The GID1-Mediated Gibberellin Perception Mechanism IsConserved in the Lycophyte Selaginella moellendorffii butNot in the Bryophyte Physcomitrella patens W
Ko Hirano,a Masatoshi Nakajima,b Kenji Asano,a Tomoaki Nishiyama,c,d Hitoshi Sakakibara,e Mikiko Kojima,e
Etsuko Katoh,f Hongyu Xiang,f Takako Tanahashi,g,h Mitsuyasu Hasebe,d,g,h Jo Ann Banks,i Motoyuki Ashikari,a
Hidemi Kitano,a Miyako Ueguchi-Tanaka,a and Makoto Matsuokaa,1
a Bioscience and Biotechnology Center, Nagoya University, Nagoya 464-8601, Japanb Department of Applied Biological Chemistry, University of Tokyo, Tokyo 113-8657, Japanc Advanced Science Research Center, Kanazawa University, Takaramachi 13-1, Kanazawa, Ishikawa 920-0934, Japand ERATO, Japan Science and Technology Agency, Takaramachi 13-1, Kanazawa, Ishikawa 920-0934, Japane Plant Science Center, RIKEN, Institute of Physical and Chemical Research, Tsurumi, Yokohama 230-0045, Japanf Division of Plant Sciences, National Institute of Agrobiological Sciences, Tsukuba, Ibaraki 305-8602, Japang National Institute for Basic Biology, Okazaki 444-8585, Japanh Department of Basic Biology, Graduate School of Life Science, Okazaki 444-8585, Japani Department of Botany and Plant Pathology, Purdue University, West Lafayette, Indiana 47907-1153
In rice (Oryza sativa) and Arabidopsis thaliana, gibberellin (GA) signaling is mediated by GIBBERELLIN-INSENSITIVE DWARF1
(GID1) and DELLA proteins in collaboration with a GA-specific F-box protein. To explore when plants evolved the ability to
perceive GA by the GID1/DELLA pathway, we examined these GA signaling components in the lycophyte Selaginella
moellendorffii and the bryophyte Physcomitrella patens. An in silico search identified several homologs of GID1, DELLA, and
GID2, a GA-specific F-box protein in rice, in both species. Sm GID1a and Sm GID1b, GID1 proteins from S. moellendorffii,
showed GA binding activity in vitro and interacted with DELLA proteins from S. moellendorffii in a GA-dependent manner in
yeast. Introduction of constitutively expressed Sm GID1a, Sm G1D1b, and Sm GID2a transgenes rescued the dwarf phenotype
of rice gid1 and gid2 mutants. Furthermore, treatment with GA4, a major GA in S. moellendorffii, caused downregulation of Sm
GID1b, Sm GA20 oxidase, and Sm GA3 oxidase and degradation of the Sm DELLA1 protein. These results demonstrate that the
homologs of GID1, DELLA, and GID2 work in a similar manner in S. moellendorffii and in flowering plants. Biochemical studies
revealed that Sm GID1s have different GA binding properties from GID1s in flowering plants. No evidence was found for the
functional conservation of these genes in P. patens, indicating that GID1/DELLA-mediated GA signaling, if present, differs from
that in vascular plants. Our results suggest that GID1/DELLA-mediated GA signaling appeared after the divergence of vascular
plants from the moss lineage.
INTRODUCTION
Gibberellins (GAs) comprise a large family of tetracyclic, diterpe-
noid plant hormones that play diverse biological roles in plant
growth, including seed germination, stem elongation, leaf expan-
sion, pollen maturation, and induction of flowering (Olszewski
et al., 2002). Through molecular genetic studies on GA-insensitive
mutants of rice (Oryza sativa) and Arabidopsis thaliana, the
mechanism underlying GA perception has been revealed (re-
viewed in Ueguchi-Tanaka et al., 2007a). A well-characterized
factor involved in the GA signaling pathway is the DELLA protein,
which belongs to the GRAS superfamily of putative transcription
factors. The DELLA protein functions as a negative regulator of
GA signaling (Peng et al., 1997; Ikeda et al., 2001; Chandler et al.,
2002) and is rapidly degraded when plants are treated with GA
(Dill et al., 2001; Silverstone et al., 2001; Gubler et al., 2002; Itoh
et al., 2002). Recent studies have identified rice GIBBERELLIN-
INSENSITIVE DWARF2 (GID2) and Arabidopsis SLEEPY1 (SLY1),
candidate F-box components of Skp1-Cullin-F box protein (SCF)
E3 ubiquitin ligases, as apparently responsible for targeting
DELLA proteins to the proteasome (McGinnis et al., 2003; Sasaki
et al., 2003). More recently, GIBBERELLIN-INSENSITIVE DWARF1
(GID1) was identified as a soluble GA receptor in rice and
Arabidopsis (Ueguchi-Tanaka et al., 2005; Griffiths et al., 2006;
Nakajima et al., 2006). Based on these observations, the current
model of GA signaling is as follows. In the absence of GA, the
DELLA protein represses GA action. In the presence of GA, the
GID1 receptor binds GA. The GID1/GA complex then interacts
with the DELLA protein. This interaction results in DELLA protein
degradation through the SCFGID2/SLY1 proteasome pathway and,
1 Address correspondence to [email protected] authors responsible for distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors (www.plantcell.org) are: Miyako Ueguchi-Tanaka ([email protected]) and Makoto Matsuoka ([email protected]).W Online version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.107.051524
The Plant Cell, Vol. 19: 3058–3079, October 2007, www.plantcell.org ª 2007 American Society of Plant Biologists
consequently, allows GA action to occur (Ueguchi-Tanaka et al.,
2007a) (Figure 1).
When the GID1/DELLA-mediated GA perception system was
established is an interesting question in the evolution of plant
growth and development. Since the late 1950s, the presence of
GAs has been reported in both seed plants and nonseed plants,
including unicellular and multicellular algae, mosses, and ferns
(Radley, 1961; Kato et al., 1962; Ergun et al., 2002). Although the
presence of GAs in nonseed plants suggested that these species
might use GA as a bioactive substance, unequivocal GA iden-
tification was not possible due to the lack of reliable chemical
methods. To clarify this issue, MacMillan (2002) proposed three
criteria for GA identification: (1) isolation of the pure product and
comparison of its physical properties with an authentic speci-
men; (2) gas chromatography–mass spectrometry (GC-MS)
comparing Kovats retention indices and m/z (and relative inten-
sities) of at least six significant ions with those of standards; and
(3) gas chromatography–single ion monitoring comparing reten-
tion times and m/z (and relative intensities) of at least six
significant ions with those of standards (http://www.plant-
hormones.info/occurrence_of_gas_in_plants.htm). When these
criteria are applied to GA identification of GAs in nonseed plants,
only a few GAs from fern are identified (Yamane et al., 1985,
1988; Yamauchi et al., 1996). Consequently, with the exception
of antheridiogen, a pheromone promoting antheridium formation
in tree ferns, it is still unclear whether nonseed plants use GAs as
bioactive substances (Yamauchi et al., 1996; Banks, 1999;
Menendez et al., 2006).
Recent progress in genome research provides an alternative
approach to determine when the GA perception system evolved.
When genes encoding proteins homologous with the GA-related
proteins in seed plants are found in the genomes of nonseed
plants, these genes can be cloned and characterized functionally
using biochemical and genetic techniques. Two model plants,
the lycophyte Selaginella moellendorffii and the moss Physco-
mitrella patens, are well suited for in silico searches for GA-
related genes because their whole-genome shotgun sequences
and EST sequences are available to the public (http://selaginella.
genomics.purdue.edu/ and http://moss.nibb.ac.jp/).
These plants are also suitable model plants for evolutionary
studies. The lycophyte S. moellendorffii has a small genome
(;100 Mb) (Weng et al., 2005), about two-thirds the size of the
Arabidopsis genome. The lycophytes form a basal group within
the vascular plants and diverged from other vascular plants
;400 million years ago (Weng et al., 2005). Leaves of the
lycophytes and other vascular plants evolved in parallel (Gifford
and Foster, 1989), so the study of S. moellendorffii will provide
insights into vascular plant evolution, particularly the develop-
ment of the plant body. The moss P. patens has been used as an
experimental organism because it exhibits a high efficiency of
homologous recombination (Schaefer and Zryd, 1997). It is
believed that mosses and vascular plants diverged ;430 million
years ago (Kenrick and Crane, 1997), and, despite large differ-
ences in morphology and life cycle between flowering plants and
mosses, P. patens contains some signal transduction systems
similar to those found in flowering plants. For example, auxins
and cytokinins are important developmental regulators for both
flowering plants and mosses (Cove et al., 2006), and the desic-
cation stress response network mediated by abscisic acid is
probably conserved in both (Knight et al., 1995). However, the
GA signaling pathway has not been studied in these model
nonseed plants. Very recently, Hayashi et al. (2006) reported the
presence of ent-kaurene synthase, an enzyme involved in GA
biosynthesis in seed plants, in P. patens. However, it remains to
be determined whether P. patens contains GAs and uses them as
growth substances.
We performed in silico screening of GA-related genes in the
genomes of S. moellendorffii and P. patens using rice genes for
the GA receptor GID1, DELLA protein genes, SLR1, and the
F-box gene GID2 as queries. Several candidate sequences were
found in both genomes, so we examined the GID1 homologs in
an in vitro GA binding experiment and studied the interaction
between GID1 and the DELLA homologs in a yeast two-hybrid
assay. We also performed complementation experiments by
introducing the candidate genes into rice plants. Our results
demonstrate that S. moellendorffii has functional homologs of
GID1, DELLA protein, and GID2, whereas P. patens does not.
RESULTS
Isolation of Putative GA Signaling Genes in S. moellendorffii
and P. patens
The amino acid sequences of rice GID1, SLR1, and GID2 were
used as queries to screen the available S. moellendorffii and
P. patens databases (http://selaginella.genomics.purdue.edu/
cgi-bin/blast_tmpl_s.cgi and http://moss.nibb.ac.jp/). Using a
TBLASTN search, candidate genes were selected for preliminary
phylogenetic analyses. In the S. moellendorffii genome data-
base, two GID1-like sequences (Sm GID1a and Sm GID1b), two
Figure 1. Current Model of GID1-Mediated GA Signaling in Flowering
Plants.
In the presence of GA, GID1 binds to DELLA, a negative regulator of GA
action, to form the GA-GID1-DELLA complex. Subsequently, DELLA
protein is degraded through the SCFGID2/SLY1 proteasome pathway, and
as a consequence, GA actions occur.
GA Perception in Nonseed Plants 3059
DELLA-like sequences (Sm DELLA1 and Sm DELLA2), and three
GID2-like sequences (Sm GID2a, Sm GID2b, and Sm GID2c)
were found. In the P. patens genome database, two GID1-like
sequences (Pp GID1L1 and Pp GID1L2), two DELLA-like se-
quences (Pp DELLAL1 and Pp DELLAL2), and three GID2-like
sequences (Pp GID2L1, Pp GID2L2, and Pp GID2L3) were found.
GID1-Like Genes
In the Sm GID1a and Sm GID1b genes, a single intron was
predicted at the same site in the N-terminal portion of the
predicted protein sequences. This predicted intron position is
the same as that found in rice (Os GID1) and three Arabidopsis
GID1 genes (At GID1a, -b, and -c) (Figure 2A, open triangle at the
top of the first block). On the other hand, Pp GID1L1 and Pp
GID1L2 contained two introns at the same site in each gene,
localized at the middle and C terminus of the predicted protein
sequences (Figure 2A, closed triangles at the bottom of the third
and fifth blocks). These sites do not correspond to the intron
position of GID1 genes in the flowering plants analyzed. The
locations and lengths of these predicted introns were confirmed
by sequencing the corresponding cDNA for each gene.
We compared the amino acid sequences of the Sm GID1 and
Pp GID1L proteins with those of Os GID1 and At GID1a, -b, and
-c. We also included the amino acid sequences of hormone-
sensitive lipases (HSLs) from Arabidopsis and rice (Figure 2A) in
the comparison, because GID1 shares homology with the con-
sensus sequence of the HSL family (Ueguchi-Tanaka et al.,
2005). The Sm GID1 and Pp GID1L proteins contain the con-
served HSL motifs HGG and GXSXG found in other GID1
proteins (Figure 2A, small closed circles at the bottom of the
second and third blocks). Of the three conserved amino acids (S,
D, and H) that form the catalytic site in the HSL family (Figure 2A,
large closed circles at the top of the third, fourth, and fifth blocks),
S and D are found at the corresponding positions in the GID1
proteins, but H is replaced with V or I in the rice and Arabidopsis
GID1 proteins (Ueguchi-Tanaka et al., 2005; Nakajima et al.,
2006). The predicted amino acid sequences for the Sm GID1 and
Pp GID1L proteins also contain S and D in the same relative posi-
tions. At the third amino acid residue, however, the Sm GID1
proteins contained V, as in other GID1 proteins, while Pp GID1L
proteins contained H, as in HSLs. A rice GID1 loss-of-function
allele, gid1-2, contains one amino acid substitution (R251T),
indicating that R251 is important for GA receptor function
(Ueguchi-Tanaka et al., 2005). This R residue is conserved in
Sm GID1a and Sm GID1b, but not in Pp GID1L1 or Pp GID1L2, in
which R is replaced by L (Figure 2A, closed diamond at the top of
the fourth block). Another rice loss-of-function allele, gid1-5,
contains E instead of G at position 169 (Ueguchi-Tanaka et al.,
2007b), whereas Sm GID1b contains C and Pp GID1L1 and -L2
contain S at that position (Figure 2A, closed diamond at the top of
the third block).
