Biochem. J. (2010) 428, 201–212 (Printed in Great Britain)
201
doi:10.1042/BJ20091777
Characterization of the AGPase large subunit isoforms from tomato
indicates that the recombinant L3 subunit is active as a monomer
Marina PETREIKOV*, Miriam EISENSTEIN†, Yelena YESELSON*, Jack PREISS‡ and Arthur A. SCHAFFER*1
*Department of Vegetable Research, Volcani Center-ARO, Bet Dagan 50250, Israel, †Chemical Research Support Unit, Weizmann Institute of Science, Rehovot 76100, Israel, and
‡Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48823, U.S.A.
The enzyme AGPase [ADP-Glc (glucose) pyrophosphorylase]
catalyses a rate-limiting step in starch synthesis in tomato
(Solanum lycopersicon) fruit, which undergoes a transient period
of starch accumulation. It has been a generally accepted paradigm
in starch metabolism that the enzyme naturally functions primarily
as a heterotetramer comprised of two large subunits (L) and two
small subunits (S). The tomato genome harbours a single gene
encoding S and three genes for L proteins, which are expressed in
both a tissue- and time-specific manner. In the present study the
allosteric contributions of the different L subunits were compared
by expressing each one in Escherichia coli, in conjunction with S
and individually, and characterizing the resulting enzyme activity.
Our results indicate different kinetic characteristics of the tomato
L1/S and L3/S heterotetramers. Surprisingly, the recombinant
L3 protein was also active when expressed alone and sizeexclusion and immunoblotting showed that it functioned as a
monomer. Subunit interaction modelling pointed to two amino
acids potentially affecting subunit interactions. However, directed
mutations did not have an impact on subunit tetramerization.
These results indicate a hitherto unknown active role for the L
subunit in the synthesis of ADP-Glc.
INTRODUCTION
However, it is not clear whether the source/sink distinction of
the regulatory properties of the different L subunits proposed
for Arabidopsis [15,16] is also applicable to plants that
have undergone significant evolutionary development for starch
accumulation in their sink tissue, such as tomato [17]. Georgelis
et al. [18] recently suggested that evolutionary constraints have
been relaxed for the AGPase L subunits, which have undergone
a high rate of evolution. Therefore it is especially likely that
the L subunits of agriculturally important crops such as potato
and tomato, that have undergone significant selection under
domestication, in addition to natural selection [19,20], may have
evolved novel modulatory characteristics. A comparison of the
regulatory properties of the different L subunits of plants with
strong starch accumulation sinks has not yet been undertaken.
Previously it had been an accepted paradigm that the eukaryotic
enzyme functions only as a heterotetramer comprised of two
S subunits, which were the sole catalytic subunits, and two L
subunits whose function was strictly regulatory [8,9]. Over the
past years, the paradigm of a strict separation of function between
the different subunits, implicating the S subunit as exclusively
catalytic and the L subunit as limited to a regulatory role, has
been challenged. Previous studies have shown that the L subunit
can be modified to perform catalysis [21–23]. Until recently, this
has only been shown in the presence of a complete heterotetramer
in which the normally catalytic S subunit was mutated to an
inactive form. The conclusion of these studies was that the L
subunit can be catalytic, but requires the S subunit for enzyme
stability, as the L subunit cannot function by itself. Most recently,
Hwang et al. [22] succeeded in characterizing the low activity of
a homotetrameric mutated potato tuber L1, indicating that an L
subunit is indeed capable of catalysis, given certain amino acid
substitutions.
AGPase [(ADP-Glc (glucose) pyrophosphorylase; E.C. 2.7.7.27)]
is the first committed step in plant starch synthesis, catalysing the
transformation of Glc-1-P to ADP-Glc. AGPase is frequently
the limiting enzyme in the pathway of starch synthesis and either
reducing its activity or increasing it has a direct impact on the
starch level in plants [1–6], including tomato [7]. Furthermore,
the enzyme exerts strong control over the rate of starch synthesis
via allosteric control of the enzyme by both small metabolites and
energy/redox status of the tissue (reviewed in [8,9]).
In general, dicots have a single gene encoding a functional S
(small) subunit, but multiple genes, generally three or four, encode
for the L (large) subunit [10]. This allows for multiple forms of
the active heterotetramer, depending on the L subunit comprising
the heterotetramer. The different L subunit genes are differentially
expressed in both a time- and tissue-specific manner. One of the L
subunit genes (referred to as L1 in tomato and potato, and APL3 in
Arabidopsis) is highly expressed in sink tissue, whereas another
(L3 in potato and tomato, APL1 in Arabidopsis) is preferentially
expressed in source tissue [11–14].
Crevillen et al. [15] and, more recently, Ventriglia et al.
[16] showed that the different L subunits of Arabidopsis confer
distinct regulatory properties to the active heterotetramer and
that these characteristics provide enzymatic plasticity adapted to
the function of the particular tissue, either source or sink. The
source tissue L subunit of Arabidopsis (APL1) conferred a higher
sensitivity to both 3-PGA (3-phosphoglyceric acid) and Pi than
did the sink-expressed L subunits, in line with the fluctuations
of these metabolites during photosynthesis, thereby effectively
serving for the fine control of carbon partitioning between sugar
and starch.
Key words: ADP-glucose pyrophosphorylase, monomer, Solanum
lycopersicon, subunit interaction, tetramer.
Abbreviations used: AGPase, ADP-Glc pyrophosphorylase; DTT, dithiothreitol; fASA, fractional solvent accessibility; Glc, glucose; IPTG, isopropyl
β-D-thiogalactoside; L, large; LB, Luria–Bertani; MS/MS, tandem MS; NIL, near isogenic line; 3-PGA, 3-phosphoglyceric acid; PGM, phosphoglucomutase;
RT, reverse transcription; S, small.
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2010 Biochemical Society
202
M. Petreikov and others
The tomato and potato are excellent models for dicot sink
starch partitioning metabolism since they both show significant
starch accumulation patterns in their developing sink tissues.
The tomato, in particular, is of interest since it is characterized
by a very temporal and transient starch accumulation period in
the young developing ovary [17]. This accumulation pattern is
accompanied by a parallel temporal expression pattern of the
sink-specific LeL1 gene in the developing fruit [7], whereas
the source leaf tissue expresses primarily the LeL3 gene [12].
In the present paper we report that the L1 and L3 subunits of the
tomato AGPase confer to the active heterotetramer distinct kinetic
characteristics, analogous to those conferred by the L subunits of
Arabidopsis. We also show for the first time that the tomato L3
protein can be functionally catalytic, even in the absence of the S
subunit protein and, furthermore, that the active L3 functions as
a monomer.
EXPERIMENTAL
Plant material
NILs (near isogenic lines) of tomato differing in the introgression
source of the distal portion of chromosome 1 and segregating
for AGPase-AgpL1 [7] served as the source of mRNA for the
cloning of the AGPase subunit genes. NILs were developed from
the interspecific cross of Solanum habrochaites (H, LA1777,
formerly Lycopersicon hirsutum [24]) and Solanum lycopersicum
(E, formerly L. esculentum) and a backcross programme to the
recurrent S. lycopersicum as described previously [7]. L2, L3 and
S genes were cloned from young green fruits of a standard tomato
breeding line 2918 and the L1 subunit gene was cloned from
its NIL 2918-L1H harbouring the AGPase-L1H allele of the wild
S. habrochaites species. The L1 cDNA was cloned from the NIL
since it is transiently expressed for a longer period during fruit
development [7]. However, sequencing of the mature native L1
protein (see below) showed that there are no sequence differences
between the mature L1 proteins of the two species.
Reagents and services
Reagents were obtained from Sigma–Aldrich and Boehringer
Mannheim. Primers were synthesized by IDT. The pGEM-T Easy
vector was purchased from Promega. Sequencing was carried out
at The Center for Genomic Technologies, The Hebrew University
of Jerusalem and at Hy Laboratories.
Cloning of cDNA and construction of plasmids
cDNA fragments encoding the four tomato AGPase subunits
proteins were obtained by RT (reverse transcription)–PCR from
total mRNA of young green fruits. RNA was isolated from three
individual tomato fruits using the EZ-RNA total RNA isolation
kit (Biological Industries), according to the manufacturer’s
instructions. RT reactions were performed as described in
[25]. For cloning cDNA fragments into the pGEM-T Easy
vector (Promega), the specific primers were designed based on
published sequences, as shown in Supplementary Table S1(a) (at
http://www.BiochemJ.org/bj/428/bj4280201add.htm).
Expression vectors pMon17335 and pMon17336, described
in [26], were used for protein expression of the cloned S
and L subunits respectively. The coding sequence for each
subunit was prepared according to the published N-terminal
sequence of potato for each of the subunits [22,24,27] as shown
in Supplementary Figure S1 (at http://www.BiochemJ.org/bj/
c The Authors Journal compilation c 2010 Biochemical Society
428/bj4280201add.htm). The appropriate restriction sites and
methionine residue were incorporated into each coding sequence
by PCR-based mutagenesis and were cloned into the pGEM-T
Easy vector cDNA using the oligonucleotides shown in
Supplemenatry Table S1(b).
Bacterial expression system
The four plasmids containing L (pMon17336-L1, pMon17336-L2
and pMon17336-L3) and S (pMon17335-S1) AGPase subunits
were expressed individually or in combinations of both L/S
subunits in Escherichia coli strain AC70R1-504, which is
deficient in endogenous AGPase activity, as described in [15,26].
For the expression of both subunits, E. coli cells already
transformed with pMon17335-S1, were converted into competent
cells using ice-cold 10 % (v/v) glycerol for cell precipitation, and
a second transformation by electroporation was performed with
the L subunits, resulting in cells that expressed both S and L
subunits. The selection of colonies was based on the kanamycinor spectinomycin-resistant genes engineered into pMon17335 or
pMon17336 plasmids respectively.
