Homeodomain-interacting protein kinases (Hipks) promote Wnt/Wg signaling through stabilization of β-catenin/Arm and stimulation of target gene expression

Metazoan development is a highly dynamic and complex process that requires the action of several key signal transduction pathways. Their activity must be tightly regulated to ensure the proper patterning and growth of tissues. Regulation of a signaling pathway can occur at any level within the pathway,from perturbation of ligand-receptor interactions to regulation of the activity of transcription factors in the nucleus. Some regulators affect the on or off state of pathways, whereas others are involved in fine-tuning, an essential aspect that ensures that accurate and physiologically necessary levels of signaling are achieved without excessive signaling, which can have deleterious effects.

The evolutionarily conserved Wnt/Wingless (Wg) signaling pathway is involved in diverse biological processes, including determination,proliferation, migration and differentiation during embryonic development and adult homeostasis (reviewed in Clevers,2006). Inappropriate activation of Wnt-dependent gene expression in mammals can lead to numerous cancers, and loss of Wnt pathway activity also profoundly affects development. The ability of this pathway to control different developmental events in a temporally and spatially specific manner requires coordination between numerous regulators. Canonical Wnt signaling controls cell fate by regulating transcription of target genes(Cadigan and Nusse, 1997). Wingless (Wg), a secreted glycoprotein, is the best characterized of the seven Drosophila Wnt ligands, and initiates the canonical pathway by binding to the Frizzled2 (Fz2) and LRP5/6/Arrow co-receptors(Bhanot et al., 1996). This leads to the activation of Dishevelled, which then inhibits the activity of the destruction complex composed of Axin, glycogen synthase kinase 3β(GSK-3β)/Zw3 and adenomatous polyposis coli (APC)(Ikeda et al., 1998; Kishida et al., 1998; Polakis, 1997). As a result,cytosolic Drosophila β-catenin called Armadillo (Arm)accumulates and enters the nucleus to interact with a Tcf/Lef(Drosophila Tcf) family transcription factor to promote target gene expression (van de Wetering et al.,1997). In the absence of Wg signaling, the Axin/GSK-3β/APC complex promotes the proteolytic degradation of Arm(Aberle et al., 1997; Willert et al., 1999; Yost et al., 1996), whereas transcriptional co-repressors bind to Tcf and repress transcription(Cavallo et al., 1998; Roose et al., 1998).

The Nemo-like kinase family (Nlk) of protein kinases can regulate activation of Tcf/Lef target genes(Ishitani et al., 1999; Rocheleau et al., 1999; Zeng and Verheyen, 2004). In Drosophila, Nemo (Nmo) inhibits Drosophila Tcf activity, and is itself a transcriptional target of the Wg pathway(Zeng and Verheyen, 2004). Recently Homeodomain-interacting protein kinase 2 (Hipk2) was proposed to participate in a kinase cascade to activate Nlk during the regulation of the Myb transcription factor (Kanei-Ishii et al., 2004). We thus sought to identify whether this regulation was perhaps more general and whether Drosophila Hipk played a role in regulating Nmo, and thus also Wg signaling. We rapidly learned that Hipk exerts a positive effect on Wg signaling, distinct from Nmo, which we have more fully characterized using the developing wing as a model system.

The patterning of the adult wing blade is a tightly regulated process involving numerous essential signaling pathways, including Wg, Notch, EGFR and TGFβ, making it an excellent tissue in which to examine regulatory and epistatic relationships between many genes involved in patterning. The adult wing blade possesses five longitudinal veins (LI-LV) that extend proximally to distally. These are connected by the anterior cross vein (ACV) and posterior crossvein (PCV). The Wg pathway acts at several stages of wing patterning and growth (reviewed by Martinez Arias,2003). Wg is expressed along the dorsal/ventral boundary, which in imaginal discs is a stripe bisecting the wing imaginal disc, and in adult wings gives rise to the wing margin and bristles that surround the edge of the wing blade. Loss of wg can lead to loss of the entire wing, to wing to notum transformations, to wing notching or to loss of bristles along the entire wing margin (Phillips and Whittle,1993; Couso et al.,1994). Wg also promotes proliferation in the wing disc and ectopic Wg can induce outgrowths from the ventral surface of the wing(Phillips and Whittle, 1993; Diaz-Benjumea and Cohen,1995).

