Plant Science 236 (2015) 260271

Contents lists available at ScienceDirect

Plant Science
journal homepage: [Link]/locate/plantsci

Genetic analyses of the interaction between abscisic acid and

gibberellins in the control of leaf development in Arabidopsis thaliana
Ming-Hau Chiang a,b , Hwei-Ling Shen b , Wan-Hsing Cheng a,b,
a
b

Graduate Institute of Life Sciences, National Defense Medical Center, Taipei, Taiwan
Institute of Plant and Microbial Biology, Academia Sinica, Taipei, Taiwan

a r t i c l e

i n f o

Article history:
Received 16 January 2015
Received in revised form 15 April 2015
Accepted 16 April 2015
Available online 23 April 2015
Keywords:
Abscisic acid
Gibberellin
Arabidopsis thaliana
Leaf growth
Transcriptome
Gene ontology

a b s t r a c t
Although abscisic acid (ABA) and gibberellins (GAs) play pivotal roles in many physiological processes in
plants, their interaction in the control of leaf growth remains elusive. In this study, genetic analyses of
ABA and GA interplay in leaf growth were performed in Arabidopsis thaliana. The results indicate that for
the ABA and GA interaction, leaf growth of both the aba2/ga20ox1 and aba2/GA20ox1 plants, which were
derived from the crosses of aba2 ga20ox1 and aba2 GA20ox1 overexpressor, respectively, exhibits partially additive effects but is similar to the aba2 mutant. Consistently, the transcriptome analysis suggests
that a substantial proportion (4565%) of the gene expression prole of aba2/ga20ox1 and aba2/GA20ox1
plants overlap and share a pattern similar to the aba2 mutant. Thus, these data suggest that ABA deciency
dominates leaf growth regardless of GA levels. Moreover, the gene ontology (GO) analysis indicates gene
enrichment in the categories of hormone response, developmental and metabolic processes, and cell wall
organization in these three genotypes. Leaf developmental genes are also involved in the ABA-GA interaction. Collectively, these data support that the genetic relationship of ABA and GA interaction involves
multiple coordinated pathways rather than a simple linear pathway for the regulation of leaf growth.
2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
Angiosperm leaves vary greatly in shape and size. The leaf is an
important aerial tissue in plants that plays pivotal roles in many
aspects of plant physiology, such as photoperception and photosynthesis. Thus, a better understanding of the mechanisms of leaf
growth is essential for increasing plant growth and productivity.
The early events of leaf development include the following three
processes: leaf primordium initiation, polarity determination, and
marginal meristem development. After development, the leaf blade
grows outward and expands until reaching its full size [1,2]. At the
initial stage of leaf development, growth is driven mostly by cell
proliferation [3]. Later, cell ceases division from the leaf top to the
base sequentially and begins to expand [4]. The nal size of the

Abbreviations: ABA, abscisic acid; GAs, gibberellins; GO, gene ontology; ABA2,
ABSCISIC ACID DEFICIENT 2; GA20ox, GA 20-oxidase; DAS, days after stratication;
EI, endoreduplication index; PA, phaseic acid; DPA, 4-dihydrophaseic acid; ABAGE,
ABA--glucosyl ester.
Corresponding author at: Institute of Plant and Microbial Biology,
Academia Sinica, 128, Section 2, Academia Road, Nankang, Taipei 115, Taiwan.
Tel.: +886 2 2787 1177; fax: +886 2 2782 7954.
E-mail address: whcheng@[Link] (W.-H. Cheng).
[Link]
0168-9452/ 2015 Elsevier Ireland Ltd. All rights reserved.

leaf is controlled by the initial cell number of the leaf primordium,

followed by cell division and cell growth. Leaf growth regulation
integrates both external environmental clues and internal signals.
The latter are involved in the intrinsic cross-regulatory network,
including phytohormones. Almost every plant hormone exerts an
effect on leaf growth to a varying degree. These plant hormones
form a complex network that synergistically or antagonistically
regulates leaf growth [59] through the spatial and temporal coordination that mediates cell division and cell expansion.
Abscisic acid (ABA) is a classic plant hormone that affects seed
dormancy, germination [10,11], stomatal closure [12], and stress
tolerance [13,14]. In addition, ABA appears to play a direct role in
the repression of certain cell cycle genes, such as CDKA [15] and
DNA replication initiation CDT1 [16]. In Arabidopsis, leaf growth is
inhibited by ABA under stress conditions [17]. Thus, ABA is considered a growth inhibitor [18]. However, ABA-decient mutants
often display a smaller plant stature than wild type [19], which suggests that endogenous physiological concentrations of ABA may act
as growth promoters. At the cellular level, diminished leaf growth
in ABA-decient mutants, such as aba1, aba2, and aba3, is due to
reduced cell area and cell number [20,21]. In fact, ABA maintains
shoot development, especially in leaf expansion, in well-watered
Arabidopsis plants [22]. The dual functions of ABA suggest that ABA

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

acts as a growth inhibitor under stress but a growth promoter under

normal conditions [23].
ABA is a C15 compound synthesized via the carotenoid pathway through enzyme-catalyzed stepwise reactions [14,18]. Among
these enzymes, short-chain dehydrogenase/reductase 1 (SDR1),
encoded by ABSCISIC ACID DEFICIENT 2 (ABA2), catalyzes the
multiple-step conversion of cytosolic xanthoxin to abscisic aldehyde [23,24], which is further oxidized to form ABA by abscisic
aldehyde oxidases (AAOs) [25] together with a molybdenum
cofactor sulfurase encoded by ABA3 [2628]. Although SDRs belong
to a superfamily, only SDR1 is involved in the ABA biosynthetic
pathway [29]. As the Arabidopsis genome contains only a single copy of ABA2, aba2 mutants thus exhibit severe defects in
ABA production, small plant size and early owering phenotypes
[23,24,27,30]. Nevertheless, transgenic plants overexpressing ABA2
result in ABA accumulation, the enhancement of seed dormancy,
and tolerance to salt stress, but resemble the wild type in size [31].
Gibberellins (GAs) are diterpene phytohormones that regulate
growth and development throughout the life cycle of plants. In
Arabidopsis, GA promotes seed germination, leaf and root growth,
oral induction, bolting stem elongation, anther and petal development, and fruit and seed development [8]. Severe GA-decient
mutants reduce the size of rosette leaves, indicating that bioactive GA promotes leaf expansion [32,33]. It has been reported that
GA controls leaf growth by affecting both cell proliferation and
expansion through degradation of GA-signaling repressor DELLA
proteins. DELLAs repress cell proliferation via the transcriptional
suppression of cell cycle inhibitors KIP-RELATED PROTEIN 2 (KRP2)
and SIAMESE (SIM) family genes [34].
In GA biosynthesis, GA 20-oxidases (GA20oxs) are the key
enzymes and comprise a ve-member gene family. GA20ox1 is
the primary gene involved in stem elongation and shoot growth.
GA20ox catalyzes the removal of C20 to form the C19 structure,
and GA3-oxidase (GA3ox) catalyzes the nal step by inserting a
3-hydroxyl group to form the biologically active product GA4
[35]. Genetic analyses indicated that severe GA-decient mutants
reduce the size of rosette leaves [32,33,36], whereas transgenic
plants overexpressing GA20ox1 exhibit GA accumulation and the
promotion of leaf growth [37,38]. Although the active forms of GA4
and GA1 are not reduced in the ga20ox1 mutant, the concentrations of GA34 and GA8 , the deactivated metabolites of GA4 and GA1 ,
respectively, are substantially reduced, producing a semi-dwarf
phenotype [39].
ABA and GA play antagonistic roles in the regulation of numerous developmental processes, including seed germination, plant
growth and owering time [9,18]. Genetic studies have established that ABA inhibits seed germination, but GA promotes it.
Lon-Kloosterziel et al. [27] used a genetic approach to screen the
Arabidopsis mutants aba2 and aba3, which allow germination on
medium containing the GA biosynthesis inhibitor paclobutrazol
[27]. Thus, it is generally believed that seed germination is largely
based on the balance of ABA to GA rather than on the individual hormone concentrations [40]. In cereal, ABA and GA exhibit
antagonistic action on -amylase expression in a model system for
studying ABA and GA interaction in monocots [41]. In this interaction, ABA activates the Ser/Thr protein kinase PKABA1, which
in turn negatively regulates GA-induced GAMyb expression [42]
and thus inhibits -amylase expression [43,44]. It is known that
various phytohormones may alter oral transition and owering
time [45]. For instance, exogenous application of GAs enhances
owering time in Arabidopsis [46], but the severe GA-decient
mutant ga1 or the defect in GA signaling in gai cause delayed
owering [47,48]. Although the severe ABA-decient mutant aba2
promotes early owering [23,24,27,30], transgenic plants overexpressing Arabidopsis ABA2 or NCED3 exhibit ABA accumulation
and owers with normal dynamics [31,49,50]. In recent years, the

