- THIS ARTICLE
-
Abstract
- Full Text (PDF)
- Alert me when this article is cited
- Alert me if a correction is posted
- SERVICES
- Similar articles in this journal
- Similar articles in PubMed
- Alert me to new issues of the journal
- Download to citation manager
- Reprints & Permissions
- CITING ARTICLES
- Citing Articles via HighWire
- Citing Articles via Google Scholar
- GOOGLE SCHOLAR
- Articles by Newfeld, S. J.
- Articles by Kumar, S.
- Search for Related Content
- PUBMED
- PubMed Citation
- Articles by Newfeld, S. J.
- Articles by Kumar, S.
Molecular Evolution of a Developmental Pathway: Phylogenetic Analyses of Transforming Growth Factor-ß Family Ligands, Receptors and Smad Signal Transducers
Stuart J. Newfelda,b, Robert G. Wisotzkeyc, and Sudhir Kumaraa Department of Biology, Arizona State University, Tempe, Arizona 85287-1501,
b Graduate Program in Molecular and Cellular Biology, Arizona State University, Tempe, Arizona 85287-1144
c AXYS Pharmaceuticals, Inc., South San Francisco, California 94080
Corresponding author: Stuart J. Newfeld, Department of Biology, Arizona State University, Tempe, AZ 85287-1501., newfeld{at}asu.edu (E-mail)
Communicating editor: A. G. CLARK
 | ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Intercellular signaling by transforming growth factor-ß (TGF-ß) proteins coordinates developmental decisions in many organisms. A receptor complex and Smad signal transducers are required for proper responses to TGF-ß signals. We have taken a phylogenetic approach to understanding the developmental evolutionary history of TGF-ß signaling pathways. We were interested in detecting evolutionary influences among the physically interacting multigene families encoding TGF-ß ligands, receptors, and Smads. Our analyses included new ligands and Smads identified from genomic sequence as well as the newest published family members. From an evolutionary perspective we find that (1) TGF-ß pathways do not predate the divergence of animals, plants, and fungi; (2) ligands of the TGF-ß/activin subfamily likely originated after the divergence of nematodes and arthropods; (3) type I receptors from Caenorhabditis elegans are distinct from other receptors and may reflect an ancestral transitional state between type I and type II receptors; and (4) the Smad family appears to be evolving faster than, and independently of, ligands and receptors. From a developmental perspective we find (1) numerous phylogenetic associations not previously detected in each multigene family; (2) that there are unidentified pathway components that discriminate between type I and type II receptors; (3) that there are more Smads to be discovered in Drosophila and mammals; and (4) that the number of C-terminal serines is the best predictor of a Smad's role in TGF-ß signal transduction. We discuss these findings with respect to the coevolution of physically interacting genes.
INTERCELLULAR signaling by growth factors is essential for proper pattern formation in metazoan development. Secreted proteins of the transforming growth factor-ß (TGF-ß) family regulate key developmental events in many organisms. For example, TGF-ß family members induce the establishment of the left-right body axis in mammals (
The mechanism by which a specific family member elicits a particular developmental response is still unclear. The current working model for TGF-ß family signaling pathways is shown in Figure 1 (reviewed in
|
|
Structurally, the Smad multigene family is characterized by N-terminal (MH1) and C-terminal (MH2) Mad homology domains, which are well conserved across species (reviewed in
Members of the receptor and Smad multigene families are also highly conserved in distant species. The functional conservation of TGF-ß receptors is impressive. For example, a complex containing a type I receptor from D. melanogaster and a type II receptor from C. elegans binds a human ligand with high affinity (
In this report, we utilize a phylogenetic approach to test hypotheses that address evolutionary and developmental questions about TGF-ß signaling pathways. Our focus is on the amino acid sequence relationships within and between the multigene families encoding ligands, receptors, and Smads. For example, we tested the hypothesis that the duplication and divergence of ligands drives the duplication and divergence of receptors and Smads. We included the newest family members in the analyses, many of whom have not been experimentally examined, to provide clues to our fellow investigators. The relationships we identify shed new light on developmental functions for individual family members. For example, our data suggest a number of new hypotheses about the mechanisms underlying the signaling specificity of receptors.
