Protein-synthesis-independent heterochromatin disruption following STAT92E phosphorylation. (a) Wild-type salivary glands were cultured with or without cycloheximide (CHX) for 1 h before addition of pervanadate for 1 h in vitro. Represented nuclei are shown. Note the disappearance of the HP1 foci in treated cells. (b) S2 cells were cultured with or without CHX for 1 h, and then stimulated with pervanadate for the indicated times. Cell extracts were blotted with the indicated antibodies. Note a greater increase in pSTAT92E (phosphorylated) and decrease in STAT92E (unphosphorylated) bands in CHX-treated samples, which is correlated with a greater decrease in H3mK9 levels. (c) A model for the role of STAT in heterochromatin stability. Unphosphorylated STAT is localized on heterochromatin in association with HP1. Increasing STAT phosphorylation (by JAK or other tyrosine kinases) reduces the amount of unphosphorylated STAT localized in heterochromatin. This in turn causes HP1 displacement from heterochromatin and heterochromatin instability. Dispersed phosphorylated STAT binds to cognitive sites in euchromatin to induce target-gene expression. Scale bars, 2 μm.

Only unphosphorylated STAT92E is localized on heterochromatin. (a) Wild-type larval salivary gland stained with anti-HP1 and anti-pSTAT92E antibodies. Note that the signals do not overlap. (b) In polytene chromosomes from sgs–Gal4<UAS–stat92E–GFP larvae, STAT92E–GFP was colocalized with either pSTAT92E (bars) or HP1 (arrowheads) signals. (c) Third-instar larval salivary glands expressing UAS–STAT92E–GFP or STAT92E^Y704F–GFP (by arm–Gal4) were treated with pervanadate ex vivo. Distribution of HP1 (red) and STAT92E–GFP or STAT92EY704F–GFP (green) was examined at the indicated time points. Note that following pervanadate stimulation, STAT92E–GFP moves away from heterochromatin (HP1 foci; 20 min) and then binds to chromosomes as distinct bands (60 min), whereas UAS–STAT92E^Y704F–GFP remains colocalized with HP1 foci (60 min). DNA was stained with Toto-3. (d) Chromatin was immunoprecipitated from S2 cells with an anti-STAT92E antibody (which recognizes both unphosphorylated and phosphorylated STAT92E) before and after pervanadate treatment. Note the enrichment of STAT92E on heterochromatin (1360; row 1) and target-gene promoter (stat92E itself) before (lane 1) and after (lane 2) pervanadate treatment. The stat92E promoter contains multiple STAT92E-binding sites that mediate stat92E positive-feedback autoregulation. Rp49 is a negative control (row 3). A goat anti-human pan SMAD antibody was used as a control IgG for goat anti-STAT92E. Scale bars, 2 μm.

STAT92E levels control heterochromatin abundance and HP1 localization. (a, b) Third-instar salivary glands overexpressing stat92E⁺ in random cells marked with GFP (green) and stained with anti-H3mK9 (magenta in a) or anti-HP1 (magenta in b) are shown partially. Note the increased levels of H3mK9 or HP1 in GFP⁺ cells, compared with GFP⁻ cells (four GFP⁻ cells are outlined). (c) A third-instar salivary gland containing stat92E^−/− cells marked with GFP (green) and stained with anti-HP1 (red) is shown partially. Note the lack of HP1 foci in the GFP⁺ cells. (d) Enrichment of HP1 on heterochromatin was detected by ChIP using an anti-HP1 antibody and PCR primers that amplify the 1360 transposon element (heterochromatin). Note the marked reduction in HP1 association with 1360 sequences in the STAT92E RNAi-depleted sample (lane 2), compared with the control (lane 1). (e, f) Cellular blastoderm-stage embryos stained with anti-HP1 (green) and propidium iodide (PI; red) to show DNA. Part of the cortical cell layer is shown. Note the concentrated HP1-staining at the apical region of the wild-type embryo (arrow in e), which is absent in the stat92E^mat− embryo (f). (g) Total protein extracts from 0–12 hold embryos of wild-type (lane 1), Nanos–Gal4 × UAS–stat92E⁺ (lane 2) and stat92E^mat− females (lane 3) were subjected to SDS–PAGE and blotted sequentially with anti-STAT92E and anti-α-tubulin antibodies (for full-length gel, see Supplementary Information, Fig. S7), or with anti-H3mK9 and anti-HP1 antibodies, respectively. The membrane was stripped between blots. Note the correlation between H3mK9 and STAT92E levels. Scale bars, 10 μm.

