Download Modifying Starch Biosynthesis with Transgenes in

Plant Physiol. (1992) 100, 1083-1086
0032-0889/92/100/1083/04/$01 .00/0
Received for publication June 22, 1992
Accepted June 22, 1992
__~
Modifying Starch Biosynthesis with Transgenes in Potatoes
Christine K. Shewmaker* and David M. Stalker
Calgene, Inc., 1920 Fifth Street, Davis, California 95616
glucan chain (Fig. 1A). Starch that primarily consists of a-i4 glucans with a very few a--1-6 branch points (approximately
1/1000) is termed amylose; starch with many branch points
(approximately 1/25) is termed amylopectin (10; see Fig. 1B
for diagrams of possible structures of amylose and amylopectin). There are several forms of starch synthases and they are
grouped into the two classes of soluble and granule-bound
starch synthases. Evidence indicates that granule-bound
starch synthase is responsible for the synthesis of amylose.
To alter carbohydrate expression in plants, three basic
technologies are needed: (a) plant transformation, (b) organspecific expression, and (c) plastid targeting. These three are
then combined with the gene of interest to yield transgenic
plants expressing the desired trait. The gene of interest may
be of bacterial, animal, or plant origin. Additionally, antisense
technology may be employed if the desired trait would result
from turning off a given endogenous gene (5). Potato was
chosen as the first system in which to alter starch biosynthesis
because all three of the above technologies were available
and because the potato tuber contains a large reservoir of
reserve starch. Transgenic potato lines can routinely be obtained using Agrobacterium tumefaciens cocultivation with a
variety of tissue sources such as leaf pieces, tuber slices, etc.
(3). Additionally, the technology can be applied to a broad
range of potato varieties.
Organ-specific (reserve starch-producing) expression is desirable so that starch biosynthesis in source tissues such as
leaves will not be disturbed. In potato, the expression of the
tuber-storage protein, patatin, has been shown to parallel
starch accumulation in the tuber (14), and chimeric constructs
using this promoter have demonstrated tuber-specific expression (26). Constitutive promoters (35S) are sometimes used,
particularly with antisense constructs, if expression in other
tissues is tolerable.
The protein product of the gene of interest may need to be
targeted to the amyloplast because this is where the final
steps of starch biosynthesis occur. Amyloplasts are ontogenically related to other plastids such as chloroplasts. Effective
chloroplast targeting has been obtained with a number of
transit peptides including the transit peptide of the nuclearencoded SSU (18). Also, interchangeability of transit peptides
for various plastids has now been demonstrated (8). Plastid
targeting is not necessary with antisense technology or if the
protein of interest is needed in the cytoplasm. In Figure 2,
the three strategies, cytoplasmic expression of a protein (1),
amyloplast targeting of a protein (2), and antisense expression
(3) are diagrammed.
Using these technologies, there are three basic ways in
A large amount of the CO2 fixed by green plants is stored
as reserve starch in plant tissues such as seeds, tubers, and
roots. At present, this starch is used either directly for food
or processed to yield starches with food and industrial uses.
Worldwide, approximately 3 to 5% of the starch crop produced is used for the production of starch. The majority of
this is processed into high fructose syrup or dextrose. Approximately one-third is used for nonfood purposes, such as
in the paper, packaging, and textile industries. Various regional and varietal differences in the utilization of starch
crops exist. For example, one-quarter of the world's potato
starch is produced in The Netherlands. Genetic engineering
of plants in conjunction with molecular biology and a knowledge of starch biosynthesis offers the promise of being able
to use this vast plant resource to produce new or altered
carbohydrates. These carbohydrates could have increased or
unique value for food or industrial purposes. In this review,
we discuss early progress that has been made in this area by
engineering transgenic potatoes (Solanum tuberosum). This
technology should be applicable to other starch crops, includ-
ing corn and rice.
