Suppressor of Hairless


Protein Interactions

The DNA-binding transcription factor Suppressor of Hairless [Su(H)] functions as an activator during Notch (N) pathway signaling, but can act as a repressor in the absence of signaling. Hairless (H), a novel Drosophila protein, binds to Su(H) and has been proposed to antagonize N signaling by inhibiting DNA binding by Su(H). In vitro, H directly binds two corepressor proteins, Groucho (Gro) and dCtBP. Reduction of gro or dCtBP function enhances H mutant phenotypes and suppresses N phenotypes in the adult mechanosensory bristle. This activity of gro is surprising, because it is directed oppositely to its traditionally defined role as a neurogenic gene. Su(H)-H complexes can bind to DNA with high efficiency in vitro. Furthermore, a H-VP16 fusion protein causes dominant-negative phenotypes in vivo, a result consistent with the proposal that H functions in transcriptional repression. Taken together, these findings indicate that 'default repression' of N pathway target genes by an unusual adaptor/corepressor complex is essential for proper cell fate specification during Drosophila peripheral nervous system development (Barolo, 2002b).

H is a novel protein, with no known vertebrate homologs. However, the H gene has been identified in three members of the order Diptera: Drosophila melanogaster, D. hydei, and the mosquito Anopheles gambiae. H is surprisingly poorly conserved among these three species: It shares 63% identity between D. melanogaster and D. hydei (diverged ~65 Mya), and 33% identity between Drosophila and Anopheles (diverged ~260 Mya). The rapid divergence of the H protein sequence readily allows the identification of short conserved motifs, which are presumably important for H function. Two such regions occur in a part of H that is required for its interaction with Su(H) in vitro (Barolo, 2002b).

Another conserved motif in the H protein is YSIxxLLG, which is perfectly conserved from Drosophila to Anopheles. This sequence resembles certain examples of the 'eh1' type of Gro-binding domain found in many transcriptional repressor proteins. Among eh1 domains, the 'octapeptide' motifs in the Pax 2/5/8 proteins, which have been shown to directly mediate repression by recruiting Gro-family corepressors, show the greatest similarity to this region of H. In addition, the extreme C-terminal sequence of H, PLNLSKH, includes a match to the consensus binding site for the CtBP corepressor, Px(D/N)LS. The PLNLS motif, fully conserved from Drosophila to Anopheles, exactly matches motifs found in four vertebrate CtBP-binding transcription factors. H also contains three lengthy alanine-repeat domains: AAAVAAAAAAAAA, AAAAAAAAAA, and AAVAAA AAAAAA. Alanine repeats and alanine-rich regions are common in transcriptional repression domains, and are found in many repressor proteins. However, these repeats are reduced or absent in the D. hydei and A. gambiae H proteins: this suggests that they may not make an essential contribution to H function (Barolo, 2002b).

A gel retardation experiment reported by Brou (1994), indicating that H can inhibit the binding of Su(H) to DNA in vitro, has strongly influenced interpretations of genetic studies of H, Su(H), and N. A DNA-binding-inhibition model of H function is indeed consistent with both loss- and gain-of-function genetic data demonstrating that H affects cell fate in a manner antagonistic to N signaling, including the N-stimulated transcriptional activation function of Su(H). However, the recent discovery of Su(H)-mediated transcriptional repression has forced a reconsideration of this simple model, since it makes incorrect predictions about the effect of H on a cell fate that is dependent on the repression function of Su(H). It is proposed that the genetic data on cell fate are instead consistent with a different role for H: facilitating transcriptional repression by Su(H) (Barolo, 2002b).

During the socket/shaft cell fate decision in adult mechanosensory bristle development, the cell that responds to N signaling takes the socket fate, while its sister cell, in which N signal transduction is blocked by the Numb protein, takes the shaft fate. Overexpression of Su(H), or loss of H function, during the socket/shaft decision causes both cells to adopt the socket fate; conversely, overexpression of H, or loss of Su(H) function, results in two shaft cells. Autorepression by Su(H) in shaft cells is important for maintaining the shaft cell fate. The corepressors Gro and dCtBP are important for specification of the shaft cell, a fate that is inhibited by N signaling and depends on both H activity and Su(H)-mediated repression. Reduction of gro or dCtBP function strongly enhances the effects of both reduction of H activity and loss of Su(H) repression, and suppresses the effects of reduced N signaling in the bristle lineage. It is therefore concluded that Gro and dCtBP, along with H and transcriptional repression mediated by Su(H), act in the opposite direction from the N signaling pathway during the socket/shaft cell fate decision, in that they promote the fate (shaft) that is inhibited by N signaling. The observation that both gro and dCtBP heterozygotes show a weak dominant (haploinsufficient) shaft-to-socket cell fate conversion phenotype is further confirmation of an important role for both corepressors in promoting the shaft cell fate. These results represent the first in vivo functional evidence for the involvement of Gro and dCtBP in transcriptional repression mediated by Su(H) (Barolo, 2002b).

Genetic analyses show that gro loss-of-function mutations enhance the effects of reduced H activity on two N-mediated cell fate decisions, the socket/shaft decision and the epidermal/SOP decision, while reduction of gro activity suppresses the effects of N loss of function on the socket/shaft and pIIA/pIIB cell fate decisions. In addition, gro has a weak haploinsufficient bristle loss phenotype, resembling an excess of N signaling. A role for gro in promoting the SOP cell fate is surprising, because gro was originally identified as a 'neurogenic' gene that acts to inhibit the SOP fate downstream of N signaling, in its capacity as a corepressor for bHLH transcriptional repressor proteins encoded by N target genes in the Enhancer of split gene complex [E(spl)-C]. In fact, gro was named after the phenotype of flies homozygous for gro1, a weak hypomorphic allele: bushy tufts of bristles over the eyes caused by a failure of N-mediated lateral inhibition of the SOP fate. At least one E(spl)-C bHLH repressor gene appears to be directly repressed by Su(H) in SOPs; the proposal that Gro promotes the SOP fate by cooperating with H to repress N target genes in this cell is currently being tested. If proved, this would represent a novel and complex form of regulation, in which Gro inhibits the SOP fate in all but one cell of the proneural cluster by partnering with the E(spl)-C bHLH repressors, and simultaneously promotes the SOP fate in one neighboring cell by preventing the expression of its own partners (Barolo, 2002b).

Potent inhibition of Su(H)'s DNA binding activity by H, as proposed by Brou (1994), is clearly incompatible with the proposition that a H/Su(H) complex directly represses N/Su(H) target genes. The mechanosensory organ lineage offers a particularly clear experimental example of this conflict. Proper specification of the shaft cell fate requires autorepression by Su(H); loss of this repression causes the shaft cell to transform its fate at substantial frequency to that of its sister, the socket cell. If H functions primarily to antagonize DNA binding by Su(H), then reduction of H activity should if anything lead to an increase in Su(H) autorepression, which should in turn stabilize the shaft fate. Instead, the shaft cell is highly sensitive to decreased H function, which readily causes its transformation to a socket cell. Thus, the genetic data on H's role in preserving the bristle shaft cell fate are irreconcilable with a simple Su(H)-DNA-binding-inhibition model for H function (Barolo, 2002b).

More generally, it has become apparent that repression of N targets by Su(H) is just as general and important a mechanism in Drosophila as it is in vertebrates. If Su(H)-mediated repression of N target genes is essential for proper specification of N-independent cell fates, as has been shown for the shaft cell, then preventing Su(H)-DNA interaction by overexpressing H would have the same effect as deleting Su(H) binding sites in N target genes: namely, their derepression, leading to N gain-of-function phenotypes (such as shaft-to-socket conversions). Instead, however, overexpression of H has been shown to repress the activity of N-regulated genes, and the phenotypic effects of gain and loss of H function suggest a strictly antagonistic relationship between H and N signaling. Therefore, it is believed that the accumulated genetic data point to a role for H in facilitating Su(H)-mediated repression of N target genes (Barolo, 2002b).

Purified Su(H) can bind efficiently to both purified H and DNA simultaneously, allowing the possibility that a H/Su(H) complex may function as a transcription factor. A weak supershift of Su(H)/DNA complexes by full-length H has been reported, and a stronger supershift by an N-terminal H fragment (H1-293). S robust supershift of Su(H)/DNA complexes occurs with full-length H, strongly supporting the notion that H/Su(H)/DNA complexes may form efficiently in vivo. The discrepancy between these results and those of previous studies may reflect different experimental conditions, such as buffer composition or protein purification protocols, or may be due to the relatively low H:Su(H) molar ratios used in these experiments. Further work will be needed to reject or confirm the possibility that H antagonizes the Su(H)-DNA interaction under physiological conditions. However, given the genetic arguments outlined above, and the consistency of the results with a view of H as a transcriptional repressor, it seems likely that DNA-binding inhibition is not the primary mechanism by which H contributes to the specification of N-independent cell fates (Barolo, 2002b).

Misexpression of wild-type and modified forms of H in the adult bristle lineage has led to three conclusions about the function of the H protein in vivo. (1) The replacement of the putative repression domains of H with a transcriptional activation domain results in a dominant-negative form of H that elicits N gain-of-function phenotypes, a result consistent with normal repression of N target genes by H. Conversion to an antimorphic form by the addition of an activation domain is a common property of transcriptional repressor proteins (Barolo, 2002b).

(2) The significant, but somewhat weakened, effects of HDeltaC relative to wild-type H suggest that the C-terminal region of H including a dCtBP-interaction domain is important for some, but not all, of H's activity in vivo. This conclusion is supported by analysis of the H22 allele, which produces mutant H protein lacking its C-terminal 69 aa, including the PLNLS motif. H22, unlike H null alleles, is a homozygous viable mutation, and its effects on N-mediated cell fate decisions, though strong, are milder than those of H null mutations. The fact that the H22/H22 phenotype, unlike the H null heterozygote phenotype, is not enhanced by loss of dCtBP, as well as the absence of any other consensus CtBP binding sites in the H protein, are consistent with the idea that H interacts with dCtBP solely via this C-terminal motif. If this is indeed the case, then both the mildness of the H22 mutant phenotype and the potency of misexpressed HDeltaC protein indicate that the dCtBP corepressor contributes some, but not all, of the repressive activity of the H protein (Barolo, 2002b).

(3) The dominant-negative activity of H-VP16 depends on the removal of the region of H containing the Gro-binding domain, suggesting that this region contributes some of the wild-type repression function of the H protein. These last two conclusions conflict with the assertion that the C-terminal dCtBP binding site is wholly responsible for H-mediated repression. Overall, the misexpression experiments provide evidence that both the region of H that binds to dCtBP and the region that binds to Gro contribute to the function of H in vivo. This is consistent with the strong phenotypic interaction between H and the genes encoding these two corepressors (Barolo, 2002b).

The current results support the hypothesis that H antagonizes N signaling by acting as an adaptor molecule between the transcription factor Su(H) and the corepressor proteins Gro and dCtBP. This model entails an unusual mechanism of repression: DNA-binding transcriptional repressors that recruit CtBP or the Gro family of corepressors generally do so via direct protein-protein interactions, although evidence for CtBP recruitment by non-DNA-binding proteins has been reported. In mammalian cells, the corepressors SMRT and CIR bind directly to the Su(H) homolog CBF1 (Barolo, 2002b).

If protein complexes containing H are important for Su(H)-mediated repression, why is H not found in vertebrates? Several possible explanations are apparent. First, vertebrate homologs of Su(H) may not make use of an adaptor/corepressor complex, but rather may recruit all corepressors (possibly including Gro and CtBP) directly, as in the case of SMRT and CIR. A second, related possibility is that vertebrate versions of Su(H) do not utilize Gro and CtBP as corepressors; in this view, H may have appeared exclusively in the protostome lineage to add to Su(H)'s corepressor repertoire. Third, it is possible that vertebrates employ not a homolog, but an analog, of H, one evolved independently after the divergence of protostomes and deuterostomes. Finally, the fact that the predicted D. melanogaster H protein is only 33% identical to its apparent ortholog in A. gambiae suggests that the vast majority of the H sequence is not under selective constraint. Thus, an ortholog of H may indeed exist in vertebrates, but be so highly diverged as to be unrecognizable by typical sequence analyses. This possibility seems less likely, since at least the Su(H)-interaction domain might be expected to be well conserved, given the strong evolutionary conservation of Su(H) itself (Barolo, 2002b).

In contrast to a DNA-binding inhibition model for H function, an adaptor/corepressor model explains why H counters NIC/Su(H)-mediated activation, but not Su(H)-mediated repression. Like previous views of H function, this model presumes competition between Su(H)-binding partners, in this case between NIC-containing activation complexes and H/Gro/dCtBP repression complexes. NIC activation complexes are likely to include the Mastermind (Mam) protein, and may also include the p300 coactivator. In the presence of N signaling, Su(H)/NIC/Mam complexes presumably replace Su(H)/H/Gro/dCtBP complexes on target genes, and convert Su(H) from a repressor to an activator. Whether this occurs by simple affinity-based competition for binding to Su(H), or by a mechanism involving active impairment of the H/Su(H) interaction, is unknown. Under an adaptor/corepressor model, the H mutant phenotype results from derepression of Su(H)/N target genes in cells lacking N pathway activity, thus mimicking an increase in N signaling. The H overexpression phenotype may be explained by the displacement of NIC-containing activation complexes by an excess of H-containing repression complexes, thus repressing NIC/Su(H) target genes in cells that respond to the N signal (Barolo, 2002b).

Similarly, it is proposed that the Su(H) overexpression phenotype, which resembles a gain of N function, is caused by a titration of H repression complexes by excess Su(H), and the subsequent derepression of Su(H) target genes. The fact that overexpression of Su(H) strongly enhances the effect of H overexpression on lateral inhibition supports this view, but is very much at odds with a DNA-binding-inhibition model for H function (Barolo, 2002b).

It has recently become apparent that the transcriptional target genes of at least six major developmental signaling pathways are in many cases subject to 'default repression'; that is, binding sites for signal-regulated transcription factors, which mediate activation during signaling events, mediate repression in the absence of signaling (for review, see Barolo, 2002a). Each of these pathways uses a different mechanism to switch from repression to activation upon stimulation of the pathway, but in each case, the effect seems to be the same: restricting the expression of pathway target genes to cells that receive active signaling. The results of this study strongly suggest that H contributes to default repression in the N pathway by directly recruiting the corepressors Gro and dCtBP to Su(H), and that formation of H/Su(H) repression complexes is crucial for the establishment of two N-inhibited cell fates, the SOP and shaft cell fates. Default repression, therefore, appears to be as important as signal-dependent activation for proper cell fate specification in this developmental context (Barolo, 2002b).

Hairless plays an important role as the major antagonist in the Notch signaling pathway in Drosophila. It appears to be a direct inhibitor of the signal transducer Su(H). Hairless encodes a pioneer protein that has been dissected in a structure-function analysis; a series of deletion constructs was tested for wild type and gain of function activity in the fly as well as for Su(H) binding. In this way, the Hairless protein was subdivided into the absolutely essential Su(H)-binding domain, important N- and C-terminal domains and a central antimorphic domain. A construct C2 that deletes the Su(H) binding domain has some activity during wing development, suggesting that Hairless has additional functions apart from Su(H) binding. For example, overexpression of the C2 deleted protein causes a novel, net-like wing phenomenon that cannot be explained by Su(H) inhibition. The central acidic domain may mark a repression domain of the Hairless protein required for silencing Hairless function, e.g. for releasing Su(H) from a H/Su(H) complex. It is speculated that the C-terminal region comprises an interactive surface for additional components involved in H function. Therefore, Hairless protein might have additional functions apart from Su(H) binding and may antagonize Notch mediated cell-cell communication in a more complex way than currently anticipated (Maier, 1997).

Su(H) binds directly to the ankyrin repeats of the intracellular domain of Notch. When Delta expressing cells interact with Notch expressing cells, cytoplasmic Notch-Su(H) interaction is relieved by direct interaction of Notch with Deltex, and Su(H) is subsequently detected mostly in the nucleus (Fortini, 1994). Hairless gene product can inhibit the DNA binding of Su(H) through direct protein-protein interaction. Thus Hairless acts as a negative regulator of Su(H) activity (Brou, 1994).

The Notch (N) pathway defines an evolutionarily conserved cell signaling mechanism that governs cell fate choices through local cell interactions. The ankyrin repeat region of the Notch receptor is essential for signaling and has been implicated in the interactions between Notch and two intracellular elements of the pathway: Deltex (Dx) and Suppressor of Hairless (Su[H]). The function of the Notch cdc10/ankyrin repeats (ANK repeats) was examined by transgenic and biochemical analysis. In vivo expression of the Notch ANK repeats reveals a cell non-autonomous effect and elicits mutant phenotypes that indicate the existence of novel downstream events in Notch signaling. The intracellular domain induces five cone cells, a phenotype consistent with the idea that this truncated form of Notch inhibits the R7 precursor which acquires R7 fate; instead, it turns into an additional cone cell. The intracellular domain induces the activation of Mdelta, a bHLH member of the E(spl) complex. In contrast, an ankyrin repeat transgene suppresses ectopic Mdelta expression. It is thought that the suppression of Mdelta by ANK repeats does not reflect a dominant negative behavior associated with an overexpressin of the ANK repeats, and instead suggests that ankyrin repeats exert their action independent of Su(H). Su(H) binds to both a subtransmembrane region that excludes the ankyrin repeats; it also binds the ANK repeats themselves. However, a peptide encompassing just the ANK repeats does not bind independently to Su(H) but is capable of binding to Deltex. The ANK repeats also mediate homotypic interactions, a property that may underly the biological function of the repeats. The simplest way to explain the non-autonomous, antagonistic action of the ANK sequences is to suggest that the ANK-expressing cells down-regulate the endogenous Delta activity. These results suggest the existence of yet unidentified Notch pathway components (Matsuno, 1997).

