Erect leaves caused by brassinosteroid deficiency increase biomass production and grain yield in rice.

Abstract

New cultivars with very erect leaves, which increase light capture for photosynthesis and nitrogen storage for grain filling, may have increased grain yields. Here we show that the erect leaf phenotype of a rice brassinosteroid-deficient mutant, osdwarf4-1, is associated with enhanced grain yields under conditions of dense planting, even without extra fertilizer. Molecular and biochemical studies reveal that two different cytochrome P450s, CYP90B2/OsDWARF4 and CYP724B1/D11, function redundantly in C-22 hydroxylation, the rate-limiting step of brassinosteroid biosynthesis. Therefore, despite the central role of brassinosteroids in plant growth and development, mutation of OsDWARF4 alone causes only limited defects in brassinosteroid biosynthesis and plant morphology. These results suggest that regulated genetic modulation of brassinosteroid biosynthesis can improve crops without the negative environmental effects of fertilizers.

Full text

Page 1

Erect leaves caused by brassinosteroid deficiency
increase biomass production and grain yield in rice
Tomoaki Sakamoto1, Yoichi Morinaka2, Toshiyuki Ohnishi3, Hidehiko Sunohara2, Shozo Fujioka4,
Miyako Ueguchi-Tanaka2, Masaharu Mizutani3, Kanzo Sakata3, Suguru Takatsuto5, Shigeo Yoshida4,
Hiroshi Tanaka6,7, Hidemi Kitano2& Makoto Matsuoka2
New cultivars with very erect leaves, which increase light
capture for photosynthesis and nitrogen storage for grain
filling, may have increased grain yields1. Here we show that
the erect leaf phenotype of a rice brassinosteroid–deficient
mutant, osdwarf4-1, is associated with enhanced grain yields
under conditions of dense planting, even without extra
fertilizer. Molecular and biochemical studies reveal that two
different cytochrome P450s, CYP90B2/OsDWARF4 and
CYP724B1/D11, function redundantly in C-22 hydroxylation,
the rate-limiting step of brassinosteroid biosynthesis.
Therefore, despite the central role of brassinosteroids in
plant growth and development, mutation of OsDWARF4 alone
causes only limited defects in brassinosteroid biosynthesis and
plant morphology. These results suggest that regulated genetic
modulation of brassinosteroid biosynthesis can improve crops
without the negative environmental effects of fertilizers.
Plant growth regulators are excellent targets for genetic modification
to enhance crop yields. Modulation of the abundance of and the
capacity to respond to a handful of growth regulators can potentially
affect agronomically important traits such as plant height, root
architecture and leaf arrangement. For instance, high-yielding, semi-
dwarf plant cultivars, produced by modulation of the synthesis of or
sensitivity to the growth regulator gibberellin, are more responsive to
fertilizer and enabled the green revolution to occur2–5. A complemen-
tary strategy for improving plant productivity is genetic manipulation
to generate leaves that are more erect. Erect leaves enhance light
capture for photosynthesis, serve as nitrogen reservoirs for grain filling
and enable more dense plantings with a higher leaf area index1, all of
which increase yield.
Brassinosteroids influence both plant height and leaf erectness in
rice6. Arabidopsis thaliana BRASSINOSTEROID INSENSITIVE1 (bri1)
mutants, which are deficient in the brassinosteroid receptor, have a
severe dwarf phenotype7. The rice mutants d61-1 and d61-2, which are
weak mutant alleles of OsBRI1, a homolog of A. thaliana BRI1, have
both a semi-dwarf phenotype and more erect leaves8. In rice, an erect
leaf phenotype has also been associated with loss-of-function, brassi-
nosteroid-deficient mutants such as brassinosteroid-deficient dwarf1
(brd1), which is defective in OsDWARF (the homolog of tomato
CYP85A1/DWARF9), and ebisu dwarf (d2), which is defective in
CYP90D2/D2 (ref. 10). However, the grain yield of d61, brd1 and d2
mutants is decreased because of morphological alterations in their
reproductive development.
To obtain erect leaf mutants with enhanced grain yield, we screened
a collection of rice mutants produced by g-ray irradiation, retro-
transposon (Tos17) mutagenesis or chemical mutagenesis. After field
experiments involving 34 candidates from the initial mutant screen,
we selected one line, osdwarf4-1, which showed erect leaves without
abnormal leaf, flower or grain morphology. The affected gene,
OsDWARF4, was identified by characterizing the site of insertion of
the Tos17 retrotransposon.
