Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
J. Plant Physiol. 158. 479 – 487 (2001) Urban & Fischer Verlag http://www.urbanfischer.de/journals/jpp The control of amylose synthesis Kay Denyer*, Philip Johnson, Samuel Zeeman, Alison M. Smith John Innes Centre, Norwich Research Park, Colney, Norfolk NR4 7UH, UK Received October 16, 2000 · Accepted January 12, 2001 Summary Starch granules are composed of two types of glucose polymer, amylose and amylopectin, that differ in size and structure. One of the most intriguing challenges in understanding starch synthesis is to explain the apparently simultaneous synthesis of two such different polymers. One isoform of starch synthase, GBSSI, is responsible for amylose synthesis but can also contribute to amylopectin synthesis. The factors which determine the partitioning of GBSSI activity between these two processes are largely unknown. Understanding the properties of GBSSI and how these differ from the properties of the amylopectin-synthesising isoforms of starch synthase are important to the understanding of the control of amylose synthesis. In this review, we will describe how the synthesis of amylose and amylopectin are integrated and what factors may determine the relative amounts of these two polymers. Key words: starch – amylose – amylopectin – starch synthase – GBSSI – malto-oligosaccharide Abbreviations: GBSSI granule-bound starch synthase I. – SS starch synthase. – ADPG adenosine 5′-diphospho-glucose. – MOS malto-oligosaccharide. – dp degree of polymerization Introduction Starch is a major storage product of many of the seeds and other plant storage organs that we are most familiar with, those we eat. Yet despite the fact the starch is commonplace, we still do not completely understand the structure of starch or how it is synthesised. If we are to manipulate starch synthesis in crop plants to provide more starch or starches with particular desirable qualities, we must first understand the process of starch synthesis and how it is controlled. One of the most intriguing challenges in understanding how starch is made is to explain how the two very different polymers which constitute starch, amylose and amylopectin, are synthesised * E-mail corresponding author: [email protected] in the same place at the same time. Although there are still questions to be answered, some significant progress towards this goal has been made recently. For recent reviews of the control of amylopectin synthesis see Myers et al. (2000) and Smith et al. (1999). In this review, we will briefly describe the structure and composition of starch and then describe in more detail what is known about the mechanism of amylose synthesis and how the amount of amylose in starch might be controlled. The structure and composition of starch The basic unit of the starch granule is a branched polymer of glucose called amylopectin. Each molecule of amylopectin is composed of many short chains of glucose molecules. These 0176-1617/01/158/04-479 $ 15.00/0 480 Kay Denyer et al. 1,4-linked chains are joined together by 1,6 branch points and the chains are arranged into clusters in which adjacent chains form double helices (for review, see Thompson 2000). Amylopectin is one of the largest natural polymers, the molecular masses of amylopectin molecules from most plants being greater than 108 (Banks and Greenwood 1975). Amylopectin molecules are arranged radially within the starch granule with the ends of the chains pointing towards the surface (French 1984). The clusters of double-helical chains in adjacent amylopectin molecules can align themselves to form semi-crystalline arrays with in the granule. However, not all of the granule is crystalline. There are alternating concentric layers of amorphous and semi-crystalline amylopectin which are visible, in some circumstances, as growth rings within the granules. We do not know the orientation or arrangement of amylopectin with in the amorphous layers. For more details of the organisation of amylopectin within starch granules, see Gallant et al. (1997). It is possible for a starch granule to be composed entirely of amylopectin. However, most starches also contain amylose, another sort of glucose polymer. During storage organ development, the amylose content usually increases with age. For example in maize seeds, the proportion of the starch that was amylose increased from 18 % to 26.5 % between 14 and 28 days after pollination (Tsai et al. 1970). Most seed storage starches have a final amylose content of 25 – 30 % (Deatherage et al. 1955). Compared to amylopectin, amylose is a smaller molecule with longer chains and a limited number of branch linkages. The long chains of amylose have a high capacity to bind iodine in solution and this imparts a blue colour to amylosecontaining starches when these are stained with iodine. Amylopectin has a much lower iodine-binding capacity and stains red-brown with iodine solution. The amylose fraction of starch comprises both linear and branched molecules (Takeda et al. 1987). The ratio of linear to branched molecules varies according to the origin of the amylose. For example, in rice starch, two thirds of the amylose is unbranched. The number of chains per branched amylose molecule varies and is usually lower for cereal starches than for non-cereal starches (e.g. branched corn amylose has on average 5.3 chains whilst that from tapioca has 17.1; Takeda et al. 1993). For rice amylose, the average number of glucose units (dp) in the branched molecules is 