Epigenetic regulation of TE activity
Although TEs occupy more than 50% of many eukaryotic genomes, the majority
are silenced via epigenetic mechanisms. The most well studied epigenetic marks include
DNA methylation, several types of histone modification and small interfering RNA
(siRNA) mediated silencing. DNA methylation involves the addition of methyl groups at
cytosine residues (in eukaryotes) and occurs in all sequence contexts (CG, CHG and
CHH) in plants. Subsequent to McClintock’s discovery that sequential breakage and
joining of chromosomes during maize development resulted in activation of TEs (class II
– Ac and Spm), research demonstrated that the reversible status of these elements was
associated with DNA methylation (McClintock B. 1958, McClintock B. 1965, Fedoroff
N, et al. 1995, Pan YB and Peterson PA 1988). Later studies in Arabidopsis showed that
several classes of TEs become hypomethylated and reactivated in lines that are mutant for
methyltransferase enzymes (Miura et al. 2001, Singer et al. 2001, Lippman et al. 2003 &
2004, Tsukahara et al. 2009). Recently, a striking example for epigenetic derepression of a
TE associated with a deleterious phenotype was found in oil palm (Ong-Abdullah et al.
2015). Hypomethylation of a Karma TE insertion within an intron of the gene
MANTLED provided an alternative splice site and a premature termination signal,
resulting in deformed oil palm fruits. Recent high-throughput silencing studies
demonstrate that TEs are enriched for DNA methylation in all three sequence contexts
(Lister et al. 2008, Zhang et al. 2006). In addition to DNA methylation, transposon
activity is often controlled by histone modifications (Bernatavichute et al. 2008). Various
modifications such as methylation, deacetylation, and biotinylation on histone
aminoterminal tails impact TE transcription. In plants, both H3K9me2 and H3K27me1
contribute to transcriptional silencing of some TEs in Arabidopsis and rice (Ebbs et al.
2005, Jacob et al. 2009&2010, Ding et al. 2007, Qin et al. 2010, Mirouze et al. 2009,
Zhang et al. 2003). In Arabidopsis, mutations in the histone deacetylase gene results in
transcriptional activation of several TE families (Lippman et al. 2003). Clearly, the
presence or absence of these histone modifications plays an important role in TE
silencing.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
In addition to these pre-transcriptional mechanisms, TE activity is controlled
posttranscriptionally via RNA silencing. Here, the host silencing system identifies RNA
molecules that were produced by transposons. TE transcripts are converted into
doublestranded RNAs (dsRNAs), which are further processed into small RNA (siRNA)
molecules that range from 21 to 35 nucleotides in length. Higher plants evolved specific
DNA-dependent RNA polymerases such as RNA polymerase IV and V to produce the
initial RNA transcripts for RNA silencing and siRNA induced methylation, respectively
(Ream et al. 2009). Therefore, being dynamic and potentially deleterious components of
the genome, evolutionary forces have produced various epigenetic ways to facilitate TE
silencing.
Stress induced TE activation
Given the diversity in TE content among closely related taxa, it seems apparent
that transposition occurs despite the presence of epigenetic silencing mechanisms. Indeed,
TEs can become activated under stressful conditions, including both genomic and
environmental stresses (Grandbastien M-A et al 1989, Pouteau S et al 1991).
Environmental stresses include both biotic (microbial infection, wounding etc.,) and
abiotic (dehydration, UV radiation) factors that can activate silenced TEs (Wessler 1996;
Kalendar et al. 2000; Grandbastien et al. 2005; Ramallo et al. 2008). New TE insertions
into coding regions are often deleterious; nevertheless, TE repression mechanisms are
often circumvented by biotic and abiotic stress factors (Mhiri et al. 1997; Grandbastien et
al. 1998; Takeda 1998; Pecinka 2010; Tittel-Elmer et al. 2010; Fujino et al. 2011; Cavrak
et al. 2014; Makarevitch et al. 2015; Finatto et al. 2015). Ac/Ds transposons in maize
(Steward et al. 2000), Tam 3 in Antirrhinum majus (Hashida et al. 2006), and mPing
DNA transposons in rice (Naito et al. 2006) are examples of TEs that are activated in
response to cold stress. A specific strain of rice, EG4 (cultivar Gimbozu), contains more
than 1,000 active mPing elements when subjected to cold or salt stress (Naito et al. 2006).