OsDWARF4 is a homolog of A. thaliana DWARF4, which encodes a
cytochrome P450, CYP90B1 (Fig. 1a). CYP90B1/DWARF4 catalyzes
C-22 hydroxylation, the rate-limiting step of brassinosteroid biosyn-
thesis11. Only one DWARF4 homolog was identified in the rice
genome. OsDWARF4 is located on the short arm of chromosome 3
(35 cM). Its predicted open reading frame consists of eight exons, and
encodes a protein of 502 amino acids (Fig. 1b,c). OsDWARF4 is most
closely related to DWARF4 (66.3% amino-acid identity; Fig. 1a). The
structure of OsDWARF4 is similar to that of DWARF4 throughout its
length: five domains found in cytochrome P450s—proline rich, A (also
referred to as dioxygen binding), B (steroid binding), C and heme
binding—are highly conserved (Fig. 1c). In wild-type rice, OsDWARF4
transcripts are most abundant in leaf blades and roots, but are also
found in all other plant parts tested (Fig. 1d). Transcript levels were
decreased by exogenous brassinolide treatment, and increased in
brassinosteroid-insensitive d61-3 and brassinosteroid-deficient brd1-1
(Fig. 1e). These results suggest that OsDWARF4 expression is
feedback regulated, as is that of A. thaliana brassinosteroid-
biosynthesis genes12.
Received 26 September; accepted 8 November; published online 20 December 2005; doi:10.1038/nbt1173
1Field Production Science Center, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Midoricho, Nishi-Tokyo, Tokyo 188-0002, Japan.
2Bioscience and Biotechnology Center, Nagoya University, Furocho, Chikusa, Nagoya, Aichi 464-8601, Japan.3Institute for Chemical Research, Kyoto University,
Gokasho, Uji, Kyoto 611-0011, Japan.4RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan.5Department of Chemistry, Joetsu University of Education,
1 Yamayashikicho, Joetsu, Niigata 943-8512, Japan.6National Institute of Agrobiological Sciences, 2-1-2 Kannondai, Tsukuba, Ibaraki 305-8602, Japan.
7Present address: Hokuriku Research Center, National Agricultural Research Center, 1-2-1 Inada, Joetsu, Niigata 943-0193, Japan. Correspondence should
be addressed to T.S. (orchardist@fm.a.u-tokyo.ac.jp).
NATURE BIOTECHNOLOGY
VOLUME 24NUMBER 1JANUARY 2006105
L E T T E R S
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology

Page 2

In osdwarf4-1, Tos17 insertion into exon 6 causes a premature
termination in domain C (arrowhead in Fig. 1b,c). Because a
premature stop codon before the heme-binding domain caused loss-
of-function of A. thaliana DWARF4 (ref. 11), we consider osdwarf4-1
to be a null allele. Unexpectedly, osdwarf4-1 plants show only weak
morphological phenotypes, such as a slightly dwarfed stature and
more erect leaves (Fig. 2a,b). Both of these phenotypes were com-
plemented by the introduction of an entire OsDWARF4 gene (data not
shown). These results seem inconsistent with the fact that rice d61-3
and brd1-1, which are loss-of-function mutants of OsBRI1 and
OsDWARF respectively, display severe dwarfing and have malformed
leaves (for example, see brd1-1 in Fig. 2a), and suggest functional
redundancy in brassinosteroid biosynthesis in rice.
OsDWARF4L1 was identified as an OsDWARF4-like gene by in silico
screening of rice DNA databases. OsDWARF4L1 is located on the long
arm of chromosome 4 (70.9 cM). A recent study demonstrated that a
rice semi-dwarf and short-grain mutant, d11, has a loss-of-function
mutation in the OsDWARF4L1 gene13. OsDWARF4L1/D11 shows the
highest homology with A. thaliana CYP724A1, and is closely related to
rice OsDWARF4 and A. thaliana DWARF4 (47.4%, 37.7% and 41.7%
amino-acid identities, respectively; Fig. 1a). In wild-type rice,
OsDWARF4L1/D11 transcripts were most abundant in vegetative
shoot apices, leaf sheaths and roots, and found at more moderate
levels in other organs (Fig. 1d). They were decreased by exogenous
brassinolide treatment, and increased in d61-3 and brd1-1 (Fig. 1e).
These results indicate that OsDWARF4L1/D11 expression is regulated
by a feedback mechanism, as is that of OsDWARF4.
