The Plant Journal (2001) 26(1), 89±100
A critical role for disproportionating enzyme in starch
breakdown is revealed by a knock-out mutation in
Arabidopsis
Joanna H. Critchley1,², Samuel C. Zeeman2,², Takeshi Takaha3, Alison M. Smith2 and Steven M. Smith1,*
1
Institute of Cell and Molecular Biology, University of Edinburgh, May®eld Road, Edinburgh EH9 3JH, UK,
2
John Innes Centre, Colney Lane, Norwich NR4 7UH, UK, and
3
Biochemical Research Laboratories, Ezaki Glico Co Ltd, Utajima, Nishiyodogawa-ku, Osaka 555, Japan
Received 10 November 2000; revised 31 January 2001; accepted 5 February 2001.
*For correspondence (fax +44 131 650 5392; e-mail [email protected]).
²
These authors contributed equally to this work.
Summary
Disproportionating enzyme (D-enzyme) is a plastidial a-1,4-glucanotransferase but its role in starch
metabolism is unclear. Using a reverse genetics approach we have isolated a mutant of Arabidopsis
thaliana in which the gene encoding this enzyme (DPE1) is disrupted by a T-DNA insertion. While Denzyme activity is eliminated in the homozygous dpe1±1 mutant, changes in activities of other enzymes
of starch metabolism are relatively small. During the diurnal cycle, the amount of leaf starch is higher in
dpe1±1 than in wild type and the amylose to amylopectin ratio is increased, but amylopectin structure is
unaltered. The amounts of starch synthesised and degraded are lower in dpe1±1 than in wild type.
However, the lower amount of starch synthesised and the higher proportion of amylose are both
eliminated when plants are completely de-starched by a period of prolonged darkness prior to the light
period. During starch degradation, a large accumulation of malto-oligosaccharides occurs in dpe1±1 but
not in wild type. These data show that D-enzyme is required for malto-oligosaccharide metabolism
during starch degradation. The slower rate of starch degradation in dpe1±1 suggests that maltooligosaccharides affect an enzyme that attacks the starch granule, or that D-enzyme itself can act directly
on starch. The effects on starch synthesis and composition in dpe1±1 under normal diurnal conditions
are probably a consequence of metabolism at the start of the light period, of the high levels of maltooligosaccharides generated during the dark period. We conclude that the primary function of D-enzyme
is in starch degradation.
Keywords: starch metabolism, disproportionating enzyme, Arabidopsis thaliana, mutant, maltooligosaccharides.
Introduction
Disproportionating enzyme (D-enzyme) is a plastidial
1,4-a-D-glucan:1,4-a-D-glucan, 4-a-D-glucanotransferase
(EC 2.4.1.25) found in many different starch-containing
organs of plants (Kakefuda et al., 1986; Lin et al., 1988;
Takaha et al., 1993). Although it is probably involved in
starch metabolism, the precise role of D-enzyme is yet to
be established. Malto-oligosaccharides (MOS) are the
preferred substrates in vitro, and D-enzyme preferentially
transfers maltosyl units from one 1,4-a-D-glucan to
another, although glucose can also act as an acceptor
(Takaha and Smith, 1999). The disproportionation reaction
ã 2001 Blackwell Science Ltd
does not result in a change in the total number of a-1,4
bonds, so D-enzyme does not directly contribute to the net
synthesis or breakdown of glucan polymers. However, the
reaction does change the size distribution of MOS molecules, potentially generating suitable substrates for other
enzymes of starch metabolism. For example, the creation
of larger MOS molecules from small ones would provide
substrates for starch-degrading enzymes such as starch
phosphorylase and b-amylase, which both act preferentially or exclusively on larger MOS molecules. Based on
these considerations, it has been proposed that D-enzyme
89
90
Joanna H. Critchley et al.
is involved primarily in starch degradation (Kakefuda and
Duke, 1989; Lin and Preiss, 1988). Elimination of more than
98% of D-enzyme activity through expression of antisense
RNA in potatoes does not result in any detectable change
in tuber starch content or structure (Takaha et al., 1998a).
However, delayed tuber sprouting and shoot growth
suggested that starch is less effectively mobilised in such
plants, consistent with the proposal that D-enzyme has a
role in starch breakdown (Takaha et al., 1998a).
Analysis of a mutant of Chlamydomonas, which speci®cally lacks D-enzyme activity (the sta11 mutant), has led to a
very different proposal for the role of D-enzyme in starch
metabolism. Colleoni et al. (1999a) discovered that the
mutant has a reduced starch content, altered starch structure and high levels of MOS when grown under limiting
nitrogen to promote starch accumulation. The authors
argue that D-enzyme acts during starch synthesis to transfer
soluble MOS onto chains of the amylopectin component of
starch (Colleoni et al., 1999b). It has been proposed that
amylopectin is initially synthesised as a highly branched
glucan, which is then `trimmed' by a debranching enzyme
with the release of MOS, creating the correct branching
structure to allow its crystallisation to form the starch
granule (Ball et al., 1996; Myers et al., 2000). Colleoni
et al. (1999b) suggest that transfer of MOS released by
`trimming' back onto amylopectin via D-enzyme would both
allow ef®cient utilisation of the MOS and contribute to the
determination of amylopectin branch length.
The conclusions from the study of Chlamydomonas are
different from those of potato, and have radical implications for our understanding of starch synthesis. To gain
de®nitive information about the role of D-enzyme in higher
plants, we isolated a mutant of Arabidopsis which completely lacks D-enzyme. Under a normal day-night cycle,
the starch in wild-type Arabidopsis leaves accumulates
without signi®cant turnover during the light period (S.C.
Zeeman, unpublished results) and is completely degraded
during the subsequent night period (Zeeman et al., 1998a;
Zeeman et al., 1998b). Examination of the impact of loss of
D-enzyme on this pattern would thus allow us to distinguish whether it is involved in starch degradation, in
starch synthesis, or both. Our results demonstrate unambiguously that D-enzyme is necessary for normal starch
degradation, but provide no evidence for a direct role in
synthesis in Arabidopsis leaves.
