Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Wide Distribution and Change of Target Specificity of R2 Non-LTR Retrotransposons in Animals

  • Kenji K. Kojima ,

    Affiliations Genetic Information Research Institute, Mountain View, CA, 94043, United States of America, Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Chiba, 277–8562, Japan, Department of Life Sciences, National Cheng Kung University, Tainan, 701, Taiwan

  • Yosuke Seto,

    Current address: Tokyo Metropolitan University, Hachioji, Tokyo, 192–0397, Japan

    Affiliation Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Chiba, 277–8562, Japan

  • Haruhiko Fujiwara

    Affiliation Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Chiba, 277–8562, Japan


Transposons, or transposable elements, are the major components of genomes in most eukaryotes. Some groups of transposons have developed target specificity that limits the integration sites to a specific nonessential sequence or a genomic region to avoid gene disruption caused by insertion into an essential gene. R2 is one of the most intensively investigated groups of sequence-specific non-LTR retrotransposons and is inserted at a specific site inside of 28S ribosomal RNA (rRNA) genes. R2 is known to be distributed among at least six animal phyla even though its occurrence is reported to be patchy. Here, in order to obtain a more detailed picture of the distribution of R2, we surveyed R2 using both in silico screening and degenerate PCR, particularly focusing on actinopterygian fish. We found two families of the R2C lineage from vertebrates, although it has previously only been found in platyhelminthes. We also revealed the apparent movement of insertion sites of a lineage of actinopterygian R2, which was likely concurrent with the acquisition of a 28S rRNA-derived sequence in their 3′ UTR. Outside of actinopterygian fish, we revealed the maintenance of a single R2 lineage in birds; the co-existence of four lineages of R2 in the leafcutter bee Megachile rotundata; the first examples of R2 in Ctenophora, Mollusca, and Hemichordata; and two families of R2 showing no target specificity. These findings indicate that R2 is relatively stable and universal, while differences in the distribution and maintenance of R2 lineages probably reflect characteristics of some combination of both R2 lineages and host organisms.


Transposons, or transposable elements, occupy considerable fractions of most eukaryotic genomes [1]. The insertion of transposons into a gene is associated with human genetic diseases and cancers [2]. Some groups of transposons, however, have developed target specificity that limits the integration sites to a specific sequence or a genomic region to avoid gene disruption. Although target specificity is reported in LTR retrotransposons and DNA transposons as well [3,4], non-LTR retrotransposons contain the widest variety of target-specific families. Their targets include almost all types of repetitive sequences both functional and non-functional: ribosomal RNA (rRNA) genes, transfer RNA (tRNA) genes, small nuclear RNA (snRNA) genes, microsatellites, telomeric repeats, and transposons [510].

R2 is one of the most intensively investigated groups of sequence-specific non-LTR retrotransposons. R2 was originally identified as an insertion sequence in the 28S rRNA genes of the fruit fly Drosophila melanogaster [11] and the domestic silkworm Bombyx mori [12], and was later characterized as a non-LTR retrotransposon [5,13]. R2 is widely distributed in arthropods [14]. Outside of arthropods, R2 was first reported in Chordata in the zebrafish Danio rerio and sea squirts Ciona intestinalis and Ciona savignyi [9]. To date, R2 has been reported in Echinodermata, Platyhelminthes, Nematoda, and Cnidaria, as well as Arthropoda and Chordata [1517].

Our previous analyses revealed that several lineages of R2 have been maintained for a long time in animals [15,16]. Four clades (supergroups) of R2: R2A, R2B, R2C, and R2D show independent lineages in the phylogenetic tree based on their reverse transcriptase sequences. They have a distinct number and type of zinc-fingers proximal to the reverse transcriptase domain. R2A has three zinc-fingers, two CCHH type and one CCHC type. R2B has two zinc-fingers, one CCHH type and one CCHC type [18]. R2C also has two zinc-fingers although both of them are of the CCHH type. R2D has only one zinc-finger, which is of the CCHH type. These zinc-fingers are responsible for target recognition, and interestingly the contribution of each zinc-finger to target recognition is different between clades [19,20]. These four clades were further classified into 11 total subclades [15].

R2 is inserted at a specific site inside of 28S rRNA genes [15]. It is dependent on the target-specific cleavage by the endonuclease encoded by R2 [21]. The bottom strand (antisense strand) cleavage is strictly determined, while the top strand (sense strand) shows some variations of cleavage sites, which are determined by the target site alterations upon insertion [16]. Two lineages of R2 have changed their target specificity within the array of rRNA genes. R8 from hydra is inserted in 18S rRNA genes while R9 from the rotifer is inserted at another site of 28S rRNA genes [16,22]. No lineage of R2 that has lost its target specificity has yet been identified.

Here, we surveyed R2 using both in silico screening and degenerate PCR. Of interest is the distribution and evolution of R2 in vertebrates. No R2 has been identified from mammals, amphibians, or chondrichthyes, while other groups of vertebrates include at least one species that possesses an R2. R2 has been reported from hagfishes [15], cyclostomes [23], actinopterygian fish [9], coelacanth [24], reptiles [15], and birds [17]. However, no systematic survey of R2 in vertebrates has been performed. Another topic is the origin and distribution of R2 in animals. R2 has been found in many animal phyla, but not yet from Mollusca and Annelida. R2 has neither been found outside the animal kingdom.

In this study, we found several independent lineages of R2 in actinopterygian fish, as well as a single lineage in birds. We also report the first R2 families from Ctenophora, Mollusca and Hemichordata. The loss of target specificity in platyhelminthes and the apparent shift of insertion sites in actinopterygian fish are observed.

