Biochemical characterization of rice trehalose-6-phosphate
phosphatases supports distinctive functions of these plant
enzymes
Shuhei Shima
1,2
, Hirokazu Matsui
2
, Satoshi Tahara
2
and Ryozo Imai
1
1 Crop Cold Tolerance Research Team, National Agricultural Research Center for Hokkaido Region, NARO, Toyohira-ku, Sapporo, Japan
2 Department of Applied Bioscience, Graduate School of Agriculture, Hokkaido University, Sapporo, Japan
Trehalose is a nonreducing disaccharide in which two
glucose units are linked by an a,a-1,1-glycosidic link-
age. The prevalent pathway for trehalose synthesis
includes two enzymatic reactions. Trehalose 6-phos-
phate (Tre6P) is generated from UDP-glucose and glu-
cose 6-phosphate (Glc6P) in a reaction catalyzed by
trehalose-6-phosphate synthase (TPS). Tre6Pis then
dephosphorylated to form trehalose via trehalose-6-
phosphate phosphatase (TPP) [1]. In yeast, trehalose
synthesis is carried out by a large enzyme complex that
is composed of four subunits, including TPS1, TPS2,
and regulatory subunits TSL1 and TPS3 [2].
Trehalose is widely distributed in nature. In bacteria,
fungi, and insects, trehalose functions as a storage car-
bohydrate or a blood sugar. In addition, trehalose can
protect cellular integrity against a variety of environ-
mental stresses associated with desiccation, heat, and
cold [3]. In plants, the presence of trehalose has been
Keywords
functional analysis; kinetic analysis; Oryza
sativa; recombinant protein; trehalose
Correspondence
R. Imai, Crop Cold Tolerance Research
Team, National Agricultural Research Center
for Hokkaido Region, National Agriculture
and Food Research Organization,
Hitsujigaoka 1, Toyohira-ku, Sapporo
0628555, Japan
Fax ⁄Tel: +81 11 857 9382
E-mail: rzi@affrc.go.jp
(Received 8 November 2006, revised 14
December 2006, accepted 19 December
2006)
doi:10.1111/j.1742-4658.2007.05658.x
Substantial levels of trehalose accumulate in bacteria, fungi, and inverte-
brates, where it serves as a storage carbohydrate or as a protectant against
environmental stresses. In higher plants, trehalose is detected at fairly low
levels; therefore, a regulatory or signaling function has been proposed for
this molecule. In many organisms, trehalose-6-phosphate phosphatase is
the enzyme governing the final step of trehalose biosynthesis. Here we
report that OsTPP1 and OsTPP2 are the two major trehalose-6-phosphate
phosphatase genes expressed in vegetative tissues of rice. Similar to results
obtained from our previous OsTPP1 study, complementation analysis of a
yeast trehalose-6-phosphate phosphatase mutant and activity measurement
of the recombinant protein demonstrated that OsTPP2 encodes a func-
tional trehalose-6-phosphate phosphatase enzyme. OsTPP2 expression is
transiently induced in response to chilling and other abiotic stresses. Enzy-
matic characterization of recombinant OsTPP1 and OsTPP2 revealed strin-
gent substrate specificity for trehalose 6-phosphate and about 10 times
lower K
m
values for trehalose 6-phosphate as compared with trehalose-
6-phosphate phosphatase enzymes from microorganisms. OsTPP1 and
OsTPP2 also clearly contrasted with microbial enzymes, in that they are
generally unstable, almost completely losing activity when subjected to heat
treatment at 50 C for 4 min. These characteristics of rice trehalose-6-
phosphate phosphatase enzymes are consistent with very low cellular sub-
strate concentration and tightly regulated gene expression. These data also
support a plant-specific function of trehalose biosynthesis in response to
environmental stresses.
Abbreviations
ABA, abcisic acid; Glc1P, glucose 1-phosphate; Glc6P, glucose 6-phosphate; GST, glutathione S-transferase; TPP, trehalose-6-phosphate
phosphatase; TPS, trehalose-6-phosphate synthase; Tre6P, trehalose 6-phosphate.
1192 FEBS Journal 274 (2007) 1192–1201 ª2007 The Authors Journal compilation ª2007 FEBS
documented in a limited number of species, including
Myrothamnus flabellifolia, a desiccation-tolerant desert
plant [4,5], and Selaginella lepidophylla, a desiccation-
tolerant moss [6]; its occurrence in many other plant
species is uncertain.