Because null alleles of d11, such as d11-4, also show a weak
phenotype13(Fig. 2a), we suspected that OsDWARF4L1/D11 and
OsDWARF4 may function redundantly. To confirm this possibility,
we produced double mutants by crossing homozygous osdwarf4-1 and
d11-4 plants. Among the F2population, 34 of 576 plants showed
severe dwarfing and malformed leaves with tortuous leaf blades,
indistinguishable from the brassinosteroid-deficient brd1-1 phenotype
(Fig. 2a,c,d). This ratio fits the theoretical 9:3:3:1 (wild-type:osdwarf4-
1:d11-4:osdwarf4-1/d11-4) segregation predicted by mendelian inheri-
tance. PCR and sequence analyses confirmed that these plants were
osdwarf4-1/d11-4 double mutants (data not shown).
Endogenous levels of brassinosteroid intermediates in brd1 clearly
indicate that OsDWARF catalyzes C-6 oxidation9. However, quantifi-
cation in d2 could not determine the step(s) catalyzed by CYP90D2/
D2, because CYP90D2/D2 functions redundantly with another gene,
CYP90D3, and loss-of-function of CYP90D2/D2 have caused only
limited defects in brassinosteroid biosynthesis10. Similarly, endo-
genous levels of brassinosteroid intermediates were obviously changed
in the osdwarf4-1/d11-4 double mutant relative to osdwarf4-1 or d11
single mutants. Although brassinolide was not detected in either the
wild-type or osdwarf4-1/d11-4, another bioactive brassinosteroid, cas-
tasterone (CS), was detected in the wild-type but not in osdwarf4-1/
d11-4, confirming that osdwarf4-1/d11-4 is brassinosteroid-deficient
CYP90C1/ROT3 (A. thaliana)
CYP90D1 (A. thaliana)
CYP85A2/BR6o×2 (A. thaliana)
CYP85A1/BR6o×1 (A. thaliana)
CYP85A1/DWARF (Tomato)
CYP724A1 (A. thaliana)
OsDWARF4L1/D11 (Rice)
CYP90B1/DWARF4 (A. thaliana)
OsDWARF (Rice)
CYP90A1/CPD (A. thaliana)
OsCPD1 (Rice)
OsCPD2 (Rice)
OsDWARF4 (Rice)
CYP90D2/D2 (Rice)
CYP90D3 (Rice)
0.1
0
ATG
Met (1)
Proline-richDomain ADomain BDomain CHeme-binding
7 kb
TAA
Stop (502)
Tos17
0
0
0
2.1e5
3.5e5
0
2.5e5
0
910111213
6.8e4
7.3e5
0
1.6e5
CR
CR
CR (without protein)
CR + CYP90B2/OsDWARF4
CR + CYP724B/D11
22-OHCR (without protein)
1.3e497
Retention time (min)
Ion intensity (cps)
22-OHCR
OsDWARF4
1.0
1.0
0.51
0.44
1.0
1.0
2.31 2.58
3.94 2.93
D11
Histone H3 gene
OsDWARF4
D11
Histone H3 gene
Metabolite
CR
Metabolite
187
m/z
97
97
187
0
187
BL–
Shoot apex
Leaf sheath Leaf blade
Stem
Root
Inflorescence
Flower
BL+
WT
d61-3
brd1-1
100
200300400 500
m/z
100
200300 400500
m/z
100
200300400500
a
b
c
d
e
f
Figure 1 Molecular characterization of OsDWARF4 and OsDWARF4L1/D11. (a) Phylogenetic relationships among brassinosteroid-biosynthetic cytochrome
P450s of rice, A. thaliana and tomato. Bar, 0.1 amino-acid substitutions per site. (b) Genomic organization of the OsDWARF4 gene. Boxes represent exons,
and lines separating boxes represent introns. The arrowhead indicates the site of Tos17 insertion in osdwarf4-1. (c) Relative positions of the major domains
in OsDWARF4. All of the major domains found in the cytochrome P450 superfamily are conserved in OsDWARF4. The arrowhead indicates the site of Tos17
insertion in osdwarf4-1. (d) Expression of OsDWARF4 and OsDWARF4L1/D11 in various organs of wild-type rice. The gene encoding Histone H3 was used as
a control. (e) Feedback regulation of OsDWARF4 and OsDWARF4L1/D11 in brassinolide-treated wild-type plants (left) and brassinosteroid-related mutants
(right). Values under each panel indicate relative amounts of each transcript. (f) Function of recombinant CYP90B2/OsDWARF4 and CYP724B1/D11
proteins. Campesterol (CR) was incubated with each recombinant protein and analyzed by GC-MS. Inset shows that the mass spectrum of the trimethylsilyl
ester of each metabolite coincided with that of the trimethylsilyl ether of synthesized (22S)-22-hydroxycampesterol (22-OHCR).
