Targets of Activity

Dorsal mesoderm induction in arthropods and ventral mesoderm induction in vertebrates are closely related processes that involve signals of the BMP family. In Drosophila, induction by Decapentaplegic of visceral mesoderm, dorsal muscles, and the heart is, at least in part, effected through the transcriptional activation and function of the homeobox gene tinman in dorsal mesodermal cells during early embryogenesis. A functional dissection has been carried out of a tinman enhancer that mediates the Dpp response. Mesoderm-specific induction of tinman requires the binding of both activators and repressors. Screens for binding factors yielded Tinman itself and the Smad4 homolog Medea. The binding and synergistic activities of Smad and Tinman proteins are critical for mesodermal tinman induction, whereas repressor binding sites prevent induction in the dorsal ectoderm and amnioserosa. Thus, integration of positive and negative regulators on enhancers of target genes appears to be an important mechanism in tissue-specific induction by TGF-beta molecules (Xu, 1998).

Functional dissection of the tinman gene has identified a 349-bp enhancer in 3'-flanking regions, tin-D, that is strictly active in dorsal portions of the mesoderm of stage 10-11 embryos. The pattern of lacZ reporter gene expression driven by tin-D closely resembles the dpp-dependent pattern of endogenous tinman expression, suggesting that tin-D functions as a Dpp response element. This notion was further supported by the observation that tin-D reporter gene activity is absent in embryos with a dpp null mutant background. Conversely, upon the ectopic expression of a constitutively active DPP type I receptor, TkvQ-D in the entire mesoderm , tin-D reporter gene expression expands into the ventral mesoderm. The observed changes of tin-D activity when the levels and spatial extents of Dpp signaling are altered closely reflect the changes seen for tinman expression under the same conditions. These observations raise the possibility that the tin-D enhancer is receiving direct inputs from the Dpp signal transduction cascade to activate tinman transcription. In addition to its dependence on dpp, dorsal mesodermal tinman expression requires the activity of tinman itself, as tinman mutant embryos show strongly reduced expression. Correspondingly, full activity of the tin-D enhancer depends on the function of tinman as well. Taken together, these results suggest that Dpp signals and autoregulation by tinman cooperate to induce full levels of tin-D enhancer activity and tinman expression in the dorsal mesoderm (Xu, 1998).

Sequence comparisons between the tin-D elements from Drosophila melanogaster and Drosophila virilis, which displayed identical activities in D. melanogaster embryos, show a high degree of sequence similarity, whereas the similarities in the 5'- and 3'-flanking regions of tin-D elements are considerably lower. The strong sequence conservation between the tin-D enhancers from the two species could reflect the functional conservation of important regulatory sequences. A first inspection of the conserved sequences reveals several candidates for regulatory sites. One of them is a sequence that is present in duplicate, TCAAGTGG, which contains a binding site consensus for homeodomain proteins of the NK family and is identical to previously identified Tinman binding sequences from a heart enhancer of the Drosophila mef2 gene. Tinman protein has specific binding affinity to these sequences in vitro. Another completely conserved sequence is potentially interesting because it contains tandemly repeated CAATGT motifs, with each of the two copies being followed by a stretch of GC-rich sequences at their 3' ends (Xu, 1998).

To define essential regulatory sequences within the tin-D enhancer, a series of derivatives were created with various deletions of the most strongly conserved sequence blocks and their activity was tested in vivo. Three of these fine deletions do not affect lacZ reporter gene expression in transgenic embryos. This indicates that the deleted sequences (nucleotide 16-47, 205-229, and 244-312) either lack any regulatory potential or contain functionally redundant regulatory sequences. In contrast, two other deletions result in a strong reduction of enhancer activity. One of them encompasses the tandemly repeated CAATGT/GC motifs (deltaD3) and causes an almost complete loss of enhancer activity. The other, deltaD6, which deletes 30 bp from the 3' end of tin-D, also yields strongly reduced activity in the dorsal mesoderm. These results show that the subelements D3 and D6 contain important regulatory sequences for the induction of tinman in the dorsal mesoderm and thus are candidates for target sites of the Dpp signaling cascade (Xu, 1998).

To test whether the putative tinman binding sites play roles in autoregulation, the activity of a tin-D derivative, tin-D-deltaD1, in which both of these sites were deleted (nucleotides 1-13 and 197-203). Deletion of these sites provoked two interesting effects. The first is a significant reduction of lacZ reporter gene expression in the mesoderm, which indicates that tinman autoregulation is required to achieve full levels of dorsal mesodermal tinman induction through these sequences. A second, more unexpected effect is observed in the ectoderm. Specifically, embryos carrying tin-D-deltaD1 show strong ectopic reporter gene expression in the dorsal ectoderm, which corresponds to the areas of dpp expression at this stage of development . Accordingly, in a dpp mutant background, both the ectodermal and the residual mesodermal activities of this mutant element are absent. These results show that upon deletion of the tinman binding sites, tin-D is still able to respond to dpp, but its response is essentially switched from the target tissue to the signaling tissue. Therefore, it is concluded that in the normal situation, Tinman binding to these sites is required in an autoregulatory fashion for full induction of tinman by the Dpp signals in the dorsal mesoderm. In addition, the Tinman binding sites appear to overlap with binding sites for an unknown repressor that normally prevents induction of tinman in the dorsal ectoderm, and these two mechanisms together apparently ensure the mesoderm-specific response to Dpp (Xu, 1998).

In the normal situation, tinman autoregulation appears to be restricted to the mesoderm, presumably because the early, twist-activated phase of tinman expression is mesoderm specific. To test whether tinman is also able to autoregulate in the ectoderm, tinman was expressed ectopically and tin-D reporter gene expression was examined under these conditions. For this purpose, tinman was expressed with the binary UAS/GAL4 system in ectodermal stripes under the control of an engrailed driver. Ectodermally expressed tinman is capable of activating tin-D in the ectoderm. Interestingly, ectodermal tin-D expression is restricted to dorsal portions of the transverse Tinman stripes, thereby demonstrating that tinman autoregulation can occur both in the mesoderm and in the ectoderm, but only in conjunction with Dpp signaling (Xu, 1998).

Because a combination of tin-D1 and tin-D3 sequences is sufficient to reproduce a virtually normal expression pattern in the dorsal mesoderm, these combined sequences would appear to contain DNA sequences that can bind the essential factors involved in this inductive process. To identify some of these factors molecularly the yeast one-hybrid system was used to screen for Drosophila cDNAs encoding proteins that specifically bind to D1 or D3 sequences. With multimeric D3 sequences as a bait, GAL4 AD fusion cDNAs were isolated containing sequences with strong similarities to DPC4/Smad4 proteins, which are effectors of various TGF-beta signaling processes. Of 54 candidate yeast clones, 8 carried these sequences and were derived from the same gene. Subsequent sequence comparisons showed that these cDNAs correspond to the Medea gene. Conceptual translation and sequence alignments with other members of this protein family indicate that the encoded GAL4 fusion proteins contain the complete amino-terminal portion of Medea but lack the carboxy-terminal portion encoded by sequences 3' to a native NotI site. It is conceivable that the screen selected against full-length clones, because previous reports have shown that the carboxyl terminus of Smad proteins has autoinhibitory activities (Xu, 1998).