1180 whilst that of the linear molecules is 740 (Takeda et al. 1993). The location of amylose within the amylopectin matrix of the granule is not entirely known. Amylose in solution very readily crystallizes but in the starch granule, amylose is not crystalline. It is supposed that amylose resides largely in the amorphous regions of the granule. In support of this, bluestaining rings have been observed in the starch granules of low-amylose potato tubers when stained with iodine (Kuipers et al. 1994). However, some amylose may also be present within the semi-crystalline areas of the starch granule (Jenkins and Donald 1994). The distribution of amylose within starch granules has been investigated by J-L Jane and colleagues. Partial gelatinization of potato starch granules showed that amylose at the periphery of the granule is more concentrated and has a smaller molecular size than amylose at the core (Jane and Shen 1993). Experiments with cross-linking agents (Jane et al. 1992) have shown that in potato and corn starch, individual amylose molecules are interspersed amongst the amylopectin molecules rather than being grouped together. When heated in water, starch forms a thick paste which makes it a valuable ingredient in many processed foods. The texture of cooked starch is influenced by its amylose content (White 1994). Starches with a normal amylose content form a paste which lacks clarity and tends to retrograde (recrystallize). This gives the starch paste very poor freeze – thaw stability. This retrogradation is due largely to the crystallization of amylose molecules. Consequently, starches with no amylose produce pastes with little tendency to retrograde. Also, amylose free starches are more easily gelatinized than starches with a normal amylose content and form clearer pastes. Starches from mutants with low starch-branching enzyme activity, such as the ae mutant of maize, have a higher than normal amylose content (50 – 60 % of the starch). These highamylose starches gelatinize at higher temperatures than normal starch and produce very viscous pastes with a strong tendency to retrograde. Separated amylose can be used to form strong transparent films and therefore high-amylose starches are being investigated as a potential source of a natural industrial polymer for the manufacture of, for example, biodegradable plastics. In the following sections we will describe in more detail the synthesis of the amylose component of starch. However, it is clear that the synthesis of amylose is intimately linked to and dependent upon the synthesis of amylopectin. The mechanism of amylose synthesis Both amylose and amylopectin are synthesised by enzymes called starch synthases. These add the glucose residue from ADPglucose to the non-reducing end of a growing glucan chain. Starch synthases can be divided into four classes on the basis of their primary amino acid sequences (SSI, SSII, SSIII and granule-bound SSI [GBSSI]). Accumulating evidence suggests that all four types of starch synthase are present in storage organs. The analysis of mutants and transgenics with reduced amounts of individual isoforms shows that each plays a unique role in the synthesis of amylose and amylopectin (Kossmann and Lloyd 2000, Smith et al. 1999). Only one of the isoforms of starch synthase, GBSSI, is required for amylose synthesis. There are well-characterized mutants of many species that lack amylose and all of these specifically lack GBSSI activity. The other isoforms of starch synthase can not substitute for GBSSI in the synthesis of amylose. They must, therefore, be primarily or entirely in- The control of amylose synthesis Figure 1. Domain comparison of starch synthase sequences from maize endosperm. Proteins sharing domains homologous to those present in maize SSIII were identified and aligned using the ProDom database of protein domain families (http://www.toulouse.inra.fr/ prodom.html). The four maize SS sequences shown are presented as examples of this alignment. The SwissProt ID of these sequences are: GBSSI, P04713; SSI, O49064; SSII, O48900; SSIII, O64923. Three domains were identified within these SS sequences by ProDom. Domain I is shown in black, domain II in grey and domain III is hatched. The length of the bars corresponds to the number of amino acids represented. volved in amylopectin synthesis. So to understand amylose synthesis we need to discover what is unique about GBSSI. In amino acid sequence, GBSSI shows strong similarity to the other isoforms of starch synthase (Ball et al. 1998, Kossman et al. 1999, Li et al. 1999). At present, there is no information to indicate which regions of GBSSI confer the unique ability to synthesise amylose or which regions of the other isoforms of starch synthase prevent this function. Sequence analysis (ProDom; Corpet et al. 1998) shows that all isoforms of starch synthase share three homologous domains (Fig. 1) and vary most in the length of the sequence at the N-terminal region, preceeding domain I. Domain I is found in no enzymes other than starch synthases. In contrast, domain II is also found, for example, in yeast α-amylase and domain III, which is a putative glucosyl transferase domain, is also found, for example, in sucrose synthase and sucrose phosphate synthase. Based on affinity labelling and site directed mutagenesis, it was shown that the lysine residue (lys 15) in a conserved motif (KTGGL) formed part of the ADPG binding site of E.coli glycogen synthase (Furukawa et al. 1990, 1993). ADPG is the glucosyl donor