Polycomb repressive complex 2-dependent and -independent functions of Jarid2 in transcriptional regulation in Drosophila

Abstract

Jarid2 was recently identified as an important component of the mammalian Polycomb repressive complex 2 (PRC2), where it has a major effect on PRC2 recruitment in mouse embryonic stem cells. Although Jarid2 is conserved in Drosophila, it has not previously been implicated in Polycomb (Pc) regulation. Therefore, we purified Drosophila Jarid2 and its associated proteins and found that Jarid2 associates with all of the known canonical PRC2 components, demonstrating a conserved physical interaction with PRC2 in flies and mammals. Furthermore, in vivo studies with Jarid2 mutants in flies demonstrate that among several histone modifications tested, only methylation of histone 3 at K27 (H3K27), the mark implemented by PRC2, was affected. Genome-wide profiling of Jarid2, Su(z)12 (Suppressor of zeste 12), and H3K27me3 occupancy by chromatin immunoprecipitation with sequencing (ChIP-seq) indicates that Jarid2 and Su(z)12 have very similar distribution patterns on chromatin. However, Jarid2 and Su(z)12 occupancy levels at some genes are significantly different, with Jarid2 being present at relatively low levels at many Pc response elements (PREs) of certain Homeobox (Hox) genes, providing a rationale for why Jarid2 was never identified in Pc screens. Gene expression analyses show that Jarid2 and E(z) (Enhancer of zeste, a canonical PRC2 component) are not only required for transcriptional repression but might also function in active transcription. Identification of Jarid2 as a conserved PRC2 interactor in flies provides an opportunity to begin to probe some of its novel functions in Drosophila development.

Fig 1

Purification of an endogenous Jarid2 complex in Drosophila. (A) Colocalization of Jarid2 and FLAG-HA-Jarid2 on polytene chromosomes from third-instar larvae showing overlap of endogenous Jarid2 with FLAG-HA-Jarid2. (A) Merged image of Jarid2 and HA antibody channels. (A′) Jarid2 antibody labeling. (A″) HA antibody labeling. (A‴) DAPI (4′,6′-diamidino-2-phenylindole) staining. (B) Silver staining of FLAG immunopurification of Jarid2. Lane 1, wild-type control FLAG purification; lane 2, Jarid2 FLAG purification. A degraded product of FLAG-HA-Jarid2 is marked with an asterisk. (C) Western blot analysis from nuclear embryonic extracts of wild-type control and FLAG-Jarid2 before and after FLAG purification. Lane 1, nuclear embryonic extract of wild-type control labeled with anti-HA (upper panel) and anti-E(z) (lower panel); lane 2, nuclear embryonic extract from the FLAG-HA-Jarid2 line labeled with anti-HA (upper panel) and anti-E(z) (lower panel); lane 3, Eluate of wild-type control labeled with anti-HA (upper panel) and anti-E(z) (lower panel); lane 4, eluate of FLAG-HA-Jarid2 line labeled with anti-HA (upper panel) and anti-E(z) (lower panel). (D) MudPIT analysis of the Jarid2 FLAG purification. All major components of PRC2, including E(z), Su(z)12, Esc, and Caf1 along with Jing, the Drosophila homolog of human AEBP2, could be copurified. The relative abundance of the purified components is depicted by distributed normalized spectrum abundance factor (dNSAF) values on the y axes. Components of the wild-type control are blotted in red, and components of the Jarid2 purification are shown in blue. (E) Colocalization of E(z) and FLAG-HA-Jarid2 on polytene chromosomes displaying only partial overlap between E(z) and Jarid2. (E) Merged image of E(z) and HA antibody channels. (E′) E(z) antibody labeling. (E″) HA antibody labeling. (E‴) DAPI staining.

Fig 2

Jarid2 modulates H3K27 methylation in vivo. (A to D) GAL4-driven overexpression of Jarid2 in the posterior part of the wing imaginal disc under the control of the engrailed (en) promoter. GFP expression in green marks the posterior part (highlighted by an arrow in the antibody-only channel) where Jarid2 is expressed (A, B, and D). Jarid2 overexpression is confirmed with an antibody against Jarid2 (A′, arrow). Global reduction of H3K27me3 (B′, arrow) is observed when Jarid2 is overexpressed, but no changes in H3K27me3 (C′, arrow) can be detected in wild-type controls. (D′) H3K27me1 levels are not affected under Jarid2-overexpressing conditions (arrow). (E) Flipase-catalyzed induction of Jarid2 mutant clones (no GFP expression) with the eye-specific eyeless (ey) promoter. Wild-type tissue is marked in green (GFP expression). (E′) Jarid2 mutant clones display increased levels of H3K27me3 (two representative clones are outlined and highlighted by arrows). (F) Overexpression of E(z)-RNAi in the posterior part of the wing imaginal disc under the control of the en promoter. GFP expression in green marks the posterior part where E(z)-RNAi is expressed. (F′) H3K27me3 is strongly reduced in the posterior compartment (white arrow). Genotypes: en-GAL4 UAS-GFP; UAS-Jarid2 (Jarid2^LA00681) (A to D), ey-FLP; Jarid2^MB00996 FRT80B/ubi-GFP FRT80B (E), and en-GAL4 UAS-GFP; UAS-E(z)-RNAi (F).