261

ABA and GA interaction in the control of owering time has been

reported to be mediated by a complex regulatory network and was
revealed by partially coordinated action [49].
ABA, GA, and ethylene act at different developmental stages
to adapt the young leaf in response to stress. Based on transcriptome analysis, ABA primarily functions in mature cells, whereas
GAs and ethylene mediate the expansion and division of cells in
response to stress [51,52]. Despite the signicance of ABA and GA
hormones, their interaction in the control of leaf growth remains
largely unknown. In this study, we demonstrated that the leaf
growth of aba2/ga20ox1 and aba2/GA20ox1, which were derived
from the crosses of aba2 ga20ox1 and aba2 GA20ox1 overexpressor, respectively, was due to a partially additive effect of ABA
and GA action but was similar to the aba2 mutant. This suggests that
ABA deciency plays a more prominent role than GA function in
the control of leaf growth. Consistent with this idea, transcriptome
analysis of aba2/ga20ox1 and aba2/GA20ox1 revealed that substantial levels of gene expression proles overlapped and largely
exhibited expression patterns that were similar to the aba2 mutant.
Moreover, GA20ox1 contained high levels of both GA and ABA and
produced the largest leaf area among the tested genotypes. These
data suggest that GA plays a dominant role in the control of leaf
growth with a normal level or the accumulation of ABA. Collectively, genetic and transcriptome analyses reveal the interactions
between ABA and GA by coordinating multiple pathways in the
control of leaf growth.
2. Materials and methods
2.1. Plant materials and growth conditions
The mutants and transgenic lines all had a background of Col-0
accession of Arabidopsis thaliana. The aba2 mutant is also known as
glucose insensitive 1-3 (gin1-3), which has a 53-bp deletion at the
beginning of the 2nd exon, resulting in a frame shift and an early
stop codon at the amino acid in the 33 position [23]. The ABA2overexpressing (35S::ABA2) transgenic plants used in this study
have been described previously [31]. The ga20ox1 (SALK 016701)
mutant, also known as ga5-3 [53], was requested from the Arabidopsis Biological Resource Center (ABRC, OH). For transgenic
plants overexpressing GA20ox1, a full-length cDNA of GA20ox1
was amplied by RT-PCR and fused to a constitutive 35S promoter (35S::GA20ox1) in the binary vector pSMAB704. Homozygous
transgenic plants (GA20ox1 overexpressors) were obtained at the
T3 generation through antibiotic screening. The combinations of
mutants and transgenic lines, i.e., aba2/ga20ox1 and aba2/GA20ox1,
were created by crossing aba2 with the ga20ox1 mutant or the
GA20ox1 overexpressor. The homozygous lines were obtained by
antibiotic screens and genotyping. Unless stated otherwise, plants
were grown at 22 C under a 16-hr light/8-hr dark regime with a
light intensity of approximately 80100 mol m2 s1 .
2.2. Leaf area measurements
Plants were grown in soil for 21 days, and then cotyledons and
rosette leaves were excised in order from plants per genotype followed by scanning at 600 dpi on a at-bed scanner. Leaf areas were
determined by capturing the leaf image using the ImageJ plugin
LeafJ [54]. For long-term observation of leaf growth, the fth rosette
leaf was harvested at two-day intervals from 17 to 39 days after
sowing.
2.3. Assay of cell size and number
The fth leaf of plants grown in soil for 21 days was excised,
xed with FAA (formalin:acetic acid:alcohol (70%) = 5:5:90, V:V:V)

262

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

and cleared in chloral hydrate solution [55]. Palisade mesophyll

cells in the center of the leaf blade on either side of the mid-vein
were observed under an Axio Imager A1 microscope (Carl Zeiss,
Germany). To determine cell size, photographs of palisade mesophyll cells were used to measure cell area with the Image Pro 6.0
software. At least 50 cells per leaf were examined, and an average
cell area was obtained from 6 leaves of each genotype. The total
number of palisade mesophyll cells was calculated by dividing the
leaf blade area by the average cell area.

2.4. Nuclear ploidy analysis

The assay of ploidy levels essentially followed the protocol described by Dolezel et al. [56]. In general, approximately
20 mg of the fth leaf tissue was chopped with a razor
blade in 1 mL nuclei isolation buffer (Tris MgCl2 buffer with
50 g/mL DNA uorochrome propidium iodide and 50 g/mL
RNase). The homogenate was ltered through a 20 m mesh.
The distribution of nuclear DNA contents was analyzed using
the MoFlo XDP Cell Sorter (Beckman Coulter). Approximately
10,000 nuclei were collected for each sample. The endoreduplication index (EI) was calculated using the following equation:
EI = [(0) + (1 %4C) + (2 %8C) + (3 %16C) + (4 %32C)]/100.

2.5. Hormone content measurement

Plants were grown in soil for 21 days, and then the aerial tissues
were harvested. The harvested tissues were frozen in liquid nitrogen and stored in a 80 C freezer until frozen dry. Each 100 mg
of dried tissues was used for hormone measurement with an LCMS/MS-based analysis to determine the levels of abscisic acid and
its catabolites [57].

2.7. Quantitative real-time PCR (qRT-PCR)

The total RNAs used for qRT-PCR were the same samples used
for microarray experiments. Superscript II Reverse Transcriptase
(Life Technologies) was used to synthesize cDNA. Real-time amplication in the presence of Power SYBR Green PCR Master Mix
(Life Technologies) was conducted on the Applied Biosystems 7500
Real-Time PCR System. All processes were performed according
to the manufacturers instructions and the guidelines proposed by
Udvardi et al. [58]. Primers were designed using Primer Express 2.0
software (Applied Biosystems). The reference gene PROTEIN PHOSPHATASE 2A SUBUNIT A2 (PP2AA2; At3g25800) or At3g12590 was
selected according to Czechowski et al. [59].

3. Results
3.1. Characterization of aba2 and ga20ox1 single mutants,
ABA2- and GA20ox1-overexpressing transgenic plants, and their
combinations
Because the aba2 and ga20ox1 single mutant and transgenic
plants overexpressing ABA2 and GA20ox1 exhibited a distinct leaf
growth morphology, these lines together with their combinations
(aba2/ga20ox1 and aba2/GA20ox1) were used to study the effects
of the ABA and GA interaction on leaf growth. Plants 21 days
after stratication (DAS) were used in this study to avoid potential inuences of bolting. At this stage, aba2 and ga20ox1 had
smaller rosette radii than the wild type. The decreased leaf size
was more profound in aba2 than that in ga20ox1. Transgenic plants
overexpressing ABA2 exhibited comparable plant stature to that
in the wild type. In contrast, the GA20ox1 overexpressor exhibited
the largest plant size among all tested genotypes (Fig. 1A). These
phenotypes were consistent with previous reports [23,31,49,53].
Moreover, the combinations of aba2 mutant allele with ga20ox1
and GA20ox1 overexpressor, i.e., aba2/ga20ox1 and aba2/GA20ox1,

2.6. Microarray analysis

Plants were grown in soil for 21 days and subsequently the aerial
parts were used for microarray analysis. Total RNA was isolated
from aerial tissues using the TRIzol Plus RNA Purication Kit (Life
Technologies) according to the manufacturers instructions. After
RNA labeling and hybridization, the ATH1 GeneChip (Affymetrix)
was scanned according to the Affymetrix standard protocol. The
resulting CEL les were analyzed by GeneSpring GX11.5 software (Agilent), and the data were processed as described below.
A MAS5.0 normalization was performed. The data were ltered
by expression level using a raw signal value of 100 as the cutoff. The ltered genes were then selected for the unpaired t-test
with a p-value cutoff of 0.05. A two-fold signal change in gene
expression was used as the threshold for aba2, aba2/ga20ox1,
ABA2, and aba2/GA20ox1. For ga20ox1 and GA20ox1, a 1.5-fold signal change was applied as the threshold. Finally, GO enrichment
analysis was performed using a p-value of 0.1 as a cutoff. Hierarchical clustering was used for the analysis of gene expression
data. The method followed an agglomerative approach, where the
most similar expression proles were joined together to form a
group. These groups were further joined in a tree structure until
all data formed a single group. For distance measures, every clustering algorithm needed to measure the similarities (differences)
between entities or conditions. The Euclidean function was used
for distance measurements according to GeneSpring GX instructions. Euclidean distance is equal to the square root of the standard
sum of the squared distance (L2-norm) between two entities, i.e.,


(xi yi )2 .

Fig. 1. Phenotypic comparison and measurement of ABA contents among the genotypes. Plants were grown in soil for 21 days, after which, tissues from the aerial
parts of plants were harvested for phenotypic comparison (A) and ABA analysis (B).
The values in (B) indicate the mean SD of four biological replicates. * P < 0.05; **
P < 0.01, Students t-test.

