Mol Genet Genomics (2008) 279:63–73 DOI 10.1007/s00438-007-0295-0

ORIGINAL PAPER

GalEa retrotransposons from galatheid squat lobsters (Decapoda, Anomura) deWne a new clade of Ty1/copia-like elements restricted to aquatic species Yves Terrat · Eric Bonnivard · Dominique Higuet

Received: 25 June 2007 / Accepted: 19 September 2007 / Published online: 10 October 2007 © Springer-Verlag 2007

Abstract Crustacean species have not been examined in great detail for their transposable elements content. Here we focus on galatheid crabs, which are one of the most diverse and widespread taxonomic groups of Decapoda. Ty1/copia retrotransposons are a diverse and taxonomically dispersed group. Using degenerate primers, we isolated several DNA fragments that show homology with Ty1/ copia retroelements reverse transcriptase gene. We named the corresponding elements from which they originated GalEa1 to GalEa3 and analyzed one of them further by isolating various clones containing segments of GalEa1. This is the Wrst LTR retrotransposon described in crustacean genome. Nucleotide sequencing of the clones revealed that GalEa1 has LTRs (124 bp) and that the internal sequence (4,421 bp) includes a single large ORF containing gag and pol regions. Further screening identiWed highly related elements in six of the nine galatheid species studied. By performing BLAST searches on genome databases, we could also identify GalEa-like elements in some Wshes and Urochordata genomes. These elements deWne a new clade of Ty1/copia retrotransposons that diVers from all other Ty1/ copia elements and that seems to be restricted to aquatic species.

Communicated by M.-A. Grandbastien. Electronic supplementary material The online version of this article (doi:10.1007/s00438-007-0295-0) contains supplementary material, which is available to authorized users. Y. Terrat · E. Bonnivard · D. Higuet (&) UMR 7138 Systématique Adaptation Evolution, Equipe Génétique et Evolution, Université Pierre and Marie Curie Paris 6, Case 5, Bât A, porte 427, 7 quai St-Bernard, 75252 Paris Cedex 05, France e-mail: [email protected]

Keywords Retrotransposon · Ty1/copia · Decapoda · Galatheid crabs

Introduction Transposable elements (TEs) constitute a large fraction of most eukaryotic genomes investigated to date (44% of the human genome or 20% for Drosophila melanogaster, Li et al. 2001; Misra et al. 2002). TEs have the ability to insert into new sites within a genome, leading to a large spectrum of genetic variations ranging from simple insertions to dramatic alterations in chromosomal structure (FedoroV 2000; Kazazian Jr 2004; Kidwell and Lisch 1997). Due to their high capacity to induce mutations, TEs are considered as major actors in genome evolution (Jordan et al. 2003; Van de Lagemaat et al. 2003). Based on their mode of transposition, TE families are classiWed into three major groups (Poulter and Goodwin 2005): (1) transposons (or class II elements; Finnegan 1989) have a DNA intermediate; (2) retrotransposons, and (3) retroposons have a RNA intermediate (class I elements), with only retrotransposons presenting a free DNA replication intermediate. Based on their structural features and their phylogenetic relationships, Wve major lineages of retrotransposons have been identiWed to date: Ty1/copia, Ty3/gypsy, BEL/Pao, Tyrosine recombinase elements, and vertebrate retroviruses (Malik and Eickbush 2001). These distinct lineages were notably revealed by the phylogeny of their most conserved domains, the reverse transcriptase (RT) and the RNase H (RH) (Jurka 1998; McCLure 1991; Xiong and Eickbush 1990). Besides Tc1/mariner transposons than have been largely studied (Robertson 1997), most of previous studies on TEs focused on model organisms. This bias in the species sampling could potentially aVect our knowledge of the

123

64

dynamics and evolution of the TEs themselves. This is particularly striking for some large phyllum such as Crustacea. To date, only a few TEs have been described in crustacean species: retroposon R2 in an isopod and in horseshoe crab (Burke et al. 1999), transposon Pokey from the piggyBac superfamily in a waterXea (Penton et al. 2002) and marinerlike transposons (Casse et al. 2000, 2006; Kvamme et al. 2005; Bui et al. 2007), jockey-like and CR1-like retroposons (Halaimia-Toumi et al. 2004) in hydrothermal and coastal crabs. No LTR retrotransposon has been described to date. These examples are not numerous enough considering the great diversity of crustacean species. For this reason, we focused our eVorts on retrieving TEs from the genomes of Galatheidae. The Galatheidae family is one of the most diverse families of Decapoda and is present worldwide in all marine ecosystems (Baba 1988), from coastal habitats to deep-sea hydrothermal vents. The phylogenetic relationships between the diVerent galatheid species have been well investigated (Machordom and Macpherson 2004). Moreover, some species show large genomes, far above the mean C-value of Crustacea (3.0 pg with a standard error of 3.85) and of Decapoda (4.91 pg with a standard error of 4.99) calculated from Animal Genome Size Database release 2 (Gregory 2007). As C-values are positively related to TE content (Kidwell 2002; Petrov et al. 2000), large genomes usually harbor a high TEs content. In this study, we focused our attention on nine galatheid species representing Wve diVerent genera. These species live either in the Western PaciWc Norfolk seamounts, in the South Atlantic Falkland Islands, close to PaciWc deep sea hydrothermal vents or along the coasts of Atlantic French Brittany. In Eumunida annulosa, the generation of several polymerase chain reaction (PCR) fragments allowed us to reconstruct a complete Ty1/copia-like retrotransposon that we named GalEa1. Using this sequence, we then retrieve several highly related elements in six other galatheid species. Here, we report the structure and sequence analysis of the new LTR retrotransposon GalEa1, the Wrst described in Crustacea. This element and the related GalEa-like elements identiWed in genome databases deWne a new clade of Ty1/copia retrotransposons that corresponds to a sister group of all other Ty1/copia elements previously described.

