MBE Advance Access originally published online on April 9, 2008
Molecular Biology and Evolution 2008 25(7):1415-1428; doi:10.1093/molbev/msn085
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
A Phylogenetic Analysis of Indel Dynamics in the Cotton Genus
* Department of Ecology, Evolution, and Organismal Biology, Iowa State University
Arizona Genomics Institute, University of Arizona
Plant Genome Mapping Laboratory, University of Georgia
E-mail: jfw{at}iastate.edu.
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
Genome size evolution is a dynamic process involving counterbalancing mechanisms whose actions vary across lineages and over time. Whereas the primary mechanism of expansion, transposable element (TE) amplification, has been widely documented, the evolutionary dynamics of genome contraction have been less thoroughly explored. To evaluate the relative impact and evolutionary stability of the mechanisms that affect genome size, we conducted a phylogenetic analysis of indel rates for 2 genomic regions in 4 Gossypium genomes: the 2 coresident genomes (AT and DT) of tetraploid cotton and its model diploid progenitors, Gossypium arboreum (A) and Gossypium raimondii (D). We determined the rates of sequence gain or loss along each branch, partitioned by mechanism, and how these changed during species divergence. In general, there has been a propensity toward growth of the diploid genomes and contraction in the polyploid. Most of the size difference between the diploid species occurred prior to polyploid divergence and was largely attributable to TE amplification in the A/AT genome. After separating from the true parents of the polyploid genomes, both diploid genomes experienced slower sequence gain than in the ancestor, due to fewer TE insertions in the A genome and a combination of increased deletions and decreased TE insertions in the D genome. Both genomes of the polyploid displayed increased rates of deletion and decreased rates of insertion, leading to a rate of near stasis in DT and overall contraction in AT resulting in polyploid genome contraction. As expected, TE insertions contributed significantly to the genome size differences; however, intrastrand homologous recombination, although rare, had the most significant impact on the rate of deletion. Small indel data for the diploids suggest the possibility of a bias as the smaller genomes add less or delete more sequence through small indels than do the larger genomes, whereas data for the polyploid suggest increased sequence turnover in general (both as small deletions and small insertions). Illegitimate recombination, although not demonstrated to be a dominant mechanism of genome size change, was biased in the polyploid toward deletions, which may provide a partial explanation of polyploid genomic downsizing.
Key Words: genome size evolution • genome evolution • cotton • illegitimate recombination • rates of genome size change
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
In recent years, there has been considerable interest in the evolutionary forces and mechanisms that underlie the extraordinary genome size variation observed within and among various groups of organisms. The primary mechanism of genome expansion, transposable element (TE) amplification, has been documented in broad surveys across angiosperm lineages (SanMiguel and Bennetzen 1998
; Vitte and Bennetzen 2006) and within genera or closely related species (Hill et al. 2005; Hawkins et al. 2006; Piegu et al. 2006; Petit et al. 2007); however, the evolutionary dynamics and primary mechanisms of genome size contraction have been less thoroughly explored. As mechanisms of deletion are more challenging to study, requiring orthologous sequence from closely related species, evidence for genome size contraction as a whole has been limited to phylogenetic inferences based on the placement of taxa having small genomes (Bennett and Leitch 1995; Leitch et al. 1998; Wendel et al. 2002; Bennett and Leitch 2005a). Analyses of deletional mechanisms thought to be most important, that is, intrastrand homologous recombination and illegitimate recombination, have produced conflicting results concerning their relative importance and whether either can affect genome size as dramatically as TE proliferation (Wicker et al. 2001, 2003; Devos et al. 2002; Ma et al. 2004; Vitte and Bennetzen 2006).
To evaluate the relative impact of mechanisms of genome size change and their evolutionary stability, we conducted a phylogenetic analysis of indels in 2 regions of the cotton (Gossypium) genome for which we had previously generated data for several species, representing approximately 95–150 kb, depending upon genome. By including orthologous sequence from the phylogenetic outgroup, Gossypioides kirkii, we were able to partition indels into insertions and deletions and study their relative rates, both overall and with respect to contributing mechanism. Gossypium is a 5–10 million year old (myo) genus whose genomes range nearly 3-fold in size, from 885 Mb in the New World diploids to over 2,570 Mb in the Australian diploids (Hendrix and Stewart 2005). Early in the history of the genus, D-genome diploids and A-genome diploids diverged, subsequently acquiring a 2-fold difference in genome size (fig. 1). These divergent genomes later became reunited with allopolyploid formation approximately 1–2 myo, leading to extant allopolyploid species that have a genome size that is slightly less than the sum of their model diploid progenitors (Hendrix and Stewart 2005). Gossypium as a genus diverged from its closest extant relative, G. kirkii, approximately 15 myo; the latter species has the smallest genome of the species studied here (590 Mb).
|
Here we present perhaps the first phylogenetic analysis of indel rates in plants, using 2 bacterial artificial chromosome (BAC)-sized genomic regions surrounding the genes cellulose synthase (CesA) (Grover et al. 2004
) and alcohol dehydrogenase A (AdhA) (Grover et al. 2007), by studying 2 diploid species representing the closest living ancestors of polyploid Gossypium, both of its genomes, and the outgroup G. kirkii. We focus on the mechanisms that gave rise to insertions and deletions and their evolutionary dynamics. Using this quintet of genomes, the direction and timing of each indel (insertion or deletion; pre- or postpolyploidization) were determined, and the rate and direction of overall genomic change for each branch were calculated. From this curated analysis among closely related species, we determined rates of sequence gain or loss along each branch and assessed rate change during species divergence. |
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
BAC Library Screening and BAC Selection
Gossypium arboreum, Gossypium raimondii, and G. kirkii BAC libraries were screened, as previously reported (Grover et al. 2004
), for clones containing the gene encoding cellulose synthase a1 (CesA) and several other predicted genes in the previously sequenced region (Grover et al. 2004). The same treatment was applied to the G. kirkii library with respect to the BAC containing the AdhA gene (Grover et al. 2007). The resulting positive clones for each marker were evaluated to facilitate selection of clones that provided maximum overlap with the previously generated BAC sequences. Polymerase chain reaction and sequencing were used to verify the presence of the desired markers on the selected BACs prior to shotgun sequencing.