The ability of bacterial colonies to synthesize glycogen was
screened by exposure to iodine solution [28] as an indicator of
active AGPase enzyme. The bacterial colonies were grown on
solid LB (Luria–Bertani) medium in the presence of appropriate
antibiotics (50 μg/ml kanamycin for a L subunit, 70 μg/ml
spectinomycin for a S subunit, and both antibiotics for the doubletransformed S/L subunits), 2 % (w/v) glucose and 15 % (w/v)
agar at 37 ◦C overnight. The colonies were stained for 2 min in
iodine solution (0.01 M I2 and 0.03 M potassium iodide). The
glycogen production in the colonies was detected by the presence
of brown staining.
Site-directed mutagenesis of the L3 subunit
Site-directed mutagenesis of the L3 subunit was performed using
pMon17336-L3 as a template. The PCRs were carried out as
follows. (i) The first reaction included primer A (Supplementary
Table S1c) with a unique HindIII restriction site incorporated
in the forward direction and primer B (Supplementary Table
S1c) with a mutation in four nucleotides that encode for two
amino acids, N317F and A322K, in the reverse direction. The
amino acids are numbered according to the potato S subunit.
(ii) The second reaction included primer C (Supplementary Table
S1c), the reverse compliment to primer B and used in the forward
direction, and primer D (Supplementary Table S1c) with an NheI
restriction site, in the reverse direction. (iii) In the third reaction,
the purified PCR products of both (i) and (ii) served as templates
with primers A and D. The product of the third reaction was ligated
into the pGEM-T Easy vector. The 700 bp fragment with mutated
amino acids within the HindIII and NheI restriction enzyme sites
was ligated into pMon17336-L3. The new plasmid pMON17336L3M was partially sequenced for conformation of mutations.
Purification of recombinant enzymes
Single colonies were inoculated in 5 ml of LB medium with
suitable antibiotics (50 μg/ml kanamycin for the L subunit,
70 μg/ml spectinomycin for the S subunit, or both antibiotics
for double-transformed S/L subunits) at 37 ◦C overnight, with
shaking. A 2 ml volume from the overnight culture was added to
200 ml of LB medium with a suitable antibiotic concentration and
were grown at 37 ◦C until the A600 reached 0.8–1.1. The induction
of protein expression was achieved by adding 1 mM IPTG
Tomato ADP-Glc pyrophosphorylase large subunits
(isopropyl β-D-thiogalactoside) for the S subunit and 5 μg/ml
nalidixic acid was added to induce L subunit expression. For the
expression of combined S/L subunits both IPTG and nalidixic
acid were added. After 15–16 h of culture at 30 ◦C, the cells were
chilled on ice for 30 min and harvested by centrifugation for
10 min at 4000 g at 4 ◦C. The pellet was kept at − 80 ◦C until
further use.
For protein extraction the pellet was resuspended in 15 ml of
cooled extraction buffer A [20 mM NaH2 PO4 /Na2 HPO4 (pH 8),
1 mM EDTA, 0.5 mM NaCl, 10 % (w/v) sucrose, 0.1 % Triton
X-100 and freshly added 1 mM PMSF and 1 mg/ml lysozyme].
DTT (dithiothreitol) was not used during the steps of purification
in order to improve stability of the AGPase complex (A. Iglesias,
personal communication). After a 30 min incubation on ice, the
suspension was disrupted by sonication, centrifuged (13 000 g
for 20 min) and the supernatant used as a crude enzyme. All
purification steps of the recombinant protein were performed at
0 – 4 ◦C.
Crude enzyme extracts were heated and, after the temperature
reached 60 ◦C, the extracts were incubated for 4 min at 60 ◦C,
chilled on ice, centrifuged (10 000 g, 30 min) and assayed for
AGPase activity. The L3 enzyme was heat-labile and lost all
activity following heat treatment, so this step was omitted in the
partial purification of L3. Crude enzyme after, or without, the heat
treatment was filtered through a 0.2 μm filter and fractionated by
passing through a 30000 molecular mass cut-off filter (Millipore).
The 15 ml Millipore tubes were centrifuged at 5000 g for 4 ◦C,
then the concentrated high-molecular-mass fraction was diluted
with buffer B [20 mM Hepes (pH 7), 5 mM MgCl2 and 5 % (w/v)
sucrose] and loaded on to a Mono Q column HR 5/5 (Pharmacia
Biotech AB) for additional purification by HPLC, as described
in [29]. The active fractions were obtained by elution with a
0 – 0.5 M KCl gradient in buffer B and assayed for AGPase
activity in the pyrophosphorolytic direction.
The active fractions were combined and dialysed against
buffer A with the addition of 1 mM PMSF and 1 mM EDTA, and
concentrated in dialysis tubes against dry PEG [poly(ethylene
glycol)] 20000. For further purification and molecular mass
determination, the active concentrated fractions were diluted
in buffer C [50 mM Hepes (pH 7.5), 5 mM MgCl2 , 5 % (w/v)
sucrose, 1 mM EDTA and 0.5 M KCl] and 0.1 ml was loaded
on to a Superdex-200 column (HR 10/300 mm; Sigma, Supelco)
for size-exclusion chromatography. Elution was at a rate of
0.5 ml/min with buffer C and 0.5 ml fractions were collected and
assayed for AGPase activity in the pyrophosphorolytic direction.
The molecular mass estimation of the eluted fractions was
calculated based on a calibration curve of the elution times
of standard proteins: blue dextran, β-amylose (200 kDa), BSA
(66 kDa) and cytochrome c (12.4 kDa).
Electrophoresis and immunoblotting
Protein of partially purified AGPase from Mono Q and
Superdex-200 chromatography was concentrated by acetone
precipitation and subjected to SDS/PAGE in a Bio-Rad minielectrophoresis system [13 % (w/v) 0.75 mm thick acrylamide
gel] according to methods described by Laemmli [30]. For
immunoblotting, specific antibodies against AGPase individual
L and S subunits, raised against purified proteins from potato
tubers, were kindly supplied by Dr T.W. Okita (Department of
Biochemistry/Biophysics and Institute of Biological Chemistry,
Washington State University, Pullman, WA, U.S.A.) [31] and
antibodies raised against both subunits from tomato fruit were
kindly supplied by Dr H. Janes (Department of Plant Biology and
203
Pathology, Rutgers University, New Brunswick, NJ, U.S.A.), as
described in [7]. The antibodies were diluted 1:5000 and 1:2000
respectively. The membranes were incubated overnight at 4 ◦C
with shaking. Precision Plus protein standards (Bio-Rad) were
used as markers for estimating the size of the separated proteins.
Bands were visualized using BCIP (5-bromo-4-chloroindol-3-yl
phosphate)/NBT (Nitro Blue Tetrazolium) (both from Promega)
according to the manufacturer’s instructions.
N-terminal and MS/MS (tandem MS) peptide sequencing
Since there are amino acid sequence differences near the
N-terminus between the L1 proteins of the cultivated tomato and
the wild species [32], which could have an effect on the enzyme,
we were interested in ascertaining the mature protein sequence
of the tomato fruit AGPase enzyme L subunit (L1), in order
to exclude these sequence differences. AGPase was extracted
from green tomato fruit as in [7] and purified by Mono Q and
Superdex-200 chromatography, as described above. The active,
desalted and concentrated enzyme was separated by SDS/PAGE
electrophoresis (as described above), blotted on to a PVDF
membrane (Immobilon-P, Millipore) in 10 mM Caps (pH 11) and
7.5 % (v/v) methanol transfer buffer, and the protein band was
visualized by Coomassie Blue staining. N-terminal sequencing
was performed on an Applied Biosystems Procise Protein
sequencer at the Protein Analysis Unit, Biological Services,
Weizmann Institute of Science (Rehovot, Israel). MS/MS peptide
sequencing was performed at Smoler Proteomic Research Center,
Department of Biology, Technion (Haifa, Israel). The samples
[1 mm thick, from a 13.5 % (w/v) SDS/PAGE Coomassie-Bluestained gel] were digested with trypsin and analysed by LCMS/MS (liquid chromatography-MS/MS) on DECA/LCQ and
identified by Pep-Miner and Sequest software against the nonredundant database. The peptides were considered as high quality
with an average Pep-Miner identification score of 97 and the
average Sequest Xcore of 4.1 for doubly charged peptides.
AGPase assay and kinetic characterization
AGPase activity was analysed essentially as previously described
[7], as below.
Pyrophosphorolytic direction
Assay 1. AGPase activity in the determination of crude activity
and for monitoring HPLC purification was assayed in the
pyrophosphorolytic direction, as described previously [7] with
minor modifications. The assay mixture contained 50 mM HepesNaOH (pH 7.8) with 5 mM MgCl2 (buffer A), 1 mM ADP-Glc,
1 mM NAD, 10 mM 3-PGA, 5μM Glc-1,6-bis-P, 10 mM freshly
prepared sodium fluoride, 1 unit of Glc-6-P dehydrogenase (from
Leuconostoc), 2 units of PGM (phosphoglucomutase) and up to
0.1 ml of enzyme extract in 0.5 ml of reaction mixture. Following
5 min incubation at 37 ◦C, the reaction was initiated by the
addition of 1 mM PPi . Preliminary analysis showed that enzyme
activity was linear for a minimum of 10 min. The production of
NADH was monitored spectrophotometrically at 340 nm. One
unit of pyrophosphorolytic activity was defined as the amount of
enzyme which catalyses the formation of 1 nmol of Glc-1-P/min.
Assay 2. For kinetic analysis in the pyrophosphorolytic direction
a two-step enzyme-linked endpoint assay was used. The Glc-1-P
production was measured in buffer A at 37 ◦C with 1 mM ADPGlc, 10 mM 3-PGA, 10 mM NaF and 20 μl of enzyme extract
c The Authors Journal compilation c 2010 Biochemical Society
204
M. Petreikov and others
in a 200 μl reaction mixture. For kinetic analysis of the L3
enzymes the 3-PGA concentration was 1 mM due to inhibition
at higher concentrations. The reaction was initiated by 1 mM PPi
and stopped after 5 min by boiling for 2 min. Preliminary analysis
showed that the enzyme activity was linear for a minimum of
10 min. After cooling on ice, a mixture containing 300 μl of
buffer A, 1 mM NAD, 10 μM Glc-1,6-bis-P, 1 unit of Glc-6P dehydrogenase (from Leuconostoc) and 1 unit of PGM was
added. After 40 min incubation at 37 ◦C, absorbance of the NADH
product was recorded at 340 nm. Concentrations from 0 to 1 mM
of the substrates ADP-Glc and PPi and the activator 3-PGA
were compared. The amount of Glc-1-P that was produced was
quantified from a standard curve of 0–100 nmol of Glc-1-P in
a 0.5 ml reaction mixture under the same assay conditions and
was expressed as an amount of enzyme necessary to produce 1
μmol of Glc-1-P per min at 30 ◦C. For comparison of enzyme heat
stability, Assay 2 in the pyrophosphorolytic direction was used.