in the reaction catalysed by glycogen and starch synthases. This motif is conserved between glycogen and starch synthases (e.g. KTGGL in domain I of GBSSI from maize). Thus, it was assumed that this motif could be the ADPG binding site in the plant enzymes also. However, there is a very similar motif in domain III of the starch synthases (e.g. STGGLV in GBSSI from maize). Site directed mutagenesis of the lysine in this motif in domain III (lys 497) of maize SSIIa suggests that it, rather than the motif in domain I, may function as the ADPG binding site in plant starch synthases (Zhong et al. 2000). GBSSI is known to differ from the other starch synthase isoforms in its location within the plastid. GBSSI is almost en- 481 tirely starch granule-bound whereas other starch synthase isoforms and also isoforms of starch-branching enzyme are located both within the granules and in the surrounding stroma (Denyer et al. 1997b). There is good evidence to suggest that GBSSI is distributed throughout the granule interior, not just attached to the surface of the granule, and that at least part of the buried GBSSI is active. For example, GBSSI protein in isolated granules is inaccessible to proteases (Rahman et al. 1995) and GBSSI activity is not removed by treating granules with salts or detergents (Tanaka et al. 1967). These results show that GBSSI is buried within the amylopectin matrix. GBSSI activity can be measured in isolated intact granules showing that some of the buried GBSSI is active and that ADPG can penetrate to the site of GBSSI within the granule. However, the activity of GBSSI is increased if the granules are broken open by grinding (Hylton et al. 1996) or chemically gelatinised with urea (Frydman and Cardini 1967) showing that not all of the GBSSI within intact starch granules is active or accessible to ADPG. Evidence that GBSSI synthesises amylose within the granule matrix in vivo came from studies of transgenic potatoes (Visser et al. 1991, Kuipers et al. 1994, Tatge et al. 1999). In potato tubers in which GBSSI was specifically reduced by expression of antisense RNA, the amylose content of the starch was much reduced. However, rather than an even distribution of the small amount of amylose within these granules, iodine staining revealed that most or all of the amylose was confined to a central core region. Careful measurements of the proportion of the granule occupied by the amylose in tubers of different sizes from young and old plants showed that the core size increased as the granule size increased (Fig. 2). This showed that amylose is synthesised deep within the granule as the granule grows by deposition of amylopectin at the surface. In the transgenics with very little GBSSI, amylose synthesis is slow and lags behind amylopectin synthesis resulting in amylose being confined to the core regions of granules. In wild type plants with a higher GBSSI activity, the synthesis of amylose probably keeps pace with the syn- Figure 2. The relationship between developmental age and the proportion of the granule containing amylose in transgenic potatoes with reduced activity of GBSSI. The diagram is based on the work of Tatge et al. (1999). The granules were stained with iodine solution. This revealed a blue-staining core-region containing amylose (indicated in grey in the diagram) and a red-staining peripheral region (indicated in white) consisting of amylopectin only. The size of the blue core increased as the volume of the granule increased indicating that amylose is synthesised inside the amylopectin matrix of the granule. 482 Kay Denyer et al. thesis of amylopectin so that the synthesis of the newest amylose molecules occurs just under the granule surface and no peripheral amylose-free zone is visible. The fact that amylose synthesis occurs within the granule whereas the bulk of the synthesis of amylopectin occurs at the surface, could potentially explain aspects of the different structures of these polymers. It is likely, for example, that glucans synthesised inside the granule are not as susceptible to branching as glucans at the surface of the granule. However, this alone can not explain the unique role of GBSSI. There are isoforms of starch synthase other than GBSSI within starch granules (Denyer et al. 1997a). Mutations that eliminate GBSSI and amylose do not eliminate these isoforms from starch granules. This shows that being buried inside the granule is not the only requirement for an amylose-synthesising starch synthase. GBSSI must have other unique properties that allow it alone to synthesise amylose. The unique properties of GBSSI To gather information about the unique properties of GBSSI, we studied its activity in starch granules isolated from developing pea embryos (Denyer et al. 1996). Granules were supplied with radiolabelled ADPG then solubilized and the polymers were size-fractionated to separate the amylose and amylopectin components. We found that rather than incorporating glucose into amylose as it would in the plant, GBSSI in isolated granules incorporated glucose largely into the amylopectin fraction. We postulated that the reason for the lack of amylose synthesis in vitro was that small soluble glucans necessary for amylose synthesis had been washed out of the granules during their isolation. Evidence to support this idea came from experiments in which granules were prepared without the usual extensive washing. These showed higher rates of incorporation of glucose into amylose relative to amylopectin. Also when pure