was observed in the pericycle (Fig 4.4 M-N). In addition, slightly stronger MUSL1-GFP signal was visible in the epidermis of the primary root tip (Fig 4 O).   Characterization of pro35S:GFP-MUSL1 expression patterns in wild-type and mus revealed that pro35S:GFP-MUSL1 did not consistently express in lateral roots.   Instead, pro35S:GFP-MUSL1 localized as a continuous band to the stele (parenchyma and vascular tissues), as well as to new division sites in the primary root tip (Fig 4.4 F – L).   107    Figure 4.4   Patterns of MUS- LIKE1 (MUSL1) expression and subcellular distribution throughout the root system  Cell walls visualized through Propidium Iodide (PI) staining (red fluorescence).  Green fluorescence represents proMUSL1:GFP (A – E), pro35S:GFP-MUSL1 (F – L) and proMUSL1:MUSL1-GFP (M – O) respectively.  All scale bars are 20 µm.  (A – E) proMUSL1-GFP expression patterns in wild-type (A – B) and mus (C – E) seedlings.   MUSL1 promoter-dependent GFP is consistently expressed in wild-type (A) and mus (B) shoot epidermises.   proMUSL1:GFP was not consistently expressed in wild-type (B) and mus (D – E) roots.  MUSL1 promoter-dependent GFP expression in the parenchyma of vascular tissue in wild-type (B) and mus (D) roots.  MUSL1 promoter-dependent GFP expression in a structure that may represent a lateral root founder cell, or lateral root primordia, in mus (D).    (F – L)  pro35S:GFP-MUSL1 expression in wild-type (F – G),  mus (H – I)  and musl1 (J – L)  seedlings. GFP-MUSL1 exhibits intense but discontinuous sub-cellular localization 108  throughout vascular tissue parenchyma in maturation and elongation zones of the primary root (F, H, J-K).   MUSL1 also occasionally exhibits weak localization to lateral root founder cell and lateral root primordia division sites (arrows in H and K).  GFP-MUSL1 is consistently localized to division planes in epidermal cells in the division zone of primary roots (G, I, L).   (M – O)   proMUSL1:MUSL1-GFP localization in wild-type.   Weak, but non-specific (diffuse and dispersed), MUSL1-GFP localization is present in vascular tissue parenchyma (M-N).   Weak MUSL1-GFP distribution to division planes is visible in a subset of cells in the division zone of primary roots (arrows, O).   4.2.3 mus mutants contain significantly fewer lateral root primordia. As proMUS:MUS-tripleGFP indicates that MUS is present in lateral root founder cell and young lateral root primordia, MUS likely participates in lateral root primordia initiation and development.  The number of early lateral root primordia stages (founder cell, FC, and Stage I, SI) in mus and wild-type were therefore quantified to identify whether MUS impacts early lateral root primordia development.  proDR5:venusYFP was utilized as a marker for lateral root primordia visualization, and quantification of early stage lateral root primordia was carried out in 2 day intervals for a maximum of 12 days after germination (2 -12 DAG).  From 2-10 DAG, the number of early stage lateral root primordia (FC + S1) in wild-type seedlings remained between 4.38 ± 0.43 (Standard error, S.E.) and 5.54 ± 0.47 lateral root primordia per cm (LRP/cm).  By 12 DAG, the number of early stage LRP/cm decreased to 1.96 ± 0.30 LRP/cm.  In mus, the number of early stage LRP/cm closely resembled that of wild-type at 2 DAG (5.26 ± 0.53). From 4 DAG - 10 DAG, however, the number of early stage lateral root primordia showed a steady decline to reach 2.32 ± 0.18 LRP/cm at 10 DAG, and finally 1.29 ± 0.09 LRP/cm at 12 DAG (Fig 4.5 A). Both raw and log transformed 2-WAY ANOVA indicate that the number of early stage LRP/cm were significantly different between wild-type and mus genotypes (F =26.1 p = 3.36E-17), as well as between a subset of time-points (DAG ) analyzed (F = 27.81 p = 6.95E-07). The decrease in early stage LRP/cm became significantly different by the 8th DAG, and the analysis further indicated a significant decrease in the early stage LRP/cm at 12 DAG compared to all other DAG.  109  As the number of LRP/cm showed a decreasing relationship with time measured in DAGs, time-series plots were created using the mean number of LRP/cm per day against time (in DAG) for wild-type and mus.  The LRP/cm counts were not normal and the regression of LRP/cm vs DAGs was not significant for wild-type (t=-2.59, p = 0.0601).  A correlation of  0.792 was indicated. The trend for wild-type indicated a decline of about 0.286 LRP/cm per day which is an approximately 5% decline in the number of LRP/cm per day.  However, in the case of mus, the regression was highly significant (t= -7.97, p = 0.00134) and the correlation was a high 0.969. The trend for mus indicated a decline of 0.352 LRP/cm per day which is approximately 6.7% decline in the number of LRP/cm per day.  4.2.4  mus, musl1 and mus musl1 double mutants contain significantly fewer emerged lateral roots.   The lower number of LRP/cm in mus might also directly impact the number emerged, mature, lateral roots per cm in (#LRs/cm) in mus.  The production of new lateral root primordia decreases 12 DAG, implying maturation of the primary root.  Thus, the number of emerged lateral roots per cm was calculated 12 DAG to determine whether mus, as well as musl1 and mus musl1, mutants affect mature lateral root numbers.    Analysis of a sample of 25 whole main roots indicated that at 12 DAG, wild-type, mus, musl1 and mus musl1 seedlings contained an average of 19.76 + 0.65, 17.40 + 1.04, 14.96 + 0.81 and 16.04 + 1.12 lateral roots per main root respectively. ANOVA demonstrated that the number of lateral roots per main roots in wild-type was significantly higher than that of both musl1 and mus musl1 (F=4.98, p= 0.0029), but not significantly higher than that of mus.  When variability caused by differences in main root lengths, however, were accounted for by using ANCOVA, the number of lateral roots per main root in wild-type was significantly higher than that of musl1, mus musl1 and also mus (F = 6.48, p=0.000484). In a parallel analysis, the number of lateral roots per whole main root was reduced to the number of lateral roots per cm of main root (LRs/cm) to account for the differences 110  in main root lengths, thereby eliminating the need for ANCOVA. The mean counts for 12 DAG wild-type seedlings contained an average of 5.00 ± 0.14 (S.E.) LRs/cm, mus contained an average of 4.19 ± 0.24 LRs/cm, musl1 contained 3.96 ± 0.14 LRs/cm, and the mus musl1 double mutant contained 3.72 ± 0.24 LRs/cm.  t-Tests indicated that wild-type  had a significantly higher number of LRs /cm compared to mus, musl1 and the mus musl1 double mutant.  There were no significant differences between mus and musl1, mus and mus musl1, or musl1 and mus musl1.  Finally, lines expressing proMUS:MUS-GFP in a mus background had a mean of 4.67 ± 0.17 LRs/cm.  There was no significant difference in the number of LRs/cm between proMUS:MUS-GFP complemented mus lines and wild-type lines (Fig 4.5 B).    111    Figure 4.5   Lateral root density is reduced in mus and musl1 mutant backgrounds  (A) Early stage lateral root primordia density (number of Founder Cell (FC) + Stage I lateral root primordia (Stage I) per cm) is significantly reduced in mus 8 -12 DAG.  Asterisks represent significant differences from wild-type.   (B) Emerged (mature) lateral root density (number of emerged lateral roots per cm) is significantly different from wild-type (asterisks) in mus, musl1 and mus musl1 lateral roots (n=30).  112  4.2.5 MUSL1 overexpression compliments the mus lateral root phenotype The impact of p35S:GFP-MUSL1 (MUSL1 OX) on the number of LRs/cm in mus, mus musl1 and wild-type backgrounds was also investigated.  mus / MUSL1 OX and mus musl1 / MUSL1 OX lines contained 4.91 ± 0.31 and 4.93 ± 0.35 LRs/cm respectively, whereas, wild-type /MUSL1 OX lines exhibited 6.58 ± 0.22 LRs/cm.  t-Tests indicated that the number of LRs/cm in wild-type / MUSL1 OX lines was significantly different from that of mus and wild-type control lines, as well as mus / MUSL1 OX and mus musl1 / MUSL1 OX lines.  In contrast, mus / MUSL1 OX and mus musl1 / MUSL1 OX lines were not significantly different from wild-type control lines. Taken together, this analysis suggests that MUSL1 can positively affect lateral root formation.  4.2.6 Stomatal density is not affected in mus, musl1¸ or mus musl1 backgrounds MUS localizes to developing and mature stomata, and mutations in MUS result in disrupted stomatal morphogenesis (Keerthisinghe et al., 2015). The reduction in lateral root formation in mus, musl1 and mus musl1 mutants prompted us to investigate whether stomatal formation is also reduced in these mutants. Stomatal density in wild-type, mus, musl1 and mus musl1 mutants was analysed using One-factor ANOVA with replication.  This determined that there was no significant difference in stomatal density (stomatal number per field) between wild-type and the three mutants, or between the mutants themselves. Thus, while MUS and MUSL1 are critical for stomatal morphology, they are not involved in stomatal formation per se.  4.2.7 Application of exogenous auxin rescues the mus lateral root phenotype We next explored the effect of auxin on the mus and musl1 mutant lateral root phenotypes. The MUS promoter contains four predicted Auxin Response Factor (ARF) binding sites, while the MUSL1 promoter does not have any ARF binding sites.  The MUSL1 promoter, however, contains one GATA light responsive element motif, which have been implicated in auxin and gibberellin signaling convergence (Davuluri et al., 2003; Richter et al., 2013).  Hypothetically, auxin treatment should not affect the reduction in lateral root number in mus, musl1 or mus musl1 double mutants.  However, auxin may through 113  upregulation of MUSL1 in mus (or MUS in musl1) restore, or partially restore, wild-type levels of lateral root formation. To test these possibilities, the number of emerged LRs/cm in seedlings transplanted 4 DAG, and grown for 8 additional days, on media containing 100 nm of auxin was observed 12 DAG in mus, musl1 and mus musl1.    Analysis of a sample of 30 main roots indicated that 12 DAG auxin-treated wild-type, mus, musl1 and mus musl1 double mutant seedlings contained an average of 17.60 + 0.83, 18.37 + 0.68,  14.76 + 0.84and 14.53 + 0.82 lateral roots per main root respectively (Fig 4.6 A). ANOVA demonstrated that the number of lateral roots per main root in wild-type was significantly higher than that of both musl1 and mus musl1 (F= 6.89 p = 0.000258) but that there were no significant differences between the number of wild-type and mus lateral roots per main root.  