Logo: to the web site of Uppsala University

Allopolyploidization, the combination of whole-genome duplication (WGD) and interspecific hybridization, is a frequent and influential event in plant evolution. Allopolyploidization potentially affects both adaptation and diversification, yet the understanding of the consequences of allopolyploidy has been obscured by several issues. First, a broadly defined phenomenon, “genomic shock”, is assumed to be a general instant outcome of allopolyploidization, but this assumption has been challenged by recent studies. Second, effects of WGD and interspecific hybridization are confounded in allopolyploidization, which hinders the understanding of the specific effects of either component. Third, in natural allopolyploid species, instant and long-term effects of allopolyploidization are mixed, masking the evolution trajectory of allopolyploid genomes. 

To address these issues, we studied the outcome of allopolyploidization in Capsella bursa-pastoris, a 100,000-year-old natural allotetraploid species. C. bursa-pastoris is a self-fertilizing weed with a worldwide distribution that originated through the hybridization between two diploid Capsella species with distinct mating systems. First, we investigated gene expression pattern in natural C. bursa-pastoris accessions with DNA- and  RNA-sequencing data. Next, we resynthesized C. bursa-pastoris-like allotetraploids, along with diploid hybrids and autotetraploids. Phenotype and gene expression patterns were compared among those synthetic Capsella plants and natural C. bursa-pastoris to (i) distinguish the instant effects of hybridization from WGD and (ii) tell apart instant effects from long-term ones.

In general, non-additive gene expression was limited in both natural and resynthesized C. bursa-pastoris. We found the original TE-mediated genomic shock hypothesis did not fit the consequences of allopolyploidization in Capsella. Instead, homoeolog expression bias and the limited non-additive gene expression in resynthesized can be better explained by homoeologous exchanges and the intergenomic interaction of regulatory elements. 

The relative gene expression pattern in resynthesized C. bursa-pastoris was mainly determined by hybridization, not WGD, but WGD still significantly affected phenotypes, likely through altering cell-size-related physical attributes. Both WGD and hybridization decrease the quality of pollen and seeds, but the two events were less deleterious when combined. In addition, the breakdown of self-incompatibility in Capsella could not be induced by pure WGD but was caused by the dominant interactions between S-alleles in hybrids.

Both gene expression patterns and phenotypes of C. bursa-pastoris were largely reshaped by long-term evolution. Almost all the transgressive gene expressions were unique to natural C. bursa-pastoris. Similarly, selfing syndrome and improvement of pollen and seed quality were likely acquired through long-term evolution. Compared to resynthesized allotetraploids, natural C. bursa-pastoris had more expression-level dominance toward the self-fertilizing parent, especially in flowers, mirroring a pronounced selfing syndrome. Nonetheless, the instant effect of allopolyploidization did contribute to gene expression patterns, as about 40% of expression level dominance in natural C. bursa-pastoris can already be found in resynthesized allotetraploids.