A phylogenetic analysis illustrated the relationships among the
Sm GID1, Pp GID1L, and GID1 proteins from seed plants (Figure
2B). Sm GID1a and -b form a monophyletic clade that is sister to
the clade of GID1 proteins from seed plants. The clade contain-
ing the S. moellendorffii and seed plant GID1 proteins is sup-
ported by a high bootstrap value and is distinct from other
proteins. This clade combined with the Pp GID1L proteins forms
another clade with weak bootstrap support. A phylogenetic
species relationship among moss, lycophyte, pine (Pinus sp),
Arabidopsis, and rice is shown in Figure 2C as a reference
(Nishiyama, 2007).
DELLA-Like Genes
DELLA proteins can be divided into two domains based on
their primary structure. DELLA/TVHYNP domains, found in the
N-terminal portion of the proteins, are unique to the DELLA
protein family and are involved in the interaction with GID1s in a
GA-dependent manner (Griffiths et al., 2006; Ueguchi-Tanaka
et al., 2007b; Willige et al., 2007). The second domain, shared
among GRAS proteins, is located in the C-terminal portion (Bolle,
2004; Itoh et al., 2005). These DELLA/TVHYNP and GRAS
domains are interrupted by highly variable sequences of various
lengths (Itoh et al., 2002). We compared the structures of the
DELLA/TVHYNP and GRAS domains separately. Comparison of
the DELLA/TVHYNP domains revealed that Sm DELLA1 contains
a sequence that is very similar to those of DELLA proteins of
Arabidopsis and rice, whereas Sm DELLA2 was less similar to
flowering plant DELLA proteins (Figure 3A). Although we could
align the amino acid sequences at the N-terminal portion of Pp
DELLAL proteins with the DELLA/TVHYNP sequences of flower-
ing plant DELLA proteins, the corresponding sequences of Pp
DELLAL1 and -L2 only partially shared the conserved motifs of
DELLA or TVHYNP (Figure 3A).
We also compared the GRAS domains of Sm DELLA and Pp
DELLAL proteins with those of Arabidopsis and rice DELLA
proteins and also with those of Os SCR, a protein in another
GRAS subfamily (Kamiya et al., 2003). The Sm DELLA and Pp
DELLAL proteins share GRAS domains, such as the leucine
heptad repeat I (LHR 1), VHIID, LHR 2, PFYRE, and SAW
domains, with high levels of similarity (Figure 3B). Recently,
Itoh et al. (2005) reported unique DELLA proteins, Os SLRL1 and
Os SLRL2, which contain regions shared with GRAS domains of
typical DELLA proteins but lack the N-terminal DELLA/TVHYNP
domains. These DELLA-like proteins lacking DELLA/TVHNYP
domains can suppress GA signaling in a GA-independent man-
ner in rice, indicating that these proteins may function as con-
stitutive suppressors in GA signaling. Phylogenetic analysis
demonstrated that Sm DELLA1, Pp DELLAL1, and Pp DELLAL2
form a clade with the DELLA proteins of flowering plants (Figure
3C). Together, these comparative analyses suggest that Sm
DELLA1 might be involved in GA signaling in a manner similar to
that of other typical DELLA proteins, since this protein contains
both DELLA/TVHYNP and GRAS domains with high similarity to
the flowering plant DELLA proteins analyzed. This hypothesis
was tested using a number of biochemical and genetic ap-
proaches, as described below.
GID2-Like Genes
In a comparative study of Arabidopsis At SLY1 and its homologs,
including Os GID2, McGinnis et al. (2003) proposed that the
At SLY1/Os GID2 proteins contain three conserved domains:
F-box, GGF, and LSL. Gomi et al. (2004) demonstrated that these
3060 The Plant Cell
Figure 2. Comparison of Deduced Amino Acid Sequences of GID1 Homologs in Land Plants.
GA Perception in Nonseed Plants 3061
three conserved domains are essential for the function of Os
GID2. Noting these important domains, we aligned the predicted
amino acid sequences of the Sm GID2 and Pp GID2L proteins
with those of At SLY1 and Os GID2 (Figure 4A). At SNZ, a
homolog of At SLY1 that also functions in GA signaling in
Arabidopsis (Strader et al., 2004), was included in the analysis.
In the phylogenetic analysis, Sm GID2a and Sm GID2b formed a
clade with Os GID2, At SLY1, At SNZ, Ta GID2L (a Triticum
aestivum GID2 homolog), and another rice gene with 91%
bootstrap support (Figure 4B). Sm GID2c and Pp GID2L proteins
were placed outside this clade. Although Sm GID2c and Pp
GID2L proteins contain an F-box similar to that of At SLY1 and Os
GID2, their sequences corresponding to the GGF and LSL
regions showed low similarity to those in At SLY1 and Os GID2
(Figure 4A). These comparative studies suggest that Sm GID2a
and Sm GID2b might function as F-box proteins in GA signaling,
whereas the other GID2-like proteins might not.
GA Binding Properties of Sm GID1 and Pp GID1L Proteins
To examine the function of the Sm GID1 and Pp GID1L proteins,
we studied the interaction between recombinant thioredoxin�histidine (Trx�His)-tagged GID1-like proteins and radioactive GA
using a nonequilibrium gel permeation technique (Ueguchi-
Tanaka et al., 2005). Because the presence of DELLA protein
stabilizes the interaction between Os GID1 and GA (Ueguchi-
Tanaka et al., 2007b), we performed the experiment in the pres-
ence and absence of DELLA proteins. Trx�His-Sm GID1a bound
to [1,2,16,17-3H4]16,17-dihydro-GA4 (3H4-H2-GA4) in the pres-
ence of glutathione S-transferase (GST)–tagged Sm DELLA1 and
Sm DELLA2, although the binding activity of Sm GID1a–Sm
DELLA1 was much higher than that of Sm GID1a–Sm DELLA2
(Figure 5A). No binding activity was seen in the absence of Sm
DELLA protein. We barely detected any GA binding activity of Sm
GID1b or Pp GID1L proteins, regardless of whether Sm DELLA1
was included in the reaction. However, since a yeast two-hybrid
experiment demonstrated that Sm GID1b interacted with GA4
with ;6-fold lower affinity than Sm GID1a (see below), we
reexamined the interaction between Sm GID1b and GA4 under a
10-fold higher level of 3H4-H2-GA4. As expected, Sm GID1b
showed low but detectable GA binding activity under the high3H4-H2-GA4 conditions, while no or very low activity of Pp GID1L
proteins was observed (Figure 5B). In each experiment, approx-
imately the same amount of Escherichia coli–produced Trx�His-
GID1 protein was used (Figure 5C). These in vitro GA binding
analyses demonstrate that Sm GID1a and Sm GID1b have GA4
binding activity only in the presence of Sm DELLA1. GA4 has a much
higher affinity for Sm GID1a than for Sm GID1b, while Pp GID1L
proteins have no or very low 3H4-H2-GA4 or GA4 binding activity.
To further elucidate whether Pp GID1L1 and -L2 can bind to
other GAs, we performed a thermodynamic analysis using iso-
thermal titration calorimetry (ITC). ITC is the most direct method
to measure the change of heat upon the formation of a complex.
Figure 5D shows the heat exchange versus molar ratio for
Trx�His-Os GID1 (40 mM) with titration of GA (final content of
500 mM) at 308C. A clear thermodynamic change was observed
in the Trx�His-Os GID1-GA4 titration experiments: the number of
binding sites per protein molecule was approximately one (n¼ 1),
and base dissociation constant was 5.16 3 106 M (Figure 5D). On
the other hand, Trx�His-Pp GID1L1 (Figure 5E) and Trx�His-Pp
GID1L2 (Figure 5F) did not show any heat change with any of the
GAs tested (GA9, GA12, and 3-epi-GA4).
Properties of the GID1–DELLA Interaction in Yeast Cells
GID1 proteins in rice and Arabidopsis can interact with DELLA
proteins in a GA-dependent manner in yeast cells (Ueguchi-
Tanaka et al., 2005; Griffiths et al., 2006; Nakajima et al., 2006).
Using a yeast two-hybrid assay, we examined the interac-
tion preference between GID1 and DELLA proteins from rice,
S. moellendorffii, and P. patens by measuring b-galactosidase
(b-gal) activity (Figure 6A). We performed the experiment in the
presence of 10�5 M GA4, which was an excess condition for the
Os GID1–SLR1 interaction in yeast (Ueguchi-Tanaka et al.,
2007b). Sm GID1a interacted with Sm DELLA1 and Os SLR1
with similar effectiveness and also interacted very weakly with
Sm DELLA2. Sm GID1b interacted actively with Sm DELLA1 and
Os SLR1 but less actively with Sm DELLA2. Os GID1 interacted
with Os SLR1 but not with Sm DELLA1 or Sm DELLA2. Pp
GID1L1 and Pp GID1L2 did not interact with any DELLA proteins.
Similarly, the Pp DELLAL proteins did not interact with any of the
GID1 proteins. We confirmed the expression of GID1 and DELLA
proteins in yeast cells by an immunoblot analysis (data not
shown), indicating that the lack of b-gal activity was caused not
by a failure in protein production in yeast cells but by a lack of
protein–protein interaction.
Figure 2. (continued).
(A) Amino acid sequences of GID1-like proteins from Arabidopsis and rice (angiosperms), S. moellendorffii (lycophyte), and P. patens (bryophyte) were
aligned with ClustalW (http://align.genome.jp/). The open triangle indicates the position of an intron shared by the true GID1 receptors (Os GID1 and
At GID1s) and Sm GID1s. Closed triangles at the bottom of the alignment are intron positions unique to Pp GID1L proteins. Small and large circles
represent the conserved residues and the catalytic triad in the HSL family, respectively. The positions of gid1-2 and gid1-5 mutations in rice mutants are
indicated by black diamonds at the top of the alignment. The amino acid residues are numbered from the first Met, and gaps (dashes) were introduced
to achieve maximum similarity. Black and gray boxes indicate identical and similar residues, respectively. At, Arabidopsis thaliana; Os, Oryza sativa; Pp,
Physcomitrella patens; Sm, Selaginella moellendorffii.
(B) Phylogenetic analysis of GID1 homologs. A maximum likelihood tree based on the JTT model (Jones et al., 1992) was obtained. The horizontal
branch lengths are proportional to the estimated number of amino acid substitutions per residue. Bootstrap values were obtained by 1000 bootstrap
replicates. Pr, Pinus radiata; Pine, Pinus taeda.
(C) A consensus phylogenetic species relationship among moss, lycophyte, pine, Arabidopsis, and rice. Note that node lengths do not reflect the
accurate divergence time of each species.
3062 The Plant Cell
Figure 3. Comparison of Deduced Amino Acid Sequences of DELLA Homologs in Land Plants.
(A) and (B) Amino acid sequence alignments of the N-terminal DELLA/TVHYNP domains (A) and C-terminal GRAS domains (B) of DELLA-like proteins
GA Perception in Nonseed Plants 3063
We then examined the effect of various GAs at 10�5 M on Sm
GID1–Sm DELLA1 interactions in yeast cells (Figures 6B and 6C).
These interactions were used as an indirect measure of the
binding preferences of Sm GID1 proteins for various kinds of
GAs. The structural features of GAs that affect the Sm GID1–Sm
DELLA1 interaction do not fit the rules established for GA
bioactivity in flowering plants. According to these rules, bioactive
GAs should contain 3b-hydroxylation of the A-ring, g-lactone
structure in the A-ring, and carboxylation of C7, whereas 2b-
hydroxylation of the A-ring causes inactivation of GA activity
(Davis, 2004). By contrast, the effect of GA37 and GA9 (10�5 M) on
the Sm GID1–Sm DELLA1 interactions was similar to that of GA4
for Sm GID1a–Sm DELLA1 and was even greater for Sm GID1b–
Sm DELLA1 (Figures 6B and 6C). Both GA37 and GA9 are
classified as inactive types because GA37 contains a d-lactone
structure in the A-ring and GA9 does not contain 3b-hydroxyla-
tion of the A-ring (Figure 6D). Furthermore, for the Sm GID1b–Sm
DELLA1 interaction, GA51 and 3-epi-GA4 showed much higher
activity than did GA1 or GA3, both of which are bioactive GAs in
flowering plants. GA51 contains 2b-hydroxylation rather than 3b-
hydroxylation in the A-ring, and 3-epi-GA4 contains 3a-hydrox-
ylation instead of 3b-hydroxylation. The unique effectiveness of
the Sm GID1–Sm DELLA1 interaction suggests that the GA
selectivity of Sm GID1 proteins differs from that of GID1 proteins
in flowering plants (see Discussion). By contrast, interactions
between Pp GID1L and Pp DELLAL proteins were not observed
in the presence of any of the GAs tested (see Supplemental
Figure 1 online).