When variability caused by differences in main root lengths were accounted for using ANCOVA, the number of lateral roots per main root in wild-type seedlings was significantly higher than that of musl1 and mus musl1 (F = 21.05 p = 6.10E-11).   In a parallel analysis, the number of lateral roots per whole main root was reduced to the number of lateral roots per cm of main root (LRs/cm) to account for the differences in main root lengths, thereby eliminating the need for ANCOVA.   12 DAG wild-type seedlings contained an average of 10.54 ± 0.48 (S.E.) LRs/cm, mus contained an average of 10.34 ± 0.40 LRs/cm, musl1 contained 8.51 ± 0.37 LRs/cm, and the mus musl1 double mutant contained 7.09 ± 0.36 LRs/cm.  ANOVA (F =17.88 ; P < 0.05) indicated that wild-type had a significantly higher number of LRs /cm of main root compared to musl1 and mus musl1, but not compared to mus. There were also significant differences in the LRs/cm between mus and musl1, as well as mus and mus musl1, but not between musl1 and mus musl1.  In summary, in the absence of exogenous auxin there is a significant difference in the number of lateral roots per main root, and the number of LRs/cm of main root, between wild-type and mus.  Auxin treatment in mus increases the number of lateral roots per main root and the number of LRs/cm of main root, such that there is no significant difference between auxin-treated wild-type and mus.  Auxin-treated mus also shows a significant 114  increase in both LRs/ main root and LRs/cm of main root compared to musl1 and mus musl1, which is in contrast to the non-significant difference between these genotypes in the absence of auxin.  Thus, auxin treatment appears to rescue the mus lateral root phenotype.   4.2.8 Auxin rescue of the mus lateral root phenotype likely results from upregulation of MUSL1 expression   Auxin- mediated rescue of the mus lateral root phenotype implies that auxin regulates a complementary pathway that works in the absence of functional MUS, since the mus allele utilized in this study (mus-1)  likely confers a loss of function mutation through a point mutation in the kinase domain.  Thus, quantitative real time PCR (qPCR) was utilized to determine the effect of auxin on MUS and MUSL1 expression levels in auxin-treated roots from mus, musl1, mus musl1 and wild-type backgrounds.    Auxin-treated wild-type roots exhibited a 11.4 ± 2.06 (S.E.) fold induction in MUS expression and a 1.50 ± 0.02 fold induction in MUSL1 expression.  In mus, MUS expression did not differ greatly between control (0.94 ± 0.25) and auxin treated (0.96 ± 0.25) roots.  Consistent with the fact that the musl1 mutation is caused by a T-DNA insertion in the ORF, MUSL1 expression was almost entirely absent in untreated musl1 (0.07 ± 0.05) and mus musl1 (0.03  ± 0.06) roots.  As well, auxin treatment slightly upregulated MUSL1 expression in musl1 (0.17± 0.17), but did not upregulate MUSL1 in mus musl1 roots (0.06  ± 0.05) (Fig 4.6 B).   Intriguingly, in untreated control lines, MUSL1 expression is downregulated (0.37 ± 0.03) in mus roots.  In contrast, MUSL1 expression reverts to, or is up-regulated to, approximately wild-type levels (1.02 ±0.17) in auxin-treated mus roots.   Additionally, auxin treatments appear to slightly down regulate MUS expression from 1.28  ±  0.16  (untreated) to  0.83 ±  0.08 (auxin treated) in musl1 roots.  Interestingly, in the untreated mus musl1 background, MUS expression was upregulated roughly 2-fold (2.14± 0.1) in roots, but was downregulated in auxin-treated roots (0.75 ± 0.24) (Fig 4.6 B).     115   Figure 4.6   Exogenous auxin application rescues the mus lateral root phenotype, but not the musl1 or mus musl1 lateral root phenotype.    A.  Lateral root density is not significantly different from wild-type in mus seedlings treated with exogenous auxin, suggesting auxin mediated rescues of the mus lateral root 116  phenotype.  However, auxin does not complement the musl1 or mus musl1 phenotypes.  Asterisks denote significant differences from wild-type B.  Effect of 8 hour auxin treatment on MUSL1 and MUS expression levels in wild-type, mus, musl1 and mus musl1 backgrounds  4.3 Discussion   4.3.1 MUS and MUSL1 likely function redundantly in lateral root formation  MUS localizes to lateral root primordia throughout lateral root development. Furthermore, mutants of MUS exhibit significantly lower numbers of LRP/cm and emerged LRs/cm when compared to wild-type, indicating that MUS influences lateral root formation.  Mutants of MUSL1, an LRR-RLK closely related to MUS, also display significant reductions in the number of emerged LRs/cm in comparison to wild-type.    Currently, only one other RLK, ACRINKLY4 (ACR4) has been shown to function during the early stages of lateral root primordia development.  Mutants in acr4 exhibit a mild lateral root defect, whereby the number of LRP/cm is significantly increased, while the number of mature emerged; LRs/cm is reduced (De Smet et al., 2008).   ACR4 is a member of the CRINKLY family, which contains four additional kinases, CRINKLY1 – CRINKLY 4 (CRR1 – CRR4), related to ACR4.   De Smet et al. (2008) suggest the increased severity of lateral root phenotypes in acr4 crr double and triple mutants demonstrate that the CRR kinases function redundantly with ACR4.  Although the differences were not significant, as shown by ANOVA , similar to acr4 crr double mutants, the mus musl1 double mutants did display a reduced number of emerged LRs/cm when compared to wild-type, indicating the MUS and MUSL1 may function redundantly. Additionally, MUSL1 over-expression was able to complement the mus musl1 phenotype, further supporting the presence of redundancy. However, although the mus musl1 double mutant contains fewer LRs/cm than both the mus and musl1 single mutants, there is no significant difference between the number of LRs/cm between the double mutant and either of the single mutants.   The mus musl1 double mutant may lack the expected severity due to the presence of three additional kinases in the MUS family (MUSL2 – MUSL4).  Future studies will focus on unmasking additional 117  redundancies in the MUS family by characterizing triple and quadruple mutants of mus, musl1 and one or more of the additional musls (musl2 – musl4). The redundant relationship between MUS and MUSL1 in lateral roots differs greatly from the antagonistic interaction of these receptors in stomatal development (Chapter 3).    Although, visualization of MUSL1 localization in lateral roots proved technically challenging,  pro35S:GFP-MUSL1 and proMUSL1:MUSL1-GFP localization was visible in primary roots.  MUSL1 localization in primary roots appears to mirror proMUS:MUS-tripleGFP localization in the primary root epidermis, suggesting that MUS and MUSL1 may function redundantly in a wide range of different tissues.    4.3.2 MUS may regulate lateral root division and differentiation MUS localizes to both the periphery of lateral root founder cells and to new division sites throughout lateral root primordia development (Stage I – Stage VII). Correspondingly, the number of young precursor cells (lateral root founder cells and Stage I lateral root primordia) is significantly reduced in mus, indicating that MUS may mediate both lateral root differentiation and division.   As in lateral roots, adventitious roots (ARs) enhance the ability of plants to locate and absorb water and nutrients.  However, in contrast to lateral roots, ARs develop from shoot tissues (leaves, stems and hypocotyls) in response to varying environmental conditions (Verstraeten et al., 2014).  Although they differentiate from distinct tissues, adventitious roots and lateral roots display shared developmental characteristics and exhibit some similarities in auxin mediated signaling modules (Verstrateten et al., 2014).   Recently Abu-Abied et al (2015) demonstrated that mutants in the Microtubule Associated Protein (MAP), MICROTUBULE ORGANIZATION 1 (MOR1), along with mutants in the microtubule severing protein BOTERO (BOT1), exhibit defects in AR formation.  Both mor1 and bot1 display reductions in the number of mature ARs, which likely results from the inability of AR primordia to transition from division to differentiation in these mutants (Abu-Abied et al., 2015).  Temperature sensitive alleles of mor1 exhibit defects in the rate of microtubule dynamics (microtubule growth) and microtubule organization (Kawamura and Wasteneys, 2008).  Previous studies have shown that mus also disrupts microtubule array 118  organization, and the direction of microtubule growth (microtubule polarity) in stomata (Keerthisinghe et al., 2015).   Although further studies are required to characterize the exact mechanism by which MUS mediates lateral root formation, MUS, like MOR1 in AR development, may participate in regulating the differentiation of lateral root founder cells into lateral root primordia and lateral roots.      4.3.3 The absence of MUSL1 fluorescence in lateral roots is likely the result of a weak promoter.  MUSL1-GFP fluorescence was not consistently visible in lateral root primordia, or lateral roots.  However, lateral root numbers were reduced in musl1, as well as mus musl1.  Additionally, although pro35S:GFP-MUSL1  (MUSL1 OX) did not localize to lateral roots, lateral root defects in mus, and mus musl1 were complemented by pro35S:GFP-MUSL1 .  Interestingly, MUSL1 OX in wild-type also increased lateral root number.   Therefore, the lack of consistent GFP fluorescence in lateral roots, does not exclude the possibility that low concentrations of endogenous MUSL1 may be present in lateral roots.  Instead, the lack of visible, or specific, fluorescence may result from low activation of the MUSL1 promoter in roots.   A small percentage of wild-type seedlings exhibited very weak proMUSL1:MUSL1-GFP in roots, while a larger percentage of seedlings also displayed faint fluorescence in the leaf epidermis, further illustrating the weakness of the MUSL1 promoter in roots, and throughout the plant.  4.3.4 MUS and MUSL1 expression are likely to be auxin dependent  Genes involved in lateral root development are characterized by their auxin-dependent regulation .   