Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
ABSTRACT MADS-box genes have been shown standing of ovule development and to serve as a guide
to play a major role in defining plant architecture. for further genetic and molecular analysis.
Recently, several MADS-box genes have been reported Several Arabidopsis mutants have been identified
that are highly expressed in the ovule. However, only for that affect ovule development [for a review, see Schneitz
the Petunia genes FBP7 and FBP11 has a function in et al., 1998a]. Almost all of them show defects in
defining ovule identity been shown. We have isolated a integument development, i.e., mutations in the
rice MADS-box gene named OsMADS13. Expression AINTEGUMENTA (ANT ) [Elliott et al., 1996; Klucher
analysis has shown that this gene is highly expressed in et al., 1996] and HUELLENLOS (HLL) [Schneitz et al.,
developing ovules. In order to facilitate a detailed 1998b] loci, which lead to almost complete absence of
characterization of rice ovule-expressed genes, a compre- integuments. Mutations in BELL1 (BEL1) [Modrusan
hensive morphological description of ovule development et al., 1994; Robinson-Beers et al., 1992; Schneitz et al.,
in rice has been performed. The predicted amino acid 1997] result in the arrest of inner integument develop-
sequence of OsMADS13 shows significant homology ment, whereas in INNER NO OUTER (INO) mutant,
with ZAG2, a maize MADS-box gene, which is also the development of the outer integument is blocked.
expressed mainly in the ovule. Mapping of the gene in These studies have led to the postulation of models for
the rice genome showed that it is located on chromo- the genetic control of ovule development [Angenent and
some 12, which is syntenic to two maize regions where Colombo, 1996; Baker et al., 1997; Schneitz et al., 1998a].
ZAG2 and its paralogous gene ZMM1 have been In Arabidopsis, genes controlling ovule identity have
mapped. Our results suggest that OsMADS13 is the not yet been characterized. However, in Petunia, two
ortholog of ZAG2 and ZMM1 and might play a role in genes, FLORAL BINDING PROTEIN7 (FBP7) and
rice ovule and seed development. Dev. Genet. 25:237– FBP11, have been identified that control ovule identity
244, 1999. r 1999 Wiley-Liss, Inc. [Angenent et al., 1995; Colombo et al., 1995]. Both
genes, which share 90% identity and are probably
Key words: MADS-box genes; rice; ovule develop- redundant, encode transcription factors of the MADS-
ment; synteny; OsMADS13; ZAG2; FBP11 box family. FBP7 and FBP11 are expressed in the
center of the gynoecium before ovule primordia are
visible; at later stages of development, their expression
can only be detected in the ovules. Simultaneous inhibi-
tion of these genes by cosuppression resulted in trans-
INTRODUCTION
formation of ovules into carpelloid structures [Ange-
The ovule is the most important part of the female nent et al., 1995], showing that these two genes are
reproductive organ. It is the site of megasporogenesis, necessary to determine ovule identity. Furthermore,
megagametogenesis, and fertilization and, as the precur- ectopic expression of FBP7 or FBP11 in Petunia in-
sor of the seed, it plays a major role in the reproductive
cycle of the plant. In angiosperms, ovules are part of the
gynoecium, which consists of one or more carpels. Ovule
development has been studied in a wide variety of Contract grant sponsor: EU-BIOTECH; Contract grant number: BIO4-
CT97-2217; Contract grant sponsor: NATO; Contract grant number:
flowering plants [Bouman, 1984; Reiser and Fischer, CRG.CRG 973222. This project was supported by the EU-BIOTECH
1993; van Tieghem, 1898; Warming, 1878; Willemse project number BIO4-CT97-2217.
and van Went, 1984]. However, because Arabidopsis is
considered to be the model species for studying develop- *Correspondence to: Lucia Colombo, Department of Genetics and
mental processes in dicots, most such research is cur- Microbiology, University of Milan, via Celoria 26, 20133 Milan, Italy.
rently done on this species. A detailed morphological E-mail: lucia.colombo@unimi.it
study of Arabidopsis ovule development was been done
[Schneitz et al., 1995] in order to gain a better under- Received 10 March 1999; Accepted 1 June 1999
Fig. 1. Light microscopy of the development of the rice ovule. A: degenerate (dm) and create space for the functional megaspore (fm) to
The developing ovule is oriented toward the abaxial side of the pistil. divide. ⫻20. The ovule is now about 100° from the vertical axis. F:
Ovule development is tenuinucelluste starting at approximately 70° Functional megaspore. ⫻100. G: The two-nucleate developing embryo
from the vertical axis and curving downward to about 160°. ⫻20. The sac showing vacuolisation at its chalazal end. ⫻40. H: The mature
inner (ii) and outer integuments (oi) have been initiated. B: The embryo sac. ⫻10. The antipodals are laterally positioned close to the
archespore (ar) develops from a subepidermal cell of the nucellus. It is chalazal pole. The egg apparatus and central cell occupy the micropy-
much larger than the other nucellus cells. ⫻100. C: The archespore lar end. I: The egg apparatus with polarized egg cell (ec) and one
has differentiated into the megaspore mother cell (mc). It shows a visible synergid (sy) ⫻40. The two polar nuclei of the central cell (cc)
characteristically less dense cytoplasm and polarity, with the nucleus have fused, shown here with two large nucleoli along a strand of
occupying the micropylar half of the cell. ⫻100. D: Linear tetrad of cytoplasm against the egg cell. The micropyle is formed by a narrow
megaspores. ⫻100. The chalazal megaspore (panel 4) becomes the opening in the inner integument, continuous with the intercellular
functional megaspore. E: Three megaspores at the micropylar end space between four large nucellus cells lining the egg apparatus.
MADS-BOX GENES AND RICE DEVELOPMENT 241
Fig. 5. In situ hybridization of developing rice florets. A: No 3-cm inflorescence showing intense signals in the developing ovule.
hybridization signal could be detected in very young spikelets. B: Arrow indicates ovule primordium. F: OsMADS13 transcripts are
Developing spikelet at a higher magnification, with no hybridization detected in the developing ovule of spikelet from a 5-cm inflorescence.
signal. C: Spikelets from a 1 cm inflorescence show strong OsMADS13 G: OsMADS13 transcripts are preferentially expressed in the differen-
signals in the ovule primordium. The arrow indicates ovule primor- tiating integuments of the ovule. H: With maturation of the embryo
dium. D: Closeup of spikelet from 1-cm inflorescence showing localiza- sac, weaker signals of OsMADS13 transcription are detected in the ovule.
tion of OsMADS13 transcripts in ovule primordium and in the inner sp, spikelet; le, lemma; ca, carpel primordium; p, palea; ov, ovule
tissue of the ovary wall (arrowhead) lining the ovule. E: Spikelet on a primordia; ow, ovary wall; in integument, es, embryo; sac, nu, nucellus.
megasporocyte results in linear or T-shaped tetrads as In situ hybridization analysis showed that Os-
is observed in most grasses. The functional megaspore MADS13 expression was detectable only in the later
in rice forms small vacuoles (Fig. 1F) which are typical stages of flower development, when the ovule starts to
of Polygonum type megagametogenesis [Maheshwari, develop. Its expression was restricted to developing
1950; Willemse and Van Went, 1984; Haig, 1990; Reiser ovules and the inner site of the carpel wall and per-
and Fischer, 1993]. The rice megagametophyte initially sisted until the ovule was mature. Comparison of the in
forms a vacuole at the chalazal pole during the two- situ data with the histological analyses (Fig. 1) showed
nucleate stage and later forms the large central vacuole that OsMADS13 expression was largely restricted to
at the late two-nucleate stage, which persists during the sporophytic parts of the ovule.
the four-nucleate stage. The division of the antipodals, Comparative sequence analysis (Fig. 3) shows that
which occurs after cellular partitioning in the rice OsMADS13 shares high homology with OsMADS3,
embryo sac, conforms with observations made on the which is considered the rice homologue of AGAMOUS
maize embryo sac. Vollbrecht and Hake [1995] counted from Arabidopsis [Kang et al., 1995; Yanofsky et al.,
up to almost 100 cells in maize after mitotic division of 1990]. Comparing the in situ hybridization data shows
the antipodal cells. that these genes are expressed differently. OsMADS3
Genomic mapping is a fundamental aspect in the is, like AGAMOUS in Arabidopsis, expressed both in
characterization of novel genes. In this respect, the stamens and carpel, whereas OsMADS13 is only ex-
availability of highly dense genetic maps that cover the pressed in the inner part of the carpel. The different
entire genome are a great advantage. This situation expression patterns of these two rice MADS-box genes
can be usefully exploited in rice, for which extensive suggests that they will not have a comparable function
genetic and physical mapping with various types of in flower development.
molecular markers have produced a large amount of The expression pattern of OsMADS13 is very similar
information regarding genome structure, gene position, to that observed for ZAG2 in maize, which is also
and the identification of quantitative trait loci control- expressed in developing ovules and in the inner carpel
ling agronomically important characters. The impor- faces [Schmidt et al., 1993]. For both monocot homologs,
tance of gene mapping in rice is also enhanced by the OsMADS13 and ZAG2, no function has yet been deter-
extensive colinearity found among the genomes of mined, although their patterns of expression suggest
grasses for which rice, with its compact genome, repre- that they may play important roles in ovule develop-
sents the species of reference [Moore et al., 1995, 1997]. ment. Experiments are currently under way to overex-
In our case, RFLP mapping proved very useful in press and suppress OsMADS13 in rice, which may help
defining the phylogenetic relationships among MADS us to understand the function of this gene. For two
box genes in maize and rice, for which sequence compari- genes of Petunia, FBP7 and FBP11, these experiments
son and analogy of expression patterns have suggested have already been done. Experiments in which FBP7
a common ancestor and functional similarity. and FBP11 expression was suppressed showed that
OsMADS13 shares the highest degree of homology these genes are necessary to determine ovule develop-
with the cDNA clones of ZAG2 [Schmidt et al., 1993] ment. Their overexpression resulted in ectopic ovule
and ZMM1 [Theissen et al., 1995]. ZAG2 and ZMM1 are formation on sepals and petals, indicating that they are
located in duplicated regions of the maize genome, on sufficient to induce ovule formation. FBP7 and FBP11
chromosomes 3 and 10, respectively. Mapping within are most likely orthologues of OsMADS13 because they
and across species of the grass family has produced a exhibit similar expression patterns and show high
consensus grass comparative map and has substanti- levels of homology (greater than 75%). Given these
ated the hypothesis that maize is allotetraploid [Devos observations, one might speculate that OsMADS13 and
and Gale, 1997]. According to this hypothesis, the ZAG2 will have a comparable role in rice and maize
duplicated regions comprising the two paralogous genes ovule development, respectively.
ZAG2 and ZMM1 derive from allotetraploidation. Our
mapping data clearly indicate that OsMADS13 is the
ortholog of the ZAG2/ZMM1 pair. OsMADS13 maps as
a single locus on the short arm of rice chromosome 12 ACKNOWLEDGMENTS
which, according to the consensus grass comparative We thank Dr. Andrè Van Lammeren for helpful
map is considered to be colinear with both maize comments, and Dr. Andrew MacCabe for critical read-
chromosomes 3 and 10, the linkage group locations of ing of the manuscript. We thank Dr. T. Sasaki, Dr. M.
ZAG2 and ZMM1, respectively. However, this consen- Yano, and Dr. A. Shomura of the Rice Genome Project at
sus map can be viewed only as indicative, unless direct Tsukuba, for providing assistance for the mapping
cross mapping of the genes of interest is performed. Our experiments. Furthermore, we thank Dr. Simona Masi-
cross-mapping data, in fact, confirm the syntenic rela- ero and Dr. Diana Rigola for their assistance in some of
tionships between maize chromosomes 3 and 10 and the experiments. Short working visits by Zenaida Lopez-
rice chromosome 12 and add strong evidence support- Dee and Dr. Peter Wittich to Holland and Italy, respec-
ing the hypothesis that OsMADS13 is phylogenetically tively, were supported by the collaborative NATO grant
related to ZAG2 and ZMM1. CRG.CRG 973222.
244 LOPEZ-DEE ET AL.