Trithorax-like


REGULATION

Targets of Activity and Protein Interactions (part 2/2)

Other Trithorax-like functions and targets

Trl sites are found in the promoters of engrailed, even-skipped, E74, fushi-tarazu, Krüppel, Ultrabithorax, hsp26, hsp70, and histone H3 and H4 genes (Soeller, 1993). TRL binds specifically to GA/CT-rich (GAGA) sequences in many promoters where it activates transcription by locally altering chromatin structure.

The proximal promoter of the Krüppel (Kr) gene consists of a 44-base pair fragment containing the RNA start sites with significant promoter activity. This minimal promoter is flanked both upstream and downstream by binding sites for the GAGA factor. The GAGA factor is the predominant sequence-specific DNA binding factor that interacts with the Kr promoter region, and the purified protein activates Kr transcription in vitro. However, strong transcriptional activation of Kr as well as of Ultrabithorax, another GAGA factor-responsive gene, requires the presence of a DNA binding transcriptional repressor. The GAGA factor is able to relieve this repression in a binding site-dependent manner, and, thus, the GAGA factor appears to function as an antirepressor, rather than an activator, of the Kr gene (Kerrigan, 1991).

Insulator DNAs and promoter competition regulate enhancer-promoter interactions within complex genetic loci. The type 1 ftz promoter contains TATA but lacks the downstream promoter element (Dpe), whereas type 2 promoters contain initiator (Inr) and/or Dpe sequences but lack TATA. Some enhancers, such as ftz autoregulatory enhancer (AE1) , preferentially activate type 1 promoters when given a choice between linked type 1 and type 2 promoters. Others, such as the rhomboid (rho) neuroectoderm enhancer (NEE), promiscuously activate both classes of promoters (Ohtsuki, 1998 and references).

A transgenic embryo assay was used to obtain evidence that the Drosophila eve promoter possesses an insulator activity that can be uncoupled from the core elements that mediate competition. The type 1 even-skipped (eve) promoter contains an optimal TATA element and a GAGA sequence. The eve promoter insulator activity can be uncoupled from the TATA, Inr, and Dpe core elements. Mutations in a GAGA element, located between TATA and the transcription start site, impair this insulator activity, so that genes residing 5' from an otherwise normal eve promoter are now activated by a 3' enhancer. Similar results were obtained in trithorax-like (trl) mutants that diminish the levels of the Trl protein. Mutations in the GAGA element do not diminish eve promoter function in competition assays. It is suggested that the Trl protein-GAGA element traps distal enhancers by stabilizing enhancer-promoter interaction (Ohtsuki, 1998).

Promoters that possess enhancer blocking activities should facilitate the orderly trafficking of cis-regulatory elements. For example, eve stripe enhancers located 3' from the transcription unit should be unable to interact with neighboring genes located 5' from eve. Similarly, the ftz promoter contains a GAGA element located 5' to TATA. This configuration of core elements should allow the ftz promoter to be both transcriptionally active and able to block distal enhancers. Perhaps the ftz promoter helps inhibit interactions between 3' Antp enhancers and 5' homeotic genes [Dfd (Deformed) and Scr] within the ANT-C. It is conceivable that many promoters possess an intrinsic enhancer blocking activity. Inspection of ~250 Drosophila promoter sequences reveals that ~15% contain at least one optimal GAGA element within 50 bp 5' of the transcription start site. An earlier analysis of one of these promoters, 1-tubulin, indicates that a GAGA sequence element helps insulate tubulin expression from position effects. The enhancer blocking activity of the eve promoter appears to be mediated by interactions of Trl with a GAGA element. Trl has been shown to recruit the NURF protein complex, which facilitates the binding of upstream activators or core polymerase II components by decondensing chromatin. Trl-GAGA might trap distal enhancers by increasing the stability of enhancer-promoter interactions through the creation of an open chromatin configuration or by increasing the occupancy of core Pol II components such as TFIID (Ohtsuki, 1998).

Genetic control elements are usually situated in local regions of chromatin that are hypersensitive to structural probes such as DNase I. Binding of the GAGA transcription factor on existing nucleosomes leads to nucleosome disruption, DNase I hypersensitivity at the TATA box and heat-shock elements, and rearrangement of adjacent nucleosomes. ATP hydrolysis facilitates this process, suggesting that an energy-dependent pathway is involved in chromatin remodeling (Tsukiyama, 1994).

The action of GAGA in activating genes is carried out in a protein array involving other proteins, including an ATPase (Wall, 1995).

Activation of the Torso RTK at the poles of the embryo sets off a phosphorylation cascade that leads to the spatially specific transcription of the tailless (tll) gene. The TOR response element (TOR-RE) in the tll promoter indicates that the key activity modulated by the tor RTK pathway is a repressor present throughout the embryo. The TOR-RE has been mapped to an 11-bp sequence. The proteins GAGA and NTF-1 (also known as Elf-1, product of the grainyhead gene) bind to the TOR-RE. NTF-1 can be phosphorylated by MAPK (mitogen-activated protein kinase). tll expression is expanded in embryos lacking maternal NTF-1 activity. These results make NTF-1 a likely target for modulation by the TOR RTK pathway in vivo. Thus activation of the TOR RTK at the poles of the embryos leads to inactivation of the repressor and therefore, to transcriptional activation (by activators present throughout the embryo) of the tll gene at the poles of the embryo (Liaw, 1995).

The hsp26 gene has a segment of GA repeats (GAGA repeats) at -135 to -85 required for full heat shock inducibility This (GA)n element appears to contribute to formation of the wild-type chromatin structure of hsp26, an organized nucleosome array that leaves the HSEs in nucleosome-free sites. There is an additional (GA)n element at -347 to -341, adjacent to the distal HSE. This distal (GA)n element (-347 to -341) contributes, along with the proximal (GA)n element (-135 to -85), and the two HSEs to the formation of the chromatin structure and to heat shock inducibility. These (GA)n repeats are associated with a nonhistone protein(s) in vivo and are bound by a purified GAGA factor, in vitro. The HSEs are required for heat-inducible expression but play only a minor role in establishing the chromatin structure of the transgenes. Prior to heat shock, these HSEs appear to be free of protein. This suggests that GAGA factor and heat shock factor play distinct roles in gene regulation: the GAGA factor establishes and/or maintains the DH sites prior to heat shock induction, while the activated heat shock factor recognizes and binds HSEs located within the nucleosome free sites to trigger transcription (Lu, 1993).

Sequence-specific transcription factors need to gain access to regulatory sequences in chromatin. Previous studies utilizing model systems have suggested many mechanisms involved in this process. It is unclear however how these findings relate to natural promoters. The Drosophila Alcohol dehydrogenase (Adh) gene distal promoter is organized into an ordered nucleosome array before multiple transcription factors recognize their sites within this nucleosomal context and activate transcription. A purified in vitro system was used to study the binding of the ubiquitous Drosophila transcription factors Adh transcription factor 1 (Adf-1) and GAGA factor to the Adh distal promoter in chromatin. Several nucleosome core particles were assembled on 150 bp DNA fragments containing the Adh distal cis-acting elements in the natural promoter context but possessing different DNA-histone environments. The Adh distal promoter regulatory sequences can position nucleosomes in the same rotational setting as observed in vivo. In one particular nucleosome position, the wrapping of the Adf-1 and adjacent GAGA factor binding sites around the histone octamer creates a unique local DNA conformation. High-affinity but non-cooperative nucleosome binding of Adf-1 and GAGA factor therefore occurs, in contrast to the inhibition of Adf-1 and GAGA factor binding in other nucleosome positions. Thus, local histone-DNA sequence contact giving rise to a specific asymmetric nucleosome structure may play important roles in modulating the affinities of transcription factors for their nucleosomal sites (Gao, 1998).

There are two modes of regulation of the hsp26 promoter: one for promoters before heat shock, and another for promoters after heat shock. Efficient heat shock induction of Drosophila hsp26 gene transcription in vivo requires binding sites for heat shock factor (HSF) and GAGA factor (GAF) close to the TATA box (proximal elements) as well as 350 bp upstream of the start site of transcription (distal elements). Transcription in extracts from unstressed embryos rely solely on GAGA elements which efficiently counteract repression by abundant non-specific DNA-binding proteins. Transcription in extracts from heat shocked embryos depends only a little on GAGA elements, relying mainly on functional HSEs. These two modes of regulation in vitro may correspond to the two functional states of the promoter before and after heat shock in vivo (Sandaltzopoulos, 1995).

While GAGA binding sites are typically composed of 3.5 GA repeats, the Drosophila hsp70 gene contains much smaller elements, some of which are as little as three bases (GAG) in length. Interestingly, the binding of GAGA protein to more distant trinucleotide elements is relatively strong and not appreciably affected by the removal of larger GA arrays in the promoter. Moreover, a simple synthetic GAG sequence is sufficient to bind GAGA protein in vitro. The affinity of GAGA protein for different sequence elements has been directly compared by immunoprecipitation and gel mobility shift analysis. Measures of the concentration of GAGA protein in vivo indicate that it is a highly abundant nuclear protein, prevalent enough to occupy a sizable fraction of correspondingly abundant trinucleotide sites (Wilkins, 1998).

A number of activators are known to increase transcription by RNA polymerase (pol) II through protein acetylation. While the physiological substrates for those acetylases are poorly defined, possible targets include general transcription factors, activator proteins and histones. Using a cell-free system to reconstitute chromatin with increased histone acetylation levels, a direct test was performed for a causal role of histone acetylation in transcription by RNA pol II. Chromatin, containing either control or acetylated histones, was reconstituted to comparable nucleosome densities and characterized by electron microscopy after psoralen cross-linking, as well as by in vitro transcription. Chromatin was reconstituted using histones from either TSA-treated CV-1 cells, which accumulates hyperacetylated histone isoforms, or from untreated cells, in which the histones are primarily non-acetylated onto a 7.75 kb plasmid containing an hsp26 minigene. This process involves the prior depletion of the endogenous histones present in the chromatin assembly extract such that chromatin is assembled quantitatively from the input, exogenous histones. The chromatin assembly reaction generates complex chromatin containing many non-histone proteins and enzymatic activities. While H1-containing control chromatin severely represses transcription of a model hsp26 gene, highly acetylated chromatin is significantly less repressive. Acetylation of histones, and particularly of histone H4, affects transcription at the level of initiation (Nightingale, 1998).

The regulatory elements of the hsp26 promoter are well-known from in vivo and in vitro studies. A proximal regulatory element includes the TATA box, proximal heat shock element (HSE) and an adjacent GAGA element, while a distal regulatory site corresponds to the distal HSE and GAGA binding sites. Monitoring the ability of the transcription machinery to associate with the promoter in chromatin, it was found that Heat shock factor, a crucial regulator of heat shock gene transcription, profits most from histone acetylation. Templates with mutated hsp26 promoters were assembled into control and acetylated chromatin and analysed for their transcription potential. A template bearing only the TATA box supports a very low level of transcription, even in the absence of chromatin; there is no discernible transcription from chromatinized templates even after prolonged exposure. Addition of the proximal HSEs however, results in significantly increased transcription from both mock assembled and chromatinized templates, confirming the important role of the activator HSF. This minimal promoter, containing only proximal HSEs and the TATA box, clearly shows increased transcription from the acetylated template. The addition of GAGA elements to the promoter enhances transcription significantly, but to a similar degree on control and acetylated chromatin templates. Interestingly, the GAGA elements do not increase transcription in the mock assembled control, confirming that GAGA factor is involved in overcoming chromatin-mediated transcriptional repression, but by a mechanism that does not profit from histone acetylation. These results suggest that the HSE and TATA box are the significant sequence elements for the increased transcription observed in acetylated chromatin. Thus histone acetylation can modulate activator access to their target sites in chromatin, and provide a causal link between histone acetylation and enhanced transcription initiation of RNA pol II in chromatin (Nightingale, 1998).

The 5' untranslated region of the gene encoding the H(+)-ATP synthase beta subunit is extremely short, and reveals the presence of multiple initiation sites of transcription. The promoters of D. melanogaster and D. virilis H(+)-ATP synthase beta-subunit genes contain a conserved region surrounding the initiation transcription sites. Nucleotide sequence analysis has revealed the absence of canonical TATA and CCAAT boxes and the presence of several putative regulatory elements in both promoter regions, including GAGA, GATA and Ets binding sites (Pena, 1995).

The alpha 1-tubulin gene has two upstream regions, tubulin element 1 and tubulin element 2, and a downstream region that activates transcription. Two of these regions bind the GAGA transcription factor and act ubiquitously to insulate the transcription process from position effects and to activate transcription (O'Donnell, 1994)

The laminin B2 chain contains a 2.1-kb of 5'-flanking DNA that contains a TATA box and two CAAT boxes. Other potential transcriptional regulatory sequences include: two reverse complementary cAMP response element sequences, two sequences that are homologous to the retinoic acid response element motifs of the mouse B1 gene, and sequences homologous to the binding sites for transcription factors dFRA and dJRA, zeste, and possibly GAGA. (Chi, 1991).

The Drosophila GAGA factor binds specifically to simple repeating d(GA.TC)n DNA sequences. These sequences are capable of forming triple-stranded DNA as well as other non-B-DNA conformations. GAGA binds to a d[CT(GA.TC)]22 intermolecular triplex with similar specificity and affinity as to a regular double-stranded B-form d(GA.TC)22 sequence. However, the interaction of GAGA with triplex DNA cannot stimulate transcription in vitro. The affinity of GAGA for triplexes of the purine motif, such as a d[AG(GA.TC)]22 intermolecular triplex, is significantly lower. The DNA binding domain of GAGA is sufficient for efficient binding to triplex DNA. Based on the reported solution structure of the complex of GAGA-DNA binding domain with double-stranded DNA, a model for its interaction with triplex DNA is proposed in which most of the protein-DNA contacts observed in duplex DNA are maintained, especially those occurring through the minor groove. The higher negative charge of the triplex is likely to also have an important contribution to both the specificity and affinity of the interaction (Jimenez-Garcia, 1998).