Results
The D-enzyme gene of Arabidopsis
We ®rst isolated cDNA clones encoding Arabidopsis Denzyme using the potato cDNA as a probe (see
Experimental procedures). A cDNA including the whole
coding region plus 5¢ and 3¢ untranslated regions was
isolated from ecotype Wasserilewskija (Ws). The open
reading frame of 572 codons includes a predicted aminoterminal plastid-targeting transit peptide sequence of 45
amino acid residues. The predicted amino acid sequences
of mature D-enzymes of potato and Arabidopsis are 74%
identical. The gene encoding D-enzyme (DPE1) is located
on chromosome 5 (GenBank accession number AB019236),
and comparison of cDNA and genomic sequences reveals
that it contains 15 introns. Available genomic DNA
sequence and Southern blotting experiments (not shown)
fail to reveal any other D-enzyme genes in Arabidopsis.
However, a gene predicted to encode a related protein is
present on chromosome 2 (GenBank accession number
AC002409). The gene encodes a putative polypeptide of up
to 738 amino acid residues, with only 18% identity to
Arabidopsis D-enzyme and no predicted plastid-targeting
sequence. Based on its predicted amino acid sequence, this
putative protein belongs to the a-amylase super-family of
enzymes, but it does not fall into any of the ®ve currently
recognised types of a-1,4-glucanotransferase classi®ed
from the Archaea, Eubacteria and Eukaryota (Takaha and
Smith, 1999). We conclude that DPE1 is the only gene
identi®ed that encodes plastidial D-enzyme.
Isolation of a D-enzyme mutant
Arabidopsis plants containing random T-DNA insertions
were screened using PCR with multiple primer combinations, for insertions in the DPE1 gene (Krysan et al., 1996).
One from the DuPont collection (Feldman, 1991) was
found to contain an insertion in the DPE1 gene. PCR
primers close to the T-DNA left border gave products with
DPE1-speci®c primers from both 3¢ and 5¢ ends of the gene
but the T-DNA right border primers gave no products,
indicating that at least two T-DNA copies had inserted in
inverted orientation relative to each other. Such arrangement of multiple T-DNA insertions is not uncommon
(McKinney et al., 1995). DNA sequence analysis con®rmed
the presence of T-DNA left border sequences at both ends,
and showed that the insertion was at the beginning of
exon 2 (Figure 1). There was a small deletion and
rearrangement of T-DNA left border sequence at the 5¢
end of the insert, and a deletion of 24 bases at the 3¢ end of
the T-DNA. The results of DNA-gel blotting experiments
with DNA from the dpe1±1 homozygous mutant were
consistent with approximately two T-DNA copies in the
DPE1 gene (not shown). In the course of isolating a
homozygous dpe1±1 mutant, a wild-type segregant was
isolated from the same transformant to serve as a control.
Expression of DPE1 and phenotype of the dpe1±1 mutant
Using a speci®c polyclonal antibody against recombinant
Arabidopsis D-enzyme, we found that D-enzyme protein
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Function of D-enzyme in starch breakdown
91
Figure 1. Structure of the dpe1±1 locus.
Structure of the DPE1 gene is depicted from 0 (translation start) to 3781 (polyadenylation site) base pairs, showing exons (black boxes), introns and noncoding sequence (lines). The vertical line shows the site of insertion of T-DNA, and nucleotide and amino acid sequences on either side show the precise
sequences of the insertion site at the beginning of exon 2. T-DNA structure is comprised of an inverted repeat, each end with partial left border (left end)
nucleotide sequences (shown in capitals), together with nucleotides of unidenti®ed origin at junction of T-DNA and DPE1 gene (italicised capitals).
was not detectable in leaves of homozygous dpe1±1 but
was present at approximately half the wild-type amount in
leaves of heterozygous dpe1±1/DPE1 plants (Figure 2). In
wild-type leaves, there is no diurnal variation in the
amount of D-enzyme protein (not shown). No morphological phenotype was apparent in dpe1±1, and ¯owering
in 10 h days occurred after 6 weeks as in wild type, but at
this time dpe1±1 fresh weight was 20% less than that of
wild type (not shown).
D-enzyme
activity is not detectable in crude extracts of
dpe1±1
D-enzyme can be assayed by measuring the rate of glucose
production during the disproportionation of maltotriose to
glucose and maltopentaose. Using this assay the activity
of D-enzyme in crude extracts of dpe1±1 leaves was
apparently 5% of that in wild-type extracts (Table 1). To
determine whether the small amount of activity in dpe1±1
leaves was genuinely D-enzyme or a different activity such
as a-glucosidase, which can also liberate glucose from
maltotriose (Sun et al., 1995), we supplied maltotriose to
de-salted extracts of leaves and analysed the products.
Extracts of wild-type leaves converted maltotriose into
glucose and larger MOS, primarily maltopentaose and
maltoheptaose, indicating D-enzyme activity (Figure 3a).
Small amounts of maltose, maltotetraose and maltohexaose were also synthesized. In extracts of mutant
leaves, maltotriose was converted solely into glucose
and maltose, indicating activity of a hydrolase such as
a-glucosidase. No larger MOS were produced (Figure 3b).
The same result was observed when 10 times more of the
mutant extract was added to the reaction mixture (to
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Figure 2. Immunoblot of leaf proteins.
Total soluble leaf protein (10 mg) from dpe1±1/dpe1±1 homozygote (lane
1), dpe1±1/DPE1 heterozygote (lane 2) and DPE1/DPE1 wild type (lane 3)
was separated by SDS-PAGE, blotted onto membrane and probed with
anti-D-enzyme polyclonal antibody. Mass of molecular weight markers is
shown (kDa) and mobility of D-enzyme shown by an arrow (DPE).
compensate in part for the lower apparent activity compared with wild type, Table 1). By mixing small amounts of
the extract of wild-type leaves with the mutant extract we
determined that this method was sensitive enough to
detect less than 1% of the wild-type activity of D-enzyme
(not shown).
A single isoform of D-enzyme, detected by native PAGE,
is missing in dpe1±1
We used native PAGE to detect isoforms of D-enzyme in
leaf extracts. Soluble proteins were separated on native
polyacrylamide gels containing 1% (w/v) glycogen, which
were then incubated at pH 7 with 5 mM maltoheptaose.
92
Joanna H. Critchley et al.
