Target site specificity of the Tos17 retrotransposon shows a preference for insertion within genes and against insertion in retrotransposon-rich regions of the genome

Abstract

Because retrotransposons are the major component of plant genomes, analysis of the target site selection of retrotransposons is important for understanding the structure and evolution of plant genomes. Here, we examined the target site specificity of the rice retrotransposon Tos17, which can be activated by tissue culture. We have produced 47,196 Tos17-induced insertion mutants of rice. This mutant population carries approximately 500,000 insertions. We analyzed >42,000 flanking sequences of newly transposed Tos17 copies from 4316 mutant lines. More than 20,000 unique loci were assigned on the rice genomic sequence. Analysis of these sequences showed that insertion events are three times more frequent in genic regions than in intergenic regions. Consistent with this result, Tos17 was shown to prefer gene-dense regions over centromeric heterochromatin regions. Analysis of insertion target sequences revealed a palindromic consensus sequence, ANGTT-TSD-AACNT, flanking the 5-bp target site duplication. Although insertion targets are distributed throughout the chromosomes, they tend to cluster, and 76% of the clusters are located in genic regions. The mechanisms of target site selection by Tos17, the utility of the mutant lines, and the knockout gene database are discussed. --The nucleotide sequence data were uploaded to the DDBJ, EMBL, and GenBank nucleotide sequence databases under accession numbers AG020727 to AG025611 and AG205093 to AG215049.

Figure 1.

Base Preferences of Tos17 Insertion Sites. Average base preferences at each position were calculated based on flanking sequences at 20,458 loci. From position 1 to position 5 is the TSD sequence. Numbers with minus and plus signs are base numbers upstream and downstream, respectively, from the TSD. The percentages of A (green), C (blue), G (black), and T (red) at each position were plotted.

Figure 2.

Frequency Distribution of the GC Content of Tos17-Inserted Regions and Published Rice Genomic Sequences. The percentages of GC for 1-kb windows centered on 20,458 insertion points were calculated, and the frequency distribution was plotted in red. For continuous rice genomic sequences, derived from PAC or BAC clones with at least 70 kb of sequence, GC contents were calculated with a 1-kb sliding window, and the frequency distribution was plotted in blue. To determine the GC content distribution of CDS, all annotated CDS sequences were first joined into one large sequence. This sequence then was split into 1-kb pieces to determine the percentage of GC and the frequency distribution (green line). To generate a frequency distribution for protein kinase and defense-related genes (pink line), each 1-kb fragment of the rice genomic sequences was searched by BLASTX against an amino acid data set containing records of protein kinase and disease resistance genes. Fragments showing matches with E values of

Figure 3.

Tos17 Insertion Map on Contigs of Rice Chromosome 1. Contigs were assembled through a combination of BLASTN searches and the EMBOSS megamerger program. Seven remaining gaps between the contigs were joined simply. The x axes represent the position on chromosome 1. The y axes show the frequency of Tos17 insertions, retrotransposons, and genes. (A) Tos17 insertions. (B) Protein kinase genes. (C) Disease resistance–related genes. (D) Retrotransposons. (E) Rice ESTs. Frequencies were calculated at 100-kb intervals along chromosome 1. The position of the centromere is shown with a box. For (B), (C), and (D), the joined sequence of chromosome 1 was split into 1-kb fragments. Each fragment was searched with BLASTX against an amino acid data set of protein kinase genes, disease resistance genes, and retrotransposons, respectively. Gene frequencies represent the number of 1-kb fragments that have matches with E values of <e-10 for BLASTX or <e-100 for BLASTN.

Figure 4.

Estimation of the Window Sizes of Tos17 Hot Spots. All insertion points were mapped onto public PAC and BAC sequences. The distances between adjacent Tos17 insertion points on the rice genomic sequences were obtained, and the frequency distribution of the distances was plotted in 100-bp increments. Each distance was calculated from continuous genomic sequences of at least 70 kb. Red, green, and blue bars indicate Tos17 insertions in coding, intergenic, and unknown regions, respectively.