• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jul 15, 2008; 105(28): 9691–9696.
Published online Jul 3, 2008. doi:  10.1073/pnas.0803981105
PMCID: PMC2474508
Evolution

Acc homoeoloci and the evolution of wheat genomes

Abstract

The DNA sequences of wheat Acc-1 and Acc-2 loci, encoding the plastid and cytosolic forms of the enzyme acetyl-CoA carboxylase, were analyzed with a view to understanding the evolution of these genes and the origin of the three genomes in modern hexaploid wheat. Acc-1 and Acc-2 loci from each of the wheats Triticum urartu (A genome), Aegilops tauschii (D genome), Triticum turgidum (AB genome), and Triticum aestivum (ABD genome), as well as two Acc-2-related pseudogenes from T. urartu were sequenced. The 2.3–2.4 Mya divergence time calculated here for the three homoeologous chromosomes, on the basis of coding and intron sequences of the Acc-1 genes, is at the low end of other estimates. Our clock was calibrated by using 60 Mya for the divergence between wheat and maize. On the same time scale, wheat and barley diverged 11.6 Mya, based on sequences of Acc and other genes. The regions flanking the Acc genes are not conserved among the A, B, and D genomes. They are conserved when comparing homoeologous genomes of diploid, tetraploid, and hexaploid wheats. Substitution rates in intergenic regions consisting primarily of repetitive sequences vary substantially along the loci and on average are 3.5-fold higher than the Acc intron substitution rates. The composition of the Acc homoeoloci suggests haplotype divergence exceeding in some cases 0.5 Mya. Such variation might result in a significant overestimate of the time since tetraploid wheat formation, which occurred no more than 0.5 Mya.

Keywords: acetyl-CoA carboxylase, Triticeae, grass

Interest in the genes encoding acetyl-CoA carboxylase (ACCase) stems from the central importance of the enzyme in lipid metabolism. In the plastid, ACCase encoded by the Acc-1 gene catalyzes the first step in fatty acid synthesis, converting acetyl-CoA to malonyl-CoA, where it is used to add two-carbon fragments to the growing fatty acid chain. In the cytoplasm, ACCase encoded by the Acc-2 gene makes malonyl-CoA for secondary metabolism including the extension of long-chain fatty acids. The gene encoding plastid ACCase in grasses, which replaced the multisubunit plastid enzyme of bacterial origin still found in dicots, arose by duplication of a gene encoding the multidomain cytosolic ACCase (1). One copy of the ancestral gene acquired a plastid-targeting signal (1, 2) and a new promoter and new regulatory elements (2, 3). The wheat Acc-1 and Acc-2 genes show no sequence similarity outside the exons encoding the mature enzyme (2, 3). For both Acc-1 and Acc-2, all three Triticum aestivum homoeologs are transcriptionally active, and each uses two nested promoters and alternative splicing of the first intron to produce transcripts with different organ specificity. Each of the homoeologs contributes equally to the total mRNA level of each transcript type (2, 3).

Analysis of genomic DNA provided evidence of a single copy of an active Acc gene in all of the homoeologous genomes of the hexaploid wheat T. aestivum (1, 48). One Acc-1 gene and one Acc-2 gene are present in rice. Sampling of wheat relatives (1, 4, 5) identified two copies of Acc-1 in diploid Aegilops speltoides. Topology of the Acc-1 phylogenetic tree (in figure 1 of ref. 4) suggests that the gene was duplicated at least once before divergence of the Aegilops and Triticum genomes, and this was followed by the loss of a different copy of the gene in different lineages. Acc-1 genes are located on the short arm of wheat group 2 chromosomes. The Acc-2 gene was also duplicated at least twice in Triticeae evolution followed by specific gene loss in some lineages (5). An Acc-2-related partially processed pseudogene has been found in all Triticeae species tested (1). Acc-2 genes are located on the long arm of group 3 chromosomes (5, 6).