Co-operative DNA binding by GAGA transcription factor requires the conserved BTB/POZ domain and reorganizes promoter topology

The POZ domain is a conserved protein-protein interaction motif present in a variety of transcription factors involved in development, chromatin remodeling and human cancers. The role of the POZ domain of the GAGA transcription factor in promoter recognition has been examined. Natural target promoters for GAGA typically contain multiple GAGA-binding elements. The POZ domain mediates strong co-operative binding to multiple sites but inhibits binding to single sites. Promoters regulated by GAGA have been identified by in vivo as well as in vitro studies. The Ultrabithorax, fushi tarazu, hsp70 and evenskipped promoters were used to compare the binding of GAGA polypeptides. All these promoters are characterized by the presence of multiple GAGA-binding sites. DNase I footprinting experiments reveal a dramatic difference in DNA-binding properties between full-length GAGA and the polypeptides lacking the POZ domain. The GAGA elements on the natural promoters are bound efficiently by full-length GAGA but not by equal molar amounts of either deltaPOZ (lacking the POZ domain) or a construct possessing only the DNA binding domain (DBD). The amount of GAGA required to bind the multiple promoter elements is significantly lower (>4- to 12-fold, depending on the promoter) than that required to bind a single site, indicative of co-operative DNA binding. The spacing of the GAGA elements in these different promoters varies considerably. However, GAGA appears to be quite flexible and able to bind co-operatively to GAGA sites located at variable distances from each other. The hsp70 promoter is generally GA rich and, at increasing GAGA concentrations, the footprints start to spread and most of the promoter DNA is protected against digestion (Katsani, 1999).

In contrast to full-length GAGA, equal molar amounts of the deltaPOZ or DBD polypeptides fail to bind the GAGA target promoters significantly. On the Ubx, ftz and eve promoters, protection of a single GAGA site by deltaPOZ and DBD can be observed. As expected, these sites are the ones that most closely resemble the optimal GAGA-binding sequence. In these experiments, deltaPOZ and DBD fail to bind to the weaker GAGA sites. This indicates that POZ-mediated co-operativity increases the binding affinity for these sites by at least one order of magnitude. Together, these DNase I footprinting experiments demonstrate that efficient binding of GAGA to its natural target promoters depends critically on the presence of the POZ domain, in addition to the DBD (Katsani, 1999).

Thus, GAGA oligomerization increases binding specificity by selecting only promoters with multiple sites. Electron microscopy reveals that GAGA binds to multiple sites as a large oligomer and induces bending of the promoter DNA. These results indicate a novel DNA binding mode by GAGA, in which a large GAGA complex binds multiple GAGA elements that are spread out over a region of a few hundred base pairs. A model is proposed in which the promoter DNA is wrapped around a GAGA multimer in a conformation that may exclude normal nucleosome formation. Since the GAGA DBD clamps almost one turn of the DNA, GAGA binding to multiple sites within a nucleosome repeat length is expected to severely compromise histone-DNA contacts. These contacts might be hampered further by DNA bending and wrapping around a GAGA oligomer. However, it is not clear whether GAGA binding leads to complete displacement of the histone core or whether some histone-DNA contacts are preserved. In summary, after transient chromatin remodelling by NURF to allow for GAGA binding, GAGA may function as an architectural factor that reorganizes the promoter DNA and maintains it in an open conformation (Katsani, 1999).

GAGA Factor-dependent transcription and establishment of DNase hypersensitivity are independent and unrelated events in vivo

The ability of GAGA factor, a putative anti-repressor, to modulate transcription-related events in the absence or presence of a bona fide activator, the Adh transcription factor 1 (Adf-1) transcription factor, has been investigated using a Drosophila transgenic system. In contrast to previous in vitro and in vivo data linking the binding of GAGA factor to the acquisition of DNase hypersensitivity at heat shock promoters, inserting multiple GAGA binding motifs adjacent to a minimal alcohol dehydrogenase (Adh) promoter leads to strongly elevated embryonic transcription without creation of a promoter-associated DNase-hypersensitive (DH) site. Establishment of DNase hypersensitivity requires the presence of both GAGA and Adf-1 binding sites and is accompanied by a further, synergistic increase in transcription. Because Adf-1 is capable neither of establishing a DH site nor of promoting efficient transcription by itself in embryos, it is likely that DH site formation depends on a GAGA factor-mediated binding of Adf-1 to chromatin, perhaps facilitated by a locally remodeled downstream promoter region. More generally it is suggested that GAGA factor-binding sequences may operate in a promoter-specific context, with transcriptional activation, polymerase pausing, and/or DH site formation critically dependent on the nature of the sequences (and their binding partners) linked in cis (Pile, 2000).

Using a Drosophila transgenic system an examination was performed of the ability of GAGA factor, a putative anti-repressor, to modulate transcription-related events in the absence or presence of a bona fide activator, the Adf-1 transcription factor. In contrast to previous in vitro and in vivo data linking the binding of GAGA factor to the acquisition of DNase hypersensitivity at heat shock promoters, it was observed that inserting multiple GAGA binding motifs adjacent to a minimal alcohol dehydrogenase (Adh) promoter, normally a target of Adf-1, leads to strongly elevated embryonic transcription without creation of a promoter-associated DNase-hypersensitive (DH) site. Establishment of DNase hypersensitivity requires the presence of both GAGA and Adf-1 binding sites and is accompanied by a further, synergistic increase in transcription. Because Adf-1 is capable neither of establishing a DH site nor of promoting efficient transcription by itself in embryos, it is likely that DH site formation depends on a GAGA factor-mediated binding of Adf-1 to chromatin, perhaps facilitated by a locally remodeled downstream promoter region. Furthermore, it is suggested that GAGA factor-binding sequences may operate in a promoter-specific context, with transcriptional activation, polymerase pausing, and/or DH site formation critically dependent on the nature of the sequences (and their binding partners) linked in cis (Pile, 2000).

A cellular memory module conveys epigenetic inheritance of hedgehog expression during Drosophila wing imaginal disc development

In Drosophila, the Trithorax-group (trxG) and Polycomb-group (PcG) proteins interact with chromosomal elements, termed Cellular Memory Modules (CMMs). By modifying chromatin, this ensures a stable heritable maintenance of the transcriptional state of developmental regulators, like the homeotic genes, that is defined embryonically. It was asked whether such CMMs could also control expression of genes involved in patterning imaginal discs during larval development. The results demonstrate that expression of the hedgehog gene, once activated, is maintained by a CMM. In addition, the experiments indicate that the switching of such CMMs to an active state during larval stages, in contrast to embryonic stages, may require specific trans-activators. Thes results suggest that the patterning of cells in particular developmental fields in the imaginal discs does not only rely on external cues from morphogens, but also depends on the previous history of the cells, as the control by CMMs ensures a preformatted gene expression pattern (Maurange, 2002).

Immunoprecipitation using cross-linked chromatin (XChIP) allows the mapping of in vivo DNA target sites of chromatin proteins. Because one Polycomb (PC, a member of the PcG) binding site on polytene chromosomes coincides with the cytological position of hh at 94E, this method was applied to ask whether there are PC and GAGA factor (GAF/Trl, a member of the trxG) binding sites in the hh genomic region. These two factors had previously been found to be hallmarks of CMMs, and the GAF has been shown to be associated with some PcG complexes and necessary for the silencing function of PREs. Initially the immunoprecipitated material was hybridized to a genomic stretch of 45 kb encompassing the hh gene. This led to the identification of PC/GAF-binding sites in regions close to the transcription unit. To further fine-map the location of the PC/GAF-binding sites, the region around the hh gene was subdivided into 1-kb-sized PCR fragments (from 4 kb upstream of the hh transcription start site according to the transcript CG4637 from Flybase, to 13.4 kb downstream to the end of the gene). Slot-blot hybridizations of immunoprecipitated material revealed two main sites where PC and GAF are strongly enriched. The first site (A) is located in a region between 0.07 and 1.06 kb upstream of the transcription start site, whereas the second binding site (B) is found in a region spanning the second exon of the hh gene and spreading about 0.4 kb on both sides of the exon. On both sites a substantial overlap was observed between PC- and GAF-binding sites. The presence of this particular arrangement of PC- and GAF-binding sites in the hh genomic region suggests that these PcG and trxG proteins directly control hh expression (Maurange, 2002).

To investigate this at the functional level, the accessibility of the hh promoter region to a trans-activating factor was assessed. It is known that a PRE placed in the vicinity of an Upstream Activating Sequence (UAS) is able to counteract GAL4 binding, preventing expression of the reporter gene (Zink, 1995; Fitzgerald, 2001). Advantage was taken of the availability of an EP line possessing a UAS site close to the endogenous hh transcription start site to test whether the hh-PREs could inhibit the activation of transcription induced by GAL4. The EP3521 line (termed here EP-hh) possesses an EP transposon containing several UAS sites, and is inserted in the hh promoter region (-0.36 kb). The endogenous hh gene is not transcribed in salivary glands. By using an hs-GAL4 line, which is known to be leaky at 25°C, weak expression of GAL4 in larval salivary glands is observed. When hs-GAL4 is crossed to a line containing UAS-hh integrated randomly in the genome, in situ stainings reveal that at 25°C, by the action of GAL4, the hh mRNA is present in high amounts in all the salivary gland cells. However, when hs-GAL4 is crossed to the EP-hh line, in which the UAS sites are juxtaposed to the presumptive PRE, hh transcription was observed in only a very few cells situated mainly at the base of the glands. It was reasoned, because in most cells transcription is inhibited, that the PcG proteins binding the PREs in the vicinity of the hh promoter block the accessibility of GAL4 to the UAS sites. Accordingly, reducing the amount of some of the PcG proteins in the cells by repeating the experiment with flies heterozygous for the Pc3 allele or with males hemizygous for the ph409 allele induces partial derepression of transcription of the endogenous hh gene in a substantial number of gland cells. These results indicate that the repression observed in most of the salivary gland cells in the EP line is caused by the action of the PcG proteins through their binding to the identified PREs. As such, these experiments demonstrate that the transcription of hh is directly repressed by the PcG proteins (Maurange, 2002).

When UAS-en is misexpressed at the D-V boundary in a wild-type genetic background using vg-GAL4, it induces hh expression in most of the cells of the wing pouch except in a stripe along the A-P boundary where hh seems to be repressed. Whereas UAS-en is strongly misexpressed at the D-V boundary, the endogenous en gene is weakly misactivated in some cells of the anterior wing pouch (Maurange, 2002).

Repeating the same experiment in a genetic background hemizygous mutant for an hypomorphic allele of polyhomeotic (ph409) leads to a broader domain of expression of hh. Remarkably, the region along the A-P boundary seems to be less refractory to activation of hh transcription, given that the territory of the repressed domain is reduced. Endogenous en is itself overexpressed in the anterior compartment. This is consistent with the findings demonstrating that en expression can be derepressed in a PcG gene mutant background. In this case in the anterior wing pouch cells, the activation of en transcription by Hh is probably more efficient than in a wild-type background because en cannot be correctly silenced by PH (Maurange, 2002).

The same experiment repeated in a genetic background now doubly heterozygote for the trxG genes trithorax (trxE2) and brahma (brm2) consistently shows that hh expression is activated at the D-V boundary, but can hardly be maintained through cell divisions in the anterior compartment, because with in situ staining, the Hh signal progressively fades away from the D-V boundary. As expected, in such a case, en expression in the anterior compartment is restricted to the D-V boundary, because Hh might not be present in a sufficient amount to activate transcription of the endogenous en gene in the subsequent wing pouch cells (Maurange, 2002).

Furthermore, it is known that PcG-mediated silencing is enhanced at higher temperature, and this hyperrepressed state can be inherited through cell divisions. Based on these observations, it was reasoned that raising embryos at 28°C instead of 18°C would make the Pc-mediated silencing more difficult to derepress, and influence the activation of hh transcription by En. vg-GAL4; UAS-en embryos were allowed to develop at 28°C until the beginning of second instar larvae, when the D-V boundary is established in wing discs and UAS-en is expressed there. As expected, stainings on third instar imaginal discs reveal ectopic clones of wing pouch cells expressing hh. However, the frequency of cells expressing hh is lower than in discs of larvae grown at 18°C, indicating that the Pc-mediated silencing was harder to erase at 28°C. Nevertheless, in contrast with trxG mutant flies, once the transcription has initially been activated in this case, it is maintained in the subsequent daughter cells as suggested by the presence of clones spreading from the D-V midline to the limits of the wing pouch (Maurange, 2002).

These experiments demonstrate that once initiated by En, the maintenance of the transcriptional state of hh to the daughter cells can be attributed to the action of the PcG and trxG proteins. It is concluded that the CMM activity of the hh upstream region described in the transgenic assay is also efficient when considered in its natural chromatin environment and is responsible for the inheritance of the initial transcriptional state of hh from the initiation to the completion of the wing pouch development (Maurange, 2002).

Identification and characterization of polyhomeotic PREs and TREs

The polyhomeotic (ph) gene is a member of the Polycomb group of genes (Pc-G), that are required for the maintenance of the spatial expression pattern of homeotic genes. In contrast to homeotic genes, ph is ubiquitously expressed and it is quantitatively regulated. ph is negatively regulated by the Pc-G genes, except Psc, and positively regulated by the antagonist trithorax group of genes (trx-G), suggesting that Pc-G and trx-G response elements (PREs and TREs) exist at the ph locus. In this study, PREs and TREs at the ph locus that function in transgenic constructs have been functionally characterized. A strong PRE and TRE has been identified in the ph proximal unit as well as a weak one in the ph distal unit. The PRE/TRE of both ph units appear atypical compared with the well-defined homeotic maintenance elements because the minimal ph proximal response element activity requires at least 2 kb of sequence and does not work at long range. Chromatin immunoprecipitation experiments on cultured cells and embryos have been used to show that Pc-G proteins are located in restricted regions, close to the ph promoters, that overlap functionally defined PRE/TREs. The data suggest that ph PRE/TREs are cis-acting DNA elements that modulate rather than silence Pc-G- and trx-G-mediated regulation, enlarging the role of these two groups of genes in transcriptional regulation (Bloyer, 2003).

Detailed analysis of the php PRE/TRE shows that this element is modular, and contains at least three regions of differential sensitivity to Pc-G and trx-G mutations. Recently, it has been proposed that homeotic 'maintenance elements'(MEs) are composed of several small DNA modules that are bound by several subsets of Pc-G and TRX-G proteins. In other PREs and MEs, the PSR is separable from silencing modules. Pc-G-dependent silencing and PSR are always linked in the ph PRE: both silencing and PSR are lost when the minimal 2-kb P{C4-812} PRE/TRE is dissected into several subfragments. It was not possible to separate PSR from silencing of white, but this may be because these studies lacked sufficient resolution (Bloyer, 2003).

