WO1994004693A2 - Novel plants and processes for obtaining them - Google Patents

Abstract

Plants with an altered starch synthesising ability are produced by incorporating into the genome of the plant at least one donor gene encoding a starch primer. The starch primer is an enzyme capable of inititiating starch synthesis, such as amylogenin and/or glycogenin and/or protoglycogenin. DNA constructs encoding a starch primer are provided.

Description

NOVEL PLANTS AND PROCESSES FOR OBTAINING THEM

This invention relates to novel plants having an altered ability to synthesise starch, and to processes for obtaining such novel plants.

Starch is an important end-product of carbon fixation during photosynthesis in leaves and is an important storage product in seeds and fruits. Starch comprises up to 75% of the grain dry weight in cereals. In economic terms, the starch produced by the edible portions of three grain crops, wheat, rice and maize, provide approximately two-thirds of the world's food calculated as calories.

Starch is synthesised in the plastid compartment (the chloroplast in photosynthetic cells or the amyloplast in non-photosyntiietic cells). The biochemical pathway of starch biosynthesis in leaves has been well-characterised and is shown in Figure 1. The abbreviations used are: G-3-P, glyceraldehyde-3-phosphate; DHAP, dihydroxyacetone phosphate; P., orthophosphate; PP., inorganic pyrophosphate; ADPG/ADPglucose, adenosine diphosphate glucose; ATP, adenosine triphosphate; UDPG/UDPglucose, uridine diphosphate glucose. The reactions are catalysed by the following enzymes: 1) phosphoglycerate kinase/glyceraldehyde-3-phosphate dehydrogenase, 2) triose-phosphate isomerase, 3) aldolase, 4) fructose-l,6-bisphosphatase, 5) hexose phosphate isomerase, 6) phosphoglucomutase, 7) ADP-glucose pyrophosphorylase, 8) starch synthase, 9) UDP-glucose pyrophosphorylase, 10) sucrose phosphate synthase, 11) sucrose phosphatase, 12) orthophosphate/ triose phosphate translocator, 13) inorganic pyrophosphatase.

Knowledge of the pathway in storage tissues was greatly improved following the development of a method for the isolation of intact amyloplasts from wheat endosperm (apRees, 1990) and our recent research with 13C labelling studies in wheat and maize (Keeling e_t l_^, 1988, Plant Physiology, 87:

311-319; Keeling, 1990, ed. CD. Boyer, J.C.

Shannon and R.C. Harrison, pp.63-78, being a presentation at the 4th Annual Penn State Symposium in Plant Physiology, May 1989). Figure 2 shows the metabolic pathway of starch biosynthesis in maize endosperm. The abbreviations used are the same as in Figure 1. The reactions are catalysed by the following enzymes: 1) sucrose synthase, 2)

UDP-glucose pyrophosphorylase, 3) hexokinase, 4) phosphoglucomutase, 5) hexose-phosphate isomerase,

6) ATP-dependent phosphofructokinase, 7)

PP.-dependent phosphofructokinase, 8) aldolase, 9) triose-phosphate isomerase, 10) hexose-phosphate translocator, 11) ADP-glucose pyrophosphorylase,

12) starch synthase, 13) sucrose phosphate synthase, 14) sucrose phosphatase.

Despite this considerable progress, the first priming molecule(s) responsible for initiating starch synthesis in plants have remained an enigma.

In the animal kingdom, glycogen is deposited rather than starch. It is now well established that glycogen synthesis is initiated on a protein primer which remains covalently attached to the mature macromolecule. Recent work on glycogen deposition has identified a new enzyme: glycogenin or self glucosylating protein (SGP) which acts as a primer. The protein primer serves as the anchor on which the oligosaccharide is constructed and on which glycogen is built (Kennedy et al, 1985, Membranes and Muscle ed. by MC Berman et al, IRL Press Oxford/ICSU Press, pp65-84). The linkage between the oligosaccharide and glycogenin involves a novel carbohydrate-protein bond, via the hydroxyl group of tyrosine (Rodriguez and Whelan, 1985, Biochem Biophys Res Commun, 132:829-836). The discovery of glycogenin as a covalent component of the glycogen molecule was followed by the discovery in muscle and other tissues of a glycogen-free form that could be revealed by glucosylation from

14 C-labelled UDPglucose (Rodriguez et al, 1986,

Proc 18th Miami Winter Symposium, ICSU Short

Reports, 4:96-99). On purification to homogeneity, the protein proved to be autocatalytic. It glucoεylates itself and in doing so constructs a maltosaccharide chain on itself that primes glycogen synthesis by glycogen synthase and branching enzyme (Lo ako et al, 1988, FASEB J,

2:3097-3103). The self-glucosylating protein (SGP) in muscle iε not present as such, but is part of a larger molecule (proglycogen) (Lomako et al, 1990,

FEBS Lett, 268:8-12; 1991, FEBS Lett, 279:223-228).

The linkage of glycogen to tyrosine in glycogenin has been confirmed by Smythe et al (1988, EMBO J,

7:2681-2686), as has the autocatalytic nature of glycogenin (Pitcher et al, 1988, Eur J Biochem,

176:391-395). The glycogenin protein from rabbit skeletal muscle has been seguenced by Campbell and

Cohen (1989, Eur J Biochem, 185:119-125).

The formation of a glycoprotein primer may be a universal feature for the synthesis of polysaccharides. Several proteins have been identified as potential priming molecules for carbohydrate synthesis. None of these proteins, however, has been associated with a polysaccharide macromolecule. Other than mammalian glycogen, the presence of a covalently bound protein (primer) in a mature carbohydrate molecule has yet to be demonstrated.

According to the present invention there is provided a method of producing a plant with altered starch synthesising ability comprising stably incorporating into the genome of a recipient plant at least one donor gene encoding a starch primer.

This invention is based on our discovery that the enzyme reactions catalyzed by glycogenin are also important in the starch biosynthetic pathway of plant organs. We have identified a new protein responsible for initiating starch synthesis in plants and have named this plant primer enzyme, amylogenin or self-glucosylating protein (SGP). We have therefore redefined the metabolic pathway of starch biosynthesis in plants to incorporate the starch priming reaction catalyzed by amylogenin. Figure 3 shows the new pathway of starch synthesis in maize endosperm. The abbreviations used are: UDPGppase (UDP-glucose pyrophosphorylase); ADPGppase (ADP-glucose pyrophosphorylase); BSS (bound starch synthase); SSSynthase (soluble starch synthase); BE (branching enzyme).

Starch (amylose and amylopectin) will be synthesized only where the initiator protein (primer) is present and in amounts corresponding to the amount of primer. Amylogenin may determine how much starch is laid down and where it is laid down. It is therefore possible to genetically manipulate the starch content of plants by altering the level and distribution of amylogenin. The genetic manipulation of starch-bearing plants of commercial importance, particularly cereals such as maize, may result in new strains that are advantageous either for direct food uses, or other uses of starch, or as sources of starch to be used as a chemical feedstock. For example, the ratio of amylose and amylopectin in corn starch may be altered which would create starches of considerable utility because of the diverse properties endowed on the different starches as thickening agents, edible food wrapping, etc. This route to new plant varieties complements the traditional and empirical processes of plant breeding.

The basic technology is of potential" application to the manipulation of any storage polysaccharide and includes the possibility to suppress polysaccharide synthesis, as well as to enhance it, thereby altering the carbohydrate profile of the plant in directions that may be more favourable for nutritional or industrial uses.

The method according to the invention is generally applicable to all plants producing or storing starch. The recipient plant may be: a cereal, in particular members of the family Gramineae such as maize (corn), wheat, rice, sorghum or barley; a fruit-producing species such as banana, apple, tomato or pear; a root crop such as cassava, potato, yam or turnip; an oilseed crop such as rapeseed, canola, sunflower, oil palm, coconut, linseed or groundnut; a meal crop such as soya, bean or pea; or any other suitable species.

Each donor gene encodes a starch primer (a protein molecule capable of initiating starch synthesis). The primer may be amylogenin and/or glycogenin, including amylogenin of plant origin and/or glycogenin or protoglycogenin of bacterial origin and/or glycogenin of fungal origin and/or glycogenin of animal origin.

A donor gene may be derived from any suitable source including a plant, a bacterium, a fungus or an animal cell. For example, suitable sources of plant genes include plants of the species Zea mays, Zea diploperennis, Zea luxurians, Zea perennis, Zea tripsacum, Zea parviglumis, Zea exicana or teosinte.

A donor gene may be incorporated into the recipient plant genome by sexual crossing of the recipient plant and a sexually compatible donor plant. Alternatively, a donor gene may be isolated and then incorporated into the recipient plant genome by genetic transformation.

The donor gene may be derived from cDNA or genomic DNA (gDNA) encoding a starch primer, or it may be synthesised ab initio using standard techniques. Typically the donor gene encodes the complete primer in the sense orientation so that the transcription product (mRNA) can be translated into the active starch primer enzyme. Alternatively, the donor gene may encode a portion of the starch primer in the sense orientation or may encode some or all of the primer in the antisense orientation so that the transcribed mRNA inhibits expression of the primer enzyme. It is possible to insert more than one copy of the donor gene into the recipient genome. At least one of the donor genes may encode a modified allelic form of the starch primer enzyme having altered characteristics (such as increased or decreased activity, or differing interactions with other enzymes) .

It can be seen that the recipient plant's starch synthesising ability may be altered in various ways depending on the particular donor gene(s) which are incorporated. The total amount of starch produced (starch yield) may be increased or decreased; the rate of starch synthesis may be increased or decreased; the starch produced may have an altered structure; the ratio of amylose to amylopectin may be altered; the temperature optimum of starch synthesis may be increased or decreased; the capacity to produce starch at an elevated or lowered temperature may be improved. These effects on starch synthesis are due directly or indirectly to altered activity of the starch primer enzyme(s). For example, interaction of the starch primer enzyme with other starch synthesising enzymes (such as soluble starch synthase or branching enzyme) may affect the degree, rate and/or type of starch production.

To increase the plant's capacity to produce starch, possible methods include: (1) Increasing the amount of the starch primer enzyme(s) in a recipient plant by the insertion of one or more additional copies of a normal donor gene encoding the primer(s).

The source of these additional copies may be the recipient line itself as the technique would simply increase the amount of enzyme available in the grain rather than change of the properties of the enzyme(s). The gene promoters and other regulatory sequences may also be modified using known techniques to achieve increased amounts of the enzyme in the recipient plant: increased gene expression may be elicited by introducing multiple copies of enhancer sequences into the 5'-untranscribed region of the donor gene. (2) Insertion of one or more donor genes encoding a starch primer enzyme with enhanced activity.

Achieving this requires the identification of a source of the enhanced activity enzyme gene. Such a source need not be a sexually compatible species as donor genes may be introduced using genetic manipulation techniques. Genetic material may thus be obtained from any source having a primer enzyme with the desired activity characteristics. Exogenous genetic material may alternatively be engineered to give the desired characteristics to the enzymes. Techniques are known in so-called "protein engineering" which can alter the characteristics of an enzyme.

The invention further provides a plant having at least one donor gene encoding a starch primer stably incorporated into the plant's genome such that the plant's ability to synthesise starch is altered. The starch primer may be amylogenin and/or glycogenin.

Presence of the donor gene(s) encoding a starch primer alters the natural process of starch production in the plant. It may alter the rate of starch production, the starch yield, the structure of the starch produced and so on. For example, such transgenic plants may have an increased capacity to produce starch and so are capable of producing higher yields and/or are capable of producing starch at a faster rate. The alteration may be an indirect effect: for example, presence of the donor gene(s) may change the natural expression ratios of the enzymes in the starch biosynthetic pathway leading to altered starch synthesis, including a different amount of starch or a different branching pattern in the starch.

The invention also provides DNA encoding amylogenin and/or glycogenin. Suitable cDNA or gDNA clones may be isolated from cDNA or gDNA libraries using standard cloning techniques. These cDNA or gDNA clones may then be used as probes^'for equivalent genes in other species, particularly starch-producing organisms. For example,_the enzyme amylogenin may be isolated by subjecting a crude plant extract containing the enzyme (eg maize endosperm) to successive purification by gel chromatography. Antibodies to the substantially pure amylogenin protein may be produced. Such antibodies may be used to screen cDNA libraries for the identification of amylogenin clones.

A cDNA clone encoding amylogenin was deposited at the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland 20852, USA under the terms of the Budapest Treaty on 19 August 1993 under the accession number ATCC 69389.