Fig. 1.
Sequenced Acc loci. The coding part of Acc genes and conserved genes located upstream of Acc-1 and immediately downstream of Acc-2 are shown in black. Pseudogenes are indicated by black stripes. Sequence similarity between Acc-1 and Acc-2 loci is limited ...

The wheat Acc genes, because of their large size and the availability of functional information, provide an excellent starting point for comparative genome analysis. The transcribed part of the wheat Acc-1 gene is ≈13.5 kb with ≈6.9 kb of coding sequence (2,311 amino acids including the putative chloroplast transit peptide) and has 33 introns with a total length of ≈6.3 kb. The transcribed part of the Acc-2 gene is ≈12.3 kb with ≈6.8 kb of coding sequence (2,258 or 2,260 amino acids) and has 29 introns with total length of ≈4.4 kb.

Evolution of Triticum and Aegilops has been the focus of studies driven by interest in wheat as a major crop and by the complexity of their evolutionary history: from the divergence of wheat diploid ancestors, through two steps of polyploidization, domestication and extensive breeding. In wheat, most active genes are highly dispersed between long rapidly changing intergenic regions consisting primarily of repetitive sequences. In addition to mutations, active genes are affected by frequent gene duplication and loss, as well as conversion and unequal crossing over between paralogs leading to hybrid genes. These processes are lineage-, genome-, and gene copy-specific. In addition to the recent polyploidization in wheat, earlier segmental duplication and whole genome duplication, such as the ancient duplication predating divergence of the major grass subfamilies, and chromosome rearrangements and segmental loss (911), add complexity to the comparative genome analysis.

The time of formation of the hexaploid wheat genome, T. aestivum (ABD), was estimated to be 10 thousand years ago (12), based on archeological artifacts. The divergence times of the wheat diploid ancestors, Triticum urartu (A genome), Aegilops tauschii (D genome), and A. speltoides (S genome, a progenitor of the B genome), of 2–4 Mya and the time of formation of tetraploid wheat, Triticum turgidum (AB genome) of <0.5 Mya, were previously estimated by using the molecular clock method (4, 13). An earlier divergence of the progenitors of the A, B, and D genomes was proposed based on analysis of the genes Sut1 and PhyC (14). All these estimates carry the possible error associated with dating fossils, which was used to place the divergence of the major grass subfamilies. Our studies on wheat evolution calibrate the clock by using 60 Mya for the divergence of Pooideae and Panicoideae. The separation of the major clades of grasses has been estimated at 55–83 Mya (1517), whereas an unpublished study of six fossils yielded a date of 52 ± 8 Mya (E. Kellogg, personal communication).

In an earlier work, we studied Acc evolution in hexaploid wheat and a broad sample of its relatives, based on the sequences of a gene fragment (1, 4, 5). Here, we analyze full-length Acc genes and their genomic context in diploid, tetraploid, and hexaploid wheat to extend our understanding of the evolution of these genes and the wheat genomes.

Results

BACs containing all of the Acc genes predicted on the basis of previously published sequence information (1, 4, 5) were isolated [Fig. 1 and supporting information (SI) Table S1], except for the A homoeolog of Acc-2 from T. turgidum. In the latter case, all three Acc-2 BACs isolated from the T. turgidum library contained only the B homoeolog. The Acc-2,2 gene was found previously only in a wheat genomic DNA library (5, 7); it could not be confirmed in any of our follow-up experiments, including the BAC analysis described in this paper. T. turgidum and T. urartu BACs containing an Acc-2-related partially processed pseudogene (ψ-Acc-2) were also found. The available T. urartu ψ-Acc-2 sequence is not complete, but it contains sequences corresponding to introns 2 and 3 of the active Acc-2 gene, one premature stop codon and one frame shift (Fig. 2). This pseudogene was previously found to be present in all homoeologous chromosomes of wheat and in its diploid relatives, including rye and barley (1). A second copy of the Acc-2 gene [previously called Acc-2 II (1)] was found in T. urartu BACs. This gene is a pseudogene (Ψ-Acc-2). It contains sequences corresponding to the coding exons and 28 introns found in active Acc-2 genes, but it has two in-frame stop codons and two frame shifts (Fig. 2). The frame shifts were confirmed by PCR amplification and sequencing of the corresponding fragment of T. urartu genomic DNA. It appears to be a “young” pseudogene, as suggested by a high ratio of synonymous to nonsynonymous substitutions. It acquired some of the deleterious mutations only very recently; the two frame shifts found in T. urartu are not present in Triticum monococcum or T. turgidum (determined by PCR amplification and sequencing of a fragment of genomic DNA). This gene was previously found, based on partial sequence, to be present in T. urartu, T. monococcum, and Triticum armeniacum (A and presumably G homoeolog); in Triticum timopheevii (presumably G homoeolog); and in T. turgidum (presumably B homoeolog) (5) but not in any other species tested by PCR cloning/sequencing or by screening genomic and BAC libraries (ref. 5 and this paper). Most notably, it appears to be absent in T. aestivum and in A. tauschii.