106 VOLUME 24 NUMBER 1JANUARY 2006 NATURE BIOTECHNOLOGY
L E T T E R S
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology

Page 3

(Fig. 2e). Levels of the other five intermediates downstream of C-22
hydroxylation—6-deoxocathasterone (6-DeoxoCT), 6-deoxoteaster-
one(6-DeoxoTE),3-dehydro-6-deoxoteasterone
6-deoxotyphasterol (6-DeoxoTY)
DeoxoCS)—were also greatly reduced in osdwarf4-1/d11-4 (Fig. 2e).
This indicates that both proteins are involved in C-22 hydroxylation. It
is unclear why the upstream compound, campestanol, did not
accumulate in osdwarf4-1/d11-4. Because A. thaliana CYP90B1/
DWARF4 also catalyzes C-22 hydroxylation of other upstream com-
pounds such as campesterol and (24R)-ergost-4-en-3-one14, it is
possible that campestanol is not the major substrate for CYP90B2/
OsDWARF4 and CYP724B1/D11 in rice.
To confirm this hypothesis, we examined enzymatic functions
in vitro. Both recombinant CYP90B2/OsDWARF4 and CYP724B1/
D11, produced in insect cells via a baculovirus system, converted
campesterol to its 22a-hydroxylation product, (22S)-22-hydroxycam-
pesterol (22-OHCR; Fig. 1f), whereas only trace levels of 6-DeoxoCT
were detected after incubation with campestanol (Supplementary
Table 1 online). Comparison of the characteristic ions, m/z, from
synthesized 22-OHCR to that of the metabolized profile confirmed
that the novel peaks observed in CYP90B2/OsDWARF4 and
CYP724B1/D11 reactions represented 22-OHCR (Fig. 1f). No meta-
bolite was obtained when CYP724B1/D11 was incubated with other
substrates such as 6-DeoxoCT, 6-DeoxoTE, 6-Deoxo3DT, 6-DeoxoTY,
6-DeoxoCS, CS or brassinolide (Supplementary Table 1 online). This
confirms that CYP724B1/D11 catalyzes C-22 hydroxylation redun-
dantly with CYP90B2/OsDWARF4. It is noteworthy that in
A. thaliana, C-22 hydroxylation is catalyzed by a single enzyme
CYP90B1/DWARF4. However, consistent with the contrasting pheno-
typic severities of the osdwarf4-1 and d11 mutations, this reaction is
primarily catalyzed by CYP724B1/D11 in rice. d11-4 plants are shorter
than osdwarf4-1 plants (Fig. 2a). In addition, d11 mutants bear small
(6-Deoxo3DT),
and6-deoxocastasterone(6-
round grains13, which are never observed in osdwarf4-1. These
phenotypes indicate that whereas OsDWARF4L1/D11 is essential to
maintain levels of bioactive brassinosteroid synthesis needed for
normal shoot elongation and reproductive development, OsDWARF4
only contributes additional levels of bioactive brassinosteroid synthesis
required for normal leaf inclination but not for reproductive deve-
lopment. Therefore osdwarf4-1 shows slight dwarfism and erect
leaves without undesirable phenotypes such as abnormal leaf, flower
or grain morphology.
Because the harvest index (grain/(grain plus straw)) of rice is about
0.5, further increases in yield potential will have to involve increases in
crop biomass driven by more net photosynthesis15. In rice, the
contribution of lower leaves to photosynthesis is significant, even
though their photosynthetic capacity is considerably less than that of
upper leaves16. Erect leaves allow greater penetration of light to lower
leaves, thereby optimizing canopy photosynthesis1. Therefore, the
most likely explanation for the increase in above-ground biomass
in osdwarf4-1 plots under dense planting conditions (more than
1.2 times that in wild-type under conventional planting condition)
is that the shade of the upper leaves is minimized, and the lower leaves
receive more light to drive higher rates of photosynthesis (Fig. 3 and
Supplementary Table 2 online). No difference was observed in the
heading date (flowering time) and the grain-filling period (time to
maturity) between wild-type and osdwarf4-1 plants (data not shown).
