Content uploaded by Joshua A Udall
Author content
All content in this area was uploaded by Joshua A Udall on Jan 04, 2014
Content may be subject to copyright.
Copyright Ó2009 by the Genetics Society of America
DOI: 10.1534/genetics.109.102608
Reciprocal Silencing, Transcriptional Bias and Functional Divergence of
Homeologs in Polyploid Cotton (Gossypium)
Bhupendra Chaudhary,*
,1
Lex Flagel,*
,1
Robert M. Stupar,
†
Joshua A. Udall,
‡
Neetu Verma,*
Nathan M. Springer
§
and Jonathan F. Wendel*
,2
*Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, Iowa 50011,
†
Departments of Agronomy and
Plant Genetics and
§
Plant Biology, University of Minnesota, St. Paul, Minnesota 55108 and
‡
Department of
Plant and Wildlife Sciences, Brigham Young University, Provo, Utah 84602
Manuscript received March 10, 2009
Accepted for publication April 2, 2009
ABSTRACT
Polyploidy is an important force in the evolution of flowering plants. Genomic merger and doubling
induce an extensive array of genomic effects, including immediate and long-term alterations in the
expression of duplicate genes (‘‘homeologs’’). Here we employed a novel high-resolution, genome-specific,
mass-spectrometry technology and a well-established phylogenetic framework to investigate relative
expression levels of each homeolog for 63 gene pairs in 24 tissues in naturally occurring allopolyploid
cotton (Gossypium L.), a synthetic allopolyploid of the same genomic composition, and models of the
diploid progenitor species. Results from a total of 2177 successful expression assays permitted us to
determine the extent of expression evolution accompanying genomic merger of divergent diploid parents,
genome doubling, and genomic coevolution in a common nucleus subsequent to polyploid formation. We
demonstrate that 40% of homeologs are transcriptionally biased in at least one stage of cotton development,
that genome merger per se has a large effect on relative expression of homeologs, and that the majority of
these alterations are caused by cis-regulatory divergence between the diploid progenitors. We describe the
scope of transcriptional subfunctionalization and 15 cases of probable neofunctionalization among 8 tissues.
To our knowledge, this study represents the first characterization of transcriptional neofunctionalization in
an allopolyploid. These results provide a novel temporal perspective on expression evolution of duplicate
genomes and add to our understanding of the importance of polyploidy in plants.
DUPLICATE genes are widespread in genomes of
almost all eukaryotes. Among flowering plants,
polyploidy (whole-genome duplication) is a primary
source of duplicate genes (Soltis and Soltis 1999;
Wendel 2000; Bowers et al. 2003; Lockton and Gaut
2005). All flowering plants either are contemporary
polyploids or harbor the evolutionary signature of
paleopolyploidy (ancient polyploidy) in their genomes.
Polyploidy may have influenced flowering plant di-
versification, as it provides raw material for the
evolution of novelty by relaxing purifying selection
on duplicate genes (Stephens 1951; Ohno 1970;
Lynch and Conery 2000; Wendel 2000). Through
genic redundancy, polyploids may be subject to an array
of evolutionary processes, including subfunctionaliza-
tion (evolution of partitioned ancestral functions
among duplicate genes) and neofunctionalization
(evolution of novel functions among duplicate genes).
Subfunctionalization and neofunctionalization have
been demonstrated in several species (Force et al. 1999;
Adams et al. 2003; Duarte et al. 2006; Cusack and
Wolfe 2007; Liu and Adams 2007; Teshima and Innan
2008). From an evolutionary perspective, both processes
can lead to the preservation of the two members of a
duplicate gene pair (Ohno 1970; Lynch and Force
2000). Because duplicate genes tend to be lost rapidly
through mutational processes (Lynch and Conery
2000), subfunctionalization is thought to be most
important shortly after gene duplication. As the age of
the duplicate pair increases, neofunctionalization be-
comes increasingly likely (Stephens 1951; Ohno 1970).
Further linking these two processes, it has been sug-
gested that subfunctionalization could serve as a preser-
vational transition state leading to neofunctionalization
(Rastogi and Liberles 2005). Thus following poly-
ploidy, both subfunctionalization and neofunctionaliza-
tion may make significant contributions to duplicate
gene retention and functional diversification.
In addition to subfunctionalization and neofunction-
alization, allopolyploid plants also generate diversity
through rapid genomic changes at various levels, in-
cluding chromosomal lesions and intergenomic ex-
changes, as in wheat (Shaked et al. 2001), Brassica
Supporting information is available online at http://www.genetics.org/
cgi/content/full/genetics.109.102608/DC1.
1
These authors contributed equally to this work.
2
Corresponding author: Department of Ecology, Evolution, and Organ-
ismal Biology, 251 Bessey Hall, Iowa State University, Ames, IA 50011.
E-mail: jfw@iastate.edu
Genetics 182: 503–517 ( June 2009)
(Song et al. 1995; Pires et al. 2004; Udall et al. 2004),
and Arabidopsis (Madlung et al. 2002), epigenetic
modifications (Lee and Chen 2001; Madlung et al.
2002; Wang et al. 2004; Salmon et al. 2005; Gaeta et al.
2007), and gene expression changes (Adams et al. 2003;
Wang et al. 2004; Bottley et al. 2006; Adams 2007;
Flagel et al. 2008). It is thought that these changes
result from ‘‘genomic shock’’ caused by the joint effects
of genome merger and genome doubling during
allopolyploid formation (Adams et al. 2004; Hegarty
et al. 2006; Flagel et al. 2008). Additionally, allopoly-
ploidy entails combining homeologous regulatory var-
iation and may lead to expression variation through
interacting cis- and trans-regulatory factors, as has been
shown for allelic variation (Wittkopp et al. 2004;
Stupar and Springer 2006; Swanson-Wagner et al.
2006). Collectively, these results demonstrate that both
genomic and genic evolutionary processes play a role in
allopolyploid evolution.
The cotton genus (Gossypium) is a useful system to
study the extent of genomic changes that accompany
genome merger and allopolyploidization (Wendel and
Cronn 2003). Allotetraploid cottons were formed by
the merger of two diploid species originating, respec-
tively, from the cotton A- and D-genome groups. This
event took place 1–2 million years ago (Percy and
Wendel 1990; Wendel and Albert 1992; Seelanan
et al. 1997; Cronn et al. 2002; Senchina et al. 2003)
(Figure 1A). The modern diploid species Gossypium
arboreum (A genome) and G. raimondii (D genome) are
extant diploids most similar to the ancestral A- and D-
genome diploids involved in the formation of natural
allotetraploids (Percy and Wendel 1990; Wendel and
Albert 1992; Seelanan et al. 1997; Cronn et al. 2002;
Senchina et al. 2003) (Figure 1A). Following formation,
the allotetraploid lineage diverged into five extant
species. Furthermore, F
1
hybrids and allotetraploids
synthetically derived from A- and D-genome species
mergers are also available (Figure 1A and Table 1).
These synthetic accessions have proved particularly
useful in teasing apart the effects of genome merger
and genome doubling during the formation of the
natural allopolyploid (Adams et al. 2004; Adams and
Wendel 2005a; Flagel et al. 2008). Although these
studies and others (Comai et al. 2000; Adams et al. 2004;
Soltis et al. 2004; Hegarty et al. 2006; Tate et al. 2006;
Chen 2007) have provided insights into the formation
and immediate genetic consequences of polyploidy,
there is still much to be learned about stabilization
and evolution of polyploid genomes following
formation.
In this study we employ a genome-specific, mass-
spectrometry technology to study relative levels of allelic
and homeologous (gene pairs duplicated by polyploidy)
gene expression in diploid and allopolyploid cotton. By
contrasting allelic and homeologous gene expression in
cotton species within an appropriate phylogenetic
framework (Figure 1A), we have detected expression
patterns consistent with subfunctionalization and neo-
functionalization (Figure 1B). Because the cotton
accessions selected for this study represent three suc-
cessive stages in allopolyploid evolution, i.e., genomic
merger of divergent parents, genome doubling, and
finally genomic coevolution in a common nucleus, we
were able to determine the extent of expression
evolution accompanying each stage.
MATERIALS AND METHODS
Maintenance of cotton germplasm and tissue collection:
Seedling tissues: Seeds of two diploid cottons, G. arboreum (A
2
)
and G. raimondii (D
5
), and a natural (G. hirsutum L. cv. Maxxa)
and synthetic [2(A
2
3D
3
)] allotetraploid cotton (Table 1),
were sown and grown in steamed potting mix in the Pohl
Conservatory at Iowa State University at 24°day/20°night
with a photoperiod of 16 hr light/8 hr dark. The synthetic
allotetraploid cotton was formed by colchicine doubling the
hybrid resulting from a cross between A
2
and the D-genome
species G. davidsonii (D
3
). Three biological replicates were
planted for each species and seedling-stage tissues were
sampled at 10 days postemergence. Additionally, a sterile F
1
hybrid (A
2
3D
5
) population has been maintain through
vegetative propagation, and was also sampled for some tissues.
The above accessions include representatives of both diploid
progenitor genomes (A and D genomes), their synthetic F
1
hybrid and synthetic allotetraploid, and a natural allopoly-
TABLE 1
Details of plant materials used
Taxon Genome designation Accession Ploidy level Location of origin
G. arboreum A
2
AKA-8401 Diploid Africa
G. raimondii D
5
Jfw Diploid Peru
G. arboreum 3G. raimondii (A
2
3D
5
) NA Diploid Laboratory
G. hirsutum AD
1
cv. Maxxa Tetraploid Mexico/Central America
G. arboreum 3G. davidsonii 2(A
2
3D
3
) NA Tetraploid Laboratory
The natural allotetraploid (G. hirsutum) was derived from hybridization, 1–2 MYA, between diploid A- and D-genome species
most similar to the modern species G. arboreum and G. raimondii. The cytoplasmic donor of G. hirsutum is its A-genome parent and
thus the F
1
cross was created in the same direction, with A
2
as the maternal parent. The synthetic allotetraploid was created by
crossing A
2
and D
3
diploid parents followed by genome doubling through colchicine treatment.
