formation of corymb-like inflorescences due to delay in bolting and
TRANSCRIPT
Plant Cell Physiol. 43(3): 298–306 (2002)
JSPP © 2002
Formation of Corymb-like Inflorescences Due to Delay in Bolting and Flower Development in the corymbosa2 Mutant of Arabidopsis
Mitsuhiro Suzuki, Taku Takahashi and Yoshibumi Komeda 1
Division of Biological Sciences, Graduate School of Science, Hokkaido University, N10, W8, Sapporo, 060-0810 Japan
;
Among the wild-type ecotypes of Arabidopsis thaliana
whose shape of inflorescence is categorized as raceme, the
ecotype Landsberg harboring the erecta (er) mutation shows
a corymb-like inflorescence, namely, a compact inflores-
cence with a flattened arrangement of flower buds at the tip.
The fact that the ER gene encodes a receptor-like protein
kinase implies the presence of a signaling cascade responsi-
ble for the inflorescence morphology of flowering plants. We
report here the characterization of another mutant with a
corymb-like inflorescence, named corymbosa2 (crm2), and
the isolation of the CRM2 gene. While the er mutation
causes a severe reduction in the length of pedicels, the crm2
mutation results in a significant delay in the initiation of
internode elongation and in the development of flowers,
despite having little effect on the timing of floral induction.
Consequently, the number of flower buds is apparently
increased at the tip of crm2 inflorescence. The crm2 er dou-
ble mutant shows an additive phenotype. These results sug-
gest that CRM2 and ER may act in different ways to gener-
ate wild-type inflorescence. The CRM2 gene was isolated by
positional cloning and appears to encode a polypeptide with
no significant homology to known sequences.
Key words: Arabidopsis thaliana — Bolting — Corymb-like
inflorescence — corymbosa2 — Flower development — Inflo-
rescence development.
Abbreviations: BAC, bacterial artificial chromosome; CaMV,
cauliflower mosaic virus; CAPS, cleaved amplified polymorphic
sequence; GUS, �-glucuronidase; NLS, nuclear localization signal;
ORF, open reading frame; RT-PCR, reverse transcription-PCR; SAM,
shoot apical meristem; SEM, scanning electron microscopy; SSLP,
simple sequence length polymorphism.
The nucleotide sequence reported in this paper has been submit-
ted to GenBank under accession number AF327068 (CORYMBOSA2).
Introduction
The inflorescence shoot architecture of flowering plants
can be divided into various types according to anatomical crite-
ria such as the identity of the shoot apical meristem (SAM), the
pattern of lateral flower primordia initiation, and the degree of
internode elongation. In Arabidopsis thaliana, transition of a
vegetative shoot meristem to an inflorescence meristem is
accompanied by continuous production of flowers on the flanks
of the SAM with a spiral phyllotaxy and by rapid elongation of
pedicels (the stalk of a single flower) and stem internodes. The
resulting shape of the inflorescence is referred to as raceme
(Weberling 1989). However, the molecular basis correspond-
ing to such terminology remains largely unexplored.
The SAM is a key source of cells for shoot growth. A
number of genes involved in determination of meristem iden-
tity and in regulation of meristem activity have been identified
in Arabidopsis. The TERMINAL FLOWER1 (TFL1) gene plays
a critical role in the maintenance of indeterminate inflorescence
meristems (Alvarez et al. 1992). While CLAVATA (CLV) genes
are involved in promoting the progression of meristem cells
toward organ initiation, SHOOT MERISTEMLESS (STM)
antagonistically functions in preventing meristem cells from
being incorporated into organ primordia (Barton and Poethig
1993, Clark et al. 1996, Laux and Schoof 1997). CLV1 has
been shown to encode a receptor-like kinase with an extracellu-
lar domain composed of tandem leucine-rich repeats (Clark et
al. 1997), and this suggested the exciting possibility that a cell-
cell communication of information on cell proliferation and
differentiation at shoot meristems might be mediated by recep-
tor-ligand interactions in a similar manner to growth-factor sig-
naling cascades in animal systems. Based on genetic evidence,
a small secreted peptide encoded by CLV3 has been identified
as a likely candidate for the CLV1 ligand (Fletcher et al. 1999).
Furthermore, wuschel (wus) mutations, which result in prema-
ture termination of shoot and floral meristems as with stm,
appear to be epistatic to clv mutations, suggesting that WUS
may function as a downstream target of CLV genes (Laux et al.
1996).
On the other hand, there is little genetic information on
the regulatory mechanisms of phyllotactic patterns of flower
primordia and the lengthening of pedicels and stem internodes,
both of which also contribute to formation of the gross mor-
phology of inflorescence. Mutation in the ERECTA (ER) locus,
which is known to be harbored by the wild-type Arabidopsis
ecotype Landsberg erecta (Ler), results in a compact inflores-
cence, short pedicels, and upright blunt siliques. As a conse-
quence, a bunch of flower buds at the tip of inflorescences con-
tinuously forms an umbellate flat surface referred to as
corymbose inflorescence. The ER gene codes for a receptor-
like kinase similar in overall structure to that of CLV1 (Torii et
1 Corresponding author: E-mail, [email protected]; Fax, +81-11-706-2739.
298
CORYMBOSA2 in inflorescence development 299
al. 1996). However, it is not known how the ER gene product
performs signal transduction. To understand the molecular
mechanisms regulating the inflorescence shoot architecture, the
genes involved must be identified. Toward this goal, we have
isolated a series of Arabidopsis mutants, named corymbosa
(crm), which show corymb-like inflorescences (Komeda et al.