To further study the GA binding properties of Sm GID1 pro-
teins, we examined the dose-dependence of the Sm GID1a– and
Sm GID1b–Sm DELLA1 interactions in yeast cells in the pres-
ence of GA4, GA1, GA3, GA9, and 3-epi-GA4 (Figure 7). For both
interactions, GA4 showed the highest affinity. The Sm GID1b–Sm
DELLA1 interaction increased as the GA4 concentration in-
creased from 10�9 to 10�5 M and finally reached a plateau, while
the saturation point for Sm GID1a–Sm DELLA1 was almost 10�7
M (Figure 7A). Consequently, the 50% saturation points of the
interactions were estimated as 8 3 10�9 M and 5 3 10�8 M,
respectively. This indicates that Sm GID1a has higher affinity for
GA4 than does Sm GID1b.
The higher GA affinity of Sm GID1a was more obvious when we
used GA1 or GA3 as a ligand (Figures 7B and 7C), whereas GA9
and 3-epi-GA4 had much stronger effectiveness in the Sm
GID1b–Sm DELLA1 interaction than did GA1 or GA3 (Figures
7D and 7E). These results demonstrate the unique GA selectivity
of Sm GID1 proteins, although the interacting affinity between
Sm GID1a and GA4 was highest (8 3 10�9 M) among any
interactions between Sm GID1 proteins and GAs, as with the
flowering plant GID1 proteins (Ueguchi-Tanaka et al., 2005;
Nakajima et al., 2006). The b-gal activity was always higher in
the Sm GID1b–Sm DELLA1 interaction than in Sm GID1a–Sm
DELLA1 at high concentrations of various GAs, even in the case
of low-affinity GAs such as GA1 and GA3. Although it is not clear
why the Sm GID1b–Sm DELLA1 interaction always causes
higher b-gal activity than the Sm GID1a–Sm DELLA1 interaction,
it is possible that the Vmax of the Sm GID1b–Sm DELLA1
interaction may be higher than that of the Sm GID1a–Sm DELLA1
interaction.
Functional Analysis of GA Signaling Genes of
S. moellendorffii and P. patens in Transgenic Rice Plants
To test whether the putative GA signaling genes in S. moellen-
dorffii possess similar functions to their flowering plant counter-
parts in vivo, we produced transgenic rice plants overexpressing
each gene under the control of the constitutive Act1 promoter
(proAct1) (McElroy et al., 1990). When we introduced Sm GID1a
and Sm GID1b into the gid1-3 background (Ueguchi-Tanaka
et al., 2005), all transgenic plants carrying the Sm GID1a (n¼ 6) or
Sm GID1b (n¼ 9) DNA sequences were taller than nontransgenic
gid1 dwarf plants (Figure 8A; see Supplemental Figure 2 online).
Sm GID1b plants were shorter than Sm GID1a or wild-type plants
(Figure 8A; see Supplemental Figure 2 online), even though the
levels of Sm GID1 mRNA were similar in both Sm GID1a and Sm
GID1b transgenic plants (data not shown). This suggests that the
GA receptor function of Sm GID1b is weaker than that of Sm
GID1a in rice cells. This is reasonable when we consider the GA
preference of Sm GID1b and that the dominant GA in the rice
plant at the vegetative stage is GA1 (Kobayashi et al., 1989): Sm
GID1b is much less sensitive than Sm GID1a to GA1 under
physiological conditions (<10�5 M) (Figure 7B).
To examine DELLA-like protein function, we introduced Sm
DELLA1, Sm DELLA2, and Pp DELLAL1 clones into wild-type
Taichung 65 (T65) rice. We used a wild-type background be-
cause the regeneration frequency of rice slr1 callus is quite low,
and if the introduced proteins had little or no DELLA function, it
would be difficult to obtain any transgenic plants. On the other
hand, transgenic plants overproducing functional DELLA pro-
teins would be expected to have a semidwarf phenotype be-
cause of the suppression of GA action by accumulated DELLA
proteins (Itoh et al., 2002). As expected, the plants overexpress-
ing Sm DELLA1 showed a dwarf phenotype; the severity of
dwarfism appeared to be associated with the level of Sm
DELLA1 expression (Figure 8B; see Supplemental Figure 2
online). On the other hand, we did not see dwarf plants among
>30 independent plants transformed with constructs for Sm
Figure 3. (continued).
from Arabidopsis and rice (angiosperms), S. moellendorffii (lycophyte), and P. patens (bryophyte) obtained using ClustalW (http://align.genome.jp/). The
conserved regions or domains are presented at the top. The amino acid residues are numbered from the first Met, and gaps (dashes) were introduced to
achieve higher similarity scores. Black and gray boxes indicate identical and similar residues, respectively. Abbreviations are as in Figure 1.
(C) Phylogenetic analysis of DELLA protein homologs. A maximum likelihood tree based on the JTT model (Jones et al., 1992) was obtained. The
horizontal branch lengths are proportional to the estimated number of amino acid substitutions per residue. Bootstrap values were obtained by 1000
bootstrap replicates. Zm, Zea mays.
3064 The Plant Cell
DELLA2 or Pp DELLAL1 overexpression. These observations
demonstrate that Sm DELLA2 and Pp DELLAL1 do not suppress
GA signaling in rice cells or that their effects are at an undetect-
able level.
Introduction of Sm GID2a partially rescued the dwarf pheno-
type of gid2-1 (Figure 8C; see Supplemental Figure 2 online),
indicating that Sm GID2a functions at least in part as a GA-
related F-box protein in rice. The lack of complete rescue of the
gid2 dwarf phenotype by transformation with Sm GID2a may be
caused by an incomplete degradation of rice DELLA protein
SLR1 by Sm GID2a, possibly because of the low interaction
between SLR1 and Sm GID2a. On the other hand, the dwarf
phenotype of gid2-1 was not rescued by overexpression of Pp
GID2L1; this result is not surprising given the sequence dif-
ferences observed (Figure 4). Again, the homologous gene in
P. patens was unable to substitute for its counterpart gene in rice.
Expression Analysis of GA-Related Genes
in S. moellendorffii
The above results demonstrate that S. moellendorffii has all the
components of GA signaling we have examined: GID1, DELLA,
and GID2 proteins. To confirm the expression of these genes in
S. moellendorffii, we performed RT-PCR using total RNA isolated
from various organs (Figure 9). For this experiment, we prepared
primers specific for Sm GID1a, Sm GID1b, Sm DELLA1, and Sm
GID2a. The primers produced almost the same amount of PCR
product for each gene when used with S. moellendorffii genomic
Figure 4. Comparison of Deduced Amino Acid Sequences of GID2 Homologs in Land Plants.
(A) Amino acid sequence alignment of GID2-like proteins from Arabidopsis and rice (angiosperms), S. moellendorffii (lycophyte), and P. patens
(bryophyte) calculated using ClustalW (http://align.genome.jp/). The conserved regions or domains are presented at the top. The amino acid residues
are numbered from the first Met, and gaps (dashes) were introduced to achieve maximum similarity. Black and gray boxes indicate identical and similar
residues, respectively. Abbreviations are as in Figure 1.
(B) Phylogenetic analysis of GID2 homologs. A maximum likelihood tree based on the amino acid alignment was obtained using the JTT model (Jones
et al., 1992). The horizontal branch lengths are proportional to the estimated number of amino acid substitutions per residue. Bootstrap values were
obtained by 1000 bootstrap replicates. Ta, Triticum aestivum.
GA Perception in Nonseed Plants 3065
DNA as the template (Figure 9A), indicating that they amplify
each product at a similar efficiency. We then compared the ex-
pression levels of these genes using cDNA produced from total
RNA of the vegetative shoots of S. moellendorffii (Figure 9A).
There was more PCR product from Sm GID1b than from the other
genes, indicating that Sm GID1b is more actively transcribed in
the shoots than the other genes. As the expression levels of each
gene differed, we selected a suitable cycle number for PCR
amplification of each gene to compare the amounts of mRNA in
various organs (Figure 9B). The products of each gene were
Figure 5. GA Binding Properties of GID1 Homologs.
(A) Affinity-purified Trx�His-Os GID1 or its homologs were incubated with 2 3 10�8 M 3H4-H2-GA4 in the presence or absence of GST-DELLA protein.
The specific binding activity (B-UB) was calculated by subtracting the nonspecific binding activity (UB), which was evaluated by the addition of 0.125
mM GA4 to the assay solution, from the total binding activity (B).
(B) GA binding activity of affinity-purified Trx�His-GID1 homologs under a higher concentration of 3H4-H2-GA4 (2 3 10�7 M).
For (A) and (B), SD values were determined from more than three measurements. ** P < 0.01, *** P < 0.001 against vector control (data not shown),
determined using the unpaired t test.
(C) SDS-PAGE profile of affinity-purified Trx�His-Os GID1 and its homologs. Circles indicate the recombinant proteins with approximately the expected
molecular sizes (Os GID1, 57.0 kD; Sm GID1a, 57.7 kD; Sm GID1b, 59.2 kD; Pp GID1L1, 55.1 kD; and Pp GID1L2, 56.1 kD). The identification of protein
corresponding to Os GID1 and its homologs was confirmed by immunoblot analysis using an Os GID1 antibody (data not shown). Although several
bands were observed in each of the Sm GID1 lanes, the immunoreacting band was considered to be Sm GID1 protein, and it was concluded that the
other bands were the result of degradation, immaturely transcribed protein, or protein unrelated to Sm GID1s. Approximately equal amounts (;16 mg)
of protein were used for the GA binding assay.
(D) to (F) Isothermal titration calorimetry analysis of the Pp GID1L–GA interaction. The heat exchange versus molar ratio for Trx�His-GID1 with titration of
GA is shown.
(D) Integrated titration curve of Trx�His-Os GID1 with GA4. The line represents the best fitting curve calculated from a single-site binding model.
(E) Integrated titration curve of Trx�His-Pp GID1L1 with epi-GA4 (red line), GA9 (blue line), and GA12 (green line).
(F) Integrated titration curve of Trx�His-Pp GID1L2 with epi-GA4 (red line), GA9 (blue line), and GA12 (green line).
3066 The Plant Cell
Figure 6. Interaction between GID1 and DELLA Homologs in Yeast Cells.
GA Perception in Nonseed Plants 3067
observed in various organs, and the levels did not appear to differ
much among organs. This lack of organ-specific expression is
similar to the expression patterns of GID1, DELLA, and GID2/
SLY1 in rice and Arabidopsis (Kaneko et al., 2003; McGinnis
et al., 2003; Gomi et al., 2004; Nakajima et al., 2006).
GA Signaling and Synthesis in S. moellendorffii
The presence of functional homologs of GID1, DELLA, and GID2
in S. moellendorffii strongly indicates that GA functions as a
bioactive substance in this species. To confirm this, we treated
S. moellendorffii bulbils with 10�5 M GA4 or 10�6 M uniconazole,
a GA biosynthesis inhibitor, and observed the effect on plant
growth at ;2 weeks after germination by measuring the length
between adjacent small leaves. The stem length of GA4-treated
plants was increased slightly compared with that of the untreated
plants, although the unpaired t test did not indicate a significant
difference (P > 0.05) (Figures 10A and 10B). By contrast, the
uniconazole treatment caused a significant dwarf phenotype
compared with the untreated plants (P < 0.01). However, treat-
ment with 10�5 M GA4 at 10 d after uniconazole treatment did not
restore its inhibitory growth effect on the plant (data not shown).
We then examined the GA signaling pathway in S. moellen-
dorffii using a molecular biological approach. GA negatively
regulates the expression of genes encoding GA-synthesizing
enzymes such as GA20 oxidase and GA3 oxidase (Chiang et al.,
Figure 6. (continued).
(A) Interaction of various combinations of GID1 and DELLA proteins from rice, S. moellendorffii, and P. patens. GID1 proteins and DELLA proteins were
used as bait and prey, respectively. b-Gal activity was detected in a liquid assay with Y187 transformants (means 6 SD; n ¼ 3). Only results in the
presence of 10�5 M GA4 are presented, since no activity > 1.4 Millar units was detected in the absence of GA4 in any combination.
(B) and (C) Effects of various GAs on the Sm GID1s–Sm DELLA1 interaction in yeast cells. A two-hybrid assay was performed using Sm GID1s as bait
and Sm DELLA1 as prey in the presence of 10�5 M of various GAs. b-Gal activity was determined as in (A) (means 6 SD; n ¼ 3).
(D) Chemical structures of GAs used in this study. Structures essential for bioactive GAs are circled in gray (free 2b- and 3b-hydroxylation of the A-ring,
g-lactone structure in the A-ring, and carboxylation of C7). The characteristic structure of each GA compared with GA4 is highlighted in gray. H2-GA4,
16,17-dihydro-GA4; GA4-Me, GA4 methyl ester; GA9-Me, GA9 methyl ester.
Figure 7. Dose-Dependence of GA4, GA1, GA3, GA9, and 3-epi-GA4 in the Sm GID1s–Sm DELLA1 Interaction in Yeast Cells.
Two-hybrid assay using Sm GID1 proteins as bait and Sm DELLA1 as prey in the presence of various concentrations of GA4 (A), GA1 (B), GA3 (C), GA9
(D), and 3-epi-GA4 (E). b-Gal activity was determined as in Figure 6A (means 6 SD; n ¼ 3). The 50% saturation points are indicated by arrows.
3068 The Plant Cell
1995; Phillips et al., 1995). To examine the downregulation of
GA20 and GA3 oxidase genes by GA in S. moellendorffii, we
screened and isolated genes homologous with the GA20 and
GA3 oxidases. We found three GA20 oxidase–like genes and one
GA3 oxidase–like gene in the S. moellendorffii genome (see
Supplemental Figure 3 online). We examined the GA20 or GA3
oxidase activity of each gene’s protein product and confirmed
that one of the GA20 oxidase candidates, Sm GA20ox, and one
GA3 oxidase–like gene, Sm GA3ox, encode functional enzymes
(see below).
