Human transforming growth factor-beta activates a receptor serine/threonine kinase from the intravascular parasite Schistosoma mansoni.

The biology of the helminth parasite Schistosoma mansoni is closely integrated with that of its mammalian host. SmRK1, a divergent type I transforming growth factor-beta (TGF-beta) receptor of unknown ligand specificity, was previously identified as a candidate for a receptor that allows schistosomes to respond to host-derived growth factors. The TGF-beta family includes activin, bone morphogenetic proteins (BMPs), and TGF-beta, all of which can play crucial roles in metazoan development. The downstream signaling protein of receptors that respond to TGF-beta and activin is Smad2, whereas the receptors that respond to BMPs signal via Smad1. When a constitutively active mutant of SmRK1 was overexpressed with either schistosome Smad1 (SmSmad1) or SmSmad2, a receptor-dependent modulation of SmSmad phosphorylation and luciferase reporter activity occurred only with SmSmad2. To evaluate potential ligand activators of SmRK1, a chimeric receptor containing the extracellular domain of SmRK1 joined to the intracellular domain of the human type I TGF-beta receptor was used. The chimeric receptor bound radiolabeled TGF-beta and could activate a luciferase reporter gene in response to both TGF-beta 1 and TGF-beta 3 but not BMP7. Confirmatory results were obtained using full-length SmRK1. These experiments implicate TGF-beta as a ligand for SmRK1 and as a potential host-derived regulator of parasite growth and development.

The biology of the helminth parasite Schistosoma mansoni is closely integrated with that of its mammalian host. SmRK1, a divergent type I transforming growth factor-␤ (TGF-␤) receptor of unknown ligand specificity, was previously identified as a candidate for a receptor that allows schistosomes to respond to hostderived growth factors. The TGF-␤ family includes activin, bone morphogenetic proteins (BMPs), and TGF-␤, all of which can play crucial roles in metazoan development. The downstream signaling protein of receptors that respond to TGF-␤ and activin is Smad2, whereas the receptors that respond to BMPs signal via Smad1. When a constitutively active mutant of SmRK1 was overexpressed with either schistosome Smad1 (SmSmad1) or SmSmad2, a receptor-dependent modulation of SmSmad phosphorylation and luciferase reporter activity occurred only with SmSmad2. To evaluate potential ligand activators of SmRK1, a chimeric receptor containing the extracellular domain of SmRK1 joined to the intracellular domain of the human type I TGF-␤ receptor was used. The chimeric receptor bound radiolabeled TGF-␤ and could activate a luciferase reporter gene in response to both TGF-␤1 and TGF-␤3 but not BMP7. Confirmatory results were obtained using full-length SmRK1. These experiments implicate TGF-␤ as a ligand for SmRK1 and as a potential host-derived regulator of parasite growth and development.
Schistosomiasis is a chronic parasitic disease that affects more than 200 million people worldwide with the majority of cases occurring in sub-Saharan Africa. Despite treatment programs, the number of individuals infected with schistosomes has not declined, and the rate of transmission is increasing, especially for Schistosoma mansoni (1). The morbidity associated with S. mansoni infection results from hepatic damage induced by parasite eggs, which are released from mature worms residing in the mesenteric veins of the host (2). During the prepatent period, the immature parasite must undergo a complex migration and developmental process within the host, and host factors have been implicated as regulators of these events (3)(4)(5). However, the mechanisms used by the parasite to recognize and respond to host-derived factors are unknown.
While investigating host-parasite interactions, our laboratory identified S. mansoni receptor kinase 1 (SmRK1) 1 , a receptor serine/threonine kinase from the transforming growth factor-␤ (TGF-␤) superfamily, and showed it to be present on the surface of the parasite (6). SmRK1 contains a conserved glycine-serine motif and shares up to 58% identity with kinase domains of other type I receptors in the TGF-␤ superfamily (6). Nevertheless, unusual features of its sequence such as an insert in the cytoplasmic juxtamembrane region and a large COOH-terminal extension make it a divergent member of the type I receptor subfamily.
