Trehalose accumulation in rice plants confers high tolerance levels to different abiotic stresses
Trehalose is a nonreducing disaccharide of glucose that functions as a compatible solute in the stabilization of biological structures under abiotic stress in bacteria, fungi, and invertebrates. With the notable exception of the desiccation-tolerant "resurrection plants," trehalose is not thought to accumulate to detectable levels in most plants. We report here the regulated overexpression of Escherichia coli trehalose biosynthetic genes (otsA and otsB) as a fusion gene for manipulating abiotic stress tolerance in rice. The fusion gene has the advantages of necessitating only a single transformation event and a higher net catalytic efficiency for trehalose formation. The expression of the transgene was under the control of either tissue-specific or stress-dependent promoters. Compared with nontransgenic rice, several independent transgenic lines exhibited sustained plant growth, less photo-oxidative damage, and more favorable mineral balance under salt, drought, and low-temperature stress conditions. Depending on growth conditions, the transgenic rice plants accumulate trehalose at levels 3-10 times that of the nontransgenic controls. The observation that peak trehalose levels remain well below 1 mgg fresh weight indicates that the primary effect of trehalose is not as a compatible solute. Rather, increased trehalose accumulation correlates with higher soluble carbohydrate levels and an elevated capacity for photosynthesis under both stress and nonstress conditions, consistent with a suggested role in modulating sugar sensing and carbohydrate metabolism. These findings demonstrate the feasibility of engineering rice for increased tolerance of abiotic stress and enhanced productivity through tissue-specific or stress-dependent overproduction of trehalose.
Trehalose accumulation in rice plants confers high
tolerance levels to different abiotic stresses
Ajay K. Garg*, Ju-Kon Kim
, Thomas G. Owens
, Anil P. Ranwala
, Yang Do Choi
, Leon V. Kochian
, and Ray J. Wu*
Departments of *Molecular Biology and Genetics,
Plant Biology, and
Horticulture, Cornell University, Ithaca, NY 14853;
Department of Biological Science,
Myongji University, Yongin, Kyonggi-Do 449-728, Korea;
School of Agricultural Biotechnology, Seoul National University, Suwon 441-744, Korea; and
U.S. Department of Agriculture–Agriculture Research Service, Plant, Soil, and Nutrition Laboratory, Cornell University, Ithaca, NY 14853
Communicated by Andre´ T. Jagendorf, Cornell University, Ithaca, NY, October 21, 2002 (received for review September 26, 2002)
Trehalose is a nonreducing disaccharide of glucose that functions
as a compatible solute in the stabilization of biological structures
under abiotic stress in bacteria, fungi, and invertebrates. With the
notable exception of the desiccation-tolerant ‘‘resurrection
plants,’’ trehalose is not thought to accumulate to detectable levels
in most plants. We report here the regulated overexpression of
Escherichia coli trehalose biosynthetic genes (otsA and otsB)asa
fusion gene for manipulating abiotic stress tolerance in rice. The
fusion gene has the advantages of necessitating only a single
transformation event and a higher net catalytic efﬁciency for
trehalose formation. The expression of the transgene was under
the control of either tissue-speciﬁc or stress-dependent promoters.
Compared with nontransgenic rice, several independent trans-
genic lines exhibited sustained plant growth, less photo-oxidative
damage, and more favorable mineral balance under salt, drought,
and low-temperature stress conditions. Depending on growth
conditions, the transgenic rice plants accumulate trehalose at levels
3–10 times that of the nontransgenic controls. The observation that
peak trehalose levels remain well below 1 mg兾g fresh weight
indicates that the primary effect of trehalose is not as a compatible
solute. Rather, increased trehalose accumulation correlates with
higher soluble carbohydrate levels and an elevated capacity for
photosynthesis under both stress and nonstress conditions, con-
sistent with a suggested role in modulating sugar sensing and
carbohydrate metabolism. These ﬁndings demonstrate the feasi-
bility of engineering rice for increased tolerance of abiotic stress
and enhanced productivity through tissue-speciﬁc or stress-depen-
dent overproduction of trehalose.
he explosive increase in world population, along with the
continuing deterioration of arable land, scarcity of fresh
water, and increasing environmental stress pose serious threats
to global agricultural production and food security. Despite
focused efforts to improve major crops for resistance to abiotic
stresses (1) such as drought, excessive salinity, and low temper-
ature by traditional breeding, success has been limited. This lack
of desirable progress is attributable to the fact that tolerance to
abiotic stress is a complex trait that is influenced by coordinated
and differential expression of a network of genes. Fortunately, it
is now possible to use transgenic approaches to improve abiotic
stress tolerance in agriculturally important crops with far fewer
target traits than had been anticipated (2).
