The Role of the Salvage Pathway in Nucleotide Sugar Biosynthesis

Home , Erythritol kinase, Fucokinase, GHMP kinase family, Glucuronokinase, Guanylyltransferase, Homoserine kinase

THE ROLE OF THE SALVAGE PATHWAY IN NUCLEOTIDE SUGAR BIOSYNTHESIS:

IDENTIFICATION OF SUGAR KINASES AND NDP-SUGAR PYROPHOSPHORYLASES

TING YANG

(Under the Direction of Maor Bar-Peled)

ABSTRACT

The synthesis of polysaccharides, glycoproteins, glycolipids, glycosylated secondary metabolites and hormones requires a large number of glycosyltransferases and a constant supply of nucleotide sugars. In plants, photosynthesis and the NDP-sugar inter-conversion pathway are the major entry points to form NDP-sugars. In addition to these pathways is the salvage pathway, a less understood metabolism that provides the flux of NDP-sugars. This latter pathway involves the hydrolysis of glycans to free sugars, sugar transport, sugar phosphorylation and nucleotidylation. The balance between glycan synthesis and recycling as well as its regulation at various plant developmental stages remains elusive as many of the molecular components are unknown. To understand how the salvage pathway contributes to the sugar flux and cell wall biosynthesis, my research focused on the functional identification of salvage pathway sugar kinases and NDP-sugar pyrophosphorylases. This research led to the first identification and enzymatic characterization of galacturonic acid kinase (GalA kinase), galactokinase (GalK), a broad UDP-sugar pyrophosphorylase (sloppy), two promiscuous UDP-GlcNAc pyrophosphorylases (GlcNAc-1-P uridylyltransferases), as well as UDP-sugar pyrophosphorylase paralogs from Trypanosoma cruzi and Leishmania major. To evaluate the salvage pathway in plant biology, we further investigated a sugar kinase mutant: galacturonic acid kinase mutant (galak) and determined if and how galak KO mutant affects the synthesis of glycans in Arabidopsis. Feeding galacturonic acid to the seedlings exhibited a 40-fold accumulation of free GalA in galak mutant, while the wild type (WT) plant readily metabolizes the fed-sugar. These findings suggest that in vivo, the GalAK gene product functions to salvage GalA in Arabidopsis. Interestingly, the galak mutant showed no visible morphological phenotype compared to WT, and the cell wall sugar composition was not affected in the mutant. Immunohistochemical analysis of galak indicated no glycan structural changes in galak. The information gained indicates that the gene encoding GalAK is required for proper recycling of GalA in plant cell. However, knocking out GalAK is not detrimental for plant cell wall synthesis, plant growth and development.

INDEX WORDS: Nucleotide sugar, Salvage pathway, Sugar kinase, Nucleotide sugar pyrophosphorylase, Galacturonic acid kinase, Galactokinase, GlcNAc-1-P uridylyltransferase,

UDP-GlcNAc pyrophosphorylase, Sloppy, Arabidopsis thaliana

THE ROLE OF THE SALVAGE PATHWAY IN NUCLEOTIDE SUGAR BIOSYNTHESIS:

IDENTIFICATION OF SUGAR KINASES AND NDP-SUGAR PYROPHOSPHORYLASES

TING YANG

B.S., Jilin University, China, 2006

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2011

TING YANG

THE ROLE OF THE SALVAGE PATHWAY IN NUCLEOTIDE SUGAR BIOSYNTHESIS:

IDENTIFICATION OF SUGAR KINASES AND NDP-SUGAR PYROPHOSPHORYLASES

TING YANG

Major Professor: Maor Bar-Peled Committee: Kelley Moremen Michael Tiemeyer Zachary Wood

Electronic Version Approved:

Maureen Grasso Dean of the Graduate School The University of Georgia August 2011

DEDICATION

I dedicate this work to my parents, for their love and encouragement; and to my husband,

Tianhao He, for his caring support in my life.

ACKNOWLEDGEMENTS

I am deeply thankful to my major professor, Dr. Maor Bar-Peled, for his endless guidance and enthusiasm motivated me to the research. I would like to give my many thanks to all my committee members: Dr. Michael Tiemeyer, Dr. Kelley Moremen and Dr. Zachary Wood, for their support and valuable advice in my research.

I am grateful for the help I received from Dr. John Glushka, for his guidance and assistance with the NMR experiments, I have learned much from him; I would also like to acknowledge Dr. Utku Avci, for his help with the immunohistochemistry assays; Dr. Yanbin

Yin, for his help with the phylogeny analysis; Dr. Siva Kuma, for his help with the glycome profiling; I would also like to thank Dr. Malcolm O‟Neill for his assistance with the glycan analysis. In addition, I would like to thank all the present and former lab members who have shared their ideas with me, especially Dr. Xiaogang Gu, Dr. Yingnan Jiang and James Amor

Smith, for their encouragement and great help in the research.

I would like to show my gratitude to many undergraduate students who have worked with me: Andy Martin, Sung G Lee, Lindsey Gebhart and Ben Mullenbach; and to my best friends.

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... v

CHAPTER

1 INTRODUCTION AND LITERATURE REVIEW ...... 1

PART I: NUCLEOTIDE SUGAR BIOSYNTHETIC SALVAGE PATHWAY ....1

PART II: THE EFFECTS OF MUTANTS IN THE SALVAGE PATHWAY ON

PLANT DEVELOPMENT ...... 31

PART III: COMPARATIVE ANALYSIS OF PROTEIN DOMAINS OF SUGAR

KINASES AND NUCLEOTIDE SUGAR PYROPHOSPHORYLASES IN

PLANTS ...... 36

PART IV: THESIS OVERVIEW ...... 54

2 IDENTIFICATION OF GALACTURONIC ACID-1-PHOSPHATE KINASE, A

NEW MEMBER OF THE GHMP KINASE SUPERFAMILY IN PLANTS, AND

COMPARISON WITH GALACTOSE-1-PHOSPHATE KINASE ...... 55

ABSTRACT ...... 56

INTRODUCTION ...... 57

EXPERIMENTAL PROCEDURES ...... 58

RESULTS ...... 66

DISCUSSION ...... 91

3 IDENTIFICATION OF A NOVEL UDP-SUGAR PYROPHOSPHORYLASE WITH

A BROAD SUBSTRATE SPECIFICITY IN TRYPANOSOMA CRUZI ...... 96

ABSTRACT ...... 97

INTRODUCTION ...... 98

EXPERIMENTAL PROCEDURES ...... 100

RESULTS ...... 105

DISCUSSION ...... 127

4 IDENTIFICATION AND CHARACTERIZATION OF A STRICT AND OF A

PROMISCUOUS N-ACETYLGLUCOSAMINE-1-P

URIDYLYLTRANSFERASES IN ARABIDOPSIS ...... 132

ABSTRACT ...... 133

INTRODUCTION ...... 134

EXPERIMENTAL PROCEDURES ...... 138

RESULTS ...... 143

DISCUSSION ...... 169

5 THE ROLE OF GALAK IN GLYCAN METABOLISM ...... 173

ABSTRACT ...... 174

INTRODUCTION ...... 175

EXPERIMENTAL PROCEDURES ...... 176

RESULTS ...... 184

DISCUSSION ...... 202

6 CONCLUSIONS...... 205

REFERENCES ...... 207

vii

APPENDICES

A SUPPLEMENTAL FIGURES AND TABLES ...... 232

B SUMMARY OF PRIMERS, ANTIBODIES, CLONING CONSTRUCTS AND

TRANSGENIC PLANTS ...... 260

viii

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

PART I: NUCLEOTIDE SUGAR BIOSYNTHETIC SALVAGE PATHWAY

Introduction - Multiple routes for the synthesis of NDP-sugars

Nucleotide sugars (Fig. 1.1, abbreviated NDP-sugars) are the substrates for the synthesis of polysaccharides, glycoproteins, glycolipids, glycoside-linked secondary metabolites and hormones. The biosynthesis of nucleotide sugars in plants is complicated and can be achieved via several pathways (Fig. 1.2): the inter-conversion pathway converts pre-existing NDP-sugars to other NDP-sugars (Reiter and Vanzin, 2001; Mohnen, 2002; Bar-Peled and O'Neill, 2010) via the selective activities of dehydrogenases, epimerases, reductases, decarboxylases, and dehydratases. The salvage pathway (Fig. 1.3 and 1.4) converts free sugars, presumably hydrolyzed from glycans, to sugar-1-phosphates by sugar kinases and subsequently into NDP- sugars by NDP-sugar pyrophosphorylases (Mohnen et al., 2008; Bar-Peled and O'Neill, 2010).

Some NDP-sugars, such as UDP-Glc can also be obtained through the conversion of sucrose by sucrose synthase (Koch, 2004). Other pathways contributing to the flux of NDP-sugars are the myo-inositol oxidation pathway, that can form UDP-GlcA (Fig. 1.3) (Young et al., 1966); and the formation of Glc-1-P, which can be obtained by a starch phosphorylase (Rathore et al., 2009) via phosphorylation of soluble heteroglycans (SHGs) (Lu et al., 2006). Lastly, a possibility that

cannot be excluded is the NDP-sugar pool can be depleted temporarily by the reverse reaction of the sugar PPases converting NDP-sugars to form sugar-1-Ps. Thus PPases could also play roles in the competition or regulation of NDP-sugar formation in plant cells. The underlying mechanism that controls the flux of specific NDP-sugars to specific glycosylation reactions at specific tissues is poorly understood. Before my PhD, only one gene/enzyme was known to function in salvage pathway. It was my role to contribute to the identification of other genes/enzymes in this pathway to better understand the function and contribution of the salvage pathway in NDP-sugar biosynthesis. During my PhD, I discovered and characterized eight genes that are involved in the salvage pathway. And the following chapters include these genes as well as others that were recently discovered.

FIGURE 1.1. Representative structures of NDP-sugars.

UDP-Glc, a UDP-sugar; GDP-Fuc, a GDP-sugar; and CMP-Kdo, an NMP-sugar. (Taken from

Bar-Peled and O‟Neill, 2010).

FIGURE 1.2. The major pathways for nucleotide sugar biosynthesis.

The pathways include the inter-conversion pathway, the salvage pathway, and the myo-inositol oxidation pathway. The unsolved pathways are indicated by dashes. (Taken from Bar-Peled and

O‟Neill, 2010).

FIGURE 1.3. Alternative pathways to generate UDP-GlcA in plants: the myo-inositol oxidation pathway, the nucleotide sugar inter-conversion pathway, and the salvage pathway.

In plants, myo-inositol synthesis initiates from the cyclization of D-Glc-6-P to myo-inositol-1-P, catalyzed by myo-inositol-1-P synthase. Subsequently, the myo-inositol-1-P is dephosphorylated by myo-inositol monophosphatase to form myo-inositol. Free myo-inositol is oxidized to D-GlcA by inositol oxygenase. Through the terminal step of myo-inositol oxidation pathway, GlcA is subsequently converted to UDP-GlcA (Young et al., 1966). Crude plant enzyme activity that converts GlcA to GlcA-1-P has been described (Neufeld et al., 1959), and recently the gene encoding such activity in Arabidopsis was identified (Pieslinger et al., 2010).

FIGURE 1.4. Partial presentation of salvage pathway metabolic enzymes in plants.

After salvaged back into the cell, sugars such as Glc, Man, or Frc are phosphorylated at the C6 position by hexokinase(s) (Wilson, 1995), and can enter different metabolic routes (e.g., glycolysis pathway), or by the action of phosphohexo-mutase to transform to sugar-1-P. On the other hand, other sugars, such as Gal, Ara, GlcA, and GalA can be directly phosphorylated at the

C1 position by specific sugar kinases to form sugar-1-Ps. The sugar-1-Ps can then be converted to nucleotide sugars in the cytosol by specific NDP-sugar pyrophosphorylases. The enzymes from Arabidopsis with known functions are listed (below the arrow) with the gene loci number.

Nucleotide sugar biosynthesis salvage pathway

In the early 1960s, D-Gal kinase, L-Ara kinase, and D-GlcA kinase activities were first identified from soluble and membrane mung bean seedling preparations (Kessler et al., 1961). D-

GalA kinase activity was detected in germinating mung bean seedlings (Neufeld et al., 1961).

Feeding experiments with radioactive D-GlcA and D-GalA in vivo showed their rapid incorporation into hemicellulose and pectin (Kessler et al., 1961; Neufeld et al., 1961). In recent years the knowledge of genes involved in the salvage pathway has increased substantially with the discovery of several sugar kinases including D-Gal kinase (Kaplan et al., 1997; Yang et al.,

2009), L-Ara kinase (Sherson et al., 1999), L-Fuc kinase (Kotake et al., 2008), D-GalA kinase

(Yang et al., 2009) and D-GlcA kinase (Pieslinger et al., 2010), and the discovery of new genes involved in the subsequent pyrophosphorylation of sugar-1-Ps to NDP-sugars. Sloppy, a UDP- sugar pyrophosphorylase (USP) converting six different sugar-1-Ps to the corresponding UDP- sugars has been identified independently by several laboratories in various plant species including Arabidopsis (Litterer et al., 2006; Kotake et al., 2007; Yang et al., 2009), pea (Kotake et al., 2004), melon (Dai et al., 2006) as well as in species other than plants, for example, parasites (Yang and Bar-Peled, 2010). In addition, other sugar-1-P specific PPases were identified (see Fig. 1.4) including UDP-Glc PPase (Chen et al., 2007; McCoy et al., 2007; Meng et al., 2008; Meng et al., 2009), UDP-GlcNAc PPase (Yang et al., 2010), ADP-Glc PPase

(Kleczkowski, 2000; Jin et al., 2005; McCoy et al., 2006), GDP-Man PPase (Lukowitz et al.,

2001; Badejo et al., 2007; Qin et al., 2008), and GDP-Fuc PPase (Kotake et al., 2008). Taken together, it is clear that the salvage pathway exists in plants, yet how the salvage pathway regulates the flux of sugars for the metabolism of diverse glycans still remains unclear.

An important question that is also unclear is what type of molecules are recycled in order for the salvage pathway to exist. In addition, we do not know how often a cell salvages such molecules. The primary wall of growing plant cells is a potential source of the recycling sugars

(see some examples in Fig. 1.5). The assembly and restructuring of this wall are believed to be required for cell expansion and growth. These processes should require numerous specific hydrolytic enzymes and the resulting free sugars could then be recycled by specific sugar transporters into the cell as the substrates for the enzymes in NDP-sugar salvage pathway.

Consistent with these, plants have a large number of genes involved in carbohydrate hydrolysis as well as a large gene family with over 50 monosaccharide transporter-like genes (Buttner,

2007). The specific function of only a few hydrolytic enzymes and transporters has been determined. Clearly, functional identification of hydrolytic enzymes and transporters will provide insight to the regulation of the salvage pathway and how often specific molecules within the cell are salvaged.

FIGURE 1.5. Numerous cellular processes are required to recycle and salvage plant cell glycans.

In the extra-cellular matrix (e.g., wall), free sugars released by the hydrolytic enzymes can enter back to the cell by passive diffusion or by sugar transporters (depicted by barrel and - - arrows).

In the cell, free sugars could be generated through the degradation of starch, glycolipids, or secondary metabolites (depicted by - - arrows). Sugars may enter different pathways. Some sugars can be directly phosphorylated at C1 position by specific sugar kinases and subsequently the sugar-1-Ps may be converted to nucleotide sugars in the cytosol (black arrow). Note: due to space limitation, cellular organelles including vacuole (where many secondary metabolites can be stored) are not shown.

A growing plant cell deposits 50% of its carbon into primary cell wall glycans, and a larger amount of wall is likely deposited during secondary wall formation, especially in woody plants, and during stem development. This process requires a constant and large supply of diverse types of nucleotide sugars. Cellular control of this supply is likely complex and may be different in distinct plant tissues, for example the tissue that stores carbohydrates (seed), versus the tissue that is able to carry out photosynthesis, or carry the plant load. Another factor that is important to recognize about cellular control of NDP-sugar flux is where in the cell NDP-sugars are made (Bar-Peled and O'Neill, 2010). It was shown that the activity of NDP-sugar metabolic pathway occurs in different subcellular organelles. For example, plants use ADP-glucose in the cytosol and in chloroplasts to produce starch, while numerous glycosylated secondary metabolites are produced in the cytosol, vacuole, mitochondria and ER (Roze et al., 2011). Some

NDP-sugar metabolic enzymes are localized in Golgi while others are found in the chloroplast, at the plasma membrane or in the cytosol. Considerable glycosylation reactions take place in the

ER for glycoprotein and glycolipid synthesis and in the Golgi for polysaccharide synthesis. Thus, specific transporters exist to move nucleotide sugars from the cytoplasm into the organelles where they are consumed. Currently, few NDP-sugar transporters have been identified in plants

(Handford et al., 2006) and it is likely that the activity of these transporters can also regulate the supply and demand of NDP-sugars in the cell.

Given that no one molecule „lives for ever‟ in a cell, we assume that the cell addresses the release of free sugars generated through glycan turnover by establishing a specific sugar salvage pathway. Although major progress was made in the past few years to identify molecular components of the salvage pathway (described below), numerous genes involved in the phosphorylation of other free sugars such as L-Gal, L-Araf, 4-O-Me-GlcA, L-Rha, Xyl, apiose,

glucosamine and corresponding pyrophosphorylases that can transform these into NDP-sugars still have not been identified. Combining genetic and bioinformatics tools with biochemistry and cell biology to identify more sugar kinases and NDP-sugar pyrophosphorylases as well as characterizing their biological functions will be instrumental to understand the salvage pathway in plants. This knowledge will help to address the role of salvage pathway genes and be useful to depict how salvage pathway is involved and regulated during wall biosynthesis, as well as how sugar flux is controlled in plants and perhaps in other organisms.

In addition to the recycling of certain wall glycans, the salvage of sugars released from glycoproteins, glycolipids, and other glyco-secondary metabolites to form nucleotide sugars during wall reconstruction and degradation is quite complicated. To better understand this complex recycling process, I will provide a brief description of the major glycan structures in plants; the hydrolytic enzymes that hydrolyze glycans; the sugar metabolism inside the cell, as well as the role of salvage pathway in plant development as depicted by the analysis of mutants.

Glycans in plants

Plants contain numerous types of glycans, including polysaccharides, glycoproteins, and glycolipids. The plant cell wall is composed of a substantial amount of glycan polymers. These glycan structures are quite complex and include over 30 different sugar residues, which include sugars and sugar derivatives. Below is a brief description of the major glycans in plants.

Polysaccharides in plants - The most abundant polysaccharides in primary cell wall include cellulose, hemicellulose, and pectin:

Cellulose is a linear polymer consisting of β-1,4 linked D-glucose units. Cellulose forms crystalline microfibrils which provide most of the strength of the cell wall and make it resistant

to hydrolysis. Cellulose is synthesized in the plasma membrane by a group of cellulose synthases

(CesA) (Mutwil et al., 2008) using UDP-Glc as a sugar donor. The source of the activated sugar is controversial, mainly as cellulose synthase assays remain to be exhaustively established. Some suggest that a plasma membrane sucrose synthase, which is associated with CesA, provides the flux of UDP-Glc from Suc. Others suggest that the UDP-Glc flux is derived from a cytosolic pool.

Hemicellulose, a term applied to the polysaccharide components of plant cell walls other than cellulose, consists of a large group of diverse heteropolymers such as mannan, xyloglucan, and xylan. The composition and the amount of these polysaccharides differ across plant species, e.g., monocot and dicot. All types of hemicellulose share a common β-1,4 linked backbone of sugar residues. Mannan (mannose-containing glycan backbone) includes mannan, glucomannan, and galactomannan. They are commonly found in fruits, seeds, and in secondary plant cell walls.

Xylan (xylose-containing glycan backbone) includes arabinoxylan, glucuronoxylan, and glucuronoarabinoxylan, which are more abundant in the secondary walls of dicots and in grasses.

The hemicellulose xyloglucans (glucose-containing glycan backbone) are most abundant in the primary wall. Some of the hemicellulose may interact with cellulose and contribute to the cell wall strength. Unlike cellulose, hemicellulose polysaccharides are composed of many different sugar residues, including Gal, Man, Xyl, Ara, GlcA, Fuc, and Glc. The synthesis of hemicellulose is believed to initiate in the Golgi and requires various types of glycosyltransferases, and numerous NDP-sugars as substrates. Newly Golgi-synthesized hemicellulose polymers are transported across the plasma membrane and assembled in the cell wall in a process yet to be determined. The physical properties of this large group of polysaccharides vary but also depend on sugar derivatives. For example, glucomannan is a

soluble polymer but when acetylated it becomes insoluble. Some of these glycans, in secondary walls, are cross-linked to other molecules such as lignin that strengthen the cell wall.

Pectic polysaccharide is predominantly found in the primary cell wall and contributes to the cell wall porosity, rigidity, and adhesion (Willats et al., 2001; Mohnen, 2008; Caffall and

Mohnen, 2009). Five different types of pectic polysaccharides are identified in the plant cell wall: homogalacturonan, rhamnogalacturonan I, rhamnogalacturonan II, xylogalacturonan and apiogalacturonan (Ridley et al., 2001) (Fig. 1.6 A-E).

Homogalacturonan (HGA) is the most abundant pectic polysaccharide and constitutes about 65% of the pectin in the primary cell wall. It is predominant in monocot and dicot plants with the structure of a linear polymer of α-1,4-D-GalA residues. HGA could be partially methylesterified at the C-6 carboxyl group, or O-acetylated at O-2 or O-3 (O'Neill and York,

2003). When C-3 of HGA backbone is substituted by β-D-xylose residues, xylogalacturonan

(XGA) is formed. Similarly, apiogalacturonans (AP) is formed by attaching β-D-apiose residue to

C-2 or C-3 of HGA backbone.

Rhamnogalacturonan I (RG-I) consists of a heteropolymer with repeating disaccharide units of α-1,2-L-Rha-α-1,4-D-GalA, comprising about 20-35% of pectin. The rhamnose residues in RG-I provide sites for further glycosylation. They may be substituted by different side chains at C-4, such as arabinan, galactan epitopes, as well as single sugar residues, forming a very complex polysaccharide family (O'Neill and York, 2003). RG-I may be covalently linked with

HGA in the cell wall matrix.

Rhamnogalacturonan II (RG-II) is present in higher plants. It is the most structurally complex pectin, comprising about 5% of pectin. RG-II has an α-1,4-linked D-GalA backbone that is substituted with four structurally distinct side chains: two oligosaccharide chains attached to

O-2 of the backbone and two disaccharide chains attached to the backbone through O-3 linkage

(O'Neill et al., 2004). The four side chains contain 12 different types of sugars, including some rare sugar residues such as D-apiose, aceric acid, Kdo, and Dha with more than 20 different linkages. RG-II usually exists as dimers: two RG-II monomers are cross-linked via a borate ester through the two side chain apiosyl residues. The RG-II dimer usually attaches to HGA and forms a macromolecular pectin network. Borate cross-link formation is regulated by pH and divalent cations. RG-II dimerization is also affected by changing side chain sugar residues. The borate

RG-II dimerization complex is critical for maintaining plant cell wall strength and normal plant development.

The synthesis of pectic polysaccharides is predicted to happen in Golgi lumen (Mohnen et al., 2008), where glycosyltransferases transfer sugar residues from nucleotide sugar donors to the glycan polymers. It has been predicted that at least 67 transferases involved in pectin biosynthesis, including glycosyltransferases, methyltransferases, and acetyltransferases (Sterling et al., 2006; Egelund et al., 2008; Caffall and Mohnen, 2009).

FIGURE 1.6. The structures of pectic polysaccharides.

A. Homogalacturonan (Taken from http://www.ccrc.uga.edu/~mao/galact/gala.htm).

B. Rhamnogalacturonan I (Taken from http://www.ccrc.uga.edu/~mao/rg1/rg1.htm).

C. Rhamnogalacturonan II (Taken from http://www.ccrc.uga.edu/~mao/rg2/intro.htm).

D. Xylogalacturonan (Taken from http://www.ccrc.uga.edu/~mao/galact/gala.htm).

E. Apiogalacturonan (Taken from http://www.ccrc.uga.edu/~mao/galact/gala.htm).

Glycolipids in plants - Plants contain many different types of lipids, including storage lipids, membrane lipids, lipids as regulators in cell signaling or as antioxidants, and lipids that serve to cover exposed surfaces (e.g., wax, cutin, and suberins) (Harwood, 1998).

Plant membrane lipids including those in plasma membranes, chloroplasts, mitochondria, glyoxysomal, ER, Golgi, and the nucleus can be divided into two major categories (Fig. 1.7): glycosylglycerides and phosphoglycerides (Harwood, 1997). The former (Fig. 1.7A) includes monogalactosyldiacylglycerol (MGDG), digalactosyldiacylglycerol (DGDG), and sulphoquinovosyldiacylglycerol (SQDG), which are mainly found in the chloroplast envelope and in the chloroplast thylakoid (Dormann and Benning, 2002); the latter includes lipids that are phosphorylated (Fig. 1.7B) and mainly found in the plasma membrane and membranes of other organelles.

FIGURE 1.7. The structures of plant membrane lipids (Harwood, 1997).

A. Glycosylglyceride, including monogalactosyldiacylglycerol (MGDG), digalactosyldiacylglycerol (DGDG), and sulphoquinovosyldiacylglycerol (SQDG).

B. Phosphotidate, the simplest phosphoglyceride.

Glycolipids in plants are predominantly found in the membranes in the form of steryl glycosides, glycosylceramides and diacylglycerol glycosides (Fig. 1.8). They help control membrane permeability, and contribute to the surface charge and hydration (Warnecke and

Heinz, 2010). Some were also found to interact with membrane proteins to facilitate their mobility and function. Some glycolipids, such as glycosphingolipids, are commonly found in lipid rafts, which contribute to the membrane trafficking and signal transduction (Hall et al.,

2010). Some preliminary evidences suggest the cellulose synthase complex is clustered in lipid rafts (Bessueille and Bulone, 2008).

FIGURE 1.8. Examples of structures of membrane glycolipids in plants (Warnecke and

Heinz, 2010).

A. Steryl glucoside, a form of steryl glycosides.

B. Glucosylceramide, a form of glycosylceramides.

C. Monogalactosyldiacylglycerol, a form of diacylglycerol glycosides.

The major sugar residues in those glycolipids are glucose or galactose. A few glycolipids contain glucuronic acid, mannose, and glucosamine residues. Carter et al. proposed the existence of GlcA in the corn phytoglycolipid (Fig. 1.9A) (Carter et al., 1969); Kaul and Lester also isolated two groups of phosphoinositol-containing sphingolipids from tobacco leaves, PSL-I and

PSL-II, which were found to contain hexuronic acid attached to inositol phosphate-ceramide

(Fig. 1.9B) (Kaul and Lester, 1975). The structure of these hexuronic acid(s) remains unknown.

GlcNAc is also found attached to the hexuronic acid in PSL-I. Inositol-containing glycosphingolipids are restricted to plant and fungi, but their functions are not completely understood. In addition, plants produce a class of sulfolipid in the chloroplast (Fig. 1.7A).

The formation of such plant glycolipids, as shown in the figure 1.7-1.9, likely requires

UDP-Glc (ADP-Glc, or GDP-Glc), UDP-Gal, and/or UDP-GlcA as sugar donors. However, how these nucleotide sugars get into the chloroplast or other organelles is not fully understood. The inner chloroplast membrane contains many transporters that transport triose-phosphate (TPT), phosphoenolpyruvate, glucose-6-P (GPT), and xylulose-5-P (XPT). It is also known that ADP-

Glc could be transported into amyloplast by a specific transporter for starch synthesis, while

UDP-Glc could be synthesized in the chloroplast by a chloroplast-localized UDP-Glc PPase

(Okazaki et al., 2009). UDP-Gal and UDP-GlcA transporters in the chloroplast were not identified. One cannot exclude the possibility that short lipid(s) on the chloroplast can be glycosylated on the cytosolic face of the organelle and flipped into the chloroplast lumen, serving as a sugar donor. Such flipping mechanisms were described in the ER of eukaryotic cells

(Lodish et al., 2000).

FIGURE 1.9. Structures of glycolipids identified in plant species containing glucuronic acid residues. GlcA (or hexuronic acid) is attached to the inositol phosphate-ceramide.

A. Structures of a corn phytoglycolipid (Kaul and Lester, 1975).

B. Structures of two groups of phosphoinositol-containing sphingolipid from tobacco leaves,

PSL-I and PSL-II (Kaul and Lester, 1975).

The synthesis of glycolipids requires lipid glycosyltransferases, some of which use NDP- sugars as sugar donors (Axelos and Peaud-Lenoel, 1978; Kelly and Dormann, 2004). UDP-Glc-

6-sulfate is shown to be required for the synthesis of the sulfolipid (Sanda et al., 2001) and UDP-

Gal is required for the synthesis of the monogalactosyldiacylglycerol (Benning and Ohta, 2005).

The enzymes that transfer the hexuronic acid to the lipid (for example, PSL-I and PSL-II in tobacco in Fig. 1.9) were not characterized. It is also possible that minor sugar components of plant glycolipids remain to be discovered. A detailed study needs to be carried out to identify the structure, metabolism, and biological function of these glycolipids.

In addition to large complex glycan polymers, plants also generate small secondary metabolites that are often glycosylated.

Glycosides in secondary metabolites and plant immunity - All plants produce secondary metabolites. In the plant kingdom, secondary metabolites include over 30,000 different substances. Some are pigments (e.g., flavonoids and carotenoids), scents (e.g., terpenes), anti-fungal molecules (e.g., sulphides and triterpenoids), and attractants (e.g., phytoestrogens). Certain secondary metabolites have specific roles as signaling molecules during host-parasite interaction or during symbiosis. Some plant secondary metabolites are glycosylated. The degree of glycosyl residues attached to the secondary metabolites also varies from a single sugar as is the case for apigenin 7-O-glucuronide, di-saccharide in naringin (the bitter compound in grapefruit) and even longer glycan structures, such as the branch chain oligosaccharide in the antifungal saponin (Fig. 1.10 A, B and C) (Morrissey and Osbourn, 1999;

Frydman et al., 2005; Noguchi et al., 2009). Flavonoids comprise of the major groups of secondary metabolites, with more than 7000 known compounds. Their roles span from growth

inhibition in bacteria and viruses, to mediating host recognition within bacteria, to signaling pathways in pathogenic fungi, and to intra-cellular events required for pollen germination.

FIGURE 1.10. Examples of glycosylated secondary metabolites in plants.

A. Apigenin 7-O-glucuronide, a flavone containing a single sugar residue (Noguchi et al., 2009).

B. Naringin, a flavanone containing a Rha-Glc disaccharide (Bar-Peled et al., 1991).

C. α-tomatine, a glycosylated alkaloid type of anti-fungal saponin from tomato/potato plants, containing a branch chain oligosaccharide (Morrissey and Osbourn, 1999).

As mentioned, some flavonoids are glycosylated. Most flavonol and flavanone glycosides contain Glc or Rha residues, as in the bitter naringin and neohesperidin from grapefruit, or the tasteless hesperidin found in oranges (Bar-Peled et al., 1991; Harborne and Williams, 2000;

Frydman et al., 2005), for example. Few of them contain GlcA residues, such as baicalin 7-O- glucuronide, a flavones, and apigenin 7-O-glucuronide (Noguchi et al., 2009). The attachment of sugars to these flavonoids is mediated by specific transferases that utilize NDP-sugars as sugar donors. For example, rhamnosyltransferase transfers Rha from UDP-Rha to make naringin or neohesperidin (Bar-Peled et al., 1991), and the specific enzyme (F7GA) that transfers GlcA from

UDP-GlcA onto the 7 position of apigenin required to make apigenin 7-O-glucuronide (Noguchi et al., 2009). In addition to Rha and Glc, other sugars were found attached to this class of metabolites including xylose and arabinose. The latter is found in secondary metabolites, such as

α-tomatine and avenacin A1 (Morrissey and Osbourn, 1999). While the most common hexuronic acids attached to the secondary metabolites are GlcA, the 4-epimer, GalA, is also found.

Examples include the flavonoid apigenin-7-galacturonic acid methyl ester in Chrysanthemum cinerariaefolium (Sashida et al., 1983; Kumar et al., 2005). Plants generate many secondary metabolites in response to pathogen attacks or stress. Some compounds are synthesized in healthy plants constitutively, building a barrier to protect themselves against the pathogen attack, while the formation of some other compounds is induced by a pathogen attack, as reviewed by

Morrissey and Osbourn (Morrissey and Osbourn, 1999). Many of the anti-microbial secondary metabolites are glycosylated, such as saponin, cyanogenic glycosides, and glucosinolates

(Morrissey and Osbourn, 1999). The sugar residues found in those molecules include Glc, GlcA,

Xyl, Rha, or Gal. Sugar residues provide the recognition site for pathogen interaction, and removal of those sugars leads to the loss of biological activity. As noted above, a significant flux

of sugars (i.e. NDP-sugars) will be required to generate the glycosides attached to the secondary metabolites. How the flux of these NDP-sugars is controlled, and whether there are specific

NDP-sugar transporters responsible for the supply of those activated sugars remain unclear.

To summarize, numerous metabolic pathways in plants require NDP-sugars, but the source of the NDP-sugars for the synthesis of wall polysaccharides, glycolipids, glycoproteins, as well as secondary metabolites is not well defined. Glycan-containing molecules are also recycled by different mechanisms. How often these glycans are salvaged and what the mechanism is that controls this process is not clear. Below we describe the enzymes involved in plant glycan degradation.

Polysaccharide hydrolytic enzymes

Polysaccharide hydrolytic enzymes can be divided into three categories: the Glycoside

Hydrolase family (GH), the Polysaccharide Lyase family (PL), and the Carbohydrate Esterase family (CE). Most of the genes encoding these activities are categorized in the CAZy database.

Below I will summarize their activities and how they are related to the salvage pathway.

Glycoside Hydrolases (GH), also referred to as glycosidases, catalyze the hydrolysis of the O-, N- and S-linked glycosides, releasing a monosaccharide or oligosaccharides. Based on amino acid sequence similarities, glycoside hydrolases are classified into 122 families in the

CAZy database. Some enzymes in the GH 28 group cleave plant cell wall pectic polymers, including for example, polygalacturonase (EC 3.2.1.15), exo-polygalacturonase (EC 3.2.1.67), exo-polygalacturonosidase (EC 3.2.1.82), and rhamnogalacturonase (EC 3.2.1.-). Some enzymes belonging to the GH family act as a „transferase‟ (GT 77). These enzymes are capable of transferring a segment of a glycan to a new acceptor, either a sugar or a glycan (e.g., 4-α-

glucanotransferase (EC 2.4.1.25)). On the other hand, other members of the GH family (GH16) can also cleave and transfer large polymers to each other. An example is xyloglucosyl transferase

(EC 2.4.1.107), which cleaves a β-(1-4) bond of a xyloglucan backbone and transfers the xyloglucan segment onto an acceptor containing a non-reducing terminal glucose residue in O-4 position (Nishitani and Tominaga, 1992).

From a cellular localization point of view, some of the GHs are localized in organelles including the ER (e.g., glucosidase) and Golgi (mannosidase), while other GHs are secreted to the cell surface, for example, xylanase (Fukuda et al., 2010) and xyloglucan endotransferase

(XET) (Davies et al., 2005). Some of these enzymes reside in the cytosol, such as starch phosphorylase (PHO), giving rise to the release of Glc-1-P from starch (Lu et al., 2006).

The GH enzymes are also classified as endo- and exo- GH. The exo-GHs release an individual sugar. An example is exo-polygalacturonase that hydrolyzes polygalacturonan to free

GalA. Polygalacturonans are often found on plant cell surface and the free GalA residues will likely be uptaken by the cell and re-activated by GalAK and sloppy via the salvage pathway to form UDP-GalA (Yang et al., 2009). In plants many of the GH activities are present in germinating seeds (Sitrit et al., 1999), in germinating and elongating pollen tubes (Dubald et al.,

1993; Holmes-Davis et al., 2005), and this may provide a source of activated sugar to the fast growing tissues. Some GHs are highly expressed in fruits (Hadfield and Bennett, 1998) to enhance N-glycoprotein turnover and start fruit ripening process. For example, some of the salvaged N-linked glycan can contribute to the flux of sugars to make UDP-sugars that are incorporated into phenolic compounds that constitute the fruit aroma of ripe tomato (Meli et al.,

2010).

Polysaccharide Lyases cleave certain glycosidic linkages in uronic acid-containing polysaccharides. The enzymatic process generates unsaturated oligosaccharides by the β- elimination mechanism. For example, pectin lyase cleaves -(1-4)-D-galacturonan methyl ester and releases an oligosaccharide with 4-deoxy-6-O-methyl--D-galact-4-enuronosyl groups at the non-reducing ends (Solbak et al., 2005). Based on amino acid sequence similarities, polysaccharide lyases are classified into 22 families in the CAZy database, including pectin lyase (EC 4.2.2.10), pectate lyase (EC 4.2.2.2), and exopolygalacturonate lyase (EC

4.2.2.9). Most of the polysaccharide lyases are cell wall associated. The genes encoding these activities are often highly expressed in pollen and ripening fruit (Marin-Rodriguez et al., 2002).

However, how the released modified-sugars are salvaged remains unclear.

Carbohydrate Esterases catalyze the de-esterification of the polysaccharides, including the de-O or de-N-acetylation of the saccharides and the demethylation of pectin (Fig. 1.11).

Based on amino acid sequence similarities, carbohydrate esterases are classified into 16 families in the CAZy database. Some esterases are cell wall associated, while some are in soluble form

(Pelloux et al., 2007). De-esterification by CE may facilitate the subsequent degradation of the polysaccharides by other hydrolytic enzymes.

Below is an example of the functions of the hydrolytic enzymes involved in pectin degradation:

FIGURE 1.11. Enzymes involved in pectin degradation.

Arrow 1 indicates the reactions catalyzed by pectinase and exo-polygalacturonase. Arrow 2 represents the reactions catalyzed by pectate lyase, exo-polygalacturonate lyase, and pectin lyase. Arrow 3 indicates the reactions catalyzed by pectin methyl esterase. R is O- or O-Me.

(Solbak et al., 2005).

Sugar flux and subcellular formation of activated-sugars

As mentioned above, free monosaccharides, disaccharides or small oligosaccharides released by the hydrolytic enzymes can enter back into the cell by passive diffusion or via sugar transporters (Fig. 1.5). The sugars that are salvaged back into the cell may go to different pathways (Fig. 1.4). Sugars, such as Glc, Man, or Frc are phosphorylated at the C6 position by hexokinase (Wilson, 1995), and enter different metabolic routes (e.g., glycolysis pathway), or by the action of phosphohexo-mutase to sugar-1-Ps. On the other hand, other sugars, such as Gal,

Ara, GlcA, and GalA can be directly phosphorylated at the C1 position by specific sugar kinases

(SK) to form sugar-1-Ps. The sugar-1-Ps may then be converted to nucleotide sugars in the cytosol (Seifert, 2004; Bar-Peled and O'Neill, 2010). The activities of some representative salvage pathway enzymes are shown in Fig. 1.4. Some cytosolic NDP-sugars will then be transported into the ER and Golgi for the supply of donor substrates for glycosylation. Several transporters importing cytosolic NDP-sugars to the Golgi have been characterized (Handford et al., 2006), although many are still missing.