1998). Here we describe the phenotypic characterization of the
crm2 mutant in comparison with er. We show that the corymb-
like phenotype of crm2 is manifested in a different fashion
from that of er. Cloning of the CRM2 gene revealed that it
encodes a novel protein.
Results
Overview of the corymb-like phenotype
In wild-type Columbia (Col) plants of A. thaliana, the api-
cal inflorescence meristem and flowers at early stages are cov-
ered with flowers at later stages (Fig. 1A). This is simply due
to a rapid upward growth of pedicels in unison with flower
organ development. In er mutants, the length of pedicels as
well as stem internodes is apparently reduced, and this results
in the corymb-like phenotype of the tip of the inflorescence
(Fig. 1B, F). The crm2 mutant was isolated by screening for
mutants with a similar appearance of inflorescence as that of er
mutants (Komeda et al. 1998). The flattened arrangement of
flower buds in the crm2-1 allele is shown in Fig. 1C. We
observed that, while the total number of flowers at stages 7–13
(stages defined by Smyth et al. 1990) at the tip of the inflores-
cence was normal in er alleles, it was increased in the crm2-1
allele compared to that of the wild type (Fig. 1E–G). Further-
more, in contrast to er alleles, in which the length of fully elon-
gated pedicels is about half of that in the wild type, crm2-1
appeared to have a moderate effect on the length of pedicels.
The number of floral organs was unaffected in crm2-1, as with
er alleles. These observations indicate that the corymb-like
phenotype in crm2-1 is attributable in part to the increased
number of flower buds at the tip.
To examine the relationship between CRM2 and ER in
inflorescence development, crm2 er double mutants were con-
structed. For the er mutation, the er-105 allele, which has an
Fig. 1 Morphology of inflorescences and flower buds at the tips in
wild-type (ecotype Columbia), er and crm2 plants. (A–D) Top views
of tips of wild-type (A), er-105 (B), crm2-1 (C) and crm2-1 er-105
double mutant (D) apical inflorescences. (E–G) Flower buds at stages
7–13 (stages according to Smyth et al. 1990) decapitated from wild-
type (E), er-105 (F) and crm2-1 (G) apical inflorescences. Bars =
1 mm.
Table 1 Measurements of wild-type and crm2-1 plants
a Days from seed germination are shown.b Total number of leaves and flowers visible without dissection and
using a dissecting microscope on the primary shoot was scored.c The mean length of the first ten internodes between nodes bearing a
single flower.
Mean values � SD in 20 wild-type (Col) and crm2-1 plants are shown.
Wild type crm2-1
Number of rosette leaves 9.9�0.8 10.1�0.6
Number of cauline leaves 2.7�0.5 4.4�0.5
Days to 1-cm bolting a 28.4�1.1 37.1�1.5
Days to first-flower opening a 31.5�1.9 42.2�3.3
Total number of lateral meristems at day 32 b 29.3�1.0 30.1�1.0
Length of stem internodes (mm) c 9.4�2.0 4.9�1.6
Length of flower organs (mm)
Sepal 1.9�0.1 1.9�0.1
Petal 2.7�0.3 2.6�0.2
Stamen 2.2�0.2 1.7�0.1
Carpel 2.3�0.1 2.2�0.2
Pedicel 6.9�0.6 4.1�0.3
CORYMBOSA2 in inflorescence development300
insertion of a 4-kb DNA fragment of unknown origin in the ER
gene and putatively represents a null allele (Torii et al. 1996),
was used. We did not detect a novel mutant phenotype in dou-
ble-mutant seedlings at the vegetative stage. After bolting,
crm2-1 er-105 double mutants showed an extremely exagger-
ated corymb-like phenotype at the tip of the inflorescence (Fig.
1D), suggesting that the effects of crm2 on inflorescence devel-
opment is additive to er.
The crm2 mutation delays the timing of bolting and flower
development
Under standard long-day conditions (see Materials and
Methods), the number of rosette leaves was normal in the
crm2-1 allele. On the other hand, crm2-1 produced approxi-
mately two cauline leaves more than the wild type (Table 1).
These results suggest that the crm2 mutation affects the flower-
ing time. To determine whether or not the timing of floral
induction is delayed in crm2-1, expression of the flower-meris-
tem-identity gene APETALA1 (AP1) was examined by reverse
transcription (RT)-PCR. AP1 is known to be expressed
throughout stage 1 and 2 flower primordia and in the first and
second whorls of developing flowers but not in vegetative and
inflorescence meristems (Mandel et al. 1992). We could not
Fig. 2 Comparison of the time courses of flower initiation, opening
and internode elongation in wild-type and crm2 plants. (A) RT-PCR
analysis of the expression of AP1 transcripts. Total RNA was prepared
from whole seedlings of the wild-type (WT) and crm2-1 plants grown
for indicated days (10–22) after germination, respectively, and sub-
jected to RT-PCR experiments (see Materials and Methods). As a con-
trol, the TFL1 transcript was amplified. (B) Time course of production
of flowers on the primary shoot axis. The number of flowers after
stage 13 (anthesis) was counted. The mean values (� SD) of 20 wild-
type (square) and crm2-1 (circle) plants are plotted. (C) Time course of
growth of stem internodes. The plant height was measured from the
day of 1-cm bolt. The mean values (� SD) of 20 wild-type (square) and
crm2-1 (circle) plants are plotted.
