Previous Article |
Table of Contents
| Next Article 
CELL BIOLOGY
Identification of the Wnt signaling activator leucine-rich repeat in Flightless interaction protein 2 by a genome-wide functional analysis
, 
*Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037; The Salk Institute for Biological Studies, P.O. Box 85800, La Jolla, CA 92186; and Genomics Institute of The Novartis Research Foundation, 10675 John Jay Hopkins Drive, San Diego, CA 92121
Contributed by Peter G. Schultz, December 17, 2004
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResults and DiscussionConclusionsReferences |
|---|
The Wnt signaling pathway acts ubiquitously in metazoans to control various aspects of embryonic development. Wnt ligands bind their receptors Frizzled and low-density lipoprotein receptor-related protein 5/6 and function through Disheveled (Dvl), Axin, adenomatous polyposis coli, glycogen synthase kinase 3
, and casein kinase (CK) 1 to stabilize -catenin and induce lymphocyte enhancer-binding factor (LEF)/T cell factor (TCF)-dependent transcriptional activities. To identify previously unrecognized Wnt signaling modulators, a genome-wide functional screen was performed using large-scale arrayed cDNA collections. From this screen, both known components and previously uncharacterized regulators of this pathway were identified, including -catenin, Dvl1, Dvl3, Fbxw-1, Cul1, CK1
, CK1
, and 
-catenin. In particular, a previously unrecognized activator, LRRFIP2 (leucine-rich repeat in Flightless interaction protein 2), was found that interacts with Dvl to increase the cellular levels of -catenin and activate -catenin/LEF/TCF-dependent transcriptional activity. The function of LRRFIP2 is blocked when a dominant negative Dvl (Xdd1) is coexpressed. Expression of LRRFIP2 in Xenopus embryos induced double axis formation and Wnt target gene expression; a dominant negative form of LRRFIP2 suppresses ectopic Wnt signaling in Xenopus embryos and partially inhibits endogenous dorsal axis formation. These data suggest that LRRFIP2 plays an important role in transducing Wnt signals.
genome-wide functional screen | Xenopus | Wnt signaling pathway | Disheveled
-catenin (1–5). In the absence of Wnt signaling, the cytoplasmic level of -catenin is kept low through its interaction with a protein complex comprised of APC, Axin, glycogen synthase kinase 3 (GSK-3), and casein kinase (CK) 1 (1, 2). Phosphorylation of -catenin by CK1 and GSK-3 results in its ubiquitination by the Skp1–Cul1–Fbox ubiquitin-conjugation complex and subsequent degradation by means of the proteasome (1, 2). Binding of the Wnt ligands to their receptors, Frizzled and low-density lipoprotein receptor-related protein 5/6, leads to activation of the adaptor protein, Disheveled (Dvl), and the inhibition of GSK-3, resulting in the accumulation of cytoplasmic -catenin. At elevated cytoplasmic levels, -catenin translocates to the nucleus, where it interacts with members of the T cell factor (TCF)/lymphocyte enhancer-binding factor (LEF) family of DNA-binding proteins. -Catenin binding to TCF/LEF converts it from a transcriptional repressor to an activator of Wnt target genes such as c-Myc, cyclin D1, and multiple developmental regulators (1, 2).
Although many of the core components required for canonical Wnt signaling are known, many details of how signals are passed from one component to the next remain obscure. For example, the cellular signal transducer Dvl is one of the key components in the pathway (1, 2), yet it is not known how Dvl is activated upon binding of Wnt ligand to its receptors (6–8) or, likewise, how Dvl transduces the signal by inhibiting the downstream effector, GSK-3. Similarly, certain Wnt ligands activate other downstream pathways, some of which also involve Dvl, but how these pathways diverge from or interact with the canonical pathway is largely unknown. Consequently, the identification of additional components that may either be essential for Wnt signaling or modulate the activity of the core components is critical to a better understanding of the role of the Wnt pathway in development and cancer. The recent availability of genome sequences of human and model organisms has provided an opportunity to study signaling pathways by using functional genome-wide approaches, including screening of large-scale small interfering RNA and cDNA collections (9–14).