The amylogenin and/or glycogenin DNA clones may be used to alter starch production and starch quality in plants by the method described above. DNA encoding amylogenin and/or glycogenin may be transferred into plants by sexual crossing or by transformation. For example, a tranεgenic maize plant containing an amylogenin DNA clone may have increased starch yield. In another example, a transgenic maize plant containing anti-sense DNA clones may have decreased starch yield but increased sugar content.

The DNA clones may also be used to express the enzyme amylogenin and/or glycogenin in a biological organism such as E coli.

The present invention will now be described, by way of illustration, by the following examples and with reference to the accompanying drawings of which:

Figure 1 shows the biochemical pathway of starch biosynthesis in leaves.

Figure 2 shows the biochemical pathway of starch biosynthesis in storage tissues.

Figure 3 shows the new biochemical pathway of starch biosynthesis in storage tissues.

Figure 4 shows the proposed mechanism of glycogen synthesis.

Figure 5 shows a model for the mechanism of glycogen biogenesis in muscle.

Figure 6 shows the separation of sweet corn amylogenin (SGP) on DEAE-Sepharose.

Figure 7 shows the fractionation of sweet corn amylogenin (SGP) on Sepharose C1-6B.

Figure 8 shows the separation of prelabelled amylogenin (SGP) on Sepharose C1-6B.

Figure 9 shows the refractionation of the combined amylogenin (SGP)-containing fractions. „ PCr/GBM/01821

11

1. BACKGROUND TO THE INVENTION

1.1 A model for the biogenesis of glycogen in animal muscle

The proposed mechanism of glycogen synthesis, wherein glucose residues from UDPglucose are fisrt added to the protein primer by glycogen initiator synthase to form oligosaccharide chains that will serve as primers for glycogen synthase and branching enzyme (Kriεman and Barengo, 1975), is illustrated in Figure 4.

Present knowledge now allows the construction of a model for the mechanism of glycogen biogenesis in muscle which is illustrated schematically in Figure 5. Following the biosynthesis of the priming protein, namely glycogenin, the first step in the biogenesiε of glycogen involveε the covalent attachment of glucose to Tyr 194 in glycogenin by a protein tyrosine glucosyltranεferase that has yet to be identified. The glucosyl-glycogenin so formed is autocatalytic and extends the covalently attached glucose into a maltosaccharide chain by adding another seven glucose units to itself. Maltooctaosyl-glycogenin then primeε the εynthesis of proglycogen (Mr 400kDa) by proglycogen εynthaεe and branching enzyme. Macromolecular glycogen (Mr 10,000kDa) iε synthesized from proglycogen by the classical glycogen synthase with the assiεtance of a branching enzyme which may or may not be the εame as that utilized in the synthesis of proglycogen.

The mechanism of the initial glucoεylation of Tyr 194 in glycogenin remains to be elucidated. While a distinct protein tyrosine glucoεyltranεferase enzyme is envisioned in catalyzing this critical first step, it iε poεsible that glycogenin itself fulfills this role and catalyzeε the glucoεylation of Tyr 194 before proceeding to elongate the initial glucose unit into a altoεaccharide chain. While the nascent glucose may be extended into a chain of eight glucose residueε, proglycogen synthase may act before the chain has reached its full length. In other words, the maltooctaose unit that is constructed on glycogenin after undergoing complete self-glucosylation may not be a prerequiεite for the priming of proglycogen εyntheεis and chains of smaller lengths also may serve as the glucan primer.

The identity of proglycogen synthaεe alεo remains to be determined. Whether it is a distinct enzyme or a form of glycogen synthase with a specificity for proglycogen has to be elucidated. If the latter iε the caεe, then the manner in which it differε from claεsical glycogen synthase will be interesting to examine and may provide additional clues to understanding the intricate regulation of glycogen metabolism.

1.2 A universal mechanism for the biogenesis of polysaccharides

The protein primer model for the biogenesiε of mammalian glycogen may extend to εtarch εyntheεis and perhaps to polysaccharide syntheεiε in general. Simple tests uεing an antibody to glycogenin have indicated that glycogenin-like material may be present in vertebrates, invertebrates, plants, yeast, fungi and bacteria containing glycogen and εtarch (Gieowar-Singh et al, 1988, FASEB J, 2(1518) :A557) . The preεence of glycogenin-like proteinε in a wide variety of lifeformε εuggeεt that εimilar protein-priming mechaniεmε for polyεaccharide synthesiε may be occurring in theεe species.

The priming mechaniεm for εtarch biosynthesis had hitherto not been elucidated and the nature of the priming molecule ij vivo remained unknown. However, the present evidence points to a protein primer.

Similar to Krisman's first observation of protein-bound glycogen syntheεiε with a high εpeed pellet from a rat liver homogenate, the synthesiε of alpha, 1-4 glucan covalently linked to protein was achieved with a potato tuber particulate preparation. A two step mechanism for the formation of the protein-bound alpha, 1-4 glucan in the potato tuber waε subsequently proposed. The first εtep involved the transfer of a glucosyl residue specifically from UDPglucose to an endogenous protein thereby forming a peptidyl- glucose bond. The enzyme that catalyzed the glucosylation waε referred to as UDPglucose : protein transglucoεylaεe 1 and it functioned at micromolar concentrationε of UDPglucoεe. In the second step, the glucosylated protein then serves aε a primer for the formation of protein-bound alpha, 1-4 glucoεidic chainε by accepting glucoεe units tranεferred from UDPglucose, ADPglucose or glucoεe 1-phoεphate. It was proposed that the chain elongation, which occured in the presence of millimolar concentrations of the glucose donor, was catalyzed by the endogenous starch synthase or phosphorylase in the preparation.

The protein undergoing glucosylation in the first step had a molecular weight of 38kDa and the added glucose was resistant to hydrolysis by alpha- and beta-amylase. Glucoa ylase, however, was capable of hydrolysing the protein-bound glucose which was εubεequently identified by paper chromatography. Furthermore, the added glucose underwent beta-elimination upon alkaline treatment in the presence of sodium borohydride and was thereby converted to sorbitol aε determined by paper electrophoreεis. The conversion of the added glucoεe to εorbitol implied that the glucosyl moiety was transfered directly to an aminoacyl reεidue on the acceptor protein and furthermore that the glucan εyntheεized during the firεt step consisted of a single glucose unit, in addition, the suεceptibility to mild alkaline hydrolyεiε suggested that the glucose was linked to a serine or threonine residue by an O-glycosidic bond. This was subsequently confirmed by reduction of the beta-elimination products with tritiated sodium borohydride. Tritiated alanine and tritiated 2- aminobutyric acid, expected by H-sodium borohydride reduction of the 2-amino acrylic acid and 2-amino crotonic acid generated by beta- elimination of carbohydrateε attached to εerine and threonine residues, were both obtained. Therefore, both serine and threonine reεidueε εeemed to be involved in the linkage to glucoεe.

The ability of the glucoεylated protein formed during the first step to serve as a primer for the synthesiε of protein-bound alpha, 1-4 glucan waε shown by conversion of the former into higher molecular weight material as the attached glucose underwent chain elongation. This waε achieved by firεt allowing the acceptor protein to undergo glucoεylation with micromolar concentrationε of UDP-[ 14Clglucose during the first step. Unlabelled

UDPglucose, ADPglucoεe or glucoεe 1-phosphate was then added to the incubation mixture in millimolar concentrations as substrates for the endogenous chain-lengthening glucosyltransferaεes in the potato tuber preparation.

The increase in the molecular weight of the radioglucosylated protein formed during the firεt step, was observed by analyεis of the final products on urea-SDS-PAGE (sodium dodecyl sulphate- polyacrylamide gel electrophoresiε) and fluorography. In addition to the 38kDa radiolabelled band representing the product formed during the firεt step a labelled band at 50kDa was observed on a 12.5% acrylamide gel for the incubation in which unlabelled UDPglucose was added during the second step. When a 10% acrylamide gel was used, the mobility of the band decreased and now migrated to the 70kDa position. Similar gel patterns were obtained with the incubations in which unlabelled ADPglucose and glucoεe 1-phoεphate were uεed as substrates for chain elongation.

Fractionation of the solubilized potato tuber preparation on DEAE-celluloεe separated the endogenouε acceptor protein, which copurified with the enzyme catalyzing its glucosylation during the first step, from the glucosyltranεferases that participated in the chain-elongation procesε in the εecond εtep. Furthermore, phoεphorylase emerged later than the acceptor protein as two overlapping peaks, each of which also contained starch synthaεe activity. The glucoεyltranεferase activities that catalyzed the chain extension of the 38kDa glucosylated acceptor protein had copurified with the first phosphorylase peak which was used to reconstitute the εynthesis of protein-bound alpha, 1-4 glucan. Chain extension was demonstrated by the increase in the molecular weight of the 38kDa radioglucosylated acceptor protein on SDS-PAGE and fluorography.

The chain lengthening enzymes contained in the first phosphorylaεe peak utilized UDPglucoεe, ADPglucoεe and glucoεe 1-phoεphate aε glucoεyl donors. Major radiolabelled bands at 43kDa and 50kDa were observed on 12.5% acrylamide gelε when UDPglucoεe and ADPglucose were used as chain- lengthening substrates. When glucose 1-phosphate was used, only the 50kDa labelled band was obtained with the 12.5% gel, which migrated as a 70kDa species on a 10% gel. The 70kDa species, following exciεion and elution from the gel, waε partially converted to a εpecieε of 38kDa upon treatment with beta amylaεe. Thiε εupported the contention that the 70kDa εpecies, or at least part of it, originated from the 38kDa glucosylated acceptor protein and was synthesized as the latter underwent chain elongation by the addition of alpha, 1-4 glucoεe moieties.

The enzymic activity that catalyzeε the initial glucoεylation of the acceptor protein waε recently purified. During purification, the glucoεyltransferase and the acceptor protein copurified to homogeneity demonstrating that both were one and the εame protein. In other wordε, the glucosyltransferase iε an autocatalytic enzyme that undergoeε self-glucosylation to form the protein primer required for alpha, 1-4 glucan synthesis in the potato tuber.

A similar autocatalytic glycosyltranεferase with a molecular weight of 40kDa has been purified from Daucuε carota L (carrot) cells. The enzyme functions in micromolar concentrationε of UDPglucose from which it tranfers glucosyl moieties onto itself. Treatment of the radioglucosylated product with weak alkali (0.1N NaOH, 24°C, 24hr) and strong alkali (IN NaOH/lM NaBH₄, 100°C, lOhr) partially εeparated the carbohydrate from the protein. Methylation of the oligoεaccharides so released produced 2,3,4,6 tetra-O-methyl-glucitol- acetate as the only labelled carbohydrate indicating that autoglucosylation added only one glucoεe residue to the protein. In addition to UDPglucose, the enzyme also utilized UDPgalactoεe as a subεtrate undergoing autogalactosylation. While thiε carrot enzyme iε εimilar to the autocatalytic protein that appears to prime alpha, 1-4 glucan syntheεiε in the potato tuber it has yet to be implicated in polysaccharide biogenesis.

The biosynthesis of beta-glucans alεo appearε to proceed through a glycoprotein intermediate. The formation of protein-linked oligoεaccharideε that εeemed to be involved in the εynthesis of beta- glucans was detected in bacteria, fungi, algae and higher plants. However, the evidence that a glycoprotein primed beta-glucan synthesis in these studies was circumstantial and a covalent linkage between a beta-glucan chain and a protein was never demonstrated. In the case of Chlorophyta Prototheca Zopfii (algae) the evidence is more convincing that beta- glucan syntheεiε iε initiated on a protein primer. An intermediate in the synthesis of beta-1,4/1,3- glucan in this alga displayed properties that were consistent with a glycoprotein nature. In addition to being insoluble in trichloroacetic acid, this intermediate was not dialyzable, strongly absorbed at 280nm and gave a poεitive reaction with phenol/sulphuric acid, Lowry and Buiret reagents. It contained glucose incorporated from UDPglucose and waε εhown to be a precurεor of alkali-inεoluble polysaccharides. This was demonstrated by the synthesiε of radiolabelled alkali-insoluble polysaccharides when the radioglucosylated intermediate was incubated with unlabelled UDPglucose and the endogenous enzymes in a fresh membrane preparation of the alga. The polysaccharideε so synthesized however seemed to be free of protein indicating that at some point in the elongation process the priming protein was discarded.

Thiε intermediate in beta-glucan εyntheεiε was subsequently iεolated and characterized. Membrane preparations from Prototheca Zopfii cells pulsed with [ 14C] proline were incubated with

UDP[ H]glucoεe. The intermediate εo formed was double-labelled containing both 3H-and 14C- radioactivity. It was susceptible to partial hydrolysiε by cellulaεe and beta-glucosidase and underwent partial degradation with protease. On gel filtration chromatography, it behaved as a 28000-

30000 Da molecule and migrated as a species of the same molecular weight on SDS-polyacrylamide gel electrophoresiε. These characteriεticε firmly established that the intermediate waε a glycoprotein with beta-glycoεidic bondε in the carbohydrate moiety.