Fig. 3 Comparison of the development of primary inflorescences in
wild-type and crm2 plants. (A) A 27-day-old wild-type inflorescence.
(B) A 36-day-old crm2-1 inflorescence. (C) A 40-day-old wild-type
inflorescence. (D) A 50-day-old crm2-1 inflorescence. Arrows indicate
variation in silique length due to reduced fertility in crm2-1. Bars =
1 mm (A, B) and 5 mm (C, D).
CORYMBOSA2 in inflorescence development 301
detect clear difference between the time courses of AP1 expres-
sion in crm2-1 and wild-type plants (Fig. 2A).
On the other hand, crm2-1 started to bolt approximately
9 d later than did the wild type (Table 1). The period from ger-
mination to anthesis (stage 13) of the first flower in crm2-1 was
also much longer than that in the wild type (Table 1; Fig. 2B).
However, the production rate of stage 13 flowers by the pri-
mary inflorescence meristem in crm2-1 was higher than that in
wild-type plants (1.6 vs. 1.2 flowers per day for the first 5 d in
the wild type; Fig. 2B). These results indicate that the crm2
mutation causes accumulation of flower buds at the tip of inflo-
rescences due to a significant delay in the initiation of bolting
and in the development of floral organs but has little effect on
the timing of floral induction. Consistent with this, the total
number of visible flowers just before bolting was apparently
increased in crm2-1 (Fig. 3A, B). However, the total number of
lateral meristems including leaves and flowers was unaffected
in crm2-1 in comparison with that in the wild type of the same
day (Table 1). At day 36, when all of the flowers in crm2-1
were at young bud stages and they were still in the center of the
rosette (Fig. 3B), wild-type plants already had flowers after
anthesis, developing siliques, and elongating internodes. Fur-
thermore, the growth of stem internodes after bolting was
decelerated in crm2-1 (Fig. 2C). The final length of an individ-
ual internode was also reduced in crm2-1 (Table 1; Fig. 3C, D).
Examination of apical inflorescence meristems by scan-
ning electron microscopy (SEM) revealed that the spiral
arrangement of flower primordia at its flanks was unaffected in
crm2-1 (Fig. 4A, B). Sections of inflorescence apices revealed
that the number of flowers at early stages was increased in
crm2-1 (Fig. 4C, D), while the apical meristem in the crm2-1
allele was normal in size and shape (Fig. 4E, F).
Reduced fertility in crm2
In addition to the corymb-like phenotype, crm2-1 exhib-
ited reduced fertility (Fig. 3C, D). Although mature siliques
with wild-type length were occasionally produced, the length
of siliques and the seed set varied along the stem. Because of
the reduced fertility and/or the delayed development of flow-
ers after stage 13, these flowers were also observed in the out-
ermost whorl of a compact flower-bud cluster in the crm2 er
double mutant (Fig. 1D), instead of developing siliques in the
er single mutant. We measured the lengths of floral organs of
stage 13 flowers (Table 1). Unlike sepals, petals, and carpels,
which exhibited normal lengths in crm2-1, the average length
of crm2-1 stamens was shorter than that of wild-type stamens.
On the other hand, crm2-1 apical inflorescence meristems con-
tinued to produce flowers for a prolonged period compared to
the wild type (Fig. 2B). As a result, the crm2-1 plants finally
had more flowers than did the wild type and exhibited a final
stature comparable to or higher than that of the wild type, in
spite of a reduction in the length of each internode. These phe-
notypes may be accounted for by reduced fertility, which often
causes a prolonged proliferative activity of apical inflores-
cence meristems (Hensel et al. 1994).
Positional cloning of CRM2
The crm2 mutation segregates as a monogenic recessive
(Komeda et al. 1998). Fine mapping experiments revealed that
the CRM2 locus is located within an 80-kb region between the
markers T13K14.160-RsaI and g13838–1.4 on chromosome IV,
which we designed based on the genomic sequence data in
GenBank (see Fig. 5A). We determined DNA sequences of all
of the predicted open reading frames (ORFs) in this region by
using the crm2-1 genome and found only a 3-bp deletion in a
putative ORF in the bacterial artificial chromosome (BAC)
clone T13K14 (Fig. 5B).
Fig. 4 Microscopic observation of wild-type and crm2 inflores-
cences. (A and B) Scanning electron micrographs of wild-type (A) and
crm2-1 (B) inflorescence apex. (C and D) Cross sections of the wild-
type (C) and crm2-1 (D) inflorescence apexes. Arrows indicate flower
buds of wild-type and crm2-1 plants at the same stage (stage 6; stages
defined by Smyth et al. 1990). Note that the number of flower buds at
early stages (asterisks) is increased in crm2-1. (E and F) Longitudinal
sections of the wild-type (E) and crm2-1 (F) inflorescence apexes.
