Full text of DOI:10.1007/s004250100571

Planta +2001) 213: 943±952
DOI 10.1007/s004250100571
ORIGINAL ARTICLE
Sugihiro Ando á Yuka Sato
Shin'ichiro Kamachi á Shingo Sakai
Isolation of a MADS-box gene ꢀERAF17 ) and correlation
of its expression with the induction of formation of female ¯owers
by ethylene in cucumber plants ꢀCucumis sativus L.)
Received: 23 November 2000 / Accepted: 7 January2001 / Published online: 30 June 2001
Ó Springer-Verlag 2001
Keywords Cucumis +¯ower formation) á Ethylene á
Flower formation á MADS-box gene
Abstract Ethylene regulates sex expression in cucumber
+Cucumis sativus L.) plants. When the apices of
monoecious cucumber seedlings +cv. Shimoshirazu-jibai)
were treated with the ethylene-releasing compound,
ethephon, female ¯owers were induced at the nodes. To
clarifythe action of ethylene in the regulation of sex
expression, we attempted to isolate genes whose ex-
pression changed during induction of the formation of
female ¯owers at the apices of these cucumber plants
upon treatment with ethephon. Using the di€erential-
displaymethod, we identi®ed 20 clones +#1 to #20) that
re¯ected di€erences in the accumulation of transcripts in
apices treated or not treated with ethephon. Sequence
analysis of cDNA fragments revealed that the cDNA
#17 had the sequence of a MADS-box gene. We isolated
the full-length cDNA and showed that it included both a
MADS box and a K box, and the corresponding gene
was designated ERAF17. We examined the expression of
ERAF17 in the apices of cv. Shimoshirazu-jibai and in
those of a gynoecious cultivar +Rensei). In these
cultivars, the timing and levels of expression of the
ERAF17 transcript were correlated with the develop-
ment of female ¯owers. Induction of the synthesis of the
ERAF17 transcript byethephon occurred within 4 h of
the start of treatment and continued for 4 days at least.
Expression of ERAF17 at apices was localized in the
¯oral buds of the gynoecious cultivar, and expression
was maintained in female ¯owers thorough their devel-
opment. Our results suggest that the induction of the
formation of female ¯owers byethylene might be regu-
lated bythe expression of ERAF17 in ¯oral buds at the
apices of cucumber plants and that expression of this
gene might also be involved in the development of
female ¯owers.
Abbreviations ACC: 1-aminocyclopropane-1-carboxy-
late á AVG: aminoethoxyvinylglycine á cv.: cultivar á
PCR:polymerase chain reaction á RACE: rapid ampli®-
cation of cDNA ends
Introduction
Cucumber +Cucumis sativus L.) plants exhibit three
typical patterns of sex expression, being monoe-
cious, gynoecious, or hermaphroditic +Malepszy and
Niemirowicz-Szczytt 1991). Monoecious cultivars pro-
duce male ¯owers at the base of the main stem, then they
produce male and female ¯owers on the middle part of
the stem, and ®nallyfemale ¯owers are produced at the
top of the stem. Gynoecious cultivars produce only fe-
male ¯owers. However, immature ¯oral buds of cu-
cumber plants initiallyhave primordia of both stamens
and pistils, and unisexualityis established bythe stage-
speci®c arrest of pre-formed sex-organ primordia in a
particular whorl +Dellaporta and Calderon-Urrea 1993;
Grant et al. 1994). Although sex expression in ¯owers of
cucumber plants is determined genetically, ethylene has
been shown to induce the formation of female ¯owers,
while gibberellin induces the formation of male ¯owers
+Peterson and Anhder 1960). Treatment of cucumber
plants with gaseous ethylene +Iwahori et al. 1970) or with
ethephon +2-chloroethylphosphonic acid; McMurray
and Miller 1968; Rudich et al. 1969), which releases
ethylene in plant tissues, changes sex expression to
promote the formation of female ¯owers. Inhibition of
ethylene biosynthesis in gynoecious cucumber plants by
aminoethoxyvinylglycine +AVG) and inhibition of the
action of ethylene by AgNO3 causes a shift in sex ex-
pression from female to male +Beyer 1976; Adams and
Yang 1979; Atsmon and Tabbak 1979). In addition,
experiments involving combination treatments with
plant hormones and inhibitors suggest that ethylene acts
S. Ando and Y. Sato contributed equallyto this work
S. Ando á Y. Sato á S. Kamachi á S. Sakai +&)
Institute of Biological Science, Universityof Tsukuba,
Ibaraki 305-8572, Japan
E-mail: ssakai@sakura.cc.tsukuba.ac.jp
Fax: +81-298-534579

944
in a more direct manner than gibberellins to in¯uence paper, we report the isolation of a MADS-box gene
the sex of cucumber ¯owers +Yin and Quinn 1995). +ERAF17) that appears likelyto be related to the in-
Furthermore, a high correlation has been observed be- duction of the formation of female ¯owers byethylene
tween the evolution of ethylene and sex expression in and/or to the development of female ¯owers at the
cucumber plants +Rudich et al. 1972). The enzyme apices of cucumber seedlings.
1-aminocyclopropane-1-carboxylate +ACC) synthase
+EC 4.4.1.14) is a key regulatory enzyme in the ethylene
biosynthetic pathway +Yang and Ho€man 1984), and it
is encoded bya multigene family. Various cDNAs for
genes for ACC synthase that appear likely to be involved
in the regulation of sex expression in cucumber plants
have been isolated independentlyin two laboratories
+Trebitsh et al. 1997; Kamachi et al. 1997). Kamachi
et al. +1997, 2000) reported that both the timing and the
levels of expression of the CS-ACS2 transcript were
correlated with the development of female ¯owers at
plant nodes. Furthermore, theyreported that the timing
of the induction of expression of the CS-ACS2 gene at
the apex corresponded to the timing of the action of
ethylene in the induction of the ®rst female ¯ower at the
apex of individual gynoecious cucumber plants. Al-
though much information has been accumulated about
the role of ethylene in the formation of female ¯owers on
cucumber plants, the mechanisms of action of ethylene
in the regulation of sex expression remain to be
explored.
