recent discoveries suggest distinct pathways and novel mechanisms

CSIRO PUBLISHING
www.publish.csiro.au/journals/fpb Functional Plant Biology, 2007, 34, 465–473
Review:
Starch breakdown: recent discoveries suggest distinct pathways
and novel mechanisms
Samuel C. Zeeman A,B , Thierry Delatte A ,Gaëlle Messerli A , Martin Umhang A , Michaela Stettler A ,
Tabea Mettler A , Sebastian Streb A , Heike Reinhold A and Oliver Kötting A
A Institute of Plant Sciences, ETH Zurich, Universitätstrasse 2, CH-8092 Zurich, Switzerland.
B Corresponding author. Email: szeeman@ethz.ch
This paper originates from a presentation at the 8th International Congress of Plant
Molecular Biology, Adelaide, Australia, August 2006.
Abstract. The aim of this article is to provide an overview of current models of starch breakdown in leaves. We summarise
the results of our recent work focusing on Arabidopsis, relating them to other work in the field. Early biochemical studies
of starch containing tissues identified numerous enzymes capable of participating in starch degradation. In the non-living
endosperms of germinated cereal seeds, starch breakdown proceeds by the combined actions of α-amylase, limit dextrinase
(debranching enzyme), β-amylase and α-glucosidase. The activities of these enzymes and the regulation of some of the
respective genes on germination have been extensively studied. In living plant cells, additional enzymes are present, such
as α-glucan phosphorylase and disproportionating enzyme, and the major pathway of starch breakdown appears to differ
from that in the cereal endosperm in some important aspects. For example, reverse-genetic studies of Arabidopsis show
that α-amylase and limit-dextrinase play minor roles and are dispensable for starch breakdown in leaves. Current data
also casts doubt on the involvement of α-glucosidase. In contrast, several lines of evidence point towards a major role
for β-amylase in leaves, which functions together with disproportionating enzyme and isoamylase (debranching enzyme)
to produce maltose and glucose. Furthermore, the characterisation of Arabidopsis mutants with elevated leaf starch has
contributed to the discovery of previously unknown proteins and metabolic steps in the pathway. In particular, it is now
apparent that glucan phosphorylation is required for normal rates of starch mobilisation to occur, although a detailed
understanding of this step is still lacking. We use this review to give a background to some of the classical genetic mutants
that have contributed to our current knowledge.
Additional keywords: amylase, amylopectin, Arabidopsis thaliana L., carbohydrate metabolism, cereal endosperm,
metabolic regulation.
Starch is central to Arabidopsis metabolism
Starch and sucrose are the primary products of photosynthesis
in the leaves of most plants. Triose-phosphates exported from
the chloroplast during the day are used to synthesise sucrose
in the cytosol. Sucrose is then loaded into the phloem of the
source tissues (e.g. mature leaves) for long-distance transport to
sink tissues (e.g. roots, developing leaves and storage organs)
where it is used to support respiration and growth. Starch
is synthesised concurrently with sucrose in the leaves. It is
made in the chloroplast and comprises glucose polymers that
are packed to form dense, osmotically inert granules (Buléon
et al. 1998; Zeeman et al. 2002). Arabidopsis plants grown in
standard growth room conditions (e.g. a 12-h photoperiod with
150 µmoles photon m−2 s−1 quantum irradiance, 20◦C, 70%
relative humidity) partition ∼40% of their newly-assimilated
carbon into starch (Zeeman and ap Rees 1999). Leaf starch is
degraded throughout the night. The sugars released are used to
provide substrates for respiration of the leaf and for continued
sucrose synthesis and export (Zeeman and ap Rees 1999).
The division of assimilates between starch and sucrose is
controlled in accordance with the environmental conditions.
Chatterton and Silvius (1980) showed that several plant species
grown under short-day conditions partition relatively more of
their photo-assimilates into starch than those grown under longday
conditions (presumably at the expense of daytime sucrose
production). This has also been shown to be the case for
Arabidopsis (Gibon et al. 2004). Thus, although the total amount
of carbon assimilated during the short day is lower than in a
long day, the amount of starch accumulated during the entire
photoperiod does not decrease to the same extent (Fig. 1A).
A greater proportion of the photo-assimilated carbon is therefore
available to sustain metabolism during the long night that follows
a short day. During the night, starch is broken down at a relatively
constant rate. The rate of breakdown is regulated in a manner
complementary to the rate of synthesis during the day. When
plants are entrained to long nights, the rate of degradation is
slow so that the amount accumulated during the short days is
exhausted only at the end of the night. Conversely, in short
© CSIRO 2007 10.1071/FP06313 1445-4408/07/060465

466 Functional Plant Biology S. C. Zeeman et al.
nights, the rate of degradation is high, but again the reserves
are exhausted only at the end of the night (Fig. 1A; Gibon et al.
2004; Lu et al. 2005).
When grown in a diurnal cycle, mutants of Arabidopsis that
cannot synthesise starch, or mutants that can synthesise starch
but cannot remobilise it, grow more slowly than wild type plants
(Caspar et al. 1985, 1991; Zeeman et al. 1998). Caspar et al.
(1985) showed that the reduction in growth rate of a starchless
line (pgm – lacking plastidial phosphoglucomutase) was more
severe in short days than in long days, whereas in continuous
A
Starch content
B
CO2
8-h photoperiod
Short days: a high %
of assimilated C
is partitioned
into starch
Chloroplast stroma
Cytosol
Long days: a low
% of assimilated C is
partitioned into starch
Calvin
cycle
Triosephosphates
Sucrose
light, its growth was equal to that of the wild type. This illustrates
the important role that starch plays in sustaining metabolism
at night.
The pathway of starch synthesis in leaves (Fig. 1B) begins
with the production of the glucosyl donor, ADPglucose
(ADPGlc), from the Calvin cycle intermediate, fructose-6phosphate.