Abiotic stresses can directly or indirectly affect the physiolog-
ical status of an organism by altering its metabolism, growth, and
development. A common response of organisms to drought,
salinity, and low-temperature stresses is the accumulation of
sugars and other compatible solutes (3). These compounds serve
as osmoprotectants and, in some cases, stabilize biomolecules
under stress conditions (3, 4). One such compound is trehalose,
a nonreducing disaccharide of glucose, which plays an important
physiological role as an abiotic stress protectant in a large
number of organisms, including bacteria, yeast, and inverte-
brates (5). Trehalose has been shown to stabilize dehydrated
enzymes, proteins, and lipid membranes efficiently, as well as
protect biological structures from damage during desiccation. In
the plant kingdom, most species do not seem to accumulate
detectable amounts of trehalose, with the notable exception of
the highly desiccation-tolerant ‘‘resurrection plants’’ (6). The
recent discovery of homologous genes for trehalose biosynthesis
in Selaginella lepidophylla, Arabidopsis thaliana, and several crop
plants suggests that the ability to synthesize trehalose may be
widely distributed in the plant kingdom (7). A putative plant
gene for trehalose-6-phosphate synthase (TPS) can complement
a ⌬tps1 mutant yeast strain, suggesting that the plant and yeast
gene products are functionally similar (8).
In bacteria and yeast, trehalose is synthesized in a two-step
process: trehalose-6-phosphate is first formed from UDP-
glucose and glucose-6-phosphate in a reaction catalyzed by TPS.
Trehalose-6-phosphate is then converted to trehalose by treha-
lose-6-phosphate phosphatase (TPP) (7). Metabolic engineering
for enhanced accumulation of trehalose in plants has been the
recent focus of attention in some model dicot plants (9–12).
However, in these previous studies, constitutive overexpression
of TPS and兾or TPP genes from yeast or Escherichia coli in
tobacco or potato plants resulted in undesirable pleiotropic
effects, including stunted growth and altered metabolism under
normal growth conditions (10–12).
Considering the importance of rice as a major crop, develop-
ing new cultivars with enhanced abiotic stress tolerance would
undoubtedly have an enormous impact on global food produc-
tion. We decided to improve abiotic stress tolerance by trans-
forming rice with a trehalose-6-phosphate synthase兾phospha-
tase (TPSP) fusion gene (13) that includes the coding regions of
the E. coli otsA and otsB genes (encoding TPS and TPP,
respectively). This approach has the dual advantages of neces-
sitating only a single transformation event and producing a
higher net catalytic efficiency for trehalose formation (13).
Because indica rice varieties represent 80% of rice grown
worldwide, we chose to transform the economically valuable
indica rice, Pusa Basmati-1 (PB-1), even though transforma-
tion and regeneration are more difficult than in japonica rice
Here, we show that engineering trehalose overproduction in
rice can be achieved by stress-inducible or tissue-specific expres-
sion of bifunctional TPSP fusion enzyme without any detrimen-
tal effect on plant growth or grain yield. During abiotic stress,
transgenic plants accumulated increased amounts of trehalose
and showed high levels of tolerance to salt, drought, and
low-temperature stresses, as compared with the nontransformed
plant. These results demonstrate the potential use of our trans-
genic approach in developing new rice cultivars with increased
abiotic stress tolerance and enhanced rice productivity.
Materials and Methods
Plasmid Constructs. Two binary plasmids, pSB109-TPSP and
pSB-RTSP, each containing a TPSP fusion gene (13), were
Abbreviations: TPS, trehalose-6-phosphate synthase; TPP, trehalose-6-phosphate phospha-
tase; TPSP, TPS兾phosphatase; PB-1, Pusa Basmati-1; ABA, abscisic acid; NTC, nontransgenic
control; NTS, nontransgenic stressed; PS II, Photosystem II.
**To whom correspondence should be addressed. E-mail: email@example.com.
December 10, 2002
no. 25 www.pnas.org兾cgi兾doi兾10.1073兾pnas.252637799
constructed in the pSB11 vector (14) by using standard cloning
and plasmid manipulation procedures. The components of the
plasmid within the T-DNA region and the selected restriction
enzyme sites are shown in Fig. 1 A and B. The expression cassette
in pSB109-TPSP consists of an abscisic acid (ABA)-inducible
promoter (15) that contains four tandem copies of ABA-
inducible element ABRC1 (0.18 kb) coupled with a minimal rice
actin 1 promoter (0.18 kb) and an HVA22 intron (0.24 kb). It is
linked to the TPSP coding region (2.2 kb), which was constructed
by fusing the otsA and otsB genes from E. coli after the stop
codon of the otsA gene had been removed by PCR (13) and then
ligated to the potato protease inhibitor II gene (pinII)3⬘
noncoding sequence (1.0 kb). The selection cassette includes the
cauliflower mosaic virus 35S promoter (0.74 kb), phosphinothri-
cin acetyltransferase gene (bar, 0.59 kb), and the nopaline
synthase gene 3⬘ noncoding sequence (Nos 3⬘, 0.28 kb). In
pSB-RTSP, a 1.3-kb fragment of the rice rbcS promoter (16) with
a chloroplast-targeting transit peptide (0.16 kb) is linked to the
TPSP coding region; the remaining components are similar to
those in pSB109-TPSP. During the cloning and ligation of an
⬇3.7-kb DNA fragment containing the rbcS promoter兾transit
peptide and TPSP fusion gene into the plasmid pSB-RTSP, three
additional restriction sites (SacI, SalI, and HindIII) were added
between TPSP and 3⬘ pin II. Both the plasmids (pSB109-TPSP
and pSB-RTSP) were separately transferred to Agrobacterium
tumefaciens strain LBA4404 harboring the pSB1 vector (14)
through triparental mating using the helper plasmid pRK2013.