Hipk2 is a member of a conserved family of serine/threonine kinases. Vertebrate species possess four Hipk proteins (Hipk1-4) that have evolved distinct functions (reviewed by Rinaldo et al., 2007). Singly mutant Hipk1 and Hipk2 mice are viable, whereas double mutant mice die before birth (Isono et al., 2006). Drosophila possesses a single Hipk ortholog (which has been referred to as both Hipk and Hipk2)(Choi et al., 2005; Link et al., 2007) that shares extensive sequence homology within the kinase domain with members of the vertebrate family.

Vertebrate Hipk2 has been the most extensively studied member of the family(reviewed by Calzado et al.,2007; Rinaldo et al.,2007). Biochemical studies have identified a growing list of Hipk2 interactors, including proteins involved in transcriptional regulation,chromatin remodeling and key components of evolutionarily conserved signaling pathways. However, the biological investigation of these interactions in multicellular organisms has been minimal. In vivo examination of murine Hipk2 protein function has thus far revealed a role in neurogenesis and homeotic transformation of the skeleton (Isono et al., 2006; Wiggins et al.,2004; Zhang et al.,2007). Studies in Drosophila using transgenic flies expressing Hipk transgenes have uncovered a role for Hipk in regulating the global corepressor Groucho (Choi et al.,2005). Using loss-of-function mutant analyses, we have also identified a role for Hipk in promoting Notch signaling during Drosophila eye development (Lee et al., 2008).

In this study, we present an analysis of the function of Hipk in Drosophila canonical Wg signaling. Genetic studies show that ectopic hipk can rescue phenotypes owing to loss-of-function wgalleles or inhibition of the pathway with a dominant-negative Fz2 receptor. Immunohistochemical studies show that hipk positively regulates expression of Wg targets, and that Hipk can act to stabilize cellular levels of the Arm protein in wing discs. Wnt reporter assays show that both Drosophila Hipk and mouse Hipk2 can promote the Wnt-responsive Topflash reporter. In addition, we have found that Hipk/Hipk2 can promote the stabilization of Arm/β-catenin in cell culture and in vivo. Our results suggest that Hipk is a positive regulator of the Wg pathway that refines Wg activity during wing development. Our findings suggest that these roles may be conserved across species.

Fly strains used in the various crosses were: omb-Gal4, 69B-Gal4,ap-Gal/Cyo:TM6B, UAS-GFP (obtained from the Bloomington Drosophila Stock Center), sd-Gal4 (sd^SD29.1), bs-gal4 (also referred to as 1348-gal4; Huppert et al., 1997), UAS-hipk (Lee et al.,2008), UAS-Arm^S2/Cyo and UAS-Arm^S10 (Pai et al., 1997), UAS-Daxin^A2-4(Willert et al., 1999), UAS-Dfz2N33/Cyo (Zhang and Carthew, 1998), UAS-nmo^C5-1e(Verheyen et al., 2001), UAS-hipk RNAi (Vienna Drosophila RNAi Center, National Institute of Genetics), nmo^DB24 (Zeng et al., 2004), nmo^adk2 (Verheyen et al., 2001), hs-Flp22; GFP, FRT79/TM6B,hipk³/TM6B, hipk⁴, FRT79/TM6B. All wild-type flies are w¹¹¹⁸, and all crosses were performed according to standard procedures at 25°C. In assays examining interaction between two UAS transgenes, control crosses were performed with UAS-lacZ to rule out suppressive effects caused by titration of Gal4. To generate hipk loss-of-function clones, hsflp.22; GFP FRT79/TM6B females were crossed to hipk⁴ FRT79/TM6Bmales and progeny were collected for 24 hours and heat shocked at 38°C for 90 minutes at 48 AEL. Wing imaginal discs were dissected from late third-instar larvae for immunohistochemistry, and adult flies were collected for phenotypic analyses.

Adult wings were dissected in 100% ethanol followed by mounting in Aquatex(EM Science).

Antibody staining was performed according to standard protocols. The following primary antibodies were used: 1:50 mouse-anti-Achaete, 1:200 mouse anti-Armadillo (concentrated supernatant; Developmental Studies Hybridoma Bank), 1:400 mouse anti-Distalless (Duncan et al., 1998), 1:2000 rabbit β-galactosidase (Cappel) and 1:1000 mouse Myc (Sigma-Aldrich). The secondary antibodies were used as follows: anti-mouse CY3 (Molecular Probes), anti-rabbit CY3 (Jackson Immunolabs) and anti-mouse HRP (Jackson Immunolabs). All secondary antibodies were used at 1:200 dilution. Wing imaginal discs were mounted in 70%glycerol.

