Transposon diversity in Arabidopsis thaliana

Abstract

Recent availability of extensive genome sequence information offers new opportunities to analyze genome organization, including transposon diversity and accumulation, at a level of resolution that was previously unattainable. In this report, we used sequence similarity search and analysis protocols to perform a fine-scale analysis of a large sample ( approximately 17.2 Mb) of the Arabidopsis thaliana (Columbia) genome for transposons. Consistent with previous studies, we report that the A. thaliana genome harbors diverse representatives of most known superfamilies of transposons. However, our survey reveals a higher density of transposons of which over one-fourth could be classified into a single novel transposon family designated as Basho, which appears unrelated to any previously known superfamily. We have also identified putative transposase-coding ORFs for miniature inverted-repeat transposable elements (MITEs), providing clues into the mechanism of mobility and origins of the most abundant transposons associated with plant genes. In addition, we provide evidence that most mined transposons have a clear distribution preference for A + T-rich sequences and show that structural variation for many mined transposons is partly due to interelement recombination. Taken together, these findings further underscore the complexity of transposons within the compact genome of A. thaliana.

Figure 1

RESites corresponding to mined Arabidopsis elements. (A) Examples of RESites for different groups of mined elements. All of the RESites illustrated were identified by computer-assisted database searches. In total, 34 RESites were found by computer-assisted database searches, and 13 were identified by cross-ecotype PCR analysis. The target sequences are underlined, and the TSDs are shaded. GenBank geninfo identifier (gi) numbers and nucleotide position on clones are indicated. *, Inserted into a Basho III element; †, inserted into a Basho III element; ‡, inserted into a MITE IX element. (B) RESites found for Basho insertions confirm mononucleotide TSD (shaded). §, Inserted into a Basho V element.

Figure 2

A majority of mined transposons show an insertion preference for A + T-rich regions. The number of transposons used in the calculations is indicated in parentheses. The average G + C content was determined by using a 20-bp sliding window over 2000 bp of sequences flanking the transposon insertion sites. Only two CACTA-like elements were used, resulting in a low signal-to-noise ratio. Individual groups (see Table 1) of class I elements showed a similar profile (data not shown) as indicated for the entire class I average.

Figure 3

Structure of an Arabidopsis Tc1/Mariner-like transposon. (A) Similarities between TIRs and TSDs (underlined) of an Arabidopsis MLE I member and Tc1/Mariner-like elements Pogo (Drosophila, gi 8354) and Tigger (human, gi 2226003). (B) Putative transposase for the Arabidopsis MLE I (gi 4262216) aligned with transposases from Drosophila melanogaster PogoR11 (gi 2133672) and from human Tigger1 (gi 2226004). Amino acid residues are shaded based on the level of structural and functional similarities. Residues conserved between three or two sequences are shaded black and gray, respectively. The arrow (⇑) indicates the predicted start of the Arabidopsis MLE I ORF as annotated in GenBank. The first methionine of the Arabidopsis MLE I transposase was inferred from the reading frame and sequence similarity with the human Tigger1 element. The stop (*) was introduced by a single nucleotide substitution (at position 85709 in gi 4262209) from GAG (glutamine) to TAG (stop).

Figure 4

MITE transposases (A). Diagram depicting structural similarities between members of MITE XI elements (gray boxes) and a member of MITE X (striped box). Black boxes represent ORFs corresponding to a putative transposase: MITE XI ORFs share >98% amino acid sequence similarity. MITE X and XI are distinct groups because no significant internal nucleotide sequence similarity is observed between MITE X and XI nor for the consensus sequence of their TIRs (5′-GG(G/T)GGTGTTATTGGTT-3′ for MITE X and 5′-GGCCCTGTTTTGTTTG-3′ for MITE XI). However, putative transposase ORFs, length of TIRs (16 bp), and TSD (5′-TTA-3′) of MITE X and XI are similar, indicating a related mechanism of transposition and possibly a common origin for both groups. GenBank gi numbers of the clones from which elements were mined are indicated to the right. Nucleotide positions of transposon termini are indicated above each element. (B) Amino acid sequence similarity between the conceptual translation for MITE X (gi 4454587, nucleotide position 4650–5865) and the corresponding region of MITE XI ORF (gi 4585884). Identical amino acids and functionally or structurally related residues are shaded in black. *, Translational stops. Boxed sequences indicate the region containing the DDE motif. (C) Alignment of conserved regions corresponding to the functionally important DDE motif found in transposases and integrases of many transposable elements (–42). Transposases are from MITE X (gi 4454587, conceptual translation of nucleotides 4650–5865), MITE XI (gi 4585884), IS5S (gi 1256580) from Synechocystis sp., IS493 (gi 1196467), and IS903 (gi 136129) from Escherichia coli. Amino acid residues are shaded based on the level of structural and functional similarities. Residues conserved between all, four, or three sequences are shaded black, dark gray, and light gray, respectively.

Figure 5

Interelement recombination in non-TIR-MULEs generates mosaic transposons. (A) Diagrammatic alignment of four members of MULE VI. Black and gray boxes delimit regions of nucleotide sequence similarity (>90%) between members of MULE groups VI. GenBank gi numbers are marked to the right. Nucleotide positions on clones are noted above each element. (B) Expanded view (as boxed in A) indicating the sequence of the putative recombination region.

Figure 6

Acquisition of a truncated cellular gene by a member of MULE-I. (A) MULE-I (gi 2182289) shares 85% nucleotide similarity with the first two exons and first intron of the homeobox gene Athb-1 (gi 6016704). Conserved nucleotides are shaded in black, and positions on clones are indicated on the right of the alignment. Sequences for the first two exons are boxed. The first ATG is underlined, and an in-frame stop codon of the MULE-I ORF is double underlined. (B) Diagram illustrating the regions of nucleotide similarity (85%) between the MULE-I and the genomic sequence of Athb-1 (shaded in gray). Positions for the Athb-1 gene and MULE-I element on their respective clones are indicated. MULE-I TIRs are represented by black triangles, and the TSD sequences are indicated flanking both termini.