Materials and Methods

In silico screening of R2

Genomic sequences of various species were obtained mostly from GenBank, and sequences of known TEs were obtained from Repbase [1] (

New R2 non-LTR retrotransposons were identified by repeated Blastn, tBlastn [25] and CENSOR [26] searches using genomic sequences of various animal species available at NCBI Blast ( and the UCSC Genome Browser ( websites with representatives of each R2 clade (R2A, B, C and D) as queries. To characterize the 5′ ends of R2 from birds, we used Blastn with the 5′ terminal sequences of several R2 families and 5′ flanking 28S rRNA gene sequences. To characterize the 3′ ends of R2 from birds, we used Blastn with the 3′ terminal sequence of R2-1_TG and the 3′ flanking 28S rRNA gene sequence. The classification was initially done by RTclass1 [27] and finally determined by phylogenetic analysis. The consensus sequences were derived using the majority rule applied to the corresponding set of multiple aligned copies of retrotransposons. All R2 sequences save for the 3′ end short fragment sequences were named following the systematic nomenclature implemented in Repbase and were submitted to Repbase [1] (

Characterization of R2 from genomic DNAs

Genomic DNA used for screening is shown in S1 Table. Genomic DNA was kindly provided by Dr. H. Mitani and Dr. S. Oda of U. Tokyo for four medaka species, by Dr. M. Nishida of U. Tokyo for other fishes, and by M. Park of U. Tokyo for reptiles. Tissue samples of salamander were kindly provided by Dr. T. Michiue, and tissues of other amphibians by Dr. M. Taira of the U. Tokyo.

To amplify R2 elements, four primers (R2IF1: 5′-AAGCARGGNGAYCCNCTNTC-3′, R2IIF1: 5′-GTNAARCARGGNGAYCCNCT-3′, R2IF2: 5′-GCYYTRGCGTTYGCNGAYGA-3′, R2IIF2: 5′-CTNGCNTTYGCNGAYGAYYT-3′) were designed from the highly conserved region of the RT domain. Four primers were also designed from the downstream 28S rRNA genes (28S_R-198: 5′- GCCTCCCACTTATYCTACACC-3′, 28S_R-147: 5′- GTCAAGCTCAACAGGGTCTTCT-3′, 28S_R-B: 5′-ATCCATTCATGCGCGTCACT-3′, 28S_R-A: 5′-TAGATGACGAGGCATTTGGC-3′). Using the R2 primer and 28S primer pair, PCR was performed with Ex-Taq (TaKaRa) for 35 cycles of 96°C for 1 min, 56°C for 20 s, and 72°C for 2 min. PCR products were cloned into the pGEM-T Easy vector (Promega) and sequenced with ABI PRISM 3130xl Genetic Analyzer (PE Applied Biosystems) and BigDye terminator v3.1 cycle sequencing kit (PE Applied Biosystems). Newly identified R2 sequences were deposited in the DNA Data Bank of Japan (DDBJ; and the accession numbers are shown in S1 Table.

Sequence alignment and phylogenetic analysis

Two types of protein sequence alignments were generated. One (alignment A; 258 sites) is the alignment of the partial RT domains spanning motif 5 to 9 of R2. The second (alignment B; 677 sites) is the alignment spanning RT motif 5 to the C-terminus of R2. In each alignment, sequences including ambiguous residues and/or deletions were excluded. All alignments were generated with the aid of MAFFT with the linsi option [28]. ProtTest was performed at the ProtTest server ( Amino acid substitution models were selected based on the Akaike Information Criterion and Bayesian Information Criterion. The models selected were LG+G+F for alignment A and LG+I+G+F for alignment B. Maximum likelihood trees were constructed by PhyML [29] with bootstrap values (100 replicates) using the respective substitution model. For each alignment, all sites or sites selected with the least strict option of Gblocks (129 sites for alignment A and 234 sites for alignment B) were used. As phylogenetic trees based on alignment with Gblocks selection showed weaker statistical support, the phylogenetic trees shown in the figures are based on all alignable sites. The phylogenetic trees were drawn with the aid of FigTree 1.3.1 (


R2 distribution in invertebrates

In addition to the six phyla from which R2 has been previously found, we found R2 families from three phyla: Ctenophora (sea walnut Mnemiopsis leidyi), Mollusca (Pacific oyster Crassostrea gigas) and Hemichordata (acorn worm Saccoglossus kowalevskii) (Fig 1A). We determined the two 3′ junctions for R2NS-1_CGi (oyster), which was previously reported as a non-sequence-specific family [30]. Here “NS” stands for the Non-sequence-Specific. One of the two copies is adjacent to the 28S rRNA gene, and thus it is not totally non-specific (S1 Fig); we renamed it R2-1_CGi (Fig 1A) to reflect this. We also found fragments of R2-like non-LTR retrotransposons from Priapulus caudatus (Priapulida), but could not determine the boundaries and thus we are not certain that it is specific for rRNA genes. We have analyzed repetitive sequences from various eukaryotes during the maintenance and expansion of Repbase [1], but have never found any R2 sequences outside animals (data not shown).

Fig 1. Junction sequences of R2 elements.

