Spatial and Temporal Transcriptomic Heredity and Asymmetry in an Artificially Constructed Allotetraploid Wheat (AADD)

Xintong Ma, Zhibin Zhang, Li Guo, Xiaowan Gou, Yao Bian, Yue Zhao, Bin Wang, Man Lang, Tianya Wang, Kun Xie, Xiaoming Liu, Liu B, Lei GongFrontiers in Plant Science 2022 · DOI

TL;DR

This study investigates how allopolyploidy, the merging of two distinct genomes, impacts gene expression in newly formed wheat. They found that while parental differences are crucial, allopolyploidy itself causes significant and asymmetric gene expression changes, particularly in the D subgenome, influencing traits like spike development.

What Problem Does This Paper Address?

Polyploidy, or whole-genome duplication, is a major driver of plant evolution, leading to new species with novel traits. Allopolyploidy, specifically, involves the hybridization of two different species followed by genome doubling. While it's known that this "transcriptome shock" dramatically alters gene expression, there's a significant knowledge gap regarding how these changes manifest across different tissues and developmental stages in the initial stages of allopolyploid formation. Understanding these early, dynamic responses is crucial for comprehending how new polyploid species establish their genetic and phenotypic identity. This paper aims to systematically characterize these immediate transcriptomic consequences in an artificially created allotetraploid wheat.

Methods

The researchers employed a multi-faceted approach:

  • RNA Sequencing (RNA-seq): Comprehensive sequencing was performed on 81 samples, covering three genotypes (two diploid parents and one synthetic allotetraploid), nine different tissues/developmental stages, and three biological replicates. This generated a massive dataset (~390 Gb) to capture a broad view of gene expression.
  • Principal Component Analysis (PCA) and Hierarchical Clustering: These standard bioinformatics techniques were used to visualize and group samples based on their overall gene expression profiles, revealing major drivers of variation (e.g., tissue type, parental legacy).
  • Differential Gene Expression (DEG) Analysis: Using DESeq2, they identified genes with significantly altered expression levels between the allotetraploid and its parental diploids.
  • Homoeolog Expression Bias (HEB) Analysis: This method quantifies the preferential expression of one homoeolog (a gene copy from a different subgenome) over the other. They classified HEB into parental legacy, convergent, and divergent categories.
  • Homoeolog-Specific Expression (HSE) Analysis: This technique distinguishes between cis-regulatory (gene-intrinsic) and trans-regulatory (external factors) contributions to expression differences between homoeologs.
  • Weighted Gene Co-expression Network Analysis (WGCNA): WGCNA was used to build networks of co-expressed genes and identify modules of genes with similar expression patterns. This allowed them to investigate homoeolog expression connectivity (HEC) and how network structures change after allopolyploidy.

Datasets

The study utilized data from:

  • Synthetic allotetraploid wheat (AADD genome, accession AT2): This is the primary subject of the study, representing a newly formed allopolyploid.
  • Diploid progenitors: Triticum urartu (AA genome, accession TMU38) and Aegilops tauschii (DD genome, accession TQ27).
  • In silico parental mix: Created by combining the RNA-seq data of the diploid parents to represent their baseline expression differences.

No new benchmark datasets were introduced; rather, the study generated a comprehensive dataset from a controlled experimental system.

Key Insights and Ideas

  • Parental Legacy is Paramount, but Allopolyploidy Adds Complexity: The study highlights that the inherent differences between the diploid parents significantly shape the expression landscape of the new allotetraploid. However, allopolyploidy itself introduces its own dramatic and asymmetric changes, not just a simple sum of parental contributions.
  • The D Subgenome is More Responsive: Genes in the D subgenome showed a more drastic response to allopolyploidy compared to the A subgenome, suggesting differential evolutionary plasticity or regulatory mechanisms.
  • Asymmetric Gene Expression Drives Phenotypic Similarity: The phenotypic similarity of the allotetraploid AT2 to one of its parents (T. urartu) is linked to a combination of stable A-subgenome gene expression and significant downregulation of its D-subgenome homoeologs, particularly in spike development.
  • Cis-Regulatory Divergence is Key for HSE: Homoeolog-specific expression patterns are largely dominated by cis-only regulation, indicating that changes in the DNA sequence of regulatory elements (like promoters) in the parental genomes are a primary driver of expression differences in the allotetraploid.
  • Network Rewiring Reflects Regulatory Changes: Allopolyploidy leads to the reconstruction of gene co-expression networks, and these changes are correlated with sequence divergence in cis elements, suggesting that regulatory evolution is a dynamic process in newly formed polyploids.

