Our attention was drawn to the S. mansoni genomic sequence of Supercontig_0018735 available at the Wellcome Trust Sanger Institute 20 while we were manually examining the splicing pattern of the gene represented by transcript SmAE C610100.1 16. Three exons of the latter mapped to bases 2032 to 2160, 6905 to 6961 and 7001 to 7231 of Supercontig_0018735. Upon examining the intron formed between bases 2160 to 6905 we detected a 4.5 kbp element that extended between bases 2164 and 6907 with one Open Reading Frame (referred to as SmTRC1-ORF in the following text) of 1,683 bp, which codes for a sequence with similarity (E-value 10^-5) to a DNA-only transposon. An inverted repeat motif was found at both extremities (the left with 54 bp and the right missing only one base), suggesting that this intron is an inserted mobile element (Figure 1B). Although the element may be considered large (4.5 kpb) in comparison, for example, to Mariner family transposons (1.3 to 2.4 kpb) 21, it is considerably smaller than CACTA elements such as Spm/En (8.3 kpb) 8 and Rim coding elements (14.1 kpb on average) 22. We detected several other copies of the element in the S. mansoni genome sequence dataset displaying a perfect 54 bp inverted repeat at both extremities, confirming the sequence and correct length of the Transposon Inverted Repeat (TIR) (Figure 2A). We named these elements SmTRC1 (an abbreviation for Schistosoma mansoni Transposon Related to CACTA transposons) because of its several similarities to transposons of the CACTA family, as described in the text below.

We found a 2 bp direct repeat suggestive of target site duplication flanking the inverted repeat in 7 out of 14 (50%) of the copies of these elements analyzed (Figure 2B). Three of these copies were found to be inserted into repetitive elements of the S. mansoni genome. Two of them were inserted into copies of the S. mansoni LTR retrotransposons Saci-1 and Saci-4 1819 and one into an unidentified repetitive element of which there are at least 40 copies in the preliminary assembly of the S. mansoni genome. When the flanking sequences of these three elements were aligned with paralogous sequence copies of the respective repetitive element [GenBank:BH202328.1, GenBank:BN000802.1, GenBank:AL620357.1] obtained from either the GSS or the nr databases at GenBank, it was clear that the original repetitive element (not having an inserted transposon) contained none of the direct repeats flanking the SmTRC1 elements (Figure 2C); one of the 2bp direct repeat motifs was missing from the original repetitive element. This suggests that the direct repeats are in fact target-site duplications created by insertion of the SmTRC1 element. It is well known that transposon integration results in the duplication of a short host sequence at the insertion site and that the length of the target-site duplication is determined by the properties of each transposase 3, therefore these data provide further evidence for the mobility of SmTRC1 elements.

Using different combinations of a set of primers designed from the extremities of the transposon sequence from Supercontig_0018735, we performed several PCR reactions to amplify genomic copies of SmTRC1. A typical result is shown in Figure 1A; the major products are approximately 2.5 kbp. Cloning and sequencing of this major band revealed a 2274 bp clone that represented a truncated copy of SmTRC1, which we designated SmTRC1d1 (Figure 1B, bottom). This element displays a truncated ORF that codes for only 98 amino acids out of the 560 deduced from the full-length transposon. The entire sequence of SmTRC1d1 aligns with the full-length genomic transposon, but it lacks part of the ORF and its 3' tandem repeat (Figure 1B), hence its short length. The diversity of lower molecular weight bands generated by PCR (Figure 1A) suggests that SmTRC1 copies of several sizes, differently truncated, must exist in the S. mansoni genome.

For one set of PCR amplifications with genomic DNA as template, we used a 16 bp region downstream from the left TIR as forward primer, and a sequence composed of 8 bp overlapping the 3'-end of the right TIR plus a 12 bp region immediately downstream as reverse primer. This primer set permitted a copy of approximately 4.7 kb to be amplified; this copy was cloned and sequenced (Figure 1B, middle). The sequence was named SmTRC1f1; it lacked 71 bp at its 5'-end, including the left TIR, owing to the design of the primer used in the amplification reaction, but otherwise it appears to represent an integral copy of SmTRC1 (Figure 1B, middle). In fact, this clone displays 99.8 % nucleotide identity (only 9 mismatches over 4675 nucleotides) with the element described in Supercontig_0018735, and one base in the left TIR is deleted in both sequences. This suggests that we had cloned from the PCR products a copy representing the sequence contained in Supercontig_0018735; the nine mismatched bases may have arisen from mutations generated either naturally in the field among the copies from individual parasites, or artificially from the Taq polymerase during the PCR amplification step.