504 B. Chaudhary et al.
ploid cotton (Wendel and Cronn 2003) (Figure 1A). All
seedlings were sampled between 9 am to 10 am to minimize
circadian effects, and tissues were flash frozen in liquid
nitrogen and stored at 80°prior to RNA isolation.
Vegetative and floral tissues: Seedlings were grown for 3–5
weeks before transfer to larger pots and maintained at 32°and
a photoperiod of 16 hr light/8 hr dark. After the emergence of
the fifth leaf, the first, third, and fifth leaves were harvested
from all five taxa on the same day and flash frozen immediately
and stored. Petioles were sampled from the fifth leaf of each
biological replicate, and midrib and lamina tissues were
harvested from young and newly emerged leaves at the same
time. After 3–4 months, flowers from all species, except D
5
,
were harvested on 0 days postanthesis (dpa) (0 dpa is the day
the flower opened). Juvenile plants from D
5
were grown
separately under a shade regime for 1 month, a treatment
necessary to induce flowering. Fully opened flowers were
collected between 9 am and 11 am to mitigate circadian effects.
All flower tissues were manually excised and immediately flash
frozen in liquid nitrogen.
Fiber: Plants from all five taxa were grown in three replicates
in the horticulture greenhouse at Iowa State University and
flowers were harvested for four different stages of fiber
development (5, 10, 20, and 25 dpa). For each replicate and
developmental time point, ovules were excised and immedi-
ately frozen in liquid nitrogen. Ovules were visually inspected
for cell damage and fibers were inspected for contaminating
tissue.
Isolation of total RNA and sample platform preparation:
RNA isolation: All 24 tissues (Table 2) from the five taxa and
three biological replicates were collected in 1.7-ml microfuge
tubes. RNAs were extracted from all seedling, vegetative, and
floral tissues using a modified QIAGEN RNA extraction
protocol according to the manufacturer’s instructions
(QIAGEN, Valencia, CA) with modifications as follows: Tissues
were ground in fresh XT buffer (Wan and Wilkins 1994) in
Figure 1.—Phylogenetic
framework and detection
of subfunctionalization and
neofunctionalization among
homeologs. (A) Phyloge-
netic history of diploid and
allopolyploid cotton (Gos-
sypium). Allopolyploidy oc-
curred 1–2 MYA by
hybridization between A-
and D-genome diploid spe-
cies, most similar to the
modern species G. arboreum
and G. raimondii. The mod-
ern F
1
hybrid and synthetic
allopolyploid, both derived
from A- and D-genome dip-
loidspecies,mimicthe stages
of genome merger and ge-
nome duplication during al-
lopolyploid formation. (B)
After genome merger regu-
latory changes may cause al-
lelic/homeologous gene
expression patterns to di-
verge. These patterns can
result in subfunctionali-
zation, the partitioning of
ancestral expression, or
neofunctionalization, oper-
ationally defined here as
the development of novel
expression patterns relative
to that of the ancestor. The
latter was detected by com-
paring ancestral expression
(1:1 mix) to the expression
found in the F
1
, synthetic,
and Maxxa.
Homeolog Expression in Cotton 505
microfuge tubes with plastic pestles and incubated at 42°for
1.5 hr. Then 2 mKCl was added and the sample was incubated
on ice for 1 hr. After incubation, the samples were transferred
to Qiashredder columns supplied with the QIAGEN Plant
RNeasy kit and all subsequent steps followed this kit’s protocol.
RNAs were extracted from fibers at each developmental
time point using a liquid nitrogen/glass bead shearing
approach following a lithium chloride hot borate protocol
(Hovav et al. 2007; Taliercio and Boykin 2007). Purified
RNA samples were quantified using a NanoDrop spectropho-
tometer (NanoDrop Technologies, Wilmington, DE) and
assayed for degradation using a BioAnalyzer (Agilent, Palo
Alto, CA).
cDNA preparation: A total of 307 tissue samples were used for
RNA isolations, each yielding 5mg of total RNA. All RNA
samples were treated with DNase following the manufacturer’s
protocol [New England Biolabs (DNase I, M0303S)], and
assayed for genomic DNA contamination by PCR amplifica-
tion with primers flanking intron eight of a Gossypium RNA
helicase with high similarity to the gene At4G00660 in
Arabidopsis (GenBank accession NM_179204). Following
DNase treatment, cDNAs were synthesized using Superscript
III reverse transcriptase, according to the manufacturer’s
instructions (Invitrogen, Carlsbad, CA). In tissues with surplus
RNA yields, cDNAs were also synthesized from equal RNA
mixes of A
2
and D
5
accessions. These mixes served as an in vitro
model for mid-parent expression within the allopolyploid and
hybrid accessions.
Probe selection for multiplex PCR: MALDI-TOF mass-spectrom-
etry assays for genome-specific expression were designed for
the Sequenom (San Diego) MassARRAY platform. Genes for
this platform were selected from 1231 cotton EST contigs
(Udall et al. 2006), derived from A
2
,D
5
, and AD
1
accessions.
These contigs were inferred to represent homeologous
relationships in AD
1
on the basis of comparisons to ortholo-
gous sequences from the A- and D-genome diploids This led to
the identification of genome-specific SNPs, which were pro-
cessed using the Sequenom probe selection software. From
these results we selected four multiplexes, each including 29
genes.
Genome-specific expression assays: For each multiplex, forward
and reverse primers from all 29 genes were pooled and used to
amplify each cDNA sample using the manufacturer’s specifi-
cations (Sequenom). Amplified cDNAs were visualized on
agarose gels to confirm amplification and loaded on a 384-well
plate in three technical replicates. Mass-spectrometry quanti-
fication of genome-specific expression ratios was performed at
the University of Minnesota genotyping facility.
Data processing, filtering, and analysis: Identification of
diagnostic assays: All expression data recovered from the
MassARRAY process were first filtered on the basis of internal
measures of assay quality, which included removing all assays
flagged as ‘‘Bad Spectra,’’ or having a frequency of uncertainty
.0.2 or an unused extension primer frequency .0.5. Next all
genes were filtered on the basis of assays of A
2
and D
5
DNA
samples, which were mixed in known ratios (4:1, 2:1, 1:1, 1:2,
and 1:4) and used to standardize the genome-specific quan-
tification procedure for each gene (Stupar and Springer
2006). All genes were required to display a strong correlation
(R
2
.0.9) between the expected and observed A
2
:D
5
DNA
ratios. Additionally, DNAs from Maxxa, the synthetic poly-
ploid, and the F
1
hybrid were also assayed as controls for
lineage-specific SNPs, which could potentially arise in these
accessions, with the expectation that good assays would yield
1:1 A- to D-genome values. Maxxa, synthetic, and F
1
hybrid
assays were excluded if these DNA control values exceeded the
expected 1:1 ratio by 625%. Following filtering, a maximum
of nine replicates (three biological 3three technical) could
potentially be recovered for each assay. Each of these replicate
pools represents one gene by tissue and evaluates the pro-
portion of A- and D-genome contribution to the transcrip-
tome. These values were stored as the percentage of D-genome
contribution to the transcriptome (% D) and outlier replicates
were identified and removed if they deviated from the median
of the replicate pool by 650% D. Next, the mean % D and
standard deviation of the remaining values were recorded and
used for all subsequent analyses; this complete data set can be
found as supporting information, File S2.
Statistical contrasts of genome-specific expression ratios: Contrasts
of A- and D-genome expression ratios were made using a t-test.
P-values were then converted to q-values using the method of
Storey and Tibshirani (2003), and individual contrasts were
considered equivalent when q.0.05.
RESULTS
Assessment of Sequenom MassARRAY performance:
Using a single nucleotide polymorphism (SNP)-based
Sequenom MassARRAY technology, we simultaneously
assayed the A- and D-genome contribution to the
transcriptome for 63 gene pairs (Table S1). These A-
TABLE 2
Number of successful genome-specific assays calculated for
all genes in all tissues
No. of genes successfully assayed
Tissues Mix F
1
hybrid Synthetic Maxxa
Seedling stage
Primary root 41 NA 33 33
Hypocotyl NA NA 7 7
Cotyledon 45 NA 38 38
Vegetative stage
First leaf NA NA 31 36
Third leaf NA NA 26 26
Fifth leaf NA NA 22 36
Petiole 50 50 42 42
Apical shoot NA 48 40 40
Leaf midrib NA 42 37 38
Leaf lamina 40 34 28 26
Floral stage
Pedicel NA 32 27 25
Bract NA 30 24 21
Calyx NA 32 22 22
Petal NA 35 26 24
Anther NA 40 24 29
Stamina tube NA 39 27 28
Pollen NA NA NA 11
Style and stigma NA 41 30 27
Ovary wall (0 dpa) 30 32 25 21
Ovule (0 dpa) 37 38 26 26
Fiber
Fiber (5 dpa) 25 NA 21 19
Fiber (10 dpa) 32 NA 26 23
Fiber (20 dpa) 37 NA 30 30
Fiber (25 dpa) 41 NA 34 32
Total 378 493 646 660
NA, tissue not available.
506 B. Chaudhary et al.
and D-genome gene pairs are hereafter termed ‘‘home-
ologous’’ in the polyploid genotypes and ‘‘allelic’’ in the
diploid F
1
hybrid genotypes (note however, in the F
1
hybrid chromosomes from the A- and D-genome chro-
mosome pairing is limited; Endrizzi et al. 1985). The
MassARRAY technology has previously been shown to
be effective in determining the relative allelic transcript
levels in hybrid maize (Stupar and Springer 2006).
Assays of cotton A- and D-genome expression were made
possible by the availability of A- and D-genome-specific
SNPs obtained from cotton EST contig assemblies
(Udall et al. 2006), which included transcripts from
the diploid members of the A and D genome (A
2
and
D
5
) and an allotetraploid (AD
1
). The presence of a
genome-specific SNP alters the molecular weight such
that the MassARRAY platform can distinguish either
variant from a mixed transcript pool and estimate
relative abundance.