Table 1. Comparison of the maximum catalytic activites of
starch metabolizing enzymes in crude extracts of leaves of wildtype and dpe1 Arabidopsis plants
Activity (nmol min±1 g±1 fresh weight)
Enzyme
Wild type
dpe1±1
a-amylase
b-amylase
Starch phosphorylase
Maltase
D-enzyme
Pullulanase
Soluble starch synthase
Branching enzymea
7.0 6 0.9
1235 6 142
77 6 7.5
39 6 3.2
264 6 17
50 6 4.0
119 6 21
4.4 6 0.6
8.0 6 1.9
2168 6 357
102 6 6.1
24 6 5.9
14 6 6.5
51 6 3.2
90 6 25
4.2 6 0.8
Measurements are the mean 6 standard error of values from ®ve
samples, each from an individual plant.
a
Activity is measured as the stimulation of the incorporation of
carbon-14 from 14C-glucose 1-phosphate into glucan by phosphorylase a and is given as micromoles per minute per gram
fresh weight of tissue.
Under these conditions, D-enzyme catalyses the transfer of
glucan units from maltoheptaose onto external chains of
the glycogen, resulting in a dark band when the gel is
stained with iodine (Colleoni et al., 1999b). One band of
D-enzyme activity was detected on gels of wild-type
extracts. This band was absent on gels of mutant extracts
(Figure 4a), and on control gels incubated without
maltoheptaose (not shown).
Two peaks of D-enzyme activity, separated by mono Q
chromatography, are both absent in dpe1±1
Earlier work showed that D-enzyme in wild-type
Arabidopsis (Columbia ecotype) can be separated into
two peaks of activity by Mono-Q anion exchange
chromatography (Lin and Preiss, 1988). We reproduced
this result using wild type of the WS ecotype (Figure 5a).
When this experiment was repeated using the mutant
dpe1±1, both peaks of D-enzyme activity were absent
(Figure 5a). This suggested that both peaks of activity are
due to the protein encoded by DPE1. Consistent with this,
the antibody raised against the DPE1 gene product recognised a 60-kDa protein in all the Mono Q fractions
containing D-enzyme activity (Figure 5b).
The mutation at DPE1 has minor effects on other
enzymes of starch metabolism
We assayed the activities of other enzymes involved in
starch metabolism in dpe1±1 and wild type (Table 1). All
assays were carried out under optimal conditions, and
were linear with respect to time and to the volume of
Figure 3. Analysis of the metabolism of maltotriose by extracts of wildtype and dpe1±1 leaves, using HPAEC-PAD.
(a) Maltotriose (®nal concentration 2 mM) was mixed with de-salted
extract from wild-type leaves and the products of the reaction after a 1-h
incubation were analysed (solid line, open symbols) and compared with
a zero-time control (dashed line, closed symbols).
(b) As described for (a) except maltotriose was mixed with de-salted
extract from dpe1±1 leaves.
Peaks representing glucose and the MOS maltose to maltoheptaose are
labelled G and G2-G7.
added extract. The activities of most of the enzymes did
not differ signi®cantly between wild type and the mutant.
However, the activity of b-amylase in the mutant was twice
that of wild type (P < 0.05), and starch phosphorylase
activity was slightly increased (P < 0.05). Native PAGE did
not reveal any changes in the isoform complement of
starch hydrolysing enzymes or starch phosphorylase
(Figure 4b,c).
dpe1±1 plants have a starch-excess phenotype and highamylose starch in normal diurnal conditions
In wild-type Arabidopsis, starch is made continuously
during the day and almost completely degraded during the
subsequent night (Zeeman et al., 1998b). When leaves
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Function of D-enzyme in starch breakdown
93
Figure 4. Native PAGE of starch metabolising enzymes in extracts of
leaves of wild type and dpe1±1.
(a) To detect D-enzyme activity: soluble proteins were extracted from
leaves and separated in polyacrylamide gels containing 1% (w/v)
glycogen. The gels were incubated in the presence of 5 mM
maltoheptaose for 2 h at 37°C, then stained with an iodine solution.
(b) To detect starch hydrolysing activities: soluble proteins were
separated in gels containing 0.1% (w/v) amylopectin. The gels were
incubated for 16 h at room temperature and stained with an iodine
solution.
(c) To detect starch phosphorylase: soluble proteins were separated in
gels containing 0.8% (w/v) glycogen. The gels were incubated for 16 h at
room temperature in the presence of 20 mM glucose 1-phosphate and
stained with an iodine solution.
Figure 6. Starch content of wild-type and dpe1±1 plants during the
diurnal cycle.
(a) Plants at the end of a 10-h light period (end of day) and a subsequent
14 h dark period (end of night) were decolourised with hot 80% (v/v)
ethanol and stained with an iodine solution to reveal the starch. Three
representative plants for each treatment are shown.
(b) Samples (0.2±0.6 g fresh weight) comprising all the leaves of
individual wild-type (solid line, open symbols) and dpe1±1 plants (dashed
line, closed symbols) were harvested and immediately frozen in liquid
N2. Leaves were ground to a powder and extracted for 30 min in ice-cold
perchloric acid. The perchloric acid was removed and the insoluble pellet
was washed three times in 80% (v/v) ethanol. The starch in the insoluble
material was measured. Each point is the mean 6 standard error of
values from six replicate samples. Where absent, error bars are smaller
than the symbols.
Figure 5. Separation of forms of D-enzyme by Mono-Q chromatography
and immunoblotting using an antibody to the DPE1 gene product.
(a) D-enzyme was partially puri®ed from extracts of wild-type (open
symbols) and dpe1±1 (closed symbols) plants by polyethylene glycol
precipitation and DEAE-Sepharose chromatography. Forms of D-enzyme
were then separated by Mono-Q chromatography. Fractions eluted from
the Mono-Q column were assayed for D-enzyme activity as described in
Experimental procedures.
(b) Proteins in each fraction eluted from the Mono-Q column of the wildtype preparation, described in (a), were separated by SDS-PAGE and
electroblotted onto nitrocellulose. The blots were developed with
antiserum to the DPE1 gene product and photographed.
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
were decolourised then stained for starch with iodine
solution, those of dpe1±1 stained more darkly than those
of wild type both at the end of the day and at the end of the
night (Figure 6a). This difference could re¯ect differences
in the rates of starch accumulation or degradation, or a
higher amylose content in dpe1±1.