Shotgun Sequencing, Assembly, and Analysis
BAC plasmid DNA was sheared at room temperature using a HydroShear (GeneMachines, Ann Arbor, MI) DNA shearing device at speed code 12 for 25 cycles. The resulting DNA fragments were end repaired using the "End-it" DNA end repair kit (Epicentre, Madison, WI) and subsequently separated on an agarose gel for size selection (range 2–6 Kb). These fragments were cloned into a pBluescript II KS+ vector (Strategene, La Jolla, CA) and sequenced with universal vector primers (T7 and T3) to an average depth of 8x. Each sequence was base called using the program Phred (Ewing and Green 1998; Ewing et al. 1998), and vector sequences were masked by CROSS_MATCH (Ewing and Green 1998; Ewing et al. 1998). Trimmed sequences were assembled by the program Phrap (Green 1999), and contigs were visualized and edited with CONSED (Gordon et al. 1998).
The newly sequenced and previously published (CesA gi: EU626442
[GenBank]
–44 & AY632359
[GenBank]
–60; AdhA gi: EU626441
[GenBank]
& EF457751
[GenBank]
–54) BACs were aligned using Multi-LAGAN (Brudno et al. 2003) with Arabidopsis thaliana-based repeat masking and with the input tree ([A AT] [D DT] Gk), where A and D refer to sequences from the diploids, AT and DT designate their counterparts in the allopolyploid Gossypium hirsutum, and Gk refers to the outgroup species. The resulting alignment was checked manually for errors using BIOEDIT (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). Predicted features from the previously sequenced genomes (Grover et al. 2004, 2007) were mapped onto the new alignment, and novel sequence (i.e., sequence unique to the newly sequenced genomes) was analyzed as previously described (Grover et al. 2004, 2007).
Gap Polarization and Analysis
Indels were partitioned into insertions or deletions and phylogenetically placed using outgroup polarization. Thus, where sequence for G. kirkii existed, if 2 genomes (A/AT or D/DT) shared sequence with the outgroup, then the gap was considered a deletion that arose along the branch leading to the other 2 genomes during divergence of the A and D genome groups; an insertion during diploid divergence, prior to polyploid formation, was called when the outgroup shared a gap with either A/AT or D/DT. If 3 genomes shared sequence or a gap with the outgroup, then a deletion or insertion, respectively, was inferred to have occurred in the remaining genome after polyploid formation. If only 1 genome shared sequence or a gap with the outgroup and the other 3 existed in the opposite state, that event was labeled as unknown because 2 separate events occurring in the outgroup and the genome sharing its state are equally parsimonious as the opposite occurring separately in 1 genome postpolyploidization and in the prepolyploidization lineage for the other 2 genomes. For regions where the outgroup lacked homologous sequence, only those gaps arising because polyploidization could be polarized. In this case, if 3 of the genomes share sequence whereas the fourth has a gap, the shared sequence is assumed to be plesiomorphic and the gap is characterized as a deletion; likewise, if 3 genomes share a gap whereas the fourth has intervening sequence, an insertion is inferred. Indels that occurred between A/AT and D/DT in regions without outgroup sequence were not polarized.
Number of nucleotides (nt) added, deleted, or missing were standardized to nucleotides per year based on previously estimated organismal divergence times for the diploid divergence (6.8 my since A–D divergence [Cronn et al. 2002; Wendel and Cronn 2003]) and polyploid formation, as estimated based on multiple nuclear gene sequences (Senchina et al. 2003). Because modern A genome diploids are a closer model of the actual A-genome donor to the polyploids than are D-genome species, by about 50%, the divergence of the extant diploid species from the polyploid ancestor genomes was calculated. Specifically, branch lengths were apportioned based on the ratio of the diploid branch length over the total branch length (Senchina et al. 2003) to the calculated time since divergence over 6.8 my (time since A–D divergence; fig. 1). These values were used to estimate rates of indel evolution along each branch of the phylogeny.
Mechanisms responsible for indel formation were hypothesized based on the sequence within and surrounding alignment gaps. Inserted sequence deemed by sequence homology to be TE in origin was considered the result of TE amplification; deletion via intrastrand homologous recombination was assumed when one genome shared a single long terminal repeat (LTR) in an orthologous position with 1 or more other genomes but no internal sequence or second LTR. Single nucleotide gaps were classified into a category of the same name. Illegitimate recombination represents a group of mechanisms that often, but not always, are associated with short, direct repeats. Several molecular mechanisms are encompassed by illegitimate recombination, including double-stranded break repair and slipstrand mispairing. For the purpose of this study, illegitimate recombination was subdivided into 3 categories based on the hallmarks of the associated gap: 1) illegitimate recombination via double-stranded break repair, 2) illegitimate recombination via double-stranded break repair or slipstrand mispairing, and 3) illegitimate recombination via slipstrand mispairing. Double-stranded break repair was inferred when the gap was flanked by short (<15 nt), direct repeats; slipstrand mispairing was inferred when the sequence in the gap was directly repeated; and, in cases where a gap met both criteria, it was placed in a separate category. Indels that could not be assigned to a group based upon sequence inspection were assigned to a category referred to as "unknown mechanisms." This category is composed of a combination of indels generated by known insertional and deletional mechanisms that have either not left identifiable and characteristic hallmarks or whose hallmarks have been wiped away by subsequent evolution, as well as mechanisms whose hallmarks have yet to be described. This category more than likely contains a number of double-stranded break repair events that used only 1 nt for repair. These were placed in the unknown category specifically because it could not be determined whether the 1 nt repeat was an indel footprint or arose by chance.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
General Description of the Sequenced Regions
A summary of the results for both regions for all genomes is in tables 1 and 2. The AdhA region includes approximately twice as much aligned sequence in the A genomes as the D genomes (94 kb in A and 89 kb in AT vs. 52 kb in D and 46 kb in DT), largely from multiple TE insertions in A/AT during its divergence from D/DT (see below and table 1). Both genomes of the tetraploid are represented by less sequence than their diploid counterparts due to a combination of net growth in both of the diploids, as well as diminished growth in DT and net loss in AT (table 1). The nearly twice as large sequence for the A and AT genomes, as well as the overall contraction of the polyploid genome, is congruent with expectations due to genome size (A = 1,697 Mbp; D = 885 Mbp; AD = 2,401 Mbp, whereas A + D = 2,582 Mbp).
|
|
As detailed in Grover et al. (2004)
, the CesA region in Gossypium is rather different from the AdhA region in that the representative sequence for the A genomes is only slightly larger than that of the D genomes (57 kb in A and 51 kb in AT vs. 49 kb in both D and DT). The length of this region in the AT genome was, as expected, slightly shorter than the region in the A genome; however, this was not true for D and DT, whose lengths were nearly identical. This region had far fewer TEs than AdhA, marked only by a single TE insertion that arose in A/AT during its divergence from D/DT (table 1).