TPS and TPP genes were functionally identified in
Arabidopsis thaliana by complementation of Saccharo-
myces cerevisiae mutants [7,8]. Homologous TPS and
TPP genes have now been identified in many other
plant species. These results suggest that trehalose
synthesis may in fact be ubiquitous among angio-
sperms, although the levels to which it accumulates
are generally low [1,9]. Attempts to increase trehalose
content in plants by overexpressing microbial TPS
and TPP genes resulted in transgenic tobacco and
potato plants with increased stress tolerance at the
tissue level [10–12]. However, these transformants
exhibited pleiotropic phenotypes, such as stunted
growth and lancet-shaped leaves [11,12]. On the
other hand, expression of an Escherichia coli TPS–
TPP fusion enzyme in transgenic rice resulted in
accumulation of 3–10 times more trehalose compared
to nontransgenic rice plants, imparting abiotic stress
tolerance without altering morphology [13,14]. There-
fore, these findings suggested that accumulation of
Tre6Pmay result in the observed morphologic alter-
ations in the tobacco and potato studies.
Although trehalose biosynthesis in higher plants has
been demonstrated, details of both the physiologic
functions and regulation of this pathway remain lar-
gely unknown. Genome sequencing of Arabidopsis and
rice has revealed complex genomic organization of
plant trehalose biosynthesis genes. Eleven putative TPS
and 10 putative TPP genes were identified within the
Arabidopsis genome, and nine putative TPS and nine
putative TPP genes were found within the rice genome.
Genetic studies have revealed that trehalose biosyn-
thesis genes function specifically in regulating plant
growth and development. An Arabidopsis knockout
mutant of AtTPS1 exhibited impaired embryo matur-
ation [15]. Further characterization of the mutant dem-
onstrated that AtTPS1 is also required for vegetative
growth and flowering [16]. A recent study established
that a maize TPP gene is involved in inflorescence
development [17]. A more specific function of trehalose
biosynthesis in the regulation of starch biosynthesis
has recently been revealed. Trehalose feeding was
found to induce expression of ApL3, encoding a large
subunit of ADP-glucose pyrophosphorylase in Arabid-
opsis [18,19]. It was demonstrated recently that Tre6P
directly regulated starch synthesis via post-transla-
tional redox activation of ADP-glucose pyrophospho-
rylase [20,21].
In our previous study, we demonstrated that expres-
sion of the rice TPP gene OsTPP1 is rapidly and tran-
siently induced by chilling stress and abcisic acid
(ABA) treatment. Induction of OsTPP1 was followed
by transient increases in total TPP activity and treha-
lose content in rice root [22]. Eight other members of
the rice TPP gene family have not yet been character-
ized, so it is not known if these members have diver-
gent functions in rice. In addition, the enzymatic
properties of plant TPPs are largely unknown.
In this article, we report the isolation of a second
TPP gene from rice, OsTPP2, and its relative tran-
scription in response to abiotic stresses, as well as the
in vivo and in vitro functionality of its translated prod-
uct. We also describe unique kinetic and biophysical
properties of the plant TPPs.
Results
Isolation of rice OsTPP2
Completion of the rice genome sequence revealed nine
putative TPP genes. To determine which of these TPP
genes are expressed in rice seedlings, RT-PCR was car-
ried out using specific primer sets designed to amplify
transcripts from all of these OsTPP genes (OsTPP1–
OsTPP9) [22]. Only mRNA for OsTPP2 was detected
in addition to that of the previously characterized
OsTPP1 after 28 cycles of PCR amplification
(Fig. 1A). OsTPP3–OsTPP9 mRNAs were not detec-
ted in root and shoot tissues after up to 35 cycles of
amplification (data not shown). These results suggested
that OsTPP1 and OsTPP2 were the major TPP genes
expressed in rice seedlings. A full-length OsTPP2
cDNA was then isolated from root tissue by RT-PCR.
The OsTPP2 gene contained an ORF encoding a
42.6 kDa protein with 382 amino acid residues. Overall
amino acid sequence homology between OsTPP2 and
OsTPP1 was 53% (Fig. 1B). Greater similarity was
observed between OsTPP2 and Arabidopsis AtTPPA
(57%). OsTPP2 contains two motifs shared by all TPP
enzymes (Fig. 1B), known as phosphatase boxes: (FIL-
MAVT)-D-(ILFRMVY)-D-(GSNDE)-(TV)-(ILVAM)-
(ATSVILMC)-X-(YFWHKR)-X-(YFWHNQ) (domain
A), and (KRHNQ)-G-D-(FYWHILVMC)-(QNH)-
(FWYGP)-D-(PSNQYW) (domain B) [23].