In addition to the requirement of activated sugars for wall biosynthesis, NDP-sugars in the cytosol could also be involved in other pathways. It is possible that the synthesis of certain glycans requires NDP-sugar made in cytosol so that it can be transported to a specific organelle, or to the nucleus; another possibility is that cytosolic NDP-sugars may be involved in the glycosylation of secondary metabolites, or the glycosylation of the lipids in organelles where the lipids are facing the cytosol. There is also a possibility that some cytosolic NDP-sugars may function as signaling molecules in plants. In human cells for example, UDP-GlcNAc serves as a glucose sensor in the cytosol and nucleus (Zeidan et al., 2010). Therefore, in addition to understanding what enzyme is required to produce a given NDP-sugar, where this enzyme

localizes and where the product functions is essential to evaluate the control and flux of sugars in a cell.

As illustrated above, the salvage pathway is likely an essential metabolic route for normal plant cell growth and development. The function of the salvage pathway possibly varies among tissues. To appreciate the role of the salvage pathway during plant growth and development, mutant studies could be an important research tool. In the following section, salvage pathway mutants will be mentioned.

PART II: THE EFFECTS OF MUTANTS IN THE SALVAGE PATHWAY ON PLANT

DEVELOPMENT

A genetic study of a UDP-sugar pyrophosphorylase (sloppy) knockout mutant generated by T-DNA insertion revealed that the activity of sloppy has a critical role in pollen development

(Schnurr et al., 2006). Homozygotic mutants of sloppy have a defect in pollen grain formation, and next generation seeds cannot be obtained. Analyses of pollen grains from wild-type and the sloppy mutant indicated that the mutation prevented the synthesis of the pectin-rich inner wall

(intine), but had no apparent effect on the outer wall (exine) (Schnurr et al., 2006). These results suggest that a salvage pathway enzyme, like sloppy, is required to provide a pool of specific

UDP-sugar(s) for the synthesis of pollen pectin-rich inner wall (intine). While sloppy is expressed in all tissues (Litterer et al., 2006), its function in Arabidopsis, appears vital only in one specific tissue (intine) at a certain developmental stage. As sloppy is capable of uridylating

(PPase) Glc-1-P, Gal-1-P, Xyl-1-P, GlcA-1-P, Ara-1-P, and GalA-1-P (Fig. 1.4), it would be of interest to determine which sugar kinase(s) are expressed in the pollen at that developmental stage, and what other genes are involved in the recycling of those sugars.

Besides sloppy, UDP-Glc can also be generated by a very specific enzyme: UDP-Glc pyrophosphorylase. UDP-Glc PPase is a protein widely conserved in plants, animals and bacteria. Its product, UDP-Glc, is the most abundant NDP-sugar and is required for the synthesis of UDP-GlcA, UDP-Gal and UDP-Rha (Mohnen et al., 2008), and see Fig. 1.2. In Arabidopsis, two genes encoding cytoplasmic UDP-Glc PPases (UGP1 and UGP2) have been identified and functionally characterized. They both display PPase activity but are expressed distinctively in different tissues: UGP1 is strongly expressed in late embryogenesis and in seeds, while UGP2 is upregulated in roots and fertilized flowers, and is suppressed in embryos and seeds. This tissue specific expression pattern could suggest that the salvage pathway may be tightly regulated in specific tissues (Meng et al., 2009). The UGPase activity is significantly reduced in ugp mutant plants. At the same time, an increase of sloppy expression was detected in ugp mutant plants, suggesting the existence of a compensatory mechanism for UDP-Glc synthesis. A mutant study of rice ugp1 suggested the enzyme product is critical for callose deposition during pollen development (Chen et al., 2007). UGP double knockout plants had greater hypocotyl and root length, and a decrease in seed production (Meng et al., 2009). However, no significant changes were detected in sucrose, starch and cell wall synthesis in double knockout plants, indicating sloppy or other UDP-Glc biosynthetic pathways should be involved in the compensation of the loss of UGP. Recently, a gene encoding another UGP (UGP3) in the chloroplast has been identified (Okazaki et al., 2009). Mutant studies of ugp3 indicated this gene is involved in sulfolipid biosynthesis.

Several other NDP-sugar pyrophosphorylases are known, including GDP-Man PPase,

GDP-Fuc PPase, ADP-Glc PPase and UDP-GlcNAc (GalNAc) PPase. GDP-Man PPase catalyzes the conversion of Man-1-P and GTP to GDP-Man (Carlson and Hansen, 1962). An

Arabidopsis GDP-Man PPase mutant showed defects in cellulose synthesis (Lukowitz et al.,

2001) and ascorbic acid synthesis (Conklin et al., 2000), and the plant was hyper-sensitive to ammonium (Qin et al., 2008), indicating its importance in plant growth. The biochemical activity of the encoded gene product (At2g39770) was confirmed in our lab and the enzyme specifically recognizes Man-1-P as a substrate.

GDP-Fuc was originally considered to be synthesized through GDP-Man in plants.

However, studies of mur1 mutants, which showed a dwarf phenotype due to the lack of fucose residues in RG-II, revealed the phenotype can be rescued by exogenously applied free L-Fuc, suggesting GDP-L-Fuc could be generated from L-Fuc via the salvage pathway (O'Neill et al.,

2001). As the enzyme participating in the salvage reaction of free L-Fuc is well characterized in mammals (Pastuszak et al., 1998) and in Bacteroides fragilis (Coyne et al., 2005), sequence similarities enabled us and another laboratory to identify a bifunctional enzyme with both L- fucokinase and GDP-L-Fuc pyrophosphorylase activities in Arabidopsis (Kotake et al., 2008).

The N-terminal region is similar to mammalian GDP-L-Fuc pyrophosphorylase and the C- terminal region contains an ATP-binding motif conserved for GHMP kinases. Such a bifunctional enzyme might efficiently drive free fucose to GDP-L-Fuc without releasing the intermediate Fuc-1-P in the cytoplasm. The fucokinase mutants (fkgp) did not show any visible phenotype compared to the WT plant (Kotake et al., 2008). The sugar composition of the total wall glycans did not change, but the mutant showed an over 40-fold accumulation of soluble fucose in the cell, indicating the salvage pathway for GDP-L-Fuc synthesis has been blocked in the fkgp mutants.

ADP-Glc PPase (AGP) is mainly involved in starch synthesis. The enzyme converts Glc-

1-P to ADP-glucose (Fig. 1.4) in the presence of ATP in amyloplast (Smith, 2001). AGP

catalyzed formation of ADP-Glc is considered to be the key step in starch biosynthesis

(Kleczkowski, 2000). AGP is composed of two large and two small subunits. Mutants in the gene encoding either the large or small subunit of AGP showed a starch deficiency phenotype

(Orzechowski, 2008). GUS staining of the Arabidopsis AGP small subunit found it is located in all plant tissues, usually associated with starch granules (Siedlecka et al., 2003). The AGP is regulated by 3-phosphoglycerate and inorganic Pi or PPi (Kleczkowski, 2000; Tiessen et al.,

2002). It is also regulated by light or sugar levels (Tiessen et al., 2002). High levels of sucrose or glucose could activate AGP through particular signaling pathways (Tiessen et al., 2003). It is worth noting that in some species, ADP-Glc is mainly generated in the cytosol by cytosolic

ADP-Glc pyrophosphorylase, and then transported into amyloplast by specific transporters for starch synthesis (Ballicora et al., 2000; Beckles et al., 2001).

Two UDP-GlcNAc(GalNAc) PPases (Fig. 1.4) were recently characterized in

Arabidopsis (Yang et al., 2010). The enzymes are also named as N-acetylglucosamine-1- phosphate uridylyltransferases (GlcNAc1pUT-1 and 2). The enzymes can utilize the 4-epimer,

UDP-GalNAc, as a substrate as well. The two enzymes share 86% amino-acid sequence identity.

However, GlcNAc1pUT-2 was found to convert in addition to GlcNAc-1-P and GalNAc-1-P,

Glc-1-P to the corresponding UDP-sugars, suggesting that subtle changes in the UT family can cause different substrate specificities. UDP-GlcNAc is the major precursor for the synthesis of many different types of glycans across species. In the cytosolic face of the endoplasmic reticulum, the attachment of GlcNAc is required to initiate N-glycan processing for the synthesis of glycoproteins and glycolipids.

An arabinokinase mutant (ara1) has been identified in Arabidopsis in early 90s (Dolezal and Cobbett, 1991). The ara1 mutant is sensitive to exogenous L-Arabinose, and mutant

seedlings failed to grow and became necrotic in the presence of exogenous L-Arabinose, whereas in the absence of L-Arabinose, they did not show any visible phenotype compared to the WT.

This suggests that free Ara residue, if accumulated, is toxic to the plant.

GlcA kinase transcript analysis, based on the eFP database, indicated that the gene is highly expressed in mature pollen (Pieslinger et al., 2010). GlcAK not only plays a role in the terminal step of myo-inositol oxidation pathway, but it is also involved in the salvage pathway for the recycling of the free GlcA (Fig. 1.3). Therefore, it is likely that GlcAK may have pivotal functions during pollen germination or pollen grain development. It would be worth studying the process of pollen development in the GlcAK mutant.

The myo-inositol pathway (Fig. 1.3) - Transcriptional analysis of the inositol oxygenase

(MIOX) gene family in Arabidopsis indicated two MIOX isoforms (MIOX1 and MIOX2) are expressed in all the tissues tested, while the other two (MIOX4 and MIOX5) are specifically expressed in mature pollen, indicating they may play an essential role in pollen formation

(Kanter et al., 2005). However, the single knock-out mutant did not show any visible phenotypes. The monosaccharide composition in the mutant is not altered compared to WT control. It is possible that the lack of one isoform could be compensated by the existence of the other isoform. Therefore, it would be interesting to analyze the double knockout mutant of the

MIOX genes in plants.

Mutant studies of those genes indicated that the salvage pathway is essential for plant growth and development. However, the enzymes in this pathway are not fully characterized.

Plants possess over 30 NDP-sugars, and it remains unknown if and how other sugars are salvaged, for example, Xyl or Rha. Therefore, more sugar kinases and NDP-sugar PPases need to be identified and functionally studied. In the following section, a detailed analysis of sugar

kinases (SK) and NDP sugar PPases (NS-PPAse) will be carried out, including the analysis of conserved functional domains of specific sugar kinases and NDP-sugar ppases, sequence alignments as well as phylogeny analysis. These analyses will help us to find other sugar kinase and nucleotide sugar pyrophosphorylase genes and orthologs across different plant species.

PART III: COMPARATIVE ANALYSIS OF PROTEIN DOMAINS OF SUGAR

KINASES AND NUCLEOTIDE SUGAR PYROPHOSPHORYLASES IN PLANTS

There are different types of monosaccharide kinases in plants. Hexokinases (Wilson,

1995) typically refer to a group of enzymes that use ATP to phosphorylate the C6-hydroxy group of Man, Frc, or Glc to form sugar-6-P. Sugar kinases (SK) refer to a group of enzymes that phosphorylate the anomeric C1-OH group to form sugar-1-P and ADP. Before starting my PhD we had little knowledge on sugar kinases in plants. Analysis of the SK protein domains described below takes into account our knowledge today.

Proteins belonging to the sugar kinase group (SK) share amino-acid sequence similarities with the proteins of the GHMP kinase superfamily. GHMP kinase family includes large and diverse groups of proteins that utilize ATP to phosphorylate sugars, amino-acids, 5-carbon and

7-carbon aromatic molecules. Original members of the GHMP include galactokinase, homoserine kinase, mevalonate kinase and phosphomevalonate kinase. Recent addition of proteins belonging to the GHMP family also includes N-acetylgalactosamine kinase, mevalonate

5-diphosphate decarboxylase, archeal shikimate kinase, 4-(cytidine 5'-diphospho)-2-C-methyl-D- erythritol kinase. Two other proteins that are catalytically inactive are also members of the

GHMP kinase family. These proteins act as signaling molecules of either developmental regulator or galactose sensor, including the C. elegans sex-fate determining protein XOL-1 and

Saccharomyces cerevisiae transcriptional activator Gal3p (Bajwa et al., 1988; Luz et al., 2003;

Timson, 2007) respectively.

The conserved functional domains of GHMP kinase superfamily

Using Pfam protein domain search analysis we were able to identify two conserved domains: ATP binding domain (annotated as GHMP_kinases_N in Pfam) and catalytic domain

(annotated as GHMP_kinases_C in Pfam) in Arabidopsis GHMP kinase proteins (see schematic presentation in Fig. 1.12). Many of the peptides that share the GHMP kinase consensus domains vary in protein length, ranging from 400 aa (GlcAK, glucuronokinase), 500 aa (GalK, galactokinase), to 600 aa (PMK, phosphomevalonate kinase), and even longer GHMP kinase proteins (>1000aa) belonging to this family are AraK and FucK (Fig. 1.12). Some of the potential SK such as GalK, GalAK and AraK also contain a sugar binding domain located in the

N-terminal region of this class of proteins and annotated in Pfam as GalKase_gal_bdg domain.

This sugar binding domain contains amino acids that specifically bind the sugar module. For example, a conserved glutamate is found in GalK proteins (see Fig. 1.13 pointed with arrow), but in GalAK from various plant species this residue is replaced by alanine (Yang et al., 2009). We have confirmed by site-directed mutagenesis that in GalAK, the non-polar alanine residue (A41) can accommodate the charged glycoses such as GalA, whereas a glutamate residue would repulse the carboxyl group of these uronic acids. In AraK, such amino acid is replaced by glycine. Whether such change allows the protein to take the pentose is unclear. No obvious sugar binding domain was determined for FucK and GlcAK. Nonetheless, the FucK enzyme contains a long domain located at the N-terminal portion (see green color in Fig. 1.12) of the protein and is termed by Pfam Fucokinase domain.

FIGURE 1.12. Schematic presentation of conserved Pfam domains identified in 13 sequences of GHMP kinases from Arabidopsis.

All GHMP kinases share a GHMP_kinases_N (ATP binding) and GHMP_kinases_C (catalysis) domains. GalK, GalAK and AraK contain a GalKase_gal_bdg domain (sugar binding), whereas fucokinase contains a Fucokinase domain.

Sequence alignment of the conserved N-terminal GalKase_gal_bdg domain from five plant genomes, including Arabidopsis (Analysis of the genome sequence of the flowering plant

Arabidopsis thaliana, 2000), rice (Goff et al., 2002; Yu et al., 2002), moss (Rensing et al., 2008), green alga (Worden et al., 2009) and red alga (Matsuzaki et al., 2004) genomes is shown in Fig.

1.13. This domain is likely involved in sugar binding, and it is specific for GalK, GalAK and

AraK. In this domain, some amino acids are highly conserved. However, some are different, for example, the glutamate in GalK is involved in the binding of C6-OH group of galactose, but in

GalAK it is substituted by alanine. Since there is no crystal structure for GalAK and AraK, it remains unclear what are the specific amino acids for the recognition of GalA or Ara in those kinases.

FIGURE 1.13. Multiple sequence alignment of 13 Pfam GalKase_gal_bdg domain regions of AraK, GalK, and GalAK in five plant genomes.

The multiple sequence alignment of the conserved Pfam GalKase_gal_bdg domain was performed using the MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The alignment was displayed using ClustalX (Larkin et al., 2007). Proteins are predicted from Arabidopsis, rice, moss, green alga and red alga genomes. Arrow represents the amino acid that is believed to be involved in the recognition of C6 of the sugar (Thoden and Holden, 2003).

Sequence alignment of the conserved GHMP_kinases_N domain from five plant species is shown in Fig. 1.14. This domain is likely involved in ATP-binding (Hartley et al., 2004). We divided this domain to eight regions: 1‟-8‟. As shown (as seen in Fig. 1.14), some regions (e.g.,

2‟, 4‟and 7‟) have gaps among different kinases. How these affect enzyme specificity remains to be studied.

FIGURE 1.14. Multiple sequence alignment of 45 GHMP_kinases_N domain regions of

GHMP kinases in five plant genomes.

The multiple sequence alignment of the conserved Pfam GHMP_kinases_N domain was performed using the MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The alignment was displayed using ClustalX (Larkin et al., 2007). The proteins are predicted from Arabidopsis, rice, moss, green alga and red alga genomes. This domain is likely involved in ATP binding (Zhou et al., 2000; Hartley et al., 2004).

Sequence alignment of the conserved GHMP_kinases_C domain from five plant species is shown in Fig. 1.15. This domain is likely involved in the enzyme catalytic function (Thoden et al., 2005). We divided this domain to five regions: 1‟-5‟. As shown some regions (e.g., 2‟ and 4‟) have gaps among different kinases. How these affect enzyme catalytic rate is yet to be studied.

FIGURE 1.15. Multiple sequence alignment of 39 GHMP_kinases_C domain regions of

GHMP kinases in five plant genomes.

The multiple sequence alignment of the conserved Pfam GHMP_kinases_C domain was performed using the MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The alignment was displayed using ClustalX (Larkin et al., 2007). The proteins presented are predicted from Arabidopsis, rice, moss, green alga and red alga genomes. This domain is likely involved in catalysis and transfers the  phosphate to the sugar (Thoden et al.,

2005). .

Phylogenic analysis of the GHMP domain structures in plants

We created three different phylogenetic trees using individual domains

(GHMP_kinases_N, GHMP_kinases_C, and GalKase_gal_bdg). Ten distinct clades appeared based on the phylogenies of the conserved GHMP_kinases_N domain and GHMP_kinases_C domain regions in the proteins from five plant species spanning from lower to higher plants (Fig.

1.16, A and B). In the phylogentic tree generated by comparing the GHMP_kinases_N domain

(i.e. the ATP binding), the five known sugar kinases (i.e. GalK, GalAK, GlcAK, FucK and

AraK) are separated into two distinct clades (c, d, see Fig. 1.16A). Proteins having ATP-binding domain belonging to MVD, MK, PMK, HSK and CDP-MEK can also be separated to two distinct clades (a, b, Fig. 1.16A).

Phylogentic tree generated by comparing the GHMP_kinases_C domain (i.e. the likely catalytic domain involved in transferring phosphate to the substrate) is complex as shown in Fig.

1.16B. And currently no conclusion can be drawn without looking at GHMP kinases from other organisms.

Phylogentic tree generated by comparing the sugar binding domain (GalKase_gal_bdg) in the proteins from five plant species is relatively simple but one cannot distinguish by sequence alone the substrate specificity, as will be shown later (Chapter 2). This GalKase_gal_bdg domain is only found in the N-terminal regions of three sugar kinases: GalK, GalAK and AraK.

Phylogenetic analysis indicated the enzymes fall into three distinct and well-separated clades

(Fig. 1.16C).

FIGURE 1.16. Phylogenies of the three conserved Pfam domain regions in plant GHMP kinases.

The multiple sequence alignment of the conserved Pfam domain was performed using the

MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The phylogeny was reconstructed using the FastTree version 2.1.3 program (Price et al., 2009). Proteins presented are predicted from Arabidopsis, rice, moss, green alga and red alga genomes.

A. Phylogeny of 45 GHMP_kinases_N domains in five plant species.

B. Phylogeny of 39 GHMP_kinases_C domains of in five plant species.

C. Phylogeny of 13 GalKase_gal_bdg domains of GalK, GalAK and AraK in five plant species.

The conserved functional domains of NDP-sugar pyrophosphorylases

NDP-sugar pyrophosphorylases (NS-PPAse) are the members of nucleotidyltransferase superfamily (EC 2.7.7) and are widely spread in eukaryotes and prokaryotes. In general, members of this group convert NTP (e.g., ATP, GTP, CTP, dTTP, UTP) and sugar-1-P to NDP- sugar and PPi. These enzymes are typically reversible, unless PPi is readily hydrolyzed to 2×Pi.

The proper name for these enzymes (NTP:sugar-1-P nucleotidyltransferase) is nucleotidyltransferases but often they are named as NDP-sugar pyrophosphorylases. The latter represents the degradation rather than the synthesis of NDP-sugar.

A protein domain search of Arabidopsis NDP-sugar pyrophosphorylases against the Pfam database found the conserved UDPGP domain in UDP-Glc PPase (UGP), UDP-GlcNAc PPase

(UAP) and sloppy proteins using E-value cutoff < 1 (Fig. 1.17). In contrast, Arabidopsis ADP

(ADP-Glc PPase like), APL (ADP-Glc PPase large subunit), APS (ADP-Glc PPase small subunit) and GMP (GDP-Man PPase) proteins were found to contain the Pfam NTP_transferase domain, suggesting they are distinct from the UDP-Glc PPase (UGP), UDP-GlcNAc PPase

(UAP) and sloppy (Fig. 1.17), although both groups, UDPGP and NTP-transferases, are nucleotidyltransferases.

FIGURE 1.17. Schematic presentation of conserved Pfam domains identified in 19 sequences of NDP-sugar pyrophosphorylases from Arabidopsis.

UDP-Glc PPase (UGP), UDP-GlcNAc PPase (UAP) and UDP-sugar ppase (sloppy) share a

UDPGP domain; ADP-Glc PPase like (ADP), ADP-Glc PPase large subunit (APL), ADP-Glc

PPase small subunit (APS) and GDP-Man PPase (GMP) share an NTP_transferase domain.

Sequence alignment of the conserved UDPGP domain of UGP, UAP and sloppy from five different plant species is shown in Fig. 1.18. Based on the crystal structure of UGP and

UAP, this domain contains amino acids that are involved in uracil binding (UB) and sugar- nucleotide (NB) binding (Peneff et al., 2001; McCoy et al., 2007; Steiner et al., 2007; Yang and

Bar-Peled, 2010; Yang et al., 2010).

FIGURE 1.18. Multiple sequence alignment of the 21 Pfam UDPGP domain regions in five plant genomes.

The multiple sequence alignment of the conserved Pfam UDPGP domain was performed using the MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The alignment was displayed using ClustalX (Larkin et al., 2007). The proteins presented are predicted from

Arabidopsis, rice, moss, green alga and red alga genomes. The amino acids that are involved in uracil binding (UB) and sugar-nucleotide (NB) binding are labeled underneath.

Phylogenetic analysis of plant NDP-sugar pyrophosphorylases

Phylogenetic analysis of the UDPGP domain regions in five plant genomes revealed four major and distinct clades (Fig. 1.19, labeled a, b, c, d). The members of the UGP clade generate

UDP-Glc, and evolutionary they are well separated from the members of UAP that generate

UDP-GlcNAc. Of interest is the sloppy clade. This clade harbors a sloppy activity capable of converting UTP and many different types of sugar-1-Ps to their corresponding UDP-sugars:

UDP-Xyl, UDP-GalA, UDP-Ara, UDP-Glc, UDP-Gal, UDP-GlcA (Fig. 1.4). The sloppy clade, as shown in Fig. 1.19, appears closer to the UGP3 clade. The closer homology shared by UGP3 and sloppy could, in part, be explained by their enzyme activity. UGP3 is a bit promiscuous (but not as sloppy) and was shown to generate UDP-Glc and to a certain degree UDP-Gal (Okazaki et al., 2009).

FIGURE 1.19. Phylogeny of 21 UDPGP domain regions in five plant species.

The multiple sequence alignment of the conserved Pfam UDPGP domain was performed using the MAFFT version 6.603 program (Katoh et al., 2005) with the L-INS-I method. The phylogeny was reconstructed using the FastTree version 2.1.3 program (Price et al., 2009). Proteins presented are predicted from Arabidopsis, rice, moss, green alga and red alga genomes. UGP,

UAP, sloppy and UGP3 are separated into four major clades. Sloppy clade is closer to UGP3 clade.

The above comparative analysis of protein domains of sugar kinases (SK) and nucleotide sugar pyrophosphorylases (NS PPAse) in plants was instrumental to identify potential enzymes in the salvage pathway in different organisms and examine their substrate specificities as will be described in the following chapters.

PART IV: THESIS OVERVIEW

Chapter 1 describes the salvage pathway.

Chapter 2 describes the functional characterization of three genes/proteins of the salvage pathway, including a galacturonic acid kinase (GalAK), a galactokinase (GalK), and a broad

UDP-sugar pyrophosphorylase (sloppy) in Arabidopsis.

Chapter 3 describes the functional characterization of two salvage pathway genes/proteins having broad UDP-sugar pyrophosphorylase (sloppy) activities in Trypanosoma Cruzi and

Leishmania Major.

Chapter 4 describes the functional characterization of two Arabidopsis N-acetylglucosamine-1-P uridylyltransferases, and the comparison of their different substrate specificities, as well as their potential roles in the NDP-sugar salvage pathway.

Chapter 5 describes the analysis of a salvage pathway mutant, galak in Arabidopsis, and the possible role of GalAK in glycan metabolism.

CHAPTER 2

IDENTIFICATION OF GALACTURONIC ACID-1-PHOSPHATE KINASE, A NEW

MEMBER OF THE GHMP KINASE SUPERFAMILY IN PLANTS, AND

COMPARISON WITH GALACTOSE-1-PHOSPHATE KINASE1

1This research was originally published in The Journal of Biological Chemistry. Reprinted here with permission of publisher. Ting Yang, Liron Bar-Peled, Lindsay Gebhart, Sung G. Lee, and

Maor Bar-Peled. J. Biol.Chem. 2009 284: 21526-21535. © The American Society for

Biochemistry and Molecular Biology.

ABSTRACT

The process of salvaging sugars released from extracellular matrix, during plant cell growth and development is not well understood, and many molecular components remain to be identified. Here we identify and functionally characterize a unique Arabidopsis gene encoding an

α-D-galacturonic acid-1-phosphate kinase (GalAK) and compare it with galactokinase. The

GalAK gene appeared to be expressed in all tissues implicating that glycose salvage is a common catabolic pathway. GalAK catalyzes the ATP-dependent conversion of α-D-galacturonic acid (D-

GalA) to α-D-galacturonic acid-1-phosphate (GalA-1-P). This sugar phosphate is then converted to UDP-GalA by a UDP-sugar pyrophosphorylase as determined by a real-time 1H-NMR-based assay. GalAK is a distinct member of the GHMP kinase family that includes galactokinase (G), homoserine kinase (H), mevalonate kinase (M), and phosphomevalonate kinase (P). Although these kinases have conserved motifs for sugar binding, nucleotide binding and catalysis, they do have subtle difference. For example, GalAK has an additional domain near the sugar-binding motif. Using site-directed mutagenesis we established that mutation in A368S reduces phosphorylation activity by 40%; A41E mutation completely abolishes GalAK activity; Y250F alters sugar specificity and allows phosphorylation of D-glucuronic acid, the 4-epimer of GalA.

Unlike many plant genes that undergo duplication, GalAK occurs as a single copy gene in vascular plants. We suggest that GalAK generates GalA-1-P from the salvaged GalA that is released during growth-dependent cell wall restructuring, or from storage tissue. The GalA-1-P itself is then available for use in the formation of UDP-GalA required for glycan synthesis.

INTRODUCTION

D-galacturonic acid (D-GalA) is a quantitatively major glycose present in numerous plant polysaccharides including pectins and arabinogalactan proteins (Mohnen, 2002). The synthesis of these polysaccharides requires a large number of glycosyltransferases and diverse nucleotide- sugar (NDP-sugar) donors (Mohnen, 2002; Mohnen, 2008). Some of these NDP-sugars are formed by inter-conversion of pre-existing NDP-sugars and by salvage pathways. For example, the main pathway for UDP-GalA formation is the 4 epimerization of UDP-GlcA, a reaction catalyzed by UDP-GlcA 4-epimerase (Gu and Bar-Peled, 2004). However, in ripening Fragaria fruit D-GalA is incorporated into pectin (Loewus, 1961). It is likely that a sugar kinase converts the D-GalA to GalA-1-P (Neufeld et al., 1961) which is then converted to UDP-GalA by a non- specific UDP-sugar pyrophosphorylase (Litterer et al., 2006). Myo-inositol may also be a source of GalA for polysaccharide biosynthesis (Young et al., 1966).

Galacturonic acid is likely to be generated by enzyme-catalyzed hydrolysis of pectic polysaccharides in plant tissues. Polysaccharide hydrolase activities are present in germinating seeds (Ren and Kermode, 2000), in germinating and elongating pollen (Holmes-Davis et al.,

2005) and in ripening fruit (Hadfield and Bennett, 1998). Thus, monosaccharide salvage pathways may be required for normal plant growth and development.

Numerous sugar-1-P kinase including D-Gal-1-P kinase (Kaplan et al., 1997), L-Ara-1-P kinase (Sherson et al., 1999), and L-Fuc-1-P kinase (Kotake et al., 2008) have been described

(Neufeld et al., 1959), but no D-GalA-1-P kinase has been identified in any species to account for the hydrolysis and recycle of pectic polymers. The subsequent pyrophosphorylation of UDP- sugars could be carried out by UDP-sugar pyrophosphorylases (Kotake et al., 2004).

Here, we report the identification and characterization of a functional galacturonic acid kinase (GalAK). We compared the activity of GalAK with a previously uncharacterized

Arabidopsis GalK and discussed the evolution of these sugar kinase members of the GHMP kinase.

EXPERIMENTAL PROCEDURES

cDNA cloning and site directed mutagenesis of Arabidopsis GalAK - Total RNA was extracted from the stems of six-week-old Arabidopsis plants and used as a template to reverse- transcribe cDNA with oligo-dT as primer (Gu and Bar-Peled, 2004). The coding sequence of

Arabidopsis GalAK was amplified by PCR using 1 unit of high-fidelity proof-reading Platinum

DNA polymerase (Invitrogen), and 0.2 M of each forward and reverse primers: 5'-AC atg tct tgg cct acg gat tct gag-3' and 5'-GG TAC CTC gag aag aac acg agc agc gtc-3'. The RT-PCR product was cloned to generate plasmid pCR4-topoTA:At3g10700#11; and sequenced (GenBank

FJ439676). The PciI-KpnI fragment (1276 bp) containing the full length AtGalAK gene without the stop codon was sub-cloned into an E. coli expression vector derived from pET28b (Gu and

Bar-Peled, 2004), generating GalAK with an extension of six histidines at its C-terminal. GalK

(At3g06580.1) and UDP-sugar pyrophosphorylase (At5g52560.1 “Sloppy”) were cloned using primer sets [GalK, S#1 5'-GCC atg gcg aaa ccg gaa gaa gta tca gtc-3' + AS#2 5'-GAT ATC TCG

AGg agg ttg aag atg gca gca c-3'] and [Sloppy, S#1 5'-atg gct tct acg gtt gat tcc-3' + AS#2 5'-gaa gag aag tcc att tgt atc ttg-3'] respectively; and subsequently used to generate expression plasmids

GalK with a 6His-extension at the C-terminal (pET28a:At3g06580#3.1) and Sloppy with a

6His-fusion at its N-terminal (pET28b:At5g52560# a73f/2#2).

Site directed mutagenesis - For mutagenesis, pET28b:At3g10700#11.3 encoding wild- type AtGalAK was used as template in the PCR-mediated QuikChange® XL site directed mutagenesis kit (Stratagene). The primer pairs 5'-gt cct tta gga gAG cac att gat cac cag gg-3' and

5'-cc ctg gtg atc aat gtg CTc tcc taa agg ac-3' (mutagenized nucleotide residues are indicated in upper case) were used to mutate codon 41A to E of GalAK (abbreviated as A41E); for mutation of Y250F of GalAK, the mutagenize primers were 5'-c aac cca gga tTt aat ctg cga gtt tct gag tg-3' and 5'-c tcg cag att aAa tcc tgg gtt ggt ggt caa cg-3'; A368S mutation of GalAK was generated with the primer pairs 5'-ga ttc agc ggt Tca ggt ttc agg gga tgt tg-3' and 5'-ct gaa acc tgA acc gct gaa tct agc tcc-3'.

For site directed mutagenesis of GalK, pET28a:At3g06580#3.1 was used as template.

The mutagenize primer pairs 5'-ctg ata gga gCg cac att gac tat gaa gga tac-3' and 5'-gtc aat gtg cGc tcc tat cag att cac tct tcc-3' were used to mutate codon 62E to A of GalK; for mutation of

Y262F in GalK, the mutagenize primers were 5'-cg gct gct aag aat tTc aat aac agg gtc gtt g-3' and 5'-cct gtt att gAa att ctt agc agc cgt gac cg-3'; to generate GalK A437S mutation, the primers were 5'-ga ctg acc gga Tct gga tgg ggc ggt tgc-3' and 5'-gcc cca tcc agA tcc ggt cag tct tgc tcc-3'; to generate GalK S206G mutation, the primers were 5'-t gga aca caa GGt ggt ggg atg gac cag gc-

3' and 5'-g gtc cat ccc acc aCC ttg tgt tcc aat gtg tc-3'. Each 25 µl site directed mutagenesis reaction consists of 10 ng plasmid, 65 ng primer pairs, 2.5U PfuTurbo DNA polymerase and

Stratagene supplied buffer and proprietary nucleotides. Reactions were carried out in a thermocycler using one cycle of 95 ºC for 0.5 min, followed by 12 cycles each (95 ºC, 0.5 min;

55 ºC, 1 min; 68 ºC for 7 min). Reaction products were treated with 10U Dpn I, subsequently plasmids were transformed to Dh5α competent cell and positive clones were selected on LB agar supplemented with kanamycin (50 µg/ml). DNA sequencing confirmed the intended single or

double based substitution of each mutant constructs, and the resulting plasmids were named pET28b:AtGalAK#11.3A41E; pET28b:AtGalAK#11.3Y250F; pET28b:AtGalAK#11.3A368S; pET28a:AtGalK#3.1E62A; pET28a:AtGalK#3.1Y262F; pET28a:AtGalK#3.1A437S; pET28a:AtGalK#3.1S206G. Expression of wild-type and mutant genes is under T7 promoter, thus each plasmid was transformed to BL21(de3)plysS-derived E. coli strain (Novagen) for gene expression.

Protein expression and purification - E. coli cells harboring each plasmid construct or an empty vector were cultured for 16 h at 37 ºC in LB medium (20 ml) supplemented with kanamycin (50 μg/ml) and chloramphenicol (34 μg/ml). A portion (8 ml) of the cultured cells was transferred into fresh LB liquid medium (250 ml) supplemented with the same antibiotics, and the cells then grown at 37 ºC at 250 rpm until the cell density reached OD600 = 0.6. The cultures were then transferred to 18 ºC and gene expression was induced by the addition of isopropyl β-D-thiogalactoside to a final concentration of 0.5 mM. After 24 h growth while shaking (250 rpm), the cells were harvested by centrifugation (6,000  g for 10 min at 4 ºC), resuspended in lysis buffer (10 ml 50 mM Tris-HCl, pH 7.5, containing 10% (v/v) glycerol, 50 mM NaCl, 1 mM MgCl2, 0.5 mM EDTA, 1 mM DTT and 0.5 mM phenylmethylsulfonyl fluoride) and lysed in an ice bath by 24 sonication cycles each (10-sec pulse; 20-sec rest) using a

Misonix S-4000 (Misonix incorporated, Farmingdale, New York) equipped with microtip probe.

The lysed cells were centrifuged at 4 °C for 30 min at 20,000  g, and the supernatant (termed s20) was recovered and kept at -20 °C.

His-tagged proteins were purified on a column (10 mm id  150 mm long) containing Ni-

Sepharose (2 ml, Qiagen) equilibrated with 50 mM sodium-phosphate, pH 7.5, containing 0.3 M

NaCl. The bound His-tagged protein was eluted with the same buffer containing increasing

concentrations of imidazole. The fractions containing GalK or GalAK activities were supplemented with equal volume of 50% (v/v) glycerol and stored in aliquots at -80 ºC, whereas

Sloppy was stored in 40% glycerol (v/v). The concentration of proteins was determined using the

Bradford reagent using bovine serum albumin (BSA) as standard.

The molecular weights of the recombinant proteins were estimated by size-exclusion chromatography using a Waters 626 LC HPLC system equipped with a photo diode array detector (PDA 996) and a Waters Millennium32 workstation. Separate solutions (0.5 ml) of

GalK, GalAK, and Sloppy or a mixture of standard proteins [10 mg each of alcohol dehydrogenase (150 kDa), ovalbumin (48.9 kDa), ribonuclease A (15.6 kDa), and cytochrome C

(12.4 kDa)] were separately chromatographed at 1 ml min-1 on a Superdex75 column (10 mm id

 300 mm long, GE) equilibrated with 0.1M sodium phosphate, pH 7.6, containing 0.1 M NaCl.

The eluant was monitored at A280nm and fractions collected every 20 sec. Fraction containing enzyme activity were pooled and kept at -80 ºC. To confirm the amino acid sequence of each recombinant protein, the purified recombinant proteins were fragmented with trypsin and the resulting peptides sequenced as described (Gu and Bar-Peled, 2004) by electrospray MS-MS analyses (data not shown).

Enzyme Assays and NMR spectroscopic Analysis - Three assays were performed:

Formation of a sugar-1-P from a sugar and ATP - GalAK reactions (50 µl final vol) were performed in 100 mM Tris-HCl, pH 7.6 (or other buffer when indicated), containing 5 mM

MgCl2, 2 mM ATP, 2 mM D-GalA and 1.5 µg recombinant GalAK. Reactions were kept at 37 ºC for up to 10 min, and then terminated by adding an equal volume of chloroform. After vortexing

(30 seconds) and centrifugation (12,000 rpm for 5 min, at room temperature), the upper aqueous phase was collected and chromatographed on a Q15 anion-exchange column (2 mm id  250 mm

long, Amersham) or a TSK-DEAE-5-pw column (7.5 mm id  75 mm long, Biorad) using an

Agilent Series 1100 HPLC system equipped with an autosampler, diode-array detector, and a

Corona (ESA, Chelmsford, MA) and ChemStation software as described (Gu and Bar-Peled,

2004).

Formation of UDP-sugars from monosaccharides using a kinase, UDP-sugar pyrophosphorylase (Sloppy) and yeast PPase - The coupled assays were performed in conjunction with HPLC, to easily detect the formation of reaction products by UV. The coupled assay has two steps. Unless otherwise noted see below for GalAK enzyme characterization, the kinase reaction (50 µl vol) conditions were as described above for recombinant GalAK for up to

10 min. The reactions were terminated in a boiling bath (1 min at 100 ºC). After cooling, 2 mM

UTP, 1 µg purified recombinant UDP-sugar pyrophosphorylase (Sloppy, At5g52560) and 1 unit of yeast PPase (Sigma) were added (final volume 60 µl). Sloppy is a reversible enzyme (i.e. can also convert PPi + UDP-sugar to sugar-1-P + UTP), thus to measure kinase activity, we included yeast PPase to drive the coupled-reaction forward towards formation of UDP-GalA by depleting the PPi to 2  Pi. After 15 min at 37 ºC, the reactions were terminated by adding chloroform and the products were analyzed by anion-exchange chromatography as described above. Nucleotides and nucleotide sugars were detected by their UV absorbance using a Waters or Agilent photodiode array detector. The maximum absorbance for uridine-nucleotides and UDP-GalA was 261.8 nm in ammonium formate. The peak area of analytes was compared to calibration curves of internal standard (UDP-Gal or UDP-GalA).

Real-time 1H-NMR analysis of sugar phosphorylation - Two NMR-based assays were performed: phosphorylation assay and coupled assay. Phosphorylation reactions were performed in 50 mM sodium phosphate, pH 7.5, in a mixture of D2O:H2O (8:1 v/v, 180 l) containing 5

mM MgCl2, 2 mM ATP, 2 mM GalA and 1.5 µg recombinant GalAK supplied in H2O-buffer. In early experiments GalA was incubated without enzyme in the reaction buffer to monitor at equilibrium the ratio between the  and  configuration of the anomeric sugar prior adding the kinase and ATP. In the coupled assay, 2 mM UTP, 1 µg recombinant Sloppy and 1U of yeast

PPase were also added. Immediately upon addition of enzyme, the reaction mixture was transferred to a 3 mm NMR tube. Real-time 1H NMR spectra were obtained using a Varian Inova

600MHz spectrometer equipped with a cryogenic probe. Data acquisition was not started until approximately 2 minutes after the addition of enzyme to the reaction mixture due to spectrometer set-up requirements (shimming). Sequential 1D proton spectra were acquired over the course of the enzymatic reaction. All spectra were referenced to the water resonance at 4.765 ppm downfield of 2,2-dimethyl-2-silapentane-5-sulphonate (DSS). Processing of the data as covariance matrices was performed with Matlab (The Mathworks, Inc.).