For cocultivation, the bacteria were grown from a single colony
in liquid AB medium containing 50 mg兾liter spectinomycin at 30°C
for 3 days and were suspended at a density of 3 ⫻ 10
cells per ml
in AAM medium (17) for rice transformation.
Generating Transgenic Rice Plants. Mature seeds of indica rice
variety PB-1 were dehusked and sterilized in 70% (vol兾vol)
ethanol for 2–3 min and then transferred into 50% (vol兾vol)
Clorox solution for 40 min with gentle shaking. The seeds were
rinsed several times with sterile water. The sterilized PB-1 seeds
were then plated for callus induction on Murashige and Skoog
(MS) medium (Sigma) supplemented with 3.0 mg/liter 2,4-
dichlorophenoxyacetic acid (2,4-D)兾0.2 mg/liter 6-benzylamin-
opurine (BAP)兾300 mg/liter casein hydrolysate (CH)兾30 g/liter
maltose兾3.0 g/liter phytagel, pH 5.8 (MSCl) and grown for 21
days at 25°C in the dark. Three weeks after callus induction from
the scutellar region of the rice embryo, 150 embryogenic calli
were immersed in A. tumefaciens suspension for 10 min. Infected
calli were cocultivated in MSCl medium supplemented with 10
M acetosyringone, pH 5.2 (MSCC). After
3 days of cocultivation, calli were washed with sterile water
containing 250 mg兾liter cefotaxime and blotted on filter paper.
The calli were immediately plated on a selection medium, MSCl
medium, supplemented with 6 mg兾liter bialaphos (a gift from
H. Anzai, Meiji Seika Kaisha, Japan) and 250 mg兾liter cefo-
taxime, pH 5.8 (MSS), and incubated at 25°C in the dark for 2–3
weeks. The microcalli that had proliferated after the initial
selection were further subcultured for two selection cycles on
fresh MSS medium every 2 weeks. The actively dividing biala-
phos-resistant calli were plated on MS plant regeneration me-
dium containing 2.5 mg/liter BAP兾1.0 mg/liter kinetin兾0.5 mg/
liter naphthaleneacetic acid (NAA)兾300 mg/liter CH兾30 g/liter
maltose兾4 mg/liter bialaphos兾250 mg/liter cefotaxime兾2.0 g/liter
phytagel, pH 5.8 (MSPR) and grown at 25°C for a 10-h light兾14-h
dark photoperiod for 3–4 weeks. The regenerated plantlets were
acclimatized hydroponically in Yoshida nutrient solution (18)
for 10 days. Later on, putative primary transformants (T
generation) were transferred to pots and tested for Basta-
herbicide resistance (19); the transgenic plants were grown to
maturity in a greenhouse for further analysis.
DNA-Blot Hybridization Analysis. Leaves from nontransformed
control (NTC) plant, and representative (T
) transformants of
nine A-lines (ABA-inducible promoter) and five R-lines (rbcS
promoter) that were transformed with the plasmid pSB109-
TPSP and pSB-RTSP, respectively, were ground in liquid nitro-
gen by using a motor and pestle. Rice genomic DNA was isolated
by the guanidine-detergent lysis method by using DNAzolES
(Molecular Research Center, Cincinnati) following the manu-
facturer’s instructions. Five micrograms of the genomic DNA
was digested overnight with HindIII restriction enzyme, frac-
tionated through 0.8% agarose gel, alkali-transferred onto Hy-
bond N⫹ nylon membrane (Amersham Pharmacia), and hybrid-
ized with an
P-labeled 2.2-kb TPSP fusion gene (13) as the
probe. DNA probe preparation, hybridization, and washing of
the membrane were performed as described (19). The
labeled membrane was exposed onto autoradiogram.
Detecting Trehalose and Soluble Carbohydrates. Soluble carbohy-
drates were extracted as described (10). Extracts from 0.5 g of
homogenized fresh leaf tissue were centrifuged (10 min at 3,220
⫻ g); supernatants were passed through ion-exchange columns
consisting of 1 ml of Amberlite IR-68 (acetate form) layered on
1 ml of Dowex 50W (hydrogen form) to remove charged
Fig. 1. Schematic representation of the expression vectors and DNA-blot
hybridization analysis. Two binary plasmids, each containing the trehalose
biosynthetic fusion gene (TPSP) that includes the coding regions of the E. coli
otsA and otsB genes (encoding TPS and TPP, respectively), were constructed
and transformed into indica rice, as described in Materials and Methods.
(A) pSB109-TPSP plasmid. (B) pSB-RTSP plasmid. Shaded boxes represent pro-
moter elements (ABA, ABA-inducible; rbcS, rice rbcS; 35S, cauliﬂower mosaic
virus 35S); RB and LB represent T-DNA border on the right and left sides,
respectively. Shown is DNA-blot hybridization analysis from nontransformed
control (NTC) plant, and representative transgenic plants of nine A-lines (C)
and ﬁve R-lines (D) that were transformed with the plasmid pSB109-TPSP and
pSB-RTSP, respectively. The rice genomic DNA was digested with HindIII (a
unique site in the plasmid pSB109-TPSP, whereas two sites are present in the
plasmid pSB-RTSP) and DNA blot hybridization analysis was performed with
the 2.2-kb TPSP fusion gene as the probe. Molecular sizes (kb) are indicated.
