Among genes in Group 2, one of the most interesting is a S. mansoni INSIG ortholog (SmINSIG), an important regulator of the mevalonate synthesis pathway.

INSIG-1 in cultured mammalian CHO cells has been shown to play an essential role in degradation of HMG-CoA reductase (a critical enzyme in the mevalonate pathway) 23. In S. mansoni, HMG-CoA reductase is vital for parasite survival and plays a physiological role in regulating egg production 2425. Egg deposition is a characteristic of platyhelminths, and in the case of S. mansoni the parasite's eggs deposited in the host circulatory system are the major cause of morbidity in the host's liver.

The full-length message of SmINSIG was obtained by RT-PCR. The reverse primer for RT-PCR was designed from the 3' end of a consensus sequence obtained by the S. mansoni Assembled EST designated SmAE C601385.1 and annotated as INSIG 9. The forward primer was designed from a genomic region 119 bp upstream from the locus where the 5' end of SmAE C601385.1 is mapped; the genomic sequence can be found in Supercontig_0000071 of the S. mansoni genome draft sequence 11 publicly available at the Wellcome Trust Sanger Institute 26. After cloning and sequencing the RT-PCR product we obtained the full-length SmINSIG message that contains 857 bases and encodes a protein of 244 amino acids. The deduced protein displays the described conserved domain of INSIG proteins (IPR009904) and the characteristic six transmembrane regions.

Figure 2A illustrates a multiple alignment of INSIGs from different deuterostomes and SmINSIG, where the conserved domain is shown and the six transmembrane regions are marked. BLASTP search against the nr database at GenBank using SmINSIG as query resulted in a best match to the zebrafish (Danio rerio) INSIG-1 ortholog ([GenBank: AAH45341.1]) with 49% identity and 68% similarity over 144 amino acids. A Maximum Likelihood tree was constructed (refer to the Methods section for details) (Figure 2B), and suggests that SmINSIG diverged before the gene duplication event responsible for emergence of the two vertebrate paralogs in zebrafish.

Real-time PCR experiments with tubulin as an internal standard showed a higher expression in egg, miracidium and cercarial stages when compared to schistosomulum (p < 0.05). The free-living forms (cercaria and miracidium) exhibit higher expression when compared to the forms that live inside the vertebrate host (schistosomulum and adult) (p < 0.05) (Figure 3A).

Despite previous studies showing induction of INSIG gene expression by insulin in rat regenerating liver 27 and the presence of genes encoding proteins with similarity to the human insulin receptor in schistosomes 928, we found that exposure of S. mansoni adult worms to exogenous recombinant human insulin peptide for 1 or 2 hours in vitro did not affect the expression of SmINSIG measured by Real-time RT-PCR while affecting the expression of a number of non-related genes (data not shown).

This work is the first published report of an INSIG ortholog in a protostome invertebrate. We have searched the S. mansoni and S. japonicum transcriptome datasets for evidence of the genes described in vertebrates as targets of INSIG, and also for the presence of homologs in the mevalonate pathway. They are listed in Table 2 and are discussed in detail below.

Schistosomes are described to be cholesterol auxotrophs (unable to synthesize cholesterol) 29, suggesting that SmINSIG might perform in S. mansoni some of the functions described in mammals for INSIGs, especially in the mevalonate synthesis pathway (a precursor of cholesterol as well as of non-sterol isoprenoids) and not in the later steps of cholesterol synthesis. Accordingly, gene sequences potentially encoding all enzymes of the mevalonate pathway described in the MetaCyc database 30 were found either in S. mansoni or in S. japonicum public datasets, indicating that this pathway is intact in schistosomes (Table 2).