Responses of OsTPP2 to chilling and other
abiotic stresses
In our previous study, we demonstrated that OsTPP1
expression is transiently induced by multiple abiotic
stresses [22]. We therefore determined whether
S. Shima et al.Rice trehalose-6-phosphate phosphatases
FEBS Journal 274 (2007) 1192–1201 ª2007 The Authors Journal compilation ª2007 FEBS 1193
OsTPP2 expression is also responsive to abiotic stres-
ses. RNA gel blot analysis was performed on total
RNA extracted from rice seedlings subjected to low
temperature (12 C), drought, and salt stresses
(Fig. 2). OsTPP2 mRNA levels were detectable prior
to stress treatments, and transiently increased in
response to low temperature, peaking at 10 h after
the initiation of treatment in both shoot and root tis-
sues. This expression pattern contrasted with the
observed rapid induction of OsTPP1 and gradual
induction of OsMEK1 in response to low-temperature
treatment [22,24]. Drought stress transiently induced
OsTPP2 expression, which peaked at 6 h in shoots
and 2 h in roots (Fig. 2). The induction of OsTPP2
expression occurs earlier during stress treatment com-
pared with expression of another drought-induced
gene (salT) [25]. Treatment with 150 mmNaCl also
induced OsTPP2 expression (Fig. 2) in roots, suggest-
ing that stresses associated with water deficit similarly
affect expression of this gene. However, in contrast to
chilling and drought stress treatments, clear induction
of OsTPP2 was not observed in shoots, whereas the
salt treatment effectively induced salT in both roots
and shoots. Slight and transient induction of OsTPP2
was observed in roots and shoots in response to exo-
genous ABA. Together, these expression analyses indi-
cated involvement of OsTPP2 in multiple stress
responses.
A
B
A
B
Fig. 1. Expression of putative OsTPP genes in young vegetative tissues, and alignment of TPP sequences. (A) Expression analysis of puta-
tive TPP genes with RT-PCR, using RNAs extracted from root and shoot tissues of rice seedlings (O. sativa L. cv. Yukihikari). (B) Alignment
of the amino acid sequences of OsTPP2, OsTPP1 (O.sativa [22]), AtTPPA and AtTPPB (A.thaliana [8]), TPS2 (Sa.cerevisae [34]) and OtsB
(E.coli [35]). Database accession numbers are: OsTPP1, BAD12596; OsTPP2, BAF34519; AtTPPA, AAC39369; AtTPPB, AAC39370; TPS2,
CAA98893; and OtsB, CAA48912. Shading reflects the degree of amino acid conservation. Black shading indicates amino acid identity. The
bars represent highly conserved domains.
Rice trehalose-6-phosphate phosphatases S. Shima et al.
1194 FEBS Journal 274 (2007) 1192–1201 ª2007 The Authors Journal compilation ª2007 FEBS
OsTPP2 complements a yeast Dtps2 mutant
To detect OsTPP2 enzyme function in vivo,aSa.cere-
visiae (YPH499) tps2 (TPP) deletion mutant [22] was
transformed with plasmid constructs based on the
pAUR123 vector (Takara). Whereas wild-type cells
grow at both 30 C and 36 C, growth of the Dtps2
mutant at 36 C was inhibited because of its inability
to synthesize trehalose (Fig. 3). The same mutant yeast
strain transformed with OsTPP2 recovered wild-type
levels of growth at 36 C, suggesting that OsTPP2 is a
functional TPP enzyme in yeast cells.
TPP activity of recombinant OsTPP2
To determine whether OsTPP2 exhibits TPP activity
in vitro, it was purified as a recombinant protein. To
accomplish this, the ORF of OsTPP2 was inserted into
a pGEX-6P-3 vector to produce a glutathione S-trans-
ferase (GST)–OsTPP2 fusion protein. After affinity
column purifications and protease digestion, OsTPP2
proteins were purified to near homogeneity. The size
of the purified recombinant enzyme was estimated to
be 45 kDa on SDS ⁄PAGE, in accordance with the size
deduced from the nucleotide sequence (42.6 kDa)
(Fig. 4A). TPP activity was then measured using this
purified recombinant enzyme. Aliquots of purified
enzyme were added to the reaction mixtures, and
conversion of Tre6Pinto trehalose was detected as a
measure of enzyme activity (Fig. 4B). Under these
same conditions, purified GST or NaCl ⁄P
i
solution
without enzyme did not result in this conversion
(Fig. 4B). We therefore concluded that OsTPP2
encodes a functional TPP enzyme.
Enzymatic properties of OsTPP1 and OsTPP2
Although genes encoding plant TPPs have been identi-
fied in several plant species, their enzymatic character-
istics have not been explored. To further characterize
the enzymatic properties of plant TPP enzymes, we
also purified recombinant OsTPP1 using the same
methods. Then, the kinetic parameters of these recom-
binant OsTPP1 and OsTPP2 enzymes were determined.