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

263

all displayed small leaf size compared with that in the wild type
(Fig. 1A).
RT-PCR was performed to conrm the genotypes of these genetic
materials. The aba2 mutant allele produced a transcript with a
53-bp deletion at the beginning of the 2nd exon, resulting in a truncated ABA2 transcript (Fig. S1), as previously described [23]. The
ga20ox1 mutant contained no GA20ox1 transcript, suggesting that
it is a knockout mutant [53]. ABA2 and GA20ox1 overexpressors led
to the accumulation of ABA2 and GA20ox1 transcripts, respectively,
compared to the wild-type levels. The combinations of aba2 mutant
alleles with the ga20ox1 allele or with transgenic plants overexpressing GA20ox1 showed a lack of the corresponding mutant allelic
transcript and the accumulation of the transgene transcript (Fig.
S1). Thus, these data further conrmed the phenotypes and genotypes of these mutants, transgenic plants, and their combinations.
3.2. Measurement of ABA among the genotypes
To further conrm the genetic materials used in this study containing various levels of ABA and its catabolites, plant hormone
assays were performed. The ABA levels in aba2 and its combinations (aba2/ga20ox1 and aba2/GA20ox1) ranged from 16 to 21% of
wild-type levels (Fig. 1B), conrming that aba2 is severely ABA decient [23]. In contrast, the ABA2-overexpressing transgenic plants
contained 40.8% higher levels of ABA than the wild type (Fig. 1B).
These results are consistent with previous reports demonstrating ABA accumulation in transgenic plants overexpressing ABA2
[31]. Interestingly, the ABA level in GA20ox1-overexpressing transgenic plants was 35.9% higher than in the wild type, but ga20ox1
revealed no signicant change (Fig. 1B). For catabolites, ABA inactivation can occur either by oxidation to phaseic acid (PA) and
4-dihydrophaseic acid (DPA) or by covalent conjugation to a glucose residue to form ABA--d-glucosyl ester (ABAGE). In general,
plants with high ABA levels normally produce high PA and DPA as
well, e.g., Col-0, ga20ox1, ABA2, and GA20ox1. However, aba2 and
its combinations (aba2/ga20ox1 and aba2/GA20ox1) exhibited very
low or undetectable levels of PA and DPA (Fig. S2). Thus, high ABA
contents are linked to high catabolic intermediates, but their net
active ABA levels remain high. The level of ABAGE in the genotypes
containing aba2 was undetectable. In ABA2 and GA20ox1, ABAGE
levels were similar to those in the wild type. However, in ga20ox1,
a signicant increase was detected (Fig. S2).
3.3. Comparison of the leaf areas of rosette leaves among ABAand GA-related genotypes
As noted above, the aba2 mutant and ABA2 overexpressor,
and the ga20ox1 mutant and GA20ox1 overexpressor exhibited
different plant and leaf sizes (Fig. 1A). To analyze leaf growth in
these different genotypes, we measured the individual leaf area
of rosette leaves from plants grown in soil for 21 DAS (Fig. 2,
Fig. S3). To unravel the mechanisms by which ABA controls leaf
growth, the aba2 mutant and the transgenic plant overexpressing
ABA2 were used for leaf-size comparisons with the wild type. The
leaf areas of aba2 were signicantly reduced by approximately
52% to 86% at all leaf positions compared with the corresponding
wild-type leaves (Fig. 2A). The ABA2 transgenic plant had similar
leaf areas in the true leaves of the rst to fourth positions. The
size decrease was only observed in leaves of the fth to eighth
positions, which were approximately 70% to 80% of the levels of the
wild type (Fig. 2B). These data suggest that ABA has a dual function
in this biological process; in cases of severe ABA deciency, such
as aba2, leaf area will be much smaller than that in the wild type.
However, ABA-accumulating plants, such as ABA2 overexpressors,
also exhibited inhibited leaf growth. To further investigate the
effect of GA on leaf growth, the ga20ox1 mutant and the transgenic

Fig. 2. Leaf areas of rosette leaves among the genotypes.

Plants were grown in soil for 21 days and then harvested for leaf area measurement.
Values are the mean SD of 8 or 9 plants (n = 8 or 9). At least two independent
experiments were performed and gave consistent results. * P < 0.05, Students t-test;
COT, cotyledon.

plant overexpressing GA20ox1 was examined. Although it has been

reported that ga20ox1 has a similar rosette radius to that in the wild
type [39], a detailed comparison of its leaf growth to that in the wild
type remains elusive. In this study, all individual leaves of ga20ox1
were only 49% to 89% of the leaf areas of the wild type (Fig. 2A);
however, the leaf size was generally larger than that in the aba2
mutant. In contrast, the GA20ox1 overexpressor had approximately
17% to 41% larger leaf areas at positions from the rst to eighth
leaves than wild type (Fig. 2C). These data suggest that leaf growth
is positively regulated by GA. The leaf size of the aba2/ga20ox1
double mutant was slightly smaller but very similar to the aba2
mutant (Fig. 2A). In contrast, the leaf size of aba2/GA20ox1 was

264

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

Fig. 3. The areas of the fth leaf at different developmental stages.

Plants were grown in soil for 17 to 39 DAS, and then, the fth leaves were harvested for the measurement of leaf area at two-day intervals. Five to nine plants were used for
each time point per genotype. Two independent experiments were performed and gave consistent results.

slightly larger but close to that of the aba2 mutant (Fig. 2C). These
data suggest that the ABA-decient aba2 mutant plays a more
important role in the control of leaf growth than the GA-defective
ga20ox1 or the GA-accumulating GA20ox1 overexpressor.
3.4. Effect of developmental stage on leaf growth
The nal leaf size is determined by the combined effects of
leaf growth rate and growth period. To analyze the kinetics of
leaf growth, the leaf area of the fth leaf from each genotype was
evaluated for long-term observation (17 to 39 DAS) because the
fth leaf was well developed in 21-day-old plants of most genotypes (Fig. 2). In general, two growth phases were observed: leaves
rst showed rapid growth at approximately 17 to 27 DAS and
then slowed growth until 39 DAS (Fig. 3). The aba2 mutant grew
much more slowly than the wild type, but the ABA2 overexpressor exhibited similar leaf-growth kinetics compared with that in
the wild type (Fig. 3A). Although ga20ox1 exhibited a smaller leaf
size than the wild type during 17-35 DAS, the mutant exhibited
comparable leaf size to the wild type after 37 DAS (Fig. 3B). These
results suggest a prolonged growth period in ga20ox1. In contrast,
GA20ox1-overexpressing transgenic plants had the largest leaf size
because it also had the fastest growth rate among the genotypes
tested from 17 to 27 DAS. For the ABA and GA interaction, the
aba2/ga20ox1 double mutant had a slightly smaller fth leaf area

than the aba2 mutant, but the pattern of leaf growth rate was similar to that in aba2 from DAS 17 to 33. After 33 DAS, no signicant
difference in leaf area was observed between these two genotypes
(Fig. 3C). Although the leaf area in GA20ox1 was much higher than
in the wild type, the aba2/GA20ox1 plants exhibited a leaf area that
was slightly larger but similar to that in the aba2 mutant (Fig. 3D).
Taken together, these data suggest that the basal level of ABA is
required for normal leaf growth. Under ABA deciency, aba2 largely
dominates GAs function. In addition, leaf growth is largely dependent on GA levels when the ABA level is normal or above normal.
3.5. ABA and GA interaction in the control of leaf growth
The regulation of leaf size includes the coordination of cell division and cell expansion. To characterize the leaf growth mechanism
controlled by ABA and GA or their interaction, the palisade cell size
and number in the fth leaf were analyzed among these genotypes
at 21 DAS (Fig. 4A). The aba2 mutant exhibited severely small leaf
size (Fig. 4A, B), which was consistent with previously data (Fig. 2).
At the cellular level, the palisade cell area and cell number in the
aba2 mutant were dramatically reduced by 56% and 53%, respectively, compared with those in the wild type (Fig. 4B). The leaf size
of the ABA2 overexpressor was reduced by 12% compared with that
in the wild type. This reduction was the result of a slightly smaller
palisade cell area but not the cell number (Fig. 4B). GA modulation

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

265

Fig. 5. Analysis of ploidy among the genotypes.