Materials and methods Biological materials Four specimens of Galatheidae (Agononida laurentae, Munida acantha, Munida thoe, and Munida zebra) and two specimens of Chirostylidae, the sister family of Galatheidae (E. annulosa, Eumunida sternomaculata) were collected south of New-Caledonia on Norfolk seamounts during the

123

Mol Genet Genomics (2008) 279:63–73

prospecting campaigns Norfolk 1 (2001, IRD Nouméa) and Norfolk 2 (2003, MUSORSTOM). In addition, the specimen of Munida gregaria (Galatheidae) was collected on the coasts of Falkland islands during the ICEFISH cruise (2004), the specimen of Munidopsis recta (Galatheidae) on the south of East PaciWc Ride during the BIOSPEEDO cruise (2004) and those of Galathea squamifera (Galatheidae) in French Brittany (2006). All specimens were kept at 4°C in 70% ethanol except for M. recta, which was conserved at ¡80°C. DNA from one individual per species was isolated using the CTAB method (Ishaq et al. 1990). Dry DNA pellets were suspended in TE BuVer (10 mM Tris and 0.1 mM EDTA). Characterization of 16S sequences Partial 16S rRNA (16S) sequences of E. sternomaculata, M. acantha, M. thoe, M. zebra, and A. laurentae (Machordom and Macpherson 2004) and 16S sequence of G. squamifera (Hiller et al. 2006) were retrieved from GenBank database (http://www.ncbi.nlm.nih.gov/Genbank, see Fig. 1 for accessions). For the three remaining species (M. gregaria, M. recta, and E. annulosa), we ampliWed 16S partial sequences by polymerase chain reaction (PCR) using primers 16Sar-L and 16Sbr-H described by Palumbi and Benzie (1991). About 50 ng of DNA was used in 25
l of total reaction volume containing 2.5 U of Taq DNA polymerase (InVitrogen, Carlsbad, CA, USA) and 10 pmol of each primer. AmpliWcations were performed in a Thermo Px2 Thermal Cycler for 40 cycles. Cycling conditions were 94°C for 45 s, 40°C for 45 s, and 72°C for 1 min with a Wnal extension step of 72°C for 10 min. The ampliWed fragments (around 500 bp) were puriWed by ethanol precipitation prior to sequencing (Genome Express). Partial 16S sequences were deposited in GenBank under accessions EF428962 (E. annulosa), EF428963 (M. gregaria), and EF428964 (M. recta). AmpliWcation of RT fragments using degenerate primers We searched for Ty1/copia-like RT fragments on a selected subset of three species (A. laurentae, E. annulosa, and E. sternomaculata). PCRs were carried out using two degenerate primers. The 5⬘ primer was designed by ourselves and corresponds to the conserved ‘KARLVA’ motif speciWc to Ty1/copia retrotransposons (5⬘-A(A/G)(A/G)GCN(A/ C)GN(T/C)TNGTNGC-3⬘). The 3⬘ primer was described by Flavell et al. (1992) and corresponds to the conserved ‘YV/LDD’ motif (5⬘-ANNAN(A/G)TC(A/G)TCNAC(A/ G)TA-3⬘). PCR ampliWcations using degenerate primers were performed for 35 cycles (94°C for 45 s, 54°C for 1 min 30 s, and 72°C for 1 min) on about 50 ng of each

Mol Genet Genomics (2008) 279:63–73

65

100 / 100 1

94 / 98 1

33 / 46

Munida subrugosa (AY050075) Munida thoe (AY351182) Munida zebra (AY351191)

83 / 96 1

0.6

Munida gregaria (EF428963)

Munida stia (AY351170)

Galatheidae

Munida notata (AY351147) 58 / 72

Munida acantha (AY351097)

1

Agononida laurentae (AY351070)

98 / 99

56 / 67

Paramunida belone (AY351199)

1

0.99

Galathea squamifera (DQ444860)

52 / 69

Munidopsis recta (EF428964)

1

Petrolisthes galathinus (DQ865303)

100/100 1 100/100 1

Petrolisthes armatus (DQ865310)

Eumunida annulosa (EF428962) Eumunida sternomaculata (AY351260)

Porcellanidae Chirostylidae

Cancer pagurus (DQ079708)

94 / 92

Carcinus maenas (AY919095)

1

0.05

Fig. 1 Phylogenetic hypothesis based on molecular data using subunit 16S rRNA partial sequences represented by a tree obtained by Bayesian inference with Brachyura species (Cancer pagurus and Carcinus maenas) used as outgroup. The best-Wt model is GTR (General Time Reversible) +
+ I. Numbers above nodes indicate bootstrap values for

parsimony and Maximum likelihood methods, respectively, separated by a slash character. Posterior probabilities for Bayesian inference are indicated above the nodes. Scale bar indicates percent sequence divergence. Accessions are given in brackets and names of species analyzed in the present study are shown in bold

DNA sample, with 2.5 U of Taq DNA polymerase (InVitrogen) and 50 pmol of each degenerate primer in a Wnal volume of 25
l. PCR products were electrophoretically segregated in 0.8% agar gel and ampliWcations of expected size (380 bp) were cut oV, puriWed with SpinX column (Costar) and ligated into pGEM-Teasy vector (Promega, Madison, WI, USA) for sequencing (Genome Express). Partial RT DNA sequences were deposited in GenBank under accessions DQ912989 to DQ912998.

we characterized the exact boundaries of the LTRs using PCR walking. For each walking step, only one clone was sequenced. For each sequence obtained, we checked for similarities with retrotransposons in the NCBI database using BLASTX. Then, all these fragments were joined together to reconstruct a complete element using the Lfasta software. Minimum overlap between two sequences was 50 bp, with a minimum DNA identity of 95%. To verify the Wnal contig, we designed three primers sets covering the entire element from 5⬘LTR-end to middle of IN (DQ913005), from middle of IN to RT-end (DQ912985) and from RT-end to 3⬘LTR-start (DQ913006). Complete sequences of LTRs and internal region of GalEa1_Ea are available in GenBank under accession (EU097705). Sequence analysis of the GalEa1 element was then performed, and conserved domains were detected using a combination of heuristics alignments (i.e., BLAST) with proper multiple alignment (i.e., ClustalX).