In the wild-type, the above-ground biomass was increased in response
to higher fertilizer application under conventional planting (22.2
plants m–2), but was significantly decreased under dense planting
with doubled fertilizer application owing to reduced photosynthesis
caused by lodging (Fig. 3a). In contrast, the above-ground biomass of
osdwarf4-1 was significantly (P o 0.01) increased by dense planting at
each fertilizer rate. Consequently, the above-ground biomass in
osdwarf4-1 plots under dense planting with standard and a 1.5-fold
Wild-type
3,000
5.0
4.0
2.0
1.0
0.0
6-DeoxoCT
2,610
2,120
0.58
0.14
0.05
0.02
0.39
4.73
1.44
0.06
0.04
0.11
0.43
N.d.N.d.
BL
N.d.
CN
Endogenous contents (ng g–1 F.W.)
2,000
1,000
6-DeoxoTE
6-Deoxo3DT
6-DeoxoTY
6-DeoxoCS
CS
0
Ib
Ib
Wild-typeosdwarf4-1
IsIs
osdwarf4-1
osdwarf4-1/d11-4brd1-1
d11-4
osdwarf4-1/d11-4
Wild-type
ab
e
cd
Figure 2 Phenotypes of rice mutants defective in brassinosteroid biosynthesis.
(a) Comparison of gross morphology between the wild-type, single mutants
(osdwarf4-1, d11-4, brd1-1), and a double mutant (osdwarf4-1/d11-4). Bar,
30 cm. (b) Erect leaf phenotype of osdwarf4-1. The degree of bending between
leaf blade and sheath of osdwarf4-1 (right) is less than that of the wild-type
(left). lb, leaf blade; ls, leaf sheath. Bar, 3 cm. (c,d) Close-up views of
the osdwarf4-1/d11-4 double mutant (c) and brd1-1 (d). Bars, 5 cm.
(e) Endogenous contents of brassinosteroid intermediates in the wild-type
and osdwarf4-1/d11-4 double mutant. CN, campestanol; 6-DeoxoCT,
6-deoxocathasterone; 6-DeoxoTE, 6-deoxoteasterone; 6-Deoxo3DT,
3-dehydro-6-deoxoteasterone; 6-DeoxoTY, 6-deoxotyphasterol; 6-DeoxoCS,
6-deoxocastasterone; CS, castasterone; BL, brassinolide.
NATURE BIOTECHNOLOGY
VOLUME 24NUMBER 1 JANUARY 2006107
L E T T E R S
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology

Page 4

increase in fertilizer application was B1.34 and 1.52 times that in
wild-type plots under conventional cultivation conditions. The
increased biomass depended mainly on accelerated formation of
tillers, which have the ability to generate a panicle (panicle number,
Fig. 3b). Both total and fertile grain numbers tended to increase in
densely planted osdwarf4-1 plots (Fig. 3c), resulting in significantly
increased grain yield even when the fertilizer application was not
increased (for example, the osdwarf4-1 yield was 6.19 t ha–1under
dense planting with standard fertilizer application whereas the wild-
type yield was 4.69 t ha–1under conventional conditions; Fig. 3d).
However, the increase in grain yield was not as much as that in above-
ground biomass, because the harvest index of osdwarf4-1 (0.41) was
slightly lower than that of the wild-type (0.44). A similar field test gave
comparable results.
Overexpression of DWARF in transgenic tomato increases plant
height17, and overexpression of DWARF4 in transgenic A. thaliana and
tobacco increases vegetative growth and seed yield18. In rice, however,
the role of brassinosteroid in regulating leaf angle has been known for
over 40 years19,20. Whereas brassinosteroid deficiency increased the
erectness of leaves, the opposite phenotype (increased leaf angle) was
observed in transgenic rice overexpressing OsDWARF4 (27.61 versus
10.01 in the wild-type, P o 0.01), and was inherited in their progeny
(data not shown). Because inclined leaves are unfavorable for photo-
synthesis when plants are grown in a group, we have to carefully
evaluate whether the overexpression of a brassinosteroid-biosynthesis
gene in transgenic rice increases grain yield as is the case in transgenic
plants overexpressing DWARF4.
