Expression of Different Glucansucrases in Potato Tubers: Implications for Starch Biosynthesis

Expression of different glucansucrases in potato tubers: Implications for starch biosynthesis

Géraldine Kok-Jacon

Promotor : Prof. Dr. R.G.F. Visser Hoogleraar in de Plantenveredeling

Co-promotor : Dr. Ir. J.-P. Vincken Universitair docent Laboratorium voor Plantenveredeling

Promotiecommissie : Prof. Dr. Ir. E. Jacobsen, Wageningen Universiteit Prof. Dr. L.H.W. van der Plas, Wageningen Universiteit Prof. Dr. L. Dijkhuizen, Rijks Universiteit Groningen Dr. M. Quanz, Bayer Bioscience, Potsdam, Duitsland

Dit onderzoek is uitgevoerd binnen de onderzoekschool Experimentele Planten Wetenschappen

Expression of different glucansucrases in potato tubers: Implications for starch biosynthesis

Géraldine Kok-Jacon

Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit, Prof. dr. M. J. Kropff, in het openbaar te verdedigen op maandag 19 september 2005 des namiddags te vier uur in de Aula.

Expression of different glucansucrases in potato tubers: Implications for starch biosynthesis. Géraldine Kok-Jacon.

Thesis Wageningen University, The Netherlands, 2005. With references - with summaries in English, Dutch and French.

ISBN 90-8504-262-3

à Sjaak,

Contents

Chapter 1 General introduction 9

Chapter 2 Production of dextran in transgenic potato plants 33

Chapter 3 Mutan produced in potato amyloplasts adheres to starch granules 53

Chapter 4 Production of alternan in transgenic potato plants 73

Chapter 5 Granule-bound mutansucrase alters melting temperature of starch granules 87

Chapter 6 General discussion 107

Summary, Samenvatting, Résumé 117

Nawoord 129

Curriculum Vitae and Publications 131

9 : C hapter 1/ General introduction

Towards a more versatile α-glucan biosynthesis in plants

Géraldine A. Kok-Jacon Jean-Paul Vincken Qin Ji Richard G.F. Visser

Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Wageningen University, P.O. Box 386, 6700 AJ Wageningen, The Netherlands.

Abstract : Starch is an important storage polysaccharide in many plants. It is composed of densely packed α-glucans, consisting of 1,4- and 1,4,6-linked glucose residues. The starch polymers are used in many industrial applications. The biosynthetic machinery for assembling the granule has been modified in many different ways in order to gain insight in the starch biosynthesis process, and to engineer starches with improved functionalities. With respect to the latter, two generic technologies with great potential are developed: (i) introduction of new linkage types in starch polymers (1,3- and 1,6-linkages), and (ii) engineering granule- boundness. The toolbox to engineer this new generation of starch polymers is discussed.

Part of this Chapter was published in: Journal of Plant Physiology 160; 765-777 (2003). Reprinted with permission of Blackwell Publishing

Key words: starch modification, glucansucrase, starch-binding domain, granule-boundness, transgenic potato

10 : Chapter 1/ General introduction

Introduction During the last 10 years, the increased need for starches with novel properties has occupied the research community, and many efforts have concentrated on unraveling the starch biosynthesis pathways. The knowledge generated in these investigations was subsequently used to produce tailor-made starches in higher plants using recombinant DNA technology. Examples of starches with new functionalities are those with a modified degree of branching (Schwall et al. 2000; Shewmaker et al. 1994; Kortstee et al. 1996) and the amylose-free starch (Visser et al. 1991a; Kuipers et al. 1994), some of which hold potential for applications in the paper-, textile-, plastics-, food and pharmaceutical industry. The accumulation of more starch has also been an objective, but this will not be discussed further here (Slattery et al. 2000). In our laboratory, we have embarked on two generic technologies with a very wide range of applicability: (i) introduction of new linkage types and structural elements using glucansucrases, and (ii) engineering granule-boundness by using microbial starch- binding domains (SBDs). It is expected that these technologies will contribute significantly to the biosynthesis of more versatile α-glucans in the near future, leading to starches with improved functionalities that cannot be obtained by conventional breeding. In this paper, these recent developments in starch modification using heterologous expression of microbial genes will be reviewed, with emphasis on the potato crop.

Starch structure and biosynthesis Starch is an important reserve carbohydrate found in many plant species, and is deposited as granules in the chloroplasts of green leaves (transitory starch) and in amyloplasts of tubers, roots and seeds (storage starch) (Kossmann and Lloyd 2000). Each starch granule has a highly organized structure defined by the succession of crystalline and amorphous lamellae (Fig. 1). The two main components of starch, amylopectin and amylose, are polymers of glucose. The major polysaccharide of the two, amylopectin (70-80%) is a highly branched molecule, mainly composed of α-1,4- linked glucosyl residues and 4-6% of α-1,4,6-linked branch points. The distribution of the branch points is not at random , enabling the unique, cluster-based structure of amylopectin (Thompson 2000). Clustering of the branch points in the amorphous lamellae enables the polymer chains to line up in a parallel fashion. The chains can

11 : Chapter 1/ General introduction

Figure 1. Overview of the various l evels of polymer organization within the starch granule, which is composed of amylopectin and amylose molecules

associate with each other to form double helices (Smith et al . 1997; Ball et al . 1998). This chain organization forms the basis of the semi-crystalline structure of the starch granule (Myers et al . 2000). In contrast to amylopectin, amylose (about 20-30%) is an essentially linear polysaccharide with less than 1% of branch points. In plant storage organs, starch biosynthesis takes place within the amyloplast (Fig. 2) and is the result of different reactions such as synthesis (polymerization of glucosyl residues), rearrangement and degradation, in which various starch synthases (E.C.2.4.1.21), transferases (branching (E.C.2.4.1.18) and disproportionating enzymes (E.C.2.4.1.25)), and hydrolytic enzymes (debranching enzyme (E.C.3.2.1.41)), respectively, play key roles. Sucrose is the starting point of starch biosynthesis, which is converted into hexose-phosphate sugars in the cytoplasm. In potato, glucose-6- phosphate (Glc-6P) is transported into the amyloplast (Kammerer et al. 1998). It is first converted to glucose-1-phosphate (Glc-1P) by plastidial phosphoglucomutase (PGM) (E.C.5.4.2.2) and, subsequently, to ADP-glucose (ADP-Glc) by ADP-glucose pyrophosphorylase (AGPase, E.C.2.7.7.27) (Tauberger et al . 2000; Müller-Röber et al . 1992; Stark et al . 1992). ADP-Glc serves as a substrate for the different starch synthase isoforms, some of which are mainly present in the soluble phase or stroma (SS-isoforms), while others are associated with the granule. The exclusively granule-

12 : Chapter 1/ General introduction

Figure 2. Schematic representation of starch biosynthesis in a potato tuber cell. A: sucrose synthase (E.C.2.4.1.13); B: UDP-Glc pyrophosphorylase (E.C.2.7.7.9); C: phosphoglucomutase; D: Glc-6P isomerase (E.C.5.3.1.9); E: hexose kinase (E.C.2.7.1.1); F: plastidial phosphoglucomutase; G: ADP-Glc pyrophosphorylase; H: many enzymes such as starch synthase, branching enzymes, debranching enzymes, etc. ; I: non-native glucansucrases. Suc = sucrose, Glc = glucose; Frc = fructose. : translocators; 1: putative sucrose transporter; 2: G6PT (Glc-6P/Pi antiporter) (Kammerer et al. 1998); 3: putative hexose transporter related to the GLUT family (Weber et al . 2000); 4: putative sucrose transporter. : transport of solutes; : reaction catalyzed by enzymes. Dotted arrows indicate putative reactions in transformed tuber cells after introduction of a bacterial glucansucrase. bound starch synthase (GBSSI) catalyzes the formation of amylose (Kuipers et al. 1994). Furthermore, it was shown that GBSSI can also contribute to amylopectin synthesis in potato, pea and the unicellular alga, Chlamydomonas reinhardtii (Baba et al. 1987; Denyer et al. 1996; van de Wal et al . 1998). All starch synthases elongate glucan chains by transferring the glucose moiety of ADP-Glc to the non-reducing end of

α-1,4 linked glucans. The branching enzymes (BE) cleave α-1,4 linkages and form α- 1,6 linkages. Schwall et al. (2000) showed that antisensing these enzymes in potato led to less branching of the forth-coming starch. Additional enzymes such as the disproportionating enzymes (D-enzyme) cleave and rejoin α-1,4 linkages in starch

13 : Chapter 1/ General introduction polymers and the debranching enzymes (DBE) hydrolyze the α-1,6 linkage at branch points.

Figure 3. Schematic representation of the starch biosynthesis pathway. The timing of the biosynthesis process is as follows: the darker the background colour, the later the event takes place. The role of the various enzymes is described in detai l in the text.

The starch biosynthesis process is summarized schematically in Fig. 3. The model shows clearly that amylopectin is synthesized first, and that amylose is formed later. Two possible mechanisms for amylose biosynthesis have been proposed: (i) the amylopectin-primed pathway, and (ii) the malto-oligosaccharides (MOS)-primed pathway (Mouille et al . 1996; Ball et al . 1998; van de Wal et al. 1998, 2000; Denyer et al. 1999; 2001). First evidence for the first pathway was provided by van de Wal et al . (1998), who showed that amylose could be synthesized in vitro by cleavage of amylopectin molecules in mutant Chlamydomonas starches. It is postulated that amylopectin is cleaved by the action of hydrolytic enzymes as α-amylase (E.C.3.2.1.1). The Arabidopsis genome sequence indicates that these enzymes may be granule- bound, because they can be equipped with a specialized region for attachment to starch granules (Coutinho and Henrissat 1999). An alternative explanation is that GBBSI has a dual activity, initially working as a polymerase, and at a particular time point as a hydrolase. Also in higher plants, the amylopectin-primed pathway seems to take place. However, in this case it is more difficult to demonstrate amylopectin- priming, due to the much lower GBSSI activity in the granules (van de Wal 2000). For the second pathway, Denyer et al . (1999) showed that amylose could be synthesized in vitro in pea granules by the processive elongation of small soluble glucans or MOS, which can diffuse into the granule. It is estimated that the MOS concentration is

14 : Chapter 1/ General introduction sufficiently high in starch-producing organs of plants, thereby providing enough acceptor substrate for amylose synthesis (Denyer et al. 2001). The MOS pool may be replenished continuously by the action of DBE and other hydrolytic enzymes. Van de Wal et al. (2000) has suggested that the two pathways can occur side-by-side. Depending on the conditions in the plant, amylose biosynthesis may switch between the two mechanisms.

Genetic modification of potato starches by using bacterial enzymes Plants are efficient production systems for heterologous proteins that lead to high-value products (Kusnadi et al. 1997; Hood and Jilka 1999). In our research, the crop plant potato has been chosen for tailoring novel starch-based polymers, because its starch granules contain only small amounts of contaminants (lipids, proteins), which is advantageous for the quality of its starch (Ellis et al . 1998). Moreover, potato transformation is a routine procedure (Visser, 1991b). Over the years, several studies have been performed aimed at turning the amyloplast into a more versatile polysaccharide factory. For this purpose, several microbial enzymes have been equipped with a plastidial targeting signal, and their influence on starch structure and functionality has been investigated (Shewmaker and Stalker 1992). Starch polymer modification was achieved for instance by targeting the Escherichia coli glycogen synthase (GLGA) and the glycogen branching enzyme (GLGB) to the potato amyloplast (Shewmaker et al. 1994; Kortstee et al. 1996). In both cases, the natural balance of chain elongation and branching was disturbed, resulting in starch granules with altered physical properties, and with more heavily branched polymers. Attachment of novel glycosyl residues to starch polymers has also been an objective. For this purpose, a Bacillus subtilis levansucrase (E.C.2.4.1.10) was introduced in potato tuber amyloplasts (Gerrits et al. 2001). Levansucrase can polymerize the fructose moiety of the donor substrate sucrose into a high molecular weight fructan. This study showed that production of fructans inside the amyloplast is possible, but that starch yield is severely compromised and that starch morphology is dramatically altered. A covalent linkage between fructosyl residues and starch polymers still needs to be demonstrated. Finally, it has been attempted to convert starch in planta into high-value cyclic oligosaccharides, which can accommodate hydrophobic substances in their apolar cavity. These cyclodextrins can be used in various food and pharmaceutical

15 : Chapter 1/ General introduction applications. A cyclodextrin glycosyltransferase (CGTase; E.C.2.4.1.19) from Klebsiella pneumoniae was introduced into potato amyloplasts (Oakes et al. 1991) for cyclodextrin production. Only 0.01% of the endogenous starch was converted to the desired product, and this product was difficult to recover from the plant material. These examples demonstrate that bacterial enzymes can be powerful tools for starch modification, but that their performance in the plant is difficult to predict beforehand. In the rest of this chapter, we will focus on two emerging generic technologies, which are currently developed in our laboratory: (i) introduction of new linkage types in starch polymers using glucansucrase genes, and (ii) tailoring granule-boundness by fusion of effector proteins to microbial starch-binding domains.

Tools for introducing new linkages in starch polymers

In principle, starch consists of polymers with little structural variation: 95% of α-1,4- linked and only approximately 5% of α-1,4,6-linked glucose residues. Certain bacteria possess an array of enzymes, so-called glucansucrases, which can attach (contiguous) 1,6-linked or 1,3-linked glucosyl residues to maltodextrins. This, together with the observation that sucrose is present inside the potato tuber amyloplast (Gerrits et al. 2001), suggests that glucansucrases are of great interest for diversifying starch structure. With few exceptions, glucansucrases are extracellular enzymes, which are produced by lactic acid bacteria such as Leuconostoc mesenteroides strains, oral Streptococci , and some species of Lactobacillus and Lactococcus (Robyt 1995; van Geel-Schutten et al . 2000) . In addition, they are produced by other bacteria such as some of the Neisseria strains (Hehre et al. 1949). These strains are involved in different processes in nature. Some of the strains colonize the oral cavity of humans and animals and can induce the formation of dental caries. Other strains can invade the throat such as the commensal Neisseria species. Some Lactobacillus species increase the viscosity of fermented milk (de Vuyst and Degeest 1999). The glucansucrases catalyse the polymerization of glucose residues from sucrose, which leads to the production of a large variety of α-glucans with different sizes and structures, and composed of diverse linkage types (Fig. 4). Some strains are able to produce more than one kind of polysaccharide. The Leuconostoc strains B-512F and B-1299 secrete a dextransucrase (DSR) (E.C.2.4.1.5) that synthesizes a water-soluble

16 : Chapter 1/ General introduction

Figure 4. Structural diversity of polysaccharides produced by dextran-, mutan- and alternansucrases. A schematic representation of the structure of three other αœglucans (glycogen, amylose, and amylopectin) is also shown.

polymer composed of mainly α-1,6 glucosidic bonds with few α-1,3 and α-1,2 branch points. The Leuconostoc strain B-1355 possesses an alternansucrase (ASR)

(E.C.2.4.1.140) that synthesizes a polysaccharide composed of alternating α-1,3 and α-1,6 glucosidic bonds. The glucansucrases of Streptococci are often referred to as glucosyltransferases (GTFs), including both dextransucrases and mutansucrases

(E.C.2.4.1.-). Mutans are water-insoluble polymers composed of mainly α-1,3 glucosidic bonds with few α-1,6 branch points. As shown in Fig. 4, the kind of linkage types in amylose, amylopectin and glycogen are different from those in alternan, dextran and mutan. The dextran-, mutan- and alternansucrases belong to family 70 of the glycosyl hydrolases and have a molecular weight between 155 and 200 kDa (Remaud-Simeon et al. 2000). Due to the large size of these enzymes, three- dimensional structures are not yet available. Most glucansucrases share a common structure composed of four different regions: a signal peptide, a variable region, a catalytic domain, and a glucan-binding domain (GBD). The modular structure of glucansucrases is summarized in Fig. 5. Amylosucrase (AS; E.C.2.4.1.4) is an exception to this, and will be discussed at the end of this section. Further, it can be seen that the sucrases can differ considerably in size, and that they can contain specific elements, particularly in their GBD. The signal peptide consists of 35-38 amino acids and is responsible for secretion of the sucrases. DSR-A, which is produced by L.

17 : Chapter 1/ General introduction mesenteroides NRRL B-1299, is the only known sucrase without a signal peptide, suggesting that it is located intracellularly (Monchois et al. 1996). The signal peptide is followed by a variable region of 140-261 amino acids. Janecek et al . (2000) showed that conserved, long repeat elements (A-like repeats) are present in the downstream part of this region of dextransucrases and alternansucrase of Leuconostoc mesenteroides (NRRL B-512F (DSRS), NRRL B-1299 (DSR-B) and NRRL B-1355 (ASR)). It has been hypothesized that these repeats can play a role in glucan binding because they share a high homology with A repeats that are present in the GBD. This region does not seem to be essential for enzyme activity because DSR-A does not possess this region and is still active.

Figure 5. Comparison of the domain structure of various glucansucrases. GtfI is a mutansucrase from Streptococcus downeii Mfe28; DsrS is a dextransucrase from Leuconostoc mesenteroides NRRL B-512F; Asr is the alternansucrase from Leuconostoc mesenteroides NRRL B-1355; AS is the amylosucrase from Neisseria polysaccharea ATCC 43768. The form er three sucrases are classified as GH70 enzymes, the latter belongs to enzyme family GH13, and shows high structural similarity to α-amylases. Repeats are shown with their consensus sequences (boldface letter type indicates the highly conserved amino acid residues).

The catalytic domain is composed of about 900 amino acids and is highly conserved within the Leuconostoc and Streptococcus species. This domain is related to those of the α-amylase family 13, consisting of a ( β/α)8 barrel except that it is circularly

18 : Chapter 1/ General introduction

Figure 6. Overview of reactions catalyzed by starch synthases and glucansucrases. A: Chain elongation by starch synthases proceeds by addition of glucosyl residues to the non-reducing end of the nascent chain. B: Most glucansucrases (mutan-, dextran-, and alternansucrase) probably extend the growing chain from the reducing terminus. The observation that amylosucrase can elongate side chains of glycogen suggests that this enzyme prefers to add glucosyl residues to the non-reducing end of a polymer. C: Acceptor reactions are postulated to be responsible for chain termination and/or branching of short chains. : Sucrose, : Glucose, : Fructose, : Acceptor molecules such as glucan chain, maltodextrins…, : Reducing end, : Sucrose-binding site, : Nucleophile, : Acceptor binding site.

permuted (MacGregor et al . 1996). This means that the barrel of the enzymes from family 70 starts with an α-helix, whereas those of family 13 start with a β-strand. The catalytic domain is also called the sucrose-binding domain because it contains a catalytic triad of aspartic and glutamic acid residues that play an important role in binding and cleavage of sucrose molecules. Site-directed mutagenesis of these residues, as shown for instance in the following sequence IRV D551 AVD (present in the

β4 strand of DSRS), can drastically reduce the enzyme activity (Monchois et al . 1997). Besides this catalytic triad, other conserved residues have important functions such as the interaction with acceptor molecules (Monchois et al. 2000a; 2000b). The elongation of glucan chains by glucansucrases is different compared to that by starch synthases (Fig. 6A). First, the preferred substrate is sucrose instead of ADP-Glc. Second, the glucose residues are added to the reducing end of a growing glucan chain by a so- called two-site insertion mechanism (Robyt 1995). As illustrated in Fig. 6B, the enzyme

19 : Chapter 1/ General introduction possesses two sucrose-binding sites with two nucleophiles (two aspartic acids) that can attack two sucrose molecules. In this way, two glucosyl residues are bound covalently to the enzyme thereby forming two glucosyl-enzyme intermediates. The C-3 (mutansucrase) or the C-6 (dextransucrase) hydroxyl group of one of these intermediates attacks the C-1 position of the other glucosyl-enzyme intermediate thereby forming an α-1,3 or an α-1,6 linkage. The free aspartic acid residue attacks another sucrose molecule forming a new glucosyl-enzyme intermediate. Subsequently, the C-3 or C-6 hydroxyl group of this new glucosyl-enzyme intermediate attacks the reducing end C-1 of the previously formed short glucan chain. This process is repeated many times, thereby elongating the α-glucan chain processively. The two-site insertion mechanism has recently been challenged by detailed characterization of the properties of particular mutant glucansucrases. For a more detailed discussion of this we refer to Monchois et al. (1997) and Monsan et al. (2001). To our knowledge, very few reports deal with invertase activity of glucansucrases, which is in contrast to levansucrases and amylosucrases (see below). This suggests that hydrolysis of sucrose might be of minor importance for glucansucrases, which makes them rather efficient polymerases. The branching of mutans, dextrans and alternan does not take place by means of a branching enzyme as in starch biosynthesis, but by a so-called acceptor reaction catalyzed by the glucansucrases themselves (Robyt, 1995). The glucansucrase is thought to contain an acceptor-binding site that can bind acceptor molecules such as the nascent glucan chains or maltodextrins (Su and Robyt, 1994). As illustrated in Fig. 6C, the hydroxyl group of the acceptor molecule can make a nucleophilic attack on the glucosyl-enzyme intermediate thereby attaching the glucose residue to the acceptor molecule. By this reaction either a single-unit side chain or a longer glucan side chain can be created. When maltodextrins are used as acceptor molecules, the glucosyl residue can be added to either the reducing or the non-reducing end. In the latter case, the side chain can be further extended (Fu and Robyt 1990, 1991). The acceptor reaction seems very advantageous for introducing novel structural elements in starch polymers, because it offers the possibility of covalent attachment of novel side chains to native structures. Therefore, glucansucrases which catalyze acceptor reactions with a high efficiency need to be selected. This is not an easy task, since the structure- function relationships underlying the acceptor reaction are not understood. In addition, the efficiency to catalyze acceptor reactions with starch polymers or maltodextrins is

20 : Chapter 1/ General introduction

Figure 7. The relationship between the relative acceptor substrate effic iency and the degree of polymerization of the maltodextrin acceptor substrates for three different glucansucrases (Fu and Robyt 1990, 1991).

poorly documented. It is important to realize that the acceptor reaction efficiency depends on the size of the acceptor molecules (Fu and Robyt 1990, 1991; Robyt 1995). This is illustrated in Fig. 7 for three glucansucrases with a range of maltodextrin acceptors. It is obvious that the glucansucrases show a different behaviour in acceptor reactions. The relative acceptor efficiency of the L. mesenteroides B512FM dextransucrase decreases with increasing chain length of the maltodextrin, whereas the relative acceptor efficiency seems to reach a constant value for S. mutans 6715 GTF-I. Also, maltotriose is a much better acceptor molecule for S. mutans 6715 GTF-I than for the other two enzymes. In the plant, amylopectin and amylose may be the acceptor molecules for the glucansucrases. However, it is uncertain whether they will participate directly in acceptor reactions because of their high molecular weight. It seems more likely that maltodextrins generated by the action of for instance DBE serve as acceptor molecules. The forth-coming branched maltodextrins might subsequently be incorporated into larger structures by the action of transferases such as BE and D- enzyme. The C-terminal part of the enzyme consists of the glucan-binding domain (GBD) covering about 500 amino acids. It is composed of repeats named A, B, C, D that are

21 : Chapter 1/ General introduction defined by a consensus sequence (Monchois et al. 1998; 1999). The N repeats are not very well conserved and novel short repeats specific for ASR have not been assigned a name (Monchois et al. 1998; Janecek et al . 2000). Giffard and Jacques (1994) showed that common YG elements were present in each repeat. From Fig. 5, it can be seen that the number and organization of these repeats is variable within the glucansucrases shown. In particular, ASR presents distinctive short repeats that are not present in the other glucansucrases. Several research groups have investigated the biochemical properties of engineered glucansucrases that lack (part of) the GBD (Monchois et al . 1998, 1999) to unravel the exact function of the GBD. To this end, these studies have not been conclusive, but a number of observations are worth mentioning. Some reports suggest that the GBD influences the structure of the glucan produced, but this needs to be investigated in more detail (Monchois et al . 1999, and references therein). The GBD of L. mesenteroides DSR-S was shown to modulate the initial velocity of dextran synthesis; truncation of the enzyme resulted in a reduced reaction rate. It seems that enzymes producing soluble glucans (like DSR-S) are more sensitive to C-terminal truncation than those producing insoluble glucans (Kato and Kuramitsu 1990; Monchois et al . 1998, 1999). Further, the GBD is probably not directly involved in mediating the actual acceptor reaction, but it has been observed that the more extensive truncations in the GBD region result in the production of larger oligosaccharides in the acceptor reaction (Monchois et al . 1998). Finally, it has been shown that the GBD is one (but not the only) factor determining whether or not the glucansucrase needs a primer to initiate polymer synthesis (Nakano and Kuramitsu, 1992). These observations have led Monchois et al. (1998) to propose that the GBD may facilitate the translation of the nascent glucans from the catalytic site in order to maintain a high catalytic efficiency in polymer synthesis. The Neisseria strains contain a glucansucrase with a different modular arrangment as discussed above: amylosucrase (AS). AS belongs to the α-amylase family (glycosyl hydrolase family 13), and catalyzes the formation of an amylose-like glucan (Büttcher et al. 1997; Coutinho and Henrissat 1999; Potocki de Montalk et al . 1999; Sarçabal et al. 2000). This enzyme has a molecular weight of about 72 kDa, much lower than that of the other sucrases. The three-dimensional structure of AS was recently elucidated, exhibiting five distinct domains (N, A, B, B‘ and C) (Mirza et al . 2001; Skov et al. 2001). Domain A represents the catalytic core, and contains the catalytic triad of Asp-Glu-Asp

22 : Chapter 1/ General introduction residues, also found in the other glucansucrases (Fig. 5). It seems likely that AS has a different mechanism of action, and extends glucan chains from the non-reducing end (Potocki de Montalk et al. 1999). In principle, this reaction is similar to that catalyzed by starch synthases (see Fig. 6A) with the exception that sucrose instead of ADP-Glc is the donor substrate. AS also catalyzes the hydrolysis of sucrose to glucose and fructose. This reaction is effectively blocked with a glycogen acceptor (Potocki de Montalk et al. 2000).

S. mutans gtfB (MT 4467) S. mutans gtfB (GS5) * S. mutans gtfB (MT 4239) I S. mutans gtfB (MT 4245) S. mutans gtfB (MT 4251) S. mutans gtfB (MT 8148) S. mutans gtfC (MT 4467) S. mutans gtfC (MT 8148) S. mutans gtfC (MT 4239) S. mutans gtfC (MT 4251) II S. mutans gtfC (MT 4245) S. mutans gtfC (GS5) * S. sobrinus gtfI (OMZ176) S. sobrinus gtfI (6715) S. sobrinus gtfI (ATCC 33478) III S. downeii gtfI (Mfe 28) * S. mutans gtfD (MT 4251) S. mutans gtfD (MT 4245) S. mutans gtfD (MT 4239) S. mutans gtfD (MT 8148) IV S. mutans gtfD (MT 4467) S. mutans gtfD (GS5) * S. gordonii gtfG (Challis) S. sal. gtfL (ATCC 25975) * S. sal. gtfN (V1477) V S. sal. gtfM (ATCC 25975) S. sal. gtfK (ATCC 25975) * S. sobrinus gtfT (OMZ176) * S. sal. gtfJ (ATCC 25975) VI S. downeii gtfS (Mfe 28) * L. mes. Dsr T (NRRL B-512F) L. mes. Dsr A (NRRL B-1299) * VII L. mes. Dsr B (NRRL B-1299) * L. mes. Dsr S (NRRL B-512F) * L. mes. Asr (NRRL B-1355) * VIII 101.7 IX 100 80 60 40 20 0

Figure 8. Phylogenetic tree of glucansucrases from Leuconostoc (L) and Streptococcus (S) bacterial strains, constructed using the complete amino ac id sequences that were available on June 2000. Each of the sequences were aligned with the Clustal method. * Bacteria selected for this study. sal = salivarius , mes = mesenteroides .

Considering the diversity of α-glucan structures produced with glucansucrases, we think that these enzymes could potentially modify the fine structure of starch. Most likely, the ability of the sucrases to catalyze acceptor reactions is a key factor for introducing new structural elements in starch. At the start of the work described in this thesis, selection of suitable candidate glucansucrases was difficult because not all of them were fully characterized, in particular their acceptor reaction efficiencies. In addition, their nomenclature was not consistent over the years. From the thirty-five different glucansucrases available at http://afmb.cnrs-mrs.fr/CAZY/, belonging to the glycoside hydrolase family 70, a phylogenetic tree was made by aligning all the known amino acid sequences (see Fig. 8), which grouped in nine different clusters. We

23 : Chapter 1/ General introduction selected glucansucrases which could give us a large diversity in the alpha (1R6)/ alpha (1R3) ratio, ranging from 100 % (1R6)-linked glucosyl residues (GTFK) to 12 % (1R6)/ 88 % (1R3)-linked glucosyl residues (GTFI) (See Table 1). The acceptor reaction of two of them (DSRS and ASR) had been characterized.

Table 1. Overview of selected glucansucrases used in this study. ND: not defined.

