Genome evolution and initial breeding of the Triticeae grass Leymus chinensis dominating the Eurasian Steppe

Abstract

Leymus chinensis, a dominant perennial grass in the Eurasian Steppe, is well known for its remarkable adaptability and forage quality. Hardly any breeding has been done on the grass, limiting its potential in ecological restoration and forage productivity. To enable genetic improvement of the untapped, important species, we obtained a 7.85-Gb high-quality genome of L. chinensis with a particularly long contig N50 (318.49 Mb). Its allotetraploid genome is estimated to originate 5.29 million years ago (MYA) from a cross between the Ns-subgenome relating to Psathyrostachys and the unknown Xm-subgenome. Multiple bursts of transposons during 0.433-1.842 MYA after genome allopolyploidization, which involved predominantly the Tekay and Angela of LTR retrotransposons, contributed to its genome expansion and complexity. With the genome resource available, we successfully developed a genetic transformation system as well as the gene-editing pipeline in L. chinensis. We knocked out the monocot-specific miR528 using CRISPR/Cas9, resulting in the improvement of yield-related traits with increases in the tiller number and growth rate. Our research provides valuable genomic resources for Triticeae evolutionary studies and presents a conceptual framework illustrating the utilization of genomic information and genome editing to accelerate the improvement of wild L. chinensis with features such as polyploidization and self-incompatibility.

Keywords: LTR retrotransposons; Leymus chinensis; forage crop; genetic breeding; genome evolution.

Fig. 1.

Morphological view and genomic profile of L. chinensis. (A) An image of L. chinensis growing on the steppe showing its robust rhizome. (B) Overview of the assembled L. chinensis genome including (i) chromosome name and length, (ii) GC content, (iii) density of genes in 2-Mb windows, (iv) distribution of transposable elements (TEs) in 2-Mb windows, and (v) the intragenomic synteny profile between homologous chromosome pairs. Synteny blocks with >10 genes are shown. (C) Karyotype analysis in a L. chinensis root tip cell. (D) Principal component analysis (PCA) based on subgenome-specific k-mers. (E) Resequencing data from Psathyrostachys juncea were mapped on the L. chinensis genome such that Ns and Xm could be distinguished by the coverage of each chromosome.

Fig. 2.

Genome evolution and the tetraploidization event of L. chinensis. (A) Phylogenetic relationships of L. chinensis with 10 other Poaceae species. Green background: Triticeae species; orange: Brachypodieae species as represented by B. distachyon; blue: Oryzeae species as represented by O. sativa. Divergence times are indicated at each node. (B) Distribution of 4DTv distances between orthologous genes. (C) Genomic synteny analysis based on collinear gene blocks across the chromosomes of L. chinensis. The chromosomal rearrangements are indicated by dotted lines. (D) Synteny between chromosomes 4 and 5 between L. chinensis and other Triticeae species. (E and F) Boxplot of TE density (E) and gene number (F) between subgenomes using a 1-Mb sliding window. (G) Boxplot of the distribution of Ka/Ks values between subgenomes.

Fig. 3.

Expansion and evolutionary analysis of LTR retrotransposons in the L. chinensis genome. (A) Genome size expansion was highly correlated with TE amplification bursts. The red dashed line shows the linear relationship between genome size and TE content. (B) Comparisons of TE components among genome-sequenced Triticeae plants. (C) Bursts of Gypsy and Copia elements in the Ns- and Xm-subgenome of L. chinensis. (D) Phylogenetic tree of the LTR retrotransposon families in L. chinensis. (E) Number of different family members in the Ns-subgenome (Upper) and Xm-subgenome (Lower).

Fig. 4.

The induced genetic modification of L. chinensis with a genome editing system. (A) Callus induction and transgenic plant regeneration from a young panicle of L. chinensis Lc6-5. (B) The conservation of miRNA families identified in L. chinensis and sequenced land plant species deposited in miRBase. The green and blue squares indicate the present families in monocots and dicots, respectively. The white squares indicate the absent ones. The red line frames the miRNA families found only in monocots. (C) The expression levels of miRNA members from monocot-conserved families in leaves of L. chinensis. (D) Genome editing of MIR528 in Lc6-5. The target site is shown in a blue font; the underlined nucleotides indicate the protospacer-adjacent motif sequence. Deletions and insertions are indicated by dashes and red letters, respectively, except for dashes in the sequences of wild type (WT). (E) Detection of mature miR528 abundance by northern blot analysis in a negative sibling control (NC) and miR528 knock-out mutants. The U6 small nuclear RNA was monitored as a loading control. (F) Phenotype of T₀ transformants of the NC and mir528 mutants (Scale bars, 10 cm). (G) The tiller number of NC and mir528 mutants. Values are shown as means ± SD. (H) Phenotype of the L. chinensis transplants from T₀ transformants of NC and mir528 mutants at 0 d, 1 mo, and 3 mo after transplanting (Scale bars, 10 cm).