Subsequent production of amylopectin, the major
component of starch, requires the actions of three classes
of enzymes. Starch synthases elongate existing chains by
transferring the glucosyl moiety of ADPGlc to an existing
16-h photoperiod
Short nights:
starch breakdown
is fast
Long nights:
starch breakdown
is slow
0 4 8 12
Time (h)
16 20 24
PGI
PGM AGPase
SS
BE
DBE
Fru6P Glc6P Glc1P ADPGlc
C
ATP P~P i
s
2 P i
Starch
Fig. 1. Starch biosynthesis in Arabidopsis leaves. (A) A scheme illustrating the pattern of
starch accumulation and utilisation in Arabidopsis leaves during different diurnal regimes.
The black line represents the near-symmetrical pattern observed under 12-h light/12-h dark
conditions. The rates of starch synthesis and starch degradation respond in complementary ways
to short-day (red) or long-day (blue) conditions, as indicated (see also Gibon et al. 2004; Lu et al.
2005). (B) The pathway of starch biosynthesis in the chloroplast stroma. PGI, phosphoglucose
isomerase; PGM, phosphoglucomutase; AGPase, ADPglucose pyrophosphorylase; SS, starch
synthase; BE, branching enzyme; DBE, debranching enzyme; Glc-1-P, glucose-1-phosphate.
(C) Transmission electron micrograph of an Arabidopsis palisade mesophyll cell, showing a
chloroplast containing five starch granules (S). Bar = 1 µm.

Pathways of starch breakdown Functional Plant Biology 467
glucan chain, creating a new α-1,4-bond. Branching enzymes
introduce branch points by cutting an existing α-1,4-linked
chain and transferring the cut portion of the chain (6 or
more glucosyl units in length) to an adjacent chain creating
an α-1,6-bond. Debranching enzymes are also required
for amylopectin synthesis, probably by selectively removing
incorrectly positioned branches. The combined actions of
starch synthases, branching enzymes and debranching enzymes
produce a glucan in which branch points are non-randomly
distributed. Linear segments of neighbouring chains form double
helices that align into crystalline layers, between which there are
non-crystalline layers containing the branch points (Fig. 1B; for
further details, see Zeeman et al. 2007 and references therein).
The resultant starch granule that forms between the thylakoid
membranes in Arabidopsis chloroplasts is an insoluble discoid
structure, 1–2 µm in diameter and 0.2–0.5 µm thick (Fig. 1C).
Although the individual functions of the starch biosynthetic
enzymes are reasonably well described, the way in which their
activities are coordinated to generate the specific branching
pattern of amylopectin and the mechanism of starch granule
initiation are not well understood.
Rethinking the initial steps in starch breakdown
Mobilisation of the starch granule involves the conversion of
the insoluble, semi-crystalline matrix formed by amylopectin
to soluble sugars that can feed into intermediary metabolism.
During starch breakdown in the endosperm of germinated cereal
seeds, α-amylase (an endoamylase) initiates starch degradation
by releasing from the granule a mixture of soluble glucans
that serve as substrates for other hydrolytic enzymes. The
aleurone layer and the scutellum – the living cells surrounding
the endosperm – secrete α-amylase into the non-living, starchcontaining
endosperm. The regulation of α-amylase gene
expression has been extensively studied in the aleurone of barley
(Fincher 1989; Gubler et al. 1995, 1999), and repression of
α-amylase gene expression in rice using RNAi has recently been
shown to delay germination (Asatsuma et al. 2005).
Investigation of α-amylase function in Arabidopsis indicates
that this enzyme has only a minor role in starch metabolism.
The Arabidopsis genome encodes three putative α-amylases
(AMY1–AMY3; Stanley et al. 2002). Arabidopsis seeds store
oil rather than starch, so there is no role for α-amylase in
seedling establishment analogous to that in cereals. However, it
was reasonable to suppose that α-amylase might be involved in
leaf starch degradation. Indeed, consistent with this supposition,
initial biochemical studies of a high starch mutant, sex4
(for starch-excess), identified a deficiency in a chloroplasttargeted
isoform of α-amylase (Zeeman et al. 1998). However,
although subsequent studies have confirmed the deficiency in
the chloroplast α-amylase (AMY3) in sex4, they have also
shown that the SEX4 gene does not encode α-amylase (Yu
et al. 2005; Niittylä et al. 2006) and that α-amylase is not
required for starch degradation. Yu et al. (2005) reported that
mutants lacking AMY3 have normal rates of starch degradation.
This was also the case for triple mutants in which all three
α-amylase genes were disrupted. Thus, there is another
mechanism for the initiation of starch degradation in leaves.
Nevertheless, α-amylase isoforms homologous to AMY3 are
widely conserved in the plant kingdom and in some situations
endoamylolysis might contribute to starch mobilisation (see
below and Delatte et al. 2006).
Glucan phosphorylation –akeystep in starch
breakdown
There is a strong body of evidence indicating that the process of
starch breakdown inside plastids requires the action of glucan,
water dikinase (GWD). This enzyme catalyses the transfer of
the β-phosphate of ATP to the C-6 position of glucosyl residues
of amylopectin (Ritte et al. 2000, 2006). Although the presence
of covalently linked phosphate in some starches has long been
known, its biological significance has come to light only recently.
The extent of phosphorylation varies considerably, depending on
the source of the starch. In Arabidopsis leaf starch, ∼1 in 2000
glucosyl residues is phosphorylated, whereas potato tuber starch
contains four times as much. Cereal endosperm starch contains
little or no phosphate.
Glucan, water dikinase was discovered by Lorberth et al.
(1998) as a protein binding to starch granules extracted from
potato tubers (the protein was initially named ‘R1’). Antisense
repression of GWD resulted in a reduction in starch-bound
phosphate. It also caused starch to accumulate in leaves
and prevented the cold-induced sweetening of stored tubers
(a process during which starch is broken down). Thus, Lorberth
et al. (1998) proposed that the phosphorylation of starch
is necessary for its subsequent mobilisation. Although the
phosphorylation of amylopectin occurs as starch is synthesised
(Nielsen et al. 1994), GWD is reported to bind to the surface
of potato leaf starch granules preferentially at night (Ritte et al.
