Control of antagonistic components of the Hedgehog signaling pathway by microRNAs in Drosophila

Hedgehog (Hh) signaling is critical for many developmental processes and for the genesis of diverse cancers. Hh signaling comprises a series of negative regulatory steps, from Hh reception to gene transcription output. Stability of antagonistic regulatory proteins, including the coreceptor Smoothened (Smo), the kinesin-like Costal-2 (Cos2), and the kinase Fused (Fu), is affected by Hh signaling activation. This study shows that the level of these three proteins is also regulated by a microRNA cluster. Indeed, the overexpression of this cluster and resulting microRNA regulation of the 3'-UTRs of smo, cos2, and fu mRNA decreases the levels of the three proteins and activates the pathway. Further, the loss of the microRNA cluster or of Dicer function modifies the 3'-UTR regulation of smo and cos2 mRNA, confirming that the mRNAs encoding the different Hh components are physiological targets of microRNAs. Nevertheless, an absence of neither the microRNA cluster nor of Dicer activity creates an hh-like phenotype, possibly due to dose compensation between the different antagonistic targets. This study reveals that a single signaling pathway can be targeted at multiple levels by the same microRNAs (Friggi-Grelin, 2009).

cos2, fu, and smo mRNA can be regulated by a cluster of microRNAs, including miR-12 and miR-283, in Drosophila wing disc. The overexpression of this cluster decreases the levels of Smo, Cos2, and Fu proteins and activates the Hh pathway, as evidenced by the induction of dpp expression in the wing imaginal discs and by the adult wing outgrowth. The experiments presented in this study with the 3'-UTR sensors of smo, fu, or cos2 are in favor of a direct binding. To constitute a real proof of a direct effect, further experiments as direct biochemical binding assay or compensatory mutation between the 3'-UTR and the miRNAs will be necessary to perform (Friggi-Grelin, 2009).

Programs that have been created to genomewide predictions of Drosophila miRNA targets provide lists of presumptive miR-12, and miR-283 regulated genes. In addition to the current in vivo validations, miR-12 binding sites are predicted on the 3'-UTR of ci and no sites were found on the 3'-UTR of the Su(fu) gene. No decrease was observed in either of these two proteins in the microRNA cluster overexpressing clones. It is interesting to note that Su(fu) mRNA, encoding another negative regulator of Hedgehog signaling, has been shown to be targeted by miR-214 in zebrafish. Absence of miR-214 results in the reduction of muscle cell types, the specification of which is dependent on Hh pathway activity. Nevertheless, the current study shows that in Drosophila wing discs an absence of microRNA does not modify the Hh pathway, raising the question of what the role of microRNAs in Drosophila Hh pathway regulation is (Friggi-Grelin, 2009).

Could the microRNAs overexpression phenotype that was identified be artifactual and simply the result of forced overexpression of the microRNA cluster in a tissue in which it should be silent? It is thought that the answer is no, because Northern blot analysis and the increase of miR-sensor in the dcr-1 mutant clones showed that the microRNA cluster is indeed expressed in this tissue. This suggests that the cluster likely has a role in this tissue in which it is normally present. Is the microRNA cluster regulation of the cos2 and smo 3'-UTRs physiological? It is thought so, because an absence of either the microRNA cluster or of Dicer in the wing imaginal disc induces an increase in the Cos2- and Smo-sensor lines. This signifies that the microRNAs expressed from the cluster regulate the cos2 and smo 3'-UTRs and thus display some functionality in the disc during larval development. Altogether, these data clearly show that an artifactual situation in which the microRNA cluster is expressed in a tissue in which it should not be present has not been created. The miRs overexpression was also tested on embryonic patterning but it did not lead to any phenotype, suggesting that the miR cluster regulation on the Hh pathway is specific to larval tissues (Friggi-Grelin, 2009).

As miR-12 and miR-283, and likely redundant miRs, are present in every cell of the wing disc, one possibility is that their normal roles are to dampen down the levels of Hh pathway components, particularly Cos2 and Smo, to prevent the accidental activation or downregulation of the pathway. Indeed, expressing both the microRNA cluster and its targets in the same tissue could provide a means of 'buffering stochastic fluctuations' in mRNA levels or in protein translation rates within the Hh signaling pathway, as has been proposed for other processes (Friggi-Grelin, 2009).

The data possibly indicate that miRNAs are able to regulate two antagonistic components of the pathway, Cos2 and Smo. It has been shown that the stability of these two proteins is 'interdependent': an increased level of Cos2 in the wing imaginal disc lowers the level of Smo, and, in the opposite direction, increased Smo decreases the level of Cos2. It is proposed that the interregulation of Cos2/Smo levels is independent of their relative activities because Cos2 effect on Smo levels is observed in posterior cells in which Cos2 activity is strongly inhibited by the constitutive activation of the pathway. Therefore, eliminating the miRNA-mediated inhibition of Cos2 and Smo in Delta3miR or dcr-1 mutant cells likely initially increased the levels of both proteins, but then the resulting higher levels of each protein presumably downregulated the other; the net variation of Cos2 and Smo levels would therefore be null. This hypothesis is favored because the independent Smo- and Cos2-sensor lines, which are unaffected by this Cos2/Smo interregulation, showed increased levels of GFP staining in Delta3miR and dcr-1 mutant animals. This suggests that the levels of both Cos2 and Smo are increased in the mutant animals but, because of the downregulation of each protein by the other, no ultimate alterations in the levels of the proteins are observed. If so, an Hh phenotype would not be expected to be seen in the miR mutant (Friggi-Grelin, 2009).

The screen created a situation in which the expression of the microRNA cluster is deregulated, ultimately destabilizing Cos2 protein levels and thereby activating Ci and Hh target gene expression. Importantly, a similar situation might be encountered during tumoral development. Aberrant Hh signaling activity is known to trigger the development of diverse cancers. While several of these tumors have been linked to mutations in Hh signaling components, not all of them have, leaving open the possibility that they are caused by other factors such as microRNA misexpression. Interestingly, more than half of the known human microRNA genes are located near chromosomal breakpoints associated with cancer, and in some documented cases the microRNAs are amplified, leading to overexpression. Some upregulated microRNAs are possibly able to bind mRNAs encoding negative regulators of Hh signaling, such as Su(fu) or Ptc, and could thus induce the misactivation of the Hh pathway, as is observed in some cancers. Therefore, a fine analysis of microRNA expression levels and the levels of known Hh components should be considered in studies of Hh pathway-related cancers (Friggi-Grelin, 2009).

What does this study add to the current knowledge about miRNA regulation? The study shows that a cluster of three microRNAs can target several antagonistic components of the same pathway in vivo. This is novel and unexpected. This raises the question of how to interpret the miRNA expression signatures observed in human tumors. Indeed, as stated above, it has been proposed that miRNAs are differentially expressed in human cancers and contribute to cancer development. The working hypothesis in the cancer/miRNAs field is that key cancer genes are regulated by aberrant expression of miRNAs. The identification of a specific miRNA:mRNA interactor pair is generally accepted as being of biological importance when the mRNA encodes a tumor suppressor or an oncogene whose expression is modified in the tumor. This study shows indirectly that this is an oversimplified view, because identifying an oncogene or tumor suppressor as a target of a miRNA may not provide a full explanation for tumor development if the same miRNA hits other antagonistic components of the same pathway that nullify the effect of the identified miRNA:mRNA interactor pair (Friggi-Grelin, 2009).

Hedgehog signaling regulates the ciliary transport of odorant receptors in Drosophila

Hedgehog (Hh) signaling is a key regulatory pathway during development and also has a functional role in mature neurons. This study shows that Hh signaling regulates the odor response in adult Drosophila olfactory sensory neurons (OSNs). This is achieved by regulating odorant receptor (OR) transport to and within the primary cilium in OSN neurons. Regulation relies on ciliary localization of the Hh signal transducer Smoothened (Smo). This study further demonstrates that the Hh- and Smo-dependent regulation of the kinesin-like protein Cos2 acts in parallel to the intraflagellar transport system (IFT) to localize ORs within the cilium compartment. These findings expand knowledge of Hh signaling to encompass chemosensory modulation and receptor trafficking (Sanchez, 2016).

This study demonstrates that the Hh pathway modulates the magnitude of the odorant response in adult Drosophila. The results show that the Hh pathway determines the level of the odorant response because it regulates the response in both the positive and negative directions. Loss of Ptc function increases the odorant response and the risk for long sustained responses, which shows that the Hh pathway limits the response potential of the OSNs and is crucial for maintaining the response at a physiological level. In addition, it was shown that the OSNs produce Hh protein, which regulates OR localization, which is interesting because autoregulation is one of the prerequisites for an adaptive mechanism. It was further shown that Hh signaling regulates the responses of OSNs that express different ORs, which demonstrates that the regulation is independent of OSN class and suggests that Hh signaling is a general regulator of the odorant response. It has been shown previously that Hh tunes nociceptive responses in both vertebrates and Drosophila (Babcock, 2011). It is not yet understood how Hh regulates the level of nociception. However, the regulation is upstream of the nociceptive receptors, which indicates that the Hh pathway is a general regulator of receptor transport and the level of sensory signaling (Sanchez, 2016).

The results show that OSN cilia have two separate OR transport systems, the Hh-regulated Cos2 and the intraflagellar transport complex B (IFT-B) together with the kinesin II system. The results show that Cos2 is required for OR transport to or within the distal cilium domain and suggest that the IFT system regulates the inflow to the cilium compartment. The two transport systems also are required for Smo cilium localization (Kuzhandaivel, 2014). This spatially divided transport of one cargo is similar to the manner in which Kif3a and Kif17 regulate distal and proximal transport in primary cilia in vertebrates. However, Cos2 is not required for the distal location of Orco or tubulin (Kuzhandaivel, 2014), indicating that, for some cargos, the IFT system functions in parallel to Cos2 (Sanchez, 2016).

Interestingly, the vertebrate Cos2 homolog Kif7 organizes the distal compartment of vertebrate primary cilia (He, 2014). Similar to the current results, Kif7 does so without affecting the IFT system, and its localization to the cilia is dependent on Hh signaling. However, the Kif7 kinesin motor function has been questioned (He, 2014). Therefore, it will be interesting to analyze whether Kif7-mediated transport of ORs and other transmembrane proteins occurs within the primary cilium compartment and whether the ciliary transport of ORs is also regulated by Hh and Smo signaling in vertebrates. To conclude, these results place the already well-studied Hh signaling pathway in the post-developmental adult nervous system and also provide an exciting putative role for Hh as a general regulator of receptor transport to and within cilia (Sanchez, 2016).

Protein Interactions

The Costa protein can be coimmunoprecipitated with antibodies to Fused. Both Fused and a hyperphosphorylated isoform of Fused designated FU-P are found in this conplex. FU-P predominates over Fused when precipitates are prepared with Cos2 antisera. Antisera against either Fused or Cos2 precipitate Cubitus interruptus protein as well. Fractionated extracts of cultured cells have two complexes larger than Fused itself, a population of about 40 million Da, and a population of greater than 700,000 Da. The 700 kDa fraction is by far the most abundant. Fused, Cos2 and Ci are enriched in microtubules formed from repolymerized tubulin. Binding of Cos2 and Fused to microtubules is barely detectable in Hedgehog treated cultured cells. These findings sustain the hypothesis that signaling from Hh releases the complexes from microtubules, which would in turn facilitate translocation of Ci to the nucleus (Robbins, 1997).

Labeling with radioactive phosphate reveals that Fused and Cos2 are phosphorylated in both cultured S2 cells and Hedgehog treated S2 cells. The phosphorylations of Fused and Cos2 is on serine. Cos2 coimmunoprecipates with kinase dead Fused mutant proteins. Thus, functional Fused kinase is probably not necessary for Fused and Cos2 to associate. There is no evidence for binding of Cos2 to the products of truncated Fused protein lacing the C-terminal domain of Fused (Robbins, 1997).

Cos2 is cytoplasmic and binds microtubules in a taxol-dependent, ATP-insensitve manner, while kinesin heavy chain binds microtubules in a toxol-dependent, ATP-insensitive manner. Ci associates with Cos2 in a large protein complex, suggesting that Cos2 directly controls the activity of Ci. This association does not involve microtubules. Elevated cytoplasmic Ci staining is seen in cos2 clones in the anterior compartment. The level of Ci staining is independent of the clone's distance from the A/P border. Nuclear Ci is not evident in the clones (Sisson, 1997).

Hedgehog (Hh) signal transduction requires a large cytoplasmic multi-protein complex that binds microtubules in an Hh-dependent manner. Three members of this complex, Costal2 (Cos2), Fused (Fu), and Cubitus interruptus (Ci), bind each other directly to form a trimeric complex. This trimeric signaling complex exists in Drosophila lacking Suppressor of Fused [Su(fu)], an extragenic suppressor of fu, indicating that Su(fu) is not required for the formation, or apparently function, of the Hh signaling complex. However, Su(fu), although not a requisite component of this complex, does form a tetrameric complex with Fu, Cos2, and Ci. This additional Su(fu)-containing Hh signaling complex does not appear to be enriched on microtubules. Additionally, it has been demonstrated that, in response to Hh, Ci accumulates in the nucleus without its various cytoplasmic binding partners, including Su(fu). A model is discussed in which Su(fu) and Cos2 each bind to Fu and Ci to exert some redundant effect on Ci such as cytoplasmic retention. This model is consistent with genetic data demonstrating that Su(fu) is not required for Hh signal transduction proper and with the elaborate genetic interactions observed among Su(fu), fu, cos2, and ci (Stegman, 2000).

Much of the understanding of the Hedgehog signaling pathway comes from Drosophila, where a gradient of Hh signaling regulates the function of the transcription factor Cubitus interruptus at three levels: protein stabilization, nuclear import, and activation. Regulation of Ci occurs in a cytoplasmic complex containing Ci, the kinesin-like protein Costal-2 (Cos2), the serine-threonine kinase Fused (Fu), and the Suppressor of Fused [Su(fu)] protein. The mechanisms by which this complex responds to different levels of Hh signaling and establishes distinct domains of gene expression are not fully understood. By sequentially mutating components from the Ci signaling complex, their roles in each aspect of Ci regulation can be analyzed. The Cos2-Ci core complex is able to mediate Hh-regulated activation of Ci but is insufficient to regulate nuclear import and cleavage. Addition of Su(fu) to the core complex blocks nuclear import while the addition of Fu restores Hh regulation of Ci nuclear import and proteolytic cleavage. Fu participates in two partially redundant pathways to regulate Ci nuclear import: the kinase function plays a positive role by inhibiting Su(fu), and the regulatory domain plays a negative role in conjunction with Cos2 (Lefers, 2001).

In fu94;Su(fu)LP mutants, it is unlikely that either Fu or Su(fu) is present in the complex: Fu protein from class II mutants fails to immunoprecipitate Cos2, and Su(fu) protein cannot be detected in Su(fu)LP mutants. In this mutant combination, the processing of Ci is not Hh regulated, and this results in uniform levels of Ci protein across the entire anterior compartment. Hh regulation of Ci nuclear import is also lost, and the Ci protein shuttles into and out of the nucleus throughout the anterior compartment. As a consequence, dpp is expressed at modest levels in all anterior compartment cells. Previous studies have shown that Cos2 is required for Ci sequestration in the cytoplasm and its proteolytic processing, but clearly Cos2 is not sufficient for all aspects of Ci regulation. In the absence of the Fu regulatory domain and Su(fu) from the complex, all anterior compartment cells behave as if they are receiving at least modest levels of Hh signaling (Lefers, 2001).

