Unexpected consequences of a sudden and massive transposon amplification on rice gene expression

High-copy-number transposable elements comprise the majority of eukaryotic genomes where they are major contributors to gene and genome evolution(1). However, it remains unclear how a host genome can survive a rapid burst of hundreds or thousands of insertions because such bursts are exceedingly rare in nature and therefore difficult to observe in real time(2). In a previous study we reported that in a few rice strains the DNA transposon mPing was increasing its copy number by similar to 40 per plant per generation(3). Here we exploit the completely sequenced rice genome to determine 1,664 insertion sites using high-throughput sequencing of 24 individual rice plants and assess the impact of insertion on the expression of 710 genes by comparative microarray analysis. We find that the vast majority of transposable element insertions either upregulate or have no detectable effect on gene transcription. This modest impact reflects a surprising avoidance of exon insertions by mPing and a preference for insertion into 5' flanking sequences of genes. Furthermore, we document the generation of new regulatory networks by a subset of mPing insertions that render adjacent genes stress inducible. As such, this study provides evidence for models first proposed previously(4-6) for the involvement of transposable elements and other repetitive sequences in genome restructuring and gene regulation.