Expression assays were filtered using a rigorous
quality-control protocol (see materials and meth-
ods), yielding the total number of successful assays
summarized in Table 2. The percentage of successful
assays varied among tissues from a maximum of 73%
in petioles to a minimum of 5% in pollen and hypocotyl
tissues (Table 2 and Figure S1). Among 63 genes and 24
tissue types examined, 660 and 646 gene-by-tissue
combinations were successful in the natural (‘‘Maxxa’’
hereafter) and synthetic (‘‘synthetic’’ hereafter) allopo-
lyploids. Due to limited tissue and sample availability in
the F
1
and 1:1 A- and D-genome mix, we examined 13
and 10 tissue types in these accessions resulting in 493
and 378 successful assays, respectively (Table 2).
Patterns of genome-specific gene expression in
cotton tissues: The primary goal of this study was to
quantify genome-specific expression among a sampling
of cotton tissues and developmental conditions in an
evolutionary context. This was accomplished by assaying
24 tissues or developmental stages, which fit into four
general categories: seedling, vegetative, and floral
tissues, as well as developing fibers (Table 2). For each
of these categories, genome-specific expression values
were extracted for the mix, F
1
, and the synthetic and
natural (Maxxa) allopolyploids and binned into five
groups, using the percentage of D-genome expression
as a metric (0–20% D, 20–40% D, 40–60% D, 60–80% D,
and 80–100% D) (Figure 2). In the F
1
, synthetic, and
Maxxa, biases indicate differential gene expression
between the A- and D-genome transcripts within the
same nucleus, whereas in the mix, which pools two
biologically different species, a bias reflects differential
gene expression between the A
2
and D
5
parents.
Overall, F
1
, synthetic, and Maxxa show an A-genome
bias in seedling and vegetative tissues, but in floral
tissues the mix shows a D bias whereas the F
1
and Maxxa
show A-genome biases and the synthetic is nearly
equivalent (Figure 2). For ‘‘floral’’ samples the mix is
represented by only ovary wall and ovule tissues, though
both individually support a D bias. Fiber expression in
the mix and Maxxa show a substantial level of A-genome
bias, whereas the synthetic is less A-genome biased
(Figure 2; note that the F
1
hybrid between A
2
and D
5
is sterile and hence fibers could not be studied). These
expression patterns are interesting, as they highlight
previous observations (Adams et al. 2003, 2004; Adams
Figure 2.—Distribution
of genome-specific expres-
sion states among acces-
sions and within different
tissue categories. Each
panel represents a tissue
category and shows histo-
grams for the mix F
1
, syn-
thetic, and Maxxa. The
expression categories cor-
respond to the following
values: strongly A biased
(0–20% D expression); A
biased (20–40% D expres-
sion); equivalent (40–60%
D expression); D biased
(60–80% D expression);
strongly D biased (80–
100% D expression). The
y-axes indicate the number
of gene-by-tissue combina-
tions that fell under each
category. NA, tissue type
not available.
Homeolog Expression in Cotton 507
and Wendel 2005a; Udall et al. 2006; Yang et al. 2006;
Flagel et al. 2008; Hovav et al. 2008b) that neither the A
nor D genome is globally dominant with regard to
genome-specific expression.
These general trends describe the overall patterns of
expression states for this sampling of genes and tissues.
At the individual gene level there is considerable
variation. An interesting example is the gene
CO131164 (a putative phytochrome-associated pro-
tein), which shows highly variable expression among
tissues and accessions. In Maxxa this gene demonstrates
nearly complete A-genome expression in anthers and
complete D-genome expression in ovary wall (Figure
3A), indicative of developmentally regulated reciprocal
silencing of alternative homeologs in different parts of
the same flower (cf. Adams et al. 2003). Additionally,
shortly after fiber initiation (5 dpa), CO131164 is
strongly A-genome biased in the synthetic, though
Maxxa shows approximately equivalent expression (Fig-
ure 3A). Another illustrative gene is CO130747 (a
putative CBL-interacting protein kinase), which shows
significant differences in tissue-specific homeolog ex-
pression between Maxxa and the synthetic during many
developmental stages (Figure 3B). The synthetic is more
A-genome biased in seedling, vegetative, and floral
stages, including almost total A-genome expression in
roots, petioles, the calyx, and all four developmental
stages of fiber. In contrast, Maxxa is only strongly A-
genome biased in 5 and 10 dpa fibers. A third example
gene illustrated (Figure 3C) is DW008528 (similar to a
putative protein with unknown function in Arabidopsis
thaliana), for which we observed equivalent A- and D-
genome homeolog expression in all tissues for the
synthetic and the F
1
, but considerable expression
variation for vegetative tissues in Maxxa. In all fiber
stages studied, both Maxxa and the synthetic show
nearly equal expression of homeologs.
Genome-specific expression biases during genome
merger and doubling: The accessions studied were
selected to provide insight into the various stages
involved in allopolyploid speciation, including diploid
divergence, genome merger, genome doubling, and
subsequent evolution and stabilization. To assess home-
olog transcriptional alteration accompanying each of
these stages, we identified all gene 3tissue combina-
tions shared by all four accessions (mix, F
1
, synthetic,
and Maxxa), as well as those shared just by the F
1
,
synthetic, and Maxxa, and finally just by the synthetic
and Maxxa. For each of these groups, we assigned all
gene 3tissue relationships as either equivalent (‘‘¼’’ ;
q-value .0.05) or nonequivalent (‘‘6¼’’ ; q-value #0.05).
Specific examples of several expression patterns
Figure 3.—Tissue-specific and
genome-specific gene expression
among three gene pairs in A
2
,
D
5
,F
1
hybrid, synthetic, and Max-
xa. Tissues are arrayed along the
x-axis while the proportion of D-
genome expression is on the y-
axis. The lines linking tissues do
not imply a strict order of plant
development; instead they serve
as a viewing aid. (A) Gene
CO131164 (a putative phyto-
chrome-associated protein). (B)
Gene CO130747 (a putative
CBL-interacting protein kinase).
(C) Gene DW008528 (a protein
of unknown function).
508 B. Chaudhary et al.
and their biological interpretation can be found in
Figure 4.
As summarized in Table 3, when comparing all four
accessions, the category that induced the most expres-
sion alteration was genome merger (implicated in 23 1
16 1919¼57 gene 3tissue events) followed by
change due to polyploid evolution (implicated in 9 116 1
619¼40 gene 3tissue events). From these results, it is
clear that genome merger and polyploidy evolution
(subsequent to formation) have the greatest effect on
homeologous gene expression, though diploid diver-
gence and genome doubling are implicated in 11 and
30 gene 3tissue events, respectively. For genes lacking
data from the mix sample, more homeolog expression
changes occurred due to polyploid evolution than
polyploidy alone, corroborating the foregoing result.
Alternatively, some of the above observations could be
due to the divergence between the model diploid
progenitors used in this study and the actual ancient
parents of natural allopolyploid cotton. Similar findings
have been reported in cotton and other polyploid
systems regarding the relative importance of genome
merger (Adams and Wendel 2005a; Wang et al. 2005;
Hegarty et al. 2006; Flagel et al. 2008) and genome
doubling (Stupar et al. 2007). However, to our knowl-
edge, this is the first study wherein the specific effects of
each of these four components (divergence, merger,
polyploidy, and polyploidy evolution) have been
disentangled.
Tissue-specific subfunctionalization and gene silenc-
ing: To address the prevalence of subfunctionalization
between homeologous genomes, we searched for pat-
terns of highly differential homeolog expression biases
between tissues from the F
1
, synthetic, and Maxxa (see
Figure 1B). We did not detect any cases of complete
reciprocal homeolog silencing (here silencing is oper-
ationally defined as the absence of detectable tran-
script) among the 63 genes assayed, but the most
subfunctionalized genes and their respective tissues
are listed in Table 4. In Maxxa, the most striking
example is the gene CO131164, where the A-genome
homeolog has been silenced in the ovary wall, but the
reverse is observed in anthers, where the A-genome
homeolog accounts for 93% of homeologous expres-
sion. Other genes showed similar patterns of subfunc-
tionalization in various tissues (Table 4). Interestingly,
among those genes displaying the largest degree of
expression subfunctionalization, it appears that repro-
ductive tissues such as anthers, style/stigma, staminal
tube, and ovary wall are often involved (these tissues
comprise 12 of 18 tissues in Table 4). This observation
mirrors similar findings from Adams et al. (2003).
In addition to subfunctionalization, hybrid and poly-
ploid plants also display genome-specific silencing
biases. For each genotype, the percentage of completely
silenced genes varied from a maximum of 6% D-
homeolog silencing in Maxxa to a minimum 0.3% A-
homeolog silencing in the synthetic (Table 5). In most
cases, complete silencing remains in each subsequent
stage along the pathway to allopolyploidy. For example,
the gene CAO23634 (a putative S-formylglutathione
hydrolase), is D silenced in petioles and apical shoot
Figure 4.—Examples of
tissue-specific expression
alteration arising from pa-
rental divergence, genomic
merger, polyploidy, and
polyploidy evolution.
Shown are the proportions
of D-genome (y-axis) ho-
meolog expression, includ-
ing the associated standard
deviation. (A) Four repre-
sentative genes from pe-
tioles, which exhibit
statistically equivalent ra-
tios in all accessions, indi-
cating little expression
evolution since divergence
between the A- and D-ge-
nome parents. (B) Four
representative genes from
leaf lamina, each showing
equivalent expression
among the F
1
hybrid, syn-
thetic, and Maxxa, which is not equivalent to the mix, indicating an expression change resulting from genomic merger. (C) Four
representative genes from petioles (genes CAO71171, CO131379, and CAO49511) and leaf lamina (gene CO131379) showing
equal genome-specific expression values in the mix and F
1
hybrid, which differ from the synthetic and Maxxa, suggesting that
the change occurred as a result of genome doubling. (D) Four genes in petioles showing no change among mix, F
1
, and synthetic,
but a new expression pattern in Maxxa, indicating a change in homeolog-specific expression during the evolution (1–2 million
years) of allopolyploid cotton.
Homeolog Expression in Cotton 509
meristems in the F
1
, and this silencing remains in the
synthetic and Maxxa polyploids (Figure 5). However, there
are also counterexamples, such as the gene CO130747.