Indel Dynamics during the Evolution of A/AT
The period of evolution between the divergence of A–D diploids and the divergence of the diploid A genome from the ancestor to the polyploid AT is marked by several TE insertions, particularly in the AdhA region, that dramatically skew the nearly 1:1 ratio of deletions:insertions toward insertions (table 1). The rate of deletion for the 2 regions combined ranged from 1.02 x 10–9 to 1.09 x 10–9 nucleotides per year, virtually identical to the non-TE insertion rate of 9.9 x 10–10 to 1.08 x 10–9 nucleotides per year. Thus, the bulk of the size change during the 5.2 my post A–D divergence and pre A–AT divergence was due to TE insertions in the AdhA region, leading to an overall gain of 6.38 x 10–8 to 6.66 x 10–8 nucleotides per year. Gain via TE insertions (98.4% of sequence added; table 2) was distantly followed by illegitimate recombination, which contributed 1.5% of the total sequence added, although 92.9% of the non-TE insertion total. With respect to deletions, illegitimate recombination was second to the unknown mechanism category (the category describing indels that did not have characteristic hallmarks of known mechanisms; see Methods) in terms of relative importance (26.9% vs. 69.1% of sequence removed, respectively), removing approximately less than half the amount of sequence in comparison to the unknown category and only about one-fourth of the total sequence removed. These trends in mechanistic preference were common to the AdhA and CesA regions.
The 2 analyzed regions were similar in number of deletions and insertions; however, the AdhA region experienced slightly more deletions than did CesA (table 1) and the average sizes of deletions and insertions (non-TE) were larger, which led to a slight bias toward deletions in the AdhA region. The rate of insertion and deletion was slightly higher for the AdhA region than for CesA; however, this contribution to sequence turnover was minimal in comparison to the increased TE activity.
Indel Dynamics during the Evolution of D/DT
The period of evolution between the divergence of the A–D diploids and the diploid D from the polyploid DT is marked by 2 TE insertions (1 each in AdhA and CesA) and 1 large (3.1 kb) illegitimate recombination–associated insertion in AdhA (table 2). These 3 insertions alone comprise over 97% of the sequence added to the regions (ca. 13.4 kb out of 13.7 kb total), and shifts what would be a 1.05:1 deletion:insertion ratio to 0.3:1, leading to a net gain of 2.94 x 10–8 to 3.18 x 10–8 nucleotides per year. Aside from the 2 TEs, insertion via illegitimate recombination had the greatest impact, contributing 23.7%, to the total inserted. DNA removal via unknown mechanisms had the largest deletional impact on the region (87.6%), followed by illegitimate recombination at 9.7%.
As with the A/AT genome, the AdhA and CesA displayed a similar number of insertions (table 1); however, unlike the A/AT genome, deletions were more frequent (over 2 times) in the CesA region than in the AdhA region. Whereas the number of insertions for AdhA and CesA were similar, the amount of sequence inserted was nearly 2 times greater in CesA. These 2 factors led to a greater rate of non-TE sequence turnover in the CesA region. The deletion mechanism that had the greatest impact (unknown category) was common to both regions; however, the insertion mechanism having the greatest impact varied across the regions (TE insertion in AdhA and unknown in CesA).
Indel Dynamics during the Evolution of A Alone
The period of evolution in the A genome after its divergence from the polyploid AT genome saw a reduction in average rate of deletion compared with the prior 5.2 my (from 1.02 x 10–9 to 1.09 x 10–9 nucleotides per year to 5.4 x 10–10 nucleotides per year; table 1), whereas the average non-TE insertion rate rose from 9.9 x 10–10 to 1.08 x 10–9 nucleotides per year to 8.71 x 10–9 nucleotides per year. No TEs were polarized in the aligned region; thus, the insertion rates derived for this genome were purely from non-TE mechanisms. As on the previous branch (A/AT), the deletion and insertion rates varied between the 2 loci; however, the amount of variation dramatically increased after the divergence of A from AT. Whereas the deletion rates between AdhA and CesA in A/AT varied slightly over 2-fold and the non-TE insertion rates 1.2-fold, the deletion rates in A varied nearly 35-fold and the insertion rates over 100-fold. The combined deletion and insertion rates yield a net gain of over 8.18 x 10–9 nucleotides per year. IR played a large role in this region as it was the largest contributor to deletions overall (table 2; 49.8% vs. 39.6% for unknown mechanisms, the second largest) and contributed nearly all inserted DNA (99.6%). This was true for both deletions and insertions in both regions with the exception of deletions in AdhA, which consisted solely of 4 single nucleotide deletions.
The 2 sequenced regions differed both in rate of sequence turnover, as well as direction of change. The AdhA region experienced far greater sequence gain than CesA without much loss, leading to a net gain of 1.72 x 10–8 nucleotides per year (table 1). The CesA region, in contrast, experienced more loss than gain, leading to a net loss of 8.8 x 10–10 nucleotides per year, a rate attributed to a small amount of gain outweighed by a nearly as small amount of loss.
Indel Dynamics during the Evolution of AT Alone
The period of evolution in the AT genome after its divergence from the diploid A genome saw a reversal from net gain (6.38 x 10–8 to 6.66 x 10–8 nucleotides per year) to net loss (–1.01 x 10–8 nucleotides per year; table 1), a reversal that was mirrored in both the AdhA and CesA regions with each experiencing sequence loss. When TE insertions are excluded from both branches as they are episodic in nature and may not have had as much opportunity to affect the AT genome in the last 1.6 my as the A/AT branch over the previous 5.2 my, the AT genome shows an even more dramatic bias toward DNA removal (adjusted loss of 2.68 x 10–8 nucleotides per year in AT vs. adjusted loss of 1 x 10–11 to 3 x 10–11 nucleotides per year in A/AT). TEs contributed the most to sequence gain and loss (via intrastrand homologous recombination; table 2) in this genome, despite their action being limited to the AdhA region.
As on the previous branch, the insertion and deletion rates varied between the 2 loci; however, the amount of variation dramatically increased after the divergence of AT from A. Whereas the deletion rates between AdhA and CesA for A/AT varied slightly over 2-fold and the non-TE insertion rates 1.3-fold, the deletion rates in AT varied nearly 19-fold (more deletions in AdhA) and the insertion rate varied 92-fold (more insertions in CesA). The major mechanisms contributing to sequence loss and gain in the AdhA region were removal and insertion of TEs, as noted above and in table 2; the CesA was not subject to either TE loss or gain, thus the largest contributor to sequence change in this region was non-TE in nature (illegitimate recombination for both loss and gain). As in the A genome, the AdhA region experienced far greater sequence turnover, in terms of nucleotides deleted, due to the large actions of the TE deletion and insertion.
