Funct Integr Genomics (2012) 12:229–248

DOI 10.1007/s10142-012-0274-3

ORIGINAL PAPER

Expression dynamics of metabolic and regulatory components

across stages of panicle and seed development in indica rice
Rita Sharma & Pinky Agarwal & Swatismita Ray &
Priyanka Deveshwar & Pooja Sharma &
Niharika Sharma & Aashima Nijhawan & Mukesh Jain &
Ashok Kumar Singh & Vijay Pal Singh &
Jitendra Paul Khurana & Akhilesh Kumar Tyagi &
Sanjay Kapoor

Received: 22 November 2011 / Revised: 2 March 2012 / Accepted: 6 March 2012 / Published online: 31 March 2012
# Springer-Verlag 2012

Abstract Carefully analyzed expression profiles can serve as a playing key role in reproductive development were elucidated
valuable reference for deciphering gene functions. We exploited after overlaying the expression data on the public databases and
the potential of whole genome microarrays to measure the manually curated list of transcription factors, respectively.
spatial and temporal expression profiles of rice genes in 19 During floral meristem differentiation (P1) and male meiosis
stages of vegetative and reproductive development. We could (P3), the genes involved in jasmonic acid and phenylpropanoid
verify expression of 22,980 genes in at least one of the tissues. biosynthesis were significantly upregulated. P6 stage of panicle,
Differential expression analysis with respect to five vegetative containing mature gametophytes, exhibited enrichment of tran-
tissues and preceding stages of development revealed reproduc- scripts involved in homogalacturonon degradation. Genes reg-
tive stage-preferential/-specific genes. By using subtractive log- ulating auxin biosynthesis were induced during early seed
ic, we identified 354 and 456 genes expressing specifically development. We validated the stage-specificity of regulatory
during panicle and seed development, respectively. The meta- regions of three panicle-specific genes, OsAGO3, OsSub42, and
bolic/hormonal pathways and transcription factor families RTS, and an early seed-specific gene, XYH, in transgenic rice.

The authors Rita Sharma and Pinky Agarwal contributed equally to this
work.
Electronic supplementary material The online version of this article
(doi:10.1007/s10142-012-0274-3) contains supplementary material,
which is available to authorized users.
R. Sharma : P. Agarwal : S. Ray : P. Deveshwar : P. Sharma : Present Address:
N. Sharma : A. Nijhawan : M. Jain : J. P. Khurana : A. K. Tyagi : P. Agarwal : M. Jain : A. K. Tyagi
S. Kapoor (*) National Institute for Plant Genome Research,
Interdisciplinary Centre for Plant Genomics and Department of Aruna Asaf Ali Marg,
Plant Molecular Biology, University of Delhi South Campus, New Delhi 110067, India
Benito Juarez Road,
New Delhi 110021, India Present Address:
e-mail: [email protected] S. Ray
Biotechnology and Bioresources Management Division, Tata
A. K. Singh : V. P. Singh Energy Research Institute,
Division of Genetics, Indian Agriculture Research Institute, Lodhi Road,
New Delhi 110012, India New Delhi 110003, India

Present Address: Present Address:

R. Sharma N. Sharma
Department of Plant Pathology, Plant Molecular Biology and Biotechnology Group, Melbourne
University of California, School of Land and Environment, University of Melbourne,
Davis, CA 95616, USA Parkville 3010 Victoria, Australia
230 Funct Integr Genomics (2012) 12:229–248

The data generated here provides a snapshot of the underlying (YL), mature leaf (ML), and shoot apical meristem (SAM)
complexity of the gene networks regulating rice reproductive revealed reproductive stage preferential/specific genes.
development. Enrichment of transcription factor coding genes and hormonal
pathways in vegetative, panicle, and seed stages implied their
Keywords Development . Expression . Meta-analysis . Metabolic significance during plant growth and development. The data
pathways . Panicle . Promoter . Seed . Transcription factors have previously been extensively validated by qPCR analysis
of candidates from various gene families encoding transcrip-
tion factors (Agarwal et al. 2007; Arora et al. 2007; Nijhawan
Introduction et al. 2008; Jain et al. 2008; Sharma et al. 2010), signal
transduction components (Jain et al. 2006; Jain and Khurana
Reproductive development is a dynamic process involving 2009; Singh et al. 2010; Jain et al. 2007), RNA interference
complex interplay of various regulatory networks. Spatial machinery (Sharma et al. 2010; Sharma et al. 2009), and stress
and temporal transcriptome profiles represent snapshot of gene responsive factors (Ray et al. 2011). Due to prior submission
activity and thus have been extensively used for deciphering in Gene Expression Omnibus database, various researchers
the role of individual genes/pathways or regulatory networks (Cao et al. 2008; Howell et al. 2009; Ma and Zhao 2010; Jiang
and plausible interactions among them (Adams 2008). In fact, et al. 2009; Li et al. 2009) have also used this dataset for
in the past decade, multitude of microarray-based studies have analyzing expression profiles of rice genes depicting the con-
been performed towards elucidating reproductive organ devel- fidence of rice community in the quality of the data. We have
opment in Arabidopsis (Alves-Ferreira et al. 2007; Becerra et also earlier used part of the data generated here to identify
al. 2006; Fait et al. 2006; Hennig et al. 2004; Wellmer et al. anther-specific transcripts by comparing the expression pro-
2006; Wellmer et al. 2004; Zhang et al. 2005; Wilson et al. files of rice genes during anther development with those of
2005b; Day et al. 2008), rice (Endo et al. 2004; Furutani et al. vegetative and seed stages analyzed here (Deveshwar et al.
2006; Hobo et al. 2008; Kondou et al. 2006; Lan et al. 2004; 2011). However, the goal of this study is to (1) identify key
Hirano et al. 2008; Suwabe et al. 2008; Jiao et al. 2009; Wang genes/pathways regulating various stages of panicle and seed
et al. 2010; Fujita et al. 2010; Li et al. 2007a, b; Wang et al. development, (2) provide an insight into magnitude and re-
2005; Deveshwar et al. 2011), maize (Grimanelli et al. 2005; dundancy in rice transcriptome during different stages of
Liu et al. 2008; Lee et al. 2002), wheat (Wilson et al. 2005a), vegetative and reproductive development, and (3) provide an
and other non-model plant systems (Hansen et al. 2009; in planta validation of stage specificity of selected genes using
Laitinen et al. 2005; Endo et al. 2002; Tebbji et al. 2010). promoter-reporter analysis. Expression profiles of four candi-
One of the limitations of these studies is that most of these date genes, exhibiting varied expression patterns, have been
scored the number of probe sets rather than unique transcripts verified by promoter-GUS analysis in transgenic rice.
as an estimate of gene expression. Moreover, these studies
mainly focused on analyzing spatial expression profiles in
various cell/tissue types or restricted time points during devel- Materials and methods
opment. Therefore, it is very difficult to make cross compar-
isons due to collection of tissues at different stages of Collection of plant material and categorization of panicle
development and lack of an internationally accepted staging and seed stages
system in rice similar to that of Arabidopsis (Smyth et al.
1990). Here, we report the time-course analysis of expression The vegetative tissues and rice panicles spanning all the stages
dynamics in indica rice encompassing the complete series of of panicle and seed development were collected from field-
reproductive development, from panicle initiation to seed mat- grown Oryza sativa indica var. IR64 plants at IARI (Indian
uration. All the analysis was performed on the list of probe IDs Agricultural Research Institute, New Delhi, India). For evalu-
representing non-TE-related unique transcripts. We believe ating the precise stages of development, we performed the
that our experimental design provides more realistic and com- histochemical analysis with the anthers collected at different
plete view of transient as well as long-term developmental stages of development and correlated it with the length of
response, which is not attainable by organ/cell-type specific panicles (for details, see Supplementary Figure S1). Based on
or single time point/stage-based studies. the stages of anther development and information available in
We categorized panicle and seed development into nine literature (Ikeda et al. 2004; Itoh et al. 2005; Raghavan 1988),
(P1–P6 and P1-I–P1-III) and five (S1–S5) categories, respec- panicles were divided into six groups (P1–P6; Supplementary
tively, and used Affymetrix arrays to generate spatial and Table S1, Figure S1 and S2). To document morphological
temporal expression profiles during rice reproductive organ details, panicles collected from all six groups were photo-
development. Itemized comparisons with five vegetative tis- graphed using a digital camera (Canon PowerShot S1 IS,
sues including 7-day-old seedlings (SD), roots (R), Y leaf Singapore). The P1 stage was further categorized into three
Funct Integr Genomics (2012) 12:229–248 231