To locate the DNA-binding domain in the Medea protein, a series of carboxy-terminal truncation and in-frame fusion constructs of Medea cDNAs with GAL4 AD coding sequences were generated, and their binding activities were tested in the yeast system, using (D3)5/lacZ as a reporter gene. The activity of Medea products increases on removal of the MH2 domain in this assay, indicating that the MH2 domain of Smad4 group proteins has an inhibitory effect on DNA binding, similar to the MH2 domain of Mad group proteins. By removing 10 carboxy-terminal amino acids from the MH1 domain, the activity drops to background levels. The MH2 domain does not display any binding activities in this assay. Thus, it appears that the MH1 domain serves as the DNA-binding domain for Medea, as it does for Mad (Xu, 1998).

DNase I footprinting assays with bacterially expressed GST fusion proteins were used to characterize the binding of Tinman and Medea to sequences of the tin-D element and to test whether Mad is also able to bind. Tinman specifically protects the two D1 sequences that contain NK homeodomain binding sites and are required for autoregulation. The MH1 domains of Medea protect three distinct sequences within tin-D. Importantly, one of them (nucleotide 95-127) overlaps with the D3 sequence that is essential and sufficient for tin-D activity and was used for the isolation of Medea. Mad shows binding to several additional sequences. Two other sites that are protected by Mad, but not Medea under the same conditions, correspond to the 3' portion of D3 and adjacent sequences. Therefore, Mad protects most of the sequence stretch between nucleotides 95 and 160, which has D3 at its core, whereas Medea protects only the 5' two-thirds of D3. Taken together, the DNase I protection data reveal a minimum of eight in vitro binding sites for Medea and Mad in the tin-D element, at least four of which are located in the essential elements D3 and D6. Moreover, it appears that Medea and Mad have overlapping, but not identical, binding specificities to tin-D sequences (Xu, 1998).

Gel retardation assays provided additional information on the DNA-binding specificities of Medea and Mad and their binding sites in the tin-D element. Both Medea MH1 and Mad MH1 bind to 32P-labeled D3 probes, and excess of unlabeled D3 DNA can compete for binding. Because D3 contains tandemly repeated CAATGT and GC-rich motifs, a test was carried out to find out which of these two sequence motifs are involved in Medea and Mad binding. Replacement of four GCs in each of the GC-rich motifs by As and Ts renders the mutated D3 sequence unable to compete for Medea and Mad binding to the wild-type D3 sequence. In contrast, in vitro mutagenesis of the CAATGT motifs does not interfere with Medea and Mad binding, as these mutated versions compete equally well as D3 wild-type DNA. These data show that the GC-rich motifs are essential for Mad and Medea binding and likely represent two distinct binding sites for these proteins in the D3 element. In summary, these in vitro DNA binding studies demonstrate that the functionally significant D3 and D6 elements contain at least four GC-rich binding sites for Medea and Mad, although Medea binds with high affinity only to those in D3. In addition, tin-D contains at least four other binding sites for Medea and Mad, all of which include GC-rich stretches (Xu, 1998).

In vitro binding sites of Medea and Mad are essential for the activity of tin-D in vivo. Combinations of single copies of the wild-type and mutated sequences of D3 and D6 similar to the ones used for the in vitro binding assays were tested for their ability to activate reporter gene expression in transgenic embryos. The activity of these sequences was tested in the context of a shortened version of tin-D (tin-D*; nucleotide 1-143 plus Tin binding site 2 and nucleotide 321-349). Tin-D* containing wild-type D3 and D6 sequences produces an almost normal pattern of dorsal mesodermal expression, although expression levels are lower than with the complete tin-D element. In contrast, a version in which both Medea/Mad binding sites in D3 are disrupted by 8 bp exchanges is almost completely inactive in vivo. Similarly, expression is nearly abolished upon disruption of the two Mad binding sites in D6 or of all four Medea/Mad binding sites in D3 and D6. Interestingly, specific disruption of the CAATGT sequences in D3 also results in a complete loss of activity in the dorsal mesoderm. Together, these data demonstrate that each of the pairs of Medea/Mad binding sites in D3 and D6 plays a critical role in the Dpp-induced activity of tin-D in the dorsal mesoderm. Moreover, the CAATGT sequences in D3 appear to be required for the binding of a different factor that is also essential during this process. An interesting difference between tin-D and tin-D* is an ectopic expression in the amnioserosa, which is observed between stage 8 and 11 of embryogenesis. This observation suggests that the region between D3 and D6, which is missing in tin-D*, contains a repressor element for this tissue. The results with mutated versions of tin-D* indicate that the Medea/Mad binding sites in D3 and D6 are necessary for amnioserosa expression, whereas the CAATGT sequences are not required (Xu, 1998).

The absolute requirement for the tandemly repeated CAATGT sequences for the activity of the Dpp response element strongly points to the existence of a second essential coactivator that binds to these sequences. The results with wild-type and mutated versions of the tin-D element predict that this factor is expressed and active in both mesoderm and ectoderm, since disruption of the CAATGT motifs abolishes both mesodermal and ectopic ectodermal induction. The close juxtaposition of these motifs with Smad binding sites in the minimal Dpp response element may suggest that the unknown binding factor also participates in protein-protein interactions with bound Smad proteins. It is interesting to note that this sequence motif is closely related to that of the binding site of Xenopus FAST-1. The forkhead domain protein FAST-1 has been shown to bind to the sequences AAATGT within an activin-response element of the Mix.2 gene and to associate with Smad2 and Smad4 . It is thus conceivable that a related member of the forkhead domain protein family plays a similar role in the tinman Dpp response element, albeit in this case in a complex with DNA-associated Smads (Xu, 1998).

The function of the Drosophila mef2 gene, a member of the MADS box supergene family of transcription factors, is critical for terminal differentiation of the three major muscle cell types, namely somatic, visceral, and cardiac. During embryogenesis, mef2 undergoes multiple phases of expression, which are characterized by initial broad mesodermal expression, followed by restricted expression in the dorsal mesoderm, specific expression in muscle progenitors, and sustained expression in the differentiated musculatures. Evidence is presented that temporally and spatially specific mef2 expression is controlled by a complex array of cis-acting regulatory modules that are responsive to different genetic signals. Functional testing of approximately 12 kb of 5' flanking region of the mef2 gene shows that the initial widespread mesodermal expression is achieved through a 280-bp twist-dependent enhancer. Subsequent dorsal mesoderm-restricted mef2 expression is mediated through a 460-bp dpp-responsive regulatory module, which involves the function of the Smad4 homolog Medea and contains several binding sites for Medea and Mad. Regulated mef2 expression in the caudal and trunk visceral mesoderm, which give rise to longitudinal and circular gut musculatures, respectively, is under the control of distinct enhancer elements. In addition, mef2 expression in the cardioblasts of the heart is dependent on at least two distinct enhancers, which are active at different periods during embryogenesis. Mef2 expressing cells are coincident with those expressing Tinman. Notably, both Mef2 and Tinman expression are in four of six cardioblasts that are present per hemisegment. The complete overlap between the two expression patterns suggests that the activity of this enhancer element could be dependent on tinman function or under similar regulatory controls as is tinman. The cardiac enhancer that functions at later stages also drives mef2 expression in the caudal visceral mesoderm as well as in the somatic mesoderm. Moreover, multiple regulatory elements are differentially activated for specific expression in presumptive muscle founders, prefusion myoblasts, and differentiated muscle fibers. Taken together, the presented data suggest that specific expression of the mef2 gene in myogenic lineages in the Drosophila embryo is the result of multiple genetic inputs that act in an additive manner on distinct enhancers in the 5' flanking region (Nguyen, 1999).