The TGF-␤ superfamily has been shown to participate in a wide variety of growth and developmental processes (7), and the signaling pathways for many of the receptors in this family have been studied in great detail (8 -10). In general, the type II receptor is responsible for binding ligand and recruiting the type I receptor into a heteromeric complex. The constitutively active kinase domain of the type II receptor phosphorylates the type I receptor, thereby activating its kinase domain. The activated type I receptor phosphorylates receptor-regulated Smad proteins (R-Smad), which can then translocate to the nucleus to modulate gene transcription. Typically, the receptors responding to TGF-␤ and activin signal through Smads 2 and 3, whereas the receptors responding to bone morphogenetic proteins (BMP) use Smads 1 and 5 (11). We have recently identified two schistosome R-Smads (SmSmad1 and SmS-mad2) and have shown that both can interact with SmRK1 in vitro as well as signal downstream of vertebrate TGF-␤ receptors expressed in mammalian cells (12). SmSmad2 was recently shown to be phosphorylated by a constitutively active SmRK1 in vitro (13), but the physiological significance of the SmRK1-SmSmad1 interaction and, most importantly, the ligand specificity of SmRK1 are unknown.
Receptor serine/threonine kinases from the TGF-␤ superfamily show a high degree of functional conservation across very different species. Because schistosomes, as parasites, live in physiological contact with their mammalian host, we wanted to address the possibility that TGF-␤-related ligands from the host might activate SmRK1 during the course of infection. In this paper, we present evidence demonstrating that SmRK1 is capable of signaling via SmSmad2 in response to human TGF-␤.

EXPERIMENTAL PROCEDURES
Expression Vectors and Reporter Constructs-The construction of pCMV2 FLAG-SmSmad1 and FLAG-SmSmad2 for NH 2 -terminal tagged SmSmads has been described previously (12). The pCMV5 T␤RIhemagglutinin vector (human type I TGF-␤ receptor), pCMV5 T␤RII-His (human type II TGF-␤ receptor), and the 3TP-Lux luciferase reporter construct (14) were provided by J. Massague (Memorial Sloan-Kettering Cancer Center, New York, NY). The 12xGCCG-Lux luciferase reporter construct (15) was provided by M. Kawabata (The Cancer Institute, Tokyo, Japan). Full-length SmRK1, which included an additional 188 amino acids at the COOH terminus relative to that originally published due to a recently discovered sequencing error (AF031557), was subcloned into pCMV5b FLAG (Sigma) in frame with a COOHterminal FLAG epitope or pSecTag2 vector (Invitrogen) without a tag. The mutant forms of SmRK1 (Q303D,K331R) were generated using a polymerase chain reaction-based site-directed mutagenesis kit (Stratagene), and mutations were confirmed by sequencing. All chimeric receptors were generated using a sequential polymerase chain reactionbased approach adapted from previously published protocols (16) using a high fidelity Taq polymerase (Roche Molecular Biochemicals) and the TOPO-TA cloning kit (Invitrogen). The resulting chimeric receptors were fully sequenced before subcloning into their final vectors. The T␤RI/SmRK1 chimera was expressed in the pSecTag2 vector and included amino acid residues 34 -186 from T␤RI and residues 286 -780 from SmRK1. The SmRK1/T␤RI chimera was also expressed in this vector and contained amino acid residues 1-132 from SmRK1 and residues 126 -503 from T␤RI. The activin receptor-like kinase (ALK2)/ T␤RI chimera was expressed in pCMV5b FLAG and included amino acid residues 1-123 from ALK2, generated by polymerase chain reaction from murine cDNA, and included residues 126 -503 from T␤RI.
Luciferase Reporter Assays-Transcriptional response assays were performed as described previously (12). R1B cells (provided by J. Massague) were plated at a density of 1 ϫ 10 5 cells/well of a 6-well plate and allowed to grow for 18 -24 h. Cells were transiently transfected with the indicated constructs as well as pSV␤-galactosidase (Promega) using FuGENE6 (Roche Molecular Biochemicals). After 24 h, cells were preincubated in reduced serum media (0.2% fetal bovine serum) for 4 h and either left untreated or stimulated for 20 h with 5 l of vehicle (40% ethanol and 0.1% trifluoroacetic acid), BMP7, or recombinant human TGF-␤ (R&D Systems). Both luciferase and ␤-galactosidase activities were measured from the same sample using a Berthold luminometer (Lumat LB9501). The results of the luciferase assay were normalized on the basis of ␤-galactosidase expression. The measurements were performed on triplicate samples/experiment, and the values reported represent pooled results from at least two independent experiments. The Student's t test was used to assess the statistical significance of these results.