We examined the expression of GA-related genes in shoots of
S. moellendorffii with or without GA4 treatment (Figure 10C). With
GA4 treatment, the expression of Sm GID1b, Sm GA20ox, and
Sm GA3ox decreased markedly, while the expression of Sm
GID1a, Sm DELLA1, Sm GID2a, and Sm GA20oxL1 was un-
changed. Downregulation of the GID1 gene by GA treatment also
occurs in rice and Arabidopsis seedlings (Griffiths et al., 2006,
M. Ueguchi-Tanaka, unpublished data), indicating that suppres-
sion of the GID1 gene by GA is a common phenomenon in vas-
cular plants. Furthermore, the downregulation of Sm GA20ox and
Sm GA3ox by GA strongly suggests that the negative feedback
regulation of GA synthesis genes mediated by the GA signaling
pathway functions in S. moellendorffii as in flowering plants.
We further examined the GA-dependent degradation of Sm
DELLA1 in young shoots of S. moellendorffii because this phe-
nomenon is one of the most direct and sensitive events under the
control of the GA signaling pathway mediated by the GID1/
DELLA system. First, we performed a protein blot analysis of the
crude extract of young shoots of S. moellendorffii with the
antibody to the rice DELLA protein, SLR1 (Itoh et al., 2002), but
this antibody did not detect Sm DELLA1 (data not shown). We
then produced a specific antibody to Sm DELLA1 and used it for
protein gel blot analysis. The antibody recognized a single band,
which migrated at the estimated molecular weight of Sm DEL-
LA1. This band was detected in a transgenic rice plant carrying
the proAct1-Sm DELLA1 construct but not in a control plant
(Figure 10D). An immunoreactive band with the same mobility
was detected in the extract of young shoots of S. moellendorffii;
this band almost disappeared within 12 h of the application of
10�4 M GA4 (Figure 10D). This result clearly demonstrates that
the GA perception pathway mediated by the GID1/DELLA sys-
tem occurs in S. moellendorffii.
Next, we directly examined the amount of bioactive GA in S.
moellendorffii (Figure 11B). Some level of GA4 and a very low
level of GA24 were detected in young shoots, while no GA1 or
GA19 was found. Identification of GA4 was confirmed by GC-MS
(see Supplemental Figure 4 online). Detection of non-13-OH
(hydroxy)-GAs (such as GA4 and GA24) but not 13-OH-GAs (such
as GA1, GA19, and GA20) suggests that S. moellendorffii may
preferentially produce non-13-OH-GAs and use GA4 as a bio-
active GA. This is reasonable given that Sm GID1a and -1b
preferentially interact with GA4 but not with GA1 (Figures 6B, 6C,
7A, and 7B). To test this possibility, we examined the enzymatic
activity of Sm GA20ox, Sm GA20oxL1, Sm GA20oxL2, and Sm
Figure 8. Complementation of the Dwarf Phenotype of Rice gid1 and
gid2 Mutants by Expression of Sm GID1s, Sm GID2a, and Pp GID2L1,
and Phenotypic Analysis of Rice Overproducers of Sm DELLAs and Pp
DELLAL1.
(A) Gross morphology of the wild type, gid1-3, and Sm GID1a and Sm
GID1b overproducers in gid1-3 mutants at the young seedling stage.
Expression of Sm GID1a and Sm GID1b completely and partially
complemented the gid1-3 dwarf phenotype, respectively.
(B) Gross morphology of the wild type and Sm DELLA1, Sm DELLA2, and
Pp DELLAL1 overproducers in wild-type T65 plants. The panels at
bottom present the results of RT-PCR analysis of each transcript of the
transgene. Higher expression of Sm DELLA1 was associated with more
severe dwarfism of transformants, while there was no dwarfism in
transformants highly expressing Sm DELLA2 or Pp DELLAL1. The Os
ACT1 gene was used as an internal standard to ensure that the same
amount of cDNA was used as the DNA template in each PCR. Results
presented are representative of three independent experiments.
(C) Gross morphology of the wild type, gid2-1, and Sm GID2a and Pp
GID2L1 overproducers in the gid2-1 mutant at the young seedling stage.
Only Sm GID2a partially complemented the gid2-1 dwarf phenotype.
GA Perception in Nonseed Plants 3069
GA3ox produced in E. coli. The recombinant Sm GA20ox cata-
lyzes the conversion of GA12 to GA15, GA24 to GA9, and GA53 to
GA44, while the conversion of GA19 to GA20 was not observed
(Figures 11A and 11C). This suggests that Sm GA20ox catalyzes
the conversion of GA12 to GA9 but converts GA53 to GA20
inefficiently if at all. Similarly, Sm GA3ox catalyzed the conver-
sion of GA9 to GA4 but not GA20 to GA1 (Figures 11A and 11C).
Two other GA20 oxidase homologs in S. moellendorffii, Sm
GA20oxL1 and Sm GA20oxL2, did not catalyze any conversions.
These results demonstrate that Sm GA20ox and Sm GA3ox
catalyze the GA20 oxidation and the GA3 oxidation of non-13-
OH-GAs but do not efficiently use 13-OH-GAs as substrates.
This observation supports the idea that S. moellendorffii prefer-
entially produces non-13-OH-GAs and uses GA4 as a bioactive
GA. We also examined the endogenous GA content (see Sup-
plemental Table 1 online) and enzymatic activity of homolog
proteins of GA20 oxidase and GA3 oxidase in P. patens (see
Supplemental Table 2 online), but we did not detect either
endogenous GA or GA oxidase activity in this species.
DISCUSSION
Evolution of GA Perception Mediated by the
GID1/DELLA System
In silico screening of the genomic DNA of S. moellendorffii and P.
patens was used to identify candidate genes for each of three GA
signal-related genes: GID1, DELLA, and GID2. To determine
whether any of these candidates are actually involved in GA
signaling, we studied their biochemical and biological properties.
We used an in vitro GA binding assay and a yeast two-hybrid
assay to examine the GA binding activity of GID1 candidate
proteins and their GA-dependent interaction with DELLA pro-
teins. We examined the ability of the GID2 candidate genes to
complement a rice gid2 mutant. We also used transgenic ex-
periments with rice mutants and wild types to study the GID1 and
DELLA candidates, respectively.
All results from these experiments demonstrate that S. moel-
lendorffii contains bioactive GID1, DELLA, and GID2 counterpart
genes. Furthermore, the expression of GA biosynthesis genes,
such as Sm GA20ox and Sm GA3ox, and the GA receptor Sm
GID1b in S. moellendorffii is downregulated by GA treatment,
similar to the GA feedback regulation in flowering plants (Chiang
et al., 1995; Phillips et al., 1995; Griffiths et al., 2006). The
GA-dependent degradation of DELLA-like protein also strongly
supports the presence of a similar GA perception system in
S. moellendorffii. However, treatment with GA4 failed to restore
the dwarf phenotype of S. moellendorffii caused by uniconazole,
suggesting either that the effect of uniconazole was not due to
the inhibition of GA biosynthesis or that the timing of our
treatment was inappropriate. Although we did not observe clear
changes in strobilus formation or sporulation when S. moellen-
dorffii was treated with GA4 or uniconazole (data not shown), it is
still possible that unusual GAs in flowering plants may be
involved in reproduction, especially in the development of sexual
organs, because some GAs (antheridiogens) are involved in the
formation of sexual organs in fern gametophytes (Yamauchi
et al., 1996; Banks, 1999; Menendez et al., 2006). Taken together,
we conclude that S. moellendorffii has a GA perception mech-
anism mediated by the GID1/DELLA system similar to that found
in flowering plants. Further studies on the biological function of GA
in S. moellendorffii should expand our knowledge of GA function.
In contrast with S. moellendorffii, we found no functional
homologs in the moss P. patens, although there are some genes
encoding proteins homologous with rice GID1, SLR1, and GID2.
The GID1 homologs in P. patens, Pp GID1L1 and Pp GID1L2, did
not show in vitro GA binding activity in two different analyses, nor
did they interact with any DELLA proteins in the presence of
various kinds of GAs in yeast cells. Transgenic rice expressing
the gene encoding Pp DELLAL1, a protein that does not contain
typical DELLA/TVHYNP domains but contains conserved GRAS
domains similar to those in seed plant DELLA proteins, did not
show any GA-insensitive phenotypes. Furthermore, the over-
expression of Pp GID2L1 in rice gid2-1 did not rescue its dwarf
phenotype. These results suggest that P. patens does not
contain a GA perception mechanism mediated by the GID1/
DELLA system, although it is possible that GAs other than those
that we tested may initiate the GID1/DELLA pathway in this
organism. Another possibility is that we did not find functional
homologs of GID1, DELLA, and GID2 due to the incomplete
sequence of the P. patens genome. However, we found several
homologs of the auxin receptor TIR1 and auxin signal compo-
nents such as Aux/IAA and ARF (Dharmasiri and Estelle, 2002)
in the P. patens genome (data not shown), suggesting that
we would have found a functional homolog to at least one
Figure 9. Expression of GA Signaling Genes in S. moellendorffii.
(A) PCR was performed using genomic DNA from S. moellendorffii or
cDNA produced from the apical part of vegetative shoots as a DNA
template. The results of genomic PCR indicate that the primers for each
gene work similarly. Compared with other genes, Sm GID1b is prefer-
entially expressed in the apical part of vegetative shoots. The number of
PCR cycles used is shown at right.
(B) RT-PCR of GA signaling genes in various organs of S. moellendorffii.
Total RNA was isolated from the organs indicated at the top, and 2 mg
was used for the RT reaction. ‘‘Roots’’ indicates a mixture of roots and
rhizophores, and ‘‘apices’’ indicates apical parts of the vegetative shoot.
Expression of the Sm 6PGD gene, an ortholog of the Sr 6PGD gene
(Tanabe et al., 2003), was used as a control. The number of PCR cycles
used is shown at right. Results presented are representative of at least
three independent experiments.
3070 The Plant Cell
of the three GA signal–related genes if it were present in the
genome.
Very recently, a similar study on the evolution of the GA/DELLA
mechanism in Arabidopsis, Selaginella kraussiana, and P. patens
was reported (Yasumura et al., 2007). Although the authors
suggest that the GA perception mechanism emerged after
bryophytes and vascular plants diverged, which is consistent
with our observations, some intriguing differences were also noted.
First, Pp GLP1 (corresponding to Pp GID1L1) and Sk DELLA
(S. kraussiana DELLA protein) interacted in a GA-independent
manner in the yeast two-hybrid assay, whereas Pp DELLAa (cor-
responding to Pp DELLAL1) did not bind to GID1 of any species
tested. Second, introducing GFP-Pp DELLAa into an Arabidop-
sis gai-t6 rga-24 ga1-3 triple mutant strain (null for the DELLA
proteins GAI and RGA and containing almost no GA) resulted in
dwarfism, whereas the triple mutant normally confers a tall
mutant phenotype. From these observations, the authors sug-
gest that the ability of GID1 to interact with DELLA, and the
growth-restraining ability of DELLA, evolved before bryophyte
and lycophyte divergence, whereas the ability of DELLA to bind
to GID1 evolved at a later stage, between the divergence of
bryophytes and lycophytes.
These results are clearly different from our findings. We did not
detect an interaction of Pp GID1L proteins with any DELLA
protein tested (Figure 6A), and overproduction of Pp DELLAL1 in
our rice plants did not cause dwarfing (Figure 8B; see Supple-
mental Figure 2 online). In addition, the unique GA preference
observed for Sm GID1 proteins (Figures 6B and 6C) was not
observed for the S. kraussiana GID1. The use of different plant
species (Arabidopsis versus rice, S. kraussiana versus S. moel-
lendorffii) could be the major cause of these differences. For Pp
DELLAL1, differences in the growth-restraining activity observed
for rice and Arabidopsis may be caused not merely by differ-
ences in plant species but may also depend on whether mutant
or wild-type plants were used (gai-t6 rga-24 ga1-3 for Arabidop-
sis, wild-type T65 for rice).
Green plants first colonized land around the mid-Ordovician
period (470 million years ago) and subsequently diverged into
various lineages (Kenrick and Crane, 1997). The mosses and
vascular plants diverged early in the Silurian period (430 million
year ago) (Kenrick and Crane, 1997). The presence of a GID1/
DELLA-mediated GA perception mechanism in S. moellendorffii,
which diverged early from the lineage of ferns and seed plants,
strongly suggests that the last common ancestor of vascular
Figure 10. Effects on Growth, Feedback Regulation of GA-Related Genes, and Degradation of Sm DELLA1 of GA4 Treatment in S. moellendorffii.
(A) Gross morphology of 10�5 M GA4-treated and 10�6 M uniconazole-treated plants. Ethanol (0.01%) solution was used as a control. The positions of
corresponding small leaves on each plant are connected with white lines. EtOH, ethanol; uni, uniconazole.
(B) Length of stem between the second and third small leaves from the bottom (6SE; n ¼ 12, 20, and 10 for GA4, ethanol, and uniconazole treatments,
respectively). ** Significant difference (P < 0.01) compared with control (ethanol) treatment from the unpaired t test analysis.