There is not a perfect correspondence between sites of Pc-G binding as detected by ChIP, and functional activity of transgenes. Peak-binding for Pc, Ph, Psc, and GAF mapped to a 0.9-kb fragment which showed no silencing activity and no sensitivity to Pc-G and trx-G mutations when tested in isolation in the P{C4-811} transgenic lines. In both cells and embryos, Pc-G and GAGA factor (GAF) binding overlaps the conserved regulatory sequences found in both the php and phd PREs. However, neither of these conserved sequences were sufficient for PRE activity in these assays. The P{C4-827} and P{C4-824} fragments overlap, and both contain the upstream 350-bp conserved region, but neither demonstrates PRE activity. Similarly, P{C4-811} that contains the downstream conserved region also lacks PRE activity. It is not known why these differences exist; but it is suggested that proteins not assayed in the ChIP experiments, and/or unconserved sequences, must contribute to PRE function. In addition, there were differences between binding of Pc-G proteins in cultured cells and in embryos, which may reflect the different physiology of developing embryos versus cultured cells or experimental variation of the technique (Bloyer, 2003).

The isolated ph PRE/TREs do not completely reproduce the ph wild type regulation even if most of the Pc-G and trx-G response elements are conserved in the minimal P{C4-812} PRE/TRE. An opposite eye color phenotype was observed in a Sce and Pc mutant background when the endogenous ph regulation was compared with transgenic lines containing isolated ph PRE/TREs. It may be simply that the effects of Pc-G mutations on ph regulation are indirect. It may be that different Pc-G complexes function differently in the endogenous and exogenous chromosomal locations. Using an in vivo-functional assay it has been shown that ph and Psc may have a different silencing function than Pc and Sce. Consistent with this, in this study, ph and Psc consistently show strong genetic effects on ph regulation. The results are in accordance with the idea that ph/Psc in one case and Pc/Sce in the other may play different roles in Pc-G silencing complexes, and that the function of these complexes may depend on chromosomal context. An alternative, but not necessarily mutually exclusive explanation for the opposing response of php PRE/TRE transgenes and the endogenous ph locus to Pc-G mutations might be that the isolated PRE/TREs lack distant cis-DNA regulatory elements that can act on ph wild type regulation within the endogenous chromosomal context. In this model, activity of the php PRE/TRE would be modulated by the distant cis-regulatory element, so that transgenes, lacking this sequence would behave differently. These hypotheses cannot be distinguished with the current data (Bloyer, 2003).

Identification in vivo of different rate-limiting steps associated with transcriptional activators in the presence and absence of a GAGA element

The impact of a GAGA element (the target of GAGA factor) was analyzed on a transgenic promoter in Drosophila that was activated by proteins composed of the Teton DNA binding domain and either the heat shock factor (HSF) activation domain or a potent subdomain of VP16. Permanganate footprinting was used to monitor polymerase II (Pol II) on the transgenic promoters in vivo. Activation by Teton-HSF but not by Teton-VP16A2 required the GAGA element; this correlated with the ability of the GAGA element to establish a paused Pol II. Although the GAGA element was not required for activation by Teton-VP16A2, the GAGA element greatly accelerated the rate of activation. The permanganate data also provided evidence that Pol II encountered different rate-limiting steps, following initiation in the presence of Teton-HSF and Teton-VP16A2. The rate-limiting step in the presence of Teton-HSF was release of Pol II paused about 20 to 40 nucleotides downstream from the start site. The rate-limiting step in the presence of Teton-VP16A2 occurred much closer to the transcription start site. Several biochemical studies have provided evidence for a structural transition shortly after Pol II initiates transcription. The behavior of Pol II in the presence of Teton-VP16A2 provides the first evidence that this transition occurs in vivo (Wang, 2005).

To learn about how the GAGA element contributes to transcription, a transgenic system was developed that allowed comparison of the function of different activators on promoters that contain or lack a GAGA element. For the present study, an activator containing amino acids 610 to 691 from HSF was compared to an activator containing three copies of a 13-amino-acid array derived from VP16. Amino acids 610 to 691 from Drosophila HSF fused to the Gal4 DNA binding domain (DBD) had previously been shown to activate transcription of a promoter located on extrachromosomal DNA. This information has been used to understand the mechanism of activation by HSF, yet it is not clear whether this information has relevance to activation in the context of the GAGA element or a paused Pol II. The three tandem repeats of the N-terminal portion of the VP16 activation domain fused to the tetracycline repressor (TET) of Escherichia coli had previously been shown to be a potent activator of a chromosomal copy of gene in human cells. It was of interest to know whether the GAGA element would have any impact on such a potent activation domain (Wang, 2005).

Each activation domain was fused to a derivative of the DBD from the tetracycline repressor that could be induced to bind DNA upon addition of the tetracycline analog, doxycycline. This provided the opportunity to analyze the effect of the GAGA element on promoter activity under steady-state conditions and on the rate of transcriptional activation. The latter was of interest because one function of the GAGA element suggested by the heat shock genes is that it provides a means for rapidly inducing a gene (Wang, 2005).

A transgenic system was developed to investigate the impact of a GAGA element on the activity of two transcriptional activation domains. This transgenic system is unique in that it allowed examinination of both steady-state and kinetic features of transcriptional activation in vivo. The use of permanganate footprinting allowed interrogation of the behavior of the Pol II on the DNA at a level not readily available to other methods, such as chromatin immunoprecipitation or nuclear run-on. Certainly, neither of these latter methods would have detected the high density of Pol II that is evident immediately downstream from the transcription start site in the presence of Teton-VP16A2. It is anticipated that further application of this overall experimental approach will provide significant insight into transcriptional control mechanisms in living cells (Wang, 2005).

This study was motivated by two main questions. First, would a small GAGA element be sufficient to establish a paused Pol II in the absence of an activator? This question is important because GAGA elements are found associated with a broad spectrum of promoters, making it important to learn if the presence of a GAGA element correlates with promoter proximal pausing. Second, what impact does the GAGA element have on activation by different activators? This information could help in the understanding of what purpose the GAGA element serves in regulating gene expression. The HSF activation domain, which had been defined by transient transfection analyses, was examined to determine if this activation domain depends on the GAGA element. The VP16 activation domain was examined because it was anticipated that its strength might obviate the requirement for the GAGA element. In addressing these two questions, convincing evidence was obtained in vivo that Pol II encounters different rate-limiting steps after initiating transcription in the presence of these two activators (Wang, 2005).

Trithorax-like interaction with SAP18, a Sin3-associated polypeptide

Drosophila SAP18 (accepted FlyBase name: Bicoid interacting protein 1), a polypeptide associated with the Sin3-HDAC co-repressor complex, has been identified in a yeast two-hybrid screen as capable of interacting with the Drosophila GAGA factor. The interaction was confirmed in vitro by glutathione S-transferase pull-down assays using recombinant proteins and crude SL2 nuclear extracts. The first 245 residues of GAGA, including the POZ domain, are necessary and sufficient to bind dSAP18. In polytene chromosomes, Drosophila SAP18 and GAGA co-localize at a few discrete sites and, in particular, at the bithorax complex where GAGA binds some silenced polycomb response elements. When the Drosophila SAP18 dose is reduced, flies heterozygous for the GAGA mutation Trl67 show the homeotic transformation of segment A6 into A5, indicating that GAGA-dSAP18 interaction contributes to the functional regulation of the iab-6 element of the bithorax complex. These results suggest that, through recruitment of the Sin3-HDAC complex, GAGA might contribute to the regulation of homeotic gene expression (Espinas, 2000).

The identity of dSAP18 with either human (hSAP18) or C. elegans (cSAP18) SAP18 is high, ~60% and 47%, respectively. The three polypeptides show high homology throughout their sequences, except for the most N- (1-15) and C-terminal (138-150) residues, and a central region (residues 32-45). Two specific regions, RI (16-31) and RII (65-89), show a very high degree of conservation with a similarity >80%. A third region, RIII (123-137), also shows significant similarity (80%), but in this case the identity is lower (47%) than for regions RI (81%) and RII (67%) (Espinas, 2000).

GAGA is organized into several functionally distinct domains. A single zinc finger is involved in nucleic acid recognition. In addition to this central DNA binding domain (DBD), GAGA carries a C-terminal glutamine-rich domain (Q-domain), which is involved in transcription activation, and a highly conserved N-terminal POZ domain, which mediates protein-protein interactions. A relatively long (140 amino acids) region of unknown function(s) links the POZ and DBD domains. Little is known about the interaction of GAGA with other nuclear proteins. The POZ domain of GAGA has been shown to support homomeric as well as heteromeric interactions with other POZ-containing proteins, such as tramtrack (ttk). The first 245 residues of GAGA are necessary and sufficient for binding dSAP18, and efficient GAGA-dSAP18 interaction requires the contribution of both the POZ and linking domains of GAGA. Not all regions of dSAP18 contribute equally to its interaction with GAGA, and residues 73-113, which include most of the highly conserved RII region, are mainly responsible for binding to GAGA (Espinas, 2000).

SAP18 was identified as a polypeptide associated with the mammalian transcriptional repressor Sin3. The core mSin3 complex contains a total of seven polypeptides, which include the histone deacetylases HDAC1 and HDAC2, RbAp48 and RbAp46, and SAP30 and SAP18. Recruitment of the Sin3-HDAC complex to specific target genes appears to rely on its interaction with sequence-specific DNA binding proteins, since none of the known components of the complex are capable of binding DNA. SAP30 and SAP18 could mediate some of these interactions. mSAP30 binds both mSin3 and N-CoR, and is required for N-CoR-mediated repression by a set of sequence-specific DNA-binding transcription factors. SAP18 could also be involved in interactions with sequence-specific DNA binding proteins. It is known that SAP18 interacts directly with mSin3. Since POZ is a highly conserved structural domain, it is likely that similar interactions would be observed with other POZ domains. Actually, several sequence-specific transcriptional repressors carry POZ domains, some of which are also found to interact with N-CoR and SMRT. Most likely, formation of a stable complex requires multiple interactions between its various components (Espinas, 2000 and references therein).

Contrary to most POZ-containing proteins, the Drosophila GAGA factor acts as a transcriptional activator. Its interaction with a component of the Sin3 co-repressor complex indicates that GAGA might also act as a repressor in some cases. In this respect, the presence of GAGA at some silenced PREs of the bithorax and antennapedia complexes might be especially revealing. Interestingly, though the immunostaining patterns of GAGA and dSAP18 show only a limited general overlapping in polytene chromosomes, the two proteins co-localize at the region of the bithorax complex (BX-C), suggesting a possible contribution of GAGA-dSAP18 interaction to BX-C regulation. Consistent with this possibility, a genetic interaction between Trl and a deficiency that uncovers dSAP18 has been observed. Flies heterozygous for the Trl67 mutation and hemizygous for Df(3R)sbd26 show a homeotic transformation of the sixth abdominal segment into the fifth as indicated by the presence in the sixth sternite of several bristles in the vast majority of the individuals. Results presented indicate that GAGA-dSAP18 interaction has a significant contribution to the functional regulation of the iab-6 element of BX-C (Espinas, 2000).

The concurrent presence of GAGA and Polycomb at some silenced PREs is surprising since, as derived from genetic analysis, these two proteins are expected to have opposing functions on the regulation of the expression of the homeotic genes. Acting at the level of the core promoter elements, GAGA is likely to activate transcription of the homeotic genes. However, functional trithorax response elements (TREs) are frequently found in the vicinity of PREs and a contribution of GAGA to the functional regulation of several segment-specific cis-regulatory regions of the bithorax complex has been reported (Espinas, 2000 and references therein).

It is still uncertain whether GAGA helps to establish the repressed or the active state of these elements. GAGA has been shown to contribute to the relief of repression at the Fab-7 element, and the homeotic transformations described here and elsewhere are also consistent with a role in activation. However, in the case of the iab-7 and bxd PREs, GAGA has been shown to contribute to silencing and the genetic interactions observed between some Pc and Trl alleles also suggest a contribution to repression. The results presented here indicate that GAGA might participate in the recruitment of the Sin3-HDAC co-repressor complex to some PREs, but that contrary to what would be anticipated for such an interaction, it contributes to the relief of repression at the iab-6 element. The same phenotype is observed in flies homozygous for the hypomorph Trl13C allele. Interestingly, some rpd3 alleles behave as enhancers of PEV, also leading to an increase in repression. It is possible that by modifying chromatin structure, GAGA-SAP18 interaction could contribute to the establishment of the domain boundaries that insulate different cis-regulatory elements, rather than to the formation of the repressed or active states themselves (Espinas, 2000).

Trithorax-like interaction with NURF301

To explore additional functions for NURF and NURF301, an analysis of proteins that interact with NURF was undertaken by identifying coimmunoprecipitating polypeptides in embryo nuclear extract fractions. Peptide sequencing of a 70 kDa species among numerous polypeptides that were substoichiometric with respect to NURF subunits reveal an unambiguous match with the embryonically expressed isoform of the GAGA transcription factor, GAGA519. A control experiment using beads coated with a nonspecific antibody showed no coimmunoprecipitation of GAGA factor. The association of GAGA factor and NURF was confirmed in binding assays. Binding to recombinant NURF complex was observed for S35-labeled GAGA519 and the alternatively spliced GAGA581 isoform, which is expressed at later stages in Drosophila development (Xiao, 2001).

By systematic deletions, the NURF-interacting sequences common to both GAGA519 and GAGA581 were determined to be in a conserved region containing the DNA binding zinc finger and flanking sequences. This finding, which is consistent with a study of the GAGA factor regions necessary for chromatin remodeling, is clearly distinguished from known interactions involving the multimerization (POZ) and transactivating (glutamine-rich) domains of GAGA factor and is also reminiscent of interactions between the zinc finger of mammalian ELKF and the SWI/SNF complex. In addition to the interactions with the GAGA factor, the NURF complex binds to S35-labeled GAL4-HSF and GAL4-VP16 activators containing the GAL4 DNA binding domain fused to strong activation regions of HSF and VP16. Binding was specific for HSF and VP16, since little or no binding was detected for the GAL4(1-147) DNA binding domain (Xiao, 2001).