Application of 100nm auxin to wild-type roots resulted in upregulation of MUS by approximately 11-fold, and MUSL1 by approximately 1.5-fold, thereby further confirming the involvement of MUS and MUSL1 in lateral root development.    Keerthisinghe et al. (2015) previously reported that MUS expression in mus was 85% of wild-type expression in whole seedlings.  In the current study, MUS expression levels in mus roots were similar to wild-type.  The differences in expression levels between the 119  two studies may be reflective of differences between the tissues used (whole seedlings vs roots).  Further, the mus mutant is generated by a point mutation in the kinase domain, which likely influences MUS signaling activity (Keerthisinghe, et al., 2015).  Kinase domain activity and conformation have been shown to be necessary for LRR-RLK function in CLV1 and BRI1(Torii, 2008).  As well, (Trotochaud et al., 2000) demonstrated that kinase activity was required for ligand binding to CLV1.  Therefore, though mus may not encode a null, or knock down, mutant, the phenotypes observed in mus may result from the presence of mus mutant transcripts encoding a mutant, inactive, kinase domain.   Application of exogenous auxin rescued the mus lateral root phenotype, but not the musl1 or mus musl1 phenotypes.  Quantification of MUS and MUSL1 expression levels revealed that musl1 is likely a transcript-null mutant, and as a consequence, musl1 expression was insensitive to auxin-mediated upregulation as observed in musl1 and mus musl1 backgrounds.  MUS expression was upregulated approximately 1.3- and 2.1-fold in untreated musl1 and mus musl1 lines.  However, MUS upregulation did not complement the lateral root phenotype in these mutant backgrounds.  The lack of complementation in musl1, despite increased MUS expression levels, demonstrates that the presence of functional MUSL1 is integral for lateral root development.    Additionally, in the absence of exogenous auxin, MUSL1 expression is downregulated in mus roots, while MUS expression is upregulated in musl1 roots. Thus, MUS and MUSL1 expression may be inter-dependent, or co-regulated.      MUS expression levels in in mus were not significantly altered by auxin application, suggesting that auxin-mediated complementation of the mus lateral root phenotype may result from the upregulation of other genes involved in lateral root formation.  Interestingly, despite down regulation of MUSL1 in untreated mus controls, MUSL1 was upregulated to approximately wild-type levels (1.02 ± 0.17 fold relative to wild-type expression levels) in auxin treated mus roots.  Thus, auxin-mediated complementation of the mus lateral root phenotype could result from MUSL1 upregulation.       120  4.4  Summary MUS localizes to lateral root founder cells, lateral root primordia and mature lateral roots.  MUS is also required for formation of early stage lateral root primordia and mature lateral root formation.  Likewise, occasionally, faint MUSL1 localization was also visible in lateral roots, and analysis of mutant lines, as well as OX lines, demonstrated that  MUSL1 participates in lateral root formation.   Thus, the current study demonstrates that two related LRR-RLKS, MUS and MUSL1, function redundantly during lateral root formation.  The correct development of lateral roots from stem cell compartments into mature organs is vital for ensuring that plants are able to adapt to environmental changes.   LRR-RLK mediated signaling pathways play integral roles in regulating the formation and maintenance of stem cell compartments in numerous plant developmental processes, including shoot apical meristem maintenance, root apical meristem maintenance, and the formation of meristemoids in the stomatal pathway.   To date, however, only one RLK, ACR4, has been shown to participate in lateral root development. Therefore, the current work is significant, as it provides evidence that two additional LRR-RLKs participate in lateral root formation.  As such, this study provides a foundation from which to build a greater understanding of how the upstream signaling pathways mediated by LRR-RLKs influence lateral root development, and also how these signaling pathways intersect with IAA-ARF auxin mediated signaling modules during lateral root formation.  Finally, numerous studies have shown that genes that participate in stomatal development also mediate many other vital developmental processes throughout the plant.  The findings of this study further illustrate the importance and value of identifying and characterizing stomatal development genes.     4.5  Materials and methods  4.5.1 Growth of plant materials Arabidopsis thaliana seeds in a wild-type, glabrous1, background (Col-0 gl1 ecotype) were sterilized in a solution of four parts 87.5% ethanol and one part 30% hydrogen 121  peroxide for thirty-five seconds.  Sterilized seeds were then dried and planted on ½ strength Murashige and Skoog (MS) media containing 0.8% w/v agar.  Seeds were germinated in a 21°C incubator under a 16 h light/8 h dark photoperiod.   4.5.2 Plant material Mutants: The mus-1 allele (referred to as mus in the text) contains a point mutation in the kinase domain, suggesting that mus-1 encodes a kinase-inactive protein (Keerthisinghe et al., 2015).  The musl1 T-DNA insertion line (SALK_04666) was obtained from the SALK institute.  Sequencing of the T-DNA line demonstrated that SALK_04666 contained a single insertion in the MUSL ORF.  mus musl1 double mutants were generated by introgressing mus-1 into musl1 (SALK_04666). Double mutants were selected by isolating seedlings with mus stomatal phenotypes and then subsequently genotyping for musl1 utilizing the following primers: Forward – 5’- AGACTTGCTAAAGCGAGTCCC - 3’ and Reverse – 5’- GAATGGTTCTCTCTATCCGCC -3’.   Marker lines : proGATA23:nuclearGFP  (De Rybel et al., 2010) and proDR5:venusYFP (Heisler et al., 2005) were introgressed into proMUS:MUS-tripleGFP (Keerthisinghe et al., 2015).  proDR5:venusYFP was also introgressed into mus.  4.5.3 Expression constructs Translational and Transcriptional fusions : The MUSL1 promoter was cloned using the following primers: proMUSL1-TOPO-F (5’-CACCTTCTCGCCGTGAAAATCGATC-3’) and proMUSL1-TOPO-R (5’- AGATCTGTGGCAGTGAACAAA-3’).  The amplified fragment was cloned into a pENTR D-TOPO ENTRY vector (Thermo Fisher Scientific), according to the manufacturer’s instructions, to generate a pENTR-proMUSL1 construct.  LR clonase II (Thermo Fisher Scientific) was then utilized to recombine pENTR-proMUSL with the pMDC107 DESTINATION (pDEST) vector to generate the proMUSL1-pMDC107 (proMUSL1:GFP) transcriptional fusion.  The MUSL1 Open Reading Frame (ORF) was cloned using the following primers:  MUSL1 Forward (5’-GGCGCGCCATGGCCATGGATATATCTCTG-3’) and MUSL1 Reverse                              (5’- GGCGCGCCTGGCCGGTGAAGGCTG AGAGGT-3’).   Both proMUSL1-pMDC107 and the 122  MUSL1 ORF were then digested with ASCI (NEB Biolabs), and the digested proMUSL1-pMDC107 vector then underwent Calf Intestinal Phosphatase (CIP) treatment (NEB biolabs).  T4 DNA ligase (NEB biolabs) was then utilized to combine the MUSL1 ORF with proMUSL1-pMDC107 to generate the proMUSL1:MUSL1-GFP translational fusion. proMUSL1:GFP and proMUSL1:MUSL1:GFP constructs were introduced into Agrobacterium tumefaciens strain GV3101, and lines resistant to 50 µg/µl kanamycin and 50 µg/µl hygromycin were used for transformation of wild-type and mus plants via the floral dip procedure (Clough and Bent, 1998).   Over-expression (OX) lines :  RNA was extracted from seedlings 15 DAG utilizing the Plant RNAEasy extraction Kit (Qiagen).  5 µg/mL of RNA and OligodT12-18 primers (Thermo Fisher Scientific) were used to generate cDNA, which was subsequently used to clone MUSL1 utilizing the following gene specific primers: MUSL1 ORF-F (5’- CAC CAT GGC CAT GGA TAT ATC TCT - 3’ )  and  MUSL1 ORF-R (5’- GGC CGG TGA AGG CTG AGA GGT-3’).  Purified DNA was recombined into the ENTRY TOPO PCR 2.1 vector (Thermo Fisher Scientific), according to manufacturer’s instructions, to generate pENTR-MUSL1.  LR clonase II (Thermo Fisher Scientific) was subsequently utilized to recombine the pENTR-MUSL1 vector into the DESTINATION (pDEST) vector pMDC83 to generate pro35S:GFP-MUSL1.  pro35S:GFP-MUSL1 was placed into Agrobacterium strain GV3101 and transformed into wild-type, mus, musl1 and the mus musl1 double mutant through the floral dip procedure (Clough and Bent, 1998).    4.5.4 Sample preparation and microscopy Staining : Cell walls in the shoot epidermis (young cotyledons and first leaves) were visualized by immersing seedlings in an aqueous solution of 5 mg/ml propidium iodide (PI) for 10 minutes, and then washing the sample with distilled water for approximately 30 seconds. Cell walls in the root epidermis were visualized by immersing seedlings in 1mg/ml PI for 30 seconds. Samples were then rinsed with distilled water for 30 seconds.   Image acquisition :  Images were obtained from a Nikon (www.nikon.com) C1 confocal laser scanning microscope equipped with 488 nm and 543 nm lasers.  FIJI was used to analyze all images, while figures were prepared on Adobe Photoshop.  Images were 123  obtained on a either a Plan Apo DIC 40X oil (N.A  1.0) or Plan Apo VC 60X oil objective (NA 1.40) on the Nikon C1 confocal.  All Confocal Z-stacks were taken at 0.5 or 0.75µm intervals.  4.5.5 Quantification of lateral root phenotypes Lateral root primordia and lateral root counts : Seeds of wild-type, mus, musl1, and mus musl1 double mutants were planted on ½ MS media. A random sample of 10 whole main roots each from wild-type and mus was examined and the number of lateral root primordia (FC+S1)/cm of main root were counted in 2 day intervals from 2-12 days after germination (DAG).  Counts of the number of emerged lateral roots per main root at 12 DAG were conducted for a sample of 25 randomly selected main roots each from wild-type, mus, musl1 and mus musl1.  Counts of the number of emerged LRs/cm at 12DAG were conducted for a sample of 30 randomly selected main roots each from wild-type, mus, musl1¸and mus musl1 to obtain mean LRs/cm counts.   As a few of the roots showed signs of contamination, 25 were randomly selected for further statistical analysis.   Exogenous Auxin treatments :  Seeds of wild-type, mus, musl1, and mus musl1 were germinated on un-supplemented ½ MS , and then transferred to  ½ MS supplemented with 100 nm of auxin (IAA) when the roots reached approximately 1.5 cm in length (roughly  4 DAG).  