Addition of Su(fu) to the Ci-Cos2 complex dramatically reduces the rate of Ci release from the complex as Ci does not accumulate in the nucleus in fu94 mutant discs that have been treated with leptomycin B (LMB), which blocks Ci nuclear export. No regulation of Ci nuclear import is observed, and processing of Ci into Ci75 is still inhibited. The block in Ci nuclear import by Su(fu) appears to be dependent on the presence of Cos2 as clones double mutant for fumH63;cos21 release Ci independent of Hh signaling (Lefers, 2001).

Addition of Fu to the Ci-Cos2 complex essentially restores Hh regulation of Ci nuclear import and the processing of Ci into Ci75. Therefore, in the absence of Hh signaling, Fu is required for both the cleavage of Ci into Ci75 and its retention in the cytoplasm. The major consequence of removing Su(fu) from the complex is a significant decrease in the overall levels of both Ci and Ci75. This decrease does not appear to significantly compromise Hh regulation (Lefers, 2001).

Although Cos2 provides an important tethering force, it apparently cannot hold Ci in the cytoplasm on its own. Addition of either the Fu regulatory domain or Su(fu) is sufficient to restore effective tethering. The requirement for Fu in Ci tethering is a new finding, since it has been shown that Fu plays a positive role in Ci nuclear entry by inhibiting Su(fu) via its kinase domain. Further examination of different classes of fu alleles demonstrates that Fu participates in Ci tethering through its regulatory domain. When a Fu class I mutant protein (kinase domain mutations) is added to the Ci-Cos2 core complex [fu1,Su(fu)LP], regulation of Ci nuclear entry is almost wild type. In contrast, when a Fu class II mutant protein (regulatory domain mutations) is present [fu94;Su(fu)LP or fuRX15;Su(fu)LP], the complex fails to tether Ci in the absence of Hh signaling. It has been shown that Fu interacts with Cos2 through its regulatory domain, and the proteins made by fu class II alleles fail to immunoprecipitate with Cos2. These results suggest that the interaction between Cos2 and the Fu regulatory domain is important for Cos2 to tether Ci in the absence of Su(fu) activity. This Cos2-Fu interaction may also be important for targeting Fu kinase regulation of Su(fu). Both fuRX15 and fu94, which delete different extents of the regulatory domain, might be expected to retain kinase function, yet Hh regulation of Su(fu) appears to have been lost and Ci is not released from the cytoplasm in either of these mutants. The simplest explanation is that by preventing Fu interaction with Cos2, Fu cannot perform its structural role in the complex nor can it regulate Su(fu). Thus, Fu plays two opposite roles in the regulation of Ci nuclear entry. Without Hh signaling, the regulatory domain in conjunction with Cos2 tethers Ci in the cytoplasm; upon Hh signaling, the kinase domain inhibits Su(fu) which, along with a change in the Cos2/Fu regulatory domain interaction, leads to the release of Ci (Lefers, 2001).

While it has been possible to clearly establish a role for Su(fu) in Ci nuclear import, its role in Ci activation and cleavage is less clear. In cells double mutant for cos2;Su(fu), Ci appears to be at least partially activated since double mutant clones away from the compartment boundary ectopically express en. A reasonable interpretation of these data is that Ci activation is inhibited by Su(fu) and signaling through Cos2 relieves such inhibition (Lefers, 2001).

But this cannot be the whole story. In Su(fu)LP discs, the expression of ptc or en is still tightly regulated and does not expand into all the cells with efficient Ci nuclear import. This regulation of Ci activity is evidently not rendered by the Fu regulatory domain, since it persists in the fu94;Su(fu)LP double mutants. It seems likely that Su(fu) is partially redundant with other factors that regulate Ci activation and that these yet to be identified factors function with Cos2 in the fu;Su(fu) double mutants. Su(fu) may also play some role in Ci cleavage. In the fu94;Su(fu)LP double mutants, the level of Ci seems significantly reduced relative to fu94 single mutants. In addition, Ci protein levels are not elevated across the entire anterior compartment in fuRX15 single mutants but are in fuRX15;Su(fu)LP double mutants. The implication of Su(fu) in these other aspects of Hh regulation suggests that while it is possible to dissect the complex and assign primary roles to the various components, the complex does normally function as a whole (Lefers, 2001).

The Drosophila protein Shaggy (Sgg, also known as Zeste-white3, Zw3) and its vertebrate ortholog glycogen synthase kinase 3 (GSK3) are inhibitory components of the Wingless (Wg) and Wnt pathways. Sgg is also a negative regulator in the Hedgehog (Hh) pathway. In Drosophila, Hh acts both by blocking the proteolytic processing of full-length Cubitus interruptus, Ci (Ci155), to generate a truncated repressor form(Ci75), and by stimulating the activity of accumulated Ci155. Loss of sgg gene function results in a cell-autonomous accumulation of high levels of Ci155 and the ectopic expression of Hh-responsive genes including decapentaplegic and wg. Simultaneous removal of sgg and Suppressor of fused, Su(fu), results in wing duplications similar to those caused by ectopic Hh signaling. Ci is phosphorylated by GSK3 after a primed phosphorylation by protein kinase A (PKA), and mutating GSK3 phosphorylation sites in Ci blocks its processing and prevents the production of the repressor form. It is proposed that Sgg/GSK3 acts in conjunction with PKA to cause hyperphosphorylation of Ci, which targets it for proteolytic processing, and that Hh opposes Ci proteolysis by promoting its dephosphorylation (Jia, 2002).

The proteolytic processing of Ci requires the activities of several intracellular Hh signaling components, including PKA and the kinesin-related protein Costal2 (Cos2). Overexpressing either Cos2 or a constitutively active form of PKA (mC*) blocks the accumulation of Ci155 induced by Hh. In contrast, wing discs overexpressing mC* or Cos2 accumulate high levels of Ci155 after treatment with 50 mMLiCl, a specific inhibitor of GSK3 kinase activity. These observations suggest that Sgg acts downstream of, or in parallel with, PKA and Cos2 to regulate Ci processing (Jia, 2002).

GSK3 is involved in multiple signaling pathways, raising the question of how its activity is selectively regulated by individual pathways. An emerging theme is that GSK3 is present, together with its substrates, in distinct complexes that are regulated by different upstream stimuli. Future study will determine whether Sgg/GSK3 forms a complex with Cos2 or Ci and whether Hh regulates Sgg/ GSK3 within the complex. In vertebrates, three Gli proteins (Gli1, Gli2 and Gli3) are implicated in transducing Hh signals. Interestingly, all three Gli proteins contain multiple GSK3-phosphorylation consensus sites adjacent to PKA sites, raising the possibility that GSK3 may regulate Gli proteins in vertebrate Hh pathways. Hh and Wnt signaling pathways act in synergy in certain developmental contexts. The finding that GSK3 is involved in both Hh and Wnt pathways raises the possibility that these two pathways might converge at GSK3 in certain developmental processes (Jia, 2002).

Altered localization of Smoothened protein activates Hedgehog signal transduction

Hedgehog (Hh) signaling is critical for many developmental events and must be restrained to prevent cancer. A transmembrane protein, Smoothened (Smo), is necessary to transcriptionally activate Hh target genes. Smo activity is blocked by the Hh transmembrane receptor Patched (Ptc). The reception of a Hh signal overcomes Ptc inhibition of Smo, activating transcription of target genes. Using Drosophila salivary gland cells in vivo and in vitro as a new assay for Hh signal transduction, the regulation of Hh-triggered Smo stabilization and relocalization was investigated. Hh causes Smo (GFP-Smo) to move from internal membranes to the cell surface. Relocalization is protein synthesis-independent and occurs within 30 min of Hh treatment. Ptc and the kinesin-related protein Costal2 (Cos2) cause internalization of Smo, a process that is dependent on both actin and microtubules. Disruption of endocytosis by dominant negative dynamin or Rab5 prevents Smo internalization. Fly versions of Smo mutants associated with human tumors are constitutively present at the cell surface. Forced localization of Smo at the plasma membrane activates Hh target gene transcription. Conversely, trapping of activated Smo mutants in the ER prevents Hh target gene activation. Control of Smo localization appears to be a crucial step in Hh signaling in Drosophila (Zhu, 2003).

Movement of Smo to the surface requires actin and tubulin components of the cytoskeleton, though the relevant motors are unknown. Cos2 is an unusual member of the kinesin family, with sequence features at odds with conventional ATPase binding site structure. Cos2 could be either a motor or a tether. Cos2 could have a role in controlling movements of vesicles that contain Smo. Overproduction of Cos2 alters GFP-Smo localization, and furthermore, prevents Hh from bringing much GFP-Smo to the surface, and the GFP-Smo that does reach the surface is located in discreet dots. Ptc also blocked Hh from bringing GFP-Smo to the surface, but no such dots were observed. Overexpression of a presumably irrelevant other motor protein, Nod, has no effect on localization of GFP-Smo. Cos2 production may therefore specifically cause the movement of Smo-containing organelles to discreet locations on the membrane, either tethering them to the cytoskeleton at specific locations or causing a coalescence effect at random locations. Cos2 has been envisioned as functioning as part of a cytoplasmic complex whose activity in processing the Ci transcription factor is controlled by Smo. The present data suggest a new function in which the complex (oralternatively, Cos2 independently of the complex), feeds back to alter Smo activity. It is interesting that both GFP-Smo (when Cos2 and Hh were coexpressed) and PtcDN-YFP exhibits a similar punctate cell surface localization pattern. PtcDN may function through competing with endogenous Ptc, raising an intriguing alternative possibility that Cos2 may interact directly with Smo to control Smo subcellular localization (Zhu, 2003).

Smoothened regulates alternative configurations of a regulatory complex that includes Fused, Costal, Suppressor of Fused and Cubitus interruptus

In the Drosophila wing, Hedgehog is made by cells of the posterior compartment and acts as a morphogen to pattern cells of the anterior compartment. High Hedgehog levels instruct L3/4 intervein fate, whereas lower levels instruct L3 vein fate. Transcriptional responses to Hedgehog are mediated by the balance between repressor and activator forms of Cubitus interruptus, CiR and CiA. Hedgehog regulates this balance through its receptor, Patched, which acts through Smoothened and thence a regulatory complex that includes Fused, Costal, Suppressor of Fused and Cubitus interruptus. It is not known how the Hedgehog signal is relayed from Smoothened to the regulatory complex nor how responses to different levels of Hedgehog are implemented. Chimeric and deleted forms of Smoothened were used to explore the signaling functions of Smoothened. A Frizzled/Smoothened chimera containing the Smo cytoplasmic tail (FFS) can induce the full spectrum of Hedgehog responses but is regulated by Wingless rather than Hedgehog. Smoothened whose cytoplasmic tail is replaced with that of Frizzled (SSF) mimics fused mutants, interfering with high Hedgehog responses but with no effect on low Hedgehog responses. The cytoplasmic tail of Smoothened with no transmembrane or extracellular domains (SmoC) interferes with high Hedgehog responses and allows endogenous Smoothened to constitutively initiate low responses. SmoC mimics costal mutants. Genetic interactions suggest that SSF interferes with high signaling by titrating out Smoothened, whereas SmoC drives constitutive low signaling by titrating out Costal. These data suggest that low and high signaling (1) are qualitatively different, (2) are mediated by distinct configurations of the regulatory complex and (3) are initiated by distinct activities of Smoothened. A model is presented where low signaling is initiated when a Costal inhibitory site on the Smoothened cytoplasmic tail shifts the regulatory complex to its low state. High signaling is initiated when cooperating Smoothened cytoplasmic tails activate Costal and Fused, driving the regulatory complex to its high state. Thus, two activities of Smoothened translate different levels of Hedgehog into distinct intracellular responses (Hooper, 2003).

Analyses of the activities of truncated and chimeric forms of Smo in a variety of genetic backgrounds yielded four principal observations. (1)The FFS chimera activates the full spectrum of Hh responses, but is regulated by Wg rather than Hh. From this, it is concluded that the Smo cytoplasmic tail initiates all intracellular responses to Hh, while the remainder of Smo regulates activity of the tail. (2) The SSF chimera interferes with high signaling but has no effect on low signaling. SSF mimics Class II fu mutants and is suppressed by increasing smo+ but not fu+ or cos+. From this, it is concluded that high Hh instructs Smo to activate Fu by a mechanism that is likely to involve dimeric/oligomeric Smo. (3) The cytoplasmic tail of Smo (SmoC) derepresses endogenous Smo activity in the absence of Hh and represses endogenous Smo activity in the presence of high Hh. That is, SmoC drives cells to the low response regardless of Hh levels. This mimics cos mutants and is suppressed by 50% increase in cos+. From this, it is concluded that low Hh instructs Smo to inactivate Cos, by a mechanism that may involve stoichiometric interaction between Cos and the Smo cytoplasmic tail. (4) Chimeras where the extracellular CRD and TM domains are mismatched fail to exhibit any activity. From this, it is concluded that these two domains act as an integrated functional unit. This leads to a model for signaling where Fz or Smo can adopt three distinct states, regulating two distinct activities and translating different levels of ligand into distinct responses. Many physical models are consistent with these genetic analyses (Hooper, 2003).

Two mutant forms of Smo have been identified that regulate downstream signaling through different activities. These mutant forms of Smo mimic phenotypes of mutants in other components of the Hh pathway, as well as normal responses to different levels of Hh. These data suggest a model where Smo can adopt three distinct states that instruct three distinct states of the Ci regulatory complex. The model further suggests that Smo regulates Ci through direct interactions between Fu, Cos and the cytoplasmic tail of Smo. This is consistent with the failure of numerous genetic screens to identify additional signaling intermediates, and with the exquisite sensitivity of low signaling to Cos dosage (Hooper, 2003).

The model proposes that Smo can adopt three states, a decision normally dictated by Hh, via Ptc. The Ci regulatory complex, which includes full-length Ci, Cos and Fu, likewise can adopt three states. (1) In the absence of Hh Smo is OFF. Its cytoplasmic aspect is unavailable for signaling. The Cos/Fu/Ci regulatory complex is anchored to microtubules and promotes efficient processing of Ci155 to CiR. (2) Low levels of Hh expose Cos inhibitory sites in the cytoplasmic tail of Smo. Cos interaction with these sites drives the Ci regulatory complex into the low state, which recruits Su(fu) and makes little CiR or CiA. (3) High levels of Hh drive a major change in Smo, possibly dimerization. This allows the cytoplasmic tails of Smo to cooperatively activate Fu and Cos. Fu* and Cos* (* indicates the activated state) then cooperate to inactivate Su(fu), to block CiR production, and to produce CiA at the expense of Ci155 (Hooper, 2003).

The OFF state is normally found deep in the anterior compartment where cells express no Hh target genes (except basal levels of Ptc). In this state, the Ci regulatory complex consists of Fu/Cos/Ci155. Cos and Fu contribute to efficient processing of Ci155 to the repressor form, CiR, presumably because the complex promotes access of PKA and the processing machinery to Ci155, correlating with microtubule binding of the complex. This state is universal in hh or smo mutants, indicating that intracellular responses to Hh cannot be activated without Smo. Therefore Smo can adopt an OFF state where it exerts no influence on downstream signaling components and the OFF state of the Ci regulatory complex is its default state (Hooper, 2003).

