expression of escherichia coli clycogen synthase in the ... · plant physiol. (1994) 104: 1159-1166...

Plant Physiol. (1994) 104: 1159-1166

Expression of Escherichia coli Clycogen Synthase in the Tubers of Transgenic Potatoes (So/anum tuberosum)

Results in a Highly Branched Starch

Christine K. Shewmaker, Charles D. Boyer', Dennis P. Wiesenborn', Donald B. Thompson3, Micheal R. Boersig, Janette V. Oakes, and David M. Stalker*

Calgene, Inc., 1920 Fifth Street, Davis, California 9561 6

A chimeric gene containing the patatin promoter and the transit- peptide region of the small-subunit carboxylase gene was utilized to direct expression of Escherichia coli glycogen synthase (glgA) to potato (Solanum tuberosum) tuber amyloplasts. Expression of the glgA gene produd in tuber amyloplasts was between 0.007 and 0.028% of total protein in independent potato lines as determined by immunoblot analysis. Tubers from four transgenic potato lines were found to have a lowered specific gravity, a 30 to 50% reduction in the percentage of starch, and a decreased amylose/ amylopedin ratio. Total soluble sugar content in these seleded lines was increased by approximately 80%. Analysis of the starch from these potato lines also indicated a reduced phosphorous content. A very high degree of branching of the amylopectin fradion was detected by comparison of high and low molecular weight carbohydrate chains after debranching with isoamylase and corresponding high-performance liquid chromatography analysis of the products. Brabender viscoamylograph analysis and differential scanning calorimetry of the starches obtained from these transgenic potato lines also indicate a composition and structure much different from typical potato starch. Brabender analysis yielded very low stable paste viscosity values (about 30% of control values), whereas differential scanning calorimetry values indicated reduced enthalpy and gelatinization properties. The above parameters indicate a nove1 potato starch based on expression of the glgA E. coli gene produd in transgenic potato.

Glycogen is a highly branched glucan molecule based on h-1,4-linked D-G~c residues with a-1,6-glucosidic branch points. Although the linkage types in glycogen are identical with those in plant starch, glycogen exhibits a different average chain length and degree of polymerization. In bac- teria, the a-1,6-glucosidic linkages constitute approximately 10% of the total linkages. In vascular plants, reserve polysac- charides are stored in roots, tubers, and seeds in the form of starch. Starch, a complex mixture of D-Glc polymers, consists

Present address: Department of Horticulture, Oregon State Uni- versity, Agricultura1 and Life Science 4017, Corvallis, OR 97331- 7304.

Present address: Department of Agricultura1 Engineering, North Dakota State University, State University Station, Fargo, ND 58105- 5626.

Present address: 114 Borland Laboratory, Pennsylvania State University, University Park, PA 16802.

* Corresponding author; fax 1-916-753-1510.

essentially of linear chain (amylose) and branched chain (amylopectin) glucans. Starches isolated from different plants are comprised of distinct amylose and amylopectin fractions (Manners, 1985; Swinkels, 1985). Typically, amylose makes up between 10 and 25% of plant starch. Linear regions of amylopectin are composed of low mo1 wt and high mo1 wt chains, with the low ranging from 5 to 30 Glc units and the high mo1 wt chains from 30 to 100 or more. The amylosel amylopectin ratio and the distribution of low and high mo1 wt D-G~c chains can affect starch granule properties such as gelatinization temperature, retrogradation, and viscosity (Blanshard, 1987).

Natural mutations that affect amylose/amylopectin ratios have been analyzed in a number of plant species (Shannon and Garwood, 1984). A well-studied example is the waxy (wx) mutant of maize. This mutant lacks GBSS and results in a maize starch containing 100% amylopectin. Waxy com starch also exhibits the highest ratio of low mo1 wt/high mo1 wt chains (3.9/1) among plant starches and exhibits viscosity and elasticity properties greater than normal (Moore et al., 1984; Hizukuri, 1985). Recently, a double mutant of maize, duwx, has been shown to result in a more highly branched starch than wx (Duxbury, 1989; Yuan et al., 1993). Mutations affecting the starch-branching enzyme in pea (Smith, 1988) result in seeds having reduced starch content and a lower proportion of amylopectin. Similar to wx maize, a potato (Solanum tuberosum) mutant (amf) has been identified that contains amylose-free starch (Hovenkamp-Hermelink et al., 1987). Recently, this mutation was characterized as a single- point mutation in potato GBSS and could be complemented by re-introduction of a gene expressing active GBSS into amf mutant potato lines (van der Leij et al., 1991a, 1991b).

Antisense experiments with the gene for the potato GBSS have been used to obtain transgenic potato lines with altered amylose/amylopectin ratios (Visser et al., 1991). Recently, a mutant Escherichia coli ADP-Glc pyrophosphorylase (g lg C) has been introduced and expressed in potato tuber amyloplasts and resulted in increased starch accumulation in the tubers (Stark et al., 1992). Conversely, antisense experiments involving a single subunit of the endogenous potato ADP- Glc pyrophosphorylase resulted in a decreased starch content and increased soluble sugar content within the tubers of

Abbreviations: GBSS, granule-bound starch synthase; SSU, small subunit; TBST, 0.02 Tris-HC1 (pH 7.5b 0.5 M NaCl, 0.1% Tween-20.

