Skip to main content
Download PDF

Evolution of Einkorn wheat centromeres is driven by the mutualistic interplay of two LTR retrotransposons

Mobile DNA volume 15, Article number: 16 (2024) Cite this article

Abstract

Background

Centromere function is highly conserved across eukaryotes, but the underlying centromeric DNA sequences vary dramatically between species. Centromeres often contain a high proportion of repetitive DNA, such as tandem repeats and/or transposable elements (TEs). Einkorn wheat centromeres lack tandem repeat arrays and are instead composed mostly of the two long terminal repeat (LTR) retrotransposon families RLG_Cereba and RLG_Quinta which specifically insert in centromeres. However, it is poorly understood how these two TE families relate to each other and if and how they contribute to centromere function and evolution.

Results

Based on conservation of diagnostic motifs (LTRs, integrase and primer binding site and polypurine-tract), we propose that RLG_Cereba and RLG_Quinta are a pair of autonomous and non-autonomous partners, in which the autonomous RLG_Cereba contributes all the proteins required for transposition, while the non-autonomous RLG_Quinta contributes GAG protein. Phylogenetic analysis of predicted GAG proteins showed that the RLG_Cereba lineage was present for at least 100 million years in monocotyledon plants. In contrast, RLG_Quinta evolved from RLG_Cereba between 28 and 35 million years ago in the common ancestor of oat and wheat. Interestingly, the integrase of RLG_Cereba is fused to a so-called CR-domain, which is hypothesized to guide the integrase to the functional centromere. Indeed, ChIP-seq data and TE population analysis show only the youngest subfamilies of RLG_Cereba and RLG_Quinta are found in the active centromeres. Importantly, the LTRs of RLG_Quinta and RLG_Cereba are strongly associated with the presence of the centromere-specific CENH3 histone variant. We hypothesize that the LTRs of RLG_Cereba and RLG_Quinta contribute to wheat centromere integrity by phasing and/or placing CENH3 nucleosomes, thus favoring their persistence in the competitive centromere-niche.

Conclusion

Our data show that RLG_Cereba cross-mobilizes the non-autonomous RLG_Quinta retrotransposons. New copies of both families are specifically integrated into functional centromeres presumably through direct binding of the integrase CR domain to CENH3 histone variants. The LTRs of newly inserted RLG_Cereba and RLG_Quinta elements, in turn, recruit and/or phase new CENH3 deposition. This mutualistic interplay between the two TE families and the plant host dynamically maintains wheat centromeres.

Background

In this study, we focus on Einkorn wheat (Triticum monococcum), a diploid and the first known wheat species to be domesticated. Einkorn wheat is a member of the Triticeae family which includes important crops such as wheat, barley and rye. The Triticeae belong to the large family of the grasses (Poaceae). Grasses originated probably over 100 million years ago and diversified 50–80 million years ago [45, 65]. The Triticeae diverged from their closest cereal relative, Avena sativa (oat) about 28 million years ago and diversified into barley, rye and the wheat group about 10 million years ago [28, 31, 42]. T. monococcum itself is a close relative of the A genome of hexaploid (bread) wheat, which diverged about 1 million years ago [28]. Because of their high transposable element (TE) content of > 90%, centromeres from Triticeae had not been sequenced completely [30, 66]. Only very recently, gap-free assemblies of Einkorn wheat centromeres were produced [2], which allowed for the detailed analysis presented in this study.

Centromeres are of fundamental importance for all eukaryotes. During cell division, sister chromatids are moved towards opposing cell poles by microtubules which are connected to chromosomes via the centromeric kinetochore complex. Functional centromeres are defined epigenetically in that the two canonical histone H3 proteins are replaced with histone variants CENH3 (CENP-A in humans) in the nucleosomes [10]. Nucleosomes, which consist of octamers of histone proteins, are placed in regular intervals along chromosomes in a process called “phasing”. Here, DNA stretches of ~ 150 bp are wrapped around each nucleosome with spacers of 10–80 bp separating individual nucleosomes [11].

Centromere sizes vary greatly between species, from “point” centromeres in yeast which are defined by a single ~ 125 bp sequence [14] to Caenorhabditis elegans “holo-centromeres” which span nearly the entire length of chromosomes [57]. Plants usually have “regional” centromeres that are several megabases (Mb) in size and typically contain large arrays of centromere-specific tandem repeated sequences [6, 35, 50, 65, 74]. A recent study in the model plant Arabidopsis thaliana showed that centromeres contain thousands of tandemly repeated units of the 178 bp sequence motif Cen178 [35, 73]. Similar tandem repeats were identified in centromeres of several grasses such as rice, maize and Brachypodium [6, 65]. Although the centromeric tandem repeat sequences are poorly conserved between species [9], they generally have a size of 150–180 bp, the necessary length for DNA to wrap around one nucleosome and allowing for spacing between nucleosomes.

In most grasses studied so far, centromeric tandem repeats are interspersed with LTR retrotransposons from highly centromere-specific families [6, 50, 65, 74]. The known centromere-specific retrotransposon families in grasses all belong to the Gypsy superfamily, and sequence homology indicates that they all evolved from a common ancestor in flowering plants (angiosperms, [38]). Occasionally, they may have been transferred horizontally between species [55]. In maize and rice, they are referred to as CRM (centromeric retrotransposons of maize) and CRR (centromeric retrotransposons of rice), respectively [39, 54], while in Triticeae they are called RLG_Cereba [46, 69]. Interestingly, a recent study found that Einkorn wheat does not have centromere-specific tandem repeats. Instead, its highly repetitive centromeres are derived almost exclusively from retrotransposons, with the RLG_Cereba and RLG_Quinta families being the most abundant [2].

Despite the low sequence conservation of centromere-specific repeats between species, the recruitment and/or phasing of centromeric (CENH3 containing) nucleosomes is likely promoted by specific DNA sequence motifs [9]. In yeast, for example, a specific ~ 125 bp sequence is essential for the establishment of its point centromeres [8]. In Arabidopsis, the Cen178 tandem repeats are strongly associated with CENH3 histone variants, while interspersed TEs in centromeres show much lower CENH3 signals [35]. Additionally, more divergent Cen178 copies were less associated with CENH3, suggesting selection pressure for specific sequence motifs [35]. In grasses, the DNA of some centromere-specific retrotransposons was shown to interact with CENH3 and their transcription might be involved in CENH3 deposition [75, 18]. In particular, parts of RLG_Cereba and RLG_Quinta retrotransposons from hexaploid wheat were shown to have strong affinity for CENH3 [22, 27, 58]. Interestingly, even the human immunodeficiency virus (HIV) was shown to have strictly placed nucleosomes in its LTR which are important for the transcriptional regulation of the retrovirus [36].

It has long been known that some retrotransposons target epigenetic marks for insertion into the genome, and that this is promoted by specific chromodomains that are fused to the C-terminus of the integrase (INT) proteins [1, 12]. Centromere-specific retrotransposons from grasses lack the classic chromodomain but instead contain a so-called CR domain [12, 37, 38]. It was therefore hypothesized that the CR domain recognizes functional centromeres and guides new insertions there. However, the exact function of the CR domain and the molecular mechanism underlying the centromere targeting are still unknown. Additionally, there may be other mechanisms that lead to accumulation of TEs in centromeric and pericentromeric regions. For example, the Athila retrotransposons in Arabidopsis are prevalent in peri/centromeric regions but do not contain CR domains [37, 43, 44]. It is unclear whether they actively target centromeres or whether they simply accumulate where they have the least deleterious effects.

Here, we analyze recently published gap-free assemblies of T. monococcum (Einkorn wheat) centromeres. We show that the centromere-specific RLG_Quinta elements are non-autonomous and most likely cross-mobilized by RLG_Cereba elements, and that new RLG_Quinta and RLG_Cereba insertions indeed target functional centromeres. Additionally, using ChIP-seq data, we identified specific sequences inside RLG_Quinta and RLG_Cereba LTRs which strongly phase CENH3-containing nucleosomes. Our combined findings allow the development of a model for the dynamic evolution of centromeres in wheat that is driven by the RLG_Quinta and RLG_Cereba retrotransposon families.

Results and discussion

RLG_Cereba is an autonomous retrotransposon that cross-mobilizes RLG_Quinta

For our previous study, we identified the boundaries of functional centromeres in Einkorn wheat accessions TA299 and TA10622 with the help of CENH3 ChIP-seq data [2]. For the current study, we focused on TA299 because its centromeres were assembled gap-free, while the centromere of chromosome 2A in the TA10622 assembly still contains a few gaps. We re-annotated centromeres with a specific focus on centromeric retrotransposons using an in-house pipeline. The known centromere-specific RLG_Cereba and RLG_Quinta retrotransposon families contribute ~ 47% and ~ 9% of the functional centromeric sequences, while they are practically absent from chromosome arms (Supplementary Fig. S1). The third known centromere-specific family, RLG_Abia [69], contributes less than 5%. Other high-copy TE families are also present, but far less abundant than outside centromeres (Supplementary Fig. S1).

Using recently described methods [71], we isolated 2,658 full-length RLG_Cereba and 935 RLG_Quinta retrotransposon copies from the TA299 T. monococcum assembly. From these we derived consensus sequences for multiple subfamilies to predict open reading frames (ORFs), encoded proteins and structural motifs such as primer binding sites (PBS) or poly-purine tracts (PPT). Analogous analyses for accession TA10622 yielded practically identical consensus sequences.