The number of emerged lateral roots (#LRs) per whole main root in a sample of 30 seedlings for each genotype was counted at 12 DAG.   4.5.6  Statistical analysis of lateral root phenotypes Lateral root primordia counts : Assumptions of normality and homoscedasticity were tested for raw and log transformed counts. Log transformed counts met both assumptions and these counts were subjected to 2-WAY ANOVA with genotype (wild-type and mus) and time (2,4,6,8,10 and 12 DAG). Time series plots showing the change in number of (FC+S1)/cm with DAGs were created using the mean of 10 number of (FC+S1)/cm values obtained from the n=10 root sample for each DAG. Lateral root /whole main root counts : All assumptions for ANOVA as well as ANCOVA were met by the raw # lateral root /whole main root counts of the wild-type, mus , musl1 and mus musl1 samples (n=25) and the data was analysed using ANOVA, followed by 124  ANCOVA to test whether removing variability caused by differences in main root lengths would improve the ANOVA results. Lateral root /cm counts : Apart from using ANCOVA, reducing the number LRs per whole main root to the number of lateral roots/cm of main root (LRs/cm) can also be used to account for the variability caused by differences in main root lengths. However the # LRs/cm samples of wild-type, mus and mus musl1 did not meet the homoscedasticity assumption Therefore, these samples were tested using Welch’s Test for unequal variances and Student’s t-Tests. Auxin treated lateral roots/ whole main root counts : Assumptions for ANOVA were not met by the raw counts, but boxcox transformed counts met the required assumptions. When ANCOVA was conducted in order to remove the variability caused by differences in main root lengths both raw and Boxcox transformed counts met all assumptions, with the exception of the independence of covariate and independent variable assumptions. Therefore ANCOVA was also utilized as a complementary analysis to ANOVA. Auxin-treated lateral roots/ cm counts : The lateral roots per whole main root counts were reduced to the number of lateral roots per cm of whole main root (LRs/cm) in order to account for variability in the main root lengths.  Both Raw and boxcox transformed counts met ANOVA assumptions.  4.5.7 Quantification and statistical analysis of pro35S:MUSL1-GFP over-expression lines   Seeds of wild-type / pro35S:GFP-MUSL1, mus / p35S:GFP-MUSL1 and mus musl1 / p35S:GFP-MUSL1  MUSL1 Over-Expression (OX) lines were planted on ½ MS media.   The effect of MUSL1 OX on lateral roots was investigated by comparing the number of lateral roots/cm for 30 randomly selected main roots each from wild-type / MUSL1 OX, mus / MUSL1 OX, and mus musl1 / MUSL1 OX lines, as well as wild-type, mus, musl1 and mus musl1 control lines.  As the samples met the normality assumption, but not the heteroscedasticity assumption, the Welch’s t-Test for unequal variances and the Student’s t-Test were used.  125   4.5.8 Comparison of stomatal density phenotypes   The total number of stomata per field (area = 0.1452 micrometers^2) were counted for 30 fields in each of 5 cotyledons from wild-type, mus, musl1 and mus musl1 seedlings 15 DAG.   The samples met normality and heteroscedasticity assumptions and one factor ANOVA with replication was conducted.  4.5.9 Quantitative real-time PCR  To quantify auxin levels,  wild-type, mus1, musl1, and mus1 musl1 seedlings (6 DAG) were incubated in either 50 mL of 100 nm IAA dissolved in dH20 (+IAA) or in  50 mL of dH20 (-IAA) for eight hours.  90-120 roots from each treatment (+IAA and –IAA) were then excised (from wild-type, mus1, musl1, and mus1 musl1 seedlings) and frozen in liquid nitrogen. RNA was extracted from the frozen roots using the Qiagen Plant RNAeasy MiniKit (Qiagen).   cDNA was synthesized from 1 microgram of RNA using an oligo(dT)18 primer and SuperScript III Reverse Transcriptase (Life Technologies). One microliter of cDNA was used for quantitative real-time PCR, utilizing the following gene specific primers  MUS (forward 5’-CGGCGTTTCTTGCTTCTC-3’; reverse, 5’-CCGTTTATGTCGTTGGTGTG-3’), MUSL1 (forward 5’-CTCCATCAATCCAACATGGTTC-3’; reverse, 5’-ATGGGCTTCAAAGTCAGCGT-3’) and ACTIN2 (forward 5’-CATTCCAGCAGATGTGGATCTC-3’; reverse, 5’-ACCCCAGCTTTTTAAGCCTTTG-3’)   .   RT-PCR employed an iQ SYBR Green Supermix (Bio-Rad) that was then analysed using an iQ5 real-time PCR machine (Bio-Rad). Three biological replicates consisting of three technical replicates each were performed.  The Arabidopsis ACTIN2 gene was used to approximate relative mRNA levels.  As primer efficiencies were not equivalent, relative changes in gene expression were quantified using the Pfaffl method (Pfaffl, 2001).    126  Chapter 5 : Conclusions and future directions   5.1  Major conclusions and significance of findings    Over the past decade much progress has been made in the field of stomatal development.   The majority of this work, however, has focused on uncovering mechanisms, gene networks and signaling pathways regulating initiation and placement of asymmetric divisions and symmetric divisions during stomatal development.  In comparison to the level of knowledge gained regarding the processes regulating asymmetric divisions and symmetric divisions, a comprehensive characterization of the last stage of stomatal morphogenesis has not yet been accomplished.  Advanced knowledge regarding the last stage of stomatal morphogenesis is vital as pore morphogenesis occurs at this stage, leading to the symmetry of guard cell shape required for the proper functioning of the stoma.  Thus, characterization of genes and mechanisms regulating pore formation and placement, as well as symmetry generation, represents one of the last unexplored frontiers of stomatal development.   Pore morphogenesis, in addition to its developmental importance,  also provides a model system through which the mechanisms governing how two distinct cells (the two guard cells) co-ordinately regulate the formation of a single complex structure (the pore), can be explored.    Currently, only three genes, the ARF-GEF GNOM (GNOM), the Dof  transcription factor STOMATAL CARPENTER 1 (SCAP1) and the LRR-RLK MUSTACHES (MUS), are known to affect stomatal morphogenesis post- symmetric division.  Moreover, the mechanisms governing how these three genes affect morphogenesis remain under-characterized. In addition, recent studies have demonstrated that several genes that act during the latter half of stomatal development, including FLP, CYCA2 and GNOM, also function in lateral root development, suggesting that genes involved in stomatal development may also represent a pool of candidates that can be mined for functions in lateral root development (Chen et al., 2015).    Thus, the aim of the current thesis was to investigate mechanisms contributing to stomatal morphogenesis, and to determine 127  whether additional stomatal developmental genes also participate in lateral root development through further characterization of the LRR-RLK MUSTACHES.   MUS is of special interest as mus mutants displayed both bilateral symmetry and pore morphogenesis defects, in conjunction with defects in microtubule organization and polarity.  However, apart from the initial characterization of MUS function, no further information has been elucidated regarding the mechanism by which MUS influences stomatal morphogenesis.    Interestingly, Arabidopsis encodes 4 LRR-RLKS, the MUS-LIKE KINASES (MUSLs), which are closely related to MUS, indicating that MUS may represent a hub in a network where microtubule organization, microtubule polarity and stomatal morphogenesis interact.  Thus, the three main objectives of the current thesis were to (1) characterize the impact of MUSTACHES on microtubule organizing centres (MTOCs), microtubules and pore morphogenesis during the guard cell morphogenesis phase of stomatal development, (2) characterize the impact of MUS-LIKE proteins (MUSLs) in stomatal development and finally (3) characterize the role of MUS, and MUSL1, in lateral root development.  The phenotype of the mus mutant leads me to posit that MUS affects stomatal morphogenesis microtubule array organization and microtubule polarity through either MTOC placement or formation.  To further explore this hypothesis the work described in Chapter 2 was performed.   The study first confirmed that the available MTOC markers (GCP2, CLASP, EB1 and NEK6) were suitable to be utilized as stomatal MTOC markers.  These markers were then used to establish a framework that characterized shifts in MTOC localization during wild-type stomatal development through live-cell imaging.  To my knowledge, this is the first study to perform such a characterization, which is significant as it will provide valuable references for future studies on stomatal morphogenesis.   Comparing MTOC localization in wild-type and mus revealed that MTOCs delocalize away from the centre of the ventral wall in mus, suggesting that MUS regulates MTOC placement.  Intriguingly, MTOC delocalization is paralleled by the initiation of a change in the polarity of microtubule growth in mus.  Time-Lapse studies of MTOC in mus demonstrated that MTOC delocalization occurs post- symmetric division, but before the morphogenesis defects.   128  Finally, this work uncovered a previously uncharacterized relationship between MUS and the MTOC-associated proteins CLASP and NEK6, as double mutants of these proteins generate a severe synergistic phenotype, which consists of a large number of ‘capsule-shaped’ stomata that lack MTOCs and are unable to form a pore.  The severe phenotype of capsule stomata in mus nek6 suggests that functional MUS and NEK6 proteins are required for MTOC formation.  In addition to the novelty of these genetic interactions, these findings are significant because they have unmasked additional cellular components that may affect MTOC formation and stomatal pore morphogenesis.  Future studies can use these proteins to uncover further mechanisms and components involved in MTOC formation.   Arabidopsis contains 4 LRR-RLKs that are related to MUS.  As LRR-RLKs often function redundantly in developmental processes, it was expected that one or more of these MUS-LIKE kinases (MUSLs) would function in conjunction with MUS.  