Garg et al. PNAS
December 10, 2002
compounds. After lyophilization, samples were dissolved in
HPLC-grade water and subjected to high-performance anion
exchange chromatography with pulsed amperometric detection
by using a Dionex DX-500 series chromatograph equipped with
a Carbopac PA-1 analytical column and a Carbopac PA-1 guard
column (Dionex). Carbohydrates were eluted at a flow rate of 1.0
ml per min at 1,400 psi with 100 mM NaOH for 34 min. Major
soluble carbohydrates present were quantified by using authentic
standard sugars (Sigma). The identity of trehalose in the plant
extracts was confirmed by incubating samples with porcine-
kidney-derived trehalase enzyme (Sigma).
Salt Stress Tolerance and Determination of Plant Mineral Nutrients.
Ten seedlings for each T
generation transgenic line (R22, R38,
R80, A05, A07, and A27) and NTC were grown hydroponically
(with modest aeration) in Yoshida nutrient solution (18) in a
growth chamber at 25 ⫾ 3°C for a 10-h light兾14-h dark photo-
period (photon flux density of 280
mol photons per m兾s) and
with relative humidity of 50–60%. After 5 weeks, 50% of the
seedlings were subjected to 100 mM NaCl stress (conductivity of
10–12 dS兾m). Nutrient solutions were replaced every week.
After 4 weeks of continuous salt stress, shoot and root samples
were separately harvested for fresh and dry weight determina-
tion. For mineral nutrient analysis, 150 mg of ground dry matter
was digested in concentrated HNO
overnight at 120°C. Samples
then were dissolved in HNO
(1:1, vol兾vol) at 220°C,
resuspended in 5% (vol兾vol) HNO
, and analyzed for elemental
composition of sodium (Na
), potassium (K
), calcium (Ca
and iron (Fe) by means of simultaneous inductively coupled
argon-plasma emission spectrometry (ICP trace analyzer; Plant,
Soil, and Nutrition Laboratory, U.S. Department of Agriculture-
Agriculture Research Service, Cornell University, Ithaca, NY).
Drought and Low-Temperature Stress Tolerance. Seedlings from six
transgenic lines and nontransformed line were
grown individually in 10-cm ⫻ 10-cm pots irrigated with Yoshida
nutrient solution for 5 weeks before performing the drought- or
low-temperature stress experiment. Drought stress (water defi-
cit) was conducted by first withholding irrigation for 3 days to
allow the soil in the pot to dry. Then, the first drought cycle of
100 h was initiated, followed by rewatering for 2 days. The
drought-stress cycle was repeated for another 100 h, and the
plants were allowed to recover by watering every day for 3 weeks.
Low-temperature stress was conducted on five-week-old seed-
lings by exposing them to 10°C for 72 h under a 10-h light兾14-h
dark photoperiod (photon flux density of 280
per m per s) and a relative humidity of 50–60%; the seedlings
were then allowed to recover under normal growth conditions at
25 ⫾ 3°C.
Protein Extraction and Immunoblotting. Proteins were extracted
from 0.2 g of homogenized fresh leaf tissue in protein extraction
buffer (20 mM Tris䡠HCl, pH 8.0兾10 mM EDTA兾30 mM NaCl兾2
mM phenylmethane sulfonyl fluoride for1hat4°C). The
homogenate was clarified by centrifugation at 12,000 ⫻ g for 15
min at 4°C. The procedure for immunoblotting was essentially
the same as described (20). The anti-TPSP protein polyclonal
antibody was used at a 1:1,500 dilution for Western blot analysis,
using an alkaline phosphatase color reaction for detection of the
protein, as per the manufacturer’s instruction (Bio-Rad).
Chlorophyll Fluorescence Parameters. Fv兾Fm and
sured by using a pulse amplitude modulated fluorometer (FMS2,
Hansatech Instruments, Pentney King’s Lynn, U.K.) to estimate
photo-oxidative damage to the Photosystem II (PS II) reaction
center and the quantum efficiency of PS II photochemistry under
ambient light conditions, respectively, as described (21). Mea-
surements were made on the youngest, fully expanded leaves.
were first determined under ambient
light; the same leaves were then dark-adapted for 10 min before
measurement of Fv兾Fm.
Results and Discussion
Transgenic Plants with Enhanced Trehalose Levels Are Phenotypically
Normal and Fertile. Two plasmid constructs, pSB109-TPSP (Fig.
1 A) and pSB-RTSP (Fig. 1B), each containing the TPSP fusion
gene, were introduced into indica rice cells of PB-1 by Agrobac-
terium-mediated gene transfer (17). In the plasmid construct
pSB109-TPSP, an ABA and stress-inducible promoter (15)
drives the fusion gene for cytosolic expression. In the other
plasmid, pSB-RTSP, the light-regulated promoter (16) of the
Rubisco small subunit gene, rbcS, from Oryza sativa with a transit
peptide drives the fusion gene for chloroplast targeting in the leaf
mesophyll cells. A large number of putative transgenic PB-1
generation) were regenerated (Table 1, which is
published as supporting information on the PNAS web site,
www.pnas.org); these plants included 28 A-lines (ABA-inducible
promoter) and 76 R-lines (rbcS promoter). Integration of the
TPSP transgene was confirmed by DNA-blot hybridization anal-
ysis (Fig. 1 C and D). Based on the T-DNA junction fragment
analysis, ⬇40% of the transgenic plants transformed with either
of the plasmids harbor a single copy, and 35–45% of plants
harbor two or three copies of the transgene.