samples of malto-oligosaccharides (MOS) consisting of 2–7 glucose units (dp 2–7) were added to isolated granules, amylose synthesis was greatly stimulated. The stimulation of amylose synthesis by MOS is not unique to isolated pea granules. MOS stimulates amylose synthesis in granules isolated from potato (Denyer et al. 1996), maize (Denyer and Keeling, unpublished results) Arabidopsis (Zeeman and Smith, unpublished results) and the unicellular alga, Chlamydomonas reinhardtii (van de Wal et al. 1998). These experiments with isolated granules suggested a possible mechanism for amylose synthesis in which GBSSI, but not other isoforms of starch synthase within the granule, synthesises amylose by elongating MOS primers. The MOS must be able to diffuse into the granules to the site of GBSSI activity. There are several pieces of evidence to support this mechanism. First, granules containing GBSSI showed MOSdependent amylose synthesis whereas granules from the lam pea mutant, which contain SSII but not GBSSI, showed no amylose synthesis in the presence or absence of MOS (De- nyer et al. 1996). Second, in experiments with radiolabelled maltose (dp 2) and unlabelled ADPG, radiolabelled amylose was formed showing that the maltose was elongated (Denyer et al. 1999 a). Thus, MOS acts as a primer for amylose synthesis rather than as an activator of GBSSI which stimulates the enzyme to elongate glucans already present in the granule. Third, it is known that granules are permeable to molecules up to approximately 1 kDa in size, roughly the size of maltoheptaose (French 1984). In these initial experiments, we looked at the insoluble products of the granule-bound isoforms of starch synthase only. In later experiments (Denyer et al. 1999 a), we also looked at soluble products made by granule-bound starch synthases during incubations of isolated granules. These soluble products consisted of elongated MOS. These experiments showed that MOS diffuses into the granules, is elongated and then some or all of the elongated MOS diffuses back out of the granules. In experiments with granules containing only SSII, the soluble products were always only one glucose unit longer than the supplied MOS e.g. if we supplied maltotriose, the product was maltotetraose. In experiments with granules containing only GBSSI, products more than one glucose longer than the supplied MOS were also observed. This suggests that the mechanism of elongation by GBSSI is different from that of SSII. When SSII binds to maltotriose, for example, it adds a single glucose unit from ADPG to make maltotetraose and then the enzyme and product dissociate. SSII must then bind to another MOS, most likely another maltotriose molecule (since the concentration of this substrate in our assays was very much higher than the concentration of the product, maltotetraose) to make a second product molecule. In contrast, when GBSSI binds to maltotriose and elongates this to give maltotetraose, it does not always then dissociate from the product. In some cases, GBSSI is able to add another glucose unit from ADPG to the maltotetraose to make maltopentaose and then another to make maltohexaose and so on. Elongation in this manner is called processive elongation. This processive reaction mechanism may, in part, explain the unique ability of GBSSI to synthesise amylose. However, there are probably other equally important factors that allow amylose synthesis to take place. First, GBSSI within starch granules has much higher affinity for MOS than does SSII (at least 20-fold higher; Denyer et al. 1999 a). This means that at the low concentrations of MOS likely to be present in vivo (see later), GBSSI will elongate MOS much more actively than SSII. Second, the limited permeability of the amylopectin matrix allows short MOS to diffuse readily in and out of the granule but longer MOS synthesised by GBSSI can not escape from the granule. The trapping of long MOS within the granule means that these molecules can be further elongated by GBSSI to eventually form mature amylose molecules. In experiments in which GBSSI was expressed in soluble form in E. coli, it was observed that, unlike the granule-bound The control of amylose synthesis 483 switched from the A-type crystal structure normally found in Chlamydomonas to the B-type. This work on Chlamydomonas starch led to an alternative model for amylose synthesis, not involving MOS (van de Wal et al. 1998, Ball et al. 1998). In this model, GBSSI elongates amylopectin to create extremely long chains. These amylosesized chains are then cleaved from the amylopectin by an as yet unidentified hydrolyase activity. What prevents the putative hydrolase from cleaving glucan chains smaller than amylose from the amylopectin is not clear. In similar experiments with starch from pea embryos, no transfer of radioactive glucose from the amylopectin to the amylose fraction was observed (Denyer et al. 1999 a). However, GBSSI in Chlamydomonas, as in higher plants, can elongate MOS in vitro to generate amylose. Thus, in Chlamydomonas, amylose can be generated in vitro by two different mechanisms whilst in higher plants, amylose synthesis from MOS only has been demonstrated. Which of these two mechanisms, if either, operates in vivo is not yet known. Figure 3. A model to explain the synthesis of amylose from malto-oligosaccharide primers. GBSSI binds tightly to the amylopectin matrix of starch granules. In this location, it has a high affinity for MOS. It elongates MOS processively to make long glucan