Strain Gene Size (aa) Glucan Acceptor Reference produced reaction efficiency S. salivarius GtfK 1599 100 % S(1> 6) ND Giffard et al ., ATCC 25975 1993 L. mesenteroides DsrS 1527 95 % S(1> 6) + Wilke- NRRL B-512F 5 % S(1> 3) Douglas et al ., 1989 S. sobrinus GtfT 1542 73 % S(1> 6) ND Hanada et al ., OMZ176 27 % S(1> 3) 1993 L. mesenteroides Asr 2057 54 % S(1> 6) + Argüello- NRRL B-1355 46 % S(1> 3) Morales et al ., 2000 S. downei GtfI 1556 12 % S(1> 6) ND Ferretti et al ., Mfe28 88 % S(1> 3) 1987

Tools for anchoring proteins inside starch granules It has been observed by many investigators that the biosynthetic enzymes are partitioned between the soluble phase or stroma and the growing crystalline granule during the biosynthesis process. Three types of enzymes can be distinguished: (i) exclusively soluble enzymes (S ENZ ), (ii) partially granule-bound enzymes (gb ENZ ), and

(iii) exclusively granule-bound enzymes (GB ENZ ). The gb ENZ s are sometimes referred to as dual-location enzymes (Cao et al. 1999). To this end, it is unknown which factors determine granule-boundness. The ability to bind to starch granules is not restricted to its biosynthetic machinery. Also, microbial starch-degrading enzymes can attach to starch granules. A large number of these enzymes have been studied extensively, and they are often composed of two or more domains (Svensson et al . 1989). Attachment of these enzymes to (raw) starch granules is mediated by a so-called starch-binding domain (SBD) or domain E. Without a SBD, these enzymes show little activity on insoluble substrates (such as the crystalline granule), whereas the activity on soluble substrates is not affected so much (Southall et

24 : Chapter 1/ Genera l introduction al. 1999). It has also been suggested that SBD plays a more active (but non-hydrolytic) role in starch degradation by disrupting the granule surface (Southall et al . 1999). Besides their affinity for starch granules, SBDs can bind maltodextrins and cyclodextrins (Svensson et al . 1989). The SBD can be present on the N-terminus or on the C-terminus of a protein and can be embedded in a multi-domain structure as in the CGTase from Bacillus circulans (Lawson et al . 1994), or it can be spatially separated from the catalytic domain by a heavily glycosylated linker peptide (the glucoamylases of Aspergillus niger and Rhizopus oryzae ). Compared to the catalytic domains (> 200 amino acids), the SBDs are usually relatively small (approximately 100 amino acids) (Svensson et al . 1989; Janecek and Łevcik 1999). From amino acid sequence alignments, it was demonstrated that a number of aromatic amino acids (Trp and Tyr) are highly conserved. Some of these play a pivotal role in binding carbohydrate structures (Penninga et al . 1996; Williamson et al . 1997). Several studies show that SBDs retain their affinity for starch, even when they are separated from the other constituent domains of the natural proteins (Belshaw and Williamson 1993; Dalmia et al . 1995; Williamson et al . 1997). Further, it has been shown that enzymes that normally do not bind to starch can acquire affinity for starch granules by fusion of an SBD to their catalytic domain (Dalmia et al . 1995; Ohdan et al. 2000). These results suggest that SBDs can be used as universal tools to target an attached biosynthetic enzyme to starch granules during the biosynthesis process. In our laboratory, we have investigated whether the SBD of CGTase from Bacillus circulans can be used for this purpose. The separate SBD (thus without a protein of interest attached) was fused to the transit peptide of potato GBSSI to allow transport of the SBD thru the amyloplast membranes (Ji et al . 2003; 2004). This artificial gene construct, driven by a tuber-specific potato GBSSI promoter, was introduced in potato plants by Agrobacterium -mediated transformation. The experiment showed that SBD proteins could be incorporated in the starch granules of the transgenic plants without compromising the physicochemical properties of the starch. In another experiment the reporter gene luciferase was inserted in frame between the sequences encoding the transit peptide and the SBD of the construct mentioned above. After transformation of potato plants, this construct yielded transgenic starch granules with luciferase activity. These results show that SBD can target effector proteins to the starch granule during the biosynthesis process, and that the effector retains its activity in the fusion protein

25: Chapter 1/ General introduction and in the crystalline starch matrix. In this thesis research, the SBD technology will be used as a tool to target glucansucrases, in particular the mutansucrase to the growing starch granule. In this way, it might be possible to bring the mutansucrase in more intimate contact with starch granules, possibly facilitating the acceptor reaction and the covalent attachment of mutan chains to starch.

Objectives and outline of this thesis Objectives of this research were to produce S-glucan polymers with novel linkages types in potato amyloplasts by overexpressing glucansucrase genes from lactic acid bacteria. At the beginning of this work, only enzymes synthesizing S-(1R3), S-(1R6) and alternating S-(1R3)/ S-(1R6) linked glucosyl residues were known and partially characterized. At that time, enzymes synthesizing S-(1R2) linked glucosyl residues were unavailable from databases; such enzymes were also of great interest for our purposes. Enzymes synthesizing S-(1R4) linked glucosyl residues were not considered because these produce polymers with linkage similarity to starch. During this thesis research, different questions were addressed: (i) is it possible to produce dextran, mutan and alternan in potato amyloplasts ?; (ii) can the amylopectin and amylose structures be modified in planta by incorporating novel linkage types, by heterologous expression of such enzymes ?; (iii) what are their subsequent impacts on starch biosynthesis, in general ?; (iv) is it possible to engineer sucrases to enhance polymer accumulation ?; (v) can the incorporation of glucan polymers in starch be facilitated, using the starch-binding domain (SBD) technology ?. Expression of the DsrS gene from Leuconostoc mesenteroides NRRL B-512F in cultivar Kardal as well as the amylose-free potato mutant led to the production of dextran, defined as S-(1R6) linked glucosyl residues, in potato juices as described in Chapter 2. Chapter 3 describes the production of mutan, S-(1R3) linked glucosyl residues, by expressing the GtfICAT gene from Streptococcus downei (truncated GtfI without glucan-binding domain), which was more efficient than the full-length GtfI . Mutan production was demonstrated by staining the starch granules with an erythrosine dye. In addition, mutan material was detached from starch granules after exo-mutanase treatment. In Chapter 4, presence of alternan, defined as alternating S-(1R3)/ S-(1R6) linked glucosyl residues, was evidenced in cv Kardal potato juices. In Chapter 5, the GtfICAT gene from which it is known that its expression led to an efficient mutan production (see Chapter 3) was fused

26: Chapter 1/ General introduction to SBD in order to produce mutan inside starch granules. Effects of the SBD position (N- or C-terminal) relating to starch modification were also studied. Implications of our research are discussed in Chapter 6.

References

gene for amylosucrase from Neisseria polysaccharea : production of a linear α-1,4- glucan. J. Bacteriol. 179, 3324œ3330. Cao, H., Imparl-Radosevich, J.M., Guan, H., Keeling, P.L., James, M.G., Myers, A.M., 1999. Identification of soluble starch synthase activities of maize endosperm. Plant Physiol. 120, 205œ215. Coutinho, P.M., Henrissat, B., 1999. Carbohydrate-active Enzymes server at URL: http://afmb.cnrs-mrs.fr/ıpedro/CAZY/db.html. Dalmia, B.K., Schütte, K., Nikolov, Z.L., 1995. Domain E of Bacillus macerans cyclodextrin glucanotransferase: An independent starch-binding domain. Biotechnol. Bioeng. 47, 575œ 584. Denyer, K., Clarke, B., Hylton, C., Tatge, H., Smith, A.M., 1996. The elongation of amylose and amylopectin chains in isolated starch granules. Plant J. 10, 1135œ1143. Denyer, K., Waite, D., Motawia, S., Moller, B.L., Smith, A.M., 1999. Granule-bound starch synthase I in isolated starch granules elongates malto-oligosaccharides processively. Biochem. J. 340, 183œ191. Denyer, K., Johnson, P., Zeeman, S., Smith, A.M., 2001. The control of amylose synthesis. J. Plant. Physiol. 158, 479œ487.

27: Chapter 1/ General introduction de Vuyst, L., Degeest, B., 1999. Heteropolysaccharides from lactic acid bacteria. FEMS Microbiol. Rev. 23, 153œ177. Ellis, R.P., Cochrane, M.P., Dale, M.F.B., Duffus, C.M., Lynn, A., Morrison, I.M., Prentice, R.D.M., Swanston, J.S., Tiller, S.A., 1998. Starch production and industrial use. J. Sci. Food Agric. 77, 289œ311. Ferretti, J.J., Gilpin, M.L., Russell, R.R.B., 1987. Nucleotide sequence of a glucosyltransferase gene from Streptococcus sobrinus Mfe28. J. Bacteriol. 169, 4271œ4278. Fu, D., Robyt, J.F., 1990. Acceptor reactions of maltodextrins with Leuconostoc mesenteroides B-512FM dextransucrase. Arch. Biochem. Biophys. 283, 379œ387. Fu, D., Robyt, J.F., 1991. Maltodextrin acceptor reactions of Streptococcus mutans 6715 glucosyltransferases. Carbohydr. Res. 217, 201œ211. Gerrits, N., Turk, S.C.H.J., van Dun, K., Hulleman, S.H.D., Visser, R.G.F., Weisbeek, P.J., Smeekens, S.C.M., 2001. Sucrose metabolism in plastids. Plant Physiol. 125, 926œ934. Giffard, P.M., Allen, D.M., Milward, C.P., Simpson, C.L., Jacques, N.A., 1993. Sequence of the gtfK gene of Streptococcus salivarius ATCC 25975 and evolution of the gtf genes of oral streptococci. J. Gen. Microbiol. 139, 1511œ1522. Giffard, P.M., Jacques, N.A., 1994. Definition of a fundamental repeating unit in streptococcal glucosyltransferase glucan-binding regions and related sequences. J. Dent. Res. 73, 1133œ1141. Hanada, N., Isobe, Y., Aizawa, Y., Katayama, T., Sato, S., Inoue, M., 1993. Nucleotide sequence analysis of the gtfT gene from Streptococcus sobrinus OMZ176. Infect. Immun. 61, 2096œ2103. Hehre, E.J., Hamilton, D.M., Carlson, A.S., 1949. Synthesis of a polysaccharide of the starch- glycogen class from sucrose by a cell-free bacterial enzyme system (amylosucrase). J. Biol. Chem. 177, 267œ279. Hood, E.E., Jilka, J.M., 1999. Plant-based production of xenogenic proteins. Curr. Opin. Biotechnol. 10, 382œ386. Janecek, S., Łevcik, J., 1999. The evolution of starch-binding domain. FEBS Lett. 456, 119œ125. Janecek, S., Svensson, B., Russell, R.R.B., 2000. Location of repeat elements in glucansucrases of Leuconostoc and Streptococcus species. FEMS Microbiol. Lett. 192, 53œ57. Ji, Q., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2003. Microbial starch-binding domains as a tool for targeting proteins to granules during starch biosynthesis. Plant Mol. Biol. 51, 789œ801. Ji, Q., 2004. Microbial starch-binding domains as a tool for modifying starch biosynthesis. Ph.D. Dissertation, Wageningen University, The Netherlands, ISBN 90-8504-022-1.

28: Chapter 1/ General introduction

Kato, C., Kuramitsu, H.K., 1990. Carboxy-terminal deletion analysis of the Streptococcus mutans glucosyltransferase-I enzyme. FEMS Microbiol. Lett. 72, 299œ302. Kammerer, B., Fischer, K., Hilpert, B., Schubert, S., Gutensohn, M., Weber, A., Flügge, U., 1998. Molecular characterization of a carbon transporter in plastids from heterotrophic tissues: the glucose 6-phosphate/phosphate antiporter. Plant Cell 10, 105œ118. Kortstee, A.J., Vermeesch, A.M.G., de Vries, B.J., Jacobsen, E., Visser, R.G.F., 1996. Expression of Escherichia coli branching enzyme in tubers of amylose-free transgenic potato leads to an increased branching degree of the amylopectin. Plant J. 10, 83œ90. Kossmann, J., Lloyd, J., 2000. Understanding and influencing starch biochemistry. Critical Rev. Plant Sci. 19, 171œ226. Kuipers, A.G.J., Jacobsen, E., Visser, R.G.F., 1994. Formation and deposition of amylose in the potato tuber are affected by the reduction of granule-bound starch synthase gene expression. Plant Cell 6, 43œ52. Kusnadi, A.R., Nikolov, Z.L., Howard, J.A., 1997. Production of recombinant proteins in transgenic plants: Practical considerations. Biotechnol. Bioeng. 56, 473œ484. Lawson, C.L., van Montfort, R., Strokopytov, B., Rozeboom, H.J., Kalk, K.H., de Vries, G.E., Penninga, D., Dijkhuizen, L., Dijkstra, B.W., 1994. Nucleotide sequence and X-ray structure of cyclodextrin glycosyltransferase from Bacillus circulans strain 251 in a maltose-dependent crystal form. J. Mol. Biol. 236, 590œ600.

MacGregor, E.A., Jespersen, H.M., Svensson, B., 1996. A circularly permuted α-amylase-type α/β-barrel structure in glucan-synthesizing glucosyltransferases. FEBS lett. 378, 263œ 266. Mirza, O., Skov, L.K., Remaud-Siméon, M., Potocki de Montalk, G., Albenne, C., Monsan, P., Gajhede, M., 2001. Crystal structures of amylosucrase from Neisseria polysaccharea in complex with D-glucose and the active site mutant Glu328Gln in complex with the natural substrate sucrose. Biochemistry 40, 9032œ9039. Monchois, V., Willemot, R-M., Remaud-Siméon, M., Croux, C., Monsan, P., 1996. Cloning and sequencing of a gene coding for a novel dextransucrase from Leuconostoc

mesenteroides NRRL B-1299 synthesizing only α (1-6) and α (1-3) linkages. Gene 182, 23œ32. Monchois, V., Remaud-Siméon, M., Russell, R.R.B., Monsan, P., Willemot, R-M., 1997. Characterization of Leuconostoc mesenteroides NRRL B-512F dextransucrase (DSRS) and identification of amino-acid residues playing a key role in enzyme activity. Appl. Microbiol. Biotechnol. 48, 465œ472.

29: Chapter 1/ General introduction

Monchois, V., Reverte, A., Remaud-Siméon, M., Monsan, P., Willemot, R-M., 1998. Effect of Leuconostoc mesenteroides NRRL B-512F dextransucrase carboxy-terminal deletions on dextran and oligosaccharide synthesis. Appl. Environ. Microbiol. 64, 1644œ1649. Monchois, V., Willemot, R-M., Monsan, P., 1999. Glucansucrases: mechanism of action and structure-function relationships. FEMS Microbiol. Rev. 23, 131œ151. Monchois, V., Vignon, M., Russell, R.R.B., 2000a. Mutagenesis of Asp-569 of glucosyltransferase I glucansucrase modulates glucan and oligosaccharide synthesis. Appl. Environ. Microbiol. 66, 1923œ1927. Monchois, V., Vignon, M., Escalier, P-C., Svensson, B., Russell, R.R.B., 2000b. Involvement of Gln937 of Streptococcus downei GTF-I glucansucrase in transition-state stabilization. Eur. J. Biochem. 267, 4127œ4136. Monsan, P., Bozonnet, S., Albenne, C., Joucla, G., Willemot, R-M., Remaud-Siméon, M., 2001. Homopolysaccharides from lactic acid bacteria. Int. Dairy J. 11, 675œ685. Mouille, G., Maddelein, M-L., Libessart, N., Talaga, P., Decq, A., Delrue, B., Ball, S., 1996. Preamylopectin processing: a mandatory step for starch biosynthesis in plants. Plant Cell 8, 1353œ1366. Müller-Röber, B., Sonnewald, U., Willmitzer, L., 1992. Inhibition of the ADP-glucose pyrophosphorylase in transgenic potatoes leads to sugar-storing tubers and influences tuber formation and expression of tuber storage protein genes. EMBO J. 11, 1229œ 1238. Myers, A.M., Morell, M.K., James, M.G., Ball, S.G., 2000. Recent progress towards understanding biosynthesis of the amylopectin crystal. Plant Physiol. 122, 989œ997. Nakano, Y.J., Kuramitsu, H.K., 1992. Mechanism of Streptococcus mutans glucosyltransferases: hybrid-enzyme analysis. J. Bacteriol. 174, 5639œ5646. Oakes, J.V., Shewmaker, C.K., Stalker, D.M., 1991. Production of cyclodextrins, a novel carbohydrate, in the tubers of transgenic potato plants. Biotechnol. 9, 982œ986. Ohdan, K., Kuriki, T., Takata, H., Kaneko, H., Okada, S., 2000. Introduction of raw starch-binding

domains into Bacillus subtilis α-amylase by fusion with the starch-binding domain of Bacillus cyclomaltodextrin glucanotransferase. Appl. Environ. Microbiol. 7, 3058œ3064. Penninga, D., van der Veen, B., Knegtel, R.M.A., van Hijum, S.A.F.T., Rozeboom, H.J., Kalk, K.H., Dijkstra, B.W., Dijkhuizen, L., 1996. The raw starch-binding domain of cyclodextrin glycosyltransferase from Bacillus circulans strain 251. J. Biol. Chem. 271, 32777œ32784. Potocki de Montalk, G., Remaud-Simeon, M., Willemot, R-M., Planchot, V., Monsan, P., 1999. Sequence analysis of the gene encoding amylosucrase from Neisseria polysaccharea and characterization of the recombinant enzyme. J. Bacteriol. 181, 375œ381.

30: Chapter 1/ General introductio n

Potocki de Montalk, G., Remaud-Simeon, M., Willemot, R-M., Monsan, P., 2000. Characterisation of the activator effect of glycogen on amylosucrase from Neisseria polysaccharea . FEMS Microbiol. Lett. 186, 103œ108. Remaud-Simeon, M., Willemot, R-M., Sarçabal, P., Potocki de Montalk, G., Monsan, P., 2000. Glucansucrases: molecular engineering and oligosaccharide synthesis. J. Mol. Catal. 10, 117œ128. Robyt, J.F., 1995. Mechanisms in the glucansucrase synthesis of polysaccharides and oligosaccharides from sucrose. Arch. Carb. Chem. Biochem. 51, 133œ168. Sarçabal, P., Remaud-Simeon, M., Willemot, R-M., Potocki de Montalk, G., Svensson, B., Monsan, P., 2000. Identification of key amino acid residues in Neisseria polysaccharea amylosucrase. FEBS Lett. 474, 33œ37. Schwall, G.P., Safford, R., Westcott, R.J., Jeffcoat, R., Tayal, A., Shi, Y-C., Gidley, M.J., Jobling, S.A., 2000. Production of very-high-amylose potato starch by inhibition of SBE A and B. Nature Biotechnol. 18, 551œ554. Shewmaker, C.K., Stalker, D.M., 1992. Modifying starch biosynthesis with transgenes in potatoes. Plant Physiol. 100, 1083œ1086. Shewmaker, C.K., Boyer, C.D., Wiesenborn, D.P., Thompson, D.B., Boersig, M.R., Oakes, J.V., Stalker, D.M., 1994. Expression of Escherichia coli glycogen synthase in the tubers of transgenic potatoes ( Solanum tuberosum ) results in a highly branched starch. Plant Physiol. 104, 1159œ1166. Skov, L.K., Mirza, O., Henriksen, A., Potocki de Montalk, G., Remaud-Simeon, M., Sarçabal, P., Willemot, R-M., Monsan, P., Gajhede, M., 2001. Amylosucrase, a glucan-

synthesizing enzyme from the α-amylase family. J. Biol. Chem. 276, 25273œ25278. Slattery, C.J., Kavakli, I.H., Okita, T.W., 2000. Engineering starch for increased quantity and quality. Trends Plant Sci. 5, 291œ298. Smith, A.M., Denyer, K., Martin, C., 1997. The synthesis of the starch granule. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48, 67œ87. Southall, S.M., Simpson, P.J., Gilbert, H.J., Williamson, G., Williamson, M.P., 1999. The starch- binding domain from glucoamylase disrupts the structure of starch. FEBS Lett. 447, 58œ60. Stark, D.M., Timmerman, K.P., Barry, G.F., Preiss, J., Kishore, G.M., 1992. Regulation of the amount of starch in plant tissues by ADP glucose pyrophosphorylase. Science 258, 287œ 292. Su, D., Robyt, J.F., 1994. Determination of the number of sucrose and acceptor binding sites for Leuconostoc mesenteroides B-512FM dextransucrase, and the confirmation of the two-site mechanism for dextran synthesis. Arch. Biochem. Biophys. 308, 471œ476.

31: Chapter 1/ General introduction

Svensson, B., Jespersen, H., Sierks, M.R., MacGregor, E.A. 1989. Sequence homology between putative raw-starch binding domains from different starch-degrading enzymes. Biochem. J. 264, 309œ311. Tauberger, E., Fernie, A.R., Emmermann, M., Renz, A., Kossmann, J., Willmitzer, L., Trethewey, R.N., 2000. Antisense inhibition of plastidial phosphoglucomutase provides compelling evidence that potato tuber amyloplasts import carbon from the cytosol in the form of glucose-6-phosphate. Plant J. 23, 43œ53. Thompson, D.B., 2000. On the non-random nature of amylopectin branching. Carbohydr. Polymers 43, 223œ239. van Geel-Schutten, G.H., 2000. Exopolysaccharide synthesis by Lactobacillus reuteri : Molecular characterization of a fructosyltransferase and a glucansucrase. PhD Thesis, Groningen University, The Netherlands, ISBN 90-9013877-3. van de Wal, M.H.B.J., D‘Hulst, C., Vincken, J-P., Buléon, A., Visser, R.G.F., Ball, S., 1998. Amylose is synthesised in vitro by extension of and cleavage from amylopectin. J. Biol. Chem. 273, 22232œ22240. van de Wal, M.H.B.J., 2000. Amylose biosynthesis in potato: interaction between substrate availability and GBSSI activity, regulated at the allelic level. PhD Thesis, Wageningen University, The Netherlands, ISBN 90-5808-224-5. Visser, R.G.F., Somhorst, I., Kuipers, G.J., Ruys, N.J., Feenstra, W.J., Jacobsen, E., 1991a. Inhibition of expression of the gene for granule-bound starch synthase in potato by antisense constructs. Mol. Gen. Genet. 225, 289œ296. Visser, R.G.F., 1991b. Regeneration and transformation of potato by Agrobacterium tumefaciens . In: Lindsey K (ed) Plant culture manual B5. Kluwer Academic Publishers, Dordrecht pp 1œ9. Weber, A., Servaites, J.C., Geiger, D.R., Kofler, H., Hille, D., Gröner, F., Hebbeker, U., Flügge, U., 2000. Identification, purification and molecular cloning of a putative plastidic glucose translocator. Plant Cell 12, 787œ801. Wilke-Douglas, M., Perchorowicz, J.T., Houck, C.M., Thomas, B.R., 1989. Methods and compositions for altering physical characteristics of fruit and fruit products. PCT patent WO 89/12386. Williamson, M.P., Le Gal-Coëffet, M-F., Sorimachi, K., Furniss, C.S.M., Archer, D.B., Williamson, G., 1997. Function of conserved tryptophans in the Aspergillus niger glucoamylase 1 starch- binding domain. Biochem. 36, 7535œ7539.

33: Chapter 2/ Production of dextran

Production of dextran in transgenic potato plants

Géraldine A. Kok-Jacon 1 Denong Wang 2,3 Jean-Paul Vincken 1 Shaoyi Liu 3 Luc C.J.M. Suurs 1 Richard G.F. Visser 1

1Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Wageningen University, P.O. Box 386, 6700 AJ Wageningen, The Netherlands 2Departments of Genetics, Neurology and Neurological Sciences, Stanford University School of Medicine, Beckman Center B006, Stanford, CA 94305-5318, The United States. 3Functional Genomics Division, Columbia Genome Center, College of Physicians and Surgeons, Columbia University, 1150 St. Nicholas Avenue, New York, NY 10032, The United States.

Abstract: The production of dextran in potato tubers and its effect on starch biosynthesis were investigated. The mature dextransucrase ( DsrS ) gene from Leuconostoc mesenteroides was fused to the chloroplastic ferredoxin signal peptide (FD) enabling amyloplast entry, which was driven by the highly tuber-expressed patatin promoter. After transformation of two potato genotypes (cv. Kardal and the amylose-free ( amf ) mutant), dextrans were detected by enzyme- linked immunosorbent assay (ELISA) in tuber juices of Kardal and amf transformants. The dextran concentration appeared two times higher in the Kardal (about 1.7 mg/g FW) than in the amf transformants. No dextran was detected by ELISA inside the starch granule. Interestingly, starch granule morphology was affected, which might be explained by the accumulation of dextran in tuber juices. In spite of that, no significant changes of the physicochemical properties of the starches were detected. Furthermore, we have observed no clear changes in chain length distributions, despite the known high acceptor efficiency of DSRS.

Published in: Transgenic research, in press (2005)

Key words: dextran, glucansucrase, transgenic potato, granule morphology

34: Chapter 2/ Production of dextran

Introduction

The use of plants for the production of novel polymers is an emerging technology of great interest (Kok-Jacon et al ., 2003). Studies based on the accumulation of fructan (Gerrits, 2000) and silk (Scheller et al ., 2001) in potato plants have indicated that plants can be used as bioreactors for the production of foreign polymers. Furthermore, modification of native polymers such as starch might be possible. For instance, attachment of fructosyl residues to starch was also investigated by expressing a sucrose-converting enzyme, the SACB levansucrase, in potato amyloplasts. Attachment of fructan to starch was demonstrated in small amounts in in vitro experiments, and it was hypothesized that this might also occur in planta (Gerrits, 2000). Biosynthesis of dextrans is mediated by Lactobacillus , Leuconostoc , and Streptococcus bacteria in the presence of sucrose. Dextrans are extracellular S-glucans composed of S-(1R6)-linked glucosyl residues in the main chain, and branched by variable proportions of S-(1R2), S-(1R3) or S-(1R4) linked glucose, depending on their origin (Jeanes et al ., 1954; Sidebotham, 1975). Dextran produced by Leuconostoc mesenteroides NRRL B-512F is a commercial polymer, which has been fermented at a large industrial scale since 1948 (Groenwall and Ingelman, 1948). It is used in several industrial applications including chromatographic media, soil conditioner (Murphy and Whistler, 1973) and biodegradable hydrogels (Hennink and van Nostrum, 2002). In the latter case, dextran-based gels are used as a delivery system for the specific targeting of drugs to the colon, where they are degraded by a secreted bacterial dextranase. The L. mesenteroides NRRL B-512F dextran polymer is water-soluble, and consists of 95 % S-(1R6) linkages in the main chain, and 5 % S-(1R3)-linked single unit side chains (van Cleve et al ., 1956). Its biosynthesis is mediated by the dextransucrase DSRS (EC 2.4.1.5), which is a 1,527 amino-acid glucosyltransferase (Wilke-Douglas et al ., 1989). This enzyme exhibits a high efficiency in bond formation, which makes it a rather attractive polymerase. Its catalytic properties can be summarized as follows: after cleavage of sucrose, a glucosyl residue can be transferred to a growing dextran chain by the so-called two-site insertion mechanism, or to acceptor molecules (Robyt, 1995; Monchois et al ., 1999). Glucosylation of acceptor molecules such as maltose and isomaltose are the most effective. Interestingly, it was recently shown that the DSRS

35: Chapter 2/ Production of dextran glucosylation reaction can also be used for the synthesis of new compounds such as oligosaccharide and surfactant derivatives, giving access to novel industrial applications (Demuth et al ., 2002; Richard et al ., 2003). In this study, the production of dextrans, is investigated in a starch-accumulating crop, the potato, after expression of DsrS . DSRS is targeted to the amyloplast, which has an estimated sucrose concentration of about 10 mM (Farré et al ., 2001). In addition, diversification of starch structures might be envisaged with DSRS, because of its high efficiency in glucosylating acceptor molecules, such as maltose, maltodextrins and starch which are present inside the amyloplast (Demuth et al ., 2002).

Materials and methods

Construction of binary plant expression vector containing the DsrS gene An expression cassette containing the patatin promoter (Wenzler et al ., 1989), the chloroplastic ferredoxin signal peptide (FD) from Silene pratensis (Pilon et al ., 1995) fused to the NOS terminator was cloned into the pBluescript SK (pBS SK) plasmid, resulting in pPF. A mature DsrS gene from L. mesenteroides NRRL B-512F (Wilke-Douglas et al ., 1989; U81374) was

Figure 1. Schematic representation of pPFS binary vector used for potato plant transformation. The mature DsrS gene fused to the FD signal peptide that allowed amyloplastic targeting, was inserted between the highly tuber-expressed patatin promoter and the NOS terminator.

ligated in frame between the signal peptide and the NOS terminator. The mature DsrS gene was amplified by PCR, with a forward primer containing a Sma I restriction site (5‘- GCCTCATTTGCTCCCGGG ACACCAAGT-3‘) and a reverse primer containing a Nru I restriction site (5‘-TGGTGGTTCGCGA GTTATGCTGACACA-3‘) using the proofreading Pfu turbo DNA polymerase (2.5 units/ µl; Stratagene, UK) and cloned into the Sma I/ EcoR V

36: Chapter 2/ Production of dextran restriction sites of pPF, resulting in pPF DsrS . FD and the fused DsrS gene were completely sequenced in one direction by Baseclear (The Netherlands) to verify the correctness of the construct. pPF DsrS was digested with Sac I and Kpn I and ligated into a pBIN20 binary vector (Hennegan and Danna, 1998), resulting in pPFS (Fig. 1).

Potato transformation pPFS was transformed into Agrobacterium tumefaciens strain LBA 4404 using electroporation (Takken et al ., 2000). Internodal stem segments from two tetraploid potato genotypes (cv. Kardal (KD) and amylose-free ( amf ) mutant (referred to as 1021-91)) were used for Agrobacterium-mediated transformation. Transformants were selected on plates with MS30 medium (Murashige and Skoog, 1962) containing kanamycin (100 mg/l). Thirty transgenic, root forming, shoots per genotype were multiplied and five plants of each transgenic line were transferred to the greenhouse for tuber development. The mature tubers were harvested after 18 weeks.

Starch isolation Potato tubers from the five plants of each transgenic line were combined, peeled and homogenized in a Sanamat Rotor (Spangenberg, The Netherlands). The resulting homogenate was allowed to settle overnight at 4°C and the potato juice was decanted and stored at œ 20°C for characterization of soluble dextran polymers. The starch pellet was washed three times with water, air-dried at room temperature for at least three days and stored at room temperature.

Immunological detection of dextrans in tuber juices and gelatinized starches Presence of dextrans was investigated with enzyme-linked immunosorbent assay (ELISA) by using monoclonal anti-S-(1R6) dextran antibodies (45.21.1 (groove-type; IgA/Kappa) and 16.4.12E BI (cavity-type; IgA/Kappa)) (Wang et al ., 2002). ELISA plates (NUNC, MAXISORP) were coated with 100 µl/well of the groove-type 45.21.1 dextran antibody at a concentration of 5 µg/ml in 1 x phosphate-buffered saline solution (PBS). After incubation at 37°C for 120 min, the plates were washed 3 x 5 min with ELISA washing buffer (1 x PBS, 0.05 % Tween 20 TM (v/v) and 0.025 % sodium azide (w/v)). The incubations were blocked by adding 200 µl of blocking buffer (1 x PBS, 1 % BSA (w/v), 0.05 % Tween 20 TM (v/v) and 0.025 % sodium azide (w/v)) per well during 60 min at 37°C. Subsequently, the plates were washed 3 x 5 min with ELISA washing buffer. 100 µl/well of blocking buffer was added to the appropriate solute (as listed below) which were incubated for 90 min at 37°C. 1. Dextran, referred to as N279 (10 µg/ml), was used for a standard curve and diluted 1:2 in blocking buffer (5; 2.5; 1.25; 0.625 µg/ml).

