Thyroid hormone receptor orthologues from invertebrate species with emphasis on Schistosoma mansoni

Background: Thyroid hormone receptors (TRs) function as molecular switches in response to thyroid hormone to regulate gene transcription. TRs were previously believed to be present only in chordates. Results: We isolated two TR genes from the Schistosoma mansoni and identified TR orthologues from other invertebrates: the platyhelminths, S. japonium and Schmidtea mediterranea, the mollusc, Lottia gigantean and the arthropod Daphnia pulex. Phylogenetic analysis of the DNA binding domain and/or ligand binding domain shows that invertebrate and vertebrate TRs cluster together, TRs from the vertebrates and from the jawless vertebrate (lamprey) clustered within separate subgroups, Platyhelminth TRs cluster outside of the vertebrate TR subgroups and that the schistosome TRs and S. mediterranea TRs clustered within separate subgroups. Alignment of the C-terminus of the A/B domain revealed a conserved TR-specific motif, termed TR 'N-terminus signature sequence', with a consensus sequence of (G/P)YIPSY(M/L)XXXGPE(D/E)X. Heterodimer formation between S. mansoni TRs and SmRXR1 suggests that the invertebrate TR protein gained the ability to form a heterodimer with RXR. ESMA analysis showed that SmTRα could bind to a conserved DNA core motif as a monomer or homodimer. Conclusion: Vertebrate TR genes originated from a common ancestor of the Bilateria. TR genes underwent duplication independently in the Protostomia and Deuterostomia. The duplication of TRs in deuterostomes occurred after the split of jawless and jawed vertebrates. In protostomes, TR genes underwent duplication in Platyhelminths, occurring independently in trematode and turbellarian lineages. Using S. mansoni TRs as an example, invertebrate TRs exhibited the ability to form a dimer with RXR prior to the emergence of the vertebrate TRs and were able to bind to vertebrate TR core DNA elements as a monomer or homodimer.


Background
Thyroid hormones (TH) play important roles in growth, development and metabolism in vertebrates. TH is synthesized in the thyroid gland under the control of thyroidstimulating hormone (TSH) secreted by the pituitary. TSH secretion is controlled by thyrotropin-releasing hormone (TRH) which is secreted from the hypothalamus. THs are lipophilic molecules able to passively cross the membrane and bind to its receptor, the thyroid hormone receptor (TR). TRs belong to a superfamily of transcription factors called nuclear receptor (NR) superfamily, based on protein sequence similarities, structural motifs and functionality [1]. TRs function as a molecular switch in response to the thyroid hormones T3 or T4 to activate or repress gene transcription depending on the promoter context and thyroid hormone binding status [2]. The typical nuclear receptor contains an N-terminal A/B domain, a conserved C domain (DNA binding domain, DBD), a D domain (hinge region) and a moderately conserved E domain (ligand binding domain, LBD). The most conserved DBD contains two zinc finger motifs (CI and CII). Like all NRs, TRs regulate transcription through its binding to the promoter region of a target gene by the DBD and they activate or repress mRNA synthesis through co-regulators bound to the LBD [1]. The specific target DNA sequence to which NRs bind is called a hormone response element (HRE). The typical HRE is a direct, inverted or everted repeat or palindrome of the DNA sequence AGGTCA. TRs can bind to the HRE as a monomer, a homodimer or as a heterodimer with RXR, another member of nuclear receptor superfamily which contributes to the specificity of the TR. TH binds to the LBD of TR which results in a conformational change in the C-terminus of the receptor. Corepressors then dissociated from the TR allowing coactivators to bind to the C-terminus of the TR in a hormone-dependent manner. TR and the coactivator complex activate the expression of the target gene [3,4].
TR was previously believed to be an innovation of chordates as the genomes of insects (Drosophila and mosquito) and nematodes (Caenorhabditis elegans and C. briggus) do not contain TR genes [5][6][7][8]. Recently, we identified two thyroid receptor homologues in the flatworm Schistosoma mansoni [9], one of which was found in the S. mansoni EST database [6,10]. The presence of TR homologues in S. mansoni demonstrated that the TR orthologue genes are present outside of chordates. However, it is still unclear whether these prostostome TRs possess the same functional domains as in vertebrate TRs. Another question is whether the TR orthologue is present in other invertebrates or just in the platyhelminth lineage? Answers to these questions will help to understand the origin of TR genes and evolution of the function of vertebrate thyroid hormone network. To begin to address these questions, we isolated cDNAs of S. mansoni TRs and demonstrated that TRs in platyhelminths are highly conserved not only in sequence similarity, but also in gene organization, protein-protein interaction and in DNA-binding ability. Furthermore, we mined the available genome data and demonstrated that TR orthologues are present in different invertebrate animals but not in Porifera or Cnidaria. Phylogenetic analysis showed that the TR orthologue likely originated from a common ancestor of the Bilateria.

