Synergistic activation of a Drosophila enhancer by HOM/EXD and DPP signaling
1997; Springer Nature; Volume: 16; Issue: 24 Linguagem: Inglês
10.1093/emboj/16.24.7402
ISSN1460-2075
Autores Tópico(s)Chromosomal and Genetic Variations
ResumoArticle15 December 1997free access Synergistic activation of a Drosophila enhancer by HOM/EXD and DPP signaling Nicole C. Grieder Nicole C. Grieder Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Thomas Marty Thomas Marty Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Hyung-Don Ryoo Hyung-Don Ryoo Department of Biochemistry and Molecular Biophysics, Center for Neurobiology and Behavior, Columbia University, 630 West 168th Street, New York, NY, 10032 USA Search for more papers by this author Richard S. Mann Richard S. Mann Department of Biochemistry and Molecular Biophysics, Center for Neurobiology and Behavior, Columbia University, 630 West 168th Street, New York, NY, 10032 USA Search for more papers by this author Markus Affolter Markus Affolter Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Nicole C. Grieder Nicole C. Grieder Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Thomas Marty Thomas Marty Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Hyung-Don Ryoo Hyung-Don Ryoo Department of Biochemistry and Molecular Biophysics, Center for Neurobiology and Behavior, Columbia University, 630 West 168th Street, New York, NY, 10032 USA Search for more papers by this author Richard S. Mann Richard S. Mann Department of Biochemistry and Molecular Biophysics, Center for Neurobiology and Behavior, Columbia University, 630 West 168th Street, New York, NY, 10032 USA Search for more papers by this author Markus Affolter Markus Affolter Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland Search for more papers by this author Author Information Nicole C. Grieder1, Thomas Marty1, Hyung-Don Ryoo2, Richard S. Mann2 and Markus Affolter1 1Abteilung Zellbiologie, Biozentrum, Universität Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland 2Department of Biochemistry and Molecular Biophysics, Center for Neurobiology and Behavior, Columbia University, 630 West 168th Street, New York, NY, 10032 USA The EMBO Journal (1997)16:7402-7410https://doi.org/10.1093/emboj/16.24.7402 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The homeotic proteins encoded by the genes of the Drosophila HOM and the vertebrate HOX complexes do not bind divergent DNA sequences with a high selectivity. In vitro, HOM (HOX) specificity can be increased by the formation of heterodimers with Extradenticle (EXD) or PBX homeodomain proteins. We have identified a single essential Labial (LAB)/EXD-binding site in a Decapentaplegic (DPP)-responsive enhancer of the homeotic gene lab which drives expression in the developing midgut. We show that LAB and EXD bind cooperatively to the site in vitro, and that the expression of the enhancer in vivo requires exd and lab function. In addition, point mutations in either the EXD or the LAB subsite compromise enhancer function, strongly suggesting that EXD and LAB bind to this site in vivo. Interestingly, we found that the activity of the enhancer is only stimulated by DPP signaling significantly upon binding of LAB and EXD. Thus, the enhancer appears to integrate positional information via the homeotic gene lab, and spatiotemporal information via DPP signaling; only when these inputs act in concert in an endodermal cell is the enhancer fully active. Our results illustrate how a tissue-specific response to DPP can be generated through synergistic effects on an enhancer carrying both DPP- and HOX-responsive sequences. Introduction During the development of multicellular organisms, cells have to be able to sense their position with respect to the major body axis, and maintain this information during subsequent cell divisions up to the adult stages. A group of genes which play a major role in conferring and maintaining positional information is located in the homeotic selector gene complex (HOM-C) in Drosophila and the homologous HOX gene complexes in vertebrates (Lewis, 1978; Hayashi and Scott, 1990; McGinnis and Krumlauf, 1992; Gehring et al., 1996). All of the homeotic genes in these complexes encode homeodomain proteins and are thought to fulfill their function by regulating distinct sets of downstream target genes (Botas, 1993; Mann, 1995; Gehring et al., 1996; Mann and Chan, 1996; Graba et al., 1997). However, the molecular mechanisms by which the different homeotic genes control the formation of specific structures in given places are far from being understood. The identification of in vivo HOM/HOX target sites has provided insight into the molecular mechanisms governing the selective regulation of genes by homeotic proteins. Guided by observations made in genetic studies which identified the extradenticle (exd) gene as an important regulator of