37: Chapter 2/ Production of dextran

2. KD-UT potato juice was used for a standard curve in which different concentrations of dextran T40 were added (50; 100; 250; 500; 750; 1000 µg/ml). 3. One ml of potato juice sample was diluted 1:10 in blocking buffer. Starch samples were used as such and heated at 100°C in a water bath for 20 min before application. After incubation, the plates were washed 3 x 5 min with ELISA washing buffer. 100 µl/well of the cavity-type 16.4.12E BI dextran antibody were added at a 1:5000 dilution in blocking buffer. After incubation at 37°C for 60 min, the plates were washed 3 x 5 min with ELISA washing buffer. 100 µl/well of biotin-avidin alkaline phosphatase (AV-AP) were added at a 1:2500 dilution in blocking buffer. After incubation at 37°C for 30 min, the plates were washed 3 x 5 min with ELISA washing buffer. 100 µl/well of substrate solution (1 tablet of alkaline phosphatase substrate/ 8.3 ml diethylamine buffer, pH 9.8) were added. Activities were detected by reading the absorbance at 405 nm.

Expression analysis of DsrS and starch synthesizing genes using semi-quantitative and real-time quantitative RT-PCR analysis RNA was isolated from 3 g (fresh weight) of potato tuber material from selected transgenic lines according to Kuipers et al . (1994). For semi-quantitative RT-PCR, 50 µg of total RNA was treated with DNAseI and purified using the Gene-elute mammalian total RNA kit (Sigma, The Netherlands). The reverse transcription was performed using 5 µg of total RNA which was incubated for 5 min at 65°C with 500 ng primer polydT (5‘-ttttttttttttttttttttttttt-3‘) and 12.5 mM of each dNTP in a final volume of 12 µl. After centrifugation (30 sec; 10,000 g), the mixture was incubated for 2 min at 42°C with 4 µl of 5 x first-strand buffer (Invitrogen, The Netherlands) and 2 µl of 0.1 M DTT. 1 µl of SuperScript II Rnase H - reverse transcriptase (200 U/µl; Invitrogen) was added and the mixture was incubated for 50 min at 42°C. Following this, the reaction was terminated by heating the sample for 15 min at 70°C. 2.5 µl of cDNA was used in a standard PCR reaction with the primer/Tm/cycle number combinations as described as below. For each combination, the cycle number was optimized in order to remain in the exponential phase of the PCR reaction. DsrSRT primers, 5‘- CGGTACGGATGCTGAGGACTT-3‘ and 5‘- GTGTCCGATTAAGTAGTCTAAAGT-3‘ (Tm=59°C, 35 cycles) were based on the DsrS gene sequence (Wilke-Douglas et al ., 1989). Ubi3 primers, 5‘-GTCAGGCCCAATTACGAAGA-3‘ and 5‘-AAGTTCCAGCACCGCACTC-3‘ (Tm=55°C, 40 cycles) were used as an internal control and were based on the ubiquitin-ribosomal protein gene sequence ( Ubi3 ) from potato (Garbarino and Belknap, 1994; L22576). Expression levels of DsrS and genes involved in starch biosynthesis were also determined in parallel by real-time quantitative RT-PCR and the corresponding primers were designed using

38: Chapter 2/ Production of dextran the Primer Express software (version 1.5, PE Applied BioSystems, CA, USA): DsrS primers, 5‘- CAAATCTCAACTGGCGTTCCA-3‘ and 5‘-GCCGCCCACTCAGTAATCTTT-3‘ were based on the same template sequence as that used previously. SuSy primers, 5‘- GAGGACGTGGCAGGTGAAA-3‘ and 5‘-GGTACACTGTGTGACGCCCAT-3‘ were based on the sucrose synthase mRNA sequence (Salanoubat and Belliard, 1987; AY205302). AGPase primers, 5‘-GCTGGGACCCGACTTTATCC-3‘ and 5‘-CGGGAATGTCAATCAGACGAT-3‘ were based on the ADP-glucose pyrophosphorylase subunit S mRNA sequence (Müller-Röber et al ., 1990; X55155). SSIII primers, 5‘-CACAGGAGGTGTCTGGAAACC-3‘ and 5‘- TGGAACTTGTGAAGGTGAGGC-3‘ were based on the starch synthase III mRNA sequence (Marshall et al ., 1996; X95759). SBEI primers, 5‘-CCGAGCCCCACGAATCTAT-3‘ and 5‘- GGCTCAGAGCTGCTCATGC-3‘ were based on the starch branching I enzyme mRNA sequence (Poulsen and Kreiberg, 1993; X69805). GBSSI primers, 5‘- GAGCTTCTGGCAGTGAACCC-3‘ and 5‘-GGCAAGTGGAGCGATTTCTTC-3‘ were based on the granule-bound starch synthase I gene sequence (van der Leij et al ., 1991; X58453). Ubi3 primers, 5‘-TTCCGACACCATCGACAATGT-3‘ and 5‘-CGACCATCCTCAAGCTGCTT-3‘ were used as an internal control as described previously. For real-time quantitative RT-PCR reactions, 1 µg of total RNA was treated with 0.5 µl DNAse I RNAse free (10 U/µl; Invitrogen) and incubated with 5 µl of 10 x Taqman RT buffer, 11 µl of 25 mM MgCl 2, 10 µl of 10 mM dNTP mix, 2.5 µl of 50 µM random hexamer primers, 1.0 µl RNAse inhibitor (20 U/µl) and H 2O until a final volume of 39 µl for 30 min at 37°C and 5 min at 75°C. For reverse transcription, the mixture was incubated 10 min at 25°C and 30 min at 48°C with 1 µl of MultiScribe reverse transcriptase (50 U/µl; Applied Biosystems). Following this, the reaction was terminated by heating the sample for 5 min at 95°C. Aliquots of 50 ng of cDNA were used in SYBR-Green PCR according to the manufacturer's protocol on the ABI PRISM7700 sequence detection system (Perkin-Elmer Applied Biosystems) with the primers mentioned above. Relative quantitation of the target RNA expression level was performed using the comparative Ct method according to the User Bulletin # 2 (ABI PRISM7700 sequence detection system, December 1997; Perkin-Elmer Applied Biosystems). The Ct value is defined as the PCR cycle at which the amount of amplified target RNA reaches a fixed threshold. The differences in the Ct values, called _Ct, between the target RNA and the endogenous Ubi3 RNA were calculated in order to normalize the differences of cDNA concentrations present in each reaction. The differences in the _Ct, called __Ct, between a transformed and an untransformed potato plant were calculated in order to relate the target RNA expression level to that of untransformed plant. The target RNA expression level of transformed plants was calculated using the equation 2-EXP (__Ct). The relative RNA expression is considered as significant when œ 0.5 < relative RNA expression > 0.5.

39: Chapter 2/ Production of dextran

Determination of morphological and physicochemical properties of starch granules Analysis of starch granule morphology was performed by light microscopy (LM) (Axiophot, Germany) equipped with a Sony colour video camera (CCD-Iris/RGB) and scanning electron microscopy (SEM, JEOL 6300F, Japan). For LM, the granules were stained with a 2 a diluted Lugol solution before visualization. For SEM, dried starch samples spread on silver tape and mounted on a brass disk were coated with a 20 nm platinum layer. Samples were then examined with a scanning electron microscope operating at an accelerating voltage of 1.5-3.5 keV. The working distance was 9 mm.

Median values of the granule size distribution (d 50 ) were determined with a Coulter Multisizer II, equipped with an orifice tube of 200 bm (Beckman-Coulter, UK). Approximately 10 mg of starch was suspended in 160 ml Isoton II and the granule size distribution was subsequently measured by counting approximately 50,000 granules. Gelatinization analysis were performed using a differential scanning calorimeter (DSC, Perkin- Elmer Pyris 1, The Netherlands) equipped with a Neslab RTE-140 cooling system. About 10 mg starch was weighed accurately into a stainless-steel pan and water was added until a starch moisture content of 20 % was obtained. The pan was sealed and equilibrated overnight at room temperature before DSC analysis. The measurements were performed at a heating rate of 10°C/min from 40°C to 100°C. Before use, the DSC was calibrated with indium and zinc, and an empty pan was used as a reference. The onset temperature of gelatinization ( T0) and the enthalpy (_H) were calculated automatically. The reported values are the average of three measurements. Amylose content was determined as described by Hovenkamp-Hermelink et al . (1988). Starch was diluted in 50 bl HClO 4 (35 %), followed by determination of the absorption at 618 nm and 550 nm after staining with Lugol solution. For starch content determination, about 50 mg of potato tuber material was homogenized in 0.5 ml of 8 M HCl (25 %). After adding 2 ml of DMSO, the samples were shaken for 1 h at 60°C. Subsequently, 0.8 ml 5 M NaOH and 3.7 ml 0.1 M citric acid buffer pH 4.6 were added. After adjusting the volume to 10 ml with water, the samples were centrifuged for 10 min at 10,000 g. 20 bl of supernatant were used for starch determination using the Boehringer kit (Boehringer, Germany) in a microplate reader (BioRad 3550-UV). Calculations were performed by using glucose release from a known amount of starch as a standard. For determination of the chain length distribution, 5 mg of (transgenic) starch was suspended in 250 bl of DMSO and gelatinized for 15 min at 100°C. After cooling down to room temperature, 700 bl of 50 mM NaAc buffer pH 4.0 was added. A sufficient amount of isoamylase (Hayashibara Biochemical Laboratories, Japan; 59,000 U/ mg protein) to debranch the starch polymers completely was added to the mixture, which was incubated for 2 h at 40°C. After

40: Chapter 2/ Production of dextran inactivation of the enzyme for 10 min at 100°C, 1 ml of 25 % DMSO was added. For high- performance size-exclusion chromatography (HPSEC), 1 ml was used as such. For high- performance anion-exchange chromatography (HPAEC), the sample was diluted 5 times with a 25 % DMSO solution. In parallel, undiluted isoamylase-treated samples were incubated with a sufficient amount of porcine pancreas S-amylase (Merck, Germany; 250 U/ mg protein) to investigate whether contiguous S-(1R6) glucosyl residues linked to amylopectin side chains were present. After incubation for 3 h at 25°C in 50 mM NaAc buffer pH 6.9 and inactivation of the enzyme for 10 min at 100°C, 1 ml was used for HPAEC analysis. HPSEC was performed on a P680 HPLC pump system (Dionex, USA) equipped with an ASI- 100 automated sample injector (Dionex) and three TSKgel columns in series (a G3000 SWXL and two G2000 SWXL; 300 x 7.5 mm; Montgomeryville, USA) in combination with a TSKgel SWXL guard column (40 x 6 mm) at 35°C. Aliquots of 100 bl were injected and eluted with 10 mM of NaAc buffer (pH 5.0) at a flow rate of 0.35 ml/min. The effluent was monitored using a RID-6A refractometer (Shimatzu, Japan). This system was calibrated with dextran standards (10, 40, 70, 150, 250, 500 kDa; Pharmacia, Sweden). Dionex Chromeleon software version 6.50 SP4 Build 1000 was used for controlling the HPLC system and data processing. In order to obtain a better separation of the smaller amylopectin side chains (in the range of 2 to 45 glucose residues), HPAEC was performed on a GP40 gradient pump system (Dionex) equipped with a CarboPac PA 100 column (4 x 250 mm; Dionex) at 35°C. The flow rate was 1.0 ml/min and 20 bl sample was injected with a Dionex AS3500 automated sampler. Two eluents were used, eluent A (100 mM NaOH) and eluent B (1 M NaAc in 100 mM NaOH) as follows: 0R 5 min (100 % eluent B; rinsing phase); 5R 20 min (100 % eluent A; conditioning phase); 20R 25 min (linear gradient 0 to 20 % eluent B; 100 to 80 % eluent A); 25R 50 min (linear gradient 20 to 35 % eluent B; 80 to 65 % eluent A); 50R 55 min (linear gradient 35 to 50 % eluent B; 65 to 50 % eluent A); 55R 60 min (50 % eluent B; 50 % eluent A). The sample was injected at 20 min. The eluent was monitored by an ED40 electrochemical detector in the pulsed amperometric mode (Dionex).

Results

Screening and selection of transgenic potato plants producing dextran FD enabling plastidic protein targeting (Gerrits et al ., 2001) was fused to the mature DsrS gene. The gene fusion was inserted between the patatin promoter (Fig. 1) allowing high tuber expression (Wenzler et al ., 1989) and the Nos terminator sequence.

41: Chapter 2/ Production of dextran

Except for the presence of two mutations at the FDcDsrS fusion (VTAMdATYKVTLITKcATP became VTAMdATYKVTLIT PcGTP, in which d represents the splice site for amyloplast entry and c the gene fusion), the inserts contained no mutations relative to the wild-type sequences (Smeekens et al ., 1985; Wilke-Douglas et al ., 1989). After Agrobacterium -mediated plant transformation, thirty independent transgenic potato clones were obtained for each genotype (Kardal (KD) and the amylose-free ( amf ) mutant). To ensure enough material, five plants of each transgenic clone were grown in the greenhouse for tuberization. Afterwards, the tubers from all plants of each clone were pooled for further characterization. Transformed potato plant series are referred to as KDDxx and amf Dxx, in which D represents the DsrS gene and xx the clone number. Untransformed genotypes are referred to as KD- UT and amf -UT. Screening for dextran accumulation in the transformants was performed by analyzing tuber juice and gelatinized starches with ELISA by using an anti-S-(1R6) dextran antibody, the cavity-type 16.4.12E BI which recognizes the terminal non-reducing end of the polysaccharide (Wang et al ., 2002). In the tuber juice of the KDD series, dextran was detected in 9 out of 30 tubers (29 %) in a concentration ranging from 0.3 to 1.7 mg g-1 FW (Fig. 2). Transformants KDD15, KDD4, KDD5 and KDD30 contained the largest amount of dextran, ranging from 1.0 to 1.7 mg g -1 FW. As expected, no dextran was detected in KD-UT plants. In the amf D series, dextran was detected in 15 out of 27 tubers (56 %), but the amount was lower (0.2 to 0.7 mg g -1 FW) (Fig. 2). No dextran was detected in amf -UT plants. Contrary to tuber juice, no dextran was detected in any of the gelatinized KDD and amf D starches. Subsequent experiments were done on the KDD series, because the accumulation of dextrans was the highest in this genotype. Further characterization of the transgenic starches was performed, because previous studies already demonstrated that the production of novel polymers in the amyloplast can have an influence on starch morphology (Gerrits 2000; Gerrits et al ., 2001). Based on the antibody screening, the transformants were divided in three classes: (œ), (+) and (++), representing no, intermediate (≤ 1.0 mg g -1 FW) and high (> 1.0 mg g -1 FW) levels of dextran, respectively. At least, two transformants of each class were selected for further characterization: KDD9 (œ), KDD20 (œ), KDD27 (+), KDD28 (+), KDD4 (++), KDD5 (++)and KDD30 (++). RNA was isolated from potato tubers and subjected to semi-

42: Chapter 2/ Production of dextran

Figure 2. Detection of dextrans accumulated in potato juices by ELISA using anti-dextran antibodies in KDD and amf D transformants. Based on the dextran concentration (in mg g -1 FW), three categories of transformants were made, where (œ), (+) and (++) represent no, intermediate and high dextran accumulation, respectively. Transgenic clones indicated with grey bars were selected for further characterization.

quantitative RT-PCR (Fig. 3A) and real-time quantitative RT-PCR analysis (Fig. 3B). The expression levels were determined for the DsrS and Ubi3 genes, of which the latter is used as a control because of its known constitutive expression (Garbarino and Belknap, 1994). Heterologous DsrS gene expression was detected in the expressers KDD27, KDD28, KDD4, KDD5 and KDD30 (Fig. 3A and 3B). For KDD20, a small amount of DsrS mRNA was detected by real-time quantitative RT-PCR that was not detected by semi-quantitative RT-PCR. No DsrS mRNA was detected in KDD9 and in the KD-UT plants. In general, results from the semi-quantitative, real-time quantitative RT-PCR and ELISA correlated very well with each other.

43: Chapter 2/ Production of dextran

Figure 3. ( A) Semi-quantitative RT-PCR analysis of the selected KDD transformants and KD - UT tuber RNA. The upper panel shows the PCR products using the DsrST primers that are based on the DsrS sequence. The lower panel shows the PCR products using the Ubi3 primers that served as an internal control. pPF DsrS : positive control. ( B) Real-time quantitative RT - PCR analysis of the selected KDD transformants and KD-UT tuber RNA. The RNA level o f the DsrS gene was expressed relative to the amount of Ubi3 RNA, as described in materials and methods.

Dextran accumulation does not affect plant morphology and tuber growth The morphology of DsrS expressing plants (green parts and tubers) showed no phenotypic alteration in comparison to KD-UT plants (data not shown). For the high expresser KDD30, the tuber number and yield were significantly decreased (about 50 %) in comparison to the other selected transformants (data not shown). However, the accumulation of dextran and decreased tuber number and yield are probably not correlated, because the high expressers KDD4 and KDD5 exhibited similar values to the (œ) class transformants.

44: Chapter 2/ Pr oduction of dextran

Granule morphology and DsrS expression are correlated Impact of dextran accumulation on starch granule morphology was investigated by LM and SEM. With LM, irregular surfaces (for description, see below) were observed more

Figure 4. SEM analysis of starch granules (x 350) from KD-UT ( A) compared to that of selected transformants (KDD9 ( B), KDD27 ( C), KDD4 ( D), KDD5 ( E) and KDD30 ( F)), and examples of starch granules (x 1,000) with altered morphology (KDD4 ( H), KDD5 ( I) and KDD30 ( J)) compared to KD -UT ( G). frequently for the highest expressers (KDD4, KDD5 and KDD30) than for the other selected transformants (data not shown). With SEM, the presence of these irregular surfaces and uncommon forms was confirmed (Fig. 4: D-E-F-H-I-J). The granules of these transformants exhibited round-protruded structures in comparison to those of KD- UT, and the granule surface was rough in contrast to that of KD-UT. Altered starch granules, such as those shown in Fig. 4 (H, I and J), were scored by analyzing a population of 100 starch granules per selected transformant in triplicate (Fig. 5). It can

45: Chapter 2/ Production of dextran

be seen that the percentage of altered starch granules was the highest in the (++) class transformants with KDD30 (16.0 ± 1.0 %), followed by KDD5 (11.0 ± 1.0 %) and KDD4

Figure 5. Percentage of granules with altered morphology for the various transgenic st arches in comparison with KD-UT.

(10.3 ± 0.6 %). Concerning the (œ) and (+) classes transformants (see Fig. 4: B-C), the frequency of altered starch granules was much lower, remaining under 7 %. From these results, it can be concluded that an altered granule phenotype coincides with dextran accumulation.

Dextran accumulation does not interfere with the physicochemical properties of the transgenic starches and the starch content

Median granule size (d 50 ), gelatinization characteristics ( T0 and _H), amylose and starch content (Table 1) were determined and no consistent differences were found for the selected transformants compared to KD-UT. Chain length distributions (HPSEC and HPAEC) (data not shown) were also investigated in order to detect the presence of S- (1R6)-linked dextran chains to starch. After complete debranching of the starch with isoamylase, no differences were observed between the transgenic starches and KD-UT upon HPSEC. Furthermore, after treating the debranched starches with S-amylase, the

46: Chapter 2/ Production of dextran

HPAEC profiles from the transgenic starches were identical to those of KD-UT. These results demonstrate that despite the reported high acceptor efficiency of DSRS (Demuth et al ., 2002), dextran chains were not covalently attached to starch, but rather present as separate chains. In general, it can be said that despite altered granule morphology, it is obvious that accumulation of dextran does not interfere with the

Transformants d50 ( Am) T0 (°C) CH (kJ/g) Amylose Starch content content (%) (mg/g FW)

KD-UT 20.9 (± 0.3) 67.0 (± 0.5) 14.8 (± 0.5) 19.5 (± 0.4) 237.2 (± 24.8)

DsrS9 (œ) 20.9 (± 0.6) 67.5 (± 0.2) 14.0 (± 0.5) 20.4 (± 0.1) 245.5 (± 21.9)

DsrS20 (œ) 22.1 (± 0.3) 66.9 (± 0.1) 14.1 (± 0.9) 20.3 (± 0.6) 230.7 (± 40.9)

DsrS27 (+) 19.8 (± 0.4) 67.7 (± 0.2) 14.1 (± 0.6) 19.7 (± 0.3) 320.7 (± 19.6)

DsrS28 (+) 19.0 (± 0.5) 66.8 (± 0.1) 14.3 (± 0.2) 20.7 (± 0.2) 257.6 (± 34.9)

DsrS4 (++) 18.7 (± 0.7) 66.8 (± 0.2) 15.1 (± 0.4) 20.3 (± 0.3) 268.1 (± 39.7)

DsrS5 (++) 19.9 (± 0.8) 68.3 (± 0.3) 14.8 (± 0.6) 19.9 (± 0.5) 389.4 (± 35.7)

DsrS30 (++) 20.6 (± 0.4) 66.1 (± 0.1) 13.0 (± 0.1) 20.1 (± 0.2) 314.0 (± 38.1)

Table 1. Summary of median granule size (d 50 ), gelatinization characteristics ( To and _ H), amylose and starch content measurements of starches from the selected transformants and KD-UT. Data (± SD) are the average of three independent measurements. physicochemical properties and starch content of the transgenic starches.

Does dextran accumulation influence the expression level of key genes involved in starch biosynthesis ? Effect of dextran accumulation on the expression level of genes involved in starch biosynthesis was performed by real-time quantitative RT-PCR analysis (Fig. 6). The following genes were selected: sucrose synthase ( Susy ), ADP-glucose pyrophosphorylase ( AGPase ), starch synthase III ( SSIII ), starch branching I ( SBEI ) and granule-bound starch synthase I ( GBSSI ). They were chosen based on their key role in starch biosynthesis (Kossmann and Lloyd, 2000).

47: Chapter 2/ Production of dextran

A slight down-regulation in the expression level of Susy , AGPase and GBSSI genes was detected, in particular for the (++) class transformant. Despite this, the physicochemical properties and starch content of the transgenic starches were not significantly changed. Concerning the SSIII and SBEI genes, their expression levels were not affected in comparison to KD-UT.

Figure 6. Real-time quantitative RT-PCR analysis of KDD9 (œ), KDD5 (++) and KDD30 (++) transformants and KD-UT tuber RNA. Expression levels of the following genes are indicated: SuSy , sucrose synthase; AGPase , ADP-glucose pyrophosphorylase subunit S; SSIII , starch synthase III; SBEI , starch branching enzyme I; GBSSI , granule-bound starch synthase I. RNA levels for each gene were expressed relative to the amount of Ubi3 RNA, as described in materials and methods.

Discussion

In this study, we describe the production of dextran in potato tubers and its effect on starch biosynthesis, mediated by the expression of the DsrS gene from L. mesenteroides B-512F . In tuber juice, soluble dextrans were detected with anti-dextran antibodies in both the amylose-containing and amylose-free series, although the Kardal plants accumulated up to 2 times more dextran than the amf ones (Fig. 2). Absence of granule-bound starch synthase I (GBSSI) in the amf background in contrast to the Kardal one, might explain this result. Due to the lack of GBSSI activity, the ADP- glucose pool size in the amyloplast of amf mutant plants might be higher in comparison to that of Kardal plants thereby increasing the osmotic potential of this compartment. Consequently, the transport of sucrose from the cytosol into the amyloplast might be hampered, which might result in a lower production of dextrans in the amf background.

48: Chapter 2/ Production of dextran

Heterologous expression of bacterial DsrS gene did not interfere with normal plant growth and development. In general, no tuber and starch yield penalties were observed for plants producing dextrans. This could be an advantage for commercialization of dextran production in plant systems. These results were different from those obtained by Gerrits et al . (2001), in which expression of the sucrose-converting levansucrase in the amyloplast triggered severe developmental alterations, even when small amounts of fructan were made. It was shown that the presence of dextrans could affect granule morphology, which correlated well with the amount of dextran accumulated in the stroma. This result was similar to that obtained by Gerrits et al . (2001) in which fructan production corresponded to altered starch morphology. One explanation for these morphological changes could be a higher viscosity around the granule during its development. Supporting this, in vitro studies reported the organization of dextran chains into a network structure that could significantly increase its viscosity (Sidebotham, 1975). Therefore, a more viscous environment around the growing granule might interfere with its ordered packing of amylopectin side chains, explaining the presence of irregular surfaces. Direct evidence of dextran accumulation inside starch granules could not be provided by the ELISA and chain length distribution experiments. This result is consistent with the unchanged physicochemical properties of the transgenic starches. Additionally, it demonstrates that despite the reported high acceptor reaction of DSRS (Demuth et al ., 2002), this reaction could not be efficiently realized inside the amyloplast. Expression of Susy , AGPase and GBSSI genes was slightly negatively affected, due to sucrose conversion of DSRS. This might influence the expression level of these genes that are known to be transcriptionally regulated by sucrose (Geigenberger, 2003; Salehuzzaman et al ., 1994). However, it needs to be mentioned that the AGPase and GBSSI down-regulations were not as significant as in Chapters 3, 4 and 5. SSIII and SBEI expression levels were not significantly changed, probably because these enzymes are less sensitive to modifications in sucrose supply. In conclusion, this study is the first report showing that it is possible to produce dextran polymers in potato tubers. Strategies for increasing dextran production could be the targeting of the DSRS enzyme to other subcellular compartments, such as the cytosol or the vacuole, containing higher sucrose concentrations (Farré et al ., 2001).

49: Chapter 2/ Production of dextran

Acknowledgements

The authors would like to thank Isolde Pereira for her assistance with the tissue culture, Dirkjan Huigen for helping with the growth of the plants in the greenhouse, and Ing. Jos Molthoff (PRI-WUR) for his assistance with SYBR-Green analysis.

References van Cleve, J.W., Schaefer, W.C., Rist, C.E., 1956. The structure of NRRL B-512 dextran, methylation studies. J. Am. Chem. Soc. 78, 4435œ4438. Demuth, K., Jördening, H.J., Buchholz, K., 2002. Oligosaccharide synthesis by dextransucrase: new unconventional acceptors. Carbohydr. Res. 337, 1811œ1820. Farré, E.M., Tiessen, A., Roessner, U., Geigenberger, P., Trethewey, R.N., Willmitzer, L., 2001. Analysis of the compartmentation of glycolytic intermediates, nucleotides, sugars, organic acids, amino acids, and sugar alcohols in potato tubers using a nonaqueous fractionation method. Plant Physiol. 127, 685œ700. Garbarino, J.E., Belknap, W.R., 1994. Isolation of a ubiquitin-ribosomal protein gene ( ubi3 ) from potato and expression of its promoter in transgenic plants. Plant Mol. Biol. 24, 119œ127. Geigenberger, P., 2003. Regulation of sucrose to starch conversion in growing potato tubers. J. Exp. Bot. 54, 457œ465. Gerrits, N., 2000. Tuber-specific fructan synthesis in potato amyloplasts. Ph.D. Dissertation, Utrecht University, The Netherlands, ISBN 90-393-2345-3. Gerrits, N., Turk, S.C.H.J., van Dun, K.P.M., Hulleman, S.H.D., Visser, R.G.F., Weisbeek, P.J., Smeekens, S.C.M., 2001. Sucrose metabolism in plastids. Plant Physiol. 125, 926œ934. Groenwall, A.J., Ingelman, B.J.A., 1948. Manufacture of infusion and injection fluids. U.S. Patent 2, 437œ518. Hennegan, K.P., Danna, K.J., 1998. pBIN20: an improved binary vector for Agrobacterium - mediated transformation. Plant Mol. Biol. Rep. 16, 129œ131. Hennink, W.E., van Nostrum, C.F., 2002. Novel crosslinking methods to design hydrogels. Adv. Drug Deliv. Rev. 54, 13œ36. Hovenkamp-Hermelink, J.H.M., de Vries, J.N., Adamse, P., Jacobsen, E., Witholt, B., Feenstra, W.J., 1988. Rapid estimation of the amylose/amylopectin ratio in small amount of tuber and leaf tissue of the potato. Potato Res. 31, 241œ246.

50: Chapter 2/ Production of d extran

Jeanes, A., Haynes, W.C., Wilham, C.A., Rankin, J.C., Melvin, E.H., Austin, M.J., Cluskey, J.E., Fisher, B.E., Tsuchiya, H.M., Rist, C.E., 1954. Characterization and classification of dextrans from ninety-six strains of bacteria. J. Am. Chem. Soc. 76, 5041œ5052. Kok-Jacon, G.A., Ji, Q., Vincken, J-P., Visser, R.G.F., 2003. Towards a more versatile S-glucan biosynthesis in plants. J. Plant Physiol. 160, 765œ777. Kossmann, J., Lloyd, J., 2000. Understanding and influencing starch biochemistry. Critical Rev. Plant Sci. 19, 171œ226. Kuipers, A.G.J., Jacobsen, E., Visser, R.G.F., 1994. Formation and deposition of amylose in the potato tuber starch granule are affected by the reduction of granule-bound starch synthase gene expression. Plant Cell 6, 43œ52. van der Leij, F.R., Visser, R.G.F., Ponstein, A.S., Jacobsen, E., Feenstra, W.J., 1991. Sequence of the structural gene for granule-bound starch synthase of potato ( Solanum tuberosum L.) and evidence for a single point deletion in the amf allele. Mol. Gen. Genet. 228, 240œ248. Marshall, J., Sidebottom, C., Debet, M., Martin, C., Smith, A.M., Edwards, A., 1996. Identification of the major starch synthase in the soluble fraction of potato tubers. Plant Cell 8, 1121œ1135. Monchois, V., Willemot, R.M., Monsan, P., 1999. Glucansucrases: mechanism of action and structure-function relationships. FEMS Microbiol. Rev. 23, 131œ151. Müller-Röber, B.T., Kossmann, J., Hannah, L.C., Willmitzer, L., Sonnewald, U., 1990. One of two different ADP-glucose pyrophosphorylase genes from potato responds strongly to elevated levels of sucrose. Mol. Gen. Genet. 224, 136œ146. Murashige, T., Skoog, F., 1962. A revised medium for rapid growth and bioassay with tobacco tissue culture. Physiol. Plant. 15, 473œ497. Murphy, P.T., Whistler, R.L., 1973. Dextrans. In: R.L. Whistler (Ed.), Industrial gums, Academic Press, New York, pp. 513œ542. Pilon, M., Wienk, H., Sips, W., de Swaaf, M., Talboom, I., van ‘t Hof, R., de Korte-Kool, G., Demel, R., Weisbeek, P., de Kruijff, B., 1995. Functional domains of the ferredoxin transit sequence involved in chloroplast import. J. Biol. Chem. 270, 3882œ3893. Poulsen, P., Kreiberg, J.D., 1993. Starch branching enzyme cDNA from Solanum tuberosum . Plant Physiol. 102, 1053œ1054. Richard, G., Morel, S., Willemot, R.M., Monsan, P., Remaud-Simeon, M., 2003. Glucosylation of S-butyl- and S-octyl-D-glucopyranosides by dextransucrase and alternansucrase from Leuconostoc mesenteroides . Carbohydr. Res. 338, 855œ864. Robyt, J.F., 1995. Mechanisms in the glucansucrase synthesis of polysaccharides and oligosaccharides from sucrose. Adv. Carbohydr. Chem. Biochem. 51, 133œ168.