After the reaction mixture, including the enzyme, was heated and
reached 60 ◦C, it was incubated at 60 ◦C for 4 min. Following
cooling and centrifugation at 4 ◦C (13 000 g), the enzyme activity
was determined as the amount of Glc-1-P produced, as above.
ADP-Glc synthesis direction
Assay 3. In the synthesis direction, the enzyme activity was
analysed as described in [33], with slight modifications. The assay
contained buffer A, 1.5 mM Glc-1-P, 3 mM ATP, 5 mM 3-PGA
and 20 μl of enzyme extract in a 200 μl reaction mixture. For
kinetic analysis of the L3 enzymes, the 3-PGA concentration was
1 mM due to inhibition at higher concentrations. After incubation
at 37 ◦C for 10 min the reaction was stopped by boiling for 2 min.
The ADP-Glc product was quantified via a HPLC reversed-phase
Adsorbosphere Nucleotide-Nucleoside column, as described in
[7].
For studies of kinetic characteristics of the purified enzymes
the substrates ADP-Glc and PPi (pyrophosphorolytic direction)
and Glc-1-P and ATP (synthesis direction) were varied at
concentrations from 0 to 5 mM and the activator 3-PGA (both
directions) was varied from 0 to 10 mM. Sm and A0.5 values
corresponding to the levels required for 50 % maximal activity
were calculated using Lineweaver–Burk plots and the regression
plot Excel program 2003. The direct linear plot method [34] was
also used for estimation of kinetic constants.
Molecular modelling and structure analysis
Model structures of the tomato L1 and L3 monomers were
constructed by homology with the X-ray structure of S [35]. The
sequence identity between the template and the models is very
high. Only a few short inserts/deletions are observed, all of which
map to loops in regions away from the tetramerization interfaces,
rendering the models accurate. Notably, the loop corresponding to
residues 90–98 of S, which is not resolved in the X-ray structure,
was not modelled. Superposition of the modelled L1 on to the S4
tetramer produced a model of the S2 /L12 tetramer. This structure
was used in the solvent accessibility computations as described in
the Results section. The InsightII molecular modelling package
(Accelrys) was used for the homology modelling, structure
verification and analysis and for the preparation of Figure 6.
Protein estimation
Protein was estimated as described in [36].
c The Authors Journal compilation c 2010 Biochemical Society
Statistical calculations
The estimation of S.E.M.s and significant statistical differences
were calculated using Excel and JMP statistical discovery
software (version 5.0; SAS Institute).
RESULTS
AGPase subunit expression and activity
The four tomato AGPase subunit genes were individually prepared
in plasmids and expressed in an E. coli mutant strain lacking
native activity. Each of the L subunits was expressed together with
the S subunit, and each of the four subunits was also expressed
individually. The coding sequence for each subunit was prepared
according to the published N-terminus of the potato protein for
each of the subunits (Supplementary Figure S1A). In addition,
since the L1 protein of the wild species S. habrochaites contains
a three amino acid duplication (EKKEKK, positions 47–52 in
Supplementary Figure S1B) near the N-terminus, we wanted to
ascertain that this duplication was not present in the mature L1
protein. The tomato fruit enzyme from the NIL harbouring the
AGPase-L1H allele was purified and the L subunit protein (L1)
was sequenced by both N-terminal sequencing as well as MS/MS
peptide sequencing. The N-terminus of the L1 began at the amino
acids VITTEN (position 61), eight amino acids downstream from
the duplication. The MS/MS sequences showed 80 % coverage
of the protein sequence with no amino acids further upstream
than the VITTEN sequence (Supplementary Figure S1B).
Accordingly, the cloned L1 sequence encoded three amino
acids (VAY, from position 57) upstream of the mature protein
N-terminus, but downstream of the EKK sequence.
Expression of the subunits in E. coli, individually and in
the various L/S combinations, followed by potassium iodide
staining for glycogen, showed that the co-expression of the
S protein and the different L proteins yielded active enzymes
(Supplementary Figure S2 at http://www.BiochemJ.org/bj/428/
bj4280201add.htm) and that the L1/S and L3/S combinations
were more active than the L2/S combination. Engineered E. coli
that expressed only individual subunits, whether S or L, did not
show visible activity following potassium iodide staining.
The protein extracts of each of the engineered E. coli
strains, including those which did not show visible activity as
determined by potassium iodide staining, were partially purified
and assayed for in vitro activity. The L1/S and L3/S enzymes
were studied in further detail since they were most active and represent the predominant native enzymes of tomato sink and
source respectively. Surprisingly, the L3 protein alone showed
measurable in vitro activity, but only if not heat-treated, which
is a standard purification step for the potato and tomato AGPase
enzymes [17,37]. Extracts from non-transformed E. coli, as well
as from E. coli transformed with only L1, L2 or S, did not show in
vitro activity, indicating that the activity in the L3-transformed E.
coli was not an artefact of contamination. Therefore the L3 extract
was partially purified only by anion-exchange chromatography
and size-exclusion chromatography, whereas the L1/S and L3/S
enzymes were partially purified by heat treatment prior to anionexchange and size-exclusion chromatography. The purification
schemes led to a 917- and 537-fold increase in activity of the L1/S
and L3/S proteins respectively, and a 143-fold increase for the L3
protein (Table 1). The calculated increase in specific activity may
partially be a function of the buffer change, from a Pi buffer
initially used in the extraction and which may reduce measurable
activity, to a Hepes buffer used for the chromatographic step.
The active L1/S and L3/S enzymes eluted in the same fraction
during their individual ion-exchange separations, whereas the L3
Tomato ADP-Glc pyrophosphorylase large subunits
Table 1
E. coli
205
Partial purification of enzymes L1/S, L3/S and L3 expressed in
Activity was not determined in the step following the molecular mass cut-off step prior to Mono
Q (see the Experimental section), and the data following Mono Q represents the purification
from both steps.
Purification steps
L1/S
Crude
60 ◦C
Mono Q
Superdex
L3/S
Crude
60 ◦C
Mono Q
Superdex
L3
Crude
Mono Q
Superdex
Activity (nmol/mg of protein per min)
Fold-increase in activity
9.6
107.1
3874.5
8780
1
11
405
917
18.5
148.1
5204.5
9959
1
8
281
538
1.0
58.6
143
1
59
143
Figure 1 HPLC ion-exchange (Mono Q) separation of AGPase-subunit
proteins L1/S, L3/S and L3
AGPase activity was measured in the pyrophosphorolytic direction. 䊐, AGPase-L1/S protein; ,
AGPase-L3/S protein; 䉬, AGPase-L3 protein. Elution was carried out with a 0–0.5 M potassium
chloride gradient. The Figure is a composite of the individual separations of the three protein
extracts under identical conditions.
protein eluted earlier, at a lower salt concentration (Figure 1).
Furthermore, the separation of the enzymes via Superdex 200
HR size-exclusion chromatography indicated that the L1/S and
L3/S enzymes were tetramers eluting at sizes of approx. 180 and
210 kDa respectively (Figures 2A–2C). In contrast, the fractions
from the L3 extract showing AGPase activity eluted at ∼ 42 kDa,
indicating that the enzyme in these fractions were active as
monomers (Figure 2D).
In order to further ascertain that the L3 extract consisted only
of L3, and not S, protein, the L3 and L3/S extracts following sizeexclusion were immunoblotted using antibodies against tomato
AGPase that recognize both L subunits and S subunits. First,
the results showed that the L3 protein is a major component
of the partially purified extract and that it did not contain any
S contaminants (Figure 3A). This is further indicated by the
immunoblot of the different enzymes following their separations
on Mono Q, which shows that only the L1/S and L3/S enzymes
yield the dimer characteristic of the S interactions under a low
reducing environment (Figure 3B). Furthermore, comparative
immunoblotting of the individual fractions from size-exclusion
showed that for the L3/S enzyme (Figure 4A), both L and
S proteins were present in the large-molecular-mass fractions
Figure 2 Molecular mass determination of partially purified AGPase
enzymes L1/S, L3/S, L3 and L3mut
The enzymes were applied to Superdex-200 HR chromatography. (A) Molecular mass (MW)
calibration curve (standards used for calibration were: 1, cytochrome c (12.4 kDa); 2, BSA
(66 kDa), 3, β-amylase, sweet potato (200 kDa); and calculated molecular masses of 4, L1/S;
5, L3/S; and 6, L3. (B) Separation of L1/S; (C) separation of L3/S; (D) separation of L3; (E)
separation of mutated L3 (L3mut ). The arrows in (B–E) indicate the point of maximum elution of
the protein. The separation was carried out at 0.5 ml/min and 0.25 ml fractions were collected.
(∼ 200 kDa, fractions 55–60), but were absent in the smallmolecular-mass fractions (∼ 50 kDa, fractions 66 –70), whereas
the L3 enzyme (Figure 4B) showed immunoreaction only in the
small molecular mass fractions.
In order to shed light on why L3 forms active monomers,
whereas reports in the literature indicate that L1 [22] and S [27,35]
can form homotetramers, we compared the sequences of the three
proteins in order to identify particular amino acids that are similar
c The Authors Journal compilation c 2010 Biochemical Society
206
M. Petreikov and others
Table 2 (fASA) values for the amino acids participating in the A:B subunit
interactions
Amino acid alignment between the tomato subunits of AGPase indicating the regions of subunit
interactions, based on Jin et al. [35]. The Table indicates the conservation of amino acids at
these positions, corresponding to the alignment with the small subunit sequence. Amino acids
in bold and underlined are those where the L3 amino acid is significantly different from both the
S and L1 amino acids. The mutated amino acids in L3 are indicated by asterisks. The column N
is the amino acid position according to the S sequence.