R2 family name, flanking 28S rRNA gene sequences (28S rDNA) with terminal sequences of R2, common name of the origin, and classification are shown from left to right. 28S rRNA gene sequences are in bold and shaded. (A) Both 5′ and 3′ junctions of non-avian R2 copies. (B) 3′ junctions of R2 copies detected by PCR. Scientific names and common English names of the origins of R2 elements are as follows: R2-1_MLe, Mnemiopsis leidyi (sea walnut); R2-1_SK, Saccoglossus kowalevskii (acorn worm); R2NS-1_CGi, Crassostrea gigas (Pacific oyster); R2-2_SMed, Schimidtea mediterranea (planarian); R2-1_SP, Strongylocentrotus purpuratus (purple sea urchin); R2-1_LV, Lythechinus variegatus (green sea urchin); R2-1_GA, Gasterosteus aculeatus (three-spined stickleback); R2-1_SSa, Salmo salar (Atlantic salmon); R2-1_AMi, Alligator mississippiensis (American alligator); R2-1_Crp, Crocodilus porosus (saltwater crocodile); R2-1_Gav, Gavialis gangeticus (gharial); R2Pp, Pungitius pungitius pungitius (nine-spined stickleback); R2Ao, Anarhichas orientalis (Bering wolffish); R2Cm, Crystallichthys matsushimae (snailfish); R2Om, Oryzias melastigma (marine medaka); R2Tcc, Tylosurus crocodilus crocodilus (houndfish); R2Tch, Theragra chalcogramma (Alaska pollock); R2Tla-B, Tanakia lanceolata (bitterling); R2Raa, Rhodeus atremius atremius (Kyushu bitterling); R2Ac, Amia calva (bowfin); R2Ar, Acipenser ruthenus (sterlet); R2Em, Eublepharis macularius (leopard gecko); R2Ec, Elaphe climacophora (Japanese rat snake); R2Sqj, Squatina japonica (Japanese angel shark).

R2 distribution in non-avian vertebrates

Thanks to the recent progress of genome sequencing of Arthropoda and Chordata, we found many R2 families from these two phyla. We focused on the distribution of R2 in fishes because the widest diversity among vertebrates is seen in actinopterygian fishes, although the genome sequences of fish are not yet available from many orders. We used genomic DNA from 34 species in 17 orders of fish (1 order of Chondrichthyes and 16 orders of Actinopterygii), and found R2 from 19 species in 11 orders (Figs 1B and 2A and S1 Table) by PCR. We also characterized R2 in silico from three species of fish, stickleback (R2-1_GA), platy (R2-1_XM), and coelacanth (R2-1_LCh). We also characterized R2 from two species of Squamata reptiles (gecko and snake) by PCR. Among non-avian reptiles, turtles and crocodilians have R2 (Fig 1A), although we could not detect R2 from the green anole genome. Moreover, we could not detect R2 from amphibians, whether frogs or salamanders.

Fig 2. The phylogenetic distributions of R2 in vertebrates and birds.

(A) The R2 distribution in vertebrates focusing on actinopterygian fish. (B) R2 distribution in birds. Orders are shown with common names in parentheses. Asterisks indicate the presence of R2 in at least one species. Order names in blue indicate groups that we analyzed by PCR. Perciformes is not monophyletic and thus shown divided. Fish phylogeny is based on [3133][34,35] while avian phylogeny is based on [36].

R2 distribution in birds

Of interest are R2 families from birds. The first R2 family from birds was reported in zebrafinch [17]. Here we screened avian R2 systematically using recently sequenced bird genomes [37]. Initially we characterized several R2 families from birds using tBlastn searches with R2 protein sequences as queries. Based on the obtained avian R2 sequences, we realized that the 5′ end sequences of avian R2 are very conserved (Fig 3A). Then we took this conservation and used it to detect more R2 families by Blastn searches with the junction sequences including both 28S rRNA gene and R2 5′ UTR as queries. We detected more R2 sequences by this method. We were also able to characterize many additional R2 fragments using Blast search with the 3′ junction sequences as queries (Fig 3B).

Fig 3. Junction sequences of avian R2 elements.

The R2 family name or accession number (if the characterized sequence was short), flanking 28S rRNA gene sequence with terminal sequence of R2, common name of the origin, and classification are shown from left to right. The 28S ribosomal RNA gene sequences are in bold and shaded. (A) 5′ junctions. (B) 3′ junctions. Scientific names and common English names of the origins of R2 elements are as follows: R2-1_TG, Taeniopygia guttata (zebrafinch); R2-1_GFo, Geospiza fortis (medium ground finch); R2-1_FAl, Ficedula albicollis (collared flycatcher); R2-1_ZA, Zonotrichia albicollis (white-throated sparrow); R2-1_CBr, Corvus brachyrhynchos (American crow); R2-1_ZLM, Zosterops lateralis melanops (silvereye); R2-1_ACh, Acanthisitta chloris (rifleman); R2-1_SCa, Serinus canaria (Atlantic canary); R2-1_MVi, Manacus vitellinus (golden-collared manakin); R2-1_MUn and R2-2_MUn, Melopsittacus undulatus (budgerigar); R2-1_FCh, Falco cherrug (saker falcon); R2-1_FPe, Falco peregrinus (peregrine falcon); R2-1_PPu, Picoides pubescens (downy woodpecker); R2-1_CSt, Colius striatus (speckled mousebird); R2-1_TAl, Tyto alba (barn owl); R2-1_ACC, Aquila chrysaetos canadensis (golden eagle); R2-1_LDi, Leptosomus discolor (cuckoo roller); R2-1_HAl and R2-2_HAl, Haliaeetus albicilla (white-tailed eagle) and H. leucocephalus (bald eagle); R2-1_CAu, Cathartes aura (Turkey vulture); R2-1_NNi and R2-2_NNi, Nipponia nippon (Asian crested ibis); R2-1_PCar, Phalacrocorax carbo (great cormorant); R2-1_EGa, Egretta garzetta (little egret); R2-1_AFo, Aptenodytes forsteri (emperor penguin); R2-1_PAd, Pygoscelis adeliae (Adelie penguin); R2-1_GSt, Gavia stellata (red-throated loon); R2-1_PLe, Phaethon lepturus (white-tailed tropicbird); R2-1_CVo, Charadrius vociferus (killdeer); R2-1_CPu, Calidris pugnax (ruff); R2-1_BRG, Balearica regulorum gibbericeps (grey crowned crane); R2-1_OHo, Opisthocomus hoazin (hoatzin); R2-1_CCa, Caprimulgus carolinensis (nightjar); R2-2_CMa, Chlamydotis macqueenii (McQueen’s bustard); R2-1_CCan and R2-2_CCan, Cuculus canorus (common cuckoo); R2-1_MUni, Mesitornis unicolor (brown mesite); R2-1_PGu, Pterocles gutturalis (yellow-throated sandgrouse); R2-1_PRR, Phoenicopterus ruber ruber (American flamingo); R2-1_PCr, Podiceps cristatus (great crested grebe); R2-1_TGut, Tinamus guttatus (white-throated tinamou).