TR orthologue genes in invertebrate animals
By an extensive search of available databases, predicted genes encoding TR orthologues were found in different invertebrate animals using the conserved DBD as a query. They include two genes in each of the platyhelminth species evaluated, the turbellarian Schmidtea mediterranea and the trematodes, Schistosoma mansoni and S. japonium, one each from the mollusc Lottia gigantean (owl limpet) and the crustacean Daphnia pulex (water flea) ( Fig. 1 and see additional file 1). No TR orthologues were identified in any species of the Radiata.

Sequence alignment analysis
Alignment of DBD amino acid sequences show that the invertebrate TR orthologues are highly conserved (Fig. 1). All TRs (from chordate and non-chordate species) possess an identical P-box sequence, which determines the specificity of target DNA binding. The methionine at position 16 of the N-terminus of the DBD (highlighted in red in Fig. 1) was non-chordate TR-specific. Instead at this position an isoleucine is conserved in all chordate TRs examined to date. Two amino acids (His and Pro) at positions 33-34 of N-terminal of DBD were chordate TR-specific (highlighted blue in Fig. 1). The second zinc finger is less conserved in non-chordate TR orthologues compared to the first zinc finger, but still shows a high degree of conservation (Fig. 1). The T-box and A-box which are located in the C-terminal extension (CTE) of the DBD are also highly conserved in non-chordate TR orthologues. They show 75-83% and 57-100% identity to the chordate Tbox and A-box, respectively. These regions in the water flea TR orthologue are highly divergent, probably due to a recombination event that happened in this region (see below). An aspartic acid in the vertebrate TRs at the 8th position of N-terminus of the T-box is highly conserved for jawed vertebrate TRs, a glutamic acid at this position is conserved in other TRs. Although the N-terminal A/B domain of nuclear receptors is divergent, there is a highly conserved motif found at the 3' end of the TR A/B domain, we termed the 'N-terminal signature sequence' (NTSS). The TR NTSS is a consensus sequence of (G/P)YIPSY(M/ L)XXXGPE(D/E) (Fig. 1). In non-chordate TR orthologues, the sequence is PYIPSYMXXXGPEEP; while in chordates, three amino acids are deleted generating a sequence of GYIPYL(D/E)KE(P/Q/L). The deleted three Sequence alignment of the DNA binding domain (DBD) of TRs Figure 1 Sequence alignment of the DNA binding domain (DBD) of TRs. There is a highly conserved motif found at the 3' end of A/B domain, termed the 'N-terminal signature sequence' (NTSS). In non-chordate TR homologues, it is PYIPSYMXXXG-PEEP; while in chordate NTSS, three amino acids are deleted (represented by dashes) generating a sequence of GYIPYL(D/ E)KE(P/Q/L). In the N-terminus of the DBD, a methionine at position 16 (highlighted in red) is non-chordate TR specific. In this position an isoleucine is conserved for all chordate TRs. Two amino acids (His and Pro for all species except Ciona) are inserted at positions 33-34 of the N-terminus of DBD in chordate TRs (highlighted in blue). An aspartic acid in vertebrate TRs at the 8th position of the T-box is highly conserved for vertebrate TRs, a glutamic acid at this position is conserved for other TRs (highlighted in red). Stars identify the conserved cysteine residues that comprise the zinc finger of the DBD. The conserved residues are highlighted in green. CiTR: Ciona intestinalis nuclear receptor 1, DpTR: water flea Daphnia pulex TR, hTRa: Human thyroid receptor alpha, hTRb: Human thyroid receptor beta, LgTR: owl limpet Lottia gigantean TR, SeTRa: turbellarian Schmidtea mediterranea TRα, SeTRb: S. mediterranea TRβ, SjTRα: blood fluke Schistosoma japonium TR alpha, SjTRβ: S. japonium TR beta, SmTRα: blood fluke S. mansoni TR alpha, SmTRβ: S. mansoni TR beta, SpTR: sea urchin Strongylocentrotu purpuratus TR. The accession numbers of the aligned human nuclear receptors can be found in Additional files 1 and 2.  *  hTRa1  GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  mTRa  GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  hTRb  GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPSYSCKY  rTRb  GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKSLHPSYSCKY  duckTRa  GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  duckTRb  GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  chickenTRa  GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  chickenTRb  GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  LizardTRa  GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  LizardTRb  GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  alligatorTRa1 GYIPSYLDKD---EQ CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  AlligatorTRb GYIPSYLDKD---EL CVVCGDKATGYHYRCIT CEGCKGFFRR TIQKNLHPTYSCKY  xTRa1  GYIPSYLDKD---EP CVVCSDKATGYHYRCIT CEGCKGFFRR TIQKNLHPSYSCKY  xTRb1 GYIPSYLDKD SmTRα and SmTRβ encode proteins of 1115 amino acids and 847 amino acids, respectively. Each protein exhibits a modular structure characteristic of the nuclear receptor superfamily with an A/B domain, a conserved DBD, a hinge region and a LBD ( Fig. 2 and 3). Alignment of the sequences demonstrated that the LBD is less conserved. Helices 1-2 in the LBD are highly divergent, similar to that of other reported schistosome NRs [11][12][13][14][15][16][17][18][19]. Although the LBD is less conserved (Fig. 2), the consensus motif I (from helix 3 to helix 6) and the consensus motif II (from the middle of helix 8 to the middle of helix 10) are conserved [20] (Fig. 2). A putative AF2 activation domain core (AF2-AD) (helix 12) is present in both SmTRs (Fig. 2). It exhibits a high degree of conservation (represented by YLHELF and YFHELF in SmTRα and SmTRβ, respectively) in comparison to the common consensus AF2-AD core structure of ΦΦXEΦΦ, where Φ denotes a hydrophobic residue [12,21,22].
The two consensus LBD motifs (motif I and motif II) were examined in other non-chordate TR orthologues when their DBD-containing contig contained these regions ( Fig.  4). Two amino acids in motif I (a glycine at position 21 and an aspartic acid at position 29) and the three amino acids in motif II (alanine at position 3, a leucine at position 9 and an isoleucine or leucine at postion 45) were found to be invertebrate TR-specific (Fig. 4).