Figure 3 Western blot analysis of L and S proteins from partially purified
enzymes L1/S, L3/S and L3
Separation under reducing (A) or non-reducing (B) conditions and immunological detection of
proteins. (A) Both subunits are present in the L3/S extract, but only the L subunit is present
in the L3 extract. Cm L3, Coomassie Blue stain of the partially purified L3 extract used for
immunoblotting, indicating that the L3 protein is the major protein in the extract. (B) The dimer
of S proteins (100 kDa) formed under non-reducing conditions (in the absence of DTT) for
L1/S and L3/S, but not L3. Separation was completed on SDS/PAGE [13 % (w/v) gel, 0.75 mm
thick, 30 μg of protein/well]. The immunological reaction was carried out with antibodies raised
against tomato fruit AGPase as described in the Experimental section. Molecular mass markers
(150 kDa, 100 kDa, 75 kDa and 50 kDa) are indicated.
Figure 4 Immunodetection profile of L3/S, L3 and L3mut enzymes eluted
from a Superdex-200 column
Fractions eluted from HPLC-Sephadex 200 were combined into three pools (represented in
Figure 2): fractions 56–60 (large molecular mass, ∼ 200 kDa), 61–65 and 66–70 (small
molecular mass, ∼ 50 kDa). S/L3 (upper panel), L3 (middle panel) and mutated L3 (L3mut ; lower
panel). Pooled fractions were separated on SDS/PAGE [13 % (w/v), 1 mm, non-reducing gel],
and immunobloted with antibodies raised against potato tuber AGPase. Separation was carried
out at 0.5 ml/min and 0.25 ml fractions were collected.
c The Authors Journal compilation c 2010 Biochemical Society
N
S
L1
L3
(fASA)
312
313
314
315
316
317
318
319
320
321
322
323
324
326
327
328
329
330
331
332
333
Y
T
Q
P
R
Y
L
P
P
S
K
M
L
A
D
V
T
D
S
V
I
Y
T
S
P
R
F
L
P
P
T
K
I
D
C
K
I
K
D
A
I
I
Y
T
S
R
R
N*
L
P
P
S
A*
I
D
S
K
I
V
D
S
I
I
0.14
0.18
<0.10
<0.10
0.21
0.55
0.18
0.36
0.70
0.28
0.27
0.47
0.42
0.31
<0.10
0.42
0.22
0.21
0.23
0.23
0.32
in L1 and S, but that differ in the L3 (Table 2). We especially
searched for amino acids that have been shown by recent structural
studies [35,38] to be involved in subunit interactions. The subunit
interactions described for the S homotetramer (α 4 , [35]) were used
to model the α 2 β 2 (heterotetramer of S together with L1 or L3)
and β 4 (homotetramers of L1 or L3) interactions. The terminology
used for the α 4 interactions (A:B, parallel β-helix and linker
loop interactions; A:B , interactions in the region of the catalytic
domain [33]) was transcribed to A:B and A:D respectively. We
modelled the effect of the sequence differences between the
different subunits, particularly on the interface interactions. The
residues which contribute to the A:B and A:D interactions are
listed in Table 2. These are residues that change their fASA
(fractional solvent accessibility) upon tetramerization by more
than 10 %. The fASA indicates which fraction of the total surface
area of a given residue is accessible to the solvent and a change
in fASA (fASA) indicates that the given residue interacts with
an adjacent subunit. Notably, the available structures of the S
tetramer of potato tuber AGPase, PDB codes 1YP2, 1YP3 and
1YP4, are incomplete and in each monomer 5–9 amino acids in
the loop 90–98 are not resolved. This loop is at the centre of
the tetramer and it may contribute to the intersubunit interaction.
Indeed, the corresponding loop in the recent structure of bacterial
AGPase, PDB code 3BRK, is resolved and participates in the
A:D and A:C interactions. However, the sequence of this loop is
significantly different in the two structures.
The A:B interface includes 28 residues whose fASA
exceeds 10 %. Of significance to the present study are those
amino acids that are conserved or similar between S and
L1, the two subunits that form active homotetramers, but that
are functionally different in the L3, which forms an active
monomer. These differences are limited to only two: a change
of the aromatic Tyr(Phe)317 to asparagine (with a fASA of
0.55), and the change from the basic Lys322 to neutral alanine
Tomato ADP-Glc pyrophosphorylase large subunits
Table 3 Kinetic and allosteric properties of partially purified enzymes L1/S,
L3/S and L3
(a) Properties measured in the pyrophosphorolytic direction. Substrate affinities and maximal
activities were determined in the presence of 3-PGA concentrations of 10 mM for L1/S and L3/S
enzymes, and 1 mM for the L3 enzyme. Values are means of at least four individual reactions
(S.E.M.). V max values presented are those derived from the Lineweaver–Burk plot of the 3-PGA
activation experiment, using 1 mM ADP-Glc and 1 mM Pi for all three enzymes. V o represents
the activity measured in the absence of 3-PGA. Inhibition by Pi was measured with 1mM Pi
in the presence or absence of 3-PGA. (b) Properties measured in the synthesis direction.
Substrate affinities and maximal activities were determined in the presence of 3-PGA
concentrations of 10 mM for L1/S and L3/S enzymes, and 1 mM for the L3 enzyme. Values are
means for two individual reactions in the synthesis direction and S.E.M.s are not presented. V max
values are derived from the activity determined from the Lineweaver–Burk plot using 3 mM ATP.
V o represents the activity measured in the absence of 3-PGA. Inhibition by Pi was measured
with 1 mM Pi in the presence or absence of 3-PGA. The L3 enzyme showed an increase (inc )
in activity in the presence of 1 mM Pi of approx. 30 %.
(a)
Properties
L1/S
L3/S
L3
ADP-Glc (S 0.5 mM)
PPi (S 0.5 mM)
V max (μmol/mg of protein per min)
3-PGA (A 0.5 mM)
Pi inhibition (%) at 0 mM 3-PGA
Pi inhibition (%) at 1 mM 3-PGA
3-PGA activation ratio (V max /V 0 )
0.45 (0.05)
0.08 (0.02)
3.15 (0.21)
0.14 (0.02)
45 (2.5)
10 (5)
2.1 (0.1)
0.72 (0.03)
0.12 (0.04)
4.00 (0.17)
0.07 (0.01)
70 (5)
20 (4)
2.3 (0.3)
1.57 (0.2)
0.41 (0.04)
0.31 (0.02)
0.66 (0.22)
5 (0.1)
2 (0.1)
1.3 (0.1)
0.34
0.29
0.49
0.40
99
91
>100
0.50
0.46
0.30
0.02
95
99
4.6
3.6
2.1
0.006
1.8
inc
inc
2.8
(b)
ATP (S 0.5 mM)
Glc-1-P (S 0.5 mM)
V max (μmol/mg of protein per min)
3-PGA (A 0.5 mM)
Pi inhibition (%) at 0 mM 3-PGA
Pi inhibition (%) at 1 mM 3-PGA
3-PGA activation ratio (V max /V 0 )
Figure 5
207
(with a fASA of 0.27). A third potentially significant
change, from the basic Lys(Arg)340 to neutral threonine has a
fASA of only 0.05. The A:D interface showed no amino
acid differences with significant fASA (Supplementary Table
S2 at http://www.BiochemJ.org/bj/428/bj4280201add.htm). As
mentioned above, the unresolved loop may contribute to this
interface. This loop is of the same length in S and L3, and one
residue shorter in L1. Interestingly, the potato L3 does have a
potentially important amino acid change in this region, an acidic
Glu(Asp)99 to serine; however, this is true only for the potato
protein, whereas the tomato L3 retains the acidic aspartate residue
(see Supplementary Table S2).
In order to test whether these two amino acid differences,
Tyr(Phe)317 to asparagine and Lys322 to alanine, play a role in
determining the subunit interactions leading to tetramerization, we
mutated the tomato L3 protein at these two sites. The L3 mutant
harbouring N317Y and A322K accordingly resembles the S and
L1 proteins at these A:B interface amino acids. The expressed
mutated protein showed activity similar to the non-mutated
protein in the crude extract prior to chromatographic separation.
However, the mutated protein did not show activity following sizeexclusion chromatography. Nevertheless, separation of the crude
extract on size-exclusion chromatography (Figure 2E) followed
by immunoblotting (Figure 4, lower panel) allowed us to test
whether the expressed protein was eluted as a monomer or
tetramer. The results clearly indicate that the mutated L3 also
does not polymerize and the expressed L3N317Y/A322K protein is
eluted in the small size fractions, similar to the non-mutated L3
protein (Figure 4, middle panel).
Enzyme kinetics
The partially purified L1/S and L3/S enzymes were characterized
with respect to substrate affinity and sensitivity to the effector
molecules 3-PGA and Pi . Results show that in both the synthesis
and pyrophosphorolytic direction, L3/S was more sensitive to
3-PGA activation profiles of partially purified AGPase enzymes
Enzyme reactions were performed in the pyrophosphorolytic (A–C) and synthesis (D–F) directions at increasing 3-PGA concentrations under saturating (S/L1 and S/L3) or nearly saturating (L3)
substrate concentrations: 1 mM ADP-Glc and 1 mM PPi in the pyrophosphorolytic direction, and 3 mM Glc-1-P and 3 mM ATP in the synthesis direction. Values are means for a minimum of four
individual reactions in the pyrophosphorolytic direction and of two individual reactions in the synthesis direction. 䊐, AGPase-S/L1 protein; , AGPase-S/L3 protein; and 䉬, AGPase- L3 protein.
c The Authors Journal compilation c 2010 Biochemical Society
208
M. Petreikov and others
3-PGA activation than was L1/S. In the synthesis direction, L3/S
exhibited significantly lower (one twentieth) A0.5 values for 3PGA (Table 3 and Figure 5). However, while the A0.5 values were
lower for the L3/S (0.02 mM compared with 0.40 mM for L1/S) in
the synthesis direction, the maximal activation of L1/S was higher
than that of L3/S, and at >1 mM 3-PGA the maximal activity of
L1/S was higher than that of L3/S. In fact, the activation ratio
(the fold-increase of activity due to 3-PGA, compared with the
activity in the absence of 3-PGA) in the synthesis direction for
L1/S was at least more than 20-fold that for L3/S. This was due
to the near complete dependence of L1/S on 3-PGA where, in
the absence of 3-PGA, there was less than 1 % of the maximal
activity observed with 10 mM 3-PGA (Figure 5D). In contrast,
the L3/S showed approx. 20 % of its maximal activity even in
the absence of 3-PGA (Figure 5E). With regard to substrate
affinities (Table 3), the L1/S showed slightly higher affinities
in the synthesis direction for both ATP and Glc-1-P, as well as
in the pyrophosphorolytic direction for the substrates ADP-Glc
and PPi . With regard to Pi inhibition, L1/S was less sensitive than
L3/S in the pyrophosphorolytic direction, but both were extremely
sensitive to > 90 % in the synthesis direction, irrespective of the
addition of 3-PGA.