Characterization of E. annulosa GalEa1 retrotransposon To characterize all parts of a full-length element in E. annulosa genome, several walking methods were used. First, we used inverse PCR (Ochman et al. 1990) to retrieve the sequences located at the edge of the RT fragment generated by degenerate primers PCR. This extended the initial fragment to the integrase (IN) domain (for the 5⬘ edge) and to the RH domain (for the 3⬘ edge). We then retrieved the remaining 3⬘ sequence by PCR walking (as described in Devic et al. 1997). As this open reading frame was interrupted 200 bp before the end of this sequence, we suspected we had reached the 3⬘LTR. Considering that both LTRs of retrotransposons are supposed to be almost identical, we designed a pair of speciWc primers: one in the suspected 3⬘LTR and one in the integrase domain that we had obtained by inverse PCR. These primers were used to amplify the missing 5⬘ part of the element. In a Wnal step,

GalEa1 elements in galatheids To investigate the presence of GalEa1 retrotransposons in the eight other species of galatheid squat lobsters, a pair of primers was designed from the E. annulosa sequence, GalEa1_1951+ (5⬘-AGATACTTACGGACAATGGG-3⬘) and GalEa1_3516- (5⬘-CACAAATCACACCCTCAA-3⬘). The expected PCR product is 1,565 bp long and corresponds to a part of IN and RT domains. PCR reactions were

123

66

run with 10 pmol of each speciWc primer in a Wnal volume of 25
l using 2.5 U of Taq DNA polymerase (InVitrogen). AmpliWcation was based on a cycle scheme of 10 min at 94°C, 35 cycles of 1 min at 94°C, 45 s at 56°C, 2 min 30 s at 72°C, and a Wnal extension step for 10 min at 72°C. PCR products of expected size were cloned using pGEM-Teasy vector protocol (Promega). One or two clones were sequenced for each species (Genome Express) and sequences deposited in GenBank under accessions DQ912982 to DQ912988. Southern hybridization About 3
g of genomic DNA of one specimen per species were digested with HindIII endonuclease (Fermentas), and separated overnight on a 0.8% agarose gel in phosphate buVer (TRIS 0.36 M, NaH2PO4 0.3 M, EDTA 0.1 M, and pH 7.5). The gel was Southern blotted on a nitrocellulose membrane (Schleicher and Schuell) according to the standard method (Sambrook et al. 1989). A probe corresponding to putative integrase and RT domains of GalEa1 retrotransposon was isolated from E. annulosa by PCR ampliWcation using GalEa1_1951+ and GalEa1_3516primers. The 1,565 bp fragment was labeled with [32P] dCTP using the High Prime DNA labeling kit (Roche, Penzberg, Germany). The probe was hybridized to the membrane overnight in 6£ SSC (Sodium and Sodium Citrate), 5£ Ficoll, Polyvinylpyrrolidone, and Glycine (FPG), 0.5% Sodium Dodecyl Sulfate (SDS) at 65°C. Membranes were washed twice 10 min at 65°C in 2£ SSC, 0.1% SDS and hybridization signals were visualized by autoradiography. Sequence analysis We used ClustalX (Thompson et al. 1997) to generate multiple sequence alignments using default parameters. All alignments were visually inspected, then ambiguously aligned sites were removed using the Gblocks software (Castresana 2000). Based on these alignments, phylogenetic analyses were conducted using Maximum likelihood, Bayesian and Parsimony methods. Maximum likelihood analyses were performed using PHYML v2.4.4 (Guindon and Gascuel 2003) with GTR model of nucleotide substitution, estimated proportion of invariable sites and one category of substitution rate. Support for individual clades was evaluated using non-parametric bootstrapping (Felsenstein 1985) obtained from 100 bootstrap replicates. Bayesian analyses were conducted using MrBayes 3.1 (Ronquist and Huelsenbeck 2003). MrModeltest v 2.2 (Nylander 2004) and ProtTest were used to select the best-Wt model and priors were set according to the suggested model. Markov chains were run for six million cycles with four chains

123

Mol Genet Genomics (2008) 279:63–73

starting from a random tree. Samples of the Markov chain were taken every 100 generations. All parameters were checked for stationary state with the program Tracer 1.2 (Rambaut and Drummond 2005) and 20% of the trees were discarded as burn-in. A majoritary consensus tree was constructed from the remaining trees. Maximum parsimony analyses were performed with PAUP* 4.0 Version 10 using heuristic search and Tree Bisection-Reconnection (TBR) with 20 random additions. Character states were considered unordered and equally weighted. Gaps were coded as missing data or excluded from the analyses. Bootstrap support for the most parsimonious topology was evaluated using 100 replicates.

Results Phylogenetic relationships based on the molecular 16S The Wrst step of our study was to ensure the identiWcation of the specimen from the hydrothermal vent we used. For this, we determined 16S partial sequence of the specimen of the Munidopsis genus (EF428964). It was veriWed to belong to M. recta, the most common species on the East PaciWc Ride. We also used 16S sequences to determine phylogenetic relationships between the nine diVerent galatheid species from our data set (Fig. 1) with sequences obtained by ourself for E. annulosa (EF428962) and M. gregaria (EF428963) or obtained from GenBank database, adding sequences from some Galatheidea (i.e., Porcellanidae). Both topology and resolution of the phylogenetic tree are well conserved whatever method is used. Species of the Munida genus form a well supported clade, which conWrms the monophyly of this group (Machordom and Macpherson 2004). M. recta clearly groups with Galatheidae and presents a basal position among them. Species of the Eumunida genus (Chirostylidae) forms a well statistically supported clade. Isolation of RT sequences using degenerate primers We looked for Ty1/copia-like retrotransposons using degenerate primers on a selected subset of three species (A. laurentae, E. annulosa, and E. sternomaculata) using primers corresponding to the “KARLVA” and “YV/LDD” conserved motifs of the RT domain. For each species, only PCR ampliWcations of the expected size (380 bp) were cloned, and three clones of the putative RT fragment were sequenced. From these nine sequences, seven present signiWcant similarities with RT domains of Ty1/copia retrotransposons. A Bayesian phylogenetic analysis revealed three families (Fig. 2), which were named GalEa1 to GalEa3 (for Galatheids Eumunida annulosa retrotransposons 1–3).