chains. These long glucans are trapped within the amylopectin matrix and are further elongated to form amylose. enzyme, soluble GBSSI has a low affinity for MOS and does not elongate processively (Denyer et al. 1999 b). However, when supplied with a small amount of amylopectin, the affinity of GBSSI for MOS increases and the enzyme becomes processive. So these important and unique kinetic properties of GBSSI may be conferred on the enzyme by its interaction with the amylopectin matrix of the granule. To summarise this work, we suggest a model to explain the unique ability of GBSSI to make amylose (Fig. 3). GBSSI can elongate amylopectin chains A number of studies of isolated starch granules have shown that GBSSI incorporates glucose into amylopectin as well as into amylose. The glucose is incorporated specifically into the longest amylopectin chains (Baba et al. 1987, Denyer et al. 1996, van de Wal et al. 1998) suggesting that, as well as synthesizing amylose, GBSSI in vivo may contribute to the determination of the structure of amylopectin. In experiments with starch granules isolated from nitrogen-starved Chlamydomonas cells, van de Wal et al. (1998) showed that the radioactive glucose initially incorporated into amylopectin was, after extended periods of incubation (6 h), transferred to a lower molecular weight fraction similar in size and structure to amylose. After even longer periods of incubation (24 – 48 h), more label was transferred from amylopectin to amylose and the crystallinity of the starch decreased from 27 % to 16 % and Are malto-oligosaccharides present in starch synthesising cells? For amylose synthesis to occur from a MOS primer in vivo, there must be a sufficient concentration of MOS available to GBSSI within the plastid (0.1mmol/L maltotriose gives half the maximal rate of amylose synthesis for isolated pea starch granules, Denyer et al. 1999 a). In isolated granules, the elongation of MOS occurs at the expense of amylopectin elongation. Thus, the extent to which GBSSI uses amylopectin or MOS as a substrate may depend upon the availability of MOS. There are two pieces of evidence that MOS is present in plant cells. First, heat-treated, soluble extracts of potato tubers stimulated amylose synthesis when added to isolated potato starch granules together with ADPG (Denyer et al. 1996). The stimulatory material in the soluble extracts was destroyed by pre-incubation with amyloglucosidase indicating that it was a glucan. Second, direct measurements of MOS have been made for several plant organs. For example, there are low levels of MOS (0.05 – 0.10 mg glucose-equivalents g –1 fresh weight), mostly in the form of maltose, in the leaves of Arabidopsis thaliana throughout the day – night cycle (Critchley et al. 2000). If certain assumptions are made (that all of the MOS is maltose, that the maltose is confined to the plastid, that 1 g of tissue has a volume of 1 mL, and that the stroma occupies 7 % of the total cell volume) then it is possible to estimate that concentration of MOS in the chloroplasts of Arabidopsis leaves is 2 – 4 mmol/L. This is probably sufficient to allow amylose synthesis to occur by MOS elongation. In a mutant of Arabidopsis which lacks disproportionating enzyme, MOS accumulates to very high levels, particularly at night (up to 15-times higher than the MOS content of the wild 484 Kay Denyer et al. type). Interestingly, the mutant accumulates more amylose in its starch than the wild type. This may indicate that the level of MOS in wild type Arabidopsis leaves is limiting for amylose synthesis. However, there are metabolic differences other than the increase in MOS concentration between the mutant and wild type leaves and these may also differentially influence amylose synthesis. MOS will be generated in vivo when starch is degraded or modified. In the plastids of starch-synthesising storage organs and in chloroplasts in the light period, significant activities of several starch-degradative enzymes are known to be present. These include amylases, starch-debranching enzymes and disproportionating enzyme (Bulpin and ap Rees 1978, Doehlert and Knutson 1991, Duffus and Rosie 1973, Takaha et al. 1993). A limited amount of starch degradation during the phase of starch accumulation in storage organs can not be ruled out and may be involved in the creation of the ordered structure of the amylopectin molecule (for review see Myers et al. 2000, Smith 1999). The trimming of branched glucans, for example by isoamylase, could contribute some of the MOS required for amylose synthesis. It is not known whether enzymes with hydrolytic activity exist within the granule. If they do exist, they might also contribute to the pool of MOS available to GBSSI. What controls the amount of amylose? In the previous section we indicated that, in some circumstances, the availability of MOS could limit amylose synthesis. In this section, we will discuss other factors that could potentially control amylose synthesis. These include the existence of different forms of GBSSI that could potentially have different properties and functions, the activity of GBSSI, the availability of the substrate ADPG and the properties of the amylopectin matrix of the starch granule. There is increasing evidence that different GBSSI genes are expressed in different parts of the plant. Examination of the granule-bound proteins in different parts of lam mutant and wild type peas plants showed that only lam embryos entirely lacked a GBSSI-like protein (Denyer et al. 