Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cuticular wax analyses with high spatial and temporal resolution lead to the identification and characterization… Hegebarth, Daniela 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_hegebarth_daniela.pdf [ 31.4MB ]
Metadata
JSON: 24-1.0308092.json
JSON-LD: 24-1.0308092-ld.json
RDF/XML (Pretty): 24-1.0308092-rdf.xml
RDF/JSON: 24-1.0308092-rdf.json
Turtle: 24-1.0308092-turtle.txt
N-Triples: 24-1.0308092-rdf-ntriples.txt
Original Record: 24-1.0308092-source.json
Full Text
24-1.0308092-fulltext.txt
Citation
24-1.0308092.ris

Full Text

CUTICULAR WAX ANALYSES WITH HIGH SPATIAL AND TEMPORAL RESOLUTION LEAD TO THE IDENTIFICATION AND CHARACTERIZATION OF NOVEL WAX BIOSYNTHESIS GENES IN ARABIDOPSIS THALIANA  by  Daniela Hegebarth   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)  August 2016  © Daniela Hegebarth, 2016 ii Abstract Plant cuticles seal above-ground organs against non-stomatal water loss, and therefore are vital for survival on land. Besides providing a transpiration barrier, cuticles have important secondary functions, for example to protect from harmful UV radiation, and to provide a self-cleaning mechanism and mechanical support. Cuticles consist of aliphatic very-long-chain wax compounds (C24 to C38) and a cutin polymer. The diversity of cuticular wax compositions across the plant kingdom but also between different organs and ontogenetic stages is remarkable, yet the regulating mechanisms and function of those chemical differences are largely unknown. In the study presented here, a new approach was used, increasing temporal and spatial resolution and integration of chemical wax analyses with analyses of gene expression patterns of wax biosynthesis genes, using Arabidopsis thaliana as a model organism. The aims of the study were to, first, monitor wax composition and gene expression as a function of leaf development as well as different epidermal cell types and, second, to use this information to identify new wax biosynthesis genes.  In the second chapter, high temporal resolution was used to follow the dynamics of wax chemistry and gene expression during development of Arabidopsis thaliana leaves, and I was able to link changes in wax chemistry to differential expression of the elongation enzyme KCS6/CER6.  In the third chapter, wax analyses and gene expression data with high spatial resolution were acquired, and I identified differences between Arabidopsis epidermal cell types in iii wax composition and gene expression. Trichomes had a higher abundance of longer chain waxes (C32 to C38) compared to pavement cells, and the KCS5, KCS8 and KCS16 elongation enzymes were identified as candidates for the elongation of C34+ waxes.   In the fourth chapter, I characterized the Arabidopsis condensing enzyme KCS16 and was able to show that it is functioning on the wax elongation pathway, elongating C34 to C38 acyl-CoA wax precursors, mainly in trichomes but also in pavement cells. iv Preface Chapter 2 is based on work conducted in collaboration with Lucas Busta, Yan Cao, and Edward Kroc. Yan Cao, Lucas Busta, Reinhard Jetter and I designed the experiments. Yan Cao performed the morphological analysis, Lucas Busta and Yan Cao performed chemical analysis of leaf waxes and I performed quantitative RT-PCR analysis. Lucas Busta, Reinhard Jetter and I analyzed the data, Edward Kroc designed and performed statistical tests, and Lucas Busta, Reinhard Jetter and I prepared the figures and wrote the manuscript.  Chapter 3 is based on work conducted in collaboration with Christopher Buschhaus and David Bird. Christopher Buschhaus, Reinhard Jetter and I designed the experiments. Christopher Buschhaus analyzed the waxes from Arabidopsis stems and leaves, David Bird and I performed SEM, and I analyzed the waxes from isolated trichomes, performed statistical analysis on the wax chemistry and screened published microarray data. Christopher Buschhaus, Reinhard Jetter and I wrote the manuscript.  Chapter 4 is based on collaboration with Jérôme Joubès and David Bird. Reinhard Jetter and I designed the experiments. Jérôme Joubès performed the heterologous yeast expression experiment, and I analyzed the chemical composition of the yeast samples. David Bird and I performed SEM. I performed the rest of the experiments and statistical analysis, made the figures and wrote the manuscript. v Table of Contents  Abstract .............................................................................................................................. ii	  Preface ............................................................................................................................... iv	  Table of Contents .............................................................................................................. v	  List of Tables ..................................................................................................................... xi	  List of Figures .................................................................................................................. xii	  List of Abbreviations ....................................................................................................... xv	  Acknowledgements ....................................................................................................... xviii	  Dedication ....................................................................................................................... xix	   Introduction .................................................................................................... 1	  Chapter 1:1.1	   Cuticle structure ..................................................................................................... 1	  1.2	   Functions of the cuticle .......................................................................................... 3	  1.3	   Biosynthesis of cutin .............................................................................................. 5	  1.4	   Biosynthesis and export of wax compounds .......................................................... 6	  1.5	   Diversity of cuticular waxes ................................................................................. 11	  1.6	   Dynamics of cuticular wax during ontogenesis ................................................... 13	  1.7	   Cuticular composition on different leaf epidermal cells ...................................... 15	  1.8	   Research goals ...................................................................................................... 17	   Changes in cuticular wax coverage and composition on developing Chapter 2:Arabidopsis thaliana leaves ............................................................................................. 20	  2.1	   Introduction .......................................................................................................... 20	  2.2	   Materials and methods .......................................................................................... 23	  vi 2.2.1	   Plant material and growth conditions ........................................................ 23	  2.2.2	   Leaf morphological analysis ...................................................................... 24	  2.2.3	   Wax sample preparation and permutation tests ......................................... 24	  2.2.4	   Wax extraction and derivatization ............................................................. 25	  2.2.5	   Wax identification and quantification ....................................................... 25	  2.2.6	   RNA extraction and gene expression analysis by quantitative RT-PCR ... 26	  2.2.7	   Adjustment for multiple comparisons ....................................................... 28	  2.3	   Results .................................................................................................................. 28	  2.3.1	   Morphological changes on developing leaves ........................................... 28	  2.3.2	   Cuticular waxes from whole Arabidopsis leaves of different ages ........... 31	  2.3.2.1	   Wax development on growing eighth rosette leaves of the trichome-free Arabidopsis mutant gl1 ......................................................................................... 31	  2.3.2.2	   Wax development on growing eighth rosette leaves of wild-type Arabidopsis  .......................................................................................................... 34	  2.3.3	   Regional distribution of wax on Arabidopsis leaves ................................. 36	  2.3.4	   Expression of wax biosynthesis genes during leaf development ............... 39	  2.4	   Discussion ............................................................................................................ 40	  2.4.1	   Pavement cell age effects on wax composition ......................................... 41	  2.4.2	   Leaf expansion effects on wax coverage ................................................... 46	  2.4.3	   Epidermal cell type effects on wax composition ....................................... 49	   The composition of surface wax of different Arabidopsis thaliana Chapter 3:epidermis cells .................................................................................................................. 53	  3.1	   Introduction .......................................................................................................... 53	  vii 3.2	   Materials and methods .......................................................................................... 56	  3.2.1	   Plant material and growth conditions ........................................................ 56	  3.2.2	   Microscopy ................................................................................................ 57	  3.2.3	   Trichome isolation ..................................................................................... 57	  3.2.4	   Wax extraction and derivatization ............................................................. 58	  3.2.5	   Wax identification and quantification ....................................................... 58	  3.2.6	   Calculation of relative expression levels of wax biosynthesis genes ........ 58	  3.3	   Results .................................................................................................................. 59	  3.3.1	   Stem waxes ................................................................................................ 60	  3.3.2	   Leaf waxes ................................................................................................. 63	  3.3.3	   Trichome waxes ......................................................................................... 66	  3.3.4	   Gene expression patterns ........................................................................... 69	  3.4	   Discussion ............................................................................................................ 72	  3.4.1	   Compound class differences between trichomes and pavement cells ....... 73	  3.4.2	   Cell-specific expression patterns of genes involved in fatty acid modification ............................................................................................................... 75	  3.4.3	   Chain length profiles of trichome and pavement cell waxes ..................... 77	  3.4.4	   Relative expression profiles of genes involved in fatty acid elongation ... 80	   Characterization of β-ketoacyl-CoA synthase 16 (KCS16), a condensing Chapter 4:enzyme involved in biosynthesis of C36 and C38 acyl-CoAs for cuticular wax formation in Arabidopsis thaliana leaf trichomes and pavement cells ........................ 85	  4.1	   Introduction .......................................................................................................... 85	  4.2	   Materials and Methods ......................................................................................... 91	  viii 4.2.1	   Plant materials and growth conditions ....................................................... 91	  4.2.2	   Wax extraction and derivatization ............................................................. 92	  4.2.3	   Wax identification and quantification ....................................................... 92	  4.2.4	   Trichome isolation from Arabidopsis wild type and ksc16 leaves ............ 92	  4.2.5	   Wax extraction from trichomes and derivatization ................................... 92	  4.2.6	   Genomic DNA extraction for PCR genotyping ......................................... 93	  4.2.7	   RNA extraction and cDNA preparation .................................................... 93	  4.2.8	   Semi-quantitative RT-PCR of Arabidopsis mutant and transgenic lines .. 94	  4.2.9	   Quantitative Real-Time-PCR of Arabidopsis wild-type plants ................. 95	  4.2.10	   Cloning of the KCS16 coding region for constitutively expression and subcellular localization ............................................................................................... 96	  4.2.11	   Agrobacterium tumefaciens mediated transformation of Arabidopsis ksc16 and Nicotiana benthamiana ........................................................................................ 97	  4.2.12	   Stable transformation of Arabidopsis with p35s:KCS16-GFP .................. 98	  4.2.13	   Transient expression of KCS16 in tobacco ............................................... 98	  4.2.14	   Laser scanning microscopy ........................................................................ 99	  4.2.15	   Cryo-scanning electron microscopy ........................................................ 100	  4.2.16	   Heterologous expression in yeast ............................................................ 100	  4.2.17	   Statistical analysis .................................................................................... 101	  4.3	   Results ................................................................................................................ 101	  4.3.1	   Expression patterns of KCS16 ................................................................. 102	  4.3.2	   Isolation and characterization of kcs16 T-DNA insertion lines .............. 104	  4.3.3	   Wax composition of isolated trichomes .................................................. 109	  ix 4.3.4	   Subcellular localization of KCS16 .......................................................... 113	  4.3.5	   Biochemical function of KCS16 .............................................................. 115	  4.4	   Discussion .......................................................................................................... 119	  4.4.1	   KCS16 is involved in wax formation through synthesis of C36 and C38 acyl-CoAs ................................................................................................................. 120	  4.4.2	   KCS16 elongates C34 acyl-CoA, but not C30 acyl-CoA .......................... 121	  4.4.3	   KCS16 extends the ER-resident wax biosynthesis pathways .................. 124	  4.4.4	   Cell type-specific function of KCS16 ...................................................... 125	  4.4.5	   Functions of KCS16 in other organs ....................................................... 128	   Major findings and conclusions ................................................................ 132	  Chapter 5:5.1	   Major findings .................................................................................................... 132	  5.1.1	   Changes in cuticular wax coverage and composition on developing Arabidopsis thaliana leaves ...................................................................................... 132	  5.1.2	   The composition of surface wax of different Arabidopsis thaliana epidermis cells .......................................................................................................... 135	  5.1.3	   KCS16 gene characterization .................................................................. 138	  5.1.4	   Function of C36 and C38 extra-long wax compounds in Arabidopsis ...... 141	  5.1.5	   Alkenes as a novel compound class in Arabidopsis ................................ 143	  5.2	   Conclusion .......................................................................................................... 145	  Bibliography .................................................................................................................. 148	  Appendices ..................................................................................................................... 167	  Appendix A Supporting information for chapter 2 ..................................................... 167	  Appendix B Supporting information for statistical analysis ....................................... 171	  x B.1	   Permutations tests .................................................................................... 171	  B.2	   Statistical analysis of the gene expression data ....................................... 171	  B.3	   R code for permutation testing ................................................................ 172	  Appendix C Supporting information for chapter 4 ..................................................... 174	  Appendix D KCS8 cuticular wax analysis .................................................................. 179	   xi List of Tables Table 2.1: Cuticular wax coverage and production rates of different Arabidopsis  organs. ............................................................................................................................. 48 Table 2.2: Cuticular wax coverage and production rates of different Arabidopsis  epidermal cell types. ........................................................................................................ 49 Table A.1: Morphological data for developing wild-type Arabidopsis eighth leaves. ... 167	  Table A.2: List of primer sequences used in time course qRT-PCR analysis. ................ 170	  Table C.1: PCR target genes and primer sequences. ....................................................... 174	  Table C.2: Expression vectors and selection medium used in the heterologous yeast expression experiment. .................................................................................................... 178	   xii List of Figures  Figure 1.1: Example of structural model of plant cuticles with wax crystals. .................... 2	  Figure 1.2: Aliphatic compounds found in cuticular waxes. ............................................... 7	  Figure 1.3: Proposed wax biosynthesis pathway in Arabidopsis.. .................................... 11	  Figure 2.1: Surface area and epidermal cell numbers on developing wild-type Arabidopsis eighth rosette leaves. ..................................................................................... 30	  Figure 2.2: Wax coverage on developing Arabidopsis gl1 and wild-type eighth rosette leaves. ................................................................................................................................ 32	  Figure 2.3: Wax composition on developing Arabidopsis gl1 mutant eighth rosette leaves. ........................................................................................................................................... 34	  Figure 2.4: Wax composition on developing Arabidopsis wild-type eighth rosette leaves. ........................................................................................................................................... 36	  Figure 2.5: Pavement cell size on segments of developing Arabidopsis wild-type eighth rosette leaves. .................................................................................................................... 37	  Figure 2.6: Wax composition on the base and tip sections of wild-type Arabidopsis eighth rosette leaves at 13 days of age. ........................................................................................ 38	  Figure 2.7: Expression of wax biosynthesis genes in developing wild-type Arabidopsis eighth leaves. ..................................................................................................................... 40	  Figure 3.1: Differences in trichome abundance between Arabidopsis gl1, wild type and cpc tcl1 etc1 etc3. .............................................................................................................. 55	  Figure 3.2: Comparison of trichome abundance on Arabidopsis gl1, wild type and cpc tcl1 etc1 etc3 stems and leaves using cryo-SEM. ............................................................. 60	  Figure 3.3: Stem wax composition of Arabidopsis mutants and wild type. ...................... 63	  xiii Figure 3.4: Leaf wax composition of Arabidopsis mutants and wild type. ....................... 66	  Figure 3.5: Leaf trichome wax composition of Arabidopsis trichome-rich mutant and wild type. ................................................................................................................................... 68	  Figure 3.6: Differential expression of wax biosynthesis genes in Arabidopsis epidermis cell types. ........................................................................................................................... 72	  Figure 3.7: Comparisons of compound amounts between Arabidopsis wax sample types. ........................................................................................................................................... 79	  Figure 4.1: Elongation of acyl-CoAs by the FAE complex. ............................................. 86	  Figure 4.2: Developmental, cell type and tissues specific expression of KCS16 in Arabidopsis, based on microarray data. .......................................................................... 103	  Figure 4.3: KCS16 gene expression relative to UBC21 in various Arabidopsis organs.  ......................................................................................................................................... 104	  Figure 4.4: Characterization of Arabidopsis kcs16 T-DNA insertion and ectopic lines.  ......................................................................................................................................... 105	  Figure 4.5: Wax composition of mature and young leaves of Arabidopsis kcs16 insertion and complementation lines expressing KCS16 ectopically in both ksc16 backgrounds.  ......................................................................................................................................... 108	  Figure 4.6: Wax composition of trichomes isolated from leaves and stems of Arabidopsis kcs16 insertion lines and wild type. ................................................................................. 110	  Figure 4.7: Cryo-SEM micrographs of Arabidopsis wild type and kcs16-1 leaf trichomes. ......................................................................................................................................... 111	  Figure 4.8: Cryo-SEM micrographs of Arabidopsis wild type and kcs16-1 stem trichomes and wax crystals. ............................................................................................................. 112	  xiv Figure 4.9: Subcellular localization of Arabidopsis KCS16 in tobacco pavement cells.  ......................................................................................................................................... 114	  Figure 4.10: Subcellular localization of KCS16 in Arabidopsis roots and wax analysis of complementation lines. .................................................................................................... 114	  Figure 4.11: Wax composition of young Arabidopsis leaves overexpressing KCS16 in the wild type background. ..................................................................................................... 116	  Figure 4.12: GC-MS analysis of fatty acid methyl ester (FAMEs) profile after extraction and derivatization from yeast. ......................................................................................... 118	  Figure 4.13: Heterologous expression of Arabidopsis KCS enzymes in yeast. .............. 119	  Figure A.1: Sampling scheme for studying cell size distributions across sections of Arabidopsis wild type leaves of different ages. .............................................................. 168	  Figure A.2: Calculated trichome wax composition. ........................................................ 169	  Figure C.1: Wax composition of Arabidopsis kcs16 stems and siliques. ........................ 175	  Figure C.2: Chain length composition of mature and young leaves of Arabidopsis kcs16 mutants and complementation lines expressing KCS16 ectopically in both kcs16 backgrounds. .................................................................................................................... 176	  Figure C.3: Wax composition of Arabidopsis lines expressing KCS16 ectopically in the wild type background ...................................................................................................... 177	  Figure D.1: Chain length composition of Arabidopsis kcs8 mutants. ............................. 179	   xv List of Abbreviations ABCG: ATP-binding cassette transporter subfamily G ABRC: Arabidopsis Biological Resource Center Asn: Asparagine ATP: Adenosine triphosphate BDG: Bodyguard BSTFA: N,O-Bis(trimethylsilyl)trifluoroacetamide CER: Eceriferum CD1: Cutin deficient1 cDNA: copy DNA CoA: Coenzyme A Cys: Cysteine dH2O: distilled water DNA: Deoxyribonucleic acid ECR: Enoyl-CoA reductase EFP: Epidermal patterning factor EGTA: Ethylene glycol tetraacetic acid ER: Endoplasmic Reticulum FAE: Fatty acid elongase FAMEs: Fatty acid methyl esters FAR: Fatty acyl-CoA reductase fdh: fiddlehead FDR: False Discovery Rate  xvi GAPDH: Glyceraldehyde 3-phosphat dehydrogenase GC-FID: Gas chromatography with Flame ionization detector GC-MS: Gas chromatography with Mass spectrometer GFP: green-fluorescent protein gl1: glabra1 gl3-sst-sim: glabra3-shapeshifter-siamese GUS: β-glucoronidase HCD: β-hydroxyacyl-CoA dehydratase HIC: High carbon dioxide His: Histidine KCS: β-ketoacyl-CoA synthase KCR: β-ketoacyl-CoA reductase LACS: Long chain acyl-CoA synthase LB: Lysogeny broth Leu: Leucine LTP: Lipid transfer protein LTPG: glycosylphosphatidylinositol LTP MAH1: Midchain alkane hydroxylase MS: Murashige and Skoog NaCl: Sodium chloride PAS2: Pasticcino2 PCR: Polymerase chain reaction PEC1: Permeable cuticle1 xvii qRT-PCR: quantitative reverse transriptase-PCR RFP: red-fluorescent protein RNA: Ribonucleic acid SEM: Scanning electron microscopy TGT: Target intensity value TMHMM: Transmembrane Helices; Hidden Markov Model Trp: Tryptophan UBC: Ubiquitin C UBQ: Polyubiqutin Ura: Uracil VLCFA: very-long-chain fatty acids WSD1: Wax synthase diacylglycerol acyltransferase WT: wild type YFP: yellow-fluorescent protein xviii Acknowledgements I would like to express my gratitude to my supervisor Reinhard Jetter for his support throughout my graduate studies. He helped me to succeed by giving advice on experimental planning, and data interpretations and by putting his time and effort to improve my writing skills. I would also like to acknowledge the great support of my supervisory committee Lacey Samuels and Jae-Hyeok Lee.  My appreciation goes also to my fellow lab members, graduate students and postdocs, Lucas Busta, Radu Racovita, Alvaro Luna, Yulin Sun, Christopher Buschhaus, Mariya Skvortsova, Steffi Fritsche, Mathias Schuetz and many more for giving me advice, sharing their expertise, and having critical and encouraging discussions about my research.  My gratitude goes to my collaborators Jérôme Joubès and David Bird for their interest and contributions to my research.  I would like to acknowledge the financial support provided by the National Sciences and Engineering Research Council, Canada, The University of British Columbia, and the Botany Department.  Special gratitude goes to my family and friends, who encouraged me to persevere with my doctoral degree and supported me throughout graduate school.   xix Dedication For my parents, who helped me achieving my goals by supporting me throughout my life’s journey.   1  Introduction Chapter 1:About 450 million years ago the first land plants evolved from aquatic species by adapting to a life in a dry terrestrial habitat. Deposition of aliphatic compounds at the plant surface formed the first cuticular structure and created a barrier to protect against desiccation, changes of temperature, and high radiation on land (Riederer, 2006; Waters, 2003). To understand how this critical barrier is formed and how the chemical composition contributes to its function, this study analyzes cuticular waxes with high spatial and temporal resolution at the chemical and genetic level, and uses this information on wax dynamics and cell-specific composition to identify and characterize new wax biosynthesis genes.  1.1 CUTICLE STRUCTURE The cuticle is a mixture of aliphatic compounds that, based on their solubility in organic solvents, can be categorized into two major hydrophobic components, the solvent-soluble lipids, which are termed waxes, and the insoluble matrix of covalently linked lipids, which are called cutin (Kolattukudy, 1980). Forming a protective cuticle is cost-intensive for plants and needs substantial carbon flux into the biosynthetic pathway. Suh et al. (2005) determined that the vast majority of aliphatic compounds produced in Arabidopsis thaliana stem epidermal cells are deposited at the stem surface as cuticular compounds, whereas only one-third is used for intracellular lipids, such as membrane and storage lipids. Within the cuticle, the abundance of waxes and cutin varies in different layers of the cuticle and, based on histo-chemical staining, it is possible to distinguish between the cuticle proper and the cuticular layer. The cuticle proper is located on top of the cutin-rich layer, and is mainly composed of waxes and cutin. Its thickness is usually less than 200 nm, and yet 99 % of the water barrier mounted by the cuticles is maintained  2 by this thin cuticle proper (Riederer & Schreiber, 1995). The cuticular layer is located at the inner part of the cuticle and consists of cutin, wax and polysaccharides (Figure 1.1). In it, polysaccharide fibers protruding from the underlying epidermal cell wall and cutin as well as wax are mixed, forming a transition zone between the cuticle and the underlying epidermal tissue (Yeats & Rose, 2013).  Figure 1.1: Example of structural model of plant cuticles with wax crystals.   Waxes can be found throughout the cuticle, with distinct layers of intracuticular wax located in the cutin matrix and epicuticular wax deposited on the outer surface of the plant cuticle (Jeffree, 2006). Epicuticular waxes are in direct contact with the environment and form wax crystals in some species, such as Arabidopsis thaliana. Wax crystal formation causes glaucous appearance of leaves and stems (Koch & Ensikat, 2008), whereas surfaces without crystals appear glossy (Yeats & Rose, 2013).   3 1.2 FUNCTIONS OF THE CUTICLE The primary function of the cuticle is to protect the plant against desiccation, yet several secondary functions are also exerted by different parts of the cuticle. Cutin polymers are highly relevant to form a physical barrier, which prevents pathogens from entering the plant body easily. However, some pathogens such as fungi developed mechanisms to overcome the barrier by mechanical rupture or secretion of digestive enzymes (Deising et al., 2000), and higher cutin amount is not always correlated with higher pathogen resistance (Bessire et al., 2007; Isaacson et al., 2009).  As stated above, the primary function of the cuticle is to provide a barrier that limits transpirational water loss and enables stomatal control of gas exchange (Riederer, 2006). Accordingly, if the cuticle structure is impaired, the transpiration barrier is negatively affected (Bessire et al., 2007). Neither thickness of the cuticle nor the amount of wax or cutin are correlated with the ability to retain water (Isaacson et al., 2009; Jetter & Riederer, 2015; Riederer & Schreiber, 2001). However, the removal of cuticular waxes from tomato fruits increased the permeability (Leide et al., 2007), showing that cuticular waxes are providing the transpiration barrier. It was also shown that the composition of waxes influences transpiration properties, where for example a high abundance of non-polar compounds correlated with relatively low water permeability, while a high abundance of non-aliphatic compounds seems to be associated with higher water permeability (Buschhaus & Jetter, 2012; Leide et al., 2007). To get a better understanding of the relationship between wax composition and transpiration barrier function of cuticles, in a recent study differences in epi- and intracuticular wax compositions were compared with differences in transpiration properties in several species. In four species (Tetrastigma voinierianum, Oreopanax guatemalensis, Monstera deliciosa and Schefflera elegantissima) the  4 intracuticular wax was providing the main transpiration barrier, and no major differences in composition between epi- and intracuticular wax were observed (Jetter & Riederer, 2015). Other species (Citrus aurantium, Euonymus japonica, Clusia flava and Garcinia spicata) contained alicyclic compounds such as triterpenoids in the intracuticular wax mixture, which coincided with an equal contribution of both layers to the transpiration barrier (Jetter & Riederer, 2015), suggesting that alicyclic compounds are associated with increased cuticle water permeability. This is consistent with a model that suggests that linear aliphatic wax compounds form crystalline zones, which are impermeable to water, whereas non-aliphatic compounds form amorphous zones through which the water can move (Riederer & Schreiber, 1995). Therefore, it seems water barrier properties of cuticles are influenced by the mixture of different wax compounds as well as their arrangement (Yeats & Rose, 2013).  Phenolic constituents of cuticular waxes may serve as protection from harmful UV radiation by improving the screening properties of cuticles (Pfündel et al., 2006). Typical phenolic compounds found in the cuticle are hydroxycinnamic acids, such as ferulic acid or p-coumaric acid (Liakopoulos et al., 2001). Based on the fluorescence properties of phenolic compounds, they were visually detected on the cuticle, and particularly high in guard cells (Karabourniotis et al., 2001).   Some species also form wax crystals on the cuticle surface, which enable water droplets to run off from the cuticle surface, taking debris and dust particles with them (lotus effect). Analyses of the cuticular surfaces of different species showed a positive correlation between wax crystal abundance and the water repellency of the cuticle, suggesting that the crystals affect the surface  5 properties by decreasing the contact surface area between droplet and cuticle (Barthlott & Neinhuis, 1997). Wax-deficient Arabidopsis mutants, such as fiddlehead, showed organ fusions, indicating that cuticular waxes are important to maintain the boundaries of tissues and organs during development (Lolle et al., 1997; Yephremov et al., 1999). Similarly, wax-deficient mutants also revealed the relevance of waxes for pollen-pistil interactions, as decreased wax amounts caused male sterility by alterations of the pollen tryphine layer (Fiebig et al., 2000; Preuss et al., 1993; Pruitt et al., 2000).   1.3 BIOSYNTHESIS OF CUTIN Cutin is a polymer of fatty acid derivatives, such as saturated and unsaturated C16 and C18 ω-hydroxyacids, polyhydroxyacids or epoxyacids (Graça et al., 2002; Xiao et al., 2004). Those compounds are esterified with each other or glycerol (Yeats & Rose, 2013), forming an insoluble matrix. Interestingly, Arabidopsis cutin has an exceptional composition comprising mainly α,ω-dicarboxylic acids (Bonaventure et al., 2004; Franke et al., 2005). Beside the compound class composition, the ratio of C16 and C18 derivatives can also vary between species, organs and developmental stages (Espelie et al., 1980; Marga et al., 2001).  The mechanism of cutin polyester formation is not well understood, but it is thought to be based on esterification of primary hydroxyl groups of C16 and C18 fatty acid monomers, followed by (ester or non-ester) cross-linking through mid-chain groups and other constituents, resulting in a three-dimensional structure (Stark & Tian, 2007; Villena et al., 2000). In recent years, progress was made in understanding the process of incorporation of monomers into the cutin polyester using Arabidopsis and tomato as model organisms. In tomato fruit, the first cutin synthase gene,  6 Cutin deficient 1 (CD1), was identified as an extracellular acyltransferase, using 2-mono-(10,16-dihydroxyhexadecanoyl)-glycerol as substrate to form ester oligomers (Girard et al., 2012; Yeats et al., 2012). Based on sequence similarity, CD1 homologs and orthologs were identified in diverse plant taxa (Volokita et al., 2011). In 2006, Kurdyukov et al. showed the importance of another enzyme called BODYGUARD (BDG) for cutin formation in Arabidopsis. BDG was identified as a member of the α/β-hydrolase family, however its exact molecular function remains unknown (Kurdyukov et al., 2006). Besides the cutin polymer, some plant species have a second polymer structure called cutan (Gupta et al., 2016). It was proposed that ether linkages of C22 to C34 alkyl compounds form the polymer, and it is the ether bonds, which makes it non-hydrolysable and therefore difficult to analyze or extract (Deshmukh et al., 2005; Schouten et al., 1998). A decrease of hydrolysable alkyl cutin compounds in older Arabidopsis stem sections suggests conversion of cutin to cutan after tissue maturation. It can further be speculated that the cutan in Arabidopsis is mainly composed of C18 alkyl compounds (Suh et al., 2005). However, because cutan is non-extractable its formation remains one of the big unknowns in cuticle biochemistry.  1.4 BIOSYNTHESIS AND EXPORT OF WAX COMPOUNDS Cuticular wax is usually comprised of a variety of aliphatic compound classes such as fatty acids, primary n-alcohols, secondary alcohols, wax esters, aldehydes, alkanes, β-diketones and terpenoids (Figure 1.2).  7  Figure 1.2: Aliphatic compounds found in cuticular waxes (Yeats & Rose, 2013), n = C16 to C38.   Great progress was made in understanding wax biosynthesis using Arabidopsis and tomato as model organisms, showing that the synthesis of waxes takes place in several steps. First, C16 and C18 fatty acid-thioesters are synthesized de novo in the plastids and hydrolyzed to free acids. The fatty acids are activated to acyl-CoAs by long chain acyl-CoA synthases (LACSs) and transported from the plastid to the endoplasmic reticulum (ER) (Pulsifer et al., 2012). There, the fatty acids are not only used for cuticular wax formation, but also as precursors for cutin, plasma membrane lipids such as sphingo- and phospholipids, and in seeds for storage triacylglycerols. At the ER, acyl-CoAs are elongated in several elongation cycles from C16 and C18 to very-long-chain fatty acids (VLCFA), which can have aliphatic chains with 24 to 34 carbons (Joubès et al., 2008) (Figure 1.3). Each elongation cycle is carried out by a fatty acid elongase (FAE), an enzyme complex catalyzing four sequential reactions, resulting in an acyl-CoA product HOAlkanesAldehydesOHOsecondary AlcoholsKetonesOHOOHOOprimary n-AlcoholsWax EstersFatty acidsO ODiketonesnnnnn nn nn nn n 8 elongated by two carbons in each cycle. The first step is catalyzed by a β-ketoacyl-CoA synthase (KCS), condensing acyl-CoA with a malonyl-CoA to form a β-ketoacyl-CoA intermediate. In the ensuing reactions, the intermediate is reduced by a β-ketoacyl-CoA reductase (KCR) to a β-D-hydroxyacyl-CoA, then a trans-Δ2-enoyl-CoA is formed in a dehydration reaction catalyzed by a β-hydroxyacyl-CoA dehydratase (HCD) enzyme, and finally a reduction catalyzed by an enoyl-CoA reductase (ECR) forms an acyl-CoA elongated by two carbons (Beaudoin et al., 2002; Domergue et al., 2000; Han et al., 2002; Kunst & Samuels, 2009; Li-Beisson et al., 2013; Millar & Kunst, 1997; Paul et al., 2006). The initial condensing reaction, catalyzed by KCS enzymes, is the rate-limiting step and determines the chain length range of substrates and products of the FAE complex (Millar & Kunst, 1997), while the other FAE enzymes, the KCR, HCD and ECR, are used ubiquitously by FAE complexes (Millar & Kunst, 1997; Zheng et al., 2005). It is likely that different FAEs co-exist within one cell, enabling a broad chain length range of wax compounds. The Arabidopsis KCS gene family consists of 21 members and several KCS enzymes showed elongation activity when heterologously expressed in yeast (Saccharomyces cerevisiae) (Blacklock & Jaworski, 2006; Franke et al., 2009; Lee et al., 2009; Paul et al., 2006; Trenkamp et al., 2004; Tresch et al., 2012). Based on ksc mutant analysis several KCS enzymes, such as KCS6/CER6 were found to be vital for cuticular wax biosynthesis. kcs6/cer6 mutants showed a severe decrease in wax amounts, as well as organ fusions and a male sterile phenotype (Preuss et al., 1993; Hülskamp et al., 1995; Trenkamp et al., 2004). KCS6/CER6 was identified to elongate C24 to C28 acyl-CoAs, and when associated with CER2-like proteins, the KCS6/CER6 FAE complex was able to elongate acyl-CoAs up to C34 ( Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015), yet the mechanism of action of CER2-like enzymes remains unknown.  