HEK293T cell culture, protein expression, immunoprecipitations and kinase assays, with Drosophila Hipk, were performed according to Zeng et al.(Zeng et al., 2007). The Hipk kinase dead construct (HA-Hipk-KD, EGFP-Hipk KD)(Choi et al., 2005) contains a K221R mutation within the catalytic ATP-binding site. For experiments using expression plasmids encoding mammalian cDNA, HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells were then transiently transfected with Polyethylenimine MW 25000(Polysciences). For co-immunoprecipitation assay, cells in 100 mm diameter plates were transfected with the expression plasmids encoding mammalian cDNA(12 μg). For reporter assays, cells in six well 35 mm diameter plates were transfected with the expression plasmids encoding mammalian cDNA (1 μg). Details on the generation of truncation constructs can be provided on request.

HEK293T and HeLa cells were cultured in six-well (35 mm diameter) plates and transiently transfected with Polyfect (Qiagen) or Polyethylenimine MW 25000 (Polysciences) according to manufacturer's manual. Cells were transfected at 24 hours after seeding with the TOPFLASH or FOPFLASH reporter gene plasmids along with each expression vector as indicated. Drosophila S2R+ cells were seeded (1×10⁶) in six-well plates and maintained at 25°C in Schneider's Drosophilamedium supplemented with 10% heat-inactivated FCS (Invitrogen) and 125μg/ml hygromycin B (Sigma). Before transfection, cells were re-seeded in S2 conditioned media or Wingless-conditioned Schneider's medium. Cells were transfected with 40 ng of Drosophila Hipk or 0.4 μg of mouse Hipk2, using the Effectene transfection reagent (Qiagen) according to the manufacturer's instructions. Eight hours post-transfection, the induction of genes under the control of the metallothionein promoter was performed by supplementing the medium with CuSO₄ at a final concentration of 0.5 mM. For both mammalian and Drosophila cells, total DNA concentration was kept constant by supplementation with empty vector DNAs. Luciferase assays were performed with the Dual Luciferase Reporter assay system (Promega)according to manufacturer's instructions and as described by Korinek et al.(Korinek et al., 1997). The renilla luciferase pRL-CMV or pRL-EF vector was used for normalizing transfection efficiencies. The values shown are the average of one representative experiment in which each transfection was performed in duplicate or triplicate as noted.

For examination of protein stability, HEK293T and HeLa cells were transfected in 60 mm plates with 1.5-2 μg per construct and empty vector was added to maintain a constant concentration of DNA. Cells were treated with 25 μg/ml cyclohexamide after 24 hours of transfection and harvested at the indicated times after treatment. Whole cell lysates were analyzed by SDS PAGE electrophoresis and immunoblotted with appropriate antibodies.

As a starting point in studying Hipk, we generated and analyzed a series of loss-of-function mutations (Lee et al.,2008). Loss of zygotic hipk resulted in pupal or larval lethality, a finding recently also described by Link et al.(Link et al., 2007). Whole-mount in situ hybridization reveals that hipk is expressed broadly in a non-uniform pattern in multiple stages of development, including all imaginal discs (data not shown). Removal of the maternal contribution caused embryonic lethality characterized by twisted embryos and head holes,showing that hipk is an essential gene for Drosophiladevelopment (W.L., unpublished). In our analyses we focused on the most severe allele, hipk⁴, which causes early larval lethality, or pupal lethality in trans to hipk³(Lee et al., 2008). Given the embryonic lethality caused by loss of maternal Hipk, we speculate that maternally contributed Hipk perdures and obscures its requirement at later stages and impacts the severity of mutant phenotypes. We used the FLP/FRT technique to generate mutant somatic clones to examine the requirements for Hipk in patterning adult structures (Xu and Rubin, 1993).

In this study, we focused on the role of hipk in the development of the wing. Clones of cells mutant for hipk⁴ show ectopic veins in the anterior region of the wing blade along LII, loss of the PCV and occasional notches in the wing margin (Fig. 1B). Reducing hipk function by expression of two independent Gal4-responsive hipk-RNAi constructs in the wing pouch with sd-gal4 (Fig. 1C)or vg-Gal4 (data not shown) also caused a wing notching phenotype reminiscent of those seen upon decreased Wg signaling(Phillips and Whittle, 1993; Couso et al., 1994; Rulifson et al., 1996).