A BLASTP search against the nr database at GenBank was performed using the protein encoded by SmTRC1-ORF as query. The highest hits produced were with two proteins of unknown function from Schistosoma japonicum [GenBank:AAW24935.1] (with E-value 10^-20) and Anopheles gambiae [GenBank:EAA01922.3] (with E-value 10^-9). The next best hits (E-value 10^-5) were with the TNPD proteins of CACTA transposons from Oryza sativa. It is worth noting that the region displaying similarity corresponds exactly to the Transposase_21 domain (Pfam 02992). In fact, as described below, global alignment of this region of the SmTRC1-ORF product with the Transposase_21 domains of CACTA transposons indicates several conserved residues.

The SmTRC1-ORF sequence was used as query to perform an additional TBLASTN search directly into several complete animal and fungal genomes using the Genomic Blast tool at NCBI. This search produced hits indicating high similarity (E-value 10^-88 to 10^-30) between the deduced SmTRC1-ORF translated sequence and translated sequences from genomes of such diverse animals as Strongylocentrotus purpuratus, Ciona intestinalis, Danio rerio and Aedes aegypti. In all these cases, practically the whole protein was aligned, not only the Transposase_21 domain. The search against fungal genomes produced hits indicating moderate similarity (E-values 10^-16 to 10^-4) between the deduced SmTRC1-ORF translated sequence and translated sequences from the genomes of e.g. Rhizopus oryzae, Coprinopsis cinerea and Phanerochaete chrysosporium. For several of the above organisms, multiple hits were generated in the TBLASTN searches against the genomic sequence, indicating that more than one copy must be present in their genomes.

As noted earlier, SmTRC1 has a 54 bp TIR sequence at its extremities, which is consistently present in all copies examined. Also, the S. mansoni TIR sequence has an internal repeat of the motif AAAGGGGAAATAAAG. A TRC element of approximately 5.2 kb was detected in the Tribolium castaneum genomic sequence [GenBank:AAJJ01002287.1], and displays at its extremities an inverted repeat of 27 bp with the sequence CCCTAGTAGCACCGAATATTTGTAAAA. In addition, we found a 10 bp inverted repeat (CCCAGTCAAC) flanking a TRC in an A. aegypti genomic segment [GenBank:AAGE02020512] that delimits a 9.2 kb element. Both elements have a 2 bp direct repeat adjacent to each inverted repeat (Figure 2B), suggesting that target site duplication occurred when the mobile elements were inserted, analogous to the situation in the S. mansoni genomic sequences.

It is interesting to note that the first 3 bases in the TIRs from the T. castaneum and A. aegypti elements are identical to those in the TIR of S. mansoni (Figure 2D), suggesting that animal TRCs may display a CCC core sequence in their TIRs analogous to the CACTA core sequence in TIRs from CACTA transposons. Comparison of these animal TIRs with those of the plant CACTA transposons showed a high similarity between the T. castaneum TIR core sequence and the CACTA core sequence of plants, with only one mismatched base (Figure 2D). No such level of similarity is found in S. mansoni or A. aegypti TIR core sequences, which display only 3 and 2 coincident bases, respectively. Owing to the low numbers of bases and examples involved, it is difficult to determine whether the matching bases in T. castaneum are coincidental or reflect an evolutionary process.

We have not been able to characterize the TIRs flanking the ORFs of TRC elements in organisms other than those described above, namely S. mansoni, A. aegypti and T. castaneum. In some animal species, the TIRs may have become unrecognizable because of mutations or have been lost through further recombination. Another possibility is that some of these elements may represent domesticated transposases 23242526, which no longer transpose but perform another cellular function instead. Further analyses of these animal elements are warranted to determine which is the case for each particular element.

Southern blot analysis using a probe from the 5' end region of SmTRC1 detected multiple bands when hybridized to EcoRI-digested genomic DNA, showing the existence of multiple copies (Figure 3). In parallel experiments, we used the same number of radioactive counts and probes of approximately the same size for SmTRC1 and Saci-2, a previously described S. mansoni retrotransposon 18. This allowed us to compare the two hybridization signals directly to estimate the number of SmTRC1 copies. The hybridization image was processed to measure the signal intensity in each lane. For each digestion, we divided the intensity from SmTRC1 by that from Saci-2. The results from both EcoRI and StuI digestions show that SmTRC1 has approximately 1/3 of the number of copies of Saci-2 (estimated at 85–850 by DeMarco et al. 18) (Figure 3). Extrapolation of these data indicates that there are approximately 30–300 copies of SmTRC1 in the S. mansoni genome. Amplification of SmTRC1 genomic DNA by PCR suggested that most of the genomic copies are not integral, implying that there are only a few full-length copies of SmTRC1 in the genome.