The L3 protein showed enzyme kinetic characteristics distinct
from both the L1/S and L3/S enzymes. It was catalytically weak
compared with the heterotetramers, and in the synthesis direction
had a 5- and 10-fold lower affinity for Glc-1-P than the L1/S
andL3/S heterotetramers respectively, and a very low affinity for
ATP (Table 3b). The L3 enzyme showed Michaelis kinetics for
both ADP-Glc and PPi in the pyrophosphorolytic direction, and
moderate affinities were observed for ADP-Glc and PPi (Table 3a).
The data of apparent maximal activity for the three enzymes
in the synthesis direction are derived from assays carried out
with 3 mM ATP, which is non-saturating for L3, thereby leading
to a calculated maximal activity of only 5 % that of the L3/S
heterotetramer. L3 did not show Michaelis kinetics for ATP, but
rather a sigmoidal response to substrate concentration (nH = 1.8),
and we did not observe signs of saturation when measured with
up to 5 mM ATP. In addition, the L3 enzyme was particularly
heat-labile, completely losing activity at 60 ◦C.
The L3 monomer was less sensitive to 3-PGA activation when
measured in either direction. Furthermore, the maximal increase
in activity of L3 due to 3-PGA was slight (∼ 25 % increase due
to 3-PGA in the phosphorolytic direction; Figure 5D), and at
concentrations above 1 mM 3-PGA we observed inhibition of
activity in the synthesis direction (Figure 5F). In the synthesis
direction, maximal 3-PGA activation was 2.8-fold at 4 mM
3-PGA, compared with 4.6-fold for L3/S and over 100-fold
for L1/S. With regard to inhibition by Pi , the L3 enzyme was
practically insensitive to Pi inhibition in either direction. In fact, in
the synthesis direction Pi increased stability and activity of the L3
monomer by approx. 30 %, in contrast with the heterotetrameric
enzymes which were strongly inhibited.
DISCUSSION
Effect of the different AGPase L subunits of tomato on
heterotetrameric enzyme activity and kinetics
The results of the present study, comparing the effects of the
different tomato L subunits on the activity of the heterotetramer,
indicate that the different sink and source L subunits of plants with
highly evolved sink starch metabolism confer analogous effector
activation characteristics as those conferred by the different
Arabidopsis L subunits [15,16]. The tomato L3/S showed higher
sensitivity to the allosteric effectors 3-PGA, as determined by
c The Authors Journal compilation c 2010 Biochemical Society
A0.5 values, and Pi than did the sink-type L1/S, analogous to the
differences reported for the respective Arabidopsis orthologues
[15,16]. Comparison of native tomato leaf and fruit AGPase
enzymes showed analogous differences in effector sensitivity
[39,40]. The similarity between the effects of PGA in response
to the different L subunits of Arabidopsis and tomato was not
obvious. While APL1 is primarily expressed in Arabidopsis
source leaves [41], the sink tissue which expresses the Arabidopsis
APL3 is the starch sheath cell layer adjacent to the vascular
bundles of the leaf, as well as developing seeds [42]. The tomato
fruit and potato tuber, on the other hand, are examples of more
extreme starch storage sink tissue and the observation that the
tomato fruit L subunits also impart distinct characteristics to
the strict fruit sink enforces the conclusion that these kinetic
characteristics indeed are distinguishing between sink and source
enzymes. Although our studies were carried out on the tomato
enzymes, the results can directly be extrapolated to include the
potato L subunit, since the orthologues of these related Solanum
species are nearly identical.
The analogous effects of the orthologous L subunits on
effector activation do not extend to the kinetic characteristics
of the apparent affinities for substrates of the reaction. In
tomato the apparent affinities for substrates were slightly higher
for the L1/S, compared with the L3/S, in contrast with the
moderate, yet opposite, differences in affinity conferred by
the Arabidopsis sink and source enzymes [15].
Focusing on the synthesis direction, which has been shown
to be the predominant, if not sole, physiological activity of the
enzyme [43,44], we observed the near complete dependence of
the tomato L1/S on 3-PGA. In the absence of 3-PGA there was
less than 1 % of the maximal activity, as compared with the L3/S
enzyme, which was more modestly activated (4.6-fold). Chen and
Janes [39] similarly reported that only negligible levels of AGPase
activity were measured in tomato fruit extracts in the absence of 3PGA and the potato tuber enzyme also shows near dependence on
PGA (∼ 30-fold activation) [27]. In light of this strong dependence
of ADP-Glc synthesis on 3-PGA levels in the tomato fruit and
potato tuber, the physiological level of 3-PGA compartmentalized
in the sink plastids is of major significance. The photosynthesizing
chloroplast of a source leaf is exposed to 3-PGA levels that are
indicative of the rate of photosynthesis and carbon balance [8,9].
However, whereas the levels of phosphorylated metabolites in the
autotrophic leaf undergo numerous fluctuations, studies of the
strictly heterotrophic potato tuber indicate that these levels do not
fluctuate as much, and generally only in response to environmental
perturbations [45]. The heterotrophic amyloplast is dependent on
the cytosol for 3-PGA. Evidence has accrued that the source for
starch synthesis in heterotrophic amyloplasts is Glc-P transported
from the sink cytosol and that the 3-PGA concentration of the
cytosol will have an impact on 3-PGA levels in the sink plastid,
thereby modulating AGPase activity [10,46].
However, in contrast with the strictly heterotrophic amyloplasts,
the tomato fruit plastids, at the period of transient starch
accumulation in the young green fruit, are functional chloroplasts
[47,48] reported to contribute ∼ 15 % to its own metabolite
pool [49]. They have also been shown to contain both triose
phosphate and hexose-P translocators [50], as well as plastidic
FBPase (fructose-1,6-bisphosphatase) activity, thereby permitting
starch synthesis from triose phosphate in the plastid [48,51–53].
Accordingly, the plastidic 3-PGA in the tomato fruit may be
derived from two sources: (i) transported triose-P, followed by
the actions of the plastidic forms of glycolytic enzymes GAPDH
(glyceraldehyde-3-phosphate dehydrogenase) and 3-PGAK (3phosphoglycerate kinase) [54], and (ii) imported Glc-P and its
plastidic glycolytic metabolism.
Tomato ADP-Glc pyrophosphorylase large subunits
Unfortunately, little is known regarding the physiological
levels of 3-PGA in tomato plastids. 3-PGA concentrations
in young tomato fruit pericarp have been estimated to be
15 nmoles per gram of fresh weight [53] and ∼ 55 nmoles
per gram of fresh weight [7], of similar magnitude to 3-PGA
concentrations in potato tubers: ∼ 20 [55], ∼ 80 [44] and ∼ 100
nmoles per gram of fresh weight [56]. However, these numbers
tell us little regarding the compartmentalized concentration in
the plastid, as these levels were determined on whole tissue
extracts. Farre et al. [44] calculated the subcellular metabolite
distribution of potato tubers, including the concentration of
3-PGA in tuber plastids. They calculated a plastidial concentration
of 0.46 mM 3-PGA (in the range of the A0.5 value of L1/S)
in the plastid, but only if assuming unrestricted diffusion
within the plastid organelle. In the case of restricted diffusion,
3-PGA concentrations may be locally increased in the immediate
environment of AGPase to 10-fold that concentration, or 4.6 mM.
Thus, depending on the compartmentation of 3-PGA within the
plastid, physiological 3-PGA concentrations will either be in the
range of the A0.5 for 3-PGA activation, and AGPase activity will
be allosterically controlled by small changes in the effector, or
3-PGA concentrations will be such that AGPase will be maximally
activated, and starch synthesis will be dependent completely on
substrate availability.
L3 is active as a monomer
The present study is the first report of an active L subunit
monomer and only the second report of an active L subunit when
expressed in the absence of an S subunit. Until recently, AGPase
activity was not detected for L subunits expressed alone and it had
been assumed that the L subunit lacks catalytic activity altogether,
serving a purely regulator role in AGPase activity [15,16,23,27].
However, the realization that the L subunit can be catalytic is not
entirely surprising, in hindsight, since the catalytic, substrate binding and effector sites are generally conserved between the L and
S subunits [9,21,23,35]. Ballicora et al. [21] showed that only two
mutations in the ATP-binding site of the potato L1, Lys44 /Thr54 ,
to the strongly conserved Arg44 and Lys54 , were sufficient to
‘restore’ the catalytic activity of L1. Arg44 and Lys54 are generally
conserved, not only in catalytic S subunits but also in the other
L subunits, including the tomato L3. In fact, Ventriglia et al. [23]
showed that the other two L subunits in Arabidopsis, which have
the conserved Arg44 and Lys54 (APL1 and APL2, homologous
with L3 and L2 in potato and tomato), can be active. However,
these previous studies indicated that for the L subunit to be
catalytic, whether a mutated L1 [21] or a natural APL1 and APL2
[23], the presence of the S protein, even if inactive, was required.
Recently, Hwang et al. [22] succeeded in showing activity
of the L subunit alone, in the total absence of the S protein.
They accomplished this by causing a third mutation at a highly
conserved site in the potato L1, S302N. This mutation led to
the solubilization of the L1 protein, which otherwise presents
itself as inclusion bodies in the E. coli expression system. Their
results show that the triple-mutated potato L1 (LRKN ) was active
as a homotetramer of ∼ 200 kDa, as determined by size-exclusion
chromatography, in a manner similar to the previously reported
active S homotetramer [27].