9  After elongation, acyl-CoAs are modified on two branch pathways, the alkane-forming pathway and the alcohol-forming pathway. On the decarbonylation pathway, the CER3/CER1 complex reduces acyl-CoAs to aldehyde intermediates (Bernard et al., 2012), followed by a decarbonylation reaction in which one carbon is lost, resulting in alkanes with odd carbon numbers (Cheesbrough & Kolattukudy, 1984; Schneider-Belhaddad & Kolattukudy, 2000) (Figure 1.3). Alkanes can be transported to the cuticle or used by MAH1 (Midchain Alkane Hydroxylase1) to form secondary alcohols and ketones (Greer et al., 2007) (Figure 1.3). On the second branch pathway, the acyl reduction pathway, fatty acyl-CoA reductases (FARs) form n-alcohols (Rowland et al., 2006) (Figure 1.3), which can either be transported to the cuticle or used by the WSD1 (Wax Synthase Diacylglycerol Acyltransferase1) enzyme for combination with fatty acids to form wax esters (Li et al., 2008) (Figure 1.3). While those two pathways are fairly well understood in Arabidopsis, it remains unknown if other wax compounds (for example fatty acids) are products of the same branch pathways or if other modification pathways exist.  All wax compounds have to be transported from the ER to the plasma membrane, yet the exact mechanism remains elusive. Two mechanisms are possible, which involve either membrane contact sites or vesicle transport across the Golgi network. In Arabidopsis, there is evidence supporting the latter, as two Golgi network mutants showed a negative effect on wax deposition and are therefore supporting a vesicular transport mechanism (McFarlane et al., 2014). For the transport across the plasma membrane, two ATP-binding cassette (ABC) half transporters were identified in Arabidopsis that are required for the transport of wax compounds to the cuticle, ABCG12 (Pighin et al., 2004) and ABCG11 (Bird et al., 2007). It was suggested that hetero- 10 dimers of ABCG12 and ABCG11 are needed for wax transport (McFarlane et al., 2010), while homo-dimers of ABCG11 or another transporter, ABCG13, could export cutin precursors (McFarlane et al., 2010; Panikashvili et al., 2011). Recently, a full ABC transporter, ABC32/PEC1, was identified that is relevant for wax and cutin synthesis during early developmental stages (Fabre et al., 2016). Several other ABC transporters were found to be necessary for the synthesis of other lipid barriers such as suberin and pollen wall lipids (Yadav et al., 2014). The transport of some wax compounds across the extracellular space and cell wall could be facilitated by lipid transfer proteins (LTPs). The Arabidopsis LTP gene family consists of 70 members (Beisson et al., 2003), and recently two glycosylphosphatidylinositol-anchored LTPs (LTPG1 and -2) were localized to the extracellular side of the plasma membrane (Debono et al., 2009). Respective ltpg1 and -2 mutants showed a decrease in cuticular wax amounts (Debono et al., 2009; Kim et al., 2012; Lee et al., 2009), suggesting that they are relevant for wax transport across the extracellular space and cell wall.  11  Figure 1.3: Proposed wax biosynthesis pathway in Arabidopsis. First, acyl-CoAs are elongated by Fatty acid elongase (FAE) complexes (black), followed by modifications in either the alkane-forming pathway (blue) or the alcohol-forming pathway (red) (Yeats & Rose, 2013).   1.5 DIVERSITY OF CUTICULAR WAXES The cuticle is a highly dynamic structure, and its diversity in compound class composition, but also chain length distribution between plant species and different organs and growth stages of the same plant is striking. For instance in leaf blades of Triticum aestivum seedlings, primary n-alcohols are the predominant compound class, whereas in flag leaf sheaths β-diketones are predominant (Wang et al., 2015). In Arabidopsis leaves and stems, alkanes are the most abundant compound class. Within the compound classes, usually a series of compounds are found that vary in the carbon number of the aliphatic chain (thus series of homologs), and for instance Arabidopsis leaves contain alkanes with a broad chain length range from C25 to C34. In contrast, C20 acyl CoAC22 acyl CoAC24 acyl CoAC26 acyl CoAC28 acyl CoAC30 acyl CoAC32 acyl CoAC34 acyl CoA C26 aldehyde C28 aldehyde C30 aldehyde C32 aldehyde C34 aldehyde C25 alkane C27 alkane C29 alkane C31 alkane C33 alkane C29 secondary alcohol C29 ketoneCER1 CER3CER1 CER3CER1 CER3CER1 CER3CER1 CER3MAH1MAH1FAEFAEFAEFAEFAEFAEFAECER4CER4CER4CER4CER4 C26 primary alcohol C28 primary alcohol C30 primary alcohol C32 primary alcohol C34 primary alcohol C42 Wax esterWSD1Cuticle Cuticle+ C16 fatty acidWSD1 + C16 fatty acidWSD1 + C16 fatty acidWSD1 + C16 fatty acidWSD1 + C16 fatty acid C44 Wax ester C46 Wax ester C48 Wax ester C50 Wax esterC16 acyl CoAC18 acyl CoAFAEFAECER4 C24 primary alcohol  C40 Wax esterWSD1 + C16 fatty acidCER4 C22 primary alcohol  C38 Wax esterWSD1 + C16 fatty acidDecarbonylation (alkane-forming) pathwayAcyl-reduction (alcohol-forming) pathwayElongation  pathway 12 Arabidopsis stems consist mainly of C29 alkanes, whereas other chain lengths are much less abundant.  The chain length distributions of various compound classes in most plant species range up to C34 or C36, including for example Cucumus sativus (Wang et al., 2015), Arabidopsis stem wax (Jenks et al., 1996), Lycopersicon esculentum (Leide et al., 2007), Capsicum annuum, and Solanum melongena (Bauer et al., 2004). However, in some species, longer wax compounds with up to 38 carbon chains were reported, for example, as trace amounts of C37 alkane (derived from C38 precursors) in leaves of Euphorbia species (Hemmers & Gülz, 1986), in Triticum aestivum, Zea mays, Lupinus angustifolius (Nadiminti et al., 2015) and Miscanthus sinensis (Gao & Huang, 2013), or in the fruit capsule of Papaver somniferum and Eschscholzia california (Jetter & Riederer, 1996), and in the foliage of Austrocedrus chilensis (Dodd et al., 1998). Interestingly, the only compound classes reported to date to have such extraordinary chain lengths beyond C34 are alkanes and alkenes.  The low abundance of C38 wax compounds makes their detection difficult, and therefore I cannot exclude the possibility that those extra-long wax compounds are present in many more plant species than reported to date. Depending on the sensitivity of the instruments, to date, about 10 % of total extracted wax from Arabidopsis remains unidentified, residing mostly in low -abundance compounds. A more detailed analysis could result in identification of new compound classes or chain lengths. Moreover, integration of chemical analyses with gene expression studies is needed to understand the underlying mechanisms, which regulate the complex chemical diversity of waxes.   13 1.6 DYNAMICS OF CUTICULAR WAX DURING ONTOGENESIS During organ development, epidermal cells (pavement cells, guard cells, trichome cells) go through a cell division and cell expansion stage, and studies in a variety of species showed differences in cuticular compositions during epidermal development. Changes in cuticular wax compositions of fruits are of particular interest, because their cuticle is suspected to be involved in fruit cracking, which it also affects storability and shelf life of fruit after harvest (Peschel et al., 2007; Domínguez et al., 2008). Interestingly, in non-climacteric fruit such as Prunus avium, cuticle deposition decreases during fruit development, likely because the surface expands extremely rapidly, and thus the risk of rain-cracks increases. The compound class composition also changes over the course of P. avium fruit development, and a decrease of triterpenes together with an increase of alkanes has been observed (Peschel et al., 2007). In climacteric fruits of other plant species, such as Solanum lycopersicum and Citrus sinensis, the wax coverage increases during development, and slight changes in compound class composition have been observed (Domínguez et al., 2008; Kosma et al., 2010; Wang et al., 2016).  The wax coverage and composition during the development of leaves have also been studied in several plant species. On Kalanchoe daigremontiana leaves, a steady increase in wax coverage and a shift in relative compound class abundance was observed (Van Maarseveen et al., 2009). A similar change was observed in leaves of Fagus sylvatica (Prasad & Gülz, 1990) and Hedera helix (Viougeas et al., 1995). Many species, such as Prunus laurocerasus (Jetter & Schäffer, 2001), Malus domestica (Bringe et al., 2006), Coffea arabica (Stocker & Ashton, 1975) and Triticum aestivum (Tulloch, 1973), have shown constant wax coverage but a shift in relative compound class distribution and a tendency to accumulate increasing amounts of wax  14 compounds with longer carbon chains. Other species such as Sesamum indicum have shown a decrease in total wax coverage during development (Kim et al., 2009) but a similar shift in compound class and chain length distribution as described above, with increasing amounts of alkanes and aldehydes, and longer chain lengths.   These studies highlight the diversity of cuticle compositions between species but also their dynamic nature during the development of different organs. Thus, the observed dynamics of cuticular wax suggest that coverages and compositions may not only be optimized for diverse functions between organs, but also during discrete developmental stages. However, the experimental design and scope of the different studies varied considerably, and hence the results are difficult to compare. For instance, some researchers focus mainly on the chemistry, making it difficult to reconstruct how the time points for the chemical analysis were chosen, while other studies integrate leaf morphology measurements with chemistry for better comparison, yet the morphology measurements taken are inconsistent between the studies.  Optimization of wax coverage and compositions at discrete developmental stages needs tight regulation of the wax biosynthesis machinery, and therefore some studies have also addressed the transcriptional dynamics of genes involved in cuticle formation. In different cherry cultivars, differential expression of wax biosynthesis genes was linked to differences in rain cracking traits on developing fruits (Balbontín et al., 2014). Transcriptional differences of wax biosynthesis genes were also found in the epidermis of growing and ripening tomato fruit (Yeats et al., 2010), indicating that the differences in wax chemistry during development are regulated by differential expression of wax biosynthesis genes. In Arabidopsis, changes in wax chemistry and changes at  15 the genetic level have been investigated on mature and expanding sections of bolting stems (Suh et al., 2005). Interestingly, some wax biosynthesis genes were found to be more highly expressed in expanding stems, yet no differences in the cuticular wax compositions between expanding and mature stems were observed. This finding may be explained by the fast expansion of stems, which might make it difficult to detect any chemical changes with the necessary spatial and/or temporal resolution. Overall, it has proved impossible to identify specific gene and enzyme classes regulating cuticle development in Arabidopsis stems.  Since Arabidopsis leaves are expanding much more slowly than stems, it might be easier to observe changes in cuticular wax compositions in the leaves. An early study on the dynamics of Arabidopsis leaf waxes found that whole rosettes from young plants have higher wax coverage, higher percentages of fatty acids, and longer alcohols than whole rosettes from older plants (Jenks et al., 1996). Since the study focused on differences of developing leaves between cuticular wax-deficient mutants and the wild type, morphological measurements were not taken, and the data are lacking the necessary temporal resolution and genetic information. A more detailed analysis of chemical changes over the course of leaf development as well as the integration of morphometric and genetic data is necessary to exploit developmental dynamics for gene discovery and to understand the genetic regulatory mechanisms underlying leaf wax development.   1.7 CUTICULAR COMPOSITION ON DIFFERENT LEAF EPIDERMAL CELLS To maximize its function, the cuticle must form a continuous layer of hydrophobic material across the entire organ surface. Therefore, it can be assumed that all epidermal cell types  16 (pavement cells, guard cells and trichomes) are covered with cuticular waxes and have independent biosynthesis machineries. Moreover, trichome development starts very early on developing leaves of many species, and at a very young age they thus already contain the full number of completely developed trichomes. Therefore, younger leaves have a higher abundance of trichomes relative to surface area compared to mature leaves, and trichomes add substantial surface area and wax coverage to younger leaves in particular. Because trichomes protrude from the leaf tissue, they likely experience greater levels of abrasion and wind-induced drying than the relatively flat pavement cells, and thus may require a unique wax composition. This notion is supported by several lines of experimental evidence suggesting distinct wax compositions on different epidermal cell types: First, ion flux across the cuticle differed between epidermal cell types, indirectly suggesting differences in cuticle composition between cell types (Schreiber, 2006). Second, some wax biosynthesis genes exhibited cell-specific expression patterns, where for example the mid-chain alkane hydroxylase MAH1 was expressed in stem pavement cells but not in trichomes or guard cells (Greer et al., 2007) and CER4 expression was observed in stem pavement cells but was restricted to trichomes for leaves (Rowland et al., 2006). Unfortunately, other studies characterizing wax biosynthesis proteins did not address their localization in specific epidermal cells, and therefore the presence or absence of most wax biosynthesis enzymes in trichomes remains unknown. However, microarray experiments comparing trichomes, pavement cells and guard cells show differential expression of some lipid-related genes, implying that the wax composition also varies between cell types (Marks et al., 2009). Most interestingly, such microarray data identified KCSs to be more highly expressed in trichomes compared to pavement cells (Jakoby et al., 2008), which make good candidates for elongation towards wax products.  17 One of the KCS expressed higher in trichomes was KCS16. Interestingly, despite the great progress in characterizing Arabidopsis KCS genes, to date no KCS has been identified that can elongate acyl-CoAs beyond C34, even if wax compounds derived from C36 and C38 acyl-CoAs were found in leaves.   1.8 RESEARCH GOALS The observed dynamics of cuticular wax in a variety of species suggest that coverage and compositions may not only be optimized for diverse functions between organs, but also during discrete developmental stages and between cell types. Therefore, the goal of my PhD work was to analyze cuticular waxes with high spatial and temporal resolution at the chemical as well as the genetic level, and to use the new information on wax dynamics and cell-specific composition to identify and characterize new wax biosynthesis genes. To this end, I analyzed changes in morphology and cuticular wax compositions at different time points during organ development and in different epidermal cell types, respectively. I chose the leaves of the model plant Arabidopsis thaliana for my studies. My specific goals were:  • to integrate morphometric and cuticular wax compositional changes during leaf development, • to identify differences in cuticular wax compositions between trichome and pavement cells, and • to determine the contribution of trichome wax to the total wax coverage.   18 After analysis of chemistry and morphology I followed up on changes in expression levels of wax biosynthesis genes during leaf development as well as in different epidermal cell types. My specific goals were:  • to identify genetic mechanisms regulating the chemical changes during leaf development, • to identify wax biosynthesis genes differentially expressed in trichome and pavement cells, and • to identify KCS genes that are expressed trichome-specifically.  My fourth chapter aimed to characterize new wax biosynthesis genes identified in the first two parts of my PhD work. I focused on the KCS16 enzyme as a candidate to be involved in the elongation pathway of cuticular waxes, hypothesizing that the condensing enzyme elongates acyl-CoAs beyond C34 in young leaves and trichomes. My specific goals were:  • to test if KCS16 is involved in the elongation pathway of cuticular waxes, • to identify the substrate and product specificity of KCS16, and • to determine if KCS16 is functioning trichome-specifically.  By addressing those goals, I was able to integrate morphometric, chemical and gene expression analyses. Specifically, in my second chapter I investigated chemical and genetic changes during Arabidopsis leaf development. To subtract the contribution of trichome cells to the wax coverage and composition, a trichome-free mutant glabra1 (gl1) was included in the analysis. The Arabidopsis GLABRA1 (GL1) gene encodes a MYB transcription factor, which is relevant for  19 trichome initiation during early trichome development (Hülskamp, 2004). In my third chapter, I investigated changes in wax chemistry and gene expression in different epidermal cell types: first, the cuticular leaf wax composition of a trichome-free Arabidopsis mutant (gl1) was compared with leaves from a trichome-rich mutant (cpc tcl1 etc1 etc3) and wild-type leaves. Second, the cuticular wax of trichomes isolated from wild-type and cpc tcl1 etc1 etc3 mutant leaves were analyzed, and third, published microarray data were screened to identify wax biosynthesis genes preferentially expressed in trichomes. In my fourth chapter, I characterized the Arabidopsis KCS16 enzyme, by analyzing the cuticular wax of young and mature leaves as well as isolated trichomes of kcs16 loss-of-function mutants. Next, I tested substrate and product specificity of KCS16 by heterologous expression in yeast as well as by constitutive expression of KCS16 in wild-type Arabidopsis plants. The subcellular localization as well as the organ and tissue-specific expression of KCS16 were analyzed.  20  Changes in cuticular wax coverage and composition on developing Chapter 2:Arabidopsis thaliana leaves  2.1 INTRODUCTION Plant organ development relies on the tightly controlled formation and expansion of various tissues based on limited resources of reduced carbon and nutrients. To protect precious new organs from adverse conditions, physical and chemical defenses are established early during development (Bennett & Wallsgrove, 2016; Fraenkel, 2016; Tian et al., 2012) and they must continuously expand to remain effective over the course of organ growth. Therefore, developing organs must continuously invest in both construction and protection of new structures (Bazzaz et al., 1987; Coley, 1985; Herms, 2016; Züst et al., 2011).  A balanced use of resources in leaf construction and protection is particularly important for epidermal cells because they form the leaf-environment interface. For example, the rapidly expanding epidermis of growing leaves must constantly protect the entire organ against physical damage, insect attack, and excessive water loss by transpiration. The plant epidermis comprises pavement cells, guard cells, and trichomes. Where present, all three epidermal cell types will thus contribute to the functions of the epidermis. Pavement cells, the most abundant epidermal cell type on all organ surfaces, are the major protective surface barrier (Ramsay & Glover, 2005). Guard cells, present on many organ surfaces in smaller numbers than pavement cells, are important for regulating gas exchange and for protecting the surface around stomata (Kearns & Assmann, 1993). Finally, trichomes have a variety of roles including UV protection, heat insulation, transpiration control, and insect deterrence (Wagner et al., 2004). To gain insight into  21 the various roles of epidermal cells, their development on leaves of various species including Arabidopsis has been studied in much detail (Glover, 2000; Guimil & Dunand, 2007).  The three epidermal cell types, together comprising the organ-environment interface, are coated by a continuous cuticle layer (Samuels et al., 2008; Yeats & Rose, 2013). Cuticular waxes, typically composed of fatty acids, primary n-alcohols, wax esters, aldehydes, and alkanes, constitutes the transpiration barrier (Haas & Schönherr, 1979; Isaacson et al., 2009; Schönherr, 1976). The amount of wax per surface area (coverage) and the composition of the wax mixture vary between tissues, organs, and species (Bringe et al., 2006; Gülza et al., 1991; Jetter & Schäffer, 2001; Kim et al., 2009; Van Maarseveen et al., 2009; Richardson et al., 2005; Salasoo, 1983; Tulloch, 1973; Viougeas et al., 1995), suggesting characteristic adaptations to optimize specific functions.  As described previously, the ontogenetic development of cuticular wax has been investigated in leaves and fruits of some species, and in most cases had been found to be dynamic with respect to time. For example, changes in wax coverage and composition during development had been reported for Kalanchoe daigremontiana (Van Maarseveen et al., 2009), Prunus laurocerasus (Jetter & Schäffer, 2001), Coffea arabica (Stocker & Ashton, 1975), Malus domestica (Bringe et al., 2006), Sesamum indicum (Kim et al., 2009), Hordeum vulgare (Richardson et al., 2005), Triticum aestivum (Tulloch, 1973), Hedera helix (Viougeas et al., 1995), and Fagus sylvatica (Prasad & Gülz, 1990). Thus, the observed dynamics of cuticular wax suggest that coverages and compositions may not only be optimized for diverse functions between organs, but also during discrete developmental stages.  22 To further our understanding of cuticular wax coverage, composition, and function on developing organs, integrated approaches combining organ morphometrics, wax chemical profiling, and gene expression analyses of a model species are required. Accordingly, cell expansion rates, wax composition, and expression levels of wax biosynthesis genes has been investigated in bolting stems of Arabidopsis (Suh et al., 2005), the species for which wax biosynthesis pathways are currently best characterized. Interestingly, expanding Arabidopsis stem sections were found to express many wax biosynthesis genes more highly than sections that had completed expansion, pointing to transcriptional regulation of wax biosynthesis (Suh et al., 2005). Nonetheless, neither wax coverage nor composition differed between the top, middle, and bottom sections of the stem, sampled as proxies for tissues of different age. However, Arabidopsis stems grow very rapidly, mainly by expanding in a very short zone near the top of the stem, rendering potential morphological, chemical, or genetic gradients within this zone only visible via analyses with very high spatial and temporal resolution.  In contrast to bolting inflorescence stems, Arabidopsis rosette leaves develop much more slowly and may therefore be good candidates for identifying developmental changes in wax coverage and/or composition that can be linked to cell expansion and gene expression. Arabidopsis leaf waxes consist mainly of alkanes and primary n-alcohols, together with fatty acids and aldehydes (Jenks et al., 1995). An early study on the dynamics of Arabidopsis leaf waxes found that whole rosettes from young plants have higher wax coverage, higher percentages of fatty acids, and longer alcohols than whole rosettes from older plants (Jenks et al., 1996). Here, I sought to first corroborate these results with higher temporal resolution using leaves from only one nodal position. Preliminary tests had shown that under our conditions leaves on nodes one to seven  23 grew to variable, relatively small sizes, while leaves on nodes eight and higher all grew to approximately the same full size. Therefore, the eighth rosette leaf was chosen for monitoring epidermal cell development, wax accumulation, and gene expression. For this, eighth rosette leaves of wild-type and trichome-free (gl1) Arabidopsis plants were harvested every four days during development, and investigated using a dissecting and confocal microscope, gas chromatography (GC) with flame ionization detection (FID) or coupled to mass spectrometry (MS), and quantitative RT-PCR.   2.2 MATERIALS AND METHODS  2.2.1 Plant material and growth conditions Arabidopsis wild type (ecotype Columbia-0) and gl1 (SALK_039478C) seeds were stratified at 4°C for three days and sterilized by shaking in 70 % ethanol for 15 min, followed by a washing step with 100 % ethanol for 2 min. The seeds were dried under sterile conditions and plated on ½ MS-agar medium. Seeds were germinated in the growth chamber at 21°C with a long-day light cycle (16 h-day/8 h-night) with 110 µE m-2 s-1 of photosynthetically active radiation. 14 day-old seedlings were transplanted into soil (Sunshine Mix4, SunPro), and grown under the same conditions as described above. The plants were watered twice per week. The sizes of eighth leaf blades and petioles were determined daily using a ruler.   24 2.2.2 Leaf morphological analysis Eighth leaves were harvested every two days between five and 21 days of age for morphometric measurements. Five leaves were segmented (Appendix Figure A.1), then each segment was stained with propidium iodide (100 mg/ml, Sigma) and images were captured with a confocal laser-scanning microscope (Radiance 2000, Bio-Rad, ex. 568 nm, em. 580-600 nm). Images were processed with ImageJ to obtain the total area and number of pavement and guard cells on each segment. The mean number and size of cells on each leaf were determined by averaging across segments. An additional five leaves were photographed under a dissecting light microscope, and the number of trichomes was counted.  2.2.3 Wax sample preparation and permutation tests Eighth leaves were harvested every four days between five and 21 days of age for wax analysis by cutting between blade and petiole. Leaf blades were pooled to have a sum surface area of at least 10 cm2 for each sample, and five independent samples were taken for each time point. To detect meaningful trends over time, R (R Core Team, 2015) was used to conduct permutation tests on the ordinary least squares slopes of the data from gl1 mutant and wild-type leaves. A total of 42 tests were performed using the gl1 mutant data (see Figure 2.3): one for each compound class (seven tests), one for the relative abundance of each homolog within the total wax mixture (26 tests), and one for the relative abundance of the sum total of compounds derived from each precursor chain length (nine tests). The same tests were also performed on the data from wild-type leaves (see Figure 2.4). Example R code used to conduct this analysis appears in Appendix B.3.   25 For leaf tip and base analyses, 30 eighth leaves at 13 days of age were harvested, cut into thirds, and then the 30 tip sections and 30 base sections were divided into four independent collections each. A total of 29 permutation tests on the mean difference between the relative abundance of each homolog in each compound class of the base and tip samples was performed to detect significant effects (see Figure 2.6). Example R code used to conduct this analysis appears in Appendix B3. Prior to extraction, each independent set of leaves was photographed alongside a ruler, and ImageJ (www.imagej.nih.gov/ij) was used to count leaf pixels to determine the surface areas extracted.  2.2.4 Wax extraction and derivatization Leaf waxes were extracted by submerging whole organs consecutively in two aliquots of chloroform (CHCl3) for 30 s each. The two solutions were pooled, and a known amount of n-tetracosane was added as an internal standard. The CHCl3 was removed under vacuum before derivatizing with 10 µl N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and 10 µl pyridine for 30 min at 70°C. The solvents were then evaporated under a gentle stream of N2 gas while heating at 50°C, before the wax was again dissolved in CHCl3.  2.2.5 Wax identification and quantification Wax constituents were separated by capillary GC (6890N, Agilent, Avondale, PA, USA; column 30 m HP-1, 0.32 mm i.d., df=0.1 µm) using the following temperature regime: on-column injection at 50°C, oven held for 2 min at 50°C, raised by 40°C min-1 to 200°C, held for 2 min at 200°C, raised by 3°C min-1 to 320°C, and held for 30 min at 320°C. For compound identification, the GC was linked to a mass spectrometric detector (5973N, Agilent) and the inlet pressure programmed  26 for a constant 1.4 ml min-1 flow of He carrier gas. For compound quantification, the GC with inlet pressure, programmed for constant flow of 2.0 ml min-1 of H2 carrier gas, was connected to a flame ionization detector (FID). The quantity (µg) of each wax constituent was established by comparison to a defined amount of n-tetracosane, the internal standard added into the total wax extracts. To determine the extracted cuticle areas, apparent surface areas were calculated with ImageJ software (www.imagej.nih.gov/ij) from digital photographs of the samples and multiplying by 2 or π for total leaf or stem surface areas, respectively. The additional surface area resulting from the three-dimensional, projecting structure of trichomes was not included in calculating the true surface area. Wax loads (µg cm-2) were determined by dividing the compound quantity by the corresponding extracted surface area.  2.2.6 RNA extraction and gene expression analysis by quantitative RT-PCR For gene expression analysis, eighth leaves were harvested every four days between five and 21 days of age and immediately frozen in liquid nitrogen. An average amount of 50 mg plant material was homogenized with Zirconia beads (2 nm diameter, BioSpec Products) at 4°C using a Precellys-24 homogenizer (Bertin, 5500 rpm, 2 x 25 sec). Total RNA was extracted using the PureLink RNA mini kit (Invitrogen) as described in the manufacturer’s protocol. On-column DNA digestions were performed using PureLink DNase Set (Invitrogen) following the manufacturer’s protocol. The integrity of extracted RNA and absence of genomic DNA was confirmed by agarose gel electrophoresis (2 %), and concentrations and purity were determined by measuring UV spectra and 260/280 and 260/230 ratios using NanoDrop 8000 Spectrophotometer (ThermoFischer Scientific) and samples with a 260/230 ratio between 1.7 and 2.2 were selected as templates for qRT-PCR. For first-strand cDNA synthesis, 5 µg total RNA  27 and Oligo(dT)20 primers (Invitrogen) were used together with SuperScript Reverse Trancriptase II (Invitrogen) following the manufacturer’s protocol, and resulting cDNA samples were stored at -20°C.  The expression level of wax biosynthesis genes was measured by quantitative RT-PCR using iQ SYBR Green Supermix (Bio-Rad) following the manufacturer’s procedure. Hard-shell 96-well PCR plates with thin walls were used (Bio-Rad) with the CFX Connect Real-time PCR Detection System (Bio-Rad) under the following PCR conditions: one cycle at 95°C for 3 min, followed by 40 cycles at 95°C for 15 sec, and at 60°C for 30 sec. A total amount of 50 ng cDNA was used in a 10 µl reaction volume. A no template control was included in each run and all samples were run as triplicates and in two biological replicates on 96 well plates (Bio-Rad). Melting curves were performed after each run to verify primer specificity. The primer efficiency was tested for each primer pair (between 93 % and 108 %) and used for normalization. CFX Connect Real-time PCR Detection System software (Bio-Rad) was used for data acquisition, and data were analyzed by adjusting the threshold cycles. Relative expression was calculated using the Pfaffl method (Pfaffl, 2001). Four genes (Actin2, Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), Ubiquitin conjugating enzyme 21 (UBC21), Polyubiquitin 10 (UBQ10)) were tested as reference genes, and UBC21 was used for normalization to determine relative gene expression.  Just as for the chemical data, permutation tests were conducted on the least squares slopes of the expression profiles of the 12 genes to detect significant changes over time. To determine which effects remained robust to tempering of possible inflation of gene expression levels at day five, a sensitivity analysis of these effects was performed. The output of these statistical procedures is  28 summarized in Figure 2.7, while R code and full details of the statistical analysis appear in Appendix B.2.  2.2.7 Adjustment for multiple comparisons To control possible inflation of the rate of Type I error due to the many statistical comparisons being made, the Benjamini-Hochberg method was used so that the study-wide expected proportion of falsely rejected null hypotheses was no more than 1 % (Benjamini & Hochberg, 1995). All effects found statistically significant according to this procedure were flagged with an asterisk.  2.3 RESULTS The goal of this chapter was to link changes in cuticular wax coverage and composition on growing Arabidopsis leaves with epidermal cell development and potential changes in the expression of underlying wax biosynthesis genes. Based on preliminary studies, the eighth rosette leaf was selected for investigation.   2.3.1 Morphological changes on developing leaves Leaf surface areas and the numbers of epidermal cells were measured as a function of leaf age. Under the growth conditions used here, leaf blades expanded steadily from 10 mm2 at day five to 138 mm2 at day 21 (Figure 2.1A), after which they did not change (Appendix Table A.1). The length of the petiole increased from 1 mm at day five to 14 mm at day 21.   29 The average surface area of pavement cells increased from ca. 490 µm2 to 3910 µm2 between days five and 21 (Figure 2.1A), and thus pavement cell expansion accounted for the majority of macroscopic leaf growth. However, leaf expansion was also partially due to pavement cell division, which led the number of pavement cells to increase from ca. 20,500 per blade at day five to ca. 29,100 by day 13 and remain roughly constant thereafter (Figure 2.1B). Concomitantly, the number of guard cell pairs increased from day five (ca. 6,700 per blade) to day 13 (10,100) and then also stayed constant (Figure 2.1B). The number of trichomes, ca. 75, was constant throughout eighth leaf development, resulting in relatively high trichome densities on younger leaves and a steady decrease in trichome density as the leaf expanded (Figure 2.1B).    30  Figure 2.1: Surface area and epidermal cell numbers on developing wild-type Arabidopsis eighth rosette leaves. A) Leaf surface area (black circles, left y-axis) and pavement cell surface area (white circles, right y-axis) measured between five and 21 days of leaf age. B) Number of pavement cells (white squares, left y-axis), guard cells (light grey squares, left y-axis), and trichome cells (dark grey squares, right y-axis) measured between five and 21 days of leaf age. Point positions and error bars indicate the mean and standard deviation of five independent measurements, respectively.  B A  31 2.3.2 Cuticular waxes from whole Arabidopsis leaves of different ages To monitor leaf wax development, leaf surfaces were extracted every four days during growth, and the resulting leaf wax mixtures were analyzed with GC-MS and GC-FID. To delineate effects arising from shifts in the relative abundance of trichome cells over time, the waxes from developing leaves of both the trichome-free mutant gl1 (2.3.2.1) and wild-type plants (2.3.2.2) were investigated.  2.3.2.1 Wax development on growing eighth rosette leaves of the trichome-free Arabidopsis mutant gl1 Five-day-old gl1 leaves were covered with 0.62 ± 0.05 µg/cm2 extractable wax (Figure 2.2), comprising fatty acids (34 % of the overall wax mixture), alkanes (26 %), branched alcohols (13 %), n-alcohols (7 %), aldehydes (2 %), alkenes (2 %), and 16 % of unidentified compounds (Figure 2.3A). At 21 days of age, gl1 leaves bore 0.64 ± 0.04 µg/cm2 wax, which was composed of alkanes (42 %), fatty acids (19 %), branched alcohols (11 %), n-alcohols (7 %), aldehydes (5 %), and alkenes (2 %), leaving 14 % unidentified. Thus, wax coverage remained constant over the course of gl1 leaf development, and after day five the composition of the mixture gradually shifted, becoming significantly less fatty acid- and more alkane-dominated.  32  Figure 2.2: Wax coverage on developing Arabidopsis gl1 and wild-type eighth rosette leaves.  The total amount of wax from gl1 and wild type is expressed as wax mass (µg) per surface area extracted (cm2). Eighth rosette leaves were harvested every four days from five to 21 days of leaf age, and their total wax coverage was measured with GC-FID. Lines connect the mean of the five independent measurements. The slope of a first order linear model fitted to the wild-type data was significantly different from zero (p < 0.0001).   Each of the chemical classes was present as a homologous series spanning chain length ranges typical of cuticular waxes. Fatty acids were present with even numbers of carbons between C22 and C32, within which the C24 and C26 homologs were the most prominent (Figure 2.3B). The n-alcohols also had predominantly even carbon numbers, but ranged from C26 to C34 with an approximately normal distribution. Three branched-chain alcohols with total carbon numbers C30, C32, and C34 were identified, with the C32 homolog being the most abundant. Four un-branched aldehydes with even carbon numbers from C28 to C34 were detected in roughly equal amounts. n-Alkanes were found with odd carbon numbers from C27 to C37, the C29, C31, and C33 0.00.20.40.60.81.01.25 7 9 11 13 15 17 19 21Age of leaf eight (days)Wax coverage (µg/cm2 )gl1WT 33 homologs being the most abundant. Finally, C35 and C37 n-alkenes were present in roughly equal amounts.  Chain length shifts were observed during gl1 leaf development, and were most pronounced within the fatty acid and alkane series. From day five to day 21, the relative abundance of C24 fatty acid decreased steadily from 13 % to 3 % of the total wax load (Figure 2.3B). Simultaneously, relative amounts of C29 and C31 alkane increased steadily from ca. 7 % to 18 % and 9 % to 18 %, respectively, while the relative abundance of C33 alkane decreased from 11 % to 3 %. Thus, a significant, combined shift from acid- to alkane-dominance and from C24 to C29 and C31 compounds was observed over the course of development in gl1 leaves.  Lastly, the overall chain length profiles across all compound classes within the wax mixture on developing gl1 leaves were calculated. For this, the relative amounts of all compounds formed by modification of the same acyl-CoA precursor were added together, e.g. C30 acid, C30 alcohols, C30 aldehyde, and C29 alkane. At five days of age, gl1 leaf wax was made up of ca. 22 % each C26 and C32 compounds, and ca. 16 % each C24, C30, and C34 compounds (Figure 2.3C). At 21 days of age, the mixture contained ca. 45 % C32 compounds, 30 % C30 compounds, and 26 % C26 compounds, together with smaller amounts of C24 and C34 compounds. Relatively minor quantities of compounds with chain lengths C22, C28, C36, and C38 were present throughout development. Overall, the most pronounced changes over time were observed as increases in the relative abundance of C32 and C30 compounds and decreases in the relative abundance of C24, C26, and C34 compounds.  34 Figure 2.3: Wax composition on developing Arabidopsis gl1 mutant eighth rosette leaves. A) Relative abundance of each compound class as a function of leaf age. B) Relative abundance of each homolog within compound classes as a function of leaf age. Labels on the x-axis indicate the carbon number and class of each identified compound. C) Relative abundance of each chain length as a function of leaf age. Odd-numbered homologs of alkanes and alkenes are known to be derived from respective even-numbered precursors with one carbon more, and both are therefore summed and shown together. Bar heights and error bars indicate the mean and standard deviation of five independent measurements, respectively. Lines indicate the mean of the five independent measurements made at each time point. Asterisks indicate significant time-dependent changes derived from a permutation test of the ordinary least squares linear fit after study-wide adjustment for multiple comparisons.   