2000). Furthermore, Ritte et al. (2004) provided evidence that
the rate of phosphorylation of the granule surface is higher
during starch breakdown than during biosynthesis. However,
Mikkelsen et al. (2005) reported that GWD is redox-regulated
and that the fraction bound to starch granules is in an inactive,
oxidised form, whereas the soluble form is reduced and active.
This observation appears inconsistent with a higher rate of nighttime
phosphorylation.
Yu et al. (2001) reported that the sex1 mutant of Arabidopsis
also lacks GWD (sex1 was initially named TC26, then TC265
by Caspar et al. 1991). Trethewey and ap Rees (1994a, 1994b)
conducted biochemical studies of sex1 and proposed that the
cause of the sex phenotype was a deficiency in the capacity
of the mutant to export glucose from the chloroplast at night.
However, as the sex phenotype was shown to be due to
the mutation in the gene encoding GWD, this conclusion
seems incorrect. The importance of glucose export for starch
degradation still awaits evaluation. Like the potato plants in
which GWD was repressed, the sex1 mutant has reduced levels
of starch-bound phosphate. In sex1 null mutants, the starch levels
are from 4- to 6-fold higher than in wild type plants at the end
of the day, and phosphate levels in the starch are reduced to
the limit of detection (Yu et al. 2001; Ritte et al. 2006). Starch
accumulates gradually in the mutant leaves, with young, newly
emerged leaves containing less than mature leaves (Zeeman and
ap Rees 1999).
A second protein, which shares sequence similarity with
GWD, is now known to phosphorylate amylopectin exclusively

468 Functional Plant Biology S. C. Zeeman et al.
on the C3 position of the glucosyl residues (Baunsgaard et al.
2005; Kötting et al. 2005; Ritte et al. 2006). This second enzyme
requires that its glucan substrate is already phosphorylated
(hence the name phosphoglucan, water dikinase – PWD) and
so is dependent on the previous action of GWD. Mutants
of Arabidopsis lacking PWD also exhibit a starch-excess
phenotype, although much less severe than that of mutants
lacking GWD. It is important to note that the mechanism
by which the presence of phosphate groups on amylopectin
permits degradation is still not clear. A plausible hypothesis is
that phosphorylation disrupts the semi-crystalline packing of
the starch granule providing ‘access points’ at which glucan
metabolising enzymes can act on the exposed amylopectin
chains. A high rate of transient phosphorylation at the surface
of the starch granule at night could thus facilitate degradation
(Fig. 2A, B). As PWD acts on glucans that have already been
phosphorylated by GWD, it is reasonable to speculate that their
combined actions result in two or more phosphorylated glucosyl
residues in close proximity to each other. Phosphate pairs or
clusters might be more effective in causing a localised disruption
of amylopectin packing than single phosphate groups. GWD
alone might distribute phosphate groups evenly, resulting in only
partial disruption of the granule surface. This could restrict the
rate of starch degradation and explain the mild starch-excess
phenotype of the pwd mutant. However, there is as yet no direct
evidence to support such a model.
It is not currently known how or when the phosphate groups
added by GWD and PWD are removed from the glucans during
starch breakdown. Phosphoglucans may be released from the
granule and metabolised in the chloroplast stroma. Alternatively,
phosphate groups may be removed directly from the granule
surface. One possibility is that the recently discovered glucanbinding
phosphatase encoded at the SEX4 locus fulfils this
role. Like sex1, sex4 was identified through screening for
mutant plants that contain starch at the end of the night,
when the wild type has exhausted its starch reserves. The
mutant has a reduced rate of starch degradation at night and
accumulates starch as a result of the imbalance between daytime
synthesis and night-time breakdown. Mature sex4 leaves
have three times as much starch as the wild type at the end
of the day (Zeeman et al. 1998; Zeeman and ap Rees 1999).
Niittylä et al. (2006) showed that SEX4 encodes a predicted dualspecificity
protein phosphatase (DSP), initially described by
Fordham-Skelton et al. (2002) as PTPKIS (for protein tyrosine
phosphatase with a kinase interaction sequence). More recently
the same protein has been described by Kerk et al. (2006) as
DSP and Sokolov et al. (2006) as DSP4. The domain of SEX4
initially described as a KIS domain is now established as a
carbohydrate-binding module (Kerk et al. 2006; Niittylä et al.
2006; Sokolov et al. 2006). Remarkably, SEX4 has homologues
in animals (called laforin), mutation of which results in Lafora
disease, a neurodegenerative condition caused by abnormal
glycogen metabolism. Worby et al. (2006) reported that laforin
can mediate the dephosphorylation of a branched glucan
(potato amylopectin) in vitro and stated that SEX4 had similar
capabilities. If further studies confirm that phosphoglucans
are indeed the substrate for SEX4, it will be necessary to
determine whether the enzymes substrate in vivo is the granule
Granule surface Stroma
BAM
GWD
PWD
BAM
ISA3
ISA3
ATP
AMP + P i
ATP
AMP + P i
A
B
C
Maltose
D
Maltooligosaccharides
Fig. 2. A hypothetical mechanism for the initiation of starch granule
breakdown involving amylopectin phosphorylation. (A) The granule surface
is depicted as a semi-crystalline layer of amylopectin. Grey boxes represent
double helices formed by adjacent unbranched glucan chains (inset).
(B) Glucan, water dikinase (GWD) phosphorylates the surface of the
starch granule (red circle). Phosphoglucan, water dikinase (PWD) adds a
second phosphate group (gold circle), probably very close to one added
by GWD. Phosphorylation might disrupt the packing of the double helices
of amylopectin. (C) β-Amylase (BAM) degrades accessible, free chains at
the granule surface, liberating maltose, but cannot work past branch points.
(D) The debranching enzyme isoamylase 3 (ISA3) removes short branches,
liberating short malto-oligosaccharides and enabling the continued action of
β-amylase.

Pathways of starch breakdown Functional Plant Biology 469
surface or a pool of soluble phosphoglucans. Furthermore,
testable models are required to explain why phosphate groups
first need to be added and subsequently removed for normal
starch breakdown to occur. It remains possible that SEX4
in fact dephosphorylates and thus regulates the activities of
proteins involved in starch degradation as well as, or instead
of, dephosphorylating amylopectin.