We next studied the effects of ectopic expression of hipk using the Gal4-UAS system (Brand and Perrimon,1993). We observed phenotypes that suggested that hipk is involved in promoting Wg signaling. Expression of one copy of hipk in the central domain of the imaginal wing discs with omb-Gal4(Fig. 5C) induced the formation of an additional wing margin and outgrowths emanating from the distal most tip of the ventral surface of the wing (Fig. 1D,E). Expression of two copies of UAS-hipkenhanced this phenotype and caused the outgrowths to extend further distally from the ventral plane of the wing (Fig. 1F). These effects phenocopied the effects observed in ectopic wg-expressing clones(Diaz-Benjumea and Cohen,1995). Similarly, expression of UAS-hipk in the wing(Fig. 2B) phenocopies the ectopic venation pattern seen both upon ectopic expression of activated Arm(Arm^S10) with bs-Gal4(Fig. 2C) and in nmo^DB24/nmo^adk2 mutants(Fig. 2D). Although control of wing vein patterning is not generally attributed to Wg signaling, we and others have observed ectopic venation upon elevated Wg signaling. For example,ectopic expression of constitutively active Arm by en-Gal4 in the posterior region of the wing, or ubiquitously with 69B-Gal4 or MS1096-Gal4, leads to disturbed and ectopic venation(Greaves et al., 1999; Lawrence et al., 2000). Moreover, loss-of-function clones of sgg/zw3 (encoding the fly homolog of GSK3β, a component of the destruction complex) induce the formation of ectopic veins (Ripoll et al.,1988). The results of these phenotypic analyses of hipkare surprising because they demonstrated that Nmo and Hipk did not act in concert to inhibit Wg signaling. Rather, hipk mutant and gain-of-function phenotypes suggest a role in promoting the Wg pathway

To explore whether Hipk is a positive regulator of the Wg pathway, as suggested by our initial analyses, genetic interaction studies were performed. During early wing development, Wg specifies the wing and disruption of signaling during the second larval instar induces a wing-to-notum transformation (Fig. 2E)(Morata and Lawrence, 1977; Ng et al., 1996). Only 14% of wg¹/wg^1-17 transheterozygotes (n=72)develop a normal pair of wings and misexpression of hipk in this genetic background suppressed the formation of the ectopic notum and restored the structure of the wing (Fig. 2F) in 41% of flies (n=39). Expressing a dominant-negative form of the Fz2 receptor Dfz2N33 ubiquitously in the wing with 69B-Gal4 caused a severe loss of wing margin phenotype in both the anterior and posterior regions of all wings examined(Fig. 2G; n=41)(Zhang and Carthew, 1998). This phenotype was entirely rescued with 100% penetrance (n=47) upon hipk co-expression (Fig. 2H). Altogether, these interactions indicated that hipkcan counteract the effects of inhibiting the Wg pathway, and, more importantly, that elevating the levels of Hipk can compensate for the reduced Wg signal.

Consistent with these observations, we find that removal of one dose of hipk can enhance phenotypes owing to inhibited Wg activity. Expression of the Dfz2N33 along the anteroposterior boundary using dpp-Gal4 induced a mild wing notching phenotype(Fig. 2I, 47%, n=15). Heterozygosity for hipk⁴ enhanced the loss of wing tissue as seen by the formation of moderate (27.8%, n=36) to severely truncated wings (Fig. 2J,30.6%), which were rarely observed in dpp>Dfz2N33 adults (6.7%, n=15). Similarly, heterozygosity for hipk³enhanced the severity of the wing notching induced by expression of Axin with sd-Gal4 (Fig. 2K,L). Thus, either a loss or a gain of hipk demonstrates that Hipk plays a positive role in transmission of the Wg signal.

Our genetic observations suggested that Hipk promoted the Wg pathway. To assess whether modulation of hipk levels could affect Wg signaling activity, we examined the expression of the Wg targets distalless(Dll), achaete (Ac) and senseless (Sens) in both hipk RNAi discs and in discs ectopically expressing Hipk. Expression of a hipk RNAi construct (sd>hipk RNAi) led to reductions in Dll (Fig. 3B), Sens(Fig. 3E) and Ac(Fig. 3H) in the wing pouch. Conversely, ectopic Hipk (omb>hipk) enhanced and expanded the expression domains of Dll (Fig. 3C), Sens (Fig. 3F)and Ac (Fig. 3I) in the wing pouch.