51: Chapter 2/ Production of dextran

Salanoubat, M., Belliard, G., 1987. Molecular cloning and sequencing of sucrose synthase cDNA from potato ( Solanum tuberosum L.): preliminary characterization of sucrose synthase mRNA distribution. Gene 60, 47œ56. Salehuzzaman, S.N.I.M., Jacobsen, E., Visser, R.G.F., 1994. Expression patterns of two starch biosynthetic genes in in vitro cultured cassava plants and their induction by sugars. Plant Sci. 98, 53œ62. Scheller, J., Gührs, K-H., Grosse, F., Conrad, U., 2001. Production of spider silk proteins in tobacco and potato. Nature Biotechnol. 19, 573œ577. Sidebotham, R.L., 1975. Dextrans. Adv. Carbohydr. Chem. Biochem. 30, 371œ444. Smeekens, S., van Binsbergen, J., Weisbeek, P., 1985. The plant ferredoxin precursor: nucleotide sequence of a full length cDNA clone. Nucleic Acids Res. 13, 3179œ3194. Takken, F.L.W., Luderer, R., Gabriëls, S.H.E.J., Westerink, N., Lu, R., de Wit, P.J.G.M., Joosten, M.H.A.J., 2000. A functional cloning strategy, based on a binary PVX- expression vector, to isolate HR-inducing cDNAs of plant pathogens. Plant J. 24, 275œ 283. Wang, D., Liu, S., Trummer, B.J., Deng, C., Wang, A., 2002. Carbohydrate microarrays for the recognition of cross-reactive molecular markers of microbes and host cells. Nature Biotechnol. 20, 275œ281. Wenzler, H.C., Mignery, A., Fisher, L.M., Park, W.D., 1989. Analysis of a chimeric class-I patatin-GUS gene in transgenic potato plants: High-level expression in tubers and sucrose-inducible expression in cultured leaf and stem explants. Plant Mol. Biol. 12, 41œ 50. Wilke-Douglas, M., Perchorowicz, J.T., Houck, C.M., Thomas, B.R., 1989. Methods and compositions for altering physical characteristics of fruit and fruit products. PCT patent WO 89/12386.

53: Chapter 3/ Production of mutan

Mutan produced in potato amyloplasts adheres to starch granules

Géraldine A. Kok-Jacon Luc C.J.M. Suurs Jean-Paul Vincken Richard G.F. Visser

Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Wageningen University, P.O. Box 386, 6700 AJ Wageningen, The Netherlands

Abstract : Production of water-insoluble mutan polymers in Kardal potato tubers was investigated after expression of a full-length ( GtfI ) and a truncated mutansucrase gene referred to as GtfICAT ( GtfI without glucan-binding domain) from Streptococcus downei. Subsequent effects on starch biosynthesis at the molecular and biochemical levels were studied. Expression of the GtfICAT gene resulted in the adhesion of mutan material on starch granules, which stained red with erythrosine, and which was hydrolyzed by exo-mutanase. In addition, GtfICAT -expressing plants exhibited a severely altered tuber phenotype and starch granule morphology in comparison to those expressing the full-length GtfI gene. In spite of that, no structural changes at the starch level were observed. Expression levels of the sucrose- regulated, AGPase and GBSSI genes were down-regulated in only the GTFICAT transformants, showing that GtfICAT expression interfered with the starch biosynthetic pathway. In accordance with the down-regulated AGPase gene, a lower starch content was observed in the GTFICAT transformants. Finally, the rheological properties of the GTFICAT starches were modified; they showed a higher retrogradation during cooling of the starch paste.

Published in: Plant Biotechnology Journal 3; 341-351 (2005). Reprinted with permission of Blackwell Publishing Ltd

Key words : mutan, glucansucrase, adhesiveness, glucan-binding domain, truncation, transgenic potato

54: Chapter 3/ Production of mutan

Introduction

In planta production of bacterial extracellular polysaccharides could generate a resource with unique environmental and commercial added-values (Moire et al ., 2003). For instance, adhesive properties of bacterial polysaccharides such as xanthan and pullulan are actually regarded as possible replacements of synthetic paper and wood adhesives (Haag et al ., 2004). Another polysaccharide with adhesive properties is mutan (see further). Besides, the in planta production of bacterial polymers may alter the properties of existing polymers such as starch, which is of particular interest to us. It could be advantageous to replace post-harvest chemical starch modifications by a more environment-friendly bioprocessing. Expression of the GtfI mutansucrase gene (Ferretti et al ., 1987), which is produced by the oral cariogenic Streptococcus downei ( S. downei ) Mfe28 bacteria, leads to the accumulation of mutan polymers in the presence of sucrose. According to in vitro experiments, GTFI catalyzes the formation of two types of chemical linkages between glucosyl residues, i.e. S-(1R3)-linkages (88 %), forming the backbone of mutan, and S- (1R6) linkages (12 %), connecting single unit glucosyl residues to the backbone (Russell et al ., 1987; Monchois et al ., 1999a). Mutan polymers account for about 70 % of the carbohydrates present in dental plaque (Loesche, 1986), the formation of which is briefly discussed below. Saliva-coated enamel surfaces are colonized by diverse oral bacteria, referred to as the early colonizers that adhere to receptors present on teeth surfaces by means of adhesin proteins. In turn, these bacteria secrete various polysaccharides such as mutans, dextrans and levans that exhibit different degrees of water-solubility (Sutherland, 2001). These polymers, together with the early colonizers, enhance the aggregation of the late colonizers, creating a biofilm, which is usually named dental plaque (Marsh, 2003). Dental plaque stains red by erythrosine (Leknes and Lie, 1988), an iodine containing molecule. Iodine is thought to be responsible for binding to the constituent polysaccharides. From the polymers that are formed, mutan is the most adhesive and water-insoluble one (Hamada and Slade, 1980). Interestingly, the chain conformation of mutan is very similar to that of cellulose, an extended helix with a two-fold symmetry, which may explain its poor solubility in water (Marchessault and Deslandes, 1980). The exact adhesion mechanism by which mutans adhere to surfaces in relation to the other dental plaque components requires further elucidation,

55: Chapter 3/ Production of mutan although there is some evidence that S-(1R6)-linked glucosyl residues might contribute to adhesive properties (Colby et al. , 1999). Because of their implications in human dental caries, different studies with genetically engineered GTFIs have been carried out in order to elucidate structure-function relationships with respect to mutan synthesis. GTFI, which belongs to family 70 of the glycoside hydrolases (http://afmb.cnrs-mrs.fr/CAZY/), has a primary structure, which is common to all glucansucrase enzymes, consisting of a signal peptide, a variable region, a catalytic domain (CAT) and a glucan-binding domain (GBD) (Monchois et al ., 1999a). Its enzymatic reaction takes place by a so-called two-site insertion mechanism in which glucose residues from sucrose molecules are polymerized in a growing mutan chain, or linked to carbohydrate acceptor molecules (Robyt, 1995). Until now, studies on GTFI acceptor reaction efficiency have not been done. Interestingly, expression of only its catalytic domain resulted in an active GTFI enzyme, with an activity of approximately 70 % of the WT enzyme (Monchois et al ., 1999b). Glucan-binding domains of GTFI share sequence homologies with the glucan-binding protein of Streptococcus mutans and the choline-binding domain of pneumococcal proteins (Wren, 1991). As their name implies, their role is to mediate the binding of the synthesized glucan chains to the enzyme. Industrial applications of mutan polymers have not been developed as far as we know. However, they could be of interest as functional food (prebiotics) (Tuohy et al ., 2003) and possibly as glues. The aim of this study is to produce mutan polymers in potato amyloplasts by expressing a full-length and a glucan-binding domain-truncated gene. Amyloplastic transport of the truncated enzyme may be enhanced, because of its smaller molecular size. Therefore, the truncated and full-length GTFIs are evaluated with respect to their efficiency in mutan production in the amyloplast. In addition, the effects of the introduction of these genes on starch biosynthesis are investigated at the molecular and biochemical levels.

Materials and methods

Preparation of constructs containing the mature and truncated GtfI genes An expression cassette containing the patatin promoter (Wenzler et al ., 1989), the chloroplastic ferredoxin signal peptide (FD) from Silene pratensis (Pilon et al ., 1995) fused to the NOS terminator was cloned into the pBluescript SK (pBS SK) plasmid, resulting in pPF. A mature mutansucrase ( GtfI ) gene from S. downei Mfe28 (Ferretti et al ., 1987; M17391) was ligated in

56: Chapter 3/ Production of mutan frame between the signal peptide FD and the NOS terminator. The mature GtfI gene was amplified by PCR, with a forward primer containing a Sma I restriction site (5‘- AGCTTGCGGCCCCGGG ACTGAAAC -3‘) and a reverse primer containing an EcoR I restriction site (5‘- GTGGTGGTGGAATTC GAGTTAGTTC -3‘) using the proofreading Pfu turbo DNA polymerase (2.5 units/ µl; Stratagene, UK) and cloned into the Sma I/ EcoR I restriction sites of pPF, resulting in pPF GtfI . FD and the fused GtfI gene were completely sequenced in one direction by Baseclear (The Netherlands) to verify the correctness of the construct. pPF GtfI was digested with Xho I and ligated into a pBIN20 binary vector (Hennegan and Danna, 1998), resulting in pPFI (Fig. 1A). For the construction of the FD-GtfICAT -NOS fusion, the mature GtfI gene truncated at position 3150 was amplified by PCR, with a forward primer containing a Sma I restriction site (5‘- AGCTTGCGGCCCCGGG ACTGAAAC -3‘) and a reverse primer containing an EcoR I restriction site (5‘- AGAAGGAATTC TCATCTTAAACATTGAGGTA -3‘) and cloned into the Sma I/ EcoR I restriction sites of pPF, resulting in pPF GtfICAT . Sequencing and cloning into the pBIN20 binary vector, resulting in pPFIC (Fig. 1B), were performed as for pPFI .

Figure 1 . Schematic representation of pPFI ( A) and pPFIC ( B) binary vectors used for potato plant transformation.

Transformation and regeneration of potato plants pPFI and pPFIC were transformed into Agrobacterium tumefaciens strain LBA 4404 using electroporation (Takken et al ., 2000). Internodal stem segments from the tetraploid potato genotype (cv. Kardal (KD)) were used for Agrobacterium-mediated transformation which was performed as described by Kok-Jacon et al . (2005).

57: Chapter 3/ Production of mutan

Starch isolation Potato tubers were peeled and homogenized in a Sanamat Rotor (Spangenberg, The Netherlands). The resulting homogenate was allowed to settle overnight at 4°C. The starch pellet was washed three times with water and finally air-dried at room temperature for at least three days. The dried starch was powdered and stored at room temperature.

Expression analysis of GtfI , GtfICAT and starch synthesizing genes using semi- quantitative and real-time quantitative RT-PCR analysis RNA was isolated from 3 g (fresh weight) of potato tuber material from selected transgenic lines according to Kuipers et al. (1994). Semi-quantitative and real-time quantitative RT-PCRs were performed as described by Kok- Jacon et al . (2005). GtfIRT primers, 5‘-CCGTGCTTACAGTACCTCAGC-3‘ and 5‘- GGTCGTTAGCATTGTAGGTGAAA-3‘ (Tm=59°C, 35 cycles) were based on the GtfI gene sequence (Ferretti et al ., 1987). RNA samples from Karnico potato tubers expressing a sense/ antisense GBSSI cDNA construct referred to as 3.8-32 and RVT34-77 (Heilersig et al ., unpublished results) were used as positive control. Extent of down-regulation was determined microscopically from which the 3.8-32 and RVT34-77 transformants were defined as strongly and completely down-regulated, respectively.

Determination of morphological and physicochemical properties of starch granules Analysis of starch granule morphology was performed by light microscopy (LM) and scanning electron microscopy (SEM) as described by Kok-Jacon et al . (2005). Mutan polymers were visualized with LM by staining starch granules with a 10 x diluted erythrosine red colouring agent (Disclosing Red-Cote solution) (American Dental Trading BV, The Netherlands). Mutan polymers and exo-mutanase were kindly provided by Dr. A. Wiater (Department of Industrial Microbiology, Lublin, Poland). Mutan polymers were produced in presence of sucrose by a mixture of streptococcal glucosyltransferase ( Streptococcus mutans 20381, S. mutans 6067 and S. sobrinus 6070) serving as a positive control (Wiater et al ., 1999). Exo-mutanase (S 1,3- glucanase, EC 3.2.1.59) was produced by Trichoderma harzianum F-470 (Wiater and Szczodrak, 2002). Exo-mutanase treatments were performed with 25 mU of exo-mutanase enzyme in 0.2 M sodium acetate buffer (pH 5.5) at 40°C for 48 h in presence of 10 mg (transgenic) starch. After brief centrifugation (1 min; 10,000 g), the supernatant was discarded and the starch granules were stained with the erythrosine red colouring solution.

Median values of the granule size distribution (d 50 ), gelatinization analysis, amylose content, starch content, chain length distributions (HPSEC, HPAEC) were determined as described by Kok-Jacon et al. (2005).

58: Chapter 3/ Production of mutan

Viscometric profiles from a 2 % starch suspension were determined by applying a small oscillating shear deformation at a frequency of 1 Hz, using a Thermo Haake rheoscope. The rheometer was equipped with parallel plate geometry (typ C70/1 Ti) and the gap size was 0.1 mm. The pasting profile of the 2 % starch-water (w/v) suspension was obtained by heating the suspension from 40°C to 90°C at a rate of 2°C/min, where it was kept for 15 min followed by cooling to 20°C at a rate of 2°C/min. After this, the starch paste was held at 20°C for 15 min. The Tg (start gelatinization temperature), Tp (peak temperature) and the corresponding viscosities were measured. Subsequently, the retrogradated sample was subjected to an amplitude sweep from 10 Pa to 1,000 Pa within 60 s to check that the measurements were made in the linear region, in which the amplitudes of stress and strain are proportional to each other.

Results

High GtfICAT mRNA level correlates with altered tuber phenotype The plastidic protein targeting sequence (FD) (Gerrits et al ., 2001) was fused in frame to the mature or GBD-truncated GtfI genes (Fig. 1). These fragments were driven by the patatin promoter allowing high tuber expression (Wenzler et al ., 1989). For each of the constructs, restriction sites were engineered at the FDcgene fusion, creating two mutations (VTAMdATYKVTLITKcDTE became VTAMdATYKVTLIT PcGTE, in which d represents the splice site for amyloplast entry and c the gene fusion). In addition, the pPFIC sequence differed from the published GTFI sequence (Ferretti et al ., 1987) at position 408 (L 408 F) which did not affect any conserved residue. Thirty independent transgenic potato clones were obtained after Agrobacterium -mediated plant transformation with pPFI and pPFIC. Five plants of each transgenic clone were grown in the greenhouse for tuberization to ensure enough material and the resulting tubers were pooled for further characterization. Transformed potato plant series are referred to as KDIxx and KDICxx, in which I and IC represent the GtfI and GtfICAT genes, respectively, and xx the clone number. The untransformed genotype is referred to as KD-UT. RT-PCR was performed on a number of transformants that were divided in classes, based on the band intensity of the different PCR products. The band intensities were compared to that of Ubi3 , which is used as an internal control (Garbarino and Belknap, 1994). From the KDI series, the KDI14 (œ), KDI2 (œ), KDI26 (+), KDI30 (+), KDI11 (++)

59: Chapter 3/ Production of mutan and KDI20 (++) transformants were selected for further characterization (Fig. 2A) in which (œ), (+) and (++) represent no, intermediate and high levels of mRNA, respectively. Because of the low Ubi3 intensity of KDI2, this transformant might also be attributed to the (+) class. In addition, KDIC1 (œ), KDIC27 (œ), KDIC22 (+), KDIC24 (+),

Figure 2 . RT-PCR analysis of the selected KDI and KDIC transformants, and KD-UT. The upper panel shows the PCR products using the primers designed on the GtfI and GtfICAT sequences f rom the KDI and KDIC transformants, respectively. The lower panel shows the PCR products using the primers designed on the Ubi3 sequence that served as an internal control. pPF GtfI plasmid: positive control.

KDIC14 (++) and KDIC15 (++) transformants were selected from the KDIC series (Fig. 2B). As expected, no GtfI mRNA was detected in the KD-UT plants. A severely altered tuber phenotype was found for the (++) class of the KDIC series in which almost all the tubers exhibited brownish/reddish discolouration (see Fig. 3), which was present inside and close to the vascular regions of the tuber. Before starch isolation, each tuber was cut in two parts in order to check the colouration. The (+) class transformants also showed such a phenotype but to a lesser extent (data not shown). The (œ) class transformants exhibited a tuber phenotype comparable to KD- UT. Thus, the colour of the tuber tissue seemed to be correlated with the level of GtfICAT expression. No phenotypical alterations were observed for the KDI series. For the KDI and KDIC series, the tuber number, yield and plant morphology were unchanged and comparable to KD-UT.

60: Chapter 3/ Production of mutan

Figure 3. Tuber phenotype of the KDIC15 (++) transformant compared to that of KD -UT.

Mutans are only visualized in the (+) and (++) classes of KDIC transformants An erythrosine red colouring solution was used for the visualization of mutans attached to starch granules. As a positive control, mutan polymers (Wiater et al ., 1999) were stained with this colouring agent (Fig. 4B); normal starch granules, containing amylopectin and amylose, did not stain with the dye. Interestingly, mutans were present on KDIC15 starch granule surfaces (Fig. 4C). A red staining was also observed on KDIC14 and KDIC22 starch granule surfaces, but to a lesser extent (data not shown), but no colouration was observed for the KDI series, which was comparable to KD-UT (Fig. 4A). When KDIC15 starch granules were treated with an exo-mutanase solution (Fig. 4D), most of the mutans were detached from the starch granules. This provided additional proof that the red-staining patches on starch granules corresponded to mutan. It also suggests that the binding only occurred at the starch granule surface.

GBD truncation correlates with a severely altered granule morphology The morphology of starch granules was determined by SEM and LM. With SEM, the presence of altered starch granules was observed for the KDI and KDIC series, particularly for the (++) class (Fig. 5: C-D-G-H-J-K-L-M) and (+) class (data not shown) of transformants. The granule morphology of the (œ) class (Fig. 5: B and F, representing

61: Chapter 3/ Production of mutan

Figure 4. LM analysis of starch granules (x 800) stained with Disclosing solution red-Cote from KD-UT ( A) and KDIC15 ( C) compared to that of mutan polymers ( B). D) Treatment of KDIC15 starch granules with 25 mU of exo -mutanase enzyme. the KDI and KDIC series, respectively) was comparable to that of KD-UT (Fig. 5: A-E). For the (++) class of the KDI series, starches contained uncommonly shaped granules with protruded forms and with small granules that associated to larger ones (Fig. 5: C- D-J-K). For the (++) class of the KDIC series, starch granules exhibited uncommon shapes with eroded and protruded forms (Fig. 5: G-H-L-M). In addition, pores in the granule surface were often observed. With LM, the presence of altered starch granules was confirmed for each of the series, in particular for the (++) and (+) classes (data not shown). It seems that expression of a GBD-truncated GtfI induces a more severe effect on starch granule morphology than the mature one. Quantification of altered starch granules number such as those represented in Figure 5 (J-K-L-M) was performed by analyzing a population of 100 starch granules in triplicate for a representative transformant of the (œ) and (+) classes and two transformants of the (++) class for each series. This quantification is shown in Figure 6 for the KDI14 (œ), KDI30 (+), KDI11 (++), KDI20 (++) and KDIC1 (œ), KDIC22 (+), KDIC14 (++), KDIC15 (++) transformants. The

62: Chapter 3/ Production of mutan

Figure 5 . SEM analysis of starch granules (x 350) from KD-UT ( A, E) compared to that of selected transformants (KDI14 ( B), KDI11 ( C), KDI20 ( D), KDIC1 ( F), KDIC14 ( G) and KDIC15 (H)), and examples of starch granules (x 1,000) with altered morphology (KDI11 ( J), KDI20 ( K), KDIC14 ( L) and KDIC15 ( M)) compared to KD-UT ( I).

Figure 6. Percentage of starch granules with altered morphology (see Fig. 5, J, K, L and M) in selected transformants compared to KD-UT ( I).

highest percentage of altered starch granules was found in the (+) and (++) classes of the KDI and KDIC transformants, ranging from 22.3 ± 2.5 % for KDI11 to 31.3 ± 2.3 % for KDIC15. For the (œ) class of transformants, the frequency of altered starch granules was about 13 % and that of KD-UT about 3 %.

63: Chapter 3/ Production of mutan

The (++) KDIC transformants exhibit decreased starch content and novel physicochemical properties Starch content (Table 2) of the (++) KDIC class was decreased, ranging from 114.5 ± 43.6 mg/g FW for KDIC14 to 81.3 ± 37.3 mg/g FW for KDIC15 in comparison to KD-UT (190.7 ± 15.2 mg/g FW). Starch content of the KDI (Table 1), the (œ) and (+) classes of the KDIC transformants (Table 2) was comparable to that of KD-UT.

Table 1: Overview of physicochemical properties and starch content of KDI transformants: Summary of granule size (d 50 : median value of the granule size distribution), gelatinization characteristics (T onset (T o): temperature of onset of starch gelatinization; _H: enthalpy released), amylose and starch content measurements of starches from the selected transformants and KD-UT. Data (± SD) are the average of two or three independent measurements.

Transformants d50 (m) T0 (°C) H(kJ/g) Amylose Starch content content (%) (mg/g FW)

KD-UT 21.5 ( ± 0.4) 66.6 ( ± 0.1) 14.3 ( ± 2.1) 19.5 ( ± 0.4) 226.4 ( ± 49.3)

KDI14 (œ) 21.8 ( ± 0.8) 67.0 ( ± 0.1) 14.4 ( ± 0.5) 19.4 ( ± 0.1) 211.5 ( ± 88.4)

KDI2 (œ) 20.3 ( ± 0.5) 64.7 ( ± 0.1) 15.7 ( ± 0.2) 19.5 ( ± 0.2) 355.6 ( ± 41.9)

KDI26 (+) 21.9 ( ± 0.6) 67.1 ( ± 0.1) 15.5 ( ± 0.1) 19.2 ( ± 0.3) 214.7 ( ± 110.8)

KDI30 (+) 18.6 ( ± 0.0) 64.8 ( ± 0.4) 13.0 ( ± 0.1) 19.3 ( ± 0.2) 286.0 ( ± 120.1)

KDI11 (++) 25.1 ( ± 0.8) 66.3 ( ± 0.1) 14.4 ( ± 0.1) 19.1 ( ± 0.1) 209.7 ( ± 115.2)

KDI20 (++) 19.6 ( ± 0.5) 66.0 ( ± 0.4) 14.2 ( ± 0.1) 19.1 ( ± 0.3) 319.3 ( ± 97.0)

Median granule size (d 50 ), gelatinization characteristics (T onset ( T0) and enthalpy released (_H)) and amylose content (Tables 1 & 2) of the transgenic starches were also determined, and no significant differences were found in comparison to KD-UT. The viscosimetric analysis showed that KDIC15 starch (Figure 7 (1)) had different behaviour from that of KDI20 (Figure 7 (2)) and KD-UT (Figure 7 (3)), particularly with respect to the end viscosity, which was about 1.7-fold higher. The presence of S-(1R3)-linked mutan chains in starch was investigated by treating the gelatinized starches with isoamylase and S-amylase, followed by chromatography of the reaction products. After complete debranching of the starch molecules with isoamylase, we did not observe differences between KD-UT and transgenic starches upon HPSEC, neither shifts in populations of amylopectin side chains, nor the

64: Chapter 3/ Producti on of mutan

Table 2: Overview of physicochemical properties and starch content of KDIC transformants : For description, see Table 1.

Transformants d50 (m) T0 (°C) H(kJ/g) Amylose Starch content content (%) (mg/g FW)

KD-UT 21.4 ( ±0.7) 66.6 ( ±0.3) 14.9 ( ±1.0) 19.5 ( ±0.4) 190.7 ( ±15.2)

KDIC1 (œ) 20.5 ( ±0.6) 67.1 ( ±0.2) 15.6 ( ±1.6) 19.5 ( ±0.2) 173.9 ( ±16.7)

KDIC27 (œ) 18.5 ( ±0.3) 66.9 ( ±0.1) 15.2 ( ±0.8) 19.1 ( ±0.0) 161.2 ( ±7.7)

KDIC22 (+) 31.1 ( ±0.9) 68.0 ( ±0.4) 14.8 ( ±0.5) 20.4 ( ±0.3) 173.0 ( ±32.4)

KDIC24 (+) 23.4 ( ±0.5) 70.2 ( ±0.1) 15.5 ( ±0.8) 18.6 ( ±0.1) 163.1 ( ±19.6)

KDIC14 (++) 22.7 ( ±0.6) 65.4 ( ±0.1) 14.9 ( ±1.4) 20.0 ( ±0.4) 114.5 ( ±43.6)

KDIC15 (++) 20.2 ( ±0.4) 67.2 ( ±0.1) 13.6 ( ±1.8) 19.7 ( ±0.2) 81.3 ( ±37.3) appearance of a new peak which could correspond to mutan. With HPAEC, which gives a better separation of the smaller amylopectin side chains, a similar result was

Figure 7. Viscosity profiles of KDIC15 (++) (1) and KDI20 (++) (2) starches compared to that of KD-UT (3). obtained; no additional peaks were seen in the chain length distribution profile. After treatment of the debranched starches with S-amylase, the HPAEC profiles obtained

65: Chapter 3/ Production of mutan with KD-UT and transgenic starches were identical. These data indicated that the mutan is predominantly present as a separate carbohydrate, which is not covalently attached to starch molecules. Summarizing, high GtfICAT gene expression seems to correlate with lower starch content and higher viscosity, in particular at the end of the gelatinization process.

Expression levels of AGPase and GBSSI are down-regulated in the (++) KDIC class Real-time quantitative RT-PCR was performed to determine the expression level of key

Figure 8. Real-time quantitative RT-PCR analysis of various KDIC ( A) and KDI ( B) transformants, and KD-UT tuber RNA. Expression levels of the following starch synthesizing genes were measured: SuSy , sucrose synthase; AGPase , ADP-glucose pyrophosphorylase subunit S; SSIII , starch synthase III; SBEI , starch branching enzyme I; GBSSI , granule-bound starch synthase I. RNA levels for each gene were expressed relative to the amount of Ubi3 RNA, as described in materials and methods. RNA samples from Karnico potato tubers expressing a sense/antisense GBSSI cDNA construct exhibiting a partial (3.8-32) as well as a complete GBSSI down-regulation (RVT34-77), were used as positive controls ( A).

66: Chapter 3/ Production of mutan genes involved in starch biosynthesis (Kossmann and Lloyd, 2000) including sucrose synthase ( SuSy ), ADP-glucose pyrophosphorylase ( AGPase ), starch synthase III (SSIII ), starch branching I ( SBEI ) and granule-bound starch synthase I ( GBSSI ) (Fig. 8). Expression levels of AGPase and GBSSI were down-regulated in the (++) KDIC class when compared to the KDI transformants and KD-UT. For GBSSI , this down- regulation was quite similar to that of the 3.8-32 transformant in which GBSSI expression is partially inhibited. In contrast, this down-regulation was more than 20 times less than for the transformant RVT34-77 in which GBSSI is completely inhibited. This implicates that the observed decreases in GBSSI expression for the (++) KDIC class were significant within the selected transformants, but relatively small with respect to the RVT34-77 transformant. Expression level of the SBEI , SuSy and SSIII genes in the (++) KDIC class remained unchanged. Concerning the KDI transformants, the expression levels of the different starch biosynthetic genes were comparable to that of KD-UT.

Discussion

This paper describes the production of mutan in potato tubers after expression of a full- length and a GBD-truncated GtfI . GtfICAT expression led to a more efficient in planta production of mutan which was higher than expected, because in vitro experiments demonstrated that GTFICAT exhibited a lower activity (Monchois et al ., 1999b). One reason for this could be that its transport to the amyloplast may be enhanced due to its decreased molecular weight. In addition, GtfICAT expression induced more severe developmental alterations in comparison to that of GtfI , which were phenotypically visible at the tuber and starch levels. Tuber browning in the GtfICAT -expressing plants could be due to a stress response, probably resulting in the production of phenolic compounds (Tomás-Barberán and Espín, 2001). Interestingly, the link between tuber browning and production of foreign polymers in potato plants was reported previously (Pilon-Smits et al ., 1996). In that study, a levansucrase gene with a vacuole-targeting sequence was expressed in potato tubers. Levansucrase and mutansucrase have in common that they consume sucrose, and consequently they can alter the sucrose pool size, and influence processes depending on sucrose. Since the vacuole contains approximately three times more sucrose than the amyloplast (Farré et al ., 2001),

67: Chapter 3/ Production of mutan vacuolar targeting of sucrose-converting enzymes may have a larger impact on plant development than plastidial targeting, at least when the activity of such enzymes is non-limiting. This is consistent with observations that the growth of levansucrase transformants was impaired, whereas that of mutansucrase transformants was not. Sucrose conversion by GTFICAT may also induce changes at the gene transcription level. Supporting this, it was shown that the starch synthesizing AGPase and GBSSI genes were down-regulated in KDIC transformants. These genes are known to be transcriptionally regulated by sucrose (Geigenberger, 2003; Salehuzzaman et al ., 1994). AGPase down-regulation might explain the lower starch content of the KDIC transformants. However, the amylose content of these granules was similar to that of KD-UT granules, which seems to disagree with their lower level of GBSSI transcript. This result differed from that of the 3.8-32 transformant in which the amylose content was lower, with an amount of GBSSI mRNA similar to that of the KDIC transformants. It needs to be mentioned that the down-regulation of GBSSI occurs in two different ways, at the transcriptional level for KDIC and at the post-transcriptional level for 3.8-32, respectively. Our result indicates that RNA degradation might affect the amylose content more severely than a less efficient transcription of the GBSSI gene. Attachment of mutan to KDIC starch granules was not observed with KDI starch granules. To a certain extent, similar mechanisms may be at play in the attachment of mutan to starch granules and teeth surfaces. However, in the latter case, adhesion forces would be strengthened due to the synergistic action of various adhesive agents present in dental plaque (Balakrishnan et al ., 2000). After exo-mutanase treatment, most of the mutan was detached from the starch granules and the remaining granule surfaces were comparable to those of KD-UT. Evidence for the presence of mutan inside starch granules was not found, although this possibility can not be excluded. Based on the chain length distributions experiments, it is not very likely that mutan is an intrinsic part of amylopectin and/or amylose molecules. The so-called acceptor reaction which is poorly documented for GTFI, is apparently of minor importance inside the amyloplast. Furthermore, it is not known from the literature whether starch is a good acceptor molecule for mutansucrase. However, it is well-known that the acceptor reaction efficiency decreases substantially with increasing molecular weight of the acceptors (Kok-Jacon et al ., 2003). Another explanation could be that the mutan

68: Chapter 3/ Production of mutan concentration is too low in comparison to that of starch, thereby limiting the detection of attached mutan chains by HPAEC analysis. Alterations in granule morphology were more severe for KDIC starches in which eroded and porous granules were commonly observed. Alterations in KDI starch granule morphology were also visualized despite the absence of stained mutan on starch granule surfaces upon erythrosine treatment. Finally, mutan production in potato amyloplasts interfered with the rheological properties of the starches. The simultaneous presence of mutan and starch gave a higher viscosity after cooling the starch paste. It is possible that the mutans with their adhesive properties firmly interact with the amylopectin and amylose chains, thereby increasing the viscosity more than with a control starch, showing normal retrogradation. This study is the first report dealing with the expression of a full-length and a GBD- truncated GtfI gene in plants. Interestingly, it demonstrates that GTFICAT, which is still active after domain truncation, is an efficient catalyst for the enhanced production of foreign polymers in plants.

