FIDEL-a retrovirus-like retrotransposon and its distinct evolutionary histories in the A- and B-genome components of cultivated peanut

Abstract

In this paper, we describe a Ty3-gypsy retrotransposon from allotetraploid peanut (Arachis hypogaea) and its putative diploid ancestors Arachis duranensis (A-genome) and Arachis ipaënsis (B-genome). The consensus sequence is 11,223 bp. The element, named FIDEL (Fairly long Inter-Dispersed Euchromatic LTR retrotransposon), is more frequent in the A- than in the B-genome, with copy numbers of about 3,000 (+/-950, A. duranensis), 820 (+/-480, A. ipaënsis), and 3,900 (+/-1,500, A. hypogaea) per haploid genome. Phylogenetic analysis of reverse transcriptase sequences showed distinct evolution of FIDEL in the ancestor species. Fluorescent in situ hybridization revealed disperse distribution in euchromatin and absence from centromeres, telomeric regions, and the nucleolar organizer region. Using paired sequences from bacterial artificial chromosomes, we showed that elements appear less likely to insert near conserved ancestral genes than near the fast evolving disease resistance gene homologs. Within the Ty3-gypsy elements, FIDEL is most closely related with the Athila/Calypso group of retrovirus-like retrotransposons. Putative transmembrane domains were identified, supporting the presence of a vestigial envelope gene. The results emphasize the importance of FIDEL in the evolution and divergence of different Arachis genomes and also may serve as an example of the role of retrotransposons in the evolution of legume genomes in general.

Fig. 1

Illustration of the characteristics of the four “bins” used to analyze the pattern of FIDEL insertion. Lines represent BAC clones with a pair of BAC end sequences each: Bin0 sequences (gray) have similarity to FIDLE and are located on the same BAC clone; Bin2 sequences (black, striped) have no similarity to FIDEL, but are paired to Bin1 sequences (light gray) that do; Bin3 sequences (black) have no similarity to FIDEL and are paired to sequences of the same group. Line interruption points out the different scales between the BAC total sequence (∼100–110 kbp) and the BAC end sequences (700 bp on average)

Fig. 2

Structure, composition, and characteristics of the Ty3-gypsy retrotransposon FIDEL, isolated from Arachis (FIDEL consensus sequence shown in ESM Supplement S2; GenBank accession no. xxxx). a Structural organization: TSD target site duplication, LTR long terminal repeat, PBS primer binding site, PPT polypurine tract used for synthesis of the second (+) DNA strand, GAG structural proteins for virion core, Protease protease for cleavage of primary translation products, RT reverse transcriptase, RNase H ribonuclease, Integrase endonuclease for integration in host genome. b Analysis of DNA complexity. Two regions corresponding to the pre-pol- and post-pol region of FIDEL (marked by vertical red dotted lines) are characterized by lower complexity. c Analysis of nucleotide identity between the two fully sequenced elements FIDEL-1 and FIDEL-2 as a measure of variability between the FIDEL copies. The red interrupted bar indicates a region between 8,000 and 9,000 bp in FIDEL-2 with a high number of insertions. These insertions were deleted to facilitate alignment and nucleotide comparison. d Amino acid sequences derived from ORFs >50 bp of four A. duranensis BAC end sequences highly similar to the end of the post-pol region (approximately 100 bp before 3′LTR). Red are 19 AA TM segments as predicted using the program Membrane Protein Explorer MPEx

Fig. 3

Estimation of copy number. a Chemilumigraph of a dot blot with dilution series of RT subclone Ah-FIDEL-9 for calibration (lanes 2 and 4; dots represent 10–10,000 copy number equivalents in 500 ng A. ipaënsis genomic DNA) and of genomic A. ipaënsis DNA (lanes 1 and 3, 1.6–500 ng), hybridized with Dig-labeled Ah-FIDEL-9. b Calibration curve based on signal intensities of Ah-FIDEL-9 dots. c Curve based on hybridization signals of A. ipaënsis dots (on the same filter). The signal intensities used for the graphs are average values from the individual dots of two dilution series each. White crosses indicate elimination of the data due to over saturation of the film