sub-groups (P1-I–P1-III) and photographed under a dissecting filter was not applied to S1 stage of seed development due to
stereo-zoom microscope (MZ12.5 with DFC320 camera; high similarity in transcript pool of P6 and S1 stages. GO
Leica Gmbh, Germany). The seed development was catego- annotations for selected datasets were downloaded from the
rized into five groups (S1–S5) based on the days after polli- rice genome annotation project (RGAP) database (http://rice.
nation (Supplementary Table S1, Figure S2). To document plantbiology.msu.edu/). The metabolic pathways induced/sup-
seed stages, spikelets at different stages of seed development pressed during reproductive development were analyzed using
were dehusked and dissected to extract embryos. The mor- data available in RiceCyc database of Gramene (http://path
phological details were observed and photographed using the way.gramene.org/expression.html; Jaiswal et al. 2006). We
dissecting stereo-zoom microscope. Three biological repli- analyzed meta-profiles of short-listed datasets using Rice
cates of five vegetative tissues, including ML, YL, R, SD, Oligonucleotide Array Database (ROAD; http://www.rice
and SAM, were also collected from field-grown plants. For array.org/expression/meta_analysis.shtml).
uniformity and simplicity, all the categories of panicle and A comprehensive list of all the transcription factor family
seed development and vegetative tissues are referred as genes was generated using HMM analysis (Madera and
stages of development in the manuscript. The details Gough 2002) as well as keyword search as described previ-
of stages used are provided in Supplementary Table S1. ously (Arora et al. 2007). The differential expression pro-
files were analyzed as described above. Hypergeometric
Microarray experiments and data analysis distribution analysis was also performed to identify tran-
scription factor families enriched in panicle/seed-specific
Affymetrix GeneChip® Rice Genome arrays containing 51,279 data sets. Cluster analysis on rows was performed for se-
rice transcripts (approximately 48,564 japonica and 1,260 ind- lected datasets, using Euclidean distance metric and Ward’s
ica transcripts) were used to study the global changes in gene Linkage rule of Hierarchical clustering.
expression during rice reproductive development. RNA was
isolated as described (Arora et al. 2007; Nijhawan et al. qPCR analysis
2008). cDNA synthesis, cRNA synthesis, labeling, and hybrid-
izations followed by scanning were carried out as per manu- The expression profiles of four selected genes were validat-
facturer’s instructions and described (Affymetrix, Santa Clara, ed using qPCR analysis as described earlier (Jain et al. 2006;
CA; (Arora et al. 2007). In total, 57 .CEL files representing Arora et al. 2007). Three biological and three technical
three biological replicates each of ML, YL, R, SD, nine panicle replicates were performed for each stage and standard error
stages (P1-I−P1-III, P1−P6), and five seed stages (S1−S5) were was calculated between them. ACTIN was used as an en-
imported in ArrayAssist 5.0.0 microarray data analysis soft- dogenous control. The data were normalized to match the
ware (Stratagene) followed by quantile normalization by using profiles with that of microarray data.
GCRMA algorithm and log2 transformated (Wu et al. 2003).
The correlation coefficient between biological replicates was Promoter-reporter constructs
analyzed. The data were separately normalized using MAS 5.0
algorithm to generate Present/Absent calls. The unique number About 1.5−2 kb putative promoter regions of selected genes
of transcripts represented on the GeneChip® was identified as (−1,560 to +16 of RTS, −1,742 to +20 of OsAGO3, −1,518 to
described previously (Deveshwar et al. 2011) and filtered data- +10 of OsSub42, −1,266 to +71 of XYH) including partial or
set was used for downstream analysis. complete 5′ UTR were PCR amplified. The list of primers is
Principal component analysis, using 19 principle compo- given in Supplementary Table S2. Amplified DNA fragments
nents, was carried out with mean centering and scaling of all the were cloned in pENTRTM/D-TOPO vector (Invitrogen Inc.,
variables to unit variance and presented as three-dimensional USA) and validated by sequencing. LR reaction was per-
view using eigenvalues, E1–E2–E3. Differential expression formed using Gateway® LR ClonaseTM II enzyme mix
analysis was carried out with respect to all four vegetative (Invitrogen, USA) as per manufacturer’s instructions to trans-
stages (SAM, YL, ML, SD), taken separately, as well as with fer the promoter DNA from pENTRTM/D-TOPO vector to
respect to preceding stage by applying an FDR correction based pMDC164, expression vector carrying the gene encoding
on Benjamini and Hochberg method and p value cut off of GUS as reporter (Arabidopsis Biological Resource Center;
≤0.05 (Benjamini and Hochberg 1995). For early panicle stages Curtis and Grossniklaus 2003).
(P1-I, P1-II, and P1-III), SAM was taken as reference and the p
values were given as uncorrected at a cut off of ≤0.005. To Rice transformations and GUS histochemical assay
identify stage-specific genes, log2 expression value cut off of
≥5.64 (normalized average expression value (NAEV) ≥50) in Final vectors were mobilized into Agrobacterium tumefaciens
the stage of interest and ≤3.90 (NAEV ≤15) in rest of the stages strain AGL1 by electroporation. The PB1 variety of indica rice
was employed. However, to identify panicle-specific genes, the was transformed using the protocols described previously
232 Funct Integr Genomics (2012) 12:229–248

(Mohanty et al. 1999; Toki et al. 2006). The list of primers used which was used for all the subsequent analyses (for details,
to confirm the presence of transgene is given in Supplementary see Supplementary Figure S3).
Table S2. GUS histochemical assay was performed to check the Genome-wide expression profiles of all 14 stages of repro-
activity of GUS (Jefferson et al. 1987). Different tissues of the ductive development, along with five stages of vegetative
transgenic as well as wild-type plants were incubated in GUS development including ML, YL, SAM, R, and SD, were
histochemical buffer containing 10 % methanol at 37°C for 16 analyzed using GeneChip® Rice Genome Arrays
−20 h followed by incubation in acetone/ethanol (1:3). The (Affymetrix). A correlation of >0.96 was obtained among
observations were recorded by using DFC320 mounted on three biological replicates of all the stages analyzed. A prin-
MZ12.5 Stereo microscope (Leica Gmbh, Germany). ciple component analysis performed on the dataset clearly
demonstrated, on one hand, the expanse and distinctness of
Genomic localization of differentially expressed genes vegetative transcriptomes and, on the other, underlined the
molecular continuum as well as progression of panicle and
To localize the genes on respective chromosomes, only those seed development paths emanating from SAM to P1-I and P6-
with NAEVof ≥50 (log2 expression value ≥5.64) in at least one S1 transitions, respectively (Fig. 1a). The distinct positioning
of the stages were analyzed. The genes exhibiting ≥2 folds of both reproductive and vegetative tissues/time points also
differential expression in any of the 19 stages were extracted. provided validation of our staging system.
In-house generated programs (written in “C” and Perl) were To comprehend the magnitude of rice transcriptome and
used to identify Gene Clusters with Similar Expression contribution of individual tissues, the number of expressed
(GCSEs) profiles. All the genes were sorted based on their genes in each stage/tissue was determined. The analysis ver-
chromosomal positions as given in RGAP database of MSU ified expression of 22,980 unique genes in at least one of the
(http://rice.plantbiology.msu.edu/). Based upon the expression stages/tissues analyzed which is comparable to the number of
profiles obtained after differential expression analysis, all the genes detected using cDNA arrays (Jiao et al. 2009). The
genes were assigned the stage preferential/specific profiles and maximum number of genes (20,421) was called “present” in
then a sequential scan was performed by comparing the ex- panicles implying the need for diverse transcript population
pression profile of each gene with the next gene in order. If during development of specialized floral organs (Fig. 1b).
expression profiles of two or more contiguous genes match up, Among vegetative tissues, seedlings and roots exhibited
the region was scored as a GCSE. To display the GCSEs on rice higher number of transcripts in comparison to leaf tissues,
chromosomes, Differential Gene Locus Mapping (DIGMAP) probably because of diverse and higher metabolic activity.
version 2 (http://geneexplorer.mc.vanderbilt.edu/DIGMAP/; Yi A comparison of expressed genes revealed that 9,021 genes
et al. 2005) was used. The per cent identity between the genes expressed in all the stages analyzed and thus might be involved
falling in the same GCSE was calculated using MegAlign in basal metabolic pathways. About 35-45 % genes from each
software 4.03 (DNASTAR Inc.). Genes sharing ≥70 % identity stage exhibited overlapping expression in multiple develop-
at protein level were considered to be tandemly duplicated. To mental windows, whereas, only 0.5 % to 3.2 % genes were
check if the genes comprised in a GCSE were involved in the specific to each developmental stage (Fig. 1b). Among repro-
same or different pathways, GOSlim assignments and putative ductive stages, P6 stage of panicle, that harbors mature pollen,
functions for all of the genes were obtained from RGAP had maximum number of specific transcripts as many of them
database (http://rice.plantbiology.msu.edu/index.shtml) and could be stored for pollen germination and tube growth
RiceCyc database of Gramene (http://pathway.gramene.org/ (Becker and Feijo 2007). Roots exhibited maximum percent-
expression.html; Jaiswal et al. 2006). age of specific genes (2.3 %, Fig. 1b) among vegetative tissues
suggesting that root being an underground part might have
developed some unique biochemical pathways (Zhang et al.
Results 2005). More than 650 genes are expressed in both panicle and
seed stages, whereas, 321 transcripts were shared between
Magnitude of rice transcriptome SAM and panicles. Panicles were found to share the maximum
number of transcriptional entities with roots (146) compared to
Since the number of probesets (57,381) on GeneChip® Rice other vegetative stages analyzed in this study.
Genome Array do not correspond to the number of annotat-
ed genes, we carried out an extensive curation exercise to Differential expression analysis during rice reproductive
filter probesets that (1) do not map to any annotated tran- development
scription unit, or (2) represent internal controls or TE-related
genes, or (3) are not the 3′ most probeset (in case of multiple Since no single method can identify all the relevant genes
probesets per transcription unit). This resulted in a dataset of with potential involvement in rice reproductive develop-
37,927 probe sets representing unique genes on the chip, ment, two strategies were adopted to identify differentially
Funct Integr Genomics (2012) 12:229–248 233

Fig. 1 a Principle component analysis of various developmental stacked bar graph. The color legend is given on the topmost left of the
stages. The graph shows all the data points projected in the three- graph. c Differential expression analysis during stages of panicle and
dimensional space formed by three coordinates after rotation. Each seed development with respect to vegetative stages. The number of
data point represents an independent tissue with green color represent- genes showing up- and downregulation in each panicle stage with
ing leaf stages (ML and YL); gray represents root (R) and seedlings respect to each of the four vegetative stages is plotted. Pink columns
(SD); purple represents SAM and early panicle stages (P1-I to P1-III); represent the commonly up- and downregulated genes with respect to
red represents panicle (P1 to P6) and blue represents seed (S1 to S5) all four vegetative stages; whereas, genes specifically upregulated in
stages. The eigenvalues E1–E2–E3 were used for plotting the data. The each stage are represented by yellow columns. The dotted line repre-
closely related developmental stages are encircled. b Overview of gene sents the pattern exhibited by differentially expressed genes with
expression during different stages of development. The number of respect to all four vegetative stages during temporal stages of repro-
genes called “Present” in each stage as well as the number of genes ductive development. SD seedling, R root, ML mature leaf, YL Y leaf,
specific to one, two, or more than two stages has been shown as the SAM shoot apical meristem, P panicle, S seed stages
234 Funct Integr Genomics (2012) 12:229–248