homeotic gene function (Peifer and Wieschaus, 1990), biochemical analyses have shown that several homeotic proteins bind to in vivo target sites cooperatively with the homeodomain protein EXD. This cooperative recognition of sequences in target enhancers contributes both to a higher affinity of the heteromeric complex (when compared with the monomeric proteins) and to a more selective recognition of target elements (Mann, 1995; Mann and Chan, 1996). One of the best characterized HOM/EXD-binding sites is present in a 20 bp oligonucleotide, repeat 3, which was identified in the 5′ autoregulatory region of the mouse Hoxb-1 gene (Pöpperl et al., 1995). Labial (LAB) protein and its mammalian ortholog HOXB-1 are both able to bind cooperatively with EXD to this site, whereas other HOM/HOX proteins, such as Ultrabithorax (UBX) or HOXB-4, cannot (Pöpperl et al., 1995; Chan and Mann, 1996; Chan et al., 1996). Furthermore, when lacZ reporter constructs containing three tandem copies of repeat 3 were introduced into either Drosophila or mouse embryos, lacZ expression patterns resembling the endogenous lab and Hoxb-1 patterns, respectively, were generated (Chan et al., 1996). However, no Drosophila cis-regulatory element has been reported to contain a functional LAB/EXD site resembling the original repeat 3 sequence. The identification of such in vivo binding sites would allow the study of the interaction of homeotic proteins with co-factors and other transcriptional regulators in their natural environment, using genetic and biochemical methods. In addition to the cell-autonomous maintenance of positional information, cells have to communicate to allow a concerted development of all tissues and organs. Members of the transforming growth factor-β (TGF-β) superfamily of signaling molecules play important roles in such intercellular communication processes (Massagué, 1996). One of the Drosophila members of this superfamily, Decapentaplegic (DPP; Gelbart, 1989), has been shown to be both upstream and downstream of the homeotic genes; on the one hand, DPP regulates the expression of certain homeotic genes, on the other hand, dpp transcription is regulated in some tissues by homeotic gene products (Bienz, 1994, 1996). Here, we report the identification of a single high affinity LAB/EXD-binding site in an enhancer of the lab gene which drives expression in endodermal cells in response to DPP signaling (Tremml, 1990; Tremml and Bienz, 1992; Eresh et al., 1997). We show that point mutations in either the LAB or the EXD subsite abolish cooperative DNA binding in vitro, and strongly reduce the activity of the enhancer in vivo. In addition, both lab and exd are genetically required for high level expression, strongly suggesting that LAB and EXD bind to these sites in vivo. Furthermore, we observe that the DPP responsiveness of the enhancer (Eresh et al., 1997) is largely abolished in the absence of EXD and LAB. Our results indicate that the activity of this lab enhancer requires a combined and synergistic input from the DPP signaling cascade on the one hand, and from the HOM selector gene product LAB on the other hand. Our results illustrate how a tissue-specific response to DPP can be generated through synergistic effects on an enhancer carrying both DPP- and HOM-responsive sequences. Results A 48 bp element is essential for the activity of a DPP-dependent enhancer of the lab gene It has been reported previously that a 550 bp fragment located 2.5 kb upstream of the Drosophila lab gene (HZ550; hereafter referred to as lab550 or 1/546) is activated in the endodermal cell layer in response to the DPP signaling molecule secreted by visceral mesoderm parasegment 7 cells (Tremml, 1990; Tremml and Bienz, 1992; Eresh et al., 1997). In an attempt to define a minimal DPP-responsive element, we made terminal deletions and examined the expression patterns driven by the shortened elements in transgenic embryos (constructs are shown in Figure 1 and expression in transgenic embryos in Figure 2). As a result of this analysis, we identified a 45 bp fragment (located between nucleotides 48 and 92 of the lab550 element) which was essential to drive high levels of expression in the endoderm. While the deletion of the first 48 bp of lab550 had no effect on its expression (Figure 2A and B), the removal of an additional 45 bp resulted in an almost complete loss of expression of the reporter gene (Figure 2C, compare 48/546 with 92/546). Two very short constructs each containing the essential 45 bp (1/95 and 48/95) were also expressed in the midgut endoderm, although at a lower level (and slightly more posterior) than the full-length element (Figure 2D and F; see also Figure 6). The removal of only 9 bp from the 3′ end of element 1/95 resulted in a very strong reduction in expression (construct 1/86; Figure 2E). Based on the results obtained with