To investigate this further, we measured the starch
content throughout the diurnal cycle (Figure 6b). The
amount of starch present at the start of the light period
was higher in the mutant than in wild type, but at the end
of the light period the starch contents of the two lines were
similar. Thus the mutant synthesised less starch during the
light period, and degraded less starch during the dark
period, than did the wild type.
To analyse starch composition, starch granules were
isolated from leaves at the end of the light period. The
94
Joanna H. Critchley et al.
Table 2. Starch accumulation in previously de-starched leaves of
wild-type and dpe1 Arabidopsis plants during a period of
extended light
Starch content
(mg g±1 fresh weight)
Time in light (h)
Wild type
dpe1±1
12
36
156
5.9 6 0.3
20.5 6 3.2
63.2 6 8.8
6.3 6 0.6
20.0 6 2.6
59.3 6 7.3
Plants were completely de-starched by an extended dark period
(36 h), then transferred to continuous light (100 mmoles photons
m±2 s±1). Samples were harvested as described in Figure 6. The
measurements are the mean 6 standard error of values from four
samples, each from an individual plant.
Figure 7. Sepharose CL2B separation of amylose and amylopectin from
starch of dpe1±1 and the wild type.
(a) Starch was extracted from plants at the end of a normal photoperiod
and dissolved at 10 mg ml±1 in 0.5 M NaOH. Samples containing 0.5 mg
applied to a 5-ml column. Fractions were eluted with 0.1 M NaOH. The
absorbance of each fraction after the addition of an iodine solution was
measured at 595 nm. To normalise each set of results, the absorbances
from each fraction were divided by the sum of the absorbance from the
sample.
(b) Starch was extracted from plants that had previously been completely
de-starched in the dark for 36 h, then allowed to photosynthesise for
12 h. The starch was analysed as described in (a).
amylose and amylopectin factions were separated by
Sepharose CL2B chromatography. Starch isolated from
the mutant contained more amylose than the starch from
wild type (Figure 7a),
Lack of D-enzyme does not in¯uence starch accumulation
and composition after plants are de-starched
Our experiments indicate that when plants are grown in a
normal light-dark regime, lack of D-enzyme affects both
starch accumulation and starch degradation. However,
they do not reveal whether D-enzyme is actually necessary
for both of these processes. A requirement for D-enzyme
for either synthesis or degradation might result in changes
in both processes in the mutant, because altered metabolism during one part of the day may in¯uence metabolism
during the other half of the day. To establish the importance of D-enzyme in accumulation and degradation in the
absence of such secondary effects, plants were kept in
darkness for 36 h prior to experimentation. This treatment
depleted starch reserves completely in both mutant and
wild type, and had no effect on D-enzyme activity in wild
type (not shown). We reasoned that this treatment would
minimise differences in metabolism between mutant and
wild type arising from an involvement of D-enzyme in
either starch accumulation or degradation.
After the 36 h of darkness, plants were transferred to
continuous light (either 100 or 175 mmol quanta m±2 s±1,
Tables 2 and 3, respectively). Under these conditions, the
amount of starch that accumulated over 12 h was the same
in mutant and wild-type leaves. The rate of accumulation
was essentially the same as that in wild-type plants in
normal diurnal conditions. Starch accumulation continued
at the same rate in mutant and wild-type leaves over the
subsequent 144 h (Table 2). The starch of mutant and wildtype plants had the same amylose to amylopectin ratio
after 12 h (Figure 7b), and this ratio was the same as that of
wild-type plants after 12 h light in normal diurnal conditions. There was also no difference between the amylose
to amylopectin ratios of starches from mutant and wildtype plants after 36 h and 256 h of the continuous light
regime (not shown). Thus D-enzyme exercises no control
over either the rate of starch accumulation or the relative
amounts of amylose and amylopectin in these conditions.
To study the effect of loss of D-enzyme on starch
degradation in the absence of potential secondary effects,
plants were kept in the dark for 36 h then transferred to
light (175 mmol quanta m±2 s±1) for 12 h. After this time,
when mutant and wild-type plants had the same starch
content and composition, plants were transferred to
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Function of D-enzyme in starch breakdown
95
Table 3. Starch degradation and malto-oligosaccharide accumulation in previously de-starched leaves of wild-type and dpe1
Arabidopsis plants allowed to photosynthesise for one day, then transferred to the dark
Malto-oligosaccharide content
(mg glucose equivalents g±1 fresh weight)
Starch content
(mg g±1 fresh weight)
Time in dark (h)
Wild type
0
2
6
12
11.7
10.2
5.7
1.2
6
6
6
6
1.7
0.8
0.6
0.1
dpe1±1
(5)
(5)
(5)
(4)
11.2
11.1
6.6
2.7
6
6
6
6
Wild type
1.0
0.9
0.4
0.4
(7)
(7)
(7)
(7)
0.033
0.070
0.051
0.035
6
6
6
6
0.004
0.016
0.015
0.006
dpe1±1
(5)
(5)
(5)
(4)
0.048
0.096
0.124
0.232
6
6
6
6
0.007
0.010
0.011
0.014
(7)
(7)
(7)
(7)
Plants were completely de-starched by an extended dark period (36 h), transferred to the light (175 mmoles photons m±2 s±1) for 12 h, then
returned to the dark for a further 12 h, during which time samples were harvested as described in Figure 6. The measurements are the
mean 6 standard error of values from between four and seven samples, each from an individual plant (n given in parentheses).
darkness. The rate of starch degradation was slower in
mutant than in wild-type plants, and at the end of the dark
period the mutant had signi®cantly more starch (P < 0.05).
Thus D-enzyme exercises some control over the rate of
starch degradation in these conditions.
dpe1±1 plants accumulate MOS during starch
degradation
To investigate the mechanism by which the loss of Denzyme affects starch metabolism we measured the MOS
content of the leaves throughout a normal diurnal cycle.
Methods for the extraction and measurement of MOS were
validated with recovery experiments in which a known
amount of maltotriose was added to one of two replicate
samples of leaf. The samples were extracted and the MOS
measured after hydrolysis to glucose. Recoveries were
calculated by comparing the MOS content of the two sets
of samples and were 91 6 2% (n = 4).