1159

https://plantphysiol.orgDownloaded on November 14, 2020. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

https://plantphysiol.org

1160 Shewmaker et al. Plant Physiol. Vol. 104, 1994

transgenic potato plants (Müller-Rober et al., 1992). Unique carbohydrates can also be synthesized from reserve plant starch in vivo. This was shown by the introduction of a cyclodextrin-glycosyltransferase into potato tuber amyloplasts and the corresponding synthesis of cyclodextrins (Oakes et al., 1991). The above experiments indicate that the structure and composition of storage carbohydrate can be manipulated by gene transfer technology.

In this communication we describe the tuber-specific expression and amyloplast localization of the E. coli glycogen synthase (&A) gene product in transgenic potatoes. This is the initial demonstration of a modified starch with altered branching characteristics created by introduction and expression of a foreign gene product. The resulting starch from these transgenic potato tubers is very highly branched, and this degree of branching correlates with the leve1 of glgA gene expression. The starch also exhibits a lower degree of phosphorylation than normal potato starch and displays unusual viscosity and thermal properties when analyzed by Brabender viscoamylography and differential scanning calorimetry, respectively.

MATERlALS AND METHODS

Cloning of the &A Gene

The gene encoding Escherichia coli glycogen synthase (EC 2.41.21) was obtained from PCR reactions as a 1.4-kb BglII to Sal1 DNA segment. Total genomic DNA from E. coli K12 strain 61 8 was utilized as a template. Synthetic oligonucleo- tides, strl and str2, corresponding to sequences flanking the glgA gene of E. coli (Kumar et al., 1986) were synthesized and contained restriction sites for BglII and SalI, respectively. The strl oligonucleotide contained the sequence AATATk GATCTAACAGGAGCGATAATGCAGGTTTT and corresponds to 24 bp of homology to the 5' end of the glgA gene (Kumar et al., 1986). The str2 oligonucleotide contained the sequence AATAATGCGTCGACTGAAAACTATTTCGAG- CATAGTAA and corresponds to the 3' end of the glgA gene (Kumar et al., 1986). The PCR products were digested with BglII and SalI, respectively, and cloned into BglII/SalI-cut pCGN789 (an ampicillin-resistant pUC-based cloning vector) similar to pUC19. The ligated DNA was transformed into competent E. coli DH5a. The cells were plated onto E. coli Luria broth containing penicillin (300 pg/mL) and isopropyl- thio-8-galactoside and X-Gal; white colonies were picked onto penicillin-containing plates, and the resultant colonies were flooded with 12/KI (0.2% I2 in 0.4% KI). Colonies producing a dark brown color (indicating excess starch production) were selected. DNA was prepared from one clone, designated glgA2, and the entire DNA sequence and translated amino acid sequence were determined. The DNA sequence of glgA2 is 98 and 96% homologous at the nucleotide and amino acid levels, respectively, to the published g1gA gene sequence (Kumar et al., 1986).

Construction of a Tuber-Specific Amyloplast-Targeted Chimeric Gene for &A Expression

Construction of plasmid pCGN2162 containing a 5' patatin-nos 3' cassette for tuber-specific expression has been

described (Oakes et al., 1991). A plasmid pCGN1132S containing the 35s promoter, transit peptide, and 16-ainino acid- coding regions of the pea ribulose bisphosphate carboxylase SSU gene has also been described (Oakes et al., 3 991). The glgA gene was excised from plasmid glgA2 and cloned as a BglII-Sal1 DNA fragment into the BamHI-SalI-ctct plasmid pCGN11,32S. This plasmid was designated pCGN1439 and provided an in-frame fusion of the glgA gene with the 16 amino acids of the mature SSU gene. pCGN1439 was then digested with XbaI and SalI, and the SSU transit peptidel glgA fusion gene was cloned into SpeI-SalI-cut patatin promoter cassette pCGN2162 to create plasmid pCGN1454. This plasmid contained the 700-bp patatin promoter, the transit peptide and mature pea SSU-coding region, the &A gene, and the 300-bp nos 3' region. The entire DNA segment carrying the chimeric gene was excised from pCGh 1454 with XhoI and blunted with Klenow polymerase. This fragment was then cloned into the XbaI-cut and Klenow-blun ted binary vector pCGN1557 (McBride and Summerfelt, 1990) in both orientations to create plasmids pCGN1457 and pCGN1457B. These plasmids were transferred to Agrobacterium i'umefaciens strain LBA4404 as described by Hoekema et al. (1'988).

Plant Transformation

Potato tuber discs (Solanunz tuberosum var Russet Burbank) were transformed with Agrobacterium LBA4404 5,train containing plasmids pCGN1457, pCGN1457B, and pCGN7001 as described previously (Sheerman and Bevan, 1988; Comai et al., 1990; Oakes et al., 1991). Transformants were multi- plied axenically and transferred to soil. Tubers from individual transformants were harvested 14 weeks after transfer to the greenhouse.