The K
m
values for Tre6Pof OsTPP1 and OsTPP2
were determined to be 0.0921 and 0.186 mm, respect-
ively, using Hanes–Woolf plots (Table 1). The k
cat
val-
ues of OsTPP1 and OsTPP2 were 6.52 and 13.4 s
)1
,
respectively. Therefore, the k
cat
⁄K
m
values of OsTPP1
and OsTPP2 were approximately the same. These
Fig. 3. Complementation of the heat-sensitive phenotype of a
Sa. cerevisiae tps2 deletion mutant by introduction of OsTPP2.A
YPH499 tps2 deletion mutant was transformed with the pAUR123
vector (Dtps2) and pAUR123-OsTPP2 (Dtps2 ⁄OsTPP2). As a posit-
ive control, YPH499 wild-type cells were transformed with the
empty pAUR123 vector (wt). These transformants were grown
overnight in YPD liquid medium supplemented with 0.5 lgÆmL
)1
aureobasidin A. The cultures were then diluted 1–1000 times. Five
microliters of each dilution was then spotted onto an YPD agar
plate supplemented with 0.5 lgÆmL
)1
aureobasidin A. These plates
were incubated at 30 Cor36C for 2 days.
A
B
C
D
Fig. 2. Expression of OsTPP2 in rice seedlings in response to
abiotic stress and exogenous ABA treatment. Total RNAs were iso-
lated from rice seedlings subjected to chilling stress (A), drought
stress (B), 150 mMNaCl stress (C), and exogenous ABA (50 lM)
solution (D). The RNA blots were hybridized with an OsTPP2 probe.
The expression of OsTPP1 is shown for comparison of expression
patterns, and those of OsMEK1 [24] and salT [25] are shown as
positive controls for these treatments. Ethidium bromide-stained
total RNA (10 lg) is presented as a loading control.
S. Shima et al.Rice trehalose-6-phosphate phosphatases
FEBS Journal 274 (2007) 1192–1201 ª2007 The Authors Journal compilation ª2007 FEBS 1195
results indicated that both enzymes exhibit similar cat-
alytic activities. It is interesting to note here that the
K
m
values for rice TPPs are more than 10 times lower
than those of bacterial TPP enzymes reported thus far.
For instance, others reported that the K
m
values for
E. coli and Mycobacterium smegmatis TPPs were
2.5 mmand 1.5 mm, respectively [26,27].
To determine the substrate specificity of these
recombinant proteins, phosphatase activities were
measured using various sugar phosphate substrates
[glucose 1-phosphate (Glc1P), Glc6P, galactose 6-
phosphate, mannose 1-phosphate, mannose 6-phos-
phate, fructose 1-phosphate, fructose 6-phosphate,
sucrose 6-phosphate, lactose 1-phosphate, and ribose
5-phosphate]. Both OsTPP1 and OsTPP2 exhibited
strong phosphatase activity upon Tre6P, but almost
no activity (less than 1% relative to Tre6P) was detec-
ted with any of the other sugar phosphates tested (data
not shown).
The pH dependences of OsTPP1 and OsTPP2
enzyme activities were determined within a pH range
of 5.5–9.0, using two different buffers (Mes ⁄NaOH,
pH 5.5–7.5; Tris ⁄HCl, pH 7.0–9.0). The pH optima of
OsTPP1 and OsTPP2 were approximately 7.0 and 6.5,
respectively, whereas the enzymes had almost no activ-
ity at pH 5.5 or 9 (Fig. 5).
The heat stabilities of the recombinant OsTPP1
and OsTPP2 were determined by measuring residual
activities after heat treatments (40–80 C) (Fig. 6).
A
B
Fig. 4. Purification of recombinant OsTPP1 and OsTPP2 and deter-
mination of their activities. Recombinant OsTPP1 and OsTPP2 were
purified according to the experimental procedure described previ-
ously. (A) SDS ⁄PAGE (12%) was run with protein standards
(lane M), crude extracts of the recombinant bacterial strains
induced without (lane 2) or with (lane 3) isopropyl thio-b-D-galacto-
side, and purified recombinant OsTPP1 or OsTPP2 (lane 4). (B)
Chromatograms detailing TPP activities of recombinant OsTPP1,
OsTPP2, and GST. These proteins (0.5 lg) were used for assays in
100 lL reaction mixtures (2 mMTre6P,2mMMgCl
2
,50mM
Tris ⁄HCl, pH 7.0). T; trehalose; T6P, trehalose 6-phosphatase.
Table 1. Enzymatic properties of recombinant OsTPP1 and
OsTPP2.
Protein
K
ma
(mM)
K
cat
(s
)1
)
K
cat
⁄K
m
(mM
)1
Æs
)1
) Reference
OsTPP1 0.0921 6.52 70.8 This study
OsTPP2 0.186 13.4 72.0 This study
E. coli TPP 2.5 14.3 5.8 [27]
M. smegmatis TPP 1.5 – – [26]
a
K
m
for Tre6P.
Rice trehalose-6-phosphate phosphatases S. Shima et al.
1196 FEBS Journal 274 (2007) 1192–1201 ª2007 The Authors Journal compilation ª2007 FEBS