Mol Genet Genomics (2008) 279:63–73

GalEa1 family (97.0 %)

67 60.1% -64.0%

Ea1 Al1 Es1 100 / 100 1.00 99 / 98 0.99

Es3 Ea3

100 / 98 0.97

GalEa3 family (90.0 %)

60.4% -62.3% Ea2

0.05 Substitution per site

Al2

55.7% -62.9%

GalEa2 family (91.5 %)

Fig. 2 Phylogenetic hypothesis of the diVerent RT partial sequences found in E. annulosa (Ea), E. sternomaculata (Es), and A. laurentae (Al). We obtained this tree by Bayesian inference. The best-Wt model was GTR (General Time Reversible) +
+ I. Bootstrap values for parsimony and Maximum likelihood methods, respectively, and posterior probabilities for Bayesian inference are indicated above the nodes. Lowest intra-family DNA identity is shown next to each group. Lowest and highest inter-family DNA identity are shown above each arrow. Scale bar indicates percent sequence divergence

The DNA identity between all RT sequences of the diVerent families ranges from 55 to 64%. The GalEa1 family contains one sequence from each species (DQ912991, DQ912992, and DQ912995), and appears to be more conserved with <3% divergence. The GalEa2 family encompasses one sequence from A. laurentae (DQ912989) and one from E. annulosa (DQ912993) that diVer by 9%. The GalEa3 family includes one sequence from E. sternomaculata (DQ912990) and one from E. annulosa (DQ912994) that diVer by 10%. IdentiWcation of GalEa1, a putative complete Ty1/copia-like retrotransposon We were interested in Wnding a TE that was largely distributed among galatheid species Therefore, the most conserved family, GalEa1, was further analyzed. We chose to characterize a complete retrotransposon in E. annulosa, because it was the only species in which we identiWed elements from the three GalEa families (Fig. 2). Using several walking steps (see Materials and methods for details), we reconstructed a 4,669 bp retrotransposon, which we named GalEa1 (Fig. 3a, EU097705). Along with the great similarities between GalEa1 and the other TEs, the analysis of its structure showed that GalEa1 belongs to the Ty1/copia group. It has a large internal domain of 4,421 bp Xanked by 124 bp LTRs. Both LTRs start with 5⬘-TG and end with CA-3⬘ as observed in many retrotransposons, but we did not Wnd any nucleotide sequence corresponding to putative TATA box or CAAT box. At position 126, the internal region carries a Primer Binding Site (PBS) that sequence (TGGTAGCAGAGC) is complementary to the 3⬘ end

region of D. melanogaster tRNAMet gene. We were also able to localize a putative PolyPurine Tract (PPT) signal (GAAGAAATGGA) at position 4,522. The central part of GalEa1 comprises a single large ORF (which harbors in our sequence a stop codon at position 1,516 and a frameshift at position 1,979) that exhibits all Wve typical ordered domains of Ty1/copia retrotransposons gag and the pol sequences (whose amino acid sequences are depicted in Fig. 3a). The Wrst domain is the nucleic acid-binding domain containing zinc-Wnger motif (CX2CX4HX4C) that is found in all retroviral gag genes. The second domain is the protease (PR) domain with the typical DSGA motif of Ty1/copia retrotransposons being substituted by a DSGC motif in GalEa1. The third identiWable domain is the IN domain with the HX4HX30CX2C and DD35E signatures. The fourth domain is the RT domain containing seven subdomains conserved in all RT sequences (Capy et al. 1997; Xiong and Eickbush 1990). The Wfth domain corresponds to RH with the highly conserved TRPDI motif. For each of these domains, we performed an alignment of GalEa1 sequence with previously described Ty1/copia retrotransposons (S1). GalEa1 elements in the genomes of other galatheid crabs To assess the distribution of GalEa1 retrotransposons among galatheid species, we performed both Southern blot and PCR analyses. For Southern blot analysis, HindIII-digested genomic DNAs isolated from single specimens were probed with a 1,565 bp PCR ampliWcation fragment of GalEa1 from E. annulosa that corresponds to part of the IN and RT domains (Fig. 3a). We performed the interspecies search on eight of the nine species studied (all but G. squamifera). Of the eight species considered (Fig. 4), six (lanes Al, Mz, Mg, Mr, Ea, and Es) show a fragment of the expected size (around 1.6 kb). In these species weaker hybridization signals of various sizes were also revealed. This suggests that deleted elements and/or elements showing Hind III site polymorphism are also present in these genomes. This is particularly striking for M. gregaria (lane Mg), where a strong signal is observed around 1.4 kb. For M. acantha (lane Ma) only a weak hybridizing fragment of a large size (around 2.5 kb) can be observed. Surprisingly, no signal is observed in M. thoe (lane Mt) although Ethidium Bromid signal in the agar gel conWrmed that an equivalent amount of DNA was loaded in each lane (data not shown). To compare the IN-RT sequences of GalEa1 elements from diVerent species, we performed direct PCR ampliWcations using a pair of speciWc primer (GalEa1_1951+ and GalEa1_3516-). As might be guessed from the Southern blot analysis, no PCR signal was observed in M. thoe and M. acantha (even with low annealing temperature). In the

123

68 Fig. 3 Structural organization of the GalEa retrotransposon. Dark gray boxes indicate Long Terminal Repeats, light gray rectangles indicate the ORF, and dotted white boxes represent undetermined part. The probe used for Southern blot is indicated by a thick solid line. Amino acid sequences coding for the diVerent motifs or signatures of Ty1/copia retrotransposons are shown above the elements. PBS primer binding site, PPT poly purine tract signal a GalEa1 retrotransposon from E. annulosa (EU097705) and b GalEa-like elements Cico1 (DQ913003), Zeco1 (DQ913001), and Olco1 (DQ913000)

Mol Genet Genomics (2008) 279:63–73

a

PBS tRNAMet

CX2CX4HX4C DSGC

HX4HX30CX2C DX49DX35E

YVDD KARLVA

PPT TRPDI 4669bp

GalEa1 124bp

b

GA2GA3TG2A

Probe (1565bp)

CA

TG

PBS tRNAMet

CX2CX4HX4C DTAC

HX4HX30CX2C DX54DX35E

YVDD KARLVA

PPT TRPDI 4513bp

Cico1 323bp

HVDD

CA

TG

A4GA4CAGA2

4807bp

Zeco1

TG

192bp CA

DX35E

SRPDV

AG3AGA 4760bp

Olco1 A14

187bp TG

CA 500bp

are deletions within the IN-RT domains shared by GalEa1 elements from diVerent species: a Wrst deletion of 18 bp (1 at position 2,763) is shared by sequences from A. laurentae, M. gregaria, G. squamifera and M. zebra and a second deletion of 54 bp (2 at position 3,332) is shared by sequences from A. laurentae and M. gregaria. Thus, sequences from A. laurentae (1,478 bp) and M. gregaria (1,494 bp) are slightly shorter than the others. Finally, the GalEa1 sequence from M. zebra is more degenerated with a size of only 1,404 bp due to many deletions. Excluding insertions and deletions, sequence comparison shows high DNA sequence identity that ranges from 69.1% for the most divergent sequences (between M. zebra and G. squamifera) to 98.8% (between A. laurentae and M. gregaria). Fig. 4 Southern analysis of GalEa1. HindIII-digested genomic DNA extracted from a single specimen of galatheid species. DNA quantity was equivalent in each lane. The PCR ampliWed fragment corresponding to IN and RT parts of GalEa1 from E. annulosa (Fig. 3a) was used as probe. Lanes: Al Agononida laurentae, Mt Munida thoe, Mz M. zebra, Ma M. acantha, Mg M. gregaria, Mr Munidopsis recta, Ea Eumunida annulosa, and Es E. sternomaculata