Northern and Western Analysis

Total FlNA was isolated from 5 g of tuber íissue and poly(A)' mRNA was processed as described previously (Facciotti et al., 1985). A nick-translated 1.4-kb BglII-Sal1 fragment from plasmid glgA2 was utilized as a probe. Blots were washed at 55OC and autoradiographed. For western analysis, extracts were prepared from 0.3 g of frozen tuber tissue ground in liquid N2.The powder was transferred to Eppendorf tubes to which was added 100 pL of :!X sample buffer (Laemmli, 1970). Samples were thawed 011 ice, vor- texed, and spun at 20008 for 5 min to remove particulate matter. Approximately 15 pL of supematant (20 /ig of total protein) was electrophoresed on SDS-polyacrylamide gels. Proteins were electroblotted onto Millipore Immobilon-NC membranes in buffer containing 0.025 M Tris, 0.2 M Gly, and 20% methanol (pH 8.2). Blots were blocked at rooin temperature for 40 min in 2% BLOTTO/TBST solution. Glycogen synthase polyclonal antiserum (a kind gift from Dr. Jack Preiss) was utilized at a 1/2000 dilution in 2% BLOTTOI TBST and incubated for 90 min. After the sample was washed twice in TBST, goat anti-rabbit antibody (United States Bio- chemical) at a 1/3000 dilution in TBST was applied. Blots were incubated for 40 min, washed twice with TBST, and developetl for 10 min using the Promega Proto Blot system



Nove1 Highly Branched Starch from Transgenic Potato 1161

in buffer containing 0.1 M Tris-HC1 (pH 9.5), 0.1 M NaCI, 0.005 M MgClz, 44 pg/mL nitroblue tetrazolium, and 33 pg/ mL 5-bromo-4-chloro-indolyl phosphate. g1gA protein standards were provided from sonicated extracts of E. coli containing plasmid glgA2. Glycogen synthase expression in these extracts was quantified by SDS-gel electrophoresis (Laemmli, 1970) and protein determinations (Bradford, 1976) of the extracts.

General Analysis of Potato Starch Samples

Specific gravity measurements were determined by weighing tubers in air, followed by a repeat weighing under HzO. Specific gravity equals the weight in air/weight in air minus the weight under H20. Phosphorous content was measured on duplicate samples of 0.5 g of purified dried starch. Phos- phates were converted to P by ashing the starch with a fixative and determining the P content on a dry weight basis as reduced phosphomolybdic acid. Starch and sugar analysis was performed on lyophilized 10-g tuber samples according to the method of Haissig and Dickson (1979). The lyophilized samples were extracted with methanol:chloroform:H20 (12:5:3). Glc in the aqueous phase was detected via a coupled enzyme system using Glc oxidase, peroxidase, and o-dianisi- dine. Suc plus Glc was detected after invertase treatment.

Analysis of Starch Branching

Potato tuber tissue was processed, starch granules were isolated as previously described (Boyer et al., 1976), and starch content was estimated on a weight basis (starch weight/fresh weight). Amylose percentages were determined by gel filtration analysis (Boyer and Liu, 1985). Chain-length distribution patterns were determined by HPLC analysis after starch debranching with isoamylase as described by Sanders et al. (1990). Amylopectin was characterized by measuring the low mo1 wt chain/high mo1 wt chain ratio (on a weight basis) according to the method of Hizukuri (1986).

Brabender Amylograph Analysis

Starch was purified from approximately 2.5 kg of potato tubers. Purified starch was added to 400 mL of distilled HzO to create a 3.25% suspension on a dry weight basis. Pasting characteristics of the suspension were determined according to the method of McComber et al. (1988) using a Brabender Visco-Amylograph VA-1B with a 700 cmg cartridge. With stirring at 75 rpm, the suspension was heated from 30 to 95OC at 1.5OC/min, held at 95OC for 15 min, and then cooled to 5OoC at 1.5OC/min. A plot of the paste viscosity in arbitrary Brabender units versus time was used to determine pasting temperature, peak viscosity, peak viscosity temperature, viscosity prior to cooling, and viscosity at the completion of cooling.

Differential Scanning Calorimetry

Starch samples of about 2 mg were weighed into aluminum pans, and HZO was added to make 30% starch by weight. The pans were sealed and allowed to stand for 1 h before heating. The samples were then heated at 10°C/min from 40

to 100OC. From the temperature curve the gelatinization onset temperature, peak temperature, and enthalpy values were calculated and averaged over three trials.

RESULTS AND DlSCUSSlON

A chimeric gene containing the potato patatin promoter, the soybean SSU transit peptide-coding region, 48 bp of mature SSU from pea, the sequence encoding the E. coli glgA gene, and the nopaline synthase (nos) transcription termina- tion region was introduced into potato via Agrobacterium- mediated transformation. The patatin class I gene has been shown to express in a tuber-specific manner under normal growing conditions (Wenzler et al., 1989; Jefferson et al., 1990), and synthesis of patatin (a major tuber storage protein) has been shown to parallel starch synthesis and accumulation in developing potato tubers (Park, 1983). Plastid targeting, utilizing the SSU transit peptide and mature coding regions, has previously been described for efficient introduction of a number of bacterial proteins to chloroplasts and amyloplasts of transgenic plants (Wasmann et al., 1986; Comai et al., 1988; Klosgen and Weil, 1991; Oakes et al., 1991). A number of transgenic potato lines were evaluated for potential altered starch content by specific gravity measurements. Four of these transgenic lines were chosen for further analysis based on these specific gravity measurements.

