Positive autoregulation of the glial promoting factor glide/gcm
1998; Springer Nature; Volume: 17; Issue: 21 Linguagem: Inglês
10.1093/emboj/17.21.6316
ISSN1460-2075
AutoresAlita A. Miller, Roberto Bernardoni, Angela Giangrande,
Tópico(s)Retinoids in leukemia and cellular processes
ResumoArticle2 November 1998free access Positive autoregulation of the glial promoting factor glide/gcm Alita A. Miller Alita A. Miller Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Roberto Bernardoni Roberto Bernardoni Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Angela Giangrande Corresponding Author Angela Giangrande Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Alita A. Miller Alita A. Miller Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Roberto Bernardoni Roberto Bernardoni Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Angela Giangrande Corresponding Author Angela Giangrande Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France Search for more papers by this author Author Information Alita A. Miller1, Roberto Bernardoni1 and Angela Giangrande 1 1Institut de Génétique et Biologie Moléculaire et Cellulaire, IGBMC/CNRS/INSERM/ULP, BP 163 67404 Illkirch, Communauté Urbaine de Strasbourg, France *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:6316-6326https://doi.org/10.1093/emboj/17.21.6316 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Fly gliogenesis depends on the glial-cell-deficient/glial-cell-missing (glide/gcm) transcription factor. glide/gcm expression is necessary and sufficient to induce the glial fate within and outside the nervous system, indicating that the activity of this gene must be tightly regulated. The current model is that glide/gcm activates the glial fate by inducing the expression of glial-specific genes that are required to maintain such a fate. Previous observations on the null glide/gcmN7-4 allele evoked the possibility that another role of glide/gcm might be to maintain and/or amplify its own expression. Here we show that glide/gcm does positively autoregulate in vitro and in vivo, and that the glide/gcmN7-4 protein is not able to do so. We thereby provide the first direct evidence of both a target and a regulator of glide/gcm. Our data also demonstrate that glide/gcm transcription is regulated at two distinct steps: initiation, which is glide/gcm-independent, and maintenance, which requires glide/gcm. Interestingly, we have found that autoregulation requires the activity of additional cell-specific cofactors. The present results suggest transcriptional autoregulation is a mechanism for glial fate induction. Introduction Cell fate determination relies on positional cues laid during the early stages of development. Since such cues are provided transiently within the cell, the assignment of a stable fate identity involves the maintenance and/or the amplification of the initial information, a step that often requires autoregulation. A typical example is provided by the segmentation genes, which establish cell fates along the antero-posterior axis in Drosophila melanogaster. Direct and indirect feedback loops have been described in both gap genes (Warrior and Levine, 1990) and pair-rule genes, transiently expressed activators and repressors that delimit the antero-posterior expression of homeotic genes (Hiromi and Gehring, 1987; Frasch et al., 1988; Lawrence and Johnston, 1989; Pick et al., 1990; Jiang et al., 1991). In addition, autoregulation is also required for the maintained expression of segment polarity genes (Siegfried et al., 1992; Yoffe et al., 1995) and that of the homeotic genes, although in this case gene expression is stable throughout development (Kuziora and McGinnis, 1988; Chouinard and Kaufman, 1991; Thuringer et al., 1993; for a review see Bienz, 1994). Interestingly, the control of gene activity through direct or indirect autoregulation is also used in pathways as diverse as those involved in sex determination (Bell et al., 1991; Keyes et al., 1992), cell cycling and proliferation (for examples see Wu et al., 1993; Johnson et al., 1994; Shan et al., 1994; Yin et al., 1994; for reviews see Pines, 1992; Lam and La Thangue, 1994), and circadian rhythms (Carter and Murphy, 1996; Foulkes et al., 1996; for a review see Sassone-Corsi, 1998). The use of such a strategy in these pathways suggests that the reinforcement of a primary signal is a crucial step in the establishment of a new cell state. Glial fate determination depends on the expression of glial cell deficient/glial cell missing (glide/gcm), a gene that is necessary and sufficient to activate the glial fate (Hosoya et al., 1995; Jones et al., 1995; Vincent et al., 1996; Bernardoni et al., 1998; for reviews see Anderson, 1995; Pfrieger and Barres, 1995; Giangrande, 1996; Klambt et al., 1996). glide/gcm is a transcription factor that binds to an