The hydrolysis of glucans at the starch granule surface
Given that α-amylase is not required for starch breakdown
in Arabidopsis, other enzymes must be able to attack the
starch granule. Results from several laboratories indicate
that β-amylase (an exoamylase which produces maltose) and
debranching enzymes may fulfil this role by acting in synergy.
Scheidig et al. (2002) first reported that antisense repression of a
chloroplast-targeted isoform of β-amylase (CT-BMY1) in potato
caused an accumulation of starch in leaves. Kaplan and Guy
(2005) showed that the repression of the Arabidopsis homologue
of CT-BMY (BAM3, also called BMY8) using RNAi also caused
a sex phenotype, a result supported by our analysis of a bam3
null mutant (S. C. Zeeman, A. M. Smith and S. M. Smith,
unpubl. data). Scheidig et al. (2002) analysed the activity of the
potato protein in vitro and showed that, in addition to activity
on soluble substrates, PCT-BMY1 was able to liberate a small
amount of maltose from isolated potato starch granules. Such
experiments can be difficult to interpret. On the one hand, a
low activity of β-amylase against intact starch granules is not
surprising as only the external linear chains of the granule will
be available for hydrolysis. β-amylase cannot work past the
branch points of amylopectin and would rapidly create a β-limit
structure consisting of short branches or ‘stubs’ on the granule
surface. The action of a debranching enzyme would be necessary
for further β-amylolysis to occur. Furthermore, if β-amylase
requires that amylopectin is first phosphorylated by GWD,
in vitro conditions may be unsuitable to show its true activity.
On the other hand, it can be difficult to purify starch granules
that are free of contaminating proteins (e.g. α-amylases), which
might provide soluble substrates for β-amylase.
By analysing the lengths of the glucan chains on the
outermost layer of potato starch granules, Ritte et al. (2004)
showed that the abundance of short chains 3–5 glucose residues
in length increases markedly in the first hour of the night. This
observation is consistent with β-amylolytic attack of the external
chains. The short branches that result are probably removed
directly from the granule surface by the debranching enzyme
isoamylase 3 (ISA3; Fig. 2C, D). Loss of this enzyme causes
starch to accumulate in Arabidopsis leaves, implicating it in
starch breakdown (Wattebled et al. 2005; Delatte et al. 2006).
Analysis of the activity of the potato ISA3 protein in vitro showed
that it was particularly active on β-limit dextrins, and also had
some action on isolated starch granules (Hussain et al. 2003).
Delatte et al. (2006) analysed the isa3 mutant phenotype of
Arabidopsis in detail and showed that the plants have a reduced
rate of starch degradation compared to the wild type and that
short chains 3–5 glucose residues in length are abundant in the
starch. This suggests that short chains are not efficiently removed
by debranching in the isa3 mutant. Furthermore, expression of
an ISA3-GFP construct in Arabidopsis leaf protoplasts showed
that the fluorescent fusion protein localises to the starch granule
surface, consistent with its proposed role there (Delatte et al.
2006). Taken together, these data indicate that the surface of
the starch granule is subject to progressive exoamylolysis and
debranching and that this process may be dependent on the
previous phosphorylation of the granule to disrupt the packing
of double helices (Fig. 2).
The mechanism described above results in the release
of maltose together with some linear malto-oligosaccharides
directly from the granule surface. Interestingly, in mutants
lacking components of this pathway (e.g. isa3 and bam3),
starch degradation still occurs, albeit at a reduced rate. This
shows that other enzymes are also able to degrade starch.
Delatte et al. (2006) investigated the effect of mutating a
second debranching enzyme, limit dextrinase (LDA), alone
and in addition to ISA3. The lda mutant displays normal
starch metabolism indicating that, like α-amylase, LDA is not
required for normal rates of starch breakdown. Interestingly,
when isa3 and lda mutations were combined, the double mutants
exhibited a more severe starch-excess phenotype than the isa3
single mutant, showing that LDA was contributing to starch
breakdown in the isa3 background. Further analysis of isa3
and of the isa3/lda double mutant provided evidence that
α-amylase might also contribute to starch degradation in these
lines. First, there was still appreciable starch degradation in
the double mutant. Second, the amount of the chloroplastic
α-amylase AMY3 was increased in both lines relative to the
wild type, as was total endoamylolytic activity. Third, branched
oligosaccharides accumulated in the double mutant, but not in
the isa3 single mutant. Delatte et al. (2006) interpreted these
data to mean that in isa3, starch degradation proceeds at least
partly through α-amylase. Liberated branched oligosaccharides
are then debranched by LDA in the stroma to yield linear
oligosaccharides, which can be metabolised further. In the
isa3/lda double mutant, branched oligosaccharides that are
released by AMY3 cannot be metabolised and so accumulate
in the stroma. Thus, there may be two distinct mechanisms by
which the surface of the starch granule is degraded: a major
pathway involving β-amylase and ISA3, and a minor pathway
involving α-amylase and LDA (Fig. 3).
Metabolism of malto-oligosaccharides
There is good evidence that the major product of starch
breakdown is maltose, which is exported by a unique chloroplast
envelope transporter, MEX1, to the cytosol (Niittylä et al. 2004;
Weise et al. 2004). In the cytosol, maltose is metabolised by
a glucosyltransferase, DPE2 (Chia et al. 2004; Lu and Sharkey
2004). In the reaction catalysed by DPE2, one glucosyl residue is
transferred to an acceptor molecule, whereas the other is released
as glucose. Glycogen, amylopectin and oligosaccharides will
all serve as acceptors for the transferred glucosyl residue in
vitro (Chia et al. 2004; Lloyd et al. 2004; Fettke et al. 2006).