To define region(s) of NURF301 that interact with GAGA factor, HSF, and VP16, the binding of GST-NURF301 segments to S35-labeled activators was analyzed in GST-pull-down assays. Two regions on NURF301 responsible for GAGA factor binding are also nucleosome binding regions. Whether the interaction of NURF301 with GAGA factor or the nucleosome is mutually exclusive is presently unknown. A separate region of NURF301 is responsible for binding GAL4-HSF and GAL4-VP16. Consistent with these results, the DeltaN301NURF complex binds to GAGA factor less efficiently (~26% of wild-type) but still binds to GAL4-HSF and GAL4-VP16. The binding of other NURF subunits to activators was also examined. ISWI, NURF55, and NURF38 show little or no binding to S35-labeled HSF and VP16. Interestingly, ISWI, but not NURF55 and NURF38, shows binding to GAGA factor, albeit at a level lower than binding to NURF (Xiao, 2001).

Trithorax-like interaction with Pipsqueak

Polycomb group proteins act through Polycomb group response elements (PREs) to maintain silencing at homeotic loci. The minimal 1.5-kb bithoraxoid (bxd) PRE of Ultrabithorax contains a region required for pairing-sensitive repression and flanking regions required for maintenance of embryonic silencing. Little is known about the identity of specific sequences necessary for function of the flanking regions. Using gel mobility shift analysis, DNA binding activities have been identified that interact specifically with a multipartite 70-bp fragment (MHS-70) downstream of the pairing-sensitive sequence. Deletion of MHS-70 in the context of a 5.1-kb bxd Polycomb group response element derepresses maintenance of silencing in embryos. A partially purified binding activity requires multiple, nonoverlapping d(GA)(3) repeats for MHS-70 binding in vitro. Mutation of d(GA)(3) repeats within MHS-70 in the context of the 5.1-kb bxd PRE destabilizes maintenance of silencing in a subset of cells in vivo but gives weaker derepression than deletion of MHS-70. These results suggest that d(GA)(3) repeats are important for silencing but that other sequences within MHS-70 also contribute to silencing. Antibody supershift assays and Western analyses show that distinct isoforms of Polyhomeotic and two proteins that recognize d(GA)(3) repeats, the Trl/GAGA factor and Pipsqueak (Psq), are present in the MHS-70 binding activity. Mutations in Trl and psq enhance homeotic phenotypes of ph, indicating that Trl/GAGA factor and Psq are enhancers of Polycomb that have sequence-specific DNA binding activity. These studies demonstrate that site-specific recognition of the bxd PRE by d(GA)(n) repeat binding activities mediates PcG-dependent silencing (Hodgson, 2001).

The results of the sequence-specific analysis suggest that d(GA)n-specific binding factors are present in a complex defined by electrophoretic mobility, termed complex 2. Therefore, antibodies directed against two nuclear factors that bind d(GA)n sequences, Trl/GAF and Psq, were tested in binding reactions with a bxd fragment termed MHS-70. In the presence of increasing amounts of antibody to Trl/GAF, the mobility of complex 2 was significantly retarded, migrating close to the sample well. Antibodies to Psq caused a modest but detectable retardation of complex 2. Neither of these antibodies alters the mobility of a second complex, complex 1. These results show that complex 2 contains detectable levels of Trl/GAF and Psq. The significantly reduced mobility of complex 2 in the presence of anti-TRL/GAF antibody presumably results from the ability to induce multimeric aggregates of DNA-TRL/GAF complexes. To show that the complexes formed by MHS-70 and the competing oligomers are equivalent, the formation of complex 2 with synthetic oligomers was tested with antibodies to Trl/GAF, Ph, and Psq. Antibodies to Ph, Trl/GAF, and Psq supershift complex 2 in synthetic oligomer binding reactions (Hodgson, 2001).

Polyhomeotic proximal (Php), Trl/GAF, and Psq have multiple isoforms. To determine which of these isoforms are potential components of complex 2, Western analysis of the Separose AS, BR0.6 and Q0.15 fractions was undertaken. Q0.15 is enriched for isoforms of Trl/GAF P67 plus Trl/GAF P54, Php105 plus Php64, and Psq P70. Taken together with the antibody supershift analysis, these results show that the distinct isoforms of Trl/GAF, Php and Psq coelute with complex 2 and suggest that these isoforms constitute potential subunits of complex 2 binding activities. It has been shown that the full-length isoform of PhP, Php-170, coimmunoprecipitates with the PcG proteins Pc, Psc, Su(z)2, and Scm. Western analyses of the three fractions described above show that there are no detectable levels of Pc, Su(z)2, Psc, or Sex Combs on Midleg (Scm) in Q0.15, indicating that these PcG proteins do not coelute with complex 1 and 2 binding activities. These results suggest that the complex 2 activity is a novel PcG activity containing distinct isoforms of PHP, TRL/GAF, and Psq (Hodgson, 2001).

Trl null or hypomorphic alleles (Trl13C, Trl62, and Trl85) do not affect maintenance of silencing in vivo by bxd5.1 UbxlacZ. Similarly, embryos mutant for psqRF13 (deletion of psq) and Df(2R)psq-lolaDelta18 (deletion of psq, termed psqlola hereafter) show wild-type bxd5.1 UbxlacZ silencing. One potential reason for these results is that maternally deposited Trl or psq protein or mRNA rescue the effects of absence of zygotic proteins on embryonic silencing (Hodgson, 2001).

Genetic interactions between PcG genes are monitored by the enhancement of PcG mutations, providing a sensitive genetic assay for genes required in PcG-mediated silencing. Therefore, the ability of Trl and Psq mutations to enhance the extra sex combs phenotype of ph was tested. Trl enhances the extra sex combs phenotype of Pc. Similarly, Trl62 enhances the extra sex combs phenotype of ph2 and ph409. The effects of psqlola-Delta18 and psq2403 on enhancement of ph2 and ph409 were tested. There is strong enhancement of the expressivity of the extra sex combs phenotype. These results are consistent with a role for Trl/GAF and Psq in PcG-mediated silencing of homeotic loci and indicate that Trl/GAF and Psq are enhancers of PC that have sequence-specific DNA binding activity (Hodgson, 2001).

The Drosophila transcription factor tramtrack (TTK) interacts with Trithorax-like (GAGA) and represses GAGA-mediated activation

This study reports the interaction of the Drosophila transcription factors Trithorax-like (GAGA) and Tramtrack (TTK). This interaction is documented both in vitro, through GST pull-down assays, as well as in vivo, in yeast and Schneider S2 cells. GAGA and TTK share in common the presence of an N-terminal POZ/BTB domain, found to be necessary and sufficient for GAGA-TTK interaction. Structural models that could account for this interaction are discussed. GAGA is known to activate the expression of many genes in Drosophila. In contrast, TTK has been proposed to act as a maternally provided repressor of several pair-rule genes, such as even-skipped (eve). As with many Drosophila genes, eve contains at its promoter region binding sites for GAGA and TTK. Transient expression experiments show that GAGA activates transcription from the eve stripe 2 promoter element, and TTK inhibits this GAGA-dependent activation. Repression by TTK of the eve promoter requires its activation by GAGA and depends on the presence of the POZ/BTB domains of TTK and GAGA. These results indicate that GAGA-TTK interaction contributes to the regulation of gene expression in Drosophila (Pagans, 2002; full text of article).

What sort of molecular interactions could account for the formation of GAGA–TTK hetero-oligomers? The POZ/BTB domain, that mediates GAGA–TTK interaction, is also responsible for the formation of GAGA–GAGA and TTK–TTK homo-oligomers. Since POZ/BTB is a highly conserved domain, structural comparative models of those present in GAGA and TTK could be built from the crystal structure of the POZ/BTB dimer of PLZF. Actually, the modelled structures of putative GAGA–GAGA and TTK–TTK homo-dimers bear a similar hydrophobic dimerisation interface as in the PLZF dimer, with a slightly lower complementarity. All the contacts involved in dimerisation are conserved and the central cavity is even more hydrophobic than in PLZF. These results strongly suggest that homomeric GAGA–GAGA and TTK–TTK interactions are likely to involve similar molecular interactions as those observed in the PLZF dimer. Moreover, similar models could also be built for heteromeric GAGA–TTK interaction, indicating that a putative GAGA–TTK hetero-dimer would be stabilised by the same interactions involved in homomeric GAGA–GAGA and TTK–TTK interactions. The extensive dimerisation interface suggests, however, that POZ/BTB-containing proteins are obligated homo-dimers. Actually, there is no experimental evidence indicating that either GAGA or TTK could ever exist as monomers in solution. A putative GAGA–TTK hetero-dimer could, therefore, arise from the respective homo-dimers by swapping of the POZ/BTB domains. However, the extension and high hydrophobic character of the contacts involved in these interactions raise the question of whether such a domain-swapping model could actually account for heteromeric GAGA–TTK interaction. In this respect, it must be noted that the crystal structure of the POZ/BTB domain of PLZF was solved independently by two different groups and, in spite of belonging to different crystallographic space groups, both structures showed conserved dimer–dimer interactions, involving the extension of the ß1/ß5'-sheet of one dimer towards the same region of the next one, keeping an antiparallel orientation. Eight hydrogen bonds connect main-chain atoms of the ß1 chains of each dimer. Each dimer buries 700 Å2 in this interface, with a high complementarity score. To address the question of whether these dimer–dimer interactions are also conserved in the POZ/BTB domains of GAGA and TTK, GAGA and TTK homo-tetramers were modelled, as well as a GAGA–TTK hetero-tetramer. The modelled tetramers have an interacting surface and complementarity similar to those of the PLZF homo-tetramer. The GAGA homo-tetramer and the GAGA–TTK hetero-tetramer keep all the eight hydrogen bonds between the ß1/ß5'-sheets, but the TTK homo-tetramer may lack those at the ends. These considerations raise the possibility that, rather than the formation of GAGA–TTK hetero-dimers, GAGA– TTK interaction would actually involve dimer–dimer interactions through the ß1/ß5'-sheets. GAGA–TTK interaction appears to be specific since, as it was found earlier, TTK does not interact in vitro with the POZ/BTB domains of either ZID or ZF5. Interestingly, the ß5-strand of the ß1/ß5' motif is fully identical in GAGA and TTK, a feature not shared by any of the other POZ/BTB-containing proteins analysed. In the crystal of the POZ/BTB domain of PLZF, this dimer–dimer interaction propagates through the lattice, suggesting that it could give rise to the formation of multimers of higher stoichiometry. Actually, bacterially expressed TTK and GAGA, but not ΔPOZGAGA, form in vitro oligomers of high apparent M. In contrast, dimer stability itself strongly relies on the ß1/ß5'-sheets which, in PLZF, involve the formation of five main chain H-bonds between the ß1 chain of one monomer and the ß5' chain of the second. The only other contact that contributes to dimer formation involves the symmetry related α1' helices of each monomer. Therefore, the ß1 residues that could mediate dimer–dimer interactions are themselves essential for dimerisation as it was found by mutational analysis of the POZ/BTB domain of PLZF (Pagans, 2002).

Pipsqueak and GAGA factor act in concert as partners at homeotic and many other loci

The Drosophila GAGA factor Trithorax like controls transcription and other chromosome functions by altering chromatin structure. A second GAGA-binding protein of Drosophila, Pipsqueak (Psq), can directly bind to Trithorax like and is associated with Trl in vivo. Genetic interaction studies provide evidence that Psq and Trl act together in the transcriptional activation and silencing of homeotic genes. A complete colocalization of Psq and Trl on polytene interphase chromosomes and mitotic chromosomes suggests that the two proteins cooperate as general partners not only at homeotic loci, but also at hundreds of other chromosomal sites (Schwendemann, 2002).

To identify chromosomal loci that are targets of both Trl and Psq, salivary gland polytene chromosomes were doubly immunostained with anti-Psq and anti-Trl antibodies. Surprisingly, the staining patterns obtained with the two antibodies appeared to be identical. Like Trl, Psq binds to hundreds of loci on the polytene chromosomes. Analysis of the binding patterns by confocal microscopy confirmed that virtually every signal derived with Trl antibody coincides with a signal derived with Psq antibody and vice versa. Occasionally, single sites appear to be stained by only one of the two antibodies. However, these sites vary between different chromosome preparations, and it is therefore believed that they do not represent binding sites truly specific for only one of the two proteins. Western analyses show that the overlapping staining patterns are not caused by cross-reaction of Psq antibody with Trl protein or Trl antibody with Psq protein (Schwendemann, 2002).

One possibility to explain the colocalization of Trl and Psq is that GAGA-binding sites are recognized in vivo by a protein complex that contains both proteins. Therefore protein extracts were prepared from salivary gland nuclei and immunoprecipitation assays were performed with the Psq antibody. This antibody efficiently coimmunoprecipitates Psq and Trl from these extracts. To test whether this effect is specific, whether two other transcription factors, dAP-4 and BR-C, were precipitated by the antibody was also analyzed. Both proteins are known to be expressed in salivary glands, from which the nuclear extract was derived. In addition, like Psq and Trl, all isoforms encoded by the BR-C contain a BTB protein interaction domain. However, neither dAP-4 nor BR-C was found to be precipitated by the Psq antibody. It is concluded that the interaction between Psq and Trl is specific (Schwendemann, 2002).

It was next asked whether the association of Psq and Trl in vivo might be due to direct binding of the two proteins to one another. To address this question, histidine (his)-tagged full-length and truncated Psq proteins were expressed in bacteria, and the ability of these proteins to retain the Trl519 isoform on a Ni2-NTA column was tested. Full-length Psq (Psq1064), but not polypeptides including only the Psq domain (Psq221) or lacking the BTB domain (Psq942), were able to efficiently bind Trl519. When Psq1064 was successively truncated from the C terminus, all resulting polypeptides were able to bind Trl519, including Psq166, which essentially consists of the BTB domain. Because previous studies have shown that the BTB domain forms a protein-protein interaction interface, this result suggests that binding is mediated by the BTB domains of both proteins. Therefore a GST fusion of the Trl BTB domain was expressed in bacteria (GST-Trl116) and tested to see whether this polypeptide can be coimmunoprecipitated with Psq331, a polypeptide that consists of the N-terminal third of Psq, which includes the BTB domain. As a control, whether Trl116 can be coimmunoprecipitated with Psq221 was tested. Psq331, but not Psq221, is bound by Trl116. Taken together, these results suggest that Psq and Trl are associated with one another in vivo through direct binding mediated by their BTB domains. The BTB domain of Trl has been shown to also mediate self-oligomerization of Trl that leads to formation of large protein complexes in vitro. These results suggest that Psq is a partner of Trl in similar complexes formed in vivo. Because both Trl and Psq have been shown to bind GA-rich sequences, it is likely that both proteins contribute to the cooperative binding of multiple GAGA elements through these complexes. The complete colocalization of Trl and Psq at hundreds of chromosomal loci predicts that, in general, Psq and Trl act as partners. Consistent with this notion, the bxd PRE of Ubx (Hodgson, 2001) and the MCP silencer of Abdominal B (Abd-B) (S. Sakonju, personal communication to Schwendemann, 2002) emerge as first examples of specific loci where Psq and Trl appear to interact (Schwendemann, 2002).