Bars = 50 �m.
CORYMBOSA2 in inflorescence development302
To determine whether it corresponds to the CRM2 gene or
not, a 6-kb wild-type genomic DNA fragment encompassing
the ORF (Fig. 5B) was used for complementation experiments.
After transformation of wild-type plants with this fragment,
they were crossed with homozygous crm2-1 mutants. Our
results showed that all of the F2 populations derived from four
independent crosses failed to segregate for kanamycin-resist-
ant plants with the mutant phenotype (data not shown).
Sequence determination of the endogenous CRM2 gene of the
kanamycin-resistant F2 plants (see Materials and Methods)
revealed that some of them were homozygous for the crm2-1
allele and phenotypically wild type (Fig. 6A).
We further transformed wild-type plants with an antisense
construct containing a part of this ORF (Fig. 5B) fused in
reverse orientation to the cauliflower mosaic virus (CaMV)
35S RNA promoter. Five of sixteen transgenic lines displayed
phenotypes similar to the original crm2 phenotype (Fig. 6B).
Taken together, these results indicate that this ORF is the
CRM2 gene.
Analysis of the CRM2 gene
Isolation of the corresponding cDNA by RT-PCR revealed
that the CRM2 gene contains nine introns (Fig. 5B). It encodes
a polypeptide of 942 amino acids with a predicted molecular
mass of 104 kDa (Fig. 5C). The crm2-1 mutation is located at
the second exon and it causes a deletion of a lysine (K) residue
in the CRM2 gene product (Fig. 5B, C). A BLAST search using
the CRM2 polypeptide sequence revealed no homology with
Fig. 5 Identification of the CRM2 gene. (A) Mapping of the CRM2
locus. CAPS markers on chromosome IV, g13838–1.4 and PG11, are
shown at the top. Asterisks represent the position of the CRM2 locus.
Four BAC clones around the CRM2 locus are shown below. Vertical
arrows indicate positions of CAPS markers that we designed according
to polymorphisms between Columbia and Landsberg ecotypes (see
Materials and Methods). Number of recombinant chromosomes from
870 crm2-1 homozygous plants (1740 chromosomes) of the F2 prog-
eny of a cross between crm2-1 and the wild type (ecotype Landsberg)
are indicated. (B) Structure of the CRM2 gene. Exons are indicated by
black boxes. The position of a 3-bp deletion in crm2-1 is indicated by
a vertical arrow. Arrows with restriction sites represent PCR primers
used for amplification of a genomic DNA fragment for transgenic
complementation experiments. An open bar represents a restriction
fragment used for the antisense construct. (C) The deduced amino acid
sequences of the CRM2 protein. The amino acid residues derived from
translation of the RT-PCR product are shown (one-letter code). The
functional translation stop site (TGA) is marked with an asterisk. Dou-
ble underlines indicate the site of the crm2-1 mutation. A region simi-
lar to the double-stranded RNA-binding motif is underlined.
Fig. 6 Analysis of transgenic Arabidopsis plants. (A) Complementa-
tion of the crm2 phenotype with the wild-type CRM2 gene. A wild-
type genomic DNA fragment containing the CRM2 locus was obtained
by PCR (see Fig. 5) and introduced into wild-type plants. crm2-1
homozygous plants with wild-type appearance were identified in the
F2 progeny of the cross with crm2-1 plants by PCR and sequencing of
the allele. (B) A phenotype conferred by antisense expression of the
CRM2 gene. A partial genomic fragment of the CRM2 gene (see Fig.
5B) was fused to the CaMV 35S RNA promoter and introduced into
wild-type plants. Bars = 1 mm.
CORYMBOSA2 in inflorescence development 303
known proteins. However, we found that a homologous
sequence is tandemly arrayed with the CRM2 gene in the BAC
clone T13K14. The CRM2 polypeptide sequence has 64.2%
identity overall with a deduced polypeptide of its neighboring
homolog (T13K14.80; accession number T10634). The PSORT
algorithm (Nakai and Kanehisa 1992) recognized a potential
nuclear targeting sequence (KKKK, amino acid residues 68 to
71), which is converted into KKK in crm2-1. The deduced
polypeptide of T13K14.80 also contains a nuclear targeting
sequence (KRKK, amino acid residues 56 to 59) in the corre-
sponding position (KKKK) of the CRM2 protein (data not
shown). Motif analysis using the Pfam HMM database
revealed a region of the CRM2 protein between amino acids 11
and 85 that displays similarity with the double-stranded RNA-
binding motif (Fig. 5C).
Because we detected no positive signals of CRM2 expres-
sion by RNA gel blot analysis, we carried out RT-PCR to deter-
mine where the CRM2 gene is expressed. The CRM2 tran-
scripts were mainly detected in young seedlings, root tissue and
flower buds with shoot apices (Fig. 7). Similar expression pat-
terns were also observed on the T13K14.80 gene (Fig. 7).