octamer sequence called the glide/gcm binding site, or GBS (Akiyama et al., 1996; Schreiber et al., 1997). Previous observations on glide/gcmN7-4, a loss-of-function mutation, suggested that glide/gcm activity might possibly be controlled by autoregulation. In this mutant, glide/gcm RNA initially accumulates at levels that are indistinguishable from those observed in wild-type embryos; however, at late developmental stages glide/gcm transcripts cease to accumulate and are extinguished earlier than in the wild-type cells (Jones et al., 1995; Bernardoni et al., 1997). In this paper we show that the glide/gcmN7-4 allele carries a point mutation in the DNA-binding domain (DBD) which abolishes DNA binding and transactivating activities. We demonstrate that five GBSs exist within the proximal 6.5 kb of the glide/gcm promoter sequence and that the binding of glide/gcm to these sites activates transcription. In addition, we show that the five sites contribute differentially to the promoter activity. Finally, we demonstrate that glide/gcm autoregulates in vivo. By characterizing the autoregulation of glide/gcm we have defined the first direct target and regulator of glide/gcm. The role of positive autoregulation as a mechanism to control glide/gcm activity during glial differentiation is discussed. Results To test the possibility that glide/gcm expression is controlled by positive autoregulation during development, we undertook three approaches. First, we characterized the glide/gcmN7-4 mutation and assessed how it may affect glide/gcm activity. Secondly, we determined whether glide/gcm-binding sites exist in the glide/gcm promoter and if its transcription is glide/gcm-dependent. Thirdly, we determined whether autoregulation occurs in vivo. glide/gcmN7-4 is a point mutation in the DNA binding domain which abolishes binding and transactivation glide/gcmN7-4, a null mutation in which glide/gcm transcript levels decrease more rapidly than in wild-type embryos (Jones et al., 1995; Bernardoni et al., 1997), has been induced by DEB treatment (Lane and Kalderon, 1993), which suggests the presence of a point mutation. Using PCR amplification and sequence analysis on genomic DNA from mutant embryos we found that this mutation consists of a single base change (G→C) within the glide/gcm transcribed sequence. This results in the conversion of the cysteine residue at position 93 to a serine (Ser93) (Figure 1A). Cys93 constitutes the first of seven cysteine residues conserved among all vertebrate homologs (Akiyama et al., 1996; Altshuller et al., 1996; M.Kammerer and A.Giangrande, in preparation) and lies within the DBD of glide/gcm (Akiyama et al., 1996; Schreiber et al., 1997). Figure 1.glide/gcmN7-4 is defective in transactivation and DNA binding activity. (A) Schematic representation of the glide/gcm protein, which shows a conserved region contained in the DBD (Schreiber et al., 1997), a nuclear localization signal (NLS) and an activation domain (AD). The sequence of the DBD is shown below it, with the conserved cysteines indicated in boxes. The glide/gcmN7-4 allele consists of a point mutation within the DBD, in which the first conserved cysteine (position 93) is converted into a serine. (B) Relative abilities of increasing amounts of the expression vectors pPAC-glide/gcm (●) and pPAC-glide/gcmN7-4 (O) to activate transcription from 1 μg of the reporter pBLCAT5-GBS in transfected S2 fly cells. (▵) CAT activity measured upon co-transfection of increasing amounts of pPAC-glide/gcm with a reporter carrying a non-specific sequence instead of the GBS, pBLCAT5-NS. (C) Western blot analysis of extracts from S2 cells transfected with 5 μg of either pPAC alone (control), pPAC-glide/gcm (WT) or pPAC-glide/gcmN7-4 (N7-4). Anti-actin was used as a loading control. Filled and open arrows indicate the bands corresponding to glide/gcm and actin products, respectively. Molecular weight is indicated in kDa. (D) Relative abilities of increasing amounts of purified glideDBD fusion proteins (0–2 μg) to bind to a labeled 30mer containing the consensus GBS. Wild-type is shown on the right and glide/gcmN7-4 on the left. Download figure Download PowerPoint Binding-site selection assays have shown that glide/gcm binds to the DNA consensus sequence 5′-AT(G/A)CGGG(T/C)-3′ (Akiyama et al., 1996; Schreiber et al., 1997) and activates transcription from this binding site in transient transfection assays (Schreiber et al., 1997). To assess the ability of glide/gcmN7-4 protein to activate transcription, we synthesized a 30mer carrying the consensus GBS 5′-ATGCGGGT-3′ in the center and cloned it into a chloramphenicol acetyl transferase (CAT) reporter plasmid which contains a thymidine kinase (tk) promoter preceding the CAT coding sequence (pBLCAT5). We then tested the ability of wild-type and mutant glide/gcm to activate