This gene is D silenced in petioles of the synthetic but
the D genome is once again expressed in Maxxa (Figure
5). Another example is gene CO131164, where there is
silencing of the A
2
diploid in ovary walls, but this gene is
expressed in the F
1
and synthetic, and then once again
the A genome is silenced in Maxxa (Figure 5).
Tissue-specific transcriptional neofunctionalization:
Neofunctionalization may be detected in our frame-
work by first indentifying all gene 3tissue assays that
lack expression of either the A- or D-genome ortholog in
the mix (i.e., not expressed in the A
2
or D
5
parent) and
which gain expression in the F
1
, synthetic, or Maxxa
(Figure 1B). It is important to note that the pattern
above can arise de novo, as a totally novel form of
expression, or as a product of the reactivation of a lost
ancestral expression regime, and our experiment can-
not distinguish between these two forms of transcrip-
tional neofunctionalization.
Using the criteria above, a total of 15 genes across
eight different tissues exhibit transcriptional neofunc-
tionalization. Additionally, by observing the range of
expression values for the 1:1 parental mixtures in all
available genes (Figure S2), it appears unlikely that
these cases of neofunctionalization are a product of an
inaccurate mix. Among the neofunctionalized genes, 10
showed substantial contributions from both genomes in
the F
1
, synthetic, and Maxxa, reinforcing the presence
of gene expression neofunctionalization (Table 6).
Genes CO108066 (a putative glyceraldehyde-3-
phosphate dehydrogenase) and CO076921 (a putative
vacuolar ATP synthase catalytic subunit) show lack of
expression of either the A or D orthologs, respectively,
in leaf lamina, but both homeologs are expressed in the
same tissue in the F
1
, synthetic, or Maxxa (Table 6). In
addition, in those cases where neofunctionalization has
occurred, it has been maintained in all genomically
merged samples (F
1
, synthetic, and Maxxa; Figure S3).
Overall, the nonfunctional alleles were usually from the
diploid A genome (11 of 14 cases), indicative of the
TABLE 3
Distribution of expression states among the mix, F
1
, synthetic, and Maxxa and their biological interpretation
Genotype comparison
Gene 3tissue combinations
showing pattern (% of total) Biological description
Mix ¼F
1
¼synthetic ¼Maxxa
with equal A–D expression
6 (6.4) No change
Mix ¼F
1
¼synthetic ¼Maxxa
with unequal A–D expression
11 (11.7) Change due to A–D divergence
Mix 6¼ F
1
¼synthetic ¼Maxxa 23 (24.4) Change due to genome merger
Mix ¼F
1
6¼ synthetic ¼Maxxa 5 (5.3) Change due to polyploidy alone
Mix ¼F
1
¼synthetic 6¼ Maxxa 9 (9.6) Change due to polyploid
evolution
Mix 6¼ F
1
6¼ synthetic 6¼ Maxxa 16 (17) Change due to all sources
Mix ¼F
1
6¼ synthetic 6¼ Maxxa 6 (6.4) Change due to polyploidy and
polyploid evolution
Mix 6¼ F
1
6¼ synthetic ¼Maxxa 9 (9.6) Change due to genome merger
and polyploidy
Mix 6¼ F
1
¼synthetic 6¼ Maxxa 9 (9.6) Change due to genome merger
and polyploid evolution
Total 94
F
1
¼synthetic ¼Maxxa 104 (34.9) No change
F
1
6¼ synthetic ¼Maxxa 57 (19.1) Change due to polyploidy alone
F
1
¼synthetic 6¼ Maxxa 67 (22.5) Change due to polyploid
evolution
F
1
6¼ synthetic 6¼ Maxxa 70 (23.4) Change due to all sources
Total 298
Synthetic ¼Maxxa 275 (50.6) No change
Synthetic 6¼ Maxxa 269 (49.4) Change due to polyploid
evolution
Total 544
The first nine rows compare the distribution of expression categories for all available gene 3tissue combi-
nations among all four taxa. The next four rows compare just F
1
, synthetic, and Maxxa, and the last two only
synthetic and Maxxa.
510 B. Chaudhary et al.
potential for a genome-of-origin bias for neofunction-
alization in cotton, albeit for a relatively small sampling
of genes.
Evolution of cis- and trans-regulatory variations in
cotton: Expression variation can originate via either
cis-ortrans-regulatory evolution, or both. By compar-
ing genome-specific expression between the mix and
F
1
it is possible to partition expression variation into
cis-andtrans-origins, using the procedures described
by Wittkopp et al. (2004) and Stupar and Springer
(2006). Our analysis of cis-andtrans-acting regulation
in cotton includes 30 genes in leaf lamina and 38 genes
in the petiole (Figure 6). Among both leaf lamina and
petiole tissues the most prevalent type of regulatory
divergence is cis-regulatory evolution (50% and 39% in
lamina and petiole, respectively) followed by a combi-
nation of cis-andtrans-factors.Thisresultissimilarto
other studies regarding the prevalence of these modes
of regulatory evolution (Wittkopp et al. 2004; Stupar
and Springer 2006; Springer and Stupar 2007a;
Zhuang and Adams 2007). Additionally this result
gives an indication that some of the expression
changes attributed to genome merger (Table 3) are
likely caused by cis-regulatory divergence between the
AandDgenomes.
DISCUSSION
Homeologous contributions to the transcriptome:
We used a mass-spectrometry-based SNP detection
technique to measure allele- and homeolog-specific
contributions to the transcriptome of diploid and
allopolyploid cotton accessions that were selected to
be informative with respect to the evolutionary stages
involved in allopolyploid speciation and subsequent
evolution (Figure 1A). Although the representative
progenitor diploid species used in this study (A
2
,D
3
,
and D
5
) are not the actual parents of natural allopoly-
ploid cotton, which formed 1–2 million years ago, a
substantial body of evidence indicates that they repre-
sent close approximations (reviewed in Wendel and
Cronn 2003). Furthermore, to evaluate differences
between D
3
and D
5
[and as a corollary species-specific
biases associated with the 2(A
2
3D
3
) synthetic allote-
traploid] we compared expression between these spe-
cies from 18 randomly selected genes in petiole tissues
and 17 genes in leaf tissues. These comparisons were
made relative to a common A
2
reference sample and
were conducted using the Sequenom platform follow-
ing the procedures outlined in materials and meth-
ods. These experiments show that D
3
and D
5
are similar
in their expression, having an average expression
TABLE 4
Proportional transcript contribution of A and D homeologs in different tissues
Genotype
GenBank
accession
Strongly
A-biased tissue
A-biased tissue
expression (A, D)
Strongly D-biased
tissue
D-biased tissue
expression (A, D)
Maxxa CO131164 Anther 93, 7 Ovary wall 0, 100
CO080701 Pollen 94, 6 10 dpa fiber 14, 86
CO077994 10 dpa fiber 77, 23 Ovary wall 0, 100
DW008528 Ovary wall 100, 0 Anther 25, 75
CO082621 Anther 71, 29 Ovary wall 0, 100
Synthetic CO124958 First leaf 100, 0 Style/stigma 25, 75
CO098920 Cotyledon 91, 9 Anther 22, 78
F
1
hybrid CO121715 Staminal tube 100, 0 Leaf lamina 25, 75
CAO23634 Leaf lamina 100, 0 Staminal tube 39, 61
Each gene (listed by GenBank accession) demonstrates nearly complete expression subfunctionalization among the two tissues
shown.
TABLE 5
Distribution of tissue-specific homeologous gene silencing events
Genotype
No. of gene 3
tissue assayed
Gene 3tissue
combinations with D
silencing
Gene 3tissue
combinations with
A silencing
%D
silenced
%A
silenced
F
1
493 7 (4 genes) 6 (3 genes) 1.41 1.21
Synthetic 646 16 (5 genes) 2 (2 genes) 2.48 0.31
Maxxa 660 42 (11 genes) 9 (5 genes) 6.36 1.36
Homeolog Expression in Cotton 511
difference of 15.5% among the 35 comparisons (Figure
S4). For comparison, the average variation between
biological replicates within D
3
and D
5
was 12.5%,
meaning that within-species variation was 81% of the
level of the difference between D
3
and D
5
. These results
indicate that species-specific differences between D
3
and D
5
are small.
Contrasting genome-specific expression in these
accessions allowed us to allocate expression alterations
to the stages of genome merger, genome doubling, and
subsequent evolution within the allopolyploid lineage,
while revealing examples of subfunctionalization and
neofunctionalization (Figure 1B).
To substantiate the MassARRAY-based interpreta-
tions, we validated these estimates of genome-specific
expression through comparisons to expression data
generated by a genome-specific microarray platform
(Udall et al. 2006). These validations were conducted
for both petals (Flagel et al. 2008) and fibers from
several developmental stages (Hovav et al. 2008a), and
demonstrate significantly positive correlations.
Allopolyploidy entails the merger of two diploid
genomes, which may contribute either equally or dis-
proportionately to the transcriptome. Data presented
here demonstrate that genomically biased expression in
cotton is a common phenomenon, occurring in vegeta-
tive and floral tissues, and also in single-celled fibers,
consistent with previous studies using other genes and
analytical methods (Adams et al. 2003, 2004; Adams and
Wendel 2005a; Udall et al. 2006; Yang et al. 2006;
Flagel et al. 2008; Hovav et al. 2008a). In this study,
among 49 homeologous genes sampled in Maxxa,
40% exhibit biased expression toward the A or D
homeolog, in all tissues examined (Figure 2). Further-
more, the extent of genome-specific bias varies sub-
stantially among tissues, from nearly equal expression to
Figure 5.—Examples of tissue-specific expres-
sion partitioning. The y-axis represents the pro-
portional transcript contribution from A and D
homeologs.
