Comparative Analysis of Miscanthus and Saccharum Reveals a Shared Whole-Genome Duplication but Different Evolutionary Fates

Abstract

Multiple polyploidizations with divergent consequences in the grass subtribe Saccharinae provide a singular opportunity to study in situ adaptation of a genome to the duplicated state, heretofore known primarily from paleogenomics. We show that allopolyploidy in a common Miscanthus-Saccharum ancestor ∼3.8 to 4.6 million years ago closely coincides in time with their divergence from the Sorghum lineage. Subsequent Saccharum-specific autopolyploidy may have created pseudo-paralogous chromosome groups with random pairing within a group but infrequent pairing between groups. High chromosome number may reduce differentiation among Saccharum pseudo-paralogs by increasing opportunities for recombinations, with the lower chromosome numbers of Miscanthus favoring the return to disomic inheritance. The widespread tendency of plant chromosome numbers to recursively return to a narrow range following genome duplication appears to be occurring now in Saccharum spontaneum based on rich polymorphism for chromosome number among genotypes, with past reductions indicated by condensations of two ancestral chromosomes in Miscanthus (now n = 19) and perhaps as many as 10 in the Narenga-Sclerostachya clade (n = 15).

Figure 1.

Distribution of Synonymous Nucleotide Substitutions Based on Illumina Sequence Data. Sb, sorghum; M, Miscanthus; So, Saccharum; Z, maize. Ks value is shown on each peak. Numbers in parentheses indicate the number of exon pairs used for Ks plotting. (A) Ks distribution between Miscanthus paralogs and orthologs among Miscanthus, sorghum, and maize. (B) Ks distribution between paralogs and orthologs from Miscanthus and Saccharum.

Figure 2.

Evolutionary Models of Polypoidizations. An extended tree structure of Miscanthus (M), Saccharum (O), and sorghum (S) is displayed with the polyploidization common to the first two species shown by a star. Two genes linked by a horizontal line are assumed identical. Assumed times of events are shown on a time line besides the tree. Five different models representing allo- and autopolyploidizations are shown. Model A: if an allopolyploidization had occurred, how a tree of duplicated genes and corresponding Ks distributions between paralogs and those between orthologs would look. A pair of duplicated genes is shown by red and blue branches. After the Miscanthus-Saccharum split, each plant has two paralogs. Model B: if an autopolyploidization had occurred and Miscanthus chromosomes recovered diploid heredity shortly after its split from Saccharum. Double arrowed lines show normal recombination between duplicated genes produced by autoploidization, which retains identical duplicates. When Saccharum retained to be an autopolyploid, two Saccharum duplicates would be identical and could not be specified to show Ks distribution. Model C: if autopolyploidy was retained in both Miscanthus and Saccharum, the duplicates in both species would be identical. Model D: if Miscanthus regained diploid heredity long after its split from Saccharum. Model E: this is a modification to Model A by considering that further autopolyploidization occurred in Saccharum.

Figure 3.

A Species Tree of Related Grasses. Stars show polyploidy events and circles indicate projected diploidization events. A time line is shown in MYA and a Ks line is also provided. Chromosome numbers of a plant are shown along branches. *, divergence of maize and sorghum; †, divergence of subtribes Sorghinae and Saccharinae; ǂ, shared duplication event between Miscanthus and Saccharum; §, divergence of Miscanthus and Saccharum.

Figure 4.

Chromosome Numbers of Three Saccharum Species Summarized from Irvine (1999) Based on Possible Basal Chromosome Numbers. Original data reported a total of 398, 96, and 497 genotypes from S. spontaneum, S. robustum, and S. officinarum, respectively; however, in this summary, a total of 392, 49, and 462 genotypes are considered from S. spontaneum, S. robustum, and S. officinarum, respectively, due to ambiguity of aneuploidies. Numbers on top of bars indicate the number of genotypes. Dashed lines indicate the hypothetical ploidy members of the series that are missing. Note that 2n = 72 could be either x = 8 or x = 9. 2n = 40 and 80 chromosome types could be either x = 8 or 10. 2n = 120 is only considered as x = 10 due to its high odd-numbered ploidy.