Mapping of the S. mansoni ESTs available at GenBank dbEST to the full-length genomic copy of SmTRC1 produced a discontinuous alignment of 64 ESTs (Figure 4B shows some representative examples), indicating that messages transcribed from this transposon are subject to alternative splicing. Moreover, several of the predicted introns contain the canonical GT-AG splicing sites at their extremities (Figure 4B). The diversity of patterns obtained for different ESTs mapping to the same locus indicates that a number of variant messages are produced by alternative splicing.

Intriguingly, no S. japonicum ESTs with homology to SmTRC1 were found by either BLASTN or TBLASTN searches using the full-length S. mansoni element as query against GenBank dbEST. This contrasts with the 70 S. mansoni ESTs (64 with discontinuous and 6 with continuous alignments) that were found with an E-value lower than 10^-10 by a BLASTN search against the same database. This leads us to hypothesize that these elements either have a much lower transcriptional activity in S. japonicum than in S. mansoni or have been eliminated from the S. japonicum genome.

To document the existence of alternatively spliced transcripts in S. mansoni better, we used the sequences at the extremities of many of the 70 ESTs known to align to SmTRC1 to design primers for amplifying additional SmTRC1 transcripts by RT-PCR using mRNA as template. The RT-PCR reaction produced a strong amplification band of approximately 900 bp and sub-products of lower molecular weight, plus a faint amplification band of higher molecular weight, which suggests that full-length transcripts may be expressed at a low level (Figure 4A). No amplification was seen in the control without reverse transcription, indicating that the amplification products were in fact derived from bona fide transcribed messages.

We cloned and sequenced these different RT-PCR products. Mapping the different sequences to the full-length SmTRC1 genomic sequence confirmed a diversity of splicing patterns, as already suggested by mapping of previously existing GenBank ESTs (Figure 4B). Two of the sequenced clones (Clones B1 and B2) are similar in size to the major band from the RT-PCR reaction, indicating that this band may contain more than one product. These two products did not overlap the ORF region, and the intron formed between their exons 1 and 2 does not display the canonical CT-AG splicing sites. However, mapping of these transcripts to the entire S. mansoni genome showed that canonical splicing sites are present at the extremities of the same intron in some truncated forms of the element (data not shown). This indicates that these transcripts probably originated from truncated copies that are apparently transcriptionally active, being responsible for a considerable fraction of the SmTRC1 transcripts detected.

Clone B1 has an ORF of 681 bp that is interrupted by a TGA stop codon at bases 439 to 441. However, alignment with the SmTRC1f1 sequence shows substitution of the TGA by CTA, which codes for isoleucine (Figure 4C). We named this hypothetical protein product SmTRC-PrA (S. mansoni TRC Protein A). In addition, a 4 bp deletion at base 37 of this clone in relation to SmTRC1f1 produces a frame shift at the beginning of the message. This suggests that the transcript was generated from a truncated copy of this element and that degeneration has produced the stop codon interrupting the ORF. Clone B2 has a shorter ORF that codes for a product very similar to the deduced N-terminal amino acids of the ORF product of clone B1; this hypothetical protein product was named SmTRC-PrB.

Although most of the ESTs exhibit a splicing pattern that does not include the SmTRC1-ORF sequence in any exon, the presence of a few ESTs mapped in the region of the SmTRC1-ORF suggests that it is actually transcribed and translated. We were not able to clone the high molecular weight products directly from the first RT-PCR using primers designed from the extremities of EST sequences, but we designed a primer from the 5'-end of the SmTRC1-ORF sequence and used a primer from the 3' extremity of the transcripts. The ensuing RT-PCR resulted in the amplification of a partial transcript of 1.3 kb (data not shown). Cloning and sequencing of this message showed that it was indeed derived from exons mapping to the SmTRC1-ORF as well as from other exons in the 3' region (Figure 4B, clone B4). It codes for an incomplete hypothetical protein product of 125 amino acids named SmTRC-PrC. A longer, complete version of the message encoding a longer SmTRC-PrC protein is expected to exist, including the Transposase_21 domain in its amino-terminal portion. Clone B4 also exhibits a second ORF coding for a protein of 299 amino acids, 175 of which are shared with SmTRC-PrA (76% of the SmTRC-PrA amino acids). In view of this level of conservation, we named this protein SmTRC-PrA2. Characterization of additional clones could eventually identify further alternatively spliced forms of SmTRC1-derived transcripts.