Immunoprecipitations and Immunoblotting-COS7 cells were transfected using LipofectAMINE (Life Technologies, Inc.) as directed. Approximately 48 h post-transfection, cells were washed and lysed, and the expressed proteins were immunoprecipitated according to protocols published previously (17). SmRK1 receptors were immunoprecipitated with the anti-M2 FLAG antibody (Sigma), and the SmSmads were immunoprecipitated with the anti-M5 FLAG antibody (Sigma). Total cell lysates and immunoprecipitated proteins were separated by SDSpolyacrylamide gel electrophoresis and transferred to Immobilon P (Millipore) for immunoblotting and, in some cases, autoradiography.
In Vitro Kinase Assay and Phosphate Labeling-In vitro kinase assays of SmRK1 immunoprecipitated from COS7 cells were performed using 50 M ATP and 30 Ci of [␥-32 P]ATP/reaction (Amersham Pharmacia Biotech) in a total volume of 40 l (18). After several washes in lysis buffer, proteins were eluted by boiling in sample buffer, separated on SDS-polyacrylamide gel electrophoresis, and visualized by autoradiography. For 32 P-phosphate labeling, cells were incubated for 2-3 h in phosphate-free minimum Eagle's medium containing 1 mCi/ml 32 P-orthophosphate (PerkinElmer Life Sciences). The labeled cells were washed and lysed, and the epitope-tagged proteins were immunoprecipitated (17). Phosphorylation of proteins was assessed both by autoradiography and phosphorimaging analysis (PhosphorImager, Molecular Dynamics).
TGF-␤ Binding and Coimmunoprecipitation-The ligand binding of [ 125 I]TGF-␤ (PerkinElmer Life Sciences) to transfected COS7 cells was performed according to previously published protocols (17). Specifically, cells were treated with 250 pM [ 125 I]TGF-␤ for either 30 min at 37°C or 3 h at 4°C, washed extensively, and then treated with 0.5 mg/ml of disuccinimidyl suberate (Pierce) for 30 min at 4°C to cross-link the ligand to the receptors. Cell lysates were subjected to immunoprecipitation with either anti-MYC (Santa Cruz Biotechnology) or anti-hemagglutinin (Roche) antibody. For coimmunoprecipitation, the ligand binding was performed as above for 30 min at 37°C, and cell lysates were immunoprecipitated with the anti-FLAG (M2) antisera. Phosphor-imaging analysis of all dried gels was performed using ImageQuant software (Molecular Dynamics).

RESULTS
Evaluating the Signaling Specificity of SmRK1-SmRK1 interacts with both SmSmad1 and SmSmad2 in vitro (12,13). The specificity of interaction between R-Smads and type I receptors is facilitated by specific structural motifs in each protein. The predominant domain in the R-Smad is a region of 17 amino acids referred to as the L3 loop (19), and both SmSmads are highly conserved in this region relative to their mammalian counterparts (12,13). A 9-amino acid segment in the kinase domain of type I receptors referred to as the L45 loop has been reported to define the signaling specificity of type I receptors for intracellular responses (16). A cluster of four amino acids in this domain is important for receptor-Smad interactions and has been used to divide type I receptors into three categories based on the L45 loop sequence and Smad interactions (20). A comparison of the L45 loop sequence from SmRK1 with those of the three categories of type I receptors revealed that SmRK1 does not completely conform to any of these categories. Of the four critical amino acids, SmRK1 shares the first two with the TGF-␤ group, the third with the BMP group, and has a non-conserved substitution in the fourth position (Fig. 1A). Hence, the signaling specificity of SmRK1 could not be predicted based upon this analysis nor could its in vitro interactions with SmSmads 1 and 2.
To begin to address the signaling specificity of SmRK1 in the absence of a known ligand, a T␤RI/SmRK1 chimera was generated. The chimera was constructed by joining the extracellular, transmembrane, and juxtamembrane domains of T␤RI with the intracellular domain of SmRK1 beginning at the glycine-serine box (Fig. 1B). Although the T␤RI/SmRK1 chimera does not include the modified juxtamembrane domain of SmRK1, it does incorporate the COOH-terminal extension, which includes 175 amino acids beyond the conserved kinase domain of other type I receptors. The T␤RI/SmRK1 chimera would permit ligand-dependent activation of the intracellular domain of SmRK1 to identify the downstream SmSmad pathway in a luciferase reporter assay.