transcription from this fragment by cotransfecting the reporter construct with an expression vector containing the wild-type or the glide/gcmN7-4 cDNA in the S2 Drosophila cell line and by determining the relative amounts of CAT activity as compared with the same reporter containing a 30mer of non-specific DNA. Using increasing amounts of expression vector, we observed that the wild-type protein transactivates gene expression from the GBS in a dosage-dependent manner. The reporter containing non-specific DNA was inactive even in the presence of the highest levels of transfected glide/gcm. Strikingly, the glide/gcmN7-4 protein is completely unable to transactivate, even when high amounts of expression vector are used (Figure 1B). Western blot analysis showed that both wild-type and mutant glide/gcm proteins are expressed at the same levels in transfected S2 cells (Figure 1C). Because of the position of the mutation, we then assessed whether the defect in transactivation is due to the inability of glide/gcmN7-4 to bind to DNA. To this end, we conducted gel-shift assays using as probes the 30mer containing the consensus GBS tested in transfection experiments. Purified recombinant GST fusion proteins carrying the first 202 aa of wild-type or mutant glide/gcm were tested for DNA binding activity. This region of the protein includes the DBD (as defined by Akiyama et al., 1996). Figure 1D shows that the wild-type GST–glideDBD protein specifically forms complexes with the GBS while DNA binding is completely abolished when the Ser93 GST–glideDBD fusion protein is used, illustrating the importance of the cysteine residue at position 93. This result is also in agreement with the recent finding that a homologous mutation in one of the mouse GCM genes, the replacement of cysteine at position 76 with a serine, completely abolishes binding (Schreiber et al., 1998). We therefore conclude that glide/gcmN7-4 is unable to transactivate the reporter in the CAT assay because it is defective in its DNA-binding activity. Thus, the decay in transcript levels in glide/gcmN7-4 embryos is probably due to lack of direct or indirect autoregulation rather than to defects in RNA stability. glide/gcm positively regulates glide/gcm transcription in vitro As a second step in defining an autoregulatory loop for glide/gcm, we sequenced the glide/gcm promoter and searched for sites to which the protein would bind. Our sequence analyses revealed the presence of five GBSs scattered throughout 6.5 kb upstream of the glide/gcm transcription start site (Figure 2A and B) whereas none were found in >5 kb of sequence downstream of it. One of the GBSs corresponds to the octamer consensus sequence (site C), while the four other sites display a nucleotide change at the seventh (site A) or at the sixth position (sites B, D and E; Figure 2B). Interestingly, the promoter of reverse polarity (repo), a putative target of glide/gcm, also contains several GBSs (Akiyama et al., 1996). Two of the 11 binding sites correspond to the consensus, while all the others display a mismatch at one of the eight positions. Figure 2.glide/gcm positively autoregulates in vitro. (A) Schematic depiction of the five glide/gcm binding sites (GBS, sites A through E) found within 6.5 kb of the glide/gcm promoter. Horizontal arrow indicates the transcription start site. (B) Sequence of the five GBSs (left column) and relative ability to support transactivation when cloned into a pBLCAT5 reporter [as determined by relative CAT activity (right columns)]. Each GBS was tested for transactivation potential in both orientations. Sites indicated by an asterisk are oriented 3′–5′ in the 6.5 kb promoter. (C) Relative amounts of CAT activity upon cotransfection of increasing amounts of pPAC-glide/gcm and the CAT5 reporter construct containing either one of the five GBSs, non-specific DNA (NS) or no additional cloned sequences (CAT5). Symbols corresponding to each reporter are shown on the right. (D) Histogram depicting results for maximum CAT activity observed for each site in the presence of 5 μg pPAC-glide/gcm. Only the orientation found in the glide/gcm promoter is shown in (C) and (D). Download figure Download PowerPoint The presence of mismatches in a binding site may induce a change in the DNA binding and in the transactivation abilities compared with those observed with the consensus sequence. For example, mutations at positions 2, 3, 6 and 7 in the GBS have a strong impact on the binding of mouse GCM (Schreiber et al., 1998). We therefore proceeded to analyze in vitro the function of the five GBSs. We synthesized 30mers carrying the different sites and cloned them individually into pBLCAT5. We then tested the ability of glide/gcm to activate transcription from each of these fragments in cotransfection experiments as above (Figure 2C and D). The five