RLG_Cereba contains one large ORF that encodes all canonical proteins necessary for its replication, namely GAG that forms virus-like particles in which reverse transcription takes place, reverse transcriptase (RT), RNAse H (RH), integrase (INT) and a protease that cleaves the polyprotein into its functional enzymes (Fig. 1).

Fig. 1
figure 1

Characterisation of centromere-specific RLG_Cereba and RLG_Quinta retrotransposons. A Comparison of sequence organization of RLG_Cereba and RLG_Quinta. Sequences conserved at the DNA level are connected with gray areas, with the shade of gray reflecting the level of sequence conservation. The region encoding the GAG domain, which shows conservation at the protein but barely any homology at the DNA level, is indicated by a yellow area. RLG_Cereba encodes a polyprotein with typical domains for autonomous retrotransposons, plus a previously described CR motif. In contrast, RLG_Quinta encodes only a GAG domain and a domain of unknown function (DUF1). Note that parts of the LTRs including diagnostic motifs such as LTR termini and primer binding site (PBS) are highly conserved between the two. ZF: Zinc finger, PRO: aspartic protease, RT: reverse transcriptase, RH: RNase H, INT: integrase, PPT: poly-purine tract. B Multiple sequence alignment of 3’ terminal regions of 5’ LTRs from RLG_Cereba and RLG_Quinta subfamilies. Predicted integrase binding site (IBS) and PBS are indicated with black horizontal bars. C Multiple sequence alignment of 5’ terminal regions of 3’ LTRs from RLG_Cereba and RLG_Quinta subfamilies. Predicted PPT and IBS are indicated with black horizontal bars. D Schematic model of how RLG_Quinta is cross-mobilized by RLG_Cereba. Both TEs are transcribed at the same time due to conservation of regulatory regions. RLG_Cereba provides proteins necessary for replication (e.g. RT and INT), while RLG_Quinta contributes GAG proteins which are needed in large quantities for the formation of virus-like particles

Full size image

Additionally, the INT protein has the CR domain that is exclusively found in centromere-specific retrotransposons [37] (see below). In contrast, RLG_Quinta retrotransposons are shorter and contain an ORF that only encodes GAG and a domain of unknown function (DUF1, Fig. 1a), while it lacks coding sequences (CDS) for RT and INT. While homology to GAG is very clear, at this point we have no hint as to the function of DUF1. Thus, we consider RLG_Quinta a non-autonomous retrotransposon that must rely on enzymes encoded elsewhere for its replication. However, the intact ORF for GAG indicates that it contributes GAG protein which can be used by both RLG_Cereba and RLG_Quinta.

It is not unusual that TE populations in plants comprise autonomous and non-autonomous elements [70], and several of the most abundant TEs in Triticeae were shown to be non-autonomous [70, 71]. Indeed, RLG_Cereba and RLG_Quinta fulfill the previously described criteria [71] for a pair of autonomous and non-autonomous retrotransposons (Fig. 1): first, the predicted primer binding site (PBS) just downstream of the 5’ LTR which is needed for the initiation of reverse transcription is identical in both families (Fig. 1B). Second, the termini of the LTRs, which serve as binding sites of the integrase protein are identical in all identified RLG_Cereba and RLG_Quinta subfamilies from Triticeae (Fig. 1B and C). Third, the second half of the LTR (which contains putative promoter and terminator sequences) is strongly conserved between RLG_Cereba and RLG_Quinta elements (Fig. 1A, Supplementary Fig. S2). The conserved 3’ half of the RLG_Cereba and RLG_Quinta LTR starts near a predicted TATA box (Fig. 1, Supplementary Fig. S2). The conserved region also contains a putative poly-adenylation signal which we identified with the help of published IsoSeq transcripts for both families [2]. RLG_Cereba has the canonical AATAAA poly-adenylation signal ~ 12 bp upstream of the start of the poly-A tail in the IsoSeq transcript (Supplementary Fig. S2B), while RLG_Quinta has two cryptic poly-adenylation sequences (AATA and ATATATAT) approximately 20–30 bp upstream of the poly-A tail. Importantly, the predicted TATA box and poly-adenylation signal are both supported by homology to the promoter of HIV. Here, the known TATA box [4] as well as the AATAAA motifs plus surrounding sequences are conserved between HIV and the two retrotransposons (Supplementary Fig. S2C). Based on this analysis, we inferred the boundaries of the typical U3, R and U5 regions of the RLG_Cereba and RLG_Quinta LTRs (Supplementary Fig. S2C). The conservation of the regulatory motifs in the 3’ half of their LTRs suggest that RLG_Cereba and RLG_Quinta may be co-expressed at the same time.

Finally, RLG_Cereba is the only retrotransposon family in the extensive datasets of wheat TEs that shares the described features with RLG_Quinta. We therefore conclude that RLG_Cereba is the autonomous partner that mobilizes RLG_Quinta elements, while RLG_Quinta contributes GAG protein. The latter makes RLG_Quinta a partially mutualistic partner in the RLG_Cereba/RLG_Quinta system, since GAG proteins are needed in large quantities for the formation of virus-like particles in which replication takes place. This is reminiscent of the “semi-autonomous” RLG_Sabrina retrotransposons in wheat which also sometimes encode GAG but lack genes for RT and INT [71].

RLG_Quinta evolved from RLG_Cereba in a common ancestor of oat and wheat

The finding that RLG_Quinta is a non-autonomous derivative of RLG_Cereba raised the question when RLG_Cereba and RLG_Quinta elements first evolved. For phylogenetic and comparative analysis, we isolated RLG_Cereba homologs from Triticeae (barley, rye, wheat) and their close relative oat (A. sativa). Additionally, we searched more distantly related grasses, Brachypodium, rice and maize, as well as the basal grasses Pharus latifolius and Streptochaeta angustifolium. We found RLG_Cereba homologs in all species studied, which dates the origin of RLG_Cereba homologs at least to a common ancestor of Streptochaeta and wheat, or approximately 100 million years ago [45], Fig. 2). In fact, RLG_Cereba homologs have been described in a wide range of plants including dicotyledons, indicating that they are an ancient retrotransposons lineage [38].

Fig. 2
figure 2

Evolutionary origin of RLG_Cereba and RLG_Quinta retrotransposons. A Phylogenetic tree of predicted GAG proteins from RLG_Quinta, RLG_Cereba and homologs from other grasses. The tree was constructed with MrBayes running for 125,000 generations. The numbers at branches indicate the probability that the taxa to the right of the branch are grouped together in all trees. The tree shows that RLG_Quinta evolved in the evolutionary lineage leading to Triticeae and oat (A. sativa). RLG_Quinta subsequently diverged into two main branches (A and B). B Sequence comparison of the two main RLG_Quinta lineages A and B using consensus sequences for two representative subfamilies from T. monococcum. The region of the GAG coding sequence shows very low sequence conservation at the DNA level, while the remaining sequence can be well aligned. This indicates that RLG_Quinta lineages underwent multiple recombination events involving GAG CDS and flanking sequences (see also Supplementary Fig. S3). C Schematic phylogeny of the grasses (Poaceae). Divergence times in million years are shown at separation nodes. Divergence times are approximate and were compiled from multiple publications (see text). Phylogenetic emergence of RLG_Cereba and RLG_Quinta retrotransposon lineages is indicated with stars

Full size image

In contrast to RLG_Cereba, we identified RLG_Quinta homologs only in Triticeae and oat, dating their emergence to the period between 35 and 28 MYA, after the Triticeae/oat ancestor diverged from Brachypodium (Fig. 2). This is also reflected in the phylogenetic tree of predicted GAG proteins where the RLG_Quinta lineage branched off after the divergence of Triticeae and oat from the other grasses (with strong branch support of 100%, Fig. 2A). We therefore conclude that RLG_Quinta evolved from a loss of the CDS for RT and INT, while the CDS for GAG was still maintained. The additional domain of unknown function encoded by RLG_Quinta may have been acquired later. Adding complexity, RLG_Quinta elements diverged early on into two main lineages (A and B) which encode two variants of GAG proteins that strongly differ from each other and from GAG encoded by RLG_Cereba (Fig. 2A and B). Interestingly, RLG_Quinta_A and RLG_Quinta_B also have highly divergent GAG genes where DNA identity is only ~ 65% (Fig. 2B, Supplementary Fig. S3), while sequences up- and downstream of the GAG genes are over 90% identical (Fig. 2B). This indicates that the RLG_Quinta A and B lineages recombined during their evolution (Supplementary Fig. S3). Such sequence exchange is well studied in retrotransposons and retroviruses and likely occurs through template switching during replication [15, 63]. We hypothesize that RLG_Quinta contributes with specific GAG variants to the RLG_Cereba/RLG_Quinta system, similar to previously described non-autonomous RLG_Sabrina and RLG_WHAM retrotransposons in wheat (Fig. 1D, [71]).

The phylogenetic tree of GAG proteins largely reflects the phylogeny of the Triticeae and their relatives, indicating no horizontal transfer between main taxa, at least since Triticeae diverged from the Brachypodium lineage (Fig. 2A).