Most of the 90 independent primary transformants (T
contained a low copy number of the transgene showed a normal
phenotype and were completely fertile. In contrast to previous
reports that used constitutive promoters driving individual TPS
and兾or TPP genes, the use of stress-inducible or tissue-specific
promoters in this work appears to minimize the negative effects
of the transgene on plant growth. The T
plants were self-
pollinated to obtain segregating T
progeny for genetic and
HPLC analysis. Forty-five transgenic lines showed a segregation
pattern of 3:1 for the basta-herbicide resistance marker gene.
HPLC analysis of leaf extracts showed that transgenic lines had
a trehalose content that was between three times and eight times
that of the nontransgenic plants (17 ⫾ 5
g of trehalose per g of
fresh weight). The identity of trehalose in the plant tissue
extracts was confirmed by incubating samples in porcine tre-
halase followed by chromatographic analysis of the monosac-
charide products (Fig. 6, which is published as supporting
information on the PNAS web site). Physiological experiments
were conducted for abiotic stress tolerance on homozygous
plants through the T
generation, because gene silencing has
been reported to occur in the T
generation, even though T
generation plants were not silenced (22). The results from
many independent transgenic lines were consistent for salt- and
drought-stress tolerance in each generation, except in few trans-
genic lines which had multiple copies of the transgene (data not
Transgenic Plants Are Salt Tolerant and Maintain Balanced Mineral
Nutrition. The T
transgenic plants with either one or two copies
of the transgene showed markedly enhanced salt tolerance
during and subsequent to 4 weeks of 100 mM NaCl treatment
under hydroponic growth conditions. Six independent transgenic
plant lines (three A-lines and three R-lines) were analyzed in
detail. For clarity of presentation, results from two representa-
tive transgenic lines (R80 and A05) are shown (Fig. 2); results for
the other four lines were very similar to the two lines presented.
After prolonged exposure to salt stress, almost all of the trans-
genic plants survived and displayed vigorous root and shoot
growth. In contrast, all of the nontransformed stressed (NTS)
plants were either dead or nearly dead because of severe salt
damage to the leaves and concomitant loss of chlorophyll.
Transgenic plants developed longer and thicker roots than NTS
plants after salt stress (Fig. 2A). Salt stress severely inhibited the
www.pnas.org兾cgi兾doi兾10.1073兾pnas.252637799 Garg et al.
growth of shoot and roots of NTS plants, as indicated by their
lower dry weights compared with NTC plants. Shoot and root
dry weights of both the transgenic lines (Fig. 2B) approached
those of NTC plants, and after removal of salt stress, the
transgenic plants were able to grow, flower, and set normal
viable seeds. To determine whether the TPSP gene product was
present in the salt-stressed plants, total protein was isolated from
the leaf samples for Western blot analysis. Immunoblot analysis
using polyclonal antibodies raised against the fusion protein
showed the presence of a protein with the expected apparent
molecular mass of 88 kDa only in the transgenic plants (Fig. 2C).
To assess how trehalose accumulation in transgenic rice
affected plant mineral nutrition during salt stress, shoot and root
mineral content for the six independent transgenic lines and two
nontransgenic lines were determined by using inductively cou-
pled plasma emission spectrometry (Table 2, which is published
as supporting information on the PNAS web site). After con-
tinuous salt stress (100 mM NaCl) for 4 weeks, NTS plants
showed a very large increase in Na
content in both shoots and
roots compared with NTC, whereas the increase in the shoots of
all of the transgenic plants was much smaller (Fig. 2D). The Na
content of transgenic plant shoots was only 30–35% of the NTS
plants after salt stress. The observed differences in shoot Na
content between transgenic and NTS plants could be caused in
part by a growth dilution because of the much faster growth rate
of the transgenic plants under salt stress. Alternatively, trehalose
might have played a direct or indirect role in maintaining ion
selectivity and, thus, facilitating cellular Na
possibility is consistent with the report that in salt-stressed rice
seedlings, the accumulation of Na
in leaf tissues was not
prevented by exogenous proline. In contrast, treatment with
exogenous trehalose significantly reduced the salt-induced ac-
cumulation of Na
in the leaves (23).
Transgenic lines R80 and A05 maintained shoot to root K
homeostasis both under nonstress and salt-stress conditions
(Table 2). After salt stress, the levels of shoot and root K
content in transgenic plants was similar to the nonstressed
controls, while a fourfold decrease in root K
in the NTS plants
was seen (Fig. 2E). Thus, the transgenic plants were able to
maintain a higher level of selectivity for K
the roots and Na
exclusion from the shoots compared with the
NTS plants. The maintenance of the Na
ratio in both shoot
and roots of transgenic plants (Fig. 2F) correlated with nearly
normal plant growth and may be the basis for minimizing Na
toxicity under salt stress. It is generally accepted that the
maintenance of Na
homeostasis is an important aspect of
salt tolerance (24, 25).