However, we obtained long protein-coding sequences of R2 only from medium ground finch (Geospiza fortis, R2-1_GFo), white-throated sparrow (Zonotrichia albicollis, R2-1_ZA), collared flycatcher (Ficedula albicollis, R2-1_FAl), and white-throated tinamou (Tinamus guttatus, R2-1_TGut). Some R2 sequences longer than 3.5 kb with both ends having disrupted open reading frames (ORFs) included R2 from Atlantic canary (Serinus canaria, R2-1_SCa), white-tailed eagle (Haliaeetus albicilla, R2-1_HAl), emperor penguin (Aptenodytes forsteri, R2-1_AFo), and Asian crested ibis (Nipponia nippon, R2-2_NNi). Most bird R2 copies are severely mutated, although R2 copies flanked by 28S rRNA gene fragments on at least one side are observed in various bird species (Fig 3). Their strong sequence similarity supports the common origin of R2 in birds. There is a tendency for R2 copies from more closely related species to show higher identity. For example, the 5′ 100-bp sequence of R2-1_TG from zebrafinch (Passeriformes) shows a 92% identity to that of R2-1_ZA from the white-throated sparrow (Passeriformes), 84% to that of R2-2_HAl from the white-tailed eagle (Falconiformes), and 79% to that of R2-1_TGut from the tinamou (Tinamiformes). However, this trend is not always clear; the 5′ 100-bp sequence of R2-1_SCa from the canary (Passeriformes) is only 76% identical to that of R2-1_TG. This may be partly due to the accumulated mutations since the inactivation of R2-1_SCa.

Finally, we found R2 copies from 25 orders spanning a wide range of bird lineages (Fig 2B). R2 is distributed among almost all of the major groups of birds, except Galloanseres (chickens and ducks). This is consistent with our previous PCR experiments that failed to detect R2 in two Galloanseres species (mallard and chicken) [9]. We found two fragments of R2 copies in budgerigar, which was one of the two other species for which we could not amplify R2 by PCR. One copy in AGAI01067519 (R2-1_MUn) is severely truncated and corresponds to the 5′ ~730-bp sequences of R2-1_AFo and R2-2_NNi (emperor penguin and Asian crested ibis). The other copy (R2-2_MUn) is an internally deleted copy, which shows similarity to the 4–517, 604–726, and 4370–4700 regions of R2-2_NNi. In both cases, the 5′ ends are flanked by a sequence with weak similarity to 28S rRNA genes. This could explain why we could not amplify R2 fragments by PCR in budgerigar.

Phylogenetic analysis and distribution of R2

We generated two phylogenetic trees of R2 families. One is based on the alignment of the partial RT domain corresponding to motifs 5 to 9 (S2 Fig). The other is based on the alignment of motif 5 to the C-terminus (Fig 4). We note that we could not recover intact ORFs for not a few R2 families due to incomplete sequencing, mutations or truncation, and thus, we could not use them in the phylogenetic analysis. The two trees showed a similar topology. We tried to determine the root using outgroup sequences, but failed to obtain a consistent result. In addition to the clusters of R2 families that can be assigned to reported subclades (vertical bars with names in Fig 4), two new clusters were supported with high statistical significance (vertical bars without names in Fig 4), consistent with the original articles on R2 from ticks, and tadpole shrimps [38,39]. These two clusters can be equivalent to subclades, although their distributions might be narrow. The subclades that include R2 families from more than three phyla are R2A1 (Chordata, Arthropoda, Nematoda), R2A3 (Chordata, Arthropoda, Platyhelminthes, Mollusca), and R2D4 (Cnidaria, Chordata, Echinodermata).

Fig 4. A phylogenetic tree of R2 families based on the protein alignment from motif 5 of the RT domain to the C-terminus.

Bootstrap values above 50% are shown at branches. R2 family names and their origins are shown as leaves. R2 families from Chordata are colored in red, those from Arthropods in blue, those from Platyhelminthes in green and those from other animals in black. Clusters of R2 families that can be assigned to reported subclades are indicated by vertical lines with names whilst clusters not assigned to reported subclades are indicated by vertical lines without names. The magenta bar indicates the cluster having a 28S rRNA gene-like sequence in the 3′ UTR.

Although R2 from Chordata is seen in six lineages, R2 from jawed vertebrates are seen only in two lineages, R2A3 and R2C. R2Tla-B and R2Ec belong to the R2C clade. All other retrotransposons belonging to the R2C clade are from trematodes (Platyhelminthes): R2Sm-A and R2Sm-B from the bloodfluke Schistosoma mansoni and R2NS-1_CSi from the liver fluke Clonorchis sinensis. To exclude the possibility of contamination, we sequenced the 3′ downstream 80-bp sequences from the R2 insertion sites. The sequences downstream of R2Tla-B and R2Ec are identical to the 28S rRNA genes from vertebrates and not identical to those from Schistosoma (S3 Fig). Trematodes are parasites infecting mostly vertebrates. Horizontal transfer of R2C from trematodes to vertebrates cannot be excluded, but the long-term maintenance of R2C since the split of trematodes and vertebrates (protostomes and deuterostomes) is another possibility.