1997a). In the other parts of the plant, we found that a GBSSI-like protein was present even in the mutant. This protein is the product of a second GBSSI gene which we called GBSSIb. GBSSIb is expressed in leaves, pods and nodules but not in pea embryos. In pods, both GBSSI and GBSSIb are present but GBSSIb accounts for most of the granule-bound starch synthase activity. Studies of wheat have also shown that GBSSIlike proteins are differentially expressed in different parts of the plant. The GBSSI in wheat endosperm is a different gene product from that in the rest of the plant (Fujita and Taira 1998, Nakamura et al. 1998). The amylose content of starches from different parts of the plant varies, with storage starches tending to have more amylose than transitory starches. For example, in pea plants, we found that the starch of leaves, pods and nodules had far less amylose than that of the embryos (Denyer et al. 1997 a). The existence of different forms of GBSSI in the various parts of a plant raises the possibility that there may be some differentiation of function within the GBSSI-class of starch synthases. However, the properties of GBSSI from pea embryo starch and GBSSIb from pod starch of the lam pea mutant are very similar. First, the enzymes are present in these starches in similar amounts and activities. Second, both GBSSI and GBSSIb can synthesise amylose in vitro when isolated granules are supplied with MOS. Thus, the differing amylose contents of starches from different parts of the plant are not obviously related to the fact that there are different forms of GBSSI. Studies of mutant and transgenic plants have shown that the activity of GBSSI in the storage organs of wild type plants does not limit amylose synthesis. For example, in maize seeds, Tsai (1974) showed that GBSSI activity was linearly proportional to the number of Wx alleles present in the endosperm. However, the amylose content was not linearly related to the activity of GBSSI. Endosperms containing two thirds of the normal GBSSI activity have only slightly lower amylose content than normal endosperms. This implies that increasing the GBSSI content above normal would result in little, if any, increase in amylose content (Fig. 4). Studies of potatoes with reduced GBSSI activity have shown a similar picture (Visser et al. 1990, Flipse et al. 1996). The GBSSI activity had to be reduced considerably below the wild type level before there was a significant effect on amylose content. Another potential factor controlling amylose synthesis is the availability of the substrate, ADPG. There is good theoretical and experimental evidence to show that the availability of ADPG could limit amylose synthesis in vivo (Clarke et al. 1999). First, GBSSI has a low affinity for ADPG (e.g. pea starch granules, apparent kM = 1.4 mmol/L) compared to that of the starch synthase isoforms responsible for amylopectin Figure 4. The dosage effect of Wx genes on the production of GBSSI activity and amylose content. Data are taken from Tsai (1974). The Wx gene encodes an active GBSSI. Wild type maize endosperm, being triploid, has three copies of the Wx gene. A mutant maize that lacked GBSSI activity (zero Wx genes) was crossed to wild-type maize to create plants with one or two copies of the Wx gene. The GBSSI activity is indicated by the circles and the amylose content by squares. The control of amylose synthesis synthesis (kM = 0.06 – 0.6 mmol/L, data from many species). Estimates of the concentration of ADPG in plastids of developing pea embryos show that the ADPG concentration (0.9 mmol/L) is much lower than that required to saturate the GBSSI reaction. However, the ADPG concentration in plastids is close to that required to saturate the reaction catalysed by the amylopectin-synthesising isoforms of starch synthase. This means that if the ADPG concentration were to fall, amylose synthesis would be reduced to a greater extent than would amylopectin synthesis. Second, work on some low-starch mutants of pea shows that reductions in ADPG content are accompanied by reductions in amylose content (Clarke et al. 1999). Mutations at the rb, rug3 and rug4 loci of pea affect enzymes involved in ADPG production and consequently, the ADPG content of these embryos is lower than that of wild type embryos. These mutations in pea do not affect enzymes directly involved in starch synthesis and yet all cause a decrease in the proportion of amylose in the starch (Fig. 5). In Chlamydomonas, mutations at the STA1 and STA5 loci affect enzymes involved in ADPG production and cause a reduction in the amylose content of the starch (van den Koornhuyse et al. 1996). Reduced amylose contents were also observed in the starches of potato tubers in which the activity of AGPase (the enzyme responsible for the synthesis of ADPG) had been reduced by the expression of antisense constructs (Lloyd et al. 1999). Although the ADPG contents of these Chlamydomonas mutants and antisense potato tubers were not measured it is assumed that the decrease in the rate of ADPG synthesis would result in a decrease in the concentration of this metabolite. This supports the idea that, at least in some circumstances, the concentration of ADPG can influence the amylose content. 