| MATERIALS AND METHODS |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Database searches and amino acid sequence predictions:
As of October 15, 1998, all available full-length amino acid sequences of TGF-ß family ligands, receptors, and Smad proteins were obtained by BLAST. For published sequences we used the National Center for Biotechnology Information web site (www.ncbi.nlm.nih.gov/BLAST/). For additional Drosophila sequences we used the Berkeley Drosophila Genome Project and the European Drosophila Genome Project web pages (fruitfly.berkeley.edu/blast/ and edgp.ebi.ac.uk/www-blast.html/). For additional C. elegans sequences we used an AXYS Pharmaceuticals BLAST server. To obtain new C. elegans and Drosophila amino acid sequences the GeneScan (
Sequence alignments and phylogenetic analyses:
Amino acid sequences were aligned using CLUSTAL V (
| RESULTS |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Phylogenetic trees generated by the neighbor joining, maximum parsimony, and Fitch-Margoliash methods were largely congruent. As expected (
75% bootstrap support in drawing inferences for two reasons. First, the bootstrap method is conservative (
Phylogenetic analysis of ligands:
Figure 2 shows a phylogeny of TGF-ß ligands. Several of the ligands are open reading frame predictions from sequences submitted by genome projects. These are indicated in Figure 2 by clone numbers rather than gene names. For example, we predict a new family member in Drosophila from genomic clone DS07149.
|
, SWISS-PROT Q04997; nodal, SWISS-PROT P43021; GDNF, SWISS-PROT P48540; dorsalin, PIR A40735; ADMP, GenBank
In its overall topology, our analysis agrees with previous studies (e.g.,
Phylogenetic analysis of type I receptors:
Figure 3 shows a phylogeny of type I receptors. The rate of amino acid substitution for type I receptors is roughly the same as for the ligands. The inferred relationships within and between clusters A, B, and C are well supported. Interestingly, both C. elegans type I receptors, C32D5.2 (Sma-6;
|
The strength of the clustering of the Drosophila Dpp receptor Sax with human ALK1 and ALK2 (Figure 3, cluster A) was unexpected. ALK1 has recently been shown to be a TGF-ß receptor (P. TEN DIJKE, unpublished data) and ALK2 is an activin receptor (
The strength of the relationships in cluster B was also unexpected. A relationship between DmATR-1 and ALK4 and ALK5 was previously shown and each of these receptors can bind activin in vitro (
The clustering of the other Dpp type I receptor Thickveins (Tkv) with the two vertebrate BMP receptors (ALK3 and ALK6;
Phylogenetic analysis of type II receptors:
Figure 4 shows a phylogeny of type II receptors. The rate of amino acid substitution is roughly the same for ligands and type I receptors. The evolutionary relationships within clusters A and B are well supported. Again a receptor from C. elegans (Daf-4) forms a distinct lineage.
|
Within cluster A, a new type II receptor from Drosophila called Wishful thinking (Wit; M. O'CONNOR, unpublished data) clusters strongly with BMPR-II, suggesting that it may signal for one of the Drosophila Dpp/BMP subfamily ligands (Dpp, 60A, or Screw). The close relationship between these two receptors and MIST-II was unexpected. The relationship in Figure 4, cluster B, between Punt (previously known as DmATR-II) and ActR-IIA/ActR-IIB (activin receptors) has been widely reported. Punt was shown to bind activin in vitro (
Phylogenetic analysis of the cytoplasmic domain of all receptors:
Given that the two ends of the receptors have completely distinct environments and roles in TGF-ß pathways (ligand binding vs. kinase signaling), we wondered if there were any difference in the evolutionary forces affecting the extracellular and the cytoplasmic domains of the receptors. To address this issue we conducted an analysis of each domain alone. The rate of amino acid substitution in the cytoplasmic domain is roughly the same as the entire receptor. The topology of the cytoplasmic tree is identical to the trees of type I receptors and type II receptors with minor differences in bootstrap values. Note that Figure 5, clusters A, B, and C, are the same as Figure 3, clusters A, B, and C, and that Figure 5, clusters D and E, are the same as Figure 4, clusters A and B.
|
The overall picture gained from this analysis is that the type I and type II receptors, both transmembrane serine-threonine kinases, form a single large lineage. The tree does not bifurcate into two monophyletic groups representing each receptor type. Within our tree the type I receptors form a group excluding type II receptors. This likely reflects the distinct roles of the receptors (type I, phosphorylate Smads; and type II, phosphorylate type I receptors). The relative placement of the receptors in our tree suggests that type I receptors diverged from an ancestral type II receptor. The C. elegans type I receptors fall at the boundary of type I and type II receptors, perhaps representing a transitional receptor type.