2.3.2.2 Wax development on growing eighth rosette leaves of wild-type Arabidopsis The cuticular wax of wild-type eighth rosette leaves was harvested and analyzed in the same way and at the same points during development as described for gl1. Coverage on wild-type leaves decreased significantly from five days of age (1.06 ± 0.02 µg/cm2) to day 21 (0.86 ± 0.06 µg/cm2) (Figure 2.2).  The wax on five-day-old wild-type eighth leaves contained alkanes (39 %), fatty acids (21 %), n-alcohols (9 %), branched alcohols (6 %), alkenes (7 %), and aldehydes (3 %), leaving 15 % of AFatty acids n-Alcohols br.-Alcohols Aldehydes Alkanes Alkenes Unidentifieds* * *010203040505 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21Age (days)Rel. abundance  (% total wax)B***** * *****0246810121416182022242622 24 26 28 30 32 26 28 30 32 34 30 32 34 28 30 32 34 27 29 31 33 35 37 35 37Rel. abundance  (% total wax) day 5day 9day 13day 17day 21Fatty Acids n-Alcohols br.-Alcohols Aldehydes Alkanes AlkenesCTotal C22 Total C24 Total C26 Total C28 Total C30 Total C32 Total C34 Total C36 Total C38* * * * * * *010203040505 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21Age (days)Rel. abundance  (% total wax)A B C  35 the wax unidentified. Alkanes were the most prominent compound class throughout development of wild-type leaves, increasing significantly from 39 % to 49 % of the total wax mixture between days five and 21 (Figure 2.4A). Over the same time interval, the relative abundance of fatty acids and alkenes decreased significantly from 20 % to 12 % and from 7 % to 2 %, respectively, while other compound classes exhibited little fluctuation.  Wild-type leaf wax comprised homologous series of fatty acids (C22 – C32), n-alcohols (C26 – C34), branched alcohols (C30 – C34), and aldehydes (C28 – C34), all with predominantly even carbon numbers, as well as alkanes (C27 – C37) and alkenes (C35 – C37) with predominantly odd carbon numbers. Overall, the same compounds were identified in the leaf waxes of wild type and the gl1 mutant. Homolog distributions within these chain length ranges were also similar on both lines. Over time, the relative abundance of C24 fatty acid decreased significantly from ca. 8 % to 2 % in wild-type wax, while C29 and C31 alkanes increased significantly from 5 % to 17 % and from 15 % to 20 %, respectively. In contrast, the relative abundance of the longer alkanes (C33, C35, and C37) as well as C35 and C37 alkenes exhibited a significant decrease. In terms of constituent chain lengths, wild-type eighth leaf wax was made up of 29 % C32 compounds, ca. 20 % C34 compounds, 13 % each C26 and C30 compounds, ca. 9 % each C24 and C36 compounds, and ca. 4 % C38 compounds at day five. At 21 days of age, the mixture consisted of 38 % C32, 28 % C30, 13 % C26, 11 % C34 compounds, ca. 2 % each C24 and C36 compounds, and 1 % C38 compounds. Thus, the relative abundance of C30 and C32 compounds increased significantly over time, while that of C24, C34, C36, and C38 compounds decreased significantly. Throughout development C22 compounds contributed relatively minor amounts to the total wax mixture.   36  Figure 2.4: Wax composition on developing Arabidopsis wild-type eighth rosette leaves. A) Relative abundance of each compound class as a function of leaf age. B) Relative abundance of each homolog within compound classes as a function of leaf age. Labels on the x-axis indicate the carbon number and class of each identified compound. C) Relative abundance of each chain length as a function of leaf age. Odd-numbered homologs of alkanes and alkenes are known to be derived from respective even-numbered precursors with one carbon more, and both are therefore summed and shown together. Bar heights and error bars indicate the mean and standard deviation of five independent measurements, respectively. Lines indicate the mean of the five independent measurements made at each time point. Asterisks indicate significant time-dependent changes derived from a permutation test of the ordinary least squares linear fit after study-wide adjustment for multiple comparisons.   2.3.3 Regional distribution of wax on Arabidopsis leaves To test whether differences in wax composition also existed between discrete leaf regions potentially differing in average cell age, leaves were hand-sectioned transversely (Appendix Figure A.1), and the size distribution of pavement cells on each section was investigated. Five-day-old leaves exhibited a pavement cell size gradient along their longitudinal axes from 300 ± 50 µm2 at the base of the blade to 900 ± 200 µm2 at the tip (Figure 2.5). A larger size difference was observed on leaves at day nine, where pavement cells varied in size from 1,000 µm2 at the AFatty acids n-Alcohols br.-Alcohols Aldehydes Alkanes Alkenes Unidentifieds* * *010203040505 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21Age (days)Rel. abundance  (% total wax)B******** ******* *0246810121416182022242622 24 26 28 30 32 26 28 30 32 34 30 32 34 28 30 32 34 27 29 31 33 35 37 35 37Rel. abundance  (% total wax) day 5day 9day 13day 17day 21Fatty Acids n-Alcohols br.-Alcohols Aldehydes Alkanes AlkenesCTotal C22 Total C24 Total C26 Total C28 Total C30 Total C32 Total C34 Total C36 Total C38* * * * * * *010203040505 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21 5 9 13 17 21Age (days)Rel. abundance  (% total wax)B C A  37 base to 3,000 µm2 at the tip. Similar gradients were observed on leaves at days 13 and 17. In contrast, at day 21, pavement cells along the entire blade were similar in size, ca. 5,000 ± 1,000 µm2. Overall, the pavement cell size distributions indicated that pavement cells at the leaf tip initiated and completed expansion before those at the leaf base. At 13 days of age leaves exhibited the largest difference in pavement cell size and age between leaf base and tip, with cells at the base most similar to those found on young leaves and cells at the tip most similar to those found on mature leaves.   Figure 2.5: Pavement cell size on segments of developing Arabidopsis wild-type eighth rosette leaves.  Pavement cell surface areas (µm2) as a function of their position along the longitudinal axis of the leaf (mm, measured from the leaf blade base) are plotted for leaves between five and 21 days of age. Point positions and error bars indicate the mean and standard deviation of five independent measurements, respectively.    38 To correlate pavement cell size and age with wax composition, 13-day-old leaves were cut into three equal-sized pieces, and the wax compositions of the base and tip segments were determined as described above for whole-leaf wax analysis. The identity and span of the classes of homologous compounds identified on both the leaf bases and tips were identical, comprising fatty acids (C22 – C34), n-alcohols (C24 – C34), branched alcohols (C30 – C34), aldehydes (C26 – C34), alkanes (C27 – C37), and alkenes (C35 – C37), but the relative abundance of some of the compounds differed between the leaf tips and bases. In particular, C24 and C26 acids made up 5 %  and 14 % of the total wax on leaf bases, respectively, while on the tips they were represented only 2 % and 8 % (Figure 2.6). Furthermore, leaf bases were covered with 9 % and 13 % C29 and C31 alkanes, respectively, while these compounds comprised 17 % and 19 % of the wax on leaf tips. All other compounds were present in similar amounts on leaf tips and bases.    Figure 2.6: Wax composition on the base and tip sections of wild-type Arabidopsis eighth rosette leaves at 13 days of age. The relative abundance of each homolog in each compound class is plotted as a percent of the total amount of extracted wax. Labels on the x-axis indicate the carbon number and class of each identified compound. Leaf base and tip samples were prepared by selecting eighth leaves at 13 days of age, cutting them into three equal segments, and independently extracting the base and tip segments. Twenty segments were pooled for each sample. Bar heights and error bars indicate the mean and standard deviation of five independent measurements, respectively. Asterisks indicate significant mean differences between base and tip derived from a permutation test and adjusted for study-wide multiple comparisons. ***** * ****02468101214161820222422 24 26 28 30 32 34 24 26 28 30 32 34 30 32 34 26 28 30 32 34 27 29 31 33 35 37 35 37Rel. abundance (% total wax)basetipFatty Acids n-Alcohols br.-Alcohols Aldehydes Alkanes Alkenes 39 2.3.4 Expression of wax biosynthesis genes during leaf development To assess possible contributions of candidate genes to wax production, their expression patterns were monitored at the same time points used for wax sampling. To account for variation in the amount of tissue harvested, the expression levels of target genes had to be normalized. Preliminary experiments showed that, of the four reference genes Actin2, GAPDH, UBQ10, and UBC21, the latter exhibited the most stable expression over all time points (data not shown) and was therefore used to normalize expression levels of target genes. Based on their function, I grouped the investigated genes as coding for 1) condensing enzymes of elongation complexes (KCSs), 2) proteins associated with the elongation complex(es), and 3) head group-modifying enzymes.  Among the KCS genes, CER6 and KCS8 were expressed at much higher levels than KCS1, KCS5, and KCS16 (Figure 2.7A). Over the course of leaf development, KCS1 and CER6 expression increased significantly, expression of KCS8 decreased significantly, and that of KCS5 and KCS16 showed a slight decrease between days five and 13. All the genes coding for proteins associated with elongation complex(es), CER10, CER8, CER2 and CER26, were expressed at relatively high and constant levels throughout leaf development (Figure 2.7B), excepting the slightly higher expression of CER10 and CER8 at day five. Among the head group-modifying enzymes, CER3 and CER1 were expressed at intermediate levels and CER4 at very low levels (Figure 2.7C). Expression of CER3 and CER1 varied little over the course of leaf development, with higher expression of CER1 at day five, while CER4 expression increased slightly but significantly throughout development. 40  Figure 2.7: Expression of wax biosynthesis genes in developing wild-type Arabidopsis eighth leaves. Expression levels were determined using qRT-PCR and normalized against the reference gene UBC21. A) Expression of selected KCS genes, encoding the FAE complex enzymes responsible for chain length control during wax precursor elongation. B) Expression of genes encoding proteins associated with the KCSs during elongation. C) Expression of genes involved in the conversion of wax precursors into wax compounds. Just as for the chemical data, changes in transcript levels as a function of leaf development between five and 21 days of age were analyzed with permutation tests on robust ordinary least squares method. Trends found to be significantly different from zero by this test after study-wide adjustment for multiple comparisons were flagged with an asterisk (*). Due to apparent inflation of gene expression measurements on day five, a sensitivity analysis was performed, and trends that were significant upon lightening or omission of day five data were flagged with a diamond (u). Details of statistical tests are presented in Appendix B.2).   2.4 DISCUSSION Overall, our morphometric results first established that trichome density decreased steadily during leaf development, that leaf tip expansion preceded leaf base expansion, and that gradients in pavement cell maturity and size between leaf tips and bases were maximized around 13 days of leaf age. Wax coverage on wild type decreased significantly with expansion, while gl1 had constant, lower coverage than the wild type. CER1Age of leaf (days)5 9 13 17 21024681012*5 9 13 17 210.00.10.20.30.4Age of leaf (days)CER4B C A CER25 9 13 17 21024681012Age of leaf (days)CER3Age of leaf (days)Gene expression (relative to UBC21)5 9 13 17 21024681012CER265 9 13 17 21024681012Age of leaf (days)5 9 13 17 210.00.51.01.52.0Age of leaf (days)Gene expression (relative to UBC21) KCS1*Age of leaf (days)5 9 13 17 210.00.51.01.52.0 KCS5*5 9 13 17 21024681012Age of leaf (days)CER6*5 9 13 17 21024681012Age of leaf (days)KCS8*5 9 13 17 210.00.51.01.52.0Age of leaf (days)KCS16*CER105 9 13 17 21024681012Age of leaf (days)Gene expression (relative to UBC21)*CER8Age of leaf (days)5 9 13 17 21024681012* 41 An age-dependent shift from C24/C26 to C30/C32 compounds was accompanied by a relative decrease in fatty acid abundance and an increase in alkane abundance on both plant lines. Wild-type leaves also exhibited a simultaneous decrease in the relative abundance of C35+ compounds. qRT-PCR analyses showed that head group-modifying enzymes were expressed at fairly constant, albeit different levels. An increase in the expression of KCS1 and CER6 was accompanied by a decrease in that of KCS5 and KCS16, whereas non-KCS elongation genes were expressed at roughly constant levels throughout leaf development. These findings can now be integrated to discuss wax dynamics in the context of pavement cell age (2.4.1), leaf expansion (2.4.2), and epidermal cell composition (2.4.3).  2.4.1 Pavement cell age effects on wax composition Arabidopsis leaf morphology was monitored to give context to our wax composition and gene expression data. The data acquired here spanned leaf development from an early stage, defined by the onset of pavement cell expansion, to a late stage when pavement cell expansion was largely complete. This occurred after 21 days of growth, similar to previous observations of Arabidopsis leaf six (Granier et al., 2002). Like other studies on leaf growth of Arabidopsis and other species (Donnelly et al., 1999; Byrne, 2005; Efroni et al., 2010; Nath et al., 2016), we found that pavement cells near the leaf tip begin to expand at a time when many cells near the base are still dividing, and that tip cells finish expansion long before base cells. Thus, the position of a pavement cell along the leaf axis is related to cell age, with the youngest cells being present at the leaf base. Accordingly, a comparison of leaf bases and tips (a spatial distribution) may serve as an orthogonal  42 method of comparing young and old leaves (a temporal distribution). Under our growth conditions, the largest difference in cell maturity between tip and base pavement cells was reached at day 13, when tip cells had reached their final size and base cells had only begun to expand, which is nearly identical to what has been reported for leaf three (Andriankaja et al., 2012).  By both spatial and temporal comparisons, leaves of gl1 and wild type exhibited decreases in the relative abundance of C24/C26 compounds and increases in that of C30/C32 compounds or their C29 and C31 alkane derivatives. A previous analysis of Arabidopsis leaves also found the relative abundance of C31 alkane to increase with age, albeit accompanied by a decrease in that of C29 alkane (Jenks et al., 1996). Furthermore, an increase in wax compound chain lengths, mostly of alkanes, has been observed on the developing leaves of the monocot Sorghum bicolor (Atkin & Hamilton, 1950) and diverse dicots including Malus domestica (Bringe et al., 2006), Sesamum indicum (Kim et al., 2009), Coffea arabica (Stocker and Ashton, 1975), and Rhododendron fortunei (Salasoo, 1983). In contrast, leaves of Kalanchoe daigremontiana (van Maarseveen et al., 2009) and several Triticum spp. exhibited alkane shortening as they developed (Tulloch, 1973; Van Maarseveen et al., 2009).  Chain length profiles of wax compounds are established by elongation in FAE complexes, within which the KCS enzymes are known to exert control over product chain length distributions (Fehling & Mukherjee, 1991; Millar & Kunst, 1997). Consequently, it seems plausible that differential expression of one or more KCS enzymes may lead to  43 the observed shift from C24/C26 to C30/C32 wax compounds. I found that the expression levels varied greatly between KCS genes, as reported previously (Joubès et al., 2008; Kim et al., 2013), and in many cases also over the course of leaf development. Expression of KCS5, KCS8, and KCS16 decreased with age, indicating that these are probably not involved in the increase of C30/C32 wax compounds. In contrast, expression of KCS1 increased with leaf age, thus paralleling the main chain length shift. However, heterologous expression of KCS1 yielded C20 - C26 products (Trenkamp et al., 2004) and, conversely, the kcs1 mutant is affected mainly in the accumulation of C26 and C28 wax compounds (Todd et al., 1999), together suggesting that this enzyme is likely not involved in the observed chain length shift. In contrast, CER6 was highly and increasingly expressed throughout leaf development. This KCS produces C30 and C32 compounds when heterologously expressed with CER2 and CER26 (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015), and the cer6 mutant accumulates C26 compounds (Fiebig et al., 2000; Hooker et al., 2002; Jenks et al., 1995). Together with these literature data, our observations now strongly suggest that the turnover of C26 acyl-CoA precursors and the resulting accumulation of C30/C32 products are largely controlled by the expression level of CER6 during Arabidopsis leaf development.  Since CER2 and the CER2-like proteins are also involved in chain elongation by allowing CER6 to produce compounds with 30 or more carbons (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015), it seemed plausible that they may also play a role in the chain length shift. However, the expression levels of the CER2-like genes did not change during leaf development, making their involvement in the shift unlikely. Their  44 high expression throughout also implies that there is an excess of CER2 and CER26 transcripts at the outset of pavement cell expansion, and that the abundance of CER6 transcripts, rather than that of CER2-likes, initially limits the production of longer chain lengths. Interestingly, CER10, also part of the FAE complex, showed significantly decreased expression, most markedly during early leaf development. However, since CER10 is equally involved in all elongation rounds, it seems unlikely that the chemical changes are driven by differential expression of CER10. Similarly, CER8 also showed only a slight decrease in expression levels mainly during early development, making it unlikely that the encoded LACS enzyme probably forming acyl-CoAs from fatty acids (Lü et al., 2009) has a role in driving the chemical changes.   In addition to the chain length shift, our temporal and spatial wax comparisons also revealed a compound class shift from fatty acids to alkanes. This trend is similar to a previous report on Arabidopsis leaf wax development (Jenks et al., 1996), except for the higher overall amounts of fatty acids found here. Many other species exhibit compound class shifts similar to that observed for Arabidopsis, including Coffea arabica (Stocker & Ashton, 1975), Prunus laurocerasus (Jetter & Schäffer, 2001), and Malus domestica (Bringe et al., 2006), though in some, such as Sorghum bicolor, both fatty acids and alkanes continue to accumulate as leaves age (Atkin & Hamilton, 1950).   The head group modification pathways generating the different compound classes all draw from the same acyl-CoA precursor pool, and the expression levels of genes encoding respective modifying enzymes are therefore prime candidates for instigators of  45 the observed compound class shift. However, expression of CER3 and CER1, the genes involved in alkane production, did not increase during leaf development. Interestingly, the expression level of CER4, encoding the reductase responsible for production of primary n-alcohols, showed a slight increase over time, despite fairly constant wax alcohol amounts. Taken together, our data suggest that a balance of head group-modifying enzymes for the alkane- and alcohol-forming pathways may be maintained by relatively constant expression of all genes encoding them. Conversely, the drastic change in chemical composition from acids to alkanes cannot be explained by differential expression of the genes coding for head group-modifying enzymes.  While it is possible that the shift from fatty acids to alkanes is the consequence of the down-regulation of a gene responsible for acid formation, no such enzyme activity has been identified. Alternatively, the increase in alkane abundance could arise from the increase in CER6 expression, if the precursors generated by CER6 are channeled preferentially into the alkane pathway instead of being equally available to the alcohol-, alkane-, and acid-forming pathways. Interestingly, such association between KCS enzymes and head group modification pathways could also explain the selective effect of kcs1 on the alcohol and acid compound classes (Todd et al., 1999).  Finally, it should be noted that, in addition to the increase in the relative abundance of C30/C32, there was also a significant decrease in C34 products (mainly in the form of C33 alkane) during development of both gl1 and wild-type leaves (compare Figure 2.3B). This trend might be due to gradually decreasing expression of an unknown KCS with C34  46 product specificity, or to a steady increase in competition between that KCS and the more highly expressed CER6. Interestingly, the latter explanation would imply that the products of the CER6 -containing FAE, C30/C32 acyl-CoAs, preferentially serve as substrates for alkane formation (by CER3 and CER1) rather than for further elongation, thus providing further support for substrate channeling from the CER6-containing FAE into the alkane-forming pathway.  2.4.2 Leaf expansion effects on wax coverage The total wax coverage on mature wild-type leaves was ca. 0.9 µg/cm2, well within the range of literature values (0.5 µg/cm2 to 1.8 µg/cm2) for this ecotype (Bernard et al., 2012; Bourdenx et al., 2011; Haslam et al., 2012; Pascal et al., 2013). We found that the glabrous leaves of gl1 had a roughly constant coverage of ca. 0.6 µg/cm2 at all stages of development, approximately 65-70 % of the macroscopic coverage of the corresponding wild type, which is consistent with previously reported ratios (Reisberg & Hildebrandt, 2012). It should be noted that the Arabidopsis wild type wax coverage presented here, like those in previous reports, were calculated based on projected, macroscopic leaf surface area, thus ignoring the contribution of trichomes to the true, microscopic surface area.  High trichome densities on young leaves and steady decreases in trichome density during leaf expansion had been reported for many species such as Bemisia tabaci and Arabidopsis (Chu et al., 2001; Mauricio, 2005). Furthermore, a decrease in macroscopic coverage with age/expansion was observed on leaves of trichome-bearing species such as  47 Arabidopsis (Jenks et al., 1996), Malus domestica (Bringe et al., 2006), Sesamum indicum (Kim et al., 2009), Hedera helix (Viougeas et al., 1995), and Fagus sylvatica (Gülz et al., 1992), but was not observed on leaves lacking trichomes, such as those of Kalanchoe daigremontia (van Maarseveen et al., 2009). Overall, these observations indicate that the decrease in macroscopic coverage on developing leaves is probably due to a decrease in trichome density.  Due to their complex geometry, the surface area of Arabidopsis leaf trichomes and their contribution to the total leaf surface area are difficult to assess accurately. We counted an average of 75 trichomes on eighth leaves regardless of their age, a number roughly similar to previous reports and well within the range of counts for different rosette leaves (Esch et al., 2003). Approximating a trichome as a cylinder (of 50 µm diameter and 150 µm height) topped with three cones (each 20 µm across and 150 µm high) (Marks, 1997), the sum surface area of all trichomes on the blade is estimated as 0.03 cm2. Consequently, trichomes may add ca. 15 % and 1 % to the projected surface area of young and mature eighth leaves, respectively. By combining the trichome and pavement cell surfaces, the overall surface area of young and mature leaves must be corrected to 0.23 cm2 and 2.83 cm2, respectively. The wax coverages of wild-type leaves may accordingly be re-calculated, resulting in coverages of ca. 0.9 µg/cm2 irrespective of age. Thus, wax coverages are roughly constant during growth of both wild type and gl1 leaves, as has been observed for growing Arabidopsis stems (Suh et al., 2005), together suggesting that wax production is synchronized with cell expansion.  48 To maintain constant wax coverage on expanding leaf surfaces, the epidermis must produce wax at a particular rate. According to published expansion rates and coverages for Arabidopsis stem segments (Suh et al., 2005), the rapidly expanding stem tops produced wax at ca. 1.1 µg/day, while the middle of the stem produced 0.15 µg/day (Table 2.1). Based on the surface expansion of wild-type eighth leaves (0.16 cm2/day), we estimate a wax production of ca. 0.15 µg/day between days five and 21 (Table 2.1). Similar calculations may be performed for specific epidermal cell types by first determining cell-specific wax coverage. Based on their approximated surface area (0.03 cm2), the difference in the absolute wax amounts on young and mature wild type leaves (0.05 µg), and the time it takes for a trichome to expand fully (1.5 days), trichome cells should produce wax at a rate of roughly 18 (pg/hr)/cell (Table 2.2). For pavement and guard cells together, this rate is approximately 0.1 (pg/hr)/cell (Table 2.2). Overall, these estimations show that rates of wax production vary considerably between cell types and organs to maintain constant coverage, suggesting that wax production is tightly linked to (cell) surface area expansion rates. The mechanisms controlling this synchrony at cell and organ-specific levels are far from clear, and may comprise genetic, biochemical, and/or physiological feedback regulation  Table 2.1: Cuticular wax coverage and production rates of different Arabidopsis organs.    !!!!!!!!!!WT#stem#middle WT#stem#top WT#leaf gl1#leafWax#coverage#(μg/cm^2) 30 30 0.9 0.6Wax#production#rate#(μg/day) 0.15 1.13 0.15 0.1Wax#production#rate#(μg/hr) 0.006 0.05 0.006 0.004 49 Table 2.2: Cuticular wax coverage and production rates of different Arabidopsis epidermal cell types.   2.4.3 Epidermal cell type effects on wax composition Fluctuations in trichome density may not only affect total wax coverage but also wax composition. We found a significant, time-dependent decrease in the relative abundance of C35 and C37 alkanes and alkenes on wild type leaves that was not observed on gl1 leaves (compare Figure 2.4 and Figure. 2.3), suggesting that these compounds may differ in their contributions to the total wax loads covering pavement cells and trichomes. This notion can be further investigated by approximating a trichome-specific wax composition by subtracting the wax composition of young (adaxial and abaxial) pavement and guard cells (i.e., the gl1 wax mixture) from the wax composition of young (adaxial and abaxial) pavement, guard, and trichome cells (i.e., the wild type wax mixture). As reported by others, we did not observe any abaxial trichomes on wild type leaves, the number of which vary with ecotype, leaf number, and light conditions (Telfer et al., 1997). Thus, this subtractive comparison will reflect the adaxial trichome wax composition, and we estimate the trichome wax to consist of n-alcohols (ca. 15 % of calculated total trichome wax), alkanes (62 %), and alkenes (ca. 15 %), with each compound class dominated by respective C32+ homologs (Appendix Figure A.2).  !!!!!!!!! !Leaf%trichomes%Leaf%pavement%and%guard%cellsWax%coverage%(μg/cm^2) 1.6 0.9Wax%production%rate%((μg/day)/leaf) 0.032 0.0675Wax%production%rate%((μg/hr)/leaf) 0.00133 0.00281Wax%production%rate%((pg/hr)/cell) 17.8 0.07 50 To evaluate chain length differences between wax mixtures on trichomes and pavement cells, the calculated trichome wax composition can be compared against the young gl1 wax composition reported here. The most abundant alkanes in the calculated trichome wax were C31 and C33 (ca. 42 % and 32 % of all alkanes, respectively, Appendix Figure A.2), while the relative abundances of these compounds were reversed in the young gl1 wax mixture (ca. 34 % and 41 %). However, the largest compositional differences were observed for C35+ alkanes and alkenes, which comprised ca. 30 % of all the alkanes and alkenes in the trichome wax mixture, but only 10 % of the alkenes and alkanes on young gl1 eighth leaves. Together, these comparisons strongly suggest that the wax mixture produced by trichome cells differs from that of pavement cells primarily in that it is enriched in C35+ alkanes and alkenes.  To identify gene candidates for the production of C33 – C37 compounds, KCSs must again be considered. However, although considerable progress has been made in the characterization of KCSs, none has been found that produces compounds longer than C34, even with the aid of CER2-like proteins. The data presented here point to KCSs involved in trichome wax elongation in particular. The decrease in trichome density was paralleled by apparent decreases in the expression levels of KCS5, KCS8, and KCS16, genes that had also been found expressed preferentially in trichomes (Marks et al., 2009). It should be noted that, since our gene expression measurements convey the average expression level in all leaf cells, it is possible that the relative expression level of these genes in trichome cells is underestimated. Consequently, KCS5, KCS8, and KCS16 may be considered as primary candidates for the production of C35+ compounds in trichomes.  51 KCS5 and CER6 share 88 % amino acid identity (Joubès et al., 2008), and the two have similar product profiles (Trenkamp et al., 2004), suggesting that KCS5, like CER6, only produces up to C30/C32 precursors. KCS8 and KCS16 show 74 % amino acid identity, are members of the same KCS subclass (Joubès et al., 2008), and have highest transcript abundance in leaves (Kim et al., 2013), though KCS16 is also highly expressed in siliques (Joubès et al., 2008). Based on the currently available data, both these KCSs remain likely candidates for the production of C35+ compounds, particularly in Arabidopsis leaf trichomes.  It should be noted that the calculated trichome wax composition contained relatively low amounts of branched alcohols. This compound class constituted ca. 14 % of the adaxial wax of gl1 leaves (Buschhaus & Jetter, 2012), but less than 1 % of the calculated trichome wax mixture (Appendix Figure A.2). Interestingly, while the chain lengths of n-alcohols on trichomes and pavement cells differ, as did the chain lengths of alkanes, branched alcohols do not follow this trend. These observations may indicate that branched alcohols are produced by a biosynthesis pathway distinct from that of the un-branched wax compound classes.  Finally, it is possible that changes in the relative abundance of guard cells may also contribute to observed coverage and/or compositional shifts, especially since wax biosynthesis gene expression within these cells can differ from that of neighboring pavement cells (Gray et al., 2000). However, guard cells are much smaller than both pavement and trichome cells, there were only half as many guard cells as pavement cells  52 present, and the ratio of guard cells to pavement cells remained constant throughout the period of leaf development studied here (compare Figure 2.1). These observations make it unlikely that guard cells contribute to the changes in wax coverage and composition reported here.  In summary, I observed shifts in both chain length and compound class composition of cuticular wax on growing Arabidopsis leaves followed the cell cycle arrest front, which and could be attributed mainly to differential expression of CER6.  53  The composition of surface wax of different Arabidopsis Chapter 3:thaliana epidermis cells  3.1 INTRODUCTION To maximize its function, the cuticle must form a continuous layer of hydrophobic material across the entire organ surface. It has therefore been assumed that all epidermal cell types are covered with cuticular waxes. Leaf and stem surface tissues consist of level pavement and guard cells, plus trichomes more or less protruding from the surfaces of many species. Arabidopsis contains only simple, non-glandular trichomes that likely inhibit insect movement, reduce mechanical abrasion, and increase the boundary layer of air to help reduce transpiration (Wagner et al., 2004).  Despite indirect evidence suggesting lateral differences in wax composition, a direct analysis of cell-specific wax composition is lacking to date. This may partly be due to the difficulty in isolating cuticular lipids free of both intracellular contamination and lipids from other cell types. In order to circumvent this problem, molecular genetic approaches combined with chemical analyses are required. Because of the size and conspicuous nature of trichomes, significant progress has been made in understanding the regulation of trichome development (Hülskamp, 2004; Ishida et al., 2008). Genetic studies have identified several genes that, singularly and in combination, regulate trichome numbers and thus also the ratio of trichomes per pavement cells (Listed in Marks et al., 2009).  54 Comparing the cuticular waxes of these lines will enable direct evidence to be acquired on cell type-specific compositions.  At one extreme, Arabidopsis mutations such as glabra1 (gl1) prevent trichome formation almost entirely (Herman & Marks, 1989) (Figure 3.1A & B). A previous study that analyzed gl1 leaf wax concluded that it was similar to that of published wild-type wax of different accessions (Sieber et al., 2000). In contrast, a fad7-1 gl1 double mutant likely reflecting the gl1 wax chemo-type was reported to have leaf wax with reduced levels of some alkanes and alcohols compared with wild type (Xia et al., 2010). Further conflicting evidence was provided in a third study, where gl1 leaf wax was found to be deficient in alkanes, but not alcohols (Reisberg & Hildebrandt, 2012). In the other extreme, Arabidopsis mutants such as cpc tcl1 etc1 etc3 cause a three-fold increase in trichome density (Figure 3.1E & F) compared to wild type (Figure 3.1C & D), thereby increasing the ratio of trichomes to pavement cells (Wang et al., 2008). However, wax analyses of such trichome-rich mutants have not yet been reported. Overall, it is thus not clear in how far trichome wax may differ from that on pavement cells.   55  Figure 3.1: Differences in trichome abundance between Arabidopsis gl1, wild type and cpc tcl1 etc1 etc3. Trichome abundance on stems, inflorescence and leaves of A and B gl1, C and D wild type and E and F cpc tcl1 etc1 etc3.   Furthermore, recent progress has also enabled the isolation of Arabidopsis trichomes in sufficient quantities for lipid analysis (Marks et al., 2008). Thus, the wax mixtures from the A E   C B D F 5 mm 5 mm 5 mm  56 surfaces of trichomes can now be isolated and analyzed in an attempt to further corroborate the results of comparative mutant analyses. The first goal of the present study was to use both approaches, the comparative analysis of various Arabidopsis lines with differing numbers of trichomes and the direct analysis of trichome waxes, to assess the cell type- specific composition of cuticular waxes. Our second goal was to compare gene expression patterns between trichomes and pavement cells, by analyzing all members of gene families known to be involved in wax biosynthesis in publicly available Arabidopsis transcriptome datasets. Ultimately, the chemical and transcriptomic data can be integrated to help shortlist gene candidates that may be involved in wax formation in trichomes.   3.2 MATERIALS AND METHODS  3.2.1 Plant material and growth conditions Seeds of wild type (Columbia-0 ecotype), trichome-free mutant gl1 (SALK_039478; (Alonso et al., 2016; Herman & Marks, 1989), and trichome-rich mutant cpc tcl1 etc1 etc3 (Wang et al., 2008) were treated and plants were grown as described above (refer to 2.2.1). Homozygous lines were confirmed by visual inspection for an absence or excess of leaf and stem trichomes. Mature leaves and stems at least one month old were harvested for trichome isolation or wax analyses.   57 3.2.2 Microscopy For cryo-SEM, segments of ~4 mm2 were sampled from the center of rosette leaves and segments ~4 mm long were sampled 2-3 cm from the stem base.  Samples were mounted onto copper stubs using PELCO water-based graphite paint (Ted Pella).  Stubs were transferred to an Emitech K1250 cryo-system (Emitech Inc.) and frozen at -120˚C.  Samples were loaded into a S4700 field emission SEM (Hitachi S4700) and held at -100˚C for 10 min to remove ice by sublimation prior to viewing.  Samples were imaged by the backscatter detector, using an accelerating voltage of 1.5 kV, a beam current of 15 mA, and a working distance of 12 mm.   3.2.3 Trichome isolation Trichome isolation from wild-type and cpc tcl1 etc1 etc3 plants largely followed Marks et al. (2008), with minor modifications: About 10 mature leaves were submerged in 15 ml 0.05 M EGTA solution and 50 mg silica beads (Silica Flash G60, Silicylce) in a 50 ml plastic tube. To remove the trichomes mechanically, the mixture was vortexed four times for 30 sec at high speed. Three tubes of the above were combined into one sample. The resulting mixture was strained through a wire mesh to remove plant organs, before filtering through a cell strainer (Corning, 100 µm, Nylon) to separate trichomes from the filtrate and other epidermal cells. The isolated trichomes were washed three times with sterile water, and trichomes were collected in a Petri dish by rinsing the cell strainer upside-down with water. Residual silica beads were removed by combining the trichome mixture with 40 ml water in a 50 ml tube. Silica beads settled faster, and the solution with floating trichomes was collected in a new 50 ml tube. This step was repeated until the majority of silica beads  58 had been removed. The trichome suspension was concentrated by centrifugation at low speed for 3 min, and residual water was removed. Collected trichomes were re-suspended in 1 ml of water and stored at 4°C until further processing.  3.2.4 Wax extraction and derivatization Leaf and stem waxes were extracted and derivatized as described above (as in 2.2.4) Trichome wax was extracted by organic solvent-water partitioning. To the aqueous suspension of trichomes an equal amount of CHCl3 was added, the two-phase system was vortexed twice for 3 min at high speed, then the CHCl3 phase was transferred into a new vial and concentrated under a stream of N2 gas. The resulting samples were derivatized as described in 2.2.4.  3.2.5 Wax identification and quantification Wax constituents were separated and quantified as describe above (see 2.2.5). GraphPad Prism v6.0 software was used for statistical analysis. Pair-wise comparisons were performed simultaneously on the entire dataset using Student’s t-tests (two-tailed, alpha = 0.05) and raw p values adjusted using a False Discovery Rate (FDR) equal to 5 %.  3.2.6 Calculation of relative expression levels of wax biosynthesis genes Transcriptomic data published by Marks et al. (2009) were used to compare the gene expression of wax biosynthesis genes in developing and mature trichomes as well as in pavement cells. The data for developing trichomes were derived from the glabra3-shapeshifter siamese (gl3-sst sim) double mutant (Marks et al., 2007), which lacks the  59 capability to progress through the normal stages of trichome development beyond stage four. In contrast, the transcriptomic data of mature trichomes captures only the genes expressed in the final developmental stage. The transcriptomic data acquired from shoots after the removal of trichomes were used as a proxy for pavement cell expression, thus ensuring that both epidermal cell types had been exposed to the same growth conditions and treatments. The resulting dataset (Marks et al., 2009) was queried for relative expression levels of wax biosynthesis genes. Transcript abundances were normalized to the reference gene UBC28 (At1g64230), which showed the most stable gene expression across the cell types.   