Enzyme properties of GalK and GalAK - To characterize the properties of recombinant

GalK and GalAK, the kinase activities were tested under a variety of conditions: with various buffers, at different temperatures, different ions, or with different potential inhibitors. For the optimal pH experiments, 1.5 µg recombinant enzyme (GalK or GalAK) was first mixed with 5 mM MgCl2, 2 mM ATP and 100 mM of each individual buffer (Tris-HCl, phosphate, MES,

MOPS, or HEPES). The optimal pH assays were initiated after the addition of specific sugar

(Gal or GalA). Inhibitor assays were performed under standard assay conditions except for the addition of various additives (sugars, nucleotides) to the reaction buffer. These kinase assays were incubated for 10 min at 37oC, and were subsequently terminated by heat (1 min at 100 ºC).

After cooling, a coupled assay was used to detect the relative activity of the kinase by adding 2 mM UTP, 1 µg purified recombinant UDP-sugar pyrophosphorylase (Sloppy) and 1 unit of yeast

PPase (Sigma). The amount of UDP-sugar formed was calculated from HPLC UV spectra. Early work in our lab has shown that the amount of Sloppy used in the assays is sufficient to provide full conversion of sugar-1-P to UDP-sugar under the variety of pH and inhibitors tested.

Similarly, the yeast PPase is fully active under these broad experimental conditions. For the experiments aimed at defining the optimal temperature the GalK or GalAK assays were performed under standard assay conditions except that reactions were incubated at different temperature for 10 min. Subsequently, the kinase activities were terminated (100 oC) and the relative kinase activity was measured by the coupled assay at 37 oC. For the experiments aimed at determining if the kinases required metals, the GalK or GalAK assays were performed with

ATP, specific sugar (GalA or Gal) with a variety of ions. After 10 min at 37 oC incubation the kinase assay was terminated by heat. After cooling, the amount of Mg2+ was adjusted to 5 mM,

Sloppy, UTP, and yeast PPase were added and the coupled-assay was carried out at 37 ºC. The amount of UDP-sugar formed was calculated from HPLC UV spectra.

For the experiments aimed at determining the ability of the kinases to utilize other sugars,

GalK or GalAK assays were performed under standard assay conditions except for substitute the sugar (Gal or GalA) with different sugars (for example, mannose, fucose etc). These kinase assays were incubated for 60 min (unless otherwise mentioned) at 37 oC, and were subsequently terminated by heat (1 min at 100 ºC). After cooling, reactions were separated by Q15 anion- exchange HPLC column (as above) and sugar or suagar-1-P peaks were monitored using Corona detector. The amount of sugar-1-P formed was calculated from HPLC Corona spectra.

Kinetics - The catalytic activity of GalAK was determined using the enzyme-coupled assay at 37 ºC for 5 min using Tris-HCl pH 7.6, containing MgCl2 (5 mM), variable concentrations of ATP (40 µM to 2 mM) with a fixed concentration of sugar (2 mM), and 0.27µg

recombinant GalAK (6 pmol). In a separate series of experiments reactions were performed with a fixed amount of ATP (2 mM) and variable concentrations of sugar (20 to 400 µM). GalK kinetic assays were performed for 5 min in the same buffer as above, using 0.5 µg recombinant enzyme (10 pmol), with variable concentrations of ATP (40 µM to 8 mM) and sugar (2 mM) or with ATP (2 mM) and variable concentrations of sugar (40 µM to 8 mM). Enzyme velocity data of the amount (µM) of phosphorylated-sugar produced per second per µg enzyme, as a function of substrate concentrations was plotted. The Solver tool (Excel version 11.5 program) was used to generate best-fit curve calculated by nonlinear regression analyses, and for calculation of Vmax and apparent Km.

Quantitative PCR of AtGalAK and AtGalK transcripts in different tissues - RNA was extracted from roots, leaves, flowers, young siliques and various sections of the stem (the lower region non elongating bottom stems, the elongating middle stem region, and the upper stem regions), of six-week-old Arabidopsis plants using RNeasy Plant Mini Kit (Qiagen). Two μg of total RNA was reverse transcribed using oligo-dT as primer in 20-μl reactions (Gu and Bar-

Peled, 2004) and the resulting cDNA was then diluted with sterile water to 200 µl prior for qPCR reaction.

qPCR reaction (final volume of 20 μl) was carried out using 2 μl of diluted cDNA, 0.5

μM each primer and 10 μl of 2  iQTM SYBR Green Supermix (BIO-RAD 170-8880) consisted of polymerase, dNTPs, company propriety buffer and SYBR Green I. The qPCR reactions, in 96- well plate format, were carried using iCycler Real-Time PCR System (Bio-Rad). The thermal cycle conditions were: 95 ºC for 3 min, and 40 cycle repeats of (95 ºC for 10 sec and 56 ºC for

30 sec), followed by 95 ºC for 1 min, and 55 ºC for 1 min. The GalAK primers for qPCR reactions were: qS#1 5'-gta tct ggg tct gcg gaa tg-3' and qAS#1 5'-caa gct cgt ggt cca aag tc-3';

for GalK, qS#2 5'-ggt gct tct ccc caa ctc tt-3' and qAS#2 5'-gaa tag cca tcg gca aca ct-3'; the qPCR primes for Actin, used as reference gene control, were qS#3 5'-ggt aac att gtg ctc agt ggt gg-3' and qAS#3 5'-aac gac ctt aat ctt cat gct gc-3'.

RESULTS

Identification, cloning and characterization of Arabidopsis GalAK - Galactokinase

(GalK) from E. coli, yeast, human, and Pyrococcus furiosus (Hartley et al., 2004), have conserved folds and amino acid domains. GalK is a member of the GHMP super family that includes galactokinase (G), homoserine kinase (H), mevalonate kinase (M), and phosphomevalonate kinase (P). All these proteins contain three conserved motifs. The motif,

PGRVNLIG(AE)HxDY, at the N-terminal region of the protein, is involved in sugar binding.

The motif GL(GS)SSA that is located ~100 aa downstream of the N-terminal motif is involved in binding the α and -phosphate groups of ADP (Hartley et al., 2004) whereas the motif,

GAGxG, located at the C-terminal region is involved in catalysis and transfer of the  phosphate to the sugar (Thoden et al., 2005).

To identify plant enzymes that contribute to the flux of monosaccharides into nucleotide- sugar metabolism, we searched the Arabidopsis genomic database

(http://www.ncbi.nlm.nih.gov/) for candidate GlcA-1-P and GalA-1-P kinases. BLAST analysis and sequence alignment of human, E. coli and yeast GalK identified two Arabidopsis genes

(At3g06580 and At3g10700).

At3g06580 encodes a protein (abbreviated GalK) that has 27% and 29% aa identity with the human and E. coli GalK respectively and has been shown to compliment a galk yeast mutant

(Kaplan et al., 1997). However, the substrate specificities of the Arabidopsis enzyme were not determined. By contrast, At3g10700 encodes a protein (abbreviated herein GalAK) that has 26%

and 27% aa sequence identity with human and E. coli GalK, respectively. Phylogenetic analysis

(Fig. 2.1A) indicated that the GalAKs from different plant species such as Arabidopsis, rice

(OsGalAK, Oryza sativa Os04g0608100), and poplar GalAK (PtGalAK, Populus trichocarpa

JGI:427630) are distinguished from GalK in all other species.

Arabidopsis GalK and GalAK have only 22% aa identity. The conserved glutamate (E62) found in GalK proteins (Fig. 2.1B), is replaced by alanine in Arabidopsis, rice, and poplar

GalAKs. We reasoned that in GalAK the non-polar alanine residue (A41) may accommodate charged glycoses such as GalA or GlcA whereas a glutamate residue would repulse the carboxyl group of these uronic acids (Fig. 2.1B). We cloned and expressed the Arabidopsis proteins in E. coli to determine if At3g10700 encodes a uronic acid kinase and to determine if At3g06580 encodes a protein with only GalK activity.

FIGURE 2.1. Sequence alignments and phylogenetic relationships of GalAK and GalK from different organisms.

A. Phylogenetic relationships of GalAK and GalK in different species. Protein sequences (see name and gene accession numbers below) were aligned and analyzed using Muscle 3.7 software

(Edgar, 2004) and the phylogenetic trees were created using MrBayes 3.1.2 software

(Huelsenbeck and Ronquist, 2001). Branch support values (more than 50%) are shown. The bar represents 0.4 protein substitutions per site.

B. The conserved motifs of the putative GalAK sugar-binding domain, ATP-binding and

catalytic domains were aligned with GalK from different species. Note: the conserved AA Glu

(E) in sugar binding motif of GalK, is altered in GalAK to Ala (A).

Arabidopsis thaliana GalA-kinase (AtGalAK, At3g10700), Oryza sativa GalAK (OsGalAK,

Os4g51880), Populus trichocarpa GalAK (PtGalAK, JGI:427630), Vitis vinifera GalAK

(VvGalAK, GSVIVT00007137001), Sorghum Bicolor GalAK (SbGalAK, Sb06g027910),

Selaginella moellendorffii GalAK (SmGalAK, JGI:82393); Arabidopsis thaliana Gal-kinase

(AtGalK, At3g06580), GalK from Saccharomyces cerevisiae (ScGalK1, 2AJ4 and ScGalK3,

NP_010292), GalK from Homo Sapiens (HsGalK1, 1WUU and HsGalK2, 2A2D), Escherichia coli (EcGalK, NP_308812), Chlamydomonas Reinhardtii (CrGalK1 XP_001701527 and

CrGalK2, XP_001691828), Physcomitrella patens GalK (PpGalK1, XP_001767042 and

PpGalK2, XP_001756018), Oryza sativa (OsGalK, Os03g0832600), Populus trichocarpa

(PtGalK1, XP_002314964 and PtGalK2, XP_002312315), Vitis vinifera (VvGalK,

GSVIVT00033170001), Sorghum Bicolor (SbGalK, Sb01g002480), Selaginella moellendorffii

(SmGalK, JGI:144964).

A highly expressed protein band (50 kDa) was detected after SDS-PAGE analysis of E. coli cells expressing GalK (Fig. 2.2A, lane 4, marked by the upper arrow). Similarly a band (46 kDa) was observed in cell overexpressing GalAK (Fig. 2.2A lane 2). GalK and GalAK were column-purified and gave distinct 50 and 46 kDa protein bands respectively (Fig. 2.2A lane 5 and lane 3) which agree with the calculated mass of the translated gene products.

Preliminary experiments demonstrated that GalAK converts GalA and ATP to GalA-1-P and ADP (data not shown). We then developed a sensitive two-step coupled assay for the quantification of GalAK activity. First the GalAK was reacted with ATP and GalA and the enzyme was then heat-inactivated. The amount of GalA-1-P produced was quantified by measuring the amount of UDP-GalA formed in the presence of UDP-sugar pyrophosphorylase

(abbreviated herein “Sloppy”). The enzymatic products of this second reaction were analyzed by anion-exchange chromatography (Fig. 2.2B, panel 1) which showed the presence of two new products with retention time characteristic of ADP (16.3 min overlap with UTP) and UDP-GalA

(15.2 min). The peak marked #1 (Fig. 2.2B, panel 1) was collected and its structure was

1 confirmed by H NMR spectroscopy as UDP-α-D-GalA (Fig. 2.2B, panel 4; Supplemental

Figures 2.1A and B) using authentic standards.

FIGURE 2.2. Expression, enzymatic activities of recombinant Arabidopsis GalA-kinase

(GalAK), Gal-kinase (GalK) and identification of reaction products.

A. SDS-PAGE of total soluble protein isolated from E. coli cell expressing Arabidopsis recombinant GalAK (lane 2), control empty vector (lane 6), or Arabidopsis GalK (lane 4); and of

Nickel-column purified recombinant GalAK (lane 3) and GalK (lane 5).

B. HPLC chromatogram of GalAK coupled-based assay. A distinct UDP-GalA peak (marked by arrow #1, in Panel 1) is detected in GalAK reaction but not in control (empty vector control cells,

Panel 2). The peak eluted at 15.2 min (Arrow #1, in panel 1) was collected and confirmed after analysis by 1H-NMR (panel 4) as UDP-α-GalA. For detailed proton NMR spectra of GalAK enzymatic product, see Supplemental Figures 2.1A-B. The HPLC peaks in B (panels 1-3) are

ATP (18 min), UTP+ADP (16.3 min), UDP-GalA (15.2 min). The minor peak at 14.8 min is

AMP contamination.

C. HPLC chromatogram of GalK coupled-based assay. A UDP-Gal peak (marked by arrow #2, in Panel 5) is detected in reaction contained GalK but not in empty vector control cells (Panel 6).

The peak eluted at 12.2 min (Arrow #2, in panel 5) was collected and confirmed after analysis by

1H-NMR (panel 8) as UDP-α-Gal. For detailed proton NMR spectra of the enzymatic product, see Supplemental Figures 2.2A-B. The HPLC peaks in C (panels 5-7) are ATP (18 min),

UTP+ADP (16.3 min), UDP-Gal (12.2 min). The minor peak at 14.8 min is AMP contamination.

Characterization and properties of GalAK - Crude recombinant GalAK is stable when stored at -20 ºC. However, the purified enzyme lost >90% of its initial activity within one week of storage at this temperature. We found that GalAK could be stabilized for several months at -

80 ºC when diluted with glycerol (to 25% v/v) and flash-frozen in liquid nitrogen.

GalAK requires Mg2+, although other divalent cations including Mn2+ and Ca2+ can substitute for magnesium. GalAK activity was 61% in the presence of Ca2+, but reduced by

~32% in the presence of Mn2+ (Table 2.1). GalAK activity is completely abolished in the presence of EDTA. The recombinant GalAK is active between pH 3.3 and pH 9.5 (Fig. 2.3, panel A), with maximum activity at pH 7.5 to 7.8 in Tris, or phosphate buffers. Activity was reduced when reactions were performed in Hepes, Mops or Mes buffers.

TABLE 2.1. GalAK and GalK require metals for activity.

Relative GalAK Relative GalK Metal (5mM) Activity % Activity %

MgCl2 100 100

MnCl2 32 42

CaCl2 61 13

CuCl2 0 0

ZnSO4 0 0

EDTA 0 0

GalAK or GalK enzyme were mixed with additive (metal, EDTA, or water control) for 10 min on ice. Subsequently, ATP and appropriate sugar were added and kinase-assay was carried out under standard conditions. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

FIGURE 2.3. The effects of buffer and pH on GalAK and GalK activities.

The activity of GalAK (Panel A) and GalK (Panel B) were analysed at different buffers (Tris-

HCl, phosphate, MES, MOPS, HEPES) at different pH. Except for the buffer, the kinase reactions were performed under standard conditions. Each value is the mean of triplicate reactions, and the values varied by no more than 5%.

GalAK is active between 4 ºC and 42 ºC, with maximum activity at 37 ºC (Table 2.2).

No GalAK activity was obtained when assays were performed at 55 ºC. Pre-incubating the recombinant GalAK with GalA or ATP at different temperature (55 to 75 ºC, for 5 min), prior to performing the activity assays at 37 ºC did not stabilize GalAK enzymatic activity.

TABLE 2.2. The effect of temperature on GalAK and GalK activities.

Relative GalAK Relative GalK Temperature ºC Activity % Activity %

4 24 26

25 53 75

30 88 101

37 100 100

42 80 100

55 0 1

65 0 0

The enzymatic reactions were performed under standard conditions for each enzyme activity except for the reaction temperature. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

To investigate the sugar specificity of GalAK, the enzyme was reacted for up to 60 min with ATP and different monosaccharides. No phosphorylation occurred with D-Gal, D-Glc, D-

Fru, D-Man, D-gluconic acid, D-Xyl, D-GalNac, D-GlcNac, D-glucosamine, D-Ara, L-Ara, L-Rha, and L-Fuc. These monosaccharides (at 4 mM) also had no effect on GalAK activity in the presence of GalA. Thus we concluded that the recombinant GalAK is specific for GalA.

We next investigated the nucleotide triphosphate phosphorylation donor specificity of

GalAK. CTP, GTP, ITP, TTP or UTP are not substrates for GalAK. The recombinant GalAK was also not inhibited by these tri-phosphate-nucleotides (4 mM). By contrast, enzyme activity was reduced by 60% in the presence of 4 mM ADP (Table 2.3). Other nucleotide diphosphates tested had no effect on activity. The specificity for GalA as substrate and ATP as nucleotide phosphate donor indicates the unique function for this GalA kinase. It is therefore likely that free

GalA residues would be recycled back into the nucleotide sugar pool by GalAK and Sloppy.

TABLE 2.3. The effect of potential inhibitors on GalAK and GalK activities.

Nucleotide Relative GalAK Relative GalK inhibitor (4mM) Activity % Activity %

NAD+ 100 111

UTP 105 112

CTP 100 112

GTP 104 107

PPi 109 110

+ NADP 93 95

ITP 104 110

ADP 60 77

UDP 105 111

TTP 101 96

AMP 101 95

Control 100 100

Inhibitors (at 4 mM), or control (water) were mixed with GalAK or GalK in 100 mM Tris-HCl, pH 7.6 for 10 min on ice prior to performing the enzymatic reaction under standard conditions for each enzyme activity. Each value is the mean of triplicate reactions, and the values varied by no more than ±10%.

Real-Time NMR analysis of GalAK - Guyett et al. (Guyett et al., 2009) have demonstrated the ability of 1H-NMR spectroscopy to monitor enzymatic reactions in real time.

This procedure can be used to monitor the dynamics of an enzymatic assay, the appearance and disappearance of intermediates, and the detection of unstable products. Thus we used real-time

1H-NMR spectroscopy to follow the reactions catalyzed by GalAK (Fig. 2.4, panel A;

Supplemental Figures 2.5A, B and C).

FIGURE 2.4. Real-Time 1H NMR based GalAK and GalK assays.

Individual enzyme (GalK or GalAK) was mixed with sugar (Gal or GalA), buffer and ATP and reaction was placed in NMR tube. Approximately, 2 min after enzyme addition and NMR shimming, NMR data were collected (bottom trace). Progression of enzyme activity at 15 and 30 min are shown. At 30 min, UTP and Sloppy were added to NMR tube and NMR spectrum were continued to be collected. The sugar anomeric region of the 1H-NMR spectrum is shown. Panel

A: NMR Peak annotation of coupled GalAK-Sloppy assay are: peak 1 (α-GalA at 5.27 ppm), peak 2 (α-GalA-1-P at 5.52 ppm), peak 3 (UDP-α-GalA at 5.65 ppm). Panel B: NMR Peak annotation of coupled GalK-Sloppy assay are: peak 4 (α-Gal, at 5.24 ppm), peak 5 (α-Gal-1-P at

5.47 ppm), peak 6 (UDP-α-Gal at 5.61 ppm).

Before enzyme was added, two NMR signals of GalA can be observed corresponding to the anomeric protons α-D-GalA (5.27 ppm) and - form of GalA (4.55 ppm) (see Supplemental

Figure 2.5A). The ratio of α to -GalA form is constant (~ 30%/70%). GalAK does not have mutarotase activity (data not shown). When GalAK is supplemented with ATP an immediate conversion of the α-D-GalA (5.27 ppm) into α-D-GalA-1-P (5.52 ppm) was observed (see time course of enzymatic progression in Fig. 2.4A, and in Supplemental Figure 2.5C). Addition of

UTP followed by recombinant UDP-sugar pyrophosphorylase (Sloppy), results in the conversion of GalA-1-P to UDP-α-D-GalA indicated by the shift (5.65 ppm) (Fig. 2.4A). During the kinase reaction GalAK converts ATP to ADP indicating that the -phosphoryl of ATP is transferred to galacturonic acid.

Arabidopsis At3g06580 is galactokinase (GalK) - To determine the activity and specificity of recombinant Arabidopsis GalK we performed experiments similar to those described for GalAK. The HPLC-based assays show that recombinant GalK in the presence of

ATP readily converts D-Gal to Gal-1-P and in the presence of Sloppy to UDP-Gal (see Fig. 2.2C panel 5 arrow #2). The UDP-Gal peak (12.2 min) was collected and the product was confirmed

1 by H-NMR spectroscopy (Fig. 2.2C panel 8; Supplemental Figures 2.2A and B) as UDP-α-D-

Gal. GalK requires divalent metals for activity. However, unlike GalAK, 42% and 13% GalK activity were observed with Mn2+ and Ca2+, respectively when compared with Mg2+ (Table 2.1).

GalK has highest activity between 30 ºC and 42 ºC (Table 2.2), with an optimal pH of 7 to 8 at phosphate and Tris buffers. GalK activity was reduced somewhat when Tris was replaced by

Hepes or Mops buffers (Fig. 2.3, panel B).

The GalK sugar specificity was also examined. In addition to D-Gal, the recombinant enzyme efficiently converts 2deoxy-D-Gal (2dGal) to 2dGal-1-P and ADP. GalK did not phosphorylate GalA, GlcA D-Glc, D-Fru, D-Man, D-Xyl, D-GalNc, D-GlcNAc, D-glucosamine, D-

Ara, L-Ara, L-Rha, and L-Fuc. Thus, we suggest that Arabidopsis At3g06580 encodes a GalK.

Purified GalK is less stable than GalAK, and within 2 days at -20 ºC the enzyme lost all its activity. However, the purified recombinant protein was stable and retained activity when stored at -80 ºC in the presence of glycerol (final 25%v/v). Interestingly, E. coli cultures containing the GalK expression plasmid and producing large amounts of GalK protein did not grow to cell densities above OD600 = 0.8. This phenomenon was not observed with E. coli expressing GalAK or Sloppy.

Comparing kinetic and catalytic properties of GalK and GalAK - Kinetics analyses of

GalAK and GalK are summarized in Table 2.4. The apparent Km values for GalAK were 71 µM

-1 -1 -1 -1 (GalA) and 195 µM (ATP), with a Vmax of 1 µM s µg (GalA) and 0.5 µM s µg (ATP), and

-1 -1 kcat/Km values (s mM ) were 36 (GalA) and 6.3 (ATP). The apparent Km values for GalK were

-1 -1 -1 -1 701 µM (Gal) and 701 µM (ATP), with Vmax values of 3.5 µM s µg (Gal) and 3.3 µM s µg

-1 -1 (ATP), and kcat/Km (s mM ) were 12.3 (Gal) and 11.8 (ATP). The Km for human GalK is 120

µM (Gal) and 350 µM (ATP). Thus, our kinetic data is comparable with kinetic analyses of other sugar kinases.

TABLE 2.4. Enzyme kinetics of GalAK and GalK.

Vmax kcat/Km kcat/Km Km (Sugar) Km (ATP) Vmax (ATP) (Sugar) (Sugar) (ATP) µM µM µM s-1µg -1 µM s-1µg -1 s-1 mM-1 s-1 mM-1

AtGalAK 70.8 195 1.0 0.5 36 6.3

AtGalK 701 701 3.5 3.3 12.3 11.8

HsGalK 120 350 81.2 81.2 568 195

EcGalK 700 100 14 14 13.8 96.7

ScGalK 600 150 55.8 55.8 89.9 360

PfGalK 270 8 43.2 41.9 105 3439

GalAK activity was measured with varied concentrations of D-GalA (0.02–0.4 mM) and ATP

(0.04–2.0 mM), after 5 min at std conditions. GalK activity was measured with varied concentrations of D-Gal (0.04-8mM) and ATP (0.04–8.0 mM), after 5 min at std conditions.

Enzyme velocities were plotted and Solver software was used to generate best-fit curve and for calculation of Vmax and apparent Km. Each value is the mean of quadruple reactions, and the values varied by no more than ±10%. The kinetics of GalK obtained from H. sapiens, E. coli, S. cerevisiae, and P. furiosus (Verhees et al., 2002), are displayed for comparison.

Substrate specificity of site-directed mutated GalAK and GalK kinases - Sequence comparisons of functional GalK and homologous proteins in different species indicated that these proteins all have a glutamic acid residue (E62), that forms a hydrogen bond with the C6 hydroxyl group of galactose as depicted by structural studies (Thoden and Holden, 2003). On the other hand, in genes encoding GalAK proteins from Arabidopsis and other plant species, the glutamate at position 41 is alanine (Fig. 2.1B). Thus, to investigate if this aa is critical for sugar binding we altered these aa residues in both GalAK and GalK using site-directed mutagenesis

A41E (see Table 2.5). GalAK did not phosphorylate either D-GalA or D-GlcA, suggesting that the

Glu-substitution is not compatible with the carboxyl group of uronic acid. Somewhat unexpectedly GalAKA41E did not phosphorylate Gal or any other sugar tested even with extended reaction times.

TABLE 2.5. The effect of selective mutation on GalAK and GalK activities.

Enzyme Relative Sugar Mutation Enzyme Substrate Site Activity %

D-GalA 100 GalAKWT D-GlcA 0

D-GalA 0 GalAKA41E D-GlcA 0

D-GalA 90 GalAKY250F D-GlcA 50

D-GalA 90 GalAKA368S D-GlcA 0

D-Gal 100 GalKWT 2-d-Gal 100

D-Gal 0 GalKE62A 2-d-Gal 0

D-Gal 80 GalKY262F 2-d-Gal 57

D-Gal 60 GalKA437S 2-d-Gal 40

D-Gal 95

GalKS206G 2-d-Gal 85

GalNac 30

Selected amino-acid of wild-type GalA and Gal proteins were altered by site directed mutagenesis (see Methods). The recombinant proteins (wt and mutants) were purified and

GalAK and GalK activities were measured under standard conditions but for 15 min. The relative activities of mutant were compared to wt (100%), and the value is the mean of duplicate reactions, and the values varied by no more than ±5%.

Analyses of the human galk gene of GALK-deficient individuals with galactosemia, has provided information on those aa residues that have a critical role in catalysis (Thoden and

Holden, 2003). Based on these studies we selected Ala368 and Ala437 for mutation. GalAKA368S and GalKA437S reduced phosphorylation activities by only ~40% and 10%, respectively when compared with native enzyme.

We next sought to convert GalAK to a GlcAK by changing aa residues we believed to interact with the C-4 hydroxyl of GalA. This assumption was based on mutation of GalK of E. coli (Hoffmeister and Thorson, 2004), which showed that a tyrosine mutation significantly changed enzyme substrate specificity. The tyrosine mutation of GalAK to phenylalanine

Y250F -1 -1 (GalAK ) resulted in phosphorylation of GlcA with kcat /Km values of 0.76 s mM .

Interestingly, the Y250F mutated enzyme was still active towards GalA but with reduced

-1 -1 catalytic efficiency kcat/Km decreased to 24 s mM (GalA), and increase Km 395 µM (GalA).

The ability of AtGalAKY250F to phosphorylate GlcA led us to explore how similar

Y262F mutations affect Arabidopsis GalK (Table 2.5). AtGalK failed to phosphorylate D-Glc but was still fully active toward Gal suggesting that other amino acids are involved in the interaction with the C-4 hydroxyl of monosaccharides. Based on mutation studies of E. coli GalK

(Hoffmeister and Thorson, 2004), and our GalAKY250F we identified Ser206 of AtGalK as a potential aa to be altered in accepting sugars with C-2 modifications. Indeed, converting Ser206 to Gly allowed AtGalKS206G to phosphorylate GalNac. However, converting the conserved Glu62 to Ala in AtGalK (GalKE62A) did not result in the phosphorylation of GalA.

Gene expression of GalAK and GalK in Arabidopsis - GalAK and GalK were shown by qPCR (see Fig. 2.5) to be expressed in roots, stems, leaves, flowers and young siliques. Stem and floral tissue had the highest expression of the mRNAs of GalK and GalAK. Interestingly, GalAK

expression was almost 1.5-fold higher in the elongating middle stem region than in the lower or upper stem region.

FIGURE 2.5. The expression of GalAK and GalK genes in different Arabidopsis tissues.

Total RNA obtained from different tissues was reverse transcribed and the relative amounts of

GalAK and GalK mRNAs compared with Actin mRNA were quantified by real-time qPCR.

Each value is the mean of triplicate reactions, and the values varied by no more than 10%.

DISCUSSION

We have described the cloning and biochemical characterization of a plant sugar kinase

(GalAK) that in the presence of ATP specifically phosphorylates α-D-GalpA to α-D-GalpA-1-P and ADP. GalA-kinase does not phosphorylate the -anomer of GalA under our assay conditions

(Supplemental Figure 2.5C). The selectivity of the GalAK for the α-anomeric configuration of monosaccharides is not unique to this kinase. The GalK form Human (Thoden et al., 2005) bacteria (24), yeast (22) as well as the Arabidopsis GalK are also selective for the α- configuration of the sugar (see Fig. 2.4B and Supplemental Figure 2.6C). At equilibrium in solution ~30% of galactose exists in the α configuration and ~64% in the -form (Supplemental

Figure 2.6A); and similar anomeric ratio is observed with GalAK (Supplemental Figure 2.5A).

In humans, a galactose mutarotase (GalM) converts the thermodynamically more stable -Galp to the bioactive α-form (Timson and Reece, 2003). Thus, GalM may be required to rapidly generate and maintain a sufficient pool of α-galactose for GalK-catalyzed conversion of Gal to

Gal-1-P. To date no Gal or GalA mutarotase activities have been identified in plants.

Nevertheless, the possibility can not be discounted that such mutarotases do exist. Indeed, plant genes with “mutarotase-like” motifs exist but their functions remain to be determined. There is also the possibility that the chelation of cations such as Ca2+ or Mg2+ by free GalA could non- enzymatically shift the anomeric equilibrium to the α-form. For example, Angyal et al. (Angyal et al., 1988) have shown that the α-anomeric configuration of GalA is stabilized in the presence of selected metal ions that form a complex with O5 and a carboxylate oxygen of the monosaccharide. Thus, plant cells may not require a GalA mutarotase if the Ca-α-GalA complex is indeed more stable than the -anomer of GalA. The fact that GalAK is active in the presence of calcium ion could explain that if such GalA-cation complex is made, it will serve a substrate

for the kinase.

Genes encoding GalK are highly conserved in Eukaryote, Bacteria, and Archaea. By contrast, genes encoding GalAK, appear to be restricted to land plants. Blast of the Arabidopsis

GalAK sequence with translated genomic data base we found other closely related sequences

(Fig. 2.1). A GalAK homolog was also found in the lycopodiophyte Selaginella moellendorffii (a member of one of the oldest extant vascular plant groups), but is not present in the moss

Physcomitrella patens, a plant that lacks a vascular system. Thus, the appearance of GalAK may have been one of the factors required for the development of the plant vascular system. However, we have also identified a sequence with low identity (36%) with GalAK in one of the two recently sequenced Micromonas. (http://www.ncbi.nlm.nih.gov/). A full understanding of the origins and evolution of GalAK will require complete genomic sequences from a wide range of embryophytes, streptophytes, and chlorophytes.

Many plant genes have undergone rounds of duplications thereby generating large gene families (Tang et al., 2008). However, GalAK exists as a single copy gene even in those plant species (e.g., poplar, maize and legumes) that are recent polyploids (Semon and Wolfe, 2007).

Interestingly, GalK has two gene copies in Populus, Physcomitrella, and Chlamydomonas, but in

Arabidopsis and rice genome for example, only one copy of GalK exist. The reason that GalAK exist as a single copy gene is not known. However, the possibility cannot be discounted that additional copies of the gene interferes with as yet unidentified aspects of plant metabolism.

Such a hypothesis can now be investigated by determining whether over-expression of GalAK has an adverse effect on plant growth and development.

UDP-GlcA 4-epimerases (Gu and Bar-Peled, 2004; Usadel et al., 2004) are believed to be the predominant enzymes involved in formation of the UDP-GalA that is required in the Golgi

for the synthesis of plant cell wall glycans. Here we have provided biochemical evidence showing that UDP-GalA is also generated from GalA in a series of reactions involving GalAK and Sloppy (a UDP-sugar PPase). Such data are consistent with early metabolic labeling studies demonstrating that plants can use exogenous GalA for the synthesis of wall polysaccharides

(Neufeld et al., 1961).

The observation that the GalAK gene is transcribed in all plant tissues suggests that most if not all plant cells have the ability to recycle GalA. Hydrolytic enzymes involved in pectin depolymerization including endo- and exo-galacturonases have been identified in numerous plant tissues including pollen tubes (Dubald et al., 1993), fruits (Pressey, 1984), seeds (Sitrit et al.,

1999), floral and leaf tissues (Hadfield and Bennett, 1998), and stem (Pressey and Avants, 1977).

Thus, free GalA formed by depolymerization of pectic polysaccharides present in plant organs and storage tissues may provide an immediate non-photosynthetic source of carbon that enters the sugar nucleotide pathway. The newly formed UDP-GalA can then be used for cell wall synthesis by the rapidly developing seedlings and during pollen tube germination and growth.

The primary wall of growing plant cells is itself a potential source of free GalA. The assembly and restructuring of this wall is believed to be required to allow plant cells to expand and grow. These processes may result in the generation of GalA that could then be recycled by specific GalA transporter into the cell as a substrate for GalAK (see a model in Fig. 2.6). While no plasmamembrane-localized transporters that import GalA into the cytosol have been identified, a large gene family with over 50 monosaccharide transporter-like genes exist in

Arabidopsis (Buttner, 2007), perhaps one of them is GalA transporter. Baluska et al. (Baluska et al., 2002) have suggested that pectic fragment can be endocytosed. Previous studies indicate that

exo-galacturonates are present in the cytosol of pollen (Barakate et al., 1993). Thus hydrolyses of pectins may provide an alternative route for the production of cytosolic GalA.

FIGURE 2.6. A model for GalA recycling during wall assembly and restructuring.

GalA hydrolytically released from wall polymers during wall assembly and restructuring or from storage tissues is transferred to the cytosol by plasmamembrane sugar transporter(s). GalA is subsequently phosphorylated by GalAK (1, At3g10700) and then converted to UDP-GalA in the cytosol by Sloppy (2, a UDP-sugar PPase, At5g52560). UDP-GalA can then be transferred by a

UDP-GalA transporter into the Golgi, where it can be directly incorporated by glycosyltransferases to glycans or enter the nucleotide-sugar interconversion pathway. UDP-

GlcA 4-epimerase (3, e.g., At2g45310) interconverts UDP-GalA and UDP-GlcA. UDP-GlcA is converted to UDP-xylose by UDP-GlcA decarboxylase (4, Uxs, e.g., At3g53520). UDP-xylose

4-epimerase (5, Uxe, e.g., At1g30620) interconverts UDP-xylose and UDP-arabinose.

The combined activity of GalAK and Sloppy results in the formation of UDP-GalA in the cytosol. Thus, a Golgi-localized UDP-GalA transporter (Fig. 2.6) will be required for this nucleotide sugar to be used in glycan synthesis. Previous studies have suggested that such a

UDP-GalA transporter exists since intact microsomes are able to take up exogenous UDP-14C-

GalA (Sterling et al., 2001).

In summary, our results demonstrate unambiguously that vascular plants have enzymes that convert GalpA to UDP-GalpA. This salvage pathway is likely common in every cell and has an important role in recycling GalA released from cell wall glycans during plant growth.

Elucidating the metabolic pathways that involve GalAK and the evolutionary origins of this plant-specific enzyme are major new challenges for plant scientists.

CHAPTER 3

IDENTIFICATION OF A NOVEL UDP-SUGAR PYROPHOSPHORYLASE WITH A

BROAD SUBSTRATE SPECIFICITY IN TRYPANOSOMA CRUZI2

2This research was originally published in Biochemical Journal. Reprinted here with permission of publisher. Ting Yang and Maor Bar-Peled. Biochem J. 2010 429(3):533-43. © The

Biochemical Society

ABSTRACT

The diverse types of glycoconjugates synthesized by trypanosomatid parasites are found to be unique compared to the host cells. These glycans are required for the parasite survival, invasion or evasion of the host immune system. The synthesis of those glycoconjugates requires a constant supply of nucleotide-sugars, yet little is known how these NDP-sugars are made and supplied. Here we report a functional gene from Trypanosoma cruzi that encodes nucleotidyltransferase, capable of transforming different types of sugar-1-Ps and NTP into NDP- sugars. In the forward reaction the enzyme catalyzes the formation of UDP-glucose, UDP- galactose, UDP-xylose and UDP-glucuronic acid, from respective monosaccharide 1-phosphates in the presence of UTP. The enzyme could also convert Glc-1-P and TTP to TDP-Glc, albeit at lower efficiency. The enzyme requires divalent ions (Mg2+ or Mn2+) for activity and is highly active between pH 6.5-8.0, and at 30-42 ºC. The apparent Km values for the forward reaction were 177 µM (Glc-1-P) and 28.4 µM (UTP) respectively. Our report of the unusual parasite enzyme having such broad substrate specificities suggests an alternative pathway that might play an essential role for nucleotide-sugar biosynthesis and for the regulation of NDP-sugar pool in the parasite.

INTRODUCTION

The trypanosomatid parasites are the causes of human and animal trypanosomiasis fatal diseases. The trypanosomatid cell surface consists of various structurally complex and diverse carbohydrates that are critical for their survival, and for the infection of the host cells and virulence (Ferguson, 1999; Guha-Niyogi et al., 2001). Major components of those glycoconjugates (glycans) including Galpyranose, Man, GlcN, Glc and GlcNAc are commonly found in Leishmania major, Trypanosoma cruzi and Trypanosoma brucei. However, the structures and sugar compositions of some glycoconjugates are fundamentally diverse among different types of trypanosomatid parasites. For example, T. cruzi exclusively produces a wide- range of unique glycan structures: the protein O-linked glycan that contains NANA and Gal residues (Turnock and Ferguson, 2007), N-linked glycans that are decorated with L-Fuc residues

(Barboza et al., 2005), and the protein phosphodiester-linked glycan (P-glycan) that contains rare

L-rhamnosyl and xylosyl-residues (Haynes et al., 1996; Macrae et al., 2005). To date these glycan structures appear distinct and were not reported in other parasites such as T. brucei and L. major. The specific glycosyl residues D-Xyl, L-Rha, D-Galfuranose and L-Fuc in T. cruzi were reported to be linked to the aa Ser/Thr-phosphate-linked carbohydrate chain of the glycoprotein gp72 (Haynes et al., 1996), which is found associated with flagella attachment. T. cruzi mutants lacking gp72, have reduced virulence in both stages of their life cycle: the insect and the mammalian host (Haynes et al., 1996). The function of these sugar residues within gp72 is lacking and so does the metabolic pathways leading for their formation.

The common glycosyl donors for the synthesis of glycoconjugates are nucleotide-sugars.

Different types of nucleotide-sugars (NDP-sugar) have been recently identified in the cell extracts of Leishmania major, Trypanosoma cruzi and Trypanosoma brucei (Turnock and

Ferguson, 2007). Some of these NDP-sugars are unique and not found in human such as UDP-

Galfuranose, making the enzymes that generate them (e.g., UDP-Gal mutase (Beverley et al.,

2005)) an attractive target for the development of anti-parasite drug. While many genes involved in NDP-sugar biosynthesis were identified in many species, less is known in T. cruzi regarding the synthesis of UDP-Xyl, UDP-Rha or GDP-Fuc, and whether their synthesis is analogous to other enzymes belonging to the interconversion pathway described in other organisms remained elusive. Some of the NDP-sugar biosynthetic genes were shown essential for the parasite invasion. UDP-Gal for example, synthesized via UDP-Glc-4-epimerase is indispensable for both bloodstream form and procyclic form T. brucei (Roper et al., 2002; Roper et al., 2005; Urbaniak et al., 2006) and is likely essential for epimastigote form T. cruzi (MacRae et al., 2006).