Fig. 2.
Structure of Acc genes and pairwise sequence comparisons at synonymous and intron sites. (A) Exons encoding ACCase highly conserved BC and CT domains are shown in the same color. Introns are shown in white. 5′- and 3′-UTRs are shown in ...

None of the sequenced loci contains more than one Acc gene or pseudogene. Acc-2 genes and Acc-2-related pseudogenes are not within a short distance of each other (same BAC insert). However, they are located in the same chromosome segment, as shown by chromosome-mapping experiments (5). This observation is also consistent with different chromosomal localizations of Acc-1 and Acc-2 genes. The structure of Acc genes is shown in Fig. 2A. The distribution of synonymous and intron substitutions along the Acc genes and nucleotide substitutions in sequences flanking the Acc genes is shown in Fig. 2 B–E and Fig. S1. Intron sequence divergence does not allow any comparisons above the Triticeae tribe level, and sequence divergence in the flanking regions does not allow any comparisons other than among homoeologs. The distribution of Acc-1 synonymous substitutions between species representing the three grass subfamilies wheat and Alopecurus (Pooideae), rice (Ehrhartoideae) and maize, and Sorghum and Setaria (Panicoideae) (Fig. 2B) varies significantly along the coding part of Acc (Fig. 2) in a pattern suggesting correlation with sequence conservation at the amino acid level; the biotin carboxylase (BC) and carboxyltransferase (CT) domains are the most conserved parts of ACCase. Their coding sequences have lower incidence of nonsynonymous substitution but also lower incidence of substitutions at synonymous sites. The pattern is less profound for Acc-2 (Fig. 2D). Note that this gene does not encode a transit peptide, which is less conserved. As a consequence, the conserved BC domain is located very close to the 5′ terminus of the gene.

Acc-1 comparisons for barley and wheat and between wheat A, B, and D homoeologs show an increased incidence of synonymous substitutions near the 3′-end of the gene (Fig. 2 B and C). For the first two-thirds of the ACCase gene (from the translation start to the beginning of the CT domain) containing all but two introns (Fig. 2A), the synonymous and intron substitution rates are very similar. The genetic distances between barley and wheat increase with distance from the translation start codon, similar to the increase of the synonymous site distance between wheat and Alopecurus, Setaria, rice, and maize (Fig. 2B). The distances along this part of the gene remain similar for all three pairs of homoeologs (Fig. 2C). The increased synonymous substitution rate in the last one-third of the gene (CT domain) is due to accumulation of homoeolog-specific substitutions. This phenomenon is most obvious for the Acc-1 B homoeolog where 14 of 36 mutations (40%) are found in a stretch of 418 nucleotides (6%). These homoeolog-specific substitutions also contribute to the observed higher genetic distance between 3′-end fragments of the gene when wheat and barley sequences are compared.

