Dynamics and adaptive benefits of protein domain emergence and arrangements during plant genome evolution.

Abstract

Plant genomes are generally very large, mostly paleopolyploid, and have numerous gene duplicates and complex genomic features such as repeats and transposable elements. Many of these features have been hypothesized to enable plants, which cannot easily escape environmental challenges, to rapidly adapt. Another mechanism, which has recently been well described as a major facilitator of rapid adaptation in bacteria, animals, and fungi but not yet for plants, is modular rearrangement of protein-coding genes. Due to the high precision of profile-based methods, rearrangements can be well captured at the protein level by characterizing the emergence, loss, and rearrangements of protein domains, their structural, functional, and evolutionary building blocks. Here, we study the dynamics of domain rearrangements and explore their adaptive benefit in 27 plant and 3 algal genomes. We use a phylogenomic approach by which we can explain the formation of 88% of all arrangements by single-step events, such as fusion, fission, and terminal loss of domains. We find many domains are lost along every lineage, but at least 500 domains are novel, that is, they are unique to green plants and emerged more or less recently. These novel domains duplicate and rearrange more readily within their genomes than ancient domains and are overproportionally involved in stress response and developmental innovations. Novel domains more often affect regulatory proteins and show a higher degree of structural disorder than ancient domains. Whereas a relatively large and well-conserved core set of single-domain proteins exists, long multi-domain arrangements tend to be species-specific. We find that duplicated genes are more often involved in rearrangements. Although fission events typically impact metabolic proteins, fusion events often create new signaling proteins essential for environmental sensing. Taken together, the high volatility of single domains and complex arrangements in plant genomes demonstrate the importance of modularity for environmental adaptability of plants.

Full text

Page 1

Dynamics and Adaptive Benefits of Protein Domain
Emergence and Arrangements during Plant Genome
Evolution
Anna R. Kersting, Erich Bornberg-Bauer, Andrew D. Moore*, and Sonja Grath*
Evolutionary Bioinformatics Group, Institute for Evolution and Biodiversity, University of Muenster (WWU), Germany
*Corresponding author: E-mail: radmoore@uni-muenster.de; s.grath@uni-muenster.de.
Accepted: 9 January 2012
Abstract
Plant genomes are generally very large, mostly paleopolyploid, and have numerous gene duplicates and complex genomic
features such as repeats and transposable elements. Many of these features have been hypothesized to enable plants, which
cannot easily escape environmental challenges, to rapidly adapt. Another mechanism, which has recently been well
described as a major facilitator of rapid adaptation in bacteria, animals, and fungi but not yet for plants, is modular
rearrangement of protein-coding genes. Due to the high precision of profile-based methods, rearrangements can be well
captured at the protein level by characterizing the emergence, loss, and rearrangements of protein domains, their structural,
functional, and evolutionary building blocks. Here, we study the dynamics of domain rearrangements and explore their
adaptive benefit in 27 plant and 3 algal genomes. We use a phylogenomic approach by which we can explain the formation
of 88% of all arrangements by single-step events, such as fusion, fission, and terminal loss of domains. We find many
domains are lost along every lineage, but at least 500 domains are novel, that is, they are unique to green plants and
emerged more or less recently. These novel domains duplicate and rearrange more readily within their genomes than ancient
domains and are overproportionally involved in stress response and developmental innovations. Novel domains more often
affect regulatory proteins and show a higher degree of structural disorder than ancient domains. Whereas a relatively large
and well-conserved core set of single-domain proteins exists, long multi-domain arrangements tend to be species-specific.
We find that duplicated genes are more often involved in rearrangements. Although fission events typically impact metabolic
proteins, fusion events often create new signaling proteins essential for environmental sensing. Taken together, the high
volatility of single domains and complex arrangements in plant genomes demonstrate the importance of modularity for
environmental adaptability of plants.
Key words: plant genome evolution, modular evolution, whole-genome duplication, evolution of stress response.
Introduction
The wealth of genomic data has governed a number of in-
sightful studies on genome evolution. To date, most studies
haveconcentratedongeneduplications,genefamilyexpan-
sion or reduction, selective sweeps or signals of selection us-
ing site-based statistics. An alternative approach to studying
genome evolution utilizes the modular nature of proteins.
Most proteins are composed of one or many protein do-
mains, which are the units of protein structure, function,
and evolution (So ¨ding and Lupas 2003; Moore et al.
2008). The majority of proteins can be described using
a small set of domains, which, despite the ever-increasing
amount of available sequence data, grows at only moderate
speed. In contrast, the number of domain arrangements,
that is, the combination of these domains in proteins, con-
tinues to rapidly grow (Levitt 2009; Yang et al. 2009). The
study of domain rearrangements across large phyla has pro-
videdadetailedunderstandingofmodularproteinevolution
(Bjo ¨rklund et al. 2005; Ekman et al. 2007; Fong et al. 2007;
Wang and Caetano-Anolles 2009; Yang et al. 2009) and has
demonstrated that domain rearrangements, paired with
the occasional formation of novel domains (Moore and
Bornberg-Bauer 2012), create an enormous degree of pro-
tein diversity (Apic et al. 2001; Levitt 2009; Yang et al.
2009). The majority of eukaryotic proteins have more than
one domain (Apic et al. 2001; Ekman et al. 2005; Yang et al.
ª The Author(s) 2012. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/
3.0), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
316
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004 Advance Access publication January 16, 2012
GBE

Page 2

2009), and while many domains are found in few arrange-
ments, only few domains are versatile and form a wide array
of different arrangements (Weiner et al. 2008; Cohen-Gihon
et al. 2011). Rearrangement events at the protein level are
easy to detect, and the key mechanisms are thought to be
fusion, fission, and terminal deletion (Bjo ¨rklund et al.
2005; Weiner et al. 2006). These events are likely fueled
by a series of underlying genetic events such as nonallelic ho-
mologous recombination, exon-shuffling, nonhomologous
end joining or transposition (Babushok et al. 2007; Buljan
et al. 2010). However, with few exceptions (e.g., Oshima
et al. 2010), traces of the genetic mechanisms of rearrange-
ment swiftly decay. Buljan et al. (2010) explored the genetic
events that facilitate domain gain events to existing arrange-
ments. Their results provide support to the notion that
domainsaretypicallyaddedateitherterminus.Thekeymech-
anism for such domain gain events involves the joining of
exons between genes or terminal exon extension. The study
ofdomaincontentevolutionineukaryoteshasillustratedthat
domain loss and gain are frequent events (Moore and
Bornberg-Bauer 2011; Zmasek and Godzik 2011). Whereas
lost domains tend to be of catalytic nature, gained domains
tend to be regulatory. Despite the diverse studies that have
explored modular evolution across many species as well as
in restricted clades, to date no study has quantitatively
addressed the topic of modularity in a set of plant species.
However, modular evolution may be of particular importance
forplants,astheyfaceachallengethatmanyotherspeciesdo
not—they cannot easily evade environmental changes
because of their sessile nature. In particular the fusion of
genes, and consequently of domain arrangements, allows
for ‘‘jumps’’ in protein evolution and may govern truly novel
genetic phenotypes. Hence such fusion proteins may exhibit
great adaptive potential. Indeed, recent findings suggest that
chimeric genes formed by gene fusion can be found in
regions of selective sweeps (Rogers and Hartl 2012).
Fusion events have been shown to be associated with
regulatory proteins such as the metazoan bHLH transcrip-
tion factors (Amoutzias et al. 2005) or the MIKC-type
MADS-box transcription factor proteins in plants (Veron
et al. 2007; Shan et al. 2009). Innovation of transcription
factor families is often the result of duplication events,
which may occur in chromosomal regions with high recom-
bination rates. Furthermore, it has been illustrated that du-
plication events in combination with high recombination
rates are strong forces in genome evolution (Lang et al.
2010).
Duplications have been more frequently described for
plants than elsewhereand plant genome evolution is special
in several aspects. First, plant genomes are repeat-rich and
transposable elements have a particularly prominent role in
creating retrocopies of genes, for example, in monocots
(Bennetzen 2005; Baucom et al. 2009; Baucom, Estill et al.
2009). Second, several whole-genome duplication (WGD)
events have created many large genomes with various de-
grees of ploidy within a relatively short period of time.
35% of all vascular plants are recent polyploids (Wood
et al. 2009). Moreover, angiosperms have undergone up to
fourroundsofWGDinroughly320Myr,withoneWGDcom-
mon to all seed plants 319 Ma and one WGD common to all
angiosperms 192 Ma (van de Peer et al. 2009; Jiao et al.
2011). Although polyploidy events pose a genomic challenge
to their host and most polyploidy events are considered
a ‘‘dead end’’ for evolution (Mayrose et al. 2011), it has been
suggested that polyploidy, be it the result of autopolyploidy or
allopolyploidy, may occasionally provide a starting point for
evolutionary innovation (Freeling et al. 2006; van de Peer
et al. 2009). The benefit of an increased amount of genetic
material might be to allow for swift adaptation to extreme
environments (van de Peer et al. 2009). For example, the in-
creased heterozygosity resulting from polyploidy impacts the
wiring of signaling cascades and can facilitate strong variation
in gene expression (Osborn et al. 2003). Numerous studies
have also explicitly explored the impact of WGD in plants
at the genomic level, for example, by exploring duplicate re-
tention rates (Hanada et al. 2008; Tang et al. 2008; Zheng
et al. 2009), gene dosage effects (Freeling et al. 2006;
Misook et al. 2007; Bekaert et al. 2011), or recombination
rates(Akhunovetal.2003).WGDsmayenhancethepotential
for diversification and speciation (van de Peer et al. 2009), yet
the details remain poorly understood.
As genomic stability is largely influenced by genome size
and repeat content (Bennetzen 2005), one might speculate
that plants have high rates of recombination and hence ex-
hibitahighnumberofdomainrearrangements.Indeed,com-
parative studies have illustrated that angiosperms exhibit
higher recombination rates than vertebrates (Kejnovsky
et al. 2009). However, to date, no study has explored the
extent of modular protein evolution in plants.
Giventheirlargegenomesize,higherrecombinationrates,
and the inability to flee upon environmental challenges, it
seems likely that plants may utilize their abundant genomic
material to facilitate rapid evolutionary innovation. Conse-
quently, the benefits of modular domain rearrangements
mightbeparticularlypronounced,sincetheabilityofmodular
evolution to swiftly implement changes to the protein reper-
toire maybe a key processinbothexploiting existingand cre-
ating functionalities. So far, all studies on the evolutionary
dynamics and the adaptive potential of domain rearrange-
mentshavebeenreportedfor bacteria(Enright andOuzounis
2001), metazoa (Ekman et al. 2007), or fungi (Cohen-Gihon
et al. 2011), but none for plants.
In this report, we explore the nature of modular evolution
in 29 green plant species (Viridiplantae) with taxa ranging
from green algae to liliopsida and eudicotyledons. Our aim
is to understand the evolutionary dynamics by studying the
frequencyofindividualmodulareventssuchasfusion,fission,
or terminal loss. We apply a maximum parsimony-based
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
317