In Arabidopsis, a copia LTR-retrotransposon, ONSEN, is activated under heat stress (Ito
et al. 2016). It has since been discovered that ONSEN acquired a heat responsive element,
resulting in transcription and production of full-length extrachromosomal DNA copies
under higher temperatures (Cavrak et al. 2014). In natural populations, stress may
increase TE amplification, as has been shown in wild barley. Plants grown in drier regions
of “evolution canyon” in Israel showed three-fold variation in copy number of a specific
retrotransposon, BARE-1, compared to plants grown in other regions of the canyon
(Kalendar et al. 2000). This suggests that a stresssensitive regulatory sequence in the
BARE-1 promoter could have activated the retrotransposon in these plants compared to
unstressed plants. A recent report identified a mutational event associated with a TE
insertion that gave rise to industrial melanism in the English peppered moth (Van’t Hoff
et al. 2016). This study found an intronic insertion of a TE that enhances expression of the
cortex gene, and this over-expression underlies the adaptive coloration in these moths that
occurred during the industrial revolution.
In addition to these environmental stresses, TEs can become activated by genomic
stresses (Baack et al. 2005, Noor and Chang 2006). Hybridization and polyploidization,
examples of genomic shock in response to the reunion of two divergent genomes in a
single nucleus, frequently release TEs from their silenced state (McClintock 1984,
Ungerer, 2006, Shan et al 2005, Madlung et al 2005). Although the precise mechanism(s)
that induces TE mobility in hybrids and polyploids is unclear, it is obvious that these
phenomena lead to global changes in DNA methylation, resulting in transcriptionally
active TEs (Liu e t al. 2004). A striking example is that of three hybrid sunflower species
that independently arose from the same two parental species, and in which the hybrid
genomes are 50% larger (~1,130 Mb additional DNA, mostly TEs) than either of the
parental genomes (Ungerer et al. 2006). Similarly, studies in polyploid Spartina, wheat,
and other plant species have reported transcriptional activation of TEs upon
polyploidization due to reduced cytosine methylation (Parisod et al. 2009, Kashkush et al.
2002, Slotkin and Martinssen 2007, Parisod et al. 2009, Madlung et al. 2002, Salmon et
al. 2005, Xu et al. 2009, DeFraia and Slotkin 2014). It is speculated that such TE
reactivation in response to hybridization and polyploidy could be due to incompatible
suppression machinery between the two donor genomes, or that unknown mechanisms
are in place that reduce genomic methylation under general stress conditions (Lisch D,
2009). At any rate, it seems likely that epigenetic regulation plays a major role in TE
reactivation during hybridization and polyploidy.
Functional consequences of TE insertions/reactivation
TE mobilization can create myriad genetic changes including the creation of
alternative expression forms by providing genes with promoters or enhancers, causing
gene movement to novel chromosomal locations, and through TE domestication.
Regulatory mutations in particular are likely to arise via TE insertions near proteincoding
genes due to the presence of regulatory modules within the TE sequence itself (Bennetzen
2014). An example of this is the insertion of an intracisternal A particle (IAP) element in
the mouse agouti locus that resulted in mice with altered coat color, obesity and diabetes
(Morgan et al. 1999). A similar effect has also been shown for TE insertions in plants
belonging to TE families such as Mutator, Helitron, CACTA, and Harbinger (Jiang et al.
2004, Paterson et al. 2009, Vogel et al. 2010, Jin and Bennetzen 1994, Kashkush and
Khasdan 2007, Kashkush et al. 2003, Butelli et al. 2012). A genome-wide survey of TE-
altered gene expression revealed ~2,000 transposon insertional polymorphisms between
two rice cultivars (japonica Nipponbare and indica 93-11), of which 10% of these
polymorphisms were located in expressed genic regions and induced a series of genetic
differences between two rice subspecies, suggesting that these insertions are an important
source of intraspecific genetic variation (Huang et al 2008). TEs can also move protein
coding genes from one location to another within a genome, as has been discovered via
synteny studies. For example, several studies have shown that non-collinear genes are
often associated with TEs due to the acquisition of these gene fragments for use as “filler
sequences” during the double strand break (DSB) repair process (Wicker et al. 2010,
Baucom et al. 2009, Jiang et al. 2004, Morgante et al. 2005). TE sequences can also be
"domesticated" as exons of new chimeric genes by fusing with nearby coding sequences.
Daysleeper, FAR1, and FHY3 transcription factors are some of the genes involved in
Arabidopsis development, which were originally derived from the transposase gene of
class II elements (Bundock and Hooykaas 2005, Hudson et al. 2003, Lin RC, et al. 2007).
Therefore, although TE activity might affect genomic integrity, TE driven genome
variability may lead to adaptation upon which selection can act (Martin et al. 2009, Naito
et al. 2009, Lockton and Gaut 2010, Fernandez et al. 2010, Tenaillon et al.
2010, Linquist et al. 2013, Oliver et al.2013, Vitte et al. 2014).
Study Systems
For the projects in this dissertation, I have used study systems from the plant
families Poaceae and Malvaceae. Below is a general overview of each family. Specific
information on the species studied and their relevance to the research objective(s) is
provided in the respective chapters.
