Figure 1. H3K9me3 is insufficient to define the large diversity of genomic Rhino domains.

(A) ChIP-seq enrichment (genome unique reads; 1-kb tiles, one representative replicate each) of H3K9me3 and Su(var)2–5 along the assembled chromosome 3 sequence in w¹¹¹⁸ ovaries. Pericentromeric heterochromatin and euchromatic chromosome arms are indicated. (B) Scatter plot comparing average log2 ChIP-seq enrichments for Rhino in ovaries from w¹¹¹⁸ (n=2) versus MTD-Gal4 >w-sh (n=3) strains (1-kb tiles separated into pericentromeric heterochromatin and chromosomal arms; piRNA clusters 38C, 42AB, and 80F are shown separately; colored 1-kb tiles correspond to example loci in panel C). (C) UCSC genome browser tracks depicting the diversity of Rhino domains. 1: heterochromatic, transposon-rich locus; 2: piRNA cluster 80F; 3: strain-specific flea insertion (ElMaghraby et al., 2019) in w¹¹¹⁸; 4: Rhino domain proximal to euchromatic headcase locus. Unless indicated otherwise, data are from w¹¹¹⁸ovaries (ChIP and RIP signal are depicted as coverage per million reads, piRNA coverage normalized to miRNA reads, data is displayed for one representative replicate). GFP-RIP-seq serves as control for non-specific mRNA binding. (D, E) Violin plots showing average log2 fold enrichment of Rhino ChIP-seq over input for 1-kb tiles (n=2) from w¹¹¹⁸ ovaries. Tiles were grouped into Rhino-bound and non-Rhino-bound based on a cutoff of fourfold enrichment (corresponding to p=0.036, Z-score=2.1) of Rhino ChIP-seq signal over input in each replicate experiment. Rhino-dependent piRNA clusters 38C, 42AB, and 80F were analyzed separately (shown as box plots due to low number of tiles). Box plots show median (center line), with interquartile range (box) and whiskers indicate 1.5x interquartile range. (F) UCSC genome browser tracks depicting a pericentromeric heterochromatin locus marked by H3K9me3 and bound by Su(var)2–5, but not Rhino (ChIP-seq signal: coverage per million reads; piRNA coverage normalized to miRNA reads; all data obtained from w¹¹¹⁸ ovaries, data is displayed for one representative replicate).

Figure 1—figure supplement 1. Relationships between methylated H3K9 and chromatin target sites of HP1 family members.

(A) Scatter plots contrasting average log2 ChIP-seq enrichment for Rhino in w¹¹¹⁸ (n=2) or MTD-Gal4 (n=3) versus iso1 (n=2) ovaries. Depicted are 1-kb tiles separated into pericentromeric heterochromatin and chromosomal arms. piRNA clusters 38C, 42AB, and 80F are excluded from the categorization into heterochromatin and euchromatin and are depicted separately. Indicated colored 1-kb tiles correspond to example loci in Figure 1F. (B, C) Violin plots showing the average log2-fold ChIP-seq enrichment of endogenous and germline specific Su(var)2–5 (transgene expressed under the control of the rhino promoter) and H3K9me2 (C) over input on 1-kb tiles. Tiles are grouped into Rhino-bound and non-Rhino-bound analogous to Figure 1D and E with Rhino-dependent piRNA clusters 38C, 42AB, and 80F depicted separately for comparison. Box plots show median (center line), with interquartile range (box) and whiskers indicate 1.5x interquartile range. (D) Heat map centered on HP1b (left) or HP1c (right) peaks, respectively, depicting the indicated ChIP-seq signal (n=1) in ovaries from the w¹¹¹⁸ strain.