Wild-type leaves contained small amounts of MOS
(0.05 mg glucose equivalents g-1 FW) throughout the
light period. In the ®rst hour of the dark period, the levels
increased two-fold (Figure 8). There were much larger
diurnal changes in the MOS content of mutant leaves.
During most of the light period the MOS content was
slightly higher than in wild type. There was a dramatic
increase over the ®rst 4 h of darkness, then no further
change. At this point the MOS content of the mutant leaves
was 15 times higher than that of wild-type leaves. During
the ®rst 4 h of the light period the MOS content of the
mutant leaves decreased rapidly.
We used HPAEC-PAD to investigate the nature of the
MOS in wild-type and mutant leaves. Neutral compounds
were obtained from the extracts assayed for MOS (above)
by passage through sequential anion and cation exchange
columns, and MOS were separated on a Carbopac PA-100
column (Dionex, Camberley, Surrey, UK). Maltose was the
major MOS in wild-type leaves (Figure 9a). Small amounts
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Figure 8. The MOS content of leaves of dpe1±1 and the wild type during
the diurnal cycle.
Samples (0.2±0.6 g fresh weight) comprising all the leaves of individual
wild-type (solid line, open symbols) and dpe1±1 plants (dashed line,
closed symbols) were harvested and immediately frozen in liquid N2.
Leaves were ground to a powder and extracted for 30 min in ice-cold
0.7 M perchloric acid. The insoluble material was removed and the
soluble extract was adjusted to pH 5.6. Sucrose and glucose were
removed from samples of this extract by adding invertase, glucose
oxidase and catalase. The samples were boiled, cooled and MOS
digested to glucose by the addition of maltase and amyloglucosidase.
The glucose content was then determined. Each point is the mean 6
standard error of values from six replicate samples which are the same
samples used for starch determination in Figure 6. Where absent, error
bars are smaller than the symbols.
of maltotriose were detected when total MOS levels were
high. The maltose content of mutant leaves was similar to
that of wild type throughout the light and dark periods.
Most of the MOS accumulated during the dark period was
maltotriose, with small amounts of maltotetraose and
maltopentaose. During the ®rst 4 h of the light period,
when the total amount of MOS was decreasing rapidly
(Figures 8 and 9e), there was an increase in the amount of
maltotetraose (Figure 9c). This could re¯ect the elongation
of maltotriose by starch synthase, using photosyntheti-
96
Joanna H. Critchley et al.
cally assimilated carbon. The sums of the different MOS
detected by HPAEC-PAD (Figure 9e) were in close agreement with the total MOS contents measured enzymatically
(Figure 8) indicating that all of the MOS had been
accounted for in the HPLC analyses.
Amylopectin structure is not affected by loss of
D-enzyme
The proposal that D-enzyme plays an important role in
the synthesis of amylopectin in Chlamydomonas (see
Introduction) led us to analyse the amylopectin structure
of starch from leaves of wild-type and dpe1±1 plants.
Starch from leaves at the end of the photoperiod was
debranched using isoamylase and subjected to ¯uorophore-assisted PAGE. The chain length distributions of
wild-type and dpe1±1 starch were not signi®cantly
different (Figure 10). Similar analyses were conducted on
starch isolated from plants that had been kept in the dark
for 36 h, followed by 12 h of light (described above).
Again, no difference in the chain length distribution of the
amylopectin was detected (data not shown). Scanning
electron micrographs revealed no morphological differences between starch granules isolated from wild-type and
mutant leaves (not shown). These results provide no
evidence for a role for D-enzyme in the synthesis of normal
amylopectin in Arabidopsis leaves.
Discussion
Figure 9. Analysis of the MOS present in leaves of wild-type and dpe1±1
plants.
Samples of the neutralised perchloric acid extracts of wild-type (solid
line, open symbols) and dpe1±1 plants (dashed line, closed symbols)
described in Figure 8 were analysed by HPAEC-PAD as described in
Experimental procedures. Peak areas were determined and compared
with standard curves for each compound. Note the different scale for
each malto-oligosaccharide.
(a) Maltose content. (b) Maltotriose content. (c) Maltotetraose content. (d)
Combined maltopentaose, maltohexaose and maltoheptaose content. (e)
Total MOS content. The sum of the MOS in (a±d).
Our results establish that D-enzyme is involved in the
metabolism of MOS during starch degradation in
Arabidopsis leaves. In the leaves of the mutant lacking
D-enzyme there is a large accumulation of MOS ± primarily
maltotriose ± at the start of the dark period. This
maltotriose is likely to be largely the product of the
debranching enzymes and a-and b-amylases known to be
present in Arabidopsis chloroplasts (Lao et al., 1999;
Zeeman et al., 1998a). The activity of plastidial starch
phosphorylase in Arabidopsis leaves is probably insuf®cient to account for the rate of starch degradation (Zeeman
et al., 1998a), and the primary residual product of phosphorolysis would be expected to be maltotetraose rather
than maltotriose (Steup and SchaÈchtele, 1981). We propose that, during starch degradation in wild-type leaves,
D-enzyme catalyses the disproportionation of small MOS
molecules, particularly maltotriose, into larger ones for
further hydrolytic and phosphorolytic breakdown, yielding
glucose and glucose-1-phosphate as ®nal products.
Our results establish that D-enzyme exerts signi®cant
control over the rate of starch degradation. The rate of
degradation was lower in mutant than in wild-type leaves
both in normal diurnal conditions and in conditions in
which secondary effects had been minimised. We offer
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Function of D-enzyme in starch breakdown
Figure 10. Analysis of the chain length distribution of amylopectin from
dpe1±1 and wild type using FA-PAGE.
Starch samples were debranched with isoamylase, derivatised with the
¯uorophore APTS and subjected to gel electrophoresis in an Applied
Biosystems DNA sequencer. Data were analysed using Genescan
software. Peak areas of chains between 3 and 41 glucose residues in
length were summed and the areas of individual peaks expressed as a
percentage of the total.
(a) Molar percentages of chains of amylopectin from wild type. Three
replicate samples of de-branched, derivatised material were prepared
from bulk starch extracted from a batch of 200 plants. The values are the
means 6 standard error of these samples.
(b) Molar percentages of chains of amylopectin from dpe1±1. Samples
were prepared as described in (a).