Phylogenetic analysis
Phylogenetic analysis of DBD sequences shows that all invertebrate TRs cluster with the vertebrate TR subgroup (Fig. 5), suggesting that they originated from a common ancestor gene. The two TR orthologues in the Platyhelminths, the turbellarian S. mediterranea and the trematodes (S. mansoni and S. japonicum) clustered outside of the vertebrate TR subgroups suggesting that they underwent duplication after the split of Prostostomia and Deuterostomia. Furthermore, each of the two duplicated trematode TR homologues cluster together within a subgroup, while the two turbellarian (S. mediterranea) TR homologues clustered within another subgroup. This result suggests that the TR duplication in the Platyhelminths occurred after the split of the trematodes and the turbellarians.
A high resolution phylogenetic tree was constructed using the DBD and LBD sequences (Fig. 6). Since the only full length LBD sequence was identified from S. mansoni TRs, the S. mansoni TRs were used to represent the protostome TRs. The results show that all jawed vertebrate TRs clustered within a group including TRα and TRβ subgroups. The two jawless vertebrate lamprey TRs clustered outside of the jawed vertebrate TRs. This suggests that the vertebrate TR gene underwent duplication after the split of jawless and jawed vertebrates. The urochordate Ciona TR and the two duplicated trematode SmTRs cluster outside of the vertebrate TR groups. These results confirmed those obtained from the DBD analysis which suggested that the TR gene duplication occurred independently in vertebrates and invertebrates.