expressed genes during panicle and seed development. In organs progressed, the number of uniquely induced transcripts
the first approach, the transcript levels of all the genes were also increased.
compared with the vegetative stages individually as well as Similarly, differential expression analysis of early panicle
collectively. In the second approach, the transcriptome of stages (P1-I to P1-III) with respect to shoot apical meristem
each stage was compared with that of preceding stage to (SAM) identified 425, 710, and 2,032 genes exhibiting ≥2
identify genes whose expression might have triggered tran- fold change in P1-I, P1-II, and P1-III stages, respectively. Of
siently in response to specific developmental cues. these, 202, 385, and 902 genes were upregulated in P1-I, P1-
II, and P1-III, respectively. In total, 49 genes were induced in
Reproductive vs. vegetative development all three stages. Among the differentially upregulated genes, a
large proportion (106, 159, and 691 genes, in P1-I, P1-II, and
The datasets were analyzed to identify up/downregulated P1-III stage, respectively) was specific to individual stages
genes in each stage with respect to all four vegetative tissues, suggesting the rapid development-dependent molecular
collectively. The resulting dataset would, therefore, include switching in these stages.
genes whose transcript levels change as a function of repro-
ductive development. Based on differential expression analysis Analysis of cascadial expression during reproductive
with respect to vegetative stages, early panicle stages (P1 and development
P2) showed higher number of differentially expressed genes
with respect to R and SD; whereas, later panicle stages (P5 and As each stage analyzed in this study represents a character-
P6) had larger number of differentially expressed genes with istic phase in flower and seed development, one would
respect to leaf stages (Fig. 1c). Conversely, the maximum expect significant upregulation of specific transcriptional
number of genes involved in seed development was differen- units with respect to preceding stage, many of which will
tially regulated with respect to young root; whereas, the seed be important for regulating the precise developmental events
tissues had the least number of differentially expressed genes and may have induced for a very short period of time. To
when compared with mature leaf. Among the seed stages, the identify these components, we performed differential ex-
number of downregulated genes with respect to vegetative pression analysis by comparing each stage with its preced-
stages gradually increased during development suggesting ing stage of development. For P1 stage, SAM was taken as
general suppression of transcriptional activity commensurate reference. This analysis revealed 2,009, 81, 905, 41, 1,854,
with the onset of the dormant phase in seeds. and 3,590 genes getting more that 2 folds upregulated (p
Among panicle stages, the maximum number of differ- value≤0.05) in P1−P6 stages, respectively (Fig. 2a). Among
entially expressed genes was found in P2 stage (that corre- seed stages, as the development progressed, the number of
sponds to initiation of male meiosis) with 1,101 and 1,053 upregulated genes declined. In total, 1,973, 1,064, 289, 137,
genes up- and downregulated, respectively, suggesting and 178 genes were upregulated in S1, S2, S3, S4, and S5
higher order genic activity at this particular stage; whereas, stages, respectively (Fig. 2a). To shortlist the most signifi-
the minimum differential transcript accumulation was ob- cant genes, the ones with NAEV of ≥50 (or log2 expression
served at the post-meiotic P4 stage (Fig. 1c). The S1 stage, value ≥5.64) in any of the vegetative stages were filtered out
corresponding to 0–2 days post-fertilization development, from respective upregulated sets resulting in a set of 65, 2,
represents the least differentially expressed gene pool, 66, 3, 79, and 415 genes uniquely upregulated in P1−P6
which increased in S2 and S3 stages involving enlargement stages, respectively and 112, 204, 33, 19, and 11 genes in S1
of organs. Using subtractive logic, the upregulated genes in −S5 seed stages, respectively (Fig. 2a). The expression
each stage were compared with those in every other stage to
identify genes that were exclusively activated in a particular
stage of reproductive development. This analysis revealed Fig. 2 a Differential expression analysis during reproductive develop-
mental stages taking preceding stage of development as reference. Black
„
13, 32, 67, 87, 81, and 248 genes to be specifically upregu-
bars represent total number of genes upregulated by ≥2 folds at p value
lated in P1–P6 stages of panicle development, and 303, 241, ≤0.05 with respect to the preceding stage. Gray bars represent the
213, 72, and 295 genes uniquely induced in S1–S5 stages of upregulated genes with NAEV ≤15 in vegetative stages (Supplementary
seed development, respectively (Fig. 1c, yellow bars). Table S3). b Major pathways induced with respect to preceding stage
during rice reproductive development. The genes exhibiting ≥2 folds
Notable of these were anther-specific proline rich protein
upregulation in reproductive stages with respect to preceding stage and
coding genes uniquely induced in P3 and P4 stages of involved in jasmonic acid, phenylpropanoid, and IAA biosynthesis have
panicle development. Transcription factors were particularly been presented with their expression profiles plotted in the form of a heat
enriched in upregulated datasets during early seed stages; map. Color bar at the base represents log expression values, with green
2
color representing low-level expression, black representing medium-
whereas, those involved in carbohydrate metabolism and
level, and red signifying high-level expression. Developmental stages
ubiquitin-mediated proteolysis were exclusively upregulated used for expression profiling are given on the top of each column (for
in later seed stages. As the development of reproductive details, see Supplementary Table S1)
Funct Integr Genomics (2012) 12:229–248 235
236 Funct Integr Genomics (2012) 12:229–248

profiles of these genes have been compiled as a heat map sunflower exhibited P6 stage-specific expression (Baltz et
(Supplementary Figure S4). al. 1992). Many of panicle stage-specific genes seem to be
Integration of upregulated dataset with the publicly avail- involved in flavonoid and lipid biosynthesis.
able RiceCyc database (http://www.gramene.org/pathway/ During seed development, the most prominent domain,
ricecyc.html) revealed metabolic and hormonal pathways S2–S5, comprised of 93 genes with high representation of
involved in reproductive development (Fig. 2b). The genes genes coding for seed storage proteins (proline and glute-
involved in jasmonic acid biosynthesis were upregulated lins), seed allergens and those involved in starch biosynthe-
during floral organ differentiation (P1), meiotic stage of sis and ubiquitin-mediated proteolysis. In total, 19, 46, 28,
anther development (P3), and pollen development (P6) and 7, and 55 genes were found to be specific to S1, S2, S3, S4,
those implicated in phenylpropanoid biosynthesis exhibited and S5 stages, respectively. The expression profiles of pan-
significant upregulation during floral organ differentiation icle and seed-specific genes have been presented in the form
(P1), male meiosis (P3), and microspore development (P5). of hierarchical cluster maps in Supplementary Figure S6.
The analysis also showed that five genes involved in Indole- We compared the dataset of panicle/seed-specific genes
3-Acetic Acid (IAA) biosynthesis were also upregulated in with two available expression data sets generated in differ-
S1 stage (Fig. 2b). The genes involved in salvage pathway ent rice subspecies using different microarray platforms
of purines were upregulated during late panicle (P5 and P6) (Sato et al. 2011; Wang et al. 2010). A higher number of
and early seed (S1) stages. panicle and seed-specific genes were common between our
The ROAD (http://www.ricearray.org) provides gene ex- dataset (IR64) and the data generated by Wang et al. (2010)
pression data across 1,867 publicly available rice microarray for other indica varieties (Zhenshan 97 and Minghui 63) in
hybridizations assembled together to perform meta-analysis. comparison to those identified in japonica (Nipponbare)
Meta-profiles of all three datasets shown in Fig. 2b, ana- rice by Sato et al. (2011; Supplementary Figure S7). Since
lyzed using ROAD database, conformed with their signifi- the data generated in earlier studies was limited to fewer
cant induction during stages of reproductive development spatial stages of panicle and seed development, many of the
(Supplementary Figure S5). genes detected specifically in our dataset could be specific
to the stages not analyzed previously.
Identification of reproduction-specific genes and their On collating the information from Gramene database, a
affiliation to metabolic pathways significant number of genes involved in biosynthesis of
gibberellins, strictosidine, and fatty acids and their elonga-
The dataset of differentially expressed genes was further tion were found to exhibit panicle-specific expression
filtered to identify genes that expressed specifically in one (Fig. 3). However, transcripts of those involved in starch
or more stages of panicle and seed development. Three degradation were specially detected in seed stages (Fig. 3).
hundred and fifty four genes exhibited panicle-specific ex- The genes involved in brassinosteroid biosynthesis were
pression (expression value ≥50 in panicle stages and ≤15 in upregulated in P5 stage, where these may be stored in starch
other stages) with 3 % of them coding for transcription granules and later released during pollination and fertiliza-
factors. Whereas, 456 genes were expressed only in seed tion (Clouse and Sasse 1998). Various genes involved in
developmental stages (expression value ≥50 in seed stages gibberellin biosynthesis were upregulated during male mei-
and ≤15 in other stages) with 9 % of them encoding tran- osis (P3), heading stage (P6) as well as early stages of seed
scription factors. Interestingly, 40 % genes in both panicle development (S1 and S2). In addition, few genes involved in
and seed-specific datasets have not even been annotated yet, ent-kaurene biosynthesis, which is a common gibberellin
probably because of narrow windows of their expression. precursor, were also upregulated at S1 stage of seed devel-
The expression of about 48 % of the panicle-specific genes opment suggestive of their involvement of GA during early
(171) initiates at P1, P2, P3, and P4 stages, which in most seed development (Davidson et al. 2003). Analysis of meta-
cases is detected till P5 or P6 stage. However, 183 genes profiles of the genes highlighted in Fig. 3 supports their
expressed only in the P6 stage harboring mature male and specific expression in limited stages of development
female gametophytes. Since male gametophytic cells out- (Supplementary Figure S8).
number the female gametophytic cells in every floret, the
bulk of these genes represent mature pollen-specific tran- Expression dynamics of transcription factors
scripts. In fact, most of these genes code for pollen aller- during reproductive development
gens, transporters, and components of cytoskeleton and
those involved in cell wall metabolism (see Supplementary The number of transcription factor (TF) coding genes in rice
Table S3 for details). These categories of genes might play is estimated to be 2,527 categorized into 65 genes families
role in pollen–pistil interactions and pollen tube germina- (Riano-Pachon et al. 2007). However, 216 of these genes are
tion. Putative homolog of pollen-specific gene, SF3, of orphans as their role in transcriptional regulation is
Funct Integr Genomics (2012) 12:229–248 237