To further test the model that Psq and Trl are partners, it was asked whether Psq is also bound to the GA-rich satellite DNA of the centromeric heterochromatin, which is occupied by Trl during mitosis. When mitotic chromosomes from larval brains are stained with Trl antibodies, strong fluorescent signals are observed in the pericentric regions. A similar staining of the pericentric regions is observed when chromosomes are stained with the Psq antibody. Double staining experiments show that this staining cannot be distinguished from the staining pattern obtained with the Trl antibody. As on the polytene chromosomes, there thus seems to be a complete overlap of the Trl and Psq binding sites. Taken together, the results of the polytene and mitotic chromosome staining strongly suggest that Psq and Trl are general partners that not only share common functions in the control of target genes at euchromatic sites, but also in heterochromatin organization and mitosis (Schwendemann, 2002).

The genetic, biochemical, and cytogenetic data presented in this study strongly suggest that Psq and Trl act together as partners in the control of homeotic and many other genes. Since both Trl and Psq are encoded by essential genes, the functions of these proteins are not redundant. Psq may even be an obligatory partner of Trl. Future studies on Trl function will therefore have to include this partner, and may thus provide novel insights into the mechanism of action of this important chromatin factor. The finding that Psq is a member of a larger family of DNA-binding proteins, which includes many Drosophila BTB proteins with previously unknown DNA-binding activity, may help to further elucidate the composition and function of Psq/Trl-containing protein complexes (Schwendemann, 2002)

Trithorax-like interaction with Lola like

Epigenetic inheritance to maintain the expression state of the genome is essential during development. In Drosophila, the cis regulatory elements, called the Polycomb Response Elements (PREs) function to mark the epigenetic cellular memory of the corresponding genomic region with the help of PcG and trxG proteins. While the PcG genes code for the repressor proteins, the trxG genes encode activator proteins. The observations that some proteins may function both as PcG and trxG members and that both these groups of proteins act upon common cis elements, indicate at least a partial functional overlap among these proteins. Trl-GAGA was initially identified as a trxG member but later was shown to be essential for PcG function on several PREs. In order to understand how Trl-GAGA functions in PcG context, the interactors of this protein were sought. lola like, aka batman, was identified as a strong interactor of GAGA factor in a yeast two-hybrid screen. lolal also interacts with polyhomeotic and, like Trl, both lolal and ph are needed for iab-7PRE mediated pairing dependent silencing of mini-white transgene. These observations suggest a possible mechanism for how Trl-GAGA plays a role in maintaining the repressed state of target genes involving lolal, which may function as a mediator to recruit PcG complexes (Mishra, 2003).

The BTB/POZ domains of several proteins have been reported to mediate homo- and hetero-dimerizations and can be functionally swapped between two proteins. It was reasoned that GAGA may carry out its diverse roles by recruiting different proteins to the target DNA sequences and the BTB domain may mediate this. A yeast two-hybrid screen using the BTB domain of GAGA, amino acids 1-245, was used as bait. The 0-16 h embryo cDNA library was screened to identify potential partners for Trl-GAGA protein. Out of the 46 adenine auxotrophs, 23 strongly activated the second reporter gene, ß-galactosidase. All the 23 positives interacted specifically with the bait upon retransformation and 14 turned out to be one single gene lola like (lolal). All the clones had the complete coding region of the gene. LOLAL also interacts with the full length GAGA protein. Quantitative assays show that the interaction of LOLAL with BTB and GAGA full-length protein is comparable. Indeed, in an independent screen using full length GAGA protein, LOLAL appears eight times. Other BTB domain containing proteins were isolated multiple times and one already reported interactor of Trl-GAGA, Sin3A. GAGA factor itself was not isolated in these screens, which may suggest that BTB domain of this protein might not form a homodimer. When tested in a directed two hybrid assay, though, BTB-BTB interaction was seen. It has been reported that Trl-GAGA exists in several isoforms. All the experiments reported here have been done with the major isoform, GAGA 519. It is not known if other isoforms can interact differently. Furthermore, pipsqueak, which binds to GAGA sequences and contains BTB domain has also been shown to be essential for sequence specific targeting of PcG protein complex. It is unclear, however, how Trl-GAGA is placed in this multi-component recruitment system working on GAGA motif of PREs. Further studies will be needed to answer these questions (Mishra, 2003).

lolal was originally identified in the Drosophila gene disruption project. Later on this mutation was also found to enhance the homeotic phenotype of polyhomeotic, ph, and renamed batman (Faucheux, 2003). lolal encodes a protein of 127 amino acids that contains a BTB domain of about 90 amino acids, leaving only few residues at both ends of the protein for any other functional motif/domain. Another PcG protein, Esc, consists almost entirely of six WD40 repeat motifs. Unlike multi-domain proteins, the ones made of a single domain alone may function as adaptor modules to bring together two different molecules/complexes. Trl-GAGA is known to activate transcription of several genes. In this context, lolal may function to inhibit this activation role of Trl-GAGA, in a way similar to MyoD-Id interaction. Further studies will be required to differentiate between these mechanisms (Mishra, 2003).

One of the key steps in the PcG/trxG mediated maintenance is the recruitment of the multi-protein complex of correct composition onto the PRE. Recent studies have shown that more than one or perhaps several recruiting processes take place in concert. It is likely that different recruitment possibilities provide the necessary variation that is needed for the establishment and maintenance of varying transcriptional states at hundreds of different loci. These studies identify a new member in this process. Trl-GAGA bound to specific sites on the PREs recruits LOLAL, which in turn, through direct or indirect means, incorporates Polyhomeotic (Ph) into the complex. This raises a question whether the other recruiting agents like pho, zeste, and others, function in cooperation or competition with each other. Also, it is not clear if the complex is assembled de novo on the PREs, a pre-assembled complex is recruited or partly assembled sub-complexes are recruited. Since large complexes of PcG proteins can be isolated, it is concluded that such structures, once assembled are stable. It is not clear though if these complexes are stable during cell division or they assemble each time a cell divides (Mishra, 2003).

Since most of the PREs contain Trl-GAGA binding sites, it is likely that at least some PcG complexes are recruited by GAGA factor through its direct interaction with LOLAL. These findings establish a molecular link between Trl-GAGA and the PcG complex. Since Trl-GAGA functions in several other processes that do not seem to be directly linked to the PRE function, it is likely that there are several interactors of Trl-GAGA. In the study of such interactors, initial observations suggest that a large number of proteins can interact with Trl-GAGA with a potential to target a repressive or activator function to different loci. It is not clear though how the selection of an appropriate partner is made. Is it in the context in which Trl-GAGA is bound or different heterodimers preexist in the nucleus and these are then recruited to appropriate loci? New assays will have to be designed to appropriately answer these questions (Mishra, 2003).

Genetic interaction studies show that lolal interacts with a variety of PcG and trxG mutations,. This underscores the important role of this protein in the regulation of developmental genes. Interestingly, lolal interactions with ph mutation leads to transformation of 2nd (and some times 3rd) leg to 1st leg, an apparent anteriorization type of homeotic transformation in thoracic but in the abdominal region same combination leads to posteriorization type of homeotic transformation, pigmentation of A4 (A4-->A5) reduction in the size of A6 (A6-->A7). However, appearance of sex comb in 2nd and 3rd legs is also known to be due to derepression of Scr in posterior segments thereby explaining this phenotype as due to loss of the repression function of the PcG proteins. Furthermore, trxG and PcG mutations upon interaction with lolal can give a similar phenotype. In lolal context, Asx and trg both show partial A6-->A5 transformation in the abdominal region. Pc is involved in pairing dependent silencing complex recruited by iab-7PRE. ph is also involved in the PS function of iab-7PRE. While it was known that lolal enhances the homeotic phenotype of ph, it is demonstrated that both ph and lolal are involved in establishing the repressive complex at the iab-7PRE. This indicates that lolal and ph function in coordination to set up a repressive complex. Taken together, these results suggest that lolal may be acting along with Trl-GAGA or with other partners in different complexes in a locus or stage specific manner. Depending on the context it could be an activator or repressor function. Since not only 'ON' or 'OFF' but also several 'levels of expression states' for a given hox gene or indeed other regulated loci are maintained, it is likely that a unique combination of trxG and PcG proteins may be needed for each varying level of expression state of a given locus (Mishra, 2003).

Affinity pull-down experiments show that GAGA, Lolal, Ph and Pc proteins coexist in a complex. This is also in agreement with the genetic interaction studies, where interaction was found of lolal with ph and Pc. Genetic and biochemical studies also suggest that, like Trl, lolal could not specify the kind of complex to be assembled. It is likely, therefore, that specificity of the GAGA partner, PcG or trxG member, does not come from lolal. It is even possible that lolal could also be a multifunctional adapter of Trl-GAGA in assembling multi protein complexes. The specificity may come from yet another factor or even from the transcriptional activity around the locus. Recent observations that transcription process itself may contribute to the cellular memory may support this view. This might bring together the ability of GAGA factor to support transcription and recruitment of multi-protein complexes and nucleosome remodeling activity in one mechanistic context (Mishra, 2003).

Trl-GAGA has been suggested to be involved in creating a nucleosome free region. The first step in establishing a PcG complex may be this Trl-GAGA mediated nucleosome remodeling of the chromatin on the PRE region to create a more accessible DNase I hypersensitive site. The recruitment of a protein complex to the accessible region may take place through GAGA factor or by other factors that can anchor the complex onto DNA. As the next step Lolal could mediate recruitment of initial complex, for example, Esc-E(z) protein complex, which modifies nearby histone tails to covalently mark the region for the recruitment of another complex, like PRC1. Consistent histone modifications and remodeling may be needed to maintain the chromatin conformation. These studies would place Lolal as the factor binding to the DNA bound GAGA even when the rest of the complex is not recruited and therefore serves to help in subsequent recruitment steps. In this context, the exact function of proteins like Lolal becomes very important. Further studies will be required to clarify these issues (Mishra, 2003).

Drosophila FACT contributes to Hox gene expression through physical and functional interactions with GAGA factor

Chromatin structure plays a critical role in the regulation of transcription. Drosophila GAGA factor directs chromatin remodeling to its binding sites. Drosophila FACT (facilitates chromatin transcription), a heterodimer of dSPT16 and dSSRP1, is associated with GAGA factor through its dSSRP1 subunit, binds to a nucleosome, and facilitates GAGA factor-directed chromatin remodeling. Moreover, genetic interactions between Trithorax-like encoding GAGA factor and spt16 implicate the GAGA factor-FACT complex in expression of Hox genes Ultrabithorax, Sex combs reduced, and Abdominal-B. Chromatin immunoprecipitation experiments indicate the presence of the GAGA factor-FACT complex in the regulatory regions of Ultrabithorax and Abdominal-B. These data illustrate a crucial role of FACT in the modulation of chromatin structure for the regulation of gene expression (Shimojima, 2003).

GAGA factor-dFACT complex was identified by co-immunoprecipitation with epitope tagged GAGA factor. GST pull-down assays show that GAGA factor makes a direct contact with dFACT through its dSSRP1 subunit. Gel electrophoresis mobility shift assays reveal that dFACT binds to the nucleosome. Furthermore, dFACT stimulates GAGA factor-directed chromatin remodeling in the embryonic extract of Drosophila. Based on these data, the following model is proposed for GAGA factor-directed site-specific chromatin remodeling. The GAGA factor-dFACT complex binds to a GAGAG sequence on DNA. dFACT binds to nucleosome and stimulates chromatin remodeling. This allows remodeling in a GAGA factor binding site-dependent manner. Because human FACT binds to histones H2A and H2B (Orphanides, 1999), and the yeast SPN complex enhances DNase I sensitivity of nucleosome in a region where H2A and H2B contact the DNA (Formosa, 2001), it is most likely that FACT binds to DNA at the entry and exit site of the nucleosome through its HMG subunit SSRP1, and then acts to destabilize and remove the H2A/H2B dimers to facilitate chromatin remodeling. However, the H2A/H2B dimers remain associated with the FACT-nucleosome complex through SPT16 such that they can quickly rebind to the H3/H4 tetramer when required. In support of this model, an acidic amino acid stretch found in histone-interacting proteins such as nucleoplasmin and NAP1 is conserved in the C-terminal tail of SPT16. Furthermore, H2B (and probably H2A) has been shown to turn over more rapidly than H3 and H4 during transcription (Shimojima, 2003).

There are many ATP-dependent chromatin remodeling factors. Which factor is responsible for the GAGA factor-dFACT complex-induced chromatin remodeling? Because an antibody against ISWI abolishs the GAGA factor-induced chromatin remodeling in the embryonic extract (Okada, 1998), remodeling factors containing ISWI as the catalytic subunit must play a role. Among them, at least NURF appears to be involved in the remodeling, because GAGA factor interacts directly with NURF (Xiao, 2001; Shimojima, 2003).

Although ISWI is essential for the expression of en and Ubx in imaginal discs, it has been suggested that ISWI is mainly involved in transcription repression in vivo. Specific acetylation of histone H4at Lys 16 by MOF counteracts the action of ISWI and leads to derepression of chromatin transcription. Interestingly, the yeast SPT16-Pob3 complex interacts with Sas3, a yeast homolog of MOF (John, 2001). It is possible that dFACT may also recruit MOF to shut out ISWI and induce a change from repression to activation. Under such conditions, remodeling factor(s) other than NURF may cooperate with the GAGA factor-dFACT complex. Brahma (BRM) remodeling complex may be a candidate for the replacer, but no genetic interaction between Trl and brm has been demonstrated. CHD1 may be another candidate because mouse, Drosophila, and yeast CHD1 have been reported to interact with SSRP1 or its yeast counterpart POB3 (Shimojima, 2003).