However, the likely acceptor in vivo is a soluble heteroglycan
present in the cytosol (Fettke et al. 2005a). Glucose liberated
by DPE2 can be phosphorylated by hexokinase and thereby
enter the hexose-phosphate pool, whereas the glucosyl residue

470 Functional Plant Biology S. C. Zeeman et al.
Chloroplast
metabolism
Glc1P
AMY3
Branched maltooligosaccharides
LDA
Starch
ISA3
Linear maltooligosaccharides
PHS1
β-amylase
pGlcT
DPE1
Glucose
Maltose
β-amylase
(BAM3)
MEX1
Chloroplast stroma
Cytosol
Fig. 3. The pathways of starch breakdown in Arabidopsis leaves. Maltose is the major product of starch
breakdown, exported to the cytosol by the chloroplast envelope transporter, MEX1. Maltose is produced by
β-amylase acting either at the granule surface (see Fig. 2) or on soluble malto-oligosaccharides. The actions of
isoamylase (ISA3, see Fig. 2) or α-amylase (AMY3) on the granule may release soluble malto-oligosaccharides,
which, if branched, can be debranched by limit dextrinase (LDA). Short linear oligosaccharides (e.g. maltotriose)
can be metabolised by disproportionating enzyme (DPE1) liberating glucose for export to the cytosol (by the
glucose transporter, pGlcT) and larger malto-oligosaccharides for continued degradation. Chloroplast α-glucan
phosphorylase (PHS1) can use linear oligosaccharides to produce Glc-1-P for chloroplast metabolism. Some steps
are depicted with dashed lines because removal of these enzymes individually does not prevent a normal rate of
starch breakdown at night.
transferred to the cytosolic acceptor might be liberated as
glucose-1-phosphate by the cytosolic α-glucan phosphorylase,
PHS2 (Lu et al. 2006a).
The loss of MEX1 or DPE2 leads to the massive accumulation
of maltose (Chia et al. 2004; Lu and Sharkey 2004; Niittylä et al.
2004). Lu et al. (2006b) confirmed that in the mex1 mutant,
the maltose accumulates in the chloroplast, whereas in the dpe2
mutant it accumulates in both the chloroplast and the cytosol.
These data are consistent with the idea that MEX1 facilitates
bi-directional diffusion of maltose across the envelope, and
that DPE2 is cytosolic. These phenotypes also suggest that
α-glucosidase (also known as maltase) does not have a
significant role in maltose metabolism in Arabidopsis leaves.
In potato, there remains some uncertainty over the cellular
location of DPE2. Initially, Lloyd et al. (2004) reported the
protein to be localised in the chloroplast. This implies a different
mechanism of maltose metabolism. However, a subsequent study
has indicated that potato DPE2 is cytosolic, as in Arabidopsis
(Fettke et al. 2005b). For additional reviews on cytosolic maltose
metabolism, see Smith et al. (2005), Lu and Sharkey (2006) and
Fettke et al. (2007).
In addition to direct release from the granule surface, maltose
can be produced by the β-amylolysis of linear oligosaccharides
in the chloroplast stroma. Linear oligosaccharides themselves
can be released from the granule by debranching enzyme
(i.e. ISA3), by α-amylase, or by the debranching of soluble
branched glucans released by α-amylase. Early biochemical
studies showed that β-amylase is able to efficiently metabolise
oligosaccharides longer than 4 glucosyl residues in length,
but that maltotriose is a poor substrate (Chapman et al.
1972). Thus, in addition to maltose, some maltotriose will be
produced when β-amylase metabolises linear chains with an
odd number of glucosyl residues. Furthermore, ISA3 might
release maltotriosyl branches from the starch granule surface
(Delatte et al. 2006).
Maltotriose is metabolised in the chloroplast stroma by
disproportionating enzyme (D-enzyme, DPE1). The loss of
D-enzyme in Arabidopsis results in maltotriose accumulation
during starch breakdown (Critchley et al. 2001). When provided
solely with maltotriose in vitro, D-enzyme transfers a maltosyl
moiety from one molecule to another, producing maltopentaose
and glucose (Jones and Whelan 1969). In vivo, glucose produced
this way could be exported from the chloroplast by the
glucose transporter (Schäfer et al. 1977; Weber et al. 2000),
and maltopentaose could serve as a substrate for β-amylase,
resulting in maltose and maltotriose. Such a cycle would result

Pathways of starch breakdown Functional Plant Biology 471
in the net conversion of maltotriose to equimolar amounts of
maltose and glucose. However, D-enzyme can catalyse a wide
range of reactions and it is likely that it acts on other maltooligosaccharides,
as well as maltotriose in vivo. Thus, it might
function to recycle short malto-oligosaccharides in general.
Interestingly, D-enzyme does not produce or metabolise maltose
(Jones and Whelan 1969; Lin and Preiss 1988). This means that
when maltose is produced during starch breakdown, it is not
re-incorporated into the malto-oligosaccharide pool and is only
metabolised after export to the cytosol.
In addition to the production of the free sugars maltose
and glucose, glucose-1-phosphate can be produced from
linear malto-oligosaccharides by the chloroplastic form of
α-glucan phosphorylase (Fig. 3). The role of phosphorolysis
in the chloroplast is not yet clear, but it probably represents
a pathway for the production of substrates for chloroplast
metabolism, as the chloroplast envelope is not permeable to
hexose-phosphates. Repression of chloroplastic phosphorylase
by antisense techniques in potato had no detectable effect on
plant growth or starch metabolism (Sonnewald et al. 1995).
Complete removal of the enzyme (PHS1) by insertional
mutagenesis in Arabidopsis had only a minor effect on leaf
starch metabolism (Zeeman et al. 2004). In phs1 mutants,
small groups of cells bordering necrotic lesions were found to
contain large amounts of starch. Necrotic lesions were more
frequently observed in phs1 mutants than in wild type plants,
and phs1 plants were particularly susceptible to acute water
and salt stresses. Plausible roles for phosphorylase include the
provision of substrates for the chloroplastic oxidative pentose
phosphate pathway (Zeeman et al. 2004). This pathway might
be important during the dark for the provision of reducing
agents necessary to tolerate stress conditions (e.g. for the
scavenging of reactive oxygen species). Phosphorylase may also
supply carbon that can replenish intermediates of the Calvin
cycle. This may be particularly important under photorespiratory
conditions where such intermediates could become depleted
(Weise et al. 2006).