Plants were grown in soil for 21 days, and then, the fth leaves were used
for polyploidy analysis. Values are the mean SD of four biological replicates.
Three independent experiments were performed and gave consistent results. E.I.,
endoreduplication index.

also affected leaf growth, as the fth leaf area was reduced by 21%
in the ga20ox1 mutant relative to that in the wild type. Similar to
the ABA2 overexpressor, this small leaf area was due to a reduction
in the palisade cell area but not in the cell number. In contrast, the
larger leaf size of the GA20ox1 overexpressor was a consequence
of both increased cell size (17%) and cell number (33%) compared
with those in the wild type (Fig. 4B).
For ABA and GA interaction, although aba2 and ga20ox1 had
smaller leaves than those in the wild type, the leaf growth of the
aba2/ga20ox1 double mutant displayed no signicantly additive
decrease; it was not different from the aba2 mutant in terms of
leaf size, cell area or cell number (Fig. 4B). The fth leaf area of
aba2/GA20ox1 was slightly larger than in aba2. The cell number
in aba2/GA20ox1 was 70% of wild type levels and was closer to
aba2 than to the GA20ox1 overexpressor. However, no signicant
change in cell area was observed between aba2 and aba2/GA20ox1.
This result suggests that the ABA-decient aba2 mutant allele plays
a more important role in leaf growth than the GA-accumulating
GA20ox1 overexpressor.
3.6. Ploidy analysis

Fig. 4. Leaf and palisade cell areas and cell number at the fth leaf.
(A) Phenotype of the fth leaves.
(B) Leaf area (upper panel), palisade cell area (middle panel), and cell number (lower
panel) of the fth leaves. Plants were grown in soil for 21 days and subjected to a phenotype comparison (A) or the measurement of leaf and cell areas or cell number (B).
Values shown in (B) indicate the mean SD of 6 fth leaves, which were derived from
one independent experiment. Three independent experiments were performed and
gave consistent results; each experiment contained 615 plants (n = 615). Fishers
least signicant difference (LSD) was used for statistical analysis. Means with the
same letter indicate no difference between them based on LSD(=0.05) values.

It has been reported that endoreduplication may cause the

enlargement of organ or leaf size [60,61]. Thus, polyploidy among
the genotypes was measured by ow cytometry (Fig. 5, Table
S1). The results indicated that aba2, ga20ox1, aba2/ga20ox1 and
aba2/GA20ox1 plants contained higher frequencies of diploid
(2C) and tetraploid (4C) genomes than the wild type. The
ABA2 and GA20ox1 overexpressors had a similar frequency of
2C + 4C genomes to the wild type. Conversely, the aba2, ga20ox1,
aba2/ga20ox1, and aba2/GA20ox1 mutants showed lower frequencies of 8C and 16C genomes, whereas the frequencies of 8C and 16C
in the ABA2 and GA20ox1 overexpressors were slightly changed
but no difference from the wild type. Consistent results were also
observed in the endoreduplication index (EI), a parameter for the
evaluation of endoreduplication activity. The EI values for aba2,
ga20ox1, aba2/ga20ox1, and aba2/GA20ox1 were close to 1, a value
slightly smaller than that of the wild type (EI = 1.17). This indicates

266

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

Fig. 7. Overlay of differential gene expression between microarray data sets. The
average signal fold change cutoff was 2 or 1.5 in three biological experiments.

Fig. 6. Transcriptome relationship among the genotypes.

Plants grown in soil for 21 DAS were used for microarray analysis. Genes (705)
that passed the ltering criteria, greater than 2-fold changes on average in three
biological experiments, were used for the analysis of the hierarchical clustering.
The values of distance measurement indicate the relationship between two clusters.
Smaller values represent closer relationships.

that the primary activity occurring in these plants is mitosis.

Although the wild type, ABA2, and GA20ox1 had EI values greater
than 1, between 1.15 and 1.21, the differences among these values
were very small. These data suggest that effect of endoreduplication
on cell size changes is very minor among the genotypes tested.
3.7. Expression prole of genes involved in leaf growth
Based on transcript proling, a total of 705 genes passed
the ltering criteria. To determine the transcriptome relationship among the genotypes, hierarchical clustering was performed
(Fig. 6) among these selected genes. The gene expression proling data showed that the aba2/ga20ox1 and aba2/GA20ox1 plants
were closely related to aba2. The aba2/ga20ox1 and aba2 plants had
65.4% (280/428) and 78.4% (280/357) overlapping genes, respectively. Similarly, the aba2/GA20ox1 and aba2 plants exhibited 44.6%
(180/403) and 50.4% (180/357) overlapping genes, respectively
(Fig. 7). A comparison of ga20ox1, GA20ox1, and ABA2 to the wild
type indicates that the differential gene expression proling in
ga20ox1 and GA20ox1 was closer to the wild type than ABA2.
ABA affects the gene expression prole, as aba2 and the ABA2
overexpressor contained 357 and 79 genes, respectively, that were
differentially expressed by at least 2-fold after normalization to

the wild type. Among aba2-regulated genes, biological processes,

such as hormone response, stress response, metabolic process, and
cell wall organization, were overrepresented in their gene ontology
(GO) enrichment analysis (Table S2). The most prominent groups of
hormone-related genes were responsive to ABA, ethylene, jasmonic
acid (JA), and salicylic acid (SA) (Table S3, Fig. S4). For example, the
expression of the JA biosynthetic genes ALLENE OXIDE SYNTHASE
(AOS, AT5G42650) and LIPOXYGENASE 2 (LOX2, AT3G45140) and
the responsive genes JASMONATE-INSENSITIVE 3 (JAI3, AT3G17860)
and EPITHIOSPECIFIER PROTEIN (ESP, AT1G54040) were decreased
in aba2, but the JA-responsive defensin PDF1.2A (AT5G44420)
was increased. Within cell wall-related genes, at least seven
expansins, four cell-wall biogenesis genes, and six xyloglucan endotransglucosylase genes were differentially expressed (Table S4,
Fig. S5). In addition, metabolic processes, such as the oxylipin
metabolic process and JA metabolism (including the AOS and
LOX2 genes) (Table S2) and the stress response (including the
COR413IM1, CORI3, and KIN1 genes) (Table S3, Fig. S4, S5), were also
enriched in this analysis. Although leaf development-associated
categories were not signicantly enriched for aba2, several related
genes passed the ltering criteria (Table 1, Fig. S6). For example,
AGAMOUS-LIKE 42 (AGL42, At5g62165), 12-OXOPHYTODIENOATE
REDUCTASE 1 (OPR1, AT1G76680/At1G76690) and protease I (pfpI)like protein YLS5 (AT2G38860) were upregulated, but CORONATINE
INDUCED 1 (CORI3, At4g23600) and COLD-REGULATED 15A (COR15A,
At2g42540) were down-regulated in the aba2 mutant plants.
These transcript proles suggest that phytohormones, the stress
response, the metabolic process, cell wall organization components, and the leaf development-associated genes may participate
in the leaf-growth reduction caused by ABA deciency. Among
ABA2-regulated genes, 18 genes were classied into the category of
developmental processes. Most genes exhibited decreased expression levels, including the homeobox gene ATH1 (AT4G32980) and
the SART-1 family gene DOT2 (AT5G16780) (Table 1). Nevertheless,
the expression of three leaf-development genes, AGL42, HIGHLY

Table 1
Expression prole of genes involved in leaf developmenta .
AGI

Target description

Fold changeb

Gene ontology/biological process

ga20ox1

aba2/ga20ox1

ABA2

GA20ox1

aba2/GA20ox1

AGL42, AGAMOUS-LIKE 42
OPR1, 12-OXOPHYTODIENOATE REDUCTASE
1c
Protease I (pfpI)-like protein YLS5

2.48
2.05

1.24
1.14

1.13
2.53

2.12
1.05

2.24
1.00

4.74
1.96

Transcription, DNA-dependent
Lipid metabolic process; response to JA or SA

1.71

1.08

2.36

1.27

1.23

1.24

1.59

1.19

1.01

1.10

1.61

2.22

1.56

1.02

1.21

1.20

1.24

2.00

AT1G75040

Actin-binding FH2 (formin homology 2) family

protein
SPL10, SQUAMOSA PROMOTER BINDING
PROTEIN-LIKE 10
PR5, PATHOGENESIS-RELATED GENE 5

IAA biosynthetic process; response to ABA or

ethylene Stimulus
Cell proliferation; actin cytoskeleton
organization
Leaf shaping

1.55

1.09

2.85

1.38

2.09

1.18

AT3G53130/AT3G53140

LUT1, Lutein-Decient 1

1.53

1.23

1.35

1.01

1.20

3.08

AT4G03550
AT2G42200

ATGSL05, a callose synthase

SPL9, SQUAMOSA PROMOTER BINDING
PROTEIN-LIKE 9
MYA1, a member of the type XI myosin protein
family
HAI1, HIGHLY ABA-INDUCED PP2 C GENE 1
ELO1, ELONGATA1