Diversification of retrotransposon subfamilies during recent evolution

Using previously described methods [71], we performed principal component analysis (PCA) of populations of RLG_Cereba and RLG_Quinta elements. Here, we aligned all individual full-length RLG_Cereba and RLG_Quinta copies to their respective consensus sequences. From these alignments, sequence variants were called that were then used for PCA. This analysis was done with retrotransposon copies from seven genomes. This included rye (S. cereale), the three subgenomes of hexaploid wheat (A, B and D genome) as well as their diploid relatives T. monococcum, Aegilops speltoides and Aegilops tauschii, respectively. Barley was excluded since its centromeres are still assembled incompletely and do not contain sufficient numbers of full-length retrotransposon copies [30].

For RLG_Cereba, the PCA largely reflects species phylogeny (Fig. 3A, Supplementary Fig. S4A for accession TA10622), with rye, Ae. speltoides and T. monococcum diverging from the core of the wheat subgenomes. The A, B and D subgenomes are in the center of the PCA possibly because the consensus sequence was done from randomly picked copies from all species, of which wheat subgenomes provide the majority. In any case, the PCA shows that distinct subfamilies diverged in the different species and subgenomes since they evolved from a common ancestor 3–6 million years ago [28, 31].

Fig. 3
figure 3

Analysis of populations of centromere-specific retrotransposons. A Principal component analysis (PCA) of full length RLG_Cereba elements using SNPs obtained from alignment of individual elements against a consensus sequence. For each species, 747 randomly picked elements were included in the analysis. Retrotransposons form separate groups which largely correspond to the species/subgenomes used (Aspe: Ae. speltoides, Atau: Ae. tauschii, Scer: S. cereale, Tmon: T. monococcum (accession TA299), Taes: T. aestivum, A, B and D subgenomes). B The same analysis as in (A) but using 274 randomly picked RLG_Quinta elements from each species/subgenome. A and B clusters include elements from all the included species. C Insertion age distributions estimated from LTR divergence of RLG_Cereba and RLG_Quinta elements. Colors correspond to those of the subfamilies shown in the PCA in A and B. For visual clarity only data points falling into the 99% percentile are shown. The most recently active RLG_Quinta elements in the T. aestivum subgenomes and T. monococcum are part of the A lineage, while the B lineage was more recently active in the genome of Ae. speltoides

Full size image

In contrast, RLG_Quinta retrotransposon populations are more diverse than those of RLG_Cereba. The first principal component (PC1) separates the two main lineages A and B which were already identified in the phylogenetic tree (Fig. 2B) and which were present in the Triticeae/oat ancestor. The clear separation is due to strong divergence in the region encoding the GAG protein domain where sequences between the A and B lineage can hardly be aligned (Fig. 2B). The second principal component (PC2) separates subfamilies mainly reflecting the different species (Fig. 3B, Supplementary Fig. S4B for accession TA10622).

In all species studied here, most RLG_Cereba and RLG_Quinta copies inserted in the past 4 million years. This is typical for TEs in grasses, because intergenic sequences are rapidly reshuffled through TE insertions and deletion of DNA [64], making it rare to find TEs older than a few million years. Interestingly, especially T. monococcum centromeres contain large numbers of very young RLG_Cereba and RLG_Quinta copies. Here, the RLG_Quinta_A lineage was more recently active, with most copies being less than 200,000 years old (Fig. 3C, Supplementary Fig. S4C). This could in part be due to the high quality of the assembly but could also reflect very recent insertion activity (see below).

RLG_Cereba and RLG_Quinta insert specifically into functional centromeres

It was proposed that the CR domain fused to the INT protein of centromere-specific retrotransposons guides their insertion to functional centromeres [37, 38]. Indeed, previous studies in wheat found that the younger RLG_Cereba elements tend to be in the middle of centromeres [71]. However, sequence assemblies in centromeric regions were not complete enough to answer the question whether they indeed actively target functional centromeres, or whether they are simply tolerated best in centromeric regions.

Here, we analyzed sequence diversity, phylogeny, insertion age and positions of 1,964 RLG_Cereba and 663 RLG_Quinta elements from T. monococcum. To exclude possible recombinant copies, we only used copies which had target site duplications (TSDs) with maximum 1 bp mismatch. By PCA, we distinguished 3 subfamilies for RLG_Cereba and 7 for RLG_Quinta, which were also reflected in their phylogenetic trees (Fig. 4). The phylogenetic trees show that the different subfamilies were active at different times during the past 4 million years. Older copies are mostly found outside of functional centromeres, while younger ones are inside (Fig. 4). Moreover, in our recent study, we found that the centromere of chromosome 4 has shifted 20,00–100,000 years ago by approximately 10 Mb [2]. This functional neocentromere contains the youngest insertions of RLG_Cereba and RLG_Quinta elements, indicating that active centromeres are indeed targets for new insertions.

Fig. 4
figure 4

Analysis of insertion age and physical location of RLG_Cereba and RLG_Quinta retrotransposons from T. monococcum (accession TA299). A Principal component analysis (PCA) of 1,964 full length RLG_Cereba elements using SNPs obtained from alignment of individual copies against a consensus sequence. B The same analysis with 663 RLG_Quinta copies. C and D Phylogenetic trees for individual RLG_Cereba and RLG_Quinta copies inferred by RAxML. E and F Association of physical to genetic distance of retrotransposon copies in centromeres. The x-axis indicates the difference in genomic position, measured by the absolute difference in the distance from the centromere midpoint. The y-axis shows the number of SNPs in pairwise alignments of individual copies. Mantel test statistics and the corresponding p-value are shown in the bottom right of each plot. The orange line shows the linear regression calculated by the function lm(). Other chromosomes for RLG_Cereba and RLG_Quinta_B are shown in Supp. Fig. S5 and S6

Full size image

We performed the Mantel test which is commonly used in population genetics to find correlations between genetic and geographical distances. We tested whether we find a correlation between sequence similarity between individual copies and their physical distance to the centromere. Here, we calculated the distance from the centromere midpoint for each TE copy. The positional difference between two copies was then defined as the absolute difference between their distances from the centromere midpoint. For example, if the centromere midpoint for a chromosome is at position 200 Mb, and two copies are located at 180 Mb and 210 Mb on this chromosome, their positional difference would be 10 Mb. Indeed, we found a strong correlation between genetic and physical distance for both RLG_Cereba and RLG_Quinta elements (Fig. 4E and F, Supplementary Fig. S5 and S6). This indicates that copies that were active at the same time were also inserted into similar physical loci (i.e. the functional centromeres), while older copies were “pushed” away from the centromeres over time. This is in line with the previous observation that the youngest RLG_Cereba elements are found almost exclusively in centromeres [2, 66].

Taken together, our analyses provide strong evidence that the RLG_Cereba integrase indeed actively targets functional centromeres.

The CR domain likely guides retrotransposon insertions toward functional centromeres

In wheat, RLG_Cereba and the much less abundant RLG_Abia are the only autonomous LTR retrotransposons strongly enriched in centromeric and peri-centromeric regions [69, 71], Supplementary Fig. S1). They are also the only TE families in the extensive wheat TE datasets which contain a CR domain fused to their INT protein (Fig. 5). A previous study on Gypsy superfamily retrotransposons showed that CRM homologs (which includes RLG_Cereba) formed a monophyletic clade, indicating that the acquisition of the CR domain was a one-time evolutionary event [38].

Fig. 5
figure 5

Analysis of integrase sequences from RLG_Cereba homologs. A Multiple alignment of predicted integrase proteins containing CR domains from 20 plant species. The previously described protein domains are indicated with colored bars above the aligned sequences. Diagnostic residues of the zinc finger, the DDE catalytic site and CR domain are shown. B Alphafold2 model of the RLG_Cereba integrase. Protein domains are indicated with the same colors as in (A) and diagnostic residues are shown as spheres. C Alphafold2 model for the interaction of histone H4 with CENH3 from T. monococcum (Tmon). Positively charged amino acids that interact with the negatively charged back bone of the DNA are shown as spheres. CENH3 interacts with H4 in the same way as human CENP-A (see Supplementary Fig. S7). The inset shows the predicted aligned error (PAE) plot. D The predicted interaction of the CR domain with CENH3. Note that the alpha helix and the RARAK motif (spheres) interact in a similar way with CENH3 as H4 (see also Supplementary Fig. S7)

Full size image

We hypothesized that the centromere-specificity is caused by a direct interaction of the CR domain with the centromeric CENH3 histone variants. We therefore compared integrase domains from centrome-specific retrotransposons from 20 plant species, covering the two major plant clades of the Monocotyledons and Dicotyledons. In all analyzed sequences, the integrase is well conserved, containing the typical domains, the HHCC Zinc finger, the integrase core domain which contains the catalytic DDE motif and a C-terminal domain (CTD, Fig. 5B). The CR domain itself is characterized by an alpha helix at which start lies a conserved TRARAR/K motif (hereafter RARAK, Fig. 5). It is separated from the canonical integrase by a poorly conserved “spacer” or “tether”.