(C) Downregulation of GA-related genes in S. moellendorffii after GA4 treatment. Plants were treated for 3 d with or without GA4 at a concentration of
10�4 M, total RNA was isolated from young shoots, and RT-PCR was performed. The Sm 6PGD gene was used as a control. Results presented are
representative of at least three biological replicates.
(D) Disappearance of Sm DELLA1 protein after GA4 treatment. S. moellendorffii plants were treated with either buffer only or 10�4 M GA4 for 12 h.
Sm DELLA1 protein was detected by immunoblot analysis of crude protein extract using the anti-Sm DELLA1 antibody. The specificity of the antibody
was confirmed using transgenic rice overexpressing Sm DELLA1 (positive control) and transgenic rice possessing vector only (negative control).
GA Perception in Nonseed Plants 3071
plants used a GID1/DELLA-mediated GA perception system.
By contrast, failure to identify functional homologs of GID1,
DELLA, GID2, GA20 oxidase, or GA3 oxidase in P. patens
suggests that bryophytes may not use GA as a bioactive sub-
stance, although the possibility that P. patens utilizes GAs that
were not tested in this study, or that it contains an alternative
GA signaling pathway, cannot be excluded. It is also necessary
to examine whether other bryophyte species (mosses, horn-
worts, and liverworts) lack the GID1/DELLA-mediated percep-
tion system and functional homologs of GA20 oxidase and GA3
oxidase. Based on our current data, we hypothesize that the
usage of GA as a bioactive substance was a key event in
the evolution of the body plans of vascular plants. However,
further studies of GA perception, response, and biosynthesis
Figure 11. GA Content and in Vitro Activity of GA20 Oxidase and GA3 Oxidase Homologs in S. moellendorffii.
(A) The late stage of GA biosynthesis. In many flowering plants, GA12 is often converted to GA53 by hydroxylation at C-13. GA12 or GA53 is converted, via
parallel pathways, to other GAs through a series of oxidations at C-20 to finally form GA9 or GA20 by GA20 oxidase. GA9 or GA20 is oxidized to the
bioactive GA4 or GA1 by GA3 oxidase. Sm GA20ox can catalyze the steps from GA12 to GA15, from GA24 to GA9, and from GA53 to GA44 (circles), but not
from GA19 to GA20 (crosses). Sm GA3ox catalyzes the step from GA9 to GA4 but not the step from GA20 to GA1.
(B) GA content of S. moellendorffii shoots. One gram of S. moellendorffii shoots was used for GA content measurement by liquid chromatography–mass
spectrometry analysis (see Methods). Tests were performed on four independent plants. Error bars indicate SD. FW, fresh weight; N.D., no expected
product was detected.
(C) In vitro enzymatic activity of GA20 oxidase and GA3 oxidase homologs in S. moellendorffii. N.D., no expected product was detected.
3072 The Plant Cell
in nonflowering land plants are necessary to evaluate this
hypothesis.
Properties of Sm GID1s
The ligand selectivity of Sm GID1 proteins, especially Sm GID1b,
differed from that of GID1 proteins in flowering plants (Figures 6
and 7). For example, the effectiveness of GA37 and GA9 at 10�5 M
was similar to that of GA4, the most effective GA in flowering
plants, for the Sm GID1a–Sm DELLA1 interaction and higher than
that of GA4 for the Sm GID1b–Sm DELLA1 interaction. By
contrast, bioactive GAs such as GA1 and GA3 had a very low
effect on the Sm GID1b–Sm DELLA1 interaction and an inter-
mediate effect on the Sm GID1a–Sm DELLA1 interaction. Fur-
thermore, GAs known to be inactive in flowering plants, such as
GA51 and 3-epi-GA4, had an apparent effect on the Sm GID1b–
Sm DELLA1 interaction.
The relatively high effectiveness of 3-epi-GA4 indicates that a
stereoscopic hydroxylation structure at the C3 site is much less
important for Sm GID1s, especially for Sm GID1b, than for GID1s
in flowering plants. The intermediate effectiveness of GA51 also
indicates that the absence of 3b-hydroxylation and the presence
of 2b-hydroxylation does not prevent the interaction of the GA
molecule with Sm GID1 proteins. Moreover, the high effective-
ness of GA37 demonstrates that the g-lactone structure in the
A-ring can be replaced with the d-lactone structure. These results
suggest that Sm GID1 proteins, especially Sm GID1b, do not
recognize the A-ring of GA as strictly as do GID1 proteins in
flowering plants. By contrast, the low or almost nonexistent
effect of GA1 and GA3 at the concentration of 10�5 M on the Sm
GID1a–Sm DELLA1 or Sm GID1b–Sm DELLA1 interaction sug-
gests that Sm GID1 proteins more strictly recognize the C-ring
structure of GA than do GID1s in flowering plants. Specifically,
Sm GID1b discriminates between GAs based on the presence
or absence of 13-hydroxylation: GA4 is active, whereas GA1
is inactive.
Such discrimination was also observed in the reaction of GA3
oxidation by Sm GA3 oxidase; that is, Sm GA3 oxidase actively
catalyzed the conversion of GA9 to GA4 but not that of GA20 to
GA1 (Figure 11C). These results indicate that S. moellendorffii
specifically metabolizes non-13-OH-GAs to produce GA4 as an
active GA and specifically perceives GA4 with Sm GID1 proteins.
In flowering plants, 13-OH- and non-13-OH-pathways are used
differentially from species to species. For example, GA1 pre-
dominates in cereals and legumes and GA4 predominates in
Arabidopsis and cucurbits (Kobayashi et al., 1989; Fleet et al.,
2003; Davis, 2004; Lange et al., 2005). Moreover, the two
pathways are also used differentially among organs within a
species or in response to different environmental conditions
(Davis, 2004). It is possible that the non-13-OH-pathway is the
default state of GA synthesis, and the end product of this
pathway, GA4, is the most active GA in terms of affinity to the
GID1/GA receptor. In this context, the 13-OH-GA synthetic
pathway may have come into existence at later stage(s) during
the evolution of vascular plants. If so, it is interesting to speculate
why vascular plants developed the 13-OH-GA synthetic pathway
during their evolution, given the biological significance of 13-OH–
type GAs such as GA1 and GA3.
METHODS
Plant Materials and Growth Conditions
Physcomitrella patens subsp patens, originally collected in Gransden
Wood (Ashton and Cove, 1977), was grown on BCDATG medium at 258C
under continuous light (Nishiyama et al., 2000). Vegetatively propagated
protonemata containing young gametophores at 13 d after inoculation
were used for RNA extraction and analysis of endogenous GA content.
Selaginella moellendorffii was grown in the laboratory at room temper-
ature. A japonica-type rice cultivar (Oryza sativa cv Taichung 65) and its
irradiation-induced mutants, gid1-3 (Ueguchi-Tanaka et al., 2005) and
gid2-1 (Sasaki et al., 2003), were used to create transgenic rice plants.
Rice plants were grown in a growth chamber at 308C under continuous
light.
Screening of GA-Related Genes in S. moellendorffii and P. patens,
and Phylogenetic Analysis
The amino acid sequences of rice GID1, SLR1, GID2, GA20ox2, and
GA3ox1 were used as queries to screen the available S. moellendorffii
and P. patens genomic databases (http://selaginella.genomics.purdue.
edu/cgi-bin/blast_tmpl_s.cgi and http://moss.nibb.ac.jp/) by TBLASTN,
and candidate genes were selected for preliminary phylogenetic analy-
ses. The genes used in the preliminary analyses along with their e-values
and bit thresholds are as follows: Sm GID1 (Æe-50, bit æ 200), Sm DELLA
(Æe-70, bit æ 250), Sm GID2 (Æe-03, bit æ 30), Sm GA20ox (Æe-30, bit æ 100),
Sm GA3ox (Æe-20, bit æ 70), Pp GID1L (Æe-20, bit æ 100), Pp DELLAL (Æe-100,
bit æ 300), Pp GID2L (Æe-04, bit æ 30), Pp GA20ox (Æe-26, bit æ 100), and Pp
GA3ox (Æe-20, bit æ 100). For GA oxidase genes, candidate sequences
were also manually checked and selected for further analyses based on
the presence of conserved amino acids important for their function.
After obtaining candidate sequences, the deduced full coding region
for each gene was PCR-amplified using cDNA of S. moellendorffii or
P. patens as the template. Exon and intron regions were confirmed by
sequencing the PCR products. Highly similar sequences from other
species were identified with PSI-BLAST (Altschul et al., 1997) from a
combined data set including the National Center for Biotechnology
Information nonredundant data set, the poplar v1.1 proteins (proteins.
Poptr1_1.JamboreeModels.fasta) data set, and the Computational Biol-
ogy and Functional Genomics Laboratory (http://compbio.dfci.harvard.
edu/tgi/plant.html). The queries, inclusion limits, and e-value thresholds
were set differently for different families (see Supplemental Table 3
online).
Tentative clusters of expressed sequence tags in pine (Pinus taeda)
were searched with TBLASTN using SLR1, GID1, and GID2 as queries,
and the translated sequences were obtained. Because GA3 oxidases
were identified as distantly related in the data set of GA20 oxidases and
the outgroup found in the search had overlaps, the GA3 oxidase and
GA20 oxidase sequences were analyzed together. The sequences of
each data set were aligned with P. patens and S. moellendorffii se-
quences determined in this study using the einsi algorithm of MAFFT
version 6.2 (Katoh et al., 2005). The unambiguously aligned regions were
manually selected, and partial sequences lacking those regions were
removed with MacClade version 4.08 (http://macclade.org/index.html). A
neighbor-joining (NJ) tree (Saitou and Nei, 1987) was obtained with
PROTDIST and NEIGHBOR in the PHYLIP version 3.65 package (http://
evolution.genetics.washington.edu/phylip.html). Bootstrap analyses
were performed by repeating the procedure on 100 data sets prepared
with SEQBOOT.
To find a maximum likelihood tree, distantly related outgroup se-
quences were reduced and ingroup sequences from rice, Arabidopsis,
pine, S. moellendorffii, and P. patens plus some well-characterized genes
were selected. A distance matrix was obtained with ProtML in the
GA Perception in Nonseed Plants 3073
MOLPHY-2.3 package (Adachi and Hasegawa, 1996) under the JTT
model (Jones et al., 1992), and NJ trees were obtained with NJDist.
Maximum likelihood trees were searched with the nearest neighbor
interchange algorithm implemented in ProtML as a local rearrangement
search starting with the NJ tree. Bootstrap analyses were performed by
repeating the procedure on 1000 data sets prepared with SEQBOOT. For
the alignments shown in Figures 2 to 4, sequences were aligned with
ClustalW version 1.81 with default parameters (Thompson et al., 1994;
http://align.genome.jp/), followed by manual alignment. Boxshade (http:
//bioweb.pasteur.fr/seqanal/interfaces/boxshade.html) was used to
draw the alignments with default parameters.
Genomic DNA Isolation, RNA Isolation, and cDNA Synthesis
For isolation of the genomic DNA from S. moellendorffii and total RNA of
S. moellendorffii and rice, plants were ground with liquid nitrogen in a
mortar and pestle. Genomic DNA was isolated using ISOPLANT (Nip-
pongene) according to the instruction manual. Total RNA of S. moellen-
dorffii and rice was isolated using Trizol reagent (Invitrogen). Total RNA
from P. patens was extracted according to Hasebe et al. (1998) and
further purified with ISOGEN-LS (Wako Pure Chemical). Total RNA was
treated with RNase-free DNase for 30 min at 378C, followed by phenol:
chloroform:isoamyl alcohol (25:24:1) extraction and ethanol precipitation.
Single-stranded cDNA was synthesized from 2 mg of total RNA using the
OmniScript reverse transcriptase kit (Qiagen), according to the instruc-
tion manual.
Plasmid Construction
The deduced full coding region of each gene was obtained by PCR using
cDNA of S. moellendorffii or P. patens as the template. For each product,
the sequence was confirmed to ensure that no mutations were intro-
duced. The primers used in this study are listed in Supplemental Table 4
online. For the GID1 genes, Sm DELLA genes, Pp DELLAL1, and Sm
GID2a, each PCR product was first cloned into the pCR4 Blunt-TOPO
vector (Invitrogen) and the plasmid was further used for subcloning into
various vectors. Pp GID2L1 and several GA20ox- and GA3ox-like genes
were cloned directly into the vector of purpose.
For the GA binding assay, GID1 genes each containing suitable
restriction enzyme sites at both ends were cloned into the pET32a vector
(Novagen/Merck Biosciences) to produce the Trx�His-GID1 plasmids.
Cloning was performed using the BamHI-HindIII site for Os GID1, Sm
GID1b, and Pp GID1L2; using SalI-NotI for Sm GID1a; and using BamHI-
XhoI for Pp GID1L1. DELLA genes with suitable restriction enzyme sites at
both ends were cloned into the pGEX-4T-1 vector (GE Healthcare), using
the EcoRI site for Os SLR1 and EcoRI-NotI for the Sm DELLA genes, to
produce the GST-DELLA plasmids.
