The repressor and activator forms of Cubitus interruptus control Hedgehog target genes through common generic gli-binding sites

Development, Jul 2000

B. Muller, K. Basler

Article PDF cannot be displayed. You can download it here:

https://dev.biologists.org/content/127/14/2999.full.pdf

The repressor and activator forms of Cubitus interruptus control Hedgehog target genes through common generic gli-binding sites

Bruno Mller Konrad Basler ) SUMMARY The Drosophila Gli homolog Cubitus interruptus (Ci) controls the transcription of Hedgehog (Hh) target genes. A repressor form of Ci arises in the absence of Hh signalling by proteolytic cleavage of intact Ci, whereas an activator form of Ci is generated in response to the Hh signal. These different activities of Ci regulate overlapping but distinct subsets of Hh target genes. To investigate the mechanisms by which the two activities of Ci exert their opposite transcriptional effect, we dissect here the imaginal disc enhancer of the dpp gene, which responds to both activities of Ci. Within a minimal disc enhancer, we identify the DNA sequences that are necessary and sufficient for the control by Ci, show that the same sequences respond to the Hedgehog signalling proteins regulate a wide variety of developmental events throughout the animal kingdom. Most, if not all, Hedgehog target genes are controlled through members of the Gli family of transcription factors (reviewed by Altaba, 1999; Ingham, 1998). Major interest is devoted therefore to the mechanisms by which Gli proteins control gene expression. Important advances have recently been made in the system of Drosophila limb development, where the function and mode of action of Hedgehog (Hh) and the Gli homolog Cubitus interruptus (Ci) are best understood. Key to our present understanding of Hh signalling was the finding that Ci is regulated post-transcriptionally. In cells that do not receive the Hh signal, a low molecular weight form of Ci, Ci-75, is generated through proteolytic cleavage of the full-length protein Ci-155 (Aza-Blanc et al., 1997). This truncated form of Ci functions as a transcriptional repressor and is here referred to as Ci[rep]. The reception and transduction of the Hh signal prevents the formation of Ci[rep] and at the same time causes the conversion of Ci-155 into an activator form (Ci[act]) (Chen et al., 1999; Mthot and Basler, 1999; Price and Kalderon, 1999; reviewed by Aza-Blanc and Kornberg, 1999). The existence of two forms of Ci with opposite transcriptional activities the formation of each being subject to Hh control allows multiple modes of target gene regulation. For example, and as described below, this setup not activator and repressor forms of Ci, and demonstrate that their activities can be replaced by a single synthetic Glibinding site. We further show that the enhancer sequences of patched, a gene responding only to the activator form of Ci, effectively integrate also the repressor activity of Ci if placed into a dpp context. These results provide in vivo evidence against the employment of distinct binding sites for the different forms of Ci and suggest that target genes responding to only one form must have acquired distant cisregulatory elements for their selective behavior. only provides the means to induce target gene expression in response to Hh but also to ensure tight repression of the same or other genes in cells that do not receive the Hh signal. Genetic analyses in imaginal discs have recently indicated that the activator and repressor activities of Ci regulate overlapping but distinct subsets of Hh target genes (Mthot and Basler, 1999). Cells of the posterior compartment are programmed by the selector gene engrailed (en) to secrete Hh (reviewed in Lawrence and Struhl, 1996). During most developmental stages, En is not active in anterior compartment cells and, as a consequence, all these cells express ci instead of hh (Schwartz et al., 1995), and are thus competent to respond to the Hh signal (Zecca et al., 1995). Only cells in the vicinity of the anteroposterior (AP) compartment boundary, however, receive the Hh signal. The activation of the Hh transduction pathway causes the upregulation of patched (ptc) expression and the induction of the decapentaplegic (dpp) gene (Basler and Struhl, 1994; Capdevila et al., 1994; Tabata and Kornberg, 1994). At late stages of wing development, high levels of Hh signalling also induce the expression of the en gene (Blair, 1992; Guillen et al., 1995; Strigini and Cohen, 1997). Mutant anterior compartment cells entirely lacking ci function fail to upregulate ptc and en expression, which indicates that these two genes are primarily controlled by the activator form of Ci (Mthot and Basler, 1999). These cells, however, express low levels of dpp and hh, irrespective of their position (Dominguez et al., 1996; Mthot and Basler, 1999). This observation led to the interpretation that dpp is controlled by both forms of Ci, whereas hh is only subject to Ci-mediated repression (Mthot and Basler, 1999). Together these findings indicated that Ci[act] and Ci[rep] can have common as well as distinct targets and raised the question of how different target genes can be differentially sensitive to the two related forms of Ci. Here we address this question experimentally by identifying regulatory DNA elements for both Ci[act] and Ci[rep] and by exploring the context in which these elements can act. We envisage three scenarios through which differential sensitivity towards Ci[act] or Ci[rep] can be explained. The first possibility is that the two forms of Ci recognize, and act through, different binding sites. Although both forms appear to contain the same zinc-finger DNA-binding domain, it is conceivable that either the C-terminal half of Ci that is lacking in Ci[rep], or some signal-induced covalent modification of Ci[act] results in distinct DNA-binding specificities. The competence of target genes to respond to only one or to both forms of Ci could thus be encoded in the nucleotide sequence of the Ci responsive elements. If these elements are identical in structure, a second possibility would be that they differ in context. Either form of Ci could be specifically associated with DNA-binding cofactors. The function of the resulting complex would depend on the presence of adjacent binding sites for Ci and such cofactors. Finally, in a third scenario, neither sequence nor local context of Ci[act]- and Ci[rep]-binding sites differ. In some target genes, the function of bound Ci may be modulated by distant cis-regulatory elements to allow input by one, but not the other form of Ci. To discriminate between these possibilities, we set out to identify regulatory elements that mediate Ci[rep] and Ci[act] input. Starting with a 4 kb fragment of the dpp gene, we narrowed these elements down to 20 bp by an unbiased, functional assay. We found that both activities of Ci are mediated via the same DNA element, that this element contains a Gli-binding site, and that a synthetic 9 bp Gli consensus binding site can substitute for both activities. In addition, we show that even the Ci responsive element of ptc, a gene that normally only responds to Ci[act] input, can confer regulation by Ci[rep] if placed into the dpp enhancer context. Finally, we demonstrate that the ability of a Gli consensus binding site to respond both to Ci (...truncated)


This is a preview of a remote PDF: https://dev.biologists.org/content/127/14/2999.full.pdf
Article home page: http://dev.biologists.org/content/127/14/2999.abstract

B. Muller, K. Basler. The repressor and activator forms of Cubitus interruptus control Hedgehog target genes through common generic gli-binding sites, Development, 2000, pp. 2999-3007, 127/14,