We have previously shown that SmSmad2 can signal downstream of T␤RI and drive the expression of luciferase from the plasminogen activator inhibitor-1 promoter region of the 3TP-Lux plasmid (12). SmSmad1 activation was assessed using the 12xGCCG-Lux reporter plasmid in R1B cells, which are a derivative of mink lung epithelial cells lacking a functional type I TGF-␤ receptor. It has been shown previously that mammalian Smad1 can drive luciferase expression from the GCCG-Lux reporter construct in this cell type in response to BMP signaling (15). Likewise, when SmSmad1 was transfected into R1B cells with the GCCG-Lux reporter, it was able to enhance luciferase expression after treatment of the cells with BMP7 ( Fig. 2A).
To test the ability of the T␤RI/SmRK1 chimera to interact with the SmSmads, R1B cells were transiently transfected with either empty vector (pCMV5), each SmSmad alone, the T␤RI/SmRK1 chimera alone, or a combination of the chimera with each SmSmad. Whereas very little change was observed in response to TGF-␤ in cells expressing only the chimera, coexpression with either SmSmad produced an increase in luciferase activity under both the basal and ligand-stimulated conditions (Fig. 2B). However, only in the case in which T␤RI/SmRK1 was coexpressed with SmSmad2 was the ligand-induced increase in luciferase activity significantly greater than that observed under basal conditions (p Ͻ 0.01).
SmRK1 Mutations Alter the Kinase Activity of the Receptor and Reveal Signaling Specificity-Because both of the SmSmads were able to increase luciferase activity in the presence of the stimulated T␤RI/SmRK1 chimera, we decided to evaluate the signaling specificity of the full-length schistosome receptor containing kinase domain mutations, which would permit the modulation of receptor activity independent of the ligand. Based on the studies of other type I receptors (18,21), mutations in the schistosome receptor were generated that were predicted to either constitutively activate (SmRK1(Q303D)) or inactivate (SmRK1(K331R)), the kinase domain of SmRK1 (Fig. 3A). Because SmRK1 is a divergent type I receptor, we initially tested the ability of these mutations to alter the kinase activity of the receptor in an in vitro kinase assay. Other type I receptors have been shown to undergo autophosphorylation in this reaction, providing an indication of the status of the receptor kinase. Whereas SmRK1 failed to autophosphorylate, it did phosphorylate an unknown substrate in the reaction mixture in a receptordependent fashion, and the level of phosphorylation was modulated according to the mutations in the kinase domain (Fig. 3B). Having confirmed that the mutations did alter SmRK1 kinase activity, the mutant receptors were tested in an in vivo phosphorylation assay using each of the SmSmads as a substrate. COS7 cells expressing either FLAG-tagged SmSmad1 or SmSmad2 alone or in combination with SmRK1(QD) or SmRK1(KR) were labeled with 32 P-orthophosphate, and the SmSmads were immunoprecipitated with an anti-FLAG(M5) antibody. Whereas the level of expression of SmSmad1 was less than that observed for SmSmad2, SmSmad1 was clearly detectable in total cell lysates, and no receptor-dependent differences in SmSmad1 phosphorylation were observed (data not shown). In contrast, the phosphorylation of SmSmad2 was increased by 60% in the presence of SmRK1(QD) and decreased by 20% in the presence of SmRK1(KR) relative to SmSmad2 expressed alone (Fig. 3C).
To obtain additional functional evidence regarding the SmSmad specificity of SmRK1, the mutant forms of the schistosome receptor were evaluated in the luciferase reporter assay. R1B cells were transfected with either empty vector, each SmSmad alone or in combination with either wild type SmRK1, SmRK1(QD), or SmRK1(KR), in the presence of the appropriate reporter construct for each SmSmad. The coexpression of SmRK1(QD) with SmSmad2 resulted in a 12-fold increase in luciferase activity, 6 times greater than that observed with SmSmad2 alone, and the response was completely reduced to background levels in the presence of SmRK1(KR) (Fig. 3D). Although a 10-fold increase in luciferase activity was observed when SmSmad1 was coexpressed with SmRK1(QD), this response was only 2-fold greater than that for SmSmad1 expressed alone and was not significantly reduced in the presence of the kinase dead form of SmRK1 (Fig. 5A). These data, obtained in the absence of ligand, are consistent with the results from the luciferase assay using the T␤RI/SmRK1 chimera and demonstrate that SmRK1 has greater signaling specificity through SmSmad2, similar to that of TGF-␤ and activin receptors.