The low state is normally found approximately five to seven cells from the compartment border, where cells are exposed to lower levels of Hh. These cells express Iro, moderate levels of dpp, no Collier and basal levels of Ptc. They accumulate Ci155, indicating that little CiA or CiR is made. Ci155 can enter nuclei but is insufficient to activate high responses. The physical state of the Ci regulatory complex in the low state has not been investigated. Cells take on the low state regardless of Hh levels when Ci is absent or when SmoC is expressed, and are strongly biased towards that state in fu(classII); Su(fu) double mutants. This state normally requires input from Smo, which becomes constitutive in the presence of SmoC. Because SmoC drives only low responses and cannot activate high responses, this identifies a low state of Smo that is distinct from both OFF and high. It is proposed that the low state is normally achieved when Smo inactivates Cos, perhaps by direct binding of Cos to Smo and dissociation of Cos from Ci155. Neither CiR nor CiA is made efficiently, and target gene expression is similar to that of ci null mutants (Hooper, 2003).

The high state is normally found in the two or three cells immediately adjacent to the compartment border where there are high levels of Hh. These cells express En, Collier, high levels of Ptc and moderate levels of Dpp. They make CiA rather than CiR, and Ci155 can enter nuclei. In this state a cytoplasmic Ci regulatory complex consists of phosphorylated Cos, phosphorylated Fu, Ci155 and Su(fu). Dissociation of Ci from the complex may not precede nuclear entry, since Cos, Fu, and Sufu are all detected in nuclei along with Ci155. Sufu favors the low state, whereas Cos and Fu cooperate to allow the high state by repressing Sufu, and also by a process independent of Sufu. This high state is the universal state in ptc mutants and requires input from Smo. As this state is specifically lost in fu mutants, Fu may be a primary target through which Smo activates the high state. SSF specifically interferes with the high state by a mechanism that is most sensitive to dosage of Smo. This suggests SSF interferes with the high activity of Smo itself. It is suggested that dimeric/oligomeric Smo is necessary for the high state, and that Smo:SSF dimers are non-productive. Cooperation between Smo cytoplasmic tails activates Fu and thence Cos. The activities of the resulting Fu* and Cos* are entirely different from their activities in the OFF state, and mediate downstream effects on Sufu and Ci (Hooper, 2003).

The cytoplasmic tail of Smo is sufficient to activate all Hh responses, and its activity is regulated through the extracellular and TM domains. This is exemplified by the FFS chimera, which retains the full range of Smo activities, but is regulated by Wg rather than Hh. The extracellular and transmembrane domains act as an integrated unit to activate the cytoplasmic tail, since all chimeras interrupting this unit fail to activate any Hh responses, despite expression levels and subcellular localization similar to those of active SSF or FFS. As is true of other serpentine receptors, a global rearrangement of the TM helices is likely to expose 'active' (Cos regulatory?) sites on the cytoplasmic face of Smo. The extracellular domain of Smo must stabilize this conformation and Ptc must destabilize it. But how? Ptc may regulate Smo through export of a small molecule, which inhibits Smo when presented at its extracellular face. Hh binding to Ptc stimulates its endocytosis and degradation, leaving Smo behind at the cell surface. Thus, Hh would separate the source of the inhibitor (Ptc) from Smo, allowing Smo to adopt the low state. Transition from low to high might require Smo hyperphosphorylation. The high state, which is likely to involve Smo oligomers, might be favored by cell surface accumulation if aggregation begins at some threshold concentration of low Smo. Alternatively, these biochemical changes may all be unnecessary for either the low or high states of Smo (Hooper, 2003).

Smoothened transduces Hedgehog signal by physically interacting with Costal2/Fused complex through its C-terminal tail

The Hedgehog (Hh) family of secreted proteins controls many aspects of growth and patterning in animal development. The seven-transmembrane protein Smoothened (Smo) transduces the Hh signal in both vertebrates and invertebrates; however, the mechanism of its action remains unknown. Smo lacking its C-terminal tail (C-tail) is inactive, whereas membrane-tethered Smo C-tail has constitutive albeit low levels of Hh signaling activity. Smo is shown to physically interact with Costal2 (Cos2) and Fused (Fu) through its C-tail. Deletion of the Cos2/Fu-binding domain from Smo abolishes its signaling activity. Moreover, overexpressing Cos2 mutants that fail to bind Fu and Ci but retain Smo-binding activity blocks Hh signaling. Taken together, these results suggest that Smo transduces the Hh signal by physically interacting with the Cos2/Fu protein complex (Jia, 2003).

The most surprising finding of this study is that the Smo C-tail suffices to induce Hh pathway activation. Overexpressing the membrane-tethered Smo C-tail (Myr-SmoCT, Sev-SmoCT) blocks Ci processing, induces dpp-lacZ expression, and stimulates nuclear translocation of Ci155. Myr-SmoCT is refractory to Ptc inhibition and activates Hh-pathway independent of endogenous Smo. Membrane tethering appears to be crucial for the Smo C-tail to activate the Hh pathway; untethered SmoCT has no signaling activity. This is consistent with observations that cell surface accumulation of Smo correlates with its activity (Jia, 2003).

Although the Smo C-tail has constitutive Hh signaling activity, it does not possess all the activities associated with full-length Smo. For example, overexpressing Myr-SmoCT in A-compartment cells away from the A/P compartment boundary does not significantly activate ptc and en, which are normally induced by high levels of Hh. In addition, Myr-SmoCT cannot substitute endogenous Smo at the A/P compartment boundary to transduce high levels of Hh signaling activity, since boundary smo mutant cells expressing Myr-SmoCT fail to express ptc in response to Hh (Jia, 2003).

The failure of the Smo C-tail to transduce high Hh signaling activity is due to its inability to antagonize Su(fu). Although Myr-SmoCT blocks Ci processing to generate Ci75, the activity of Ci155 accumulated in Myr-SmoCT-expressing cells is still blocked by Su(fu); removal of Su(fu) function from Myr-SmoCT-expressing cells allows Ci155 to activate ptc to high levels. Because Myr-SmoCT stimulates nuclear translocation of Ci155, the inhibition of Ci155 by Su(fu) in Myr-SmoCT-expressing cells must rely on a mechanism that is independent of impeding Ci nuclear translocation (Jia, 2003).

Several observations prompted a determination of whether Smo can transduce the Hh signal by physically interacting with the Cos2/Fu complex: (1) although Smo is related to G-protein-coupled receptors, no genetic or pharmacological evidence has been obtained to support the involvement of a G-protein in a physiological Hh signaling process; (2) Myr-SmoCT can interfere with the ability of endogenous Smo to transduce high levels of Hh signaling activity, which can be offset by increasing the amount of full-length Smo. This implies that Myr-SmoCT may compete with full-length Smo for binding to limiting amounts of downstream signaling components. (3) Extensive genetic screens failed to identify Hh signaling components that may link Smo to the Cos2/Fu complex (Jia, 2003).

Using a coimmunoprecipitation assay, it was demonstrated that Smo interacts with the Cos2/Fu complex both in S2 cells and in wing imaginal discs, and the Smo C-tail appears to be both necessary and sufficient to mediate this interaction. The Cos2/Fu-binding domain was narrowed down to the C-terminal half of the Smo C-tail (between amino acids 818 and 1035). Furthermore, both the microtubule-binding domain (amino acids 1-389) and the C-terminal tail (amino acids 990-1201) of Cos2 interact with Smo. Since none of these Cos2 domains binds Fu, this implies that the Cos2/Smo interaction is not mediated through Fu. Ci is also dispensable for Smo/Cos2/Fu interaction; Smo binds Cos2/Fu in S2 cells in which Ci is not expressed. However, the results did not rule out the possibility that Smo could interact with the Cos2/Fu complex through multiple contacts. For example, Smo could simultaneously contact Cos2 and Fu. Nor was it demonstrated that binding of Cos2 to Smo is direct. Indeed, no protein-protein interaction between Smo and Cos2 was detected in yeast. It is possible that a bridging molecule(s) is required to link Smo to the Cos2/Fu complex. Alternatively, Smo needs to be modified in vivo in order to bind Cos2. It has been shown that Hh stimulates phosphorylation of Smo; hence, it is possible that phosphorylation of Smo might be essential for recruiting the Cos2/Fu complex (Jia, 2003).

Several lines of evidence suggest that Smo/Cos2/Fu interaction is important for Hh signal transduction. (1) Deletion of the Cos2-binding domain from Smo, either in the context of full-length Smo or the Smo C-tail, abolishes Smo signaling activity. (2) Overexpressing Cos2 deletion mutants that no longer bind Fu and Ci but retain a Smo-binding domain intercept Hh signal transduction. Genetic evidence has been provided that Cos2 has a positive role in transducing Hh signal in addition to its negative influence on the Hh pathway, since Ci155 is no longer stimulated into labile and hyperactivity forms by high levels of Hh in cos2 mutant cells. In light of the finding that Smo interacts with Cos2/Fu, the simplest interpretation for a positive role of Cos2 is that it recruits Fu to Smo and allows Fu to be activated by Smo in response to Hh (Jia, 2003).

Of note, interaction between SmoCT and Cos2/Fu per se is not sufficient for triggering Hh pathway activation. For example, Myr-SmoCTDelta625-818, which binds Cos2/Fu to the same extent as Myr-SmoCT, does not possess Hh signaling activity. The fact that Myr-SmoDeltaCT625-730 and Myr-Smo730-1035 can activate the Hh pathway suggests that Smo sequence between amino acids 730 and 818 is essential. This domain may recruit factors other than Cos2/Fu to achieve Hh pathway activation. Alternatively, it might target SmoCT to an appropriate signaling environment (Jia, 2003).

An important property of Hh family members in development is that they can elicit distinct biological responses via different concentrations. How different thresholds of Hh signal are transduced by Smo to generate distinct transcriptional outputs is not understood. The results suggest that Smo can function as a molecular sensor that converts quantitatively different Hh signals into qualitatively distinct outputs. In the absence of Hh, the cell surface levels of Smo are low. In addition, the Smo C-tail may adopt a 'closed' conformation that prevents it from binding to Cos2/Fu. Low levels of Hh partially inhibit Ptc, leading to an increase of Smo on the cell surface. In addition, the Smo C-tail may adopt an 'open' conformation, which allows Smo to bind the Cos2/Fu complex and inhibit its Ci-processing activity. Low levels of Hh signaling activity can be mimicked by overexpression of either full-length Smo or membrane-tethered forms of the Smo C-tail. High levels of Hh completely inhibit Ptc, resulting in a further increase in Smo signaling activity. Hyperactive Smo stimulates the phosphorylation and activity of bound Fu, which in turn antagonizes Su(fu) to activate Ci155. Consistent with this, Fu bound to Myc-Smo was found to became phosphorylated in response to ectopic Hh (Jia, 2003).

The Smo sequence N terminus to SmoCT (SmoN) appears to be essential for conferring high Smo activities. It is not clear how SmoN modulates the activity of SmoCT. SmoN might recruit additional effector(s) or target SmoCT to a microdomain with a more favorable signaling environment. Alternatively, SmoN might function as a dimerization domain that facilitates interaction between two SmoCTs, as in the case of receptor tyrosine kinases. It is also not clear how Smo/Cos2/Fu interaction inhibits Ci processing. One possibility is that Smo/Cos2 interaction may cause disassembly of the Cos2/Ci complex, which could prevent Ci from being hyperphosphorylated; Cos2/Ci complex formation might be essential for targeting Ci to its kinases. Consistent with this view, Ci is barely detectable in the Cos2/Fu complex bound to Smo (Jia, 2003).

Physical association of the receptor complex with a downstream signaling component has also been demonstrated for the canonic Wnt pathway whereby the Wnt coreceptor LRP-5 interacts with Axin, a molecular scaffold in the Wnt pathway. Hence, Hh and Wnt/Wg pathways appear to use a similar mechanism to transmit signal downstream of their receptor complexes (Jia, 2003).

Identification of a functional interaction between the Smoothened and Costal2

Hedgehog signal transduction is initiated when Hh binds to its receptor Patched (Ptc), activating the transmembrane protein Smoothened (Smo). Smo transmits its activation signal to a microtubule-associated Hedgehog signaling complex (HSC). At a minimum, the HSC consists of the Kinesin-related protein Costal2 (Cos2), the protein kinase Fused (Fu), and the transcription factor Cubitus interruptus (Ci). In response to HSC activation, the ratio between repressor and activator forms of Ci is altered, determining the expression levels of various Hh target genes. The steps between Smo activation and signaling to the HSC have not been described. A functional interaction is described between Smo and Cos2 that is necessary for Hh signaling. It is proposed that this interaction is direct and allows for activation of Ci in response to Hh. This work fills in the last major gap in the understanding of the Hh signal transduction pathway by suggesting that no intermediate signal is required to connect Smo to the HSC (Ogden, 2003).

To determine whether Cos2 and Smo could interact directly, a directed yeast two-hybrid assay was used. The cytoplasmic carboxyl-terminal domain of Smo (SmoC) was used in the two-hybrid assay, since the signaling capabilities of Smo appear to reside within this domain. The carboxyl-terminal domain of Smo interacts with Cos2, though this interaction appears less efficient than that of Cos2 with Fu. This interaction is specific and reproducible, since there is no growth when the open reading frame of Cos2 is inserted in the reverse orientation. These results demonstrate that the carboxyl-terminal domain of Smo is sufficient to associate with Cos2 and that this association appears to be direct. Combined with immunoprecipitation and immunofluorescence data, the yeast two-hybrid results provide strong evidence that Smo and Cos2 directly associate and that the association occurs within the intracellular signaling portion of Smo (Ogden, 2003).

To determine whether Hh signaling would affect the Cos2-Smo interaction, Smo was immunoprecipitated from S2 cell lysates prepared from cells transfected with Hh expression or control vectors. Cos2 and Fu coimmunoprecipitate with Smo at similar levels regardless of Hh activation status. Phosphorylation-induced mobility shifts of Cos2 occurs in Hh-transfected cells, verifying that Hh signaling is intact. The modest increase observed in Cos2 immunoprecipitating with Smo in response to Hh stimulation may be accounted for by Smo protein stabilization in response to Hh. These results suggest that interactions between Smo, Cos2, and Fu are relatively stable and independent of Hh activation status (Ogden, 2003).

To verify that Hh activation does not modify Smo-Cos2 association in vivo, Smo immunoprecipitations were performed from embryos engineered to overexpress Ptc, Hh, or neither. Embryos overexpressing Ptc serve as a source of cells in which Hh signaling is inactive due to repression of Smo by Ptc, while embryos overexpressing Hh serve as a source of Hh-activated cells. Mobility shifts of Cos2, Fu, and Smo, which have previously been attributed to Hh-induced phosphorylation, confirm that Hh or Ptc have turned Hh signaling on or off in these embryos. An equal amount of Cos2 was observed coimmunoprecipitating with Smo from wild-type, Ptc, and Hh embryo lysates. In two separate experiments, it was estimated that 3% of Cos2, 4% of Fu, and 3%-8% of Smo were recovered in coimmunoprecipitates. By contrast, 50% of Fu was recovered by Cos2 immunoprecipitation, while negligible amounts of Fu, Cos2, or Smo were recovered in Fz immunoprecipitates. These results demonstrate that a small percentage of total Cos2 and Smo are associated in a high-affinity association, and the percentage associated does not change due to Hh signaling (Ogden, 2003).