The double-labelled glycoprotein intermediate was resistant to beta-elimination (0.15M NaOH, IM NaBH., 100°C, 5hr) indicating that the carbohydrate moiety was not. linked by an O-glycosidic bond to serine/threonine or by an N-glycosidic bond to asparagine. Treatment with saturated barium hydroxide (0.2M Ba(OH)₂, 100°C, 6hr) however, liberated a double-labelled aminoacyl- oligoεaccharide from the glycoprotein. Thiε waε conεiεtent with an O-glycoεidic linkage to hydroxyproline or hydroxylyεine whereby cleavage of the peptide bondε to the glycoεylated amino acid occurε. The association of 14C-label with the liberated oligosaccharide moiety suggested that the carbohydrate was linked to hydroxyproline derived from the radioactive proline that was fed to the cells.

Cellulaεe digestion of the double-labelled product obtained by barium hydroxide treatment produced a smaller double-labelled aminoacyl oligoεaccharide which upon acid hydrolysis and subsequent reduction with tritiated sodium borohydride yielded [ 14C] hydroxyproline and [3H] sorbitol. The formation of sorbitol as the only alditol following reduction indicated that glucoεe waε preεent at the reducing end of the oligoεaccharide chain. Thiε εuggeεted that the carbohydrate-protein linkage in the glycoprotein intermediate waε an O-glycosidic one between hydroxyproline and glucose. 20

2. PURIFICATION OF AMYLOGENIN PROTEIN FROM MAIZE ENDOSPERM

The εource of the amylogenin enzyme (or εelf- glucosylating protein, SGP) was a US inbred line (B73) of Zea mays.

2.1 Preparation of corn extracts

The kernels were cut from the cobs directly into a food grinder which functioned by continuously extruding the ground material. The ground kernels were collected and expressed through cheese cloth into a beaker surrounded by ice and containing a concentrated extracting solution which, after dilution by the corn juice, gave a final concentration of 50mM Tris-HCL pH 7.4, 5mM EDTA (ethylene diamine tetra acetate), O.lmM EGTA, 14mM mercaptoethanol, O.lmM PMSF

(phenylmethaneεulphonyl fluoride), ImM benzamidine and O.lmg/liter of both leupeptin and pepstatin. Following this, the mixture was centrifuged at 3000 x g for 15 minutes at 4°C and the supernatant retained as the cell-free corn extract.

The amount of concentrated extracting solution needed for each preparation was estimated at the outset. The exact volume of the extract could not be predetermined because of variability in the size of the cobε. Therefore, an average volume of 70ml of extract/cob was used for eεtimating the amount of concentrated extracting solution required. After the final volume of the mixture was measured, the quantity of extracting solution was adjusted to give the desired concentration. The use of a concentrated extracting solution facilitated in keeping the volume of the extract to a minimum, which from a typical preparation with 15 cobs of corn, ranged between 1100-1300 milliliters.

2.2 Detection and quantitation of Self- Glucosylating Protein (SGP)

SGP was aεεayed by incubating a sample of the enzyme with 2uM UDP-t 14C]glucose, 5mM MnCl₂ and

50mM Tris-HCL buffer, pH 7.4 in a lOOul reaction mixture at room temperature for 30 minutes (the reaction mixture was made up by adding, in the following order, lOul each of 10X εolutionε of

Triε-HCL pH 7.4 buffer, MnCl, and UDPglucose to a

70ul solution of the enzyme). The reaction was stopped by the addition of 1ml of 10% TCA

(trichloroacetic acid). After standing for 10 minutes the sample was paεεed through a ^~ nitrocelluloεe filter, type HA 0.45uM, washed with about 3ml 10% TCA and approximately 10ml _of water and the filter dried under an infrared lamp.

Following this, the filter was placed in 4ml of scintillation liquid (Ecolume) and measured for 14C radioactivity in a Searle Delta 300 6890 liquid scintillation counter.

The radioactivity measured on the filter represents 14C-glucose that was incorporated into

TCA-precipitable material (assuming no conversion to the glucose molecule into any other compound).

In the case of SGP, which functionε by transfering a single glucose residue from UDPglucose onto itself, the assay provides a means of quantitating

'naked' SGP. Endogenously glucoεylated SGP would not be detectable and could not be quantitated by thiε method. In corn, as in many other plant species possessing SGP-like proteins, the assay also detects, in a crude extract, the incorporation of radioactivity into a TCA-precipitable, lipid-like substance. This lipid-like substance is soluble in methanol, methanol/chloroform and ethanol. In a crude extract of corn, this lipid-like substance representε about 80-90% of the TCA-precipitable, radiolabelled material measured by the asεay. Aε SGP is purified away from this subεtance, however, the assay provides an accurate meaεurement of SGP activity.

In a modification of the procedure and in an attempt to meaεure SGP activity in crude extracts, lml of 95% ethanol waε uεed in place of 1ml 10% TCA to precipitate protein. After vortexing well to disεolve the lipid-like εubεtance and standing on ice for 30 min to allow protein precipitation, the εample was centrifuged at 7000 rpm for 15 min with a bench top centrifuge. The protein pellet waε resuspended in 10% TCA and processed as in the original assay. This provided a way of separating the two activities and allowing SGP activity to be measured in crude extractε.

In addition to scintillation counting, SGP was detected by SDS-PAGE, Western blot analysis and radioautography/fluorography. Theεe procedures, however, were used only for qualitative analyses.

SDS-PAGE was performed with a Mini-Protean 11 electrophoresiε εyεtem (BioRad) on 10% acrylamide gelε according to Laemmli [108]. Samples were electrophoresed in electrode buffer, pH 8.3 (25mM Triε, 192mM glycine, 0.001% SDS) with a 25mA current for approximately 1 hour. Protein bandε were viεualized by εtaining gels with either Coomaεsie blue R-250 or silver (BioRad silver stain kit). Following this, the εtained gelε were εoaked in Amplify for 15 min [109] and dried onto Whatman 3MM paper in a εlab dryer (BioRad) at 60°C.

Western blotting waε performed in a Mini tranε-blot electrophorectic tranεfer cell (BioRad). Following SDS-PAGE, proteinε from unstained gels were transfered onto nitrocelluloεe paper in 25mM Tris, 192mM glycine, 20% methanol (v/v) pH 8.3 buffer. The transfer was carried out by applying a 150mA current for 1 hr. After transfer, the nitrocellulose blot was blocked in 5% Carnation lowfat milk dissolved in Tris-buffered saline (20mM Tris-HCL, 0.5M NaCl, pH 7.4). After blocking, the blot was washed 3 times with Tris-buffered εaline (TBS) and then incubated overnight in TBS containing 1% gelatin with goat anti-glycogenin IgG (1:1000 dilution of original stock). Next, the blot was washed 3 times with TBS and then incubated with alkaline phosphatase-conjugated rabbit anti-goat IgG. The antigen-bound protein bands were detected by color development with solutions of BCIP (5- bromo-4-chloro-3indolyl phosphate) and NBT (p- nitroblue tetrazoliu chloride) aε inεtructed by the εupplier.

Radioautography waε performed on dried gelε, immunoblotε and thin layer εilica plateε containing radiolabelled materialε by expoεure to Kodak X-OMAT XAR-5 filmε at -70°C. (Fluorography was also performed on dried gels in the same manner). The films were developed with an automated Kodak X-OMAT film processor. 24

2.3 Purification of Self-Glucosylating Protein from Sweet Corn

The following steps were carried out, at 4°C, to obtain homogeneous SGP from εweet corn.

Step 1- Acid precipitation

After preparing a cell-free extract from 15 cobε of εweet corn, the pH waε adjuεted from 7.4 to 5.0 with IM ammonium acetate pH 5.0 buffer. Precipitation waε allowed to occur for 1 hr. The supernatant, which contained the protein glucosylating activity, was collected by centrifugation at 4000 x g for 15 min. Most of the lipid-glucosylating activity remained in the pellet.

Step 2- Fractionation by εalt precipitation Solid ammonium εulphate was added to the pH 5.0 supernatant to give a 50% saturated εolution. Protein was allowed to precipitate overnight and collected by centrifugation at 10,000 x g. The protein pellet was resuspended in 50mM Tris-HCL pH

7.4 buffer containing 2mM CHAPS (TC) and dialyzed against the same buffer. Following dialyεis, the solution was centrifuged to remove the considerable amount of undissolved material.

Step 3- Fractionation by ion exchange chromatography

The dialyzed solution waε applied to a DEAE- Sepharoεe column (6 x 15cm) that was equilibrated with TC. After pasεing equilibrating buffer through the column to remove unbound material, the protein was desorbed with an increasing concentration gradient of NaCl, from 0 to IM, formed from 500ml 25

of TC and 500ml of TC containing IM NaCl (Figure 6). Fractions of 11ml were collected and assayed for protein-glucosylating activity. The fractions posεeεsing the highest activity were pooled and concentrated to approximately 25ml by centrifugation using centriprep-10 concentrators (Amicon).

Step 4- Fractionation by gel filtration chromatography

The concentrated material was applied, in two equal portions during separate runs, to a Sepharose Cl 6B column (3 x 85cm) that waε equilibrated with TC and eluted with the εame buffer (Figure 7). The protein-glucoεylating activity was found in a single, but broad, peak that began at the void colume. The material that emerged at the front7 represented a single protein specieε on a εilver- εtained gel after SDS-PAGE and poεεeεεed aoth the glucoεylating activity and the glucosylation site. Theεe early fractionε containing homogenouε protein represented about 10% of the total glucosylating activity applied to the column. The later fractions contained more activity but were increasingly less pure with each additional fraction.

2.4 Purification of Prelabelled SGP from Sweet Corn

The purification of the self-glucoεylating protein from εweet corn waε accompliεhed by εucceεεive precipitation with acid and salt followed by column chromatography on DEAE-Sepharose and Sepharose C1-6B. 26

TABLE 1 Purification of sweet corn amylogenin (SGP)

A 3% yield (Table 1) repreεenting approximately 2 nanomoleε or 80ug of the protein was attained from the preparation with 15 cobs of corn. The homogenous SGP εo obtained, however, waε difficult to proceεε and required additional purification. It appeared that the protein was complexed to lipid. Additional purification to separate the protein away from the contaminating lipid would further reduce the already low yield of the enzyme and therefore, in order to obtain a sufficient quantity of the protein for amino acid and εequence analyses, it was deemed necessary to purify a larger amount of the SGP.

Two approaches were envisioned to obtain an adequate amount of SGP. One was to purify several extracts individually and combine the homogenouε protein from each preparation. The other was to purify, in addition to the homogenous protein at the front of Sepharose C1-6B, the impure SGP that elutes later. The first approach was attempted and approximately 4 nanomoles of electrophoretically homogenouε SGP waε obtained from two preparationε with 15 cobε of corn each. In a series of experiments to find an appropriate method to purify the protein from the lipid, most of the material waε loεt.

In the next attempt to isolate sufficient protein, it was decided to try the second approach by purifying the SGP that elutes later on Sepharose C1-6B together with the portion that emerges at the front as the εingle protein species. It was not necesεary to purify the protein in an active state since the objective was to obtain the amino acid composition and amino acid sequence information on the protein. Therefore, the strategy was to partially purify the SGP, then prelabel the enzyme by allowing it to undergo self-glucoεylation and finally isolate the prelabelled protein. The prelabelling of the SGP prevented losses from inactivation of the enzyme during subεequent purification and allowed procedures that involve denaturing conditions to be employed.

The following steps were performed to isolate the self-glucosylating protein from εweet corn in a prelabelled form and waε accomplished in two phaseε.

Phaεe 1 - Partial purification of the self- glucosylating protein and subsequent prelabelling by autoglucosylation:

Step 1 - Acid precipitation; Step 2 - Salt precipitatio;n Step 3 - Fractionation by ion exchange chromatography.

Steps 1-3 are exactly the first three stepε in the protocol for purifying active SGP (εection 2.3) and were carried out in the εame manner. Following the fractionation on DEAE-Sepharose and concentration of the SGP containing fractions, the enzyme was allowed to undergo self-glucosylation, 25% with UDP-[ 14C]glucose and the remaining 75% with unlabelled UDPglucoεe. The prelabelled protein waε then purified further.

Phaεe 2 - Purification of the prelabelled SGP: Step 4 - Fractionation by gel filtration chromatograph .