3.3 RESULTS The current investigations aimed to assess whether trichomes are covered by cuticular waxes differing from those on pavement cells, and thus have (at least partially) autonomous wax biosynthesis machinery. To address this, first the stem wax compositions were compared between three Arabidopsis lines with drastically differing numbers of trichomes, followed by detailed analyses of the leaf waxes from the same lines. The results from these whole-organ wax analyses were then corroborated by comparative investigations into the waxes of isolated leaf trichomes. Finally, transcriptome data from developing and mature trichomes were compared to shoots lacking trichomes, taken as proxy for pavement cell composition, to help interpret the differences in wax composition between the epidermis cell types and to identify new wax biosynthesis gene candidates.   60 3.3.1 Stem waxes The trichome abundance of the Arabidopsis lines studied here was first assessed by cryo-SEM. The gl1 stem surface was found completely devoid of trichomes, while wild-type stems showed low trichome density, and cpc tcl1 etc1 etc3 mutant stems in contrast were covered with large numbers of trichomes (Figure 3.2A - C). The sizes and shapes of trichomes did not differ between the wild type and the cpc tcl1 etc1 etc3 mutant.   Figure 3.2: Comparison of trichome abundance on Arabidopsis gl1, wild type and cpc tcl1 etc1 etc3 stems and leaves using cryo-SEM. Trichome abundance on stems of (A) gl1, (B) wild type, (C) cpc tcl1 etc1 etc3. Trichome abundance on mature leaves of (D) gl1, (E) wild type, (F) cpc tcl1 etc1 etc3.   The cuticular wax extracted from the stems of gl1 mutants totaled 30 ± 4 µg cm-2. The large majority of compounds were identified as VLC fatty acid derivatives and triterpenoids, leaving only 7 % of the wax mixture (2.2 ± 0.2 µg cm-2) unidentified (Figure 3.3A). 1 mm 1 mm 1 mm A D E F 500 µm 500 µm 500 µm B C  61 Alkanes were by far the most abundant compound class (11 ± 2 µg cm-2), accompanied by ketone (8 ± 1 µg cm-2), secondary alcohols (3.2 ± 0.6 µg cm-2) and primary n-alcohols (3.1 ± 0.4 µg cm-2), and lesser quantities of esters (1.6 ± 0.7 µg cm-2), free fatty acids (0.2 ± 0.1 µg cm-2), aldehydes (0.2 ± 0.1 µg cm-2) and triterpenoids (0.2 ± 0.2 µg cm-2).  The chain length distributions within each compound class in gl1 stem wax were further examined (Figure 3.3B). Within the alkane fraction, compounds ranged from 27 to 33 carbons in length, with nonacosane (C29) dominating (94 ± 1 %). While odd-numbered alkane homologs were most abundant, trace quantities (<0.5 %) of octacosane (C28) and triacontane (C30) were also observed (data not shown). A single C29 ketone was present (nonacosan-15-one), and two isomeric C29 secondary alcohols (nonacosan-14-ol and nonacosan-15-ol) were found in approximately equal abundance. All other compound classes were dominated by even-numbered chain lengths. Only few free fatty acids and aldehydes were detected within both fractions as C28 and C30 homologs and present in approximately equal amounts. The primary n-alcohols were all un-branched, ranging in length from 24 to 32 carbons and peaking at C28. The alkyl ester chain lengths ranged from 40 to 50 carbons, peaking at C44, and were composed predominantly of hexadecanoic or octadecanoic acid plus the various complementary alcohols (data not shown). Finally, the triterpenoids β-amyrin and trinorlupeol constituted approximately 1 % each of the total wax.  The total wax coverage of wild-type stems (29 ± 8 µg cm-2) did not differ significantly from that of gl1. Accordingly, the compound class composition of the wild-type stem wax  62 (both absolute and relative quantities) equaled that of gl1 as well (Figure 3.3A). Moreover, no differences were observed in the relative chain length distributions of compounds within their respective classes between wild type and gl1 (Figure 3.3B). The dramatic increase in the number of trichomes in the cpc tcl1 etc1 etc3 mutant was not paralleled by a change in the total stem wax load (28 ± 12 µg cm-2), compared with the wild type or gl1. Furthermore, coverages of individual compound classes did not differ between the trichome-rich mutant and both other lines investigated (Figure 3.3A).  The overall chain length ranges within stem wax compound classes were the same for cpc tcl1 etc1 etc3 as for gl1 and wild type, but the relative distribution of homologs differed slightly and significantly greater coverages of C31 and C33 alkanes were observed in cpc tcl1 etc1 etc3 as compared to gl1 and wild type (Figure 3.3B). Similarly, a higher coverage of C32 primary n-alcohol was found in the trichome-rich mutant in comparison with gl1.   63  Figure 3.3: Stem wax composition of Arabidopsis mutants and wild type. A) Coverage of wax compound classes within the wax mixture on stems of trichome-free mutant (gl1), wild type, and trichome-rich mutant (cpc tcl1 etc1 etc3). B) Coverage of single compounds within each compound class of stem wax of trichome-free mutant, wild type, and trichome-rich mutant. x-Axis labels indicate the carbon chain lengths of compounds. Average values are given with standard deviations (n = 5). Asterisks indicate significant differences between the relative wax compositions determined by using the False Discovery Rate approach (* = FDR with 5 %).   3.3.2 Leaf waxes Cryo-SEM showed that gl1 leaves were lacking trichomes, while wild-type leaves exhibited low trichome density (ca. 75 trichomes per leaf) and cpc tcl1 etc1 etc3 leaves showed a three-fold increase in trichome numbers compared to wild-type leaves (Figure 3.2D – F). The total amount of wax extracted from the leaves of gl1 equalled 0.92 ± 0.15 µg cm-2. Six compound classes were identified within this mixture, leaving 0.19 ± 0.07 µg A B  64 cm-2 unidentified (Figure 3.4A). Alkanes were the most abundant fraction within the gl1 leaf wax mixture (0.33 ± 0.08 µg cm-2), accompanied by smaller amounts of alkenes (0.01 ± 0.002 µg cm-2), primary n-alcohols (0.15 ± 0.07 µg cm-2), branched alcohols (0.10 ± 0.03 µg cm-2), free fatty acids (0.16 ± 0.05 µg cm-2) and aldehydes (0.06 ± 0.02 µg cm-2).  The chain lengths within the gl1 leaf alkanes ranged from 27 to 37 carbons, with odd-numbered homologs predominating (Figure 3.4B) and only small quantities (2 %) of C30 and C32 alkanes present (data not shown). The overall alkane profile peaked at C31 (45 %) and C29 (35 %). Two alkene homologs C33 and C35 were detected, the former largely dominating. All other wax fractions contained mainly compounds with even numbers of carbons, accompanied by traces of compounds with odd carbon numbers. The (un-branched) n-alcohols exhibited a broad chain length distribution from C26 to C34, with a maximum at C30 (12 % of the class). In contrast, branched alcohols were only detected ranging from 30 to 34 carbons (including methyl branch carbons) and peaking at C32 (33 % of the class). Free fatty acids spanned the range from C24 to C34, with C26 acid predominating (40 % of the class) and all other homologs present at similar, lower levels. Finally, aldehyde homologs with 28 to 34 carbons were detected, with a maximum at C32 (36 % of the class).  The wax mixture of wild-type leaves had similar total coverage (1.1 ± 0.2 µg cm-2) as gl1, and also very similar absolute and relative quantities of compound classes (Figure 3.4A). Furthermore, the relative distributions of individual compounds within the diverse  65 compound classes of wild type leaf wax were also not significantly different from those in gl1 leaf wax (Figure 3.4B).  The total wax load of cpc tcl1 etc1 etc3 leaves (1.4 ± 0.4 µg cm-2) was similar to that of wild type or gl1, and also the compound class distribution within the leaf wax of the trichome-rich mutant resembled those of the other two lines (Figure 3.4A). However, the coverage of alkenes in the trichome-rich cpc tcl1 etc1 etc3 mutant was found to be significantly higher than in the gl1 and wild type leaf wax mixtures.  Significant differences were found between the chain length profiles within the alkane, alkene and primary n-alcohol fractions of cpc tcl1 etc1 etc3 leaf wax and those of the other two lines (Figure 3.4B). The C33, C35 and C37 alkanes were found at significantly higher levels in the trichome-rich mutant than in the trichome-free mutant or in the wild type. Similarly, C35 alkene accumulated to significantly higher amounts in the trichome-rich mutant compared to the trichome-free mutant, and C37 alkene accumulated to significantly higher coverage in the trichome-rich mutant compared to both other lines. Finally, the C32 and C34 n-alcohols were also present in higher coverages in the leaf wax of cpc tcl1 etc1 etc3 than in gl1 or wild type. In contrast to the alkanes, alkenes and primary n-alcohols, the chain length profiles of the aldehydes, branched alcohols and fatty acids did not differ between the leaf wax of the trichome-rich cpc tcl1 etc1 etc3 mutant and the other two lines.  66  Figure 3.4: Leaf wax composition of Arabidopsis mutants and wild type. A) Coverage of wax compound classes within the wax mixture on leaves of trichome-free mutant (gl1), wild type, and trichome-rich mutant (cpc tcl1 etc1 etc3). B) Coverage of single compounds within each compound class of leaf wax of trichome-free mutant, wild type, and trichome-rich mutant. Average values are given with standard deviations (n = 5). x-Axis labels indicate the carbon chain lengths of compounds. Asterisks indicate discovery of significant differences between coverages based on Student’s t-test (* = p < 0.05).   3.3.3 Trichome waxes To probe the epidermis cell type-specific wax composition further, I isolated trichomes and extracted their surface wax with chloroform. Preliminary experiments had shown that the trichome quantities necessary for GC analyses of waxes could be isolated from leaves of Arabidopsis wild type and the trichome-rich cpc tcl1 etc1 etc3 mutant. Trichomes could also be prepared from stems of the same lines, but were found to be contaminated with A B  67 epicuticular wax crystals from neighboring pavement cells, preventing the selective analysis of trichome wax in the scope of the current work.  Wax from wild-type leaf trichomes comprised only four compound classes (Figure 3.5A). Alkanes dominated the mixture (64 ± 9 %), while alkenes contributed 3 ± 1 %, un-branched primary n-alcohols 6 ± 3 % and branched primary alcohols 0.6 ± 0.4 %. Due to small sample amounts available, 27 ± 9 % of the wax could not be identified. Within the alkane fraction, chain lengths ranged from 27 to 37 carbons (Figure 3.5B), and the C31 (44 %) and C33 (32 %) homologs largely dominated. The primary n-alcohol homologs from 26 to 34 carbons in length were detected, with the C32 alcohol most abundant (31 %). Branched alcohols with total carbon numbers C30, C32 and C34 were found, but homolog distribution within this fraction could not be accurately quantified in all samples due to low amounts.  The wax extracted from isolated trichomes of cpc tcl1 etc1 etc3 contained the same compound classes in similar quantities as wild type trichomes (Figure 3.5A). Only the alkenes were found in higher amounts in the trichome wax of the mutant compared to wild-type trichomes. The chain length distributions were also similar between wax mixtures from both lines (Figure 3.5B), albeit with slightly lower percentages of C27, C29 and C31 alkanes on the trichome-rich mutant, and concomitantly higher amounts of C33, C35 and C37 alkanes. Of note, relatively minor quantities of glycerolipids, hexadecanoic and octadecanoic acids were detected in any of the lipid samples extracted from isolated  68 trichomes. This confirmed that membrane and other intracellular lipids were not extracted, showing that the methods used here were specific to cuticular waxes.  Figure 3.5: Leaf trichome wax composition of Arabidopsis trichome-rich mutant and wild type. A) Percentages of wax compound classes within the wax mixture on leaf trichomes of wild type and trichome–rich mutant (cpc tcl1 etc1 etc3) (n = 5). B) Percentages of single compounds within each compound class of leaf trichome wax of wild type and trichome-rich mutant (cpc tcl1 etc1 etc3). x-Axis labels indicate the carbon chain lengths of compounds. Average values are given with standard deviations (n = 5). Asterisks indicate discovery of significant differences between coverages based on Student’s t-test (* = p < 0.05). A B  69 3.3.4 Gene expression patterns The trichome transcriptome dataset of Marks et al. (2009) was analyzed to identify candidate genes for cell type-specific wax formation. To this end, expression signals of all genes were normalized within each cell type to Ubiquitin C28 (UBC28), a reference gene found to have particularly stable expression levels across the entire dataset. The reference data for gene expression in whole shoots (Marks et al., 2009) was included in the comparisons as a proxy for pavement cell expression. The Arabidopsis shoot epidermis consists nearly entirely of pavement cells, and only very few guard cells or trichomes. Overall, genes known to encode wax-forming enzymes were abundantly expressed, along with diverse other members of the same gene families that have not yet (or only partially) been characterized. Among the wax biosynthesis genes (or gene families), those involved in fatty acid elongation or in head group modification can be distinguished respectively determining chain length profiles or compound class distributions within the product wax mixtures.  The ketoacyl-CoA synthetases (KCSs) are known to confer chain length specificity to the fatty acid elongase (FAE) complexes involved in wax biosynthesis. Of the 21 KCS homologs in the Arabidopsis genome, 16 were found expressed above the detection limit in the trichome transcriptome (Figure 3.6A). Among them, KCS3, KCS4, KCS11, KCS12, KCS13/HIC, KCS14 and KCS19 showed relatively low expression signals, and no major differences between the developing and mature trichomes as well as pavement cells. Two genes (KCS2/DAISY and KCS16) were also weakly expressed, albeit only in developing trichomes. Four genes (KCS1, KCS5/CER60, KCS8 and KCS10) had high expression  70 signals, in developing trichomes in particular. The three remaining genes (KCS6/CER6, KCS9 and KCS20) were characterized by high expression signals in pavement cells and trichomes.  To put the KCS expression levels into context, the transcript levels of the three other genes encoding FAE enzymes had to be assessed (Figure 3.6A). One ketoacyl-CoA reductase gene (KCR1) was found expressed highly in pavement cells and even more in developing trichomes. Interestingly, its homolog (KCR2) was expressed at lower level only in developing trichomes, but not in pavement cells. Both the β-hydroxyacyl-CoA dehydratase gene (HCD/PAS2) and the enoyl-CoA reductase gene (ECR/CER10) were expressed in all epidermis cells at levels intermediate between those of the various KCSs.   Transcripts encoding other proteins implicated in chain length specificity of the FAE complex, CER2 and its homologs, were also detected in trichomes (Figure 3.6B). CER2 and CER2-like2 were expressed at similar, low levels in pavement cells and in developing or mature trichomes. Expression signals of CER2-like1 (CER26) were particularly high in developing trichomes, in stark contrast to mature trichomes or pavement cells.  Since long-chain acyl-CoA synthases have been implicated in wax formation, the expression levels of all members of this gene family were compared between trichomes and pavement cells as well (Figure 3.6B). Three of them were expressed mainly in trichomes, either in both developing and mature trichomes (LACS1), or in developing (LACS2) or mature trichomes only (LACS3). All other LACS genes were expressed at  71 similar levels in trichomes and pavement cells, from relatively low (LACS7 and LACS9) to intermediate (LACS8) and fairly high levels (LACS4 and LACS6).  Finally, the expression levels of genes encoding modifying enzymes (responsible for compound class distribution) could be assessed. Firstly, substantial concentrations of transcripts were detected for those genes encoding the two major enzymes involved in alkane formation, CER3 and CER1. In particular, CER3 was found highly expressed in developing trichomes, and at lower levels in mature trichomes and pavement cells (Figure 3.6B). Transcripts encoding CER1 were detected at intermediate levels in developing trichomes, mature trichomes and pavement cells, and hence in much lower amounts than CER3 in all three tissues. The genes encoding both the homologous protein (CER1-like1) and the mid-chain hydroxylase (MAH1) were expressed solely in developing trichomes, at moderate levels. Secondly, the expression levels of enzymes catalyzing the formation of primary n-alcohols and their esterification could also be compared between trichomes and pavement cells (Figure 3.6B). The gene of the fatty acyl-CoA reductase known to form wax primary n-alcohols (FAR3/CER4) was expressed at low levels in all three tissues tested. Two homologous genes, FAR1 and FAR4, were expressed weakly in trichomes, but not in pavement cells. Finally, expression signals of the wax ester synthase involved in formation of VLC alkyl-acyl esters (WSD1) were found at intermediate levels in developing trichomes, in only very low amounts in mature trichomes and not in pavement cells.  72  Figure 3.6: Differential expression of wax biosynthesis genes in Arabidopsis epidermis cell types. Transcriptome data (Marks et al., 2009) for shoots (and thus largely pavement cells) as well as developing and mature trichomes are shown relative to expression levels of UBC28. A) Genes involved in fatty acid elongation. B) Genes involved in head group modification. Average values are given with standard deviations (n = 3).   3.4 DISCUSSION In order to probe the cell type specificity of wax compositions, waxes from a trichome-free mutant (gl1), wild type, and a trichome-rich quadruple mutant (cpc tcl1 etc1 etc3) were examined as well as waxes on trichomes isolated from wild type and cpc tcl1 etc1 etc3. Both on stems and on leaves, cell type-specific differences in wax compositions occurred at two levels, namely for compound classes and compound chain length distributions. The compound class differences can be related to cell-specific expression patterns of those genes known to be involved in fatty acid modification, as A B  73 their corresponding gene products establish the compound class distributions. On the other hand, the chain length profiles of trichome and pavement cell waxes can be related to the relative expression profiles of those genes involved in fatty acid elongation.  3.4.1 Compound class differences between trichomes and pavement cells Wax from non-trichome epidermal cells was investigated directly using the trichome-free mutant gl1, and the stem wax of this line was found to contain eight classes of compounds, each contributing up to 40 % to the mixture. In contrast to previous reports, none of the compound classes were found to accumulate to levels different from wild type (Xia et al., 2010; Reisberg et al., 2012). Similarly, the coverages of the different compound classes in the stem wax of the trichome-rich mutant cpc tcl1 etc1 etc3 did not differ from the two other lines.  In leaves, the comparison of wax compositions from the trichome-rich and the wild type and trichome-free mutant revealed an increase in alkenes in cpc tcl1 etc1 etc3 compared to the wild type or gl1, suggesting that trichomes have more alkenes than pavement cells. This conclusion is partially confirmed by the analysis of wax from isolated leaf trichomes (cpc tcl1 etc1 etc3 and wild type). As compared to the wax on gl1, and hence on pavement cells, wax from isolated trichomes contained mainly alkanes accompanied by alkenes, primary n-alcohols, and trace amounts of branched primary alcohols. Leaf pavement cells show a much greater diversity in their wax compound classes than leaf trichomes, and a corresponding decrease in alkane dominance.  It must be noted that the original cpc mutant was generated in the Wassilewskija ecotype. Subsequent crossings with the other three mutants (Columbia-0 ecotypes) diminished but likely did not eliminate the genetic contributions of this parent towards wax biosynthesis. Moreover,  74 mutations causing trichome density reductions frequently also affect root hair numbers, and root hairs constitute only 3 % of epidermal cells in cpc tcl1 etc1 etc3 as compared to 42 % for wild type (Wang et al., 2008). A decrease in root hairs may generate drought stress, which in turn may influence wax composition on aboveground organs. For these two reasons, trichomes from both wild-type and cpc tcl1 etc1 etc3 leaves were isolated in sufficient quantity to enable direct comparative analyses of their waxes. Interestingly, cpc tcl1 etc1 etc3 trichome wax showed an increase of longer chain lengths (C33, C35, C37) compared to the wild-type trichome wax. This shift in wax composition might be due to genetic differences between the ecotype backgrounds involved, or it may reflect physiological responses due changes in root architecture and thus decreased water uptake efficiency.  The compound class differences between trichomes and pavement cells might lead to lateral diffusion along the surface of these cells. Is there exchange of surface wax between trichomes and neighboring pavement cells that might lead to mitigation of compositional differences? To address possible lateral diffusion processes, the mobility of compounds within wax mixtures can be assessed. The diffusion coefficients of C26/27 acyl compounds (regardless of class) within wax are approximately 10-21 m2 s-1 (Schreiber, 2006). After 30 days (the approximate time from trichome initiation to wax analysis), and assuming no additional wax biosynthesis, the compound would reach a concentration of e-1 times the initial amount at a distance of only circa 0.1 µm. Increasing compound length by ten carbons would decrease the distance to approximately 0.01 µm. Considering that trichome stalks are around 100 µm in length, it can be concluded that diffusion is not fast enough to significantly alter the trichome wax composition. Moreover, if diffusion were to occur, it would only lessen the observed differences instead of causing or enhancing them.  75 Consequently, the detected wax compositions reflect the minimum differences between epidermal cell types.  Overall, the comparisons of bulk waxes from the three lines as well as direct leaf trichome wax analyses mutually confirm each other. All our results hence suggest that the wax mixtures on both stem and leaf trichomes contain fewer compound classes and increased amounts of compounds with especially long chain lengths compared to stem and leaf pavement waxes, respectively.  3.4.2 Cell-specific expression patterns of genes involved in fatty acid modification Since trichome and pavement waxes have distinct compositions, it can be concluded that trichomes must be autonomous in wax biosynthesis, and that they likely contain the entire suite of enzymes required for wax elongation and compound class elaboration. The diverse acyl derivatives, and hence compound classes, are known to arise via two parallel branch pathways, one generating primary n-alcohols and their esters, and the other generating aldehydes, alkanes, secondary alcohols and ketones (Yeats & Rose, 2013). Compound classes of both pathways had concentrations differing between Arabidopsis pavement cells and trichomes, and thus genes encoding enzymes on both branch pathways may be differentially expressed in both epidermis cell types.  Our finding that trichome waxes have high alkane concentrations is matched by relatively high trichome expression of the two genes, CER3 and CER1, encoding enzymes of the alkane-forming pathway (Bernard et al., 2012; Chen et al., 2003). Both genes are also expressed at moderate levels in pavement cells, in line with the assumption that the alkane-forming machinery of both epidermis cell types is thus largely identical.   76 While substantial amounts of alkanes are exported to the epidermis surface, a portion of them may also serve as substrates for modification into secondary alcohols and ketones (Wen & Jetter, 2009). The mid-chain hydroxylase, MAH1, is the single enzyme responsible for catalyzing the formation of both these compound classes, and it had been found expressed only in pavement cells but not in trichomes (Greer et al., 2007). On the one hand, this is confirmed by our chemical analyses, showing that trichome waxes are lacking secondary alcohols and ketones. On the other hand, our transcriptome analyses contradicted the previous report to some degree, revealing low expression of MAH1 in developing but not mature trichomes.  The presence of primary n-alcohols in stem and leaf pavement cell wax likely reflects the activity of one or more fatty acyl-CoA reductases (FARs). However, only one wax-specific FAR has been identified to date (FAR3/CER4), and previous evidence using a 2.1 kb CER4 promoter::GUS fusion suggested that it is expressed in leaf trichomes but not in pavement cells of rosette leaves (Rowland et al., 2006). Transcriptome analysis now suggest that CER4 is expressed in rosette leaf pavement cells, albeit at low level, and that it is the only FAR present there. Expression patterns of plant genes can involve cis elements within introns and the 3’-noncoding region (Taylor, 1997), explaining why the original promoter::GUS analysis may not have revealed the entire CER4 expression pattern.  Instead, it seems likely that this enzyme is responsible for n-alcohol formation in the leaf pavement cells. In trichomes, on the other hand, CER4 is expressed together with FAR1 and FAR4. Of the latter, FAR1 had been found expressed in leaves before, but not FAR4 (Domergue et al., 2010). The FAR1 and FAR4 proteins had been shown to form mainly C22 and C20 primary n-alcohols, respectively, and little or no C24 and C26 primary n-alcohols (Domergue et al., 2010). However, their activities on longer-chain substrates have not yet been fully explored,  77 and thus it cannot be ruled out that FAR1 and/or FAR4 may still be involved in the formation of C32 and C34 primary n-alcohols in trichomes. While awaiting further characterization of these two FARs, it nonetheless seems plausible that CER4 is the major enzyme also responsible for the formation of the wax alcohols up to very high chain lengths in trichomes.  In this context it is noteworthy that branched alcohols, although present in equal amounts as un-branched alcohols on leaf pavement cells, were detected only in trace amounts on trichomes. This finding suggests that branched-chain acyl-CoA precursors are formed in much lower amounts in the trichomes than in pavement cells. It will be interesting to test this hypothesis further, and to exploit the difference between both cell types for investigations into the mechanisms of branched-chain elongation and corresponding head group modification.  The final product of the n-alcohol biosynthesis pathway, alkyl esters, accumulated to slightly lower concentrations in the trichome-rich mutant than in gl1, suggesting that this compound class may also be restricted to pavement cells. This was confirmed by our trichome wax analyses, where esters could not be detected. However, the chemical data are in contrast to gene transcript analyses, showing the ester-forming enzyme WSD1 (Li et al., 2008) expressed at low levels in developing trichomes but not in pavement cells.  3.4.3 Chain length profiles of trichome and pavement cell waxes To assess the differences in chain length profiles between the trichome-rich and trichome-free mutants, the ratios of relative percentages of each compound within its class were calculated. To better visualize the results, only ratios ≥1 are plotted, either for the ratio of cpc tcl1 etc1 etc3 to gl1, or for the inverse. In this analysis, bars pointing downwards indicate higher percentages in non- 78 trichome epidermal cells, while upward bars reflect higher percentages in trichome wax. Within the stem waxes, compounds with shorter chain lengths were enriched on trichome-free surfaces, while longer compounds were relatively more abundant on trichome-rich surfaces (Figure 3.7A). The change-over between more pavement-enriched to more trichome-enriched compounds occurred between C28 and C30 free acids and aldehydes, C29 and C31 alkanes, and C30 and C32 primary n-alcohols. The corresponding shift was more broadly distributed between the C42 and C48 ester homologs, where the C42 and C44 esters are known to contain mainly C26 and C28 n-alcohols (linked to C16 fatty acid), respectively, and the esters from C46 onwards mainly C>30 n-alcohols (linked to various acids) (Lai et al., 2007). I conclude that the trichome wax was enriched in esters containing C32 n-alcohol, and that the change-over between more pavement-enriched to more trichome-enriched homologs in the esters closely matched that of the free n-alcohols.  For the leaf waxes, the ratios between cell types showed similar trends within compound classes as in the stems (Figure 3.7B). Specifically, the change from more pavement-enriched to more trichome-enriched compounds appeared between C28 and C30 free fatty acids and straight-chain n-alcohols, C30 and C32 aldehydes, and C31 and C33 alkanes. The branched alcohols had ratios between both lines near 1:1 for all chain lengths.  In summary, both stem and leaf trichome wax compounds displayed an increase in average chain length as compared to pavement cell wax. In stem wax, compounds with more than 30 carbons accumulated to higher percentages in cpc tcl1 etc1 etc3 wax than in gl1. In leaf waxes, compounds longer than 32 carbons showed a clear increase in the trichome-rich mutant over gl1. These findings are matched by analyses of the isolated leaf trichome waxes that showed an increase in  79 longer chain lengths relative to pavement cells (as seen in gl1 leaf wax; Figure 3.7C). The distribution of C29 – C35 alkanes and of C26 – C32 alcohols in the waxes of isolated trichomes paralleled the ratios found by Ebert et al. (2010) in a trichome metabolomics study.   Figure 3.7: Comparisons of compound amounts between Arabidopsis wax sample types. Ratios of percentages for single compounds within respective compound classes are given A) between stem waxes of trichome-rich (cpc tcl1 etc1 etc3) and trichome-free (gl1) Arabidopsis lines, B) between leaf waxes of trichome-rich (cpc tcl1 etc1 etc3) and trichome-free (gl1) Arabidopsis lines, and C) between trichome and pavement cell waxes. Ratios for each compound were calculated from data given in Figure 3.3 for stems, Figure 3.4 for leaves and Figure 3.5 for trichomes. Note that ratios plotted in the top half of each panel denote compounds accumulating preferentially in trichomes (left y-axes), while inverse ratios plotted in the bottom half of each panel refer to those compounds accumulating less in trichomes than in pavement cells (right y-axes). Thus, all ratios are shown as values  > 1. x-Axis labels indicate the carbon numbers of compounds.  B A C  80 3.4.4 Relative expression profiles of genes involved in fatty acid elongation The higher concentration of longer homologs in most compound classes of trichome waxes, relative to pavement cells, points to substantial differences in the fatty acid elongation machineries of both cell types. Most likely, trichomes may contain FAE complexes (Haslam & Kunst, 2013) that can carry out additional elongation rounds rather than releasing their acyl-CoA products into the alkane- and alcohol-forming pathways. Alternatively, but much less likely, chain extension to these extremely long chain lengths may by catalyzed by alternative elongases such as the four ELO-like enzymes encoded in the Arabidopsis genome (Dunn et al., 2004).  Chain length specificity of FAE complexes is known to be largely conferred by the keto-acyl-CoA synthetase (KCS) enzyme involved (Lassner et al., 1996; Millar & Kunst, 1997). Thus, it seems plausible that trichomes may possess one or more KCSs able to elongate acyl-CoA substrates to chain lengths of C34 – C38. In addition, other KCSs will be crucial for earlier elongation rounds, and these may either be identical to the KCS complement in pavement cells or differ from it as well. Overall, 16 KCS genes were found expressed in trichomes, and based on their relative transcript levels roles in trichome wax formation can now be hypothesized for some of them.  The three KCS genes that were highly expressed in both pavement cells and trichomes (KCS6/CER6, KCS9 and KCS20) had all been characterized, and shown to be involved in chain length elongation from C24 to C28, C22 to C24, and C20 to C22, respectively (Haslam et al., 2012; Kim et al., 2013; Lee et al., 2009). However, while loss-of-function mutants cer6 were found to have drastically reduced wax amounts (Jenks et al., 1995; Millar et al., 1999; Fiebig et al., 2000), those of the other two genes had relatively mild phenotypes (Kim et al., 2013) pointing to substantial redundancy of biochemical functions. Redundancy was confirmed for KCS20 by  81 kcs20/kcs2 double mutants analysis, which resulted in a severe phenotype compared to kcs20 or ksc2 single mutant analysis (Lee at al., 2009). Our results now suggest that this redundancy likely occurs within all epidermal cell types, and thus on the single-cell level. KCS1, KCS5, KCS8 and KCS10 were expressed most strongly in trichome cells, and hence all four genes may well be involved in trichome wax formation. Of these, KCS1 is known to participate in the formation of precursors for VLC wax compounds (Todd et al., 1999), possibly by catalyzing elongation steps from C18 to C20 and beyond (Blacklock & Jaworski, 2006). Our transcriptome analysis showed that KCS1 is highly expressed in trichomes and moderately in pavement cells, suggesting that it may be involved in early steps of elongation in both cell types. Next, KCS5 shares high sequence similarity with CER6 (Fiebig et al., 2000) and both enzymes have similar substrate and product profiles (Trenkamp et al., 2004; Tresch et al., 2012). Therefore, it seems plausible that both are important for elongation from C26 to C30 in trichomes (Fiebig et al., 2000; Millar et al., 1999; Weidenhamer et al., 1993). However, such a role of KCS5 has not been reported yet, possibly because kcs5 mutant phenotypes are restricted to chain length shifts only in their trichome waxes, and because the presence of CER6 in the prevalent pavement cells might mask such phenotypes. KCS8, on the other hand, has also not been characterized to date, and it will be interesting to study its involvement in trichome wax formation as well. Finally, KCS10 was shown to be involved in plant surface formation (Lolle et al., 1997; Yephremov et al., 1999; Pruitt et al., 2000), however its role in the elongation of aliphatic compounds is not well understood. Interestingly, Yephremov et al. (1999) observed defects in trichome differentiation in kcs10 mutants, leading to reduced numbers of leaf trichomes. It is currently not clear whether KCS10 is involved in cuticular wax biosynthesis, possibly in trichomes.  82 Finally, nine other KCSs were found expressed at low levels in trichomes. Seven of them, KCS3, KCS4, KCS11-14 and KCS19, were equally expressed in trichomes and pavement cells, making it unlikely that any of them is involved in formation of either the abundant C26 – C34 compounds or the extraordinary C35 – C38 compounds present in trichome wax. The two remaining genes, KCS2/DAISY and KCS16, were found expressed in developing trichomes only, albeit at very low levels. KCS2/DAISY is known to be involved in elongation to C22 fatty acid in suberin formation (Franke et al., 2009; Lee et al., 2009), and is thus likely not responsible for the formation of the especially long (saturated) trichome wax compounds. On the other hand, KCS16 has not been characterized to date, and may thus serve as another candidate for a trichome-specific elongation enzyme. At its low expression level, it seems plausible that this KCS would be involved in formation of the small, but significant amounts of C36 – C38 compounds found in trichome wax.  Interestingly, the alkenes found in the waxes extracted from whole leaves or isolated trichomes showed the same trend as alkanes, with a shift towards the C35 homolog in trichomes. It is therefore plausible that the alkenes are formed on a pathway largely paralleling that of alkanes, also involving special KCS enzymes for the formation of longer homologs. In this context it should be noted that KCS18/FAE1, the only KCS previously shown to elongate unsaturated fatty acyl precursors during seed oil formation (Todd et al., 1999), was not expressed in leaf trichomes. It is therefore not clear whether another KCS, with similar preference for acyls containing C=C double bonds, might be involved in elongation of unsaturated acyls to generate precursors en route to the alkenes (of trichomes and pavement cells). Alternatively, one or more of the KCSs thought to be involved in elongation of saturated substrates might also accept unsaturated substrates.   83 The other three enzymes associated with KCSs in the FAE complex (Haslam & Kunst, 2013; Yeats & Rose, 2013) all had intermediate transcript levels. The gene encoding the KCR homolog known to be involved in wax biosynthesis, KCR1 (Beaudoin et al., 2009) was expressed highly in pavement cells as well as trichomes. In contrast, KCR2 was found weakly expressed only in developing trichomes, explaining why corresponding mutants exhibited no bulk wax phenotype (Beaudoin et al., 2009). Nonetheless, it is possible, based on our results, that KCR2 is involved in trichome wax formation. Interestingly, the other two FAE enzymes, HCD/PAS2 and ECR/CER10 (Bach et al., 2008; Zheng et al., 2005), had only moderate transcript levels in both pavement cells and trichomes. They are known to be the sole genes encoding the dehydratase and enoyl-CoA reductase, respectively, required in each round of elongation irrespective of KCSs and chain lengths involved (Bach et al., 2008; Zheng et al., 2005), and should thus be expressed at levels similar to the KCSs and KCR. The relatively low expression levels of HCD and ECR then point either to lower transcript or protein turnover.  Also worth noting are the relatively low expression levels of CER2 and its homologs CER2-like1 and -2 in pavement cells, encoding three other proteins thought to be associated with the FAE complex and thus controlling overall chain length profiles (Haslam et al., 2012; Pascal et al., 2013: Haslam et al., 2015). Presence of CER2 enables the formation of compounds up to C30, whereas the CER2-likes in combination with FAE can yield up to C34 (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015). In this context it is interesting that CER2-like1/CER26 was very highly expressed in trichomes, much more than CER2, leading us to hypothesize that CER26 may enable the formation of compounds up to C38 if associated with a special, trichome-specific KCS in its FAE.  84 Finally, some long-chain acyl-CoA synthetases (LACSs) have also been implicated in wax biosynthesis, with mutant wax phenotypes described for lacs1, lacs2 and lacs4 (Jessen et al., 2011; Lü et al., 2009). Interestingly, of these only LACS4 was highly expressed in pavement cells, whereas the other two showed highest transcript levels in trichomes, suggesting a partial segregation of the homologs into different epidermis cell types. To date, only LACS1, LACS2 and LACS8 have been localized to the ER (Pulsifer et al., 2012; Zhao et al., 2010), and thus the sub-cellular compartment of wax biosynthesis, whereas LACS6 and LACS7 were found in the peroxisomes (Radakovits et al., 2010) and LACS9 in the plastids (Zhao et al., 2010). A sunflower LACS8 homolog showed substrate specificity towards 16:0 and 18:2 fatty acids, suggesting a role in oil seed or phospholipid rather than wax biosynthesis (Aznar-Moreno et al., 2014). LACS1, LACS2 and LACS3 are able to facilitate the uptake of C16 to C30 fatty acids into yeast (Pulsifer et al., 2011). Taken together with our expression data, this leaves LACS3 as the best candidate for further homologs to be involved in wax biosynthesis, either in the trichomes or the pavement cells.  