The regions of GAGA factor, dSPT16, and dSSRP1 have been identified that are responsible for interactions of these proteins with GAGA factor. GAGA factor interacts with dSSRP1 through the region containing the Zn-finger domain and its flanking sequences. This finding is consistent with the observation that both BTB/POZ and Q-rich domains are not required for the GAGA factor-induced chromatin remodeling in the embryonic extract. It also suggests that GAGA factor can form an oligomer through its BTB/POZ and/or Q-rich domains and bind cooperatively to clusters of its binding sites, just as does the GAGA factor-dFACT complex. The cooperative and stable binding of GAGA factor-dFACT complex to chromatin may be important for the epigenetic maintenance of the active state (Shimojima, 2003).

The GAGA factor-interacting region of dSSRP1 contains the HMG box and its N-terminal flanking sequence that overlaps with the dSPT16-binding region of dSSRP1. However, the presence of equimolar complex of GAGA factor, dSSRP1, and dSPT16 in the embryonic nuclear extract indicates that the overlapped region in dSSRP1 does not interfere with the simultaneous binding of GAGA factor and dSPT16. Interestingly, dSSRP1 binds to naked DNA, but addition of increasing amounts of dSPT16 decreases the DNA binding of dSSRP1 in a dose-dependent manner. This observation suggests that dSPT16 suppresses the binding of dSSRP1 to naked DNA through its interaction with the HMG box region without affecting the affinity for nucleosome (Shimojima, 2003).

The dSSRP1-binding sequence of dSPT16 was defined as the C-terminal highly conserved region. This is in good agreement with the observation that expression of the corresponding region of yeast SPT16 is required to rescue yeast temperature-sensitive mutants of spt16 (Shimojima, 2003 and references therein).

The most interesting finding in this study is the involvement of the GAGA factor-dFACT complex in the regulation of gene expression. The anterior transformation of T3 and A6 in Deltaspt16 Trl double heterozygotes and the binding of the GAGA factor-dFACT complex to the bxd region of Ubx and the iab-6 element of Abd-B in vivo indicate that the complex contributes to the epigenetic maintenance of Hox gene expression. Based on these data, the following scheme is envisioned for the maintenance of the active state. The GAGA factor-dFACT complex induces chromatin remodeling in the regulatory regions of various GAGA factor-dependent genes and potentiates transcription. Whereas the expression of ftz and hsp70 is transient, the active state is maintained in Hox genes such as Ubx, Scr, and Abd-B with the aid of other trx group gene products (Shimojima, 2003).

What is the mechanism underlying the maintenance? Among trx group proteins, BRM constitutes an SWI/SNF-type chromatin remodeling complex. This type of chromatin remodeler possesses a unique ability to act on condensed mitotic chromatin. A sequence-specific regulator, Zeste, has been shown to recruit the BRM complex to its target sites. Functionally distinct chromatin remodeling induced by the GAGA-dFACT and Zeste-BRM complexes may be important to keep the active state through many rounds of cell cycle. In addition to the GAGA factor-dFACT and the BRM complexes, three trx group protein complexes have been identified to date. One is TAC1 consisting of Trx, dCBP, and Sbf1, which acetylates core histones in nucleosomes. Mutations in trx or nejire encoding dCBP have been shown to reduce the expression of Ubx. The others are ASH1 and ASH2 complexes. ASH1 also has been known to interact directly with dCBP. These data suggest that acetylation of core histones or other proteins plays a crucial role in the maintenance of the active state. In support of this hypothesis, hyper-acetylation of H4 has been shown to be a heritable epigenetic mark of the active state. The finding that a counteracting Pc group complex ESC/E(Z) contains histone deacetylase RPD3 is also consistent with this hypothesis. Chromatin remodeling induced by the GAGA factor-d-FACT and the Zeste-BRM complexes might be essential for maintenance of the hyperacetylated state of H4 (Shimojima, 2003).

Drosophila GAGA factor directs histone H3.3 replacement that prevents the heterochromatin spreading

Epigenetic maintenance of the expression state of the genome is critical for development. Drosophila GAGA factor interacts with FACT and modulates chromatin structure for the maintenance of gene expression. This study shows that the GAGA factor-FACT complex (Fact is a heterodimer of dSPT16 and dSSRP1; Shimojima, 2003) and its binding site just downstream from the white gene are crucial for position effect variegation. Interestingly there is a dip of histone H3 Lys 9 methylation and a peak of H3 Lys 4 methylation at this site. The GAGA factor and FACT direct replacement of histone H3 by H3.3 through association of HIRA at this site, and maintain white expression under the heterochromatin environment. Based on these findings it is proposed that the GAGA factor and FACT-dependent replacement of Lys 9-methylated histone H3 by H3.3 counteracts the spreading of silent chromatin (Nakayama, 2007).

This study shows that the GAGA factor-FACT complex is present on the GAGA factor-binding DNase-hypersensitive site d1, a site just downstream from w, and participates in PEV. d1 appears to be a peculiar site where histone H3 K4 methylation peaks and H3 K9 methylation dips, and necessary and sufficient to counteract the heterochromatin spreading. The GAGA factor and FACT contribute to replacement of histone H3 by H3.3 through association of histone H3.3 chaperone HIRA to d1, and maintains w expression under the heterochromatin environment. Based on these data, the following model is proposed for the maintenance of the active state against the spreading of silent chromatin. Heterochromatin is marked by K9-methylated histone H3 and its binding protein HP1, and has a tendency to spread into neighboring regions. Histone H3.3 replacement is thought to be achieved through either eviction of a nucleosome and deposition of a H3.3-containing nucleosome or stepwise disassembly-reassembly without eviction of a nucleosome. Since the GAGA factor-FACT complex facilitates chromatin remodeling and the GAGA factor is known to generate a nucleosome-free region around its binding site, it is most likely that eviction or disassembly of a nucleosome occurs at the DNase-hypersensitive site of d1. The GAGA factor and FACT participate association of HIRA to d1 and the histone replacement would be accomplished by subsequent deposition or reassembly of a H3.3-containing nucleosome. This process would be repeated constantly to eliminate K9-methylated histone H3 at d1 and counteract the spreading of silent chromatin (Nakayama, 2007).

It has been reported that histone H3.3 replacement is triggered by transcription elongation. However, genome-wide profiling has shown histone H3.3 replacement from upstream of to downstream from transcription units. Although some of the replacement may be explained by elongation during intergenic transcription, the histone H3.3 replacement at d1 appears to occur independent of transcription elongation. Thus, the present study indicates a distinct pathway for histone H3.3 replacement (Nakayama, 2007).

Transcription of the w adjacent gene CG32795 has been reported to start immediately after the GAGA factor-binding sequence of d1, suggesting that d1 is a part of the promoter region of CG32795. Therefore, the effect was examined of Trl and spt16 mutations on expression of CG32795. The reduction of a single dose of Trl or spt16 affect the CG32795 expression in the wm4 context but not in the w+ context. These data are consistent with the idea that d1 is a functional promoter element of CG32795 in the w+ context, although Trl and spt16 become haplo-insufficient only when the accessibility of the GAGA factor-FACT complex to d1 decreased under the heterochromatin environment. This raises the possibility that the protection from heterochromatin spreading by the GAGA factor and FACT at d1 is a consequence of their function within the CG32795 promoter. However, conventional promoters do not have a barrier function against heterochromatin silencing. For example, the presence of GAL4 (or E2F) on a promoter carrying GAL4 (or E2F)-binding site did not modify PEV of the attached reporter gene. Genome-wide profiling of H3.3 replacement in Drosophila has revealed the clear dip of H3.3-containing nucleosomes at immediately upstream of the transcription start sites of active genes. This is in sharp contrast with the case of d1, where peaks were observed of both the H3.3/H3 ratio and the actual H3.3 level, and illuminates the difference between d1 and ordinary promoters. Furthermore, the GAGA factor-dependent histone H3.3 replacement was detected also at the DNase HS1 in the Fab-7 boundary of Abd-B, where no promoter activity has been demonstrated. These findings indicate that the GAGA factor and FACT-dependent histone H3.3 replacement can occur without promoter functions. Nevertheless, the barrier function could be assisted by the putative promoter activity of d1 such as formation of a transcription initiation complex (Nakayama, 2007).

The GAGA factor-binding sequence at d1 consists of (GA)8. Since GAGA factor forms an oligomer through its BTB domain, the factor can make a cooperative and stable binding to closely spaced GAGAG elements. This is presumably the reason why d1 gave a prominent signal among the GAGAG sequences around w in the ChIP assay. Because the GAGA factor occupies many closely spaced GAGAG sequences within the Drosophila genome including the Polycomb/trithorax response elements of Hox genes, the proposed mechanism may operate not only in loci juxtaposed with heterochromatin but also in other loci such as the regulatory regions of Hox genes. Indeed GAGA factor and FACT-dependent histone H3.3 replacement were observed in the Fab-7 boundary of Abd-B. High levels of histone H3.3 have been also reported at the locus control region of the chicken folate receptor gene, suggesting that the barrier function against the chromatin silencing via histone H3.3 replacement may be evolutionarily conserved up to vertebrates (Nakayama, 2007).

GAGA facilitates binding of Pleiohomeotic to a chromatinized Polycomb response element

Polycomb response elements (PREs) are chromosomal elements, typically comprising thousands of base pairs of poorly defined sequences that confer the maintenance of gene expression patterns by Polycomb group (PcG) repressors and trithorax group (trxG) activators. Genetic studies have indicated a synergistic requirement for the trxG protein GAGA and the PcG protein Pleiohomeotic (PHO) in silencing at several PREs. However, the molecular basis of this cooperation remains unknown. Using DNaseI footprinting analysis, a high-resolution map is provided of sites for the sequence-specific DNA-binding PcG protein PHO, trxG proteins GAGA and Zeste and the gap protein Hunchback (HB) on the 1.6 kb Ultrabithorax (Ubx) PRE. Although these binding elements are present throughout the PRE, they display clear patterns of clustering, suggestive of functional collaboration at the level of PRE binding. While GAGA can efficiently bind to a chromatinized PRE, PHO alone is incapable of binding to chromatin. However, PHO binding to chromatin, but not naked DNA, is strongly facilitated by GAGA, indicating interdependence between GAGA and PHO already at the level of PRE binding. These results provide a biochemical explanation for the in vivo cooperation between GAGA and PHO and suggest that PRE function involves the integrated activities of genetically antagonistic trxG and PcG proteins (Mahmoudi, 2003).

This study has determined the precise distribution within the Ubx PRE of the recognition elements for four sequence-specific DNA-binding proteins that have all been implicated in Ubx regulation in vivo: PcG protein PHO, gap protein HB and trxG proteins GAGA and Zeste. The results indicate that, rather than a random collection, the binding site distribution within the Ubx PRE reflects a functional arrangement, allowing cooperation between distinct PRE binding proteins. Of particular interest is the observation that chromatin binding by the PcG protein PHO is strongly facilitated by the trxG protein GAGA. This finding provides a molecular mechanism for the requirement for both factors during PRE-directed silencing in vivo, and suggests that PHO and GAGA elements together may form a functional module (Mahmoudi, 2003).

Several independent genetic studies have pointed to a concurrent requirement for GAGA and PHO during gene silencing directed by distinct PREs. The PcG-dependent silencing conferred by a 230 bp fragment of the iab-7 PRE is dependent on both GAGA and PHO binding. Similarly, a 138 bp fragment of the MCP silencer, which was found to be sufficient for maintenance of embryonic silencing, contains PHO and GAGA sites. Mutations in either PHO or GAGA sites compromised silencing and revealed cooperation between both proteins. Particularly relevant for the current study are results that support a critical role in PcG silencing for GAGA and PHO sites within the Ubx PRE (Mahmoudi, 2003).

Functional dissection of the Ubx PRE has revealed that a Pc-dependent PRE silencer is contained in the central 567 bp fragment from position 577 to 1143, which includes all PHO and the highest density of GAGA sites. Another study showed that an oligomerized subfragment, corresponding to positions 890-1079 within PRE C, harboring two PHO and five GAGA elements, is able to confer PcG silencing in vivo. Finally, deletion of a 160 bp region corresponding to positions 851-1011 within PRE C impairs maintenance of silencing. The large extent of overlap between the DNA fragments identified in these independent studies strongly suggests that the common region within PRE C represents the critical core of the Ubx PRE. The most noticeable feature of this region is the many alternating GAGA and PHO binding elements. Moreover, it is of interest to note that footprinting analysis revealed the presence of Zeste as well as HB sites within this region, which may also contribute to the in vivo maintenance of repression (Mahmoudi, 2003).

The identification of Zeste as a component of the PRC1 PcG complex, suggests that it may play a direct role in PcG complex recruitment to the Ubx PRE. Further evidence for the involvement of Zeste in the maintenance of Ubx repression as well as activation has been provided by transgene experiments. Finally, the presence of HB sites within the Ubx PRE suggests a potential role for HB, not only during the initiation of Ubx repression, but also during the transition from establishment to maintenance. One attractive possibility is that this transition involves dMi-2 recruitment by HB. It should be noted that in the absence of initiating activation and repression elements, HB-independent PcG repression of the Ubx promoter has been documented (Mahmoudi, 2003).

Although there is substantial evidence for the notion that the proteins discussed above are involved in PcG silencing of homeotic genes, it remains unclear whether they can be sufficient for targeting or whether additional factors are required. One way to determine a minimal set of protein recognition sequences that can mediate PcG silencing will be the generation of synthetic PREs, which should be tested in vivo. The results suggest that, within such a PRE, PHO sites will need to be flanked by GAGA sites in order to facilitate chromatin binding. The proteins GAGA and Zeste may be particularly well adapted for such a purpose. Both GAGA and Zeste form large homo-oligomers that bind cooperatively to the multiple sites present in their natural response elements, such as the Ubx PRE and promoter. This cooperative mode of DNA-binding may allow these proteins to first bind an accessible site within a nucleosomal array and then progressively displace histones during binding to flanking sites. In addition, GAGA and Zeste have both been shown to recruit selective ATP-dependent chromatin remodeling factors. The process of targeting of remodelers to specific DNA elements may enable GAGA and Zeste to create nucleosome-free or remodeled areas, thus facilitating binding of other regulators. It is considerede likely that the remodeling complexes present in the chromatin preparations used in assays, are involved in the observed synergistic binding between PHO and either GAGA or Zeste (Mahmoudi, 2003).