Acknowledgements

We would like to thank Isolde Pereira for her assistance with the tissue culture, Dirkjan Huigen for helping with the growth of the plants in the greenhouse and Ing. Jos Molthoff (PRI-WUR) for his assistance with SYBR-Green analysis. In addition, we are very grateful to Dr. Adrian Wiater (Department of Industrial Microbiology, Lublin, Poland) for providing the mutan polymers and the mutanase enzyme.

References

Balakrishnan, M., Simmonds, R.S., Tagg, J.R., 2000. Dental caries is a preventable infectious disease. Aus. Dent. J. 45, 235œ245. Colby, S.M., McLaughlin, R.E., Ferretti, J.J., Russell, R.R.B., 1999. Effect of inactivation of gtf genes on adherence of Streptococcus downei . Oral Microbiol. Immunol. 14, 27œ32. Farré, E.M., Tiessen, A., Roessner, U., Geigenberger, P., Trethewey, R.N., Willmitzer, L., 2001. Analysis of the compartmentation of glycolytic intermediates, nucleotides, sugars, organic acids, amino acids, and sugar alcohols in potato tubers using a nonaqueous fractionation method. Plant Physiol. 127, 685œ700.

69: Chapter 3/ Production of mutan

Ferretti, J.J., Gilpin, M.L., Russell, R.R.B., 1987. Nucleotide sequence of a glucosyltransferase gene from Streptococcus sobrinus Mfe28. J. Bacteriol. 169, 4271œ4278. Garbarino, J.E., Belknap, W.R., 1994. Isolation of a ubiquitin-ribosomal protein gene ( ubi3 ) from potato and expression of its promoter in transgenic plants. Plant Mol. Biol. 24, 119œ127. Geigenberger, P., 2003. Regulation of sucrose to starch conversion in growing potato tubers. J. Exp. Bot. 54, 457œ465. Gerrits, N., Turk, S.C.H.J., van Dun, K.P.M., Hulleman, S.H.D., Visser, R.G.F., Weisbeek, P.J., Smeekens, S.C.M., 2001. Sucrose metabolism in plastids. Plant Physiol. 125, 926œ934. Haag, A.P., Maier, R.M., Combie, J., Geesey, G.G., 2004. Bacterially derived biopolymers as wood adhesives. Int. J. Adhesion Adhesives 24, 495œ502. Hamada, S., Slade, H.D., 1980. Biology, immunology, and cariogenicity of Streptococcus mutans . Microbiol. Rev. 44, 331œ384. Hennegan, K.P., Danna, K.J., 1998. pBIN20: an improved binary vector for Agrobacterium - mediated transformation. Plant Mol. Biol. Rep. 16, 129œ131. Kok-Jacon, G.A., Ji, Q., Vincken, J-P., Visser, R.G.F., 2003. Towards a more versatile S-glucan biosynthesis in plants. J. Plant Physiol. 160, 765œ777 (this thesis, Chapter 1). Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Wang, D., Liu, S., Visser, R.G.F., 2005. Production of dextran in transgenic potato plants. Transgenic Res. in press (this thesis, Chapter 2). Kossmann, J., Lloyd, J., 2000. Understanding and influencing starch biochemistry. Critical Rev. Plant Sci. 19, 171œ226. Kuipers, A.G.J., Jacobsen, E., Visser, R.G.F., 1994. Formation and deposition of amylose in the potato tuber starch granule are affected by the reduction of granule-bound starch synthase gene expression. Plant Cell 6, 43œ52. Leknes, K.N., Lie, T., 1988. Erythrosin staining in clinical disclosure of plaque. Quintessence Int. 19, 199œ204. Loesche, W.J., 1986. Role of Streptococcus mutans in human dental decay. Microbiol. Rev. 50, 353œ380. Marchessault, R.H., Deslandes, Y., 1980. Texture and crystal structure of fungal polysaccharides. ACS Symp . 126, 221œ250. Marsh, P.D., 2003. Are dental diseases examples of ecological catastrophes? Microbiol. 149, 279œ294. Moire, L., Rezzonico, E., Poirier, Y., 2003. Synthesis of novel biomaterials in plants. J. Plant Physiol. 160, 831œ839.

70: Chapter 3/ Production of mutan

Monchois, V., Willemot, R.M., Monsan, P., 1999a. Glucansucrases: mechanism of action and structure-function relationships. FEMS Microbiol. Rev. 23, 131œ151. Monchois, V., Argüello-Morales, M., Russell, R.R.B., 1999b. Isolation of an active catalytic core of Streptococcus downei Mfe28 Gtf-I glucosyltransferase. J. Bacteriol. 181, 2290œ2292. Pilon, M., Wienk, H., Sips, W., de Swaaf, M., Talboom, I., van ‘t Hof, R., de Korte-Kool, G., Demel, R., Weisbeek, P., de Kruijff, B., 1995. Functional domains of the ferredoxin transit sequence involved in chloroplast import. J. Biol. Chem. 270, 3882œ3893. Pilon-Smits, E.A.H., Ebskamp, M.J.M., Jeuken, M.J.W., van der Meer, I.M., Visser, R.G.F., Weisbeek, P., Smeekens, S.C.M., 1996. Microbial fructan production in transgenic potato plants and tubers. Ind. Crops Prod. 5, 35œ46. Robyt, J.F., 1995. Mechanisms in the glucansucrase synthesis of polysaccharides and oligosaccharides from sucrose. Adv. Carbohydr. Chem. Biochem. 51, 133œ168. Russell, R.R.B., Gilpin, M.L., Mukasa, H., Dougan, G., 1987. Characterization of glucosyltransferase expressed from a Streptococcus sobrinus gene cloned in Escherichia coli . J. Gen. Microbiol. 133, 935œ944. Salehuzzaman, S.N.I.M., Jacobsen, E., Visser, R.G.F., 1994. Expression patterns of two starch biosynthetic genes in in vitro cultured cassava plants and their induction by sugars. Plant Sci. 98, 53œ62. Sutherland, I.W., 2001. Biofilm exopolysaccharides: a strong and sticky framework. Microbiol. 147, 3œ9. Takken, F.L.W., Luderer, R., Gabriëls, S.H.E.J., Westerink, N., Lu, R., de Wit, P.J.G.M., Joosten, M.H.A.J., 2000. A functional cloning strategy, based on a binary PVX- expression vector, to isolate HR-inducing cDNAs of plant pathogens. Plant J. 24, 275œ 283. Tomás-Barberán, F.A., Espín, J.C., 2001. Phenolic compounds and related enzymes as determinants of quality in fruits and vegetables. J. Sci. Food Agric. 81, 853œ876. Tuohy, K.M., Probert, H.M., Smejkal, C.W., Gibson, G.R., 2003. Using probiotics and prebiotics to improve gut health. Drug Discovery Today. 8, 692œ700. Wenzler, H.C., Mignery, A., Fisher, L.M., Park, W.D., 1989. Analysis of a chimeric class-I patatin-GUS gene in transgenic potato plants: High-level expression in tubers and sucrose-inducible expression in cultured leaf and stem explants. Plant Mol. Biol. 12, 41œ 50. Wiater, A., Choma, A., Szczodrak, J., 1999. Insoluble glucans synthesized by cariogenic streptococci: a structural study. J. Basic Microbiol. 39, 265œ273. Wiater, A., Szczodrak, J., 2002. Selection of strain and optimization of mutanase production in submerged cultures of Trichoderma Harzianum . Acta Biol. Hung. 53, 389œ401.

71: Chapter 3/ Production of mutan

Wren, B.W., 1991. A family of clostridial and streptococcal ligand-binding proteins with conserved C-terminal repeat sequences. Mol. Microbiol. 5, 797œ803.

73: Chapter 4/ Production of alternan

Production of alternan in transgenic potato plants

Géraldine A. Kok-Jacon 1 Denong Wang 2,3 Jean-Paul Vincken 1 Shaoyi Liu 3 Luc C.J.M. Suurs 1 Richard G.F. Visser 1

1Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Wageningen University, P.O. Box 386, 6700 AJ Wageningen, The Netherlands. 2Departments of Genetics, Neurology and Neurological Sciences, Stanford University School of Medicine, Beckman Center B006, Stanford, CA 94305-5318, The United States. 3Functional Genomics Division, Columbia Genome Center, College of Physicians and Surgeons, Columbia University, 1150 St. Nicholas Avenue, New York, NY 10032, The United States.

Abstract: Alternan, which consists of alternating S-(1R3)/S-(1R6)-linked glucosyl residues, was produced in amyloplasts by expressing a mature alternansucrase ( Asr ) gene from Leuconostoc mesenteroides NRRL B-1355 in potato. Detection of alternan was performed by enzyme-linked immunosorbent assay in tuber juices, revealing a concentration between 0.3 and 1.2 mg g -1 FW; the covalent attachment of alternan to starch was not evidenced. The Asr transcript levels correlated well with alternan accumulation in tuber juices. It appeared that the expression of several sucrose-regulated starch-synthesizing genes was down-regulated. Especially for ADP-glucose pyrophosphorylase subunit S and granule-bound starch synthase I, the down-regulation was considerable in the higher and intermediate expressors. Despite this, the physicochemical properties of the transgenic starches were unaltered. These results are compared to those obtained with other transgenic potato plants expressing mutansucrase or dextransucrase.

To be submitted

Key words : alternan, glucansucrase, transgenic potato, polymer solubility, sucrose regulation

74: Chapter 4/ Production of alternan

Introduction

Production of novel polymers in plants by genetic modification is a great opportunity to obtain plants with unique properties that cannot be generated by conventional breeding (Kok-Jacon et al ., 2003). In addition, modifications of native polymers in planta could also generate crops with added nutritional, environmental or commercial value. For instance, production of biodegradable plastics in crops such as flax offers new perspectives for the replacement of oil-derived plastics (Wróbel et al ., 2004). Another example is the production of a freeze-thaw-stable potato starch exhibiting novel physicochemical properties, thereby increasing the number of industrial applications (Jobling et al ., 2002). Alternan is a unique polymer, which is produced by three known Leuconostoc mesenteroides strains: the NRRL B-1355, NRRL B-1498 and NRRL B-1501 (Jeanes et al ., 1954). Alternan synthesized by L. mesenteroides NRRL B-1355 is mediated by the alternansucrase ASR (EC 2.4.1.140), which is a large glucansucrase of 2,057 amino- acids (Argüello-Morales et al ., 2000). Its C-terminal domain (also referred to as glucan- binding domain or GBD) exhibits short repeats specific for ASR, which could contribute to its distinct features (Janećek et al ., 2000). The resulting polymer has a unique structure with alternating S-(1R3)/ S-(1R6)-linked glucose residues, present for 46 % and 54 %, respectively. Due to this structure, alternan is a highly soluble and low viscous polymer, which is resistant to microbial and mammalian enzymes making it suitable for the production of ingredients for functional foods such as prebiotics (Côté, 1992). Also, novel industrial applications were investigated by hydrolyzing native alternan polymers with isolates of Penicillium bacterial strains, creating potential replacers of commercial gum arabic (Leathers et al ., 2002; 2003). Furthermore, ASR is an attractive enzyme because of its efficiency in bond formation, which is higher than that of the dextransucrase (DSRS) (Richard et al ., 2003). In addition, mutated ASR enzymes showed a high efficiency in glucosylating acceptor molecules (cellobiose, S- alkylglucosides) in comparison to native ASR and DSRS enzymes, which might enable novel industrial applications (Argüello-Morales et al ., 2001; Richard et al ., 2003). In this work, we describe the production of alternan in potato amyloplasts by expressing ASR. Modification of starch structure was envisaged with ASR, because of its high acceptor reaction efficiency. The effect of ASR on starch biosynthesis was studied at

75: Chapter 4/ Production of alternan the microscopical, molecular and biochemical level, and compared to the effects of the dextransucrase (DSRS) and mutansucrase (GTFI), producing less soluble polymers, such as dextran and mutan (Kok-Jacon et al ., 2005a; 2005b).

Materials and methods

Construction of binary plant expression vector containing the Asr gene An expression cassette containing the patatin promoter (Wenzler et al ., 1989), the chloroplastic ferredoxin signal peptide (FD) from Silene pratensis (Pilon et al ., 1995) fused to the NOS terminator was cloned into the pBluescript SK (pBS SK) plasmid, resulting in pPF that was used as starting material for cloning the alternansucrase ( Asr ) gene. A mature Asr gene from L. mesenteroides NRRL B-1355 (Argüello-Morales et al ., 2000; AJ250173) was ligated in frame between the signal peptide FD and the NOS terminator. The mature Asr gene was amplified by PCR, with a forward primer containing a Sma I restriction site (5‘- CATCAGGGCCCCGGG GATACAAAT-3‘) and a reverse primer containing a Nru I restriction site (5‘-CTCCTTTCGCGA ATCCTTCCCTTA-3‘) using the proofreading Pfu turbo DNA polymerase

(2.5 units/ µl; Stratagene, UK) and cloned into the Sma I/ Nru I restriction sites of pPF, resulting in pPF Asr . FD and the fused Asr gene were completely sequenced in one direction by Baseclear (The Netherlands) to verify the correctness of the construct. pPF Asr was digested with Sac I and Sal I and subsequently ligated into a pBIN20 binary vector (Hennegan and Danna, 1998), resulting in pPFA (Fig. 1).

Figure 1. Schematic representation of pPFA binary vector used for potato plant transformation.

Transformation and regeneration of potato plants pPFA was transformed into Agrobacterium tumefaciens strain LBA 4404 using electroporation (Takken et al ., 2000). Internodal stem segments from the tetraploid potato genotype (cv. Kardal

76: Chapter 4/ Production of alternan

(KD)) were used for Agrobacterium-mediated transformation, which was performed as described by Kok-Jacon et al. (2005a).

Starch isolation Potato tubers were peeled and homogenized in a Sanamat Rotor (Spangenberg, The Netherlands). The resulting homogenate was allowed to settle overnight at 4°C and the potato juice was decanted and stored at œ 20°C for characterization of soluble alternan. The starch pellet was washed three times with water, air-dried at room temperature for at least three days and stored at room temperature.

Immunological detection of alternans in tuber juices and gelatinized starches Presence of alternans was investigated with enzyme-linked immunosorbent assay (ELISA) as described by Kok-Jacon et al. (2005a), using monoclonal anti-S-(1R6) dextran antibodies (45.21.1 (groove-type; IgA/Kappa) and 16.4.12E BI (cavity-type; IgA/Kappa)) (Wang et al ., 2002) with tuber juices and gelatinized starches.

Expression analysis of Asr and genes involved in starch biosynthesis using semi- quantitative and real-time quantitative RT-PCR analysis RNA was isolated from 3 g (fresh weight) of potato tuber material from selected transgenic lines according to Kuipers et al . (1994). Semi-quantitative and real-time quantitative RT-PCR‘s were performed as described by Kok- Jacon et al. (2005a). AsrRT primers, 5‘-ACCGGTTCCATCAACTAATAAT-3‘ and 5‘- GACATCTCGGAAGGATCCC- 3‘ (Tm=55°C, 35 cycles) were based on the Asr gene sequence (Argüello-Morales et al ., 2000). RNA sample from Karnico potato tubers expressing a sense/ antisense GBSSI cDNA inverted-repeat construct referred to as RVT34-77 (Heilersig et al ., unpublished results) was used as a positive control, because its GBSSI expression level was completely down-regulated.

Determination of morphological and physicochemical starch properties Analysis of starch granule morphology was performed by light microscopy (LM) and scanning electron microscopy (SEM) as described by Kok-Jacon et al. (2005a).

77: Chapter 4/ Production of alternan

Results

Detection of alternan in transgenic potato juices To enable plastidic protein targeting, the mature Asr gene was fused to the ferredoxin (FD) signal peptide (Gerrits et al ., 2001). The resulting gene fusion was inserted between the patatin promoter (Fig. 1) allowing high-tuber expression (Wenzler et al ., 1989) and the Nos terminator sequence. At the FDcAsr fusion, two mutations were present because a Sma I restriction site was engineered at this position (VTAMdATYKVTLITKcADT became VTAMdATYKVTLIT PcGDT, in which d represents the splice site for amyloplast entry and c the gene fusion). Furthermore, differences from the published ASR sequence (Argüello-Morales et al ., 2000) were found at three positions (Y 208 H, D 221 G and G 1092 S), but these did not affect conserved residues. After Agrobacterium -mediated plant transformation, thirty independent transgenic potato clones were obtained using the Kardal (KD) genotype. Five plants of each transgenic clone were grown in the greenhouse from which the tubers were pooled for further characterization. KDAxx referred to the transformed potato plant serie in which A represents the Asr gene and xx the clone number. The untransformed genotype is referred to as KD-UT.

Figure 2. Detection of alternans accumulated in potato juices by ELISA using anti-dextrans antibodies. Based on the alternan concentration (in mg g -1 FW), three categories of transformants were made, where (œ), (+) and (++) represent no, intermediate and high alternan accumulation, respectively. Tran sgenic clones indicated with grey bars were selected for further characterization.

78: Chapter 4/ Production of alternan

Detection of alternan was performed by analyzing tuber juices of the transformants with ELISA using anti-dextran antibodies (Wang et al ., 2002). Alternan was detected in 4 out of 29 tubers (about 14 %) in a concentration ranging from 0.3 to 1.2 mg g -1 FW (Fig. 2) in the transformants KDA16, KDA19, KDA27 and KDA13. As expected, no alternan was found in KD-UT plants. According to the tuber juice results, the KDA transformants were divided in three classes: (œ), (+) and (++), representing no, intermediate (≤ 1.0 mg g-1 FW) and high (> 1.0 mg g -1 FW) levels of alternan, respectively. All the transformants containing alternan and two from the (œ) class were selected for further characterization: KDA13 (++), KDA16 (+), KDA19 (+), KDA27 (+), KDA1 (œ) and KDA24 (œ). RNA was isolated from potato tubers and subjected to RT-PCR analysis. The expression levels were determined for the Asr and Ubi3 genes, of which the latter is used as a control because of its constitutive expression (Garbarino and Belknap, 1994) (Fig. 3). Heterologous Asr gene expression was detected in the expressers KDA13, KDA16, KDA19, KDA27. No Asr mRNA was detected in the (œ) class transformants and in the KD-UT plants. The Asr expression levels correlated well with the ELISA results described above.

Figure 3. RT-PCR analysis of the selected KDA transformants and KD-UT tuber RNA. The upper panel shows the PCR products using the primers designed on the Asr sequence. Th e lower panel shows the PCR products using the primers designed on the Ubi3 sequence that served as an internal control. pPF Asr plasmid: positive control.

Alternan accumulation does not interfere with plant, tuber and starch morphologies Asr expressing plants (green parts and tubers) did not exhibit any morphological changes in comparison to KD-UT plants (data not shown). In addition, starch morphology of Asr expressing plants was quite similar to that of KD-UT. With SEM, the only detected difference was the rough granule surface present on some of the (++)

79: Chapter 4/ Production of alternan class transformant granules (Fig. 4B & F) in contrast to those of the KD-UT (Fig. 4A & E). In general, starch granules from the (+) and (œ) class transformants were similar to those of the KD-UT (data not shown). Starch granules comparable to those illustrated in Fig. 4F were scored by analyzing a population of 100 granules in triplicate for each selected transformant (data not shown). KDA13, belonging to the (++) class transformant, exhibited the highest number of altered starch granules (12 % ± 1.0), followed by the (+) class transformant (KDA19 (9.3 % ± 0.6); KDA27 (8.3 % ± 0.6)). For the (œ) class transformant and KD-UT, the frequency of altered granules was lower, which was around the 7 %.

Figure 4. SEM analysis of starch granules (x 350: upper panel) and (x 1,000: lower panel) from KD-UT ( A, E) compared to that of selected transformants producing foreign polymers with decreasing water-solubility (KDA13 ( B and F; ++: highly soluble (S)), KDD30 ( C and G; +: soluble (L)) and KDIC15 ( D and H; œ: insoluble (I)). Degrees of polymer solubility were defined according to Robyt et al ., (1996) in which class S = more soluble referring to glucans precipitated by 40-44 % ethanol, L = less soluble referring to glucans precipitated by 34-37 % ethanol and I = water-insoluble.

The physicochemical properties and starch content of KDA transformants remain unchanged

Median granule size (d 50 ), gelatinization characteristics ( T0 and _H), amylose and starch content measurements were performed on selected transformants (Table 1). From these results, it can be seen that no consistent changes were detected for the different classes of transformants. Furthermore, chain length distribution experiments

80: Chapter 4/ Production of alternan

Table 1. Summary of granule size (d 50 : median value of the granule size distribution), gelatinization characteristics (T onset (T o): temperature of onset of starch gelatinization; _H: enthalpy released), amylose and starch content measurements of starches from the selected

Transformants d50 (m) T0 (°C) H (kJ/g) Amylose Starch content content (%) (mg/g FW)

KD-UT 26.5 ( ± 0.3) 67.9 ( ± 0.1) 14.5 ( ± 0.1) 22.3 ( ± 0.2) 214.8 ( ± 117.5)

KDA1 (œ) 24.4 ( ± 0.2) 68.1 ( ± 0.1) 17.0 ( ± 0.1) 22.2 (± 0.2) 103.4 ( ± 66.3)

KDA24 (œ) 25.0 ( ± 0.2) 68.0 ( ± 0.1) 16.3 ( ± 1.2) 21.3 ( ± 0.4) 86.7 ( ± 41.9)

KDA16 (+) 24.9 ( ± 0.3) 67.9 ( ± 0.2) 16.4 ( ± 1.3) 22.2 ( ± 0.1) 140.0 ( ± 88.2)

KDA19 (+) 27.9 ( ± 0.2) 67.7 ( ± 0.0) 15.2 ( ± 0.1) 23.0 ( ± 0.2) 137.1 ( ± 38.2)

KDA27 (+) 22.8 ( ± 0.7) 67.7 ( ± 0.2) 16.2 ( ± 0.5) 22.2 ( ± 0.4) 289.3 ( ± 39.7)

KDA13 (++) 24.0 ( ± 0.1) 67.8 ( ± 0.1) 16.0 ( ± 0.7) 22.2 ( ± 0.5) 107.2 ( ± 49.4)

(HPSEC and HPAEC) were also done, particularly because ASR exhibits a high acceptor reaction efficiency. After complete debranching of starch with isoamylase, no consistent changes were found with HPSEC and HPAEC in comparison to KD-UT starches (data not shown). In addition, debranched starches, which were further treated with S-amylase, were analyzed with HPAEC in order to detect the presence of novel structural elements on starch molecules such as alternating S-(1R3)/S-(1R6) linkages. Again, no consistent changes were detected with HPAEC in comparison to KD-UT starches (data not shown).

Expression levels of AGPase and GBSSI genes are down-regulated in the (+) and (++) KDA class The expression levels of key genes involved in starch biosynthesis such as sucrose synthase ( SuSy ), ADP-glucose pyrophosphorylase subunit S ( AGPase ), starch synthase III ( SSIII ), starch branching enzyme I ( SBEI ) and granule-bound starch synthase I ( GBSSI ) were monitored by real-time quantitative RT-PCR (Fig. 5). All these genes seemed to be down-regulated, particularly the AGPase and GBSSI genes. In most cases, the extent of AGPase and GBSSI down-regulation corresponded well with the amount of alternan that was accumulated in the potato tubers. The GBSSI down-

81: Chapter 4/ Production of alternan regulation was about 20 times less than for the transformant RVT34-77 in which GBSSI is completely inhibited. A partially inhibited GBSSI antisense transformant contained a comparable amount of GBSSI messenger as KDA27 (data not shown), but a lower amylose content. Typically, no reduction in amylose content was observed for the KDA transformants (Table 1), irrespective of their GBSSI messenger RNA level. Thus, the observed reduction in GBSSI expression for the (+) and (++) KDA classes were significant within the selected transformants, but relatively small with respect to the RVT34-77 transformant.

Figure 5. Real-time quantitative RT-PCR analysis of KDA24 (œ), KD A27 (+) and KDA13 (++) transformants and KD-UT tuber RNA using the following specific primers: SuSy , sucrose synthase; AGPase , ADP-glucose pyrophosphorylase subunit S; SSIII , starch synthase III; SBEI , starch branching enzyme I; GBSSI , granule-bound starch synthase I. RNA levels for each gene were expressed relative to the amount of Ubi3 RNA, as described in materials and methods. RNA sample from Karnico potato tubers expressing a sense/ antisense GBSSI cDNA construct exhibiting a complete GBSSI down-regulation (RVT34-77), was used as a positive control.

82: Chapter 4/ Production of alternan

Discussion

This report is the first study on the production of alternan in potato tubers. Their presence in potato juices was demonstrated by ELISA using anti-dextran antibodies. Expression of ASR did not interfere with plant growth and development, and tuber and starch yield penalties were not observed. These results were similar to those obtained with the dextransucrase (DSRS) expression (Kok-Jacon et al. , 2005a), but not to those obtained with the mutansucrase (GTFI) expression in which the tuber phenotype was significantly affected (Kok-Jacon et al., 2005b). The amount of alternan accumulated in potato tubers (1.2 mg g -1 FW) was lower than that of dextran (1.7 mg g -1 FW) (Kok-Jacon et al. , 2005a). The lower accumulation of alternan could be due to the large size of ASR, which might reduce the efficiency with which the enzyme is transported through the amyloplast membrane. The mature ASR consists of 2,057 amino-acids, whereas DSRS has only 1,527. It has been shown that the size of ASR can be reduced (by removal of the C-terminal GBD) without compromising its activity and its ability to glycosylate acceptor molecules (Joucla et al ., 2001). If the size of the protein is indeed a critical factor, than this truncation may offer opportunities for enhancing alternan synthesis in the amyloplast. Such an approach was already employed successfully for the Streptococcus downei mutansucrase GTFI (Kok-Jacon et al., 2005b). We have directed a mature and a GBD-truncated GTFI protein to potato amyloplasts, and found that the truncated form synthesized a larger amount of mutan, and had a much more pronounced effect on starch granule morphology. Although ASR is known to be efficient in catalyzing acceptor reactions (Richard et al ., 2003), no evidence was found for the covalent attachment of novel, alternan-based structural elements to starch molecules. Also with dextransucrase and mutansucrase we have not been able to introduce different glycosyl linkage patterns in starch (Kok- Jacon et al., 2005a; 2005b). To this end, acceptor reactions of glucansucrases with starch or maltodextrins are not studied in much detail. It has been observed that the efficiency of acceptor reaction decreases with increasing length of maltodextrins (reviewed in Kok-Jacon et al ., 2003). The acceptor reactions seem to follow a general pattern in that glucosyl residues can only be attached to either the reducing or non- reducing terminus of the acceptor molecule, and not to a glucosyl residue with a mid-

83: Chapter 4/ Production of alternan chain position. Only those molecules having a glucosyl residue attached to the C-6 of the non-reducing Glc of the maltodextrin acceptor can serve as an acceptor molecule in subsequent acceptor reactions, but probably this depends on the glucansucrase used. We had anticipated that the nascent starch polymers would be poor acceptors for the glucansucrases. However, during starch biosynthesis, potential acceptors (small maltodextrins) are thought to be generated through the action of, for instance, debranching enzymes (or isoamylases). If such a small acceptor is mutanylated, alternanylated, or dextranylated at the non-reducing end, then these novel structures might be incorporated into starch polymers through the action of certain transferases such as, for instance, branching enzyme. Apparently, this does not happen, or at a very low (undetectable) frequency, but the reason for this is unclear. Starch morphology in the ASR transformants was not significantly altered in comparison to that of dextran and mutan-accumulating plants (Fig. 4). This might be related to the fact that alternan is more water-soluble than dextran and mutan. An indication of the water-solubility of the three polysaccharides is given in Fig. 4; the more ethanol is required for precipitation, the higher the water-solubility. The water-solubility decreases in the order of alternan, dextran and mutan. We hypothesize that the co- synthesis of water-insoluble mutan and starch leads to co-crystallization of the two polymers, as a result of which the granule is packed in a less orderly fashion. This comparison should be approached with caution. For alternan and dextran, the observed differences in starch morphology may also be related to the fact that more dextran than alternan was accumulated in the potato tubers; for mutan, we have not been able to quantify the amount accumulated in the tubers. Therefore, it can not be excluded that the observed effects are related to the amount of foreign polymer produced. Interestingly, co-synthesis of levan, a water-soluble fructosyl-based polymer, and starch resulted in a dramatically altered starch granule morphology (Gerrits, 2000). However, it should be noted that much higher levels of levan which were estimated to be 66 mg.g -1 FW (Gerrits et al ., 2001; Cairns et al ., 2003) were produced in comparison with alternan (1.2 mg g -1 FW) or dextran (1.7 mg g -1 FW), and that the starch granules contained approximately 5% of levan. This result contrasts with that of alternan- and dextran-accumulating plants in which foreign polymers were only found in the stroma. Taking together the results of potato transformants expressing glucan- or levansucrases in amyloplasts, it seems that the site of accumulation of the

84: Chapter 4/ Product ion of alternan foreign polymer (granule or stroma), the solubility of the foreign polymer, and the amount of foreign polymer that is actually produced are important factors in determining starch granule morphology.