Fig. 3 Schematic representation of major pathways involving panicle/ expression values, where green color represents low-level expression,
seed-specific genes. Nine major pathways represented by panicle/seed- black shows medium-level expression, and red signifies high-level ex-
specific genes are highlighted. The expression profiles of the genes are pression. Developmental stages used for expression profiling are given on
shown in the form of heat maps. Color bar at the base represents log2 the top of each column (for details, see Supplementary Table S1)
238 Funct Integr Genomics (2012) 12:229–248

ambiguous. Here, we reanalyzed the rice genome using expression profiles of the transcription factor coding genes
name search and HMM analysis and identified a total of expressing in a developmental stage specific manner is
2,314 transcription factor genes belonging to 68 gene fam- presented in Supplementary Figure S10.
ilies. Except for a few gene families, additional members
were identified in each gene family (Supplementary Table Gene Clusters with Similar Expression profiles
S5). Of these, 2,100 TF genes are represented on rice
GeneChip® array. Differential expression analysis of these Previous studies have shown correlation between expression
genes with respect to vegetative stages (ML, YL, R, and SD) profiles and physical location of genes in eukaryotic
revealed 204 and 246 genes exhibiting ≥2 folds upregulation genomes including Drosophila, nematode, mouse, and
(p value≤0.05) in at least one of the panicle and seed stages, humans (Michalak 2008). Such a phenomenon has also been
respectively. Only 80 genes were upregulated in both panicle identified in rice and Arabidopsis (Ren et al. 2005; Ren et al.
and seed stages, whereas, the rest were exclusive to either 2007; Zhan et al. 2006). To get a genome-wide perspective
panicle or seed development. Conversely, 55 genes were of the influence of physical proximity of genes on their
downregulated in both panicle and seed stages. Differential expression profiles, we selected 18,180 differentially
expression analysis revealed 18 TF families including SBP, expressed genes (≥2 folds with NAEV ≥50) and analyzed
GRF, and DOF exhibited higher number of upregulated genes their physical location on rice chromosomes as described in
in panicles. Further, hypergeometric distribution analysis the “Materials and methods” section. This analysis revealed
(Supplementary Table S5) revealed that 14 families, namely 1,278 GCSE profiles comprising of 2,792 genes
BBR_GAGA, BTB/POZ, bZIP, GRF, HORMA, MADS, (Supplementary Tables S6 and S7). In total, 25 varied ex-
NAM, PWWP, SBP, TRIHELIX, WRKY, YABBY, C2H2, pression patterns were observed (Supplementary Table S8),
and DOF, were particularly enriched during panicle develop- of which 8 major patterns comprising of more than 80 % of
ment. Out of these, bZIP, MADS, and C2H2 families had more the data are shown in Fig. 5. The maximum number of
than 10 upregulated members. Twenty-seven families includ- clusters was found to exhibit SAM + panicle-preferential
ing MYB, NAM, HSF, MADS, POZ, and bZIP had relatively expression (209) and vegetative stages-preferential expres-
higher number of genes upregulated in seed stages (Fig. 4). sion (200). One hundred and eighty-five clusters exhibited
Nine TF families, namely, bHLH, BTB/POZ, bZIP, HSF, panicle-preferential expression; whereas, 147 GCSEs were
MADS, SRS, TRIHELIX, YABBY, and C2H2, were signifi- found to be seed-preferential. Most of the GCSEs (1,099)
cantly enriched, with five of them having ten or more upregu- comprised of two genes. The largest GCSE included 12
lated members. Of these, we have earlier shown the detailed genes on chromosome 10 (10_43) exhibiting root-
expression patterns of bZIP, MADS and C2H2 families, preferential expression followed by a cluster of 7 genes on
wherein certain members are panicle/seed-specific (Agarwal chromosome 1 (Supplementary Table S6). We did not ob-
et al. 2007; Arora et al. 2007; Nijhawan et al. 2008). Similar serve any chromosomal bias in the distribution of GCSEs.
analysis of downregulated genes revealed 14 gene families To check if the genes comprising a GCSE were involved in
including WRKY, C2H2, and AP2 with higher number of similar function, information related to their affiliation to
members downregulated in panicles; whereas, 27 gene fami- various biochemical pathways was retrieved from Gramene
lies including bHLH, TUBBY and C3H had higher number of database (http://pathway.gramene.org/expression.html).
downregulated genes in seeds (Fig. 4). An enrichment analy- Apparently, of the 35 clusters (74 genes), for which pathway
sis of downregulated members lays emphasis on genes whose data was available, genes in 18 clusters belonged to same
expression is probably not an essential feature for the devel- biochemical pathways. In ten of these GCSEs, however,
opment process (Supplementary Table S5). some of the genes were found to have >70 % identity at
In total, 52 TF genes including members of bHLH, bZIP, amino acid level, which could suggest their origin from
B3, C2H2, HOMEOBOX, MADS, MYB, NAM, SBP, and tandem duplication events. The annotated genes in other
WRKY gene families had vegetative stages-specific expres- eight GCSEs were involved in different pathways and
sion with fifteen of them exhibiting SAM-specific and ten exhibited <16 % identity in their protein sequences. In nine
genes exhibiting root-specific expression in contrast to only GCSEs, all the members of a cluster had not yet been
one gene specific to each leaf and seedlings. Panicle stages annotated, but, based on the available information; they
were found to have specific expression of only eleven TF seem to be involved in similar pathways.
genes having representation of two genes each from bHLH
and DOF gene families. Forty-seven TF genes exhibited Molecular characterization of candidate promoters
seed-specific expression with a majority of them being
specific to the S2 stage (2–5 DAP), suggesting that S2 might Based on their expression profiles, we selected four genes
represent an important developmental hot spot that should with stage-specific/preferential expression in panicle and
be investigated in detail. A hierarchical cluster map showing seed stages namely, OsAGO3 (LOC_Os04g52550),
Funct Integr Genomics (2012) 12:229–248 239

Fig. 4 Graphical representation of a up- and b downregulated tran- downregulated genes, respectively, in panicle (red) and seed (green)
scription factor family genes in panicle and seed stages. The total stages. The axis on the left represents the number of differentially
number of genes of each transcription factor family represented on expressed genes (shown in bar chart); whereas, broken axis with
the GeneChip® is shown in the form of area graph (gray). The graph a different data ranges on the right represents total number of genes
depicts the number of upregulated and graph b depicts the number of represented on the GeneChip® (shown by the area graph)
240 Funct Integr Genomics (2012) 12:229–248

Fig. 5 Chromosomal localization of co-expressed genes in rice. chromosomes. Each profile is shown by a different color. The color
Microarray-based expression profiles of co-expressed genes at 19 legend has been given at the base. The chromosome numbers are
stages of vegetative and reproductive development were extracted indicated on the left and Locus ID of first and last gene comprising a
and plotted in sequential order based on their location on rice cluster on each chromosome has been given

OsSub42 (LOC_Os04g45960), RTS (LOC_Os01g70440),

and XYH (AK108917) exhibiting P3–P4, P4–P5, P5–P6,
Fig. 6 Expression profiles of candidate genes and respective promoters.
The panel on the left presents the expression profiles exhibited by candi-„
and S1 specific expression, respectively, for in planta vali-
dation of promoter activities. Analysis of meta-profiles of date genes. Black bars represent the values obtained using microarray and
these genes in Affymetrix data during developmental stages white bars represent qPCR data. The qPCR values have been normalized
showed consistency in expression profiles of these genes to those obtained from the microarray dataset. The error bars represent
the standard error between three biological replicates of each stage. The
with the publicly available data (Supplementary Figure gene names are given in the top left corner of each graph. Right panel
S9a). The respective expression profiles were validated us- represents the expression of GUS reporter derived by promoters of
ing qPCR and promoter-reporter constructs were prepared candidate genes in different stages of development in rice. Stages ana-
using GUS as reporter (Fig. 6). OsAGO3 is an argonaute lyzed are as below: a a mature leaf, b root, c T.S. stem, d seed, e, f floret at
P3, g floret at P4, h floret at P5, i, j floret at P6, k anther of floret at P3,
family gene containing PIWI domain (Sharma et al. 2010) l anther of floret at P4, m anther of floret at P6, n mature gynoecium. Scale
with high transcript accumulation in P3 and P4 stages. The bar (a–j) 1 mm, (k–n) 0.1 mm. b a mature leaf, b root, c stem, d mature
promoter activity was observed specifically in anthers at seed, e, i florets at P2, f, j florets at P3, g, k florets at P4 stage, h, l mature
meiotic and tetrad stage suggesting its involvement in florets. Scale bar, 0.5 mm. c a mature leaf, b T.S. stem, c root, d mature
seed, e floret at P3, f floret at P4, g, h floret at P5, i floret at P6, j mature
microsprorogenesis (Fig. 6). Analysis of its meta-profile floret at P6 after anthesis, k dissected P6 floret showing GUS stained
during consolidated data from microarray experiments cov- anthers, l anther at P6, m stigma at P6, n isolated pollen at P6. Scale bar
ering anatomical stages showed that its transcripts are par- (a–k) 1 mm, (l–n) 0.1 mm. d a callus, b mature leaf, c root at pre-
ticularly enriched in 0.7 to 1.0 mm anthers (Supplementary pollination stage, d pollen at P6, e anther at P6, f floret at P6, g gynoecium
at P6, h seed at S1 stage, i 0 DAP seed (S1), j, k 1 DAP seed (S1), l 2 DAP
Figure S9b). The second gene, OsSub42 encodes a seed (S1), m seed at S2 stage, n seed at S3 stage, o seed at S5 stage. Scale
subtilisin-like serine protease detected in P4 and P5 stages. bar, 0.5 mm. “1” represents wild type and “2” denotes the transgenic
The GUS activity was observed mainly in anthers, specifically plant. The scale bar is equal to 0.5 mm
Funct Integr Genomics (2012) 12:229–248 241
242 Funct Integr Genomics (2012) 12:229–248