sites found in the glide/gcm promoter do indeed display different activation potentials. Sites C and A induce the highest levels of CAT activity, 100- and 25-fold, respectively, compared with the activity observed with control reporter plasmids not containing the GBS or containing a 30mer of random DNA sequence. Sites B, D and E are poorly active, their levels of CAT activation being 2.4-, 5.2- and 2.2-fold, respectively, compared with the control. From these data we conclude that sites C and A are the main sites for transcriptional activation, and that this activation is mediated through the activity of the glide/gcm protein. We have observed that sites A and D reside in the opposite orientation with respect to that of B, C and E. Interestingly, sites present in the opposite orientation were also observed in three of the 11 GBSs found in the repo promoter. To determine whether activation of transcription is dependent on the orientation of the GBS, we tested the transactivation potential of each site in its reverse orientation relative to the reporter gene. The results in Figure 2B show that the level of glide/gcm-mediated transcriptional activation varies only slightly depending on the orientation of the binding site. Binding properties of the glide/gcm binding sites in the glide/gcm promoter In order to further analyze the role of the five GBSs found in the glide/gcm promoter we also determined their relative binding affinity. Binding assays were performed using the purified recombinant GST–glideDBD fusion protein and the six 30mers already tested in transfection experiments. Figure 3A shows that GST–glideDBD forms complexes with notably different affinities for the five sites. Site C, which corresponds to the consensus site as defined by binding site selection (Akiyama et al., 1996; Schreiber et al., 1997), displays the highest affinity. The glideDBD fusion protein binds to site A with ∼9-fold lower affinity than to site C. Interestingly, site A contains an A instead of a G, a change observed in only 2% of the cases in the binding-site selection assay. Sites B, D and E, which present a mismatch at the sixth position, display very poor (sites E and D) or no (site B) affinity for glideDBD. This is in agreement with the observations that the glide/gcm binding sites identified through site selection assays always contained a G residue at the sixth position and that sites mutagenized at this position are inactive (Schreiber et al., 1998). Interestingly, site A (ATGCGGAC) is still active, while a site that introduces a T at the seventh position is not (Schreiber et al., 1998), suggesting that the activity of a GBS also depends on the type of residue at a given position. Figure 3.glide/gcm displays different affinities for the five GBSs found within 6.5 kb of the glide/gcm promoter. (A) Gel-shift assay to determine the ability of purified glideDBD fusion protein to bind to labeled 30mers containing the five different GBSs or non-specific DNA. Site C, which contains the consensus GBS, displays the highest level of binding, followed by site A, with ∼9-fold lower affinity. ‘−’ and ‘+’ indicate the absence and presence of the fusion protein, respectively. (B) Competition gel-shift assay on sites C and A. One- to one-hundred-fold excess of non-specific (NS) or specific (S) cold DNA competitor was preincubated with 500 ng of purified glideDBD fusion protein, followed by incubation with labeled DNA. (C) Gel-shift assay to determine the ability of full-length glide/gcm to bind to the A, B and C GBSs or to non-specific DNA. Nuclear extracts from S2 cells stably transformed with glide/gcm were prepared after treatment in both inducing (+) and non-inducing (−) conditions. One microgram of extract was incubated with labeled 30mers as in (A). Arrow indicates the band which is specific for glide/gcm binding. Note that full-length glide/gcm displays the same relative affinities for the five GBSs as the purified glideDBD fusion protein. (D) Competition gel-shift assay on site C using induced nuclear extracts from stably transformed cells, performed as in (B). Download figure Download PowerPoint The specificity of the binding was confirmed by challenging the glide/gcm–DNA complex with increasing amounts of specific and non-specific DNA competitors. In these experiments, the specific competitor was the 30mer used as probe. The protein–DNA complex between glide/gcm and site C is affected upon the addition of an equal amount of cold specific competitor DNA and nearly completely disrupted upon the addition of 10-fold excess competitor (Figure 3B). In addition, the complex is able to withstand the addition of up to 100-fold of non-specific competitor. Therefore, the glide/gcm protein displays a very high affinity for its consensus binding site. Similar results were obtained using the A-binding