(c) Percentage molar difference plot derived by subtracting the molar
percentages of wild-type amylopectin chain lengths (a) from those of
dpe1±1 (b). The standard errors were added together.
two possible explanations for this observation. First, the
high levels of MOS in mutant plants may inhibit the activity
of an enzyme directly involved in attacking the starch
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
97
granule, for example an amylase. Second, D-enzyme may
have a direct role in degradation of the starch granule
(Takaha et al., 1998b). It can potentially transfer linear
glucan chains from starch to an acceptor such as glucose,
creating MOS, which can be attacked either by amylases or
by phosphorylase (Takaha and Smith, 1999; Takaha et al.,
1993). It could also catalyse the release of highly soluble
cyclic glucans from starch, which could then be broken
down by hydrolysis (Takaha et al., 1996; Takaha et al.,
1998b).
Our experiments indicate that D-enzyme is not directly
involved in starch synthesis in Arabidopsis leaves. In
normal diurnal conditions, the mutant accumulated less
starch during the light period than wild type and the
amylose content of its starch was higher than that of wild
type. However, these differences were not apparent when
plants had been kept in the dark for a prolonged period to
remove starch prior to the light period. After this treatment, starch accumulated at the same rate, and had the
same amylose content, in mutant and wild-type leaves.
From these observations we propose that the differences
in starch synthesis between mutant and wild-type plants
observed in normal diurnal conditions do not re¯ect an
involvement of D-enzyme in the process of starch synthesis. Rather, they are secondary effects of differences in
starch degradation during the preceding dark period. We
suggest that the large difference in MOS content of mutant
and wild-type leaves at the start of the light period is a
likely cause of the differences in starch synthesis.
Elongation of maltotriose in mutant leaves by starch
synthases might both limit the availability of these
enzymes for synthesis of starch granules and provide
substrates for degrading enzymes, thus consuming
ADPglucose in a futile cycle. The transient increase in
maltotetraose at the start of the light period when MOS
content is declining rapidly is consistent with this suggestion. The presence of high levels of MOS during starch
synthesis in the mutant may also explain the higher
amylose content of its starch. MOS acts as a substrate
for amylose synthesis in starch granules isolated from pea,
potato (Denyer et al., 1996), Chlamydomonas (van de Wal
et al., 1998) and Arabidopsis (S.C. Zeeman and A.M. Smith,
unpublished results).
Our results show that D-enzyme is not necessary for the
synthesis of amylopectin with a normal chain-length
distribution in Arabidopsis leaves. The chain length
distribution of amylopectin was the same in starch from
wild-type and mutant leaves regardless of the growth
conditions. Similarly, in potato plants in which the activity
of D-enzyme in tubers was reduced by more than 98%
through expression of antisense RNA, starch content and
amylopectin structure were unaffected (Takaha et al.,
1998a). The situation is different in Chlamydomonas
grown on acetate under nitrogen-limited conditions. In
98
Joanna H. Critchley et al.
cells of the sta11 mutant, in which D-enzyme activity is
drastically reduced, amylopectin had a higher proportion
of short chains than amylopectin from wild-type cells
(Colleoni et al., 1999a). The mutant cells also accumulated
less starch and more MOS than wild-type cells, and the
starch had a higher amylose content. These effects led
Colleoni et al. (1999a) to propose that D-enzyme plays a
direct role in amylopectin synthesis. The differences
between the effects of the dpe1±1 and the sta11 mutations
may re¯ect the very different physiological conditions in
which starch is synthesised in Arabidopsis and
Chlamydomonas. Taken together, the phenotypes of the
two mutants indicate that D-enzyme is not necessary for
the synthesis of amylopectin, but it can, in some conditions, in¯uence the structure of amylopectin during its
synthesis.
Experimental procedures
Soluble starch synthase was assayed using the resin method
described by Jenner et al. (1994). Starch branching enzyme was
assayed according to Denyer et al. (1993). All assays were
optimised for extracts of Ws wild-type Arabidopsis with respect
to pH and all components.
D-enzyme was assayed according to Takaha et al. (1993). Leaves
were extracted in 20 mM MES, pH 6.2. The 500 ml reaction mixture
contained 50 mM MOPS, pH 6.8, 60 mM maltotriose. Reactions
were stopped by boiling and assayed for glucose. To distinguish
between D-enzyme and maltase (a-glucosidase) activities, 15 ml of
desalted extract was incubated in 200 ml containing 2 mM
maltotriose, 50 mM MOPS, pH 7.0 at 30°C for 1 h. After boiling
and centrifugation, the products were analysed by high performance anion exchange chromatography with pulsed amperometric
detection (HPAEC-PAD) on a Carbopac PA-100 column. Eluants
were 100 mM NaOH (eluant A), and 150 mM NaOH, 500 mM Na
acetate (eluant B). Gradients of the eluants were: 0±1 min, 85%
eluant A/15% eluant B; 1±22 min, a linear gradient between 85%
eluant A/15% eluant B and 60% eluant A/40% eluant B; 22±
22.5 min, a linear gradient between 60% eluant A/40% eluant B
and 85% eluant A/15% eluant B; 22.5±26 min, 85% eluant A/15%
eluant B. For each run, 20±50 ml of sample was injected onto the
column.
Plants
Seeds from 7100 Arabidopsis thaliana ecotype Wasserilewskija
(Ws) plants, which had been transformed with T-DNA (Feldmann,
1991), were kindly supplied by Tim Caspar (E. I. DuPont de
Nemours and Company, Wilmington, DE, USA). DNA was
prepared from these plants in pools of 100 for PCR screening as
described by Krysan et al. (1996). Unless otherwise stated, plants
for physiological analysis were grown at 20°C, with 75% relative
humidity, and a 12-h photoperiod. The quantum irradiance was
50 mmol m±2 sec±1 for the ®rst 3 weeks and 200 mmol m±2 sec±1
thereafter. Plants for measurement of starch and MOS were
grown as above but with an irradiance of 175 mmol m±2 sec±1.
Enzymes and biochemicals
Isoamylase, MOS standards and Starch Azure were from Sigma
(Poole, Dorset, UK), and b-amylase, maltase and pullulan were
from Megazyme (Bray, Ireland). Other biochemicals were from
Roche (Lewes, East Sussex, UK).