GAGA oligomerization may also promote the communication between the Ubx PRE and promoter. Both elements, which are separated by ~24 kb of intervening DNA, contain a preponderance of binding sites for GAGA. GAGA oligomerization through its POZ domain allows it to form a protein bridge that directs long-range enhancer-promoter association. In fact, GAGA could even mediate enhancer function in trans by simultaneous binding of two separate DNA fragments. Thus, it is tempting to speculate that GAGA may link the Ubx PRE to the Ubx promoter. It should be noted that both the chromatin remodeling and long-range bridging functions of GAGA might accommodate PRE-mediated activation as well as repression (Mahmoudi, 2003).

The interdependence between proteins belonging to antagonistic genetic groups for efficient chromatin binding described it this study will have to be taken into account when interpreting mutational analysis of PRE function. Thus, removal of recognition sequences for the trxG protein GAGA may block its activation function but could also affect binding of the PcG protein PHO. Moreover, recent results suggest additional opportunities for cross-talk during recruitment of non-DNA-binding PcG complexes. Although a clear consensus between different studies is still lacking, there is experimental evidence for PcG complex recruitment by PHO, GAGA and Zeste. Because binding sites for either one of these proteins alone do not confer PRE function, it appears likely that they work in a combinatorial fashion. Depending on their context, the multitude of distinct binding elements that constitute a PRE might be redundant, cooperative or antagonistic to each other. Furthermore, distinct PREs may require different sets of PRE-binding proteins, and additional recruiters may be involved in PcG-silencing. Attractive candidates are GAGA-related factors batman and the PHO-related factor PHO-like (Mahmoudi, 2003).

In conclusion, current evidence suggests that PRE-directed maintenance of gene activation or repression is not achieved by a simple binary switch set by competing trxG and PcG proteins. Although their relative ratios vary considerably and correlate with transcription levels, they coexist at PREs during gene activation as well as repression. Likewise, genetic suppressor studies indicated extensive cross-talk between PcG and trxG proteins. This study has shown that, already at the level of PRE binding, there is strong interdependence between trxG protein GAGA and PcG protein PHO. The results demonstrate a direct biochemical mechanism for the cooperation between PcG and trxG proteins during PRE binding (Mahmoudi, 2003).

Trithorax-like interaction with Corto

In Drosophila, PcG complexes provide heritable transcriptional silencing of target genes. Among them, the ESC/E(Z) complex is thought to play a role in the initiation of silencing whereas other complexes such as the PRC1 complex are thought to maintain it. PcG complexes are thought to be recruited to DNA through interaction with DNA binding proteins such as the GAGA factor, but no direct interactions between the constituents of PcG complexes and the GAGA factor have been reported so far. The Drosophila corto gene interacts with E(z) as well as with genes encoding members of maintenance complexes, suggesting that it could play a role in the transition between the initiation and maintenance of PcG silencing. Moreover, corto also interacts genetically with Trl, which encodes the GAGA factor, suggesting that it may serve as a mediator in recruiting PcG complexes. Corto bears a chromo domain, and evidence is provided for in vivo association of Corto with ESC and with PC in embryos. Moreover, GST pull-down and two-hybrid experiments show that that Corto binds to E(Z), ESC, PH, SCM and GAGA and co-localizes with these proteins on a few sites on polytene chromosomes. These results reinforce the idea that Corto plays a role in PcG silencing, perhaps by confering target specificity (Salvaing, 2003).

dSAP18 and dHDAC1 contribute to the functional regulation of the Drosophila Fab-7 element

The Drosophila GAGA factor [Trithorax-like (Trl)] interacts with dSAP18, which, in mammals, is a component of the Sin3-HDAC co-repressor complex. GAGA-dSAP18 interaction has been proposed to contribute to the functional regulation of the bithorax complex (BX-C). Mutant alleles of Trl, dsap18 and drpd3/hdac1 enhance A6-to-A5 transformation indicating a contribution to the regulation of Abd-B expression at A6. In A6, expression of Abd-B is driven by the iab-6 enhancer, which is insulated from iab-7 by the Fab-7 element. GAGA, dSAP18 and dRPD3/HDAC1 co-localize to ectopic Fab-7 sites in polytene chromosomes, and mutant Trl, dsap18 and drpd3/hdac1 alleles affect Fab-7-dependent silencing. Consistent with these findings, chromatin immunoprecipitation analysis shows that, in Drosophila embryos, the endogenous Fab-7 element is hypoacetylated at histones H3 and H4. These results indicate a contribution of GAGA, dSAP18 and dRPD3/HDAC1 to the regulation of Fab-7 function (Canudas, 2005).

The conclusion that GAGA, dSAP18 and dRPD3/HDAC1 contribute to the function of the Fab-7 element of BX-C is based on the following observations:

  1. the localization of GAGA, dSAP18 and dRPD3/HDAC1 at ectopic Fab-7 elements (Canudas, 2005).
  2. the effects of Trl, dsap18 and drpd3/hdac1 mutations on Fab-7-dependent silencing. Ectopic Fab-7 constructs are known to mediate silencing of flanking reporter genes both in cis, as in heterozygous GCD6 flies, as well as in trans, as in 5F24 flies, where silencing is pairing-sensitive being observed only when the transgene is in a homozygous state. This study shows that Trl, dsap18 and drpd3/hdac1 mutations affect both cis- and trans-silencing mediated by Fab-7 (Canudas, 2005).
  3. the homeotic A6-to-A5 transformation observed in flies heterozygous for various Trl, dsap18 and drpd3/hdac1 mutant alleles and hemizygous for Df(3R)sbd45, which uncovers dsap18. This homeotic transformation results from the ectopic repression of the iab-6 enhancer at A6 that is insulated from the repressed iab-7 enhancer by the Fab-7 element. The fact that this homeotic transformation is very infrequent in hemizygous Df(3R)sbd45 flies, as well as in the heterozygous mutants, demonstrates that it is directly associated to the Trl, dsap18 and drpd3/hdac1 mutations. Moreover, a single copy of a transgene expressing dsap18 significantly rescues this phenotype. The results also indicate that an unidentified element(s) contained within Df(3R)sbd45 is also contributing to the establishment of the phenotype. In addition to sap18, Df(3R)sbd45 uncovers at least 11 other genes including the trithorax gene, taranis. However, the homeotic transformation described in this study does not appear to be associated to a loss of taranis function since no transformation is observed in flies trans-heterozygous for a null taranis allele and Trl, dsap18 or drpd3/hdac1 mutations (Canudas, 2005).

Together, these results indicate a contribution of GAGA, dSAP18 and dRPD3/HDAC1 to the structural and functional properties of Fab-7. What could this contribution be? Several models might account for these results. Fab-7 is known to contain two functional elements: a PRE, which is required for Pc-dependent silencing, and an adjacent boundary element that insulates iab-6 from iab-7. The finding that, in heterozygous GCD6 flies, mutant Trl, dsap18 and drpd3/hdac1 alleles enhance cis-silencing imposed by Fab-7 suggests that their functions might antagonize Pc-dependent silencing. Several observations, however, make this hypothesis unlikely: (1) at some PREs, GAGA helps recruitment of PcG complexes and contributes to silencing; (2) dRPD3/HDAC1 was shown to be a component of several PcG complexes, and genetic analysis indicates a contribution to homeotic silencing; (3) in mammals, SAP18 acts as a repressor when targeted to an active promoter (Canudas, 2005).

An alternative possibility is that GAGA, dSAP18 and dRPD3/HDAC1 contribute to the function of the Fab-7 boundary element. In fact, the Fab-7 boundary contains several GAGA-binding sites that are required for its enhancer blocking activity and, it is hypoacetylated at histones H3 and H4. In GCD-6 flies, the Fab-7 boundary element is located proximal to the reporter mini-white gene with respect to the PRE so that it might help to insulate the reporter gene from repression by the PRE. In this context, mutations that affect boundary function would result in a less efficient insulation and, therefore, would enhance silencing (Canudas, 2005).

In contrast to the enhancer effect observed in heterozygous GCD6 flies, mutations in Trl, dsap18 and drpd3/hdac1 suppress pairing-dependent trans-silencing in transgenic 5F24(25,2) flies. A contribution to boundary-functions might also account for this effect. Pairing-sensitive trans-silencing results from long-distance chromosomal interactions that involve the association of the transgenes with each other and with the endogenous Fab-7 element, even when located in different chromosomes. These long-distance interactions that require the contribution of PcG proteins might be facilitated by a functional boundary element as has been described for the gypsy insulator (Canudas, 2005).

The incomplete A6-to-A5 homeotic transformation observed in the presence of Trl, dsap18 and drpd3/hdac1 mutations might also reflect a contribution to the boundary function of Fab-7 as, in the mutant conditions, it might not properly insulate the iab-6 enhancer from the repressing activity of the Fab-7 PRE, thereby becoming partially inactivated. Interestingly, mutations that delete the Fab-7 boundary but not the PRE produce, in addition to strong A6-to-A7 transformation, incomplete A6-to-A5 transformation. Moreover, replacement of the Fab-7 boundary by the gypsy or the scs insulator (both of which are not functional in the context of BX-C) results in complete A6-to-A5 transformation (Canudas, 2005).

The results indicate that GAGA, dSAP18 and dRPD3/HDAC1 have similar effects on the functional properties of Fab-7 suggesting a functional link. A physical interaction between GAGA and dSAP18 has been reported. Moreover, in mammals, SAP18 is associated with the Sin3-HDAC co-repressor complex and, in Drosophila, dSAP18 modulates bicoid activity through the recruitment of dRPD3/HDAC1 and it is required to suppress bicoid activity in the anterior tip of the embryo. In this context, it is tempting to speculate that GAGA helps in the recruitment of dSAP18 and dRPD3/HDAC1 to Fab-7 resulting in a concerted contribution to its boundary function (Canudas, 2005).

In mammals, SAP18 is also associated with ASAP, a protein complex involved in RNA processing. In Drosophila, dSAP18 may also participate in RNA processing; in cultured S2 cells, a large proportion of dSAP18 co-immunoprecipitates with factors that participate in RNA processing. It is possible that, in response to cellular signals, the association of dSAP18 to different protein complexes would be regulated during development and/or cell cycle progression (Canudas, 2005).

Transcriptional activation by GAGA factor is through its direct interaction with dmTAF3

The GAGA factor (GAF), encoded by the Trithorax like gene (Trl) is a multifunctional protein involved in gene activation, Polycomb-dependent repression, chromatin remodeling and is a component of chromatin domain boundaries. Although first isolated as transcriptional activator of the Drosophila homeotic gene Ultrabithorax (Ubx), the molecular basis of this GAF activity is unknown. This study shows that dmTAF3 (also known as BIP2 and dTAFII155), a component of TFIID, interacts directly with GAF. Mutations were generated in dmTAF3; in Trl mutant background, they affect transcription of Ubx leading to enhancement of Ubx phenotype. These results reveal that the gene activation pathway involving GAF is through its direct interaction with dmTAF3 (Chopra, 2008)

GAF has been shown to be a transcriptional activator of many genes. Recent studies suggested that both GAF and TFIID are necessary for formation of the appropriate chromatin structure at the hsp26 promoter indicating a mechanism in which GAF binding precedes and contributes to the recruitment of TFIID. However, the question of how GAF leads to the recruitment of TFIID remained unanswered. The finding that GAF interacts directly with dmTAF3 reveals a possible mechanism for how GAF could recruit the TFIID complex to carry out transcriptional activation. It is known that GAF carries out functions other than the activation of transcription. Since GAF does not interact with large number of transcription factors directly, as is evident from genome scale interaction screen, it is likely that recruitment of transcription machinery to activate a promoter is mediated through specific factors, and the current results show that dmTAF3 plays a major role in this. Colocalization and ChIP results also suggest that interaction of GAF with dmTAF3 is very likely to be dependent on the genomic context. This may reflect the fact that such loci may be involved in functions that do not require activator proteins, for example, the loci where GAF functions as repressor protein and recruits Polycomb group members. GAF functions that are independent of dmTAF3 would also include chromatin domain boundary elements where GAF is known to play a role (Chopra, 2008).

The fact that a 3-fold reduction in dmTAF3 expression is sufficient to enhance the phenotype of Ubx mutation in a sensitized background -- heterozygous for the Trl gene encoding GAGA factor -- is functional evidence that TAF3 is a direct partner of GAF in the activation pathway. The effect is likely to be at the level of transcription as shown by the modified expression of a Ubx-lacZ transgene in dmTAF3 and Trl mutant background and confirmed by the reduced level of UBX in the double mutant context. A simple model proposes that GAF contributes to Ubx transcription by its binding to specific sites (via its Zinc fingers) near the promoter and then recruits the transcriptional machinery by interaction of its BTB/POZ domain with dmTAF3 (Chopra, 2008).

GAF can remodel the nucleosomes with the help of NURF complexes and facilitate access to activator or repressor components to such remodeled cis elements. The bound GAF can interact with dmTAF3 which can help recruit TFIID complex and maintain active state of the target gene. In contrast, when a target gene needs to be repressed the bound GAF could recruit PcG complexes to maintain repressed state. GAF binding sites have been found in promoters as well as in PREs. The interaction of GAF either with transcription factors like dmTAF3 or with PcG repressor proteins such as LOLAL, raises the possibility that GAF functions like a switch that could recruit either activator or repressor complexes at a target promoter and then maintain the transcriptional state. In cell types in which a gene needs to be active, GAF bound at the promoter sites would interact with activators and maintain the active state. By contrast, in cell types where a promoter needs to be silenced, GAF would interact with PcG proteins, associated with even distant PREs, by a looping mechanism, and bring about a repressive chromatin context, which probably involves histone tail modifications. It is likely that additional DNA binding factors and their interacting partners contribute to these cross talks of delicately regulated loci with activation and repression machinery. Further studies will be needed to understand this complex network of regulatory events (Chopra, 2008)

The Torso signaling pathway modulates a dual transcriptional switch to regulate tailless expression

The Torso (Tor) signaling pathway activates tailless (tll) expression by relieving tll repression. None of the repressors identified so far, such as Capicuo, Groucho and Tramtrack69 (Ttk69), bind to the tor response element (tor-RE) or fully elucidate tll repression. In this study, an expanded tll expression pattern was shown in embryos with reduced heat shock factor (hsf) and Trithorax-like (Trl) activities. The GAGA factor, GAF encoded by Trl, bound weakly to the tor-RE, and this binding was enhanced by both Hsf and Ttk69. A similar extent of expansion of tll expression was observed in embryos with simultaneous knockdown of hsf, Trl and ttk69 activities, and in embryos with constitutively active Tor. Hsf is a substrate of mitogen-activated protein kinase and S378 is the major phosphorylation site. Phosphorylation converts Hsf from a repressor to an activator that works with GAF to activate tll expression. In conclusion, the GAF/Hsf/Ttk69 complex binding to the tor-RE remodels local chromatin structure to repress tll expression and the Tor signaling pathway activate tll expression by modulating a dual transcriptional switch (Chen, 2009).