We then used Alphafold2 to model possible interactions between the CR domain and CENH3-containing nucleosomes. The structure of human centromeric nucleosomes is well known [59], Supplementary Fig. S7) and shows direct interaction of CENP-A (the human CENH3 homolog) with histone H4. As a control, we modeled the interaction of CENH3 and histone H4 from T. monococcum, which looked practically identical with the dimer of the two human proteins (Fig. 7C). Interestingly, replacing H4 with the CR domain of RLG_Cereba resulted in a very similar dimer in which, most notably, the positively charged R and K residues of the RARAK motif were in the same position as R and K residues in H4, near the negatively charged back bone of the DNA wrapping around the nucleosome (Fig. 5D, Supplementary Fig. S7).

We are aware of the uncertainty of predictions of protein–protein interactions. Nevertheless, our results suggest that the CR domain may interact with CENH3 in a manner similar to H4. It is, for example, possible that the CR domain competes with H4 when nucleosomes are assembled during DNA replication [53]. This would “anchor” the integrase/dsDNA complex and guide the insertion to a nearby location, for example to one of the neighboring coils of the chromatin fiber (see below).

The presence of CENH3 is strongly associated with RLG_Cereba and RLG_Quinta LTRs

If RLG_Cereba INT indeed guides insertions toward the functional centromere, it would explain why other centromeric repeats would over time be out-competed. We therefore hypothesized that RLG_Cereba and/or RLG_Quinta should also functionally replace sequences where CENH3 containing nucleosomes are positioned. For our previous study, we produced ChIP-seq data to localize the boundaries of functional centromeres [2]. The main ChIP-seq experiment used MNase digested DNA and enrichment for CENH3 by antibody, while the H3K4me3 histone modification (which characterizes open chromatin) was used as control. As negative controls, ChIP-seq sequencing was done with MNase digested DNA without the use of antibodies. In this study, we used the ChIP-seq data to search for sequence motifs which promote deposition and/or phasing of CENH3 containing nucleosomes. We mapped CENH3 ChIP-seq reads on T. monococcum chromosomes, allowing multi-mapping reads to include regions that are highly conserved within TE families (see methods). To minimize erroneous mappings, we only allowed perfect alignment matches of at least 150 bp.

We first aimed at identifying loci which showed a high coverage with CENH3 ChIP-seq reads but low coverage in control experiments (see methods). In this “sequence agnostic” approach, we identified 1,051 sequences with an average size of 146 bp inside functional centromeres. A total of 933 (~ 89%) of them had homology to TEs. For the H3K4me3 control, we identified 675 loci, of which 495 (~ 73%) were TE-derived (Fig. 6). This indicates that in wheat centromeres, only a fraction of nucleosomes contains CENH3. It can also be expected that CENH3 localization varies somewhat between individual cells, or that individual nucleosomes are heterotypic (i.e. they may contain one H3 and one CENH3 histone variant). With our method, we selected those loci which were consistently associated with CENH3 while having very low signals in control data.

Fig. 6
figure 6

Analysis of ChIP-seq data in centromeric TEs from T. monococcum. A Numbers of ChIP-seq peaks found in TE sequences. A ChIP-seq peak was defined as a region of 80–300 bp which showed high read coverage in the experimental set (i.e. CENH3 and H3K4me3) but no signal in controls (see methods). B The same data normalized to the abundance and size of TE families. C and D ChIP-seq read coverage mapped on RLG_Quinta and RLG_Cereba consensus sequences. Here, ChIP-seq read coverage of all available full-length copies was compiled. The individual tracks show read coverage from different experiments and mock controls (MNase digested DNA without the use of antibodies). The top track shows the ratio between CENH3 and its mock inoculation. E Detailed mapping of the terminal 2000 bp of RLG_Quinta. ChIP-seq read coverage was used to infer positioning of nucleosomes. Nucleosomes and DNAse hypersensitive regions (DHS) were named analogous to those in human immunodeficiency virus (HIV [36])

Full size image

We mapped the 1,051 CENH3-associated sequences on wheat TE consensus sequences and found that they are highly enriched in RLG_Cereba and RLG_Quinta sequences, while other TE families showed only very low coverage (Fig. 6). In contrast, the 675 sequences found associated with H3K4me3 were enriched in various TE families other than RLG_Cereba and RLG_Quinta (Fig. 6A and B). Normalization for TE abundance shows that TE families other than RLG_Cereba and RLG_Quinta are associated more often with the H3K4me3 control (Fig. 6B), while RLG_Cereba and RLG_Quinta are under-represented. Taken together, these data indicate that a large majority of CENH3 signals in centromeres are associated with RLG_Cereba and RLG_Quinta, and that the two families have a particular affinity to CENH3 containing nucleosomes. Similar observations were made in hexaploid wheat, where both families showed co-localization with CENH3 and ChIP enrichment [22, 27]. Outside centromeres, we found no particular enrichment of TE families in either experiment (Suppl Fig. S8). Furthermore, the CENH3 phasing satellite repeats T566 and T550 previously described in hexaploid wheat [58] were found only in few copies and predominantly outside of centromeres, reflecting the previous finding that T. monococcum has lost practically all centromeric tandem repeats [2].

Since CENH3 reads were enriched for RLG_Cereba and RLG_Quinta sequences, we cross-matched sequence read coverage of all ChIP-seq experiments with our annotation of full-length TEs. This was then used to map read coverage on consensus sequences of the two families allowing examination of read coverage at a near base pair resolution. Interestingly, CENH3-associated sequences came particularly often from RLG_Cereba and RLG_Quinta LTRs (Fig. 6C and D). Additionally, the ratio between CENH3 and CENH3 mock was highest in LTRs (Fig. 6C). These data indicate that RLG_Cereba and RLG_Quinta LTRs have a particular affinity for CENH3 containing nucleosomes (Fig. 6C and D).

A 162 bp motif in the RLG_Quinta LTR is associated with strict placement of nucleosomes

MNase treatment is known to be very sensitive to over-digestion of DNA [51] and some levels of over-digestion have to be expected. This can lead to a general enrichment of sequence motifs that are (i) MNase resistant and (ii) strongly phasing nucleosomes (i.e. strictly keeping nucleosomes in place once they associate with the given sequence). In our case, this led to the identification of a particular ~ 162 bp motif in the 5’ half of the RLG_Quinta LTR (the region which is not conserved between RLG_Cereba and RLG_Quinta LTRs, Supplementary Fig. S9A, see Fig. 1A). This sequence (hereafter called QuinCent) showed 10–15 times the average read coverage in all samples, including controls and mock inoculations (Fig. 6D). This indicates that the QuinCent motif generally interacts with nucleosomes and holds them strictly in place. It also suggests that this sequence is highly resistant to MNase treatment, which would explain its high abundance in the two mock controls. The QuinCent motif has a cryptic inverted repeat structure over most of its length (Suppl. Fig. S9B). We speculate that this could be a functional feature, as it could make binding of nucleosomes strand independent. There are various studies on conserved sequence motifs which promote nucleosome positioning (e.g. [24, 60, 61]). However, we did not find any of the previously described motifs. It is thus possible that the QuinCent sequence represents a novel type of nucleosome positioning sequence.

A strict positioning/phasing of nucleosomes should, consequently, influence the placements of neighboring nucleosomes. This is indeed visible in our mapping of ChIP-seq coverage on the RLG_Quinta consensus sequence (Fig. 6D and E): the QuinCent motif presumably acts as an “anchor” point which determines the placement of nearby nucleosomes, which is visible in weaker but periodic neighboring peaks in ChIP-seq read coverage (Fig. 6D and E. The proposed placing/phasing of CENH3 containing nucleosomes in the LTR of RLG_Quinta is surprisingly similar to the situation described for the LTR of HIV where nucleosomes Nuc-0, Nuc-1 and Nuc-2 are strictly placed in the LTR [36], defining a region accessible to transcription factors between Nuc-0 and Nuc-1. The proposed nucleosome positioning in RLG_Quinta places Nuc-0 and Nuc-1 up- and downstream of the region that shows homology to the HIV promoter, and which contains the predicted TATA box (Fig. 6E).

Because the Nuc-0 binding QuinCent motif lies in a region where RLG_Quinta and RLG_Cereba LTRs are completely different (see Fig. 1), we assumed that RLG_Quinta has acquired this motif from another genomic source. Indeed, we found a homologous sequence inside the LTRs of the Copia retrotransposon family RLC_Gisela (Suppl Fig. 9C). We propose that RLG_Quinta acquired this motif, for example through gene conversion, thereby replacing the ancestral LTR segment.

Conclusions

Our study provided detailed insight into the role of centromere-specific retrotransposons in the function and evolution of centromeres of Einkorn wheat (T. monococcum). The diploid T. monococcum has diverged from the A genome of hexaploid wheat less than 1 million years ago [28, 31], suggesting the situation to be very similar in the two. Additionally, previous studies found that parts of RLG_Cereba and RLG_Quinta retrotransposons have a strong propensity to associate with or phase CENH3 in the A, B and D subgenomes of hexaploid wheat [22, 27, 58]. Furthermore, genomic organization and TE content of all three wheat subgenomes are very similar [66, 69]. We therefore suggest that our findings also hold true for the three subgenomes of hexaploid wheat, despite them having diverged 3–7 million years ago [28, 31].