In the absence of Wg pathway activation, the destruction complex targets Arm for degradation. Upon Wg signaling, Arm is stabilized and accumulates in the cytoplasm, serving as a measure of Wg activity. In wild-type wing discs,cytoplasmic Arm is stabilized in cells adjacent to the DV boundary in which Wg is active (Peifer et al.,1994) (Fig. 4A). We determined the status of stabilized Arm in hipk mutant clones. We found that Arm is reduced in hipk clones at the DV boundary (arrows in Fig. 4B-D). These mutant cells showed a reduction of cytoplasmic Arm (Fig. 4B″-4D″),whereas the adherens junction pool of Arm appeared normal. A similar decrease in Arm is seen in discs expressing hipk RNAi(Fig. 4E).

We then assessed whether Hipk could promote Arm stabilization by ectopically expressing Hipk using omb-gal4. Endogenous Arm protein levels are expanded in omb>hipk wing discs(Fig. 5B), compared with wild type (Fig. 5A). Consistent with these observations, western blot analysis of omb>hipk imaginal discs revealed higher levels of Arm than lysates obtained from wild-type larvae (Fig. 9A). These results suggested that elevated levels of Hipk can promote more Arm stabilization even in regions of the wing disc receiving lower levels of Wg signaling and that the stabilized Arm is active for Wg signaling, as it can induce target gene expression.

To further address the role of Hipk in Arm stabilization, the consequences of reducing hipk were assessed on the stability of a series of UAS-Arm constructs expressed throughout the wing pouch in hipk³/hipk⁴ mutant wing discs. The Arm constructs UAS-Arm^S2 and UAS-Arm^S10(Pai et al., 1997) have been used to dissect the regulatory mechanisms of Arm localization and stability in the embryo (Tolwinski and Wieschaus,2001) and wing (Bajpai et al.,2004). UAS-Myc-Arm^S2 (encoding Myc-tagged full-length Arm) was expressed in a broad domain in the central region of the wing pouch and in the ventral periphery using omb-gal4(Fig. 5C). Wing discs stained with anti-Myc antibody showed stabilization of Myc-Arm^S2 along the DV boundary and the dorsal hinge primordia, similar to what was seen with endogenous Arm (Fig. 5D). In hipk mutant discs expressing omb>myc-Arm^S2, we observed a failure to accumulate Arm protein (arrow in Fig. 5E), suggesting that in these discs Arm protein is not efficiently stabilized owing to the reduction of Hipk.

Expression of a UAS-Myc-Arm^S10 construct lacking the N-terminal target sites of the destruction complex induced a wggain-of-function phenotype characterized by ectopic bristles within the wing blade (Couso et al., 1994;data not shown). omb>UAS-Arm^S10 discs show stabilized Myc-tagged Arm^S10 throughout the omb expression domain(Fig. 5F). The accumulation of the degradation-resistant form of Arm is unchanged in discs lacking hipk (Fig. 5G). These data show that Hipk promotes the stabilization of wild-type Arm, whereas degradation-resistant Arm bypasses the need for Hipk-promoted stabilization.

As Hipk could modulate Arm stability and expression of Wg target genes, we examined whether Hipk interacted with the core components of the transcriptional complex. In HEK293T cells transfected with HA-tagged Hipk and Myc-tagged Drosophila Tcf, Hipk was co-immunoprecipitated in a complex with Drosophila Tcf (Fig. 6A). Myc-Hipk and HA-tagged Arm also formed a complex in HEK293T whole cell lysates (Fig. 6B). These complexes were also seen in HeLa cell lysates transfected with mouse Flag-tagged Hipk2 and β-catenin or human T7-tagged Lef1(Fig. 6C).

We next investigated whether β-catenin was required for the interaction between Lef1 and Hipk2. Hipk2 binds a Lef1ΔN mutant, lacking the β-catenin binding domain, suggesting that β-catenin is not required for the interaction between Lef1 and Hipk2(Fig. 6C, lane 4). Furthermore,Hipk2 co-immunoprecipitated with Lef1ΔC, a construct lacking the conserved DNA binding HMG box, suggesting this region was also dispensable for the Lef1/Hipk2 interaction (Fig. 6C, lane 6). Although the β-catenin and Hipk2 interaction does not require Lef1, we observe that complex formation is enhanced in the presence of Lef1 (Fig. 6C,lanes 8, 9).