Studies of ¯oral homeotic mutants of Antirrhinum
and Arabidopsis led directlyto the isolation of the ®rst
members of the plant familyof MADS-box genes
+Sommer et al. 1990; Yanofskyet al. 1990). Further
studies of such mutants and of related members of the
MADS-box familyof genes resulted in the establishment
of a universal model for the speci®cation of ¯oral organ
identitywithin a hermaphroditic ¯ower. This model,
known as the ABC model, is based on the activities of
three classes of genes, most of which have been identi®ed
as MADS-box genes, in the determination of four types
of ¯oral organ. The ABC model and the genes respon-
sible for functions of A, B and C have been reviewed by
Ma +1994), Weigel and Meyerowitz +1994), and Yanof-
sky+1995). The various MADS-box genes encode
transcription factors with a stronglyconserved DNA-
binding domain, the MADS box, and a moderatelywell-
conserved domain called the K box, which is able to
form amphipathic helices that allow dimerization of
these transcription factors +Ma et al. 1991; Davies and
Schwarz-Sommer 1994). Three MADS-box genes were
isolated from cucumber plants as homologs of AGA-
MOUS +Filipecki et al. 1997; Perl-Treves et al. 1998;
Kater et al. 1998). Two of them were shown to mediate
function C in experiments with transgenic petunia +Ka-
ter et al. 1998). However, in analyses of these genes in
cucumber plants, no e€ects of treatments with ethylene
or gibberellin were detected +Perl-Treves et al. 1998).
Thus, the involvement of these genes in sex determina-
tion of cucumber ¯owers appears unlikely.
Materials and methods
Plant materials
A monoecious cultivar of cucumber +Cucumis sativus L., cv.
a gynoecious cultivar of cucumber
Shimoshirazu-jibai) and
+Cucumis sativus L., cv. Rensei) were grown in soil-®lled pots in a
phytotron at 25 °C with 14 h of light and at 20 °C with 10 h of
darkness daily. Apices were excised from seedlings just above the
youngest leaf at the indicated stages of growth, frozen immediately
in liquid nitrogen and stored at ±80 °C prior to extraction of
nucleic acids.
Treatments
The apices of monoecious cucumber seedlings +cv. Shimoshirazu-
jibai) were treated with 500 lM ethephon that contained 0.02%
+v/v) Tween 20 at the indicated stages of growth. The apices of
gynoecious cucumber seedlings +cv. Rensei) were treated with
100 lM AVG that contained 0.02% +v/v) Tween 20. In control
experiments, apices were treated with 0.02% Tween 20. Each apex
was treated byplacing a piece of absorbent cotton that had been
soaked in the appropriate solution on the apex of a growing cu-
cumber plant. Treatments were repeated on three consecutive days
at the indicated stages after planting. After further growth of
seedlings, the sex of each ¯ower on the ®rst 20 nodes was examined
and classi®ed as male or female. A node was designated male if it
had at least one male ¯ower and it was designated female if it had
onlyfemale ¯owers. After treatments, apices were excised from
some seedlings and frozen in liquid nitrogen for extraction of
nucleic acids.
Isolation of RNA
Total RNA was extracted from the apices of cucumber seedlings as
described byPrescott and Martin +1987), and then the RNA was
puri®ed bysuccessive precipitations in lithium chloride and in
ethanol for use in RNA gel blot analysis. For analysis by the dif-
ferential-displaymethod, the extracted total RNA was puri®ed by
ultracentrifugation at 200,000 g for 16 h on a cesium chloride
densitygradient. Then the puri®ed RNA was treated with DNase
+RQ1 RNase-Free DNase; Promega, Madison, Wis., USA) to re-
move anycontaminating DNA. For 5 ¢-RACE +rapid ampli®cation
of cDNA ends), poly+A)+RNA was isolated from the total RNA
after ultracentrifugation on a spun column +mRNA Puri®cation
kit; Amersham Pharmacia Biotech, Bucks., UK).
Di€erential-displaytechnique
The di€erential-displaytechnique was applied as described by
Liang et al. +1993). For each condition, four sets of duplicate
reactions were performed according to the manufacturer's protocol
+RNAmap, mRNA di€erential-displayssytem; GenHunter Co.,
Brookline, Mass., USA). Each reaction mixture +total volume,
60 ll) contained 0.6 lg of total RNA, one of the primers T12MA,
T12MT, T12MG, or T12MC +where M stands for a mixture of G, A,
and C), 20 lM dNTP mixture, 50 mM Tris-HCl +pH 8.3), 75 mM
KCl, 3 mM MgCl2 and 10 mM +‹)-dithiothreitol. Reaction mix-
tures were heated at 65 °C for 5 min and then at 37 °C for 10 min,
after which 150 units of Moloneymurine leukemia virus reverse
We have initiated a studyof ethylene-regulated genes
that might be related to sex expression in cucumber
plants using the di€erential-displaymethod. In this

945
transcriptase +StrataScript Reverse Transcriptase; StrataGene, La Isolation of full-length cDNA of ERAF17
Jolla, Calif., USA) were added. Reaction mixtures were incubated
at 37 °C for a further 50 min and then reactions were stopped by Full-length cDNA corresponding to ERAF17 was ampli®ed by5 ¢-
heating at 95 °C for 5 min. One-tenth of each mixture after reverse RACE. All reactions were performed with a Marathon cDNA
transcription +RT-mixture) was used as the template for polymer- Ampli®cation Kit bythe method described in the Protocol and
ase chain reaction +PCR) in a reaction mixture that contained Reference Manual from the manufacturer +CLONTECH Labora-
the corresponding T12MN primer in combination with one of 20 tories Japan, Tokyo, Japan). The cDNA was reverse-transcribed
arbitrary10-mer primers for each reaction.
from poly+A)+RNA that had been isolated from apices of 18-day-
Ampli®cation byPCR was performed using Taq polymerase old Shimoshirazu-jibai seedlings, treated with ethephon as de-
+Perkin-Elmer Japan Co., Urayasu, Japan) in 20 ll of a reaction scribed above, and it was ligated to the Marathon cDNA adaptor.