Page 3

approach to reconstruct events placing this study into a phy-
logenomic framework and quantitatively address the role of
domain emergence and domain rearrangements. Further-
more, we explore the speed with which new domains,
and their arrangements, are gained and lost; how many of
these events are clade or species-specific and whether event
‘‘hotspots’’ can be found amongst the phylogenies of the
considered species. Finally, we employ several functional
analyses based on the Gene Ontology (GO) classification
(Ashburner and Lewis 2002) to shed light on the potential
adaptive benefits of domain emergence and rearrangements
during plant genome evolution.
Materials and Methods
Proteomes and Domain Annotation
Comparative analyses of protein domains and their arrange-
ments were performed on the following 29 plant genomes:
Arabidopsis thaliana v9.0 (The Arabidopsis Initiative 2000);
Arabidopsis lyrata v1.0 (Hu et al. 2011); Carica papaya v1.0
(Mingetal.2008);Citrussinensisv1.0(SweetOrangeGenome
Project 2010); Citrus clementine v0.9 (Haploid Clementine
Genome International Citrus Genome Consortium 2011);
Eucalyptus grandis v1.0 (Eucalyptus grandis Genome Project
2010);Mimulusguttatusv1.1;Aquilegiacoerulea;Theobroma
cacao v1.0 (Argout et al. 2011); Glycine max v1.0 (Schmutz
et al. 2010); Medicago truncatula v3.0 (Young et al. 2005);
Lotus japonica v1.0 (Young et al. 2005); Populus trichocarpa
v2.0 (Tuskan et al. 2006); Ricinus communis v1.0 (Chan et al.
2010); Manihot esculenta v1.1; Malus domestica (Velasco
et al. 2010); Prunus persica v1.0 (International Peach Genome
Initiative 2010); Cucumus sativa v1.0 (Huang et al. 2009); Vitis
vinifera v1.0 (Jaillon et al. 2007); Setaria italica v2.0 (Setaria
italica Genome Sequencing Project 2011); Zea mays v4a.53
(Schnable et al. 2009); Sorghum bicolor v1.4 (Dubchak
et al. 2009); Oryza sativa v6.1 (Go et al. 2002); Brachypodium
distachyon v1.0 (Vogel et al. 2010); Phoenix dactylifera v2.0
(Al-Dous et al. 2011); Selaginella moellendorffii v1.0 (Banks
et al. 2011); Physcomitrella patens v1.5 (Rensing et al.
2008); Chlamydomonas reinhardtii v4.0 (Merchant et al.
2007); Ostreococcus lucimarinus v2.0 (Palenik et al. 2007);
and Micromonas pusilla v3.0 (Worden et al. 2009).
Werootedthetree;1.700MabyincludingTrichoplaxad-
haerens v1.0 (Srivastava et al. 2008), Rhizopus oryzae (Ma
et al. 2009) and Drosophila melanogaster v5.11 (Adams
et al. 2000). Phylogenetic relationships for all 32 species
(29 plants and 3 outgroups) used for this study are given
in supplementary figure 1 (Supplementary Material online).
If several splice variants were present for one protein, we
excluded all but the longest transcript. All proteomes were
scanned for domains with the pfam_scan utility and
HMMER3.0 (Eddy 2011) against the Pfam-A and Pfam-B
models obtained from Pfam (v.24) (Finn et al. 2008).
For the annotation of Pfam-A domains, we used the
model-definedgatheringthresholdandquerysequenceswere
required to match at least 30% of the defining model (Buljan
et al.2010).Pfam-Bdomainswere annotatedusinganEvalue
cutoffof0.001(Ekmanetal.2007).Pfam-Adomainswithclan
membership were mapped to their clans and domains of type
‘‘repeat’’ or ‘‘motif’’ were collapsed into one large domain
instance (Ekman et al. 2005; Forslund et al. 2007).
Reconstruction of the Ancestral Domain State; Domain
Gain, Loss, and Emergence
We reconstructed ancestral domain contents using a maxi-
mum parsimony approach as follows: the tree (see fig. 1B)
was traversed twice, first from leaves to root then from root
toleaves.Domainpresenceorabsenceisdeterminedbyma-
jority rule. During first traversal (leaves / root), the state of
domain d is set to present at a node n, if d is present in the
majority of leaves of the subtree rooted in n (leaves of n).
Similarly, d is set to absent at n, if d is absent in the majority
of leaves of n. If there is no state majority for d in the child
nodes of n (i.e., there is an identical proportion of presence
and absence states for d in the leaves), the state of d at n is
set to unknown. As traversal continues toward the root, d is
set to present (absent) at n as soon as the majority of leaves
of n exhibit the present (absent) state. Ergo, present and un-
known are resolved to present, while unknown and absent
are resolved to absent. The first traversal terminates at the
root node. All unknown states at the root node are set to
present (note that this root includes the outgroups). During
the second traversal (root / leaves), unknown states are
resolved by setting them to the state of their ancestor.
We used a combination of custom-made python scripts
and the ETE2 package (Huerta-Cepas et al. 2010) for tree
traversal. Branch lengths of the tree (Soltis et al. 2002; Choi
et al. 2004; Magallo ´n and Sanderson 2005; Hedges et al.
2006; Cartwright and Collins 2007; Anderson and Janßen
2009; Bhattacharya et al. 2009; Bremer et al. 2009;
Forest and Chase 2009; Herron et al. 2009; Wang and
Caetano-Anolles 2009; Lang et al. 2010; Reineke et al.
2011) and whole-genome duplication events (Blanc and
Wolfe 2004; Schnable et al. 2009; van de Peer et al.
2009; Jiao et al. 2011) were extracted from the literature.
We performed a Blast (Altschul et al. 1997) search to
identify recently duplicated proteins. Proteins with a similar-
ity of 75% or more and an E value ? 10?20were considered
to be paralogs. We employed a synteny analysis to distin-
guish between tandem and segmental duplications. Two
genes wereconsidered to be tandem duplicates if they were
five or less genes apart. Paralogs with more than five genes
betweenthemwereconsideredtobearesultofasegmental
duplication event (Hanada et al. 2008).
Domain gain and loss events along branches were calcu-
latedbycomparisonofdomaincontentata givennodewith
Kersting et al.
GBE
318
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004 Advance Access publication January 16, 2012