In vitro kinase assays were performed to determine whether Hipk phosphorylates components of the transcriptional complex. We found that Hipk phosphorylates Arm, but not Drosophila Tcf(Fig. 7A, lane 1; data not shown). The kinase activity was crucial for this event, as a kinase dead Hipk protein (Hipk KD) was unable to phosphorylate Arm(Fig. 7A, lane 2). We next tested the ability of Hipk to phosphorylate Arm truncations(Fig. 7A, lanes 4, 11; Fig. 7B). We find that Hipk can phosphorylate a number of Arm truncations, namely Arm-N (encoding only the N terminus, Fig. 7A, lane 11),Arm ΔN (missing the N terminus, data not shown) and Arm ΔC(missing most of the C terminus, Fig. 7A, lane 4), but not Arm-R (composed of just the central repeat region, data not shown), suggesting that phosphorylation sites map to both the N and C termini of Arm. Hipk proteins are known to phosphorylate numerous targets on their substrates, and further detailed analyses will reveal the functional significance of each of these phosphorylation events.

We examined whether Hipk affected expression of the Tcf-responsive Topflash transcriptional reporter (Korinek et al.,1997). Transfection of HEK293T cells with Hipk enhanced the transcriptional activity induced by Drosophila Tcf/Arm nearly tenfold compared with transfection with Drosophila Tcf and Arm alone(Fig. 8A, lanes 2, 3). This effect was not observed upon transfection with Hipk KD(Fig. 8A, lane 4), suggesting that the kinase activity of Hipk is required for this effect. We also assessed the combined effect of Nmo and Hipk on Topflash. As expected, expression of the Tcf antagonist Nmo suppressed Topflash activation(Fig. 8A, lane 5). This inhibitory effect was partially relieved when HEK293T cells were co-transfected with Hipk and Nmo (Fig. 8A, lane 6).

Next we examined the ability of Hipk to promote Topflash through interactions with the endogenous proteins(Fig. 8C) in Drosophila S2R+ cells that express the Fz2 receptor, which can be activated with Wg-conditioned media. Transfection of Hipk resulted in activation of Topflash in the absence of Wg-conditioned medium(Fig. 8C, lanes 1, 2). Upon addition of Wg-conditioned media, cells transfected with Hipk showed a robust induction of Topflash (Fig. 8C,lanes 4, 5).

Although Hipk enhances Topflash in HEK293T cells and genetically promotes Wg signaling, transfection of Hipk2 led to an inhibition of Topflash in HEK293T cells (data not shown) (Wei et al., 2007). This effect of Hipk2 may be attributed to additional mammalian Hipk homologues or to the cellular context. To investigate this, we performed similar assays with Hipk2 in HeLa cells and Drosophila S2R+cells. Transfection of Hipk2 in HeLa cells led to an enhancement of Topflash compared with control cells that were solely treated with Wnt3a-(Fig. 8B, lanes 5, 6) or Wnt1-conditioned media (Fig. 8B, lanes 9, 10). Hipk2 KD did not cause an increase in the transcriptional response, indicating this effect is kinase dependent(Fig. 8B, lanes 7,11). Strikingly, addition of fly Hipk to mammalian cells also enhanced the transcriptional response (Fig. 8B, lanes 8, 12). Consistent with the HeLa cell data, Hipk2 was able to induce Topflash in the presence or absence of Wg-conditioned media in Drosophila S2R+ cells (Fig. 8C, lanes 3, 6). These experiments revealed that Hipk2 and Hipk perform conserved functions in multiple cellular contexts.

Our data suggested that Hipk is required to stabilize both endogenous and overexpressed full-length Arm. We confirmed this by western blotting of protein extracts from omb>hipk wing discs and observed dramatically elevated levels of Arm protein(Fig. 9A), consistent with the elevated levels of Arm seen in discs in Fig. 5B. Arm and β-catenin protein levels are also increased in S2R+ cells(Fig. 9B) and HeLa cells(Fig. 9C) transfected with Hipk2, respectively.

Further cell culture assays were performed to confirm this role. Degradation of Arm was assessed in HEK293T cells expressing HA-Arm that were treated with the protein synthesis inhibitor, cycloheximide (Chx)(Abou Elela and Nazar, 1997). In these cells, addition of Hipk prolonged the stability of Arm(Fig. 9D). Similar results were obtained in HeLa cells (data not shown). The presence of HA-Hipk KD partially promoted the accumulation of Arm levels, but not to the extent seen with Hipk WT.