Overall Indel Dynamics during the Evolution of D Alone
The period of evolution in the D genome after its divergence from the polyploid DT genome saw an increase in average rate of deletion and non-TE insertion compared with the prior 4.2 my (from 2.52 x 10–9 to 2.55 x 10–9 nucleotides per year to 8.77 x 10–9 nucleotides per year for deletions and from 8.22 x 10–10 to 8.37 x 10–10 to 1.94 x 10–8 for insertions; table 1), as well as a decrease in the TE insertion rates (by 1.54 x 10–8 to 1.69 x 10–8 nucleotides per year). Insertion and deletion rates varied between the 2 regions, with the CesA region experiencing 34-fold more nucleotides deleted but less than half the amount of non-TE insertions. This difference in insertion and deletion rates led to a net gain in AdhA and a net loss in CesA. Combined, the regions experienced a slightly smaller net gain than experienced in D/DT (1.92 x 10–8 nucleotides per year). This rate was largely attributed to the combined effects of a single TE insertion (30.9%; table 2) and 2 large illegitimate recombination–associated insertions (68.7% together) but was slightly relieved by the longer and more frequent deletions (primarily of unknown mechanism) in the CesA region. Overall, deletion via unknown mechanisms (88.7%) and insertion via illegitimate recombination (68.9%) had the greatest effects, an observation common to both regions for insertions but not deletions (deletion via illegitimate recombination had the most impact in AdhA).
Indel Dynamics during the Evolution of DT Alone
The period of evolution in the DT genome since divergence from the D genome saw a slight decrease in the rate of deletion (1.05 x 10–9 nucleotides per year in DT vs. 2.52 x 10–9 to 2.55 x 10–9 nucleotides per year in D/DT; table 1) and a substantial reduction of the insertion rate (by nearly 90%). The 2 regions displayed opposite changes in rates, with AdhA experiencing loss and CesA experiencing gain leading, to an overall rate of sequence gain equivalent to 1.55 x 10–9 nucleotides per year (down from 2.94 x 10–8 to 3.18 x 10–8 nucleotides per year in the ancestor, D/DT). Neither region was affected by TE proliferation, and thus the difference in direction of genome size change reflects other indel dynamics. In general, this genome experienced less turnover than the ancestral D/DT and the D genome, with the exception of slightly more deletions in the AdhA region than experienced by the other 2 genomes. Illegitimate recombination and the unknown mechanism category impacted the genome approximately equally with respect to deletions (table 2), whereas the latter was the major contributor to insertions in those regions (91%). These regions experienced mechanistic biases in this genome as well, with the unknown mechanisms contributing the most sequence loss to AdhA and the most sequence gain to CesA and single nucleotide insertions and deletion via illegitimate recombination having the greatest impact in AdhA and CesA, respectively.
Unpolarized Indels
The number of unpolarized gaps (between A/AT and D/DT, A and AT, and D and DT) ranged in number and size across the genomes and served to expand the range in possible rates for these genomes (table 1). One hundred and fifty two gaps were unpolarized between A/AT and D/DT, accounting for 1.78 x 10–9 to 2.00 x 10–9 nucleotides per year and 2.63 x 10–8 to 2.84 x 10–8 nucleotides per year missing from A/AT and D/DT, respectively. This increases the range in the rate of overall genome size expansion in A/AT from 6.38 x 10–8 to 6.66 x 10–8 nucleotides per year to 6.20 x 10–8 to 9.50 x 10–8 nucleotides per year, a range shift that is mostly toward further expansion. Similarly, the range in the rate of genome size expansion for the D/DT branch increased from 2.94 x 10–8 to 3.18 x 10–8 nucleotides per year to 1.00 x 10–9 to 3.38 x 10–8 nucleotides per year, a range shift which suggests that the rate of growth due to polarized indels is likely an overestimate for this branch. The unpolarized gaps between A and AT represent more missing sequence in A than AT (5.83 x 10–9 vs. 4.91 x 10–9 nucleotides per year); however, even taking this into consideration, the range in overall rate of sequence change remains positive in A gain (2.35 x 10–9 to 1.31 x 10–8 nucleotides per year) and negative in AT (loss of 4.25 x 10–9 to 1.50 x 10–8 nucleotides per year). The unpolarized gaps between D and DT represent slightly more sequence missing in D than in DT (3.4 x 10–10 vs. 2.1 x 10–10 nucleotides per year) and created a small range in rates for each (1.89 x 10–8 to 1.94 x 10–8 nucleotides per year for D and 1.30 x 10–9 to 1.85 x 10–9 nucleotides per year for DT).
Analysis of Indels <400 nt
Previously, we reported that deletions in the AdhA region were consistent in size and frequency with the expectations of small indel bias and genome size (i.e., more and longer deletions in the smaller D genome [Grover et al. 2007]). Furthermore, we noted a higher rate of deletion in the polyploid compared with the diploid ancestors, noting that this observation is congruent with the idea of nonadditivity of polyploid genome sizes relative to their diploid antecedents. By adding sequence from the outgroup, we are now able to evaluate, and for a much larger data set, the rate of small indel formation to include the longer period prior to polyploid formation. Contrary to expectations, deletions were more than twice as frequent in A/AT than in D/DT (table 3); however, in accordance with expectations, the deletions were over 1.5-fold larger in the smaller D/DT genome (table 4). Small insertions in A/AT versus D/DT were congruent with the expectations of a small indel bias, in that they were more frequent and larger in the larger A/AT genome.
|
|
We also previously reported that the spectra of small indel sizes, unpolarized, were nearly equivalent with respect to size and frequency between the AT and DT genomes in the CesA region (Grover et al. 2004
). Upon polarization, a slight bias with regard to frequency appeared for both deletions and insertions (deletions 1.1-fold more frequent in AT; insertions 1.1-fold more frequent in DT; table 3); however, the amount of sequence affected was more variable, particularly for deletions (2.6-fold more sequence deleted in AT per my). The diploid A and D genomes displayed a similar pattern for small insertions (more in the D genome but smaller in size); however, small deletions were fully consistent with a small indel bias, with more frequent and larger deletions in the D genome. The average deletion and insertion sizes among the genomes (table 4) did not mirror the results of AdhA. The A genome deletions were on average smaller than those from the D genome (6.93 vs. 33.47 nt), as expected based on the previously analyzed AdhA region. Deletions in the polyploid, however, did not mirror its diploid counterparts, but instead was characterized by an acceleration in deletions in the AT genome and a deceleration in DT. The pattern between A/AT and D/DT, in the case of the CesA region, more closely resembled what the small indel bias would predict in terms of average deletion size and frequency (deletions in D/DT 1.5-fold as frequent and twice the size of those in A/AT). Conversely, the pattern of insertions is contrary to what the small indel bias would predict, with insertions in A/AT being 1.1 times as frequent as and more than 50% smaller than in D/DT.