Expression of a chimera of SmoC fused to a myristate membrane-targeting sequence (Myr-SmoC) induces phenotypes in Drosophila similar to cos2 loss-of-function mutations; weak Hh responses are activated, while strong Hh responses are inhibited. Myr-SmoC drives all Hh responses to a weak activation state in Drosophila and requires endogenous Smo to do so. Although the mechanism by which Myr-SmoC acts is unknown, dosage dependence of the effect suggests that it interferes with signaling by competing with endogenous Smo for Cos2. A similar epitope-tagged construct was expressed in cultured cells to test the hypothesis that Myr-SmoC interferes with signaling by binding to Cos2. Using a Myc epitope tag to specifically immunoprecipitate Myr-SmoC, it was found that both Cos2 and Fu associate with Myr-SmoC. These data support the directed two-hybrid experiment, showing that the carboxyl-terminal domain of Smo is sufficient to interact with Cos2. Further, Myr-SmoC functions was found to be a potent inhibitor of Hh signaling, able to inhibit Hh-dependent transcription in a dose-dependent fashion. These results indicate that even in the absence of Hh, Ci activity is effectively reduced by Myr-SmoC. Thus, Myr-SmoC does not constitutively activate Ci in this reporter assay. It is proposed that Myr-SmoC can act as a dominant negative by binding endogenous Cos2. This argument is bolstered by genetic evidence showing that increasing Cos2 levels in vivo can suppress the overgrowth phenotype associated with expressing Myr-SmoC in flies. These results are consistent with the hypothesis that association between Smo and Cos2 is necessary for Hh signaling to be propagated to its ultimate effector, the transcription factor Ci (Ogden, 2003).

Two scenarios are proposed that may account for the observation that Smo and Cos2 association is not altered in response to Hh. The first possibility is that Smo and Cos2 may be held in an associated but inactive state in the absence of Hh stimulation, presumably through the function of Ptc. Hh stimulation would relieve Ptc-mediated repression of the Smo-Cos2 complex to allow Smo relocalization to the plasma membrane. The Kinesin-like properties of Cos2 and its direct interaction with Smo may facilitate this relocalization. A second possibility is that the dynamics of association are changed in response to Hh, such that Smo and Cos2 association turns over more rapidly in the process of creating the active form of Ci (Ogden, 2003).

Hedgehog signal transduction via Smoothened association with a cytoplasmic complex scaffolded by the atypical kinesin, Costal-2

The seven-transmembrane protein Smoothened (Smo) transduces extracellular activation of the Hedgehog (Hh) pathway by an unknown mechanism to increase transcriptional activity of the latent cytoplasmic transcription factor Ci (Cubitus interruptus). Evidence is presented that Smo associates directly with a Ci-containing complex that is scaffolded and stabilized by the atypical kinesin, Costal-2 (Cos2). This complex constitutively suppresses pathway activity, but Hh signaling reverses its regulatory effect to promote Ci-mediated transcription. In response to Hh activation of Smo, Cos2 mediates accumulation and phosphorylation of Smo at the membrane as well as phosphorylation of the cytoplasmic components Fu and Su(fu). Positive response of Cos2 to Hh stimulation requires a portion of the Smo cytoplasmic tail and the Cos2 cargo domain, which interacts directly with Smo (Lum, 2003).

Early studies of Cos2 suggested primarily a negative role for Cos2 in pathway regulation, as manifested by phenotypic analysis of cos2 mutations and by a requirement for Cos2 in cytoplasmic retention of Ci and in its proteolytic processing to produce CiR. More recent studies have suggested a potential positive role based on a requirement for Cos2 in transcriptional activation of gene targets associated with highest levels of pathway activity. This study extends the evidence for such a positive role by demonstrating: (1) a requirement for Cos2 in mediating a series of Hh-induced biochemical changes in pathway components; (2) an association between the Cos2/Fu/Ci complex and Smo; (3) a direct interaction between Smo and Cos2, and (4) a requirement for Cos2 in highest level Hh pathway response in cultured cell reporter assays and Hh-induced morphogenesis in the dorsal cuticle (Lum, 2003).

Based on the sequence relationship between Smo and GPCRs, previous speculation and experimental work has focused on the possibility that Smo may interact with heterotrimeric G proteins. G protein components have been systematically targeted using RNAi in a cultured cell signaling assay, and no significant role has been found for G proteins in transcriptional regulation via Ci. A potential role cannot be ruled out for G proteins or other mediators in cellular responses to Hh signaling that do not involve transcriptional regulation via Ci/Gli. For example, a recently described chemoattractant activity for Shh in axon guidance appears to be mediated by Smo, yet proceeds in a short time scale and with a local cell polarity that suggests a possible nonnuclear mechanism of response (Lum, 2003).

In searching for other mediators of information transfer from membrane to cytoplasm, it was surprising to find cytoplasmic components copurifying with Smo. Since these were the only Drosophila proteins identified in the bands excised, it is concluded that these complexes were highly pure, and that Smo associates stably with components of the cytoplasmic complex in vivo. It is further demonstrated that Smo interacts directly with Cos2, which scaffolds this complex. Consistent with these findings, articles presenting genetic evidence for a role of the Smo cytoplasmic tail in Hh signaling and evidence suggesting a physical interaction between Smo and Cos2 were published during final review of this work. Additional direct association of Smo with other complex components have not been ruled out (Lum, 2003 and references therein).

The identification of a complex that includes both Ci and Smo immediately suggested that recruitment of the cytoplasmic complex to Smo upon Hh stimulation might be critical for pathway activity. Cos2 plays a central role in mediating this association, functioning both as a scaffold that brings together cytoplasmic components (the Cos2 complex) and as a sensor that monitors the state of pathway activation by interacting with Smo at the membrane (Lum, 2003).

In the unstimulated state, Smo levels are low, and most of the Cos2 complex therefore is not associated with or influenced by Smo; even the small fraction of Cos2 complex associated with Smo may be negative, since Smo may not be in an active state. The negative form of the Cos2 complex presumably mediates production of CiR and prevents nuclear accumulation of Ci, resulting in a net suppression of transcriptional targets. In the intermediate state, present after a few minutes of stimulation, Smo protein has become activated by Hh stimulation but has not yet accumulated. Therefore, despite a positive state for the specific Cos2 complexes affected by Smo, the low level of activated Smo protein is insufficient for interaction with most of the Cos2 complexes. The net outcome with regard to transcriptional targets thus remains negative (Lum, 2003).

In stimulated cells, activated Smo exerts a pervasive influence on the Cos2 complex, either through a stable association or through a transient association with enduring effects. This association presumably involves a direct binding interaction between a portion of the Smo cytotail and the Cos2 cargo and stalk domains. The evidence suggests that both activation and accumulation of Smo are critical, as evidenced by the observation that moderate Smo overexpression alone is unable to fully activate the pathway in the absence of Hh stimulation. Similarly, a cycloheximide block of Smo accumulation dramatically limits the biochemical changes normally induced by Hh stimulation, and cotransfection experiments demonstrate that transition of Cos2 from a pathway suppressor to activator requires adequate levels of activated Smo. The activation of transcriptional targets resulting from pathway stimulation presumably result from the positive action of Cos2 on Fu and from loss of the ability to produce CiR (Lum, 2003).

It is the dual action of Cos2 in promoting formation of CiR and in stabilization and possible activation of Fu that leads to the apparent dual negative and positive roles of Cos2 in pathway regulation. Pathway activation induced by loss of Cos2 thus results from loss of CiR, but this activation is only partial because Fu is also destabilized, and the pathway-suppressing action of Su(fu) is unchecked. Consistent with this interpretation, a combined loss of Cos2 and Su(fu) results in maximal pathway activation irrespective of the presence of Hh (Lum, 2003).

How does Cos2 function? Motor proteins recently have been found to play a role in regulation of transmembrane receptors, as in the case of rhodopsin and mannose-6-phosphate receptor. These motor proteins apparently regulate receptor localization but do not play a direct role in receptor function. A recent report suggests that Cos2 overexpression indeed may influence Smo localization. The evidence, however, suggests that Cos2 also functions as a primary sensor of the state of pathway activation by interacting with Smo at the membrane and by scaffolding and stabilizing downstream pathway components (Lum, 2003).

Kinesins likely share an evolutionary origin with G proteins and myosins, and all three types of proteins use a conserved mechanism to couple nucleotide hydrolysis to dramatic conformational changes in protein structure. Kinesins utilize this mechanism to generate force that allows movement on microtubules. Sequences essential for microtubule binding, for nucleotide binding, and for motor function in other kinesins, however, are not conserved among Cos2 sequences in the Diptera and are not required for Cos2 function in cultured cells. It is still possible, however, that Cos2 retains the capacity for a conformational shift that may be triggered by Smo activation. If so, then this conformational change may involve the Cos2 cargo domain, which binds to Smo at the cytoplasmic tail and is likely required for Smo responsiveness (Lum, 2003).

Whatever its mechanism of activation, evidence points to a role for the atypical kinesin Cos2 as a scaffold and sensor that functions as the pivotal component in transduction of pathway activation from the seven transmembrane receptor Smo to the latent cytoplasmic transcription factor Ci. In addition to stabilizing Fu and mediating forward signaling events that affect other cytoplasmic components and Ci, Cos2 is also required for the accumulation of activated Smo, a critical aspect of producing a full response to Hh signal. Cos2 thus functions not just as a passive sensor for the state of pathway activation at the membrane, but is also an active participant in the cellular dynamics of transition from the unstimulated to the stimulated state. These activities are all the more remarkable in view of the well-recognized role played by Cos2 in maintaining the unstimulated state of the Hh pathway, and together these findings suggest that Cos2 dynamics are a critical determinant of intracellular Hh pathway response and regulation (Lum, 2003).

Sxl is in a complex that contains all of the known Hh cytoplasmic components: Hh promotes the entry of Sxl into the nucleus in the wing disc

The sex determination master switch, Sex-lethal (Sxl), controls sexual development as a splicing and translational regulator. Hedgehog (Hh) is a secreted protein that specifies cell fate during development. Sxl is in a complex that contains all of the known Hh cytoplasmic components, including Cubitus interruptus (Ci) the only known target of Hh signaling. Hh promotes the entry of Sxl into the nucleus in the wing disc. In the anterior compartment, the Hh receptor Patched (Ptc) is required for this effect, revealing Ptc as a positive effector of Hh. Some of the downstream components of the Hh signaling pathway also alter the rate of Sxl nuclear entry. Mutations in Suppressor of Fused or Fused with altered ability to anchor Ci are also impaired in anchoring Sxl in the cytoplasm. The levels, and consequently, the ability of Sxl to translationally repress downstream targets in the sex determination pathway, can also be adversely affected by mutations in Hh signaling genes. Conversely, overexpression of Sxl in the domain that Hh patterns negatively affects wing patterning. These data suggest that the Hh pathway impacts on the sex determination process and vice versa and that the pathway may serve more functions than the regulation of Ci (Horabin, 2003).

Sxl co-immunoprecipitates with Cos2 and Fu in the female germline. Since Ci is not expressed in germ cells, it is probable that a different Hh cytoplasmic complex might exist in germ cells. In somatic cells, Sxl is expressed in all female cells while Ci is expressed in only a subset. To test whether the Hh pathway differentiates between the two proteins in somatic cells, Sxl was immunoprecipitated from embryonic extracts and the immunoprecipitates probed for the various Hh cytoplasmic components. The immunoprecipitates showed that Cos2, Fu and Ci are complexed with Sxl. The specificity of this association of Sxl with the Hh pathway components was verified using antibodies to either Ci or Su(fu), and testing the immunoprecipitates for the presence of Sxl. Both co-immunoprecipitated with Sxl. The Ci immunoprecipitate was also tested for another Hh cytoplasmic component, Fu, which was present as expected. These interactions are maintained in a Su(fu)LP background (protein null allele). An IP of Ci from Su(fu)LP embryos brought down Sxl, as well as Fu and Cos2. Taken together, these data suggest that cells that express Ci and Sxl have both proteins in the same complex with the known cytoplasmic components of the Hh signaling pathway (Horabin, 2003).

Previous work on the germline has suggested that the Hh signaling pathway affects the intracellular trafficking Sxl. The cross talk between these two developmental pathways has been analyzed in tissues where both Hh targets can be present in the same cell. While analysis of embryos only uncovered an effect of Cos2 on Sxl, analysis of wing discs allowed several specific effects to be uncovered. At least three new functional aspects of the Hh pathway are suggested:

  1. More than one 'target' protein can exist in the Hh cytoplasmic complex.
    Immunoprecipitation experiments using extracts from embryos indicate that Sex-lethal and the known Hh signaling target Ci are in the same complex. The two proteins can co-immunoprecipitate each other as well as other known members of the Hh cytoplasmic complex. Even when Su(fu), the cytoplasmic component that most strongly anchors Sxl in the cytoplasm, is removed, Sxl can still be co-immunoprecipitated with both Ci and Fu. As a whole, these results suggest that at least some proportion of the two Hh 'target' proteins are in a common complex within the cell. Additionally, the wing defects produced when Sxl is overexpressed in the Hh signaling region suggest that their relative concentrations are important for their normal functioning (Horabin, 2003).
  2. The Hh targets can be affected differentially.
    The presence of two 'targets' within the Hh cytoplasmic complex, raises the question of how they can be differentially affected. The data show that the various members of the Hh pathway do not affect Sx1 and Ci similarly. Smo appears to be dispensable for the transmission of the Hh signal in promoting Sx1 nuclear entry, while Smo is critical for the activation of Ci. Conversely, while Ptc is essential for the effect of Hh on Sxl, it is dispensable for the activation of Ci. The Fu kinase (fumH63 background) also appears to have no role in Hh signaling with respect to Sxl, while it is critical for the activation of Ci. By contrast, both Su(fu) and the Fu regulatory domain act similarly on Sxl and Ci, serving to anchor them in the cytoplasm (Horabin, 2003).

    Taken together, these data suggest that the presence of Hh can be relayed to the cytoplasmic components differentially and, while the data do not address the point, suggest how different outcomes might be achieved. Ptc has been proposed to be a transmembrane transporter protein that functions catalytically in the inhibition of Smo via a diffusible small molecule. The stimulation of Sxl nuclear entry by the binding of Hh to Ptc might also involve a change in the internal cell milieu, but in this case the Hh cytoplasmic complex may be affected independently, not requiring a change in the activity of Smo or the Fu kinase (Horabin, 2003).

  3. Ptc can signal the presence of the Hh ligand in a positive manner.
    Several experiments indicate that Hh bound to Ptc enhances the nuclear entry of Sxl. That Smo has no role in transmitting the Hh signal is most clearly demonstrated by expressing the PtcD584 protein in both the anterior and posterior compartments of the dorsal half of the wing disc. PtcD584 acts as a dominant negative and so activates Ci in the anterior compartment, but it fails to enhance the levels of nuclear Sxl in the anterior because it sequesters Hh in the posterior compartment. The double mutant condition of ptc clones in a hhMRT background clearly places Ptc downstream of Hh, while showing Ptc can act positively in transmitting the Hh signal (Horabin, 2003).

A positive role for Ptc, but in this case in conjunction with Smo, in promoting cell proliferation during head development has recently been reported. In this situation, however, Hh acts negatively on both Ptc and Smo in their activation of the Activin type I receptor, suggesting an even greater variance from the canonical Hh signaling process (Horabin, 2003).

While the effects on Sxl in the anterior compartment show a dependence on the known Hh signaling components, it is not clear what promotes the rapid nuclear entry of Sxl in the posterior compartment. Su(fu) is expressed uniformly across the disc so it does not appear to be responsible for the AP differences, and ptc clones have no effect (and Ptc RNA and protein are not detected in the posterior compartment). Removal of Hh, however, reduces the nuclear entry rate of Sxl in both compartments. In this regard, the parallel between Hh pathway activation and Sxl nuclear entry in the posterior compartment is worth noting. Fu is also activated in the posterior compartment in a Hh-dependent manner, even though Ptc is not present. It is not clear what mediates between Hh and Fu (Horabin, 2003).