Page 4

the domain content of its ancestor. We distinguish between
‘‘gained’’ domains, which are all domains found present at
a node while absent in its ancestor, and ‘‘emerged’’ do-
mains, which are gained domains which can only be found
within Viridiplantae. Ergo, emerged domains are a subset of
the gained domains. Emerged domains were determined by
scanning gained domains with HMMER3.0 against NCBI NR
and Integr8 (Kersey et al. 2005). Gained domains, which are
not present in the outgroups were also scanned against
NCBI NR to determine the kingdoms where these domains
are present(supplementary table 6,Supplementary Material
online). Domain event rates (gain and loss) were calculated
by dividing the number of events predicted to occur along
a given branch by the branch length (in million years).
Given the evidence that novel domains are frequently en-
riched in structural disorder (Buljan et al. 2010; Moore and
Bornberg-Bauer 2012), we predicted disorder in domains
classified as emerging. VSL2.0 (Obradovic et al. 2005)
was used to detect structural disorder in domain sequences.
Emerged domains were divided into four bins (Viridiplantae,
Embryophyta, Tracheophyta, and Magnoliophyta), corre-
sponding to their emergence nodes. Domains that emerged
after the Magnoliophyta node were pooled into one ‘‘RE-
CENT’’ bin. To compare disorder of emerged domains with
old domains (i.e., domains that exist at the root), a bin
‘‘OLD’’ was constructed consisting of 500 randomly picked
domains occurring in the root. In addition, we constructed
a ‘‘RANDOM’’ bin consisting of 100 randomly selected do-
mains, which exist at the root. To account for sampling bias,
werepeatedthe randomselection 100times.Statistical infer-
encewasconductedwiththekruskalmctestoftheRpackage
pgirmess (Siegel and Castellan 1988; R Development Core
Team 2008).
We quantified domain emergence and explored a set of
attributes (Moore and Bornberg-Bauer 2012). Domain
frequency, d(f), is defined as the absolute frequency of
FIG. 1.—Domain gain, loss, emergence and proteome coverage of 26 plant genomes. (A) Correlation of domain gain and loss with branch length.
Both gain and loss correlate significantly with branch length (gain: q 5 0.6, P , 0.001; loss: q 5 0.63, P , 0.001). (B) Phylogenetic relationship of all
species used in this study. For each branch, the size of the green circle corresponds to the number of domain emergence events along the branch.
Branches colored in red indicate that the gain and/or loss at this branch is higher than the average gain and/or loss rates. Exact values for domain gain,
loss, and emergence are given in supplementary table 2 (Supplementary Material online). (C) Domain coverage for proteins. The lower axis (percentage
of proteins with domains) displays the proportion of proteins with only Pfam-A domains (red), only Pfam-B domains (dark blue), both Pfam-A and
Pfam-B domains (light blue), and without any protein domain annotation (yellow). The upper axis displays proteome size indicated as vertical black line
for each species. Statistics for three species (Setaria italica, Prunus persica, and Mimulus guttatus) that are still under Fort Lauderdale restriction are not
provided.
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
319

Page 5

a domain across all plant genomes used for the analysis. The
domain rate x(d) of domain d is defined as the domain fre-
quency divided by the number of plants in which d occurs.
The domain success rate corresponds to the domain rate di-
vided by the node age (in million years) at which the domain
first emerged. The prevalence P(d) of a domain d is the num-
berofplantswithddividedbythenumberofplantswiththe
emergence node of d as an ancestor.
Functional Analysis of Domains
Where available, GO (Ashburner and Lewis 2002) annota-
tion of proteomes was obtained from PLAZA 2.0 (Proost
et al. 2009); Blast2GO (Go ¨tz et al. 2008) with default set-
tings was used to functionally annotate the remaining pro-
teomes. Comparative functional analyses were performed
by assessing GO-term overrepresentation (overrepresenta-
tion analysis, ORA) in two separate steps. First, for emerging
domains, we performed the functional analysis indirectly by
using the GO annotation of arrangements that harbor at
least one emerging domain, similar to a previous approach
(Mooreand Bornberg-Bauer 2012). Statistical inference was
conducted using the R package TopGo (Alexa et al. 2006).
As universe, we used the GO annotation of all proteins in
our data set; the sample consisted of arrangements with
emerging domains. Second, for assessing functional over-
representation of arrangements in events (such as fusion
or fission), we again conducted an ORA using the proteins
GO annotation, however, our sample here was the arrange-
ment set that results from a specific event (e.g., all gained
arrangements explainable by a fusion event). P value
transformed TermClouds were created by logarithmic
transformation of the False Discovery Rate (FDR)-corrected
(Benjamini and Hochberg 1995) P value obtained from the
ORA, such that term size represents the significance of the
GO term. Visualization was created using Wordle (http://
www.wordle.net/)with the transformedP value asa custom
scaling factor.
Reconstruction of the Ancestral Domain Arrangements
State, Arrangement Gain, and Loss
We defined domain arrangements as ordered sets of
domains for each protein. For the analysis of arrangements
in this study, only Pfam-A domains were used. Ancestral
states for arrangements were reconstructed as previously
described. Similarly, arrangement gain and loss was
determined by comparing current and ancestral states.
Determination of Arrangement Rates
For each gained arrangement, we applied a search algo-
rithm to determine the possible mechanism that led to its
formation. We considered the four most important mecha-
nisms of modular rearrangements—fusion, fission, terminal
deletion, and domain addition (Bjo ¨rklund et al. 2005; Pasek
et al. 2006; Weiner et al. 2006; Buljan et al. 2010). The al-
gorithm assigns a fusion event when two ancestral arrange-
ments can be fused to form the gained arrangement. A
gained arrangement is considered to be the result of fission
if an ancestral arrangement can be split to give rise to the
new arrangement; both products of the split are required to
be present in the current node. In contrast, for terminal de-
letion, only one product of the split (the gained arrange-
ment) may be present in the current node (the other
product is considered to be lost). The algorithm counts a do-
main addition event when the newly gained arrangement
contains a domain that is absent in the ancestral node.
Note that in general, any new arrangement can be ex-
plained by a sufficiently large ‘‘chain’’ of events. However,
since the likelihood of events is not available, we make
no assumptions about the relative costs of each mechanism
and therefore are not able to determine the most likely
chain. Instead, we focus on single-step solutions, that is,
on cases where a newly gained arrangement can be ex-
plained by a single event. Using this strategy, we can differ-
entiate between arrangements with exact solution (i.e., the
formation can be explained by exactly one mechanism), ar-
rangements with nonambiguous solution (i.e., only one
mechanism explains the arrangement but there are several
events possible) and arrangements with ambiguous solution
(i.e., conflicting solutions of different types). All arrange-
ments with solution are referred to as ‘‘simple gains,’’
whereas all other arrangements are considered to be
‘‘complex gains.’’
Results
Domain Coverage
In plants, on average, 50% of the proteome residues were
found to be covered by domain annotation; the residue cov-
erage ranges from 30% to 70% (supplementary table 1 and
fig. 2, Supplementary Material online). For an average of
35% of the residues, for each plant, a Pfam-A domain
can be detected, whereas Pfam-B domains affect 15% of
all residues. Residue coverage levels for all species are given
in supplementary table 1 (Supplementary Material online).
At the protein level, the coverage distribution is more di-
verse (supplementary table 1 and fig. 2, Supplementary
Material online). On average, 70% of the proteins for one
plant species have at least one Pfam-A or Pfam-B domain.
Fifty percent of the proteins contain only Pfam-A domains,
14% contain only Pfam-B domains, and 6% contain both
Pfam-A and Pfam-B domains (fig. 1C). All protein coverage
values are given in supplementary table 1 (Supplementary
Material online). The total number of proteins containing
Pfam-A and Pfam-B domains is highly variable between
the different proteomes (fig. 1C, supplementary table 1,
Supplementary Material online).
Kersting et al.
GBE
320
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004 Advance Access publication January 16, 2012