Su(H) mutations were first isolated as dominant suppressors of the Hairless dominant, haploinsufficient loss-of-function mutation. Su(H) binds to the RBP consensus-binding site of Hairy Enhancer of Split gene [See Enhancer of Split for information about this mammalian homolog of E(spl)]. Hairless acts as a negative regulator of this DNA binding activity of Su(H). Negative regulation by Hairless results from the formation of H-Su(H) complexes that can prevent the in vivo binding of Su(H) to the DNA (Brou, 1994).

Cell-cell signaling mediated by the receptor Notch regulates the differentiation of a wide variety of cell types in invertebrate and vertebrate species, but the mechanism for signal transduction following receptor activation is unknown. A recent model proposes that ligand binding induces intracellular processing of Notch; the processed intracellular form of Notch then translocates to the nucleus and interacts with DNA-bound Suppressor of Hairless [Su(H)], a transcription factor required for target gene expression. Intracellular cleaveage has been suggested to occur within either the transmembrane domin or the first 10 amino acids of the cytoplasmic domain. As intracellular processing of endogenous Notch has so far escaped immunodetection, a sensitive nuclear-activity assay was devised to monitor indirectly the processing of an engineered Notch in vivo. First, the non-membrane-tethered intracellular domain of Notch, fused to the DNA-binding domain of Gal4, regulates transcription in a Delta-independent manner. This transcriptional regulation requires Su(H) activity, suggesting that Su(H) may not only target the Notch intracellular domain to the DNA but may also have an additional function. For instance, Su(H) may be required to protect processed Notch from degradation, or participate in transcriptional activation together with processed Notch. Subsequently, full-length Notch, containing the Gal4 DNA-binding domain inserted 27 amino acids carboxy-terminal to the transmembrane domain, activates transcription in a Delta-dependent manner. These results provide indirect evidence for a ligand-dependent intracellular processing event in vivo, supporting the view that Su(H)-dependent Notch signaling involves intracellular cleavage, and transcriptional regulation by processed Notch (Lecourtois, 1998).

In the development of the socket cell of the mechanosensory organ, Su(H) does not co-localize with Notch and Deltex at the apicolateral membrane. Instead Su(H) protein appears to be distributed evenly in the cytoplasm of the socket cell. This result is paradoxical since Notch is believed to regulate the cytoplamic retention of Su(H). Implied is a dynamic equilibrium between membrane-associated and cytoplasmic Su(H) that is largely in favor of the accumulation of Su(H) in the cytoplasm (Gho, 1996).

The role of the Notch signaling pathway has been examined in the transcriptional regulation of two Drosophila Enhancer of split [E(spl)] genes. Using a reporter assay system in Drosophila tissue culture cells, a significant induction of E(spl) m gamma and m delta expression is observed after cotransfection with activated Notch. Characterization of the 5' regulatory regions of these two genes led to the identification of a number of target sites for the Suppressor of Hairless [Su(H)] protein, a transcription factor activated by Notch signaling. Su(H) binding sites are present in the upstream regions of both E(spl) genes. Notch-inducible expression of E(spl) m gamma and m delta, both in cultured cells and in vivo, is dependent on functional Su(H). Although overexpression of Su(H) augments the level of induction of the reporter genes by activated Notch, Su(H) alone is insufficient to produce high levels of transcriptional activation. Despite the synergy observed between activated Notch and Su(H), the former affects neither the nuclear localization nor the DNA binding activity of the latter. The behavior of Drosophila Notch is consistent with a mechanism whereby N activates Su(H) by covalent modification. It is unlikely that N functions to sequester Su(H) in the cytoplasm, since Su(H) is nuclear. There also are no apparent differences in strength in reporter gene activation in Drosophila between nuclear and membrane bound forms of activated Notch. In the covalent modification hypothesis, N-Dl binding could result in the binding and/or activation of a modifying enzyme (such as a kinase or methylase) which could act on Notch-bound Su(H) (Eastman, 1997).

Numb protein, known to interact with the cytoplasmic domain of Notch, interfers with the ability of Notch to cause the nuclear translocation of Suppressor of hairless. Both the C-terminal half of the highly conserved phosphotyrosine binding domain of Numb and the C-terminus of Numb are required to inhibit Notch. Numb prevents the colocalization to the nucleus of cells of the Notch intracellular domain with Su(H) resulting in a cytoplasmic localization. Overexpression of Numb during wing development, which is sensitive to Notch dosage, reveals that Numb is also able to inhibit the Notch receptor in vivo. In the external sense organ lineage, the phosphotyrosine binding domain of Numb is found to be essential for the function but not for asymmetric localization of Numb. These results suggest that Numb determines daughter cell fates in the external sense organ lineage by inhibiting Notch signaling (Frise, 1996).

Ligand-induced cleavage and regulation of nuclear entry of Notch in Drosophila melanogaster embryos

Drosophila Notch is processed in a ligand-dependent fashion to generate phosphorylated, soluble intracellular derivatives. During most of Drosophila embryogenesis, two size classes of N proteins are coimmunoprecipitated by antibodies against Su(H). These include full-length N proteins and, to a greater extent, phosphoproteins of ~114-kD, Npp114. Unlike mammalian systems in which N exists predominantly as a heterodimer, during Drosophila embryogenesis, the bulk of N exists as the full-length form. When dephosphorylated, Npp114 resolves into three proteins, each ~100 kD: Np100A, Np100B, and Np100C. Through most of embryogenesis, the most abundant of these proteins is Np100B, Np100C being found only late in development. The size difference between the two proteins might be because Np100C has been cleaved further than Np100B, or the two proteins may both have the same amino termini, but Np100C might have been additionally cleaved at the carboxyl terminus. It is also possible that there is a precursor-product relationship between the two. In any case, the occurrence of Np100C only late in embryogenesis suggests that production of these forms of N is under developmental control. Throughout most of embryogenesis, the majority of processed N proteins that are associated with Su(H) show some level of phosphorylation. Full-length N has been shown previously to be phosphorylated on serines. It is not known how the latter relates to the phosphorylation described here, although the presence of hypophosphorylated forms of N bound to Su(H) suggests that the two events are unrelated. How this phosphorylation is effected and how it influences N function is not known. There are two lines of evidence that suggest that phosphorylation is not an immediate consequence of ligand binding and cleavage. (1) Most if not all of NIntra1790 (a constructed soluble intracellular domain of Notch) is phosphorylated -- none of which has been produced as a result of ligand binding and cleavage of N. (2) Overexpression of Dl induces at least one processed form of N which is hypophosphorylated. In addition, phosphorylation of NIntra1790 is not dependent on the presence of Su(H). Because most, if not all, of NIntra1790 is phosphorylated and there is an enrichment of Npp114 in the soluble fraction, perhaps phosphorylation is related to the release of cleaved intracellular N from the membrane. Alternatively, phosphorylation may promote nuclear translocation or association with Su(H), or both. There is some salt extractable Npp114 associated with Su(H) in the membrane fraction. Finding the intracellular domain of N, which contains functional nuclear localization signals either in the membrane or cytoplasmic fractions, indicates that the cell contains mechanisms to restrain the nuclear entry of N cleavage products. Because it has been demonstrated that the cdc10 repeats of N mediate homodimerization, newly produced intracellular forms of N may be retained by full-length forms of N at the membrane. This association might be particularly favored if, as believed, the receptor is presented at the cell surface as a dimer. It is also conceivable that Npp114 is retained on the membrane by a complex of Su(H) and full-length N (Kidd, 1998).

Su(H) may regulate nuclear entry of N. With respect to cytoplasmic retention of Su(H)/Npp114 complexes, regulation may come from Su(H) itself. Whereas coexpressing high levels of NIntra along with Su(H) in S2 cells results in both proteins translocating to nuclei, when low levels of NIntra are coexpressed along with Su(H) in S2 cells, there is retention of NIntra in the cytoplasm. This suggests that excess Su(H) can promote cytoplasmic localization of soluble, intracellular forms of N. Given that there are multiple binding sites for Su(H) in the cytoplasmic domain of N, differences in subcellular localization could reflect the number of Su(H) molecules bound to N, with changes in stoichiometry resulting from increased levels of intracellular N in response to ligand. Because in vivo levels of Su(H) appear to be in excess of soluble Notch product due to sufficient Su(H) to bind to ectopically expressed NIntra and generate gain of function phenotypes, the cytoplasmic retention observed in Su(H)+ embryos is expected from the S2 cell studies. Further supporting the view that Su(H) can retain soluble N in the cytoplasm in vivo, it has been found that lowering the dose of Su(H) promotes nuclear localization of NIntra in embryos (Kidd, 1998).

When tethered directly to DNA, the cytoplasmic domain of N can activate transcription. Conversely, a viral activator fused to Su(H) can substitute for at least some N functions during embryogenesis. It is suggested that one function of soluble forms of N is to bind to Su(H), and in the nucleus, to act directly as a transcriptional transactivator of the latter protein. The data presented here suggest that the prime function of the sequences downstream of the cdc10 repeats is to provide transactivator activity. In accord with this, the cytoplasmic domain of N has many features that are found in transcriptional activators. Although it is possible that N indirectly confers activating ability on Su(H), given the finding of appropriately processed N proteins, which contain functional nuclear localization signals preferentially associated with Su(H), the simplest interpretation of thes results is that one function of N is to bind to Su(H) and in the nucleus to directly act as its transcriptional transactivator. Recently it has been suggested that N activates transcription by disrupting the formation of a repressor complex between Su(H) and a histone deacetylase complex (SMRT/HDAC-1). The data here suggest that rather than simply disrupting the Su(H)/SMRT/HDAC-1 complex, Npp114 plays a more active role in providing transactivator activity to Su(H) (Kidd, 1998 and references).

Subcellular localization of Hairless protein shows a major focus of activity within the nucleus

Hairless, a major antagonist of the Notch signaling-pathway in Drosophila, associates with Suppressor of Hairless [Su(H)], thereby inhibiting trans-activation of Notch target genes. These molecular interactions could occur either at the step of signal transduction in the cytoplasm or during implementation of the signal within the nucleus. The subcellular distribution of Hairless was examined. Hairless is a low abundant, ubiquitous protein that is cytosolic as well as nuclear. High levels of Hairless cause nuclear retention of Su(H); loss of Hairless reduces the amount of Su(H) in the nucleus (Maier, 1999).

In order to examine the Hairless expression pattern, antisera were raised that were directed against, respectively, the central portion of Hairless, termed anti-A, and the C-terminal part, termed anti-B. Upon forced expression, both antisera detect ectopic Hairless with high specificity almost exclusively within nuclei. Induction of Hairless from a heat shock controlled transgene results in a strong nuclear accumulation of Hairless in somatic tissues, like the follicle cells of ovaries. Nuclear accumulation of Hairless is also observed, when ectopic expression is forced from a UAS-H transgene in a pair rule pattern in the embryo or along the antero-posterior boundary in wing imaginal discs. Thus, although anti-A does not recognize nuclear Hairless in wild type, it does detect ectopically expressed Hairless in a sub-cellular distribution identical to that of anti-B. This implies that nuclear Hairless at wild type levels is modified or bound by another nuclear protein, which is present in limiting amounts and able to mask the anti-A epitopes upon binding to Hairless. Su(H) has been shown previously to bind to Hairless in the vicinity of the A-domain. It was then asked whether the anti-A epitope of Hairless is demasked in cells mutant for Su(H). However, in embryos devoid of Su(H) protein, anti-A exclusively detects cytoplasmic Hairless. The same is observed in Su(H) clones induced in imaginal discs. Thus, Su(H) binding plays no major role in the recognition of Hairless by anti-A, and it remains undecided, whether masking of Hairless by protein(s) other than Su(H) or its modification accounts for the differential recognition by the two antisera (Maier, 1999).

The reverse experiment was performed, asking whether Su(H) protein distribution depends on Hairless. Mutant cell clones devoid of Hairless show a markedly reduced amount of nuclear Su(H) protein. Conversely, ectopic expression of Hairless causes Su(H) to accumulate strongly in the nuclei of Hairless overexpressing cells. Calculations on Su(H) distribution show that this accumulation is not simply due to an overall increase of Su(H) in the Hairless expressing cells but mainly caused by an increased nuclear vs. cytoplasmic ratio, compared to the surrounding wild type cells. In situ hybridization of such discs with a Su(H) probe reveals no ectopic Su(H) transcription. Thus, the increase of nuclear Su(H) in response to Hairless must occur primarily at the level of protein. As a result of the ectopic Hairless expression along the antero-posterior border, adult wings are notched at the tip, resulting from interference of Hairless with the Notch-dependent process of wing margin formation (Maier, 1999).

Hairless subcellular distribution is consistent with antagonizing Su(H) activity in either compartment. The finding that levels of nuclear Su(H) accumulation are linked to Hairless levels places this interaction within the nucleus. It remains for future studies to determine whether nuclear retention prevents Su(H) from shuttling back to the membrane, recharging another Notch receptor or whether nuclear accumulation reflects a dual function of Su(H) in Drosophila, as suggested for the mammalian homolog, which acts in repression and activation of Notch target genes (Maier, 1999).

Transcriptional repression by Suppressor of Hairless involves the binding of a Hairless-dCtBP complex in Drosophila

Notch signal transduction involves the presenilin-dependent intracellular processing of Notch and the nuclear translocation of the intracellular domain of Notch, NICD. NICD associates with Suppressor of Hairless [Su(H)], a DNA binding protein, and Mastermind (Mam), a transcriptional coactivator. In the absence of Notch signaling, Su(H) acts as a transcriptional repressor. Repression by Su(H) is relieved by the activation of Notch. In the Drosophila embryo, this transcriptional switch from repression to activation is important for patterning the expression of the single-minded (sim) gene along the dorsoventral axis. The mechanisms by which Su(H) inhibits the expression of Notch target genes in Drosophila has been investigated. Hairless, an antagonist of Notch signaling, is required to repress the transcription of the sim gene. Hairless forms a DNA-bound complex with Su(H). Furthermore, it directly binds the Drosophila C-terminal Binding Protein (dCtBP), which acts as a transcriptional corepressor. The dCtBP binding motif of Hairless is essential for the function of Hairless in vivo. It is proposed that Hairless mediates transcriptional repression by Su(H) via the recruitment of dCtBP (Morel, 2001).

The sim gene is expressed in a single row of cells abutting the mesoderm in the Drosophila embryo at the cellular blastoderm stage. Sim confers to these cells a mesectodermal identity. Su(H) has a dual function in the regulation of sim expression: (1) Su(H) directly inhibits the expression of the sim gene in the neuroectoderm -- in Su(H) mutant embryos derived from germ-line clones (GLC), both endogenous sim and a sim-lacZ transgene that mimics the expression of sim are ectopically expressed in 2-3 rows of neuroectodermal cells; (2) Su(H) upregulates the expression of sim in the mesectoderm. Transcriptional activation by Su(H) depends on Notch signaling, but repression by Su(H) is independent of Notch activity (Morel, 2001).

Hairless is a nuclear protein that binds Su(H) and antagonizes Notch activity in numerous cell fate decisions. It is, however, unclear how Hairless inhibits transcriptional activation by Su(H). One hypothesis is that Hairless promotes repression by Su(H). Analysis of the sim promoter has allowed for a test of whether Hairless is required for Su(H)-dependent repression. The expression of the sim gene was analyzed in embryos that lack both the maternal and the zygotic contributions of Hairless. In these Hairless GLC embryos, sim-lacZ is ectopically expressed in cells located dorsally to the mesectoderm. This ectopic expression in the neuroectoderm is similar to the one observed in Su(H) mutant embryos. Thus, Hairless is required for the repression of sim in the same cells that also require Su(H) for repression (Morel, 2001).

Hairless binds Su(H) and inhibits transcriptional activation by Su(H). However, because Su(H)-Hairless complexes do not bind DNA in vitro, Hairless has been proposed to inhibit the DNA binding activity of Su(H). In this model, Hairless inhibits Notch signaling by titrating Su(H) and not by repressing Notch target genes. This model does not explain, however, why the loss of Hairless activity leads to the same phenotype as the loss of Su(H) activity, i.e., derepression of sim in neuroectodermal cells. It is therefore envisaged that Hairless and Su(H) act together to repress transcription. This implies that Hairless must bind to a DNA bound form of Su(H). Whether Hairless and Su(H) can form a DNA bound complex has been reexamined in a gel retardation assay. Hairless and a truncated version of Hairless that binds Su(H) (H[1-293]) are shown to supershift a Su(H)-oligonucleotide complex. It is concluded that Hairless associates with DNA via Su(H) (Morel, 2001).

To test whether Hairless is able to cooperate with Su(H) in vivo, an assay based on the expression of Hairless and Su(H) during pupal development was used. Lateral inhibition mediated by Notch signaling controls the spacing of bristle sensory organs on the dorsal thorax of the fly. Increasing the level of Notch signaling results in the determination of a reduced number of sense organs. Conversely, decreasing the level of Notch signaling leads to an increased density of sense organs. Similarly, overexpression of Su(H) under heat-shock control decreases sense organ density, while overexpression of Hairless has the opposite effect. Control flies in which the expression of either Su(H) or Hairless was induced under mild heat shock conditions (30 min at 37°C) display only a weakly decreased or increased bristle density, respectively. The titration model for Hairless predicts that when Su(H) and Hairless are simultaneously overexpressed, they should counteract each other's activity to produce an intermediate phenotype. In contrast to this prediction, low-level expression of Hairless and Su(H), under the same mild heat shock conditions, leads to a dramatic increase in sense organ density on the dorsal thorax. This shows that Su(H) and Hairless strongly synergize to inhibit Notch signaling in this experimental situation. This synergy between Hairless and Su(H) was also seen for the regulation of sim expression in the mesectoderm, as well as during wing development. These findings suggest that this synergy represents a general feature of the function of these two genes. These observations are not consistent with the titration model described above, but rather they support the hypothesis that Hairless acts in a Su(H)-dependent manner to antagonize Notch signaling activity (Morel, 2001).