1.45
1.36

1.19
1.29

1.04
1.01

1.28
1.17

1.50
1.11

1.41
2.77

1.30

1.01

1.27

1.78

1.67

1.03

1.28
1.25

1.38
1.42

1.11
1.46

2.30
1.22

1.73
1.66

1.64
1.39

1.19

1.09

1.20

2.04

1.64

1.35

AT4G20910

DOT2, DEFECTIVELY ORGANIZED TRIBUTARIES

2
HEN1, an enhancer of hua1 and hua2

1.17

1.14

1.16

1.30

1.60

1.26

AT2G29890
AT4G32980

VILLIN 1, an actin binding protein

ATH1, HOMEOBOX GENE 1

1.26
1.47

1.02
1.15

2.05
2.06

2.16
2.14

1.24
1.09

1.22
1.81

AT5G54630
AT1G44446

Zinc nger protein-related

CAO, CHLOROPHYLL A OXYGENASE

1.82
1.84

1.30
1.24

2.02
2.33

1.22
1.41

1.14
1.01

1.64
1.70

AT5G44190
AT3G59400

GLK2, Golden2-like 2
GUN4, GENOMES UNCOUPLED 4, GUN4

1.97
2.08

1.38
1.55

1.95
1.89

1.06
1.31

1.48
1.40

2.32
2.35

AT2G42530
AT1G54040

COR15B, COLD REGULATED 15B

ESP, EPITHIOSPECIFYING SENESCENCE
REGULATOR
CORI3; CORONATINE INDUCED 1
COR78, COLD REGULATED 78; RD29A
SVP, SHORT VEGETATIVE PHASE
COR15A, COLD-REGULATED 15A

2.28
3.30

1.03
1.14

1.65
4.16

1.20
1.38

1.09
1.28

1.98
1.43

3.56
4.09
6.30
12.60

1.41
1.06
1.02
1.21

2.40
4.37
3.18
9.21

1.13
1.11
1.00
1.82

1.52
1.45
2.65
1.17

1.73
3.20
5.89
6.17

AT5G62165
AT1G76680/AT1G76690
AT2G38860
AT5G48360
AT1G27370

AT1G17580
AT5G59220
AT3G11220
AT5G16780

AT4G23600
AT5G52310
AT2G22540
AT2G42540

Cell differentiation; leaf morphogenesis;

defense response
Cell differentiation; carotenoid biosynthetic
process
Leaf morphology, leaf senescence
Regulation of leaf formation
Actin lament-based movement; cell division;
cell growth
Abscisic acid-mediated signaling pathway
Cell proliferation; response to ABA or auxin
stimulus
Leaf development
Leaf proximal/distal pattern formation; leaf
morphogenesis
Microtubule cytoskeleton organization
Photomorphogenesis; JA biosynthetic
process; regulation of GA biosynthetic
process
Leaf morphogenesis; cell differentiation
Leaf morphogenesis; chlorophyll biosynthetic
process
chloroplast organization; leaf morphogenesis
Leaf morphogenesis; chlorophyll
biosynthetic process
Response to cold; leaf senescence
Defense response; leaf senescence; response to
JA stimulus
ACC biosynthetic process; cell differentiation
Response to ABA or stress; leaf senescence
Response to JA and SA; leaf morphogenesis
Response to stress or hormones (ABA, SA,
and JA); leaf senescence

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

aba2

Plants were grown in soil for 21 days, and subsequently the aerial parts of the plants were used for the microarray assay.
Fold change represents the signal fold change when normalized to that of the wild type. The values in red or in green indicate upregulation or downregulation over 1.5 fold, respectively.
c
The bold type indicates genes that were veried by qRT-PCR and are listed in Supplementary Fig. S6.
Genes are referred to have signicant change with signal fold change greater than 2.0 for ABA-related mutants or transgenic plants or at least 1.5-fold changes for the ga20ox1 mutant or GA20ox1 transgenic plants on average in
three biological experiments. The raw data are available in the GEO database with accession no. GSE 61344.
b

267

268

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

ABA-INDUCED PP2C GENE 1 (HAI1, AT5G59220), and COR15A, were

slightly upregulated (Table 1). This result suggests that these
development-related genes might, at least in part, confer the slight
reduction in leaf growth observed in the ABA-accumulating ABA2
overexpressor.
GA affected the gene expression prole only slightly, as the list
of differentially expressed genes in ga20ox1 and GA20ox1 includes
only 16 and 26 genes, respectively, when a 2-fold change in gene
expression was used for ltering. No biological processes were indicated by these genes in the GO enrichment analysis. When the
threshold was lowered to a 1.5-fold change, the gene numbers for
ga20ox1 and GA20ox1 increased to 115 and 243 genes, respectively
(Fig. 7). Among the ga20ox1-regulated genes, the GO enrichment
analysis showed that only defense response-related genes were
enriched. However, the differentially expressed genes of GA20ox1
did not belong to any potentially associated category. Several leaf
development-associated genes were differentially expressed in
GA20ox1 (Table 1). Thus, the increased leaf size of GA20ox1 might
also be associated with leaf development-related genes.
For ABA and GA interactions, more genes in the aba2 mutant
are in the categories of hormone response, stress response,
metabolic process, and cell wall organization; aba2/ga20ox1 and
aba2/GA20ox1 largely exhibited similar expression patterns to the
aba2 mutant in these categories (Table 1, Tables S2S4, Figs.
S4S6). For example, PLANT DEFENSIN 1.2 (PDF1.2, AT5G44420),
PYRABACTIN RESISTANCE (PYL4, AT2G38310) and EXPANSIN 1 (EXP1,
AT1G69530) were upregulated and COLD REGULATED 314 INNER
MEMBRANE 1 (COR413IM1, AT1G29395), COR15A (AT2G42540), and
KINASE 1 (KIN1, AT1G14370) were downregulated in these three
genotypes. Although several experiments were conducted in this
study using the fth leaf to characterize leaf growth, the expression
prole was derived from the rosette leaves of 21-day-old plants.
To determine whether the gene expression proles in the fth leaf
and all rosette leaves are comparable, the expression proles of ve
genes (COR15A, COR413IM1, EXPA15, KIN1, and XYL4) were tested
from the fth leaf among the genotypes. The results indicated that
these ve gene expression proles (Fig. S7) were largely comparable to those derived from the rosette leaves (Figs. S4S6).

wild type. The differences of EI values in the wild type and the ABA2
and GA20ox1 overexpressors (Fig. 5, Table S1), which showed larger
leaf size, were very small. Thus, this indicates that endoreduplication does not play a major role in cell size changes among the tested
genotypes.
A basal level of ABA is essential for shoot growth largely due to
ethylene suppression [22]. Further analysis demonstrated that the
ABA-decient mutant aba2 produces a higher ethylene level than
in the wild type [22,64]. Here, our data showed that basal levels
of ABA were required to maintain leaf growth. Thus, the leaf area
of the double mutant aba2/ga20ox1 exhibited a slight reduction
but no signicant difference from that in the aba2 single mutant
(Fig. 2A). Similar results were observed in the palisade cell area
and the cell number of the fth leaf on 21 DAS (Fig. 4). Although
the leaf area of aba2/GA20ox1 was slightly larger than in aba2, it was
closer to the leaf area in aba2 than to that in GA20ox1 (Fig. 3). This
difference is due to a larger cell number but not cell area (Fig. 4).
This result suggests that ABA deciency plays a more important
role than the GA-promoted leaf growth observed in GA20ox1. Thus,
our results showing the effect of the interaction between ABA and
GA indicate that ABA dominates leaf growth when ABA biosynthesis is interrupted. Under ABA-accumulating conditions, ABA2
had a slightly reduced leaf area (Fig. 2). This reduction was associated with reduced cell area but not cell number (Fig. 4). Thus,
ABA has a dual function in the control of leaf growth; in other
words, both ABA-deciency and ABA overproduction may reduce
leaf growth. A similar reduction in leaf growth was observed in the
GA defective mutant ga20ox1. Notably, in addition to GA accumulation [37,38], the GA20ox1 plants accumulated high levels of ABA
(Fig. 1B) but displayed the largest leaf areas, cell areas, and cell numbers (Figs. 2 and 4) among the genotypes tested in this study. These
data further support the interpretation that under normal or ABAaccumulating conditions, GA plays a positive role in the control of
leaf growth. In general, the additive effect of ABA and GA on leaf
growth only partially occurred in aba2/ga20ox1 and aba2/GA20ox1.
It is likely that one pathway (such as ABA deciency) might partially
mask the phenotypes of the others (GA defect or accumulation in
this case). Taken together, these data indicate that both ABA and GA
are essential for but play distinct roles in the control of leaf growth.