We found that the non-autonomous RLG_Quinta family evolved from autonomous RLG_Cereba retrotransposons at least 28 million years ago in the ancestor of Triticeae and oat [42]. We propose that RLG_Quinta relies on the RT and INT proteins of RLG_Cereba, but presumably also contributes GAG proteins which are needed in high numbers for the virus-like particles in which reverse transcription takes place. During their evolution, the LTRs of RLG_Cereba and RLG_Quinta evolved the function to phase and/or position CENH3-containing nucleosomes, rendering the ancestral centromere-specific tandem repeats unnecessary and leading to their eventual loss in the T. monococcum genome. Here, RLG_Quinta went through a particular evolutionary step when it acquired the novel QuintCent sequence motif from another retrotransposon. This gave it a strong propensity to precisely place nucleosomes in a manner that is strikingly similar to the strict placing of nucleosomes in the LTR of HIV [36]. Retroviruses are widely believed to have evolved from Gypsy retrotransposons in animals [41]. However, it is still intriguing that nucleosome phasing in their LTRs has remained so similar despite HIV and plant centromere-specific retrotransposons having diverged hundreds of millions of years ago.

We therefore suggest that RLG_Quinta has evolved beyond being purely a parasite of RLG_Cereba. On one hand it contributes to the replication process of both RLG_Cereba and RLG_Quinta through the contribution of GAG proteins. On the other hand it contributes to the function of the host centromere. We propose that, in this way, RLG_Quinta evolved from a purely parasitic to mutualistic genomic element.

We combined our findings in a model for the evolution and current dynamics of wheat centromeres (Fig. 7). Our data on insertion ages indicate that RLG_Cereba retrotransposons are consistently active, providing a steady flow of new insertions into active centromeres. RLG_Cereba retrotransposons also cross-mobilize RLG_Quinta retrotransposons (Fig. 7A, step 1). Here, the two retrotransposons presumably have complementary functions: first, RLG_Cereba contributes enzymes for replication and insertion, while RLG_Quinta provides GAG proteins in the required high quantities for the formation of virus-like particles in which replication takes place. Second, the CR domain attached to the RLG_Cereba INT enzyme ensures that new RLG_Cereba and RLG_Quinta copies are inserted into the functional centromere. Third, newly inserted copies then promote phasing/placement of CENH3-containing nucleosomes to their LTR sequences (Fig. 7A, step 2). Over time, the physical growth of the functional centromere through new retrotransposon insertions is compensated through loss of CENH3 in its distal regions (Fig. 7B, step 3). It is possible that centromere size is simply determined by the amount of available CENH3 protein [68]. Additional CENH3 may be accumulated in the functional centromere over time through subsequent epigenetic processes [32, 33].

Fig. 7
figure 7

Models for the function and dynamics of retrotransposons in T. monococcum centromeres. A Schematic simplified model of the RLG_Cereba integrase complex. The CR domain specifically binds directly to CENH3 in nucleosomes, thereby ensuring selective insertion into functional centromeres. The CR domain is connected through a  tether/linker to the integrase core enzyme, directing the insertion of the double-stranded DNA (dsDNA) to a nearby loop in the chromatin fiber. B Sequence turnover and maintenance of centromeres. Step 1: The nucleosomes in and around centromeres are shown simplified by colored circles. The autonomous RLG_Cereba retrotransposons (blue boxes) replicate and insert copies of themselves into centromeres. Additionally, they also cross-mobilize RLG_Quinta retrotransposons. New RLG_Cereba and RLG_Quinta copies are inserted near nucleosomes that contain CENH3 histone variants. Step 2: The presence of QuinCent sequence motifs (see Fig. 6) in RLG_Quinta promotes recruitment of new CENH3 histone variants through unknown mechanisms. Step 3: Over time, CENH3 variants in distal parts are lost, thereby maintaining size the functional centromere. Additional CENH3 may be deposited in the functional centromere over time

Full size image

The precise molecular mechanism of CENH3 deposition in plants is still an active research field, and our study can only contribute evidence that specific sequences (such as QuinCent) may be essential. Indeed, previous work found a strong association of RLG_Quinta sequences with CENH3 deposition [22]. We also emphasize that we only provide bioinformatical evidence for the interaction of the CR domain with CENH3. It was beyond the scope and resources of this study to provide functional proof, and extensive wet lab experiments such as Co-immuno precipitation will be needed to conclusively demonstrate a direct protein–protein interaction. Despite its limitations, our study highlights the complex interplay of a pair of autonomous and non-autonomous retrotransposons in the environment of plant centromeres. Whether this quasi-symbiotic relationship is an exception that only arose in Triticeae or whether mutualistic pairs of retrotransposons shape the centromeres of other plants remains an open question.

Methods

Software sources

Unless stated otherwise in the methods section, bioinformatics software was obtained from Ubuntu repositories (ubuntu.com).

Transposable element annotation

Full-length RLG_Cereba and RLG_Quinta retrotransposons were identified and annotated with the previously described TEpop pipeline [71] which uses multiple consensus sequences of LTRs and searches for occurrences of LTRs in the same orientation and in the distance from each other that is roughly expected from the length of the consensus sequence of the respective retrotransposon family. In a second step, candidate full-length copies are screened for the presence of the predicted coding sequences (CDS) to discard cases in which two LTRs were found at the right distance by chance. For RLG_Cereba retrotransposons, we selected copies ranging in size from 7,700–8,000 bp, to also allow for copies that contain small insertions or deletion. The RLG_Quinta populations are more complex, as they contain two groups that differ in size due to variable lengths of the predicted UTR. Here, we size-selected for copies of 4,300–4,500 bp and 4,700–4,800 bp, respectively. Full-length RLG_Cereba and RLG_Quinta retrotransposon homologs were isolated from the T. monococcum TA299 and T10622 genome assemblies [2], as well as from the previously published genome assemblies of barley [29], Brachypodium [65], oat [20], rye [47], maize [17], rice [21], Pharus latifolius [26] and Streptochaeta angustifolia [52].

Retrotransposon insertion age estimates

Insertion ages of individual retrotransposon copies were estimated by aligning the two LTRs of each copy with the EMBOSS program Water (obtained from Ubuntu repositories, ubuntu.com), using a gap opening penalty of 10 and a gap extension penalty of 0.5. Nucleotide differences between LTRs were counted and transitions and transversions were distinguished for molecular dating as previously described [5]. For all molecular dating, a rate of 1.3E-8 per site per year proposed for intergenic regions in grasses was used [25]. Molecular dating of insertions times was automated with the in-house Perl script date_pair.

Analyses of populations of RLG_Cereba and RLG_Quinta retrotransposons

All identified full-length RLG_Cereba and RLG_Quinta retrotransposon copies were aligned to their respective consensus sequences using the EMBOSS program Water, using a gap opening penalty of 50 and a gap extension penalty of 0.1. The alignments were then transformed into a variant call file (vcf) with the previously described Perl script pair_to_vcf [71]. Sequence variants were used if they occur in at least 1% of all retrotransposon copies (i.e. minor allele frequency of 1%). Insertions in retrotransposon copies were ignored, and deletions were treated as missing data, with a missing data cutoff at 90%. The vcf file was then used for principal component analysis (PCAs) using the R libraries gdsfmt, SNPRelate, ggplot2 and magrittr. Consensus sequences for defined subfamilies were constructed from 30 randomly picked full-length copies, which were aligned with Clustalw at default settings.

The consensus sequences for individual subfamilies were then used for subsequent analyses which included: (i) prediction of hypothetical encoded proteins, (ii) comparisons between RLG_Cereba and RLG_Quinta retrotransposons (shown in Fig. 1a), and (iii) multiple alignments of LTRs and PBS and PPT regions (shown in Fig. 1B and C). Hypothetical proteins were used for sequence comparisons and phylogenetic analyses (Fig. 2A, Supplementary Fig. S9).

Analysis of CENH3 ChIP-seq data

Mapped reads for CENH3, H3K4me2 and their respective controls were downloaded from: (https://doi.org/10.5061/dryad.0p2ngf24b). The samtools depth command was used to calculate per base sequence coverage. The CENH3 read depth coverage was then cross-matched with our TE annotations. For the calculation of average CENH3 read coverage per TE-copy, all annotated TEs of at least 1000 bp were used. For the identification of specific TE regions with high CENH3 read coverage, the individual annotated copies were aligned to the consensus sequence of the respective TE family, omitting insertions in the aligned copies in order to map CENH3 hot spots on the TE consensus sequence.

For the identification of CENH3-specific hot spots, the ChIP-seq read mappings were screened for segments of at least 100 bp where CENH3 read coverage was at least 5 times the average CENH3 coverage, and at least 5 times higher than the sum of the read coverage of all control and mock experiments. As control, we also searched the terminal 100 Mb of chromosome 1A for such peaks. Because of high enrichment of CENH3 in centromeres, the average CENH3 coverage was calculated once only for all predicted centromeric regions and for the control for the data from the terminal 100 Mb of chromosome 1A. The same procedure was used for C1_H3K4me3 ChIP-seq data. The sequence peaks identified in this way where then used in blastn searches against the TREP database (www.botinst.uzh.ch/en/research/genetics/thomasWicker/trep-db.html). Segments were classified as belonging to a given TE family, if they produced blastn alignments >= 90 bp and had at least 70% sequence identity.