Interestingly, INSIG-1 has been shown to play an essential role in sterol-mediated degradation of HMG-CoA reductase (a critical enzyme in the mevalonate pathway) in cultured mammalian CHO cells 23. Knockout mice have been generated in which the INSIG-1 and INSIG-2 genes are disrupted in the liver through recombination 31. These mice have uncontrolled synthesis of fatty acids, cholesterol and a marked increase in HMG-CoA reductase protein, resulting from a decreased degradation of the reductase in the absence of INSIGs 31. HMG-CoA reductase has an ortholog already characterized in S. mansoni 32 and its inhibition compromises egg production, as well as the survival of early stages of the parasite 24. In addition, HMG-CoA reductase inhibition in schistosomes blocks egg production, which is induced by cholesterol precursors, such as mevalonate and farnesol; these precursors can also reverse the mevinolin-induced inhibition of egg production 25. When expression was compared between male and female adult worms, a 10-fold higher expression of SmINSIG was found in females (Figure 3B). Egg production in female worms is a very efficient process and a putative negative control exerted by SmINSIG could be part of a modulatory mechanism of mevalonate synthesis in order to tune it to the rate of egg production. Miracidia and cercariae (larval forms) are short-lived and highly specialized forms, and the high levels of SmINSIG (Figure 3A) and the resulting inhibition of mevalonate synthesis in those stages when compared to adults would be consistent with a more complex physiology of the latter.

A recent report showed that in mammalian cells a fraction of INSIG-1 molecules are bound constitutively to a protein named gp78 or autocrine motility factor receptor (also present in schistosomes, see Table 2) through binding of transmembrane domains of the two proteins 33. Evidence indicates that this sterol-triggered reductase/INSIG-1/gp78 complex is essential for ubiquitination and degradation of the reductase 33. Based on the presence of orthologs in the S. mansoni transcriptome database, this evolutionarily conserved complex could also occur in schistosomes.

In line with the known S. mansoni auxotrophy for cholesterol, we could not identify either in the S. mansoni or in the S. japonicum transcriptomes any of the enzymes specifically involved in cholesterol synthesis from mevalonate. Interestingly, we found a S. mansoni EST with similarity to a short partial segment of M. musculus SCAP ortholog (Table 2); SCAP has been described in mammals as a

S

c

a

p

Interfering with the mevalonate pathway in schistosomes could be a promising novel drug target approach because of the known adverse clinical consequences of eggs to the patients and of the biological importance of egg deposition in the parasite's life cycle.

Another gene from Group 2 that we have selected for further scrutiny is represented by S. mansoni Assembled EST (SmAE) C605907.1. Its consensus sequence had good (56%) similarity to vasohibin-like proteins (recently named vasohibin 2).

Vasohibins constitute a recently described family of endothelium-derived angiogenesis inhibitors in humans 3536. The presence of a Vasohibin ortholog in schistosomes might suggest a potential angiogenesis inhibition process in the host, mediated by schistosomes' molecules. Complete sequencing of clone MA3-9999U-M294-F04-U.B, the longest clone in SmAE cluster C605907.1, showed a transcript with 975 bases (Figure 4A) that we named SmVASL for

S

Vas

l

To search for possible additional portions of SmVASL, the sequence of the original clone was mapped to the draft sequence of the S. mansoni genome 11. We found that SmVASL maps with splicing to Supercontig _0000046, from base 1,255,010 to 1,259,701 (Figure 4A). In a genomic region of ~2,000 bases upstream from the SmVASL gene we found a sequence apparently encoding a peptide that displays similarity to vasohibin-like proteins (not shown). Next, we designed a forward and a reverse primer based on the sequence of this region, as indicated in Figure 4A (green arrows). These primers were used in a RT-PCR reaction, which resulted in a complex profile of amplification with one band of ~650 bp and a smear between 750–900 bp (Figure 4C). No amplicon was obtained in a parallel reaction without reverse transcription, to control for the absence of DNA contamination, indicating that the diverse range of products (Figure 4C) was derived from bona fide mRNA messages. This result suggested multiple alternatively spliced forms at the 5' end of full-length SmVASL.