Expression of the E. coli glycogen synthase in tubers was monitored by northern and western analysis as indicated in Figure 1. Northern analysis shows that the range of expression for three of the transgenic lines, 57-4, 57-17, and 57B- 15, is approximately 0.0001 to 0.0003% of the total poly(A)' mRNA. The line 57-17 has lower expression than the other two lines. There is no detectable glgA mRNA in nontrans- formed control tissue. The glgA mRNA is also the expected size, approximately 1.9 kb. This RNA encompasses the 1.4- kb glgA structural gene, 0.3 kb encoding the SSU transit peptide/mature protein region and 0.2 kb of nos 3' untrans- lated sequence. This low level of chimeric mRNA production utilizing the patatin promoter in a gene reconstruction format has previously been observed (Oakes et al., 1991).

Western analysis on the four transgenic potato lines revealed the presence of authentic E. coli glycogen synthase when probed with glgA-specific antisera (Fig. 1B). Lines 57- 4 and 57-17 contain glycogen synthase present at approximately 0.007%, and lines 57-18 and 57B-15 contain 0.028% of total tuber protein. The observation that line 57-4 ex- presses a higher level of protein but a lower mRNA level compared to line 57B-15 may reflect a lack of correlation after the glgA gene product is localized in the plastid. It should also be noted that the glgA signal in the transgenic tuber tissue exhibits a slightly higher mo1 wt than the glycogen synthase control protein synthesized in E. coli. The normal E. coli-produced enzyme is 477 amino acids in length, producing a 52,600 mo1 wt polypeptide (Kumar et al., 1986).

In the above described chimeric gene construction, the glgA gene product is longer by 21 amino acids. We do not know how these amino-terminal additions directly affect glgA in plants, although various amino-terminal modifications do not affect the enzyme when expressed in E. coli (data not shown). These additional residues are derived from the 16




s *It in in in of i*"4.4kb

2.4kb

1.4kb

B1 2 3 4 5 6 7 8 9

69 kd

46 kd

Figure 1. A, Northern blot analysis run as described in "Materialsand Methods." Total RNA (20 ^g) was loaded per lane. Control andtransgenic lines are indicated and positions of mol wt standards areshown. Copy number standards were generated utilizing the 1.4-kb DMA fragment encoding the glycogen synthase. B, Western blotanalysis. Samples were prepared and run on 10% Laemmli gels asdescribed in "Materials and Methods." Mol wt standards are indi-cated. Lane 1, RB43 control; lane 2, 7001 transformed control; lane3. RB43 control spiked with 8 ng of E. coli glycogen synthase; lane4. RB43 control spiked with 2 ng of E. coli glycogen synthase; lane5. line 57-4; lane 6, line 57-17; lane 7, line 57-18; lane 8, line 57B-15; lane 9, E. coli extract containing 16 ng of glycogen synthase.

N-terminal amino acids of the pea SSU mature protein fusedto five additional amino acids provided by the 5' PCR primerutilized to clone the g/gA gene from E. coli. When the SSUtransit peptide is cleaved during amyloplast targeting, themature glycogen synthase in the transgenic lines is 498 aminoacids in length, resulting in a protein of approximately 55,000mol wt. This slight increase is apparent in the blot andindicates correct processing of the amyloplast-targeted gly-cogen synthase. The extraction method utilized for the west-ern analysis procedure would preferentially remove solubleamyloplast proteins but should not remove proteins such asGBSS, which are tightly bound to the starch. Immunoblotanalysis indicates that some, if not all, of the gfgA-encodedprotein is located in the soluble fraction of the amyloplast.This observation may help to explain the alteration of thestarch in transgenic tubers.

The four transgenic potato tuber lines exhibit specific grav-ity measurements considerably lower than the Russet Bur-bank control tubers (Table I). These specific gravity valuescorrelate with the dramatic increase in soluble sugar contentin the transgenic tubers. This 60 to 80% increase in solublesugar content suggests unincorporated Sue and Glc mole-cules. These values also correlate quite nicely with the de-creased starch content found in the transgenic tuber lines.

Normal potato tubers are usually 17 to 20% starch by weight(Swinkels, 1985). Lines 57-4, 57-18, and 57B-15 are reducedapproximately 50% for their total starch content. Line 57-17is reduced approximately 20% in total starch content. Starchisolated from these transgenic tubers exhibits a reduced mois-ture content (data not shown), and the dried starch is quite abit lower with respect to the endogenous P content.

