Down, then up: non-parallel genome size changes and a descending chromosome series in a recent radiation of the Australian allotetraploid plant species, Nicotiana section Suaveolentes (Solanaceae)

Abstract

Background and aims: The extent to which genome size and chromosome numbers evolve in concert is little understood, particularly after polyploidy (whole-genome duplication), when a genome returns to a diploid-like condition (diploidization). We study this phenomenon in 46 species of allotetraploid Nicotiana section Suaveolentes (Solanaceae), which formed <6 million years ago and radiated in the arid centre of Australia.

Methods: We analysed newly assessed genome sizes and chromosome numbers within the context of a restriction site-associated nuclear DNA (RADseq) phylogenetic framework.

Key results: RADseq generated a well-supported phylogenetic tree, in which multiple accessions from each species formed unique genetic clusters. Chromosome numbers and genome sizes vary from n = 2x = 15 to 24 and 2.7 to 5.8 pg/1C nucleus, respectively. Decreases in both genome size and chromosome number occur, although neither consistently nor in parallel. Species with the lowest chromosome numbers (n = 15-18) do not possess the smallest genome sizes and, although N. heterantha has retained the ancestral chromosome complement, n = 2x = 24, it nonetheless has the smallest genome size, even smaller than that of the modern representatives of ancestral diploids.

Conclusions: The results indicate that decreases in genome size and chromosome number occur in parallel down to a chromosome number threshold, n = 20, below which genome size increases, a phenomenon potentially explained by decreasing rates of recombination over fewer chromosomes. We hypothesize that, more generally in plants, major decreases in genome size post-polyploidization take place while chromosome numbers are still high because in these stages elimination of retrotransposons and other repetitive elements is more efficient. Once such major genome size change has been accomplished, then dysploid chromosome reductions take place to reorganize these smaller genomes, producing species with small genomes and low chromosome numbers such as those observed in many annual angiosperms, including Arabidopsis.

Keywords: Nicotiana benthamiana; Nicotiana sect. Suaveolentes; Allotetraploid evolution; Australian endemics; C-value; Solanaceae; WGD; diploidization; dysploidy; epigenetics; model organism; polyploidy.

Fig. 1.

Global distribution of Nicotiana with distribution of N. sect. Suaveolentes in red and that of the New World species in yellow. Map base from Wikimedia Commons.

Fig. 2.

Co-ancestry heatmap3, constructed based on genotype likelihoods. Darker tones represent higher pairwise relatedness; estimates on the diagonal have been excluded.

Fig. 3.

RADseq tree, sub-tree A (top) and sub-tree B (bottom). RAxML-derived phylogenetic tree based on 457 382 single nucleotide polymorphisms (SNPs).

Fig. 3.

RADseq tree, sub-tree A (top) and sub-tree B (bottom). RAxML-derived phylogenetic tree based on 457 382 single nucleotide polymorphisms (SNPs).

Fig. 4.

Present state and ancestral reconstruction of chromosome number and genome size. (A) The summary tree of species relationships (with locality names for undescribed new species, as in Fig. 3), as estimated with ChromEvol with chromosome number increases prohibited. (B) Genome sizes (triangles, upper x-scale) and haploid chromosome number (circles, lower X-scale). (C) Genome size evolution as estimated using BayesTree using the summary tree of species relationships. Chromosome number according to the embedded colour legend.

Fig. 5.

Scatter plots showing the relationships between (A) genome size and chromosome number; and (B) genome size and chromosome number using phylogenetic independent contrasts (PICs). The dashed trend line indicates the estimated slope from a linear regression, which is not significant in either (A) or (B).