Morphogen gradients control body pattern by differentially regulating cellular behavior. Molecular events underlying the primary response to the Dpp/BMP morphogen have been analyzed in Drosophila. Throughout development, Dpp transduction causes the graded transcriptional downregulation of the brinker (brk) gene. Significance for the brk expression gradient is provided by showing that different Brk levels repress distinct combinations of wing genes expressed at different distances from Dpp-secreting cells. The brk regulatory region has been dissected and two separable elements have been identified with opposite properties, a constitutive enhancer and a Dpp morphogen-regulated silencer. Furthermore, genetic and biochemical evidence is presented that the brk silencer serves as a direct target for a protein complex consisting of the Smad homologs Mad/Medea and the zinc finger protein Schnurri. Together, these results provide the molecular framework for a mechanism by which the extracellular Dpp/BMP morphogen establishes a finely tuned, graded read-out of transcriptional repression (Müller, 2003).

The Dpp signaling system shapes an inverse profile of Brk expression, which serves as a mold for casting the spatial domains of Dpp target genes. Thus, the question of how the Dpp morphogen gradient is converted into transcriptional outputs can be largely reduced to the question of how Dpp generates an inverse transcriptional gradient of brk expression. An unbiased approach was applied to this problem by isolating the regulatory elements of brk. A protein complex has been identified and characterized that binds to and regulates the activity of these elements in a Dpp dose-dependent manner (Müller, 2003).

Dissection of the brk locus reveals two separable elements with opposite properties: a constitutive enhancer and a morphogen-regulated silencer. Both elements have a direct effect on the level of brk expression, and it is the net sum of their opposing forces that dictates the transcriptional activity of brk in any given cell. In this sense, expression of the brk gene behaves like a spring that is compressed by Dpp signaling. Its silencer and enhancer embody the variable compressing and constant restoring forces, respectively. As stated by Hooke's law, an increased elastic constant (e.g., two copies of the constitutive enhancer) either shifts the brk levels toward those normally present at more lateral positions or necessitates a correspondingly higher compressing force (e.g., more silencer elements or higher levels of Dpp signaling). Given the central role Brk plays in controlling growth and pattern together with the direct impact of the two regulatory elements on brk levels, it appears inevitable that their quantitative properties must exhibit a fine-tuned evolutionary relationship with each other and with those of the Dpp transduction system. It appears, furthermore, that both the brk enhancer as well as the brk silencer elements represent ideal substrates for evolutionary changes in morphology (Müller, 2003).

Based on combined genetic and biochemical analysis, it is proposed that upon Dpp signaling the following key players meet at the brk silencer elements to execute repression: the Smad proteins Mad and Med and the zinc finger protein Shn. The role of Shn must be to direct the signaling input provided by Mad and Med into transcriptional silencing. In principle, two scenarios can be envisaged by which Shn fulfills this task. Shn could possess repressor activity (presumably via recruitment of corepressors) but lack the ability to bind the brk silencer and, hence, depend on Mad/Med for being targeted to its site of action. Alternatively, Shn could be prebound to the silencer, but only be capable of recruiting corepressors upon interaction with Mad/Med. Based on the observation that a Shn/DNA complex cannot be detected in the absence of Mad/Med, the first of these two possibilities is favored. The molecular architecture of the protein complex binding to the brk silencer as well as the DNA sequences providing the specificity for the local setup of this complex remain to be determined in detail (Müller, 2003).

An additional protein, which appears to influence the events at the brk silencer, is Brk itself. Genetic experiments indicate that Brk negatively modulates its own expression, forming a short regulatory loop that contributes to the final shape of the Brk gradient. This autoregulatory action occurs also via the brk silencer element, suggesting that Brk directly participates in the protein-protein or protein-DNA interactions at this site (Müller, 2003).

Most regulatory events ascribed to Smad proteins to date concern signaling-induced activation of target gene transcription. In the case of the brk silencer Shn could be regarded as a 'switch factor' that converts an inherently activating property of Smad proteins into transcriptional repression activity. Indeed, it has been shown that Smad proteins have the ability to recruit general coactivators with histone acetyl transferase activity. However, in an alternative and more general view, Smad proteins per se may provide no bias toward activation or repression. Their main function may be to assemble transcriptional regulatory complexes involving other DNA binding proteins and endow these complexes with additional DNA binding capacity. Such associated DNA binding factors would not only determine target site specificity, but, by their recruitment of either coactivator or corepressor proteins, also define the kind of regulatory influence exerted on nearby promoters. Since Shn directs Mad/Med activity toward repression, the existence of at least one other such Mad/Med partner in Drosophila is hypothesized to account for Mad/Med-mediated activation of gene expression. Such Mad/Med-mediated activation appears to be required for peak levels of sal and vg transcription, as well as for defining gene expression patterns in domains where brk expression is completely repressed, e.g., close to the Dpp source of the dorsal embryonic ectoderm (Müller, 2003).

At the heart of the model is the direct causal relationship between the formation of a Shn/Mad/Med/brk-silencer complex and the silencing of brk gene transcription. Although the two observations have been derived from different experimental data sets (biochemical versus genetic, respectively), there is a firm correlation between the requirements for either event to occur. brk is not repressed when either (1) the brk silencer elements are lacking or mutated; (2) or when Dpp input is prevented (and hence Mad is neither phosphorylated, nor nuclearly localized, nor associated with Med), or when (3) Shn is not present or is deprived of its C-terminal zinc fingers. The same set of requirements was observed for the formation of the Shn/Mad/Med/brk complex. Moreover, it is the concurrence of all three of these conditions that appears to provide the exquisite specificity to the Dpp-regulated silencing of gene transcription. (1) It only occurs in conjunction with a functional brk silencer, or an equivalent element. (2) There is an absolute requirement for Dpp input in Shn-mediated silencing. Not even a partial repressor activity of Shn was observed in cells that do not receive Dpp signal (e.g., loss of shn function in cells situated in lateral-most positions of the wing disc does not cause a further upregulation of brk transcription). (3) Shn represents only one of several zinc finger proteins expressed in Dpp receiving cells, yet none of the other proteins is able to substitute for Dpp-mediated repression. A major determinant for the specificity with which Shn engages in the signaling-dependent protein/DNA complex appears to be the triple zinc-finger motif. Although it is likely that this structural feature is required for contacting specific nucleotides on the brk silencer, the possibility cannot be not excluded that some of the zinc fingers mediate protein-protein interactions between Shn and Mad, Med or other cofactors (Müller, 2003).

During germ band elongation, widespread dpp expression in the dorsal ectoderm patterns the underlying mesoderm. These Dpp signals specify cardial and pericardial cell fates in the developing heart. At maximum germ band extension, dpp dorsal ectoderm expression becomes restricted to the dorsal-most or leading edge cells (LE). A second round of Dpp signaling then specifies cell shape changes in ectodermal cells leading to dorsal closure. A third round of dpp dorsal ectoderm expression initiates during germ band retraction. This round of dpp expression is also restricted to LE cells but Dpp signaling specifies the repression of the transcription factor Zfh-1 in a subset of pericardial cells in the underlying mesoderm. Surprisingly, cis-regulatory sequences that activate the third round of dpp dorsal ectoderm expression are found in the dpp disc region. The activation of this round of dpp expression is dependent upon prior Dpp signals, the signal transducer Medea, and possibly release from dTCF-mediated repression. These results demonstrate that a second round of Dpp signaling from the dorsal ectoderm to the mesoderm is required to pattern the developing heart and that this round of dpp expression may be activated by combinatorial interactions between Dpp and Wingless (Johnson, 2003).