Genomic organization
The gene organization of the TR homologue genes in S. mansoni (SmTRα and SmTRβ) was identified by alignment of the cDNA sequence with the genomic sequence (Fig. 3 The ORF of SmTRα is encoded by seven exons spanning over 32 kb and the ORF of SmTRβ is encoded by six exons spanning over 21 kb (Fig. 3). All splice donor and acceptor sites for the two genes fit the GT-AG role. The A/B domain, hinge and E domains (LBD) are each encoded by 2-3 exons, respectively. The DBD of both proteins is encoded by one exon (Fig. 3). The 5' UTR and 3'UTR in both SmTRα and SmTRβ were not determined due to a short 3'UTR in SmTRα which seems not to be full length and due to the genomic contig not covering the 5'-and 3'UTR for SmTRβ.
It is known that there are four conserved splice junctions in nuclear receptor genes [3,4,9,20]. The first conserved junction position (JP1) is within the DBD encoding    region, the second (JP2) is in the region which encodes the 3' end of the DBD, the third (JP3) is in the LBD motif I (also known as the LBD signature sequence, Tτ) encoding region [20] and the fourth one (JP4) is in the LBD motif II encoding region [20]. JP1 exhibits diversity among nuclear receptor genes but is conserved for certain gene groups including vertebrate TR genes [9]. JP2 is highly conserved usually in the region encoding the fourth amino acid of the hinge region, but most of the Drosophila NR genes have lost this junction [9]. JP3 is NR gene-specific [20] and JP4 is at the same position in all vertebrate NR genes [20]. All identified non-chordate TR orthologues possess the four JPs, except JP1 was lost in the schistsome TR genes (SmTRα, SmTRβ, SjTRα and SjTRβ) and JP2 was lost in DpTR. JP3 of all invertebrate TR genes is at the same position as that found in vertebrate TR genes (NR1A) (Fig. 7). JP4 in all invertebrate TRs is located at the same conserved position as found in all analyzed NRs [20]. The conserved splice junctions in these invertebrate TRs suggest that the organization of the TR genes was ancient and has been maintained throughout evolution. It adds further support to the notion that vertebrate and invertebrate TR orthologues originated from a common ancestor gene.

Protein-protein interaction
One property of vertebrate TRs is the ability to form a heterodimer with RXR [4]. To determine if a non-chordate TR would form a heterodimer with RXR, the interaction between SmRXR1 with SmTRα and SmTRβ was tested by GST pull-down experiments. Either GST-SmTRα-LBD or GST-SmTRβ-LBD fusion protein was incubated with in vitro translated 35 S-SmRXR1. GST-bound resin was used as a negative control to assess non-specific background binding. The results showed that both SmTRα-LBD and SmTRβ-LBD could form a heterodimer with SmRXR1 in vitro. The results demonstrate that non-chordate TRs are capable of forming a heterodimer with RXR (Fig. 8).
Electrphoretic mobility shift assays (EMSAs) were performed to determine DNA binding specificity of SmTRα and SmTRβ. A DNA element containing the half-site AGGTCA, a direct repeat of the half-site spaced with 0-5 nucleic acids (DR0-DR5) and palindrome repeat of the half-site not separated by nucleic acids (Pal0) were employed. A gel shift was observed when γ-32 P labelled half-site, DR0-DR5 or Pal0 were added to SmTRα. SmTRα could bind to all the tested oligonucleotides either as a monomer or as a homodimer (Fig. 9A). SmTRβ demon-  I  II  III  IV  V  VI  VII   I  II  III  IV  V  VI