Acknowledgements

References

85: Chapter 4/ Production of alternan

Jeanes, A., Haynes, W.C., Wilham, C.A., Rankin, J.C., Melvin, E.H., Austin, M.J., Cluskey, J.E., Fisher, B.E., Tsuchiya, H.M., Rist, C.E., 1954. Characterization and classification of dextrans from ninety-six strains of bacteria. J. Am. Chem. Soc. 76, 5041œ5052. Jobling, S.A., Westcott, R.J., Tayal, A., Jeffcoat, R., Schwall, G.P., 2002. Production of a freeze-thaw-stable potato starch by antisense inhibition of three starch synthase genes. Nature Biotechnol. 20, 295œ299. Joucla, G., Argüello-Morales, M.A., Pizzut, S., Willemot, R-M., Monsan, P., Remaud-Simeon, M., 2001. Sequence analysis of alternansucrase gene from Leuconostoc mesenteroides NRRL B-1355 and expression of entire and genetically truncated forms in Escherichia coli . In: Proceedings 4th Carbohydrate Bioengineering Meeting, Stockholm, Sweden. Kok-Jacon, G.A., Ji, Q., Vincken, J-P., Visser, R.G.F., 2003. Towards a more versatile S-glucan biosynthesis in plants. J. Plant Physiol. 160, 765œ777 (this thesis, Chapter 1). Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Wang, D., Liu, S., Visser, R.G.F., 2005a. Production of dextran in transgenic potato plants. Transgenic Res. in press (this thesis, Chapter 2). Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2005b. Mutan produced in potato amyloplasts adheres to starch granules. Plant Biotechnol. J. 3, 341œ351 (this thesis, Chapter 3). Kuipers, A.G.J., Jacobsen, E., Visser, R.G.F., 1994. Formation and deposition of amylose in the potato tuber starch granule are affected by the reduction of granule-bound starch synthase gene expression. Plant Cell 6, 43œ52. Leathers, T.D., Nunnally, M.S., Côté, G.L., 2002. Modification of alternan by novel Penicillium spp. J. Ind. Microbiol. Biotechnol. 29, 177œ180. Leathers, T.D., Nunnally, M.S., Ahlgren, J.A., Côté, G.L., 2003. Characterization of a novel modified alternan. Carbohydr. Polym. 54, 107œ113. Pilon, M., Wienk, H., Sips, W., de Swaaf, M., Talboom, I., van ‘t Hof, R., de Korte-Kool, G., Demel, R., Weisbeek, P., de Kruijff, B., 1995. Functional domains of the ferredoxin transit sequence involved in chloroplast import. J. Biol. Chem. 270, 3882œ3893. Richard, G., Morel, S., Willemot, R.M., Monsan, P., Remaud-Simeon, M., 2003. Glucosylation of S-butyl- and S-octyl-D-glucopyranosides by dextransucrase and alternansucrase from Leuconostoc mesenteroides . Carbohydr. Res. 338, 855œ864. Robyt, J.F., 1996. Mechanism and action of glucansucrases, in Enzymes for carbohydrate engineering, Elsevier, Amsterdam, Park, K.H., Robyt, J.F., Choi, T.D. (Eds.) pp. 1œ22. Takken, F.L.W., Luderer, R., Gabriëls, S.H.E.J., Westerink, N., Lu, R., de Wit, P.J.G.M., Joosten, M.H.A.J., 2000. A functional cloning strategy, based on a binary PVX-

86: Chapter 4/ Production of alternan

expression vector, to isolate HR-inducing cDNAs of plant pathogens. Plant J. 24, 275œ 283. Wang, D., Liu, S., Trummer, B.J., Deng, C., Wang, A., 2002. Carbohydrate microarrays for the recognition of cross-reactive molecular markers of microbes and host cells. Nature Biotechnol. 20, 275œ281. Wenzler, H.C., Mignery, A., Fisher, L.M., Park, W.D., 1989. Analysis of a chimeric class-I patatin-GUS gene in transgenic potato plants: High-level expression in tubers and sucrose-inducible expression in cultured leaf and stem explants. Plant Mol. Biol. 12, 41œ 50. Wróbel, M., Zebrowski, J., Szopa, J., 2004. Polyhydroxybutyrate synthesis in transgenic flax. J. Biotechnol. 107, 41œ54.

87: Chapter 5/ Production of mutan and SBD technology

Granule-bound mutansucrase alters melting temperature of starch granules

Géraldine A. Kok-Jacon Luc C.J.M. Suurs Jean-Paul Vincken Richard G.F. Visser Qin Ji

Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Wageningen University, P.O. Box 386, 6700 AJ Wageningen, The Netherlands.

Abstract : It has been shown previously that mutan can be co-synthesized with starch when a truncated mutansucrase (GTFICAT) is directed to potato tuber amyloplasts. The mutan seemed to adhere to isolated starch granules, but it was not incorporated in the starch granule. In this study, GtfICAT was fused to the N- or C-terminus of a starch-binding domain (SBD). These constructs were introduced in potato plants (cv. Kardal), in order to bring GTFICAT in more intimate contact with growing starch granules, and to facilitate the incorporation of mutan polymers in starch. For the high SBD-GTFICAT expressors, it was found that mutan was present inside starch granules, demonstrating that granule-targeting of GTFICAT was successful. Interestingly, the granules of these transformants had spongious surfaces, a higher melting temperature, and a more pronounced retrogradation behaviour, compared to those from controls. Except for the T onset, these alterations were less pronounced than those observed in transformants expressing GtfICAT gene without appended SBD. In vitro production of mutan by incubating starch granules from transformants with an excess of sucrose was not evidenced. Our results show that expression of granule-bound and —soluble“ GTFICAT can affect starch biosynthesis differently, and that the appended SBD inhibits the activity of GTFICAT in the engineered fusion protein.

To be submitted

Key words : mutan, glucansucrase, starch-binding domain, granule-boundness, transgenic potato

88: Chapter 5/ Production of mutan and SBD technology

Introduction

In a previous study, we have shown that expression of the truncated mutansucrase gene GtfICAT ( GtfI without a glucan-binding domain) in potato amyloplasts led to a more efficient production of mutan in comparison to that of the full-length GtfI (Kok- Jacon et al ., 2005a). Mutan production by GTFICAT was accompagnied with pronounced morphological and physicochemical alterations at the tuber and starch levels. However, from this study, the presence of mutan inside starch granules was not evidenced. We hypothesized that even more dramatic effects on starch granule morphology and properties might be obtained, if mutan could be incorporated in the starch granule. This might be achieved by bringing GTFICAT and the growing starch granule in intimate contact with each other. For this purpose, we have engineered two genes in which a starch-binding domain (SBD) was fused to either the 5' or the 3' end of GTFICAT. It has been shown before that SBD is an efficient tool for targeting effector proteins to the growing starch granule (Ji et al., 2003; 2004b). The use of SBD technology may offer three advantages. (i) Mutan will be produced close to the granule surface, which will increase the chance that it co-crystallizes with starch during granule formation. (ii) The more intimate contact between GTFICAT and starch may facilitate the so-called acceptor reaction of the enzyme (Monchois et al ., 2000), which in turn may lead to covalent attachment of mutan to starch. (iii) Mutan may be produced post- harvest by supplying sucrose, a cheap and abundant substrate, to transgenic starch granules containing the fusion proteins (Kok-Jacon et al ., 2003). In this study, we investigate if starch granule properties can be more severely affected with a granule-bound GTFICAT than with a "soluble" GTFICAT. For this, SBD was fused to the N- or C-terminus of GTFICAT, since our previous studies have pointed out that the position of the SBD in the fusion protein can influence the activity of the appended effector. Although previous studies (Ji et al ., 2003; 2004b) have shown that accumulation of SBD was most efficient in the amylose-free ( amf ) potato background, we have chosen to express the engineered genes in the Kardal genotype. In this way, a direct comparison with the GTFICAT potato transformants that had already been obtained is possible (Kok-Jacon et al ., 2005a). Post-harvest experiments with transgenic granules and sucrose were also performed, in order to alter the polymer composition of the granules by the in vitro production of mutan.

89: Chapter 5/ Production of mutan an d SBD technology

Materials and methods

Preparation of constructs containing the GtfICAT gene and the SBD fragment cloned at the N- and C-terminal positions pPF and pPF GtfICAT (Kok-Jacon et al ., 2005a) were used as starting material for cloning the SBD-linker and linker-SBD fragments resulting in pPF SBDGtfICAT and pPF GtfICATSBD, respectively. The SBD-linker and linker-SBD fragments were obtained from the pTrcHisB/SBD2 plasmid (Ji et al ., 2004a) that was used as a template for PCR amplification. The SBD fragment originated from the CGTase gene of Bacillus circulans strain 251 (Lawson et al ., 1994) and the linker fragment is similar to the Pro-Thr-rich linker of the Cellulomonas fimi exoglucanase ( Cex ) (Gilkes et al ., 1991). For the construction of pPF SBDGtfICAT , the SBD-linker fragment was obtained by PCR amplification with a forward primer, containing a Sca I restriction site (5'- AGTACT ATGGCCGGAGATCAGGTC- 3'), and a reverse primer, containing a Nru I restriction site (5'- TCGCGA CGACGGGGTC- 3'), and cloned into the Sma I restriction site of pPF GtfICAT .

Figure 1 . Schematic representation of pPFSBDIC ( A) and pPFICSBD ( B) binary vectors used for potato plant transformation.

For the construction of pPF GtfICATSBD , the linker-SBD fragment was obtained by PCR amplification with a forward primer, containing an EcoR V restriction site (5'- GATATC TCCGACGCCGACGC- 3'), and a reverse primer, containing a Sma I restriction site (5'- CCCCGGG ATCCACCAAAAC- 3'), and cloned into the EcoR V restriction site of pPF resulting in pPFSBD. In order to remove the stop codon in the original GtfICAT sequence, the GtfICAT fragment was amplified with a forward primer, containing a Sma I restriction site (5'- CCCGGG ACTGAAAACTGTTAG- 3'), and a reverse primer, containing a Nru I restriction site

90: Chapter 5/ Production of mutan and SBD technology

(5'- TCGCGA ACATTGAGGTACTTG- 3'), and cloned into the Sma I restriction site of pPFSBD, resulting in pPF GtfICATSBD . pPF SBDGtfICAT and pPF GtfICATSBD were completely sequenced in one direction by Baseclear (The Netherlands) to verify the correctness of the construct. pPF SBDGtfICAT and pPF GtfICATSBD were digested with Xho I and ligated into a pBIN20 binary vector (Hennegan and Danna, 1998), resulting in pPFSBDIC and pPFICSBD (Fig. 1).

Transformation and regeneration of potato plants pPFSBDIC and pPFICSBD were transformed into Agrobacterium tumefaciens strain LBA 4404 using electroporation (Takken et al ., 2000). Internodal stem segments from the tetraploid potato genotype (cv. Kardal (KD)) were used for Agrobacterium-mediated transformation which was performed as described by Kok-Jacon et al . (2005b).

Starch isolation Starch isolation was performed as described by Kok-Jacon et al . (2005a).

Expression analysis of GtfICAT and genes involved in starch synthesis, using semi- quantitative and real-time quantitative RT-PCR analysis RNA was isolated from 3 g (fresh weight) of potato tuber material from selected transgenic lines according to Kuipers et al. (1994). Semi-quantitative and real-time quantitative RT-PCR were performed as described by Kok- Jacon et al. (2005b). GtfIRT primers, 5‘-CCGTGCTTACAGTACCTCAGC-3‘ and 5‘- GGTCGTTAGCATTGTAGGTGAAA-3‘ (Tm=59°C, 35 cycles) were based on the GtfI gene sequence (Ferretti et al ., 1987). RNA sample from the RVT34-77 transformant was used as described by Kok-Jacon et al . (2005a).

91: Chapter 5/ Production of mutan and SBD technology

Determination of SBD content of transgenic starches by dot blot analysis Dot blot analysis was performed according to the method described by Ji et al. (2003). A 12.5 % sodium dodecylsulphate-polyacrylamide gel (50 mm þ 50 mm þ 3 mm), with nine equally spaced holes (9 mm diameter), was placed in contact with a similarly sized Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech, UK). Twenty milligrams of (transgenic) starch was boiled for 5 min with 200 bl of 2 þ SDS sample buffer (Murashige and Skoog, 1962). After cooling to room temperature, the starch gel was transferred into one of the holes. SBD proteins from transgenic starch gels were blotted onto the membrane with PhastSystem (Pharmacia, Sweden; 20 V, 25 mA, 15°C, 45 min). The blot was incubated overnight in a 1 % blocking solution (10 ml 10 þ western blocking reagent; Roche, Germany) in 90 ml TBS (20 mM Tris, 500 mM NaCl pH 7.5) at room temperature. Subsequently, the blot was washed in TTBS (0.05 % Tween-20 in TBS) for 5 min, and incubated for 2 h at room temperature with a 1:500 dilution of the primary antibody (antiSBD) in a 0.96 % blocking solution in TTBS. After this, the blot was washed twice in TTBS for 5 min, and incubated for 1 h at room temperature with a 1:2000 dilution of Goat Anti-Rabbit IgG (H + L) Alkaline Phosphatase Conjugate (BioRad, USA) in a 0.64 % blocking solution in TTBS. The blot was washed twice in TTBS, and once in TBS for 5 min. Finally, the blot was stained with a 0.1 M

NaHCO 3 solution pH 9.8 containing 1 % NBT/BCIP (Roche Molecular Biochemicals, Germany), and 0.01 M MgCl 2.

Post-harvest experiments Post-harvest experiments were performed using 10 mg of starch, which was incubated for 66 h in 1 ml of 50 mM Tris/HCl pH 7.0 and 1 M sucrose at 37°C and 45°C. After centrifugation (1 min; 10,000 g), the supernatant was submitted to HPAEC analysis. HPAEC was performed as described previously except that the column temperature was 28°C and a different gradient was used. Three eluents were used, eluent A (100 mM NaOH), eluent

B (1 M NaAc in 100 mM NaOH) and eluent C (H 2O) as follows: 0R 12 min (linear gradient 25 to 85 % eluent A; 75 to 15 % eluent C); 12R 25 min (linear gradient 0 to 10 % eluent B; 15 to 5 % eluent C); 25 R 25.1 min (linear gradient 85 to 0 % eluent A; 10 to 100 % eluent B; 5 to 0 % eluent C); 25.1 minR 30 min (100 % eluent B; rinsing phase); 30R 45 min (0 to 25 % eluent A; 100 to 0 % eluent B; 0 to 75 % eluent C; equilibration). The eluents were monitored by an ED40 electrochemical detector in the pulsed amperometric mode (Dionex).

92: Chapter 5/ Production of mutan and SBD technology

Results

Expression of SBD-GtfICAT results in altered tuber phenotype The SBD-GtfICAT and GtfICAT-SBD fragments were cloned in frame to the plastidic protein targeting sequence (FD) (Gerrits et al ., 2001), that was driven by the highly tuber-expressed patatin promoter (Wenzler et al ., 1989) resulting in the pPFSBDIC and pPFICSBD constructs (Fig. 1). These constructs were used for Agrobacterium - mediated transformation of potato cv. Kardal (KD) from which thirty independent transgenic potato clones were obtained that were named KDSICxx and KDICSxx, respectively. SIC and ICS represent the SBD-GtfICAT and GtfICAT-SBD genes, respectively, and xx the clone number. The untransformed genotype is referred to as KD-UT. During growth, the transgenic plants were morphological similar to the controls (data not shown).

Figure 2 . RT-PCR analysis of the selected KDSIC and KDICS transformants, and KD-UT. The upper panel shows the PCR products using the primers designed on the GtfI sequence and the lower panel, the PCR products using the primers designed on the Ubi3 sequence that served as an internal control.

GtfICAT expression was monitored by RT-PCR analysis, from which a number of transformants were selected and divided in classes, based on the band intensity of the PCR products. In parallel, the ubiquitin-ribosomal gene expression ( Ubi3 ), which is known as constitutive (Garbarino and Belknap, 1994), was used as an internal control. The (œ), (+) and (++) classes were defined as no, intermediate and high levels of

93: Chapter 5/ Production of mutan and SBD technology mRNA, respectively. From the KDSIC series, the KDSIC30 (œ), KDSIC15 (+), KDSIC19 (+), KDSIC1 (++), KDSIC2 (++) and KDSIC10 (++) transformants were selected for further characterization (See Fig. 2A). In addition, transformants from the KDICS series such as KDICS10 (œ), KDICS25 (œ), KDICS5 (+), KDICS28 (+), KDICS4 (++) and KDICS27 (++), were also selected (See Fig. 2B). Due to the low Ubi3 intensity of KDICS5, this transformant could be also assigned to the (++) class. As shown in Figure 2, no GtfICAT mRNA was detected in the KD-UT plants. Tuber phenotype of the (++) (Fig. 3) and (+) classes (data not shown) of the KDSIC series was severely affected, showing brownish/reddish discolourations. Concerning the (œ) class of transformants, the tuber phenotype was comparable to that of KD-UT. No phenotypical differences were observed for tubers of the KDICS series, which was comparable to KD-UT. In general, tuber number and yield were unchanged for the KDSIC and KDICS series, and were similar to KD-UT.

Figure 3. Tuber phenotype of KDSIC1 (++) and KDIC15 (++) transformants compared to that of KD-UT.

Accumulation of SBD-containing protein in the KDSIC10 transformant Accumulation of SBD-containing protein in the starch granules was monitored by Western dot blot analysis, using gelatinized starches from all the selected transformants. Determination of SBD protein concentration was performed by using calibrated positive controls from the (+), (++) and (+++) classes of the amf S series of transformants described by Ji et al . (2003). Subsequently, the dot intensities from all the gelatinized starches were compared to these positive controls. The KDSIC10 transformant exhibited a dot intensity comparable to that of the (+) class (data not

94: Chap ter 5/ Production of mutan and SBD technology shown), showing that SBD proteins accumulated inside or at the starch granule surfaces. Concerning the other selected transformants, the dot intensity was comparable to that of KD-UT, in which accumulation of SBD proteins was too low to be detected or absent (data not shown).

Detection of mutan in only the (++) and (+) classes of KDSIC transformants As described previously (Kok-Jacon et al ., 2005a), mutan can be visualized using an erythrosine red dye (Fig. 4). This staining was also used on a positive control,

Figure 4. LM analysis of starch granules (x 800) stained with Disclosing solution red-Cote from KD-UT ( A) and KDSIC10 ( C) compared to that of mutan ( B). Treatment of KDSIC10 ( D) and KDIC15 ( E) starch granules with 25 mU of mutanase enzyme. consisting of purified mutan (Wiater et al ., 1999) (Fig. 4B). A number of starch granule surfaces from the (++) class, illustrated in Figure 4C, and (+) classes (data not shown) of the KDSIC series coloured red, which demonstrated the presence of mutan. Starch granules from the (œ) class transformants did not stain with the erythrosine dye (data not shown) similar to those of KD-UT (Fig. 4A). For none of the KDICS starch granules, a red colouration was observed upon erythrosine treatment (data not shown). KDSIC10 starch granules were exhaustively treated with an exo-mutanase solution in order to

95: Chapter 5/ Production of mutan and SBD technology detach the mutan from the granule surface. From previous results (Kok-Jacon et al ., 2005a), it was shown that the 48 h incubation time was sufficient for detaching mutan from the granule surfaces (Fig. 4E). From Figure 4D, it can be seen that most of the mutan remained attached to the granule surface, which suggests that a proportion of the mutan is present inside the starch granules, and therefore inaccessible to the exo- mutanase.

Altered granule phenotype in the KDSIC series SEM analysis was performed on the selected transgenic starches. With SEM, the presence of altered starch granules was visualized in the (++) class of the KDSIC series (Fig. 5: B-F-G-H) in contrast to the (œ) class (data not shown), which was comparable to that of KD-UT (Fig. 5A). From Figure 5 (F, G and H), it can be seen that surfaces of starch granules were spongy and irregular. In addition, small protrusions on the starch granule were also present. Such granule phenotypes were not observed in the (++) class of the KDIC transformants (Fig. 5D and I) demonstrating that N-terminal

Figure 5 . SEM analysis of starch granules (x 350) from KD-UT ( A) compared to that of selected transformants (KDSIC10 ( B), KDICS27 ( C), KDIC15 ( D)) and examples of starch granules (x 1,000) with altered morphology (KDSIC10 ( F), KDSIC10 ( G), KDIC15 ( I) and KDSIC10 (x 5,000) ( H)) compared to KD -UT ( E). fusion of SBD to the GtfICAT gene interferes differently with starch granule morphology. For the (++) class of the KDICS series (Fig. 5C), starch granules were not

96: Chapter 5/ Production of mutan and SBD technology significantly altered, and similar to those of KD-UT. From the SEM pictures, the number of altered granules was scored by counting a population of 100 starch granules in triplicate for one transformant of the (œ) and two transformants of the (++) class of each series. This quantification is shown in Figure 6 for the KDSIC30 (œ), KDSIC1 (++), KDSIC10 (++) and KDICS25 (œ), KDICS4 (++), KDICS27 (++) transformants. The

Figure 6. Percentage of starch granules with altered morphology comparable to those of Figure 5 ( F, G and H) in selected transformants and KD-UT plants.

highest number of altered starch granules was found in the (++) class of the KDSIC series, ranging from 16.7 ± 3.1 for KDSIC1 to 24.7 ± 3.8 for KDSIC10. In the (++) class of the KDICS series, the number of altered starch granules was similar as in KD-UT (5.7 ± 0.6), ranging from 9.3 ± 0.6 for KDICS4 to 10.7 ± 2.3 for KDICS27. For the (œ) class of transformants, the frequency of altered starch granules was about 7 %. In general, it seems that the SBD-GTFICAT fusion protein affects the starch granule phenotype more dramatically than GTFICAT-SBD. However, in the first case, the starch granule phenotype was less affected than that of (++) KDIC transformants in which 31.3 ± 2.3 % of the starch granules exhibited altered morphologies (Kok-Jacon et al ., 2005a).

SBD-GTFICAT expression alters viscosity profile

Median granule size (d 50 ), amylose and starch contents of the transgenic starches were determined, from which no significant differences were found in comparison to KD-UT (Table 1). A direct correlation between tuber browning and decreased starch content

97: Chapter 5/ Production of mutan and SBD technology

Table 1: Summary of SBD expression (classification according to Ji et al. , 2003), granule size (d 50 : median value of the granule size distribution), gelatinization characteristics ( To = T onset: temperature of onset of starch gelatinizati on), amylose and starch contents measurements of starches from the KDSIC10 (++), KDICS27 (++), KDIC15 (++; Kok-Jacon et al ., 2005a) transformants and KD-UT. Data (± SD) are the average of two or three independent measurements.

Transformants SBD d50 (m) T0 (°C) Amylose Starch content expression content (%) (mg/g FW) KD-UT œ 28.8 ( ± 0.5) 66.7 ( ± 0.1) 20.1 ( ± 0.5) 242.9 ( ± 93.5) KDSIC10 (++) + 25.1 ( ± 0.8) 69.5 ( ± 0.1) 18.0 ( ± 0.2) 179.7 ( ± 38.3) KDICS27 (++) œ 28.6 ( ± 0.2) 67.4 ( ± 0.1) 20.0 ( ± 0.2) 191.6 ( ± 29.4) KDIC15 (++) œ 20.2 ( ± 0.4) 67.2 ( ± 0.1) 19.7 ( ± 0.2) 81.3 ( ± 37.3) could not be found, in contrast to transformants only expressing GTFICAT (Kok-Jacon et al ., 2005a). Changes in the temperature of onset of starch gelatinization (T onset

(T0)) were only detected for the KDSIC10 transformant (Table 1), exhibiting a T onset of 69.5°C which is about 3°C higher than that of KD-UT (66.7°C). Viscosimetric analysis showed that the KDSIC10 profile (Fig. 7 (2)) was different from that of KDICS27 (Fig. 7

Figure 7. Viscosity profiles from KDIC15 (++) (1), KDSIC10 (++) (2) and KDICS27 (++) (3) transformants compared to that of KD-UT (4).

(3)) and KD-UT (Fig. 7 (4)). The T onset was about 3°C higher than that of KD-UT and KDICS27, which is in accordance with the DSC results. In addition, the end viscosity

98: Chapter 5/ Production of mutan and SBD technology was about 1.5-fold higher in comparison to KD-UT and KDICS27, but not as high as that of KDIC15 (Fig. 7 (1)). These results illustrate that N-terminal fusion of SBD to GtfICAT induced more severe effects in the physicochemical properties of the starch compared to the C-terminal SBD fusion, but not as pronounced as GTFICAT alone. It should be noted that some variation in the values for peak viscosity was observed, in contrast to those for end viscosities, which were very consistent. Within the KDSIC series, also the starches from KDSIC30 (-), KDSIC1 (++), and KDSIC2 (++) were subjected to viscosimetric analysis. KDSIC30 starch showed a behaviour similar to KD- UT starch, whereas KDSIC1 and KDSIC2 starch gave a pattern intermediate between KDSIC30 and KDSIC10. This is consistent with the results from RT-PCR and Western blot analysis.

GTFICAT with appended SBD does not seem to alter starch fine structure The chain length distribution was determined in order to detect deviations in starch structure, which may indicate the presence of S-(1R3)-linked glucose residues. After complete debranching of starch with isoamylase, no consistent changes were detected with HPSEC and HPAEC in comparison to KD-UT starches (data not shown). In addition, isoamylase debranched starches were further incubated with S-amylase and analyzed with HPAEC in order to detect the presence of possible S-(1R3) linkages. Despite the pronounced alterations of the rheological properties, in particular for KDSIC10, no consistent changes in amylopectin fine structure were detected in comparison to KD-UT starches (data not shown). These results suggest that mutan is not covalently attached to the starch polymers, but rather is present as a separate carbohydrate.

Expression levels of AGPase and GBSSI genes are down-regulated in the (++) KDSIC class and that of SuSy in the (++) KDICS class As described previously (Chapters 2, 3 and 4), key enzymes involved in starch biosynthesis such as sucrose synthase (SuSy), ADP-glucose pyrophosphorylase (AGPase), starch synthase III (SSIII), starch branching enzyme I (SBEI) and granule- bound starch synthase I (GBSSI) were selected and their expression levels were monitored by real-time quantitative RT-PCR (Fig. 8). In the KDSIC series, the expression levels of AGPase and GBSSI genes were down-regulated in the (++) class

99: Chapter 5/ Production of mutan and SBD technology

Figure 8. Real-time quantitative RT-PCR analysis of KDSIC30 (œ), KDSIC1 (++), KDSIC10 (++), KDICS25 (œ), KDICS4 (++), KDICS27 (++) transformants and KD-UT tuber RNA. Expression levels of the following starch synthesizing genes are indicated: SuSy , sucrose synthase; AGPase , ADP-glucose pyrophosphorylase subunit S; SSIII , starch synthase III; SBEI , starch branching enzyme I; GBSSI , granule-bound starch synthase I. RVT34-77, RNA from Karnico potato tubers expressing a sense/antisense GBSSI cDNA construct leading to complete GBSSI gene down-regulation. RNA levels for each gene were expressed relative to the amount of Ubi3 RNA, as described in materials and methods. in contrast to the KDICS transformants and KD-UT. For GBSSI , this down-regulation was 20 times less than for a completely down-regulated GBSSI transformant named RVT34-77. Typically, the SuSy expression in both high GTFICAT-SBD expressors of the KDICS series was down-regulated, contrary to that in the KDSIC series.

100: Chapter 5/ Production of mutan and SBD technology

Post-harvest mutan synthesis with GTFICAT-containing starch granules In an effort to produce more mutan, transgenic starch granules were incubated with an excess of sucrose at 37°C, which is the optimal temperature for GTFICAT, and at 45°C. For this, starch granules from KDSIC10 were used because of their known SBD accumulation inside starch granules, together with the high GtfICAT expression in potato tubers. KDIC15, which is a high expressor (Kok-Jacon et al ., 2005a), was selected as a control to investigate if granules from this plant can produce mutan post- harvest. Another reason for selecting granules of KDSIC10 and KDIC15 was their spongious appearance (see Fig. 5), which might facilitate the diffusion of sucrose to the granule interior. After 66 h of incubation, the production of fructose and glucose was determined by HPAEC. The release of fructose is indicative for the amount of GTFICAT activity, whereas the release of glucose is not; glucose can either be released as such (hydrolytic activity of the enzyme) or as part of the mutan that is formed (polymerizing activity of the enzyme). From Table 2, it can be seen that GTFICAT was active, but at a low level. After 66 h, fructose was released in a higher amount for the KDIC15 (33 µg/ml at 45°C) transformant than for KDSIC10 (20 µg/ml at 45°C), in contrast to KD-UT for which no fructose was released at all. The glucose concentration was similar to that of fructose, suggesting that the enzyme inside starch granules mainly catalyzes a hydrolysis reaction and no polymerization. This is in line with our observations that no increased amounts of mutan were visualized upon light microscopy analysis of transgenic starch granules, which were stained with erythrosine after 66 h of incubation with sucrose.

Table 2: Summary of post-harvest experiments from KDSIC10, KDIC15 and KD-UT starches by measuring the release of fructose and glucose concentrations in µg/ml by HPAEC at 37°C and 45°C.