The Spm element of Maize has been shown to display four different transcripts (TnpA-D) generated by alternative splicing, each coding for a different protein product 7. One of these transcripts (TnpD) spans the entire region comprising ORFs 1 and 2 predicted in the spm DNA, which explains the selection pressure that maintains such ORFs in the transposon structure. Similarly, we predict the existence of a spliced transcript spanning all the TRC1-ORF to explain the maintenance of such a large ORF and the conserved amino acid sequence observed on comparison with elements from other animals. The diverse splicing pattern of the transcribed S. mansoni messages results in distinct ORFs coding for different proteins; only the Transposase_21 domain of proteins encoded by these transposons has a detectable similarity with proteins from CACTA transposons. The other portions of the proteins encoded by these alternatively spliced S. mansoni transcripts appear not to be detectably conserved.

Spm TnpA, a short alternatively spliced transcript that lacks the Transposase_21 domain, is apparently more abundantly transcribed in plants than the other, longer transcripts. This reflects the results of RT-PCR experiments suggesting that shorter TRC transcripts are more abundantly transcribed in S. mansoni than the longer alternatively spliced transcript that includes the Transposase_21 domain. Spm TnpD displays two ORFs in tandem, one coding for TNPD and the other for TNPA 8. Likewise, clone B4 also exhibits two ORFs in tandem, one coding for SmTRC-PrC and the other for a SmTRC-PrA-like protein.

Multiple protein sequence alignment (Figure 5) of the Transposase_21 domain (PFAM# PF02992) was performed using sequences from known CACTA transposons from several plants 222728, together with the related domain in the deduced ORFs from S. mansoni and several other metazoan and fungal elements identified in the present work by BLAST analysis (as described above). Although there is a visible divergence between the domains from elements derived from different phyla, conservation of several residues is apparent (Figure 5). It is also interesting to note that the TRC-like sequence derived from the fungus Cryptococcus neoformans has several characteristics that distinguish it from the other fungal sequences, which are apparently very similar to one another.

Three different conserved motifs can be discerned, marked I-III in the aligned proteins (Figure 5). In most of the elements, a first conserved motif I/L/V-X-I/L/V/F-X-I/L/V/F-X₂-D-G-X₃-F/Y-X_7–9-W-P-I/L/V of Transposase_21 domain is present (Figure 5, box I); however, in 3 out of 4 fungal sequences this motif shows an interchange between tryptophan and glycine residues. This suggests that these two residues have an important role and that only the position of one residue relative to the other is essential for activity of this protein. In the second conserved domain (Figure 5, box II), the fungal proteins show a level of conservation comparable to Magnoliophyta and higher than Eumetazoa. Thus, only proteins from Fungi and Magnoliophyta have a proline immediately after the conserved glycine residue (Figure 5, box II), and an additional P-L/I conserved motif in the middle of this domain; both are absent from Eumetazoa. Interestingly, we identified a L/V/I-D-X-L/M-H-X₃-L-G motif in the Transposase_21 domain that is present in the eumetazoan TRC elements (Figure 5, box A) and also in 3 out of the 4 fungal TRC proteins, but is absent in proteins from Magnoliophyta.

Among the conserved residues there are two aspartyl residues (DD), one in the first and one in the third conserved domain, separated by 80 residues; this is very similar to the distance between the two aspartyl residues in the DDE motifs in Tn3 transposases 29. There is also a glutamyl residue (E) in the conserved domain 2 in all but three sequences, two of which have this residue in adjacent positions and one in a position 2 residues away. It is possible that conservation of such an amino acid triad reflects a similar catalytic mechanism in the DDE motif despite the different arrangement of residues. In this case its function would be similar to that described for the DDE motif, which is presumed to coordinate divalent metal ions to promote catalysis of DNA cleavage and ligation 3. Moreover, there is a conserved CXXC motif in the conserved domain 3 that is identical to the configuration of cysteines in the zinc-finger-like motifs (HHCC domains) of retroviral integrases, suggesting that TRC cysteines may also be involved in DNA binding.

A phylogenetic tree (Figure 6) was generated from the three conserved regions (I-III) shown in the alignment of Figure 5. Although the analysis does not permit a clear inference of phylogeny within the Eumetazoan elements, it clearly shows the separation of the Eumetazoa and Plant branches (Figure 6). Interestingly, three of the Fungi elements appear to be distinct from the others, but one of them, the C. neoformans element, appears to be at a basal position in relation to plant transposons. With the exception of the latter, the analysis clearly shows the separation between Plants, Fungi and Eumetazoa.