Page 6

Domain Emergence
To investigate domain gain, loss, and emergence across the
considered plants, we reconstructed the ancestral domain
content at each internal node of the tree (see also Materials
and Methods; supplementary fig. 1, Supplementary Material
online). In total, 545 domains emergedin the plant kingdom,
that is, these domains are exclusively found in Viridiplantae.
The largest amount of domain emergence within plants
occurs along the branch leading to Embryophyta, which sees
the emergence of 262 domains (fig. 1B). A total of 114 and
66 domains emerge along the branches to Magnoliophyta
and Tracheophyta, respectively. Fifty-one domains emerged
prior to the split of Embryophyta and the green algae and
52 domains are the result of recent emergence events and
can only be found within Magnoliophyta (see also Discussion
below) (fig. 1B).
Radiation and Functional Impact of Emerging Domains
Next, we assessed whether emerged domains confer spe-
cific functionalities and whether these might provide adap-
tive benefit. We assessed functional overrepresentation
using GO categories and TopGO (Alexa et al. 2006) (see
Materials and Methods for details). We find that GO terms
prefixed by response_to are overrepresented along with
functionalities related to reproduction, developmental
mechanisms, and metabolic processes (fig. 2).
We binned emerging domains according to their point of
emergence (for details, see Materials and Methods) and
ranked them by their frequency d(f). The 5% highest ranked
domains from each age bin (supplementary table 3, Supple-
mentaryMaterialonline)weresubjecttofurtherinvestigation
as these can be considered to be particularly ‘‘successful’’
emerging domains. Among these, we find domains with
plant-specific functionssuch asfloweringcontrol,auxinreg-
ulation, fruit development, cell wall development, and plant
organelle recognition. Furthermore, we detected domains
related to the F-box protein family, to transcription factors
and to DNA binding. For the majority of emerging domains,
direct functional annotation is difficult—the largest propor-
tion (85%) of all emerging domains in plants are domains of
unknown function (DUFs) or belong to the set of poorly an-
notated Pfam-B domains. We assessed functional overrepre-
sentationusingthefunctionof proteinsthat obtainemerging
domains—wearehencenotexploringwhichfunctionalmod-
ules emerge but rather which protein functionalities undergo
innovation (by the addition of an emerging domain).
There is increasing evidence that young domains can ex-
hibit higher levels of structural disorder than established do-
mains (Buljan et al. 2010; Moore and Bornberg-Bauer
2012). We examined the degree of structural disorder in
emerging domains. The results indicate that emerging do-
mains are significantly enriched in intrinsic disorder, more
than in randomly chosen domains (see Materials and Meth-
ods; supplementary fig. 3, Supplementary Material online).
Furthermore, the younger a domain, the higher the degree
of disorder.
Domain Gain and Loss
Domain gain and loss are frequent events in plant evolution,
and there is a strong variation between different branches
(fig.1A).Nevertheless,bothgainandlossratescorrelatesig-
nificantly with branch length (Spearman rank correlation,
gain: q 5 0.6, P , 0.001; loss: q 5 0.63, P , 0.001). On
average, plants have a domain gain rate of 6.64/Myr and
a domain loss rate of 6.11/Myr (fig. 1A, supplementary
table 2 and fig. 9, Supplementary Material online). In
monocots, the average domain gain rate (6.7/Myr) is
lower than the domain loss rate (7.4/Myr), whereas in eu-
dicots the situation is reversed; eudicots show a loss rate
of 7.4/Myr and a gain rate of 8.3/Myr (supplementary
table 2 and fig. 9, Supplementary Material online). Some
branches exhibit very high loss rates, such as the branch
leading to P. dactylifera, the branches to the two Fabaceae
M. truncatula and L. japonica, and the branches to the
two Andropogoneae Z. mays and S. bicolor (fig. 1B).
Gain, Loss, and Distribution of Arrangements
We next explored the dynamics of arrangement gain and
loss. After determining the presence/absence of arrange-
ments at ancestral nodes (for details, see Materials and
FIG. 2.—Gene Ontology (GO) terms associated with emerging domains. GO terms affected by emergence were tested for overrepresentation
using the TopGO package and all terms present in plants as universe (for details, see Materials and Methods). The font size corresponds to the value of
significance obtained for this term. Significance was determined after correction for multiple testing using FDR (Benjamini and Hochberg 1995) correction
at P , 0.01. The vast majority of GO terms is related to stimulus response, development, reproduction, regulation, and plant-specific metabolic processes.
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
321

Page 7

Methods), wecompared arrangementcontentat eachnode
with the content at the corresponding parent node to de-
termine arrangement gain and loss. As expected, both gain
and loss rates correlate significantly with branch length
(Spearman rank correlation, gain: q 5 0:56, P , 0.001; loss:
q 5 0.38, P 5 0.003, supplementary fig. 5, Supplementary
Material online). Overall, arrangement gain rate is higher
than arrangement loss rate. However, both rates correlate
significantly with each other (q 5 0.56, P , 0.001). By
far,thelargestamountofarrangementgain(2,814arrange-
ments) occurs along the branch to M. domestica followed
by the branch to R. communis (1,018). Large amounts of
arrangementlosscanbefoundalong thebranchestoP.dac-
tylifera (1,028) and L. japonica (680); both plants also
showed a high amount of domain loss. All values for ar-
rangement gain and loss are given in supplementary table
4 (Supplementary Material online).
We investigated the amount of arrangements shared by
different plants species (fig. 3). The distribution is bimodal,
with the largest number of arrangements being either spe-
cific to one species (;7,000) or shared by all (;1,000); only
a very small amount of arrangements is shared by 10–20
species. Although by far the largest proportion of arrange-
mentssharedbyallspeciesconsistsofsingle-domainproteins,
the contrary is true for species-specific arrangements. Here,
the largest number of arrangements tends to be composed
of more than one domain, with a large proportion contain-
ing seven or more domains. This indicates that the longer an
arrangement is, the more likely it is species-specific.
Modular Rearrangements
Usingasimplemodelofmodularrearrangement(fordetails,
see Materials and Methods), we next explored the mecha-
nisms that can facilitate the formation of novel arrange-
ments. For this, we considered fusion, fission, terminal
deletion, and domain addition. The results illustrate that
70% of all gained arrangements can be explained by exactly
one solution (exact solutions). Of the gained arrangements,
14% can be explained by one particular mechanism, how-
ever, with a number of different possible solutions (nonam-
biguous solutions); only 4% have conflicting solutions
(ambiguous solutions). The remaining 12% of all new ar-
rangements are complex gains that likely arose by a chain
of events (see Materials and Methods; fig. 4). The different
events were found to occur with different frequencies
(table 1). Fusion events makeup the largest proportion of
exact solutions, followed by domain addition, fission, and
terminal deletion. Fusion events occur with a frequency
of 4.59/Myr, followed by fission with 1.98/Myr, and gain
with 1.89/Myr. Domain deletion events can be split in
C-terminal and N-terminal domain deletion; both events
have a frequency of 0.7/Myr. All rates were averaged across
all branches. We further explored event frequencies across
different age bins. At the Embryophyta node, 68% of
new arrangements are affected by domain addition and
26% by fusion. Domain deletion (4%) and fission (3%)
are less prevalent at this node. Over time, the frequency
of domain deletion and fission increases up to 13% and
21% in recent rearrangements, whereas domain additions
FIG. 3.—Arrangements shared between species. The dashed line represents the number of arrangements shared by the different numbers of
species (right axis). The distribution of unique arrangements is roughly bimodal with the majority of arrangements shared by either few or all species.
The left axis and barplots display the frequency of arrangements with a certain length (one, two, three, four, five, six, and seven or more domains).
Although single-domain arrangements tend to occur in all species, longer arrangements are often species-specific.
Kersting et al.
GBE
322
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004 Advance Access publication January 16, 2012

Page 8

decrease to a frequency of 24%. The largest fraction of re-
cently gained arrangements (49%) can be explained by fu-
sion events (fig. 4).
Discussion
Domain Emergence
The increasing availability of plant genomes has allowed us
to conduct a comparative domain analysis between a set of
diverse plant species. Here, we reconstruct the ancestral
states of domain content and arrangement and investigate
the functional impact of domain emergence and domain re-
arrangements across a comprehensive set of 29 genomes
dating back ;800 Myr. However, the considered clade still
contains a number of species for which genome sequences
are missing,suchasthegymnospermsorthecharophyta.As
these genomes become available, a more comprehensive
picture of modular evolution in plants will emerge.
In contrast to animals, plants are sessile organisms that
are unable to escape strong environmental shifts and must
rather adapt to such variation. Hence, plants, more so than
animals, are required to evolve mechanisms in order to deal
with biotic and abiotic stresses. Here, we illustrate that the
emergence of new domains can provide an important strat-
egy for evolving stress response. More than 500 domains
emerged within Viridiplantae of which more than 100 do-
mains are unique for Tracheophyta (fig. 1). We recently as-
sessed the impact of domain emergence in a set of
insects, where only 30 domains emerged within 19 insect ge-
nomes spanning roughly 300 Myr of evolution (Moore and
Bornberg-Bauer 2011). Hence, it would seem that plants ex-
hibit a large amount of domain innovation. One might spec-
ulate that plants at least partly address the challenge of
a sessile lifestyle by means of domain innovation. The inves-
tigationofGOtermsofproteinscontainingemergeddomains
further supports this notion. A large number of terms are re-
latedtoplant-specificprocesses,suchasmegagametogenesis
and development of plant-specific organs. This is not surpris-
ing as the reproductive system and morphology of plants not
only differ strongly from other kingdoms but are also highly
variable between plant species (Endress 2001; Bennici 2005;
Williams 2008; Kawakita and Kato 2009). Besides these
plant-specific functions, a number of overrepresented GO
termscorrespondtoresponse_tocategoriesandtosecondary
metabolite pathways related tostress response, such as auxin
and jasmonic acid. Such secondary metabolites are strongly
related to the defense and response mechanisms in plants
(Grace and Logan 2000; Pateraki and Kanellis 2010; Kerchev
et al. 2012). As the composition of these compounds is
variable between plant species and also within species
(Kroymann 2011), such secondary metabolites may provide
a strong flexible basis for improving adaptation and defense.
Functional links tophotosynthesis arenotfound amongst
emerged domains (fig. 2). This is likely explained by
Table 1
Contribution of Fusion, Fission, C-Terminal Deletion, N-Terminal
Deletion and Domain Addition to Simple Arrangement Gains
Fusion Fission C-DelN-DelAdd
Total number
Average number/Myr
9,669
4.59
4,073
1.98
1,283
0.7
1,424
0.7
4,848
1.89
NOTE.—Del, deletion; Add, addition.
FIG. 4.—Mechanisms of rearrangement across different clades. We applied a search algorithm to assess the mechanisms that might account for
newly gained arrangements (for details, see Materials and Methods). Only 12% of all gained arrangements cannot be explained by a one-step event
(complex gains). The remaining 88% of simple gains can be further differentiated into exact solutions where only one particular mechanism (fusion,
fission, terminal deletion, or domain addition) was necessary to explain the arrangement gain event (70%). All proteomes were divided into four
different age bins: Embryophyta, Tracheophyta, Magnoliophyta, and Recent Nodes. The frequencies of fusion, fission, and terminal deletion increase
over time, whereas the frequency of domain addition decreases.
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
323