4. Discussion
4.1. ABA and GA are essential for but play distinct roles in the
control of leaf growth
Plant leaf size is highly controlled by the integration of environmental stimuli and an intrinsic growth-regulatory network that
spatially and temporally coordinates to mediate cell division and
cell expansion. The intrinsic growth-regulatory network includes
functions of plant hormones. In many physiological processes,
hormone-signaling pathways do not serve as separate entities;
instead, they interact at various levels to ensure an appropriate biological response [62]. Transgenic plants overexpressing ve
genes (GRF5, AVP1, JAW, BRI1, and GA20ox1) belonging to different functional classes show increased leaf area in these lines,
which is dependent on leaf position and growth conditions. Moreover, the enlargement of leaf size is different in these lines but is
largely associated with an increase in leaf cell number [63]. In this
study, leaf size was correlated to leaf position (Fig. 2), leaf-growth
period (Fig. 3), and leaf cell area and cell size (Fig. 4). Leaf size
and shape can be altered when plants are grown under different
growth conditions [2,63]. Thus, our results are largely in agreement
with previous reports. Consistent with the small-leaf-size phenotype, the aba2, ga20ox1, aba2/ga20ox1, and aba2/GA20ox1 plants all
showed an increased frequency of 2C + 4C genomes but a reduced
frequency of 8C + 16C genomes. This result reects that endoreduplication is impaired in these genotypes compared with that in the

4.2. Transcriptome analysis reveals multiple pathways for the

control of leaf growth
To better understand the molecular basis of the ABA and GA
interaction, a transcriptome analysis of the genotypes used in this
study was performed. The total numbers of genes showing differential expression in aba2, aba2/ga20ox1, and aba2/GA20ox1 were
357, 428, and 403, respectively, that had a signal fold change
over 2. However, using the same ltering criteria, the total numbers of altered gene expression were 16 and 26 in ga20ox1 and
GA20ox1, respectively. Thus, GA signicantly induced fewer target genes. Consistent with this observation, it was reported that
a single hormone may specically affect numerous target genes;
however, in the case of GA-induced genes, no specic robust targets were identied. Studies have speculated that interactions
with other hormones may play major roles in GA action [6,65].
Although more differentially expressed genes were observed in
aba2, aba2/ga20ox1, and aba2/GA20ox1, 65.4% (280/428) of genes
in aba2/ga20ox1 and 44.6% (180/403) in aba2/GA20ox1 overlapped
with aba2 (Fig. 7). These overlapping genes largely exhibited similar
expression patterns to those in aba2 (Fig. 6, Table 1, Tables S3 and
S4). These data further support that the small leaf growth observed
in these three genotypes is correlated with their transcriptome similarity. It is noteworthy that in this study, we did not exclude the
possible effect of post-transcriptional or post-translational levels
on the control of leaf growth.

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

It has long been accepted that plant hormones form a complex

network to coordinate the regulation of specic development processes [5,8,9]. Thus, changing one hormone will alter the levels or
signaling of other hormones. In this study, aba2, aba2/ga20ox1, and
aba2/GA20ox1 altered ABA levels and led to enriched GO biological
processes, including hormone response, stress response, cell wall
organization, and metabolic processes (Tables S2S4). The gene
expression which is under the controls of major phytohormones
(ABA, ethylene, JA, or SA) was also altered in these genotypes. In
addition to palisade cell size, these three lines all produced lower
cell numbers in the leaf blade (Fig. 4B). Thus, these data suggest
that plant hormones may affect plant growth through the control of cell cycle progress as reported previously [22,52,6669]
and that the enhancement or reduction of organ growth is largely
governed by different or independent pathways [63]. It has been
well-documented that plant growth is inhibited under stress conditions. Thus, the induction of stress responsive genes by changes
in ABA and GA contents may restrict cell or tissue growth. Cellwall loosening due to expansin activity has been linked to acidic
growth [70]. Microarray data indicate that most of the expansins,
XYLOGLUCAN ENDOTRANSGLUCOSYLASES/HYDROLASES (XTHs),
and pectinacetylesterase were downregulated (Table S4 and Fig.
S5), which might impair cell-wall loosening or expansion. These
data support the importance of the cell wall in the growth adaption of Arabidopsis leaves under phytohormone alteration. The GO
enrichment analysis also indicated that several metabolic pathways were involved in leaf growth, such as oxylipin and avonoid
metabolism. Oxylipins comprise a group of biologically active compounds derived from the metabolism of hydroperoxy fatty acids.
Their study has focused mainly on the biosynthesis of JA and its
role in developmental and defense-related processes [71]. JAs have
been implicated in plant growth and mitosis [68]. Flavonoids are
a group of plant polyphenolic secondary metabolites that includes
anthocyanins. They participate in a wide range of functions, such
as the stress response, male sterility, and auxin transport [72]. It
was reported that GA exerts an effect on Arabidopsis rosette leaf
growth through the coordination of metabolic changes, including
cell wall extension, secondary metabolism, and lipid metabolism
[73]. In agreement with this concept, these data suggest that altered
ABA or GA levels may cause metabolic changes that are involved in
Arabidopsis rosette leaf growth.
Endoreduplication is a process that bypasses mitosis activity and
increases DNA content and is frequently correlated with increased
cell or tissue size. The EI, an index of endoreduplication, represents the average number of endocycles undergone by a given
nucleus. This parameter typically increases during leaf development [68]. Many different types of specialized plant cells, such as
xylem, pavement cells, and epidermal hair cells, have a correlation
between DNA content and cell size in such cells [20,61,74,75]. In
this study, the differences of ploidy levels between two genotypes,
particularly in the wild type vs. ABA2 and wild type vs. GA20ox1
overexpressor, were too small to explain their cell size differences.
This suggests that endoreduplication play a very minor role in cell
size changes among the examined genotypes. Consistently, genes of
the core cell cycle in leaf growth were not observed in this microarray analysis when the gene signal fold change threshold was set at
2. When the threshold was lowered to 1.5, several genes, such as
CYCB2;2 (At4g35620), REGULATORY PARTICLE NON-ATPASE SUBUNIT
5B (RPN5B, At5g64760), CYCB2;4 (At1g76310), E2Fa (At2g36010),
and CKS1 (At2g27960) passed the ltering. One possibility is that
some genes of the core cell cycle might be regulated at the posttranscriptional or post-translational levels, which are not reected
in these microarray data sets. As reported previously [61,76,77],
increased ploidy levels as a result of endoreduplication appear
to not be a unique determinant of plant cell size. Other factors,
such as hormone responsive genes, the stress response, cell wall

269

organization, and metabolic processes, also affect the plant organ

size. In addition, at least twenty-six genes that correlated with
leaf growth (Table 1, Fig. S6) exhibited differential gene expression among the genotypes. These genes were largely involved in
the hormone response, metabolic processes, the stress response,
and cytoskeleton organization. Thus, the ABA and GA interaction
is not a linear pathway but appears to be composed of complex
cross-regulatory networks that synergistically or antagonistically
regulate leaf growth.
Acknowledgements
We thank the Arabidopsis Biological Research Center (ABRC, OH)
for providing T-DNA inserted mutants and the Plant Biotechnology
Institute (PBI) and the National Research Council of Canada for help
with the analyses of ABA metabolites. We are also grateful to the
Affymetrix Gene Expression Service Lab of the Scientic Instrument
Center (Academia Sinica, Taipei, Taiwan) for undertaking the ATH1
GeneChip analysis, the DNA Microarray Core Laboratory (Institute
of Plant and Microbial Biology (IPMB), Academia Sinica, Taipei,
Taiwan) for providing GeneSpring GX11.5 software, and the Flow
Cytometry Core Facility (IPMB) for technical support. This work
was supported by the Academia Sinica (grant no. AS-97-FP-L19-3
to W.-H.C.).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at [Link]
009
References
[1] R.E. Rodriguez, J.M. Debernardi, J.F. Palatnik, Morphogenesis of simple leaves:
regulation of leaf size and shape, WIREs Dev. Biol. 3 (2014) 4157.
[2] H. Tsukaya, Leaf development, Arabidopsis Book 11 (2013) e0163.
[3] P.M. Donnelly, D. Bonetta, H. Tsukaya, R.E. Dengler, N.G. Dengler, Cell cycling
and cell enlargement in developing leaves of Arabidopsis, Dev. Biol. 215 (1999)
407419.
[4] G.T. Beemster, L. De Veylder, S. Vercruysse, G. West, D. Rombaut, P. Van Hummelen, A. Galichet, W. Gruissem, D. Inze, M. Vuylsteke, Genome-wide analysis
of gene expression proles associated with cell cycle transitions in growing
organs of Arabidopsis, Plant Physiol. 138 (2005) 734743.
[5] L. Alvey, N.P. Harberd, DELLA proteins: integrators of multiple plant growth
regulatory inputs? Physiol. Plant. 123 (2005) 153160.
[6] J.L. Nemhauser, F. Hong, J. Chory, Different plant hormones regulate similar
processes through largely nonoverlapping transcriptional responses, Cell 126
(2006) 467475.
[7] R. Sablowski, M. Carnier Dornelas, Interplay between cell growth and cell cycle
in plants, J. Exp. Bot. 65 (2014) 27032714.
[8] T.P. Sun, Gibberellin metabolism, perception and signaling pathways in Arabidopsis, Arabidopsis Book 6 (2008) e0103.
[9] D. Weiss, N. Ori, Mechanisms of cross talk between gibberellin and other hormones, Plant Physiol. 144 (2007) 12401246.
[10] C.M. Karssen, D.L. Brinkhorst-van der Swan, A.E. Breekland, M. Koornneef,
Induction of dormancy during seed development by endogenous abscisic acid:
studies on abscisic acid decient genotypes of Arabidopsis thaliana (L.), Heynh
Planta 157 (1983) 158165.
[11] M. Koornneef, C.J. Hanhart, H.W. Hilhorst, C.M. Karssen, In vivo inhibition of seed
development and reserve protein accumulation in recombinants of abscisic
acid biosynthesis and responsiveness mutants in Arabidopsis thaliana, Plant
Physiol. 90 (1989) 463469.
[12] J. Leung, J. Giraudat, Abscisic acid signal transduction, Annu. Rev. Plant Physiol.
Plant Mol. Biol. 49 (1998) 199222.
[13] X. Qin, J.A. Zeevaart, Overexpression of a 9-cis-epoxycarotenoid dioxygenase
gene in Nicotiana plumbaginifolia increases abscisic acid and phaseic acid levels
and enhances drought tolerance, Plant Physiol. 128 (2002) 544551.
[14] L. Xiong, J.K. Zhu, Regulation of abscisic acid biosynthesis, Plant Physiol. 133
(2003) 2936.
[15] A. Hemerly, A. Engler Jde, C. Bergounioux, M. Van Montagu, G. Engler, D. Inze, P.
Ferreira, Dominant negative mutants of the Cdc2 kinase uncouple cell division
from iterative plant development, EMBO J. 14 (1995) 39253936.
[16] M.M. Castellano, J.C. del Pozo, E. Ramirez-Parra, S. Brown, C. Gutierrez, Expression and stability of Arabidopsis CDC6 are associated with endoreplication, Plant
Cell 13 (2001) 26712686.