The Wg/Wnt pathway is crucial for the initiation and maintenance of developmental programs in multicellular organisms across the animal kingdom. Alterations in signaling activity can have dire consequences on cell fate, in the most severe circumstances causing the initiation and progression of tumorigenesis. In this study we reveal that Hipk possesses an intrinsic ability to promote Wg pathway activity and this regulatory function for Hipk is conserved in both Drosophila and mammalian cells. Through a combination of genetic and biochemical analyses, our data reveal that Hipk proteins promote Tcf/Lef1-mediated transcription. Additionally, Hipks enhances the stabilization of Arm/β-catenin in several cell lines and hipk mutant clones in the wing disc have diminished Arm protein levels. Overexpression of Hipk induces a broader domain of stabilized Arm,suggesting Hipk is required to maintain the signaling pool of cytosolic Arm. We propose a model in which Hipk promotes the Wnt/Wg signal via its regulation of Arm stabilization.

Nlk is a conserved antagonist of the Lef1/β-catenin transcriptional complex. Our research has shown that Nmo is also an inducible antagonist of the Wg signal in the developing Drosophila wing (Zeng et al., 2004). It was previously reported that Wnt1 induces the activation of a putative Tak1-Hipk2-Nlk kinase cascade to promote the degradation of the Myb transcription factor (Kanei-Ishii et al.,2004). We sought to delineate the physiological relevance of this potential kinase cascade; specifically, we determined whether these interactions played a role in the regulation of the Wg pathway. Our data reveal that in Drosophila Nmo and Hipk do not form a kinase cascade in this context, rather they exert opposing effects on the same pathway,probably through distinct mechanisms.

The Wg morphogen can bring forth a spectrum of biological processes. Sustaining maximal levels of signaling could be accomplished through the amplification and enhancement of the signal in the wing margin. In support of this model, we found that overexpression of Hipk expands the expression of Wg targets such as Dll, Sens and Ac. Transcriptional assays reveal that both Drosophila Hipk and mouse Hipk2 enhance the transcriptional activity of Tcf and Lef1, respectively, in a kinase-dependent manner. These findings strongly suggest that Hipk and Hipk2 function to enhance the activity of the transcriptional complex to promote the Wg/Wnt signal.

Accumulation of stabilized Arm is paramount to effective Wg signaling. Failure to escape the destruction complex results in Arm degradation and inhibition of Tcf-mediated gene activation. Thus understanding the regulation of Arm is central to the global understanding of how the Wg signal is modulated. In our studies, we reveal a role for Hipk in Arm stabilization. This feature is highlighted by the loss of stabilized Arm in hipkmutant clones. Additionally, in hipk mutant discs, overexpressed wild-type Arm fails to accumulate, despite its expression in domains of high Wg signaling. These findings demonstrate that Hipk plays an important role in Arm stabilization. Hipk may reduce the ability of Arm to interact with destruction complex components or may increase the nuclear retention of Arm. In the absence of Hipk, either of these scenarios would give the destruction complex more access to Arm. In agreement with such a model, we find that increasing Hipk activity in the wing surpasses the inhibitory effects of the degradation machinery and expands the perimeter of stabilized Arm. Furthermore, we have found that the presence of Hipk or Hipk2 in cell culture stabilizes Arm/β-catenin. Thus, the enhanced transcriptional activity is probably due to the elevated availability of Arm protein.

It is crucial for normal development to maintain the proper amounts ofβ-catenin, as elevated levels of β-catenin can lead to cancer(reviewed by Clevers, 2006). Elaborate regulatory networks in the cytoplasm and nucleus are vital to maintaining appropriate levels of β-catenin. It is well documented that phosphorylation in the N terminus of β-catenin is crucial for its negative regulation (reviewed by Daugherty and Gottardi, 2007). A chain of phosphorylation events begins when Casein Kinase I (CKI) primes β-catenin for successive modifications by GSK-3β (Amit et al., 2002; Liu et al., 2002; Yanagawa et al., 2002). Central to this event is Axin, which provides a scaffold for APC, CK1,GSK-3β and β-catenin (Amit et al., 2002; Hart et al.,1998; Hinoi et al.,2000; Ikeda et al.,1998). N-terminally phosphorylated β-catenin is ubiquitinated by βTrCP ubiquitin ligase and targeted for degradation via the proteasome.