Conclusions
The findings summarised in this mini-review illustrate
both the progress made recently in understanding starch
degradation, and the number of major questions that remain
unanswered. Although the evidence supports the existence
of distinct mechanisms of starch granule degradation in
Arabidopsis – a minor endoamylolytic pathway and a
major exoamylolytic/debranching pathway – the functional
distinctions between the two pathways is not clear. The degrees
to which the results from Arabidopsis can be extended to
other species also remain unclear. As starch is the form
in which carbohydrate is stored in leaves, it seems likely
that its metabolism may respond to diverse physiological and
environmental parameters. The narrow range of environmental
conditions typically used in laboratory experiments may not
bring all such parameters to the fore. It is particularly important
that further progress is made in understanding how the
phosphorylation of amylopectin influences starch breakdown.
The severe phenotype that results from the loss of GWD implies
that it exerts significant control over the process. It will be
very interesting to determine how the phosphatase SEX4 is
involved and to discover the extent of the similarities between
starch metabolism and animal glycogen metabolism. Finally, a
complete understanding of the pathways of starch breakdown
will enable us to study how they are regulated and integrated
with the rest of plant primary metabolism.
Acknowledgements
We offer particular thanks to our close collaborators, Professor Alison
M. Smith and Professor Steven M. Smith, and to our colleagues in the
Arabidopsis Starch Metabolism Network (www.starchmetnet.org, accessed
17 March 2007). We also thank Martine Trevisan and Simona Eicke for
excellent technical support, past and present, in our laboratory. Our research
is funded by the Roche Research Foundation, the Swiss National Science
Foundation (Nation Centre for Competence in Research – Plant Survival and
Grant 3100–067312.01/1), and the EMBO Young Investigator Programme.
References
Asatsuma S, Sawada C, Itoh K, Okito M, Kitajima A, Mitsui T (2005)
Involvement of α-amylase I-1 in starch degradation in rice chloroplasts.
Plant & Cell Physiology 46, 858–869. doi: 10.1093/pcp/pci091
Baunsgaard L, Lütken H, Mikkelsen R, Glaring MA, Pham TT, Blennow A
(2005) A novel isoform of glucan, water dikinase phosphorylates
pre-phosphorylated α-glucans and is involved in starch degradation
in Arabidopsis. The Plant Journal 41, 595–605. doi: 10.1111/j.1365-
313X.2004.02322.x
Buléon A, Colonna P, Planchot V, Ball S (1998) Starch granules: structure
and biosynthesis. International Journal of Biological Macromolecules
23, 85–112. doi: 10.1016/S0141-8130(98)00040-3
Caspar T, Huber SC, Somerville C (1985) Alterations in growth,
photosynthesis, and respiration in a starchless mutant of Arabidopsis
thaliana (L.) deficient in chloroplast phosphoglucomutase activity.
Plant Physiology 79, 11–17.
Caspar T, Lin T-P, Kakefuda G, Benbow L, Preiss J, Somerville C (1991)
Mutants of Arabidopsis with altered regulation of starch degradation.
Plant Physiology 95, 1181–1188.
Chapman GW Jr, Pallas JE Jr, Mendicino J (1972) The hydrolysis of
maltodextrins by a β-amylase isolated from leaves of Vicia faba.
Biochimica et Biophysica Acta 276, 491–507. doi: 10.1016/0005-
2744(72)91010-8
Chatterton NJ, Silvius JE (1980) Photosynthate partitioning into leaf
starch as affected by daily photosynthetic period duration in 6
species. Physiologia Plantarum 49, 141–144. doi: 10.1111/j.1399-
3054.1980.tb02642.x
Chia T, Thorneycroft D, Chapple A, Messerli G, Chen J, Zeeman SC,
Smith SM, Smith AM (2004) A cytosolic glucosyltransferase is
required for conversion of starch to sucrose in Arabidopsis leaves
at night. The Plant Journal 37, 853–863. doi: 10.1111/j.1365-
313X.2003.02012.x
Critchley JH, Zeeman SC, Takaha T, Smith AM, Smith SM (2001) A critical
role for disproportionating enzyme in starch breakdown is revealed by
a knock-out mutation in Arabidopsis. The Plant Journal 26, 89–100.
doi: 10.1046/j.1365-313x.2001.01012.x
Delatte T, Umhang M, Trevisan M, Eicke S, Thorneycroft D, Smith SM,
Zeeman SC (2006) Evidence for distinct mechanisms of starch granule
breakdown in plants. The Journal of Biological Chemistry 281,
12 050–12 059. doi: 10.1074/jbc.M513661200
Fettke J, Eckermann N, Tiessen A, Geigenberger P, Steup M (2005a)
Identification, subcellular localization and biochemical characterization
of water-soluble heteroglycans (SHG) in leaves of Arabidopsis
thaliana L.: distinct SHG reside in the cytosol and in the
apoplast. The Plant Journal 43, 568–585. doi: 10.1111/j.1365-
313X.2005.02475.x

472 Functional Plant Biology S. C. Zeeman et al.
Fettke J, Poeste S, Eckermann N, Tiessen A, Pauly M, Geigenberger P,
Steup M (2005b) Analysis of cytosolic heteroglycans from leaves
of transgenic potato (Solanum tuberosum L.) plants that under- or
overexpress the Pho 2 phosphorylase isozyme. Plant & Cell Physiology
46, 1987–2004. doi: 10.1093/pcp/pci214
Fettke J, Chia T, Eckermann N, Smith A, Steup M (2006) A transglucosidase
necessary for starch degradation and maltose metabolism in leaves at
night acts on cytosolic heteroglycans (SHG). The Plant Journal 46,
668–684. doi: 10.1111/j.1365-313X.2006.02732.x
Fettke J, Eckermann N, Kötting O, Ritte G, Steup M (2007) Novel starchrelated
enzymes and carbohydrates. Cellular and Molecular Biology 52,
OL883–OL904.