Potato starch is unique in that it is the only commercialstarch containing appreciable amounts of bound phosphateesters. The majority of the phosphate groups are bound tothe carbon-6 of Glc units of amylopectin and range from onephosphate group per 200 to 400 Glc units. This phosphateconfers on potato amylopectin polyelectrolyte propertieswhen dispersed in aqueous solutions (Swinkels, 1985). Thereduced phosphate content is probably indicative of the loweroverall starch content in and the reduced chain length of thestarches isolated from transgenic tubers, because phosphatemolecules normally link to longer Glc chains. Other factorsmay be relevant, because phosphate incorporation into potatostarch is poorly understood. In any event, tubers expressingE. coli glycogen synthase in their amyloplasts result in anoverall reduced starch yield with corresponding effects onother intrinsic properties.

The most striking effect of g/gA gene expression in trans-genic tubers was on the amylose/amylopectin ratios and thedegree of amylopectin branching. The data presented in TableII indicate that tubers from transgenic plants contain a re-duced amylose content and a significant change in the amy-lopectin portion of the starch. Normal potato starch containsapproximately 20 to 24% amylose, whereas three of thetransgenic lines exhibit reduced values between 12 and 8%amylose content. Lines 57-18 and 57B-15 also exhibit thelowest amylose values, which correlate with the highest levelsof g/gA gene expression. These two lines also contain a highdegree of branching in their amylopectin fractions.

The average chain length for the total amylopectin fractionsof lines 57-18 and 57B-15 was 26 Glc residues compared toaverage chain lengths of 32, 28, and 33 Glc residues for lines57-17, 57-4, and the RB-43 control, respectively. Normalpotato amylopectin contains low mol wt chains (between 10and 40 Glc residues in length) in a 2:1 ratio with high molwt chains (between 40 and 150 Glc residues). This ratioreflects the degree of branching of the amylopectin fractionof starch (Hizukuri, 1985; Manners, 1985). Thus, the differ-ence in the chain lengths of amylopectin fractions from

Table I. General analysis of potato samplesP was measured on dry starch samples as described in "Materials

and Methods."

Potato Line

RB 43 control57-457-1757-1857B-15

a Soluble sugars

SpecificGravity

1.0811.0601.0691.0601.053

Total Sugars

mg/g dry wt

11.527.518.028.528.0

Starch

%

17.111.014.611.89.0

P

ppm

842574nt

559599

were measured as Sue plus Clc.



Novel Highly Branched Starch from Transgenic Potato 1163

Table II. Analysis of starch branchingThe designations A, B1, B2, and B3 refer to individual carbohy-

drate chains within the amylopectin molecules as revealed byenzymic digestion of starch and HPLC fractionation. These desig-nations conform to the "cluster model" for amylopectin (Hizukuri,1986). Fractions A and B1 to B3 are the A and B chains that bind atthe carbon-6 of the other chains through their reducing residues(A-chains carry no chains and B-chains carry A- or other B-chains).The chains in fractions A and B1 make a single cluster, and thechains in fractions B2 and B3 extend into two and three clusters,respectively. Representative average chain lengths for fractions B1,B2, and B3 are in the ranges 20 to 24, 42 to 48, and 69 to 75,respectively. Average lengths for A-chains are 12 to 16 residues.

Potato Line

RB 43 control57-457-1757-1857B-15

PercentAmylose

23122489

Chains HPLCB2 -1- B3%

Longer Chains

3320281515

Chains HPLCA + Bt%

Shorter Chains

6680728585

RatioA + B1%B2 + B3%

2.04.02.55.75.7

transgenic plants was not due to changes in size of the lowmol wt and high mol wt chains but the relative ratio of thesechains in the starch molecules.

Lines 57-18 and 57B-15, the highest glgA expressors, ex-hibit a very high degree of branching with a low mol wt/high mol wt chain ratio of 5.7. The most highly branchedamylopectin fraction that has previously been described isfrom waxy maize lines with a low mol wt/high mol wt chainratio of 3.9 (Moore et al., 1984; Hizukuri, 1985). Extensivebranching could contribute to the differences in other param-eters such as reduced tuber-specific gravity and starch yieldand low phosphate content. The increase in total soluble

sugar content indicates a pool of unincorporated sugars inthe tubers, which may result from incomplete granule syn-thesis as a result of extensive branching.

This observation is also reflected in the starch distributionwithin the tuber as shown by taking transgenic tubers andsoaking the sliced tubers in a solution of I2/KI (Fig. 2). Thecontrol RB43 and transformed control tubers (7001) displayeda complete iodine staining pattern that was very dark blue,indicating binding of iodine to amylose. Transgenic tubersfrom lines 57-18, 57-4, and 57B-15 displayed a lighter mar-beled staining pattern that is slightly purple and/or brownish.The intensity of blueness and distribution of the color reflecta reduced amylose content. The marbled pattern could resultfrom changes in amylose/amylopectin ratios, distribution ofglycogen synthase as determined by the spatial pattern of thepatatin promoter, reduced starch yield/tuber, or a combina-tion of these changes. Starch granules from normal tubertissue are usually large granules of two types: round granuleswith diameters of 20 to 30 ^m and oval granules with lengthsfrom 20 to 80 nm. The distribution of these two types ofgranules is normally in a 50:50 ratio. The most strikingdifference that was noticed when comparing granules pre-pared from transgenic and control plants was the size of thegranules. Starches prepared from transgenic plants weresmaller: round granules of less than 10 nm in diameter andoval granules with lengths up to 50 fim (data not shown). Inaddition, the round granules made up more than 80% of thesample. Some of the oval granules were also unusuallyshaped with single angular points or irregular surfaces. Thesegranules are somewhat similar to granules observed in cerealswith a high percentage of amylopectin (Swinkels, 1985).