The data also suggest that the Hh cytoplasmic complex may have slightly different compositions in different tissues and/or at developmental stages. In the female germline and in embryos, the absence of Cos2 leads to a severe reduction in Sxl levels. However, in wing discs when mutant clones are made using the same cos2 allele, there is no effect on Sxl. It is suggested that between the third instar larval stage and eclosion, the composition of the Hh cytoplasmic complex may change again to make Sxl more sensitive to Cos2. This would explain why removal of Cos2 can produce sex transformations of the foreleg even though mutant clones in wing discs (and also leg discs) show no alterations in Sxl levels (Horabin, 2003).

A similar argument might apply to the weak sex transformations of forelegs produced by PKA clones. Alternatively, PKA may have a very weak effect but the assay on wing discs is not sufficiently sensitive to allow detection of small effects; PKA was found to have a modest effect on Sxl nuclear entry in the germline. Sxl is sufficiently small (38-40 kDa) to freely diffuse into the nucleus, or the protein may enter the nucleus as a complex with splicing components. This may account for the limited sex transformations caused by removal of Hh pathway components (Horabin, 2003).

Removal of several of the Hh pathway components, such as smo, gives the same weak sex transformation phenotype, even though smo has no effect on Sxl nuclear entry. Additionally, there is no correlation between a positive and a negative Hh signaling component and whether there is a resulting phenotype. Changing the dynamics of the activation state of the Hh cytoplasmic complex may perturb the normal functioning of Sxl, since Sxl appears to be in the same complex as Ci. For example, if the Hh pathway is fully activated because of a mutant condition, the relative amounts of Sxl in the cytoplasm versus nucleus at any given time, may be different from the wild-type condition. Perturbing the usual cytoplasmic-nuclear balance could compromise the various processes that Sxl protein regulates. Sxl acts both positively and negatively on its own expression through splicing and translation control and, additionally, regulates the downstream sex differentiation targets. The latter could also be responsible for the weak sex transformations seen, in view of the recent demonstration that doublesex affects the AP organizer and sex-specific growth in the genital disc (Horabin, 2003).

With the exception of Cos2, which can produce relatively substantial effects on Sxl levels in embryos as well as sex transformations in the foreleg, the effects of removal of any of the other Hh pathway components are generally not large. The strong effects of Cos2 on Sxl could be because it affects the stability of Sxl. However, Sxl depends on an autoregulatory splicing feedback loop for its maintenance making the protein susceptible to a variety of regulatory breakdowns. If Cos2 altered the nuclear entry of Sxl, for example, its removal could compromise the female-specific splicing of Sxl transcripts by reducing the amounts of nuclear Sxl. Splicing of Sxl transcripts would progressively fall into the male mode to eventually result in a loss of Sxl protein (Horabin, 2003).

Cos2 and Fu have been reported to shuttle into and out of the nucleus, and their rate of nuclear entry is not dependent on the Hh signal. That Ci and Sxl are complexed with the same Hh pathway cytoplasmic components, and share and yet have unique intracellular trafficking responses to mutations in the pathway, makes it tempting to speculate that the Hh cytoplasmic components may have had a functional origin related to intracellular trafficking that preceded the two proteins. Whether this reflects a more expanded role in regulated nuclear entry remains to be determined (Horabin, 2003).

Hedgehog-regulated Costal2-kinase complexes control phosphorylation and proteolytic processing of Cubitus interruptus

Hedgehog (Hh) proteins control animal development by regulating the Gli/Ci family of transcription factors. In Drosophila, Hh counteracts phosphorylation by PKA, GSK3, and CKI to prevent Cubitus interruptus (Ci) processing through unknown mechanisms. These kinases physically interact with the kinesin-like protein Costal2 (Cos2) to control Ci processing and Hh inhibits such interaction. Cos2 is required for Ci phosphorylation in vivo, and Cos2-immunocomplexes (Cos2IPs) phosphorylate Ci and contain PKA, GSK3, and CKI. By using a Kinesin-Cos2 chimeric protein that carries Cos2-interacting proteins to the microtubule plus end, it was demonstrated that these kinases bind Cos2 in intact cells. PKA, GSK3, and CKI directly bind the N- and C-terminal regions of Cos2, both of which are essential for Ci processing. Finally, it was shown that Hh signaling inhibits Cos2-kinase complex formation. It is proposed that Cos2 recruits multiple kinases to efficiently phosphorylate Ci and that Hh inhibits Ci phosphorylation by specifically interfering with kinase recruitment (Zhang, 2005).

To facilitate detection of protein-protein interaction between Cos2 and its binding partners in vivo, a Kinesin/Cos2 chimeric protein (Kinco) was generated in which the microtubule binding domain of Cos2 is replaced by the motor domain of Drosophila KHC. Kinco moves to the microtubule plus end and accumulates near the basal surface of imaginal disc epithelial cells. Strikingly, Kinco carries all the known Cos2 binding proteins to the same subcellular compartment, leading to colocalization. PKAc, GSK3, and two CKI isoforms, CKIα and CKIϵ, all colocalize with Kinco at the microtubule plus end, demonstrating that these kinases associate with Cos2 in intact cells. Hence, Kinco provides a powerful tool to determine if a protein interacts with Cos2 in vivo. In addition, Kinco colocalizes with Cos2-interacting proteins in cultured Drosophila cells such as S2 and cl8 cells. It is conceivable that one can use such a cell-based colocalization assay to identify additional proteins that form a complex with Cos2. Furthermore, it is also possible to extend this approach to other protein complexes by generating appropriate Kinesin chimeric 'bait' proteins (Zhang, 2005).

By using immunoprecipitation and GST pull-down assays, the kinase interaction domains were mapped to the microtubule binding (MB) and C-terminal (CT) of Cos2. GST fusion proteins containing either of these domains bind purified recombinant PKAc, GSK3, and CKI, suggesting that these kinases directly bind Cos2. However, the possibility cannot be rule out that these kinases may have additional contacts with other components in the Cos2 complex. Indeed, it was found that CKI can bind Ci in yeast (Zhang, 2005).

Several lines of evidence suggest that Cos2/kinase interaction plays an important role in regulating Ci phosphorylation and processing: (1) Ci phosphorylation is compromised in cos2 mutants; (2) the kinase-interacting domains in Cos2 are essential for Ci processing; (3) overexpressing multiple kinases can bypass the requirement of Cos2 for Ci processing (Zhang, 2005).

PKAc, GSK3, and CKI appear to bind competitively to Cos2; however, since Cos2 can dimerize and each Cos2 protein contains two kinase binding domains, a Cos2 dimer could in principle bind all three kinases simultaneously. It is possible that these kinases might not form a tight complex with Cos2 in a stoichiometric manner, which could explain why purification of endogenous Cos2 complexes failed to identify any of these kinases. However, by using in vitro kinase assay and Western blot analysis, the association of PKAc, GSK3, and CKI with endogenous Cos2 was detected. It is likely that interactions between Cos2 and kinases are transient; however, such interactions could increase local concentrations of these kinases; this greatly facilitates Ci hyperphosphorylation (Zhang, 2005).

It has been demonstrated that Hh induces Ci dephosphorylation in cl8 cells; however, it is not clear whether Hh blocks Ci phosphorylation by all three kinases or a subset of them. By using an antibody that specifically recognizes a phosphorylated PKA site in Ci, it was found that Hh partially inhibits PKA phosphorylation of Ci in wing discs. Consistent with this, Hh only partially blocks Cos2/PKA interaction. In contrast, Hh appears to have a more profound influence on the interaction between Cos2 and CKI or GSK3. Furthermore, CKI and GSK3 kinase activities associated with endogenous Cos2 diminishes upon Hh stimulation and Cos2IPs phosphorylates Ci to a lesser extent after Hh treatment. These observations suggest that Ci phosphorylation by CKI and GSK3 is likely to be inhibited by Hh in vivo (Zhang, 2005).

Several mechanisms may contribute to the regulation of Cos2-Ci-kinase complex formation by Hh. (1) The finding that PKAc, GSK3, and CKI bind Cos2 domains that also interact with Smo raises a possibility that Smo/Cos2 interaction may exclude kinases from binding to Cos2. Indeed, a membrane-tethered form of SmoCT (Myr-SmoCT) interferes with Cos2-Ci-kinase complex formation. (2) Smo/Cos2 interaction at the cell surface may induce conformational change in Cos2, which could mask its kinase interacting domains. (3) Cos2 is phosphorylated in response to Hh. Phosphorylation of Cos2 could regulate its interaction with one or more kinases. (4) There is evidence that Hh induces dissociation of Ci from Cos2, which may further decrease the accessibility of Ci to the kinases. This may explain why Hh induces more significant dissociation of PKAc from Ci than from Cos2. (5) Hh induces degradation of Cos2 in P compartment cells as well as in cells immediately adjacent to the A/P boundary; this may lead to a chronic destruction of Cos2-Ci-kinase complexes. However, it appears that only high levels of Hh induce Cos2 degradation in vivo. Low levels of Hh may prevent Ci phosphorylation with different mechanisms such as those described above (Zhang, 2005).

The following model is proposed for the regulation of Ci phosphorylation by Cos2 and Hh. In the absence of Hh, Cos2 scaffolds multiple kinases and Ci into proximity, thus increasing the accessibility of Ci to these kinases and facilitating extensive phosphorylation of Ci. Upon Hh stimulation, Cos2 complexes are recruited to the cell surface via Smo, leading to disassembly of Cos2-Ci-kinase complexes. As a consequence, Ci phosphorylation is compromised and Ci processing is blocked. This model has several interesting parallels to that proposed for the Wnt pathway. In quiescent cells, both pathways employ large protein complexes to bring kinases and their substrates in close proximity, resulting in phosphorylation and proteolysis of the transcription factor (Ci) or effector (β-catenin). Upon ligand stimulation, both pathways recruit the cytoplasmic signaling complex to the cell surface and cause dissociation of the complex, leading to dephosphorylation and stabilization of the transcription factor/effector. Interestingly, both pathways use common kinases, including GSK3 and CKI. However, these kinases together with their substrates form distinct signaling complexes assembled by pathway-specific scaffolding proteins (Cos2 and Axin in the Hh and Wnt pathways, respectively). Pathway activation is achieved by a specific interaction between the receptor system and the scaffolding protein (Smo/Cos2 interaction in the Hh pathway and LPR5/6/Axin interaction in the Wnt pathway). Thus, each pathway only controls the pool of kinases in the same complex with the pathway effector, leading to pathway-specific regulation of substrate phosphorylation. The combinatorial mechanism by which pathway-specific scaffolds bring common kinases into proximity with their substrates thus appears to be a general one and may apply to other signaling pathways that utilize a common set of kinases (Zhang, 2005).

Divergence of hedgehog signal transduction mechanism between Drosophila and mammals

The Hedgehog (Hh) signaling pathway has conserved roles in development of species ranging from Drosophila to humans. Responses to Hh are mediated by the transcription factor Cubitus interruptus (Ci; GLIs 1-3 in mammals), and constitutive activation of Hh target gene expression has been linked to several types of human cancer. In Drosophila, the kinesin-like protein Costal2 (Cos2), which associates directly with the Hh receptor component Smoothened (Smo), is essential for suppression of the transcriptional activity of Ci in the absence of ligand. Another protein, Suppressor of Fused [Su(Fu)], exerts a weak negative influence on Ci activity. Based on analysis of functional and sequence conservation of Cos2 orthologs, Su(Fu), Smo, and Ci/GLI proteins, Drosophila and mammalian Hh signaling mechanisms have been found to diverge; in mouse cells, major Cos2-like activities are absent and the inhibition of the Hh pathway in the absence of ligand critically depends on Su(Fu) (Varjosalo, 2006).

The evidence indicates that a significant divergence in the mechanism of Shh signal transduction has occurred between vertebrates and invertebrates at the level of Smo, Cos2, and Su(Fu). The results indicate that major Cos2-like activities are absent in mouse cells based on four observations: (1) domains in Smo that are required in Drosophila to bind to Cos2 are not required for mSmo function; (2) mouse Shh signaling is insensitive to expression of Drosophila Cos2, but can be rendered Cos2 sensitive by replacing the mSmo C-terminal domain with the dSmo C-terminal domain; (3) expression of the Smo C-terminal domain which, in Drosophila, inactivates Cos2 has no effect in the mouse in vivo or in vitro; (4) overexpression or RNAi-mediated suppression of mouse Cos2 homologs has no effect on Hh signaling, even under sensitized conditions. These results are also consistent with divergence of the sequence of domains involved in Cos2 binding in Ci/GLI proteins and Smo between insects and mammals (Varjosalo, 2006).

Although the RNAi experiments targeting Cos2 orthologs Kif7 and Kif27 were performed under conditions in which negative regulators of GLI2 were limiting, they could be argued to be consistent with a model in which multiple kinesins with Cos2-like activity would act in a redundant fashion in mammals. By loss-of-function studies of individual kinesins in cell culture or in mice it would be difficult to obtain conclusive evidence against such a model due to the potential redundancy of multiple members of the kinesin family. However, several other in vitro and in vivo experiments that were presented directly contradict such a model. These include RNAi analyses targeting multiple Kif proteins, the analysis of loss of function of mSmo domains, and the lack of effect of overexpression of myristoylated-mSmoC and the Cos2 orthologs Kif7 and Kif27. In addition, no kinesin with Cos2-like activity could be found by extending the analyses to several other kinesins, which show homology to Cos2 but have different fly orthologs (Varjosalo, 2006).

In contrast to the case in Drosophila, Su(Fu) has a critical role in suppression of the mammalian Hh pathway in the absence of ligand, and loss of Su(Fu) function results in dramatic induction of GLI transcriptional activity. The results are also consistent with the studies that show that loss of Su(Fu) in mouse embryos results in complete activation of the Hh pathway, in a fashion similar to the loss of Ptc. These results are particularly surprising in light of the central role of Cos2 and a minor role of Su(Fu) in Drosophila. Together, these results also clearly show that mouse cells and embryos lack a Cos2-like activity that, in Drosophila, can completely suppress the Hh pathway in the absence of Su(Fu). However, the results should not be taken as evidence against novel proteins (including kinesins not orthologous to Cos2) acting in mammalian cells between Smo and GLI proteins with mechanisms that are distinct from those used by Drosophila Cos2. Several reports have, in fact, described such vertebrate-specific regulators of Hh signaling, including SIL, Iguana, Rab23, Kif3a, IFT88, IFT172, MIM/BEG4, and β-arrestin2 (Varjosalo, 2006).

The results also shed light on some known differences in the function of the Hh pathway in Drosophila and mammals. Mutations and small molecules affecting conformation of Smo transmembrane domains have a strong effect in mammals, but they have little effect in Drosophila. Interestingly, the Smo transmembrane domain is required for regulation of Su(Fu) activity, whereas the Smo C-terminal domain is critical for inhibition of Cos2 activity. Thus, based on the data, manipulations that affect the Smo transmembrane domain would be predicted to affect Su(Fu) and therefore to have a limited role in Drosophila and a major effect in mammals (Varjosalo, 2006).

Although there are differences in mouse and Drosophila Smo functional domains, and a lack of conservation of Smo phosphorylation sites, conservation of Smo function at a level not involving Cos2 is supported by the observation that mutation of a conserved isoleucine (I573A in mSmo) results in loss of both mouse and Drosophila Smo activity, yet does not result in a loss of Cos2 binding to dSmo. In addition, dSmo proteins that are activated by phosphomimetic mutations are constitutively stabilized; yet, they are partially responsive to Hh, suggesting that, in addition to stabilization and phosphorylation, other, potentially conserved mechanisms could be required to generate fully active Smo in Drosophila as well (Varjosalo, 2006).