Fincher GB (1989) Molecular and cellular biology associated with
endosperm mobilization in germinating cereal grains. Annual Review
of Plant Physiology and Plant Molecular Biology 40, 305–346.
doi: 10.1146/annurev.pp.40.060189.001513
Fordham-Skelton AP, Chilley P, Lumbreras V, Reignoux S, Fenton TR,
Dahm CC, Pages M, Gatehouse JA (2002) A novel higher plant protein
tyrosine phosphatase interacts with SNF1-related protein kinases via
a KIS (kinase interaction sequence) domain. The Plant Journal 29,
705–715. doi: 10.1046/j.1365-313X.2002.01250.x
Gibon Y, Bläsing OE, Palacios-Rojas N, Pankovic D, Hendriks JHM,
Fisahn J, Höhne M, Günther M, Stitt M (2004) Adjustment of
diurnal starch turnover to short days: depletion of sugar during the
night leads to a temporary inhibition of carbohydrate utilization,
accumulation of sugars and post-translational activation of ADP-glucose
pyrophosphorylase in the following light period. The Plant Journal 39,
847–862. doi: 10.1111/j.1365-313X.2004.02173.x
Gubler F, Kalla R, Roberts JK, Jacobsen JV (1995) Gibberellin-regulated
expression of a MYB gene in barley aleurone cells – evidence for MYB
transactivation of a high-pI α-amylase gene promoter. The Plant Cell 7,
1879–1891. doi: 10.1105/tpc.7.11.1879
Gubler F, Raventos D, Keys M, Watts R, Mundy J, Jacobsen JV (1999) Target
genes and regulatory domains of the GAMYB transcriptional activator
in cereal aleurone. The Plant Journal 17, 1–9. doi: 10.1046/j.1365-
313X.1999.00346.x
Hussain H, Mant A, Seale R, Zeeman SC, Hinchliffe E et al. (2003)
Three isoforms of isoamylase contribute different catalytic properties
for the debranching of potato glucans. The Plant Cell 15, 133–149.
doi: 10.1105/tpc.006635
Jones G, Whelan WJ (1969) The action pattern of D-enzyme, a
transmaltodextrinylase from potato. Carbohydrate Research 9, 483–490.
doi: 10.1016/S0008-6215(00)80033-6
Kaplan F, Guy CL (2005) RNA interference of Arabidopsis β-amylase
8 prevents maltose accumulation upon cold shock and increases
sensitivity of PSII photochemical efficiency to freezing stress. The Plant
Journal 44, 730–743. doi: 10.1111/j.1365-313X.2005.02565.x
Kerk D, Conley TR, Rodriguez FA, Tran HT, Nimick M, Muench DG,
Moorhead GBG (2006) A chloroplast-localized dual-specificity protein
phosphatase in Arabidopsis contains a phylogenetically dispersed and
ancient carbohydrate-binding domain, which binds the polysaccharide
starch. The Plant Journal 46, 400–413. doi: 10.1111/j.1365-
313X.2006.02704.x
Kötting O, Pusch K, Tiessen A, Geigenberger P, Steup M, Ritte G
(2005) Identification of a novel enzyme required for starch metabolism
in Arabidopsis leaves. The phosphoglucan, water dikinase. Plant
Physiology 137, 242–252. doi: 10.1104/pp.104.055954
Lin T-P, Preiss J (1988) Characterisation of D-enzyme (4-αglucanotransferase)
in Arabidopsis leaf. Plant Physiology 86, 260–265.
Lloyd JR, Blennow A, Burhenne K, Kossmann J (2004) Repression
of a novel isoform of disproportionating enzyme (stDPE2) in
potato leads to inhibition of starch degradation in leaves but not
tubers stored at low temperature. Plant Physiology 134, 1347–1354.
doi: 10.1104/pp.103.038026
Lorberth R, Ritte G, Willmitzer L, Kossmann J (1998) Inhibition of a
starch-granule-bound protein leads to modified starch and repression
of cold sweetening. Nature Biotechnology 16, 473–477. doi: 10.1038/
nbt0598-473
Lu Y, Sharkey TD (2004) The role of amylomaltase in maltose
metabolism in the cytosol of photosynthetic cells. Planta 218, 466–473.
doi: 10.1007/s00425-003-1127-z
Lu Y, Gehan JP, Sharkey TD (2005) Daylength and circadian effects on starch
degradation and maltose metabolism. Plant Physiology 138, 2280–2291.
doi: 10.1104/pp.105.061903
Lu Y, Sharkey TD (2006) The importance of maltose in transitory
starch breakdown. Plant, Cell & Environment 29, 353–366.
doi: 10.1111/j.1365-3040.2005.01480.x
Lu Y, Steichen JM, Yao J, Sharkey TD (2006a) The role of cytosolic
α-glucan phosphorylase in maltose metabolism and the comparison of
amylomaltase in Arabidopsis and Escherichia coli. Plant Physiology
142, 878–889. doi: 10.1104/pp.106.086850
Lu Y, Steichen JM, Weise SE, Sharkey TD (2006b) Cellular and organ
level localization of maltose in maltose-excess Arabidopsis mutants.
Planta 224, 935–943. doi: 10.1007/s00425-006-0263-7
Mikkelsen R, Mutenda KE, Mant A, Schürmann P, Blennow A (2005)
α-Glucan, water dikinase (GWD): a plastidic enzyme with redoxregulated
and coordinated catalytic activity and binding affinity.
Proceedings of the National Academy of Sciences USA 102, 1785–1790.
doi: 10.1073/pnas.0406674102
Nielsen TH, Wischmann B, Enevoldsen K, Møller BL (1994) Starch
phosphorylation in potato tubers proceeds concurrently with de novo
biosynthesis of starch. Plant Physiology 105, 111–117.
Niittylä T, Messerli G, Trevisan M, Chen J, Smith AM, Zeeman SC
(2004) A previously unknown maltose transporter essential for starch
degradation in leaves. Science 303, 87–89. doi: 10.1126/science.