TABLE 6
Expression neofunctionalization
Proportion genomic expression
Gene Tissue Mix (A, D) F
1
(A, D) Synthetic (A, D) Maxxa (A, D)
CAO62858 Root 0, 100 NA 32, 68 41, 59
Ovule 0, 100 30, 70 25, 75 32, 68
CO111212 Leaf petiole 0, 100 – 27, 73 33, 67
5 dpa fiber 0, 100 NA 34, 66 49, 51
10 dpa fiber 0, 100 NA 29, 71 42, 58
CO108066 Leaf lamina 0, 100 49, 51 66, 34 63, 37
Ovule 0, 100 59, 41 68, 32 34, 66
CO076921 Leaf lamina 100, 0 61, 39 93, 07 57, 43
AAP41846 Ovule 0, 100 53, 47 50, 50 48, 52
CO081422 Ovule 0, 100 48, 52 40, 60 –
CAO71171 Ovule 0, 100 61, 59 40, 60 47, 53
AAK69758 Ovule 0, 100 49, 51 22, 78 38, 62
CO077994 20 dpa fiber 100, 0 NA 63, 37 58, 42
CO093729 25 dpa fiber 100, 0 NA 72, 28 86, 14
Each gene exhibited differential expression between the diploids (shown by the 1:1 parental mix), but ex-
pression from both the A and D genomes in the F
1
, synthetic, and natural allopolyploid (Maxxa). NA, tissue not
available. –, value could not be determined.
512 B. Chaudhary et al.
complete silencing (here again, silencing refers to an
absence of detectable transcript). The accumulated
results from this study and others noted above indicate
that among hybrid and allopolyploid cotton both the A
and D genome contribute unequally to the transcript
pool, but that neither genome displays an overall
expression preference. This result differs from natural
and synthetic allotetraploids in Arabidopsis, which show
a global downregulation of the A. thaliana genome in
favor of the A. arenosa genome (Wang et al. 2006; Chen
et al. 2008).
Although genomic preference was not detected at a
global scale, relative transcript abundance from indi-
vidual genes varied greatly. Genome-specific silencing
was observed in 4 genes in the F
1
hybrid, 5 genes in the
synthetic, and 11 genes in Maxxa, noting that this
differed widely among tissue types for many of those
genes (Table 5). These results indicate that silencing is
most prevalent in the natural allopolyploid, following 1–
2 MY of allopolyploid evolution. Furthermore, in
Maxxa, silencing is more prevalent among D-genome
homeologs than among A-genome homeologs (Table
5). Both of these findings regarding the enhancement
of silencing in Maxxa and a greater level of D-genome
silencing mirror the findings of Flagel et al. (2008),
though their study was limited to only petal tissues.
Though the phenotypic effects of homeolog silencing
in cotton are unknown, it is possible that tissue-specific
Figure 6.—Plots of A- and D-genome parental
mix (mix) vs.F
1
hybrid (F
1
) for leaf lamina (A)
and petiole (B). In principle, genome-specific ex-
pression differences initiated by cis-regulatory di-
vergence are expected to share this difference
between both the mix and the F
1
and will accord-
ingly fall on a 1:1 diagonal when plotted against
one another (red points), whereas trans-
regulatory divergence will equilibrate genome-
specific expression when coresident in the F
1
nu-
cleus and instead fall on equivalently expressed
horizontal line for the F
1
only (blue points).
Genes that fall along neither of these lines are in-
ferred to be regulated by a combination of cis-
and trans-factors (Wittkopp et al. 2004; Stupar
and Springer 2006) (green points). Finally,
genes with divergence only in the F
1
(purple
points) or no expression divergence (gray
points) offer no insight into cis-ortrans-expres-
sion evolution.
Homeolog Expression in Cotton 513
homeolog silencing has had an impact on the evolution
of allotetraploid cotton. For the AdhA gene in G.
hirsutum,Liu and Adams (2007) have shown that
homeologous expression biases can occur as a response
to abiotic stress. These findings of altered homeologous
expression patterns in response to genomic stress may
hint at the adaptive potential of polyploidy. In this vein,
our findings shed additional light on the extensive
breadth and diversity of homeolog expression patterns
in natural allotetraploid cotton.
Distinguishing the effects of genome merger, ge-
nome doubling, and polyploid evolution on gene
expression: By partitioning genome-specific expression
changes within a selected framework of cotton acces-
sions (Figure 1A), we were able to determine that
genome merger has the largest impact on biased
expression of homeologs along the pathway to poly-
ploidy in cotton (Table 3). Allelic expression differ-
ences, detectable immediately in the F
1
hybrid, likely
arise as a result of the merger of the divergent regulatory
machinery of the A and D genomes within cotton. As
many expression biases are shared with ancient allo-
polyploid cotton, the early establishment of expression
patterns may play a role in gene expression evolution
during the formation and subsequent evolution of
natural cotton allopolyploids (Adams 2007; Chen
2007). Similar results have been previously noted in
cotton (Adams and Wendel 2005a; Flagel et al. 2008),
as well as Senecio (Hegarty et al. 2006) and Brassica
(Albertin et al. 2006). These authors all found that a
considerable portion of gene expression alteration took
place at the F
1
hybrid stage when compared to resynthe-
sized allopolyploids. In Senecio and Brassica the effect
of genome merger was, in fact, found to contribute a
majority of the observed expression changes. Hegarty
et al. (2006) classified this result as an example of
‘‘genomic shock,’’ a phenomenon which has often been
observed in plant hybrids, but remains poorly under-
stood at the molecular level. Some insight may derive
from estimating the relative roles of cis- and trans-
regulation within the F
1
(Figure 6), and in this respect
our data indicate that cis-evolutionary factors (those
arising from A- and D-genome cis-regulatory diver-
gence), appear to be most prevalent. Taken together,
these data indicate that reuniting divergent cis-regula-
tory domains may be a major component of genomic
shock as it pertains to cotton hybrids and allopolyploids.
Following genomic merger, we found that allopoly-
ploid evolution was the next most prevalent contributor
to expression evolution (Table 3). This result is in-
teresting as it implicates a significant role for the action
of long-term evolutionary processes, such as sub- and
neofunctionalization. Furthermore, changes that occur
via allopolyploid evolution are more prevalent than
those occurring via genome duplication alone (40 vs. 30
gene 3tissue events; Table 3). This result indicates that
genomic duplication alone may play a less significant
role in altering homeologous gene expression states in
cotton, possibly affecting only those homeologs with
dosage-regulated expression (Osborn et al. 2003).
Mechanisms of functional divergence and retention
of homeologs following allopolyploidy: Tissue-specific
and developmental expression variation between co-
resident genomes may occur via several mechanisms,
including altered regulatory interactions, epigenetic
modifications, and gene dosage changes (Comai et al.
2000; Birchler et al. 2003; Osborn et al. 2003; Riddle
and Birchler 2003; Adams and Wendel 2005b). At
present, we lack an explanation of the underlying
mechanisms of allelic and homeologous gene expression
biases, though our results indicate that both short-
(genome merger) and long-term (duplicate gene evo-
lution) evolutionary processes play a role in determining
homeolog expression states in allopolyploid cotton.
Recent work in allotetraploid Arabidopsis has shown
that genome-specific methylation may play a crucial role
in establishing homeolog expression patterns (Chen
et al. 2008). Using RNAi to silence met1, a cytosine
methyltransferase, Chen et al. (2008) demonstrated that
many previously identified cases of genome-specific
gene silencing were caused by or connected to methyl-
ation. Though these results may offer a promising
mechanistic explanation of our findings of genome-
specific biased expression and silencing, changes in
methylation do not appear to accompany allopolyploidy
in cotton (Liu et al. 2001). This difference between
Arabidopsis and cotton indicates that there may be no
single unifying factor that governs genome-specific
expression biases in allopolyploid plant species; instead
genome-specific expression evolution may occur via a
unique and ad hoc mixture of genetic and epigenetic
regulatory mechanisms within different species.
Following allopolyploid establishment, several mech-
anisms may affect the fate of homeologous genes
(Leitch and Bennett 1997; Matzke et al. 1999;
Wendel 2000; Levy and Feldman 2002; Liu and
Wendel 2002; Soltis et al. 2004; Comai 2005; Chen
and Ni2006). One model of homeologous gene re-
tention is subfunctionalization, which is the partition-
ing of ancestral function and/or expression domains
between duplicated genes, such that both copies con-
tinue to be necessary (Ohno 1970; Force et al. 1999;
Lynch and Force 2000). Various studies of subfunc-
tionalization, including MADS-box genes in Arabidopsis
(Duarte et al. 2006), germin genes in barley (Federico
et al. 2006), ZMM1 and ZAG1 genes in maize (Mena et al.
1996), and the AdhA gene in cotton (Adams et al. 2003),
have shown that expression subfunctionalization occurs
in plants. Here we show that instantaneous expression
subfunctionalization may occur immediately following
genomic merger (Table 4). Because of this, the preser-
vational forces of subfunctionalization may be immedi-
ately initiated for a significant number of genes within
allopolyploid cotton, as previously suggested (Adams
514 B. Chaudhary et al.
et al. 2003; Adams and Wendel 2005a; Flagel et al.
2008). Recent genomic analyses comparing homeolo-
gous regions in G. hirsutum lend support to this claim, as
homeologous gene loss appears to be rare (Grover et al.
2004, 2007).
During allopolyploid evolution, duplicate genes not
subject to subfunctionalization may still be retained if
one copy evolves a novel function via neofunctionaliza-
tion (Force et al. 1999; Lynch et al. 2001). Several
studies have identified neofunctionalization among
duplicate genes in diploid plants, including lectins in
legumes (Van Damme et al. 2007), MADS-box genes in
Physalis (Heand Saedler 2005) and Arabidopsis
(Duarte et al. 2006), LEAFY paralogs in Idahoa scapigera
(Brassicaceae) (Sliwinski et al. 2007), and diterpene
synthase paralogs in conifers (Keeling et al. 2008).
Expression neofunctionalization was also detected in
this study, which makes this the first example of neo-
functionalization in an allopolyploid, as far as we are
aware. We found 15 genes in eight different tissues where
expression was undetectable in one of the parental
diploids but appeared in the F
1
, synthetic, and Maxxa.
This pattern, which indicates an expansion of ancestral
expression domains, is consistent with expression
neofunctionalization.
