Analysis of the giant genomes of Fritillaria (Liliaceae) indicates that a lack of DNA removal characterizes extreme expansions in genome size

Abstract

Plants exhibit an extraordinary range of genome sizes, varying by > 2000-fold between the smallest and largest recorded values. In the absence of polyploidy, changes in the amount of repetitive DNA (transposable elements and tandem repeats) are primarily responsible for genome size differences between species. However, there is ongoing debate regarding the relative importance of amplification of repetitive DNA versus its deletion in governing genome size. Using data from 454 sequencing, we analysed the most repetitive fraction of some of the largest known genomes for diploid plant species, from members of Fritillaria. We revealed that genomic expansion has not resulted from the recent massive amplification of just a handful of repeat families, as shown in species with smaller genomes. Instead, the bulk of these immense genomes is composed of highly heterogeneous, relatively low-abundance repeat-derived DNA, supporting a scenario where amplified repeats continually accumulate due to infrequent DNA removal. Our results indicate that a lack of deletion and low turnover of repetitive DNA are major contributors to the evolution of extremely large genomes and show that their size cannot simply be accounted for by the activity of a small number of high-abundance repeat families.

Keywords: DNA deletion; Fritillaria; Liliaceae; genome size evolution; genome turnover; repetitive DNA; transposable elements (TEs).

Figure 1

Ancestral genome size reconstruction and evidence for genome expansion in Fritillaria. Phylogenetic relationships between species of Fritillaria and related genera are shown; values above branches indicate node support (posterior probabilities of ≥ 0.95/bootstrap percentages ≥ 70). Ancestral genome sizes for the most recent common ancestor (MRCA) of each major Fritillaria clade are shown; 95% confidence intervals are given in parentheses. Closed circles indicate monoploid genome size (1Cx‐values in Gb) for extant species; dashed lines indicate the ancestral genome sizes for the MRCA of the F. affinis (blue) and F. imperialis (red) clades. For each species of Fritillaria, the increase or decrease in genome size relative to the MRCA of its clade is indicated.

Figure 2

Cumulative abundance of the most common repeat families from Fritillaria affinis and Fritillaria imperialis. For each species, the abundance in their genome of the top repeat families identified from F. affinis (upper bar) and F. imperialis (lower bar) is shown in megabases (Mb). Repeat families are ordered from left to right according to their abundance in F. affinis (upper bar in each pair) and F. imperialis (lower bar in each pair) and coloured according to repeat type; LTR, long terminal repeat retrotransposon. The summary of relationships between the 10 species is derived from the phylogenetic tree shown in Fig. 1.

Figure 3

Intrafamily heterogeneity of repeats in Fritillaria. (a) Histogram of average edge weights from graphs of all top repeat families from Fritillaria affinis (n = 47). (b–e) Histograms of percentage sequence similarity for read pairs from selected repeat families representing a range of different edge weights from F. affinis, illustrating that repeat families with average edge weights of < 450, which comprise the vast majority of the top families, show an absence of peaks of very high similarity read pairs (i.e. ≥ 98% sequence similarity). (f) Histogram of average edge weights from graphs of all top repeat families from F. imperialis (n = 41). (g–j) Histograms of percentage sequence similarity for read pairs from selected repeat families representing a range of different edge weights from F. imperialis, showing a similar pattern to that described above for F. affinis. Cluster names and repeat types follow those listed in Supporting Information Tables S5 and S6; see Notes S2 for further explanation.