This study suggests that a second round of dpp LE expression leads to a second round of Dpp dorsal ectoderm to mesoderm signaling. The role of the second round of Dpp LE signaling appears to be the specification of a novel subset of pericardial cells by repressing the expression of the transcription factor Zfh-1. Sequences in the dpp disc region and the signal transducer Medea are required for the second round of dpp LE expression (Johnson, 2003).

Genetic analysis demonstrates that some of the enhancers responsible for the second round of dpp LE expression are contained within the portion of BS3.21 that is not shared with BS3.22. An interspecific comparison shows that this region contains a stretch of 96 nucleotides that is 95% identical between two evolutionarily distant Drosophila species. An examination of this 96-base pair stretch reveals a candidate Mad/Med binding site that is 100% conserved between the species. This is consistent with results showing that BS3.21 expression requires dpp disc region sequences and that dpp151H expression requires Medea (Johnson, 2003).

Is the second round of Dpp LE signaling conserved in vertebrates? The initial round of Dpp dorsal ectoderm to mesoderm signaling (the specification of heart progenitors via the maintenance of tinman expression) is conserved in vertebrate heart development. The mammalian homolog of Dpp (bone morphogenetic protein-2, BMP2) specifies cell fate in the cardiac mesoderm by regulating the expression of the tinman homolog Nkx2.5. Further, the interaction between Dpp and tinman and between BMP-2 and Nkx2.5 is via evolutionarily conserved enhancers recognized by the homologous signal transducers Medea and Smad4 (Johnson, 2003).

Mad and Medea silence bam to establish asymmetric divisions of germline stem cells

Stem cells execute self-renewing and asymmetric cell divisions in close association with stromal cells that form a niche. The mechanisms that link stromal cell signaling to self-renewal and asymmetry are only beginning to be identified, but Drosophila oogenic germline stem cells (GSCs) have emerged as an important model for studying stem cell niches. Decapentaplegic sustains ovarian GSCs by suppressing differentiation in the stem cell niche. Dpp overexpression expands the niche, blocks germ cell differentiation, and causes GSC hyperplasty. The bag-of-marbles (bam) differentiation factor is the principal target of Dpp signaling in GSCs; ectopic bam expression restores differentiation even when Dpp is overexpressed. The transcriptional silencer element in the bam gene integrates Dpp control of bam expression. Finally and most significantly, this study demonstrates that Dpp signaling regulates bam expression directly since the bam silencer element (SE) is a strong binding site for the Drosophila Smads, Mad and Medea. These studies provide a simple mechanistic explanation for how stromal cell signals regulate both the self-renewal and asymmetric fates of the products of stem cell division (Chen, 2003).

GSCs divide in the anterior/posterior axis, and this division produces daughters with different fates. The anterior cell of a GSC division retains contact with the stromal cap cells, maintains high levels of Dpp signaling, and continues as a stem cell. The posterior stem cell daughter dissociates from the cap cells and becomes a cystoblast (CB). The CB divides precisely four times with incomplete cytokinesis, giving rise to a cyst of 16 interconnected cells that differentiate further into 1 oocyte and 15 nurse cells. The progress of cyst formation can be followed by monitoring the morphogenesis of a dynamic organelle, termed the fusome, that grows and branches with each cyst cell mitosis (Chen, 2003).

Cystoblasts require the product of the bag-of-marbles (bam) gene, which is both necessary and sufficient for differentiation. Germ cells lacking bam fail to differentiate into cystoblasts and continue to divide with full cytokinesis, producing germ cell hyperplasia. The failure of these proliferating germ cells to differentiate can be recognized by following the fusome since it remains spherical instead of growing into a branched structure. Superficially, therefore, bam mutant cells behave like GSCs, but molecular markers to determine the stage of arrest have been lacking (Chen, 2003).

Studies characterizing bam transcription demonstrate that bam is tightly regulated such that it is off in GSCs and on in CBs. Thus, it was possible to determine if bam mutant cells are 'stem cell-like' since the activity of a bam reporter transgene would distinguish GSCs from CBs precisely. GFP expression was examined in bam mutant animals carrying a transgene with the bam promoter fused to GFP; most germ cells were observed to be GFP positive. Thus, unlike GSCs, germ cells lacking bam advanced to a state of differentiation sufficient to activate bam transcription (Chen, 2003).

Two features of GSC divisions demand molecular explanation: how is the anterior daughter of the GSC division retained as a stem cell (self-renewal) and what causes the posterior daughter to differentiate into a CB (asymmetric division)? Stromal cells, including the cap and inner sheath cells, at the germarial tip express several signaling molecules and are a likely source of dpp. dpp signaling has been shown to be required to maintain GSCs, and transcriptional control over Bam has been shown to distinguish GSCs from CBs. These two phenomena can now be linked directly (Chen, 2003).

In GSCs, in which dpp signaling and pMad levels are highest, the Mad:Med complex binds to the bam SE and prevents bam transcription. GSCs self-renew because association of the anterior daughter with stromal cells permits sufficient dpp signaling to block CB differentiation by assembling a repressor complex on the bam SE element. This complex is likely to include other factors required for transcriptional antagonism, such as TGIF, a homeodomain-containing transcriptional corepressor of TGF-beta-dependent gene expression, or Ski/Sno factors, which can recruit histone-modifying enzymes. The complex may also contain Schnurri, a negatively acting Mad cofactor, since shn mutant GSCs also differentiate precociously (Chen, 2003).

During division, a GSC daughter cell is displaced away from the cap cells and into a region of diminished dpp signaling. This cell, the CB precursor (pre-CB), expresses lower pMad levels that would cause the concentration of Mad:Med complexes to fall. Declining occupancy levels of the bam SE would produce derepression of bam transcription and concomittant activation of the CB differentiation program (Chen, 2003).

Embryonic stem cells are considered totipotent because they can populate any of the adult niches. Although the degree of adult stem cell plasticity is currently receiving much attention, assembly of stem cells into signaling niches during postembryonic development might impose differentiation limits. What are the specific effects of stromal cell niches on captured stem cells? In the case of the Drosophila ovarian niche, GSCs are maintained as 'CBs-in-waiting' because a stromal cell signal represses the expression of one key factor (i.e., Bam). Perhaps other types of stem cells are similarly differentiated but blocked by stromal cell signals and require the expression of only one or a few key molecules to resume development (Chen, 2003).

The Drosophila ovary is an attractive system to study how niches control stem cell self-renewal and differentiation. The niche for germline stem cells (GSCs) provides a Dpp/Bmp signal, which is essential for GSC maintenance. bam is both necessary and sufficient for the differentiation of immediate GSC daughters (cystoblasts). Bmp signals directly repress bam transcription in GSCs in the Drosophila ovary. Similar to dpp, gbb encodes another Bmp niche signal that is essential for maintaining GSCs. The expression of phosphorylated Mad (pMad), a Bmp signaling indicator, is restricted to GSCs and some cystoblasts, which have repressed bam expression. Both Dpp and Gbb signals contribute to pMad production. bam transcription is upregulated in GSCs mutant for dpp and gbb. In marked GSCs mutant for two essential Bmp signal transducers (Med and punt) bam transcription is also elevated. Finally, Med and Mad are shown to directly bind to the bam silencer in vitro. This study demonstrates that Bmp signals maintain the undifferentiated or self-renewal state of GSCs, and directly repress bam expression in GSCs by functioning as short-range signals. Thus, niche signals directly repress differentiation-promoting genes in stem cells in order to maintain stem cell self-renewal (Song, 2004).