In summary, the data suggest that trichomes have autonomous wax biosynthesis, and that their biosynthetic machinery comprises enzymes similar to those in pavement cells. However, trichomes may contain (at least some) unique wax biosynthesis enzymes, and regulation of individual pathways may differ from pavement cells. In particular, the higher abundance of extremely long fatty acid derivatives found in trichome waxes are likely due to differential expression of elongating enzymes (KCSs). Currently, nothing is known about elongation to the longest chain lengths observed in leaf trichomes (C35-C37), but they likely result from one or more of the uncharacterized KCSs.  85  Characterization of β-ketoacyl-CoA synthase 16 (KCS16), a Chapter 4:condensing enzyme involved in biosynthesis of C36 and C38 acyl-CoAs for cuticular wax formation in Arabidopsis thaliana leaf trichomes and pavement cells  4.1 INTRODUCTION Not only cuticular waxes, but also various lipids, have hydrocarbon chains derived from very-long-chain fatty acids (VLCFAs). Among them, sphingolipids are ubiquitously present in the membranes of most organisms, typically with chain lengths up to C24. In addition, some plants have C20 – C24 acyls in their seed storage triacylglycerols. Those plant VLCFAs are formed by ER-localized fatty acid elongase (FAE) complexes, comprising four enzymes catalyzing a cyclic reaction sequence that effects the elongation of an acyl chain by two carbons. Repeat cycles, carried out by the same or different FAEs, lead to a range of acyl chain lengths. The FAE cycle starts by joining the incoming acyl substrate with a C2 unit derived from malonate, in a condensation catalyzed by a β-ketoacyl-CoA synthase (KCS) enzyme (Figure 4.1) (Joubès et al., 2008). In the ensuing reactions, a ketoacyl-CoA reductase (KCR) first reduces the resulting β-keto group into a hydroxyl function, a hydroxyacyl-CoA dehydratase (HCD) then forms the corresponding enoyl-CoA, and an enoyl-CoA reductase (ECR) finally saturates the C=C double bond to form the elongated acyl product (Figure 4.1) (Beaudoin et al., 2002; Domergue et al., 2000; Han et al., 2002; Millar & Kunst, 1997; Paul et al., 2006). In Arabidopsis, all FAE complexes share the same KCR, HCD and ECR enzymes, but their KCS components may differ and are thought to determine substrate and product chain length specificities of the overall complex (Millar & Kunst, 1997; Zheng et al., 2005). 86  Figure 4.1: Elongation of acyl-CoAs by the FAE complex (Haslam & Kunst, 2013).   The genomes of two species of green algae, Chlamydomonas reinhardtii (Merchant et al., 2007) and Volvox cartii (Prochnik et al., 2010), contain only one KCS gene, whereas the liverwort Marchantia polymorpha has three KCS homologs (Kajikawa et al., 2003). In contrast, the genome of the moss Physcomitrella patens contains 15 KCS genes, similar to those of gymnosperms; angiosperms tend to have even bigger numbers of KCSs (Guo et al., 2015). For example, the Arabidopsis genome encodes 21 KCSs that can be divided into eight subfamilies based on phylogenetic analysis. The genes are distributed unevenly over the Arabidopsis genome, across all chromosomes. Some of the Arabidopsis KCS genes lack introns, similar to most algal genes (Guo et al., 2015). All the current genome information taken together suggests that the earliest plant lineages had only one ancestral KCS gene, and that repeated duplication of KCS genes occurred during the evolution of first land plants and then continued, at lower rates,  87 throughout the evolution of vascular plants. Interestingly, the wide-spread occurrence of relatively large KCS gene families throughout the plant kingdom is contrasting with the presence of single-copy KCSs in all other eukaryotes investigated to date (Guo et al., 2015).  Plant KCS proteins comprise a 3-ketoacyl-synthase-III domain, located at the C-terminus, and a type-III polyketide synthase-like domain at the N-terminus (Guo et al., 2015). All KCS enzymes share the same four-layered structure α1-β2-α3-α4. The first α1 fold consists of one or two α-helices, which may act as membrane anchors in almost, if not all, KCS. The other folds, β2-α3-α4, comprise the soluble part of the KCS enzyme, and the active-site residues located in the α3 fold are thought to form a cavity running through the enzyme. This cavity can be closed by a lid provided by the α4 fold (Joubès et al., 2008), and it contains a catalytic triad of the amino acids Cys258, His425 and Asn458 in almost all KCSs. While regions containing the catalytic residues are highly conserved across the KCS gene family, the N-terminal domains (including the membrane-spanning domain) share little identity and do not contain residues relevant for catalysis (Blacklock & Jaworski, 2006). Accordingly, substitution of an amino acid in this region of a chimeric Arabidopsis thaliana/Brassica napus FAE1 altered the substrate specificity but not the activity of the enzyme (Blacklock & Jaworski, 2002).  Over the years, the biochemical activities of many KCS enzymes were tested, and associated with their biological functions in some cases. Heterologous expression in yeast (Saccharomyces cerevisiae) was used to test the substrate activity of the Arabidopsis KCSs, and nine out of the 21 KCS family members (KCS1, KCS2, KCS5, KCS6, KCS9, KCS11, KCS17, KCS18/FAE1 and KCS20) were found to be active in this heterologous system (Blacklock & Jaworski, 2006;  88 Franke et al., 2009; Lee et al., 2009; Paul et al., 2006; Trenkamp et al., 2004; Tresch et al., 2012). The first of these KCS enzymes to be characterized was FAE1/KCS18, involved in elongating acyl-CoAs to C20, C20:1, and C22:1 acyl-CoAs in seeds (Rossak et al., 2001). Several FAE1 homologs have been found in other species, e.g. Brassica napus (Fourmann et al., 1998), Simmondsia chinensis (Lassner et al., 1996) and also in an alga, Dunaliella salina (Azachi et al., 2002). The KCS11 enzyme showed in vitro activity towards saturated and unsaturated acyl-CoA substrates ranging from C16 to C20 (Blacklock & Jaworski, 2006) and was found expressed in all Arabidopsis organs (Joubès et al., 2008), while KCS17 accepted C16 to C22 saturated precursors (Blacklock & Jaworski, 2006) and was expressed in siliques only (Joubès et al., 2008). However, the chain lengths of the elongation products have not yet been determined and the biological function of these two KCSs can therefore not be assigned.  Interestingly, several of Arabidopsis kcs loss-of-function mutants showed cuticle alterations, highlighting the importance of KCSs in the formation of cuticular waxes. In particular, KCS2/DAISY and KCS20 are thought to have redundant functions, elongating VLCFA from C20 to C22, to provide substrates for the biosynthesis of cuticular waxes in several organs, as well as for suberin formation in roots (Lee & Suh, 2015). Based on mutant phenotype analysis, KCS9 is involved in elongating C22 acyl substrate to C24 product (Kim et al., 2013), thus contributing to the formation of wax, suberin and sphingolipids, and kcs1 mutants showed a decrease in wax fatty acids and primary alcohols compared to wild type (Todd et al., 1999). Mutation of kcs10/fiddlehead (fdh) caused organ fusions and male sterility, due to alteration in the epidermis, causing impaired contact mediated cell interactions (Lolle et al., 1997; Pruitt et al., 2000; Yephremov et al., 1999). Interestingly, KCS10/FIDDLEHEAD did not show elongation activity  89 towards saturated and unsaturated C20 to C30 substrates (Trenkamp et al., 2004). Therefore its function remains unknown.  KCS6/CER6 has been recognized as a pivotal enzyme for wax biosynthesis in Arabidopsis, since ksc6/cer6 loss-of-function mutants exhibited a severe decrease of cuticular wax compounds in leaves and stems as well as organ fusions and male sterility due to inhibited pollen hydration (Hülskamp et al., 1995; Preuss et al., 1993). KCS6/CER6 has high amino acid sequence similarity with KCS5/CER60 and, based on heterologous expression experiments in yeast, both enzymes have similar biochemical functions, elongating mainly C24 acyl-CoA substrate to C28 acyl-CoA product (Trenkamp et al., 2004). KCS6/CER6 is highly expressed in various organs throughout plant development, whereas KCS5/CER60 transcript levels are much lower and likely restricted to a few organs and epidermis cell types (Hooker et al., 2002). In this context it should be noted that CER2-like proteins were recently found to be associated with KCS6/CER6-containing FAE complexes, enhancing the capacity of KCS6/CER6 to elongate acyl-CoAs beyond C28. For instance, association of KCS6 with CER2 resulted in elongation up to C30, and association with CER26 even up to C34, however the underlying mechanisms are currently unknown (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015).  Despite great progress in characterizing Arabidopsis KCSs, the biological functions of several members of this gene family remain elusive, and for some of them neither substrate nor product chain length specificities have been reported. On the other hand, several elongation steps could not be assigned to specific KCSs so far, especially those initiating elongation with relatively short precursors including C16 acyl-CoA and those extending chain lengths beyond C34.  90 However, substantial amounts of C35 and C37 wax compounds are found in Arabidopsis waxes, likely formed from C36 and C38 acyl-CoAs along the decarbonylation pathway (Bernard et al., 2012; Cheesbrough & Kolattukudy, 1984; Schneider-Belhaddad & Kolattukudy, 2000). In particular, the wax mixtures of trichome-rich young Arabidopsis leaves (compare Figure 2.2) and isolated leaf trichomes (compare Figure 3.5) were found to contain such extra-long compounds, and therefore I hypothesized that trichomes contain a cell type-specific KCS enzyme dedicated to the formation of C36 and C38 acyl-CoAs. Analysis of microarray data (Jakoby et al., 2008, Marks et al., 2009), combined with gene expression studies in young leaves (compare Figure 2.7), pointed to KCS16 as a primary candidate for the trichome-specific elongation beyond C34.  Sequence analysis of KCS16 showed high sequence similarity of 74 % with KCS8 at the protein level, and KCS16 clusters with both FAE1/KCS18 and KCS8 in the subclass β (Joubès et al., 2008). The KCS16 protein is predicted to consist of 496 amino acids, including two trans-membrane domains spanning residues 10 to 32 and 52 to 74 predicted by TMHMM (http://www.cbs.dtu.dk/services/TMHMM/). The KCS16 sequence shares all the characteristic features of other Arabidopsis KCSs, leading to the prediction that it has the four-layered structure α1-β2-α3-α4 similar to other KCSs, with a catalytic triad located in the α3 fold and a cavity running through the enzyme (Joubès et al., 2008). All this information taken together, KCS16 may be expected to be a functional KCS. However, the biochemical and biological functions of this enzyme have not been studied, except for one report showing that the gene is highly expressed in siliques and also substantially in leaves and other organs (Joubès et al., 2008), and the finding that the enzyme showed no in vitro elongation activity towards C16 – C20 substrates (Blacklock & Jaworski, 2006). The scant experimental evidence available so far thus  91 left it open whether KCS16 might be involved in the formation of extra-long leaf wax constituents.  Therefore, the goal of this chapter was to investigate if KCS16 is involved in cuticular wax formation, to characterize the enzyme in terms of substrate and product specificities, and to test whether it may have trichome-specific functions. To this end, quantitative RT-PCR and analysis of published microarray data were used to assess KCS16 expression in different organs and tissues, respectively, GC-MS and GC-FID were used to analyze the cuticular waxes on isolated trichomes, young and mature leaves as well as stems and siliques of kcs16 loss-of-function and overexpressor lines. The subcellular localization of the enzyme was tested by GFP tagging, and heterologous expression in yeast was used to test substrate and product profiles of KCS16.   4.2 MATERIALS AND METHODS  4.2.1 Plant materials and growth conditions Arabidopsis kcs16-1 (SALK_110690C; Col-0 ecotype), and kcs16-2 (Salk_35139; Columbia-0 ecotype) knockout mutant lines were obtained from ABRC (www.arabidopsis.org). Wild type (Col-0 ecotype) and mutant seeds were treated as described above (see 2.2.1). The plants were grown under the same conditions as stated above (see 2.2.1).   92 4.2.2 Wax extraction and derivatization Rosette leaves (15 leaves per replicate) and stems (4 stems per replicate) from five week-old plants were harvested and photographed, the surface area was measured using the Image J programme. The wax was extracted as described previously (see 2.2.3).  4.2.3 Wax identification and quantification Waxes were analyzed following the same procedures as stated above (see 2.2.5).  4.2.4 Trichome isolation from Arabidopsis wild type and ksc16 leaves Trichomes were isolated from rosette leaves (30 leaves per replicate) of five weeks old Arabidopsis wild-type, kcs16-1 and kcs16-2 mutant lines as well as from lines constitutively expressing KCS16, as described by Marks et al, 2008. Minor modifications were made as stated previously (as in 3.2.3).  4.2.5 Wax extraction from trichomes and derivatization Two-phase extraction was used to extract the wax from trichomes with chloroform (CHCl3) as organic solvent. The water-solvent phases were mixed by vortexing three times for 30 sec. The CHCl3 phase was transferred into a glass vial and concentrated under a N2 stream at 50°C. The samples were derivatized by adding 10 µl pyridine (Fisher-Scientific) and 10 µl N,O-bis(trimethylsilyl)trifluoroacetamid (BSTFA), (Sigma-Aldrich) and incubated at 70°C for 15 min. The samples were dried completely under a stream of N2 at 50°C and the wax was re-dissolved in 10 µl CHCl3. For quantitative and qualitative analysis, the same GC-FID and GC-MS instrumentations and methods were used as described above.  93  4.2.6 Genomic DNA extraction for PCR genotyping Tissue from single plants was harvested, and genomic DNA was extracted by adding 50 µl of extraction buffer (Sigma Aldrich), incubation at 90°C for 10 min, and adding 50 µl of dilution buffer (Sigma Aldrich). Samples were mixed by vortexing and stored at -20°C. All SALK lines were genotyped by PCR, using Phusion High-Fidelity DNA Polymerase (New England Biolabs). The T-DNA insertion was amplified using LbaI forward primer 5’-TGGTTCACGTAGTGGGCCATCG-3’ and a gene specific reverse primer KCS16 (At4g34150) 5’-AAGTTTTGAGTGATGTTCTCC-3’. As a control, gene specific forward and reverse primer, flanking the T-DNA insertion were used KCS16 forward 5’-GCACCTTTCTAAGCCGTTATCTG-3’ and KCS16 reverse 5’-AAGTTTTGAGTGATGTTCTCC-3’. The PCR conditions were as followed: Initial denaturation at 98°C for 30 sec, 30 cycles of a denaturation step at 98°C for 8 sec, annealing at 60°C for 20 sec, and elongation at 72°C for 1 min and 50 sec, followed by a final extension step at 72°C for 8 min. Homozygous T-DNA insertion lines were selected based on successful amplification of the T-DNA insertion region and unsuccessful amplification of the wild type region flanking the T-DNA.  4.2.7 RNA extraction and cDNA preparation Total RNA was extracted from different plant organs and lines to measure the expression level of KCS16 in Arabidopsis wild type (WT) (Col-0 ecotype), as well as kcs16-1, kcs16-2, KCS16OE in WT (line 1 and 2), and KCS16OE in kcs16-1 and -2. Tissues (50 mg per replicate) from five week-old plants were harvested and immediately stored in liquid nitrogen. Samples were ground  94 up at 4°C using a Precellys-24 homogenizer (Bertin) at a speed of 5500 rpm for 2 x 25 sec and stored immediately in liquid nitrogen until usage. RNA was extracted using the PureLink RNA Mini Kit (Ambion), following the manufacturer’s instructions (https://tools.thermofisher.com/content/sfs/manuals/purelink_rna_mini_kit_man.pdf). On column DNaseI (Invitrogen) digestion was performed as described in the manufacturer’s manual. RNA integrity and absence of genomic DNA was confirmed by 2 % agarose gel electrophoresis, purity and quantity was determined by measuring UV spectra and 260/280 and 260/230 ratios using NanoDrop 8000 Spectrophotometer (ThermoFischer Scientific) and samples with a ratio between 1.7 and 2.2 were selected as templates for qRT-PCR. A total of 1 µg RNA was used to prepare cDNA, using the reverse transcriptase SuperScript Reverse Transcriptase III (Invitrogen) and Oligo(dT)20 primers (Invitrogen), as described in the manufacturer’s manual, and stored at -20°C.  4.2.8 Semi-quantitative RT-PCR of Arabidopsis mutant and transgenic lines Semi-quantitative RT-PCR was used to determine the expression level of KCS16 in leaves in the mutant lines kcs16-1, kcs16-2, as well as in the KCS16OE in WT (line 1 and 2), and KCS16OE in kcs16-1 and -2 lines. Col-0 wild-type cDNA was used as a reference and Actin2 (AT3G18780) as a control. For the PCR 1 µg of cDNA was used as a template, and the KCS16 gene (At4G34250) was amplified with gene specific primers KCS16 forward 5’-GCACCTTTCTAAGCCGTTATCTG-3’ and KCS16 reverse 5’-AAGTTTTGAGTGATGTTCTCC-3’, Actin2 was amplified using gene specific primers forward 5’- CAC GAG ACA ACC TAT AAC TCA AT-3’ and reverse 3-GTG ATT TCT TTG CTC ATA CGG TCA-5’ (Appendix Table C.1), using Phusion High-Fidelity DNA Polymerase (New  95 England Biolabs). The PCR conditions were as followes: initial denaturation step at 98°C for 30 sec, followed by 25 cycles of denaturation at 98°C for 10 sec, annealing at 52°C for 20 sec, and elongation at 72°C for 50 sec. The last step was the final extension at 72°C for 8 min. The samples were analyzed by agarose gel electrophoresis (1 %).  4.2.9 Quantitative Real-Time-PCR of Arabidopsis wild-type plants The relative expression level of KCS16 was measured in different organs of Arabidopsis (ecotype Col-0). Four genes (GAPDH, UBC21, Actin2 and UBQ10) (Appendix Table C.1) were tested as reference genes, and UBC21 was used for normalization to determine relative gene expression. For each sample KCS16 (At4G34250) and UBC21 were PCR amplified, using gene specific primers qRT-PCR KCS16 forward 5’-CCCATAAGCGAGAAGTTTCA-3’, qRT-PCR KCS16 reverse 5’-CCAACTCGTACCAAATGGAG-3’, qRT-PCR UBC21 forward 5'-GAATGCTTGGAGTCCTGCTT-3', qRT-PCR UBC21 reverse 5'-GGCGAGGCGTGTATACATTT-3' and 1x SYBR Green Master Mix (Bio-Rad). The total amount of 2 ng of cDNA was used in a final volume of 10 µl reaction mixture, following the manufacturer’s protocol. Real-time PCR was performed using the CFX Connect Real-time PCR Detection System (Bio-Rad, Hercules, CA, USA) with the following conditions: 1 cycle at 95°C for 3 min followed by 40 cycles at 95°C for 15 sec, and at 60°C for 30 sec. For data acquisition and analysis, the CFX Connect Real-time PCR Detection System software (version 3.0a, Bio-Rad) was used. A no template control was included in each run and all samples were run as triplicates and in two biological replicates on 96 well plates (Bio-Rad). Melting curves were performed after each run to verify primer specificity. PCR efficiency was tested by preparing a 10-fold dilution series and confirmed to be between 95 % and 103 %. For data analysis,  96 threshold cycles were adjusted and relative expression was calculated using the Pfaffl method for correction of primer efficiency (Pfaffl, 2001).  4.2.10 Cloning of the KCS16 coding region for constitutively expression and subcellular localization The KCS16 coding region was cloned using the GATEWAY method, to be able to constitutively express KCS16 and to determine subcellular localization by GFP-tagging. Genomic DNA was used to PCR amplify the coding region using the forward primer KCS16 GTW fw 5’-GGG GAC AAG TTT GTA CAA AAA AGC AGA TGG ATT ACC CCA TGA AGA AGG TAA AAA TCT TTT TCA ACT-3’ and reverse primer KCS16 GTW re 5’-GGG GAC CAC TTT GTA CAA AGC TGT CAC TCT TTT AAA TCT ATA TCG ATC TCA ACT GGA TAT TTG T-3’. Phusion High-Fidelity DNA Polymerase (New England Biolabs) was used with the following thermal conditions: initial denaturation at 98°C for 30 sec, 30 cycles of denaturation at 98°C for 8 sec, annealing at 57.7°C for 8 sec, and elongation at 72°C for 30 sec, one final extension step at 72°C for 5 min. The PCR amplified KCS16 coding region (1482 bp) was inserted into the pDONR221 vectors using the Gateway BP Clonase II enzyme Mix (Invitrogen), according to the manufacturer’s manual. Chemically competent Top10 E. coli cells were transformed with the KCS16-pDONR221 construct and selected on LB agar medium enriched with 50 µg/ml kanamycin (w/v) by incubating overnight at 37°C. Colonies were picked and liquid LB medium containing 50 µg/ml kanamycin (w/v) was used to grow transformed cells overnight at 37°C. The KCS16-pDONR221 plasmid was extracted using PureLink Quick Plasmid Miniprep Kit (Invitrogen) and the construct was confirmed by sequencing. Confirmed recombinant plasmids were used for glycerol stocks and stored at -80°C. For constitutively expression of KCS16 the  97 pMDC32 expression vector containing the 2x35S-promoter was chosen. The KCS16 coding region was introduced from the KCS16-pDONR221 entry vector into pMDC32, using Gateway LR Clonase II enzyme Mix (Invitrogen), according to the manufacturer’s manual to generate the expression vectors pMDC32-2xp35S:KCS16. Chemical competent Top10 E. coli cells were transformed, selected with 50 µg/ml kanamycin (w/v) and 50 µg/ml hygromycin (w/v) and the expression vector was sequenced as described above. The subcellular localization of KCS16 was determined by C- and N- terminus tagging with GFP. The KCS16 coding region was transferred from the pDONR22-KCS16 entry vector into the GFP-containing pGWB5 and pGWB6 expression vectors, using the Gateway LR Clonase II enzyme Mix (Invitrogen), according to the manufacturer’s manual to generate the expression vectors pGWB5-p35S:KCS16-GFP (C-sGFP) and pGWB6-p35S:GFP-KCS16 (N-sGFP). Chemical competent Top10 E. coli cells were transformed, selected and the expression vector was sequenced as described above (4.2.10).  4.2.11 Agrobacterium tumefaciens mediated transformation of Arabidopsis ksc16 and Nicotiana benthamiana The Agrobacterium strain GV3101 (pMP90) was transformed with the expression vectors pMDC32-2xp35S:KCS16, pGWB5-p35S:KCS16-GFP (C-sGFP), pGWB6-p35S:GFP-KCS16 (N-sGFP) by the triparental mating method (Wise et al., 2006). For transformation preparations Agrobacterium was grown on LB agar medium containing 50 µg/ml gentamycin (w/v) for two days at 28°C, E coli containing the expression vectors were grown overnight at 37°C on LB agar medium containing 50 µg/ml kanamycin (w/v) and 50 µg/ml hygromycin (w/v) and helper E. coli cells were grown on LB agar medium containing 50 µg/ml kanamycin (w/v). Cells from each plate were collected and dissolved in 20 µl dH2O and mixed on LB agar medium without  98 antibiotics. Plates were wrapped in tinfoil and incubated overnight at 28°C. Cell mixture was scraped off the plate and transferred into 5 ml LB containing 50 µg/ml hygromycin (w/v), 50 µg/ml gentamycin (w/v) and 50 µg/ml kanamycin (w/v). Transformed cells were selected for 4 hours on the shaker at 28°C and incubated over night at 28°C on LB agar medium with 50 µg/ml hygromycin (w/v), 50 µg/ml gentamycin  (w/v) and 50 µg/ml kanamycin (w/v). Colonies were screened for the expression vectors by colony PCR using gene specific primers KCS16 forward 5’-GCACCTTTCTAAGCCGTTATCTG-3’ and KCS16 reverse 5’-AAGTTTTGAGTGATGTTCTCC-3’ for the coding region. The PCR conditions were: initial denaturation at 98°C for 10 min, 30 cycles of denaturation at 98°C for 30 sec, annealing at 60°C for 20 sec and extension at 72°C for 30 sec, the final extension step was at 72°C for 8 min, using Phusion High-Fidelity DNA Polymerase (New England Biolabs).  4.2.12 Stable transformation of Arabidopsis with p35s:KCS16-GFP  Arabidopsis wild-type (ecotype Col-0) plants, as well as kcs16-1, and kcs16-2 were transformed with Agrobacterium harboring the pMDC32-2xp35S:KCS16 construct. The floral dip method was used (Clough & Bent, 1998), and T1 and T2 seeds of were selected on LB agar medium containing 20 µg/ml hygromycin (w/v). T2 plants were used for chemical wax analysis, as well as semi-quantitative RT-PCR.  4.2.13 Transient expression of KCS16 in tobacco Four Agrobacterium lines harboring one of the following constructs: pGWB5-p35S:KCS16-GFP (C-sGFP), pGWB6-p35S:GFP-KCS16 (N-sGFP), the ER specific marker p35S:HDEL-RFP or the control construct 35S::YFP-WBC11 (Bird et al., 2007b), were grown over night from  99 glycerol stocks in 5 ml LB containing respective antibiotics for selection. The cells were collected by centrifugation for 5 min at 3000 rpm, and the LB was removed. A washing-step with dH2O was included before re-suspending the cells in 4 ml dH2O. Three weeks old tobacco plants (five per construct) were prepared for infiltration by scratching the abaxial leaf site with a sterile needle. A 4 ml syringe was used to infiltrate tobacco plants with Agrobacterium harboring pGWB5-p35S:KCS16-GFP (C-sGFP) or pGWB6-p35S:GFP-KCS16 (N-sGFP) a long with a ER specific marker p35S:HDEL-RFP. Infiltration of the control construct 35S::YFP-WBC11 (Bird et al., 2007) localizing to the plasma membrane, along with the ER specific marker construct was done in parallel. The marker construct, tobacco seeds and the control construct were kindly provided by Dr. Mathias Schuetz, (Samuels lab, Department of Botany, University of British Columbia). The infiltrated plants were transferred into the growth chamber and tested for expression after three days post-infiltration using a confocal microscope (BioImaging Facility, Department of Botany, University of British Columbia).  4.2.14 Laser scanning microscopy Tobacco plants transiently expressing KCS16 tagged with GFP and Arabidopsis plants constitutively expressing KCS16 tagged with GFP were analyzed for GFP expression using laser scanning microscopy. Arabidopsis samples were stained with an ER specific stain, according to manufacturer’s protocol (Hexyl rhodamine B, ThermoFisher Scientific). Tobacco and Arabidopsis leaves were visualized with an Olympus multi-photon confocal microscope (FV1000MPE, BioImaging Facility, Department of Botany, University of British Columbia) and tested for co-localization of ER specific markers and gene constructs tagged with GFP. GFP was visualized with an excitation of 488 nm and emission was collected at 509 nm, RFP with an  100 excitation of 543 nm and emission collection at 588 nm, and YFP with an excitation of 514 nm and emission collection at 535 nm. Olympus FluoView FV1000 software was used for analysis.  4.2.15 Cryo-scanning electron microscopy For cryo-SEM, segments of ~4 mm2 were sampled from the center of rosette leaves and segments ~4 mm long were sampled 2-3 cm from the stem base. Samples were mounted onto copper stubs using PELCO water-based graphite paint (Ted Pella). Stubs were transferred to an Emitech K1250 cryo-system (Emitech Inc.) and frozen at -120˚C.  Samples were loaded into a S4700 field emission SEM (Hitachi) and held at -100˚C for 10 min to remove ice by sublimation prior to viewing. Samples were imaged by the backscatter detector, using an accelerating voltage of 1.5 kV, a beam current of 15 mA, and a working distance of 12 mm.  4.2.16 Heterologous expression in yeast The KCS and CER2-LIKE/CER26 (At4g13840) open reading frames cloned into the pDONR221 ENTRY vector (Joubès et al., 2008; Pascal et al., 2013) were transferred into yeast expression vectors as detailed in Appendix Table C.2. Saccharomyces cerevisiae strain INVSc1 [MATa/MATα his3D1/his3D1 leu2/leu2 trp1-289/trp1-289 ura3-52/ura3-52] cells were transformed with different combinations of constructs by a poly-ethylene glycol/lithium acetate protocol (Ausubel et al., 2003) and grown on minimal medium lacking appropriate amino acids as indicated in Appendix Table C.2. For fatty acyl chain analyses, yeasts were grown in 25 ml appropriate liquid minimal medium supplemented with 2 % galactose. Expression cultures were grown for one week at 30°C. Yeast cells were pelleted and washed in 10 ml of NaCl (2.5 %, w/v) before fatty acyl chain analyses. Fatty acid methyl esters were obtained by transmethylation  101 at 85°C for 3 h of yeast cell sediments with 4 ml 0.5 M sulphuric acid in methanol containing 2 % v/v dimethoxypropane. After cooling, 4 ml of NaCl 2.5 % (w/v) was added, and fatty acyl chains were extracted in 4 ml hexane and then dried under a gentle stream of nitrogen. For qualitative analysis the yeast samples were loaded on a GC connected to a mass spectrometric detector (5973N, Agilent; column 30 m HP-1, 0.32 mm i.d., df = 0.1 µm, Agilent) using temperature-programmed on-column injection at 50°C, oven for 2 min at 50°C, then raised by 40°C/min to 200°C, held for 2 min at 200°C, then raised by 3°C/min to 320°C, and finally held at 320°C for 30 min, with helium gas inlet pressure, programmed to 1.4 ml/min. Peaks were identified by comparison with characteristic fragmentation peaks and compounds were quantified counting the abundance of molecular ions specific for each homolog within the respective peak (C26 = m/z 410; C28 = m/z 438; C30 = m/z 466; C32 = m/z 494; C34 = m/z 522; C36 = m/z 550; C38 = m/z 578). Relative abundance was calculated as % of total of molecular ions of all eight homolog peaks.  4.2.17 Statistical analysis GraphPad Prism v6.0 software was used for statistical analysis. Pair-wise comparisons were performed simultaneously on the entire dataset using Student’s t-tests (two-tailed, alpha = 0.05) and raw p values adjusted using a False Discovery Rate (FDR) equal to 5 %.  4.3 RESULTS  The current chapter aimed to test the role of the Arabidopsis KCS16 gene in the formation of cuticular wax compounds. To this end, I analyzed (4.3.1) the expression patterns of KCS16 in  102 various organs and tissues, (4.3.2) the wax mixtures on various organs of ksc16 T-DNA insertion lines, (4.3.3) the cell type-specific composition of waxes on leaf and stem trichomes of ksc16 insertion lines and the wild type, (4.3.4) the subcellular localization of the KCS16 protein, and (4.3.5) its biochemical function using overexpression in Arabidopsis as well as heterologous expression in yeast.  4.3.1 Expression patterns of KCS16 To correlate gene expression with the amounts of potential wax products formed by the KCS16 gene product, I first analyzed its relative expression levels in various Arabidopsis organs, tissues and cell types. An inspection of the available in silico data (http://bar.utoronto.ca) revealed that KCS16 is expressed ubiquitously in various organs of the plant, with high expression in seeds, the vegetative shoot apex, the vegetative rosette, and developing leaves (Figure 4.2A). At the tissue level, KCS16 was more highly expressed in root and stem epidermis compared to the endodermis, indicating a cuticle-related function of KCS16 (Figure 4.2B). Cell type-specific analysis of KCS16 expression in leaves showed higher expression in guard cells and trichomes compared to mesophyll cells (Figure 4.2B). Furthermore, KCS16 was expressed in carpels, and in ovary and stigma tissues (Figure 4.2B. To verify the in silico results, organ-specific expression levels of KCS16 were tested by quantitative RT-PCR. Preliminary tests showed that, of the four reference genes included in the analysis (Actin2, GAPDH, UBC21, UBQ10), UBC21 showed the most stable expression across samples. It was therefore used for normalization of the KCS16 results. KCS16 was expressed most highly in rosette and cauline leaves, with high variation in cauline leaves. KCS16 was expressed at medium levels in flowers and siliques (Figure 4.3). In contrast, only very low amounts of KCS16 transcript were detected in roots and stems, with  103 slightly higher levels in the bottom parts of the stems relative to the top. Overall, our experimental results thus confirm the in silico data.   Figure 4.2: Developmental, cell type and tissues specific expression of KCS16 in Arabidopsis, based on microarray data (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi; TGT = Target intensity value). A) Expression of KCS16 during development. Numbers indicate stages of seed development: 3 = siliques with seeds, mid globular to early heart embryo; 4 = siliques with seeds, early heart embryo to late heart embryo; 5 = siliques with seeds, late to mid torpedo heart embryo; 6 = mid to late torpedo embryo; 7 = late torpedo to early walking-stick embryo; 8 = walking-stick to early curled-cotyledons embryo; 9 = curled-cotyledons to early green cotyledons embryo; 10 = green cotyledons embryo. B) Cell type and tissue specific expression of KCS16. Error bars indicate standard deviation of a total of three replicates.   It was therefore used for normalization of the KCS16 results. KCS16 was expressed most highly in rosette and cauline leaves, and at medium levels in flowers and siliques (Figure 4.3). In contrast, only very low amounts of KCS16 transcript were detected in roots and stems, with slightly higher levels in the bottom parts of the stems relative to the top. Overall, our experimental results thus confirm the in silico data.   w/ Sil 3w/ Sil 4w/ Sil 5 6 7 8 9 10HypocotylCotyledonVegetative rosetteEntire rosetteDevelopingMatureSenescingCauline leaf010020030040050060070080090010001100Expression signal (TGT = 100)Rosette leafSeed stageA Whole rootWhole stemStem epidermisXylemMesophyll cellsGuard cellsTrichomesDeveloping trichomesLeaves no trichomesCarpelsPetalsSepalsStamenPedicelsOvary tissueStigma tissueMature pollen0255075100125150175200225250Expression signal (TGT = 100)B  104  Figure 4.3: KCS16 gene expression relative to UBC21 in various Arabidopsis organs. The amount of mRNA was analyzed by qRT-PCR in three replicates and two in dependent experiments. Error bars indicate standard deviation of a total of three replicates.   4.3.2 Isolation and characterization of kcs16 T-DNA insertion lines In a second set of experiments, Arabidopsis kcs16 loss-of-function lines were isolated and studied to test the involvement of KCS16 in wax biosynthesis. The kcs16-1 (SALK_110690C) and kcs16-2 (SALK_035139C) mutant lines were predicted to have T-DNA insertions in the promoter region of KCS16 (Figure 4.4A). Semi-quantitative RT-PCR showed that leaves of homozygous kcs16-1 and kcs16-2 plants had very low amounts of KCS16 transcript expression RootsLower stemUpper stemRosette leavesCauline leavesFlowersSiliques0.000.020.040.060.080.100.120.140.160.180.20Gene expression (relative to UBC21) 105 compared to the wild type (Figure 4.4B). It should be noted that neither of the kcs16 insertion lines showed any visual phenotypic differences compared to wild-type plants (data not shown).   Figure 4.4: Characterization of Arabidopsis kcs16 T-DNA insertion and ectopic lines. A) T-DNA insertion sites of kcs16-1 (SALK_0110690C) and kcs16-2 (SALK_035139C) mutant lines on the KCS16 gene. B) Semi-quantitative RT-PCR of KCS16 gene expression (30 cycles) in Arabidopsis rosette leaves of (1) Col-0, (2) kcs16-1, (3) kcs16-2, (4) and (5) p35S:KCS16 in Col-0 line 1 and –line 2, (6) p35S:KCS16 in kcs16-1, (7) p35S:KCS16 in kcs16-2. Actin2 was used as reference.   To further characterize the kcs16-1 and kcs16-2 insertion lines, the cuticular waxes covering their stem, silique and leaf surfaces were analyzed using GC-FID and GC-MS. The total wax loads of kcs16-1 and kcs16-2 stems (16.5 ± 2.2 µg/cm2 and 15.5 ± 2.4 µg/cm2, respectively) were not significantly different from the wild type (18.6 ± 1.4 µg/cm2). Similarly, also the compound class and chain length compositions of the kcs16-1 and kcs16-2 stem waxes showed no significant difference to those of the wild type (Appendix Figure C.1A & B). Furthermore, the total wax loads of kcs16-1 and kcs16-2 siliques (5.5 ± 0.9 µg/cm2 and 4.9 ± 1.1 µg/cm2, respectively) were not significantly different from the wild type (5.3 ± 0.8 µg/cm2) and the compound class and A B 5’#$ #3’$T"DNA&inser,on&kcs16&1'(SALK_&110690C)$$UTR&(1&"&41bp)& UTR&(1524&"&1788bp)&T"DNA&inser,on&kcs16&2'(SALK_&035139C)$$KCS16'(42&"&1523bp)& 106 chain length distributions of the kcs16 silique waxes were identical to those of the wild type (Appendix Figure C.1C & D).  The total wax loads on mature leaves of kcs16-1 and kcs16-2 (1.4 ± 0.3 µg/cm2 and 1.3 ± 0.3 µg/cm2, respectively) were not significantly different from the wild type (1.5 ± 0.1 µg/cm2). The compound class distribution within the kcs16-2 leaf wax showed a slight decrease in alkenes, while kcs16-1 leaf alkenes were not significantly different to the wild type (Figure 4.5A). The chain length profiles within most compound classes did not differ between the kcs16 insertion lines and the wild type (Appendix Figure C.2A). However, kcs16-2 leaf wax contained significantly less C35 alkane than the wild type, while maintaining the same amounts of all other alkane homologs (Figure 4.5B). Similarly, the kcs16-1 and ksc16-2 leaf waxes had substantially lower concentrations of C35 and C37 alkenes than the wild type (Figure 4.4B).  To test the effect of KCS16 in an early stage of organ ontogenesis, when trichomes are already formed while pavement and guard cells are still developing through division and expansion, I analyzed the wax composition on young leaves of kcs16-1 and kcs16-2. Rosette leaves were harvested when they were ca. 30 % of their final size, after approximately seven days of individual leaf development. The young kcs16-1 and kcs16-2 leaves had total wax loads of 1.2 ± 0.4 µg/cm2 and 1.2 ± 0.3 µg/cm2, respectively, and thus very similar to young wild-type leaf coverage (1.1 ± 0.3 µg/cm2). Also, the compound class distribution of both kcs16 insertion lines was not significantly altered compared to the wild type (Figure 4.5C). The chain length distributions within most compound classes were identical between the mutant and wild-type leaves, similar to mature leaves (Appendix Figure C.2B). However, young leaves of both kcs16  107 insertion lines showed significant alterations in their C33 - C37 hydrocarbon profiles (Figure 4.5D). In particular, the amounts of C35 and C37 alkanes were significantly reduced in ksc16-1 and ksc16-2, relative to wild type. Similarly, the concentrations of C35 and C37 alkenes were also drastically decreased in young kcs16 leaves, relative to wild type leaves of the same age. In sharp contrast, the C33 alkene was found at increased levels in ksc16-1 and ksc16-2 compared to the wild type (Figure 4.5D).  Finally, to verify that the observed wax phenotypes were caused by sole mutation of the kcs16 gene, I tested whether ectopic expression of wild-type 35S::KCS16 in the kcs16 background could rescue the mutant phenotype. Semi-quantitative RT-PCR showed that the transgene was expressed in both kcs16-1 and kcs16-2 insertion lines at higher levels than in the wild type (Figure 4.4B). Young leaves of the transgenics had total wax loads (1.5 ± 0.3 µg/cm2 and 1.1 ± 0.2 µg/cm2, respectively) identical to the wild type (1.1 ± 0.2 µg/cm2), and also compound class distributions indistinguishable from it (Figure 4.5E). Both ectopic lines, expressing KCS16 in either ksc16-1 or ksc16-2, had levels of C35 and C37 alkanes as well as alkenes restored to wild-type concentrations (Figure 4.5F & Appendix Figure C.2C). The amounts of C33 alkenes remained slightly increased relative to the wild type, similar to the corresponding ksc16-2 line. In summary, the ksc16 wax phenotype observed in young leaves was successfully rescued by expression of KCS16. Overall, our kcs16 loss-of-function mutant characterization shows that KCS16 is involved in the formation of wax compounds with extremely long chain lengths up to C37.     108  Figure 4.5: Wax composition of mature and young leaves of Arabidopsis kcs16 insertion and complementation lines expressing KCS16 ectopically in both ksc16 backgrounds. A) Compound class composition of total wax of mature leaves. B) Amount of alkanes and alkenes in mature leaves of kcs16 mutants. Labels on x-axis indicate the carbon number of identified compounds. C) Compound class composition of total wax of young leaves. D) Amount of alkanes and alkenes. Labels on x-axis indicate the carbon number of identified compounds. E) Compound class distribution of ectopic Arabidopsis lines expressing KCS16 in the kcs16-1 and kcs16-2 backgrounds. F) Amount of alkanes and alkenes in the ectopic complementation lines expressing KCS16 in both kcs16 backgrounds. Labels on x-axis indicate the carbon number of identified compounds. Error bars represent the mean standard deviation of a total of five replicates. Error bars represent the mean standard deviation of a total of five replicates. Asterisks indicate discovery of significant differences of coverage between wild type and ksc16 lines based on Student’s t-test (* = p < 0.05). Fatty acidsAldehydesAlkanesAlkenessec. n-AlcoholsKetoneprim. n-Alcoholsprim. br. AlcoholsSterolsNot identified0.00.10.20.30.40.50.60.70.80.9Coverage (µg/cm2 )Fatt acidsAldehydesAlkanesAlkenessec. n-AlcoholsKetoneprim.n-Alcoholsprim. br. AlcoholsSterolsNot identified0.00.10.20.30.40.50.60.70.80.9Coverage (µg/cm2 )Col-0kcs16-1kcs16-2*A B D C F E Fatty acidsAldehydesAlkanesAlkenessec.n-AlcoholsKetoneprim.n-Alcoholsprim. br. AlcoholsSterolsNot identified0.00.10.20.30.40.50.60.70.80.9Coverage (µg/cm2 )p35S:KCS16 in kcs16-1Col-0p35S:KCS16 in kcs16-234 35 36 37 33 35 370.000.010.020.030.040.050.060.07Alkanes Alkenes** * **Coverage (µg/cm2 )27 28 29 30 31 32 33 34 35 36 37 33 35 370.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Alkanes Alkenes* ** * *34 35 36 37 33 35 370.000.010.020.030.040.050.060.07Alkanes Alkenes*****Coverage (µg/cm2 )***27 28 29 30 31 32 33 34 35 36 37 33 35 370.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Alkanes Alkenes* ** * ** ****z 34 35 36 37 33 35 370.000.010.020.030.040.050.060.07Alkanes AlkenesCoverage (µg/cm2 )*27 28 29 30 31 32 33 34 35 36 37 33 35 370.