In starch-storing organs such as the potato tuber, starch is
stored in plastids called amyloplasts. The exact pathway for
starch conversion is still open to debate but a general pathway
can be outlined (see refs. 16 and 17 for comprehensive
reviews). Photoassimilate as sucrose is delivered to the tuber
cell via the phloem. Once in the cell, both hexose-phosphates
and triose-phosphates are formed. There is no general agreement as to whether hexose-phosphates such as glucose-6phosphate or glucose-i-phosphate are transported to the
amyloplasts or whether triose-phosphates are transported
into the amyloplast. Once the carbon source is in the amyloplast, the major building block of starch, ADP-Gl is formed
from glucose-1-phosphate and ATP via the enzyme ADP-GPP (EC 2.7.7.27). All plant ADP-G-PPs studied to date have
been shown to be activated by 3-phosphoglycerate and inhibited by inorganic phosphate. Many studies have indicated
that this enzyme is a key control point in starch biosynthesis.
Once the ADP-G is formed, it is added to the growing
starch chain by starch synthase (EC 2.7.7.9). This enzyme
forms linear chains of a-1-4-linked glucopyranose units (Fig.
1A) by adding ADP-G to a growing a-1-4 glucan chain. The
last enzyme in starch biosynthesis, starch-branching enzyme
(EC 2.4.1.11), introduces a-1-6 branch points in the a-1-4
1
Abbreviations: ADP-G, ADP-glucose; ADP-G-PP, ADP-glucosepyrophosphorylase; CD, cyclodextrin; CGT, cyclodextringlycosyl
transferase; SSU, small subunit of ribulosebisphosphate carboxylase.
1083
Downloaded from on June 15, 2017 - Published by www.plantphysiol.org
Copyright © 1992 American Society of Plant Biologists. All rights reserved.
A
H20H
A
44 H2OH
209
1,6-linkage
H2
CH20H
CH2
CH2 H
1,4-linkage
B
Plant Physiol. Vol. 100, 1992
SHEWMAKER AND STALKER
1 084
Amylose
0
Amylopectin
0
Figure 1. A, Structure of a-1-4 and a-1-6 glucan linkages in starch.
B, Diagrammatic representation of possible structures of amylose
and amylopectin.
which to alter carbohydrate metabolism. These include (a)
production of novel carbohydrates, (b) production of starches
with altered amylose/amylopectin ratios and chain lengths,
and (c) production of more starch (increased yield).
range of compounds. As the physical chemical characteristics
of CDs, such as pore size and solubility, vary with the type
of CD, different forms of CDs form complexes with different
compounds. The formation of inclusion complexes can
change both the physical and chemical properties of the guest
compounds. These effects include increased stability due to
protection against oxidation, light, heat, etc., and enhanced
solubility. As such, CDs have a broad range of uses including
delivery systems for pharmaceuticals and agrochemicals; stabilization of flavors, odors, and other components in food;
and removal of given compounds from a complex mixture
(15). The proposed uses for CDs have risen exponentially in
the past few years. Recently, it has been suggested that CDs
or their derivatives could be used for removal of undesirable
compounds from food such as caffeine and cholesterol.
CDs are formed by the action of CGTs on preformed starch
(Fig. 3). CGTs have only been found in bacteria, specifically
Bacillus and Klebsiella. The products of various CGTs vary in
the ratio of a:f:'y CDs formed. For expression of CDs in
potato tubers, a chimeric gene consisting of 5'-patatin promoter-SSU leader-CGT gene from Klebsiella pneumoniaenos-3' was transferred to potato via A. tumefaciens (13).
Small but detectable amounts of CDs were seen in the potato
tuber and the ratio of the a:/3 CDs was similar to that
produced by this particular enzyme in vitro. The expression,
as determined by northern analysis, was seen only in the
tuber. Amyloplast targeting was inferred because there is no
starch present for the enzyme to act upon in the cytoplasm.
The level of CDs produced in these tubers was low and needs
to be improved for broad commercial usage. However, the
three technologies discussed above, potato transformation,
tuber-specific expression, and plastid targeting came together
to produce a novel carbohydrate from the gene of interest.