Phylogenetic analysis of the extracellular domain of all receptors:
In general, the phylogeny of the extracellular domain (Figure 6) has low bootstrap values on many of the branchpoints. This suggests that there is more divergence between the receptor sequences on the ligand-binding side than on the signaling side. The rate of amino acid substitution in the extracellular domain is roughly the same as for the entire receptor. With two exceptions, the overall topology of the relevant portion of the extracellular tree (top half of Figure 6) is congruent with the tree of type I receptors (Figure 3). One exception is that Sax breaks the monophyly of ALK1 and ALK2 (Figure 6, cluster A). However, this unique relationship is not strongly supported. Another exception is the secondary connection of the DmATR-1 cluster (Figure 6, cluster B) with the Tkv cluster (Figure 6, cluster C). In the type I tree the DmATR-1 cluster is closest to the Sax cluster. Compare Figure 3 (clusters A and B) with Figure 6 (clusters B and C). While the bootstrap values of these secondary/tertiary clusters are low, the CP values are high (cluster B and C node: CP = 86; cluster A node with cluster B/C: CP = 85).
|
A number of terminal clusters in the extracellular tree (Figure 6) are different from the tree of type II receptors (Figure 4). In the extracellular tree, Wit is clustered with MIST-II (CP = 91; Figure 6, cluster D) instead of with BMPR-II (Figure 4, cluster A). The two activin receptors and BMPR-II are clustered (CP = 76; Figure 6, cluster E) instead of being distantly related (Figure 4, cluster A vs. cluster B). In Figure 6, Punt is distinct from a node containing clusters D and E rather than branching specifically from the two activin receptors I (Figure 4, cluster B).
Just as in the cytoplasmic tree (Figure 5), the type I receptors form a group of closely related sequences exclusive of the type II receptors. Why do type I and type II receptors that bind the same ligand not cluster together? Note the phylogenetic distance between the Dpp-signaling heteromeric partners Tkv/Punt and Sax/Punt. Perhaps there are factors that bind to only one receptor type that prevent the sequence convergence of type I and type II receptors with a common ligand.
Phylogenetic analysis of Smad signal transducers:
Figure 7 shows a phylogeny of Smad signal transducers. Note that the rate of amino acid substitution in the Smad family is roughly 2.5-fold faster than for ligands and receptors. The analysis includes several Smads whose relationships to other family members have not been reported. For example, DSmad2, a new family member in Drosophila (
|
In its overall topology, our tree generally agrees with previous reports demonstrating the existence of three distinct subfamilies (e.g.,
Other family members derive from open reading frame predictions generated by the recently completed C. elegans genome project. Two of these predictions have been connected to genes. These are R05D11.1, which is Daf-8 (D. RIDDLE, unpublished data), and F01G10.8, which is Daf-14 (J. THOMAS, unpublished data). These C. elegans Smads do not show close evolutionary relationships to other family members. As unique lineages, both Daf-8 and Daf-14 fall roughly between cluster D and cluster E (Figure 7). We predict that F37D6.a is another family member in C. elegans. This open reading frame is mispredicted in the database (see Figure 7 legend). F37D6.a clusters well with antagonist Smads. Thus, C. elegans has 7 Smad family members. The fact that there are 8 mammalian Smads and 4 Smads in Drosophila suggests that there are more Smads to be found in these experimental systems. This proposal is supported by comparing the number of mammalian ligands (24) with those in nematodes (4) as shown in Figure 2.
In many new multigene families the identification of the same gene in different species (homologs) is often difficult. This results in the occasional misidentification of genes. We noted one such instance during our analysis involving Xenopus Smad8 (
Phylogenetic analysis of the conserved domains of Smad signal transducers:
The Smad family is characterized by N-terminal (MH1) and C-terminal (MH2) domains that are well conserved between species. Given that the two domains appear to have distinct roles in signal transduction (MH1, DNA binding/transcriptional activation; MH2, Smad complex formation) we wondered if there were any difference in the evolutionary forces affecting the two domains. To address this issue we conducted an analysis of each domain alone. As with the receptors above, we were interested to see whether the phylogenetic relationships among the individual domains showed any differences from each other or from the Smad tree.