Transformants Fructose (µg/ml) Glucose (µg/ml)

37 °C 45°C 37 °C 45°C KD-UT 0 0 0 6

KDSIC10 (++) 10 20 7 16

KDIC15 (++) 13 33 11 28

101: Chapter 5/ Production of mutan and SBD technology

Discussion

In this study, a microbial SBD was fused to the N- and C-terminal parts of GTFICAT in order to bring this effector enzyme in more intimate contact with starch granules. An appended SBD at the N-terminal part of GTFICAT was the most efficient way to produce mutan inside starch granules. Besides its presence on granule surfaces, it was also trapped within the granule matrix, as visualized after exhaustive treatment with exo-mutanase. This result differed from that obtained with GTFICAT alone (Kok-Jacon et al ., 2005a), showing that a SBD fused to the N-terminus of GTFICAT can target the enzyme to starch granules. In contrast, SBD fusion to the C-terminus of GTFICAT did not promote the binding of GTFICAT to starch granules and the production of mutan, because the obtained results were very similar to those of untransformed plants. Ji et al . (2004b) also indicated that it is difficult to predict the prefered position for SBD in a fusion protein beforehand, and that both the N- and C-terminal position should be investigated. Several results showed that GTFICAT with an appended SBD had less pronounced effects on starch biosynthesis than GTFICAT alone. (i) Granule morphology was more affected in KDIC starch granules (Fig. 5I and Kok-Jacon et al ., 2005a) than in those of KDSIC and KDICS. (ii) The end viscosity of starch pastes was higher for the selected KDIC15 transformant in comparison to that of KDSIC and KDICS transformants (Fig. 7). (iii) Expression of the AGPase and GBSSI genes were more significantly down- regulated in the KDIC (++) than in the KDSIC (++) and KDICS (++) transformants. This is probably due to a higher sucrose conversion rate with GTFICAT alone, by which the expression level of the sucrose-regulated AGPase and GBSSI genes is more affected (Geigenberger, 2003; Salehuzzaman et al ., 1994). In accordance with these results, the starch content of the tubers from the KDIC (++) transformants was lower than that from the KDSIC (++) and KDICS (++) transformants. (iv) The post-harvest experiments demonstrated that granules from KDIC (++) were more active in presence of excess sucrose than those from KDSIC (++). In itself this was an unexpected observation, since GTFICAT is thought not to have an affinity for the starch granule of its own. However, it is possible that plastidial enzymes are coincidently entrapped in the granule matrix, as has been shown for other —soluble“ enzymes such as the potato starch synthase III, that are involved in starch biosynthesis (Marshall et al ., 1996).

102: Chapter 5/ Production of mutan and SBD technology

The less pronounced effects of GTFICAT with appended SBD in comparison with GTFICAT alone might be explained by hypothesizing either that the appended SBD hinders accumulation of fusion protein in the amyloplast, or that the appended SBD reduces the activity of the enzyme. Although the former explanation can not be excluded, we favour the latter, particularly because this is most consistent with the post-harvest experiments; we consider it unlikely that more —soluble“ than granule- bound GFTICAT can be accumulated in the starch granule. GTFICAT, which is an enzyme of large molecular weight, might interfere with the proper folding of SBD, and vice versa. Alternatively, the two domains might interact with each other in such a way, that the accessibility of the catalytic site of the enzyme is limited for substrates, or that the aromatic amino acids responsible for the binding of SBD to starch are not exposed anymore (Kok-Jacon et al ., 2003). The performance of the fusion proteins might be improved by engineering a different inter-domain linker, which keeps the two domains further apart so that they could function more independently. This remains to be investigated further. This paper provides the first evidence that starch with different physical properties can be obtained by fusing SBD to an effector (GTFICAT). One of the transformant lines (KDSIC10) produced starch with a 3 °C higher T onset. This increase was very consistent with other observations for this line, including a high transcript level with RT- PCR analysis, detection of SBD by Western blot analysis, the most severely altered granule morphology of the described series, a large amount of mutan in starch preparations, the most deviating viscosity profile, and a large influence on the expression of sucrose-regulated genes as shown by real-time quantitative RT-PCR. Interestingly, these results demonstrate that a covalent attachment of mutan to starch (see debranching analysis) did not seem to be a prerequisite for altering starch properties. The higher T onset of KDSIC10 starch, together with our speculation that GTFICAT alone is probably more active than SBD-GTFICAT, suggests that the site of deposition of the foreign polymer might be more important for altering starch properties than the amount that is actually produced. It is not understood why the T onset of KDSIC10 starch is increased. We think that the presence of mutan would interfere with granule packing, leading to a less ordered structure, and had therefore expected a lower T onset. Since mutan is not well soluble in water, it is possible that the mutan chains interact, forming a co-existing network with starch, which might reinforce the

1 03: Chapter 5/ Production of mutan and SBD technology granule, and increase the melting temperature. It can not be excluded that starch polymers and mutan interact with each other inside the granule. The increased viscosity (retrogradation) upon cooling a paste of KDSIC10 starch compared to KD-UT may indicate this. This investigation has demonstrated that it is possible to alter starch granule properties by expression of SBD-GTFICAT in potato tubers, the impact of which is larger than with GTFICAT alone. In previous studies, we have shown that more SBD-containing proteins can be accumulated in the amf background (Ji et al. , 2003, 2004b). Therefore, GTFICAT with an appended SBD has also been introduced in amf potatoes, the production of which is in progress. It is expected that the properties of starch derived from these plants are more severely affected than in the Kardal background.

Acknowledgements

The authors would like to thank Isolde Pereira for her assistance with the tissue culture, Dirkjan Huigen for helping with the growth of the plants in the greenhouse and Ing. Jos Molthoff (PRI-WUR) for his assistance with SYBR-Green analysis. In addition, we are very grateful to Dr. Adrian Wiater (Department of Industrial Microbiology, Lublin, Poland) for providing the mutan polymers and the mutanase enzyme.

References

104: Chapter 5/ Production of mutan and SBD technology

Hennegan, K.P., Danna, K.J., 1998. pBIN20: an improved binary vector for Agrobacterium - mediated transformation. Plant Mol. Biol. Rep. 16, 129œ131. Ji, Q., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2003. Microbial starch-binding domains as a tool for targeting proteins to granules during starch biosynthesis. Plant Mol. Biol. 51, 789œ801. Ji, Q., Oomen, R.J.F.J., Vincken, J-P., Bolam, D.N., Gilbert, H.J., Suurs, L.C.J.M., Visser, R.G.F., 2004a. Reduction of starch granule size by expression of an engineered tandem starch-binding domain in potato plants. Plant Biotechnol. J. 2, 251œ260. Ji, Q., 2004b. Microbial starch-binding domains as a tool for modifying starch biosynthesis. Ph.D. Dissertation, Wageningen University, The Netherlands, ISBN 90-8504-022-1. Kok-Jacon, G.A., Ji, Q., Vincken, J-P., Visser, R.G.F., 2003. Towards a more versatile S-glucan biosynthesis in plants. J. Plant Physiol. 160, 765œ777 (this thesis, Chapter 1). Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2005a. Mutan produced in potato amyloplasts adheres to starch granules. Plant Biotechnol. J. 3, 341œ351 (this thesis, Chapter 3). Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Wang, D., Liu, S., Visser, R.G.F., 2005b. Production of dextran in transgenic potato plants. Transgenic Res. in press (this thesis, Chapter 2). Kuipers, A.G.J., Jacobsen, E., Visser, R.G.F., 1994. Formation and deposition of amylose in the potato tuber starch granule are affected by the reduction of granule-bound starch synthase gene expression. Plant Cell 6, 43œ52. Lawson, C.L., van Montfort, R., Strokopytov, B., Rozeboom, H.J., Kalk, K.H., de Vries, G.E., Penninga, D., Dijkhuizen, L., Dijkstra, B.W., 1994. Nucleotide sequence and X-ray structure of cyclodextrin glycosyltransferase from Bacillus circulans strain 251 in a maltose-dependent crystal form. J. Mol. Biol. 236, 590œ600. Marshall, J., Sidebottom, C., Debet, M., Martin, C., Smith, A.M., Edwards, A., 1996. Identification of the major starch synthase in the soluble fraction of potato tubers. Plant Cell 8, 1121œ1135. Monchois, V., Vignon, M., Escalier, P.C., Svensson, B., Russell, R.R.B., 2000. Involvement of Gln937 of Streptococcus downei GTF-I glucansucrase in transition-state stabilization. Eur. J. Biochem. 267, 4127œ 4136. Murashige, T., Skoog, F., 1962. A revised medium for rapid growth and bioassay with tobacco tissue culture. Physiol. Plant. 15, 473œ497. Salehuzzaman, S.N.I.M., Jacobsen, E., Visser, R.G.F., 1994. Expression patterns of two starch biosynthetic genes in in vitro cultured cassava plants and their induction by sugars. Plant Sci. 98, 53œ62.

105: Chapter 5/ Production of mutan and SBD technology

Takken, F.L.W., Luderer, R., Gabriëls, S.H.E.J., Westerink, N., Lu, R., de Wit, P.J.G.M., Joosten, M.H.A.J., 2000. A functional cloning strategy, based on a binary PVX- expression vector, to isolate HR-inducing cDNAs of plant pathogens. Plant J. 24, 275œ 283. Wenzler, H.C., Mignery, A., Fisher, L.M., Park, W.D., 1989. Analysis of a chimeric class-I patatin-GUS gene in transgenic potato plants: High-level expression in tubers and sucrose-inducible expression in cultured leaf and stem explants. Plant Mol. Biol. 12, 41œ 50. Wiater, A., Choma, A., Szczodrak, J., 1999. Insoluble glucans synthesized by cariogenic streptococci: a structural study. J. Basic Microbiol. 39, 265œ273.

107: Chapter 6/ General discussion

General discussion

108: Chapter 6/ General discussion

This study described the co-expression of various glucansucrases in potato tuber amyloplasts together with the native starch biosynthetic machinery. Experiments have been performed to explore the possibilities to confer novel properties to potato starch by genetic engineering. Particularly, attention was focused on the potential to produce novel types of glucan polymers, linked by S-(1R3), S-(1R6) and alternating S-(1R3)/ S-(1R6) glucosyl residues in potato amyloplasts. In addition, possibilities to modify the fine structure of starch, by making use of the so-called acceptor reaction of glucansucrases, were also investigated. This study is a contribution to a better understanding of the production of foreign polymers in plants in general.

Selection of glucansucrases for the diversification of linkages types in starch At the start of the work described in this thesis, selection of candidate glucansucrases had been based on sequences available at http://afmb.cnrs-mrs.fr/CAZY/. Our choice was focused on glucansucrases synthesizing polymers with a large diversity in types of linkages, consisting of S-(1R3), S-(1R6) and alternating S-(1R3)/ S-(1R6)-linked glucosyl residues (see Chapter 1, Figure 8). At that time, glucansucrases synthesizing S-(1R2)-linked glucosyl residues were not available in the databases. Recently, a glucansucrase named DSR-E with an unique primary structure, synthesizing a polymer with S-(1R2)-linked glucosyl residues, was reported in the literature (Bozonnet et al ., 2002; Fabre et al ., 2005). Sucrases producing polymers with an S-(1R4) glucosidic type have also been reported (Kralj et al ., 2002). However, such a glucansucrase was a priori not suitable for our purposes because it could not diversify starch linkage types. Concerning the acceptor reaction efficiency, only few glucansucrases were well characterized (Côté and Robyt 1982; Fu and Robyt 1990, 1991; Heincke et al ., 1999; Mukasa et al ., 2000). Our selection for the dextransucrase (DSRS) (Wilke-Douglas et al ., 1989) and alternansucrase (ASR) (Argüello-Morales et al ., 2000) was based on previous biochemical characterization performed on acceptor reaction, showing that such enzymes exhibited a high acceptor reaction efficiency in presence of acceptor molecules, such as maltodextrins (Côté and Robyt 1982; Heincke et al ., 1999; Demuth et al ., 2002; Richard et al ., 2003). Concerning the mutansucrase (GTFI) from Streptococcus downei Mfe28 (Ferretti et al ., 1987), studies on acceptor reaction efficiency have not been performed until now. Selection of this glucansucrase was based on its direct availability and on the fact that the glucan produced was linked by S-

109: Chapter 6/ General discussion

(1R3) bonds and exhibited properties such as adhesiveness and water-insolubility that are unique to this polymer (Hamada and Slade, 1980). In addition, domain truncation was possible with this enzyme (Monchois et al ., 1999); and thus it might be an advantage to have a smaller mutansucrase with respect to transport through the amyloplastic membrane. Finally, the dextransucrases GTFK (100 % S(1R6)) and GTFT (73 % S-(1R6), 27 % S-(1R3)) were also used for potato transformation and further analyses at the molecular and biochemical level are currently in progress.

Incorporation of novel linkage types in starch At the beginning of this work, one of the main goals was to modify the fine structure of potato starch. For this, the glucansucrases need to be directed to the amyloplast, although it was realized that the donor substrate sucrose might be more abundant in other subcellular compartments. From all the analyzed transgenic starches, we have not obtained evidence that starch structure was diversified by novel types of linkages. In spite of that, dextran, mutan or alternan were co-produced with starch, without interfering directly with the starch structure, although the morphology of the granules was altered with a number of the sucrases tested. Therefore, we postulated that the acceptor reaction does not seem to play a major role inside amyloplasts, possibly because nascent starch polymers are poor acceptors for the selected glucansucrases and/or the concentration of acceptor molecules is too low. Use of the starch-binding domain technology to bring glucansucrases in more intimate contact with growing starch granules was therefore applied. Conferring granule-boundness to GTFICAT (see Chapter 5) might increase the probability to incorporate mutan polymers to starch and to favour the acceptor reaction inside starch granules. From our results, it was shown that granule-boundness was conferred unambigiously to only one transformant and that the activity of GTFICAT was inhibited by the appended SBD. Despite the use of the starch-binding domain technology, it was concluded that the acceptor reaction was not particularly favoured inside starch granules because alterations of starch structures were not detected by chain length distribution analysis. From the work of Ji et al. (2003; 2004), it was shown that more SBD-containing proteins can be accumulated in the amf potato than in an amylose-containing background. Transformation of the amf background with the same constructs has already been performed, and these transformants will be analyzed very soon.

110: Chapter 6/ General discussion

Effects of the co-production of starch and foreign polymers on potato development and starch properties In this work, it has been demonstrated that it is possible to produce polysaccharides from lactic acid bacteria, inside potato amyloplasts. These co-products were coupled to changes at different levels in the potato plants. These involved amongst others: (i) Tuber browning and decrease of starch content in high-expressing GtfICAT plants (Chapter 3). (ii) Altered starch granules in high-expressing DsrS -, GtfICAT and SBD- GtfICAT plants (Chapters 2, 3 and 5). (iii) Down-regulation of ADP-glucose pyrophosphorylase subunit S and granule-bound starch synthase I in high-expressing GtfICAT -, Asr - and SBD-GtfICAT plants (Chapters 3, 4 and 5) and sucrose synthase in high-expressing GtfICAT -SBD (Chapter 5) and (iv) interference with the rheological properties of the starches from the high-expressing GtfICAT and SBD-GtfICAT plants, exhibiting a higher end viscosity compared to control plants (Chapters 3 and 5). In general, effects on potato development and starch properties were considered as significant and were compared to previous studies performed with bacterial sucrose- converting enzymes targeted to potato amyloplasts, in particular those with levansucrase (Gerrits, 2000; Gerrits et al ., 2001). In these studies, it was shown that dramatical changes occurred at the plant and starch levels due to levan production in potato tubers. The amounts of different polymers varied considerably from about 66 mg g-1 FW in levansucrase expressing plants to 1.2 mg g -1 FW alternan (Chapter 4) and 1.7 mg g -1 FW dextran (Chapter 2). The observed effects on plant development and starch biosynthesis due to levan production seemed to be related to the amount that was produced in potato tubers. In addition, the starch granules from the levan- producing plants contained approximately 5 % of levan, which was not the case for dextran- and alternan-accumulating plants. Assuming that the sucrose concentration in the levansucrase and glucansucrase transformants is the same, it seems as if the levansucrase is more efficient in producing polymers than the glucansucrases. The reason for these differences is still unknown; it could be that the levansucrases have more favourable characteristics than glucansucrases for application in potato, although the Km's and pH optima of the various enzymes, especially for the dextransucrase, are more or less the same (levansucrase: Km = 27 mM, pH = 6.0; Ebskamp, 1994; dextransucrase: Km = 26 mM, pH = 5.2; Monchois et al ., 1997). Although the full-length mutansucrase GTFI has a higher affinity for sucrose, (Km value of about 4.5 mM;

111: Chapter 6/ General discussion

Wright et al ., 2002) than levansucrase, mutan accumulation in potato tubers was lower than that of levan. This suggests that enzyme characterictics other than the Km are responsible for the better performance of the levansucrase. Possibly, the pH optimum is a factor of importance, which is about 6.5 for the mutansucrase. Data on the Km and pH optimum of the alternansucrase are not available. Alternatively, the higher accumulation of levan in potato amyloplasts might be due to the lower molecular weight of levansucrase compared to that of glucansucrases; levansucrase has a molecular weight of about 50 kDa (Ebskamp, 1994), whereas that of glucansucrases ranges from about 155 kDa (DSRS and GTFI) to 200 kDa (ASR) (Remaud-Simeon et al ., 2000). Therefore, the transport of levansucrase through the amyloplastic membrane might be easier, leading to higher plastidial enzyme concentrations, and possibly more accumulation of the polymer. This might be in accordance with our results on the mutan-accumulating potato plants, where the smaller, GBD-truncated form of GtfI accumulated more mutan than for the full-length GtfI (Kok-Jacon et al ., 2005).

Strategies to increase the amount of glucan polymers produced in plants Our primary goal was not to produce a maximum amount of glucan polymers in potato plants. However, it can be envisaged that a higher yield could also be an objective, independent of that of modifying starch structure. In the past decade, many studies have been performed to produce foreign polymers in planta (Gerrits, 2000; Romano, 2002; Cairns, 2003). However, it is still common that their yield remains low . Different strategies can be employed to improve their production. Targeting to subcellular compartments with a higher amount of sucrose might be a possibility to increase glucan yield in potato plants. In tubers, concentrations of sucrose are higher in the vacuole and the cytosol, estimated to be 35 mM and 41 mM, respectively, in comparison to that of the amyloplast of only 10 mM (Farré et al ., 2001). Van der Meer et al . (1994) showed that it was possible to produce fructans by expressing the sucrose-converting SACB levansucrase in the vacuolar compartment of potato tubers at a level of about 10 mg g -1 FW (see Table 1), which is six times lower than that of levan (60 mg g -1 FW) produced in potato amyloplasts (Gerrits, 2000; Gerrits et al ., 2001). These results demonstrate that a higher amount of sucrose is not the only prerequisite for a higher production of foreign polymers in plants. To this end, it is unclear what the other critical factors for the production of polymers from sucrose are. It

112: Chapter 6/ General discussion might be possible that vacuolar accumulation could influence the carbon partitioning

Table 1. Summary of parameters associated with fructan or glucan accumulation in potato tubers, transformed with various bacterial sucrases.

Gene Bacterial Targeted Promoter Targeting Endpoint Reference origin compartment signal fructan/glucan (mg g -1) SacB Bacillus subtilis Plastid Tuber Ferredoxin 66 Gerrits et specific chloroplast al ., 2001 patatin DsrS Leuconostoc Plastid Tuber Ferredoxin 1.7 (dextran) This thesis, mesenteroides specific chloroplast Chapter 2 patatin NRRL B-512F Asr Leuconostoc Plastid Tuber Ferredoxin 1.2 (alternan) This thesis, mesenteroides specific chloroplast Chapter 4 patatin NRRL B-1355 SacB Bacillus subtilis Vacuole Constitutive Vacuolar 10 van der CaMV 35S cpy Meer et al ., 1994 SacB Bacillus subtilis Vacuole Constitutive Vacuolar 11 Pilon-Smits CaMV 35S cpy et al ., 1996 SacB Bacillus Cytoplasm Tuber Untargeted 9.8 Caimi et amyloliquifaciens specific al ., 1997 patatin within the plant more drastically than amyloplastic accumulation, thereby limiting the levan production. It should be noted that levan accumulation in vacuolar as well as in amyloplastic compartments was correlated with severe alterations in plant development such as amongst others, a lower starch content, a reduction of tuber yield and detrimental plant phenotypes. Another strategy for increasing the amount of glucan polymers might be the use of a sugar-accumulating crop such as the sugar beet. Sugar beet, that can be transformed by a polyethylene glycol-based technique (Hall et al ., 1996), contains about 0.7 M of sucrose in its roots (Lohaus et al ., 1994), which is at least 15 times more than in potato tubers (Farré et al ., 2001). The use of the sugar beet for fructan production has already been investigated; it appeared that sucrose could be converted efficiently into onion- type fructans (Weyens et al ., 2004). It was demonstrated that about 100 mg g -1 fructo- oligosaccharides with a degree of polymerization larger than three accumulated in the taproots, compared to only 4 mg g -1 in control plants. From these results, it is apparent that glucan polymers can be accumulated more efficiently in sugar beet than in potato tubers.

113: Chapter 6/ General discussion

Final remarks This study has described the potential of potato plants to be used as a recipient for the production of bacterial polymers such as dextran, mutan and alternan, conferring added-value to this crop. Additional studies need to be performed in order to enhance their in planta production, and to enable the addition of novel linkage types to starch. Elucidation of the three-dimensional structure of glucansucrases will certainly give better insight on the structure-function relationships of these enzymes, which might enable the identification of essential residues (or regions) responsible for acceptor reactions. By site-directed mutagenesis it may become possible to engineer a second generation of glucansucrases which have improved characteristics with respect to acceptor reactions with maltodextrins. It is thought that such enzymes may provide a better starting point for diversifying starch structure. Even though the generated potato transformants have not been tested for industrial applications, some of them might be interesting for this, especially the mutan- accumulating plants. Simultaneous presence of mutan with starch might generate environmentally friendly compounds, such as adhesives. For instance, adhesiveness of starch might be increased for paper coating, replacing the currently used polyvinyl alcohol and acrylic latex (Haag et al ., 2004). In addition, the presence of mutan might improve the moisture resistance of paper products, due to the capacity of mutan to exclude water from its structure (Sutherland, 2001). Interestingly, it was shown that mutan interfered with the rheological properties of the starches, in which a higher viscosity was obtained after cooling the starch paste (Kok-Jacon et al ., 2005). The significance of this finding with respect to applications needs to be further investigated.

114: Chapter 6/ General discussion

References

Argüello-Morales, M.A., Remaud-Simeon, M., Pizzut, S., Sarçabal, P., Willemot, R.M., Monsan, P., 2000. Sequence analysis of the gene encoding alternansucrase, a sucrase glucosyltransferase from Leuconostoc mesenteroides NRRL B-1355. FEMS Microbiol. Lett. 182, 81œ85. Bozonnet, S., Dols-Laffargue, M., Fabre, E., Pizzut, S., Remaud-Simeon, M., Monsan, P., Willemot, R-M., 2002. Molecular characterization of DSR-E, an S-1,2 linkage- synthesizing dextransucrase with two catalytic domains. J. Bacteriol. 184, 5753œ5761. Caimi, P.G., McCole, L.M., Klein, T.M., Hershey, H.P., 1997. Cytosolic expression of the Bacillus amyloliquefaciens SacB protein inhibits tissue development in transgenic tobacco and potato. New Phytologist 136, 19œ28. Cairns, A.J., 2003. Fructan biosynthesis in transgenic plants. J. Exp. Bot. 54, 549œ567. Côté, G.L., Robyt, J.F., 1982. Acceptor reactions of alternansucrase from Leuconostoc mesenteroides NRRL B-1355. Carbohydr. Res. 111, 127œ142. Demuth, K., Jördening, H.J., Buchholz, K., 2002. Oligosaccharide synthesis by dextransucrase: new unconventional acceptors. Carbohydr. Res. 337, 1811œ1820. Ebskamp, M., 1994. Fructan accumulation in transgenic plants. Ph.D. Dissertation, Utrecht University, The Netherlands, ISBN 90-393-0623-0. Fabre, E., Bozonnet, S., Arcache, A., Willemot, R-M., Vignon, M., Monsan, P., Remaud- Simeon, M., 2005. Role of the two catalytic domains of DSR-E dextransucrase and their involvement in the formation of highly S-1,2 branched dextran. J. Bacteriol. 187, 296œ 303. Farré, E.M., Tiessen, A., Roessner, U., Geigenberger, P., Trethewey, R.N., Willmitzer, L., 2001. Analysis of the compartmentation of glycolytic intermediates, nucleotides, sugars, organic acids, amino acids, and sugar alcohols in potato tubers using a nonaqueous fractionation method. Plant Physiol. 127, 685œ700. Ferretti, J.J., Gilpin, M.L., Russell, R.R.B., 1987. Nucleotide sequence of a glucosyltransferase gene from Streptococcus sobrinus Mfe28. J. Bacteriol. 169, 4271œ4278. Fu, D., Robyt, J.F., 1990. Acceptor reactions of maltodextrins with Leuconostoc mesenteroides B-512FM dextransucrase. Arch. Biochem. Biophys. 283, 379œ387. Fu, D., Robyt, J.F., 1991. Maltodextrin acceptor reactions of Streptococcus mutans 6715 glucosyltransferases. Carbohydr. Res. 217, 201œ211. Gerrits, N., 2000. Tuber-specific fructan synthesis in potato amyloplasts. Ph.D. Dissertation, Utrecht University, The Netherlands, ISBN 90-393-2345-3.

115: Chapter 6/ General discussion

Gerrits, N., Turk, S.C.H.J., van Dun, K.P.M., Hulleman, S.H.D., Visser, R.G.F., Weisbeek, P.J., Smeekens, S.C.M., 2001. Sucrose metabolism in plastids. Plant Physiol. 125, 926œ934. Haag, A.P., Maier, R.M., Combie, J., Geesey, G.G., 2004. Bacterially derived biopolymers as wood adhesives. Int. J. Adhesion Adhesives 24, 495œ502. Hall, R.D., Riksen-Bruinsma, T., Weyens, G.J., Rosquin, I.J., Denys, P.N., Evans, I.J., Lathouwers, J.E., Lefèbvre, M.P., Dunwell, J.M., van Tunen, A., Krens, F.A., 1996. A high efficiency technique for the generation of transgenic sugar beets from stomatal guard cells. Nature Biotechnol. 14, 1133œ1138. Hamada, S., Slade, H.D., 1980. Biology, immunology, and cariogenicity of Streptococcus mutans . Microbiol. Rev. 44, 331œ384. Heincke, K., Demuth, B., Jördening, H-J., Buchholz, K., 1999. Kinetics of the dextransucrase acceptor reaction with maltoseœ-experimental results and modeling. Enzyme Microbial Technol. 24, 523œ534. Ji, Q., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2003. Microbial starch-binding domains as a tool for targeting proteins to granules during starch biosynthesis. Plant Mol. Biol. 51, 789œ801. Ji, Q., 2004. Microbial starch-binding domains as a tool for modifying starch biosynthesis. Ph.D. Dissertation, Wageningen University, The Netherlands, ISBN 90-8504-022-1. Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2005. Mutan produced in potato amyloplasts adheres to starch granules. Plant Biotechnol. J. 3, 341œ351 (this thesis, Chapter 3). Kralj, S., van Geel-Schutten, G.H., Rahaoui, H., Leer, R.J., Faber, E.J., van der Maarel, M.J.E.C., Dijkhuizen, L., 2002. Molecular characterization of a novel glucosyltransferase from Lactobacillus reuteri strain 121 synthesizing a unique, highly branched glucan with S-(1R4) and S-(1R6) glucosidic bonds. Appl. Environ. Microbiol. 68, 4283œ4291. Lohaus, G., Burba, M., Heldt, H.W., 1994. Comparison of the contents of sucrose and amino- acids in the leaves, phloem sap and taproots of high and low sugar-producing hybrids of sugar-beet ( beta-vulgaris L ). J. Exp. Bot. 45, 1097œ1101. Monchois, V., Remaud-Simeon, M., Russell, R.R.B., Monsan, P., Willemot, R-M., 1997. Characterization of Leuconostoc mesenteroides NRRL B-512F dextransucrase (DSRS) and identification of amino-acid residues playing a key role in enzyme activity. Appl. Microbiol. Biotechnol. 48, 465œ472. Monchois, V., Argüello-Morales, M., Russell, R.R.B., 1999. Isolation of an active catalytic core of Streptococcus downei Mfe28 Gtf-I glucosyltransferase. J. Bacteriol. 181, 2290œ2292. Mukasa, H., Shimamura, A., Tsumori, H., 2000. Nigerooligosaccharide acceptor reaction of Streptococcus sobrinus glucosyltransferase GTF-I. Carbohydr. Res. 326, 98œ103.

116: Chapter 6/ General discussion

Pilon-Smits, E.A.H., Ebskamp, M.J.M., Jeuken, M.J.W., van der Meer, I.M., Visser, R.G.F., Weisbeek, P., Smeekens, S.C.M., 1996. Microbial fructan production in transgenic potato plants and tubers. Ind. Crops Prod. 5, 35œ46. Remaud-Simeon, M., Willemot, R-M., Sarçabal, P., Potocki de Montalk, G., Monsan, P., 2000. Glucansucrases: molecular engineering and oligosaccharide synthesis. J. Mol. Cat. 10, 117œ128. Richard, G., Morel, S., Willemot, R-M., Monsan, P., Remaud-Simeon, M., 2003. Glucosylation of S-butyl- and S-octyl-D-glucopyranosides by dextransucrase and alternansucrase from Leuconostoc mesenteroides . Carbohydr. Res. 338, 855œ864. Romano, A., 2002. Production of polyhydroxyalkanoates (PHAs) in transgenic potato. Ph.D. Dissertation, Wageningen University, The Netherlands, ISBN 90-5808-726-3. Sutherland, I.W., 2001. Biofilm exopolysaccharides: a strong and sticky framework. Microbiol. 147, 3œ9. Van der Meer, I.M., Ebskamp, M.J.M., Visser, R.G.F., Weisbeek, P.J., Smeekens, S.C.M., 1994. Fructan as a new carbohydrate sink in transgenic potato plants. Plant Cell 6, 561œ570. Weyens, G., Ritsema, T., van Dun, K., Meyer, D., Lommel, M., Lathouwers, J., Rosquin, I., Denys, P., Tossens, A., Nijs, M., Turk, S., Gerrits, N., Bink, S., Walraven, B., Lefèbvre, M., Smeekens, S., 2004. Production of tailor-made fructans in sugar beet by expression of onion fructosyltransferase genes. Plant Biotech. J. 2, 321œ327. Wilke-Douglas, M., Perchorowicz, J.T., Houck, C.M., Thomas, B.R., 1989. Methods and compositions for altering physical characteristics of fruit and fruit products. PCT patent WO 89/12386. Wright, W.G., Thelwell, C., Svensson, B., Russell, R.R.B., 2002. Inhibition of catalytic and glucan-binding activities of a streptococcal GTF forming insoluble glucans. Caries Res. 36, 353œ359.