Smad affinity can direct distinct readouts of the embryonic extracellular Dpp gradient in Drosophila

The TGF-β signaling molecule Dpp is an essential morphogen that patterns many tissues during Drosophila development, including the embryonic dorsal ectoderm and larval wing imaginal disc. An activity gradient of Dpp specifies distinct cell fates in the dorsal ectoderm of the embryo through the activation of different transcriptional threshold responses. The gene Race, which is expressed in response to peak levels of Dpp signaling in gastrulating embryos, was analyzed. The Smad transcription factors, which are intracellular transducers of Dpp signaling, are essential activators of Race in vivo. Furthermore, increasing the affinity of the Smad binding sites in the Race enhancer broadens the expression pattern of a linked reporter gene and alters its behavior in mutant embryos to that characteristic of a distinct threshold response. It is concluded that Smad activator affinity is a critical determinant of the threshold response to the extracellular Dpp gradient in the embryo. These results identify a mechanism for interpreting the Dpp gradient in the embryo which is different fromthe reciprocal repressor gradient model proposed for the wing disc. It is suggested that transcription factor binding site affinity will be a general strategy used in the interpretation of other extracellular morphogen gradients (Wharton, 2004).

A 533 bp enhancer has been identified that directs expression of a linked lacZ reporter gene in transgenic embryos in a similar pattern to that of endogenous Race. Race expression is dependent on peak Dpp signaling, and the Race enhancer contains three binding sites for the Zerknüllt (Zen) transcription factor. Zen, which is itself activated by Dpp signaling in the presumptive amnioserosa of gastrulating embryos, is essential for Race activation but not sufficient in the absence of Dpp signaling. This raised the possiblity that the Mad and Medea transcription factors, which are the intracellular transducers of Dpp signaling, may function with Zen to directly activate Race. Therefore, whether Mad and Medea could bind directly to the Race enhancer was tested in vitro. DNase I footprinting assays with bacterially expressed GST fusion proteins were used to characterize the binding of Mad and Medea to three overlapping fragments of the Race enhancer. Mad and Medea bind to two of the three enhancer fragments tested. Three areas are protected from DNase digestion in the presence of Mad/Medea fusion proteins. One of these footprints (site A) encompasses nucleotides 28-41, and the other two sites, B and C, are adjacent to each other (nucleotides 464-483 and 484-502, respectively). These adjacent binding sites are flanked by the previously identified Zen binding sites. Mad and Medea recognize the same sites within the Race enhancer, consistent with previous observations that Mad and Medea have overlapping binding specificities (Wharton, 2004).

These binding sites were confirmed by gel retardation assays using oligonucleotide probes containing the protected regions and a GST-Mad fusion protein. It appears that sites A and B are low-affinity binding sites, whereas site C is higher affinity; an increased amount of the retarded complex is observed with probe C and Mad, even though equimolar amounts of probe and the same amount of Mad protein were used. Since similar binding data were obtained for Medea, and given the overlapping binding specificities documented for Mad and Medea, subsequent in vitro binding assays were only performed using Mad (Wharton, 2004).

Mad/Med regulate decapentaplegic expression during Drosophila wing veins pupal development

The differentiation of veins in the Drosophila wing relies on localised expression of decapentaplegic in pro-vein territories during pupal development. The expression of dpp in the pupal veins requires the integrity of the shortvein region (shv), localised 5' to the coding region. It is likely that this DNA integrates positive and negative regulatory signals directing dpp transcription during pupal development. A minimal 0.9 kb fragment has been identified giving localised expression in the vein L5 and a 0.5 kb fragment giving expression in all longitudinal veins. Using a combination of in vivo expression of reporter genes regulated by shv sequences, in vitro binding assays, and sequence comparisons between the shv region of different Drosophila species, binding sites were found for the vein-specific transciption factors Araucan, Knirps and Ventral veinless, as well as binding sites for the Dpp pathway effectors Mad and Med. It is concluded that conserved vein-specific enhancers regulated by transcription factors expressed in individual veins collaborate with general vein and intervein regulators to establish and maintain the expression of dpp confined to the veins during pupal development (Sotillos, 2006).

The expression of dpp in the wing disc is restricted to a narrow stripe of anterior cells localised at the anterior/posterior compartment boundary. This expression is regulated by sequences localised 3′ to the dpp coding region, and the function of the gene at this stage is required for the growth and patterning of the wing. The expression of dpp is still detected at the A/P boundary during the 8 h of pupal development. Later, at 14 h APF novel domains of dpp expression appear corresponding to the developing wing veins. The function of dpp during pupal development requires the integrity of the shv region, which is localised 5′ to the dpp coding region. There are two different transcripts expressed during pupal development, transcripts dpp-RA and dpp-RC, whose promoters (P5 and P3, respectively) are separated by approximately 20 kb of DNA. This DNA includes the first exon of transcript dpp-RC and corresponds to the place where all dpps alleles map. Because the strength of dpps alleles correlates with their distance to the P3 promoter, it is likely that dpp function in pupal development is mediated mainly by transcript dpp-RC. This suggests that dpps mutations affect regulatory sequences necessary to drive dpp expression in presumptive vein territories during pupal development. This possibility was confirmed by analysing the expression of a 8.5 kb construct containing most of the shv region fused to the reporter gene lacZ (shv8.5lacZ). The expression of βGal in shv8.5lacZ is detected exclusively in the pupal veins, indicating that this region includes all dpp wing veins regulatory regions (Sotillos, 2006).

Several constructs were made using different sub-fragments from the original 8.5 kb dpps DNA to identify with more precision the sequences that regulate dpp expression during pupal development. These fragments were cloned in front of a dicistronic lacZGal4 reporter gene and the activity of these constructs was analysed by looking at the expression of βGal in pupal wings from transgenic flies. In addition, to amplify the signal of the dicistronic lacZGal4 gene, the expression was monitored of a reporter gene regulated by UAS sequences. This expression should reveal all places where the Gal4 protein is present. Several regulatory regions were detected that control dpp expression in the veins during pupal development. One regulatory sequence is localised in a 1.1 kb fragment localised 6.5 kb from P3, and drives high levels of expression in most pupal veins and low levels of expression in some interveins. Additional regulatory sequences that efficiently drive expression in most veins are localised in an adjacent 0.5 kb fragment, and further vein-specific regulatory sequences for L5 are localised in the 0.9 kb SalI/KpnI fragment (Sotillos, 2006).

The expression of dpp during embryogenesis is highly dynamic and several independent regulatory regions controlling embryonic dpp expression have already been identified. The shv constructs included in the 8.5 kb EcoRI fragment drive reporter expression during embryonic development from stage 12/13 mainly in three regions of the mesoderm: the oesophagus, gastric caeca and midgut. Regulatory regions controlling dpp expression in the oesophagus appear to be duplicated, because they are localised in the 2.7 kb EcoRI/SalI fragment and also in the 3 kb KpnI/KpnI fragment. Similarly, regions controlling dpp expression in the gastric caeca are also present in the two adjacent fragments 0.9 kb SalI/KpnI and 3 kb KpnI/KpnI. The regions driving reporter expression in the gut are localised in the 3′ end of the shv region (Sotillos, 2006).

To better understand the regulation of dpp expression during vein development, the interactions were analyzed between a 2.5 kb region including wing veins pupal enhancers and several proteins involved in the regulatory network controlling the formation of veins. For this purpose, the 2.5 kb region was subdivided into overlapping fragments of 250-300 bp used as probes to detect the binding of different transcription factors by Electrophoretic Mobility-Shift Assays (EMSAs). Both prepattern specific genes that control vein development, such as Ventral veinless (Vvl) and the Araucan protein (Ara), and transcription factors belonging to the Dpp pathway (Mad and Medea) were analyzed (Sotillos, 2006).