In addition to the processes described above, cis-and
trans-regulatory changes provide insight into the evolu-
tion of regulatory networks in cotton. We observed that
most variation in gene expression following genome
merger is the result of cis-regulatory variation. This
finding suggests a mechanism for additive expression
patterns detected for many genes in a microarray study
of the F
1
hybrid (Flagel et al. 2008). Additionally, cis-
regulatory variation has been found to be a prevalent
mechanism for generating expression differences in F
1
maize hybrids (Stupar and Springer 2006; Swanson-
Wagner et al. 2006). While cis-regulatory evolution may
be more common, it is also possible that trans-regulatory
effects may affect gene expression, and even profoundly
so. For example, reactivation of a silenced gene copy in a
hybrid background, due to a trans-effect, may generate
novel expression cascades that have evolutionary con-
sequences. Mechanistic studies that determine the exact
nature of important cis-changes would be of tremen-
dous help in advancing our understanding of under-
pinnings of the observation of a prevalence of cis-
regulatory in the divergence in hybrid and allopolyploid
plants.
Evolutionary consequences of homeologous gene
expression in cotton: Recurrent polyploidization has
played a significant role in adding genetic variation to
the genomes of plant species. It has been demonstrated
that most duplicate genes are lost quickly on evolution-
ary times scales (Lynch and Conery 2000; Kellis et al.
2004; Thomas et al. 2006). Despite these rapid losses
some homeologous genes are retained, and various
explanations have been put forth to explain this re-
tention, including dosage sensitivity (Thomas et al.
2006) and gene function (Blanc and Wolfe 2004).
For example, among the retained homeologs in A.
thaliana, transcription factors and signal transduction
genes have been preferentially retained, whereas genes
performing enzymatic functions have not (Blanc and
Wolfe 2004). It has also been suggested that alteration
in duplicate gene expression patterns may enhance
retention (Adams et al. 2003; Flagel et al. 2008). In
cotton, this form of duplicate gene retention may be
facilitated by expression subfunctionalization and neo-
functionalization. These forms of divergence can occur
rapidly after polyploidization; indeed we show here that
many changes occur immediately in synthetic F
1
hybrids
and allopolyploids. From an evolutionary perspective,
this immediate form of expression divergence can
enhance expression variation and phenotypic diversifi-
cation in the short-term with the long-term consequence
of homeolog retention. Together these processes may
add to genetic and phenotypic variation with a species,
thus enhancing the future potential for natural selection
to lead to adaptive evolution.
The authors thank two anonymous reviewers for their helpful
comments. This project was supported by the National Research
Initiative of the United States Department of Agriculture Cooperative
State Research, Education and Extension Service, grant no. 2005-
35301-15700. B.C. received financial assistance from the Department
of Biotechnology, India. L.F. received financial assistance through a
graduate fellowship from the Plant Sciences Institute at Iowa State
University.
LITERATURE CITED
Adams, K. L., 2007 Evolution of duplicate gene expression in poly-
ploid and hybrid plants. J. Hered. 98: 136–141.
Adams, K. L., R. Cronn,R.Percifield and J. F. Wendel,
2003 Genes duplicated by polyploidy show unequal contribu-
tions to the transcriptome and organ-specific reciprocal silenc-
ing. Proc. Natl. Acad. Sci. USA 100: 4649–4654.
Adams, K. L., R. Percifield and J. F. Wendel, 2004 Organ-specific
silencing of duplicated genes in a newly synthesized cotton allo-
tetraploid. Genetics 168: 2217–2226.
Adams, K. L., and J. F. Wendel, 2005a Allele-specific, bidirectional
silencing of an alcohol dehydrogenase gene in different organs
of interspecific diploid cotton hybrids. Genetics 171: 2139–2142.
Adams, K. L., and J. F. Wendel, 2005b Polyploidy and genome evo-
lution in plants. Curr. Opin. Plant Biol. 8: 135–141.
Albertin, W., T. Balliau,P.Brabant, A.-M. Chevre,F.Eber et al.,
2006 Numerous and rapid nonstochastic modifications of gene
products in newly synthesized Brassica napus allotetraploids. Ge-
netics 173: 1101–1113.
Birchler, J. A., D. L. Auger and N. C. Riddle, 2003 In search of the
molecular basis of heterosis. Plant Cell 15: 2236–2239.
Blanc, G., and K. H. Wolfe, 2004 Functional divergence of dupli-
cated genes formed by polyploidy during Arabidopsis evolution.
Plant Cell 16: 1679–1691.
Bottley, A., G. M. Xia and R. M. D. Koebner, 2006 Homoeologous
gene silencing in hexaploid wheat. Plant J. 47: 897–906.
Bowers, J. E., B. A. Chapman,J.Rong and A. H. Paterson,
2003 Unravelling angiosperm genome evolution by phyloge-
netic analysis of chromosomal duplication events. Nature 422:
433–438.
Chen, M., M. Ha,E.Lackey,J.Wang and Z. J. Chen, 2008 RNAi of
met1 reduces DNA methylation and induces genome-specific
changes in gene expression and centromeric small RNA accumu-
lation in Arabidopsis allopolyploids. Genetics 178: 1845–1858.
Homeolog Expression in Cotton 515
Chen, Z. J., 2007 Genetic and epigenetic mechanisms for gene ex-
pression and phenotypic variation in plant polyploids. Ann. Rev.
Plant Biol. 58: 377–406.
Chen, Z. J., and Z. Ni, 2006 Mechanisms of genomic rearrange-
ments and gene expression changes in plant polyploids. Bio-
Essays 28: 240–252.
Comai, L., 2005 The advantages and disadvantages of being poly-
ploid. Nat. Rev. Genet. 6: 836–846.
Comai, L., A. P. Tyagi,K.Winter,R.Holmes-Davis,S.H.Reynolds
et al., 2000 Phenotypic instability and rapid gene silencing in
newly formed Arabidopsis allotetraploids. Plant Cell 12: 1551–
1568.
Cronn, R. C., R. L. Small,T.Haselkorn and J. F. Wendel,
2002 Rapid diversification of the cotton genus (Gossypium: Mal-
vaceae) revealed by analysis of sixteen nuclear and chloroplast
genes. Am. J. Bot. 89: 707–725.
Cusack, B. P., and K. H. Wolfe, 2007 When gene marriages don’t
work out: divorce by subfunctionalization. Trends Genet. 23:
270–272.
Duarte, J. M., L. Cui,P.K.Wall,Q.Zhang,X.Zhang et al.,
2006 Expression pattern shifts following duplication indicative
of subfunctionalization and neofunctionalization in regulatory
genes of Arabidopsis. Mol. Biol. Evol. 23: 469–478.
Endrizzi, J. E., E. L. Turcotte and R. J. Kohel, 1985 Genetics, cy-
togenetics, and evolution of Gossypium. Adv. Genet. 23: 271–375.
Federico, M. L., F. L. Iniguez-Luy,R.W.Skadsen and H. F.
Kaeppler, 2006 Spatial and temporal divergence of expression
in duplicated Barley germin-like protein-encoding genes. Genet-
ics 174: 179–190.
Flagel, L., J. A. Udall,D.Nettleton and J. F. Wendel,
2008 Duplicate gene expression in allopolyploid Gossypium re-
veals two temporally distinct phases of expression evolution.
BMC Biol. 6: 16.
Force, A., M. Lynch,F.B.Pickett,A.Amores, Y.-L. Yan et al.,
1999 Preservation of duplicate genes by complementary, de-
generative mutations. Genetics 151: 1531–1545.
Gaeta, R. T., J. C. Pires,F.Iniguez-Luy,E.Leon and T. C. Osborn,
2007 Genomic changes in resynthesized Brassica napus and
their effect on gene expression and phenotype. Plant Cell 19:
3403–3417.
Grover, C. E., H. Kim,R.A.Wing,A.H.Paterson and J. F. Wendel,
2004 Incongruent patterns of local and global genome size evo-
lution in cotton. Genome Res. 14: 1474–1482.
Grover, C. E., H. Kim,R.A.Wing,A.H.Paterson and J. F. Wendel,
2007 Microcolinearity and genome evolution in the AdhA re-
gion of diploid and polyploid cotton (Gossypium). Plant J. 50:
995–1006.
He, C., and H. Saedler, 2005 Heterotopic expression of MPF2 is
the key to the evolution of the Chinese lantern of Physalis, a mor-
phological novelty in Solanaceae. Proc. Natl. Acad. Sci. USA 102:
5779–5784.
Hegarty, M. J., G. L. Barker,I.D.Wilson,R.J.Abbott,K.J.
Edwards et al., 2006 Transcriptome shock after interspecific
hybridization in senecio is ameliorated by genome duplication.
Curr. Biol. 16: 1652–1659.
Hovav, R., J. A. Udall,B.Chaudhary,E.Hovav,L.Flagel et al.,
2008a The evolution of spinable cotton fiber entailed natural
selection for prolonged development and a novel metabolism.
PLoS Genet. 4(2): e25.
Hovav, R., J. A. Udall,B.Chaudhary,R.Rapp,L.Flagel et al.,
2008b Partitioned expression of duplicated genes during devel-
opment and evolution of a single cell in a polyploid plant. Proc.
Natl. Acad. Sci. USA 105: 6191–6195.
Hovav, R., J. A. Udall,E.Hovav,R.A.Rapp,L.Flagel et al.,
2007 A majority of genes are expressed in the single-celled seed
trichome of cotton. Planta 227: 319–329.
Keeling,C.I.,S.Weisshaar,R.P.C.Lin and J. Bohlmann,
2008 Functional plasticity of paralogous diterpene synthases in-
volvedin coniferdefense.Proc.Natl. Acad.Sci.USA 105: 1085–1090.
Kellis, M., B. W. Birren and E. S. Lander, 2004 Proof and evolu-
tionary analysis of ancient genome duplication in the yeast Sac-
charomyces cerevisiae. Nature 428: 617–624.
Lee, H.-S., and Z. J. Chen, 2001 Protein-coding genes are epigenet-
ically regulated in Arabidopsis polyploids. Proc. Natl. Acad. Sci.
USA 98: 6753–6758.