At the end of the prelabelling reaction, the incubation mixture waε applied to a Sepharose C1-6B column (3 x 85cm) equilibrated with 50mM Tris-HCL pH 7.4 buffer containing 0.02% triton and eluted with the same buffer (Figure 8). Fractions were collected in a volume of 5.5ml each. The prelabelled SGP was detected by measuring lOOul aliquotε from alternating fractionε for 14C- radioactivity. The prelabelled protein emerged in a broad peak similar to that observed for the native unglucosylated enzyme. A peak of 14C-radioactivity detected at the total volume of the column repreεented unreacted radiolabelled UDPglucoεe from the prelabelling reaction.

The purity of the prelabelled SGP waε examined by SDS-PAGE in which the protein bandε were viεualized by silver staining. Fifty microliter

(50ul) aliquots from every third fraction containing prelabelled SGP were analyzed in this manner. The fractions containing the prelabelled protein were then separated into two setε baεed on their protein content depicted by the silver εtained polyacrylamide gel. The earlier fractionε (43-51) containing purer SGP were pooled together and the later fractions (52-60) possessing SGP in a less pure state were combined into a εecond pool. The pools of prelabelled SGP were concentrated by centrifugation using centriprep-10 concentrators (Amicon) .

Step 5 - Fractionation by Reverse-Phaεe HPLC.

Following the εeparation on Sepharoεe C1-6B, the prelabelled SGP waε purified next by reverεe- phase HPLC on a Perkin ELmer HPLC syεtem uεing a Vydac C . column (8 x 40mm). The sample was fractionated on the column in several portionε. During each εeparation 450ul aliquotε containing 0.26mg of total protein were applied to the column. Thiε waε done to obtain the best resolution of the SGP containing peak. It waε observed, during the purification of prelabelled SGP from B73 commercial dent corn, that a larger amount of protein on this column resulted in a lower resolution. The protein was desorbed from the column with linear gradientε of acetonitrile generated by uεing 0.1% TFA (trifluoroacetic acid) in water (εolution A) and 0.09% TFA in water containing 84% acetonitrile (solution B) at a flow rate of lml/min. The first two separations were performed successfully with a linear gradient of 0-100% solution B applied over a period of 50 minutes followed by an additional 10 minuteε with 100% εolution B. In an effort to reduce the time for each run, further separations were carried out with a gradient composed of the following linear segments : 0-30 minutes, 0-70% solution B; 30-31 minutes, 70%-100% solution B; 31- 41 minutes, 100% solution B. The altered gradient worked with the same succesε. Protein waε detected by monitoring the abεorbance of the effluent at 30

215nm and 295nm simultaneously. Peaks containing protein were collected manually and the prelabelled SGP emerged from the column in a single peak.

The fractions containing prelabelled SGP from each run were pooled together. The combined fractions had a volume of 10.7ml and was concentrated to approximately half its volume on a εpeed-vac. The concentrated εample was then diluted to its original volume with water thereby reducing its acetonitrile concentration by approximately half. The sample waε then refractionated on the C, column in two portionε (Figure 9). In the firεt run, 5.7ml of the sample was applied to the column by injecting consecutively six aliquotε of 950ul each. The remainder of the εample waε fractionated in a εecond run by injecting five 950ul aliquotε onto the column. Protein waε eluted in the εame manner aε before and the fractionε containing prelabelled SGP from both runε were combined giving a volume of 3.3ml. Protein determination on the combined fractions by amino acid analysis revealed that 149.2ug of protein was present. An aliquot containing 3ug of protein was analyzed on SDS-PAGE for purity. A single homogenouε protein band representing prelabelled SGP was observed.

2.5 Reduction and Alkylation of purified SGP

The sample containing 149.2ug of protein in a volume of 3.3ml, was prepared for reductive carboxymethylation by adding to it 400ul of 5M guanidium-HCL, 0.5M Triε-HCL pH 8.6 buffer. After mixing well the solution was concentrated to 0.4ml on a speed-vac. Reduction and carboxymethylation were performed on the protein in the following manner: 31

Step 1 - Reduction with DTT (20 umol/mg protein) .

Five microliters (5ul) of 1.0M DTT was added to the sample and the reaction tube then flushed with argon. The reaction was allowed to proceed by incubation at 50°C for 2 hr.

Step 2 - Carboxymethylation with iodoacetic acid (60 umol/mg protein).

Next, 15ul of IM iodoacetic acid was added and after flushing the tube with argon the sample was incubated at room temperature for 30 min in the dark.

Step 3 - Reduction with DTT (30 umol/mg protein) .

Seven and one half microliters (7.5ul) of 1.0M DTT was added to the sample and then incubated at 50°C for 2 hrs after fluεhing the tube with argon.

Step 4 - Carboxymethylation with iodoacetic acid (60 umol/mg protein).

Next, 15ul of IM iodoacetic acid was added and the reaction allowed to proceed at room temperature for 30 min in the dark under argon.

The reaction was terminated by adding 17.5ul of 1.0M DTT to the incubation mixture. Following thiε, the protein waε dialyzed against 50mM Tris HCL pH 7.6 buffer containing ImM CaCl₂ on a Pierce Microdialysiε System 500. The progreεε of dialyεis was monitored by following the decreaεe in conductivity of the εample. This was achieved by taking 2.5ul aliquotε from the sample at regular intervals and, after diluting in 10ml of water, measuring the conductivity. Dialysis was stopped after a 100 fold decrease in conductivity was obtained over 2.5 hr. After dialysis, the volume of the sample had increased to 550ul and its pH was 7.8. 3. AMINO ACID COMPOSITION AND SEQUENCING OF AMYLOGENIN (SGP)

3.1 Amino acid composition of sweet corn SGP

An accurate determination of the amino acid composition of sweet corn SGP waε made by calculating the expected yield of each amino acid after 100 hr of hydrolysis. This was done by performing amino acid hydrolyseε on εamples of the protein at 110°C for 24 hr, 48 hr and 72 hr respectively. From a εtatistical analysiε of the yield for each amino acid at each time interval an extrapolation waε made for the expected yield at 100 hr and the amino acid composition calculated therefrom. A compariεon between the amino acid compoεitionε of εweet corn SGP and glycogenin (Table 2) indicates a general similarity between the two proteins. Other than serine, hiεtidine, threonine and iεoleucine, the proportion of each amino acid in the two proteinε are comparatively εimilar. In particular, the ratio of glycine, arginine, alanine, and valine in each protein differ by leεs than 10% while that of leucine, phenylalanine and lysine are less than 13% different.

TABLE 2 Comparison between the amino acid composition of glycogenin and amylogenin

3.2 Preparation and isolation of tryptic peptides from purified SGP

The sample of purified prelabelled SGP contained 149.2ug of protein. On previous occasionε, it waε found that the partially pure protein waε reεiεtant to trypsinolyεiε with 70% proteolyεis occurring over a 48 hr incubation period. To make the purified protein more susceptible to proteolysiε, reduction and carboxymethylation waε performed. The protein waε then digested with trypsin in an enzyme/subεtrate ratio of 1:25 for 15 hr. The tryptic digeεt waε fractionated on a C-,₈ column by reverεe-phase HPLC and peptide peaks collected separately. An examination of the fractions for 14C-radioactivity revealed the presence of 3 radiolabelled peaks. 34

Nine of the peptide peaks (labelled P1-P9) including the 3 radioactive peaks were further purified by fractionation on a smaller C.g column at pH 6.0. Upon refractionation of peaks P1-P9 the following number of tryptic peptides were obtained from each.

Peak Number No of tryptic peptide peaks obtained

P 1 (fraction 29+30) 4

P 2 (fraction 33+34) 3

P 3 (fraction 37+38) 5*

P 4 (fraction 42) 2

P 5 (fraction 54) 4

P 6 (fraction 55-57) 5*

P 7 (fraction 59) 3

P 8 (fraction 60+61) 1*

P 9 (fraction 62) 1 * 14C-radioactivity containing peaks

The fractionation of P3 and P6 each gave rise to one radiolabelled tryptic peptide peak. The separation of P8, which had very little radioactivity (12000cpm in total), produced only a εingle peptide peak which was not radiolabelled. No radioactivity was detected in any of the fractions. It appears that the recovery of radioactivity was too small to be detected.

3.3 Amino Acid Sequence of Tryptic Peptides

An attempt to sequence the N-terminus of sweet corn SGP was made by applying a sample of the purified protein for amino acid sequencing. It waε found however that the N-terminus was blocked. Nine of the purified tryptic peptides were chosen for amino acid εequence analyεeε (labelled T1-T9). The amino acid εequences of these peptides are summarized below.

T l (P I) - Y V N A V M T I P K

T 2 (P 3) - E G A N F V X G Y P F S L R *

T 3 (P 4) - Y X X M W A G W T V K

T 4 (P 4) - E G A H T A V S H G L W L N I P D Y D A

P T Q L V K P K T 5 (P 5) - L G D A M V T W I E A W D E L N P S T P

A A A D G K T 6 (P 6) - L G D A M V T D I E A A D E L N P A G P

X X X X K T 7 (P 6) - N L L S P S T P F F F N T L Y D P Y R E

G A N F V X G Y P F S L R * 1 8 (P 7) - G I F W Q E D I I P F F Q N V T I P K T 9 (P 9) - N L D F L E M W R P F F Q P Y H L I I V

Q D G D P T K

* radioglucosylated tryptic peptides

The radiolabelled peptide T2 is a proteolytic fragment^' of the larger radiolabelled peptide T6. It is likely therefore that the radiolabelled peptide in P8 waε εimply a larger peptide comprising T2 and T6 formed from incomplete proteolysis. The emergence of the peptide later than T2 and T6 εuggest that it has a larger size. The small quantity of radioactivity (12000cpm) further suggest that the peptide may have arisen from incomplete hydrolysis and more εo that trypsinolysis was performed for only 15 hr. In the amino acid sequence of T2 and T6 no phenylthiohydantoin amino acid was detected at position 7 and 26 reεpectively and moεt of the applied radioactivity remained bound to the glaεε fibre diεc, indicating that the [ 14C]-glucoεylated amino acid was not extracted by the sequencer solventε. Thiε implied that the radiolabelled glucose was attached to the amino acid at this position.

3.4 Determination of the amino acid involved in the protein-carbohydrate linkage

In order to characterize the protein- carbohydrate linkage in the sweet corn SGP, it was necesεary to determine the amino acid to which the glucoεe waε attached. Thiε waε achieved by performing amino acid analyεiε and maεs spectrometry on the radioglucosylated peptide, T2.

Amino acid analysis of T2 revealed that the peptide had the following empirical composition

^D1.0 ^E1.0 ^S1.2 ^G2.8 ^R2.0 ^A1.2 ^Pnd ^Yl.l ^Vl.0 1.5 (nd « not determined). The composition of the peptide from sequence analysiε waε D- E. S, G₂ R.

A- P- Y. V. L. X.. When the two compositions were compared, the most εtriking difference waε in the amount of arginine. Sequence analyεiε had found only one arginine residue but amino acid analysis indicated that two molecular proportions were preεent in the peptide. By inference, therefore, the unknown amino acid, X, at poεition 7 in the peptide must be arginine.

The peptide T2 was submitted to ion-spray mass spectrometry in order to determine itε preciεe molecular mass. The ion-spray maεs spectrum of the peptide yielded only one prominant signal at m/z (mass/charge ratio) 888.9 corresponding to a molecular mass of 1777.8 Da for a doubly protonated ion. The molecular masε of peptide T2 calculated from itε amino acid composition by assuming an arginine residue in poεition 7 iε 1776.97. By 37

inference, therefore the εignal muεt represent peptide T2 and in combination with the amino acid analysis data confirms that the amino acid to which the glucose is attached at position 7 is arginine. It should be noted that minor signalε were obεerved in the ion spray mass spectrum. These may repreεent contaminants in the εample aε well as sodium and potassium salts of the peptide.

Tandem ion-εpray maεs εpectrometry waε alεo performed on the peptide. A daughter ion εpectrum obtained by colliεional activation in which the doubly protonated parent ion was bombarded with argon, showed two signals, a prominant one at m/z 889.0 and a minor one at m/z 807.5. The major εignal representε the doubly protonated parent ion correεponding to the glucoεylated peptide T2. The minor signal is due to a doubly charged daughter ion generated by fragmentation of the parent ion. The mass difference between the parent ion and the daughter ion is 163 Da which corresponds to a glucosyl fragment. The presence of only one daughter ion with a loεε of a 163 Da fragment from the parent ion εtrongly εuggests that only one glucose residue was attached to the peptide. This implies, therefore, that during the self- glucoεylation reaction by the εweet corn SGP only one glucoεe reεidue iε tranεferred onto the protein directly to an arginine reεidue.