A previous study concluded that more erect leaves improve not only
light distribution but also the size and use of leaf nitrogen reservoirs
for grain growth1. Our results clearly demonstrate that the erect leaf
phenotype caused by brassinosteroid deficiency improved biomass
production under field conditions, resulting
in increased grain yield without the negative
environmental effects associated with fertili-
zers. We have recently identified rice homo-
logs of A. thaliana BAS1 that encodes the
brassinosteroid catabolic enzyme, CYP734A1
(ref. 21). Overexpression of these rice homo-
logs in transgenic rice successfully induced
thebrassinosteroid-deficient
(unpublishedresults).
expression of these genes should be restricted
phenotypes
ideally,Although
to particular tissues and organs such as leaf laminae, it should be
possible to generate erect leaf varieties by the introduction of a single
transgene. In conclusion, engineering more erect leaves modulating
brassinosteroid metabolism can be combined with dense planting to
improve rice grain yields.
METHODS
Plant materials. Seeds of osdwarf4-1, the Tos17 insertion knockout line
of OsDWARF4 (NE7040), were kindly provided by the Rice Genome
Resource Center (RGRC) at the National Institute of Agrobiological Sciences
(NIAS). Seedlings of wild-type rice (Oryza sativa L. cv. ‘Nipponbare’) and
mutants were grown in a greenhouse at 28 1C under ambient light conditions.
Nipponbare is the original strain for all mutants used in this study including
osdwarf4-1, d11-4, brd1-1 and d61-3. For expression analysis, wild-type seeds
were sown on an agar medium containing 106M brassinolide and grown for
1 week. For yield evaluation, the wild-type and osdwarf4-1 were grown in the
paddy field by transplanting one plant per hill at a distance of 30 ? 15 cm
(22.2 plants m–2) or 15 ? 15 cm (44.4 plants m–2) at three nitrogen fertilizer
concentrations (6, 9 and 12 g m–2). Out of total fertilizer, three-fifths was
applied by basal dressing with an ordinary chemical fertilizer, and remaining
two-fifths was by the top dressing with a slow-release fertilizer. Each treatment
was replicated three times in randomized blocks, and each plot was 2 ? 2 m.
Three neighboring individuals were harvested from each plot, excluding
marginal plants.
Sequence analysis. BLAST search, mapping and phylogenetic analysis were
performed as previously described22. Accession numbers are listed in Supple-
mentary Table 3 online.
Expression analysis. RT-PCR was performed with DNase-treated total RNAs
separately prepared from various organs of rice by using the Advantage RT-for-
PCR Kit (Clontech). The primer sequences are 5¢-AGTCGCGTGCTGC
CATTCTCGGAGTAATAG-3¢
and5¢-AGCAAGCTCAGCAAGAGGTCCAG
GATTTGC-3¢
for OsDWARF4, and5¢-TTGGGTCATGGCATGGCAAGA
2.0
0.8
0.5
Panicle dry weight (kg m–2)
Estimated grain yield (t ha–1)
0.0
1.0
de
wx
wx
fgh
def
def
efg
wx
wx
bc
w
yz
cd
h
xy
xy
wx
a
v
b
yz
gh
h
z
Above-ground biomass (kg m–2)
0.0
Density
Fertilizer
W
22.2
×1
×1.5
×2
44.4
M W M W M W M
22.2
W M W M
22.2 44.4
W M
44.4
7.0
5.0
e
cde
cd
cd
cd
de
bcdbcd
abc
ab
a
f
0.0
Density
Fertilizer
W
22.2
×1
×1.5
×2
44.4
M W MW M W M
22.2
W M W M
22.2 44.4
W M
44.4
a
Panicle number (m–2)
Panicle length (cm)
cd
cd
bc
bc
bc bc
b
z
c
ab
yz
yz
yz
yz
yz
yz
d
y
y
y
y
a
a
y
450
400
200
0
Density
Fertilizer
W
22.2
×1
×1.5
×2
44.4
M W M W M W M
22.2
W M W M
22.2 44.4
M
0
10
14
44.4
b
d
30,000
20,000
10,000
0
Grain number (m–2)
bc
yz
abc
a
a
z
y
abc
abc
abc
abc
ab
abc
ab
yzyz
yz
yz
yz
yz
yz
yz
c
z
Density
Fertilizer
W
22.2
×1
×1.5
×2
44.4
M W MW M W M
22.2
W M W M
22.244.4
W M
44.4
c
Figure 3 Characterization of a field-grown
osdwarf4-1 mutant. (a) Comparison of biomass
production between the wild-type and osdwarf4-1.
White and black bars indicate above-ground and
panicle dry weight, respectively. (b) Comparison
of panicle number (white bar) and panicle
length (black bar). (c) Comparison of total
(white bar) and fertile (black bar) grain number.