these deletion constructs, we concluded that a cis-acting element essential for expression in the central midgut endoderm must be located within this 45 bp region. Figure 1.Transgenic constructs and mutations in a LAB/EXD-binding sequence. (A) Wild-type and deletion constructs were named according to the extent of sequences present. The wild-type enhancer is 546 bp long and corresponds to HZ550 in Tremml and Bienz (1990) (see Materials and methods). We refer to this construct as lab550 or 1/546. The level of expression in the central midgut endoderm has been determined semi-quantitatively in several lines for each construct; strongest expression was seen with the wild-type construct, whereas 1/95 and 48/95 were significantly weaker but clearly detectable in the midgut. The sequence of an essential 48 bp region is shown below the constructs. A 10 bp sequence corresponding to a putative LAB/EXD-binding site is displayed in bold. The exact endpoints of deletions 1/86 and 92/546 are indicated with open arrows. (B) Point mutations introduced into the LAB/EXD site are shown, along with the nomenclature of the corresponding transgenic constructs. Mutated residues are printed in bold and further indicated with an asterisk. Download figure Download PowerPoint Figure 2.Expression of transgenic constructs in stage 14 embryos. β-Gal expression driven by the different constructs. Strong expression in all cells of the expression domain is seen for constructs 1/546 (A) and 48/546 (B). Only a few cells (on average 1–4 cells per embryo) are detected in embryos transgenic for construct 92/546 (C). Expression in the developing midgut is clearly detected with the short elements 1/95 (D) and 48/95 (F), but expression levels are lower than with constructs 1/546 and 48/546, and endodermal cells within the expression domain are not stained uniformly. Construct 1/86 (E) does not drive expression in the midgut. All constructs are also expressed in cells of the amnioserosa and in lateral epidermal cells which served as a control for the presence of the construct in the stained embryos. Download figure Download PowerPoint Expression of the lab enhancer requires lab and exd function Sequence inspection of the essential 45 bp region suggested that it could be involved in conferring homeotic and signaling-dependent regulation to lab550. We identified a sequence element (TGATGGATTG; shown in bold and italics on the bottom strand in Figure 1) which is identical in nine out of ten positions to a sequence element of the mouse Hoxb-1 upstream regulatory region (TGATGGATGG; Pöpperl et al., 1995) that previously has been shown to be capable (in an oligomerized form) of conferring LAB/EXD-dependent expression in the Drosophila embryo, both in the midgut and in the head (Chan et al., 1996, 1997). This element might indeed be involved in the activation of lab550, since the 9 bp deletion from construct 1/95 to 1/86 (which abolishes expression of the element) removes part of the putative LAB/EXD site (Figures 1 and 2). The recent finding that the nuclear localization of the EXD homeodomain protein in the developing endoderm is regulated by DPP (Mann and Abu-Shaar, 1996) suggests that the LAB/EXD-binding site might mediate part of the DPP responsiveness of the lab550 enhancer. To investigate whether lab550 as well as the short 48 bp enhancer (48/95) were indeed lab dependent, we analyzed the expression pattern driven by these elements in a lab null mutant background (Diederich et al., 1989). Expression of lab550 was strongly reduced in stage 14 embryos in the absence of LAB, and remained detectable at low levels in 2–4 cells only (Figure 3B; compare to Figure 3A). Expression driven by the 48 bp element was not detectable at all in the midgut of lab mutant embryos, demonstrating an absolute requirement for lab function in the activation of 48/95 (Figure 3C). In contrast, the residual expression of the deletion construct 92/546 (1–4 cells, occasionally, in the central midgut endoderm; Figure 2C) was not affected by the absence of lab (data not shown). This finding is in line with the absence of the putative LAB/EXD-binding site in construct 92/546. Figure 3.LAB and EXD protein are required for high levels of expression of the lab enhancer. To identify lab mutants, embryos shown in (A), (B) and (C) were stained in brown for β-gal and in blue for the endogenous LAB protein. (A) Wild-type embryo transgenic for construct 1/546. (B) Expression of construct 1/546 in an embryo lacking LAB protein; only weak expression in a few cells is detectable. In stage 16 lab mutant embryos, expression levels are somewhat increased (see also Tremml and Bienz, 1992); apparently, DPP-inducible enhancer activity in late embryos is less dependent on lab (data not shown). (C) The short element 48/95 is not expressed at detectable levels in the endoderm in lab mutants. (D) Only