Wnt signaling promotes the accumulation of β-catenin; however, some of the mechanisms governing this process remain enigmatic. Although overexpression of wild-type β-catenin/Arm is unable to overcome the effects of the degradation machinery(Mohit et al., 2003; Pai et al., 1997),Wnt-stimulated β-catenin can resist the activity of the destruction complex. Although achieving stabilized pools of β-catenin represents the core goal of the Wnt pathway, high levels of β-catenin are not always coupled with elevated transcription (Guger and Gumbiner, 2000; Staal et al., 2002). For example, in Xenopus, alanine substitution of one of the GSK3 target residues leads to elevated β-catenin levels,without causing an increase in Tcf-mediated transcription(Guger and Gumbiner, 2000). Thus, further posttranslational modifications of β-catenin are necessary to potentiate its signaling activity.

Furthermore, phosphorylation can affect β-catenin stability by affecting protein-protein interactions that regulate protein turnover and activity. Phosphorylation of β-catenin by Cdk5 inhibits APC binding toβ-catenin (Munoz et al.,2007; Ryo et al.,2001), whereas phosphorylation by CK2 promotes β-catenin stability and transcriptional activity(Song et al., 2003).

Recent advances have begun to unravel the molecular complexity that controls β-catenin-mediated transcription within the nucleus. Upon pathway activation, Tcf recruits Arm to the enhancers of Wg-responsive genes where Arm forms multiple transcriptional complexes along its length(Hoffmans et al., 2005; Kramps et al., 2002; Thompson et al., 2002; Hecht et al., 1999; Mosimann et al., 2006). Formation of these transcriptional units is needed for the transmission of the Wg/Wnt signal. Recent studies have shown that phosphorylation (distinct from the N-terminal phosphorylation that triggers β-catenin destruction) may modulate its ability to recruit these co-factors (reviewed by Daugherty and Gottardi, 2007). We have found that the Hipk-dependent stabilized form of Arm is transcriptionally active and induces the expression of Wg targets, suggesting modification by Hipk may promote protein interactions.

APC and the cell-adhesion molecule E-cadherin compete with Tcf for overlapping binding sites on β-catenin(Hulsken et al., 1994; von Kries et al., 2000). Competition between proteins may play an important role in the regulation of the Wnt signaling pathway (reviewed by Xu and Kimelman, 2007). We propose that Hipks may promote the stability of Arm/β-catenin by excluding further interactions with other proteins, including those that antagonize Arm/β-catenin. Given that Hipks can also bind to Tcf/Lef1, we predict that these proteins may act synergistically to displace the inhibitory partners of β-catenin. In agreement with such a role, we observe that Lef1 enhances the interaction between Hipk2 and β-catenin, and these interactions may insulateβ-catenin from components of the degradation machinery.

Although Hipk phosphorylates Arm, the functional significance of this modification has yet to be determined. Hipk might facilitate the interactions between Arm and its transcriptional co-factors, as Hipk2 phosphorylation has been shown to affect gene regulation by modifying the composition of various transcriptional complexes (reviewed by Calzado et al., 2007). Hipk may also enhance the formation of the β-catenin/Tcf transcriptional complex by inducing a conformational change and/or reducing the affinity of possible inhibitors for β-catenin through the phosphorylation of β-catenin. Recently, it has been reported that Hipk2 could antagonizeβ-catenin/Tcf-mediated transcription in a kinase-independent manner(Wei et al., 2007). Although these data appear in conflict with our findings, we have observed that the effect of Hipk2 on transcription is very cell type- and target gene-dependent,suggesting Hipk2 function is affected by its cellular context, most probably owing to the availability of targets and co-factors.

The dynamic localization of Hipk2 in the nucleus, nucleoplasm and in cytosolic speckles suggests that the protein may carry out distinct roles in each site (reviewed by Calzado et al.,2007; Rinaldo et al.,2007). Given the growing list of interacting proteins, it is tempting to speculate that specific Hipk function is determined in part through its particular localization. It is also possible that Hipk/Hipk2 may act as a scaffolding protein, bringing together multiple binding partners. Ongoing biochemical studies will further uncover the molecular significance of these interactions. Hipk proteins are emerging as important components of multiple signaling networks. Our studies describe the roles of Hipk and Hipk2 as Wnt/Wg regulators and shed light on the regulatory mechanisms governing this conserved pathway.