The results of the present study indicate that the tomato L3 can
be active alone and as a monomer, without additional mutations.
In contrast with the mutated potato L1, the tomato and potato L3
proteins both have the three conserved amino acids, Arg44 , Lys54
and Asn302 . In fact, of all the L subunit proteins whose sequences
are available, only the potato and tomato L1 proteins have Ser302 ,
rather than the Asn302 , as recognized by Hwang et al. [22]. As
209
practically all previous research on potato L subunits has been
carried out on the major tuber L subunit protein L1, the activation
and solubilization of this L subunit required the triple mutations
to yield a soluble and active protein, which was not necessary for
the tomato L3.
However, in addition to the tomato L3 being active and soluble,
the tomato protein is also active as a monomer, which is perhaps
the most surprising result of the present study. We repeated
the expression and extraction studies five times and each time
activity was recovered, although activity was variable. We were
successful twice in obtaining enough active enzyme from the
E. coli expression of L3 for partial purification, indicating that
the protein is labile, but we do not know what determines its
stability. The L3 monomer is inefficiently catalytic, with unusual
characteristics, all nearly identical with those of the mutated L
subunit homotetramer studied by Hwang et al. [22]. Like the
triple-mutated potato L1, the tomato L3 shows, (i) an increase
in activity in the presence of Pi , (ii) an inhibitory response to
high 3-PGA concentrations, and (iii) poor substrate affinities. The
similarity of these characteristics indicates that they are not due
to the monomeric status of the enzyme but, perhaps, to the loss of
synergism in the absence of L and S interactions [22,57].
The enzyme is heat-labile, similar to the mutated L subunit
homotetramer reported by Hwang et al. [22], as well as to the
active heterotetramer containing an S with a truncated N-terminus
lacking the Cys12 [27], which contributes to the interactions
between the two S subunits [35]. Apparently, the novel heat
stability which characterizes numerous dicot AGPases [37] is
related to the stabilizing effect of the Cys12 disulfide bond.
We hypothesized that the ability of the natural tomato L3 to
function as an active monomer may be due to the many additional
differences between the triple-mutated potato L1 and the tomato
L3, particularly in the regions contributing to subunit interactions.
Alignment of the two proteins indicates 83 % identity and we
could make an intelligent guess as to which amino acids may allow
monomer activity based on the structure and subunit interactions
of the heterotetramer [35,58].
Jin et al. [35] showed that besides the Cys12 disulfide bond
that leads to the interaction of the two S proteins, and which
are absent in the L proteins, the different subunit monomers
also interact by end-to-end stacking of the β-helix domain (A:B
interactions), as well as the area of the linker loop between the Nand C-terminal domains (A:D interactions). The A:B interface
consists mostly of backbone interactions between the β-helix
subunits of monomers A and B. In contrast, the smaller A:D
interface mostly involves side-chain interactions. The sequence
similarity between S and the different L subunit isoforms at both
the A:B and A:D interface regions is very high, suggesting that
the subunit interactions at play in the S4 homotetramer are similar
to those in the native heterotetramer, as well as in the L subunit
homotetramers described by Hwang et al. [22]. We hypothesized
that amino acids in these regions, which contribute to the subunit
interactions as predicted by fASA values, and that are conserved
or similar between the two subunits that have been reported to
form active homotetramers (S [26] and L1 [21,22]) but that are
different in the L3 which forms an active monomer, are likely
to be candidates for a controlling role in subunit interactions.
These differences are limited to only three: a change of the
aromatic Tyr(Phe)317 to asparagine, the basic Lys322 to neutral
alanine and the basic Lys(Arg)340 to neutral threonine. Figure 6
presents two views of the A:B interface in the S4 homotetramer,
with residues Tyr(Phe)317 , Lys322 and Lys(Arg)340 emphasized.
Lys(Arg)340 is mostly exposed in the monomer and in the tetramer
(fASA=0.05) and the replacement by threonine in L3 is not
likely to have a significant effect. We therefore mutated the
c The Authors Journal compilation c 2010 Biochemical Society
210
Figure 6
M. Petreikov and others
A:B interface of the AGPase tetramer
Two views of the A:B interface indicating the positions of Lys340 and Lys322 (left-hand panel) and Tyr317 (right-hand panel). The indicated residues are shown as space-filling models coloured by
atom type, green for carbon, blue for nitrogen and red for oxygen. The emphasis is on residues conserved or similar between S and L1, and not conserved in L3. The solvent-accessible surface of
subunit A is shown coloured by the electrostatic potential, with blue for positive and red for negative potential. The second subunit is shown as a green ribbon.
remaining two amino acids together in order to test whether they
contribute to the subunit interactions.
The results of the mutation study indicate that these two
particular amino acid differences together are not responsible
for the different results of interactions between L3 (the present
study) and L1 [22] subunits. The fASA for Lys322 is 0.27 and
although this residue is not involved in inter-monomer hydrogen
bonding, it makes significant hydrophobic interactions that are
mostly lost upon mutation to alanine. Both lysine residues do not
contribute to electrostatic inter- or intra-monomer interactions
(note the exposed Nζ atoms). The A:B interface is mostly neutral,
hence replacement of the positively charged residues to neutral
residues does not necessarily hamper tetramerization. Tyr317 of S is
replaced conservatively by phenylalanine in L1, but by asparagine
in L3. Figure 6 shows that Tyr317 makes extensive contacts with
the adjacent subunit leading to a fASA of 0.55. However, this
residue is not involved in hydrogen bonding and its hydroxy
hydrogen is exposed in the tetramer. Our results are in line with
those of a recent study which suggests that although the side
chains of Tyr317 and Lys322 contribute to the binding energy of the
A:B interface, most of the binding energy is due to interactions
via backbone atoms [59].
We did not mutate any of the amino acids involved in the
A:D interface (amino acids 75–135) since there were no clear
differences that distinguished L3 from the L1 and S in tomato.
The A:D interface includes 16 residues whose fASA exceeds
10 % (Supplementary Table S2). This interface is more conserved
between S and L1 than the A:B interface, in line with the observation that most of the interactions here involve the side chains.
Residue Glu99 of S is conservatively replaced with aspartate in
L1, and non-conservatively replaced with serine in L3 of potato;
however, it remains an aspartate residue in the tomato L3, one of
the few differences between the potato and tomato L3 sequences.
This residue occurs next to the unresolved segment in the structure
[35] and its exact position may therefore be compromised. A
c The Authors Journal compilation c 2010 Biochemical Society
change in the side-chain conformation of this negative residue
can lead to formation of an ionic bridge with Arg87 of the adjacent
monomer, whose disruption may hamper tetramerization. The
corresponding position in the bacterial AGPase [38] is occupied
by Asn89 , which makes a hydrogen bond with the backbone of
subunit C, contributing to this weak interface. This observation,
together with the retention of the acidic aspartate residue in
the Glu99 position of the tomato L3 lead to the conclusion that
differences in the A:D interface are not involved in the novel
absence of tetramerization in the present study.
In summary, very few interface positions are occupied by the
same (or very similar) residue in S and L1 and by a different
residue in L3. Our mutation results indicate that these two
particular amino acid differences do not contribute to determining
the subunit interactions leading to the monomer or tetramer state.
Non-conserved residues will probably play a role in determining
activity, even if the solvent accessibility is not significantly altered.
Georgelis et al. [60] recently reported the effects of mutations
in the maize L subunit interface regions on the kinetic properties
of the maize tetrameric enzymes. One of the positions mutated in
the present study (N317F) was also mutated in the maize study
(C382F) and the resulting tetramer exhibited reduced activity. At
this stage we cannot explain why both S and mutant-L1 proteins
can form tetramers and are active only in the tetrameric form
[22,26], whereas L3 is active only as a monomer and does not
form tetramers.
AUTHOR CONTRIBUTION
Marina Petreikov, Jack Preiss and Arthur Schaffer designed the experiments. Miriam
Eisenstein contributed the protein structure portions of the paper. Marina Petreikov and
Yelena Yeselson carried out the research. Marina Petreikov, Miriam Eisenstein and Arthur
Schaffer were responsible for writing the manuscript.
Tomato ADP-Glc pyrophosphorylase large subunits
ACKNOWLEDGEMENTS
We thank Mr Shmuel Shen and Mr Lev Gelfandbein for expert technical assistance. We
thank Dr Thomas W. Okita and Dr Harry W. Janes for gifts of antibodies.
FUNDING
This work was supported by the United States–Israel Binational Agricultural Research and
Development Fund [grant number IS-3733-05 (to A.A.S and J.P.)].
REFERENCES
1 Tsai, C. Y. and Nelson, O. E. (1966) Starch-deficient maize mutant lacking adenosine
diphosphate glucose pyrophosphorylase activity. Science 151, 341–343
2 Lin, T. P., Caspar, T., Somerville, C. and Preiss, J. (1988) Isolation and characterization of
a starchless mutant of Arabidopsis thaliana (L.) Heynh lacking ADPglucose
pyrophosphorylase activity. Plant Physiol. 86, 1131–1135
3 Stark, D. M., Timmerman, K. P., Barry, G. F., Preiss, J. and Kishore, G. M. (1992)
Regulation of the amount of starch in plant tissues by ADP glucose pyrophosphorylase.