The DNA-binding activity of Drosophila Smad proteins, Mad and Medea, is crucial for the expression of Dpp target genes. The expression of phosphorylated Mad (p-Mad), the activated form of the Mad protein, is restricted to the developing veins during pupal development. The efficiency of ectopic dpp expression to direct vein differentiation depends on the integrity of the shv region, suggesting that Dpp signalling is sufficient, directly or indirectly, for driving the expression of additional dpp transcription via the shv enhancer. Therefore, whether the Smads proteins can bind to the dpps enhancer was studied. Specific binding was obtained using all probes as shown by competition both with cold DNA and with specific oligonucleotides containing consensus binding sequences for Medea and for Mad and Medea. Other oligonucleotides with consensus for the transcription factor Nubbin did not compete the binding. The three main regions of competition with Mad and Medea binding included in the S9 and S10 probes correspond to GC rich sequences characteristic of Smad-response elements. However, only a single consensus sequence for Mad/Med (GCGGCTGT) in S10 is localised in a highly conserved region of different Drosophilids (see below) (Sotillos, 2006).

The pattern of four longitudinal veins is very similar in all Drosophilids despite the differences in wing size and pigmentation that exist between species. This conservation suggests that the mechanisms underlying vein pattern formation are conserved. The availability of the genomic sequence for different Drosophila species allows a direct comparison between their dpps regions. Two Drosophila species from the melanogaster group (D. melanogaster and D. ananassae), one Drosophila from the obscura group (D. pseudoobscura) and D. virilis from the virilis group were compared. It is expected that sequence similarity in non-coding regions corresponds to functional regulatory DNA. In the 2.5 kb region analysed several clusters of sequence conservation were found including most of the binding sites identified by in vitro analysis. Thus, there are four highly conserved regions corresponding to the S1, S4-5, S7-8 and S9-S10 probes containing conserved binding sequences for Vvl, Mad, Med and Ara. This conservation reinforces the importance of these regions to regulate the expression of dpp in the pupal veins. In the case of Vvl specific DNA binding to all probes was shown. However, the putative Vvl binding sequences localised in the conserved regions are only in S1, S3, S7, S8 and S10. In the case of the Dpp pathway transcription factors Mad and Med, putative binding sites are present throughout the enhancer, and accordingly binding of them to all probes was shown. However, only the S5 and S10 probes contain putative binding sites in regions of high sequence conservation. Interestingly, these conserved Mad/Med binding regions contain overlapping binding consensus for the Brinker repressor. This suggests a competition mechanism between Mad/Med and Brinker for binding to the shv enhancer. Competition mechanisms between activator and repressor also occur in several Dpp-downstream genes such as zen and omb. Four consensus binding sequences were found for the transcription factors of the Knirps-complex. The kni genes are expressed in the L2 vein, where they are required for its formation. Three Kni-binding sites were found in the 1.1 kb KpnI/SacII enhancer and one in the 0.5 SacII regions. Only two of the sequences located in the 1.1 kb KpnI/SacII enhancer present some conservation between Drosophilids. Interestingly, the 0.9 kb SalI/KpnI enhancer responsible of dpp expression in the L5 veins does not contain any putative Knirps binding sequence. Although whether Kni binds directly to the shv enhancer has not been analyzed, the presence of Kni-binding sites in conserved regions of the enhancer suggests that, in addition to its role during imaginal development, Kni might also activate dpp transcription during pupal development (Sotillos, 2006).

Therefore, regulatory sequences that drive dpp expression in the pupal veins in 2.5 kb of the dpps region have been found. This regulatory DNA can be subdivided into three fragments, a 1.1 kb fragment that recapitulates almost completely the pupal expression of dpp, a 0.9 kb upstream fragment, which drives expression in the proximal part of L5, and a 0.5 kb fragment that directs expression in all veins. Binding sites were found in these fragments for general transcription factors involved in the development of all veins (Vvl) and for the downstream activators of the dpp pathway, Mad and Medea. The regulatory region also contains binding sites for transcription factors expressed and required only in specific veins, such as Ara (L3 and L5) and Kni (L2). Most of these sequences are located in highly conserved regions of the dpp gene in different Drosophila species, indicating a general conservation of dpp regulation in the Drosophilids (Sotillos, 2006).

Decapentaplegic-responsive silencers contain overlapping mad-binding sites

Smad proteins regulate transcription in response to transforming growth factor-β signaling pathways by binding to two distinct types of DNA sites. The sequence GTCT is recognized by all receptor-activated Smads and by Smad4. The subset of Smads that responds to bone morphogenetic protein signaling recognizes a distinct class of GC-rich sites in addition to GTCT. Recent work has shown that Drosophila Mad protein, the homologue of bone morphogenetic protein rSmads, binds to GRCGNC sites through the same MH1 domain β-hairpin interface used to contact GTCT sites. However, binding to GRCGNC requires base-specific contact by two Mad proteins, and this study provides evidence that this is achieved by contact of the two Mad subunits that overlap across the two central base pairs of the site. This topology is supported by results indicating that His-93, which is located at the tip of the Mad β-hairpin, is in close proximity to base pairs 2 and 5. Also consistent with the model is disruption of binding by mutation of Glu-39 and Glu-40, which are predicted to lie at the interface of the two overlapping Mad MH1 domains. As predicted from the overlapping model, binding is disrupted by insertion of 1 bp in the middle of the site, whereas insertion of 2 bp creates abutting sites that can be bound by the Mad-Medea heterotrimer without requiring Glu-39 and Glu-40. Overlapping Mad sites predominate in Decapentaplegic response elements, consistent with a high degree of specificity in response to signaling (Gao, 2006).

DNA contact by Smad proteins has been shown to play an important role in many instances of target regulation by TGFβ pathways. For the consensus Smad3/Smad4-binding site, GTCT can also be bound by Mad- and BMP-specific Smad1, but Smad3 does not bind to GC-rich Mad/Smad1-binding sites, leaving open the question of whether such sites are contacted by a different mechanism. Recent work had shown that Mad-binding sites within the brk and bam silencers are bound by two Mad subunits and that in each case two Mad MH1 domains simultaneously contact a single 6-bp site using the same three β-hairpin residues that are responsible for base-specific contact by Smad3. By using mutational analysis and directly measuring binding, this study provides evidence that two Mad MH1 domains bind to the 6-bp site by overlapping across the two central base pairs. Smad1-binding sites match this 6-bp motif, a likely indication that overlap is also a feature of BMP-response elements (Gao, 2006).

The overlapping structure of Mad sites explains the seeming discrepancy between the Mad consensus and that of Smad3/Smad4. Smad3 differs from Mad at two positions that influence binding to brkS. Arg-58 at the C terminus of helix 2 is absolutely essential for binding to brkS; in Smad3 this position is a threonine, whereas the adjacent Lys-59 of Mad is absent in Smad3. Glu-39 at the N terminus of helix 2 contributes substantially to binding affinity for brkS, and Smad3 has instead a glutamine at this position. In addition, the loop between helices 1 and 2 is three residues shorter in Mad than in Smad3, a difference that modeling suggests will affect the structure of the alpha-carbon backbone and side chains near the N terminus of helix 2. Each of these differences is conserved between Smad3 and Smad2 and between Mad and the vertebrate BMP-specific rSmads (Gao, 2006).

Mutational analysis indicated the optimal sequence for an overlapping Mad site is GGCGCC, meaning each Mad MH1 prefers GGCG in the context of overlap. However, even when the two Mad sites are spaced such that they do not overlap, GGCG is still bound by Mad with about the same affinity as GTCT. The structural basis for this compatibility with two distinct sites remains to be determined, but the differential effects of helix 2 alanine substitutions suggest distinct docking geometries. Individual GGCG motifs occur in Dpp and BMP-response elements, and the results indicate that these are likely sites for contact by Mad and Smad1 (Gao, 2006).