The question arises as to why glycogen synthase, a bacterialchain-extending enzyme functionally analogous to plantstarch synthase, results in a highly branched starch whenexpressed in tuber amyloplasts. Antisense RNA inhibition of

RB 43L. 1 4 5 7 - 1 8 7 0 0 1 1 4 5 7 - 4 1 4 5 7 B - 1 5

Figure 2. Potato tubers stained with iodine. Representative greenhouse-grown tubers for the five lines indicated weresliced in half and soaked with a solution of 0.2% I2 in 0.4% Kl at room temperature for 10 min. The tubers were washedin distilled H2O, and photographs were taken.




endogenous potato GBSS can be ruled out because of the fact that the bacterial glgA and potato GBSS genes are approximately 15% homologous at the nucleotide leve1 (van der Leij et al., 1991b). Additionally, transgenic potato plants in which GBSS was inhibited by an antisense mechanism show no decrease in overall dry matter content (specific gravity) with respect to control plants (Kuipers et al., 1992).

This is not the case with the &A-expressing transgenic potato plants described here. Glycogen synthase may act as a soluble synthetic enzyme that quickly extends the a-1,6 branch points of the developing starch molecule (i.e. an increase of the soluble starch synthase pool). This is sup- ported by the observation that glycogen synthase may exist as a soluble enzyme as detected by westem blots. AItema- tively, glycogen synthase may intercede in biochemical com- plexes of starch synthase and starch branching enzyme required for starch granule synthesis, thereby altering the elongation/branching ratio and resulting in a highly branched starch and reduced yield. It is also possible, although highly unlikely, that E. coli glycogen synthase acts as a branching enzyme when expressed in a plant plastid background. What- ever the mechanism, it is clear that expression of glgA in potato tubers results in a highly branched nove1 polysaccha- ride with unique properties.

The purified starches from the four transgenic lines were subjected to further analysis by Brabender viscoamylography and differential scanning calorimetry. When granules are heated in H20, initial swelling takes place in amorphous regions of the granule, disrupting weak bonding between starch molecules during the hydration. Brabender analysis is a method for following the swelling and viscosity changes during cooking of a starch paste, and viscosity curves are characteristic for different types of starches. The temperature at which measurable viscosity changes occur is the initial pasting temperature. The peak viscosity is the highest viscosity encountered during starch paste preparation and is a measure of starch thickening power, and the cooling part of the viscosity curve is a measure of starch paste retrogradation due to the association of starch molecules. The initial pasting temperatures for starch pastes derived from control and transgenic potato starches were relatively similar, although the initial temperature was significantly higher for sample 57-18 (Table 111). Paste viscosity measurements at a11 phases of the pasting process are significantly lower and more stable

Table IV. Differential scanning calorimetry parameters 01' potato star'ch

All values are means f SD, n = 3. Enthalpy values were corrected for dry stalrch values by taking into account the moisture content of the wet starch values.

Maximum Temp

Starch Sample Onset Temp Erithalpy

"C "C ca l k RB43 control 63.62 f 0.20 66.89 f 0.15 3.9;' -+ 0.12 57-4 59.67 f 0.28 64.77 k 0.23 3.64 0.08 57-1 7 60.61 f 0.62 64.09 f 0.68 3.84 f 0.26 57-18 54.49 f 0.26 59.46 f 0.34 3.111 f 0.12 57B-15 54.41 f 0.40 61.41 f 0.45 2.8;! f 0.05

with time for the three transgenic lines (57-4, 57-18, and 57B-15) measured.

A signifcant and unusual observation was that no peak viscosities or temperatures could be determined for the starch samples from transgenic plants. The amylograms for these samples slhowed only a slow rise after initial pasting. These results could imply very little granule swelling and thickening of the starch paste or rapid swelling and granule clisintegra- tion. The reduced phosphate and amylose content and the increased branching of the amylopectin could jointly contribute to the lower paste viscosities. These highly branched potato starches resemble cereal starches, such as vrheat and rice, with respect to their viscoamylograms and also behave in a fashion similar to certain chemically cross-linked indus- trial starches.

Gelatinization, the process applied to the loss of polariza- tion and concurrent initiation of granule swelling, of the starch can also be monitored by differential scanning calorimetry, and the results are shown in Table IV. Gelatinization began at sxgnificantly lower temperatures, with respect to the control value, for a11 four transgenic samples. Starches from lines 57-18 and 57B-15, which express glycogen synthase at 0.028% of total protein, exhibited the lowest initial gelatinization temperatures and the lowest enthalpy values. Ther- mograms (not shown) were broadest for these two samples, and the gelatinization peaks were asymmetrical, tailing off in the high-tlemperature region. These starch granules may be composed of a more heterogenous granule population than

Table 111. Viscosity analysis of potato starches Temperature for the peak viscosity measurement for the RB43 control sample was 95°C. No peak

viscosities or temperatures could be determined for the three samples prepared from transgenic plants.