Science 258, 287–292
4 Muller-Rober, B., Sonnewald, U. and Willmitzer, L. (1992) Inhibition of the ADP-glucose
pyrophosphorylase in transgenic potatoes leads to sugar-storing tubers and influences
tuber formation and expression of tuber storage protein genes. EMBO J. 11, 1229–1238
5 Smith, A. M., Bettey, M. and Bedford, I. D. (1989) Evidence that the rb locus alters the
starch content of developing pea embryos through an effect on ADP-glucose
pyrophosphorylase. Plant Physiol. 89, 1279–1284
6 Smidansky, E. D., Clancy, M., Meyer, F. D., Lanning, S. P., Blake, N. K., Talbert, L. E. and
Giroux, M. J. (2002) Enhanced ADP-glucose pyrophosphorylase activity in wheat
endosperm increases seed yield. Proc. Natl. Acad. Sci. U.S.A. 99, 1724–1729
7 Petreikov, M., Shen, S., Yeselson, Y., Levin, I., Bar, M. and Schaffer, A. A. (2006)
Temporally extended gene expression of the ADP-Glc pyrophosphorylase large subunit
(AgpL1) leads to increased enzyme activity in developing tomato fruit. Planta 224,
1465–1479
8 Ballicora, M. A., Iglesias, A. A. and Preiss, J. (2004) ADP-Glucose pyrophosphorylase: a
regulatory enzyme for plant starch synthesis. Photosynth. Res. 79, 1–24
9 Preiss, J. and Sivak, M. N. (1996) Starch synthesis in sink and sources. In
Photoassimilate Distribution in Plants Crops. (Zamski, E. and Schaffer, A. A., eds),
pp. 63–96, Marcel Dekker, New York
10 Smith-White, B. J. and Preiss, J. (1992) Comparison of proteins of ADP-glucose
pyrophosphorylase from diverse sources. J. Mol. Evol. 34, 449–464
11 Crevillen, P., Ventriglia, T., Pinto, F., Orea, A., Merida, A. and Romero, J. M. (2005)
Differential pattern of expression and sugar regulation of Arabidopsis thaliana
ADP-glucose pyrophosphorylase-encoding genes. J. Biol. Chem. 280, 8143–8149
12 Park, S. W. and Chung, W. I. (1998) Molecular cloning and organ-specific expression of
three isoforms of tomato ADP-glucose pyrophosphorylase gene. Gene 206, 215–221
13 Chen, B. Y., Janes, H. W. and Gianfagna, T. (1998) PCR cloning and characterization of
multiple ADP-glucose pyrophosphorylase cDNAs from tomato. Plant Sci. 136, 59–67
14 Singh, S., Choi, S. B., Modi, M. K. and Okita, T. W. (2002) Isolation and characterization
of cDNA clones encoding ADP-glucose pyrophosphorylase (AGPase) large and small
subunits from chickpea (Cicer arietinum L.). Phytochemistry 59, 261–268
15 Crevillen, P., Ballicora, M. A., Merida, A., Preiss, J. and Romero, J. M. (2003) The
different large subunit isoforms of Arabidopsis thaliana ADP-glucose pyrophosphorylase
confer distinct kinetic and regulatory properties to the heterotetrameric enzyme. J. Biol.
Chem. 278, 28508–28515
16 Ventriglia, T., Ballicora, M. A., Crevillen, P., Preiss, J. and Romero, J. M. (2007)
Regulatory properties of potato-Arabidopsis hybrid ADP-glucose pyrophosphorylase.
Plant Cell Physiol. 48, 875–880
17 Schaffer, A. A. and Petreikov, M. (1997) Sucrose-to-starch metabolism in tomato fruit
undergoing transient starch accumulation. Plant Physiol. 113, 739–746
18 Georgelis, N., Braun, E. L. and Hannah, L. C. (2008) Duplications and functional
divergence of ADP-glucose pyrophosphorylase genes in plants. BMC Evol. Biol. 8, 232
19 Ranc, N., Munos, S., Santoni, S. and Causse, M. (2008) A clarified position for Solanum
lycopersicum var. cerasiforme in the evolutionary history of tomatoes (solanaceae). BMC
Plant Biol. 8, 130
20 Spooner, D. M., McLean, K., Ramsay, G., Waugh, R. and Bryan, G. J. (2005) A single
domestication for potato based on multilocus amplified fragment length polymorphism
genotyping. Proc. Natl. Acad. Sci. U.S.A. 102, 14694–14699
21 Ballicora, M. A., Dubay, J. R., Devillers, C. H. and Preiss, J. (2005) Resurrecting the
ancestral enzymatic role of a modulatory subunit. J. Biol. Chem. 280, 10189–10195
211
22 Hwang, S. K., Nagai, Y., Kim, D. and Okita, T. W. (2008) Direct appraisal of the potato
tuber ADP-glucose pyrophosphorylase large subunit in enzyme function by study of a
novel mutant form. J. Biol. Chem. 283, 6640–6647
23 Ventriglia, T., Kuhn, M. L., Ruiz, M. T., Ribeiro-Pedro, M., Valverde, F., Ballicora, M. A.,
Preiss, J. and Romero, J. M. (2008) Two Arabidopsis ADP-glucose pyrophosphorylase
large subunits (APL1 and APL2) are catalytic. Plant Physiol. 148, 65–76
24 Spooner D.M., Peralta I. E. and Knapp S. (2005) Comparison of AFLPs with other markers
for phylogenetic inference in wild tomatoes [Solanum L. section Lycopersicon (Mill.)
Wettst.]. Taxon 54, 43–61
25 Miron, D., Petreikov, M., Carmi, N., Shen, S., Levin, I., Granot, D., Zamski, E. and
Schaffer, A. A. (2002) Sucrose uptake, invertase localization and gene expression in
developing fruit of Lycopersicon esculentum and the sucrose-accumulating Lycopersicon
hirsutum . Physiol. Plant 115, 35–47
26 Iglesias, A. A., Barry, G. F., Meyer, C., Bloksberg, L., Nakata, P. A., Greene, T., Laughlin,
M. J., Okita, T. W., Kishore, G. M. and Preiss, J. (1993) Expression of the potato tuber
ADP-glucose pyrophosphorylase in Escherichia coli . J. Biol. Chem. 268, 1081–1086
27 Ballicora, M. A., Laughlin, M. J., Fu, Y. B., Okita, T. W., Barry, G. F. and Preiss, J. (1995)
Adenosine 5 -diphosphate-glucose pyrophosphorylase from potato-tuber - significance
of the N-terminus of the small-subunit for catalytic properties and heat-stability. Plant
Physiol. 109, 245–251
28 Govons, S., Vinopal, R., Ingraham, J. and Preiss, J. (1969) Isolation of mutants of
Escherichia coli B altered in their ability to synthesize glycogen. J. Bacteriol. 97,
970–972
29 Petreikov, M., Dai, N., Granot, D. and Schaffer, A. A. (2001) Characterization of native
and yeast-expressed tomato fruit fructokinase enzymes. Phytochemistry 58, 841–847
30 Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature 227, 680–685
31 Nakata, P. A. and Okita, T. W. (1995) Differential regulation of ADP-glucose
pyrophosphorylase in the sink and source tissues of potato. Plant Physiol. 108,
361–368
32 Schaffer, A. A., Levin, I., Oguz, I., Petreikov, M., Cincarevsky, F., Yeselson, L., Shen, S.,
Gilboa, N. and Bar, M. (2000) ADP-glucose pyrophosphorylase activity and starch
accumulation in immature tomato fruit: the effect of a Lycopersicon hirsutum -derived
introgression encoding for the large subunit. Plant Sci. 152, 135–144
33 Tiessen, A., Hendriks, J. H., Stitt, M., Branscheid, A., Gibon, Y., Farre, E. M. and
Geigenberger, P. (2002) Starch synthesis in potato tubers is regulated by
post-translational redox modification of ADP-glucose pyrophosphorylase: a novel
regulatory mechanism linking starch synthesis to the sucrose supply. Plant Cell 14,
2191–2213
34 Eisenthal, R. and Cornish-Bowden, A. (1974) The direct linear plot. A new graphical
procedure for estimating enzyme kinetic parameters. Biochem. J. 139, 715–720
35 Jin, X., Ballicora, M. A., Preiss, J. and Geiger, J. H. (2005) Crystal structure of potato
tuber ADP-glucose pyrophosphorylase. EMBO J. 24, 694–704
36 Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,
248–254
37 Sowokinos, J. R. and Preiss, J. (1982) Pyrophosphorylases in Solanum tuberosum : III.
Purification, physical, and catalytic properties of ADPglucose pyrophosphorylase in
potatoes. Plant Physiol. 69, 1459–1466
38 Cupp-Vickery, J. R., Igarashi, R. Y., Perez, M., Poland, M. and Meyer, C. R. (2008)
Structural analysis of ADP-glucose pyrophosphorylase from the bacterium Agrobacterium
tumefaciens. Biochemistry 47, 4439–4451
39 Chen, B. Y. and Janes, H. W. (1997) Multiple forms of ADP-glucose pyrophosphorylase
from tomato fruit. Plant Physiol. 113, 235–241
40 Chen, B. Y. and Janes, H. W. (1998) Multiple forms of ADP-glucose pyrophosphorylase
from tomato leaf. Physiol. Plant 103, 491–496
41 Kavakli, I. H., Kato, C., Choi, S. B., Kim, K. H., Salamone, P. R., Ito, H. and Okita, T. W.
(2002) Generation, characterization, and heterologous expression of wild-type and
up-regulated forms of Arabidopsis thaliana leaf ADP-glucose pyrophosphorylase. Planta
215, 430–439
42 Rook, F., Corke, F., Card, R., Munz, G., Smith, C. and Bevan, M. W. (2001) Impaired
sucrose-induction mutants reveal the modulation of sugar-induced starch biosynthetic
gene expression by abscisic acid signalling. Plant J. 26, 421–433
43 Sweetlove, L. J., Muller-Rober, B., Willmitzer, L. and Hill, S. A. (1999) The contribution of
adenosine 5 -diphosphoglucose pyrophosphorylase to the control of starch synthesis in
potato tubers. Planta 209, 330–337
44 Farre, E. M., Tiessen, A., Roessner, U., Geigenberger, P., Trethewey, R. N. and Willmitzer,
L. (2001) Analysis of the compartmentation of glycolytic intermediates, nucleotides,
sugars, organic acids, amino acids, and sugar alcohols in potato tubers using a
nonaqueous fractionation method. Plant Physiol. 127, 685–700
45 Geigenberger, P. (2003) Regulation of sucrose to starch conversion in growing potato
tubers. J. Exp. Bot. 54, 457–465
c The Authors Journal compilation c 2010 Biochemical Society
212
M. Petreikov and others
46 Neuhaus, H. E. and Emes, M. J. (2000) Nonphotosynthetic metabolism in plastids. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 51, 111–140
47 Piechulla, B., Glick, R. E., Bahl, H., Melis, A. and Gruissem, W. (1987) Changes in
photosynthetic capacity and photosynthetic protein pattern during tomato fruit ripening.
