Inducible degradation of the Drosophila Mediator subunit Med19 reveals its role in regulating developmental but not constitutively-expressed genes

Abstract

The multi-subunit Mediator complex plays a critical role in gene expression by bridging enhancer-bound transcription factors and the RNA polymerase II machinery. Although experimental case studies suggest differential roles of Mediator subunits, a comprehensive view of the specific set of genes regulated by individual subunits in a developing tissue is still missing. Here we address this fundamental question by focusing on the Med19 subunit and using the Drosophila wing imaginal disc as a developmental model. By coupling auxin-inducible degradation of endogenous Med19 in vivo with RNA-seq, we got access to the early consequences of Med19 elimination on gene expression. Differential gene expression analysis reveals that Med19 is not globally required for mRNA transcription but specifically regulates positively or negatively less than a quarter of the expressed genes. By crossing our transcriptomic data with those of Drosophila gene expression profile database, we found that Med19-dependent genes are highly enriched with spatially-regulated genes while the expression of most constitutively expressed genes is not affected upon Med19 loss. Whereas globally downregulation does not exceed upregulation, we identified a functional class of genes encoding spatially-regulated transcription factors, and more generally developmental regulators, responding unidirectionally to Med19 loss with an expression collapse. Moreover, we show in vivo that the Notch-responsive wingless and the E(spl)-C genes require Med19 for their expression. Combined with experimental evidences suggesting that Med19 could function as a direct transcriptional effector of Notch signaling, our data support a model in which Med19 plays a critical role in the transcriptional activation of developmental genes in response to cell signaling pathways.

Fig 1. K-in of the Auxin Induced Degron (AID) for fast and deep depletion of Med19 in the Drosophila wing imaginal disc.

(A) Principle of auxin-dependent degradation of Med19^AID in Drosophila. The Med19 endogenous gene located on the third chromosome has been engineered using CRISPR-Cas9 system in order to generate Med19 fusion protein with AID and GFP tags. The expression of the auxin-dependent and AID-specific F-box protein TIR1 is controlled by the UAS/Gal4 system. The Gal4 driver is located in the second chromosome, and the UAS-TIR1 construct in the third chromosome. The F-box TIR1 incorporates the endogenous SCF (Skip1 Cul1 F-box) E3 ubiquitin ligase complex. Auxin (added to the fly food) triggers poly-ubiquitination of Med19^AID by the SCF-TIR1 targeting the Mediator subunit to degradation by the proteasome. (B) PCR genotyping for the insertion of the AID-GFP coding sequence in the Med19 locus using the primers couple indicated in (A) with genomic DNA extracted from WT control (Med19⁺ left lane), Med19⁺/Med19^AID heterozygous (middle lane), or Med19^AID homozygous (right lane) adult flies. WT and Med19^AID alleles theoretically produce a 270 and 1200 pb PCR product, respectively. (C) Ubiquitous Auxin-dependent degradation of Med19^AID in the wing imaginal disc. Images showing DAPI and GFP (detection of Med19^AID) signals from ub>TIR1; Med19^AID wing discs dissected from larvae not exposed to auxin (no aux.), or fed with auxin for 1 hour (1H), or 4 hours (4H). (D) GFP signal quantification (Arbitrary Unit), as measurement of Med19^AID depletion, in ub>TIR1; Med19^AID wing imaginal discs (images shown in (C)) obtained from larvae not exposed to auxin (no Aux., n = 9), or fed with auxin during 1 hour (1H, n = 11), or 4 hours (4H, n = 12). (E) Auxin dependent degradation results in loss of both C-terminal GFP and N-terminal Med19 moieties of Med19^AID. Images showing en>TIR1 Med19^AID wing discs, in which auxin-induced degradation in the posterior compartment was performed for 4 hours, co-immunostained using anti-GFP and Med19 antibodies. (F) Auxin dependent degradation of Med19^AID does not affect Med12 expression. Wing discs from en>TIR1 Med19^AID larvae not exposed to auxin (no auxin- upper panel.), or fed with auxin (auxin -bottom panels). Co-immunostaining as in (E) except antibodies against Med12 were used.

Fig 2. Med19^AID removal results in the deregulation a limited part of the wing imaginal disc transcriptome.

(A) Scheme depicting the experimental setup used to produce wing imaginal discs from ub>TIR1 Med19^AID larvae in which Med19^AID degradation has been triggered, or not (no Auxin control- see GFP signal), upstream of the RNA-seq analysis. (B) Differential gene expression analysis volcano plot showing the expression fold change after Med19^AID degradation and associated adjusted p-values. 8753 genes found expressed in the wing imaginal disc are depicted. Genes with significative fold change (adjusted p-value <0,05) are shown in blue for the down-regulated (log₂FC <-0,5) and in red for the up-regulated (log₂FC >0,5).