270

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

[17] R. Munns, G.R. Cramer, Is coordination of leaf and root growth mediated by
abscisic acid? Opin. Plant Soil 185 (1996) 3349.
[18] R. Finkelstein, Abscisic acid synthesis and response, Arabidopsis Book 11 (2013)
e0166.
[19] M. Koornneef, M.L. Jorna, D.L. Brinkhorst-van der Swan, C.M. Karssen, The isolation of abscisic acid (ABA) decient mutants by selection of induced revertants
in non-germinating gibberellin sensitive lines of Arabidopsis thaliana (L.) Heynh,
Theor. Appl. Genet. 61 (1982) 385393.
[20] G. Horiguchi, U. Fujikura, A. Ferjani, N. Ishikawa, H. Tsukaya, Large-scale histological analysis of leaf mutants using two simple leaf observation methods:
identication of novel genetic pathways governing the size and shape of leaves,
Plant J. 48 (2006) 638644.
[21] J.M. Barrero, P. Piqueras, M. Gonzalez-Guzman, R. Serrano, P.L. Rodriguez, M.R.
Ponce, J.L. Micol, A mutational analysis of the ABA1 gene of Arabidopsis thaliana
highlights the involvement of ABA in vegetative development, J. Exp. Bot. 56
(2005) 20712083.
[22] M.E. LeNoble, W.G. Spollen, R.E. Sharp, Maintenance of shoot growth by endogenous ABA: genetic assessment of the involvement of ethylene suppression, J.
Exp. Bot. 55 (2004) 237245.
[23] W.H. Cheng, A. Endo, L. Zhou, J. Penney, H.C. Chen, A. Arroyo, P. Leon, E. Nambara,
T. Asami, M. Seo, T. Koshiba, J. Sheen, A unique short-chain dehydrogenase/reductase in Arabidopsis glucose signaling and abscisic acid biosynthesis
and functions, Plant Cell 14 (2002) 27232743.
[24] M. Gonzlez-Guzmn, The short-chain alcohol dehydrogenase ABA2 catalyzes the conversion of xanthoxin to abscisic aldehyde, Plant Cell 14 (2002)
18331846.
[25] M. Seo, A.J. Peeters, H. Koiwai, T. Oritani, A. Marion-Poll, J.A. Zeevaart, M. Koornneef, Y. Kamiya, T. Koshiba, The Arabidopsis aldehyde oxidase 3 (AAO3) gene
product catalyzes the nal step in abscisic acid biosynthesis in leaves, Proc.
Natl. Acad. Sci. U.S.A. 97 (2000) 1290812913.
[26] F. Bittner, M. Oreb, R.R. Mendel, ABA3 is a molybdenum cofactor sulfurase
required for activation of aldehyde oxidase and xanthine dehydrogenase in
Arabidopsis thaliana, J. Biol. Chem. 276 (2001) 4038140384.
[27] K.M. Lon-Kloosterziel, M.A. Gil, G.J. Ruijs, S.E. Jacobsen, N.E. Olszewski,
S.H. Schwartz, J.A. Zeevaart, M. Koornneef, Isolation and characterization of
abscisic acid-decient Arabidopsis mutants at two new loci, Plant J. 10 (1996)
655661.
[28] L. Xiong, M. Ishitani, H. Lee, J.K. Zhu, The Arabidopsis LOS5/ABA3 locus encodes a
molybdenum cofactor sulfurase and modulates cold stress- and osmotic stressresponsive gene expression, Plant Cell 13 (2001) 20632083.
[29] S.G. Hwang, N.C. Lin, Y.Y. Hsiao, C.H. Kuo, P.F. Chang, W.L. Deng, M.H. Chiang,
H.L. Shen, C.Y. Chen, W.H. Cheng, The Arabidopsis short-chain dehydrogenase/reductase 3, an abscisic acid decient 2 homolog, is involved in plant
defense responses but not in ABA biosynthesis, Plant Physiol. Biochem. 51
(2012) 6373.
[30] F. Rook, F. Corke, R. Card, G. Munz, C. Smith, M.W. Bevan, Impaired sucroseinduction mutants reveal the modulation of sugar-induced starch biosynthetic
gene expression by abscisic acid signalling, Plant J. 26 (2001) 421433.
[31] P.C. Lin, S.G. Hwang, A. Endo, M. Okamoto, T. Koshiba, W.H. Cheng, Ectopic
expression of ABSCISIC ACID 2/GLUCOSE INSENSITIVE 1 in Arabidopsis promotes
seed dormancy and stress tolerance, Plant Physiol. 143 (2007) 745758.
[32] M. Koornneef, J.H. van der Veen, Induction and analysis of gibberellin sensitive mutants in Arabidopsis thaliana (L.) Heynh, Theor. Appl. Genet. 58 (1980)
257263.
[33] A.L. Silverstone, C. Chang, E. Krol, T.P. Sun, Developmental regulation of the
gibberellin biosynthetic gene GA1 in Arabidopsis thaliana, Plant J. 12 (1997)
919.
[34] P. Achard, A. Gusti, S. Cheminant, M. Alioua, S. Dhondt, F. Coppens, G.T.
Beemster, P. Genschik, Gibberellin signaling controls cell proliferation rate in
Arabidopsis, Curr. Biol. 19 (2009) 11881193.
[35] P. Hedden, A.L. Phillips, Gibberellin metabolism: new insights revealed by the
genes, Trends Plant Sci. 5 (2000) 523530.
[36] J. Grifths, K. Murase, I. Rieu, R. Zentella, Z.L. Zhang, S.J. Powers, F. Gong, A.L.
Phillips, P. Hedden, T.P. Sun, S.G. Thomas, Genetic characterization and functional analysis of the GID1 gibberellin receptors in Arabidopsis, Plant Cell 18
(2006) 33993414.
[37] J.P. Coles, A.L. Phillips, S.J. Croker, R. Garcia-Lepe, M.J. Lewis, P. Hedden, Modication of gibberellin production and plant development in Arabidopsis by
sense and antisense expression of gibberellin 20-oxidase genes, Plant J. 17 (1999)
547556.
[38] S. Huang, A.S. Raman, J.E. Ream, H. Fujiwara, R.E. Cerny, S.M. Brown, Overexpression of 20-oxidase confers a gibberellin-overproduction phenotype in
Arabidopsis, Plant Physiol. 118 (1998) 773781.
[39] I. Rieu, O. Ruiz-Rivero, N. Fernandez-Garcia, J. Grifths, S.J. Powers, F. Gong, T.
Linhartova, S. Eriksson, O. Nilsson, S.G. Thomas, A.L. Phillips, P. Hedden, The
gibberellin biosynthetic genes AtGA20ox1 and AtGA20ox2 act, partially redundantly, to promote growth and development throughout the Arabidopsis life
cycle, Plant J. 53 (2008) 488504.
[40] R. Finkelstein, W. Reeves, T. Ariizumi, C. Steber, Molecular aspects of seed dormancy, Annu. Rev. Plant Biol. 59 (2008) 387415.
[41] J.C. Rogers, S.W. Rogers, Denition and functional implications of gibberellin
and abscisic acid cis-acting hormone response complexes, Plant Cell 4 (1992)
14431451.
[42] F. Gubler, R. Kalla, J.K. Roberts, J.V. Jacobsen, Gibberellin-regulated expression
of a myb gene in barley aleurone cells: evidence for Myb transactivation of a
high-pI alpha-amylase gene promoter, Plant Cell 7 (1995) 18791891.