The RT-PCR amplification products were cloned and sequenced. A total of approximately 300 clones were sequenced from both ends, disclosing fourteen different isoforms (SmVASLv1-13, with two SmVASLv6 isoforms). Mapping of these isoforms to the S. mansoni genomic sequence confirmed that the different clones are a result of alternative splicing (Figure 4A). Variation is caused by a number of alternative splicing events, such as 5'-deletion, exon skipping, and junction of two exons by intron retention. Two different lengths for the second exon were observed: exon 2 with 79 bp for variants SmVASLv1, 2, 6a, 6b, 7, 10, 12 and 13; and exon 8 with 65 bp, resulting from a 5'-deletion in exon 2, for SmVASLv3, 4, 5, 8, 9 and 11 (Figure 4A). The latter splicing form caused an early stop, and the resulting isoforms encode a putative short 57 amino acid protein. The isoforms have different 3'-UTR ends, which may be related to different stability of the messages.

Exon skipping was observed in SmVASLv5, 7, 8, 11 and 12 (Figure 4A). Intron retention was detected in isoforms SmVASLv1, 2, 4, 12 and 13, where exon 4 is the junction of exons 10 and 11 (that are present for example in isoform 5). Intron retention was also detected in SmVASLv1, 4, 7, 9 and 11, where exon 3 is the junction of exons 6 and 7. Finally, the intron retention observed in SmVASLv3 results from junction of exons 6, 7, 10 and 11.

We have identified seven different polypeptides that are encoded by these variants, representing different deduced protein isoforms (Figure 4A). As noted earlier, isoforms SmVASLv3, 4, 5, 8, 9 and 11 encode the same 57 amino acid protein. Due to the absence of a stop codon in the SmVASLv6 sequence in the segment amplified by RT-PCR, a 3'-RACE experiment was conducted to extend SmVASLv6 sequence, which resulted in two additional longer SmVASLv6 isoforms, SmVASLv6a and SmVASLv6b. The longer isoform, SmVASLv6a encodes a 337 amino acid protein, the longest protein isoform detected here. When compared through BLASTP with the Swiss-Prot/TrEMBL database, the highest match was a protein of S. japonicum of unknown function ([Swissprot: Q5DB91]). The second best match was to human vasohibin-like protein (vasohibin 2, [Swissprot: Q86V25]) (Figure 4B) to which SmVASLv6a protein aligns with 37% identity and 57% similarity over 252 amino acids (71% coverage of the human protein). Alignment of SmVASLv6a and human vasohibin 2 reveals a C-terminal divergence between these two proteins (Figure 4B). 5'-RACE experiments under several conditions were performed, without success. A multiple sequence alignment and a phylogenetic tree of SmVASLv6a and several other orthologs are represented in Figure 5A and 5B, respectively.

Interestingly, several alternatively spliced isoforms are predicted for human vasohibin 2 ([Swissprot: Q86V25]), some of them encoding a shorter protein that lacks the carboxyl-terminal end. Similarly, all SmVASL but SmVASLv6a and SmVASL6b isoforms encode shorter proteins, lacking different portions of the carboxyl-terminal end. Absence of the C-terminal end has been also described for one of the isoforms of vasohibin 1, where the shorter isoform abolishes the anti-angiogenic activity of the full-length protein 35. Conservation of such a mechanism for generation of short isoforms lacking the carboxyl-terminal region of the protein in distantly related species such as H. sapiens and S. mansoni indicates that the C-terminal portion of vasohibin should represent a conserved functional domain.