Analysis of Iillegitimate Recombination
Previously, we reported that illegitimate recombination may be a key player in Gossypium genome size evolution, particularly in the polyploid. The data reported here (table 2) provide the ability to assess not only the rate of illegitimate recombination for each genome since polyploid formation but also how those rates compare with the ancestral rates. The combined data suggest that every lineage (A, AT, D, and DT) has experienced accelerations (to varying degrees) in the rate of deletion via illegitimate recombination since the diploid–polyploid divergence, from the near doubling in D to the over 4.5-fold increase in AT; however, this increase was not for both regions in every genome. Whereas the D and DT genomes displayed an increase in illegitimate recombination for both regions, both A and AT had a slight rate decrease in AdhA that was compensated for by the increased rate in CesA. It may be tempting to attribute the observed increase in illegitimate recombination in the polyploid to the smaller temporal distance between A–AT and D–DT (vs. A–D divergence); however, due to the extraordinary conservation of sequence flanking the indels that occurred in the A/AT and D/DT ancestors between the 4 extant genomes, we believe the impact of temporal distance is, in this case, minimal.
The rates attributable to insertion via illegitimate recombination display opposite effects in the diploids and polyploids (table 2), whereas the diploids experienced an overall increase in the rate of IR-associated insertions, both genomes of the polyploid experienced decreases. Again, this overall trend was not equivalent in the 2 regions. Whereas the AdhA region mirrored the overall results, in the CesA region the A genome experienced a decrease in rate, whereas the DT genome experienced an increase. Overall, the amount of sequence deleted via illegitimate recombination was less than the total amount of sequence inserted for the diploid lineages (added 1.04 x 10–9 to 1.84 x 10–8 nucleotides per year) and more for the polyploid lineages (deleted –5.5 x 10–10 and –2.0 x 10–10 nucleotides per year for AT and DT, respectively).
The relative impact of illegitimate recombination varied by genome and by region (table 2). In the AdhA region, most polarized deletions were attributed to the unknown mechanism category (A/AT: 73.7%; D/DT: 50.6%; DT: 67.7%), single nucleotide deletions (A: 100%), or LTR recombination (AT: 99.7%). Only in the D genome were most deletions attributed to IR (76.3%). In the CesA region, the most deleted nucleotides were attributed to illegitimate recombination in some genomes (A: 51%; AT: 88.9%; DT: 53%), whereas along other branches the unknown mechanism was dominant (A/AT: 58.4%; D/DT: 90.5%; D: 90.8%). When viewed together, the A genome was the only one where more nucleotides were deleted via IR than from any other mechanism (52.5%).
The relative impact of insertion via illegitimate recombination was small in comparison to the amount of sequence inserted by TEs in the A/AT, D/DT, and AT genomes (98%, 74.2%, and 97.4%, respectively) and unknown mechanisms in DT (93.9%). Illegitimate recombination only had a major impact on insertions in the diploid A and D genomes (at 99.6% and 68.9%, respectively). This trend was largely similar between both regions, with the only difference occurring in the CesA region of AT genome, where illegitimate recombination represented 86.7% of the nucleotides inserted, a difference attributed to the lack of TE insertions in this region.
|
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
Genome size evolution is a dynamic process, reflecting the net effects of counterbalancing mechanisms whose actions vary across a genomic landscape, across lineages, and over time. The potential for the primary mechanism of genome size change, TE proliferation, to effect genome size has become evident, although its catalysts are less clear. Mechanisms of deletion, by their nature, are more difficult to study; whereas TE proliferation can be gauged by simply evaluating the extent of TE sequence in a genome, deletional mechanisms can only be identified and evaluated by comparison to nondeleted sequence. Compounding this problem is rapid evolution, which may quickly erase by superimposed mutations the hallmarks of deletional mechanisms that leave small footprints, such as illegitimate recombination. Comparisons of long, orthologous tracts of sequence between closely related species that are polarized by an outgroup provides a potentially powerful means to evaluate the relative effects of different mechanisms influencing genome size change (both growth and reduction).
Rate of Sequence Loss and Gain on 6 Branches of the Gossypium Phylogeny
The combined rates of DNA deletion and insertion are ultimately what determine genome size change. Comparisons of extant genome sizes and their TE contents provide important information on the probable direction and nature of genome size change, but a phylogenetic perspective adds insights into the tempo, details, and dynamics of genomic divergence. One might imagine 2 species with the same genome size and TE composition that have different genomic histories; for example, 1 species may have acquired its genome size through slow and steady TE accumulation, whereas the other taxon has achieved a similar genome size via rapid flux of intergenic space (nearly concurrent insertions and deletions). Comparative genomic sequencing of closely related species, as exemplified here, provides the opportunity to illuminate this history and similar nuances of genome evolution.
The general trend unveiled by comparative sequencing of the AdhA and CesA regions is that there has been an overall propensity toward growth of the diploid genomes and an overall contraction of the polyploid (fig. 1; table 1), whereas each of these 2 regions display heterogeneous rates of sequence gain and loss at different times in the evolutionary histories of the genomes studied, possibly linked to genomically regional properties. All genomes experienced growth except for the AT genome; however, the AdhA region experienced contraction in the diploid D (in addition to AT) and the CesA region displayed contraction for the A and D genomes in addition to the AT genome. In addition, there appears to be a regional bias dependent on lineage and, possibly, ploidy level. For the 4 purely diploid branches (A/AT, D/DT, A, and D), the AdhA region experienced more gain than did the CesA region (A/AT and D/DT) or, in the case of A and D, gain versus loss. The converse was seen for the polyploid lineages, where both AT and DT experienced loss, or more loss, in AdhA compared with CesA and the diploid branches experienced gain.
In general, the A/AT branch is marked by large sequence gains, primarily TE in origin (tables 1 and 2). The rate of deletion barely outweighs the rate of non-TE insertion, indicating that genome growth along this branch has primarily been due to the action of TEs. The AdhA region of the A/AT genome gained sequence at a rate 4.5-fold higher than did the CesA region, due to TE proliferation and possibly indicating insertional preferences or exclusion. The D/DT branch also experienced most of its sequence gain related to TE insertions; however, the rate of deletions better outweighed the non-TE insertions in this genome. The combined deletion rate in D/DT was 2.5 times the rate of A/AT, consistent with a hypothesis that small genomes differ from large genomes in part due to their inherently higher deletion rates. We note, however, that the rate of deletion was still only about 1/13th the total rate of insertion (vs. 1/64rd in A/AT). These results indicate that the trend for the majority of the genome size divergence between the A and D genome species, having taken place on the A/AT and D/DT branches, is one of genome growth, with the rate in A/AT 2-fold higher than in D/DT and both being largely dependent upon the rate of insertion.