at single-celled microspore and bicellular pollen stage with establishment of polarity (FE of Ipomea; Iwasaki and
very low-level expression in nodal portion of the stem, stigma Nitasaka 2006), sexual fate of meristems (TS2 of maize;
and embryo (Fig. 6). Upon dissection of anthers at uninucleate DeLong et al. 1993), and brassinosteroid response (BIM2 of
stage, GUS activity was observed in the anther wall as well as Arabidopsis; Yin et al. 2005) were specifically upregulated in
in developing pollen. Meta-profile analysis during ana- P1-I stage; whereas, those showing similarity with genes
tomical stages showed accumulation of its transcripts in associated with circadian clock (LHY and ELF3 of
1.2 to 1.5 mm anthers (Supplementary Figure S9b). The Arabidopsis; Ding et al. 2007; Liu et al. 2001), flowering time
third gene, RTS is an anther-specific gene that expresses control (FPF1 and CO of Arabidopsis; Kania et al. 1997; An
exclusively in tapetum and has been shown to be essen- et al. 2004), panicle branch initiation (RFL of rice; Kyozuka et
tial for pollen development in rice (Luo et al. 2006). The al. 1998), and organ primordia formation (ZmOCL1; Ingram
RTS promoter derived specific expression in post-meiotic et al. 2000) were specifically upregulated in P1-II stage. The
anthers with low level GUS activity in stigma. Meta- list of genes showing P1-II specific upregulation also included
profile analysis revealed its specific expression in 1.6 those involved in meristem organization (TSO1 of
to 2.0 mm anthers (Supplementary Figure S9b). XYH Arabidopsis; Song et al. 2000) and cell specialization in
exhibiting S1 stage-specific expression encodes 1, 4- anthers (TPD1 of Arabidopsis; Yang et al. 2003). Induction
beta-D xylan xylanohydrolase. The GUS activity was of 80 genes was detected in both P1-I and P1-II stages includ-
observed specifically in ovary and style in pollinated ing genes involved in initiation and maintenance of rachis and
panicles suggesting that it might have role during pene- branch meristem (LAX and, OSH1 of rice and JUBEL2 of
tration of pollen tube. Meta-profile analysis during ana- barley; Komatsu et al. 2001; Sentoku et al. 1999; Muller et al.
tomical stages conforms its specific expression in ovaries 2001), panicle branching (OsTB1 of rice; Takeda et al. 2003),
at 1 day after fertilization (Supplementary Figure S9b). whereas those showing upregulation in P1-II–PI-III window
The consistency in the expression profiles of these genes (146 in number) included genes involved in flowering time
in specific developmental events/stages and organs control and floral organ identity (OsMADS1, 2, 6, 7, 8, and 58
strengthens the reliability of our microarray data as well of rice; Agrawal et al. 2005; Chung et al. 1994; Chung et al.
as staging system used. 1995; Greco et al. 1997; Jeon et al. 2008; Kang et al. 1997;
Lee et al. 2003; Prasad et al. 2005; Prasad et al. 2001; Prasad
and Vijayraghavan 2003; Yamaguchi and Hirano 2006;
Discussion Yamaguchi et al. 2006; Jeon et al. 2000), auxiliary meristem
initiation (SPT of Arabidopsis; Komatsu et al. 2003), and
The reproductive phase in rice commences with the transition determination of floral organ number and shape (PNH of
of shoot apical meristem to floral meristem and culminates at Arabidopsis; Lynn et al. 1999).
mature seed about a month after fertilization (Itoh et al. 2005). Similarly, overlaying the data of differentially expressed
Since genes regulating panicle and seed development are genes on RiceCyc database provided important insights into
major determinants of yield, understanding gene regulation genes and pathways that are specifically altered during rice
is crucial for improving yield potential of rice. Here, we reproductive development. For instance, jasmonic acid and
analyzed the genome-wide spatial and temporal expression phenylpropanoid biosynthesis pathways are mainly upregu-
profiles of rice genes during nine stages of panicle and five lated during panicle development and IAA biosynthesis
stages of seed development, demarcated based on the land- during early seed development. Since several genes in-
mark events involved. The transcriptome of each develop- volved in jasmonic acid and phenylpropanoid biosynthesis
mental stage was compared with the vegetative tissues as have already been implicated in anther dehiscence and pol-
well as the preceding stage of development to investigate len development (Ma 2005; Yang et al. 2007; Millar et al.
stage-specific regulation of gene expression. Analysis of 1999; Wilson and Zhang 2009), detailed investigation of the
stage-specific/preferential gene expression along with litera- novel candidate genes, related to these pathways, identified
ture mining revealed that the upregulated genes in our study, in this study would be useful.
during panicle initiation stages (P1-I, P1-II, and P1-III) with Seed development in rice involves embryo and endo-
respect to vegetative controls have previously been implicated sperm development, encompassing cell division, organ ini-
in flowering time control (OsFTL1, OsMADS5, OsMADS14, tiation, and maturation (Agarwal et al. 2011; Itoh et al.
and 15 of rice; Komiya et al. 2008; Jia et al. 2000; Lee et al. 2005). The genes homologous to cell differentiation protein
2003; Moon et al. 1999; Kang and An 1997), meristem RCD1 of wheat and FIE2 of Arabidopsis exhibit S2–S3–S4
organization (CLAVATA1 of Arabidopsis; Clark et al. 1997), stage-specific expression (Danilevskaya et al. 2003; Drea et
and regulating symmetry (DIV of Antirrhinum; Galego and al. 2005; Okazaki et al. 1998). The homolog of Arabidopsis
Almeida 2002), thus adding credibility to our data. Homologs late embryogenesis protein D-34 expresses specifically in
of previously characterized genes for involvement in S2–S3–S4–S5 stages (Hundertmark and Hincha 2008). Rice
Funct Integr Genomics (2012) 12:229–248 243

genes homologous to Arabidopsis MFT and DC-8 of carrot, exhibiting P6 stage-specific expression encode allergens.
implicated in grain maturation and OlE-5 of coffee involved Switching off these genes by knock out mutation/RNAi
in stabilization of oil bodies in embryos were detected approaches can have significant impact on society by reduc-
specifically in S3–S4–S5 stages (Chardon and Damerval ing the allergy cases (Bhalla et al. 2001). Similarly, seed-
2005; Cheng et al. 1996; Simkin et al. 2006). It would be specific promoters could have important implications in
interesting to examine if these genes in rice have similar improving grain quality and yields in cereal crops and have
functions to the ones already deduced in other plants. Also, been exploited for the production of biologically/commer-
the genes involved in auxin distribution and transport have cially relevant products (Aluru et al. 2008; Furtado and
been shown to play a role in establishing embryonic pattern Henry 2005; Furtado et al. 2008; Lamacchia et al. 2001;
in wheat (Fischer-Iglesias et al. 2001). Functional character- Qu le and Takaiwa 2004; Russell and Fromm 1997;
ization of IAA biosynthesis genes identified in this study Sunilkumar et al. 2002).
might, therefore, prove useful in understanding the role of The identification of reproductive stage-specific genes is
IAA in embryo establishment. not trivial. In past few years, many groups have attempted to
The role of gibberellins in floral induction, bolting and understand transcriptional dynamics during rice panicle and
development of anther, pistil, and seed has been well docu- seed development (Endo et al. 2002; Fujita et al. 2010;
mented (Hu et al. 2008). Upregulation of genes involved in Furutani et al. 2006; Jiao et al. 2009; Kondou et al. 2006;
gibberellins biosynthesis during panicle and seed develop- Ma et al. 2005; Sato et al. 2011; Wang et al. 2010).
mental stages reiterates its role in rice reproductive develop- However, very less overlap is observed among the datasets
ment. Some of the genes controlling same step of betanidin generated across different studies. This is largely because of
degradation, brassinosteroid biosynthesis, flavonoid biosyn- differences in tissue sampling and subspecies/varieties used
thesis, homogalacturonon degradation, cytokinin glucoside in different experiments. Moreover, tissue/stage-specific ex-
biosynthesis, and sucrose degradation exhibited either pression is strongly influenced by choice of microarray
panicle- or seed-specific expression suggesting that probably platform and parameters used for data analysis. For instance,
different genes may have taken up similar functions in differ- when we compared our panicle and seed-specific datasets
ent tissue types. with those extracted by Wang et al. (2010) and Sato et al.
The regulation of these metabolic and hormonal pathways (2011), fewer genes were common between datasets gener-
at transcriptional level is partly executed by specific transcrip- ated from different rice subspecies. Since we used prepro-
tion factors. Therefore, spatial and temporal expression pat- cessed datasets by both the groups for comparison, different
terns of genes encoding transcription factors can reveal the methodologies used would be another cause of less redun-
target genes serving as nucleation points for building gene dancy in these datasets. For extracting specific datasets, both
regulatory networks. The list of seed-specific transcription Wang et al. (2010) and Sato et al. (2011) have used Shannon
factor genes included those belonging to MYB and NAM entropy values [<3 and <4.5, respectively] as a measure to
families followed by noteworthy representation from AP2, assess tissue-specificity. Sato et al. (2011) further filtered the
AUX_IAA, bZIP, C3H, HSF, and MADS-box families. data by applying a stringent minimum expression value (8 in
Many genes belonging to these families are critical for the log2) cut-off in at least one of stages interrogated. Since
process of seed development (Agarwal et al. 2011). Similarly, Shannon entropy values do not discriminate between bio-
significant enrichment of bZIP, MADS, and C2H2 families logically relevant stages, a large number of genes with
during panicle development highlights potential sites of their relatively high expression in non-reproductive stages have
function. Therefore, this analysis has revealed a large dataset also been shortlisted in both these reports. We, however,
of candidate genes, which apparently play important roles in worked on log2 normalized expression value cut-offs of <15
different stages of panicle and seed development. (3.9 in log2, for a gene to be called not-expressed) and >50
The gene expression dataset presented here would also be (5.6 in log2, for a gene to be called expressed). Selection of
a useful resource to mine candidates for promoter function these cut-offs was based on normalized signal intensity
validation. The genes exhibiting P3–P6 specific high ex- values of non-rice probes (Magnaporthe grisea) on the rice
pression are likely to have anther-specific expression. The Affymetrix chip (Deveshwar et al. 2011), which was used to
promoters of these genes once verified in transgenic system discriminate between expressed and non-expressed genes.
can be used to generate male-sterile lines for hybrid seed In spite of all these differences, significant number of genes
production (Gupta et al. 2007; Perez-Prat and van Lookeren was common between the panicle/seed-specific datasets
Campagne 2002; Roque et al. 2007; Twell et al. 1990; Xu et extracted in all three studies (Supplementary Figure S7).
al. 2006). Since meiosis is the main source of variation These genes would likely have higher reproducibility and
because of recombination, halting meiosis by targeting P2 thus would serve as candidates for detailed investigation
and P3-specific genes could be another approach to multiply (Supplementary Table S4). By generating meta-profiles of
the elite varieties without variations. Many of the genes shortlisted genes/datasets, we have shown that large-scale
244 Funct Integr Genomics (2012) 12:229–248