mixture that contained 2 ll of RT-mixture, 1 lM corresponding The 5¢-cDNA fragment was ampli®ed byPCR with an ERAF17-
T
12MN primer, 0.2 lM arbitrary10-mer, 2 lM dNTP, 0.2 ll speci®c primer +5¢-ACGGCTGTGCTCCTCTTGCAAAC-3¢), the
a-[32P]dCTP +111 TBq/mmol; ICN Biomedicals Inc., Costa Mesa, adaptor primer that was supplied with the kit, and the adaptor-
Calif., USA), 10 mM Tris-HCl +pH 8.3), 50 mM KCl, 1.5 mM ligated cDNA as the template. PCR was performed with TaKaRa
MgCl2 and 1 unit of AmpliTaq polymerase +Perkin-Elmer Japan Ex Taq +TAKARA SHUZO Co., Tokyo, Japan) in the presence of
Co.). The parameters for PCR were 40 cycles of heating at 94 °C TaqStart Antibody+CLONTECH). A ``hot start'' involved incu-
for 30 s, at 40 °C for 2 min and at 72 °C for 30 s, followed by bation at 94 °C for 1 min and was followed byPCR with 30 cycles
extension at 72 °C for 5 min in a GeneAmp PCR system +model of heating at 94 °C for 30 s, at 61 °C for 30 s and at 68 °C for
9600; Perkin-Elmer Japan Co.). Aliquots of duplicate reaction 4 min. The products of PCR were cloned into pCRII.
mixtures after PCR were subjected to electrophoresis on a 6%
polyacrylamide/8 M urea sequencing gel to separate ampli®ed
cDNAs.
We recognized several bands that di€ered both in intensityand
Sequencing of DNA
The cDNA that had been cloned into pCRII was sequenced with an
automated DNA sequencer +model 377; Perkin-Elmer) bythe di-
deoxysequencing method +Sanger et al. 1977) using a Taq Dye
Primer Cycle Sequencing kit +Perkin-Elmer). Nucleotide and amino
acid sequences were analyzed with GENETYX-MAC software,
version 10.1.1 +Software Development Co., Tokyo, Japan). Data-
bases were searched with NCBI BLAST +National Center for
BiotechnologyInformation; Basic Local Alignment Search Tool
system 2.0). The predicted amino acid sequence encoded by
ERAF17 and portions of representative MADS-box genes from
di€erent plants that included the same part of the sequence +the
MADS fragment and the adjacent 110 amino acids) were used for
phylogenetic analysis as described by Theissen et al. +1996). A
neighbor-joining tree was produced from the results of 1,000
bootstrap replicates using the CLUSTAL W program, version 1.8,
of the DNA Data Bank of Japan +Mishima, Japan). The phylo-
genetic tree was displayed by TreeView software version 1.6 +Page
1996).
mobility, and we analyzed bands of cDNAs that were di€erentially
ampli®ed in a consistent manner in further detail. Regions of a gel
that contained di€erentiallyexpressed cDNAs were excised from
the dried gel and placed individuallyin 100 ll of distilled water.
The cDNAs that di€used from the gel fragments were re-ampli®ed
with the appropriate pair of primers. The primers used in the ex-
periment that led to the isolation of cDNA #17, reported in this
paper, were T12MT +5¢-TTTTTTTTTTTTMT-3¢) and AP-17
+5¢-GCAAGGAGTC-3¢). The ampli®ed fragment was puri®ed
from a 5% polyacrylamide gel. The puri®ed product of PCR was
cloned into the pCRII vector +Invitrogen Co., San Diego, Calif.,
USA) bythe method described in the TA Cloning Instruction
Manual from Invitrogen.
Preparation of the cDNA probe
cDNA was cleaved by EcoRI from pCRII that had been ampli®ed
in Escherichia coli JM109, puri®ed bygel electrophoresis and re-
covered. In all of the present experiments, the 3¢ sequence of cDNA
#17 that had been obtained bythe di€erential-displaymethod was
used as the cDNA probe. The cDNA was labeled with a-[32P]dCTP
bythe random-priming method with a Multiprime DNA Labeling
System +Amersham Pharmacia Biotech) for use as the probe.
Results
Sex conversion in ¯owers byethephon
For identi®cation of ethylene-responsive genes that
might be related to the induction of the formation of
female ¯owers byethephon, we examined the conditions
under which ethephon e€ectivelyinduced the conversion
of ¯owers from male to female, using the monoecious
cultivar Shimoshirazu-jibai as plant material. As shown
RNA gel blot analysis
Total RNA was isolated from the apices of the two cultivars +cv.
Shimoshirazu-jibai and Rensei) at the indicated stages after
planting. Total RNA was subjected to electrophoresis on a 1.17%
agarose gel that contained 0.66 M formaldehyde and then trans-
ferred to a GeneScreen Plus membrane +NEN Life Science Prod- in Fig. 1, Shimoshirazu-jibai exhibited stable maleness at
ucts, Boston, Mass., USA) bycapillaryaction in 10 ´SSC +1´SSC is
0.15 M NaCl, 15 mM sodium citrate), as recommended bythe
manufacturer of the membrane. After baking at 80 °C, the mem-
brane was incubated in 1 M NaCl, 1% SDS and 10% dextran
sulfate +sodium salt) at 60 °C for 1 h. The denatured 32P-labeled
probe and denatured salmon-sperm DNA were then added to the
pre-hybridization solution, and the membrane was incubated at
60 °C for 18 h. Post-hybridization washes were performed once for
2 min with 2´SSC at room temperature and twice, successively, for
15 min each with 2´SSC, 1% SDS at 60 °C. The washed membrane
was subjected to autoradiographywith an intensifying screen or the
BioImaging Analyzer +BAS 5000; Fuji Photo Film Co., Tokyo,
Japan). Then the membrane was washed with boiling 1´SSC, 0.1%
SDS to remove the probe and blots were re-hybridized with
ribosomal RNA to ensure that equal amounts of total RNA had
been loaded in each lane.
the lower nodes of the main stem and no female ¯owers
were formed on the lower nodes. The apices of Shimo-
shirazu-jibai were treated with 500 lM ethephon, an
ethylene-releasing compound, at various stages of
growth for 3 days. As shown in Fig. 1, no female ¯owers
were induced on plants that were treated with ethephon
6 days +cotyledon to one-leaf stage) after planting, as
was the case on control plants. We could con®rm no
¯ower formation at the apices of 6-day-old plants under
the light microscope +data not shown). When the apices
of 11-day-old plants +one- to two-leaf stage) were treated
with ethephon, female ¯owers were induced on lower
nodes, including the ®rst ¯ower. This result indicated

946
di€erential-displayreactions were performed using
combinations of 20 arbitrary10-mers and 4 anchor
primers. We detected 92 di€erentiallyexpressed bands of
cDNA +promoted, 62; suppressed, 30) in a series of ex-
periments. We performed RNA gel blot analysis with
same samples of RNA as those used for the di€erential
displayto remove false-positive clones. As the result, we
identi®ed 20 clones +#1 to #20) for which there was a
clear di€erence in the accumulation of transcripts in
apices treated with or without ethephon, as determined
byRNA gel blot analysis +data not shown).