We found only one family of R2 in each vertebrate species we analyzed with the sole exception of alfonsino (Beryx splendens), which contains two closely related families (R2Bs-A and R2Bs-B). However, we may have missed R2 subfamilies or lineages that are difficult to be amplified by PCR with our primer sets. However, the relationships among R2 families from two medaka species (Oryzias melastigma, R2Om and Oryzias latipes, R2Ol-A) and from two stickleback species (Pungitius pungitius pungitius, R2Pp and Gasterosteus aculeatus, R2-1_GA) indicate the maintenance of two R2 lineages in their ancestral species. Another explanation for these relationships may be the occurrence of horizontal transfer. It is likely that more than one lineage of R2 has been maintained in vertebrates, as reported in arthropods [40,41].

Newly identified R2 families from insects were clustered with reported insect R2 families R2A2, R2B, R2D2, and R2D5. In arthropods, it is more obvious than in vertebrates that several lineages of R2 have been maintained in a single species. The leafcutter bee Megachile rotundata is the outlier case in that all four lineages that were found in insects (R2A2, R2B, R2D2, and R2D5) were observed.

The phylogenetic position of R2-1_MLe from the sea walnut is of interest since Ctenophore is one basal lineage of animals. Although we could not determine its phylogenetic position with high statistical confidence, its capacity of encoding a protein with a single CCHH zinc finger at the N-terminus indicates that R2-1_MLe is inside the R2D clade.

Non-target specific R2 families

We found two families of R2 that are not specifically inserted into 28S ribosomal RNA genes. One is R2NS-1_SMed from the Mediterranean planaria Schmidtea mediterranea, belonging to the R2D clade, and the other is R2NS-1_CSi from the liver fluke Clonorchis sinensis, belonging to the R2C clade. We confirmed their non-specific integration by analyzing their flanking sequences (S1 Fig). Some copies are >98% identical to their respective consensuses, eliminating the possibility that recombination contributed to their apparent lack of sequence specificity. Thus the loss of target specificity has occurred independently in two different families from the R2 lineage.

We could not determine the boundaries of R2NS-1_PMi from the bat star Patiria miniata, as we did not find rRNA genes in the scaffold sequence (AKZP01104910). At present we have no evidence for the target specificity of R2NS-1_PMi, but it is possible that other copies are inserted into 28S rRNA genes.

Apparent movement of insertion sites of R2 families.

Seven R2 families show a similarity to the 28S rRNA gene sequences in their 3′ UTR (Fig 5A). The identity of this region is not as high as that downstream of the canonical insertion site of R2. This indicates that these 28S-like sequences are parts of the 3′ UTR of R2. Among them, five (R2Gs, R2Pt, R2Bs-A, R2Bs-B, and R2Sb) are phylogenetically closely related (Fig 4, magenta bar). The lineage that includes these five families also includes one more family that was identified from the sequenced genome, R2-1_XM from the platyfish Xiphophorus maculatus. We also found a fragment sequence of R2 from the Amazon molly Poecilia formosa (AYCK01024837). Unfortunately we could not characterize their 3′ termini since all of these sequences were either fragmented or unsequenced, and not flanked with 28S rRNA genes. R2Tch does not have a sequence similar to 28S rRNA genes in its 3′ terminus, unlike its sister lineage.

Fig 5. Apparent movement of 5′ insertion sites of R2.

(A) Alignment of 5′ and 3′ junction sequences. Nucleotides of 28S rRNA genes are in bold and shaded while nucleotides identical to 28S rRNA genes inside of R2 elements are only shaded. Patterned boxes above the sequences correspond to the regions with the same patterned boxes in (B). R2Gs, R2Pt, R2Bs-B, R2Bs-A, and R2Sb are phylogenetically closely related and it is likely that their 3′ 28S-like sequences were acquired in their common ancestor. (B) Schematic diagram around the insertion sites of canonical R2 and R2Gs. Patterned boxes indicate the regions of 28S rRNA genes. Scientific names and common English names of the origins of R2 elements are as follows: R2Gs, Grammistes sexlineatus (goldenstriped soapfish); R2Pt, Parapristipoma trilineatum (chicken grunt); R2Bs-A and R2Bs-B, Beryx splendens (splendid alfonsino); R2Sb, Stethojulis bandanesis (red shoulder wrasse); R2Kp, Konosirus punctatus (dotted gizzard shad); R2Op, Oxyurichthys papuensis (arrowfin gobies).

We were able to characterize the 5′ junctions of three of these R2 families (R2Gs, R2Bs-A, and R2Bs-B). Their 5′ junctions are upstream from the canonical insertion site of R2—around 110 bp away in the case of R2Gs, and around 130 bp away in the cases of R2Bs-A and R2Bs-B (Fig 5A). This indicates that upon the insertion of these three families of R2, around 110–130 bps of 28S rRNA gene sequence is replaced by R2 (Fig 5B).


The distribution of R2

Nine animal phyla (Arthropoda, Chordata, Echinodermata, Hemichordata, Nematoda, Platyhelminthes, Mollusca, Cnidaria, Ctenophora) have maintained at least one of the R2 lineages. Cnidaria and Ctenophora are basal lineages of animals. Four major groups of bilaterian animals, Deuterostoma (Chordata, Echinodermata, and Hemichordata), Ecdysozoa (Arthropoda and Nematoda), Platyzoa (Platyhelminthes) and Lophotrochozoa (Mollusca), all include R2-harboring species. R9, which is a derivative of R2, was reported from Rotifera in Platyzoa [22]. We found an R2-like sequence in Priapulida in Ecdysozoa. It is now clear that R2 is very widely distributed in animals. However, considering sublineages (subclades) of R2, the distribution is apparently patchy. Most of the R2 subclades were reported only from one phylum. We consider that it is partially due to the sampling bias for Arthropoda and Chordata. It should also be mentioned that we may have failed to amplify R2 sequences due to the sequence differences from the primer sets we used and there is a possibility that each R2 lineage may be more widely distributed.