3D protein modelling and visualization

3 dimensional protein models were generated using ColabFold/Alphafold2 (https://doi.org/10.1038/s41592-022-01488-1 / https://doi.org/10.1038/s41586-021-03819-2), using the following settings: template_mode = none, mas_mode:MMseqs2(UniRef + Environmental), num_recycles = 3. Protein structures were visualized usingPyMOL(https://pymol.org). The schematic model for the complex of RLG_Cereba integrase with the chromatin fiber (shown in Fig. 7B) was created using Blender (http://www.blender.org).

Identification of RLG_Cereba and RLG_Quinta homologs in species other than wheat

The predicted protein sequences containing the GAG protein from T. monococcum RLG_Cereba and RLG_Quinta consensus proteins were used in tblastn searches against the genomes of Brachypodium, barley, oat and rice. All regions producing blast hits > 90 aa and > 35% protein sequence identity were extracted from the respective genomes adding 5000 bp of flanking sequence. The extracted sequences were then aligned using Clustalw. The alignments were then trimmed, and a consensus sequence was inferred using the in-house script visual_clustal. The consensus sequences were then visually inspected by dotplot e.g. for completeness of LTRs.

Consensus sequences of CR domain containing retrotransposons

CR domains of retrotransposons were downloaded from [38]. The CR domains were taken as a seed for blastn queries against their respective genome of origin [3, 7, 13, 16, 17, 19, 21, 23, 34, 40, 48, 49, 56, 62, 67, 72, 76,77,78] regions producing blast hits were extracted and processed as described above to generate consensus sequences.

Phylogenetic analyses of consensus sequences

DNA or protein sequences were aligned with CustalW. Multiple alignments were converted to nexus format with Clustalx. Alignments were visually inspected for proper alignment of sequences. Phylogenetic trees were constructed with MrBayes using the mcmc algorithm. Generations were added until the average standard deviation of split frequencies dropped below 0.01. For trees of nucleotide sequences, the option lset nst = 6 rates = invgamma was used. For all trees, a burn-in of 25% was used. Phylogenetic trees were visualized with FigTree.

Phylogenetic analyses of individual TE copies of Einkorn

To exclude recombined elements we first filtered the isolated TE copies, based on the similarity between the 5’ and 3’ target site duplication, allowing for 1 mismatch. The copies passing this filtering were then aligned with MAFFT version 7.525 with the following parameters: –reorder –maxiterate 1000 –nomemsave –leavegappyregion –6merpair. Maximum likelihood trees were then estimated with RAxML version 8.2.12 as described in (https://doi.org/10.1371/journal.ppat.1011130). Briefly, 10 maximum likelihood trees were estimated with: raxmlHPC-PTHREADS-SSE3 -m GTRGAMMA -p 12,345 -# 10 –print-identical-sequences -s [alignment].phy. Then we performed boostrap analysis on the best tree using the parameters: raxmlHPC-PTHREADS-SSE3 -m GTRGAMMA -p 12,345 -b 12,345 -# 50 –print-identical-sequences -s [alignment].phy -t [best_tree] and the bipartition were added to the best tree with: raxmlHPC-PTHREADS-SSE3 -m GTRGAMMA -p 12,345 -f b –print-identical-sequences -s [alignment].phy -t [best_tree_with_bootstrap]. Information on whether an individual copy is part of a centromere was obtained from [2]. To determine whether a copy likely inserted into the newly formed portion of the Chr4Am centromere, copies had to be located in the genomic region of 281.1 Mb-284 Mb and have an estimated insertion ago of 100′000 years or younger, as the centromere shift was dated to approximately 30′000–100′000 years ago. Phylogenetic trees with the corresponding annotations were visualized in R using the packages: ggplot2, gtree, ggtreeExtra and ggnewscale.

Mantel test for association between genetic distance and genomic position of TEs

The vcf file obtained from TEpop was converted to hapmap and then reformatted using TASSEL5 and R. Pairwise comparison of all individual elements to count SNPs was done using the dist.gene function from the ape package. The difference of genomic positions was calculated in relation to the midpoint of the centromere of the respective chromosome where a TE copy is situated. This means that if the centromere midpoint of a chromosome is at 250 Mb and a TE copy is at 245 Mb the genomic position in relation to the centromere midpoint was counted as 5 Mb. Another copy situated at 255 Mb would also be counted as 5 Mb meaning that they have no positional difference for this analysis. Association between positional and genetic distance was then assessed using the R function mantel() with the parameters: method = ”spearman”, permutations = 9999, na.rm = TRUE. Linear regression was calculated using the R function lm() with the formula x ~ y.

Availability of data and materials

Source data not otherwise publicly available can be accessed at: https://github.com/Wicker-Lab/wheat_centromere_materials.

References

  1. Abascal-Palacios G, Jochem L, Pla-Prats C, et al. Structural basis of Ty3 retrotransposon integration at RNA Polymerase III-transcribed genes. Nat Commun. 2021;12:1–11. https://doi.org/10.1038/s41467-021-27338-w.

    Article  CAS  Google Scholar 

  2. Ahmed HI, Heuberger M, Schoen A, et al. Einkorn genomics sheds light on history of the oldest domesticated wheat. Nature. 2023;620:830–8. https://doi.org/10.1038/s41586-023-06389-7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Bennetzen JL, Schmutz J, Wang H, et al. (2012) Reference genome sequence of the model plant Setaria. Nat Biotechnol. 2012;306(30):555–61. https://doi.org/10.1038/nbt.2196.

    Article  CAS  Google Scholar 

  4. Berkhoutt B, Jeang K-T. Functional roles for the TATA promoter and enhancers in basal and Tat-induced expression of the human immunodeficiency virus type 1 long terminal repeat. J Virol. 1992;66:139–49. https://doi.org/10.1128/JVI.66.1.139-149.1992.

    Article  Google Scholar 

  5. Buchmann JP, Matsumoto T, Stein N, et al. Inter-species sequence comparison of Brachypodium reveals how transposon activity corrodes genome colinearity. Plant J. 2012;71:550–63. https://doi.org/10.1111/j.1365-313X.2012.05007.x.

    Article  CAS  PubMed  Google Scholar 

  6. Cheng Z, Dong F, Langdon T, et al. Functional rice centromeres are marked by a satellite repeat and a centromere-specific retrotransposon. Plant Cell. 2002;14:1691–704. https://doi.org/10.1105/tpc.003079.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Cooper EA, Brenton ZW, Flinn BS, et al. A new reference genome for Sorghum bicolor reveals high levels of sequence similarity between sweet and grain genotypes: implications for the genetics of sugar metabolism. BMC Genomics. 2019;20:1–13. https://doi.org/10.1186/S12864-019-5734-X/FIGURES/5.

    Article  CAS  Google Scholar 

  8. Cumberledge S, Carbon J. Mutational analysis of meiotic and mitotic centromere function in Saccharomyces cerevisiae. Genetics. 1987;117:203–12. https://doi.org/10.1093/genetics/117.2.203.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Dumont M, Fachinetti D. DNA sequences in centromere formation and function. Prog Mol Subcell Biol. 2017;56:305–36. https://doi.org/10.1007/978-3-319-58592-5_13.

    Article  CAS  PubMed  Google Scholar 

  10. Earnshaw WC. Discovering centromere proteins: from cold white hands to the A, B, C of CENPs. Nat Rev Mol Cell Biol. 2015;16:443–9. https://doi.org/10.1038/nrm4001.

    Article  CAS  PubMed  Google Scholar 

  11. Felsenfeld G. Groudine M (2003) Controlling the double helix. Nat. 2003;4216921(421):448–53. https://doi.org/10.1038/nature01411.

    Article  CAS  Google Scholar 

  12. Gao X, Hou Y, Ebina H, et al. Chromodomains direct integration of retrotransposons to heterochromatin. Genome Res. 2008;18:359–69. https://doi.org/10.1101/gr.7146408.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Hellsten U, Wright KM, Jenkins J, et al. Fine-scale variation in meiotic recombination in Mimulus inferred from population shotgun sequencing. Proc Natl Acad Sci U S A. 2013;110:19478–82. https://doi.org/10.1073/PNAS.1319032110/SUPPL_FILE/PNAS.201319032SI.PDF.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Henikoff S, Henikoff JG. “Point” centromeres of saccharomyces harbor single centromere-specific nucleosomes. Genetics. 2012;190:1575–7. https://doi.org/10.1534/GENETICS.111.137711.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Hu WS, Temin HM. Retroviral recombination and reverse transcription. Science. 1990;250(4985):1227–33. https://doi.org/10.1126/science.1700865.

    Article  CAS  PubMed  Google Scholar 

  16. Hu TT, Pattyn P, Bakker EG, et al. (2011) The Arabidopsis lyrata genome sequence and the basis of rapid genome size change. Nat Genet. 2011;435(43):476–81. https://doi.org/10.1038/ng.807.

    Article  CAS  Google Scholar 

  17. Hufford MB, Seetharam AS, Woodhouse MR, et al. De novo assembly, annotation, and comparative analysis of 26 diverse maize genomes. Science. 2021;373:655–62. https://doi.org/10.1126/SCIENCE.ABG5289/SUPPL_FILE/SCIENCE.ABG5289_MDAR_REPRODUCIBILITY_CHECKLIST.PDF.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Jiang J, Birchler JA, Parrott WA, Dawe RK. A molecular view of plant centromeres. Trends Plant Sci. 2003;8:570–5. https://doi.org/10.1016/J.TPLANTS.2003.10.011.