485 The nature of the amylopectin matrix could influence the amylose content of the granule in several ways. We have shown that amylopectin influences GBSSI activity directly, conferring kinetic properties which favour the elongation of MOS. The permeability of the matrix to the substrates, ADPG and MOS may also influence GBSSI activity. For example, the concentration of soluble substrates for GBSSI may decrease with increasing distance from the surface of the granule limiting its activity particularly in the centre of the granule. The space available within the amylopectin matrix may also be important. This could set an upper limit on the amount of amylose which can be synthesised. This may be the reason why wild type storage starches all have an amylose content of approximately 25 – 30 % and no higher (Flipse et al. 1996). Conclusion We have learnt much about the unique properties of GBSSI that enable this enzyme to synthesise amylose. However, there are still some remaining questions about the role of GBSSI. It is clear that GBSSI can elongate amylopectin chains within the granule as well as synthesising amylose. Whether these amylopectin chains or soluble MOS derived from them are the primers for GBSSI in vivo has still to be resolved. The study of starch synthesis in different plant organs makes it clear that starch granules containing GBSSI activity do not necessarily synthesize amylose. In granules that accumulate very little amylose, GBSSI may contribute primarily to the synthesis of amylopectin or it may contribute little to the synthesis of any component of the starch granule. Further studies of the roles of GBSSI in different sorts of starch synthesising plant organs are required to answer these questions. Acknowledgements. Financial support by the Agri-Food directorate of the BBSRC, U.K. (SZ) and Dupont Agricultural products (PJ) is gratefully acknowledged. References Baba T, Yoshii M, Kainuma K (1987) Acceptor molecule of granulebound starch synthase from sweet-potato roots. Starch/Stärke 39: 52 – 56 Figure 5. Measurements of the amylose and ADPG contents of developing wild type and mutant pea embryos. Data are taken from Clarke et al. (1999). The mutations affect enzymes involved in the supply of ADPG for starch synthesis. The rb mutant embryos have a reduced activity of ADPglucose pyrophosphorylase (AGPase; squares), rug3 embryos lack plastidial phosphoglucomutase activity (PGM; triangles) and rug4 embryos have a reduced activity of sucrose synthase (Susy; circles). The values for wild type embryos are indicated by diamonds. A logarithmic curve was fitted to the combined data (R2 = 0.56). Ball S, van de Wal MHBJ, Visser RGF (1998) Progress in understanding the biosynthesis of amylose. Trends Plant Sci 3: 462 – 467 Banks W, Greenwood CT (1975) Starch and its components. Edinburgh University Press. Bulpin PV, ap Rees T (1978) Starch breakdown in the spadix of Arum maculatum. Phytochemistry 17: 391– 396 Clarke BR, Denyer K, Jenner CF, Smith AM (1999) The relationship between the rate of starch synthesis, the adenosine 5′-diphosphoglucose concentration and the amylose content of starch in developing pea embryos. Planta 209: 324 – 329 486 Kay Denyer et al. Corpet F, Gouzy J, Kahn D (1999) The ProDom database of protein domain families. Nucl Acids Res 26: 323 – 326 Critchley JH, Zeeman SC, Takada T, Smith AM, Smith SM (2000) A critical role for disproportionating enzyme in starch breakdown but not in starch synthesis is revealed by a knock-out mutation in Arabidopsis (submitted) Deatherage WL, MacMasters MM, Rist CE (1955) A partial survey of amylose content in starch from domestic and foreign varieties of corn, wheat and sorghum and from some other starch bearing plants. Trans Am Assoc Cereal Chem 13: 31– 42 Denyer K, Clarke B, Hylton C, Tatge H, Smith AM (1996) The elongation of amylose and amylopectin chains in isolated starch granules. Plant J 10: 1135–1143 Denyer K, Barber LM, Edwards EA, Smith AM, Wang TL (1997 a) Two isoforms of the GBSSI class of granule-bound starch synthase are differentially expressed in the pea plant. Plant, Cell Environ 20: 1566–1572 Denyer K, Edwards A, Martin C, Smith AM (1997b) The mechanism of amylose synthesis. In: Frazier PJ, Donald AM, Richmond P (eds) Starch: structure and functionality. The Royal Society of Chemistry, Cambridge, UK, pp 222 – 229 Denyer K, Waite D, Motawia S, Moller B-L, Smith AM (1999 a) Granule-bound starch synthase I in isolated starch granules elongates malto-oligosaccharides processively. Biochem J 340: 183–191 Denyer K, Waite D, Edwards A, Martin C, Smith AM (1999 b) Interaction with amylopectin influences the ability of granule-bound starch synthase I to elongate malto-oligosaccharides. Biochem J 342: 647– 653 Doehlert D, Knutson CA (1991) Two classes of starch debranching enzymes from developing maize kernels. J Plant Physiol 138: 566 – 572 Duffus C, Rosie R (1973) Starch hydrolysing enzymes in the developing barley grain. Planta 109: 153–160 Flipse E, Keetels CJAM, Jacobsen E, Visser RGF (1996) The dosage effect of the wildtype GBSS allele is linear for GBSS activity but not for amylose content: absence of amylose has a distinct influence on the physico-chemical properties of starch. Theor Appl Genet 92: 121–127 French D (1984) Organisation of starch granules. In: Whistler RL, BeMiller JN, Paschall JF (eds) Starch: Chemistry and Technology. Academic Press, Orlando, FL, pp 183 – 248 Frydman RB, Cardini CE (1967) Studies on the biosynthesis of starch. II. Some properties of the adenosine diphosphate glucose:starch glucosyltransferase bound to the starch granule. J Biol Chem 242: 312 – 317 Fujita N, Taira T (1998) A 56-kDa protein is a novel granule-bound starch synthase existing in the pericarps, aleurone layers, and embryos of immature seed in