3.5 Trypsinolysis of purified SGP

Following carboxymethylation of the purified SGP the protein was digested with trypsin to generate peptides for εubεequent εequencing. For this purpose, TPCK-treated trypsin modified by reductive alkylation (Promega) was used. The modified trypsin is extremely resiεtant to autolytic digestion. The lyophillized TPCK- treated, modified trypsin was dissolved first in 50mM acetic acid to a concentration of lug/ul. Fifteen microliters (15ul) of this εolution was added to the sample containing 144ug of purified SGP in a volume of 530ul giving a protease : protein ratio of 1 : 25 (w/w) . The digest was incubated at 37^βC for 15 hr.

3.6 Purification of Tryptic Peptides from SGP by Reverse-Phase HPLC

The peptides generated by trypsinolysis on purified SGP were isolated by reverse-phase HPLC on a Perkin Elmer HPLC syεtem using a Vydac C,« column (100 x 4mm). At the end of proteolyεis the whoie digest was applied to the column. Peptides were eluted with a linear gradient between 0.1% TFA in water (εolution A) and 0.09% TFA in water containing 84% acetonitrile (solution B) applied over a period of 100 min at a flow rate of 0.4ml/min. Peptides were detected by monitoring the absorbance of the effluent at 215nm and 295nm εimultaneouεly and collected manually.

Individual peakε were then choεen and purified further by refractionation at pH 6.0 on a smaller C._Q column (50 x 2mm) with the same HPLC system. Each peak was concentrated to half its volume on a εpeed-vac and then diluted to itε original volume with water. Thiε served to reduce the acetonitrile concentration in each sample by approximately half. The samples were refractionated individually by applying each to the C,₈ column. Peptides were eluted from the column with a linear gradient of acetonitrile from 0-84% generated from lOmM ammonium acetate pH 6.0 and 10mM ammonium acetate pH 6.0 containing 84% acetonitrile. The gradient was applied over a period of 60 min with a flow rate of 0.15ml/min. Peptides were detected as before and collected manually. After this round of purification chosen peptides were then subjected to amino acid sequence analyseε.

3.7 Amino Acid and Sequence Analyses

Amino acid analyses was performed on both the intact prelabelled SGP and tryptic peptideε generated therefrom. This was done by the PTC derivitization procedure on an automated amino acid analyzer. Samples were dried on a speed-vac and then hydrolyzed in 6N HCL at 150°C for 70 min. Following hydrolysiε the εamples were dried again on the speed-vac and then 2 nanomoles of norleucine added to each as an internal standard. The εamples were then applied to the amino acid analyzer which quantitates the amino acids by reverse-phaεe HPLC against amino acid standards. The HPLC solutions were 50mM εodium acetate pH 5.6 (buffer A) and 30mM εodium acetate pH 4.6 containing 70% acetonitrile (buffer B).

In order to obtain the moεt accurate amino acid composition of the intact SGP, samples were hydrolyzed in 6N HCL at 110^βC for 24 hr, 48 hr and 72 hr reεpectively. The amino acid compoεition waε then determined by extrapolation for 100 hr of hydrolysis.

Amino acid sequence analyεeε were performed by the Edman degradation procedure on an Applied Bioεyεtemε 470A gas-phase εequencer equipped with an on line model 120 phenylthiohydantoin analyzer using the 03R phenylthiohydantoin standard program.

3.8 Mass Spectrometry and Tandem Mass Spectrometry of Radioglucosylated Tryptic Peptide from Prelabelled SGP

Ion-spray mass spectrometry and ion-spray tandem mass spectrometry were performed on the radioglucosylated tryptic peptide isolated from purified prelabelled SGP. The ion-spray mass spectra and ion-spray tandem mass spectra were recorded on a Sciex (Toronto) API 111 triple- quadrupole maεs spectrometer with 2400-Da masε range equipped with an ion spray source. The masε spectrometer was operated under unit-maεε reεolution conditions for all determinations. The ion spray voltage was 5kV.

The samples were dissolved in 1% formic acid/methanol prepared in a 1 : 1 (vol/vol) ratio and introduced into the ion spray source at a constant flow rate of 5ul/min with a microliter syringe by using a medical infusion pump (Harvard

Apparatus). Argon at a target gas thickness of approximately 1.2 x 10 14 atoms per cm2 was used as collision gas for tandem mass spectrometry. The collision energies ranged from 40 to 60 eV.

4. NATURE OF THE CARBOHYDRATE COMPONENT OF SWEET CORN SGP

The self-glucosylating protein in rabbit muscle, as isolated, containε at least one glucose unit. Self-glucosylation consiεts in lengthening this chain to eight units by the addition of 1,4- linked alpha-glucose residues. The added glucose 41

reεidues are therefore removable by glycogenolytic enzymes. In an attempt to characterize the carbohydrate component of the sweet corn SGP the following studies were performed.

4.1 Acid hydrolysis of prelabelled SGP

The first step was to determine whether the radioactivity incorporated into the SGP waε in fact radiolabelled glucoεe and not a byproduct of a conversion reaction. To examine this, SGP Was allowed to undergo radiolabelling and then acid hydrolysis performed on the product. An autoradiograph of the hydrolysate after thin layer chromatography included a radiolabelled glucoεe standard. Under mild hydrolysiε, after boiling in 2M trifluoroacetic acid for 30 minuteε, a faint radiolabelled band iε seen at the position of glucoεe. In addition, three εlower moving radiolabelled bands of stronger intensity were observed. Under stronger conditions, by boiling for 2 hrs, the εame radiolabelled bandε are obεerved. Now, however, the radiolabelled band at the poεition of glucose was much more intense and the two other bandε had a εlightly leεε intenεity than before. The appearance of the radiolabelled band at the glucoεe marker after acid hydrolysis indicates that the radioactivity incorporated into SGP is radiolabelled glucose. Furthermore, the incorporated glucose was not bound by an acid labile bond since strong acid hydrolysiε was necessary to release a εignificant amount of the carbohydrate from the protein. It is unclear what the slower moving radiolabelled bands represent, but perhaps these may be radiolabelled peptide fragments from the protein. The marginally lighter intensity observed for εtrong acid hydrolysis 42

supports this hypothesis.

4.2 Digestion of prelabelled SGP with glycogenolytic enzymes

The radiolabelled glucose attached to SGP after undergoing radioglucosylation is not removable by glycogenolytic enzymes, namely alpha- amylase, beta-amylase and glucoamylaεe. It is removable, however, by beta-glucosidaεe. When a εample waε incubated with beta-glucoεidase only 50% of the radiolabelled glucose was removed, and only after 72 hr of digestion. The difficulty in hydrolysing the glucose by beta-glucosidaεe suggested that there may be steric hindrance to the enzyme. Therefore, it waε decided to teεt the beta- glucoεidaεe on a prelabelled tryptic peptide. When thiε waε done, beta-glucoεidaεe removed all the -

14 C-labelled glucose from the radiolabelled tryptic glucopeptide isolated from the prelabelled SGP.

Alpha-Glucosidaεe failed to remove any glucose from the εame glucopeptide and in the intact prelabelled

SGP.

The inability of alpha-amylaεe, beta-amylase and glucoamylaεe to remove the added glucose to SGP indicated that multiple alpha-glucose residueε were not attached to the protein. The removal of the glucoεe by beta-glucoεidaεe indicateε that it is attached via a beta-linkage. These data, in combination with other analyses, suggeεt that the glucoεe iε attached in a beta-linkage to an arginine residue on the protein.

4.3 Beta-elimination of radioglucosylated SGP

The alkaline borohydride procedure which beta- eliminates carbohydrate attached to serine or 43

threonine residueε on proteinε was attempted on radioglucosylated SGP from sweet corn. In this experiment prelabelled SGP from rabbit heart was uεed aε a control since its carbohydrate is attached to tyroεine. Surpriεingly, the glucoεe attached to both the εweet corn and heart proteins beta-eliminated or appeared to do so. After 42 hr of incubation, 63% of the radiolabel from the εweet corn protein and 89% from the heart protein became TCA-soluble. This observation suggested that solubilization of the radiolabel in TCA occurred not by beta-elimination but perhapε by fractionation of the protein into TCA-εoluble peptideε. It waε concluded, therefore, that the conditions of this procedure were to strong for the two proteins.

A second procedure commonly used for beta- elimination by incubation in 0.5M NaOH for 24 hr at 4°C waε performed on the sweet corn SGP. In thiε caεe, the glucoεe attached to the protein failed to beta-eliminate suggesting that the carbohydrate was not attached to a serine or threonine residue. This result is consiεtent with the finding that the glucoεe iε attached to an arginine residue.

4.4 Acid hydrolysis and Thin Layer Chromatography of Prelabelled SGP

A sample of prelabelled SGP from sweet corn was prepared by allowing the enzyme to undergo self glucosylation with UDP-[ 14C]glucoεe. Thiε waε accomplished by incubating 250ul of a 16 fold purified sample of the protein with 50mM Tris-HCl pH 7.4 buffer, 5mM MnCl₂ and 2uM UDP-[¹ C]glucoεe in a 500ul reaction mixture at room temperature for

30 min. The reaction waε εtopped by extenεive dialysis of the incubation mixture against 20mM Tris-HCl pH 7.4 buffer containing 2mM CHAPS. Two aliquots of the prelabelled protein (12000cpm each) were each precipitated with an equal volume of 20% TCA and the pellets collected by centrifugation. The precipitates were washed twice in ether to remove residual TCA. After drying in air, 1.0ml of 2M trifluoroacetic acid was added to the precipitates. For mild hydrolysiε one sample was heated for 30 minuteε in a boiling water bath. For strong hydrolysis the other sample was heated for 2 hr instead. Following hydrolysis, the sampleε were lyophilized and redissolved in lOOul of water.

After hydrolyεiε, lOul of each hydrolysate was spotted on thin layer silica plates (the plates were heated to 60°C for about 5 mins to remove moisture before use). After drying, the plates were placed in tanks containing a εolvent composed of butanol, water and ethanol in a ratio of 5 : 4 : 5. After 24 hr the plate was taken out, dried in an oven and then replaced for a second run. After an additional 24 hr the plate was taken out dried and placed for radioautography.

4.5 Digestion of Prelabelled Sweet Corn SGP with Glycogenolytic Enzymes

A sample of prelabelled SGP from sweet corn waε prepared. Five hundred microliters (500ul) of a 16 fold purified sample of the protein was allowed to self glucosylate with UDP[ 14C]glucoεe in a lOOOul reaction mixture. The prelabelled SGP εo formed was used in the following digestionε.

4.5.1 ALPHA-AMYLASE DIGESTION

Three hundred microliters (300ul) containing

22680cpm of prelabelled SGP was incubated with 45

lOOmM sodium acetate pH 7.0 buffer, lOmM CaCl₂ and 2U of alpha-amylase in a 400ul reaction mixture at room temperature. Hydrolysis of glucose attached to the SGP was monitored by measuring any losε of TCA precipitable countε. One hundred microliter (lOOul) aliquotε were taken at 1 hr, 3 hr and 18 hr time intervals and precipitated with 1ml of 10% TCA. The precipitate was filtered on nitrocellulose, washed and meaεured for C radioactivity.

4.5.2 BETA-AMYLASE DIGESTION

Three hundred microliterε (300ul) containing 22680cpm of prelabelled SGP was incubated with lOOmM sodium acetate pH 5.0 buffer and 0.5mM dithiothreitol with approximately 2U of beta- amylase in a 400ul reaction mixture at room temperature. Hydrolysiε of glucose attached to the SGP was monitored in exactly the same manner as described before.

4.5.3 GLUCOAMYLASE DIGESTION

Three hundred microliterε (300ul) containing 22680cpm of prelabelled SGP waε incubated with lOOmM sodium acetate pH 5.0 buffer and 4U of glucoamylase in a 400ul reaction mixture at room temperature. Hydrolysis of glucose attached to the SGP was monitored in exactly the same manner as described before.

4.6 Digestion of Prelabelled Muscle SGP with Glycogenolytic Enzymes

This experiment was performed as a control for the glycosidase digestion of prelabelled sweet corn SGP. A εample of prelabelled SGP from rabbit muεcle was prepared in a similar manner to that from sweet corn. Two hundred microliters (200ul) of a 20 fold purified sample of the muscle enzyme was allowed to undergo autoglucosylation in a 400ul reaction 46

mixture for this purpose. The prelabelled muscle SGP was incubated with glycogenolytic enzymes in the following digestions.

4.6.1 ALPHA-AMYLASE DIGESTION

Fifty microliters (50ul) containing 4907cpm of prelabelled muscle SGP was incubated with lOOmM sodium acetate pH 7.0 buffer, lOmM CaCl₂ and 2U of alpha-amylase in a lOOul reaction mixture at room temperature. Hydrolysis of glucose attached to muscle SGP was monitored by loss of TCA precipitable counts. After 18 hr of incubation the entire digest was precipitated with 1ml of 10% TCA.