In the time after divergence from the polyploid AT, the rate of sequence gain experienced in the diploid A was less than 13% of that experienced prior to diploid–polyploid divergence (fig. 1; table 1), primarily due to the lack of TE insertions. The overall rate of sequence gain in A, however, still outweighed that of deletion due to the higher number of insertions and fewer deletions found in AdhA. The situation for the D genomes is far less exaggerated in this respect; in the time after divergence from the polyploid DT, the rate of gain in the diploid D decreased to about 65% of the ancestral rate in D/DT. This rate reflects a combination of increased deletions and the combination of a slight increase in non-TE insertions and decreased TE insertions. It is important to note that the individual rate changes that ultimately led to the attenuation of sequence gain in the diploid species were more than likely not instantaneous and not necessarily a consequence of divergence from the parents of the polyploid species; rather, it is more probable that the calculated rates reflect changes that may have been occurring in the diploid ancestors and continued through the point at which the diploid and polyploid parent genomes diverged or that they represent changes initiated in the diploids themselves. The potential causes of these rate changes in the diploid lineages are many and varied and warrant further exploration.
The impetus to change rates of indel evolution and, consequently, genome size can come from many and varied sources, one of which being the union of 2 divergent genomes in an allopolyploid nucleus. Polyploidization has been implicated in numerous genetic and genomic changes (reviewed in [Adams and Wendel 2005; Chen and Ni 2006]), and the resulting genome size of the polyploid species is often less than the sum of the 2 parental genomes (Soltis DE and Soltis PS 1999; Ozkan et al. 2003; Bennett and Leitch 2005b). This phenomenon of "genomic downsizing" has been explored in several other cases (Chantret et al. 2005; Gu et al. 2006), but to our knowledge this is the first phylogenetically informed evaluation of changes in deletion and insertion rate that accompany polyploidization using large contiguous tracts of orthologous and homoeologous sequences. Both genomes of the polyploid show an increase in the rate of deletion (more dramatic in the case of AT) and a reduction in the rate of insertion when compared with their ancestral lineages (fig. 1; table 1). The shifting balance from insertions to deletions produced a rate of near stasis in DT and an overall rate of contraction in AT, leading to a combined shrinkage of the polyploid genome. Interregion variability was also present in the polyploid genomes, serving to shrink the AdhA region in AT 10-fold more than CesA and contracting the AdhA region in DT (compared with the moderate gain experienced in CesA). Because the DT genome spent half of its time since divergence from the diploid D as a diploid itself (fig. 1), the rate of loss in the polyploid DT may be, in part, an underestimate masked by gains (primarily in the CesA region) that could have occurred during the 1.3 my spent as a diploid. These data suggest that the polyploid genome has, in fact, been experiencing genomic shrinkage in the 1–2 my postpolyploidization instead of the alternative (growth in the diploids relative to slower growth or stasis in the polyploid). Further analyses in Gossypium and other polyploids are required to test the generality of these observations.
Mechanisms Affecting the Rate of Sequence Loss and Gain
TE proliferation is thought to be responsible for most genome size growth in angiosperms (Bennetzen 2000, 2002; Kidwell 2002). This leads to the a priori hypothesis that a majority of the size difference between extant genomes reflects differential proliferation of TEs in a manner congruent with genome size (i.e., the A/AT lineage will have accrued TEs twice as fast as the D/DT lineage, as would A vs. D). The bias in TE proliferation observed between A/AT and D/DT is in the direction that is expected and is slightly more exaggerated than expected (fig. 1; table 1). The A/AT lineage gained TE sequence at a rate that was over 2.5-fold greater than the D/DT lineage. In the time that the A and D genomes evolved independent from the polyploid genomes (ca. 1–2 my), however, the A genome has not gained TE sequence in either of these regions, whereas the D genome has gained TE sequence at a rate of approximately one-third the ancestral rate (8.64 x 10–9 nucleotides per year due to a single insertion).
The higher rate of TE gain in the ancestral lineages may reflect several nonmutually exclusive factors. The subsequent reductions in observed TE insertion rates may be explained by the episodic nature of TE proliferation (Wicker and Keller 2007; Hawkins et al. 2008), as well as by types of TEs that may have proliferated in the individual genomes since diploid–polyploid divergence. For example, the TE population in these genomes may be concentrated in regions that have not been surveyed, potentially due to different integrational or targeting requirements of the types of elements that have proliferated. An interesting alternative, however, is that 1 or more of these genomes may have become less permissive of TE proliferation.
Just as TEs have been implicated in genome size growth, their amplification and genomic presence can also lead to genome size contraction via intrastrand homologous recombination. Intrastrand homologous recombination has been demonstrated in many systems and at various levels (Kalendar et al. 2000; Devos et al. 2002; Vitte and Panaud 2003; Wicker et al. 2003; Vitte and Bennetzen 2006), raising questions about how constraints on or stimulation of LTR recombination varies among species. The data from Gossypium indicate that intrastrand homologous recombination may be rare, as only one solo-LTR was observed (vs. 13 intact elements in the 4 genomes; table 2); however, data indicate that even a rare event can greatly impact the rate of deletion. The AdhA region experienced a rate of deletion that was over 397-fold greater due to the single intrastrand homologous recombination event, ultimately leading to net contraction for the region; similarly, this single deletion increased the overall deletion rate across the 2 regions nearly 19-fold and reversing what would be an overall net gain of 1.57 x 10–8 nucleotides per year to a contraction of 1.01 x 10–8 nucleotides per year.
Biased accumulation of small indels has been promoted (Petrov 1997, 2002), as well as criticized (Gregory 2003), as a solution to the discordance between the phylogenetic placement of plants possessing small genomes and the potential of deletions to shrink genomes (Vitte and Bennetzen 2006). The indel bias proposal is that, on average, smaller genomes will acquire more frequent and larger deletions (<400 nt) than larger genomes, thus slowly and stochastically shrinking the size of the genome more in the smaller genomes. Although the data presented here provide some support for this notion, this is limited to the period of evolution since polyploidization. The branches that we would expect to show the most bias (i.e., the branches where the most differential genome size change likely took place, A/AT and D/DT) were contradictory over the 2 regions in this respect, with the average deletion size less than twice the average size in D/DT and the insertion size greater for A/AT in only 1 of the 2 sequenced regions. Overall, the data support a slightly larger average deletion size for D/DT, while also being less frequent, yet a slightly larger average insertion size for D/DT, although also less frequent. Taken together, the data for the diploid genomes suggest the possibility of a small indel bias as the smaller genomes tend to add less sequence (D/DT) or delete more sequence (D) through small indels than the larger genomes, while noting regional biases in small indel formation. The data also suggest that polyploid genome has, in general, experienced increased sequence turnover (both as small deletions and small insertions).