The mechanism by which Hairless might regulate transcription was investigated. Sequence analysis of Hairless identifies a putative binding site for the Drosophila C-terminal Binding Protein (dCtBP). This site is located at the very C terminus of the Hairless protein. In Drosophila and mammals, CtBP is a transcriptional corepressor. It was therefore tested whether Hairless binds to dCtBP. The full-length Hairless protein, H[1-1076], interacts with dCtBP in a yeast two-hybrid assay. In contrast, a truncated version of Hairless in which the last 15 amino acids had been deleted, H[1-1061], does not bind to dCtBP. This shows that the Hairless-dCtBP interaction strictly depends on the conserved C-terminal part of Hairless that contains the dCtBP binding site. Furthermore, a small C-terminal peptide, H[1052-1076], is sufficient to bind dCtBP. Finally, a specific interaction between Hairless and dCtBP is also observed in vitro with a GST pull-down assay. H[1-1076], but not H[1-710] or H[1-1061], is efficiently retained by a GST-dCtBP fusion protein. This in vitro interaction indicates that the Hairless-dCtBP interaction is likely to be direct. It is concluded that the conserved C-terminal part of Hairless contains a motif necessary and sufficient to bind dCtBP (Morel, 2001).

To test the functional significance of this binding site, an in vivo assay was used. The expression of a truncated version of Hairless that does not bind dCtBP, H[1-1061], does not lead to an increased density of sense organs and does not rescue the loss of Hairless function. Thus, the last 15 amino acids of Hairless are required for the activity of the protein. Interestingly, flies overexpressing both H[1-1061] and Su(H) display a wild-type phenotype. This shows that H[1-1061] is unable to cooperate with Su(H) to block Notch signaling. Nevertheless, H[1-1061] expression suppresses the loss-of-bristle phenotype associated with increased levels of Su(H). Since H[1-1061] binds Su(H), it is possible that H[1-1061] proteins form nonproductive complexes with Su(H). Accordingly, the residual activity of the mutant HRP1 protein, which carries a 68 amino acid C-terminal deletion, might result from its ability to sequester Su(H) without actively repressing transcription. These results therefore suggest that Hairless requires the binding of dCtBP to repress the expression of Notch target genes (Morel, 2001).

In summary, these findings indicate that Hairless links Su(H) to the dCtBP corepressor. It is therefore proposed that Hairless antagonizes Notch signaling activity by recruiting dCtBP to repress Notch target gene expression. The activation of the Notch receptor would then lead to a competition between NICD and Hairless to assemble DNA-bound regulatory complexes of opposite activities (Morel, 2001).

Hairless-mediated repression of Notch target genes requires the combined activity of Groucho and CtBP corepressors

Notch signal transduction centers on a conserved DNA-binding protein called Suppressor of Hairless [Su(H)] in Drosophila species. In the absence of Notch activation, target genes are repressed by Su(H) acting in conjunction with a partner, Hairless, which contains binding motifs for two global corepressors, CtBP and Groucho (Gro). Usually these corepressors are thought to act via different mechanisms; complexed with other transcriptional regulators, they function independently and/or redundantly. This study investigated the requirement for Gro and CtBP in Hairless-mediated repression. Unexpectedly, it was found that mutations inactivating one or the other binding motif can have detrimental effects on Hairless similar to those of mutations that inactivate both motifs. These results argue that recruitment of one or the other corepressor is not sufficient to confer repression in the context of the Hairless-Su(H) complex; Gro and CtBP need to function in combination. In addition, this study demonstrates that Hairless has a second mode of repression that antagonizes Notch intracellular domain and is independent of Gro or CtBP binding (Nagel, 2005).

To test the repressive effects of Hairless in the absence of NICD, Hairless ability to inhibit transcription in the presence of Grainyhead (Grh) was tested. The Notch response (NRE) reporter contains binding sites for the transcriptional activator Grh that stimulate transcription fourfold in the absence of NICD and increase the stimulation seen in the presence of NICD. Addition of full-length Hairless inhibits these effects, reducing transcription in the presence of Grh alone by 50%. Furthermore, this inhibitory effect is dependent on Su(H), as indicated by a lack of repression of HDeltaS, and requires both CtBP and Gro, since Hairless proteins with either interaction domain mutated (HDeltaC, H*C, HDeltaG, H*G) lose most of their repressive activity. Again, the levels of activity with the single mutants are similar to the levels seen with the double-mutant forms of the protein (HDeltaGC, H*GC) and all resulted in >90% of the expression seen with Grh. These experiments suggest that Hairless has two modes of repression, one that operates by repressing the transcriptional machinery through its recruitment of global corepressors and a second that operates by directly antagonizing NICD (Nagel, 2005).

These data confirm therefore that both Gro and CtBP can function as corepressors with Hairless, and indeed both factors are necessary for full repression by Hairless on the NRE; preventing the interaction with one or the other factor severely compromises Hairless activity. This is in apparent contrast to the effects on vgBE-LacZ, for which only Gro appears essential. Furthermore, the two cofactors appear to act together, since Hairless proteins lacking both interaction motifs retains a level of repression that is comparable to the results seen upon removing either alone (Nagel, 2005).

Previous studies of CtBP and Gro have argued that they mediate repression in qualitatively different ways, although both are thought to recruit histone deacetylases. Gro has predominantly been associated with so-called long-range repression, as it operates to dominantly silence modular enhancers. In contrast, CtBP appears to act in a local way to inhibit activators that are bound nearby. However, these models do not appear compatible with a combined requirement for Gro and CtBP in Hairless-mediated repression. Furthermore, direct fusion of a Gro interaction domain to the Su(H) protein is sufficient to convert it into a potent repressor, as described for other transcriptional regulators. Why should Gro and CtBP therefore be interdependent in the context of Hairless recruitment? One simple explanation would be that one or the other corepressor is needed to specifically counteract NICD activation. For example, CtBP interferes with recruitment of p300, a histone acetyltransferase that is reported to interact with mammalian NICD. However, the data suggest that CtBP and Gro are both needed to repress Grh even in the absence of NICD, arguing that each corepressor can only perform a subset of its functions in the context of Hairless. Maybe the two corepressors recruit different enzymatic activities that are needed together to promote repression. If the Hairless complex were incompatible with oligomerization of Gro, which is reported to be important for stable repression, Gro might be able to recruit histone deacetylases but not to promote spreading of the repression complex. And if CtBP, which in mammals has been found complexed with methyl transferases as well as deacetylases, could recruit only histone methyl transferases, the corepressors would each confer a critical component on the Hairless complex. A more complete understanding of the molecular functions of Gro and CtBP in the context of chromatin dynamics and transcription complexes will be needed to determine why Hairless requires their coordinate activities in many developmental scenarios, as has been shown in this study (Nagel, 2005).

Protein Interactions: Supressor of Hairless and Mastermind

During signaling by the Notch receptor, Notch's intracellular domain is cleaved, moves to the nucleus and associates with a DNA-binding protein of the CSL class (CSL for CBF1, Suppressor of Hairless [Su(H)], LAG-1); as a result, target genes are transcriptionally activated. In C. elegans, a glutamine-rich protein called LAG-3 forms a ternary complex with the Notch intracellular domain and LAG-1 and appears to serve as a transcriptional activator that is critical for signaling. Although database searches have failed to identify a LAG-3-related protein, it has been surmised that Notch signaling in other organisms might involve an analogous activity. To search for a LAG-3-like activity in mice, a modified yeast two-hybrid screen was used similar to that used to identify LAG-3. Briefly, a complex bait was used to screen a library of mouse cDNAs fused to the Gal4 activation domain (Clontech). That bait included mouse CBF1 fused to the Gal4 DNA-binding domain (GD) as well as the intracellular domain of mouse Notch1. The bait proteins were co-expressed from a pBridge vector. Out of 6 million transformants, one positive with similarity to Drosophila Mastermind and human KIAA0200 was recovered. A focus was placed on this clone because Drosophila Mastermind is known to be critical for Notch signaling. The murine ortholog of Mastermind is called mMam1, and the human one hMam1. The mMam1 fragment recovered in the two-hybrid screen consists of 62 amino acids and included a conserved region present in both fly and human Mastermind proteins (Petcherski, 2000).

To explore the idea that Mastermind might have a role similar to LAG-3 in Notch signaling, a series of two-hybrid assays was conducted. mMam1 binds mCBF1-GD in the presence of either Notch1 or Notch3, but not in their absence. It was next asked whether Drosophila Mastermind might participate in a similar complex in flies. A fusion protein was used carrying the Gal4 activation domain and the amino-terminal 198 amino acids of fly Mastermind (dMam [1-198]; henceforth called dMam), which includes the conserved region of Mastermind that is critical for complex formation among mouse components. dMam was found to bind Su(H) strongly in the presence of the fly Notch intracellular domain, but not in its absence (Petcherski, 2000).

The interchangeability of proteins from different species was examined. Remarkably, the fly protein, dMam, interacts with murine Notch1 or Notch3 and murine CBF1, and mMam1 interacts with fly Notch and Su(H). In contrast, C. elegans LAG-3 does not form a complex with either murine or fly components, and mMam and dMam do not complex with worm components. It is concluded that both fly and murine Mastermind proteins form a ternary complex with either fly or murine receptors and CSL proteins. This interchangeability underscores the similarity between the fly and murine Notch pathways. Although murine Mastermind is not described, a full-length cDNA sequence for human Mastermind is available. Comparison of human and fly Mastermind sequences reveals only one short region of significant similarity that is limited to 60 amino acids at the amino terminus. Therefore, despite a low overall sequence similarity between mouse and Drosophila Mastermind proteins, the region crucial for complex formation is conserved (Petcherski, 2000).

The importance of Notch's ankyrin repeats for complex formation was examined. In C. elegans, formation of the ternary complex is dependent on the ankyrin repeats of the Notch-related receptor GLP-1. To ask whether the same situation holds for the murine complex, two missense mutants, M1 and M2, were used, each of which bears amino-acid substitutions in the fourth ankyrin repeat of mNotch1. Consistent with results in C. elegans, both M1 and M2 compromise interactions among Notch1, CBF1 and either mMam1 or dMam (Petcherski, 2000).

What is the role of Mastermind in Notch signaling? Previous studies have suggested a role in transcriptional control. In Drosophila, Mastermind is a nuclear protein and is bound to chromatin. Furthermore, in Drosophila, Mastermind acts downstream of Notch in signaling. The amino-acid sequences of both human and fly Mastermind proteins are rich in glutamine and proline, a common feature in transcriptional activators. In the work reported here, a physical link between Mastermind and the major CSL transcription factor of the Notch pathway is described. The interaction of both mMam and dMam with the Notch intracellular domain and CBF1 relies on the receptor's ankyrin repeats. These repeats are essential for Notch signaling and the transcriptional response. In C. elegans, point mutations in the ankyrin repeats severely compromise signaling by the Notch-related receptor GLP-1. In tissue culture cells, the M1 and M2 point mutations abolish receptor function and compromise the activation of transcription by Notch signaling. The simplest explanation for all these findings is that Mastermind functions as a transcriptional activator for Notch signaling (Petcherski, 2000).

Important parallels exist between LAG-3 in C. elegans and Mastermind in Drosophila and mammals. (1) All of these proteins form a ternary complex with an intracellular fragment of Notch and a CSL DNA-binding protein. (2) Mutations in the fourth ankyrin repeat of the receptor compromise ternary complex formation for C. elegans and mouse proteins, as is reported here. (3) All three proteins are rich in glutamine and proline: 27.6% in LAG-3, 29.4% in dMam and 22% in hMam1. (4) LAG-3 and Mastermind function downstream of Notch in C. elegans and Drosophila, respectively. It is proposed that LAG-3 and Mastermind perform analogous functions as activators for Notch (Petcherski, 2000).

What is the evolutionary relationship between LAG-3 and Mastermind? An intriguing idea is that LAG-3 and Mastermind share a common ancestor. The conservation in amino-acid sequence between Mastermind orthologs is much lower than is found for other components of the pathway: whereas hMam1 and dMam share similarity only in a stretch of 60 amino acids within a much larger protein, Notch and CSL proteins show high similarity (44.8% and 74.5% identity for hNotch1/dNotch and hCBF1/Su[H], respectively) over most of their length between these same species. It therefore seems plausible that the absence of similarity between LAG-3 and Mastermind may reflect a high rate of amino-acid substitution in these proteins rather than a distinct evolutionary origin (Petcherski, 2000).

Inducible RNA interference uncovers the Drosophila protein Bx42 as an essential nuclear cofactor involved in Notch signal transduction

The UAS/GAL4 two component system was used to induce mRNA interference (mRNAi) during Drosophila development. In the adult eye the expression from white transgenes or the resident white locus is significantly repressed by the induction of UAS-wRNAi using different GAL4 expressing strains. By induced RNAi it was demonstrated that the conserved nuclear protein Bx42 is essential for the development of many tissues. Phenotypically the effects of Bx42 RNAi resemble those obtained for certain classes of Notch mutants, pointing to an involvement of Bx42 in the Notch signal transduction pathway. The wing phenotype following overexpression of Suppressor of Hairless is strongly enhanced by simultaneous Bx42 RNAi induction in the same tissue. Target genes of Notch signaling like cut and Enhancer of split m8 were suppressed by induction of Bx42 RNAi (Negeri, 2002).

Phenotypically, the consequences of Bx42RNAi often resemble effects obtained by interference with components of the Notch pathway. Studies of protein interaction in vitro suggest an involvement of Bx42 and its human homolog Skip in the Notch signal transduction. Both Skip and Bx42 were found to interact with Notch-IC, CBF1 and components of the CBF1 corepressor complex like SMRT, N-CoR, CIR, Sin3a and HDAC2 proteins (Zhou, 2000a; Zhou, 2000b; Zhang, 2001). By yeast two hybrid interaction and coimmunoprecipitation it was found that Bx42 physically interacts with the Drosophila CBF1-homolog Su(H) and with its antagonist Hairless, for which so far no vertebrate counterpart is known. The current study presents evidence that these interactions are biologically meaningful. Ubiquitous early induction of Bx42 RNAi results in embryos with dorsal cuticle only, a phenotype similar to Notch mutations. Induction of Bx42 RNAi in the eye disc results in an eye to antenna transformation as is observed following overexpression of dominant negative forms of Notch in the same tissue. Both effects could be interpreted that Bx42 normally functions as a coactivator of the Notch pathway. However, Other studies have demonstrated that overexpression of Notch-IC in the eye-antennal discs results in the formation of ectopic antennae too, but only if the eyeless function is reduced by a hypomorphic mutation. eyeless is one of the master regulators for eye development and functions in a cross-regulatory circuitry together with six other master regulatory genes. One of them is dachshund, whose human homolog Ski is a known interactor for the Bx42 human homolog Skip (Dahl, 1998). Thus, the observed eye antennal transformation by Bx42 RNAi could also be interpreted as a downregulation of eye master regulatory genes via diminished dachshund activity and a simultaneous derepression of the Notch pathway. Negative interference of Bx42 with Notch signaling is consistent with the results of ptc-GAL4 driven induction of Bx42 RNAi. A similar loss of scutellar bristles is observed on ptc-GAL4-driven overexpression of the Notch-ankyrin repeats, a part of Notch-IC involved in active signal transductio. However, this is not fully understood, since overexpression of Notch extracellular domains missing certain EGF repeats results in a similar antineurogenic phenotype. Moreover, Notch mutant clones result in a loss of bristles as well. A negative role of Bx42 in Notch signaling is suggested by the wing phenotype of dpp-GAL4/Bx42RNAi. Following reduction of Bx42 the veins are thinner or missing consistent with a Notch gain of function (Negeri, 2002).

Ectopic expression of Su(H) prevents sensory organ development in a similar manner to activated Notch. This may reflect an excess of lateral inhibition or it may be due to interference with the establishment of the correct fates in the progeny of the sensory organ precursor cells. The failure of sensory organ formation in the scutellum and the wing following local Bx42 RNAi may be related to a gain of Su(H) function. The enhancement of the Su(H) overexpression phenotype by Bx42 RNAi and the similar effects of Bx42 RNAi and Su(H) overexpression on Notch target genes strongly support this argument and suggest a functional relationship between both proteins (Negeri, 2002).

Although the data suggest a negative role, it is not believed that Bx42 protein acts as a repressor within the Notch pathway. Earlier work, which suggested an activation function for Su(H) was at odds with data demonstrating a repressive role for its mammalian homolog CBF1. More recent work provides evidence for Su(H) acting as a switch between repression and activation of Notch target genes. It is proposed that Bx42 contributes to the switch provided by the Su(H) protein. How this is accomplished can only be speculated at the moment. By its direct interaction with Su(H) Bx42 may stabilize a switching complex. By its direct interaction Bx42 could recruit Hairless into the complex contributing to its repressive function. Modification or removal of Bx42 protein (as by RNAi) would result in a destabilization of this repressive complex allowing to switch to the active state. Similarly, due to squelching by a large excess of overexpressed Su(H), only a reduced amount of Bx42 protein would be available to stabilize the repression complex. It is important to emphasize that Bx42 is able to physically interact with Notch-IC (Zhou, 2000a) and may act as a switching protein in the recruitment of activators as well (Negeri, 2002).

An important role in Notch signaling and functional relation between Su(H) and Bx42 were also suggested when the effects of reducing Bx42 on Notch target gene expression were studied. Both cut and E(spl)m8 were suppressed by Bx42 RNAi as was the vestigial quadrant element (vgQE) enhancer, indicating a Bx42 activating function for these genes in wild type. wingless expression was not affected under these conditions, excluding Bx42 RNAi induced cell death as an explanation for the observed effects. The effects on Notch target gene expression are consistent with the proposed role for Bx42 as part of a switch. Following Su(H) overexpression repression of cut, E(spl)m8 and the vgQE enhancer was demonstrated. wingless, on the other hand, was not affected. The similar effects of Su(H) overexpression and Bx42 RNAi on Notch target gene expression underscores the close relationship in the function of both proteins (Negeri, 2002).

Besides its involvement in Notch signaling the observed phenotypic effects of Bx42 RNAi suggest that the Bx42 protein is involved in other signaling pathways as well. Data from its vertebrate homolog, which indicate that Bx42 takes part in nuclear receptor pathways, are supported by observations on chromosomal binding to sites occupied by the ecdysone receptor complex. A possible interaction with Dachshund, one of the master regulators in eye development, which is widely expressed in the nervous system, has already been mentioned. It remains to be established how Bx42 is involved in these other pathways (Negeri, 2002).