site, but in this case binding was completely abolished by adding an equal amount of specific competitor due to the lower affinity of site A for glide/gcm (Figure 3B). To eliminate the possibility that the differences in affinity arise from the use of a truncated glide/gcm protein, we also tested the entire protein. To this end, DNA probes were incubated with nuclear extracts from a stably transformed S2 cell line in which glide/gcm expression is under the control of a metal-inducible promoter. DNA-binding assays were performed under both inducing and non-inducing conditions (Figure 3C). Expression of glide/gcm under inducing conditions was verified by Western blot analysis (data not shown). For the consensus GBS (site C), a specific band appears upon incubation with induced nuclear extracts that is not present in non-induced samples, or when induced nuclear extracts are incubated with a 30mer containing non-specific DNA. As in the case of the glideDBD fusion protein, the five GBSs display different binding affinities for the full-length glide/gcm protein, with site C being the strongest followed by weak binding observed for site A. Sites B, D and E are totally inactive in this test (Figure 3C; data not shown). Importantly, the binding affinity of either the full-length glide/gcm or a fusion protein containing only its DBD for the GBSs corresponds well to the degree of their transactivation abilities revealed in the above CAT assays (compare with Figure 2C and D). The specificity of the binding was also confirmed by adding specific and non-specific competitor DNA (Figure 3D). The addition of 10-fold excess of specific competitor completely eliminates the protein–DNA complex (low mobility band), while this same complex is not affected by the addition of 10-fold excess of non-specific competitor. glide/gcm-dependent activation of the glide/gcm promoter In order to determine the contribution of each GBS to autoregulation within the context of the glide/gcm promoter, we assessed the transactivation potential of the entire 6.5 kb fragment. Since this fragment carries the glide/gcm promoter and the transcription start site, a CAT reporter vector devoid of tk promoter sequences (pBLCAT6) was used instead of pBLCAT5. The 6.5 kb fragment induces a basal level of CAT activity in the absence of transfected glide/gcm (Figure 4A). This activity is due to the sequences present in the 6.5 kb because no activity was observed for the pBLCAT6 reporter vector itself (Figure 4A). Such basal activity may depend on the presence of positive transcription factors that act on the glide/gcm promoter and/or to the presence of endogenous glide/gcm, since anti-glide/gcm antibodies detect a product in non-transfected S2 cells in Western blot assays (Figure 1C). The intensity of the signal detected in non-transfected cells, however, is much weaker than that detected upon transfection with a glide/gcm expression vector. In agreement with this, we have found that a 3.3-fold increase of CAT activity takes place upon cotransfection of the 6.5 kb reporter vector with a glide/gcm expression vector while cotransfection with the expression vector devoid of the glide/gcm cDNA or carrying the glide/gcmN7-4 sequences does not result in any increase in enzymatic activity (Figure 4A). Figure 4.glide/gcm activates transcription from the glide/gcm promoter in vitro. (A) A reporter construct containing 6.5 kb of glide/gcm promoter sequence (WT-CAT6) was tested for transactivation potential when co-transfected in S2 cells with 5 μg of the following: carrier DNA (none, column 2), pPAC (column 3), pPAC-glide/gcm (pPAC-WT, column 4) or pPAC-glide/gcmN7-4 (pPAC-N7-4, column 5). CAT6 was co-transfected with 5 μg of pPAC-glide/gcm as a control (column 1). The relative CAT activity for each reporter is shown. (B) Reporter constructs with individual (ΔA–CAT6 through ΔE–CAT6, grey columns 2–6) or all (ΔA–E-CAT6, grey column 7) GBSs abolished by mutation were tested for transactivation potential when co-transfected with 5 μg of pPAC-glide/gcm. The degree of activation of the wild-type promoter (WT-CAT6, grey column 1) is arbitrarily assigned 100%. The basal activity was also evaluated for each reporter (−glide/gcm) and is represented in the white columns. Download figure Download PowerPoint To determine the contribution of each GBS to overall transcriptional activation in the context of the entire 6.5 kb promoter, we created five reporter constructs (ΔA through ΔE), each with one GBS replaced by the sequence AGGGAAAC (Figure 4B). Gel-shift assays previously showed that glide/gcm is not capable of binding to this sequence alone (Figure 3). The mutation of sites E and B did not significantly affect the transactivation potential of the 6.5 kb promoter, which we arbitrarily define as 100% (Figure 4B). In contrast, mutation of site