Several other changes in plant mineral status that may have
played indirect roles in stress tolerance were seen in the trans-
genic lines compared with the NTCs. It was found that salt stress
led to a significant increase in root and shoot Ca
content in the
NTS lines, whereas in the transgenic lines, this Na-mediated
increase in Ca
content was only observed in the shoots and not
the roots (Table 2). This rise in Ca
may be caused by
alterations in the ion selectivity of the transporters at high
concentrations of Na
(25). We also found significantly higher
levels of shoot Fe ion content in the transgenic lines compared
with the NTCs (Table 2). It has been well documented that Fe,
Cu, and Zn ions are essential for the function of critical
antioxidant enzymes such as the superoxide dismutases that play
a role in scavenging reactive oxygen species during a number of
abiotic stresses (25, 26). In general, the relationship between salt
stress and plant mineral content is complex, and the links
between elevated trehalose content and improved mineral status
during salt stress are not known.
Transgenic Plants Are Drought Tolerant. To study drought tolerance,
5-week-old nontransformed and transgenic seedlings grown in soil
were subjected to two cycles of 100 h of drought stress. After the
drought treatments, all 15 plants of each line showed wilting and
drought-induced rolling of the young leaves. Nontransgenic plants
exhibited rolling of leaves within 48 h of the stress as compared with
considerably fewer visual symptoms in transgenic plants during the
same time period. After two cycles of 100 h of drought stress and
subsequent watering for 3 weeks, the growth of both the transgenic
lines, R80 and A05 (Fig. 3B), were almost identical to nonstressed
control plant (Fig. 3A). In contrast, the growth of the drought-
stressed NTS was severely inhibited (Fig. 3B).
Transgenic Plants Produced Increased Amounts of Trehalose and
Other Soluble Carbohydrates. To evaluate our hypothesis that
trehalose accumulation in plants might act as a positive regulator
of stress tolerance, we measured the levels of trehalose and other
soluble carbohydrates (Table 3, which is published as supporting
information on the PNAS web site). A low but significant
amount of trehalose was detected in the shoots (17
weight) of NTC plants; these levels increased significantly under
salt or drought stresses. The transgenic plants grown under
control conditions exhibited trehalose levels comparable with
the NTS plants (Fig. 4). After salt stress, the transgenic lines
(R80 and A05) showed 2.5–3 times higher shoot trehalose levels
compared with NTS plants, whereas after drought stress, treha-
lose levels in the transgenic lines increased 3- to 9-fold (Fig. 4).
Despite the similarities in tolerance levels exhibited by trans-
genic plants engineered to increase trehalose synthesis in either
the cytosol or chloroplast, R-lines showed considerable protec-
Fig. 2. Salt tolerance of rice plants and changes in mineral nutrition caused
by salt stress. (A) Plant roots after 4 weeks of continuous 100 mM NaCl stress;
the plants were not stressed in NTC. (B) Dry weight of shoots (black bars) and
roots (white bars) of plants grown under salt stress (NTS, R80, and A05) or no
stress (NTC) conditions. (C) Western blots of leaf extracts (20
g of proteins)
immediately after salt stress of plants. (D–F) Plant mineral nutrient content in
shoots (black bars) and roots (white bars) under salt stress (NTS, R80, and A05)
or no stress (NTC) conditions. (D)Na
ratio. The ionic
concentration is presented as mg兾g dry weight. Values are the means ⫾ SD
(n ⫽ 5).
Garg et al. PNAS
December 10, 2002
tion at much lower trehalose concentrations during drought
stress (Table 3). In general, there was no obvious relationship
between trehalose accumulation and stress tolerance among the
transgenic lines evaluated (data not shown). On the other hand,
the difference in trehalose levels between the transgenic and
nontransgenic lines clearly correlates with increased tolerance to
Transgenic Plants Show Improved Photosystem II Function. During
many different abiotic stresses, a reduction in photosynthesis and
the subsequent production of reactive oxygen species are
thought to be a major contributor to decreased plant perfor-
mance and photooxidative damage. The effects of increased
trehalose accumulation on photosynthesis during drought
stress were assessed by determination of the quantum yield of
PS II photochemistry (⌽
) by using in vivo chlorophyll
fluorescence techniques (21). ⌽
is a measure of the pho-
tosynthetic performance of the plant under ambient light
conditions. After the first cycle of 100 h of drought stress, the
quantum yield of PS II photochemistry in NTS plants de-
creased by ⬇68%, whereas the activity of the two best-
performing transgenic lines (R80 and A05) only decreased by
29–37% compared with the nonstressed controls (Fig. 3C).