Supressor of Hairless, Strawberry notch and Ebi

The Notch and Epidermal growth factor receptor (Egfr) pathways both regulate proliferation and differentiation, and the cellular response to each is often influenced by the other. A mechanism is described that links them in a sequential fashion, in the developing compound eye of Drosophila. Egfr activation induces photoreceptor (R cell) differentiation and promotes R cell expression of Delta. This Notch ligand then induces neighboring cells to become nonneuronal cone cells. ebi and strawberry notch (sno) regulate Egfr-dependent Delta transcription by antagonizing a repressor function of Suppressor of Hairless [Su(H)]. Sno binds to Su(H), and Ebi, an F-box/WD40 protein, forms a complex with Su(H) and the corepressor Smrter. Egfr-activated transcriptional derepression requires ebi and sno, is proteasome-dependent, and correlates with the translocation of Smrter to the cytoplasm (Tsuda, 2002).

The Notch signaling pathway plays multiple roles in eye development. At the morphogenetic furrow, the proneural protein Atonal facilitates the expression of Dl in the R8 cell. The first step of ommatidial assembly involves lateral inhibition between equivalent cells, but successive steps are inductive, arising from an already differentiated cell to its uncommitted neighbors. The Notch pathway is involved in the regulation of both of these processes. Similarly, the Egfr ligand, Spi, expressed in R8, activates the receptor in neighbors allowing them to assume their respective R1–R7 cell fates. Subsequently, these R cells express Spi, and as described in this study, they also express Dl in response to Egfr activation. The cone cells receive an Egfr signal and a Notch signal from the R cells and this combination is critical for the assumption of their fate. Later, after their fate is determined, these cone cells, too, will express Delta, which is important for pigment cell induction. Presumably, the level of the Egfr signal rises in the cone cells with time, and as a threshold of Egfr activation is surpassed, the proteasome mediated arm of the pathway becomes effective causing derepression of Su(H) and expression of functional levels of Dl sufficient for pigment cell development. Thus, a temporally and spatially positioned combination of parallel and sequential Egfr/Notch signals is important for the successive induction of cell types in the eye (Tsuda, 2002).

An interesting interplay between Egfr and Notch pathways is also seen during vulval induction in C. elegans. Cells that are close to the anchor cell assume the primary developmental fate, while those farther away become secondary cells. The development of the secondary cell fate shows many similarities with cone cell development. Both secondary and cone cells primarily require high levels of Notch signal and a low-level activation of the Egfr signaling pathway. Genetic studies support one of two alternative models for the development of the secondary cell fate. In the first model, the graded activation of Egfr (Let23) mediated by the expression of its ligand Lin3 in the anchor cell and lateral Notch (Lin12) signaling imparts a secondary cell fate. Alternatively, the signal mediated by Lin3 is required for the specification of the primary cell, which in turn induces secondary cells through the Notch pathway. The latter model is similar to the sequential activation mechanism describe in this study for cone cell development. It will be interesting to determine if in C. elegans, the Egfr (Let23) pathway activates an as yet unidentified Notch (Lin12) ligand in primary cells that is then used to induce secondary cell fate (Tsuda, 2002 and references therein).

Evidence from mammalian systems has suggested that CBF1, the mammalian homolog of Su(H), is a component of a large repressor complex. The activation function of CBF1 results from a displacement of repressive components (such as HDAC) by the intracellular domain of Notch which converts Su(H) into a transcriptional activator. Genetic analysis of the embryonic midline and the pupal bristle complexes in Drosophila have also supported a switch from Su(H)-mediated repression to activation. A second mechanism for relieving Su(H) mediated repression is through Sno, Ebi, and the Egfr pathway. In response to the Egfr signal, Ebi, an F-box protein, presumably causes a proteasome-mediated degradation of an unknown component of the Su(H) inhibitory complex. Mammalian TBL1 (Ebi) can function downstream of the tumor suppressor gene, p53, in the degradation of the ß-catenin protein in a novel ubiquitin-dependent degradation pathway involving Siah, the mammalian homolog of the Drosophila Sina protein. Similarly, Drosophila Ebi can also act in combination with Sina to degrade protein targets. More generally, phosphorylation by MAPK downstream of RTK pathways is known to trigger proteasome-mediated degradation of target proteins. In addition to Ebi, a core component of the proteasome, encoded by l(3)73Ai gene, is also important for expression of Dl. The simplest model is that in response to Egfr signaling, one or more of the many components in the large Su(H)/SMRTER repression complex becomes a target of a proteasome-mediated degradation process (Tsuda, 2002).

The studies presented here also show that the corepressor SMRTER is redistributed from the nucleus to the cytoplasm in an Egfr/Sno/Ebi dependent manner. These results are in complete agreement with the role of the corresponding mammalian protein SMRT in its function as a repressor. Like Su(H), nuclear hormone receptors such as retinoic acid receptor and thyroid hormone receptor can function as both repressors and activators. SMRT has been shown to be phosphorylated in response to an RTK signal. This leads to translocation of SMRT out of the nucleus. Thus, steroid hormone receptors lose their ability to repress but not activate transcription. In an in vivo example, the Egfr/Sno/Ebi pathway promotes the dissociation of the Su(H)/SMRTER repressor complex and causes the nuclear export of SMRTER. As a result, target genes such as Dl are derepressed (Tsuda, 2002).

Notch signaling can take place between cells that are equivalent at the time the signal initiates, or it can occur between a signaling cell that is different from the cell receiving the signal. Traditionally, the first kind of process has been referred to as lateral inhibitory Notch signaling and the second as an inductive Notch pathway. These studies suggest that the fundamental difference between these two processes is not due to differences in molecular components of the pathway downstream from activated Notch, but rather due to the mechanism that controls the expression of the ligand, Dl. For lateral inhibitory Notch pathways, a mechanism involving a feedback loop and proneural genes is at the core of Dl/Notch regulation. Starting with an equipotent group, an asymmetric signaling system is created, in which the signaling cell expressing high levels of Dl, assumes a differentiated fate and prevents its neighbors from adopting an identical fate. All available evidence suggests that the Egfr pathway, Sno, and Ebi do not control Dl expression in such lateral inhibitory processes mediated by Notch. In contrast, this study shows that in inductive processes controlled by Notch signaling, Dl expression is controlled by Egfr, Ebi, and Sno and apparently not by proneural genes. For example, no known proneural gene (Ac/sc, amos, or atonal) is expressed in R cells that contact the cone cells (i.e., R1-R7) and express Dl. This is also true for cells at the dorsoventral boundary of the wing disc where Notch signaling directly activates vestigial expression through Su(H) binding to the enhancer and in the mesectodermal cells of stage 6 embryos where the Notch pathway has been implicated in controlling the expression of single minded at the midline. Instead, all of these cells in the eye, wing, and embryo receive an Egfr signal that likely controls Dl expression. Indeed, the late expression of Dl in R cells does not involve feedback from the cone cells but instead involves the derepression of Dl expression in a Notch-independent manner. This is different from the early expression of Dl that is required for the selection of R8 cells at the furrow through a lateral inhibitory signal (Tsuda, 2002).

This study highlights the function of two unusual proteins, Sno and Ebi, in controlling the expression of Dl. Mammalian Ebi (TBL1) interacts with a SMRT/HDAC complex as also supported by this study in Drosophila. There are two human and three mouse genes similar to Sno identified by genome projects. The function of the mammalian Sno proteins is unknown. Whether the mammalian proteins also function upstream of the Notch pathway, as they do in Drosophila, remains to be established. Given the conservation of developmental pathways between Drosophila and mammals, this may not be an unreasonable expectation (Tsuda, 2002).

A DNA transcription code for cell-specific gene activation by notch signaling

Cell-specific gene regulation is often controlled by specific combinations of DNA binding sites in target enhancers or promoters. A key question is whether these sites are randomly arranged or if there is an organizational pattern or 'architecture' within such regulatory modules. During Notch signaling in Drosophila proneural clusters, cell-specific activation of certain Notch target genes is known to require transcriptional synergy between the Notch intracellular domain (NICD) complexed with CSL proteins bound to 'S' DNA sites and proneural bHLH activator proteins bound to nearby 'A' DNA sites. Previous studies have implied that arbitrary combinations of S and A DNA binding sites (an 'S+A' transcription code) can mediate the Notch-proneural transcriptional synergy. By contrast, this study shows that the Notch-proneural transcriptional synergy critically requires a particular DNA site architecture ('SPS'), which consists of a pair of specifically-oriented S binding sites. Native and synthetic promoter analysis shows that the SPS architecture in combination with proneural A sites creates a minimal DNA regulatory code, 'SPS+A', that is both sufficient and critical for mediating the Notch-proneural synergy. Transgenic Drosophila analysis confirms the SPS orientation requirement during Notch signaling in proneural clusters. Evidence that CSL interacts directly with the proneural Daughterless protein, thus providing a molecular mechanism for this synergy. It is concluded that the SPS architecture functions to mediate or enable the Notch-proneural transcriptional synergy which drives Notch target gene activation in specific cells. Thus, SPS+A is an architectural DNA transcription code that programs a cell-specific pattern of gene expression (Cave, 2005).

The functional significance of the SPS element has not been determined, but initially, it was proposed that the arrangement of the S binding sites in the SPS may function to mediate cooperative DNA binding by CSL proteins, or it may be necessary for the recruitment of other proteins to the promoter. Subsequent studies, though, showed that CSL, NICD, and Mam "ternary complexes" can assemble on single S sites. To date, no studies have experimentally addressed whether there are significant functional differences between SPS elements and single S or other non-SPS binding site configurations, and the mechanistic function of the SPS element is not known (Cave, 2005).

In Drosophila, five of the seven bHLH repressor genes in the E(spl)-Complex contain an SPS element in their promoter regions, and four of these bHLH R genes contain both SPS and proneural bHLH A protein binding (A) sites. These four bHLH R genes (the m7, m8, , and genes, collectively referred to as the 'SPS+A bHLH R' genes have been shown genetically to depend upon proneural bHLH A genes for expression. In addition, transcription assays in Drosophila cells with at least two of these four genes (m8 and ) have shown that there is strong transcriptional synergy when NICD and proneural proteins are expressed in combination. These SPS+A bHLH R genes also have similar patterns of cell-specific expression within proneural clusters. Following determination of the neural precursor cell from within a proneural cluster of cells, Notch-mediated lateral inhibition is initiated and these SPS+A bHLH R genes are specifically upregulated in all of the nonprecursor cells but not in the precursor cell. The absence of NICD, and the presence of specific repressor proteins such as Senseless, prevent upregulation of SPS+A bHLH R genes in the precursor cells (Cave, 2005).

This study shows that there are important functional differences between the SPS architecture and non-SPS configurations of S binding sites. The SPS architecture is critical for synergistic activation of the m8 SPS+A bHLH R gene by Notch pathway and proneural proteins. Whereas previous studies have focused on which regulatory genes and proteins function combinatorially to activate SPS+A bHLH R gene expression, this study focuses on the underlying DNA transcription code that programs the Notch-proneural transcriptional synergy that drives cell-specific gene transcription. The results of previous studies have implied that an apparently arbitrary combination of S and A binding sites (S+A transcription code) is sufficient for transcriptional activation of SPS+A bHLH R genes. By contrast, this study shows that a minimal transcription code, SPS+A, is sufficient and critical for mediating Notch-proneural synergistic activation of these genes. The SPS+A code is composed of the specific SPS binding site architecture in combination with proneural A binding sites. Furthermore, evidence is presented that direct physical interactions between the Drosophila Su(H) and Daughterless protein mediate the transcriptional synergy, thus providing a molecular mechanism for the Notch-proneural synergy. Together, these studies show that the SPS architecture functions to mediate or enable the transcriptional synergy between Notch pathway and proneural proteins and that SPS+A is an architectural transcription code sufficient for cell-specific target gene activation during Notch signaling (Cave, 2005).

To test whether the SPS binding site architecture is important for Notch-proneural synergy, the ability of Drosophila NICD (dNICD) and proneural bHLH A proteins, such as Achaete and Daughterless (Ac/Da) to synergistically activate the wild-type native m8 promoter and SPS architecture variants was examined. Whereas the native m8 promoter carries the wild-type SPS architecture of S binding sites, the m8 promoter variants contain either a disrupted S site, leaving a single functional S site (SF-X or X-SR), or orientation variants in which the orientation of one or both S sites have been reversed (SR-SF, SF- SF, and SR-SR) (Cave, 2005).

The native m8 promoter is synergistically activated in transcription assays by coexpression of dNICD and Ac/Da, but it is only weakly activated by expression of dNICD or proneural Ac/Da proteins alone. However, neither promoter with a single S binding site (SF-X or X-SR) can mediate synergistic interactions between dNICD and proneural proteins. In fact, both single S site promoters are only weakly activated when proneural and dNICD proteins are expressed individually or together. Thus, single S sites are not sufficient to mediate Notch-proneural synergy in these contexts, even though they are in the same position as the SPS in the wild-type m8 promoter (Cave, 2005).

When the number of S binding sites are maintained, but the orientation of these sites within the SPS is varied (SR-SF, SF-SF, and SR-SR), only the wild-type (SF-SR) SPS orientation is synergistically activated by coexpression of dNICD and proneural Ac/Da proteins. Thus, the wild-type SPS architecture of S binding sites is clearly necessary for the m8 promoter to mediate transcriptional synergy between NICD and the proneural protein complexes assembled on the SPS and A sites, respectively (Cave, 2005).

The transcriptional synergy between NICD and proneural proteins mediated by the SPS element is crucial for the coactivation by the Mastermind (Mam) protein. Coexpression of Mam with both dNICD and proneural proteins provides a strong coactivation of transcription of the wild-type m8 promoter. However, this strong coactivation is not observed with any of the non-wild-type m8 SPS variants, which also cannot mediate Notch-proneural synergy. Thus, coactivation by both the NICD and Mam cofactors is strongly dependent on synergistic interactions with proneural combinatorial cofactors, and the specific SPS architecture is critical for mediating this synergy (Cave, 2005).

The native m8 promoter studies tested whether the organization of the S binding sites in the SPS are necessary to mediate the Notch-proneural synergy. In order to test which of these architectural features are sufficient to mediate that synergy, a set of synthetic promoters was created carrying the same SPS variants mentioned above in combination with A sites (SPS-4A reporter). These synthetic promoters thus contain the sites predicted to mediate the synergy but lack the other sites present in the native m8 promoter, which might also be necessary. This reductionist approach allows for the identification of a minimal promoter that contains only those sites that are necessary and sufficient to mediate the Notch-proneural synergy. All of these synthetic reporters are modestly activated by expression of proneural proteins alone, but expression of dNICD alone gives no activation. By contrast, only the SPS-4A reporter containing the wild-type SPS (SF-SR) mediates clear synergistic activation when dNICD and proneural proteins are coexpressed, and none of the SPS variants do so (Cave, 2005).

Given that functional CSL/NICD/Mam ternary complexes have been shown to assemble on single S sites and activate transcription, it was expected that promoters with single S sites could be activated at low levels by expression of dNICD in the absence of the proneural proteins and that promoters with two S sites might have more activity than single S sites. However, it was surprising to observe that all of the m8 and synthetic promoters, even with the wild-type SPS element, have very low or no activity when dNICD is expressed alone. Thus, the SPS binding site architecture does not appear to facilitate recruitment of functional NICD coactivator. This argues against previous proposals that suggested that the SPS architecture might function to recruit other proteins to the promoter. Thus, given that the wild-type SPS architecture is necessary and sufficient for Notch-proneural synergy, these results indicate that the function of the SPS element is to enable synergistic interactions with proneural proteins (Cave, 2005).

The synthetic promoters do not carry bHLH R sites, which are present in all E(spl)-C gene promoters. Thus, these sites clearly are not necessary for Notch-proneural synergy, although they may modulate it in vivo. It has been proposed that other repressor proteins bind the and SPS+A bHLH R gene promoters to restrict their expression to a subset of proneural clusters. Although these hypothetical repressor binding sites may be necessary to program the full and gene expression pattern, the current results indicate that they are not necessary for the Notch-proneural synergy that drives nonprecursor cell-specific upregulation (Cave, 2005).

Both the m8 and SPS-4A synthetic reporter contain a hexamer sequence that has been coconserved with the SPS element. Elimination of that hexamer site in a synthetic promoter does not disrupt Notch-proneural, suggesting that Notch-proneural synergy in vivo is not dependent on the hexamer site (Cave, 2005).

Together, the synthetic and m8 promoter results indicate that SPS+A is a minimal transcription code that is both necessary and sufficient for Notch-proneural synergy in Drosophila. The results with the promoters that were tested show that Notch-proneural transcriptional synergy requires the specific organization or architecture of the SPS element, in addition to its combination with proneural A binding sites. All of the promoters with SPS variants failed to mediate this synergy. This clearly indicates that arbitrary combinations of S and A binding sites are not sufficient to mediate Notch-proneural synergy (Cave, 2005).

An important question is whether there are other DNA binding transcription factors that can combinatorially synergize with CSL/NICD transcription complexes. Previous studies have shown that Notch pathway factors can synergize with a nonproneural transcription factor, Grainyhead, suggesting that synergy with the CSL/NICD transcription complexes could be very general or nonspecific. To test whether a general coactivator, the VP16 transcription activation domain, can synergistically interact with dNICD, an essentially identical wild-type SPS-containing synthetic promoter was created in which the A sites were replaced by UAS binding sites for the yeast Gal4 transcription factor (SPS-5U). Expression of a fusion protein containing the Gal4 DNA binding domain and the constitutively active VP16 activation domain can activate the synthetic SPS-5U promoter. However, the Gal4-VP16 fusion protein does not synergize with NICD. Thus, CSL/NICD complexes do not synergize with every nearby DNA bound transcription factor, and there is at least some specificity to the synergy with bHLH A proteins. This interaction specificity could contribute significantly to selective activation of Notch target genes. Further studies will be required to determine whether other DNA binding transcription factors can combinatorially synergize with Notch signaling and whether such factors fall into distinct classes (Cave, 2005).