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Alkanes Alkenes*z  109 4.3.3 Wax composition of isolated trichomes To test if KCS16 is involved in wax biosynthesis in trichome cells in particular, I analyzed the cuticular wax of trichomes isolated from both wild-type and kcs16 leaves and stems. The wax mixture on the wild-type leaf trichomes comprised 64 % alkanes, 6 % primary n-alcohols, 3 % alkenes, <1 % branched alcohols and 24 % not identified (Figure 4.6A). Similarly, the main compound classes in the ksc16-1 and ksc16-2 leaf trichome waxes were alkanes (51 % and 65 % of total wax compounds, respectively), together with primary n-alcohols (7 % and 4 %), alkenes (2 % for both mutant lines), small amounts of branched alcohols (<1 % for both mutant lines) and 40 % and 28 % not identified for ksc16-1 and -2 respectively (Figure 4.6A). The chain length distributions within the n-alcohols and branched alcohols were identical between wild-type and kcs16 leaf trichome waxes (Figure 4.6B). However, the chain length profiles of the alkanes and alkenes were altered in the kcs16 leaf trichomes (Figure 4.6B), with a significant decrease of C35 alkane concentrations relative to the wild type. The C37 alkane and alkene could not be detected at all in the ksc16 leaf trichome waxes, in sharp contrast to the wild type.  The wax composition of wild-type stem trichomes comprised 41 % alkanes, 12 % ketone, 5 % primary n-alcohols, 4 % secondary alcohols, 3 % aldehydes and 33 % not identified (Figure 4.6C). Similarly, stem trichome waxes of ksc16-1 and ksc16-2 were composed of 46 % and 48 % alkanes respectively, 17 % and 16 % ketone, 5 % and 6 % primary n-alcohols, 7 % and 5 % secondary alcohols, 3 % and 5 % aldehydes, and 23 % and 20 % not identified respectively. The chain length profiles of aldehydes, ketone, secondary alcohols and primary n-alcohols were similar between wild-type and ksc16 stem trichomes (Figure 4.6D). The C27 - C31 alkanes also showed a similar distribution between wild type and ksc16 insertion lines, however C33 alkane  110 was significantly decreased in the kcs16-2 insertion line compared to wild type (Figure 4.6D). Taken together, the trichome wax analyses showed that leaf and stem trichomes have cuticular waxes resembling those of adjacent pavement cells on either organ, but with a shift towards longer chain lengths in both trichome types.   Figure 4.6: Wax composition of trichomes isolated from leaves and stems of Arabidopsis kcs16 insertion lines and wild type. A) Relative compound class composition of total identified wax in leaf trichomes. B) Carbon chain length distribution within compound classes in leaf trichomes. Labels on x-axis indicate the carbon number of identified compounds. C) Relative compound class composition of total identified wax in stem trichomes. D) Carbon chain length distribution within compound classes in stem trichomes. Labels on x-axis indicate the carbon number of identified compounds. Error bars represent the mean standard deviation of a total of five replicates for the leaves and three replicates fro the stems. Asterisks indicate discovery of significant differences of coverage between wild type and ksc16 lines based on Student’s t-test (* = p < 0.05). AlkanesAlkenesprim.n-Alcoholsprim. br. AlcoholsNot identified01020304050607080Relativee compositions(% of total wax) Col-0kcs16-1kcs16-2B C D 26 28 30 32 27 29 31 33 29 29 24 26 28 30 3205101520253035404550Relative composition(% of total wax)AlkanesAldehydes prim. n-AlcoholsKetonesec. n-Alcohols*AldehydesAlkanessec. n-AlcoholsKetoneprim. n-AlcoholsNot identified01020304050607080Relative composition(% of total wax)A 27 29 31 33 35 37 33 35 37 26 28 30 32 34 30 32 3405101520253035404550Relative compositions(% of total wax) Alkanes Alkenes prim. n-Alcoholsprim. br. Alcohols** ** **35 37 33 35 370.00.51.01.52.02.53.03.54.0Relative compositions(% of total wax) **** ** 111 To test if the observed alterations in the kcs16 waxes have an effect on respective cuticle structures, kcs16 leaf and stem surfaces were investigated using cryo-SEM. The micro-relief of leaf pavement cells and leaf trichomes did not appear to be different between the wild type (Figure 4.7A) and ksc16-1 (Figure 4.7B). Moreover, the trichome surfaces also appeared identical between the wild type and ksc16-1 (Figure 4.7C & D, respectively). The surfaces of stems were also identical between the wild type and kcs16-1, in their overall appearance (Figure 4.8A & B, respectively) and in the surface details of trichomes (Figure 4.8C & E and Figure 4.8D & F, respectively) and pavement cells (Figure 4.8G & H, respectively).   Figure 4.7: Cryo-SEM micrographs of Arabidopsis wild type and kcs16-1 leaf trichomes. A) & B) Trichome on wild type and ksc16-1 leaf, respectively. C) & D) Surface structure of wild type and kcs16-1 leaf trichomes, respectively.   300 µm 20 µm  300 µm  20 µm  A B C D  112   Figure 4.8: Cryo-SEM micrographs of Arabidopsis wild type and kcs16-1 stem trichomes and wax crystals. Trichome on wild-type (A) and ksc16-1 (B) stems. Surface structure of wild type (C) and kcs16-1 (D) stem trichomes. Trichome base and wax crystals on wild type (E) and kcs16-1 (F) stems. Wax crystals on wild type (G) and kcs16-1 (H) stems. 200 µm 200 µm 20 µm 20 µm 20 µm 20 µm 10 µm 10 µm A B C D E F G H  113 4.3.4 Subcellular localization of KCS16 The subcellular localization of KCS16 was determined by tagging with GFP at the C- and N-termini and expression under control of a constitutive promoter (p35S:GFP-KCS16 and p35S:KCS16-GFP) in either tobacco leaves or Arabidopsis ksc16-1 plants. First, both KCS16-GFP constructs were transiently expressed in tobacco (Nicotiana benthamiana) leaves together with an ER-specific marker construct. Visualization by fluorescence microscopy showed that the KCS16-GFP signal coincided with the ER-specific marker construct p35S:HDEL-RFP (Figure 4.9). Second, Arabidopsis kcs16-1 plants were transformed with either p35S:GFP-KCS16 or p35S:KCS16-GFP, and GFP expression was visualized by fluorescence microscopy after staining with the ER-specific marker Hexylrhodamine B. Similar to the transient expression experiment, stable Arabidopsis transformation showed that KCS16-GFP constructs were co-localized with the ER-specific marker (Figure 4.10A-C).  Finally, to test functionality of the KCS16-GFP fusion constructs, the wax compositions of kcs16-1 harboring either p35S:KCS16-GFP or p35S:GFP-KCS16 constructs were analyzed and compared to the wild type. Young leaves of kcs16-1 plants expressing either p35S:KCS16-GFP or p35S:GFP-KCS16 had levels of C35 and C37 hydrocarbons restored to normal from the kcs16 phenotype, while p35S:GFP-KCS16 had significantly more C33 alkenes compared to the wild type (Figure 4.10D & Appendix Figure C.2C). This result shows that both KCS16-GFP constructs were functional, and thus confirms that the active KCS16 enzyme must be residing in the ER. Hence, KCS16 is localized to the subcellular compartment known to harbor wax biosynthesis, and thus may have access to VLCFA-CoA intermediates necessary for its putative elongase function.  114  Figure 4.9: Subcellular localization of Arabidopsis KCS16 in tobacco pavement cells. Transient expression of p35S:KCS16-GFP (A) and the ER specific p35S:HDEL-RFP construct (B). (C) shows the merge of (A) and (B). Scale bars = 40 µm.    Figure 4.10: Subcellular localization of KCS16 in Arabidopsis ksc16-1 roots and wax analysis of complementation lines. Stable transformation of p35S:KCS16-GFP (A) in kcs16-1 Arabidopsis roots. ER specific stain Hexyl Rhodamine B (B) and as a merge (C). Scale bars = 10 µm. D) Complementation of the wild type amount of alkanes and alkenes in p35S:KCS16-GFP and p35S:GFP-KCS16 lines expressed in kcs16-1. Labels on x-axis indicate the carbon number of identified compounds. Error bars represent the mean standard deviation of a total of five replicates. Asterisks indicate discovery of significant differences of coverage between wild type and p35S:KCS16-GFP and p35S:GFP-KCS16 lines based on Student’s t-test (* = p < 0.05). A B CA B C D 34 35 36 37 33 35 370.000.010.020.030.040.050.060.07Alkanes AlkenesCoverage (µg/cm2 )*27 28 29 30 31 32 33 34 35 36 37 33 35 370.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Alkanes Alkenes*Col-0p35S:KCS16-GFP in kcs16-1p35S:GFP-KCS16 in kcs16-1 34 35 36 37 33 35 370.000.010.020.030.040.050.060.07 Col-0p35S:KCS16-GFP in kcs16-1Alkanes Alkenesp35S: FP-KCS16 in kcs16-1* 115 4.3.5 Biochemical function of KCS16  Two experiments were carried out to test the biochemical activity of KCS16, using Arabidopsis gain-of-function mutants and heterologous expression in yeast. First, two independent lines expressing KCS16 under the 35S promoter in the Arabidopsis wild-type background were isolated and analyzed. Semi-quantitative RT-PCR confirmed significantly increased transcript abundance of KCS16 in both overexpressor lines compared to the wild type (Figure 4.4B). The total wax load on the stems of both transgenic lines (15.8 ± 1.2 µg/cm2 and 17.1 ± 1.1 µg/cm2, respectively) was not significantly different from the wild-type wax load (18.5 ± 1.4 µg/cm2). Also, the chain length distributions within compound classes of the stems were identical to those of the wild type (Appendix Figure C.3A & B).  For both overexpressor lines, the total wax loads of mature leaves (1.6 ± 0.3 µg/cm2 and 1.2 ± 0.3 µg/cm2, respectively) and young leaves (1.5 ± 0.3 µg/cm2 and 1.2 ± 0.2 µg/cm2, respectively) were not significantly different from those of mature (1.5 ± 0.14 µg/cm2) and young (1.1 ± 0.3 µg/cm2) wild-type leaves. The compound classes and chain length distribution within compound classes was not significantly different between mature or young leaves of the transgenic lines and wild type (Appendix Figure C.3C & D and Figure 4.11A). However, the young leaves of the p35S:KCS16 line 1 showed a significant increase of C37 alkenes, with 0.0231 ± 0.0035 µg/cm2 compared to the wild type coverage of 0.0055 ± 0.0015 µg/cm2 (Figure 4.11B). Thus, a weak gain-of-function phenotype was discovered, showing an increase in one of the longest hydrocarbon products (C37) and therefore a change in wax composition opposite to respective loss-of-function lines. Interestingly, C34 branched n-alcohols were also significant increased in both overexpressor lines compared to the wild type (Figure 4.11B).  116    Figure 4.11: Wax composition of young Arabidopsis leaves overexpressing KCS16 in the wild type background. A) Compound class distribution of two independent Arabidopsis overexpressor lines expressing KCS16 in the wild-type background. B) Amount of alkanes and alkenes of two independent overexpressor lines. Labels on x-axis indicate the carbon number of identified compounds. Error bars represent the mean standard deviation of a total of five replicates. Asterisks indicate discovery of significant differences of coverage between wild type and p35S:KCS16 lines based on Student’s t-test (* = p < 0.05).   In a second experiment, expression of Arabidopsis KCS16 in yeast alone did not alter the fatty acid profile compared to wild-type yeast (data not shown), likely due to lack of fatty acid substrates with chain lengths beyond C26. To provide such substrates in vivo, CER6 was expressed together with CER2 or CER26 (CER2-like1) in yeast, which resulted in elongated fatty acids (Figure 4.13). Expression of CER6 alone or CER6 together with KCS16 did result in C28 fatty acids and very little C30, while expressing CER6 together with CER2 yielded to increased amounts of C30 fatty acids. However, a combination of CER6, CER2 and KCS16 resulted in a similar fatty acid profile as observed when expressing CER6 and CER2 alone. Expression of CER6 and CER26 (CER2-like1) resulted in increased amounts of C32 and C34 fatty acids (Figure 4.12A; Figure 4.13). Finally, co-expression of CER6 and CER26 (CER2-like1) together with KCS16 yielded C36 and C38 fatty acids which were not observed in the CER6, CER26 control 34 35 36 37 33 35 370.000.010.020.030.040.050.060.07Alkanes Alkenes*27 28 29 30 31 32 33 34 35 36 37 33 35 370.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 ) Alkanes Alkenes*A B  Fatty acidsAldehydesAlkanesAlkenessec.n-AlcoholsKetoneprim. n-Alcoholsprim. br. AlcoholsSterolsNot identified0.00.10.20.30.40.50.60.70.80.9Coverage (µg/cm2 )Col-0p35S:KCS16 in WT line 1p35S:KCS16 in WT line 2 117 (Figure 4.12B, Figure 4.13). In a further control experiment, yeast expressing KCS16 in combination with CER26 (CER2-like1) alone had fatty acid chain length profiles identical to the wild type (data not shown). Overall, our biochemical characterization experiments thus showed that KCS16 elongates fatty acids up to chain lengths of C38, which in planta can be converted into alkanes and alkenes up to C37.     118                                     Figure 4.12: GC-MS analysis of fatty acid methyl ester (FAMEs) profile after extraction and derivatization from yeast. Peaks were identified by comparison with characteristic fragmentation peaks and compounds were quantified counting the abundance of molecular ions specific for each homolog within the respective peak (C26 = m/z 410; C28 = m/z 438; C30 = m/z 466; C32 = m/z 494; C34 = m/z 522; C36 = m/z 550; C38 = m/z 578). A) FAMEs profile from yeast expressing CER6 and CER26. B) FAMEs profile form yeast expressing CER6, CER26, and KCS16.     20 25 30 35 40 45 50050001000015000200002500030000Relative abundance of specific ionsRetention time (min)C28 C30C32C34 C36 C38C26 CER6 + CER26 + KCS1620 25 30 35 40 45 50050000100000150000200000250000300000Retention time (min)Relative abundance of specific ionsCER6 + CER26C26C28C30C32C34m/z 410m/z 438m/z 466m/z 494m/z 522m/z 550m/z 578A B  119  Figure 4.13: Heterologous expression of Arabidopsis KCS enzymes in yeast. Complete (fatty acid methyl esters) FAMEs profile of yeast expressing various combinations of Arabidopsis genes shown in % of total specific ions counted. Abundance was calculated as % of combined total abundances of molecular ions of all eight homolog peaks (compare Figure 4.12). Black box shows zoom in of FAMEs with C32 to C38 of yeast expressing CER6 and CER26 and KCS16, Labels on x-axis indicate the carbon number of identified compounds. Error bars indicate mean of standard deviation of a total of four replicates (eight replicates for CER6, CER26, and CER6, CER26, KCS16). Asterisks indicate discovery of significant differences in abundance of FAMES between yeast expressing CER6, CER26 and yeast expressing CER6, CER26 and KCS16 based on Student’s t-test (* = p < 0.05).   4.4 DISCUSSION  This study aimed at characterizing KCS16 as a novel KCS enzyme participating in precursor elongation towards cuticular wax formation in Arabidopsis. A series of experiments were carried out to determine 1) the product spectrum of KCS16, 2) the substrate specificity of this enzyme, 3) the sub-cellular localization of the enzyme, 4) the cell-type-specific activity of the enzyme, 24 26 28 30 32 34 36 380102030405060708090100Abundance (% of total ion counts)CER6CER6, KCS16CER6, CER2CER6, CER2, KCS16CER6, CER26CER6, CER26, KCS16* *FAMEs32 34 36 380246810Abundance (% of total ion counts)* *FAMEs 120 and 5) its expression patterns across various Arabidopsis organs. All five aspects can now be discussed in detail.  4.4.1 KCS16 is involved in wax formation through synthesis of C36 and C38 acyl-CoAs In the current work, three experiments tested the involvement of the Arabidopsis KCS16 protein in cuticular wax biosynthesis. In one of these experiments, the wax compositions of two independent ksc16 loss-of-function mutants were analyzed. Leaf waxes were of primary interest since the gene was found to be expressed most highly in the rosette, in particular in trichomes. KCS16 expression seems to be more variable in cauline leaves compared to rosette leaves, which might reflect a broader range of stressors affecting cauline leaves. The leaves of both ksc16 mutant lines had significantly lower amounts of C35 and C37 alkanes and alkenes than the wild type, but identical amounts of all other wax compounds. Isolated trichomes showed a very similar but more severe phenotype, almost completely devoid of C35 and C37 hydrocarbons in contrast to the wild type. Conversely, expression of KCS16 in the ksc16 background restored the wild-type wax composition, thus further confirming the involvement of KCS16 in formation of the wax compounds with 35 or more carbons. Since alkanes are thought to be derived from acyl-CoA precursors by reduction to aldehydes and further decarbonylation in which one carbon is lost, it seems plausible that C35 and C37 alkanes are derived from C36 and C38 acyl-CoAs, respectively. Our kcs16 mutant wax analyses therefore suggested that this KCS is involved in elongation of wax precursors to C36 and C38 acyl-CoAs. This conclusion was confirmed by a second experiment, in which young leaves overexpressing KCS16 in the wild type background were found to have increased amounts of C38-derived wax compounds compared to the wild type (even though compounds derived from C36 were not affected).  121  Based on the mutant analysis alone, I could not distinguish whether KCS16 participates in late elongation rounds, such as those leading from C34 to C36 and C38, or whether it is involved in earlier rounds that might specifically supply substrate for elongation to C38. To determine in which elongation step(s) this condensing enzyme is active, the product specificity of KCS16 was determined in a third experiment. In this, first only the Arabidopsis wax biosynthesis genes CER6 and CER26 were expressed in yeast, yielding acyl products up to C32 and C34 as reported in the literature (Haslam et al., 2015). In contrast to this, co-expression of CER6 and CER26 together with KCS16 in yeast resulted in accumulation of small but significant amounts of C36 and C38 fatty acyl products, thus unambiguously demonstrating that KCS16 is crucial for late elongation rounds leading from C34 to C36 and on to C38. It is therefore an active keto-acyl-CoA synthase condensing enzyme directly involved in Arabidopsis leaf wax formation. Like other KCS enzymes, it is very likely functioning in the context of a fatty acid elongase (FAE) complex, together with the same three enzymes (ketoacyl-Co reductase, hydroxyacyl-CoA dehydratase, enoyl-CoA reductase) as other FAEs involved in earlier elongation rounds.  4.4.2 KCS16 elongates C34 acyl-CoA, but not C30 acyl-CoA My yeast expression experiments can be used to assess not only the product profile but also the substrate specificity of KCS16. To this end, KCS16 was co-expressed together with various other genes involved in fatty acid elongation, combined here to provide acyl-CoA substrates for KCS16 up to certain chain lengths. First, yeast expressing KCS16 alone or together with CER2 or CER26 was found to have fatty acid chain length profiles identical with respective controls without KCS16, showing that the longest acyl substrate available in the wild-type yeast, C26, did  122 not serve as substrate for KCS16. Similarly, co-expression of KCS16, CER6 and CER2 also did not change the acid chain length distribution compared to respective controls, including substantial amounts of C30 and traces of C32 acyls, further revealing that KCS16 in this gene combination cannot use substrates with chain lengths up to C32. In sharp contrast, co-expression of KCS16, CER6 and CER26 led to the formation of C36 and C38 products, while the corresponding control yeast harboring only CER6 and CER26 contained mainly acyls up to C32 (and traces of C34). This finding, taken together with our results for heterologous expression of KCS16 together with various other genes, can be explained in two different ways. On the one hand, it may be concluded that KCS16 accepts C34 acyl-CoA substrate provided by CER6 and CER26, but not the acyls up to C30 formed by CER6 and CER2. On the other hand, it is also possible that KCS16 requires the presence of CER26, but not of CER2, to be functional. In this scenario, KCS16 would still be expected to accept C34 acyl-CoA substrate, but shorter substrate chain lengths cannot be ruled out based on the evidence thus far. From both scenarios, it follows that KCS16 is able to accept C34 substrate and, taken together with its product spectrum, it may thus be concluded that this KCS enzyme can catalyze condensation in at least two elongation rounds.  It is not clear at this point if KCS16 is associated with CER26 in planta, and if both proteins are needed to perform elongation cycles beyond C34. However, based on our current understanding where association of FAE complexes with CER2-like enzymes enhances the elongation capacity of KCS enzymes, especially in elongation rounds > C30 (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015), it seems likely that the KCS16-containing FAE complex is associated with a CER2-like enzyme to assist in handling extraordinary chain lengths beyond C34. Three CER2- 123 like homologs have been identified and characterized in Arabidopsis (CER2, CER2-LIKE1/CER26 and CER2-LIKE2/CER26-like) (Pascal et al., 2013; Haslam et al., 2015). It was suggested that CER2-like enzymes are necessary to stabilize the complex or to assist in handling compounds with extraordinary chain lengths (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015).  Previous reports had shown that CER6, a KCS enzyme similar to KCS16, is functioning in association with either of the three CER2-like proteins, leading to different product chain length profiles (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015). If an association of KCS16 with CER2-like proteins is assumed, then it is interesting to assess whether this KCS also may function with more than one of the CER2-like proteins. Our yeast results firstly suggest that KCS16 is not affected by the presence of CER2. This finding is consistent with wax composition data and gene expression profiles for Arabidopsis stems, showing a complete lack of wax constituents beyond C34, even though CER2 is highly expressed in stems (Haslam et al., 2012; Pascal et al., 2013; Haslam et al., 2015) and KCS16 too is expressed to some degree in the stem epidermis. The CER26-like protein resembled CER2 in its effect on CER6, enhancing elongation to C30 in yeast, making it unlikely that CER26-like may enhance elongation beyond C34. Furthermore, CER26-like is mainly expressed in flowers but not leaves, also ruling out that it is functioning together with KCS16 in formation of compounds with especially long chain lengths found in leaf waxes. Overall, this leaves CER26 as the only CER2-like protein likely to be associated with KCS16.   124 4.4.3 KCS16 extends the ER-resident wax biosynthesis pathways My experiments on the sub-cellular localization of KCS16 demonstrate that the protein resides in the ER, as shown by co-localization of KCS16-GFP fusion proteins with an ER-specific marker. Meanwhile, expression of the fusion protein restored the kcs16 mutant to wild-type leaf wax composition, thus confirming that the enzyme is functional in the ER. Of note, both the C- and the N-terminal fusions of the KCS16 protein with GFP were equally active, as judged by mutant wax phenotype complementation, highlighting that those terminal regions are not relevant for the KCS16 enzyme function. The ER localization of KCS16, as an enzyme functioning in wax precursor elongation, is in agreement with our current understanding of the sub-cellular organization of wax biosynthesis (Lessire et al.,1985). It is well established that elongation of C16 fatty acyl precursors to C34 acyl intermediates takes place at the ER (Samuels et al., 2008), and the other enzymes needed to complete the elongation rounds together with KCS16, likely in a FAE complex, are therefore co-localized in the same compartment (Kunst & Samuels, 2003; Xu et al., 2002; Zheng et al., 2005). Co-localization with other FAE complexes catalyzing earlier elongation rounds also implies that they can provide substrate for KCS16 in the same compartment.  The enzymes involved in transforming elongated acyl-CoAs into final wax products such as alkanes are also known to reside in the ER and may thus modify the KCS16 acyl products within the same compartment. Overall, KCS16 may hence be described as a new enzyme catalyzing additional elongation rounds en route to wax products, acquiring its substrates from and delivering its products to well-established wax biosynthesis enzymes, all within the ER. KCS16 is effectively interspersing additional steps into the wax biosynthesis pathway.  125  In light of the intermediary role of KCS16 in the ER-based formation of wax compounds, the results from heterologous expression of KCS16 in yeast and its overexpression in Arabidopsis must be compared. First, my yeast experiments showed that KCS16 yields relatively small (albeit significant) amounts of C36 and C38 acyl products, a finding that may be due to negative feed-back regulation, auto-toxicity or partial incompatibility between KCS16 and the yeast elongase complex enzymes. Second, overexpression of KCS16 in Arabidopsis led to a relatively small increase of C38-derived wax compounds in young leaves, and an even weaker (not significant) effect in mature leaves, again pointing to possible feedback regulation or toxicity. However, the in planta results may further suggest that KCS16 is an enzyme with relatively low reaction rate, compared to other KCS, possibly due to the extremely long substrate and product chain lengths. Alternatively, the amount of C34 substrates available to KCS16 might be limited in Arabidopsis leaves, perhaps due to channeling of C30 - C34 acyl-CoA products of the CER6-containing FAE (in association with CER2 or CER26) into alkane formation, possibly by physical interaction between the CER6-FAE and CER3 or a CER3-CER1 complex. Finally, also the CER3 and CER1 enzymes may have substantially higher affinity for C30 – C34 substrates than for those with C36 and beyond (assuming that both enzymes are involved in modification of all acyls up to C38).  4.4.4 Cell-type-specific function of KCS16 A sub-set of our results can be interpreted to assess how far KCS16 is involved in wax biosynthesis of certain epidermal cell types.  The high expression level of KCS16 in leaf trichomes (Jakoby et al., 2008) and the relatively high expression of KCS16 in young leaves  126 followed by a steady decline of expression over the course of leaf expansion and maturation (compare Figure 2.7) both suggested its involvement in acyl-CoA elongation within trichomes. To test this hypothesis, I first performed comparative analyses of the waxes covering kcs16 and wild-type leaf trichomes. The results revealed that the mutant trichomes almost completely lack compounds derived from C36 and C38 acyl-CoAs, such as C35 and C37 alkanes, in sharp contrast to the wild type trichomes (compare Figure 3.5B). This severe mutant phenotype suggests that KCS16 is likely to be the sole condensing enzyme involved in elongation from C34 to C38 in trichomes, and that there is no functional overlap with other KCS enzymes in this specific type of epidermis cells.  Second, I compared the wax compositions of young and mature kcs16 mutant leaves. Young kcs16 leaves exhibited a lower level of compounds derived from C36 and C38 acyl-CoA relative to the wild type, while mature leaves showed a similar but weaker phenotype. Since young leaves have a higher density of trichomes relative to the leaf surface area and therefore a higher contribution of trichome wax to the overall wax composition it is likely that the age-dependent decline in the severity of the kcs16 mutant phenotype is largely due to decreasing trichome density, lending further support to the conclusion that KCS16 alone is responsible for elongation of acyls beyond C34 in Arabidopsis leaf trichomes.  Conversely, the previous analysis of trichome-free gl1 leaves showed that also pavement cells are producing C36 and C38 wax constituents but at an overall lower abundance compared to trichome cells (compare Figure 2.3 & Figure 3.4). Since the spatial dilution of trichomes over time increases the proportion of pavement cells, they will contribute more C36+ compounds to the  127 overall wax mixture on mature leaves than on young leaves. In light of this, the majority of the C36- and C38–derived wax compounds on mature leaves of the kcs16 loss-of-function mutant is likely to reside on pavement cells, and KCS16 may consequently not be the only homolog responsible for elongation beyond C34 in pavement cells. From this, I conclude that a different, currently unknown KCS enzyme exists in pavement cells, whose substrate and product specificities are redundant or at least partially overlapping with the mostly trichome-specific KCS16.  Interestingly, the decrease of C35 and C37 alkanes in the kcs16 mutant and during leaf development was paralleled by a similar decline of C35 and C37 alkenes, suggesting that precursors for both compound classes are both elongated by KCS16. Alkenes had not been reported as constituents of Arabidopsis wax mixtures before, possibly because they are most apparent in leaf trichome wax and on young leaves and present only as a minor compound class on mature leaves. Alkenes were identified previously in the cuticular waxes of cucumber fruits and stems, barley leaves and tomato fruits, but were also found in maize pollen lipids and olive oil (Bianchi et al., 1990; Bortolomeazzi et al., 2001; Leide, 2008; Wang et al., 2015; Von Wettstein-Knowles, 2007). However, despite their fairly wide-spread occurrence, their exact structures including the double bond position and geometry remain elusive for most species. Only the structure of the barley alkenes was investigated in detail, leading to the identification of C23 to C35 alkenes primarily with 9-ene structures, together with minor amounts of 11-, 13-, 15-ene isomers (Von Wettstein-Knowles, 2007).   128 The biosynthesis of plant wax alkenes has not been investigated in any detail, and thus it is not known when on the biosynthetic pathway the double bond is introduced. In this context, it is remarkable that only alkenes with especially long chains (C33, C35 and C37) were detected in Arabidopsis waxes. Interestingly, a marked shift was observed from C35 as the main alkene in young leaves to C33 in mature leaves, thus matching the trend towards (relatively) shorter chain lengths in the alkanes over the course of leaf development, driven pavement cell expansion. To further corroborate the chain length shift of the alkenes, the wax of mature Arabidopsis leaves was carefully inspected, confirming the presence of traces of C31 alkene only. All these results reveal characteristic similarities between the chain length profiles, developmental trends and trichome-specific formation of alkanes and alkenes, suggesting common biosynthetic pathways likely involving KCS16 to both classes. Further investigation to elucidate the structure and biosynthesis of alkenes in Arabidopsis are currently underway in our laboratory.  4.4.5 Functions of KCS16 in other organs After firmly establishing the direct involvement of KCS16 in wax biosynthesis of leaf trichomes and pavement cells, now the possible roles of this enzyme in other organs can also be assessed. The low but significant expression of KCS16 in stems suggested a somewhat minor function of the enzyme in stems. However, no significant difference was observed between kcs16 and wild type stem waxes, in accordance with the finding that whole stem wax does not contain detectable amounts of wax compounds beyond C34. This finding on whole stems might be due to the prevalence of pavement cells on the stem surface, raising the question whether KCS16 might be involved in wax formation in stem trichome cells, analogous to its function in leaf trichomes. Stem trichome wax had not been analyzed before, mainly due to difficulties with potential  129 contamination by epicuticular wax crystals on adjacent pavement cells. Our present analyses revealed a stem trichome wax composition similar to whole stem wax, and thus to stem pavement cells, but with a shift towards longer-chain compounds. This result is in accordance with previous results on stem wax from the trichome-rich Arabidopsis mutant cpc tcl1 etc1 etc3 that showed a similar increase of longer chain lengths compared with wild type (compare Figure 3.5) making it likely that our direct wax analyses do accurately reflect the stem trichome wax composition, with relatively minor contamination from pavement cells. Based on this, it is interesting to note that the stem trichome wax did not contain detectable amounts of compounds derived from C36+ acyls, in contrast to leaf trichomes. However, it cannot be ruled out that KCS16 in stems may still function, like in leaves, in the formation of especially long wax compounds, albeit in trace quantities below our detection limit.  The high abundance of KCS16 transcripts in flowers suggested an important role of the condensing enzyme in floral organs, however ksc16 mutants did not show visual defects in the flowers. Wax analyses of whole ksc16 flowers did not reveal wax compounds  > C32 or significant differences between wild type and ksc16 mutants (data not shown). Analysis of tissue-specific microarray data showed KCS16 expression only in specific parts of the flower, in particular the carpels, the ovary tissue and the stigma (Le et al., 2010), suggesting that KCS16 might be involved in formation of waxes destined for only small portions of the flower surface and possible mutant effects might be hard to detect on whole flowers. This leaves the possibility that KCS16, and its acyl products with especially long chains, might still serve important functions in plant reproduction. In line with this hypothesis, the importance of cuticular lipids for the integrity and function of reproductive organs has been demonstrated repeatedly, for example  130 showing the importance of cuticle structure and composition for anther development (Aarts et al., 1995; Chen et al., 2003; Millar et al., 1999), or male sterility due to reduction of waxes in pollen tryphine layers and disrupted pollen-pistil interactions (Preuss et al., 1993).  Finally, KCS16 showed also high transcript abundance in siliques, and microarray data revealed that the gene is highly expressed in seed endosperms during late stages of seed development (Le et al., 2010), pointing to a possible involvement of KCS16 in the production of cuticular compounds in the latter. Interestingly, De Giorgi et al. (2015) showed that seeds indeed do contain a cuticular film covering the endosperm, raising the possibility that seeds may contain waxes. KCS16 might be participating in the formation of such seed waxes, should they exist. Overall, the high expression of KCS16 in reproductive organs and seeds suggested additional functions of KCS16 besides elongation of acyl-CoAs for cuticular waxes in leaves. Further studies are needed to understand the function of KCS16 in reproductive organs as well as seed coats.  In summary, I was able to show that KCS16 is involved in wax biosynthesis in Arabidopsis leaves, through the elongation of C34 acyl-CoA to C36 and C38 acyl-CoA products that can be further modified to C35 and C37 alkanes. Furthermore, the enzyme is also involved in the elongation towards corresponding C35 and C37 alkenes. It seems likely that KCS16 is associated with CER26 for elongation of especially long substrates, however the direct involvement of this helper protein cannot be tested at the moment. KCS16 is the sole condensing enzyme responsible for biosynthesis of C36 and C38 acyl-CoAs in leaf trichome cells, and partially involved in their formation in adjacent pavement cells. Finally, based on gene expression analysis, I hypothesize  131 that KCS16 may also be relevant for wax production in other organs and tissues, such as stem trichomes, seed coats and carpels.  132  Major findings and conclusions Chapter 5: The goal of my PhD work was to improve the spatial and temporal resolution of cuticular waxes at the chemical and genetic level in order to identify and characterize new wax biosynthesis genes. I used Arabidopsis as a model organism to analyze changes in morphology, cuticular wax compositions and gene expression of wax biosynthesis genes at different time points during leaf development and in different epidermal cell types.   5.1 MAJOR FINDINGS  5.1.1 Changes in cuticular wax coverage and composition on developing Arabidopsis thaliana leaves To investigate the processes involved in wax formation with improved temporal resolution, the development of Arabidopsis leaf surfaces was monitored using a combination of morphological, chemical, and gene expression measurements. Our morphological measurements showed that during the early stages of development (between five and 21 days of age) leaf expansion was driven by increasing pavement cell numbers and therefore by cell division, whereas after 13 days of age leaf expansion was mainly driven by pavement cell expansion. In the intermittent period, half way through leaf expansion, leaf bases were comprised of dividing cells while most cells at the leaf tip were expanding. Trichome numbers stayed constant over the time of development, resulting in high trichome density in young leaves, which decreased with leaf expansion.   133 Chemical analysis of cuticular waxes at various leaf development stages showed a decrease in fatty acids and an increase of alkanes in the trichome-free mutant and wild type during ontogenesis. The shift in compound class distribution was accompanied by a chain length shift in wild type and gl1, where C24/C26 compounds decreased while C30/C32 compounds increased. Next, expression of selected wax biosynthesis genes was monitored during leaf development. Despite the observed change in compound class compositions, wax biosynthesis enzymes involved in head-group modifications were expressed at fairly stable levels, with much higher expression of CER3 and CER1 compared to CER4. CER2-like enzymes, which enable elongation to relatively long chains when associated with KCSs, were also expressed at fairly stable but overall high levels. The expression levels of KCSs elongation genes were variable, with KCS6/CER6 and KCS8 expressed at much higher levels compared to KCS1, KCS5, and KCS16. Interestingly, expression of KCS1 and KCS6/CER6 increased steadily with leaf expansion, whereas expression of KCS1, KCS5, and KCS16 decreased over time. By monitoring the development of Arabidopsis leaf cuticles at three different levels, It was possible to integrate morphology with wax chemistry and gene expression analysis. The main finding was that observed shifts in compound class and chain length distributions were largely driven by differential expression of KCS genes. In particular, expression of KCS6/CER6 increased throughout leaf development, underlining the importance of this particular enzyme in orchestrating chemical shifts. Since KCS6/CER6 is known to elongate wax precursors up to C32 in leaves, it seems plausible that the observed chain length shift towards C30/C32 compounds is caused entirely by increased expression of KCS6/CER6.  However, the further correlation with a shift towards alkanes (but not other compound classes) cannot be explained by KCS6/CER6  134 expression levels alone, but instead points to preferred channeling of the products from the KCS6/CER6-containing FAE complex into the alkane pathway  In summary, the production and compositional changes of cuticular waxes during ontogenesis of Arabidopsis leaves are highly regulated and synchronized with epidermal cell expansion. In this context, it should be noted that changes in production rates must occur within a few days, if not hours, in order to be coordinated with cell expansion. Therefore, also the production of wax molecules, from initial elongation of its carbon chain to final elaboration of the functional group and export to the cuticle may take a few hours at most. Furthermore, my findings also imply that the delay between gene expression and product accumulation in the cuticle is rather small.  