The R2 families from basal groups of animals (Cnidaria and Ctenophore) appear not to be the basal lineages of R2. R2Nvec-A belongs to the R2D4 subclade. R8Hm-A and R8Hm-B are positioned inside of the R2A clade. R2-1_MLe is probably a distinct lineage inside of the R2D clade. Their phylogenetic positions indicate that the origin of R2 predates the birth of metazoa. At present, there is no reason to introduce horizontal transfer to explain the distribution and phylogeny of R2, but there is a possibility that ancient horizontal transfer between different metazoan lineages complicated the R2 distribution further. It is noteworthy that even if horizontal transfer has occurred, the very ancient origin of R2 is evident.

In vertebrates, the R2A3 subclade is the dominant lineage, even though some other lineages (R2C, R2D3 and R2-1_PM) are also present in some species. The phylogenetic relationships in R2A3 are not always consistent with the host vertebrate phylogeny, and the finding of more than two families of R2 in some fish species or genera indicates that multiple lineages of R2A3 have been maintained in some groups of vertebrates. Considering the presence of multiple lineages of R2 in insects, this is quite likely.

From teleost fishes, we recovered R2 sequences mainly belonging to the R2A3 subclade, flanked with 28S rRNA genes. R2 families characterized from birds in this study also belong to the R2A3 subclade. We could find R2 sequences from most of the bird genomes we analyzed. However, many R2 sequences were disrupted or not yet completely sequenced. Only some R2 copies are intact and encode a full-length protein. Some R2 copies seem full-length but contain disrupted protein-coding regions. It is likely that many R2 sequences were remnants of anciently active R2 elements. It can explain why many avian R2 sequences are flanked with fragmented 28S rRNA gene sequences or non-rRNA gene sequences.

The phylogeny of R2 proteins from birds is consistent with the host phylogeny. The closest relatives of bird R2 sequences are R2 families from crocodilians (R2-1_Gav from gharial in Fig 4). This suggests that only a single lineage of R2 is present in birds.

The conservation of the 5′ end sequences of avian R2 appears extraordinary. We could not quantify their conservation in terms of identity, but the 5′ ends of bird R2 appear to be more conserved than those of Drosophila R2. It may indicate the importance of the 5′ UTR sequences in R2. It is reported that some R2 families contain an HDV-like ribozyme in their 5′ UTR which contributes to the 5′ processing and translation initiation [42,43]. In R2Dm from Drosophila melanogaster, the most conserved sequence, CCUCCUCGUGG, is positioned at nucleotides 135–145, while the identical sequence is positioned at nucleotides 37–47 in R2-1_TG from zebrafinch. Thus, natural selection for retention of ribozyme function likely contributes to the higher conservation in the 5′ 100-bp sequence of avian R2 than in Drosophila R2.

Change of target specificity of R2

To date, only two lineages in the R2 superclade, R8 and R9, have been reported to have different target specificity from canonical R2 families. Here, we report two clearly non-target-specific R2 families: R2NS-1_SMed and R2NS-1_CSi. Their phylogenetic positions are distinct, indicating independent loss of target specificity. Since all other clades of non-LTR retrotransposons that show target specificity also include some non-target-specific families, it shows that R2 is not an exception in this facet. The conserved target specificity of R2 is still exceptional. One of the reasons for this maintenance of target specificity is certainly the suitability of its target, rRNA genes, which are highly conserved sequences with high copy numbers. It is unclear why two R2NS families (R2NS-1_SMed and R2NS-1_CSi) lost their target specificity. The genome of the Mediterranean planaria S. mediterranea also contains two R2 families showing canonical target specificity (R2-1_SM, and R2-2_SMed). It is also noteworthy that the genome of S. mediterranea contains several families of target-specific non-LTR retrotransposons, belonging to the NeSL clade [10]. The targets of three of these families (LIN9_SM, LIN24_SM, and LIN26_SM) are 28S rRNA genes and thus, the loss of target specificity of R2NS-1_SMed may be caused by the competition of target 28S rRNA genes with other target-specific non-LTR retrotransposon families. No canonical R2 family has been characterized from the genome of the liver fluke C. sinensis.

One vertebrate R2 lineage includes families having 28S rRNA gene-derived sequences in their 3′ UTR (Fig 4, magenta bar). This seems related to the apparent movement of 5′ boundaries of R2 insertions upstream compared to those in canonical R2 families (Fig 5). A similar case is observed in R2Hm-B [16]. Based on the 3′ junction sequences, R2Hm-B appeared to be inserted 15 bp upstream relative to R2Hm-A. One interpretation for this is that R2Hm-B has a 15-bp sequence identical to 28S rRNA in its 3′ terminus. In the case of families of fish R2, the 3’ UTR sequences are distinct from 28S rRNA genes and thus we can exclude the possibility of the movement of 3′ junctions, which likely correspond to the bottom strand cleavage site. Rather, the situation indicates that a region over 100 bp long of the 28S rRNA gene is replaced by R2 upon integration (Fig 5B). This corresponds to the movement of the top strand cleavage site. Since the top strand cleavage site, unlike the bottom strand cleavage site, is not strictly determined, this is a likely explanation, though further investigation is necessary. The similarity of the 3′ UTR with the 28S rRNA gene may contribute to stabilizing the transposition intermediate through the binding between R2 mRNA and 28S rRNA genes, like other target-specific non-LTR retrotransposons [44,45].