Fig 3. Spatially and temporally regulated genes are more prone to expression change than constitutively expressed genes upon Med19^AID depletion.

(A) Heat map depicting the spatio temporal expression profiles (modENCODE) of genes belonging to three different ranges of expression fold change (L2FC specified above the horizontal axis) in our transcriptomic analysis. Each column contains the color-coded expression score (modENCODE) of a gene in 58 different tissues and developmental stages (left panel and S2 Table for details) extracted from the modENCODE Drosophila gene expression database the expression profiles of each gene we found expressed in the wing imaginal disc, compiling expression scores in 58 different developmental stages and tissues. (B) Box plot analyzing the count of the number of null expression (0 and 1 expression score occurrences) among the 58 expression values for the genes belonging to |L2FC|>1 (n = 1185) or -0.1

Fig 4. Med19 promotes the expression of the spatially regulated developmental genes.

(A) Scatter plot depicting the distribution of the Log2FC values of the 573 transcription factors found significantly expressed in the wing imaginal disc from the RNAseq data. The transcription factors with a demonstrated spatial transcriptional regulation (SR TFs) in the wing imaginal disc are shown in purple (n = 122). (B) Dot plot 10x10 showing the proportion of the non-regulated (dark grey), upregulated (red), or downregulated (blue) genes among the 122 SR transcription factors. (C) Bar graph made from RT-qPCR data showing the relative level of transcripts of 6 genes known to be spatially regulated (wg, cut, E(Spl)m8, sal or ubiquitous (Myosin, E2F1) in the wing discs of ub>; Med19^AID-GFP larvae after Med19^AID depletion was carried out for 4 hours in all the disc cells (percentage of the control in which Med19^AID degradation was not triggered). The graph shows the mean and standard deviation of 3 independent experiments. (D) Confocal images of the wing pouch region of dissected en>; Med19^AID wing discs that were subjected to Med19AID degradation (en> 3H Aux), or not (control: No Gal4 driver, and en> no Aux), and in which wingless (wg) pre-mRNA were visualized by smiFISH using intronic probes and Med19 degradation by direct GFP fluorescence. top images:Med19^AID GFP signal. bottom images: wg smiFISH signal. (E) Wing disc D/V boundary posterior to anterior wg smiFISH signal ratio established in en>; Med19^AID wing discs: No auxin control (no Aux. n = 5), 1H auxin (n = 6), and 3H auxin (n = 6). _****: P value <0.0001 (Unpaired t test). (F) Hierarchical clustering tree of the 30 best FDR of the GO terms associated with the downregulated genes. The size of the solid circles is proportional to the enrichment False Discovery Rate (FDR). GO clusters appear related to development (purple), nucleic acid synthesis (blue), and cell cycle regulation (yellow).

Fig 5. Med19 is required for the expression of Notch responsive genes in the wing imaginal disc.

(A) Scatter plot and median of the expression fold change (Log2FC), obtained from the Med19^AID RNAseq DGE analysis, of the 11 Enhancer of Split complex (E(spl)-C) genes, and genes encoding TFs involved in the transcriptional regulation of the E(spl)-C genes, namely Suppressor of Hairless Su(H), Hairless (H), Groucho (Gro), Mastermind (Mam), Notch (N), and CtBP. The red dotted lines delineate the -0.5<Log2FC<0.5 expression fold change zone.(B, C) Maximum projection confocal microscopy images of Med19^AID GFP signal and transcripts detection using smiFISH probes in the pouch of wing imaginal discs in which Med19^AID has been regionally degraded, or not. Driving the degradation of Med19^AID for 3 hours in the dorsal domain of the wing disc using the ap>; Med19^AID genetic set-up severely affects the level of E(spl)m3 (B), and E(spl)mbeta (C) transcripts specifically in the area where Med19 is depleted. (D) Western blot analysis of a GST pull-down experiment showing that in vitro translated HA-tagged Su(H)–detected using an anti-HA antibody–physically interacts with GST-Med19, but not with GST. Input: 5% of the starting material.

Fig 6. Model, derived from the analysis of the nature of the genes deregulated or not after Med19^AID degradation, showing how Med19 may function.

(A) Med19 acts as a critical cofactor of transcription factors whose activity is regulated by signaling pathways, in negative or positive control of transcription. (B) Med19 is not a Mediator subunit required for the expression of the constitutively expressed genes.