Leitch, I. J., and M. D. Bennett, 1997 Polyploidy in angiosperms.
Trends Plant Sci. 2: 470–476.
Levy, A. A., and M. Feldman, 2002 The impact of polyploidy on
grass genome evolution. Plant Physiol. 130: 1587–1593.
Liu, B., C. L. Brubaker,G.Mergeai,R.C.Cronn and J. F. Wendel,
2001 Polyploid formation in cotton is not accompanied by
rapid genomic changes. Genome 44: 321–330.
Liu, B., and J. Wendel, 2002 Non-Mendelian phenomenon in allo-
polyploid genome evolution. Curr. Genomics 3: 489–505.
Liu, Z., and K. L. Adams, 2007 Expression partitioning between
genes duplicated by polyploidy under abiotic stress and during
organ development. Curr. Biol. 17: 1669–1674.
Lockton, S., and B. S. Gaut, 2005 Plant conserved non-coding se-
quences and paralogue evolution. Trends Genet. 21: 60–65.
Lynch, M., and J. S. Conery, 2000 The evolutionary fate and con-
sequences of duplicate genes. Science 290: 1151–1155.
Lynch, M., and A. Force, 2000 The probability of duplicate gene
preservation by subfunctionalization. Genetics 154: 459–473.
Lynch, M., M. O’Hely,B.Walsh and A. Force, 2001 The proba-
bility of preservation of a newly arisen gene duplicate. Genetics
159: 1789–1804.
Madlung, A., R. W. Masuelli,B.Watson,S.H.Reynolds,
J. Davison et al., 2002 Remodeling of DNA methylation and
phenotypic and transcriptional changes in synthetic Arabidopsis
allotetraploids. Plant Physiol. 129: 733–746.
Matzke, M. A., S. O. Mittelsten and A. J. Matzke, 1999 Rapid
structural and epigenetic changes in polyploid and aneuploid ge-
nomes. BioEssays 21: 761–767.
Mena, M., B. A. Ambrose,R.B.Meeley,S.P.Briggs,M.F.Yanofsky
et al., 1996 Diversification of C-function activity in maize flower
development. Science 274: 1537–1540.
Ohno, S., 1970 Evolution by Gene Duplication. Springer-Verlag, New
York.
Osborn, T. C., J. Chris Pires,J.A.Birchler,D.L.Auger,Z.Jeffery
Chen et al., 2003 Understanding mechanisms of novel gene ex-
pression in polyploids. Trends Genet. 19: 141–147.
Percy, R. G., and J. F. Wendel, 1990 Allozyme evidence for the or-
igin and diversification of Gossypium barbadense L. Theor. Appl.
Genet. 79: 529–542.
Pires, J. C., J. Zhao,M.E.Schranz,E.J.Leon,P.A.Quijada et al.,
2004 Flowering time divergence and genomic rearrangements
in resynthesized Brassica polyploids (Brassicaceae). Bio. J. Linn.
Soc. 82: 675–688.
Rastogi, S., and D. A. Liberles, 2005 Subfunctionalization of du-
plicated genes as a transition state to neofunctionalization. BMC
Evol. Biol. 5: 28.
Riddle, N. C., and J. A. Birchler, 2003 Effects of reunited diverged
regulatory hierarchies in allopolyploids and species hybrids.
Trends Genet. 19: 597–600.
Salmon, A., M. L. Ainouche and J. F. Wendel, 2005 Genetic and
epigenetic consequences of recent hybridization and polyploidy
in Spartina (Poaceae). Mol. Ecol. 14: 1163–1175.
Seelanan, T., A. Schnabel and J. F. Wendel, 1997 Congruence
and consensus in the cotton tribe (Malvaceae). Syst. Bot. 22:
259–290.
Senchina, D. S., I. Alvarez,R.C.Cronn,B.Liu,J.Rong et al.,
2003 Rate variation among nuclear genes and the age of poly-
ploidy in Gossypium. Mol. Biol. Evol. 20: 633–643.
Shaked, H., K. Kashkush,H.Ozkan,M.Feldman and A. A. Levy,
2001 Sequence elimination and cytosine methylation are rapid
and reproducible responses of the genome to wide hybridization
and allopolyploidy in wheat. Plant Cell 13: 1749–1759.
Sliwinski, M. K., J. A. Bosch, H.-S. Yoon,M.v.Balthazar and D. A.
Baum, 2007 The role of two LEAFY paralogs from Idahoa scapi-
gera (Brassicaceae) in the evolution of a derived plant architec-
ture. Plant J. 51: 211–219.
Soltis, D. E., and P. S. Soltis, 1999 Polyploidy: recurrent formation
and genome evolution. Trends Ecol. Evol. 14: 348–352.
Soltis, D. E., P. S. Soltis and J. A. Tate, 2004 Advances in the study
of polyploidy since plant speciation. New Phytol. 161: 173–191.
Song, K., P. Lu,K.Tang and T. C. Osborn, 1995 Rapid genome
change in synthetic polyploids of Brassica and its implications
for polyploid evolution. Proc. Natl. Acad. Sci. USA 92: 7719–
7723.
516 B. Chaudhary et al.
Springer, N. M., and R. M. Stupar, 2007a Allele-specific expression
patterns reveal biases and embryo-specific parent-of-origin effects
in hybrid maize. Plant Cell 19: 2391–2402.
Stephens, S. G., 1951 Possible significances of duplication in evolu-
tion. Adv. Genet. 4: 247–265.
Storey, J. D., and R. Tibshirani, 2003 Statistical significance for ge-
nomewide studies. Proc. Natl. Acad. Sci. USA 100: 9440–9445.
Stupar, R. M., P. B. Bhaskar,B.S.Yandell,W.A.Rensink,A.L.
Hart et al., 2007 Phenotypic and transcriptomic changes asso-
ciated with potato autopolyploidization. Genetics 176: 2055–
2067.
Stupar, R. M., and N. M. Springer, 2006 Cis-transcriptional varia-
tion in maize inbred lines B73 and Mo17 leads to additive expres-
sion patterns in the F1 hybrid. Genetics 173: 2199–2210.
Swanson-Wagner, R. A., Y. Jia,R.DeCook,L.A.Borsuk,
D. Nettleton et al., 2006 All possible modes of gene action
are observed in a global comparison of gene expression in a
maize F1 hybrid and its inbred parents. Proc. Natl. Acad. Sci.
USA 103: 6805–6810.
Taliercio, E. W., and D. Boykin, 2007 Analysis of gene expression
in cotton fiber initials. BMC Plant Biol. 7: 22.
Tate, J. A., Z. Ni,A.C.Scheen,J.Koh,C.A.Gilbert et al.,
2006 Evolution and expression of homeologous loci in Tragopo-
gon miscellus (Asteraceae), a recent and reciprocally formed allo-
polyploid. Genetics 173: 1599–1611.
Teshima, K. M., and H. Innan, 2008 Neofunctionalization of dupli-
cated genes under the pressure of gene conversion. Genetics
178: 1385–1398.
Thomas, B. C., B. Pedersen and M. Freeling, 2006 Following tetra-
ploidy in an Arabidopsis ancestor, genes were removed preferen-
tially from one homeolog leaving clusters enriched in dose-sensitive
genes. Genome Res. 16: 934–946.
Udall, J. A., P. A. Quijada,H.Polewicz,R.Vogelzang and T. C.
Osborn, 2004 Phenotypic effects of introgressing chinese win-
ter and resynthesized Brassica napus L. germplasm into hybrid
spring canola. Crop Sci. 44: 1990–1996.
Udall, J. A., J. M. Swanson,D.Nettleton,R.J.Percifield and J. F.
Wendel, 2006 A novel approach for characterizing expression
levels of genes duplicated by polyploidy. Genetics 173: 1823–
1827.
Van Damme, E. J. M., R. Culerrier,A.Barre,R.Alvarez,P.Rouge
et al., 2007 A novel family of lectins evolutionarily related to
class V chitinases: an example of neofunctionalization in le-
gumes. Plant Physiol. 144: 662–672.
Wan, C., and T. Wilkins, 1994 A modified hot borate method sig-
nificantly enhances the yield of high-quality RNA from cotton
(Gossypium hirsutum L.). Anal. Biochem. 223: 7–12.
Wang, J., L. Tian,A.Madlung, H.-S. Lee,M.Chen et al.,
2004 Stochastic and epigenetic changes of gene expression
in Arabidopsis polyploids. Genetics 167: 1961–1973.
Wang, J., L. Tian, H.-S. Lee,N.E.Wei,H.Jiang et al.,
2006 Genomewide nonadditive gene regulation in Arabidopsis
allotetraploids. Genetics 172: 507–517.
Wang, Y.-M., Z.-Y. Dong, Z.-J. Zhang, X.-Y. Lin,Y.Shen et al.,
2005 Extensive de novo genomic variation in rice induced by in-
trogression from wild rice (Zizania latifolia Griseb.). Genetics 170:
1945–1956.
Wendel, J. F., 2000 Genome evolution in polyploids. Plant Mol.
Biol. 42: 225–249.
Wendel, J. F., and V. A. Albert, 1992 Phylogenetics of the cotton
genus Gossypium: character-state weighted parsimony analysis of
chloroplast-DNA restriction site data and its systematic and bio-
geographic implications. Syst. Bot. 17: 115–143.
Wendel, J. F., and R. C. Cronn, 2003 Polyploidy and the evolution-
ary history of cotton. Adv. Agron. 78: 139–186.
Wittkopp, P. J., B. K. Haerum and A. G. Clark, 2004 Evolutionary
changes in cis and trans gene regulation. Nature 430: 85–88.
Yang, S., F. Cheung,J.J.Lee,M.Ha,N.E.Wei et al.,
2006 Accumulation of genome-specific transcripts, transcrip-
tion factors and phytohormonal regulators during early stages
of fiber cell development in allotetraploid cotton. Plant J. 47:
761–775.
Zhuang, Y., and K. L. Adams, 2007 Extensive allelic variation in
gene expression in Populus F1 hybrids. Genetics 177: 1987–1996.