meta-analysis by integrating data generated in independent Agrawal GK, Abe K, Yamazaki M, Miyao A, Hirochika H (2005)
studies can provide cross validation of gene expression. Conservation of the E-function for floral organ identity in rice
revealed by the analysis of tissue culture-induced loss-of-function
The next challenge would be to decipher genetic regula- mutants of the OsMADS1 gene. Plant Mol Biol 59:125–135
tory networks by integrating existing information resources. Aluru M, Xu Y, Guo R, Wang Z, Li S, White W, Wang K, Rodermel S
Coupling multiple datasets with co-expression analysis will (2008) Generation of transgenic maize with enhanced provitamin
help in elucidating gene clusters underlying biological pro- A content. J Exp Bot 59:3551–3562
Alves-Ferreira M, Wellmer F, Banhara A, Kumar V, Riechmann JL,
cesses. As a first step towards this, we have made an attempt Meyerowitz EM (2007) Global expression profiling applied to the
to analyze the correlation between expression and physical analysis of Arabidopsis stamen development. Plant Physiol
location of genes on the genome. The close placement of 145:747–762
2,792 genes on rice chromosomes comprising a total of An H, Roussot C, Suarez-Lopez P, Corbesier L, Vincent C, Pineiro M,
Hepworth S, Mouradov A, Justin S, Turnbull C, Coupland G
1,278 GCSEs suggests a link between genome organization (2004) CONSTANS acts in the phloem to regulate a systemic
and regulation of expression. Though many of these regions signal that induces photoperiodic flowering of Arabidopsis. De-
could also have resulted due to tandem duplications, pres- velopment 131:3615–3626
ence of different set of genes in a cluster suggested that Arora R, Agarwal P, Ray S, Singh AK, Singh VP, Tyagi AK, Kapoor S
(2007) MADS-box gene family in rice: genome-wide identifica-
these might have co-evolved for regulation of similar bio- tion, organization and expression profiling during reproductive
logical processes. Several groups have tried to explain the development and stress. BMC Genomics 8:242
phenomenon based on tandem duplications, overlapping Baltz R, Domon C, Pillay DT, Steinmetz A (1992) Characterization of
genes, chromatin domains, shared regulatory elements, clus- a pollen-specific cDNA from sunflower encoding a zinc finger
protein. Plant J 2:713–721
tering of genes involved in similar metabolic pathways, etc. Becerra C, Puigdomenech P, Vicient CM (2006) Computational and
but so far no single factor could explain it (Michalak 2008). experimental analysis identifies Arabidopsis genes specifically
Further analysis of these genes will reveal if they are in- expressed during early seed development. BMC Genomics 7:38
volved in a regulatory cascade or their expression could be Becker JD, Feijo JA (2007) How many genes are needed to make a
pollen tube? Lessons from transcriptomics. Ann Bot (Lond)
attributed to the ripple effect of transcriptional activation, 100:1117–1123
where intensive transcription activity at one locus may cause Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate-a
upregulation of its physically neighboring loci (Carninci practical and powerful approach to multiple testing. J R Stat Soc
2008; Ebisuya et al. 2008). Ser B 57:289–300
Bhalla PL, Swoboda I, Singh MB (2001) Reduction in allergenicity of
grass pollen by genetic engineering. Int Arch Allergy Immunol
Acknowledgements This work was supported by Department of 124:51–54
Biotechnology, Ministry of Science & Technology, Government of Cao PJ, Bartley LE, Jung KH, Ronald PC (2008) Construction of a rice
India (Project No. BT/AB/FG-I(PH-II)(4)/2009). We acknowledge glycosyltransferase phylogenomic database and identification of
Dr. Ramesh Hariharan and his team at Strand LS Bengaluru, India rice-diverged glycosyltransferases. Mol Plant 1:858–877
for their help in microarray data analysis and Ms. Manupriya for Carninci P (2008) Non-coding RNA transcription: turning on neigh-
providing the list of transcription factor family genes in rice. Senior bours. Nat Cell Biol 10:1023–1024
Research fellowship by the Council for Scientific and Industrial Re- Chardon F, Damerval C (2005) Phylogenomic analysis of the PEBP
search (CSIR) to R.S., S.R., P.D, M.J., A.N., and P.S. and University gene family in cereals. J Mol Evol 61:579–590
Grants Commissions (UGC) fellowship to P.A. are acknowledged. Cheng JC, Seeley KA, Goupil P, Sung ZR (1996) Expression of DC8 is
Microarray data used in this study have been deposited in the Gene associated with, but not dependent on embryogenesis. Plant Mol
Expression Omnibus database at the National Center for Biotechnolo- Biol 31:127–141
gy Information under the accession nos. GSE6893 and GSE6901. All Chung YY, Kim SR, Finkel D, Yanofsky MF, An G (1994) Early
the datasets shortlisted in this manuscript including list of panicle or flowering and reduced apical dominance result from ectopic ex-
seed-specific genes, differentially expressed genes during panicle or pression of a rice MADS box gene. Plant Mol Biol 26:657–665
seed development with respect to all four vegetative stages and unique Chung YY, Kim SR, Kang HG, Noh YS, Park MC, Finkel D, An G
genes upregulated with respect to preceding stage of development are (1995) Characterization of two MADS box genes homologous to
given in Supplementary Table S3. GLOBOSA. Plant Sci 109:45–56
Clark SE, Williams RW, Meyerowitz EM (1997) The CLAVATA1 gene
encodes a putative receptor kinase that controls shoot and floral
meristem size in Arabidopsis. Cell 89:575–585
References Clouse SD, Sasse JM (1998) Brassinosteroids: essential regulators of
plant growth and development. Annu Rev Plant Physiol Plant Mol
Biol 49:427–451
Adams (2008) Transcriptome: connecting the genome to gene func- Curtis MD, Grossniklaus U (2003) A gateway cloning vector set for
tion. Nature Education 1 high-throughput functional analysis of genes in planta. Plant
Agarwal P, Arora R, Ray S, Singh AK, Singh VP, Takatsuji H, Kapoor Physiol 133:462–469
S, Tyagi AK (2007) Genome-wide identification of C2H2 zinc- Danilevskaya ON, Hermon P, Hantke S, Muszynski MG, Kollipara K,
finger gene family in rice and their phylogeny and expression Ananiev EV (2003) Duplicated fie genes in maize: expression pattern
analysis. Plant Mol Biol 65:467–485 and imprinting suggest distinct functions. Plant Cell 15:425–438
Agarwal P, Kapoor S, Tyagi AK (2011) Transcription factors regulat- Davidson SE, Elliott RC, Helliwell CA, Poole AT, Reid JB (2003) The
ing the progression of monocot and dicot seed development. pea gene NA encodes ent-kaurenoic acid oxidase. Plant Physiol
Bioessays 33:189–202 131:335–344
Funct Integr Genomics (2012) 12:229–248 245

Day RC, Herridge RP, Ambrose BA, Macknight RC (2008) Tran- ordinated transcriptional networks of the developing grain of
scriptome analysis of proliferating Arabidopsis endosperm reveals field-grown barley. J Exp Bot 60:153–167
biological implications for the control of syncytial division, cyto- Hennig L, Gruissem W, Grossniklaus U, Kohler C (2004) Transcrip-
kinin signaling, and gene expression regulation. Plant Physiol tional programs of early reproductive stages in Arabidopsis. Plant
148:1964–1984 Physiol 135:1765–1775
DeLong A, Calderon-Urrea A, Dellaporta SL (1993) Sex determination Hirano K, Aya K, Hobo T, Sakakibara H, Kojima M, Shim RA,
gene TASSELSEED2 of maize encodes a short-chain alcohol Hasegawa Y, Ueguchi-Tanaka M, Matsuoka M (2008) Compre-
dehydrogenase required for stage-specific floral organ abortion. hensive transcriptome analysis of phytohormone biosynthesis and
Cell 74:757–768 signaling genes in microspore/pollen and tapetum of rice. Plant
Deveshwar P, Bovill WD, Sharma R, Able JA, Kapoor S (2011) Analysis Cell Physiol 49:1429–1450
of anther transcriptomes to identify genes contributing to meiosis Hobo T, Suwabe K, Aya K, Suzuki G, Yano K, Ishimizu T, Fujita M,
and male gametophyte development in rice. BMC Plant Biol 11:78 Kikuchi S, Hamada K, Miyano M, Fujioka T, Kaneko F, Kazama
Ding Z, Doyle MR, Amasino RM, Davis SJ (2007) A complex genetic T, Mizuta Y, Takahashi H, Shiono K, Nakazono M, Tsutsumi N,
interaction between Arabidopsis thaliana TOC1 and CCA1/LHY Nagamura Y, Kurata N, Watanabe M, Matsuoka M (2008) Various
in driving the circadian clock and in output regulation. Genetics spatiotemporal expression profiles of anther-expressed genes in
176:1501–1510 rice. Plant Cell Physiol 49:1417–1428
Drea S, Leader DJ, Arnold BC, Shaw P, Dolan L, Doonan JH (2005) Howell KA, Narsai R, Carroll A, Ivanova A, Lohse M, Usadel B,
Systematic spatial analysis of gene expression during wheat cary- Millar AH, Whelan J (2009) Mapping metabolic and transcript
opsis development. Plant Cell 17:2172–2185 temporal switches during germination in rice highlights specific
Ebisuya M, Yamamoto T, Nakajima M, Nishida E (2008) Ripples from transcription factors and the role of RNA instability in the germi-
neighbouring transcription. Nat Cell Biol 10:1106–1113 nation process. Plant Physiol 149:961–980
Endo M, Matsubara H, Kokubun T, Masuko H, Takahata Y, Tsuchiya Hu J, Mitchum MG, Barnaby N, Ayele BT, Ogawa M, Nam E, Lai
T, Fukuda H, Demura T, Watanabe M (2002) The advantages of WC, Hanada A, Alonso JM, Ecker JR, Swain SM, Yamaguchi S,
cDNA microarray as an effective tool for identification of repro- Kamiya Y, Sun TP (2008) Potential sites of bioactive gibberellin
ductive organ-specific genes in a model legume, Lotus japonicus. production during reproductive growth in Arabidopsis. Plant Cell
FEBS Lett 514:229–237 20:320–336
Endo M, Tsuchiya T, Saito H, Matsubara H, Hakozaki H, Masuko H, Hundertmark M, Hincha DK (2008) LEA (late embryogenesis abun-
Kamada M, Higashitani A, Takahashi H, Fukuda H, Demura T, dant) proteins and their encoding genes in Arabidopsis thaliana.
Watanabe M (2004) Identification and molecular characterization BMC Genomics 9:118
of novel anther-specific genes in Oryza sativa L. by using cDNA Ikeda K, Sunohara H, Nagato Y (2004) Developmental course of
microarray. Genes Genet Syst 79:213–226 inflorescence and spikelet in rice. Breed Sci 54:147–156
Fait A, Angelovici R, Less H, Ohad I, Urbanczyk-Wochniak E, Fernie Ingram GC, Boisnard-Lorig C, Dumas C, Rogowsky PM (2000) Ex-
AR, Galili G (2006) Arabidopsis seed development and germina- pression patterns of genes encoding HD-ZipIV homeo domain
tion is associated with temporally distinct metabolic switches. proteins define specific domains in maize embryos and meris-
Plant Physiol 142:839–854 tems. Plant J 22:401–414
Fischer-Iglesias C, Sundberg B, Neuhaus G, Jones AM (2001) Auxin Itoh J, Nonomura K, Ikeda K, Yamaki S, Inukai Y, Yamagishi H,
distribution and transport during embryonic pattern formation in Kitano H, Nagato Y (2005) Rice plant development: from zygote
wheat. Plant J 26:115–129 to spikelet. Plant Cell Physiol 46:23–47
Fujita M, Horiuchi Y, Ueda Y, Mizuta Y, Kubo T, Yano K, Yamaki S, Iwasaki M, Nitasaka E (2006) The FEATHERED gene is required for
Tsuda K, Nagata T, Niihama M, Kato H, Kikuchi S, Hamada K, polarity establishment in lateral organs especially flowers of the
Mochizuki T, Ishimizu T, Iwai H, Tsutsumi N, Kurata N (2010) Japanese morning glory (Ipomoea nil). Plant Mol Biol 62:913–925
Rice expression atlas in reproductive development. Plant Cell Jain M, Khurana JP (2009) Transcript profiling reveals diverse roles of
Physiol 51:2060–2081 auxin-responsive genes during reproductive development and
Furtado A, Henry RJ (2005) The wheat Em promoter drives reporter abiotic stress in rice. FEBS J 276:3148–3162
gene expression in embryo and aleurone tissue of transgenic Jain M, Kaur N, Garg R, Thakur JK, Tyagi AK, Khurana JP (2006)
barley and rice. Plant Biotechnol J 3:421–434 Structure and expression analysis of early auxin-responsive Aux/
Furtado A, Henry RJ, Takaiwa F (2008) Comparison of promoters in IAA gene family in rice (Oryza sativa). Funct Integr Genomics
transgenic rice. Plant Biotechnol J 6:679–693 6:47–59
Furutani I, Sukegawa S, Kyozuka J (2006) Genome-wide analysis of Jain M, Nijhawan A, Arora R, Agarwal P, Ray S, Sharma P, Kapoor S,
spatial and temporal gene expression in rice panicle development. Tyagi AK, Khurana JP (2007) F-box proteins in rice. Genome-
Plant J 46:503–511 wide analysis, classification, temporal and spatial gene expression
Galego L, Almeida J (2002) Role of DIVARICATA in the control of during panicle and seed development, and regulation by light and
dorsoventral asymmetry in Antirrhinum flowers. Genes Dev abiotic stress. Plant Physiol 143:1467–1483
16:880–891 Jain M, Tyagi AK, Khurana JP (2008) Genome-wide identification,
Greco R, Stagi L, Colombo L, Angenent GC, Sari-Gorla M, Pe ME classification, evolutionary expansion and expression analyses of
(1997) MADS box genes expressed in developing inflorescences homeobox genes in rice. FEBS J 275:2845–2861
of rice and sorghum. Mol Gen Genet 253:615–623 Jaiswal P, Ni J, Yap I, Ware D, Spooner W, Youens-Clark K, Ren L,
Grimanelli D, Perotti E, Ramirez J, Leblanc O (2005) Timing of the Liang C, Zhao W, Ratnapu K, Faga B, Canaran P, Fogleman M,
maternal-to-zygotic transition during early seed development in Hebbard C, Avraham S, Schmidt S, Casstevens TM, Buckler ES,
maize. Plant Cell 17:1061–1072 Stein L, McCouch S (2006) Gramene: a bird’s eye view of cereal
Gupta V, Khurana R, Tyagi AK (2007) Promoters of two anther- genomes. Nucleic Acids Res 34:D717–D723
specific genes confer organ-specific gene expression in a stage- Jefferson RA, Kavanagh TA, Bevan MW (1987) GUS fusions: beta-
specific manner in transgenic systems. Plant Cell Rep 26:1919– glucuronidase as a sensitive and versatile gene fusion marker in
1931 higher plants. EMBO J 6:3901–3907
Hansen M, Friis C, Bowra S, Holm PB, Vincze E (2009) A pathway- Jeon JS, Lee S, Jung KH, Yang WS, Yi GH, Oh BG, An G (2000)
specific microarray analysis highlights the complex and co- Production of transgenic rice plants showing reduced heading
246 Funct Integr Genomics (2012) 12:229–248