Therefore, in chapter 3, the relationship between MUS and the MUSLs was investigated.  Of the four MUSLs, MUSL1 proved to be of particular interest, as MUSL1 OX lines displayed sub-cellular localization patterns that closely resembled that of native MUS, except for one key difference.  In contrast to MUS, which never localizes to ventral walls in young guard cells, both MUSL1 OX and translational MUSL1-GFP lines display symmetric division localization in young guard cells.  Additionally, when expressed in a mus background MUSL1-OX results in a gain of function phenotype which causes an inversion of guard cell polarity, referred to as the butterfly (bf) phenotype. The opposing MUS and MUSL1 localization patterns, along with the gain of function bf phenotype in mus backgrounds, lead to the development of a hypothetical model for regulation of pore placement, whereby MUS and MUSL1 contribute to pore morphogenesis through opposing signaling pathways.   These findings are significant as they describe a potential mode of action between MUS and MUSL1 genes.  Discovery of the relationship between TMM and ER more than a decade ago laid the foundation for the subsequent studies, which ultimately revealed the intricate relationships between genes regulating the initiation of stomatal development.  Likewise, it is hoped that the discovery of the MUS and MUSL1 gene module model will lead to the further deepening of our 129  understanding of the mechanisms and process by which pore placement and morphogenesis are maintained in the future.   Finally, several stomatal genes functioning during the guard cell morphogenesis phase have been shown to be involved in lateral root development.   MUS is a late acting stomatal development gene, which was shown by public expression databases to be highly expressed in lateral root primordia.   Therefore,  chapter 4 describes studies performed to determine whether MUS and MUSL1, contribute to lateral root development.  MUS localizes to lateral roots during all stages of lateral root development, and mus mutants exhibit a significantly reduced number of lateral root primordia compared to wild-type.  In addition mus, musl1 and mus musl1 exhibit reduced emerged lateral root numbers compared to wild-type.   Although significant differences were not found, the reduction in lateral root number was compounded in mus musl1, which showed a slightly greater decrease in the mean number of lateral roots/cm  compared to all other genotypes checked.  The lateral root phenotypes, in conjunction with the ability of the MUSL1 OX line to complement the mus phenotype, suggest that MUS and MUSL1 function redundantly in lateral root development.  Thus, in contrast to stomata, in which MUS and MUSL1 appear to function through opposing pathways, MUS and MUSL1 function in the same process (either as part of one pathway, or as part of different pathways) in lateral root development.   This work confirmed that a pair of redundant kinases mediates developmental events in early lateral root development.   This finding is significant because, although lateral root formation is an important developmental process, to date very few kinases have been found to function in lateral root development.   As such, MUS and MUSL1 represent a novel genetic pathway, the further exploration of which will lead to a greater understanding of upstream signaling processes involved in lateral root development.    Thus, in conclusion, the current thesis provides several significant insights into the function of MUS, and members of the MUS-LIKE family (MUSL1) throughout the plant.   First, the work in this thesis furthered our understanding of MUS function by demonstrating that MUS is likely responsible for regulating both the symmetrical placement of MTOCs, and also the subsequently occurring polarization of microtubule movement, thereby enabling 130  the formation of a symmetrical wild-type stomata.  As well, this work demonstrated that MUS may function in parallel with CLASP and NEK6 in order to ensure MTOC formation during stomatal morphogenesis. Second, a potential model describing how MUS and MUSL1 may co-ordinate pore placement and morphogenesis through opposing signaling pathways was elucidated.  Finally, a previously undescribed function of the MUS and MUSL1 kinases in lateral root development was characterized.      5.2  Future directions  5.2.1 Chapter 2 :  the role of MUSTACHES in the regulation of microtubule organizing centers and pore morphogenesis in stomata  The work described in chapter 2 confirmed that MUS is responsible for the symmetrical placement of MTOCs in the centre of the ventral wall.   However, this finding raises several new questions that must be further evaluated in order to gain a complete understanding of MUS function in this process.   First, it needs to be determined whether MUS influences MTOC localization by directly phosphorylating one, or several γ-TURC components, or whether MUS indirectly influences MTOC localization by phosphorylating a target upstream of the γ-TURC.   To do this will require identifying the target substrates of MUS, as well as additional downstream components of the MUS pathway.  Additionally, identifying the MUS ligand remains of great interest.  Finally, the role of the MUS-NEK6-CLASP pathway(s) in regulation of MTOC formation requires further refining.     GCP2, CLASP and NEK6 are associated with the guard cell  MTOC (Chapter 2), and therefore could be MUS substrates.  Additional candidates for MUS targets include the γ-TURC components NEDD1 and GCP3 INTERACTING PROTEIN 1 / MOZART1 (GIP1/MOZART1).  Phosphorylation of NEDD1 orthologues in Xenopus and human cells results in the targeting of γ-TURCs to spindles and centrosomes (Sdelci et al., 2012).   In addition, Nakamura et al (2012) demonstrated that GIP1/MOZART1 is present in a subset of Arabidopsis γ-TURCs that lack NEDD1.  Nakamura et al ( 2012) postulate that GIP1/MOZART1 and NEDD1 may represent plant ‘accessory factors’ that mediate the 131  nucleation of plant microtubules in different cellular locations.  Thus, both NEDD1 and GIP1/MOZART1 represent promising candidates for MUS interactions.    Kinase assays can be utilized to assess whether MTOC-related proteins, and/or components of the γ-TURC are directly phosphorylated by MUS.   Kinase assays involve the purification of both the kinase and the potential substrate and their incubation with ATP.  Confirmation of substrate phosphorylation is subsequently achieved either through utilizing radioactively labelled ATP and autoradiography, or through detection of increased fluorescence levels following phosphorylation in fluorescently tagged substrates.  Additionally, phosphoproteomics assays on mus and wild-type samples can be utilized to determine which proteins lack phosphorylation in mus, thereby leading to the identification of interactions between MTOC components ,or MTOC associated proteins, and MUS.  Identifying proteins that function in signaling pathways is often achieved through isolating mutants with similar phenotypes.   However, to date, no other mutants with similar phenotypes similar to mus have been described.  Therefore, it may be of value to perform Ethyl MethaneSulfonate (EMS) mutagenesis screens to identify additional MUS substrates through identification of mus suppressors.  MUS enhancers, such as mutations in a MUS substrate would likely have a phenotype similar to, or indistinguishable from, mus.  Therefore, it would be more efficient to concentrate on identifying mus suppressors, such as those resulting in the constitutive activation of a MUS substrate.   EMS mutagenesis screens should be performed in mus lines expressing proGCP2:GCP2-GFP to simultaneously assess the effect any suppressors or enhancers of mus on MTOC localization.  Further, EMS mutagenesis of proMUS:MUS-tripleGFP expressing lines should also be undertaken to uncover genes that impact MUS localization to the periphery of young guard mother cells and guard cells, as this peripheral localization appears to be critical for MUS function (Chapter 3).   In order to compensate for difficulties in isolating ligands, it would be of value to employ multiple ligand discovery methods.   Members of previously characterized ligand families should be examined for involvement in the MUS signaling pathway.  Several members of the EPF/EPFL family, including EPF1, EPF2, CHALLAH/EPFL6 and STOMAGEN/EPFL9, have already been shown to regulate asymmetric divisions during 132  stomatal development.  Expression patterns of the uncharacterized EPF members should be examined to determine whether one, or more, of the EPFs/EPFLs may function in guard mother cells or guard cells.  As well, Jun et al (2010) demonstrated that a member of the CLAVATA3/ ENDOSPERM SURROUNDING REGION (CLE) family of ligands, CLE9, is expressed specifically in stomata lineage cells throughout the meristemoid to young guard cell stages of stomatal development.    MUS is also expressed during meristemoid- young guard cell stages, and due to the overlap in expression patterns between MUS and CLE9, CLE9 could represent a potential candidate for a MUS ligand.    Finally, MUS ligands might be identified by cross-linking proteins in a proMUS:MUS-tripleGFP background followed by Co-Immunoprecipitation (Co-IP) of MUS-ligand complexes through application of cross-liked samples to affinity columns containing GFP specific antibodies.  Lee et al (2012) recently utilized Co-IP experiments to successfully validate binding interactions between the ER/ERL family, TMM and the EPF ligands, suggesting that Co-IP may also be successfully employed to identify MUS ligands.   Potential MUS ligands, as well as other proteins obtained through Co-IP, would then be identified through Mass Spectrometry (MS) analysis.    A deeper understanding of the impact of MUS, CLASP and NEK6 on MTOC formation would likely greatly expand the current understanding of pore and guard cell morphogenesis in stomata.    To further elucidate relationships between these proteins, clasp nek6 double mutants, as well as clasp mus nek6 triple mutants should be generated and examined.   In addition, the mus nek6 double mutant appears to exhibit weaker GCP2 fluorescence, implying that the expression levels of GCP2, and potentially other MTOC components, are downregulated in the double mutant.  