Isolation and sequence analysis
of full-length cDNA #17
We isolated full-length cDNA #17 by5 ¢-RACE.
Figure 2A shows the complete nucleotide sequence of the
cloned cDNA and the amino acid sequence deduced from
the nucleotide sequence. The cDNA was 819 bp long and
contained an open reading frame of 609 bp that encoded a
putative polypeptide of 203 amino acids. A homology
search revealed that the cDNA corresponded to a gene in
the MADS-box gene family, having a MADS box and a K
box +Fig. 2A, B). This gene was designated ERAF17
+ethylene-responsive gene associated with the formation
of female ¯owers 17; GenBank accession number
AB046596). The deduced amino acid sequence encoded
by ERAF17 was found to be approximately50% ho-
mologous to that encoded by TM8, which is a gene for a
MADS-box protein in tomato +Lycopersicon esculentum;
Fig. 1 The timing of the conversion of ¯owers of cucumber
+Cucumis sativus) from male to female upon treatment of apices
with ethephon. Apices of seedlings +cv. Shimoshirazu-jibai) were
treated with or without ethephon +500 lM) for 3 days from the
indicated dayafter sowing. After further growth of seedlings, the
sex of each ¯ower on the ®rst 20 nodes was examined and classi®ed
as male or female. The node number indicates the position of
individual nodes along the main stem. Open circles Nodes with
male ¯owers, closed circles nodes with female ¯owers, no circles
vegetative nodes. Data from three plants are presented in each case
that the ¯ower formation was initiated up to node 7 at Pnueli et al. 1991). The genomic southern analysis re-
the apices of 11-day-old plants and the sex of ¯ower buds vealed that the ERAF17 gene was present as a single copy
was changed byethephon. The apices of cucumber
in both monoecious and gynoecious cucumber plants and
plants were also treated with ethephon at later growth that the probe used in this experiment was speci®c for the
stages. In the case of 15-day-old plants +two- to three-leaf ERAF17 gene +data not shown).
stage) and 19-day-old plants +three- to four-leaf stage),
the plants produced female ¯owers on higher nodes than
did the plants treated with ethephon 11 days after
of the familyof MADS-box genes in plants
Phylogenetic relationships to other members
planting. These results indicated that ¯oral buds at nodes
2±13 and nodes 2±20 or more were present at the nodes
The familyof MADS-box genes is composed of several
of 15-day-old and 19-day-old plants, respectively. Fur-
de®ned clades of genes whose members have similar
thermore, these results indicated that the sex of ¯ower
respective patterns of expression and stronglyrelated
buds on nodes 2±6 and 2±13 at the apices of 15-day-old
functions, as revealed by phylogenetic analysis +Doyle
and 19-day-old plants, respectively, had already been
1994; Purugganan et al. 1995; Theissen et al. 1996, 2000).
determined, while that of ¯ower buds on higher nodes
Therefore, we performed cluster analysis of ERAF17
di€erentiating at the apices had not yet been determined
and representative MADS-box genes. Our phylogenetic
and could be changed bytreatment with ethephon.
analysis indicated that ERAF17 did not belong to any
distinguishable clade of anysubfamilyof MADS-box
genes that have been included as participants in the ABC
model +Fig. 3). The most similar gene to ERAF17 was
TM8 from tomato +Pnueli et al. 1991).
Di€erential display
Seedlings at the two- to three-leaf stage +15 days old)
were chosen for treatment with ethephon because of the
e€ective induction byethephon of female ¯owers at this The e€ects of ethephon and AVG on the expression
stage. The apices of 15-day-old seedlings were treated or of ERAF17 and the sex of ¯owers
not treated with ethephon for 3 days and apices were cut
from seedlings the dayafter the last treatment. Total Figure 4 shows the expression of ERAF17 tran-
RNA was extracted from each sample and 80 sets of scripts in the apices of the monoecious cultivar

947
Fig. 3 Dendrogram of MADS-box genes from various plants. The
tree is based on the ``MADS + 110 domain'' +Theissen et al. 1996),
which represents the complete MADS-domain and the immediately
adjacent I- and K-domains. All known MADS-box genes of
Arabidopsis and cucumber, representative MADS-box genes of
snapdragon, tomato, maize, sorrel and white campion, and ®ve
genes of the other species that fell near clades of ERAF17 are
included in the phylogenetic tree. Horizontal branch lengths are
proportional to the estimated number of amino acid +a.a.)
substitutions per residue +bar = 0.1 a.a. substitution per residue).
Names of species from which the respective genes were isolated are
given in parenthesis next to the names of proteins. ERAF17 is
indicated byan open box. The numbers next to some nodes give
bootstrap percentages, which are shown onlywhen above 50%.
The names of sub-families that include participants in the ABC
model are indicated on the right
Fig. 2 A Nucleotide sequence of ERAF17 from cucumber and the
deduced amino acid sequence. The putative ERAF17 protein is
shown in the one-letter amino acid code. The MADS box is
represented byan open box and the K box is represented bya gray
box. Within the K box, hydrophobic amino acids +L, I, V, and M)
are indicated byan asterisk. The speci®c primer for RACE-PCR is
indicated byan arrow. The nucleotide sequence that was isolated by
di€erential displayis underlined. This region was used as the cDNA
probe for RNA and DNA gel blot analysis. B Alignment of the
amino acid sequences of MADS domains encoded byvarious
MADS-box genes. Amino acids identical to those in a consensus
sequence based on ERAF17 are shaded in black. The amino acid
sequences of the following MADS domains are aligned: ERAF17
+this study); Arabidopsis AG +GenBank accession number,
S10933), AP1 +Z16421), AP3 +A42095), and PI +D30807); tomato
TM8 +X60760); and cucumber CUM10 +AF035439), CUM1
+AF035438), and CUS1 +X97801)
of ethephon treatment. However, the expression of
ERAF17 was induced in 11-, 15- and 19-day-old plants
byethephon treatment +Fig. 4). The induction of the
synthesis of ERAF17 mRNA byethephon was corre-
lated with the induction byethephon of female ¯owers
on the stems of Shimoshirazu-jibai plants +Fig. 1).
However, a weak expression of ERAF17 was detected in
+Shimoshirazu-jibai) after treatment in the presence or 6-day-old plants. It seems that there are a few female
absence ethephon at the indicated stages of growth. The ¯owers formed in some apices of 6-day-old plants.
transcript was not detectable in the apices of Shimo-
We treated the apices of gynoecious plants +cv.
shirazu-jibai plants at anystage examined in the absence Rensei), which formed onlyfemale ¯owers, with a

948
solution of AVG, an inhibitor of ethylene synthesis, at
two stages of growth. As shown in Fig. 5A, no male
¯owers were induced on plants that had been treated
with AVG 9 days +cotyledon to one-leaf stage) after
planting, as was the case on control plants. When the
apices of 18-day-old plants +two- to three-leaf stage)
were treated with AVG, the sex of ¯owers on lower
nodes was unchanged but plants produced male ¯owers
on higher nodes. These results are explained bythe facts
that the sex of ¯ower buds on the lower nodes had al-
readybeen determined and that on higher nodes had not
yet been determined and could be changed with AVG as
the results of ethephon treatment on Shimoshirazu-jibai
Fig. 4 Stage-dependent accumulation of ERAF17 mRNA in the +Fig. 1). No expression of ERAF17 was detected in the
apices of monoecious cucumber and the e€ects of ethephon on such
accumulation. The apices of seedlings +cv. Shimoshirazu-jibai) were
on such expression +Fig. 5B). These results are explained
treated in the presence or absence ethephon +500 lM) for 3 days
apices of 9-day-old Rensei plants and AVG had no e€ect
bythe observation that no ¯ower formation could be
from the indicated dayafter sowing. The apices were collected the
con®rmed at this stage under the light microscope.
When the apices of 18-day-old plants were treated with
AVG, the expression of ERAF17 was inhibited com-
pared to that in control plants. These results showed
that the conversion of ¯owers from female to male by
AVG was correlated with inhibition of the expression of
ERAF17 in apices.
dayafter the third treatment in each case. Total RNA +20 lg) was
extracted from apices, fractionated on a denaturing agarose gel,
transferred to a nylon membrane, and allowed to hybridize with the
ERAF17 probe. The blot was re-hybridized with an rRNA probe to
con®rm loading of equal amounts of RNA in each lane
Fig. 5A The e€ects of AVG on sex expression. Apices of
cucumber seedlings +cv. Rensei) were treated or not treated with
AVG +100 lM) for 3 days from the indicated day after sowing.
After further growth of seedlings, the sex of each ¯ower on the ®rst
20 nodes was examined and classi®ed as male or female. The node
number indicates the position of individual nodes along the main
stem. Closed circles Nodes with female ¯owers, open circles nodes
with male ¯owers, no circles vegetative nodes. Data from three
plants are presented in each case. B Stage-dependent accumulation
of ERAF17 mRNA in gynoecious cucumber plants +cv. Rensei) and
the e€ects of AVG on such accumulation. Apices of seedlings +cv.
Rensei) were treated or not treated with AVG +100 lM) for 3 days
from the indicated dayafter sowing. Apices were collected the day
after the last treatment. Total RNA +20 lg) was extracted from
apices, fractionated on a denaturing agarose gel, transferred to a
nylon membrane, and allowed to hybridize with the ERAF17
probe. The blot was re-hybridized with an rRNA probe as
described in the legend to Fig. 4
Time course of the induction of expression
of ERAF17 byethephon
In the above experiment +Fig. 4), we examined the ex-
pression of ERAF17 in the apices of Shimoshirazu-jibai
plants using plants that had been treated or not treated
with ethephon for three successive days. Therefore, we
next examined the time course of the expression of
ERAF17 transcripts in the apices of monoecious cu-
cumber +cv. Shimoshirazu-jibai) after a single treatment
with ethephon. As shown in Fig. 6A, B, the ERAF17
transcript was induced byethephon 4 h after the start of

949
Fig. 6A±D Time course of ex-
pression of ERAF17 transcripts
at the apices of cucumber cv.
Shimoshirazu-jibai seedlings af-
ter treatment with ethephon. A
short time-course +A) and a
long time-course +C) are shown.
Results of quanti®cation of da-
ta using Science Lab 98 Image
Gauge +Fuji Photo Film Co.,
Tokyo, Japan; software version
3.1) are shown in B and D,
respectively. Ethephon
+500 lM; containing 0.1%
Tween 20) or 0.1% Tween 20
was used to treat Shimoshirazu-
jibai apices 18 days after plant-
ing. Apices were harvested at
the indicated times after treat-
ment. Total RNA +20 lg) was
extracted from apices, fraction-
ated on a denaturing agarose
gel, transferred to a nylon
membrane, and allowed to hy-
bridize with the ERAF17 probe.
The blot was re-hybridized with
an rRNA probe as described in
the legend to Fig. 4
a single treatment, but it was not detected in controls. extracted from ¯owers at various stages. As shown
Furthermore, the level of the expression remained high in Fig. 8, transcripts of ERAF17 were barelydetectable
for at least 4 days after the treatment +Fig. 6C, D).
in male ¯owers of Shimoshirazu-jibai plants through-
out development. Bycontrast, strong expression of
ERAF17 was detected in female ¯owers of Rensei plants
throughout development as far as anthesis.
Localization of the expression of ERAF17 mRNA
at the apices of cucumber plants
Expression of ERAF17 in ¯oral organs
As shown in Figs. 1 and 5A, the monoecious cultivar
Shimoshirazu-jibai produced onlymale ¯owers and the
gynoecious cultivar Rensei produced only female ¯owers
on its lower nodes. Since ¯ower buds were formed at the
apices of the cucumber plants, we used these two cultivars
to examine the sites of expression of the ERAF17 tran-
script in apices. As shown in Fig. 7A, expression was
detected in apices onlyof ethephon-treated Shimoshir-
azu-jibai plants and in those of Rensei plants treated or
not treated with ethephon. The apices included ¯oral buds
and unexpanded leaves and we separated those tissues
under a light microscope and then extracted the total
RNA from each tissue. No expression of the ERAF17
transcript was detected in unexpanded leaves from the
apices of either Shimoshirazu-jibai or Rensei plants. The
ERAF17 transcript was also not detected in the ¯oral buds
in the apices of Shimoshirazu-jibai plants, but the tran-
script was detected in those of Rensei plants +Fig. 7B).
We next examined the expression of ERAF17 in the
¯oral organs of mature ¯owers +harvested at anthesis
and the preceding day) by northern blot analysis. As
shown in Fig. 9, the expression of ERAF17 was only
detected in sepals, petals and ovaries of female ¯owers,
with the strongest signal being detected in sepals, and an
expression in the ovaries was stronger than in the petals.
Discussion
In the present study, we isolated a cDNA clone for a
gene designated ERAF17, whose expression was induced
bytreatment with ethephon, an ethleyne-releasing
compound +Fig. 4), in a monoecious cucumber plant.
ERAF17 was stronglyhomologous to other MADS-box
genes, being most similar to the TM8 gene from tomato.
Phylogenetic analysis showed that ERAF17 was distinct
from ¯oral homeotic MADS-box genes +Fig. 3). Ac-
cording to the phylogenetic tree +Fig. 3), it seems that
ERAF17 and TM8 belong to the same subfamily. The
TM8 gene has been considered an orphan gene or a new
subfamily+Doyle 1994; Purugganan et al. 1995; Theissen
Expression of ERAF17 in ¯owers during development
In order to examine the stage-speci®c expression of the
ERAF17 gene during ¯oral development, we performed
RNA gel blot analysis using total RNA that had been

950
Fig. 7A, B Tissue-speci®c expression of ERAF17 and responsive- et al. 1996). It was reported that TM8 is expressed at an
ness to ethylene. A Whole cucumber plants were treated or not
treated with ethephon +500 lM) byspraying and byapplication of
developed ¯owers but the function of TM8 is not known
a piece of absorbent cotton that had been soaked in the appropriate
earlystage of ¯ower development and is repressed in
+Pnueli et al. 1991). Furthermore, no orthologs to TM8
solution on apices 1 day+17 days after sowing) before harvesting.
from other plants have been characterized. We could not
judge whether ERAF17 and TM8 have anycommon
features from our phylogenetic tree alone. In this anal-
ysis, while we did not obtain any information about the
actual role of the product of ERAF17, we were able to
show that ERAF17 belongs to a new categoryof
MADS-box genes.
Total RNA was extracted from apices, unexpanded +small) leaves,
expanded +large) leaves and stems of 18-day-old seedlings of
Shimoshirazu-jibai and Rensei. B Unexpanded leaves and ¯oral
buds were separated from each apex of 18-day-old seedlings +three-
leaf stage) of Shimoshirazu-jibai and Rensei under
a light
microscope. Floral buds were excised from the fourth to eighth
nodes from the apices. Floral buds were 2±4 mm or 1.5±3 mm long,
respectively. Total RNA +20 lg in A and 10 lg in B) was extracted
from each tissue, fractionated on a denaturing agarose gel,
transferred to a nylon membrane, and allowed to hybridize with
the ERAF17 probe. The blot was re-hybridized with an rRNA
probe +see legend to Fig. 4)
To examine the relationship between ERAF17 and
sex expression, we performed RNA gel blot analysis. As
shown in Fig. 4, the expression of ERAF17 at apices of
the monoecious cultivar Shimoshirazu-jibai depended
on ethephon treatment and on the growth stage of the
plant. The timing of changes in the level of the ERAF17
transcript appeared to coincide with the timing of the
action of ethylene in the formation of female ¯owers
+Fig. 1). When AVG, an inhibitor of ethylene biosyn-
thesis, was used to treat the apices of the gynoecious
cultivar Rensei, expression of ERAF17 was suppressed
during the induction of the formation of male ¯owers
+Fig. 5A, B). Moreover, ERAF17 was expressed onlyin
the ¯oral buds of the gynoecious cultivar +Fig. 7A, B).
Recently, Kamachi et al. +2000) reported that the
CS-ACS2 transcript is expressed onlyin a limited
number of ¯oral buds that will develop into female
¯owers. These results suggest that expression of
ERAF17 is stronglycorrelated with induction of the
formation of female ¯owers byethylene. Furthermore,
the ERAF17 transcript was detected 4 h after a single
treatment with ethephon +Fig. 6A, B). In cucumber
¯owers, morphological changes of ¯oral buds to female
¯owers are observed in ovarydevelopment +Atsmon and
Galun 1960). However, this morphological change was
undetectable in any¯oral buds at this time under the
Fig. 8 Expression of ERAF17 during ¯oral development. Floral
buds of cucumber cv. Shimoshirazu-jibai +male) and Rensei
+female) were collected and grouped according to respective size
and external markers +yc corolla becoming yellow) and anthesis.
The apices of the two cultivars were also collected 18 days after
sowing. Total RNA +12 lg) was extracted from each sample,
fractionated on a denaturing agarose gel, transferred to a nylon
membrane, and allowed to hybridize with the ERAF17 probe. The
blot was re-hybridized with an rRNA probe +see legend to Fig. 4) light microscope +data not shown). Therefore, the

951
pion +Silene latifolia; Ainsworth et al. 1995, and
Hardenack et al. 1994, respectively). In sorrel, the
transcripts of B-function genes, RAD1 and RAD2,
which were expressed in the stamen whorl, were absent
from arrested stamen primordia in females. The tran-
script of the C-function gene RAP1 was expressed in the
inner two whorls in veryyoung ¯ower primordia of both
males and females. However, such expression was not
detected in arrested fourth-whorl regions of male ¯owers
and in the stamen primordia of female ¯owers. Although
such coincidence might suggest a causal relationship, it
is still impossible to determine whether the decrease in
level of the transcript caused arrest or vice versa
+Ainsworth et al. 1995). In white campion, primordia of
opposite sex develop to a more advanced stage, and no
di€erences between male and female ¯owers were
detected in terms of the expression of SLM1 +a PLENA
homolog) from the initiation of ¯owers to meiosis
+Hardenack et al. 1994). In the case of monoecious
plants, it was reported that ZMM2, a putative C-class
gene in maize, was expressed in stamen and carpel pri-
mordia throughout their development, even though
stamen primordia abort in female in¯orescences and
Fig. 9 Expression of ERAF17 in ¯oral organs. Floral organs of
male ¯owers +sepals, petals and stamens; cv. Shimoshirazu-jibai)
and of female ¯owers +sepals, petals, pistils and ovaries; cv. Rensei)
were collected from mature ¯owers +harvested at anthesis and
preceding day). Total RNA +12 lg) was extracted from each
sample, fractionated on a denaturing agarose gel, transferred to a
nylon membrane, and allowed to hybridize with the ERAF17
probe. The blot was re-hybridized with an rRNA probe +see legend
to Fig. 4)
carpel primordia abort in male ones +Cacharron et al.
Â
1999). Furthermore, some homologs of AGAMOUS
genes were also isolated from cucumber plants and an-
alyzed +Perl-Treves et al. 1998). However, they have not
been implicated as genes involved in the sex-expression
cascade because of their insensitivityto hormonal
stimuli. These various observations implythat MADS-
box genes that are involved in the ABC model are not
involved in sex di€erentiation in these species.
Cluster analysis of ERAF17 and other MADS-box
genes indicated that ERAF17 belongs to a clade distin-
guishable from that of genes in the ABC model +Fig. 3).
Moreover the induction of the accumulation of its
transcript was detected soon +4 h) after ethephon treat-
ment +Fig. 6A, B). These features are clearlydi€erent
from those of other MADS-box genes from cucumber.
The expression of ERAF17 was correlated with the
formation of female ¯owers and, thus, it is possible that
ERAF17 might act in a feminizing cascade. In cucumber
plants, ethylene might control sex di€erentiation via the
direct or indirect regulation of the expression of
ERAF17. We recentlyfound that the sequence of a
partial cDNA of the CUS3 gene, isolated from embryo-
genic cucumber callus +Filipecki 2000; published onlyin
the database, GenBank accession number AJ278013), is
identical to the sequence of part of ERAF17.
``early'' expression of ERAF17 appears not to be a result
of the development of female ¯owers but represents a
factor that endows femaleness after exposure of pri-
mordia to ethylene. The expression of ERAF17 re-
mained at a high level at least for 4 days after a single
treatment with ethephon +Fig. 6C, D), and the transcript
was also expressed in female ¯ower buds throughout
their development +Fig. 8). It is possible that the role of
``late'' expression might be di€erent from that of ``early''
expression. It is well established that the expression of
MADS-box genes persists at later stages, when their
products perform developmental, or maintenance func-
tions +Bowman et al. 1991; Schwarz-Sommer et al.
1992). Moreover, the patterns of expression of several
MADS-box genes suggest that theyare involved in
embryogenesis and/or in the development of the ovule or
fruit +Angenent et al. 1995; Colombo et al. 1995; Heck
et al. 1995; Flanagan et al. 1996; Filipecki et al. 1997).
Thus, late expression of ERAF17 might be involved in
the development of female ¯owers and this possibilityis
supported bythe expression of ERAF17 in ovaries
+Fig. 9). Interestingly, the expression of ERAF17 was
detected in sepals and petals of female ¯owers but not in
those of male ¯owers +Fig. 9); however, we could not
explain this di€erence between male and female ¯owers.
The possibilitythat a MADS-box gene might control
sex di€erentiation has been considered in the past.
Various studies of class-B and class-C genes +in the ABC
model) that are expressed in reproductive organs have
been undertaken in dioecious and monoecious plants.
Researchers have examined whether MADS-box genes
that control the development of reproductive organs are
expressed in arrested organ primordia in two dioecious
plants, namely, sorrel +Rumex acetosa) and white cam-
References
Adams DO, Yang SF +1979) Ethylene biosynthesis: identi®cation
of 1-aminocyclopropane-1-carboxylic acid as an intermediate in
the conversion of methionine to ethylene. Proc Natl Acad Sci
USA 76:170±174
Ainsworth C, CrossleyS, Buchanan-Wollaston V, Thangavelu M
+1995) Male and female ¯owers of the dioecious plant sorrel
show di€erent patterns of MADS box gene expression. Plant
Cell 7:1583±1598

952
Angenent GC, Franken J, Busscher M, van Dijken A, van Went MalepszyS, Niemirowicz-Szczytt K +1991) Sex determination in
JL, Dons HJM, van Tunen AJ +1995) A novel class of MADS
box genes is involved in ovule development in petunia. Plant
Cell 7:1569±1582
cucumber +Cucumis sativus) as a model system for molecular
biology. Plant Sci 80:39±47
McMurrayAL, Miller CH +1968) Cucumber sex expression mod-
i®ed by2-chloroethanephosphonic acid. Science 162:1397±1398
pistillate and hermaphrodite ¯owers in Cucumis sativus +L.). Page RDM +1996) TREEVIEW: an application to displayphylo-
Atsmon D, Galun E +1960) A morphogenetic studyof staminate,
Phytomorphology 10:110±115
Atsmon D, Tabbak C +1979) Comparative e€ects of gibberellin,
genetic trees on personal computers. Comput Appl Biosci
12:357±358
silver nitrate and aminoethoxyvinyl glycine on sexual tendency Perl-Treves R, Kahana A, Rosenman N, Xiang Y, Silberstein L
and ethylene evolution in the cucumber plant +Cucumis sativus
L.). Plant Cell Physiol 20:1547±1555
Beyer EJ +1976) Silver ion: a potent antiethylene agent in cucumber
and tomato. HortScience 11:195±196
Bowman JL, Drews GN, Meyerowitz EM +1991) Expression of the
Arabidopsis homeotic gene AGAMOUS is restricted to speci®c
cell types late in ¯ower development. Plant Cell 3:749±758
+1998) Expression of multiple AGAMOUS-like genes in male
and female ¯ower of cucumber +Cucumis sativus L.) Plant Cell
Physiol 39:701±710
Peterson CE, Anhder LD +1960) Induction of staminate ¯owers on
gynoecious cucumbers with gibberellin A3. Science 131:1673±
1674
Pnueli L, Abu-Abeid M, Zamir D, Nacken W, Schwarz-Sommer Z,
Lifschitz E +1991) The MADS box gene familyin tomato:
temporal expression during ¯oral development, conserved sec-
ondarystructures and homologywith homeotic genes from
Antirrhinum and Arabidopsis. Plant J 1:255±266
Cacharron J, Saedler H, Theissen G +1999) Expression of MADS
Â
box genes ZMM8 and ZMM14 during in¯orescence develop-
ment of Zea mays discriminates between the upper and lower
¯oret of each spikelet. Dev Genes Evol 209:411±420
Colombo L, Franken J, Koetje E, van Went JL, Dons HJM, Prescott A, Martin C +1987) A rapid method for the quantitative
Angenent GC, van Tunen AJ +1995) The petunia MADS box assessment of levels of speci®c mRNAs in plants. Plant Mol
gene FBP11 determines ovule identity. Plant Cell 7:1859±1868 Biol Rep 4:219±224
Davies B, Schwarz-Sommer Z +1994) Control of ¯oral organ identity Purugganan MD, RounsleySD, Schmidt RJ, YanofskyMF +1995)
byhomeotic MADS-box transcription factors. In: Nover L +ed)
Results and problems in cell di€erentiation 20. Plant promoters
and transcription factors. Springer, Berlin, pp 235±258
Molecular evolution of ¯ower development: diversi®cation of
the plant MADS-box regulatorygene family. Genetics 140:345±
356
Dellaporta SL, Calderon-Urrea A +1993) Sex determination in Rudich J, HalevyAH, Kedar N +1969) Increase in femaleness of
¯owering plants. Plant Cell 5:1241±1251
Doyle JJ +1994) Evolution of a plant homeotic multigene family:
three cucurbits bytreatment with ethrel, an ethylene releasing
compound. Planta 86:69±76
toward connecting molecular systematics and molecular devel- Rudich J, HalevyAH, Kedar N +1972) Ethylene evolution from
opmental genetics. Syst Biol 43:307±328
cucumber plants as related to sex expression. Plant Physiol
49:998±999
CUS1 is expressed during cucumber somatic embryogenesis. Sanger FS, Nicklen S, Coulson AR +1977) DNA sequencing with
Filipecki MK, Sommer H, MalepszyS +1997) The MADS-box gene
Plant Sci 125:63±74
Flanagan CA, Hu Y, Ma H +1996) Speci®c expression of the AGL1
chain-terminating inhibitor. Proc Natl Acad Sci USA 74:5463±
5467
MADS-box gene suggests regulatoryfunctions in Arabidopsis Schwarz-Sommer Z, Hue I, Huijser P, Flor PJ, Hansen R, Tetens
gynoecium and ovule development. Plant J 10:343±353
Grant S, Houben A, Vyskot B, Siroky J, Pan WH, Macas J, Sae-
dler H +1994) Genetics of sex determination in ¯owering plants.
Dev Genet 15:214±230
F, Lonnig W-E, Saedler H, Sommer H +1992) Characterization
È
of the ¯oral homeotic MADS-box gene de®ciens: evidence for
DNA binding and autoregulation of its persistent expression
throughout ¯ower development. EMBO J 11:1687±1688
Hardenack S, Ye D, Saedler H, Grant S +1994) Comparison of Sommer H, Beltran J-P, Huijser P, Pape H, Lonnig W-E, Saedler
Â
È
MADS box gene expression in developing male and female
¯owers of the dioecious plant white campion. Plant Cell 6:1775±
1787
H, Schwarz-Sommer Z +1990) De®ciens, a homeotic gene in-
volved in the control of ¯ower morphogenesis in Antirrhinum
majus: the protein shows homologyto transcription factors.
EMBO J 9:605±613
Heck GR, PerrySE, Nichols KW, Fernandez DE +1995) AGL15, a
MADS domain protein expressed in developing embryos. Plant Theissen G, Kim JT, Saedler H +1996) Classi®cation and phylogeny
Cell 7:1271±1282
of the MADS-box multigene familysuggest de®ned roles of
MADS-box gene subfamilies in morphological evolution of
eukaryotes. J Mol Evol 43:484±516
Iwahori S, Lyons JM, Smith OE +1970) Sex expression in cucumber
plants as a€ected by 2-chloroethylphosphonic acid, ethylene,
and growth regulators. Plant Physiol 46:412±415
Theissen G, Becker A, Rosa AD, Kanno A, Kim JT, Munster T,
È
Kamachi S, Sekimoto H, Kondo N, Sakai S +1997) Cloning of a Winter KU, Saedler H +2000) A short historyof MADS-box
cDNA for a 1-aminocyclopropane-1-carboxylate synthase that genes in plants. Plant Mol Biol 42:115±149
is expressed during development of female ¯owers at the apices Trebitsh T, Staub JE, O'Neill SD +1997) Identi®cation of a 1-am-
of Cucumis sativus L. Plant Cell Physiol 38:1197±1206
Kamachi S, Mizusawa H, Matsuura S, Sakai S +2000) Expression
of two 1-aminocyclopropane-1-carboxylate synthase genes, CS-
inocyclopropane-1-carboxylic acid synthase gene linked to the
Female +F) locus that enhances female sex expression in
cucumber. Plant Physiol 113:987±995
ACS1 and CS-ACS2, correlated with sex phenotypes in cu- Weigel D, Meyerowitz EM +1994) The ABCs of ¯oral homeotic
cumber plants +Cucumis sativus L.). Plant Biotechnol 17:69±74 genes. Cell 78:203±209
Kater MM, Colombo L, Franken J, Busscher M, Masiero S, YanofskyMF +1995) Floral meristems to ¯oral organs: genes
Campagne MMVL, Angenent GC +1998) Multiple AGAMOUS
homologs from cucumber and petunia di€er in their abilityto
induce reproductive organ fate. Plant Cell 10:171±182
controlling earlyevents in Arabidopsis ¯ower development.
Annu Rev Plant Physiol Plant Mol Biol 46:167±188
YanofskyMF, Ma H, Bowman JL, Drews GN, Feldmann KA,
Meyerowitz EM +1990) The protein encoded by the Arabidopsis
homeotic gene agamous resembles transcription factors. Nature
346:35±39
Liang P, Averboukh L, Pardee AB +1993) Distribution and cloning
of eukaryotic mRNAs by means of di€erential display: re®ne-
ments and optimization. Nucleic Acids Res 21:3269±3275
Ma H +1994) The unfolding drama of ¯ower development: recent Yang SF, Ho€man NE +1984) Ethylene biosynthesis and its
results from genetic and molecular analyses. Genes Dev 8:745± regulation in higher plants. Annu Rev Plant Physiol 35:155±
756 189
Ma H, YanofskyMF, Meyerowitz EM +1991) AGL1-AGL6, an Yin T, Quinn JA +1995) Tests of a mechanistic model of one hor-
Arabidopsis gene familywith similarityto ¯oral homeotic and
transcription factor genes. Genes Dev 5:484±495
mone regulating both sexes in Cucumis sativus +Cucurbitaceae).
Am J Bot 82:1537±1546