Page 9

photosynthesis not being unique to plants; photosynthetic
processes can be found in algae and in many species of bac-
teria (Olson 1970, 2001). Indeed, photosynthesis-related
GO terms can be detected by investigating gained domains
which are absent in the outgroups (supplementary fig. 6,
Supplementary Material online), as well as response_to
terms and a number of plant-specific functionalities related
to development, similar to those terms found in proteins
containing emerged domains.
Emerged domains seem to be evolutionarily important as
they have a high prevalence of 0.9–1, indicating that their
occurrence is strongly conserved. Besides their widespread
occurrence in nearly all leaves, such emerged domains often
occur in high copy numbers (supplementary table 3,
Supplementary Material online).
Investigating the most successful emerged domains un-
coversconnectionstokeyfunctionalcategoriessuchastran-
scription factors, binding-related processes, and secondary
metabolites, including auxin and jasmonic acid (supplemen-
tary table 3, Supplementary Material online). Indeed, a burst
of transcription factors and their constituent domains,
which are assumed to be correlated with increasing com-
plexity in plant evolution (Lang et al. 2010), has been found
in angiosperms. The increase of plant complexity with du-
plicationevents(Freelingetal.2006)maypartlybetheresult
of duplication facilitating increasingly complex regulatory
networks (Veron et al. 2007).
Emerging domains exhibit an increased amount of intrin-
sicdisorder;themorerecenttheemergenceevent,themore
likely the domain in question exhibits intrinsic disorder. Dis-
ordered sequences may increase the binding affinity of pro-
teins(DysonandWright2005).Highintrinsicdisorderpaired
with the fact that emerged domains are significantly under-
represented in single-domain proteins (hypergeometric test,
P , 0.01), leads us to the speculation that emerging do-
mains may have higher interaction potential, which in turn
may increase their viability and result in higher prevalence
and frequency. Indeed, some of the most successful emerg-
ing domains have links to binding-related processes.
Arrangement Mechanisms
In plants, roughly 70% of the domain-containing proteins
are single domain (supplementary fig. 4, Supplementary
Material online). This high percentage of single-domain pro-
teins can be an artifact of low domain coverage or ‘‘eroded-
domains,’’ which have diverged beyond detection (Weiner
etal.2006).Recentrearrangementscanmostlybeexplained
by the fusion of two single or two domain proteins. The
relative rates of fusion and fission are similar to previously
reported rates (Kummerfeld and Teichmann 2005). GO
terms overrepresented in proteins, which arose via fusion,
are stress-, defense-, and adaptation-related as well as
related to the reproduction system (supplementary fig. 7,
Supplementary Material online). In contrast, proteins formed
byfissionmainlyplayaroleinmetabolicandbiosynthesispro-
cesses(supplementaryfig.7,SupplementaryMaterialonline).
Proteins shaped by domain deletion are mainly related to ba-
sic functions such as the primary metabolism, and only a mi-
nor part of these proteins are stress–response related
(supplementary fig. 7, Supplementary Material online).
Our results provide further evidence that duplication im-
pacts rates of modular rearrangement (Buljan and Bateman
2009). We find that proteins affected by rearrangement
eventsareoverrepresentedin duplicatedgenes(supplemen-
tary table 6, Supplementary Material online). Furthermore,
we find indication that species with recent WGD have high-
er rates of fusion and fission in comparison to species with-
out recent WGD (supplementary table 7, Supplementary
Material online). In general, duplicates are thought to un-
dergooneofthreedifferentscenarios: subfunctionalization,
where the two duplicates share ancestral gene function;
neofunctionalization, where one copy retains the ancestral
function and the other copy diverges toward a novel func-
tion; and pseudogenization, where one copy is not ex-
pressed and is subsequently lost (Walsh 2003). One
explanation for sub- or neofunctionalization is the loss or
change of regulatory regions (Ganko et al. 2007). As the
conservation of noncoding sequences follows an exponen-
tial decay rate (Reineke et al. 2011), the retention of both
duplicates might be the result of the change of one of the
gene’s regulatory region under relaxed selectional con-
straints. The high retention rate of proteins that result from
a fusion event might be explained by the conservation of at
least one regulatory element in the upstream region,
whereas after fission, one arising protein may lose a regula-
tory region and undergo pseudogenization followed by
gene loss. A further reason for sub- and neofunctionaliza-
tion after duplication might be domain rearrangements in
one paralog or differential domain loss (Buljan et al. 2010).
We further illustrate the impact of protein domain rear-
rangements on an organism’s protein repertoire (fig. 5). The
emerging domains PAN_2 (emerged in the Tracheophyta)
and S_locus_glycop (Embryophyta) often co-occur together
with the B-lectin domain. Arrangements containing the two
emergeddomainsS_locus_glycopandPAN_2arefrequently
rearrangedwithinparalogousgenes(fig.5)andobtainacat-
alytic function through the addition of kinase domains. Pro-
teins that consist of arrangements with these two emerged
domains have GO functions related to the recognition of
pollen, protein phosphorylation, and cell recognition. Al-
though we observed fusion events in tandemly duplicated
genes in ourcase study, fusionevents are not generallyover-
represented in tandemly duplicated genes (supplementary
table 5, Supplementary Material online). After fusion, dupli-
cates might be difficult to recognize as paralogs. One might
therefore speculate that in tandemly duplicated proteins
fused arrangements are harder to detect. The increased
Kersting et al.
GBE
324
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012