For the yeast two-hybrid assay, Sm GID1 and Pp GID1L genes
containing appropriate restriction sites at both ends were cloned into a
pGBKT7 DNA-BD shuttle vector (Clontech), using the NdeI-SmaI site for
Sm GID1a and Pp GID1L1, NcoI-SmaI for Sm GID1b, and NdeI-EcoRI for
Pp GID1L2, to produce pGBKT7 DNA-BD-GID1 bait plasmids. Similarly,
the entire coding region of each Sm DELLA and Pp DELLAL sequence
containing appropriate restriction sites at both ends was cloned into the
pGADT7 AD vector (Clontech) using the NdeI-SmaI site to produce
pGADT7 AD-DELLA prey plasmids. The Os GID1 bait plasmid and the Os
SLR1 prey plasmid were constructed as described previously (Ueguchi-
Tanaka et al., 2005).
To construct vectors for the production of transgenic rice, Sm GID1,
Sm DELLA, and Sm GID2a genes containing appropriate restriction sites
were introduced at the site between proAct1 (McElroy et al., 1990) and the
NOS terminator of the binary vector pActNos/Hm2 (Sentoku et al., 2000)
to produce proAct1-Sm GID1, -Sm DELLA, and -Sm GID2 transformation
vectors. The XbaI-SmaI site was used for cloning the Sm GID1, Sm
DELLA, and Sm GID2a genes, and the XbaI site was used for Pp GID2L1.
For Pp DELLAL1, the PCR product cloned in pCR4-TOPO (Invitrogen)
was digested with NotI/SmaI, blunt-ended using a DNA blunting kit
(Takara), and cloned into the SmaI site of the vector.
The full-length coding regions for the GA20 oxidases and GA3 oxidases
containing appropriate restriction sites were cloned into pMAL-c2x (New
England Biolabs), using BamHI-HindIII sites for Sm GA20oxL1, BamHI-
XhoI for Sm GA20oxL2 and Pp GA20oxL1, HindIII for Sm GA20ox, and
NcoI-BamHI for Sm GA3ox and Pp GA3oxL, to produce MBP fusion
plasmids.
Production of Recombinant Proteins
Escherichia coli BL21 (DE3) pLysS Rosseta-gami 2 (Novagen) was used
as a host strain for recombinant protein production. To produce recom-
binant Trx�His-GID1 proteins for use in the GA binding assay, 10 mL of
precultured cells was added to 500 mL of Luria-Bertani medium in a 2-liter
flask and cultured at 378C until the OD600 was 0.4 to 0.6. Induction of
recombinant proteins was performed by the addition of 0.01 mM
isopropyl-b-D-thiogalactopyranoside (IPTG) and further incubated at
168C for 18 h. Cells were harvested and resuspended with buffer A (50
mM Tris-HCl, pH 8.0, 100 mM NaCl, 10 mM imidazole, and 0.1% Triton
X-100). The cells were lysed by sonication (20 kHz, 10 s 3 20 times). Each
lysate was centrifuged at 16,000g for 30 min, and the supernatants were
mixed with 400 mL of TALON Metal Affinity Resin (Clontech) and rotated
for 2 h at 48C. The resin was washed five times with buffer A and eluted five
times with 400 mL of 500 mM imidazole in buffer A. Five portions of eluate
were gathered and desalted using a PD-10 column (GE Healthcare)
equilibrated with buffer B (20 mM Tris-HCl, pH 7.5, 0.15 M NaCl, and
2 mM 2-mercaptoethanol [2-ME]). Preparation of the Trx�His-GID1 pro-
teins for the ITC assay was identical except that buffer A was adjusted to
pH 9.4 and ITC buffer containing 20 mM PBS, pH 9.4, 100 mM NaCl, and
2.5 mM 2-ME was used in the PD-10 column desalting procedure.
For the production of recombinant GST-DELLA proteins, the cell cul-
ture and induction were performed as for Trx�His-GID1, except that the
induction temperature was 248C instead of 168C. Cells were harvested
and resuspended with buffer C (50 mM Tris-HCl, pH 8.0, 50 mM NaCl,
1 mM EDTA, and 1 mM DTT). The cells were lysed by sonication (20 kHz,
10 s 3 20 times) and 1% Triton X-100 was added. The lysates were
centrifuged at 16,000g for 30 min, and the supernatants were mixed with
2 mL of glutathione Sepharose 4B beads (GE Healthcare) and rotated for
2 h at 48C. The beads were washed five times with PBS containing 1%
Triton X-100 and eluted five times with 400 mL of 20 mM glutathione in
buffer C. Five portions of the eluate were gathered and desalted using a
PD-10 column (GE Healthcare) equilibrated with buffer B. The production
of recombinant proteins was confirmed by 7.5% SDS-PAGE. Further-
more, GID1 proteins were confirmed by immunoblot analysis using the Os
GID1 antibody as described previously (Ueguchi-Tanaka et al., 2005).
For the production of recombinant MBP-GA20 oxidases and MBP-GA3
oxidases, cell culture and induction were performed as for Trx�His-GID1,
except that 0.4 mM IPTG was used instead of 0.01 mM IPTG. Cells were
harvested and resuspended with buffer containing PBS, pH 7.4, 10 mM
2-ME, and 0.1 mg/mL lysozyme. After incubation for 30 min on ice, the
suspension was frozen and kept at �808C overnight, thawed on ice, and
sonicated (20 kHz, 10 s 3 20 times). The lysates were centrifuged at
16,000g for 30 min, and the supernatants were collected by centrifugation
and used as crude extracts for each enzyme assay.
GA Binding Assay of GID1s
For the binding assay, 3H4-H2-GA4 was used as the labeled form of GA.
GA binding was performed as reported previously (Nakajima et al., 1997)
3074 The Plant Cell
with the following modifications. One hundred microliters of purified
Trx�His-GID1s (16 mg) was incubated at room temperature with 100 mL of3H4-H2-GA4 (6 or 60 pmol), either with an excess of unlabeled GA4 (at a
final concentration of 0.125 mM) for nonspecific binding or without
unlabeled GA4 for total binding. After 20 min, 100 mL of GST-DELLA or
GST alone (16 mg) was added to the solution and incubated for another
40 min. One hundred microliters of the mixture was then fractionated on a
NAP-5 column (GE Healthcare). After discarding a void volume binding
buffer eluate (600 mL), a 200-mL fraction was collected and its radioac-
tivity was measured. The specific binding activity, which reflected the
number of replaceable GA binding sites, was calculated by subtraction of
nonspecific binding from total binding.
ITC experiments were performed with a VP-ITC microcalorimeter
(MicroCal). The instrument design and its operation have been described
in detail elsewhere (Wiseman et al., 1989). The instrument was allowed to
equilibrate overnight. Since Pp GID1L proteins are stable only at basic
pH, the ITC analysis of Trx�His-Os GID1 and Trx�His-Pp GID1L proteins
with GAs was performed at pH 9.4. GAs were prepared by dissolving in
ITC buffer (see above) and injected in 8-mL increments (final content of
500 mM) into the sample cell containing 1.4482 mL of GID1 solution (40
mM) at 308C. The stirrer was kept rotating at 400 rpm during the
experiments. The baseline was judged to have reached stability when
root mean square noise was <5 ncal/s. The heat produced in the GA-GID1
binding experiment was subtracted from the heat produced in two control
experiments. For the controls, injection of GA solution into buffer solution
and injection of buffer solution into GID1 solution were performed under
the same conditions used in the GA-GID1 binding experiment. Nonlinear
fitting of the data was performed using MicroCal Origin 7.0 (Origin-Lab).
Yeast Two-Hybrid Assay
The yeast two-hybrid assay was performed as described previously
(Ueguchi-Tanaka et al., 2005) using the BD Matchmaker Two-Hybrid
System 3 (Clontech). Vector cassettes for DNA-BD and DNA-AD were
used as negative controls, and Saccharomyces cerevisiae strain Y187
was used as the host. GAs dissolved in ethanol, or ethanol only, were
added to the culture medium at a dilution rate of 1:1000. Expression of
DNA-BD and DNA-AD fusion proteins was confirmed by immunoblot
analysis using anti-c-Myc (Clontech) and anti-HA (Sigma-Aldrich) anti-
bodies. Details of the methods used for the yeast assays can be found in
the manufacturer’s instructions (Yeast Protocols Handbook PT3024-1;
http://www.clontech.com/). Experiments were independently repeated
at least three times.
Overexpression of GID1-Like, DELLA-Like, and GID2-Like Genes
of S. moellendorffii and P. patens in Rice
proAct1-Sm GID1s, proAct1 DELLAs, and proAct1 GID2s were intro-
duced into rice gid1-3 mutant plants (Ueguchi-Tanaka et al., 2005), wild-
type T65 plants, and gid2-1 mutant plants (Sasaki et al., 2003), respectively,
by Agrobacterium tumefaciens–mediated transformation (Hiei et al.,
1994). Expression of these transgenes in rice shoots was confirmed by
RT-PCR as described below. For each transgenic plant, the length of the
second leaf sheath was measured.
RT-PCR
cDNAs prepared from various plant parts of S. moellendorffii were used
for RT-PCR analysis. Young stems (excluding microphylls), apical parts of
vegetative shoots (consisting mainly of microphylls), and a mixture of
roots and rhizophores were obtained from plants <6 cm in height. Strobili
and old stems were obtained from plants >15 cm in height. RT-PCR was
performed in a 50-mL solution containing a 2.5-mL aliquot of cDNA as the
DNA template, 0.2 mM gene-specific primers (see Supplemental Table 4
online), 10 mM deoxynucleotide triphosphates, 1 unit of ExTaq DNA
polymerase (Takara), and reaction buffer. Amplifications of Sm 6PGD and
Os ACT1 cDNAs were performed as controls for S. moellendorffii and
rice, respectively, to ensure that equal amounts of cDNA were added to
each PCR. The reaction included an initial 5-min denaturation at 948C,
followed by 25 to 31 cycles of PCR (948C for 30 s, 568C for 30 s, and 728C
for 30 s), and a final 10-min extension at 728C. The number of cycles used
for amplification with each primer pair was adjusted to be in the linear
range. All RT-PCR data are representative of at least three independent
experiments.
Antibody Production
GST-Sm DELLA1 recombinant protein was produced in E. coli and
subsequently purified by glutathione beads by the same method de-
scribed above. This protein was used for the production of antibodies
after exchanging the buffer for PBS using a PD-10 desalting column (GE
Healthcare). Sm DELLA1 polyclonal antibody was produced by immuni-
zation of a rabbit (Operon Biotechnologies).
Immunoblot Analysis of the Sm DELLA1 Protein
Crude protein extracts of young S. moellendorffii shoots and seedlings of
transgenic rice were prepared by grinding with liquid nitrogen in the
presence of sea sand (425 to 850 mm; Wako Pure Chemical) followed by
an equal volume of 23 sample buffer (13 sample buffer is 67.5 mM Tris-
HCl, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromphenol blue, and 0.1 M
DTT) and boiling for 5 min. Protein samples were separated by 7.5% SDS-
PAGE and transferred to a Hybond enhanced chemiluminescence nitro-
cellulose membrane (GE Healthcare) by semidry blotting. The blots were
treated with 5% skim milk in TBST (0.1% Tween 20 in 2 mM Tris-HCl, pH
7.6, and 13.7 mM NaCl) for 1 h and subsequently incubated with anti-Os
SLR1 antiserum (Itoh et al., 2002) or anti-Sm DELLA1 antiserum for 1 h.
Blots were washed three times with TBST for 15 min each. Goat anti-
rabbit IgG horseradish peroxidase–conjugated secondary antibody was
incubated for 45 min, and blots were washed following the same proce-
dure described above. All reactions were conducted at room tempera-
ture. Detection of peroxidase activity was performed according to the
instruction manual from Pierce.
GA Treatment of S. moellendorffii
Young S. moellendorffii plants were grown in pots to no more than 5 cm in
height. Pots were submerged in water containing 0.1% ethanol, up to
about one-third from the bottom, either with or without 10�4 M GA4. To
ensure complete infiltration of GA4, plants were sprayed with 10�4 M GA4
solution (containing 0.1% ethanol and 0.02% Tween 20) at 24 and 72 h
after the start of submergence. One hour after the second spray, total
RNA was isolated from young shoots.
For the DELLA protein disappearance experiment, two different con-
ditions were used, and both showed the disappearance of Sm DELLA1
protein in GA4-treated plants. Plants were either treated as described
above for 1 month or dipped completely in a solution containing 10�4 M
GA4, 0.1% ethanol, and 0.02% Tween 20 with continuous shaking using a
seesaw shaker for 12 h. Plants not treated with GA solution were used as
negative controls. Crude protein was extracted as described above.
To observe the effects of GA4 and uniconazole on plant growth, bulbils
of S. moellendorffii were treated with either GA4 or uniconazole (each
containing 0.01% ethanol) at concentrations of 10�5 M and 10�6 M,
respectively. Ethanol solution (0.01%) was used as a negative control.
GA Perception in Nonseed Plants 3075
Pots were filled with vermiculite, and Rockwool was placed on top. Pots
were then submerged in each solution up to about one-third from the
bottom, and the same solution was used to soak the Rockwool. Bulbils of
S. moellendorffii were placed on the surface of the Rockwool to germi-
nate. Growth effects were assessed by measuring the length between the
second and third small leaves.