1091811
Niittylä T, Comparot-Moss S, Lue W-L, Messerli G, Trevisan M et al. (2006)
Similar protein phosphatases control starch metabolism in plants and
glycogen metabolism in mammals. Journal of Biological Chemistry 281,
11 815–11 818. doi: 10.1074/jbc.M600519200
Ritte G, Lorberth R, Steup M (2000) Reversible binding of the starch-related
R1 protein to the surface of transitory starch granules. The Plant Journal
21, 387–391. doi: 10.1046/j.1365-313x.2000.00683.x
Ritte G, Scharf A, Eckermann N, Haebel S, Steup M (2004) Phosphorylation
of transitory starch is increased during degradation. Plant Physiology
135, 2068–2077. doi: 10.1104/pp.104.041301
Ritte G, Heydenreich M, Mahlow S, Haebel S, Kötting O, Steup M
(2006) Phosphorylation of C6- and C3-positions of glucosyl residues in
starch is catalysed by distinct dikinases. FEBS Letters 580, 4872–4876.
doi: 10.1016/j.febslet.2006.07.085
Schäfer G, Heber U, Heldt HW (1977) Glucose transport into spinach
chloroplasts. Plant Physiology 60, 286–289.
Scheidig A, Fröhlich A, Schulze S, Lloyd JR, Kossmann J (2002)
Down regulation of a chloroplast-targeted β-amylase leads to a
starch-excess phenotype in leaves. The Plant Journal 30, 581–591.
doi: 10.1046/j.1365-313X.2002.01317.x
Smith AM, Zeeman SC, Smith SM (2005) Starch degradation.
Annual Review of Plant Biology 56, 73–98. doi: 10.1146/annurev.
arplant.56.032604.144257
Sokolov LN, Dominguez-Solis JR, Allary A-L, Buchanan BB, Luan S
(2006) A redox-regulated chloroplast protein phosphatase binds to
starch diurnally and functions in its accumulation. Proceedings
of the National Academy of Sciences USA 103, 9732–9737.
doi: 10.1073/pnas.0603329103
Sonnewald U, Basner A, Greve B, Steup M (1995) A second L-type
isozyme of potato glucan phosphorylase: cloning, antisense inhibition
and expression analysis. Plant Molecular Biology 27, 567–576.
doi: 10.1007/BF00019322

Pathways of starch breakdown Functional Plant Biology 473
Stanley D, Fitzgerald AM, Farnden KJF, McRae EA (2002) Characterization
of putative amylases from apple (Malus domestica) and Arabidopsis
thaliana. Biologia 57(Supp), 137–148.
Trethewey RN, ap Rees T (1994a) A mutant of Arabidopsis thaliana
lacking the ability to transport glucose across the chloroplast envelope.
The Biochemical Journal 301, 449–454.
Trethewey RN, ap Rees T (1994b) The role of the hexose transporter
in the chloroplasts of Arabidopsis thaliana L. Planta 195, 168–174.
doi: 10.1007/BF00199675
Wattebled F, Dong Y, Dumez S, Delvallé D, Planchot R et al. (2005)
Mutants of Arabidopsis lacking a chloroplastic isoamylase accumulate
phytoglycogen and an abnormal form of amylopectin. Plant Physiology
138, 184–195. doi: 10.1104/pp.105.059295
Weber A, Servaites JC, Geiger DR, Kofler H, Hille D, Gröner F, Hebbeker U,
Flügge U-I (2000) Identification, purification, and molecular cloning of
a putative plastidic glucose translocator. The Plant Cell 12, 787–801.
doi: 10.1105/tpc.12.5.787
Weise SE, Weber APM, Sharkey TD (2004) Maltose is the major form of
carbon exported from the chloroplast at night. Planta 218, 474–482.
doi: 10.1007/s00425-003-1128-y
Weise SE, Schrader SM, Kleinbeck KR, Sharkey TD (2006) Carbon
balance and circadian regulation of hydrolytic and phosphorolytic
breakdown of transitory starch. Plant Physiology 141, 879–886.
doi: 10.1104/pp.106.081174
Worby CA, Gentry MS, Dixon JE (2006) Laforin: a dual specificity
phosphatase that dephosphorylates complex carbohydrates.
Journal of Biological Chemistry 281, 30 412–30 418. doi: 10.1074/
jbc.M606117200
http://www.publish.csiro.au/journals/fpb
Yu T-S, Kofler H, Häusler RE, Hille D, Flügge U-I et al. (2001)
SEX1 is a general regulator of starch degradation in plants and not
the chloroplast hexose transporter. The Plant Cell 13, 1907–1918.
doi: 10.1105/tpc.13.8.1907
Yu T-S, Zeeman SC, Thorneycroft D, Fulton DC, Dunstan H et al. (2005)
α-Amylase is not required for breakdown of transitory starch in
Arabidopsis leaves. Journal of Biological Chemistry 280, 9773–9779.
doi: 10.1074/jbc.M413638200
Zeeman SC, Northrop F, Smith AM, ap Rees T (1998) A starch-accumulating
mutant of Arabidopsis thaliana deficient in a chloroplastic starchhydrolyzing
enzyme. The Plant Journal 15, 357–365. doi: 10.1046/
j.1365-313X.1998.00213.x
Zeeman SC, ap Rees T (1999) Changes in carbohydrate metabolism
and assimilate partitioning in starch-excess mutants of Arabidopsis.
Plant, Cell & Environment 22, 1445–1453. doi: 10.1046/j.1365-
3040.1999.00503.x
Zeeman SC, Pilling E, Tiessen A, Kato L, Donald AM, Smith AM (2002)
Starch synthesis in Arabidopsis; granule synthesis, composition and
structure. Plant Physiology 129, 516–529. doi: 10.1104/pp.003756
Zeeman SC, Thorneycroft D, Schupp N, Chapple A, Weck M, Dunstan H,
Haldimann P, Bechtold N, Smith AM, Smith SM (2004) The role of
plastidial α-glucan phosphorylase in starch degradation and tolerance
of abiotic stress in Arabidopsis leaves. Plant Physiology 135, 849–858.
doi: 10.1104/pp.103.032631
Zeeman SC, Smith SM, Smith AM (2007) The diurnal metabolism of leaf
starch. The Biochemical Journal 401, 13–28. doi: 10.1042/BJ20061393
Manuscript received 28 November 2006, accepted 7 February 2007