Plant centromeric retrotransposons: a structural and cytogenetic perspective

Abstract

Background: The centromeric and pericentromeric regions of plant chromosomes are colonized by Ty3/gypsy retrotransposons, which, on the basis of their reverse transcriptase sequences, form the chromovirus CRM clade. Despite their potential importance for centromere evolution and function, they have remained poorly characterized. In this work, we aimed to carry out a comprehensive survey of CRM clade elements with an emphasis on their diversity, structure, chromosomal distribution and transcriptional activity.

Results: We have surveyed a set of 190 CRM elements belonging to 81 different retrotransposon families, derived from 33 host species and falling into 12 plant families. The sequences at the C-terminus of their integrases were unexpectedly heterogeneous, despite the understanding that they are responsible for targeting to the centromere. This variation allowed the division of the CRM clade into the three groups A, B and C, and the members of each differed considerably with respect to their chromosomal distribution. The differences in chromosomal distribution coincided with variation in the integrase C-terminus sequences possessing a putative targeting domain (PTD). A majority of the group A elements possess the CR motif and are concentrated in the centromeric region, while members of group C have the type II chromodomain and are dispersed throughout the genome. Although representatives of the group B lack a PTD of any type, they appeared to be localized preferentially in the centromeres of tested species. All tested elements were found to be transcriptionally active.

Conclusions: Comprehensive analysis of the CRM clade elements showed that genuinely centromeric retrotransposons represent only a fraction of the CRM clade (group A). These centromeric retrotransposons represent an active component of centromeres of a wide range of angiosperm species, implying that they play an important role in plant centromere evolution. In addition, their transcriptional activity is consistent with the notion that the transcription of centromeric retrotransposons has a role in normal centromere function.

Figure 1

Diversity of CRM families and their species of origin. (A) Neighbor-joining tree inferred from a comparison of reverse transcriptase (RT) domain sequences. The non-chromovirus element Tat4-1 was used as an outgroup, while members of the Tekay, Reina and Galadriel clades were included as representatives of other plant chromoviruses. Alignment of the RT domains is provided in Additional file 3: Alignment of RT domains. On the basis of differences at the C-terminus of integrase, the CRM families were divided into groups A, B and C (Figure 3). Previously described CRM members are shown in purple (see also Additional file 1: Origin and structural features of sequences used in this work). Families with confirmed centromeric localization are marked with orange stars (fluorescence in situ hybridization results) or green stars (in silico localization). Families having a dispersed chromosomal distribution are labeled with orange or green hexagons. Bootstrap values are shown only for the major nodes. Elements belonging to the Tekay, Reina and Galadriel clades are listed in Additional file 1: Origin and structural features of sequences used in this work. It should be noted that because of the limitations of the neighbor-joining method and the lack of representatives from a wider range of evolutionarily distant species, the tree topology may not fully reflect real phylogenetic relationships between different groups of CRM elements. (B) Taxonomy classification of the species containing the CRM elements. Dates of divergence between major groups of plants are from the work by Chaw et al. [105]. The names of CRM families present in the species are shown in brackets.

Figure 2

Graphical representation of the conserved portion of the integrase protein sequence. Integrase sequences extracted from CRM, Tekay, Reina and Galadriel chromoviruses aligned using the Muscle program are shown as sequence logo plots [96]. CRM clade members are shown in the upper part of the figure, and those from the other clades are shown in the lower part. Despite the overall high level of sequence similarity, several amino acid residues are conserved only within the CRM clade. The HHCC and DD35E motifs are indicated by green and red stars, respectively.

Figure 3

Alignments of sequences at the C-terminus integrase. Group A elements possess a CR motif with several strongly conserved amino acid residues near the N-terminus. These residues are not present in Beetle1, Beetle2 or SilL1, but are well conserved in the grass species elements (see bottom part of the alignment). The numbers between two aligned blocks specify the number of amino acid residues not shown. No putative targeting domain (PTD) is encoded by group B elements. The type II chromodomain of group C elements shares sequence similarity with Tekay, Galadriel and Reina clade members. The dotted line above each alignment marks a region conserved among all plant chromoviruses. The arrow shows the portion of the integrase lying within the 3' long terminal repeat(LTR). Asterisks indicate stop codons at the end of each open reading frame. A highly conserved GPY/F motif [36,37] is indicated by a black trapezoid above the beginning of each alignment.

Figure 4

Structural analysis of CRM elements. (A) Polyprotein coding (white boxes), noncoding (gray boxes), putative targeting domain (PTD) (hatched boxes) and long terminal repeats (LTRs) (arrowed). Pbs, primer binding site; ppt, polypurine tract; aaa, A-rich stretch. The group A member coding region extends into the 3' LTR, which encodes the CR motif. Group B elements lack any PTD. Group C possesses a type II chromodomain-coding domain which terminates close to the 5' end of the 3' LTR. The graph is not drawn in proportion to segment lengths in base pairs. (B) Most elements share TGATG and T/CATCA inverted repeats at, respectively, the 5' and 3' end of the LTR. The primer binding site complementary to the 3' end of tRNA^Met differs in sequence between various families. Group A elements contain highly conserved polypurine tract sequences. (C) A protein similarity plot shows that the CRM polyproteins are highly conserved, varying mainly within their C-terminal PTD regions. Individual polyprotein domains: GAG, capsid domain, similar to pfam37032; ZF, nucleocapsid GAG protein zinc finger; PRO, protease; RT, reverse transcriptase RNase H; IN, integrase; PTD, putative targeting domain.

Figure 5

Fluorescence in situ hybridization (FISH)-based visualization of the intrachromosomal distribution of chromoviruses. (A) Pea chromosomes hybridized with PiSat1 (group A). (B and C) Black cottonwood interphase nucleus and metaphase chromosomes hybridized with PopT2 (group A). Note that most of the signal is associated with chromocenters (bright 4',6-diamidino-2-phenylindole (DAPI)-stained spots in the interphase nucleus). Three metaphase chromosomes were enlarged to allow a clearer localization of PopT2 to the centromeric region (C1-C3). (D and E) White campion chromosomes hybridized with SilL1 and SilL2 (group A). (F) Banana chromosomes hybridized with MusA1 (group B). Since all of the banana chromosomes are metacentric or submetacentric, signals located around the center of the chromosome are taken to reflect loci near or within the centromere. Three of the chromosomes with identifiable centromeres were enlarged (F1-F3). (G and H) Norway spruce chromosomes counterstained with DAPI and hybridized with a Spdl-like sequence (group C). (I) Pea chromosomes hybridized with Peabody (Tekay clade). Positive hybridization signals are shown in green, and DAPI-counterstained DNA appears in red.