Examination of gene expression levels through RNA sequencing (RNAseq) or microarrays in mus nek6 may help in the identification of down-regulated genes, thereby providing additional insight into proteins which participate in MUS-NEK6-CLASP mediated MTOC formation. Niederhuth et al., (2013) utilized RNAseq in hae hsl2 double mutants to identify potential downstream targets of the HAE-HSL2 mediated floral organ abscission pathway, thereby demonstrating the utility of RNAseq in identifying potential RLK substrates and targets.  Further, it is interesting to note that gnom (gn) mutants display mutant stomata that are strikingly similar to mus nek6 133  double mutants, in that they lack a pore and are abnormally expanded (capsule shaped) (Le et al., 2014a).  As GNOM is responsible for recycling of PIN proteins to cellular membranes, the overlap in gnom and mus nek6 phenotypes may suggest that auxin participates in mediating MTOC formation.  The role of auxin in MTOC formation could represent an exciting area of research for future studies.    5.2.2  Chapter 3 : MUSTACHES, and the closely related kinase MUS-LIKE1, may regulate positioning of the stomatal pore  In chapter 3, I postulated that MUSL and MUSL1 pathways are antagonistic, or in opposition to each other, and that this antagonistic relationship results in the symmetrical placement of the pore.  This model needs to be validated through further characterization of the MUS and MUSL1 signaling pathways.  Nevertheless, several interesting research questions emerge from this model, including whether the differential localization patterns between MUS and MUSL1 contribute to their distinct functions, whether the bf phenotype is an artifact of the MUSL1-GFP construct, why null musl1 mutants do not display a stomatal phenotype, and whether MUSL1 and MUS act on the same substrates.     Domain swaps between MUS and MUSL1 should be attempted in order to further validate the afore-mentioned model of MUS-MUSL1 function.  Before domain swaps are attempted, individual domains of MUS and MUSL1 (signal sequence, extracellular domain, trans-membrane domain, kinase domain and combinations thereof) should be fused to GFP and expressed in Arabidopsis to determine which domains are required for MUS and MUSL1 to localize to the stomatal periphery and ventral wall respectively. These domains should then be swapped between MUS and MUSL1, and the resulting constructs expressed in mus musl1 background.  Domain-swapped MUSL1 proteins would likely localize to the stomatal periphery and domain-swapped MUS proteins would likely localize to the symmetric division.  Therefore plants harboring domain-swapped constructs would be expected to consist of predominately bf stomata, as MUS would emit a signal preventing MTOCs from localizing to the ventral wall, while MUSL1 would emit a signal which prevented MTOCs from localizing to the ventral wall, and/or enhance localization of MTOCs to the cell periphery.   134   The bf phenotype may be a result of dominant negative interactions of the MUSL1-GFP fusion protein generated by interference of the attached GFP with the native function of the protein.  In these cases the GFP may prevent binding of a partner protein, inhibit kinase domain activity (in the case of C-terminal GFP attachments) or interfere with ligand binding (in the case of N-terminal GFP attachments).  The bf phenotype is unlikely to be due to such a dominant negative interaction, as mis-expression lines consisting of GFP fused to either a LRR or kinase domain deleted MUSL1 exhibited localization to the symmetric division, but failed to generate bf phenotypes.  These results suggest that it is MUSL1, and not the attached GFP, which generates the bf phenotype (Tsuyoshi Nakagawa, personal communication).  However, in order to further validate that the bf phenotype is due to MUSL1 OX, a control line lacking GFP (pro35S:MUSL1) should be generated and transformed into a variety of backgrounds.  The role of MUSL1 will be verified by the generation of bf phenotypes in these control lines.   Despite the involvement of MUSL1 OX in generating the bf phenotype, the musl1 mutant does not display a stomatal phenotype.   As expression of the native MUSL1 promoter is weak, it may be that the MUS signaling pathway is sufficient to maintain stomatal symmetry in the absence of MUSL1, or alternatively this may indicate the presence of proteins that function redundantly with MUSL1 in stomatal development.  Split ubiquitin assays of membrane bound proteins in Arabidopsis revealed that MUSL1 interacts with several LRR-RLKs (Jones et al., 2014).  Higher order mutants of the MUSL1 interactors, as well as MUSL2-MUSL4 should be analyzed to validate the presence or absence of redundancies.  Additionally, putative musl2 – musl4 mutants should be subjected to qPCR in order to verify whether they are null or knock-down mutants.   As well, it would be beneficial to express a kinase-defective, or kinase-null MUSL1 in wild-type, and mus backgrounds to determine if a kinase-defective version of MUSL1 would generate a stomatal phenotype.   Finally, it would be of interest to determine if MUS and MUSL1 signal through the same substrates, including γ-TURC subunits.  Again, kinase assays between MUSL1 and the MTOC associated proteins and γ -TURC components, EMS screens, or Co-IP followed by MS 135  can be utilized to identify MUSL1 substrates.    Additionally, Tsuyoshi Nakagawa (personal communication) has generated seedlings containing very high percentages of bf phenotypes by mis-expressing MUSL1 through a variety of early stage stomatal promoters, including that of TMM.   RNAseq utilizing these MUSL1 mis-expression lines may also be useful in identifying potential MUSL1 substrates through downstream targeted genes.  It is not expected that MUSL1 mis-expression directly influences transcriptional regulation of down-stream targets.  However, it is possible that mis-expression, and OX, of MUSL1 would increase MUSL1 activity, which in turn may indirectly stimulate the transcription of potential MUSL1 targets.  Finally, a publicly available co-expression database, ATTED-II identifies the MICROTUBULE ASSOCIATED PROTEIN 65 (MAP65) as the gene that is most highly co-expressed with MUSL1.   The MAP65 family of proteins impacts organization of cortical microtubule arrays through mediation of microtubule bundling (Lucas et al., 2011).   Since both the mus and bf phenotypes likely result from delocalization and misplacement of the MTOC, it would be of further interest to explore the relationship between MUSL1 and MAP65.     5.2.3 Chapter 4 : the role of MUS and MUS-LIKE1 in lateral root development   In chapter 4, the involvement of MUS and MUSL1 in lateral root development was characterized.   As this chapter represents an initial characterization of MUS and MUSL1 localization patterns and mutant phenotypes, several aspects MUS and MUSL1 function in lateral root development require further study.  A key area of future research will likely involve defining whether MUS and MUSL1 represent a novel, independent, gene-module regulating lateral root development, or whether MUS and MUSL1 function in conjunction within previously described lateral root gene modules.  Also, it may be necessary to better define MUSL1 localization in lateral roots.   Additionally, further exploration of the presence of redundant interactions between MUS and MUSL1 in other parts of the plant, specifically the primary root, is required. Reverse Transcriptase - Quantitive Polymerase Chain Reaction (RT-QPCR), in combination with double mutant analysis, should be utilized to assess whether MUS and MUSL1 function with previously described lateral root genes, including the RLK ACR4 and 136  various ARF/IAA gene modules.  Further candidates of interest for MUS-MUSL1 interactors may include stomatal developmental and morphogenesis genes which also have established roles in lateral root development, such as GNOM and FLP.   Additional MUS and MUSL1 interactors may be identified through Co-IP followed by MS analysis. Visualization of MUSL1 localization in the lateral roots was difficult to observe in both native promoter and 35S OX lines.  However, characterization of MUSL1 localization in lateral roots will be vital for defining its relationship with MUS during lateral root development.  Thus, fusion of the native MUSL1 promoter, and the MUSL1 ORF, to a triple GFP construct may enhance visualization of MUSL1 localization in lateral roots.  Moreover, as auxin appears to upregulate MUSL1 expression, the strength and location of expression in proMUSL1:MUSL1-GFP, or pro35S:MUSL1-GFP seedlings exposed to prolonged applications of auxin should be observed to determine if clearer visualization of MUSL1 is possible post auxin treatment.  Alternately, techniques that enable detection of weakly expressed proteins, such as immuno-localization or in-situ hybridization, may also be useful in determining native MUSL1 expression and localization patterns.   MUS, as well as MUSL1, localizes to new epidermal divisions throughout the primary root, suggesting that MUS and MUSL1 may also function redundantly in the primary root.  Thus, it may be beneficial to carefully assess whether mus musl1 double mutants display primary root defects.  Further, Co-IP and MS analysis should be performed on whole root tissues, to obtain candidates for potential MUS and MUSL1 interactors in the root.  Several genes that function in lateral roots, including ACR4/CCR4 and the GRAS transcription factor family members SCARECROW (SCRW) and SHORT ROOT (SHR) also function in primary root development (Tian et al., 2014).   Thus, any interactors uncovered through Co-IP of whole root tissues might also be candidates for MUS and MUSL1 interactors in lateral roots.    The current work outlines several previously unknown facets of MUS and MUSL1 function throughout the plant.  It is hoped that this work will provide a framework, or base, from which future studies can build a comprehensive understanding of not only mechanisms governing the vital process of pore morphogenesis in stomata, but also  the relationships between genes regulating stomatal and lateral root development.     