The role of the MH2 domain in Smad complex formation is a highly conserved function in all Smad family members (reviewed in
The MH1 domain shows wide sequence variation among Smad family members. Figure 8 shows our alignment of this domain using 19 Smad sequences. In the alignment, some portion of the MH1 domain (particularly amino acid numbers 150–160) is recognizable in every sequence except for Daf-14, F37D6.a, and Dad. Smad6 aligns well and is not missing all or part of the MH1 domain as previously reported. The domain is divided into subregions by unique insertions in a number of Smads. The biological role of these insertions, if any, is unknown.
|
There are only two absolutely invariant amino acids in the MH1 domain. One is cysteine 45 and the other is proline 153 of the alignment. These amino acids are affected in mutant alleles of Med (
A phylogeny generated from the MH1 alignment is shown in Figure 9. Several clusters are present in the Smad and the MH1 trees. For example, Figure 9, clusters A (Med), B (DSmads), and E (signal transducing), are the same as Figure 7, clusters A, B, and E. There are several differences between the trees. In the MH1 tree, Daf-8 and Daf-14 form a cluster and appear more divergent than in the Smad and MH2 trees. Their placement near the outgroup in the MH1 tree likely reflects their lack of (Daf-14) or a minimal (Daf-8) MH1 domain. Interestingly, F37D6.a, which also does not have an obvious MH1 domain, does not move away from its strong secondary cluster (Figure 7 and Figure 9, cluster D) with the antagonists Smad6 and Smad 7, both of which have recognizable MH1 domains. Daf-3 clusters very strongly with the Med subfamily in the MH1 tree (Figure 9, cluster C) instead of as a highly divergent monophyletic lineage. Interestingly, both sma-2 and sma-3 move into cluster C (Mad) in the MH1 tree.
|
| DISCUSSION |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Our phylogenetic analyses of three physically interacting multigene families involved in TGF-ß signaling provide a number of new insights into the molecular evolution and developmental biology of intercellular communication. One valuable set of observations is derived from our studies of the newest family members, many of whom have not been experimentally examined. These include seven new ligands, two new receptors, and five new Smads. Another set of results with wide relevance are the relationships we detect between sequences not previously connected to each other, such as the clustering of the type I receptors Sax/ALK1/ALK2.
Calibrating our phylogenies for ligands (Figure 2) and Smads (Figure 7) with an arthropod-nematode divergence of 1.1 billion years ago (
One important evolutionary question that can be addressed by our data is whether the duplication and divergence of ligands drives the duplication and divergence of receptors and Smads. Our data indicate that the answer to this question is complex. Examination of the recent history of the receptor and Smad families (e.g., terminal clusters) indicates that each appears to be evolving independently or under the influence of factors other than the ligand. Evidence for this comes from three sources: the clustering of receptor extracellular domains by type and not by ligand and the clustering of Smads with different ligands or different roles.
Early in the history of TGF-ß signaling, ligand duplication and subsequent sequence diversification may have been a powerful force in shaping the molecular evolution of receptors and Smads. Our data show no convincing evidence for a TGF-ß/activin subfamily signaling pathway in C. elegans. The bootstrap values for branches placing Daf-7 in the TGF-ß/activin subfamily are extremely low (Figure 2) and there are no nematode receptors (Figure 3 and Figure 4) or Smads (Figure 7) in clusters known to signal for TGF-ß/activin subfamily ligands. Two explanations are possible. First, the TGF-ß/activin subfamily arose after the divergence of nematodes and arthropods. Second, a preexisting TGF-ß/activin subfamily signaling pathway was lost in the C. elegans lineage. We suggest that TGF-ß/activin subfamily signaling arose after the separation of arthropods and nematodes but before the separation of arthropods and vertebrates (1.1 billion vs. 950 million years ago;
A global phylogenetic tree containing a clade of arthropods and nematodes diverging from a common ancestor with vertebrates (
A subset of TGF-ß signaling pathways in nematodes may approximate an ancestral TGF-ß pathway. Evidence for this hypothesis comes from our analysis of C. elegans receptors. Two receptors, Daf-1 (type I;
The possibility that Daf-1 and Daf-4 resemble an ancestral "nonspecialized" receptor type led us to compare the literature on other TGF-ß family receptors with our data. Perhaps we could identify other receptors that appear "less specialized." Given that receptors have two domains, there are at least three ways in which a receptor can be "nonspecific." A receptor could have relationships with two ligands, two Smads, or bind a single ligand but signal through another ligand's Smad. We identified two such receptors in Drosophila.