Functional Interactions of the Extracellular Domain of SmRK1 with Potential Ligands-Having investigated the signaling specificity of the intracellular domain of SmRK1, we next directed our attention to the extracellular domain of the receptor and its ability to interact with ligands from the TGF-␤ superfamily. Using a chimeric receptor approach similar to that employed for ALK1 (22), two chimeric receptors were constructed: 1) the extracellular domain of either SmRK1 or 2) the type I receptor for BMP7, ALK2, was fused to the transmembrane and intracellular domain of T␤RI, resulting in a SmRK1/T␤RI chimera and an ALK2/T␤RI chimera, respectively (Fig. 1B). Each of these chimeras could be expressed at high levels in COS7 cells (data not shown). In R1B cells, the ALK2/T␤RI chimera served as both a positive and negative control in the luciferase reporter assays. A similar chimera when expressed in R1B cells has been shown previously to be unresponsive to TGF-␤ in luciferase reporter assays using a plasminogen activator inhibitor-1 reporter construct (16). However, we expected this chimera to activate the 3TP-Lux reporter when the R1B cells were treated with BMP7, as BMP7 has been shown to activate ALK2 receptors (23).
To confirm the activity of the chimeric signaling system, R1B cells expressing either empty vector or the ALK2/T␤RI receptor were stimulated overnight with TGF-␤1 or BMP7. Although minor fluctuations were observed with the empty vector control, the ALK2/T␤RI chimera showed no response when treated with TGF-␤1 but demonstrated a 2.5-fold increase in luciferase activity in response to BMP7. When the SmRK1/T␤RI chimera was expressed in R1B cells under these same conditions, a 2.5-fold increase in luciferase activity was observed after treatment with TGF-␤1, and no change in activity occurred after stimulation with BMP7 (Fig. 4A). In an attempt to improve the magnitude of the luciferase response, the chimeric receptors were coexpressed with SmSmad2. The coexpression of SmS-mad2 with T␤RI was previously shown to enhance luciferase reporter activity in R1B cells treated with TGF-␤ compared with T␤RI expressed alone (12). In the presence of SmSmad2, the SmRK1/T␤RI chimera showed a 4-fold increase in luciferase activity in response to both TGF-␤1 and TGF-␤3 but showed no change after stimulation with BMP7 (Fig. 4B). Similarly, R1B cells coexpressing the ALK2/T␤RI chimera and SmSmad2 demonstrated a 4-fold increase in luciferase activity in response to BMP7, whereas the treatment with either isoform of TGF-␤ produced levels of luciferase activity similar to those for the pCMV5 control (data not shown).
The results of these ligand interaction studies suggest that the extracellular domain of SmRK1 can specifically interact with TGF-␤ and lead to the activation of the SmRK1/T␤RI chimera. However, the studies of mammalian TGF-␤ signaling have shown that type I TGF-␤ receptors like T␤RI are unable to bind ligand in the absence of the type II TGF-␤ receptor (24), suggesting that productive signaling for SmRK1/T␤RI in R1B cells requires endogenous T␤RII. To directly assess the ability of SmRK1 to interact with TGF-␤ independent of T␤RII, [ 125 I]TGF-␤ labeling of COS7 cells expressing either the SmRK1/T␤RI chimera alone or in combination with T␤RII was performed. As a positive control for TGF-␤ binding, T␤RI and T␤RII were coexpressed in COS7 cells. Type I receptors were immunoprecipitated with either anti-MYC or anti-hemagglutinin antibody directed against the COOH-terminal tagged epitope, and the presence of ligand-bound types I and II receptors was evaluated. Similar to other type I TGF-␤ receptors, the chimeric receptor containing the extracellular domain of SmRK1 was unable to bind ligand in the absence of the type II TGF-␤ receptor. However, when the SmRK1/T␤RI chimeric receptor was coexpressed with T␤RII, both receptors were present in the immunoprecipitate labeled with the radioactive TGF-␤ (Fig. 4C). Control samples transfected with empty vector (pCMV5) or T␤RII alone and immunoprecipitated with the MYC antisera were negative for ligand binding.