137  References  Abu-Abied, M., Rogovoy, O., Mordehaev, I., Grumberg, M., Elbaum, R., Wasteneys, G.O., and Sadot, E. (2015). Dissecting the contribution of microtubule behaviour in adventitious root induction. J. Exp. Bot. 66: 2813–2824. Ambrose, C., Allard, J.F., Cytrynbaum, E., and Wasteneys, G. (2011). A CLASP-modulated cell edge barrier mechanism drives cell-wide cortical microtubule organization in Arabidopsis. Nat. Commun. 2: 412–430. Ambrose, C. and Wasteneys, G.O. (2011). Cell edges accumulate gamma tubulin complex components and nucleate microtubules following cytokinesis in Arabidopsis thaliana. PLoS One 6: e27423. Ambrose, C. and Wasteneys, G.O. (2014). Microtubule initiation from the nuclear surface controls cortical microtubule growth polarity and orientation in Arabidopsis thaliana. Plant Cell Physiol. 55: 1636–1645. Ambrose, J.C., Shoji, T., Kotzer, A.M., Pighin, J. a, and Wasteneys, G.O. (2007). The Arabidopsis CLASP gene encodes a microtubule-associated protein involved in cell expansion and division. Plant Cell 19: 2763–75. Bella, J., Hindle, K.L., McEwan, P.A., and Lovell, S.C. (2008). The leucine-rich repeat structure. Cell. Mol. Life Sci. 65: 2307–2333. Bergmann, D.C., Lukowitz, W., and Somerville, C.R. (2004). Stomatal development and pattern controlled by a MAPKK kinase. Science 304: 1494–7. Bergmann, D.C. and Sack, F.D. (2007). Stomatal development. Annu. Rev. Plant Biol. 58: 163–81. Bisgrove, S.R., Lee, Y.-R.J., Liu, B., Peters, N.T., and Kropf, D.L. (2008). The microtubule plus-end binding protein EB1 functions in root responses to touch and gravity signals in Arabidopsis. Plant Cell 20: 396–410. Castells, E. and Casacuberta, J.M. (2007). Signaling through kinase-defective domains: the prevalence of atypical receptor-like kinases in plants. J. Exp. Bot. 58: 3503–11. Chen, Q. et al. (2015). A coherent transcriptional feed-forward motif model for mediating auxin-sensitive PIN3 expression during lateral root development. Nat. Commun. 6: 8821. Chen, X., Grandont, L., Li, H., Hauschild, R., Paque, S., Abuzeineh, A., Rakusová, H., Benkova, E., Perrot-Rechenmann, C., and Friml, J. (2014). Inhibition of cell expansion by rapid ABP1-mediated auxin effect on microtubules. Nature 516: 90–3. Clark, S., Running, M., and Meyerowitz, E. (1993). CLAVATA1, a regulator of meristem and 138  flower development in Arabidopsis. Development 119: 397–418. Clark, S.E., Williams, R.W., and Meyerowitz, E.M. (1997). The CLAVATA1Gene Encodes a Putative Receptor Kinase That Controls Shoot and Floral Meristem Size in Arabidopsis. Cell 89: 575–585. Clough, S.J. and Bent, A.F. (1998). Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16: 735–43. Cui, H., Kong, D., Wei, P., Hao, Y., Torii, K.U., Lee, J.S., and Li, J. (2014). SPINDLY, ERECTA, and Its Ligand STOMAGEN Have a Role in Redox-Mediated Cortex Proliferation in the Arabidopsis Root. Mol. Plant 7: 1727–1739. Van Damme, D., De Rybel, B., Gudesblat, G., Demidov, D., Grunewald, W., De Smet, I., Houben, A., Beeckman, T., and Russinova, E. (2011). Arabidopsis α Aurora kinases function in formative cell division plane orientation. Plant Cell 23: 4013–24. Davuluri, R. V, Sun, H., Palaniswamy, S.K., Matthews, N., Molina, C., Kurtz, M., and Grotewold, E. (2003). AGRIS: Arabidopsis gene regulatory information server, an information resource of Arabidopsis cis-regulatory elements and transcription factors. BMC Bioinformatics 4: 25. DeYoung, B.J., Bickle, K.L., Schrage, K.J., Muskett, P., Patel, K., and Clark, S.E. (2006). The CLAVATA1-related BAM1, BAM2 and BAM3 receptor kinase-like proteins are required for meristem function in Arabidopsis. Plant J. 45: 1–16. Dharmasiri, N., Dharmasiri, S., and Estelle, M. (2005). The F-box protein TIR1 is an auxin receptor. Nature 435: 441–5. Dong, J., MacAlister, C.A., and Bergmann, D.C. (2009). BASL controls asymmetric cell division in Arabidopsis. Cell 137: 1320–30. Dyachok, J., Zhu, L., Liao, F., He, J., Huq, E., and Blancaflor, E.B. (2011). SCAR mediates light-induced root elongation in Arabidopsis through photoreceptors and proteasomes. Plant Cell 23: 3610–26. Fu, Y., Gu, Y., Zheng, Z., Wasteneys, G., and Yang, Z. (2005). Arabidopsis interdigitating cell growth requires two antagonistic pathways with opposing action on cell morphogenesis. Cell 120: 687–700. Fu, Y., Xu, T., Zhu, L., Wen, M., and Yang, Z. (2009). A ROP GTPase signaling pathway controls cortical microtubule ordering and cell expansion in Arabidopsis. Curr. Biol. 19: 1827–32. Fujita, M., Himmelspach, R., Hocart, C.H., Williamson, R.E., Mansfield, S.D., and Wasteneys, G.O. (2011). Cortical microtubules optimize cell-wall crystallinity to drive unidirectional growth in Arabidopsis. Plant J. 66: 915–28. 139  Galatis, B. and Apostolakos, P. (2004). The role of the cytoskeleton in the morphogenesis and function of stomatal complexes. New Phytol.: 613–639. Gao, X. and Guo, Y. (2012). CLE peptides in plants: proteolytic processing, structure-activity relationship, and ligand-receptor interaction. J. Integr. Plant Biol. 54: 738–45. Goff, K.E. and Ramonell, K.M. (2007). The role and regulation of receptor-like kinases in plant defense. Gene Regul. Syst. Bio. 1: 167–75. Goh, T., Joi, S., Mimura, T., and Fukaki, H. (2012). The establishment of asymmetry in Arabidopsis lateral root founder cells is regulated by LBD16/ASL18 and related LBD/ASL proteins. Development 139: 883–893. Gómez-Gómez, L. and Boller, T. (2000). FLS2: an LRR receptor-like kinase involved in the perception of the bacterial elicitor flagellin in Arabidopsis. Mol. Cell 5: 1003–1011. Hanks, S. and Hunter, T. (1995). Protein kinases 6. The eukaryotic protein kinase superfamily: kinase (catalytic) domain structure and classification. FASEB J 9: 576–596. Hara, K., Kajita, R., Torii, K.U., Bergmann, D.C., and Kakimoto, T. (2007). The secretory peptide gene EPF1 enforces the stomatal one-cell-spacing rule. Genes Dev. 21: 1720–5. Heisler, M.G., Ohno, C., Das, P., Sieber, P., Reddy, G. V, Long, J.A., and Meyerowitz, E.M. (2005). Patterns of auxin transport and gene expression during primordium development revealed by live imaging of the Arabidopsis inflorescence meristem. Curr. Biol. 15: 1899–911. Hunt, L. and Gray, J.E. (2009). The signaling peptide EPF2 controls asymmetric cell divisions during stomatal development. Curr. Biol. 19: 864–9. J. Marc, Y.M. and B.A.P. (1989). A planar microtubule-organizing zone in guard cells of Allium: experimental depolymerization and reassembly of microtubules on JSTOR. Planta: 530–540. Jeong, S., Trotochaud, A.E., and Clark, S.E. (1999). The Arabidopsis CLAVATA2 gene encodes a receptor-like protein required for the stability of the CLAVATA1 receptor-like kinase. Plant Cell 11: 1925–34. Jones, A.M. et al. (2014). Border control--a membrane-linked interactome of Arabidopsis. Science 344: 711–6. Kanaoka, M.M., Pillitteri, L.J., Fujii, H., Yoshida, Y., Bogenschutz, N.L., Takabayashi, J., Zhu, J.-K., and Torii, K.U. (2008). SCREAM/ICE1 and SCREAM2 Specify Three Cell-State Transitional Steps Leading to Arabidopsis Stomatal Differentiation. PLANT CELL ONLINE 20: 1775–1785. Kawamura, E. and Wasteneys, G.O. (2008). MOR1, the Arabidopsis thaliana homologue of Xenopus MAP215, promotes rapid growth and shrinkage, and suppresses the pausing 140  of microtubules in vivo. J. Cell Sci. 121: 4114–23. Keerthisinghe, S., Nadeau, J.A., Lucas, J.R., Nakagawa, T., and Sack, F.D. (2015). The Arabidopsis leucine-rich repeat receptor-like kinase MUSTACHES enforces stomatal bilateral symmetry in Arabidopsis. Plant J. 81: 684–94. Kollman, J.M., Merdes, A., Mourey, L., and Agard, D.A. (2011). Microtubule nucleation by γ-tubulin complexes. Nat. Rev. Mol. Cell Biol. 12: 709–721. Křeček, P., Skůpa, P., Libus, J., Naramoto, S., Tejos, R., Friml, J., and Zažímalová, E. (2009). The PIN-FORMED (PIN) protein family of auxin transporters. Genome Biol. 10: 249. Kumpf, R.P., Shi, C.-L., Larrieu, A., Stø, I.M., Butenko, M.A., Péret, B., Riiser, E.S., Bennett, M.J., and Aalen, R.B. (2013). Floral organ abscission peptide IDA and its HAE/HSL2 receptors control cell separation during lateral root emergence. Proc. Natl. Acad. Sci. U. S. A. 110: 5235–40. Lai, L.B. (2005). The Arabidopsis R2R3 MYB Proteins FOUR LIPS and MYB88 Restrict Divisions Late in the Stomatal Cell Lineage. PLANT CELL ONLINE 17: 2754–2767. Lau, O.S., Davies, K.A., Chang, J., Adrian, J., Rowe, M.H., Ballenger, C.E., and Bergmann, D.C. (2014). Direct roles of SPEECHLESS in the specification of stomatal self-renewing cells. Science 345: 1605–9. Lavenus, J., Goh, T., Roberts, I., Guyomarc’h, S., Lucas, M., De Smet, I., Fukaki, H., Beeckman, T., Bennett, M., and Laplaze, L. (2013). Lateral root development in Arabidopsis: Fifty shades of auxin. Trends Plant Sci. 18: 1360–1385. Le, J. et al. (2014a). Auxin transport and activity regulate stomatal patterning and development. Nat. Commun. 5: 3090. Le, J., Mallery, E.L., Zhang, C., Brankle, S., and Szymanski, D.B. (2006). Arabidopsis BRICK1/HSPC300 is an essential WAVE-complex subunit that selectively stabilizes the Arp2/3 activator SCAR2. Curr. Biol. 16: 895–901. Le, J., Zou, J., Yang, K., and Wang, M. (2014b). Signaling to stomatal initiation and cell division. Front. Plant Sci. 5: 297. Lee, E., Lucas, J.R., Goodrich, J., and Sack, F.D. (2014a). Arabidopsis guard cell integrity involves the epigenetic stabilization of the FLP and FAMA transcription factor genes. Plant J. 78: 566–77. Lee, E., Lucas, J.R., and Sack, F.D. (2014b). Deep functional redundancy between FAMA and FOUR LIPS in stomatal development. Plant J. 78: 555–65. Li, J. and Chory, J. (1997). A putative leucine-rich repeat receptor kinase involved in brassinosteroid signal transduction. Cell 90: 929–38. 141  Li, M. and Sack, F.D. (2014). Myrosin idioblast cell fate and development are regulated by the Arabidopsis transcription factor FAMA, the auxin pathway, and vesicular trafficking. Plant Cell 26: 4053–66. Lucas, J.R., Nadeau, J. a, and Sack, F.D. (2006). Microtubule arrays and Arabidopsis stomatal development. J. Exp. Bot. 57: 71–9. Lucas, M. et al. (2013). Lateral root morphogenesis is dependent on the mechanical properties of the overlaying tissues. Proc. Natl. Acad. Sci. U. S. A. 110: 5229–34. Makkena, S., Lee, E., Sack, F.D., and Lamb, R.S. (2012). The R2R3 MYB Transcription Factors FOUR LIPS and MYB88 Regulate Female Reproductive Development. J. Exp. Bot. 63: 5545–5558. Malamy, J.E. and Benfey, P.N. (1997). Organization and cell differentiation in lateral roots of Arabidopsis thaliana. 44: 33–44. Malamy, J.E. and Ryan, K.S. (2001). Environmental regulation of lateral root initiation in Arabidopsis. Plant Physiol. 