Supporting Information

S1 Fig. Flanking sequences of non-target-specific R2 families.

The top 30 hits with 3′ termini in the Censor search are shown. R2 is inserted at “|”. The positions of R2 copies are shown in parentheses. 28S rRNA sequences are in bold.


S2 Fig. A phylogenetic tree of R2 families based on the protein alignment from motif 5 to 9 of the RT domain.

Bootstrap values above 50% are shown at branches. R2 family names and their origins are shown at leaves. R2 families from Chordata are colored in red, those from Arthropods in blue, those from Platyhelminthes in green and those from other animals in black. Clusters of R2 families that can be assigned to reported subclades are indicated by vertical lines with names and clusters not assigned to reported subclades are indicated by vertical lines but without names.


S3 Fig. The downstream 100 bp of R2Tla-B and R2Ec.

Nucleotides identical to the 28S rRNA genes from humans are shown by dots (.).



We thank Jun Tamefusa for the help of sequencing R2Ol-A, and Karolina Walichiewicz for the critical reading of the manuscript.

Author Contributions

  1. Conceptualization: KKK HF.
  2. Data curation: KKK YS.
  3. Funding acquisition: HF.
  4. Investigation: KKK YS.
  5. Project administration: KKK HF.
  6. Resources: KKK YS.
  7. Supervision: KKK HF.
  8. Visualization: KKK.
  9. Writing – original draft: KKK.
  10. Writing – review & editing: KKK YS HF.