Fig. 4

Cytogenetic analysis using GISH and FISH.a–d Genomic in situ hybridization of a metaphase preparation of an amphidiploid hybrid of A. duranensis × A. ipaënsis probed with genomic DNA of both parents. a DAPI counterstain showing 40 chromosomes, 20 of which with centromeric bands typical for Arachis A-genome chromosomes. b A. duranensis genomic probe. c A. ipaënsis genomic probe. d Superposition of b and c. Note chromosome pair A9 without significant hybridization signal. In contrast to the typical A. hypogaea karyotype, two A-genome chromosome pairs (A10 and A2) exhibit (predominantly) extended secondary constrictions (see text). The short arm and the proximal segment of the long arm are indicated by an asterisk, and the separated satellites are marked by a degree sign. e–k FISH using rDNA and LTR sequences of the FIDEL element as probes. e, f Chromosome spreads of A. duranensis. e Localization of one sub-centromeric 5S rDNA sites (green) and two 18S–5.8S–25S sites (red). Note the distance from satellites in A10 from the proximal arm segment (chromosome annotation according to Seijo et al. 2004). f Rehybridization with a Dig-labeled LTR-probe. Chromosome A10a bearing a 18S–5.8S–25S site shows only weak hybridization (arrows). g, h FISH of chromosome spread of A. stenosperma (A-genome). g Localization of one sub-centromeric 5S rDNA sites (green). h Rehybridization with Dig-labeled LTR probe. Note the lack of or less intensive hybridization to centromeres and telomeric regions (arrows) as well as to chromosome pair A9. i–k FISH of chromosome spread of A. hypogaea. i DAPI staining enables discrimination between A-genome (DAPI bands) and B-genome chromosomes. j Localization of two 5S rDNA sites (red). k Dig-labeled LTR probe (green) showing hybridization preferentially to A-genome chromosomes. B-genome chromosomes with 5S rDNA site exhibiting stronger signals as compared to the others. Scale bar, 5 µm

Fig. 5

Neighbor-joining tree based on nucleic acid sequences of 132 FIDEL rt sequences derived from cloned PCR products from A. duranensis, A. ipaënsis, and A. hypogaea (Ad-FIDEL-x, Ai-FIDEL-x, Ah-FIDEL-x) and from homologous BAC end sequences (ADURxxxTF/TR, AIPAxxxTF/TR). Also included are the rt sequences from the isolated complete elements Ad-185P1FIDEL1 and -2, which have an estimated age of insertion 1.7 and 2.8 Mya (sequences in ESM Supplement S1). Prior to phylogenetic analysis, the sequences were aligned using ClustalW. Bootstrap values from 1,000 replicates are shown next to the branches as percentages (values lower than 50% are hidden). Sequence names were color-labeled to facilitate discrimination of the host species: A. duranensis (green dots), A. ipaënsis (red dots), and A. hypogaea (yellow dots)

Fig. 6

Neighbor-joining tree based on amino acid sequences of reverse trancriptases of Ty3-gypsy class LTR retrotransposons from higher plants. Bootstrap values from 1,000 replicates are shown next to the branches as percentages (values lower than 50% are hidden). The tree is rooted to gypsy (P10401). Apart from MEGY-1 (AC146683), PIGY-1 (AY299398, nucleotides 26–13419), Ogre (AY299398, nucleotides 14501–37253; the RT sequences of the Athila/Calypso and Tat group were derived from supplemental material to Wright and Voytas (2002). The GenBank accession nos. of other gypsy-like elements included in the analysis (lower clade) are as follows (from top to bottom): IFG7: AJ004945; Lore1a: AJ966990; Reina: U69258; Lore2A: AB430320; Lore2B: AB430231; RIRE7: BAA89466; Dea1: Y12432; del1: X13886; RIRE3: AB014738; RIRE8A: AB014746. Phylogenetic analyses were conducted in MEGA4