Real-time PCR experiments using primers designed to detect most of the isoforms (except isoform 3), shows that cercaria and adult have the highest SmVASL expression levels, whereas schistosomulum has the lowest SmVASL levels (Figure 5C). Considering the angiogenesis control role exerted by vasohibin in vertebrates and the high conservation level between SmVASL and human vasohibin, together with vascular location of adult schistosomes, it is tempting to hypothesize a role of SmVASL in modulating human angiogenesis. Considering the presence of a vasohibin homolog in planarian (Additional file 2) and the high SmVASL expression in miracidium, the S. mansoni life form that invades the snail (an invertebrate host that has an open circulatory system), we speculate that this protein could perform a different (possibly endogenous) role in this life stage. Loeffler et al. 37 reported that schistosomes exert a positive effect on angiogenesis because soluble egg antigen (SEA) induces angiogenesis-related processes by up-regulating VEGF in human endothelial cells 37. It is hypothesized that neovascularization in the schistosome granuloma may be necessary to maintain oxygen and nutrient levels, in a similar way as in rapidly growing tumors 3738. Negative regulation of host angiogenesis may be provided by the S. mansoni well conserved vasohibin ortholog identified here, which would counterbalance the previously reported positive effects of S. mansoni on angiogenesis 37, thus helping to maintain host hemostasis. Functional experiments are warranted to determine the biological roles of the SmVASL isoforms reported here.

The third and final gene analyzed in the present study encodes an Interferon Regulatory Factor (IRF) protein (represented by SmAE C603512.1 and named SmIRF).

IRFs are important molecules involved in immune response and a schistosome ortholog could be related to the recognition and response to host immune processes. Primers were designed from the extremities of this assembled sequence and a single 1330 bp amplicon was generated by RT-PCR; cloning and sequencing confirmed its identity. In addition, a 3'-RACE with primers designed from SmAE C603512.1 permitted cloning and sequencing of the 3' end of this mRNA. SmIRF full-length sequence has 2297 bp and encodes a 476 amino acid protein that shares 27% identity and 44% similarity over 434 amino acids with the interferon regulatory factor 4 of Gallus gallus ([GenBank: AAK08198]) [see Additional file 3]. A multiple sequence alignment and a phylogenetic tree of SmIRF and several other orthologs are represented in Figure 6A and 6B, respectively. Each IRF contains a well-conserved DNA-binding domain with ~120 amino acids at the amino terminus that folds into a helix-turn-helix DNA-binding motif (Smart SM00348). The C-terminal end of IRFs is generally more variable among family members 39 and contains a SMAD/FHA domain, commonly found in transcription factors and responsible for interaction with other phosphorylated molecules 40. In silico analysis of SmIRF using NetPhos 41 revealed the presence of several serine phosphorylation sites [Additional file 3]; in mammalians, IRF phosphorylation is important for its function and regulation 42.

IRFs constitute a complex family of transcription factors with broad functionality in mammalians, which was recently reviewed 43. Honda and Taniguchi 43 pointed to several mammalian genes regulated by IRFs and we have searched for orthologs of these targets in S. mansoni. Two genes regulated by IRFs (TAP1 and LMP2), involved in processing and transport of peptides appear to have homologs in schistosomes. Using human TAP1 gene encoded protein as query we found in S. mansoni an ABC transporter gene, SMDR1 [GenBank: AAA66476.1] (35% identity and 61% coverage) that has been previously described as possible transporter of peptides 44. When using human LMP2 ([GenBank: CAA60784]) as query, we found a schistosome ortholog in our database (C602473.1, with 55% identity and 88% coverage). Human LMP2 encodes a proteasome subunit and replaces the Y subunit in the proteasome structure upon IFN signaling and this change results in a different proteasome proteolytic activity 45. Functional proteasomes were recently shown to be essential for schistosome development in the vertebrate host 46. IRFs also regulate caspase-1 expression 43. We found a caspase in S. mansoni that has higher similarity with caspase-3 than with caspase-1 (44% identity and 85% coverage); in mammalians both caspases are involved with apoptosis control; however, the apoptosis pathway appears to be incomplete in the parasite 9.

Besides the inherent capacity of IRFs to function as transcription factors, they selectively bind to a group of proteins, namely immunophilins. Interaction between IRF-4 and immunophilin FKBP52 in mammalians results in a conformational change, abolishing DNA-binding activity and partial IRF-4 proteolysis 47. One immunophilin named p50 [GenBank: AAA69867.1] was already identified in schistosomes, having a high similarity with vertebrate FKBP immunophilins 4849.