Page 10

rates of events along more recent branches might be ex-
plained by WGD which have taken place in angiosperms
(De Bodt et al. 2005; Freeling et al. 2006; Shoemaker
et al. 2006; van de Peer et al. 2009; Paterson et al.
2010).Indeed,inapairwisecomparisonoffusionandfission
rates between plant pairs, which differ by one recent WGD,
we find increased rates in plants with more recent WGD
(supplementary table 7, Supplementary Material online).
Roughly one-third of all vascular plants have undergone
recent WGDs (Wood et al. 2009).
Arrangement Distribution
We investigated the distribution of shared arrangements
among the plant species. The majority of domain arrange-
ments are either species-specific or universal (fig. 3). This
bimodal distribution is even stronger when we consider only
a well-annotated subset of our species and exclude the
green algae (supplementary fig. 8, Supplementary Material
online). In particular, proteins with two or three domains are
often species-specific. In combination with the observation
that roughly 70% of all domain-containing proteins are
single-domain proteins (supplementary fig. 4, Supplemen-
tary Material online), this can lead to the assumption that
the fusion of single-domain proteins is a powerful mecha-
nism to obtain species-specific proteins with new function-
alities. This distribution suggests that only very few long
arrangements are highly conserved; long arrangements
are possibly more often affected by fission events. Proteins
with arrangements shared by several but not all species are
overrepresented in GO terms related to basic functions such
as primary metabolism, cellulose biosynthetic process, and cell
wall organization. In proteins with arrangements shared by
a subset of between 5 and 24 proteomes, innate_immune_r-
esponse is significantly overrepresented, suggesting that
there might be different pathogens affecting different sub-
clades.Proteins withGOterms relatedto reproduction, signal
transduction, and prefixed with response_to are overrepre-
sented in species-specific arrangements or those shared by
only few species. The high number of species-specific ar-
rangements observed here is in accordance with the observa-
tion that, within a set of five angiosperm species, around
20% of proteins do not align to an orthologous group (Pa-
terson et al. 2010). The high amount of species-specific ar-
rangements and genes might also be a consequence of
frequent duplication events followed by lineage-specific re-
tention (Paterson et al. 2010). This supports the hypothesis
that plants have many flexible genetic mechanisms to pro-
duce species-specific adaptation (Bomblies 2010).
Gain and Loss of Domains and Arrangements
We investigated gain and loss at the levels of domains and
domain arrangements by reconstructing the ancestral states
based on maximum parsimony. We observe that gain and
loss can frequently be found in all clades in plant evolution
atbothdomainandarrangementlevels.Thisisinagreement
with Buljan and Bateman (2009), who found an equal event
distribution after speciation and duplication within animals
and a high amount of change in arrangements after dupli-
cation events. As we here do not conduct a direct compar-
ison of paralogs, but instead compare presence/absence
patterns of domains and their arrangements across pro-
teomes, our results only support the notion that domain
gain and loss can be found along all branches and that both
have a significant correlation with each other and with
FIG. 5.—Example of two emergent domains at the Tracheophyta node (PAN_2) and Embryophyta node (S_locus_glycope). The evolution of
example arrangements over time is shown in five different species (Arabidopsis thaliana [AT], Oryza sativa [OS], Populus trichocarpa [PT], Ricinus
communis [RC], Vitis vinifera [VV]). The observable diversity in arrangements within this family is explainable by simple one-step events of fusion, fission,
terminal deletion, and domain gain.
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
325

Page 11

branch length (fig. 1A). Branches with an increased loss rate
have, on average, a higher domain gain rate. This high gain
andlossrate,branch-specificvariationandthelargenumber
of species-specific arrangements show the high variability
and flexibility with which single-step mechanisms can create
evolutionary novelties. One might speculate that the high
gain rate of arrangements in M. domestica (supplementary
table 4, Supplementary Material online) is caused by the re-
cent polyploidy event or hybridization as consequence of
domestication (Velasco et al. 2010). The large amount of
domain loss in P. dactylifera might also be the consequence
of low sequence coverage (Al-Dous et al. 2011). Differences
in gain and loss between the different branches might also
be a consequence of variation in generation time between
plants. Evidence from studies in Fugu and Tetraodon sug-
geststhatintronlossisincreasedinspecieswithshortergen-
eration time (Loh et al. 2008). Similar patterns have been
found in Arabidopsis and rice (Roy and Penny 2007).
Coverage
The average domain residue coverage is 50%. Protein cov-
erage varies strongly even between closely related species
(fig. 1C). Three plants belonging to the Fabaceae clade
are included in this study, G. max, L. japonica, and M. trun-
catula. Their branches split around 50–60 Ma (Reineke et al.
2011). Several events of WGD have been found within the
Fabaceae clade; all three species share a common WGD fol-
lowed by additional independent WGDs (Blanc and Wolfe
2004). These WGDs in connection with different retention
and pseudogenization rates might explain the variance in
coverage within this clade. It is also possible that a number
of plant-specific domains are still not yet described, as the
number of sequenced plant genomes is still considerably
lower than the currently available animal genomes. In the
Fabaceae family, for example, a unique conserved disor-
dered region has been described in sieve element occlusion
genes (Ruping et al. 2010; Ernst et al. 2011). Many of these
family-specific conserved functional sequencesmight be still
not covered by Pfam. It should also be considered that ge-
nome quality between the investigated genomes varies,
which might be the cause of differences in domain cover-
age; the most recently sequenced genomes exhibit low cov-
erage in comparison to longer established genomes such as
M. truncatula or O. sativa (fig. 1C).
Conclusions
The results presented here provide, from a phylogenomic
perspective,multipleinsightsintotheevolutionarydynamics
of modular rearrangements and the potential adaptive
benefits in plant genomes. Although around 70% of all pro-
teinsaresingle-domainproteinsandalargefractionofthese
are shared by many species, we observe a very high vola-
tility of novel domains and arrangements in general. Most
strikingly, the majority of all arrangements is species-specific
or restricted to a very small clade. Our phylogeny-based ap-
proachunravelsthatthemajorityofnovelarrangementscan
be explained by single-step events such as fusion, fission,
and terminal loss. Several events of accelerated activity of
rearrangements and domain emergence could be associ-
ated to the respective changes in stress adaptation and mor-
phogenesis. This is particularly pronounced for fusion in
regulatory proteins. We thus observe a dominant effect
of rearrangements on adaptation, which is partly driven
by the high volatility of novel domains.
Taken together, this study illustrates another layer of
complexity, which explains how modularity helps plants
to both create and exploit their abundant genetic material
in order to accomplish rapid adaptation in response to en-
vironmental challenges. We propose these results will fuel
furtherlarge-scaleexperiments.Recentexperimentsinfungi
using recombination of libraries of domains from signaling
proteins (Peisajovich et al. 2010) and the expansion of do-
main repeats in self-recognition molecules (Chevanne et al.
2010) have already highlighted the enormous evolutionary
potential of modularity in protein evolution. Along these
lines, experiments on plant adaptation should be more
explicitly geared at furthering our understanding of how
protein modularity facilitates rapid adaptation.
Supplementary Material
Supplementary figures 1–9 and tables 1–7 are available at
Genome Biology and Evolution online (http://www.gbe.
oxfordjournals.org/).
Acknowledgments
This work was supported by the Deutsche Forschungs
Gemeinschaft, grant BO 2455/4-1 and by the Human
Frontier Science Program (R6P0033/2006-C) to E.B.B.;
A.D.M. acknowledges support by the WWU-PAS stipend.
The authors would like to thank two anonymous reviewers
for critical reading and helpful comments on the manuscript.
Literature Cited
Adams MD, et al. 2000. The genome sequence of Drosophila
melanogaster. Science 287:2185–2195.
Akhunov ED, et al. 2003. The organization and rate of evolution of
wheat genomes are correlated with recombination rates along
chromosome arms. Genome Res. 13:753–763.
Al-Dous EK, et al. 2011. De novo genome sequencing and comparative
genomics of date palm (Phoenix dactylifera). Nat Biotechnol.
29:521–527.
Alexa A, Rahnenfu ¨hrer J, Lengauer T. 2006. Improved scoring of
functional groups from gene expression data by decorrelating GO
graph structure. Bioinformatics 22:1600–1607.
Altschul SF, et al. 1997. Gapped BLASTand PSI-BLAST: a new generation
of proteindatabase search
25:3389–3402.
programs. NucleicAcidsRes.
Kersting et al.
GBE
326
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012