In addition to nucleotide-sugar interconversion pathway, the role of salvage pathway in supplying NDP-sugars to different metabolic processes remains largely unknown. In this pathway, free sugars generated from the degradation of polysaccharide, glycoprotein, and glycolipid could be recycled into the cell by specific sugar transporters. These sugars can be phosphorylated at C1 position by specific kinases to form sugar-1-Ps. Subsequently a group of nucleotidyltransferases, also known as nucleotide-sugar pyrophosphorylases (PPase), could transfer a nucleotidyl-residue to form NDP-sugars. Two functional PPase genes were identified in trypanosomatid parasites: the UDP-Glc PPase from Leishmania major (Lamerz et al., 2006), and UDP-N-acetylglucosamine pyrophosphorylase from Trypanosoma brucei (Stokes et al.,

2008). Knockout of the latter gene is lethal, however, a conditional mutant of this gene significantly reduces the amount of poly-N-acetylglucosamine structure and lead to underglycosylation of T. brucei glycoprotein, suggesting that other NDP-sugar pathways (i.e. interconversion) cannot substitute in forming UDP-GlcNAc. We are interested in identifying

potential enzymes that contribute to the synthesis of NDP-sugars to better understand how the formation of diverse glycans provides the organism fitness with its surrounding. This led among others, to the isolation of a plant gene fondly named Sloppy (Yang et al., 2009): a unique PPase enzyme that utilizes diverse sugar-1-Ps and UTP to form UDP-sugars (Kotake et al., 2004;

Litterer et al., 2006; Yang et al., 2009). Based on genome data to date few organisms carry such

PPase. Here we first report the identification and characterization of a functional broad range

UDP-sugar pyrophosphorylase in Trypanosoma cruzi that could metabolize various sugar-1- phosphates: Glc-1-P, Gal-1-P, GlcA-1-P and Xyl-1-P into their corresponding UDP-sugars. The

TcSloppy may explain alternative pathways for the synthesis of UDP-Xyl, UDP-Gal, UDP-GlcA and perhaps UDP-Rha in this organism.

EXPERIMENTAL PROCEDURES

cDNA cloning of Trypanosoma cruzi Sloppy - Total genomic DNA of Trypanosoma cruzi was a gift from Dr. Robert Sabatini (UGA). The coding sequence of Trypanosoma cruzi Sloppy was amplified by PCR using 1 unit of high-fidelity proof-reading Platinum DNA polymerase

(Invitrogen), and 0.2 µM of each forward and reverse primers: 5'- TC atg aag atg gtg cct gac g -3' and 5'- GG ATC cta aag ctt cgc atg atg -3' using genomic DNA of Trypanosoma cruzi as template. The PCR product was cloned to generate plasmid pCR4-topoTA:TcSloppy; and DNA was sequenced (GenBank GU443973, Tc00.1047053511761.10). The BspHI-KpnI fragment containing the full length Sloppy gene without the stop codon was sub-cloned into an E. coli expression vector derived from pET28b (Gu and Bar-Peled, 2004), generating Sloppy with an extension of six histidines at its N-terminal. Expression of Sloppy gene is under T7 promoter,

100

and the plasmid was transformed to BL21(de3)plysS-derived E. coli strain (Novagen) for gene expression.

Protein expression and purification - E. coli cells harboring the plasmid construct or an empty vector were cultured for 16 h at 37 ºC in LB medium (20 ml) supplemented with kanamycin (50 µg/ml) and chloramphenicol (34 µg/ml). A portion (8 ml) of the cultured cells was transferred into fresh LB liquid medium (250 ml) supplemented with the same antibiotics, and the cells then grown at 37 ºC at 250 rpm until the cell density reached OD600 = 0.6. The cultures were then transferred to 18 ºC and gene expression was induced by the addition of isopropyl β-D-thiogalactoside to a final concentration of 0.5 mM. After 24 h growth while shaking (250 rpm), the cells were harvested by centrifugation (6,000  g for 10 min at 4 ºC), resuspended in lysis buffer (10 ml 50 mM Tris-HCl, pH 7.6, containing 10% (v/v) glycerol, 50 mM NaCl, 1 mM MgCl2, 0.5 mM EDTA, 1 mM DTT and 0.5 mM phenylmethylsulfonyl fluoride) and lysed in an ice bath by 24 sonication cycles each (10-sec pulse; 20-sec rest) using a

Misonix S-4000 (Misonix incorporated, Farmingdale, New York) equipped with microtip probe.

The lysed cells were centrifuged at 4 ºC for 30 min at 20,000  g, and the supernatant (termed s20) was recovered and kept at -20 ºC. His-tagged proteins were purified on a column (10 mm id

 150 mm long) containing Ni-Sepharose (2 ml, Qiagen) equilibrated with 50 mM sodium- phosphate, pH 7.6, containing 0.3 M NaCl. The bound His-tagged protein was eluted with the same buffer containing increasing concentrations of imidazole. The fractions containing Sloppy activities were stored in aliquots at -80 ºC. The concentration of protein was determined using bovine serum albumin (BSA) as standard. The molecular weight of the recombinant protein was estimated by size-exclusion chromatography using a Waters 626 LC HPLC system equipped with a photo diode array detector (PDA 996) and a Waters Millennium32 workstation. Separate

101

solutions (0.5 ml) of TcSloppy or a mixture of standard proteins [10 mg each of alcohol dehydrogenase (157 kDa), bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), and cytochrome C (12.4 kDa)] were separately chromatographed at 0.5 ml min-1 on a Superdex200 column (10 mm id  300 mm long, GE) equilibrated with 0.1M sodium phosphate, pH 7.6, containing 0.1 M NaCl. The eluant was monitored at A280 nm and fractions were collected every

15 sec. Fractions containing enzyme activity were pooled and kept at -80 ºC.

Enzyme assays - The typical forward HPLC-based reactions for the formation of NDP- sugars were carried out in a final volume of 50 µl and consisted of 1 mM sugar-1-P, 1 mM UTP

(or other NTPs), 5 mM MgCl2, 100 mM Tris-HCl pH 7.6, 1U of yeast inorganic pyrophosphatase (Sigma), and recombinant TcSloppy. After 15 min incubation at 37 ºC, reactions were terminated (1 min at 100 ºC); equal volume of chloroform was added and the products were analyzed by anion-exchange chromatography using TSK-DEAD-5-PW column

(7.5mm inner diameter × 75 mm long, Bio-Rad) and ammonium formate HPLC gradient system

(Yang et al., 2009). Nucleotides and nucleotide-sugars were detected by their UV absorbance using photodiode array detector that was connected to the HPLC system. The maximum absorbance for uridine nucleotides and UDP-sugars were 261.8 nm in ammonium formate. The peak area of analytes was determined based on standard calibration curves. HPLC-based reverse reactions were carried out in a similar manner and included 1 mM PPi, 1 mM UDP-sugar, 5 mM

MgCl2, 100 mM Tris-HCl pH 7.6, and 45 ng TcSloppy. After 15 min at 37 ºC, reactions were terminated and the amount of UTP produced was determined from a standard calibration curve.

Real-time 1H-NMR analysis of NDP-sugar pyrophosphorylation - Individual pyrophosphorylation reaction in final volumes of 180 µl at the mixture of D2O:H2O of 8:1 (v/v) consisted of 100 mM sodium phosphate, pH 6, 5 mM MgCl2, 1 mM UTP, 1 mM sugar-1-P, and

102

enzyme: 0.9 µg recombinant Trypanosoma cruzi Sloppy was supplied in H2O-buffer.

Immediately upon addition of enzyme, the reaction mixture was transferred to a 3 mm NMR tube. In the combined UDP-sugar reaction, assays were as described above but included 4 mM

UTP and 1 mM of each UDP-sugars, and 1U of yeast inorganic pyrophosphatase. Real-time 1H

NMR spectra were obtained using a Varian Inova 600MHz spectrometer equipped with a cryogenic probe. Data acquisition was not started until approximately 2 minutes after the addition of enzyme to the reaction mixture due to spectrometer set-up requirements (shimming).

Sequential 1D proton spectra were acquired over the course of the enzymatic reaction. All spectra were referenced to the water resonance at 4.765 ppm downfield of 2,2-dimethyl-2- silapentane-5-sulphonate (DSS). Processing of the data as covariance matrices was performed with Matlab (The Mathworks, Inc.).

Enzyme properties and inhibition assays - The forward pyrophosphorylase activity of

TcSloppy was measured with various buffers, at different temperatures, different ions, or with different potential inhibitors. For the optimal pH experiments, 45 ng recombinant enzyme was first mixed with 5 mM MgCl2 and 100 mM of each individual buffer (Tris-HCl, phosphate, 4- morpholineethanesulfonic acid (MES), 4-morpholinepropanesulfonic acid (MOPS) or 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)). The optimal pH assays were initiated after the addition of specific sugar-1-P and UTP. Inhibitor assays were performed under standard assay conditions except for the addition of various additives (sugars, nucleotides) to the reaction buffer. These assays were incubated for 15 min at 37 ºC, and were subsequently terminated by heat (1 min at 100 ºC). The amount of UDP-sugar formed was calculated from a calibration curve of HPLC UV spectra of standards. For the experiments aimed at defining the optimal temperature, assays were performed under standard assay conditions except that reactions were

103

incubated at different temperatures for 15 min. Subsequently, the activities were terminated (100

ºC). For the experiments aimed at determining if TcSloppy required metals, assays were performed with UTP, specific sugar-1-P (Glc-1-P) with a variety of ions. After 15 min at 37 ºC incubation, the PPase activity was terminated by heat. The amount of UDP-sugar formed was calculated from HPLC UV spectra of standards.

For the experiments aimed at determining the ability of TcSloppy to utilize other sugar-1-

Ps, Sloppy assays were performed under standard assay conditions except for substituting the sugar-1-P (Glc-1-P) with different sugar-1-Ps (for example, GlcA-1-P, Gal-1-P, etc). These assays were incubated for 60 min (unless otherwise mentioned) at 37 ºC, and were subsequently terminated by heat (1 min at 100 ºC). The amount of UDP-sugar formed was calculated from

HPLC UV spectra.

Kinetics - The forward pyrophosphorylation catalytic activity of TcSloppy was determined at 37 ºC for 5 min. Reactions contained 100 mM Tris-HCl pH 7.6, MgCl2 (5 mM), 9 ng recombinant enzyme (0.13 pmol), and variable concentrations of UTP (40 µM to 8 mM) and

Glc-1-P (1 mM) or with UTP (1 mM) and variable concentrations of Glc-1-P (40 µM to 8 mM).

The forward reaction was also carried out with variable concentrations of Gal-1-P or Xyl-1-P (40

µM to 8 mM) and UTP (1 mM), using 74 ng and 220 ng recombinant TcSloppy respectively.

Kinetics for the forward reaction of TcSloppy with Glc-1-P and TTP were carried out the same way as described for UTP, except for using 880 ng recombinant TcSloppy. The kinetic assays were also carried out with 2 units of yeast inorganic pyrophosphatase to deplete PPi. Kinetics for the reverse reactions were performed in the same condition as above, with a fixed concentration of PPi (1 mM) and variable concentrations of UDP-Glc (40 µM to 8 mM), and 9 ng recombinant

TcSloppy (0.13 pmol). In a separate series of reverse-reaction experiments, assays were

104

performed with a fixed amount of UDP-Glc (1 mM) and variable concentrations of PPi (40 µM to 8 mM). Enzyme velocity data of the amount (µM) of UDP-sugar produced per second, as a function of substrate concentrations were plotted. The Solver tool (Excel version 11.5 program) was used to generate best-fit curve calculated by nonlinear regression analyses, and for the calculation of Vmax and apparent Km.

RESULTS

Identification, cloning and characterization of Trypanosoma cruzi (T. cruzi) Sloppy -

Nucleotide-sugar pyrophosphorylases (PPase), also known as NTP:sugar-1-P nucleotidyltransferases, are very specific for their nucleotides. While some utilize ATP (Jin et al., 2005), others are specific for GTP, UTP, CTP or TTP (Tanner, 2001). Until recently, most

PPases were shown to be specific for their sugar-1-Ps as well. Some recognize Man-1-P; others recognize Glc-1-P or GlcNAc-1-P (Brown et al., 1999; Badejo et al., 2007; McCoy et al., 2007).

Few of the PPases, most notably the plant and bacterial ADP-glucose PPase, are allosterically regulated by intermediates of the carbon assimilation pathways (McCoy et al., 2006). The discovery of a “Sloppy” UDP-sugar PPase in plants (Kotake et al., 2004; Litterer et al., 2006;

Yang et al., 2009), an enzyme that can utilize diverse sugar-1-Ps with UTP to form UDP-sugars, provides an alternative route to explain how NDP-sugars can be made. The existence of such enzyme supports the essential role of the glycan salvage pathway for normal cellular function.

The cellular regulation of each specific NDP-sugar PPase compared with the Sloppy PPase remained unclear.

Although all NDP-sugar PPases belong to the same family of nucleotidyl-transferases, they are very diverse with low aa sequence identity, although based on structural analyses

105

(Peneff et al., 2001; McCoy et al., 2007; Steiner et al., 2007) they appear to have a conserved fold. The nucleotidyltransferases also vary in length: in eukaryotes, UDP-Glc PPase range in sizes across species from 470 to 510 aa, while TDP-Glc PPase in prokaryotes is much shorter

~300 aa, and is more closely homologous to GDP-Man PPase. The eukaryote UDP-GlcNAc

PPase is 505 aa and the plant UDP-sugar PPase, “Sloppy”, is 614 aa long. CMP-NeuAc- and

CMP-KDO-PPase are also members of the family (Ghalambor and Heath, 1966; Kean and

Roseman, 1966). What are the structural alterations in the ancestral PPase that provide this class of enzymes their strict NTP and sugar specificities as oppose to being “sloppy” remained unclear since only plant Sloppy-like enzymes were characterized.

To identify other potential “Sloppy-like PPases” across species we compared the

Arabidopsis protein sequence (ABC55066.1) with NR-sequence database. BLAST analyses of homologous proteins from different species revealed that “Sloppy” shares overall a low sequence identity with UDP-Glc PPase and UDP-GlcNAc PPase (23% and 26% respectively), suggesting they may share a similar protein fold and conserved catalytic residues. Sequence alignment of three PPases found two consensus motifs (Fig. 3.1A): the N-terminal region of Sloppy comprised of a putative nucleotide binding (NB) motif „GG(L/Q)G(E/T)(R/T)(L/M)GX3(I/P)K‟ (starting at aa 136) and the „PXGHGD(V/I)HX2(L/I)‟ (starting at aa 251) motif probably involved in

„uridine-binding‟ (UB). Interestingly and unexpected was the identification of a relatively close

“Sloppy-like” homolog protein in Trypanosoma cruzi: an organism that is evolutionary far removed from plant kingdom. The T. cruzi Sloppy-like gene encodes protein with 35% aa identity with the Arabidopsis Sloppy. Phylogenetic analysis (Fig. 3.1B) indicated that the

Sloppy-like proteins from different species are distinguished from UDP-Glc- and UDP-GlcNAc-

PPases. To determine however, if the Tc-gene encodes a PPase and more specifically to establish

106

if it has different or similar range of sugar-1-Ps and NTP specificity, the gene was cloned and the recombinant protein expressed in E. coli was analyzed.

107

108

FIGURE 3.1. Sequence alignments and phylogenetic relationships of UDP-Glc PPase

(UGP), UDP-GlcNAc PPase (UAP) and UDP-sugar PPases (Sloppy) from different organisms.

A. Sequences of UGP, UAP and Sloppy (see gene name below) were aligned using T-coffee

(Notredame et al., 2000) software with G-block (Castresana, 2000). The conserved motifs presumably involved in nucleotide-sugar binding (NB) and the uracil binding (UB) are labeled in bold. Potential amino acids that are conserved in UGP, UAP and Sloppy are highlighted in grey, based on sequence alignment.

109

B. Phylogenetic relationships of UGP, UAP and Sloppy in different species. Protein sequences

(see name and gene accession numbers below) were aligned and analyzed using T-coffee

(Notredame et al., 2000) software with G-block (Castresana, 2000) and the phylogenetic tree was created using MrBayes 3.1.2 software (Huelsenbeck and Ronquist, 2001; Dereeper et al., 2008).

Branch support values (more than 50%) are shown. The bar represents 0.2 protein substitutions per site.

Arabidopsis thaliana Sloppy (AtSloppy, ABC55066.1), Trypanosoma cruzi Sloppy (TcSloppy,

GU443973, Tc00.1047053511761.10), Thalassiosira pseudonana Sloppy (TpSloppy,

XP_002291538), Leishmania major Sloppy (LmSloppy, GU443974, LmjF17.1160), Pisum sativum Sloppy (PsSloppy, Q5W915), Paramecium tetraurelia Sloppy (PtSloppy,

XP_001430540), Cryptosporidium muris Sloppy (CmSloppy, XP_002141351); UDP-Glc PPase from Homo sapiens (HsUGP, NP_006750), Arabidopsis thaliana (AtUGP, NP_186975),

Saccharomyces cerevisiae (ScUGP, NP_012889), Trypanosoma cruzi (TcUGP, XP_808700),

Trypanosoma brucei (TbUGP, XP_827798), Leishmania major (LmUGP, XP_001682505)

Cryptococcus neoformans, (CnUGP, XP_569599); UDP-GlcNAc PPase from Homo sapiens

(HsUAP, NP_003106), Arabidopsis thaliana (AtUAP, NP_564372), Saccharomyces cerevisiae

(ScUAP, CAY78406), Trypanosoma cruzi (TcUAP, XP_820911), Trypanosoma brucei

(TbUAP, XP_828335), Leishmania major (LmUAP, XP_001686013), Cryptococcus neoformans,

(CnUAP, XP_571302).

110

A highly expressed protein band (67 kDa) was detected after SDS-PAGE analysis of E. coli cells expressing Trypanosoma cruzi Sloppy (TcSloppy) (Fig. S3.1, lane 2 and 4; marked by the upper arrow). The mass of the column-purified protein is in agreement with the calculated mass of the translated gene product fused at the N-terminal portion to 6His. Preliminary experiments have shown that in the presence of Mg2+, the recombinant Tc-protein converts Glc-

1-P and UTP to a new UDP-sugar peak that eluted at 12.3 min (see peak #1, in Fig. 3.2, panel 1).

This UDP-sugar was eluted with the same retention time as UDP-Glc standard. To determine the identity of the UDP-sugar peak, it was collected from the column, and analyzed by 1H-NMR.

The NMR spectrum (Fig. 3.3, panel 1 and supplementary section Fig. S3.2A) provided chemical shifts consistent with UDP--D-glucose. The diagnostic J1, 2 value of 3.5 Hz and J2, 3, J3, 4, J4, 5, J5,

6, J6a, 6b values of 9.7, 9.7, 9.7, 3, 12 Hz respectively, indicate an α-glucopyranose configuration, along with the distinct chemical shift of H1 (5.59 ppm). The linkage of the sugar moiety to the phosphate is given by the coupling constant value of 7 Hz for J1, P of the proton anomeric Glc residue and a coupling value of 3 Hz for J2, P. These data confirmed that the Tc-enzyme is a

PPase.

111

112

FIGURE 3.2. HPLC-based assays of recombinant T. cruzi Sloppy activity.

The assays for the formation of NDP-sugars, included UTP and different sugar-1-Ps with either

TcSloppy (panels 1-4) or with control protein (empty vector control cells, Panels 6-9). Panel 1 shows the formation of UDP-Glc (marked by arrow #1); Panel 2 shows the formation of UDP-

Gal (marked by arrow #2); the formation of UDP-GlcA (marked by arrow #3, in Panel 3); and the formation of UDP-Xyl (marked by arrow #4, in Panel 4). TcSloppy was also incubated with

TTP and Glc-1-P; TDP-Glc was formed (marked by arrow #5, in Panel 5) when compared with control (empty vector control cells, Panel 10). The HPLC peaks (panels 1-5) are UTP (16.3 min),

TTP (16.3 min), UDP-Glc (12.3 min), UDP-Gal (12.2 min), UDP-GlcA (16.0 min), UDP-Xyl

(12.5 min), and TDP-Glc (12.3 min). The minor peak marked as “a” at 14.8 min is UDP contamination in the UTP reagent. Panel stdA and stdB show the elution time of standards:

UDP-Gal (12.2 min), UDP-Glc (12.3 min), TDP-Glc (12.3 min), UDP-Xyl (12.5 min), UDP-

GlcA (16.0 min), and UTP (16.3 min).

113

FIGURE 3.3. Product analyses of HPLC-based assays by NMR, confirming that recombinant Tc enzyme has Sloppy NDP-sugar PPase activity.

Proton NMR spectra of TcSloppy enzymatic products. Each peak eluted from the Bio-Rad column (see Fig. 3.2, Arrow #1, #2, #4, #5) was collected, lyophilized, dissolved in D2O and analyzed by 1H-NMR. The assay shown in panel 3 (Fig. 3.2, Arrow #3) was chromatographed on a Q15 column as it separates UDP-GlcA from UTP. The UDP-GlcA peak was collected and analyzed by NMR. The NMR spectra (0-9 ppm) of each individual TcSloppy reaction product are shown and the identity of the product is labeled on the upper right of each panel. Detail NMR data that cover the sugar anomeric region (5.5-6 ppm) and an expansion spectra of the NDP- sugar ring protons (3.4-4.4 ppm) are provided in Supplemental section, Fig. S3.2.

114

We found that the recombinant enzyme also converts UTP and Gal-1-P to UDP-Gal (12.2 min, see Fig. 3.2, panel 2), UTP and GlcA-1-P to UDP-GlcA (16.0 min, Fig. 3.2, panel 3), UTP and Xyl-1-P to UDP-Xyl (12.5 min, Fig. 3.2, panel 4), based on the retention time of standards.

Control cells expressing empty vector had no detectable activity. To unambiguously determine the identity of each enzymatic product, the individual peak marked #2 and #4 (Fig. 3.2) were

1 collected from the column, and their structures were confirmed by H NMR as UDP--D- galactose, and UDP--D-xylose respectively (Fig. 3.3 panel 2 and 4; Fig. S3.2, B and D). Since the enzymatic product of UTP and GlcA-1-P co-eluted with UTP (Fig. 3.2 panel 3), the reaction mixture was chromatographed on a Q15 column as it separates UDP-GlcA from UTP. The peak eluted from the Q-column was analyzed by proton NMR (Fig. 3.3 panel 3 and Fig. S3.2C) and confirmed as UDP--D-glucuronic acid. Interestingly and unlike the Arabidopsis-Sloppy, the Tc- enzyme was capable of converting TTP and Glc-1-P to TDP-Glc (12.3 min, Fig. 3.2 panel 5 and

Fig. 3.3 panel 5) as well, albeit at lower rate. Therefore, we propose that Tc-enzyme is a PPase and displays like the plant Sloppy, a broad uridylyltransferase activity with various sugar-1-Ps as substrates. Just like the plant Sloppy, TcSloppy was unable to convert GlcNAc-1-P and UTP.

Characterization and properties of T. cruzi Sloppy - Crude recombinant TcSloppy is stable when stored at -20 ºC. Purified TcSloppy could be stabilized for several months at -80 ºC when flash-frozen in liquid nitrogen. TcSloppy requires Mg2+, although Mn2+ can substitute for magnesium (Table 3.1), and the activity, as expected, is completely abolished in the presence of

EDTA. The recombinant Sloppy is active between pH 3.3 and pH 9.0 (Fig. 3.4), with maximum activity at pH 7.5 to 7.8 in Tris, or at pH 6 in phosphate buffer. The enzyme is also active when reactions were performed in HEPES, MOPS or MES buffers. TcSloppy is active between 25 ºC and 55 ºC, with maximum activity at 30-42 ºC (Table 3.2).

115

FIGURE 3.4. The effects of different buffers and pH on TcSloppy activity.

TcSloppy activity was determined in different buffers at different pH values. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

116

TABLE 3.1. TcSloppy requires metal for activity.

Relative TcSloppy Additive (5 mM) Activity (%)

MgCl2 100

MnCl2 104

CaCl2 4

ZnSO4 2

EDTA 3 water 2

TcSloppy was mixed with different metal salts, EDTA, or water as a control for 10 min on ice.

Subsequently, UTP and appropriate sugar-1-P were added and PPase assay was carried out under standard conditions. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

117

TABLE 3.2. The effect of temperature on TcSloppy activity.

Relative TcSloppy Temperature (ºC) Activity (%)

4 5

25 34

30 102

37 100

42 115

55 26

65 7

118

We next investigated the nucleotide triphosphates (NTPs) specificity of TcSloppy. CTP,

GTP, ITP and ATP are not substrates for TcSloppy when using Glc-1-P as substrate. Several commercially available sugar-1-Ps were tested as substrates for TcSloppy with different nucleotides (e.g., GTP, CTP, UTP, ATP). No activity was observed even when the standard Tc- assay expanded longer incubation time (up to 1h). To determine if TcSloppy may recognize other NTPs we performed standard assays in the presence of competing nucleotides such as

(ATP, CTP, GTP, ITP at 0.5 mM each). In all cases Glc-1-P was readily uridylated, suggesting that the enzyme other than UTP is not recognizing the above NTPs. In addition to UTP, the enzyme could also convert Glc-1-P to TDP-Glc in the presence of TTP, but at much lower rate

(Fig. 3.2, lower panel). To determine if the enzyme recognizes and binds nucleotide diphosphates (NDP such as UDP or ADP), prior to the standard assay, the enzyme was incubated with NDPs. These NDPs as well as other nucleotides tested (e.g., NMP, NAD, NADH) had no effect on Sloppy activity. By contrast, enzyme activity was reduced by 65% in the presence of

0.5 mM PPi (Table S3.1) when assays were conducted in the forward reaction.

Real-Time NMR analysis of T. cruzi Sloppy - To monitor the dynamics of the enzymatic reaction and the substrate preference of TcSloppy, we used real-time 1H-NMR spectroscopy

(Fig. 3.5). These assays were carried out in phosphate buffer to avoid the proton signals from

Tris. In the NMR reactions presented in Fig. 3.5, all sugar-1-phosphates were combined along with four mole equivalents of UTP. As shown in the time dependent enzymatic progression, a faster conversion of the Glc-1-P (5.46 ppm) into UDP-α-D-Glc (5.59 ppm) and Gal-1-P (5.50 ppm) into UDP-α-D-Gal (5.63 ppm) were observed when compared with the conversion of GlcA-

1-P (5.48 ppm) to UDP-α-D-GlcA (5.61 ppm) and Xyl-1-P (5.41 ppm) to UDP-α-D-Xyl (5.54

119

ppm) (Fig. 3.5). After peak deconvolution, the rate of specific sugar-1-P conversion was UDP-

Glc=UDP-Gal>UDP-Xyl>UDP-GlcA.

120

FIGURE 3.5. Real-Time 1H NMR based Sloppy assays.

In the real-time NMR assay, recombinant Sloppy was mixed with all sugar-1-Ps (Glc-1-P, Gal-1-

P, GlcA-1-P and Xyl-1-P, 1 mM each), buffer and 4 mM UTP. Approximately, 2 min after enzyme addition and NMR shimming, data were collected. Progressions of enzyme activity covering the anomeric region of the proton NMR spectrum are shown. The signal for the anomeric proton of the sugar-1-P and the UDP-sugar has a quadruplet peak form: peak 1 (5.41 ppm) is Xyl-1-P, peak 2 (5.46 ppm) is Glc-1-P, peak 3 (GlcA-1-P at 5.48 ppm), peak 4 (Gal-1-P at 5.50 ppm), peak 5 (UDP-α-D-Xyl at 5.54 ppm), peak 6 (UDP-α-D-Glc at 5.59 ppm), peak 7

(UDP-α-D-GlcA at 5.61 ppm), peak 8 (UDP-α-D-Gal at 5.63 ppm). Note: Glc-1-P and Gal-1-P peaks overlap in part with GlcA-1-P. Horizontal line shows the ppm spanning the quadruplet shape peak for each of the substrate and the enzymatic product.

121

We also performed similar real-time NMR assays with the individual sugar-1-P in the absence of the pyrophosphatase, to obtain equilibrium between the forward and reverse reactions.

The results of these assays are shown in Fig. 3.6 and Fig. S3.3 panel A-C, and summarized in Fig.

S3.3 panel D. At equilibrium the ratio of sugar-1-P to UDP-sugar were as follows: 1.8 for UDP-

Glc, 1.3 for UDP-Gal, 2.8 for UDP-GlcA, 2.8 for UDP-Xyl. These data clearly show the preference for the reverse reaction, i.e. conversion of UDP-sugar to sugar-1-P. This preference is common to almost all PPases, such as UDP-GlcNAc PPase (UAP), UDP-Glc PPase (UGP) and

TDP-Glc PPase. In the forward reaction, the sugar-1-P preference of T. cruzi Sloppy is in contrast to the plant Sloppy from Arabidopsis, where UDP-GlcA was the preferred substrate and the conversion rate was UDP-GlcA>UDP-Glc>UDP-Gal>UDP-GalA>UDP-Xyl (Yang et al.,

2009) (and T. Yang, unpublished).

122

FIGURE 3.6. Real-Time 1H NMR based Sloppy assays with Glc-1-P.

A time dependent 1H-NMR spectrum arrays for the formation of UDP-Glc by TcSloppy. Glc-1-P

(1 mM) was separately reacted with 1 mM UTP and TcSloppy. Approximately, 2 min after enzyme addition and NMR shimming, data were collected. The progression of enzyme activity monitored by following changes in the spectrum of the sugar anomeric proton region (from 5.4 to 5.7 ppm) is shown. To visualize changes in product with time, each spectrum at a given time is plotted sequentially. Each time point reflects the amount of UDP-Glc formed (Uglc, ) in the forward reaction, and the decreased the amount of Glc-1-P (glc1P, ).

123

Kinetic and catalytic properties of T. cruzi Sloppy - Kinetics analyses of the enzyme are summarized in Table 3.3A. The apparent Km values for the forward reaction were 177 µM (Glc-

-1 -1 -1 1-P) and 28.4 µM (UTP), with Vmax values of 0.07 µM s , and kcat/Km (s µM ) were 0.15 (Glc-

1-P) and 0.92 (UTP). The kinetics for the reverse reaction, had an apparent Km values of 26 µM

-1 -1 -1 (UDP-Glc) and 134 µM (PPi), with Vmax values of 0.08 µM s , and kcat/Km (s µM ) of 1.15

(UDP-Glc) and 0.22 (PPi). Further kinetic data for the forward reaction of TcSloppy with different sugar-1-Ps and nucleotides are summarized in Table 3.3B. The relatively lower activity of TcSloppy for the conversion of TTP and Glc-1-P to TDP-Glc can be explained by the affinity to the nucleotide. The Km values of TcSloppy for UTP and TTP are different (28.4 µM and

2540.8 µM, respectively). In addition, the difference in catalytic efficiency (kcat/Km) between the two substrates of the enzyme toward UDP-Glc and TDP-Glc reflects in part the different affinity for the respective nucleotide-sugar but more significantly due to the actual rate of catalysis. The kinetic data of TcSloppy are comparable with Sloppy from Arabidopsis (Litterer et al., 2006) and

Leishmania.

124

TABLE 3.3. Enzyme kinetics of TcSloppy.

A. Forward reaction k /K K (Glc-1-P) K (UTP) cat m k /K (UTP) m m (Glc-1-P) cat m µM µM µM -1 s-1 µM -1 s-1 LmSloppy 3605 322 0.02 0.05 TcSloppy 177 28.4 0.15 0.92 AtSloppy 420 190 0.16 0.57 Reverse reaction k /K K (UDP-Glc) K (PPi) cat m k /K (PPi) m m (UDP-Glc) cat m µM µM µM -1 s-1 µM -1 s-1 LmSloppy 657 659 0.1 0.1 TcSloppy 26 134 1.15 0.22 AtSloppy 720 160 0.08 0.35

B. Forward reaction

Km kcat kcat/Km µM s-1 s-1 µM -1 Glc-1-P 177 26 0.15 Gal-1-P 818.8 35.2 0.04 Xyl-1-P 2538.8 8.6 0.034 Glc-1-P (TTP) 1760 0.12 6.8E-05 TTP 2540.8 0.085 3.34E-05

A. The forward kinetic TcSloppy reactions were measured for 5 min with different concentrations of Glc-1-P and UTP. In the reverse reaction, the enzymatic activity was measured with different concentrations of UDP-Glc and PPi.

B. TcSloppy kinetics using various sugar-1-Ps, UTP and TTP. Reactions with Glc-1-P, Gal-1-P, and Xyl-1-P were carried out with UTP. The reactions labeled Glc-1-P (TTP) were carried out with fixed concentration of TTP as the nucleotide. The Km for TTP was measured with variable concentration of the nucleotide and fix concentration of the Glc-1-P. Enzyme velocities were plotted and Solver software was used to generate best-fit curve and for calculation of Vmax and apparent Km. Each value is the mean of triplicate reactions, and the values varied by no more

125

than ±10%. The gene encoding L. major Sloppy (GenBank GU443974, LmjF17.1160) was cloned, expressed, purified and the kinetic data were carried out as described for TcSloppy. The kinetics of Arabidopsis Sloppy are also provided for comparison (Litterer et al., 2006).

126

DISCUSSION

We have cloned and biochemically characterized a T. cruzi UDP-sugar pyrophosphorylase that in the presence of Mg2+ and UTP specifically uridylates a broad range of sugar-1-Ps with higher efficiency towards Glc-1-P and Gal-1-P and decreased efficiency for Xyl-

1-P and GlcA-1-P. We have also cloned and expressed the Sloppy like gene from Leishmania major (GenBank GU443974, LmjF17.1160). The LmSloppy is active towards UDP-Glc (Table

3.3), UDP-Gal, UDP-Xyl and UDP-GlcA but unlike TcSloppy it is also active in vitro, with

UDP-GalA (Yang et al, unpublished).

The Tc-PPase enzyme is reversible and in the presence of UDP-Glc and PPi for example, will form Glc-1-P and UTP. The physiological significance of the reverse reaction is unclear, as

PPi in normal cell is thought to be readily hydrolyzed by pyrophosphatase to 2Pi (Chen et al.,

1990), hence prevents the hydrolysis of NDP-sugar to sugar-1-P. Comparing the activities of

Sloppy from different organisms shows altered specificities. The Arabidopsis UDP-sugar PPase

(Sloppy) can form at least six different UDP-sugars, and it is possible that TcSloppy may have other sugar-1-P as substrates, for example rhamnose-1-P. The analyses of these different Sloppy- like proteins also illustrate that functional, biochemical analysis is essential; and that homology is an insufficient criterion to infer functional specificity.

Sloppy-like sequences are found in the genome of several protozoan parasites such as L. major, Cryptosporidium muris, Paramecium tetraurelia and in the marine phytoplankton,

Thalassiosira pseudonana. Interestingly, the genome of the parasite T. brucei does not harbor

Sloppy-like gene and so far, no xylose, rhamnose, and arabinose residues are reported to exist in the glycans of that organism. What advantage Sloppy provides L. major and T. cruzi but not T. brucei remains to be investigated. It would be of interest to express Sloppy in T. brucei and

127

examine if it alters invasion, host preference or survival. Alternatively, the lack of Sloppy in T. brucei could be an advantage and allow it to escape the immune system of the host.

Amino acid sequence alignment of several Sloppy-like sequences along with UDP-Glc

PPases and UDP-GlcNAc PPases shares very low sequence identity. Yet most of the PPases appear to have conserved binding and catalytic motifs, suggesting these enzymes have maintained a conserved fold throughout evolution, but the sequences were altered and evolved.

Common consensus motifs in PPases are the glycine rich motif for nucleotide binding (NB, based on the nomenclature of Steiner et al., 2007 (Steiner et al., 2007)) and a consensus motif for uracil binding. Some amino acids outside of the two motifs, presumably involved in substrate binding and catalytic function are also highly conserved in those PPases. It is likely that insertion of loops between the conserved structural domains generated many changes during evolution to allow the specific enzyme to be either strict to the substrate or to accept different substrates. The functional groups attached to the sugar carbon atoms (C2, C4, C6) appear critical for the recognition of PPases. For example, Peneff et al (Peneff et al., 2001) pointed to Asn223 in UAP, and suggested its involvement in the interaction with the acetyl group linked to C2 of GlcNAc.

Sequence alignment indicates this amino acid is replaced by His230 in TcSloppy, and this may explain the inability of Sloppy to uridylate GlcNAc-1-P. In UGP, the recognition of the C6-OH group of glucose likely occurs via Lys380 and Asn219. Sequence alignment indicates that several loops were inserted in this region in TcSloppy, and this may explain the ability of Sloppy to uridylate pentose-, hexose- and uronate- 1-Ps. The contribution of these loop elements within

Sloppy remained to be determined. In addition to the above examples, a significant difference among UDP-Glc PPase, UDP-GlcNAc PPase and Sloppy is also found in the C-terminal region.

Sloppy has several additional domains between aa 500 and 603. These domains may either be

128

structural, regulatory, or other functional elements. Insertion of loops between structural domains may provide alternative binding towards different sugar-1-P substrates or perhaps different nucleotides. Current work is underway to mutagenize and to crystallize TcSloppy with different ligands (e.g., Glc-1-P, GlcA-1-P, and Gal-1-P) and UTP to identify how subtle changes in amino acids with certain loops contribute to its “sloppiness” and be able to accept not only sugar-1-P with difference at the C4-epimer (i.e. Gluco- vs Galacto-configurations) but also sugar-1-P with altered groups attached to C5: carboxylate (COOH), primary alcohol (CH2OH), proton (H), and conceivably methyl group (CH3). Based on the sloppiness nature of this enzyme, it would suggest that functional groups at the C2, C3 and the C1-P portion of the sugar-1-P are likely the only regions that are critical for recognition by these Sloppy enzymes. Interestingly, TcSloppy like many PPases is not inhibited by UDP and UMP (Table S3.1), suggesting that the gamma- phosphate group of UTP is essential for binding of the nucleotide and perhaps inducing the conformational change of the enzyme.

Unlike L. major, T. cruzi consists of both UDP-Glc dehydrogenase-like gene (UGDH) and UDP-GlcA decarboxylase-like gene (UXS) with high sequence identity to these functional genes isolated in other organisms. Although the specific functions of these Tc genes were not determined yet, it suggests that T. cruzi can convert UDP-glucose to UDP-GlcA and subsequently to UDP-xylose (See Fig. 3.7), as originally proposed by Turnock and Ferguson

(Turnock and Ferguson, 2007). The presence of TcSloppy in T. cruzi is therefore complex, suggesting that different metabolic pathways may be involved for the production of these NDP- sugars in different stages of the parasite life cycle. Another possibility is that recycling of monosaccharides released from catabolism of glycans is mediated by the salvage pathway, and

129

requires Sloppy. One cannot discount that Sloppy in T. cruzi rather than synthesizing UDP-sugar is actually depleting them by conversion them to sugar-1-P.

As Sloppy-like genes do not appear in human and animal, it would be worth perusing knockout of the genes and examine if inhibition of Sloppy could be an effective drug to clear the parasite from its host.

130

FIGURE 3.7. A model for sugar recycling, salvage, and nucleotide-sugar formation in T. cruzi.