AMYLOSE/AMYLOPECTIN
The starch that is deposited in amyloplasts varies with the
species of plants as to the size of the granules, amylose/
amylopectin ratio, and the nature of contaminating compounds such as water and lipids (23). All of these factors can
NOVEL CARBOHYDRATES
Novel carbohydrates are carbohydrates not normally produced in the transgenic starch organ, i.e. potato tubers. Novel
carbohydrates could be produced by the action of an enzyme
on preformed starch, by production of alternate polymers to
a-1-4 and a-1-6 glucans through the introduction of one or
more new synthetic enzymes, or by the production of alternate carbohydrate monomers or dimers, etc., that could be
used as is or as precursors for chemical processes. Microbial
germplasm should be a good source of some of these genes
since bacteria produce a much wider variety of carbohydrates
than is found in plants.
As a first demonstration of the possibility of engineering
potatoes to produce novel carbohydrates, we engineered
potato tubers to produce CDs (13), which are cyclic oligosaccarides containing six (a), seven (,B), and eight ('y) a-1-4linked glucopyranose units (24). The exterior of the cyclic
molecule is hydrophilic and the interior is apolar. These
molecules, therefore, form inclusion complexes with a broad
X
0
Figure 2. Strategies for altering starch biosynthesis. 1, Cytoplasmic
expression of a protein; 2, amyloplast targeting of a protein; 3,
antisense inhibition of expression.
Downloaded from on June 15, 2017 - Published by www.plantphysiol.org
Copyright © 1992 American Society of Plant Biologists. All rights reserved.
STARCH MODIFICATION IN POTATOES
CGT|
Figure 3. Diagrammatic representation of the action of CGTs
starch.
on
affect the final properties of the starch. The granules vary in
size from 2 to 100 ,m, with potato typically having large
granules. The starch itself, as discussed above, is composed
of amylose and amylopectin. The amylopectin is responsible
for the crystalline nature of the starch, and minor differences
in the chain length within the amylopectin can affect the
ultimate structure of the starch (2). When heated, starches
will go into solution. Amylose readily forms hydrogen bonds
in solution and thus forms a rigid gel, whereas amylopectin,
due to its branched structure, has limited hydrogen bonding
and remains more fluid. Thus, amylopectin is much less prone
to retrogradation or shrinking after heating and cooling than
amylose. Beyond the range of functionalities that is imparted
by the natural variation in starch, additional characteristics
can be obtained by chemical modification and cross-linking.
Natural mutations that affect amylose/amylopectin ratios
have been found in corn, rice, etc. (19). A well-studied
example is the waxy (wx) mutant of maize. This mutant lacks
granule-bound starch synthase and the starch from this mutant corn is almost 100% amylopectin. Waxy corn starch has
a heavy viscosity and a higher elasticity (11) than normal
com starch, which makes it useful in thickeners, etc. Recently,
it has been demonstrated that a double starch mutant in corn,
duwx, has a slightly more branched polymer than wx corn
starch, and this imparts characteristics to the starch that make
it similar to a chemically modified starch (4). Thus, the ability
to alter amylose/amylopectin ratios via genetic engineering
may offer tailor-made starches (7) or starches that mimic
chemically modified starches.
1085
No naturally occurring amylose free (amf) mutations (waxy
type) exist in potato; however, mutagenesis followed by
laborious screening of haploid lines has produced amf potatoes (6). A genetic engineering approach using antisense
technology has also produced potatoes that are amylose free
(25). In this experiment, a granule-bound starch synthase
cDNA from potato was expressed in an antisense orientation
from a constitutive promoter (35S). Lines that yielded tubers
with no amylose and lines with amounts of amylose intermediate between the wild type and the amf lines were found.
This demonstrates part of the power of genetic engineering
for starch biosynthesis in that a range of possibly useful
phenotypes can be formed from one chimeric construct.
Additionally, the construct can then easily be transferred to
many varieties of potato.
In addition to natural mutations with high amylopectin,
mutants have been found in a variety of species such as corn
and peas that have an increased amount of amylose relative
to amylopectin. In the embryos of wrinkled peas, the percentage of amylopectin has dropped from 60 to 70% to only
30% and this relative decrease in amylopectin has been
associated with decreased branching activity and loss of one
isoform of the branching enzyme (20). With the cloning of a
cDNA for a branching enzyme from potato, it is possible that
an antisense strategy with this cDNA may lead to altered
amylose/amylopectin ratio (9).