The IRF interactions are highly complex and not fully elucidated, even in mammalians, with new interactions and functionalities being recently reported 43. In this work we have pointed to the existence in schistosomes of some possible players. Other experiments (e.g. protein-protein interaction) are warranted to identify functions and other possible SmIRF interaction partners. It would be interesting to demonstrate if SmIRF is involved in schistosome stress responses (viral infections, for example), or in host-parasite interaction, being part of the pathway that interplays with host immune molecules.

S. mansoni EST sequences generated by our group 9 were used as the query dataset of S. mansoni sequences in this work. Local BLAST 51 databases were formatted with all the sequences available from arthropods, nematodes and deuterostomes in the non redundant (nr) nucleotide section of GenBank database (updated as of May/2007) and were used as subject in BLAST searches. In addition, all the publicly available nucleotide sequences from non model arthropods, nematodes and from the planarian S. mediterranea were used. The planarian dataset we used is composed of all the publicly available 171,483 nucleotide sequences (97,901 CoreNucleotide and 73,582 ESTs, on May 2007) together with the assembled ESTs obtained by a large scale EST sequencing project 12. The planarian dataset was further supplemented with the EST assembly data provided by the authors (10,485 assembled sequences; 6,488 contigs and 3,997 singlets) 12. In order to reduce the EST sequence redundancy and consequently the time complexity of the process, the sequences were processed using CD-HIT 52 prior to database formatting. After identifying the S. mansoni genes of interest, we have considered their downstream effectors described in the literature for vertebrates and we searched for the presence of orthologs which could act as molecular partners or members of downstream pathways in S. mansoni. When any of these orthologs were missing in the S. mansoni GenBank database the S. japonicum sequences were alternatively searched.

The S. mansoni genome sequence partial assembly (v. 3) at the Wellcome Trust Sanger Institute 26 was used.

We used the BLASTX and TBLASTX programs 51 to search for S. mansoni (a platyhelminth) translated sequences that have considerable similarity to proteins and translated ESTs from deuterostomes, nematodes and arthropods. Perl scripts together with the Zerg parser 53 and BioPerl 54 were used to perform and process BLAST searches.

BLAST alignments with bit scores higher than 100 and lower than 50 were considered as true matches and non-significant hits, respectively. Bit scores were used to facilitate cross-database comparisons. Hits with intermediary bit scores (between 50 and 100) were manually inspected, basically to eliminate hits derived from unmasked low complexity regions where matches occur through a single amino acid at repeated intervals. Overall, the sequences included in Additional File 2 have scores higher than 80.

The S. mansoni genes without matches and Schistosoma specific genes were filtered out. The resulting genes were further assigned to one of the following groups (Table 1 and Additional file 1) when a given gene was detected in at least one species belonging to the corresponding clade: (1) present in S. mansoni, deuterostomes, arthropods and nematodes (all groups); (2) present in S. mansoni and deuterostomes, but not in arthropods and nematodes; (3) present in S. mansoni and arthropods, but not in deuterostomes and nematodes; (4) present in S. mansoni and nematodes, but not in arthropods and deuterostomes); (5) present in S. mansoni, arthropods and nematodes, but not in deuterostomes; (6) present in S. mansoni, arthropods and deuterostomes, but not in nematodes; (7) present in S. mansoni, nematodes and deuterostomes, but not in arthropods.

To evaluate the significance of the gene gains/losses under each hypothesis, we have generated 100,000 bootstrapped samples and performed the Wilcoxon test. Simulations and the significance test were performed in the R environment.

TMHMM 55 and MINNOU 56 were used to predict transmembrane domains, NetPhos 41 to identify possible phosphorylation sites, with SignalP 3.0 57 to predict signal peptides, and with InterProScan 58 to predict conserved domains.