date and plant height by ectopic expression of rice MADS-box Li Z, Zhang H, Ge S, Gu X, Gao G, Luo J (2009) Expression pattern
genes. Mol Breed 6:581–592 divergence of duplicated genes in rice. BMC Bioinforma 10
Jeon JS, Lee S, An G (2008) Intragenic control of expression of a rice (Suppl 6):S8
MADS box gene OsMADS1. Mol Cells 26:474–480 Liu XL, Covington MF, Fankhauser C, Chory J, Wagner DR (2001)
Jia H, Chen R, Cong B, Cao K, Sun C, Luo D (2000) Characterization ELF3 encodes a circadian clock-regulated nuclear protein that
and transcriptional profiles of two rice MADS-box genes. Plant functions in an Arabidopsis PHYB signal transduction pathway.
Sci 155:115–122 Plant Cell 13:1293–1304
Jiang SY, Christoffels A, Ramamoorthy R, Ramachandran S (2009) Liu X, Fu J, Gu D, Liu W, Liu T, Peng Y, Wang J, Wang G (2008)
Expansion mechanisms and functional annotations of hypotheti- Genome-wide analysis of gene expression profiles during the
cal genes in the rice genome. Plant Physiol 150:1997–2008 kernel development of maize (Zea mays L.). Genomics 91:378–
Jiao Y, Tausta SL, Gandotra N, Sun N, Liu T, Clay NK, Ceserani T, Chen 387
M, Ma L, Holford M, Zhang HY, Zhao H, Deng XW, Nelson T Luo H, Lee JY, Hu Q, Nelson-Vasilchik K, Eitas TK, Lickwar C,
(2009) A transcriptome atlas of rice cell types uncovers cellular, Kausch AP, Chandlee JM, Hodges TK (2006) RTS, a rice
functional and developmental hierarchies. Nat Genet 41:258–263 anther-specific gene is required for male fertility and its promoter
Kang HG, An G (1997) Isolation and characterization of a rice MADS sequence directs tissue-specific gene expression in different plant
box gene belonging to the AGL2 gene family. Mol Cells 7:45–51 species. Plant Mol Biol 62:397–408
Kang HG, Jang S, Chung JE, Cho YG, An G (1997) Characterization Lynn K, Fernandez A, Aida M, Sedbrook J, Tasaka M, Masson P,
of two rice MADS box genes that control flowering time. Mol Barton MK (1999) The PINHEAD/ZWILLE gene acts pleiotropi-
Cells 7:559–566 cally in Arabidopsis development and has overlapping functions
Kania T, Russenberger D, Peng S, Apel K, Melzer S (1997) FPF1 with the ARGONAUTE1 gene. Development 126:469–481
promotes flowering in Arabidopsis. Plant Cell 9:1327–1338 Ma H (2005) Molecular genetic analyses of microsporogenesis and
Komatsu M, Maekawa M, Shimamoto K, Kyozuka J (2001) The LAX1 microgametogenesis in flowering plants. Annu Rev Plant Biol
and FRIZZY PANICLE 2 genes determine the inflorescence archi- 56:393–434
tecture of rice by controlling rachis-branch and spikelet develop- Ma H, Zhao J (2010) Genome-wide identification, classification, and
ment. Dev Biol 231:364–373 expression analysis of the arabinogalactan protein gene family in
Komatsu K, Maekawa M, Ujiie S, Satake Y, Furutani I, Okamoto H, rice (Oryza sativa L.). J Exp Bot 61:2647–2668
Shimamoto K, Kyozuka J (2003) LAX and SPA: major regulators of Ma L, Chen C, Liu X, Jiao Y, Su N, Li L, Wang X, Cao M, Sun N,
shoot branching in rice. Proc Natl Acad Sci USA 100:11765–11770 Zhang X, Bao J, Li J, Pedersen S, Bolund L, Zhao H, Yuan L,
Komiya R, Ikegami A, Tamaki S, Yokoi S, Shimamoto K (2008) Hd3a and Wong GK, Wang J, Deng XW (2005) A microarray analysis of
RFT1 are essential for flowering in rice. Development 135:767–774 the rice transcriptome and its comparison to Arabidopsis. Genome
Kondou H, Ooka H, Yamada H, Satoh K, Kikuchi S, Takahara Y, Res 15:1274–1283
Yamamoto K (2006) Microarray analysis of gene expression at Madera M, Gough J (2002) A comparison of profile hidden Markov
initial stages of rice seed development. Breed Sci 56:235–242 model procedures for remote homology detection. Nuc Acids Res
Kyozuka J, Konishi S, Nemoto K, Izawa T, Shimamoto K (1998) 30:4321–4328
Down-regulation of RFL, the FLO/LFY homolog of rice, accom- Michalak P (2008) Coexpression, coregulation, and cofunctionality of
panied with panicle branch initiation. Proc Natl Acad Sci USA neighboring genes in eukaryotic genomes. Genomics 91:243–248
95:1979–1982 Millar AA, Clemens S, Zachgo S, Giblin EM, Taylor DC, Kunst L
Laitinen RA, Immanen J, Auvinen P, Rudd S, Alatalo E, Paulin L, (1999) CUT1, an Arabidopsis gene required for cuticular wax
Ainasoja M, Kotilainen M, Koskela S, Teeri TH, Elomaa P (2005) biosynthesis and pollen fertility, encodes a very-long-chain fatty
Analysis of the floral transcriptome uncovers new regulators of acid condensing enzyme. Plant Cell 11:825–838
organ determination and gene families related to flower organ Mohanty A, Sarma NP, Tyagi AK (1999) Agrobacterium-mediated
differentiation in Gerbera hybrida (Asteraceae). Genome Res high frequency transformation of an elite indica rice variety Pusa
15:475–486 Basmati 1 and transmission of the transgene to R2 progeny. Plant
Lamacchia C, Shewry PR, Di Fonzo N, Forsyth JL, Harris N, Lazzeri Sci 147:125–135
PA, Napier JA, Halford NG, Barcelo P (2001) Endosperm- Moon YH, Kang HG, Jung JY, Jeon JS, Sung SK, An G (1999)
specific activity of a storage protein gene promoter in transgenic Determination of the motif responsible for interaction between
wheat seed. J Exp Bot 52:243–250 the rice APETALA1/AGAMOUS-LIKE9 family proteins using a
Lan L, Chen W, Lai Y, Suo J, Kong Z, Li C, Lu Y, Zhang Y, Zhao X, yeast two-hybrid system. Plant Physiol 120:1193–1204
Zhang X, Han B, Cheng J, Xue Y (2004) Monitoring of gene Muller J, Wang Y, Franzen R, Santi L, Salamini F, Rohde W (2001) In
expression profiles and isolation of candidate genes involved in vitro interactions between barley TALE homeodomain proteins
pollination and fertilization in rice (Oryza sativa L.) with a 10 K suggest a role for protein-protein associations in the regulation of
cDNA microarray. Plant Mol Biol 54:471–487 Knox gene function. Plant J 27:13–23
Lee JM, Williams ME, Tingey SV, Rafalski JA (2002) DNA array Nijhawan A, Jain M, Tyagi AK, Khurana JP (2008) Genomic survey
profiling of gene expression changes during maize embryo devel- and gene expression analysis of the basic leucine zipper transcrip-
opment. Funct Integr Genomics 2:13–27 tion factor family in rice. Plant Physiol 146:333–350
Lee S, Kim J, Son JS, Nam J, Jeong DH, Lee K, Jang S, Yoo J, Lee J, Lee Okazaki N, Okazaki K, Watanabe Y, Kato-Hayashi M, Yamamoto M,
DY, Kang HG, An G (2003) Systematic reverse genetic screening of Okayama H (1998) Novel factor highly conserved among eukar-
T-DNA tagged genes in rice for functional genomic analyses: yotes controls sexual development in fission yeast. Mol Cell Biol
MADS-box genes as a test case. Plant Cell Physiol 44:1403–1411 18:887–895
Li L, Wang X, Sasidharan R, Stolc V, Deng W, He H, Korbel J, Chen Perez-Prat E, van Lookeren Campagne MM (2002) Hybrid seed pro-
X, Tongprasit W, Ronald P, Chen R, Gerstein M, Deng XW duction and the challenge of propagating male-sterile plants.
(2007a) Global identification and characterization of transcrip- Trends Plant Sci 7:199–203
tionally active regions in the rice genome. PLoS One 2:e294 Prasad K, Vijayraghavan U (2003) Double-stranded RNA interference
Li M, Xu W, Yang W, Kong Z, Xue Y (2007b) Genome-wide gene of a rice PI/GLO paralog, OsMADS2, uncovers its second-whorl-
expression profiling reveals conserved and novel molecular func- specific function in floral organ patterning. Genetics 165:2301–
tions of the stigma in rice. Plant Physiol 144:1797–1812 2305
Funct Integr Genomics (2012) 12:229–248 247