Temp Viscosity Viscosity Viscosity at Start at End at Start at End Peak

of of of of Viscosity' Starch Sample

Pasting Heating" Cooling" Cooling"

RB43 control 68.2 550 870 970 990 57-4 69.4 230 270 375 - 57-18 74.5 165 215 315 - 578-15 69.4 240 290 400 -

a Brabender units.



Novel Highly Branched Starch from Transgenic Potato 1165

for normal starch, with most of the granules gelatinizing at lower temperatures than normal starch. Alternatively, the granule population may be homogenous, but the individual granules may undergo less cooperative gelatinization. The magnitude of these differences between control samples and lines 57-18 and 57B-15 is on the order of comparison between wx and aewx maize amylopectins (Yuan et al., 1993). In any event, the differences between control and transgenic starch differential scanning calorimetry values are consistent with the Brabender values obtained from the viscoamylograms on the starches obtained from the four transgenic lines. These values in turn are consistent with the degree of amylopectin branching and ultimately with the level of glgA gene expression in the independent transgenic potato lines.

This is the first description of gross changes in amylopectin branching and altered starch granule morphology in plants as a result of the introduction and expression of a heterolo- gous, albeit functionally closely related, bacterial gene. It is possible that genetic engineering will be able to produce “natural starches” with properties analogous to chemically modified starches now in use. It will be interesting to study further the properties and potential commercial utility of this unique starch and to see what other modifications can be made to reserve starch in plants by introducing genes encoding enzymes that provide more diverse activities (different glycosidic linkages, altemative carbohydrate polymers) than glycogen synthase.

ACKNOWLEDCMENTS

The authors wish to thank Ray Sheehy and Lori Malyi for plasmid constructions, Dr. Jack Preiss for the glycogen synthase antiserum, and Kristin Summerfelt for potato transformations.

Received November 16, 1993; accepted January 16, 1994. Copyright Clearance Center: 0032-0889/94/104/1159/08.

LITERATURE CITED

Blanshard JMV (1987) Starch granule structure and function: a physicochemical approach. In T Galliard, ed, Starch: Properties and Potential. John Wdey & Sons, New York, pp 17-53

Boyer CD, Liu KC (1985) The interaction of endosperm genotype and genetic background. I. Differences in chromatographic profiles of starches from nonmutant and mutant endosperms. Starch Starke 37: 73-79

Boyer CD, Shannon JC, Garwood DL, Creech RG (1976) Changes in starch granule size and amylose percentage during kemel development in severa1 Zea mays L. genotypes. Cereal Chem 53: 3 2 7-33 7

Bradford MM (1976) A rapid and sensitive method for the quanti- tation of microgram quantities for protein using the principal of protein-dye binding. Ana1 Biochem 72: 248-253

Comai L, Larson-Kelly N, Kiser J, Mau CJD, Pokalsky AR, Shew- maker CK, McBride K, Jones A, Stalker DM (1988) Chloroplast transport of a ribulose bisphosphate carboxylase small subunit-5- enolpyruvyl 3-phosphoshikimate synthase chimeric protein re- quires part of the mature small subunit in addition to the transit peptide. J Biol Chem 263 15104-15109

Comai L, Moran P, Maslyar D (1990) Novel and useful properties of a chimeric plant promoter combining CaMV 35s and MAS elements. Plant Mo1 Biol 15 373-381

Duxbury DD (1989) Modified starch functionalities-no chemicals or enzymes. Food Proc 50 35-37

Facciotti D, O’Neal JK, Lee S, Shewmaker CK (1985) Light-induc-

ible expression of a chimeric gene in soybean tissue transformed with Agrobacterium. Biotechnology 3: 241-246

Haissig BE, Dickson RE (1979) Starch measurement in plant tissue using enzymatic hydrolysis. Physiol Plant 47: 151-157

Hizukuri S (1985) Relationships between the distribution of the chain length of amylopectin and the crystalline structure of starch granules. Carbohydr Res 141: 295-306

Hizukuri S (1986) Polymodal distribution of the chain lengths of amylopectins, and its significance. Carbohydr Res 147: 342-350

Hoekema A, Hirsch PR, Hooykaas PJJ, Schilperoot RA (1988) A binary vector based on separation of vir and T-region of the Agrobacterium tumefaciens Ti plasmids. Nature 303 179-181

Hovenkamp-Hermelink JHM, Jacobsen E, Ponsstein AS, Visser RGF, Vos-Scheperkeuter GH, Bijmolt EW, de Vries JN, Witholt B, Feensta WJ (1987) Isolation of an amylose-free starch mutant of the potato (Solanum Tuberosum L.). Theor Appl Genet 75

Jefferson R, Goldsbrough A, Bevan M (1990) Transcriptional regulation of a patatin-1 gene in potato. Plant Mo1 Biol 14 995-1006