This study has shown that the interaction of GAF with Hsf and Ttk69 plays a critical role in regulation of tll expression. At locations where the Tor pathway is inactive, GAF, Hsf, and Ttk69 constitute a protein complex that binds to the tor-RE tightly. The protein complex recruits other co-repressors and chromatin remodeling factors containing Rpd3 to the organization of a high-order local chromatin structure. tll expression is off in this scenario (Chen, 2009).

Both Hsf and Ttk69 have been shown to be substrates of Mapk. At locations where the Tor pathway is active, activated Mapk phosphorylates both Hsf and Ttk69. The phosphorylated Ttk69 is degraded. The phosphorylation converts Hsf into an activator that leads to an increase in tll. In addition, the expanded lacZ expression patterns in hsf4 embryos at the nonpermissive temperature suggested that other activators, such as Stat, are apparently required for full activation of tll (Chen, 2009).

A protein–protein interaction network, including GAF, Hsf, Rpd3 and Ttk69, is required for tll repression. Studies show that Capicuo interacts with Groucho, which associates with Rpd3 and Sin3A to repress tll expression. Furthermore, CtBP is essential for Ttk69 to suppress neuronal cell fate. Therefore, these factors may also be recruited to repress tll expression (Chen, 2009).

Multiple GAFs and Ttk69s bind to the flanking regions of the tor-RE. The results from the DNaseI footprinting experiments showed that GAF binds to four sites in the tll-MRR, including the tor-RE. DNA sequences in footprints a and b match the consensus sequence bound by GAF, 3.5 GA repeats. Although DNA sequence in footprint c does not match the consensus sequence, this site contains a GAGA tandem repeat with 1-bp spacing. These three sites are well protected by GAF from DNaseI digestion, without or with little influence by Hsf. Therefore, GAF oligomer binds to these sites with high affinity and likely assists its own binding to the tor-RE. Similarly, binding of Even-skipped protein to low-affinity sites was assisted by its own binding to high-affinity sites at a distance (Chen, 2009).

Ttk69 acts as a co-repressor to increase GAF/Hsf binding to the tor-RE. In another EMSA experiment, a probe containing both the tor-RE and TC5 was used. Ttk69 binds to TC5 and assists GAF/Hsf/Ttk69 binding to the tor-RE. This is consistent with a previous report that base substitution a TCCT element (TC5) at the 3' flanking region of the tor-RE clearly affects the initiation of tll repression. Additionally, the binding of Ttk69 to TC2 might facilitate tll repression (Chen, 2009).

Results from the DNaseI footprintings showed different patterns over the tor-RE protected by either GAF/Hsf or GAF alone. Unexpectedly, these different tor-RE complexes showed the same mobility in results from EMSA experiments, which might be explained by the altered binding property of GAF in the presence of Hsf. BTB domains in GAF and human Promyelocytic Leukemia Zinc Finger (PLZF) proteins have been shown to belong to the Ttk subfamily, and BTB homodimer is a unit of PLZF oligomer. Since GAF oligomer presumably assisted itself in binding to the tor-RE, GAF homodimer could bind the 5'-end as well as the 3'-end of the tor-RE. When Hsf was added in addition to GAF, the interaction of GAF with Hsf influenced GAF binding to the tor-RE, leading to the alteration of the footprinting pattern. However, it remains unclear whether Hsf contacts DNA when it exists in the DNA–protein complex. Nevertheless, the results from shift-western blotting clearly demonstrated the presence of both GAF and Hsf in the DNA–protein complex (Chen, 2009).

Displacement of GAF monomer by Ttk69 to form a GAF/Hsf/Ttk69 complex might explain the unchanged band-shift patterns when the three proteins were added to the binding mixture. Molecular weights between GAF and Ttk69 are slightly different and BTB domains in both GAF and Ttk69 proteins belong to the same subfamily. Molecular modeling with crystal structure of the BTB domain in human BACH1 (PDB code: 2ihc) as a template was used to test the displacement hypothesis. Results showed that buried areas and free energies (ΔGs) of engaged interfaces among GAF and Ttk69 homodimers and GAF/Ttk69 heterodimer were similar, suggesting that formation of a GAF/Ttk69 heterodimer is possible. This explanation is partially supported by the detection of both Ttk69 and Hsf in the DNA–protein complex. In addition, Ttk69 has been shown to inhibit GAF activation in the absence of Ttk69 binding sites, and addition of Ttk69 significantly increases GAF binding to the 173-bp probe that contains multiple GAF binding sites. However, results from EMSA show that mobility of the DNA–protein complex is slightly affected. In conclusion, these data plus the existence of GAF, Hsf and Ttk69 in the DNA–protein complex indicate that the interaction of GAF with Hsf and Ttk69 form a protein complex binding to the tor-RE tightly (Chen, 2009).

In summary, multiple factors bind to the tor-RE and its flanking regions to form a large repression complex, consistent with the finding that the 240-bp cis-regulatory region of tll, but not the tor-RE itself, silences from a heterologous promoter (Chen, 2009).

Several studies on gene repression, i.e., bacterial bipA and eno genes, fly decapentaplegic and human and GP91phox genes, have revealed that multiple low-affinity sites are bound by a repressor to regulate gene expression. Likewise, multiple weak binding sites that cluster within a short range of a cis-regulatory region are reported to facilitate the cooperative binding of factors, leading to a sharp definition in expression patterns. Reduction of repressor concentration leads to a loss in the definition at the edge of expression domains. In this study, the binding affinity of GAF to the tor-RE is low, and multiple tor-REs are present in the tll cis-regulatory region. This explains the poorly-defined boundary of the tll expression patterns in embryos that have reduced hsf and Trl activities (Chen, 2009).

Footprint b was marginally affected by supplementing Hsf. The DNA sequence in this site contains one binding site each for GAF and Hsf (−130 GAGAGAG and −115 GAATCCTGCGGAA), located in regions O and P, respectively. The sequence for GAF binding matches the consensus sequence. Although this putative Hsf binding sequence does not match the consensus sequence, it has been shown that Hsf is able to bind a sequence with two GAAs spaced by 7 bp. Interestingly, in contrast to the role of tor-RE, deletion of either region O or P from the tll-MRR results in a drastic reduction of lacZ mRNA levels, but no changes in the expression patterns. These results not only further support the notion that Hsf and GAF are required for tll activation, but also provide an explanation for (1) the different tll expression patterns in embryos with reduced hsf and Trl activities by either removing one copy of the genes or using RNAi to knockdown activities of the genes, and (2) the different tll expression levels resulting from either base substitutions in the tor-RE or reduction of gene activities (Chen, 2009).

These results showed a moderate effect of Hsf on GAF binding to footprint c. In addition to the GAGA repeat, this site also contains two GAA repeats with the 7-bp spacing. Results from the DNaseI footprinting experiments showed that GAF bound to this site less efficiently. Furthermore, deletion of this site from the tll-MRR results in a low level and slightly expanded lacZ expression pattern. In addition, footprint d at the 5'-end of the tor-RE was a weak site bound by GAF and the footprint pattern protected by GAF was significantly affected by Hsf. Base substitutions to this site severely damaged tll repression. These data supported the notion that the low and high affinity sites bound by GAF are the major contributors to tll repression and activation (Chen, 2009).

All signaling pathways regulate, at least in part, specific factors to activate the expression of target genes. In most well-studied signal pathways, the signals directly activate factors. In some cases, the signals switch on expression of their target genes from a repressed state. For example, in the absence of Notch, Wnt, Hedgehog or nuclear receptor signaling, the expression of target genes is repressed. There is a common mechanism among these cases. A specific factor or complex, such as Su(H)/CBF1 in the Notch pathway, Lef/Tcf in the Wnt pathway, Ci/Gli in the Hh pathway or the nuclear receptors themselves, bind to a specific DNA sequence to prevent target genes from being transcribed. When the signaling pathway is active, the repression is relieved by the processed receptor, activated activator, co-activators or by the nuclear receptor itself. Results from this study indicated that both GAF-associated proteins, Hsf and Ttk69, constitute a dual al switch for tll expression that includes degradation of the Ttk69 co-repressor and conversion of the Hsf repressor into an activator after Mapk phosphorylation, where Mapk is a downstream effector of the Tor pathway (Chen, 2009).

Drosophila GAGA factor polyglutamine domains exhibit prion-like behavior

The Drosophila GAGA factor (GAF) participates in nucleosome remodeling to activate genes, acts as an antirepressor and is associated with heterochromatin, contributing to gene repression. GAF functions are intimately associated to chromatin-based epigenetic control, linking basic transcriptional regulation to heritable long-term maintenance of gene expression. These diverse functions require GAF to interact with different partners in different multiprotein complexes. The two isoforms of GAF depict highly conserved glutamine-rich C-terminal domains (Q domain), which have been implicated in complex formation. This study shows that the Q domains exhibit prion-like properties. In an established yeast test system the two GAF Q domains convey prion activities comparable to well known yeast prions. The Q domains stably maintain two distinct conformational states imposing functional constraints on the fused yeast reporter protein. The prion-like phenotype can be reversibly cured in the presence of guanidine HCl or by over-expression of the Hsp104 chaperone protein. Additionally, when fused to GFP, the Q domains form aggregates in yeast cells. It is concluded that prion-like behavior of the GAF Q domain suggests that this C-terminal structure may perform stable conformational switches. Such a self-perpetuating change in the conformation could assist GAF executing its diverse epigenetic functions of gene control in Drosophila (Tariq, 2013).

By making use of well-characterized genetic assays determining prion-like characteristics of glutamine-rich domains in different proteins, this study has identified the Q domains of both the GAF isoforms as prion-like domains. The fusion proteins in which GAF-Q domains were introduced in place of the Sup35p prion domain could support distinct physical and functional prion states that recapitulated the translation termination defect associated with [PSI+]. Importantly, the nonsense suppression prion-like state exhibited by the GAFQ-SupC fusion was cured by growth in the presence of GuHCl. Similar to the [PSI+] prion aggregated non-functional state of yeast prion Sup35p, the nonsense suppression phenotype by GAFQ-SupC could also be cured by the over-expression of Hsp104. The GAF-Q domains fused to GFP also formed visible aggregates resembling those of GFP labeled Sup35p in [PSI+], which also depended on prion [PIN+]. Many sequences with high Q content (as high as that of yeast prions) including human polyglutamine expansion disease proteins, form visible aggregates when overexpressed in yeast as GFP fusions. However, only a limited number of Q/N rich sequences are bone fide prion domains capable of propagating these aggregates over multiple cell generations even when expressed at low levels. Construction of a synthetic prion revealed that pathogenically expanded stretch of 62 Qs (Q62) fused to Sup35C or GFP could mimic prion-like behavior, however, 22Q did not show such characteristics. The prion-like behavior of 519Q similar to Sup35 is of significance because it contains only a short Q stretch as compared to Q62. Importantly, computational assessment of GAF519 and GAF582 using prion aggregation prediction algorithm reveals that both proteins have propensity to make prion (Tariq, 2013).

As compared to other eukaryotes analyzed, a surprisingly large number of proteins in Drosophila have extended Q-rich tracts, remarkably similar to those found in the prion-forming domains of yeast proteins. In in vitro studies and in transient assays in cell culture fusions of the Q domains with the Gal4 DNA binding domain activate by stabilizing the transcriptional complex. However, in transgenic flies chromatin binding and transcriptional activation activity by GAF was found to be independent of Q domains, leaving open the designation of the exact molecular function. So far, the combined results suggest that the Q domains are mostly involved in the formation of larger GAF complexes. The associated prion-like activity might thus provide an ability to GAF to attain distinct conformational states that may be heritable. The high conservation of the C terminal Q-rich domain of GAF in insects, suggests that there is a strong evolutionary preference to maintain such associated structure and function (Tariq, 2013).

The analysis of GAFQ-SupC fusions in yeast provides an interesting analogy between GAFQ and the Sup35 prion domain, consistent with the previous findings, revealing that the GAFQ domain is essential for the formation amyloid fibers in vitro. The results also support the previous findings that oligomerization of GAF found in Drosophila cells may be facilitated by the long Q stretches in GAF (Wilkins, 1999). It is emphasized that GAF may not be a bone fide prion but Q domains in GAF may induce conformational switch reminiscent of prion-like behavior. In yeast, prions are not pathogenic but rather act as an epigenetic regulator of cell physiology and several epigenetically heritable traits are found to depend on a prion mechanism (Shorter, 2005). Evidence for regulation of gene expression patterns by propagation of Swi1 and Cy8 proteins as prions has provided a novel link between chromatin remodeling proteins and prion formation and it has revealed an additional mechanism for controlling global gene transcription that is based on an inherited self-perpetuating change in the conformation. The current results indicate that the possibility of such an intricate link between chromatin associated proteins, prion formation and epigenetic inheritance of gene expression might also apply in higher eukaryotes. Intriguingly, a large majority of the identified Drosophila proteins with Q-rich domains are essential developmental proteins including chromatin regulating proteins from PcG and TrxG involved in epigenetic inheritance. It could be envisaged that GAF-Q domains provide an inherent plasticity which may lead to a conformational switch in GAF in a changing environment. Such a Q domain dependent conformation switch in GAF may be regulated by some specific post-translational modifications of GAF and facilitated by molecular chaperones. This could result in modulated gene expression patterns that may contribute to phenotypic variation. It is suggested that GAF-Q domain may act as prion-like domain in Drosophila and support the notion that oligomeriztaion of GAF and other PcG/TrxG proteins, which is known to be crucial for the function of these proteins, may be facilitated by such prion-like domains (Tariq, 2013).

Return: see Trithorax-like Targets of Activity part 1/2


Trithorax-like: Biological Overview | Evolutionary Homologs | Developmental Biology | Effects of Mutation | References

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

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