The type I receptor Sax signals for Dpp in vivo, although its role in Dpp signaling is less significant than Tkv (
A type II receptor with multiple reported ligands is Punt. This receptor binds activin with high affinity in vitro, but subsequently was shown to function in Dpp signaling in vivo. In the type II tree (Figure 4) and the cytoplasmic tree (Figure 5), Punt clusters with activin receptors with a bootstrap value of 100%. However, in the extracellular tree (Figure 6), Punt appears as a divergent lineage. The data suggest that Punt binds ligands from both subfamilies and is able to signal to both Dpp/BMP and TGF-ß/activin subfamily type I receptors. Overall, our data for receptors suggest more "non-specificity" in ligand binding and more flexibility in signaling than had been suspected previously.
Given this potential versatility of individual receptors, how are specific instructive signals transduced? One possibility is that additional pathway components play an important role in ensuring signal specificity. Members of the recently identified SARA family of cytoplasmic proteins (
The Smad family appears to be evolving independently of the ligands and receptors. Perhaps Smad molecular evolution is instead influenced by interactions with other TGF-ß signaling pathway components such as the SARA proteins or transcriptional factors. Alternatively, Smad family evolution may be driven by interactions with components of other signaling pathways. For example, a recent study in Drosophila (
Smad family members do not cluster by ligand as evidenced by the grouping of Smad8 (TGF-ß/activin signaling) with Smad1 and Smad5 (BMP signaling) in the Smad tree (Figure 7). Nor do they cluster by role as evidenced by the grouping of the antagonist Daf-3 with Smads that signal for multiple ligands such as Smad4 in the MH1 tree (Figure 9). A comparison of our Smad tree (Figure 7) and the MH2 tree (data not shown) indicates that the relationships identified in the Smad tree are almost universally reflective of those in the MH2 tree. The MH2 domain facilitates the formation of complexes between Smads. The extensive similarity between the Smad and MH2 trees suggests that the ability to form multi-Smad complexes is the most fundamental feature of Smad function.
In Smads with different roles the most informative sequence difference is the number of C-terminal serines in the MH2 domain. There are two or three serines in Smads dedicated to one ligand and zero or one serine in Smads signaling for multiple ligands and antagonist Smads. The absence of the MH1 domain is not an accurate predictor of Smad function. The C. elegans antagonist Daf-3 has an identifiable MH1 domain and one C-terminal serine, while the positively signaling Daf-14 has no MH1 domain and two serines. The function of the MH1 domain appears to be DNA binding/transcriptional activation in both ligand-specific and multiple-ligand Smads. This common function is reflected in the Smad (Figure 7; CP = 95) and MH1 trees (Figure 9). Both trees have secondary clusters of ligand-specific (e.g., Mad; cluster E) and multiple-ligand Smads (e.g., Smad4; cluster A).
In conclusion, we hope that these results will stimulate new experimental directions for both evolutionary and developmental biologists. As new multigene families that participate in TGF-ß signaling are discovered we plan to extend our analyses to test hypotheses reported here. We believe this study illustrates the value of applying statistical methods in the analysis of developmentally important signaling pathways.
| ACKNOWLEDGMENTS |
|---|
We thank M. O'Connor, D. Riddle, P. ten Dijke, and J. Thomas for communicating data prior to publication. S.K. thanks Chandra Laneback for technical assistance. Supported in part by institutional funds from Arizona State University (S.K.) and a Basil O'Connor Starter Scholar Research Award from the March of Dimes and a Research Incentive Award from Arizona State University to S.J.N.
Manuscript received December 3, 1998; Accepted for publication March 2, 1999.
| LITERATURE CITED |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
AGUINALDO, A., J. TURBEVILLE, L. LINFORD, M. RIVERA, and J. GAREY et al., 1997 Evidence for a clade of nematodes, arthropods and other moulting animals. Nature 387:489-493[Medline].
BRUMMEL, T. J., V. TWOMBLY, G. MARQUES, J. WRANA, and S. J. NEWFELD et al., 1994 Characterization and relationship of Dpp receptors encoded by the saxophone and thickveins genes in Drosophila.. Cell 78:251-262[Medline].
BRUMMEL, T. J., S. ABDOLLAH, T. HAERRY, M. SHIMELL, and J. MERRIAM et al., 1999 The Drosophila activin receptor Baboon signals through DSmad2 and controls cell proliferation but not patterning during larval development. Genes Dev. 13:98-111
BURGE, C. and S. KARLIN, 1997 Prediction of complete gene structures in human genomic DNA. J. Mol. Biol. 268:78-94[Medline].