other seven species, including G. squamifera, we ampliWed and sequenced a fragment corresponding to GalEa1 elements. When two clones were sequenced (E. annulosa, A. laurentae, and M. recta) sequences were 100% identical. The seven sequences were deposited in GenBank as DQ912982 to DQ912988. Only GalEa1 sequence from M. recta does not present any stop codon or frameshift (1,568 bp). In E. sternomaculata and G. squamifera, the sequences are, respectively, 1,568 and 1,582 bp in size but are highly corrupted by multiple frameshifts and/or stop codons. Interestingly, compared to GalEa1 from E. annulosa (for alignments see supplementary materials S2), there

123

Searching for GalEa-like elements in other genomes In order to examine the distribution of GalEa1 related elements, we retrieved homologous TEs with NCBI-BLAST through a TBLASTN search using the GalEa1 pol protein domains as query. High signiWcant hits (e-value from 0.00 to 2 £ 10¡97) were found in the Whole-Genome-Shotgun (wgs) database for the urochordate Ciona intestinalis and Wshes Danio rerio and Oryzias latipes. This allowed us to characterize four new retrotransposons: Cico1 (DQ913003) and Cico2 (DQ913004) in C. intestinalis, Zeco1 (DQ913001) in D. rerio and Olco1 (DQ913000) in O. latipes. For Cico2, we were unable to precise the ends of the element as subsequent analyses of the hits did not allow to recover the 5⬘ part of the element. Olco1 was reconstructed using two distinct sequences, respectively, 5⬘ and 3⬘ parts of the element, and so present, according to alignments with other GalEa-like elements, a missing part of 55 amino

Mol Genet Genomics (2008) 279:63–73

acid in IN (including Wnal CX2C of the HX4HX30CX2C and Wrst D of the DD35E signatures, Fig. 3b). Except for Cico2, all the coding sequences obtained are corrupted by frameshifts (between IN and RT for Cico1, in gag gene for Zeco1 and in RH for Olco1) and stop codons (at positions 926, 2,044, 3,397 for Zeco1 and 2,997 for Olco1), suggesting that the copies described are no longer active in host genomes. Principal features of these GalEa-like retrotransposons are presented in Fig. 3b. Element lengths (4,513– 4,760–4,807 bp) are similar to those of GalEa1 from E. annulosa even if their LTRs are longer (187 to 323 bp). These elements share numerous characteristics with GalEa1, such as LTRs bordered by 5⬘-TG and CA-3⬘; a PBS homologous to tRNAMet; a large single ORF containing a zinc-Wnger motif (CX2CX4HX4C) in the gag region, the HHCC and DD35E signatures of the integrase amino acid sequence (only partial signatures observed for Olco1) and the KARLVA motif of the RT. However, each element presents some particularities. All four have an uncommon DTAC motif in the PR coding region, Wshes elements Zeco1 and Olco1 have a HVDD (instead of YVDD) motif in RT, and Olco1 has a SRPDV (instead of TRPDI) motif in RH. The most variable part is the putative PPT sequence that consists in (A4GA4CAGA2) for Cico1, (A6GA5GA6) for Cico2, (AG3AGA) for Zeco1, and (A)14 for Olco1. Average GC content is about 0.41 for all GalEa-like sequences (GalEA1 = 0.46, Zeco1 = 0.41, and Olco1 = 0.41) except for Cico1 and Cico2 (Cico1 = 0.36; Cico2 = 0.37). The lower GC content observed for C. intestinalis seems to be related to the global GC content of the genome.

Discussion We characterized the Wrst LTR retrotransposon in crustaceans, GalEa1 from E. annulosa, which sequence shows features that are characteristic of Ty1/copia retrotransposons. As GalEa1 was reconstructed from several PCR fragments, we were unable to compare the two LTRs of a copy and, therefore, to determine the nature of the target site duplication. Nevertheless, we could get access to such information using complete copies of GalEa-like elements that we retrieved from databases. Comparison of 5⬘ and 3⬘ LTR sequences revealed 100% of identity for Cico1 and Zeco1 suggesting that such elements may have been recently active. This is supported by the fact that cDNAs from Cico1 are also detected in databases. Nucleotide sequence analyses of seven Zeco1 copies revealed that they are Xanked by a target-site duplication of 5 bp that is common for Ty1/copia retrotransposons. Moreover, we found that GalEa1 is a widespread element in galatheid crabs. Of the nine galatheid species studied, seven harbor highly related copies. An additional species, M. acantha, seems to present some traces of

69

more distantly related copies according to Southern blot analyses. Only one species, M. thoe, does not harbor any traces of GalEa1. As this TE is present in most of the galatheid species studied, the most parsimonious hypothesis is to consider that GalEa1 has been lost in M. thoe. However, we cannot exclude the possibility that the element was present but has diverged to the point that it is no more detectable by hybridization methods. In addition, it appears that GalEa1 elements are highly conserved at the DNA level in geographically widespread species. Such pattern is similar to the distribution and copy conservation observed for the Sea Urchin Retroviral-Like (SURL) elements (Gonzalez and Lessios 1999) or mariner-like elements in marine crustacean genomes (Casse et al. 2006). Moreover, the high conservation level of GalEa1 between Galatheidae and Chirostylidae raises the question of its distribution among Galatheidea or, more largely, among Anomoura. To assess the relationship between the diVerent GalEalike elements, we performed a phylogenetic analysis based on RT-RH DNA sequences using the copies that were molecularly characterized in galatheid crabs and the copies extracted from BLAST searches (Fig. 5). We observed two monophyletic groups with a strong statistical support, as indicated by high bootstrap values and posterior probabilities. The Wrst cluster contains all GalEa1 elements from galatheid crabs and the second harbors the GalEa-like elements from Chordata (i.e., C. intestinalis and Wshes). Within galatheids, GalEa1 from M. recta groups with elements from E. annulosa and E. sternomaculata and none of them present the 1 or 2 deletions. The observation of the 1 deletion in all other sequences suggests that the corresponding mutation event corresponds to an insertion in this group. GalEa1 from A. laurentae groups with elements from M. gregaria and M. zebra. They all present the 1 deletion but only the two Wrst share the 2 deletion, which suggests that the corresponding deletion event occurred in this clade. The element from G. squamifera forms a third clade that is located at a basal position. The topology observed is not congruent with the species phylogeny (Fig. 1). However, as each species was represented by a partial sequence of a single copy, we cannot discriminate whether the lack of congruence between phylogenies reXects possible horizontal transfers, or intraspeciWc polymorphism of the GalEa1 elements. Within Chordata, GalEa-like elements form a cluster that is strengthened by the share of several signatures as illustrated by the DATC protease motif. TEs from teleosteans form a monophyletic group, when TEs from C. intestinalis are more divergent and form a paraphyletic group. Since genomes from C. intestinalis, D. rerio, and O. latipes are partially or totally assembled, it is possible to use genome databases to estimate the number of copies of GalEa-like TEs by running BLASTN searches on these genomes using a speciWc copy of each element (i.e., Cico1