diploid wheat (Triticum monococcum L.). Planta 207: 125–132 Furukawa K, Tagaya M, Inouye M, Preiss J, Fukui T (1990) Identification of lysine 15 at the active site in Escherichia coli glycogen synthase. J Biol Chem 265: 2086 – 2090 Furukawa K, Tagaya M, Tanizawa K, Fukui T (1993) Role of the conserved Lys-X-Gly-Gly sequence at the ADP-glucose-binding site in Escherichia coli glycogen synthase. J Biol Chem 268: 23837– 23842 Gallant DJ, Bouchet B, Baldwin PM (1997) Microscopy of starch: evidence of a new level of granule organization. Carbohydr Polymers 32: 177–191 Hylton CM, Denyer K, Keeling PL, Chang M-T, Smith AM (1996) The effect of waxy mutations on the granule-bound starch synthases of barley and maize endosperms. Planta 198: 230 – 237 Jane J-L, Xu A, Radosavljevic M, Seib PA (1992) Location of amylose in normal starch granules. I. Susceptibility of amylose and amylopectin to cross-linking reagents. Cereal Chem 69: 405 – 409 Jane J-L, Shen JJ (1993) Internal structure of the potato starch granule revealed by chemical gelatinization. Carbohydr Res 247: 279 – 290 Jenkins PJ, Donald AM (1994) The influence of amylose on starch granule structure. Int J Biol Macromol 17: 315 – 321 Kossmann J, Abel GJW, Springer F, Lloyd JR, Willmitzer L (1999) Cloning and functional analysis of a cDNA encoding a starch synthase from potato (Solanum tuberosum L.) that is predominantly expressed in leaf tissue. Planta 208: 503 – 511 Kossmann J, Lloyd J (2000) Understanding and influencing starch biochemistry. Critical Rev in Plant Sci 19: 171– 226 Kuipers AGJ, Jacobsen E, Visser RGF (1994) Formation and deposition of amylose in the potato tuber are affected by the reduction of granulebound starch synthase gene expression. Plant Cell 6: 43 – 52 Li Z, Chu X, Mouille G, Yan L, Kosar-Hashemi B, Hey S, Napier J, Shewry P, Clarke B, Appels R, Morell MK, Rahman S (1999) The localization and expression of the class II starch synthases of wheat. Plant Physiol 120: 1147–1155 Lloyd JR, Springer F, Buléon A, Müller-Rober B, Willmitzer L, Kossmann J (1999) The influence of alterations in ADP glucose pyrophosphorylase activities on starch structure and composition in potato tubers. Planta 209: 230 – 238 Myers AM, Morell MK, James MG, Ball SG (2000) Recent progress towards understanding biosynthesis of the amylopectin crystal. Plant Physiol 122: 989 – 997 Nakamura T, Vrinten P, Hayakawa K, Ikeda J (1998) Characterization of a granule-bound starch synthase isoform found in the pericarp of wheat. Plant Physiol 118: 125–132 Rahman S, Kosar-Hashemi B, Samuel MS, Hill A, Abbot DC, Skerritt JH, Preiss J, Appels R, Morell MK (1995) The major proteins of wheat endosperm starch granules. Aust J Plant Physiol 22: 793 – 803 Smith AM (1999) Making starch. Curr Opin Plant Biol 2: 223 – 229 Smith AM, Denyer K, Zeeman SC, Edwards A, Martin C (1999) The synthesis of the starch granule. In: Burrell M, Bryant J, Kruger N (eds) Plant Carbohydrate Biochemistry. BIOS Scientific Publishers Ltd, Oxford UK, pp 1–11 Takaha T, Yanase M, Okada S, Smith SM (1993) Disproportionating enzyme (4-α-glucanotransferase; EC 2.4.1.25) of potato. J Biol Chem 268: 1392–1396 Takeda Y, Hizukuri S, Takeda C, Suzuki A (1987) Structures of branched molecules of amyloses of various origins, and molecular fractions of branched and unbranched molecules. Carbohydr Res 165: 139–145 Takeda Y, Tomooka S, Hizukuri S (1993) Structures of branched and linear molecules of rice amylose. Carbohydr Res 246: 267– 272 Tanaka Y, Minagawa S, Akazawa T (1967) Association of enzyme proteins with starch granules in rice grains. Starch/Stärke 7: 206 – 212 Tatge H, Marshall J, Martin C, Edwards EA, Smith AM (1999) Evidence that amylose synthesis occurs within the matrix of the starch granule in potato tubers. Plant, Cell Environ 22: 543 – 550 Thompson DB (2000) On the non-random nature of amylopectin branching. Carbohydr Polymers 43: 223 – 239 Tsai C-Y (1974) The function of the waxy locus in starch synthesis in maize endosperm. Biochem Genet 11: 83 – 96 The control of amylose synthesis Tsai CY, Salamini F, Nelson OE (1970) Enzymes of carbohydrate metabolism in the developing endosperm of maize. Plant Physiol 46: 299 – 306 van den Koornhuyse N, Libessart N, Delrue B, Zabawinski C, Decq A, Iglesias A, Carton A, Preiss J, Ball S (1996) Control of starch composition and structure through substrate supply in the monocellular alga Chlamydomonas reinhardtii. J Biol Chem 271: 16281– 16287 van de Wal M, D’Hulst C, Vincken J-P, Buléon A, Visser R, Ball S (1998) Amylose is synthesised in vitro by extension of and cleavage from amylopectin. J Biol Chem 273: 22232 – 22240 Visser RGF, Feenstra WJ, Jacobsen E (1990) Manipulation of granulebound starch synthase activity and amylose content in potato by 487 antisense genes. In: van der Krol AR (ed) Application of antisense nucleic acids. Marcel Dekker Inc, New York, pp 141–155 Visser RGF, Somhorst I, Kuipers GJ, Ruys NJ, Feenstra WJ, Jacobsen E (1991) Inhibition of expression of the gene for granule-bound starch synthase in potato by antisense constructs. Mol Gen Genet 225: 289 – 296 White P (1994) Properties of corn starch. In: Hallauer AR (ed) Specialty corns. CRC press Inc, Boca Raton FL, pp 29 – 54 Zhong G, Shibles R, Keeling P, Guan H (2000) Lysine 497 instead of 193 of the conserved KTGGL motif functions as the ADP-glucose binding site of maize starch synthase IIa. Final Program and abstract supplement, Plant Biology 2000, the Annual Meeting of the American Society of Plant Physiologists. Poster 600: 127