Discussion
There is an astonishing variety of forms of flower-bearing
branching systems, which contain different modifications of
the same basic plan often observed within the same family. As
one approach toward unravelling the molecular basis for these
divergent forms of inflorescence in flowering plants, we
focused on the inflorescence morphology in A. thaliana. As
described in detail by Weberling (1989), the raceme inflores-
cence of the family Cruciferae, to which Arabidopsis belongs,
is characterized by elongation of stem internodes of its main
axis and development of pedicels in each flower. Although er
and crm2 mutant inflorescences take on the form of a raceme at
later stages of development, these continuously keep a corymb-
like flattened arrangement of flower buds at the tip. Our analy-
sis revealed that the corymb-like form in crm2 is generated by
an essentially different mechanism from that in er.
Fig. 8 illustrates how corymb-like phenotypes are mani-
fested in these mutants. Because the timing of bolting and
flower opening was normal in er-105, which represents a puta-
tive null allele (data not shown), the er phenotype is primarily
attributable to a reduction in the pedicel length. On the other
hand, the concurrent delay in flower organ development and
bolting in crm2-1 resulted in the accumulation of flower buds
at the tips of inflorescences. Although we could not detect any
differences between AP1 expression profiles in the wild type
and crm2-1, crm2-1 produced more cauline leaves than did the
wild type, suggesting that the timing of floral induction may be
only slightly delayed. These phenotypes might be interpreted,
at least in part, by the delayed cell division rate at the cellular
level. It is less likely that the rate of initiation of flower primor-
dia is increased in crm2-1, because the total number of leaves
and flowers was unchanged in young flowering plants of the
mutant compared to wild-type plants of the same day (Table 1).
However, the interpretation of the crm2 phenotype is somewhat
complicated by the fact that the rate of production of stage 13
flowers was higher in the mutant than that in the wild type (Fig.
2B). Assuming that the rate of initiation of flower primordia is
normal in crm2-1, this result suggests that the duration of
flower organ development may be gradually restored in late
flowers and, accordingly, the phenotype must become moder-
ated. To clarify this point, further analysis of the development
of individual flowers will be needed. Smyth et al. (1990)
defined 12 stages of Arabidopsis flower development from ini-
tiation until the bud opens. Although the possibility that the
duration of certain stages in crm2 may be particularly pro-
longed cannot be ruled out, we observed no obvious accumula-
tion of flower buds of specific stages in the mutant bud cluster
(Fig. 1, 4), suggesting that the crm2 mutation may affect all
stages of flower development.
Fig. 7 RT-PCR analysis of CRM2 and its neighboring homolog
T13K14.80. Total RNA was isolated from 7-day-old seedlings (Se),
root tissue (R), rosette leaves (L), stem internodes (St), flowers con-
taining shoot apical meristems (F) and siliques (Si). One �g of total
RNA was used in each reaction. The PCR products were blotted onto
nylon membranes and hybridized with each labeled probe. As a con-
trol, the TUBULIN �9 (TUB9; Snustad et al. 1992) transcript was
amplified.
Fig. 8 Schematic representation of the inflorescence shoot architec-
ture of A. thaliana. (A–C) Overviews of the tips of inflorescences in
wild-type Columbia (A), er (B) and crm2 (C) plants of the same ages
are shown. Arrowheads indicate apical inflorescence meristems. Indi-
vidual flower buds are represented by circles.
CORYMBOSA2 in inflorescence development304
The crm2 phenotype suggests a tight continuity between
internode elongation and flower organ development in Arabi-
dopsis. However, considering the fact that growth of stem
internodes can occur without flower development in the leafy
mutant of Arabidopsis (Schultz and Haughn 1991) and that
some mutants have been identified with a defect in internode
elongation but not in flower organ development (Fridborg et al.
1999, Hanzawa et al. 2000), a causal relationship between the
delay in bolting and that in flower organ development is
unlikely. Reduction in the final length of an individual inter-
node in crm2-1 (Table 1; Fig. 3C, D) may reflect the effect of
the mutation on the elongation rate of stem internodes after
bolting (Fig. 2C). However, as a consequence of flower pro-
duction for a prolonged period, which may be related to
reduced fertility, the final stature of crm2 plants is comparable
to or higher than that of the wild type. Taken as a whole,
CRM2 is likely to be involved in the acceleration of growth of
flower organs and stem internodes. It is well known that gib-
berellins (GAs) play a critical role in many aspects of plant
development, including internode elongation and flower open-
ing, as a plant growth regulator (Pharis and King 1985).
Although crm2 plants do not exhibit pleiotropic phenotypes
caused by GA deficiency and the crm2 phenotype can not be
restored by exogenous application of GAs (data not shown),
suggesting that crm2 plants have normal levels of endogenous
GAs, the possibility cannot be excluded that CRM2 may be
involved in GA signal transduction pathways.