Wild Type SmRK1 Signals through SmSmad2 in Response to TGF-␤-Having evaluated the function of the intracellular and extracellular domains of SmRK1 independently, we wanted to test the ability of the full-length receptor to signal through the schistosome Smads in response to ligand stimulation in the luciferase reporter assay. R1B cells were transfected with ei- ther empty vector, each SmSmad alone, wild type SmRK1 alone, or in combination with each SmSmad, all in the presence of the appropriate reporter construct for the SmSmad. We first evaluated SmRK1-SmSmad1 interactions in response to both TGF-␤ and BMP7. When R1B cells expressing both SmRK1 and SmSmad1 were treated with TGF-␤1, a ligand-dependent increase in luciferase activity resulted, but the increase was not significantly different from untreated cells coexpressing SmRK1 and SmSmad1 (p ϭ 0.12) (Fig. 5B). Likewise, BMP7 treatment of cells coexpressing SmRK1 and SmSmad1 did not produce an increase in luciferase activity above that observed for untreated cells coexpressing these two proteins ( Fig. 5A and data not shown).
We next examined whether the full-length receptor could signal through SmSmad2 in response to human TGF-␤. When SmRK1 was coexpressed with SmSmad2, the luciferase activity increased 3-fold relative to SmSmad2 alone, and this level of activity doubled after the treatment of cells with TGF-␤1 (p Ͻ 0.01) (Fig. 6A). Thus, the results obtained with the full-length receptor confirm the findings that the activated intracellular domain of SmRK1 interacts with SmSmad2 (Figs. 2B and 3C), and that the extracellular domain of SmRK1 functionally interacts with TGF-␤ (Fig. 4).
Because the ligand binding and type II receptor interaction studies were performed with the SmRK1/T␤RI chimera, fulllength SmRK1 was tested for its ability to bind TGF-␤ and form a complex with T␤RII and SmSmad2 in vivo. Previous studies have shown that the phosphorylation of Smad2 promotes its release from the type I receptor, but Smad2 can be immunoprecipitated with the receptor complex in the absence of phosphorylation (25). Hence, we tested the ability of the kinase dead form of SmRK1 along with T␤RII to be coimmunoprecipitated by SmSmad2 in transfected COS7 cells. COS7 cells expressing SmSmad2 and T␤RII either alone or in the presence of the untagged SmRK1(KR) were labeled with [ 125 I]TGF-␤ for 30 min at 37°C. After cross-linking and cell lysis, SmSmad2 was immunoprecipitated from the lysates with the anti-FLAG antibody. Although similar quantities of SmSmad2 were present in the total cell lysates, only those cells expressing the kinase dead form of SmRK1 demonstrated labeled receptor complexes after the SmSmad2 immunoprecipitation (Fig. 6B). Taken together, the data obtained from the studies of the full-length receptor support the hypothesis that human TGF-␤ is capable of stimulating SmRK1 and activating a signal transduction pathway, which uses SmSmad2.

DISCUSSION
Invertebrate BMP-like signaling pathways are well characterized with BMP receptors and ligands present in both nematodes and insects. However, evidence for TGF-␤/activin type pathways in this group has emerged only recently from studies of Drosophila melanogaster (26,27). Here we show that schistosomes which are members of the Platyhelminthes, a phylum that in the course of metazoan evolution diverged before the Nematoda, possess a type I receptor serine/threonine kinase that can be activated by human TGF-␤ and signal via schistosome Smad2. Because schistosomes are intravascular parasites of human beings, the activation of SmRK1 by a host-derived factor may have physiological relevance to the development of the parasite as well as the clinical course of the disease in humans.
In the in vitro culture systems we have used to study SmRK1 signaling, the type II TGF-␤ receptor appears to facilitate TGF-␤ binding to SmRK1 and may promote the activation of SmRK1 in response to ligand. Nevertheless, the chimeric receptors used to evaluate functional receptor-ligand interac- tions, SmRK1/T␤RI and ALK2/T␤RI, demonstrated the affinity of SmRK1 for TGF-␤. If a nonspecific interaction were occurring between T␤RII and the chimeric receptors, increased luciferase activity in response to TGF-␤ would have been observed for the ALK2/T␤RI chimera. Functional ligand interactions were not limited to the chimeric receptor as fulllength SmRK1 was activated by TGF-␤ in the luciferase assay, and the parasite receptor could be found in a complex with T␤RII, SmSmad2, and radiolabeled TGF-␤ immunoprecipitated from COS7 cells.