[43] A. Gomez-Cadenas, S.D. Verhey, L.D. Holappa, Q. Shen, T.H. Ho, M.K. WalkerSimmons, An abscisic acid-induced protein kinase, PKABA1, mediates abscisic
acid-suppressed gene expression in barley aleurone layers, Proc. Natl. Acad. Sci.
U.S.A. 96 (1999) 17671772.
[44] A. Gomez-Cadenas, R. Zentella, M.K. Walker-Simmons, T.H. Ho, Gibberellin/abscisic acid antagonism in barley aleurone cells: site of action of the
protein kinase PKABA1 in relation to gibberellin signaling molecules, Plant Cell
13 (2001) 667679.
[45] S.J. Davis, Integrating hormones into the oral-transition pathway of Arabidopsis thaliana, Plant Cell Environ. 32 (2009) 12011210.
[46] J. Langridge, Effect of day-length and gibberellic acid on the owering of Arabidopsis, Nature 180 (1957) 3637.
[47] R.N. Wilson, J.W. Heckman, C.R. Somerville, Gibberellin is required for owering in Arabidopsis thaliana under short days, Plant Physiol. 100 (1992)
403408.
[48] R.N. Wilson, C.R. Somerville, Phenotypic suppression of the gibberellininsensitive mutant (gai) of Arabidopsis, Plant Physiol. 108 (1995) 495502.
[49] M.A. Domagalska, E. Sarnowska, F. Nagy, S.J. Davis, Genetic analyses of interactions among gibberellin, abscisic acid, and brassinosteroids in the control of
owering time in Arabidopsis thaliana, PLoS ONE 5 (2010) e14012.
[50] S. Iuchi, M. Kobayashi, T. Taji, M. Naramoto, M. Seki, T. Kato, S. Tabata,
Y. Kakubari, K. Yamaguchi-Shinozaki, K. Shinozaki, Regulation of drought
tolerance by gene manipulation of 9-cis-epoxycarotenoid dioxygenase, a
key enzyme in abscisic acid biosynthesis in Arabidopsis, Plant J. 27 (2001)
325333.
[51] A. Skirycz, S. De Bodt, T. Obata, I. De Clercq, H. Claeys, R. De Rycke, M. Andriankaja, O. Van Aken, F. Van Breusegem, A.R. Fernie, D. Inze, Developmental
stage specicity and the role of mitochondrial metabolism in the response of
Arabidopsis leaves to prolonged mild osmotic stress, Plant Physiol. 152 (2010)
226244.
[52] W. Verelst, A. Skirycz, D. Inze, Abscisic acid, ethylene and gibberellic acid act
at different developmental stages to instruct the adaptation of young leaves to
stress, Plant Signal. Behav. 5 (2010) 473475.
[53] T. Hisamatsu, R.W. King, C.A. Helliwell, M. Koshioka, The involvement of
gibberellin 20-oxidase genes in phytochrome-regulated petiole elongation of
Arabidopsis, Plant Physiol. 138 (2005) 11061116.
[54] J.N. Maloof, K. Nozue, M.R. Mumbach, C.M. Palmer, LeafJ: an ImageJ plugin for
semi-automated leaf shape measurement, J. Vis. Exp. 71 (2013) e50028.
[55] T. Tsuge, H. Tsukaya, H. Uchimiya, Two independent and polarized processes of
cell elongation regulate leaf blade expansion in Arabidopsis thaliana (L.) Heynh,
Development 122 (1996) 15891600.
[56] J. Dolezel, J. Greilhuber, J. Suda, Estimation of nuclear DNA content in plants
using ow cytometry, Nat. Protoc. 2 (2007) 22332244.
[57] S.D. Chiwocha, S.R. Abrams, S.J. Ambrose, A.J. Cutler, M. Loewen, A.R. Ross,
A.R. Kermode, A method for proling classes of plant hormones and their
metabolites using liquid chromatographyelectrospray ionization tandem
mass spectrometry: an analysis of hormone regulation of thermodormancy
of lettuce (Lactuca sativa L.) seeds, Plant J. 35 (2003) 405417.
[58] M.K. Udvardi, T. Czechowski, W.R. Scheible, Eleven golden rules of quantitative
RT-PCR, Plant Cell 20 (2008) 17361737.
[59] T. Czechowski, M. Stitt, T. Altmann, M.K. Udvardi, W.R. Scheible, Genome-wide
identication and testing of superior reference genes for transcript normalization in Arabidopsis, Plant Physiol. 139 (2005) 517.
[60] H. Breuninger, M. Lenhard, Control of tissue and organ growth in plants, Curr.
Top. Dev. Biol. 91 (2010) 185220.
[61] K. Sugimoto-Shirasu, K. Roberts, Big it up: endoreduplication and cell-size
control in plants, Curr. Opin. Plant Biol. 6 (2003) 544553.
[62] S. Gazzarrini, P. McCourt, Genetic interactions between ABA, ethylene and sugar
signaling pathways, Curr, Opin. Plant Biol. 4 (2001) 387391.
[63] N. Gonzalez, S. De Bodt, R. Sulpice, Y. Jikumaru, E. Chae, S. Dhondt, T. Van Daele,
L. De Milde, D. Weigel, Y. Kamiya, M. Stitt, G.T.S. Beemster, D. Inze, Increased
leaf size: different means to an end, Plant Physiol. 153 (2010) 12611279.
[64] W.H. Cheng, M.H. Chiang, S.G. Hwang, P.C. Lin, Antagonism between abscisic
acid and ethylene in Arabidopsis acts in parallel with the reciprocal regulation
of their metabolism and signaling pathways, Plant Mol. Biol. 71 (2009) 6180.
[65] R. Zentella, Z.L. Zhang, M. Park, S.G. Thomas, A. Endo, K. Murase, C.M. Fleet,
Y. Jikumaru, E. Nambara, Y. Kamiya, T.P. Sun, Global analysis of DELLA direct
targets in early gibberellin signaling in Arabidopsis, Plant Cell 19 (2007)
30373057.
[66] W.M. Gray, I. Estelle, Function of the ubiquitin-proteasome pathway in auxin
response, Trends Biochem. Sci. 25 (2000) 133138.
[67] C. Gutierrez, The Arabidopsis cell division cycle, Arabidopsis Book 7 (2009)
e0120.
[68] S. Noir, M. Bomer, N. Takahashi, T. Ishida, T.L. Tsui, V. Balbi, H. Shanahan, K.
Sugimoto, A. Devoto, Jasmonate controls leaf growth by repressing cell proliferation and the onset of endoreduplication while maintaining a potential
stand-by mode, Plant Physiol. 161 (2013) 19301951.
[69] A. Skirycz, D. Inze, More from less: plant growth under limited water, Curr.
Opin. Biotechnol. 21 (2010) 197203.
[70] D.J. Cosgrove, Loosening of plant cell walls by expansins, Nature 407 (2000)
321326.
[71] G.A. Howe, A.L. Schilmiller, Oxylipin metabolism in response to stress, Curr.
Opin. Plant Biol. 5 (2002) 230236.
[72] E. Petrussa, E. Braidot, M. Zancani, C. Peresson, A. Bertolini, S. Patui, A. Vianello,
Plant avonoidsbiosynthesis, transport and involvement in stress responses,
Int. J. Mol. Sci. 14 (2013) 1495014973.

M.-H. Chiang et al. / Plant Science 236 (2015) 260271

[73] D.M. Ribeiro, W.L. Araujo, A.R. Fernie, J.H. Schippers, B. Mueller-Roeber, Translatome and metabolome effects triggered by gibberellins during rosette growth
in Arabidopsis, J. Exp. Bot. 63 (2012) 27692786.
[74] J.E. Melaragno, B. Mehrotra, A.W. Coleman, Relationship between endopolyploidy and cell size in epidermal tissue of Arabidopsis, Plant Cell 5 (1993)
16611668.
[75] J. Traas, M. Hulskamp, E. Gendreau, H. Hofte, Endoreduplication and
development: rule without dividing? Curr. Opin. Plant Biol. 1 (1998)
498503.

271

[76] C. Massonnet, D. Vile, J. Fabre, M.A. Hannah, C. Caldana, J. Lisec, G.T. Beemster,
R.C. Meyer, G. Messerli, J.T. Gronlund, J. Perkovic, E. Wigmore, S. May, M.W.
Bevan, C. Meyer, S. Rubio-Diaz, D. Weigel, J.L. Micol, V. Buchanan-Wollaston,
F. Fiorani, S. Walsh, B. Rinn, W. Gruissem, P. Hilson, L. Hennig, L. Willmitzer, C.
Granier, Probing the reproducibility of leaf growth and molecular phenotypes:
a comparison of three Arabidopsis accessions cultivated in ten laboratories,
Plant Physiol. 152 (2010) 21422157.
[77] H. Tsukaya, Does ploidy level directly control cell sizes? Couterevidence from
Arabidopsis genetics, PLoS ONE 8 (2013) e83729.