C resulted in a decrease in transcriptional activation to 43%, whereas a decrease in activity to 67% was observed upon the mutation of site A (Figure 4B), once the basal activity is taken into account. Site D behaved like site A, although with higher variability (see standard deviation bars in Figure 4B). We also tested a reporter with all five GBSs mutated (ΔA–E) and a construct that contained mutations in sites A, B and C (ΔA–C). Most strikingly, the ability to respond to cotransfected glide/gcm is almost completely abolished when the ΔA–E construct is used, with an observed CAT activity of 16% compared with the wild-type promoter (Figure 4B). Similar results were obtained with the ΔA–C construct (data not shown). These data altogether clearly demonstrate that (i) the activation of glide/gcm promoter depends on autoregulation; (ii) the most important GBSs for autoregulation are C, the consensus site, and to a minor extent, sites A and D; and (iii) glide/gcm activation through either site C or A takes place independently from the other site. The ΔA–E construct still retains some glide/gcm-dependent transcriptional activation, despite the absence of any site through which glide/gcm can directly mediate its effects (Figure 4). This could imply that transfected or endogenous glide/gcm activates downstream genes, which in turn would positively regulate the glide/gcm promoter at sites other than the GBSs. Alternatively, glide/gcm may act in concert with other factors to positively regulate its transcription. The role of glide/gcm in this activation is also in agreement with the observation that the ΔA–E construct is less active in the absence of transfected glide/gcm. Finally, part of the basal activation observed in the absence of the GBSs and in the absence of transfected glide/gcm may depend on the sole activity of factors acting on other cis elements present throughout the glide/gcm promoter. glide/gcm autoregulates in vivo The previous results indicate that direct autoregulation of glide/gcm takes place in a cell line. To demonstrate that glide/gcm regulates its own expression in vivo, we used two approaches. First, we analyzed the accumulation of the glide/gcm product in glide/gcmN7-4 embryos. Whilst in wild-type embryos glide/gcm is expressed in all glial precursor cells (Figure 5A and B), in mutant embryos its expression is limited to a few cells at the periphery of the ventral cord (Figure 5C and D). In agreement with the in situ hybridization results and as predicted from a mutation affecting autoregulation, the difference in the pattern of glide/gcm expression between wild-type and mutant embryos was not detectable at early developmental stages (data not shown). Together with the data obtained in vitro, this result strongly suggests that autoregulation does take place during glial differentiation. Figure 5.Expression of glide/gcm protein is defective in glide/gcmN7-4 mutant embryos. Ventral views of late stage 12 embryos, total projections of confocal images. Anti-glide/gcm labeling in wild-type (A and B) and mutant embryos (C and D). Vertical white lines in (A) and (C) indicate region that is shown in (B) and (D), respectively. Bar: 25 μm in (A and C); 50 μm in (B and D). Download figure Download PowerPoint As a second approach, we expressed glide/gcm ectopically and determined whether this can induce autoregulation using the original enhancer trap line, rA87, in which the lacZ gene is under the control of the glide/gcm promoter (Hosoya et al., 1995; Jones et al., 1995; Vincent et al., 1996), to follow glide/gcm expression. We employed the upstream activating sequences (UAS)-GAL4 system (Brand and Perrimon, 1993) to express glide/gcm ectopically and determined whether this would induce ectopic β-galactosidase (β-gal) expression in a rA87 background. rA87 was therefore crossed with UAS-glideM21G (M21G) (Bernardoni et al., 1998), a transgenic line which carries glide/gcm under the control of the UAS, the targets of the GAL4 transcription factor (Fischer et al., 1988). The rA87/+; M21G/+ progeny were then crossed with a scabrous-GAL4 (sca-GAL4) driver which mimicks the scabrous profile of expression in the whole neurogenic region (Mlodzik et al., 1990; data not shown). Ectopic glide/gcm expression does indeed promote activation of the endogenous glide/gcm promoter since massive β-gal labeling was observed throughout the ventral cord (Figure 6a). We and others have shown previously that ectopic glide/gcm leads to ectopic activation of the glial differentiation program (Hosoya et al., 1995; Jones et al., 1995; Bernardoni et al., 1998). The colocalization between the glial-specific repo product and β-gal indicates that glide/gcm autoregulation takes place in cells that take the glial fate at ectopic positions. Figure 6.In vivo analysis of glide/gcm aut
Referência(s)