Key Results

  • Differential Gene Expression: Thousands of differentially expressed genes (DEGs) were identified between the allotetraploid and its parents. The D subgenome exhibited a higher proportion of downregulated DEGs, especially in spike tissues.
  • Homoeolog Expression Bias (HEB): While most HEB patterns were inherited from parents, a significant portion showed de novo generation, primarily driven by asymmetrical differential gene expression. The D subgenome showed a higher proportion of biased homoeologs.
  • Homoeolog-Specific Expression (HSE): The cis-only regulatory pattern was predominant, indicating that parental cis-element divergence is a major factor in HSE.
  • Homoeolog Expression Connectivity (HEC): HEC was significantly correlated with sequence divergence in cis elements, suggesting that changes in regulatory sequences influence how homoeologs are co-expressed.
  • Spike Development: A transcriptome atlas of spike development revealed that the phenotypic similarity of AT2 to T. urartu is likely due to stable A-subgenome gene expression and drastic downregulation of D-subgenome homoeologs, particularly for genes involved in cell differentiation.

Limitations and Future Work

  • Focus on Early Stages: The study focuses on the immediate effects of allopolyploidy. Long-term evolutionary changes and stability over many generations are not addressed.
  • Artificial System: While valuable for controlled study, the findings from an artificially constructed allotetraploid may not perfectly mirror the complex evolutionary trajectories of natural allopolyploids.
  • Future Directions: The authors suggest that future research should explore subgenome dominance by accounting for inherent parental divergence. They also propose using advanced techniques like in situ assays and CRISPR-Cas9 genome editing to further dissect the molecular regulatory pathways of spike development and identify candidate genes for yield improvement. The role of epigenetic modifications in regulating homoeolog expression bias also warrants further investigation.

Summary updated

Abstract

Polyploidy, or whole-genome duplication (WGD), often induces dramatic changes in gene expression due to "transcriptome shock. " However, questions remain about how allopolyploidy (the merging of multiple nuclear genomes in the same nucleus) affects gene expression within and across multiple tissues and developmental stages during the initial foundation of allopolyploid plants. Here, we systematically investigated the immediate effect of allopolyploidy on gene expression variation in an artificial allopolyploidy system consisting of a constructed allotetraploid wheat (AADD genome, accession AT2) and its diploid progenitors <i>Triticum urartu</i> and <i>Aegilops tauschii</i>. We performed comprehensive RNA sequencing of 81 samples from different genotypes, tissues, and developmental stages. First, we found that intrinsic interspecific differences between the diploid parents played a major role in establishing the expression architecture of the allopolyploid. Nonetheless, allopolyploidy <i>per se</i> also induced dramatic and asymmetric patterns of differential gene expression between the subgenomes, and genes from the D subgenome exhibited a more drastic response. Second, analysis of homoeolog expression bias (HEB) revealed that the D subgenome exhibited significant expression bias and that <i>de novo</i>-generated HEB was attributed mainly to asymmetrical differential gene expression. Homoeolog-specific expression (HSE) analyses showed that the <i>cis</i>-only regulatory pattern was predominant in AT2, reflecting significant divergence between the parents. Co-expression network analysis revealed that homoeolog expression connectivity (HEC) was significantly correlated with sequence divergence in <i>cis</i> elements between subgenomes. Interestingly, allopolyploidy-induced reconstruction of network modules was also associated with different HSE patterns. Finally, a transcriptome atlas of spike development demonstrated that the phenotypic similarity of AT2 to <i>T. urartu</i> may be attributed to the combination of relatively stable expression of A-subgenome genes and drastic downregulation of their D-subgenome homoeologs. These findings provide a broad, multidimensional characterization of allopolyploidy-induced transcriptomic responses and suggest that allopolyploidy can have immediate and complex regulatory effects on the expression of nuclear genes.
BibTeX
@inproceedings{ma2022spatial,
  author = {Xintong Ma and Zhibin Zhang and Li Guo and Xiaowan Gou and Yao Bian and Yue Zhao and Bin Wang and Man Lang and Tianya Wang and Kun Xie and Xiaoming Liu and Liu B and Lei Gong},
  title = {{Spatial and Temporal Transcriptomic Heredity and Asymmetry in an Artificially Constructed Allotetraploid Wheat (AADD)}},
  booktitle = {Frontiers in Plant Science},
  year = {2022},
  doi = {10.3389/fpls.2022.887133},
}
Figure