117: Summary, Samenvatting, Résumé

Summary, Samenvatting, Résumé

118: Summary

Summary Starch consists mainly of two polymers of glucose, amylose and amylopectin. It is deposited as granules in a very ordered structure, with the succession of crystalline and amorphous lamellae. Amylose, which is present for about 20-30 %, is essentially linear in contrast to amylopectin (70-80 %) which is highly branched. Modification of starch composition in potato by genetic engineering has been extensively studied during the last decade and a number of transgenic starches have been generated, including the amylose-free starch. Such approach can be a way to broaden the number of applications for the starch industry, such as the paper, textile, adhesives, pharmaceutical, and food industries. In order to modify the physicochemical properties of potato starch, and to investigate the possibilities of introducing novel linkage types in starch, glucansucrases were introduced in potato, and targeted to amyloplasts. Glucansucrases are bacterial enzymes that use sucrose as a donor substrate and polymerize the glucosyl residues by S-(1R2), S-(1R3), S-(1R4) or S-(1R6) linkages. In addition, such enzymes can modify the structure of carbohydrate acceptor molecules by covalently adding glucosyl residues. This thesis describes the expression of various glucansucrase genes in potato plants, in particular of the dextransucrase ( DsrS ) gene leading to the production of dextran (polymers with mainly S-(1R6)-linked glucosyl residues) (Chapter 2), the mutansucrase (GtfI and GtfICAT (a truncated form of GtfI )) genes leading to the production of mutan (polymers with mainly S-(1R3)-linked glucosyl residues) (Chapter 3). Furthermore, a hybrid form of these polymers, alternan (a polymer with alternating S-(1R3)/ S-(1R6)- linked glucosyl residues) was produced by expressing the alternansucrase ( Asr ) gene in potato (Chapter 4). Finally, the use of starch-binding domain (SBD) technology for engineering granule-boundness of glucansucrases was investigated by fusing GtfICAT to the N- or C-terminus of SBD (Chapter 5). The production of these novel polymers in potato tubers, and their subsequent effect on starch biosynthesis, was studied at the molecular and biochemical levels. The selected bacterial glucansucrase genes were cloned in frame behind the chloroplastic ferredoxin signal peptide (FD) enabling amyloplast entry. These fragments were driven by the highly tuber-expressed patatin promoter. For the production of dextran (Chapter 2), the Kardal and the amylose-free ( amf ) mutant genotypes were used for potato transformation. In the Kardal background, the production of dextran

119: Summary was the highest, which was demonstrated by enzyme-linked immunosorbent assay (ELISA) with anti-dextran antibodies. Heterologous expression of the DsrS gene was also demonstrated by RT-PCR and real-time quantitative RT-PCR analysis, which correlated very well with the ELISA results. Despite morphological changes of the starch granules, the physicochemical properties of the transgenic starches were identical to those of untransformed potato plants. In addition, covalent binding of S- (1R6)-linked glucosyl residues was not demonstrated by chain length distribution, although DSRS is known to catalyze acceptor reactions at high efficiency. Finally, sucrose conversion of DSRS led to a slight down-regulation of the sucrose-regulated Susy , AGPase and GBSSI starch synthesizing genes. Production of mutan (Chapter 3) was achieved by expressing the GtfI gene in Kardal potato plants, which were shown to be most appropriate for dextran accumulation. Furthermore, in an effort to improve the GTFI targeting to the amyloplast compartment, GTFI was shortened by truncating its glucan-binding domain, leading to GTFICAT, from which it is known from in vitro studies that 70 % of the enzyme activity was retained after domain truncation. Expression of GtfICAT led to a higher production of mutan, correlating with pronounced alterations at the tuber and starch levels, in comparison to that of GtfI . GtfICAT expressing plants exhibited an impaired tuber phenotype with brownish/reddish regions. Furthermore, it was observed that mutan was deposited on starch granules, as visualized by erythrosine staining. Granule morphology was dramatically affected, showing eroded surfaces with holes. For the high expressers, the starch content was decreased, which correlated with a down-regulated AGPase gene expression. In addition, the expression level of the sucrose-regulated GBSSI gene was down-regulated; however, this did not result in a lower amylose content. The rheological properties of the GTFICAT starches appeared to be modified; the starches showed more pronounced retrogradation after gelatinization, and had an about 1.7-fold higher viscosity after cooling than untransformed and GTFI starches. Apparently, the simultaneous presence of mutan, amylopectin, and amylose can create a more viscous network of polymers in solution. Production of alternan (Chapter 4) was demonstrated by means of ELISA experiments by expressing the Asr gene in Kardal potato plants. The morphological and physicochemical properties of the ASR starch granules were not significantly changed and were comparable to those of the untransformed plants. However, down-regulation

120: Summary of the sucrose-regulated genes, AGPase and GBSSI, were observed in particular in the higher asr expressers. Novel structural elements such as alternating S-(1R3)/ S-(1R6)- linked glucosyl residues were not found to be attached to starch molecules, despite the known high acceptor reaction efficiency of ASR. Possibly, polymer solubility and alteration of starch granule morphology are correlated for the Asr -, DsrS - and GtfICAT - expressing plants. Alternan is a highly water-soluble polymer in comparison to dextran which is less soluble in water, and mutan which is water-insoluble. There appears to be a trend that the accumulation of water-insoluble polymers can affect the starch granule morphology more dramatically than that of a more water-soluble one. It was also investigated whether new structural elements could be introduced into starch by bringing GTFICAT and the growing starch granule in more intimate contact with each other during starch biosynthesis. For this, SBD was fused to the N- or C- terminus of GTFICAT (Chapter 5). After expression of the SBD-containing gene fusion in Kardal potato plants, the presence of SBD inside starch granules was demonstrated in only the SBD-GTFICAT starches, by using specific anti-SBD antibodies. Furthermore, the high expressers of this series of transformants showed less pronounced morphological and physicochemical starch alterations than the transformants expressing only the GtfICAT gene. Therefore, it seems as if the appended SBD negatively influenced the activity of GtfICAT in the fusion protein . Finally, starch granules of the best SBD-GTFICAT transformant was incubated with an excess of sucrose, but the in vitro production of mutan was not evidenced. Further research is needed to investigate the novel properties of the transgenic starches, in particular from the mutan-accumulating potato plants. During this work, it was possible to influence starch biosynthesis by expressing glucansucrase genes in potato tubers, in particular with GtfICAT and to a lesser extent with DsrS . Furthermore, introduction of novel linkage types in starch was not evidenced. Although ASR and DSRS are known to have a high acceptor efficiency, attachment of novel side chains to starch requires different enzyme characteristics. More research is needed to investigate the possibilities of engineering a second generation of glucansucrases with improved acceptor reactions towards starch.

121: Samenvatting

Samenvatting Zetmeel bestaat uit twee polymeren van glucose: amylose en amylopectine. Het wordt afgezet in de vorm van korrels met een zeer georganiseerde structuur, waarin kristallijne en amorfe ringen elkaar opvolgen. Amylose (20-30 %) is lineair, terwijl amylopectine (70-80 %) een sterk vertakt molecuul is. Genetische modificatie van de zetmeel-samenstelling werd gedurende de laatste 10 jaren intensief bestudeerd met als doel zetmeelstructuur en œeigenschappen te veranderen. Dit heeft geleid tot een aantal transgene zetmelen, waaronder het amylose-vrije ( amf ) zetmeel. Genetische modificatie is een manier om het aantal industriële toepassingen van zetmeel te vergroten, zoals in papier, textiel, en hechtmiddel, en in de farmaceutische- en levensmiddelenindustrie. In dit onderzoek hebben we geprobeerd om de fysisch- chemische eigenschappen van aardappel-zetmeel te veranderen door nieuwe vertakkingen (bindingstypen) in de zetmeelpolymeren aan te brengen door glucaansucrasen naar de amyloplast te dirigeren. Glucaansucrasen zijn bacteriële enzymen die sucrose nodig hebben voor de synthese van glucanen met S-(1R2), S- (1R3), S-(1R4) of S-(1R6) gebonden glucosyl eenheden. Bovendien kunnen zulke enzymen glucoseresiduen koppelen aan acceptor-moleculen zoals b.v. maltose. Dit proefschrift beschrijft de genetische modificatie van aardappel planten m.b.v. glucaansucrase genen, zoals het dextraansucrase ( DsrS ) gen dat leidt tot de productie van dextranen (polymeren met S-(1R6) gebonden glucosyl eenheden) (Hoofdstuk 2), en de mutaansucrase ( GtfI en GtfICAT (een getrunceerde vorm van GtfI )) genen die leiden tot de productie van mutaan (polymeren met S-(1R3) gebonden glucosyl- eenheden) (Hoofdstuk 3). Bovendien werd het alternaansucrase ( Asr ) gen gebruikt om alternanen, opgebouwd uit glucose-eenheden die alternerend S-(1R3) en S-(1R6) zijn gebonden, in aardappelen te produceren (Hoofdstuk 4). Als laatste, werd het gebruik van een zetmeel-bindend domain (SBD) onderzocht om een kunstmatig korrel- gebonden GtfICAT te vervaardigen (Hoofdstuk 5). De productie van deze nieuwe polymeren in aardappel knollen en hun effect op de zetmeelbiosynthese werden bestudeerd op moleculair en biochemisch niveau. Om transport van de verschillende glucaansucrasen over de amyloplast membraan te verzekeren, werd het chloroplast transitpeptide van ferredoxine (FD) gebruikt. Om een hoge expressie in knollen te bewerkstelligen, werd gebruik gemaakt van de promoter van het aardappel patatine gen. Voor productie van dextranen werd zowel Kardal als

122: Samenvatting het amylose-vrije ( amf) mutante aardappel genotype getransformeerd (Hoofdstuk 2). Met behulp van "enzyme-linked immunosorbent assay" (ELISA), gebruikmakend van anti-dextraan antilichamen, werd de ophoping van dextranen in met name de Kardal achtergrond aangetoond. Heterologe expressie van het DsrS gen werd aangetoond d.m.v. "RT-PCR" en "real-time quantitative RT-PCR" analyse, die goed met de ELISA resultaten overeenkwamen. Ondanks veranderingen in de morfologie van de zetmeelkorrels, waren de fysisch-chemische eigenschappen van de transgene zetmelen vergelijkbaar met die van de wildtype zetmelen. Bepaling van de ketenlengteverdeling van de transgene zetmelen suggereerde dat er geen extra S- (1R6)-gebonden glucosyl-eenheden in het zetmeel geïncorporeerd waren, ondanks het feit dat DSRS in de literatuur bekend staat om zijn hoge efficiëntie in acceptorreacties. De omzetting van sucrose door DSRS leidde tot een lichte reductie van de expressieniveaus van de Susy , AGPase en GBSSI genen, waarvan bekend is dat ze sucrose-gereguleerd zijn. Mutanen werden geproduceerd (Hoofdstuk 3) door expressie van het GtfI gen in het Kardal genotype, dat voor de productie van dextranen het meest geschikt was gebleken (zie Hoofdstuk 2). Bovendien werd, om het transport van GTFI naar de amyloplast te verbeteren, de "glucan-binding domain" van GTFI verwijderd; dit getrunceerde eiwit werd GTFICAT genoemd. Uit in vitro experimenten was bekend dat 70 % van GTFI´s enzym activiteit bewaard bleef na truncatie. Expressie van GtfICAT leidde tot een hogere accumulatie van mutanen, en liet meer veranderingen in knollen en zetmeel zien dan expressie van GtfI . De knollen van GTFICAT transformanten met een hoog expressieniveau hadden bruin/ rode gebieden in hun dwarsdoorsnede. Bovendien werd met erythrosine kleuring aangetoond, dat mutanen aan het oppervlakte van zetmeelkorrels kunnen hechten. Ook de morfologie van de zetmeelkorrels was in sommige gevallen sterk veranderd; sommige zetmeelkorrels leken geërodeerd en vertoonden een aantal kleine gaatjes. De hoge GTFICAT expressors hadden een verlaagde zetmeelopbrengst, hetgeen correleerde met een reductie van het expressieniveau van het AGPase gen. Ook het expressieniveau van het GBSSI gen was lager, maar dit was niet gecorreleerd met een lagere hoeveelheid amylose. De rheologische eigenschappen van de GTFICAT zetmelen bleken sterk veranderd te zijn; deze zetmelen lieten een sterke retrogradatie na verstijfseling zien, met een ongeveer 1.7 maal hogere viscositeit na afkoeling dan die van zetmelen

123: Samenvatting verkregen uit ongetransformeerde planten en GTFI transformanten. Dit laat zien dat de gelijktijdige aanwezigheid van mutanen, amylopectine, en amylose een visceuzer netwerk van polymeren in oplossing kan vormen. Voor de productie van alternan werd het Asr gen in aardappelplanten van het Kardal genotype geïntroduceerd. De ophoping van alternanen in aardappelknollen (Hoofdstuk 4) werd d.m.v. ELISA experimenten gedemonstreerd. De morfologische en fysisch- chemische eigenschappen van de zetmeelkorrels van deze transgene planten waren niet veranderd en vergelijkbaar met die van ongetransformeerde planten. Een reductie van de expressieniveau´s van AGPase en GBSSI werd aangetoond in met name de hoge ASR expressors. Nieuwe vertakkingen, zoals alternerende S-(1R3)/ S-(1R6)- gebonden glucosyl-eenheden, werden niet in het zetmeel gevonden, ondanks de hoge efficiëntie van ASR in acceptor-reacties. Onze experimenten suggereren dat er een verband is tussen de oplosbaarheid van het nieuwe polymeer (alternaan, dextraan, mutaan) en de morfologische veranderingen aan de zetmeelkorrels in de ASR-, DSRS- en GTFICAT transformanten. Alternanen zijn goed oplosbaar in water, terwijl dextranen minder goed water-oplosbaar zijn; mutanen zijn onoplosbaar in water. Het blijkt dat ophoping van onoplosbare polymeren van grotere invloed op de morfologie van zetmeelkorrels is dan die van oplosbare polymeren. Tevens werd onderzocht of het mogelijk is om nieuwe vertakkingselementen aan zetmeel te koppelen door gebruik te maken van zetmeel-bindende domeinen (SBD) in combinatie met GTFICAT. SBD werd zowel aan de N- als aan de C-terminus van GTFICAT gefuseerd (Hoofdstuk 5). Kardal aardappel planten werden getransformeerd met constructen coderend voor SBD-bevattende fusie-eiwitten. De aanwezigheid van SBD werd alleen in SBD-GTFICAT zetmelen aangetoond, d.m.v. specifieke anti-SBD antilichamen. Bovendien vertoonden de hoge expressors van de SBD-GTFICAT serie weinig morfologische en fysisch-chemische veranderingen in het zetmeel in vergelijking met de hogere expressors van de GTFICAT serie. Het lijkt alsof fusie aan SBD de activiteit van GtfICAT negatief beïnvloedt. Zetmeelkorrels van de hoogste expressor van de SBD-GTFICAT serie werden tevens met een overmaat aan sucrose geïncubeerd, maar de in vitro productie van mutanen werd niet aangetoond. Verder onderzoek zal nodig zijn om de nieuwe eigenschappen van de transgene zetmelen, in het bijzonder de GTFICAT zetmelen, te bepalen. In deze studie hebben we laten zien dat het mogelijk is om de zetmeelbiosynthese te beïnvloeden door

124: Résumé expressie van glucaansucrase genen in aardappelknollen, met name voor GtfICAT en in mindere mate voor DsrS . Covalente koppeling van nieuwe vertakkingselementen aan zetmeel werd niet aangetoond, ondanks de hoge efficiëntie van ASR en DSRS in acceptor-reacties. Om dit te bewerkstelligen dienen de enzymen blijkbaar over andere karakteristieken te beschikken. Meer onderzoek is nodig om een tweede generatie van glucaansucrasen te creëren, met een verbeterde efficiëntie in de acceptor-reactie met zetmeel-moleculen.

Résumé L'amidon se compose principalement de deux polymères de glucose: l'amylose et l'amylopectine. L'amidon est déposé sous forme de granules de manière très ordonnée, en formant une succession de lamelles crystallines et amorphes. L'amylose, qui est présente aux environs de 20-30 %, est essentiellement de structure linéaire tandis que l'amylopectine (70-80 %) est largement ramifiée. Les altérations de la composition de l'amidon de pomme de terre par modification génétique ont été largement étudiées pendant les dix dernières années et de nombreuses sortes d'amidons transgéniques ont été produites, incluant l'amidon dépourvu d'amylose. Ce type d'approche peut être un moyen d'élargir les applications industrielles de l'amidon, comme celles du papier, textile, adhésifs, industries pharmaceutiques et alimentaires. Afin de modifier les propriétés physico-chimiques de l'amidon de pomme de terre, et d'investir les possibilités d'introduire de nouveaux types de ramifications dans l'amidon, les glucansucrases ont été introduites dans les tubercules de pomme de terre et transportées vers les amyloplastes, site de synthèse de l'amidon. Les glucansucrases sont des enzymes bactériennes qui utilisent le sucrose comme substrat et polymérisent les résidus de glucose par des liaisons S-(1R2), S-(1R3), S-(1R4) ou S-(1R6). De plus, ces enzymes peuvent modifier la structure de molécules accepteurs comme celle du maltose par addition covalente de résidus glycosidiques. Cette thèse décrit l‘expression de différents gènes de glucansucrases dans la pomme de terre, en particulier ceux de la dextransucrase ( DsrS ) qui produit des dextranes (polymères avec principalement des liens S-(1R6)) (Chapitre 2), ceux de la mutansucrase ( GtfI et GtfICAT (forme tronquée de GtfI )) qui produit des mutanes (polymères avec principalement des liens S-(1R3)) (Chapitre 3). De plus, une forme

125: Résumé hybride de ces polymères, les alternanes (polymères avec des liens S-(1R3) alternant avec des S-(1R6)), a été produite dans les tubercules de pomme de terre, en exprimant le gène de l‘alternansucrase ( Asr ) (Chapitre 4). Finalement, lútilisation de la technique dénommée —Starch-Binding Domain (SBD)“ pour l‘ancrage d‘enzymes à l‘intérieur de granules d‘amidon a été explorée en fusionnant ce domaine tant aux extrémités N- que C- terminal de GtfICAT (Chapitre 5). La production de ces nouvelles sortes de polymères dans les tubercules de pomme de terre, et leurs effets sur la synthèse de lámidon, ont été étudiés tant au niveau moléculaire que biochimique. Les gènes sélectionnés de glucansucrases ont été clonés en aval du peptide signal ferredoxine dórigine chloroplastique (FD), permettant le transport vers lámyloplaste. La transcription de ces fragments dÁDN a été conduite par le promoteur de patatine qui séxprime fortement dans les tubercules de pomme de terre. Pour la production de dextranes (Chapitre 2), les génotypes Kardal et amf , le dernier étant dépourvu dámylose, ont été utilisés pour la transformation des pommes de terre. La production des dextranes était la plus importante dans le génotype Kardal, qui a été démontrée au moyen d´ —enzyme-linked immunosorbent assay— (ELISA) à láide dánticorps anti- dextranes. Léxpression du gène DsrS a été analysée par —RT-PCR— et —real-time quantitative RT-PCR— et les résultats concordent avec ceux obtenus par ELISA. Malgré des changements morphologiques des granules dámidon, les propriétés physico- chimiques des amidons transgèniques étaient identiques à celles de pommes de terre non transformées. De plus, la présence de liaisons covalentes S-(1R6) n´était pas démontrée par — chain length distribution“, bien quíl est connu que DSRS catalyse des réactions dáccepteurs avec grande efficacité. Finalement, la conversion de sucrose par DSRS a abouti à une légère diminution du niveau déxpression des gènes Susy , AGPase et GBSSI , qui sont régulés par le sucrose lors de la synthèse de lámidon. La production de mutanes (Chapitre 3) a été effectuée en exprimant le gène GtfI dans le génotype Kardal dont il est connu (Chapitre 2) que la production de dextranes a été la plus importante. De plus, afin dáméliorer le transport de GTFI vers lámyloplaste, GTFI a été tronquée en éliminant son domaine C-terminal, créant GTFICAT, dont des analyses in vitro ont montré que son activité est maintenue à 70 % après élimination de ce domaine. Léxpression de GtfICAT a abouti à une production plus importante de mutanes, en corrélation avec des altérations prononcées aux niveaux des tubercules et de lámidon, en comparaison de GtfI . Les plantes exprimant GtfICAT avaient des

126: Résumé tubercules montrant des régions brunâtres/ rougeâtres. De plus, il a été observé que les mutanes se déposaient sur les granules de lámidon, visualisés au moyen dúne solution rouge d´érythrosine. La morphologie des granules était dramatiquement altérée, montrant des surfaces érodées avec des trous. Pour les plantes exprimant fortement GtfICAT , la quantité dámidon était diminuée à cause d'une baisse du niveau déxpression du gène AGPase . De plus, le niveau déxpression du gène GBSSI , qui est aussi régulé par le sucrose, était diminué; cependant, cela ná pas abouti à une diminution de la quantité dámylose. Les propriétés rhéologiques des amidons GTFICAT semblaient être modifiées. Elles montraient une rétrogradation importante après gélatinisation, et avaient une viscosité accrue dénviron 1.7 après refroidissement par rapport aux amidons non transformés et à ceux issus de GTFI. Apparemment, la présence simultanée de mutane, dámylopectine, et dámylose semble créer un réseau plus visqueux de polymères en solution. La production dálternanes (Chapitre 4) a été démontrée au moyen déxpériences dÉLISA en exprimant le gène Asr dans le génotype Kardal. Les propriétés morphologiques et physico-chimiques des granules dámidon dÁSR n´étaient pas modifiées de manière significative et étaient comparables à celles de plantes non transformées. Cependant, léxpression du gène Asr réprimait celle des gènes AGPase et GBSSI . De nouveaux éléments structuraux comme des liaisons S-(1R3) alternant avec des S-(1R6) n´étaient pas trouvés attachés aux molécules dámidon, malgré que ASR catalyse des réactions dáccepteurs avec grande efficacité. Il est possible que la solubilité des polymères et láltération de la morphologie des granules dámidon soient corrélées pour les plantes exprimant Asr -, DsrS - and GtfICAT . Les alternanes sont très solubles dans léau en comparaison des dextranes (moyennement solubles), et des mutanes (insolubles). Lors de nos investigations, il est apparu que láccumulation de polymères insolubles dans léau affectait plus dramatiquement la morphologie des granules dámidon que celle de polymères plus solubles. Il a aussi été envisagé díntroduire de nouveaux éléments structuraux aux molécules dámidon en rapprochant GTFICAT et les granules dámidon en contact plus intime lún de láutre lors de la synthèse dámidon. Pour cela, SBD a été fusionné aux extrémités N- ou C-terminal de GTFICAT (Chapitre 5). Après avoir exprimé les deux différentes fusions dans le génotype Kardal, la présence de SBD à líntérieur des granules dámidon a été démontrée seulement dans les amidons SBD-GTFICAT, en

127: Résumé utilisant des anticorps spécifiques contre SBD. Additionellement, les amidons contenant le plus de SBD-GTFICAT montraient des altérations morphologiques et physico-chimiques moins importantes que ceux ne contenant que GTFICAT. Il semble que la fusion SBD influence de manière négative láctivité enzymatique de GTFICAT. Finalement, les granules dámidon exprimant le plus de SBD-GtfICAT ont été incubés avec une grande quantité de sucrose; à la suite de cela, la production in vitro de mutanes ná pas été démontrée. Des recherches supplémentaires ont besoin d´être effectuées afin dánalyser les nouvelles propriétés des amidons transgéniques, en particulier celles des plantes accumulant les mutanes. Durant ce travail, il a été possible dínfluencer la synthèse de lámidon en exprimant des gènes de glucansucrases dans les tubercules de pomme de terre, en particulier celui de GtfICAT et à un moindre degré celui de DsrS . De plus, líntroduction de nouveaux types de liaisons aux molécules dámidon ná pas été démontrée. Bien que ASR et DSRS catalysent des réactions dáccepteurs avec grande efficacité, il semble que láttachement de nouvelles liaisons glycosidiques demande des charactéristiques enzymatiques différentes de celles des glucansucrases. Il faudrait créer par modification génétique une seconde génération de glucansucrases exhibant des réactions dáccepteurs améliorées par rapport aux molécules dámidon.

129: Nawoord

Nawoord

En nu is dan ook het moment gekomen om iedereen, die op een of andere manier heeft bijgedragen aan de totstandkoming van dit boekje, te bedanken. Allereerst mijn promotor Richard Visser en co-promotor Jean-Paul Vincken. Richard, bedankt voor jouw vertrouwen in mij, je inspanningen tijdens het project en je morele steun tijdens de moeilijke momenten. Jean-Paul, bedankt voor je enthousiasme en je dynamisme gedurende het project, voor je positieve instelling en de goede samenwerking waarvan ik altijd heb kunnen genoten. Richard en Jean-Paul, jullie hebben ook een zeer belangrijke rol gespeeld in de totstandkoming van mijn eerste artikelen en het is te danken aan jullie "professionnalisme". Luc en Heleen, bedankt voor jullie gezelligheden en praktische suggesties in het lab. Luc, ook bedankt voor de leuke resultaten met de rheoscoop. Qin, you was not only my office mate at Binnenhaven but you also contributed for a large part to the SBD work. Thanks again and good luck in China. Farhad, thanks for your collaboration at the end of my PhD. Special thanks go to Denong Wang and Shaoyi Liu (Columbia University, New-York) who provided the tools to demonstrate the presence of dextran and alternan in potato juices. Thanks for the good collaboration and the fruitful results. In addition, I would like to thank Adrian Wiater (Department of Industrial Microbiology, Lublin, Poland) who provided me the mutan polymers and mutanase enzyme that I needed during this project. I would like to thank Bayer Bioscience (Germany) for its contribution and interest for this research. Zonder planten, geen onderzoek. Daardoor, wil ik Isolde, Dirk-Jan en de mensen van Unifarm bedanken voor het goed zorgen van de aardappelplanten. En dan mijn beide paranimfen Annie Marchal en Elly Janssen. Annie, nog bedankt voor je betrokkenheid en je warmte tijdens deze periode. Ook voor de leuke fietstochten in de Betuwe. Elly, ook bedankt voor je inspanningen tijdens mijn verhuizing naar België. Je stond altijd voor mij klaar en ik heb het altijd op prijs gesteld. Nu is het ook het moment gekomen om collega´s en vrienden te bedanken voor de gezellige momenten tijdens het werk en etentjes. Kortom, Jean-Paul & Marian, Elly & John, Monique & Maarten, Berlinda & Erik, Luisa & Rob, Ronald & Ruurd, Jaap, Wole,

130: Nawoord

Vivi, Erna, Sylvestre, Irma, Fien, Petra, Marian, Anne-Marie, Christian & Bea, Asun, Bjorn, Sanwen, Yuling, Jiang, Manolis. Bedankt ook aan Annie, Letty en Theo voor al die verschillende verrichte werkzaamheden tijdens het onderzoek. En nu wil ik ook vele mensen uit de Hyacintenstraat en Rooseveltlaan bedanken voor hun betrokkenheden en ondersteuningen: Tinecke, Arie, Linda, Rob, Teun, Oma, Marijke, Angela, Silvia, Ali, Riet en Gem. Siebe en Coby, ook heel veel dank voor jullie steun en eigenlijk schieten de woorden te kort. En nu Sjaak, ik heb dit proefschrift aan jou opgedragen want je hebt het verdiend; Je hebt alles voor mij gedaan om de studie mogelijk te kunnen maken, vanwege allerlei weerstanden. Je stond altijd achter mij en je hebt altijd mijn studiekeuze gewaardeerd. Je bent er niet meer, maar ik weet dat een stukje van jou in mij nog steeds leeft en ik wil het vandaag laten weten. Vandaag, is het ook de gelegenheid om mijn familie uit Frankrijk te bedanken, vooral oomen en tanten, Christophe, Karine, Gatiën, Maud, Jeanne en vader. Pour vous dire que ça était une longue route pour en arriver jusque là, mais j´ai enfin concrétisé ce projet complètement fou ! Het leven gaat door en ik ben nu met een andere stukje van mijn leven begonnen. Alain, je ne te connais pas depuis longtemps mais j´ai quand même voulu inscrire ton nom dans ce petit livre, car je préfère regarder devant moi…

131: Curriculum vitae & Publications

Curriculum vitae

Géraldine Armelle Kok-Jacon was born on March 26 th 1968 in Valence, France. In 1987 she completed High School in the Lycée la Saulaie, Saint Marcellin (France). Between 1987 and 1990, she studied Life and Natural sciences at the University Joseph Fourier in Grenoble. In 1994, she moved to the Netherlands and studied Laboratory Sciences at the Highschool of Amsterdam. In 1995, she started her University degree in Molecular and Cellular Biology at the Free University in Amsterdam. From 1998 to 1999, she carried out a research project on the —Post-transcriptional gene silencing mediated by promoterless inverted repeats in Arabidopsis thaliana at the Laboratory of Developmental Genetics of the Free University. In March 2000, she started her PhD research project which is described in this thesis, at the Laboratory of Plant Breeding, Wageningen University. In September 2004, she started as Post-Doc at the University of Louvain-la-Neuve (Belgium), à l´Unité de Biochimie physiologique.

Publications

Kok-Jacon, G.A., Ji, Q., Vincken, J-P., Visser, R.G.F., 2003. Towards a more versatile S-glucan biosynthesis in plants. J. Plant Physiol. 160, 765œ777. Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Wang, D., Liu, S., Visser, R.G.F., 2005. Production of dextran in transgenic potato plants. Transgenic Res. in press. Kok-Jacon, G.A., Vincken, J-P., Suurs, L.C.J.M., Visser, R.G.F., 2005a. Mutan produced in potato amyloplasts adheres to starch granules. Plant Biotechnol. J. 3, 341œ351.

133

Training and supervision plan of the Graduate School Experimental Plant Sciences

• Participation in postgraduate courses and workshops

a. Bioinformatics b. Protein engineering in Agro- and Food biotechnology c. Advanced food analysis d. Summer course Glycosciences e. Bioinformation Technology I f. English for PhDs

• Participation in international meetings

a. 10th International Symposium Plant Polysaccharides, Wageningen, the Netherlands (2000), attendance.

b. 12th European Carbohydrate Symposium, Grenoble, France (2003), attendance.

• Participation in national meetings

a. Annual EPS theme symposia (2000-2003), attendance.

b. ALW (Earth and Life Sciences) meetings (2000-2003), attendance.

c. EPS PhD students day (2001-2004), attendance.

d. Plant Breeding Seminar Series (2000-2004), attendance and presentations.

134

The work described in this thesis was carried out in the Graduate School Experimental Plant Sciences at the Laboratory of Plant Breeding, Wageningen University. The J.E. Jurriaanse Stichting financially supported the reproduction of this thesis.

Cover: Un jour de printemps dans les Vosges (painted by Sjaak).

Printed at Ponsen & Looijen B.V., Wageningen.