123

70

Mol Genet Genomics (2008) 279:63–73 ∆1

GalEa1_G.squamifera (DQ912988)

*

GalEa1 _E.sternomaculata (DQ912987) GalEa1_ M.recta (DQ912982)

* 96 / 93 * 100 / 98

1.00

GalEa1_A.laurentae (DQ912986)

∆1 ∆2

GalEa1 _M.gregaria (DQ912983)

∆1 ∆2

GalEa1_ M.zebra (DQ912984)

1.00

Galatheidea

*

*

GalEa1 _E.annulosa (DQ912985)

∆1

Olco1_ O.latipes (DQ913000) 65 / 72

Zeco1_D.rerio (DQ913001)

0.54 91 / 87

*

1.00

Cico1 _C.intestinalis (DQ913003)

Chordata

*

Cico2_C.intestinalis (DQ913004) Copia_D.melanogaster (X04456) *

Mtanga _A.gambie (AF387862) Yokozuna_B.mori (AB014676)

0.1

Substitution per site

Fig. 5 Phylogenetic hypothesis based on RT and RNase H DNA sequences of GalEa-like elements represented by a tree obtained by Bayesian inference using Ty1/copia elements from Drosophila melanogaster, Anopheles gambiae, and Bombyx mori as outgroup. The best-Wt model was GTR (General Time Reversible) +
+ I. Bootstrap values for parsimony and Maximum likelihood methods, respectively,

separated by a slash character, and posterior probabilities for Bayesian inference are indicated above nodes. Nodes with a bootstrap support of 100% and a posterior probability of 1 are indicated with a star. Scale bar indicates percent sequence divergence. Accessions are given in brackets and “1” and “2” refer to shared deletions compared to reference sequence GalEa1

and Cico2 for C. intestinalis, Zeco1 for D. rerio, and Olco1 for O. latipes as query). We focused on D. rerio, because the sequencing project of its genome is the most advanced. All sequences homologous to Zeco1 were isolated and Xanking regions were compared using Dot Plot analyses to avoid redundancy. This allowed to detect about ten distinct copies: two full-length elements, two partial sequences corresponding to rearranged or highly deleted elements and seven solo LTRs. The number of copies was more diYcult to estimate for Cico1, Cico2, and Olco1. We found a maximum of three to four diVerent copies for each of them. So, even if this numbers are probably underestimated, they favor the idea that GalEa-like elements are present in few copies in these genomes, as expected for compact genomes such as that of Wsh (VolV et al. 2003). The observation of such low number of copies could explain why these TEs were previously not described despite intensive research on these model organisms. Despite their low number of copies, GalEa-like elements may be still active in these genomes as suggested by complete LTRs identity (Cico1, Zeco1) or detection of cDNA copies in databases (Cico1 and Cico2). Altogether, our data suggest that the response of Wshes and C. intestinalis genomes to GalEa-like elements are similar to those observed for other classes of TEs previously described, i.e., low number of copies but high sequence turnover (VolV et al. 2003). In squat lobsters, genomic number of copies of GalEa1 elements can be evaluated

through Southern blot analysis (Fig. 4) and seems to be higher. For example, M. recta harbors a various subset of GalEa1 copies and a strong intensity for the internal fragment signal. To a lesser extent, this is also observed for E. annulosa, E. sternomaculata, and M. gregaria. However, thorough estimations of the numbers of copies in galatheid species, including inter-population variability S-SAP (Waugh et al. 1997) analyses, are needed to determine whether GalEa1 elements are frequent or not in Galatheidae genomes. To analyze the relatedness of GalEa-like elements with other retrotransposons, we conducted a phylogenetic analysis using multiple alignments of protein RT-RH sequences (Fig. 6). Because topologies constructed from each domain were congruent with one another (data not shown), we used them simultaneously to increase the resolution and statistical support of the topology (Malik et al. 1999). The core data set is composed of 26 Ty1/copia sequences from divers organisms and three related Ty3/gypsy retrotransposons that were used as outgroup to perform a family wide comparison. It is noteworthy to mention that we observed the same topology and node statistical support using either Parsimony, Bayesian or Maximum likelihood methods. The resulting tree (Fig. 6) conWrms that GalEa-like elements groups with the Ty1/copia retrotransposons. However, the topology found reveals two clearly distinct groups: GalEalike retrotransposons are separated from all other Ty1/copia

123

Mol Genet Genomics (2008) 279:63–73

71

Fig. 6 Phylogenetic hypothesis based on RT and RH amino acid sequences represented by a tree obtained by Bayesian inference under the best-Wt model of protein evolution rtREV +
+ I using Ty3/gypsy elements from Ciona intestinalis and Danio rerio as outgroup. Bootstrap values for parsimony and Maximum likelihood methods, respectively, separated by a slash character, and posterior probabilities for Bayesian inference are indicated above nodes. Nodes with a bootstrap support below 75% were collapsed, nodes with a bootstrap support of 100% and a posterior probability of 1 are indicated with a star. Scale bar indicates percent sequence divergence. We labeled each element with the name of the retrotransposon (when identiWed) followed by the species name. Accessions are given in brackets

Anopheles gambiae (AY009101) Salto_A.gambiae (AF295692)

83 / 69 0.89 100 / 90 1.00

mtanga_A.gambiae (AF187862) Sorgum bicolor (AF466199) Ipomoea trifida (AH013750) 1731_D.melanogaster (X07656) A.gambiae (COPIA5-I_AG) Sto-4_Zea mays (AF082133) Tnt 1_Nicotiana tabacum (X13777)