For the first two-thirds of the Acc-1 gene (segment Ea and I, Fig. 3A), the synonymous substitution rate and the intron substitution rate are equal for the three pairs of homoeologs. The synonymous substitution rate in the last third of Acc-1 (segment Eb Fig. 3A) is three times higher, reflecting the accumulation of synonymous substitutions near the end of the gene (above). The substitution rates are also 3-fold higher in the conserved segments at the 5′-end of the gene (including putative promoters and regulatory elements, 5′-UTR, and the first intron) and immediately downstream of the gene (R segments, Fig. 3A). The rate remains the same or is only 50% higher in the 3′-end segment including the last intron, 3′-UTR, and short stretch of highly conserved sequence following the transcription termination site (Fig. 3A).

Fig. 3.
Nucleotide substitution rates in different parts of the Acc-1 (A–D) and Acc-2 (E–H) loci. Synonymous substitution rates are shown for the coding sequences: Ea, from translation start codon to the beginning of the CT domain; Eb, CT domain ...

Very few nucleotide substitutions were found in the 70-kb fragment of the Acc-1 B locus in T. turgidum and T. aestivum (Fig. 3B and Fig. S1A), consistent with the formation of the ABD hexaploid from an AB tetraploid as recently as ten thousand years ago. The distances, however, are much higher among the A and D Acc-1 homoeologs (Fig. 3 C and D). This difference may be explained by haplotype variation and, for the T. urartu and T. turgidum pair, by the age of the AB tetraploid. The nucleotide substitution rates upstream and downstream of the Acc-1 gene are significantly higher (2- to 8-fold) than either intron or synonymous rates within the gene (L and R segments, Fig. 3 C and D). These rates vary significantly across the locus (Fig. S1A).

For the first two-thirds of Acc-2 (segment Ea and I, Fig. 3E), the synonymous substitution rate and the intron substitution rate are equal for the three pairs of homoeologs. The synonymous substitution rate is two times higher in the last third of the Acc-2 gene and in the gene located immediately downstream (segment Eb and e, Fig. 3E). Average synonymous and intron rates are two to three times higher in the Acc-2 than in Acc-1 genes. Significantly more homoeolog-specific substitutions accumulated in the A and B Acc-2 genes. Uneven distribution of the synonymous substitutions along the gene is most striking for the A to B comparison (Fig. 2E).

The Acc-2 B locus in T. turgidum and T. aestivum is characterized by a significant divergence upstream of Acc-2 and low divergence downstream of the gene. This observation might be explained by haplotype variation coupled with a recombination event whose cross-over was located downstream of Acc-2 (Fig. 3F and Fig. S1B). For the Acc-2 A locus in T. turgidum and T. aestivum, only the upstream flanking region was available for comparison (Fig. 3G). The Acc-2 D locus shows the same correlation as the B locus, except that in this case, the rate difference between the upstream and downstream flanking regions is only 2-fold (Fig. 3H). With the exception of the downstream region at the B locus, the nucleotide substitution rates in the intergenic region are 2- to 9-fold higher than either intron or synonymous rates within the Acc-2 gene. Similar to the Acc-1 locus, these rates vary significantly across the locus (Fig. S1B).

Pairwise distances at synonymous sites and intron distances are shown in Table 1 to describe quantitatively some of the observations presented above and to estimate divergence times. The molecular clock was set by using 60 Mya for divergence of Pooideae and Panicoideae. The molecular clock for introns, which can be aligned reliably only at the tribe level or below, was calibrated by using dates for divergence of wheat and barley calculated from synonymous substitutions in each gene (Table 1). Two distances were calculated for the Acc-1 gene, one based on the full-length sequence and one based on the first two-thirds (segment Ea) of the coding sequence. The two distances are different for A, B, and D homoeologs and between wheat and barley but are identical (within error) for all other sequence pairs. This result is consistent with increased incidence of substitutions in the terminal third of the Acc-1 coding sequence (segment Eb) observed only at the homoeolog level and affecting significantly only the distances between pairs of A, B, and D genes and, to a lesser degree, distances between barley and wheat A, B, and D sequences. This phenomenon does not appear to be a general property of the Acc-1 gene. The sequence of segment Ea was also used to compare distances at synonymous and intron sites. Intron distances are very similar to distances at synonymous sites in neighboring exons (Fig. 3). The divergence times calculated by using the two data sets are consistent (Table 1).