Gene organization of SmTRα and SmTRβ
strated a weak gel shift with the same labelled oligonucleotides as template (Fig. 9B). Heterodimers of either SmTRα or SmTRβ with SmRXR1 did not cause a gel shift ( Fig. 9A and 9B). SmRXR1 alone could bind to the tested oligonucleotides ( Fig. 9A and 9B) as we previously demonstrated [19].
The above result showed that SmTRα could bind to a half site, DR0-DR5 and Pal0 element as a monomer or a homodimer. To test whether the DBD domain (which contains a weaker-dimer interface in the D-box) could bind to the oligonucleotides as a homodimer, in vitro synthesized SmTRα (Gln 182 -Ala 288 ) (containing 20 aa at the 5' end of the DBD, the DBD and 40 aa at the 3' end of the DBD) was tested. SmTRα (Gln 182 -Ala 288 ) bound to all tested oligonucleotides as either a monomer or a homodimer (Fig. 10). The results suggest that there is a dimer interface located in DBD region of SmTRα [25][26][27]. Specificity was demonstrated by the ability of cold specific competitor but not the non-specific competitor to prevent SmTRα binding to the oligonucleotide template.  Maximum Likelihood phylogenetic tree derived from amino sequences of DBD and LBD Figure 6 Maximum Likelihood phylogenetic tree derived from amino sequences of DBD and LBD. The phylogenetic relationship among TRs was examined by the Maximum Likelihood method (ML) and Bayesian inference with the same methods as in Fig. 6. Support values for ML tree were obtained by bootstrapping a 100 replicates and are indicated above each branch. Branches under the bootstrap value of 50 were shown as polytomies. Bayesian inference was running 3 million generations. The PPs are shown below each branch or after the ML bootstrapping value separated by a slash. Star indicates the node obtained form by Bayesian inference which was different from that obtained by the ML method. SmTRα: blood fluke S. mansoni TR alpha, SmTRβ: S. mansoni TR beta. The accession number of each sequence used for phylogenetic analysis can be found in Additional files 1 and 2.

Discussion
An extensive search of whole genomic sequence (WGS) databases extracted from NCBI and online resources revealed the presence of TR orthologues in the Platyhelminths, Mollusca, Crustacea, and Echinodermata. However, no TR orthologues nor sequences encoding a TR P-box sequence of CEGCKG followed by the amino acid sequence FFRR (CEGCKGFFRR), which is unique to NR subfamily 1 including TR, was found in sponges (Reniera sp) nor cnidarians (Hydra magnipapillata and Nematostella vectensis). The results suggested that the TR orthologue gene originated in a common ancestor of the Bilateria. The gene duplication of TRs was deduced from a phylogenetic analysis employing DBD and LBD sequences. The analyses demonstrated that the TR orthologue genes underwent duplication independently in invertebrates and vertebrates. Furthermore, the TR orthologues in the Platyhelminths underwent duplication after the split of the turbellaria and trematodes. In vertebrates, the TR genes underwent a duplication event in lampreys, a jawless vertebrate, after the lamprey split from a common ancestor of jawed vertebrates. This result is consistent with the duplication of Hox gene clusters in lampreys [28]. Although the A/B domain of NR is highly divergent, we identified a conserved NTSS (N-terminal signature sequence, (G/P)YIPSY(M/L)XXXGPE(D/E)) located in the C-terminus of the A/B domain of TRs. Analysis of other NRs (data not shown) demonstrates this motif is found only in TRs suggesting it is TR-specific. The NTSS of TRs was found to be sequence-specific for the chordate and for the non-chordate TRs, three amino acids present in nonchordate TRs are missing in chordate NTSS. In the DBD, a methionine at position 16 of the N-terminus was nonchordate-specific, while two amino acids inserted at positions 33-34 of the N-terminus were chordate-specific, These differences which are conserved, support the notion that TR genes duplicated independently in non-chordate and chordate species.
That the JP3 conserved in all invertebrate TRs are also conserved in vertebrate TR genes support the notion that the vertebrate and invertebrate TR orthologue genes originated from a common ancestor gene. Further, that the four JPs are conserved in both invertebrate and vertebrate TR orthologue genes suggest that the gene organization of vertebrate TR genes is ancient and has been maintained through out evolution of TRs.
GST pull-down experiments showed that SmTRs exhibits similarity to vertebrate TR proteins as they both can form a heterodimer with RXR (SmRXR1). RXRs have been characterized in other invertebrates such as Mollusca and Arthropoda [29,30]. The formation of SmTRs with SmRXR1 indicates that other invertebrate TRs can possibly form a heterodimer with RXR. The ability of vertebrate NR to bind to a DNA element on a target gene was described as the 1-2-3-4-5 rule [31,32]. Vertebrate TRs are known to bind to half-sites as a monomer, to DR0 as a heterodimer with RXR [33] and to DR4 as a homodimer or heterodimer [32,[34][35][36][37][38][39][40][41][42][43]. The binding of the TR monomer or homodimer is weak, which is due to a rapid dissociation of the TR-TRE complex [44]. Our EMSA show that SmTRα can bind to a half-site as a monomer and bind to DR0-DR5 as a homodimer. The results suggested that TR gained the ability to bind to conserved half-site repeats in a common ancestor, but did not subsequently evolve a distinct specificity as regards spacing between half sites. Although the DBD sequence between SmTRα and SmTRβ are highly similar, SmTRβ showed weak binding to the DNA elements compared to that of SmTRα. This suggests that other domains may have a role in determining the ability of SmTRβ to bind to the DNA element, such as has been shown for another schistosome nuclear receptor, SmNR1 [19]. Although both SmTRs could form a heterodimer with SmRXR1 as determined in a GST pulldown assay, SmTRα/SmRXR1 or SmTRβ/SmRXR1 did not bind to the DNA elements tested. This may due to the flanking or spacing of nucleotides between the HRE, as it is known that these nucleotides also determine the ability of the protein to bind to it. Identification of the HRE in a target gene will help to understand the evolution of TR DNA binding properties. Mammalian TRs are liganddependent transcription factors, they repress basal transcription through association with a variety of corepressors in the absence of T3. A structural change occurs when Protein-protein interaction of SmTRs with SmRXR1 Figure 8 Protein-protein interaction of SmTRs with SmRXR1. GST pull down shows that both SmTRα and SmTRβ can form a heterodimer with SmRXR1 in vitro. 35 S-labeled SmRXR1 was synthesized in vitro using pCITE-SmRXR1 as template and then incubated with GST-SmTRα(LBD), GST-SmTRβ(LBD) or GST (negative control) protein affixed to glutathione-Sepharose beads. The beads were collected, washed and the bound protein was resolved on 10% SDS acrylamide gel and visualized by autoradiography.