D-enzyme
gene isolation and characterisation
The potato D-enzyme cDNA (Takaha et al., 1993) was used as a
probe to screen the PRL2 cDNA library from Arabidopsis thaliana
L. ecotype Columbia (Nottingham Arabidopsis Stock Centre, UK).
The longest cDNA isolated (1.3 kbp) lacked the 5¢ end. This cDNA
was introduced into pTrc99A (Promega, Southampton, UK) for
expression in Escherichia coli, and the resultant 40 kDa partial Denzyme protein puri®ed by gel electrophoresis for antibody
production in rabbits. A full-length D-enzyme cDNA (2.1 kbp)
was prepared from Ws RNA using PCR with primers designed
from partial cDNA together with DPE1 DNA sequence (accession
number AB019236).
Enzyme measurements and D-enzyme product analysis
Maltase, b-amylase, starch phosphorylase, a-amylase and
pullulanase were assayed according to Zeeman et al. (1998a).
a-amylase was assayed by the Starch Azure method at pH 7.2.
Native PAGE
Native PAGE of D-enzyme was a method adapted from Colleoni
et al. (1999b). The resolving gel contained 7.5% (w/v) acrylamide
(30 : 0.8 acrylamide/bisacrylamide), 375 mM Tris±HCl, pH 8.8 and
1% (w/v) oyster glycogen. The stacking gel contained 3.3% (w/v)
acrylamide and 63 mM Tris, pH 6.8. After electrophoresis at 12 mA
and 4°C, the gel was washed twice with 40 ml of 100 mM Tris,
pH 7.0, 1 mM MgCl2, 1 mM EDTA and 1 mM DTT for 15 min then
incubated for 2 h at 37°C in 20 ml of this medium plus 5 mM
maltoheptaose. The gel was stained with 0.67% (w/v) I2 and 3.33%
(w/v) KI. Native PAGE on amylopectin- and glycogen-containing
gels was according to Zeeman et al. (1998a). Photographs of the
gels are presented.
Chromatographic separation of D-enzyme forms
Forms of D-enzyme in soluble extracts of leaves were separated
according to Lin and Preiss (1988), except that DE-52 chromatography was omitted. Proteins precipitating between 3 and 10%
(w/v) polyethylene glycol (Mw 6000) were re-dissolved and
subjected to DEAE-Sepharose chromatography (2.6 3 25 cm
column, 60 ml bed volume). Fractions containing D-enzyme
activity were pooled, concentrated and dialysed against low salt
buffer. For dpe1±1 extracts, where no D-enzyme activity was
detected, the same fractions were taken as for wild type. The
dialysed sample was subjected to Mono-Q chromatography on
a 1-ml FPLC column (Promega). Fractions were assayed for
D-enzyme activity and used for SDS-PAGE (Laemmli, 1970) and
immunoblots (Denyer et al., 1993).
Extraction and measurements of carbohydrates
For measurement of total MOS, leaf tissue (0.2±0.6 g) was
harvested and frozen in liquid N2. The frozen leaves were
powdered and extracted in 1.5 ml 0.7 M perchloric acid for
30 min on ice. After centrifugation (3000 g, 10 min, 4°C), 1 ml of
the extract was adjusted to pH 5 by the addition of 0.5±0.6 ml 2 M
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
Function of D-enzyme in starch breakdown
KOH, 0.4 M MES, 0.4 M KCl. Precipitated potassium perchlorate
was removed by centrifugation (20 000 g, 10 min, 4°C).
Glucose and sucrose were ®rst removed from the extract as
follows: a 0.5-ml sample was incubated in 0.12 M MES, pH 5.6,
containing 7.5 units b-fructosidase, 1 unit glucose oxidase and
650 units catalase, for 2 h at 32°C. The sample was then divided in
two, 1 unit of maltase and 7 units of amyloglucosidase added to
half and glucose determined after further incubation at 37°C for
2 h. MOS content, as glucose equivalents, was calculated from
the difference in glucose content between the two halves.
For the analysis of MOS by HPAEC-PAD, samples of extracts
prepared as above were applied to sequential 0.5-ml columns of
Dowex 50 W and Dowex 1. The neutral compounds were eluted
with water (Harley and Beevers, 1963), lyophilized and redissolved in water. MOS were separated by HPAEC-PAD. Peaks were
identi®ed by adding known amounts of MOS standards, and
susceptibility of the compounds to digestion by a-glucosidase
con®med. Peak areas were determined using Peaknet software
(Dionex) and compared to standard curves for known amounts of
MOS.
Analysis of starch composition and amylopectin structure
Sepharose CL2B analysis of starch was according to Denyer et al.
(1995), except that 0.35-ml fractions were collected at a rate of one
fraction per 2 min. Analysis of the chain length distribution of
amylopectin was according to Edwards et al. (1999).
Acknowledgements
We thank Tim Caspar (E.I. DuPont de Nemours and Company,
Wilmington, DE, USA) for access to T-DNA tagged plants, and
Sarah Sherson, Veronique Germain and Susan Forbes for help
with PCR screening, and Mr K. Ezaki (Ezaki Glico Co. Ltd)
for continued support. J.H.C. gratefully acknowledges the
Biotechnology and Biological Sciences Research Council
(BBSRC) for a research studentship. S.M.S. was funded by EC
contract BIO-CT96±0311 and BBSRC grant RASP 07677. S.Z. was
funded by BBSRC grant 208/D11090 and by the Gatsby Charitable
Foundation. The John Innes Centre is funded by a competitive
Strategic Grant from the BBSRC.
References
Ball, S., Guan, H.P., James, M., Myers, A., Keeling, P., Mouille, G.,
BuleÂon, A., Colonna, P. and Preiss, J. (1996) From glycogen to
amylopectin ± a model for the biogenesis of the plant starch
granule. Cell, 86, 349±352.
Colleoni, C., DauvilleÂe, D., Mouille, G. et al. (1999a) Genetic and
biochemical evidence for the involvement of a-1,4
glucanotransferases in amylopectin synthesis. Plant Physiol.
120, 993±1003.