Communicating editor: A. H. Paterson
Homeolog Expression in Cotton 517
Supporting Information
http://www.genetics.org/cgi/content/full/genetics.109.102608/DC1
Reciprocal Silencing, Transcriptional Bias and Functional Divergence
of Homeologs in Polyploid Cotton (Gossypium)
Bhupendra Chaudhary, Lex Flagel, Robert M. Stupar, Joshua A. Udall, Neetu Verma,
Nathan M. Springer, and Jonathan F. Wendel
Copyright © 2009 by the Genetics Society of America
DOI: 10.1534/genetics.109.102608
B. Chaudhary et al.
2 SI!
File S1
Raw data is available in a compressed file at http://www.genetics.org/cgi/content/full/genetics.109.102608/DC1.
B. Chaudhary et al.
3 SI!
FIGURE S1.—Number of informative genome-specific assays for each tissue type.
B. Chaudhary et al.
4 SI!
FIGURE S2.—Expression values of mix samples in 8 tissues with relative contribution of A-and D-homoeologs.
B. Chaudhary et al.
5 SI!
FIGURE S3.—Relative genome-specific expression states for all 9 neo-functionalized genes in all available tissues of F1,
synthetic and Maxxa.
B. Chaudhary et al.
6 SI!
gene
GENBANK
tissue
D3_%D_exp
D5_%D_exp
absolutediff.D3vsD5
COTTON16_00001_192_2122
CO071793
Leaf
0.359
0.643
0.285
COTTON16_00004_01_779
CO108066
Leaf
0.117
0.352
0.235
COTTON16_00056_02_720
AAP41846
Leaf
0.447
0.693
0.246
COTTON16_00069_04_828
CO076921
Leaf
0.053
0.044
0.008
COTTON16_00173_04_1494
CO122994
Leaf
0.701
0.821
0.12
COTTON16_00285_02_685
CO131379
Leaf
0.529
0.46
0.069
COTTON16_00373_02_1079
CO125131
Leaf
0.476
0.444
0.032
COTTON16_00727_02_140
CAO62858
Leaf
0.698
0.375
0.323
COTTON16_01040_02_1017
CO111355
Leaf
0.326
0.306
0.019
COTTON16_01391_01_705
DT461656
Leaf
0.583
0.717
0.134
COTTON16_02074_02_678
CO081139
Leaf
0.608
0.413
0.195
COTTON16_07872_01_1017
CO118336
Leaf
0.613
0.885
0.272
COTTON16_07872_01_1185
EE592712
Leaf
0.578
0.839
0.261
COTTON16_09095_01_1544
DT546602
Leaf
0.605
0.537
0.068
COTTON16_21601_01_747
CO101293
Leaf
0.366
0.22
0.146
COTTON16_24663_01_633
CO073158
Leaf
0.148
0.297
0.149
COTTON16_36070_01_1103
CO115511
Leaf
0.389
0.364
0.024
COTTON16_00001_192_2122
CO071793
Petiole
0.733
0.932
0.199
COTTON16_00004_01_779
CO108066
Petiole
0.476
0.459
0.017
COTTON16_00056_02_720
AAP41846
Petiole
0.828
0.703
0.125
COTTON16_00069_04_828
CO076921
Petiole
0.589
0
0.589
COTTON16_00173_04_1494
CO122994
Petiole
0.798
0.898
0.1
COTTON16_00285_02_685
CO131379
Petiole
0.703
0.581
0.122
COTTON16_00373_02_1079
CO125131
Petiole
0.602
0.508
0.093
COTTON16_00727_02_140
CAO62858
Petiole
0.904
0.765
0.139
COTTON16_01040_02_1017
CO111355
Petiole
0.543
0.344
0.198
COTTON16_01391_01_705
DT461656
Petiole
0.794
0.705
0.089
COTTON16_02074_02_678
CO081139
Petiole
0.706
0.597
0.109
COTTON16_07872_01_1017
CO118336
Petiole
0.9
0.97
0.07
COTTON16_07872_01_1185
EE592712
Petiole
0.821
0.941
0.119
COTTON16_09095_01_1544
DT546602
Petiole
0.747
0.682
0.066
COTTON16_17428_01_832
ABA95925
Petiole
0.726
0.891
0.165
COTTON16_19657_01_224
NP_196352
Petiole
0.533
0.893
0.36
COTTON16_24663_01_633
CO073158
Petiole
0.436
0.622
0.185
COTTON16_36070_01_1103
CO115511
Petiole
0.603
0.51
0.093
average
0.572
0.583
0.155
B. Chaudhary et al.
7 SI!
FIGURE S4.— Quantification of expression divergence between G. davidsonii (D3) and G. raimondii (D5) relative to a
common G. arboreum (A2) reference sample.
B. Chaudhary et al.
8 SI!
TABLE S1
Genes used in this study
Contig
SNP_position
Reference_GeneBank_Accession
Gene_description
Cotton16_00001_034
832
CO112436
putative small GTP.binding protein
Cotton16_00001_056
1492
CO130933
clathrin adaptor medium chain protein MU1B, putative
Cotton16_00001_062
1928
CO124017
vacuolar proton.ATPase subunit.like protein
Cotton16_00001_192
2122
CO071793
acetyl.CoA carboxylase
Cotton16_00001_452
1635
CO131164
phytochrome.associated protein 1
Cotton16_00004_01
779
CO108066
glyceraldehyde.3.phosphate dehydrogenase
Cotton16_00012_02
1315
CAO71171
hydroxyproline.rich glycoprotein.like protein
Cotton16_00013_06
1353
CO094037
putative cystathionine gamma.synthase
Cotton16_00024_03
2070
CO105110
calmodulin.like domain protein kinase
Cotton16_00025_07
221
CO081422
Os05g0455600 [Oryza sativa ]
Cotton16_00056_02
720
AAP41846
cysteine protease
Cotton16_00069_04
828
CO076921
VATA_GOSHI Vacuolar ATP synthase catalytic subunit A
Cotton16_00071_01
540
CO121715
AUX/IAA protein
Cotton16_00075_03
432
CO117833
aldehyde dehydrogenase family 7 member A1
Cotton16_00076_06
860
CO102987
ethylene transcription factor
Cotton16_00156_04
492
CO105072
putative beta.1,3.glucanase
Cotton16_00173_04
1494
CO122994
TBA3_ELEIN Tubulin alpha.3 chain
Cotton16_00174_02
802
CO111918
unknown protein
Cotton16_00197_01
52
CO090037
TPIC_SPIOL Triosephosphate isomerase
Cotton16_00285_02
685
CO131379
serine acetyltransferase 7
Cotton16_00373_02
1079
CO125131
ADT1_GOSHI ADP,ATP carrier protein 1
Cotton16_00479_01
592
CO093729
ATP binding
Cotton16_00690_02
916
CO118820
Beta.COP.like protein
Cotton16_00727_02
140
CAO62858
nucleoporin family protein
Cotton16_00922_01
949
CO124958
phosphoinositide.specific phospholipase C P13
Cotton16_01040_02
1017
CO111355
TBB5_GOSHI Tubulin beta.5 chain (Beta.5 tubulin)
Cotton16_01121_02
632
CO129884
2.phosphoglycerate kinase.related
Cotton16_01189_01
1178
CO130747
CBL.interacting protein kinase 16
Cotton16_01218_01
872
CO096546
LEJ2 (LOSS OF THE TIMING OF ET AND JA BIOSYNTHESIS 2)
Cotton16_01391_01
705
DT461656
phosphorybosyl anthranilate transferase 1
Cotton16_01436_02
169
CO102224
t.complex polypeptide 1
Cotton16_01499_01
472
CO110993
phosphate.responsive 1 family protein
Cotton16_01704_01
1415
DW008528
putative protein
Cotton16_01766_02
410
CO106047
oligosaccharide transporter
Cotton16_01818_02
625
CO085186
tetratricopeptide repeat (TPR).containing protein
B. Chaudhary et al.
9 SI!
Cotton16_02074_02
678
CO081139
transcription factor Hap5a
Cotton16_02786_01
891
CO084845
1,4.alpha.glucan branching enzyme
Cotton16_03680_01
309
CO085733
universal stress protein (USP) family protein
Cotton16_04501_01
681
CAO49511
EMB1417 (EMBRYO DEFECTIVE 1417)
Cotton16_06427_01
504
CO080453
unknown protein
Cotton16_07872_01
1017
CO118336
TBA4_GOSHI Tubulin alpha.4 chain (Alpha.4 tubulin)
Cotton16_09095_01
1294
CO111212
Ubiquitin.associated protein
Cotton16_09331_01
537
CO102525
leucine.rich repeat family protein
Cotton16_14442_01
395
CAO23634
S.formylglutathione hydrolase
Cotton16_15666_01
268
CO077994
putative c.myc binding protein MM.1
Cotton16_17428_01
832
ABA95925
Amidase
Cotton16_19029_01
267
CO075025
acyl carrier protein
Cotton16_19620_01
376
CO080701
Tubby; Di.trans.poly.cis.decaprenylcistransferase
Cotton16_19657_01
224
NP_196352
4SNc.Tudor domain protein
Cotton16_21601_01
747
CO101293
unnamed protein product
Cotton16_21697_01
393
CO080172
putative lateral suppressor region D protein
Cotton16_22170_01
513
CO098920
GTP.binding protein GB2
Cotton16_24663_01
633
CO073158
protein kinase family protein
Cotton16_25466_01
1125
CO129204
Calmodulin.binding transcription activator 2
Cotton16_26306_01
942
CO089285
putative secretory carrier.associated membrane protein
Cotton16_27501_01
1173
CO122743
predicted proline.rich protein
Cotton16_28738_01
753
CO122313
hypothetical protein MtrDRAFT_AC136288g4v1
Cotton16_32946_01
1145
CO087191
6.phosphogluconate dehydrogenase
Cotton16_34102_01
940
AAK69758
unknown protein
Cotton16_34627_01
1149
CO082621
putative protein kinase
Cotton16_35244_01
334
CO072998
calcineurin B.like protein
Cotton16_35439_01
812
CO131697
mutant cincinnata
Cotton16_36070_01
1103
CO115511
ethylene signaling protein