Given that Notch signaling and neural bHLH A proteins have been conserved between Drosophila and mammals, it was next asked whether the transcriptional synergy between these proteins is also conserved in mammalian cells. Using the same set of synthetic promoters as mentioned above, activation following expression of the mammalian NICD and neural bHLH A protein homologs (Notch-1 ICD [mNICD] and MASH1/E47, respectively) was tested in murine NIH 3T3 cells. As in the Drosophila system, expression of MASH1/E47 proteins alone produces modest activation of the wild-type (SF-SR) SPS-4A promoter, and mNICD alone does not produce any significant activation of the promoter. However, clear transcriptional synergy is observed with the wild-type SPS promoter when both mNICD and neural bHLH A proteins are coexpressed. Moreover, SPS-mediated synergy requires nearly the same organizational features of S binding sites as observed in Drosophila. Neither of the single S site promoters can mediate that synergy, nor can most of the orientation variants. Although the SR-SR promoter is activated following coexpression of both the mNICD and bHLH A proteins, it is not activated by mNICD alone (Cave, 2005).

These results indicate that the potential for transcriptional synergy between NICD and neural bHLH A proteins has been conserved in a mammalian cell system and that the SPS+A code is sufficient and critical for mediating that transcriptional synergy. This raises the possibility that there may be mammalian genes that are regulated by neural bHLH A proteins and Notch signaling via this code. Although there is an SPS element conserved in the HES-1 promoter, HES-1 does not have an A site in its proximal promoter region, and HES-1 is not activated by expression of bHLH A genes. Thus, HES-1 appears to be similar to the Drosophila E(spl)-C m3 bHLH R gene, which also has an SPS but no obvious nearby A site. Whole-genome searches are being performed for genes in mammalian systems that may be regulated by the SPS+A code (Cave, 2005).

It has been proposed that the architecture of the SPS element may mediate cooperative binding of a second CSL protein once an initial CSL protein binds the DNA. Using electromobility gel shift assays to test for cooperative binding, the ability was compared of bacterially expressed and partially purified Drosophila Su(H) protein to bind DNA probes containing either the wild-type m8 SPS or an m8 SPS with one S site mutated. If there is cooperativity, one would expect to observe the band corresponding to two DNA bound CSL proteins to be as strong or stronger than the band corresponding to a single CSL protein bound to DNA. The single S site probe serves as a control because it cannot be cooperatively bound by two Su(H) proteins, and it also serves to identify the band corresponding to a single Su(H) protein bound to the wild-type SPS probe. Similar amounts of Su(H) protein bind strongly to the wild-type probe and to the single-site probe. In particular, because single protein binding to the wild-type DNA probe did not facilitate or stabilize simultaneous binding of two S proteins, Su(H) does not appear to bind cooperatively to the two S sites in the wild-type probe. These results suggest that CSL proteins do not bind cooperatively to the SPS in vivo, although posttranslational modifications in vivo could affect these binding properties Cave, 2005).

In addition, the protein binding affinity for the SF-SR and SR-SF probes appears to be comparable, although the reversed orientation of the two S sites would have likely disrupted cooperative binding if it were present. This result strongly suggests that the complete lack of activation by SR-SF sites in all of the promoters tested is not due simply to decreased ability of Su(H) protein to bind to the SR-SF orientation variant Cave, 2005).

To test the in vivo relevance of the conserved S binding site orientation in SPS elements, transgenic flies were created carrying β-galactosidase reporter genes driven by native m8 promoters containing either the wild-type (SF-SR) or SR-SF variant SPS elements. Wing and eye imaginal discs containing m8 promoters with the wild-type SPS element produced strong expression in proneural cluster regions, similar to the pattern described for endogenous m8. By contrast, comparably stained wing and eye discs carrying the m8 promoter reporters with the SR-SF SPS variant showed no expression or very low levels of expression, respectively. Extended staining of discs containing the SR-SF element revealed clear but weak expression in a pattern of single cells that resembles the distribution of neural precursors in the wing discs and eye discs. This is likely due to activation via the A site by proneural proteins because proneural levels are highest in the precursor cells. However, there was no expression in the surrounding nonprecursor cells within the proneural clusters even though Notch signaling is activated in these cells. Similar neural precursor-specific m8 reporter expression patterns have been observed when the S binding sites are eliminated, indicating that reversal of the S binding site orientations is functionally equivalent to eliminating them for this aspect of Notch target gene expression. These in vivo results confirm that the conserved orientation of the S binding sites in the wild-type SPS element is essential for nonprecursor cell specific upregulation of the SPS+A bHLH R m8 genes in response to Notch signaling in proneural clusters (Cave, 2005).

To gain an insight into the molecular mechanism underlying the strong transcriptional synergy between Notch signaling and bHLH A proteins on the m8 and SPS-4A promoters, whether this synergy involves a direct physical interaction was tested by using yeast two-hybrid assays with the Drosophila proteins. These experiments revealed that the Daughterless N-terminal domain directly and specifically interacts with the Su(H) protein in the absence of the bHLH domain and C terminus (Cave, 2005).

Using transcription assays in Drosophila cells, whether the Da N terminus (DaN construct), which contains a transcription activation domain, can synergistically activate the m8 promoter was tested in the absence of both its bHLH DNA binding domain and a heterodimerization partner, like Ac. The Da N-terminal protein synergistically activates the m8 promoter when dNICD is coexpressed, apparently by direct binding of the DaN protein to endogenous CSL bound to the SPS element. These results indicate that the Notch-proneural transcriptional synergy is not mediated by cooperative DNA binding interactions between the Su(H) and proneural proteins, although such cooperative binding may mediate transcriptional synergy between some combinatorial cofactors. These results suggest that a direct interaction between Su(H) and the Da N-terminal fragment, which can occur independent of NICD, facilitates the formation of an active transcription complex when NICD is also present during Notch signaling (Cave, 2005).

These results suggest that the SPS architecture functions to enable a direct physical interaction between Su(H) and Da proteins, thus providing a molecular mechanism for the observed Notch-proneural synergy that is mediated by the SPS element. This interaction could stabilize the recruitment or functional activity of NICD, which then recruits Mam, and could explain the strong dependence of both NICD and Mam coactivation functions on the presence of proneural proteins (Cave, 2005).

In previous studies, it has been proposed that neither the synergistic activation nor the transcriptional repression mediated by CSL protein complexes imply direct interactions between CSL and DNA bound combinatorial cofactors; rather, it is likely that CSL proteins exert their effects through the recruitment of non-DNA binding cofactors, such as chromatin modifying enzymes. While this might be the case for some Notch target gene promoters, in the case of m8, the results indicate that the mechanism underlying the synergistic interactions between CSL/NICD and bHLH A proteins does involve direct physical interactions (Cave, 2005).

A mechanistic model is proposed for programming Notch-proneural synergy with the SPS+A transcription code. These studies demonstrate that there are important functional differences between SPS and non-SPS organizations of S binding sites. The critical role of the SPS binding site architecture is not predicted or explained by the previous models for Notch target gene transcription. Previous models suggest that transcription is promoted by the binding of NICD to CSL, which displaces CSL bound corepressors, thus allowing transcriptional synergy with other DNA bound combinatorial cofactors. These models have not distinguished between Notch target genes with regulatory modules that contain SPS or non-SPS configurations of S binding sites, nor do they explain or predict the critical function of the SPS binding site architecture in mediating Notch-proneural transcriptional synergy (Cave, 2005).

A revised model is proposed that incorporates the essential requirement for the specific SPS binding site architecture in combination with the proneural A binding sites for transcriptional activation of m8 and the other SPS+A bHLH R genes. These genes each contain an SPS+A module and exhibit similar cell-specific upregulation in nonprecursor cells in proneural clusters. In this new model, the specific architecture of the S sites in the SPS element directs the oriented binding of Su(H) so that it is in the proper orientation and/or conformation to enable a direct interaction with Da. This interaction is an essential prerequisite for subsequent recruitment and/or functional coactivation by NICD during Notch signaling. This Notch-proneural complex is then further activated by subsequent recruitment of Mam (Cave, 2005).

It is interesting to note that the mammalian homologs of each of the Su(H), NICD, and Da proteins have been shown to interact with the p300 coactivator; thus, when complexed together, these proteins could potentially function combinatorially to recruit p300 or a related coactivator (Cave, 2005).

In Drosophila and mammals, Notch signaling is used throughout development to activate many different target genes, and in multiple developmental pathways. Thus, it is of paramount importance that the proper target genes are selectively activated in the proper cell-specific patterns. It is known that Notch signaling can activate genes through non-SPS configurations of S sites in certain other target genes. For example, expression of the Drosophila genes single minded, Su(H), and vestigal have all been shown to be regulated by Notch signaling, and all have single S sites or multiple unpaired S sites but no SPS elements in their promoter and/or enhancer regions (Cave, 2005).

The results show that for essentially every promoter tested, NICD cannot activate in the absence of neural bHLH A combinatorial cofactors, suggesting that NICD may always require a combinatorial cofactor to activate target genes. If so, the non-SPS Notch target genes are likely also to have specific combinatorial cofactors. The results also clearly show that the Notch-proneural combinatorial synergy requires a specific configuration of S sites, the SPS. There may be other specific configurations of S binding sites that mediate synergy for different classes of combinatorial cofactors for Notch signaling (Cave, 2005).

Together, these observations suggest that specific, but unknown, non-SPS configurations of sites may program the interactions between Notch complexes and the proper combinatorial cofactors. It is speculated that these non-SPS configurations might be unique to each target gene, or it is possible that there are specific patterns or classes of S binding site configurations -- an 'S binding site subcode' -- that determine cofactor specificity. Thus, the results suggest that selective Notch target gene activation may be programmed by distinct Notch transcription codes in which specific configurations of S binding sites mediate selective interactions with specific combinatorial cofactors (Cave, 2005).

Elucidating the various transcription codes controlling target gene activation during Notch signaling will be an important goal for future studies. The results have clearly shown that the architecture of transcription factor binding sites can be crucial for control of cell-specific Notch target gene activation. The studies presented here give a glimpse into the molecular mechanisms by which a one dimensional pattern of DNA binding sites can program cell-specific patterns of gene expression (Cave, 2005).

Gene-specific targeting of the histone chaperone asf1 to mediate silencing: Asf1 can be coprecipitated with Su(H), the corepressor Hairless, and Groucho

The histone chaperone Asf1 assists in chromatin assembly and remodeling during replication, transcription activation, and gene silencing. However, it has been unclear to what extent Asf1 could be targeted to specific loci via interactions with sequence-specific DNA-binding proteins. This study shows that Asf1 contributes to the repression of Notch target genes, as depletion of Asf1 in cells by RNAi causes derepression of the E(spl) Notch-inducible genes. Conversely, overexpression of Asf1 in vivo results in decreased expression of target genes and produces phenotypes that are strongly modified (enhanced and suppressed) by mutations affecting the Notch pathway, but not by mutations in other signaling pathways. Asf1 can be coprecipitated with the DNA-binding protein Su(H) and the corepressor Hairless and interacts directly with two components of this complex, Hairless and SKIP. Thus, in addition to playing more general roles in chromatin dynamics, Asf1 is directed via interactions with sequence-specific complexes to mediate silencing of specific target genes (Goodfellow, 2007).

Modulation of the chromatin structure is a key feature in transcriptional regulation. Chromatin remodeling by ATP-dependent enzymes and posttranslational histone modifications are two important mechanisms that affect transcriptional activity, by influencing the accessibility of upstream regions and promoters. A third mechanism involves the breakdown and reassembly of nucleosomes on the DNA, a process that also allows for the incorporation of histone variants, such as H3.3. Histone chaperones, which bind to histone heterodimers, are required both for nucleosome assembly and for their disassembly. They include the H3/H4 chaperone Anti-silencing factor 1 (Asf1), which has roles in replication-dependent and replication-independent chromatin dynamics (Goodfellow, 2007).

In yeast, extensive Asf1-mediated exchange of histones that is independent of replication and of transcription has been detected at gene promoters and is likely to be highly significant in maintaining the balance between induction and silencing of genes. Indeed, there are now several examples of yeast Asf1 contributing to chromatin disassembly at promoters to facilitate binding of the RNA-polymerase complex. Conversely, Asf1 also plays important roles in gene silencing when the reassembly of nucleosomes accompanies transcriptional repression. For example, in the absence of Asf1, there is a delay in promoter closure at the PHO5 gene. However, it remains unclear whether Asf1-mediated nucleosome reassembly occurs via a targeted mechanism, involving sequence-specific DNA-binding proteins, or whether it occurs constitutively by default (Goodfellow, 2007).

A strong correlation between histone loss and gene activation has emerged from genome-wide studies in Drosophila, as it has in yeast, suggesting that transcription in higher eukaryotes is also likely to be regulated by histone loss and replacement at the promoter. However, thus far, the contribution of Asf1 to dynamic gene regulation during cell signaling in multicellular organisms has not been examined. One cell-signaling pathway with very direct effects on transcription is the highly conserved Notch pathway. Activation of the receptor results in the release of a nuclear-targeted intracellular fragment (Nicd), which binds directly to the CSL DNA-binding protein (Suppressor of Hairless, Su(H), in Drosophila) and recruits the coactivator Mastermind, resulting in the activation of target genes. CSL proteins also contribute to the silencing of target genes in the absence of Nicd, through adaptor-mediated recruitment of corepressors such as Groucho (Gro), CtBP, and SMRT. Previous analysis indicates that the activity of Notch target genes correlates with a reduction in histone H3 density, suggesting that nucleosome disassembly and reassembly is likely to be involved in their regulation, and prompting an investigatation of whether Asf1 could play a role (Goodfellow, 2007).

This study shows that Asf1 contributes to the repression of Notch target genes, and that it is recruited to the DNA through interactions with the Su(H)/H complex. Thus, Asf1 is targeted to specific loci by binding to sequence-specific DNA-binding complexes, where it can promote gene silencing during development (Goodfellow, 2007).

To investigate whether Asf1 contributes to the regulation of inducible genes in Drosophila, RNA interference (RNAi) was used to deplete S2-N cells and the levels of transcription were analyzed from the 11 well-characterized Notch target genes clustered in the E(spl) complex. Conditions were established for activating Notch in these cells and it was shown that activation results in Su(H)-dependent stimulation of E(spl) gene transcription (Goodfellow, 2007).

Unlike knockdown of the other chromatin regulators tested, depletion of Asf1 led to a 4-fold increase in E(spl)m7 mRNA levels, but it had no effect on the housekeeping genes rp49 and EF2B. More extensive analysis revealed that mRNA levels for all E(spl) genes were increased after Asf1 depletion in the absence of Notch activation; some showed a greater than 10-fold change in expression, suggesting that these Notch targets are derepressed as they are when the corepressor Hairless is depleted. In contrast, there was little effect of Asf1 depletion on several other repressed genes, including a phagocytosis receptor gene, nimrod. In addition to the derepression observed in resting cells, Asf1 depletion also altered the responsiveness to Notch activation. Many more of the E(spl) genes were susceptible to Notch activation in Asf1-depleted cells; for example, 5 of the 11 genes were expressed at greater than 20-fold higher levels after Asf1 RNAi. There was comparatively little change at the genes, such as E(spl)m3, which normally has the most robust response to Notch and is depleted for histones. Thus, it appears that Asf1 makes important contributions to the silencing of Notch target genes (Goodfellow, 2007).

Previous studies showed that overexpression of Asf1 in the Drosophila eye (ey::Gal4 UAS::asf1/+) causes a 'small-eye' phenotype in which the eye is reduced in size and ommatidia are disorganized. If these small-eye phenotypes are a consequence of Asf1 altering the transcription of Notch targets, they may be modified when combined with mutations in the Notch pathway. To investigate this possibility, flies overexpressing Asf1 were crossed to alleles affecting genes central to Notch or to other signaling pathways, and the eye size was analyzed in the heterozygous progeny (Goodfellow, 2007).

The first dramatic result was that the heterozygous combination of a Notch loss-of-function allele (N55e11) and Asf1 overexpression caused a severe reduction in the eye/head capsule ('pin-head') and resulted in lethality. Thus, the effects of Asf1 overexpression were strongly enhanced by a decrease in Notch function. Significant enhancement of the Asf1 phenotype also occurred with Delta loss-of-function alleles, but not with alleles affecting Hedgehog (smo), EGF-R (Egfr), or Wingless (arm, arrow) pathways or with alleles affecting the SET domain protein Trithorax-related (trr), the histone exchange factor Domino (dom), or the cell adhesion protein Pawn (pwn). Complementary results were obtained by using mutant alleles that increase Notch signaling: both a loss-of-function Hairless (H) allele and a gain-of-function Notch allele (NMcd1) suppressed the small-eye defect caused by Asf1 overexpression. These findings are fully consistent with the results of RNAi-mediated Asf1 depletion, and they suggest that Asf1 is involved in repression of Notch target genes. As asf1 mutant cells failed to proliferate, it was not possible to obtain clones of homozygous mutant cells to test the effects of eliminating Asf1 on Notch target genes in the eye (Goodfellow, 2007).

To investigate whether interactions between Notch and Asf1 occur in other tissues, it was asked whether Asf1 overexpression also perturbed Notch function in the Drosophila wing. Expression of Asf1 in the developing wing pouch (sd::Gal4/+; UAS::asf1/+) resulted in margin loss/wing nicks and mild vein thickening, characteristics of reduced Notch function (Notch/+ heterozygous flies have mild wing nicks due to reduced signaling at the dorsal/ventral (d/v) organizer of the wing). The Asf1 overexpression phenotypes were strongly enhanced when the levels of Notch were reduced; thus, wings had extensive scalloping/margin loss and more extensive vein thickening. Wing phenotypes, similar to the eye phenotypes, produced by Asf1 expression were thus enhanced by reduced Notch (Goodfellow, 2007).