Colleoni, C., DauvilleÂe, D., Mouille, G., Morell, M., Samuel, M.,
Slomiany, M.-C., LieÂnard, L., Wattebled, F., d'Hulst, C. and Ball,
S. (1999b) Biochemical characterization of the Chlamydomonas
reinhardtii a-1,4,glucanotransferase supports a direct function
in amylopectin biosynthesis. Plant Physiol. 120, 1005±1014.
Denyer, K., Sidebottom, C., Hylton, C.M. and Smith, A.M. (1993)
Soluble isoforms of starch synthase and starch branching
enzyme occur within starch granules in developing pea
embryos. Plant J. 4, 191±198.
Denyer, K., Barber, L.M., Burton, R. et al. (1995) The isolation and
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100
99
characterisation of novel low amylose mutants of Pisum
sativum L. Plant Cell Environ. 18, 1019±1026.
Denyer, K., Clarke, B., Hylton, C., Tatge, H. and Smith, A.M. (1996)
The elongation of amylose and amylopectin chains in isolated
starch granules. Plant J. 10, 1135±1143.
Edwards, A., Fulton, D.C., Hylton, C.M., Jobling, S.A., Gidley, M.,
RoÈssner, U., Martin, C. and Smith, A.M. (1999) A
combined reduction in activity of starch synthases II and III of
potato has novel effects on the starch of tubers. Plant J. 17,
251±261.
Feldmann, K.A. (1991) T-DNA insertion mutagenesis in
Arabidopsis: mutational spectrum. Plant J. 1, 71±82.
Harley, J.L. and Beevers, H. (1963) Acetate utilisation by maize
roots. Plant Physiol. 38, 117±123.
Jenner, C.F., Denyer, K. and Hawker, J.S. (1994) Caution on the
use of the generally accepted methanol precipitation technique
for the assay of soluble starch synthase in crude extracts of
plant-tissues. Aust. J. Plant Physiol. 21, 17±22.
Kakefuda, G. and Duke, S.H. (1989) Charactersiation of pea
chloroplast D-enzyme (4-a-glucanotransferase). Plant Physiol.
91, 136±143.
Kakefuda, G., Duke, S.H. and Hostak, M.S. (1986) Chloroplast and
extrachloroplastic starch-degrading enzymes in Pisum sativum
L. Planta, 168, 175±182.
Krysan, P.J., Young, J.C., Tax, F. and Sussman, M.R. (1996)
Identi®cation of transferred DNA insertions within Arabidopsis
genes involved in signal transduction and ion transport. Proc.
Natl Acad. Sci. USA, 93, 8145±8150.
Lao, N.T., Schoneveld, O., Mould, R.M., Hibberd, J.M., Gray, J.C.
and Kavanagh, T.A. (1999) An Arabidopsis gene encoding a
chloroplast-targeted b-amylase. Plant J. 20, 510±527.
Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature, 227,
680±685.
Lin, T. and Preiss, J. (1988) Characterisation of D-enzyme (4-aglucanotransferase) in Arabidopsis leaf. Plant Physiol. 86,
260±265.
Lin, T., Spilatro, S.R. and Preiss, J. (1988) Subcellular localisation
and characterisation of amylases in Arabidopsis leaf. Plant
Physiol. 86, 251±259.
McKinney, E.C., Ali, N., Traut, A., Feldmann, K.A., Belostotsky,
D.A., McDowell, J.M. and Meagher, R.B. (1995) Sequencebased identi®cation of T-DNA insertion mutations in
Arabidopsis ± actin mutants act2±1 and act4±1. Plant J. 8,
613±622.
Myers, A.M., Morell, M.K., James, M.G. and Ball, S.G. (2000)
Recent progress toward understanding biosynthesis of the
amylopectin crystal. Plant Physiol. 122, 989±997.
Steup, M. and SchaÈchtele, C. (1981) Mode of glucan degradation
by puri®ed phosphorylase forms from spinach leaves. Planta,
153, 351±361.
Sun, Z., Duke, S.H. and Henson, C.A. (1995) The role of the
chloroplast a-glucosidase in transitory starch degradation.
Plant Physiol. 108, 211±217.
Takaha, T., Yanase, M., Okada, S. and Smith, S.M. (1993)
Disproportionating enzyme (4-a-glucanotransferase ± EC
2.4.1.25) of potato -puri®cation, molecular cloning, and
potential role in starch metabolism. J. Biol. Chem, 268,
1391±1396.
Takaha, T., Yanase, M., Takata, H., Okada, S. and Smith, S.M.
(1996) Potato D-enzyme catalyzes the cyclization of amylose to
produce cycloamylose, a novel cyclic glucan. J. Biol. Chem,
271, 2902±2908.
Takaha, T., Critchley, J., Okada, S. and Smith, S.M. (1998a)
100
Joanna H. Critchley et al.
Normal starch content and composition in tubers of antisense
potato plants lacking D-enzyme (4-a-glucanotransferase).
Planta, 205, 445±451.
Takaha, T., Yanase, M., Takata, H., Okada, S. and Smith, S.M.
(1998b) Cyclic glucans produced by the intramolecular
transglycosylation activity of potato D-enzyme on
amylopectin. Biochem. Biophys. Res. Comm. 247, 493±497.
Takaha, T. and Smith, S.M. (1999) The functions of 4-aglucanotransferases and their use for the production of cyclic
glucans. Biotech. Gen. Eng. Rev. 16, 257±279.
van de Wal, M., d'Hulst, C., Vincken, J.P., BuleÂon, A., Visser, R.
and Ball, S. (1998) Amylose is synthesised in vitro by extension
of and cleavage from amylopectin. J. Biol. Chem. 273,
22232±22240.
Zeeman, S.C., Northrop, F., Smith, A.M. and ap Rees, T. (1998a) A
starch-accumulating mutant of Arabidopsis thaliana de®cient in
a chloroplastic starch-hydrolysing enzyme. Plant J. 15, 357±365.
Zeeman, S.C., Umemoto, T., Lue, W.L., Au-Yeung, P., Martin, C.,
Smith, A.M. and Chen, J. (1998b) A mutant of Arabidopsis
lacking a chloroplastic isoamylase accumulates both starch and
phytoglycogen. Plant Cell, 10, 1699±1711.
GenBank accession numbers AC002409 and AB019236.
ã Blackwell Science Ltd, The Plant Journal, (2001), 26, 89±100