The enzymes and candidate genes are shown in italics. Note: the human or yeast GalT-like sequences were not found in the T. cruzi genome. The plant UDP-Glc 4,6-dehydratase like sequences were found in T. cruzi genome, however, the subsequent UDP-6-deoxy-Glc 3,5- epimerase/4-reductase like sequences for the formation of UDP-Rha were not found in T. cruzi genome. Whether Gal or other sugar (GlcA, Xyl, etc) is internalized into the cell or a by-product of glycan degradation in the cell, remained to be determined.

131

CHAPTER 4

IDENTIFICATION AND CHARACTERIZATION OF A STRICT AND OF A

PROMISCUOUS N-ACETYLGLUCOSAMINE-1-P URIDYLYLTRANSFERASES IN

ARABIDOPSIS3

3This research was originally published in Biochemical Journal. Reprinted here with permission of publisher. Ting Yang, Merritt Echols, Andy Martin and Maor Bar-Peled. Biochem J. 2010

430(2):275-84. © The Biochemical Society

132

ABSTRACT

UDP-GlcNAc is an essential precursor for glycoprotein and glycolipid synthesis. A functional nucleotidyltransferase gene from Arabidopsis encoding a 58.3 kDa N- acetylglucosamine-1-phosphate uridylyltransferase (GlcNAc1pUT-1) was identified. In the forward reaction the enzyme catalyzes the formation of UDP-N-acetylglucosamine and PPi, from the respective monosaccharide 1-phosphate and UTP. The enzyme can utilize the 4-epimer,

UDP-GalNAc as a substrate as well. The enzyme requires divalent ions (Mg2+ or Mn2+) for activity and is highly active between pH 6.5-8.0, and at 30-37 ºC. The apparent Km values for the forward reaction were 337 µM (GlcNAc-1-P) and 295 µM (UTP) respectively. Another

GlcNAc1pUT-2, which shares 86% amino-acid sequence identity with GlcNAc1pUT-1, was found to convert in addition to GlcNAc-1-P and GalNAc-1-P, Glc-1-P to corresponding UDP- sugars, suggesting that subtle changes in the UT family cause different substrate specificities. A

3D-protein structure model using the human AGX1 as template, showed a conserved catalytic fold and helped identify key conserved motifs, albeit the high sequence divergence. The identification of these strict and promiscuous gene products open a window to indentify new roles of amino-sugar metabolism in plants and specifically their role as signaling molecules. The ability of GlcNAc1pUT-2 to utilize three different substrates may provide further understanding why biological systems have plasticity.

133

INTRODUCTION

N-acetyl-glucosamine (GlcNAc) is a major sugar residue found in different types of glycans across species. In the endoplasmic reticulum, the attachment of GlcNAc is required to initiate N-glycan processing for the synthesis of glycoproteins (Stanley et al., 2008), and separately to initiate the synthesis of the core for GPI anchor-linked protein synthesis (Hancock,

2004), and the synthesis of glycolipids (Schnaar et al., 2008). In certain bacteria, the GlcNAc residue is major components of the peptidoglycan in periplasm and in the lipopolysaccharide of lipid A (endotoxin) (Raetz and Whitfield, 2002). Both in plants and mammalian, the specific attachment and removal of GlcNAc-residue from cytosolic and nuclear proteins play a signaling function for the modification of regulatory proteins. Such posttranslational modification by either an OGT (O-GlcNAc-transferase) or protein-kinase (Wells et al., 2004) is very common. Specific

OGT catalyzes the transfer of the GlcNAc residue from UDP-GlcNAc to a serine or threonine within a substrate protein (Hart et al., 1989), and the mutation of plant OGT is crucial for gamete and seed development (Jacobsen et al., 1996; Hartweck et al., 2002; Silverstone et al., 2007).

Lastly in plants, several proteins modified with terminal N-acetylglucosamine (GlcNAc) are present at the nuclear rim and at the nuclear pore complex of tobacco, the role of these sugar modification at the pore remain elusive (Heese-Peck and Raikhel, 1998).

Very little is known about the 4-epimer of GlcNAc, N-acetylgalactosamine (GalNAc) residue in plants. In animal GalNAc is largely found in the core structure of O-linked glycan.

GalNAc residue was reported however, in the polysaccharide of green algae (Brosch-Salomon et al., 1998). Certain plant UDP-D-glucose-4-epimerase isoforms can in-vitro interconvert UDP-

GlcNAc and UDP-GalNAc (Zhang et al., 2006). The metabolite, UDP-GalNAc was isolated from Dahlia and Squash plants (Gonzalez and Pontis, 1963), and more recently from Arabidopsis

134

suspension culture and from the endosperm of Fenugreek plants (Alonso et al., 2010). But how these sugar residues are made still remained unknown, as the genes encoding these activities were not identified.

The biosynthesis of UDP-N-acetylglucosamine, UDP-GlcNAc, differs in different organism. In prokaryote the transamination of Fru-6-phosphate and glutamine into glucosamine-

6-P by a glutamine:Fru-6-P amidotransferase initiates the pathway. GlcN-6-P is subsequently phosphoisomerized into GlcN-1-P by GlcN phosphate mutase. A single bifunctional enzyme glmU, comprised of acetyltransferase domain transfers acetate from Ac-CoA to GlcN-1-P, producing GlcNAc-1-P, and the nucleotidyltransferase domain of glmU uridylates GlcNAc-1-P with UTP to form UDP-GlcNAc (Mio et al., 1998) (Fig. 4.1A). Unlike in prokaryotes, in eukaryotes (Fig. 4.1B), glucosamine-6-P is first N-acetylated by an acetylase to GlcNAc-6-P and then converted to GlcNAc-1-P by phospho-N-acetylglucosamine mutase. The final step, first identified in 1954 by EEB Smith and GT Mills (Smith and Mills, 1954) is carried out by a specific nucleotidyltransferase, UTP:N-acetyl-glucosamine-1-P uridylyltransferase, (Abbr,

GlcNAc1pUT) that in the forward reaction catalyzes the formation of UDP-GlcNAc and PPi from GlcNAc-1-P and UTP. This type of uridylation activity belongs to a large family of enzymes that also catalyze the formation of other specific NDP-sugars such as UDP-Glc, UDP-

GalA, UDP-GlcA, ADP-Glc, GDP-Man and many others. Common for this family is the metal- dependent transfer of the nucleotidyl moiety and the specific phosphate acceptors. These catalysis involves the cleavage of the pyrophosphate linkage (P-alpha : P-beta) in the nucleotide substrate (Feingold and Avigad, 1980). In the reverse reaction, PPi and NDP-Sugar are converted to sugar-1-P and NTP (Mio et al., 1998; Milewski et al., 2006). The sugar-1-P specificities of

GlcNAc1pUT across organisms are unresolved. As some enzymes may use Glc-1-P and

135

GlcNAc-1-P as substrates (Mio et al., 1998), while others may use GalNAc-1-P and GlcNAc-1-P

(Peneff et al., 2001). The gene encoding the plant homologous activity was not identified due to low sequence identity, and the specificities of the plant enzyme are not clear.

136

FIGURE 4.1. UDP-N-acetylglucosamine (UDP-GlcNAc) biosynthesis pathways in prokaryotes and eukaryotes.

A. UDP-GlcNAc biosynthesis pathway in prokaryotes. The enzymes and candidate genes are shown in italics.

B. UDP-GlcNAc biosynthesis pathway in eukaryotes. The enzymes and candidate genes are shown in italics.

137

Here we first report the identification and characterization of two different UDP-GlcNAc nucleotidyltransferases in Arabidopsis that converts GlcNAc-1-P and GalNAc-1-P into their corresponding UDP-sugars. The Arabidopsis GlcNAc1pUT provides an opportunity to understand how pool of UDP-GlcNAc and perhaps UDP-GalNAc are controlled and affect signaling event in the cytosol and nucleus.

EXPERIMENTAL PROCEDURES

cDNA cloning of Arabidopsis GlcNAc-1-P uridylyltransferase (GlcNAc1pUT) - Total

RNA from young rosette leaves of Arabidopsis thaliana Columbia ecotype was extracted using

Trizol reagent (Invitrogen) and used as template for reverse transcription reaction using 0.2 µM

Oligo-dT17V primer [(v=dG/dA/or dC)], dNTPs and 200 units of SuperScript II-reverse transcriptase (Invitrogen). BLAST analyses with human GlcNAc-1-P uridylyltransferase reveals several potential homologous proteins in plants and subsequently primers for cloning these putative genes were designed. The coding sequence of one gene corresponding to At1g31070 loci (herein abbr, GlcNAc1pUT-1) was amplified by PCR using 1 unit of high-fidelity proof- reading Platinum DNA polymerase (Invitrogen), and 0.2 µM of each forward and reverse primers: 5'- C ACC atg Gta gaa ccg tcg atg gag aga g -3' and 5'- GGA TCC agg gaa att tca caa ggt gca tg -3' using cDNA of Arabidopsis as template (the primer sequence in uppercase represents added bases to facilitate cloning). The PCR product was cloned to generate plasmid pCR4-topoTA:at1g31070.11#6; and DNA was sequenced (GenBank GU937393). The NcoI-

BamHI fragment containing the full length GlcNAc1pUT-1 gene without the stop codon was sub-cloned into a pET28b E. coli expression vector so that upon expression the recombinant

138

enzyme will have a six-histidine extension at its C-terminal. Similarly pET28c:at2g35020#13

(GlcNAc1pUT-2) was generated with 6H-extension at the N-terminus region.

Protein expression and purification - E. coli cells, BL21(de3)plysS derived strain, harboring the GlcNAc1pUT expression construct or an empty vector control, were cultured for

16 h at 37 ºC in LB medium (20 ml) supplemented with kanamycin (50 µg/ml) and chloramphenicol (34 µg/ml). A portion (8 ml) of the cultured cells was transferred into fresh 250 ml LB liquid medium supplemented with the same antibiotics, and the cells then grown at 37 ºC at 250 rpm until the cell density reached A600 = 0.6. The cultures were then transferred to 18 ºC and gene expression was induced by the addition of isopropyl β-D-thiogalactoside to a final concentration of 0.5 mM. After 24 h growth while shaking (250 rpm), the cells were centrifuged

(6,000  g for 10 min at 4 ºC), resuspended in lysis buffer (10 ml 50 mM Tris-HCl, pH 7.6, containing 10% (v/v) glycerol, 50 mM NaCl, 0.5 mM EDTA, 1 mM DTT and 0.5 mM phenylmethylsulfonyl fluoride) and lysed in an ice bath by 24 sonication cycles each (10-sec pulse; 20-sec rest) using a Misonix S-4000 (Misonix incorporated, Farmingdale, New York) equipped with 1/8” microtip probe. The lysed cells were centrifuged at 4 ºC for 30 min at 20,000

 g, and the supernatant (termed s20) was recovered and kept at -20 ºC. His-tagged proteins were purified on a column (10 mm id  150 mm long) containing Ni-Sepharose (2 ml, Qiagen) that was previously equilibrated with a buffer: 50 mM sodium-phosphate, pH 7.6, 0.3 M NaCl. The bound His-tagged protein was eluted with the same buffer containing increasing concentrations of imidazole. The fractions containing GlcNAc1pUT activities were stored in aliquots at -80 ºC.

The concentration of protein was determined using bovine serum albumin (BSA) as standard.

The molecular weight of the recombinant protein was estimated by size-exclusion chromatography using Superdex-200 column (1 cm i.d  30 cm long, GE Healthcare) with buffer

139

composed of 0.1M sodium phosphate, pH 7.6 and 0.1 M NaCl. Separate solutions (0.5 ml) of recombinant GlcNAc1pUT or a mixture of standard proteins [10 mg each of alcohol dehydrogenase (157 kDa), bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), and cytochrome C (12.4 kDa)] were separately injected via a Waters 626 LC HPLC system equipped with a photodiode array detector (PDA 996) and a Waters Millennium32 workstation, at 0.5 ml

-1 min . The eluant was monitored at A280 nm and the fractions were collected every 15 sec.

Fractions containing enzyme activity were analyzed and kept at -80 ºC. The native molecular mass of recombinant GlcNAc1pUT was determined based on the elution time and the calibration curve of the above standard proteins.

Enzyme assays - Unless otherwise mentioned, the typical forward HPLC-based reactions for the formation of UDP-GlcNAc were carried out in a final volume of 50 µl and consisted of 1 mM GlcNAc-1-P, 1 mM UTP, 5 mM MgCl2, 100 mM Tris-HCl pH 7.6, 1U of yeast inorganic pyrophosphatase (Sigma), and recombinant GlcNAc1pUT (420 ng). After 10 min incubation at

37 ºC, reactions were terminated (1 min at 100 ºC). Chloroform (50 µl) was added and after vortexing (30 seconds) and centrifugation (12,000 rpm for 5 min, at room temperature), the entire upper aqueous phase was collected and subjected to chromatography on anion-exchange chromatography using TSK-DEAD-5-PW column (7.5mm inner diameter × 75 mm long, Bio-

Rad) and ammonium formate HPLC gradient system (Yang et al., 2009). Nucleotides and nucleotide sugars were detected by their UV absorbance using photodiode array detector that was connected to the HPLC system. The maximum absorbance for uridine nucleotides and UDP- sugars were 261.8 nm in ammonium formate. The peak area of analytes was determined based on standard calibration curves. HPLC-based reverse reactions were carried out in a similar

140

manner and included 1 mM PPi, 1 mM UDP-GlcNAc, 5 mM MgCl2, 100 mM Tris-HCl pH 7.6, and 420 ng GlcNAc1pUT.

Real-time 1H-NMR analysis of GlcNAc1pUT-1 activities - The forward reaction in the final volumes of 180 µl at mixture of D2O:H2O of 8:1 (v/v) consisted of 0.1 M sodium phosphate, pH 7.6, 5 mM MgCl2, 2 mM UTP, 2 mM GlcNAc-1-P, and enzymes: 1.05 µg recombinant GlcNAc1pUT-1 and 1U of yeast inorganic pyrophosphatase in H2O-buffer. For the reverse PPase reaction, 2.1 µg recombinant GlcNAc1pUT-1 was incubated with 2 mM PPi, 2 mM UDP-GalNAc (or UDP-GlcNAc). Immediately upon addition of the enzyme, the reaction mixture was transferred to a 3 mm NMR tube. Real-time 1H-NMR spectra were obtained using a

Varian Inova 600MHz spectrometer equipped with a cryogenic probe. Data acquisition was not started until approximately 2 minutes after the addition of enzyme to the reaction mixture due to spectrometer set-up requirements (shimming). Sequential 1D proton spectra were acquired over the course of the enzymatic reaction. All spectra were referenced to the water resonance at 4.765 ppm downfield of 2,2-dimethyl-2-silapentane-5-sulphonate (DSS). Processing of the data as covariance matrices was performed with Matlab (The Mathworks, Inc.).

Enzyme properties and inhibition assays - The forward nucleotidyltransferase activity of

GlcNAc1pUT-1 was measured with various buffers, at different temperatures, different ions, or with different potential inhibitors. For the optimal pH experiments, 420 ng recombinant enzyme was first mixed with 5 mM MgCl2 and 100 mM of each individual buffer (Tris-HCl, phosphate,

4-morpholineethanesulfonic acid (MES), 4-morpholinepropanesulfonic acid (MOPS) or 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)). The optimal pH assays were initiated after the addition of specific GlcNAc-1-P and UTP. Inhibitor assays were performed under the standard assay conditions except for the addition of various additives (sugars, nucleotides, and

141

antibiotics) to the reaction buffer. These assays were incubated for 10 min at 37 ºC, and were subsequently terminated by heat (1 min at 100 ºC). The amount of UDP-sugar formed was calculated from a calibration curve of HPLC UV spectra of standards. For the experiments aimed at defining the optimal temperature, assays were performed under standard assay conditions except that reactions were incubated at different temperatures for 10 min. Subsequently, the activities were terminated (100 ºC). For the experiments aimed at determining if GlcNAc1pUT-

1 required metals, assays were performed with UTP, GlcNAc-1-P with a variety of ions. After 10 min at 37 ºC incubation the activity was terminated by heat. The amount of UDP-GlcNAc formed was calculated from HPLC UV spectra of standards.

For the experiments aimed at determining the ability of GlcNAc1pUT to utilize other sugar-1-Ps, the std assay was modified by substituting the GlcNAc-1-P with different sugar-1-Ps

(for example, GlcA-1-P, Gal-1-P, etc), and reaction were incubated for 60 min at 37 ºC.

Kinetics - The forward GlcNAc1pUT catalytic activity was determined at 37 ºC for 4 min using 0.1 M Tris-HCl pH 7.6, 5 mM MgCl2, 1 mM GlcNAc-1-P, recombinant GlcNAc1pUT-1

(0.9 pmol, 50 ng) or recombinant GlcNAc1pUT-2 (0.6 pmol, 38 ng), and variable concentrations of UTP (20 µM to 4 mM) or with fixed UTP (1 mM) and variable concentrations of GlcNAc-1-P

(20 µM to 4 mM). The forward kinetic assays included 2 units of yeast inorganic pyrophosphatase to deplete PPi. Kinetics for the reverse reactions were performed in the same condition as above, with a fixed concentration of PPi (1 mM) and variable concentrations of

UDP-GlcNAc (20 µM to 4 mM), and recombinant GlcNAc1pUT-1 (0.9 pmol, 50 ng) or recombinant GlcNAc1pUT-2 (0.3 pmol, 19 ng). In a separate series of reverse-reaction experiments assays were performed at 37 ºC for 10 min with a fixed concentration of PPi (1 mM) and variable concentrations of UDP-GalNAc (200 µM to 10 mM), and recombinant

142

GlcNAc1pUT-1 (1.8 pmol, 100 ng) or recombinant GlcNAc1pUT-2 (0.6 pmol, 38 ng). The kinetics of GlcNAc1pUT-2 for Glc-1-P was essentially described as above. Enzyme velocity data of the amount (µM) of UDP-sugar produced per second, as a function of substrate concentrations was plotted. The Solver tool (Excel version 11.5 program) was used to generate best-fit curve calculated by nonlinear regression analyses, and for the calculation of Vmax and apparent Km. The amount of enzyme used and the incubation time were adjusted so that the reactions were in the linear range.

Structural Prediction for GlcNAc1pUT - Three-dimensional structure models for

GlcNAc1pUT were predicted using both the “Protein Homology/analogY Recognition Engine”

PHYRE web (sbg.bio.ic.ac.uk/~phyre/) server (Kelley and Sternberg, 2009) and the SWISS-

MODEL comparative protein-modeling server (swissmodel.expasy.org/). Structural models were superimposed with known crystal structures from human AGX1 (1jv1a.pdb) (Peneff et al.,

2001). All models and protein structures were visualized using PyMol. Amino acid alignments of

GlcNAc1pUT sequence to fold library protein were performed using the PHYRE program.

RESULTS

Identification, cloning and characterization of Arabidopsis GlcNAc-1-P uridylyltransferase (GlcNAc1pUT-1) - The enzyme nomenclature for NDP-sugar pyrophosphorylase (abbreviated PPase) is the same as for sugar-1-P nucleotidyltransferases. To identify potential functional GlcNAc-1-P nucleotidyltransferase homolog in plants, we compared the human AGX1 protein sequence (NP_003106) with NR-sequence database. BLAST analyses of homologous proteins from different species revealed that Arabidopsis genome loci At1g31070 shares overall a low sequence identity with functional UDP-Glc PPase and UDP-sugar PPase

143

(24% and 26% respectively), and about 34% and 40% aa sequence identity with human and yeast functional UDP-GlcNAc PPase respectively (Mio et al., 1998; Peneff et al., 2001). Interestingly, the functional bacterial glmU (UDP-GlcNAc PPase) shares sequence similarity with plant GDP- man PPase (At2g39770), rather than At1g31070. This suggests that sequence alone is not sufficient to predict function, nevertheless after amino acid sequence alignment of these homologous proteins with At1g31070 (herein, abbr. GlcNAc1pUT-1) we have identified two consensus motifs (Fig. S4.1). The N-terminal region of GlcNAc1pUT-1 comprised of a putative nucleotide binding (NB) motif „LXGG(L/Q)G(E/T)(R/T)(L/M)GX3(I/P)K‟, with a predicted coil structure starting at aa 134 and a „PXGXG‟ motif probably involved in „uridine-binding‟ (UB) starting at aa 250. Phylogenetic analysis indicated that the UDP-GlcNAc PPase-like proteins from different species are distinguished from other nucleotidyltransferases involved in UDP-Glc and UDP-sugar synthesis (data not shown). In addition to At1g31070, another gene, At2g35020, with 86% identity to UDP-GlcNAc PPase was also identified (herein, abbr. GlcNAc1pUT-2). To determine however, if the Arabidopsis GlcNAc1pUT encode such activity, despite the low sequence identity to the human, yeast and bacterial enzymes, and more specifically to establish if the enzymes have different or similar range of sugar-1-Ps and NTP specificity, the genes were cloned and the recombinant proteins expressed in E. coli were analyzed.

A highly expressed protein band (58 kDa) was detected after SDS-PAGE analysis of E. coli cells expressing GlcNAc1pUT-1 (Fig. 4.2A, lane 2; marked by the arrow) when compared with control cells expressing empty vector (Fig. 4.2A, lane 3). The apparent mass of the column- purified enzyme as determined by SDS-PAGE analysis (Fig. 4.2A, lane 4) is in agreement with the calculated mass of the translated gene product fused at the C-terminal portion to 6His.

Preliminary HPLC-based forward assays (Fig. 4.2B, panel 3) demonstrated that the recombinant

144

GlcNAc1pUT-1 converts GlcNAc-1-P and UTP in the presence of Mg2+ to UDP-GlcNAc. The enzymatic product eluted at 11.7 min (Fig. 4.2B, panel 3) had the same retention time as standard

UDP-GlcNAc, while control cells expressing empty vector had no detectable activity (Fig. 4.2B, panel 4). In the reverse assay, recombinant enzyme converts UDP-GlcNAc and PPi to UTP (Fig.

4.2B, panel 5), while control had no activity (Fig. 4.2B, panel 6). The HPLC-peak marked #1

(Fig. 4.2B) was collected from the column, and its structure and identity were confirmed by 1H-

NMR spectroscopy (Fig. S4.2). The chemical shift assignments for UDP-GlcNAc are summarized in Table 4.1. The diagnostic J1, 2 coupling value of 3 Hz and J2, 3, J3, 4, J4, 5 values of

10, 9.7, 9.7 Hz respectively, indicates α-glucopyranose configuration. The linkage of the anomeric GlcNAc residue with the phosphate is given by the coupling constant values 7 and 3

Hz for JP beta, H1 and JP beta, H2 respectively, and also supported by the chemical shift of H1 (5.50 ppm). In addition, the NAc-C=O resonance was assigned 2 ppm and the H2 chemical shift (3.98 ppm) is consistent with the diagnostic 2-acetamido-2-deoxy group. Collectively these results provide unambiguous evidence that the enzyme product of GlcNAc1pUT-1 is UDP-α-D-

GlcNAc.

145

146

FIGURE 4.2. Recombinant GlcNAc1pUT-1 Protein purification and HPLC-based assays.

A. SDS-PAGE of total soluble protein isolated from E. coli cell expressing recombinant

GlcNAc1pUT-1 (lane 2), control empty vector (lane 3), and Nickel-column purified recombinant

GlcNAc1pUT-1 (lane 4), and control empty vector (lane 5).

B. HPLC-based assays of GlcNAc1pUT-1. Forward assays included UTP and GlcNAc-1-P with either recombinant GlcNAc1pUT-1 (panel 3) or with protein extracted from cell expressing

147

empty vector (panel 4). Reverse assays included PPi and UDP-GlcNAc or UDP-GalNAc with either GlcNAc1pUT-1 (panel 5 and 7) or control empty vector protein (Panel 6 and 8). Panel 1 shows chromatography of UDP-GlcNAc and UTP standards; Panel 2 shows UDP-GalNAc and

UTP standards.

The HPLC peaks in the panels, based on retention time are UDP-GlcNAc (11.7 min, marked as

#1), UDP-GalNAc (11.7 min) and UTP (16.1 min). The minor peak marked as “a” at 14.5 min is

UDP contamination from the UTP reagent. The peaks labeled by the arrows represent the nucleotide-based enzymatic reaction products. PPi and GlcNAc-1-P are not UV-visible.

148

TABLE 4.1. Proton chemical shifts and coupling constants of UDP-GlcNAc formed from

GlcNAc-1-P by recombinant GlcNAc1pUT.

Proton H1 H2 H3 H4 H5 H6, 6

GlcNAc

Chemical Shift,  (ppma) 5.50 3.98 3.82 3.54 3.92 3.86, 3.79

J coupling constants J1, P 7 J1, 2 3 J2, 3 10 J2, P 3 J3, 4 9.7 J4, 5 9.7 J5, 6 3.5 J6, 6 12

Rib

Chemical Shift,  (ppma) 5.97 4.37 4.36 4.28 4.24, 4.18 -

J coupling constants J1‟, 2‟ 4.5 - - - - -

Uracil

Chemical Shift,  (ppma) - - - - 5.96 7.95

J coupling constants - - - - J5”, 6” 8.2 -

GlcNAc1pUT enzymatic product was purified from HPLC column, essentially as described in

Fig. 4.2. The product peak was collected, lyophilized, resuspended in D2O and analyzed by proton NMR. a Chemical shifts are in ppm relative to internal DSS signal set at 0 ppm. GlcNAc proton-proton coupling constants in hertz are indicated as well as the J1, P and J2, P coupling values between phosphate and the H1 and H2 protons of GlcNAc.

149

Interestingly, the enzyme can also utilize the 4-epimer UDP-GalNAc as substrate (Fig.

4.2B, panel 7). The forward reaction with GalNAc-1-P could not be performed, as this amino- sugar-1-P is commercially unavailable. Based on these analyses we propose that GlcNAc1pUT-1 is a uridylyltransferase (UT) or depending on other enzyme nomenclature a pyrophosphorylase

(PPase), and like the human AGX1 enzyme, has same specific activities with UDP-GlcNAc and

UDP-GalNAc as substrates.

Characterization and properties of GlcNAc1pUT-1 - GlcNAc1pUT-1 requires divalent cations such as Mg2+ or Mn2+ (Table 4.2), and the activity, as expected, is abolished in the presence of EDTA. However, cations like Ca2+ for example, cannot substitute for magnesium.

The recombinant GlcNAc1pUT-1 catalyze the conversion of GlcNAc-1-P to UDP-GlcNAc over a wide range of pH (3.3 to pH 9.0, see Fig. 4.3), with maximum activity observed at pH 7.6 to

8.0 in either Tris-HCl, or at pH 7.6 in phosphate buffer. The enzyme is also active when reactions were performed in HEPES, MOPS or MES buffers (data not shown) at a similar range of neutral pH. GlcNAc1pUT-1 is also active over a broad range of temperature (0 ºC and 65 ºC), but has maximum activity at 30 - 37 ºC (Table 4.3).

150

FIGURE 4.3. The effects of buffer and pH on GlcNAc1pUT-1 activity.

GlcNAc1pUT-1 activity was determined at different buffers at different pH. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

151

TABLE 4.2. GlcNAc1pUT-1 requires metal for activity.

Relative GlcNAc1pUT Activity Additive (5 mM) Forward reaction (%)

MgCl2 100

MnCl2 126

CaCl2 8

ZnSO4 9

CuCl2 8

EDTA 11 water 8

GlcNAc1pUT-1 was mixed with additive (metal, EDTA, or water control) for 10 min on ice.

Subsequently, UTP and GlcNAc-1-P were added and assays were carried out under standard conditions. Each value is the mean of duplicate reactions, and the values varied by no more than

±5%.

152

TABLE 4.3. The effect of temperature on GlcNAc1pUT-1 activity.

Relative GlcNAc1pUT Temperature (ºC) Activity (%)

4 29

25 69

30 100

37 100

42 60

55 24

65 15

The enzymatic reactions were performed under standard conditions except for the reaction temperature. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

153

We next investigated the nucleotide triphosphates (NTPs) specificity of GlcNAc1pUT-1.

Neither CTP, GTP, ITP, nor ATP is the substrate for the recombinant enzyme when using

GlcNAc-1-P as substrate. Several commercially available sugar-1-Ps were tested as substrates for GlcNAc1pUT-1 with UTP. Reaction carried out with Glc-1-P, GlcA-1-P, Xyl-1-P, Gal-1-P,

GalA-1-P, Fuc-1-P, Man-1-P, GalN-1-P, Fru-1-P, Mannitol-1-P, Glc-6-P as substrates failed to show any conversion even when assays expanded for longer incubation time. To determine if

GlcNAc1pUT-1 may recognize other NTPs we performed our standard assays in the presence of competing nucleotides such as (ATP, CTP, GTP, and ITP at 0.5 mM each). In all cases GlcNAc-

1-P was readily uridylated, suggesting that the enzyme other than UTP cannot recognize other

NTPs. These experiments demonstrate the enzyme preferences for GlcNAc-1-P and GalNAc-1-P as its substrates along with UTP. To determine if the enzyme recognizes and binds nucleotide diphosphates (NDP such as UDP or ADP), prior to the standard assay, the enzyme was incubated with NDPs. These NDPs as well as other nucleotides tested (e.g., NMP, NAD, and NADH) had no effect on GlcNAc1pUT-1 activity. By contrast, the forward enzyme activity was reduced by

56% in the presence of 0.5 mM PPi (Table S4.1). Many antimicrobial reagents are aminoglycoside-derivatives. To determine if the GlcNAc1pUT-1 activity is affected by these aminoglycosides, the enzyme was first incubated with individual antibiotics (gentamycin, kanamycin, hygromycin and streptomycin) and subsequently the std assay was performed.

Interestingly, the enzyme activity was dramatically reduced by 36% and 74% in the presence of

0.5 mg/ml hygromycin and streptomycin respectively (Table S4.1), while on the other hand gentamycin and kanamycin had no affect on plant GlcNAc1pUT-1 activity.

Real-Time 1H-NMR analysis of GlcNAc1pUT-1 - To monitor the dynamics of the enzymatic reaction and the substrate preference of GlcNAc1pUT-1, we used real-time 1H-NMR

154

spectroscopy (Fig. 4.4). These assays were carried out in phosphate buffer to avoid the proton signals from the Tris reagent. As shown in the time dependent enzymatic progression, conversion of the GlcNAc-1-P (5.34 ppm) into UDP-α-D-GlcNAc (5.50 ppm) is observed (Fig.

4.4A). The NMR-based assay also provides unambiguous evidence that the enzyme can also utilize UDP-GalNAc in the presence of PPi as a substrate (see conversion of the UDP-α-D-

GalNAc (5.54 ppm) into GalNAc-1-P (5.37 ppm)); as shown in Fig. 4.4B.

155

156

FIGURE 4.4. Real-Time 1H-NMR based GlcNAc1pUT-1 assays.

A. GlcNAc1pUT-1 forward activity. The protein was mixed with 2 mM GlcNAc-1-P, phosphate buffer and 2 mM UTP. Approximately, 2 min after enzyme addition and NMR shimming, data were collected. Progressions of enzyme activity covering the anomeric region of the proton

NMR spectrum are shown. The „peak‟ shape for sugar-1-P and UDP-sugar has a quadruple form: the chemical shift for GlcNAc-1-P is 5.34 ppm, and 5.50 ppm for UDP-GlcNAc.

B. GlcNAc1pUT-1 reverse activity. The protein was mixed with 2 mM UDP-GalNAc, phosphate buffer and 2 mM PPi. Approximately, 2 min after enzyme addition and NMR shimming, data were collected. Progressions of enzyme activity covering the anomeric region of the proton

NMR spectrum are shown. The chemical shifts are: GalNAc-1-P at 5.37 ppm, and UDP-GalNAc at 5.54 ppm.

157

GlcNAc1pUT-2 is promiscuous and converts GlcNAc-1-P. GalNAc-1-P and Glc-1-P to corresponding nucleotide sugar - Similar to GlcNAc1pUT-1, a highly expressed protein band

(60 kDa) was detected after SDS-PAGE analysis of E. coli cells expressing GlcNAc1pUT-2

(Fig. 4.5A, lane 2; marked by the arrow) when compared with control cells expressing empty vector (Fig. 4.5A, lane 3). The apparent mass of the column-purified enzyme as determined by

SDS-PAGE analysis (Fig. 4.5A, lane 4) is in agreement with the calculated mass of the translated gene product fused at the N-terminal portion to 6His. HPLC-based forward assays (Fig. 4.5B, panel 3) demonstrated that the recombinant GlcNAc1pUT-2 converts GlcNAc-1-P and UTP in the presence of Mg2+ to UDP-GlcNAc. The enzymatic product eluted at 11.7 min (Fig. 4.5B, panel 3) had the same retention time as standard UDP-GlcNAc, while control cells expressing empty vector had no detectable activity (Fig. 4.5B, panel 4). In the reverse assay, recombinant enzyme converts UDP-GalNAc and PPi to UTP (Fig. 4.5B, panel 5), while control had no activity (Fig. 4.5B, panel 6). Interestingly, and unlike GlcNAc1pUT-1, recombinant

GlcNAc1pUT-2 can also utilize the Glc-1-P as substrate, converting Glc-1-P and UTP in the presence of Mg2+ to UDP-Glc. (Fig. 4.5B, panel 7). The HPLC-peak marked #2 (Fig. 4.5B) was collected from the column, and its structure and identity was confirmed as UDP-α-D-glucose by

1H-NMR spectroscopy (Fig. S4.3).

158

159

FIGURE 4.5. Recombinant GlcNAc1pUT-2 Protein purification and HPLC-based assays.

A. SDS-PAGE of total soluble protein isolated from E. coli cell expressing recombinant

GlcNAc1pUT-2 (lane 2), control empty vector (lane 3), and Nickel-column purified recombinant

GlcNAc1pUT-2 (lane 4), and control empty vector (lane 5).

B. HPLC-based assays of GlcNAc1pUT-2. Forward assays included UTP and GlcNAc-1-P or

Glc-1-P with either recombinant GlcNAc1pUT-2 (panel 3 and 7) or with protein extracted from

160

cell expressing empty vector (panel 4 and 8). Reverse assays included PPi and UDP-GalNAc with either GlcNAc1pUT-2 (panel 5) or control empty vector protein (Panel 6). Panel 1 shows

UDP-GlcNAc and UDP-Glc standards; Panel 2 shows UDP-GalNAc and UTP standards.

The HPLC peaks in the panels, based on retention time are UDP-GlcNAc (11.7 min), UDP-Glc

(12.3 min, marked as #2), UDP-GalNAc (11.7 min) and UTP (16.1 min). The minor peak marked as “a” at 14.5 min is UDP contamination from the UTP reagent. The peaks labeled by the arrows represent the enzymatic products.

161

Determination of kinetic parameters of GlcNAc1pUT - Kinetics analyses of the two enzymes are summarized in Table 4.4. The data from experiments with GlcNAc-1-P as the variable substrate and fixed concentration of UTP (or vise versa) fit well to the Michaelis-

Menten model. For GlcNAc1pUT-1, the apparent Km values for the forward reaction were 337

-1 -1 -1 µM (GlcNAc-1-P) and 295 µM (UTP), with Vmax values of 0.69 µM s , and kcat/Km (µM s ) were 0.11 (GlcNAc-1-P) and 0.12 (UTP). The kinetic for the reverse reaction, had an apparent

Km values of 219 µM (UDP-GlcNAc) and 2,768 µM (UDP-GalNAc), with Vmax values of 0.93

-1 -1 -1 -1 µM s (UDP-GlcNAc) and 0.75 µM s (UDP-GalNAc). The kcat/Km (µM s ) values are 0.22

(UDP-GlcNAc) and 0.015 (UDP-GalNAc). For GlcNAc1pUT-2, the apparent Km values for the

- forward reaction were 180 µM (GlcNAc-1-P) and 203 µM (UTP), with Vmax values of 0.36 µM s

1 -1 -1 , and kcat/Km (µM s ) were 0.16 (GlcNAc-1-P) and 0.14 (UTP). The kinetic for the reverse reaction, had an apparent Km values of 65 µM (UDP-GlcNAc) and 808 µM (UDP-GalNAc), with

-1 -1 -1 - Vmax values of 0.19 µM s (UDP-GlcNAc) and 0.18 µM s (UDP-GalNAc). The kcat/Km (µM s

1) values are 0.46 (UDP-GlcNAc) and 0.018 (UDP-GalNAc).

162

TABLE 4.4. Enzyme kinetics of GlcNAc1pUT.

Forward reaction

Km Km kcat/Km kcat/Km

(GlcNAc-1-P) (UTP) (GlcNAc-1-P) (UTP)

µM µM µM -1 s-1 µM -1 s-1

GlcNAc1pUT-1 337 295 0.11 0.12

GlcNAc1pUT-2 180 203 0.16 0.14

Reverse reaction

Km Km kcat/Km kcat/Km

(UDP-GlcNAc) (UDP-GalNAc) (UDP-GlcNAc) (UDP-GalNAc)

µM µM µM -1 s-1 µM -1 s-1

GlcNAc1pUT-1 219 2768 0.22 0.015

GlcNAc1pUT-2 65 808 0.46 0.018

The forward kinetic GlcNAc1pUT reactions were measured for 4 min with varied concentrations of GlcNAc-1-P (0.02-4 mM) and UTP (0.02-4 mM). In the reverse reaction, the enzymatic activity was measured with varied concentrations of UDP-GlcNAc (0.02-4 mM) for 4 min and

UDP-GalNAc (0.2-10 mM) for 10 min. Enzyme velocities were plotted and Solver software was used to generate best-fit curve and for the calculation of Vmax and apparent Km. Each value is the mean of triplicate reactions, and the values varied by no more than ±10%. The Km of

GlcNAc1pUT-2 for UTP is 928 µM when Glc-1-P is the substrate, with kcat/Km value of 3.1E-04

µM -1 s-1.

163

Structure prediction for GlcNAc1pUT - Chromatography of GlcNAc1pUT-1 on gel filtration column showed that the recombinant enzyme has an apparent MW of 60 kDa indicating that the native enzyme is a monomer in solution since the predicted molecular weight for monomeric protein is 58.3 kDa. Whereas the human AGX1 and yeast QRI1 appears to form a complex as a dimer in solution. Three-dimensional model of Arabidopsis GlcNAc1pUT-1 was generated by SWISS-MODEL (Arnold et al., 2006) and PHYRE prediction, using human

GlcNAc1p nucleotidyltransferase (AGX1) complexed with UDP-GlcNAc (1jv1.pdb) as the most appropriate template (Fig. 4.6A). The overall structure of the model is similar to the fold of the human AGX1. Based on the model, the predicted amino acid residues of the plant GlcNAc1pUT-

1 involved in binding of the uracil, ribose, as well as GlcNAc are shown in Table 4.5. The positions of those amino acids are identical to the corresponding amino acids in the crystal structure of AGX1. This suggests that although GlcNAc1pUT-1 shares only 32% amino acid identity to AGX1, its structure is conserved, albeit significant divergence of amino acids between the two enzymes. A three-dimensional model of Arabidopsis GlcNAc1pUT-2 was also generated using AGX1 as the template. The two Arabidopsis GlcNAc1pUTs have very similar structural fold (Fig. 4.6C), with some changes in various loop regions, see Fig. 4.6C, a-e. For example, the loop designated „e‟ (between amino acid Pro312 and Gly317) on GlcNAc1pUT-2 in approximate to the GlcNAc sugar moiety is modeled differently compared with the GlcNAc1pUT-1. Other loop variations are also observed between the two Arabidopsis proteins.