Database search
Whole genomic sequences (WGS) were extracted from the GenBank public ftp site [45] and imported into the Star-Blast program (DNASTAR) to build a local database [18] which was searched by tblastn using the sequence of the DBD of SmTRα as the query. Any sequence that contained a zinc finger structure of the DBD (Cys-X2-Cys-X13-Cys-X2-Cys or Cys-X5-Cys-X9-Cys-X2-Cys) was retained.
Sequence walking was carried out to assemble the contigs. Website databases of GenBank (nr, EST_human, EST_mouse and EST_other databases) Genbank blast [46], European Bioinformatics Institute [47] and Swiss-Prot [48] were also mined by tblasn or blastp using the same query sequence as above [18].

BAC library screening and BAC DNA sequencing
S. mansoni BAC clones containing SmTRα and SmTRβ were identified by screening S. mansoni SmBAC1 or CHOR-1 BAC library with methods previously described [23]. For BAC DNA sequencing, the BAC clone was grown in 100 ml LB medium (12.5 µg/ml choramphenicol) and the BAC DNA was purified using Plasmid Midi kit (Qiagen) and sequenced on an ABI-377 automatic sequencing machine.

Phylogenetic tree construction
Phylogenetic trees were constructed from deduced sequences of the DBD and the sequence of DBD plus LBD, respectively. Sequences were aligned with ClustalW [49]. Phylogenetic analysis of the data set was carried out using Maximum Likelihood method under Jones-Taylor-Thornton (JTT) substitution model [50] with a gamma distribution of rates between sites (eight categories, parameter alpha, estimated by the program) using PHYML (v2.4.4)) [51] with support values obtained by bootstrapping a 100 replicates. The same data set was also tested by Bayesian inference using MrBAYES v3.1.1 with a JTT amino acid replacement model + gamma rates [52]. The trees were started randomly; four simultaneous Markov chains were run for 5 million generations for the DBD data set and 3 million generations for the DBD+LBD data set, respectively. The trees were sampled every 100 generations, the Bayesian posterior probabilities (PPs) were calculated using a Markov chain Monte Carlo (MCMC) sampling approach implemented in MrBAYES v3.1.1, with a burnin value setting at 12,500 for DBD data set and 7,500 generations for the DBD+LBD data set, respectively.
For pull-down experiments, a 50 µl reaction that contained 2 µl of the in vitro translation reaction, SmTRα(LBD) or SmTRβ(LBD) GST fusion protein, or GST protein (negative control) affixed to glutathione-Sepharose beads (about 2 µg) and binding buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10% glycerol, 0.15% Nonidet P40) was used [19]. The reaction was incubated overnight at 4°C and then washed three times with binding buffer. The bound proteins were analyzed by 10% SDS-PAGE and autoradiography.