Plant Physiol. 84, 911–917
48 Buker, M., Schunemann, D. and Borchert, S. (1998) Enzymic properties and capacities of
developing tomato (Lycopersicon esculentum L.) fruit plastids. J. Exp. Bot. 49,
681–691
49 Tanaka, A., Fujita, K. and Shioya, M. (1974) Nutrio-physiological studies on the tomato
plant. II. Translocation of photosynthesis. Soil Sci. Plant Nutr. 20, 163–171
50 Schunemann, D., Schott, K., Borchert, S. and Heldt, H. W. (1996) Evidence for the
expression of the triosephosphate translocator gene in green and non-green tissue of
tomato and potato. Plant Mol. Biol. 31, 101–111
51 Entwistle, G. and ap Rees, T. A. (1990) Lack of fructose-1,6-bisphosphatase in a range of
higher plants that store starch. Biochem. J. 271, 467–472
52 Thorbjornsen, T., Asp, T., Jorgensen, K. and Nielsen, T. H. (2002) Starch biosynthesis
from triose-phosphate in transgenic potato tubers expressing plastidic fructose-1,6bisphosphatase. Planta 214, 616–624
53 Obiadalla-Ali, H., Fernie, A. R., Lytovchenko, A., Kossmann, J. and Lloyd, J. R. (2004)
Inhibition of chloroplastic fructose 1,6-bisphosphatase in tomato fruits leads to decreased
fruit size, but only small changes in carbohydrate metabolism. Planta 219, 533–540
Received 23 November 2009/11 March 2010; accepted 17 March 2010
Published as BJ Immediate Publication 17 March 2010, doi:10.1042/BJ20091777
c The Authors Journal compilation c 2010 Biochemical Society
54 Plaxton, W. C. (1996) The organization and regulation of plant glycolysis. Annu. Rev.
Plant Physiol. Plant Mol. Biol. 47, 185–214
55 Sweetlove, L. J., Kruger, N. J. and Hill, S. A. (2001) Starch synthesis in transgenic potato
tubers with increased 3-phosphoglyceric acid content as a consequence of increased
6-phosphofructokinase activity. Planta 213, 478–482
56 Westram, A., Lloyd J.R., Roessner, U., Riesmeier, J.W. and Kossman, J. (2002) Increases
in 3-phosphoglyceric acid in potato plants through antisense reduction of cytoplasmic
phosphoglycerate mutase impairs photosynthesis and growth but does not increase
starch synthesis. Plant Cell Environ. 25, 1133–1143
57 Hwang, S. K., Salamone, P. R. and Okita, T. W. (2005) Allosteric regulation of the higher
plant ADP-glucose pyrophosphorylase is a product of synergy between the two subunits.
FEBS Lett. 579, 983–990
58 Tuncel, A., Kavakli, I. H. and Keskin, O. (2008) Insights into subunit interactions in the
heterotetrameric structure of potato ADP-glucose pyrophosphorylase. Biophys. J. 95,
3628–3639
59 Baris, I., Tuncel, A., Ozber, N., Keskin, O. and Kavakli, H. K. (2009) Investigation of the
interaction between the large and small subunits of potato ADP-glucose
pyrophosphorylase. PLoS Comput. Biol. 5, 1–14
60 Georgelis, N., Shaw, J. R. and Hannah, L. C. (2009) Phylogenetic analysis of
ADP-glucose pyrophosphorylase subunits reveals a role of subunit interfaces in the
allosteric properties of the enzyme. Plant Physiol. 151, 67–77
Biochem. J. (2010) 428, 201–212 (Printed in Great Britain)
doi:10.1042/BJ20091777
SUPPLEMENTARY ONLINE DATA
Characterization of the AGPase large subunit isoforms from tomato
indicates that the recombinant L3 subunit is active as a monomer
Marina PETREIKOV*, Miriam EISENSTEIN†, Yelena YESELSON*, Jack PREISS‡ and Arthur A. SCHAFFER*1
*Department of Vegetable Research, Volcani Center-ARO, Bet Dagan 50250, Israel, †Chemical Research Support Unit, Weizmann Institute of Science, Rehovot 76100, Israel, and
‡Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48823, U.S.A.
Table S1
Primers for cloning the tomato AGPase subunits
(a) Primers for cloning cDNA fragments into the pGEM-T Easy vector. (b) Primers for cloning cDNA fragments into the expression vectors pMon17335 and pMon17336. Restriction enzyme sites are
underlined. (c) Primers used for site-directed mutagenesis.
(a)
Gene
Direction
Primer
L1 (DQ322684)
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
5 -GGATACTTGTTGTGCGGCT-3
5 -CCTAAGTTGAGCAATAGATAAAC-3
5 -CACAGAAATCTCAAGAGAGTCC-3
5 -CTCGATACCAATAATTAACATACC-3
5 -CAGCCATTTCAGTACTCTTGAT-3
5 -GTGTCATCAGCTAGTTTATTGAA-3
5 -GGAAGTGCAATCACACTCTAC-3
5 -CAGCAACCAGGTAATGAGAAG-3
Direction
Primer
Restriction enzyme
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
5 -GAAGATTAAACCTGCCATGGCTTACKCTGTGATC-3
5 -CCACTAGTTCAACAAGAGAAGCTTYYCTAGTTCA-3
5 -GTAAAAATGTTTTCATGACTGTTCTCACCCGTGA-3
5 -CCAACTAGCAAGAACACTTAGCAAGCTTTCATATG-3
5 -GTGAAGCCTCATATCTACATGTCACTTACAACTGATATTG-3
5 -CTTGAAAGAGAAGACAAAAAGACATTGGTACCCTTGT-3
5 -GTGTCTCCCATGGCTGTTTCTGATTCG-3
5 -CCAAGTTATAATGAGCTCCTTTAAATGACGATTC-3
NcoI, V64M
HindIII
PagI, S55M, A56T
HindIII
PciI
KpnI
NcoI, S71M
SacI
L2 (U81034)
L3 (U85497)
S (L41126)
(b)
Gene
L1
L2
L3
S
1
(c)
Primer name
Sequence
Primer A
Primer B
Primer C
Primer D
5 -CCATATCAAGCTTGCCAATA-3
5 -GCTATTATCAATCTTTGATGGAGGTAAGAATCTCCTTGACGT-3
5 -GCTATTATCAATCTTTGATGGAGGTAAGAATCTCCTTGACGT-3
5 -GCACTTTTC CTTCAGCTAGCT-3
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2010 Biochemical Society
M. Petreikov and others
Figure S1
Sequences used for the expression of the tomato AGPase subunits
(A) Homology of the four tomato AGPase subunits indicating the N-termini according to which the individual subunits were cloned. Tomato protein sequence accession numbers were from the
GenBank® database: L1 (AF184345), L2 (U85496), L3 (U85497) and S (L41126). The amino acids are numbered according to the potato small subunit. Underlined letters in red indicate the
N-terminus of the expressed protein, as described by Ballicora et al. [1] (S), Singh et al. [2], derived from Park and Chung [3] (L). (B) Results from N-terminal sequencing (large arrow) and MS/MS
peptide sequencing of the AGPase large subunit (Lycopersicon hirsutum ) protein. The N-terminus of the tomato fruit L1 began at the amino acids VITTEN and the MS/MS sequences showed 80 %
coverage of the protein sequence with no amino acids further upstream than the VITTEN sequence.
Table S2 fASA values for the amino acids participating in the A:D subunit
interactions
These are residues that change their fASA upon tetramerization by more than 10 %. The Table
indicates the conservation of amino acids at these positions, corresponding to the alignment
with the S sequence. For the L1 and L3 columns, single-amino-acid codes are used.
Figure S2 Iodine staining of different combinations of tomato or potato
AGPase L and S subunits expressed in E. coli AC70RI-504
Sp/Lp, potato small and large L1 subunits; S/L1, tomato small AgpS1 and large AgpL1 subunits;
S/L2, tomato small AgpS1 and large AgpL2 subunits; S/L3, tomato small AgpS1 and large AgpL3
subunits; L1, tomato large AgpL1 subunit; L2, tomato large AgpL2 subunit; L3, tomato large
AgpL3 subunit; S1, tomato small AgpS1 subunit. The S1 streak grew denser than the L1, L2
and L3 streaks so that the black background does not show through, in contrast with the L1,
L2 and L3 streaks, which appear darker due to the black cloth background.
c The Authors Journal compilation c 2010 Biochemical Society
S
fASA
L1
L3
Gln109
Ala79
Trp129
Asn82
Arg83
Ser86
Pro111
Glu103
Glu133
Leu130
Glu99
Gln126
Gly100
Tyr127
Asn77
Ala106
0.52
0.46
0.42
0.31
0.30
0.26
0.26
0.21
0.21
0.19
0.17
0.17
0.16
0.15
0.13
0.10
Q
A
W
N
R
A
P
E
D
V
D
K
G
F
N
A
Q
A
W
N
R
A
P
E
D
L
D (tomato); S (potato)
Q
G
F
N
A
Tomato ADP-Glc pyrophosphorylase large subunits
REFERENCES
1 Ballicora, M. A., Laughlin, M. J., Fu, Y. B., Okita, T. W., Barry, G. F. and Preiss, J. (1995)
Adenosine 5 -diphosphate-glucose pyrophosphorylase from potato-tuber - significance of
the N-terminus of the small-subunit for catalytic properties and heat-stability. Plant
Physiol. 109, 245–251
2 Singh, S., Choi, S. B., Modi, M. K. and Okita, T. W. (2002) Isolation and characterization of
cDNA clones encoding ADP-glucose pyrophosphorylase (AGPase) large and small
subunits from chickpea (Cicer arietinum L.). Phytochemistry 59, 261–268
3 Park, S. W. and Chung, W. I. (1998) Molecular cloning and organ-specific expression of
three isoforms of tomato ADP-glucose pyrophosphorylase gene. Gene 206,
215–221
Received 23 November 2009/11 March 2010; accepted 17 March 2010
Published as BJ Immediate Publication 17 March 2010, doi:10.1042/BJ20091777
c The Authors Journal compilation c 2010 Biochemical Society