95 / 89

Tto1_N.tabacum (D83003)

0.87

Arabidopsis thaliana (AB018112) Ta1-3_A.thaliana (X13291) Osser_Volvox carteri (X69552) Chlamydomonas reinhardtii (Copia1-I_CR) Lueckenbuesser_V. carteri (U90320) A.thaliana (AL161556) Hopscotch_Z.mays (U12626)

79 / 83 0.90

98 / 98 *

1.00

99 / 100 0.98

Cico2_C.intestinalis (DQ913004)

76 / 77 0.80

Olco1_ O.latipes (DQ913000)

* *

AtRE1_A.thaliana (AB021265) Glycine max (AF053008) Rire1_O.sativa (D85597) Yokozuna_B.mori (AB014676) GalEa1 _E.annulosa (DQ912985) Cico1 _C.intestinalis (DQ913003)

98 / 96 0.96

GalEa-like retrotransposons

Zeco1_D.rerio (DQ913001)

C.intestinalis (Cigr-1-I_CI) D.rerio (Gypsy18-I_DR) D.rerio (Gypsy10-I_DR)

0.2

Substitution/site

elements and form a particular monophyletic group (including GalEa2 and GalEa3, data not shown). So, they deWned a special new clade of Ty1/copia-like elements that is largely dispersed among animal species, from crustaceans to Wshes. Among all the features of the GalEa-like elements, one of the most striking is their complex distribution that presents three characteristics: they are present in phylogenetically distant species, in a greatly discontinuous way and limited, for what we see with our results, to aquatic environments. Such pattern can be compared to those observed for other elements. For instance, Abyss1-like elements, that were initially characterized in anemones, show highly related members in Wsh, sea urchins and molluscs (Greenwood et al. 2005). L2-like elements are present in echinoderms and Wshes in a conserved form, and show highly degenerate copies in most Tetrapoda (Lovsin et al. 2001). This is also the case for a Gypsy group that shows surprising amino acid sequence similarity among members from herring, sea urchin, starWsh, and a tunicate (Britten et al. 1995), as do Penelope-related elements from Wsh, sea urchin, shrimp, frogs and trematode (VolV et al. 2001a). The distribution of GalEa-like elements is highly discontinuous because restricted to some distant clades of bilateria as urochordates and teleosts. But discontinuity also appears within teleosts, as GalEa-like TEs are not found in the well-studied genomes of Takifugu rubipes or Tetraodon nigroviridis, which are closely related to O. latipes. We noticed the same

pattern in urochordates as C. intestinalis harbors GalEa-like copies whereas no copy was detected in C. savignyii. NonLTR retrotransposon Rex3 seems to follow the same discontinuous distribution inside the Wsh lineage (VolV et al. 2001b). Such pattern can be explained by at least two hypotheses: horizontal transfers and/or stochastic loss. VolV and Schartl (2000) had suggested the possibility that retroposons are capable of horizontal transfer among distantly related taxa. It could also be the case for retrotransposons and could be more frequent than previously suspected. DiVerential loss can also be considered according to the low number of copies detected in the diVerent genomes. Whatever the hypothesis considered, for now, the vision we get is greatly inXuenced by the spectrum of the phyla for which large genome sequences are currently available in databanks. Therefore, for a better understanding of mechanisms underlying GalEa-like elements evolution, it is still necessary to precise GalEa-like elements distribution. This will furthermore allow to conWrm the “aquatic nature” of GalEalike elements. Such characteristic, which seems to be shared by some of the elements cited above (e.g., Abyss1), remains diYcult to explain, in particular since only galatheid crabs and C. intestinalis are marine species. It suggests that aquatic environment may facilitate horizontal transfers, or that some speciWc evolutionary forces (selection pressure, eVective population size (Ne), and migration) may act in such ecosystems.

123

72 Acknowledgments We would like to thank Claude Bazin, Clémentine Vitte, Guillaume Achaz, and Philippe Lopez for helpful comments on manuscript and Malcom Eden for English revision. We also thank Enrique MacPherson for biological and phylogenetical informations on squat lobsters species, and Marie-Catherine Boisselier, Michel Descombes, Guillaume Lecointre, and Bertrand Richer de Forges for providing biological material.

References Baba K (1988) Chirostylidae and Galatheidae Crustaceans (Decapoda: Anomura) of the Albatross Philippine Expedition, 1907–1910. Res Crust 2:1–203 Britten RJ, McCormack TJ, Mears TL, Davidson EH (1995) Gypsy/ Ty3-class retrotransposons integrated in the DNA of herring, tunicate and echinoderms. J Mol Evol 40:13–24 Bui Q-T, Delaurière L, Casse N, Nicolas V, Laulier M, Chénais B (2007) Molecular characterization and phylogenetic position of a new mariner-like element in the coastal crab, Pachygrapsus marmoratus. Gene 396:248–256 Burke WD, Malik HS, Jones JP, Eickbush TH (1999) The domain structure and retrotransposition mechanism of R2 elements are conserved throughout arthropods. Mol Biol Evol 12(4):502–511 Capy P, Bazin C, Higuet D, Langin T (1997) Dynamics and evolution of transposable elements. R. G. Landes Company, Austin, TX Casse N, Pradier E, Loiseau C, Bigot Y, Laulier M (2000) Mariner, a mobile DNA transposon in the genomes of several hydrothermal invertebrates. InterRidge News 9:15–17 Casse N, Bui QT, Nicolas V, Renault S, Bigot Y, Laulier M (2006) Species sympatry and horizontal transfers of Mariner transposons in marine crustacean genomes. Mol Phylogenet Evol 40(2):609–619 Castresana J (2000) Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol 17(4):540–552 Devic M, Albert S, Delseny M, Roscoe TJ (1997) EYcient PCR walking on plant genomic DNA. Plant Physiol Biochem 35:331–339 FedoroV N (2000) Transposons and genome evolution in plants. Proc Natl Acad Sci USA 97(13):7002–7007 Felsenstein J (1985) ConWdence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791 Finnegan DJ (1989) Eukaryotic transposable elements and genome evolution. Trends Genet 5(4):103–107 Flavell AJ, Dunbar E, Anderson R, Pearce SR, Hartley R, Kumar A (1992) Ty1-copia group retrotransposons are ubiquitous and heterogeneous in higher plants. Nucleic Acids Res 20(14):3639–3644 Gonzalez P, Lessios HA (1999) Evolution of sea urchin retroviral-like (SURL) elements: evidence from 40 echinoid species. Mol Biol Evol 16(7):938–952 Gregory TR (2007) Animal genome size database. http://www.genomesize.com. Cited Jan 2007 Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52(5):696–704 Greenwood AD, Leib-Mösch C, Seifarth W (2005) Abyss1: a novel L2like non-LTR retroelement of the snakelocks nemone (Anemonia sulcata). Cytogenet Genome Res 110:553–558 Halaimia-Toumi N, Casse N, Demattei MV, Renault S, Pradier E, Bigot Y, Laulier M (2004) The GC-rich transposon Bytmar1 from the deep-sea hydrothermal crab, Bythograea thermydron, may encode three transposase isoforms from a single ORF. J Mol Evol 59(6):747–760 Hiller A, Holger K, Almon M, Werding B (2006) The Petrolisthes galathinus complex: species boundaries based on color pattern, morphology and molecules, and evolutionary interrelationships