Table 1.
Pairwise divergence times of grass genomes based on ACC genes

Discussion

Our analysis indicates a departure from neutral evolution at both Acc loci; substitution rates at synonymous and intron sites vary along the Acc genes (Fig. 2) and, on average, these rates are lower than substitution rates outside of the coding part of the genes (Figs. 2 and and3,3, Fig. S1). The significance of these differences is difficult to gauge because evolution of noncoding gene sequences as well as intergenic sequences is poorly understood. Furthermore, comparisons among homoeologous loci, the only level at which intergenic sequences can be compared, may be affected by haplotype variation. On the other hand, substitution rates at synonymous and intron sites in the same part of the gene are very similar (Figs. 2 and and3,3, Table 1). The number of synonymous substitutions among homoeologs is low for a reliable estimate, even for genes as large as Acc. Even then, synonymous and intron rates are very similar, except for A homoeologs of Acc-1 (Table 1, Fig. 3C).

The promoter regions of the Acc genes are very compact: no more than 1 kb, Fig. 2A. This is evident from sequence comparisons of the homoeologous genes. In the case of the A homoeologs of Acc-1 and Acc-2, the upstream sequences were changed completely by large indels within a short distance of the translation start codon (Fig. 1). Large blocks of conserved sequence are present at 5′- and 3′- ends of both Acc-1 and Acc-2 when A, B, and D homoeologs are compared, but only Acc-1 sequences could be used in rate calculations (Figs. 2C and and33A). Such comparisons for the Acc-2 genes would require significant manual editing of sequence alignments. Substitution rate in the noncoding 5′- and 3′-ends of Acc-1 is on average 3.5-fold higher than either synonymous or intron rate.

Comparisons of intergenic sequences at the Acc loci were only possible for pairs of homoeologs (Fig. S1); there is no sequence similarity between A, B, and D homoeologs outside of the genes shown in Fig. 1. Comparison of gene and intergenic sequences poses another problem; the intergenic rates vary significantly across the locus (Fig. S1), reflecting, most likely, a combination of DNA composition (active genes, various repetitive sequences, DNA methylation) as well as haplotype variation. There are few or no genes other than Acc predicted for the loci (Fig. 1 and gene models shown in the annotation files at www.tigr.org/tdb/e2k1/tae1). The average GC content of regions flanking Acc genes is 45% compared with 33% in introns. The analysis is further complicated by low information content for synonymous sites and absence of introns from the 3′-end of the Acc coding sequence. Haplotype variation and recombination, such as the postulated recombinations with cross-over sites located within or near the Acc-2 B and D genes (above), could lead to either over- or underestimation of relative rates. The average substitution rate in intergenic regions is 3.5 times higher than at intron sites, based on the data shown in Fig. 3 C, D, and G.

Assuming the intron substitution rate in Acc genes in Triticeae to be 3.6 × 10−9 substitutions per site per year (average intron substitution rate between barley and wheat is 0.084 Table 1; and divergence time is 11.6 Mya, Fig. 4), the average substitution rate in the intergenic region calculated as the number of substitutions per site per year is 13 × 10−9, similar to the rate reported previously (18). Clearly, synonymous and intron substitution rates do not directly reveal nucleotide substitution rates in the flanking intergenic regions.

Fig. 4.
Time line of wheat evolution. Divergence times were calculated by using 60 Mya for the divergence of Pooideae from Panicoideae and 11.6 Mya for the divergence of barley from wheat to calibrate the molecular clock. The latter calibration point was used ...