BURT, D. and A. LAW, 1994 Evolution of the TGF-ß superfamily. Prog. Growth Factor Res. 5:99-118[Medline].
CHEN, Y., A. BHUSHAN, and W. VALE, 1997 Smad8 mediates the signaling of the receptor serine kinase. Proc. Natl. Acad. Sci. USA 94:12938-12943
CHILDS, S., J. WRANA, K. ARORA, L. ATTISANO, and M. O'CONNOR et al., 1993 Identification of a Drosophila activin receptor. Proc. Natl. Acad. Sci. USA 90:9475-9479
COLLIGNON, J., I. VARLET, and E. ROBERTSON, 1996 Relationship between asymmetric nodal expression and the direction of embryonic turning. Nature 381:155-158[Medline].
DERYNCK, R., Y. ZHANG, and X. FENG, 1998 Smads: transcriptional activators of TGF-ß responses. Cell 95:737-740[Medline].
DOPAZO, J., 1994 Estimating errors and confidence intervals for branch lengths in phylogenetic trees by a bootstrap approach. J. Mol. Evol. 38:300-304[Medline].
EPPERT, K., S. SCHERER, H. OZCELIK, R. PIRONE, and P. HOODLESS et al., 1996 MADR2 maps to 18q21 and encodes a TGF-ß regulated MAD-related protein that is functionally mutated in colorectal carcinoma. Cell 86:543-552[Medline].
ESTEVEZ, M., L. ATTISANO, J. WRANA, P. ALBERT, and J. MASSAGUÉ et al., 1993 The daf-4 gene encodes a BMP receptor controlling C. elegans dauer larva development. Nature 365:644-649[Medline].
FELSENSTEIN, J., 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791.
FELSENSTEIN, J., 1993 PHYLIP version 3.5c. University of Washington, Seattle.
FITCH, W. and E. MARGOLIASH, 1970 The usefulness of amino acid and nucleotide sequences in evolutionary studies. Evol. Biol. 4:67-109.
FRISCH, A. and C. WRIGHT, 1998 XBMPR-II, a novel Xenopus Type II receptor mediating BMP signaling in embryonic tissues. Development 125:431-442[Abstract].
GEORGI, L., P. ALBERT, and D. RIDDLE, 1990 daf-1: a C. elegans gene controlling dauer larva development encodes a novel receptor protein kinase. Cell 61:635-645[Medline].
GRAFF, J., R. THIES, J. SONG, A. CELESTE, and D. MELTON, 1994 Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo.. Cell 79:169-179[Medline].
GRAFF, J., A. BANSAL, and D. MELTON, 1996 Xenopus MAD proteins transduce distinct subsets of signals for the TGF-ß family. Cell 85:479-487[Medline].
HAERRY, T., O. KHALSA, M. O'CONNOR, and K. WHARTON, 1998 Synergistic signaling by two BMP ligands through the SAX and TKV receptors controls wing growth and patterning in Drosophila. Development 125:3977-3987[Abstract].
HAHN, S., M. SCHUTTE, A. HOQUE, C. MOSKALUK, and L. DA COSTA et al., 1996 DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271:350-353[Abstract].
HIGGINS, D., A. BLEASBY, and R. FUCHS, 1992 CLUSTAL V: improved software for multiple sequence alignment. Comput. Appl. Biosci. 8:189-191
HILLIS, D. and J. BULL, 1993 An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42:182-192.
HOODLESS, P., T. HAERRY, S. ABDOLLAH, M. STAPLETON, and M. O'CONNOR et al., 1996 MADR1, a Mad-related protein that functions in BMP2 signaling pathways. Cell 85:489-500[Medline].
IMAMURA, T., M. TAKASE, A. NISHIHARA, E. OEDA, and J. HANAI et al., 1997 Smad6 inhibits signaling by the TGF-ß superfamily. Nature 389:622-626[Medline].
JACKSON, S., H. NAKATO, M. SUGIURA, A. JANNUZI, and R. OAKES et al., 1997 dally, a Drosophila glypican controls cellular responses to the TGF-ß related morphogen Dpp. Development 124:4113-4120[Abstract].
JING, S., D. WEN, Y. YU, P. HOLST, and Y. LUO et al., 1996 GDNF-induced activation of the ret protein tyrosine kinase is mediated by a novel receptor for GDNF. Cell 85:1113-1124[Medline].