We cloned the CRM2 gene by a chromosome walk. Final
proof of identification of the CRM2 gene was obtained by com-
plementation of the crm2 mutant phenotype with the wild-type
gene. RT-PCR experiments revealed that the CRM2 transcripts
are present mainly in seedlings, roots and flower buds. We
could not detect any signals of the CRM2 transcript by in situ
hybridization analysis. Given the mutant phenotype, CRM2
may be expressed in shoot and root meristems, although the
crm2-1 allele showed normal growth of the root tissue (data not
shown). The crm2-1 allele contains a 3-bp deletion correspond-
ing to one amino acid that is located in a putative nuclear local-
ization signal (NLS) of the deduced protein. Expression of the
mutated transcript was detected in crm2-1 plants (data not
shown), suggesting that it may be translated. It remains to be
determined whether the CRM2 protein is indeed targeted to the
nucleus or not. If so, deletion of the NLS in crm2-1 might pre-
vent nuclear localization of the mutated CRM2. Whatever the
functional significance of the amino acid deleted in crm2-1 is,
the possibility that the mutated CRM2 protein still retains its
function to some degree and results in the observed phenotype
cannot be ruled out. A homologous sequence was found to be
present next to the CRM2 locus. This raises another possibility
that the growth of internodes and the development of flowers in
crm2-1 may be accomplished by redundant functions of this
homologous gene. Identification of additional mutant alleles of
CRM2 and its homolog will be of considerable importance in
defining the functions of these gene products during inflores-
cence development. Motif analysis of the CRM2 protein also
shows homology to the double-stranded RNA-binding motif at
the N-terminal region. Double-stranded RNA-binding proteins
are involved in diverse cellular functions and provide further
examples of post-transcriptional gene regulation by RNA-bind-
ing proteins (Burd and Dreyfuss 1994). It remains to be deter-
mined whether CRM2 binds double-stranded RNA or not.
In conclusion, our study revealed the importance of the
timing of bolting and the balance between the rate of flower
primordia initiation and that of flower opening in determining
the inflorescence morphology. Now that the CRM2 gene has
been cloned, it should provide a useful tool for further study
and genetic manipulation of the inflorescence shoot architec-
ture of other plant species.
Materials and Methods
Plant material and growth conditions
The original crm2-1 allele was identified among the M2 progeny
of fast neutron-irradiated seeds in the Columbia (Col) ecotype (Lehle
Seeds, Round Rock, TX, U.S.A.), as previously described (Komeda et
al. 1998). The mutant was backcrossed more than three times to the
wild type before analysis. For the er mutant, the er-105 allele in the
Col background (Torii et al. 1996) was used for phenotypic and
genetic analyses.
All plants were grown on rockwool bricks supplemented with a
vermiculite in a growth chamber at 22�C under illumination with
white fluorescent light. The photoperiodic cycle was 16 h light/8 h
dark of long-day conditions.
Microscopy
For light microscopy, tissues were fixed in FAA (45% [v/v] etha-
nol, 5% [v/v] formaldehyde, and 5% [v/v] acetic acid) overnight at
room temperature, dehydrated in a graded ethanol series, and embed-
ded in Technovit 7100 resin (Heraeus Kulzer GmbH, Wehrheim, Ger-
many). Sections, 5 �m thick, were cut on a microtome (Yamato-Koki,
Asaka, Japan), attached to slide glass, and stained for 30 s in a 0.1%
(w/v) toluidine blue solution. They were examined and photographed
using a microscope (Nikon, Tokyo, Japan).
For SEM, samples were fixed in FAA and dehydrated in a graded
ethanol series and isoamyl acetate. They were then critical-point-dried
with liquid CO2. The dried materials were mounted and coated with
gold-palladium in a sputter coater (Hitachi, Tokyo, Japan). Specimens
were viewed under a scanning electron microscope (JEOL, Tokyo,
Japan) with an accelerating voltage of 25 kV.
Positional cloning
The CRM2 locus was mapped on chromosome IV by segregation
analyses using cleaved amplified polymorphic sequence (CAPS) mark-
ers (Konieczny and Ausubel 1993) and simple sequence length poly-
morphism (SSLP) markers (Bell and Ecker 1994). Further fine mapping
experiments delimited the locus to BAC clones T13K14 (accession
number AL080282) and F21C20 (accession number AL080254). The
primers we designed for CAPS on these BAC clones (see Fig. 5A) and
restriction enzymes used were as follows: T13K14–10-MseI-1, 5�-
ATTCC AGCAA CATGA GTGCC-3�; T13K14–10-MseI-2, 5�-TCTGC
AGCTG CAGAT TCACG-3�; T13K14.80-HapII-1, 5�-TCTCC ATTTC
CTCTT TGACC-3�; T13K14.80-HapII-2, 5�-TAAGT AGGCT TTTCG
ACGAG-3�; T13K14.160-RsaI-1, 5�-ACATA CTCAA GAACA
TTAAG C-3�; T13K14.160-RsaI-2, 5�-ACCGG TCTTG AATTT
CORYMBOSA2 in inflorescence development 305
ATCAT CTC-3�; T13K14.180-DdeI-1, 5�-ATGGC GGCCA TGCAA
CTAAC-3�; T13K14.180-DdeI-2, 5�-TCTTC CTGTG ATCTT CCACG-
3�; F7J7.10-EcoRI-1 (accession number AL021960); 5�-TGTTA GTTAT
GCGGT GGAGG-3�; F7J7.10-EcoRI-2, 5�-AGCTG CAACT TATTCT
CCGG-3�; T6K22.80-EarI-1 (accession number AL031187), 5�-
CATGA ACTCA TTGGA ATCCC AA-3�; and T6K22.80-EarI-2, 5�-
GCCAT TGCAT ATTTG CATGT AC-3�.
Nucleotide sequences of putative open reading frames within
the T13K14 region in crm2-1 were determined by using an
ABIPRISM377 DNA Sequencer (Applied Biosystems/Perkin Elmer,
CA) and were compared to the wild-type sequences in GenBank using
the BLAST program (www.ncbi.nlm.nih.gov/BLAST/).
RNA extraction and RT-PCR analysis
Total RNA was extracted according to the SDS-phenol method
(Takahashi et al. 1992). RT-PCR was conducted by using a TaKaRa
RNA LA PCR Kit (AMV) Ver.1.1 (TaKaRa, Kyoto, Japan) with 1 �g of
total RNA. For the first-strand cDNA synthesis, an oligo (dT)-adaptor
primer was used. Primers for AP1 (Mandel et al. 1992), TFL1
(Ohshima et al. 1997), CRM2 (accession number AF327068),
T13K14.80 and TUB9 (Snustad et al. 1992) were as follows: AP1-1, 5�-
ATGGA GAATA TGGGA ACTAG-3�; AP1-2, 5�-TCATG CGGCG
GAGGA GCCA-3�; TFL1-1, 5�-ATGGA GAATA TGGGA ACTAG-3�;
TFL1-2, 5�-GACAA TCATG AAACT AGCGT-3�; CRM2-1, 5�-TATGT
TGCGG TATGT CACCC-3�; CRM2-2, 5�-TCTGG TAGAT ACTAC
TCTAG TCAAA G-3�; T13K14.80–1, 5�-TAAGT AGGCT TTTCG
ACGAG-3�; T13K14.80–2, 5�-TCTCC ATTTC CTCTT TGACC-3�;
TUB9-1, 5�-TCTTC ATATT CAAGG AGGTC AG-3�; and TUB9-2, 5�-
TTAGG CTTCT TCTTC TTCTT CG-3�. The amplification protocol for
AP1 and TFL1 was 25 cycles of 30 s at 94�C, 30 s at 50�C, and 3 min at
72�C. The PCR products were fractionated in 1.5% (w/v) agarose gels,
stained with ethidium bromide, and visualized using UV light.
RT-PCR conditions for CRM2, T13K14.80 and TUB9 were the
same as described above except for 14 cycles for CRM2, 16 cycles for
T13K14.80, and 12 cycles for TUB9. The PCR products were fraction-
ated in 0.8% (w/v) agarose gels, transferred to GeneScreen Plus (NEN
Life Science Products, Boston, MA, U.S.A.). Blots were hybridized at
42�C with probes labeled to high specific activity by random priming
with [�-32P]dCTP using BcaBEST Labeling Kit (Takara, Kyoto,
Japan), washed at 65�C twice each with 2� SSC and 1% (w/v) SDS for
20 min, and once with 0.1� SSC for 20 min.
Plant transformation
For transgenic complementation, a 6-kb fragment containing the
CRM2 genomic sequence was amplified by PCR using the wild-type
genomic DNA as a template and sequence-specific primers with incor-
porated restriction sites. LA Taq polymerase (Takara, Kyoto, Japan),
which has a high replication fidelity, was used for PCR. The amplified
DNA was cloned as a SalI-KpnI fragment into pUC18 and sequenced.
The binary Ti vector was made from pBI101 (Clontech, Palo Alto,
CA, U.S.A.) by replacing the XbaI–SacI �-glucuronidase (GUS) gene
fragment with the multicloning site of pUC18 and then introducing the
6-kb SalI–KpnI genomic fragment. The resulting construct was intro-
duced into wild-type plants (Col) by the floral dip protocol of Clough
and Bent (1998). Transformants were selected on MS medium contain-
ing 60 �g ml–1 kanamycin and subsequently crossed with crm2-1
mutant plants to see whether the mutant phenotype could be restored
by the introduced gene in the F2 generation. The F2 plants
homozygous for the crm2-1 allele were identified by PCR that specifi-
cally amplifies the endogenous CRM2 gene and the subsequent
sequence determination of PCR products.
For the antisense expression of CRM2, the GUS gene in the
binary vector pBI121 (Clontech) was replaced with an XbaI–SacI frag-
ment from the multicloning site of pUC18. Then, a 2-kb XbaI genomic
DNA fragment containing the 5� UTR of the CRM2 gene (see Fig. 5B)
was cloned in the antisense orientation with regard to the CaMV 35S
RNA promoter. The resulting construct was introduced into wild-type
plants (Col).
Acknowledgments
We thank Masashi Hashimoto for help with plant transformation
experiments and Masahiko Ozaki for help with mapping crm muta-
tions. This work was supported in part by a grant-in-aid from the Min-
istry of Education, Science and Culture of Japan.
References
Alvarez, J., Guli, C.L., Yu, X-H. and Smyth, D.R. (1992) terminal flower: a
gene affecting inflorescence development in Arabidopsis thaliana. Plant J. 2:
103–116.
Barton, M.K. and Poethig, R.S. (1993) Formation of the shoot apical meristem
in Arabidopsis thaliana: an analysis of development in the wild type and in
the shoot meristemless mutant. Development 119: 823–831.
Bell, C.J. and Ecker, J.R. (1994) Assignment of 30 microsatellite loci to the
linkage map of Arabidopsis. Genomics 19: 137–144.
Burd, C.G. and Dreyfuss, G. (1994) Conserved structures and diversity of func-
tions of RNA-binding proteins. Science 265: 615–621.
Clark, S.E., Jacobsen, S.E., Levin, J.Z. and Meyerowitz, E.M. (1996) The CLA-
VATA and SHOOT MERISTEMLESS loci competitively regulate meristem
activity in Arabidopsis. Development 122: 1567–1575.
Clark, S.E., Williams, R.W. and Meyerowitz, E.M. (1997) The CLAVATA1 gene
encodes a putative receptor kinase that controls shoot and floral meristem size
in Arabidopsis. Cell 89: 575–585.
Clough, S.J. and Bent, A.F. (1998) Floral dip: a simplified method for Agrobac-
terium-mediated transformation of Arabidopsis thaliana. Plant J. 16: 735–
743.
Fletcher, J.C., Brand, U., Running, M.P., Simon, R. and Meyerowitz, E.M.
(1999) Signaling of cell fate decisions by CLAVATA3 in Arabidopsis shoot
meristems. Science 283: 1911–1914.
Fridborg, I., Kuusk, S., Moritz, T. and Sundberg, E. (1999) The Arabidopsis
dwarf mutant shi exhibits reduced gibberellin responses conferred by overex-
pression of a new putative zinc finger protein. Plant Cell 11: 1019–1031.
Hanzawa, Y., Takahashi, T., Michael, A.J., Burtin, D., Long, D., Pineiro, M.,
Coupland, G. and Komeda, Y. (2000) ACAULIS5, an Arabidopsis gene
required for stem elongation, encodes a spermine synthase. EMBO J. 19:
4248–4256.
Hensel, L.L., Nelson, M.A., Richmond, T.A. and Bleecker, A.B. (1994) The fate
of inflorescence meristems is controlled by developing fruits in Arabidopsis.
Plant Physiol. 106: 863–876.
Komeda, Y., Takahashi, T. and Hanzawa, Y. (1998) Development of inflores-
cences in Arabidopsis thaliana. J. Plant Res. 111: 283–288.
Konieczny, A. and Ausubel, F.M. (1993) A procedure for mapping Arabidopsis
mutations using co-dominant ecotype-specific PCR-based markers. Plant J.
4: 403–410.
Laux, T., Mayer, K.F., Berger. J. and Jurgens, G. (1996) The WUSCHEL gene is
required for shoot and floral meristem integrity in Arabidopsis. Development
122: 87–96.
Laux, T. and Schoof, H. (1997) Maintaining the shoot meristem – the role of
CLAVATA1. Trends Plant Sci. 2: 325–327.
Mandel, M.A., Gustafson-Brown, C., Savidge, B. and Yanofsky, M.F. (1992)
Molecular characterization of the Arabidopsis floral homeotic gene
APETALA1. Nature 360: 273–277.
Nakai, K. and Kanehisa, M. (1992) A knowledge base for predicting protein
localization sites in eukaryotic cells. Genomics 14: 897–911.
Ohshima, S., Murata, M., Sakamoto, W., Ogura, Y. and Motoyoshi, F. (1997)
Cloning and molecular analysis of the Arabidopsis gene Terminal Flower1.
Mol. Gen. Genet. 254: 186–194.
Pharis, R.P. and King, R.W. (1985) Gibberellins and reproductive development
in seed plants. Ann. Rev. Plant Physiol. 36: 517–568.
Schultz, E.A. and Haughn, G.W. (1991) LEAFY, a homeotic gene that regulates
CORYMBOSA2 in inflorescence development306
inflorescence development in Arabidopsis. Plant Cell 3: 771–781.
Smyth, D.R., Bowman, J.L. and Meyerowitz, E.M. (1990) Early flower devel-
opment in Arabidopsis. Plant Cell 2: 755–767.
Snustad, D.P., Haas, N.A., Kopczak, S.D. and Silflow, C.D. (1992) The small
genome of Arabidopsis contains at least nine expressed �-tubulin genes.
Plant Cell 4: 549–556.
Takahashi, T., Naito, S. and Komeda, Y. (1992) Isolation and analysis of the
expression of two genes for the 81-kilodalton heat shock proteins from Arabi-
dopsis. Plant Physiol. 99: 383–390.
Torii, K.U., Mitsukawa, N., Oosumi, T., Matsuura, Y., Yokoyama, R., Whittier,
R.F. and Komeda, Y. (1996) The Arabidopsis ERECTA gene encodes a puta-
tive receptor protein kinase with extracellular leucine-rich repeats. Plant Cell
8: 735–746.
Weberling, F. (1989) Morphology of Flowers and Inflorescences. Cambridge
University Press, Cambridge.
(Received December 10, 2001; Accepted January 4, 2002)