Although the natural brkS element was specific for the Mad-Medea heterotrimer, changing the Mad site to abutting SBEs allowed binding by Mad alone or by Medea alone. The ability of such a site to be bound by Medea oligomers (putatively homotrimers) without Dpp signaling seemingly would make it ill-suited to function as a Dpp-response element, although signaling-induced activation was observed by reporter analysis (perhaps an indication that Medea alone is a poor activator). However, the brkS derivative with abutting GGCG sites (i.e. GGCGCGCC) shows little or no Medea binding in the absence of active Mad, is able to recruit Shn, and causes repression in response to signaling. Similar sites in the Dpp-response element of Race (GACGCGAC), which does not respond to repression by Brk protein, and in a BMP-response element of Smad7 (GGCGCGCC) appear to be examples of functional nonoverlapping Mad/Smad1 sites. In Drosophila a potentially significant difference between overlapping and nonoverlapping Mad sites is that the overlapping motif allows for competitive binding by the Brinker protein and thus dual control of Dpp targets, whereas the nonoverlapping motif does not. This may account for the predominance of overlapping Mad sites in Drosophila. The predominance of overlapping sites in BMP-response elements may reflect specificity for Smad1 but not Smad3 (Gao, 2006).

Collaboration between Smads and a Hox protein in target gene repression

Hox proteins control the differentiation of serially iterated structures in arthropods and chordates by differentially regulating many target genes. It is yet unclear to what extent Hox target gene selection is dependent upon other regulatory factors and how these interactions might affect target gene activation or repression. Two Smad proteins, effectors of the Drosophila Dpp/TGF-ß pathway, that are genetically required for the activation of the spalt (sal) gene in the wing, collaborate with the Hox protein Ultrabithorax (Ubx) to directly repress sal in the haltere. The repression of sal is integrated by a cis-regulatory element (CRE) through a remarkably conserved set of Smad binding sites flanked by Ubx binding sites. If the Ubx binding sites are relocated at a distance from the Smad binding sites, the proteins no longer collaborate to repress gene expression. These results support an emerging view of Hox proteins acting in collaboration with a much more diverse set of transcription factors than has generally been appreciated (Walsh, 2007).

The activation of sal in the wing and its repression in the haltere are regulated by a 1.1 kb CRE, sal1.1 (Galant, 2002). Previous studies have shown that sal1.1 is directly repressed by Ubx in the haltere (Galant, 2002). In order to test whether Mad/Med binds to and directly represses the activity of the sal1.1 CRE in the haltere, candidate Mad/Med binding sites were sought in the sal1.1 CRE. One candidate Mad/Med binding site, M1 (5'-AGACGGGCAC-3'), was identified that lies between Ubx binding sites 5 and 6 in sal1.1, using binding site prediction and electrophoretic mobility shift assays (EMSAs). The sequence of M1 deviates somewhat from published Mad/Med silencer consensus binding sites (5'-AGAC-5 bp-GNCGYC-3') (Gao, 2005; Pyrowolakis, 2004), and Mad and Med bound with >10-fold and >25-fold lower affinities, respectively, to the M1 site than to the bam (Gao, 2005) and brk (Pyrowolakis, 2004) silencer elements (Walsh, 2007).

In order to test whether Mad/Med bound specifically to the M1 site, a series of point mutations were introduced within the M1 site, and their effect on protein binding was examined in vitro. Of four point mutations to the M1 site, the single mutation at position 808 reduced the binding of a Med fusion protein (GST-MedMH1) to M1 as compared with the wild-type sequence. The remaining three point mutations did not affect the affinity of GST-MedMH1 for the probe. These results suggest that Med might contact the sequence 5'-AGAC-3' in sal1.1. By contrast, the four individual point mutations each decreased, but did not abolish, binding of a Mad fusion protein (GST-MadN) in vitro, with the point mutation at bp 814 having the strongest effect. The weaker effect of the individual point mutations in M1 on Mad binding affinity in vitro is likely to be due to the affinity of MadN for both 5'-AGAC-3' Smad sites and GC-rich sequence. Combining these four mutations (sal798-824 kM1) had the greatest effect on GST-MadN binding to the probe. This analysis of individual point mutations indicates a putative orientation for a Mad/Med compound-binding site in the sal1.1 CRE (Walsh, 2007).

Most importantly, in transgenic flies, each point mutation of M1 introduced into an otherwise wild-type sal1.1 reporter construct caused derepression of the reporter gene lacZ in the haltere imaginal disc. The strength of derepression correlates with the decreased affinity of Mad for its binding site with the pm814 mutation, the strongest point mutation in vitro, showing the strongest level of derepression in vivo. Full derepression was observed when all four point mutations were combined into a sal1.1 reporter construct. No effect of mutations in M1 were observed on sal1.1-driven reporter gene expression in the wing as compared with the wild-type sal1.1 element or with endogenous sal expression, indicating that this site is not required for gene activation in the wing or haltere disc. Together, the biochemical, reporter gene and genetic evidence indicate that Mad/Med/Shn are directly required for sal repression in the haltere imaginal disc (Walsh, 2007).

This study demonstrates that Mad/Med and Ubx bind to adjacent sites in the sal1.1 CRE and that each protein is required for the direct repression of sal expression in the haltere. Furthermore, the sequence and spacing of Ubx and Smad binding sites are highly conserved and their proximity is required for target gene repression in the haltere. Because no evidence was found that these proteins interact directly, it is suggested this is an example of 'collaboration' or target gene co-regulation without direct cooperative interaction. These results have general implications for understanding how Hox proteins regulate diverse sets of target genes in animal development (Walsh, 2007).

The direct role for Smads in the repression of sal in the haltere is surprising in the light of previous genetic and molecular studies that had indicated that the Dpp pathway and Mad/Med were involved in sal activation in the wing. No direct evidence was found that this is the case and the fact that sal is activated in Mad and Med clones in the haltere indicates that sal is activated independently of Mad/Med in the flight appendages. The requirement for Mad/Med/Shn in shaping the pattern of sal expression in the wing appears to be indirect -- the protein complex represses the expression of brk, a repressor of sal, in cells in the central region of the developing wing and thereby permits sal expression (Walsh, 2007).

The Mad-Med-Shn complex is also active within cells in the central region of the haltere as a consequence of Dpp signaling. However, whereas sal is expressed and the sal1.1CRE is active in the wing, sal and the sal1.1 CRE are repressed in the haltere. These observations raise the question of how the Mad-Med-Shn complex selectively represses sal in the haltere but not in the wing disc? The results suggest that there are two key determinants in the selective repression of sal in the haltere. The first is collaboration with Ubx, which is expressed in the haltere and not in the wing disc. The second key determinant might be the affinity of Mad/Med binding to the sal CRE (Walsh, 2007).

The different responses of the brk and sal genes to Mad/Med/Shn suggests how the different affinities of proteins for binding sites might determine how available transcriptional regulatory inputs are integrated by CREs. Mad/Med binding to the brk CRE is of high affinity (Pyrowolakis, 2004) and apparently sufficient to impart repression, whereas that to the sal CRE is of much lower affinity and insufficient to impart repression in the wing. In the haltere, although Mad-Med-Shn or Ubx binding are alone insufficient, they act together either via simultaneous or sequential occupancy of their binding sites to repress sal (Walsh, 2007).