To identify additional components and modulators of the Wnt signaling pathway, we have screened a large arrayed and addressable cDNA library (
20,000 cDNAs) for clones that modulate -catenin-dependent activities in cultured cells. Clones isolated in this screen encode known components of the Wnt pathway as well as previously unrecognized potential regulators of Wnt signaling. In particular, one of the clones encodes human LRRFIP2 (leucine-rich repeat in Flightless interaction protein 2), a protein of previously unknown function. We show that LRRFIP2 is able to activate the canonical Wnt pathway in both cultured mammalian cells and in Xenopus embryos. LRRFIP2 binds to Dvl, and the region involved in binding is required for its activity both in cultured cells and in embryos. Finally, we show that a putative dominant negative form of LRRFIP2 can suppress ectopic Wnt signaling in Xenopus embryos and partially inhibits endogenous dorsal axis formation. These data suggest strongly that LRRFIP2 is a previously unrecognized component of the Wnt signaling pathway that mediates or modulates Wnt signaling through interactions with Dvl.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResults and DiscussionConclusionsReferences |
|---|
20,000 cDNAs from the Origene collection and Mammalian Genome Collection as described in refs. 9, 10, and 12. Briefly, each individual cDNA plasmid (62.5 ng) in the collection was spotted into a single well of 384-well cell culture plates. Serum-free DMEM (20 µl) containing a -catenin/TCF/LEF-responsive TOPflash reporter (Upstate Biotechnology, Lake Placid, NY) and FuGENE6 (Roche Diagnostics) was allocated into each well before transfection. After incubation at room temperature for 30 min, 20% FBS DMEM (20 µl) containing 104 HEK293T cells was plated into each well. Cells were then cultured in a humidified incubator at 37°C in 5% CO2. After 48 h, BrightGlo (Promega) reagent (40 µl) was added to each well, and luciferase luminescence was measured with an Acquest plate reader (LJL Biosystems, Sunnyvale, CA). Constructs. Human Dvl3 was cloned by RT-PCR from human embryonic kidney 293T cells and inserted into a Myc-tagged mammalian expression vector, pMyc-cyto (Invitrogen). The Dvl mutants were generated by PCR amplification and inserted into the same vector. The mutant constructs of human LRRFIP2 were generated with PCRs and cloned into the pCS2-Flag or pMyc-cyto vectors. Xenopus LRRFIP2 was cloned by RT-PCR from the stage-20 embryos and inserted into the pCS2-Flag vector. The sequences of the primer sets used in this report are available upon request.
Immunoprecipitation and Western Blot Analysis. Coimmunoprecipitation and Western blot analysis were performed as described in ref. 15. Briefly, whole-cell lysates were pretreated with mouse IgG and protein G agarose. Flag-tagged proteins were immunoprecipitated from supernatants by mouse anti-Flag monoclonal antibody (M2, Sigma) conjugated to agarose (Sigma). For immunoblots, immunoprecipitants or cellular proteins were mixed with Laemmli SDS sample buffer. Total cellular proteins (40 µg per lane) were separated by 4–20% gradient Tris-glycine SDS/PAGE (Invitrogen) and transferred to a nitrocellulose membrane. Proteins were detected with primary antibodies and horseradish peroxidase-conjugated secondary antibodies by using an enhanced chemiluminescence kit (Amersham Biosciences). Primary antibodies used were mouse anti-c-Myc monoclonal antibody (9E10, Santa Cruz Biotechnology), mouse anti-Flag monoclonal antibody (M2, Sigma), mouse anti--tubulin monoclonal antibody (Sigma), and mouse anti--catenin monoclonal antibody (C-19220, Transduction Laboratories, Lexington, KY).
Xenopus Microinjection. Xenopus eggs were in vitro fertilized following established methods (16). Embryos were raised in 0.1x Marc's modified Ringer's solution and staged according to the methods of Nieuwkoop and Faber (17). Capped RNAs were transcribed in vitro by using the mMessage-mMachine kit (Ambion, Austin, TX) according to the manufacturer's instructions. Synthetic RNAs were microinjected into the two- or four-cell-stage embryos. After 48 h, the axis duplication phenotype was scored. For the animal cap explant experiment, synthetic RNAs were microinjected into the animal pole of two-cell stage embryos. At stage 8.5, animal caps were dissected and allowed to develop to stage 10.5. Explants were harvested, and RNAs were extracted by using TRIzol (Invitrogen). The first-strand cDNAs were synthesized by using the SuperScript III First Strand System (Invitrogen) according to the manufacturer's instructions. One to two microliters of the cDNA-synthesizing reaction was used for PCRs. The sequences of the primers used for RT-PCR are available upon request. Histology section and staining were performed as described in ref. 16.
|
Results and Discussion |
|---|
|
TopAbstractMaterials and MethodsResults and DiscussionConclusionsReferences |
|---|
20,000 cDNAs (described in detail in Materials and Methods). Briefly, cDNAs were prespotted on 384-well cell culture plates as individual clones and then assayed by transfection into HEK293T cells along with a TOPflash reporter. The level of TOPflash reporter activity was used as an indicator of TCF/LEF-dependent transcriptional activity and, indirectly, -catenin levels (18). Each well that scored positive from a luminescence readout of reporter luciferase activity was retested on both TOPflash and on a negative control FOPflash construct carrying mutations in the -catenin/TCF-responsive elements (18).