164

165

FIGURE 4.6. Structural alignment of AGX1 and homology 3D-model of GlcNAc1pUT-1 and GlcNAc1pUT-2

A. Ribbon represents the superimposed structure of the UDP-GlcNAc-complexed form of AGX1 and GlcNAc1pUT-1. AGX1 is colored in salmon and GlcNAc1pUT-1 is colored in cyan. The non-conserved structure between AGX1 and GlcNAc1pUT-1 are highlighted in magenta and blue respectively. The UB loop involved in uracil binding is highlighted in green; and the NB loop involved in nucleotide binding is highlighted in yellow.

B. Predicted amino acid residues in the active site of GlcNAc1pUT-1 complexed with UDP-

GlcNAc. Amino acid residues involved in the NB loop are colored yellow; aa involved in the UB loop are colored green; and aa involved in the binding of hexosamine are colored blue.

166

Magnesium probably coordinates the transition by interacting with the alpha and beta phosphate of UDP-GlcNAc.

C. Ribbon represents the superimposed structure of the UDP-GlcNAc-complexed form of

GlcNAc1pUT-1 and GlcNAc1pUT-2. GlcNAc1pUT-1 is colored in Cyan and GlcNAc1pUT-2 is colored in Orange. The non-conserved structure between GlcNAc1pUT-1 and GlcNAc1pUT-2 are highlighted in blue and yellow respectively, and labeled a-e.

167

TABLE 4.5. The predicted UDP-GlcNAc binding site of GlcNAc1pUT; the protein atoms that are presumably hydrogen-bonded to the nucleotide-sugar in the structure of

GlcNAc1pUT are listed.

Interacting atoms aa (GlcNAc1pUT-1) aa (GlcNAc1pUT-2) Distance Å Ribose O2‟ G137 G133 3.18 L134 L130 3.30 Ribose O3‟ L134 L130 2.73 V282 V279 3.24 Uracil O2” M195 M192 3.40 Q226 Q223 3.47 G136 G132 2.87 Uracil O4” G252 G249 3.00 Q226 Q223 3.3 P250 P247 3.17 -phosphate N253 N250 3.46 -phosphate Y336 Y333 2.77 K432 K429 2.57 GlcNAc O2 N253 N250 2.76 E335 E332 3.50 GlcNAc O3 E335 E332 2.74 N360 N357 2.73 G320 F319 3.37 /1.26 GlcNAc O4 N360 N357 2.65 G320 G317 2.96/3.63 V319 V318 3.02/1.97 F319 2.77 GlcNAc O6 K432 K429 3.05

168

DISCUSSION

We have cloned and biochemically characterized an Arabidopsis N-acetylglucosamine-1-

P uridylyltransferase-1 that in the presence of Mg2+ and UTP specifically uridylates GlcNAc-1-P and GalNAc-1-P. The GlcNAc1pUT-1 enzyme is reversible and in the presence of UDP-GlcNAc and PPi for example, will form GlcNAc-1-P and UTP. We also cloned and characterized

At2g35020, abbr. GlcNAc1pUT-2, that encodes an enzyme which uridylates GlcNAc-1-P,

GalNAc-1-P, and in addition Glc-1-P (See Fig. 4.5). Although the two Arabidopsis genes share

86% sequence identity, the latter GlcNAc1pUT-2 has broader substrate specificity. Comparing the activities of GlcNAc1pUT from different organisms, yeast GlcNAc1pUT (i.e. QRI1) can also utilize Glc-1-P as substrate and able to form UDP-Glc (Mio et al., 1998), just like Arabidopsis

GlcNAc1pUT-2. The human AGX1 is more specific for the hexosamines and has both

GlcNAc1p-UT and GalNAc1p-UT activity, just like Arabidopsis GlcNAc1pUT-1 (Peneff et al.,

2001).

In Arabidopsis, two genes encoding UDP-Glc PPases (UGP1, At3g03250 and UGP2,

At5g17310) that share 96% amino acid sequence similarity were described (McCoy et al., 2007;

Meng et al., 2008). Single or double T-DNA knockout mutants of these UGP genes revealed no phenotype. This may be explained in part, by the overexpression of a gene encoding Sloppy, a

UDP-sugar PPase (Litterer et al., 2006; Kotake et al., 2007) as suggested by Meng et al. (Meng et al., 2009). However, we cannot exclude the possibility that in these mutants background, the promiscuous activity of GlcNAc1pUT-2 can contribute to the formation of UDP-glucose as well.

In the context of plant metabolism and physiology, the ability of the Arabidopsis enzyme to utilize UDP-GalNAc in vivo as a substrate requires further studies. There are a few reports that

GalNAc may be a metabolite in plants. For example, in-vitro studies have shown, that some plant

169

UDP-4-glucose epimerases including the UGE from barley (Zhang et al., 2006) interconvert

UDP-GlcNAc and UDP-GalNAc. Relative large amount of UDP-GalNAc have been reported to be present in the leaves of squash (Tolstikov and Fiehn, 2002) and in dahlia tubers (Gonzalez and Pontis, 1963). A recent metabolomic studies clearly demonstrate the existence of both UDP-

GlcNAc and UDP-GalNAc in Arabidopsis and Fenugreek (Alonso et al., 2010). In-vitro assay also demonstrated that some plant hexosaminidases have a minor activity toward p-nitrophenyl-

N-acetylgalactosaminide as a substrate (Gutternigg et al., 2007). However, the existence of

GalNAc residue in the glycans from land plants has not been reported to our knowledge. On the other hand, report from Desmidiaceae, a green algae, described GalNAc moiety in the polysaccharide as determined by lectin binding (Brosch-Salomon et al., 1998). It is possible that the levels of GalNAc residue are very low in plant. As GalNAc and GlcNAc have identical mass, mass spectrometry of plant oligosaccharides released from glycoprotein would not distinguish these isomers. Therefore, more detailed sugar analyses studies will be required to address the occurrence of these glycosyl-residues in plants.

The symbiotic bacteria and plant pathogens (bacteria, insects and nematodes) synthesize glycans that contains GalNAc and GlcNAc residues. Thus it is feasible that plant produce hydrolytic enzymes that cleave such glycans and re-use the monosaccharides released. This may explain the need for a dual plant GalNAc-1-P and GlcNAc-1-P UTs as described in this report

(Fig. 4.2). Clearly further exploration to elucidate the metabolic pathway that may utilize

GalNAc residues in plants is needed. Another important aspect of normal plant metabolism is the recycling of sugars including hexosamines from salvaged plant glycoproteins. How a plant recycles the glycoproteins, cleaves the sugars, and re-utilizes the hexosamines remain to be studied. Current work is underway to understand this cellular process.

170

GlcNAc1pUT-like sequences are found in the genome of prokaryotes and eukaryotes.

Amino acid sequence alignment of several GlcNAc1pUT-like sequences along with UDP-Glc

PPases and UDP-sugar PPases shares very low sequence identity, indicating that sequence alone is not the best criteria to determine the specific function of this class of enzymes. The biochemical data provided here also demonstrate that sequence alone is not even sufficient to predict if an enzyme is selective or has broader substrate specificities. Although the amino-acid sequence is highly divergent in the UT-family of proteins, they appear to have conserved binding and catalytic motifs, suggesting these UT-enzymes have maintained a conserved fold and most likely similar catalytic mechanism throughout evolution. Putative structural models were created using human AGX1 as a template (PDB ID: 1jv1). The two enzymes share a highly conserved structural fold consists of a central domain with / structure resembling the Rossmann fold, with the N-terminal and C-terminal domains flanking at each end of the polypeptide chain (Fig.

4.6A).

Common consensus regions are the glycine rich (NB) motif for nucleotide binding and the uracil binding (UB) motif (Fig. S4.1). The amino-acids in these two motifs that are involved in binding are shown in the 3D-structure homology modeling (Fig. 4.6A) of GlcNAc1pUT-1 using AGX1 as template. The amino acids likely involved in the recognition of the hexosamine moiety are shown in the homology model (Fig. 4.6B). The conserved E355, N360, G320 are likely involved in recognition of C3, and the aa N360, G320, V319 are involved in recognition of

C4 (See Table 4.5). The NAc-sugar residues of hexosamine linked via C2 is likely critical for the recognition of GlcNAc-1-P, and be different than UT that recognized for example, Glc-1-P. This is in agreement with the conserved Asn223 in AGX1 UAP (Peneff et al (Peneff et al., 2001)); in

Arabidopsis UT-1 and UT-2, Asn253 (UT1) and Asn250 (UT2) respectively could be one of the

171

aa that involved in the recognition of NAc and -phosphate. The residues involved in the C-6 recognition include the conserved Lys432 in UT1 and Lys 429 (UT2). We have tried to use

UDP-GlcNAc-uronic acid or UDP-XylNAc as substrates, but the Arabidopsis GlcNAc1pUT-1, 2 failed in utilizing them indicating these Arabidopsis enzymes have a strict binding for the gluco-

C6 sugar moiety. Current work is underway to mutagenize and to crystallize the enzyme with different ligands (e.g., GalNAc-1-P, GlcNAc-1-P, UTP, PPi, Mg) to identify how subtle changes in amino acids allow it to be able to accept the C4-epimer (i.e. Gluco- vs Galacto-configurations).

Comparison of the homology model between GlcNAc1pUT-1 and GlcNAc1pUT-2 indicated both enzymes share a similar fold except for some variations in some loop region (Fig. 4.6C).

The loop in approximate to the sugar moiety between amino acid Pro312 and Gly317 on

GlcNAc1pUT-2 appears different to GlcNAc1pUT-1. Whether this is the actual cause of the

GlcNAc1pUT-2 to take UDP-Glc as substrate remained unknown. Crystal structure is required to solve the promiscuity of GlcNAc1pUT-2. The promiscuity of GlcNAc1pUT-2 to both Glc-1-P and GlcNAc-1-P as substrates is not a unique feature of Arabidopsis, as the yeast PPase (Mio et al., 1998) utilizes both substrates as well.

The biological significance of GlcNAc1p-UT is critical not only as post-translational modification of regulatory proteins, but also as a component of glycolipid and glycoprotein. So far T-DNA insertion in At1g31070 loci has no visible phenotype (Yang, unpublished).

Considering the key role of UDP-GlcNAc in development and regulation, and the fact that two genes encoding this activity appears to be expressed in most Arabidopsis tissues, suggest that both genes need to be knockout for functional analyses. However, in many organisms, the lack of

UDP-GlcNAc is lethal.

172

CHAPTER 5

THE ROLE OF GALAK IN GLYCAN METABOLISM4

4Ting Yang and Maor Bar-Peled, to be submitted to Plant Physiol

173

ABSTRACT

We previously characterized in vitro a recombinant Arabidopsis enzyme (GalAK) that with ATP phosphorylates galacturonic acid to GalA-1-P. To gain insights into the role of salvage pathway, we analyzed the in vivo function of this Arabidopsis gene mutant. galak shows no visible morphological phenotype compared to WT Arabidopsis Landsberg plants grown under laboratory conditions. Moreover, sugar composition of total cell wall polysaccharides isolated from wild type and galak mutant were essentially similar. No gross changes were observed by immunohistochemistry studies when comparing the sections of galak stems, flowers and pollen with WT. On the other hand, feeding galacturonic acid to Arabidopsis seedlings germinating on the agar plates, resulted in a 40 fold accumulation of the monosaccharide GalA in the galak mutant when compared to WT. This suggests that i) WT harbors a functional GalAK gene with the ability to convert GalA to other pathways; ii) the galak mutants are unable, at the seedling stage, to shunt GalA to other pathways; iii) the GalA accumulation is apparently non-toxic to seedlings. GalA released by glycan hydrolysis during germination can serve to provide the flux of sugars to form UDP-GalA and in principle UDP-GlcA, UDP-Xyl, UDP-Arap and UDP-Api.

The fact that seed germination and seedling growth in galak is normal and comparable to WT suggests that UDP-GalA is sufficiently made by other routes, i.e. the inter-conversion pathway.

However, it is possible that total sugar analysis of wall is not sensitive enough to identify slight changes in GalA content. Lastly, we cannot discount the possibility that GalA in seed or seedling is required for the synthesis of non-wall metabolites. Therefore, future analysis will be primarily focused on the metabolome differences of glycolipids and small glycosides between the galak mutant and WT.

174

INTRODUCTION

In plants, D-galacturonic acid (D-GalA) is present in numerous polysaccharides including homogalacturonan (HGA), rhamnogalacturonan I (RG-I), rhamnogalacturonan II (RG-II), xylogalacturonan, apiogalacturonan, at the reducing end of xylan core structure, xyloglucan

(O'Neill and York, 2003), and arabinogalactan proteins (Mohnen, 2002) (Fig. 1.6). GalA containing glycans have significant physiological functions for plant growth and development: they are involved in mediating wall strength (Hu and Brown, 1994; Ishii et al., 2001; O'Neill et al., 2001; Iwai et al., 2002; Iwai et al., 2006), in cell-cell adhesion (Willats et al., 2001; Mohnen,

2008), as well as in intercellular signal transduction and plant morphogenesis (Simpson et al.,

1998; Norman et al., 1999; Casasoli et al., 2008; Falasca et al., 2008; Caffall and Mohnen, 2009).

The synthesis of GalA-containing glycans requires UDP-GalA as a substrate (Sterling et al., 2006). In plants, the biosynthesis of UDP-GalA is achieved through the NDP-sugar interconversion pathway (Gu and Bar-Peled, 2004) and/or, the salvage pathway (Yang et al.,

2009). The former activity occurs inside the Golgi and is catalyzed by various type II membrane- bound reversible UDP-GlcA 4-epimerase isoforms (UGlcAEs) (Gu and Bar-Peled, 2004; Gu,

2009). The salvage pathway, on the other hand, suggests that UDP-GalA is made in the cytosol

(Bar-Peled and O'Neill, 2010) by an ATP-dependent kinase that phosphorylates GalA to GalA-1-

P (Yang et al., 2009). The salvage pathway likely occurs in several steps: GalA is hydrolyzed from extracellular GalA-containing glycans, and then is taken up by the cell (Fig. 1.5).

Subsequently, GalA in the cytosol is phosphorylated by a GalAK and then converted to UDP-

GalA by a UDP-sugar pyrophosphorylase (Sloppy, USP) (Yang et al., 2009).

While two pathways exist for the synthesis of UDP-GalA in different compartments (i.e.

Golgi and cytosol), no information has been reported about the relative contribution of each

175

pathway for the synthesis of wall GalA-containing glycans or for the synthesis of secondary metabolites containing GalA. It is also unclear if the metabolism of UDP-GalA varies in different plant tissues. We initially proposed several hypotheses to explain the role of GalA- salvage pathway: 1) some glycan glycosylation reactions occur at the cytosolic faces of organelles (for example, Golgi, ER, chloroplast, vacuole) for which a cytosolic pool of UDP-

GalA is required; 2) certain glycan synthesis requires UDP-GalA made in the cytosol, so that it can be transported to specific organelles, for example, the Golgi or nucleolus; 3) glycosylation of secondary metabolites requires a pool of cytosolic UDP-GalA; 4) GalA serves as a sensor metabolite, for example, to monitor the ATP, sugar or NDP-sugar levels; 5) The biological role of GalAK differs in different cells, for example, in pollen tube growth and seed germination.

To clarify the metabolism of GalA in plants, and to investigate the role of the salvage pathway in the synthesis of plant glycans, we analyzed plant mutants that lack the expression of

GalAK. We compared the phenotype of the galak mutant and WT plants grown under various conditions including: humid growth chamber, under drought, and when challenged by pathogenic fungus (Botryotinia fuckeliana). We further studied the glycan compositions in different tissues (including flowers, roots, stems, and leaves) in galak and WT plants. We also fed Arabidopsis with GalA to check whether its metabolism was affected in the mutant. And lastly, we complemented the galak mutant line with the 35SP:GalAK gene and examined the phenotype/chemotype and its response to feeding with galacturonic acid.

EXPERIMENTAL PROCEDURES

Isolation and analysis of Arabidopsis galak mutant - galak mutant (GT8007) with the insertion of Ac-Ds transposable element containing GUS reporter gene in Arabidopsis

176

Landsberg (Ler) was obtained from Cold Spring Harbor Laboratory (GSHL). This transposable element mutant construct simultaneously disrupts gene function and monitors the gene expression (Sundaresan et al., 1995; Martienssen, 1998). Total RNA was extracted using Trizol reagent. GT8007 galak mRNA expression level was determined by RT-PCR using GalAK cDNA cloning primers (Primer #1 and #2, see Appendix B, Table B1) and compared with the transcript of GalK (At3g06580) as the control.

GalAK knockdown (KD) mutants were also generated in Populus, using RNAi methodology (McGinnis et al., 2005). These transgenes were generated via the BESC pipeline and the sequence region that was utilized for the RNAi construct spans from nt 1380 to 1580 bp.

Unfortunately, the transgenic RNAi galak Populus plants start to wake up from dormancy on

Apr 2011, so the mutant has not yet been analyzed.

Generation of transgenic plants that overexpress GalAK and the complementation of galak mutants - the generation of plasmid pCR4-topoTA:At3g10700#11 harboring the coding sequence of GalAK from Arabidopsis Columbia cDNA was described in Chapter 2. The PciI-

KpnI fragment from pCR4-topoTA:At3g10700#11 was subcloned into the NcoI/KpnI cloning site of pCAM35stl-EYFP (Pattathil et al., 2005) and into the pCAM35stl-6×His to generate an

EYFP or 6×His tagged fusion protein driven under a strong 35S promoter respectively

(pCAM35stl:GalAK-EYFP and pCAM35stl:GalAK-6×His).

The plasmid constructs were transformed individually to Agrobacterium tumefaciens

GV3101/pM90 strain (Pattathil et al., 2005) and the clones were selected on LB/Kan (50µg/ml).

The GV3101/pM90 strain harboring the specific construct was grown overnight in the liquid media comprised of LB for two days at 28 ºC. Agrobacteria cells were harvested, resuspended in

0.5 L water solution supplemented with 5% Suc and with 0.02% Silwet L-77. WT Arabidopsis

177

Landsberg or galak mutants were grown for 3 weeks until flowers started to emerge. The plant aerial region was dipped (Bechtold and Pelletier, 1998) into 0.5 L solution of the Agrobacteria.

Following transformation, plants were allowed to seed. The dry seeds were harvested, and 2000 seeds (equal in weight to 40 mg) were resuspended in 4 ml 0.1% agarose, and plated on 15 cm

MS-agar plates supplemented with 0.25% Suc and 15 μg/ml hygromycin. The agar on the plate was allowed to „dry‟ and the plates were cold treated (4 ºC) for 2 days then moved to growth chamber and grown under continuous influorescent light for 10 days at 22 C. The hygromycin- resistant seedlings (2 weeks post germination) were then transferred to soil and grown to maturity in order to collect F1 transgenic seeds.

Molecular characterization of GalAK transgenic lines - Total RNA from 6-week-old

Arabidopsis leaves (derived from WT or transgenic lines) was extracted as described (Gu and

Bar-Peled, 2004). GalAK-EYFP and GalAK-6×His mRNA expression levels in plants (WT and transgenic lines) were determined by RT-PCR using GalAK cDNA cloning primers and EYFP primers (see primer list in Appendix B, Table B1). Two to four individual plant lines were selected for further analyses (see list of plant lines in Appendix B, Table B4)

Preparation of polyclonal antibodies against Arabidopsis Sloppy and GalAK -

Recombinant proteins (sloppy and GalAK) were generated by E. coli, extracted and purified as described ((Yang et al., 2009) and in Chapter 2). Individually purified proteins (1 mg/ml, total 10 ml) were shipped to Covance and Fisher Scientific for the generation of polyclonal antibodies in rabbit. Based on the Fisher Scientific protocol of immunization, each rabbit was initially injected with 0.5 mg of protein mixed with adjuvant and subsequently the rabbits received 3 boosts (250

µg antigen each over a period of 70 days). In the Covance protocol, the rabbit was injected with three boosts of 125 µg antigen, and over a period of 77 days. Blood was withdrawn and the titer

178

against the antigen (sloppy or GalAK) was quantified by ELISA at the company site. Upon receiving the bleed (i.e. Ab), the specificity of the Ab was determined by Western blot and compared with the control (pre-immunized bleed).

Western blot analyses - Various types of plant tissues were collected from 6-week-old

Arabidopsis plants (tissues including the portion of roots, leaves, stems and some flowers).

Tissues (1 g) were pulverized in liquid nitrogen with a mortar and pestle, and 3 ml cold extraction buffer (50 mM Tris-HCl at pH 7.6, 50 mM NaCl, 1 mM MgCl2, 0.5 mM EDTA-

NaOH, 10% glycerol, supplemented with fresh 1 mM DTT and 0.1 M PMSF) was added. After 5 min homogenization on ice, the homogenate was centrifuged at 1,000 g for 10 min at 4 C and the supernatant layer (S1) was re-centrifuged at 100,000 g for 60 min at 4 C. The resulting supernatant (termed S100) and the pellet (P100, microsomes) were separated. The microsome fraction (P100) was further resuspended with TX-extraction buffer (1% Triton X-100 in 50 mM

HEPES at pH 7.6). The amount of total proteins was quantified using Bradford reagent and the bovine serum albumin (BSA) protein was used as the standard. The activity of GalAK was performed as described (Yang et al., 2009), and the relative amount of the GalAK protein in tissues was examined using Western blot analyses with the known amount of pure recombinant protein as std.

For Western blot analyses, protein samples (80 µg per lane) were first separated on 0.1%

SDS-10% PAGE (Laemmli, 1970). Proteins in gels were electro-transferred onto 0.45 μm PVDF

(polyvinylidene fluoride) membranes in a Bio-Rad wet blot apparatus using transfer solution (12 mM Tris, 96 mM glycine and 20% methanol) at 35 V (constant voltage) for 3 hours at 4 ºC. After transfer, membranes were incubated for 2 hours with the blocking reagent [5% (w/v) dry milk in

TBST (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 0.1% Tween-20)], and washed four times with

179

TBST. Blots were subsequently reacted with polyclonal anti-recombinant AtGalAK serum diluted 1:1,000 in the blocking reagent for 2 h. Blots were then washed four times with TBST, and reacted for 1 h with goat-anti rabbit secondary antibodies linked to alkaline phosphatase

(Sigma) at 1:5,000 dilution in the blocking reagent. Blots were washed four times with TBST, and then incubated with alkaline phosphatase color reagents (NBT/BCIP, Pierce) for up to 10 min. After color was developed, blots were washed with water, dried, and scanned digitally with

EPSON scanner 1240 or digital picture of the blot was taken with iPhone version 3GS.

Glycosyl residue composition analyses and the preparation of plant material for wall and for 80%-soluble carbohydrates - 6-week-old Arabidopsis Landsberg tissues (stems, roots, leaves, or flowers) (100 mg) were pulverized in liquid nitrogen with a mortar and pestle. The frozen powder was transferred to an eppendorf tube. 1 ml cold 80% ethanol was added and vortexed at room temperature for 30 sec to homogenize the sample. After centrifugation (10,000 rpm at room temperature for 5 min), the supernatant was collected and termed „80%-soluble carbohydrates‟. This fraction contains most of the monosaccharide. The 80% ETOH extraction and centrifugation step was repeated twice, and the supernatants were combined. Subsequently 1 ml ice-cold 100% ETOH was added to the pellet and centrifuged, the supernatant was discarded.

Then 1 ml cold Methanol-Chloroform (Vol:Vol=1:1) was added to the pellet, vortexed and centrifuged, the supernatant was discarded. Finally 1 ml cold acetone was added to the pellet, vortexed and centrifuged, the supernatant was discarded. The residual pellet after the organic extraction, (termed alcohol insoluble residue, AIR) was then dried under stream of air in the chemical hood at room temperature for overnight.

Determination of monosaccharide contents and composition of Arabidopsis wall glycans by TMS - For wall analyses, different tissues were used including roots, stems, leaves, and

180

flowers from 6-week-old Arabidopsis WT Landsberg and galak mutant, and from 10-day-old seedlings grown on the MS-agar plates. Wall from each plant sample was obtained by sequential extraction using 80% and 100% ETOH, Methanol-Chloroform=1:1 (v/v) and acetone. Wall (1 mg, supplemented with 20 µg inositol as internal reference) was hydrolyzed with 1 ml reagent of

1 M Methanolic-HCl (80 ºC for 16 h), and the bulk reagent was removed under a stream of filtered air (40 oC) followed by three cycles of rinse (100 µl MeOH) and evaporation. The released methyl-glycosides were acetylated for 45 min in the solution of 200 µl anhydrous

MeOH, 100 µl pyridine, 100 µl acetic anhydride at 50 ºC as described (York et al., 1986). After air-stream evaporation, acetylated sugars were converted to their trimethylsilylation derivative with 150 µl of Tri-Sil reagent (Pierce) for 20 min at 80 ºC. After air evaporation at 22 ºC, sample was dissolved in hexane, sonicated, filtered via glass wool and analyzed by GLC equipped with

FID or MS detectors. Sample (1 µl) was injected via GLC [Agilent 7890A equipped with an auto sampler (Agilent 7693)] onto a DB-1 column (30 m Long × 250 μm Dia ×0.25 μm Film, J&W

122-1032) developed with helium as the gas carrier at a constant flow rate of 1.5 ml/min. The

GC chromatography program started with the column oven temperature held at 80 ºC for 2 min, followed by an increase to 140 ºC at a rate of 20 ºC/min, held at 140 ºC for 2 min; then the temperature was ramped from 140 to 200 ºC (at 2 ºC/min) and held for 0 min; then the temperature was ramped from 200 to 250 ºC (at 30 ºC/min) and held for 5 min. The GC was reconditioned (post run) to 80 ºC and held at 80 ºC for 5 min before next sample injection. The injector was operated in the split mode (1:25), the temperature of the injection port and of the

FID detector was kept at 280 ºC, and FID operated with constant air flow (air at 400 ml/min, H2 at 30 ml/min and N2 at 28.5 ml/min). TMS-derivatives of monosaccharide standards (50 µg each of arabinose, rhamnose, fucose, xylose, GlcA, GalA, mannose, glucose, and GlcNAc) were

181

prepared under the same condition as the samples, and FID or MS was used to identify retention times of individual sugar peaks. In the EI-MS analysis a single monosaccharide provides several

TMS-peaks. Hence, the sum areas of the various peaks belonging to one sugar-derivative were used for the quantization of a sugar in a sample.

Arabidopsis growth condition - Arabidopsis Landsberg seeds were sterilized by 70%

ETOH for 1 min, followed by 10 min immersion in 20% Clorox (household bleach) containing

(0.01% Tween-20), and then rinsed six times with sterilized double de-ionized water (DDW).

Seeds were kept in sterilized-DDW at 4 C for 24 h, and then germinated and grown in sterile 8 cm petri-dish containing 0.8% agar, Murashige and Skoog basal salt media 4.3 g/L (Invitrogen,

11117-074), and 0.25% Suc. The media is adjusted to pH 5.7 by 1 M KOH before adding the agar. The petri-dish was vertically oriented and kept in 22 C for 10 days under 24 h fluorescent light. Seedlings grown on the plates were then transferred to the pots (6×6×6 cm filed with

Fafard 3B soil mix) and grown in a controlled growth chamber at a constant temperature 22 C with a 16-h-light/8-h-dark. Plants were irrigated with water every other day and supplemented once a week with professional all purpose plant food (Peters, NPK).

Growing plants in the presence of various sugar additives - Arabidopsis seeds (WT, galak mutant, overexpression lines, complementation lines) were germinated and grown in the

MS-agar solid media supplemented with either 0.01% GalA (or 0.01% GlcA, or 0.05% myo- inositol) for 10 days, seedlings were then collected and washed twice in DDW. The glycosyl- residue composition of the plant cell wall or of the 80%-ethanol soluble carbohydrate fraction was determined by GC-MS of the derived TMS-methylglycosides.

Tissue fixation and immunohistochemistry - Tissue parts derived from 6-week-old

Arabidopsis Landsberg stems, leaves, flowers, and 10-day-old Arabidopsis seedlings were cut

182

with a razor blade (aprx. 1 cm) and fixed for 16 h at 4 ºC in fixative solution [4% (v/v) paraformaldehyde (Electron Microscopy Sciences Cat#15710), 0.5% (w/v) glutaraldehyde

(Electron Microscopy Sciences Cat#16120), 0.02% Triton X-100, in 25 mM sodium phosphate buffer, pH 7.2]. Samples were washed with 25 mM sodium phosphate buffer (pH 7.2) for three times, 15 min each, followed by two times wash in DDW, 15 min each, and gradually dehydrated by incubating for 25 min each time with different increasing concentrations of ethanol (35%, 50%, 75%, 95%, twice 100%). The dehydrated samples were then moved to 4 C and gradually infiltrated with cold LR White embedding resin (Ladd Research, Cat#12394) starting with 50% resin in 100% ethanol for 24 h, followed by 100% resin for 24 h. The 100% resin step was repeated three times. Tissues were embedded into gel caps containing 100% fresh resin, and the resin was polymerized by exposing the capsules to 365-nm UV light at 4 C for 48 h.

Thin sections (250 nm) were cut by an ultramicrotome (Leica EM UC6) and mounted on coated glass slides (colorfrost/plus; Fisher Scientific) (Pattathil et al., 2010). Immunolabeling was performed at room temperature. Thin sections were blocked with 10 µl of 3% (w/v) non fat milk in KPBS buffer (10 mM potassium phosphate, pH 7.2, 0.5 M NaCl) for 30 min. The blocking reagent was removed and 10 µl of undiluted mAbs against cell wall glycans (Pattathil et al., 2010) were applied to the sections, and incubated for 2 h (monoclonal antibodies used were summarized in Appendix B, Table B3 and obtained from (Pattathil et al., 2010)). Sections were washed with KPBS three times, 5 min each, and subsequently goat-anti-mouse IgG secondary antibody (Invitrogen) diluted 1:100 in KPBS was added and incubated in the dark for 1 h.

Sections were then washed twice with KPBS for 5 min and rinsed with DDW for 5 min. A tiny

183

drop (~1 µl) of citiflour antifade mounting medium AF1 (Electron Microscopy Sciences) was applied on each section. The sections were covered with a cover slip.

Fluorescent light microscopy was performed using an Eclipse 80i microscope (Nikon) equipped with epifluorescence optics. Images were captured using Nikon DS-Ri1 camera via the

NIS-Elements Basic Research software, and were assembled by Adobe Photoshop CS4. The microscopy operation was carried with the help of Dr. Utku Avci in CCRC.

RESULTS

Analyses of galak knockout mutant - GalAK KO mutant was obtained from a collection generated at Cold Spring Harbor Laboratory (Nakayama et al., 2005). The mutant lacking the expression of the native GalAK transcript has been validated by RT-PCR (Fig. 5.1). The galak mutant showed no visible morphological phenotype compared to the WT Arabidopsis Landsberg plants grown under laboratory conditions. Transgenic plants were constructed for overexpression of GalAK using the strong 35S promoter in wild type (i.e. GalAK-YFP and GalAK-

6×His) and in mutant (i.e. GalAK-6×His). The transgenic plant lines were validated for the expression of the recombinant GalAK by RT-PCR (Fig. 5.2A) and for protein expression (e.g.,

GalAK fused to YFP and to 6×His) by Western blot analysis with an anti GalAK antibody (Fig.

5.2B). Under laboratory conditions, the transgenic plants carrying GalAK fusion protein showed no visible morphological phenotype compared to the WT Landsberg plants.

184

FIGURE 5.1. Analysis of galak transposon insertion mutant.

A. Illustration of the transposable element (TE) insertion in GalAK (At3g10700) loci.

The TE insertion site was identified through DNA sequencing of the genomic region

(Sundaresan et al., 1995). The relative location and orientation of GalAK and TE (bottom panel) are shown. TE was inserted into the first intron of GalAK gene. Bottom panel indicates the molecular markers spanning the TE insertion cassette: RR5, 5‟ repeat region (Ds mobile elements); NPT, neomycin phosphotransferase; uidA, beta-glucuronidase (GUS); RR3, 3‟ repeat region (Ds mobile elements).

185

B. Gene expression analyses of WT Arabidopsis Landsberg (Ler) and galak mutant [GalAK KO line: GT8007]. i) Total RNA (2 µg) from 3-week-old Arabidopsis Landsberg leaves was reverse transcribed with oligo-dT primer and PCR-amplified using GalAK primers #1 and #2 (lane 1 and 2, see primer set i [solid arrow ] in panel A, also see primer list in Appendix B, Table B1), and with

GalK primers #3 and #4 (lane 3 and 4, primers see primer list in Appendix B, Table B1) as an internal control. Note: Although GalAK gene expression can not be detected by PCR, one can not exclude the possibility that the „full length GalAK‟ is made in the mutant due to the splicing of the transposable element. ii) Total RNA (2 µg) from 3-week-old Arabidopsis Landsberg leaves was reverse transcribed with oligo-dT primer and PCR-amplified using GalAK primers #21 and #22 (lane 1 and 2, see primer set ii [dashed arrow ] in panel A, also see primer list in Appendix B, Table B1), PCR reaction was performed for 30 cycles and analyzed on the agarose gel stained with ethidium bromide. Note: the GalAK gene is interrupted but a transcript spanning from nt 25 to the end can be made. However, it lacks a functional 5‟ region of the GalAK.

186

187

FIGURE 5.2. Semi-quantitative RT-PCR (A) and Western blot (B) of GalAK transgenic lines: overexpression and complementation of galak.

A. Total RNA (2 µg) was extracted from 4-week-old Arabidopsis leaves and stems and reverse transcribed and amplified by PCR using GalAK primers #1 and #2 (lane a, See Appendix B for list of primers), primers #1 and #28 (lane b), and with GalK Primers #3 and #4 (lane c) as an internal control. PCR reaction was performed for 30 cycles and analyzed on the agarose gel stained with ethidium bromide.

B. Proteins were extracted from 4-week-old Arabidopsis leaves and stems. Equal amount of proteins (80 µg) were separated on 10% SDS-PAGE blotted onto PVDP membrane and analyzed by Western blot using anti-GalAK serum. i) GalAK-6His overexpression lines in WT background: 35Sp:GalAK-6His 10-2 and 15-3 (lane

1 and 2); WT Ler (lane 3); GalAK-EYFP overexpression lines 35Sp:GalAK-EYFP #1, #2, #3

(lane 4-6). ii) GalAK-6His complementation lines in galak background: 35Sp:GalAK-6His C1-4, C2-3, C2-

4 and C3-4 (lane 1-4); WT Ler (lane 5); galak (lane 6).

188

Sugar composition analysis of cell wall polysaccharides in galak and WT - Wall polysaccharides were isolated from different plant tissues. TMS analysis was used as it provides means to detect neutral and acidic monosaccharides, such as GlcA, GalA, DHA, and KDO in a plant wall sample. The wall sugar composition of galak and WT is shown in Fig. 5.3, and the contents of individual monosaccharides are listed in Table 5.1. No statistical changes in sugar compositions between galak mutant and WT were observed.

189

40 35 WT Root 30 galak Root 25 20

15 µg/mg AIR µg/mg 10 5 0 Ara Rha Fuc Xyl GalA Man Gal Glc

40 WT Stem 35 30 galak Stem 25 20 15 µg/mg AIR µg/mg 10 5 0 Ara Rha Fuc Xyl GalA Man Gal Glc

35 WT Leaf 30 galak Leaf 25 20

15 µg/mg AIR µg/mg 10 5 0 Ara Rha Fuc Xyl GalA Man Gal Glc

190

16 WT Flower 14 galak Flower 12 10 8 6 µg/mg AIR µg/mg 4 2 0 Ara Rha Fuc Xyl GalA Man Gal Glc

FIGURE 5.3. Monosaccharide composition of non-cellulosic wall polysaccharides in galak mutant and WT.

TMS sugar composition analysis from 6-week-old Arabidopsis Landsberg roots (A), stems (B), leaves (C) and flowers (D) of galak and WT plants. The amounts of sugars released from wall polysaccharides are presented as mean values of micrograms of sugar/milligrams of dry weight

(n = 4) ± the standard deviation.

191

TABLE 5.1. Monosaccharide composition of non-cellulosic wall polysaccharides of galak and WT.

Root Stem Leaf Flower

WT galak WT galak WT galak WT galak

Ara 14.54±5.27 13.44±0.67 4.90±0.87 5.17±0.06 3.02±0.09 3.15±0.1 12.67±1.84 12.29±0.53 Rha 7.47±0.24 8.31±1.13 4.62±1.02 4.93±0.98 3.38±0.98 3.05±0.57 6.14±0.71 5.96±0.35 Fuc 2.26±1.31 1.70±0.42 0.78±0.08 0.71±0.09 0.72±0.01 0.61±0.24 1.04±0.24 0.97±0.21 Xyl 31.35±5.61 26.42±1.18 27.96±1.95 28.94±5.46 4.46±1.36 3.81±0.23 7.10±1.57 6.74±1.1 GalA 17.47±2.09 15.02±1.57 11.58±0.89 12.11±0.56 7.24±0.49 6.41±0.58 10.89±0.66 10.87±0.23 Man 2.17±0.54 1.91±1.05 2.97±0.98 3.18±0.25 1.08±0.02 1.23±0.72 3.05±0.43 2.93±0.21 Gal 13.53±2.97 8.89±6.66 7.35±0.49 7.21±0.96 4.17±0.67 4.12±0.33 13.01±0.79 11.63±1.23 Glc 3.73±1.04 3.66±2.08 4.93±0.21 3.47±1.59 19.52±12.97 6.54±1.74 8.62±3.77 7.27±0.41

TMS sugar composition analysis from 6-week-old roots, stems, leaves and flowers of galak and

WT Arabidopsis Landsberg plants. The amounts of sugars are presented as mean values of micrograms of sugar/milligrams of dry weight (n = 4) ± the standard deviation. Note: the glucose levels may vary due to the possible contaminations from the media or from the residual starch.

Therefore, the statistical analysis of this sugar residue may be inaccurate.

192

galak mutant seedling is unable to metabolize GalA - The experiments in Chapter 2 has demonstrated that recombinant GalAK, in vitro, is able to convert GalA to GalA-1-P, which is subsequently converted to UDP-GalA by sloppy (Yang et al., 2009). To determine if in vivo the native GalAK gene encodes similar function and to better understand GalA metabolism, we designed a GalA feeding experiment. In this experiment, galak and WT seeds were germinated on MS-agar plates containing 0.01% GalA or 0.25% Suc (control). After 10-day growth, the seedlings were harvested and washed twice with DDW. The sugar composition of the 80%- alcohol soluble fractions and the alcohol insoluble wall residues (AIR) were analyzed (Fig. 5.4) and the contents are listed in Table 5.2. The monosaccharide composition of wall glycans does not have an obvious statistical difference between galak and WT. In the 80%-alcohol soluble fraction, no differences could be detected for the seedlings grown in the media containing 0.25%

Suc. In contrast, galak mutant plants grown in the media with 0.01% GalA, appeared to accumulate large amount of GalA (over 40 fold) when compared to WT (see the arrow in Fig.

5.4), indicating galak mutant seedlings are unable to metabolize free GalA.

193

30 WT Suc 25 galak Suc WT GalA 20 galak GalA

15 µg/mg AIR µg/mg 10

0 Ara Rha Fuc Xyl GalA Man Gal Glc

WT Suc 25 galak Suc WT GalA 20 galak GalA

15 µg/mg AIR µg/mg 10

0 Ara Rha Fuc Xyl GalA Man Gal Glc

194

FIGURE 5.4. Monosaccharide composition of wall glycans and of 80%-alcohol soluble fractions from galak and WT.

WT Arabidopsis and galak seedling at 10 days after germination in MS-agar media containing either 0.25% Suc or 0.01% GalA were harvested. Alcohol soluble fractions were collected, dried, and the sugar composition was analyzed by TMS analysis (A). Alcohol insoluble residues (AIR, mainly wall polysaccharide) were also analyzed by TMS (B). The amounts of sugars are presented as mean values of micrograms of sugar/milligrams of dry weight (n = 4) ± the standard deviation. Arrow represents the difference of GalA content between WT and galak.