While it is plausible that the growing leaf needs to be protected against transpirational water loss and therefore should keep constant wax coverage throughout development, the relevance of differences in chemical composition remains unknown. Differences in chemistry could alter the overall cuticle structure and therefore physiological properties. One could argue that higher abundance of shorter chain lengths might be beneficial for expanding leaf structures. To test such ideas, differences in physiological properties of young and mature cuticles, such as transpiration rates or water repellency, should be investigated in the future. Similarly, differences in secondary cuticle functions between young and mature leaves, such as their susceptibility to pathogens or reflection of UV light, should be studied to improve our understanding how the chemical composition translates into cuticle function, and to address the question why wax composition changes during ontogenesis. The present findings of developmental variations in chemical compositions and their tight genetic regulation highlight the potential of time-resolved studies. In particular, such  135 comparisons of young, expanding and fully-grown leaves may serve to identify new wax biosynthesis genes relevant for cuticle formation only during early ontogenetic stages. Analysis of cuticular waxes of young leaves of selected Arabidopsis mutants, in comparison with the data presented here and using the methods established in my work, could identify wax phenotypes that would be missed if only analyzing mature leaves.  5.1.2 The composition of surface wax of different Arabidopsis thaliana epidermis cells Studying cuticular waxes during ontogenesis of Arabidopsis leaves showed a steady decrease of total wax coverage of wild type leaves over time, and young wild-type leaves showed a relatively high abundance of C35+ compounds, which decreased in the course of leaf expansion. On the contrary, the wax coverage on leaves of trichome-free gl1 mutants remained constant, and had an overall low abundance of C35+ compounds, which remained constant throughout leaf development. The fact that I observed a decrease of C35+ compounds on expanding wild-type leaves but not on trichome-free leaves indicates that the observed effect in the wild type is likely caused by the presence of trichomes. Since young wild-type leaves have a higher trichome density than mature leaves, the contribution of trichome wax to the overall wax mixture is higher in young leaves. This also suggests that the trichome wax is different from pavement cell wax, with a higher abundance of C35+ compounds, and differences of cuticular wax compositions between epidermal cell types therefore became a second focus point of my work. Two approaches were taken to further test for differences in wax formation and composition between epidermal cell types. First, the cuticular wax composition of the trichome-free Arabidopsis mutant gl1 was compared with composition of a trichome-rich mutant (cpc tcl1 etc1  136 etc3). The main findings were that stems of the trichome-rich mutant had a higher abundance of wax compounds with aliphatic chains of 32 or more carbons compared to stems of the trichome-free mutant. Similarly, the abundance of C32+ wax compounds was also increased in leaves of the trichome-rich mutant compared to gl1. In a second set of experiments, the cuticular wax of isolated leaf trichomes was analyzed and found to contain fewer compound classes compared to pavement cells. Most importantly, the trichome wax had a higher abundance of C32+ compounds compared to pavement cell wax (as judged by the composition of the gl1 leaf wax), thus confirming the results from both the mutant comparison and the comparative analyses of developing wild-type leaves. Taken together, the results from various experiments presented here thus clearly establish that Arabidopsis trichomes have cuticular waxes distinct from those on pavement cells. The differences in cuticular wax compositions suggest very specific functions of cuticles on both epidermal cell types. However, in order to further probe into such functional specializations, the chemical data must be compared between both cell types on a quantitative basis. However, despite the great progress in analyzing cuticular waxes at the cellular level, procedural limitations thus far made it impossible to quantify trichome wax coverage. For this, it would be necessary to measure the trichome surface area extracted for wax analysis, and thus to count the number of isolated trichomes and measure their average surface area. It may be possible to develop microscopic techniques or to adopt a cell counting system to determine trichome numbers, and careful electron microscopic surveys may provide an estimate of the trichome surface areas.  137 It will be interesting to also monitor Arabidopsis trichome wax coverage as a function of development. In the work presented here, it was found that pavement cells maintain constant wax coverage during leaf expansion, and their coverage was estimated to be roughly comparable with trichome wax coverage. It would therefore be interesting to determine if the total wax coverage of trichome cells also remains constant during surface expansion, even though they are formed very rapidly and are fully developed ca. 1.5 days after leaf initiation. While their rapid development may make it difficult to harvest sufficient numbers of young trichomes for cuticular wax analysis, the results and methods presented here for the first time enable the design of such experiments. As one promising solution to these experimental challenges, I propose that the trichomes be harvested at the leaf base of 5-day-old leaves, to select for rapidly expanding trichomes and avoid those at the adjacent leaf tip that are already mature. From the chemical results, it follows that trichome and pavement cells must have at least partially autonomous wax biosynthesis machineries, with some enzymes functioning cell type specifically, but likely also with some functioning in both cell types. To assess the likely differences in corresponding gene expression levels, I analyzed published microarray data for wax biosynthesis genes in trichomes and pavement cells. I was able to identify several wax biosynthesis genes differentially expressed in the two epidermal cell types. Interestingly, I found KCS5, KCS8 and KCS16 more highly expressed in trichomes compared to pavement cells, making them primary candidates to be involved in elongating precursors into extra-long compounds found in trichome wax (see below).  138 5.1.3 KCS16 gene characterization The goal of the fourth chapter was to test the involvement of some of the KCS gene candidates (KCS5, KCS8 and KCS16) in elongation of wax compounds beyond C32, mainly in trichomes. Among these, KCS5 has high sequence similarity with KCS6/CER6 and, based on yeast expression, similar chain length specificity as KCS6/CER6 (elongating up to C34). It is therefore likely that KCS5 and KCS6/CER6 have redundant functions, but that KCS5 active in trichomes while KCS6 functions mostly in pavement cells (but also in trichomes). However, all the current evidence taken together, KCS5 was not a good candidate for elongating wax compounds beyond C32. On the other hand, the functions or biochemical activities of both KCS8 and KCS16 were unknown, and these candidates were thus chosen for detailed characterization. Wax analysis of two confirmed homozygous T-DNA insertion lines of ksc8 did not show differences in wax compositions of stems, mature or young leaves or trichomes compared to the wild type (Appendix Figure D.1), despite the high expression of this gene in young leaves and trichomes. Thus, mutant analysis did not suggest involvement of KCS8 in the elongation of wax compounds beyond C32. However, in order to confirm that the analyzed kcs8 mutant alleles are truly null mutants, the transcript abundance will have to be tested in both alleles using semi-quantitative or quantitative RT-PCR. Perhaps KCS8 has overlapping functions with another KCS.  However, based on the strong decrease of C36 and C38 wax compounds in trichomes of ksc16 mutants, it seems unlikely that KCS8 has redundant functions with KCS16 in trichomes. Instead, the weaker phenotype of ksc16 mutants in leaves may suggest redundant function of both KCSs in pavement cells, which could be tested by analyzing cuticular waxes of mature leaves of kcs8ksc16 double mutants.  139 The remaining candidate gene, KCS16, was found expressed at overall low levels, but higher in trichomes and young leaves. Analysis of kcs16 mutants showed a drastic decrease of wax compounds derived from C36 and C38 acyl-CoAs in young leaves and isolated leaf trichomes, whereas in mature leaves the decrease was less severe. Heterologous expression of KCS16 together with KCS6/CER6 and CER26 in yeast resulted in accumulation of C36 and C38 fatty acids, which were not observed when KCS6/CER6 was expressed with CER26 alone. All these findings showed that KCS16 is elongating wax precursors, most likely C34 acyl-CoA, to C36 and C38 products, and that it is the sole enzyme performing this elongation step in trichomes. The less severe phenotype in mature leaves suggests that KCS16 is also involved in wax formation in pavement cells, but likely in parallel with another KCS. Overall, my gene characterization experiments showed that KCS16 encodes a novel Arabidopsis condensing enzyme involved in the cuticular wax biosynthesis pathway. Moreover, I found that CER6 and KCS16 together are sufficient to elongate acyl-CoAs from C26 all the way up to C38, demonstrating that a very small complement of KCSs may accomplish a large number of elongation rounds. I was able to show that KCS16 can use C34 substrates, but I cannot exclude that KCS16 is also able to use substrates with 32 or fewer carbons. However, with the currently available tools for ectopic or heterologous expression, the exact substrate specificity of this enzyme cannot be determined. Therefore, to further understand the function of KCS16, it will be important to test if the protein is interacting physically with CER26 in planta and whether, thus, CER26 is needed for the KCS16 activity. Physical interaction of proteins can be tested by a variety of methods, such as co-immunoprecipitation or using yeast two-hybrid screening (Ausubel et al., 2003). It is likely that KCS16 is interacting with CER26, as described previously for KCS6/CER6 and CER26.  140 The finding that the Arabidopsis KCS16 is highly expressed in reproductive organs and seeds raises interesting questions regarding the enzyme function in those organs. High KCS16 expression in seed coats suggests a role of the C36 and C38 wax compounds produced by this enzyme. Yet, the function of waxes in seeds coats is unknown, and it has only once been suggested that seeds contain a cuticle like structure (De Giorgi et al., 2015). High KCS16 expression levels were also found in certain parts of the flower, such as the carpel, ovary tissue and stigma (Le et al., 2010), again suggesting a specific function of the C36 and C38 products in plant reproduction. This hypothesis matches previous observations, where mutation of other wax biosynthesis genes affected the proper function of reproductive organs, through alterations in anther development, pollen tryphine layers, interrupting pollen-pistil interactions and causing male sterility (Aarts et al., 1995; Chen et al., 2003; Millar et al., 1999; Preuss et al., 1993). It therefore seems possible that extra-long wax compounds may also be relevant for plant reproduction, yet their exact function remains unknown and should be the focus of future research.  However, analysis of whole flowers would likely dilute possible ksc16 phenotypes affecting only specific parts of the flower. Therefore, wax analyses will have to be performed on single floral whorls, making it challenging to sample sufficient amounts for wax analysis by GC-MS. The higher abundance of compounds derived from C36 and C38 wax compounds in certain cell types, such as trichomes, raises the question weather they are positively correlated with certain physiological traits, such as providing a better transpiration barrier. In this context, I would expect enzymes generating longer wax compounds, such as KCS16, to be increasingly expressed under drought or high temperature conditions, yet Joubès et al. (2015) found decreased expression of KCS16 in seedlings exposed to drought conditions. Conversely, in silico data show  141 a strong increase of KCS16 expression in seedlings when exposed to drought and heat, and slightly less increased expression when exposed to cold (https://bar.utoronto.ca/eplant/). It is difficult to draw conclusions from the published data, since only developing seedlings were investigated, and it was not clear weather they featured substantial amounts of trichomes, yet KCS16 is mainly functioning in trichomes. To better understand how KCS16 expression is regulated under different environmental conditions, further studies are needed, in which isolated trichomes or mature tissues are included in the analysis.  5.1.4 Function of C36 and C38 extra-long wax compounds in Arabidopsis Here, for the first time we were able to detect and quantify low-abundance C35+ wax compounds in Arabidopsis. This required improved analytical methods and avoiding ubiquitous contamination such as paraffin, which are common in laboratory settings. C36 and C38 wax compounds had been described in other species before, such as Euphorbia characias, E. cyparissias, E. lathyris, E. niccaensis, and E. peplus (Hemmers & Gülz, 1986), Triticum aestivum, Zea mays and Lupinus angustifolius (Nadiminti et al., 2015), Miscanthus sinensis (Gao & Huang, 2013), Papaver somniferum and Eschscholzia california (Jetter & Riederer, 1996) and Austrocedrus chilensis (Richardson et al., 1998), yet the function of those extra-long wax compounds remains unknown.  What possible advantages may wax constituents with longer chain lengths confer to the plant through changes in physico-chemical structure and properties of the wax mixture of leaves and trichomes? A study on adaptation of cuticles in Austrocedrus chilensis of different climatic zones showed correlation of higher temperatures with longer alkane carbon chains (C34 – C37), and of  142 higher humidity with shorter carbon chains (high levels of C33) (Dodd et al., 1998). Similarly, in Arabidopsis increasing amounts of longer alkanes were observed under drought stress (Kosma et al., 2009), suggesting that wax compounds with longer carbon chain may provide better transpiration barriers. The cuticular changes observed in several studies (Dodd et al., 1998; Kosma et al., 2009) highlight that cuticular compositions are dynamic, changing not only during developmental stages but also as a result of changing short-term environmental conditions, such as changes and long-term conditions, such as different climatic habitats. The higher melting temperatures of longer compounds might also be an advantage in maintaining cuticle structure under high ambient temperatures. This seems especially relevant for trichomes with their higher cell surface area-to-volume ratio as compared to pavement cells, and because trichomes, protruding outwards from the leaf surface, are highly exposed to mechanical stress and extreme temperatures. Accumulation of longer alkanes in mature leaves could provide a better transpiration barrier with increasing leaf surface area and sun exposure, which should be further investigated in the future.  Future studies are needed to clarify which functional advantage longer-chain-length compounds indeed confer. For example, using Arabidopsis wax mutants, such as cer6, cer2 or cer26, which are not able to elongate wax compounds beyond C26 or C30, can be used to investigate the effect of chain length distributions on water retention or UV light reflection of isolated trichomes. Conversely, KCS16 will prove to be a useful tool to increase the abundance of extra-long wax compounds in Arabidopsis leaves, to analyze changes of physical properties. Unfortunately, overexpression of KCS16 in leaves so far did not show substantially increased amounts of extra-long wax compounds, likely because the availability of KCS16 substrates is limited.  143 Overexpression of both, CER26 and KCS16 in wild-type Arabidopsis plants could potentially increase the substrate availability and therefore result in higher abundance on C36 and C38 derived wax compounds.   5.1.5 Alkenes as a novel compound class in Arabidopsis Throughout my research, I was able to identify and quantify alkenes in leaf and trichome wax, which had not been described in Arabidopsis before. Alkenes were described in other species, such as cucumber fruits and stems, barley leaves, tomato fruit, maize pollen or olive oil (Bianchi et al., 1990; Bortolomeazzi et al., 2001; Leide, 2008; Wang et al., 2015; Von Wettstein-Knowles, 2007). Several studies analyzed cuticular alkene structures in detail, to determine the double bond position(s), and to quantify isomers and homologous distributions. For example, the cuticular alkenes on Hordeum vulgare spikes ranged from C23 to C35 (Von Wettstein-Knowles, 2007), whereas Rosa damascena flower alkenes were found as a homologous series peaking at C29 (Wollrab, 1968). In Arabidopsis, I detected alkenes with odd carbon numbers ranging from C33 to C37. Young leaves had a higher abundance of longer alkenes compared to mature leaves, similar as for alkanes. Moreover, both alkanes and alkenes were affected equally in ksc16 mutants, suggesting that this KCS enzyme elongates both compound classes. The double bond positions of wax alkenes have rarely been investigated. In one detailed study, the cuticular alkenes of H. vulgare were found to be predominantly 9-ene isomers, accompanied by 11-, 13-, 15-enes (Von Wettstein-Knowles, 2007). In contrast, the flower wax of R. damascena was reported to contain mainly 5- and 7-enes (Wollrab, 1968). In light of the sparse  144 information of isomer distributions of wax alkenes, it will be important to analyze the double bond position(s) of the Arabidopsis alkenes. The identification of alkenes in the model species Arabidopsis raises interesting questions regarding their biosynthesis and function. The pathway leading to alkenes, and thus the sequence of reactions generating the double bond and elongating the hydrocarbon chain remain unknown. Double bond formation could occur before, during or after elongation, and with or without the help of a dedicated enzyme such as a desaturase. Several desaturase gene candidates are available in Arabidopsis (Heilmann, 2004; Kachroo et al., 2007; Shanklin & Somerville, 1991). As an alternative, Von Wettstein-Knowles (2007) suggested that a trans-Δ2 double bond may be formed during a FAE-catalyzed elongation cycle, and may be retained either if the enoyl-CoA reductase is absent or if it is incompatible with an alternative cis-Δ3 enoyl intermediate. None of the above hypotheses for alkene formation can be ruled out or favoured for Arabidopsis at this point, and therefore it will be very interesting to address wax alkene biosynthesis in this species in the future. As one approach, I suggest investigating differences in alkene abundance in already characterized Arabidopsis wax mutants, since their alkene profiles have not been reported to date. Analysis of alkenes in elongation mutants such as cer6, cer2, or cer26 could test if alkenes are elongated through the same pathway as saturated wax compounds. Since the Arabidopsis wax alkenes have odd carbon numbers similar to the accompanying alkanes, it would be interesting to see if both compound classes are synthesized by the same decarbonylation pathway enzymes CER3 and CER1, and therefore alkene analysis of cer3 mutants should also be attempted.  145 The narrow chain length range of alkenes (C33 to C37) is intriguing, and the question arises what functions those particular alkenes have. The higher abundance of alkenes in trichomes and young leaves compared to mature leaves indicates a cell type-specific function. The kink in the aliphatic chain caused by the relatively rigid double bond could alter the cuticle structure by disturbing crystalline packing order and enhancing amorphous domains within the wax mixture. One could speculate that adding unsaturated compounds such as alkenes might thus affect the transpiration barrier and change the physiological function of the cuticle (Reynhardt, 1997; Riederer & Schreiber, 1995). Both the crystallinity and the water permeability of waxes containing various amounts of alkenes could be assessed experimentally in the future.  5.2 CONCLUSION In conclusion, my experiments analyzing the Arabidopsis cuticle both chemically and genetically with increased temporal and spatial resolution led to the finding that a novel KCS enzyme is involved in the elongation pathway of wax compounds. Diverse further investigations may now build on the newly developed protocols and my results, and many new questions can be addressed in the near future.  One example for a new research direction would be to expand the investigation of different epidermal cell types to guard cells. After I determined differences in cuticular wax compositions between trichome and pavement cells, the question arises if guard cells also have a different composition. Analysis of the Arabidopsis cer1 and cer6 mutants showed a 43 % and 30 % increase in stomata abundance, respectively (Gray et al., 2000). In the same study, High Carbon Dioxide  146 (HIC), and novel Arabidopsis gene expressed exclusively in guard cells, was reported to encode a KCS, and hic mutants also showed an increase in stomata numbers (Gray et al., 2000). This indicates that cuticular wax composition influences stomata development, yet the mechanism underlying this phenomenon is unknown. There is some evidence that guard cells may have a distinct wax composition, based on differences of fluorescence emission after UV-light exposure between guard and other epidermal cells, possibly caused by wax-bound phenolic compounds or a thicker epicuticular wax layer found on guard cells. Wax removal led to decreased fluorescence intensities from guard cells on Olea europaea, Vici faba, and Triticum aestivum leaves (Karabourniotis, 2001). However, direct evidence testing whether guard cells have wax compositions differing from neighboring pavement cells is missing to date. In my studies, it was not possible to distinguish between the wax composition of pavement and guard cells, due to experimental limitations. The number and size of guard cells is much smaller compared to pavement cells, and the contribution of guard cell wax to the overall wax mixture is necessarily minor. Therefore, chemical differences could not be concluded based on the small differences in guard cell numbers observed in various comparative studies performed in the course of this work. Instead, other mutants with dramatic alterations in stomata density will have to be used to distinguish between guard cell and pavement cell wax. In Arabidopsis, several mutants with increased stomata density were described previously, including tmm (Geisler et al., 2000; Nadeau & Sack, 2002; Yang & Sack, 1995), sdd1 (Berger & Altmann, 2000; Von Groll et al., 2002), and yda (Gray & Hetherington, 2004; Hunt & Gray, 2009). Similarly, overexpression of EPIDERMAL PATTERNING FACTOR (EPF) genes also increases the density of stomata on the abaxial side of Arabidopsis leaves, while epf mutants are lacking stomata almost completely and have an increased number of pavement cells instead (Hara et al., 2007; Hunt & Gray, 2009).  147 Comparative analyses of the cuticular waxes on these Arabidopsis mutants and overexpressor with vastly differing stomata numbers may well hold the answer to the question on guard cell wax autonomy and also function.     148 Bibliography Aarts, M. G., Keijzer, C. J., Stiekema, W. J., & Pereira,  A. (1995). Molecular characterization of the CER1 gene of Arabidopsis involved in epicuticular wax biosynthesis and pollen fertility. The Plant Cell, 7(12), 2115–27.  Alonso, J. M., Stepanova, A. N., Leisse, T. J., Kim, C. J., Chen, H., Shinn, P., Stevenson, D. K., et al. (2016). Genome-wide insertional mutagenesis of Arabidopsis thaliana. American Association for the Advancement of Science, (301)5633, 653–657. Andriankaja, M., Dhondt, S., DeBodt, S., Vanhaeren, H., Coppens, F., DeMilde, L., Mühlenbock, P., et al. (2012). Exit from proliferation during leaf development in Arabidopsis thaliana: A Not-So-Gradual Process. Developmental Cell, 22(1), 64–78. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A, Struhl, K., et al. (2003). Current Protocols in Molecular Biology. John Wiley & Sons, Inc., New Jersey, USA. Azachi, M., Sadka, A., Fisher, M., Goldshlag, P., Gokhman, I., Zamir, A., Chemistry, B., et al. (2002). Salt induction of fatty acid elongase and membrane lipid modifications in the extreme halotolerant alga Dunaliella salina. Plant Physilogy, 129(July), 1320–1329. Aznar-Moreno, J. A, Venegas Calerón, M., Martínez-Force, E., Garcés, R., Mullen, R., Gidda, S. K., & Salas, J. J. (2014). Sunflower (Helianthus annuus) long-chain acyl-coenzyme A synthetases expressed at high levels in developing seeds. Physiologia Plantarum, 150(3), 363–373. Bach, L., Michaelson, L., Haslam, R., Bellec, Y., Gissot, L., Marion, J., & Da Costa, M. (2008). The very-long-chain hydroxy fatty acyl-CoA dehydratase PASTICCINO2 is essential and limiting for plant development. Proceedings of the Nationa Academy of Sciences of the United States of America, 105(38), 14727–14731. Balbontín, C., Ayala, H., Rubilar, J., Cote, J., & Figueroa, C. R. (2014). Transcriptional analysis of cell wall and cuticle related genes during fruit development of two sweet cherry cultivars with contrasting levels of cracking tolerance. Chilean Journal of Agricultural Research, 74(2), 162–169. Barthlott, W., & Neinhuis, C. (1997). Purity of the sacred lotus, or escape from contamination in biological surfaces. Planta, 202(1), 1–8. Bauer, S., Schulte, E., & Thier, H.-P. (2004). Composition of the surface waxes from bell pepper  149 and eggplant. European Food Research and Technology, 220(1), 5–10.  Bazzaz, F. A, Chiariello, N. R., Coley, P. D., & Pitelka, L. F. (1987). Resources to allocating defense and new assessments of the costs and benefits of allocation patterns in plants are relating ecological roles to resource use. American Institut of Biological Sciences, 37(1), 58–67. Beaudoin, F. R., Gable, K., Sayanova, O., Dunn, T., & Napier, J. A. (2002). A Saccharomyces cerevisiae gene required for heterologous fatty acid elongase activity encodes a microsomal β-keto-reductase. Journal of Biological Chemistry, 277(13), 11481–11488. Beaudoin, F., Wu, X., Li, F., Haslam, R. P., Markham, J. E., Zheng, H., Napier, J. A., et al. (2009). Functional characterization of the Arabidopsis beta-ketoacyl-coenzyme A reductase candidates of the fatty acid elongase. Plant Physiology, 150(3), 1174–1191. Beisson, F., Koo, A. J. K., Ruuska, S., Pollard, M., Thelen, J. J., Paddock, T., Salas, J., et al. (2003). Arabidopsis genes involved in acyl lipid metabolism. Plant Physiology, 132(3), 681–697. Benjamini, Y., & Hochberg, Y. (1995). Controlling the False Discovery Rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B, 57(1), 289–300.  Bennett, R. N., & Wallsgrove, R. M. (2016). Secondary metabolites in plant defence mechanisms. New Phytologist, 127(72), 617–633. Berger, D., & Altmann, T. (2000). A subtilisin-like serine protease involved in the regulation of stomatal density and distribution in Arabidopsis thaliana. Genes & Development, 14(9), 1119–1131.  Bernard, A., Domergue, F., Pascal, S., Jetter, R., Renne, C., Faure, J.-D., Haslam, R. P., et al. (2012). Reconstitution of plant alkane biosynthesis in yeast demonstrates that Arabidopsis ECERIFERUM1 and ECERIFERUM3 are core components of a very-long-chain alkane synthesis complex. The Plant Cell, (24)7, 1–14.  Bessire, M., Borel, S., Fabre, G., Carraça, L., Efremova, N., Yephremov, A., Cao, Y., et al. (2011). A member of the PLEIOTROPIC DRUG RESISTANCE family of ATP binding cassette transporters is required for the formation of a functional cuticle in Arabidopsis. The Plant Cell, 23(5), 1958–70.  Bessire, M., Chassot, C., Jacquat, A.-C., Humphry, M., Borel, S., MacDonald-Comber Petetot,  150 J., Metraux, J.-P., et al. (2007). A permeable cuticle in Aarabidopsis leads to a strong resitance to Botrytis cinera. European Molecular Biology Organization, 26(8), 2158–2168. Bianchi, G., Murelli, C., & Ottaviano, E. (1990). Maize pollen lipids. Phytochemistry, 29(3), 739–744. Bird, D., Beisson, F., Brigham, A., Shin, J., Greer, S., Jetter, R., Kunst, L., et al. (2007). Characterization of Arabidopsis ABCG11/WBC11, an ATP binding cassette (ABC) transporter that is required for cuticular lipid secretion. The Plant Journal, 52(3), 485–98. Blacklock, B. J., & Jaworski, J. G. (2002). Studies into factors contributing to substrate specificity of membrane-bound 3-ketoacyl-CoA synthases. European Journal of Biochemistry, 269(19), 4789–4798. Blacklock, B. J., & Jaworski, J. (2006). Substrate specificity of Arabidopsis 3-ketoacyl-CoA synthases. Biochemical and Biophysical Research Communications, 346, 583–590. Bonaventure, G., Beisson, F., Ohlrogge, J., & Pollard, M. (2004). Analysis of the aliphatic monomer composition of polyesters associated with Arabidopsis epidermis: occurrence of octadeca-cis-6, cis-9-diene-1,18-dioate as the major component. The Plant Journal, 40(6), 920–30.  Bortolomeazzi, R., Berno, P., Pizzale, L., & Conte, L. S. (2001). Sesquiterpene, alkene, and alkane hydrocarbons in virgin olive oils of different varieties and geographical origins. Journal of Agricultural and Food Chemistry, 49(7), 3278–3283. Bourdenx, B., Bernard, A., Domergue, F., Pascal, S., Léger, A., Roby, D., Pervent, M., et al. (2011). Overexpression of Arabidopsis ECERIFERUM1 promotes wax very-long-chain alkane biosynthesis and influences plant response to biotic and abiotic stresses. Plant Physiology, 156(1), 29–45.  Bringe, K., Schumacher, C. F. A., Schmitz-Eiberger, M., Steiner, U., & Oerke, E.-C. (2006). Ontogenetic variation in chemical and physical characteristics of adaxial apple leaf surfaces. Phytochemistry, 67(2), 161–70. Buschhaus, C., & Jetter, R. (2012). Composition and physiological function of the wax layers coating Arabidopsis leaves: β-amyrin negatively affects the intracuticular water barrier. Plant Physiology, 160(2), 1120–9.  Byrne, M. E. (2005). Networks in leaf development. Current Opinion in Plant Biology, 8(1), 59–66.  151 Cheesbrough, T. M., & Kolattukudy, P. E. (1984). Alkane biosynthesis by decarbonylation of aldehydes catalyzed by a particulate preparation from Pisum sativum. Proceedings of the National Academy of Sciences of the United States of America, 81(21), 6613–6617. Chen, X., Goodwin, S. M., Boroff, V. L., Liu, X., & Jenks, M. A. (2003). Cloning and characterization of the WAX2 Gene of Arabidopsis involved in cuticle membrane and wax production. The Plant Cell, 15(5), 1170–1185. Chu, C.-C., Freeman, T. P., Buckner, J. S., Henneberry, T. J., Nelson, D. R., & Natwick, E. T. (2001). Susceptibility of upland cotton cultivars to Bemisia tabaci Biotype B (Homoptera: Aleyrodidae) in relation to leaf age and trichome density. Annals of the Entomological Society of America, 94(5), 743–749. Clough, S. J., & Bent, A. F. (1998). Floral dip: A simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant Journal, 16, 735–743. Coley, P. D. (1985). Resource availability and plant antiherbivore defense. Advancement Of Science, 230(4728), 895–899. Debono, A., Yeats, T. H., Rose, J. K. C., Bird, D., Jetter, R., Kunst, L., & Samuels, L. (2009). Arabidopsis LTPG is a glycosylphosphatidylinositol-anchored lipid transfer protein required for export of lipids to the plant surface. The Plant Cell, 21(4), 1230–8. Deising, H. B., Werner, S., & Wernitz, M. (2000). The role of fungal appressoria in plant infection. Microbes and Infection, 2(13), 1631–1641. Deshmukh, A. P., Simpson, A. J., Hadad, C. M., & Hatcher, P. G. (2005). Insights into the structure of cutin and cutan from Agave americana leaf cuticle using HRMAS NMR spectroscopy. Organic Geochemistry, 36(7), 1072–1085. Dodd, R. S., Rafii, Z. A., & Power, A. B. (1998). Ecotypic adaptation in Austrocedrus chilensis in cuticular hydrocarbon composition. New Phytologist, 138(4), 699–708. Domergue, F., Chevalier, S., Creach, A., Cassagne, C., & Lessire, R. (2000). Purification of the acyl-CoA elongase complex from developing rapeseed and characterization of the 3-ketoacyl-CoA synthase and the 3-hydroxyacyl-CoA dehydratase. Lipids, 35(5), 487–494.  Domergue, F., Vishwanath, S. J., Joubès, J., Ono, J., Lee, J. A., Bourdon, M., Alhattab, R., et al. (2010). Three Arabidopsis fatty acyl-coenzyme A reductases, FAR1, FAR4, and FAR5, generate primary fatty alcohols associated with suberin deposition. Plant Physiology, 153(4), 1539–54.   152 Domínguez, E., López-Casado, G., Cuartero, J., & Heredia, A. (2008). Development of fruit cuticle in cherry tomato (Solanum lycopersicum). Functional Plant Biology, 35(5), 403–411. Donnelly, P. M., Bonetta, D., Tsukaya, H., Dengler, R. E., & Dengler, N. G. (1999). Cell cycling and cell enlargement in developing leaves of Arabidopsis. Developmental Biology, 215(2), 407–19. Dunn, T. M., Lynch, D. V., Michaelson, L. V., & Napier, J. A. (2004). A post-genomic approach to understanding sphingolipid metabolism in Arabidopsis thaliana. Annals of Botany, 93(5), 483–497. Ebert, B., Zöller, D., Erban, A., Fehrle, I., Hartmann, J., Niehl, A., Kopka, J., et al. (2010). Metabolic profiling of Arabidopsis thaliana epidermal cells. Journal of Experimental Botany, 61(5), 1321–1335. Efroni, I., Eshed, Y., & Lifschitz, E. (2010). Morphogenesis of simple and compound leaves: A critical review. The Plant Cell, 22(4), 1019–1032. Esch, J. J., Chen, M., Sanders, M., Hillestad, M., Ndkium, S., Idelkope, B., Neizer, J., et al. (2003). A contradictory GLABRA3 allele helps define gene interactions controlling trichome development in Arabidopsis. Development, 130, 5885–5894. Espelie, K. E., Davis, R. W., & Kolattukudy, P. E. (1980). Plant inner seed coat of Grapefruit. Planta, 511, 498–511. Fabre, G., Garroum, I., Mazurek, S., Daraspe, J., Mucciolo, A., Sankar, M., Humbel, B. M., et al. (2016). The ABCG transporter PEC1/ABCG32 is required for the formation of the developing leaf cuticle in Arabidopsis. New Phytologist, 209(1), 192–201. Fehling, E., & Mukherjee, K. D. (1991). Acyl-CoA elongase from a higher plant (Lunaria annua): metabolic intermediates of very-long-chain acyl-CoA products and substrate specificity. Biochimica et Biophysica Acta, 1082(3), 239–246. Fiebig, A., Mayfield, J. A., Miley, N. L., Chau, S., Fischer, R. L., & Preuss, D. (2000). Alterations in CER6, a gene identical to CUT1, differentially affect long-chain lipid content on the surface of pollen and stems. The Plant Cell, 12(10), 2001–2008. Fourmann, M., Barret, P., Renard, M., Pelletier, G., Delourme, R., & Brunel, D. (1998). The two genes homologous to Arabidopsis FAE1 co-segregate with the two loci governing erucic acid content in Brassica napus. Theoretical and Applied Genetics, 96(6-7), 852–858.   153 Fraenkel, G. S. (2016). The raison d ’ Être of secondary plant substances. American Association for the Advancemnet of Science, 129(3361), 1466–1470. Franke, R., Briesen, I., Wojciechowski, T., Faust, A., Yephremov, A., Nawrath, C., & Schreiber, L. (2005). Apoplastic polyesters in Arabidopsis surface tissues - A typical suberin and a particular cutin. Phytochemistry, 66(22), 2643–2658. Franke, R., Höfer, R., Briesen, I., Emsermann, M., Efremova, N., Yephremov, A., & Schreiber, L. (2009). The DAISY gene from Arabidopsis encodes a fatty acid elongase condensing enzyme involved in the biosynthesis of aliphatic suberin in roots and the chalaza-micropyle region of seeds. The Plant Journal  , 57(1), 80–95. Gao, L., & Huang, Y. (2013). Inverse gradients in leaf wax δD and δ13C values along grass blades of Miscanthus sinensis: Implications for leaf wax reproduction and plant physiology. Oecologia, 172(2), 347–357. Geisler, M., Nadeau, J., & Sack, F. D. (2000). Oriented asymmetric divisions that generate the stomatal spacing pattern in arabidopsis are disrupted by the too many mouths mutation. The Plant Cell, 12(11), 2075–2086. Gibbons, J. D., & Chakraborti, S. (2011). International Encyclopedia of Statistical Science. Springer Berlin Heidelberg, 977–979. De Giorgi, J., Piskurewicz, U., Loubery, S., Utz-Pugin, A., Bailly, C., Mène-Saffrané, L., & Lopez-Molina, L. (2015). An endosperm-associated cuticle is required for Arabidopsis seed viability, dormancy and early control of germination. PLoS Genetics, 11(12), 1–32. Girard, A.-L., Mounet, F., Lemaire-Chamley, M., Gaillard, C., Elmorjani, K., Vivancos, J., Runavot, J.-L., et al. (2012). Tomato GDSL1 is required for cutin deposition in the fruit cuticle. The Plant Cell, 24(7), 3119–34.  Glover, B. J. (2000). Differentiation in plant epidermal cells. Journal of Experimental Botany, 51(344), 497–505. Graça, J., Schreiber, L., Rodrigues, J., & Pereira, H. (2002). Glycerol and glyceryl esters of ω-hydroxyacids in cutins. Phytochemistry, 61(2), 205–215. Granier, C., Massonnet, C., Turc, O., Muller, B., Chenu, K., & Tardieu, F. (2002). Individual leaf development in Arabidopsis thaliana: A stable thermal-time-based programme. Annals of Botany, 89(5), 595–604. Gray, J. E., Holroyd, G. H., van der Lee, F. M., Bahrami, A. R., Sijmons, P. C., Woodward, F. I.,  154 Schuch, W., et al. (2000). The HIC signalling pathway links CO2 perception to stomatal development. Nature, 408(6813), 713–716. Gray, J. E., & Hetherington, A. M. (2004). Plant development: YODA the stomatal switch. Current Biology, 14(12), 488–490. Greer, S., Wen, M., Bird, D., Wu, X., Samuels, A. L., Kunst, L., & Jetter, R. (2007). The cytochrome P450 enzyme CYP96A15 is the midchain alkane hydroxylase responsible for formation of secondary alcohols and ketones in stem cuticular wax of Arabidopsis. Plant Physiology, 145(3), 653–667.  Von Groll, U., Berger, D., & Altmann, T. (2002). The subtilisin-like serine protease SDD1 mediates cell-to-cell signaling during Arabidopsis stomatal development. The Plant Cell, 14(7), 1527–39. Guimil, S., & Dunand, C. (2007). Cell growth and differentiation in Arabidopsis epidermal cells. Journal of Experimental Botany, 58(14), 3828–3240. Gülza, P., Prasad, R. B. N., & Müllera, E. (1991). Surface structure and chemical composition of epicuticular waxes during leaf development of Tilia tomentosa. Moench, Verlag der Zeitschrift für Naturforschung (47), 190–196. Guo, H.-S., Zhang, Y.-M., Sun, X.-Q., Li, M.-M., Hang, Y.-Y., & Xue, J.-Y. (2015). Evolution of the KCS gene family in plants: the history of gene duplication, sub/neofunctionalization and redundancy. Molecular Genetics and Genomics, (291)2, 739–752. Gupta, N. S., Collinson, M. E., Briggs, D. E. G., Evershed, R. P., & Pancost, R. D. (2016). Reinvestigation of the occurrence of cutan in plants: Implications for the leaf fossil records. Paleobiolgy, (32)3, 432–449. Haas, K., & Schönherr, J. (1979). Composition of soluble cuticular lipids and water permeability of cuticular membranes from Citrus leaves. Planta, 146(4), 399–403. Han, G., Gable, K., Kohlwein, S. D., Beaudoin, F., Napier, J. A., & Dunn, T. M. (2002). The Saccharomyces cerevisiae YBR159w gene encodes the 3-ketoreductase of the microsomal fatty acid elongase. Journal of Biological Chemistry, 277(38), 35440–35449. Haslam, T. M., Mañas Fernández, A., Zhao, L., & Kunst, L. (2012). Arabidopsis ECERIFERUM2 is a component of the fatty acid elongation machinery required for fatty acid extension to exceptional lengths. Plant Physiology.  Haslam, T. M., & Kunst, L. (2013). Extending the story of very-long-chain fatty acid elongation.  155 Plant Science, 210, 93–107. Haslam, T. M., Haslam, R., Thoraval, D., Pascal, S., Delude, C., Domergue, F., Fernández, A. M., et al. (2015). ECERIFERUM2-LIKE proteins have unique biochemical and physiological functions in very-long-chain fatty acid elongation. Plant Physiology, 167(3), 682–92. Heilmann, I. (2004). Identification of the Arabidopsis palmitoyl-monogalactosyldiacylglycerol  7-desaturase gene FAD5, and effects of plastidial retargeting of Arabidopsis desaturases on the fad5 mutant phenotype. Plant Physiology, 136(4), 4237–4245. Hemmers, H; Gulz, P. G. (1986). Waxes from leaves of five hemmers Euphorbia species. Phytochemistry, 25(9), 2103–2107. Herman, P. L., & Marks, M. D. (1989). Isolation and Complementation of the GLABROUS1 Gene. Plant Molecular Biology, 1, 1051–1055. Herms, D. A. (2016). The dilemma of plants: to grow or defend. The Quarterly Review of Biology, 67(3), 283–335. Hooker, T. S., Millar, A. A., & Kunst, L. (2002). Significance of the expression of the CER6 condensing enzyme for cuticular wax production in Arabidopsis. Plant Physiology 129(August), 1568–1580. Hülskamp, M. (2004). Plant trichomes: a model for cell differentiation. Nature Reviews, Molecular Cell Biology, 5(6), 471–80.  Hülskamp, M., Kopczak, S. D., Horejsi, T. F., Kihl, B. K., & Pruitt, R. E. (1995). Identification of genes required for pollen-stigma recognition in Arabidopsis thaliana. The Plant Journal, 8(5), 703–714. Hunt, L., & Gray, J. E. (2009). The signaling peptide EPF2 controls asymmetric cell divisions during stomatal development. Current Biology, 19(10), 864–869. Isaacson, T., Kosma, D. K., Matas, A. J., Buda, G. J., He, Y., Yu, B., Pravitasari, A., et al. (2009). Cutin deficiency in the tomato fruit cuticle consistently affects resistance to microbial infection and biomechanical properties, but not transpirational water loss. The Plant Journal, 60(2), 363–377. Ishida, T., Kurata, T., Okada, K., & Wada, T. (2008). A genetic regulatory network in the development of trichomes and root hairs. Annual Review of Plant Biology, 59(1), 365–386.  