The average distance at synonymous sites for full-length Acc-1 is 0.58, 0.56, and 0.50 when Pooideae and Panicoideae, rice and Panicoideae, and rice and Pooideae sequences, respectively, are compared (Table 1). For Acc-2, the corresponding distances are 0.58, 0.57, and 0.53. These numbers are very similar to the multigene averages of 0.59, 0.56, and 0.48, respectively, calculated by us (Table S2) as well as distances previously determined from genomic surveys of multiple maize, rice, sorghum, and barley genes (9, 11). The corresponding multigene average substitution rate (including Acc genes) is 4.8 × 10−9 substitutions per synonymous site per year, assuming 60 Mya for the time Pooideae (wheat, barley) and Ehrhartoideae (rice) diverged from Panicoideae (maize, sorghum). This rate is lower than a commonly used previous estimate of 6.0 × 10−9 (19), but the lower estimate for the time of divergence of maize and rice (50 Mya) used in that study to calibrate the molecular clock accounts for most of the difference.

The 50 Mya time of divergence of rice from Triticeae estimated from sequences of multiple genes (Fig. 4) is consistent with the prevailing belief that wheat and rice diverged from each other only after they diverged from maize. Full-length Acc sequences were also used to estimate the divergence time of Aveneae (Alopecurus) from Triticeae at 31 Mya, Brachypodieae (Brachypodium) from Triticeae at 35 Mya, and Paniceae (Setaria) from Andropogoneae (maize) at 27 Mya (Table 1). The Brachypodium–Triticeae divergence time is within the range of a previous multigene estimate [35–40 Mya (20)]. The Setaria–maize divergence time is higher than the 24 Mya obtained previously from partial Acc-1 sequences of Panicum virgatum and maize (21). Finally, the divergence time of the two copies of maize Acc-1 gene at 16 Mya and the divergence time of maize and Sorghum at 13 Mya are consistent with the previously proposed scenario of evolution of these species including a whole genome duplication (polyploidization) of the maize genome (10).

Barley is an important outgroup for molecular evolution studies of the Triticum/Aegilops complex. The use of the divergence time between wheat and barley to set the molecular clock to date the subsequent events in wheat evolution may help minimize effects of substitution rate variability over longer distances and in different less closely related lineages and should avoid problems with the clock calibration for individual genes caused by differential loss of duplicated genes. In addition, intronic sequences can often be aligned with high confidence and used for the evolutionary analysis of the Triticeae species. For Acc genes, intron and synonymous substitution rates produce consistent estimates of divergence times when neighboring introns and exons are compared, but introns have a much higher information content (Fig. 3 and Table 1).

The distance between wheat and barley calculated for synonymous sites is 0.107 for Acc-1 (Ea segment) and 0.100 for Acc-2 (full-length), and the corresponding divergence times are 11.8 and 10.6 Mya (Table 1). These numbers are very similar to the multigene average estimates of 0.114 and 11.6 Mya, respectively (Fig. 4 and Table S3). The wheat–barley divergence time estimates vary between 6.7 and 16.2 Mya, reflecting properties and evolutionary history of individual genes.

By using 11.6 Mya for the time of wheat divergence from barley and coding and intron sequences of the first two-thirds of the Acc-1 genes, the three divergence times calculated for the A, B, and D homoeologous genomes are 2.3–2.4 Mya, consistent with previous estimates based on a short fragment of the gene and by using two alternative evolutionary models (4, 14). Note that for the Acc-1 gene, these estimates do not differ from the estimates using a much more distant event to calibrate the clock (Table 1).

We analyzed a group of genes for which barley, rice, maize, and multiple wheat sequences, putative homoeologous sequences, are available (Fig. 4 and Tables S2 and Table S3). Divergence times between these wheat genes, calculated relative to the divergence of the same barley gene, vary between 2 and 9 Mya. Only several of these genes have been studied more extensively, allowing phylogenetic analysis. The divergence times for these genes vary between 2 and 6 Mya, with the exception of A-B and A-D divergence of Acc-2 at 9 and 8 Mya, respectively. We argue that the estimates at the lower end of the spectrum (2–3 Mya) reveal more accurately the divergence time of the ancestors of the A, B, and D genomes of wheat and its relatives. The higher estimates probably capture haplotype variation, gene duplication, and specific loss, including the possibility of formation of chimeric genes by unequal crossing over between duplicated genes.