In the mSmo C terminus, six residues between amino acids 570 and 580 were identified that resulted in significant loss of mSmo activity. The predicted secondary structure for this region is an α helix, in which these residues would reside on the same side, raising the possibility that, together with the third Smo intracellular loop, this region may form an interaction surface involved in inactivation of Su(Fu) or activation of Ci/GLI (Varjosalo, 2006).

Recent results have indicated that Su(Fu) acts as a tumor suppressor in medulloblastoma, and it has been suggested that medulloblastomas associated with loss of Su(Fu) result, in part, from activation of the Wnt pathway. However, consistent with the lack of a Wnt phenotype of Su(Fu) mutations in Drosophila, in the current experiments, a Wnt pathway-specific reporter is not activated by shRNAs targeting Su(Fu). Given observations that Su(Fu) is critically important in the suppression of the mammalian Hh pathway in the absence of ligand, and the fact that Hh pathway activation is required for growth of a form of medulloblastoma induced by mutations in Patched, it is likely that constitutive activation of the Hh pathway is also essential for growth of medulloblastomas associated with the loss of Su(Fu) (Varjosalo, 2006).

In a wider context, the results demonstrate that signal transduction mechanisms of even the major signaling pathways are not immutable, but that they can be subject to evolutionary change. The divergence may have occurred after the separation of the vertebrate and invertebrate lineages. However, some evidence also suggests that functional divergence may have occurred much later in evolution. Although mutants of Fused or Cos2 orthologs of zebrafish have not been identified, zebrafish homologs of Fused and Cos2 act in the Hh pathway based on morpholino antisense injections. In contrast to these findings, mice deficient in mouse ortholog of Drosophila Fused do not have a Hh-related phenotype, and mouse orthologs of Cos2 do not affect Hh signaling. Hh-related phenotypes can be observed in zebrafish by morpholino-mediated targeting of other genes as well, such as β-arrestin2, whose loss in mice does not result in a Hh-related phenotype. It is widely appreciated that multiple types of embryonic insults result in Hh-like phenotypes, such as holoprosencephaly. Thus, it is possible that the zebrafish phenotypes observed are caused by the morpholino injection process itself. Alternatively, there may also be significant differences between the mechanism of Hh signaling between vertebrate species (Varjosalo, 2006).

Because of the strong conservation of Su(Fu) in both invertebrate and vertebrate phyla, the presence of a Cos2 binding domain only in insect Smo, and the divergence of the Cos2 proteins from the kinesin family, the simplest explanation of the data is that Su(Fu) represents the primordial Ci/GLI repressor, and that the Cos2-like functionality has evolved specifically in the invertebrate lineage. The results, thus, also raise the possibility that multicomponent pathways evolve, in part, by insertion of novel proteins between existing pathway components. This mechanism potentially explains a challenging aspect of evolutionary biology regarding the emergence of signaling pathways with multiple specific components (Varjosalo, 2006).

Smoothened interacts with Cos2 to regulate activator and repressor functions of Hedgehog signaling via two distinct mechanisms

The secreted protein Hedgehog (Hh) plays an important role in metazoan development and as a survival factor for many human tumors. In both cases, Hh signaling proceeds through the activation of the seven-transmembrane protein Smoothened (Smo), which is thought to convert the Gli family of transcription factors from transcriptional repressors to transcriptional activators. This study provides evidence that Smo signals to the Hh signaling complex, which consists of the kinesin-related protein Costal2 (Cos2), the protein kinase Fused (Fu), and the Drosophila Gli homolog cubitus interruptus (Ci), in two distinct manners. Many of the commonly observed molecular events following Hh signaling are not transmitted in a linear fashion but instead are activated through two signals that bifurcate at Smo to independently affect activator and repressor pools of Ci (Ogden, 2006).

This work demonstrates that targeting the association between Smo and the Cos2 cargo domain functionally separates the known molecular markers of the Hh pathway into two distinct categories: those events dependent on a direct association between the Cos2 cargo domain and Smo and those not dependent on this direct association. The Hh-induced readouts requiring direct Smo-Cos2 association include Smo phosphorylation, stabilization, and translocation to the plasma membrane, which facilitate intermediate to high level activation of Ci. Hh-induced Fu and Cos2 hyperphosphorylation, Hedgehog signaling complex relocalization from vesicular membranes to the cytoplasm, and Ci stabilization do not appear to require a direct Smo-Cos2 cargo domain association. Thus, although Smo is necessary for all aspects of Hh signaling, only the molecular events grouped with Ci activation appear to require direct association between Cos2 and Smo. In vivo, carboxyl-terminal Smo binding domain expression is also capable of attenuating Hh signaling. This observation is consistent with in vitro observation that carboxyl-terminal Smo binding domain inhibits critical requirement(s) for pathway activation (Ogden, 2006).

A model has been proposed suggesting the existence of two independently regulated pools of the Hedghog signaling complex (HSC), one involved in pathway repression (HSC-R), and one involved in activation (HSC-A). HSC-R is dedicated to priming Ci for processing into the Ci75 transcriptional repressor, whereas HSC-A is dedicated to activation of stabilized Ci155 in response to Hh. this study provides evidence that the effects of these two HSCs can be functionally separated by specifically targeting the interaction between Smo and the Cos2 cargo domain. Moreover, distinct molecular markers were identified for each HSC. It is proposed that in HSC-R, the membrane vesicle tethered Cos2 functions as a scaffold to recruit protein kinase A, glycogen synthase kinase 3ß, and casein kinase I, which in turn phosphorylate Ci. Hyperphosphorylated Ci is then targeted to the proteasome by the F-box protein supernumerary limbs (Slimb), where it is converted into Ci75. In response to Hh, Fu and Cos2 are phosphorylated and dissociate from vesicular membranes and microtubules, which is suggested to result in the attenuation of HSC-R function. This allows for the subsequent accumulation of full-length Ci. The mechanism by which HSC-R function is inhibited by Hh-activated Smo is not clear but appears to require the carboxyl-terminal tail of Smo and, by this analysis, appears to occur independently of a direct Smo-Cos2 cargo domain association. However, the direct Cos2-Smo association is critical for regulation of HSC-A. In the absence of Hh, HSC-A is tethered to vesicular membranes, through Smo, where it is kept in an inactive state. In the presence of Hh, Cos2 bound directly to Smo acts as a scaffold for the phosphorylation of Smo by protein kinase A, glycogen synthase kinase 3ß, and casein kinase I. Phosphorylation of Smo triggers its stabilization and relocalization to the plasma membrane with HSC-A, where Ci is proposed to be activated. Thus, Cos2 plays a similar role in both HSC-R and HSC-A. In the former case, coupling protein kinase A, glycogen synthase kinase 3ß, and casein kinase I with Ci and, in the latter case, coupling the same protein kinases with the carboxyl-terminal tail of Smo (Ogden, 2006).

An alternative interpretation of these data is that disruption of the Cos2 cargo domain-Smo association separates high and low level Hh signaling. It has been suggested that a second, low affinity Smo binding domain may reside within the coiled-coil domain of Cos2. Thus, high level signaling, where all aspects of the Hh pathway are activated may require both Cos2 interaction domains to be directly bound to Smo. In either scenario, HSC-R function would be regulated independently of HSC-A function (Ogden, 2006).

It is concluded that targeted disruption of Cos2 cargo domain-Smo binding by CSBD is able to functionally separate the activities ascribed to the two HSC model. This two-switch system is amenable to the formation of a gradient of Hh signaling activity across a field of cells, in that the relative activity of HSC-R to HSC-A is directly proportional to the level of Hh stimulation a cell receives. The opposing functional effects of the two complexes can then establish unique ratios of Ci75 to activated Ci, resulting in distinct levels of pathway activation on a per cell basis (Ogden, 2006).

Phosphorylation of the atypical kinesin Costal2 by the kinase Fused induces the partial disassembly of the Smoothened-Fused-Costal2-Cubitus interruptus complex in Hedgehog signalling

The Hedgehog (Hh) family of secreted proteins is involved both in developmental and tumorigenic processes. Although many members of this important pathway are known, the mechanism of Hh signal transduction is still poorly understood. In this study, the regulation of the kinesin-like protein Costal2 (Cos2) by Hh was analyzed. A residue on Cos2, serine 572 (Ser572), is necessary for normal transduction of the Hh signal from the transmembrane protein Smoothened (Smo) to the transcriptional mediator Cubitus interruptus (Ci). This residue is located in the serine/threonine kinase Fused (Fu)-binding domain and is phosphorylated as a consequence of Fu activation. Although Ser572 does not overlap with known Smo- or Ci-binding domains, the expression of a Cos2 variant mimicking constitutive phosphorylation and the use of a specific antibody to phosphorylated Ser572 showed a reduction in the association of phosphorylated Cos2 with Smo and Ci, both in vitro and in vivo. Moreover, Cos2 proteins with an Ala or Asp substitution of Ser572 were impaired in their regulation of Ci activity. It is proposed that, after activation of Smo, the Fu kinase induces a conformational change in Cos2 that allows the disassembly of the Smo-Fu-Cos2-Ci complex and consequent activation of Hh target genes. This study provides new insight into the mechanistic regulation of the protein complex that mediates Hh signalling and a unique antibody tool for directly monitoring Hh receptor activity in all activated cells (Reul, 2007).

These data show that phosphorylation of Cos2 residue Ser572 is necessary for the full activation of Hh signalling, and that this phosphorylation is dependent on the kinase Fu. It is likely that Fu directly phosphorylates Cos2 on Ser572, but it was not possible to purify an activated Fu kinase to confirm this. The phosphorylation of this residue strongly decreased the association of Cos2 with both Ci and Smo, an important step in the regulation of the cytoplasmic anchoring of Ci. By contrast, Cos2-572A, a Cos2 mutant that cannot be phosphorylated at Ser 572, remained associated with Smo and Ci but was much less sensitive to Hh regulation; this is because both its restraining activity on Ci and its association with Ci were only minimally sensitive to the presence of Fu and to the activation of Hh signalling. Phosphorylation of Ser572 of Cos2 induces the partial disassembly of the protein complex (Reul, 2007).

The data show that Cos2 phosphorylated on Ser572 does not bind Smo. However, previous studies have shown that Cos2 is phosphorylated and is pulled down by Smo in response to Hh stimulation. How can these data be reconciled? First, it is possible that not all Cos2 proteins that bind to Smo are phosphorylated. Indeed, only a limited fraction of Cos2 and Fu are sensitive to Hh activation. This is clearly observed with Fu (only 50% of the protein undergoes an electromobility shift upon Hh activation), but is more difficult to quantify with Cos2 because of its very small and diffused electromobility shift. Nevertheless, if Cos2 behaves similarly to Fu, it would mean that 50% of the total Cos2 (corresponding to the non-modified protein in Hh-treated cells) should be able to bring enough Smo down to be detectable in immunoprecipitates. Second, it is possible that Smo still binds to phosphorylated Cos2 on Ser572, but with much less affinity. Third, phosphorylation on Ser572 is not responsible for all Cos2 mobility shift, because Cos2-572A still shifts upon OA treatment, suggesting that other phosphorylated sites are present. Therefore, some phosphorylated isoforms that are not phosphorylated on Ser572 might also be associated with Smo. It is thus possible that this study has revealed only one of a series of sequential phosphorylation events on Cos2 that ultimately lead to the complete dissociation of Cos2 from Smo. Finally, it is worth mentioning that more Smo is present in the Cos2 IP from Hh-treated cells than in non-Hh treated cells. This is thought to simply reflect an increased level of Smo resulting from Hh signalling activation, and not the Hh-dependent regulation of the efficiency of the interaction of Smo with Cos2 (Reul, 2007).

The role of the Cos2 protein in the complex is to serve as a platform to allow both positive and negative regulators to be brought into close proximity with Smo and Ci. Thus, the role of Cos2 in transmitting a response can be masked by the role of Cos2 in limiting pathway activity in the absence of Hh. At low concentrations, it is able to stimulate Hh reporter activity in vitro and engrailed expression in vivo. But in Cos2-572A-expressing cells, engrailed expression was lower than in wild-type discs, and the in vitro stimulation of Hh signalling could not be potentiated by Fu activity. Moreover, the restraining activity of Cos2-572A on Ci could not be counteracted by Hh or Fu in vitro. Therefore, it is proposed that the Ser572 to Ala substitution on Cos2 rendered Cos2 less sensitive to Hh and Fu regulation. Because Cos2-572A still binds to its partners, it could bring Fu into proximity with its other targets. Indeed, it is likely that Fu activation leads not only to the direct phosphorylation of Cos2 but also to direct changes in Ci and/or other partners, such as Sufu. This explains why Cos2-572A is still able to stimulate Hh signalling, albeit not to its highest level (Reul, 2007).

From the Cos2-572A results, one could wonder why Cos2-572D did not constitutively activate the pathway. Because the Cos2-572D form is in a 'frozen' state compared with the wild-type form, cycles of phosphorylation/dephosphorylation are blocked and thus Cos-572D cannot participate in the Hh complex signalling anymore. The data show that constitutively phosphorylated Cos2 and endogenous phospho-Cos2 are bound to Fu but are dissociated from Smo and Ci. Therefore, Fu bound to phosphorylated Cos2 would be absent from the complex, preventing the release of all the cytoplasmic anchors from Ci (Reul, 2007).

Because the Cos2 Ser572 residue is not part of the Ci- or Smo-binding domains, but phosphorylation of this site nevertheless leads to the dissociation of these two proteins from Cos2, it is proposed that the Fu-mediated modification of Cos2 induces the protein to undergo a conformational change that leads to the disassembly of the complex. The disassembly is partial because phosphorylated Cos2 and Fu are still associated. Interestingly, it has been proposed that the binding of Cos2, Sufu and Fu to Ci masks a nuclear localisation site on Ci (Ci-NLS). A conformational change that supports this idea: that disassembly of the complex is necessary to expose the Ci-NLS and for consequent nuclear translocation (Reul, 2007).

Costal 2 interactions with Cubitus interruptus underlying Hedgehog-regulated Ci processing

Extracellular Hedgehog (Hh) proteins alter cellular behaviours from flies to man by regulating the activities of Gli/Ci family transcription factors. A major component of this response in Drosophila is the inhibition of proteolytic processing of the latent transcriptional activator Ci-155 to a shorter Ci-75 repressor form. Processing is thought to rely on binding of the kinesin-family protein Cos2 directly to Ci-155 domains known as CDN and CORD, allowing Cos2-associated protein kinases to phosphorylate Ci-155 efficiently and create a binding site for an E3 ubiquitin ligase complex. This study shows that the last three zinc fingers of Ci-155 also bind Cos2 in vitro and that the zinc finger region, rather than the CDN domain, functions redundantly with the CORD domain to promote Hh-regulated Ci-155 proteolysis in wing discs. Evidence was also found for a unique function of Cos2 binding to CORD. Cos2 binding to CORD, but not to other regions of Ci, is potentiated by nucleotides and abrogated by the nucleotide binding variant Cos2 S182N. Removal of the CORD region alone enhances processing under a variety of conditions. Most strikingly, CORD region deletion allows Cos2 S182N to stimulate efficient Ci processing. It is deduced that the CORD region has a second function distinct from Cos2 binding that inhibits Ci processing, and that Cos2 binding to CORD relieves this inhibition. It is suggested that this regulatory activity of Cos2 depends on a specific nucleotide-bound conformation that may be regulated by Hh (Zhou, 2010).