Enzyme Assay
Crude extracts were incubated at 308C with GA substrates (1 mg each) in
200 mL of 100 mM Tris-HCl, pH 7.9, 4 mM ascorbic acid, 4 mM
2-oxoglutaric acid, 0.5 mM FeSO4, 4 mM DTT, 2 mg/mL BSA, and
1 mg/mL catalase. The reactions were stopped after overnight incubation
by adding 25 mL of acetic acid. The solution was passed through a C18-
HD high-performance extraction disk cartridge (1 mL; Empore). After the
column was washed with 3 mL of water, substances retained on the
column were eluted with 500 mL of methanol. The methanol eluate was
evaporated with dry N2 gas. After trimethylsilyl (TMSi) ester–TMSi ether
derivatization with N-methyl-N-trimethylsilyl-trifluoroacetamide, products
were analyzed by full-scan GC-MS, and identical ion peaks were com-
pared with standard GAs (GA12, GA24, GA9, GA4, GA53, GA44, GA19, GA20,
and GA1) and published data (GA15) (Gaskin and MacMillan 1992).
GC-MS Analysis of the Enzyme Assay
Full-scan GC-MS analysis of GAs was performed using a mass spec-
trometer (JMS-K9; JEOL) connected to a gas chromatograph (6890N;
Agilent Technology). The trimethylsilylated derivatives (TMSi ester–TMSi
ether) were injected (2508C) into an HP-5 MS column (0.32 mm i.d. 3 30
m, 0.25 mm film thickness; Agilent Technology). The column temperature
was kept at 1008C for 2 min, then increased at a rate of 308C/min to 2608C
and held for 1 min, and then increased at a rate of 308C/min to 3008C. The
flow rate of the carrier He gas was 1.5 mL/min, and mass spectra were
acquired by scanning from m/z 50 to 750 at 70 eV.
Endogenous GA Analysis
Tissue samples (;1 g fresh weight) were ground to a fine powder under
liquid nitrogen and then soaked in 5 mL of extraction solvent (methanol:
formic acid:water, 15:1:4). For the internal standards, 5 pmol of stable
isotope GAs (2H2-GA1, 2H2-GA4, 2H2-GA9, 2H2-GA19, 2H2-GA20, and 2H2-
GA24) was added to the extract. To remove interfering compounds, the
extract was first passed through a Sep-Pak Vac tC18 cartridge (Waters).
The pass fraction was dried and reconstituted with 1 M formic acid.
The pass fraction was then further fractionated using an MCX column
(Waters), and the eluate was recovered by methanol in the solid-phase
extraction (Dobrev and Kaminek, 2002). The fraction was dried and
reconstituted with water. Subsequently, compounds contained in a
fraction were further purified using DEAE-Sephadex (GE Healthcare).
GAs were eluted from the DEAE-Sephadex with 1% acetic acid.
The GA contents were measured using a liquid chromatography–mass
spectrometry system (UPLC/Quattro Ultima Pt; Waters) with an ODS
column (AQUITY UPLC BEH C18, 1.7 mm, 2.1 3 50 mm; Waters) at a flow
rate of 0.25 mL/min. The gradients of solvent A (0.05% formic acid) and
solvent B (0.05% formic acid in acetonitrile) were applied at a flow rate of
0.25 mL/min according to the following profile: 0 min, 99% A þ 1% B; 13
min, 66% A þ 34% B; 15 min, 99% A þ 1% B. Quantification was
performed in the multiple reaction monitoring mode. The mother and
daughter ions for the detection of each GA type were as follows: m/z 345
and 259 for GA1, m/z 331.2 and 243 for GA4, m/z 315.2 and 271 for GA9,
m/z 361.1 and 243 for GA19, m/z 331.2 and 287 for GA20, and m/z 345.2
and 257 for GA24, respectively. Cone voltage and collision energy were 85
V and 18 eV for GA1, 88 V and 17 eV for GA4, 95 V and 18 eV for GA9, 85 V
and 19 eV for GA19, 85 V and 19 eV for GA20, and 80 V and 23 eV for GA24,
respectively. Capillary voltage was 3.12 kV. For the identification of GA4,
mass spectra were obtained by daughter ion scanning of negative ions
from m/z 50 to 400 with collision energy at 25 eV and compared with those
of a GA4 standard.
GAs
3H4-H2-GA4 was custom ordered from DuPont–New England Nuclear.
GA4, GA9, and 3-epi-GA4 used in the yeast two-hybrid assay and GAs
used in the enzyme assay were purchased from L.N. Mander (Australian
National University). 2H2-GA1, 2H2-GA4, 2H2-GA9, 2H2-GA19, 2H2-GA20,
and 2H2-GA24 were purchased from OlChemIm.
Accession Numbers
GenBank/EMBL accession numbers and Arabidopsis Genome Initiative
locus identifiers for the genes mentioned in this article are as follows: Os
GID1 (Q6L545), Os HSL1 (ABA92266), Os SLR1 (BAE96289), Os SLRL1
(AAR31213), Os SLRL2 (AAT69589), Os SCR (ABA91267), Os GID2
(Q7XAK4), Os GA20ox1 (AAP21386), Os GA20ox2 (Q8RVF5), Os
GA20ox3 (BAB90378), Os GA20ox4 (AAT44252), Os GA3ox1 (AAT77356),
Os GA3ox2 (BAB17075), Os GA2ox1 (AAV43914), Os GA2ox2
(BAD53498), Os GA2ox3 (AAU03107), Os GA2ox4 (AAU03107), Os
ACT1 (CT831215), Sr 6PGD (AB086022), At GID1a (AT3G05120), At
GID1b (AT3G63010), At GID1c (AT5G27320), At HSL1 (AT5G23530), At
GAI (AT1G14920), At RGA (AT2G01570), At SLY1 (AT4G24210), At SNZ
(AT5G48170), At SCL3 (AT1G50420), At LAS (AT1G55580), At PAT
(AT5G48150), At SCL1 (AT2G29060), At GA20ox1 (AT4G25420), At
GA20ox2 (AT5G51810), At GA20ox3 (AT5G07200), At GA20ox4
(AT1G60980), At GA3ox1 (AT1G15550), At GA3ox2 (AT1G80340), At
GA3ox3 (AT1G80330), At GA3ox4 (AT4G21690), At GA2ox1
(AT1G78440), At GA2ox2 (AT1G30040), At GA2ox3 (AT2G34555), At
GA2ox4 (AT1G47990), At GA2ox5 (AT1G02400), Pr MC3 (AAD04946), Zm
dwarf plant9 (ABI84225), Zm DWARF8 (Q9ST48), Ta GID2L (ABK79908),
and Medicago truncatula DELLA (ABE77443). Sequence data obtained
from the Computational Biology and Functional Genomics Laboratory
(http://compbio.dfci.harvard.edu/tgi/plant.html) can be found under the
following accession numbers: pine GID1L1 (TC73510), pine GID1L2
(TC57783), pine GID1L3 (TC76887) pine DELLA1 (TC59241), and pine
DELLAL1 (TC60068).
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. Effects of Various GAs on the Pp GID1L–Pp
DELLAL Interactions in Yeast Cells.
Supplemental Figure 2. Length of the Second Leaf Sheath of
Transgenic Rice Plants.
Supplemental Figure 3. Phylogenetic Analysis of GA Oxidase
Homologs Using the Maximum Likelihood Method.
Supplemental Figure 4. GC-MS Spectrum of S. moellendorffii
Endogeneous GA4.
Supplemental Figure 5. Phylogenetic Analysis of GID1 Homologs
Using the NJ Method.
Supplemental Figure 6. Phylogenetic Analysis of DELLA Homologs
Using the NJ Method.
Supplemental Figure 7. Phylogenetic Analysis of GID2 Homologs
Using the NJ Method.
Supplemental Figure 8. Phylogenetic Analysis of GA Oxidase
Homologs Using the NJ Method.
3076 The Plant Cell
Supplemental Table 1. Endogenous GA Content in P. patens.
Supplemental Table 2. In Vitro Activity of GA20 Oxidase and GA3
Oxidase Homologs in P. patens.
Supplemental Table 3. Inclusion Limits and e-Value Thresholds Used
for Phylogenetic Analyses Using the NJ Method.
Supplemental Table 4. Primers Used.
Supplemental Data Set 1. Text File of Amino Acid Alignments
Presented in Figure 2A.
Supplemental Data Set 2. Text File of Amino Acid Alignments
Presented in Figure 2B.
Supplemental Data Set 3. Text File of Amino Acid Alignments
Presented in Figure 3A.
Supplemental Data Set 4. Text File of Amino Acid Alignments
Presented in Figure 3B.
Supplemental Data Set 5. Text File of Amino Acid Alignments
Presented in Figure 3C.
Supplemental Data Set 6. Text File of Amino Acid Alignments
Presented in Figure 4A.
Supplemental Data Set 7. Text File of Amino Acid Alignments
Presented in Figure 4B.
Supplemental Data Set 8. Text File of Amino Acid Alignments Used
to Construct the Phylogenetic Tree Presented in Supplemental Figure
3 Online.
Supplemental Data Set 9. Text File of Amino Acid Alignments Used
to Construct the Phylogenetic Tree Presented in Supplemental Figure
5 Online.
Supplemental Data Set 10. Text File of Amino Acid Alignments Used
to Construct the Phylogenetic Tree Presented in Supplemental Figure
6 Online.
Supplemental Data Set 11. Text File of Amino Acid Alignments Used
to Construct the Phylogenetic Tree Presented in Supplemental Figure
7 Online.
Supplemental Data Set 12. Text File of Amino Acid Alignments Used
to Construct the Phylogenetic Tree Presented in Supplemental Figure
8 Online.
ACKNOWLEDGMENTS
We thank Hitomi Kihara, Mayuko Kawamura, Kazuko Kigaku, Kozue
Ohmae, and Hiroko Ohmiya for their technical assistance, Naomi
Sumikawa for cultivation of S. moellendorffii, and Yuji Hiwatashi for P.
patens RNA. This work was supported by the Ministry of Education,
Culture, Sports, Science, and Technology of Japan (M.M., M.N., M.U.-T.,
T.T., T.N., and M.H.), the Ministry of Agriculture, Forestry, and Fisheries
of Japan (Green Technology Project IP1003; M.M. and M.A), and by
Exploratory Research for Advanced Technology, Japan Science and
Technology (T.N. and M.H.).
Received May 8, 2007; revised September 20, 2007; accepted October 6,
2007; published October 26, 2007.
REFERENCES
Adachi, J., and Hasegawa, M. (1996). MOLPHY version 2.3: Programs
for molecular phylogenetics based on maximum likelihood. Comput.
Sci. Monogr. 28: 1–150.
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z.,
Miller, W., and Lipman, D.J. (1997). Gapped BLAST and PSI-BLAST:
A new generation of protein database search programs. Nucleic Acids
Res. 25: 3389–3402.
Ashton, N.W., and Cove, D.J. (1977). The isolation and preliminary
characterisation of auxotrophic and analogue resistant mutants in the
moss Physcomitrella patens. Mol. Gen. Genet. 154: 87–95.
Banks, J.A. (1999). Gametophyte development in ferns. Annu. Rev.
Plant Physiol. Plant Mol. Biol. 50: 163–186.
Bolle, C. (2004). The role of GRAS proteins in plant signal transduction
and development. Planta 218: 683–692.
Chandler, P., Marion-Poll, A., Ellis, M., and Gubler, F. (2002). Mutants
at the Slender1 locus of barley cv Himalaya. Molecular and physio-
logical characterization. Plant Physiol. 129: 181–190.
Chiang, H., Hwang, I., and Goodman, H. (1995). Isolation of the
Arabidopsis GA4 locus. Plant Cell 7: 195–201.
Cove, D., Benzanilla, M., Harries, P., and Quatrano, R. (2006).
Mosses as model systems for the study of metabolism and develop-
ment. Annu. Rev. Plant Biol. 57: 497–520.
Davis, P. J. (2004). Plant Hormones: Biosynthesis, Signal Transduction,
Action! (Dordrecht, The Netherlands: Kluwer Academic Publishers).
Dharmasiri, S., and Estelle, M. (2002). The role of regulated protein
degradation in auxin response. Plant Mol. Biol. 49: 401–409.
Dill, A., Jung, H., and Sun, T. (2001). The DELLA motif is essential for
gibberellin-induced degradation of RGA. Proc. Natl. Acad. Sci. USA
98: 14162–14167.
Dobrev, P.I., and Kaminek, M. (2002). Fast and efficient separation of
cytokinins from auxin and abscisic acid and their purification using
mixed-mode solid-phase extraction. J. Chromatogr. A 950: 21–29.
Ergun, N., Topcuoglu, Sx.F., and Yildiz, A. (2002). Auxin (indole-3-
acetic acid), gibberellic acid (GA3), abscisic acid (ABA) and cytokinin
(zeatin) production by some species of mosses and lichens. Turk. J.
Bot. 26: 13–18.
Fleet, C.M., Yamaguchi, S., Hanada, A., Kawaide, H., David, C.J.,
Kamiya, Y., and Sun, T.P. (2003). Overexpression of AtCPS and
AtKS in Arabidopsis confers increased ent-kaurene production but no
increase in bioactive gibberellins. Plant Physiol. 132: 830–839.
Gaskin, P., and MacMillan, J. (1992). GC-MS of the Gibberellins and
Related Compounds: Methodology and a Library of Spectra. (Bristol,
UK: Cantocks Enterprises).
Gifford, E.M., and Foster, A.S. (1989). Morphology and Evolution of
Vascular Plants (New York: Freeman and Company).