To further assess whether Asf1 affects expression of target genes regulated by Notch (e.g., cut) or by other pathways (e.g., spalt), the effects of overexpressing Asf1 in wing discs was analyzed. In wild-type discs, Notch-dependent expression of Cut is detected in a stripe along the d/v boundary. This was interrupted and reduced in discs in which Asf1 was overexpressed. In contrast, there was no visible effect on Spalt under these conditions. Similar results were obtained when Asf1 was expressed in a more limited domain (by using ptc::Gal4), where a local loss of Cut, but not Spalt, expression was seen. Stronger expression of Asf1 resulted in more pronounced Notch-like phenotypes and loss of Cut expression, which could be rescued by a reduction in Hairless function. Under these conditions, where Asf1 was expressed more strongly, some more generalized effects of Asf1 were sometimes detected, compatible with its proposed role as a histone chaperone during replication. The replication defects became more severe at even higher levels of expression (29°C). Similarly, clones of cells mutant for asf1 failed to proliferate. Thus, as in yeast, Asf1 appears to have roles in replication-dependent as well as replication-independent chromatin dynamics in Drosophila. By moderating the levels of Asf1 expression, it was possible to uncouple these requirements, revealing a contribution to repression of Notch target genes (Goodfellow, 2007).

Complexes implicated in repression at Notch targets are formed by the CSL/Su(H) DNA-binding protein in conjunction with adaptor proteins, such as SKIP and Hairless, which recruit general corepressors, including SMTR or Gro and CtBP. On polytene chromosomes from Drosophila salivary glands, Asf1 is detected at most Su(H)-enriched sites, suggesting that these proteins are present at the same loci. Asf1 is also bound at many other loci, and it is strongly enriched at centromeres and telomeres, reflecting its multiple roles in chromatin dynamics (Goodfellow, 2007).

The colocalization of Su(H) and Asf1 on polytene chromosomes prompted a test of whether Su(H) and/or associated factors could copurify with Asf1 in immunoprecipitation (IP) experiments. For these experiments, extracts prepared from Drosophila embryos, and Su(H) or Asf1 was immunoprecipitated by using moderate salt conditions. Under these conditions, Asf1 was detected in Su(H) IP experiments, and, conversely, Su(H) was precipitated with Asf1, as was the corepressor Gro, but not CtBP. To exclude the possibility that the interaction between Asf1 and the Su(H) complex was mediated by the independent binding of both protein complexes to DNA, IP experiments were performed in the presence of ethidium bromide (EtBr), a DNA-intercalating drug that dissociates proteins from DNA. This treatment did not affect the interaction of Asf1 with Su(H). Thus, these data suggest that Asf1 is present in protein complexes containing the sequence-specific DNA-binding protein Su(H) and the Gro corepressor. A significant suppression of the Asf1-induced small-eye phenotype was observed in flies that were also heterozygous for a strong gro allele (groE48) and an enhancement was seen by Hairless proteins that retained a Gro-binding domain, agreeing with a model linking Gro to Asf1-mediated repression. Therefore whether any of the proteins in the Su(H) repression complex are able to bind to bacterially produced Asf1 (fused to glutathione S-transferase, GST), was examined. Of those tested, both Hairless and the adaptor protein SKIP were bound to GST-Asf1, but not to GST alone or to GST-CAF1p55 (a component of chromatin assembly factor 1). Neither Gro nor Su(H) itself showed direct interactions with Asf1 in this assay (Goodfellow, 2007).

Finally, to test whether Hairless contributes to the recruitment of Asf1 in vivo, chromatin immunoprecipitation (ChIP) was performed with anti-Asf1 antibodies in cells with and without RNAi-mediated depletion of Hairless and association with two E(spl) genes, m3 and m7, was assayed. The E(spl)m7 gene is silenced in the S2 cells and is strongly affected by Asf1 depletion, whereas E(spl)m3 is expressed in S2 cells, is highly induced by Notch activation, and is more mildly affected by Asf1 depletion. Of the two genes, the greatest effects were seen for E(spl)m7; binding of Asf1 to both enhancer and ORF fragments strongly decreased in ChIP after Hairless depletion. A decrease was also seen at the E(spl)m3 ORF region, but not at the E(spl)m3 enhancer. This enhancer is found to have very low histone coverage in these cells, and it was found that it shows only small Asf1 occupancy levels. The decrease in Asf1 from ORFs of both E(spl)m3 and E(spl)m7 after Hairless depletion may indicate that Asf1 spreads from the site of recruitment. Binding of Asf1 to E(spl)m7 and E(spl)m3 regions was confirmed by using affinity-purified anti-Asf1 antibodies raised in a different species. Loss of Hairless does not affect the binding of Asf1 to other loci that do not require Su(H)/H for their regulation, such as eiger or snRNP69D. Similarly, there was no change in the levels of Polycomb protein associated with bxd-PRE after Hairless knockdown. Together, these data support the model that recruitment of Asf1 to Notch targets requires Hairless (Goodfellow, 2007).

The density and precise positioning of nucleosomes are important factors in determining the transcriptional activity of a locus. It is now evident that most nonnucleosomal histones in cells are likely to be complexed with chaperones. It is therefore not surprising that the histone chaperone Asf1 is important for chromatin dynamics and has been shown to have multiple roles in transcription as well as in the disassembly and reassembly of chromatin during replication. These include gene-specific roles in repression, activation, and transcription elongation. For example, Asf1 is required for nucleosome disassembly and transcription activation at the yeast PHO5, PHO8, ADY2, and ADH2 promoters. However, the mechanisms responsible for targeting Asf1 to these loci remain unclear. This study has demonstrated that Asf1 can be specifically recruited to target loci by interactions with sequence-specific DNA-binding transcription factors. Asf1 is present in a complex with Su(H), the central DNA-binding protein in the Notch pathway, and that it interacts directly with two proteins found in CSL complexes, Hairless and SKIP. Importantly, it was found that Asf1 plays a significant role in the repression of Notch target genes. Thus, contrary to effects at many of the inducible loci examined in yeast, these data demonstrate a requirement for Asf1 in silencing rather than in activation of these inducible genes (Goodfellow, 2007).

As the global corepressor Gro is also coprecipitated with Asf1 and is implicated in Asf1-mediated repression through genetic interactions, Gro and Asf1 may cooperate in the repression of Notch target genes. Gro has been postulated to exert long-range repressive effects by nucleating a transcriptionally silent chromatin state, in a similar manner to its yeast relative Tup1. For example, at the STE6 locus, Tup1 recruitment results in increased nucleosomal density and local nucleosome positioning. The recruitment of the histone chaperone Asf1 with Gro to Su(H)/H DNA-binding complexes could facilitate a similar localized increase in histone deposition and participate in the spreading of repressed chromatin. Furthermore, since (H)/H complexes engage in comparatively low-stability interactions with target loci, it is suggested that Asf1 could be critical for translating these transient interactions into stable silencing. However, thus far, the analysis has focused on relatively few targets and tissues; thus, it remains to be determined whether Asf1 is recruited to all targets regulated by Su(H)/H, or whether there are additional factors that influence its recruitment at specific loci. Similarly, it will be important to determine whether other sequence-specific complexes are able to bind directly to Asf1 (Goodfellow, 2007).

In conclusion, these results show that the histone H3/H4 chaperone Asf1 contributes to selective silencing of genes in Drosophila, through interactions with the Su(H)/H DNA-binding protein complexes. In this way, chaperones can act as gene-selective regulators that contribute to the control of gene expression by developmental signaling pathways (Goodfellow, 2007).

Histone chaperones ASF1 and NAP1 differentially modulate removal of active histone marks by LID-RPD3 complexes during NOTCH silencing

Histone chaperones are involved in a variety of chromatin transactions. By a proteomics survey, the interaction networks of histone chaperones ASF1 (Anti-silencing factor 1), CAF1, HIRA, and NAP1 were identified. This study analyzed the cooperation of H3/H4 chaperone ASF1 and H2A/H2B chaperone NAP1 with two closely related silencing complexes: LAF and RLAF. NAP1 binds RPD3 and LID-associated factors (RLAF) comprising histone deacetylase RPD3, histone H3K4 demethylase LID/KDM5, SIN3A, PF1, EMSY, and MRG15. ASF1 binds LAF, a similar complex lacking RPD3. ASF1 and NAP1 link, respectively, LAF and RLAF to the DNA-binding Su(H)/Hairless complex, which targets the E(spl) Notch-regulated genes. ASF1 facilitates gene-selective removal of the H3K4me3 mark by LAF but has no effect on H3 deacetylation. NAP1 directs high nucleosome density near E(spl) control elements and mediates both H3 deacetylation and H3K4me3 demethylation by RLAF. It is concluded that histone chaperones ASF1 and NAP1 differentially modulate local chromatin structure during gene-selective silencing (Moshkin, 2009).

Regulated modulation of the chromatin structure is essential for the transmission, maintenance, and expression of the eukaryotic genome. The combined actions of ATP-dependent chromatin-remodeling factors (remodelers), histone chaperones, and histone-modifying enzymes drive chromatin dynamics. Histones are subjected to a wide range of reversible posttranslational modifications, including acetylation, phosphorylation, methylation, and ubiquitylation. Histone modifications, in turn, can promote the recruitment of selective regulatory factors and modulate chromatin accessibility. Chromatin remodelers control DNA accessibility by mediating nucleosome mobilization either through sliding or by nucleosome (dis)assembly (Moshkin, 2009).

Whereas originally considered mainly as mere chaperones, it has become clear that histone chaperones play diverse roles during chromatin transactions. Histone chaperones bind selective histones and include the highly conserved H3/H4 chaperones ASF1, CAF1, HIRA, and Spt6 and the H2A/H2B chaperones NAP1, Nucleoplasmin, and FACT. Although their biochemical activity, binding and release of histones, appears rather mundane, in conjunction with other factors, histone chaperones participate in a variety of chromatin transactions and other cellular tasks. For example, yeast NAP1 participates in an extensive interaction network including a diverse set of transcription initiation/elongation factors, chromatin remodelers, RNA-processing factors, cell-cycle regulators, and other proteins (Moshkin, 2009).

ASF1 is one of the major H3/H4 chaperones, and through association with other proteins, it contributes to diverse chromatin transactions. (1) In conjunction with CAF1 and the MCM2-7 DNA helicase, ASF1 participates in replication-coupled chromatin assembly. (2) When associated with HIRA, ASF1 participates in replication-independent chromatin assembly and histone replacement. (3) DNA-repair-associated chromatin assembly requires the cooperation between ASF1 and the H3K56 acetyltransferase Rtt109. (4) ASF1 functionally cooperates with the Drosophila BRM chromatin remodeler, and (5) interaction of ASF1 with transcription activators stimulates histone eviction from promoter areas and facilitates recruitment of chromatin-specific coactivator complexes. (6) ASF1 itself is one of the targets of Tousled-like kinase (TLK), which controls cell-cycle progression and chromatin dynamics. (7) Finally, ASF1 is involved in developmental gene expression control by mediating transcriptional repression of Notch target genes. ASF1 is recruited to E(spl) genes by the sequence-specific DNA-binding protein Su(H) and its associated corepressor complex, harboring Hairless (H) and SKIP (Moshkin, 2009).

Notch is the central component of a highly conserved developmental signaling pathway that is present in all metazoans. Notch is a single-pass transmembrane protein that is activated through ligand binding, resulting in the release of the Notch intracellular domain (Nicd), which is targeted to the nucleus to activate gene expression. The CSL (CBF1, Su(H), and Lag1) family of sequence-specific DNA-binding proteins is the key targeting factor of Nicd and coactivators and, in the absence of Nicd, corepressors. The repression of Notch target genes involves multiple chromatin-modifying activities including histone deacetylases, H3K9 methyltransferases, CtBP, NcoR/SMRT, and Goucho (GRO). In the absence of the Nicd, loss of ASF1 leads to derepression of the E(spl) genes, revealing its essential role in silencing (Moshkin, 2009).

The molecular mechanism by which ASF1 achieves gene-specific transcription repression and the potential roles of other histone chaperones in developmental gene regulation remains largely unknown. To address these issues, a proteomics survey was performed of the protein interaction networks of ASF1, CAF1, HIRA, and NAP1 in Drosophila embryos. This analysis revealed that ASF1 and NAP1 interact with two related but distinct corepressor complexes: LAF and RLAF. LAF, comprising LID/KDM5 SIN3A, PF1, EMSY, and MRG15, associates with ASF1 (forming LAF-A). RLAF, comprising LAF plus RPD3, interacts with NAP1 (forming RLAF-N). Through a combination of biochemistry and developmental genetics, it was established that LAF-A and RLAF-N are tethered to Notch target genes by the Su(H)/H complex and mediate gene-selective silencing. Both ASF1 and NAP1 are required for the targeted removal of the positive H3K4me3 mark by facilitating LID/KDM5 recruitment to chromatin. Furthermore, NAP1 mediates nucleosome assembly at regulatory elements of Notch target genes and histone deacetylation by RLAF. These results uncover extensive crosstalk between distinct histone chaperones and histone-modifying enzymes in developmental gene regulation (Moshkin, 2009).

These results emphasize that, rather than generic, redundant factors, histone chaperones play highly specialized roles in gene-specific regulation. This study has dissected the molecular mechanism underpinning coordinate silencing of Notch target genes by the histone H3/H4 chaperone ASF1 and the H2A/H2B chaperone NAP1. ASF1 interacts with LAF, comprising SIN3A, PF1, EMSY, MRG15, and the histone H3K4me2/3 demethylase LID/KDM5, forming LAF-A. A closely related complex, RLAF that includes the deacetylase RPD3, does not bind ASF1. Instead, RLAF associates with NAP1, forming RLAF-N. The chaperones ASF1 and NAP1 link, respectively, LAF and RLAF to the Su(H)/H DNA-binding complex, tethering them to the E(spl) genes. Both ASF1 and NAP1 bind the SKIP subunit of the Su(H)/H complex (Goodfellow, 2007). Thus, at least in part, ASF1 and NAP1 facilitate H3K4me3 demethylation activity at the E(spl) genes through LID recruitment. Other LAFs might provide additional links to the Su(H)/H complex by contacting GRO and CtBP, which themselves associate with the Su(H)/H complex. For example, mammalian PF1, MRG15, and SIN3A have been reported to bind GRO. This study identified CtBP in LID, PF1, and NAP1 immunopurifications, providing an additional contact between the Su(H)/H complex and (R)LAF (Moshkin, 2009).

ASF1 does not bind RLAF and has no effect on histone H3 deacetylation by RPD3. In contrast, NAP1 does associate with RLAF and stimulates both H3K4 demethylation by LID and H3 deacetylation by RPD3. SIN3A had a mild effect, but the other LAF subunits played no apparent role in deacetylation. Finally, NAP1 depletion caused a dramatic loss of histones at the E(spl) regulatory elements, whereas ASF1 depletion had no effect on local histone density (Moshkin, 2009).

ASF1 has been proposed to function in chromatin assembly by acting as a donor that hands off the H3/H4 tetramer to either CAF1 or HIRA (De Koning, 2007). Because LAF-A does not associate with either CAF1 or HIRA, this might explain that ASF1 does not modulate nucleosome density at the E(spl) genes. In conclusion, the H3/H4 chaperone ASF1 mediates silencing of Notch target genes by (1) providing a connection between LAF and the Su(H)/H tether and (2) facilitating H3K4 demethylation by LID. The H2A/H2B chaperone NAP1 participates in E(spl) silencing by (1) linking RLAF to Su(H)/H, (2) facilitating H3K4 demethylation by LID, (3) facilitating H3 deacetylation by RPD3, and (4) directing high nucleosome density at repressed loci. The functioning of the H2A/H2B chaperone NAP1 in demethylation and deacetylation of histone H3 provides an example of trans-histone regulation (Moshkin, 2009).

LID and its interacting factors appear to work in a context-dependent manner. For example, LID facilitates activation of dMYC target genes in a manner independent of its demethylase activity. Suggestively, this study observed a genetic interaction between ASF1 and dMYC, indicating a potential role for LAF-A. Recently, it has been suggested that selective RLAF subunits could interact with a homolog of GATA zinc-finger domain-containing protein 1 to facilitate expression of targets by inhibition of RPD3 activity. In mammalian cells, LID homolog RBP2 and MRG15 have been implicated in transcription elongation by restricting H3K4me3 levels within transcribed regions. Identification of SIN3A as a LAF and RLAF subunit provides a molecular explanation for the recent observation that SIN3A is involved in genome-wide removal of both H3K4 methyl and acetyl marks. Collectively, these findings suggest that LID and RPD3 enzymatic activities can be modulated through association with specific partners. The proteomics analysis of the LID, PF1, and EMSY interaction networks further emphasizes the diverse involvement of LAFs in regulation of chromatin dynamics (Moshkin, 2009).

In conclusion, these results emphasize the close interconnectivity between distinct chromatin transactions and reveal cooperation between histone chaperones and targeted histone modifications during developmental gene control. The proteomic survey of ASF1, CAF1, HIRA, and NAP1 provides a starting point for the functional analysis of the regulatory networks in which these chaperones participate. As illustrated by the analysis of LAF-A and RLAF-N, specific protein-protein associations and gene targeting provide an intricate network of combinatorial gene expression control (Moshkin, 2009).

The Daughterless N-terminus directly mediates synergistic interactions with Notch transcription complexes via the SPS+A DNA transcription code

Cell-specific expression of a subset of Enhancer of split (E(spl)-C) genes in proneural clusters is mediated by synergistic interactions between bHLH A (basic Helix-Loop-Helix Activator) and Notch-signalling transcription complex (NTC) proteins. For a some of these E(spl)-C genes, such as m8, these synergistic interactions are programmed by an "SPS+A" transcription code in the cis-regulatory regions. However, the molecular mechanisms underlying this synergistic interaction between NTCs and proneural bHLH A proteins are not fully understood. Using cell transcription assays, it was shown that the N-terminal region of the Daughterless (Da) bHLH A protein is critical for synergistic interactions with NTCs that activate the E(spl)-C m8 promoter. These assays also show that this interaction is dependent on the specific inverted repeat architecture of Suppressor of Hairless (Su(H)) binding sites in the SPS+A transcription code. Using protein-protein interaction assays, it was shown that two distinct regions within the Da N-terminus make a direct physical interaction with the NTC protein Su(H). Deletion of these interaction domains in Da creates a dominant negative protein that eliminates NTC-bHLH A transcriptional synergy on the m8 promoter. In addition, over-expression of this dominant negative Da protein disrupts Notch-mediated lateral inhibition during mechanosensory bristle neurogenesis in vivo. These findings indicate that direct physical interactions between Da-N and Su(H) are critical for the transcriptional synergy between NTC and bHLH A proteins on the m8 promoter. These results also indicate that the orientation of the Su(H) binding sites in the SPS+A transcription code are critical for programming the interaction between Da-N and Su(H) proteins. Together, these findings provide insight into the molecular mechanisms by which the NTC synergistically interacts with bHLH A proteins to mediate Notch target gene expression in proneural clusters (Cave, 2009).