Prasad K, Sriram P, Kumar CS, Kushalappa K, Vijayraghavan U Suwabe K, Suzuki G, Takahashi H, Shiono K, Endo M, Yano K,
(2001) Ectopic expression of rice OsMADS1 reveals a role in Fujita M, Masuko H, Saito H, Fujioka T, Kaneko F, Kazama
specifying the lemma and palea, grass floral organs analogous to T, Mizuta Y, Kawagishi-Kobayashi M, Tsutsumi N, Kurata
sepals. Dev Genes Evol 211:281–290 N, Nakazono M, Watanabe M (2008) Separated transcrip-
Prasad K, Parameswaran S, Vijayraghavan U (2005) OsMADS1, a rice tomes of male gametophyte and tapetum in rice: validity of
MADS-box factor, controls differentiation of specific cell types in a laser microdissection (LM) microarray. Plant Cell Physiol
the lemma and palea and is an early-acting regulator of inner 49:1407–1416
floral organs. Plant J 43:915–928 Takeda T, Suwa Y, Suzuki M, Kitano H, Ueguchi-Tanaka M, Ashikari
Qu le Q, Takaiwa F (2004) Evaluation of tissue specificity and expres- M, Matsuoka M, Ueguchi C (2003) The OsTB1 gene negatively
sion strength of rice seed component gene promoters in transgenic regulates lateral branching in rice. Plant J 33:513–520
rice. Plant Biotechnol J 2:113–125 Tebbji F, Nantel A, Matton DP (2010) Transcription profiling of
Raghavan V (1988) Anther and pollen development in rice (Oryza fertilization and early seed development events in a solanaceous
sativa). Amer J Bot 75:183–196 species using a 7.7 K cDNA microarray from Solanum chacoense
Ray S, Dansana PK, Giri J, Deveshwar P, Arora R, Agarwal P, Khurana ovules. BMC Plant Biol 10:174
JP, Kapoor S, Tyagi AK (2011) Modulation of transcription factor Toki S, Hara N, Ono K, Onodera H, Tagiri A, Oka S, Tanaka H (2006)
and metabolic pathway genes in response to water-deficit stress in Early infection of scutellum tissue with Agrobacterium allows
rice. Funct Integr Genomics 11:157–178 high-speed transformation of rice. Plant J 47:969–976
Ren XY, Fiers MW, Stiekema WJ, Nap JP (2005) Local coexpression Twell D, Yamaguchi J, McCormick S (1990) Pollen-specific gene
domains of two to four genes in the genome of Arabidopsis. Plant expression in transgenic plants: coordinate regulation of two
Physiol 138:923–934 different tomato gene promoters during microsporogenesis. De-
Ren XY, Stiekema WJ, Nap JP (2007) Local coexpression domains in velopment 109:705–713
the genome of rice show no microsynteny with Arabidopsis Wang Z, Liang Y, Li C, Xu Y, Lan L, Zhao D, Chen C, Xu Z, Xue Y,
domains. Plant Mol Biol 65:205–217 Chong K (2005) Microarray analysis of gene expression involved
Riano-Pachon DM, Ruzicic S, Dreyer I, Mueller-Roeber B (2007) in anther development in rice (Oryza sativa L.). Plant Mol Biol
PlnTFDB: an integrative plant transcription factor database. 58:721–737
BMC Bioinforma 8:42 Wang L, Xie W, Chen Y, Tang W, Yang J, Ye R, Liu L, Lin Y,
Roque E, Gomez MD, Ellul P, Wallbraun M, Madueno F, Beltran JP, Xu C, Xiao J, Zhang Q (2010) A dynamic gene expression
Canas LA (2007) The PsEND1 promoter: a novel tool to produce atlas covering the entire life cycle of rice. Plant J 61:752–
genetically engineered male-sterile plants by early anther ablation. 766
Plant Cell Rep 26:313–325 Wellmer F, Riechmann JL, Alves-Ferreira M, Meyerowitz EM (2004)
Russell DA, Fromm ME (1997) Tissue-specific expression in trans- Genome-wide analysis of spatial gene expression in Arabidopsis
genic maize of four endosperm promoters from maize and rice. flowers. Plant Cell 16:1314–1326
Transgenic Res 6:157–168 Wellmer F, Alves-Ferreira M, Dubois A, Riechmann JL, Meyerowitz
Sato Y, Antonio B, Namiki N, Motoyama R, Sugimoto K, Takehisa H, EM (2006) Genome-wide analysis of gene expression during
Minami H, Kamatsuki K, Kusaba M, Hirochika H, Nagamura Y early Arabidopsis flower development. PLoS Genet 2:e117
(2011) Field transcriptome revealed critical developmental and Wilson ZA, Zhang DB (2009) From Arabidopsis to rice: pathways in
physiological transitions involved in the expression of growth pollen development. J Exp Bot 60:1479–1492
potential in japonica rice. BMC Plant Biol 11:10 Wilson ID, Barker GL, Lu C, Coghill JA, Beswick RW, Lenton
Sentoku N, Sato Y, Kurata N, Ito Y, Kitano H, Matsuoka M (1999) Regional JR, Edwards KJ (2005a) Alteration of the embryo transcrip-
expression of the rice KN1-type homeobox gene family during em- tome of hexaploid winter wheat (Triticum aestivum cv.
bryo, shoot, and flower development. Plant Cell 11:1651–1664 Mercia) during maturation and germination. Funct Integr
Sharma R, Mohan Singh RK, Malik G, Deveshwar P, Tyagi AK, Genomics 5:144–154
Kapoor S, Kapoor M (2009) Rice cytosine DNA methyltrans- Wilson IW, Kennedy GC, Peacock JW, Dennis ES (2005b) Micro-
ferases—gene expression profiling during reproductive develop- array analysis reveals vegetative molecular phenotypes of
ment and abiotic stress. FEBS J 276:6301–6311 Arabidopsis flowering-time mutants. Plant Cell Physiol
Sharma R, Kapoor M, Tyagi AK, Kapoor S (2010) Comparative 46:1190–1201
transcript profiling of TCP family genes provide insight into gene Wu Z, Irizarry RA, Gentleman R, Murillo FM, Spencer F (2003) A
functions and diversification in rice and Arabidopsis. J Plant Mol model based background adjustment for oligonucleotide expres-
Biol Biotechnol 1:24–38 sion arrays. Technical Report, Department of Biostatistics. Work-
Simkin AJ, Qian T, Caillet V, Michoux F, Ben Amor M, Lin C, Tanksley ing Papers, Baltimore, MD.
S, McCarthy J (2006) Oleosin gene family of Coffea canephora: Xu BY, Liu G, Jin ZQ (2006) Isolation, sequencing analysis and
quantitative expression analysis of five oleosin genes in developing characterization of the promoter of banana lectin gene. Sheng
and germinating coffee grain. J Plant Physiol 163:691–708 Wu Gong Cheng Xue Bao 22:945–949
Singh A, Giri J, Kapoor S, Tyagi AK, Pandey GK (2010) Protein Yamaguchi T, Hirano HY (2006) Function and diversification of
phosphatase complement in rice: genome-wide identification MADS-box genes in rice. Scientific World J 6:1923–1932
and transcriptional analysis under abiotic stress conditions and Yamaguchi T, Lee DY, Miyao A, Hirochika H, An G, Hirano HY
reproductive development. BMC Genomics 11:435 (2006) Functional diversification of the two C-class MADS box
Smyth DR, Bowman JL, Meyerowitz EM (1990) Early flower devel- genes OsMADS3 and OsMADS58 in Oryza sativa. Plant Cell
opment in Arabidopsis. Plant Cell 2:755–767 18:15–28
Song JY, Leung T, Ehler LK, Wang C, Liu Z (2000) Regulation of Yang SL, Xie LF, Mao HZ, Puah CS, Yang WC, Jiang L, Sundar-
meristem organization and cell division by TSO1, an Arabidopsis esan V, Ye D (2003) Tapetum determinant1 is required for cell
gene with cysteine-rich repeats. Development 127:2207–2217 specialization in the Arabidopsis anther. Plant Cell 15:2792–
Sunilkumar G, Connell JP, Smith CW, Reddy AS, Rathore KS (2002) 2804
Cotton alpha-globulin promoter: isolation and functional character- Yang C, Vizcay-Barrena G, Conner K, Wilson ZA (2007) MALE
ization in transgenic cotton, Arabidopsis, and tobacco. Transgenic STERILITY1 is required for tapetal development and pollen wall
Res 11:347–359 biosynthesis. Plant Cell 19:3530–3548
248 Funct Integr Genomics (2012) 12:229–248

Yi Y, Mirosevich J, Shyr Y, Matusik R, George AL Jr (2005) Coupled Zhan S, Horrocks J, Lukens LN (2006) Islands of co-expressed neigh-
analysis of gene expression and chromosomal location. Genomics bouring genes in Arabidopsis thaliana suggest higher-order chro-
85:401–412 mosome domains. Plant J 45:347–357
Yin Y, Vafeados D, Tao Y, Yoshida S, Asami T, Chory J (2005) Zhang X, Feng B, Zhang Q, Zhang D, Altman N, Ma H (2005)
A new class of transcription factors mediates brassinosteroid- Genome-wide expression profiling and identification of gene
regulated gene expression in Arabidopsis. Cell 120:249– activities during early flower development in Arabidopsis. Plant
259 Mol Biol 58:401–419