The large size of the Acc genes allowed us to identify sequence variations of unknown origin affecting divergence time estimates: the increased incidence of synonymous substitutions in the last third of the Acc-1 gene and a significant variation of the substitution rate along the Acc-2 gene. In the former case, the extra substitutions increase the divergence time between A, B, and D homoeologs by only ≈0.5 Mya, if a proper barley reference sequence is used for the molecular clock calibration. In the latter case, the divergence time estimates could differ by as much as 4-fold, if 1-kb coding sequences from different parts of the gene are compared. Different haplotypes or gene copies are a possible explanation of the observed sequence variation. Haplotype divergence for A, B, and D genomes is significant and can reach 1 million years at some loci (refs. 1, 4, 5, 13, and 22 and this work). The haplotype divergence observed for A. speltoides, the progenitor of the B genome, is even greater and can exceed the divergence between A, B, and D homoeologs (23). Both Acc-1 and Acc-2 genes were duplicated at least once after divergence of Triticum/Aegilops from barley (1, 5) followed by specific loss of one gene copy in different species.

The composition of the Acc homoeoloci suggests a significant haplotype divergence, which in some cases reaches or even exceeds 0.5 million years (Table 1). Such variation might result in a significant overestimate of the time of tetraploid wheat formation, which, based on the Acc-1 estimates shown in Table 1, occurred no more than ≈0.5 Mya. This upper estimate is consistent with results published previously (4, 13). The date was deduced from the comparison of the Acc-1 A homoeolog in T. urartu to T. turgidum and T. aestivum (average of two synonymous and two intron estimates). The estimates based on synonymous substitutions are only approximate due to very low sequence divergence. The estimate based on the intronic sequences, which have a higher information content, is slightly higher: ≈0.7 Mya.

Materials and Methods

Details of the BAC sequencing strategy can be found in the SI Text, along with the methods used for assembly and alignment. Substitution rates and divergence times were calculated by using standard methods. Additional references are given in the SI Text.

Supplementary Material

Supporting Information:

Acknowledgments.

We thank A. R. Akhunova and J. Dvorak (University of California, Davis) and W. Li and B. Gill (Kansas State University, Manhattan, KS) for providing BACs; J. Carroll and W. J. Buikema (University of Chicago CRC DNA Sequencing Facility, Chicago) for BAC sequencing; and W. Zhu (The Institute for Genomic Research, Rockville, MD) for annotation of BAC sequences. The hexaploid wheat BAC library is a resource obtained by the Genoplante consortium (www.genoplante.com). This project was supported in part by a grant from U.S. Department of Agriculture, National Research Initiative Competitive Grants Program.