Prior to this study it was thought that Cos2 regulates Ci by binding to specific protein kinases and directly to Ci-155 via two regions, CDN and CORD to promote Ci-155 phosphorylation. Experiments with tissue culture cells suggested that CDN and CORD regions of Ci-155 act largely redundantly to promote Ci-155 processing. The current investigations have modified these views in two significant ways. First, studies in the physiological setting of Drosophila wing discs confirm some functional redundancy of two Cos2-binding regions on Ci to promote Ci proteolysis, but the region acting together with the CORD binding site comprises the last three zinc fingers of Ci, not the CDN region. Second, it was found that the CORD region has an additional unique function of inhibiting Ci proteolysis unless it binds to Cos2. Furthermore, the potential was uncovered for Cos2-CORD association to be regulated by nucleotide binding to Cos2 and evidence was uncovered that Hh signalling may modulate Cos2 function in at least two ways to regulate its interaction with Ci (Zhou, 2010).

Three new observations lead to the deducition that Cos2 binding to CORD has an important non-redundant role in promoting Ci processing in the absence of Hh. First, it was found that removing the entire CORD region enhanced Ci proteolysis in a number of settings, revealing an inhibitory role for CORD. Similar criteria suggested an inhibitory role also for the CDN region of Ci. Second, it was found that Cos2 S182N fails to bind the CORD region of Ci but binds normally to a Ci region including the zinc fingers and CDN in vitro. Third, it was found that Cos2 S182N promotes efficient proteolysis of Ci only when the CORD region is absent. The restoration of proteolysis was specific to the S182N substitution and deletion of the CORD domain. Loss of CORD did not allow proteolysis by Cos2 S572D and loss of CDN did not allow proteolysis by Cos2 S182N. It is concluded that the strong defect of Cos2 S182N in supporting wild-type Ci processing results principally from an inability to bind to CORD and thereby relieve the inhibitory effect of CORD on Ci-155 processing. The importance of Cos2-CORD binding was not apparent by simple deletion of the CORD region because that deletion simultaneously eliminates Cos2 binding and the need for Cos2 binding, while sparing the zinc fingers of Ci as an alternative means to recruit Cos2. While Cos2 S182N mediated the proteolysis of Ci molecules lacking the CORD region remarkably efficiently, wild-type Cos2 was consistently better, implying that Cos2 S182N does have a deficit beyond CORD binding that is relevant to Ci proteolysis. That, relatively minor, deficit may stem from the failure of Cos2 S182N to move normally along microtubules (Farzan, 2008; Zhou, 2010).

What is the nature of the inhibitory influence of the CORD region on Ci proteolysis? A variety of segments of Ci, including the zinc fingers, CORD and phosphorylation regions, have been found to bind to each other in vitro. It is therefore speculated that the CORD region may interact, intra- or inter-molecularly, with other regions of Ci, to limit exposure of either key phosphorylation sites to protein kinases, or of the zinc finger and CORD regions to Cos2. The CDN region of Ci also appears to contribute to interactions that make Ci less accessible to one or more steps directing its proteolysis. Relief of CDN inhibition does not, however, appear to depend on Cos2 binding to CORD (because Ci?CORD is efficiently processed by Cos2 S182N) and is apparent in the presence or absence of either CORD or zinc finger Cos2-binding domains (Zhou, 2010).

In addition to a unique function of Cos2 binding to CORD, this association also has a function that can alternatively be executed by the zinc finger region. This assertion is deduced simply from the defective proteolysis of CiδZnδCORD compared to the efficient proteolysis of both CiδZn and CiδCORD (whether assayed in the presence or absence of CDN). Most likely this function is the recruitment of Cos2-associated protein kinases to Ci (Zhou, 2010).

What properties are conferred by the two partially overlapping functions of Cos2-Ci binding and the two Ci domains capable of recruiting Cos2? An obvious hypothesis is that this diversifies the means by which Hh can regulate Ci processing through Cos2, perhaps to extend the range of Hh sensitivity or to produce a more robust Hh response. Specific mechanisms are considered in the next section but the general hypothesis can be investigated by simply eliminating specific modes of Cos2-Ci interaction. It has not been possible to probe the consequences of eliminating Ci zinc fingers in detail because loss of DNA binding prevents execution of normal Ci functions. However, the regulation of CiδCORD and CiδCDN?CORD appeared to be remarkably normal. High levels of Hh in posterior cells fully inhibited Ci processing and elevated full-length Ci protein levels extended over roughly the normal range at the AP border, suggesting that sensitivity to significant inhibition by low Hh levels is also retained. The sensitivity of proteolysis of Ci lacking zinc fingers to low Hh levels also appeared to be roughly normal. Therefore the idea is favoured that the multiple Cos2-Ci interactions are each subject to Hh regulation over a similar range of sensitivity, and that the mechanisms for regulating Cos2-Ci interactions by Hh, like the interactions themselves, are largely redundant, resulting in a very robust regulatory response that is resistant to single genetic perturbations (Zhou, 2010).

Evidence has previously been presented that Hh causes some degree of dissociation between Cos2 and the protein kinases, PKA, CK1 and GSK3, as well as reduced association between Cos2 and Ci. These mechanisms are, of course, not exclusive and their quantitative contributions remain unresolved because definitive physiological measurements of association are very difficult. More importantly, the upstream instigators of these proposed dissociations are not at all clear. The current studies suggest that a specific nucleotide-dependent conformation of Cos2 may be one important mediator of Hh signalling (Zhou, 2010).

It was found that nucleotides stimulated binding of Cos2 derived from cell extracts to GST-Ci CORD, presumably by increasing the proportion of Cos2 molecules that are nucleotide bound. Conventional kinesins are not readily isolated in a nucleotide-free state and their properties are generally altered by exchanging one bound nucleotide for another. It is therefore surprising that it was possible to alter Cos2 properties by adding excess nucleotide rather than by altering the nature of the excess nucleotide. Cos2 has been noted to differ from conventional kinesins in a number of conserved residues but retains residues S182 and G175 within the conserved P-loop that interacts with the β-phosphate of bound nucleotides. There are no reliable means to predict whether Cos2 S182N or G175A would be defective for binding specific nucleotides, all nucleotides or nucleotide hydrolysis, and those properties have not been measured directly for Cos2 or Cos2 variants. Nevertheless, the observation that both Cos2 S182N and G175A showed no evidence of binding CORD suggests two complementary assertions. First, Cos2 S182N and G175A are unable to adopt a nucleotide-bound conformation that is stringently required for binding CORD. Second, the binding of wild-type Cos2 in the absence of added nucleotide is most likely due to a minor proportion of Cos2 molecules bound to nucleotides rather than due to a lower affinity interaction of a nucleotide-free conformation of Cos2. Thus, it is hypothesised that distinct Cos2 conformations, influenced by nucleotide binding, constitute a clean on/off switch for binding the CORD region of Ci (Zhou, 2010).

Conformational changes couple nucleotide binding and microtubule association in kinesins. Hence, the previously observed Hh-induced dissociation of Cos2 from microtubules supports the hypothesis that Hh induces a conformational change in Cos2 that alters nucleotide binding, CORD association and microtubule binding. The speculated Hh-induced conformational change is most likely brought about by the known direct association of Cos2 with Smo. Smo is related to G-protein coupled receptors (GPCRs), suggesting that the actions of Smo on Cos2 could conceivably be analogous to the nucleotide exchange factor activity of GPCRs, which is normally directed to regulating G-protein conformation and activity (Zhou, 2010).

While modulation of Cos2-CORD interaction through an altered Cos2 conformation phenocopied by Cos2 S182N provides a potential mechanism for Hh to influence the efficiency of Ci proteolysis, it cannot be the sole mechanism because CiδCORD (and CiδCDN?CORD) proteolysis is still extensively regulated by Hh. The Cos2 S572D variant was created previously to mimic Hh-stimulated phosphorylation of Cos2 by Fused and was shown to have reduced ability to co-localise with Ci in embryos and co-precipitate Ci from tissue culture cells. This study found no significant defect in binding of Cos2 S572D to CORD or zinc finger regions of Ci in vitro, suggesting that the biochemical deficits of Cos2 S572D and Cos2 S182N are distinct. This was confirmed by in vivo studies showing that, unlike Cos2 S182N, Cos2 S572D did not preferentially promote proteolysis of Ci lacking the CORD region. Thus, current evidence indicates that Hh-stimulated Cos2 phosphorylation by Fu may provide a second, potentially redundant, mechanism for Hh to regulate Cos2-Ci interactions and the consequent processing of Ci-155. Whether the Hh-regulated association of PKA, CK1 and GSK3 with Cos2 is mediated by either Cos2 phosphorylation or nucleotide-dependent Cos2 conformational changes, or by a third, distinct mechanism, remains to be investigated (Zhou, 2010).

Phosphorylation of Ci at specific sites in the CORD domain (PKA site S962) and the Slimb-binding region (especially GSK3 sites primed by PKA site S892) reduced binding of the CORD region to Cos2 in vitro. That investigation was prompted by prior knowledge that loss of PKA sites in the CORD region appeared to enhance Ci activity. However, that observation would more readily be explained by increased, rather than decreased, Cos2-Ci binding in response to Ci phosphorylation. An alternative hypothesis is that the observed dependence of Cos2-CORD binding on Ci phosphorylation might contribute to extending a graded Hh response. Where Hh levels are high, Ci-155 will be less phosphorylated and would bind Cos2 more readily, requiring a strong Hh signal to disrupt Cos2-CORD association. At the edge of Hh signalling territory Ci-155 will be more highly phosphorylated, would bind Cos2 less readily and hence allow only a very low level of Hh to disrupt Cos2-CORD association and inhibit Ci-155 processing. Currently there has not been any in vivo evidence testing that hypothesis (Zhou, 2010).

While large regions of Ci (CDN and CORD) could be deleted without impairing proteolysis, implying that Ci is composed largely of independently folding domains, two other deletions (of residues 6-339 and 1286-1377) were identified with significant effects on proteolysis. Ci lacking C-terminal residues did not generate any detectable Ci-75 repressor. CiδC strongly induced ptc-lacZ, implying that binding to CBP, which has been mapped to an adjacent region of Ci and is required for Ci-155 activity, was not affected. How the C-terminus of Ci contributes to proteolysis remains a mystery since there is no evidence from binding assays or co-localization studies in tissue culture cells showing association with Cos2 or Cos2-associated factors (Zhou, 2010).

A study using Kc tissue culture cells previously identified the extreme C-terminus of Ci as essential for Ci processing (Wang, 2008). That study also found that the zinc fingers of Ci were not essential for processing, provided they were substituted by a stably folded domain that contributes to the arrest of proteasome digestion. That result is consistent with the observation that CiδZn is efficiently proteolyzed in wing discs. However, in contrast to the observation of very efficient processing of CiδCDNδCORD in wing discs, it was reported that one of these two domains must be present for efficient Ci processing in Kc cells. In fact, the key Ci substrate assayed also lacked residues 1-345 and is therefore virtually identical to the CiδNδCDNδCORD variant (rather than CiδCDNδCORD), which is also processed with reduced efficiency in wing discs (Zhou, 2010).

Removing the N-terminal region (residues 6-339) from Ci strongly raised anterior full-length Ci levels but did not appear to block proteolysis completely because loss of PKA was found to increase the activity of CiδN, presumably by completely eliminating proteolysis. Su(fu) is known to bind within the first 346 residues of Ci and has the potential to recruit Fu-Cos2 complexes to Ci indirectly. In Drosophila, loss of Su(fu) results in strongly reduced Ci-155 and Ci-75 levels but still permits Hh or loss of PKA to increase Ci-155 levels and Hh to inhibit Ci-75 repressor formation. Thus, Su(fu) is certainly not essential for Ci processing or its regulation. The substantial effects of Su(fu) on Ci-155 and Ci-75 levels are thought to involve a different proteolytic mechanism but it remains possible that Su(fu) might also modulate Ci processing efficiency. It is therefore similarly possible that the impaired proteolysis of CiδN results from a failure of Su(fu) to facilitate recruitment of Cos2-Fu complexes to Ci (Zhou, 2010).

In summary, this study has found that Ci-155 has at least two domains (CORD and zinc fingers) functionally capable of recruiting Cos2 directly, that Cos2 binding to the CORD domain additionally prevents that region from inhibiting proteolysis, and that the Cos2-CORD interaction might be regulated physiologically via a specific nucleotide-bound conformation of Cos2. Evidence was also found indicating that Cos2 might additionally be recruited indirectly to Ci, that Hh regulates productive Cos2-Ci engagement through multiple, potentially redundant, mechanisms, and that two terminal Ci-155 domains contribute to processing through mechanisms that are not yet understood (Zhou, 2010).

Hedgehog activates fused through phosphorylation to elicit a full spectrum of pathway responses

In flies and mammals, extracellular Hedgehog (Hh) molecules alter cell fates and proliferation by regulating the levels and activities of Ci/Gli family transcription factors. How Hh-induced activation of transmembrane Smoothened (Smo) proteins reverses Ci/Gli inhibition by Suppressor of Fused (SuFu) and kinesin family protein (Cos2/Kif7) binding partners is a major unanswered question. This study shows that the Fused (Fu) protein kinase is activated by Smo and Cos2 via Fu- and CK1-dependent phosphorylation. Activated Fu can recapitulate a full Hh response, stabilizing full-length Ci via Cos2 phosphorylation and activating full-length Ci by antagonizing Su(fu) and by other mechanisms. It is proposed that Smo/Cos2 interactions stimulate Fu autoactivation by concentrating Fu at the membrane. Autoactivation primes Fu for additional CK1-dependent phosphorylation, which further enhances kinase activity. In this model, Smo acts like many transmembrane receptors associated with cytoplasmic kinases, such that pathway activation is mediated by kinase oligomerization and trans-phosphorylation (Zhou, 2011).

This study has shown that Fu is activated by phosphorylation in a Hh-initiated positive feedback loop and that Fu kinase activity alone can provoke the two key outcomes of Hh signaling in Drosophila, namely Ci-155 stabilization and Ci-155 activation. This previously unrecognized central thread of the Drosophila Hh pathway is strikingly similar to receptor tyrosine kinase (RTK) pathways or cytokine pathways, where the transmembrane receptor itself or an associated cytoplasmic tyrosine kinase initiates signal transduction via intermolecular phosphorylation. In Hh signaling, engagement of the Ptc receptor leads indirectly to changes in Smo conformation, and perhaps oligomerization that are relayed to Fu via a mutual binding partner, Cos2 (Zhou, 2011).

Three activation loop residues were identified as critical for normal Fu activity. Fu with acidic residues at T151 and T154 (Fu-EE) was not active at physiological levels in the absence of Hh but could initiate Fu activation in three different ways. First, increasing Fu- EE levels induces the full spectrum of Hh target genes and responses in wing discs and is accompanied by extensive phosphorylation, undoubtedly including S159, indicating that phosphorylation can fully activate Fu. Second, low levels of a Fu-EE derivative could synergize with an excess of wild-type Fu, provided the latter molecule had an intact activation loop and was kinase-competent, indicating that a feedback phosphorylation loop could initiate Fu activation even from a ground state containing no phosphorylated residues or their mimics. Third, Hh could activate Fu-EE or wild-type Fu, but this, unlike the above mechanisms, required Cos2 and the Cos2-binding region of Fu. Activation by Hh alters Smo conformation and increases the plasma membrane concentration of Smo-Cos2 complexes, suggesting that the role of activated Smo-Cos2 complexes may simply be to aggregate Fu molecules (Zhou, 2011).