    Article  CAS  PubMed  Google Scholar 

  19. Kamal N, Mun T, Reid D, et al. Insights into the evolution of symbiosis gene copy number and distribution from a chromosome-scale Lotus japonicus Gifu genome sequence. DNA Res. 2020;27. https://doi.org/10.1093/DNARES/DSAA015.

  20. Kamal N, TsardakasRenhuldt N, Bentzer J, et al. The mosaic oat genome gives insights into a uniquely healthy cereal crop. Nature. 2022;606:113–9. https://doi.org/10.1038/s41586-022-04732-y.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Kawahara Y, de la Bastide M, Hamilton JP, et al. Improvement of the oryza sativa nipponbare reference genome using next generation sequence and optical map data. Rice. 2013;6:3–10. https://doi.org/10.1186/1939-8433-6-4/FIGURES/2.

    Article  Google Scholar 

  22. Li B, Choulet F, Heng Y, et al. Wheat centromeric retrotransposons: the new ones take a major role in centromeric structure. Plant J. 2013;73:952–65. https://doi.org/10.1111/tpj.12086.

    Article  CAS  PubMed  Google Scholar 

  23. Lovell JT, Jenkins J, Lowry DB, et al. The genomic landscape of molecular responses to natural drought stress in Panicum hallii. Nat Commun. 2018;9:5213–5213. https://doi.org/10.1038/S41467-018-07669-X.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Lowary PT, Widom J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning. J Mol Biol. 1998;276:19–42. https://doi.org/10.1006/JMBI.1997.1494.

    Article  CAS  PubMed  Google Scholar 

  25. Ma J, Bennetzen JL. Rapid recent growth and divergence of rice nuclear genomes. Proc Natl Acad Sci U S A. 2004;101:12404–10. https://doi.org/10.1073/pnas.0403715101.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Ma PF, Liu YL, Jin GH, et al. The Pharus latifolius genome bridges the gap of early grass evolution. Plant Cell. 2021;33:846–64. https://doi.org/10.1093/PLCELL/KOAB015.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Ma H, Ding W, Chen Y, et al. Centromere plasticity with evolutionary conservation and divergence uncovered by wheat 10+ genomes. Mol Biol Evol. 2023;40(8):msad176. https://doi.org/10.1093/molbev/msad176.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Marcussen T, Sandve SR, Heier L, et al. Ancient hybridizations among the ancestral genomes of bread wheat. Science. 2014;345. https://doi.org/10.1126/science.1250092.

  29. Mascher M, Gundlach H, Himmelbach A, et al. A chromosome conformation capture ordered sequence of the barley genome. Nature. 2017;544:427–33. https://doi.org/10.1038/nature22043.

    Article  CAS  PubMed  Google Scholar 

  30. Mascher M, Wicker T, Jenkins J, et al. Long-read sequence assembly: a technical evaluation in barley. Plant Cell. 2021. https://doi.org/10.1093/plcell/koab077.

  31. Middleton CP, Senerchia N, Stein N, et al. Sequencing of chloroplast genomes from wheat, barley, rye and their relatives provides a detailed insight into the evolution of the triticeae tribe. PLoS One. 2014;9. https://doi.org/10.1371/journal.pone.0085761.

  32. Müller S, Almouzni G. A network of players in H3 histone variant deposition and maintenance at centromeres. Biochim Biophys Acta - Gene Regul Mech. 2014;1839:241–50. https://doi.org/10.1016/j.bbagrm.2013.11.008.

    Article  CAS  Google Scholar 

  33. Müller S, Almouzni G. Chromatin dynamics during the cell cycle at centromeres. Nat Rev Genet. 2017;18:192–208. https://doi.org/10.1038/nrg.2016.157.

    Article  CAS  PubMed  Google Scholar 

  34. Myburg AA, Grattapaglia D, Tuskan GA, et al. (2014) The genome of Eucalyptus grandis. Nat. 2014;5107505(510):356–62. https://doi.org/10.1038/nature13308.

    Article  CAS  Google Scholar 

  35. Naish M, Alonge M, Wlodzimierz P, et al. The genetic and epigenetic landscape of the Arabidopsis centromeres. Science. 2021;374:7489. https://doi.org/10.1126/SCIENCE.ABI7489/SUPPL_FILE/SCIENCE.ABI7489_MDAR_REPRODUCIBILITY_CHECKLIST.PDF.

    Article  Google Scholar 

  36. Ne E, Palstra RJ, Mahmoudi T. Transcription: insights from the HIV-1 promoter. Int Rev Cell Mol Biol. 2018;335:191–243. https://doi.org/10.1016/BS.IRCMB.2017.07.011.

    Article  CAS  PubMed  Google Scholar 

  37. Neumann P, Navrátilová A, Koblížková A, et al. Plant centromeric retrotransposons: a structural and cytogenetic perspective. Mob DNA. 2011;2:1–16. https://doi.org/10.1186/1759-8753-2-4.

    Article  CAS  Google Scholar 

  38. Neumann P, Novák P, Hoštáková N, MacAs J. Systematic survey of plant LTR-retrotransposons elucidates phylogenetic relationships of their polyprotein domains and provides a reference for element classification. Mob DNA. 2019;10:1–17. https://doi.org/10.1186/s13100-018-0144-1.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Neumann P, Yan H, Jiang J. The centromeric retrotransposons of rice are transcribed and differentially processed by RNA interference. Genetics. 2007;176:749–61. https://doi.org/10.1534/GENETICS.107.071902.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Pecrix Y, Staton SE, Sallet E, et al. (2018) Whole-genome landscape of Medicago truncatula symbiotic genes. Nat Plants. 2018;412(4):1017–25. https://doi.org/10.1038/s41477-018-0286-7.

    Article  CAS  Google Scholar 

  41. Pélisson A, Teysset L, Chalvet F, et al. About the origin of retroviruses and the co-evolution of the gypsy retrovirus with the Drosophila flamenco host gene. Genetica. 1997;100:29–37. https://doi.org/10.1023/A:1018336303298.

    Article  PubMed  Google Scholar 

  42. Peng Y, Yan H, Guo L, et al. Reference genome assemblies reveal the origin and evolution of allohexaploid oat. Nat Genet. 2022;54:1248–58. https://doi.org/10.1038/s41588-022-01127-7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Pereira V. Insertion bias and purifying selection of retrotransposons in the Arabidopsis thaliana genome. Genome Biol. 2004;5(10):R79. https://doi.org/10.1186/gb-2004-5-10-r79.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Peterson-Burch BD, Nettleton D, Voytas DF. Genomic neighborhoods for Arabidopsis retrotransposons: a role for targeted integration in the distribution of the Metaviridae. Genome Biol. 2004;5(10):R78. https://doi.org/10.1186/gb-2004-5-10-r78. Epub 2004 Sep 29.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Prasad V, Strömberg CAE, Leaché AD, et al. (2011) Late Cretaceous origin of the rice tribe provides evidence for early diversification in Poaceae. Nat Commun. 2011;21(2):1–9. https://doi.org/10.1038/ncomms1482.

    Article  CAS  Google Scholar 

  46. Presting GG, Malysheva L, Fuchs J, Schubert I. A TY3/GYPSY retrotransposon-like sequence localizes to the centromeric regions of cereal chromosomes. Plant J. 1998;16:721–8. https://doi.org/10.1046/J.1365-313X.1998.00341.X.

    Article  CAS  PubMed  Google Scholar 

  47. Rabanus-Wallace MT, Hackauf B, Mascher M, et al. Chromosome-scale genome assembly provides insights into rye biology, evolution and agronomic potential. Nat Genet. 2021;53:564–73. https://doi.org/10.1038/s41588-021-00807-0.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Schmutz J, Cannon SB, Schlueter J, et al. (2010) Genome sequence of the palaeopolyploid soybean. Nat. 2010;4637278(463):178–83. https://doi.org/10.1038/nature08670.

    Article  CAS  Google Scholar 

  49. Schmutz J, McClean PE, Mamidi S, et al. (2014) A reference genome for common bean and genome-wide analysis of dual domestications. Nat Genet. 2014;467(46):707–13. https://doi.org/10.1038/ng.3008.

    Article  CAS  Google Scholar 

  50. Schnable PS, Pasternak S, Liang C, et al. The B73 Maize Genome: Complexity, Diversity, and Dynamics. Science. 2009;80(326):1112–5.

    Article  Google Scholar 

  51. Schwartz U, Németh A, Diermeier S, et al. Characterizing the nuclease accessibility of DNA in human cells to map higher order structures of chromatin. Nucleic Acids Res. 2019;47:1239–54. https://doi.org/10.1093/NAR/GKY1203.

    Article  CAS  PubMed  Google Scholar 

  52. Seetharam AS, Yu Y, Bélanger S, et al. The Streptochaeta Genome and the Evolution of the Grasses. Front Plant Sci. 2021;12:710383. https://doi.org/10.3389/FPLS.2021.710383/BIBTEX.