weak and scattered expression of the lab550 enhancer (1/546) is detected in embryos derived from females carrying exd mutant germline clones. (E) Expression pattern of an embryo transgenic for a construct (5CRE) carrying oligomerized CRE-binding sites; note the expression in the anterior and the central midgut. (F) Expression of the 5CRE construct is also evident in embryos lacking EXD protein; the expression domain is altered but follows the pattern of dpp expression in the visceral mesoderm of exd mutants (data not shown; see Rauskolb et al., 1994). Download figure Download PowerPoint We have also analyzed the expression pattern of the full-length construct in the absence of the exd gene product. To test the requirement for exd, which is a highly expressed maternal gene (Rauskolb et al., 1993), we generated females with mosaic exd mutant germlines. When crossed to males carrying the lab550 element, virtually no expression was observed in embryos derived from exd mutant germline clones (Figure 3D). These experiments demonstrate that the expression of the DPP-dependent lab550 enhancer critically depends on the presence of the two homeodomain-containing transcriptional regulators EXD and LAB. Activity of the lab550 enhancer requires the LAB and the EXD half-sites To find out whether the genetic requirement for lab and exd was linked to the single putative LAB/EXD site in lab550, we tested whether the introduction of point mutations in the EXD or LAB subsites in the context of the full-length lab550 enhancer would impair expression of the latter. The strategy to inactivate either the EXD or the LAB half-site, exclusively, was based on previous biochemical analysis (Chan and Mann, 1996; Chan et al., 1996, 1997). The mutations we introduced are shown in Figure 1. Indeed, inactivation of either the LAB subsite by a single mutation (labm1, Figure 4B), or of the EXD subsite by the mutation of one (exdm1; Figure 4C) or two adjacent bases (exdm2; Figure 4D) resulted in a severe reduction in expression, mimicking the expression of the full-length (wild-type) element in lab or exd mutant backgrounds (compare with Figure 3B and D). Mutation of both the EXD and the LAB subsite resulted in an even stronger reduction of expression (exdm1labm1; Figure 4E). Figure 4.Point mutations in the putative LAB/EXD-binding site strongly reduce expression of the lab enhancer. (A) Expression of the full-length enhancer (lab550 or 1/546) is compared with the expression of constructs harboring one or two mutations in the context of the full-length element. Expression levels were severely reduced upon the introduction of a mutation in the LAB subsite (B, labm1) or the EXD subsite (C, exdm1), or upon the introduction of two point mutations in the EXD site (D, exdm2) or the LAB/EXD site (E, exdm1labm1). Changing the two base pairs in the center of the LAB/EXD site from GG to TA somewhat reduced the expression level, but did not alter the domain of expression (F, labm2). Note that no expression is detectable in the head region with construct labm2 (F). Download figure Download PowerPoint To confirm that EXD and LAB indeed bind cooperatively to the LAB/EXD site which we identified in lab550, we performed band shift experiments with labeled oligonucleotide probes and purified proteins (Figure 5A). The TGATGGATTG wild-type site formed complexes with EXD and LAB, but not with EXD or LAB alone. The reduced affinity of the point mutants in the EXD or the LAB subsite for LAB–EXD complexes was confirmed by band shift experiments; both mutations abolish complex formation (Figure 5A). Taken together, these results strongly suggest that EXD and LAB are binding directly to the lab enhancer element in vivo, consistent with the genetic requirement for exd and lab function. Figure 5.Cooperative binding of LAB and EXD requires both the LAB and EXD half-sites. 32P-Labeled oligonucleotides containing the wild-type (WT), exdm1, exdm2, labm1 or labm2 LAB/EXD-binding sites were mixed with EXD and LAB (A) or EXD and DFD (B) proteins as indicated, and complexes were resolved on native polyacrylamide gels. LAB–EXD complex formation is reduced by a single base pair change (exdm1) and abolished by a two base pair change (exdm2) in the EXD half-site. LAB–EXD complex formation is abolished by a single base pair change in the LAB half-site (labm1) but only weakly reduced by the GG→TA change in the LAB half-site (labm2). As was previously observed (Chan et al., 1997), DFD/EXD prefers to bind the sequence 5′TGATTAAT rather than 5′TGATGGAT. Download figure Download PowerPoint Figure 6.Expression of the wild-type lab550 enhancer is, in contrast to isolated sub-elements, strongly DPP responsive. Expression of the wild-type lab550 enhancer (1/546) (A–D, horizontal lines), and constructs 95/546 (E–H), 1/95 (I–L) and 48/95 (M–P) were analyzed in wild-type embryos (A, E, I and M; vertical lines), dpps4 mutant embryos (B, F, J and N), abdA mutants (C, G, K and O) and upon ectopic expression of DPP (D, H, L and P). Note that both small constructs (I and M) are expressed in cells posterior to the second midgut constriction (arrow); these cells do not express the endogenous lab gene during these stages. The expression of all constructs strictly depends on DPP since none of them is active in dpps4 mutant embryos which lack DPP in the visceral mesoderm (B, F, J and N). The lab550 enhancer is strongly expressed in most endodermal cells posterior to the first midgut constriction in abdA mutant embryos; these endodermal cells are adjacent to DPP-expressing visceral mesodermal cells (Immerglück et al., 1990; Reuter et al., 1990). The deletion construct 95/546 which lacks the LAB/EXD-binding site is also expressed in posterior endodermal cells (G), although it is weaker and more patchy than the lab550 element. However, the short elements 1/95 (K) and 48/95 (O) are not expressed in abdA mutants and therefore do not respond to DPP in this mutant background. Element 1/95 drives weak expression in the most posterior endodermal cells in stage 15 embryos. lab550 is strongly induced by ectopic DPP in most endodermal cells (D). The sub-elements 92/546 (H), 1/95 (L) and 48/95 (P) respond weakly to DPP in the posterior midgut, but cannot be induced by DPP in endodermal cells anterior to the normal expression domain (see H and L). Download figure Download PowerPoint It has been reported previously that a two base pair change (GG→TA) in the minimal LAB/EXD-responsive element repeat 3 can switch the in vivo specificity from lab to Deformed (Dfd) (Chan et al., 1997). To test whether we could similarly alter the expression pattern of the lab550 enhancer, we introduced the same two base pair mutation in the LAB/EXD site of lab550, and analyzed the expression pattern in transgenic embryos. Midgut expression was clearly still detectable (although at somewhat reduced levels), and we never observed expression in the head region (labm2; Figure 4F). Thus, although DFD can bind to the TA site in vitro (see Figure 5), it is incapable of activating the TA-containing lab550 enhancer in transgenic embryos in the DFD expression domain. This suggests that factors in addition to LAB and EXD determine the germ layer specificity of the lab550 enhancer element (see also Discussion). This is also evident from the observation that construct 92/546 (which lacks the LAB/EXD site) is still expressed in the endoderm (Figure 2C). Strong induction of the lab enhancer by DPP requires LAB and EXD In several previous reports, it has been shown that lab550 represents a DPP-responsive enhancer (Tremml, 1990; Tremml and Bienz, 1992; Eresh et al., 1997). We have shown above that in lab and exd mutant embryos, the lab550 enhancer is not strongly expressed (Figure 3), although DPP is present in lab mutants and (albeit in an altered pattern) in the visceral mesoderm of exd germline clone-derived embryos (Rauskolb and Wieschaus, 1994). In contrast to lab550, other DPP-dependent enhancers (such as the cAMP response element. 5CRE; Eresh et al., 1997) do respond to DPP in lab mutant embryos (data not shown) and in embryos lacking maternal and zygotic exd, confirming that DPP signaling per se is not impaired in the endoderm in the absence of lab or exd (Figure 2G and H; see also Discussion). Since the endogenous lab gene is strongly induced by DPP (Immerglück et al., 1990; Reuter et al., 1990; Grieder et al., 1995), the DPP responsiveness of lab550 expression could, in principle, be entirely due to an indirect autoregulatory feedback loop through the LAB/EXD site. However, more recent results suggest that CREB-binding sites act as targets for DPP signaling in lab550 (Eresh et al., 1997). To assess the role of DPP signaling in the expression and regulation of various sub-elements of lab550, we have analyzed different enhancer constructs with respect to their DPP responsiveness. As reported previously (Tremml and Bienz, 1992; Eresh et al., 1997), lab550 is not expressed in dpps4 mutants which lack dpp expression in the visceral mesoderm (Figure 6B). In abdA mutants, in which dpp is expressed in the entire posterior midgut, lab550 is expanded posteriorly (Immerglück et al., 1990) (Figure 6C). In addition, lab550 can be induced to high levels in most endodermal cells upon ubiquitous DPP expression (Figure 6D). Thus, the lab550 enhancer not only strictly requires DPP for expression, but responds strongly to DPP upon genetic manipulation of the expression pattern of dpp. In all of the experiments shown in Figure 6, the expression of the lab550 enhancer closely follows the expression of the endogenous lab gene. We next analyzed the residual expression of the deletion construct 92/546 in the same genetic backgrounds. Because the