It is not necessary to restrict experiments to antisensing
endogenous genes to alter amylose/amylopectin ratios. It
should be possible to overexpress either plant or bacterial
starch-synthesizing genes in the manner depicted in Figure
2, strategies 1 and 2, and thus alter the amylose/amylopectin
ratio. It is, for example, possible that the bacterial glycogen
synthase gene, g1gA, when targeted to the amyloplast, might
alter the ratio of synthesizing/branching enzyme and thus
alter the amylose/amylopectin ratio. As with novel carbohydrates, bacterial genes offer a good source of variety because
they may have properties that are slightly different from
their plant counterparts and thus create different or novel
phenotypes.
YIELD
Another way to alter starch biosynthesis is to have the
starch-accumulating organ accumulate more starch. A variety
of strategies exists for altering flow of photosynthate, which
may alter source-sink relationships and thus affect yield (see
ref. 21). Within the three reactions that are involved in starch
biosynthesis per se, the enzyme ADP-G-PP has been postulated to be a key control point (16, 17). This has recently
been confirmed using transgenic potatoes. When a cDNA
encoding one of the subunits of ADP-G-PP was expressed in
the antisense orientation from the 35S promoter, dramatically
lower levels of ADP-G-PP were seen in the tubers. This lower
level of ADP-G-PP correlated with very decreased starch and
substantially increased sugars in the tubers (12). Thus, it
seems reasonable that increasing activity levels of ADP-GPP might increase starch accumulation. This might be accomplished with either a plant or bacterial gene. The ADP-G-PP
from Escherichia coli (termed glgC) has been well characterized and consists of only one subunit, as opposed to two for
Downloaded from on June 15, 2017 - Published by www.plantphysiol.org
Copyright © 1992 American Society of Plant Biologists. All rights reserved.
SHEWMAKER AND STALKER
1 086
plants. Unlike the plant enzyme that is subject to activation
by 3-phosphoglycerate and inhibition by Pi, the bacterial
enzyme is subject to activation by fructose-1,6-bisphosphate
(1). Additionally, a mutant E. coli glgC is known that is less
sensitive to fructose-1,6-bisphosphate activation. Recently,
expression of this glgC in potato from a patatin promoter and
SSU leader has been reported to lead to increased starch
accumulation in tubers (22).
In conclusion, the above studies show that it is indeed
possible to modify starch biosynthesis in potatoes. Novel
carbohydrates, starches with altered amylose/amylopectin
ratios, and increased starch have been produced in potatoes
using transgenes. Although the traits may need to be optimized for commercialization, the first steps in using genetic
engineering to broaden the range of carbohydrates made in
potatoes have clearly been successful. The range of traits and
products produced also can be clearly extended by using the
wide variety of genes available from both prokaryotic and
eukaryotic germplasm and/or by introducing two or more
expression chimeras in one potato line. Additionally, the
technologies to transfer these traits to other starch-producing
crops such as corn, wheat, and rice are being developed. In
the next 10 to 20 years, the technologies described herein
should yield a broad range of natural carbohydrate products
that are produced in plants.
10.
11.
12.
13.
14.
15.
16.
17.
18.
LITERATURE CITED
1. Anderson JM, Okita TW, Preiss J (1990) Enhancing carbon
flow into starch: the role of ADPglucose pyrophosphorylase.
In ME Vayda, WD Park, eds, The Molecular and Cellular
Biology of the Potato. CAB International, Wallingford, UK, pp
159-180
2. Blanshard JMV (1987) Starch granule structure and function: a
physicochemical approach. In T Galliard, ed, Starch: Properties and Potential. John Wiley & Sons, New York, pp 17-53
3. DeBlock M (1988) Genotype-independent leaf disc transformation of potato (Solanum tuberosum) using Agrobacterium
tumefaciens. Theor Appl Genet 76: 767-774
4. Duxbury DD (1989) Modified starch functionalities-no chemicals or enzymes. Food Processing 50: 35-37
5. Hiatt WR, Kramer M, Sheehy RE (1989) The application of
antisense RNA technology to plants. In JK Setlow, A Hollaender, eds, Genetic Engineering: Principles and Methods, Vol 1.