Illegitimate recombination is attractive as a method for genome contraction, despite its tendency to create small rather than large deletions (Petrov 2002; Bennetzen et al. 2005), due to its presumed global nature and the idea that the effects of a slow, consistent "genomic leak" would outweigh episodic TE amplification over time. The data presented here fail to provide support for IR as a key determinant of genome size variation in Gossypium. Just as with other mechanisms of genome size evolution, illegitimate recombination may operate heterogeneously within genomes, affecting some genomic regions more than others and perhaps linked to regional features such as level of chromatin unwinding. This heterogeneity is evident in the regions and genomes studied here, where illegitimate recombination added sequence in about half of the cases but deleted sequence in the polyploid. Finally, for most regions harboring TEs, TEs played a larger role in genome size increase than could be compensated for by illegitimate recombination, and removal of a single TE (as observed in the AdhA region of AT) creates a much greater sequence reduction than deletion via illegitimate recombination (over 900-fold more for this region).
Although IR does not appear to be a major mechanism of genome size change in the regions and genomes studied here, our data suggest that polyploidy induced a shift in the bias of illegitimate recombination toward deletions over insertions. Although this bias toward contraction may be thwarted by TE insertions, as mentioned above, it is suggestive, as previously reported (Grover et al. 2007), of a mechanism to partially explain the phenomenon of genomic downsizing in polyploids.
The category encompassing unknown mechanisms (see Methods) had a surprisingly large impact on indel rate in some regions (CesA) and genomes (e.g., DT). For some of the evaluated indels, it was clear that evolution had simply erased any potential evidence of mechanistic hallmarks as the sequence bordering the indel had clearly been subject to subsequent mutations. In many cases, however, the sequence proximal to and distal to the indel was nearly perfectly conserved, yet it did not display the motifs commonly associated with mechanisms of genome downsizing. This indicates 2 not mutually exclusive possibilities. First, the motifs typically associated with mechanisms that generate indels, such as the repeats associated with illegitimate recombination, may not consistently be generated. Although this is true for a fraction of the "unknown" indels, it is also likely that there are mechanisms operating to add and remove DNA from the genome that do not leave evidence of their action. This possibility warrants further exploration and may be better addressed as more genomic data from closely related species, which enables such fine-scale analyses, become available.
|
Concluding Remarks |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
The data presented here highlight the instability of the rates and mechanisms of genome size change on an evolutionary timescale corresponding to divergence of species within a single angiosperm genus. Although much research has focused on mechanisms of genome size change, less is known concerning rates of DNA removal and gain due to specific mechanisms, and, to our knowledge, none have addressed the issue of how rates of the various mechanisms governing genome size expansion or contraction change over time. The heterogeneous nature of genome size evolution elucidated here is underscored by both the differences in genome contraction and growth experienced by regions within a single genome and by genomes over time. The complexities revealed here underscore the dynamics of genome size evolution that may be revealed by focused phylogenetic analyses.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary tables 1–4 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
We thank R. Percifield for technical assistance, the National Science Foundation Plant Genome Program, and the National Science Foundation BAC Program for financial support.
|
Footnotes |
|---|
Charles Delwiche, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionConcluding RemarksSupplementary MaterialAcknowledgementsReferences |
|---|
Adams KL, Wendel JF. Polyploidy and genome evolution. Curr Opin Plant Biol (2005) 8:135–141.[CrossRef][Web of Science][Medline]
Bennett MD, Leitch IJ. Nuclear DNA amounts in angiosperms. Ann Bot (1995) 76:113–176.
Bennett MD, Leitch IJ. Nuclear DNA amounts in angiosperms: progress, problems and prospects. Ann Bot (2005a) 95:45–90.
Bennett MD, Leitch IJ. Genome size evolution in plants. In: The evolution of the genome—Gregory TR, ed. (2005b) San Diego (CA): Elsevier. 89–162.
Bennetzen JL. Transposable element contributions to plant gene and genome evolution. Plant Mol Biol (2000) 42:251–269.[CrossRef][Web of Science][Medline]
Bennetzen JL. Mechanisms and rates of genome expansion and contraction in flowering plants. Genetica (2002) 115:29–36.[CrossRef][Web of Science][Medline]
Bennetzen JL, Ma J, Devos KM. Mechanisms of recent genome size variation in flowering plants. Ann Bot (2005) 95:127–132.
Brudno M, Do C, Cooper G, Kim MF, Davydov E, Green ED, Sidow A, Batzoglou S. LAGAN and Multi-LAGAN: efficient tools for large scale multiple alignment of genomic DNA. Genome Res (2003) 13:721–731.
Chantret N, Salse J, Sabot F, et al, (19 co-authors). Molecular basis of evolutionary events that shaped the hardness locus in diploid and polyploid wheat species (Triticum and Aegilops). Plant Cell (2005) 17:1033–1045.
Chen ZJ, Ni Z. Mechanisms of genomic rearrangements and gene expression changes in plant polyploids. Bioessays (2006) 28:240–252.[CrossRef][Web of Science][Medline]
Cronn RC, Small RL, Haselkorn T, Wendel JF. Rapid diversification of the cotton genus (Gossypium: malvaceae) revealed by analysis of sixteen nuclear and chloroplast genes. Am J Bot (2002) 89:707–725.
Devos KM, Brown J, Bennetzen JL. Genome size reduction through illegitimate recombination counteracts genome expansion in Arabidopsis. Genome Res (2002) 12:1075–1079.
Ewing B, Green P. Basecalling of automated sequencer traces using phred. II. Error probabilities. Genome Res (1998) 8:186–194.
Ewing B, Hillier L, Wendl MC, Green P. Base-calling of automated sequences traces using phred. I. Accuracy assessment. Genome Res (1998) 8:175–185.
Gordon D, Abajian C, Green P. Consed: a graphical tool for sequence finishing. Genome Res (1998) 8:195–202.
Green P. Phrap documentation [Internet] (1999) [cited 2008 April 23]. Available from: http://www.phrap.org/phrap.docs/phrap.html.