(d) Comparison of estimated grain yields between
wild-type (white bar) and osdwarf4-1 (black bar).
W, wild-type, M, osdwarf4-1. Density indicates
conventional planting (22.2 plants m–2) or dense
planting (44.4 plants m–2). Fertilizer indicates
the level of nitrogen fertilizer. Conventional
condition is ?1 (6 g m–2), and two increased
conditions are ?1.5 and ?2 (9 and 12 g m–2
respectively). Lowercase letters indicate
significant differences at the level of P o 0.01
within a parameter (Tukey’s Honest Significant
Difference test).
108 VOLUME 24NUMBER 1 JANUARY 2006NATURE BIOTECHNOLOGY
L E T T E R S
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology

Page 5

GCAAGGA-3¢ and 5¢-TTGTTGCTGGAGCCAGCATTCCTCCTCT-3¢ for OsD-
WARF4L1/D11. These primers specifically amplified the target gene sequences
(data not shown). Accumulation of OsDWARF4 and OsDWARF4L1/D11
transcripts in brassinolide-treated seedlings and brassinosteroid-related
mutants was analyzed by real-time PCR with an iCycler iQ real-time PCR
system (Bio-Rad).
Brassinosteroid analysis. Shoots from wild-type and mutant plants were
harvested at 4 weeks after germination. Brassinosteroid was extracted, purified
and quantified as previously described9.
Enzyme assay. The full-length cDNA of OsDWARF4L1/D11 (AK106528)
was kindly provided by RGRC at NIAS. The cDNAs of CYP90B2/OsDWARF4
and CYP724B1/D11 were cloned into the pFastBac1 vector (Life Technologies),
and heterologous expression in a baculovirus-insect cell system was performed
as previously described23. Each activity of CYP90B2 and CYP724B1 was assayed
in a reaction mixture consisting of the microsomes (50 pmol of the
P450), NADPH:cytochrome P450 reductase, 100 mM potassium phosphate
(pH 7.25), 1 mM NADPH and 100 mM of brassinosteroid intermediate
compounds. The reaction was carried out at 30 1C for 2 h, and the reaction
products were extracted three times with a half-volume of ethyl acetate. The
organic phase was collected and evaporated. The residue was treated with 10 ml
of N-methyl-N-trimethylsilyltrifluoroacetamide and 40 ml of pyridine at 80 1C
for 30 min and analyzed by gas chromatography–mass spectrometry (GC-MS)
as described24.
Plasmid constructs and plant transformation. OsDWARF4 was overexpressed
in transgenic rice, an OsDWARF4 cDNA was inserted between the rice actin
promoter and the nopaline synthase polyadenylation signal of hygromycin-
resistant binary vector pAct-Hm2. Agrobacterium tumefaciens–mediated trans-
formation was performed as described previously25.
Note: Supplementary information is available on the Nature Biotechnology website.
ACKNOWLEDGMENTS
T.S. was supported by a grant from the Ministry of Agriculture, Forestry and
Fisheries of Japan (Rice Genome Project IP-1010) and a Grant-in-Aid for Young
Scientists (B) from the Ministry of Education, Culture, Sports, Science and
Technology (MEXT). M.U.-T. and H.K. were supported by a grant from
the Program for Promotion of Basic Research Activities for Innovation of
Biosciences, and M. Matsuoka was supported by a Grant-in-Aid for Centers of
Excellence from MEXT. We thank M. Sekimoto, H. Hayashi, N. Kato, K. Izumi
and K. Yatsuda for their technical assistance.
AUTHORS’ CONTRIBUTIONS
T.S. conceived the experiment, and together with Y.M., M.U.-T. and
M. Matsuoka carried it out; S.F., S.T. and S.Y. carried out the brassinosteroid
analysis; T.O., M. Mizutani, and K.S. carried out the enzyme assay; H.S.,
H.T. and H.K. generated the transformants and double mutants; T.S. and
M. Matsuoka cowrote the paper.
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Published online at http://www.nature.com/naturebiotechnology/
Reprints and permissions information is available online at http://npg.nature.com/
reprintsandpermissions/
1. Sinclair, T.R. & Sheehy, J.E. Erect leaves and photosynthesis in rice. Science 283,
1455 (1999).
2. Evans, L.T. Crop Evolution, Adaptation and Yield (Cambridge Univ. Press, Cambridge,
1993).
3. Peng, J. et al. ‘Green revolution’ genes encode mutant gibberellin response modulators.
Nature 400, 256–261 (1999).