The precipitate was filtered on nitrocelluloεe, washed and measured for 14C radioactivity.

4.6.2 BETA-AMYLASE DIGESTION

Fifty microliterε (50ul) containing 4907cpm of prelabelled muεcle SGP was incubated with lOOmM εodium acetate pH 5.0 buffer and 0.5mM dithiothreitol with approximately 2U of beta- amylaεe in a lOOul reaction mixture at room temperature. Hydrolyεis of glucose was monitored in exactly the same manner as described before.

4.6.3 GLUCOAMYLASE DIGESTION

Fifty microliters (50ul) containing 4907cpm of prelabelled muscle SGP was incubated with lOOmM sodium acetate pH 5.0 buffer and 4U of glucoamylase in a lOOul reaction mixture at room temperature. Hydrolysis of glucose was monitored in exactly the same manner as deεcribed before.

4.7 Digeεtion of Prelabelled SGP with alpha-and beta-Glucosidase

Prelabelled SGP was prepared from 500ul of al6 fold purified εample of the enzyme in a lOOOul reaction mixture and uεed for the following digestionε. 47

4.7.1 ALPHA-GLUCOSIDASE DIGESTION

Three hundred microliters (300ul) containing 25400cpm of prelabelled SGP was incubated with lOOmM sodium phosphate pH 7.0 buffer and 5U of alpha-glucosidaεe at room temperature. At 18 hr, 48 hr and 72 hr time intervalε, lOOul aliquots of the digest were taken and precipitated with 1ml 10%

TCA. The precipitate was filtered on nitrocellulose, washed and counted for 14C radioactivity.

4.7.2 BETA-GLUCOSIDASE DIGESTION

Three hundred microliters (300ul) containing 25400cpm of prelabelled SGP was incubated with lOOmM sodium acetate pH 5.0 buffer and 5U of beta- glucosidaεe at room temperature. One hundred microliter (lOOul) aliquotε were taken from the digest at 18 hr, 48 hr and 72 hr time intervals and precipitated with 1ml 10% TCA. The precipitate was filtered on nitrocellulose, washed and counted for

14 C radioactivity.

4.8 Digestion of Radioglucosylated Tryptic Peptide from SGP with alpha- and beta-glucosidase

4.8.1 TRYPSINOLYSIS OF PARTIALLY PURIFIED PRELABELLED SGP

A radiolabelled tryptic peptide was prepared from prelabelled SGP, obtained with a 16 fold purified sample of the enzyme in the following manner. Three hundred microliterε (300ul) containing 25400cpm of the prelabelled protein was precipitated with an equal volume of 20% TCA and collected by centrifugation. After washing the pellet twice with ether, it was dissolved in 50ul of 8M urea containing 0.4M ammonium bicarbonate at pH 8.0. The sample was heated at 55°C for 15 min 48

and then diluted to 200ul with water. TPCK-trypsin (Pierce) was then added to the sample to give a 1 : 25 protease : protein ratio and incubated at 37°C for 24 hr. Following this, the same amount of trypsin was added again and the sample incubated for a second 24 hr period.

4.8.2 RECOVERY OF RADIOGLUCOSYLATED TRYPTIC PEPTIDE After incubation was completed, the tryptic digest was applied to a Biogel P-4 column (80 x 0.8cm) equilibrated with 50mM ammonium bicarbonate buffer and eluted with the same buffer. Fractions of 0.5ml were collected and 50ul aliquots taken for measurement of 14C-radioactivity.

4.8.3 ENZYME HYDROLYSIS OF RADIOGLUCOSYLATED TRYPTIC PEPTIDE

Samples of the radioglucosylated tryptic peptide (5350cpm each) were digested with alpha-and beta-glucosidaεe separately, and each digest fractionated on the same P-4 column. Fractions of the same size as in (b) were collected and measured for 14C-radioactivity.

4.9 Beta-elimination of prelabelled SGP

Two procedures for beta-elimination of carbohydrate from serine/threonine linked proteins were performed on prelabelled SGP from sweet corn. 4.9.1 PROCEDURE 1

In this procedure, 210ul containing 15321cpm of prelabelled SGP from sweet corn (prepared from 16 fold purified enzyme) was incubated with 0.2M sodium hydroxide and 0.3M sodium borohydride in a 300ul digest at 45°C. Beta-elimination of glucose was determined by measuring any losε of TCA precipitable countε. At 15 hr and 42 hr time intervals 150ul aliquots were taken and precipitated with 1ml of 20% TCA. The precipitate 49

waε filtered on nitrocellulose, washed with 10%

TCA and water and the filter paper measured for

14 C-radioactivity. As a control, 155ul containing

15213cpm of prelabelled SGP from rabbit heart

(prepared from 20 fold purified enzyme) was incubated with 0.2M sodium hydroxide and 0.3M sodium borohydride in a 300ul digest at 45°C. Beta elimination of glucose was determined in exactly the same manner.

4.9.2 PROCEDURE 2

A sample containing 200ul of prelabelled SGP

(14591cpm) from sweet corn, prepared from 16 fold purified enzyme waε precipitated with an equal volume of 20% TCA. The pellet waε collected by centrifugation and waεhed twice with 1.5ml of ether. The pellet was redisεolved in 0.5ml of 0.5M

NaOH and a 250ul aliquot taken and precipitated with 1ml of 20% TCA. The precipitate was filtered on nitrocellulose, washed with 10% TCA and water an —d the filter paper measured for 14C- radioactivity. The rest of the solution was incubated at 4°C for 24 hr. After incubation, the digest waε meaεured for TCA-precipitable countε in exactly the same manner to determine whether any beta-elimination of glucose occurred.

5. ENZYMOLOGY OF SWEET CORN SGP

5.1 pH optima of sweet corn SGP

Most enzymes function within a limited pH range and exhibit maximal activity at a specific hydrogen ion concentration. A few operate in a broad range of pH. The sweet corn SGP not only functions over a wide range of hydrogen ion concentrations but exhibits maximal activity peaks 50

at two separate pH values 4.5 and 8.5. In other words, the enzyme has a double pH optima. Furthermore, the enzyme is about 3 times more active at pH 8.5 than at pH 4.5.

This uncommon pH-activity curve for the sweet corn SGP, especially the difference of 4 units between the two maximal activity peaks, suggested that perhaps two separate activitieε were being meaεured. The meaεurements were performed with partly pure enzyme that was 80 fold purified. The sample, however, were free from any lipid- glucosylating activity and when radioglucosylated, only the 42kDa SGP is radiolabelled. When the sample was fractionated on Sepharose C1-6B there was an enrichment of the pH 8.5 activity in the early fractions of the SGP peak suggesting that, perhaps there were two separate but εimilar SGP'ε. However, analyεis of the glucosylated protein after self glucosylation at pH 4.5 and 8.5 revealed that the products, at leaεt, were εimilar in that both were εtable to alkali and the added glucose were removable by beta-glucoεidaεe. When both products were analyzed by SDS-PAGE each had the same molecular weight. It appears, therefore, that there is only one SGP that has an uncommon double- pH optima. The result of the enrichment of the pH 8.5 activity in the early fractions of C1-6B remains unreεolved.

5.2 Inhibition of εweet corn SGP

When a cell-free extract is asεayed, an almoεt negligible amount of glucosylating activity is detected. The glucosylating activity is only expresεed after the extract is dialyzed. Apparently a low-molecular weight inhibitor(ε) is preεent in 51

the extract that inhibitε both the self- glucoεylating protein and the lipid-glucoεylating activity. Furthermore, the inhibitor is stable to heat as a heated extract inhibits the glucosylating activity in a dialyzed extract.

The endogenouε inhibitor in the sweet corn extract was not identified. However, ATP was found to be a potent inhibitor to the εweet corn SGP. Rabbit muεcle SGP iε inhibited by ATP with 50% inhibition observed at a concentration of 3.5mM. ATP is a 30-fold more potent inhibitor of sweet corn SGP with 50% inhibition occuring at a concentration of lOOuM.

The self-glucosylating protein in rabbit muscle is inhibited by p-nitrophenyl-saccharides, namely p-nitrophenylglucose, p-nitro- phenylmaltoεe, p-nitrophenylmaltotrioεe and p- nitrophenyl- maltotetrose. These synthetic compoundε inhibit by acting as alternative substrates. The rabbit muscle SGP functionε by tranεferring glucose from UDPglucose to a pre- exiεting oligoεaccharide chain attached to itεelf via tyrosine, extending the chain to eight glucose unitε long. Theεe p-nitrophenylεaccharides, by mimicking the acceptor site, can act as acceptors for the tranεferred glucoεe. In addition, thiε behaviour of the p-nitrophenylsaccharides further demonstrates the function of rabbit muscle SGP in syntheεizing alpha-1,4 glucoεidic bondε. p- Nitrophenol, p-nitrophenyl-alpha-glucoside and p- nitrophenyl-beta-glucoside were tested on sweet corn SGP and no effect on its activity was observed. This observation indicated that there was a fundamental difference in the self- 52

glucosylating activity between the muscle and sweet corn proteins. In addition arginine and a synthetic peptide comprising the same amino acid sequence as the glucosylated tryptic peptide T2 were examined for any inhibitory effect on the activity of the sweet corn SGP. No effect was observed, indicating that the enzyme is unable to utilize these as subεtrateε.

5.3 Dependence of εweet corn SGP on divalent cations

The muεcle SGP iε εtimulated by manganeεe ionε. Sweet corn SGP iε completely dependent on divalent cationε for itε activity. When treated with EDTA the sweet corn SGP looses all itε activity. The concentration dependence of sweet corn SGP on various divalent cations was measured. Manganese has the most potent stimulating effect at a concentration of 5mM. Magnesium also activated SGP but to a lesεer extent than manganese, achieving a maximum stimulating effect of about 50% at lOmM concentration. Calcium had no stimulatory effect on SGP activity.

5.4 Radioglucosylation in the absence of MnCl₂ When a cell-free extract is assayed for SGP in the absence of Mn ion, in addition to the SGP, a number of proteins of higher molecular weight became radiolabelled. Of these, the most prominent radiolabelled band occurs in the region of approximately 70kDa. This form of apparent radioglucosylation occurs within the 30 min incubation period allowed for the SGP assay. The activity, however, was preεent only εometimes and in most preparations was abεent. 3 -3

A similar form of radioglucosylation has been detected with partially purified extracts (16 fold) in which proteins of molecular weight higher than SGP become radiolabelled in the absence of manganese ions. However, this form of radioglucoεylation iε detectable only after prolonged incubation and occurε over a 24 hr period. In addition the reaction proceedε in the preεence of ImM MnCl₂ but is inhibited by 5mM of the divalent cation. Furthermore, UDPglucose as well as ADPglucose serve as subεtrateε for the reaction. Thiε iε unlike SGP which utilizeε UDPglucoεe as a substrate but not ADPglucose.

The radiolabelled proteins formed during this 24 hr incubation period range in molecular weight between 42kDa and 200kDa and appear as sharp distinct bands on a radioautograph following SDS- electrophoresis. Evenly spaced radiolabelled bands appear between 42kDa and 70kDa separated by approximately 3kDa.

The formation of these higher molecular weight products could arise from elongation of the carbohydrate component of SGP. In thiε caεe, the radiolabelled bandε would represent SGP with different amounts of carbohydrate. In order to test this hypothesiε, the radiolabelled productε were incubated with glycogenolytic enzymeε in an attempt to remove the radiolabel from the proteinε. It waε found that the radiolabel waε resistant to hydrolyεiε by alpha-amylaεe, glucoamylaεe, phosphorylase and isoamylaεe. Conεequently the radiolabelled productε were incubated with cellulaεe to determine whether the radioactive glucose was attached in a beta-linkage. Cellulaεe 54

failed to remove the radiolabel. It appearε therefore that theεe higher molecular weight productε are not εtarch-like or celluloεe-like molecules and represents some other form of glucosylation. There is no evidence that these proteins are related to SGP.

6. PRIMING ABILITY OF SWEET CORN AMYLOGENIN (SGP)

6.1 Experimental strategy

If the sweet corn SGP was the protein primer upon which the starch molecule waε conεtructed, then it may be poεsible to synthesize starch in vitro by incubating the primer with the starch synthesizing enzymes, and the glucose donor εubεtrateε (UDPglucoεe and ADPglucoεe). The εyntheεiε of εtarch on the protein primer could be detected by monitoring the increase in the molecular weight of the protein as the carbohydrate increases in size. Thiε could be accompliεhed by SDS-PAGE which can demonstrate any change in the molecular weight of the protein.