  1. 1. Bao W, Kojima KK, Kohany O (2015) Repbase Update, a database of repetitive elements in eukaryotic genomes. Mob DNA 6: 11. pmid:26045719
  2. 2. Babushok DV, Kazazian HH Jr. (2007) Progress in understanding the biology of the human mutagen LINE-1. Hum Mutat 28: 527–539. pmid:17309057
  3. 3. Kojima KK, Jurka J (2013) A superfamily of DNA transposons targeting multicopy small RNA genes. PLoS One 8: e68260. pmid:23874566
  4. 4. Eagle SH, Crease TJ (2016) Distribution of the DNA transposon family, Pokey in the Daphnia pulex species complex. Mob DNA 7: 11. pmid:27330569
  5. 5. Burke WD, Calalang CC, Eickbush TH (1987) The site-specific ribosomal insertion element type II of Bombyx mori (R2Bm) contains the coding sequence for a reverse transcriptase-like enzyme. Mol Cell Biol 7: 2221–2230. pmid:2439905
  6. 6. Xiong Y, Eickbush TH (1993) Dong, a non-long terminal repeat (non-LTR) retrotransposable element from Bombyx mori. Nucleic Acids Res 21: 1318. pmid:8385316
  7. 7. Okazaki S, Ishikawa H, Fujiwara H (1995) Structural analysis of TRAS1, a novel family of telomeric repeat-associated retrotransposons in the silkworm, Bombyx mori. Mol Cell Biol 15: 4545–4552. pmid:7623845
  8. 8. Kojima KK, Fujiwara H (2003) Evolution of target specificity in R1 clade non-LTR retrotransposons. Mol Biol Evol 20: 351–361. pmid:12644555
  9. 9. Kojima KK, Fujiwara H (2004) Cross-genome screening of novel sequence-specific non-LTR retrotransposons: various multicopy RNA genes and microsatellites are selected as targets. Mol Biol Evol 21: 207–217. pmid:12949131
  10. 10. Kojima KK, Jurka J (2015) Ancient Origin of the U2 Small Nuclear RNA Gene-Targeting Non-LTR Retrotransposons Utopia. PLoS One 10: e0140084. pmid:26556480
  11. 11. Roiha H, Glover DM (1981) Duplicated rDNA sequences of variable lengths flanking the short type I insertions in the rDNA of Drosophila melanogaster. Nucleic Acids Res 9: 5521–5532. pmid:6273796
  12. 12. Fujiwara H, Ogura T, Takada N, Miyajima N, Ishikawa H, Maekawa H (1984) Introns and their flanking sequences of Bombyx mori rDNA. Nucleic Acids Res 12: 6861–6869. pmid:6091041
  13. 13. Jakubczak JL, Xiong Y, Eickbush TH (1990) Type I (R1) and type II (R2) ribosomal DNA insertions of Drosophila melanogaster are retrotransposable elements closely related to those of Bombyx mori. J Mol Biol 212: 37–52. pmid:1690812
  14. 14. Burke WD, Malik HS, Lathe WC 3rd, Eickbush TH (1998) Are retrotransposons long-term hitchhikers? Nature 392: 141–142. pmid:9515960
  15. 15. Kojima KK, Fujiwara H (2005) Long-term inheritance of the 28S rDNA-specific retrotransposon R2. Mol Biol Evol 22: 2157–2165. pmid:16014872
  16. 16. Kojima KK, Kuma K, Toh H, Fujiwara H (2006) Identification of rDNA-specific non-LTR retrotransposons in Cnidaria. Mol Biol Evol 23: 1984–1993. pmid:16870681
  17. 17. Kapitonov VV, Jurka J (2009) R2 non-LTR retrotransposons in the bird genome. Repbase Reports 9: 1329.
  18. 18. Luchetti A, Mantovani B (2013) Non-LTR R2 element evolutionary patterns: phylogenetic incongruences, rapid radiation and the maintenance of multiple lineages. PLoS One 8: e57076. pmid:23451148
  19. 19. Christensen SM, Bibillo A, Eickbush TH (2005) Role of the Bombyx mori R2 element N-terminal domain in the target-primed reverse transcription (TPRT) reaction. Nucleic Acids Res 33: 6461–6468. pmid:16284201
  20. 20. Thompson BK, Christensen SM (2011) Independently derived targeting of 28S rDNA by A- and D-clade R2 retrotransposons: Plasticity of integration mechanism. Mob Genet Elements 1: 29–37. pmid:22016843
  21. 21. Xiong YE, Eickbush TH (1988) Functional expression of a sequence-specific endonuclease encoded by the retrotransposon R2Bm. Cell 55: 235–246. pmid:2844414
  22. 22. Gladyshev EA, Arkhipova IR (2009) Rotifer rDNA-specific R9 retrotransposable elements generate an exceptionally long target site duplication upon insertion. Gene 448: 145–150. pmid:19744548
  23. 23. Kapitonov VV, Jurka J R2 retrotransposons in the lamprey genome. Repbase Reports 9: 1168.
  24. 24. Kojima KK, Jurka J (2013) Non-LTR retrotransposons from the coelacanth genome. Repbase Reports 13: 3184.
  25. 25. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402. pmid:9254694
  26. 26. Kohany O, Gentles AJ, Hankus L, Jurka J (2006) Annotation, submission and screening of repetitive elements in Repbase: RepbaseSubmitter and Censor. BMC Bioinformatics 7: 474. pmid:17064419
  27. 27. Kapitonov VV, Tempel S, Jurka J (2009) Simple and fast classification of non-LTR retrotransposons based on phylogeny of their RT domain protein sequences. Gene 448: 207–213. pmid:19651192
  28. 28. Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol 30: 772–780. pmid:23329690
  29. 29. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O (2010) New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: 307–321. pmid:20525638
  30. 30. Kojima KK, Jurka J (2013) R2 non-LTR retrotransposons not inserted into ribosomal RNA genes. Repbase Reports 13: 1215–1217.
  31. 31. Inoue JG, Miya M, Tsukamoto K, Nishida M (2003) Basal actinopterygian relationships: a mitogenomic perspective on the phylogeny of the “ancient fish”. Mol Phylogenet Evol 26: 110–120. pmid:12470943
  32. 32. Mabuchi K, Miya M, Azuma Y, Nishida M (2007) Independent evolution of the specialized pharyngeal jaw apparatus in cichlid and labrid fishes. BMC Evol Biol 7: 10. pmid:17263894
  33. 33. Yamanoue Y, Miya M, Inoue JG, Matsuura K, Nishida M (2006) The mitochondrial genome of spotted green pufferfish Tetraodon nigroviridis (Teleostei: Tetraodontiformes) and divergence time estimation among model organisms in fishes. Genes Genet Syst 81: 29–39. pmid:16607039
  34. 34. Miya M, Satoh TP, N M. (2005) The phylogenetic position of toadfishes (order Batrachoidiformes) in the higher ray-finned fish as inferred from partitioned Bayesian analysis of 102 whole mitochondrial genome sequences. Biological Journal of the Linnean Society 85: 289–306.
  35. 35. Smith WL, Craing MT (2007) Casting the Percomorph Net Widely: The Importance of Broad Taxonomic Sampling in the Search for the Placement of Serranid and Percid Fishes. Copia 1: 35–55.
  36. 36. Jarvis ED, Mirarab S, Aberer AJ, Li B, Houde P, Li C et al. (2014) Whole-genome analyses resolve early branches in the tree of life of modern birds. Science 346: 1320–1331. pmid:25504713
  37. 37. Zhang G, Li C, Li Q, Li B, Larkin DM, Lee C et al. (2014) Comparative genomics reveals insights into avian genome evolution and adaptation. Science 346: 1311–1320. pmid:25504712
  38. 38. Bunikis J, Barbour AG (2005) Ticks have R2 retrotransposons but not the consensus transposon target site of other arthropods. Insect Mol Biol 14: 465–474. pmid:16164602
  39. 39. Luchetti A, Mingazzini V, Mantovani B (2012) 28S junctions and chimeric elements of the rDNA targeting non-LTR retrotransposon R2 in crustacean living fossils (Branchiopoda, Notostraca). Genomics 100: 51–56. pmid:22564473
  40. 40. Jakubczak JL, Burke WD, Eickbush TH (1991) Retrotransposable elements R1 and R2 interrupt the rRNA genes of most insects. Proc Natl Acad Sci U S A 88: 3295–3299. pmid:1849649
  41. 41. Stage DE, Eickbush TH (2010) Maintenance of multiple lineages of R1 and R2 retrotransposable elements in the ribosomal RNA gene loci of Nasonia. Insect Mol Biol 19 Suppl 1: 37–48. pmid:20167016
  42. 42. Ruminski DJ, Webb CH, Riccitelli NJ, Luptak A (2011) Processing and translation initiation of non-long terminal repeat retrotransposons by hepatitis delta virus (HDV)-like self-cleaving ribozymes. J Biol Chem 286: 41286–41295. pmid:21994949
  43. 43. Eickbush DG, Eickbush TH (2010) R2 retrotransposons encode a self-cleaving ribozyme for processing from an rRNA cotranscript. Mol Cell Biol 30: 3142–3150. pmid:20421411
  44. 44. Osanai M, Takahashi H, Kojima KK, Hamada M, Fujiwara H (2004) Essential motifs in the 3’ untranslated region required for retrotransposition and the precise start of reverse transcription in non-long-terminal-repeat retrotransposon SART1. Mol Cell Biol 24: 7902–7913. pmid:15340053
  45. 45. Anzai T, Osanai M, Hamada M, Fujiwara H (2005) Functional roles of 3’-terminal structures of template RNA during in vivo retrotransposition of non-LTR retrotransposon, R1Bm. Nucleic Acids Res 33: 1993–2002. pmid:15814816