JONES, C., M. KUEHN, B. HOGAN, J. SMITH, and C. WRIGHT, 1995 Nodal-related signals induce axial mesoderm and dorsalize mesoderm in gastrulation. Development 121:3651-3662[Abstract].
KINGSLEY, D., 1994 The TGF-ß superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 8:133-146
KOENIG, B., J. COOK, D. WOLSING, J. TING, and J. TIESMAN et al., 1994 Characterization and cloning of a receptor for BMP2 and BMP4 from NIH 3T3 cells. Mol. Cell. Biol. 14:5961-5974
KRISHNA, S., L. MADUZIA, and R. PADGETT, 1999 Specificity of TGF-ß signaling is conferred by distinct Type I receptors and their associated Smad proteins in C. elegans.. Development 126:251-260[Abstract].
KUMAR, S., K. TAMURA and M. NEI, 1993 MEGA: Molecular Evolutionary Genetic Analysis. Pennsylvania State University Press, University Park, PA.
KUTTY, G., R. KUTTY, W. SAMUEL, T. DUNCAN, and C. JAWORSKI et al., 1998 Identification of a new member of the TGF-ß family in Drosophila: the first invertebrate activin gene. Biochem. Biophys. Res. Commun. 246:644-649[Medline].
LEVIN, M., R. JOHNSON, R. STERN, M. KUEHN, and C. TABIN, 1995 A molecular pathway determining left-right asymmetry in chick embryogenesis. Cell 82:803-814[Medline].
LIN, L., D. DOHERTY, J. LILE, S. BEKTESH, and F. COLLINS, 1993 GDNF: a glial cell line derived neurotrophic factor for midbrain dopaminergic neurons. Science 260:1130-1132
LIU, F., F. VENTURA, J. DOODY, and J. MASSAGUÉ, 1995 Human type II receptor for BMPs: extension of the two-kinase receptor model to the BMPs. Mol. Cell. Biol. 15:3479-3486[Abstract].
MASSAGUÉ, J., L. ATTISANO, and J. WRANA, 1994 The TGF-ß family and its composite receptors. Trends Cell Biol. 4:172-178[Medline].
MIYA, T., K. MORITA, A. SUZUKI, N. UENO, and N. SATOH, 1997 Functional analysis of an ascidian homolog of vertebrate BMP2/BMP4 suggests its role in the inhibition of neuralfate specification. Development 124:5149-5159[Abstract].
NAKAYAMA, T., M. SNYDER, S. GREWEL, K. TSUNEIZUMI, and T. TABATA et al., 1998 Xenopus Smad8 acts downstream of BMP4 to modulate its activity during vertebrate embryonic patterning. Development 125:857-867[Abstract].
NEI, M., S. KUMAR, and K. TAKAHASHI, 1998 The optimization principle in phylogenetic analysis tends to give incorrect topologies when the number of nucleotides or amino acids used is small. Proc. Natl. Acad. Sci. USA 95:12390-12397
NEWFELD, S. J. and W. M. GELBART, 1995 Identification of two Drosophila TGF-ß family members in the grasshopper Schistocerca americana.. J. Mol. Evol. 41:155-160[Medline].
NEWFELD, S. J., E. CHARTOFF, J. GRAFF, D. MELTON, and W. GELBART, 1996 Mothers against dpp encodes a conserved cytoplasmic protein required in DPP/TGF-ß responsive cells. Development 122:2099-2108[Abstract].
NGUYEN, M., S. PARK, G. MARQUES, and K. ARORA, 1998 Interpretation of a BMP activity gradient in Drosophila embryos depends on synergystic signaling by two Type I receptors, SAX and TKV. Cell 95:495-506[Medline].
PADGETT, R., J. WOZNEY, and W. GELBART, 1993 Human BMP sequences confer normal dorsal ventral patterning in the Drosophila embryo. Proc. Natl. Acad. Sci. USA 90:2905-2909
PATTERSON, G., A. KOWEEK, A. WONG, Y. LIU, and G. RUVKUN, 1997 The daf-3 Smad protein antagonizes TGF-ß related receptor signaling in the C. elegans dauer pathway. Genes Dev. 11:2679-2690
POSAKONY, L. G., L. A. RAFTERY, and W. M. GELBART, 1991 Wing formation in Drosophila requires decapentaplegic function along the anterior-posterior compartment boundary. Mech. Dev. 33:69-82