123

Mol Genet Genomics (2008) 279:63–73 between this complex and other Porcellanidae (Crustacea: Decapoda: Anomura). Mol Phylogenet Evol 33(2):259–279 Ishaq M, Wolf B, Ritter C (1990) Large-scale isolation of plasmid DNA using cetyltrimethylammonium bromide. Biotechniques 9(1):19–24 Jordan IK, Rogozin IB, Glazko GV, Koonin EV (2003) Origin of a substantial fraction of human regulatory sequences from transposable elements. Trends Genet 19(2):68–72 Jurka J (1998) Repeats in genomic DNA: mining and meaning. Curr Opin Struct Biol 8(3):333–337 Kazazian HH Jr (2004) Mobile elements: drivers of genome evolution. Science 12:1626–1632 Kidwell MG, Lisch D (1997) Transposable elements as sources of variation in animals and plants. Proc Natl Acad Sci USA 94(15):7704–7711 Kidwell MG (2002) Transposable elements and the evolution of genome size in eukaryotes. Genetica 115:49–63 Kvamme BJ, Kongshaug H, Nilsen F (2005) Organisation of trypsin genes in the salmon louse (Lepeophtheirus salmonis, Crustacea, copepoda) genome. Gene 352:63–74 Li WH, Gu Z, Wang H, Nekrutenko A (2001) Evolutionary analysis of the human genome. Nature 409:847–849 Lovsin N, Gubensek F, Kordi D (2001) Evolutionary dynamics in a novel L2 clade of non-LTR retrotransposons in Deuterostomia. Mol Biol Evol 18:2213–2224 McClure MA (1991) Evolution of retroposons by acquisition or deletion of retrovirus-like genes. Mol Biol Evol 8(6):835–856 Machordom A, Macpherson E (2004) Rapid radiation and cryptic speciation in squat lobsters of the genus Munida (Crustacea, Decapoda) and related genera in the South West PaciWc: molecular and morphological evidence. Mol Phylogenet Evol 33(2):259–279 Malik HS, Burke WD, Eickbush TH (1999) Modular evolution of the integrase domain in the Ty3/Gypsy class of LTR retrotransposons. J Virol 73(6):5186–5190 Malik HS, Eickbush TH (2001) TH Phylogenetic analysis of ribonuclease H domains suggests a late, chimerical origin of LTR retrotransposable elements and retroviruses. Genome Res 11(7):1187– 1197 Misra S, Crosby MA, Mungall CJ et al (26 co-authors) (2002) Annotation of the Drosophila melanogaster euchromatic genome: a systematic review. Genome Biol 3(12):1–22 Nylander JAA (2004) MrAIC.pl. Program distributed by the author. Evolutionary Biology Centre, Uppsala University Ochman H, Ajioka JW, Garza D, Hartl DL (1990) Inverse polymerase chain reaction. Biotechnology (NY) 8:759–760 Palumbi SR, Benzie J (1991) Large mitochondrial DNA diVerences between morphologically similar Penaeid shrimp. Mol Mar Biol Biotechnol 1(1):27–34 Penton EH, Sullender BW, Crease TJ (2002) Pokey, a new DNA transposon in Daphnia (cladocera: crustacea). J Mol Evol 55(6):664– 673 Petrov DA, Sangster TA, Johnston JS, Hartl DL, Shaw KL (2000) Evidence for DNA loss as a determinant of genome size. Science 287:1060–1062 Poulter RT, Goodwin TJ (2005) DIRS-1 and the other tyrosine recombinase retrotransposons. Cytogenet Genome Res 110:575–588 Rambaut A, Drummond A (2005) Tracer version 1.2.1. Computer program distributed by the authors. Department of Zoology, University of Oxford, UK Robertson HM (1997) Multiple mariner transposons in Xatworms and hydras are elated to those of insects. J Hered 88:195–201 Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

Mol Genet Genomics (2008) 279:63–73 Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The CLUSTAL_X windows interface: Xexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25(24):4876–4882 Van de Lagemaat LN, Landry JR, Mager DL, Medstrand P (2003) Transposable elements in mammals promote regulatory variation and diversiWcation of genes with specialized functions. Trends Genet 19:530–536 VolV JN, Schartl M (2000) Multiple lineages of the non-LTR retrotransposon Rex1 with varying success in invading Wsh genomes. Mol Biol Evol 17:1673–1684 VolV JN, Hornung H, Schartl M (2001a) Fish retotransposons to the Penelope element of Drosophila virilis deWne a new group of retrotransposable elements. Mol Genet Genomics 265:711–720

73 VolV JN, Körting C, Meyr A, Schartl M (2001b) Evolution and discontinuous distribution of Rex3 retrotransposons in Wsh. Mol Biol Evol 18(3):427–431 VolV JN, Bouneau L, Ozouf-Costaz C, Fischer C (2003) Diversity of retrotransposable elements in compact puVerWsh genome. Trends Genet 19:674–678 Waugh R, McLean K, Flavell AJ, Pearce SR, Kumar A, Thomas BBT, Powell W (1997) Genetic distribution of Bare-1-like retrotransposable elements in the barley genome revealed by sequence-speciWc ampliWcation polymorphisms (S-SAP). Mol Gen Genet 253:687–694 Xiong Y, Eickbush TH (1990) Origin and evolution of retroelements based upon their reverse transcriptase sequences. EMBO J 9(10):3353–3362

123