Plenum Press, New York, pp 49-63
6. 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: 217-221
7. Kennedy JF, Cabalda VM, White CA (1988) Enzymic starch
utilization and genetic engineering. Tibtech 6: 184-189
8. Klosgen RB, Weil JH (1991) Subcellular location and expression
level of a chimeric protein consisting of the maize waxy transit
peptide and the ,B-glucuronidase of Escherichia coli in transgenic potato plants. Mol Gen Genet 225: 297-304
9. Kossmann J, Visser RGF, Muller-Rober B, Willmitzer L, Son-
19.
20.
21.
22.
23.
24.
25.
26.
Plant Physiol. Vol. 100, 1992
newald U (1991) Cloning and expression analysis of a potato
cDNA that encodes branching enzyme: evidence for coexpression of starch biosynthetic genes. Mol Gen Genet 230:
39-44
Manners DJ (1985) Starch. In PM Dey, RA Dixon, eds, Biochemistry of Storage Carbohydrates in Green Plants. Academic
Press, London, pp 149-203
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
Muller-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: (in
press)
Oakes JV, Shewmaker CK, Stalker DM (1991) Production of
cyclodextrins, a novel carbohydrate, in the tubers of transgenic
potato plants. Bio/Technology 9: 982-986
Park WD (1983) Tuber proteins of potato-a new and surprising
molecular system. Plant Mol Biol Rep 1: 61-66
Parrish MA (1987) Cyclodextrins-a review. Speciality Chemicals 11: 366-380
Preiss J (1988) Biosynthesis of starch and its regulation. In The
Biochemistry of Plants, Vol 14, Academic Press, Orlando, FL,
pp 181-254
Preiss J (1991) Biology and molecular biology of starch synthesis
and its regulation. In BJ Miflin, HF Miflin, eds, Oxford Surveys
of Plant Molecular and Cell Biology, Vol 7. Oxford University
Press, Oxford, UK, pp 59-114
Schreier PH, Seftor EA, Schell J, Bohnert HJ (1985) The use of
nuclear-encoded sequences to direct the light-regulated synthesis and transport of a foreign protein into plant chloroplasts.
EMBO J 4: 25-32
Shannon JC, Garwood DL (1984) Genetics and physiology of
starch development. In RL Whistler, JN Bemiller, EF Paschall,
eds, Starch: Chemistry and Technology, Ed 2. Academic Press,
Orlando, FL, pp 25-86
Smith AM (1988) Major differences in isoforms of starchbranching enzyme between developing embryos of roundand wrinkled-seeded peas (Pisum sativum L.). Planta 175:
270-279
Sonnewald U, Willmitzer L (1992) Molecular approaches to
sink-source interactions. Plant Physiol 99: 1267-1270
Stark DM, Barry GF, Muskopf YM, Timmerman KP, Kishore
GM (1991) Increased starch and dry matter deposition in
trangenic Russet Burbank potato tubers (poster abstract No.
714). Third International Congress of Plant Molecular Biology:
Molecular Biology of Plant Growth and Development
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
Szejtli J (1988) Cyclodextrin Technology, Kluwer Academic
Publishers, Dordrecht, The Netherlands
Visser RGF, Somhorst I, Kuipers GJ, Ruys NJ, Feenstra WJ,
Jacobsen E (1991) Inhibition of the expression of the gene for
granule-bound starch synthase in potato by antisense constructs. Mol Gen Genet 225: 289-296
Wenzler HC, Mignery GA, Fisher LM, Park WD (1989) Analysis of a chimeric class-I patatin-GUS gene in transgenic potato
plants: high level expression in tubers and sucrose-inducible
expression in cultured leaf and stem explants. Plant Mol Biol
12: 41-50
Downloaded from on June 15, 2017 - Published by www.plantphysiol.org
Copyright © 1992 American Society of Plant Biologists. All rights reserved.