This approach was used to determine whether εweet corn SGP could act as a primer for the syntheεis of polysaccharide. The SGP was prelabelled by allowing the protein to undergo self-glucoεylation with UDP-[ 14Clglucoεe. Thiε permitted the detection of the protein by radioautography. Cell-free extractε of εweet corn were uεed aε a εource for the starch synthesizing enzymes. It was poεsible that, in addition to starch synthaεe and branching enzyme, other enzymes were involved in the synthesis of starch. If SGP was a primer, a 'chain extender' enzyme may be 55

necesεary to extend the single glucoεe into an oligoεaccharide before εtarch εynthase and branching enzyme can act.

6.2 Sweet corn SGP is a primer

The prelabelled protein, therefore, was incubated with cell-free extracts of corn and millimolar concentrations of UDPglucose and/or ADPglucose. At different time intervals aliquots of the incubation mixture were analyzed by SDS-PAGE and radioautography. After 1 hr of incubation, the appearance of a radiolabelled band at the top of the gel occurred and this was also observed for 2 hr and 3 hr time periods. In addition, there was a decrease in the amount of radiolabel at 42kDa, the position of prelabelled SGP. The decrease in intensity of the SGP band was moreso at 2 hr and 3 hr.

The appearance of the radiolabelled band at the top of the gel could only originate from the prelabelled SGP. It appears, therefore, that the SGP was converted into a large molecular weight molecule by incubation with both UDPglucose and ADPglucoεe suggesting that the protein can function as a primer in polysaccharide syntheεis.

6.3 Demonstration of the Priming Ability of Sweet Corn SGP

The priming ability of sweet corn SGP waε examined by molecular-weight-shift experiments carried out in 3 stageε. 6.3.1 PREPARATION OF PRELABELLED SGP

Prelabelled SGP for thiε purpoεe waε prepared from a 16 fold purified extract by incubating a εample of the enzyme with 2uM UDP-[ 14C]glucoεe, 5mM MnCl₂ and 50mM Tris-HCl pH 7.4 buffer at room temperature for 20 min. Following incubation, the reaction mixture was applied to a Sephadex G-25 column (6 x 1.5cm) equilibrated with 50mM Tris-HCl pH 7.4 buffer containing 2mM CHAPS and eluted with the same buffer. Fractions (1ml) were collected and the prelabelled SGP was detected by measuring 50ul aliquotε for C-radioactivity.

6.3.2 INCUBATION OF PRELABELLED SGP WITH 1MM UDPG AND/OR ADPG

The prelabelled SGP, (typically about lOOOOcpm) was incubated with ImM UDPglucose or ImM ADPglucose or both and an equal volume of dialyzed or undialyzed sweet corn extract at room temperature in 1ml digeεts. At 1 hr, 2 hr, 3 hr and 24 hr time intervals 250ul aliquots were taken and frozen to stop the reaction. The protein was precipitated with an equal volume of 20% TCA and collected by centrifugation. The pellets -were waεhed twice with ether and dried.

6.3.3 ANALYSIS BY SDS-PAGE AND FLUOROGRAPHY

The dried pellets were desεolved in 50ul of εolubilizing buffer by incubation for 24 hr followed by εtrong vortexing. SDS-PAGE were performed on lOul of the εample followed by fluorography.

7. EXPERIMENTAL BACKGROUND TO THE INVENTION

In order to examine in detail the relationεhip between amylogenin and εtarch synthesiε rate and other εtarch εynthetic enzymes, experiments were conducted on maize plants growing under defined temperature regimes. In these experiments the maize endosperm developmentally regulated expresεion of amylogenin was found to parallel that 57

of starch deposition, with minimal amylogenin activity early in endoεperm development and increasing amylogenin expresεion as εtarch depoεition progreεεed until physiological maturity of the grain.

Results demonstrated that the initiation of starch synthesis in maize endosperm iε attributable to the co-ordinated expreεεion of εeveral enzymeε involved in the pathway of εtarch εynthesis. Continued starch deposition is dependent on a continued expresεion of amylogenin which actε aε an important control-point in the pathway.

7.1 Developmental aspects of the formation of corn SGP in relation to starch synthesis

The appearance of SGP during development must precede or coincide with the onεet of starch production if the protein is a primer for starch syntheεis. The formation of SGP at a later time will certainly preclude it from any involvement in the biogeneεis of starch. Therefore an examination of the formation of SGP during development, particularly the period during which the polyεaccharide iε laid down in the cell, would provide conεiderable εupport for or against the hypothesiε that the protein iε a precursor of starch.

An examination of the formation of SGP and different components of the starch εynthesizing system was performed on the endosperm of B73 commercial dent corn sampled at 6-30 days after pollination. The enzyme activitieε of SGP, ADPG pyrophosphorylase, ATP-dependent fructokinase, ATP- dependent glucokinase, branching enzyme, bound starch synthaεe, ATP-dependent phosphofructokinase, PPI-dependent phosphofructokinase, phosphoglucomutaεe, εoluble εtarch synthase, sucrose synthase, UDPG pyrophosphorylase and UTP- dependent fructokinase were measured in terms of enzyme concentration, as unitε/gram dry weight and enzyme content aε unitε/endoεperm with the reεultε expreεsed as a percentage of the maximum value attained. The starch content, dry weight of the endoεperm, the number of endosperm cells, number of starch granules, the albumin content and concentration and the rateε of starch accumulation were also measured.

It was found that SGP was already present at 6 days after pollination where other enzymes important in εtarch εyntheεiε εuch aε εucrose synthase, bound starch synthase, soluble starch synthase and branching enzyme are low in activity at this time. Theεe enzymeε reach their peak along with SGP at 10 days after pollination. The relative rate of starch accumulation is greatest between 10 and 14 days after pollination.

The appearance of SGP well before the accumulation of starch and its peak concentration occurring together with the other starch syntheεizing enzymes is consiεtent with the protein being a precurεor of εtarch.

7.2 Formation of SGP and enzymeε involved with the εtarch synthesizing system during development in B73 commercial dent corn

B73 commercial dent corn was harvested at different times between 6-30 days after pollination inclusively. The enzyme activities of SGP, ADPG 59

pyrophosphorylase, ATP-dependent fructokinase, ATP- dependent glucokinase, branching enzyme, bound εtarch synthase, ATP-dependent phosphofructokinase, PPI-dependent phosphofructokinase, phosphoglucomutase, soluble starch synthaεe, εucrose synthase, UDPG pyrophosphorylase and UTP- dependent fructokinase were meaεured in termε of enzyme concentration, aε unitε/endoεperm. Theεe were expressed as a percentage of the maximum value measured. In addition, the starch content, dry weight of the endosperm, the number of endoεperm cellε, number of εtarch granules, the albumin content and concentration and the rates of starch accumulation were also determined.

7.3 Purification of Prelabelled SGP from B73 Commercial Dent Corn

The isolation of prelabelled self- glucosylating protein from B73 commercial dent corn was achieved in a 800 fold purification over the cell-free extract. The protein was partially purified in an active state by εuccessive precipitationε with acid and εalt followed by fractionation on DEAE-Sepharoεe. At thiε εtage the 27 fold purified enzyme waε allowed to undergo εelf-glucosylation and the prelabelled protein purified further.

The prelabelled protein was purified by εuccessive fractionations on Sephadex S300, Mono Q and C₄. The degree of purification and yield at each stage of the protocol is listed in Table 3. 60

TABLE 3 Purification of εweet corn amylogenin

The purified protein had a εpecific activity of 3.5 x 10 cpm/mg protein. Only 25% of the glucoεe, however, waε radiolabelled. Therefore the calculated εpecific activity of the B73 corn SGP with all the incorporated glucoεe being radiolabelled would be 1.4 x 10 cpm/mg protein. The εpecific activity of the radiolabelled glucose was 4.4 x 10 14cpm/mole of glucose. Thiε implied that the B73 dent corn SGP had incorporated 0.13 mole of glucoεe into each mole of protein.

7.3 Purification of Prelabelled SGP from B73 Commercial Dent Corn

The iεolation of self-glucosylating protein from B73 commercial dent corn waε carried out with cobε εampled at 18 and 22 dayε after pollination. The protein waε iεolated in a prelabelled form and 61

waε achieved in two phases.

7.3.1 PHASE 1

This phase of the purification was carried out in exactly the same manner as described for phaεe 1 in the purification of prelabelled SGP from εweet corn. The protein waε partially purified and allowed to undergo autoglucoεylation, 25% with UDP[ C]glucoεe and the remaining 75% with unlabelled UDPglucoεe. The prelabelled protein waε then purified further.

7.3.2 PHASE 2, PURIFICATION OF THE PRELABELLED SGP Step 4: Fractionation by gel filtration chromatography

The prelabelling reaction mixture was applied to a Sephadex S-300 column (2 x 60cm) equilibrated with 50mM Tris-HCl pH 7.4 buffer containing 0.02% triton and eluted with the εame buffer. Fractionε of 3ml each were collected and the prelabelled SGP detected by measuring 50ul aliquots from 14 alternating fractions for C-radioactivity. The fractionε containing the prelabelled SGP were analyzed by SDS-PAGE and thoεe that εhowed the highest protein purity were combined. The pooled fractions were concentrated by centrifugation uεing centriprep-10 concentratorε.

Step 5: Fractionation by faεt protein liquid chromatograpy (FPLC) on Mono Q

The concentrated εample waε εubmitted to faεt protein liquid chromatography (FPLC) on a Mono Q column. Protein waε eluted with a linear gradient of NaCl, from 0-0.4M, formed from 50mM Triε-HCl pH 7.4, 0.02% triton (εolution A) and 50mM Tris-HCl pH 7.4, 0.02% triton, 1.0M NaCl (εolution B). The gradient waε applied over a period of 40 min at a flow rate of lml/min. Fractions of 1ml were collected. The prelabelled SGP was detected by measuring the fractions for 14C-radioactivity and its purity examined by SDS-PAGE. The fractions containing the best purity were combined and concentrated.

Step 6: Fractionation by Reverse-Phaεe HPLC on

^C4

The prelabelled SGP waε purified next by reverse-phase HPLC using a Vydac C₄ column (8 x

40mm) . The sample was fractionated in portions of

0.16mg each. When a larger portion containing approximately 0.32mg was applied to the column a reduction in the resolution of the SGP containing peak waε observed. Protein was eluted with a linear gradient of acetonitrile generated by using

0.1% TFA (solution A) and 0.09% TFA containing 84% acetonitrile (solution B) at a flow rate of lml/min. The gradient of 0-100% solution B was applied over a period of 50 min. Protein was detected by monitoring the abεorbance of the effluent at 215nm and 295nm simultaneously. Peaks containing protein were collected manually. The prelabelled SGP emerged from the column in a single peak.

7.4 Compariεon of the propertieε of εweet corn SGP (amylogenin) and muεcle SGP (glycogenin)

8. GENERATION OF ANTIBODIES TO THE ENZYME

Polyclonal antibodies to purified amylogenin were generated in goats by the procedure of Lomako et al (1988, FASEB J, 2:3097-3103). Sufficient protein to enable the immunisation was obtained by pooling the enzyme subunit iεolated as described above from a number of separate experiments. Antisera were prepared against the 39 kD amylogenin polypeptide. The antibodies were then tested for specificity to the amylogenin polypeptides.

The immune sera obtained precipitated amylogenin activity from crude extractε of maize endoεperm. Following incubation of extractε with immune εerum and centrifugation of the enzyme-Igγ-protein A-Sepharose conjugates, the amylogenin activity was detectable in the pellet fractionε when the waεhed enzyme-Igγ-protein A-Sepharoεe conjugateε were assayed directly for enzyme activity.

In control experiments, pre-immune sera collected from animals prior to their primary immunisation with amylogenin did not partition amylogenin activity.

Western blot analysiε of total maize_endoεperm soluble protein using anti-maize endosperm amylogenin serum revealed the presence of a single 39 kD immunoreactive polypeptide.

Two further types of polyclonal antibody have been made. One was generated against rabbit-muscle glycogenin obtained from rabbit-muscle glycogen. The other was generated from three synthetic peptides coupled to bovine serum albumin using glutaraldehyde. The peptides were: GDAMVTDIEAADE REGANFVRGYPFSLR NLDFLEMWRPFFQP 03

9. IDENTIFICATION OF AN AMYLOGENIN CDNA CLONE FROM MAIZE ENDOSPERM

The antibodies produced above may be used to screen a maize endosperm cDNA library for clones derived from the mRNAε for amylogenin in an in vitro tranεcription/ tranεlation εyεtem.