Fig 3

The role of Jarid2 in recruitment of Polycomb repressive complexes 2 and 1 and modulation of H3K27me3 on chromatin. (A to L) Squashes of polytene chromosomes from salivary glands of wild-type, Jarid2 mutant, Jarid2-overexpressing (UAS-Jarid2), and E(z) mutant third-instar larvae labeled with various antibodies. Panels show individual and merged images of antibody and DAPI channels. Panels A′ to L′ reflect the antibody-only channel, and panels A′ to D′ show results of Su(z)12 antibody labeling. (A′) Su(z)12 binding pattern in wild-type controls. Su(z)12 localization is not changed in Jarid2 mutants (B′) but is lowered in Jarid2-overexpressing (C′) and E(z) mutant (D′) animals. (E′ to H′) H3K27me3 antibody labeling. (E′) H3K27me3 binding pattern in wild-type controls. H3K27me3 is not visibly changed in Jarid2 mutants (F′). A reduction of H3K27me3 can be observed when Jarid2 is overexpressed (G′) or E(z) is removed (H′). (I′ to L′) Pc antibody labeling. (I′) Pc binding pattern in wild-type controls. No effects on Pc localization can be observed in either the Jarid2 mutants (J′) or under Jarid2-overexpressing (K′) or E(z) mutant (L′) conditions. Genotypes: Oregon R (A, E, and I), Jarid2^MB00996 (B, F, and J), act-GAL4/UAS-Jarid2 (Jarid2^LA00681) (C, G, and K), and E(z) TS [E(z)⁶¹, temperature shifted to 29°C for 24 h] (D, H, and L).

Fig 4

Genome-wide localization patterns by ChIP-seq of H3K27me3, Su(z)12, and Jarid2 in wild-type and Jarid2 mutant tissue. (A) Pie charts showing the genome-wide distribution of Jarid2 and Su(z)12 peaks in wild-type larvae. Gene localizations are represented as indicated on the figure. (B) The scatter plot shows the correlation of Jarid2 occupancy and Su(z)12 occupancy for each transcript from Ensembl, version 63, measured as the maximum normalized fragment count within 50 bp of the TSS. Highlighted in red are transcripts that fall within the lower quartile of Jarid2 occupancy (≤5.41) and the upper quartile of Su(z)12 occupancy (≥10.43) for Jarid2- and Su(z)12-bound transcripts, respectively. (C) The Bithorax complex on chromosome 3R showing one replicate of the wild type and Jarid2^MB0096 mutant from each chromatin immunoprecipitation as indicated. Significant changes in Su(z)12 binding between Jarid2^MB00996 mutants and the wild type are indicated by arrows but do not appear to significantly affect H3K27me3 levels within the region.

Fig 5

PRC2-dependent function of Jarid2 in active transcription and repression. (A) MA plot of the gene expression analysis in Jarid2^MB00996 mutant larvae compared to the wild type. The y axis shows the fold change of expression levels of the Jarid2^MB00996 mutant divided by wild type in log₂ scale; the x axis shows the average expression level (Affymetrix probe intensity) for each probe. Colored dots correspond to genes bound by Jarid2/Su(z)12 in the wild type and which display significant transcriptional changes in Jarid^MB00996 mutants. Genes that are cobound by Jarid2/Su(z)12 and are at least 1.5-fold upregulated in the Jarid2^MB00996 mutant are highlighted in blue; cobound genes with a greater than 1.5-fold loss are shown in red. (B) MA plot, similar to that shown in panel A, comparing E(z)-RNAi and wild-type gene expression in larvae. Colored dots from panel A are now shown on the E(z)-RNAi MA plot. (C and D) Individual examples from panels A and B that are coregulated and cobound by Jarid2 and PRC2. Displayed are fold changes in log₂ scale from the gene expression analyses in panels A and B. (C) Genes upregulated in Jarid2^MB00996 mutant and E(z)-RNAi larvae which are cobound by Jarid2/Su(z)12 in wild-type larvae. (D) Genes downregulated in Jarid2^MB00996 mutant and E(z)-RNAi larvae which are cobound by Jarid2/Su(z)12 in wild-type larvae. (E) All transcripts of genes upregulated (blue) or downregulated (red) in both Jarid2^MB00996 mutant and E(z)-RNAi larvae which are cobound by Jarid2/Su(z)12 in wild-type larvae plotted according to wild-type (x axis) and Jarid2 mutant (y axis) Su(z)12 enrichment levels. (F) All transcripts of genes upregulated (blue) or downregulated (red) in both Jarid2^MB00996 mutants and E(z)-RNAi larvae which are cobound by Jarid2/Su(z)12 in wild-type larvae plotted according to wild-type (x axis) and Jarid2 mutant (y axis) H3K27me3 enrichment levels. WCE, whole-cell extract.