    Article  PubMed  PubMed Central  Google Scholar 

  53. Serra-Cardona A, Zhang Z. Replication-coupled nucleosome assembly in the passage of epigenetic information and cell identity. Trends Biochem Sci. 2018;43:136–48. https://doi.org/10.1016/J.TIBS.2017.12.003.

    Article  CAS  PubMed  Google Scholar 

  54. Sharma A, Presting GG. Centromeric retrotransposon lineages predate the maize/rice divergence and differ in abundance and activity. Mol Genet Genomics. 2008;279:133–47. https://doi.org/10.1007/S00438-007-0302-5.

    Article  CAS  PubMed  Google Scholar 

  55. Sharma A, Presting GG. Evolution of centromeric retrotransposons in grasses. Genome Biol Evol. 2014;6:1335–52. https://doi.org/10.1093/gbe/evu096.

    Article  PubMed  PubMed Central  Google Scholar 

  56. Shi X, Cao S, Wang X, et al. The complete reference genome for grapevine (Vitis vinifera L.) genetics and breeding. Hortic Res. 2023;10. https://doi.org/10.1093/HR/UHAD061.

  57. Steiner FA, Henikoff S. Holocentromeres are dispersed point centromeres localized at transcription factor hotspots. Elife. 2014;2014. https://doi.org/10.7554/ELIFE.02025.

  58. Su H, Liu Y, Liu C, et al. (2019) Centromere satellite repeats have undergone rapid changes in polyploid wheat subgenomes. Plant Cell. 2019;31(9):2035–51. https://doi.org/10.1105/tpc.19.00133.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Tachiwana H, Kagawa W, Shiga T, et al. Crystal structure of the human centromeric nucleosome containing CENP-A. Nature. 2011;476:232–5. https://doi.org/10.1038/nature10258.

    Article  CAS  PubMed  Google Scholar 

  60. Thåström A, Lowary PT, Widlund HR, et al. Sequence motifs and free energies of selected natural and non-natural nucleosome positioning DNA sequences. J Mol Biol. 1999;288:213–29. https://doi.org/10.1006/JMBI.1999.2686.

    Article  PubMed  Google Scholar 

  61. Trifonov EN. Cracking the chromatin code: Precise rule of nucleosome positioning. Phys Life Rev. 2011;8:39–50. https://doi.org/10.1016/J.PLREV.2011.01.004.

    Article  PubMed  Google Scholar 

  62. Tuskan GA, DiFazio S, Jansson S, et al. The genome of black cottonwood, Populus trichocarpa (Torr & Gray). Science. 2006;80(313):1596–604. https://doi.org/10.1126/SCIENCE.1128691.

    Article  Google Scholar 

  63. Vicient CM, Kalendar R, Schulman AH. Variability, recombination, and mosaic evolution of the barley BARE-1 retrotransposon. J Mol Evol. 2005;61(3):275–91. https://doi.org/10.1007/s00239-004-0168-7.

    Article  CAS  PubMed  Google Scholar 

  64. Vitte C, Panaud O. LTR retrotransposons and flowering plant genome size: emergence of the increase/decrease model. Cytogenet Genome Res. 2005;110:91–107. https://doi.org/10.1159/000084941.

    Article  CAS  PubMed  Google Scholar 

  65. Vogel JP, Garvin DF, Mockler TC, et al. Genome sequencing and analysis of the model grass Brachypodium distachyon. Nature. 2010;463:763–8. https://doi.org/10.1038/nature08747.

    Article  CAS  Google Scholar 

  66. Walkowiak S, Gao L, Monat C, et al. Multiple wheat genomes reveal global variation in modern breeding. Nature. 2020;588:277–83. https://doi.org/10.1038/s41586-020-2961-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. Wang M, Li J, Qi Z, et al. (2022) Genomic innovation and regulatory rewiring during evolution of the cotton genus Gossypium. Nat Genet. 2022;5412(54):1959–71. https://doi.org/10.1038/s41588-022-01237-2.

    Article  CAS  Google Scholar 

  68. Wang N, Dawe RK. Centromere size and its relationship to haploid formation in plants. Mol Plant. 2018;11:398–406. https://doi.org/10.1016/j.molp.2017.12.009.

    Article  CAS  PubMed  Google Scholar 

  69. Wicker T, Gundlach H, Spannagl M, et al. Impact of transposable elements on genome structure and evolution in bread wheat. Genome Biol. 2018;19:103. https://doi.org/10.1186/s13059-018-1479-0.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Wicker T, Schulman AH, Tanskanen J, et al. The repetitive landscape of the 5100 Mbp barley genome. Mob DNA. 2017;8:1–16. https://doi.org/10.1186/s13100-017-0102-3.

    Article  CAS  Google Scholar 

  71. Wicker T, Stritt C, Sotiropoulos AG, et al. Transposable Element Populations Shed Light on the Evolutionary History of Wheat and the Complex Co‐Evolution of Autonomous and Non‐Autonomous Retrotransposons. Adv Genet. 2022;2100022. https://doi.org/10.1002/ggn2.202100022.

  72. Wise K. Diseases of corn: northern corn leaf blight. Purdue Ext. 2011;6:1–3.

    Google Scholar 

  73. Wlodzimierz P, Rabanal FA, Burns R, et al. Cycles of satellite and transposon evolution in Arabidopsis centromeres. Nature. 2023;618(7965):557–65. https://doi.org/10.1038/s41586-023-06062-z.

    Article  CAS  PubMed  Google Scholar 

  74. Wolfgruber TK, Sharma A, Schneider KL, et al. Maize centromere structure and evolution: Sequence analysis of centromeres 2 and 5 reveals dynamic loci shaped primarily by retrotransposons. PLoS Genet. 2009;5:13–6. https://doi.org/10.1371/journal.pgen.1000743.

    Article  CAS  Google Scholar 

  75. Xiaoyan Zhong C, Marshall JB, Topp C, et al. Centromeric retroelements and satellites interact with maize kinetochore protein CENH3. Plant Cell. 2002;14:2825–36. https://doi.org/10.1105/tpc.006106.

    Article  CAS  Google Scholar 

  76. Yang R, Jarvis DE, Chen H, et al. The reference genome of the halophytic plant Eutrema salsugineum. Front Plant Sci. 2013;4:45219. https://doi.org/10.3389/FPLS.2013.00046/ABSTRACT.

    Article  Google Scholar 

  77. Zhang L, Cai X, Wu J, et al. (2018) Improved Brassica rapa reference genome by single-molecule sequencing and chromosome conformation capture technologies. Hortic Res. 2018;51(5):1–11. https://doi.org/10.1038/s41438-018-0071-9.

    Article  CAS  Google Scholar 

  78. Zhou R, Macaya-Sanz D, Carlson CH, et al. A willow sex chromosome reveals convergent evolution of complex palindromic repeats. Genome Biol. 2020;21:1–19. https://doi.org/10.1186/S13059-020-1952-4/FIGURES/8.

    Article  CAS  Google Scholar 

Download references

Acknowledgements

This work was supported by the following sources: TW and MH were supported by University of Zurich core funding. MH acknowledges support from Swiss National Science Foundation grant 310030_204165 awarded to Beat Keller. This work was supported by the King Abdullah University of Science and Technology. We are grateful to Ahmed et al. for providing pre-publication access to the T. monococcum genome assemblies.

Funding

M.H. was supported by the Swiss National Science Foundation grant 310030_212428. T.W. was supported by University of Zurich core funding. This work was supported by the King Abdullah University of Science and Technology.

Author information

Authors and Affiliations

  1. Department of Plant and Microbial Biology, University of Zurich, Zurich, Switzerland

    Matthias Heuberger & Thomas Wicker

  2. Wheat Genetics Resource Center and Department of Plant Pathology, Kansas State University, Manhattan, KS, 66506, USA

    Dal-Hoe Koo

  3. Plant Science Program, Biological and Environmental Science and Engineering Division (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal, 23955-6900, Saudi Arabia

    Hanin Ibrahim Ahmed, Michael Abrouk, Jesse Poland & Simon G. Krattinger

  4. Centre d’Anthropobiologie et de Génomique de Toulouse (CAGT), Université Paul Sabatier, Toulouse, France

    Hanin Ibrahim Ahmed

  5. Department of Plant Science and Landscape Architecture, University of Maryland, College Park, MD, 20724, USA

    Vijay K. Tiwari

Authors
  1. Matthias Heuberger
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Dal-Hoe Koo
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Hanin Ibrahim Ahmed
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Vijay K. Tiwari
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Michael Abrouk
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Jesse Poland
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Simon G. Krattinger
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Thomas Wicker
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

T.W., M.H., S.G.K., J.P., M.A., H.I.A., and V.K.T. designed the research. M.H. and T.W. conducted the main bioinformatic analyses and prepared the figures. D.-H.K. and J.P. processed data from CHiP experiments. M.H. and T.W. wrote the manuscript with input from S.G.K., J.P., V.K.T., M.A., H.I.A. and D.-H.K.

Corresponding author

Correspondence to Thomas Wicker.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Heuberger, M., Koo, DH., Ahmed, H.I. et al. Evolution of Einkorn wheat centromeres is driven by the mutualistic interplay of two LTR retrotransposons. Mobile DNA 15, 16 (2024). https://doi.org/10.1186/s13100-024-00326-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13100-024-00326-9

Keywords

Mobile DNA

ISSN: 1759-8753