195

TABLE 5.2-A. Monosaccharide composition of alcohol soluble fractions from galak and

WT 10-day seedlings in Suc or GalA media.

Suc GalA

WT galak WT galak

Ara 2.47±1.69 5.98±5.11 4.27±0.08 4.55±0.15

Rha 0.89±0.2 0.73±0.17 1.58±0.19 1.25±0.31

Fuc 0.14±0.14 0.10±0.1 0.25±0.06 0.27±0.07

Xyl 0.27±0.06 0.20±0.11 0.28±0.04 0.65±0.06

GalA 0.06±0.06 0.06±0.06 0.07±0.07 2.86±1.06

Man 0.57±0.09 0.60±0.04 0.82±0.32 1.07±0.16

Gal 6.28±0.07 5.83±0.63 6.98±1.50 8.30±0.16

Glc 14.23±2.82 17.50±5.96 14.15±0.33 19.88±3.06

196

TABLE 5.2-B. Monosaccharide composition of alcohol insoluble residues from galak and

WT 10-day seedlings in Suc or GalA media.

Suc GalA

WT galak WT galak

Ara 9.25±0.87 9.69±0.16 7.78±0.64 8.69±0.99

Rha 5.74±1.11 6.25±1.01 6.59±0.08 6.69±0.98

Fuc 1.99±0.33 1.85±0.25 1.36±0.1 1.49±0.26

Xyl 7.07±0.68 7.51±0.01 5.87±0.09 6.39±0.78

GalA 7.16±0.14 7.78±0.02 9.10±0.09 8.69±0.06

Man 1.23±0.88 1.55±1.04 2.12±0.49 2.27±0.31

Gal 16.55±0.58 20.53±6.32 21.61±0.12 24.97±3.41

Glc 8.13±1.09 11.13±1.03 5.22±0.06 5.13±1.10

TMS sugar composition analysis from galak and wild-type seedlings at 10 days after germination in MS-agar media containing either 0.25% Suc or 0.01% GalA. A, alcohol soluble fraction, the amounts of sugars are presented as mean values of micrograms of sugar/milligrams of dry weight (n = 4) ± the standard error; B, alcohol insoluble residue. The amounts of sugars are presented as mean values of micrograms of sugar/milligrams of dry weight (n = 4) ± the standard deviation.

197

Immunohistochemical analyses of WT and galak plant tissues with antibodies generated to different glycan epitopes - Total monosaccharide analysis by TMS was not sensitive enough to evaluate if there is any structural defects in glycans isolated from galak mutant plants compared with WT. We therefore decided to monitor any alternation in glycan structures or quantities by immunohistochemical analysis. Monoclonal antibodies to selective glycan epitopes were proven a powerful tool to distinguish wall mutant from WT plants (Pattathil et al., 2010). Different antibodies were used [kindly supplied by Dr. Michael Hahn, (Pattathil et al., 2010)] and the immunohistochemical studies were carried out on various tissues of 6-week-old Arabidopsis

Landsberg WT and galak mutant. The tissues analyzed were bottom and inflorescence stems, flowers (pollen), and 10-day-old seedlings (Fig. 5.5). The different monoclonal antibodies used in this study were generated against pectin polymers (e.g., HG and RG-I), xylan, or xyloglucan

(see figure legend for detail). No obvious differences were observed in the pattern of antibody staining between WT and mutant.

198

199

FIGURE 5.5. Immunohistochemistry of different polysaccharides in galak mutant and WT.

Specific monoclonal antibodies against distinct wall glycan epitopes were used to stain thin sections of various plant tissues: 10-day-old seedling hypocotyl (A), 6-week-old Arabidopsis

Landsberg lower stem (B), upper stem (C), anther and pollen (D), and flower (E) sections of WT

(upper) and galak mutant (lower). Antibodies are indicated in the upper left corner of the figure, and the recognizing epitope modules are summarized in Appendix B.

A. Immunofluorescent labeling of Arabidopsis hypocotyl with monoclonal antibodies. Sections were taken from the hypocotyl of 10-day-old Arabidopsis Landsberg seedlings. Immunolabeling was carried out using two non-fucosylated xyloglucan antibodies (CCRC-M88, CCRC-M100),

200

two pectin binding antibodies (CCRC-M35, CCRC-M38), and one fucosylated xyloglucan binding antibody (CCRC-M1). Bars = 50 µm.

B. Immunofluorescent labeling of Arabidopsis stems with monoclonal antibodies. Sections were taken from the base of inflorescence stems of 6-week-old Arabidopsis Landsberg plants.

Immunolabeling was carried out using one xylan binding antibody (CCRC-M138), three pectin binding antibodies (CCRC-M34, CCRC-M35, CCRC-M38), and one fucosylated xyloglucan binding antibody (CCRC-M1). Bars = 50 µm.

C. Immunofluorescent labeling of Arabidopsis stems with monoclonal antibodies. Sections were taken from the upper inflorescence stems of 6-week-old Arabidopsis Landsberg plants.

Immunolabeling was carried out using two xylan binding antibodies (CCRC-M137, CCRC-

M149), two pectin binding antibodies (CCRC-M35, CCRC-M38), and one non-fucosylated xyloglucan binding antibody (CCRC-M93). Bars = 50 µm.

D. Immunofluorescent labeling of Arabidopsis flowers with monoclonal antibodies. Sections were taken from the flowers (anther and pollen portion) of 6-week-old Arabidopsis Landsberg plants. Immunolabeling was carried out using one callose antibody (LAMP); one xylan binding antibody (CCRC-M138), and three pectin binding antibodies (CCRC-M14, CCRC-M35, CCRC-

M38). Bars = 50 µm.

E. Immunofluorescent labeling of Arabidopsis flowers with monoclonal antibodies. Sections were taken from the flower (stigma and style) of 6-week-old Arabidopsis Landsberg plants.

Immunolabeling was carried out using pectin binding antibody (CCRC-M38). Bars = 100 µm.

201

DISCUSSION

The work presented in this chapter has shown that 1) in vivo GalAK has the ability to shunt free galacturonic acid to other metabolic pathway(s); 2) the salvage pathway exists in plants; 3) GalAK function in vivo is the same as in vitro; 4) no alternative plant gene product can compensate for GalAK.

Feeding Arabidopsis seedlings with galacturonic acid caused a large amount (>40 fold) of free GalA accumulated in galak mutants when compared with WT control. Such accumulation suggests that in vivo the expression of GalAK is essential for the metabolism of galacturonic acid. Interestingly however, no visible growth defects were detected in galak mutant seedlings and the mutant plant developed normally and was able to set flowers and viable seeds just like the wild type plant. It was shown that over-dosing Arabidopsis with galactose is toxic for seedlings (Dormann and Benning, 1998). Whether GalA is toxic to all plants is unclear, nonetheless based on our experimental setup GalA is not toxic to the Arabidopsis seedlings.

Knocking out the GalAK gene in Arabidopsis had no phenotypic effect. A possible explanation may be the existence of duplicate genes, where the effects of knocking out a gene are compensated due to a duplicate copy. This explanation doesn‟t conincide with the data showing accumulation of GalA in the sugar feeding experiments of the mutant (see arrow, Fig.

5.4). The second explanation for no phenotype in galak mutant is the existence of alternative pathways. The lack of UDP-GalA in galak can be overcome from the activity of UDP-GlcA 4- epimerase in Golgi (see Fig. 2.6, Chapter 2). However, this suggests that UDP-GalA made in

Golgi can exit into the cytosol. The data on such transporters is currently lacking. Another possibility for no phenotype is that the product of this metabolic pathway is only required when challenged by specific biotic or abiotic factors. However, no phenotype of galak mutant was

202

detected when it was grown under various conditions including: humid growth chamber, under drought, and when challenged by pathogenic fungus (Botryotinia fuckeliana) compared to WT plants (data not shown).

UDP-GalA generated by the salvage pathway in the cytosol may enter the Golgi for the synthesis of wall polysaccharide. Previous studies have suggested that such a UDP-GalA transporter exists since intact microsomes are able to take up exogenous UDP-14C-GalA (Sterling et al., 2001). If this is the function of GalAK, then we would anticipate cell wall polysaccharide synthesis to be affected in galak mutant. However, the monosaccharide composition of cell wall polysaccharides in various tissues were not affected in the galak mutant, nor were the glycan structures, since the biosynthesis of UDP-GalA could be achieved by other pathways e.g., interconversion pathway. It has been shown that in Arabidopsis, there are 6 isoforms of UGlcAE in

Golgi (Gu, 2009). It is also proposed that the synthesis of the major wall polysaccharides happens in the Golgi compartments. Therefore, the inter-conversion pathway may substitute the salvage pathway in the supply of UDP-GalA for the GalA-containing polysaccharide synthesis

(see Fig. 2.6, Chapter 2).

Another possibility that cannot be excluded is the UDP-GalA, produced by GalAK, may be transported to other organelles for the synthesis of glycolipids or small glycosidic- metabolites, e.g., flavonoids, that may not be essential when the mutant is growing in a laboratory setup. In plants, some glycolipids in chlorophyll and mitochondria were reported to contain galactose, glucosamine and hexuronic sugar residues (Carter et al., 1969; Kaul and

Lester, 1975). The nature of the „hexuronic‟ sugar(s) was not studied in Arabidopsis and will require further work. Some secondary metabolites are glycosylated. The glycosyl residues in some plants were found to contain Glc, Rha, GalA or GlcA residues (Sashida et al., 1983; Bar-

203

Peled et al., 1991; Morrissey and Osbourn, 1999; Kumar et al., 2005; Noguchi et al., 2009).

Further studies will be carried out for the analysis of glycolipids and small glycosides metabolism in galak mutant and WT.

204

CHAPTER 6

CONCLUSIONS

Nucleotide sugar biosynthesis is the key metabolic process for the formation of polysaccharides, glycolipids, glycopropteins, many glycosylated secondary metabolites and hormones. There are several metabolic routes to form NDP-sugars and the least understood or characterized was the salvage pathway. To understand how salvage pathway contributes to the flux of NDP-sugars, my PhD research was first focused on the identification of the enzymes and corresponding genes that convert free sugars to NDP-sugars in plants. This research includes the characterization of several new sugar kinases and NDP-sugar pyrophosphorylases. The second focus of my research was to determine by genetic approaches, how critical the salvage pathway is in plants. Lastly, I have tried to take bioinformatic and microbial approaches to globally understand how salvage pathway is evolved and controlled in different organisms. I have identified several sugar kinases and nucleotide sugar pyrophosphorylases in various species, including Arabidopsis, Populus, and Protozoan (Yang Ph. D. Dissertation Chapter 1, 2, 3, 4). I functionally characterized genes encoding galacturonic acid kinase (GalA kinase), galactokinase

(GalK), UDP-sugar PPase (Sloppy) (Yang Ph. D. Dissertation Chapter 2), UDP-sugar pyrophosphorylase paralogs from Trypanosoma cruzi and Leishmania major (Yang Ph. D.

Dissertation Chapter 3), as well as two promiscuous UDP-GlcNAc pyrophosphorylases

(GlcNAc-1-P uridylyltransferases) (Yang Ph. D. Dissertation Chapter 4). Phylogenetic analysis

205

of sugar kinase and NDP-sugar PPase (Yang Ph. D. Dissertation Chapter 1) proteins indicates that kinase enzymes, due to their substrate specificities, can be divided into different clades, and the same is correct for NDP-sugar PPases. Amino acid sequence alignment of sugar kinases reveals conserved motifs in ATP binding and catalytic activity (GHMP_kinase_N and C domain), and the potential amino acid residues for sugar binding. Likewise, amino acid sequence alignment of PPases indicates that UDP-Glc PPase, UDP-GlcNAc PPase and sloppy share a

UDPGP domain, but ADP-Glc PPase and GDP-Man PPase share an NTP-transferase domain.

Enzymatic characterization of these PPases revealed that some are promiscuous and capable of using more that one sugar-1-P as substrates: e.g., the two UDP-GlcNAc PPases (Yang Ph. D.

Dissertation Chapter 4) and Sloppy (Yang Ph. D. Dissertation Chapter 2 and 3), whereas other

PPases have strict substrate specificities, e.g., UDP-Glc PPase and GDP-Man PPase.

To further investigate the role of salvage pathway and determine if and how it involves the synthesis of plant glycans, I analyzed a mutant plant lacking the activity of galacturonic acid kinase (galak) (Yang Ph. D. Dissertation Chapter 5). The data shown clearly provided strong evidence that without functional GalAK gene in Arabidopsis plants, seedlings can accumulate large amount of galacturonic acid. This accumulation, however, does not appear to adversely affect plant growth and development when they are grown under laboratory setup. Analysis of wall glycans (sugar composition and immunohistochemistry) from galak mutant showed no major differences when compared with WT plants. Nonetheless, it is possible that GalA in seed or seedling is required for the synthesis of non-wall metabolites. Some glycolipids and secondary metabolites harbor uronic acids, and subsequent analysis would focus on the metabolism of glycolipids and small glycosides in galak mutant and WT.

206

REFERENCES

Alonso AP, Piasecki RJ, Wang Y, LaClair RW, Shachar-Hill Y (2010) Quantifying the

labeling and the levels of plant cell wall precursors using ion chromatography tandem

mass spectrometry. Plant Physiol 153: 915-924

Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. (2000) Nature

408: 796-815

Angyal SJ, Greeves D, Littlemore L (1988) The structure of the complex formed from D-

galacturonate ions and cations in solution Carbohydrate research 174: 121-131

Arnold K, Bordoli L, Kopp J, Schwede T (2006) The SWISS-MODEL workspace: a web-

based environment for protein structure homology modelling. Bioinformatics 22: 195-

201

Axelos M, Peaud-Lenoel C (1978) Glycosyl transfers from UDP-sugars to lipids of plant

membranes: identification and specificity of the transferases. Biochimie 60: 35-44

Badejo AA, Jeong ST, Goto-Yamamoto N, Esaka M (2007) Cloning and expression of GDP-

D-mannose pyrophosphorylase gene and ascorbic acid content of acerola (Malpighia

glabra L.) fruit at ripening stages. Plant Physiol Biochem 45: 665-672

Bajwa W, Torchia TE, Hopper JE (1988) Yeast regulatory gene GAL3: carbon regulation;

UASGal elements in common with GAL1, GAL2, GAL7, GAL10, GAL80, and MEL1;

encoded protein strikingly similar to yeast and Escherichia coli galactokinases. Mol Cell

Biol 8: 3439-3447

207

Ballicora MA, Frueauf JB, Fu Y, Schurmann P, Preiss J (2000) Activation of the potato tuber

ADP-glucose pyrophosphorylase by thioredoxin. J Biol Chem 275: 1315-1320

Baluska F, Hlavacka A, Samaj J, Palme K, Robinson DG, Matoh T, McCurdy DW, Menzel

D, Volkmann D (2002) F-actin-dependent endocytosis of cell wall pectins in

meristematic root cells. Insights from brefeldin A-induced compartments. Plant Physiol

130: 422-431

Bar-Peled M, Lewinsohn E, Fluhr R, Gressel J (1991) UDP-rhamnose:flavanone-7-O-

glucoside-2''-O-rhamnosyltransferase. Purification and characterization of an enzyme

catalyzing the production of bitter compounds in citrus. J Biol Chem 266: 20953-20959

Bar-Peled M, O'Neill MA (2010) Plant Nucleotide Sugar Formation, Interconversion, and

Salvage by Sugar Recycling. Annu Rev Plant Biol

Barakate A, Martin W, Quigley F, Mache R (1993) Characterization of a multigene family

encoding an exopolygalacturonase in maize. J Mol Biol 229: 797-801

Barboza M, Duschak VG, Fukuyama Y, Nonami H, Erra-Balsells R, Cazzulo JJ, Couto AS

(2005) Structural analysis of the N-glycans of the major cysteine proteinase of

Trypanosoma cruzi. Identification of sulfated high-mannose type oligosaccharides. FEBS

J 272: 3803-3815

Bechtold N, Pelletier G (1998) In planta Agrobacterium-mediated transformation of adult

Arabidopsis thaliana plants by vacuum infiltration. Methods Mol Biol 82: 259-266

Beckles DM, Smith AM, ap Rees T (2001) A cytosolic ADP-glucose pyrophosphorylase is a

feature of graminaceous endosperms, but not of other starch-storing organs. Plant Physiol

125: 818-827

208

Benning C, Ohta H (2005) Three enzyme systems for galactoglycerolipid biosynthesis are

coordinately regulated in plants. J Biol Chem 280: 2397-2400

Bessueille L, Bulone V (2008) A survey of cellulose biosynthesis in higher plants. Plant

Biotechnology 25: 315-322

Beverley SM, Owens KL, Showalter M, Griffith CL, Doering TL, Jones VC, McNeil MR

(2005) Eukaryotic UDP-galactopyranose mutase (GLF gene) in microbial and metazoal

pathogens. Eukaryot Cell 4: 1147-1154

Brosch-Salomon S, Hoftberger M, Holzinger A, Lutz-Meindl U (1998) Ultrastructural

localization of polysaccharides and N-actetyl-D-galactosamine in the secretory pathway

of green algae (Desmidiaceae). Journal of Experimental Botany 49: 145-153

Brown K, Pompeo F, Dixon S, Mengin-Lecreulx D, Cambillau C, Bourne Y (1999) Crystal

structure of the bifunctional N-acetylglucosamine 1-phosphate uridyltransferase from

Escherichia coli: a paradigm for the related pyrophosphorylase superfamily. EMBO J 18:

4096-4107

Buttner M (2007) The monosaccharide transporter(-like) gene family in Arabidopsis. FEBS Lett

581: 2318-2324

Caffall KH, Mohnen D (2009) The structure, function, and biosynthesis of plant cell wall pectic

polysaccharides. Carbohydr Res

Carlson DM, Hansen RG (1962) The isolation and synthesis of guanosine diphosphate glucose.

J Biol Chem 237: 1260-1265

Carter HE, Strobach DR, Hawthorne JN (1969) Biochemistry of the sphingolipids. 18.

Complete structure of tetrasaccharide phytoglycolipid. Biochemistry 8: 383-388

209

Casasoli M, Spadoni S, Lilley KS, Cervone F, De Lorenzo G, Mattei B (2008) Identification

by 2-D DIGE of apoplastic proteins regulated by oligogalacturonides in Arabidopsis

thaliana. Proteomics 8: 1042-1054

Castresana J (2000) Selection of conserved blocks from multiple alignments for their use in

phylogenetic analysis. Mol Biol Evol 17: 540-552

Chen J, Brevet A, Fromant M, Leveque F, Schmitter JM, Blanquet S, Plateau P (1990)

Pyrophosphatase is essential for growth of Escherichia coli. J Bacteriol 172: 5686-5689

Chen R, Zhao X, Shao Z, Wei Z, Wang Y, Zhu L, Zhao J, Sun M, He R, He G (2007) Rice

UDP-glucose pyrophosphorylase1 is essential for pollen callose deposition and its

cosuppression results in a new type of thermosensitive genic male sterility. Plant Cell 19:

847-861

Conklin PL, Saracco SA, Norris SR, Last RL (2000) Identification of ascorbic acid-deficient

Arabidopsis thaliana mutants. Genetics 154: 847-856

Coyne MJ, Reinap B, Lee MM, Comstock LE (2005) Human symbionts use a host-like

pathway for surface fucosylation. Science 307: 1778-1781

Dai N, Petreikov M, Portnoy V, Katzir N, Pharr DM, Schaffer AA (2006) Cloning and

expression analysis of a UDP-galactose/glucose pyrophosphorylase from melon fruit

provides evidence for the major metabolic pathway of galactose metabolism in raffinose

oligosaccharide metabolizing plants. Plant Physiol 142: 294-304

Davies GJ, Gloster TM, Henrissat B (2005) Recent structural insights into the expanding

world of carbohydrate-active enzymes. Curr Opin Struct Biol 15: 637-645

210

Dereeper A, Guignon V, Blanc G, Audic S, Buffet S, Chevenet F, Dufayard JF, Guindon S,

Lefort V, Lescot M, Claverie JM, Gascuel O (2008) Phylogeny.fr: robust phylogenetic

analysis for the non-specialist. Nucleic Acids Res 36: W465-469

Dolezal O, Cobbett CS (1991) Arabinose Kinase-Deficient Mutant of Arabidopsis thaliana.

Plant Physiol 96: 1255-1260

Dormann P, Benning C (1998) The role of UDP-glucose epimerase in carbohydrate metabolism

of Arabidopsis. Plant J 13: 641-652

Dormann P, Benning C (2002) Galactolipids rule in seed plants. Trends Plant Sci 7: 112-118

Dubald M, Barakate A, Mandaron P, Mache R (1993) The ubiquitous presence of

exopolygalacturonase in maize suggests a fundamental cellular function for this enzyme.

Plant J 4: 781-791

Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high

throughput. Nucleic Acids Res 32: 1792-1797

Egelund J, Damager I, Faber K, Olsen CE, Ulvskov P, Petersen BL (2008) Functional

characterisation of a putative rhamnogalacturonan II specific xylosyltransferase. FEBS

Lett 582: 3217-3222

Falasca G, Capitani F, Della Rovere F, Zaghi D, Franchin C, Biondi S, Altamura MM

(2008) Oligogalacturonides enhance cytokinin-induced vegetative shoot formation in

tobacco explants, inhibit polyamine biosynthetic gene expression, and promote long-term

remobilisation of cell calcium. Planta 227: 835-852

Feingold DS, Avigad G (1980) Sugar nucleotide transformations in plants. In The Biochemistry

of Plants, Vol 3. Academic Press, New York pp 101-170

211

Ferguson MA (1999) The structure, biosynthesis and functions of glycosylphosphatidylinositol

anchors, and the contributions of trypanosome research. J Cell Sci 112 ( Pt 17): 2799-

2809

Frydman A, Weisshaus O, Huhman DV, Sumner LW, Bar-Peled M, Lewinsohn E, Fluhr

R, Gressel J, Eyal Y (2005) Metabolic engineering of plant cells for biotransformation

of hesperedin into neohesperidin, a substrate for production of the low-calorie sweetener

and flavor enhancer NHDC. J Agric Food Chem 53: 9708-9712

Fukuda M, Watanabe S, Yoshida S, Itoh H, Itoh Y, Kamio Y, Kaneko J (2010) Cell surface

xylanases of the glycoside hydrolase family 10 are essential for xylan utilization by

Paenibacillus sp. W-61 as generators of xylo-oligosaccharide inducers for the xylanase

genes. J Bacteriol 192: 2210-2219

Ghalambor MA, Heath EC (1966) The biosynthesis of cell wall lipopolysaccharide in

Escherichia coli. IV. Purification and properties of cytidine monophosphate 3-deoxy-d-

manno-octulosonate synthetase. J Biol Chem 241: 3216-3221

Goff SA, Ricke D, Lan TH, Presting G, Wang R, Dunn M, Glazebrook J, Sessions A, Oeller

P, Varma H, Hadley D, Hutchison D, Martin C, Katagiri F, Lange BM, Moughamer

T, Xia Y, Budworth P, Zhong J, Miguel T, Paszkowski U, Zhang S, Colbert M, Sun

WL, Chen L, Cooper B, Park S, Wood TC, Mao L, Quail P, Wing R, Dean R, Yu Y,

Zharkikh A, Shen R, Sahasrabudhe S, Thomas A, Cannings R, Gutin A, Pruss D,

Reid J, Tavtigian S, Mitchell J, Eldredge G, Scholl T, Miller RM, Bhatnagar S,

Adey N, Rubano T, Tusneem N, Robinson R, Feldhaus J, Macalma T, Oliphant A,

Briggs S (2002) A draft sequence of the rice genome (Oryza sativa L. ssp. japonica).

Science 296: 92-100

212

Gonzalez NS, Pontis HG (1963) Uridine diphosphate fructose and uridine diphosphate

acetylgalactosamine from dahlia tubers. Biochim Biophys Acta 69: 179-181

Gu X (2009) CHARACTERIZATION OF PLANT UDP-GLUCURONIC ACID 4-

EPIMERASES. The University of Georgia, Athens

Gu X, Bar-Peled M (2004) The biosynthesis of UDP-galacturonic acid in plants. Functional

cloning and characterization of Arabidopsis UDP-D-glucuronic acid 4-epimerase. Plant

Physiol 136: 4256-4264

Guha-Niyogi A, Sullivan DR, Turco SJ (2001) Glycoconjugate structures of parasitic protozoa.

Glycobiology 11: 45R-59R

Gutternigg M, Kretschmer-Lubich D, Paschinger K, Rendic D, Hader J, Geier P, Ranftl R,

Jantsch V, Lochnit G, Wilson IB (2007) Biosynthesis of truncated N-linked

oligosaccharides results from non-orthologous hexosaminidase-mediated mechanisms in

nematodes, plants, and insects. J Biol Chem 282: 27825-27840

Guyett P, Glushka J, Gu X, Bar-Peled M (2009) Real-time NMR monitoring of intermediates

and labile products of the bifunctional enzyme UDP-apiose/UDP-xylose synthase.

Carbohydr Res 344: 1072-1078

Hadfield KA, Bennett AB (1998) Polygalacturonases: many genes in search of a function. Plant

Physiol 117: 337-343

Hall A, Rog T, Karttunen M, Vattulainen I (2010) Role of glycolipids in lipid rafts: a view

through atomistic molecular dynamics simulations with galactosylceramide. J Phys Chem

B 114: 7797-7807

Hancock JF (2004) GPI-anchor synthesis: Ras takes charge. Dev Cell 6: 743-745

213

Handford M, Rodriguez-Furlan C, Orellana A (2006) Nucleotide-sugar transporters:

structure, function and roles in vivo. Braz J Med Biol Res 39: 1149-1158

Harborne JB, Williams CA (2000) Advances in flavonoid research since 1992. Phytochemistry

55: 481-504

Hart GW, Haltiwanger RS, Holt GD, Kelly WG (1989) Glycosylation in the nucleus and

cytoplasm. Annu Rev Biochem 58: 841-874

Hartley A, Glynn SE, Barynin V, Baker PJ, Sedelnikova SE, Verhees C, de Geus D, van

der Oost J, Timson DJ, Reece RJ, Rice DW (2004) Substrate specificity and

mechanism from the structure of Pyrococcus furiosus galactokinase. J Mol Biol 337:

387-398

Hartweck LM, Scott CL, Olszewski NE (2002) Two O-linked N-acetylglucosamine transferase

genes of Arabidopsis thaliana L. Heynh. have overlapping functions necessary for gamete

and seed development. Genetics 161: 1279-1291

Harwood JL (1997) Plant lipid metabolism. In PM Dey, JM Harborne, eds, Plant Biochemistry.

Academic Press, San Diego, pp 237-271

Harwood JL (1998) Plant lipid biosynthesis : fundamentals and agricultural applications.

Cambridge University Press, New York

Haynes PA, Ferguson MA, Cross GA (1996) Structural characterization of novel

oligosaccharides of cell-surface glycoproteins of Trypanosoma cruzi. Glycobiology 6:

869-878

Haynes PA, Russell DG, Cross GA (1996) Subcellular localization of Trypanosoma cruzi

glycoprotein Gp72. J Cell Sci 109 ( Pt 13): 2979-2988

214

Heese-Peck A, Raikhel NV (1998) A glycoprotein modified with terminal N-acetylglucosamine

and localized at the nuclear rim shows sequence similarity to aldose-1-epimerases. Plant

Cell 10: 599-612

Hoffmeister D, Thorson JS (2004) Mechanistic implications of Escherichia coli galactokinase

structure-based engineering. Chembiochem 5: 989-992

Holmes-Davis R, Tanaka CK, Vensel WH, Hurkman WJ, McCormick S (2005) Proteome

mapping of mature pollen of Arabidopsis thaliana. Proteomics 5: 4864-4884

Hu H, Brown PH (1994) Localization of Boron in Cell Walls of Squash and Tobacco and Its

Association with Pectin (Evidence for a Structural Role of Boron in the Cell Wall). Plant

Physiol 105: 681-689

Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetic trees.

Bioinformatics 17: 754-755

Ishii T, Matsunaga T, Hayashi N (2001) Formation of rhamnogalacturonan II-borate dimer in

pectin determines cell wall thickness of pumpkin tissue. Plant Physiol 126: 1698-1705

Iwai H, Hokura A, Oishi M, Chida H, Ishii T, Sakai S, Satoh S (2006) The gene responsible

for borate cross-linking of pectin Rhamnogalacturonan-II is required for plant

reproductive tissue development and fertilization. Proc Natl Acad Sci U S A 103: 16592-

16597

Iwai H, Masaoka N, Ishii T, Satoh S (2002) A pectin glucuronyltransferase gene is essential for

intercellular attachment in the plant meristem. Proc Natl Acad Sci U S A 99: 16319-

16324

215

Jacobsen SE, Binkowski KA, Olszewski NE (1996) SPINDLY, a tetratricopeptide repeat

protein involved in gibberellin signal transduction in Arabidopsis. Proc Natl Acad Sci U

S A 93: 9292-9296

Jin X, Ballicora MA, Preiss J, Geiger JH (2005) Crystal structure of potato tuber ADP-glucose

pyrophosphorylase. EMBO J 24: 694-704

Kanter U, Usadel B, Guerineau F, Li Y, Pauly M, Tenhaken R (2005) The inositol oxygenase

gene family of Arabidopsis is involved in the biosynthesis of nucleotide sugar precursors

for cell-wall matrix polysaccharides. Planta 221: 243-254

Kaplan CP, Tugal HB, Baker A (1997) Isolation of a cDNA encoding an Arabidopsis

galactokinase by functional expression in yeast. Plant Mol Biol 34: 497-506

Katoh K, Kuma K, Toh H, Miyata T (2005) MAFFT version 5: improvement in accuracy of

multiple sequence alignment. Nucleic Acids Res 33: 511-518

Kaul K, Lester RL (1975) Characterization of Inositol-containing Phosphosphingolipids from

Tobacco Leaves: Isolation and Identification of Two Novel, Major Lipids: N-

Acetylglucosamidoglucuronidoinositol Phosphorylceramide and

Glucosamidoglucuronidoinositol Phosphorylceramide. Plant Physiol 55: 120-129

Kean EL, Roseman S (1966) The sialic acids. X. Purification and properties of cytidine 5'-

monophosphosialic acid synthetase. J Biol Chem 241: 5643-5650

Kelley LA, Sternberg MJ (2009) Protein structure prediction on the Web: a case study using

the Phyre server. Nat Protoc 4: 363-371

Kelly AA, Dormann P (2004) Green light for galactolipid trafficking. Curr Opin Plant Biol 7:

262-269

216

Kessler G, Neufeld EF, Feingold DS, Hassid WZ (1961) Metabolism of D-glucuronic acid and

D-galacturonic acid by Phaseolus aureus seedlings. J Biol Chem 236: 308-312

Kleczkowski LA (2000) Is leaf ADP-glucose pyrophosphorylase an allosteric enzyme? Biochim

Biophys Acta 1476: 103-108

Koch K (2004) Sucrose metabolism: regulatory mechanisms and pivotal roles in sugar sensing

and plant development. Curr Opin Plant Biol 7: 235-246

Kotake T, Hojo S, Tajima N, Matsuoka K, Koyama T, Tsumuraya Y (2008) A bifunctional

enzyme with L-fucokinase and GDP-L-fucose pyrophosphorylase activities salvages free

L-fucose in Arabidopsis. J Biol Chem 283: 8125-8135

Kotake T, Hojo S, Yamaguchi D, Aohara T, Konishi T, Tsumuraya Y (2007) Properties and

physiological functions of UDP-sugar pyrophosphorylase in Arabidopsis. Biosci

Biotechnol Biochem 71: 761-771

Kotake T, Yamaguchi D, Ohzono H, Hojo S, Kaneko S, Ishida HK, Tsumuraya Y (2004)

UDP-sugar pyrophosphorylase with broad substrate specificity toward various

monosaccharide 1-phosphates from pea sprouts. J Biol Chem 279: 45728-45736

Kumar A, Singh SP, Bhakuni RS (2005) Secondary metabolites of Chrysanthemum genus and

their biological activities. Current Science 89: 1489-1501

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of

bacteriophage T4. Nature 227: 680-685

Lamerz AC, Haselhorst T, Bergfeld AK, von Itzstein M, Gerardy-Schahn R (2006)

Molecular cloning of the Leishmania major UDP-glucose pyrophosphorylase, functional

characterization, and ligand binding analyses using NMR spectroscopy. J Biol Chem

281: 16314-16322

217

Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H,

Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG

(2007) Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947-2948

Litterer LA, Schnurr JA, Plaisance KL, Storey KK, Gronwald JW, Somers DA (2006)

Characterization and expression of Arabidopsis UDP-sugar pyrophosphorylase. Plant

Physiol Biochem 44: 171-180

Lodish H, Berk A, Zipursky SL, Matsudaira P, Baltimore D, Darnell J (2000) Protein

Glycosylation in the ER and Golgi Complex. In Molecular Cell Biology. W.H. Freeman,

New York

Loewus FA (1961) Aspects of ascorbic acid biosynthesis in plants. Ann N Y Acad Sci 92: 57-78

Lu Y, Steichen JM, Yao J, Sharkey TD (2006) The role of cytosolic alpha-glucan

phosphorylase in maltose metabolism and the comparison of amylomaltase in

Arabidopsis and Escherichia coli. Plant Physiol 142: 878-889

Lukowitz W, Nickle TC, Meinke DW, Last RL, Conklin PL, Somerville CR (2001)

Arabidopsis cyt1 mutants are deficient in a mannose-1-phosphate guanylyltransferase and

point to a requirement of N-linked glycosylation for cellulose biosynthesis. Proc Natl

Acad Sci U S A 98: 2262-2267

Luz JG, Hassig CA, Pickle C, Godzik A, Meyer BJ, Wilson IA (2003) XOL-1, primary

determinant of sexual fate in C. elegans, is a GHMP kinase family member and a

structural prototype for a class of developmental regulators. Genes Dev 17: 977-990

Macrae JI, Acosta-Serrano A, Morrice NA, Mehlert A, Ferguson MA (2005) Structural

characterization of NETNES, a novel glycoconjugate in Trypanosoma cruzi

epimastigotes. J Biol Chem 280: 12201-12211

218

MacRae JI, Obado SO, Turnock DC, Roper JR, Kierans M, Kelly JM, Ferguson MA

(2006) The suppression of galactose metabolism in Trypanosoma cruzi epimastigotes

causes changes in cell surface molecular architecture and cell morphology. Mol Biochem

Parasitol 147: 126-136

Marin-Rodriguez MC, Orchard J, Seymour GB (2002) Pectate lyases, cell wall degradation

and fruit softening. J Exp Bot 53: 2115-2119

Martienssen R (1998) Transposons, DNA methylation and gene control. Trends Genet 14: 263-

264

Matsuzaki M, Misumi O, Shin IT, Maruyama S, Takahara M, Miyagishima SY, Mori T,

Nishida K, Yagisawa F, Yoshida Y, Nishimura Y, Nakao S, Kobayashi T,

Momoyama Y, Higashiyama T, Minoda A, Sano M, Nomoto H, Oishi K, Hayashi H,

Ohta F, Nishizaka S, Haga S, Miura S, Morishita T, Kabeya Y, Terasawa K, Suzuki

Y, Ishii Y, Asakawa S, Takano H, Ohta N, Kuroiwa H, Tanaka K, Shimizu N,

Sugano S, Sato N, Nozaki H, Ogasawara N, Kohara Y, Kuroiwa T (2004) Genome

sequence of the ultrasmall unicellular red alga Cyanidioschyzon merolae 10D. Nature

428: 653-657

McCoy JG, Arabshahi A, Bitto E, Bingman CA, Ruzicka FJ, Frey PA, Phillips GN, Jr.

(2006) Structure and mechanism of an ADP-glucose phosphorylase from Arabidopsis

thaliana. Biochemistry 45: 3154-3162

McCoy JG, Bitto E, Bingman CA, Wesenberg GE, Bannen RM, Kondrashov DA, Phillips

GN, Jr. (2007) Structure and dynamics of UDP-glucose pyrophosphorylase from

Arabidopsis thaliana with bound UDP-glucose and UTP. J Mol Biol 366: 830-841

219

McGinnis K, Chandler V, Cone K, Kaeppler H, Kaeppler S, Kerschen A, Pikaard C,

Richards E, Sidorenko L, Smith T, Springer N, Wulan T (2005) Transgene-induced

RNA interference as a tool for plant functional genomics. Methods Enzymol 392: 1-24

Meli VS, Ghosh S, Prabha TN, Chakraborty N, Chakraborty S, Datta A (2010)

Enhancement of fruit shelf life by suppressing N-glycan processing enzymes. Proc Natl

Acad Sci U S A 107: 2413-2418

Meng M, Geisler M, Johansson H, Harholt J, Scheller HV, Mellerowicz EJ, Kleczkowski

LA (2009) UDP-glucose pyrophosphorylase is not rate limiting, but is essential in

Arabidopsis. Plant Cell Physiol 50: 998-1011

Meng M, Wilczynska M, Kleczkowski LA (2008) Molecular and kinetic characterization of

two UDP-glucose pyrophosphorylases, products of distinct genes, from Arabidopsis.

Biochim Biophys Acta 1784: 967-972

Milewski S, Gabriel I, Olchowy J (2006) Enzymes of UDP-GlcNAc biosynthesis in yeast.