156 Jakoby, M. J., Falkenhan, D., Mader, M. T., Brininstool, G., Wischnitzki, E., Platz, N., Hudson, A., et al. (2008). Transcriptional profiling of mature Arabidopsis trichomes reveals that NOECK encodes the MIXTA-like transcriptional regulator MYB106. Plant Physiology, 148(3), 1583–1602. Jeffree, C. E. (2006). The fine structure of the plant cuticle. Annual Plant Reviews Volume 23: Biology of the Plant Cuticle, 11–125. Jenks, M. A., Rashotte,  A. M., Tuttle, H. A., & Feldmann, K. A. (1996). Mutants in Arabidopsis thaliana altered in epicuticular wax and leaf morphology. Plant Physiology, 110(2), 377–385. Jenks, M. A., Tuttle, H. A., Eigenbrode, S. D., & Feldmann, K. A. (1995). Leaf epicuticular waxes of the ECERIFERUM mutants in Arabidopsis. Plant Physiology, 108(1), 369–377. Jenks, M. A., Tuttle, H. A., & Feldmann, K. A. (1996). Changes in epicuticular waxes on wildtype and ECERIFERUM mutants in Arabidopsis during development. Phytochemistry, 42(1), 29–34. Jessen, D., Olbrich, A., Knüfer, J., Krüger, A., Hopert, M., Pole, A., & Fulda, M. (2011). Combines activity of LACS1 and LACS4 is required for proper pollen coat fromation in Arabidopsis. The Plant Journal, 68, 715–726. Jetter, R., & Riederer, M. (1996). Cuticular waxes from the leaves and fruit capsules of eight Papaveraceae species. Canadian Journal of Botany, 74(3), 419–430. Jetter, R., & Schäffer, S. (2001). Chemical composition of the Prunus laurocerasus leaf surface. Dynamic changes of the epicuticular wax film during leaf development. Plant Physiology, 126(4), 1725–37. Jetter, R., & Riederer, M. (2015). Localization of the transpiration barrier in the epi- and intracuticular waxes of eight plant species: water transport resistances are associated with fatty acyl rather than alicyclic components. Plant Physiology, 170(2), 921–934. Joubes, J., Raffalele, S., Bourdenx, B., Garcia, C., Laroche-Traineau, J., Morea, P., Domergue, F., et al. (2008). The VLCFA elongase gene family in Arabidopsis thaliana: phylogenetic analysis, 3D modelling and expression profiling. Plant Molecular Biology, 67, 547–566. Kachroo, A., Shanklin, J., Whittle, E., Lapchyk, L., Hildebrand, D., & Kachroo, P. (2007). The Arabidopsis stearoyl-acyl carrier protein-desaturase family and the contribution of leaf isoforms to oleic acid synthesis. Plant Molecular Biology, 63(2), 257–271.  157 Kajikawa, M., Yamaoka, S., Yamato, K. T., Kanamaru, H., Sakuradani, E., Shimizu, S., Fukuzawa, H., et al. (2003). Functional analysis of a beta-ketoacyl-CoA synthase gene, MpFAE2, by gene silencing in the liverwort Marchantia polymorpha L. Bioscience, Biotechnology, and Biochemistry, 67(3), 605–612.  Karabourniotis, G. (2001). Epicuticular phenolics over guard cells: exploitation for in situ stomatal counting by fluorescence microscopy and combined image analysis. Annals of Botany, 87(5), 631–639. Kearns, E. V., & Assmann, S. M. (1993). The guard cell-environment connection. Plant Physiology, 102, 711–715. Kim, M.-S., Shim, K.-B., Park, S.-H., & Kim, K.-S. (2009). Changes in cuticular waxes of developing leaves in sesame (Sesamum indicum L.). Journal of Crop Science and Biotechnology, 12(3), 161–167. Kim, H., Lee, S. B., Kim, H. J., Min, M. K., Hwang, I., & Suh, M. C. (2012). Characterization of glycosylphosphatidylinositol-anchored lipid transfer protein 2 (LTPG2) and overlapping function between LTPG/LTPG1 and LTPG2 in cuticular wax export or accumulation in Arabidopsis thaliana. Plant & Cell Physiology, 53(8), 1391–403. Kim, J., Jung, J. H., Lee, S. B., Go, Y. S., Kim, H. J., Cahoon, R., Markham, J. E., et al. (2013). Arabidopsis 3-ketoacyl-coenzyme a synthase9 is involved in the synthesis of tetracosanoic acids as precursors of cuticular waxes, suberins, sphingolipids, and phospholipids. Plant Physiology, 162(2), 567–80. Koch, K., & Ensikat, H.-J. (2008). The hydrophobic coatings of plant surfaces: epicuticular wax crystals and their morphologies, crystallinity and molecular self-assembly. Micron, 39(7), 759–72.  Kolattukudy, P. E. (1980). Biopolyester membranes of plants: cutin and suberin. Science, 208(4447), 990–1000. Kosma, D. K., Bourdenx, B., Bernard, A., Parsons, E. P., Lü, S., Joubès, J., & Jenks, M. A. (2009). The impact of water deficiency on leaf cuticle lipids of Arabidopsis. Plant Physiology, 151(4), 1918–29. Kosma, D. K., Parsons, E. P., Isaacson, T., Lü, S., Rose, J. K. C., & Jenks, M. A. (2010). Fruit cuticle lipid composition during development in tomato ripening mutants. Physiologia Plantarum, 139(1), 107–117.  158 Kunst, L., & Samuels, A. L. (2003). Biosynthesis and secretion of plant cuticular wax. Progress in Lipid Research, 42, 51–80.  Kunst, L., & Samuels, A. L. (2009). Plant cuticles shine: advances in wax biosynthesis and export. Current Opinion in Plant Biology, 12(6), 721–7.  Kurdyukov, S., Faust, A., Nawrath, C., Bär, S., Voisin, D., Efremova, N., Franke, R., et al. (2006). The epidermis-specific extracellular BODYGUARD controls cuticle development and morphogenesis in Arabidopsis. The Plant Cell, 18(February), 321–339. Lai, C., Kunst, L., & Jetter, R. (2007). Composition of alkyl esters in the cuticular wax on inflorescence stems of Arabidopsis thaliana cer mutants. The Plant Journal, 50(2), 189–96.  Lassner, M. W., Lardizabal, K., & Metz, J. G. (1996). A jojoba beta-Ketoacyl-CoA synthase cDNA complements the canola fatty acid elongation mutation in transgenic plants. The Plant Cell, 8(2), 281–92. Le, B. H., Cheng, C., Bui, A. Q., Wagmaister, J. a, Henry, K. F., Pelletier, J., Kwong, L., et al. (2010). Global analysis of gene activity during Arabidopsis seed development and identification of seed-specific transcription factors. Proceedings of the National Academy of Sciences of the United States of America, 107(18), 8063–70.  Lee, S. B., Go, Y. S., Bae, H.-J., Park, J. H., Cho, S. H., Cho, H. J., Lee, D. S., et al. (2009). Disruption of glycosylphosphatidylinositol-anchored lipid transfer protein gene altered cuticular lipid composition, increased plastoglobules, and enhanced susceptibility to infection by the fungal pathogen Alternaria brassicicola. Plant Physiology, 150(1), 42–54.  Lee, S. B., & Suh, M. C. (2015). Advances in the understanding of cuticular waxes in Arabidopsis thaliana and crop species. Plant Cell Reports, 34(4), 557–572. Lee, S.-B., Jung, S.-J., Go, Y.-S., Kim, H.-U., Kim, J.-K., Cho, H.-J., Park, O. K., et al. (2009). Two Arabidopsis 3-ketoacyl CoA synthase genes, KCS20 and KCS2/DAISY, are functionally redundant in cuticular wax and root suberin biosynthesis, but differentially controlled by osmotic stress. The Plant Journal, 60(3), 462–75.  Leide, J. (2008). Cuticular wax biosynthesis of Lycopersicon esculentum and its impact on transpiration barrier properties during fruit development (Doctoral dissertation). Bayerischen Julius-Maximilians-Universität Würzburg. Leide, J., Hildebrandt, U., Reussing, K., Riederer, M., & Vogg, G. (2007). The developmental pattern of tomato fruit wax accumulation and its impact on cuticular transpiration barrier  159 properties: effects of a deficiency in a beta-ketoacyl-coenzyme A synthase (LeCER6). Plant Physiology, 144(3), 1667–1679. Lessire, R., Bessoule, J. J., & Cassagne, C. (1985). Solubilization of C18-Co A and C20-CoA elongases from Allium porrum L. epidermal cell microsomes. FEBS Letters, 187(2), 314–320. Liakopoulos G., Stavrianakou S., Karabourniotis G. (2001). Analysis of epicuticular phenolics in Prunus persica and Olea europaea leaves: evidence for the chemical origin of the UV-induced blue fluorescence of stomata. Annals of Botany 87, 641–648. Li-Beisson, Y., Shorrosh, B., Beisson, F., Andersson, M. X., Arondel, V., Bates, P. D., Baud, S., et al. (2013). Acyl-lipid metabolism. The Arabidopsis book/American Society of Plant Biologists, volume 11: e0161. Li, F., Wu, X., Lam, P., Bird, D., Zheng, H., Samuels, A. L., Jetter, R., et al. (2008). Identification of the wax ester synthase/acyl-coenzyme A: diacylglycerol acyltransferase WSD1 required for stem wax ester biosynthesis in Arabidopsis. Plant Physiology, 148(1), 97–107.  Lolle, S. J., Berlyn, G. P., Engstrom, E. M., Krolikowski, K. a, Reiter, W. D., & Pruitt, R. E. (1997). Developmental regulation of cell interactions in the Arabidopsis fiddlehead-1 mutant: a role for the epidermal cell wall and cuticle. Developmental Biology, 189(2), 311–21.  Lü, S., Song, T., Kosma, D., Parson, E., Rowland, O., & Jenks, M. (2009). Arabidopsis CER8 encodes LONG-CHAIN ACYL-COA SYNTHETASE 1(LACS1) that has overlapping functions with LACS in plant wax cutin synthesis. The Plant Journal, 59, 553–564.  Van Maarseveen, C., Han, H., & Jetter, R. (2009). Development of the cuticular wax during growth of Kalanchoe daigremontiana (Hamet et Perr. de la Bathie) leaves. Plant, Cell and Environment, 32(1), 73–81. Marga, F., Pesacreta, T. C., & Hasenstein, K. H. (2001). Biochemical analysis of elastic and rigid cuticles of Cirsium horridulum. Planta, 213(6), 841–848. Marks, M. D. (1997). Molecular Genetic analysis of trichome development in Arabidopsis. Annual Review of Plant Physiology and Plant Molecular Biology, 48, 137–163. Marks, M. D., Betancur, L., Gilding, E., Chen, F., Bauer, S., Wenger, J. P., Dixon, R. a., et al. (2008). A new method for isolating large quantities of Arabidopsis trichomes for  160 transcriptome, cell wall and other types of analyses. The Plant Journal, 56(3), 483–492. Marks, M. D., Wenger, J. P., Gilding, E., Jilk, R., & Dixon, R. a. (2009). Transcriptome analysis of arabidopsis wild-type and gl3-sst sim trichomes identifies four additional genes required for trichome development. Molecular Plant, 2(4), 803–822. Mauricio, R. (2005). Ontogenetics of QTL: The genetic architecture of trichome density over time in Arabidopsis thaliana. Genetica, 123(1-2), 75–85. McFarlane, H. E., Shin, J. J. H., Bird, D. a, & Samuels, A. L. (2010). Arabidopsis ABCG transporters, which are required for export of diverse cuticular lipids, dimerize in different combinations. The Plant Cell, 22(9), 3066–75.  McFarlane, H. E., Watanabe, Y., Yang, W., Huang, Y., Ohlrogge, J., & Samuels, A. L. (2014). Golgi- and trans-Golgi network-mediated vesicle trafficking is required for wax secretion from epidermal cells. Plant Physiology, 164(3), 1250–1260. Merchant, S. S., Prochnik, S. E., Vallon, O., Harris, E. H., Karpowicz, J., Witman, G. B., Terry, A., et al. (2007). The Chlamydomonas genome reveals the evolution of key animal and plant functions. Science, 318(5848), 245–250. Millar,  A. A., & Kunst, L. (1997). Very-long-chain fatty acid biosynthesis is controlled through the expression and specificity of the condensing enzyme. The Plant Journal, 12(1), 121–131. Millar, A. A., Clemens, S., Zachgo, S., Giblin, E. M., Taylor, D. C., & Kunst, L. (1999). CUT1, an Arabidopsis gene required for cuticular wax biosynthesis and pollen fertility, encodes a very-long-chain fatty acid condensing enzyme. The Plant Cell, 11(5), 825–38.  Nadeau, J. a., & Sack, F. D. (2002). Stomatal development in Arabidopsis. The Arabidopsis book/American Society of Plant Biologists volume 1: e0066. Nadiminti, P. P., Rookes, J. E., Boyd, B. J., & Cahill, D. M. (2015). Confocal laser scanning microscopy elucidation of the micromorphology of the leaf cuticle and analysis of its chemical composition. Protoplasma, 252(6), 1475–1486. Nath, U., Crawford, B. C. W., Carpenter R.,Coen E. (2016). Genetic control of surface curvature. American Society for the Advancement of Sciences, 299(5611), 1404–1407. Panikashvili, D., Shi, J. X., Schreiber, L., & Aharoni, A. (2011). The Arabidopsis ABCG13 transporter is required for flower cuticle sectretion and patterning of the petal epidermis. New Phytologist, 190(11), 113–114.  161 Pascal, S., Bernard, A., Sorel, M., Pervent, M., Vile, D., Haslam, R. P., Napier, J. A., et al. (2013). The Arabidopsis cer26 mutant, like the cer2 mutant, is specifically affected in the very-long-chain fatty acid elongation process. The Plant Journal, 73(5), 733–746.  Paul, S., Gable, K., Beaudoin, F., Cahoon, E., Jaworski, J., Napier, J. A., & Dunn, T. M. (2006). Members of the arabidopsis FAE1-like 3-ketoacyl-CoA synthase gene family substitute for the elop proteins of Saccharomyces cerevisiae. Journal of Biological Chemistry, 281(14), 9018–9029. Peschel, S., Franke, R., Schreiber, L., & Knoche, M. (2007). Composition of the cuticle of developing sweet cherry fruit. Phytochemistry, 68(7), 1017–1025. Pfaffl, M. W. (2001). A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Research, 29(9), 2002–2007. Pfündel, E. E., Agati, G., & Cerovic, Z. G. (2006). Optical properties of plant surfaces. Biology of the Plant Cuticle (pp. 216–249). Pighin, J. A., Zheng, H., Balakshin, L. J., Goodman, I. P., Western, T. L., Jetter, R., Kunst, L., et al. (2004). Plant cuticular lipid export requires an ABC transporter. Science, 306(5696), 702–704.  Prasad, R. B. N., & Gülz, P. (1990). Developmental and seasonal variations in the epicuticular waxes of beech leaves (Fagus sylvatica L.). Zeitschrift für Naturforschung C, 45(7-8), 805–812. Preuss, D., Lemieux, B., Yen, G., & Davis, R. W. (1993). A conditional sterile mutation eliminates surface components from Arabidopsis pollen and disrupts cell signaling during fertilization. Genes and Development, 7(6), 974–985. Prochnik, S. E., Umen, J., Nedelcu, A. M., Hallmann, A., Miller, M., Nishii, I., Ferris, P., et al. (2010). Genomic analysis of organismal complexity in the multicellular green alga Volvox carteri. Science, 329(5988), 223–226. Pruitt, R. E., Vielle-Calzada, J. P., Ploense, S. E., Grossniklaus, U., & Lolle, S. J. (2000). FIDDLEHEAD, a gene required to suppress epidermal cell interactions in Arabidopsis, encodes a putative lipid biosynthetic enzyme. Proceedings of the National Academy of Sciences of the United States of America, 97(3), 1311–1316. Pulsifer, I., Kluge, S., & Rowland, O. (2012). Arabidopsis LONG-CHAIN ACYL-COA SYNTHETASE 1 (LACS1), LACS2, and LACS3 facilitate fatty acid uptake in yeast. Plant  162 Physiology and Biochemistry, 51(2012), 31–39.   Radakovits, R., Jinkerson, R. E., Darzins, A., & Posewitz, M. C. (2010). Genetic engineering of algae for enhanced biofuel production. Eukaryotic Cell, 9(4), 486–501.  Ramsay, N. A., & Glover, B. J. (2005). MYB-bHLH-WD40 protein complex and the evolution of cellular diversity. Trends in Plant Science, 10(2), 63–70. Reisberg, E., & Hildebrandt, U. (2012). Phyllosphere bacterial communities of trichome-bearing and trichomeless Arabidopsis thaliana leaves. Antonie van Leeuwenhoek, Journal of Microbiology, 101, 551–560.  Reynhardt, E. C. (1997). The role of hydrogen bonding in the cuticular wax of Hordeum vulgare L. European Biophysics Journal with Biophysics Letters, 26(2), 195–201. Richardson, A., Franke, R., Kerstiens, G., Jarvis, M., Schreiber, L., & Fricke, W. (2005). Cuticular wax deposition in growing barley (Hordeum vulgare) leaves commences in relation to the point of emergence of epidermal cells from the sheaths of older leaves. Planta, 222(3), 472–83. Riederer, M. (2006). Thermodynamics of the water permeability of plant cuticles: characterization of the polar pathway. Journal of Experimental Botany, 57(12), 2937–2942.  Riederer, M., & Schreiber, L. (1995). Waxes: The transport barriers of plant cuticles. In R.J. Hamilton (Ed), Waxes: Chemistry, Molecular Biology and Functions. The Oily Press, Dundee, Scotland, pp 130–156. Riederer, M., & Schreiber, L. (2001). Protecting against water loss: analysis of the barrier properties of plant cuticles. Journal of Experimental Botany, 52(363), 2023–2032. Rossak, M., Smith, M., & Kunst, L. (2001). Expression of the FAE1 gene and FAE1 promoter activity in developing seeds of Arabidopsis thaliana. Plant Molecular Biology, 46, 717–725.  Rowland, O., Zheng, H., Hepworth, S. R., Lam, P., Jetter, R., & Kunst, L. (2006). CER4 encodes an alcohol-forming fatty acyl-coenzyme A reductase involved in cuticular wax production in Arabidopsis. Plant Physiology, 142(3), 866–77.  Salasoo, I. (1983). Effect of leaf age on epicuticular wax alkanes in Rhododendron. Phytochemistry, 22(2), 461–463. Samuels, A. L., Kunst, L., & Jetter, R. (2008). Sealing plant surfaces: cuticular wax formation by epidermal cells. Annual Review of Plant Biology, 59, 683–707.  163 Schneider-Belhaddad, F., & Kolattukudy, P. (2000). Solubilization, partial purification, and characterization of a fatty aldehyde decarbonylase from a higher plant, Pisum sativum. Archives of Biochemistry and Biophysics, 377(2), 341–9.  Schönherr, J. (1976). Water permeability of isolated cuticular membranes: The effect of cuticular waxes on diffusion of water. Planta, 131(2), 159–164. Schouten, S., Moerkerken, P., Gelin, F., Baas, M., De Leeuw, J. W., & Sinninghe Damsté, J. S. (1998). Structural characterization of aliphatic, nonhydrolyzable biopolymers in freshwater algae and a leaf cuticle using ruthenium tetroxide degradation. Phytochemistry, 49(4), 987–993. Schreiber, L. (2006). Review of sorption and diffusion of lipophilic molecules in cuticular waxes and the effects of accelerators on solute mobilities. Journal of Experimental Botany, 57(11), 2515–23.  Shanklin, J., & Somerville, C. (1991). Stearoyl-acyl-carrier-protein desaturase from higher plants is structurally unrelated to the animal and fungal homologs. Proceedings of the National Academy of Sciences of the United States of America, 88(6), 2510–4. Sieber, P., Schorderet, M., Ryser, U., Buchala, A., Kolattukudy, P., Métraux, J. P., & Nawrath, C. (2000). Transgenic Arabidopsis plants expressing a fungal cutinase show alterations in the structure and properties of the cuticle and postgenital organ fusions. The Plant Cell, 12(5), 721–738. Stark, R. E., & Tian, S. (2007). The cutin biopolymer matrix. Annual Plant Reviews, 23, 126–144. Stocker, H., & Ashton, L. (1975). Changes in the composition of coffee wax with development. Phytochemistry, 14, 1919–1920. Suh, M. C., Samuels, A. L., Jetter, R., Kunst, L., Pollard, M., Ohlrogge, J., & Beisson, F. (2005). Cuticular lipid composition, surface structure, and gene expression in Arabidopsis stem epidermis. Plant Physiology, 139(December), 1649–1665. Taylor, C. B. (1997). Promoter fusion analysis: an insufficient measure of gene expression. The Plant Cell, 9(3), 273–275. Telfer, A., Bollman, K. M., Poehtig, R. S. (1997). Phase change and the trichome distribution in Arabidopsis thaliana. Development, 124(3), 645–654. Tian, D., Tooker, J., Peiffer, M., Chung, S. H., & Felton, G. W. (2012). Role of trichomes in  164 defense against herbivores: Comparison of herbivore response to woolly and hairless trichome mutants in tomato (Solanum lycopersicum). Planta, 236(4), 1053–1066. Todd, J., Post-Beittenmiller, D., & Jaworski, J. G. (1999). KCS1 encodes a fatty acid elongase 3-ketoacyl-CoA synthase affecting wax biosynthesis in Arabidopsis thaliana. The Plant Journal, 17(July), 119–130. Trenkamp, S., Martin, W., & Tietjen, K. (2004). Specific and differential inhibition of very-long-chain fatty acid elongases from Arabidopsis thaliana by different herbicides. Proceedings of the National Academy of Sciences of the United States of America, 101(32), 11903–11908. Tresch, S., Heilmann, M., Christiansen, N., Looser, R., & Grossmann, K. (2012). Inhibition of saturated very-long-chain fatty acid biosynthesis by mefluidide and perfluidoe, selective inhibitors of 3-ketyacyl-CoA synthases. Phytochemistry, 76, 162–171. Tulloch, P. (1973). Composition of leaf surface waxes of Triticum species: variation with age and tissue. Phytochemistry, 12(1972), 2225–2232. Villena, J. F., Dominguez, E., & Heredia, A. (2000). Monitoring biopolymers present in plant cuticles by FT-IR spectroscopy. Journal of Plant Physiology, 156, 419–422. Viougeas, M. A., Rohr, R., & Chamel, A. (1995). Structural changes and permeability of ivy (Hedera helix L.) leaf cuticles in relation to leaf development and after selective chemical treatments. New Phytologist, 130(3), 337–348. Volokita, M., Rosilio-Brami, T., Rivkin, N., & Zik, M. (2011). Combining comparative sequence and genomic data to ascertain phylogenetic relationships and explore the evolution of the large GDSL-lipase family in land plants. Molecular Biology and Evolution, 28(1), 551–65.  Wagner, G. J., Wang, E., & Shepherd, R. W. (2004). New approaches for studying and exploiting an old protuberance, the plant trichome. Annals of Botany, 93(1), 3–11. Wang, J., Sun, L., Xie, L., He, Y., Luo, T., Sheng, L., Luo, Y., et al. (2016). Regulation of cuticle formation during fruit development and ripening in “ Newhall ” navel orange (Citrus sinensis Osbeck) revealed by transcriptomic and metabolomic profiling. Plant Science, 243, 131–144. Wang, S., Hubbard, L., Chang, Y., Guo, J., Schiefelbein, J., & Chen, J.-G. (2008). Comprehensive analysis of single-repeat R3 MYB proteins in epidermal cell patterning and their transcriptional regulation in Arabidopsis. BMC Plant Biology, 8, 81–93. Wang, W., Liu, X., Gai, X., Ren, J., Liu, X., Cai, Y., Wang, Q., et al. (2015). Cucumis sativus L.  165 WAX2 plays a pivotal role in wax biosynthesis, influencing pollen fertility and plant biotic and abiotic stress responses. Plant & Cell Physiology, 56(7), 1339–1354. Wang, Y., Wang, M., Sun, Y., Hegebarth, D., Li, T., Jetter, R., & Wang, Z. (2015). Molecular characterization of TaFAR1 involved in primary alcohol biosynthesis of cuticular wax in hexaploid wheat. Plant and Cell Physiology, 56(10), 1944–1961. Waters, E. R. (2003). Molecular adaptation and the origin of land plants. Molecular Phylogenetics and Evolution, 29(3), 456–463. Weidenhamer, J. D., Macias, F. a., Fischer, N. H., & Williamson, G. B. (1993). Just how insoluable are monoterpenes? Journal of Chemical Ecology, 19(8), 1799–1807. Wen, M., & Jetter, R. (2009). Composition of secondary alcohols, ketones, alkanediols, and ketols in Arabidopsis thaliana cuticular waxes. Journal of Experimental Botany, 60(6), 1811–1821. Von Wettstein-Knowles, P. (2007). Analyses of barley spike mutant waxes identify alkenes, cyclopropanes and internally branched alkanes with dominating isomers at carbon 9. The Plant Journal, 49(2), 250–264. Wise, A. a, Liu, Z., & Binns, A. N. (2006). Three methods for the introduction of foreign DNA into Agrobacterium. Methods in Molecular Biology, 343, 43–53. Xia, Y., Yu, K., Navarre, D., Seebold, K., Kachroo, A., & Kachroo, P. (2010). The glabra1 mutation affects cuticle formation and plant responses to microbes. Plant Physiology, 154(2), 833–46.  Xiao, F., Goodwin, S. M., Xiao, Y., Sun, Z., Baker, D., Tang, X., Jenks, M. A., et al. (2004). Arabidopsis CYP86A2 represses Pseudomonas syringae type III genes and is required for cuticle development. The EMBO Journal, 23(14), 2903–2913. Xu, X., Dietrich, C. R., Lessire, R., Nikolau, B. J., & Schnable, P. S. (2002). The endoplasmic reticulum-associated maize GL8 protein is a component of the acyl-coenzyme A elongase ivolved in the production of cuticular waxes. Plant Physiology, 128(3), 924–34. Yadav, V., Molina, I., Ranathunge, K., Castillo, I. Q., Rothstein, S. J., & Reed, J. W. (2014). ABCG transporters are required for suberin and pollen wall extracellular barriers in Arabidopsis. The Plant Cell, 1–21.  Yang, M., & Sack, F. D. (1995). The too many mouths and four lips mutations affect stomatal production in Arabidopsis. The Plant Cell, 7(12), 2227–2239.  166 Yeats, T. H., Howe, K. J., Matas, A. J., Buda, G. J., Thannhauser, T. W., & Rose, J. K. C. (2010). Mining the surface proteome of tomato (Solanum lycopersicum) fruit for proteins associated with cuticle biogenesis. Journal of Experimental Botany, 61(13), 3759–71.  Yeats, T. H., Martin, L. B. B., Viart, H. M.-F., Isaacson, T., He, Y., Zhao, L., Matas, A. J., et al. (2012). The identification of cutin synthase: formation of the plant polyester cutin. Nature Chemical Biology, 8, 8–10.  Yeats, T. H., & Rose, J. K. C. (2013). The formation and function of plant cuticles. Plant Physiology, 163(1), 5–20.  Yephremov, A., Wisman, E., Huijser, P., Huijser, C., Wellesen, K., & Saedler, H. (1999). Characterization of the FIDDLEHEAD gene of Arabidopsis reveals a link between adhesion response and cell differentiation in the epidermis. The Plant Cell, 11(11), 2187–2201. Zhao, L., Katavic, F., Haughn, G., & Kunst, L. (2010). Insertional mutant analysis reveals that long/chain avyl-CoA synthetase 1 (LACS1), but not LACS8, functonally overlaps with LACS9 in Arabidopsis seed oil biosynthesis. The Plant Journal, 64, 1048–1058. Zheng, H., Rowland, O., & Kunst, L. (2005). Disruptions of the Arabidopsis Enoyl-CoA reductase gene reveal an essential role for very-long-chain fatty acid synthesis in cell expansion during plant morphogenesis. The Plant Cell, 17(5), 1467–1481. Züst, T., Joseph, B., Shimizu, K. K., Kliebenstein, D. J., & Turnbull, L. a. (2011). Using knockout mutants to reveal the growth costs of defensive traits. Proceedings of the Royal Society of London B: Biological Sciences, 278(1718), 2598–2603.   167 Appendices  Appendix A  Supporting information for chapter 2  Table A.1: Morphological data for developing wild-type Arabidopsis eighth leaves.  Values are means ± standard deviation from five independent leaves. Leaf surface area is the macroscopic surface area of both sides of the blade, ignoring contributions from trichomes.      Plant Leaf)8 Petiole Blade20 5 1 4 74 ± 3 6675 ± 734 20521 ± 1259 0.20 ± 0.02 487 ± 3022 7 2 5 75 ± 2 7502 ± 1339 22064 ± 2297 0.36 ± 0.02 809 ± 8424 9 4 7 77 ± 6 7549 ± 1176 25106 ± 2028 0.80 ± 0.06 1586 ± 12826 11 6 9 75 ± 3 8283 ± 1023 25180 ± 2911 0.98 ± 0.08 1958 ± 22628 13 7 12 78 ± 6 10070 ± 1214 29097 ± 3173 1.6 ± 0.1 2808 ± 30630 15 10 13 75 ± 4 9892 ± 858 30013 ± 2220 2.1 ± 0.06 3456 ± 25632 17 11 14 75 ± 5 10411 ± 994 30660 ± 1821 2.4 ± 0.1 3914 ± 23236 21 14 16 72 ± 6 10366 ± 1156 29602 ± 2681 2.8 ± 0.08 4646 ± 42140 25 17 18 70 ± 8 9885 ± 1242 30446 ± 2303 3.2 ± 0.1 5230 ± 3960.004 # 0.00040.0072 # 0.00040.016 # 0.00120.0196 # 0.00160.0328 # 0.0020.0416 # 0.00120.048 # 0.00240.0552 # 0.00160.0636 # 0.002original5(mm2/blade)10 ± 118 ± 140 ± 349 ± 482 ± 5104 ± 3120 ± 6138 ± 4159 ± 5Surface)areaLeaf)(cm2) Pavement)cell)(μm2cell91))Age)(days) Length)(mm) Number)of)cells)(on)one)side)of)the)leaf)Trichome Guard)(pairs) Pavement 168  Figure A.1: Sampling scheme for studying cell size distributions across sections of Arabidopsis wild type leaves of different ages. Leaf age (days) and blade length (mm) are shown at the left side of each oval. Dashed lines delineate sections of equal length for cell size counting. The length of each section along the longitudinal leaf axis is denoted underneath each corresponding double-headed arrow.   5 days / 4 mm9 days / 7 mm13 days / 12 mm17 days / 14 mm21 days / 16 mmleaf base leaf tip1 mm1.75 mm2 mm2.3 mm2.6 mm 169  Figure A.2: Calculated trichome wax composition. The relative amount of each compound class, as calculated by subtracting gl1 composition from wild type composition, is plotted as a percent of the total wax mixture. Within each compound class the amount of compounds C<30 are shown in black, C32 are shown in dark grey, C34 are shown in light grey, and C36-C38 are shown in white.     170 Table A.2: List of primer sequences used in time course qRT-PCR analysis.  Gene	   Forward	  primer	  sequence	   Reverse	  primer	  sequence CER10	   5'-­‐GTGACTCCTGGATCGAAGGACAAAC-­‐3'	   5'-­‐GGCTGAACCGATGTACGAAAAACG-­‐3' KCS1	   5'-­‐ACGTGGCATAACTTCAACGC-­‐3'	   5'-­‐ACATAACGGCTTCAGCTTCG-­‐3' KCS5	   5'-­‐AACTCCGCCGTTTGGAAATG-­‐3'	   5'-­‐TCGTTCGATACAATCCGACCAC-­‐3' CER6	   5'-­‐TGGCCTTCGAACACTTTTGA-­‐3'	   5'-­‐ATTCTTGAGGCCTCAACGTG-­‐3' KCS8	   5'-­‐TATCGCAACTTTGGGTCCTC-­‐3'	   5'-­‐CACCAGCGTGGATACAGAAA-­‐3' KCS16	   5'-­‐CCCATAAGCGCGAAGTTTCA-­‐3'	   5'-­‐CCAACTCGTACCAAATGGAG-­‐3' CER2	   5'-­‐ACTCCAATCTTATCACCACGTCTCAG-­‐3'	   5'-­‐TGGAAATAGCTTCACCATTACGA-­‐3' CER26	   5'-­‐ATGCAATCTCACTGTGGAGGAG-­‐3'	   5'-­‐CGCATTTAAACCGGGTCATCTG-­‐3' CER8	   5'-­‐TGGGGATCTAGGGTGTTAATTGTCCT-­‐3'	   5'-­‐GCCGCTGGTGTACATTATGGTGCATA-­‐3' CER3	   5'-­‐TATCCTCTGTGTCGTAGCTGGGTGTC-­‐3'	   5'-­‐CATGGCCGACATTACATCAACTCC-­‐3' CER1	   5'-­‐AGCATACTTCATCCTCTTCGCGATAC-­‐3'	   5'-­‐GGTATATGGATTCTGGCGTCGTCAG-­‐3' CER4	   5'-­‐GCTCATGCGATGATCCACCAAGTTGA-­‐3'	   5'-­‐CTCGTAGTTGATGTCTAAACCGGTGG-­‐3' WSD1	   5'-­‐GTTTGTTTGGTTTGATCTTAG-­‐3'	   5'-­‐GGTGATCTTTTGTTATGCCC-­‐3' UBC	   5'-­‐GAATGCTTGGAGTCCTGCTT-­‐3'	   5'-­‐GGCGAGGCGTGTATACATTT-­‐3' Actin2	   5'-­‐CCAGAAGGATGCATATGTTGGTGA-­‐3'	   5'-­‐GAGGAGCCTCGGTAAGAAGA-­‐3' UBQ10	   5'-­‐ACCCTAACGGGAAAGACGAT-­‐3'	   5'-­‐AGCCTGAGAACAAGATGAAGG-­‐3' GAPDH	   5'-­‐GGCATTGTTGAGGGACTCAT-­‐3'	   5'-­‐TACTCGTGCTCGTTGACACC-­‐3'     171 Appendix B  Supporting information for statistical analysis  B.1 Permutations tests Permutation tests were used throughout the analysis in place of standard parametric procedures for two successive reasons: (1) measurements were made on a limited number of subjects raised in a growth chamber that could not be considered as drawn at random from some hypothetical large population; consequently, (2) typical assumptions required for parametric procedures, like those about random sampling, were untenable. Permutation tests require no sampling assumptions, nor do they require as restrictive assumptions about the distributions of the errors of the observed processes as do their parametric counterparts. The theoretical details behind permutation testing can be found in any introductory text on nonparametric statistics (for example, Gibbons & Chakraborti, 2011). Trends in both the chemical and gene expression data that were found to be significant by these tests were marked with an asterisk (*).  B.2 Statistical analysis of the gene expression data The gene expression data (Figure 2.7) appear consistently inflated at day 5. This suggests the possible presence of some unknown biasing factor or agent. Due to the small size of five-day-old eighth leaves, RNA extractions and gene expression level are difficult to calibrate and measure consistently in a pooled sample of leaves. This could plausibly lead to a type of systematic measurement error that could account for the apparent inflation of the gene expression data acquired on day five. However, it is also possible that the apparent inflation on day five is simply due to some real biochemical phenomenon. Consequently, the assumptions under which statistical hypothesis testing could be performed are unclear.  172 To deal with this uncertainty, a sensitivity analysis was performed that investigated which trends were sensitive to possible inflation on day five. A robust least squares fit (Tukey's bisquare) was conducted on a lightened data set, wherein only a single data point from day five was used, selected at random from the three day five candidates at each permutation sampled for the test of hypothesis. Approximately 96 % of these robust fits converged in 20 or fewer iterations. Next, an ordinary least squares fit was conducted on a revised data set completely omitting the data from day 5. Trends that held up across all three analyses can be considered insensitive to the presence of possible inflation at day five, and so are robust to analytical assumptions. After accounting for the study-wide false discovery rate, such trends were found among KCS1, CER6, KCS8, and CER10. The trends in the KCS5, CER8 and CER1 data, flagged by the initial robust least squares fit, were not present in the lightened data. These trends, along with the one associated with KCS16, were also not present when data from day five was omitted completely. Thus, these three trends should be viewed as only suggested by the data under the assumption of no inflation at day 5. Finally, CER2 and CER4 exhibited significant trends using both the lightened data and when omitting data from day five, but not when using the full dataset, indicating a probable effect assuming the presence of inflation at day five. Trends that are significant under the assumption of inflation at day five are flagged with a diamond (u) in Figure 2.7.  B.3 R code for permutation testing The following is example R code (R Core Team, 2015) for a permutation test of the ordinary least squares slopes of each of N classes over time. All permutation tests associated with Figure 2.3 or Figure 2.4 were run using this base code, adjusted only by setting the number, N, of tests being sequentially performed and the number of permutations (M = 10000 or M = 25000) used  173 to generate each null distribution. Permutation tests associated with Figure 2.6 were adjusted to test a hypothesis on the difference of means between two classes (with M = 40000); those associated with Figure 2.7 used a robust least squares fit (with M = 10000).  ##permutation tests on the slopes of N classes over time, given n sample points from each class## ##the data frame "pv" collects the p-values of the N permutation tests## slope <- data.frame(matrix(ncol = N, nrow = M)) pv <- data.frame(matrix(ncol = N, nrow = 1))  for (j in 1 : N) { for (i in 1 : M) {  class <- sample(data[1:n,j])  ols <- lm(class ~ data$age)  slope[i,j] <- ols$coef[2] } ols.t <- lm(data[1:n,j] ~ data$age) fn <- ecdf(slope[,j]) if(ols.t$coef[2] < 0){pv[j] <- 2*fn(ols.t$coef[2])} else{pv[j] <- 2*(1 - fn(ols.t$coef[2]))}}     174 Appendix C  Supporting information for chapter 4  Table C.1: PCR target genes and primer sequences.  Target'gene locus forward reverseGenotyping LbaI N/A 5’;TGGTTCACGTAGTGGGCCATCG;3’ N/AKCS16 AT4G34250 5’;GCACCTTTCTAAGCCGTTATCTG;3’' 5’;AAGTTTTGAGTGATGTTCTCC;3’Semi;quantitative'RT;PCR KCS16 AT4G34250 5’;GCACCTTTCTAAGCCGTTATCTG;3’' 5’;AAGTTTTGAGTGATGTTCTCC;3Actin2 AT3G18780 5’;'CAC'GAG'ACA'ACC'TAT'AAC'TCA'AT;3’' 3;GTG'ATT'TCT'TTG'CTC'ATA'CGG'TCA;5’Quantitative'RT;PCR KCS16 AT4G34250 5';CCCATAAGCGCGAAGTTTCA;3' 5';CCAACTCGTACCAAATGGAG;3'UBC21 AT5G25760 5';GAATGCTTGGAGTCCTGCTT;3' 5';GGCGAGGCGTGTATACATTT;3'Actin2 AT3G18780 5';CCAGAAGGATGCATATGTTGGTGA;3' 5';GAGGAGCCTCGGTAAGAAGA;3'UBQ10 AT4G05320' 5';ACCCTAACGGGAAAGACGAT;3' 5';AGCCTGAGAACAAGATGAAGG;3'GAPDH AT1G13440' 5';GGCATTGTTGAGGGACTCAT;3' 5';TACTCGTGCTCGTTGACACC;3'Cloning KCS168GTW8coding8region AT4G342505’;GGG'GAC'AAG'TTT'GTA'CAA'AAA'AGC'AGA'TGG'ATT'ACC'CCA'TGA'AGA'AGG'TAA'AAA'TCT'TTT'TCA'ACT;3’'5’;GGG'GAC'CAC'TTT'GTA'CAA'AGC'TGT'CAC'TCT'TTT'AAA'TCT'ATA'TCG'ATC'TCA'ACT'GGA'TAT'TTG'T;3’Colony'PCR KCS168coding8region AT4G34250 5’;GCACCTTTCTAAGCCGTTATCTG;3’' 5’;AAGTTTTGAGTGATGTTCTCC;3 175  Figure C.1: Wax composition of Arabidopsis kcs16 stems and siliques. A) Compound class composition of kcs16-1 and-2 stems. B) Carbon chain length distribution within compound classes in kcs16-1 and-2 stems. Numbers at the x-axis indicate the carbon chain length. C) Compound class composition of kcs16-1 and-2 siliques. D) Carbon chain length distribution within compound classes in kcs16-1 and-2 in siliques. Error bars indicate standard deviation as a mean of five replicates.  Fatty acidsAldehydesAlkanessec. n-AlcoholsKetoneprim. n-AlcoholsWax estersSterolsNot identified0123456789Coverage (µg/cm2 )Col-0kcs16-1kcs16-224 26 28 30 26 28 30 32 25 26 27 28 29 30 31 32 33 3429 sec.n-Alcohol29 Ketone 22 24 26 28 30 32 38 40 42 44 46 48 50 52 54 560.00.51.01.52.02.53.0Coverage (µg/cm2 )Fatty acidsprim. n-AlcoholsWax EstersAldehydes AlkanesFatty acidsAldehydesAlkanessec. n-AlcoholsKetoneprim. n-AlcoholsWax estersSterolsNot identified0.00.51.01.52.02.53.03.54.0Coverage (µg/cm2 )A B C D 24 26 28 30 26 28 30 32 25 26 27 28 29 30 31 32 33 3429 sec. n-Alcohol29 Ketone 24 26 28 30 32 34 38 40 42 44 46 48 50 52012345678Coverage (µg/cm2 )Fatty acidsprim.  n-AlcoholsAldehydes Alkanes Wax esters 176  Figure C.2: Chain length composition of mature and young leaves of Arabidopsis kcs16 mutants and complementation lines expressing KCS16 ectopically in both kcs16 backgrounds. A) Carbon chain length distribution within compound classes in kcs16-1 and -2 mature leaves. B) Carbon chain length distribution within compound classes in kcs16-1 and -2 young leaves. C) Carbon chain length distribution within compound classes in young leaves of complementation lines expressing KCS16 in both ksc16 backgrounds. Labels on x-axis indicate the carbon number of identified compounds. Error bars represent the mean standard deviation of a total of five replicates for the leaves and three replicates fro the stems. Asterisks indicate discovery of significant differences of coverage between wild type and ksc16 lines or p35S:KCS16 lines expressed in ksc16 backgrounds respectively based on Student’s t-test (* = p < 0.05). 24 26 28 30 32 24 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 sec.n-Alcohols29 Ketone 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Col-0p35S:KCS16 in kcs16-1p35S:KCS16 in kcs16-2Fatty acidsAldehydes Alkanes Alkenes  n-        br.   Alcoholsprim.      *24 26 28 30 32 34 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 sec.n- Alcohol29 Ketone 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Fatty acids n-         br. AlcoholsAldehydes Alkanes Alkenes* *kcs16-1Col-0kcs16-2 prim.** * *A B C 24 26 28 30 32 24 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 sec. n-Alcohol29 Ketone 22 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Fatty acids   n-         br. AlcoholsAldehydes Alkanes Alkenes prim. ** ******** 177  Figure C.3: Wax composition of Arabidopsis lines expressing KCS16 ectopically in the wild type background. A) Stem compound class distribution of ectopic lines expressing KCS16 in the wild type background in stems. B) Stem carbon chain length distribution within compound classes of ectopic lines expressing KCS16 in the wild type background. C) Carbon chain length distribution within compound classes of ectopic lines expressing KCS16 in the wild type background in mature leaves. D) Carbon chain length distribution within compound classes of ectopic lines expressing KCS16 in the wild type background in mature leaves. Numbers at the x-axis indicate the carbon chain length. Error bars indicate standard deviation as a mean of five replicates. Asterisks indicate discovery of significant differences of coverage between wild type and p35S:KCS16 lines based on Student’s t-test (* = p < 0.05).    A B C D 24 26 28 30 32 34 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 sec.n- Alcohol29 Ketone 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 )Fatty acids n-         br. AlcoholsAldehydes Alkanes Alkenes prim.24 26 28 30 26 28 30 32 25 26 27 28 29 30 31 32 33 3429 sec. n-Alcohol29 Ketone 24 26 28 30 32 34 38 40 42 44 46 48 50 52012345678Coverage (µg/cm2 )Fatty acidsprim.  n-AlcoholsAldehydes Alkanes Wax esters24 26 28 30 32 24 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 sec. n-Alcohol29 Ketone 22 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.45Coverage (µg/cm2 ) Fatty acids   n-         br. AlcoholsAldehydes Alkanes Alkenes prim. ***Fatty acidsAldehydesAlkanessec. n-AlcoholsKetoneprim. n-AlcoholsWax estersSterolsNot identified0123456789Coverage (µg/cm2 )Col-0p35S:KCS16 in WT line1 p35S:KCS16 in WT line2 178 Table C.2: Expression vectors and selection medium used in the heterologous yeast expression experiment. GENES expressed Expression Vectors Selection medium  CER6 pYES2GW:CER6 pYES3GW pVTLEUGW -URA-TRP-LEU CER6 + CER2 pYES2GW:CER6 pYES3GW:CER2 pVTLEUGW -URA-TRP-LEU CER6 + CER26 pYES2GW:CER6 pYES3GW:CER26 pVTLEUGW -URA-TRP-LEU CER6 + KCS16 pYES2GW:CER6 pYES3GW pVTLEUGW:KCS16 -URA-TRP-LEU CER6 + CER2 + KCS16 pYES2GW:CER6 pYES3GW:CER2 pVTLEUGW:KCS16 -URA-TRP-LEU CER6 + CER26 + KCS16 pYES2GW:CER6 pYES3GW:CER26 pVTLEUGW:KCS16 -URA-TRP-LEU     179 Appendix D  KCS8 cuticular wax analysis  Figure D.1: Chain length composition of Arabidopsis kcs8 mutants. A) Carbon chain length distribution within compound classes in kcs8-1 and -2 young leaves. B) Carbon chain length distribution within compound classes in kcs8-1 and -2 mature leaves. C) Carbon chain length distribution within compound classes in ksc8-1 and -2 stems. Error bars represent the mean standard deviation of a total of four replicates.  24 26 28 30 32 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 Sec. Alcohol29 Ketone 22 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.450.50Coverage (µg/cm2 )Fatty acidsAldehydesAlkanes Alkenes prim. n-Alcoholsbr.-Alcohols24 26 28 30 32 26 28 30 32 34 25 26 27 28 29 30 31 32 33 34 35 36 37 33 35 3729 Ketone29 Sec. Alcohol 22 24 26 28 30 32 34 30 32 340.000.050.100.150.200.250.300.350.400.450.50Coverage (µg/cm2 )Col-0kcs8-1 SALK_055908Ckcs8-2 SALK_112610CFatty acids Aldehydes Alkanes Alkenes prim. n-Alcoholsbr.-Alcohols24 26 28 30 26 28 30 32 25 26 27 28 29 30 31 32 33 3429 Sec. Alcohol29 Ketone 22 24 26 28 30 32 38 40 42 44 46 48 50 52012345678910Coverage (µg/cm2 )Fatty AcidsAldehydes Alkanesprim. n-AlcoholsWax esters

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0308092/manifest

Comment

Related Items