The cDNA library may alεo be εcreened with oligonucleotide εequenceε derived from the amino acid εequence of amylogenin. Numerous oligonucleotide probes have been synthesised.

An amplified maize endosperm cDNA library has been constructed. Double-stranded cDNA was prepared from oligo dT-cellulose-purified maize endosperm RNA by a method employing RNaseH and E coli DNA polymerase for the εynthesis of the εecond εtrand, without prior purification of εingle-stranded cDNA (Gubler and Hoffman, 1983). The antibody and oligonucleotide probes are used to screen this library to identify cloneε containing cDNA encoding maize endoεperm amylogenin.

Cloneε producing εtrong poεitive εignalε are plaque-purified during additional rounds of screening. DNA is prepared from the putative maize endosperm amylogenin cDNA clones. Following restriction endonuclease digeεtion of the DNAε with EcoRl and agarose gel electrophoresis, the sizes of the cDNA insertε of theεe clones are determined. The cDNA insert is labelled with 32P by nick-translation and used to probe a northern blot of poly(A)-containing RNA from maize endosperm.

The size of the mRNA bands to which thiε probe hybridiεeε indicateε the size of the maize endoεperm amylogenin mRNA. This should be in close agreement with the eεtimateε from western blots of maize endosperm amylogenin.

9.1 Molecular cloning of the amylogenin cDNA

An amplified cDNA library (10 plaque-forming units/ml) from corn endoεperm constructed on lambda ZAPII was used as the DNA template for PCR. A sense primer CC(T/C/A/G)TTCTTCTT(C/T)AA(C/T)AC was designed from the partial amino acid sequence of the protein while the antisense primer, TCGACGGTATCGATAAGC, was from lambda ZAPII nucleotide sequence. After 30 cycles (95°C for 1 min; 45°C for 2 min and 72°C for 2 min followed by a final extension step for 6 min at 72°C), the PCR product was cloned into a PCR 1000 vector (invitrogen) and sequenced uεing Sequenase Version 2.0 (USB). A probe about 500 base pairs in length was obtained containing the deduced sequence for three peptides already sequenced directly. Digoxigenin was then incorporated into the cloned fragment aε digoxigenin-11-dUPT by meanε of PCR, uεing the modified nucleotide in a 1:2 ratio with dTTP. This served as a probe for screening the cDNA library. Approximately 10 plaque-forming units of lambda ZAPII phage were used to infect E coil strain XLl-Blue (Stratagene) and after plating the plaques were transferred to nylon membranes in duplicate. The hybridisation buffer contained 5xSSC (chloride-citrate), 0.5% blocking reagent (Boehringer), 0.1% N-laurylsarcoεine, 0.02% SDS and 15 ng/ml of DNA probe. After overnight hybridiεation at 65°C, the membraneε were waεhed with O.δxSSC and 0.1% SDS at 65°C. Putative positive clones were detected by the Genius kit (Boehringer), subjected to three additional rounds of purification and excised from lambda ZAPII to yield recombinant pBluescript phagemid carrying ampicillin reεiεtance, using ExAssist helper phage in the iτι vivo excision protocol provided by Stratagene, which utiliseε the E coli SOLR εtrain.

10. SEQUENCE ANALYSIS OF THE AMYLOGENIN cDNA CLONES

The cDNA inεert from each clone may be sequenced using standard techniques.

Each clone will encode a polypeptide equivalent to a partial or a complete amylogenin protein. The hydropathy index (Kyte and Doolittle, 1982) and secondary structure predictions (Chou and Fasman, 1978) may be calculated for the encoded protein using the programmes of Devereux et al (1987). The Kyte and Doolittle (1982) hydropathy profile indicates the hydrophobic and hydrophilic domainε within the protein and εhould be conεistent with its location in the cell (Kaiser and Basεham, 1979). The secondary structure predictions of Chou and Fasman (1978) may be used in compariεons of the maize endosperm and animal glycogenin and/or amylogenin proteins.

The various amylogenin and/or glycogenin cDNA sequences may be compared to indicate the extent of homology. A detailed comparison of the nucleotide sequences of the cDNA will identify base substitutions, insertions or deletions, and hence the amino acid changes produced in the derived protein sequences.

One amylogenin cDNA clone has been sequenced by the dideoxy method using Sequenase Version 2.0 DNA Sequencing Kit (USB). The cDNA clone encoding amylogenin was deposited at the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland 20852, USA under the terms of the Budapest Treaty on 19 August 1993 under the accesεion number ATCC 69389.

Partial cDNA εequenceε correεponding to amylogenin from B73 maize are given below in Tableε 4 and 5. Theεe match-up with the amino acid εequences of some of the peptide fragmentε shown above.

TABLE 4

SEQUENCE (I)

TGAACTTGGCCTTTGACCGTGAGCTCATTGGTCCGGCTATGTACTTCGGTC

TCCTGGGTGATGGTCAGCCTATTGGTCGCTACGACGATATGTGGGCTGGGT

GGTGTGTCAAGGTGATCTGTGATCATTTGGGATTGGGAGTGAAGACGGGTC

TTCCCTACATCTACCACAGCAAGGCGAGCAACCCATTTGTGAACCTGAAGA

AGGAGTACAAGGGAATTTTCTGGCAGGAGGACATCATGCCTTTCTTCCAGA

GTGCAAAGCTCTCGAAAGAAGCTGTGACGGTTCAACAATGCTACATTGAGC

TGTCCAAGATGGTGAAGGAGAAGCTTAGCGCCATTGATCCTTACTTTGACA

AGCTTGCTGATGCTATGGTGACTTGGATTGACGCTTGGGATGTGCTTAACC

CGGCCACATAAG

TABLE 5

SEQUENCE (II)

CTTCCGTTCTTCTTTAACACCTTGTACGATCCCTACCGTGAAGGTGCTGAC

TTCGTCCGTGGATACCCTTTCAGTCTCCGTGAGGGTGTTTCCACTGCTGTT

TCTCACGGTCTCGGGCTCAACATCCCTGATTACGACGCCCCAACTCAACTC

GTCAAGCCTAAGGAAAGAAACACAAGGTATGTGGATGCTGTCATGACCATC

CCAAAGGAACACCTTTGGCCAATTGTGTGGCATGAACTGCC The cDNA clones may also be compared by restriction mapping. Restriction maps of the maize endosperm amylogenin cDNAs are constructed to demonstrate their relatedness.

Sequences may also be compared to other glycogenin and/or amylogenin polypeptideε. For example, the amino acid εequences derived from the maize amylogenin cDNAs may be aligned with the amino acid εequenceε derived from an animal glycogenin cDNA (Cohen et al, 1990). The homologies between amino acid εequences are subsequently calculated on the basiε of thiε alignment.

11. IDENTIFICATION OF A MAIZE ENDOSPERM AMYLOGENIN gDNA CLONE

The cDNAs derived as above may be used to probe a maize genomic library to isolate maize amylogenin genomic DNAs.

In addition, the oligonucleotide probeε generated from the amylogenin N-terminal amino acid εequence information may be uεed to εcreen the maize genomic library.

12. CONSTRUCTION OF TRANSFORMATION CASSETTE

The glycogenin and/or amylogenin gene tranεformation conεtruct requires the presence of an amyloplast tranεit peptide to enεure its correct localisation in the amyloplast. It is believed that chloroplast transit peptides have similar sequences but other potential sources are available such as that attached to ADPG pyrophosphorylaεe (Plant Mol Biol Reporter, 1991, 9:104-126). Other potential tranεit peptideε are thoεe of εmall subunit RUBISCO, acetolactate synthase, glyceraldehyde-3P-hydrogenase and nitrite reductase. For example, consenεus sequence of the transit peptide of small subunit RUBISCO from many genotypes has the sequence:

MASSMLSSAAV—ATRTNPAQAS MVAPFTGLKSAAFPVSRK QNLDITSIA SNGGRVQC; the corn small subunit RUBISCO has the sequence: MAPTVMMASSAT-ATRTNPAQAS AVAPFQGLKSTASLPVARR SSRSLGNVA SNGGRIRC; the tranεit peptide of leaf εtarch εynthaεe from corn has the sequence:

MA ALATSQLVAT RAGLGVPDAS TFRRGAAQGL RGARASAAAD TLSMRTASARA APRHQQQARR GGRFPSLWC; the transit peptide of leaf glyceraldehyde-3P- dehydrogenase from corn has the sequence: MAQILAPS TQWQMRITKT SPCATPITSK MWSSLVMKQT KKVAHSAKFR VMAVNSENGT; the putative transit peptide from ADPG pyrophoεphorylaεe from wheat haε the sequence: RASPPSESRA PLRAPQRSAT RQHQARQGPR RMC.

It is posεible to expreεs the glycogenin and/or amylogenin enzyme constitutively uεing one of the well-known constitutive promoters such as CaMV35S but there may be biochemical penalties in the plant resulting from increased starch deposition throughout the entire plant. Deposition in the endosperm is preferred. Posεible promoterε for use in the invention include the promoters of the starch synthase gene, ADPG pyrophosphorylase gene, and the sucroεe εynthaεe gene. 13. INSERTION OF GLYCOGENIN AND/OR AMYLOGENIN DNA INTO MAIZE

The glycogenin and/or amylogenin DNA may be transformed into either protoplastε or other tissueε of a maize inbred line or population. The exiεting gene promoterε in genomic DNA ensure that the extra genes are expreεεed only in the developing endoεperm at the correct developmental time. The protein εequences likewise ensure that the enzymes are inserted into the amyloplast.

14. TRANSGENIC PLANTS AND PROGENY

Transgenic maize plants are regenerated and the endosperms of these plants are tested for increased glycogenin and/or amylogenin enzyme activity. The kernels are also tested for enhanced rate of starch syntheεiε at different temperatureε, εtarch yield, etc.

The plantε are then included in a breeding program to produce new maize hybridε with altered εtarch εyntheεiε.

Claims

1. A method of producing a plant with altered starch syntheεiεing ability compriεing εtably incorporating into the genome of a recipient plant at leaεt one donor gene encoding a εtarch primer.

2. A method aε claimed in claim 1 in which the plant haε an increaεed capacity to produce εtarch.

3. A method aε claimed in claim 1 or claim 2 in which the plant haε an ability to εyntheεiεe starch with an altered structure.

4. A method as claimed in claim 1 in which at leaεt one of the donor genes encodes amylogenin.

5. A method as claimed in claim 1 in which at least one of the donor genes encodes glycogenin.

6. A method as claimed in claim 1 in which at leaεt one of the donor geneε encodes protoglycogenin.

7. A method as claimed in claim 1 in which at least one of the donor genes is in the senεe orientation and encodeε all or part of the εtarch primer.

8. A method as claimed in claim 1 in which at least one of the donor genes iε in the antiεenεe orientation and encodeε all or part of the εtarch primer.

9. A method aε claimed in claim 1 in which at leaεt one of the donor geneε encodeε a modified allelic form of the enzyme.

10. A method as claimed in claim 1 in which at least one of the donor genes is derived from a plant.

11. A method as claimed in claim 1 in which the recipient plant is of the family Gramineae.

12. A method as claimed in claim 11 in which the recipient plant is of the specieε Zea mayε.

13. A method aε claimed in claim 1 in which the recipient plant iε a tomato.

14. A plant and progeny thereof having at leaεt one donor gene encoding a starch primer stably incorporated into the plant's genome such that the plant's ability to εynthesise εtarch iε altered.

15. A plant aε claimed in claim 14 in which at leaεt one of the_. donor geneε encodeε amylogenin.

16. A plant aε claimed in claim 14 in which at leaεt one of the donor genes encodes glycogenin.

17. A plant as claimed in claim 14 in which at least one of the donor genes encodeε protoglycogenin.

18. A plant as claimed in claim 14 which is of the family Gramineae.

19. A plant as claimed in claim 18 which is of the εpecies Zea mays.

20. A plant as claimed in claim 14 which is a tomato.

21. Seed of a plant as claimed in claim 14. 74

22. A plant genome which is homozygous for the donor gene or genes encoding the said starch primer.

23. A hybrid plant of which at least one parent is a plant as claimed in claim 22.

24. A DNA construct comprising a DNA sequence encoding amylogenin.

25. A DNA construct comprising a DNA sequence encoding glycogenin.

26. A DNA construct comprising a DNA sequence encoding protoglycogenin.

27. A DNA construct comprising a DNA having the nucleotide sequence shown in Table 4.

28. A DNA construct comprising a DNA having the nucleotide sequence shown in Table 5.

29. A DNA construct comprising all or part of a cDNA sequence as deposited at the American Type Culture Collection on 19 August 1993 under the accesεion number ATCC 69389.