Similarly, drought-induced decreases in the fluorescence pa-
rameter Fv兾Fm, which is a measure of accumulated photo-
oxidative damage to PS II, were considerably smaller in the
transgenic lines than in the NTS plants (Fig. 3D). In other
independent experiments, similar results were obtained for
both low-temperature stress (Fig. 7, which is published as
supporting information on the PNAS web site) and salt stress
(data not shown), indicating the common role that mainte-
nance of photosynthetic capacity plays in tolerance to these
Transgenic Plants Have Increased Photosynthetic Capacity Under
Nonstress Conditions. Improved photosynthesis under abiotic
stress conditions is known to limit photo-oxidative damage and
permit continued growth (27) and is clearly suggested by the data
in Fig. 3. Under the same conditions, transgenic plants exhibited
soluble carbohydrate levels that were ⬇20% higher than those of
corresponding NTC plants, including subtle changes in levels of
glucose, fructose, and sucrose (Table 3). Both of these results are
consistent with the suggestion that trehalose may be involved in
sugar sensing and modulating carbon metabolism (7, 28). The
ability of trehalose to modulate photosynthetic capacity has been
demonstrated recently (29) in transgenic tobacco plants express-
ing E. coli trehalose biosynthetic genes. Plants with enhanced
TPS expression exhibited a higher photosynthesis per unit of leaf
area than nontransgenic controls, whereas those over-expressing
TPP showed diminished rates of photosynthesis. These data lead
Fig. 5. Photosystem II electron transport rate in nontransformed and two
independent, ﬁfth generation transgenic plants grown under control condi-
tions. The electron transport rate under increasing irradiance was calculated
from chlorophyll ﬂuorescence measurements on the youngest fully expanded
leaf of NTC (
), R80 (
), and A05 (
) at 360 ppm of CO
,25°C, and 50% relative
humidity after 10 weeks of growth. Values are the means ⫾ SD (n ⫽ 9). Data
are normalized to the average light-saturated rate of the nontransgenic
Fig. 3. Appearance of plants and chlorophyll ﬂuorescence parameters
during drought stress. Five-week-old nontransformed and T
transgenic (R80 and A05) seedlings grown in soil were subjected to two cycles
of 100 h of drought stress followed by watering for 3 weeks. (A) Plants grown
under well watered conditions (NTC, nontransgenic plants). (B) Plants of the
same age after two cycles of drought-stress treatment (NTS, nontransgenic
plants after drought stress). (C and D) Chlorophyll ﬂuorescence measurements
on young, fully expanded leaves during the ﬁrst cycle of 100 h of continuous
drought stress. (C) ⌽
, a measure of the efﬁciency of PS II photochemistry
under ambient growth conditions. (D) Decreases in Fv兾Fm are a measure of
photooxidative damage to PS II.
, nontransformed plants;
Dotted lines represent the range of values for nonstressed control plants of all
lines. Data represent means ⫾ SD (n ⫽ 5) from independent plants.
Fig. 4. Trehalose content in shoots of transgenic (R80 and A05) and non-
transgenic plants with or without stress. Trehalose accumulation under non-
stressed (white bars), salt-stressed (100 mM NaCl for 4 weeks, hatched bars), or
drought-stressed (100 h, black bars) conditions.
www.pnas.org兾cgi兾doi兾10.1073兾pnas.252637799 Garg et al.
them to conclude that it is trehalose-6-P and not trehalose that
is modulating photosynthetic capacity (29).
Fig. 5 shows the light intensity dependence of PS II electron
transport rates, as determined by
measurements (21) for
nontransgenic rice and transgenic lines R80 and A05 measured
under control (nonstress) conditions. Although the differences
in photosynthesis are small at limiting light intensities, at light
saturation, the rates of photosynthesis in the transgenic plants
are 5–15% higher than in the NTCs. At light saturation, pho-
tosynthetic rate is limited by the capacity of the dark reactions,
in particular the Calvin cycle and triose phosphate utilization in
the cytoplasm (27). Together with the observed higher levels of
soluble carbohydrate under both stress and nonstress conditions
(Table 3), the elevated levels of light-saturated photosynthesis in
the transgenic plants supports the suggestion that in plants,
trehalose acts as a regulator of sugar sensing and, thus, the
expression of genes associated with carbon metabolism (29). The
presence of a higher capacity for photosynthesis before stress
provides a larger sink for the products of photosynthesis during
stress, thus limiting the extent of excess-light-induced photooxi-
dative damage and accounting, in part, for the more vigorous
growth of the transgenic lines during stress. Interestingly, the
higher efficiency of trehalose synthesis by the TPSP fusion gene
product (13) would suggest that trehalose, rather than trehalose-
6-P is leading the enhanced capacity for photosynthesis.
We have demonstrated that regulated overexpression of treha-
lose biosynthetic genes in rice has considerable potential for
improving abiotic stress tolerance and, at the same time, aug-
menting productivity under both stress and nonstress conditions.
This work showed successful conferment of tolerance to multiple
abiotic stresses by means of overexpression of trehalose biosyn-
thesis without the negative pleiotropic effects seen in previous
studies. The modest increase in trehalose levels in transgenic
lines, using either the tissue-specific or stress-dependent pro-
moters, resulted in a higher capacity for photosynthesis and a
concomitant decrease in the extent of photo-oxidative damage
during stress. In addition, trehalose must be interacting with
other physiological processes to account for changes in ion
uptake and partitioning during salt stress. Because other cereal
crops, like rice, are also sensitive to abiotic stresses, it is likely
that overexpression of trehalose biosynthetic genes in maize and
wheat may also confer high levels of abiotic stress tolerance.