In all of the above situations there is likely an important contribution of binding between the catalytic and regulatory regions of pairs of Fu molecules to allow cross-phosphorylation, as suggested by the impotence of the Fu-EE 1-305 kinase domain alone. The sites of inferred cross-phosphorylation, T151, S159, and S482 might most simply be direct Fu auto-phosphorylation sites but they may involve the participation of an intermediate kinase. Importantly, because Fu is the key activating stimulus and Fu is the key target for activation, there is no need to postulate additional upstream regulatory inputs into a hypothetical intermediary protein kinase. Phosphorylated residues in positions analogous to Fu S159 generally stabilize the active form of the protein kinase, whereas unphosphorylated residues at other positions, closer to the DFG motif may also, or exclusively, stabilize specific inactive conformations. By analogy, phosphorylated T151, T154, and S159 are likely to serve independent, additive functions, all of which are required to generate fully active Fu kinase. There are clearly additional phosphorylated residues on Fu, including the cluster at S482, S485, and T486. These residues are not essential for Hh or Fu-EE to generate fully active Fu when Fu is expressed at high levels. However, S485A/T486A substitutions did suppress activation of GAP-Fu in wing discs and in Kc cells, suggesting that stimulation of physiological levels of Fu, perhaps by lower levels of Hh uses S482, S485, and T486 phosphorylation to favor an active conformation of Fu or productive engagement of Fu molecules. Because the S482 region may be recognized directly as a substrate by the Fu catalytic site, this region may initially mask the catalytic site (in cis or in trans) and then reduce its affinity for the catalytic site once it is phosphorylated, permitting further phosphorylation of Fu in its activation loop (Zhou, 2011).

For a long time it was thought that Fu kinase acts only to prevent inhibition of Ci-155 by Su(fu), and Fu was postulated to accomplish this by phosphorylating Su(fu). This study mapped the sites responsible for the previously observed Hh- and Fu-stimulated phosphorylation of Su(fu) and showed that they were not important for regulating Hh pathway activity. It was found that CK1, like Fu, was required for Hh to oppose Su(fu) inhibition of Ci-155 and because each of the Fu-dependent phosphorylation sites in Fu and Su(fu) that were mapped in this study prime CK1 sites it is suspected that the critical unidentified Fu and CK1 sites for antagonizing Su(fu) will be found in the same molecule, with Ci-155 itself being a prime candidate (Zhou, 2011).

This study found that Fu does considerably more than just antagonize Su(fu). It was unexpectedly found that Fu kinase can also stabilize Ci-155 via phosphorylation of Cos2 on S572, which likely leads to reduced association of Cos2 Ci-155 activation independently of Su(fu), even when Ci-155 processing was blocked by other means (Zhou, 2011).

Some insight was gained into the key regulatory role that Fu plays in Hh signaling. The truncated partially activated Fu derivative, Fu-EE 1-473, exhibited constitutive activity when expressed at high levels but, unlike full-length Fu-EE, it was not activated by Hh. Importantly, a level of Fu-EE 1-473 expression could not be found in fumH63 mutant wing discs where Hh target genes were induced at the AP border but not ectopically. Hence, Hh regulation of Fu activity appears to be essential for normal Hh signaling. This contrasts with the normal Hh signaling observed in animals lacking Su(fu) and emphasizes that Fu is a key regulatory component that has essential actions beyond antagonizing Su(fu) (Zhou, 2011).

In mice, SUFU increases Gli protein levels and inhibits Gli activators in a manner that can be overcome by Hh, much as Su(fu) affects Ci levels and activity in flies. However, in mammalian Hh signaling there is no satisfactory mechanistic model connecting Smo activation and SUFU antagonism. This study found that mouse SUFU can substitute for all of the activities of Su(fu) in flies, including a dependence on both Fu and CK1 for Hh to antagonize silencing of Ci-155. These findings, and the observation that Drosophila Su(fu) can partially substitute for murine SUFU in mouse embryo fibroblasts, suggest that SUFU silencing of Gli proteins in mice is also likely to be sensitive to analogous changes in phosphorylation produced by at least one Hh-stimulated protein kinase. Even though the murine protein kinase most similar in sequence to Drosophila Fu is not required for Hh signaling at least three other protein kinases (MAP3K10, Cdc2l1, and ULK3) have been found to contribute positively to Hh responses in cultured mammalian cells. It will be of great interest to see if these or other protein kinases are activated by Hedgehog ligands, perhaps promoted by association with Smo-Kif7 complexes in a positive feedback loop, and whether they can antagonize mSUFU to activate Gli proteins, and perhaps even stabilize Gli proteins via Kif7 phosphorylation (Zhou, 2011).

The Hedgehog-induced Smoothened conformational switch assembles a signaling complex that activates Fused by promoting its dimerization and phosphorylation

Hedgehog (Hh) transduces signal by regulating the subcellular localization and conformational state of the GPCR-like protein Smoothened (Smo) but how Smo relays the signal to cytoplasmic signaling components remains poorly understood. This study shows that Hh-induced Smo conformational change recruits Costal2 (Cos2)/Fused (Fu) and promotes Fu kinase domain dimerization. Induced dimerization through the Fu kinase domain activates Fu by inducing multi-site phosphorylation of its activation loop (AL) and phospho-mimetic mutations of AL activate the Hh pathway. Interestingly, it was observed that graded Hh signals progressively increase Fu kinase domain dimerization and AL phosphorylation, suggesting that Hh activates Fu in a dose-dependent manner. Moreover, it was found that activated Fu regulates Cubitus interruptus (Ci) by both promoting its transcriptional activator activity and inhibiting its proteolysis into a repressor form. Evidence is provided that activated Fu exerts these regulations by interfering with the formation of Ci-Sufu and Ci-Cos2-kinase complexes that normally inhibit Ci activity and promote its processing. Taken together, these results suggest that Hh-induced Smo conformational change facilitates the assembly of active Smo-Cos2-Fu signaling complexes that promote Fu kinase domain dimerization, phosphorylation and activation, and that Fu regulates both the activator and repressor forms of Ci (Shi, 2011).

How Hh signal is transduced from the GPCR-like receptor Smo to the transcription factor Ci/Gli is still poorly understood. A major unsolved issue is how a change in the Smo activation state is translated into a change in the activity of intracellular signaling complexes, which ultimately changes the balance between CiR/GliR and CiA/GliA. The current study suggests that Hh-induced conformational change of Smo exposes a Cos2 docking site(s) near the Smo C terminus that facilitates the assembly of an active Smo-Cos2-Fu complex, and that Smo activates Fu by promoting its kinase domain dimerization and phosphorylation. Evidence is provided that graded Hh signals progressively increase Fu kinase domain dimerization and phosphorylation, which may generate a Fu activity gradient, and that activated Fu regulates both CiR and CiA by controlling Ci-Sufu and Ci-Cos2-kinase complex formation (Shi, 2011).

Previous immunoprecipitation studies have revealed that Smo pulled down Cos2/Fu in both quiescent cells and Hh-stimulated cells, suggesting that Smo can form a complex with Cos2/Fu even in the absence of Hh. Furthermore, deletion analyses have indicated that both a membrane proximal domain and a C-terminal region of Smo C-tail can mediate the interaction between Smo and Cos2/Fu. Intriguingly, deleting the C-terminal region impaired, whereas deleting the membrane proximal domain potentiated, Smo activity in vivo. Further study suggested that the membrane proximal domain recruits Cos2/PP4 to inhibit Smo phosphorylation and cell-surface accumulation, which is released by Fu-mediated phosphorylation of Cos2 Ser572 in response to Hh. These observations suggest that Smo-Cos2-Fu interaction is likely to be dynamic and that distinct complexes may exist depending on the Hh signaling status. For example, Cos2 may associate with the membrane proximal region of Smo to inhibit Smo phosphorylation in quiescent cells. Upon Hh stimulation, Cos2/Fu may interact with the C-terminal region of Smo to transduce the Hh signal. In support of this model, it was found that Hh stimulated the recruitment of Cos2/Fu to the C-terminal region rather than the membrane proximal region of the Smo C tail. The increased binding depends on phosphorylation-induced conformational change of Smo C-tail that may expose the C-terminal Cos2 binding pocket(s) (Shi, 2011).

Hh signaling induces Fu kinase domain dimerization in a dose-dependent manner, most probably as a consequence of phosphorylation-induced conformational change and dimerization of Smo C tails. In addition, Hh-induced Fu dimerization depends on Cos2. Importantly, dimerization through the Fu kinase domain (CC-Fu) triggers Fu activation both in vitro and in vivo. Furthermore, CC-Fu can activate Ci in smo mutant clones and restore high levels of Hh signaling activity in cos2 mutant discs. Taken together, these results support a model in which Hh-induced Fu dimerization via Smo/Cos2 leads to Fu activation (Shi, 2011).

Both Fu dimerization and Hh stimulation induce phosphorylation of multiple Thr/Ser residues in the Fu activation loop that are important for Fu activation. Fu phosphorylation depends on its kinase activity and Fu can trans-phosphorylate itself, suggesting that Hh and dimerization may induce Fu autophosphorylation, although the results do not exclude the involvement of additional kinase(s). CC-induced dimerization does not fully activate Fu, suggesting that Smo may promote Fu activation through additional mechanisms. Activated Fu can promote phosphorylation of its C-terminal regulatory fragment, raising a possibility that Fu activation may also involve phosphorylation of its regulatory domain. Indeed, while this manuscript was under review, Zhou and Kalderon provided evidence that phosphorylation of several Ser/Thr residues in the Fu regulatory domain, likely by CK1, modulates the activity of an activated form of Fu (Zhou, 2011; Shi, 2011 and references therein).

The involvement of multiple phosphorylation events in Fu activation may provide a mechanism for fine-tuning Fu activity in response to different levels of Hh. Indeed, the efficiency of Fu dimerization and the level of activation loop phosphorylation correlate with the level of Hh signaling. Furthermore, the level of Fu activity correlates with the level of its activation loop phosphorylation. Thus, graded Hh signals may generate a Fu activity gradient by progressively increasing its dimerization and phosphorylation in response to a gradual increase in Smo phosphorylation and C-tail dimerization (Shi, 2011).

The conventional view is that Fu is required for high levels of Hh signaling by converting CiF into CiA. In support of this notion, fu mutations only affect the high, but not low, threshold Hh responsive genes. However, Fu function could have been underestimated because none of the fu mutations examined so far represents a null mutation. In addition, the existence of paralleled mechanisms, such as Gαi activation, could mask the contribution of Fu to low levels of Hh signaling. Nevertheless, a recent study using the phospho-specific antibody against Cos2 Ser572 revealed that Fu kinase activity could be induced by low levels of Hh, raising an interesting possibility that Fu may contribute to all levels of Hh signaling (Raisin, 2010). However, the lack of a fu-null mutation and the involvement of Fu in promoting Ci processing, probably through a structural role, make it difficult to directly demonstrate a role of Fu in blocking Ci processing. Using an in vivo assay for Ci processing, it was demonstrated that activated forms of Fu block Ci processing into CiR. In addition, this study found that activated Fu attenuates the association between Cos2 and Ci, as well as their association with PKA/CK1/GSK3, probably by phosphorylating Cos2, suggesting that activated Fu may block Ci processing by impeding the formation of the kinase complex required for efficient Ci phosphorylation (Shi, 2011).

Evidence is provided that activated Fu attenuates Ci/Sufu interaction. Because Sufu impedes Ci nuclear localization and may recruit a co-repressor(s) to further inhibit Ci activity in the nucleus, dissociation of Ci from Sufu may lead to the conversion of CiF to CiA. Interestingly, recent studies using mammalian cultured cells revealed that Shh signaling induces dissociation of full-length Gli proteins from Sufu (Humke, 2010; Tukachinsky, 2010), suggesting that inhibition of Sufu-Ci/Gli complex formation could be a conserved mechanism for Ci/Gli activation. Although activated forms of Fu promote Sufu phosphorylation, phospho-deficient and phospho-mimetic forms of Sufu behaved in a similar manner to wild-type Sufu in functional assays (Zhou, 2011), implying that phosphorylation of Sufu might not be a major mechanism through which Fu activates Ci. It has been shown that Shh also induces phosphorylation of full-length Gli3 that correlates with its nuclear localization (Humke, 2010). Furthermore, a Fu-related kinase Ulk3 can phosphorylate Gli proteins and promote their transcriptional activities (Maloverjan, 2010). Thus, Fu may activate Ci by promoting its phosphorylation, an interesting possibility that awaits further investigation (Shi, 2011).

Cilia-mediated Hedgehog signaling in Drosophila

Cilia mediate Hedgehog (Hh) signaling in vertebrates and Hh deregulation results in several clinical manifestations, such as obesity, cognitive disabilities, developmental malformations, and various cancers. Drosophila cells are nonciliated during development, which has led to the assumption that cilia-mediated Hh signaling is restricted to vertebrates. This study identified and characterized a cilia-mediated Hh pathway in Drosophila olfactory sensory neurons. Several fundamental key aspects of the vertebrate cilia pathway, such as ciliary localization of Smoothened and the requirement of the intraflagellar transport system, are present in Drosophila. Cos2 and Fused are required for the ciliary transport of Smoothened and cilia mediate the expression of the Hh pathway target genes. Taken together, these data demonstrate that Hh signaling in Drosophila can be mediated by two pathways and that the ciliary Hh pathway is conserved from Drosophila to vertebrates (Kuzhandaivel, 2014).

The existence of this second cilia-dependent Hh pathway in Drosophila shows that Hh signaling can be mediated via two pathways within a single organism. The results further demonstrate that the core components are shared between the two Hh pathways in Drosophila. The function of Cos2 as a putative kinesin in the ciliary compartment indicates that the ancestral Hh signaling pathway may have been cilia specific and that invertebrate cells did not maintain this specialization. Interestingly, not all vertebrate cells have primary cilia, and different types of tumors react differently to Shh depending on whether they are ciliated, indicating that there might be a second, overlooked nonciliary pathway in vertebrates (Kuzhandaivel, 2014).

Genetic in vivo analysis of Smo ciliary localization revealed that, as in vertebrates, the ciliary IFT system and a ciliary localization signal are required for localization of Smo to cilia in Drosophila. The results further show that the Hh receptor Ptc regulates Smo stability and that ciliary localization depends on the activation of the kinesin-like protein Cos2. In the Drosophila wing disc, Fu regulates Cos2 function and is required for most aspects of Hh signaling. The current data show that Fu is also required for Cos2 ciliary localization and Smo transport within the cilia. However, Fu is not essential for mammalian Hh signaling, and in zebrafish, loss of Fu results in weak Hh-related morphological phenotypes. These differences from the Drosophila pathway and vertebrate ciliary signaling could be explained by the existence of a second, as yet unidentified kinase with an analogous function. Cell culture and in vivo studies in vertebrates led to the identification of four kinases with phenotypes related to Fu: Ulk3, Kif11, Map3K10, and Dyrk2. Further investigation is required to determine whether these kinases control the ciliary transport of Smo and whether Cos2 Smo transport is conserved in vertebrates. Yet, the current results demonstrate that cilia-mediated Hh signaling does occur in Drosophila and that this pathway is conserved in vertebrates, which makes the Drosophila OSN a powerful in vivo model for studying Hh signaling and its ciliary transport regulation (Kuzhandaivel, 2014).

costa/costal2: Biological Overview | Evolutionary Homologs | Developmental Biology | Effects of Mutation | References

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.