Multiple Sequence Alignments in the paper were generated using MUSCLE 59 and edited using JalView 60. Curated alignments were then used to generate Maximum Likelihood phylogenetic trees with PhyML package 61. Significance of the results was estimated by building 1000 bootstrapped trees. The NEWICK files generated by PhyML were then displayed with MEGA 62 producing the trees displayed in the results section.

mRNA was obtained from adult parasites conserved in RNALater (Ambion, Austin, TX, USA) by extraction of tissue with MACs mRNA isolation kits (Miltenyi Biotec, Bergisch Gladbach, Germany). 200 ng of mRNA were treated with 5 U of RQ1 RNAse-free DNAse (Promega, Madison, WI, USA) for 30 min at 37°C. Reverse transcription was performed with oligo dT primers using the protocol of Superscript first strand system for RT-PCR (Invitrogen, Carlsbad, CA, USA). PCR was performed with Advantage II (BD Biosciences, Palo Alto, CA, USA) with the buffer supplied by the manufacturer, and 200 nM of each specific primer using the following cycling program: 95°C for 1 min plus 35 cycles each at 95°C for 30 s, 55°C for 30 s, and 68°C for 3 min, followed by a final extension at 68°C for 3 min. The products were analyzed in 1.2% agarose gel and cloned in pGEM-T vector (Promega) for further sequencing. Primers are listed in Additional file 4.

mRNA was obtained from adult parasites conserved in RNALater (Ambion) by extraction of tissue with MACs mRNA isolation kits (Miltenyi Biotec). 200 ng of mRNA was used for reverse transcription using the protocol of the 3'RACE system kit for rapid amplification of cDNA ends (Invitrogen) and specific primers. To perform 3' RACE of vasohibin gene, 3 μg of total RNA from miracidium was used. Reverse transcription was performed using 1 μl of Super Scrip III (Invitrogen) at 65°C for 5 min, 55°C for 50 min and 85°C for 5 min.

PCR reaction was performed with Advantage II polymerase (BD Biosciences) with buffer supplied by the manufacturer, 200 μM dNTPs and 200 nM of each primer using the following cycling program: 95°C for 1 min plus 35 cycles each at 95°C for 30 s, 55°C for 30 s, 68°C for 3 min, followed by a final extension at 68°C for 3 min. The products were analyzed in 1.2% agarose gel and cloned in pGem-T vector (Promega) for further sequencing. Primers are listed in Additional file 4.

mRNA was obtained from male or female adult parasites conserved in RNALater (Ambion) by extraction of tissue with MACs mRNA isolation kits (Miltenyi Biotec). 200 ng of mRNA were treated with RQ1 RNAse-free DNAse (Promega), using 1 U in 10 μl of reaction, for 30 min at 37°C. The DNAse was inactivated at 65°C for 10 min.

Three micrograms of total RNA from each of five stages was treated with RQ1 RNAse-free DNAse (Promega), using 4 U in 10 μl of reaction, for 1 hour at 37°C. The resulting products were reverse transcribed with random hexamer primers using the protocol of Superscript first strand system for RT-PCR (Invitrogen). Control reactions without addition of reverse transcriptase were run in parallel, and were used as templates for PCR negative control, to control for the absence of possible DNA contaminants. Primers for real-time RT-PCR (listed in Additional file 4) were designed using the Primer Express program version 2.0.0 (Applied Biosystems, Foster City, CA, USA) with default parameters. Real-time RT-PCR reactions were performed using SYBR Green PCR master mix (Applied Biosystems) and the specific primers in a GenAmp 5700 Sequence Detection System (Applied Biosystems). P-value was calculated using Student's t-test with 95% confidence interval.

All sequences determined in this work were deposited at EMBL under the following numbers: SmINSIG, [EMBL: AM493258]; SmIRF, [EMBL: AM493259]; SmVASL isoforms, [EMBL: AM493260 – AM493273].