References

1. Huang SX, et al. Phylogenetic analysis of the acetyl-CoA carboxylase and 3- phosphoglycerate kinase loci in wheat and other grasses. Plant Mol Biol. 2002;48:805–820. [PubMed]
2. Zuther E, et al. Complex nested promoters control tissue-specific expression of acetyl-CoA carboxylase genes in wheat. Proc Natl Acad Sci USA. 2004;101:1403–1408. [PMC free article] [PubMed]
3. Podkowinski J, et al. Expression of cytosolic and plastid acetyl-CoA carboxylase genes in young wheat plants. Plant Physiol. 2003;131:763–772. [PMC free article] [PubMed]
4. Huang S, et al. Genes encoding plastid acetyl-CoA carboxylase and 3- phosphoglycerate kinase of the Triticum/Aegilops complex and the evolutionary history of polyploid wheat. Proc Natl Acad Sci USA. 2002;99:8133–8138. [PMC free article] [PubMed]
5. Faris J, Sirikhachornkit A, Haselkorn R, Gill B, Gornicki P. Chromosome mapping and phylogenetic analysis of the cytosolic acetyl-CoA carboxylase loci in wheat. Mol Biol Evol. 2001;18:1720–1733. [PubMed]
6. Gornicki P, et al. Plastid-localized acetyl-CoA carboxylase of bread wheat is encoded by a single gene on each of the three ancestral chromosome sets. Proc Natl Acad Sci USA. 1997;94:14179–14185. [PMC free article] [PubMed]
7. Podkowinski J, Sroga GE, Haselkorn R, Gornicki P. Structure of a gene encoding a cytosolic acetyl-CoA carboxylase of hexaploid wheat. Proc Natl Acad Sci USA. 1996;93:1870–1874. [PMC free article] [PubMed]
8. Gornicki P, et al. Wheat acetyl-CoA carboxylase: cDNA and protein structure. Proc Natl Acad Sci USA. 1994;91:6860–6864. [PMC free article] [PubMed]
9. Wang X, Shi X, Hao B, Ge S, Luo J. Duplication and DNA segmental loss in the rice genome: implications for diploidization. New Phytol. 2005;165:937–946. [PubMed]
10. Wei F, et al. Physical and Genetic Structure of the Maize Genome Reflects Its Complex Evolutionary History. PLoS Genet. 2007;3:e123. [PMC free article] [PubMed]
11. Paterson AH, Bowers JE, Chapman BA. Ancient polyploidization predating divergence of the cereals, and its consequences for comparative genomics. Proc Natl Acad Sci USA. 2004;101:9903–9908. [PMC free article] [PubMed]
12. Nesbitt M, Samuel D. In: Padulosi S, Hammer K, Heller J, editors. Proceedings of the First International Workshop on Hulled Wheats; Rome: International Plant Genetic Resources Institute; 1996. pp. 41–100.
13. Dvorak J, Akhunov ED. Tempos of gene locus deletions and duplications and their relationship to recombination rate during diploid and polyploid evolution in the Aegilops–Triticum alliance. Genetics. 2005;171:323–332. [PMC free article] [PubMed]
14. Devos KM, Beales J, Ogihara Y, Doust AN. Comparative sequence analysis of the phytochrome C gene and its upstream region in allohexaploid wheat reveals new data on the evolution of its three constituent genomes. Plant Mol Biol. 2005;58:625–641. [PubMed]
15. Crepet WL, Feldman GD. The earliest remains of grasses in the fossil record. Am J Bot. 1991;78:1010–1014.
16. Prasad V, Stromberg CA, Alimohammadian H, Sahni A. Dinosaur coprolites and the early evolution of grasses and grazers. Science. 2005;310:1177–1180. [PubMed]
17. Linder HP. The evolutionary history of the Poales/Restionales—A hypothesis. Kew Bull. 1986;42:297–318.
18. Dvorak J, Akhunov ED, Akhunov AR, Deal KR, Luo MC. Molecular characterization of a diagnostic DNA marker for domesticated tetraploid wheat provides evidence for gene flow from wild tetraploid wheat to hexaploid wheat. Mol Biol Evol. 2006;23:1386–1396. [PubMed]
19. Gaut BS. Molecular clocks and nucleotide substitution rates in higher plants. Evol Biol. 1998;30:93–120.
20. Bossolini E, Wicker T, Knobel PA, Keller B. Comparison of orthologous loci from small grass genomes Brachypodium and rice: implications for wheat genomics and grass genome annotation. Plant J. 2007;49:704–717. [PubMed]
21. Huang S, Su X, Haselkorn R, Gornicki P. Evolution of switchgrass (Panicum virgatum L.) based on the sequence of the nuclear gene encoding plastid acetyl-CoA carboxylase. Plant Sci. 2003;164:43–49.
22. Dubcovsky J, Dvorak J. Genome plasticity a key factor in the success of polyploid wheat under domestication. Science. 2007;316:1862–1866. [PubMed]
23. Kilian B, et al. Independent wheat B and G genome origins in outcrossing Aegilops progenitor haplotypes. Mol Biol Evol. 2007;24:217–227. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...