127: 899–909. Matos, J.L., Lau, O.S., Hachez, C., Cruz-Ramírez, A., Scheres, B., and Bergmann, D.C. (2014). Irreversible fate commitment in the Arabidopsis stomatal lineage requires a FAMA and RETINOBLASTOMA-RELATED module. Elife 3: e03271. McDonald, A.R., Liu, B., Joshi, H., and Palevitz, B. (1993). ?-Tubulin is associated with a cortical-microtubule-organizing zone in the developing guard cells of Allium cepa L. Planta 191. Meng, X., Chen, X., Mang, H., Liu, C., Yu, X., Gao, X., Torii, K.U., He, P., and Shan, L. (2015). Differential Function of Arabidopsis SERK Family Receptor-like Kinases in Stomatal Patterning. Curr. Biol. 25: 2361–72. Motose, H., Hamada, T., Yoshimoto, K., Murata, T., Hasebe, M., Watanabe, Y., Hashimoto, T., Sakai, T., and Takahashi, T. (2011). NIMA-related kinases 6, 4, and 5 interact with each other to regulate microtubule organization during epidermal cell expansion in Arabidopsis thaliana. Plant J. 67: 993–1005. Motose, H., Tominaga, R., Wada, T., Sugiyama, M., and Watanabe, Y. (2008). A NIMA-related protein kinase suppresses ectopic outgrowth of epidermal cells through its kinase activity and the association with microtubules. Plant J. 54: 829–844. Nadeau, J.A. (2009). Stomatal development: new signals and fate determinants. Curr. Opin. Plant Biol. 12: 29–35. Nadeau, J.A. and Sack, F.D. (2002). Control of stomatal distribution on the Arabidopsis leaf surface. Science 296: 1697–700. Nakamura, M., Ehrhardt, D.W., and Hashimoto, T. (2010). Microtubule and katanin-142  dependent dynamics of microtubule nucleation complexes in the acentrosomal Arabidopsis cortical array. Nat. Cell Biol. 12: 1064–70. Nakamura, M. and Hashimoto, T. (2009). A mutation in the Arabidopsis gamma-tubulin-containing complex causes helical growth and abnormal microtubule branching. J. Cell Sci. 122: 2208–17. Nakamura, M., Yagi, N., Kato, T., Fujita, S., Kawashima, N., Ehrhardt, D.W., and Hashimoto, T. (2012). Arabidopsis GCP3-interacting protein 1/MOZART 1 is an integral component of the γ-tubulin-containing microtubule nucleating complex. Plant J. 71: 216–25. Negi, J., Moriwaki, K., Konishi, M., Yokoyama, R., Nakano, T., Kusumi, K., Hashimoto-Sugimoto, M., Schroeder, J.I., Nishitani, K., Yanagisawa, S., and Iba, K. (2013). A Dof transcription factor, SCAP1, is essential for the development of functional stomata in Arabidopsis. Curr. Biol. 23: 479–84. Niederhuth, C.E., Patharkar, O.R., and Walker, J.C. (2013). Transcriptional profiling of the Arabidopsis abscission mutant hae hsl2 by RNA-Seq. BMC Genomics 14: 37. Ohashi-Ito, K. and Bergmann, D.C. (2006). Arabidopsis FAMA controls the final proliferation/differentiation switch during stomatal development. Plant Cell 18: 2493–505. Okushima, Y., Fukaki, H., Onoda, M., Theologis, A., and Tasaka, M. (2007). ARF7 and ARF19 regulate lateral root formation via direct activation of LBD/ASL genes in Arabidopsis. Plant Cell 19: 118–30. Overvoorde, P., Fukaki, H., and Beeckman, T. (2010). Auxin control of root development. Cold Spring Harb. Perspect. Biol. 2: a001537. Péret, B. et al. (2012). AUX/LAX Genes Encode a Family of Auxin Influx Transporters That Perform Distinct Functions during Arabidopsis Development. Plant Cell 24: 2874–2885. Péret, B., De Rybel, B., Casimiro, I., Benková, E., Swarup, R., Laplaze, L., Beeckman, T., and Bennett, M.J. (2009). Arabidopsis lateral root development: an emerging story. Trends Plant Sci. 14: 399–408. Peterson, K.M., Rychel, A.L., and Torii, K.U. (2010). Out of the mouths of plants: the molecular basis of the evolution and diversity of stomatal development. Plant Cell 22: 296–306. Peterson, K.M. and Torii, K.U. (2012). Long-term, high-resolution confocal time lapse imaging of Arabidopsis cotyledon epidermis during germination. J. Vis. Exp. Pfaffl, M.W. (2001). A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29: e45. 143  Pillitteri, L.J., Peterson, K.M., Horst, R.J., and Torii, K.U. (2011). Molecular profiling of stomatal meristemoids reveals new component of asymmetric cell division and commonalities among stem cell populations in Arabidopsis. Plant Cell 23: 3260–75. Pillitteri, L.J., Sloan, D.B., Bogenschutz, N.L., and Torii, K.U. (2007). Termination of asymmetric cell division and differentiation of stomata. Nature 445: 501–5. Richter, R., Behringer, C., Zourelidou, M., and Schwechheimer, C. (2013). Convergence of auxin and gibberellin signaling on the regulation of the GATA transcription factors GNC and GNL in Arabidopsis thaliana. Proc. Natl. Acad. Sci. U. S. A. 110: 13192–7. De Rybel, B. et al. (2010a). A novel Aux/IAA28 signaling cascade activates GATA23-dependent specification of lateral root founder cell identity. Curr. Biol. 20: 1697–1706. De Rybel, B. et al. (2010b). A novel aux/IAA28 signaling cascade activates GATA23-dependent specification of lateral root founder cell identity. Curr. Biol. 20: 1697–706. Schatten, H. ed (2012). The Centrosome (Humana Press: Totowa, NJ). Sdelci, S., Schütz, M., Pinyol, R., Bertran, M.T., Regué, L., Caelles, C., Vernos, I., and Roig, J. (2012). Nek9 phosphorylation of NEDD1/GCP-WD contributes to Plk1 control of γ-tubulin recruitment to the mitotic centrosome. Curr. Biol. 22: 1516–23. Serna, L. (2009). Emerging parallels between stomatal and muscle cell lineages. Plant Physiol. 149: 1625–31. Shirakawa, M., Ueda, H., Nagano, A.J., Shimada, T., Kohchi, T., and Hara-Nishimura, I. (2014). FAMA is an essential component for the differentiation of two distinct cell types, myrosin cells and guard cells, in Arabidopsis. Plant Cell 26: 4039–52. Shiu, S.H. and Bleecker, A.B. (2003). Expansion of the receptor-like kinase/Pelle gene family and receptor-like proteins in Arabidopsis. Plant Physiol. 132: 530–43. Shiu, S.H. and Bleecker, A.B. (2001). Receptor-like kinases from Arabidopsis form a monophyletic gene family related to animal receptor kinases. Proc. Natl. Acad. Sci. U. S. A. 98: 10763–8. Shpak, E.D., McAbee, J.M., Pillitteri, L.J., and Torii, K.U. (2005). Stomatal patterning and differentiation by synergistic interactions of receptor kinases. Science 309: 290–3. De Smet, I. et al. (2007). Auxin-dependent regulation of lateral root positioning in the basal meristem of Arabidopsis. Development 134: 681–90. Smet, I. De (2012). Minireview Lateral root initiation : one step at a time. 6: 867–873. De Smet, I. et al. (2008). Receptor-like kinase ACR4 restricts formative cell divisions in the Arabidopsis root. Science 322: 594–7. 144  Stone, J.M. and Walker, J.C. (1995). Plant protein kinase families and signal transduction. Plant Physiol. 108: 451–7. Sugano, S.S., Shimada, T., Imai, Y., Okawa, K., Tamai, A., Mori, M., and Hara-Nishimura, I. (2010). Stomagen positively regulates stomatal density in Arabidopsis. Nature 463: 241–4. Takatani, S., Otani, K., Kanazawa, M., Takahashi, T., and Motose, H. (2015). Structure, function, and evolution of plant NIMA-related kinases: implication for phosphorylation-dependent microtubule regulation. J. Plant Res. 128: 875–891. Tian, H., Jia, Y., Niu, T., Yu, Q., and Ding, Z. (2014). The key players of the primary root growth and development also function in lateral roots in Arabidopsis. Plant Cell Rep. 33: 745–53. Torii, K.U. (2008). Intracellular Signaling in Plants Z. Yang, ed (Wiley-Blackwell: Oxford, UK). Trotochaud, A.E., Jeong, S., and Clark, S.E. (2000). CLAVATA3, a multimeric ligand for the CLAVATA1 receptor-kinase. Science 289: 613–7. Vanneste, S. et al. (2011). Developmental regulation of CYCA2s contributes to tissue-specific proliferation in Arabidopsis. EMBO J. 30: 3430–41. Verstraeten, I., Schotte, S., and Geelen, D. (2014). Hypocotyl adventitious root organogenesis differs from lateral root development. Front. Plant Sci. 5: 495. Vilches-Barro, A. and Maizel, A. (2015). Talking through walls: mechanisms of lateral root emergence in Arabidopsis thaliana. Curr. Opin. Plant Biol. 23: 31–8. Walia, A., Nakamura, M., Moss, D., Kirik, V., Hashimoto, T., and Ehrhardt, D.W. (2014). GCP-WD mediates γ-TuRC recruitment and the geometry of microtubule nucleation in interphase arrays of Arabidopsis. Curr. Biol. 24: 2548–55. Wang, H.-Z. et al. (2015). Transcriptional regulation of PIN genes by FOUR LIPS and MYB88 during Arabidopsis root gravitropism. Nat. Commun. 6: 8822. Wang, Z.Y., Seto, H., Fujioka, S., Yoshida, S., and Chory, J. (2001). BRI1 is a critical component of a plasma-membrane receptor for plant steroids. Nature 410: 380–3. Wasteneys, G.O. and Ambrose, J.C. (2009). Spatial organization of plant cortical microtubules: close encounters of the 2D kind. Trends Cell Biol. 19: 62–71. Wierzba, M.P. and Tax, F.E. (2013). Notes from the Underground: Receptor-like kinases in arabidopsis root development. J. Integr. Plant Biol. 55: 1224–1237. Willmer, C. and Fricker, M. (1996). Stomata (Springer Science & Business Media). Xie, Z., Lee, E., Lucas, J.R., Morohashi, K., Li, D., Murray, J.A.H., Sack, F.D., and Grotewold, 145  E. (2010). Regulation of Cell Proliferation in the Stomatal Lineage by the Arabidopsis MYB FOUR LIPS via Direct Targeting of Core Cell Cycle Genes. PLANT CELL ONLINE 22: 2306–2321. Xu, T., Wen, M., Nagawa, S., Fu, Y., Chen, J.-G., Wu, M.-J., Perrot-Rechenmann, C., Friml, J., Jones, A.M., and Yang, Z. (2010). Cell surface- and rho GTPase-based auxin signaling controls cellular interdigitation in Arabidopsis. Cell 143: 99–110. Yokota, T. (1997). The structure, biosynthesis and function of brassinosteroids. Trends Plant Sci. 2: 137–143. Zeng, C.J.T., Lee, Y.-R.J., and Liu, B. (2009). The WD40 repeat protein NEDD1 functions in microtubule organization during cell division in Arabidopsis thaliana. Plant Cell 21: 1129–40. Zhang, X., Chen, Q., Feng, J., Hou, J., Yang, F., Liu, J., Jiang, Q., and Zhang, C. (2009). Sequential phosphorylation of Nedd1 by Cdk1 and Plk1 is required for targeting of the gammaTuRC to the centrosome. J. Cell Sci. 122: 2240–2251. Zhao, L. and Sack, F.D. (1999). Ultrastructure of stomatal development in Arabidopsis (Brassicaceae) leaves. Am. J. Bot. 86: 929–939. Zhou, A., Wang, H., Walker, J.C., and Li, J. (2004). BRL1, a leucine-rich repeat receptor-like protein kinase, is functionally redundant with BRI1 in regulating Arabidopsis brassinosteroid signaling. Plant J. 40: 399–409.  

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:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0303134/manifest

Comment

Related Items