expression of this construct does not depend on lab function, its activity is not influenced by dpp indirectly due to the induction of the endogenous lab gene. In dpps4 mutants, expression driven by the 92/546 construct is abolished, demonstrating that the residual, lab-independent expression depends on DPP (Figure 6F). Consistent with this observation, low levels of expression are seen in the entire posterior midgut in abdA mutants (Figure 6G); upon ectopic expression of DPP, scattered cells expressing the 92/546 construct were found in the posterior midgut and (rarely) in the anterior midgut (Figure 6H). This demonstrates that the sequences 3′ to the LAB/EXD site behave, by themselves, as a weak DPP response element. Similar conclusions were reached based on the expression of a lab550 subfragment extending from approximately position 95 to position 300 (Tremml and Bienz, 1992). We have also investigated the behavior of construct labm1exdm1 in the same genetic backgrounds. This construct behaved in a virtually identical manner to 92/546 (data not shown), a finding which is consistent with the requirement for LAB and EXD for strong DPP inducibility. The low expression levels and the weak DPP inducibility of construct 92/546 (and exdm1labm1) suggested that the expression of the lab550 element might be driven mainly by the 48 bp element which contains the LAB/EXD site, and that the adjacent sequences play only a moderate role in the control of the expression domain and/or expression level. To test this possibility, we have analyzed the expression domain of 48/95, and tested its response to DPP. Although the expression of the short element in the central midgut strictly depends on lab function (Figure 3C), its expression domain is shifted posteriorly with respect to lab550 and to the expression domain of the endogenous LAB protein (Figure 6M and legend). Surprisingly, the 48 bp element did not respond to DPP very efficiently; only weak activation was observed in the posterior midgut upon ubiquitous expression of DPP, and no DPP-dependent induction was observed in anterior endodermal cells (Figure 6P). More surprisingly, the expression of 48/95 was lost in abdA mutants, demonstrating that the element requires a signaling input from the visceral mesoderm in addition to DPP; this additional input is presumably lost in abdA mutants. Similar observations were made when the 1/95 element was analyzed in the same genetic backgrounds (Figure 6I–L). These results contrast with those obtained with lab550, which is strongly responsive to DPP in most cells of the midgut (Figure 6C and D), and demonstrate that although the 48 bp element containing the LAB/EXD site is essential for expression of lab550, it is not sufficient to reproduce the properties of the latter. Our experiments demonstrate that by themselves, neither the 5′ sub-elements (1/95 and 48/95) nor the 3′ sub-element (92/546) of lab550 can recapitulate the behavior of the latter, either with respect to the expression domain and the expression levels, or with respect to the DPP inducibility. Therefore, it appears that input through both sub-elements synergizes to allow for a response to homeotic gene products and to DPP signaling (see Discussion). Discussion A lab enhancer represents a direct autoregulatory target We have identified a single binding site for a LAB/EXD heterodimer in an autoregulatory element of the lab gene (lab550). Based on the following observations, we suggest that this site represents an in vivo target for the homeotic protein LAB and its co-factor EXD: (i) genetically, the expression of lab550 depends strongly on the presence of LAB and EXD (Figure 3); (ii) an essential 48 bp region of the lab550 element contains a site to which LAB/EXD bind cooperatively in vitro, and which is identical at nine out of 10 positions to a HOXB-1/PBX site identified in the 5′ promoter region of the mouse Hoxb-1 gene (Pöpperl et al., 1995); (iii) point mutations in the LAB and EXD subsites abolished cooperative DNA binding in vitro (Figure 5) and resulted in strong reduction of the full-length lab550 enhancer activity in vivo (Figure 4). Taken together, the results of the genetic, the reverse genetic and the biochemical analyses support the view that lab550 is a direct in vivo target for the LAB/EXD complex. Lab550 represents the first bona fide Drosophila target for the LAB/EXD protein–protein complex, which has served as a paradigm to study how HOM/HOX genes regulate their targets. In contrast to most other HOM/HOX targets previously characterized, the lab enhancer represents a very simple situation in being regulated by the LAB/EXD complex through one single, essential site. The identification of an in vivo target site for a LAB/EXD complex in Drosophila allowed us to dissect the function of a HOM gene product on a natural target and its int
Referência(s)