4. Sasaki, A. et al. Green revolution: a mutant gibberellin-synthesis gene in rice. Nature
416, 701–702 (2002).
5. Sakamoto, T. et al. Genetic manipulation of gibberellin metabolism in transgenic rice.
Nat. Biotechnol. 21, 909–913 (2003).
6. Sakamoto, T. & Matsuoka, M. Generating high-yielding varieties by genetic manipula-
tion of plant architecture. Curr. Opin. Biotech. 15, 144–147 (2004).
7. Li, J. & Chory, J. A putative leucine-rich repeat receptor kinase involved in brassinos-
teroid signal transduction. Cell 90, 929–938 (1997).
8. Yamamuro, C. et al. Loss of function of a rice brassinosteroid insensitive1 homolog
prevents internode elongation and bending of the lamina joint. Plant Cell 12, 1591–
1606 (2000).
9. Hong, Z. et al. Loss-of-function of a rice brassinosteroid biosynthetic enzyme, C-6
oxidase, prevents the organized arrangement and polar elongation of cells in the leaves
and stem. Plant J. 32, 495–508 (2002).
10.Hong, Z. et al. A rice brassinosteroid-deficient mutant, ebisu dwarf (d2), is caused by a
loss of function of a new member of cytochrome P450. Plant Cell 15, 2900–2910
(2003).
11.Choe, S. et al. The DWF4 gene of Arabidopsis encodes a cytochrome P450 that
mediates multiple 22a-hydroxylation steps in brassinosteroid biosynthesis. Plant Cell
10, 231–243 (1998).
12.Bancos, S. et al. Regulation of transcript levels of the Arabidopsis cytochrome P450
genes involved in brassinosteroid biosynthesis. Plant Physiol. 130, 504–513 (2002).
13.Tanabe, S. et al. A novel cytochrome P450 is implicated in brassinosteroid biosynthesis
via the characterization of a rice dwarf mutant, dwarf11, with reduced seed length.
Plant Cell 17, 776–790 (2005).
14.Fujioka, S., Takatsuto, S. & Yoshida, S. An early C-22 oxidation branch in brassinos-
teroid biosynthetic pathway. Plant Physiol. 130, 930–939 (2002).
15.Mann, C.C. Crop scientists seek a new revolution. Science 283, 310–314 (1999).
16.Horton, P. Prospects for crop improvement through the genetic manipulation of
photosynthesis: morphological and biochemical aspects of light capture. J. Exp. Bot.
51, 475–485 (2000).
17.Bishop, G.J. et al. The tomato DWARF enzyme catalyses C-6 oxidation in brassinoster-
oid biosynthesis. Proc. Natl. Acad. Sci. USA 96, 1761–1766 (1999).
18.Choe, S. et al. Overexpression of DWARF4 in the brassinosteroid biosynthetic pathway
results in increased vegetative growth and seed yield in Arabidopsis. Plant J. 26,
573–582 (2001).
19.Maeda, E. Rate of lamina inclination in excised rice leaves. Physiol. Plant. 18,
813–827 (1965).
20.Wada, K. et al. Brassinolide and homobrassinolide promotion of lamina inclination of
rice seedlings. Plant Cell Physiol. 22, 323–325 (1981).
21.Neff, M.M. et al. BAS1: A gene regulating brassinosteroid levels and light responsive-
ness in Arabidopsis. Proc. Natl. Acad. Sci. USA 96, 15316–15323 (1999).
22.Sakamoto, T. et al. An overview of gibberellin metabolism enzyme genes and their
related mutants in rice. Plant Physiol. 134, 1642–1653 (2004).
23.Saito, S. et al. Arabidopsis CYP707As encode (+)-abscisic acid 8¢-hydroxylase, a key
enzyme in the oxidative catabolism of abscisic acid. Plant Physiol. 134, 1439–1449
(2004).
24.Fujita, S. et al. Arabidopsis CYP90B1 catalyzes the early C-22 hydroxylation of C27,
C28, and C29 sterols. Plant J. 45, in press.
25.Hiei, Y., Ohta, S., Komari, T. & Kumashiro, T. Efficient transformation of rice (Oryza
sativa L.) mediated by Agrobacterium and sequence analysis of boundaries of the
T-DNA. Plant J. 6, 271–282 (1994).
NATURE BIOTECHNOLOGY
VOLUME 24NUMBER 1JANUARY 2006109
L E T T E R S
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology