Functional Characterization of Lysine-rich Arabinogalactan-Proteins (AGPs) and an AG

Peptide in Arabidopsis

A dissertation presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree

Doctor of Philosophy

Yizhu Zhang

November 2008

© 2008 Yizhu Zhang. All Rights Reserved. 2

This dissertation titled

Functional Characterization of Lysine-rich Arabinogalactan-Proteins (AGPs) and an AG

Peptide in Arabidopsis

by

YIZHU ZHANG

has been approved for

the Department of Environmental and Plant Biology

and the College of Arts and Sciences by

Allan M. Showalter

Professor of Environmental and Plant Biology

Benjamin M. Ogles

Dean, College of Arts and Sciences 3

ABSTRACT

ZHANG, YIZHU, Ph.D., November 2008, Environmental and Plant Biology

Functional Characterization of Lysine-rich Arabinogalactan-Proteins (AGPs) and an AG

Peptide in Arabidopsis (160 pp.)

Director of Dissertation: Allan M. Showalter

The plant cell wall is composed of complex polysaccharides and a small amount

of structural proteins and cell wall enzymes. Arabinogalactan-proteins (AGPs) are highly glycosylated, hydroxyproline-rich structural proteins that play important roles in plant growth and development. AtAGP17, 18 and 19 comprise the lysine-rich classical AGP family in Arabidopsis. They consist of an N-terminal signal peptide, a classical AGP domain disrupted by a short basic lysine-rich subdomain and a C-terminal glycosylphosphatidylinositol (GPI) anchor addition sequence. A previous study showed a null T-DNA insertion mutant of AtAGP19 displayed pleiotropic phenotypes. Here, a microarray approach was employed to elucidate changes in gene expression associated with the atagp19 mutant. The expression of several genes related to cell expansion were found to change significantly. Interestingly, one gene (At1g68720, cytidine/deoxycytidylate deaminase family protein) adjacent to AtAGP19 was found to be down-regulated about 50 fold and RT-PCR showed the absence of mRNA for this gene in the atagp19 mutant. Furthermore, complementation with the 3’ portion of the At1g68720

gene can fully restore all the wild type phenotypes, indicating this region is critical for the functions revealed by the agp19 mutant. To examine cellular localization of the lysine- rich AGPs, GFP-AtAGP17/18/19 fusion proteins as well as a GFP control were 4

overexpressed in Arabidopsis plants and the fusion proteins were present on the plant cell

surface. Plasmolysis of leaf trichome cells further determined the localization of the

fusion proteins at the plasma membrane. Moreover, in vitro pollen germination showed

that AtAGP17, unlike LeAGP-1 (the lysine-rich AGP in tomato), was not associated with pollen tube elongation. To further elucidate AtAGP17/18/19 function(s), transgenic

Arabidopsis overexpressing AtAGP17/18/19 without the GFP tag were produced.

AtAGP18 overexpressors displayed several phenotypes distinct from the wild type plants: they were short and bushy, had short roots and produced less viable seeds. In contrast, the vector control transformants as well as the AtAGP17/19 overexpressors had the same phenotypes as the wild type plants. Furthermore, AtAGP18 was down-regulated by the plant hormone ABA, indicating ABA may be involved in AtAGP18 function(s). Finally, the expression pattern of AtAGP14 (At5g56540), an AG peptide in Arabidopsis was examined. AtAGP14 was highly expressed in flowers and young roots and moderately expressed in seedlings, stems and rosette leaves. A plate-based phenotypic analysis was also carried out for the T-DNA insertional mutant of AtAGP14 and wild type Arabidopsis but no significant differences were observed with respect to germination rate, true leaf numbers, primary root length and lateral root numbers.

Approved: ______

Allan M. Showalter

Professor of Environmental and Plant Biology 5

ACKNOWLEDGMENTS

First of all, I express my gratitude to my advisor, Dr. Allan M. Showalter, for his

constant guidance, help and encouragement in my research and his patience to help me

improve my writing skills. I thank Drs. Ahmed Faik, Marcia J. Kieliszewski and Sarah E.

Wyatt for serving on my doctoral advisory committee and their support and valuable

advice on my work.

The Department of Environmental and Plant Biology, the Molecular and Cellular

Biology Program and the Department of Biological Sciences have provided me with financial assistantship and a good research environment so that I can finish my dissertation. Specifically, I thank our department administrative, Connie Pollard, for her assistance in my lab work and dissertation. I thank Jeffrey Thuma for the confocal microscopy orientation, Darron Luesse for his valuable suggestions on Arabidopsis transformation techniques, John Withers for the insightful discussion of constructs design, Li Tan for the GFP control construct and Vijay Nadella for sequencing service from the Ohio University genomics facility. My labmates, Dr. Ming Chen, Dr. Wenxian

Sun, Dr. Harjinder S. Sardar, Dr. Jie Yang, Yan Liang and Brian Keppler have offered me generous help in the lab. Working with them is a pleasant and unforgettable experience. I also thank Rebecca Vondrell for her assistance in screening transgenic plants and phenotype analysis.

Finally, I am indebted to my relatives and friends for their understanding, help and support. 6

TABLE OF CONTENTS

Page

ABSTRACT...... 3 ACKNOWLEDGEMENTS...... 5 LIST OF TABLES...... 11 LIST OF FIGURES ...... 12 LIST OF ABBREVIATIONS...... 14 CHAPTER 1 INTRODUCTION ...... 16 Plant cell wall proteins...... 16

Arabinogalactan proteins (AGPs)...... 18

AGP structure...... 19

Protein backbone...... 19 Carbohydrate...... 21 The Hyp contiguity hypothesis ...... 21 Molecular structure of AGPs ...... 22 AGP classification ...... 23

Cellular localization of AGPs ...... 24

Physiological function of AGPs...... 25

Role of AGPs in embryogenesis...... 26 Role of AGPs in xylem development ...... 27 Role of AGPs in reproduction...... 28 Role of AGPs in cell division and expansion ...... 31 Role of AGPs in cell signaling...... 32 Lysine-rich AGPs...... 35

Specific aims of this dissertation research...... 37

CHAPTER 2 T-DNA MUTANT STUDY OF atagp19 ...... 39 7

Summary...... 39

Introduction...... 40

Materials and methods ...... 42

Plant materials and growth conditions...... 42 Microarray experiments...... 42 Quantitative PCR (Real-time PCR) ...... 45 Breaking force measurement ...... 47 Complementation construct design with the AtAGP19 promoter sequence ...... 47 DNA extraction and amplification of AtAGP19 promoter ...... 48 Restriction digestion ...... 49 Ligation and transformation...... 49 Construct verification by restriction digestion...... 49 Preparation of Agrobacterium competent cells...... 50 Transformation of Agrobacterium using electroporation ...... 50 Floral dip transformation of Arabidopsis...... 51 Screening of transgenic plants ...... 52 Verification of transgenic plants by DNA extraction and PCR...... 53 RNA extraction from transgenic plants and RT-PCR...... 53 PCR amplification for At1g68720 sequencing ...... 54 Results...... 55

Quality assessment of RNA samples by Agilent Bioanalyzer...... 55 Overview of changes in gene expression in the atagp19 mutant...... 57 Breaking force was different in atagp19 mutant and WT ...... 63 Real-time PCR analyses and RT-PCR...... 63 AtAGP19 promoter sequence complements atagp19 phenotypes ...... 65 Gene At1g68720 is disrupted in the atagp19 mutant...... 68 Discussion...... 69 8

Real-time PCR results are reliable...... 69 Expression of AtAGP17 and AtAGP18 did not change in the mutant..... 70 Gene At1g68720 and atagp19 mutant phenotypes ...... 71 Cell wall-related genes were down-regulated in the mutant...... 74 Genes encoding wall-associated kinases (WAKs), receptor-like kinases (RLKs) and other protein kinases display changes in expression...... 77 Expression of Genes Related to Transcriptional Regulation Showed Changes...... 79 CHAPTER 3 OVEREXPRESSION OF AtAGP17, 18 AND 19 ...... 81 Summary...... 81

Introduction...... 82

Materials and methods ...... 83

Overexpression constructs design...... 83 Construction of the overexpression vectors...... 89 Sequencing...... 90 Agrobacterium transformation, floral dip transformation and screening of transgenic plants...... 90 Genetic analyses of transgenic plants ...... 91 Total protein extraction...... 92 Protein electrophoresis and transfer...... 93 Coomassie blue staining ...... 94 Western blotting...... 94 Confocal laser scanning microscopy ...... 95 Sectioning ...... 95 Pollen germination...... 95 Pollen protoplast preparation ...... 96 Plasmolysis ...... 96 Hormone treated Arabidopsis seedlings, RNA extraction and RT-PCR .. 96

Hormone treated PAtAGP18:GUS transgenic plants and GUS staining ...... 97 9

Organ specific expression of AtAGP17...... 98

PAtAGP17:GUS construct and GUS staining...... 98 Results...... 99

GFP-AtAGP17 was expressed in transgenic plants...... 99 AtAGP17 is localized to the plasma membrane ...... 103 AtAGP17 is not involved in pollen tube elongation...... 104 Organ specific expression of AtAGP17...... 104 GFP-AtAGP18/19 were expressed in transgenic plants ...... 107 AtAGP17 overexpressors showed no phenotype...... 110 AtAGP18 overexpressor had several phenotypes...... 111 AtAGP19 overexpressors showed no phenotype...... 116 ABA down-regulated AtAGP18 expression...... 117 Discussion...... 119

Subcellular localization of AtAGP17/18/19 ...... 119 AtAGP17 and pollen germination ...... 120 AtAGP18 and overexpression phenotypes ...... 120 CHAPTER 4 EXPRESSION AND MUTANT STUDY OF AtAGP14...... 125 Summary...... 125

Introduction...... 126

Materials and methods ...... 127

Plant growth, RNA extraction and RT-PCR...... 127 Northern blotting...... 127 Bioinformatics analysis...... 128 Screening of T-DNA homozygous line ...... 129 Plate-based phenotypic analysis ...... 130 Results...... 131

Organ-specific expression analysis and Microarray expression data ..... 131 Isolation of homozygous line from SALK_096806...... 134 10

Phenotypic analyses showed no significant differences between mutant and wild type...... 136 Discussion...... 138

CHAPTER 5 CONCLUSIONS ...... 141 T-DNA mutant study of AtAGP19...... 141

Overexpression of AtAGP17, 18 and 19 ...... 142

Expression and mutant study of AtAGP14...... 144

REFERENCES ...... 145 APPENDIX A: SIGNAL STRENGTH DATA OF THE CANDIDATE GENES...... 154 APPENDIX B: SEQUENCE INFORMATION FOR OVEREXPRESSION CONSTRUCTS...... 157

11

LIST OF TABLES Page

Table 2.1 Primers used for quantitative PCR...... 46 Table 2.2 Summary of genes with mRNA levels up-regulated or down-regulated at least 2-fold in atagp19 mutant plants compared to wild type control plants...57 Table 2.3 Genes up-regulated or down-regulated at least 2-fold in atagp19...... 59 Table 2.4 Comparison of the changes in gene expression (i.e., mRNA levels) detected by microarray analysis and by quantitative PCR (QPCR), and signal strength for each gene...... 65 Table 2.5 Analysis of WT, atagp19 mutant and complemented plants...... 67 Table 3.1 Analysis of WT, 18NG and VC plants ...... 114 Table 4.1 Plate-based growth stage phenotype analysis of WT and atagp14 seedlings..... ……………………………………………………………………………………...... 136

12

LIST OF FIGURES Page

Figure 1.1 Glycosylphosphatidylinositol (GPI) anchor structure and the consensus sequence for GPI anchor addition...... 20 Figure 1.2 Two models for molecular shapes of AGPs...... 23 Figure 1.3 Schematic illustration of different classes of AGPs in Arabidopsis...... 24 Figure 1.4 Models of AGP signaling mechanism...... 34 Figure 2.1 AtAGP19 gene structure and the atagp19 mutant ...... 43 Figure 2.2 Experimental strategy to elucidate RNA expression levels in wild type Arabidopsis and atagp19 mutant plants by microarray analysis ...... 45 Figure 2.3 Electropherogram summary of RNA samples by Agilent Bioanalyzer ...... 56 Figure 2.4 Volcano plot of the microarray results ...... 58 Figure 2.5 Genes with more than a two-fold change categorized by their predicted cellular location...... 58 Figure 2.6 Breaking force measurements of stems from wild type and atagp19 mutant plants...... 63 Figure 2.7 RT-PCR of candidate genes in WT and atagp19 mutant...... 64 Figure 2.8 Genetic analyses of AtAGP19 in complemented atagp19 mutant...... 66 Figure 2.9 Genetic analyses of At1g68720 in complemented atagp19 mutant...... 67 Figure 2.10 Complementation of atagp19 mutant with the 19 promoter region...... 67 Figure 2.11 A schematic model showing the gene location of AtAGP19 and At1g68720, and 8 pairs of sequencing primers ...... 68 Figure 2.12 A schematic model showing the gene location of AtAGP19 and At1g68720, complementation sequences, and primers used for RT-PCR analysis of At1g68720...... 71 Figure 3.1 A summary of the structure of the recombinant constructs used in this research ...... 85 Figure 3.2 Figure illustration of recombinant constructs...... 87 13

Figure 3.3 A schematic model showing primers used in RT-PCR of the AtAGP17/18/19...... 92 Figure 3.4 GFP-AtAGP17 is expressed in transgenic plants...... 101 Figure 3.5 AtAGP17 is localized on the plant cell surface...... 102 Figure 3.6 AtAGP17 is localized to the plasma membrane...... 103 Figure 3.7 GFP-AtAGP17 is absent in pollen tubes during pollen germination ...... 105 Figure 3.8 Organ specific expression of AtAGP17...... 106 Figure 3.9 GFP-AtAGP18 is expressed in transgenic plants...... 108 Figure 3.10 GFP-AtAGP19 is expressed in transgenic plants...... 110 Figure 3.11 Genetic analyses of AtAGP17 overexpression plants...... 111 Figure 3.12 Genetic analyses of AtAGP18 overexpression (18NG) and vector control (VC) plants ...... 112 Figure 3.13 Phenotypes of AtAGP18 (18NG) overexpression plants ...... 113 Figure 3.14 Reproductive phenotypes of AtAGP18 (18NG) overexpressing plants....115 Figure 3.15 Root phenotypes of AtAGP18 (18NG) overexpressing plants...... 116 Figure 3.16 Genetic analyses of AtAGP19 overexpressing plants...... 117 Figure 3.17 AtAGP18 expression is down-regulated by ABA treatment ...... 118 Figure 4.1 An illustration of the PCR screening rationale...... 130 Figure 4.2 RT-PCR and northern blotting results of AtAGP14 mRNA expression levels in different organs...... 132 Figure 4.3 Expression profiles of AtAGP14 from microarray data in Arabidopsis...... 133 Figure 4.4 Expression profiles of AtAGP13 from microarray data in Arabidopsis...... 134 Figure 4.5 SALK_096806 was a null knockout mutant of AtAGP14...... 135 Figure 4.6 atagp14 had no obvious phenotypes in roots ...... 137

14

LIST OF ABBREVIATIONS ABA: abscisic acid ABRC: Arabidopsis Biological Resource Center AG: arabinogalactan AGP: arabinogalactan-protein bp: base pair BY-2: bright yellow-2 CaMV35S promoter: 35S cauliflower mosaic virus promoter d: day EGFP: enhanced green fluorescence protein EXT: extensin FLAs: fasciclin-like AGPs GFP: green fluorescence protein GPI: glycosylphosphatidylinositol GUS: β-glucuronidase h: hour HRGP: hydroxyproline-rich proteins LB medium: Luria broth medium min: minute MS medium: Murashige and Skoog medium MW: molecular weight NOS terminator: nopaline synthase terminator nt: nucleotide PBS: phosphate buffered saline PCR: polymerase chain reaction PE: Pectinesterase PLC: phosphotidylinositol-specific phospholipase C PLD: phosphotidylinositol-specific phospholipase D PM: plasma membrane RLK: Receptor-like kinase 15

RT: room temperature RT-PCR: reverse transcriptase-polymerase chain reaction SDS: sodium dodecyl sulphate SDS-PAGE: sodium dodecyl sulphate-polyacrylamide gel electrophoresis TBS: Tris buffered saline T-DNA: transfer DNA VC: vector control WAK: wall-associated kinase WT: wild type XTH: Xyloglucan endotransglycosylases/hydrolase 17 over: transgenic plant overexpressing GFP-AtAGP17 fusion protein 18 over: transgenic plant overexpressing GFP-AtAGP18 fusion protein 19 over: transgenic plant overexpressing GFP-AtAGP19 fusion protein 17NG: transgenic plant overexpressing AtAGP17 (no GFP) 18NG: transgenic plant overexpressing AtAGP18 (no GFP) 19NG: transgenic plant overexpressing AtAGP19 (no GFP)

16

CHAPTER 1 INTRODUCTION

Plant cell wall proteins

Unlike animal cells, plant cells have strong and flexible cell walls surrounding

them. In growing plant cells, the cell wall is usually a thin layer that is composed of

complex polysaccharides and a small amount of structural proteins and cell wall enzymes

(Varner and Lin, 1989; Cosgrove, 2005). This thin, yet strong layer not only gives shape

to the plant cells, but also has key roles in plant growth and development, intercellular

communication, water movement and defense (Cosgrove, 2005).

The main polysaccharides in plant cell wall are cellulose, hemicellulose and

pectin. Cellulose is composed of 1,4,-β-D-glucose. Many glucan chains are bundled

together to form cellulose microfibrils that have strong mechanical properties (Cosgrove,

2005). Hemicellulose, mainly consists of xyloglucan and arabinoxylan, is highly

branched and crosslinks cellulose microfibrils to form a network (Carpita and Gibeaut,

1993; Cosgrove, 2005). This cellulose-hemicellulose network is embedded in gel-like matrix of pectins, which are complex polysaccharides including homogalacturonan,

rhamnogalacturonans I and II, galactans, arabinans and other polysaccharides (Cosgrove,

2005).

Glycine (Gly)-rich proteins (GRPs) and hydroxyproline (Hyp)-rich glycoproteins

(HRGPs) are the most abundant structural proteins in the plant cell wall (Keller, 1993;

Showalter, 1993). GRPs, as the name indicates, are rich in glycine. The amino acid

sequence of GRPs is usually arranged in Gly-X repeats, where X is most frequently Gly 17

but can also be alanine (Ala) or Serine (Ser) (Varner and Lin, 1989; Showalter, 1993).

Three major groups constitute the HRGP superfamily: proline(Pro)-rich proteins (PRPs),

extensins (EXTs) and arabinogalactan-proteins (AGPs) (Kieliszewski and Shpak, 2001),

and sometimes a fourth group, solanaceous lectins are also included (Showalter, 1993;

Sommer-Knudsen, et al., 1998).

PRPs are basic and lightly glycosylated. They have a characteristic Pro-Pro repeats that are embedded in some larger repeat units, such as Pro-Pro-X-Y-Lys, Pro-Pro-

Pro-X-Y-Lys, Pro-Pro-X-Lys and Pro-Pro-X-Y-Pro-Pro, where X and Y can be valine

(Val), tyrosine (Tyr), histidine (His) and glutamic acid (Glu) (Showalter, 1993; Sommer-

Knudsen, et al., 1998).

EXTs are basic proteins and are moderately glycosylated. They usually contain pentapeptide Ser(Hyp)4 repeats. The Hyp residues are usually glycosylated with 1-4 arabinosyl residues, and some Ser residues are glycosylated with a single galactose unit

(Showalter, 1993; Kieliszewski and Lamport, 1994).

Solanaceous lectins are chimeric proteins containing one carbohydrate-binding lectin domain and one extensin-like domain. The extensin-like domain probably reflects a close evolutionary relationship between the extensins and the solanaceous lectins

(Showalter, 1993; Sommer-Knudsen, et al., 1998).

AGPs are acidic and heavily glycosylated with arabinose and galactose

(Kieliszewski and Lamport, 1994). They often contain Ala/Ser-Hyp repeats (Gleeson et al., 1989). The sugar residues are usually linked to the Hyp residues in the core protein

(reviewed in Kreuger and van Holst, 1996). 18

Arabinogalactan proteins (AGPs)

Arabinogalactan-proteins (AGPs) are hyperglycosylated hydroxyproline-rich glycoproteins (HRGPs) which decorate plant cell surfaces in different plant species

(Fincher et al., 1983). They have molecular weights of 100 to 200 kDa (the molecular weights of the core proteins based on amino acid sequence are 13 to 15 kDa) and their isoelectric points vary from pH 2 to pH 5. They have stable chemical properties, such as resistance to heat, cold alkali treatment, and even resistance to proteolysis in their native state (Showalter, 1993; Kreuger and van Holst, 1996).

An AGP consists of a hydroxyproline-rich core protein which is decorated by arabinose and galactose-rich polysaccharide units. AGPs are highly glycosylated. Over

90% of an AGP is carbohydrate and less than 10% is protein (Kreuger and van Holst,

1996). It is suggested that the sugar moiety on the protein backbone has certain structures so they can act as signal molecules during plant development (Kreuger and van Holst,

1996).

AGPs can be specifically precipitated by the synthetic phenylazoglycoside dyes, the β–glucosyl Yariv reagents, to form a red-orange complex. This characteristic has provided a good way to study the localization of AGPs and to purify and quantify AGPs.

Moreover, Yariv reagents were also used in cell cultures and plants to perturb AGPs and to examine their functions (Guan and Nothnagel, 2004; Sardar et al., 2006).

An example of a commercially useful AGP is gum arabic, an exudate from the

Acacia senegal tree native to Africa. Gum arabic is used in food industry as emulsifier, in 19 medicinal uses, and in cosmetic and lithography industries (Majewska-Sawka and

Nothnagel, 2000).

AGP structure

Protein backbone

The composition of the protein backbone was determined by AGPs purification, deglycosylation and amino acid component analysis. The protein backbone is usually rich in hydroxyproline, alanine, glycine, threonine and serine. However, some AGPs have different amino acid compositions; for example, some are hydroxyproline-poor AGPs

(Showalter, 1993; Kreuger and van Holst, 1996; Sommer-Knudsen et al., 1998;

Showalter, 2001).

A special sequence for some of the AGPs is the C-terminal hydrophobic tail, which can be cleaved off and replaced by a glycosylphosphatidylinositol (GPI) anchor during posttranslational modification. A GPI anchor has an inositol phospholipid (Figure

1.1) which is embedded in the plasma membrane so the AGP is localized to the outer face of the plasma membrane (Oxley and Bacic, 1999; Showalter, 2001). Recently, a search program termed “GPT” (GPI-anchoring Prediction Tool) was devised by Borner et al.

(2002) to identify GPI-anchored proteins (GAPs) based on their common sequences. In

Arabidopsis, 210 proteins were identified as GAPs and later were validated by a 20

proteomic analysis. Among these GAPs, 13 were classical AGPs and 9 were AG peptides

(Borner et al., 2002; Borner et al., 2003).

a

b

Figure 1.1 Glycosylphosphatidylinositol (GPI) anchor structure and the consensus sequence for GPI anchor addition. (From Sun et al., 2004a and Showalter, 2001). a) GPI anchor structure for a pear plasma membrane-bound AGP. The GPI anchor is embedded in the outer leaflet of the plasma membrane. A partial β-galactosyl substitution (*) on the core oligosaccharide is shown. Potential cleavage sites of phosphotidylinositol- specific phospholipase C (PLC) and D (PLD) are also shown. Et, ethanolamine; Man, mannose; Gal, galactose; GlcN, N-acetylglucosamine. b) Consensus sequence in classical AGPs for the addition of a GPI anchor to the C- terminal of the AGP core protein. The amino acid residue designated ω is the site of GPI anchor addition, during anchor addition, the remaining C-terminal residues are removed.

The cleavage site (ω site) for GPI anchor addition has some common features.

Generally, the ω site and amino acids adjacent to it (ω-1…. ω+2) are small residues.

Upstream of the ω site is a polar and flexible linker region of about 11 residues (ω-11….

ω-1). Downstream of the ω site is a flexible spacer region (ω+3…. ω+9) of moderately

polar residues. Following the spacer region is the hydrophobic region (Eisenhaber et al.,

1998; Eisenhaber et al., 2003). To predict possible GPI anchor signals, Eisenhaber et al. 21

developed the “big-II plant” bioinformatics program and offered the program in the form

of a web server (http://mendel.imp.univie.ac.at/gpi/plants/gpi―plants.html) so that plant

researchers can conveniently analyse their protein using this new service (Eisenhaber et

al., 2003).

Carbohydrate

The sugar components in AGPs are mainly arabinose and galactose. The ratio of

arabinose:galactose varies between 10:90 and 15:85 (Fincher et al., 1983). They also

contain other less abundant sugars, such as glucuronic acid, rhamnose, mannose, xylose,

glucose, fucose, glucosamine and galacturonic acid. The carbohydrate chain is composed

of a 1,3-β-linked galactan backbone with 1,6-β-linked galactan side chains, which are

modified by 1,3-linked arabinose and other less abundant sugars. The size of this

polysaccharide chain varies from 30 to 150 sugar residues. In addition to the

polysaccharide chain there are also short arabinose oligosaccharide chains in some AGPs

(Fincher et al., 1983; Sommer-Knudsen, et al., 1998; Showalter, 2001).

The Hyp contiguity hypothesis

The carbohydrate chains are attached to the Hyp residues, and possibly to Ser

residues by O-linked glycosylation (Showalter, 2001). Different types of carbohydrate chains are attached to the Hyp residues at different locations based on the amino acid sequence, which can be summarized by the Hyp contiguity hypothesis (Kieliszewski and 22

Shpak, 2001). This hypothesis states that contiguous Hyp residues are glycosylated with short oligoarabinosides, while clustered noncontiguous Hyp residues are glycosylated with large polysaccharide chains. So far, this hypothesis has been tested in a few cases including some endogenous HRGPs and synthetic genes and the results supported the prediction (Kieliszewski and Shpak, 2001).

Molecular structure of AGPs

Two models have been proposed to describe the molecular shape of AGPs: the

“wattle blossom” model and the “twisted hairy rope” model (Figure 1.2) (Fincher et al.,

1983; Showalter, 2001). In the first model, the “wattle blossom” model, each polysaccharide unit on the protein backbone is folded into a sphere like structure and as a result, the whole AGP molecule has a spheroidal shape. While in the “twisted hairy rope” model, the polysaccharides and the oligoarabinosides wrap around the core protein resulting in a rod like structure (Qi et al., 1991). Both models were proved to be correct based on transmission electron microscopic imaging of different AGPs. For example, the structure of the Hyp-poor carrot AGPs and the non-classical tobacco AGP TTS supports the “wattle blossom” model, and the structure of a gum Arabic AGP corroborate the

“twisted hairy rope” model (Qi et al., 1991; Baldwin et al., 1993; Cheung et al., 1995;

Showalter, 2001). 23

Figure 1.2 Two models for molecular shapes of AGPs. (From Showalter, 2001). a) The “wattle blossom” model. b) The “twisted hairy rope” model.

AGP classification

There are many structurally diverse members in the AGP family. Traditionally,

AGPs are divided into two classes based on their core protein: “classical” AGPs and

“nonclassical” AGPs (or chimeric AGPs). Classical AGPs mainly contain Hyp, Ala, Ser,

Thr and Gly. They are composed of an N-terminal signal peptide, a Hyp-rich core sequence, and a GPI anchor addition sequence. Nonclassical AGPs (or chimeric AGPs)

have some other regions besides the hydroxyproline-rich core sequence, for example,

Cys-rich AGPs, Asn-rich AGPs, AGPs with fasciclin-like regions, etc. They don’t have

the C-terminal hydrophobic tail (Showalter, 2001). With our knowledge of AGPs

increased, the classification also changed. More recently, using a bioinformatics approach 24

involving searching for Pro-, Ala-, Ser- and Thr-rich proteins encoded by the Arabidopsis genome, Schultz et al. (2002) identified 47 putative AGP genes. Based on their protein sequences, these AGPs can be divided into several classes, including the classical AGPs, lysine-rich classical AGPs, AGP peptides, fasciclin-like AGPs (FLAs), and other chimeric AGPs (Figure 1.3).

Figure 1.3 Schematic illustration of different classes of AGPs in Arabidopsis. (Adapted and modified from Gaspar et al., 2001).

Cellular localization of AGPs

AGPs are widely expressed in the plant kindom, from lower plants such as

bryophytes and algae to higher plants such as angiosperms. At the organ level, AGPs are

present in leaves, stems, roots, flowers and seeds in higher plants. At the tissue level,

AGPs are found in various tissues, such as xylem, stylar transmitting tissue as well as 25

cultured cells derived from embryo, endosperm, root and leaf tissue (Fincher et al., 1983;

Showalter, 2001).

The subcellular localization of AGPs has been revealed by transmission electron

microscopy and subcellular fractionation methods (Showalter, 2001). In general, AGPs

are localized in plasma membranes, cell walls and intercellular spaces. As mentioned

before, a special structure in some AGPs that accounts for their localization is the C- terminal hydrophobic tail, which can be cleaved and replaced by GPI anchor when it is mature. So the AGPs are localized on the plasma membrane. AGPs can then be cleaved and released to the extracellular matrix by phospholipase C (PLC) or phospholipase D

(PLD) (Oxley and Bacic, 1999). In this case, these AGPs have multiple surface locations and their release is a dynamic process. During plant development, such release may be important for cell-cell signaling. The released AGPs, as soluble signals, may set up a

concentration gradient that provides distinct positional information to neighboring cells

for pattern formation (Seifert and Roberts, 2007).

Physiological function of AGPs

The complex carbohydrate structure on AGP molecules, the variety and abundance

of AGPs in the plant kingdom, and the localization of AGPs on cell surface imply that

AGPs may be involved in various and multiple phases of plant growth and development.

Although specific functions of AGPs remain uncertain, it is suggested that AGPs play important roles in many biological processes, including cell division and expansion, cell 26 differentiation, cell adhesion, pattern formation, fertilization, plant-microbe interactions and programmed cell death (Majewska-Sawka and Nothnagel, 2000; Showalter, 2001;

Faik et al., 2006; Seifert and Roberts, 2007). These studies often used antibodies against the carbohydrate chain or protein backbone to examine the AGPs expression patterns, and used β-Yariv reagent that specifically bind AGPs to study the biological effects of this AGP inhibitor on cell cultures. More recently, molecular genetics approaches, such as overexpression, RNAi techniques, and insertional mutants have also been employed to explore function(s) of AGPs.

Role of AGPs in embryogenesis

Pattern formation includes embryo pattern formation and postembryonic pattern formation. Early studies on carrot (Daucus carota L.) showed that by using one of the monoclonal antibodies against a specific carrot AGP epitope, the epitope was found to be expressed in specific patterns during development of somatic embryos in the cell cultures of carrot. This result indicates that expression of certain AGPs may have a role in establishing embryo pattern formation (Stacey et al., 1990; Kreuger and van Holst, 1993).

A further experiment supporting this idea is that the addition of carrot-seed AGPs re- induced the embryogenic potential in a two-year-old, non-embryogenic cell line or increased the percentage of embryogenic cells in the embryogenic cell line (Kreuger and van Holst, 1993). Similarly, in Norway spruce (Picea abies), the AGP fraction of the seed extract could stimulate the less developed somatic embryos to develop further into well 27 developed embryos (Egertsdotter and Arnold, 1995). Moreover, somatic embryogenesis was induced from the root tissues of Cichorium, but addition of β-Yariv reagent to the culture medium stopped somatic embryogenesis. When β-Yariv reagent was removed, somatic embryogenesis was re-induced similar to the control roots (Chapman et al.,

2000). These results have provided further evidence that AGPs are involved in embryo pattern formation.

Role of AGPs in xylem development

AGPs are also associated with postembryonic pattern formation, such as root, shoot and flower formation. JIM13 labeled AGP expression was spatially and temporarily related to xylem differentiation in Arabidopsis roots. The JIM13 AGP epitope appeared in a single xylem initial at first. Then during further development, the epitope was expressed in xylem parenchyma cells and the endodermis (Dolan et al., 1995; Seifert and

Roberts, 2007). Interestingly, further studies using the JIM13 antibody in other plants showed that the labeling pattern was somewhat species specific. In dicotyledons such as carrot, radish, pea and Arabidopsis, JIM13 labeled cells associated with xylem, while in monocotyledons such as onion and maize, JIM13 labeled cells associated with phloem

(Casero et al., 1998; Samaj et al., 1998; Majewska-Sawka and Nothnagel, 2000).

In addition to antibodies specific to carbohydrate chains on AGPs, antibodies raised against the protein region were also developed with the cloning of several cDNAs encoding AGPs. PtaAGP6 is a lysine-rich, classical AGP in loblolly pine (Pinus taeda 28

L.). In the typical AGP domain, it has a short lysine-rich region which is not glycosylated and can be used to produce antibodies. Immunolocalization studies with antibodies against this peptide region showed that PtaAGP6 expression was restricted to the cells that just precede secondary cell wall thickening, suggesting its role in xylem development

(Zhang et al., 2003).

AtFLA11, a fasciclin-like AGP in Arabidopsis, is also likely involved in xylem differentiation. Among the 48 AGPs in Arabidopsis, AtFLA11 had the highest expression level in inflorescence stems, which made researchers focus on this gene in the stem.

Immunostaining showed that AtFLA11 was specifically expressed in the sclerenchyma cells of stems and siliques, and its expression fluctuated during the cell maturation process, indicating a potential role in secondary wall formation (Ito et al., 2005).

Role of AGPs in reproduction

Upon pollination, pollen grains become hydrated and germinate to produce pollen tubes. The pollen tubes penetrate the stigmatic surface and grow through the stylar transmitting tract towards the ovary to reach the ovules where the egg cells are located.

AGPs are expressed abundantly in these reproductive tissues and they are likely to have important roles during the reproduction process (Cheung and Wu, 1999).

The adhesive stigmatic exudates are rich in AGPs, which imply that AGPs may have a role in adhesion when pollen grains contact the stigmatic surface. AGPs may also function in cell-cell recognition since the stigmatic surface allows the compatible pollen tube to penetrate and rejects the incompatible pollen sources (Fincher et al., 1983; 29

Cheung and Wu, 1999). A stigma AGP RT35 (cDNA clone AGPNa3) from Nicotiana alata was purified and characterized as a “nonclassical AGP” since the C-terminus contains a cysteine-rich domain. The expression level of AGPNa3 in pistils peaked during maturation and declined quickly afterwards to a very low level 5 days later, suggesting a specific, yet unclear role of RT35 in the pistil (Du et al., 1996; Cheung and

Wu, 1999).

The stylar transmitting extracellular matrix is enriched in lipids, polysaccharides and glycoproteins including AGPs. When pollen tubes elongate through the transmitting tract, this extracellular matrix is proposed to promote pollen tube elongation (Cheung,

1996; Cheung and Wu, 1999). This idea was supported by an in vitro pollen germination experiment. Pollen germination and pollen tube elongation can be accomplished either in vivo or in vitro on germination media. However, the in vitro pollen tube elongation was much slower than the in vivo process. When exudates from the lily transmitting tract were applied to the germination media, pollen tubes grown in vitro adhered to this media at their tip region and they elongated considerably faster than control pollen tube grown in the media without exudates (Jauh et al., 1997).

Although the above experiment provided evidence that the transmitting tissue extracellular matrix enhances pollen tube elongation, researchers still need to identify which component in the matrix has this function. An AGP from the styles of Nicotiana alata named RT25 (cDNA clone AGPNa1) was purified and characterized (Du et al.,

1994). Unlike RT35, RT25 is a “classical AGP” with a C-terminal hydrophobic domain.

In the mature protein, this domain was absent and replaced with a GPI-anchor (Youl et 30

al., 1998). Similar to AGPNa1 protein, LeAGP-1, a “classical AGP” from tomato with

high expression in the stylar transmitting extracellular matrix, was also a GPI-anchored

protein (Li and Showalter, 1996; Sun et al, 2004a). Compared to the other “nonclassical

AGPs” in the reproductive tissue, it will be interesting to examine if these GPI-anchored

proteins have distinct functions from other AGPs because of this special structure

(Cheung and Wu, 1999).

Another stylar transmitting-tissue specific AGP from Nicotiana tabacum, TTS

protein, was identified as a “nonclassical AGP” with a cysteine-rich domain. In vitro

pollen tube growth assays showed that TTS protein enhances pollen tube growth in vitro

and attracts pollen tubes grown in a semi-in vivo culture system. Moreover, transgenic

plants expressing the TTS antisense gene down-regulated TTS protein expression and as

a result, reduced the pollen tube growth rate as well as seed production (Cheung et al.,

1995). Interestingly, the sugar residues on the TTS proteins are specific to the stylar

transmitting tissue. Overexpression of TTS resulted in fully glycosylated TTS protein

only in the style, while in all other tissues, TTS polypeptides accumulated without proper

glycosylation. Transgenic plants overexpressing Agamous, a gene that regulates both

TTS expression and glycosylation, produced abnormal sepals that overexpressed

glycosylated TTS (Cheung and Wu, 1999; Majewska-Sawka and Nothnagel, 2000).

In Nicotiana tabacum, the ovaries have less AGPs than the stigmas and styles

(Cheung and Wu, 1999). Although AGPs in the ovary have speculative functions in

guiding pollen tubes into the ovules, purification and characterization of these AGPs has not been reported. Recently, an RNAi technique was used to silence AtAGP18, a 31

classical lysine-rich AGP in Arabidopsis. AtAGP18-silenced plants had a phenotype of

ovule abortion since functional megaspores failed to enlarge and divide, indicating a role

for AtAGP18 in initiation of female gametogenesis (Acosta-Garcia and Vielle-Calzada,

2004).

In addition to female reproductive tissues, AGPs were also present in male

gametophytes. The localization of AGPs in the male reproductive tissues of different species can be different. For example, in tobacco pollen tubes, AGPs were localized in

the cell wall along elongating pollen tubes but not at the pollen tube tip, while lily pollen

tubes had AGPs at the tip region (Li et al., 1992; Roy et al., 1998). Applying β-Yariv

reagent, a synthetic phenylazoglycoside dye that binds to AGPs, to in vitro germination

media had no effect on tobacco pollen tube growth, but inhibited lily pollen tube

elongation. It was suggested that β-Yariv reagent bound AGPs at the pollen tip and

interfered with cell wall assembly, thereby arresting pollen tube growth (Roy et al., 1998;

Cheung and Wu, 1999).

Role of AGPs in cell division and expansion

At the cellular level, AGPs can be involved in cell division and expansion.

Applying β-Yariv reagent to rose and carrot cell suspension culture affected cell

proliferation by inhibiting cell division (Serpe and Nothnagel, 1994; Thompson and

Knox, 1998). In undifferentiated sugar beet protoplasts, addition of the JIM13 antibody

also inhibited proliferation as the JIM13 antibody bound AGPs and presumably blocked 32 their functions (Butowt et al., 1999). Moreover, β-Yariv treatment of tomato whole seedlings inhibited root growth dramatically. In the affected roots, both root cell numbers and epidermal root cell elongation were reduced, suggesting that both cell division and expansion were compromised (Lu et al., 2001).

Several papers have shown the role of AGPs in cell expansion. In tobacco cell culture, addition of β-Yariv reagent strongly inhibited cell elongation (Vissenberg et al.,

2001). β-Yariv reagent also reduced cell elongation in Arabidopsis seedlings, causing a root epidermal cell bulging phenotype similar to the Arabidopsis reb1-1 mutant, which contained less AGPs in the root (Willats and Knox, 1996; Ding and Zhu, 1997). More recently, overexpression of CsAGP1, a classical AGP isolated from cucumber hypocotyls, in tobacco resulted in taller plants, suggesting its role in stem elongation

(Park et al., 2003). Moreover, in the moss Physcomitrella patens, several approaches, including β-Yariv reagent treatment, immunostaining, and gene knockout approaches, have demonstrated AGPs are required for apical cell extension (Lee et al., 2005).

Role of AGPs in cell signaling

The carbohydrate chains of AGPs are highly complex structures. Most regions of the AGP protein cores are modified by these carbohydrate chains and only limited regions may be accessible. These accessible regions on the protein moieties together with the information-rich carbohydrate moieties could serve as potential chemical signals in cellular signaling processes (Majewska-Sawka and Nothnagel, 2000; Showalter, 2001). 33

Five models of AGP signaling mechanism were proposed (Figure 1.4) based on limited experimental information and analogous molecules in animal models (Showalter, 2001).

First, oligosaccharides released from the carbohydrate chains could bind to a plasma membrane receptor to start cell signaling (Figure 1.4 a). Second, AGPs could directly bind to a plasma membrane receptor (Figure 1.4 b, left) or interact with ligand molecules for effective presentation to a ligand receptor on the plasma membrane (Figure 1.4 b, right). Third, AGPs could interact with a plasma membrane receptor on a neighboring cell to start intercellular signaling (Figure 1.4 c). Fourth, signal molecules derived from the GPI anchor such as phosphatidyl-inositol or ceramide after phospholipase cleavage could start intracellular signaling (Figure 1.4 d). Fifth, AGPs may aggregate on the plasma membrane and serve as adhesion molecules for plant growth and development

(Figure 1.4 e). These models, however, are hypothetical and speculative and need further experimentation to support them.

34

Figure 1.4 Models of AGP signaling mechanism. (From Showalter, 2001). a) Oligosaccharides (small squares) released from the sugar chains bind to a plasma membrane receptor. b) AGPs directly bind to a receptor or collect ligand molecules (diamonds) for presentation to a receptor. c) AGPs directly bind to a receptor on a adjacent cell to start cell-cell signaling. d) Cleavage of AGPs by phospholipase produces signal molecules from the GPI anchor. e) Aggregation of AGPs on the plasma membrane and serve as adhesion molecules to form a plasma membrane-cell wall connection.

35

Lysine-rich AGPs

Lysine-rich AGPs, as mentioned previously, belong to the classical AGPs. They

contain a characteristic lysine-rich subdomain in the hydroxyproline-rich AGP domain.

In 1996, a lysine-rich AGP gene for tomato, LeAGP-1, was cloned and sequenced.

Northern blot analysis in tomato indicated this gene was highly expressed in stem and flowers, moderately expressed in root and green fruit, and weakly expressed in leaves and red fruit. In stems, expression levels were high in young internodes and low in old internodes (Li and Showalter, 1996). Following that, LeAGP-1 was isolated from tomato cell culture and verified to be a bona fide AGP by β-Yariv precipitation and amino acid composition analysis (Gao et al., 1999). The basic lysine-rich subdomain, which is not glycosylated, was used to generate an antibody (PAP antibody) against this region.

Immunolocalization studies with this antibody showed LeAGP-1 was present in certain cells associated with cell wall thickening and lignification, and in stylar transmitting tissue cells (Gao et al., 1999; Gao and Showalter, 2000).

Several experiments, such as immunolocalization of LeAGP-1 on tomato protoplasts, expression of a GFP-LeAGP-1 fusion protein in tobacco cell culture, and isolation of plasma membrane followed by western blotting, supported the hypothesis that LeAGP-1 was a GPI-anchored protein localized on the plasma membrane (Zhao et al., 2002; Sun et al., 2004a). The subcellular localization of LeAGP-1 to the plasma membrane and its developmentally regulated expression suggested several functions in growth and development. To elucidate specific function(s) of LeAGP-1, GFP-LeAGP-1 36

fusion protein was expressed under the control of the 35S cauliflower mosaic virus

(CaMV35S) promoter in tomato. Transgenic tomato plants were shorter, highly branched

and produced less seeds, suggesting LeAGP-1 may have functions in vegetative growth

and reproduction (Sun et al., 2004b).

Besides LeAGP-1, lysine-rich AGPs in other species were also identified and

examined, such as NaAGP4 in tobacco, AtAGP17, 18 and 19 in Arabidopsis, CsAGP1 in

cucumber and PtAGP6 in pine (Gilson et al., 2001; Schultz et al., 2002; Park et al., 2003;

Zhang et al., 2003; Sun et al., 2005). The model plant Arabidopsis has several advantages: it has a small genome and a short life cycle, it is easy to manipulate genetically, and a pool of DNA insertion mutants exist for this plant. These advantages made researchers become interested in the lysine-rich AGPs in Arabidopsis, namely

AtAGP17, 18 and 19. The rat1 mutant (resistant to Agrobacterium tumefaciens) had a T-

DNA insertion in the promoter of AtAGP17 that resulted in a decreased efficiency of

Agrobacterium root transformation, suggesting a role of AtAGP17 in Agrobacterium binding to roots (Gaspar et al., 2004). In addition, using RNAi, AtAGP18 was implicated to function in female gametogenesis (Acosta-Garcia and Vielle-Calzada, 2004).

Moreover, a recent study on a T-DNA mutant of AtAGP19 showed pleiotropic

phenotypes in this mutant (Yang et al., 2007). These studies have provided valuable clues

about specific functions of Arabidopsis lysine-rich AGPs, which could also guide

functional study of lysine-rich AGPs in other species.

37

Specific aims of this dissertation research

DNA microarray analysis is a high-throughput technology which can examine

expression levels of thousands of genes in one experiment. With complete sequencing of

the Arabidopsis genome, standard Arabidopsis microarray chips became commercially

available (for example, the Affymetrix ATH1 chip) and greatly assisted research in

Arabidopsis. As already mentioned, atagp19, a mutant with a T-DNA insertion in

AtAGP19, has several phenotypes. To elucidate which genes change the expression levels in the mutant, it is appropriate to use a microarray approach to monitor the expression of more than twenty thousand genes in Arabidopsis.

Since AtAGP17, 18 and 19 were identified by a bioinformatics approach, their expression patterns, subcellular localization in tobacco cell culture, and carbohydrate composition and linkage were studied (Gaspar et al., 2004; Sun et al., 2005; Yang et al.,

2007; Yang and Showalter, 2007). However, the exact functions of these AGPs were not

determined. Here, I plan to overexpress these AGPs in Arabidopsis and study their

localization and functions in transgenic Arabidopsis plants.

Ten AG peptides in Arabidopsis were identified in 2002 (Schultz et al., 2002).

Since then, no AG peptide mutant work has been carried out on these AG peptides. In

this research, one of the AG peptides, AtAGP14 was selected because at the beginning of

this study only AtAGP14 had a mutant with the T-DNA insertion site at the exon region.

38

This dissertation research is focused on the following questions:

1. Which genes change their expression levels in the atagp19 mutant compared to

the wild type Arabidopsis? How are they related to the mutant phenotypes?

2. Where are the GFP-AtAGP17/18/19 fusion proteins located when they are transformed into Arabidopsis? When AtAGP17/18/19 are overexpressed in Arabidopsis, do these transgenic plants have any phenotypes?

3. What is the expression pattern of AtAGP14? Is there any phenotype in the atagp14 T-DNA mutant?

39

CHAPTER 2 T-DNA MUTANT STUDY OF ATAGP19

Summary

AtAGP19 is a member of the lysine-rich classical arabinogalactan-protein (AGP)

gene subfamily in Arabidopsis. AtAGP19 encodes an AGP with a glycosylphosphatidylinositol (GPI) anchor for subcellular localization to the plasma membrane and is hypothesized to act as a signal transduction molecule. Utilizing a null

T-DNA insertion mutant of AtAGP19 which displays pleiotropic phenotypes, a microarray approach was employed to elucidate changes in gene expression associated with the atagp19 mutant. The expression levels of two homologous lysine- rich classical

AGP genes, AtAGP17 and AtAGP18, did not change significantly, indicating these two genes do not compensate for the loss of AtAGP19. In contrast, the expression of several genes related to cell expansion were found to change significantly. Interestingly, one gene (At1g68720, cytidine/deoxycytidylate deaminase family protein) adjacent to

AtAGP19 was found to be down-regulated about 50 fold and an RT-PCR experiment confirmed an alteration of this gene in the atagp19 mutant. Furthermore, complementation with the 3’ portion of the At1g68720 gene (1.8 kb) can fully restore all the wild type phenotypes, indicating this partial region is critical for the functions revealed in the atagp19 mutant. 40

Introduction

AtAGP17 (At2g23130), AtAGP18 (At4g37450) and AtAGP19 (At1g68725) comprise the lysine-rich classical AGP family. Each of these family members encodes an N- terminal signal peptide, a central Pro/Hyp-rich AGP domain that is interrupted by a short basic lysine-rich region near its C-terminus, and a C-terminal glycosylphosphatidylinositol (GPI) anchor domain (Figure 2.1a). Several Arabidopsis mutants were examined in order to elucidate functions of these AGPs. The Arabidopsis rat1 (resistant to Agrobacterium transformation) mutant with a T-DNA insertion in the promoter region of AtAGP17, exhibits reduced binding of Agrobacterium, resulting in resistance to Agrobacterium tumefaciens root transformation (Gaspar et al., 2004). The

AtAGP18 RNAi mutant demonstrates defective ovule development as functional megaspores fail to enlarge and divide (Acosta-Garcia and Vielle-Calzada, 2004).

In contrast to the above two mutants, the atagp19 T-DNA knockout mutant is a pleiotropic mutant. Compared to wild type plants, the atagp19 mutant had: 1) lighter green leaves containing less chlorophyll and anthocyanins, 2) smaller and rounder leaves, with shorter petioles, 3) reduced hypocotyls length, 4) shorter and thinner inflorescence stems, 5) slower growth with delayed and reduced flowering, 6) fewer siliques and seeds and 7) fewer lateral roots (Yang et al., 2007). Microscopic examination of the mutant reveals these phenotypes are the result of reduced cell numbers and cell size in several organs, indicating compromised cell division and expansion in the mutant. Furthermore, complementation of this mutant with the wild type AtAGP19 gene restored all the wild type phenotypes (Yang et al., 2007). 41

How can a mutation in AtAGP19 elicit these phenotypic changes? In order to begin to address this question, a microarray approach was used to elucidate changes in gene expression associated with the atagp19 mutant. Here, these changes in the gene expression, particularly those observed for genes that are related to cell expansion, are reported and discussed as they provide insight to a potential signaling network that links

AtAGP19 function to cell expansion. Surprisingly, the gene (At1g68720) adjacent to

AtAGP19 is also down-regulated dramatically, and complementation with the AtAGP19 promoter region (1.8 kb upstream of the AtAGP19 start codon, including a portion of this gene) can fully restore all the wild type phenotypes. Here, the possible explanation of the role of the gene (At1g68720) and its relationship to the mutant phenotypes are also discussed. 42

Materials and methods

Plant materials and growth conditions

Wild type Arabidopsis (Columbia-0 ecotype) plants were grown side by side with homozygous atagp19 mutant plants obtained from Arabidopsis Biological Resource

Center (ABRC) under identical environmental conditions, namely 24°C, with 16h day/8h night cycle with illumination (100µE) for 14 days (Figure 2.1b).

Microarray experiments

Total RNA was extracted from the aerial portions of 14 day-old wild type and atagp19 plants using the Qiagen RNeasy Plant Mini Kit. The plants were grown in three batches separately under identical growth conditions. Each batch included wild type plants and atagp19 mutant plants, and as a result, 6 RNA samples were prepared as 3 biological replicates for the microarray experiments. RT-PCR was performed for each

RNA sample to ensure AtAGP19 mRNA was absent in atagp19 mutant samples (Figure

2.1c). RNA samples were sent to the UCI DNA & Protein MicroArray Facility

(University of California, Irvine) for microarray analysis. Total RNA samples were quality assessed on an Agilent Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA).

Double-stranded cDNA was synthesized from the total RNA with poly (T)-nucleotide primers using the SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen Corp.,

Carlsbad, CA). The BioArray High-Yield RNA Transcript Labeling Kit (T7) (Enzo 43

c

Figure 2.1 AtAGP19 gene structure and the atagp19 mutant. a) Schematic illustration of AtAGP19 gene structure. The location of the T-DNA insertion mutation in atagp19 (SALK_038728) is indicated with the triangle. The colored boxes represent exons encoding different protein domains as indicated in the key. The line separating the boxes represents the lone intron in this gene. b) Two-week-old Arabidopsis wild type and atagp19 mutant seedlings grown under identical conditions. RNA was harvested from the aerial portions of these seedlings and used for microarray analysis. Bar represents 1cm. c) RT-PCR results of AtAGP19 expression in three batches of RNA samples. Lane1, WT- B12; lane 2, 19M-B12; lane 3, WT-B14; lane 4, 19M-B14; lane 5, WT-B15; lane 6, 19M- B15.

44

Diagnostics Inc., Farmingdale, NY) was used to produce biotin-tagged cRNA from the cDNA. Fifteen μg of the cRNA was fragmented to the length of 35-200 bases, and 10 μg

of this fragmented cRNA was hybridized to Arabidopsis ATH1 GeneChip expression

arrays in an Affymetrix GeneChip Hybridization Oven 640 at 45°C with rotation for 16

h. The arrays were washed, stained with SAPE (streptavidin-phycoerythrin) on an

Affymetrix Fluidics Station 450 (Affymetrix, Santa Clara, CA), and scanned on a

GeneChip Scanner 3000 (Affymetrix). The results were quantified and analyzed by

GCOS 1.1.1 software (Affymetrix) and ArrayAssist 3.0 software (Stratagene, La Jolla,

CA) using default values (Figure 2.2). Briefly, GeneChip files generated by GCOS 1.1.1

software were provided by the UCI DNA & Protein MicroArray Facility and were

analyzed by ArrayAssist 3.0 software. Further, background correction and normalization

were performed using the MAS5 algorithm (Affymetrix). Wild type data were chosen to

define the baseline and a two-class unpaired T-test was performed for significance

analysis. A volcano plot was generated from the result of this significance analysis. 45

Figure 2.2 Experimental strategy to elucidate RNA expression levels in wild type Arabidopsis and atagp19 mutant plants by microarray analysis.

Quantitative PCR (Real-time PCR)

Total RNA, the same as that used in the microarray experiments, was reverse- transcribed into cDNA and labeled with 2-Step Brilliant SYBR Green QRT-PCR Master

Mix (Stratagene). Quantitative PCR was performed using Mx3000P QPCR System

(Stratagene) and CT values were determined using this system. Relative quantitation was 46 calculated based on actin expression by the method (Livak and Schmittgen,

2001). Briefly, each sample was run in triplicate and the CT value was averaged. The

ΔΔCT value was calculated based on the equation ΔΔCT = (CT,Candidate gene − CT,Actin) Mutant

− (CT,Candidate gene − CT,Actin) Wild type. The fold change for each candidate gene was equal to

. Gene-specific primers were designed with IDT (Integrated DNA Technologies,

Coralville, IA) PrimerQuest software and Vector NTI software as shown in Table 2.1.

Table 2.1 Primers used for quantitative PCR

Primers AGI Locus Forward Reverse At2g15020 5'-GGACTTGGGAATCTTGGTTTGAA-3' 5'-GACCATGAGCTTCTGTGGCATAC-3' At1g72490 5'-GCGAGTAATCCGTGAGGCCAAT-3' 5'-TTTCGGGTGGATGCAGAACAAG-3' At3g28240 5'-CGTCGTGACCAACCATGGAGAA-3' 5'-TCGGCCGCTGGATTTGCTAC-3' At2g03540 5'-ACACGCACAACGGCGAACAA-3' 5'-TGCCATGTTTGCGCTCGTTAC-3' At1g02820 5'-GCCGGGTCAATCTCCTCAGAAAC-3' 5'-CTCGCTAACGCTAAGATCCAATCC-3' At1g25550 5'-GTGGAACCAATCACCAGATCCACA-3' 5'-TTGAAACGGTTGAAACGCACCAGC-3' At3g04640 5'-TTCGATTACGTCACCGTCTGGCTT-3' 5'-AACTGCAACACCACTTTGCTCACG-3' At4g00050 5'-TGGAACGTTGGAGTCGATAGTGGA-3' 5'-TGCTCGTGTACTAGGTTGGAGCAA-3' At4g23260 5'-TCTGGACCAGGCAGATGCTTTCTT-3' 5'-TGGATCCATGACGAGCAATCCAGA-3' At1g78270 5'-CGCTGCGTCAATTGTGAAGCTCAT-3’ 5'-AGACGTCGACGCTAAGCAAGACA-3’ At3g14310 5'-AGCTCGTCCAGCGTCTCATCAATA-3' 5'-TCGAGGCATCGGTGAACCTAACAA-3' At5g57550 5'-CAGGGACGAGTTTGAGCTGCATG-3' 5'-TTCGCAGGAACGTTCGACACAG-3' At1g21270 5'-GATGTGGCAATGTCGCAGTTGAG-3' 5'-CTGGCATGTTGCCAAAGAAGAGC-3' At5g18690 5'-GAAGAAACCGGCGGAGAAGCAA-3' 5'-CGCTTTACCTTCAGACGATGGCA-3' At1g75780 5'-TCGTTACGTTCCTCGTGCTGTTCT-3' 5'-TTTCGCCCAATTATTTCCGGCACC-3' At1g10550 5'-GGAGGATTTGTCGAGAGTGAGCTT-3' 5'-CGTGACACGCCTCACAGATCAA-3' At5g58670 5'-GAACCGCAAAGGAGGGTTGAAGAA-3' 5'-AGTCTCAAGCCACTGCTCACTCAT-3' At2g37620 5'-GTGCTCGACTCTGGAGATGGTGTG-3' 5'-CGGCGATTCCAGGGAACATTGTGG-3' (actin)

47

Breaking force measurement

An Ohaus pull-spring scale (capacity 500g; Fisher, Waltham, MA) was used to

measure breaking force of inflorescence stems from the wild type and atagp19 mutant

plants. The top 5cm of the inflorescence stems from 5-week, 7-week and 9-week old

plants were used. One end of the stem was fixed on a clamp attached to the pull-spring

scale, and a pulling force was applied manually on the other end of the stem until it was

broken. Readings for the breaking force were recorded from the scale.

Complementation construct design with the AtAGP19 promoter sequence

The primers for AtAGP19 promoter (1836 nucleotide upstream of the start codon)

were designed to introduce a restriction enzymes site PacI (TTAATTAA) in the forward

primer 5’- CCG GTT AAT TAA TGA AAC TGC CTA GTC GGA ACC TGA -3’ and a

SacI (GAGCTC) site in the reverse primer 5’- AAT CAG AGC TCT GTG TTG TGG

AGG AAG CTA CAA GA -3’. The AtAGP19 promoter sequence was amplified. The

PCR products and pMDC110 binary vector (Curtis and Grossniklaus, 2003) were

digested with PacI (TTAATTAA) and SacI (GAGCTC) and ligated to form the complementation construct.

48

DNA extraction and amplification of AtAGP19 promoter

Genomic DNA was extracted from wild type Arabidopsis whole plants with the

Plant DNAzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions with minor modifications. Briefly, 100 mg plant material was ground in liquid nitrogen and DNA was extracted with 0.3 ml Plant DNAzol and 0.3 ml chloroform. Then the DNA in the aqueous phase was precipitated with 0.225 ml of 100% ethanol. The DNA pellet was first washed with 0.3 ml Plant DNAzol-ethanol wash solution (Plant DNAzol:ethanol = 4volumes:3volumes) and then 0.3 ml 75% ethanol. The

DNA pellet was air-dried and dissolved in 50 μl EB (elution buffer, 10 mM Tris·Cl, pH8.5).

The DNA sample was diluted and quantified on a Biophotometer (Eppendorf,

Hamburg, Germany). Absorbance at 260nm (A260) and 280nm (A280) was measured.

DNA concentration was calculated assuming one A260 unit equals 50 ng of double-

stranded DNA/μl. DNA quality was determined by the A260/ A280 ratio.

The AtAGP19 promoter sequence was amplified using AccuPrime Pfx DNA

polymerase (Invitrogen, Carlsbad, CA) with a final concentration of Mg2+ at 1.0 mM. To

reduce the chance of getting mutations during amplification, 100-200 ng genomic DNA

was used as template and PCR was performed for 25-28 cycles. The PCR products were

purified with the QIAquick PCR Purification Kit (QIAGEN, Valencia, CA).

49

Restriction digestion

The purified PCR products were digested with PacI and SacI (New England

BioLabs, Beverly, MA) at 37°C in a waterbath for 2 h, then purified with the QIAquick

PCR Purification Kit. The vector pMDC110 was also digested with PacI and SacI and the large fragment was purified with the PureLink Gel Extraction Kit (Invitrogen,

Carlsbad, CA). The purified DNA samples were quantified by the A260/ A280 ratio and

also on agarose gels by comparing brightness of sample bands to a 1kb DNA ladder

(New England BioLabs, Beverly, MA) of known quantities.

Ligation and transformation

The purified vector and insert were ligated at a molar ratio of 1:3 with the Quick

Ligation Kit (New England BioLabs, Beverly, MA). The amount of vector was 50 ng.

The ligation products were transformed into Subcloning Efficiency DH5α chemically

competent cells (Invitrogen, Carlsbad, CA) using heat shock transformation (protocol

according to New England BioLabs, Beverly, MA).

Construct verification by restriction digestion

Single colonies were picked from the plate with toothpicks and cultured overnight

in 3 ml LB media containing 50 μg/ml kanamycin. Plasmid DNA from the culture was 50

extracted with QIAprep Spin Miniprep Kit (QIAGEN, Valencia, CA) and digested with

different restriction enzyme combinations to confirm the construct was correct (the

enzyme combinations were: PacI and SacI; HindIII and SacI; SphI and HindIII).

Preparation of Agrobacterium competent cells

Five ml of a fresh stationary culture of Agrobacterium strain LBA4404 was inoculated into 500 ml LB medium and incubated at 28ºC with vigorous agitation overnight until OD600 reached 0.5-0.8. The culture was then chilled on ice and centrifuged at 4000 g for 10 min at 4°C. The pelleted cells were washed and centrifuged

three times with 500 ml, 250 ml and 50 ml of ice-cold sterile distilled H2O, respectively.

Finally, cells were resuspended in 5 ml of 10% (v/v) ice-cold sterile glycerol, aliquoted

and stored at -80°C (Weigel and Glazebrook, 2002) until use.

Transformation of Agrobacterium using electroporation

Agrobacterium competent cells were thawed on ice and mixed with different

constructs, respectively (50 μl of competent cells and 1 μl of construct made from

plasmid miniprep for each transformation). The mixture was transferred to a prechilled 2

mm cuvette and electroporation was carried out using the GENE PULSER II system

(Bio-Rad, Hercules, CA) with the following conditions: capacitance, 25 μF; voltage, 2.4

kV; resistance, 200 Ohm; pulse length, 5 msec. The cells were resuspended to 1 ml of LB 51

medium and incubated at 28°C for 4 h with gentle shaking. The cells were then centrifuged at 3000 rpm for 1 min and spread on LB plates containing kanamycin (25

μg/ml) and streptomycin (25 μg/ml). The plates were incubated at 28°C for 3 d.

Single colonies were picked from the plate with toothpicks and cultured overnight in 3 ml LB media with 25 μg/ml kanamycin and 25 μg/ml streptomycin. PCR and plasmid minipreps of Agrobacterium were carried out to verify the presence of the desired constructs.

Verified Agrobacterium cultures containing the proper constructs were mixed

with sterile 60% glycerol to reach a final glycerol concentration of 15%, and stored at

-80°C. After retrieving Agrobacterium stocks from the freezer, PCR was performed again to verify the presence of the correct constructs.

Floral dip transformation of Arabidopsis

The floral dip transformation was performed according to Clough and Bent’s method (Clough and Bent, 1998) with minor modifications. Briefly, Arabidopsis plants

were grown in soil pots covered with bridal veil, under either long day or short day

conditions until many immature flower buds appeared (the first bolts were clipped to

encourage more secondary bolts). The Agrobacterium carrying the correct construct was

inoculated in 500 ml LB medium with 25 μg/ml kanamycin and 25 μg/ml streptomycin

and grown at 28°C with vigorous shaking for 16 to 24 h until OD600 reached 0.8 or

higher. The Agrobacterium cells were spun down by centrifugation at 5,000 g for 10 min 52

and resuspended in a 5% sucrose solution to OD600 = 0.8. Silwet L-77 (LEHLE SEEDS,

Round Rock, TX) was added to the Agrobacterium solution to a concentration of 0.02%

or 0.05%. Arabidopsis inflorescence stems and rosettes were dipped in Agrobacterium

solution for 10-25 sec with gentle agitation, laid on their sides, covered with plastic wrap

and put in a cool place (such as under the bench) for 24 h. The plants were then transferred back to the growth room to grow normally. Sometimes, a second transformation was performed a week after the initial dip. After about one month and half, seeds were collected and dried in paper bags for 3 weeks.

Screening of transgenic plants

Kanamycin, streptomycin and hygromycin (Research Products International,

Mount Prospect, IL) stocks were prepared in ddH2O to a final concentration from 20

mg/ml to 50 mg/ml and sterilized with 0.20 µm filter (Fisher Scientific, Ireland). Sterile

stock solutions were aliquoted and stored at -20°C.

For hygromycin selection, seeds (T1 generation) from the dipped plants were

washed briefly in 70% ethanol and then incubated in 30% bleach supplemented with

0.1% Triton X-100 for 10 min with gentle shaking. Sterilized seeds were rinsed in sterile

ddH2O at least 6 times and then sown on MS plates containing 20 μg/ml hygromycin.

After stratification in the dark for 3 d at 4°C, the plates were exposed to light for 3 h to

promote germination and then kept in the dark at room temperature for 3 d. Hygromycin-

resistant seedlings had long hypocotyls with closed cotyledons, while hygromycin- 53

sensitive plants had short hypocotyls with open cotyledons. The potential transformants

containing the hygromycin-resistant gene were transferred to MS medium containing

0.5% sucrose to grow for 7 d without hygromycin before transferring them to soil.

Verification of transgenic plants by DNA extraction and PCR

Leaves from 4 week old potential transgenic plants were used to extract DNA

with the Extract-N-Amp Plant PCR Kit (Sigma-Aldrich, St. Louis, MO). Alternatively,

the CTAB method was used to extract genomic DNA (Weigel and Glazebrook, 2002).

PCR with primers corresponding to the hygromycin resistant gene was performed to

verify the presence of the construct. The primers specific to the hygromycin resistance

gene were: 5’-GGT TTC CAC TAT CGG CGA GTA CTT CTA C-3’ and 5’-AGA TCG

TTA TGT TTA TCG GCA CTT TGC ATC-3’.

RNA extraction from transgenic plants and RT-PCR

Total RNA was extracted from each transgenic plant using the RNeasy Plant Mini

Kit (QIAGEN, Valencia, CA). RNA samples were diluted and quantified on a

Biophotometer (Eppendorf, Hamburg, Germany). Absorbance at 260nm (A260) and

280nm (A280) was measured. RNA concentration was calculated assuming one A260 unit equals 40 ng of RNA/μl. RNA quality was determined by the A260/ A280 ratio. 54

RT-PCR was performed with the OneStep RT-PCR Kit (QIAGEN, Valencia, CA) according to the manufacturer’s instructions. The primers for AtAGP19 were 5' - CTG

CTC TCA TCT CTT CCT TTA GTG - 3' and 5' - ATT GAG CCA CAT TAC TGC TCT

TCC - 3'. The primers for At1g68720 were 5’- GCT GTG CTT GTT CAT GAT GG-3’ and 5’-GGA TGG AAT GGA TGA ACT GG -3’. The primers for actin as an internal control were 5' - GTG CTC GAC TCT GGA GAT GGT GTG - 3' and 5' - CGG CGA

TTC CAG GGA ACA TTG TGG - 3'.

The PCR conditions were: 50°C for 30 min, 95°C for 15 min, followed by 35 cycles of 94°C, 30 sec; 50°C, 30 sec; 72°C, 1 min and a final extension at 72°C for 10 min.

PCR amplification for At1g68720 sequencing

Eight pairs of primers were designed to sequence gene At1g68720 in the atagp19 mutant. The primer sequences were: (F1) 5' - GAA TCA ACA CAA GCG TAG ACA

ATG GAC C - 3', (R1) 5' - CGG TCC GGT TTG ATT GAG TTG GTA - 3', (F2) 5' -

CGA GTT TAT CTG AAG AAG CCC AAC - 3', (R2) 5' - CCT TAA TCT CCT TTC

GCC TAC CAA - 3', (F3) 5' - GTT GTG GTG GTA GAA GTT GTT CAG - 3', (R3) 5' -

CAC TGC TAT AAC GAG TCT CCA ATC - 3', (F4) 5' - GGA AAT TCA TGA GGT

CCA TGT CAA CG - 3', (R4) 5' - GGC TGT GGT ATA ACC ACT TTC TGA AGT

TCC - 3', (F5) 5' - TTC CTC CTT CTT CTC AGC TGG TAA - 3', (R5) 5' - AAG CTG

TTT CAG GAA GAC GGA ACT - 3', (F6) 5' - TTA GAG GGT GGC TCA GTA TCG 55

AAT - 3', (R6) 5' - CAT CAG CAC CCC AAA ACG TAA TTC - 3', (F7) 5' - TTG TGT

TTC AGA GTA GAG GAG CTT CG - 3', (R7) 5' - GAA GCC TCT GAT CCA TTT

CCT TCG - 3', (F8) 5' - AAT GTG TGC GGG AGC AAT ACT TCA - 3', (R8) 5' - CAT

TGT GTT GTG GAG GAA GCT ACA AG - 3'.

Results

Quality assessment of RNA samples by Agilent Bioanalyzer

Six RNA samples (3 wild type and 3 mutant samples) were quality assessed on an

Agilent Bioanalyzer prior to the microarray experiment. All 6 samples had two sharp 18S

and 28S peaks, and low signal in between the peaks and in the lower molecular range,

indicating integrity of all the RNA samples (Figure 2.3). In addition, the area of 28S:18S

was close to 2 in each sample from the Electropherogram Result table (data not shown),

further indicating good RNA quality. 56

Figure 2.3 Electropherogram summary of RNA samples by Agilent Bioanalyzer. The two sharp peaks (18S and 28S) indicate the integrity of mRNA. The 18S:28S rRNA ratio of each sample is close to 2, indicating good RNA quality.

57

Overview of changes in gene expression in the atagp19 mutant

Two-week-old atagp19 mutant and wild type Arabidopsis seedlings were chosen

for microarray analysis because this was the earliest time point when the mutant

phenotypes became obvious and the plants were of sufficient size for harvesting the aerial

portions for RNA extraction. At later times, changes in gene expression are increasingly

likely to be caused by downstream effects. In analyzing the microarray data, a threshold level of a 2-fold change in gene expression in the mutant compared to the wild type was set (Figure 2.4). With this criterion, 72 genes were found to be up-regulated and 39 genes were found to be down-regulated in the atagp19 mutant (Table 2.2). These genes were categorized by Affymetrix’s annotation of gene functions. Figure 2.5 shows the candidate genes categorized by their predicted cellular location. However, a large number of these

genes had no annotation. Particular attention was given to annotated genes likely to be

involved with cell wall biosynthesis and modification and cell expansion (Table 2.3). The predicted cellular location of most of these genes is in the endomembrane system (Figure

2.5).

Table 2.2 Summary of genes with mRNA levels up-regulated or down-regulated at least 2-fold in atagp19 mutant plants compared to wild type control plants Number of genes Fold change Up-regulated Down-regulated >10 6 3 5-10 5 2 4-5 8 2 3-4 10 10 2-3 50 15 Total (≥2) 79 32

58

Figure 2.4 Volcano plot of the microarray results. Differential expression values (log2 ratio of mutant to wild type) of each gene were plotted against p-values. A threshold level of a 2-fold change in differential expression and a 0.05 p-value was set. Genes which fall in the lower left (down-regulated) and lower right (up-regulated) region have a 2-fold change or greater change with p-values of <0.05.

endomembrane system, 13%

unclear, 38% unclear, 41% endomembrane system, 44% chloroplast, 23%

microtubule, 3% peroxisome, 3% mitochondrion, chloroplast, 9% nucleus, 4% nucleus, 3% 16% mitochondrion, 3% Up-regulated Down-regulated Figure 2.5 Genes with more than a two-fold change categorized by their predicted cellular location.

59

Table 2.3 Genes up-regulated or down-regulated at least 2-fold in atagp19 (continued on pp. 60-62)

Probe Set ID AGI Locus Gene Title Fold change

265892_at At2g15020 Expressed protein + 40.0

260456_at At1g72490 Hypothetical protein + 17.0

256634_s_at At3g28240 Hypothetical protein + 16.0

265709_at At2g03540 Hypothetical protein predicted by genscan + 13.2

262113_at At1g02820 Late embryogenesis abundant 3 family protein / LEA3 family protein + 11.9

261996_at At1g33830 Avirulence-responsive family protein / avirulence induced gene (AIG1) family protein + 11.3

255177_at At4g08040 1-aminocyclopropane-1-carboxylate synthase, putative / ACC synthase, putative + 7.0

264514_at At1g09500 Cinnamyl-alcohol dehydrogenase family / CAD family + 6.2

264866_at At1g24140 Matrixin family protein + 5.8

252411_at At3g47430 Peroxisomal biogenesis factor 11 family protein / PEX11 family protein + 5.3

256526_at At1g66090 Disease resistance protein (TIR-NBS class), putative + 5.0

256096_at At1g13650 Expressed protein + 4.8

247452_at At5g62430 Dof-type zinc finger domain-containing protein + 4.6

251745_at At3g55980 Zinc finger (CCCH-type) family protein + 4.5

245119_at At2g41640 Expressed protein + 4.5

265668_at At2g32020 GCN5-related N-acetyltransferase (GNAT) family protein + 4.4

251028_at At5g02230 Haloacid dehalogenase-like hydrolase family protein + 4.2

263779_at At2g46340 Phytochrome A supressor spa1 (SPA1) + 4.1

256427_at At3g11090 LOB domain family protein / lateral organ boundaries domain family protein (LBD21) + 4.0

260603_at At1g55960 Expressed protein + 3.9

255016_at At4g10120 Sucrose-phosphate synthase, putative + 3.8

249774_at At5g24150 Squalene monooxygenase 1,1 / squalene epoxidase 1,1 (SQP1,1) + 3.8

251727_at At3g56290 Expressed protein + 3.5

267505_at At2g45560 Cytochrome P450 family protein + 3.3

251658_at At3g57020 Strictosidine synthase family protein + 3.3

257485_at At1g63580 Protein kinase-related + 3.1 60

249942_at At5g22300 Nitrilase 4 (NIT4) + 3.1

265184_at At1g23710 Expressed protein + 3.0

267138_s_at At2g38210 Ethylene-responsive protein, putative + 3.0

263796_at At2g24540 Kelch repeat-containing F-box family protein + 2.96

246998_at At5g67370 Expressed protein + 2.94

253971_at At4g26530 Fructose-bisphosphate aldolase, putative + 2.93

247323_at At5g64170 Dentin sialophosphoprotein-related + 2.92

267010_at At2g39250 AP2 domain-containing transcription factor, putative + 2.86

258321_at At3g22840 Chlorophyll A-B binding family protein / early light-induced protein (ELIP) + 2.84

266097_at At2g37970 SOUL heme-binding family protein + 2.83

256548_at At3g14770 Nodulin MtN3 family protein + 2.75

262526_at At1g17050 Geranyl diphosphate synthase, putative / GPPS, putative / dimethylallyltransferase, + 2.75 putative / prenyl transferase, putative 245015_at Atcg00490 Large subunit of riblose-1,5-bisphosphate carboxylase/oxygenase + 2.75

255734_at At1g25550 Myb family transcription factor + 2.73

267516_at At2g30520 Signal transducer of phototropic response (RPT2) + 2.63

247694_at At5g59750 Riboflavin biosynthesis protein, putative + 2.59

254447_at At4g20860 FAD-binding domain-containing protein + 2.57

258792_at At3g04640 Glycine-rich protein + 2.52

253048_at At4g37560 Formamidase, putative / formamide amidohydrolase, putative + 2.45

252520_at At3g46370 Leucine-rich repeat protein kinase, putative + 2.43

264096_at At1g78995 Expressed protein + 2.38

245724_at At1g73390 Expressed protein + 2.35

260309_at At1g70580 Glutamate:glyoxylate aminotransferase 2 (GGT2) + 2.33

253174_at At4g35090 Catalase 2 + 2.33

244999_at Atcg00190 Chloroplast DNA-dependent RNA polymerase beta subunit + 2.33

255694_at At4g00050 Basic helix-loop-helix (bHLH) family protein + 2.32

246411_at At1g57770 Amine oxidase family + 2.31

244998_at Atcg00180 RNA polymerase beta subunit + 2.30

254247_at At4g23260 Protein kinase family protein + 2.28 61

247585_at At5g60680 Expressed protein + 2.28

260308_at At1g70610 ABC transporter (TAP1) + 2.27

254158_at At4g24380 Expressed protein + 2.26

258217_at At3g17990 Phosphoethanolamine N-methyltransferase 1 / PEAMT 1 (NMT1) + 2.25

262973_at At1g75600 Histone H3.2, putative + 2.22

256336_at At1g72030 GCN5-related N-acetyltransferase (GNAT) family protein + 2.22

258452_at At3g22370 Alternative oxidase 1a, mitochondrial (AOX1A) + 2.21

264738_at At1g62250 Expressed protein + 2.20

258807_at At3g04030 Myb family transcription factor + 2.20

260155_at At1g52870 Peroxisomal membrane protein-related + 2.17

253043_at At4g37540 LOB domain protein 39 / lateral organ boundaries domain protein 39 (LBD39) + 2.17

250429_at At5g10470 Kinesin motor protein-related + 2.16

254691_at At4g17840 Expressed protein + 2.13

246540_at At5g15600 Expressed protein + 2.13

266572_at At2g23840 HNH endonuclease domain-containing protein + 2.11

259418_at At1g02390 Phospholipid/glycerol acyltransferase family protein + 2.10

258025_at At3g19480 D-3-phosphoglycerate dehydrogenase, putative / 3-PGDH, putative + 2.09

255764_at At1g16720 Expressed protein + 2.05

246313_at At1g31920 Pentatricopeptide (PPR) repeat-containing protein + 2.04

250399_at At5g10750 Expressed protein + 2.03

256541_at At1g42540 Glutamate receptor family protein (GLR3.3) + 2.02

251237_at At3g62420 CP12 domain-containing protein + 2.02

262878_at At1g64770 Expressed protein + 2.01

253425_at At4g32190 Centromeric protein-related + 2.00

250002_at At5g18690 Hydroxyproline-rich glycoprotein family protein - 2.00

262978_at At1g75780 Tubulin beta-1 chain (TUB1) - 2.10

255856_at At1g66940 Protein kinase-related - 2.10

259794_at At1g64330 Myosin heavy chain-related - 2.10

248309_at At5g52540 Expressed protein - 2.10 62

248276_at At5g53550 Transporter, putative - 2.20

260264_at At1g68500 Expressed protein - 2.40

257858_at At3g12920 Expressed protein - 2.40

258369_at At3g14310 Pectinesterase family protein - 2.40

247866_at At5g57550 Xyloglucan:xyloglucosyl transferase / xyloglucan endotransglycosylase / endo- - 2.50 xyloglucan transferase (XTR3) 264987_at At1g27030 Expressed protein - 2.60

263498_at At2g42610 Expressed protein - 2.60

263207_at At1g10550 Xyloglucan:xyloglucosyl transferase, putative / xyloglucan endotransglycosylase, - 2.80 putative / endo-xyloglucan transferase, putative 260211_at At1g74440 Expressed protein - 2.80

255148_at At4g08470 Mitogen-activated protein kinase, putative - 2.90

245074_at At2g23200 Protein kinase family protein - 3.0

258332_at At3g16180 Proton-dependent oligopeptide transport (POT) family protein - 3.1

254915_s_at At4g11310 Cysteine proteinase, putative - 3.2

260799_at At1g78270 U-P-glucose glucosyltransferase, putative - 3.3

259789_at At1g29395 Stress-responsive protein, putative - 3.3

247794_at At5g58670 Phosphoinositide-specific phospholipase C (PLC1) - 3.4

257636_at At3g26200 Cytochrome P450 71B22, putative (CYP71B22) - 3.5

252537_at At3g45710 Proton-dependent oligopeptide transport (POT) family protein - 3.8

249867_at At5g23020 2-isopropylmalate synthase 2 (IMS2) - 3.8

258100_at At3g23550 MATE efflux family protein - 3.9

259560_at At1g21270 Wall-associated kinase 2 (WAK2) - 4.3

250640_at At5g07150 Leucine-rich repeat family protein - 4.9

257119_at At3g20190 Leucine-rich repeat transmembrane protein kinase, putative - 5.3

264214_s_at At1g65330 MADS-box protein (AGL38) - 7.4

245275_at At4g15210 Beta-amylase (BMY1) / 1,4-alpha-D-glucan maltohydrolase - 10.6

264346_at At1g12010 1-aminocyclopropane-1-carboxylate oxidase, putative / ACC oxidase, putative - 11.4

262274_at At1g68720 Cytidine/deoxycytidylate deaminase family protein - 47.5

+ up-regulated, - down-regulated

63

Breaking force was different in atagp19 mutant and WT

At each time point, 12 wild type plants and 12 atagp19 mutant plants were tested

for their inflorescence stem breaking force. The breaking force of both the wild type and

the atagp19 mutant increased over time. However, at each time point, the breaking force

of the wild type is significantly higher than that of the atagp19 mutant (Figure 2.6).

Figure 2.6 Breaking force measurements of stems from wild type and atagp19 mutant plants. Error bars show the standard deviation of 12 samples at each time-point.

Real-time PCR analyses and RT-PCR

To test the reliability of the microarray data, mRNA levels corresponding to several genes were tested by quantitative PCR. These genes were chosen either because 64

they had the most fold changes or they are related to cell wall modification. Comparison

of the results for the two methods revealed that the fold changes of some genes

determined by quantitative PCR were not as large as those determined by microarray

analysis; nonetheless, quantitative PCR confirmed the up-regulated/down-regulated

changes in gene expression observed by microarray analysis (Table 2.4). Furthermore,

RT-PCR (Reverse-Transcription PCR) of some of these candidate genes also confirmed

these changes (Figure 2.7).

a b

Figure 2.7 RT-PCR of candidate genes in WT and atagp19 mutant. a) Up-regulated genes. b) Down-regulated genes. Actin is the internal loading control. For the identification/annotations of each gene, see Table 2.3.

65

Table 2.4 Comparison of the changes in gene expression (i.e., mRNA levels) detected by microarray analysis and by quantitative PCR (QPCR), and signal strength for each gene Fold Change Gene Title AGI Locus Signal Detection Microarray QPCR Expressed protein At2g15020 +40.0 +24.3 188.7 P Hypothetical protein At1g72490 +17.0 +3.2 24.5 A Up- 7.7 A regulated Hypothetical protein At3g28240 +16.0 +3.1 Hypothetical protein predicted At2g03540 +13.2 +3.4 56.0 A LEA3 family protein At1g02820 +11.9 +11.3 1283.7 P Myb family transcription factor At1g25550 +2.7 +2.7 2910.0 P Glycine-rich protein At3g04640 +2.5 +2.7 1284.0 P Basic helix-loop-helix family 325.5 P protein At4g00050 +2.3 +2.8 Protein kinase family protein At4g23260 +2.3 +2.7 496.2 P UDP-glucose glucosyltransferase At1g78270 -3.3 -1.4 172.5 P Down- Pectinesterase family protein At3g14310 -2.4 -3.3 3065.1 P regulated Endo-xyloglucan transferase At5g57550 -2.5 -3.1 168.6 P Wall-associated kinase 2 (WAK2) At1g21270 -4.3 -2.5 263.3 P Hydroxyproline-rich glycoprotein At5g18690 -2.0 -1.6 69.3 P family protein Tubulin beta-1 chain (TUB1) At1g75780 -2.1 -1.9 250.7 P Endo-xyloglucan transferase, At1g10550 -2.8 -3.3 71.2 M putative Phosphoinositide-specific At5g58670 -3.4 -4.9 301.6 P phospholipase C (PLC1) + up-regulated, - down-regulated, P present call, A absent call, M marginal call Signal for each gene is from a single microarray chip

AtAGP19 promoter sequence complements atagp19 phenotypes

After hygromycin screening, PCR was performed on the potential transformants to confirm the presence of the complementation construct (Figure 2.8a and Figure 2.9a).

Since the complementation construct only contained the promoter sequence of AtAGP19, 66

not its coding sequence, AtAGP19 mRNA expression was not detected in the

complemented plants by RT-PCR (Figure 2.8b).

a

b

Figure 2.8 Genetic analyses of AtAGP19 in complemented atagp19 mutant. a) PCR results of the complemented plants. Lane 1, 2, WT; lane 3, atagp19 mutant; lane 4, PAtAGP19:AtAGP19 complemented atagp19 mutant; lane5-12, PAtAGP19 complemented atagp19 mutant plants. b) RT-PCR results of the complemented plants. Lane 1, 2, WT; lane 3, atagp19 mutant; lane 4-16, PAtAGP19 complemented atagp19 mutant plants.

The AtAGP19 promoter sequence contained part of the At1g68720 gene (1.5 kb) adjacent to AtAGP19. RT-PCR result of the transformants showed the expression of this gene (At1g68720), which is absent in the atagp19 mutant (Figure 2.9b). In the 30

independent transgenic lines observed in this study, about 20% of the complemented

plants restored the wild type phenotypes, including rosette leaf size, color, plant height

and seed production (Figure 2.10 and Table 2.5). 67

a

b

Figure 2.9 Genetic analyses of At1g68720 in complemented atagp19 mutant. a) PCR results of the complemented plants. Lane 1, 2, WT; lane 3, atagp19 mutant; lane 4, PAtAGP19:AtAGP19 complemented atagp19 mutant; lane5-12, PAtAGP19 complemented atagp19 mutant. b) RT-PCR results of the complemented plants. Lane 1, WT; lane 2, atagp19 mutant; lane 3, PAtAGP19:AtAGP19 complemented atagp19 mutant; lane4-6, PAtAGP19 complemented atagp19 mutant.

Figure 2.10 Complementation of atagp19 mutant with the 19 promoter region. 3-week-old WT (left), atagp19 mutant (middle) and complemented (right) plants. Bar represents 1cm.

Table 2.5 Analysis of WT, atagp19 mutant and complemented plants Parameters WT atagp19 complemented plants Height (cm) 40.3±3.6 20.1±5.1 38.5±2.6 Rosette radius (cm) 7.7±1.4 4.3±0.7 7.5±1.4 Silique length (mm) 12.5±1.0 10.7±1.3 13.5±1.3 Values are expressed as mean±SD Data are from 10 wild type, 10 atagp19 mutant and 7 complemented plants

68

Gene At1g68720 is disrupted in the atagp19 mutant

Since gene At1g68720 is not expressed in the atagp19 mutant, 8 pairs of primers

(F1 and R1, F2 and R2 and so on until F8 and R8) were designed to sequence this gene

(Figure 2.11). In the atagp19 mutant, two pairs of primers (F1 and R1, F2 and R2) gave the same size PCR products as in the wild type, while the other pairs of primers (F3 and

R3 and so on until F8 and R8) gave no PCR products, indicating this region is disrupted.

Furthermore, F2 and R8, F8 and RBa1 (on the right border of T-DNA) were also used as primers in PCR experiments and no product was produced in the atagp19 mutant.

1kb

Figure 2.11 A schematic model showing the gene location of AtAGP19 and At1g68720, and 8 pairs of sequencing primers. Arrow box, gene and its direction; arrowhead, T-DNA insertion; short black arrow, primers.

69

Discussion

Real-time PCR results are reliable

There is a good correlation between signal strength of a specific gene on the

microarray chip and the reliability of the fold change for this gene, i.e. when the signal

for the gene is high, the fold change from the real-time PCR data agrees well with the

microarray data. As shown in Table 2.4, At1g72490, At3g28240 and At2g03540, whose

real-time PCR data are not consistent with the microarray results, also have low signals in

the microarray raw data such that their detection is shown as “A” (Absent). In contrast,

the other 14 genes which have real-time PCR data more consistent with the microarray

results have higher signals in the microarray raw data (Table 2.4 and Supplemental Table

1). Based on these results, genes with high signals usually have reliable microarray data.

Among the 111 candidate genes (72 up-regulated and 39 down-regulated), 68 genes

(61%) are within the lower fold change range of +3.0 to +2.0 and –3.0 to –2.0 based on

microarray data. One may think the microarray data is speculative since more than half of

the candidate genes are in the lower fold change range. In order to prove the reliability of

the microarray result, the signal strength of the 68 genes was examined and only 2 genes show “Absent” (Supplemental Table 1). Therefore, based on the correlation between signal strength and the reliability of the microarray data we discussed above, the fold 70

changes of these 68 candidate genes are robust and meaningful though they are in the

lower fold change range.

Expression of AtAGP17 and AtAGP18 did not change in the mutant

The other two members of the lysine-rich classical AGP family, namely AtAGP17

and AtAGP18, are structurally similar to AtAGP19 (i.e., all three contain a small lysine-

rich subdomain interrupting the classical AGP domain). Furthermore, AtAGP18 has a

similar organ-specific expression pattern to AtAGP19 (Yang et al., 2007). This information might lead one to believe that AtAGP17 and AtAGP18 would compensate for the loss of AtAGP19 in the mutant. However, the expression levels of AtAGP17 and

AtAGP18 did not change in the mutant. Consequently, it is likely that these two lysine- rich AGP genes are independently regulated and function independently with respect to

AtAGP19 as their expression does not appear to compensate for the loss of AtAGP19 in the atagp19 mutant. Moreover, the glycosylation pattern of AtAGP19 is predicted to be different from AtAGP17 and AtAGP18 in terms of its greater number of arabinosides modifying the protein (Sun et al., 2005), which may change the molecular surface of

AtAGP19, and as a result, may affect its function by changing its molecular interactions.

71

Gene At1g68720 and atagp19 mutant phenotypes

Previous study on atagp19 complementation took the AtAGP19 gene and its promoter

sequence to complement atagp19 mutant. The promoter region begins from the 1836

nucleotide upstream of the AtAGP19 start codon so it also contains part of the gene

At1g68720 (1.5 kb), which is next to AtAGP19 (At1g68725) (Figure 2.12). At1g68720

(4565 bp in length) is a cytidine/deoxycytidylate deaminase family protein which has 4

exons and 3 introns. The protein it encodes (1307 aa) is predicted to have hydrolyse

activity and may have a function in riboflavin biosynthesis.

1kb

FP RP

Figure 2.12 A schematic model showing the gene location of AtAGP19 and At1g68720, complementation sequences, and primers used for RT-PCR analysis of At1g68720. Arrow box, gene and its direction; arrowhead, T-DNA insertion (90 bp downstream of the AtAGP19 start codon); PAtAGP19:AtAGP19, sequence (2669 bp, AtAGP19 promoter plus AtAGP19 gene) to complement atagp19 mutant; PAtAGP19, control sequence (1836 bp, AtAGP19 promoter) to complement atagp19 mutant; FP, forward primer; RP, reverse primer.

At1g68720 was also absent in the atagp19 mutant. In order to examine which gene (At1g68725 or At1g68720) determines the mutant phenotypes, the AtAGP19 72

promoter sequence was used to complement the mutant. Surprisingly, this promoter

region (contains only part of At1g68720 and without start codon or promoter) was

transcribed into mRNA in the transgenic plants, as RT-PCR had a band with the

predicted size amplified from mRNA. Moreover, AtAGP19 mRNA was not expressed in

the transgenic plants, suggesting complementation of the mutant phenotypes was due to

part of gene At1g68720, not AtAGP19, and this small part of gene At1g68720 is critical

for the gene function. At1g68720 is also studied by Liming Xiong’s research group

(Donald Danforth Plant Science Center, St. Louis, MO). They had several mutant lines in

At1g68720 and they all had similar phenotypes as atagp19, supporting the idea that

At1g68720 causes the phenotypes.

In the PAtAGP19:AtAGP19 complementation experiment, more than 50 transgenic

lines were observed and about 20% of these plants showed wild type phenotypes (Yang,

personal communication). This is about the same percentage as the PAtAGP19 complementation experiment. In this study, 30 lines of the PAtAGP19 complementation

plants were observed and 7 plants (about 20%) restored all the wild type phenotypes. In

these 7 plants, RT-PCR of At1g68720 mRNA using primers shown in Figure 2.12

produced a PCR product with the predicted size (Figure 2.9 b and data not shown). This

PCR product is from mRNA, not genomic DNA, because the primers used were designed

to flank introns in this gene to distinguish mRNA from genomic DNA. Similarly, RT-

PCR of PAtAGP19:AtAGP19 complementation plants also produced a band of the same size,

suggesting they had a mRNA transcript from the At1g68720 gene. Based on these results, the mRNA transcript from the At1g68720 gene likely results in the complementation of 73

the phenotypes in both the PAtAGP19:AtAGP19 complementation plants and the PAtAGP19 complementation plants.

At1g68720 was not expressed in the atagp19 mutant, indicating the DNA sequence of this gene may be disrupted. In order to sequence this gene and see whether this gene was disrupted, 8 pairs of primers and a primer from the T-DNA right border were used in different combinations as described in the materials and methods. Only two pairs of primers (F1 and R1, F2 and R2) produced PCR products in the atagp19 mutant.

Sequencing results of these two PCR products showed this region had the same DNA sequence as the wild type, suggesting this region was intact but the rest of the gene was disrupted in the mutant.

The T-DNA insertion site was 569 bp downstream of the stop codon of

At1g68720. Several studies have reported that T-DNA insertions often induce genomic

DNA rearrangements, especially near the insertion site. The Arabidopsis T-DNA mutagenized line ACL4 had two T-DNA insertions. A reciprocal translocation between chromosome 2 and 3 near the T-DNA insertion site was observed (Nacry et al., 1998). In the Arabidopsis T-DNA mutant mgoun2, evidence for a DNA inversion between two T-

DNAs was reported (Laufs et al., 1999). Moreover, in the Arabidopsis T-DNA mutant hosoba toge toge (hot), there was a 75.8 kbp deletion at the insertion site (Kaya et al.,

2000). In a similar way, the T-DNA insertion in the atagp19 mutant has likely resulted in the DNA rearrangement of the At1g68720 gene.

It is still a mystery how this fragment of the At1g68720 gene without a start codon and promoter can be expressed and how this gene fragment serves to complement the 74

mutant phenotypes. One explanation, given random insertion into the plant genome, is

that the fragment can be inserted by chance after a promoter region and expressed under the control of this promoter. Another explanation is the possibility of one or more microRNAs encoded in this region. MicroRNAs are short (about 22 nucleotide) single- stranded forms of RNA generated from endogenous hairpin-shaped RNA transcribed from the genomic DNA. They bind to the complementary sequence on the target mRNA to inhibit protein translation. The online tool “miRBase”

(http://microrna.sanger.ac.uk/sequences/search.shtml) provides a way to search for miRNA candidates from DNA sequences. The PAtAGP19 sequence results in 13 miRNA

candidates. It is possible that miRNA generated from this region down-regulates the

expression of some target protein(s) which is important in plant development. Disruption

of this region in the mutant would thus interfere with the regulation of target protein

expression, resulting in mutant phenotypes.

Cell wall-related genes were down-regulated in the mutant

The hypocotyl cell length in atagp19 mutant seedlings is shorter than in the wild

type. At later developmental stages, the mutant also shows smaller rosette epidermal cells

and more regularly shaped spongy mesophyll cells. All these phenotypes indicate some

defects in cell expansion in the mutant (Yang et al., 2007).

Cell expansion is a complex process that requires both loosening of the cell wall and

new wall material deposition. Xyloglucan is the major hemicellulosic polysaccharide in 75

the primary cell wall that forms hydrogen bonds with cellulose microfibrils (Valent and

Albersheim, 1974). During cell expansion, xyloglucan endotransglycosylases/hydrolases

(XTHs) cleave xyloglucan chains and form covalent polysaccharide–enzyme intermediates (Steele and Fry, 1999). Cleaved xyloglucan molecules are then transferred to free nonreducing ends of other xyloglucan chains to form new bonds (Vissenberg et al., 2000). In this way, the cell wall is loosened, and the cell expands by turgor pressure.

Two XTHs genes, At5g57550 and At1g10550, were down-regulated in the mutant, indicating these genes may be related to the compromised cell expansion observed in the atagp19 mutant.

Pectin is another major component of plant cell walls. Pectin is postulated to be synthesized in the Golgi apparatus in a highly methyl esterified form and then exported to the cell wall. In the cell wall, pectin is partially de-esterified by pectinesterase (PE) in a pH dependent manner (Pressey, 1984), exposing the carboxyl groups that participate in the formation of Ca2+ cross-bridges between and among adjacent pectin chains to stiffen the cell wall. The atagp19 mutant stem is considerably weaker than wild type. This was observed when mutant stems were sectioned previously for microscopic analysis. The mutant stem is more likely to break than the wild type stem when a given force is applied

(Yang, personal communication). To substantiate this observation, a breaking force measurement experiment was carried out to measure the force needed to break apart inflorescence stems of the wild type and mutant plants (Fig. 2.6). At different development stages, the mutant stems consistently require significantly less breaking force than the wild type stems. This mechanical strength defect may be the result of a 76

down-regulated gene (At3g14310) encoding a pectinesterase family protein.

Pectinesterase also functions in promoting cell elongation in different organs, such as

Arabidopsis pollen tubes (Jiang et al., 2005), pea roots (Wen et al., 1999) and potato stems (Pilling et al., 2000). Down-regulation of this pectinesterase gene may thus contribute to the cell expansion defect observed in the mutant.

Microtubules are also associated with cell expansion. Cortical microtubules orient

newly synthesized cellulose microfibrils and guide cell expansion (Baskin, 2001).

Tubulin beta-1 chain (TUB1), a microtubule subunit, was down-regulated 2.1 fold in our

study. Interestingly, the TUB1 gene in a T-DNA knockout mutant (dim mutant) which

has an impaired cell elongation phenotype, was also down-regulated (Takahashi, 1995),

providing additional evidence for a possible relationship between TUB1 and cell

expansion.

Hydroxyproline-rich glycoproteins (HRGPs), including AGPs, extensins and

proline-rich proteins, are the major protein components found in plant cell walls. In this

regard, it was interesting to note that one HRGP (At5g18690, AtAGP25) was down- regulated 2 fold. AtAGP25 is a classical AGP predicted to be located on the plasma membrane. Although the specific function of AtAGP25 is unknown, considering its localization to the plasma membrane, it may sense the extracellular environment and direct a signal cascade in a way similar to that proposed for AtAGP19. It was also noted that no other AGP genes had their expression levels change by a factor of two or more, although expression of some AGP genes changed slightly over one fold. Expression of 77

the two homologous genes to AtAGP19, namely AtAGP17 and AtAGP18, did not change

significantly in the mutant, as discussed above, indicating their functions are independent.

Although AGP gene expression changed little in the atagp19 mutant, a phosphoinositide-specific phospholipase C (PLC1, At5g58670) related to cell signaling was down-regulated 3.4 fold. Phospholipase C cleaves the GPI anchor in GPI anchored proteins including AGPs to release proteins from the plasma membrane (Udenfriend and

Kodukula, 1995). At the same time, cleavage of the GPI anchor also produces intracellular messengers (i.e., phosphatidyl-inositol, inositol phosphoglycan and ceramides) (Schultz et al., 1998) from the GPI anchor. The normal expression level of the phospholipase may be essential to release various signal messengers to regulate plant growth and development at different stages.

Overall, genes relating to cell wall biosynthesis or modification identified in this microarray assay using the 2-fold criterion, were generally found to be down-regulated. It

is hypothesized that the reduced expression levels of these genes have a direct effect on producing the mutant phenotypes, mainly with regard to compromised cell expansion.

Genes encoding wall-associated kinases (WAKs), receptor-like kinases (RLKs) and other protein kinases display changes in expression

Another cell wall-related gene not mentioned above, wall-associated kinase2

(WAK2, At1g21270), was down-regulated 4.3 fold. There are five WAKs in Arabidopsis,

and all contain a cytoplasmic kinase domain, a transmembrane region, and a domain 78

extending into the cell wall. The extracellular domain contains regions homologous to the

vertebrate epidermal growth factor (EGF)-like domain (He et al., 1996). WAKs are

tightly associated with cell wall components, in particular, pectin and glycine-rich proteins (GRP) (Anderson et al., 2001). WAKs are involved in various biological functions, including pathogen resistance (He et al., 1998), heavy-metal tolerance

(Sivaguru et al., 2003), and plant development (Lally et al., 2001; Wagner and Kohorn,

2001). The expression levels of WAK1 and WAK2 are highest among the five members, and they are expressed in vegetative meristems, organ junctions, and areas of cell

expansion (Anderson et al., 2001). A WAK2 antisense gene was introduced into wild

type Arabidopsis and the resulting transgenic plants exhibit a loss of cell expansion in

leaves (Wagner and Kohorn, 2001). The expression and antisense RNA data indicate

WAK2 functions in cell expansion. The mechanism of how WAK2 regulates cell expansion is unclear. Based on the structure and subcellular localization of WAKs, they may serve as signal transduction molecules between the cell wall and the cytoplasm

(Kohorn, 2001).

WAKs belong to the receptor-like kinases (RLKs) superfamily in Arabidopsis with

at least 610 members (Shiu and Bleecker, 2001). RLKs span the plasma membrane with one domain in the extracellular matrix and a kinase domain in the cytosol (Cock et al.,

2002). As physical links between the extracellular matrix and the cytoplasm, these RLKs

may sense extracellular signals and trigger downstream cellular responses. A RLK gene

encoding a putative leucine-rich repeat transmembrane protein kinase (At3g20190) was

down-regulated 5.3 fold. In addition, genes encoding several other protein kinases 79

demonstrated changes in expression. Three protein kinases (At1g63580, At3g46370,

At4g23260) were up-regulated, while another three protein kinases (At1g66940,

At4g08470, At2g23200) were down-regulated. Changes in the expression of these genes in the mutant point to protein kinases being involved with a potential signal transduction pathway linking At1g68720 to cell expansion.

Expression of Genes Related to Transcriptional Regulation Showed Changes

In order to provide insight to the early events in the mutant leading to its characteristic phenotypes, genes encoding transcription factors were of particular interest.

Two myb family transcription factors were up-regulated. Different myb proteins have

distinct functions in plants. Some function in secondary metabolism, some control

cellular morphogenesis, and some are involved in signal pathways responding to plant

growth regulators (Martin and Paz-ares, 1997). One Dof-type zinc finger domain-

containing protein was up-regulated 4.6 fold. The Dof motif is a novel motif identified in

plants. Several members of this protein family are involved in the transcriptional control

of developmental gene expression (Takatsuji, 1998). One basic helix-loop-helix (bHLH)

family protein was up-regulated 2.3 fold. Among the 147 bHLH-encoding genes in

Arabidopsis, 14 have established biological functions, including phytochrome signaling, gynoecium development, flavonoid biosynthesis, trichrome differentiation, microspore

development, abscisic acid–induced gene expression, tryptophan biosynthesis, and

brassinosteroid signaling. These various functions indicate that the bHLH family 80 probably regulates a broad range of growth and developmental processes at different developmental stages (Toledo-Ortiz et al., 2003).

The microarray data revealed that some transcription factors related to plant growth and development changed significantly, potentially leading to abnormal phenotypes in the mutant. However, how and which of these transcription factors regulate the expression of genes related to cell expansion remains unclear.

Originally, we thought the atagp19 mutant phenotypes were due to the loss of

AtAGP19, since complementation of AtAGP19 restored all the phenotypes. Later, we found partial sequence of At1g68720 was enough to complement the mutant phenotypes.

This is an interesting phenomenon since one gene usually needs its complete sequence to be expressed and functional. However, the microarray data is still valid in discussing candidate genes and mutant phenotypes, and their possible relation to At1g68720. 81

CHAPTER 3 OVEREXPRESSION OF ATAGP17, 18 AND 19

Summary

AtAGP17, 18 and 19 comprise the lysine-rich AGP family in Arabidopsis. GFP-

AtAGP17/18/19 fusion proteins as well as a GFP control were overexpressed in

Arabidopsis and the fusion proteins were found to be present on the plant cell surface of different organs. Subcellular localization of the fusion proteins to the plasma membrane was further determined by plasmolysis of leaf trichome cells. Moreover, in vitro pollen germination showed that AtAGP17, unlike LeAGP-1, was not involved in pollen tube elongation.

To elucidate AtAGP17/18/19 function(s), AtAGP17/18/19 were expressed without the GFP tag under the control of 35S promoter. At the same time, a vector control construct was also introduced into Arabidopsis plants. The transformants were

examined by PCR to ensure the proper constructs were incorporated and RT-PCR was

used to examine mRNA levels of AtAGP17/18/19. In contrast to AtAGP17 and AtAGP19

overexpressors which showed no different phenotypes from the wild type plants,

AtAGP18 overexpressors with high mRNA levels displayed several phenotypes distinct

from the wild type plants: they were shorter, produced more branches, had shorter roots

and produced less viable seeds. In contrast, the vector control transformants had the same

phenotypes as the wild type plants. The AtAGP18 overexpressors had similar phenotypes

exhibited by LeAGP-1 overexpressors (shorter stems, more branches, shorter roots and

less viable seeds), suggesting these two AGP genes may have similar function(s). 82

Furthermore, AtAGP18 transcripts was down-regulated by the plant hormone ABA, indicating ABA was likely involved in the signaling pathway(s).

Introduction

Overexpression of a target gene is widely used in gene function studies.

Previously, overexpression of LeAGP-1, a lysine-rich AGP in tomato, caused an overbranching phenotype. The transgenic tomatoes were short, had multiple branches and produced less seeds than the control plants. As cytokinin-overexpressing plants also have a similar bushy phenotype, and cytokinins and other plant growth hormones (auxins and

ABA) can regulate LeAGP-1 mRNA expression, LeAGP-1 probably functions in plant growth and development involving cytokinin/auxin signaling (Sun et al., 2004b). To examine the subcellular localization of LeAGP-1, GFP-LeAGP-1 fusion protein was expressed in tobacco (Nicotiana tabacum BY-2) cells and the fusion protein was found to be localized to the plasma membrane and in Hechtian threads.

As my laboratory turned its attention to the model plant Arabidopsis to extend research on lysine-rich AGPs, I was interested in using this overexpression method to explore the functions of the three homologous genes, namely AtAGP17, AtAGP18 and

AtAGP19. Here, a series of GFP fusion protein constructs 35S-ss-GFP-AtAGP17/18/19

(35S, CaMV35S promoter; ss, LeAGP-1 signal sequence; GFP, green fluorescence protein) as well as a GFP only control (35S-ss-GFP) were delivered into Arabidopsis plants to study the subcellular localization of these AGPs. Another series of constructs 83

(35S-AtAGP17/18/19) as well as the empty vector control were constructed and

overexpressed in Arabidopsis plants to observe phenotypes.

Materials and methods

Overexpression constructs design

A schematic representation of all the overexpression constructs and their related

information are shown in Figure 3.1.

35S-ss-GFP-AtAGP17/18/19

The 35S-ss-GFP-AtAGP17 chimeric gene construct was produced as described previously (Sun et al., 2005). The construct was built in binary vector pBI121 and contained CaMV35S promoter (35S), LeAGP-1 signal sequence (ss), green fluorescence protein (GFP) and AtAGP17 gene without the signal sequence. The GFP used in this construct was an enhanced version (EGFP).

The 35S-ss-GFP-AtAGP18/19 chimeric gene constructs were produced as described previously (Yang, 2006). The constructs were built in binary vector pART27 and contained the same promoter, signal sequence, GFP as 35S-ss-GFP-AtAGP17 and

AtAGP18/19 gene without the signal sequence. 84

The above constructs were obtained in Agrobacterium and used to transform

Arabidopsis wild type plants to overexpress the chimeric genes in this study. 85

Name of transgenic Main purpose Construct Construct plants harboring the of made by construct transformation 35S-ss-GFP Subcellular GFP control Yizhu Zhang localization

35S-ss-GFP-AtAGP17 Subcellular Dr. Wenxian 17 over localization Sun

35S-ss-GFP-AtAGP18 Subcellular 18 over Dr. Jie Yang localization

35S-ss-GFP-AtAGP19 Subcellular 19 over Dr. Jie Yang localization

Vector control Control for VC overexpression Yizhu Zhang phenotypes 35S-AtAGP17 Observe 17NG overexpression Yizhu Zhang phenotypes 35S-AtAGP18 Observe 18NG overexpression Yizhu Zhang phenotypes 35S-AtAGP19 Observe 19NG overexpression Yizhu Zhang phenotypes Figure 3.1 A summary of the structure of the recombinant constructs used in this research. The schematic illustrations of the constructs are not drawn in scale. 17 over, transgenic plant overexpressing GFP-AtAGP17 fusion protein; 18 over, transgenic plant overexpressing GFP-AtAGP18 fusion protein; 19 over, transgenic plant overexpressing GFP-AtAGP19 fusion protein; VC, vector control; 17NG, transgenic plant overexpressing AtAGP17 (no GFP); 18NG, transgenic plant overexpressing AtAGP18 (no GFP); 19NG, transgenic plant overexpressing AtAGP19 (no GFP).

86

35S-ss-GFP

The pUC-ss-GFP construct was obtained from Dr. Li Tan in the Department of

Chemistry and Biochemistry at Ohio University and sequenced in both directions at Ohio

University Genomics Facility using M13 forward primer and M13 reverse primer. The construct contained LeAGP-1 signal sequence (ss) and GFP. The pUC-ss-GFP construct was digested with restriction enzyme BamHI and SacI. The smaller fragment containing ss-GFP was purified and subcloned into binary vector pBI121 between the BamHI site and the SacI site to replace the GUS gene (Figure 3.2a). The recombinant pBI121 vector

(35S-ss-GFP construct) was introduced into Agrobacterium strain LBA4404 by electroporation and then transformed Arabidopsis wild type plants (Agrobacterium and

Arabidopsis transformation methods are described in Chapter 2, section “Transformation of Agrobacterium using electroporation” and “Floral dip transformation of Arabidopsis”) as a control since it only overexpress GFP, not the chimeric AGP genes. 87

a

b

c nos terminator Pst I RB

hygromycin resistance

Vector control LB

kanamycin resistance

Figure 3.2 Figure illustration of recombinant constructs. (continued on p.88) 88 a) 35S-ss-GFP construct originates from binary vector pBI121. ss-GFP was inserted into binary vector pBI121 between the BamHI site and the SacI site. b) 35S-AtAGP17/18/19 construct originates from binary vector pMDC45. AtAGP17/18/19 coding sequence was inserted into binary vector pMDC45 between the KpnI site and the PacI site. c) Empty vector control construct originates from binary vector pMDC45. pMDC45 binary vector was digested with restriction enzyme PstI and the small fragment containing useless DNA sequence (such as ccdB gene, CMr gene, GFP6 gene) between the two PstI site was removed. The remaining fragment was self-ligated to create the empty vector control construct.

35S-AtAGP17/18/19

This set of constructs was created to produce the transgenic plants that overexpress AtAGP17/18/19 without the GFP tag. To create these constructs, the

AtAGP17/18/19 coding sequence was inserted between the KpnI site and the PacI site on binary vector pMDC45 (Figure 3.2b). Primers used for amplifying the AtAGP17/18/19 coding sequence were designed to contain a KpnI site (GGTACC) in the 5’ end of the forward primer and a PacI site (TTAATTAA) in the 5’ end of the reverse primer. The

AtAGP17 coding sequence was amplified by the primers: (AGP17 CDS KpnI F) 5' -

GGG GTA CCC CTC TCA ACT AAT TAC AAA TTA TGA CTC G - 3' and (AGP17

CDS PacI R) 5' - CCT TAA TTA AGG TTA TTA GAA GGC TAG AAC AAG TAG

AG - 3'. The AtAGP18 coding sequence was amplified by the primers: (AGP18 CDS

KpnI F) 5' - GGG GTA CCC CTC CAA ATT TTA ACA AAA TTA TGG ATC - 3' and

(AGP18 CDS PacI R) 5' - CCT TAA TTA AGG TTA GAA TGC CAT AAC GAG AAC

G - 3'. The AtAGP19 coding sequence was amplified by the primers: (AGP19 CDS KpnI 89

F) 5' - GGG GTA CCT AGC TTC CTC CAC AAC ACA ATG - 3' and (AGP19 CDS

PacI R) 5' - CCT TAA TTA AGG TTA GGC TGT CAT AGC AAG TAG - 3'.

Empty vector control

An empty vector construct containing only a hygromycin resistant gene was also transformed into wild type Arabidopsis as a control. To build this construct, pMDC45 binary vector was digested with restriction enzyme PstI and the small fragment containing useless DNA sequence between the two PstI site was removed. The remaining fragment was self-ligated to create the empty vector control construct (Figure 3.2c).

Construction of the overexpression vectors

Genomic DNA was extracted and AtAGP17, 18 and 19 genomic DNA was amplified using either AccuPrime Pfx DNA polymerase (Invitrogen, Carlsbad, CA) or

Platinum Pfx DNA polymerase (Invitrogen, Carlsbad, CA) as described in chapter 2, section “DNA extraction and amplification of AtAGP19 promoter”. The purified PCR products and the vector pMDC45 were digested with KpnI and PacI (New England

BioLabs, Beverly, MA). The following ligation, transformation and restriction digestion diagnosis method were the same as described in chapter 2, section “Restriction digestion” and “Ligation and transformation”. The constructs containing AtAGP17/18/19 genomic sequence were saved for sequencing. 90

Sequencing

Two primers flanking the insert region in the constructs were used in the PCR

reaction to amplify the AtAGP17, 18 and 19 genomic DNA sequence. The PCR products

were purified with QIAquick PCR Purification Kit and sent to Ohio University Genomics

Facility for sequencing. The same primers in the PCR reaction were used in sequencing

reactions, in both directions. The forward primer was 5' - ATC CTT CGC AAG ACC

CTT CCT C - 3' and the reverse primer was 5' - CCG GCA ACA GGA TTC AAT CTT

AAG - 3'. Correct constructs were chosen for Agrobacterium transformation.

Agrobacterium transformation, floral dip transformation and screening of transgenic plants

Methods for Agrobacterium transformation, floral dip transformation and screening of transgenic plants (hygromycin selection) were described in chapter 2, section “Transformation of Agrobacterium using electroporation”, “Floral dip transformation of Arabidopsis” and “Screening of transgenic plants”.

For kanamycin selection, seeds sterilization methods were the same as

hygromycin selection, and then the seeds were sown on MS plates containing 50 μg/ml

kanamycin. The plates were kept in dark at 4°C to stratify the seeds for 2-3 days then

transferred to the tissue culture room at 22°C with 16 h day/8 h night condition. After 10

days, kanamycin-resistant seedlings were green with true leaves, while kanamycin- 91

sensitive seedlings were pale in color with no development of true leaves. The potential

transformants were transferred to soil.

Genetic analyses of transgenic plants

Methods for DNA extraction and PCR, RNA extraction and RT-PCR were

described in chapter 2, section “Verification of transgenic plants by DNA extraction and

PCR” and “RNA extraction from transgenic plants and RT-PCR”.

For transgenic plants harboring either the 35S-AtAGP17 or 35S-AtAGP19

construct, young leaves were used in RNA extraction since expression levels of AtAGP17

and AtAGP19 in young leaves are low and young leaves usually produce good RNA

quality. For transgenic plants harboring 35S-AtAGP18 construct, old leaves were used in

RNA extraction since old leaves have a low endogenous expression level of AtAGP18

mRNA.

Primers for mRNA expression level of each AGP gene were designed to flank the

intron region so the PCR product from RNA and contaminating DNA were of different

sizes (Figure 3.3). The primers for AtAGP17 were 5' - ATA AGC CTA AAC CCA CAT

CTC CCG - 3' and 5' - CAG CTC CCA CCA TTT GTA TCA - 3'. The primers for

AtAGP18 were 5' - GCT CCG GCG AAA ACT CCA ACT G - 3' and 5' - AAC CGC

TCC CAC CGC TAC ATT C - 3'. The primers for AtAGP19 were 5' - CTG CTC TCA

TCT CTT CCT TTA GTG - 3' and 5' - ATT GAG CCA CAT TAC TGC TCT TCC - 3'. 92

The primers for actin as an internal control were 5' - GTG CTC GAC TCT GGA GAT

GGT GTG - 3' and 5' - CGG CGA TTC CAG GGA ACA TTG TGG - 3'.

Figure 3.3 A schematic model showing primers used in RT-PCR of the AtAGP17/18/19. Primers were designed to flank an intron of the gene. RT-PCR using RNA as template produced a smaller product, while RT-PCR using DNA as template produced a larger product.

For the actin control, 10 ng total RNA was used as template. For the

AtAGP17/18/19 genes, 50 ng total RNA was used as template. The PCR conditions were:

50°C for 30 min, 95°C for 15 min, followed by 35 cycles of 94°C, 30 sec; 50°C, 30 sec;

72°C, 1 min and a final extension at 72°C for 10 min.

Total protein extraction

Leaves from wild type and transgenic plants were ground in liquid nitrogen with a mortar and pestle. Then protein extraction buffer (50 mM Tris-HCl pH 7, 10 mM potassium chloride, 1 mM EDTA, 0.1 mM magnesium chloride, 8% sucrose, 1 mM 93

phenylmethanesulphonylfluoride, maintained at 4°C for up to 1 month) was added to the

leaf powder and centrifuged at 10,000 g for 10 min at 4°C. The supernatant was collected and quantified by the Lowry assay using DC Protein Assay (Bio-Rad, Hercules, CA) according to the instruction manual provided by the supplier.

Protein electrophoresis and transfer

Protein samples were mixed with 2X SDS-PAGE gel loading buffer (50 mM Tris-

Cl, pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100mM dithiothreitol) at a

volume ratio of 1:1 and heated in boiling water bath for 3 min. Protein samples (30 µg

each lane) were run on SDS-PAGE gels (5% stacking gel and 10% resolving gel) with

Precision Plus Protein Standards (Bio-Rad, Hercules, CA) using a Mini-PROTEAN 3

Electrophoresis Cell (Bio-Rad, Hercules, CA).

The PVDF membrane (Bio-Rad, Hercules, CA) was first wet with methanol for

30 seconds, then equilibrated in transfer buffer (25 mM Tris, 192 mM glycine, pH 8.3).

Proteins on the gel were then transferred to the PVDF membrane in transfer buffer for 1 h

at 100 V in a Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad, Hercules, CA).

94

Coomassie blue staining

The PVDF membrane was stained in the staining solution (0.1% Coomassie blue

R-250, 40% methanol) for 1 min, then destained in 50% methanol until the background color turned light blue, and the bands became clear.

Western blotting

The PVDF membrane was rinsed in ddH2O three times before incubation in

TBST (0.05% Tween 20, 1XTBS) supplemented with 5% non-fat dry milk for 1 h with gentle shaking. The membrane was then rinsed three times in TBST and incubated with

GFP Living Colors A.v. Peptide Antibody (Clontech, Mountain View, CA) diluted

1:1000 in blocking buffer (3% BSA, 1XTBST) for 1 h with gentle shaking. After the membrane was rinsed three times in TBST, it was incubated with goat anti-rabbit alkaline phosphatase-conjugated secondary antibody (Sigma-Aldrich, St. Louis, MO) diluted

1:500 in blocking buffer for 1 h with gentle shaking. The membrane was again rinsed three times in TBST and another time with TBS. After that, the membrane was incubated in freshly prepared color substrate solution (0.5mM NBT, 0.5mM BCIP, 0.1 M Tris-HCl,

0.1 M NaCl, pH 9.5) without shaking in the dark until color developed. Then the membrane was soaked in ddH2O.

95

Confocal laser scanning microscopy

Different organs from wild type and transgenic plants harboring 35S-ss-GFP-

AGP17/18/19 constructs were observed under a Nikon fluorescence microscope or a

Zeiss confocal laser scanning microscope LSM510 (Zeiss, Germany). The settings for the

LSM510 were as follows: 488 nm argon laser with 75% power, 510 to 550 nm emission filter, pinhole 150-200 µm. Images were processed with the Zeiss LSM image browser.

Sectioning

Stems from wild type and transgenic plants harboring 35S-ss-GFP-AGP17/18/19 constructs were embedded in 6% low melt agarose (Fisher Scientific, Hampton, NH), and sectioned either with a razor blade manually or with the vibratome (Vibratome Company,

St. Louis, MO). The section thickness for the vibratome was set between 80 μm and 100

μm.

Pollen germination

Open Arabidopsis flowers were picked and shaken above the Pollen Germination

Medium (18% sucrose, 0.01% boric acid, 2 mM CaCl2, 1 mM MgSO4, pH7.0, filter

sterilized) in a Flat Bottom Microtest Plate (Becton Dickinson and Company, Oxnard,

CA) to collect pollen grains on the surface of the medium, the plate was then covered 96

with lid to maintain high humidity and placed on the bench overnight for pollen grains to

germinate.

Pollen protoplast preparation

Flowers were dissected and incubated in enzyme solution [0.1% MES, pH5.8,

0.65 M mannitol, 50 mM CaCl2, 1% cellulase (Sigma, Saint Louis, MO), 1% pectinase

(Sigma, Saint Louis, MO)] at 28°C on a shaker at 135 rpm for 2 h. The enzyme solution containing pollen protoplasts released from the pollen grains was mounted on the slide and viewed under a microscope.

Plasmolysis

Different organs (flowers, roots, leaf epidermal peels) from Arabidopsis were incubated in 4% NaCl for 10-15 min. Sometimes the NaCl solution with the plant organs was degassed in a vacuum desiccator for 2 min to facilitate infiltration.

Hormone treated Arabidopsis seedlings, RNA extraction and RT-PCR

Arabidopsis wild type seeds were sterilized and sown on half strength MS plates containing 0.5% sucrose. The plates were kept in dark at 4°C to stratify the seeds for 2-3 days and then transferred to the tissue culture room at 22°C with a 16 h day/8 h night 97 condition. On certain days, the seedlings were transferred to MS liquid media containing

20 µM ABA or no ABA (mock) and allowed to culture on a shaker at 120 rpm for 2 days.

The expression levels of AtAGP18 in the seedlings were examined by total RNA extraction and RT-PCR method as described previously in Chapter 2, section “RNA extraction from transgenic plants and RT-PCR”. The primers were 5' - GCT CCG GCG

AAA ACT CCA ACT G - 3' and 5' - AAC CGC TCC CAC CGC TAC ATT C - 3'.

Hormone treated PAtAGP18:GUS transgenic plants and GUS staining

The transgenic Arabidopsis plants harboring the PAtAGP18:GUS construct was obtained from Dr. Jie Yang (Yang, 2006). The T2 plants were grown and treated with

ABA as described above.

GUS staining was performed with the β-Glucuronidase (GUS) Reporter Gene

Staining Kit (Sigma, Saint Louis, MO) according to the manufacturer’s instructions with minor modifications. The seedlings (including wild type plants as a negative control) were soaked in staining solution (50 mM sodium phosphate, 0.1 mM potassium ferricynide, 0.1 mM potassium ferrocynide, 0.5% Triton X-100 and 700 µg/ml X-GlcA, pH 7.0) in a 1.5 ml or 15 ml vial. The vial was degassed in a vacuum desiccator for 2 min and incubated at 37°C in the dark for 6 to 12 h. When the seedlings were properly stained, they were destained in 95% ethanol to remove the green chlorophyll and stored in ethanol.

Images were taken with the Nikon binocular stereo dissecting microscope SMZ1500.

98

Organ specific expression of AtAGP17

Arabidopsis wild type plants were grown in soil in the growth room under long day condition (16h day/8h night cycle) for 2 weeks to harvest seedlings. The plants were grown for 5 weeks to harvest flowers, stems and rosette leaves. Plants grown on MS vertical plates with 1% sucrose for 2 weeks were used to harvest roots.

The above different organs were used for RNA extraction with the RNeasy Plant

Mini Kit (QIAGEN, Valencia, CA). The RNAs were treated with DNase I, amplification grade (Sigma, St. Louis, MO) according to the manufacturer’s instructions to avoid DNA contamination and used for RT-PCR with the OneStep RT-PCR Kit (QIAGEN, Valencia,

CA). The primers for AtAGP17 expression were 5’- ATA AGC CTA AAC CCA CAT

CTC CCG -3’ and 5’- CAT GAG ACA AAT GGG AGA GGA TCA -3’. The primers for actin as an internal control were 5' - GTG CTC GAC TCT GGA GAT GGT GTG - 3' and

5' - CGG CGA TTC CAG GGA ACA TTG TGG - 3'. The RT-PCR conditions for

AtAGP17 expression were: 50°C for 30 min, 95°C for 15 min, followed by 35 cycles of

94°C, 30 sec; 53°C, 30 sec; 72°C, 1 min and a final extension at 72°C for 10 min. For actin, 28 cycles were used in the RT-PCR.

PAtAGP17:GUS construct and GUS staining

The primers for AtAGP17 promoter were designed to introduce restriction enzyme sites for PacI (TTAATTAA) in the forward primer 5’- CGG CTT AAT TAA GTG TAT 99

ATG GTT GCT ACG TGC -3’ and AscI (GGCGCGCC) in the reverse primer 5’- AAT

TGG CGC GCC AAT TTG TAA TTA GTT GAG AGA -3’. The AtAGP17 promoter sequence (1.2 kb upstream of the start codon) was amplified with this pair of primers.

The PCR products and pMDC164 binary vector (Curtis and Grossniklaus, 2003) were

digested by PacI (TTAATTAA) and AscI (GGCGCGCC) and ligated. The recombinant pMDC164 (PAtAGP17:GUS construct) was first introduced into Agrobacterium by

electroporation and then into Arabidopsis wild type plants by the floral dip method as

described previously in Chapter 2, section “Transformation of Agrobacterium using

electroporation” and “Floral dip transformation of Arabidopsis”.

The transgenic T1 plants were screened by hygromycin selection and identified

by PCR. GUS staining was performed as described above on different organs of the

transgenic plants.

Results

GFP-AtAGP17 was expressed in transgenic plants

The 35S-ss-GFP-AtAGP17 construct was delivered into Arabidopsis wild type

plants through floral dip transformation to overexpress AtAGP17. For convenience,

transgenic plants harboring 35S-ss-GFP-AtAGP17 were named 17 over. Fifteen

independent transgenic lines were identified to contain this construct. To verify AtAGP17

was overexpressed in 17 over, total RNA was extracted from wild type and 17 over 100

rosette leaves at day 35 and RT-PCR was performed using actin as an internal loading

control (Figure 3.4a). In 17 over, AtAGP17 mRNA was from both the endogenous

AtAGP17 gene and the GFP-AtAGP17 transgene. As a result, the RT-PCR product of

AtAGP17 mRNA from 17 over was much more than that from wild type.

To verify presence of the GFP-AtAGP17 fusion protein, western blotting was

carried out using GFP antibody. Protein samples were boiled for 3 minutes prior to

loading to release the GFP from the GFP-AtAGP17 fusion protein since a previous study showed heating can release the GFP from the fusion protein (Sun et al., 2004a). As

Figure 3.4b shows, GFP (about 25kD) was present in 17 over, but was absent in WT.

Green fluorescence was also observed in different lines of 17 over, indicating expression of the GFP-AtAGP17 fusion protein. GFP-AtAGP17 was found to be localized to the cell surface in stems, roots, leaf epidermis, stamen filaments and pollen grains (Figure 3.5).

101

a

b

Figure 3.4 GFP-AtAGP17 is expressed in transgenic plants. a) RT-PCR results of AtAGP17 mRNA expression in wild type and AtAGP17 overexpression plants. b) Coomassie staining of protein samples (left) and western blot using the GFP antibody (right). Protein samples were boiled for 3 min prior to loading to release free GFP. GFP, an indicator of the fusion protein, is present in the overexpression plants.

102

Figure 3.5 (continued on p. 103) AtAGP17 is localized on the plant cell surface. a) and b) Stem sections of WT and 17 over plants, respectively. c) and d) Roots of WT and 17 over plants. e) and f) Leaf peels of WT and 17 over plants. g) and h) Stamen filaments of WT and 17 over plants. 103

i) to k) Fluorescent image (i), bright field image (j) and superimposed image (k) of WT pollen grains. l) to n) Fluorescent image (l), bright field image (m) and superimposed image (n) of 17 over pollen grains. Bars = 50 µm (a to h) and 20 µm (i to n).

AtAGP17 is localized to the plasma membrane

The GFP-AtAGP17 fusion protein was found to be localized on the plant cell

surface in different plant organs. However, its subcellular localization can not be seen

unless the cell is plasmolyzed. After plasmolysis, GFP staining was observed on the

plasma membrane of leaf trichome cells (Figure 3.6).

Figure 3.6 AtAGP17 is localized to the plasma membrane. a) and b) Fluorescent image (a) and bright field image (b) of 17 over trichome. c) and d) Fluorescent image (c) and bright field image (d) of plasmolysed (4% NaCl for 15 min) 17 over trichome. GFP-AtAGP17 is distributed on the plasma membrane. Bars = 20 µm 104

AtAGP17 is not involved in pollen tube elongation

Pollen grains from 17 over plants were germinated in vitro on Pollen Germination

Medium. Confocal images of representative pollen grains at different time points were

taken (Figure 3.7). At time point 0 h, green fluorescence (indicates GFP-AtAGP17) was

distributed homogenously on the surface of the pollen grain. At 6 h, when pollen grains began to germinate, green fluorescence became absent in the area where the pollen tube

began to grow. At 12 h and 20 h, green fluorescence was still absent in the pollen tube

area as they continued to extend. The above results showed GFP-AtAGP17 was not

involved during pollen tube elongation process.

Organ specific expression of AtAGP17

RT-PCR results showed AtAGP17 was expressed in seedlings, flowers, stems and

rosette leaves, and weakly expressed in roots (Figure 3.8a). These results are in

agreement with Yang’s northern blot analyses and are consistent with the microarray data

from Genevestigator (Yang, 2006; https://www.genevestigator.ethz.ch).

105

Figure 3.7 (continued on p. 106) GFP-AtAGP17 is absent in pollen tubes during pollen germination. 106

a) to c) Fluorescent image (a), bright field image (b) and superimposed image (c) of 17 over pollen grain at 0 h. d) to f) Fluorescent image (a), bright field image (b) and superimposed image (c) of 17 over pollen grain at 6 h. g) to i) Fluorescent image (a), bright field image (b) and superimposed image (c) of 17 over pollen grain at 12 h. j) to l) Fluorescent image (a), bright field image (b) and superimposed image (c) of 17 over pollen grain at 20 h. Bars = 20 µm

a

b

Figure 3.8 Organ specific expression of AtAGP17 a) RT-PCR results of AtAGP17 expression. SL, seedlings; FL, flowers; ST, stems; RT, roots; RL, rosette leaves. b) PCR results with hygromycin resistance gene primers of PAtAGP17:GUS construct in transgenic plants. M, 100 bp DNA ladder; lane 1, WT; lane 2 and 3, different transgenic lines harboring PAtAGP17:GUS construct; Hyg, hygromycin resistant gene.

In order to examine AtAGP17 expression at the tissue level, a PAtAGP17:GUS

construct was made and introduced into Arabidopsis wild type plants. The promoter

sequence used in this construct was 1.2 kb upstream of the start codon since a previous

study showed 0.5 kb or 0.9 kb of this promoter region was not enough to drive the

expression of AtAGP17 (Sun and Yang, personal communication). Two T1 transgenic 107

lines were identified by PCR (Figure 3.8b). However, GUS staining was not detected in

T1 transgenic lines in different organs, while GUS staining of positive controls

(PAtAGP18:GUS and PAtAGP19:GUS T2 plants) showed nice staining pattern, indicating this

1.2 kb promoter region was not enough to control AtAGP17 expression.

GFP-AtAGP18/19 were expressed in transgenic plants

To overexpress AtAGP18/19, the 35S-ss-GFP-AtAGP18/19 constructs were introduced into Arabidopsis and the transgenic plants were named 18 over and 19 over, respectively. Two 18 over lines and five 19 over lines were identified (Figure 3.9a and b and data not shown) and green fluorescence was detected in all these lines. Like GFP-

AtAGP17, GFP-AtAGP18/19 were localized to the cell surface in different plant organs

(Figure 3.9c-l and Figure 3.10). 108

(Figure continued on p. 109) a b

109

Figure 3.9 GFP-AtAGP18 is expressed in transgenic plants. a) PCR results with AtAGP18 primers confirming the presence of the construct in transgenic plants. M, 100 bp DNA ladder; lane 1, WT; lane 2 and 3, different transgenic lines harboring 35S-ss-GFP-AtAGP18 construct. b) RT-PCR results of AtAGP18 mRNA expression in wild type and AtAGP18 overexpression plants. c) and d) Roots of WT and 18 over plants. e) and f) Leaf peels of WT and 18 over plants. g) and h) Stamen filaments of WT and 18 over plants. i) and j) Petals of WT and 18 over plants. k) and l) Seeds of WT (left) and 18 over (right) plants. m) and n) Fluorescent image (m) and bright field image (n) of 18 over trichome. o) and p) Fluorescent image (o) and bright field image (p) of plasmolysed (4% NaCl for 15 min) 18 over trichome. GFP-AtAGP18 is distributed on the plasma membrane. Bars = 50 µm (c to j), 1 mm (k) and 20 µm (l to p).

110

Figure 3.10 GFP-AtAGP19 is expressed in transgenic plants and localized on the plant cell surface. a) and b) Leaf peels of WT and 19 over plants. c) and d) Roots of WT and 19 over plants. e) and f) Stamen filaments of WT and 17 over plants. g) and h) Petals of WT and 17 over plants.

AtAGP17 overexpressors showed no phenotype

To overexpress AtAGP17/18/19 without the GFP tag, the 35S-AtAGP17/18/19 constructs as well as the vector control were transformed into Arabidopsis wild type plants. For convenience, the transformants were named 17NG (17 overexpressor, No

GFP tag), 18NG, 19NG and VC (Vector Control), respectively. Ten 17NG lines were identified by PCR (Figure 3.11a and data not shown). Phenotypes of these lines were not different from those of wild type plants with respect to plant height, branch number, and 111

reproduction, though these lines had different mRNA expression levels of AtAGP17

(Figure 3.11b).

a

b

Figure 3.11 Genetic analyses of AtAGP17 overexpression plants. a) PCR results with hygromycin resistance gene primers of AtAGP17 overexpression (17NG) plants. M, 100 bp DNA ladder; lane 2, 5 and 8, WT; lane 1, 3, 4, 6 and 7, different AtAGP17 overexpression lines. b) RT- PCR results showing the expression levels of AtAGP17 mRNA in AtAGP17 overexpressing (17NG) plants. M, 100 bp DNA ladder; lane 1, 3, 4 and 8, different transgenic lines with high expression of AtAGP17; lane 6, 10 and 11, different transgenic lines with moderate expression of AtAGP17; lane 9, transgenic line with low expression of AtAGP17; lane 2, 5, 7 and 12, WT.

AtAGP18 overexpressor had several phenotypes

Thirty-two 18NG lines were identified to contain the 35S-AtAGP18 construct.

Among these lines, 18 lines (slightly more than half) displayed several phenotypes distinct from the wild type plants. In contrast, 7 VC (Vector Control) lines containing the correct construct had the same phenotypes as the wild type plants (Figure 3.12a). 112

Furthermore, RNA was extracted from 50 days old leaves of individual plants and RT-

PCR of AtAGP18 were conducted. The 18NG lines had high AtAGP18 expression, while the VC lines had low AtAGP18 expression as did wild type (Figure 3.12b). The various phenotypes of 18NG lines are discussed below.

a

b

Figure 3.12 Genetic analyses of AtAGP18 overexpression (18NG) and vector control (VC) plants. a) PCR results with hygromycin resistance gene primers of 18NG and VC plants. Lane 1, 2 and 10, WT; lane 3-9, 11 and 12, 18NG lines; lane 13-16, VC lines. b) RT-PCR results showing the expression levels of AtAGP18 mRNA in 18NG and VC plants. Lane 1, 2 and 10, WT; lane 3-9 and 11, 18NG lines with high expression of AtAGP18; lane 12, 18NG line with low expression of AtAGP18; lane 13-16, VC lines.

AtAGP18 overexpressors had shorter stems, more branches and smaller rosettes

There was not much difference between 18NG lines and the wild type plants

during the first 3 weeks. During the 4th week and later, when the wild type plants developed long inflorescence stems, 18NG plants developed multiple but short branches 113

from the rosette (Figure3.13 and Table 3.1). The length of these stems was usually less

than one fifth of the wild type stem length, and branching number was usually more than

twice the number of the wild type, resulting in multiple flower buds crowded near the

rosette. Moreover, the rosette of 18NG plants was smaller than that of the wild type, and

some rosette leaves were curling abnormally (Figure 3.13).

a

b

c

Figure 3.13 Phenotypes of AtAGP18 (18NG) overexpression plants. a) Three-week old WT (left), 18NG (middle) and VC (right) plants. The 18NG plant has smaller rosette. b) Seven-week old WT (first), VC (second) and 18NG (third and fourth) plants. Insert shows a magnified picture of 18NG plant. 18NG plants are shorter with more branches. c) Leaf of WT (left) and 18NG (right) plants. 18NG leaves are curling abnormally.

114

Table 3.1 Analysis of WT, 18NG and VC plants Parameters WT 18NG VC Height (cm) 40.3±3.6 3.0±2.0 43.3±6.7 Rosette radius (cm) 7.7±1.4 4.7±1.2 7.3±0.9 Silique length (mm) 12.5±1.0 3.0±0.5 12.8±1.2 Root length (cm) 9.2±2.0 1.5±0.5 9.6±1.3 Branch number 4.5±1.3 11.9±3.8 5.0±0.8 Values are expressed as mean±SD Data are from 10 wild type, 10 18NG and 7 VC plants

AtAGP18 overexpressors had fewer viable seeds

The reproductive system in 18NG plants was also affected. Although 18NG plants had the same flowering time as the wild type, they usually produced smaller and sterile siliques (Figure 3.14a). As a result, most of the 18NG lines were sterile, even the partially fertile lines still had less seeds than the wild type. When flowers were dissected, flowers from 18NG plants were found to have short stamens such that apparently can’t function normally (Figure 3.14b).

115

a

b

Figure 3.14 Reproductive phenotypes of AtAGP18 (18NG) overexpressing plants. a) Siliques of WT (left) and 18NG (right) plants. 18NG siliques are shorter. b) Dissected flowers of WT (left) and 18NG (right) plants. 18NG flowers have shorter stamens so they can not fertilize normally.

AtAGP18 overexpressors had shorter primary roots and less lateral roots

Several 18NG lines died 4-6 weeks after germination. When these plants were taken out from the soil pots, they were found to have shorter primary roots and less lateral roots. These plants without the normal root system might have difficulty obtaining water and nutrition from the soil, and as a result, they are likely to die. Statistical data showed the 18NG lines (including the viable lines) had shorter primary roots than the wild type and VC lines (Figure 3.15). 116 WT 18NG

Figure 3.15 Root phenotypes of AtAGP18 overexpressing plants (18NG). Mature roots of WT (left) and 18NG (right) plants. 18NG roots are shorter.

AtAGP19 overexpressors showed no phenotype

Overexpression of AtAGP19 without the GFP tag resulted in 17 independent

19NG lines (Figure 3.16a), most of which had high expression of AtAGP19 (Figure

3.16b). These lines had indistinguishable phenotypes from the wild type. There was only one 19NG line (less than 10%) which was smaller and shorter compared with the wild type plants. This might be due to the disruption of some important genes by random insertion of the transgene. 117

a

b

Figure 3.16 Genetic analyses of AtAGP19 overexpressing plants. a) PCR results with hygromycin resistance gene primers of AtAGP19 overexpressing (19NG) plants. Lane 1 and 2, WT; lane 3-19, different 19NG lines. b) RT- PCR results showing the expression levels of AtAGP19 mRNA in 19NG plants. Lane 1 and 2, WT; lane 3-8, 10, 11, 14 and 16, 19NG lines with high expression of AtAGP19; lane 9, 12, 13 and 15, 19NG lines with low expression of AtAGP19.

ABA down-regulated AtAGP18 expression

Microarray data of hormone treatment from the TAIR website

(www.arabidopsis.org) showed that ABA down-regulated AtAGP18 expression (Figure

3.17a). In my experiment, seedlings treated with ABA appeared a yellow color (Figure

3.17b). This is consistent with the function of ABA in abscission. Further, RT-PCR

results showed that AtAGP18 mRNA levels decreased after ABA treatment (Figure

3.17c), which is consistent with the on-line microarray data. Moreover, GUS staining of

ABA treated seedlings showed lighter staining at the hypocotyl region compared to the

mock treatment (data not shown).

118

a

b

c

Figure 3.17 AtAGP18 expression is down-regulated by ABA treatment. a) Response of AtAGP18 expression to ABA treatment from the TAIR database. Y axis is the signal strength from the microarray data. b) Twelve day seedlings treated with ABA for 2 days (right). The same seedlings were treated with the same MS media without ABA (left, mock). Bar = 1 cm. Seedlings treated with 20 µM ABA became yellow. c) RT-PCR results of AtAGP18 expression by ABA treatment. The left two lanes were 6 d) seedlings; the right two lanes were 12 d seedlings.

119

Discussion

Subcellular localization of AtAGP17/18/19

Lysine-rich AGPs in Arabidopsis (AtAGP17/18/19) belong to the classical AGPs and are predicted to have a GPI-anchor addition sequence (Schultz et al., 2002). This sequence encodes a C-terminal hydrophobic domain. During post-translational modification, this domain is removed and replaced by a GPI lipid anchor for attachment to the plasma membrane. Biochemical analyses were carried out to support this hypothesis in tobacco and pear (Youl et al., 1998, Oxley and Bacic, 1999) and rose

(Svetek et al., 1999). Later, Sun et al. (2004a) did several experiments to prove the presence of GPI-anchor in LeAGP-1, a lysine-rich AGP in tomato. In this study, GFP-

AtAGP17/18/19 were expressed in Arabidopsis successfully, as the strong green fluorescence observed in different organs of the transformants demonstrated expression and cellular localization of these fusion proteins. To examine whether the fusion proteins are localized to the plasma membrane, leaf trichome cells were plasmolyzed with 4%

NaCl. The fusion protein was present on the plasma membrane, but no Hechtian strands were observed in trichome cells. Furthermore, protoplasts without cell walls made from pollen grains by enzyme digestion were observed and the fusion proteins were present on the protoplast plasma membrane (data not shown). The special localization of these lysine-rich AGPs to the plasma membrane provides some insights in exploring their possible functions, probably involved in cell signaling. 120

AtAGP17 and pollen germination

In vitro pollen germination is a convenient way to monitor the process of pollen tube elongation and the distribution of proteins on the cell surface. Previously, wild type tomato pollen grains were germinated in vitro and the PAP antibody (the antibody specific to the lysine-rich region on LeAGP-1) was used to localize LeAGP-1. As the pollen tube elongated, LeAGP-1 was present only at the pollen tube tip region, indicating

LeAGP-1 may be involved in cell wall material deposition at the pollen tip as it extends.

Here, in vitro germination of GFP-AtAGP17 pollen grains showed no fluorescence in the pollen tube region. There are several possible explanations: First, AtAGP17 is only one member of the lysine-rich AGPs in Arabidopsis. The other two members, namely

AtAGP18 and AtAGP19, may be involved in this process. Second, unlike tomato, lysine- rich AGPs in Arabidopsis are not involved in pollen germination. Which one is correct needs further experimentation.

AtAGP18 and overexpression phenotypes

AGPs play important roles in various phases of plant growth and development.

However, specific functions of any single AGP remains mostly elusive. To examine the role of LeAGP-1, a lysine-rich AGP in tomato, transgenic tomato plants were produced which expressed GFP-LeAGP-1 under the control of CaMV35S promoter. The transgenic 121 lines overexpressing LeAGP-1 were shorter and bushy, had less fruit and smaller seeds compared with wild type and GFP control plants (Sun et al., 2004b). Here, a similar method was used to overexpress GFP-AtAGP17/18/19 in Arabidopsis. At the same time,

GFP was also overexpressed in Arabidopsis as a control. Transgenic plants overexpressing GFP-AtAGP17 or GFP-AtAGP19 were not significantly different from

GFP control plants, while a transgenic plant overexpressing GFP-AtAGP18 had the phenotypes similar to the GFP-LeAGP-1 overexpressors in tomato.

Although these Arabidopsis transgenic plants provided some clues to the functions of these lysine-rich AGPs, there are still some problems. First, the number of independent transgenic lines of GFP-AtAGP18 was limited. Only one line was screened and showed phenotypes. Second and most important, although GFP control plants showed no phenotypes, the possibility that GFP in the fusion protein affected the plant phenotypes can not be excluded. To exclude the effect of GFP on transformants phenotypes, AtAGP17/18/19 without a GFP tag were overexpressed in Arabidopsis. A vector control construct was delivered into Arabidopsis as well. While 7 VC transformants as well as 10 lines of 17NG and 16 lines of 19NG displayed the same phenotypes as the wild type plants, 18 lines of 18NG (slightly more than 50% of all

18NG transformants) had several phenotypes different from the wild type, with respect to development of inflorescence stems and root system, stem branching and reproduction.

Moreover, these phenotypes were insignificant at early stages of plant development, but became obvious after 3 weeks, when inflorescence stems began to develop. 122

Screening of 18NG plants resulted in 32 plants containing the 35S-AtAGP18

construct. Eighteen of of the 32 plants displayed phenotypes distinct from the wild type plants. With these 18 plants, except for 3 plants that died within 5 weeks, RT-PCR of

AtAGP18 was performed and all 15 plants showed high levels of AtAGP18 mRNA. RT-

PCR was also performed on some of these transgenic plants that had wild type phenotypes. Some of them (4 plants) had endogenous expression levels similar to the wild type, while others (3 plants) had high expression levels. Thus, the phenotypes were not completely correlated with the expression levels. One explanation for this is that the

AtAGP18 expression was detected at the RNA level, but not at the protein level, and high mRNA expression does not necessarily guarantee high protein expression.

Among the three overexpressors (AtAGP17/18/19), only AtAGP18 had phenotypes distinguishing it from the wild type. This is probably because the endogenous expression level of AtAGP18 is the highest among these three genes in wild type plants according to the microarray data (Yang, 2006). In addition, northern blot analysis shows the expression pattern of these three AGPs are organ-specific. AtAGP17 is expressed moderately in leaves and stems, but not in flowers and roots. In contrast, the expression pattern of AtAGP18 shows a high level in flowers and moderate level in roots, stems and young leaves, which is nearly identical to the expression pattern of LeAGP-1 (Li and

Showalter, 1996; Sun et al., 2005). The AtAGP19 expression pattern is similar to

AtAGP18 but much weaker (Sun et al., 2005). Considering all the above expression information about these three AGPs, transgenic plants overexpressing AtAGP18 are most likely to have phenotypes similar to transgenic plants overexpressing LeAGP-1. 123

The Arabidopsis AtAGP18 overexpressors and tomato LeAGP-1 overexpressors

both have a bushy phenotype similar to transgenic tobacco plants overproducing

cytokinin (Li et al., 1992; Sun et al., 2004b). Cytokinins are an important class of plant hormones involved in many plant growth and development processes, such as cell growth and division, differentiation, and other physiological processes (Sakakibara, 2006).

Therefore, Sun et al. (2004b) proposed LeAGP-1 might function in concert with the cytokinin signal transduction pathway. Since the overexpression phenotypes of AtAGP18 are similar to LeAGP-1, AtAGP18 is also likely related to the cytokinin signal transduction pathway.

A number of genes co-expressed with AtAGP18 were obtained through the

Arabidopsis thaliana Co-Response Database (http://csbdb.mpimp-

golm.mpg.de/csbdb/dbcor/ath.html). These genes belong to different families and are

involved in various functions (Yang and Showalter, 2007). Interestingly, these genes

include some receptors on the plasma membrane (such as receptor-like kinases) which

may be related to hormone signaling.

The involvement of AtAGP18 and some transmembrane receptors with cytokinin signal transduction suggests this signaling process may be associated with lipid rafts. In plants as well as in animals, there are cholesterol-enriched, detergent-resistant plasma membrane microdomains called lipid rafts (Seifert and Roberts 2007). Lipid rafts are

thought to be involved in signal transduction since a number of cell signaling related

proteins such as AGPs, other GPI-anchored proteins and some transmembrane proteins

are associated with such microdomains (Pike 2003, Johnson and Ingram 2005). The 124 accumulation of these proteins in microdomains indicates that probably there are interactions between these proteins in sensing extracelluar signals which lead to various intracellular events. AGPs may be associated with receptor-like kinases (RLKs) since immunofluorescence labeling showed colocalization of AGP epitopes and some RLKs

(Gens et al. 2000). The complex structure and special subcellular localization of AGPs in lipid rafts together with other signal related proteins provide a potential signaling role of

AtAGP18. AtAGP18 may be involved in a signal transduction cascade by binding some plant growth factors (such as cytokinins) and bringing these growth factors to the transmembrane receptor in lipid rafts to facilitate their interaction. While further evidence is required to support this hypothesis, it is clear that AtAGP18 has functional roles in vegetative growth and reproduction. 125

CHAPTER 4 EXPRESSION AND MUTANT STUDY OF ATAGP14

Summary

AtAGP14 (At5g56540) encodes an AG peptide in Arabidopsis with an N-terminal signal sequence, a classical AGP domain and a C-terminal GPI anchor addition sequence.

Northern blotting and Reverse Transcription-PCR (RT-PCR) studies with AtAGP14 revealed high mRNA levels in flowers and young roots, and moderate mRNA levels in seedlings, stems and rosette leaves. These results correspond with the microarray based expression profiles for AtAGP14. In order to explore the functions of AtAGP14, a T-

DNA insertional mutant of AtAGP14 was obtained from ABRC and a PCR-based screening method was used to select homozygous (HM) mutants. Several individual homozygous mutant plants were obtained and RT-PCR analysis showed the absence of

AtAGP14 mRNA in the homozygous mutant. A plate-based phenotypic analysis was carried out for the homozygous mutant using wild type Arabidopsis as a control.

However, no significant differences between wild type and the homozygous mutant were observed with respect to germination rate, true leaf numbers, primary root length and lateral root numbers. The lack of a discernable phenotype for atagp14 may be attributed to gene redundancy, given that AtAGP13 is a homologous gene which has a similar expression pattern to AtAGP14. 126

Introduction

In the AGP family, there is a group of AGPs called AG peptides. Like classical

AGPs, these AG peptides also contain an N-terminal signal sequence, a typical AGP domain and a C-terminal GPI anchor addition sequence. Interestingly, the AGP domains of these AG peptides are so short that after processing, the mature AG peptides only consist of 10 to 15 amino acids (Schultz et al., 2002). In Arabidopsis, 10 AG peptides were identified using a bioinformatics approach based on the amino acid composition

(Schultz et al., 2002). It is interesting that the nature creates so many and such small AG peptides in this model plant. To date, little research has been carried out on the functions of these peptides. In order to initiate a functional study on an AG peptide, a T-DNA mutant of AtAGP14 was obtained and examined; this T-DNA mutant occurs in the exon region and was the only mutant available for this gene at the beginning of this study.

Here, homozygous mutant plants were selected by the PCR-based screening, and further phenotype analyses of the homozygous mutant were carried out based on the expression pattern of this gene by northern blotting and RT-PCR. 127

Materials and methods

Plant growth, RNA extraction and RT-PCR

Plant materials were obtained and used for RNA extraction and then RT-PCR as described in chapter 3 on organ specific expression of AtAGP17. The primers for

AtAGP14 expression were 5’- AAC AAC AAC ACA ACA ATA AGC ACC -3’ and 5’-

CGA ATG AGT CAA ATT CAA CAG CG -3’. The primers for actin as an internal control were same as described before.

Northern blotting

Northern blotting was performed according to Sun et al. (2005) with minor modifications. Total RNA was concentrated by a Lyph. Lock6 Freeze dry/ Shell freeze system (Labconco Corporation, Kansas, MO) to a final concentration of 3 μg/μl and 5 μl total RNA (15 μg) was mixed with 15.5 μl fresh sample buffer (mix 40 μl of 10 X MOPS,

70 μl of formaldehyde and 200 μl of formamide) and 1μl ethidium bromide before loading on a 1% denaturing agarose-formaldehyde gel. The RNA samples were electrophoresed at no more than 60 volts for about 5 h and transferred onto Zeta-Probe

Genomic Tested Blotting Membrane (Bio-Rad, Hercules, CA) or Nitrocellulose Transfer

Membrane (Micron Separations Inc., Westboro, MA) in 10X SSC for 12-16 h. 128

Gene-specific probes were amplified by PCR and purified with QIAquick PCR

Purification Kit (QIAGEN, Valencia, CA). The purified probes were labeled with α-32P- dCTP using the Prime-a-gene labeling system (Promega, Madison, WI) according to the manufacturer’s instructions.

The membrane was pre-hybridized in 1X hybridization buffer [7% sodium dodecyl sulphate (SDS), 0.25 M Na2HPO4, pH 7.2] at 65°C for 5 min and then hybridized in 1X hybridization buffer with appropriate probe(s) at 65°C overnight with gentle shaking. The membrane was then washed with 1/2X hybridization buffer once, followed by 1/4X hybridization buffer twice and finally 1/7X hybridization buffer once.

All the washing steps were at 65°C for 15 min with gentle shaking. The membrane was then wrapped in saran wrap and exposed to Kodak Biomax MS film in a cassette at -80°C for 48 to 72 h. The film was developed with Kodak GBX developer and fixer (Fisher

Scientific, Hampton, NH).

The membrane was stripped in stripping buffer (0.5% SDS, 0.1X SSC) by boiling three times, 15 min each time. The membrane was then exposed to the film for 2 days to check the stripping efficiency. Stripping steps were repeated if radioactivity was still detected. The stripped membrane was used for re-hybridization.

Bioinformatics analysis

The microarray-based gene expression profiles were obtained from

Genevestigator database (https://www.genevestigator.ethz.ch/) with the AGI locus 129

numbers of AtAGP14 (At5g56540) and AtAGP13 (At4g26320). The array type was

“ATH1: 22k array” and “High quality arrays only” was chosen for quality control.

Screening of T-DNA homozygous line

A T-DNA insertion line of AtAGP14 (SALK_096806) was obtained from

Arabidopsis Biological Resource Center (ABRC) as the T3 generation. The seeds contained wild type, homozygous mutant and heterozygous mutant.

Seeds were sowed in soil and grown in the growth room for three weeks for DNA extraction from individual plants with the Extract-N-Amp Plant PCR Kit (Sigma-Aldrich,

St. Louis, MO). PCR was performed to identify the homozygous mutants. An illustration of the PCR screening rationale is shown in Figure 4.1. A pair of gene specific primers FP and RP (the same as those used in RT-PCR of AtAGP14 expression) was designed to flank the T-DNA insertion site and this pair of primers can amplify a DNA fragment of a certain size. The LBa1 primer (5’-TGG TTC ACG TAG TGG GCC ATC G-3’) on the T-

DNA left border was also used and the LBa1 primer with either FP or RP (since the T-

DNA can be inserted in both directions) can amplify a DNA fragment of a different size.

In the PCR reaction, when the above three primers were added, wild type plants produced a DNA band from FP and RP; homozygous mutant plants produced a DNA band from

Lba1 and FP/RP, but no band from FP and RP since the T-DNA is too large; and heterozygous mutant plants produced both DNA bands.

To examine the expression level of AtAGP14 in the mutant, RNA extraction from wild type plants and homozygous mutants and then RT-PCR (the same primers for 130

AtAGP14 expression and actin control) were performed as described earlier in this chapter.

LBa1

Exon

UTR T-DNA FP ATG TAA RP Primer

Figure 4.1 An illustration of the PCR screening rationale. The T-DNA insertion is not drawn to scale. LBa1, T-DNA primer; FP, forward gene specific primer; RP, reverse gene specific primer; ATG, start codon; TAA, stop codon. Wild type plants amplified a DNA band from FP and RP; homozygous mutant plants amplified a DNA band of a different size from LBa1 and RP; and heterozygous mutant plants amplified both DNA bands.

Plate-based phenotypic analysis

Wild type and AtAGP14 homozygous mutant seeds were surface sterilized and

sown on MS plate supplied with 1% sucrose. Eight seeds were aligned and grown per 10

cm square plate with the same space between the seeds. After stratification, the plates

were placed vertically in the tissue culture room at 22°C with 16h day/8h night cycle for

2 weeks. Different statistical data, such as days required to develop into certain growth

stages, germination rates, true leaf numbers, primary root length and lateral root numbers

were recorded according to Boyes et al. (2001).

131

Results

Organ-specific expression analysis and Microarray expression data

RT-PCR results from different organs revealed that AtAGP14 was expressed most

in roots, followed by flowers. It is also expressed in seedlings, stems and rosette leaves at

a lower level. The expression pattern deduced from northern blotting data was consistent

with the RT-PCR results (Figure 4.2).

Microarray based expression profile was also obtained from Genevestigator

database to confirm RT-PCR and northern blot results (Figure 4.3). AtAGP14 expression

was high in roots, moderate in flowers (specifically in petals) and seeds, and low in other

organs, which supported our experimental data.

Expression profiles of AtAGP13 from Genevestigator were also examined since

this gene is most related to AtAGP14 by phylogenetic analyses (Sardar, 2007).

Interestingly, the expression pattern was quite similar to AtAGP14 (Figure 4.4), with

highest expression in roots and much lower expression in other organs, suggesting these

two genes may have complementary functions.

132

a

b

Figure 4.2 RT-PCR and northern blotting results of AtAGP14 mRNA expression levels in different organs. mRNA expression levels of different organs were normalized with actin and relative mRNA abundance was also shown. a) RT-PCR results of AtAGP14. M, 100 bp DNA ladder; SL, seedlings; FL, flowers; ST, stems; RT, roots; RL, rosette leaves. b) Northern blotting of AtAGP14. SL, seedlings; FL, flowers; ST, stems; RT, roots; RL, rosette leaves.

133

Figure 4.3 Expression profiles of AtAGP14 from microarray data in Arabidopsis. The data was obtained from Genevestigator database (https://www.genevestigator.ethz.ch/) under “Anatomy” tab.

134

Figure 4.4 Expression profiles of AtAGP13 from microarray data in Arabidopsis. The data was obtained from Genevestigator database (https://www.genevestigator.ethz.ch/) under “Anatomy” tab.

Isolation of homozygous line from SALK_096806

The information of available T-DNA lines was obtained from the SIGnAL "T-

DNA Express" Arabidopsis Gene Mapping Tool (http://signal.salk.edu/cgi-

bin/tdnaexpress). Only one T-DNA line (SALK_096806) of AtAGP14 had a T-DNA

insertion in the exon region of this gene. 135

From the PCR screening results, wild type had a 400 bp PCR product, the

homozygous mutant had a single band at 550 bp, and heterozygous mutant had both

bands (Figure 4.5a). A total of 17 plants were screened and 4 plants were homozygous

mutants. This result was generally consistent with the Mendel’s law (4:13=1:3.25, close

to 1:3 ratio). Next, RT-PCR results showed there was no AtAGP14 mRNA expression in the homozygous mutant (Figure 4.5b).

a

b

Figure 4.5 SALK_096806 was a null knockout mutant of AtAGP14. a) PCR screening of WT, homozygous and heterozygous plants. M, 100 bp DNA ladder; WT, wild type; HM, homozygous; HZ, heterozygous. b) RT-PCR result showed that AtAGP14 mRNA was present in the wild type plant but absent in the homozygous mutant. Actin was used as the internal control.

136

Phenotypic analyses showed no significant differences between mutant and wild type

Seeds from the homozygous mutants were collected and planted for phenotypic analyses. The growth stage phenotypes of atagp14 were not significantly different from

those of wild type (Table 4.1). Since AtAGP14 expression was highest in roots, root

phenotypes (i.e. primary root length and lateral root number) of atagp14 and the wild

type were compared and no differences were observed (Figure 4.6).

Table 4.1 Plate-based growth stage phenotype analysis of WT and atagp14 seedlings Stage Description WT (Days)* atagp14 (Days) 0.1 Stratification 3.0 3.0 0.5 Radicle emergence 4.3±0.3 4.4±0.3 0.7 Hypocotyl and cotyledon emergence 5.2±0.4 5.2±0.4 1.0 Cotyledons fully opened 6.1±0.5 6.1±0.5 1.02 2 rosette leaves >1 mm 10.1±0.6 10.0±0.6 1.04 4 rosette leaves >1 mm 14.2±0.7 14.1±0.6 *Average days±SD of more than 80 plants.

137 a

b

Figure 4.6 atagp14 had no obvious phenotypes in roots. a) Primary root length of WT and atagp14 seedlings at day 5, 7, 10 and 14. Error bars indicate SD (n > 40). b) Lateral root number of WT and atagp14 seedlings at day 10 and 12. Error bars indicate SD (n > 40).

138

Discussion

Ten AG peptides were identified in Arabidopsis (Schultz et al., 2002). The

`number of their expressed sequence tags (ESTs) were summarized in Schultz’s paper

since this information shows how abundant each gene is expressed in general to provide

an “expression summary” of each gene (Schultz et al., 2002). From the EST expression summary, several AG peptides were abundantly expressed in roots and flowers.

Interestingly, the mature AG peptides are so short that they only contain 10 to 15 amino

acids. The abundance of these small peptides indicates that they have important, yet not

clear functions. To begin some studies of these AG peptides, I did an online search for

the available T-DNA mutants of these genes from Arabidopsis Biological Resource

Center (ABRC) since this reverse genetics approach provides us a convenient way to

study gene function in Arabidopsis. Among the mutants of these genes, atagp14 has a T-

DNA insertion in the exon region and was the only mutant available for this gene at the

beginning of this study.

Expression profiles for AtAGP14 from Genevestigator database indicate it is

highly expressed in roots (especially in stele, endodermis and root hair), and moderately expressed in flowers (especially in petal). Here, RT-PCR and northern blotting results

were generally consistent with the microarray-based expression profiles. When relative

mRNA abundance of each organ was measured by image J software, northern blotting

results showed a better correlation to the microarray-based expression profiles,

suggesting norther blotting may be more reliable. 139

A T-DNA insertion line in the exon region of AtAGP14 was obtained in this study and homozygous mutant plants were screened successfully. As expected, the mRNA expression of AtAGP14 was not detected by RT-PCR in the homozygous mutant, since the T-DNA insertion is in the exon region. As microarray data and the experiment results showed, AtAGP14 had the highest expression in roots, as this gene might have an important function(s) in root growth and development. It is expected that loss of this gene in the homozygous mutant may have a dramatic effect on roots. To identify if there are any phenotypic aberrations in the atagp14 mutant, especially in the roots, a plate based phenotypic analysis was carried out since this is a convenient method to examine root growth and development. To avoid any variations in the growth chamber conditions and in the growth media, wild type plants and homozygous mutant plants were grown and analyzed side by side on the same plate. To achieve statistically significant results, at least 40 wild type plants and 40 mutant plants were analyzed with respect to days required to develop into certain growth stages, germination rates, true leaf numbers, primary root length and lateral root numbers. However, the atagp14 mutant grew normally under standard growth conditions. Its root system developed like the wild type without any observable defects. This is likely because of redundancy of a homologous gene, AtAGP13, which has both a similar gene structure and a similar expression pattern.

To prove this idea, further experimentation can be done to examine if the mRNA expression level of AtAGP13 increases in the atagp14 mutant. To continue the study on

AtAGP14, the following work can be carried out: Search atagp13 mutant and make a double mutant with atagp14 to see if there is any phenotype in this double mutant. 140

Another way is to do some phenotypic analyses of atagp14 under certain stress conditions to see if there is any difference. 141

CHAPTER 5 CONCLUSIONS

Arabinogalactan-proteins (AGPs) are a class of highly glycosylated, hydroxyproline-rich structural proteins that are wildly distributed in the plant kingdom, from lower plants such as bryophytes and algae to higher plants such as angiosperms

(Fincher et al., 1983; Showalter, 2001). AGPs in Arabidopsis can be categorized into five groups: the classical AGPs, lysine-rich classical AGPs, AG peptides, fasciclin-like AGPs

(FLAs), and other chimeric AGPs (Schultz et al., 2002). This dissertation research focused on the lysine-rich classical AGPs and an AG peptide in Arabidopsis. Various molecular biology approaches, such as a functional genomics approach (microarray), mutant phenotype analysis, and overexpression phenotype analysis were employed to discover the functions of these AGPs in plant growth and development.

T-DNA mutant study of AtAGP19

Lysine-rich classical AGPs consist of an N-terminal signal peptide, a classical AGP domain disrupted by a short basic lysine-rich subdomain and a C-terminal glycosylphosphatidylinositol (GPI) anchor addition sequence. AtAGP19 is a member of this lysine-rich classical AGP gene subfamily. A null T-DNA insertion mutant of

AtAGP19 which displays pleiotropic phenotypes was obtained from the Arabidopsis

Biological Resource Center (ABRC). Compared to wild type plants, the atagp19 mutant had: 1) lighter green leaves, 2) smaller and rounder leaves, with shorter petioles, 3) 142

shorter hypocotyls, 4) shorter and thinner inflorescence stems, 5) slower growth with

delayed and reduced flowering, 6) fewer siliques and seeds and 7) fewer lateral roots.

Complementation of this mutant with the wild type AtAGP19 gene restored all the wild type phenotypes (Yang et al., 2007). Here, a microarray approach was employed to elucidate changes in gene expression associated with the atagp19 mutant. The expression

levels of two homologous lysine-rich classical AGP genes, namely AtAGP17 and

AtAGP18, did not change significantly, indicating these two genes do not compensate for

the loss of AtAGP19. In contrast, several genes related to cell expansion and cell wall

modification such as xyloglucan endotransglycosylases/hydrolases (XTHs),

pectinesterase (PE), wall-associated kinases (WAKs) and receptor-like kinases (RLKs)

had significantly different expression levels in the mutant compared to the wild type.

Interestingly, one gene (At1g68720, cytidine/deoxycytidylate deaminase family protein)

adjacent to AtAGP19 was found to be down-regulated about 50 fold. Furthermore,

complementation with the 3’ portion of the At1g68720 gene can fully restore all the wild

type phenotypes, indicating this partial region is critical for the functions/phenotypes

observed in the atagp19 mutant. However, how this fragment of a gene without a start

codon and promoter can be expressed in the complemented plants is still unclear. It is

possible that this fragment is inserted by chance after a promoter region and expressed

under the control of this promoter.

Overexpression of AtAGP17, 18 and 19

143

This research project focused on the functions of three members (AtAGP17, 18 and 19) in the lysine-rich AGP subfamily using an overexpression approach. GFP-

AtAGP17/18/19 fusion proteins as well as a GFP control were overexpressed in

Arabidopsis plants to examine the subcellular localizations of these lysine-rich AGPs.

The fusion proteins were present on the plant cell surface of different organs. Confocal laser scanning microscopy of plasmolysed leaf trichome cells suggested that AtAGP17,

18 and 19 are localized on the plasma membrane, supporting the prediction that these

AGPs contain a glycosylphosphatidylinositol (GPI) anchor. The subcellular localization of lysine-rich AGPs to the plasma membrane implies they may have roles in cell signaling.

To elucidate AtAGP17/18/19 function(s), and to exclude the artificial effects of

GFP, AtAGP17/18/19 were expressed in transgenic plants without the GFP tag under the control of 35S promoter. At the same time, a vector control construct was also introduced into Arabidopsis plants. mRNA levels of AtAGP17/18/19 were examined in the transgenic plants. In contrast to AtAGP17 and AtAGP19 overexpressors which showed no observable phenotypes from both the wild type plants and the vector control transgenic plants, AtAGP18 overexpressors with high mRNA levels displayed several phenotypes distinct from the wild type plants: they were shorter, produced more branches from the rosettes, had shorter roots and had less seed production. Interestingly, a previous

study on LeAGP-1, a lysine-rich AGP in tomato, showed LeAGP-1 overexpressors also

had an overbranching phenotype similar to AtAGP18 overexpressor (Sun et al., 2004b),

suggesting these two AGP genes may have similar function(s). Considering the plasma 144 membrane localization of AtAGP18, it may serve as a co-receptor to bind some signal molecules and bring them to the plasma membrane receptor to initiate cell signaling.

Expression and mutant study of AtAGP14

AtAGP14 belongs to the AG peptide subfamily. In this study, northern blotting and Reverse Transcription-PCR (RT-PCR) were used to examine the AtAGP14 expression pattern. AtAGP14 was highly expressed in young roots and flowers, and moderately expressed in seedlings, stems and rosette leaves. These results are consistent with the microarray based expression profiles for AtAGP14. To begin a functional study of AtAGP14, a T-DNA insertional mutant of AtAGP14 was obtained from ABRC and homozygous mutants were selected by a PCR-based screening method. A plate-based phenotypic analysis was carried out to examine if there are any phenotypic aberrations in the atagp14 homozygous mutant, especially in the roots where the expression of

AtAGP14 is the highest. However, no significant differences between wild type and the homozygous mutant were observed. Since AtAGP14 has a homologous gene, AtAGP13, which has a similar expression pattern to AtAGP14, it is possible that gene redundancy accounts for the lack of phenotypes. 145

REFERENCES

Acosta-Garcia, G. & Vielle-Calzada, J. (2004), 'A classical arabinogalactan protein is essential for the initiation of female gametogenesis in Arabidopsis.', Plant Cell 16(10), 2614--2628.

Akita, M.; Nielsen, E. & Keegstra, K. (1997), 'Identification of protein transport complexes in the chloroplastic envelope membranes via chemical cross-linking', The Journal of Cell Biology 136(5), 983--994.

Akiyama, Y. & Kato, K. (1981), 'An extracellular arabinogalactan-protein from Nicotiana tabacum', Phytochemistry 20, 2507--2510.

Anderson, C.; Wagner, T.; Perret, M.; He, Z.; He, D. & Kohorn, B. (2001), 'WAKs: cell wall-associated kinases linking the cytoplasm to the extracellular matrix', Plant Molecular Biology 47(1), 197--206.

Baldwin, T.; McCann, M. & Roberts, K. (1993), 'A novel hydroxyproline-deficient arabinogalactan protein secreted by suspension-cultured cells of Daucus carota (purification and partial characterization).', Plant Physiology 103(1), 115.

Baskin, T. (2001), 'On the alignment of cellulose microfibrils by cortical microtubules: A review and a model', Protoplasma 215(1), 150--171.

Borner, G.; Lilley, K.; Stevens, T. & Dupree, P. (2003), 'Identification of glycosylphosphatidylinositol-anchored proteins in Arabidopsis. A proteomic and genomic analysis', Plant Physiology 132(2), 568.

Borner, G.; Sherrier, D.; Stevens, T.; Arkin, I. & Dupree, P. (2002), 'Prediction of glycosylphosphatidylinositol-anchored proteins in Arabidopsis. A genomic analysis', Plant Physiol 129(2), 486--499.

Boyes, D.; Zayed, A.; Ascenzi, R.; McCaskill, A.; Hoffman, N.; Davis, K. & Gцrlach, J. (2001), 'Growth stage--based phenotypic analysis of Arabidopsis', The Plant Cell Online 13, 1499--1510.

Butowt, R.; Niklas, A.; Rodriguez-Garcia, M. & Majewska-Sawka, A. (1999), 'Involvement of JIM13-and JIM8-responsive carbohydrate epitopes in early stages of cell wall formation', Journal of Plant Research 112(1), 107--116.

Carpita, N. & Gibeaut, D. (1993), 'Structural models of primary cell walls in flowering plants: consistency of molecular structure with the physical properties of the walls during growth.', Plant J 3(1), 1--30.

146

Casero, P.; Casimiro, I. & Knox, J. (1998), 'Occurrence of cell surface arabinogalactan- protein and extensin epitopes in relation to pericycle and vascular tissue development in the root apex of four species', Planta 204(2), 252--259.

Chapman, A.; Blervacq, A.; Vasseur, J. & Hilbert, J. (2000), 'Arabinogalactan-proteins in Cichorium somatic embryogenesis: effect of beta-glucosyl Yariv reagent and epitope localisation during embryo development', Planta 211(3), 305--314.

Cheung, A. (1996), 'Pollen—pistil interactions during pollen-tube growth', Trends in Plant Science 1(2), 45--51.

Cheung, A.; Wang, H. & Wu, H. (1995), 'A floral transmitting tissue-specific glycoprotein attracts pollen tubes and stimulates their growth', Cell 82(3), 383--393.

Cheung, A. & Wu, H. (1999), 'Arabinogalactan proteins in plant sexual reproduction', Protoplasma 208(1), 87--98.

Clough, S. & Bent, A. (1998), 'Floral dip: a simplified method for Agrobacterium- mediated transformation of Arabidopsis thaliana', Plant Journal 16(6), 735--743.

Cock, J.; Vanoosthuyse, V. & Gaude, T. (2002), 'Receptor kinase signalling in plants and animals: distinct molecular systems with mechanistic similarities', Current Opinion in Cell Biology 14(2), 230--236.

Cosgrove, D. (2005), 'Growth of the plant cell wall', Nat Rev Mol Cell Biol 6, 850--861.

Curtis, M. & Grossniklaus, U. (2003), 'A gateway cloning vector set for high-throughput functional analysis of genes in planta', Plant Physiology 133(2), 462.

Ding, L. & Zhu, J. (1997), 'A role for arabinogalactan-proteins in root epidermal cell expansion', Planta 203(3), 289--294.

Dolan, L.; Linstead, P. & Roberts, K. (1995), 'An AGP epitope distinguishes a central metaxylem initial from other vascular initials in the Arabidopsis root', Protoplasma 189(3), 149--155.

Du, H.; Simpson, R.; Clarke, A. & Bacic, A. (1996), 'Molecular characterization of a stigma-specific gene encoding an arabinogalactan-protein (AGP) from Nicotiana alata', The Plant Journal 9(3), 313--323.

Du, H.; Simpson, R.; Moritz, R.; Clarke, A. & Bacic, A. (1994), 'Isolation of the protein backbone of an arabinogalactan-protein from the styles of Nicotiana alata and characterization of a corresponding cDNA', Plant Cell 6(11), 1643--53.

147

Egertsdotter, U. & Arnold, S. (1995), 'Importance of arabinogalactan proteins for the development of somatic embryos of Norway spruce (Picea abies)', Physiologia Plantarum 93(2), 334--345.

Eisenhaber, B.; Bork, P. & Eisenhaber, F. (1998), 'Sequence properties of GPI-anchored proteins near the omega-site: constraints for the polypeptide binding site of the putative transamidase.', Protein Eng 11(12), 1155--61.

Eisenhaber, B.; Wildpaner, M.; Schultz, C.; Borner, G.; Dupree, P. & Eisenhaber, F. (2003), 'Glycosylphosphatidylinositol lipid anchoring of plant proteins. Sensitive prediction from sequence-and genome-wide studies for Arabidopsis and rice', Plant Physiology 133(4), 1691.

Faik, A.; Abouzouhair, J. & Sarhan, F. (2006), 'Putative fasciclin-like arabinogalactan- proteins (FLA) in wheat (Triticum aestivum) and rice (Oryza sativa): identification and bioinformatic analyses', Molecular Genetics and Genomics 276(5), 478--494.

Fincher, G.; Stone, B. & Clarke, A. (1983), 'Arabinogalactan-proteins: structure, biosynthesis, and function', Annual Reviews in Plant Physiology 34(1), 47--70.

Gao, M.; Kieliszewski, M.; Lamport, D. & Showalter, A. (1999), 'Isolation, characterization and immunolocalization of a novel, modular tomato arabinogalactan- protein corresponding to the LeAGP-1 gene', The Plant Journal 18(1), 43--55.

Gao, M. & Showalter, A. (2000), 'Immunolocalization of LeAGP-1, a modular arabinogalactan-protein, reveals its developmentally regulated expression in tomato', Planta 210(6), 865--874.

Gaspar, Y.; Nam, J.; Schultz, C.; Lee, L.; Gilson, P.; Gelvin, S. & Bacic, A. (2004), 'Characterization of the Arabidopsis lysine-rich arabinogalactan-protein AtAGP17 mutant (rat1) that results in a decreased efficiency of Agrobacterium transformation', Plant Physiology.

Gilson, P.; Gaspar, Y.; Oxley, D.; Youl, J. & Bacic, A. (2001), 'NaAGP4 is an arabinogalactan protein whose expression is suppressed by wounding and fungal infection in Nicotiana alata', Protoplasma 215(1), 128--139.

Gleeson, P.; McNamara, M.; Wettenhall, R.; Stone, B. & Fincher, G. (1989), 'Characterization of the hydroxyproline-rich protein core of an arabinogalactan-protein secreted from suspension-cultured Lolium multiflorum (Italian ryegrass) endosperm cells.', Biochemical Journal 264(3), 857.

148

Guan, Y. & Nothnagel, E. (2004), 'Binding of arabinogalactan proteins by Yariv phenylglycoside triggers wound-like responses in Arabidopsis cell cultures', Plant Physiology.

He, Z.; Fujiki, M. & Kohorn, B. (1996), 'A cell wall-associated, receptor-like protein kinase', Journal of Biological Chemistry 271(33), 19789.

He, Z.; He, D. & Kohorn, B. (1998), 'Requirement for the induced expression of a cell wall associated receptor kinase for survival during the pathogen response', The Plant Journal 14(1), 55--63.

Ito, S.; Suzuki, Y.; Miyamoto, K.; Ueda, J. & Yamaguchi, I. (2005), 'AtFLA11, a fasciclin-like arabinogalactan-protein, specifically localized in screlenchyma cells', Bioscience, Biotechnology, and Biochemistry 69(10), 1963--1969.

Jauh, G.; Eckard, K.; Nothnagel, E. & Lord, E. (1997), 'Adhesion of lily pollen tubes on an artificial matrix', Sexual Plant Reproduction 10(3), 173--180.

Jiang, L.; Yang, S.; Xie, L.; Puah, C.; Zhang, X.; Yang, W.; Sundaresan, V. & Ye, D. (2005), 'VANGUARD1 encodes a pectin methylesterase that enhances pollen tube growth in the Arabidopsis style and transmitting tract', Plant Cell 17(2), 584--596.

Johnson, K. & Ingram, G. (2005), 'Sending the right signals: regulating receptor kinase activity', Current Opinion in Plant Biology 8(6), 648--656.

Kaya, H.; Sato, S.; Tabata, S.; Kobayashi, Y.; Iwabuchi, M. & Araki, T. (2000), 'hosoba toge toge, a Syndrome Caused by a Large Chromosomal Deletion Associated with a T- DNA Insertion in Arabidopsis', Plant and Cell Physiology 41(9), 1055--1066.

Keller, B. (1993), 'Structural cell wall proteins.', Plant Physiology 101(4), 1127.

Kieliszewski, M. & Lamport, D. (1994), 'Extensin: repetitive motifs, functional sites, post-translational codes, and phylogeny.', Plant J 5(2), 157--72.

Kieliszewski, M. & Shpak, E. (2001), 'Synthetic genes for the elucidation of glycosylation codes for arabinogalactan-proteins and other hydroxyproline-rich glycoproteins', Cellular and Molecular Life Sciences (CMLS) 58(10), 1386--1398.

Kohorn, B. (2001), 'WAKs; cell wall associated kinases', Current Opinion in Cell Biology 13(5), 529--533.

Kreuger, M. & Holst, G. (1996), 'Arabinogalactan proteins and plant differentiation', Plant Molecular Biology 30(6), 1077--1086.

149

Kreuger, M. & Holst, G. (1993), 'Arabinogalactan proteins are essential in somatic embryogenesis of Daucus carota L.', Planta 189(2), 243--248.

Lally, D.; Ingmire, P.; Tong, H. & He, Z. (2001), 'Antisense expression of a cell wall- associated protein kinase, WAK4, inhibits cell elongation and alters morphology', The Plant Cell Online 13(6), 1317.

Laufs, P.; Autran, D. & Traas, J. (1999), 'A chromosomal paracentric inversion associated with T-DNA integration in Arabidopsis', The Plant Journal 18(2), 131--139.

Lee, K.; Sakata, Y.; Mau, S.; Pettolino, F.; Bacic, A.; Quatrano, R.; Knight, C. & Knox, J. (2005), 'Arabinogalactan proteins are required for apical cell extension in the moss Physcomitrella patens', The Plant Cell 17(11), 3051.

Li, S. & Showalter, A. (1996), 'Cloning and developmental/stress-regulated expression of a gene encoding a tomato arabinogalactan protein', Plant Molecular Biology 32(4), 641-- 652.

Li, Y.; Bruun, L.; Pierson, E. & Cresti, M. (1992), 'Periodic deposition of arabinogalactan epitopes in the cell wall of pollen tubes of Nicotiana tabacum L.', Planta 188(4), 532-- 538.

Li, Y.; Hagen, G. & Guilfoyle, T. (1992), 'Altered morphology in transgenic tobacco plants that overproduce cytokinins in specific tissues and organs.', Dev Biol 153(2), 386-- 95.

Livak, K. & Schmittgen, T. (2001), 'Analysis of relative gene expression data using real- time quantitative PCR and the 2(-delta delta C(T)) method', Methods 25(4), 402--408.

Lu, H.; Chen, M. & Showalter, A. (2001), 'Developmental expression and perturbation of arabinogalactan-proteins during seed germination and seedling growth in tomato', Physiol Plant 112(3), 442--450.

Majewska-Sawka, A. & Nothnagel, E. (2000), 'The multiple roles of arabinogalactan proteins in plant development', Plant Physiology 122(1), 3.

Martin, C. & Paz-Ares, J. (1997), 'MYB transcription factors in plants', Trends in Genetics 13(2), 67--73.

Nacry, P.; Camilleri, C.; Courtial, B.; Caboche, M. & Bouchez, D. (1998), 'Major Chromosomal Rearrangements Induced by T-DNA Transformation in Arabidopsis', Genetics 149(2), 641--650.

150

Oxley, D. & Bacic, A. (1999), 'Structure of the glycosylphosphatidylinositol anchor of an arabinogalactan protein from Pyrus communis suspension-cultured cells', Proceedings of the National Academy of Sciences of the United States of America 96(25), 14246.

Park, M.; Suzuki, Y.; Chono, M.; Knox, J. & Yamaguchi, I. (2003), 'CsAGP1, a gibberellin-responsive gene from cucumber hypocotyls, encodes a classical arabinogalactan protein and is involved in stem elongation', Plant Physiol 131(3), 1450-- 1459.

Pike, L. (2003), 'Lipid rafts', Journal of Lipid Research 44, 655--667.

Pilling, J.; Willmitzer, L. & Fisahn, J. (2000), 'Expression of a Petunia inflata pectin methyl esterase in Solanum tuberosum L. enhances stem elongation and modifies cation distribution', Planta 210(3), 391--399.

Pressey, R. (1984), 'Role of pectinesterase in pH-dependent interactions between pea cell wall polymers', Plant Physiology 76(2), 547.

Qi, W.; Fong, C. & Lamport, D. (1991), 'Gum arabic glycoprotein is a twisted hairy rope: a new model based on O-galactosylhydroxyproline as the polysaccharide attachment site', Plant Physiology 96(3), 848.

Roy, S.; Jauh, G.; Hepler, P. & Lord, E. (1998), 'Effects of Yariv phenylglycoside on cell wall assembly in the lily pollen tube', Planta 204(4), 450--458.

Sakakibara, H. (2006), 'Cytokinins: activity, biosynthesis, and translocation', Annu. Rev. Plant Biol 57, 431--449.

Samaj, J.; Baluska, F. & Volkmann, D. (1998), 'Cell-specific expression of two arabinogalactan protein epitopes recognized by monoclonal antobodies JIM8 and JIM13 in maize roots', Protoplasma 204(1), 1--12.

Sardar, H. (2007), 'Molecular interactions of arabinogalactan-proteins (AGPs) in tobacco Bright Yellow-2 cultured cells and functional identification of four classical AGPs in Arabidopsis', In Biological Sciences (Athens, OH: Ohio University).

Sardar, H.; Yang, J. & Showalter, A. (2006), 'Molecular interactions of arabinogalactan proteins with cortical microtubules and F-actin in Bright Yellow-2 tobacco cultured cells', Plant Physiology 142(4), 1469.

Schultz, C.; Gilson, P.; Oxley, D.; Youl, J. & Bacic, A. (1998), 'GPI-anchors on arabinogalactan-proteins: implications for signalling in plants', Trends in Plant Science 3(11), 426--431.

151

Schultz, C.; Rumsewicz, M.; Johnson, K.; Jones, B.; Gaspar, Y. & Bacic, A. (2002), 'Using genomic resources to guide research directions. The arabinogalactan protein gene family as a test case', Plant Physiol 129(4), 1448--63.

Scott Gens, J.; Fujiki, M. & Pickard, B. (2000), 'Arabinogalactan protein and wall- associated kinase in a plasmalemmal reticulum with specialized vertices', Protoplasma 212(1), 115--134.

Seifert, G. & Roberts, K. (2007), 'The biology of arabinogalactan proteins', ANNUAL REVIEW OF PLANT BIOLOGY 58, 137.

Serpe, M. & Nothnagel, E. (1994), 'Effects of Yariv phenylglycosides on Rosa cell suspensions: Evidence for the involvement of arabinogalactan-proteins in cell proliferation', Planta 193(4), 542--550.

Shiu, S. & Bleecker, A. (2001), 'Receptor-like kinases from Arabidopsis form a monophyletic gene family related to animal receptor kinases', Proceedings of the National Academy of Sciences 98(19), 10763.

Showalter, A. (2001), 'Arabinogalactan-proteins: structure, expression and function', Cellular and Molecular Life Sciences (CMLS) 58(10), 1399--1417.

Showalter, A. (1993), 'Structure and function of plant cell wall proteins.', The Plant Cell 5(1), 9.

Showalter, A. & Rumeau, D. (1990), 'Molecular biology of plant cell wall hydroxyproline-rich glycoproteins', Organization and Assembly of Plant and Animal Extracellular Matrix, 247--281.

Sivaguru, M.; Ezaki, B.; He, Z.; Tong, H.; Osawa, H.; Baluska, F.; Volkmann, D. & Matsumoto, H. (2003), 'Aluminum-induced gene expression and protein localization of a cell wall-associated receptor kinase in Arabidopsis', Plant Physiology 132(4), 2256.

Sommer-Knudsen, J.; Bacic, A. & Clarke, A. (1998), 'Hydroxyproline-rich plant glycoproteins', Phytochemistry 47(4), 483--497.

Stacey, N.; Roberts, K. & Knox, J. (1990), 'Patterns of expression of the JIM4 arabinogalactan-protein epitope in cell cultures and during somatic embryogenesis in Daucus carota L.', Planta 180(2), 285--292.

Steele, N. & Fry, S. (1999), 'Purification of xyloglucan endotransglycosylases (XETs): a generally applicable and simple method based on reversible formation of an enzyme- substrate complex', Biochem. J 340, 207--211.

152

Sun, W.; Kieliszewski, M. & Showalter, A. (2004b), 'Overexpression of tomato LeAGP- 1 arabinogalactan-protein promotes lateral branching and hampers reproductive development', The Plant Journal 40(6), 870--881.

Sun, W.; Xu, J.; Yang, J.; Kieliszewski, M. & Showalter, A. (2005), 'The lysine-rich arabinogalactan-protein subfamily in Arabidopsis: gene expression, glycoprotein purification and biochemical characterization', Plant and Cell Physiology 46(6), 975-- 984.

Sun, W.; Zhao, Z.; Hare, M.; Kieliszewski, M. & Showalter, A. (2004a), 'Tomato LeAGP-1 is a plasma membrane-bound, glycosylphosphatidylinositol-anchored arabinogalactan-protein', Physiologia Plantarum 120(2), 319--327.

Svetek, J.; Yadav, M. & Nothnagel, E. (1999), 'Presence of a glycosylphosphatidylinositol lipid anchor on rose arabinogalactan proteins', Journal of Biological Chemistry 274(21), 14724--14733.

Takahashi, T.; Gasch, A.; Nishizawa, N. & Chua, N. (1995), 'The DIMINUTO gene of Arabidopsis is involved in regulating cell elongation.', Genes & Development 9(1), 97.

Takatsuji, H. (1998), 'Zinc-finger transcription factors in plants', Cellular and Molecular Life Sciences (CMLS) 54(6), 582--596.

Thompson, H. & Knox, J. (1998), 'Stage-specific responses of embryogenic carrot cell suspension cultures to arabinogalactan protein-binding beta-glucosyl Yariv reagent', Planta 205(1), 32--38.

Toledo-Ortiz, G.; Huq, E. & Quail, P. (2003), 'The Arabidopsis basic/helix-loop-helix transcription factor family', The Plant Cell Online 15(8), 1749.

Udenfriend, S. & Kodukula, K. (1995), 'How glycosyl-phosphatidylinositol-anchored membrane proteins are made', Annual Reviews in Biochemistry 64(1), 563--591.

Valent, B. & Albersheim, P. (1974), 'The structure of plant cell walls. V. On the binding of xyloglucan to cellulose fibers', Plant Physiol 54(1), 105--108.

Varner, J. & Lin, L. (1989), 'Plant cell wall architecture.', Cell 56(2), 231--9.

Vissenberg, K.; Feijo, J.; Weisenseel, M. & Verbelen, J. (2001), 'Ion fluxes, auxin and the induction of elongation growth in Nicotiana tabacum cells', Journal of Experimental Botany 52(364), 2161--2167.

153

Vissenberg, K.; Martinez-Vilchez, I.; Verbelen, J.; Miller, J. & Fry, S. (2000), 'In vivo colocalization of xyloglucan endotransglycosylase activity and its donor substrate in the elongation zone of Arabidopsis roots', The Plant Cell Online 12, 1229--1238.

Wagner, T. & Kohorn, B. (2001), 'Wall-associated kinases are expressed throughout plant development and are required for cell expansion', The Plant Cell Online 13(2), 303.

Weigel, D. & Glazebrook, J. (2002), Arabidopsis: A Laboratory Manual, Cold Spring Harbor Laboratory Press.

Wen, F.; Zhu, Y. & Hawes, M. (1999), 'Effect of pectin methylesterase gene expression on pea root development', The Plant Cell Online 11(6), 1129.

Willats, W. & Knox, J. (1996), 'A role for arabinogalactan-proteins in plant cell expansion: evidence from studies on the interaction of beta-glucosyl Yariv reagent with seedlings of Arabidopsis thaliana', The Plant Journal 9(6), 919--925.

Yang, J. (2006), 'Functional identification of three lysine-rich arabinogalactan-proteins (AGPs) in Arabidopsis', In Biological Sciences (Athens, OH: Ohio University).

Yang, J.; Sardar, H.; McGovern, K.; Zhang, Y. & Showalter, A. (2007), 'A lysine-rich arabinogalactan protein in Arabidopsis is essential for plant growth and development, including cell division and expansion', The Plant Journal 49(4), 629--640.

Yang, J. & Showalter, A. (2007), 'Expression and localization of AtAGP18, a lysine-rich arabinogalactan-protein in Arabidopsis', Planta 226(1), 169--179.

Youl, J.; Bacic, A. & Oxley, D. (1998), 'Arabinogalactan-proteins from Nicotiana alata and Pyrus communis contain glycosylphosphatidylinositol membrane anchors', Proc Natl Acad Sci US A 95(14), 7921--7926.

Zhang, Y.; Brown, G.; Whetten, R.; Loopstra, C.; Neale, D.; Kieliszewski, M. & Sederoff, R. (2003), 'An arabinogalactan protein associated with secondary cell wall formation in differentiating xylem of loblolly pine', Plant Molecular Biology 52(1), 91-- 102.

Zhao, Z.; Tan, L.; Showalter, A.; Lamport, D. & Kieliszewski, M. (2002), 'Tomato LeAGP-1 arabinogalactan-protein purified from transgenic tobacco corroborates the Hyp contiguity hypothesis', The Plant Journal 31(4), 431--444.

154

APPENDIX A: SIGNAL STRENGTH DATA OF THE CANDIDATE GENES

Supplemental Table 1 Signal strength data of the 111 candidate genes on a single microarray chip Probe Set ID AGI Fold Signal Detection p-value Locus change 265892_at At2g15020 + 40.0 188.7 P 0.001953 260456_at At1g72490 + 17.0 24.5 A 0.303711 256634_s_at At3g28240 + 16.0 7.7 A 0.780518 265709_at At2g03540 + 13.2 56.0 A 0.246094 262113_at At1g02820 + 11.9 1283.7 P 0.000732 261996_at At1g33830 + 11.3 13.8 A 0.696289 255177_at At4g08040 + 7.0 93.3 P 0.008057 264514_at At1g09500 + 6.2 633.8 P 0.000244 264866_at At1g24140 + 5.8 75.6 A 0.219482 252411_at At3g47430 + 5.3 757.1 P 0.000244 256526_at At1g66090 + 5.0 401.2 P 0.000244 256096_at At1g13650 + 4.8 2343.8 P 0.000244 247452_at At5g62430 + 4.6 700.5 P 0.000732 251745_at At3g55980 + 4.5 3813.9 P 0.000244 245119_at At2g41640 + 4.5 779.5 P 0.000732 265668_at At2g32020 + 4.4 73.3 A 0.171387 251028_at At5g02230 + 4.2 487.1 P 0.000244 263779_at At2g46340 + 4.1 1425.5 P 0.000244 256427_at At3g11090 + 4.0 617.0 P 0.000244 260603_at At1g55960 + 3.9 2204.9 P 0.000244 255016_at At4g10120 + 3.8 2290.9 P 0.000244 249774_at At5g24150 + 3.8 2761.8 P 0.000244 251727_at At3g56290 + 3.5 2132.5 P 0.000244 267505_at At2g45560 + 3.3 890.7 P 0.000244 251658_at At3g57020 + 3.3 1522.2 P 0.000244 257485_at At1g63580 + 3.1 74.7 P 0.023926 249942_at At5g22300 + 3.1 551.3 P 0.002930 265184_at At1g23710 + 3.0 458.6 P 0.000244 267138_s_at At2g38210 + 3.0 5059.7 P 0.000244 263796_at At2g24540 + 2.96 260.8 P 0.014160 246998_at At5g67370 + 2.94 1655.2 P 0.000244 253971_at At4g26530 + 2.93 16911.6 P 0.000244 247323_at At5g64170 + 2.92 921.8 P 0.000244 267010_at At2g39250 + 2.86 345.7 P 0.001953 258321_at At3g22840 + 2.84 112.3 P 0.004150 266097_at At2g37970 + 2.83 784.9 P 0.000244 256548_at At3g14770 + 2.75 642.1 P 0.000244 155

262526_at At1g17050 + 2.75 901.6 P 0.000244 245015_at Atcg00490 + 2.75 3408.1 P 0.000244 255734_at At1g25550 + 2.73 2910.0 P 0.000244 267516_at At2g30520 + 2.63 3109.4 P 0.000244 247694_at At5g59750 + 2.59 1406.6 P 0.000244 254447_at At4g20860 + 2.57 482.0 P 0.000244 258792_at At3g04640 + 2.52 1284.0 P 0.000732 253048_at At4g37560 + 2.45 826.9 P 0.000244 252520_at At3g46370 + 2.43 247.5 P 0.001221 264096_at At1g78995 + 2.38 1042.6 P 0.000244 245724_at At1g73390 + 2.35 540.1 P 0.000732 260309_at At1g70580 + 2.33 1688.2 P 0.000244 253174_at At4g35090 + 2.33 20488.6 P 0.000244 244999_at Atcg00190 + 2.33 595.2 P 0.001221 255694_at At4g00050 + 2.32 325.5 P 0.000244 246411_at At1g57770 + 2.31 1459.5 P 0.000244 244998_at Atcg00180 + 2.30 6019.8 P 0.000244 254247_at At4g23260 + 2.28 496.2 P 0.001221 247585_at At5g60680 + 2.28 3227.8 P 0.000244 260308_at At1g70610 + 2.27 996.8 P 0.000244 254158_at At4g24380 + 2.26 602.6 P 0.000244 258217_at At3g17990 + 2.25 1646.4 P 0.000244 262973_at At1g75600 + 2.22 1.9 A 0.943848 256336_at At1g72030 + 2.22 953.6 P 0.000244 258452_at At3g22370 + 2.21 1380.5 P 0.000732 264738_at At1g62250 + 2.20 943.5 P 0.000244 258807_at At3g04030 + 2.20 45.1 A 0.129639 260155_at At1g52870 + 2.17 6703.6 P 0.000244 253043_at At4g37540 + 2.17 1149.8 P 0.000244 250429_at At5g10470 + 2.16 1150.1 P 0.000244 254691_at At4g17840 + 2.13 1702.9 P 0.000244 246540_at At5g15600 + 2.13 161.6 P 0.004150 266572_at At2g23840 + 2.11 1096.7 P 0.000244 259418_at At1g02390 + 2.10 279.7 P 0.000732 258025_at At3g19480 + 2.09 1248.8 P 0.000244 255764_at At1g16720 + 2.05 6603.2 P 0.000244 246313_at At1g31920 + 2.04 466.1 P 0.000244 250399_at At5g10750 + 2.03 547.4 P 0.000244 256541_at At1g42540 + 2.02 414.4 P 0.001953 251237_at At3g62420 + 2.02 4730.1 P 0.000244 262878_at At1g64770 + 2.01 3653.8 P 0.000244 253425_at At4g32190 + 2.00 1117.2 P 0.000244 250002_at At5g18690 - 2.00 69.3 P 0.037598 156

262978_at At1g75780 - 2.10 250.7 P 0.000244 255856_at At1g66940 - 2.10 875.9 P 0.000732 259794_at At1g64330 - 2.10 138.3 P 0.000244 248309_at At5g52540 - 2.10 440.3 P 0.000244 248276_at At5g53550 - 2.20 767.6 P 0.000244 260264_at At1g68500 - 2.40 194.7 P 0.000244 257858_at At3g12920 - 2.40 432.8 P 0.000732 258369_at At3g14310 - 2.40 3065.1 P 0.000244 247866_at At5g57550 - 2.50 168.6 P 0.000244 264987_at At1g27030 - 2.60 253.1 P 0.000244 263498_at At2g42610 - 2.60 212.0 P 0.004150 263207_at At1g10550 - 2.80 71.2 M 0.056152 260211_at At1g74440 - 2.80 116.5 P 0.014160 255148_at At4g08470 - 2.90 342.9 P 0.000732 245074_at At2g23200 - 3.0 120.5 P 0.000244 258332_at At3g16180 - 3.1 361.8 P 0.000244 254915_s_at At4g11310 - 3.2 219.4 P 0.000244 260799_at At1g78270 - 3.3 172.5 P 0.023926 259789_at At1g29395 - 3.3 694.9 P 0.000244 247794_at At5g58670 - 3.4 301.6 P 0.000732 257636_at At3g26200 - 3.5 85.9 A 0.067627 252537_at At3g45710 - 3.8 7.9 A 0.725830 249867_at At5g23020 - 3.8 98.2 P 0.001953 258100_at At3g23550 - 3.9 98.1 A 0.095215 259560_at At1g21270 - 4.3 263.3 P 0.000244 250640_at At5g07150 - 4.9 7.7 A 0.567627 257119_at At3g20190 - 5.3 3.8 A 0.805420 264214_s_at At1g65330 - 7.4 5.3 A 0.953857 245275_at At4g15210 - 10.6 4.8 A 0.828613 264346_at At1g12010 - 11.4 11.5 A 0.567627 262274_at At1g68720 - 47.5 10.3 A 0.828613 P present, A absent, M marginal 157

APPENDIX B: SEQUENCE INFORMATION FOR OVEREXPRESSION

CONSTRUCTS a pMDC45KpnI pMDC45PacI

b Construct 35S-AtAGP17

35s-AtAGP17---AGTTCATTTCATTTGGAGAGGACCTCGACTCTAGAGGATCCCCGGGTACCCCTCTCA pMDC45KpnI----NNNNNNNNNNNTTTGGAGANG-NNNCGACTCTAGAGGANCCCCGGGTACCCCTCTCA pMDC45PacI------

ACTAATTACAAATTATGACTCGCAATATTCTCTTGACGGTTACATTGATCTGCATTGTTTTCATCACCGTC ACTAATTACAAATTATGACTCGCAATATTCTCTTGACGGTTACATTGATCTGCATTGTTTTCATCACCGTC -----NNNNANNNTNNGACTCGCAANNTCNN--TGACNNTNN-ATTNNTNNNNNNNNNNNNNNNNNNNG-N

GGTGGCCAATCTCCGGCCACCGCACCGATCCATTCTCCTTCTACATCTCCTCATAAGCCTAAACCCACATC GGTGGCCAATCTCCGGCCACCGCACCGATCCATTCTCCTTCTACATCTCCTCATAAGCCTAAACCCACATC GGNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNTNNNTCTACATCTCNTCANAANCCT-NNCCCACATC

TCCCGCCATTTCTCCAGCTGCTCCAACGCCGGAATCCACGGAGGCTCCGGCGAAGACTCCGGTAGAAGCAC TCCCGCCATTTCTCCAGCTGCTCCAACGCCGGAATCCACGGAGGCTCCGGCGAAGACTCCGGTAGAAGCAC TCCCGCCATTTCTCCAGCTGNTCCAACGCNGGAATCCACGGAGGNTCCGGCGAAGACTCCGGTAGAAGCAC

CTGTAGAGGCTCCTCCTTCTCCTACACCTGCTTCTACACCTCAGATCTCTCCGCCCGCTCCTTCACCGGAG CTGTAGAGGCTCCTCCTTCTCCTACACCTGCTTCTACACCTCAGATCTCTCCGCCCGCTCCTTCACCGGAG CTGTAGAGGCTCCTCCTTCTCCTACACCTGCTTCTACACCTCAGATCTCTCCGCCCGCTCCTTCACCGGAG

GCTGATACTCCTTCAGCTCCGGAGATTGCACCTTCAGCTGATGTTCCGGCTCCAGCTCTGACAAAACATAA GCTGATACTCCTTCAGCTCCGGAGATTGCACCTTCAGCTGATGTTCCGGCTCCAGCTCTGACAAAACATAA GCTGATACTCCTTCAGCTCCGGAGATTGCACCTTCAGCTGATGTTCCGGCTCCAGCTCTGACAAAACATAA

GAAGAAGACAAAGAAGCATAAGACCGCTCCGGCTCCTGGACCAGCTTCGGAGCTTTTAAGCCCGCCTGCTC GAAGAAGACAAAGAAGCATAAGACCGCTCCGGCTCCTGGACCAGCTTCGGAGCTTTTAAGCCCGCCTGCTC GAAGAAGACAAAGAAGCATAAGACCGCTCCGGCTCCTGGACCAGCTTCGGAGCTTTTAAGCCCGCCTGCTC

CGCCCGGAGAAGCTCCTGGTCCTGGACCAAGCGATGCTTTCTCTCCTGCCGCTGACGATCAGGTAAATAAT CGCCCGGAGAAGCTCCTGGTCCTGGACCAAGCGATGCTTTCTCTCCTGCCGCTGACGATCAGGTAAATAAT CGCCCGGAGAAGCTCCTGGTCCTGGACCAAGCGATGCTTTCTCTCCTGCCGCTGACGATCAGGTAAATAAT

ATATGATTTAACCTAAACAATATGGTAAATTAATTTAGGTAATATGTAAAAATATCACTTAAGTAAAAACA ATATGATTTAACCTAAACAATATGGTAAATTAATTTAGGTAATATGTAAAAATATCACTTAAGTAAAAACA ATATGATTTAACCTAAACAATATGGTAAATTAATTTAGGTAATATGTAAAAATATCACTTAAGTAAAAACA

AATGTAAATTAGTCCCATCTTTTGACTATAGTTTAGGTTATTCGTTATAATTTAAAAATGTCTGAAATCAC AATGTAAATTAGTCCCATCTTTTGACTATAGTTTAGGTTATTCGTTATAATTTAAAAATGTCTGAAATCAC AATGTAAATTAGTCCCATCTTTTGACTATAGTTTAGGTTATTCGTTATAATTTAAAAATGTCTGAAATCAC

158

AATTTTTTTGTTAGATTTGATTCTTTCTAGTTTCTAGAACTTAAAATTTCGAATTATTTTACTGCAGAGCG AATTTTTTTGTTAGATTTGATTCTTTCTAGTTTCTAGAACTTAAAATTTCGAATTATTTTACTGCAGAGCG AATTTTTTTGTTAGATTTGATTCTTTCTAGTTTCTAGAACTTAAAATTTCGAATTATTTTACTGCAGAGCG

GAGCACAAAGAATAAGTGTTGTGATACAAATGGTGGGAGCTGCGGCAATCGCATGGTCTCTACTTGTTCTA GAGCACAAAGAATAAGTGTTGTGATACAAATGGTGGGAGCTGCGGCAATCGCATGGTCTCTACTTGTTCTA GAGCACAAAGAATAAGTGTTGTGATACAAATGGTGGGAGCTGCGGCAATCGCATGGTCTCTACTTGTTCTA

GCCTTCTAATAACCTTAATTAACTAGTTCTAGAGCGGCCGCCACCGCGGTGGAGCTCGAATTTCCCCGATC GCCTTCTAATAACCTTAATTAACTAGTTCTAGAGCGGCCGCCACCGCGGTGGAGCTCGAATTTCCCCGATC GCCTTCTAATAACCTTAATTAACTAGTTCTAGAGCGGCCGCCACCGCGGTGGAGCTCGANNNNN-CCGATC

GTTCAAACATTTGGCAATAAAGTTTCTTAAGATTGAAT------35s-AtAGP17 GTTCAANNNTTTGGNANNNN--TT--NNAN—-TNNNNN------pMDC45KpnI GT-CAAACNNN-GNNNNN------pMDC45PacI

Construct 35S-AtAGP18

35s-AtAGP18----TTTGGAGAGGACCTCGACTCTAGAGGATCCCCGGGTACCCCTCCAAATTTTAACAA pMDC45KpnI-----NNNNNNNNNNANCTCGACTCTAGA-NNNNNCCGGGTACCCCTCCAAATTTTAACAA pMDC45PacI------

AATTATGGATCGCAATTTCCTCCTAACAGTTACATTGATCTGCATTGTCGTCGCCGGTGTCGGTGGCCAAT AATTATGGATCGCAATTTCCTCCTAACAGTTACATTGATCTGCATTGTCGTCGCCGGTGTCGGTGGCCAAT ------NNNNNNNTNNNNN

CTCCTATCTCTTCTCCGACCAAATCTCCCACCACTCCTTCTGCTCCAACTACTTCCCCTACTAAATCCCCC CTCCTATCTCTTCTCCGACCAAATCTCCCACCACTCCTTCTGCTCCAACTACTTCCCCTACTAAATCCCCC NTTTTTTTTNNNNNAANNNNNNNNNNCCCCCCCCCCNNNNNNNTNNAACTACTTCCCCTACTAANTCCCCC

GCCGTTACTTCTCCCACTACGGCTCCGGCGAAAACTCCAACTGCTTCGGCTTCTTCACCGGTTGAGTCACC GCCGTTACTTCTCCCACTACGGCTCCGGCGAAAACTCCAACTGCTTCGGCTTCTTCACCGGTTGAGTCACC GCCGTTACTTCTCCCACTACGGCTCCGGCGAAAACTCCAACTGCTTCGGCTTCTTCACCGGTTGAGTCACC

AAAATCTCCGGCTCCTGTTAGCGAGTCGTCTCCACCACCGACACCTGTTCCAGAGAGCTCTCCTCCGGTTC AAAATCTCCGGCTCCTGTTAGCGAGTCGTCTCCACCACCGACACCTGTTCCAGAGAGCTCTCCTCCGGTTC AAAATCTCCGGCTCCTGTTAGCGAGTCGTCTCCACCACCGACACCTGTTCCAGAGAGCTCTCCTCCGGTTC

CTGCACCAATGGTTTCTTCTCCAGTGAGCTCTCCACCGGTTCCAGCACCAGTAGCTGATTCTCCTCCAGCT CTGCACCAATGGTTTCTTCTCCAGTGAGCTCTCCACCGGTTCCAGCACCAGTAGCTGATTCTCCTCCAGCT CTGCACCAATGGTTTCTTCTCCAGTGAGCTCTCCACCGGTTCCAGCACCAGTAGCTGATTCTCCTCCAGCT

CCGGTAGCCGCTCCGGTTGCTGATGTACCGGCTCCTGCTCCAAGCAAGCATAAGAAGACTACAAAGAAATC CCGGTAGCCGCTCCGGTTGCTGATGTACCGGCTCCTGCTCCAAGCAAGCATAAGAAGACTACAAAGAAATC CCGGTAGCCGCTCCGGTTGCTGATGTACCGGCTCCTGCTCCAAGCAAGCATAAGAAGACTACAAAGAAATC

GAAAAAGCATCAAGCTGCACCTGCTCCGGCTCCGGAACTTCTCGGTCCACCTGCACCACCGACTGAATCTC GAAAAAGCATCAAGCTGCACCTGCTCCGGCTCCGGAACTTCTCGGTCCACCTGCACCACCGACTGAATCTC GAAAAAGCATCAAGCTGCACCTGCTCCGGCTCCGGAACTTCTCGGTCCACCTGCACCACCGACTGAATCTC

CCGGACCTAACTCCGACGCTTTTTCTCCCGGTCCTTCCGCCGACGATCAGGTACATTAGCTTATATCAACA CCGGACCTAACTCCGACGCTTTTTCTCCCGGTCCTTCCGCCGACGATCAGGTACATTAGCTTATATCAACA CCGGACCTAACTCCGACGCTTTTTCTCCCGGTCCTTCCGCCGACGATCAGGTACATTAGCTTATATCAACA 159

ATGTTGATATTTATCTACCGTCACAAGTTCGAGTATCAATGTTAAAAATAACCAATGAGACATTTCAAAAC NNNNNNATATTTATCTACCGTCACAAGTTCGAGTATCAATGTTAAAAATAACCAATGAGACATTTCAAAAC ATGTTGATATTTATCTACCGTCACAAGTTCGAGTATCAATGTTAAAAATAACCAATGAGACATTTCAAAAC

AGTGAACATAGATTTTCAATTTTTGTTATATTGATTTAGTAATTACATATTCGATTAACTAAAATGAAAAT AGTGAACATAGATTTTCAATTTTTGTTATATTGATTTAGTAATTACATATTCGATTAACTAAAATGAAAAT AGTGAACATAGATTTTCAATTTTTGTTATATTGATTTAGTAATTACATATTCGATTAACTAAAATGAAAAT

ATGGAATTGAATGCAGAGCGGAGCAGCGAGCACAAGGGTGTTGAGGAATGTAGCGGTGGGAGCGGTTGCAA ATGGAATTGAATGCAGAGCGGANCAGCGAGCACNANGGTGTTGAGGAATGTAGCGGTGGGAGCGGTTGCAA ATGGAATTGAATGCAGAGCGGAGCAGCGAGCACAAGGGTGTTGAGGAATGTAGCGGTGGGAGCGGTTGCAA

CCGCATGGGCCGTTCTCGTTATGGCATTCTAACCTTAATTAACTAGTTCTAGAGCGGCCGCCACCGCGGTG CCGCATGGGCCGTTCTCGTTATGGCATTCTAACCTTAATTAACTAGTTCTAGAGCGGCNGCCACCGCGGNG CCGCATGGGCCGTTCTCGTTATGGCATTCTAACCTTAATTAACTAGTTCTAGAGCGGCCGCCACCGCGGTG

GAGCTCGAATTTCCCCGATCGTTCAAACATTTGGCAATAAAGTTTCTTAAGATTG-----35s-AtAGP18 GAGCTCGAATTTCCCCGATCGTTCAANCNTTTGGCAN—AANTNNNNNNNNNNNNN------pMDC45KpnI GAGCTCGNNNNNN-CCGATCGTC-AAACNNNNNNNNN------pMDC45PacI

Construct 35S-AtAGP19

35s-AtAGP19—---TTCATTTCATTTGGAGAGGACCTCGACTCTAGAGGATCCCCGGGTACCTAGCTTCC pMDC45KpnI-----NNNNNNNNNNNNNNNGANGANN-CGACTCTAGAGGANCCCCGGGTACCTAGCTTCC pMDC45PacI------

TCCACAACACAATGGAATCAAATTCAATCATTTGGTCTCTTCTTTTGGCTTCTGCTCTCATCTCTTCCTTT TCCACAACACAATGGAATCAAATTCAATCATTTGGTCTCTTCTTTTGGCTTCTGCTCTCATCTCTTCCTTT ------

AGTGTAAATGCACAAGGACCTGCTGCTTCACCAGTAACCTCCACAACCACTGCTCCGCCTCCAACAACTGC AGTGTAAATGCACAAGGACCTGCTGCTTCACCAGTAACCTCCACAACCACTGCTCCGCCTCCAACAACTGC ------

TGCTCCTCCAACCACTGCTGCTCCGCCTCCAACTACCACTACTCCACCCGTTTCAGCTGCGCAGCCACCAG TGCTCCTCCAACCACTGCTGCTCCGCCTCCAACTACCACTACTCCACCCGTTTCAGCTGCGCAGCCACCAG ------

CATCTCCGGTTACACCTCCACCAGCAGTTACTCCAACTTCACCACCAGCTCCAAAAGTTGCACCAGTAATC CATCTCCGGTTACACCTCCACCAGCAGTTACTCCAACTTCACCACCAGCTCCAAAAGTTGCACCAGTAATC ------

AGCCCCGCAACTCCGCCTCCACAACCACCACAAAGCCCGCCCGCTTCAGCTCCAACCGTCTCACCACCACC AGCCCCGCAACTCCGCCTCCACAACCACCACAAAGCCCGCCCGCTTCAGCTCCAACCGTCTCACCACCACC NNNN--GNNNNNNNNCCTNNNNAACCANCNNAAAGCCCGCCNNNTTCAGNTCCAACNGTNTCACCACCACC

TGTATCACCACCACCAGCACCGACGTCTCCCCCACCCACACCAGCTTCACCACCACCTGCACCAGCTTCTC TGTATCACCACCACCAGCACCGACGTCTCCCCCACCCACACCAGCTTCACCACCACCTGCACCAGCTTCTC TGTATCACCACCACCAGCACCGACGTCTCCCCCACCCACACCAGCTTCACCACCACCTGCACCAGCTTCTC

CACCTCCTGCACCAGCTTCACCGCCACCTGCACCAGTTTCCCCACCCCCTGTACAAGCGCCATCACCAATA CACCTCCTGCACCAGCTTCACCGCCACCTGCACCAGTTTCCCCACCCCCTGTACAAGCGCCATCACCAATA 160

CACCTCCTGCACCAGCTTCACCGCCACCTGCACCAGTTTCCCCACCCCCTGTACAAGCGCCATCACCAATA

AGTTTACCACCAGCTCCAGCACCGGCTCCTACCAAGCACAAGAGAAAGCACAAACACAAAAGGCATCACCA AGTTTACCACCAGCTCCAGCACCGGCTCCTACCAAGCACAAGAGAAAGCACAAACACAAAAGGCATCACCA AGTTTACCACCAGCTCCAGCACCGGCTCCTACCAAGCACAAGAGAAAGCACAAACACAAAAGGCATCACCA

TGCCCCAGCTCCAGCACCAATTCCCCCAAGCCCTCCATCTCCTCCAGTTCTAACAGATCCACAGGACACGG TGCCCCAGCTCCAGCACCAATTCCCCCAAGCCCTCCATCTCCTCCAGTTCTAACAGATCCACAGGACACGG TGCCCCAGCTCCAGCACCAATTCCCCCAAGCCCTCCATCTCCTCCAGTTCTAACAGATCCACAGGACACGG

CTCCAGCACCATCACCAAACACGGTAACAATTCAATACTAAACAGCAATCTCCCCAATCACTGCCATGCTT CTCCNGCACCATCACCAAACACGGTAACAATTCAATACTAAACAGCAATCTCCCCAATCACTGCCATGCTT CTCCAGCACCATCACCAAACACGGTAACAATTCAATACTAAACAGCAATCTCCCCAATCACTGCCATGCTT

ATATGATTCTAACCTTCTTTTATTTTTCTGGTTTACAGAATGGAGGAAATGCCTTAAATCAGCTTAAAGGA ATATGATTCTAACCTTCTTTTATTTTTCTGGNTTACAGAATGGAGGAAATGNCTTAAATCAGCTTAAAGGA ATATGATTCTAACCTTCTTTTATTTTTCTGGTTTACAGAATGGAGGAAATGCCTTAAATCAGCTTAAAGGA

AGAGCAGTAATGTGGCTCAATACTGGACTAGTAATCCTCTTTCTACTTGCTATGACAGCCTAACCTTAATT ANAGCAGTAATGNGGCTCNATACTGGACTAGTAATCCTCTTTCTACTTGCTATGANAGCCTAACCNTAATT AGAGCAGTAATGTGGCTCAATACTGGACTAGTAATCCTCTTTCTACTTGCTATGACAGCCTAACCTTAATT

AACTAGTTCTAGAGCGGCCGCCACCGCGGTGGAGCTCGAATTTCCCCGATCGT------35s-AtAGP19 AACTAGTTCTAGAGNGGCCGNNNNNNCNGNGGANCTCGAN-TTCCCCNATCGT------pMDC45KpnI AACTAGTTCTAGAGCGGCCGCCACCGCGGTGGAGCTCGANNNN-CCCGNTCGT------pMDC45PacI

Supplemental Figure 1 Sequence information for constructs 35S-AtAGP17/18/19 a) Schematic illustration of the location of sequencing primers. Arrows show forward sequencing primer (pMDC45KpnI) and reverse sequencing primer (pMDC45PacI). The constructs are not drawn in scale. b) Sequencing results of constructs 35S-AtAGP17/18/19. Sequencing results of forward sequencing primer (pMDC45KpnI) and reverse sequencing primer (pMDC45PacI) were aligned with the predicted sequences. Start codons and stop codons were shown in bold letters.