A variety of known components/modulators of the Wnt pathway were identified in the screen (Table 1, which is published as supporting information on the PNAS web site), including -catenin, Dvl1, Dvl3, Fbxw-1, CK1, CK1, and -catenin (1, 2). Cul1 and the ubiquitin-conjugating enzyme E2, which are components of the Skp1–Cul1–Fbox complex that regulates the ubiquitination of -catenin (19), also were identified. In addition, the screen identified previously unrecognized proteins that have not been associated with -catenin regulation and/or Wnt signaling, including KIAA1190, ecotropic viral insertion site 3 (Evi3), Ring finger gene 4 (Rnf4), and LRRFIP2 (Table 1). KIAA1190 is an uncharacterized protein containing eight putative zinc fingers. Evi3, which contains 30 zinc fingers, is frequently up-regulated in mouse B cell lymphomas and functions in B cell development and oncogenesis (20). Rnf4 is a Ring finger-containing factor that is involved in transcription factor activation, nuclear trafficking, and ubiquitin conjugation (21–23). These proteins may play a role in -catenin nuclear translocation and formation of active TCF/LEF transcription complexes.
The clone encoding LRRFIP2 was one of the strongest positives identified in the screen, inducing the TOPflash reporter activity 20- to 30-fold over background and, therefore, was characterized more fully. Human LRRFIP2 was originally identified in a two-hybrid screen using human Flightless-I (24), an actin-binding protein (25), as a bait. However, the significance of this potential interaction and the genetic and biochemical function of LRRFIP2 are unknown. LRRFIP2 was shown to be expressed as a long and short form, representing alternatively spliced variants (24). Although the screen only identified the long form, subsequent analysis showed that the short form has similar albeit weaker activity in the TOPflash reporter assay. The obvious structural features of LRRFIP2 are a predicted coiled-coil domain at its carboxyl terminus (24) and a Ser-rich region at the amino terminus, a potential site for posttranslational modification. A search of public databases did not reveal homologous orthologs of LRRFIP2 in either Caenorhabditis elegans or Drosophila, although the genes CG8578-PA (accession no. AAF48525 [GenBank] in Drosophila and F57B9.7 (accession no. AAM48538 [GenBank] in C. elegans are partially homologous to the carboxyl terminus of LRRFIP2. However, we were able to isolate a cDNA encoding the long form of Xenopus LRRFIP2 from an embryonic cDNA library by PCR. Alignment of the frog and human proteins indicates a high level of overall amino acid conservation (75%), with particularly high levels in the amino terminus (92%) as well as in the predicted coiled-coiled domain (94%) (Fig. 6, which is published as supporting information on the PNAS web site), suggesting that these regions encode conserved essential functions.
Overexpression of LRRFIP2 in HEK293T cells robustly activates the TOPflash reporter activity in a dose-dependent manner (Fig. 1A). The specificity of this activation was confirmed by using the negative control FOPflash reporter as well as other reporter constructs, such as a p53-responsive reporter and a sonic hedgehog-responsive reporter (data not shown). No activity was detected with LRRFIP2 using either of these reporters. To test whether LRRFIP2 activates the canonical Wnt signaling pathway by modulating the level of free -catenin (1, 2), LRRFIP2 was transfected into HEK293T cells and the levels of cytoplasmic -catenin were measured by Western blot analysis. As shown in Fig. 1D, expression of LRRFIP2 in HEK293T cells results in a significant increase in the cytoplasmic level of -catenin, suggesting that the induction of the TOPflash reporter activity by LRRFIP2 is mediated through -catenin.
|
Activation of canonical Wnt signaling in early embryos produces a secondary dorsal axis by inducing the expression of Spemann organizer genes, including Siamois and Xnr3 (1, 2). To test whether LRRFIP2 also promotes dorsal axis formation by inducing Wnt target gene expression, the induction of Siamois and Xnr3 RNA by LRRFIP2 was measured in animal cap assays. Animal caps were isolated at stage 8.5 from embryos injected at the two-cell stage with LRRFIP2 or GFP RNA, cultured to stage 10.5, and then assayed by using RT-PCR for Siamois and Xnr3 RNA. As shown in Fig. 1E, LRRFIP2 strongly induces the expression of Siamois and Xnr3 to levels comparable to those induced by -catenin; injection of a GFP RNA control had no effect. These data suggest that LRRFIP2 can indeed function as a Wnt signaling activator in Xenopus.
