Identifying targets of the Sox domain protein Dichaete in the Drosophila CNS via targeted expression of dominant negative proteins

Abstract

Background: Group B Sox domain transcription factors play important roles in metazoan central nervous system development. They are, however, difficult to study as mutations often have pleiotropic effects and other Sox family members can mask phenotypes due to functional compensation. In Drosophila melanogaster, the Sox gene Dichaete is dynamically expressed in the embryonic CNS, where it is known to have functional roles in neuroblasts and the ventral midline. In this study, we use inducible dominant negative proteins in combination with ChIP, immunohistochemistry and genome-wide expression profiling to further dissect the role of Dichaete in these two tissues.

Results: We generated two dominant negative Dichaete constructs, one lacking a DNA binding domain and the other fused to the Engrailed transcriptional repressor domain. We expressed these tissue-specifically in the midline and in neuroblasts using the UAS/GAL4 system, validating their use at the phenotypic level and with known target genes. Using ChIP and immunohistochemistry, we identified two new likely direct Dichaete target genes, commisureless in the midline and asense in the neuroectoderm. We performed genome-wide expression profiling in stage 8-9 embryos, identifying almost a thousand potential tissue-specific Dichaete targets, with half of these genes showing evidence of Dichaete binding in vivo. These include a number of genes with known roles in CNS development, including several components of the Notch, Wnt and EGFR signalling pathways.

Conclusions: As well as identifying commisureless as a target, our data indicate that Dichaete helps establish its expression during early midline development but has less effect on its established later expression, highlighting Dichaete action on tissue specific enhancers. An analysis of the broader range of candidate Dichaete targets indicates that Dichaete plays diverse roles in CNS development, with the 500 or so Dichaete-bound putative targets including a number of transcription factors, signalling pathway components and terminal differentiation genes. In the early neurectoderm we implicate Dichaete in the lateral inhibition pathway and show that Dichaete acts to repress the proneural gene asense. Our analysis also reveals that dominant negatives cause off-target effects, highlighting the need to use other experimental data for validating findings from dominant negative studies.

Figure 1

Dominant negative proteins mimic Dichaete phenotypes. A-F) BP102 staining reveals the ventral nerve cord of stage 16 embryos. All ventral views with anterior to the top. A) sim-GAL4;UAS-D; B) sim-GAL4;UAS-D^ΔHMG; C) sim-GAL4;UAS-D^EnRep; D) pros-GAL4;UAS-D; E) pros-GAL4;UAS-D^ΔHMG; F) pros-GAL4;UAS-D^EnRep. Arrows indicate breaks in the longitudinal connectives, arrowheads indicate collapse of commissures. G-L) FasII staining revealing the major longitudinal fascicles G) Wild type; H) D^r72/Df(3L)fz-GS1a; I) sim-GAL4;UAS-D^ΔHMG; J) sim-GAL4;UAS-D^EnRep; K) sim-GAL4;UAS-mSox2^ΔHMG; L) sim-GAL4;UAS-mSox2^EnRep. M-P) Anti-Slit staining to reveal the midline glia. M and N) Wild type at stage 13 (M) and 15 (N). O and P) sim-GAL4;UAS-D^ΔHMG at stage 13 (O) and 15 (P). M and O are ventral views with anterior to the left, N and P are lateral views with anterior to the left. Grey arrowheads indicate loss of Slit expression.

Figure 2

Dominant negative expression in a Dichaete mutant background. A-E) BP102 staining, all embryos oriented with anterior to the top. A) Wild type; B) D^r72/Df(3L)GS1a; C) D^r72/Df(3L)GS1a, simGAL4/UAS-mSox2; D) D^r72/Df(3L)GS1a; simGAL4/UAS-mSox2^ΔHMG; E) D^r72/Df(3L)GS1a; simGAL4/UAS-mSox2^EnRep .

Figure 3

Proneural gene expression. Ventral views of the medial and intermediate columns of the stage 11 embryonic neuroectoderm stained with anti-Achaete, anterior is to the top. The ventral midline is indicated by the dashed line and neuromeres in individual hemisegments indicated by brackets. A) Wild type; B) D^r72/Df(3L)fz-GS1a; C) pros-GAL4; UAS-D^ΔHMG; D) pros-GAL4; UAS-Sox2^ΔHMG; E) pros-GAL4; UAS-Sox2^EnRep .

Figure 4

Commisureless is a Dichaete target in the midline. Ventral views of stage 12 (A-E) and close-up of the midline in stage 14 (A’ – E’) embryos stained with anti-Comm. Anterior is to the left and the white arrowheads indicate the midline. A and A’) Wild type; B and B’) D^r72/Df(3L)fz-GS1a; C and C’) sim-GAL4; UAS-D^ΔHMG; D and D’) sim-GAL4; UAS-mSox2^EnRep; E and E’) Maternal-GAL4^VP16; UAS-mSox2^EnRep .

Figure 5

Effects of Dichaete on Asense expression. A – J) Ventral views of anti-Ase staining at the indicated stages from 9 to 16 in wild type (A, C, E, G and I) and D^r72/Df(3L)fz-GS1a (B, D, F, H and J). Anterior is to the left. K and L) Ventral views of stage 12 embryos stained with anti-Ase. Anterior is to the top and the midline is indicated by the dashed line. K) Wild type, L) pros-GAL4; UAS-D^ΔHMG .

Figure 6

ChIP assays detect Dichaete binding. Binding profiles of Dichaete in the early embryo (stage 5) generated by the Berkeley Drosophila Transcription Network Project [23] are shown in each graph with bound regions identified at 1% and 25% FDR shown as black bars above the ChIP profile. The grey bars represent 1% and 25% FDR regions identified by modENCODE (embryonic stages 1–12). The location of each amplicon assayed by specific ChIP-PCR assays is indicated by the numbered grey boxes below the GBrowse gene model for each region. Gene models above the chromosome scale line are transcribed from left to right and those below from right to left as indicated by the arrows in the introns. The PCR enrichments from anti-Dichaete and control immunopurification reactions are shown in the gel images below the gene models with the numbers indicating the amplicons. A) Approximately 55 kb encompassing the sli gene. The location of the known sli midline enhancer (fragment 13) is indicated by an asterisk and a grey bar above the gene model. B) 25 kb around the ase gene. C) 25 kb around comm.

Figure 7

Gene expression profiling DN mutants. A) Overlap between gene lists in the sim-GAL4 and pros-GAL4 experiments. B) Hierarchical clustering of genes with significant expression changes when DN proteins are driven in the neuroectoderm with pros-GAL4. C) Dichaete binding associated with genes changing expression in the pros-GAL4 study: dark (1% FDR) and light blue (25% FDR) bars represent binding intervals identified in the BDTNP Dichaete ChIP study [23]. D) Gene Expression Maps from FlyExpress [37] .