We thank A. Jagendorf, M. Hanson, W. B. Miller, and T. L. Setter for
critical review of the manuscript. We also thank J. Lee, H. Manslank, and
A. S. Stolfi for technical assistance. This research was supported in part
by Rockefeller Foundation Grant RF 98001-606 (to R.J.W.), by a
postdoctoral fellowship (to A.K.G.) from the Rockefeller Foundation,
and by grants from the Ministry of Science and Technology of Korea
through the Crop Functional Genomics Center (to J.-K.K. and Y.D.C.).
1. Boyer, J. S. (1982) Science 218, 443–448.
2. Zhang, H. X., Hodson, J. N., Williams, J. P. & Blumwald, E. (2001) Proc. Natl.
Acad. Sci. USA 98, 12832–12836.
3. Hare, P. D., Cress, W. A. & van Staden, J. (1998) Plant Cell Environ. 21,
4. Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D. & Somero, G. N. (1982)
Science 217, 1214–1222.
5. Crowe, J. H., Hoekstra, F. A. & Crowe, L. M. (1992) Annu. Rev. Physiol. 54, 579–599.
6. Wingler, A. (2002) Phytochemistry 60, 437–440.
7. Goddijn, O. J. & van Dun, K. (1999) Trends Plant Sci. 4, 315–319.
8. Zentella, R., Gallardo, J. O. M., Van Dijck, P., Mallol, J. F., Bonini, B.,
Van Vaeck, C., Gaxiola, R., Covarrubias, A. A., Sotelo, J. N., Thevelein, J. M.
& Iturriaga, G. (1999) Plant Physiol. 119, 1473–1482.
9. Holmstrom, K. O., Mantyla, E., Welin, B., Mandal, A. & Palva, E. T. (1996)
Nature 379, 683– 684.
10. Goddijn, O. J., Verwoerd, T. C., Voogd, E., Krutwagen, R. W., de Graaf, P. T.,
van Dunn, K., Poels, J., Ponstein, A. S., Damm, B. & Pen, J. (1997) Plant
Physiol. 113, 181–190.
11. Romero, C., Belles, J. M., Vaya, J. L., Serrano, R. & Culianez-Macia, F. A.
(1997) Planta 201, 293–297.
12. Pilon-Smits, E. A. H., Terry, N., Sears, T., Kim, H., Zayed, A., Hwang, S.,
van Dun, K., Voogd, E., Verwoerd, T. C., Krutwagen, R. W. & Goddijn,
O. J. M. (1998) J. Plant Physiol. 152, 525–532.
13. Seo, H. S., Koo, Y. J., Lim, J. Y., Song, J. T., Kim, C. H., Kim, J. K., Lee, J. S.
& Choi, Y. D. (2000) Appl. Environ. Microbiol. 66, 2484–2490.
14. Komari, T., Hiei, Y., Saito, Y., Murai, N. & Kumashiro, T. (1996) Plant J. 10,
15. Su, J., Shen, Q., David Ho, T. H. & Wu, R. (1998) Plant Physiol. 117, 913– 922.
16. Kyozuka, J., McElroy, D., Hayakawa, T., Xie, Y., Wu, R. & Shimamoto, K.
(1993) Plant Physiol. 102, 991–1000.
17. Hiei, Y., Ohta, S., Komari, T. & Kumashiro, T. (1994) Plant J. 2, 271–282.
18. Yoshida, S., Forno, D. A., Cook, J. H. & Gomez, K. A. (1976) in Laboratory
Manual for Physiological Studies of Rice (International Rice Research Institute,
Philippines), pp. 61–66.
19. Roy, M. & Wu, R. (2001) Plant Sci. 160, 869–875.
20. Xu, D., Duan, X., Wang, B., Hong, B., Ho, T. & Wu, R. (1996) Plant Physiol.
21. Saijo, Y., Hata, S., Kyozuka, J., Shimamoto, K. & Izui, K. (2000) Plant J. 23,
22. Iyer, L. M., Kumpatla, S. P., Chandrasekharan, M. B. & Hall, T. C. (2000) Plant
Mol. Biol. 43, 323–346.
23. Garcia, A. B., de Engler, J. A., Iyer, S., Gerats, T., van Montagu, M. & Caplan,
A. (1997) Plant Physiol. 115, 159–169.
24. Rus, A., Yokoi, S., Sharkhuu, A., Reddy, M., Lee, B. H., Matsumoto, T. K.,
Koiwa, H., Zhu, J. K., Bressan, R. A. & Hasegawa, P. M. (2001) Proc. Natl.
Acad. Sci. USA 98, 14150–14155.
25. Epstein, E. (1998) Science 280, 1906–1907.
26. Alscher, R. G., Erturk, N. & Heath, L. S. (2002) J. Exp. Bot. 53, 1331-
27. Owens, T. G. (1996) in Photosynthesis and the Environment, ed. Baker, N. R.
(Kluwer, Dordrecht, The Netherlands), pp. 1–23.
28. Thevelein, J. M. & Hohmann, S. (1995) Trends Biochem. Sci. 20, 3–10.
29. Paul, M., Pellny, T. & Goddijn, O. (2001) Trends Plant Sci. 6, 197–200.
Garg et al. PNAS
December 10, 2002