Grasses
The economic and ecological importance of grasses has paved the way for many
whole genome-sequencing efforts, and to date has resulted in completed genomes of
important grass species such as Oryza sativa (rice) (Matsumoto et al. 2005, Yu et al.
2002), Sorghum bicolor (sorghum) (Paterson et al. 2009), Zea mays (maize) (Schnable et
al. 2009), Brachypodium distachyon (purple false broom) (Vogel et al. 2010), Setaria
italica (foxtail millet) (Bennetzen et al. 2012, Zhang et al. 2012), and Hordeum vulgare
(barley) (Mascher et al. 2017). These sequenced taxa represent three subfamilies of
Poaceae and have enabled some of the most comprehensive comparative genomic
analyses possible, providing extensive knowledge of the organization and evolution of
grass genomes.
Grasses evolved from a common ancestor with a base chromosome number of five
(~90 mya) that then underwent a series of whole genome duplications, segmental
duplications, chromosome fusions and translocations to produce a 12-chromosome
intermediate common ancestor around 70 mya (Salse et al. 2008). Even though most
current day grass genomes are functionally diploid, all grasses are ancient polyploids that
have gone through a tremendous amount of chromosome fusions and rearrangements
resulting in chromosome number reductions. Hence, it is fascinating to study grass
genomes from an evolutionary perspective, because of the variation found in ploidy level,
genome size, chromosome number, and repetitive DNA content. For example, Tritricum
aestivum (Bread wheat) is a hexaploid (2n=42) with a genome size of ~17 Gb that is
>80% repetitive, Zea mays (Maize) is a diploid (2n=20) with a genome size of ~2.5 Gb
that is 85% repetitive, and diploid rice (2n=24) has a much smaller genome of 0.4 Gb in
which 35% is repetitive DNA (Figure 2). In the most comprehensive comparative studies
to date, it has been shown that gene content and order is highly conserved between
grasses, even after 50 million years of independent divergence (Dubcovsky et al. 2001,
Paterson et al. 2004; Bowers et al. 2003 & 2005); however, with the exception of a very
small amount of conserved non-coding sequences, most of the intergenic space is
relatively distinct, even between recently diverged species such as sorghum and maize.
From these studies it is clear that a considerable amount of genome variation in grasses
can be ascribed to repeat variability.
For Chapters 2 and 3, I have chosen two independently domesticated grasses,
maize (Zea mays) and sorghum (Sorghum bicolor) in addition to their close wild relatives
as study systems. Both genera belong to the tribe Andropogoneae and differ in
morphology, genome size, TE content, and ploidy level. In addition to the difference in
their TE proportion and content, the distribution of TEs within each genome varies. For
example, sorghum shows a strong separation in the distribution of genes and repeats
compared to the maize genome (Paterson et al. 2009, Schnable et al. 2009). Such
variation in lineage-specific TE content provides an excellent phylogenetic framework for
determining the impact of TEs on genome structure and function in plants.
Gossypium
The cotton genus, Gossypium L (Malvaceae) comprises about 50 species and is
distributed worldwide with several primary centers of diversity in the arid/semi-arid
tropics and subtropics (Fryxell 1979, Wendel 2009). Species-rich regions include
northwestern Australia, central and southern Mexico, the Horn of Africa and the southern
Arabian Peninsula. The genus is extraordinarily diverse and differentiated cytogenetically
into eight genome groups (A through G, and K) that differ in DNA content and
chromosome size but not in chromosome number (Endrizzi, Turcotte and Kohel 1985;
Stewart 1995). The genus contains three-fold variations in genome size: from 800 Mb
(1C) in the D-genome to ~2,500 Mb in the K genome. There are three major lineages of
diploid (2n=26) species corresponding to three continents: Australia (C, G, and K
genomes), the Americas (D genome), and Africa/Arabia (A, B, E and F genomes). Apart
from the diploids, this genus contains five recent tetraploids (2n=52) that emerged from a
single interspecific hybridization between the A and D genomes approximately 1-2
million years ago. These five tetraploids (AD) are: G. hirsutum, G.tomentosum, G.
barbendense, G. darwinii, and G. mustelinum. G. hirsutum and G. barabendense are
domesticated crops with G. hirsutum dominating the world’s textile industry for its
superior yield, whereas G. barbendense is known for its higher quality. Apart from these
two-tetraploid species, there are two diploid A genome species that are also cultivated for
textile: G. arboreum and G. herbaceum. All four-cultivated species were independently
domesticated for their specialized fibers and have their own unique history of
domestication, diversification and utilization. Given the wide-range of genome size
variation reflecting the copy numbers of repeat DNA sequences and a recent
polyploidization event that may have contributed to the improvement in cotton fiber yield
and quality, I have used Gossypium as a model system in Chapter 4 to understand the
evolutionary dynamics of repeat components in host genome evolution.