STAT92E colocalizes and physically associates with HP1. (a–c) Third-instar wild-type salivary glands were stained with anti-HP1 (red) and anti-STAT92E (green) antibodies in whole-mount tissues or squashed polytene chromosomes. Representative giant nuclei and squashed chromosomes are shown. (a) In wild-type larvae, the HP1 foci in the salivary gland nuclei, which mark heterochromatin, colocalize with STAT92E. In squashed chromosomes, STAT92E and HP1 colocalization is shown in the telomere (b) and chromocentre (c). (d) Drosophila embryo extracts were incubated with bacterially expressed GST–HP1, which was then purified by glutathione beads, subjected to SDS–PAGE and blotted with an anti-STAT92E antibody. (e) Drosophila S2 cells transfected with or without STAT92E–V5 variants or His–HP1 were immunoprecipitated with an anti-V5 antibody and subjected to SDS–PAGE. The associated His–HP1 was detected with an anti-His antibody. Note the double-mutant STAT92E–V5 (lane 7) was unable to bind to HP1. The HP1-binding consensus motif, the two sequences present in STAT92E and the mutated sequences are shown under the blots. (f) Embryo extracts were immunoprecipitated with an anti-STAT92E antibody, subjected to SDS–PAGE and blotted with anti-HP1. Note that a greater amount of the endogenous HP1 was co-immunoprecipitated with endogenous STAT92E from wild-type than hop^Tum–l/+ embryos. Also note that depleting STAT92E using an antibody did not seem to affect HP1 levels in the lysate (lanes 2, 6), suggesting that perhaps only a small portion of HP1 is normally associated with STAT92E. (g) S2 cells transfected with or without STAT92E–V5 or Hop (lane 3) were immunoprecipitated with anti-V5 and subjected to SDS–PAGE. The associated endogenous HP1 was detected with an anti-HP1 antibody. Note that HP1 is co-immunoprecipitated with STAT92E–V5 (lane 2) and a reduced amount of HP1 is detected in the presence of Hop (lane 3). Scale bars, 2 μm.

STAT92E is required for heterochromatic gene silencing. (a–c) Effects of mutations in hop and stat92E on white⁺ variegation induced by transgene repeats, DX1 (a) or a control transgene (b) are shown as changes in red eye-pigmentation in representative images. DX1 consists of seven tandem copies of a P[white⁺] reporter transgene inserted in a euchromatic region, which induce heterochromatin formation at the insertion locus²⁶. In(1)w^m4 is caused by X chromosomal inversion, juxtaposing the white⁺ gene to centromeric heterochromatin¹². (c) Red eye-pigment levels were determined by measuring absorbance at a wavelength of 480 nm. Data are mean ± s. d, **P < 0.001 when compared with +/+ control, Student’s t-test (n = 40 for each group). 6–2 is a single P[white⁺] element inserted at the same chromosomal location as DX1 but that does not induce heterochromatin formation^26,27. (d–k) Effects of altering dosages of stat92E⁺ and/or HP1⁺ on white⁺ variegation induced by chromosomal inversion (In(1)w^m4) are shown as changes in red eye-pigmentation in representative images. (d) In(1)w^m4 in otherwise wild-type background. Note increased white⁺ expression (more red pigments) in stat92E (e, h, k) or HP1 heterozygotes (g–i) and decreased white⁺ expression when stat92E⁺ (f, i) or HP1⁺ is overexpressed (j, k). The combination of hsp70–HP1⁺ and 3 × stat92E⁺ was not viable. Flies bearing hsp70–HP1⁺ were raised at 29°C. All others were raised at 25°C. All genotypes are in In(1)^wm4/+ or w^−/− backgrounds. (hop^GOF=hop^Tum–l, hop⁻=hop³, stat92E⁻=stat92E⁶³⁴⁶, HP1⁻=Su(var)205⁰⁵, 3 × stat92E⁺=Dp(3;3)MRS/+). Similar results were observed when another chromosomal duplication, Dp(3;3)M95A[+]13, that duplicates stat92E⁺ was used (data not shown), or by a stat92E⁺ transgene (Supplementary Information, Fig. S1).