Page 12

Amoutzias GD, Weiner J, Bornberg-Bauer E. 2005. Phylogenetic
profiling of protein interaction networks in eukaryotic transcription
factors reveals focal proteins being ancestral to hubs. Gene
347:247–253.
Anderson CL, Janßen T. 2009. Monocots. In: Hedges S, Kumar S,
editors. The timetree of life. New York: Oxford University Press. p.
203–212.
Apic G, Gough J, Teichmann SA. 2001. Domain combinations in
archaeal, eubacterial and eukaryotic proteomes. J Mol Biol.
310:311–325.
Argout X, et al. 2011. The genome of Theobroma cacao. Nat Genet.
43:101–108.
Ashburner M, Lewis S. 2002. On ontologies for biologists: the gene
ontology—untangling the web. Novartis Found Symp. 247:66–80;
discussion 80–90, 244–252.
Babushok D, Ostertag EE, Kazazian HH. 2007. Current topics in genome
evolution: molecular mechanisms of new gene formation. Cell Mol
Life Sci. 64:542–554.
Banks JA, et al. 2011. The Selaginella genome identifies genetic changes
associated with the evolution of vascular plants. Science 332:
960–963.
Baucom RS, Estill JC, Leebens-Mack J, Bennetzen JL. 2009. Natural
selection on gene function drives the evolution of LTR retrotrans-
poson families in the rice genome. Genome Res. 19:243–254.
Baucom RS, et al. 2009. Exceptional diversity, non-random distribution,
and rapid evolution of retroelements in the b73 maize genome.
PLoS Genet. 5:e1000732.
Bekaert M, Edger PP, Pires JC, Conant GC. 2011. Two-phase resolution
of polyploidy in the Arabidopsis metabolic network gives rise to
relative and absolute dosage constraints. Plant Cell 23:1719–1728.
Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate:
a practical and powerful approach to multiple testing. J R Stat Soc
Ser B Biol Sci (Methodological). 57:289–300.
Bennetzen JL. 2005. Transposable elements, gene creation and genome
rearrangement inflowering
15:621–627.
Bennici A. 2005. A fundamental plant evolutionary problem: the origin
of land-plant sporophyte; is a new hypothesis possible? Riv Biol.
98:469–480.
Bhattacharya D, Yoon HS, Hedges SB, Hackett JD. 2009. Eukaryotes
(eukaryota). In: Hedges S, Kumar S, editors. The timetree of life.
New York: Oxford University Press. p. 116–120.
Bjo ¨rklund AK, Ekman D, Light S, Frey-Sko ¨tt J, Elofsson A. 2005. Domain
rearrangements in protein evolution. J Mol Biol. 353:911–923.
Blanc G, Wolfe KH. 2004. Widespread paleopolyploidy in model plant
species inferred from age distributions of duplicate genes. Plant Cell
16:1667–1678.
Bomblies K. 2010. Doomed lovers: mechanisms of isolation and
incompatibility in plants. Annu Rev Plant Biol. 61:109–124.
Bremer B, et al. 2009. An update of the angiosperm phylogeny group
classification for the orders and families of flowering plants: APG III.
Bot J Linn Soc. 161:105–121.
Buljan M, Bateman A. 2009. The evolution of protein domain families.
Biochem Soc Trans. 37:751–755.
Buljan M, Frankish A, Bateman A. 2010. Quantifying the mechanisms of
domain gain in animal proteins. Genome Biol. 11:R74.
Cartwright P, Collins A. 2007. Fossils and phylogenies: integrating
multiple lines of evidence to investigate the origin of early major
metazoan lineages. Integr Comp Biol. 47:744–751.
Chan AP, et al. 2010. Draft genome sequence of the oilseed species
Ricinus communis. Nat Biotechnol. 28:951–956.
plants.CurrOpinGenetDev.
Chevanne D, Saupe SJ, Clave C, Paoletti M. 2010. WD-repeat instability
and diversication of the Podospora anserina hnwd non-self
recognition gene family. BMC Evol Biol. 10:134.
Choi HK, et al. 2004. Estimating genome conservation between crop
andmodellegumespecies.
101:15289–15294.
Cohen-Gihon I, et al. 2011. Evolution of domain promiscuity in
eukaryotic genomes—a perspective from the inferred ancestral
domain architectures. Mol Biosyst. 7:784–792.
De Bodt S, Maere S, van de Peer Y. 2005. Genome duplication and the
origin of angiosperms. Trends Ecol Evol. 20:591–597.
Dubchak I, et al. 2009. The Sorghum bicolor genome and the
diversification of grasses. Nature 457:551–556.
Dyson HJ, Wright PE. 2005. Intrinsically unstructured proteins and their
functions. Nat Rev Mol Cell Biol. 6:197–208.
Eddy SR. 2011. Accelerated profile HMM searches. PLoS Comp Biol.
7:e1002195.
Ekman D, Bjo ¨rklund AK, Elofsson A. 2007. Quantification of the
elevated rate of domain rearrangements in metazoa. J Mol Biol.
372:1337–1348.
Ekman D, Bjo ¨rklund AK, Frey-Sko ¨tt J, Elofsson A. 2005. Multi-domain
proteins in the three kingdoms of life: orphan domains and other
unassigned regions. J Mol Biol. 348:231–243.
Endress PK. 2001. Origins of flower morphology. J Exp Zool.
291:105–115.
Enright AJ, Ouzounis CA. 2001. Functional associations of proteins in
entire genomes by means of exhaustive detection of gene fusions.
Genome Biol. 2:RESEARCH0034.
Ernst AM, et al. 2011. The sieve element occlusion gene family in
dicotyledonous plants. Plant Signal Behav. 6:151–153.
Eucalyptus grandis Genome Project. 2010.
Available from: http://www.phytozome.net/eucalyptus.
Finn RD, et al. 2008. The Pfam protein families database. Nucleic Acids
Res. 38:D211–D222.
Fong JH, Geer LY, Panchenko AR, Bryant SH. 2007. Modeling the
evolution of protein domain architectures using maximum parsi-
mony. J Mol Biol. 366:307–315.
Forest F, Chase MW. 2009. Eurosid I. In: Hedges S, Kumar S, editors. The
timetree of life. New York: Oxford University Press. p. 188–196.
Forslund K, Henricson A, Hollich V, Sonnhammer ELL. 2007. Domain
tree based analysis of protein architecture evolution. Mol Biol Evol.
25:254–264.
Freeling M, Thomas BC, Freeling M, Thomas BC. 2006. Gene-balanced
duplications, like tetraploidy, provide predictable drive to increase
morphological complexity. Genome Res. 16:805–814.
Ganko EW, Meyers BC, Vision TJ. 2007. Divergence in expression between
duplicated genes in Arabidopsis. Mol Biol Evol. 24:2298–2309.
Go SA, et al. 2002. A draft sequence of the rice genome (Oryza sativa L.
ssp). Science 296:92–100.
Go ¨tz S, et al. 2008. High-throughput functional annotation and data
mining with the Blast2GO suite. Nucleic Acids Res. 36:3420–3435.
Grace SC, Logan BA. 2000. Energy dissipation and radical scavenging by
the plant phenylpropanoid pathway. Philos Trans R Soc Lond B Biol
Sci. 355:1499–1510.
Hanada K, Zou C, Lehti-Shiu MD, Shinozaki K, Shiu SH. 2008.
Importance of lineage-specific expansion of plant tandem duplicates
in the adaptive response to environmental stimuli. Plant Physiol.
148:993–1003.
Haploid Clementine Genome International Citrus Genome Consortium.
2011. [cited 2012 Jan 24]. Available from: http://www.phytozome.
net/clementine.
Proc NatlAcadSciUSA.
[cited 2012 Jan 24].
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
327