Yeast 23: 1-14

Mio T, Yabe T, Arisawa M, Yamada-Okabe H (1998) The eukaryotic UDP-N-

acetylglucosamine pyrophosphorylases. Gene cloning, protein expression, and catalytic

mechanism. J Biol Chem 273: 14392-14397

Mohnen D (2002) Biosynthesis of Pectins. In Seymour GB, Knox JP, eds, Pectins and their

manipulation, Vol 84. Blackwell Oxford, pp 52-98

Mohnen D (2008) Pectin structure and biosynthesis. Curr Opin Plant Biol 11: 266-277

Mohnen D, Bar-Peled M, Somerville CR (2008) Biomass Recalcitrance. In H M, ed,

Biosynthesis of plant cell walls Vol 5. Blackwell Publishing, Oxford, pp 94-187

220

Morrissey JP, Osbourn AE (1999) Fungal resistance to plant antibiotics as a mechanism of

pathogenesis. Microbiol Mol Biol Rev 63: 708-724

Mutwil M, Debolt S, Persson S (2008) Cellulose synthesis: a complex complex. Curr Opin

Plant Biol 11: 252-257

Nakayama N, Arroyo JM, Simorowski J, May B, Martienssen R, Irish VF (2005) Gene trap

lines define domains of gene regulation in Arabidopsis petals and stamens. Plant Cell 17:

2486-2506

Neufeld EF, Feingold DS, Hassid WZ (1959) Enzymic phosphorylation of D-glucuronic acid

by extracts from seedlings of Phaseolus aureus. Arch Biochem Biophys 83: 96-100

Neufeld EF, Feingold DS, Ilves SM, Kessler G, Hassid WZ (1961) Phosphorylation of D-

galacturonic acid by extracts from germinating seeds of Phaseolus aureus. J Biol Chem

236: 3102-3105

Nishitani K, Tominaga R (1992) Endo-xyloglucan transferase, a novel class of

glycosyltransferase that catalyzes transfer of a segment of xyloglucan molecule to

another xyloglucan molecule. J Biol Chem 267: 21058-21064

Noguchi A, Horikawa M, Fukui Y, Fukuchi-Mizutani M, Iuchi-Okada A, Ishiguro M, Kiso

Y, Nakayama T, Ono E (2009) Local differentiation of sugar donor specificity of

flavonoid glycosyltransferase in Lamiales. Plant Cell 21: 1556-1572

Norman C, Vidal S, Palva ET (1999) Oligogalacturonide-mediated induction of a gene

involved in jasmonic acid synthesis in response to the cell-wall-degrading enzymes of the

plant pathogen Erwinia carotovora. Mol Plant Microbe Interact 12: 640-644

Notredame C, Higgins DG, Heringa J (2000) T-Coffee: A novel method for fast and accurate

multiple sequence alignment. J Mol Biol 302: 205-217

221

O'Neill MA, Eberhard S, Albersheim P, Darvill AG (2001) Requirement of borate cross-

linking of cell wall rhamnogalacturonan II for Arabidopsis growth. Science 294: 846-849

O'Neill MA, Ishii T, Albersheim P, Darvill AG (2004) Rhamnogalacturonan II: structure and

function of a borate cross-linked cell wall pectic polysaccharide. Annu Rev Plant Biol 55:

109-139

O'Neill MA, York WS (2003) The composition and structure of plant primary walls. In JK

Rose, ed, The Plant Cell Wall Blackwell, Oxford, pp 1-54

Okazaki Y, Shimojima M, Sawada Y, Toyooka K, Narisawa T, Mochida K, Tanaka H,

Matsuda F, Hirai A, Hirai MY, Ohta H, Saito K (2009) A chloroplastic UDP-glucose

pyrophosphorylase from Arabidopsis is the committed enzyme for the first step of

sulfolipid biosynthesis. Plant Cell 21: 892-909

Orzechowski S (2008) Starch metabolism in leaves. Acta Biochim Pol 55: 435-445

Pastuszak I, Ketchum C, Hermanson G, Sjoberg EJ, Drake R, Elbein AD (1998) GDP-L-

fucose pyrophosphorylase. Purification, cDNA cloning, and properties of the enzyme. J

Biol Chem 273: 30165-30174

Pattathil S, Avci U, Baldwin D, Swennes AG, McGill JA, Popper Z, Bootten T, Albert A,

Davis RH, Chennareddy C, Dong R, O'Shea B, Rossi R, Leoff C, Freshour G, Narra

R, O'Neil M, York WS, Hahn MG (2010) A comprehensive toolkit of plant cell wall

glycan-directed monoclonal antibodies. Plant Physiol 153: 514-525

Pattathil S, Harper AD, Bar-Peled M (2005) Biosynthesis of UDP-xylose: characterization of

membrane-bound AtUxs2. Planta 221: 538-548

Pelloux J, Rusterucci C, Mellerowicz EJ (2007) New insights into pectin methylesterase

structure and function. Trends Plant Sci 12: 267-277

222

Peneff C, Ferrari P, Charrier V, Taburet Y, Monnier C, Zamboni V, Winter J, Harnois M,

Fassy F, Bourne Y (2001) Crystal structures of two human pyrophosphorylase isoforms

in complexes with UDPGlc(Gal)NAc: role of the alternatively spliced insert in the

enzyme oligomeric assembly and active site architecture. EMBO J 20: 6191-6202

Pieslinger AM, Hoepflinger MC, Tenhaken R (2010) Cloning of Glucuronokinase from

Arabidopsis thaliana, the last missing enzyme of the myo-inositol oxygenase pathway to

nucleotide sugars. J Biol Chem 285: 2902-2910

Pressey R (1984) Purification and characterization of tomato polygalacturonase converter. Eur J

Biochem 144: 217-221

Pressey R, Avants JK (1977) Occurrence and Properties of Polygalacturonase in Avena and

Other Plants. Plant Physiol 60: 548-553

Price MN, Dehal PS, Arkin AP (2009) FastTree: computing large minimum evolution trees

with profiles instead of a distance matrix. Mol Biol Evol 26: 1641-1650

Qin C, Qian W, Wang W, Wu Y, Yu C, Jiang X, Wang D, Wu P (2008) GDP-mannose

pyrophosphorylase is a genetic determinant of ammonium sensitivity in Arabidopsis

thaliana. Proc Natl Acad Sci U S A 105: 18308-18313

Raetz CR, Whitfield C (2002) Lipopolysaccharide endotoxins. Annu Rev Biochem 71: 635-700

Rathore RS, Garg N, Garg S, Kumar A (2009) Starch phosphorylase: Role in starch

metabolism and biotechnological applications. Crit Rev Biotechnol 29: 214-224

Reiter WD, Vanzin GF (2001) Molecular genetics of nucleotide sugar interconversion

pathways in plants. Plant Mol Biol 47: 95-113

Ren C, Kermode AR (2000) An increase in pectin methyl esterase activity accompanies

dormancy breakage and germination of yellow cedar seeds. Plant Physiol 124: 231-242

223

Rensing SA, Lang D, Zimmer AD, Terry A, Salamov A, Shapiro H, Nishiyama T, Perroud

PF, Lindquist EA, Kamisugi Y, Tanahashi T, Sakakibara K, Fujita T, Oishi K, Shin

IT, Kuroki Y, Toyoda A, Suzuki Y, Hashimoto S, Yamaguchi K, Sugano S, Kohara

Y, Fujiyama A, Anterola A, Aoki S, Ashton N, Barbazuk WB, Barker E, Bennetzen

JL, Blankenship R, Cho SH, Dutcher SK, Estelle M, Fawcett JA, Gundlach H,

Hanada K, Heyl A, Hicks KA, Hughes J, Lohr M, Mayer K, Melkozernov A,

Murata T, Nelson DR, Pils B, Prigge M, Reiss B, Renner T, Rombauts S, Rushton

PJ, Sanderfoot A, Schween G, Shiu SH, Stueber K, Theodoulou FL, Tu H, Van de

Peer Y, Verrier PJ, Waters E, Wood A, Yang L, Cove D, Cuming AC, Hasebe M,

Lucas S, Mishler BD, Reski R, Grigoriev IV, Quatrano RS, Boore JL (2008) The

Physcomitrella genome reveals evolutionary insights into the conquest of land by plants.

Science 319: 64-69

Ridley BL, O'Neill MA, Mohnen D (2001) Pectins: structure, biosynthesis, and

oligogalacturonide-related signaling. Phytochemistry 57: 929-967

Roper JR, Guther ML, Macrae JI, Prescott AR, Hallyburton I, Acosta-Serrano A,

Ferguson MA (2005) The suppression of galactose metabolism in procylic form

Trypanosoma brucei causes cessation of cell growth and alters procyclin glycoprotein

structure and copy number. J Biol Chem 280: 19728-19736

Roper JR, Guther ML, Milne KG, Ferguson MA (2002) Galactose metabolism is essential for

the African sleeping sickness parasite Trypanosoma brucei. Proc Natl Acad Sci U S A

99: 5884-5889

224

Roze LV, Chanda A, Linz JE (2011) Compartmentalization and molecular traffic in secondary

metabolism: a new understanding of established cellular processes. Fungal Genet Biol

48: 35-48

Sanda S, Leustek T, Theisen MJ, Garavito RM, Benning C (2001) Recombinant Arabidopsis

SQD1 converts udp-glucose and sulfite to the sulfolipid head group precursor UDP-

sulfoquinovose in vitro. J Biol Chem 276: 3941-3946

Sashida Y, Nakata H, Shimomura H, Kagaya M (1983) Sesquiterpene Lactones From

Pyrethrum Flowers. Phytochemistry 22: 1219-1222

Schnaar R, Suzuki A, Stanley P (2008) Glycosphingolipids. In Varki A, Cummings RD, Esko

JD, Freeze HH, Stanley P, Bertozzi CR, Hart GW, Etzler ME, eds, Essentials of

Glycobiology. Cold Spring Harbor Laboratory Press, New York

Schnurr JA, Storey KK, Jung HJ, Somers DA, Gronwald JW (2006) UDP-sugar

pyrophosphorylase is essential for pollen development in Arabidopsis. Planta 224: 520-

532

Seifert GJ (2004) Nucleotide sugar interconversions and cell wall biosynthesis: how to bring the

inside to the outside. Curr Opin Plant Biol 7: 277-284

Semon M, Wolfe KH (2007) Consequences of genome duplication. Curr Opin Genet Dev 17:

505-512

Sherson S, Gy I, Medd J, Schmidt R, Dean C, Kreis M, Lecharny A, Cobbett C (1999) The

arabinose kinase, ARA1, gene of Arabidopsis is a novel member of the galactose kinase

gene family. Plant Mol Biol 39: 1003-1012

Siedlecka A, Ciereszko I, Mellerowicz E, Martz F, Chen J, Kleczkowski LA (2003) The

small subunit ADP-glucose pyrophosphorylase ( ApS) promoter mediates okadaic acid-

225

sensitive uidA expression in starch-synthesizing tissues and cells in Arabidopsis. Planta

217: 184-192

Silverstone AL, Tseng TS, Swain SM, Dill A, Jeong SY, Olszewski NE, Sun TP (2007)

Functional analysis of SPINDLY in gibberellin signaling in Arabidopsis. Plant Physiol

143: 987-1000

Simpson SD, Ashford DA, Harvey DJ, Bowles DJ (1998) Short chain oligogalacturonides

induce ethylene production and expression of the gene encoding aminocyclopropane 1-

carboxylic acid oxidase in tomato plants. Glycobiology 8: 579-583

Sitrit Y, Hadfield KA, Bennett AB, Bradford KJ, Downie AB (1999) Expression of a

polygalacturonase associated with tomato seed germination. Plant Physiol 121: 419-428

Smith AM (2001) The biosynthesis of starch granules. Biomacromolecules 2: 335-341

Smith EE, Mills GT (1954) Uridine nucleotide compounds of liver. Biochim Biophys Acta 13:

386-400

Solbak AI, Richardson TH, McCann RT, Kline KA, Bartnek F, Tomlinson G, Tan X,

Parra-Gessert L, Frey GJ, Podar M, Luginbuhl P, Gray KA, Mathur EJ, Robertson

DE, Burk MJ, Hazlewood GP, Short JM, Kerovuo J (2005) Discovery of pectin-

degrading enzymes and directed evolution of a novel pectate lyase for processing cotton

fabric. J Biol Chem 280: 9431-9438

Stanley P, Schachter H, Taniguchi N (2008) N-Glycans. In Varki A, Cummings RD, Esko JD,

Freeze HH, Stanley P, Bertozzi CR, Hart GW, Etzler ME, eds, Essentials of

Glycobiology. Cold Spring Harbor Laboratory Press, New York

226

Steiner T, Lamerz AC, Hess P, Breithaupt C, Krapp S, Bourenkov G, Huber R, Gerardy-

Schahn R, Jacob U (2007) Open and closed structures of the UDP-glucose

pyrophosphorylase from Leishmania major. J Biol Chem 282: 13003-13010

Sterling JD, Atmodjo MA, Inwood SE, Kumar Kolli VS, Quigley HF, Hahn MG, Mohnen

D (2006) Functional identification of an Arabidopsis pectin biosynthetic

homogalacturonan galacturonosyltransferase. Proc Natl Acad Sci U S A 103: 5236-5241

Sterling JD, Quigley HF, Orellana A, Mohnen D (2001) The catalytic site of the pectin

biosynthetic enzyme alpha-1,4-galacturonosyltransferase is located in the lumen of the

Golgi. Plant Physiol 127: 360-371

Stokes MJ, Guther ML, Turnock DC, Prescott AR, Martin KL, Alphey MS, Ferguson MA

(2008) The synthesis of UDP-N-acetylglucosamine is essential for bloodstream form

trypanosoma brucei in vitro and in vivo and UDP-N-acetylglucosamine starvation reveals

a hierarchy in parasite protein glycosylation. J Biol Chem 283: 16147-16161

Sundaresan V, Springer P, Volpe T, Haward S, Jones JD, Dean C, Ma H, Martienssen R

(1995) Patterns of gene action in plant development revealed by enhancer trap and gene

trap transposable elements. Genes Dev 9: 1797-1810

Tang H, Wang X, Bowers JE, Ming R, Alam M, Paterson AH (2008) Unraveling ancient

hexaploidy through multiply-aligned angiosperm gene maps. Genome Res 18: 1944-1954

Tanner ME (2001) Sugar Nucleotide-Modifying Enzymes. Current Organic Chemistry 5: 24

Thoden JB, Holden HM (2003) Molecular structure of galactokinase. J Biol Chem 278: 33305-

33311

Thoden JB, Timson DJ, Reece RJ, Holden HM (2005) Molecular structure of human

galactokinase: implications for type II galactosemia. J Biol Chem 280: 9662-9670

227

Tiessen A, Hendriks JH, Stitt M, Branscheid A, Gibon Y, Farre EM, Geigenberger P

(2002) Starch synthesis in potato tubers is regulated by post-translational redox

modification of ADP-glucose pyrophosphorylase: a novel regulatory mechanism linking

starch synthesis to the sucrose supply. Plant Cell 14: 2191-2213

Tiessen A, Prescha K, Branscheid A, Palacios N, McKibbin R, Halford NG, Geigenberger

P (2003) Evidence that SNF1-related kinase and hexokinase are involved in separate

sugar-signalling pathways modulating post-translational redox activation of ADP-glucose

pyrophosphorylase in potato tubers. Plant J 35: 490-500

Timson DJ (2007) GHMP Kinases – Structures, Mechanisms and Potential for Therapeutically

Relevant Inhibition. Current Enzyme Inhibition 3: 77-94

Timson DJ, Reece RJ (2003) Sugar recognition by human galactokinase. BMC Biochem 4: 16

Tolstikov VV, Fiehn O (2002) Analysis of highly polar compounds of plant origin: combination

of hydrophilic interaction chromatography and electrospray ion trap mass spectrometry.

Anal Biochem 301: 298-307

Turnock DC, Ferguson MA (2007) Sugar nucleotide pools of Trypanosoma brucei,

Trypanosoma cruzi, and Leishmania major. Eukaryot Cell 6: 1450-1463

Urbaniak MD, Turnock DC, Ferguson MA (2006) Galactose starvation in a bloodstream form

Trypanosoma brucei UDP-glucose 4'-epimerase conditional null mutant. Eukaryot Cell 5:

1906-1913

Usadel B, Schluter U, Molhoj M, Gipmans M, Verma R, Kossmann J, Reiter WD, Pauly M

(2004) Identification and characterization of a UDP-D-glucuronate 4-epimerase in

Arabidopsis. FEBS Lett 569: 327-331

228

Verhees CH, Koot DG, Ettema TJ, Dijkema C, de Vos WM, van der Oost J (2002)

Biochemical adaptations of two sugar kinases from the hyperthermophilic archaeon

Pyrococcus furiosus. Biochem J 366: 121-127

Warnecke D, Heinz E (2010) Glycolipid headgroup replacement: a new approach for the

analysis of specific functions of glycolipids in vivo. Eur J Cell Biol 89: 53-61

Wells L, Kreppel LK, Comer FI, Wadzinski BE, Hart GW (2004) O-GlcNAc transferase is in

a functional complex with protein phosphatase 1 catalytic subunits. J Biol Chem 279:

38466-38470

Willats WG, McCartney L, Mackie W, Knox JP (2001) Pectin: cell biology and prospects for

functional analysis. Plant Mol Biol 47: 9-27

Wilson JE (1995) Hexokinases. Rev Physiol Biochem Pharmacol 126: 65-198

Worden AZ, Lee JH, Mock T, Rouze P, Simmons MP, Aerts AL, Allen AE, Cuvelier ML,

Derelle E, Everett MV, Foulon E, Grimwood J, Gundlach H, Henrissat B, Napoli C,

McDonald SM, Parker MS, Rombauts S, Salamov A, Von Dassow P, Badger JH,

Coutinho PM, Demir E, Dubchak I, Gentemann C, Eikrem W, Gready JE, John U,

Lanier W, Lindquist EA, Lucas S, Mayer KF, Moreau H, Not F, Otillar R, Panaud

O, Pangilinan J, Paulsen I, Piegu B, Poliakov A, Robbens S, Schmutz J, Toulza E,

Wyss T, Zelensky A, Zhou K, Armbrust EV, Bhattacharya D, Goodenough UW,

Van de Peer Y, Grigoriev IV (2009) Green evolution and dynamic adaptations revealed

by genomes of the marine picoeukaryotes Micromonas. Science 324: 268-272

Yang T, Bar-Peled L, Gebhart L, Lee SG, Bar-Peled M (2009) Identification of Galacturonic

Acid-1-phosphate Kinase, a New Member of the GHMP Kinase Superfamily in Plants,

and Comparison with Galactose-1-phosphate Kinase. J Biol Chem 284: 21526-21535

229

Yang T, Bar-Peled M (2010) Identification of a novel UDP-sugar pyrophosphorylase with a

broad substrate specificity in Trypanosoma cruzi. Biochem J 429: 533-543

Yang T, Echols M, Martin A, Bar-Peled M (2010) Identification and characterization of a

strict and a promiscuous N-acetylglucosamine-1-P uridylyltransferase in Arabidopsis.

Biochem J 430: 275-284

York WS, Darvill AG, NcNeil M, Stevenson TT, Albersheim P (1986) Isolation and

Characterization of plant cell walls and cell wall components. In NP Colowick, NP

Kaplan, A Weissbach, H Weissbach, eds, Method in enzymology, Vol 118. Academic

Press, Orlando, Florida

Young LCT, Stanley RG, Loewus FA (1966) Myo-inositol-2-t Incorporation by Germinating

Pollen. Nature 209: 530-531

Yu J, Hu S, Wang J, Wong GK, Li S, Liu B, Deng Y, Dai L, Zhou Y, Zhang X, Cao M, Liu

J, Sun J, Tang J, Chen Y, Huang X, Lin W, Ye C, Tong W, Cong L, Geng J, Han Y,

Li L, Li W, Hu G, Li J, Liu Z, Qi Q, Li T, Wang X, Lu H, Wu T, Zhu M, Ni P, Han

H, Dong W, Ren X, Feng X, Cui P, Li X, Wang H, Xu X, Zhai W, Xu Z, Zhang J, He

S, Xu J, Zhang K, Zheng X, Dong J, Zeng W, Tao L, Ye J, Tan J, Chen X, He J, Liu

D, Tian W, Tian C, Xia H, Bao Q, Li G, Gao H, Cao T, Zhao W, Li P, Chen W,

Zhang Y, Hu J, Liu S, Yang J, Zhang G, Xiong Y, Li Z, Mao L, Zhou C, Zhu Z,

Chen R, Hao B, Zheng W, Chen S, Guo W, Tao M, Zhu L, Yuan L, Yang H (2002)

A draft sequence of the rice genome (Oryza sativa L. ssp. indica). Science 296: 79-92

Zeidan Q, Wang Z, De Maio A, Hart GW (2010) O-GlcNAc cycling enzymes associate with

the translational machinery and modify core ribosomal proteins. Mol Biol Cell 21: 1922-

1936

230

Zhang Q, Hrmova M, Shirley NJ, Lahnstein J, Fincher GB (2006) Gene expression patterns

and catalytic properties of UDP-D-glucose 4-epimerases from barley (Hordeum vulgare

L.). Biochem J 394: 115-124

Zhou T, Daugherty M, Grishin NV, Osterman AL, Zhang H (2000) Structure and mechanism

of homoserine kinase: prototype for the GHMP kinase superfamily. Structure 8: 1247-

1257

231

APPENDIX A

SUPPLEMENTAL FIGURES AND TABLES

FIGURE S2.1. Full 1H-NMR spectrum of UDP-GalA derived from reaction products of

GalAK.

A. Expanded proton NMR spectra that covers the sugar anomeric region (5.5-6 ppm).

B. Expended view of the proton NMR spectra that covers the nucleotide-sugar carbon ring of

UDP-GalA between 3.8 and 4.6 ppm.

232

FIGURE S2.2. Full 1H-NMR spectrum of UDP-Gal derived from reaction products of

GalK.

A. Expanded proton NMR spectra that covers the sugar anomeric region (5.5-6 ppm).

B. Expended view of the proton NMR spectra that covers the nucleotide-sugar carbon ring of

UDP-Gal between 3.8 and 4.4 ppm.

233

FIGURE S2.3. Expression and purification of recombinant wild type (wt) and mutants of

GalK.

Panel A. SDS-PAGE analyses of total soluble protein (S20) and inclusion bodies (P20) derived from cell expressing WT or mutant GalK altered in the following aa (E62A; Y262F; A437S; S206G).

Panel B. SDS-PAGE analyses of column-purified wt and mutants GalK recombinant proteins.

234

FIGURE S2.4. Expression and purification of recombinant wild type (wt) and mutants of

GalAK.

Panel A. SDS-PAGE analyses of total soluble protein (S20) and inclusion bodies (P20) derived from cell expressing WT or mutant GalAK altered in the following aa (A41E; Y250F; A368S). Panel B.

SDS-PAGE analyses of column-purified wt and mutants GalAK recombinant proteins.

235

FIGURE S2.5. Selected portions of the 1H-NMR spectra corresponding to α-GalA and β-

GalA peaks before and during GalAK conversion of α-GalA to α-GalA-1-P.

A. Portions of the 1H-NMR spectrum (no enzyme) showing the anomeric proton (H-1) of α-GalA

(5.27 ppm), the anomeric proton (H-1) of the β-GalA (4.55 ppm), and the H-5 peak of β-GalA

(4.03 ppm). D-GalA (2 mM) was mixed with 50 mM phosphate buffer (pH/pD = 7.6), 5 mM

1 MgCl2 and D2O and monitored continuously by H-NMR spectroscopy. The spectra shown were obtained after 2, 6, and 20 min.

236

B. After ~ 20 min, ATP (2 mM) was added and monitored continuously by 1H-NMR spectroscopy. The spectra shown (+ATP) were obtained after 2, 6, and 20 min. The panels are showing no change in the α and β configuration at time 2, 6, and 20 min post addition of ATP.

NOTE: signal from the ribose protons (H-2‟, 3‟) of ATP are overlap with H-1 of β-GalA (β1).

C. After ~ 20 min in the presence of ATP, the recombinant kinase (GalAK, 3 µg) was added and the reaction monitored continuously by 1H-NMR spectroscopy. The spectra shown (+ enzyme) were obtained after 2, 6, and 20 min, and show the enzymatic phosphorylation of the α-GalA configuration and its conversion to α-GalA-1-P.

Abbreviation of the NMR spectrum are as follows: α1, peak corresponding to H-1 of α-GalA

(5.27 ppm); β1, peak corresponding to H-1 of β-GalA (4.55 ppm); β5, peak corresponding to H-5 of β-GalA (4.03 ppm); α1P, peak corresponding to H-1 of α-GalA-1-P, (5.52 ppm). Note the β1 peaks in panels B, overlap with ribose protons of ATP; the β1 peaks in panels C, overlap with the ribose protons from ADP/ATP thus creating additional shift when compared with B.

237

FIGURE S2.6. Selected portions of the 1H-NMR spectra corresponding to α-Gal and β-Gal peaks before and during GalK conversion of α-Gal to α-Gal-1-P.

A. Portions of the 1H-NMR spectrum (no enzyme) showing the anomeric proton (H-1) of α-Gal

(5.24 ppm), the anomeric proton (H-1) of the β-Gal (4.56 ppm), and the H-4 peak of β-Gal (3.90 ppm). D-Gal (2 mM) was mixed with 50 mM phosphate buffer (pH/pD = 7.6), 5 mM MgCl2 and

1 D2O and monitored continuously by H-NMR spectroscopy. The spectra shown were obtained after 2, 6, and 20 min.

238

NOTE: signal from the ribose protons (H-2‟, 3‟) of ATP are overlap with H-1 of β-Gal (β1).

C. After ~ 20 min in the presence of ATP, the recombinant kinase (GalK, 3 µg) was added and the reaction monitored continuously by 1H-NMR spectroscopy. The spectra shown (+ enzyme) were obtained after 2, 6, and 20 min, and show the enzymatic phosphorylation of the α-Gal configuration and its conversion to α-Gal-1-P.

Abbreviation of the NMR spectrum are as follows: α1, peak corresponding to H-1 of α-Gal (5.24 ppm); β1, peak corresponding to H-1 of β-Gal (4.56 ppm); β4, peak corresponding to H-4 of β-

Gal (3.90 ppm); α1P, peak corresponding to H-1 of α-Gal-1-P, (5.47 ppm). Note the β1 peaks in panels B, overlap with ribose protons of ATP; the β1 peaks in panels C, overlap with the ribose protons from ADP/ATP thus creating additional shift when compared with B.

239

FIGURE S3.1. Expression of recombinant T. cruzi Sloppy.

SDS-PAGE of total soluble protein isolated from E. coli cell expressing T. cruzi recombinant

Sloppy (lane 2), control empty vector (lane 3), Nickel-column purified recombinant TcSloppy

(lane 4), Nickel-column purified control empty vector (lane 5).

240

241

242

243

FIGURE S3.2. Product analyses of HPLC-based assays by NMR, confirming the recombinant Tc enzyme has Sloppy UDP-sugar PPase activity.

Proton NMR spectra of the products formed by TcSloppy.

A. 1H-NMR spectrum of UDP-Glc.

244

B. 1H-NMR spectrum of UDP-Gal.

C. 1H-NMR spectrum of UDP-GlcA.

D. 1H-NMR spectrum of UDP-Xyl.

Each peak eluted from HPLC column (see Fig. 3.2, Arrow #1, #2, #4) was collected, lyophilized,

1 dissolved in D2O and analyzed by H-NMR. Assay in panel 3 was chromatographed on a Q15

(as described Yang et al., [15]) that separates UDP-GlcA from UTP, and the UDP-GlcA was collected and analyzed by NMR. The spectra shown in A-D are individual UDP-sugar produced by TcSloppy. Panel “a” covers the sugar anomeric region (5.5-6 ppm) and panel “b” is the expansion view spectrum of the UDP-sugar carbon ring (3.4-4.4 ppm).

245

246

247

FIGURE S3.3. Real-Time 1H NMR based Sloppy assays with different sugar-1-Ps.

Gal-1-P, GlcA-1-P or Xyl-1-P (1 mM) was separately reacted with 1 mM UTP and TcSloppy.

Approximately, 2 min after enzyme addition and NMR shimming, data were collected. The progression of enzyme activity monitored by following changes in the spectrum of the sugar anomeric proton region (from 5.4 to 5.7 ppm) is shown. To visualize changes in product with time, each spectrum at a given time is plotted sequentially. Each time point reflects the amount of UDP-sugar formed (Usugar, ) in the forward reaction, and the decreased the amount of sugar-1-P (sugar1P, ).

A. A time dependent 1H-NMR spectrum arrays for the formation of UDP-Gal by TcSloppy.

B. A time dependent 1H-NMR spectrum arrays for the formation of UDP-GlcA by TcSloppy.

C. A time dependent 1H-NMR spectrum arrays for the formation of UDP-Xyl by TcSloppy.

D. Integral of the quadruplet peak of the anomeric proton signal of UDP-sugar at each time point was calculated using the NMR software and plotted against time. The relative amount of individual UDP-sugar made during the time course of the enzyme reaction in NMR is shown.

248

249

FIGURE S4.1. Sequence alignments of N-acetylglucosamine-1-phosphate uridylyltransferase (GlcNAc1pUT) from different organisms.

Sequences of GlcNAc1pUT and Arabidopsis UDP-Glc Pyrophosphorylase (see gene names below) were aligned using T-coffee software (Notredame et al., 2000) with G-block (Castresana,

2000). The conserved motifs presumably involved in nucleotide sugar binding (NB) and the uracil binding (UB) are labeled in bold. Potential amino acids that are conserved in

GlcNAc1pUT and AtUGP are highlighted in grey, based on sequence alignment.

Arabidopsis thaliana GlcNAc1pUT-1 (GU937393, At1g31070), Arabidopsis thaliana

GlcNAc1pUT-2 (GU937394, At2g35020), Homo sapiens GlcNAc1pUT (AGX1,1jv1),

Saccharomyces cerevisiae GlcNAc1pUT (QRI1, NP_010180.1), UDP-Glc PPase from

Arabidopsis thaliana (AtUGP, NP_186975).

250

FIGURE S4.2. Product analyses by NMR, confirm the specific activity of recombinant

GlcNAc1pUT-1.

251

Proton NMR spectra of GlcNAc1pUT-1 enzymatic products. Peak eluted from the Bio-Rad

1 column (see Fig. 4.2B, #1) was collected, lyophilized, dissolved in D2O and analyzed by H-

NMR. The full NMR spectrum (2-8 ppm) of GlcNAc1pUT-1 reaction product is shown. A more detailed spectrum that covers the sugar anomeric region (5-6 ppm) and the NDP-sugar carbon ring (3.5-4.4 ppm) are shown in panel “a” and panel “b”. The 1H chemical shift values of specific protons along the UDP-α-D-GlcNAc structure are indicated by underline. H refer to protons belong to the GlcNAc ring; H‟ refer to protons belong to the ribose ring, and H” refer to protons belongs to the uracil ring. The peak labeled with “*” is the signal from column contaminants.

252

FIGURE S4.3. The effects of buffer and pH on GlcNAc1pUT-1 activity.

GlcNAc1pUT-1 activity was determined at different buffers at different pH. Each value is the mean of duplicate reactions, and the values varied by no more than ±5%.

253

FIGURE S4.4. Product analyses by NMR, confirm the specific activity of recombinant

GlcNAc1pUT-2.

Proton NMR spectra of GlcNAc1pUT-2 enzymatic products. Peak eluted from the Bio-Rad

1 column (see Fig. 4.4B, #2) was collected, lyophilized, dissolved in D2O and analyzed by H-

NMR. The full NMR spectrum (3.5-8 ppm) of GlcNAc1pUT-2 reaction product is shown. A more detailed spectrum that covers the sugar anomeric region (5-6 ppm) and the NDP-sugar

254

carbon ring (3.4-4.4 ppm) are shown in panel “a” and panel “b”. The 1H chemical shift values of specific protons along the UDP-α-D-Glc structure are indicated by underline. H refer to protons belong to the Glc ring; H‟ refer to protons belong to the ribose ring, and H” refer to protons belongs to the uracil ring. The peak labeled with “*” is the signal from column contaminants.

255

TABLE S3.1. The effect of potential inhibitors on TcSloppy activity.

Additive (0.5 mM) Relative TcSloppy Activity (%)

UMP 101

NADH 105

CTP 101

Fru-1-P 108

Man-1-P 98

Fuc-1-P 101

Glc-6-P 99

GalN-1-P 100

Mannitol-1-P 106

TTP 101

ATP 99

NAD+ 101

GTP 98

ADP 85

AMP 98

NADP + 98

PPi 64

ITP 99

UDP 109

Glc-1,6-di-P 99

Fru 100

256

Control 100

Inhibitors (at 0.5 mM), or control (water) were mixed with TcSloppy in 100 mM Tris-HCl, pH

7.6 for 10 min on ice prior to performing the enzymatic reaction. Each value is the mean of duplicate reactions, and the values varied by no more than ±10%.

257

TABLE S4.1. The effect of potential inhibitors on GlcNAc1pUT-1 activity.

Relative GlcNAc1pUT Activity Additive Forward reaction (%)

UMP 97

NADH 96

CTP 99

TTP 98

ATP 99

NAD+ 96

GTP 100

ADP 100

NADP + 96

PPi 56

ITP 96

UDP 101

Glc-1-P 99

Gal-1-P 99

GlcA-1-P 97

GalA-1-P 100

GalN-1-P 98

Xyl-1-P 100

Fuc-1-P 98

GalNAc 97

258

GlcNAc 101

Man-1-P 99

Gentamycin 93

Kanamycin 102

Streptomycin 74

Hygromycin 37

Control 100

Sugar-1-Ps and nucleotides (at final concentration of 0.5 mM), antibiotics (at 1mM) or control

(water) were mixed with GlcNAc1pUT-1 in 100 mM Tris-HCl, pH 7.6 for 10 min on ice prior to performing the enzymatic reaction. Each value is the mean of duplicate reactions, and the values varied by no more than ±10%.

259

APPENDIX B

SUMMARY OF PRIMERS, ANTIBODIES, CLONING CONSTRUCTS AND

TRANSGENIC PLANTS

TABLE B1. Summary of primers.

Oligo # Oligo Name 5'-3' sequence location

1 GalAK S AC atg tct tgg cct acg gat tct gag 1

2 GalAK AS GG TAC CTC gag aag aac acg agc agc gtc 1275

3 GalK S GCC atg gcg aaa ccg gaa gaa gta tca gtc 1

4 GalK AS GAT ATC TCG AGg agg ttg aag atg gca gca c 1491

5 Sloppy S atg gct tct acg gtt gat tcc 1

6 Sloppy AS gaa gag aag tcc att tgt atc ttg 1845

7 GalAK A41E S gt cct tta gga gAG cac att gat cac cag gg 110

8 GalAK A41E AS cc ctg gtg atc aat gtg CTc tcc taa agg ac 140

9 GalAK Y250F S c aac cca gga tTt aat ctg cga gtt tct gag tg 738

10 GalAK Y250F AS c tcg cag att aAa tcc tgg gtt ggt ggt caa cg 760

260

11 GalAK A368S S ga ttc agc ggt Tca ggt ttc agg gga tgt tg 1091

12 GalAK A368S AS ct gaa acc tgA acc gct gaa tct agc tcc 1112

13 GalK E62A S ctg ata gga gCg cac att gac tat gaa gga tac 175

14 GalK E62A AS gtc aat gtg cGc tcc tat cag att cac tct tcc 195

15 GalK Y262F S cg gct gct aag aat tTc aat aac agg gtc gtt g 770

16 GalK Y262F AS cct gtt att gAa att ctt agc agc cgt gac cg 795

17 GalK A437S S ga ctg acc gga Tct gga tgg ggc ggt tgc 1298

18 GalK A437S AS gcc cca tcc agA tcc ggt cag tct tgc tcc 1320

19 GalK S206G S t gga aca caa GGt ggt ggg atg gac cag gc 606

20 GalK S206G AS g gtc cat ccc acc aCC ttg tgt tcc aat gtg tc 631

21 GalAK QS gta tct ggg tct gcg gaa tg 561

22 GalAK QAS caa gct cgt ggt cca aag tc 664

23 GalK QS ggt gct tct ccc caa ctc tt 130

24 GalK QAS gaa tag cca tcg gca aca ct 229

25 Actin QS ggt aac att gtg ctc agt ggt gg -

26 Actin QAS aac gac ctt aat ctt cat gct gc -

27 EYFP CGT TTA CGT CGC CGT CCA GCT C -

261

28 GalAK 3UTR aa ctc cag gat tga TCA GAG 1289

29 GalAK Ext F CAT CCT CAG ACG TGC ATA AGC 1st intron

30 GalAK Int F GAA ATC CAG CAC CCA ATA GGC 262

31 GlcNAc1pUT-1 S C ACC atg Gta gaa ccg tcg atg gag aga g 1

32 GlcNAc1pUT-1 AS GGA TCC agg gaa att tca caa ggt gca tg 1518

33 TcSloppy S TC atg aag atg gtg cct gac g 1

34 TcSloppy AS GG ATC cta aag ctt cgc atg atg 1812

35 LmSloppy S TC atg acg aac ccg tcc aac 1

36 LmSloppy AS GCGG CCG cta caa ctt tgc cgg gtc 1893

37 Sloppy QS#1 CTTTTGTTCTGGTTGCTGGTG 389

38 Sloppy QAS#1 GCATGTCCCTGTAGTTGTCTC 480

39 AraK QS#1 TGGTGTGCGTTTTGAGGATAG 2028

40 AraK QAS#1 CGATACTCAGTCCATGTGCAG 2161

41 FucoK QS#1 TCGCCACAGCTAATTTCAGAG 2719

42 FucoK QAS#1 CCGTCAATCGCTTAATGCTTG 2866

43 GlcAK QS#1 GACATTCAACTTCACCAAGCC 325

44 GlcAK QAS#1 AAGTCGAGAAGGCAGTTAAGG 443

262

45 GalK2 QS#1 CAAGATTACTGGAGGCGGAAG 2700

46 GalK2 QAS#1 GATCCTTCAAACACATACGGC 2837

47 AtUAP1 QS#1 GGGTTTCGACAGTAGATGGTAG 302

48 AtUAP1 QAS#1 TCCTTTTGGATCTGAGCTTCC 447

49 UGlcAE3 QS#1 TCAAGCTGGTGTTAGATACGC 513

50 UGlcAE3 QAS#1 TGAAGCCCAAACTATCGCC 639

51 UGlcAE5 QS#1 TCTACGCCTGGGAAATTCAAG 25

52 UGlcAE5 QAS#1 TGGAGATGGGAGATCGGTAG 172

263

TABLE B2. Summary of constructs.

Construct # Construct name

1 pCR4-topoTA:At3g10700#11

2 pET28b:At3g10700#11.3

3 pET28a:At3g06580#3.1

4 pET28b:At5g52560# a73f/2#2

5 pET28b:AtGalAK#11.3A41E

6 pET28b:AtGalAK#11.3Y250F

7 pET28b:AtGalAK#11.3A368S

8 pET28a:AtGalK#3.1E62A

9 pET28a:AtGalK#3.1Y262F

10 pET28a:AtGalK#3.1A437S

11 pET28a:AtGalK#3.1S206G

12 pCR4-topoTA:TcSloppy

13 pET28b:TcSloppy

14 pCR4-topoTA:LmSloppy

15 pET28b:LmSloppy

264

16 pCR4-topoTA:at1g31070.11#6

17 pET28b: at1g31070.11#6

18 pET28c:at2g35020#13

19 pCAM35stl:GalAK-EYFP

20 pCAM35stl:GalAK-6×His

265

TABLE B3. Summary of antibodies used in the immunohistochemistry.

Name Epitopes

CCRC-M1 Fuc-Xyloglucan

CCRC-M14 RG-I backbone

CCRC-M34 HG

CCRC-M35 RG-I backbone

CCRC-M38 Pectin (Homogalacturonan de-esterified)

CCRC-M48 Non-Fuc Xyloglucan (XXLG+XLLG)

CCRC-M88 Non-Fuc Xyloglucan (XXLG+XLLG+XLXG+XXFG)

CCRC-M93 Non-Fuc Xyloglucan

CCRC-M100 Non-Fuc Xyloglucan (XXXG)

CCRC-M138 Xylan (Xylohexose)

CCRC-M149 Xylan (Xylotetrose)

LAMP Callose (β 1-3 glucan) goat anti mouse (Invitrogen) 2nd Ab goat anti rabbit (Invitrogen) 2nd Ab

266

TABLE B4. List of plants, transgenic plants and origin. plants and transgenic plants origin

Arabidopsis Columbia WT

Arabidopsis Landsberg WT galak KO (Arabidopsis Landsberg) Mutant

35S-GalAK-EYFP#1 (WT Arabidopsis Columbia) GalAK-EYFP overexpression

35S-GalAK-EYFP#2 (WT Arabidopsis Columbia) GalAK-EYFP overexpression

35S-GalAK-EYFP#3 (WT Arabidopsis Columbia) GalAK-EYFP overexpression

35S-GalAK-6His#10-2 (WT Arabidopsis Landsberg) GalAK-6His overexpression

35S-GalAK-6His#15-3 (WT Arabidopsis Landsberg) GalAK-6His overexpression

35S-GalAK-6His#C1-4 (galak Arabidopsis Landsberg) GalAK-6His Complementation

35S-GalAK-6His#C2-3 (galak Arabidopsis Landsberg) GalAK-6His Complementation

35S-GalAK-6His#C2-4 (galak Arabidopsis Landsberg) GalAK-6His Complementation

35S-GalAK-6His#C3-4 (galak Arabidopsis Landsberg) GalAK-6His Complementation

267