Organellar transcripts dominate the cellular mRNA pool across plants of varying ploidy levels

Abstract

Mitochondrial and plastid functions depend on coordinated expression of proteins encoded by genomic compartments that have radical differences in copy number of organellar and nuclear genomes. In polyploids, doubling of the nuclear genome may add challenges to maintaining balanced expression of proteins involved in cytonuclear interactions. Here, we use ribo-depleted RNA-seq to analyze transcript abundance for nuclear and organellar genomes in leaf tissue from four different polyploid angiosperms and their close diploid relatives. We find that, even though plastid genomes contain <1% of the number of genes in the nuclear genome, they generate the majority (69.9 - 82.3%) of mRNA transcripts in the cell. Mitochondrial genes are responsible for a much smaller percentage (1.3 - 3.7%) of the leaf mRNA pool but still produce much higher transcript abundances per gene compared to nuclear genome. Nuclear genes encoding proteins that functionally interact with mitochondrial or plastid gene products exhibit mRNA expression levels that are consistently more than ten-fold lower than their organellar counterparts, indicating an extreme cytonuclear imbalance at the RNA level despite the predominance of equimolar interactions at the protein level. Nevertheless, interacting nuclear and organellar genes show strongly correlated transcript abundances across functional categories, suggesting that the observed mRNA stoichiometric imbalance does not preclude coordination of cytonuclear expression. Finally, we show that nuclear genome doubling does not alter the cytonuclear expression ratios observed in diploid relatives in consistent or systematic ways, indicating that successful polyploid plants are able to compensate for cytonuclear perturbations associated with nuclear genome doubling.

Variation in cytonuclear expression accommodation among allopolyploid plants

“…Because the aim of this study was to characterize cytonuclear accommodation at the level of gene expression in polyploid species, total RNA was extracted for each accession and ribodepletion was used to remove ribosomal RNAs, circumventing the bias of polyA-selection protocols that exclude some organellar transcripts ( Slomovic et al 2006 , 2008 ; Smith 2013 ). As expected, transcripts from the organelles were abundant ( Supplementary Table 3 and analyzed in Forsythe et al 2022 ); however, sufficient nuclear transcriptome coverage was achieved, ranging from 4 to 26 M reads per sample (mean reads per sample were Arabidopsis = 13 M, Arachis = 11 M, Chenopodium = 7 M, Gossypium = 20 M). One replicate each for A. hypogaea and Arachis IpaDur1 was removed due to low overall mapping rates (i.e.…”

“…Allopolyploids face a complex array of challenges stemming both from whole-genome duplication and from hybridization of divergent genomes. These challenges include maintaining stoichiometric balance among interacting molecules ( Birchler and Veitia 2010 , 2012 , 2014 , 2021 ; Forsythe et al 2022 ), which may be even more problematic for interactions between the biparentally inherited, organelle-targeted genes and those occurring in the maternally coevolved organelles ( Wolf and Hager 2006 ; Sharbrough et al 2017 ). These potential cytonuclear incompatibilities may underlie observations of rapid and repeated return to single copy for organelle-targeted genes in polyploid species ( De Smet et al 2013 ; Li et al 2016 ) and the expectation that any paternal cytonuclear homoeologs that exhibit deleterious interactions should evolve rapidly when not immediately lost ( Rand et al 2004 ; Bock et al 2014 ; Sloan et al 2014 ).…”

The evolutionary constraints on angiosperm chloroplast adaptation

“…Relationship between the rate of molecular evolution and gene expression for plastid encoded protein-coding genes. A-C ) Scatterplots of mRNA abundance (data from (Forsythe, et al 2022)) versus the rate of nucleotide evolution, protein evolution and protein sequence constraint, respectively. Where appropriate, simple linear regressions are shown as black lines with their associated R 2 and P -values.…”

“…The GC 3 excluded methionine and tryptophan codons as these amino acids are encoded by a single codon, and thus have no flexibility in base identity at the 3 rd position. mRNA abundance and protein abundance mRNA abundance data were obtained from (Forsythe, et al 2022) and values were available for all 69 genes analysed in this study. Integrated protein abundance data for Arabidopsis thaliana with a proteome coverage of 76 % was obtained from the Protein Abundance Database (PaxDb) (Wang, et al 2012).…”

“…A Welch's t-test identified no significant difference between the two categories (P > 0.05). (Forsythe, et al 2022)) versus the rate of nucleotide evolution, protein evolution and protein sequence constraint, respectively. Where appropriate, simple linear regressions are shown as black lines with their associated R 2 and P-values.…”