Page 13

Hedges SB, Dudley J, Kumar S. 2006. Timetree: a public knowledge-
base ofdivergencetimesamong
22:2971–2972.
Herron MD, Hackett JD, Aylward FO, Michod RE. 2009. Triassic origin
and early radiation of multicellular volvocine algae. Proc Natl Acad
Sci U S A. 106:3254–3258.
Hu TT, et al. 2011. The Arabidopsis lyrata genome sequence and the
basis of rapid genome size change. Nat Genet. 43:476–481.
Huang SW, et al. 2009. The genome of the cucumber, Cucumis sativus.
Nat Genet. 41:1275–1281.
Huerta-Cepas J, Dopazo J, Gabaldo ´n T. 2010. ETE: a python environ-
ment for tree exploration. BMC Bioinformatics 11:24.
International Peach Genome Initiative. 2010.
Available from: http://www.phytozome.net/peach.
Jaillon O, et al. 2007. The grapevine genome sequence suggests ancestral
hexaploidization in major angiosperm phyla. Nature 449:463–467.
Jiao Y, et al. 2011. Ancestral polyploidy in seed plants and angiosperms.
Nature 473:97–100.
Kawakita A, Kato M. 2009. Repeated independent evolution of obligate
pollination mutualism in the phyllantheae-epicephala association.
Proc Biol Sci. 276:417–426.
Kejnovsky E, Leitch IJ, Leitch AR. 2009. Contrasting evolutionary
dynamics between angiosperm and mammalian genomes. Trends
Ecol Evol. 24:572–582.
Kerchev PI, Fenton B, Foyer CH, Hancock RD. 2012. Plant responses to
insect herbivory: interactions between photosynthesis, reactive
oxygen species and hormonal signalling pathways. Plant Cell
Environ. 35:441–453.
Kersey P, et al. 2005. Integr8 and genome reviews: integrated views of
complete genomes and proteomes. Nucleic Acids Res. 33:D297–D302.
Kroymann J. 2011. Natural diversity and adaptation in plant secondary
metabolism. Curr Opin Plant Biol. 14:246–251.
Kummerfeld SK, Teichmann SA. 2005. Relative rates of gene fusion and
fission in multi-domain proteins. Trends Genet. 21:25–30.
Lang D, et al. 2010. Genome-wide phylogenetic comparative analysis of
plant transcriptional regulation: a timeline of loss, gain, expansion,
and correlation with complexity. Genome Biol Evol. 2:488–503.
Levitt M. 2009. Nature of the protein universe. Proc Natl Acad Sci U S A.
106:11079–11084.
Loh YH, Brenner S, Venkatesh B. 2008. Investigation of loss and gain of
introns in the compact genomes of pufferfishes (fugu and
tetraodon). Mol Biol Evol. 25:526–535.
Ma LJ, et al. 2009. Genomic analysis of the basal lineage fungus
Rhizopus oryzae reveals a whole-genome duplication. PLoS Genet.
5:e1000549.
Magallo ´n SA, Sanderson MJ. 2005. Angiosperm divergence times: the
effect of genes, codon positions, and time constraints. Evolution
59:1653–1670.
Mayrose I, et al. 2011. Recently formed polyploid plants diversify at
lower rates. Science 333:1257.
Merchant SS, et al. 2007. The Chlamydomonas genome reveals the
evolution of key animal and plant functions. Science 318:245–250.
Ming R, et al. 2008. The draft genome of the transgenic tropical fruit
tree papaya (Carica papaya linnaeus). Nature 452:991–996.
Misook H, Wen-Hsiung L, Chen CJ. 2007. External factors accelerate
expression divergence between duplicate genes. Trends Genet.
23:162–166.
Moore AD, Bjo ¨rklund AK, Ekman D, Bornberg-Bauer E, Elofsson A.
2008. Arrangements in the modular evolution of proteins. Trends
Biochem Sci. 33:444–451.
organisms.Bioinformatics
[cited 2012 Jan 24].
Moore AD, Bornberg-Bauer E. 2012. The dynamics and evolutionary
potential of domain loss and emergence. Mol Biol Evol. 29:
787–796.
Obradovic Z, Peng K, Vucetic S, Radivojac P, Dunker AK. 2005.
Exploiting heterogeneous sequence properties improves prediction
of protein disorder. Proteins 61(Suppl 7):176–182.
Olson JM. 1970. The evolution of photosynthesis. Science 168:438–446.
Olson JM. 2001. ‘Evolution of photosynthesis’ (1970), re-examined
thirty years later. Photosynth Res. 68:95–112.
Osborn TC, et al. 2003. Understanding mechanisms of novel gene
expression in polyploids. Trends Genet. 19:141–147.
Oshima A, et al. 2010. Asymmetric configurations and N-terminal
rearrangements in connexin26 gap junction channels. J Mol Biol.
405:724–735.
Palenik B, et al. 2007. The tiny eukaryote Ostreococcus provides
genomic insights into the paradox of plankton speciation. Proc Natl
Acad Sci U S A. 104:7705–7710.
Pasek S, Risler JL, Bre ´zellec P. 2006. Gene fusion/fission is a major
contributor to evolution of multi-domain bacterial proteins.
Bioinformatics 22:1418–1423.
Pateraki I, Kanellis AK. 2010. Stress and developmental responses of
terpenoid biosynthetic genes in Cistus creticus subsp. creticus. Plant
Cell Rep. 29:629–641.
Paterson AH, Freeling M, Tang H, Wang X. 2010. Insights from the
comparison of plant genome sequences. Annu Rev Plant Biol.
61:349–372.
Peisajovich SG, Garbarino JE, Wei P, Lim WA. 2010. Rapid diversification
of cell signaling phenotypes by modular domain recombination.
Science 328:368–372.
Proost S, et al. 2009. PLAZA: a comparative genomics resource to study
gene and genome evolution in plants. Plant Cell 21:3718–3731.
Reineke AR, Bornberg-Bauer E, Gu J. 2011. Evolutionary divergence and
limits of conserved non-coding sequence detection in plant
genomes. Nucleic Acids Res. 39:6029–6043.
Rensing SA, Lang D, Zimmer AD, et al. 2008. The Physcomitrella
genome reveals evolutionary insights into the conquest of land by
plants. Science 319:64–69.
Rogers RL, Hartl DL. 2012. Chimeric genes as a source of rapid evolution
in Drosophila melanogaster. Mol Biol Evol. 29:517–529.
Roy SW, Penny D. 2007. Patterns of intron loss and gain in plants: intron
loss-dominated evolution and genome-wide comparison of O. sativa
and A. thaliana. Mol Biol Evol. 24:171–181.
Ruping B, et al. 2010. Molecular and phylogenetic characterization of
the sieve element occlusion gene family in Fabaceae and non-
Fabaceae plants. BMC Plant Biol. 10:219.
Schmutz J, et al. 2010. Genome sequence of the palaeopolyploid
soybean. Nature 463:178–183.
Schnable PS, et al. 2009. The b73 maize genome: complexity, diversity,
and dynamics. Science 326:1112–1115.
Setaria italica Genome Sequencing Project. 2011. These sequence data
were produced by the US Department of Energy Joint Genome
Institute [Internet]. [cited 2012 Jan 24] Available from: http://
www.phytozome.net/foxtailmillet.
Shan H, et al. 2009. Evolution of plant MADS box transcription factors:
evidence for shifts in selection associated with early angiosperm
diversification and concerted gene duplications. Mol Biol Evol.
26:2229–2244.
Shoemaker RC, Schlueter J, Doyle JJ. 2006. Paleopolyploidy and gene
duplication in soybean and other legumes. Curr Opin Plant Biol.
9:104–109.
Kersting et al.
GBE
328
Genome Biol. Evol. 4(3):316–329.doi:10.1093/gbe/evs004Advance Access publication January 16, 2012

Page 14

Siegel S, Castellan N Jr. 1988. Nonparametric statistics for the behavioral
sciences, 2nd ed. Boston: McGraw-Hill Humanities.
So ¨ding J, Lupas AN. 2003. More than the sum of their parts: on the
evolution of proteins from peptides. Bioessays 25:837–846.
Soltis DE, Soltis PS, Zanis MJ. 2002. Phylogeny of seed plants based on
evidence from eight genes. Am J Bot. 89:1670–1681.
Srivastava M, et al. 2008. The Trichoplax genome and the nature of
placozoans. Nature 454:955–960.
Sweet Orange Genome Project. 2010. [cited 2012 Jan 24]. Available
from: www.phytozome.net/citrus.
Tang H, et al. 2008. Unraveling ancient hexaploidy through multiply-
aligned angiosperm gene maps. Genome Res. 18:1944–1954.
The Arabidopsis Initiative. 2000. Analysis of the genome sequence of
the flowering plant Arabidopsis thaliana. Nature 408:796–815.
R Development Core Team. 2008. R: a language and environment for
statistical computing. Vienna (Austria): R Foundation for Statistical
Computing.
Tuskan GA, et al. 2006. The genome of black cottonwood, Populus
trichcarpa (Torr. & Gray). Science 313:1596–1604.
van de Peer Y, Maere S, Meyer A. 2009. The evolutionary significance of
ancient genome duplications. Nat Rev Genet. 10:725–732.
Velasco R, et al. 2010. The genome of the domesticated apple (Malus ?
domestica Borkh.). Nat Genet. 42:833–839.
Veron AS, Kaufmann K, Bornberg-Bauer E. 2007. Evidence of
interaction network evolution by whole-genome duplications: a case
study in MADS-box proteins. Mol Biol Evol. 24:670–678.
Vogel JP, et al. 2010. Genome sequencing and analysis of the model
grass Brachypodium distachyon. Nature 463:763–768.
Walsh B. 2003. Population-genetic models of the fates of duplicate
genes. Genetica 118:279–294.
Wang M, Caetano-Anolles G. 2009. The evolutionary mechanics of
domain organization in proteomes and the rise of modularity in the
protein world. Structure 17:66–78.
Weiner J, Beaussart F, Bornberg-Bauer E. 2006. Domain deletions and
substitutionsin themodular
273:2037–2047.
Weiner J, Moore AD, Bornberg-Bauer E. 2008. Just how versatile are
domains? BMC Evol Biol. 8:285.
Williams JH. 2008. Novelties of the flowering plant pollen tube underlie
diversification of a key life history stage. Proc Natl Acad Sci U S A.
105:11259–11263.
Wood TE, et al. 2009. The frequency of polyploid speciation in vascular
plants. Proc Natl Acad Sci U S A. 106:13875–13879.
Worden AZ, et al. 2009. Green evolution and dynamic adaptations
revealed by genomes of the marine picoeukaryotes Micromonas.
Science 324:268–272.
Yang X, Jawdy S, Tschaplinski TJ, Tuskan GA. 2009. Genome-wide
identication of lineage-specific genes in Arabidopsis, Oryza and
Populus. Genomics 93:473–480.
Young ND, et al. 2005. Sequencing the genespaces of Medicago
truncatula and Lotus japonicus. Plant Physiol. 137:1174–1181.
Zheng C, et al. 2009. Gene loss under neighborhood selection following
whole genome duplication and the reconstruction of the ancestral
Populus genome. J Bioinform Comput Biol. 7:499–520.
Zmasek CM, Godzik A. 2011. Strong functional patterns in the
evolution of eukaryotic genomes revealed by the reconstruction of
ancestral protein domain repertoires. Genome Biol. 12:R4.
protein evolution. FEBSJ.
Associate editor: Yves van de Peer
Modular Evolution in Plants
GBE
Genome Biol. Evol. 4(3):316–329. doi:10.1093/gbe/evs004Advance Access publication January 16, 2012
329