The Amino Terminus of LRRFIP2 Functions as a Dominant Negative Form. To begin to analyze the molecular function of LRRFIP2, a series of deletion mutants was constructed as shown in Fig. 2 and then tested for their ability to activate the TOPflash reporter when expressed in HEK293T cells. The ability of LRRFIP2 to activate the TOPflash reporter was severely reduced after truncation of either the carboxyl terminus or the amino terminus, which encode the coiled-coil and Ser-rich domains, respectively. In contrast, deletion of the central region, which presumably mediates the interaction with Flightless-I (24), had little or no effect on LRRFIP2 activity.
|
LRRFIP2 Interacts with Dvl3. To explore the mechanism by which LRRFIP2 activates Wnt signaling, the binding of LRRFIP2 to components in the Wnt pathway was assayed by coimmunoprecipitation. A Flag-tagged form of LRRFIP2 was introduced into HEK293T cells, along with Myc-tagged forms of Dvl3, Axin, or GSK-3 by cotransfection. After 48 h, Myc-tagged and associated proteins were recovered from cell lysates by immunoprecipitation and then probed by Western blots for Flag-LRRFIP2 with an anti-Flag antibody (Fig. 3A). The results show that LRRFIP2 can be coimmunoprecipated in a complex with Dvl3 but not detectably with Axin or GSK-3 (data not shown).
|
The carboxyl terminus of LRRFIP2, which does not detectably interact with Dvl3, also seems to be required for its function (Fig. 2), presumably through interactions with other components of the Wnt pathway (or yet unknown factors), which function synergistically with LRRFIP2 to activate Dvl. It is worth noting that Western blot analysis of the LRRFIP2 truncated mutant afforded high molecular weight bands at trace levels, raising the possibility that LRRFIP2 is posttranslationally modified and that this modification could modulate its ability to interact with other proteins.
Dvl3 mediates Wnt signaling through three key domains: the DIX domain at the amino terminus, the PDZ domain in the central region, and the DEP domain near the carboxyl terminus (1, 2). Each of these domains likely mediates protein–protein interactions that allow the Dvl proteins to act as scaffolding proteins during Wnt signaling (27–36). To determine which of these domains binds LRRFIP2, a series of truncation mutants of Dvl3 (Fig. 7A, which is published as supporting information on the PNAS web site) was assayed with the coimmunoprecipitation protocol described above. Extracts were prepared from HEK293T cells coexpressing Flag-tagged LRRFIP2-M5 and Myc-tagged deletions of Dvl3, subjected to immunoprecipitation with an anti-Flag antibody, and analyzed by Western blots to detect the bound Myc-tagged Dvl3 proteins. As shown in Fig. 7B, the region containing both the PDZ and DEP domains is responsible for the interaction with LRRFIP2-M5. Each domain alone cannot bind LRRFIP2-M5 efficiently. Similar results were obtained when the Dvl3 mutant constructs were cotransfected with the full-length Flag-tagged LRRFIP2 (data not shown). These experiments suggest that binding of LRRFIP2 to Dvl3 involves residues from both the PDZ and DEP domains and/or the hinge region between the two. Moreover, forms of Dvl3 lacking the DIX domain were more strongly associated with LRRFIP2, suggesting that the DIX domain may be inhibitory for this interaction.
In vitro GST pull-down assays also were carried out to test the direct interaction between LRRFIP2 and Dvl. GST-LRRFIP2 (1–370) indeed could pull down in vitro translated Dvl3 (80–510) but with much lower efficiency than immunoprecipitation in HEK293T cells (data not shown). It is possible that phosphorylation of Dvl is required for the direct binding, because Dvl is heavily phosphorylated when expressed in cells, whereas the in vitro translated Dvl is not.
It was next determined whether LRRFIP2 functions within the Wnt pathway at the level of Dvl. A dominant negative form of Dvl (Xdd1) that lacks the PDZ domain (37) was assayed for its ability to inhibit LRRFIP2-induced TOPflash reporter activity in cultured cells. As shown in Fig. 4A, coexpression of Xdd1 totally abolished the activity of LRRFIP2. Furthermore, we tested whether the dominant negative mutant, LRRFIP2-M5, blocks the ability of Dvl3 to activate the TOPflash reporter in cultured cells. Again, dominant negative LRRFIP2 dramatically reduced Dvl3-activated reporter activity (Fig. 4B) but not the reporter activity promoted by -catenin (data not shown). Taken together, these data suggest that LRRFIP2 may function in the Wnt pathway in association with Dvl3, and vice versa, both acting upstream of -catenin.
|
|
-catenin (data not shown). These data suggest that LRRFIP2 may indeed be required for Wnt signaling and operate upstream of -catenin, as suggested by the results in tissue culture cells described above.
