Architecture and evolution of a minute plant genome

Abstract

It has been argued that the evolution of plant genome size is principally unidirectional and increasing owing to the varied action of whole-genome duplications (WGDs) and mobile element proliferation. However, extreme genome size reductions have been reported in the angiosperm family tree. Here we report the sequence of the 82-megabase genome of the carnivorous bladderwort plant Utricularia gibba. Despite its tiny size, the U. gibba genome accommodates a typical number of genes for a plant, with the main difference from other plant genomes arising from a drastic reduction in non-genic DNA. Unexpectedly, we identified at least three rounds of WGD in U. gibba since common ancestry with tomato (Solanum) and grape (Vitis). The compressed architecture of the U. gibba genome indicates that a small fraction of intergenic DNA, with few or no active retrotransposons, is sufficient to regulate and integrate all the processes required for the development and reproduction of a complex organism.

Figure 1. Syntenic analysis of the Utricularia gibba genome

a, Whole-genome duplication (WGD) history highlighting the phylogenetic position of U. gibba. Vitis, Arabidopsis and Carica papaya are rosids; Arabidopsis has had two WGDs since the paleohexaploid (Phex) core eudicot ancestor. Tomato (Solanum), Mimulus and U. gibba are asterids; tomato has a mix of duplicated and triplicated regions; U. gibba has had three WGDs since common ancestry with tomato and the Phex ancestor. Mimulus has had a single WGD that may also be the most ancient WGD observed for U. gibba (see Supplementary Information section 7.1.3). U. gibba flowers are similar to those of Mimulus (that is, like snapdragons); tiny suction traps are borne on highly divided branching structures (insets, clockwise from left). b, A microsyntenic analysis shows that U. gibba (U) is 8:2:1 relative to homologous tomato (T) and Vitis (V) regions, respectively. As such, U. gibba is a 16-ploid with respect to Vitis, and the polyploidy of tomato is entirely independent (Supplementary Information section 7). Coloured lines connect high-scoring segment pairs (HSPs) on genome blocks masked for non-coding sequences. Gene models lie in the centres of each block, below the HSPs. This analysis may be regenerated by CoGe at http://genomeevolution.org/r/4wvh. c, Fractionation in a given U. gibba region can be massive with respect to tomato; the regions shown include an over 3 Mb block of the tomato genome (top), strongly syntenic and colinear to an approximately 130-kb block of U. gibba, representing an approximately 20:1 difference in total DNA. This analysis may be regenerated by CoGe at http://genomevolution.org/r/5cet.

Figure 2. Architecture of the Utricularia gibba genome

a, U. gibba gene islands are more compact than in Arabidopsis, and much higher in gene density than tomato or grape. For example, the Arabidopsis LEAFY gene lies directly in the middle of the second block from the top, which is an approximately 100-kb region from Arabidopsis chromosome 5. There are 28 genes in this view. In the corresponding U. gibba block (top), there are 34 genes within the same-sized region, which is therefore approximately 18% more densely packed. In tomato (3rd block) and grape (4th), there are many fewer genes (14 and 17, respectively) for a much lower density of gene space. b, Promoter spaces in U. gibba can be very short. Shown is part of a scaffold (scf00089), the sequence of which was verified by PCR walking. Four promoter regions (blue) showed reproducible activity in transient expression experiments (see Supplementary Information section 3). For example, the short bidirectional promoter between a divergent gene pair is approximately 400 bp. Other gene arrangements, tandem and convergent, can be seen in this example. c, Solo LTR remains of ectopically recombined mobile elements can be identified in the U. gibba genome. This example shows two blocks from U. gibba, the Solo LTR in the bottom block being homologous to the LTR pair present in the top block. In a, syntenic HSPs are shown as coloured lines connecting particular gene models (purple). Results from a and c can be regenerated at http://genomevolution.org/r/5kv5 and http://genomevolution.org/r/8lvv, respectively. See Supplementary Information for further discussion of b and c.

Figure 3. A model of genome size reduction and the plant genome size evolutionary spectrum

a, The initial diploid genome has 10 genes. b, c, After one WGD (b), there are 20 genes in the tetraploid, which fractionate into 16 genes (c). d–g, After another round of WGD (d), the octoploid genome (32 genes) fractionates again to yield 16 genes (e), which duplicate (to 32 genes) in yet another WGD (f), after which fractionation yields 16 genes in the 16-ploid (g). The resulting number of genes is the same as in the fractionated genome resulting from the first WGD (c), with only 6 more genes than the original diploid ancestor (a). h, The resulting genome after intergenic DNA contraction at any stage (a–g) has thus survived a high deletion rate via the net accrual of very few gene duplicates following sequential WGDs. U. gibba has in fact fractionated down to single copy two-thirds of its genes syntenic to tomato genes since its three WGDs. i, An interplay of deletion and retroelement proliferation rates relates to a continuum of plant genome size evolution, with WGDs providing short-term buffering against loss of crucial gene functions in small genomes affected by high endogenous deletion rates. Small genomes result when the recombinational deletion rate is high relative to retroelement proliferation and WGD, vice versa with large genomes.