Histone demethylase KDM5A is an integral part of the core Notch-RBP-J repressor complex

Timely acquisition of cell fates and the elaborate control of growth in numerous organs depend on Notch signaling. Upon ligand binding, the core transcription factor RBP-J activates transcription of Notch target genes. In the absence of signaling, RBP-J switches off target gene expression, assuring the tight spatiotemporal control of the response by a mechanism incompletely understood. This study shows that the histone demethylase KDM5A (Little imaginal discs in Drosophila) is an integral, conserved component of Notch/RBP-J gene silencing. Methylation of histone H3 Lys 4 is dynamically erased and re-established at RBP-J sites upon inhibition and reactivation of Notch signaling. KDM5A interacts physically with RBP-J; this interaction is conserved in Drosophila and is crucial for Notch-induced growth and tumorigenesis responses (Liefke, 2010).

Histone lysine demethylases reversibly remove methyl marks, thus facilitating changes in chromatin formation and transcriptional regulation. Histone demethylases have therefore been proposed as promising therapeutic targets of human diseases, including cancer, that are often associated with aberrant histone methylation. This study identified KDM5A as an enzyme responsible for the removal of H3K4me3 at Notch target genes; KDM5A interacts directly with RBP-J via a domain located between PHD2 and PHD3 domain and its C-terminal PHD3 domain. Interestingly, the PHD3 domain was shown recently to bind to H3K4me3 (Wang, 2009). Although it has been suggested that the Arid domain of KDM5A can bind to a short DNA sequence, CCGCCC, the importance of this finding is challenged by the fact that this sequence is very common in CGIs, yet KDM5A is found only at a small number of genes in ChIP-on-chip experiments. Morover, 11 of these putative KDM5A DNA-binding sites can be found at the CGI of Deltex-1, but only one is present in the Deltex-1 enhancer. Therefore, no correlation exists between the position with a high density of putative KDM5A-binding sites and the dynamically regulated H3K4 trimethylation site at the Deltex-1 gene. Furthermore, in EMSA assays no corroborate binding of KDM5A was seen at the proposed sites. Thus, it is hypothesized that the PHD3 domain of KDM5A binds to H3K4m3 at active promoters, and once the H3K4me3 substrate is demethylated, KDM5A is released (Liefke, 2010). The Polycomb group (PcG) proteins play important roles in maintaining gene silencing during development and adult tissue homeostasis, and recent studies have shown that histone demethylase KDM5A is part of a Polycomb complex. Thus, a KDM5A/Polycomb complex, recruited to Notch target genes, could facilitate the removal of active mark H3K4me3 and the subsequent addition of the repressive H3K27me3 marks. Paradoxically, original genetic analysis of Drosophila lid mutations classified Lid as a member of the Trithorax group of genes. Biochemical experiments suggest that Drosophila Lid affects the HDAC activity of Rpd3 (Lee, 2009), and molecular and genetic data show that Lid facilitates activation of dMYC target genes in a demethylase-independent manner, explaining in part the original classification of lid as a positive transcriptional regulator. However, more recent data and this study support a key role for KDM5A demethylase in the dynamics of gene silencing (Liefke, 2010).

Human KDM5A-containing and Drosophila KDM5A/Lid-containing complexes have been analyzed by several groups. KDM5A is found to be part of an MRG15-containing complex comprising multiple subunits, including Sin3B, HDAC1/2, and RbAp46, and two histone acetyltrasferases, TRRAP and Tip60 (Hayakawa, 2007). Interestingly, similar to the current data, these studies unveiled effects of KDM5A/RBP2 on H3K4 trimethylation away from the TSSs in intergenic regions. Moreover, the current findings on dynamic removal of H3K4me3 associated with changes in acetylated H3K9 also point to a combined action of KDM5A and histone deacetylases. Importantly, Drosophila KDM5A/Lid complexes also contain histone deacetylase activity along with histone chaperones ASF1 and NAP1. In agreement with with the current data, several of these Drosophila corepressors affect Notch target gene expression (Moshkin, 2009). However, although all of these data clearly point to a repressive role of KDM5A/Lid, how these enzymes silence specific genes was unknown (Liefke, 2010).

KDM5A is a member of the KDM5 family, which consists of four proteins (KDM5A-D) in mammals. Particularly, KDM5A is highly expressed in the hematopoietic system. KDM5A (RBP2)-deficient mice appear grossly normal but display a mild hematopoietic phenotype, especially in the myeloid compartment. The relatively mild phenotype of KDM5A mouse knockout might be explained in part by some redundancy between the KDM5 paralogs. Yet, one of the up-regulated genes in the KDM5A knockout microarray is Ifi2004, a Notch target gene, indicating that KDM5 paralogs might play redundant as well as specific roles. Loss of KDM5 orthologs in organisms that encode a single KDM5 gene show more severe phenotypes, underscoring the important role of KDM5 demethylases in development. Thus, mutations in Drosophila KDM5A homolog lid often result in lethality before hatching, some animals show a small optic brain lobe and small imaginal discs, and functional inactivation of the C. elegans KDM5A ortholog, Rbr-2, results in undeveloped vulvas or a multivulval phenotype (Liefke, 2010).

Although the discovery of histone demethylases implicates a reversible state of epigenetic gene silencing, it was unanticipated that these chromatin-modifying enzymes exert pathway-specific effects on gene regulation. The dynamic switch-off (and back on) system for Notch target gene expression used in this study allowed revealing of dynamic changes of histone modifications at Notch target genes. It was found that the H3K4me3 is removed with a half-time of about 4 h after inhibition of the Notch pathway by the γ-secretase inhibito GSI. These 4 h cannot be explained by out-dilution through cell cycling; the cell division time here is 24 h. Modulation of H3K4 methylation has also been observed in other biological systems, such as the circadian variation of the transcription of the albumin D-element-binding protein gene in the mouse liver or the X inactivation in early embryonic development, where loss of H3K4me3 is one of the earliest and most characteristic features of chromosome-wide silencing (Liefke, 2010).

Disruption of Notch signaling results in a reduction of H3K4me3 at RBP-J sites, while reactivation re-establishes H3K4me3 levels. This suggests that switch-off and switch-on of Notch target genes depends on a tightly controlled balance of histone H3 methylation and demethylation. This study further identifies the histone demethylase KDM5A as a fundamental element in the switch-off process. For re-establishing H3K4me3 levels, Notch-IC could recruit an H3K4me3 methyltransferase to RBP-J sites of Notch target genes. In contrast to Drosophila genes, ~70% of mammalian genes possess CpG islands (CGIs), including many Notch target genes like Deltex-1, Hes-1, Hes-5, Nrarp, and Hey-1. Genome-wide studies proposed that H3K4 trimethylation remains very constant at CGI-containing promoters in different cell types. This study showed that H3K4me3 stays stable at the CGIs of Deltex-1 and Hes-1 after switching off Notch, but is regulated at the RBP-J-binding site. This finding shows for the first time that H3K4me3 does not necessarily have to be removed from the entire promoter to facilitate gene silencing. Instead, modulation of H3K4me3 at specific regulator elements could be sufficient to regulate gene expression. It will be of interest if the dynamic versus constant H3K4me3 is a more common feature of CGI-containing promoters. Recent technical advances in analyzing histone modifications genome-wide might help to address this question (Liefke, 2010).

In summary, this study unveils that histone methylation is dynamically regulated by Notch signaling: Inhibition of Notch leads to a reduction of H3K4me3 levels at regulatory RBP-J sites, while reactivation of signaling re-establishes high levels of H3K4me3. These biochemical and in vivo data support a role for the histone H3K4me3 demethylase KDM5A/Lid in facilitating the switch from activation to repression state via Su(H)/RBP-J in both Drosophila and mammals. Thus, the histone lysine demethylase KDM5A/Lid is a crucial factor in the silencing process. With the in vivo evidence of Drosophila lid/KDM5A in Notch-induced tumorigenesis, this study suggests a pathway-specific tumor suppressor role of KDM5A in cancer, and provides the basis for studies in novel strategies to manipulate Notch-mediated carcinogenesis (Liefke, 2010).

Requirements for mediator complex subunits distinguish three classes of notch target genes at the Drosophila wing margin

Spatial and temporal gene regulation relies on a combinatorial code of sequence-specific transcription factors that must be integrated by the general transcriptional machinery. A key link between the two is the mediator complex, which consists of a core complex that reversibly associates with the accessory kinase module. Genes activated by Notch signaling at the dorsal-ventral boundary of the Drosophila wing disc fall into three classes that are affected differently by the loss of kinase module subunits. One class requires all four kinase module subunits for activation, while the others require only Med12 and Med13, either for activation or for repression. These distinctions do not result from different requirements for the Notch coactivator Mastermind or the corepressors Hairless and Groucho. It is proposed that interactions with the kinase module through distinct cofactors allow the DNA-binding protein Suppressor of Hairless to carry out both its activator and repressor functions (Janody, 2011).

Intercellular signaling pathways drive many processes during development. Their activation results in changes in transcription factor activity that lead to the activation or repression of specific target genes. An important goal is to understand the transcriptional regulatory codes that allow the combinations of proteins bound to enhancer elements to direct precise patterns of gene expression. One well-characterized developmental paradigm is the specification of the Drosophila wing margin by Notch signaling. The Notch receptor is specifically activated at the dorsal-ventral boundary of the larval wing imaginal disc, due to the restricted expression of its ligands Delta and Serrate and of the glycosyltransferase Fringe. Notch activation results in expression of the target genes Enhancer of split m8 (E(spl)m8), cut, wingless (wg), and vestigial (vg), the last through a specific enhancer element known as the boundary enhancer (vgBE). Wg signaling then leads to the differentiation of characteristic sensory bristles adjacent to the margin of the adult wing (Janody, 2011).

Upon ligand binding, Notch is cleaved by the γ-secretase complex, and its intracellular domain (Nintra) enters the nucleus, where it interacts with the DNA-binding protein Suppressor of Hairless (Su(H)). In the absence of Notch activation, Su(H) represses target gene expression through interactions with the corepressor Hairless (H), which binds to Groucho (Gro) and C-terminal binding protein (CtBP). Nintra displaces these corepressors from Su(H) and recruits coactivators such as Mastermind (Mam). It has been proposed that only a subset of Notch target genes require Su(H) to recruit coactivators, while others require Notch signaling only to relieve Su(H)-mediated repression, allowing transcription to be activated by other factors. However, the mechanisms by which Su(H) directs both activation and repression are not fully understood (Janody, 2011).

The mediator complex is thought to promote transcriptional activation by recruiting RNA polymerase II (Pol II), the general transcriptional machinery, and the histone acetyltransferase p300 to promoters, and by stimulating transcriptional elongation by Pol II molecules paused downstream of the promoter. The 'head' and 'middle' modules of the core complex bind to Pol II and general transcription factors, while the 'tail' module consists largely of adaptor subunits that bind to sequence-specific transcription factors. This core complex reversibly associates with a fourth 'kinase' module that consists of the four subunits Med12, Med13, Cdk8, and Cyclin C (CycC). Several studies have implicated the kinase module in transcriptional repression, which can be mediated by phosphorylation of Pol II and other factors by Cdk8, by histone methyltransferase recruitment, and by occlusion of the Pol II binding site. However, this module also appears to function in activation in some contexts; for example, it promotes Wnt target gene expression during Drosophila and mouse development, in mammalian cells, and in colon cancer. Although all four subunits have very similar mutant phenotypes in yeast, loss of Med12 or Med13 has more severe effects on Drosophila development than loss of Cdk8 or CycC, suggesting that Med12 and Med13 have evolved additional functions in higher eukaryotes (Janody, 2011).

This study shows that Notch target genes at the wing margin can be divided into three classes based on their requirements for kinase module subunits. An E(spl)m8 reporter requires all four subunits for its activation, cut requires only Med12 and Med13 (known as Kohtalo [Kto] and Skuld [Skd], respectively, in Drosophila) for its activation, and wg and the vgBE enhancer require Med12 and Med13 for their repression in cells close to the wing margin. Because Med12 and Med13 coimmunoprecipitate with Su(H), regulate an artificial reporter driven by Su(H) binding sites, and can be replaced by a VP16 activation domain or a WRPW repression signal fused to Su(H), it is proposed that the kinase module directly regulates Notch target genes. All four Notch target genes fail to be expressed in the absence of Mam and are similarly affected by the loss of Hairless or Gro, suggesting that other more specific cofactors might recruit kinase module subunits to these genes (Janody, 2011).

The kinase module of the mediator complex is conserved throughout eukaryotes, yet its functions in transcription remain poorly understood. In yeast, loss of any of the four subunits has a very similar effect. In Drosophila, however, loss of Med12 or Med13 has more dramatic effects than loss of Cdk8 or CycC. The kinase module was originally thought to be primarily important for transcriptional repression, mediated by the kinase activity of Cdk8. However, Med12 and Med13 appear to directly activate genes regulated by Wnt signaling in Drosophila and mammalian systems, and also play a positive role in gene activation by the Gli3 and Nanog transcription factors. The data presented in this study confirm that Med12 and Med13 have functions distinct from Cdk8 and CycC. In addition, evidence is provided that all four kinase module subunits contribute to the activation of E(spl)m8 (Janody, 2011).

The human Mastermind homologue MAM has been shown to recruit Cdk8 and CycC to promoters of Notch target genes, where Cdk8 phosphorylates the intracellular domain of Notch, leading to its ubiquitination by the Fbw7 ligase and degradation (Fryer, 2004). This mechanism would be expected to reduce Notch target gene expression, consistent with the increase in E(spl)mβ expression seen in clones lacking the Drosophila Fbw7 homologue Archipelago (Nicholson, 2011); thus it cannot explain the positive effects of Cdk8 and CycC on E(spl)m8. A function for Cdk8 and CycC in Notch-mediated activation would be analogous to recent findings showing that Cdk8 phosphorylation of Smad transcription factors and of histone H3 promotes activation. Cdk8 phosphorylation of RNA polymerase II (Pol II) is also important for transcriptional elongation (Janody, 2011).

Of interest, the current data also suggest that Med12 and Med13 are involved in the repression of wg and the vgBE enhancer in the absence of Notch signaling. The kinase module has been proposed to inhibit transcription through steric hindrance of Pol II binding, independently of Cdk8 kinase activity. Removal of this module on the C/EBP promoter is thought to convert the mediator complex to its active form. In contrast, this study find that wg and vgBE require Med12 and Med13 for their repression but not their activation, while cut and E(spl)m8 require Med12 and Med13 only for their activation, arguing that the two functions occur on different promoters. It cannot be ruled out that Med12 and Med13 have only indirect effects on some of the genes examined; however, their physical association with Su(H) and the requirement for Su(H) binding sites for misexpression of an artificial reporter in skd and kto mutant clones are consistent with a direct effect of Med12 and Med13 on the Su(H) complex (Janody, 2011).

Med12 and Med13 are found associated with both active and inactive promoters in genome-wide chromatin immunoprecipitation studies, suggesting that they can have different effects on transcription when bound to distinct interaction partners. Although both are very large proteins, they contain no domains predicted to have enzymatic activity, and may instead act as scaffolds for the assembly of transcriptional complexes (Janody, 2011).

It has been proposed that Notch target genes could be categorized into two classes: permissive genes, for which the primary function of Notch is to relieve repression by the Su(H) complex, and instructive genes, for which Notch plays an essential role in activation by recruiting specific coactivators. These differences presumably depend on the combinatorial code of transcription factors that regulate each promoter. This study shows that vgBE, an enhancer previously placed in the permissive category, as well as wg, require Med12 and Med13 for their repression but not their activation. During eye development, the proneural gene atonal is likewise regulated permissively by Notch, and ectopically expressed in skd or kto mutant clones. Unexpectedly, this study found that Gro, previously thought to be a cofactor through which Hairless mediates repression, is not required for the repression of vgBE or wg. Hairless may repress target genes at the wing margin through CtBP, its other binding partner. Alternatively, Gro may affect the expression of other upstream regulators of wing margin fate, masking its repressive effect on the genes that were examined (Janody, 2011).

It was also show in this study that instructive Notch target genes can be further subdivided into two classes based on their requirement for kinase module subunits; E(spl)m8 requires all four subunits, while cut requires Med12 and Med13, but not Cdk8 and CycC. Cdk8 and CycC may simply increase the ability of the mediator complex to recruit Pol II or promote transcriptional initiation; this model would suggest that E(spl)m8 has a higher activation threshold than cut. Alternatively, Cdk8 and CycC might enhance the function of a transcription factor that is specifically required for the expression of E(spl)m8 but not cut. Good candidates for such factors would be the proneural proteins Achaete or Scute or their partner Daughterless (Janody, 2011).

The mechanism by which the kinase module is recruited to promote the activation of instructive target genes is not yet clear. Although Mam proteins are well-characterized coactivators for Nintra, this study found that Mam is necessary for the activation of both instructive and permissive genes. It may thus have a general function in transcriptional activation, such as recruiting histone acetyltransferases or stabilizing the Notch-Su(H) complex. A coactivator that recruits Med12 and Med13 specifically to instructive target genes to promote activation may remain to be identified. The current results, like recent reports demonstrating that the arrangement of Su(H) binding sites can affect the interactions between Notch and its coactivators, highlight the complexity in the mechanisms through which promoter elements respond to Notch signaling (Janody, 2011).