Gomi, K., Sasaki, A., Itoh, H., Ueguchi-Tanaka, M., Ashikari, M.,
Kitano, H., and Matsuoka, M. (2004). GID2, an F-box subunit of the
SCF E3 complex, specifically interacts with phosphorylated SLR1
protein and regulates the gibberellin-dependent degradation of SLR1
in rice. Plant J. 37: 626–634.
Griffiths, J., Murase, K., Rieu, I., Zentella, R., Zhang, Z.L., Powers,
S.J., Gong, F., Phillips, A.L., Hedden, P., Sun, T.P., and Thomas,
S.G. (2006). Genetic characterization and functional analysis of the
GID1 gibberellin receptors in Arabidopsis. Plant Cell 18: 3399–3414.
Gubler, F., Chandler, P.M., White, R.G., Llewellyn, D.J., and
Jacobsen, J.V. (2002). Gibberellin signaling in barley aleurone cells.
Control of SLN1 and GAMYB expression. Plant Physiol. 129: 191–200.
Hasebe, M., Wen, C.-K., Kato, M., and Banks, J.A. (1998). Charac-
terization of MADS homeotic genes in the fern Ceratopteris richardii.
Proc. Natl. Acad. Sci. USA 95: 6222–6227.
Hayashi, K., Kawaide, H., Notomi, M., Sakigi, Y., Matsuo, A., and
Nozaki, H. (2006). Identification and functional analysis of bifunctional
ent-kaurene synthase from the moss Physcomitrella patens. FEBS
Lett. 580: 6175–6181.
Hiei, Y., Ohta, S., Komari, T., and Kumashiro, T. (1994). Efficient
transformation of rice (Oryza sativa L.) mediated by Agrobacterium
GA Perception in Nonseed Plants 3077
and sequence analysis of the boundaries of the T-DNA. Plant J. 6:
271–282.
Ikeda, A., Ueguchi-Tanaka, M., Sonoda, Y., Kitano, H., Koshioka, M.,
Futsuhara, Y., Matsuoka, M., and Yamaguchi, J. (2001). slender
rice, a constitutive gibberellin response mutant, is caused by a null
mutation of the SLR1 gene, an ortholog of the height-regulating gene
GAI/RGA/RHT/D8. Plant Cell 13: 999–1010.
Itoh, H., Shimada, A., Ueguchi-Tanaka, M., Kamiya, N., Hasegawa,
Y., Ashikari, M., and Matsuoka, M. (2005). Overexpression of a
GRAS protein lacking the DELLA domain confers altered gibberellin
responses in rice. Plant J. 44: 669–679.
Itoh, H., Ueguchi-Tanaka, M., Sato, Y., Ashikari, M., and Matsuoka,
M. (2002). The gibberellin signaling pathway is regulated by the
appearance and disappearance of SLENDER RICE1 in nuclei. Plant
Cell 14: 57–70.
Jones, D.T., Taylor, W.R., and Thornton, J.M. (1992). The rapid gen-
eration of mutation data matrices from protein sequences. Comput.
Appl. Biosci. 8: 275–282.
Kamiya, N., Itoh, J., Morikami, A., Nagato, Y., and Matsuoka, M.
(2003). The SCARECROW gene’s role in asymmetric cell divisions in
rice plants. Plant J. 36: 45–54.
Kaneko, M., Itoh, H., Inukai, Y., Sakamoto, T., Ueguchi-Tanaka, M.,
Ashikari, M., and Matsuoka, M. (2003). Where do gibberellin bio-
synthesis and gibberellin signaling occur in rice plants? Plant J. 35:
104–115.
Kato, J., Purves, W.K., and Phinney, B.O. (1962). Gibberellin-like
substances in plants. Nature 196: 687–688.
Katoh, K., Kuma, K., Toh, H., and Miyata, T. (2005). MAFFT version 5:
Improvement in accuracy of multiple sequence alignment. Nucleic
Acids Res. 33: 511–518.
Kenrick, P., and Crane, P.R. (1997). The origin and early evolution of
plants on land. Nature 389: 33–39.
Knight, C.D., Sehgal, A., Atwal, K., Wallace, J.C., Cove, D.J., Coates,
D., Quatrano, R.S., Bahadur, S., Stockley, P.G., and Cuming, A.C.
(1995). Molecular responses to abscisic acid and stress are con-
served between moss and cereals. Plant Cell 7: 499–506.
Kobayashi, M., Sakurai, A., Saka, H., and Takahashi, N. (1989).
Quantitative analysis of endogenous gibberellins in normal and dwarf
cultivars of rice. Plant Cell Physiol. 30: 963–969.
Lange, T., Kappler, J., Fischer, A., Frisse, A., Padeffke, T.,
Schmidtke, S., and Lange, M.J. (2005). Gibberellin biosynthesis in
developing pumpkin seedlings. Plant Physiol. 139: 213–223.
MacMillan, J. (2002). Occurrence of gibberellins in vascular plants,
fungi, and bacteria. J. Plant Growth Regul. 20: 387–442.
McElroy, D., Zhang, W., Cao, J., and Wu, R. (1990). Isolation of an
efficient actin promoter for use in rice transformation. Plant Cell 2:
163–171.
McGinnis, K.M., Thomas, S.G., Soule, J.D., Strader, L.C., Zale, J.M.,
Sun, T., and Steber, C.M. (2003). The Arabidopsis SLEEPY1 gene
encodes a putative F-box subunit of an SCF E3 ubiquitin ligase. Plant
Cell 15: 1120–1130.
Menendez, V., Revilla, M., Bernard, P., Gotor, V., and Fernandez, H.
(2006). Gibberellins and antheridiogen on sex in Blechnum spicant L.
Plant Cell Rep. 25: 1104–1110.
Nakajima, M., et al. (2006). Identification and characterization of
Arabidopsis gibberellin receptors. Plant J. 46: 880–889.
Nakajima, M., Takita, K., Wada, H., Mihara, K., Hasegawa, M.,
Yamaguchi, I., and Murofushi, N. (1997). Partial purification and
characterization of a gibberellin-binding protein from seedlings of
Azukia angularis. Biochem. Biophys. Res. Commun. 241: 782–786.
Nishiyama, T. (2007). Evolutionary developmental biology of nonflow-
ering land plants. Int. J. Plant Sci. 168: 37–47.
Nishiyama, T., Hiwatashi, Y., Sakakibara, K., Kato, M., and Hasebe,
M. (2000). Tagged mutagenesis and gene-trap in the moss, Physco-
mitrella patens by shuttle mutagenesis. DNA Res. 7: 9–17.
Olszewski, N., Sun, T., and Gubler, F. (2002). Gibberellin signaling:
Biosynthesis, catabolism, and response pathways. Plant Cell 14
(suppl.), S61–S80.
Peng, J., Carol, P., Richards, D.E., King, K.E., Cowling, R.J., Murphy,
G.P., and Harberd, N.P. (1997). The Arabidopsis GAI gene defines a
signaling pathway that negatively regulates gibberellin responses.
Genes Dev. 11: 3194–3205.
Phillips, A.L., Ward, D.A., Uknes, S., Appleford, N.E.J., Lange, T.,
Huttly, A.K., Gaskin, P., Graebe, J.E., and Hedden, P. (1995).
Isolation and expression of three gibberellin 20-oxidase cDNA clones
from Arabidopsis. Plant Physiol. 108: 1049–1057.
Radley, M. (1961). Gibberellin-like substances in plants. Nature 191:
684–685.
Saitou, N., and Nei, M. (1987). The neighbor-joining method: A
new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:
406–425.
Sasaki, A., Itoh, H., Gomi, K., Ueguchi-Tanaka, M., Ishiyama, K.,
Kobayashi, M., Jeong, D., An, G., Kitano, H., Ashikari, M., and
Matsuoka, M. (2003). Accumulation of phosphorylated repressor for
gibberellin signaling in an F-box mutant. Science 299: 1896–1898.
Schaefer, D.G., and Zryd, J. (1997). Efficient gene targeting in the moss
Physcomitrella patens. Plant J. 11: 1195–1206.
Sentoku, N., Sato, Y., and Matsuoka, M. (2000). Overexpression of
rice OSH genes induces ectopic shoots on leaf sheaths of transgenic
rice plants. Dev. Biol. 220: 358–364.
Silverstone, A.L., Jung, H., Dill, A., Kawaide, H., Kamiya, Y., and Sun,
T. (2001). Repressing a repressor: Gibberellin-induced rapid reduction
of the RGA protein in Arabidopsis. Plant Cell 13: 1555–1565.
Strader, L.C., Ritchie, S., Soule, J.D., McGinnis, K.M., and Steber,
C.M. (2004). Recessive-interfering mutations in the gibberellin signal-
ing gene SLEEPY1 are rescued by overexpression of its homologue,
SNEEZY. Proc. Natl. Acad. Sci. USA 101: 12771–12776.
Tanabe, Y., Uchida, M., Hasebe, M., and Ito, M. (2003). Character-
ization of the Selaginella remotifolia MADS-box gene. J. Plant Res.
116: 71–75.
Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994). CLUSTAL W:
Improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and
weight matrix choice. Nucleic Acids Res. 22: 4673–4680.
Ueguchi-Tanaka, M., Ashikari, M., Nakajima, M., Itoh, H., Katoh, E.,
Kobayashi, M., Chow, T., Hsing, Y., Kitano, H., Yamaguchi, I., and
Matsuoka, M. (2005). GIBBERELLIN INSENSITIVE DWARF1 encodes
a soluble receptor for gibberellin. Nature 437: 693–698.
Ueguchi-Tanaka, M., Nakajima, M., Ashikari, M., and Matsuoka, M.
(2007a). Gibberellin receptor and its role in gibberellin signaling in
plants. Annu. Rev. Plant Biol. 58: 183–198.
Ueguchi-Tanaka, M., Nakajima, M., Katoh, E., Ohmiya, H., Asano,
K., Saji, S., Hongyu, X., Ashikari, M., Kitano, H., Yamaguchi, I., and
Matsuoka, M. (2007b). Molecular interactions of a soluble gibberellin
receptor, GID1, with a rice DELLA protein, SLR1, and gibberellin. Plant
Cell 19: 2140–2155.
Weng, J.K., Tanurdzic, M., and Chapple, C. (2005). Functional anal-
ysis and comparative genomics of expressed sequence tags from the
lycophyte Selaginella moellendorffii. BMC Plant Biol. 6: 85.
Willige, B.C., Ghosh, S., Nill, C., Zourelidou, M., Dohmann, E.M.,
Maier, A., and Schwechheimer, C. (2007). The DELLA domain of GA
INSENSITIVE mediates the interaction with the GA INSENSITIVE
DWARF1A gibberellin receptor of Arabidopsis. Plant Cell 19: 1209–1220.
Wiseman, T., Williston, S., Brandts, J.F., and Lin, L.N. (1989). Rapid
measurement of binding constants and heats of binding using a new
titration calorimeter. Anal. Biochem. 179: 131–137.
3078 The Plant Cell
Yamane, H., Fujioka, S., Spray, C.R., Phinney, B.O., Macmillan, J.M.,
Gaskin, P., and Takahashi, N. (1988). Endogenous gibberellins from
sporophytes of two tree ferns, Cibotium glaucum and Dicksonia
antarctica. Plant Physiol. 86: 857–862.
Yamane, H., Yamaguchi, I., Kobayashi, M., Takahashi, M., Sato, Y.,
Takahashi, N., Iwatsuki, K., Phinney, B.O., Spray, C.R., Gaskin, P.,
and Macmillan, J. (1985). Identification of ten gibberellins from spo-
rophytes of the tree fern, Cyathea australis. Plant Physiol. 78: 899–903.
Yamauchi, T., Oyama, N., Yamane, H., Murofushi, N., Schraudolf, H.,
Pour, M., Furber, M., and Mander, L.N. (1996). Identification of
antheridiogens in Lygodium circinnatum and Lygodium flexuosum.
Plant Physiol. 111: 741–745.
Yasumura, Y., Crumpton-Taylor, M., Fuentes, S., and Harberd, N.
(2007). Step-by-step acquisition of the gibberellin-DELLA growth-
regulatory mechanism during land-plant evolution. Curr. Biol. 17:
1225–1230.
GA Perception in Nonseed Plants 3079
DOI 10.1105/tpc.107.051524; originally published online October 26, 2007; 2007;19;3058-3079Plant Cell
Ashikari, Hidemi Kitano, Miyako Ueguchi-Tanaka and Makoto MatsuokaKojima, Etsuko Katoh, Hongyu Xiang, Takako Tanahashi, Mitsuyasu Hasebe, Jo Ann Banks, Motoyuki
Ko Hirano, Masatoshi Nakajima, Kenji Asano, Tomoaki Nishiyama, Hitoshi Sakakibara, MikikoPhyscomitrella patens but Not in the Bryophyte moellendorffii
SelaginellaThe GID1-Mediated Gibberellin Perception Mechanism Is Conserved in the Lycophyte
This information is current as of May 27, 2021
Supplemental Data /content/suppl/2007/10/26/tpc.107.051524.DC1.html
References /content/19/10/3058.full.html#ref-list-1
This article cites 61 articles, 24 of which can be accessed free at:
Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X
eTOCs http://www.plantcell.org/cgi/alerts/ctmain
Sign up for eTOCs at:
CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain
Sign up for CiteTrack Alerts at:
Subscription Information http://www.aspb.org/publications/subscriptions.cfm
is available at:Plant Physiology and The Plant CellSubscription Information for
ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists