Biology Reference
In-Depth Information
interesting model in which to study a variety of developmental aspects relating to
cell specification, cell polarity, signaling, cell differentiation, double fertilization,
genomic imprinting, and apomixis [1-3].
Out of the 28,974 predicted open reading frames of Arabidopsis thaliana,
a few thousand genes are predicted to be involved in embryo sac development
[1,4]. These genes can be grouped into two major classes: genes that are necessary
during female gametogenesis and genes that impose maternal effects through the
female gametophyte, and thus play essential roles for seed development. To date,
loss-of-function mutational analyses have identified just over 100 genes in Ara-
bidopsis that belong to these two classes [5-14]. However, only a small number
of genes have been characterized in depth. Cell cycle genes (for instance, PRO-
LIFERA, APC2 [ANAPHASE PROMOTING COMPLEX 2], NOMEGA, and
RBR1 [RETINOBLASTOMA RELATED 1]), transcription factors (for instance,
MYB98 and AGL80 [AGAMOUS-LIKE-80]), and others (including CKI1
[CYTOKININ INDEPENDENT 1], GFA2 [GAMETOPHYTIC FACTOR
2], SWA1 [SLOW WALKER 1] and LPAT2 [LYSOPHOSPHATIDYL ACYL-
TRANSFERASE 2]) are essential during embryo sac development [6,15-23].
Maternal effect genes include those of the FIS (FERTILIZATION INDEPEN-
DENT SEED) class and many others that are less well characterized [9,13,24].
FIS genes are epigenetic regulators of the Polycomb group and control cell pro-
liferation during endosperm development and embryogenesis [7,10,12,25,26].
Ultimately, the molecular components of cell specification and cell differentiation
during megagametogenesis and double fertilization remain largely unknown, and
alternate strategies are required for a high-throughput identification of candidate
genes expressed during embryo sac development.
Although transcriptome profiling of Arabidopsis floral organs [27,28], whole
flowers and seed [29], and male gametophytes [30-33] have been reported in
previous studies, large-scale identification of genes expressed during female game-
tophyte development remains cumbersome because of the microscopic nature of
the embryo sac. Given the dearth of transcriptome data, we attempted to ex-
plore the Arabidopsis embryo sac transcriptome using genetic subtraction and
microarray-based comparative profiling between the wild type and a sporophytic
mutant, coatlique (coa), which lacks an embryo sac. Using such a genetic subtrac-
tion, genes whose transcripts were present in the wild type at levels higher than
in coa could be regarded as embryo sac expressed candidate genes. While our
work was in progress, Yu and coworkers [34] reported a similar genetic approach
to reveal the identity of 204 genes expressed in mature embryo sacs. However,
their analysis of the embryo sac transcriptome was not exhaustive because they
used different statistical methodology in their data analysis. Thus, we combined
their dataset with ours for statistical analyses using three statistical packages in