The requirement for two or more regulators to act together to control gene expression, i.e. combinatorial regulation, is fundamental to the generation of the great diversity of gene expression patterns by a finite set of transcription factors. Several previous studies have revealed the dual requirement for Hox and Smad functions for the activation of a target gene. Studies have suggested a general combinatorial mechanism for gene activation in which apparently separate transcriptional inputs act synergistically in gene activation and, in at least one case, the Hox response element and Dpp response element are separable. In this study, however, a requirement was observed for strict evolutionary conservation of the close topology of Hox and Smad binding sites in the sal CRE. It is suggested that collaboration is a distinct mode of combinatorial regulation in which two or more regulatory proteins must bind to nearby sites, but not necessarily to each other (Walsh, 2007).

The integration of Hox and Smad inputs could work through a number of possible mechanisms in the absence of direct physical interaction. One appealing possibility that might explain the requirement for the close proximity of binding sites is that Ubx and Mad-Med-Shn might interact with, and could therefore cooperatively recruit, the same co-repressor(s) for the repression of sal. Alternatively, if Mad-Med-Shn and Ubx bind sequentially to sal1.1, they might recruit different co-repressors and thereby orchestrate the assembly of a co-repressor complex. A third possibility is that because the Ubx and Mad/Med sites are embedded within a larger block of conserved regulatory DNA sequence in the sal1.1 CRE, the binding of other interacting transcription factors might also be involved in the repression of sal by Ubx and Mad-Med-Shn (Walsh, 2007).

These and recent results raise the question of whether collaboration is a general feature of target gene selection by Hox proteins. It is suggested that collaboration might be a widespread requirement for Hox function in vivo. This proposal is prompted by three observations: (1) Hox proteins alone have low DNA-binding specificity; (2) some, and perhaps all, Hox proteins might act as both repressors and activators; (3) Hox proteins regulate a great diversity of target genes that are also regulated by other transcription factors. In order to be such versatile regulators, it would be too great a constraint to require that Hox proteins always interact cooperatively with the diverse repertoire of transcription factors with which they act. Indeed, it may be argued that too much weight has been ascribed to the cooperative binding of Hox proteins and co-factors to DNA (Walsh, 2007).

Previously, much attention has focused on Exd and Hth, which interact with Hox proteins and bind cooperatively to DNA, thereby increasing Hox DNA-binding selectivity. However, it was only recently shown that the binding of these complexes alone was not sufficient to regulate target gene expression. Rather, Hox-Exd-Hth collaborate with and require the segmentation proteins Slp and En to repress the target gene Dll. This study has shown that the Exd- and Hth-independent target gene repression of sal requires collaboration between Ubx and Mad-Med-Shn. Although still a tiny sample of target genes, cases of transcription factors of various structural types acting as collaborators with Hox proteins are now available. The picture of Hox proteins relying on dedicated interacting co-factors such as Exd and Hth is expanding to a larger pool of collaborating transcription factors that modulate target gene selection (Walsh, 2007).

Indeed, collaboration might be the key to another unresolved mystery of the Hox proteins - the regulation of Hox protein activity. Some Hox proteins appear to act in both gene activation and repression; this is certainly the case for Ubx. This versatility would appear to be crucial to their role as sculptors of major features of body patterns, but how does the same transcription factor act positively in some contexts but negatively in others? There is evidence to suggest that the identity of the collaborating proteins and/or CRE sequences determines the 'sign' of Hox action (Walsh, 2007).

For instance, there is no evidence that the mere binding of Hox-Exd-Hth to a site determines the sign of Hox activity. These co-factors are involved in both Hox target gene activation (e.g., dpp in the midgut) and target repression (e.g.,Dll in the embryonic abdomen). But, in the latter case, En and Slp, two proteins that each harbor motifs for interaction with the co-repressor Groucho, are required collaborators for Dll repression. The roles of En and Slp in this instance might not be so much a matter of facilitating Hox target selection, but rather in regulating the sign of the output of the collaboration (Walsh, 2007).

Similar to the Hox proteins, the Smads can either activate or repress target genes. Furthermore, it has been demonstrated that the topology of Smad binding sites on DNA appears to be critical for determining whether a target gene is activated or repressed. In Drosophila, the topology of Mad and Med binding sites is critical for the recruitment of the co-repressor Shn. The recruitment of Shn was shown here to be necessary for sal repression. These two examples suggest that the positive or negative regulatory activity of a Hox protein depends on the context of surrounding binding sites and how they influence the activity of collaborating factors (Walsh, 2007).

The dependence of Hox proteins upon co-factors and collaborators indicates that, at the molecular level, Hox proteins are not 'master' regulatory proteins that dictate how target genes behave. Rather, they exert their great influence by virtue of their simple binding specificity, broad domains of expression and versatile, collaborative properties (Walsh, 2007).

A conserved activation element in BMP signaling during Drosophila development

The TGF-β family member Decapentaplegic (Dpp) is a key regulator of patterning and growth in Drosophila development. Previous studies have identified a short DNA motif called the silencer element (SE), which recruits a trimeric Smad complex and the repressor Schnurri to downregulate target enhancers upon Dpp signaling. The minimal enhancer of the dad gene was isolated, and a short motif was discovered that was termed the activating element (AE). The AE is similar to the SE and recruits the Smad proteins via a conserved mechanism. However, the AE and SE differ at important nucleotide positions. As a consequence, the AE does not recruit Schnurri but rather integrates repressive input by the default repressor Brinker and activating input by the Smad signal transducers Mothers against Dpp (Mad) and Medea via competitive DNA binding. The AE allows the identification of hitherto unknown direct Dpp targets and is functionally conserved in vertebrates (Weiss, 2010).

A 520-bp fragment was discovered within the second intron of the dad gene. This fragment induces an expression pattern very similar to that of the endogenous dad gene in embryonic, larval and adult tissues and contains evolutionarily conserved and largely overlapping binding sites for Smad and Brk proteins within a short sequence element that was called activating element (AE). The Smad and Brk proteins bind in a competitive manner to the AE, a mechanism similarly proposed for zen and Ubx enhancer elements. By precise targeted mutations, Brk binding was selectively abolished, and it was possible to unlink Smad and Brk input. Notably, the AE assembles a high-affinity trimeric complex of full-length Mad and Medea proteins. In Drosophila, such complexes have so far only been demonstrated for a so-called silencer element (SE) (Pyrowolakis, 2004). Therefore, this study presents the first example of such complex formation on a short-sequence element in the context of a gene activated by Dpp (Weiss, 2010).

The AE very closely resembles the SE, but despite their analogy, AE and SE differ in several key aspects. Because of the arrangement of the Smad binding sites, they are both able to recruit a complex of Mad and Medea. However, only the SE includes the second thymidine, which is essential for the recruitment of the repressor Shn (Pyrowolakis, 2004). Furthermore, the AE identified in the dad enhancer is able to interact with the Brk repressor. Brk competes with Mad for binding to the AE, which fulfills the consensus sequence derived from analysis of the SE with regard to Mad binding (GRCGNC) as well as the sequence for Brk binding (TGGCGYY). In contrast, Brk does not bind to the SEs described (Pyrowolakis, 2004). Thus, the AE and SE use a very similar sequence to exert opposite effects. These results provide a striking molecular scenario for Dpp signaling readout, based on the assembly of a trimeric Smad complex and its recruitment of a corepressor (Shn) or its competition with a dedicated repressor of the pathway (Brk) (Weiss, 2010).

Medea: Biological Overview | Evolutionary Homologs | Protein Interactions | 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.