The BH isolates of S. mansoni were maintained in the laboratory by routine passage through mice and snails. Adult parasites were obtained by portal perfusion of hamsters 7 to 8 weeks after infection. Tissues were conserved in RNALater (Ambion) according to the manufacturer's instructions. Tissue mRNAs were extracted using MAC isolation kits (Miltenyi Biotec). mRNA samples were treated with RQ1 RNase-free DNase (1 U/10 μl; Promega) for 30 minutes at 37°C. cDNAs were prepared using the Superscript II first strand synthesis system for RT-PCR (Invitrogen), following the manufacturer's instructions. A parallel control reaction was performed without the addition of reverse transcriptase and used as template for a PCR reaction to detect any genomic DNA contamination. The PCR step was performed using Advantage II polymerase (Clontech) with the buffer supplied by the manufacturer, 200 μM dNTPs, and 200 nM of each primer using the following program: 95°C (1 min); 35 cycles of 95°C (30 s), 55°C (30 s) and 68°C (4 min); and final extension at 68°C (3 min). The reaction products were cloned into pGem-T vector (Promega) and sequenced.

Genomic DNA was extracted from 800 mg of adult worms (approximately 3,000 worms) using the protocol described by Ausubel et al. 35. PCR reactions were performed using the same conditions as described in the previous paragraph.

Southern blot experiments for estimating the number of copies of SmTRC1 were performed according to the protocol described by DeMarco et al. 18, except that StuI was used in place of BamHI and that Saci-2 was used as a benchmark. The total signal intensity for each lane was calculated using ImageQuant v5.1 (Molecular Dynamics), with a fixed area rectangle to delimit the area for each sample analyzed.

Using the deduced protein sequence of the Transposase_21 domain from SmTRC1, we retrieved several other sequences that showed significant similarity (E-values less than 10^-4) in a tBLASTn search against genomes of Metazoa and Fungi at NCBI. We used these sequences along with the sequences of known CACTA transposons from Magnoliophyta to perform an alignment using ClustalX v1.83 36. Alignments were imported to the GeneDoc program V2.6.002 for shading of conserved residues. Further analysis with Clustal X and the neighbor-joining method, using the three conserved regions indicated by boxes I to III in Figure 5 and excluding positions with gaps, resulted in the phylogenetic tree shown in Figure 6. The confidence of the branches was evaluated by bootstrap analysis using 1,000 samplings. Phylogenetic trees were drawn using Treeview (version 1.6.6) 37. The GenBank sequences and accession numbers utilized for construction of alignments and phylogenetic trees are as follows: (1) Transposase family tnp2 members – Arabdopsis thaliana, [GenBank:CAB80813.1]; Brassica rapa, [GenBank:BAA85462]; Daucus carota, [GenBank:BAA20532.1], Oryza sativa, [GenBank:DAA02106.1]; Zea mays, [GenBank:AAA66266.1]; Ipomoea trifida, [GenBank:AAS79612.1]; (2) Whole genome DNA sequences – Aedes aegypti, [GenBank:AAGE02020512]; Danio rerio, [GenBank:CAAK02057130.1]; Strongylocentrotus purpuratus, [GenBank:AAGJ01221697.1]; Ciona intestinalis, [GenBank:AABS01001120.1]; Tribolium castaneum, [GenBank:AAJJ01002287.1]; Drosophila pseudoobscura, [GenBank:AADE01004520.1]; Glomus intraradices, [GenBank:AC156590]; Phakopsora pachyrhizi, [GenBank:AC149399]; Cryptococcus neoformans, [GenBank:EAL18770.1]; Rhizopus oryzae, [GenBank:AACW02000214.1].

We have deposited all sequences obtained in this work at EMBL under the following numbers: SmTRC1f1, [EMBL:AM268206]; SmTRC1d1, [EMBL:AM268205]; SmTRC-PrA, [EMBL:AM268207]; SmTRC-PrB, [EMBL:AM268208]; SmTRC-PrC, [EMBL:AM268209]. We have also deposited Third Party Annotations (TPAs) at EMBL for Transposase family Tnp2 members, under the following TPA numbers: Aedes aegypti, [EMBL:BN000947]; Danio rerio, [EMBL:BN000951]; Strongylocentrotus purpuratus, [EMBL:BN000955]; Ciona intestinalis, [EMBL:BN000948]; Tribolium castaneum, [EMBL:BN000946]; Drosophila pseudoobscura, [EMBL:BN000950]; Glomus intraradices, [EMBL:BN000952]; Phakopsora pachyrhizi, [EMBL:BN000953]; Rhizopus oryzae, [EMBL:BN000954]; Cryptococcus neoformans, [EMBL:BN000949].