Molecular analysis of the notch repressor-complex in Drosophila: characterization of potential hairless binding sites on suppressor of hairless

The Notch signalling pathway mediates cell-cell communication in a wide variety of organisms. The major components, as well as the basic mechanisms of Notch signal transduction, are remarkably well conserved among vertebrates and invertebrates. Notch signalling results in transcriptional activation of Notch target genes, which is mediated by an activator complex composed of the DNA binding protein CSL, the intracellular domain of the Notch receptor, and the transcriptional coactivator Mastermind. In the absence of active signalling, CSL represses transcription from Notch target genes by the recruitment of corepressors. The Notch activator complex is extremely well conserved and has been studied in great detail. However, Notch repressor complexes are far less understood. In Drosophila, the CSL protein is termed Suppressor of Hairless [Su(H)]. Su(H) functions as a transcriptional repressor by binding Hairless, the major antagonist of Notch signalling in Drosophila, which in turn recruits two general corepressors--Groucho and C-terminal binding protein CtBP. Recently, the C-terminal domain (CTD) of Su(H) was shown to bind Hairless and a single site in Hairless was identified that is essential for contacting Su(H) (Maier, 2011). This study presents additional biochemical and in vivo studies aimed at mapping the residues in Su(H) that contact Hairless. Focusing on surface exposed residues in the CTD, two sites were identified that affect Hairless binding in biochemical assays. Mutation of these sites neither affects binding to DNA nor to Notch. Subsequently, these Su(H) mutants were found to function normally in cellular and in vivo assays using transgenic flies. However, these experiments rely on Su(H) overexpression, which does not allow for detection of quantitative or subtle differences in activity (Kurth, 2011; full text of article).

Structural and functional analysis of the repressor complex in the Notch signaling pathway of Drosophila melanogaster

CSL is the nuclear effector of the Notch signaling pathway and is required for both repression and activation of transcription from Notch target genes. In the absence of a signal, CSL functions as a transcriptional repressor by interacting with corepressor proteins, such as SHARP, SMRT/NCoR, KyoT2, and CIR. CSL-corepressor interactions function to localize histone deacetylase and histone demethylase activity at Notch target genes, which converts the local chromatin into a condensed, transcriptionally silent state. On pathway activation, the ICN binds CSL, and together with Mam, forms a transcriptionally active ternary complex that ultimately displaces corepressors from CSL and upregulates transcription from Notch target genes. In mammals, a number of corepressors have been shown to interact with the BTD of CSL, similar to ICN, which provides a model in which ICN displaces or outcompetes corepressors for binding to CSL. Thus, there are potentially two modes of repression mediated by corepressors: 1) at the transcription or chromatin level, in which the recruitment of HDAC/HDM-containing complexes by corepressors silences gene expression -- this mode of transcriptional repression is independent of Notch; and 2) at the protein level, by which corepressors and ICN compete for binding to CSL (Maier, 2011).

Although several of the mammalian corepressors have fly orthologues, these molecules do not seem to be generally involved in repressing transcription from Notch target genes in flies. A complex involving the SMRT homologue SMRTER negatively regulates Notch signaling during the specification of a subset of nonneuronal cell types in the developing Drosophila retina. However, mammalian SMRT is believed to contact CBF1 directly, whereas SMRTER does not bind Su(H) on its own. The Drosophila orthologue of SHARP/MINT, termed Spen, which genetically inhibits Notch signaling in the context of eye development, is presumably not a transcriptional repressor of Notch target genes in this process. Recently it was shown that Spen is required for the activation rather than the repression of Notch target genes during the development of hemocytes. Moreover, the region of SHARP/MINT that has been defined to interact with CSL is not conserved in Spen. Although formally SHARP/MINT might act as a functional Hairless analogue in mammals, the role of the structurally related Spen proteins seems largely diverse in different organisms (Maier, 2011).

In flies, the major antagonist of Notch signaling is the transcriptional corepressor Hairless, which is ubiquitously expressed in all tissues. Hairless binds the transcription factor Su(H), as well as the corepressors Groucho and CtBP, which serves to localize the transcriptional repression machinery in the nucleus to Notch target genes, thereby repressing gene expression. Removal of the Groucho and CtBP-binding sites from Hairless does not completely eliminate its activity as a repressor, suggesting that, similar to other corepressors, Hairless might compet.e with ICN for binding Su(H). However, the molecular mechanism by which Su(H) is converted from a repressor to an activator complex is unclear (Maier, 2011).

This work investigated the molecular details of the Notch repressor complex in Drosophila. The analysis was multidisciplinary in nature, using biophysical, biochemical, cellular, and in vivo assays to characterize the protein-protein interface between Hairless and Su(H). Hairless was shown to forms a high-affinity 1:1 complex with Su(H) (~1 nM Kd) but only interacts with the CTD, which is in stark contrast to mammalian CSL-corepressor interactions, which are largely mediated through BTD contacts. Previous ITC binding studies of the mammalian Notch components Notch1 (ICN) and RBP-J showed that the Kd of the ICN/RBP-J complex is ~10 nM, suggesting that Su(H)-Hairless and Su(H)-Notch interactions are likely of comparable affinity (Maier, 2011).

Given the similar affinities of ICN and Hairless for Su(H), the question arises whether ICN and Hairless compete for binding to CSL. On one hand, gel-shift assays with purified protein components showed that ICN can displace Hairless from Su(H) independent of Mastermind. On the other hand, it was shown that residues on Su(H) that are important for Notch ANK and Mam binding to CTD do not affect interactions with Hairless. These data suggest that the ICN- and Hairless-binding sites on Su(H) do not overlap. If the ANK domain of ICN and Hairless are competing for binding to the CTD of Su(H), then there is an additional factor to consider: based on binding studies of the mammalian proteins, the vast majority of the binding energy for the Su(H)-ICN complex comes from the RAM domain interaction with the BTD of Su(H), whereas the isolated CTD-ANK interaction is of very low affinity. This represents at least a 10,000-fold difference in the affinities of ANK and Hairless for CTD, which suggests that the ANK domain of ICN would seem to be a very poor competitor for removing Hairless from Su(H) (Maier, 2011).

How then is ICN able to supplant Hairless from Su(H) in order to activate transcription from Notch target genes? Certainly additional experiments will be required to fully address this question; however, at present the following hypothesis is favored: the binding of ICN to Su(H), that is, the RAM-BTD interaction, results in allosteric changes in Su(H) that decreases its overall affinity for Hairless, thereby making ANK a more effective competitor for CTD. Consistent with this notion, gel-shift experiments showed that Hairless was far less effective at removing ICN from Su(H), even when Hairless was present in vast excess (Maier, 2011).

The studies also analyzed two absolutely conserved residues in Hairless (L235 and F243) for their contribution to binding Su(H). Whereas F243 was dispensable for binding, L235 was absolutely required for binding Su(H) in vitro. Mutation of this site to aspartate abrogated binding but did not change the secondary structure content of Hairless, which suggests that L235 lies at the Su(H)-Hairless interface. Given the conservation of F243 but its dispensability for Su(H) binding, this perhaps suggests that this residue is important for interacting with other nuclear factors. Consistent with in vitro binding results, the Hairless mutant L235D failed to assemble a repressor complex with Su(H) in cellular and in vivo assays in the fly. In fact, the L235D mutant was as deficient in repression as the Hairless deletion mutant ΔNT, which removes residues 232-270, emphasizing the importance of this contact in forming the Su(H)-Hairless complex (Maier, 2011).

In conclusion, the fly Notch repressor complex shows similarities and differences compared with the mammalian complex. Despite the high degree of sequence and likely structural conservation, Su(H) in Drosophila differs from mammalian CBF1/RBP-J in that it has no repressor activity on its own; overexpression of Su(H) in cell culture and in vivo results in a Notch gain-of-function phenotype. It is not until the binding of Hairless that Su(H) is transformed into a repressor. Of interest, this study showed that Hairless bound the mammalian CSL orthologue CBF1 nearly as avidly as Su(H), which suggests that the Hairless-binding site on the CTD has been conserved in mammals. In accordance, a potent repression of Notch transcriptional activity was found in cultured mammalian cells by the Hairless NT construct. This raises the possibility of identifying Hairless homologues in other organisms or potentially other transcriptional coregulators that use the Hairless-binding site on CTD, which may be indicative of an as-yet-unidentified mode in mammals to repress Notch signaling. Nonetheless, detailed knowledge of Su(H)-Hairless interactions can now be used to develop molecules that target Notch transcription complexes and either enforce or disrupt their activity, thereby opening new therapeutic avenues (Maier, 2011). omain-containing protein associates with heterochromatin and represses transcription. J Cell Sci 124: 3149-3163. PubMed ID: 21914818

The BEN domain is a novel sequence-specific DNA-binding domain conserved in neural transcriptional repressors

Drosophila Insensitive (Insv) has been shown to promote sensory organ development and has activity as a nuclear corepressor for the Notch transcription factor Suppressor of Hairless [Su(H)]. Insv lacks domains of known biochemical function but contains a single BEN domain (i.e., a 'BEN-solo' protein). Chromatin immunoprecipitation (ChIP) sequencing (ChIP-seq) analysis confirmed binding of Insensitive to Su(H) target genes in the Enhancer of split gene complex [E(spl)-C]; however, de novo motif analysis revealed a novel site strongly enriched in Insv peaks (TCYAATHRGAA). Binding was validated of endogenous Insv to genomic regions bearing such sites, whose associated genes are enriched for neural functions and are functionally repressed by Insv. Unexpectedly, it was found that the Insv BEN domain binds specifically to this sequence motif and that Insv directly regulates transcription via this motif. The crystal structure of the BEN-DNA target complex was determined, revealing homodimeric binding of the BEN domain and extensive nucleotide contacts via alpha helices and a C-terminal loop. Point mutations in key DNA-contacting residues severely impair DNA binding in vitro and capacity for transcriptional regulation in vivo. DNA-binding and repression activities was further demonstrated by the mammalian neural BEN-solo protein BEND5. Altogether, this study defines novel DNA-binding activity in a conserved family of transcriptional repressors, opening a molecular window on this extensive gene family (Dai, 2013).

This study used molecular genetics and ChIP-seq analysis of the neural regulator Insv to delineate two strategies by which it associates with chromatin to mediate transcriptional repression. First, Insv functions as a corepressor for the transcription factor Su(H). In this mode of action, it is recruited to Notch target genes and can oppose the action of Notch signaling in activating these targets (Duan, 2011). Second, it was found that the Insv BEN domain comprises a novel DNA-binding fold that directly recruits Insv homodimers to hundreds of target genes via a palindromic binding site. These dual activities permit Insv to act as a transcriptional repressor during neurogenesis by Su(H)-dependent and Su(H)-independent mechanisms, endowing it with functional versatility. As a corepressor tethered to DNA via Su(H), it tunes the appropriate transcriptional output of Notch target genes (Duan, 2011). It has long been appreciated that cell fate decisions regulated by the Notch pathway are exquisitely sensitive to the precise level of signaling. This appears to be manifest not only by the direct mechanism by which the Notch receptor transduces an extracellular signal into the nucleus, and is in fact obligately destroyed during this process, but also by the fact that transcriptional activation by Notch is balanced by multiple mechanisms of transcriptional repression actively mediated by Notch transcription factors of the CSL family. In Drosophila, Hairless serves as a major direct corepressor for Su(H), and multiple stages of peripheral nervous system development require Hairless function, including during specification of the SOP and at multiple steps in the SOP lineage. However, misexpression of Insv is sufficient to fully rescue SOP specification in Hairless-null clones and can partially rescue their lineage divisions, attesting to its identity as a genuine Su(H) corepressor (Dai, 2013).

This study shows that Notch target genes comprise only a minor fraction of the totality of endogenous Insv target genes. Nevertheless, one of the most prominent categories of genes that contain Insv-occupied target sites are those involved in nervous system development. Indeed, Insv binding and functional repression of a host of genes that regulate neural development was confirmed, including vestigial, mir-263a, repo, fringe, grainyhead, hamlet, tramtrack, Synaptotagmin1, fne, and chinmo, among others. Direct transcriptional regulation by Insv provides a strategy for it to regulate neurogenesis independently of Notch signaling. In fact, the specific expression of Insv in SOPs (and within the lineage), its requirement for fully stable commitment to the SOP fate, and its capacity to promote ectopic peripheral sensory organs when misexpressed (Duan, 2011) are all consistent with the viewpoint that Insv plays an active role in directing neurogenesis upstream of and/or in parallel to its role in inhibiting Notch signaling (Dai, 2013).

The experimental data extensively support that Insv functions as a transcriptional repressor. Consistent with this, the catalog of validated Insv repression targets includes many genes that oppose neurogenesis (e.g., tramtrack), promote late temporal events in neural lineages (e.g., hamlet, grainyhead, and chinmo), or explicitly promote nonneural cell fates in the nervous system (e.g., the proglial factor repo). These observations support the notion that Insv-mediated repression helps guide proper neural specification and cell fate determination within the neural lineage (Dai, 2013).

It is clear that the BEN consensus is modest at best, with no amino acids actually shared across all instances of this domain. Nevertheless, structural studies draw particular attention to Insv residues that participate in direct contacts with the DNA backbone or that recognize specific DNA base side chains. This knowledge allows identification of certain BEN domains as candidates for Insv-like binding specificity. For example, the BEN domain in the Drosophila Mod(mdg4)-PC (Abhiman, 2008) isoform retains few of the base-specific contacts of Insv, but two other Drosophila BEN-solo domain proteins bear striking homology with Insv across multiple residues that were defined as functionally critical for the DNA-binding activity of Insv (Dai, 2013).

It is predicted that CG9883 and Bsg25A are good candidates for harboring sequence-specific DNA-binding activity. Mammalian genomes encode many BEN domain proteins (Abhiman, 2008), but BEN domains per se have been little studied to date. This study has found that mammalian BEND5 is noticeably related to Insv across several functionally important DNA-binding residues, including at many positions within the C-terminal tail region. Although it is possible that BEND5 may have a different optimal binding site, it was possible to demonstrate specific binding of BEND5 to the Insv target site. Therefore, it can be concluded that the BEN domain is a conserved DNA-binding domain. Other functional parallels between Insv and BEND5 have been identified, including their specific expression in neurons and their capacity to function as transcriptional repressors. Given that Insv and BEND5 are fairly unrelated in primary sequence, these findings suggest that some of the other >100 BEN proteins are likely to be transcription factors. Curiously, some other BEN proteins, including the BTB/POZ-BEN factor NAC-1 and the quadruple BEN protein BEND3, have previously been linked to transcriptional repression (Korutla, 2009; Sathyan, 2011). Potentially, their BEN domains are involved in recruiting these factors to their appropriate chromatin targets (Dai, 2013).

Finally, not only are BEN domains widely distributed across metazoans, they are also encoded by several viral genomes (Abhiman, 2008). The data suggest that BEN proteins may potentially prove to add to the catalog of cellular transcription factors that have been co-opted by viruses. Overall, our findings now provide molecular direction to the study of this large and enigmatic family of proteins (Dai, 2013)

Gain of function notch phenotypes associated with ectopic expression of the Su(H) C-terminal domain illustrate separability of Notch and hairless-mediated activities

The Notch signaling pathway is instrumental for cell fate decisions. Signals from the Notch receptor are transduced by CSL-type DNA-binding proteins. In Drosophila, this protein is named Suppressor of Hairless [Su(H)]. Together with the intracellular domain of the activated Notch receptor ICN, Su(H) assembles a transcriptional activator complex on Notch target genes. Hairless acts as the major antagonist of the Notch signaling pathway in Drosophila by means of the formation of a repressor complex together with Su(H) and several co-repressors. Su(H) is characterized by three domains, the N-terminal domain NTD, the beta-trefoil domain BTD and the C-terminal domain CTD. NTD and BTD bind to the DNA, whereas BTD and CTD bind to ICN. Hairless binds to the CTD, however, to sites different from ICN. This work addresses the question of competition and availability of Su(H) for ICN and Hairless binding in vivo. To this end, the CTD was overexpressed during fly development. A strong activation of Notch signaling processes was observed in various tissues, which may be explained by an interference of CTD with Hairless corepressor activity. Accordingly, a combined overexpression of CTD together with Hairless ameliorated the effects, unlike Su(H) which strongly enhances repression when overexpressed concomitantly with Hairless. Interestingly, in the combined overexpression CTD accumulated in the nucleus together with Hairless, whereas it is predominantly cytoplasmic on its own (Maier, 2013).

Thermodynamic binding analysis of Notch transcription complexes from D. melanogaster

Notch is an intercellular signaling pathway that is highly conserved in metazoans and is essential for proper cellular specification during development and in the adult organism. Misregulated Notch signaling underlies or contributes to the pathogenesis of many human diseases, most notably cancer. Signaling through the Notch pathway ultimately results in changes in gene expression, which is regulated by the transcription factor CSL. Upon pathway activation, CSL forms a ternary complex with the intracellular domain of the Notch receptor (NICD) and the transcriptional coactivator Mastermind (MAM) that activates transcription from Notch target genes. While detailed in vitro studies have been conducted with mammalian and worm orthologous proteins, less is known regarding the molecular details of the Notch ternary complex in Drosophila. This study thermodynamically characterized the assembly of the fly ternary complex using isothermal titration calorimetry. The data reveals striking differences in the way the RAM (RBP-J associated molecule) and ANK (ankyrin) domains of NICD interact with CSL that is specific to the fly. Additional analysis using cross-species experiments suggest that these differences are primarily due to fly CSL, while experiments using point mutants show that the interface between fly CSL and ANK is likely similar to the mammalian and worm interface. The binding of the fly RAM domain to CSL does not affect interactions of the corepressor Hairless with CSL. Taken together, these data suggests species-specific differences in ternary complex assembly that may be significant in understanding how CSL regulates transcription in different organisms (Contreras, 2015).

Suppressor of Hairless: Biological Overview | Evolutionary Homologs | Regulation | 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.