Gregory TR. Is small indel bias a determinant of genome size? Trends Genet (2003) 19:485–488.[CrossRef][Web of Science][Medline]
Grover CE, Kim H, Wing RA, Paterson AH, Wendel JF. Microcolinearity and genome evolution in the AdhA region of diploid and polyploid cotton (Gossypium). Plant J (2007) 50:995–1006.[CrossRef][Web of Science][Medline]
Grover CE, Kim H, Wing RA, Paterson AH, Wendel JF. Incongruent patterns of local and global genome size evolution in cotton. Genome Res (2004) 14:1474–1482.
Gu YQ, Salse J, Coleman-Derr D, et al, (13 co-authors). Types and rates of sequence evolution at the high-molecular-weight glutenin locus in hexaploid wheat and its ancestral genomes. Genetics (2006) 174:1493–1504.
Hawkins JS, Hu G, Rapp RA, Grafenberg JL, Wendel JF. Phylogenetic determination of the pace of transposable element proliferation in plants: copia and LINE-like elements in Gossypium. Genome (2008) 51:11–18.[Medline]
Hawkins JS, Kim H, Nason JD, Wing RA, Wendel JF. Differential lineage-specific amplification of transposable elements is responsible for genome size variation in Gossypium. Genome Res (2006) 16:1252–1261.
Hendrix B, Stewart JM. Estimation of the nuclear DNA content of Gossypium species. Ann Bot (2005) 95:789–797.
Hill P, Burford D, Martin D, Flavell AJ. Retrotransposon populations of Vicia species with varying genome size. Mol Genet Genomics (2005) 273:371–381.[CrossRef][Web of Science][Medline]
Kalendar R, Tanskanen J, Immonen S, Nevo E, Schulman AH. Genome evolution in wild barley (Hordeum spontaneum) by BARE-1 retrotransposon dynamics in response to sharp microclimatic divergence. Proc Natl Acad Sci USA (2000) 97:6603–6607.
Kidwell MG. Transposable elements and the evolution of genome size in eukaryotes. Genetica (2002) 115:49–63.[CrossRef][Web of Science][Medline]
Leitch IJ, Chase MW, Bennett MD. Phylogenetic analysis of DNA C-values provides evidence for a small ancestral genome size in flowering plants. Ann Bot (1998) 82(Suppl A):85–94.
Ma J, Devos KM, Bennetzen JL. Analyses of LTR-retrotransposon structures reveal recent and rapid genomic DNA loss in rice. Genome Res (2004) 14:860–869.
Ozkan H, Tuna M, Arumuganathan K. Nonadditive changes in genome size during allopolyploidization in the wheat group (Aegilops-Triticum) group. J Hered (2003) 94:260–264.
Petit M, Lim KY, Julio E, Poncet C, de Borne FD, Kovarik A, Leitch AR, Grandbastien M, Mhiri C. Differential impact of retrotransposon populations on the genome of allotetraploid tobacco (Nicotiana tabacum). Mol Genet Genomics (2007) 278:1.[CrossRef][Web of Science][Medline]
Petrov D. Slow but steady: reduction of genome size through biased mutation. Plant Cell (1997) 9:1900–1901.[CrossRef][Web of Science][Medline]
Petrov DA. Mutational equilibrium model of genome size evolution. Theor Popul Biol (2002) 61:531–544.[CrossRef][Web of Science][Medline]
Piegu B, Guyot R, Picault N, et al, (11 co-authors). Doubling genome size without polyploidization: dynamics of retrotransposition-driven genomic expansions in Oryza australiensis, a wild relative of rice. Genome Res (2006) 16:1262–1269.
SanMiguel P, Bennetzen JL. Evidence that a recent increase in maize genome size was caused by the massive amplification of intergene retrotransposons. Ann Bot (1998) 82:37–44.
Senchina DS, Alvarez I, Cronn RC, Liu B, Rong JK, Noyes RD, Paterson AH, Wing RA, Wilkins TA, Wendel JF. Rate variation among nuclear genes and the age of polyploidy in Gossypium. Mol Biol Evol (2003) 20:633–643.
Soltis DE, Soltis PS. Polyploidy: recurrent formation and genome evolution. Trends Ecol Evol (1999) 9:348–352.
Vitte C, Bennetzen JL. Analysis of retrotransposon structural diversity uncovers properties and propensities in angiosperm genome evolution. Proc Natl Acad Sci USA (2006) 103:17638–17643.
Vitte C, Panaud O. Formation of solo-LTRs through unequal homologous recombination counterbalances amplifications of LTR retrotransposons in rice Oryza sativa L. Mol Biol Evol (2003) 20:528–540.
Wendel JF, Cronn RC. Polyploidy and the evolutionary history of cotton. Adv Agron (2003) 78:139–186.[CrossRef]
Wendel JF, Cronn RC, Johnston JS, Price HJ. Feast and famine in plant genomes. Genetica (2002) 115:37–47.[CrossRef][Web of Science][Medline]
Wicker T, Keller B. Genome-wide comparative analyses of copia retrotransposons in Tritaceae, rice, and Arabidopsis reveals conserved ancient evolutionary lineages and distinct dynamics of individual copia families. Genome Res (2007) 17:1072–1081.
Wicker T, Stein N, Albar L, Feuillet C, Schlagenhauf E, Keller B. Analysis of a contiguous 211 kb sequence in diploid wheat (Triticum monococcum L.) reveals multiple mechanisms of genome evolution. Plant J (2001) 26:307–316.[CrossRef][Web of Science][Medline]
Wicker T, Yahiaoui N, Guyot R, Schlagenhauf E, Liu Z-D, Dubcovsky J, Keller B. Rapid genome divergence at orthologous low molecular weight glutenin loci of the A and Am genomes of wheat. Plant Cell (2003) 15:1186–1197.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. S.S. Ammiraju, F. Lu, A. Sanyal, Y. Yu, X. Song, N. Jiang, A. C. Pontaroli, T. Rambo, J. Currie, K. Collura, et al. Dynamic Evolution of Oryza Genomes Is Revealed by Comparative Genomic Analysis of a Genus-Wide Vertical Data Set PLANT CELL, December 1, 2008; 20(12): 3191 - 3209. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Wawrzynski, T. Ashfield, N. W.G. Chen, J. Mammadov, A. Nguyen, R. Podicheti, S. B. Cannon, V. Thareau, C. Ameline-Torregrosa, E. Cannon, et al. Replication of Nonautonomous Retroelements in Soybean Appears to Be Both Recent and Common Plant Physiology, December 1, 2008; 148(4): 1760 - 1771. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Charles, H. Belcram, J. Just, C. Huneau, A. Viollet, A. Couloux, B. Segurens, M. Carter, V. Huteau, O. Coriton, et al. Dynamics and Differential Proliferation of Transposable Elements During the Evolution of the B and A Genomes of Wheat Genetics, October 1, 2008; 180(2): 1071 - 1086. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||