The specificity of the interaction between SmRK1 and T␤RII may also extend to the intracellular domain of the receptors. Previous studies using chimeric receptors have shown that only certain type I receptor intracellular domains can be activated by T␤RII. For example, the chimeric receptor containing the extracellular domain of T␤RI and the intracellular domain of the type I activin receptor, ActR-IB, was not activated by three different isoforms of TGF-␤ as measured by a luciferase reporter assay in R1B cells (22). Whereas similar results have been obtained with T␤RI/BMP receptors in R1B cells (28), the T␤RI/ActR-IB findings are particularly interesting given that the kinase domains of T␤RI and ActR-IB are 89% identical (T␤RI and SmRK1 share only 57% identity). Taken together, these results imply that SmRK1 has both an affinity for TGF-␤ as well as a high degree of compatibility with T␤RII. Further studies will be necessary to determine the means of TGF-␤ binding to the parasite receptor in vivo.
An alternative hypothesis to explain the mechanism of activation of SmRK1 is that the receptor exhibits kinase activity independent of a type II receptor. It has recently been shown in Caenorhabditis elegans that the type I receptor DAF-1 maintains signaling activity in the absence of DAF-4, the only type II receptor present in the C. elegans genome (29). Gunther and colleagues suggest that the activity associated with DAF-1 may represent an evolutionary remnant owing to the origin of the receptor near the divergence of type I and type II receptors. We have previously shown in a phylogenetic analysis of SmRK1 that the schistosome receptor probably diverged before DAF-1 (6), implying that SmRK1 may also be able to signal independently of a type II receptor. Additionally, the sequence differences present in the juxtamembrane and COOH terminus of SmRK1 are reminiscent of alternatively spliced isoforms of the mammalian type II receptors ActR-IIB (activin type II receptor B) and BRK-3 (BMP receptor kinase 3), respectively (30,31). Although type I receptor serine/threonine kinases have been described in other parasitic helminths (32), there have been no reports thus far of type II receptors. Although we have performed yeast two-hybrid screens to identify an interacting type II receptor for SmRK1 and other interacting proteins have been cloned and characterized, no type II receptor has been identified to date (33,34). The completion of the genome sequencing project for S. mansoni will provide valuable information regarding the presence or absence of additional TGF-␤-signaling components in schistosomes. Whereas our studies and those of others (13) suggest that SmRK1 signals through SmSmad2, there is also an interaction between SmRK1 and SmSmad1. When the two proteins were coexpressed in R1B cells, moderate increases in luciferase expression from the GCCG-Lux reporter plasmid were observed relative to SmSmad1 expressed alone. These findings may be an artifact resulting from the overexpression of proteins in mammalian cells, as similar observations have been made for T␤RI and Smad1 coexpressed in COS cells (35). The observed enhancement in luciferase activity appeared to be independent of the activation state of the receptor as indicated by the experiments using the mutant forms of SmRK1. Likewise, the ligand-dependent activation of SmRK1 as either the T␤RI/ SmRK1 chimera or the wild type receptor was unable to significantly alter the activation state of SmSmad1 as measured by the luciferase reporter assay. Our inability to detect significant -fold increases in the presence of SmRK1 and ligand may be because of the participation and perhaps sequestration of SmSmad1 in the endogenous BMP pathway in R1B cells. Whether or not there is a physiologically relevant interaction between SmRK1 and SmSmad1 remains to be elucidated as does the possibility of another schistosome receptor for BMPtype ligands and SmSmad1 signaling.
In these experiments, we have used a heterologous system to study the signaling specificity of a divergent type I TGF-␤ receptor from the trematode parasite, S. mansoni. Although there are several caveats to using these types of systems to interpret the function of a signal transduction pathway, we are limited to these methods, because schistosomes are not amenable to transfection or other means of genetic manipulation at this time. Although these experiments have been a useful means to initiate investigations of signaling in schistosomes, the predictive value of these systems can only be confirmed by studies of signal transduction within the parasite itself. We now have a hypothesis which can be tested in the parasite as well as the mouse model of schistosomiasis to determine the effects of TGF-␤ signaling through the SmRK1-SmSmad2 pathway on the development of the parasite and the clinical course of infection.