To determine whether LRRFIP2 is involved in endogenous axis formation, RNA encoding the dominant negative LRRFIP2 was injected at multiple positions along the dorsal ventral axis. A small but reproducible number of injected embryos showed hypodorsalized phenotypes, such as reduced head structures, indicative of reduced Wnt signaling (Fig. 5C). These phenotypes, although weak, were not observed in control embryos injected with a control -gal RNA, indicating that LRRFIP2 might be at least partially required for dorsal axis formation in Xenopus.
It cannot be ruled out that LRRFIP2 may function by means of its interactions with Dvl in other arms of the Wnt signaling pathway. In this respect, we have observed defects in gastrulation that might be consistent with the role of the Wnt pathway in convergent extension movements. The additional roles of LRRFIP2 in normal Xenopus development await further study.
|
Conclusions |
|---|
|
TopAbstractMaterials and MethodsResults and DiscussionConclusionsReferences |
|---|
-catenin and activate TCF/LEF-dependent gene transcription. Both the amino and carboxyl termini of LRRFIP2 are required for its function. The amino terminus of LRRFIP2 is responsible for interaction with the central region of Dvl, which contains the PDZ and DEP domains; a mutant of LRRFIP2 containing only the amino terminus acts as a dominant negative form. One possibility is that LRRFIP2 interacts with Dvl and acts as a scaffold protein to facilitate the activation of Dvl, leading to the stabilization of -catenin; disruption of this interaction with a dominant negative LRRFIP2 abolishes the activities of both. LRRFIP2 promotes the formation of a second dorsal axis in Xenopus embryos and induces the expression of organizer genes in animal caps. The dominant negative LRRFIP2 mutant suppresses xWnt8 dorsalizing activity, partially inhibits dorsal axis formation, and, in some embryos, reduces the levels of embryonic myogenesis, a process that is also known to require Wnt signaling within the dorsal and ventral marginal zone after the midblastula transition. Taken together, these results provide compelling evidence that LRRFIP2 transduces Wnt signals through the canonical pathway.
The lack of homologs of LRRFIP2 in invertebrates may be an indication that the primary function of LRRFIP2 in vertebrates may not be as a core component of Wnt signaling but as a modulator that is used to expand the signaling potential of the pathway, thus adding a new level of regulation during embryonic development. In this respect, LRRFIP2 could be akin to other signal modulators, such as Dkk1, Frodo, and Dapper, that modulate Wnt signaling but are unique to vertebrates (30, 31, 38).
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Abbreviations: LRRFIP2, leucine-rich repeat in Flightless interaction protein 2; GSK-3, glycogen synthase kinase 3; CK, casein kinase; Dvl, Disheveled; LEF, lymphocyte enhancer-binding factor; TCF, T cell factor; HEK, human embryonic kidney.
To whom correspondence should be addressed. E-mail: schultz{at}scripps.edu.
© 2005 by The National Academy of Sciences of the USA
|
References |
|---|
|
TopAbstractMaterials and MethodsResults and DiscussionConclusionsReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg What's this?
This article has been cited by other articles in HighWire Press-hosted journals:
|
|
 |
|
  S. Muramatsu, M. Wakabayashi, T. Ohno, K. Amano, R. Ooishi, T. Sugahara, S. Shiojiri, K. Tashiro, Y. Suzuki, R. Nishimura, et al. Functional Gene Screening System Identified TRPV4 as a Regulator of Chondrogenic Differentiation J. Biol. Chem., November 2, 2007; 282(44): 32158 - 32167. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-H. Lee and M. R. Stallcup Interplay of Fli-I and FLAP1 for regulation of {beta}-catenin dependent transcription Nucleic Acids Res., October 6, 2006; (2006) gkl652v3. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Park, Y. Hu, T. V. S. Murthy, F. Vannberg, B. Shen, A. Rolfs, J. E. Hutti, L. C. Cantley, J. LaBaer, E. Harlow, et al. Building a human kinase gene repository: Bioinformatics, molecular cloning, and functional validation PNAS, June 7, 2005; 102(23): 8114 - 8119. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
serves as a loading control.