Kliisgen RB, Weil JH (1991) Subcellular location and expression level of a chimeric protein consisting of the maize waxy transit peptide and the P-glucuronidase of Escherichia coli in transgenic potato plants. Mo1 Gen Genet 225 297-304

Kuipers GJ, Vreem JTM, Meyer H, Jacobson E, Feenstra WJ, Visser RGF (1992) Field evaluation of antisense RNA mediated inhibition of GBSS gene expression in potato. Euphytica 5 9 83-91

Kumar A, Larsen CE, Preiss J (1986) Biosynthesis of bacterial glycogen. Primary structure of Escherichia coli ADP-glucose a1,4- glucan, 4-glycosyltransferase as deduced from the nucleotide sequence of the glgA gene. J Biol Chem 261: 16256-16259

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685

Manners DJ (1985) Starch. In PM Dey, RA Dixon, eds, Biochemistry of Storage Carbohydrates in Green Plants. Academic Press, Lon- don, pp 149-203

McBride KE, Summerfelt K (1990) Improved binary vectors for Agrobacterium-mediated plant transformation. Plant Mo1 Biol 14

McComber DR, Osman EM, Lohnes RA (1988) Factors related to potato mealiness. J Food Sci 53 1423-1426

Moore CO, Tuschhoff JV, Hastings CW, Schanefelt RV (1984) Applications of starches in foods. In RL Whistler, JN Bemiller, EF Paschall, eds, Starch: Chemistry and Technology, Ed 2. Academic Press, Orlando, FL, pp 575-591

Miiller-Rober B, Sonnewald U, Willmitzer L (1992) Inhibition of the ADP-glucose pyrophosphorylase in transgenic potatoes leads to sugar-storing tubers and influences tuber formation and expression of tuber storage protein genes. EMBO J 11: 1229-1238

Oakes JV, Shewmaker CK, Stalker DM (1991) Production of cyclodextrins, a nove1 carbohydrate, in the tubers of transgenic potato plants. Biotechnology 9 982-986

Park WD (1983) Tuber proteins of potato-a new and surprising molecular system. Plant Mo1 Biol Rep 1: 61-66

Sanders EB, Thompson DB, Boyer CD (1990) Thermal behavior during gelatinization and amylopectin fine structure for selected maize genotypes as expressed in four inbred lines. Cereal Chem

Shannon JC, Garwood DL (1984) Genetics and physiology of starch development. In RL Whistler, JN Bermiller, EF Paschall, eds, Starch: Chemistry and Technology, Ed 2. Academic Press, Or- lando, FL, pp 25-86

Sheerman S, Bevan MW (1988) A rapid transformation method for Solanum tuberosum using binary Agrobacterium tumefaciens vectors. Plant Cell Rep 7: 13-16

Smith AM (1988) Major differences in isoforms of starch-branching enzyme between developing embryos of round- and wrinkled- seeded peas (Pisum sativum L). Planta 175 270-279

Stark DM, Timmerman KP, Barry GF, Preiss J, Kishore GM (1992) Regulation of the amount of starch in plant tissues by ADP-glucose pyrophosphorylase. Science 258: 287-292

Swinkels JJM (1985) Sources of starch, its chemistry and physics. In GMA Van Beynum, JA Roels, eds, Starch Conversion Technology. Marcel Dekker, New York, pp 15-46

217-221

269-276

67: 594-602




van der Leij FR, Visser RGF, Oosterhaven K, van der Kop DAM, Jacobsen E, Feenstra WJ (1991a) Complementation of the amylose-free starch mutant of potato (Solanum tuberosum) by the gene encoding granule-bound starch synthase. Theor Appl Genet 82:

van der Leij FR, Visser RGF, Ponstein AS, Jacobsen E, Feenstra WJ (1991b) Sequence of the structural gene for granule-bound starch synthase of potato (Solanum tuberosum L.) and evidence for a single point deletion in the amf allele. Mo1 Gen Genet 228:

Visser RGF, Somhorst I, Kuipers GJ, Ruys NJ, Feenstra WJ, Jacob- sen E (1991) Inhibition of the expression of the gene for granule-

289-295

240-248

bound starch synthase in potato by antisense constructs. Mo1 Gen Genet 22!5 289-296

Wasmann CC, Reiss B, Bartlett SG, Bohnert HJ (1986) ' f i e impor- tance of ithe transit peptides and the transported protein for protein import irito chloroplasts. Mo1 Gen Genet 205 446-453

Wenzler HIC, Mignery GA, Fisher LM, Park WD (198'3) Analysis of a chinieric class-I patatin-GUS gene in transgenic potato plants: high leve1 expression in tubers and sucrose-inducible expression in cultured leaf and stem explants. Plant Mo1 Biol12: 41-50

Yuan RC, Thompson DB, Boyer CD (1993) Fine structcire of amylopectin in relation to gelatinization and retrogradation behavior of maize starches from three wx-containing genotypes in two inbred h e s . Cereal Chem 70 81-89



expression of escherichia coli clycogen synthase in the ... · plant physiol. (1994) 104: 1159-1166...

Documents